[nato science for peace and security series b: physics and biophysics] terahertz and mid infrared...

AB 3

NATO Science for Peace and Security Series - B: Physics and Biophysics

Terahertz and Mid Infrared Radiation: Detection of

Explosives and CBRN (Using Terahertz)

Edited byMauro F. PereiraOleksiy Shulika

Terahertz and Mid Infrared Radiation: Detection of Explosives and CBRN (Using Terahertz)

NATO Science for Peace and Security Series

This Series presents the results of scientifi c meetings supported under the NATO Programme: Science for Peace and Security (SPS).

The NATO SPS Programme supports meetings in the following Key Priority areas: (1) Defence Against Terrorism; (2) Countering other Threats to Security and (3) NATO, Partner and Mediterranean Dialogue Country Priorities. The types of meeting supported are generally “ Advanced Study Institutes ” and “ Advanced Research Workshops.” The NATO SPS Series collects together the results of these meetings. The meetings are co-organized by scientists from NATO countries and scientists from NATO’s “ Partner ” or “ Mediterranean Dialogue ” countries. The observations and recommendations made at the meetings, as well as the contents of the volumes in the Series, refl ect those of participants and contributors only; they should not necessarily be regarded as refl ecting NATO views or policy.

Advanced Study Institutes (ASI) are high-level tutorial courses to convey the latest developments in a subject to an advanced-level audience

Advanced Research Workshops (ARW) are expert meetings where an intense but informal exchange of views at the frontiers of a subject aims at identifying directions for future action

Following a transformation of the programme in 2006 the Series has been re-named and re-organised. Recent volumes on topics not related to security, which result from meetings supported under the programme earlier, may be found in the NATO Science Series.

The Series is published by IOS Press, Amsterdam, and Springer, Dordrecht, in conjunction with the NATO Emerging Security Challenges Division.

Sub-Series

A. Chemistry and Biology SpringerB. Physics and Biophysics SpringerC. Environmental Security SpringerD. Information and Communication Security IOS PressE. Human and Societal Dynamics IOS Press

http://www.nato.int/sciencehttp://www.springer.comhttp://www.iospress.nl

Series B: Physics and Biophysics

Terahertz and Mid Infrared Radiation: Detection of Explosives and CBRN (Using Terahertz)

edited by

Mauro F. Pereira Materials and Engineering Research Institute

Sheffield Hallam University

Sheffield, UK

and

Oleksiy Shulika Universidad de Guanajuato

Salamanca, Guanajuato , Mexico

Published in Cooperation with NATO Emerging Security Challenges Division

Proceedings of the NATO Advanced Research Workshop on

Detection of Explosives and CBRN (Using Terahertz)

Cesme, Izmir, Turkey

3–6 November 2012

Library of Congress Control Number: 2014933670

ISBN 978-94-017-8583-9 (PB)

ISBN 978-94-017-8571-6 (HB)

ISBN 978-94-017-8572-3 (eBook)

DOI 10.1007/978-94-017-8572-3

Published by Springer,

P.O. Box 17, 3300 AA Dordrecht, The Netherlands.

www.springer.com

Printed on acid-free paper

All Rights Reserved

© Springer Science+Business Media Dordrecht 2014

This work is subject to copyright. All rights are reserved by the Publisher, whether

the whole or part of the material is concerned, specifi cally the rights of translation,

reprinting, reuse of illustrations, recitation, broadcasting, reproduction on microfi lms or

in any other physical way, and transmission or information storage and retrieval,

electronic adaptation, computer software, or by similar or dissimilar methodology now

known or hereafter developed. Exempted from this legal reservation are brief excerpts

in connection with reviews or scholarly analysis or material supplied specifi cally for the

purpose of being entered and executed on a computer system, for exclusive use by

the purchaser of the work. Duplication of this publication or parts thereof is permitted

only under the provisions of the Copyright Law of the Publisher’s location, in its current

version, and permission for use must always be obtained from Springer. Permissions

for use may be obtained through RightsLink at the Copyright Clearance Center.

Violations are liable to prosecution under the respective Copyright Law.

The use of general descriptive names, registered names, trademarks, service

marks, etc. in this publication does not imply, even in the absence of a specifi c

statement, that such names are exempt from the relevant protective laws and

regulations and therefore free for general use.

While the advice and information in this book are believed to be true and accurate

at the date of publication, neither the authors nor the editors nor the publisher

can accept any legal responsibility for any errors or omissions that may be made.

The publisher makes no warranty, express or implied, with respect to the material

contained herein.

v

Foreword

Many substances exhibit rotational and vibrational transitions in the terahertz (THz- 0.3 THz up to 10 THz) and mid infrared (MIR-15 THz to 120 THz), jointly called here “TERA-MIR”, hence giving access to a spectroscopic analysis of a large variety of molecules which play a key role in security as well as various other areas, e.g. air pollution, climate research, industrial process control, agriculture, food industry, workplace safety and medical diagnostics can be monitored by sensing and identifying them via MIR and THz absorption “fi nger prints”. Most plastics, textiles and paper are nearly transparent for THz radiation. Therefore, CBRN agents, explosives or illegal drugs can be detected by their characteristic absorp-tion spectra at THz frequencies with high selectivity and resolution in application fi elds as industrial quality inspection control, customs inspection and security screening. Moreover, MIR and THz radiation has no endangering effects on human beings and enables higher contrast for “soft matter” than X-rays. In comparison to standard optical technologies for wavelengths up to about 2 μm, sources and detec-tors for MIR and THz have not yet reached this level of maturity and there is still a large gap for features like wavelength tunability, spectral purity, high power and room temperature operation, which all are necessary for commercial applications.

Plastics or ceramics are detected by X-rays very poorly especially against a background of human body. Unlike X-rays, THz (or T-wave) is not a dangerous radiation, and in some cases T-wave sensors can reveal not only the shape of a hidden object but also its chemical composition. This unique combination of traits make T-waves perfect for effective applications like explosive detection and security applications. Besides, Т-rays have high resolution in 3D space in case of THz ultrashort pulses. The possibility to analyze chemical composition of sub-stances by spectroscopic methods is of big interest. Even in case the substance is in the plastic tank or under the cloth. However, there are many open problems on the path to practical and routine use of THz.

The NATO Advanced Research Workshop on Detection of Explosives and CBRN (Using Terahertz), which took place in Cesme, Izmir, Turkey, 3–6 November 2012, delivered a timely update of new THz sources and detection schemes as well

vi

as concrete applications to the detection of explosives and CBRN. The workshop characteristic feature was a stronger emphasis on the mathematical and physical aspects of the research, together with a detailed analysis of the application problems. The next paragraphs summarize some of the main results presented in the workshop.

Among direct application methods and devices, we can highlight that it was shown that both passive and active systems can be used for the detection of concealed objects at stand-off distances, however that it was diffi cult to identify the material, such as an explosive, at these distances behind barrier materials. Typically passive systems work at certain frequency bands where the background emissivity and refl ectivity in regard to the object of interest is quite different allowing for better image contrast. On the other hand, active systems eliminate the background and can detect the object and identify at the cost of an increase in system complexity. In such an active system, improvements in receiver design, read-out noise and overall system integration (amplifi ers, waveguides, etc.) can better active system based imaging technologies that use off-the-shelf components. These issues were discussed in the context of developing an active imaging system working at frequencies near the F-band and above 300GHz based on Schottky-diode multipliers and heterodyne receivers.

A method which can identify hidden RDX-based explosives (pure and plastic ones) has been presented. The method takes into account only part of an impulse refl ected from the sample, which is analyzed in frequency domain by Fourier Transform. This has been complemented by the demonstration of improvement of the quality of the images captured with commercially available THz passive cam-eras. The approach is on application of novel spatial fi lters and algorithms, devel-oped for computer processing of passive images produced by the THz camera. The presented examples show the big potential for the detection of small hidden objects from long distances (6–10 m) and observation of the difference in temperature on the human body, which is caused by different temperatures inside the body.

A spectrometer under development that led to detection of explosive substances (NG, TNT, RDX, etc.) in sub-THz, based on registration of certain marker- molecules that evaporated from surfaces of samples, has been presented, complementing the usual THz range. Experimental results have demonstrated a presence of nitric oxides, ammonia, acetone, etc. in explosives vapors. Furthermore, a new direction has been shown with the demonstration that transmission pulsed imaging utilizing this source and a Schottky-detector. It was particularly interesting to see that images utilizing propagation delay are much more clear and informative than those based on sub-THz wave attenuation.

Complementing the results in the short-wavelength range, laser spectroscopy with mid infrared, room temperature, continuous wave, DFB laser diodes and high-performance DFB QCLs have been demonstrated to be a relevant tool for the development of sensor technologies for environmental monitoring, medical diagnostics, and industrial and security applications.

From the new source point of view, a number of systems have been presented – from superconductors to semiconductors, e.g. detection of terahertz waves from

Foreword

vii

superconducting Bi 2 Sr 2 CaCu 2 O 8 + δ intrinsic Josephson junctions. The quest for a compact, room temperature THz source and the recent advances in high-power mid- IR QCLs led to the development of a semiconductor THz source based on intracavity difference frequency generation. With this technique, 65 μW output power at room temperature at a frequency of 4 THz has been achieved. Lithographical tuning is also demonstrated from 1 to 4.6 THz. These demonstrations pave the way for compact, reliable and cost effective solutions for detection of CBRN agents, explosives and illegal drugs. The mJ-level ultrashort THz pulses by optical rectifi ca-tion have been demonstrated and proposed as source for investigation of different materials. Coherent imaging hybrid system based on a THz quantum cascade laser (QCL) phase-locked to a near-IR fs-laser comb has been demonstrated, leading to raster scan coherent imaging using a QCL emitting at 2.5 THz. At this frequency, the detection noise fl oor of our system is of 1 pW/Hz. THz emitters based on strong light-matter interaction have been discussed. As the struggle to achieve room temperature (or thermoelectric cooler compatible temperature) operation of THz Quantum Cascade Lasers continues, alternative electrically pumped THz sources are worth investigating. One way relies on the high emission effi ciency predicted for polaritonic states in the ultra-strong coupling regime. Electroluminescent devices in this regime were demonstrated. Furthermore, antipolaritons in dispersive media were discussed and different aspects of the interaction of THz radiation with biomatter were presented. A fundamental discussion of free carrier absorption for THz radiation in heterostructures has been given. Altogether, the presentations and discussions provided during the workshop in the frontier of terahertz (THZ) and mid infrared (MIR) basic science and applications can potentially stimulate joint research and projects for designing new materials and devices.

In summary, this meeting allowed the attendees to get a global picture of the state of the art in TERA-MIR generation, detection and applications. We had an excellent opportunity to discuss further proposal possibilities and we have high hopes that a few meaningful collaboration projects will be submitted after this meeting.

Sheffi eld, UK Mauro F. Pereira Salamanca, GTO, Mexico Oleksiy Shulika March, 2013

Foreword

ix

Acknowledgements

We start our acknowledgement list with our thanks to NATO and the Science for Peace project for the very generous fi nancial support and continuous support with all necessary details which made the realization of The NATO Advanced Research Workshop on Detection of Explosives and CBRN (Using Terahertz) and this book possible. The co-chairs Mauro Pereira and Igor Sukhoivanov are grateful to NATO, which provided a fantastic opportunity for all of us to meet in Cesme, Izmir. Joint proposals and scientifi c collaborations are already evolving thanks to this opportu-nity, and we hope that they will have an impact in the development of research in the fascinating TERA-MIR range.

We further acknowledge COST ACTION MP1204 for the logistic and organiza-tional support delivered and Lutfi Ozyuzer, Management Committee member of MP1204 representing Turkey and his team for their local support.

We thank staff and administration of Altinus Yunus Hotel in Cesme, which extended every courtesy to the attendees and gave us an opportunity to meet in a stunning location by the sea. They did everything in their power to help us with all logistic issues related to bringing people from all over the world to this meeting and helped create the perfect atmosphere for this meeting.

Our big thanks to Chris Hughes and the fi nance team at Sheffi eld Hallam University who helped us go through the many fi nancial details needed to organize this meeting and for taking care of travel for people scattered around the world very effi ciently.

The other committee members played a major role in helping us in selecting the speakers and reaching a fi nal program conclusion, so here is a statement of our appreciation for support given by Igor Sukhoivanov, Romuald Brazis, Guido Giuliani, Martin Koch, Marian Marciniak, and Ekaterina Orlova.

In preparing this book we have relied on the timely contribution of the authors. Without their expert insight, motivation and commitment, the publication of this volume would not have been possible. We, thus, extend our appreciation to all the authors. We also convey our thanks to Springer for the opportunity of publishing this volume.

xi

Contents

1 THz Hybrid Metamaterial-Liquid Crystal Based Structures with Large Tunability ........................................................ 1 N. Chikhi , M. Lisitskiy , G. Molis , A. Urbanovic , and A. Andreone

2 High-Resolution THz Spectroscopy of Biomolecules and Bioparticles: Concentration Methods .......................................... 7 E. R. Brown , W. Zhang , L. K. Viveros , E. A. Mendoza , Y. Kuznetsova , S. R. J. Brueck , K. P. Burris , R. J. Millwood , and C. N. Stewart

3 Intervalence THz Antipolaritons ......................................................... 19 I. A. Faragai and M. F. Pereira Jr.

4 Terahertz Aperiodic Multilayered Structure Arranged According to the Kolakoski Sequence ............................... 25 Volodymyr I. Fesenko , Vladimir R. Tuz , and Igor A. Sukhoivanov

5 Population Dynamic in Quantum Cascade GaN/AlGaN Photodetector Structure ................................................. 33 S. V. Gryshchenko , Oleksiy Shulika , V. V. Lysak , and I. A. Sukhoivanov

6 Design of Metamaterial Photonic Crystals for Explosives Detection ....................................................................... 39 H. Hamdouni , F. Ouerghi , F. Abdelmalek , and H. Bouchriha

7 New Type High-Q THz Planar All-Dielectric Metamaterial ............. 47 Vyacheslav V. Khardikov and Sergey L. Prosvirnin

8 Bipolar THz-Lasing Structures Based on InAs-GaSb Coupled Quantum Wells and Their Potential for Security Checks .............................................................................. 53 L. D. Shvartsman and Boris Laikhtman

xii

9 Coherent Quantum Control of Donor States in Silicon with THz and MIR Light: A Route Towards a Scalable Quantum Computing Architecture ..................................................... 63 Stephen A. Lynch

10 THz Bio-chemical Sensing Capabilities with High Performance SIW Based Sensor on nL-Volume Liquids in Capillary .................... 75 V. Matvejev , J. Stiens , C. De Tandt , and D. Mangelings

11 A Theoretical Study on Monitoring Explosives Degradation by Pentaerythritol Tetranitrate Reductase Using THz Spectroscopy .................................................... 81 Maria Mernea and Dan Florin Mihailescu

12 Area Dependence of Josephson Critical Current Density in Superconducting Bi 2 Sr 2 CaCu 2 O8+d for Terahertz Emission ......................................................................... 87 H. Saglam , Y. Demirhan , K. Kadowaki , N. Miyakawa , and L. Ozyuzer

13 Inhomogenity of Micron-Sized Triple Terahertz Emitters Fabricated from Intrinsic Josephson Junctions in Single Crystal Bi2Sr2CaCu2O8+δ ...................................................................... 95 Yasemin Demirhan , F. Turkoglu , H. Koseoglu , H. Saglam , N. Miyakawa , K. Kadowaki , and L. Ozyuzer

14 The Fourier Transformed MIR Microspectroscopy to Reveal a Morphological and Spectral Markers of a Cervical Cancer Cells ................................................................... 103 A. A. Paiziev

15 THz Diffractive Optical Element for Passive Imaging ...................... 109 A. Czerwinski , P. Zagrajek , E. Rurka, N. Palka , M. Szustakowski , J. Suszek , A. Siemion , M. Makowski , and M. Sypek

16 Intersubband Dispersive Gain Media ................................................. 117 Mauro F. Pereira

17 Recent Advances in IR Laser Diodes with High Power, High WPE, Single Mode, CW Operation at RT ................................. 123 Manijeh Razeghi , Neelanjan Bandyopadhyay , Quanyong Lu , Yanbo Bai , Steven Slivken , and David Heydari

18 Characterization of Selenide, Sulfide and Telluride Materials by Terahertz Time-Domain Spectroscopy ......................... 129 R. M. Sardarly , F. Garet , M. Bernier , and J. -L. Coutaz

Contents

xiii

19 Ultrashort Electromagnetic Modes in the Low Frequency Region of the Spectrum in a Nanocylinder Array ............................. 135 L. Sirbu , V. Sergentu , R. Muller , V. Ursaki , and I. M. Tiginyanu

20 Influence of Mesa-Fabrication-Dependent Waveguide-Sidewall Roughness on Threshold Current and Slope Efficiency of AlGaAs/GaAs Mid-Infrared Quantum-Cascade Lasers .................................................................... 143 Anna Szerling , Kamil Kosiel , Piotr Karbownik , Anna Wójcik-Jedlińska , and Mariusz Płuska

21 Mid-infrared Laser Based Gas Sensor Technologies for Environmental Monitoring, Medical Diagnostics, Industrial and Security Applications .................................................. 153 Frank K. Tittel , Rafał Lewicki , Mohammad Jahjah , Briana Foxworth , Yufei Ma , Lei Dong , Robert Griffi n , Karol Krzempek , Przemyslaw Stefanski , and Jan Tarka

22 Computer Processing of Images Captured with a Commercially Available THz Camera at Long Distances .............. 167 Vyacheslav A. Trofi mov , Vladislav V. Trofi mov , Norbert Palka , and Marcin Kowalski

23 Transmission Subterahertz Imaging Utilizing Milliwatt-Range Nanosecond Pulses from Miniature, Collapsing-Domain-Based Avalanche Source .................................................................................. 175 S. N. Vainshtein and J. T. Kostamovaara

24 Sub-THz Spectroscopy for Security Related Gas Detection ............. 189 V. Vaks , E. Domracheva , E. Sobakinskaya , and M. Chernyaeva

Contents

1M.F. Pereira and O. Shulika (eds.), Terahertz and Mid Infrared Radiation: Detection of Explosives and CBRN (Using Terahertz), NATO Science for Peace and Security Series B: Physics and Biophysics, DOI 10.1007/978-94-017-8572-3_1,© Springer Science+Business Media Dordrecht 2014

Abstract Large frequency tunability is achieved by combining a planar metamaterial- based device with a liquid crystal (LC) having a relatively high birefringence. The device is based on the exploitation of the LC molecule reorientation under an applied electric fi eld to change the permittivity of different capacitors present in the sub-wavelength unit cells. The whole system is designed to obtain a maximum signal frequency shift close to 10 % around the operational frequency of 1 THz.

Keywords Metamaterial • Liquid crystal • THz technology

1.1 Introduction

There are still restrictions limiting the full exploitation of fruitful applications covering the THz region. An increasing number of devices and systems are fabricated, and many applications have been developed in this frequency range. Amongst others, THz imaging [ 1 ], chemical and biological sensing [ 2 , 3 ], THz sources and detectors [ 4 ], semiconductor characterization [ 5 ], as well as homeland security systems [ 6 ]. Nevertheless, compared to the well established neighboring infrared and microwave

Chapter 1 THz Hybrid Metamaterial-Liquid Crystal Based Structures with Large Tunability

N. Chikhi , M. Lisitskiy , G. Molis , A. Urbanovic , and A. Andreone

N. Chikhi Department of Physics , University of Naples “Federico II” , Naples , Italy

M. Lisitskiy CNR-IC “E. Caianiello” , Pozzuoli , NA , Italy

G. Molis • A. Urbanovic TERAVIL Ltd , Vilnius , Lithuania

A. Andreone (*) CNR-SPIN and Department of Physics , University of Naples “Federico II” , Naples , Italy e-mail: [email protected]

2

regions, further improvements are required especially in terms of material properties in order to develop novel devices that can come over the so called “THz gap” [ 7 ].

The challenge to overcome those restrictions arises mainly from the diffi culty to fi nd naturally occurring materials that have a usable electronic response in this fre-quency range. The electromagnetic (EM) properties of natural media are dictated by their microscopic composition that typically varies over the entire spectrum, from the visible spectrum to radio frequencies. There is however a relatively new class of artifi cial materials called metamaterials [ 8 ] where the EM response is mainly func-tion of its macroscopic composition. The prefi x meta, which in Greek means beyond, is used to indicate the class of engineered materials that have properties that do not occur naturally, such as the ability to bend light the wrong way. Metamaterials are made by periodically arranging resonant sub-wavelength metallic inclusions called “meta-atoms”. The ability of meta-atoms to resonantly couple to the electric and magnetic components of the EM fi eld is at the basis of the intriguing concept of negative index media [ 9 , 10 ] that opened the doors to a new fi eld of modern optics.

Metamaterials are naturally fi ltering devices, and their electromagnetic response can be tailored to transmit, refl ect and absorb light in a narrow frequency band, hence they can be used as impedance matching layers, fi lters, absorbers, etc. Moreover, being their electromagnetic response scale invariant, metamaterials can be scaled down and used at any wavelength of interest from microwaves to optics. With an appropriate design such as Split Ring Resonators (SRR) [ 11 ], they can be therefore successfully exploited in the Terahertz range for the development of novel devices operating in this frequency region and therefore to fi ll the THz gap.

The area is now entering the stage of maturity where the basic exotic phenomena discovered in the early years are supporting the current research efforts toward more advanced results. A number of prominent potential applications can be realized with the proper exploitation of the ability to dynamically control the material properties or tune them in real time, through either direct external tuning or nonlinear response. Tunability in metamaterials has been already demonstrated using different mecha-nisms, such as MEMS [ 12 ], Schottky gate [ 13 ], photoexcitation [ 14 ]. We show here that strong tunability on a metamaterial embedded with liquid crystal is also possi-ble. Birefringence of liquid crystals in the THz band can be exploited to control the electromagnetic narrowband response of the metamaterial [ 15 ]. The tuning mecha-nism is based on the LC molecule reorientation under an applied voltage to change the permittivity of different capacitors present in the unit cells based on split ring resonators. The whole system is designed to obtain a maximum signal frequency shift close to 10 % around the operational frequency of 1 THz.

1.2 Design and Simulation Results

Here we present the results of computational studies on a planar hybrid tunable THz metamaterial based on split ring resonators, that can achieve a large frequency shift using liquid crystal as tuning mechanism.

We designed a THz planar structure operating at around 1 THz and composed of arrays of SRR unit cells covered using liquid crystal. Each unit cell consists of a

N. Chikhi et al.

3

normal square ring with bended edges of 40 μm lateral sides, 5 μm width, 0.2 μm thickness and a 7 μm gap in each side, connected to the adjacent ones using metallic wire needed for the LC polarization (see Fig. 1.1a ). The whole metamaterial is placed on lossy silicon substrate ɛ = 11.9, using a lattice parameter a = 50 μm.

The proposed tuning system is based on the birefringence of the liquid crystal. Its properties can be controlled reorienting the LC molecular director, described by the angle θ, in respect to the oscillating electric fi eld direction. Our idea was based on the creation of different capacitors over the ring gaps (as they cover the most sensitive area) and use of the LC to change the overall permittivity. Therefore, a suspended metallic cap was made in order to have cantilevers that overlap each side of the ring gaps, after the infi ltration of the LC. Each gap side will form a capacitor with cantilever on the top with LC as dielectric medium (see Fig. 1.1b ). In order to use the electric fi eld to polarize the LC, we designed metallic connections between the different part of the SRRs, and in this confi guration an ITO glass was put on the top of the structure to cover the LC and connect the caps to each other.

The tunability of the designed system is represented by the shift of the resonance frequency plotted in the S-parameter curve. Several simulations were performed using CST, a commercial electromagnetic code, in order to study the device response in the required frequency region. For those calculations, the LC under evaluation is represented as an anisotropic materials with an ordinary optical index n o = 1.62, an extraordinary optical index n e = 1.83, and thus a birefringence of 0.21. Therefore, in our simulations the LC layer permittivity is considered as [3.3489, 2.6244, 2.6244] and [2.6244, 3.3489, 2.6244] respectively for two orientations: θ = 0° and 90°. We also used a variable parameter s representing the difference between the ordinary and the extraordinary permittivity, so that the LC layer is treated as [ɛ x − s, ɛ x + s, ɛ z ]. The results are displayed in Fig. 1.2 and show that the metamaterials response shifts from f 1 = 1.04 THz to f 2 = 0.96 THz with Δf = 0.08 THz. This translates in a high frequency shift up to 8 % around the central frequency, compared to less than 4 % bandwidth.

Fig. 1.1 Hybrid metamaterial-liquid crystal based structure. ( a ) metamaterial array based on SRR unit cells with l = 40 μm, w = 5 μm, and g = 7 μm, with 3 μm connection wire; ( b ) plan cut of the unit cell where the gap is present, showing the LC interaction with the SRR

1 THz Hybrid Metamaterial-Liquid Crystal Based Structures with Large Tunability

4

1.3 Fabrication

The array of SRRs is fabricated by UV photolithography using the Al technology. The technological process is based on the deposition and patterning of two Al layers. The fi rst Al layer consists of the array of SRRs designed as shown in Fig. 1.1a . An effective array area is of 3 × 3 mm 2 . A 200 nm thick fi rst Al layer was deposited on the 1 × 1 cm 2 Si substrate by the dc sputtering magnetron technique and then patterned by a lift-off process. The suspended metallic caps (Fig. 1.1b ) was fabricated using sacrifi cial photoresist layer used as a support for the structure. A second 600 nm thick Al layer was deposited on the sacrifi cial photoresist, followed by the direct lithography process and Al wet etching to pattern the second Al layer in the form of suspended metallic caps. The device fabrication was completed after removing the sacrifi cial photoresist.

1.4 Experimental Results

Preliminary measurements have been carried out on the base layer structure (the SSR array), using conventional Time Domain Spectroscopy based on a fi ber laser operating at 1,050 nm and a low temperature GaAs emission/recording system. The frequency

0.6 0.8 1.0 1.2 1.4-40

-30

-20

-10Tr

ansm

ittan

ce S

21 (

dB)

Frequency (THz)

Without LC polarizationWith LC polarization

Fig. 1.2 Simulation results of the hybrid metamaterial-liquid crystal based structure. The trans-mittance S 21 response to the THz radiation when the LC is oriented on the θ = 0° direction is shown in the black curve . The red curve shows instead the metamaterial response when the LC is reoriented to θ = 90°. A red shift close to 10 % is observed in this confi guration

N. Chikhi et al.

5

dependent transmittance and refl ectance at normal incidence was characterized from 0.5 to 2.5 THz. Linearly polarized light was used with the THz electric fi eld perpendicular to the capacitive gaps. A THz beam was focused at the sample surface to a spot of diameter similar to the dimension of the metamaterial sample.

An LC resonance is observed in the spectrum at around 1.5 THz. No other reso-nances are observed up to 2.5 THz. The resonance frequency is higher than the value for the overall (quasi-three dimensional) full hybrid structure because of the absence of the vertical capacitive gaps. This is what is expected from the simulations on the electromagnetic response of the base layer only. Figure 1.3 shows that the experimen-tal data, even if noisy and lossy, nicely match the results of the simulation (red curve).

1.5 Conclusion

A planar metamaterial structure for modulation of the THz radiation has been designed. Its electromagnetic response has been numerically simulated. The struc-ture has been fabricated using Al technology and preliminary characterized in the frequency range 0.5–2.5 THz. The transmission experimental response shows a pronounced dip at around 1.5 THz, which is close to the intrinsic resonance frequency of the array of the split-ring resonators.

Next step will be the characterization of the full hybrid structure, namely the metamaterial SRR array with the insertion of liquid crystal. Simulations indicate that the full hybrid structure can potentially show a THz signal modulation depth up to 20 dB at the central frequency.

Fig. 1.3 Transmittance spectrum of the metamaterial structure. Experimental data ( blue curve ) are compared with the results of the simulation ( red curve )

1 THz Hybrid Metamaterial-Liquid Crystal Based Structures with Large Tunability

6

References

1. Hu BB, Nuss MC (1995) Imaging with terahertz waves. Opt Lett 20:1716–1718 2. Jacobsen RH, Mittleman DM, Nuss MC (1996) Chemical recognition of gases and gas

mixtures with terahertz waves. Opt Lett 21:2011–2013 3. Crowe TW, Globus T, Woolard DL, Hesler JL (2004) Terahertz sources and detectors and their

application to biological sensing. Philos Trans R Soc Lond A 362:365–377 4. Ferguson B, Zhang X-C (2002) Materials for terahertz science and technology. Nat Mater

1:26–33 5. Mittleman DM, Cunningham J, Nuss MC, Geva M (1997) Noncontact semiconductor wafer

characterization with the terahertz Hall effect. Appl Phys Lett 71:16–18 6. Zimdars D (2003) Fiber-pigtailed terahertz time domain spectroscopy instrumentation for

package inspection and security imaging. Proc SPIE 5070:108–116 7. Tonouchi M (2007) Cutting-edge terahertz technology. Nat Photon 1:97 8. Zheludev NI (2010) The road ahead for metamaterials. Science 328:5978 9. Veselago VG (1968) The electrodynamics of substances with simultaneously negative values

of ε and μ. Sov Phys USP 10:509–514 10. Smith DR, Padilla WJ, Vier DC, Nemat-Nasser SC, Schultz S (2000) Composite medium with

simultaneously negative permeability and permittivity. Phys Rev Lett 84:4184–4187 11. Pendry JB, Holden AJ, Robbins DJ, Stewart WJ (1999) Magnetism from conductors and

enhanced nonlinear phenomena. IEEE Trans Microw Theory Tech 47:2075 12. Hand T, Cummer S (2007) Characterization of tunable metamaterial elements using MEMS

switches. IEEE Antennas Wirel Propag Lett 6:401 13. Chen H-T, Padilla WJ, Zide JMO, Gossard AC, Taylor AJ, Averitt RD (2006) Active terahertz

metamaterial devices. Nature 444:597 14. Chen H-T, O’Hara JF, Azad AK, Taylor AJ, Averitt RD, Shrekenhamer DB, Padilla WJ (2008)

Experimental demonstration of frequency-agile terahertz metamaterials. Nat Photon 2:295 15. Chicki N, Di Gennaro E, Esposito E, Andreone A (2010) A study of tunable metamaterial

devices for the THz region. In: Terahertz and Mid-infrared radiation: generation, detection and applications, NATO science for peace and security series. B, physics and biophysics. Springer, Dordrecht, pp 9–13

N. Chikhi et al.


Abstract During the past several years we have utilized fl uidic-chip and waveguide- concentrator technology in combination with high-resolution frequency-domain THz spectroscopy to detect absorption signatures in biomolecules and bioparticles of vari-ous types, especially the nucleic acids and bacterial spores. Some of the signatures have been surprisingly narrow (<20 GHz FWHM), leading to the hypothesis that the fl uidic chips can enhance certain vibrational resonances because of their concentrating and linearizing effects. For solid or moist bio-samples, circular waveguide coupling allows signature detection of small quantities with some degradation of sensitivity but no loss of resolution. It concentrates the radiation, not the biomaterial. This method was used to demonstrate strong THz signatures in bacterial spores (e.g., Bacillus thuringiensis ).

Keywords THz radiation • THz coherent spectroscopy • Vibrational signatures • Nucleic-acid molecules • Lambda DNA • Nanofl uidic chips • Circular waveguide concentrator • Conical horns • Bacillus thuringiensis

Chapter 2 High-Resolution THz Spectroscopy of Biomolecules and Bioparticles: Concentration Methods

E. R. Brown , W. Zhang , L. K. Viveros , E. A. Mendoza , Y. Kuznetsova , S. R. J. Brueck , K. P. Burris , R. J. Millwood , and C. N. Stewart

E. R. Brown (*) Wright State University , Dayton , OH , USA

Physical Domains , LLC , Glendale , CA , USA e-mail: [email protected]

W. Zhang • L. K. Viveros Wright State University , Dayton , OH , USA

E. A. Mendoza Redondo Optics, Inc. , Redondo Beach , CA , USA

Y. Kuznetsova • S. R. J. Brueck Center for High Technology Materials and Department of Electrical and Computer Engineering, and Physics and Astronomy , University of New Mexico , Albuquerque , NM , USA

K. P. Burris • R. J. Millwood • C. N. Stewart Department of Plant Sciences , University of Tennessee , Knoxville , TN , USA

8

2.1 Background

THz radiation remains of great interest for biomolecular sensing because it interacts resonantly with collective vibrations that involve hundreds or more atomic constituents. They are therefore more specifi c to the primary structure of the biomolecules, compared to the interatomic vibrations sensed by traditional IR spectroscopy (e.g., FTIR). In addition, THz is inherently non-ionizing to biomole-cules, and like IR spectroscopy, can detect “label free” targets. Sub-THz vibrational modes in the nucleic acids DNA and RNA have been investigated for several decades going back to seminal work by Van Zandt [ 1 ] and Wittlin [ 2 ]. The former emphasized the existence of sub-THz polar optical phonon modes in aqueous DNA, while the latter emphasized an optically active hydrogen bond between the base pairs. Later, the study of localized vibrational modes became of interest, and in more recent work narrow signatures have been observed [ 3 – 4 ].

2.2 Introduction

2.2.1 Sample Presentation

More recently the importance of sample presentation and concentration has become better appreciated, leading to our investigation of new techniques for THz measure-ment of biomolecules. This paper reviews two such techniques, displayed schemati-cally in the system block diagram of Fig. 2.1 . The fi rst is a nanochannel, fl uidic-chip platform for the aqueous nucleic acid samples and proteins. The second technique concentrates the THz radiation into circular metal waveguide structures where vibrations can be excited effi ciently in sub-wavelength dry samples, such as pow-ders. The second technique concentrates the THz radiation into circular metal wave-guide structures where vibrations can be excited effi ciently in sub-wavelength dry-powder or grainy samples. In other words, the fi rst technique is concentrating the biomaterial, and the second technique is concentrating the THz electric fi eld.

2.2.2 High-Sensitivity THz Spectrometry

Our spectrometer of choice is the state-of-the-art coherent frequency-domain transceiver shown in Fig. 2.1 [ 5 – 7 ]. Because of spectral and spatial coherence, it is amenable to both sample-presentation techniques just described. Like time-domain spectroscopy, photomixing utilizes ultrafast photoconductive emitters and detectors but driven by single-frequency temperature-tunable distributed-feedback (DFB) semiconductor lasers rather than mode-locked lasers. It provides a continuously tunable coherent tone from below 100 GHz to ~2.0 THz with instantaneous linewidth

E.R. Brown et al.

9

of ~100 MHz or better [ 8 ]. The photomixers are fabricated at the center of an ultra-compact planar square-spiral antenna which radiates a primarily circular polarized beam above ~200 GHz. Including antenna impedance effects, the band-width of each photomixer is approximately 1.0 THz. One photomixer acts as the transmitter and the other acts as the receiver. The radiation from the transmit photo-mixer is coupled from the antenna to free space through a high-resistivity silicon hyperhemispherical lens. The THz beam is then collimated using an aspherical optic, usually an off-axis paraboloid. The reciprocal process occurs between free space and the receive photomixer.

Because the lasers driving receive and transmit photomixers are mutually coher-ent, the THz beam into the receive photomixer is mixed down in frequency by homodyne conversion. A simple amplitude modulation on the transmit photomixer then allows for dc offset and straightforward synchronous detection with all the benefi ts of traditional homodyne transceivers. With no samples in the THz path, this spectrometer produces a high dynamic range, typically in the range 70–80 dB at 100 GHz, and 50–60 dB at 1.0 THz. This range is achieved without evacuation of the free-space portion of the spectrometer so that water-vapor absorption lines do appear, especially when the ambient humidity is high. Unless one of these water

TransimpedanceLock-In Amp

+

SiliconLens

TransmitPhotomixer

Off-AxisParaboloid

Off-AxisParaboloid

Nanofluidic Cell (Top View)

Fixed DFBLaser ≈780 nm

TunableDFB Laser

>780 nm

Wavemeter

+

Isolator

FocusingLens

BeamCombiner

THzPath

THzPath

ReceivePhotomixer

NanofluidicChip.

THz circularpolarization

SiO2Channel

SiO2Wall

Conical-Horn

Coupler

Isolator

Fig. 2.1 Block diagram of THz coherent transceiver showing optimum location of the nanofl uidic chip adjacent to the transmit photomixer, and the conical-horn waveguide coupler at the middle of the THz path (only one method used at a time)

2 High-Resolution THz Spectroscopy of Biomolecules and Bioparticles…

10

vapor lines coincides with a signature from the sample-of-interest, this does not pose a problem. On the contrary, because the water vapor lines are highly resolved, their known shape and strength is often used to monitor the frequency metrology and dynamic range of the instrument.

The samples of interest are mounted in one of the three positions shown in Fig. 2.1 . The fl uidic chips are usually located at the lens port of the transmit or receive photomixer where the THz beam diameter is ≈3 mm, which underfi lls the nanochannel aperture of the chip. As shown in Fig. 2.2a , the chip is mounted on a precision dovetail slide for accurate and repeatable placement in the THz beam path. The circular-waveguide coupler is always located at the half-way point between the transmit and receive photomixers where the beam is collimated with a spot size of ≈1 cm. The insertion loss of the fl uidic chip is ≈2–3 dB, whereas the waveguide coupler loss is >20 dB. Nevertheless, the waveguide coupler transmits enough power to maintain signal strengths at least ~20 dB above the noise fl oor out to 1.0 THz. This takes advantage of the high-dynamic range of the instrument, which is typically 80 dB at 200 GHz, 60 dB at 1.0 THz, and 40 dB at 1.6 THz.

2.3 Sample Methods and Results

2.3.1 Fluidic Chips

The intent of the fl uidic chips is to concentrate and possibly linearize nucleic-acid molecules in nanochannel arrays, while maintaining effi cient coupling to free-space THz radiation. To fabricate the nanochannels at the micron scale and below, we have applied submicron-interference lithography and silica-nanoparticle calcination tech-niques [ 9 ]. In the present samples the channels were approximately 800 nm wide by

Fig. 2.2 ( a ) Fluidic chip lying on precision dovetail slide. ( b ) Close-up of fl uidic chip showing the transparent quartz substrate, metal strips (connected by electrodes), and gaps between the strips through which the THz propagates

E.R. Brown et al.

11

1,000 nm deep, on a pitch of ~1,200 nm, and separated from opposing reservoirs by about 5 mm. It takes ~15 min to fi ll the channels by capillary action [ 10 ], and the fi lling and concentrating effects have been confi rmed by fl uorescence imaging [ 10 ]. The THz transmission is then measured with the beam aligned at the center of the nanochannel array. The spatially-coherent, focused THz beam is typically 3 mm in diameter, so much smaller than the ~5 × 5-mm aperture of the fl uidic chip.

Historically, this fl uidic-chip platform quickly displayed strong signatures in the smaller nucleic-acid samples, such as those for small-interfering RNA shown in Fig. 2.3a [ 11 ]. This particular sample was a mixture of 17-, 21-, and 25-bp ds- siRNA molecules. The signatures are the strongest and narrowest (<20 GHz FWHM) we have ever observed in nucleic acid solutions. And three of them stood out in terms of absorption depth, labeled S1, S2, and S3 in Fig. 2.3a , and centered at 916, 962, and 1,034 GHz, respectively. In the meantime we have searched for signatures of larger molecules, starting with double-stranded DNA in the range between 50 and 1,000 bp. Most of these appeared similar to the transmission spectrum for 50-bp DNA shown in Fig. 2.3b . Three prominent sig-natures tended to appear between 850 and 1,000 GHz, the three in Fig. 2.3b being centered at ≈870, 938, and 993 GHz, and labeled in correspondence with those of si-RNA. Note however that the signatures in Fig. 2.3b are signifi cantly weaker and broader than those of Fig. 2.3a (which is why the vertical scale in Fig. 2.3a is logarithmic and in Fig. 2.3b linear).

The largest nucleic-acid specimen studied to date is Lambda DNA – the bacte-riophage for the Escherichia coli bacterium, which consists of 48,502 base-pairs [ 12 ]. A typical transmission spectrum for Lambda DNA is shown in Fig. 2.3c between 0.8 and 1.1 THz. Again, there are three prominent signatures, but they are even weaker and broader than those in Fig. 2.3b , and not very distinguishable from the typical transmission-spectrum undulations. However, it is compelling to com-bine all of the center frequencies in Fig. 2.3 to compose the trend chart shown in Fig. 2.4 . Here we see the tendency for the center frequency to shift downward between the smallest DNA (17-bp) and the largest (48.5-kbp) tested.

2.3.2 Waveguide Concentrator

2.3.2.1 Dry Control Sample

The circular-waveguide coupler is displayed in Fig. 2.5a , consisting of two conical horn antennas aligned back-to-back and separated by a thin circular waveguide sec-tion located near the throat of each horn. The thin waveguide section contains the biosample with two very thin (~12 μm thick) plastic windows held taut by the wave-guide fl anges. Because of the high radiative condensing capability of horn antennas, this technique is well suited to coupling limited THz power to a small volume fi lled with a solid sample. Our fi rst investigation of this method utilized a 500-μm-diameter circular-waveguide section that was 380 μm thick. This corresponds to a sample


12

1.2

1.1

1

0.9

0.8

0.7700 750 800 850 900 950 1000 1050 1100

normalized transmisson

b

1

0.95

0.9

0.85

0.8

0.75

0.7

0.65

0.6

0.55800 850 900 950 1000 1050 1100

c

aT

rans

mis

sion

Tra

nsm

issi

on

Frequency [GHz]

Frequency [GHz]

S1

S1

S2

S2

S3

S3

Fig. 2.3 THz transmission measured through fl uidic chips fi lled with solutions of ( a ) si-RNA, ( b ) 50-bp DNA, and ( c ) Lambda DNA. S1, S2, and S3 denote the strongest signatures, assumed to be common between the three samples

E.R. Brown et al.

13

Fig. 2.4 The three signature center frequencies, S1, S2, and S3 from Fig. 2.3 plotted vs size of the nucleic acid sample in units of base-pairs [bp]

a

b

c

10-2

10-3

10-4

10-5

10-6

10-7

300 400 500 600 700 800 900 1000

200 300

FWHMª 30 GHz

530-GHzLine

400 500 600 700

1100 1200200

Frequency [GHz]

Frequency [GHz]

Tra

nsm

issi

on

Tra

nsm

issi

on

10-1

100

TE11 TE21 TE22TM01 TM21 TM02TE12

TM11,TE01

Fig. 2.5 ( a ) Exploded view of circular waveguide concentrator with conical horn coupling, ( b ) Background transmission through waveguide concentrator relative to free-space propagation, ( c ) Transmission through lactose monohydrate


14

volume of ≈77 μl, which is small enough that dangerous or expensive samples can be used with relative impunity.

The experimental protocol was to start with the horn coupler located as shown in Fig. 2.1 but without any sample, and then measure transmission relative to a background consisting of the open-path spectrum. As always the noise fl oor was measured with the THz path blocked, resulting in the horn-coupler transmission spectrum shown in Fig. 2.5b . There are two notable aspects of this spectrum. The fi rst is a precipitous turn-on in the transmission around 320 GHz that rises by about 40 dB starting from just above the noise fl oor. This is the expected cut-off frequency f c of the TE 11 mode of the 500-μm diameter circular waveguide at the throat of the horn. Electromagnetic theory [ 13 ] predicts this as f c = (1.841/d)* c/(π) = 346 GHz, assuming d = 0.020 in. (508 μm), the nominal radius of the circular waveguide according to the machine tools used. The discrepancy between the experiment and theory is explained by the machining tolerance of ≈+5 %. Adding this to the nomi-nal diameter, we get d = 0.021 in (r = 533 μm), or f c = 330 GHz, as shown by the vertical dashed line in Fig. 2.5b and in good agreement with experiment.

The second notable aspect is a roll-off in transmission with frequency up to 1.0 THz, and a maximum transmission of only ≈ 7 × 10 −3 around 340 GHz. The gradual roll-off in transmission is attributed to excitation of higher-order modes in the circular waveguide. The cut-off frequency for each successive mode is shown in Fig. 2.5b (assuming d = 508 μm). The fi rst debilitating drop in transmission occurs around 1.0 THz where the TM 21 , TE 12 and TM 02 cut-off frequencies are in close proximity. The low maximum transmission is attributed primarily to mismatch and misalignment between the quasi-Gaussian beams in the coherent THz spectrometer and the horn coupler. This misalignment is a consequence of too many degrees-of- freedom. The guided-wave portion of the horn coupler is mounted in a gimbal, so must be adjusted in both azimuth and elevation, which is diffi cult to optimize. In spite of the mismatch loss, the instrument still produces the radiative concentration effect we seek, and achieves useful transmission levels because of the high dynamic reserve of the coherent photomixing spectrometer.

To demonstrate spectroscopic capability on solids, we fi lled the circular- waveguide section with lactose monohydrate powder using a needle. The monohy-drate-crystallized form of lactose (milk sugar) has become a standard THz absorber because of its two strong and narrow signatures centered at 530 and 1,369 GHz, respectively [ 14 , 15 ]. In our experiment, the transmission through the lactose was computed relative to a background scan through the empty horn coupler, and to the instrument noise fl oor. As displayed in Fig. 2.5c , the lactose signature centered at 530 GHz is fully resolved in depth and width. This is believed to be the smallest volume of lactose monohydrate ever detected in the THz region.

2.3.2.2 Wet Sample

The interest in THz bacterial spectroscopy started in earnest with the anthrax pathogen ( Bacillus anthracis ) domestic-terror events in the U.S. shortly after 9/11/2001. This event proved that a simple postal letter could disguise the presence

E.R. Brown et al.

15

of the deadly anthrax spores from conventional biosensors. The high transparency of paper and similar dry insulating materials to THz radiation begged the question of whether Anthrax offered any radiative signatures. So some experimental studies were conducted on an innocuous Anthrax simulant, Bacillus subtilis (Bg), and signatures were measured at or near 415 and 1,035 GHz [ 16 , 17 ]. However, these signatures were rather weak and diffi cult to reproduce, in addition to the fact that (Bg) was diffi cult to obtain outside of a biological laboratory.

Therefore, more recently we investigated another species of genus Bacillus, B. thuringiensis (Bt), also a candidate surrogate for anthrax, but much easier to obtain and less controlled. In fact, it is a common “green” pesticide used through-out the world, and widely cultured for agriculture and plant sciences. Our fi rst sample had a pasty consistency so was readily smeared into Whatman TM -5 fi lter paper. To our delight, this provided the interesting THz transmission spectrum shown in Fig. 2.6a , with unmistakable and reproducible signatures centered around 410, 915, and 1,050 GHz [ 18 ]. The signature at 915 GHz was, in fact, the strongest THz biosignature we had ever measured up to that point. To elucidate the origin of the strong signature, we obtained a scanning-electron micrograph of the sample as shown in Fig. 2.6b . The spores are clearly visible as prolate spheroids, and occupy approximately 50 % of the biomass, the remainder of which is mostly obliterated cell-wall and cytoplasmic material.

To determine whether the signatures in Fig. 2.6a are independent of the fi lter- paper method, we took a very small amount of the moist, pasty Bt material and located it in the sample holder of the waveguide coupler of Fig. 2.5a . Although tedious and imprecise, the waveguide-coupler was able to maintain the moisture level for a long time (~1 h). We then obtained the spectrum in the lower part Fig. 2.7 showing the strongest signature centered around 910 GHz, which is reduced below that in Fig. 2.6a by more than the frequency uncertainty in our experiment (0.5 GHz). Analysis of this signature has led to its model as a surface-phonon polariton

200 400 600 800 1000

Water Lines

Btsignature

a b100

410

915

105010−1

10−2Tra

nsm

issi

on

Frequency [GHz]

120010−3

Fig. 2.6 ( a ) THz transmission spectrum through Bacillus thuringiensis (Bt) paste smeared on fi lter paper. ( b ) Scanning electron micrograph of Bt paste


16

resonance on the outer coat of the Bacillus spore. Although this was fi rst proposed in 2006 for B. subtilis [ 19 ], it could never be proven because of the relative weakness of the signatures.

Believing that the hydration level could be responsible for the frequency shift, we maintained the same sample in the waveguide coupler for ~44 min just to de hydrate with no other changes in the experiment. The spectrum shown in the upper part of Fig. 2.7 was obtained. As expected, with less hydration there was signifi cantly higher transmission through the sample across the entire band since liquid water is very absorbing in the THz region. However, to our surprise the strong signature at 910 GHz disappeared, proving that hydration level affects the signature strength signifi cantly. This may also be consistent with the surface-phonon polariton mechanism since in the moist state liquid water will likely bind to the spore coat, polarizing its surface. And when dried, the surface proteins on the spore collapse their normal structure.

2.4 Conclusion and Future Work

Our research has focused on the detection of signatures in biomolecules and bioparticles, especially nucleic acids and bacterial particles (i.e., spores). Our quest has been aided signifi cantly by the development of concentration methods both for the biomaterial itself and the THz radiation. The preferred concentration method for

Fig. 2.7 THz transmission through small quantity of Bt paste in circular waveguide coupler while moist ( lower , red ), and 44 min later after drying ( upper , blue )

E.R. Brown et al.

17

liquid samples has been the fl uidic chip fabricated in quartz. This was useful in detecting strong signatures in double-stranded nucleic acids up to 48.5-kbp, although the detectability appears to decrease with molecular size. The preferred radiative concentration method has been the circular waveguide with conical-horn coupling. This was useful in detecting the first signatures in bacterial spores ( B. thuringiensis ) and demonstrating their sensitivity to hydration.

Future work will focus fi rst on improving these sample presentation methods, starting with the use of electrophoresis on the fl uidic chips to further concentrate and linearize the biomolecules, and also establish concentration control. The wave-guide concentrator will be improved by reducing its insertion loss via quasi-optical (i.e., lens) coupling between the spectrometer THz beam and the conical horns. Then we will be in a position to assess these methods and the associated THz signa-tures as a viable sensor technology. This must include estimates of the sensitivity with respect to sample concentration, and selectivity with respect to other signatures and the background clutter that is so pervasive in THz spectroscopy.

Acknowledgments This material is based upon work supported by, or in part by, the U. S. Army Research Laboratory and the U. S. Army Research Offi ce under contract numbers W911NF-11-1-0024 and W911NF-11-C-0080.

References

1. Van Zandt LL, Prohofsky EW, Kohli M (1980) Microwave absorption by double-helical DNA. Int J Quant Chem Quant Biol Symp 7:35–38

2. Wittlin A, Genzel L, Kremer F, Häseler D, Poglitsh A (1986) Far-infrared spectroscopy on oriented fi lms of dry and hydrated DNA. Phys Rev A 34:493–500

3. Van Zandt LL, Saxena VK (1994) Vibrational local modes in DNA polymers. J Biomol Struct Dyn 11:1149

4. Woolard DL, Kosica T, Rhodes DL, Cui HL, Pastore RA, Jensen JO, Jensen JL, Loerop WR, Jacobsen RH, Mittleman D, Nuss MC (1997) Millimeter wave-induced vibrational modes in DNA as a possible alternative to animal tests to probe for carcinogenic mutations. J Appl Toxicol 17(4):243–246

5. Verghese S, McIntosh KA, Calawa S, Dinatale WF, Duerr EK, Molvar KA (1998) Generation and detection of coherent terahertz waves using two photomixers. Appl Phys Lett 73:3824

6. Bjarnason JE, Brown ER (2005) Sensitivity measurement and analysis of an ErAs:GaAs coherent photomixing transceiver. Appl Phys Lett 87:134105

7. Demers JR, Logan RT Jr., Bergeron NJ, Brown ER (2008) A coherent frequency-domain THz spectrometer with a signal-to-noise ratio 60 dB at 1 THz. In: Proceedings of the SPIE, Paper# 6949-8, Orlando, 16–20 March 2008

8. Brown ER, Bjarnason J, Chan TLJ, Driscoll DC, Hanson M, Gossard AC (2004) Room temperature, THz photomixing sweep oscillator and its application to spectroscopic transmis-sion through organic materials. Rev Sci Inst 75:5333

9. Xia D, Brueck SRJ (2005) Fabrication of enclosed nanochannels using silica nanoparticles. J Vac Sci Tech B23:2694–2699

10. Xia D et al (2008) DNA transport in hierarchically-structured colloidal nanoparticle porous- wall nanochannels. Nano Lett 8:1610–1618

11. Brown ER, Mendoza EA, Xia D-Y, Brueck SRJ (2010) Narrow THz spectral signatures through DNA and RNA in nanofl uidic channels. IEEE Sensors J 10(755)


18

12. Sanger F, Coulson AR, Hong GF, Hill DF, Petersen GB (1982) Nucleotide sequence of bacteriophage λ DNA. J Mol Biol 162(4):729–773

13. Pozar DM (2005) Microwave engineering, 3rd Ed. Wiley, Hoboken, NJ 14. Brown ER, Bjarnason JE, Fedor AM, Korter TM (2007) On the strong and narrow absorption

signature in lactose at 0.53 THz. Appl Phys Lett 90:061908 15. Roggenbuck A, Schmitz H, Deninger A, Camara Mayorga I, Hemberger J, Gusten R,

Gruninger M (2010) Coherent, broadband continuous-wave THz spectroscopy on solid-state samples. New J Phys 12:043017

16. Globus TR, Woolard DL, Khromova T, Crowe TW, Bykhovskaia M, Gelmont BL, Hesler JL, Samuels AC (2003) THz-spectroscopy of biological molecules. J Biol Phys 29:89–100

17. Brown ER et al (2004) Optical attenuation signatures of Bacillus Subtilis in the THz region. Appl Phys Lett 84(18):3438–3440

18. Zhang W, Brown ER, Viveros L, Burris K, Stewart N (2013) Narrow terahertz attenuation signatures in Bacillus thuringiensis . J Biophoton. doi: 10.1002/jbio.201300042

19. Brown ER, Khromova TB, Globus T, Woolard DL, Jensen JO, Majewski A (2006) THz regime attenuation signatures in Bacillus subtilis and a model based on surface polariton effects. IEEE Sens J 6(5):1076–1083

E.R. Brown et al.

http://dx.doi.org/10.1002/jbio.201300042


Abstract In this paper we investigate how a valence band excitation couples with TE-polarized THz radiation in a microcavity. Nonequilibrium Many Body calcula-tions delivering an optical response beyond the Hartree-Fock approximation are used as input for a dielectric approach to the polariton/antipolariton problem. Analytical approximations that can be easily programmed are given and applied to GaAs-AlGaAs structures.

Keywords THz • Polaritons • Antipolaritons • Nonparabolicity • Many-body effects • Intervalence band transitions

3.1 Introduction

Devices based on intersubband (ISB) transitions can unleash the potential of many applications for the detection of chemical, biological, radiological and nuclear (CBRN) agents, explosives and illegal drugs. These can be detected by their charac-teristic absorption spectra at THz frequencies with high selectivity and resolution in applications fields such as industrial quality inspection control, customs inspection and security screening. Moreover, THz radiation has no endangering effects on human beings and enables higher contrast for “soft matter” than x-rays [1]. Intersubband polariton/antipolaritons have attracted strong interest recently [2–8]. Intersubband polaritons allow for external control of the system parameters. For example electri-cal injection can be achieved through resonant tunneling [7] and so the coupling

Chapter 3Intervalence THz Antipolaritons

I.A. Faragai and M.F. Pereira Jr.

I.A. Faragai • M.F. Pereira Jr. (*)Materials and Engineering Research Institute, Sheffield Hallam University, Sheffield, UKe-mail: [email protected]

20

strength can be improved. Recently, THz polariton emitters [8] have been reported. Since most of the work in intersubband optics in the literature focuses on inter-conduction band transitions, it is timely to investigate the possibilities that the valence band can deliver in this field. Thus, in this work we predict how a valence intersubband material excitation in either the absorption or gain cases couples with a THz TE-polarized cavity photon.

3.2 Main Equations

To obtain the energy dispersion relations for the THz polaritons and antipolaritons we consider as input to the dielectric function formalism, an optical susceptibility obtained from a selfconsistent evaluation of nonequilibrium many body green func-tions. The numerical method used can be applied to both intersubband [9–16] and interband [17–20] cases. The resulting susceptibility is adjusted to a simple model given (for one transition) by:

χ ωπ ω ω σ ω ω σ

( ) = −− +

−+ +

1

4 0 0

Λ Λi i

,

(3.1)

where ω and ω0 are the photons and transition frequencies between two arbitrary subbands, Λ measures the oscillator strength due to dipole coupling and σ is the line-shape broadening (dephasing) term. The optical susceptibility and the dielec-tric function of the medium are connected by the equations

e w e plc w( ) = + ( )b 4 ,

(3.2)

where εb and λ are respectively the background dielectric function and the effective medium factor for the cavity. Details of this approach can be found in [5]. The microcavity mode stems from using the dielectric function in Eq. 3.2. in Maxwell’s equations. This leads to the following wave equation with a transverse electric (TE) mode polarized electric field

k k

cy z2 2

2

2+ = ( )w

e w .

(3.3)

here ky and kz, are respectively the components of the electric fields wave vectors lying in-plane of the layers and the quantum well growth direction. Confinement of electric field is achieved in the GaAs/AlGaAs quantum well cavity with AlAs layers at the top and bottom of the QWs which trap the incident beam due to highly refractive prism in-between the air and GaAs cap. The nonzero absorption or gain imply that kz and ε(ω) are both complex numbers. The microcavity mode, thus reads


21

k

w c

w

cy

c

ci r

22

2

4

4

2

22

2

24+ - ( ) = ( )p w

pe w

we w ,

(3.4)

where wc is the width of the cavity and εr are εi the real and imaginary parts of the dielectric function, kz → g + ix and g = π/wc. The wave vector ky is related to the incident angle θ

by the Snell’s law kc

nc

ny cap b b= =w

qw

qsin sin , nb is the background refractive

index of the cavity medium. The refractive index in the cap layer is ncap cap= e ,

Fig. 3.1. By substituting the real and imaginary parts of Eq. (3.3) into Eq. (3.4) and neglecting the last term on the LHS of Eq. (3.4) due to the fact that εr ≫ εi, the polari-ton/antipolariton dispersion is finally obtained as

Fig. 3.1 Sketch of the microcavity resonator suitable for investigations with a TE-polarized electric field. The polariton/antipolaritons dispersions in Figs. 3.2 and 2.3 are given as a function of incident angle θ of a THz beam. The incident angle θ, the angle inside the GaAs cap layer θcap and the angle in the cavity θb (not shown in Fig. 3.1) are related by the Snell’s law. We assume the prism to have the same refractive index as the cap layer so that θ = θcap

3 Intervalence THz Antipolaritons

http://dx.doi.org/10.1007/978-94-017-8572-3_2

22

yA A A

2

2 2 2 2 2 2 2

2

1 2 1 2 4 1

2 1=

- + + ± - + +( ) + -( )-( )

¢ ¢x D l x D l x

x,

(3.5)

where dimensionless variable have been used; y = ω/ω0, ξ = εcap/εb ⋅ sin2θ,

A c b= ( )p w e/ 0 and Δ′ = Λ/(εbω0). Here ω0 is the resonance frequency.

3.3 Numerical Results and Discussion

The numerical results presented in this paper are for a cavity core of total length 37.9 nm with 165 GaAs/AlGaAs QWs. The well and the barrier width are respec-tively chosen to be 10 and 220 nm each, with 985 nm (2 × 492.5 nm) AlAs at the top and bottom to provide the required mirror cladding. Figure 3.1 shows the complete microcavity structure grown on a GaAs substrate. Photon confinement is achieved with lower refractive index AlAs layers on surrounding the cavity core.

Figure 3.2a depicts intervalence band polariton dispersion relations as a function of incident angle θ for the case of absorption media in the THz spectral region, based on TE-mode polarization. The carrier density is assumed to be indepen-dently thermalized in each subband at 300 K, with population difference between initial and final subbands δN = 2 × 10− 11 cm− 2 for absorption (leading to a polariton). Figure 3.2b has a population difference δN = − 2 × 10− 11 cm− 2, which means that the upper subband is occupied and the lower subband is empty. This corresponds to gain (giving rise to an antipolariton. A clear anticrossing, with a cavity resonance frequency at approximately 5.0 meV which is a fraction of the transition frequency (10.0 meV) is obtained, defining clearly the mixed cavity and material excitation modes. The background dielectric function is numerically calculated to be εb = 10.028

Fig. 3.2 (a) Intervalence band THz polariton dispersion due to absorption media for the microcav-ity geometry of Fig.3.1, in TE mode. (b) Intervalence band THz antipolaritons due to inverted media (in the gain regime) based on TE mode microcavity geometry shown in Fig. 3.1


23

and the anisotropic constant λ = 0.0424. For details of the method used see Ref. [5]. Further investigations will be conducted to incorporate all relevant scattering process [14] and the influence of dephasing as well.

3.4 Conclusion

In conclusion, this paper shows the relevance of valence bands based design to study the interaction of TE-polarized THz radiation with intervalence band excita-tions in a microcavity resonator. The system proposed has potential for easier mea-surement of polaritons/antipolaritons without the need to grow the resonator on top a slanted prism, as in usual TM mode, conduction band realizations. A prism on top of the sample would be enough to perform the measurements.

Acknowledgments The authors acknowledges support from MPNS COST ACTION MP1204 – TERA-MIR Radiation: Materials, Generation, Detection and Applications. I.A. Faragai’s research is supported by the KUST-Wudil under TETFUND, Nigeria.

References

1. Liu H-B, Chen Y, Bastiaans GJ, Zhang X-C (2006) Detection and identification of explosive RDX by THz diffuse reflection spectroscopy. Opt Express 14:415

2. Dini D, Köhler R, Tredicucci A, Biasiol G, Sorba L (2003) Microcavity polariton splitting of intersubband transitions. Phys Rev Lett 90:116401

3. de Liberato S, Ciutti C (2009) Stimulated scattering and lasing of intersubband cavity polari-tons. Phys Rev Lett 102:136403

4. Zanotto S, Biasiol G, Degl’innocenti R, Sorba L, Tredicucci A (2010) Intersubband polaritons in one dimensional plasmonic crystal. Appl Phys Lett 97:231123

5. Pereira MF Jr (2007) Intersubband antipolaritons: microscopic approach. Phys Rev B 75:195301

6. Todorov Y, Sirtori C (2012) Intersubband polaritons in the electrical dipole gauge. Phys Rev B 85:045304

7. Sapienza L, Vasanelli A, Colombelli R, Ciuti C, Chassagneux Y, Manquest C, Gennner U, Cirtori C (2008) Electrically injected cavity polaritons. Phys Rev Lett 100:136806

8. Geiser M, Scalari G, Castellano F, Beck M, Faist J (2012) Room temperature THz polariton emitter. Appl Phys Lett 101(141118)

9. Kubis T, Raj Mehrotra S, Klimeck G (2010) Design concept of THz quantum cascade lasers: proposal for THz laser efficiency improvement. Appl Phys Lett 97:261106

10. Winge DO, Lindskog M, Wacker A (2012) Nonlinear response of quantum cascade structures. Appl Phys Lett 101:211113

11. Pereira MF Jr (2011) Microscopic approach for intersubband-based thermophotovoltaic struc-tures in the THz and mid infrared. JOSA B 28(8):2014

12. Pereira MF, Tomić S (2011) Intersubband gain without global inversion through dilute nitride band engineering. Appl Phys Lett 98:061101

13. Schmielau T, Pereira MF (2009) Nonequilibrium many body theory for quantum transport in terahertz quantum cascade lasers. Appl Phys Lett 95:231111

3 Intervalence THz Antipolaritons

24

14. Schmielau T, Pereira MF (2009) Impact of momentum dependent matrix elements on scatter-ing effects in quantum cascade lasers. Phys Status Solidi b 246:329

15. Schmielau T, Pereira MF (2009) Momentum dependent scattering matrix elements in quantum cascade laser transport. Microelectron J 40:869

16. Pereira MF Jr, Nelander R, Wacker A, Revin DG, Soulby MR, Wilson LR, Cockburn JW, Krysa AB, Roberts JS, Airey RJ (2007) Characterization of intersubband devices combining a nonequilibrium many body theory with transmission spectroscopy experiments. J Mater Sci Mater Electron 18:689

17. Chow WW, Pereira MF Jr, Koch SW (1992) Many-body treatment on the modulation response in a strained quantum well semiconductor laser medium. Appl Phys Lett 61:758

18. Pereira MF Jr, Henneberger K (1998) Microscopic theory for the optical properties of coulomb- correlated semiconductors. Phys Status Solidi (b) 206(477)

19. Pereira MF Jr, Henneberger K (1997) Gain mechanisms and lasing in II-VI compounds. Phys Status Solidi b B202:751

20. Pereira MF Jr (1995) Analytical solutions for the optical absorption of superlattices. Phys Rev B52:1978



Abstract We report on a novel type of an aperiodic one-dimensional multilayered structure which can be used for terahertz radiation manipulations. The studied structure is formed by stacking together dielectric layers according to the Kolakoski self-generation scheme. Numerical simulations are carried out for different configu-rations of the structure to reveal the dependence of its optical characteristics on the generation stage, frequency, and the angle of wave incidence. The dependence of the number and width of omnidirectional bandgaps on the refractive indexes and the thicknesses of the dielectric materials is studied.

Keywords Aperiodic photonic structures • Kolakoski sequence • Photonic bandgap • Transparent states • Omnidirectional reflection • Transfer matrix method

Chapter 4Terahertz Aperiodic Multilayered Structure Arranged According to the Kolakoski Sequence

Volodymyr I. Fesenko, Vladimir R. Tuz, and Igor A. Sukhoivanov

V.I. Fesenko (*) Institute of Radio Astronomy of NASU, Kharkiv, Ukraine

Laboratory “Photonics”, Kharkiv National University of Radio Electronics, Kharkiv, Ukraine e-mail: [email protected]

V.R. Tuz Institute of Radio Astronomy of NASU, Kharkiv, Ukraine

School of Radio Physics, Karazin Kharkiv National University, Kharkiv, Ukrainee-mail: [email protected]

I.A. Sukhoivanov Departamento de Electronica, FIMEE, University of Guanajuato, Salamanca, Mexicoe-mail: [email protected]

26

4.1 Introduction

The aperiodic design of photonic structures (quasiperiodic or fractal arrangements of their constituents) gives an additional degree of freedom which allows achieving an enhancement of the optical performance in some specific applications [1]. These prospects occur due to several unique characteristics of aperiodic structures, namely, the formation of multi-fractal bandgaps, existence of critical modes, field localiza-tion, scaling and transport properties. For example, it is revealed that these features can be useful in designing optical filters, omnidirectional mirrors, thermal emission controllers, number recognition systems and data storage devices, etc. At the same time, the clear experimental indications on certain advantages of aperiodic systems over periodic ones were previously reported in the nonlinear optics field [2, 3], where it was shown that the second or third harmonic generation processes are more efficient in quasiperiodic systems in virtue of their richer Fourier spectrum structure. The nonlinear properties of aperiodic optical heterostructures can also be used to fabricate compact-sized compressors for laser pulses due to their larger group velocity dispersion [4].

On the other hand, there is a certain number of publications which are focused on using aperiodic design in the porous silicon (PSi) based systems [5]. It is expected that these photonic structures are capable to operate in the terahertz range as chemi-cal and biological sensors. Such sensors are proposed and tested for the detection of gases, liquids and biological molecules. Their main sensing mechanism is based on the refractive index variation of the PSi due to the partial substitution of the air in the pores on exposure to biochemical substances. Particularly, in a recent study [6] the sensitivities of the resonant optical biochemical sensors constructed on both periodic and aperiodic PSi-based multilayers were compared. The measurements indicate that the aperiodic structures provide a higher filling capability with respect to the periodic one, due to the lower number of interfaces, and this feature makes them more sensitive. The recent studies in the area of chemical and biological sensors based on PSi and polymer nanocomposites for sensing/detection of CBRN agents are also presented in [7].

In accordance with the above mentioned, there is of interest to search for new designs of aperiodic structures which can satisfactory cope with certain physical requirements necessary for the fabrication of improved devices operating in the terahertz range. Pursuing this goal, in the present paper we propose a novel type of quasiperiodic structures suitable for terahertz applications. It is based on the PSi layered system constructed in accordance to the generation rules of the Kolakoski sequence [8]. The main focus is directed on the study of its spectral behaviors and frequency bands where omnidirectional reflection is possible.

V.I. Fesenko et al.

27

4.1.1 Kolakoski Sequence Generation Rules

A Kolakoski self-generating sequence (KSGS) is defined by the property that it equals the sequence of its run lengths, where a run is a maximal subword consisting identical letters [9]. A one-sided infinite sequence ξ

x

x= …

… =2 2 1 1 2 1 2 2 1 2 2 1 1 2

2 2 1 1 2 1 2 2 1123123123123123123123123123

. (4.1)

over the alphabet A = {1, 2} is called the Kolakoski sequence. The sequence

x x′ = = …1 12211212212211211 (4.2)

is yet another type of the Kolakoski sequence over this alphabet. So, the sequence ξ which is started from digit ‘2’ is named K(2, 1), while another one ξ′ which is started from ‘1’ is named K(1, 2) [10]. Note, that generally, the K(2, 1) and K(1, 2) sequences have different number of letters for the same generation stage.

4.1.2 Aperiodic Structure Configuration

Further, for realization of the Kolakoski multilayer in the terahertz range, we assume that Ψ and ϒ are two dielectric layers which correspond to digits ‘1’ and ‘2’ in the K(2, 1) and K(1, 2) sequences, respectively (see Fig. 4.1). The number of generation

Fig. 4.1 Two types of the aperiodic Kolakoski multilayered structure over the alphabet A = {1, 2}

4 Terahertz Aperiodic Multilayered Structure Arranged According to the Kolakoski…

28

stage of the Kolakoski sequence is defined as σ and the total number of layers in the system is N.

It is assumed that the letters Ψ and ϒ denote two different layers with refractive indexes n1 and n2, respectively. The optical thicknesses of both layers Ψ and ϒ satisfy a quarter-wave condition and denoted as d1n1 = d2n2 = D, where D = λ0/4. The total length of the whole structure is Λ. The incident medium and substrate are homogeneous, isotropic and have refractive indexes n0 and n3, respectively.

The incident field is a plane monochromatic wave of frequency ω with perpen-dicular (electric field vector E is perpendicular to the plane of incidence) or parallel (electric-field vector E is parallel to the plane of incidence) polarization (s- and p-polarized waves). The direction of the wave propagation in the input isotropic medium z ≤ 0 is defined by angle θ relative to the z axis.

4.2 Numerical Results: Solution Analysis

In order to perform simulations of light propagation through the investigated structure, the transfer matrix formulation proposed earlier [11–14] is adopted. Another treatment which can be used for simulation of PSi layered structures is the computational scheme based on the electromagnetic field scalarization method in coupling with the finite-difference procedure [15].

4.2.1 Spectral Behaviors

In this section we draw an attention to the properties of transmission spectra of the PSi-based Kolakoski multilayer, and, especially, to the features of completely transparent states and photonic bandgaps positions. For this reason we consider the optical characteristics of both K(1, 2) and K(2, 1) structure configurations for the different stage σ of their generation. In our numerical simulation the typical param-eters of the PSi-based multilayers in the THz range [16] are used. The transmission spectra calculated for the different generation stage σ (and, accordingly, for the different number of layers N) of the studied structure are summarized in Fig. 4.2. These spectra have interleaved bands with high and low average level of transmis-sion, which are referred to the pseudo-stopbands (PBGs) and pseudo-passbands (the PBGs are marked in figures as a shaded area). In each such band there is a certain frequency v0 with respect to which the spectrum is symmetrical. All PBGs have the same bandwidth (about 0.1 THz), and the central frequencies of these bandgaps satisfy the condition vc = v0 ± 0.4 THz.

The interference with the wave reflected from the outside boundaries of the structure gives small-scale oscillations in each pseudo-passbands. Obviously the number of these small-scale oscillations around the central frequency v0, their

V.I. Fesenko et al.

29

spectral density, width, position and magnitude are dependent on the Kolakoski sequence generation stage σ. Furthermore, as it is typical for most of aperiodic systems [17], as the number of layers N increases, the PBGs boundaries become sharper and more resonant transmission peaks appear in the PBGs. These peaks get narrower and their magnitudes verge towards 1.

One can see that the spectral position of the transmission peaks, their quantity and the distance between them are also different for the K(1, 2) and K(2, 1) struc-tures. Moreover, the spectra are different for the multilayers of the same type but with odd (N-) and even (N+) total number of layers in the system. It is found that both types of the Kolakoski structure with even number of layers demonstrate the presence of the resonant peak at the central frequency v0 while this peak is absent in the spectra of systems with odd number of layers.

Fig. 4.2 Transmission spectra of PSi-based K(1, 2) (a, b) and K(2, 1) (c, d) multilayers on the different generation stage σ for the normally incident (θ = 0) electromagnetic wave. The system consists of odd N- (a, c) and even N+ (b, d) number of layers; n1 = 1.56, n2 = 2.06


30

4.2.2 Omnidirectional Reflection

It is known that under proper conditions, periodic multilayers can reflect light from all incidence angles and any polarization [18]. To date, such systems are referred to omnidirectional mirrors – reflectors which are able to reflect light at any polarization, any incidence angle, and over a wide range of wavelengths [1]. However, in the periodical systems only one omnidirectional PBG exists within a period of the reciprocal space. This limitation does not apply to aperiodic multilayers possessing a much more complex structure in the reciprocal space. Thus, the presence of several omnidirectional PBGs in a period of the reciprocal space has been theoretically and experimentally confirmed for a number of aperiodic multilayer configurations [19–22], including those ones made of PSi [23].

Thus, in this section, the PBGs of the Kolakoski aperiodic multilayer system are studied for oblique incidence of the exciting wave. In our calculation we use the same values of material parameters of the structure as in the theoretical work [20] to maintain continuity.

According to Fig. 4.3, both K(1, 2) and K(2, 1) systems can have multiple PBGs. When r is small, there is not any omnidirectional PBG in the spectra. If r is large enough, omnidirectional PBGs appear for both s- and p-polarizations. The width of these omnidirectional PBGs is exactly determined by the refractive index and the thickness of the dielectric materials. Remarkably that the width and position of the omnidirectional PBGs are the same for both K(1, 2) and K(2, 1) structure configurations. Comparison of the calculated PBGs for the Kolakoski dielectric multilayer with obtained results for Thue-Morse structure [20] shows that in the Kolakoski multilayer the width of the omnidirectional PBGs is wider than those ones in the Thue-Morse structure at the same number of layers, material and structural parameters.

4.3 Conclusion

In conclusion, we have proposed and studied optical properties of new type of a PSi-based aperiodic structure constructed in accordance to the Kolakoski sequence generation rules. It is shown that the proposed structure has completely transparent states in the transmission spectra and pronounced omnidirectional reflectance. We suppose that this kind of aperiodic structure can be used for the fabrication of omnidirectional mirrors, multifrequency laser cavities, optical filters, and sensors with potential for explosives and CBRN agent detection.

Acknowledgments This work was partially supported (V. R. Tuz) by Ministry of Education and Science of Ukraine under the Program “Electrodynamics of layered composites with chiral proper-ties and multifunctional planar systems”, Project No. 0112 U 000561.

V.I. Fesenko et al.

31

References

1. Macia E (2012) Exploiting aperiodic designs in nanophotonic devices. Rep Prog Phys 75(3):036502

2. Zhu S-N, Zhu Y-Y, Ming N-B (1997) Quasi-phase-matched third-harmonic generation in a quasi-periodic optical superlattice. Science 278:843–846

3. Sibilia C, Tropea F, Bertolotti M (1998) Enhanced nonlinear optical response of a Cantor-like and Fibonacci-like quasi-periodic structures. J Mod Opt 45(11):2255–2267

4. Makarava L, Nazarov M, Ozheredov I, Shkurinov A, Smirnov A, Zhukovsky S (2007) Fibonacci-like photonic structure for femtosecond pulse compression. Phys Rev E 75:036609

5. Agarwal V, Mora-Ramos ME (2007) Optical characterization of polytype Fibonacci and Thue–Morse quasiregular dielectric structures made of porous silicon multilayers. J Phys D Appl Phys 40(10):3203

Fig. 4.3 Calculated PBGs for Kolakoski dielectric multilayers K(1, 2) (a, b) and K(2, 1) (c, d) containing N = 32 layers, with n1 = 1.47 and different values of the ratio r = n1/n2; (a, c) r = 1.6; (b, d) r = 3.2


32

6. Moretti L, Rea I, De Stefano L, Rendina I (2007) Periodic versus aperiodic: enhancing the sensitivity of porous silicon based optical sensors. Appl Phys Lett 90:191112

7. Vaseashta A, Khudaverdyan S (eds) (2013) Advanced sensors for safety and security. Springer, Dordrecht, p 372

8. Tuz V, Fesenko V, Sukhoivanov I (2013) Optical characterization of the aperiodic multilayered anisotropic structure based on Kolakoski sequence. In: Integrated optics: physics and simula-tions, vol 8781, Proceedings of SPIE. SPIE, Bellingham, P. 87811C

9. Kolakoski W (1965) Self-generating runs, problem 5304. Am Math Mon 72:674 10. Sing B (2004) Kolakoski sequences – an example of aperiodic order. J Non-Cryst Solids

334:100–104 11. Tuz V, Kazanskiy V (2009) Electromagnetic scattering by a quasiperiodic generalized multi-

layer Fibonacci structure with grates of magnetodielectric bars. Waves Rand Complex Media 19(3):501–508

12. Tuz V (2009) Optical properties of a quasiperiodic generalized Fibonacci structure of chiral and material layers. J Opt Soc Am B 26(4):627–632

13. Tuz V (2009) A peculiarity of localized mode transfiguration of a Cantor-like chiral multilayer. J Opt A Pure Appl Opt 11(12):125103

14. Tuz V, Batrakov O (2010) Localization and polarization transformation of waves by a sym-metric and asymmetric modified Fibonacci chiral multilayer. J Mod Opt 57(21):2114–2122

15. Fesenko V, Sukhoivanov I, Shul’ga S, Andrade Lucio J (2013) Propagation of electromagnetic waves in anisotropic photonic structures. In: Passaro V (ed) Advances in photonic crystals. InTech, Rijeka, pp 79–105

16. Lo S-Z, Murphy T (2009) Nanoporous silicon multilayers for terahertz filtering. Opt Lett 34(19):2921–2923

17. Hsueh WJ, Wun SJ, Lin ZJ, Cheng YH (2011) Features of the perfect transmission in Thue–Morse dielectric multilayers. J Opt Soc Am B 28(11):2584–2591

18. Fink Y, Winn JN, Fan S, Chen C, Michel J, Joannopoulos JD, Thomas EL (1998) A dielectric omnidirectional reflector. Science 282:1679–1682

19. Lusk D, Abdulhalim I, Placido F (2001) Omnidirectional reflection from Fibonacci quasiperiodic one-dimensional photonic crystal. Opt Commun 198:273–279

20. Qiu F, Peng RW, Huang XQ, Hu XF, Mu Wang, Hu A, Jiang SS, Feng D (2004) Omnidirectional reflection of electromagnetic waves on Thue-Morse dielectric multilayers. Europhys Lett 68(5):658–663

21. Ben Abdelaziz K, Zaghdoudi J, Kanzari M, Rezig B (2005) A broad omnidirectional reflection band obtained from deformed Fibonacci quasi-periodic one dimensional photonic crystals. J Opt A Pure Appl Opt 7:544–549

22. Hsueh WJ, Chen CT, Chen CH (2008) Omnidirectional band gap in Fibonacci photonic crys-tals with metamaterials using a band-edge formalism. Phys Rev A 78:013836

23. Dal Negro L, Stolfi M, Yi Y, Michel J, Duan X, Kimerling LC, LeBlanc J, Haavisto J (2004) Photon band gap properties and omnidirectional reflectance in Si/SiO2 Thue–Morse quasicrys-tals. Appl Phys Lett 84(25):5186–5188

V.I. Fesenko et al.


Abstract Based on the density matrix theory we have developed the theoretical model electron transport in GaN/AlGaN photodetectors including both coherent and incoherent constituents. The lifetimes utilized in the model are computed using Fermi’s Golden Rule approach. We also use Ben-Daniel-Duke equation to obtain the envelope functions and single-electron energy spectrum.

Keywords Photodetection • Quantum cascade • GaN/AlGaN • Density matrix • Electron transport • Temporal response

5.1 Introduction

CBRN – weaponized or non-weaponized Chemical, Biological, Radiological and Nuclear materials – can cause great harm and pose significant threats in the hands of terrorists. Several different technologies are used today to detect chemical agents

Chapter 5Population Dynamic in Quantum Cascade GaN/AlGaN Photodetector Structure

S.V. Gryshchenko, Oleksiy Shulika, V.V. Lysak, and I.A. Sukhoivanov

S.V. Gryshchenko (*) Laboratory “Photonics”, Kharkov National University of Radio Electronics, Lenin av.14, 61166 Kharkov, Ukrainee-mail: [email protected]; [email protected]

O. ShulikaUniversidad de Guanajuato, Salamanca, Guanajuato, Mexicoe-mail: [email protected]

I.A. Sukhoivanov Department of Electronics Engineering, DICIS, University of Guanajuato, Comunidad de Palo Blanco, C.P. 36885 Salamanca, GTO, Mexicoe-mail: [email protected]

V.V. Lysak Semiconductor Physics Department, Chonbuk National University, 664-14, Deokjin-dong, Jeonju 651-756, Republic of Koreae-mail: [email protected]

34

(CAs). CAs are defined as chemicals intended to kill or seriously injure human beings as nerve agents, blister agents, and arsenical vesicants. A large variety of equipment using terahertz and mid infrared radiation is used identifying liquid droplets of CAs on surfaces and in vapors. Using quantum cascade detectors in these applications looks promising due to ultrafast response and narrow spectral band. The modelling of responsivity of nitride quantum cascade detector due to high spontaneous polarization field faced some difficulties. Adequate modeling of absorption and photocurrent should take into account many body effects and band coupling [1]. Also quantum efficiency is strongly dependent on phonon mediated electron transport in the extrac-tor region [2, 3]. Diffusion length and lifetimes of electrons are critical parameters to device performance. Recently, the III-V nitride alloys have been applied for to archive required configuration of energy states [4]. The internal piezoelectric and spontaneous polarizations in these materials allow producing required staircase energy spectrum with constant-thickness semiconductor layers as is shown in Fig. 5.1.

In this chapter, we present a density matrix model for the quantum cascade detec-tor taking into account optical transitions in the active region as well as electron transport in the extractor. Nonequilibrium Green’s Functions methods including full momentum and frequency dependence give a more accurate description of scatter-ing phenomena in cascaded structures [5]. But the thermalization time of distribu-tion function is very low due to room temperature work regime. So we use lifetimes calculated using the Fermi Golden Rule [3], which are consistent with our density matrix approach. In this work, our goal is obtaining lifetimes of the carriers in nitride quantum cascade structure.

5.2 Modeling

The considered GaN/AlGaN quantum cascade detector should operate at 1.7 m (0.73 eV). The absorption occurs in GaN active layer generating electron transition between subbands. Excited carrier travels to second cascade through series of

Fig. 5.1 Band diagram and wave functions

S.V. Gryshchenko et al.

35

relaxation events with participation of LO phonons. The quantum cascade consists from 40 cascades each of them starts from GaN absorbing well following by five barrier/well stack made from AlN/Al0,25Ga0,75N.

Piezoelectric polarizations in binary semiconductor materials have been computed using formulas taken from [6]. For ternary semiconductors, the piezo-electric polarization have been computed using linear interpolation formulas. The coordinate dependence of the piezoelectric polarization Ppiezo allow to obtain corre-sponding electric charge distribution ppiezo in the structure [5]:

p Ppiezo piezo= -Ñ

(5.1)

If the charge distribution is known, one can compute the potential solving the Poisson equation. Usually, the periodic boundary conditions for the potential is applied. In this case, the potential drop along one period of the structure is equal zero.

We use Ben-Daniel-Duke [7] equation to obtain the envelope functions and energies of the electron states. Within this approximation electron dispersion is sup-posed to be parabolic one [8]. It is good approximation for the conduction band. The eigen-value problem reads:

- ¶¶ ( )

¶¶

( ) + ( ) + ( )æ

èçç

ö

ø÷÷ ( ) = ( )

2 2 2

2

1

2z m z zz E z

k

zz zc

II

y y eym

(5.2)

where ψ(z) is the electron envelope function, Ec(z) is the potential profile of the conduction band edge, m is the effective mass for the growth direction, mII is the in-plane effective mass, k is the in-plane wave vector.

To calculate electron transport in the structure we use a simplified density matrix model developed by Dupont et al. [9]. The averaging of k-dependent scattering rates in this model eliminates the in-plane wave vector dependence of the density matrix elements and decreases significantly the number of kinetic equations. The band non-parabolicity is relatively small but this does not lead to inaccuracy [10, 11].

As basis states for the density matrix, the Wannier-Stark states have been chosen. Since we consider only seven states of the quantum-cascade detector, the density matrix ρ has dimension 7 × 7. Time evolution of a density matrix element is the solu-tion of the Liuville-Neuman equation:

dp t

dtRij

ij

( )=

(5.3)

Right side of the Eq. 5.3. is an element of the matrix R which reads:

R i H H S= ( ) -( ) -1 0 0/ r r

(5.4)

5 Population Dynamic in Quantum Cascade GaN/AlGaN Photodetector Structure

36

where H0 – 7 × 7 dimension Hamiltonian. Electron transport in the extractor is fully uncoherent being determined by phonon-assisted scattering rates. Electron transi-tions in active quantum well are mediated by photon absorption mainly. Phonon energy is not enough to induce energy transition between states 1 and 2.

Computing of defined above kinetic equations requires knowing of electron scat-tering rates 1/τij. Phonon-assisted scattering rates can be computed using Harrison’s approach [2] based on Fermi’s Golden Rule. For estimation of these terms we use 3D phonon approximation for the phonon modes and allow only polar-optical pho-non scattering. Equation (4) means phonon-dependent form factor:

G K z e z dzij z f

iK zi

z( ) = ( ) ( )∫ −ψ ψ

(5.5)

5.3 Results

Figure 5.2 shows form factors for transitions in active well and extractor. As well as the form factor mathematically consist of wave functions overlap integral this parameter give us all information about structure geometry.

We put in the model arbitrary distribution of electrons on each energy level. After some time most of electrons are relaxed to the ground state of absorbing quantum well. Then we apply a femtosecond optical pulse (120 ps) to the structure which excites the electrons. After optical excitation, the quantum structure relaxes to its initial state.

The structure with Al0,25Ga0,75N quantum wells in extractor shows high response time near 50 ps (solid curve on Fig. 5.3). Next we increase the Al concentration in to 0.35 in first two wells (left to right) (dashed curve). The changing of energy levels than results in decreasing dipole matrix element. On other hand this will enhance coupling between ground level of absorbing well and wells in extractor due to decreasing depth of the well. Therefore we will obtain higher lifetimes and slower relaxation as it happens on a dashed curve.

The relaxation rate defines response of the device. To produce high-speed photodetector the relaxation rate should be minimized.

Fig. 5.2 Form factors for different transitions in cascade


37

5.4 Conclusion

Developed model allows to predict the results of femtosecond pump-probe spectroscopy applied to the GaN/AlGaN quantum-cascade structures taking into account coherent and incoherent electron transport between subbands. The computa-tional efficiency of the model is caused by a series of approximations, the most cru-cial of which is the thermalization model for electron distribution functions belonging to subbands. Obtained results evidence that depending on the molar fraction of the Al in AlGaN semiconductor alloy the electron transport can be governed by the coher-ent process or incoherent one basing on the electron-phonon scatterings mainly.

References

1. Pereira MF Jr, Binder R, Koch SW (1994) Theory of nonlinear absorption in coupled band quantum wells with many-body effects. Appl Phys Lett 64:279

2. Harrison P (2005) Quantum wells, wires and dots, 2nd edn. Wiley, Hoboken 3. Giorgetta F, Baumann E, Graf M, Yang Q, Manz C, Kohler K, Beere H, Ritchie D, Lin_eld E,

Davies A, Fedoryshyn Y, Jackel H, Fischer M, Faist J, Hofstetter D (2009) Quantum cascade detectors. IEEE J Quant Electron 45(8):1039

4. Piprek J (2007) Nitride semiconductor devices. Wiley, Weinheim 5. Schmielau T, Pereira MF Jr (2009) Nonequilibrium many body theory for quantum transport

in terahertz quantum cascade lasers. Appl Phys Lett 95:231111 6. Fiorentini V, Bernardini F, Ambacher O (2002) Evidence for nonlinear macroscopic polariza-

tion in III-V nitride alloy heterostructures. Appl Phys Lett 80:1204–1206 7. Bastard G (1988) Wave mechanics applied to semiconductor heterostructures. Wiley- Interscience,

New York 8. Klymenko M, Shulika A, Sukhoivanov I (2011) Theoretical study of optical transition matrix

elements in InGaN/GaN SQW subject to indium surface segregation. J Sel Top Quant Electron 17(5):1374–1380

Fig. 5.3 Population dynamics with femtosecond pump pulse. Solid line – Al0,25Ga0,75N wells in extractor, dashed line Al0,25Ga0,75N mixed with Al0,35Ga0,75N wells

5 Population Dynamic in Quantum Cascade GaN/AlGaN Photodetector Structure

38

9. Dupont E, Fathololoumi S, Liu H (2010) Simplified density matrix model applied to three well terahertz quantum cascade lasers. Phys Rev B 81:205311

10. Gryshchenko SV, Klymenko MV, Lysak VV, Shulika OV, Sukhoivanov IA (2011) Temperature dependence of electron transport in GaN/AlGaN quantum cascade detectors In: Proceedings of SPIE (Conference proceedings), vol 8155, ISSN 0277-786X, ISBN 9780819487650, pp 81550O-1–81550O-6

11. Gryshchenko SV, Klymenko MV, Lysak VV, Sukhoivanov IA (2011) Modeling of optical spectral characteristics of nitrides-based quantum-cascade detectors. In: Pereira M, Shulika O (eds) Terahertz and mid infrared radiation. Generation, detection and applications, NATO series B: physics and biophysics. Springer, Dordrecht, pp 59–64



Abstract Recently, there has been an increasing demand to use Terahertz (THz) spectroscopy for screening and detection of CBRN agents and dangerous materials. Detection using THz radiation is possible due to structural absorption in this region. In this paper, we focus on the design of a novel detector based on metamaterials embedded in a one-dimensional photonic crystal. By optimizing the structural parameters of the metamaterial, the transmission of the proposed structure ranges from 3 to 8 THz. The result shows the appearance of a peak in the transmission band, which is located at 5.25 THz. By comparing this frequency to those of some explosives, we find that the presented structure based device can be used to identify explosives.

Keywords Metamaterials • THz radiation • Photonic crystals • Explosives • Transmission band

Chapter 6Design of Metamaterial Photonic Crystals for Explosives Detection

H. Hamdouni, F. Ouerghi, F. Abdelmalek, and H. Bouchriha

H. Hamdouni (*) National Institute of Applied Science and Technology, BP 676 Cedex 1080, University of Carthage, Tunis, Tunisia

Department of Physics, Quantum Physics and Photonics Laboratory, Faculty of Science at Tunis, University of El Manar, Tunis, Tunisiae-mail: [email protected]

F. Ouerghi School of Technology and Computer Science, University of Carthage, Tunis, Tunisia

F. Abdelmalek National Institute of Applied Science and Technology, BP 676 Cedex 1080, University of Carthage, Tunis, Tunisia

H. Bouchriha Department of Physics, Quantum Physics and Photonics Laboratory, Faculty of Science at Tunis, University of El Manar, Tunis, Tunisia

40

6.1 Introduction

Terahertz and Mid Infrared radiation – THz (0.3–10 THz) and Mid Infrared (10–100 THz) radiation sources and detectors are being intensively investigated for the detection of explosives and CBRN agents. CBRN is initialism for chemical, biological, radiological and nuclear. CBRN agents, explosives and illegal drugs can be detected by their characteristic absorption spectra at THz frequencies with high selectivity and resolution in applications fields as industrial quality inspection control, customs inspection and security screening [1]. Furthermore, THz radiation has no endangering effects on human beings and enables higher contrast for “soft matter” than x-rays. The explosives such as TNT, RDX, HMX and biochemical weapons decompose at atmospheric pressure making it possible to have the gases resulting from the decomposition to be used as signatures, since they absorb radiation in the THz range [2–5]. THz radiation is able to penetrate non conducting materials without producing molecules ionization. The devices using this new technology can play a great role in improving the security measurements and providing accurate data base worldwide for sensitive gas tracing. Those data are able to provide sophisticated and accurate automated detection systems capable of challenging the growing multi-threat in public transport and airports and for other industrial and medical applications. In order to identify these dangerous materials it is necessary to develop devices that are able to detect either the transmitted or reflected THz waves selectively. Fine tuning is required to scan busy spectral lines. Thus, in this paper we deliver simulations for a filter based on a photonic crystal with an embedded metamaterial [6]. These simula-tions can be combined, e.g. with simulations for Quantum Cascade Lasers [7, 8] or other promising solid state sources [9, 10] to help propose complete sets of robust and reliable tunable sources and detectors for security applications.

6.2 Design and Analysis

Our proposed structure as shown in Fig. 6.1, contains a pair of dielectrics which are SiO2 and Ta2O5 and a metamaterial of thickness L.

In this study, we consider a one-dimensional photonic crystal containing a meta-material layer as shown in Fig. 6.1. The photonic crystal is composed of an alternate

Fig. 6.1 One-dimensional photonic crystal including a metamaterial layer

H. Hamdouni et al.

41

distribution of SiO2 and Ta2O5 pairs, their thicknesses are respectively 9.96 and 6.6 μm and their refractive indices are 1.46 and 2.46. However the thickness of the metamaterial was set to 634 nm. The transfer matrix method [11, 12] was used to calculate the transmission of the proposed structure. The chromatic dispersion of the metamaterial is considered through its permittivity, which is given by εmetm = 1 − ωp

2/(ω2 − iνω). However the permeability is constant μmetm = 2.3, where ν stands for the losses due to collisions frequency and ωp is the plasma frequency, which is fixed at 2,100 THz [13]. The concept of operation of the device consists of using THz radiation containing distinct frequencies to scan either luggage or clothing and the reflected light with different frequencies will be monitored and detected. In second phase the detected frequencies will be identified and compared to explosive frequen-cies, the contrast indicates the presence or absence of explosives. Thus if a broadband source is used to generate the radiation to cover a broad range of explosives, selective filters as the one proposed here are necessary to allow for a selective excitation of a given spectral line and thus uniquely identify a given threat with a small error margin. This is crucial because most explosives have THz signatures in the same range as harmless materials and fine tune filtering is required. The probe light, a TE polarized wave, was incident in a direction perpendicular to the SiO2 layer.

6.2.1 Effect of the Length on the Transmission Band

In order to study the effect of the length on the optical properties of the proposed structure, we fix the number of period to N = 5 and we change the length; the results of our simulation are reported in Fig. 6.2.

It is clear from these figures that the transmission peak is located at around 5.17 THz. The transmission decreases when the length L increases. This is obvious in Fig. 6.2d, which shows the variation of the transmission versus L. The transmis-sion varies inversion proportional to L, the variation is rapid when L is less than 634 nm, thus in this range the proposed structure is very sensitive to the environ-mental medium. However, when L exceeds 634 nm the variation flattens. Through the rest of the paper the length is kept fix to 634 nm. From these results, the length influences the transmission intensity, however, the frequency of maximum peak doesn’t shift and remains located at the same position regardless the variation of L.

The miniaturization and integration of optical devices require a small size that can be achieved by reducing the number of periods in our case. We vary the number of dielectric pairs and see their effect on the transmission of spectrum. In Fig. 6.3 we report transmission variation with frequency when the number of period equals 6. We notice that the transmission peak is still located at f = 5.17 THz and the maximum is about 45 %. However, this maximum drops to 4 % when the number of period is set to 7 and the maximum is at 5.17 THz.

From these curves it is clearly obvious that the transmission decreases when the number of layer increases. This result seems interesting for the design of miniaturized devices with high transmission.

6 Design of Metamaterial Photonic Crystals for Explosives Detection

42

1

0.8

0.6

Tra

nsm

issi

onT

rans

mis

sion

Tra

nsm

issi

onT

rans

mis

sion

a b

dc

T=98.85%

T=11.79%

F=517 THz

F=517 THz

T=46.86%

F=517 TH z

L=534 nm

N=5

L=834 nm

N=5

L=634 nm

N= 5

0.4

0.2

0.2

0.1

0.15

0.05

0

0 0

1.0

0.8

0.6

0.4

0.2

0.0

500 600 700

L(nm)

800 900 1000

0.2

0.1

0.3

0.4

0.5

3 4 5

Frequency (THz)

Frequency (THz)

Frequency (THz)

6 7 8

3 4 5 6 7 8

3 4 5 6 7 8

Fig. 6.2 Transmission variation with frequency when (a) N = 5 and L = 534 nm, (b) N = 5 and L = 634 nm, (c) N = 5 and L = 834 nm and (d) transmission versus the length L

0.5a bT=41.77%

F=5.17THz

T=4.45%L=634 nm

N=6

L=634 nm

N=7F=5.17TH z0.4

0.3

0.2

Tra

nsm

issi

on

Tra

nsm

issi

on

0.1

0

0.05

0.04

0.03

0.02

0.01

03 4 5

Frequency (THz)

6 7 8 3 4 5

Frequency (THz)

6 7 8

Fig. 6.3 Transmission spectrum when L is fixed to 634 nm, (a) N = 6 and (b) N = 7

H. Hamdouni et al.

43

6.2.2 Effects of Gold Layer on the Spectrum Properties of the Structure

In this section a gold (Au) layer is introduced inside the structure reported in Fig. 6.1. The Au film is of a thickness L1. The structure with gold is shown in Fig. 6.4. The transmission is calculated using the transfer matrix method. The permittivity of metal is described as the following:

e w e

e ewt

swe

e c wswe

e e¥¥

¥( ) = +--

+ = + ( ) + = -¢ ²s

i i ii

1 0 0

,

(6.1)

where ω is the angular frequency, εs represents the static relative permittivity, τ is the relaxation time, σ is the conductivity, and ε′ and ε″ are the real and imaginary part of the experimental determined relative permittivity at the angular frequency ω [14].

In order to study its effect on the spectrum properties, the thickness L1 of the metal is varied. The length L of the entire structure is fixed to 950 nm, the transmis-sion is calculated and shown in Fig. 6.5. The Fig. 6.5a shows transmission with frequency when the thickness of Au layer is L1 = 18 nm, it is clear that the transmis-sion curve becomes narrower and its peak is slightly shifted to low frequency range that is located at f = 5.12 THz. This shift is due to the plasmon sensitivity towards geometry changes. We increase L1 to 20 nm, the peak remains located at the same frequency, however, the transmission maximum drops to around 79 %, this shows that as the structure length increases the transmission decreases, the length of the structure affects the spectrum of the device. This result agrees with what our previ-ous investigation. We keep increasing L, in this case it is set to 25 nm, and the result is reported in Fig. 6.5c it is obvious that the transmission maximum is significantly reduced compared to that shown in Fig. 6.5b, whilst the frequency doesn’t shift and remain located at f = 5.12 THz.

Figure 6.5c shows that the transmission maximum drops to 49.76 % which two time less than that obtained when L1 = 18 nm. In both cases which are structures

Fig. 6.4 Proposed structure involving the gold layer


44

with and without gold layer, the transmission is reduced when the device size increases. Our study shows that the proposed structure may be used in small circuit and may have potential applications.

6.3 Conclusion

In summary, we report the design of photonic crystal metamaterial structure that exhibits a spectral band from 3 to 8 THz. The results show that a high peak appeared in the band located at 5.17 THz. This frequency matches the absorption frequency of HMX explosive. By adding the gold layer, the transmission peak shifted towards low frequency, f = 5.12 THz. The optimization of the thickness of gold layer features interesting effects on designing devices with a large frequency range.

1

0.8

0.6

0.4

0.2

0.8

0.6

0.4

0.2

0

0

1

0.8

0.6

0.4

0.2

0

Tra

nsm

issi

on

Tra

nsm

issi

on

Tra

nsm

issi

on

a b

c

3 4 5Frequency (THz)

6 7 8

3 4 5

Frequency (THz)

6 7 8

3 4 5

Frequency (THz)

6 7 8

T=49.76 nm

F=5.12 TH z

T=98.24%

F=5.12 THz

T=79.01%

F=5.12 TH z

L=950nm

L1=18 nm

L=950 nm

L1=20nm

L=950 nm

L1=25 nm

Fig. 6.5 Transmission versus the frequency for different thicknesses of Au layer, (a) L1 = 18 nm, (b) L1 = 20 nm, (c) L1 = 25 nm

H. Hamdouni et al.

45

References

1. Wallin S, Pettersson A, Östmark H, Hobro A (2009) Laser-based standoff detection of explosives: a critical review. Anal Bioanal Chem 395:259–274

2. Federici JF, Schulkin B, Huang F, Gary D, Barat R, Oliveira F, Zimdars D (2005) THz imaging and sensing for security applications – explosives, weapons, and drugs. Semicond Sci Technol 20:S260–S280

3. Huang F, Schulkin B, Altan H, Federici J, Gary D, Barrat R, Zimdars D, Chen M, Tanner D (2004) Terahertz study of 1,3,5-trinitro-s-triazine (RDX) by tune domain spectroscopy and FTIR. Appl Phys Lett 85:5535

4. Oxley JC, Smith JL, Shinde K, Moran J (2005) Determination of the vapor density of triace-tonetriperoxide (TATP) using a gas chromatography headspace technique. Propellants Explos Pyrotech 30:127

5. Oxley JC, Smith J (2006) Peroxide explosives. In: Schubert H, Kuznetsov A (eds) Proceedings of the NATO advanced workshop on detection and disposal of improvised explosives, NATO security, though science series –B: physics biophysics. Springer, Dordrecht, pp 113–121

6. Fang YT, Zhou J, Pun EYB (2007) High-Q filters based on one- dimensional photonic crystals using epsilon-negative material. Appl Phys B 86:587

7. Pereira MF Jr, Nelander R, Wacker A, Revin DG, Soulby MR, Wilson LR, Cockburn JW, Krysa AB, Roberts JS, Airey RJ (2007) Characterization of intersubband devices combining a nonequilibrium many body theory with transmission spectroscopy experiments. J Mater Sci Mater Electron 18:689

8. Schmielau T, Pereira MF (2009) Nonequilibrium many body theory for quantum transport in terahertz quantum cascade lasers. Appl Phys Lett 95:231111

9. Pereira MF Jr (2008) Intervalence transverse-electric mode terahertz lasing without population inversion. Phys Rev B78:245305

10. Pereira MF, Tomić S (2011) Intersubband gain without global inversion through dilute nitride band engineering. Appl Phys Lett 98:061101

11. Čtyroký J, AbdelMalek F, Usbeck K, Eck W (1999) Modeling of surface plasmon resonance waveguide sensors with Bragg gratings. Opt Quantum Electron 31:927

12. Yariv A, Yeh P (1988) Optical waves in layered media. Wiley, New York 13. Rao XS, Ong CK (2003) Amplification of evanescent waves in a lossy left-handed material

slab. Phys Rev B 68:113103 14. Kunz KS, Luebbers RJ (1993) The finite difference time domain method for electromagnetics.

CRC Press/LLC, Boca Raton, pp 123–162



Abstract Results of research of resonant phenomena in planar periodic structures consisted of dielectric elements are presented. An existence of high Q-factor trapped mode resonances is revealed in structure which a periodic cell is composed of two dielectric bars with different lengths or width. Besides higher Q-factor essential dif-ference of studied structures in contrast to the planar structure with a single dielec-tric bar in the unit cell is appearing of a great red shift of a trapped mode resonant wavelength. This property is very attractive to design a new type high-Q dielectric metamaterials for THz and Infrared.

Keywords Trapped mode • Fano resonance • All-dielectric metamaterials • THz and infrared diffraction

7.1 Introduction

Planar double-periodic structures are widely used in microwave, terahertz and optical wavelengths because of their manufacturing simplicity and the multiplicity of remark-able electromagnetic properties of such surfaces. In microwaves, these structures are used as absorbing and scattering covers, frequency-selective and polarization- selective surfaces [ 1 ]. The planar double periodic structures are also a basis of metamaterials,

Chapter 7 New Type High-Q THz Planar All-Dielectric Metamaterial

Vyacheslav V. Khardikov and Sergey L. Prosvirnin

V. V. Khardikov (*) Institute of Radio Astronomy of NASU , Kharkiv , Ukraine

School of Radio Physics , Karazin Kharkiv National University , Kharkiv , Ukraine e-mail: [email protected]

S. L. Prosvirnin Institute of Radio Astronomy of NASU , Kharkiv , Ukraine e-mail: [email protected]

48

i.e. artifi cial materials engineered to have properties which are not found in natural media [ 2 ]. The planar metamaterials (also known as metafi lms) are an impressive modern object, which is driven by certain fascinating facilities such as, e.g. the anom-alous refl ection and refraction [ 3 ], the unusual asymmetric transmission [ 4 ], and light refl ection that does not change the phase of the incident wave [ 5 ]. All mentioned above and many other exciting metafi lm properties are extremely attractive for devel-oper of new types of effective and robust THz components.

The metafi lms are usually planar periodic structures of complex shaped metal strips. The complex shape of metal strips allows metafi lms to attain features of resonance transmission and refl ection of incident wave without any high diffraction orders form-ing. Unfortunately Q-factor of the resonances excited in such structures are extremely limited because a high level of radiation losses and energy dissipation in metals. On the other hand there are numerous important applications of metalayers which require a strong confi nement of electromagnetic fi eld inside the periodic structure.

The usual resonance fi eld enhancement inside a planar metamaterial may be greatly increased by involving structures which bear the so-called trapped modes [ 6 , 7 ]. The excitation of the trapped mode resonances in planar double periodic structures with broken symmetry was found both theoretically [ 8 , 9 ] and experimentally [ 10 , 11 ] in microwaves. The radiation losses decreasing in such materials are reached through some destructive interference of the radiation by the antiphased currents in metallic elements of a subwavelength periodic cell. Now, the existence conditions and the spec-tral properties of trapped mode resonances are investigated in detail in the suitably structured planar plasmonic metamaterials developed for the near-IR range [ 12 , 13 ].

However, the high level of THz and IR radiation dissipation in metals remains a major limiting factor for fi eld enhancement in such type meta-layers. Fortunately there is a way of signifi cantly decreasing of energy dissipation. This way lies in using of all-dielectric metalayers supporting the trapped-mode resonances. The existence of such structures for near IR was recently shown in [ 14 ]. The fi eld enhancement level in all-dielectric metamaterials in trapped mode regime is signifi -cantly higher than in a plasmonic one.

This paper is aimed involving in the family of the high-Q trapped mode planar metamaterials for THz and Mid Infrared the new and highly desirable low-loss structures made up of entirely dielectric elements.

7.2 Numerical Results: Solution Analysis

The resonances in the plasmonic and dielectric structures have a considerably different nature. A plasmon–polariton excitation propagating along metallic surfaces replicates their shape. Therefore, the complex-shaped metallic elements can be used to provide a resonant interaction of light with a periodic structure in a deeply sub-wavelength range. On the contrary, complications of the shape of dielectric elements do not involve a substantial increase of the resonant wavelength, like in the case of metallic elements. In the case of all-dielectric metamaterials the dielectric elements

V.V. Khardikov and S.L. Prosvirnin

49

of array act as open dielectric resonators. Thus, one requires using a dielectric with relatively high refractive index as a material for elements of the subwavelength dou-ble periodic planar array. A semiconductors are very promising materials for manu-facturing of periodic array elements because they are widely using in microchip technologies, i.e. they are favorable for metafi lms manufacturing, and have transpar-ency windows in the terahertz and infrared regions. Within transparency windows the semiconductors act as a good dielectric with small dielectric loss tangent, which does not exceed 10 –3 . In this paper the germanium and silicon are considered as a suitable material for the array elements. The silicon and germanium transparency windows lie within 1.2–14 μm and 1.9–18 μm, respectively, and their refractive indexes vary from 3.4 to 3.6 and from 3.9 to 4.2 through these windows [ 15 ].

As it was shown in [ 14 ] the trapped mode resonance may be excited in a planar dielectric structure with two asymmetric dielectric bars in a periodic cell. Thus the problem of light diffraction by double-periodic planar array of dielectric parallel-epipeds placed on a dielectric substrate with thickness L is considered (see Fig. 7.1 ). A periodic cell of the array includes a pair of dielectric parallelepipeds which have different lengths or width but the same height ( L a ) and materials. The length of dielectric elements are h 1 and h 2 and their widths are t 1 and t 2 . Sizes of the square periodic cell are chosen to be d x = d y = d . The periodic cell is symmetric relatively to line which is parallel to axis Oy and passing through its centre. The normal inci-dence of linearly x-polarized plane wave is considered. The substrate material is assumed to be synthetic fused silica (amorphous silicon dioxide).

To solve the diffraction problems, the numerical method proposed in [ 16 ] is used. This method is based on the both mapped PSTD method [ 17 ] and transfer matrix theory. In this paper the dispersion of dielectrics does not take into account. However the materials dispersion may be taken into account without any problems in the context of used method.

Fig. 7.1 A sketch of the unite cell of the double-periodic planar structure. The all-dielectric array composed of two dielectric bars per a periodic cell is placed on the synthetic fused silica layer

7 New Type High-Q THz Planar All-Dielectric Metamaterial

50

7.2.1 The Germanium High-Q Planar Metamaterial for Near Infrared

Let us consider the planar double periodic structure with a pair of asymmetric germanium bars in the periodic cell. The structure asymmetry is achieved by using the dielectric bars with different length. The longer bar length is h 2 and the shorter h 1 . The structure parameters are chosen as follows d = 975 nm, t 1 = t 2 = L a = 195 nm, h 2 = 877 nm and L = ∞ . The distance between the longer and the shorter bars of the pair is 195 nm. The refractive index of germanium and fussed silica are assumed to be equal to 4.12 and 1.44 respectively that correspond to the wavelength range from 1,850 to 1,950 nm.

Figure 7.2 illustrates the wavelength dependence of the refl ection coeffi cient of the arrays with different length h 1 . If two bars of different length are combined in the periodic cell, some additional refl ection resonances are excited (see Fig. 7.2 , lines 2–3). It is a typically trapped mode resonance and has a shape inherent to Fano- shape sharp resonance with a specifi c wavelength dependence of the refl ection coef-fi cient rolled over trough to peak.

Let us notice that Q-factor of trapped mode resonance shown in Fig. 7.2 are equals to 200 and 1,100 that much bigger than one may expect in the case of plas-monic metamaterial using. The Q-factor increase with decreasing of the structure asymmetry as it was predicted theoretically for non-dissipative structures.

Besides the enhanced Q-factor, the main distinctive feature of the trapped mode resonance of the two-element dielectric array is a great red shift of its wavelength relative to the resonant wavelengths of the corresponding one-element arrays (see Fig. 7.2 line 1). Thus the coupling between the dielectric bars of the two-element periodic array results in an extremely large increase of the resonant wave-length. This property of the entirely dielectric trapped mode arrays is quite important in view of possible applications in the fi eld of all-dielectric metamaterials. First, the ratio of the array pitch to wavelength may be decreased to design more homogeneous

Fig. 7.2 The wavelength dependences of the refl ection coeffi cients of the arrays of paired germanium bars. h 2 = 877 nm. Lines 1 , 2 and 3 correspond to h 1 = 877 nm, h 1 = 838 nm and h 1 = 780 nm respectively


51

metamaterials. Second, the increase of the resonant wavelength results in an enhancement of the confi nement of the fi eld intensity due to a decrease of radiation losses. It is important to the design of nonlinear and active artifi cial media.

Moreover it is evident that a decrease of the asymmetry degree characterized by the value h 1 / h 2 results in an increase of the red shift of the trapped mode resonance (see Fig. 7.2 lines 2 and 3).

7.2.2 The Silicon High-Q Planar Metamaterial for THz Range

Let us consider the planar double periodic structure with a pair of asymmetric Silicon bars in the periodic cell. The structure asymmetry is achieved by using the dielectric bars with different width. The structure parameters are chosen as follows d = 0.9 mm, h 1 = h 2 = 0.8 mm, t 1 = 0.16 mm, t 2 = 0.26 mm, L a = 0.36 mm and L = 1 mm. The distance between the longer and the shorter bars of the pair is 0.24 mm. The refractive index of the silicon and quartz, which used as structure substrate, are assumed to be equal to 3.4 and 2 respectively that correspond to the wavelength range from 1.8 to 2.1 mm.

Figure 7.3 illustrates the wavelength dependence of the refl ection coeffi cient of the considered metalayer. In this case trapped mode resonance appears as additional refl ection resonances against the background of interference resonances in substrate (see Fig. 7.3a ).

It is clear that the silicon and quartz dissipative losses will strong effect on the properties of the trapped mode resonance. To evaluate such infl uence a model of dielectric with frequency-dependent losses was used for Silicon and quartz. One can see that the real losses in dielectric and semiconductor result in decreasing of trapped mode resonance amplitude (see Fig. 7.3b ). It needs to notice that chosen

Fig. 7.3 The wavelength dependences of the refl ection coeffi cients of the arrays of paired silicon bars in wide range (a) and at the trapped mode resonance wavelength (b). It is assumed that dielectric loss tangent of the silicon and quartz have values tg δ Sil = tg δ Q = 0 ( line 1 ) and tg δ Sil = 3 ⋅ 10 − 3 , tg δ Q = 0.9 ⋅ 10 − 3 ( line 2 )

7 New Type High-Q THz Planar All-Dielectric Metamaterial

52

values of the silicon and quartz dielectric loss tangent are signifi cantly bigger that occur in practice. Moreover it is evident that the degree of the real losses infl uence on the trapped mode resonance is strongly depends on Q-factor of resonance, i.e. on the structure asymmetry.

Thus the existence of the high-Q trapped mode resonance has been shown in low-loss entirely dielectric structures. One can design highly desirable deep- subwavelength low-loss planar metamaterials in the wide range from terahertz to near-infrared wavelength using the semiconductors.

References

1. Munk BA (2000) Frequency selective surfaces: theory and design. Wiley, New York 2. Smith DR, Pendry JB, Wiltshire MCK (2004) Metamaterials and negative refractive index.

Science 305:788–792 3. Yu N et al (2011) Light propagation with phase discontinuities: generalized laws of refl ection

and refraction. Science 344:333–337 4. Fedotov VA, Mladyonov PL, Prosvirnin SL, Rogacheva AV, Chen Y, Zheludev NI (2006)

Asymmetric propagation of electromagnetic waves through a planar chiral structure. Phys Rev Lett 97:167401

5. Schwanecke AS, Fedotov VA, Khardikov VV, Prosvirnin SL, Chen Y, Zheludev NI (2007) Optical magnetic mirrors. Opt J A Pure Appl Opt 9:L1–L2

6. Stockman MI, Faleev SV, Bergman DJ (2001) Localization versus delocalization of surface plasmons in nanosystems: can one state have both characteristics? Phys Rev Lett 87:167401

7. Liu HC, Yariv A (2009) Grating induced transparency (GIT) and the dark mode in optical waveguides. Opt Expr 17:11710–11718

8. Prosvirnin SL, Zouhdi S (2001) Multi-layered arrays of conducting strips: switchable photonic band gap structures. Int J Electron Commun (AEU) 55:260–265

9. Blackburn JF, Arnaut LR (2004) High performance split ring FSS for WLAN bands. In: Proceedings of the 27th ESA antenna technology workshop on innovative periodic antennas: electromagnetic bandgap, left-handed material, fractal and frequency selective surfaces, 2004, pp 329–336

10. Fedotov VA, Schwanecke AS, Zheludev NI, Khardikov VV, Prosvirnin SL (2007) Asymmetric transmission of light and enantiomerically sensitive plasmon resonance in planar chiral nano-structures. Nano Lett 7(7):1996–1999

11. Fedotov VA, Rose M, Prosvirnin SL, Papasimakis N, Zheludev NI (2007) Sharp trapped-mode resonances in planar metamaterials with a broken structural symmetry. Phys Rev Lett 99:147401

12. Khardikov VV, Iarko EO, Prosvirnin SL (2010) Trapping of light by metal arrays. J Opt 12:045102

13. Dong ZG et al (2010) Plasmonically induced transparent magnetic resonance in a metallic metamaterial composed of asymmetric double bars. Opt Expr 18:18229–18234

14. Khardikov VV, Iarko EO, Prosvirnin SL (2012) A giant red shift and enhancement of the light confi nement in a planar array of dielectric bars. J Opt 14:035103(7)

15. Li HH (1980) Refractive index of silicon and germanium and its wavelength and temperature derivatives. J Phys Chem Ref Data 9:561–658

16. Khardikov VV, Iarko EO, Prosvirnin SL (2008) Using of transmission matrixes and pseudo-spectral method in time domain to investigate light diffraction on planar periodic structures. Radiophys Radioastron 13:146–158

17. Gao Xian. Mirotznik MS, Shi S, Prather DW (2004) Applying a mapped pseudospectral time-domain method in simulating diffractive optical elements. J Opt Soc Am A 21(5):777



Abstract Development of compact low-cost sources of a coherent radiation with a high efficiency in the range of several THz remains one of the key challenges of the modern security technologies. In spite of rather impressive achievements in inter-subband quantum cascade lasers their current parameters are still far from the needs of practical implementation. We present here a bird-eye view on the current state of the field. We compare theoretical prospects of the optimal gain of lasers in the range of few THz for two cases: intersubband GaAs-based quantum cascade lasers and interband laser based on coupled quantum wells InAs-GaSb. We show that the last case promises the gain that could be three orders of magnitude higher.

Keywords TeraHertz • Nanostructures • Bipolar lasing • Quantum cascade lasers

8.1 Introduction

The very attractive idea of harmless THz based security checks still remains a dream. The key challenge on the way of its practical implementation is a lack of efficient sources of coherent radiation in the range of 1–5 THz. The best achieve-ments in this field are probably presented by GaAs-based quantum cascade lasers [1]; however they are still limited by relatively low gain. This is the most evident reason for appearing of new concepts and new materials. In this paper we present our analysis of the key physical factors limiting the efficiency of nanostructure- based THz lasers. We believe that for intersubband monopolar lasing these restrictions are of fundamental nature. We also present a comparison of our

Chapter 8Bipolar THz-Lasing Structures Based on InAs-GaSb Coupled Quantum Wells and Their Potential for Security Checks

L.D. Shvartsman and Boris Laikhtman

L.D. Shvartsman (*) • B. Laikhtman Racah Institute of Physics, Hebrew University, Jerusalem 91904, Israele-mail: [email protected]

54

estimations of gain in two scenarios: existing GaAs-based quantum cascade lasers (QCL) [1] and proposed interband lasers based on coupled quantum wells InAs-GaSb [2, 3]. We present an ideology of such a comparison of two different systems that is reduced to the following. The most typical design of intersubband GaAs-based QCL is compared with a typical InAs-GaSb coupled quantum well laser (CQWL) designed for the same frequency range. Our comparison based on detailed density matrix calculations shows that the maximal possible gain for CQWL can be three orders of magnitude higher than for QCL while from the technological point of view QCLs, of course, still look much more attractive. Thus the problem is reduced to the old dilemma “monopolar intersubband lasing” vs. “interband bipolar lasing”.

This problem may be also considered in a wider context. Historically, the domi-nant tendency in the development of semiconductor lasers was the following: lower-ing the frequency from visible light to infrared and further on to THz, resulted in evolution from interband lasing to the intersubband one. It happened because of very obvious reason: typical semiconductor gaps are much higher than, say, THz photon energy. Nevertheless, our theoretical analysis shows that it might be reason-able to consider an option of “going back” to interband lasing at least for the case of few-THz spectral range. (May be it is worth to notice in this context that even for MIR range alike alternatives showed recently a considerable promise [4]). In the following subsection of this introduction we list the key physical factors affecting the gain value and formalize the problem.

We adopt the following structure of this paper:

In the second section we describe the main physical factors defining the gain of nanostructure-based THz lasers and present the expressions for gain both for intersubband and interband lasers.

In the third section we analyse the LO-phonon scattering and its limitations affect-ing the depopulation selectivity for single QW-based intersubband laser in the THz range.

In the fourth section we briefly review two major designs for InAs-GaSb coupled quantum wells-based bipolar interband lasers and explain the physics of the appearance of THz gaps in each scenario. The relevant physical restrictions of their efficiency in THz range are analysed as well. We also analyse the resulting expression for gain of nanostructure-based THz lasers and compare theoretical predictions for GaAs-based intersubband lasers and CQWs InAs-GaSb based interband lasers.

The following Sect. 8.5 compares the more realistic design of GaAs-based QCLs with the major single QW intersubband laser. Additional time-scales introduced by spatial separation of lasing and LO-phonon emission substantially improve the situation, however from the point of view of high gain the interband CQW lasers still have much better promise. The methodology of our analysis of the basic gain restrictions of THz GaAs-based QCLs is based on density matrix calculations, and it reveals the most general relevant time scales that limit the population inversion for nearly any design.

L.D. Shvartsman and B. Laikhtman

55

8.2 Key Physical Factors Affecting the Gain of THz Lasing Heterostructures

Intersubband lasing essentially involves at least three nearly parallel subbands. Say, the lasing itself occurs between the initial subband, “i”, and the final subband, “f”, while from the subband “f” the carriers are removed by the resonant optical phonon emission. The gain in this case may be expressed as:

ge

m

M

n cLn n

E Eisb

if

effi f

i f

= -( )- -( ) + ( )

8 2

02

2

2 2

pw

G t

w t

/

/

(8.1)

where ni and nf are electron concentrations at the initial and final lasing levels, Ei and Ef are the energies of the levels and Mif is the matrix element of transition between them.

For the interband case the expression for the gain can be written in a similar structure as well:

ge

m

M

n cL

mf f Eib

cv

effe h g= ( ) + ( ) -éë ùû -( )

2

02

2

2

41

wG p

e e J ww w

(8.2)

where m0 is the free electron mass, Γ is the optical confinement factor, L is the width of the quantum well where the radiative transitions take place, Mcv is the matrix ele-ment of the electron momentum between the conduction and valence band, neff is the effective refractive index, εeω and εhω are the energies of an electron and hole partici-pating in the optical transition, m is reduced mass of electron and hole, and f are their distribution functions.

The last factors in both expressions represent the reduced density of states. Therefore the alternative “intersubband” vs “interband” lasing, in cases when such an alternative exists, can be basically reduced to the interplay between the following physical factors:

– radiative matrix elements; – reduced densities of states; – possibility to maintain a population inversion.

The radiative matrix elements in the list above obviously favor an interband las-ing. Indeed, the matrix element of the electron momentum is of the order of ћ/l where l is the smallest spatial size of the electron wave functions. Indeed, for inter-band transitions the matrix element Mcv is calculated between the Bloch amplitudes of the conduction and valence band and the smallest spatial size above is of the order of the lattice constant a. In the case of intersubband transitions the momentum matrix element is calculated between the envelop functions in the well or superlat-tice, and the smallest length scale is of the order of L/j, where L characterizes the extension of the wave function, and j is the number of the level. That is Mcv/Mif ~ L/ja. In III-V compounds a ≈ 0.5 nm. Reduction of the radiation frequency inevitably

8 Bipolar THz-Lasing Structures Based on InAs-GaSb Coupled Quantum Wells…

56

leads to increase of L. On the other hand, it is impossible to use very high levels because electron confinement there is not so good. In fabricated so far THz intersub-band lasers L ~ 30–50 nm and j = 3–5. That is Mcv/Mif ≥ 12 and the ratio of the gains contains the ratio of the matrix elements squared.

As for reduced density of states the advantage of an intersubband lasing is rather obvious. Because of the subband parallelism the reduced density of states is nearly sin-gular δ(Ei − Ef − ħω). Finite life time results in a broadening of this singularity (the last factor in (8.2)), and the maximal achievable intersubband gain may be expressed as:

ge

m

M

n cLn nisb

if

effi f= -( )8 2

02

2

pw

G t

(8.3)

Thus, the dilemma “intersubband lasing vs interband lasing” seems to be reduced to the choice “radiative matrix element” vs. “reduced density of states”. Interband lasing promises higher matrix element while an intersubband one promises higher reduced density of states because of slightly broadened nearly parallel subbands. However, in THz range the situation becomes less trivial since: 1. an actual broad-ening may become comparable with an emitted frequency, and 2. for the type II structures an interband matrix element is considerably reduced because the transi-tion becomes indirect in real space. Then, the last factor, the prospect to maintain a high level of population inversion, plays the key role in this comparison.

To maintain the population inversion in intersubband lasing, one employs the LO-phonons. Hence, depopulation selectivity becomes a crucial issue. It is consid-ered in the following sections.

8.3 LO-Phonon Scattering and Its Influence on Depopulation Selectivity

For the case of QW intersubband lasing the depopulation of the final state “f” essen-tially entails the LO phonon emission. In the THz range such a mechanism cannot provide good enough depopulation selectivity. It means that if the phonon assisted scattering rate of the state “f” is tuned to be high enough, the phonons are going to cause rather efficient depopulation of the initial state “i” as well. It challenges the pumping process and puts rather significant limitations on the achievable gain of intersubband lasing in THz range.

In the Fig. 8.1 we present our calculations of the LO-phonon scattering for single QW Al0.15Ga0.85As-GaAs-Al0.15Ga0.85As. There are various kinds of LO-phonons participating in the scattering. We adopt the following classification:

– Quantum well phonons, i.e. LO GaAs phonons; – Interface phonons that originate at each interface GaAs-Al0.15Ga0.85As; – Barrier phonons of Al0.15Ga0.85As that essentially combine both the phonons of

GaAs nature and the phonons of AlAs nature.


57

The resulting output of each kind of phonons depends on the particular design but the total width of the phonon peaks is always comparable or higher than the THz frequencies (Fig. 8.1). It makes sense to notice that for more advanced designs including several quantum wells or chirped superlattices LO-phonon scattering involves even more kinds of phonons. For instance, for the typical design shown in the Fig. 8.1 the scattering process in the right QW will be influenced by the phonons originated from each additional remote interface, i.e.: the interfaces of the left QW. Our estimations show that the widths of these additional peaks of scattering can be of the same order of magnitude.

These essential features of LO-phonon scattering have affect directly the maxi-mal value of gain of intersubband THz lasers . The factor of population inversion in (8.3) can be expressed as:

n n Ji f if

if

- = -æ

èçç

ö

ø÷÷t

t

t1 0

(8.4)

where J is the pumping, 1/τi = 1/τi0 + 1/τif. Here τif corresponds to electron-electron scattering that is usually an order of magnitude longer than LO-phonon assisted times τi0 and τf0. Therefore τi is close to τi0 and to τf0. Thus, the total value of popula-tion inversion is limited from above, and this limit can be estimated as:

J t ee 4/ .

This limitation of the intersubband gain can be overcome basically by the two ways considered below: decoupling in space the process of lasing and LO-phonon emission (the QCL scenario) or considering the interband lasing as an alternative. Let us start our consideration from the interband lasing.

Fig. 8.1 LO-phonon emission rate in the single QW and typical QCL design: lasing and phonon emission are spatially separated


58

8.4 Interband THz Lasing in InAs-GaSb CQWs

In InAs-GaSb CQWs-based structures the lasing mechanism is different. Here we actually have an interband lasing in a type II heterostructure where InAs conduction band and GaSb valence band have an initial overlap. Its value in the bulk is around 150 meV. The required THz gaps may be arranged by two various ways:

1. For narrow enough QWs, when the overlap is removed by size quantization. the in-plane dispersion law in this case resembles V for InAs electrons or inverted V for GaSb holes. We further call it the V-structure. The simplest V-structure that can be used for terahertz lasing consists of adjacent InAs and GaSb wells embed-ded in between AlSb that presents a good barrier for both electron and holes. Playing with the widths of InAs and GaSb wells, it is possible to tune the separa-tion between the ground electron level in InAs well, e1, and ground hole level in the GaSb well, hh1, from negative values of more than 100 meV to positive values of the same order of magnitude. That is, in principle, the separation between e1 and hh1 levels can be adjusted for generation of radiation from a few micrometers to hundreds of micrometers, i.e. to the THz range as well (insert “a” in the Fig. 8.2).

2. For wider QWs the initial overlap is not removed, but the hybridization gap is formed [5]. This gap naturally lies in the THz range [2, 3]. The in-plane dispersion law in this case resembles W electrons or inverted W holes, and we call “the W-structure” (insert “b” in the Fig. 8.2).

Fig. 8.2 Interband lasing in THz range gives much higher gain promise


59

In the Fig. 8.2 we present our calculation of the ratio of gib/gisb for the V-structure as a function of pumping. These calculations [2] are based on the Eqs. (8.1, 8.2, 8.3, and 8.4). In spite of the fact that the interband radiative matrix element is reduced by the indirect in real space optical transition in InAs-GaSb coupled quantum wells the interband gain still has almost three order of magnitude better promise than the intersubband gain for single GaAs based quantum well. This huge advantage results mostly from the higher interband matrix element and problems in depopulation selectivity for intersubband THz lasers.

For the W-structure, the situation can be even further improved if the design is sophisticated enough. Because of substantial for the large in-plane momentums anisotropy of hole dispersion in GaSb QW, the energetic positions of above men-tioned hybridization gaps can significantly differ for various in-plane directions [5] (Fig. 8.4). This realization is not the best for THz lasing. That is why the thickness of GaSb QW has to be small enough. For narrow enough GaSb QW, the hybridiza-tion gap opens within the pocket of light holes in the ground valence subband when the in-plane anisotropy is negligibly small [6]. In this case that has the best promise for THz lasing, the arising W-dispersion for both holes and electrons leads to a peculiarity in the reduced density of states. Thus, for interband lasing in W-design one can uniquely combine both the advantage of the high interband matrix elements and the high reduced density of states. Our calculations show that this W-design can provide an additional factor of 2.5 in the value of gain for properly chosen param-eters of InAs-GaSb CQWs.

One more opportunity to combine the advantages of interband and intersubband lasing in InAs-GaSb coupled quantum wells is to arrange the k-space cascade shown in the Fig. 8.2a, b for both the V-structure and the W-structure. Our simulations show that if in the InAs QW to set up a certain step one can arrange the intersubband distance between e3 and e2 to be in resonance with the interband distance between e1 and hh1. In this scenario the same carrier initially pumped into e3 can perform the following jumps: lasing e3->e2, LO phonon emission e2->e1, lasing e1->hh1. We suggest the following design as an example: AlSb barrier/(AlSb)(InAs) step of 349 A/InAs QW of 274A/GaSb QW of 18 A/AlSb barrier.

8.5 Intersubband Lasing Based on Adjacent QWs: Factors Limiting Efficiency of QCLs in THz Range

To complete our analysis let us consider a more realistic design of two adjacent QWs that is a basis of QCLs. The most common design of THz working QCL employs the following idea: Two key processes, photon emission and resonant pho-non emission have to be spatially separated though coupled by tunneling. Therefore any design essentially has to include two parts: the left part for optical emission that supplies the required THz gaps and the right part that supplies subband structure supporting renonant optical phonon emission thus responsible for selective


60

depopulation. The simplest design of this kind is presented in the Fig. 8.1, where THz lasing takes place in the left QW between the subbans 4 and 3, while LO pho-nons emitted in the right QW between the subbands 2 and 1. Wells are separated by the barrier that has to be large enough to suppress unwanted phonon emission from the upper lasing state 4 to the state 1 but transparent enough to assure an effective tunnelling process 3->2. In actual designs both left “photon” part and right “pho-non” part may be realized by more than one quantum wells or by chirped superlat-tices, and therefore more quantum subbands may be involved. Nevertheless, the basic idea stays the same.

We consider the key processes in QCLs in the following manner:The system that can be at levels: “left”, “right”, “ground”. By the state “left” we

can consider either the initial lasing state or the final lasing state in the left QW (Fig. 8.3). Statistical state of the system is described by the (3 × 3) density matrix. Initially the system is in an excited state, i.e., the only non-zero off-diagonal terms are those that couple the states “left” and “right”. An interaction of the system with a thermal bath presents a relaxation mechanism that induces transitions from level “right” to the ground state level while transitions from the state “left” to the ground state can be neglected (Fig. 8.3). The relaxation mechanism leads to depopulation of level “right”, i.e., relaxation of ρrr and destruction of the coherence of the system state, i.e., relaxation of elements ρrl and ρlr.

Detailed calculations show that spatial separation of lasing and phonon emission substantially improves the gain in comparison with the single QW. This improve-ment can be expressed by the following factor “r”:

ree ee

=+

-æ

èç

ö

ø÷

4142

42

32tt t

tt

The time scales τ42 and τ32 depend on both the LO-phonon emission time and the period of Rabi oscillations between the wells. The typical dependence of this factor

Fig. 8.3 Key physical processes in QCLs: tunneling couplings F32 and F42 are of the same order for THz lasing


61

on the tunneling coupling strength between the well is shown in the Fig. 8.5. One can see that with the lowering of the lasing frequency this improvement factor goes down and requires much more delicate tuning. In any case this improvement by the factor of 1.5 or 2 does not compete with the advantage of two-three orders of mag-nitude that interband lasing promises.

145

140

135

130

125

120

115

120

110

100

90

80

70

60

50

-1.2-1.0-0.8-0.6-0.4-0.2 0.0 0.2

k vector [π/Lv]k vector [π/Lv]

0.4 0.8 1.0 1.20.6

Ene

rgy

[meV

]

Ene

rgy

[meV

]

-0.6-0.5-0.4-0.3-0.2-0.1 0.1 0.2 0.3 0.4 0.5 0.60.0

<11> <11><01> <01>

Fig. 8.4 The influence of valence subband anisotropy on the position of hybridisation gap

Fig. 8.5 The improvement in the gain for QCL design vs. tunneling coupling


62

8.6 Conclusion

Though historically, the reduction of the lasing frequencies from visible light to near and mif-IR led to domination of intersubband lasing over interband, the further decrease of frequencies to THz favors the reverse process, especially, if the high gain is targetted. In intersubband THz lasing the problems with selective depopula-tion and low matrix element are of basic nature and weakly depend on the details of the design. The usage of resonant tunneling scheme in QCL may improve the gain by the factor of two while the CQW InAs-GaSb-based interband alternative gives the 2–3 order of magnitude better theoretical promise.

Acknowledgments We acknowledge the support of Yissum, the technology transfer company of the Hebrew University.

References

1. Kumar S, Williams BS, Kohen S, Hu Q, Reno JL (2004) Continuous-wave operation of terahertz quantum-cascade lasers above liquid-nitrogen temperature. Appl Phys Lett 84:2494

2. Shvartsman LD, Laikhtman B (2008) InAs-GaSb laser: prospects for efficient terahertz emission. Appl Phys Lett 93:131104

3. Laikhtman B, Shvartsman LD (2005) THz emitter based on InAs/GaSb coupled quantum wells: new prospects for THz photonics. In: Proceedings of SPIE, vol 5727. SPIE, Bellingham, p 54

4. Belenky G, Donetsky D, Kipshidze G, Wang D, Shterengas L et al (2011) Properties of unrelaxed InAs1-xSbx alloys grown on compositionally graded buffers. Appl Phys Lett 99:141116

5. de Leon S, Shvartsman LD, Laikhtman B (1999) Band structure of coupled InAs-GaSb quantum wells. Phys Rev B60:1861

6. Shvartsman LD (1983) Subband structure and kinetic characteristics of thin films of gapless semiconductors. Solid State Commun 46:787



Abstract Silicon doped with group-V donor atoms such as phosphorus is a particularly interesting material system from the point of view of quantum control. Experiments demonstrating coherent optical control of this donor center with THz laser light have recently been demonstrated, and are discussed here. The prospect for achiev-ing quantum control in other silicon-based materials with mid-infrared laser light, such as the silicon-chalcogen donor system, is also explored.

Keywords TERA-MIR*2012 • Quantum control • Silicon • Phosphorus • Sulfur • Selenium • Terahertz • Mid-infrared • Transient pump-probe spectroscopy • Photon echo

9.1 Introduction

Quantum computing is an expression that has made its way into common scientifi c usage. Why is this the case? For some, its emergence can be traced back to a proposal by Shor, that quantum computation would make the task of factorizing large num-bers tractable [ 1 ]. This could potentially compromise the integrity of public- key encryption systems commonly used to communicate sensitive data such as bank transactions. For others, however, quantum computing is merely an inevitable con-sequence of Moore’s Law. As the number of transistors on an integrated chip dou-bles every 18 months, so the power needed to switch each of these transistors needs

Chapter 9 Coherent Quantum Control of Donor States in Silicon with THz and MIR Light: A Route Towards a Scalable Quantum Computing Architecture

Stephen A. Lynch

S. A. Lynch (*) School of Physics and Astronomy , Cardiff University , Queen’s Buildings , The Parade, Cardiff CF24 3AA , UK e-mail: [email protected]

64

to fall. It has been predicted that by 2020 the number of electrons needed to switch a transistor will fall to just one. The consequence of this prediction is that electronics will be operating in the quantum regime irrespective of the driving application. Should we blindly follow the semiconductor industry’s roadmap and trust to new technological tricks, or do we seek a radically new technology: a wholly novel quantum component to operate alongside existing silicon components? Given the obvious bottleneck imposed by single electron switching, the later approach seems to make more sense.

9.1.1 Why Is Silicon Attractive for Quantum Computing?

There have already been some elegant proof-of-principle demonstrations of quantum computation using exotic systems such as ultra cold ions [ 2 , 3 ]. According to the now famous DiVincenzo criteria, however, one of the chief design characteristics of a realistic quantum logic gate is that the architecture must be scalable [ 4 ]. Quantum devices are nearly always nanoscale devices, so that it is diffi cult to imagine how such scalability can be achieved through current lithographic techniques without using the solid state. It is also a certainty that any quantum chip will eventually need to talk to the outside world, which will necessitate integrating quantum components with existing classical electronic control elements. This means there are good argu-ments for trying to realize any novel quantum component directly in silicon because we can exploit the enormous nano-scale fabrication know-how of the silicon semi-conductor industry. Silicon can also be made purer than any other man-made mate-rial. It is for this reason that the Avogadro Project set out to redefi ne the kilogram by using mono-isotopic single crystal silicon [ 5 ]. This is not to say that silicon is with-out any signifi cant disadvantages: as a solid, it is subject to laws of thermodynamics so that there will always be lattice vibrations characterized by a phonon spectrum at fi nite temperature. The presence of these phonons is a potential decoherence mecha-nism in any solid-sate quantum system.

9.1.2 Donor Atoms in Silicon: Nature’s Quantum Dots

Crystalline silicon can be conveniently doped by substituting a small number of sili-con atoms at lattice sites with elements from the adjacent pnictogen group-V col-umn of the periodic table. At low temperatures the extra electron left over after bonding remains loosely bound to the positive core. This object looks and behaves like an isolated hydrogen atom. There is an analogous Rydberg series of narrow lines in the absorption spectrum but they are shifted towards much lower energies. Whereas the Rydberg for an isolated hydrogen atom is 13.6 eV, the corresponding Rydberg for a group-V donor in silicon is 10s of meV. Thus, the spectral signatures

S.A. Lynch

65

lie at THz frequencies. Figure 9.1 shows a typical absorption spectrum for phosphorus donors recorded at low temperatures.

Silicon doped with phosphorus is a particularly interesting material from the point of view of quantum control, and this has lead to a dramatic resurgence of activity in the research fi eld. Much of this renewed interest stemmed from a pro-posal by Kane that silicon doped with group-V donors might be exploited to realize a quantum computer [ 6 ]. A related scheme involving group-V donors in silicon was also proposed by Stoneham [ 7 ]. A number of experiments designed to investigate the feasibility of Stoneham’s quantum computing scheme have now been performed using THz laser light, and are described in this chapter. THz pump-probe measure-ments reveal the lifetimes of the excited states, while THz photon echo experiments show how these states can be manipulated coherently at a quantum level. This was also the fi rst demonstration of a THz photon echo. All of the experiments were per-formed using the FELIX free electron laser at the FOM institute at Nieuwegein in the Netherlands. The silicon phosphorus lifetime results are discussed in greater detail in [ 8 ] and [ 9 ], while a full description of the THz pump-probe technique be found in [ 10 ]. A more complete discussion describing the discovery of a THz pho-ton echo and how it was exploited to demonstrate quantum control can be found in [ 11 ]. The main cogent points, however, will now be discussed.

Fig. 9.1 FT-IR absorbance spectrum for a sample of high-resistivity n-type phosphorus doped silicon recorded at 4.2 K. Two Rydberg series of lines, np 0 , and np ± , can be clearly observed at meV (THz) energies. The energy level structure has been superimposed over the spectrum in the left hand corner for reference. The Rydberg levels are labeled to the right of each level, while the relative position of the levels below the conduction band edge are directly to the left of each level. The energies (in meV) of some of the optically allowed transitions observed in the absorbance spectrum are also shown

9 Coherent Quantum Control of Donor States in Silicon with THz and MIR Light…

66

9.2 Silicon Donor Excited State Dynamics

Quantum information differs from conventional digital data in that the unit of information is a quantum bit, or qubit. A qubit can be generalized as a coupled two state system, normally existing in a superposition of states until a measure-ment is performed. When dealing with such a quantum object, it is essential to understand the timescale governing the underlying dynamics. This sets an upper bound on one of the critical timescales during which quantum operations can be performed before the system naturally relaxes back to the ground state. Most relaxation processes in solids take place on relatively fast timescales: certainly faster than any conventional THz detector can measure in real time. Thus, we have to resort to an indirect method to establish the temporal dynamics, i.e. time-resolved pump-probe spectroscopy.

9.2.1 THz Pump-Probe Lifetime Experiments Using FELIX

The pump-probe technique allows temporal phenomenon on a fast time-scale to be studied using a relatively slow THz detector. The basic principle involves exciting electrons into an upper state in the material being probed using a very intense pump pulse, and then monitoring the transmitted intensity of a much weaker probe pulse as the population in the ground state recovers. The free electron laser at the FOM Institute in the Netherlands (FELIX) was used as a source of short wavelength tun-able optical (THz) pulses. Figure 9.2 shows a schematic diagram of the pump-probe experimental setup. The beam path of the stronger pump pulse is shown as a solid line, while the beam path of the weaker probe pulse is shown as a dashed line. The pump traverses an optical delay line, allowing the relative delay between the pump and probe to be controlled. The polarization of the probe is rotated by 90 0 in order to discriminate it from the pump at the detector. A portion of the probe beam also passes through a much longer optical delay (~20 ns). All beam paths were confi ned in a dry nitrogen atmosphere to avoid water vapor absorption. The recorded probe signal is derived from the difference between the directly measured probe and refer-ence signal, resulting in a background free signal.

Figure 9.3 shows some typical pump-probe transmission data for three different pump pulse energies from [ 8 ]. Fits to this data with a simple exponential decay gave a value for the lifetime of T 1 = 205 ± 18 ps. This corresponds to a linewidth of 1/T 1 = 0.026 cm −1 , which is less, but not very much less, than the lowest value reported for this transition of 0.034 cm −1 , which was obtained in an isotopically enriched 28 Si sample [ 12 ]. This is signifi cant because it shows that the lifetime of the donor is not detrimentally affected by the quality of the surrounding crystal host. The donor atom is in effect isolated from the crystal.

S.A. Lynch

67

ReferenceChopper /Beam block

PolarisationRotation

FE

LIX

HeN

e

Attenuator Box

Pump

ProbeAttenuator Box

SampleBeamdump

Analyser

Ge:Ga

Detector

Iris

Iris

Iris

Iris Iris

CC2

CC1

dB1

P

dB2

BS2 BS1

Fig. 9.2 Schematic diagram showing the layout of the pump-probe experimental setup used to measure the lifetimes of the excited donor Rydberg states. A solid line shows the path of the pump beam, while a dashed line shows the path of the probe beam

10-1

10-2

10-3

10-4

0

Delay between pump and probe pulses (ps)

Pro

be T

rans

mis

sion

(ar

b. u

nits

)

200 400

16.7 nJ5.3 nJ1.1 nJ

600

pump

probe

800

Fig. 9.3 The change in probe transmission induced by the pump as a function of the time delay between pump and probe, observed in the Si:P sample for the 1s(A 1 ) → 2p 0 transition at a sample temperature, T, of 10 K, and a pump and probe photon energy of 34.1 meV. The rise of the leading edge indicates the pulse duration, which was 10 ps. The laser pump powers used correspond to the micro-pulse energies shown on the fi gure. The lowest pump pulse energy (1.1 nJ) corresponds to a focused photon fl uence of 10 17 photons/m 2 . Also shown are fi ts using a single exponential decay where the decay parameter is the spontaneous relaxation rate 1/T 1 . ( Inset ) Transient pump-probe experimental geometry


68

Very recently a further set of lifetime measurements were fi nally made on an isotopically enriched 28 Si sample [ 9 ]. The lifetime of the 2p 0 state in 28 Si:P was found to be 235 ps, which is 16 % longer than the lifetime of the reference Si:P sample with a natural isotope composition. It was concluded that the interaction of the 2p 0 state with inter-valley g-type longitudinal acoustic and f-type transverse acoustic phonons ultimately determines its lifetime. This interaction, which depends on the homogeneity of the crystal, becomes weaker in 28 Si because of its more perfect crystal lattice compared to natural Si, and this leads to a longer lifetime. The difference between the linewidths of the 1s(A 1 ) → 2p 0 transition in 28 Si:P and natural Si:P is more than a factor of two. It follows therefore that linewidth broaden-ing due to isotopic composition is an inhomogeneous process.

9.3 Coherent Control of the Donor Rydberg States

For the donor center to be useful, however, it is necessary to demonstrate that any desired quantum superposition of the ground and excited state can be generated. Put in simpler terms, it must be possible to both de-excite the donor electron from the upper state as well as excite it, in order to initialize the qubit. This ability is usually termed ‘coherent control’, and it is a necessary condition for the operation of any qubit. One elegant way of demonstrating coherent control is to exploit a phenome-non known as the photon echo as tool to explore the state of the donor electron. Before doing this, however, it is necessary to demonstrate that it is possible generate a THz photon echo.

9.3.1 Demonstrating the First THz Photon Echo

In order to provide conclusive proof of a true THz photon echo, both the directional property and expected timing of the phenomenon need to be established. This fi rst component of the experiment for demonstrating an echo is based on the pump-probe setup just described. This ‘front-end’ allows us to control the arrival timing of the THz pulses incident on the sample. Two different dedicated ‘back-ends’ allowed us to study the direction and temporal properties of the echo. These modifi cations to the setup are as shown in Fig. 9.4a, b respectively.

The directional property of the echo is defi ned by the wavevector ( k E = 2 k 2 − k 1 ) and was established by studying the angular distribution of the beams. In this experi-ment the pump ( k 1 ) and rephasing ( k 2 ) beams intersect at an angle of −5 0 . Simple vector algebra dictates that the echo should emerge at an angle +5 0 with respect to the direction of the rephasing beam k 2 . Figure 9.5a shows the intensities of the three beams exiting the sample as a function of angle. This graph proves that the echo emerges at the predicted angle with respect to the direction of the rephasing beam k 2 .

S.A. Lynch

69

The second important point for consideration is the echo arrival time. This has to be determined indirectly because there are currently no suffi ciently fast detectors in the THz spectral range. A reference pulse was split from the rephasing pulse using a beamsplitter and an extra optical delay line. The transmitted pump, rephasing, and emitted echo pulses, as well as the reference pulse are all focused onto the detector through a pinhole to produce a characteristic interference pattern in time. The angu-lar dispersion of the pump, rephasing and echo pulses is exploited to block all but one of them, thereby obtaining the interference patterns of the reference beam with the pump, rephasing and echo beams separately. By subtracting the mean intensity

BS3BS2BS1

BS1

R1

a

b

A

ANGLE

sampleP1 P2

R1

P1 P2sample

MM

D

A

FELIX

FELIX

D

TIMINGR2

kE

k2

k1

kE

k2

k1

Fig. 9.4 ( a ) Experimental setup used to determine the angular direction of the photon echo beam and ( b ) the setup used to determine the timing of the photon echo


70

1.0a

b

174.6 nJ

141.0 nJ

122.5 nJ

102.6 nJ

82.6 nJ

63.6 nJ

46.5 nJ

32.0 nJ

12.0 nJ

0.8

0.6

0.4

Sig

nal I

nten

sity

(a.

u.)

Det

ecto

r S

igna

l (a.

u.)

Pul

se In

tens

ity (

a.u.

)

Time (ps)

Angle (degrees)

0.2

k2

k1 kecho

0.0

-8 -4 0 4 8

Pump K1

Rephasing K2

Echo KF

Pump K1

Rephasing K2

Echo KF

6030-30 06030-30 0

0.5

0.0

1.0

0.5

0.0

1.0

0.5

0.0

1.0

T12

T2E

0.0

-1.0

+1.0

0.0

-1.0

+1.0

0.0

-1.0

+1.0

Fig. 9.5 ( a ) Angle resolved echo. The intensities of the angle resolved signals were recorded by translating the detector across the far-fi eld which shows that ( k E = 2 k 2 − k 1 ) as predicted. ( b ) Time resolved echo. On the left is the detector signal showing the interference patterns with the pump, rephasing and echo beams. A moving average has been subtracted, in order to remove the back-ground and laser drift. The pump, rephasing and echo temporal profi les were obtained from the square of these interference patterns, as shown on the right , where the pump rephasing beam time interval τ 12 and the rephasing beam-echo time interval τ 2E are also shown

and squaring the result, the arrival times and shapes of the pump, rephasing and echo pulses can then be determined as a function of time as shown in Fig. 9.5b . All three pulses take the form of well-defi ned peaks, with the maxima occurring at the times anticipated for echoes.

S.A. Lynch

71

9.3.2 Using the Echo to Demonstrate Coherent Quantum Control

The photon echo can be exploited as an experimental tool to investigate the quantum coherence properties of the excited donor states. The photon echo is used to directly observe Rabi oscillations produced by coherent optical excitation of phosphorus donors in silicon with intense THz pulses from the free-electron laser. Figure 9.6 shows the time-integrated photon echo signal S as a function of pump peak pulse area A P for a rephasing peak pulse area of 0.54π and a pulse length of 6.79 ps. The dotted line is the ideal theoretical result showing Rabi oscillations. The thin black line shows the corrected prediction when including the non-uniform spa-tial profi le of the laser beam, and the thick grey line includes the effect of both photoionization and the beam profi le. The theory lines were calculated using values for μ 12 , Γ 0 , σ 2p0 , and σ e , which were found from a global fi t of many experimental data sets like the one shown here. A full explanation of how this was done can be found in the supplemental information of Greenland’s paper [ 11 ]. The experimental results for the same conditions are shown as points. The normalization factor for the ordinate of the experiment relative to the theory was found by a global comparison of many similar experiments with different pulse lengths and rephasing pulse areas. The error bars shown indicate the standard deviation of the normalization factor (systematic for an individual experiment such the one in this fi gure) and dominate the statistical errors of the measurements.

Fig. 9.6 The time-integrated photon echo signal S as a function of pump peak pulse area AP for a

rephasing peak pulse area of 0.54π and a pulse length of 6.79 ps. The dotted line is the ideal theoretical result showing Rabi oscillations. The thin black line shows the corrected prediction when including the non-uniform spatial profi le of the laser beam, and the thick grey line includes the effect of both photoionization and the beam profi le. The experimental results for the same conditions are shown as points. The error bars shown indicate the standard deviation of the normalization factor (systematic for an individual experiment such the one in this fi gure) and dominate the statistical errors of the measurements


72

This shows that by varying the intensity of the initial optical pulse, that the probability amplitude of fi nding the donor electron in either the ground or excited state cycles through 0 and 1 as the pump intensity is increased. In other words, we have shown that it is possible prepare any desired coherent superposition of the two donor electron states by choosing the appropriate THz pump intensity.

9.4 The Future of Optically Controlled Donor Atoms in Silicon

The major downside of the group-V donors, however, is their small binding energy. This fact leads to the requirement for tunable THz laser light and also the need for very low temperatures. Are there other candidate donor atoms in silicon that might be better for this particular application?

9.4.1 Chalcogen Donors as an Alternative Single Photon Center

Nature provides us with another possibility. Moving one further column to the right on the periodic table leads us to the chalcogen elements, sulfur, selenium, and tel-lurium. Early spectroscopic work showed that these elements have much larger binding energies and that they leave sharp optical signatures in the mid-infrared band. Furthermore, the silicon can be co-doped with an acceptor atom such as boron, leaving a natural quantum dot with an even higher binding energy. This leads to the tantalizing prospect of a deep single photon center that could be coherently controlled with near infrared light, permitting experiments to be performed with a conventional bench-top OPA laser system, rather than necessitating a large shared user facility such as FELIX.

9.4.2 Technical Challenges of Introducing Chalcogens into Silicon

Doping silicon with chalcogen donors, such as sulfur, is not without its technologi-cal challenges. The very small solid solubility of the chalcogen elements in silicon precludes the conventional route of doping the silicon melt before pulling the single crystal. This has severely hampered the production of suitable samples, and stunted research in this area. It has recently been shown, however, that it is possible to intro-duce the chalcogen donors into silicon by high-pressure diffusion at elevated tem-peratures, while controlling the ratio of atomic and molecular chalcogen centers [ 13 ].

S.A. Lynch

73

High quality silicon samples with suffi ciently large chalcogen donor densities to perform infrared spectroscopy have now been demonstrated, as shown in Fig. 9.7 . It is likely that with further work the chalcogen donor density in the silicon can be increased to the level needed to perform quantum optics experiments.

9.5 Conclusion

The ability to coherently control the quantum state of a donor electron is potentially a game changer for quantum computing technology. While scientists are still some time away from delivering a fully operational qubit exploiting single donor centers in silicon such as phosphorus, the path ahead is now clearer. Some technological hurdles still need to be overcome. It is not clear whether THz or mid-infrared light is the best solution for optically switching the donor state. While the technology for fabricating the sort of nanoscale devices to harness a single phosphorus atom is maturing, the lack of THz laser sources with the appropriate spectral characteristics is a still a problem. Chalcogen donors such as sulphur and selenium might be easier to control optically but there are still signifi cant material challenges to be overcome before we can think of building nanoscale devices based on these elements.

4.14.24.34.44.54.64.7

260 270 280 290 300 310

0.30

0.26

0.22

0.18

Abs

orba

nce

Wavelength (µm)

Energy (meV)

Si:Se (atomic)

190 195 200 205 210

0.26

0.24

0.22

0.20

0.18

Abs

orba

nce

Wavelength (µm)

Energy (meV)

Si:Se2 (molecular)

6.4 6.3 6.2 6.1 6.0

170 175 180 1901850.06

0.08

0.10

0.12

0.14

Energy (meV) Energy (meV)

Wavelength (µm)

4.14.24.34.4 4.0 3.9

Si:S (atomic)

Wavelength (µm)

Si:S2 (molecular)0.20

0.18

0.16

0.14

0.12280 290 300 310 320

7.2 7.1 7.0 6.9 6.8 6.7 6.6

Fig. 9.7 Mid-infrared absorbance spectra of silicon doped with sulfur and selenium. The samples were made by Yuri Astrov’s group at the Ioffe Institute in St. Petersburg using the technique described in Ref. [ 13 ]. The spectra show a range of sharp absorption features crossing the mid- infrared band associated single atomic and diatomic chalcogen centers


74

Acknowledgments I acknowledge the fi nancial support of NWO and EPSRC (Advanced Research Fellowship EP/E061265/1 and COMPASSS, Grant Ref EP/H026622/1). I am also grate-ful to the Royal Society for their support through research grant RG110228.

References

1. Shor PW (1997) Polynomial-time algorithms for prime factorization and discrete logarithms on a quantum computer. Siam J Comput 26:1484–1509

2. Schmidt-Kaler F, Haffner H, Riebe M, Gulde S, Lancaster GPT, Deuschle T et al (2003) Realization of the Cirac-Zoller controlled-NOT quantum gate. Nature 422:408–411

3. Leibfried D, DeMarco B, Meyer V, Lucas D, Barrett M, Britton J et al (2003) Experimental demonstration of a robust, high-fi delity geometric two ion-qubit phase gate. Nature 422:412–415

4. DiVincenzo DP (2000) The physical implementation of quantum computation. Fortschritte Der Physik-Prog Phys 48:771–783

5. Andreas B, Azuma Y, Bartl G, Becker P, Bettin H, Borys M et al (2011) Determination of the Avogadro constant by counting the atoms in a Si-28 crystal. Phys Rev Lett 106(3):030801 [4 pages]

6. Kane BE (1998) A silicon-based nuclear spin quantum computer. Nature 393:133–137 7. Stoneham AM, Fisher AJ, Greenland PT (2003) Optically driven silicon-based quantum gates

with potential for high-temperature operation. J Phys Condens Matter 15:L447–L451 8. Vinh NQ, Greenland PT, Litvinenko K, Redlich B, van der Meer AFG, Lynch SA et al (2008)

Silicon as a model ion trap: time domain measurements of donor Rydberg states. Proc Natl Acad Sci U S A 105:10649–10653

9. Hubers H-W, Pavlov SG, Lynch SA, Greenland T, Litvinenko KL, Murdin B et al (2013) Isotope effect on the lifetime of the 2p0 state in phosphorus-doped silicon. Phys Rev B 88:035201

10. Lynch SA, Matmon G, Pavlov SG, Litvinenko KL, Redlich B, van der Meer AFG et al (2010) Inhomogeneous broadening of phosphorus donor lines in the far-infrared spectra of single- crystalline SiGe. Phys Rev B 82:245206

11. Greenland PT, Lynch SA, van der Meer AFG, Murdin BN, Pidgeon CR, Redlich B et al (2010) Coherent control of Rydberg states in silicon. Nature 465:1057–1061

12. Karaiskaj D, Stotz JAH, Meyer T, Thewalt MLW, Cardona M (2003) Impurity absorption spectroscopy in Si-28: the importance of inhomogeneous isotope broadening. Phys Rev Lett 90(18):186402 [4 pages]

13. Astrov YA, Lynch SA, Shuman VB, Portsel LM, Makhova AA, Lodygin AN (2013) Silicon with an increased content of monoatomic sulfur centers: sample fabrication and optical spectroscopy. Semiconductors 47:247–251

S.A. Lynch


Abstract Label-free and online bio-chemical detection proved to be possible with THz waves, which would open alternative ways to a myriad of applications in the chemical, biological and medical fi elds such as integrated micro-reactor monitors, PCR monitoring, faster and cheaper drug discovery. However some technological challenges need to be surmounted, especially low sensitivity and large sample quan-tity, before practical applications can be implemented. These challenges are tackled here by an integrated sensor approach, whereby a specially developed substrate integrated waveguide is used in combination with capillary tube to maximally exploit the dynamic range of the measurement system. The sensor performance is benchmarked with water/alcohol mixtures. A selection of measurements on biologi-cal substances is presented to demonstrate the bio-chemical sensing capabilities. The developed approach shows outstanding sensitivity performance for extremely small sample quantities.

Keywords Terahertz • Millimeter waves • Sensing • Liquid • Biomolecule • Protein • DNA • Cells • Binding • Chemical • Substrate integrated waveguide • Capillary tube

Chapter 10 THz Bio-chemical Sensing Capabilities with High Performance SIW Based Sensor on nL-Volume Liquids in Capillary

V. Matvejev , J. Stiens , C. De Tandt , and D. Mangelings

V. Matvejev • C. De Tandt Laboratory of Micro- and Photoelectronics, LAMI-ETRO , Vrije Universiteit Brussel , Pleinlaan 2 , Brussel 1050 , Belgium

J. Stiens (*) Laboratory of Micro- and Photoelectronics, LAMI-ETRO , Vrije Universiteit Brussel , Pleinlaan 2 , Brussel 1050 , Belgium

Unit SSET, Department HIM, Group RFCDM , IMEC , Kapeldreef 75 , Leuven 3001 , Belgium e-mail: [email protected]

D. Mangelings Department of Analytical Chemistry and Pharmaceutical Technology , Vrije Universiteit Brussel , Laarbeeklaan 103 , Brussel 1090 , Belgium

76

10.1 Introduction

THz wave sensitivity to weak intermolecular interactions, which govern many biological processes, has high application potential in the pharmaceutical industry (or drug discovery). The promising advantage of THz technology in this fi eld is label- free, immobilization-free, online and automated detection.

However the detection with THz waves has several technological challenges, preventing from the commercialization. Biological samples are often highly diluted with water and introduce high losses. The most common THz measurement approach is free-space, which requires intolerable sample amount. Large amount of highly absorbing sample implies demanding dynamic range on the measurement system. Not to mention the price of reagents. Therefore the trend is obvious: to scale down the sample and increase the THz wave interaction with sample.

The integrated sensors reported to date utilize planar transmission lines (micro- strip line, single wire transmission – Goubau line). They offer signifi cant sample volume reduction, however sensitivity was improved only moderately. Here we make use of the hollow metallic pipe waveguide, which has superior performance in terms of losses. The capillary tube inserted into the waveguide increases the THz wave interaction with the liquid enclosed in it therefore increasing the sensitivity.

10.2 Results and Discussion

The sensor comprises a capillary tube and a substrate integrated waveguide see Fig. 10.1 . The low-loss substrate integrated waveguide (SIW) is made with micro-machining techniques in silicon bulk, which is then metalized and enclosed to form

Fig. 10.1 Sensor confi guration cut-away view: substrate integrated waveguide, commercial capillary tube

V. Matvejev et al.

77

a hollow metallic pipe waveguide with hexagonal cross-section [ 1 ]. The SIW is then provided with additional holes to insert the capillary tube (CT) inside. The CT axis is located in the center and perpendicular to waveguide H-plane, thus allowing for maximal EM-wave and liquid interaction [ 2 ]. The CT dimensions are selected (rela-tive to the waveguide) to increase the sensor response in either transmission (S 21 ) or refl ection (S 11 ) modes. In refl ection mode the sensor exhibits resonance behavior and therefore is more sensitive.

The sensors shows outstanding performance, see Fig. 10.2 . The sensor, designed to operate in WR-3.4 band (220–330 GHz), shows the following characteristics. In transmission mode at 280 GHz, the measured S 21 signal change between water and methanol fi lled CT is 15 dB. The simulated S 21 signal change is of the same order of magnitude. The smallest detectable methanol molar fraction in water, in transmis-sion mode, is 10 −4 and 10 −2 for the signal change of 0.1 and 1.0 dB, respectively. The refl ection mode response has still to be measured, however simulations suggest sig-nifi cant performance increase. The refl ection signal sensitivity to water-methanol is 53 dB. The smallest detectable methanol molar fraction is 3.3 and 4.3 × 10 −5 with a refl ection signal 0.1 and 1 dB resolution, respectively. The actual liquid volume considered required (waveguide and CT intersection) for our sensor operation is 40 and 4 nL for transmission and refl ection mode operation, respectively. The sensor performance is summarized in Table 10.1 .

In order to compare the sensor performance with other THz detection tech-niques it is preferable to express mass concentration. This way the molecule size is

Fig. 10.2 Simulated and measured response curves for sensors transmission and refl ection opera-tion modes. The test liquid is water-methanol mixture

10 THz Bio-chemical Sensing Capabilities with High Performance…

78

taken into account. The measured smallest detectable methanol molar fraction x(MeOH) = 10 −4 translates to ρ(MeOH) = 178 μg/mL mass concentration. The expected smallest detectable mass concentration of methanol in refl ection mode is ρ(MeOH) = 58.6 μg/mL. The spectrometer based on 10 W peak power p-Ge laser source operating in 1–4 THz [ 3 ] was used in combination with liquid cells in free- space confi guration. The minimum sensitivities obtained with such high power source: 1.28 mg/mL for Ubiquitin [ 4 ], 6.70 mg/mL for HSA [ 5 ], 38.7 mg/mL for NAGA [ 6 ] and 240 mg/mL for lactose [ 7 ]. Free space spectroscopy tool in combi-nation with liquid cells achieves 57.5 mg/mL sensitivity for HEWL [ 8 ]. Spectroscopic technique based on planar single wire transmission line achieves 26.7 mg/mL with ethanol [ 9 ], in addition with the advantage of reduced sample volume. Hence our actual THz SIW sensor design outperforms competitive struc-tures with one to three orders of magnitude.

Protein-ligand interaction detection in physiological buffer (mM TRIS buffer) is demonstrated with streptavidin – protein and biotin – ligand specifi c to the protein. First, two solutions of streptavidin (Mr = 60 kDa, supplier: Sigma-Aldrich) and bio-tin (Mr = 244 Da, supplier: Sigma-Aldrich) are prepared in buffer with 2 mM (2 × 10 −3 mol/L) concentration. Then, 2 mM streptavidin and 2 mM biotin solutions of the same volume are mixed together to result in solution of 1 mM streptavidin with 1 mM biotin. At last, two solutions of 2 mM streptavidin and 2 mM biotin are diluted 1:1 with buffer to result in 1 mM solutions of streptavidin and biotin. The resulting solutions are measured in the following sequence (as shown in Fig. 10.3 ): buffer, 1 mM streptavidin, 1 mM biotin, 1 mM streptavidin with 1 mM biotin. The sequence is important to minimize the risk of contamination of one sample with another one measured previously.

The response of the refl ection mode sensor for streptavidin and biotin solutions is presented in Fig. 10.3 . The measured refl ection minimum of the sensor responses of buffer, streptavidin, biotin and streptavidin with biotin are −67.04 ± 0.65 dB, −54.49 ± 0.13 dB, −68.00 ± 0.72 dB and −55.83 ± 0.12 dB, respectively. The biotin alone does not make a big contrast, the response difference between biotin and buffer is 0.96 dB. The response curves of biotin and buffer are not distinguishable. While when biotin is added to streptavidin the streptavidin response changes by 1.24 dB from initial value. The difference between pure streptavidin and streptavidin and biotin overall response curves is evident (with the maximum difference of 3.44 dB).

Refl ection minimum for streptavidin with biotin solution is lower than for pure streptavidin, which could correspond to a lower concentration of streptavidin.

Table 10.1 The smallest detectable methanol concentration in water and water-methanol sensitivity (full range)

Mode Volume (nL)

ΔS = 0.1 dB ΔS = 1.0 dB

Full range (dB) x MeOH x MeOH

S21 40 10 −4 (Meas) 10 −2 (Meas) 15 S11 4 3.3 × 10 −5 (Sim) 4.3 × 10 −5 (Sim) 53

V. Matvejev et al.

79

This difference is explained by the change in the hydration shell of the protein. Streptavidin molecule has a number of “pockets” on its surface, which are comple-mentary to biotin molecule. When biotin specifi cally binds to streptavidin the hydration water is released from the pocket and the total amount of hydrated water is reduced. During protein-ligand binding some part water changes its state, going from hydrated (retarded) to bulk like. Similar conclusions are drawn from experi-mental and theoretical studies of enzyme-substrate interaction [ 10 ]. Protein- ligand interaction accompanied by hydration water changes are refl ected in the dielectric permittivity at THz frequencies.

The developed high performance THz sensor will be explored for its applicability for R&D in the pharmaceutical fi eld. It will be used to measure: concentrations, biomolecules’ hydration shells, bio-molecular interactions and conformation evolu-tions. Some technology demonstrative cases will be presented.

10.3 Conclusion

In summary, an integrated THz sensor with outstanding sensitivity is developed. The current prototype operating in transmission mode at 220–330 GHz frequency range is capable of detecting 178 μg/mL concentration, which is an improvement of ×10 in sensitivity compared to the state-of-the-art solutions. The next generation sensor operating in the refl ection mode is expected to detect 58.6 μg/mL. Moreover

Fig. 10.3 Sensor refl ection response to wild-type and denatured HEWL concentration

10 THz Bio-chemical Sensing Capabilities with High Performance…

80

the sample volume required for carrying out the measurements is reduced to 40 and 4 nL with transmission and refl ection operating sensors in the 220–330 GHz fre-quency band. The demonstration of the sensor capabilities to detect biologically and chemically relevant samples will follow.

References

1. Matvejev V, De Tandt C, Ranson W, Stiens J (2010 ) Wet silicon bulk micromachined THz waveguides for low-loss integrated sensor applications. In: The 35th international conference on infrared, millimeter and THz waves, Rome

2. Matvejev V, de Tandt C, Ranson W, Stiens J, Vounckx R, Mangelings D (2011) Integrated waveguide structure for highly sensitive THz spectroscopy of nano-liter liquids in capillary tubes. Prog Electromagn Res 21:89–101

3. Leitner D, Gruebele M, Havenith M (2008) Solvation dynamics of biomolecules: modeling and terahertz experiments. HFSP J 2:314–323

4. Heyden M, Havenith M (2010) Combining THz spectroscopy and MD simulations to study protein-hydration coupling. Methods 52:74–83

5. Luong TQ, Verma PK, Mitra RK, Havenith M (2011) Do hydration dynamics follow the struc-tural perturbation during thermal denaturation of a protein: a terahertz absorption study. Biophys J 101:925–933

6. Born B, Weingaertner H, Bruendermann E, Havenith MB (2009) Solvation dynamics of model peptides probed by terahertz spectroscopy. Observation of the onset of collective network motions. J Am Chem Soc 131:3752–3755

7. Heyden M, Bruendermann E, Heugen U, Niehues G, Leitner DM, Havenith M (2008) Long- range infl uence of carbohydrates on the solvation dynamics of water-answers from tera-hertz absorption measurements and molecular modeling simulations. J Am Chem Soc 130:5773–5779

8. Vinh NQ, Allen SJ, Plaxco KW (2011) Dielectric spectroscopy of proteins as a quantitative experimental test of computational models of their low-frequency harmonic motions. J Am Chem Soc 133:8942–8947

9. Laurette S, Treizebre A, Affouard F, Bocquet B (2010) Subterahertz characterization of etha-nol hydration layers by microfl uidic system. Appl Phys Lett 97:111904

10. Grossman M, Born B, Heyden M, Tworowski D, Fields GB, Sagi I, Havenith M (2011) Correlated structural kinetics and retarded solvent dynamics at the metalloprotease active site. Nat Struct Mol Biol 18:1102–1108

V. Matvejev et al.


Abstract Pentaerythritol tetranitrate (PETN) reductase is an enzyme produced by bacteria ( Enterobacter cloacae PB2) that have the ability to use nitroaromatic (trinitrotoluene – TNT, picric acid) or nitroester (glycerol trinitrate – GTN, PETN) explosives as a sole nitrogen source for growth. The reaction of PETN reductase comprises a reductive and an oxidative half-reaction, explosives being degraded during the oxidative half-reaction. Since the side chains of the residues that form the active site of PETN reductase assume different conformations during the two half- reactions, we investigated whether THz spectroscopy could discriminate between the different states of PETN reductase and whether their discrimination could be useful in the detection of explosives and consequent monitoring of explosives deg-radation process. In order to answer these questions, we simulated and compared the THz spectra of the hydrated enzyme in the reduced state, in the oxidized state bound to an inert substrate, namely a thiocyanate ion and in the oxidized state bound to picric acid. Our results show that PETN reductase structures in reduced and oxi-dized states present a signifi cantly different THz absorption, the enzyme in reduced state being the one that absorbs more THz radiation. In oxidized state, the enzyme bound to thiocyanate absorbs less THz radiation then the enzyme bound to picric acid. By discriminating between the conformations of PETN reductase during the two half-reactions involved in explosives degradation, THz spectroscopy should allow the evaluation of explosives degradation state.

Keywords THz spectroscopy • Spectra simulation • PETN reductase • Picric acid • Degradation of explosives

Chapter 11 A Theoretical Study on Monitoring Explosives Degradation by Pentaerythritol Tetranitrate Reductase Using THz Spectroscopy

Maria Mernea and Dan Florin Mihailescu

M. Mernea (*) • D. F. Mihailescu Department of Anatomy, Animal Physiology and Biophysics, Faculty of Biology , University of Bucharest , Bucharest , Romania e-mail: [email protected]

82

11.1 Introduction

Pentaerythritol tetranitrate (PETN) reductase is a NADPH-dependent fl avoenzyme produced by bacteria ( Enterobacter cloacae PB2) that have the ability to use nitroaromatic (trinitrotoluene – TNT, picric acid) or nitroester (glycerol trinitrate – GTN, PETN) explosives as a sole nitrogen source for growth [ 1 , 2 ]. The reaction of PETN reductase comprises two half-reactions. In the reductive half-reaction, NADPH reduces the fl avin mononucleotide (FMN) cofactor of PETN reductase, while in the oxidative half-reaction, FMN is oxidised by a nitro-containing explosive. Explosives degradation occurs during the oxidative half-reaction due to the reductive hydride addition to the aromatic nucleus [ 1 , 3 ].

The side chains of residues that form the active site of PETN reductase assume different conformations during the two half-reactions [ 4 ]. Since Terahertz (THz) spectroscopy is a technique highly sensitive to proteins structure and conforma-tion, we investigated whether THz spectroscopy could discriminate between the different states of PETN reductase and whether their discrimination could be use-ful in the detection of explosives and consequent monitoring of the explosives degradation process. In order to answer these questions, we simulated and com-pared the THz spectra of the enzyme in the reduced state [ 5 ], in the oxidized state bound to an inert substrate, namely a thiocyanate ion [ 5 ] and in the oxidized state bound to picric acid [ 4 ].

11.2 Methods

The coordinates of the enzyme in the three states were retrieved from the Protein Data Bank, as follows: (1) the structure 1H63 [ 5 ], presenting the reduced form of the enzyme, bound to FMN; (2) the structure 1H51 [ 5 ], representing the oxidized enzyme bound to FMN and thiocyanate ion and (3) the structure 1VYR [ 4 ] presenting PETN reductase in the oxidized state, bound to FMN and picric acid. When picric acid is present in the enzyme active site, the side chain of Trp102 can assume at least two conformations. These conformations of Trp102 further infl uence the side chains of other residues found both close and distal to the active site. Since 1VYR structure comprises information on two conformations (conformation A and B) of PETN reductase, both of them were taken into account.

The THz spectra of PETN reductase were simulated based on the normal modes analysis (NMA) performed with CHARMM [ 6 ], using the building-block approach of Tama et all [ 7 ]. In order to apply the harmonic approximation, the structures were energy minimized for the minimum number of minimization steps required in order to generate only positive normal modes frequencies. The normal modes of vibration were calculated considering all water molecules present in the crystal structures 1H63 (586 molecules) and 1H51 (476 molecules). In the case of 1VYR structure, from the784 water molecules present, we considered only the 538 molecules from

M. Mernea and D.F. Mihailescu

83

the fi rst two hydration layers. FMN was modeled based on the topology and parameter fi les developed by P.L. Freddolino et al. [ 8 ], the thiocyanate ion and picric acid were modeled using the parameters we developed using the SwissParam web service [ 9 ].

The theoretical THz spectra of the structures were calculated using the method described elsewhere [ 10 , 11 ]. The method involves calculating the nor-mal modes intensities based on the dipole derivatives during normal modes and on the eigenvectors obtained by diagonalizing the systems mass-weighted Hessian. In order to obtain spectra that can be compared with the experimen-tally recorded spectra, the intensity lines are then uniformly broadened using a Lorentzian function.

11.3 Results

Prior to energy minimization, the backbones of PETN reductase structures were very similar, the root mean squared deviation (RMSD) for the backbones of the structures 1H51 and 1H63 compared to 1VYR (the 3D structure is presented in Fig. 11.1a ) being 0.18 and 0.26 Å. After the energy minimization of PETN reduc-tase in oxidized and reduced states bound to FMN, the appropriate substrate and surrounded by the water molecules from the crystal structures, their backbones changed to a similar extent of ~0.5 Å.

The THz spectra simulated for the structures in reduced state, in oxidized state bound to the thiocyanate ion and in oxidized state bound to picric acid (both confor-mations) are presented in Fig. 11.1b . The theoretical THz absorption of all structures increases linearly in the frequency range comprised between 0.25 and 0.6 THz, therefore in this frequency range we compared the spectra by comparing the slopes of the lines that fi t them.

Calculated slope values are 18.9 ± 0.3 (R 2 = 0.99) for PETN reductase in the reduced state, 10.9 ± 0.3 (R 2 = 0.98) for oxidized PETN reductase bound to thiocya-nate ion, 16.5 ± 0.2 (R 2 = 0.99) for oxidized PETN reductase bound to picric acid in conformation A and 15.5 ± 0.4 (R 2 = 0.98) for oxidized PETN reductase bound to picric acid in conformation B. These values show that in the 0.25–0.6 THz fre-quency range, PETN reductase structures in reduced and oxidized states present a signifi cantly different THz absorption, the enzyme in reduced state being the one that absorbs more THz radiation. In oxidized state, the enzyme bound to thiocyanate absorbs less THz radiation then the enzyme bound to picric acid. The calculated slope values show that even in the case of the enzyme bound to picric acid, we should still see differences between the two conformational states identifi ed by X-ray crystallography [ 4 ].

In the frequency range comprised between 0.6 and 1 THz, the reduced enzyme and the oxidized enzyme bound to thiocyanate linearly absorb THz radiation up to 0.8 and 0.7 THz. At 0.8 THz, the spectrum of PETN reductase in reduced state

11 A Theoretical Study on Monitoring Explosives Degradation by Pentaerythritol…

84

presents a discrete shoulder, while the oxidized enzyme bound to thiocyanate presents two weak shoulders at ~0.7 and ~0.9 THz. The spectra of the two confor-mations of oxidized PETN reductase bound to picric acid show a plateau between 0.6 and 0.8 THz for conformation A and between 0.6 and 0.86 THz for conforma-tion B. At higher frequencies, their THz absorption increases and in the case of conformation A, a discrete shoulder can be seen at 0.95 THz.

In order to explain the absorption features that can be seen on the simulated spectra, in Fig. 11.1c, d we represented the normal modes with high dipole deriva-tives calculated for PETN reductase in all conformations. As can be seen, PETN reductase in reduced state presents a higher density of normal modes with high dipole derivatives which leads to the highest THz absorption, while the structure that present the lowest THz absorption, namely oxidized PETN reductase bound to thiocyanate, presents the lowest density of normal modes with high dipole derivatives.

Fig. 11.1 ( a ) Structure of PETN reductase from 1VYR [ 4 ] crystal structure. Protein backbone is represented in white. Its cofactor – the fl avin mononucleotide (FMN) molecule and its substrate – the picric acid molecule are represented in black bonds and are labeled on the fi gure. ( b ) The THz spectra of PETN reductase in reduced form (reduced + FMN), in oxidized state bound to thiocya-nate ion (oxidized + FMN + SCN) and in the oxidized state bound to picric acid (oxidized + FMN + picric acid), in both conformations identifi ed form 1VYR structure (conf A and Conf B). ( c ) and ( d ) The structures normal modes that present high dipole derivatives


85

11.4 Conclusion

Here we present simulated THz spectra of hydrated PETN reductase in reduced state, bound to its cofactor (FMN) and in oxidized states, bound its cofactor and to picric acid or to thiocyanate ion. The differences between the simulated spectra show that THz spectroscopy should be able to discriminate between the oxidized and the reduced forms of PETN reductase. THz spectroscopy should be able to identify the enzyme bound to picric acid and even more, it should be able to dis-criminate between the two conformations assumed by the enzyme bound to picric acid. By discriminating between the conformations of PETN reductase during the two half-reactions involved in explosives degradation, THz spectroscopy could allow the evaluation of explosives degradation state.

Acknowledgments The authors would like to acknowledge the fi nancial support of the Romanian Ministry of Education, Research, Youth and Sport through the “IDEAS” project 137/2011 (Protein three-dimensional structure and conformational transitions determination by high-power narrowband THz radiation and by molecular modeling).

References

1. Binks PR, French CE, Nicklin S, Bruce NC (1996) Degradation of pentaerythritol tetranitrate by Enterobacter cloacae PB2. Appl Environ Microbiol 62(4):1214–1219

2. French CE, Nicklin S, Bruce NC (1998) Aerobic degradation of 2,4,6-trinitrotoluene by Enterobacter cloacae PB2 and by pentaerythritol tetranitrate reductase. Appl Environ Microbiol 64(8):2864–2868

3. Khan H, Harris RJ, Barna T, Craig DH, Bruce NC, Munro AW, Moody PC, Scrutton NS (2002) Kinetic and structural basis of reactivity of pentaerythritol tetranitrate reductase with NADPH, 2-cyclohexenone, nitroesters, and nitroaromatic explosives. J Biol Chem 277(24):21906–21912, doi: 10.1074/jbc.M200637200 M200637200 [pii]

4. Khan H, Barna T, Harris RJ, Bruce NC, Barsukov I, Munro AW, Moody PCE, Scrutton NS (2004) Atomic resolution structures and solution behavior of enzyme-substrate complexes of Enterobacter cloacae PB2 pentaerythritol tetranitrate reductase. J Biol Chem 279(29):30563–30572. doi: 10.1074/jbc.M403541200

5. Barna TM, Khan H, Bruce NC, Barsukov I, Scrutton NS, Moody PCE (2001) Crystal structure of pentaerythritol tetranitrate reductase: “fl ipped” binding geometries for steroid substrates in different redox states of the enzyme. J Mol Biol 310(2):433–447. doi: 10.1006/jmbi.2001.4779

6. Brooks BR, Brooks CL 3rd, Mackerell AD Jr, Nilsson L, Petrella RJ, Roux B, Won Y, Archontis G, Bartels C, Boresch S, Cafl isch A, Caves L, Cui Q, Dinner AR, Feig M, Fischer S, Gao J, Hodoscek M, Im W, Kuczera K, Lazaridis T, Ma J, Ovchinnikov V, Paci E, Pastor RW, Post CB, Pu JZ, Schaefer M, Tidor B, Venable RM, Woodcock HL, Wu X, Yang W, York DM, Karplus M (2009) CHARMM: the biomolecular simulation program. J Comput Chem 30(10):1545–1614. doi: 10.1002/jcc.21287

7. Tama F, Gadea FX, Marques O, Sanejouand YH (2000) Building-block approach for determining low-frequency normal modes of macromolecules. Proteins 41(1):1–7

8. Freddolino PL, Dittrich M, Schulten K (2006) Dynamic switching mechanisms in LOV1 and LOV2 domains of plant phototropins. Biophys J 91(10):3630–3639, doi:S0006- 3495(06)72075-6 [pii] 10.1529/biophysj.106.088609

11 A Theoretical Study on Monitoring Explosives Degradation by Pentaerythritol…

http://dx.doi.org/10.1074/jbc.M200637200

http://dx.doi.org/10.1074/jbc.M403541200

http://dx.doi.org/10.1006/jmbi.2001.4779

http://dx.doi.org/10.1002/jcc.21287

http://dx.doi.org/10.1529/biophysj.106.088609

86

9. Zoete V, Cuendet MA, Grosdidier A, Michielin O (2011) SwissParam: a fast force fi eld generation tool for small organic molecules. J Comput Chem 32(11):2359–2368. doi: 10.1002/jcc.21816

10. Mernea M, Calborean O, Petrescu L, Dinca MP, Leca A et al (2010) The fl exibility of hydrated bovine serum albumin investigated by THz spectroscopy and molecular modeling. In: Proceedings of SPIE 7469, ROMOPTO 2009: ninth conference on optics: micro- to nanopho-tonics II, 74690N, Sibiu, 17 May. doi: 10.1117/12.861821

11. Mernea M, Leca A, Dascalu T, Mihailescu D (2011) Bovine serum albumin 3D structure determination by THz spectroscopy and molecular modeling. In: Pereira MFF, Shulika O (eds) Terahertz and Mid infrared radiation, NATO science for peace and security series B: physics and biophysics. Springer, Dordrecht, pp 101–105


http://dx.doi.org/10.1002/jcc.21816

http://dx.doi.org/10.1117/12.861821


Abstract Rapid increase in applications of the electromagnetic waves in the terahertz frequency range requires new techniques to obtain continues-wave terahertz radiation sources. It is shown that rectangular shaped mesa structure from Bi 2 Sr 2 CaCu 2 O 8+d (Bi2212) is a good source for generation of terahertz radiation. In this study, we have fabricated triple mesa structures using e-beam lithography and argon ion beam etching techniques. Our aim is to fi nd the critical current density dependence of mesa area to obtain maximum emission power for the terahertz radiation.

Keywords Intrinsic Josephson junction • Terahertz radiation • Bi 2 Sr 2 CaCu 2 O 8+d single crystal

12.1 Introduction

There is a growing interest in the research on high power, low cost and tunable terahertz radiation sources. Because, electromagnetic waves in this frequency region (0.1–10 THz) have many benefi cial applications such as imaging, sensing and spectroscopy. In contrast to that, there is still lack of powerful, continuous-wave, compact solid-state terahertz sources [ 1 ]. Therefore, the research has gone towards the novel THz sources, which include technology of high temperature

Chapter 12 Area Dependence of Josephson Critical Current Density in Superconducting Bi 2 Sr 2 CaCu 2 O 8+d for Terahertz Emission

H. Saglam , Y. Demirhan , K. Kadowaki , N. Miyakawa , and L. Ozyuzer

H. Saglam (*) • Y. Demirhan • L. Ozyuzer Department of Physics , Izmir Institute of Technology , Urla , Izmir 35430 , Turkey e-mail: [email protected]

K. Kadowaki University of Tsukuba , Tsukuba , Japan

N. Miyakawa Tokyo University of Science , Tokyo , Japan

88

superconductors (HTSs). The most signifi cant reason making HTSs good source for the generation of THz radiation is their layered structure, which makes possible the propagation of electromagnetic waves by Josephson plasma oscillations and the frequency of the Josephson plasma is in the THz ranges [ 2 ]. Furthermore, the stacks of intrinsic Josephson junctions (IJJs) in Bi 2 Sr 2 CaCu 2 O 8+d (Bi2212) can be used as voltage- frequency converter [ 3 ]. Ozyuzer et al. obtained continuous, coherent and monochromatic electromagnetic terahertz radiation from rectangular mesa-shaped samples of the high temperature superconductor Bi2212, in which electromagnetic cavity resonance synchronizes almost all of the IJJ [ 4 , 5 ]. Therefore, in our study, we are searching for proper critical current density (which is closely related to the heating effects) dependence of mesa area to obtain maximum emission power for the terahertz radiation.

12.2 Experimental

In the previous studies, it has been shown that the THz emitting mesas are below a certain under-doped level, which has relatively small critical current unlike optimally doped and over-doped Bi2212 samples [ 6 ]. Therefore, we fabricated a set of samples with various doping levels and the heat treatment duration is varied to change the T c and critical current of crystals. Annealing system for underdoping of Bi2212 can be seen in Fig. 12.1 . First, these annealed single crystals were glued on sapphire substrates using good thermal and electrical conductor silver epoxy. Then we obtained smooth and clean surface of Bi2212 by cleaving process. After that 100 nm of gold layer is deposited onto cleaved crystals to protect them from chemical reactions. To obtain natural IJJ stacks with various size and height, mesa on Bi2212 have been fabricated using e-beam lithography and argon ion beam etching techniques. Since obtained mesas have small area, we have deposited CaF 2 insulating layer onto them.

Fig. 12.1 Annealing system for underdoping of Bi2212

H. Saglam et al.

89

Then using lift-off technique by e-beam lithography, the gold stripes with the width of 30 μm were deposited on insulating layer and mesa. Finally, three gold wires were placed using silver epoxy. The fabricated triple mesa structure with three gold contacts can be seen in Fig. 12.2 .

After the mesa fabrication, surface profi lometer and atomic force microscope were used to obtain the exact dimensions of each mesa. In order to characterize the Bi2212 mesas, three-probe contact were taken to measure c-axis resistance versus temperature (R–T), and current–voltage behavior (I–V) were measured in a He fl ow cryostat.


In this study, SEM and surface profi lometer were used to determine dimensions of the fabricated mesa structures on the same crystal, respectively.

Figure 12.3 represents the resistance versus temperature (R-T) behaviors of three mesa structures with areas of 300 × 50, 200 × 50 and 100 × 50 μm 2 . From R-T measurements, we have observed that all fabricated mesas shows typical resistance versus temperature properties of c-axis of Bi2212 high temperature superconductors. It is seen from the graphs that the resistance of mesas exponentially increases when the temperature decrease from room temperature to 20 K (Fig. 12.3 ). While the onsets of the critical temperatures are approximately 92 K, R(T c )/R(300 K) values are nearly 1.20. The R(T c )/R(contact) values are 2.60, 4.30 and 9.25 for the mesas with areas of 300 × 50, 200 × 50 and 100 × 50 μm 2 , respectively (Fig. 12.3 ). Therefore, we can conclude that when the mesa area is increasing the measured resistance of sample is decreasing because of the surface area of the mesas. In addition, when we look at our R-T results, we see that the contact resistance below transition temperature is still fi nite.

Fig. 12.2 Optical images of fabricated triple mesa structure

12 Area Dependence of Josephson Critical Current Density in Superconducting…

90

This is because of the interface between gold layer and Bi2212 crystal and the resistance is increasing below critical temperature since the interface acts like a tunneling barrier at low temperatures.

In this work, we have studied change in Josephson critical current density of mesas with different dimensions. For this reason, the I–V measurements (as seen in Figs. 12.4 , 12.5 , and 12.6 ) were taken at 20 K to fi nd the Josephson critical current values to obtain the area dependence of Josephson critical current density in super-conducting Bi2212 mesas for terahertz emission. For our triple mea structure, the magnitudes of Josephson critical currents of Bi2212 single crystal are about 21.8, 18.4, 9.0 mA for mesa with areas of 300 × 50, 200 × 50, 100 × 50 μm 2 as seen

0

5

10

15

20

25

30

35

40

0 50 100 150 200 250 300 350

R (

W)

T (K)

(a)

(b)

(c)

Fig. 12.3 Comparison of resistance versus temperature behavior for three mesas with areas of 300 × 50 ( a ), 200 × 50 ( b ), 100 × 50 μm 2 ( c )

-40

-30

-20

-10

0

10

20

30

40

-1200 -800 -400 0 400 800 1200

I (m

A)

V (mV)

T = 20 K Fig. 12.4 I–V measurement of mesa with area of 300 × 50 μm 2

H. Saglam et al.

91

in Figs. 12.4 , 12.5 , and 12.6 , respectively. Then we have calculated the critical current densities of each mesa as 140, 160 and 180 A/cm 2 , respectively. Therefore, we can conclude that the critical current density is decreasing when area of mesa is increasing.

When we look at the I–V curves in Figs. 12.4 , 12.5 , and 12.6 , it is clearly seen that there exist some quasiparticle branches. In addition, we have seen that while the mesa dimension is increasing, the back bending of the current-voltage curve is seen due to the large volume of the mesa causes self-heating [ 7 , 8 ]. The result of heating effects can be seen in all I–V curves as backbending. Furthermore, it is obviously seen from the I–V curves that the back bending voltage points are increasing,

-40

-30

-20

-10

0

10

20

30

40

-1200 -800 -400 0 400 800 1200

I (m

A)

V (mV)


-40

-30

-20

-10

0

10

20

30

40

-1200 -800 -400 0 400 800 1200

I (m

A)

V (mV)



92

which means heating of mesas, when the area of mesas are decreasing. Also due to the heating, the spaces between quasiparticle branches become smaller when the bias voltage goes to high values.

Kurter et al. showed that the back bending of the I–V curves due to the heating is coming from the shape of the R qp (T) (the quasiparticle resistance) in Bi2212.

Therefore, in order to fi nd the mesa temperatures along the I–V curves, the V/I and the R qp (T) values are compared and they obtained disappearance of backbending above ~60 K [ 9 ]. Kurter et al. also showed that the high and sharp peaks in the conductance of intrinsic Josephson junctions in Bi2212 mesas are a result of heating in spite of the fact that they were regarding as superconducting energy gap [ 10 ]. Therefore, we can conclude that heating is very important for generation of terahertz radiation. Here, the large current density and close proximity of neighboring junctions in Bi2212 junction arrays are interpreted as the cause of such heating effects. Some researchers have used several methods to reduce the heating effects such as reduction of the current density by intercalation of some molecules such as HgB 2 , HgI 2 , I 2 within the Bi-O bilayer, which increase the c-axis resistance. In addition, using short pulses and decreasing the mesa dimensions are other methods to decrease the self-heating effects [ 9 ].

12.4 Conclusion

We studied area dependence of Josephson critical current density in superconducting Bi2212 mesas for terahertz emission. In the R-T measurements, sharp phase transitions to superconducting state were seen. From the R-T graphs, we see that samples show nearly underdoped behaviour. We observed heating effects on the current-voltage curves due to the large volume of the mesas. In the I–V measure-ments, hysterical tunneling behavior of Bi2212 and many number of quasiparticle branches were seen. Finally, we see that critical current density is decreasing with increasing mesa area due to the mesa heating, which has very signifi cant effects on mesa properties.

Acknowledgments This research is partially supported by TUBITAK (Scientifi c and Technical Research Council of Turkey) project number 110T248.

References

1. Tonouchi M (2007) Cutting-edge terahertz technology. Nat Photon 1(2):97–105 2. Tachiki M, Koyama T, Takahashi S (1994) Electromagnetic phenomena related to a low

frequency plasma in cuprate superconductors. Phys Rev B 50:7065 3. Josephson BD (1962) Possible new effects in superconductive tunneling. Phys Lett 1:251–253 4. Ozyuzer L, Koshelev AE, Kurter C, Gopalsami N, Li Q, Tachiki M, Kadowaki K, Yamamoto

T, Minami H, Yamaguchi H, Tachiki T, Gray KE, Kwok WK, Welp U (2007) Emission of coherent THz radiation from superconductors. Science 318:1291

H. Saglam et al.

93

5. Turkoglu F, Koseoglu H, Demirhan Y, Ozyuzer L, Preu S, Malzer S, Simsek Y, Muller P, Yamamoto T, Kadowaki K (2012) Interferometer measurements of terahertz waves from Bi 2 Sr 2 CaCu 2 O 8+d mesas. Supercond Sci Technol 25:125004

6. Ozyuzer L, Simsek Y, Koseoglu H, Turkoglu F, Kurter C, Welp U, Koshelev AE, Gray KE, Kwok WK, Yamamoto T, Kadowaki K, Koval Y, Wang HB, Müller P (2009) Terahertz wave emission from intrinsic Josephson junctions in high-Tc superconductors. Supercond Sci Technol 22:114009

7. Zhu XB, Wei YF, Zhao SP, Chen GH, Yang HF, Jin AZ, Gu CZ (2006) Intrinsic tunneling spectroscopy of Bi2Sr2CaCu2O8+δ: the junction-size dependence of self-heating. Phys Rev B 73:224501

8. Suzuki M, Hamatani T, Yamada Y, Anagawa K, Watanabe T (2009) Signifi cantly doping- dependent Josephson critical current – inhomogeneity in real space or heterogeneity in k-space. J Phys Conf Ser 150:052252

9. Kurter C, Gray KE, Zasadzinski JF, Ozyuzer L, Koshelev AE, Li Q, Yamamoto T, Kadowaki K, Kwok W-K, Tachiki M, Welp U (2009) Thermal management in large Bi2212 mesas used for terahertz sources. IEEE Trans Appl Supercond 19:428

10. Kurter C, Ozyuzer L, Proslier T, Zasadzinski JF, Hinks DG, Gray KE (2010) Counterintuitive consequence of heating in strongly-driven intrinsic junctions of Bi2Sr2CaCu2O8 mesas. Phys Rev B 81:224518



Abstract The as-grown Bi 2 Sr 2 CaCu 2 O 8+δ (Bi2212) crystals were heat-treated at various temperatures either in Argon atmosphere or in vacuum. Bi2212 triple mesa structures were fabricated by electron beam lithography and Ar ion beam etching processes. A set of samples has identical size (50 × 300 μm 2 ) are vertically mounted in vacuum on the cold fi nger of an optical LHe continuous-fl ow cryostat and resistance versus temperature and current-voltage measurements achieved. We investigated and compared characteristics of three mesas which are on same chip and next to each other. By this way, we searched the crystal inhomogenity in triple mesa structures and studied how critical current density varies with the doping conditions.

Keywords Terahertz radiation • Terahertz emitters • Intrinsic Josephson junctions • Josephson plasma • Bi2212 single crystals • Ac-Josephson effect • High-Tc super-conductors • Continuous THz sources

13.1 Introduction

Rapidly increasing applications of the electromagnetic waves in the under developed terahertz frequency range requires a well understood technique of effi cient terahertz (THz) wave generation. Terahertz technology is an extremely attractive research fi eld

Chapter 13 Inhomogenity of Micron-Sized Triple Terahertz Emitters Fabricated from Intrinsic Josephson Junctions in Single Crystal Bi 2 Sr 2 CaCu 2 O 8+δ

Yasemin Demirhan , F. Turkoglu , H. Koseoglu , H. Saglam, N. Miyakawa , K. Kadowaki , and L. Ozyuzer

Y. Demirhan (*) • F. Turkoglu • H. Koseoglu • H. Saglam • L. Ozyuzer Department of Physics , Izmir Institute of Technology , Urla, Izmir 35430 , Turkey e-mail: [email protected]

N. Miyakawa University of Tsukuba , Tsukuba , Japan

K. Kadowaki Department of Applied Physics , Tokyo University of Science , Tokyo , Japan

96

but there is still lack of compact solid state sources. Improving THz source effi ciency will benefi t all application areas, including imaging, spectroscopy, information technology, medical diagnosis etc [ 1 ]. The Josephson junction which could serve for this purpose is composed of two superconductors that has been separated by an insulator at nanometer order. An application of 1 mV results in a radiation with a frequency of 0.48 THz. Thus, a Josephson junction is a simple device which could convert DC voltage to THz electromagnetic waves. However, the radiation strength of a single Josephson junction is at the order of pW. Layered high temperature superconductor Bi 2 Sr 2 CaCu 2 O 8+δ (Bi2212) has the natural stacks of Josephson junctions, called intrinsic Josephson junctions [ 2 ]. In pursuit of the preliminary investigations on THz emission from intrinsic Josephson junctions (IJJ) in HTS [ 3 – 4 ], Ozyuzer et al. [ 5 ] was successful in observing the strong emission of electromag-netic waves directly where they have made use of a mesa of Bi2212 crystal without any external magnetic fi eld. THz emitting device which is shaped in the form of a mesa and which has a quite large power operates as a dc voltage to high frequency converter. Along the c-axis of the mesa structure an external current is applied which gives rise to a cavity resonance mode of wave by the ac Josephson current in the resistive state and ultimately some of the electromagnetic energy at resonance is transformed to THz radiated devices at the resonance frequency. Dated from discovery of the THz emission from IJJs, promotive progress has been encountered both from experimental and theoretical points of view. From the theoretical side synchronization mechanisms for the IJJs have been proposed namely Hu, Lin, and Koshelev put forward the following new mechanism [ 7 – 8 ]. In the case of strong inductive interaction between the superconducting CuO layers in Bi2212, kink structures take place in the phase difference of superconducting order parameter between the superconducting layers. The phase kinks induce cavity resonance modes of the Josephson plasma. This is a new dynamic state caused by the nonlinear effect special in the IJJ system.

Experimental investigations have been successful in obtaining power up to >30 μW and 2.5 THz from Bi2212 crystals [ 9 ]. Moreover, recent progress indicates that it is feasible to obtain coherent and large power THz emission when the Josephson oscillations of these stacked junctions can be synchronized. On the other hand, for the time being the mechanism of the emission is not clarifi ed yet and in addition to this, self heating effect leads to desynchronization which ends up with dramatic drop in emission power. Hot spot detection by low temperature laser scanning microscopy (LSTEM) is a crucial phase in realizing the mechanism. Existence of wave structures in the stacks is verifi ed by LSTEM [ 10 – 11 ] and moreover indicating the formation of electrothermal domains in such structures [ 12 ]. The THz emission properties are substantially affected by this heating phenomenon and thus control over its behavior is major interest. Using the charge injection method, superconducting properties of Bi2212 stacks can be tuned [ 13 ]. Bi2212 crystals with stacked Josephson junctions have a large energy gap and wide doping range [ 14 ] which is suitable for the THz region.

Optimizing the doping level, to signifi cantly improve the control over the super-conducting properties for powerful THz emission was the main objective of this

Y. Demirhan et al.

97

research. By the annealing procedure that we performed in our experiments we can change the superconducting properties as well as the critical supercurrent density of an intrinsic junction stack. It is already shown that the THz emitting mesas are below a certain underdoped level, which has relatively small critical current in contrast to optimally doped and overdoped Bi2212 [ 15 ]. Because of small critical current, large area mesas fabricated from underdoped crystals cause less heating and backbending occurs after the cavity resonance in voltage scale. So, powerful THz radiation can be obtained before heating severely affects the local mesa temperature.

13.2 Sample Fabrication

In this work, experiments were performed on as-grown Bi2212 single crystals which were grown using travelling solvent fl oating zone technique. In this work, in order to obtain various doping levels, we annealed the high temperature super-conducting Bi2212 single crystals in vacuum or Argon atmosphere with different heat treatment recipes. The heat treatment duration is varied to change the Tc and critical current of crystals. In the fi rst set up which has been constituted from Argon gas, fl owmeter, gettering furnace and annealing furnace, after placing the crucible inside the quartz tube; 100 sccm argon gas has been send to the annealing furnace right after purifi cation.

In the second annealing set up, Fig. 13.1b , we have used only the vacuum pump and the furnace. At fi rst we have reduced the pressure of the quartz tube in which we have located the crystals up to 10 −3 Torr using the turbomolecular pump. Afterwards we have operated the furnace which has been previously programmed to desired temperature and time in order to establish annealing at different temperatures. For further processing, single crystal of Bi2212 is glued onto a sapphire substrate from its smooth a-b surface by silver epoxy. In order to get a fresh and smooth surface on Bi2212, the crystal was then cleaved with an adhesive tape and Au layer

Flowmeter

Argon

GetteringFurnace

AnnealingFurnace Turbomolecular pump Annealing furnace

turbopump

Fig. 13.1 Annealing system with Ar fl ow ( left ), annealing system in vacuum ( right )

13 Inhomogenity of Micron-Sized Triple Terahertz Emitters Fabricated…

98

with the thickness of 100 nm was thermally deposited on the cleaved crystal surface immediately to prevent the chemical reactions. To obtain natural IJJ stacks with various size and height, mesas on Bi2212 crystals have been fabricated using e-beam lithography and argon ion beam etching techniques. Because of the diffi culties in making a contact on small area of the mesa, fi rstly CaF 2 insulating layer is deposited by evaporation onto crystal and a gold stripe is composed by e-beam lithography on insulating layer and mesa. Finally three gold probe wires are connected on two path and mesa by silver epoxy. After the mesa fabrication, the exact dimensions of the mesas were obtained using surface profi lometer and atomic force microscope. The number of Josephson junctions were determined which gives emission voltage. The electrical characterization of the mesas was obtained room temperature through low temperatures. In order to characterize the Bi2212 mesas, c-axis resis-tance versus temperature (R–T), and current–voltage behavior (I–V ) were measured in a He fl ow cryostat (Fig . 13.2 ).


13.3.1 Temperature Dependence of the Mesa Resistance

Figure 13.3a presents 50 μm width mesa structures annealed at various conditions, the R(T c )/R(300 K) values give idea about the doping level. We observe that the overall magnitude of c-axis resistivity of the crystal increases with decreasing doping level. At variation of oxygen level from the optimally doped to overdoped, c-axis resistivity of the crystal gradually start to exhibit metallic behavior. For underdoped samples, R–T characteristics exhibit semiconductor like behavior above T c [ 14 ].

The curves exhibit a typical temperature dependence of the c-axis resistance of Bi2212, which tends to increase with decreasing temperature around room temperature and turns to sharp increase near T c . Mesa resistance is fi nite even below T c . This is because of the contact surface resistance due to three terminal measurement.

Fig. 13.2 SEM images of 300 × 50 μm 2 triple mesa structures on Bi2212 crystal fabricated on Bi2212 crystal with different magnifi cations

Y. Demirhan et al.

99

However, the contact resistance is usually smaller than a few ohms so that it seems not to infl uence the measurement, but it may contribute to the heating, which will be discussed below.

Watanabe et al. have systematically studied the oxygen doping level dependence of the c-axis resistivity of Bi2212 and, R(T c )/R(300 K) >4 indicates a necessity of a certain doping range of δ ~ 0.22 for THz emission [ 11 ]. As seen in Fig. 13.3b , the R(T c )/R(300 K) value for one of the as-grown and underdoped crystals is 2 and 4.2 respectively. This shows the importance of oxygen doping level for THz emission (Fig. 13.4 ).

0

20

40

60

80

100

0 50 100 150 200 250 300

R(o

hm)

T(K)

0

20

40

60

80

100

0 50 100 150 200 250 300

R(o

hm)

T(K)

As-grown

Underdoped

a b

Fig. 13.3 ( a ) R–T measurements of 300 × 50 μm 2 mesa structures with various doping levels. ( b ) R–T measurements of 300 × 50 μm 2 as-grown and underdoped mesa structure

Fig. 13.4 R–T measurements of 300 × 50 μm 2 triple mesa structure on the same crystal 100 μm apart from each other


100

Our samples so far prepared have T c between 75 and 90 K, which indicate that the mesa is in the under-doped condition. The transition width, ΔT c , is sharp and is mostly 1–2 K. We have seen that R(T c )/R(300 K) value is roughly 2.5 for this sample. This indicates that fabricated mesa is near optimally doped. For all THz emitting mesas, this value is larger than 4 [ 13 – 15 ].

13.3.2 The I–V Curves

I–V characteristic of Bi2212 mesas show a number of characteristic features of the multi stacked intrinsic Josephson junction. They includes several branches due to individual switching of each junction from Josephson state to quasipar-ticle state when the bias current exceeds the individual critical current of each junction.

Figure 13.5 shows a comparison of I–V measurements of THz emitting mesas with same dimensions. Although they are on the same crystal, the backbending voltage and critical current values of the mesa structures whose critical tempera-tures exhibit discrepancy are different from each other due to the inhomogenity of the crystal. We can also observe from the transitions in resistance temperature graphics that the crystal is inhomogenous. In nonuniform mesas doping levels of the intrinsic Josephson junctions could be slightly different. In the triple mesa structures that we have changed doping, critical current densities are 180, 186 and 200 A/cm 2 for 50 ×300 μm 2 mesas respectively.

-60

-40

-20

0

20

40

60

-1200 -800 -400 0 400 800 1200

I(m

A)

V(mV)

Fig. 13.5 I–V characteristics of 300 × 50 μm 2 triple mesa structure on same crystal at 20 K

Y. Demirhan et al.

101

13.4 Conclusion

In this study, we presented data on triple mesa structures fabricated on Bi2212 crystals. The optimum conditions to make high power THz emitting Bi2212 devices were studied. By the annealing procedure, with different heat recipes we changed the doping level of the crystals. The estimated heat reduction is improved in com-parison with the previous reports. The measurements revealed distinct features of intrinsic Josephson junctions and the mesas exhibit usually underdoped behavior. We observed hysteretic quasiparticle branches in the performed I–V measurements are all in agreement with the mechanism of terahertz emission. Currently, the effect of crystal inhomogenity on THz emission is not known and under investigation. We realize the fact that our results are strongly infl uenced by crystal defects and impurities. To obtain homogenous doping level we should fabricate the mesa structures on the homogenous side of the crystal.

Acknowledgments This research was supported in part by the TUBITAK (Scientifi c and Technical Council of Turkey) project number 110T248.

References

1. Tonouchi M (2007) Cutting-edge terahertz technology. Nat Photon 1(2):97–105 2. Kleiner R, Steinmeyer F, Kunkel G, Müller P (1992) Intrinsic Josephson effects in

Bi 2 Sr 2 CaCu 2 O 8 single crystals. Phys Rev Lett 68:2394 3. Tachiki M, Iizuka M, Minami K, Tejima S, Nakamura H (2005) Emission of continuous

coherent terahertz waves with tunable frequency by intrinsic Josephson junctions. Phys Rev B 71:134515–1–134515–5

4. Bae MH, Lee HJ, Choi JH (2007) Josephson-vortex-fl ow terahertz emission in layered high-Tc superconducting single crystals. Phys Rev Lett 98:027002

5. Ozyuzer L, Koshelev AE, Kurter C, Gopalsami N, Li Q, Tachiki M, Kadowaki K, Tamamoto T, Minami H, Yamaguchi H, Tachiki T, Gray KE, Kwok WK, Welp U (2007) Emission of coherent THz radiation from superconductors. Science 318:1291

6. Koshelev AE, Bulaevskii LN (2008) Resonant electromagnetic emission from intrinsic Josephson-junction stacks with laterally modulated Josephson critical current. Phys Rev B 77:014530

7. Hu X, Lin S (2008) Three-dimensional phase-kink state in a thick stack of Josephson junctions and terahertz radiation. Phys Rev B 78:134510–134511

8. Kadowaki K, Tsujimoto M, Yamaki K, Yamamoto T, Kashiwagi T, Minami H, Tachiki M, Klemm RA (2010) Evidence for a dual-source mechanism of terahertz radiation from rectangular mesas of single crystalline Bi 2 Sr 2 CaCu 2 O 8+δ intrinsic Josephson junctions. J Phys Soc Jpn 79:023703

9. Wang HB, Guénon S, Yuan J, Iishi A, Arisawa S, Hatano T, Yamashita T, Koelle D, Kleiner R (2009) Hot spots and waves in Bi 2 Sr 2 CaCu 2 O 8 intrinsic Josephson junction stacks – a study by low temperature scanning laser microscopy. Phys Rev Lett 102:017006

10. Wang HB, Guénon S, Gross B, Yuan J, Jiang ZG, Zhong YY, Gruenzweig M, Iishi A, Wu PH, Hatano T et al (2010) Coherent terahertz emission of intrinsic Josephson junction stacks in the hot spot regime. Phys Rev Lett 105:057002


102

11. Guénon S, Grünzweig M, Gross B, Yuan J, Jiang Z, Zhong Y, Iishi A, Wu P, Hatano T, Koelle D et al (2010) Interaction of hot spots and THz waves in Bi 2 Sr 2 CaCu 2 O 8 intrinsic Josephson junction stacks of various geometry. Phys Rev B 82:214506

12. Koval Y, Jin X, Bergmann C, Simsek Y, Özyüzer L, Müller P, Wang HB, Behr G, Büchner B (2010) Tuning superconductivity by carrier injection. Appl Phys Lett 96:082507

13. Watanabe T, Fuji T, Matsuda A (1997) Anisotropic resistivities of precisely oxygen controlled single-crystal Bi 2 Sr 2 CaCu 2 O 8+δ : systematic study on “Spin Gap” effect. Phy Rev Lett 79:2113

14. Ozyuzer L, Simsek Y, Koseoglu H, Turkoglu F, Kurter C, Welp U, Koshelev AE, Gray KE, Kwok WK, Yamamoto T, Kadowaki K, Koval Y, Wang HB, Muller P (2009) Terahertz wave emission from intrinsic Josephson junctions in high-Tc superconductors. Supercond Sci Technol 22:114009

15. Turkoglu F, Koseoglu H, Demirhan Y, Ozyuzer L, Preu S, Malzer S, Simsek Y, Muller P, Yamamoto T, Kadowaki K (2012) Interferometer measurements of terahertz waves from Bi 2 Sr 2 CaCu 2 O 8+d mesas. Supercond Sci Technol 25:125004

Y. Demirhan et al.


Abstract A technique for revealing surface morphology of human cervical cancer cells has been developed to facilitate early diagnostics of a pre-cancer and cancer cells under refl ected light microscopy. To measure spectral features of morphological markers of cervical cancer cells, so named disperse lightened particles (DLP), we used Synchrotron based Fourier Transformed IR Micro-spectroscopy (SB FTIRM) in the mid-IR range (2–25 mkm wavelength). We used point-by-point IR microspectroscopy analysis in confocal geometry for high resolu-tion for cervical cancer cells.

Keywords Fourier transformed MIR Microspectroscopy • Spectral markers • Cancer cells • Cell morphology • Light microscopy • Cytology

14.1 Introduction

THz and Mid Infrared Radiation are very relevant tools to detect and image explosives and CBRNs. In present paper we are using combination of NIR and Mid Infrared Radiation to investigate cancer cells can be potentially adapted to image explosives and CBRN agents too.

Cervical cancer is the second most common cancer in women worldwide. More than 80 % of cervical cancers occur in the developing world where the least resources exist for management. Present diagnostic technologies to detect early cervical cancer cells use the Pap-test on cervical smears in stationary diagnostic laboratories in hospitals or diagnostic centers. But there is a major problem with analyzing Pap

Chapter 14 The Fourier Transformed MIR Microspectroscopy to Reveal a Morphological and Spectral Markers of a Cervical Cancer Cells

A. A. Paiziev

A. A. Paiziev (*) Institute of Ion-Plasma and Laser Technologies , Uzbek Academy of Science , Durmon Yuli 33 , 100125 Tashkent , Uzbekistan e-mail: [email protected]

104

smears, connected with drying changes in cells due to delays in applying the fi xing procedure. It is very diffi cult to evaluate smears that have dried in air and still distinguish abnormal cells. Sample handling and its evaluation is a time consuming procedure and demands a sophisticated testing infrastructure and highly trained professionals to evaluate the cytological test in a stationary environment.

The new method (named AKPa-test according to core patent authors names: Abdullakhodjaeva-Krakhmalev-Paiziev) [ 2 ], has been developed in the Institute of Electronics, Uzbek Academy of Science, and successfully piloted in Pathology Institute in Uzbekistan. The method is low cost, quick time-to-answer, has good sensitivity and specifi city, and is able to be used by minimally trained personnel. Furthermore, the new AKPa-test allows evaluation of the tissue without the need for chemical reagents and equipments. Thus the new test provides the ability to be mobile, allowing implementation of diagnostic services on-site.

But up to now we do not know nature and chemical compound of DLP content. We suppose that suitable method to detect and identify content of DLP is FTIR microspectroscopy based on synchrotron radiation. It may be hypothesized that the and normal cells may be exploited by means of FTIR microspectroscopy provided that clear IR absorption signal can be detected from DLP formations. DLP formations located on cell cytoplasm around nuclear. The size of DLP on subcel-lular level and it demands to use of non-conventional IR sources, in particular the use of synchrotrons. The broadband and high brightness electromagnetic radiation that may be obtained from SR should help towards the characterization of these observed morphomarkers (DLP) on cervical cancer cell membrane.

The potential of new diagnostic platform is based on visualization of morpho-logical features of cervical cancer cell membrane instead of intracellular morphological features as in the standard Pap-test. We are expecting to determine spectral parameters of sub-cellular formations of cervical cancer cells at early stage of cancerogenesis and identify biochemical status of DLP. IR microspectroscopy beamlines allows to use this technology in biomedical applications where the area of great interest is cancer diagnostics.

14.2 Experimental Methods

Previously we have used refl ected light microscope (RLM) to observe DLP on a cancer cell membrane. Smears from uterine neck were collected for patients with diagnosis cervical cancer. Smears were collected by Ayre’s spatula after exposing the cervix by a Cusco’s speculum. Samples collected were transferred to glass slides. Glass slides were preliminarily cleaned thoroughly in a 2 vol.% detergent followed by repeated washing in pure deionized water. Two set of slides were prepared for each patient and fi xed by 95 % ethanol. Relevant information was obtained from the patient and recorded on a specially designed proforma. The fi rst set of marked slides were then sent to cytology laboratory to view at high magnifi -cation with RLM without fi xing and any treatment by dyes. This native smears has

A.A. Paiziev

105

been observed under optical refl ected microscope Neophot-2. For each smear under investigation from 4 to 10–13 fi eld of view was captured. In this case we watched a morphology and topology of cell membrane and captured this images on digital photocamera. Both practitioner and patient can then see smears image in color. The results are then used as a basis for prescribing supplements. The second set of smears was prepared on glass and fi xed by 95 % ethanol and stained with Papani-colaou and each slide was then carefully examined by a cytopathologist to distinguish normal cells from leased one. Relevant information was recorded on a specially designed proforma on PC and was marked on the slides. It was clearly specifi ed whether smear was satisfactory or not. Slides showing some abnormal changes in the cellular pattern were further scrutinized by a cytopathologist. Images has been observed on optical microscope JENAMED-2 and captured by digital high Resolution Microscopy Camera AxioCam. To measure spectral features of DLP need IR microspectroscopy beamline for diffraction -limited spatial resolution and spectral optimum brightness in the mid-IR range (2–25 mkm wavelength). There are two modes of operation. one of them is point-by-point IR microspectroscopy analysis in confocal geometry for high resolution what can provide the highest sensitivity IR spectra at the single point in the mid-IR and it is suitable to probe subcellular formations what we have observed before on cancer cell membrane (DLP).


14.3.1 Morphologic Markers

Neoplastic transformed cells have numerous essential features which may be identi-fi ed after staining by dyes (Pap-test, Fig. 14.1a ). But the above mentioned cancer cell signs are situated inside of the cell (nuclear, cytoplasm. Fig. 14.1a ).

Unlike the Pap-test, AKPa-test will handle the morphological signs of cancer cell membrane. These features are connected with exocytose when metabolism products of cancer cells move up to membrane inside the cell. Consequently in patients with early stage cancer, there are numerous protrusions of the cell membrane surface with pronounced light refractivity. The core innovation of our proposal is connected with the discovery of new cancer cell membrane surface morphological markers. In micrographs these markers look like disperse lightened particles (DLPs) which we observe under a conventional light microscope (Fig. 14.1c, d ).

14.3.2 Spectral Markers

To measure spectral features of DLP we used Synchrotron based Fourier Transformed IR microspectroscopy (Fig. 14.2 ) in the mid-IR range (2 to 25 mkm wavelength).

14 The Fourier Transformed MIR Microspectroscopy to Reveal a Morphological…

106

Fig. 14.1 Cervical cancer cells: ( a ) staining according to Pap-test, ( b ) AFM image of cervical cancer cells [ 1 ], ( c ) Morphological markers of cervical cancer cells at early stage ( c ) and last stage ( d ) of cell membrane destruction

Fig. 14.2 Synchrotron based Fourier transformed infrared spectra I abs ( a ) on difference points (5 points) of cancer cell membrane ( b )

A.A. Paiziev

107

We used point-by-point IR microspectroscopy analysis in confocal geometry for high resolution for cervical cancer cells (Fig. 14.2a ). Screening methods based on IR spectroscopy offer potential advantages over screening by conventional cytology, which include faster and less subjective analysis, reduced cost and the potential for automation. In micrographs these protrusions look like disperse lightened particles (DLPs) which we observe under a conventional light microscope (Fig. 14.1c, d ). Here we can see similar protrusions observed under AFM too [ 1 ] (Fig. 14.1b ).

14.4 Conclusion

Much is yet to be known about the nature of endometrial cancer cells and until now there has not been a reliable, simple method for visualizing cell topography in air, at high resolution without fi xation and dehydration. Being able to directly view membrane structures regulated by exocytosis will enable researchers to analyze the secretory nature and response of cells, yielding insights into drug responses and effects. Considerable variability in the sizes of dimple depressions and ruptures, as well as dynamic formation and grouping of these structures around the nucleus, illustrates that cells have diverse morphologies.

In summary, this paper shows that in comparison of Pap-test offered AKPa-test have (see Table 14.1 ):

1. Better sensitivity (about 100 %) 2. Two time cheaper in comparison with Pap-test 3. Much more express. To test one sample need 5–10 min (Pap-test take 0.5 day) 4. Easy-to-work. Two stage evaluate of sample. (Pap-test is many-stage treatment

and evaluation (5–6 steps)) 5. A little number of smears diagnostic signatures (3) against Pap-smear (8–10)

Acknowledgments Author gratefully acknowledge partial funding by the PSI for access to the SLS-facility for the experiments, UNFPA and OSCE for travel support.

Table 14.1 Comparison of Pap- and AKPa-test parameters

Test parameters PAP-test AKPa-test

1. Sensitivity 70 % 100 % 2. Cost ~$10 ~$5–6 3 Test duration 3–4 h 10 min 4. Complicacy (preparing steps) 5–6 2–3 5. Number of diagnostic signs 10–15 3–4 6. Electronic data processing No Yes 7. Qualifi ed personal requirement Yes No 8. Laboratory infrastructure requirement Yes No

14 The Fourier Transformed MIR Microspectroscopy to Reveal a Morphological…

108

References

1. Muys JJ, Alkaisi MM, Melville DOS, Nagase J, Sykes P, Parguez et al (2006) Cellular transfer and AFM imaging of cancer cells using Bioimprint. J Nanobiotechnol 4:1–10

2. Paiziev A, Krakhmalev V (2010) Color image of red blood cells on the solid surface. Int J Lab Hematol 32(Suppl 1):1–180

A.A. Paiziev


Abstract We propose a specially designed THz diffractive element for imaging purposes. The designing process was performed to obtain low attenuation and small weight. The element is optimized for broadband application, especially for remote detection of harmful materials. It provides low both chromatic and geometric aberrations.

Keywords Diffractive lens • High Order Kinoform • Imaging • Numerical modeling • Security • Terahertz

15.1 Introduction

Passive THz surveillance systems are constantly awaited by the industry. Recent development of terahertz technology has led to the construction of passive cameras working at the frequencies within the range of 0.05–0.3 THz, for example by ThruVision and Brijot corp. The ability to reconstruct an image of an object, based on its own natural radiation is one of the most signifi cant advantages of passive systems. Hence the external THz illumination source is not necessary, leading to smaller over-all costs of ownership. The exploited fact is that all objects in ambient temperatures emit radiation in a wide range of frequencies with a maximum in far infrared (8–14 μm). However, these objects radiate also in terahertz part of the spectrum.

Chapter 15 THz Diffractive Optical Element for Passive Imaging

A. Czerwinski , P. Zagrajek , E. Rurka , N. Palka , M. Szustakowski , J. Suszek , A. Siemion , M. Makowski , and M. Sypek

A. Czerwinski (*) • J. Suszek • A. Siemion • M. Makowski • M. Sypek Faculty of Physics , Warsaw University of Technology , Koszykowa 75 , 00-662 Warsaw , Poland e-mail: [email protected]

P. Zagrajek • E. Rurka • N. Palka • M. Szustakowski Military University of Technology , Gen. Sylwestra Kaliskiego 2 , 00-908 Warsaw , Poland

110

Passive systems detect this weak signal and convert it into the usable electrical signal. Due to a very low energy, THz cameras must deal with very high level of noise.

In order to achieve a satisfactory signal to noise ratio (SNR), one of the solutions is the application of large-aperture, low-attenuating optical elements like lenses [ 1 , 2 ] or mirrors [ 3 ] with typical parameters: focal length/diameter ≈1 and diameters from 200 to 500 mm. The advantage of mirrors is the achromaticity and smaller loses in comparison with lenses. On the other hand mirror optics is very expensive to manu-facture and maintain. One must also remember that mirrors are very sensitive to geometrical aberrations. Plastic lenses have larger losses than mirrors, nevertheless they are lighter, cheaper, easier to produce and maintain. They can be also cor-rected for geometrical and chromatic aberrations in a certain range of wavelengths.

Here we present the results of imaging in the THz range using a High Density Polyethylene (HDPE) diffractive element with a F/D equal to 1 and the ThruVision TS4 scanner.

15.2 Design and Experimental Setup

In order to meet the requirements of the optical system for passive cameras mentioned in the previous section, a double-sided diffractive element in the form of a High-Order Kinoform (HOK) [ 4 ] was designed. It had a diameter of 300 mm and a focal length equal to 300 mm. Such a large diameter is used in order to direct as much energy as possible onto the detectors. Moreover, small thickness (center thickness was 17.6 mm, edge thickness was 2 mm, back focal length was 291 mm) provides low losses due to attenuation of the lens material (α = 1.89 cm −1 at wavelength λ = 1 mm). The design wavelength was λ d = 1 mm. The refractive index of HDPE for λ d is 1.528 (Fig . 15.1 ).

Fig. 15.1 Overview of the manufactured HOK lens

A. Czerwinski et al.

111

The lens was not a classic Fresnel Lens, where the difference in the thickness of the zones is much bigger than the design wavelength. In contrast, the lens belonged to the class of High-Order Kinoforms (HOK). Figure 15.2 shows the principal difference in the defi nition and the dimensions of the HOK in comparison with FL.

The experimental setup was built in order to confi rm the imaging capabilities of the HOK lens. It consisted of a THz radiation source – a black body set to temperature T = 45 °C, the ThruVision TS4 scanner and a transparency with two hot holes, as shown in Fig. 15.3 . This test object was cut from the paper covered with aluminum duct tape. Figure 15.3 shows the experimental setup with the details of the imaged transparency object.

The black body used in experiment is a thermally stabilized device with emissivity equal to 0.99. It was used to obtain higher contrast and better signal to noise ratio.

The ThruVision TS4 scanner is a system with a scanning mirror, based on a GaAs Schottky mixer. The detected radiation is converted to the image with resolu-tion of 80 × 150 pixels. The TS4 camera provides relatively sharp image of the object located in the distance greater than 3 m from the system. In a normal working confi guration it has an object plane resolution of 1 cm. The functional scheme of the TS4 scanner is shown in Fig. 15.4 .

In the system from Fig. 15.3 , the theoretical image plane of the lens was located at the same distance to the TS4 scanner as the object in a normal working

Fig. 15.2 Design and dimensions of Fresnel Lens ( FL ) and High Order Kinoform ( HOK )

Fig. 15.3 Test transparency object

15 THz Diffractive Optical Element for Passive Imaging

112

confi guration without any additional lenses. Afterwards, the location of the lens and object plane was estimated using the lens equation 1/x + 1/y = 1/f, where x and y are the distances from the lens to the object and to image plane respectively, and f is the focal length of the lens. The best image sharpness was achieved when the theo-retical magnifi cation of the system reached about 1.83 (accuracy ±0.04). The images created in the theoretical image plane were obviously real and inverted.

15.3 Results of the Numerical Modeling

The numerical experiment was conducted based on the described confi guration, aimed at the extraction of point-spread functions (PSF) for different angles of incidence of the radiation and for different wavelengths. The confi guration of the virtual experiment is shown in Fig. 15.5 .

Modeling was performed using off-axis modifi cation of the convolution method [ 5 ]. The lens was simulated as a thick non-attenuating phase-only element. All simulations were carried out for the wavelength λ = 1 mm. The results of the modeling are shown in Fig. 15.6 .

The evaluated PSFs for the angle of 12° fall within the size of a single detector, therefore the resolution will not be affected by the increase of the PSF spot in this case. Therefore the proposed lens can be used successfully for high numerical aperture systems. Obviously for shorter wavelengths the PSF size decreases, which potentially enables the imaging with a proportionally higher resolution, provided that the detectors are smaller.

Fig. 15.4 The functional scheme of the used ThruVision TS4 scanner


Fig. 15.5 Confi guration of the numerical experiment to measure the PSF spots for a variable wavelength and incidence angles

1000 µm

600 µm

300 µm

200 µm12 deg6 deg0 deg

Fig. 15.6 The point spread function spots for a variable wavelength and incidence angles

114


Results obtained in the experimental setup presented in Fig. 15.3 are shown in Fig. 15.7 . The results prove the imaging capabilities of the proposed lens. The test transparency is visible in three different orientations, which gives the clue about the resolution in variable lateral directions. Figure 15.7d shows the imaging result when the additional HOK lens is removed and the optimal imaging distance is set. In the last case the object is hardly visible, due to low signal to noise ratio. Moreover the imaging resolution is not big enough.

The purpose of this paper is to report the imaging performance of the designed lens. The TS4 camera was used as an imaging tool. The result in Fig. 15.7d is presented for comparison purposes only. The experimental results showed that the HOK lens provides enough radiation focusing capabilities to obtain suffi cient

Fig. 15.7 The experimental results of imaging with 3 orientations of the test transparency: (a) horizontal; (b) vertical; (c) 45 degrees; (d) imaging without the HOK lens – in the normal working confi guration


115

SNR and image size for ThruVision scanner to create clear image. The presented structure was not designed as an optical system for improvement of the TS4 resolution and imaging performance.

15.5 Conclusions

The experiments showed that terahertz passive imaging system (“a THz magnifying glass”) was successfully designed, manufactured and tested. The lens offers a low- attenuating radiation focusing with an optimized shape, providing lower geometric and chromatic aberrations. Due to a double-sided design the lens is lightweight, cheap and easy to manufacture in a widely available plastic material.

Acknowledgments This work was supported by the Polish Ministry of Science and Higher Education under the Project O N515 020140 and by the Polish National Science Center under grant N N515 498840, with a complementary support from the European Social Fund implemented under the Human Capital Programme (POKL).

References

1. Richter J, Hofmann A, Schmidt LP (2001) Proceedings of the 31st European Microwave conference, London

2. Sypek M, Makowski M, Hérault E, Siemion A, Siemion A, Suszek J, Garet F, Coutaz JL (2012) Highly effi cient broadband double-sided Fresnel lens for THz range. Opt Lett 37:2214–2216

3. Bruckner C, Notni G, Tunnermann A (2010) Optimal arrangement of 90° off-axis parabolic mirrors in THz setups. Optik 121:113–119

4. Sypek M, Makowski M, Hérault E, Siemion A, Siemion A, Suszek J, Garet F, Coutaz JL (2012) Highly effi cient broadband double-sided Fresnel lens for THz range. Opt Lett 37:2214–2216

5. Sypek M (1995) Light propagation in the Fresnel region. New numerical approach. Opt Commun 116:43–48

15 THz Diffractive Optical Element for Passive Imaging


Abstract Semiconductor intersubband gain media can be engineered to deliver dispersive gain shape, leading to interesting consequences for the coupling with light in microcavities giving rise to coupled polaritons and antipolaritons. This may open interesting possibilities for polaritonic devices in the THz range.

Keywords Many body effects • Nonequilibrium transport and optics • Intersubband transitions • Intersubband gain without inversion • Dilute nitrides • Dispersive gain • Polartions and antipolaritons

16.1 Introduction

Intersubband (ISB) transitions can lead to lasing without global inversion [1, 2] and dispersive gain [3–5]. In this paper we investigate the implications of the dispersive gain in the coupling between light in a microcavity and the material excitation that leads to coherently coupled polaritons [6] and antipolaritons [7–9] branches. Since possible lasing of intersubband polaritons is a current topic of interest [10] the coupled polariton/antipolariton can lead to new or improved sources for the THz detection of explosives and CBRNs, which have absorption signatures in this range.

In dilute nitride quantum wells, the strong interaction between the N resonant states and the conduction band edge means that the conventional eight-band k·p method cannot be applied directly to GaAsN and related heterostructures. The interaction between the N resonant states and the conduction band edge to describe the variation of the zone-center conduction band edge energy with N, leading to a modified ten-band Hamiltonian, which describes on average the interaction of the

Chapter 16Intersubband Dispersive Gain Media

Mauro F. Pereira

M.F. Pereira (*) Materials and Engineering Research Institute, Sheffield Hallam University, Sheffield, UKe-mail: [email protected]

118

all N atoms with the host material [5]. Indeed, starting with this Hamiltonian, we have recently demonstrated that by adding a small concentration of Nitrogen in GaAs quantum wells, the bandstructure can be engineered to deliver strong intersubband gain without global population inversion. Depending on the resulting effective mass, a rather symmetric dispersive gain shape can be obtained [5]. The flexibility in engineering the gain shape in these systems makes them ideal candidates for this investigation. The system is globally out of equilibrium but the electrons are assumed to be independently thermalised within each subband with occupation functions characterised by temperatures which can be extremely different from the lattice temperature, similarly to the case of electrons in conduction-band based QCLs as found in micro-probe photoluminescence experiments [11]. The total number of electrons in each subband can be controlled in practice by optical pumping, selective doping or a combination of both methods.

In the numerics shown here there is no global population inversion, with the same total number of electrons in the upper (16.2) and lower (16.1) subbands, i.e. N1 = N2. As described in detail in Ref. [5], gain without global inversion is obtained by engineering the conduction band dispersion relations with the effective masses in the upper bands lower than those below. If the difference between masses is not very large, strongly dispersive gain is observed. Thus, here we combine both conditions, N1 ≈ N2 and m2 ≾ m1and of course allowed transition dipole moments, ℘12 ≠ 0. In the notation o Ref. [5] the transitions are between subbands e

32+ ( ) and e

11- ( ) with

average effective masses given respectively by m2 = 0.109 and m1 = 0.125 obtained for a 7 nm Ga0.98N0.02As-Al0.3Ga0.7As quantum well. Note that these energy levels lead to transition energies higher than the usual first two quantum well interconduction band transitions found in the literature.

16.2 Numerical Results and Discussion

The absorption α(ω) and gain spectra g(ω) = − α(ω) are evaluated within the context of a NGF approach.

a wpw

c w c w Ã cm n

mn nm( ) = ( ){ } ( ) = ( ) ( )¹å4 2

cn Vk k

b k

Im , .,

(16.1)

The susceptibility χ is directly related to the carriers Green’s functions G, which satisfies a Dyson equation [12–16]. The first step of the numerical scheme is the solution of the ten-band k·p Hamiltonian which includes the dilute nitrogen levels responsible for the extra nonparabolicity that gives rise to strong differences in effective subband masses. The Green’s functions and self-energies are expanded using eigenstates and eigenvalues of this Hamiltonian. The model system actually investigated in this paper is globally out of equilibrium but the electrons are assumed to be thermalised within each subband. For a feasibility study we control the total number of electrons in each subband, which can in practice be achieved, e.g.,

M.F. Pereira

119

by optical pumping, selective doping or a combination of both methods. Thus, the full NGF scheme is simplified and reduces to the self-consistent evaluation of chemical potentials and self-energy matrix elements which lead to subband energy renormalization, dephasing constants, and occupation functions. Finally, absorption and gain are given by the solution of the integro-differential equation for the optical susceptibility obtained from the carriers Green’s function in linear response. The numerical solution of the resulting equation is similar for both intersubband and interband cases [17–19]. The imaginary part of the optical susceptibility used as input after a simple analytical model adjusted to the exact numerical solution for the Ga0.98N0.02As/Al0.3Ga0.7As QW of Ref. [5] with the same global occupation in two subbands. The density for the top e2+ subband is the same in all curves, n = 1.0 × 1011 cm−2. The high energy compared to typical GaAs intersubband studies is due to the fact that higher subbands are involved in the study. For details see Ref. [5]. The sample geometry is depicted in Fig. 16.1.

The dispersion relations (ω vs θ) as shown in Fig. 16.2 are calculated as a function of incidence angle θ around the normal to the prism facets, from the solution of the resulting wave equation, following the approach of Ref. [7],

w hw e

ew

ee

2 2 24

4

2

21

4+ -

æ

èç

ö

ø÷ =W

WcI

b

R

b

,

(16.2)

Fig. 16.1 Diagram indicating the microcavity geometry considered. Light reflected at the bottom of the GaAs substrate reaches the AlAs layers with an incidence angle θ. The light is confined through internal reflection at the sample surface in one side and in the other side at the interface with the low index AlAs layer. The dispersions in the next figures are given as a function of θ following the convention used in the experiments described in Ref. [6]

16 Intersubband Dispersive Gain Media

120

where the dielectric constant is calculated using the susceptibility in Eq. (16.1) and the background dielectric constant εb, ε(ω) = εIR + iεI = εb + 4πχ(ω). The remaining parameters are η2 = (εs/εb)sin 2θ and Ωc

2 = π2c2/(Lc2εb), where the cavity length and

substrate dielectric function are given respectively by Lc and εs.The Fig. 16.2 shows a prediction of how a material excitation with a dispersive

gain profile couples with light in a microcavity. The density for the bottom and top subbands is the same, n = 1 × 1011cm− 2.

Even a single transition yields an interacting set of coupled intersubband antipo-lariton and polariton branches which can potentially lead to a new level of all- optical control and switching in a microcavity. In contrast to the conventional antipolariton case with an imaginary Rabi frequency, the dilute nitride quantum wells chosen show ample flexibility to engineer the dispersive gain shape without global inversion and have potential for real Rabi frequencies with measurable oscillations, suggesting this as the medium of choice to investigate the coupling of photonic modes with an excitation delivering dispersive gain.

16.3 Conclusion

In summary, this paper shows the impact of dispersive medium in laser linewidth and coupling of light in a microcavity. In forthcoming research, we shall extend the results for different structures in an effort to push the gain towards the THz domain,

Fig. 16.2 Dispersions of coupled polaritons and antipolariton branches resulting from the dispersive gain medium. The imaginary part of the optical susceptibility used as input after a simple analytical model adjusted to the exact numerical solution for the Ga0.98N0.02As/Al0.3Ga0.7As QW of Ref. [5] with the same global occupation in two subbands

M.F. Pereira

121

where polaritonic/antipolaritonic devices may have an impact for detection of different substances, including explosives and CBRN agents.

Acknowledgments The author acknowledges support from MPNS COST ACTION MP1204 – TERA-MIR Radiation: Materials, Generation, Detection and Applications and input dilute nitride bandstructure provided by S. Tomić.

References

1. Faist J, Capasso F, Sirtori C, Sivco DL, Hutchinson AL, Hybertsen MS, Cho AY (1996) Quantum cascade lasers without intersubband population inversion. Phys Rev Lett 76(3):411–415

2. Pereira MF Jr (2008) Intervalence transverse-electric mode terahertz lasing without population inversion. Phys Rev B 78(24):245305-1–245305-5

3. Wacker A (2007) Coexistence of gain and absorption. Nat Phys 3:298 4. Revin DG, Soulby MR, Cockburn JW, Yang Q, Manz C, Wagner J (2008) Dispersive gain and

loss in midinfrared quantum cascade laser. Appl Phys Lett 92:081110 5. Pereira MF, Tomić S (2011) Intersubband gain without global inversion through dilute nitride

band engineering. Appl Phys Lett 98:061101 6. Dini D, Köhler R, Tredicucci A, Biasiol G, Sorba L (2003) Microcavity polariton splitting of

intersubband transitions. Phys Rev Lett 90:116401 7. Pereira MF Jr (2007) Intersubband antipolaritons: microscopic approach. Phys Rev B 75:195301 8. Pereira MF (2009) The influence of dephasing in the coupling of light with intersubband

transitions. Microelectron J 40:841 9. Pereira MF (2008) Intersubband vs interband-light coupling in semiconductors. Opt Quant

Electron 40:325 10. de Liberato S, Ciutti C (2009) Stimulated scattering and lasing of intersubband cavity polaritons.

Phys Rev Lett 102:136403 11. Vitiello MS, Iotti RC, Rossi F, Mahler L, Tredicucci A, Beere HE, Ritchie DA, Hu Q,

Scamarcio G (2012) Appl Phys Lett 100:091101 12. Nelander R, Wacker A, Pereira MF Jr, Revin DG, Soulby MR, Wilson LR, Cockburn JW,

Krysa AB, Roberts JS, Airey RJ (2007) Fingerprints of spatial charge transfer in quantum cascade lasers. J Appl Phys 102:113104

13. Pereira MF Jr, Nelander R, Wacker A (2007) Characterization of intersubband devices combining a nonequilibrium many body theory with transmission spectroscopy experiments. J Mater Sci Mater Electron 18:689

14. Pereira MF Jr (2011) Microscopic approach for intersubband-based thermophotovoltaic structures in the terahertz and mid-infrared. J Opt Soc Am B 28:2014

15. Pereira MF Jr, Henneberger K (1998) Microscopic theory for the optical properties of coulomb- correlated semiconductors. Phys Status Solidi B 206:477

16. Schmielau T, Pereira MF Jr (2009) Nonequilibrium many body theory for quantum transport in terahertz quantum cascade lasers. Appl Phys Lett 95:231111

17. Pereira MF Jr, Henneberger K (1997) Gain mechanisms and lasing in II-VI compounds. Phys Status Solidi B 202:751

18. Pereira MF Jr, Binder R, Koch SW (1994) Theory of nonlinear absorption in coupled band quantum wells with many-body effects. Appl Phys Lett 64:279

19. Chow WW, Pereira MF Jr, Koch SW (1992) Many-body treatment on the modulation response in a strained quantum well semiconductor laser medium. Appl Phys Lett 61:758

16 Intersubband Dispersive Gain Media


Abstract Quantum cascade lasers (QCLs) are important practical sources for infrared spectroscopy. Recent research in mid-IR QCLs has resulted in record high wallplug effi ciency, output power, single mode operation and wide tunability. The spectroscopic range of QCL has been extended to as short as 3 μm, opening up possibility of detection of a wide range of molecules. Wide tuning based on dual section sampled grating distributed feedback QCLs has resulted in individual tuning of 50 cm −1 and 24 dB side mode suppression ratio with continuous wave operation greater than 100 mW. Need for compact, room temperature sources, in spectral range constituting the THz gap, has led to QCL based THz sources by difference frequency generation with modal or Cherenkov phase matching schemes. Single mode emission from 1 to 4.6 THz with side mode suppression ratio up to 40 dB and THz power of 65 µW at 4 THz are demonstrated.

Keywords Springerlink • TERA-MIR 2012 • Quantum cascade laser • Mid- infrared • Terahertz • Room temperature • Continuous wave • Single mode • Tunability

17.1 Introduction

Compact semiconductor source and detector of THz (0.3–10 THz) and mid-infrared (mid-IR) (30–100 THz) radiation are in high demand for applications in environment monitoring, pollution control, trace gas sensing and stand-off explosive detection. This is because many chemicals have their spectroscopic fi ngerprints in these regions. Though

Chapter 17 Recent Advances in IR Laser Diodes with High Power, High WPE, Single Mode, CW Operation at RT

Manijeh Razeghi , Neelanjan Bandyopadhyay , Quanyong Lu , Yanbo Bai , Steven Slivken , and David Heydari

M. Razeghi (*) • N. Bandyopadhyay • Q. Lu • Y. Bai • S. Slivken • D. Heydari Center for Quantum Devices, Department of Electrical Engineering and Computer Science , Northwestern University , Illinois 60208 , USA e-mail: [email protected]

124

some of these sources are now commercially available there is still a lot of room for improvement in terms of high power continuous wave (CW), single mode, widely tunable operation at room temperature (RT) with high reliability. The THz radiation has huge potential in medical diagnostics due to its non-ionizing character in contrast to x-ray.

17.2 Recent Advances in Mid-IR QCL

QCLs operating in upper part (4.5–5 μm) of mid infrared atmospheric window (3–5 μm) with record performance have been developed in recent years [ 1 ]. Wall plug effi ciency (WPE), as high as 27 % (21 %) in pulsed (continuous) mode of operation at RT [ 2 ] and 53 % at cryogenic temperature [ 3 ], has been demonstrated. The maximum power in CW operation at RT is 5.1 W [ 2 ]. Optimization of strained and lattice matched material growth with sharper interfaces, improved active region design, better device processing followed by device packaging and thermal management has contributed to a signifi cant improvement of the QCL performance. Figure 17.1 shows current, power, effi ciency records for a mid-IR QCL at RT.

Many QCL based applications, including chemical imaging and remote or photo acoustic chemical sensing prefer single mode spectrum, stable beam quality, which can be accomplished by incorporating a distributed feedback (DFB) grating into a standard QCL waveguide. 2.4 W CW power was obtained at RT from a narrow ridge DFB [ 4 ]. Angled cavity DFB with grating lines parallel to laser facets has also been demonstrated to achieve single mode diffraction limited emission [ 5 ]. A method to improve both beam quality and spectral purity of broad area QCL is the

Fig. 17.1 RT operation of a 4.9 μm QCL in pulsed ( dotted curve ) and continuous wave ( solid curve ) operation. Maximum WPE in pulsed (CW) operation is 27 % (21 %). Maximum output power in CW is 5.1 W

M. Razeghi et al.

125

implementation of two dimensional photonic crystal features along the cavity. A high peak power of 34 W has been demonstrated with PCDFB [ 6 ]. To decrease diver-gence of the far fi eld, which is inversely proportional to the size of the emitting aperture in the near fi eld, a surface emitting ring laser structure with higher aper-ture size is used. With an epi-down bonding scheme, a substrate emitting quantum cascade ring laser with RT CW operation of 0.51 W [ 7 ] was obtained. Incorporation of DFB structure, either as one-dimensional lines or two dimensional arrays or a ring shaped pattern provides suffi cient feedbacks.

Even with narrow linewidth and a high side mode suppression ratio (SMSR), wide tuning is required for spectroscopic applications. Unfortunately a traditional DFB only gives a tuning range of 5 cm −1 which may limit its usefulness in scanning for spectroscopic tuning. Monolithically integrated multiple DFB QCLs with slightly shifted lasing wavelength to form an array of DFB QCL may be used as widely tunable source. CW operation ensures that in addition to high average power there is no need for complex driving circuit for pulse operation. For an array of 20 QCLs a total tuning of 100 cm −1 was obtained around 4.5–4.7 um [ 8 ]. Another way of increasing the wavelength range is the use of dual section sampled grating distributed feedback QCLs. It resulted in a individual laser tuning of 50 cm −1 and SMSR of 24 dB with continuous wave operation with greater than 100 mW optical power [ 9 ]. Additionally, a broad spectral coverage of 350 cm −1 on a single chip was demonstrated, which is

Fig. 17.2 Single-mode emission spectra for an electrically-tuned, sampled grating laser. Upper fi gures show both the side modes suppression and continuous wave output power as a function of emission wavelength

17 Recent Advances in IR Laser Diodes with High Power, High WPE…

126

equivalent to 87.5 % of the gain band width [ 10 ] of that QCL. Single-mode emission spectra, for an electrically-tuned, sampled grating laser, are shown in Fig. 17.2 .

Many hydrocarbons have their fundamental vibrational modes less than 3.5 μm. So detection of the molecules requires effi cient QCL sources working in these wave-lengths. However, the major challenges in developing these QCLs working in CW at RT are indirect valley transfer, thermal leakage to the continuum and growth of high quality material at high strain level. A QCL emitting ~500 mW CW power at RT has been obtained [ 11 ]. A QCL operating in RT at a wavelength of 3 μm, which is the shortest wavelength of QCL operating in CW at RT, has also been demonstrated [ 12 ].

17.3 THz Sources Based on Intracavity Generation in Mid- infrared Laser

On the longer wavelength side of electromagnetic spectrum, the terahertz spectral range is important for many applications in imaging, spectroscopy and biological engineering. A compact, room temperature source will make the

a b

c d

Fig. 17.3 ( a ) P-I-V characterization for the two wavelengths. ( b ) THz power and mid-IR power product as functions of current. Inset: measured room temperature THz spectra at 4.0 THz. ( c ) The mid-IR spectra of different DFB designs with varied frequency differences between λ 1 and λ 2 and the EL spectrum. ( d ) The THz spectra of the devices based on Cerenkov phase matching scheme emitting from 1.0 to 4.6 THz

M. Razeghi et al.

127

implementation of these applications feasible possible. A possible solution is a THz source based on intra-cavity frequency difference generation along with an integrated dual-period distributed feedback grating in waveguide to select a single mode [ 13 ]. Maximum power at 3.3–4.6 THz in single mode operation by using dual-period grating and modal phase matching scheme is 65 μW [ 14 ]. Another approach is to use Cerenkov phase matching scheme along with dual-period distributed feedback grating. Wide tuning of emissions from 1 to 4.6 THz with side mode suppression ration of 40 dB and power up to 32 μW is obtained [ 15 ]. Figure 17.3 shows the THz output power and mid-IR power product as a function of current along with tuning of mi-IR and THz emissions from 1 to 4.6 THz based on Cerenkov phase matching scheme.

17.4 Conclusion

In summary, InP based quantum cascade laser technology has demonstrated its versatility as a source across wide spectrum from mid-infrared to THz. Thus this technology is currently observing a rapid progress to achieve even more effi cient compact, room temperature, spectrally pure sources for different applications.

References

1. Razeghi M (2009) High-performance InP-based Mid-IR quantum cascade lasers. IEEE J Sel Top Quantum Electron 15:941

2. Bai Y, Bandyopadhyay N, Tsao S, Slivken S, Razeghi M (2011) Room temperature quantum cascade lasers with 27 % wall plug effi ciency. Appl Phys Lett 98:181102

3. Bai Y, Livken S, Kuboya S, Darvish SR, Razeghi M (2010) Quantum cascade lasers that emit more light than heat. Nat Photonics 4:99

4. Lu QY, Bai Y, Bandyopadhyay N, Slivken S, Razeghi M (2011) 2.4 W room temperature continuous wave operation of distributed feedback quantum cascade lasers. Appl Phys Lett 98:181106

5. Bai Y, Slivken S, Lu QY, Bandyopadhyay N, Razeghi M (2012) Angled cavity broad area quantum cascade lasers. Appl Phys Lett 101:081106

6. Goekden B, Bai Y, Bandyopadhyay N, Slivken S, Razeghi M (2010) Broad area photonic crystal distributed feedback quantum cascade lasers emitting 34 W at λ – 4.36 μm. Appl Phys Lett 97:131112

7. Bai Y, Tsao S, Bandyopadhyay N, Slivken S, Lu QY, Caffey D, Pushkarsla M, Day T, Razeghi M (2011) High power, continuous wave, quantum cascade ring laser. Appl Phys Lett 99:261104

8. Razeghi M, Gokden B, Tsao S, Haddadi A, Bandyopadhyay N, Slivken S (2011) Widely tunable single-mode high power quantum cascade lasers. Proc SPIE 8069:806905

9. Slivken S, Bandyopadhyay N, Tsao S, Nida S, Bai Y, Lu QY, Razeghi M (2012) Sampled grating, distributed feedback quantum cascade lasers with broad tunability and continuous operation at room temperature. Appl Phys Lett 100:261112

10. Slivken S, Bandyopadhyay N, Tsao S, Nida S, Bai Y, Lu QY, Razeghi M (2013) Dual section quantum cascade lasers with wide electrical tuning. Proc SPIE 8631:86310P

17 Recent Advances in IR Laser Diodes with High Power, High WPE…

128

11. Bandyopadhyay N, Slivken S, Bai Y, Razeghi M (2012) High power continuous wave, room temperature operation of λ 3.4 μm and λ 3.55 μm InP-based quantum cascade lasers. Appl Phys Lett 100:212104

12. Bandyopadhyay N, Bai Y, Tsao S, Nida S, Slivken S, Razeghi M (2012) Room temperature continuous wave operation of λ ~ 3–3.2 μm quantum cascade lasers. Appl Phys Lett 101:241110

13. Lu QY, Bandyopadhyay N, Slivken S, Bai Y, Razeghi M (2011) Room temperature single-mode terahertz sources based on intracavity difference-frequency generation in quantum cascade lasers. Appl Phys Lett 99:131106

14. Lu QY, Bandyopadhyay N, Slivken S, Bai Y, Razeghi M (2013) High performance terahertz quantum cascade laser sources based on intracavity difference frequency generation. Opt Express 21:968

15. Lu QY, Bandyopadhyay N, Slivken S, Bai Y, Razeghi M (2012) Widely tuned room tempera-ture terahertz quantum cascade laser sources based on difference frequency generation. Appl Phys Lett 101:251121

M. Razeghi et al.


Abstract We characterize, by terahertz time-domain spectroscopy, the far infrared properties of different selenide, sulfi de and telluride compounds and crystals. We propose a new method to extract, from the experimental data, the refractive index and absorption of such materials that exhibit strong absorption lines. We illustrate this method by measuring the terahertz response of doped TlInS 2 crystals. We report the observation of a strong absorption peak at 1.242 THz in InSe, which could be attributed to the excitation of TO phonon (half layer shear mode).

Keywords Terahertz • Time-domain spectroscopy • Selenide • Sulfi de • Telluride

18.1 Introduction

Telluride and selenide materials exhibit amazing properties in view of various promising applications like storage of electricity, nonlinear optics, etc. Some of them show specifi c spectral features in the terahertz (THz) range, while some others can be used to generate THz waves by optical rectifi cation. Therefore, a precise characterization of the THz response of such materials is necessary for developing optimized THz devices, and also to have a better knowledge of the physical phenomena occurring at THz frequencies in the materials. Thanks to its large

Chapter 18 Characterization of Selenide, Sulfi de and Telluride Materials by Terahertz Time- Domain Spectroscopy

R. M. Sardarly , F. Garet , M. Bernier , and J.-L. Coutaz

R. M. Sardarly (*) Institute of Radiation Problems , National Academy of Sciences of Azerbaijan , B. Vahabzade 9 , Baku AZ1143 , Azerbaijan e-mail: [email protected]

F. Garet • M. Bernier • J.-L. Coutaz IMEP-LAHC, UMR CNRS 5130 , Université of Savoie , 73376 Le Bourget du Lac , France e-mail: [email protected]

130

achievable bandwidth and its great dynamics, THz time-domain spectroscopy (TDS) is a valuable tool to perform this characterization. Nevertheless, we are obliged to modify the classical code of extraction of parameters (refractive index n and coeffi cient of absorption α of the materials) as these materials present spectral bands of high absorption, in which almost no THz signal is transmitted through the samples. We apply this modifi ed procedure to study low dimensional A III B III C 2 VI compounds [ 1 ], like TlInS 2 , and crystals, like GaSe and InSe.

18.2 Principle of the Extraction THz-TDS Technique

In classical THz-TDS, one records the temporal waveforms impinging onto and transmitted by a sample. Then a numerical fast Fourier transform (FFT) of both signals is performed. The ratio of the transmitted and incident FFT spectra gives the transmission coeffi cient of the sample. As the origin of time is preserved between the two requested measurements, both modulus and relative phase of the transmis-sion coeffi cient are obtained. If the sample is a slab with parallel sides, the index of refraction and the coeffi cient of absorption could be accurately determined using inverse electromagnetic code [ 2 ]. This code, as well as derived ones, requires input-ting the absolute transmission phase, which is obtained by a linearization of the measured phase at low frequencies. In materials exhibiting a high absorption band, the transmitted signal within this band could be weaker than noise. In this case, the transmission coeffi cient is almost zero in modulus, and its phase is unknown, which makes extraction codes inoperative. The usual solution to this problem is to perform THz-TDS in refl ection. However, the disadvantage of the refl ection technique is its weak precision due to diffi culty to get a good reference signal. This latter is supplied by a metallic mirror located at the position of the sample, whose coeffi cient of refl ection is supposed to be 100 %. A small error of the sample position, for example a few μm shift from the position of the mirror, leads to dramatic errors, mostly regarding the absorption of the sample.

Thus we developed a combined technique [ 3 ], which takes benefi t of both transmission and refl ection THz-TDS’s. The basic idea is to roughly estimate n from refl ection data over the whole THz spectrum. In the fi rst spectral region of transparency, n is more precisely obtained from transmission data. Then we correct the effect of any error in position of the refl ection experiment by equalizing the n values obtained in refl ection and in transmission. In the second band of trans-parency, and similarly in the other transparency regions, the missing phase in transmission is retrieved by comparing the n values extracted from corrected-refl ection and transmission. Thus n is nicely determined from transmission data in the regions of transparency, while we save the refl ection values in the regions of absorption. As well, absorption id determined from transmission in the transparency bands, while we keep the corrected refl ection values in the absorption peaks. However, in the present study, materials are not transparent at lower frequencies and the extraction method [ 3 ] must be modifi ed. The refractive index is estimated

R.M. Sardarly et al.

131

over the whole spectrum by performing a Kramers-Kronig transformation of the absorption obtained in transmission. The error due to the saturated transmission in the absorption bands or to missing absorption peaks located outside the experimental spectral window, is spread over the whole spectrum. Then the phase correction in refl ection and the phase retrieval in transmission are performed as explained previ-ously. Of course, the method can be applied only if the error in the Kramers-Kronig refractive index, due to the missing absorption data, is smaller than the error induced by a 2 π phase error.


We applied our method to determine n and α of low dimensional A III B III C 2 VI compounds, like doped TlInS 2 , TlGaTe 2 , etc. These materials grow as fi lms of nanofi bers. The fi bers exhibit a hexagonal structure made of layers of A and C atoms, which form A III,3 + C 2 VI,2 − nanochains, separated by a layer of B atoms. Electrical hoping-type conductivity occurs in the metallic B layer, and thus this conductivity shows a strong anisotropy. Below a threshold temperature, the experimental DC-conductivity of the samples varies linearly as 1/ T , which is typical of ionic-type conductivity [ 1 ]. Above this temperature, the conductivity increases dramatically: this phenomenon is called super-ionic conductivity. In the case for example of TlGaTe 2 , this sharp increase of conductivity happens due to the phase transition accompanying the disordering of the Tl + sublattice. Figure 18.1 shows the coeffi cient of transmission of TlInS 2 :Co versus frequency. Several absorption peaks in the spectrum could be attributed to the excitation of phonons (A 2u and B u ). The absorption

Fig. 18.1 Transmission coeffi cient of a TlGaTe 2 crystal sample versus frequency for 2 THz E-fi eld polarizations

18 Characterization of Selenide, Sulfi de and Telluride Materials…

132

line at approximately 0.2 THz, seen only when the THz fi eld is aligned along the nanofi bers, occurs at a frequency lower than the lowest phonon peak (A 2u ). Therefore, this resonance is probably related to the libration oscillation of the Ga 3+ Te 2− nanofi bers [ 4 ].

Figures 18.2 and 18.3 present respectively the n and α spectra of TlInS 2 :Co. In the transparency regions, n is precisely obtained from transmission TDS while,

Fig. 18.2 Refractive index of TlInS 2 :Co versus frequency obtained with the different THz-TDS methods explained in the text

Fig. 18.3 Absorption coeffi cient of TlInS 2 :Co versus frequency obtained with the different THz- TDS methods explained in the text

R.M. Sardarly et al.

133

in the absorption peaks, its value deduced from refl ection is noisier. Absorption obtained from transmission TDS exhibits a strong saturation at the resonances, while values derived from refl ection data are affected by a strong error due to a mispositionning of the reference mirror, which is corrected using our method (thick curve).

n and α spectra of InSe and GaSe are depicted in Fig. 18.4 . The rigid layer phonon mode E’ (2) of GaSe at 0.596 THz is clearly observed. In InSe, we can see a similar peak at 1.242 THz, which can be attributed to a TO half-layer shear phonon-mode, as already reported from Raman spectroscopy [ 5 ]. To the best of our knowledge, our observation is the fi rst one in THz spectroscopy, as published far infrared studies of InSe [ 6 , 7 ] were performed with FTIR instruments that do not reach the 1-THz range.

References

1. Panich AM, Sardarly RM (2010) Physical properties of the low dimensional A III B III C 2 VI compounds. Nova Science Publishers, Inc., New York

2. Duvillaret L, Garet F, Coutaz J-L (1996) IEEE J Sel Top Quantum Electron 2:739 3. Bernier M, Garet F, Coutaz J-L (2013) IEEE THz Sci Technol 3:295 4. Sardarly R, Samedov O, Abdullayev A, Salmanov F, Urbanovic A, Garet F, Coutaz J-L (2011)

Jap J Appl Phys 50:05FC09 5. Imai K, Kumazaki K, Seto S, Yamaya K, Abe Y (1990) Proc Indian Acad Sci (Chem Sci)

102:601 6. Julien C, Edrieff M, Balkanski M, Chevy A (1992) Phys Rev B46:2435 7. Benramdane N, Bouzidi A, Tabet-derraz H, Kebbab Z, Latreche M (2000) Microelectron

Eng 51–52:645

Fig. 18.4 n and α versus frequency of InSe and GaSe

18 Characterization of Selenide, Sulfi de and Telluride Materials…


Abstract In this work we present the theoretical model of existence of ultrashort modes in low frequency region of the electromagnetic spectrum in a system of nanocilinders array. These modes have no analogue in a spectrum with only one nanocylinder. For nanodot deposited or filled-in pores produced from semiconductor or conductive polymers, monomers, composites etc., the SPR may be found in MIR range or in a range of frequencies with much higher wavelengths, like THz even GHz.

Chapter 19Ultrashort Electromagnetic Modes in the Low Frequency Region of the Spectrum in a Nanocylinder Array

L. Sirbu, V. Sergentu, R. Muller, V. Ursaki, and I.M. Tiginyanu

L. Sirbu (*) D. Ghitu Institute of Electronic Engineering and Nanotechnologies, 3/3, Academiei str., Chisinau, Moldovae-mail: [email protected]

V. Sergentu Institute of Applied Physics, Academy of Science of Moldova, 5, Academiei str., Chisinau, Moldovae-mail: [email protected]

R. Muller National Institute for R&D in Microtechnologies, IMT Bucharest, Erou Iancu Nicolae 126 A Str., Bucharest, Romaniae-mail: [email protected]

V. Ursaki Institute of Applied Physics, Academy of Science of Moldova, 5, Academiei str., Chisinau, Moldova

Moldova Technical University of Moldova, 168, Stefan cel Mare str., Chisinau, Moldovae-mail: [email protected]

I.M. Tiginyanu D. Ghitu Institute of Electronic Engineering and Nanotechnologies, 3/3, Academiei str., Chisinau, Moldova

Moldova Technical University of Moldova, 168, Stefan cel Mare str., Chisinau, Moldova

136

Keywords Nanocylinder • Nanotube • Ultrashort • Electromagnetic • Array • Dispersion • Maxwell

19.1 Introduction

Electrochemically formed porous materials usually show a narrow pore size distribution and a certain pore density, which allows us to determine the ratio of pore volume to the total volume, the porosity. In most cases, the pore distribution at the electrode surface is random; however, in certain cases, like for example, for porous InP (Fig. 19.1), a short-range order may be observed.

The transmission of radiation through subwavelength hole (nanotubes, nanocylinders) arrays made in a semiconductor film apart from its fundamental interest has potential applications in near field microscopy, wavelength-tunable filters, optical modulators, bio-sensors, and THz emitters and the detection of CBRN agents and explosives.

The downsizing of photonic devices towards nanoscale dimensions meets a major difficulty, since the fundamental limit of diffraction restricts the minimum lateral dimensions of dielectric optical elements and waveguides to about half the effective light wavelength. The optics based on surface plasmons (SP) allows us to overcome this limitation. Because of the small wavelength of the plasmon wave, a much higher spatial resolution can be obtained, which can provide a new nano- fabrication or nano-storage approach by using optical light with a long wave-length. The SPs which come from metallical dots are in range of visible light or near IR. But using dots or composite material made from semiconductor or conductive polymers, monomers etc., we can have an SPR in MIR or in range of frequencies with much higher wave lengths such THz even GHz.

Fig. 19.1 Porous InP impregnated by coordinative metaloorganic compounds

L. Sirbu et al.

137

19.2 Problem Statement

The fabrication of nanoporous templates within technological processes allow to obtain various kinds of cylinders arrays. Between them one can distinguish metalized and simple pores, nanocylinders, and nonotubes, respectively. Our purpose is to deduce the convenient analytical formulas and their solution in order to receive and optimize the dispersion law for a nanocylinder array system built in a periodic two-dimensional lattice. For a simple representation we will treat a case when the electric field intensity component

E is along the axis of multilayered nanocylinder and it is not equal to zero, and the intensity of the magnetic field

H is perpendicular to the axes, and radiations frequency ω (the wavelength λ in vacuum) is small (is great). The calculation was carried for an example of filling of nanopores. The basic criterion which will allow to receive further simple enough formulas, is

l w p>> ~ 2 /a c a( )

(19.1)

where a is the distance between two pores.

19.3 The Dispersion Equations for the Elementary Kind of Nanotubes

At first we will consider the propagation process of electromagnetic waves in nanocylinder array system, built in an infinite lattice (we will consider only square one). Let’s consider following ideas: the nanotubes are physical objects with well defined radius and extremely small wall thickness (carbon nanotubes are a physical embodiment of similar objects [1]). From all possible physical parameters, the nanotubes have only the radius R and the superficial complex conductivity σzz.

The electric current induced in nanotubes can be directed only along the axis of tubes. The description of the field and the material components (currents and charges) that are influencing on the field will be carried out on the basis Maxwell’s standard equations system [2]. The solution of the above equations is searched in a form F i hz t

ρ ω( ) −( )( )exp , where F

ρ( ) is a function describing distribution along the plane (X, Y) for any of values of (

E Hρ ρ( ) ( ), etc.), h is a mode wave vector along the axis of tubes (Z axis),

ρ = ( )x,y .The dispersion equation obtained coincides with formula 58 from ref. [3]. The

dispersion equations for a nanotubes system obtained by means of dual Green functions (the formula (11) from Ref. [4]). The following analytical expression is obtained for the dispersion relations taking into account only the motion along the nanotube axis

19 Ultrashort Electromagnetic Modes in the Low Frequency Region…

138

211

2

1 02

1

20 0

πε σw

kG

aw

i

z k Rzz

−( )

= − ,

(19.2)

where w k h1 1 02 2= −ε is the value of the wavevector for the two-dimensional

motion of nanotube perpendicularly to the nanotube axis, k0 = ω/c is the wavevector in the vacuum, ε1 is the relative dielectric permittivity of the medium in which the tubes are placed, z0 0 0= µ ε/ =376.6 Ω is the vacuum impedance, R is the nano-pore radius. Note that in fact G = G(R/a,w1). We obtain a numerical estimation G ~ 2.2∙10−2, for R = 0.35a and |w1| << π/a.

19.4 The Dispersion Equations for an Arbitrary Type of Nanotubes (Nanopores)

A method for deducing the dispersion equation for periodic systems consisting from multilayered cylinders is described in [5]. As a result, we obtain a set of functions {Dm} completely describing the dispersion process of the external electromagnetic wave on a single nanotube for any order m of cylindrical functions which can be used to deduce the dispersion law of a single nanotube (nanopore). Further, these functions are used in the equation considering the interaction between various cylinders (formula (11) ref. [5]). We will use a similar procedure, but for a long- wavelength range. In our case, in contrast to [5], it is necessary to take into account also the movement along the axis of the nanotube (nanopore). However, this circumstance does not influence essentially the deduced equation if we consider only the zero order of cylindrical functions m = 0, and introduce the wave vector h for the movement along the nanotube. For our case, the solution is searched in form of

E rz

( ) = ( ) + ( ) ( )C I Diw K iw ihz0 0 1 0 0 1ρ ρ exp .

(19.3)

Let’s consider the first stage of the deduction of the dispersion law for a single nanotube (nanopore). The dispersion law for the movement along the axis of a single nanotube (nanopore) is defined by one equation taking in account only a zero order of cylindrical functions

1 00/ .D = (19.4)

For the above described elementary type of nanotubes we receive

D ik z w

d

dRK iw R I iw R

ik z K iw Rzz

zz

0 0 0

12

12 0 1 0 1

0 0 0 1

0

= −( ) ( )( )

+σ

ε

σ/

/k

(( ) ( )( )

/

.

I iw R0 1

(19.5)

L. Sirbu et al.

139

By substituting (19.5) in (19.4) and by using formulas from [6] it is possible to obtain equation (58) form ref. [3]. We will not give the result for the value of D for conventional nanopores or nanotubes. Let’s consider the second stage of the deduction of the dispersion law, but for a multiwall nanotube system. Basically it would be possible to use the results presented in [7]. However, the analytical formula 11 in ref. [7] are rather complex and inconvenient in calculations. Therefore, we will use the formula (19.2). The complex conductivity σzz is replaced by another value σzz

eff received from the following relation

ik zw

d

dRI iw R

w

d

dRI iwzz

eff0 0

12

12 0 1

22

22 0 2

0 0σε ε

= ( )( ) −k k

ln ln RR ik z zz( )( ) + 0 0σ

(19.6)

were w k h1 2 1 2 02 2

( ) ( )= -ε represents the wave vector outside (inside) of nanopores,

ε1(2) is the dielectric constant outside (inside) of nanopores. Equation (19.6) is derived in a way assuring the equality of the scattering efficiency D0

eff of a virtual simplest type nanotube with an effective complex conductivity σzz

eff to the scattering efficiency D0 of a real nanopores.

19.5 Results

Let’s consider the results for dispersion equations obtained with this approach. Our results obtained for a common single nanopore or nanotube coincide with the well- known one [8], which means the absence of searched modes in common widely used materials (ε1, ε2 > 0). Substantially different results are obtained for a nanopore (nanocylinder) system. Several fundamentally different types of modes are obtained in an array system. Most of them don’t have any analogues in a single nanopore or nanotube. The graphs of the dispersion laws in Fig. 19.2 clearly show the existence of different branches for ε1 = 6, ε2 = 1. New branches of oscillations with ω ~ 2πc/a emerge in addition to the usual long-wavelength oscillations with a speed of propa-gation v ~ c. The electromagnetic modes can exhibit fundamentally new optical properties in the neighborhood of these oscillations, for instance their velocity of propagation is |v| << c within a narrow frequency range Δω ~ 2πc/a, while the value of the effective refractive index can be very low or even negative. Apart from that, a new type of undamped pure electric ultrashort modes emerge with h ~ 2π/a, which can exist at low frequencies ω << 2πc/a (19.1) for ε1 > ε2. The group velocity of such waves is v > c, which means that these modes can’t transport electromagnetic energy. The existence of ultrashort modes indicates a significant restructuring of the part responsible for the interaction between the charges inside of the nanopores system, which is usually associated with the longitudinal electric fields. The introduction of a metal into the system by choosing the plasma frequency allows one to control the properties of the above mentioned electromagnetic modes. For instance, by choosing the plasma frequency ωo << 2πc/a we can totally remodel the properties


140

of utrashort modes. The Fig. 19.2 for the metalized nanopores (metal thickness d = 0.1 ∙ R) suggests that this mode is converted into a plasma mode, and therefore can be used for the transport of electromagnetic energy. The group velocity of such waves is v << c. However, this relation is valid only in a narrow frequency range.

19.6 Conclusion

In summary, the results of this work demonstrate the existence of ultrashort modes at low frequencies in porous systems. These modes can be used for the control of optical properties of materials (nonlinear etc.) and for the creation of optical devices (MOEMS, lab-on-chip, biosensors, etc.).

Acknowledgments This work was supported by Romanian(IMT&INFLPR)-Moldavian Bilateral project, young scientist project 12.819.15.20A, partially supported by the FP7 project MOLD- ERA (Grant no 266515), and NATO Advanced research Workshop on Detection of Explosives and CBRN (Using Terahertz).

References

1. Ilyinsky S, Slepyan GY, Slepyan AY (1993) Propagation, scattering and dissipation of electro-magnetic waves. Peter Peregrinus, London

2. Jackson D (ed) (1962) Classical electrodynamics. Wiley, New York3. Slepyan GY, Maksimenko SA, Lakhtakia A, Yevtushenko O, Gusakov AV (1999)

Electrodynamics of carbon nanotubes: dynamic conductivity, impedance boundary conditions, and surface wave propagation. Phys Rev B 60:17136–17149. doi:10.1103/PhysRevB.60.17136

4. Nefedov IS (2010) Electromagnetic waves propagating in a periodic array of parallel metallic carbon nanotubes. Phys Rev B 82:155423–155430. doi:10.1103/PhysRevB.82.155423

Fig. 19.2 Dispersion laws for nanopore, and for metallic coated nanopore

L. Sirbu et al.

http://dx.doi.org/10.1103/PhysRevB.60.17136


141

5. Zhang X, Li L-M, Zhang Z-Q, Chan CT (2001) Surface states in two-dimensional metallodielectric photonic crystals studied by a multiple-scattering method. Phys Rev B 63:125114–125121. doi:10.1103/PhysRevB.63.125114

6. Katsenelenbaum BZ (1966) High-frequency electrodynamics. Nauka, Moscow7. Nefedov I, Tretyakov S (2011) Effective medium model for a periodic array of metallic carbon

nanotubes and eigenwaves propagating. In: A finite-thickness carbon nanotube slab. Proceedings of the international conference NANOMEETING-2011. World Scientific Publishing Co. Pte. Ltd, Minsk, pp 267–269

8. Weinstein LA (1988) Electromagnetic waves, 2nd edn. Radio & Sviaz, Moscow




Abstract The paper deals with the AlGaAs/GaAs mid-infrared (MIR) quantum- cascade laser (QCL) double-trench mesa wet-etching fabrication conditions, experimentally tested and selected in order to ensure the smooth-sidewall mesa waveguide. The other important properties of etching solutions, like etching unifor-mity in plane of the wafer, process velocity and cost of the process were analyzed, as well. The huge importance of nearly ideal sidewall smoothness of mesa for high performance of the laser, in terms of its low threshold current density and its high slope efficiency, is emphasized. The impact of the unintentional mesa-sidewall roughness on QCLs’ performance is quantitatively studied – it was investigated how the mesa- sidewall roughness standard deviation value increasing in the range of σrough from 0 μm up to 1.42 μm increases the scattering losses, and thus, negatively affects laser’s parameters; it was found that the excess roughness of mesa-sidewall, i.e., the σrough reaching ≥1.8 μm, was eventually a reason of a total deterioration of laser’s performance, i.e., it caused a lack of lasing.

Keywords AlGaAs/GaAs QCLs • Wet etching • Mesa fabrication • Threshold current • Laser’s efficiency • Scattering losses

20.1 Introduction

There are known several processes, like carrier losses, current spreading, wave-guide internal and scattering losses, mirror losses and problems with heat dissipa-tion, that can negatively affect the semiconductor laser performance in terms of its threshold current density (Jth), slope efficiency (ηS) or T0 parameter. The solutions

Chapter 20Influence of Mesa-Fabrication-Dependent Waveguide-Sidewall Roughness on Threshold Current and Slope Efficiency of AlGaAs/GaAs Mid-Infrared Quantum-Cascade Lasers

Anna Szerling, Kamil Kosiel, Piotr Karbownik, Anna Wójcik-Jedlińska, and Mariusz Płuska

A. Szerling (*) • K. Kosiel • P. Karbownik • A. Wójcik-Jedlińska • M. Płuska Institute of Electron Technology, Al. Lotników 32/46, 02-668 Warsaw, Polande-mail: [email protected]

144

against the mentioned above parasitic processes may be found generally in the fields of device construction or technology. In this paper we focus on problems connected with scattering losses caused by unintentional roughness of mesa-waveguide sidewalls, that may be found in AlGaAs/GaAs MIR QCLs. The smooth enough morphology of mesa sidewalls is necessary for lasing, as the rough walls of mesas are the reason of scattering losses, and hence they cause strong deterio-ration of laser’s performance; scattering losses depend quadratically on mesa-sidewall roughness (Eq. 20.1), and because of that the increasing roughness is a reason of fast increase of threshold current density (Eq. 20.2) and fast decrease of slope efficiency of the laser (Eq. 20.3).

α σ λπσ

λθθscatt rough

rough tp

,( ) =

+

4 1

2

12 3

1

cos

sin

+

−1

2

1

p

(20.1)

J

g gth roughm w scatt roughσ

α α α σ( ) = +

+( )

Γ Γ (20.2)

η σ σλη

αα α α σS rough

outrough i

m

m w scatt rough

dP

dI e

hc( ) = ( ) =+ + ( )

1 Γ

(20.3)

where: αscatt is the scattering losses in the waveguide, σrough is the roughness standard deviation of waveguide (mesa) sidewall, λ is the wavelength in the waveguide, θ is the angle of ray propagation in the waveguide, t is the waveguide thickness, 1/p1,2 are the penetration depths of modes into the cladding layers, αw is the intrinsic wave-guide losses (mostly due to absorption phenomena), αm is the mirror losses of the laser cavity, g is the gain coefficient of the active core, Γ is the confinement factor of the mode for the active part of the waveguide, ηi is the laser’s internal efficiency, h is the Planck constant, c is the speed of light in vacuum, and e is the elementary charge [1, 2].

Similar mesa-roughness problems have been analyzed experimentally by Toor et al. in [1, 2] – however, the specially designed masks were prepared in those works in order to be used for the strictly managed experiment, in which the controlled roughness was introduced on purpose to the fabricated laser structures. On the con-trary, here we analyse the lasers we fabricated during the development of our QCL technology, i.e., when several mesa-etching techniques and etching conditions were tested by us in order to obtain the maximum improvement of the mesa sidewall morphology. The experiment described here relates to development of fabrication technology of AlGaAs/GaAs (λ ~ 9.5-μm) QCLs, which description – together with details on the resulting final devices’ – is presented elsewhere [3, 4]. The elaboration of perfectly smooth mesa walls turned out to be one of the sine qua non conditions of high performance lasers, that finally were successfully applied in trace-gas detection system [5].

A. Szerling et al.

145

20.2 Selection of Conditions of Mesa Fabrication Towards Smooth Mesa-Wall Morphologies

The shape of mesa, its size and depth, strongly influence the electric-current flow pattern, and hence these factors influence the heat dissipation pattern, as well. It was theoretically found, for example, that rectangular mesa results in current accumula-tion points at the edges of mesa [6], so in order to avoid this disadvantageous effect the circular-shaped mesas must be fabricated. What is more, a theoretical research approach shows the importance of mesa depth and that by deep enough etching of mesa – and that means etching exactly under the laser’s active region – the laser threshold current may be reduced by several times [7]. For this reason our QCL mesas were ~6 μm-deep; they were etched through the topmost micrometer-sized GaAs waveguide layer, and through the whole AlxGa1-xAs/GaAs (with x = 45 %) multilayer active stack of nanometer-sized layers. Because of heat management requirements, however, the excessively deep mesa etching was avoided.

The standard technology of mesa formation was performed, i.e., photolithography and wet etching of mesa. This type of process, in contrast to dry etching method, is less expensive and generally doesn’t need application of hard masks – which usage is associated with additional technological operations. The risk of damage of the laser structure, during the wet-fabrication of mesas, was also found relatively low. A particular attention was paid in order to achieve low roughness of mesa sidewalls, that was necessary to minimize scattering losses of the laser. In order to optimize the etching process, a wide range of compositions of etching solutions, as well as their different temperatures and stirring rates were tested. Several particularly interesting examples of etchants are presented in Table 20.1. The etchants’ compositions are given in terms of volumetric proportions. Ammonium hydroxide, acids, and hydrogen peroxide, as the substrates for preparation of etching solutions – these were taken in their commercially available “concentrated” forms.

A criterion of selection of an appropriate solution that should be taken for mesa etching technology is not only the resulting mesa sidewall (smooth) morphology, but also the etching-process uniformity in plane of the wafer, etching-process rate, time consumption of the overall technology and the total number of operations involved, and cost of the process.

The investigated etching aqueous solutions, are based on bases or acids and contain hydrogen peroxide; deionized water is used as a solvent. Some solutions, like for example these ones based on sulphuric acid, produce rough mesa sidewalls as well as poor etching uniformity in plane of the wafer and hence their further use for technology was rejected. Most of solutions presented in Table 20.1, however, like the etchants based on ammonium hydroxide and solutions based on hydrochloric acid, provide smooth sidewalls of mesas. For many of them, however, the problem was not sufficient spatial uniformity of etching, that was a reason of the resulting not uniform etching in plane of the wafer. On the other hand, the tested phosphoric acid-based etchant provides too slow etch rate to be practically useful.

20 Influence of Mesa-Fabrication-Dependent Waveguide-Sidewall…

146

The smooth-sidewall etching process, taking into account that the concurrent obtainment of fast enough and uniform in plane of the wafer etching is a must, was elaborated with use of the hydrochloric acid and hydrogen peroxide based aqueous solution, with volumetric proportions of HCl:H2O2:H2O = 40:4:1. Because the control of mesa shape (etching process spatial uniformity) and control of sidewall morphology are ensured only when not excessive velocity of etching is applied, this etchant has to be kept at the temperature T = 20 °C. As the etching-process rate strongly increases with temperature – only a few degrees higher temperature would be a reason of twice faster etching – the solution temperature control is truly crucial (Fig. 20.1).

That is why the freshly prepared solution was iced-cooled down to the necessary temperature. Such a rapid cooling procedure allows for avoidance of

Table 20.1 The tested etching solutions used for AlGaAs/GaAs-QCL mesa wet fabrication

Solution What else was tested

Etching rate (nm/s); (solution temperature ~20 °C, stirring rate ~60 rpm) Properties

NH4OH: H2O2: H2O

Stirring rate effect. 26 nm/s Hard mask is required; Fast enough etching rate; Low etching uniformity in plane of the wafer; Smooth mesa sidewalls

20:7:73

NH4OH: H2O2: H2O

Stirring rate effect. 9 nm/s Hard mask is required; Too slow etching rate; Smooth mesa sidewalls20:12:480

H2SO4: H2O2 Stabilization and stirring rate effect.

27 nm/s Fast enough etching rate; Low etching uniformity in plane of the wafer; Rough mesa sidewalls

4:1

H3PO4: H2O2: H2O

Stabilization and aging of solution.

3 nm/s Too slow etching rate; Smooth mesa sidewalls

1:1:25HCl: H2O2:

H2OStirring rate and solution

temperature effect. Stabilization and aging of solution.

25 nm/s Fast enough etching rate; Low etching uniformity in plane of the wafer; Smooth mesa sidewalls

80:4:1

HCl: H2O2: H2O

Stirring rate and solution temperature effect. Stabilization and aging of solution.

18 nm/s Fast enough etching rate; Low etching uniformity in plane of the wafer; Smooth mesa sidewalls

60:1:1

HCl: H2O2: H2O

Stirring rate and solution temperature effect. Stabilization and aging of solution.

25 nm/s Fast enough etching rate; High etching uniformity in plane of the wafer; Smooth mesa sidewalls

40:4:1

A. Szerling et al.

147

solution degradation. Fast enough stirring (~60 rpm) of the solution during the etching process is crucial, as well – as it allows for obtainment of uniform con-centration of reactants, and hence allows for spatially uniform etching rate.

The final result of the proper etching process performed in controlled conditions is presented in Fig. 20.2a, in which the double-trench circular-shaped mesa with fully controlled shape and with perfectly smooth walls is shown.

In contrast to this outstanding result, the rough sidewalls of mesa were obtained – as it was mentioned previously – till the conditions of mesa formation were appro-priately controlled (Fig. 20.2b, c). It was mentioned, as well, that such rough morphologies were a cause of decreasing of lasers’ performance, leading to a total deterioration in extreme cases. The quantitative analysis of this phenomenon is a subject of the next section.

Fig. 20.1 Etching rates for the selected HCl:H2O2:H2O etching solution at two different temperatures

Fig. 20.2 Scanning electron microscopic images of (a) the smooth-wall mesa waveguides (b, c) the mesa waveguides with unintentional roughness of walls


148

20.3 Analysis of the Influence of Mesa-Sidewall Roughness on Laser Performance

Only the mesas with local changes (disturbances) of width spreading continuously for the whole height of mesa were analyzed below; this type of roughness, only, may be taken into account within the Eqs. (20.1), (20.2), and (20.3). Information about roughness standard deviation (RMS roughness) of given mesa sidewall was obtained by taking top view optical images of consecutive fragments of mesa- waveguide through the Olympus microscope, and by analyzing the mesa edges using special processing toolbar (analySIS docu). Typically the edge of mesa is composed of several fragments of different roughness values. That is why, the overall roughness value (the average standard deviation of the mesa edge) σroughAll was calculated, according to Eq. (20.4):

σ σroughAlli

n

rough ai

i

a

L=

=∑

1

22

(20.4)

where: ai is the length of the i-th fragment, σrough ai is the roughness of the i-th fragment, n is the number of fragments of mesa, and L is the resonator (mesa) length.

The analyzed lasers were made of the same heterostructure, so we can accept they had the same values of gain coefficient and confinement factor. We may also assume the intrinsic and scattering waveguide losses of the lasers used in the experi-ment are nearly independent on their mesa widths [8].

A comparison of L-I-V characteristics is presented in Fig. 20.3a for selected two lasers with the same geometry of resonator, but with different mesa-edge morpholo-gies (with roughness values: 0 and 1.42 μm). The current-voltage characteristics for smooth-sidewall mesa and for rough-sidewall mesa lasers are essentially identical indicating the same electrical properties of the mesas. In contrast to this, the light output power versus current characteristics of these two lasers are very different – with significantly lower threshold current density and higher slope efficiency for smooth-edge mesa laser. We link this dependence with scattering losses, which are higher in case of rougher mesa sidewalls.

The light output power versus current characteristics measured for several lasers with different resonator geometries as well as different mesa-edge morphologies, i.e., characterized by different values of roughness standard deviation, are presented in Fig. 20.3b. The values of relative threshold current densities and slope efficiencies were extracted from these data, and the dependencies of these two parameters on the mesa-edge roughness are shown in Fig. 20.4.

The threshold current density dependencies on mesa-edge roughness standard deviation were fitted with the quadratic function of roughness (Eq. 20.5):

J A Cth rough= + σ 2

(20.5)

A. Szerling et al.

149

The fits were made separately for each group of lasers joining the devices with the same mesa length (Fig. 20.4a) – in order to take into account the possible influ-ence of different values of mirror losses for the lasers with different resonator lengths, and the relative parameters C and A were extracted. Similarly, the slope efficiency dependencies on mesa-edge σrough were fitted according to the Eq. (20.6):

ησS

rough

P

P P=

+1

2 3 2

(20.6)

Fig. 20.3 (a) L-I-V characteristics for two AlGaAs/GaAs QCLs with different morphologies of mesa sidewalls, i.e., with σrough = 0 μm and σrough = 1.42 μm; (b) L-I characteristics for AlGaAs/GaAs QCLs for waveguides with different standard deviation measured in pulse mode at T = 77 K


150

separately for each group of lasers with given resonator length (Fig. 20.4b), and the relative parameters P1, P2, P3 were extracted.

It is clearly seen, that the lasers’ threshold currents increase and the efficiencies decrease with mesa-edge roughness for every resonator length. The lasers with mesa-sidewall σrough ≥ 1.8 μm didn’t lase at all.

It is seen, particularly, that the P1 parameter values found by us are much lower than for the lasers with similar wavelengths reported by Toor et al. [1, 2], that is a possible result of different construction of laser’s heterostructure.

20.4 Conclusion

The AlGaAs/GaAs QCL mesa wet etching conditions necessary for the obtainment of proper shape of mesa and its smooth sidewall morphology were investigated and specified. The excessively time-consuming or cost-consuming etching procedures were rejected. The etching process main parameters under the investigation were: the etching solution temperature and the solution stirring rate. A special attention was paid in order to obtain low roughness of mesa sidewalls that is necessary to minimize the scattering losses of the laser in order to improve its performance. The scattering losses, increasing with mesa sidewall roughness, were found to be the reason of decline of lasers’ performance. It was experimentally checked, how the unintentional mesa-sidewall surface roughness – that arose for cases of not optimized etching conditions – affects the laser threshold current density and laser efficiency. The declining lasers’ performance for unintentional mesa-sidewall roughness σrough increasing in the range of 0–1.42 μm was presented. It was found that the devices with σrough ≥ 1.8 μm didn’t lase.

Fig. 20.4 Experimental data (points) and fits made for groups of lasers with given resonator lengths (solid lines), collected for AlGaAs/GaAs QCLs: (a) threshold current density versus mesa- sidewall roughness standard deviation – the fits are based on Eq. 20.5; (b) slope efficiency versus mesa-sidewall roughness standard deviation – the fits are based on Eq. 20.6.

A. Szerling et al.

151

Acknowledgments The work was financially supported by Polish National Science Center by project 5028/B/T02/2011/40 and project 2011/03/D/ST7/03146.

References

1. Toor F et al (2008) Effect of waveguide sidewall roughness on the threshold current density and slope efficiency of quantum cascade lasers. Appl Phys Lett 93:031104-1–031104-3. doi:10.1063/1.2962984

2. Toor F et al (2009) Effect of waveguide sidewall roughness on the performance of quantum cascade lasers. Proc SPIE 7230:72301P-1–72301P-10. doi:10.1117/12.808227

3. Kosiel K et al (2012) Multi-step interrupted-growth MBE technology for GaAs/AlGaAs (∼9.4 μm) room temperature operating quantum-cascade lasers. Opto Electron Rev 20(3):239–246. doi:10.2478/s11772-012-0029-7

4. Kosiel K et al (2011) Room temperature AlGaAs/GaAs quantum cascade lasers. Photonic Lett Poland 3(2):55–57

5. Kosiel K et al (2011) AlGaAs/GaAs quantum cascade lasers for gas detection systems. Infrared, Millimeter and Terahertz Waves (IRMMW-THz), 2011, 36th international conference on. IEEE, , doi: 10.1109/irmmw-THz.2011.6105118

6. Gniazdowski Z. et al Data not published7. Wasiak M, Sarzała R Private communication8. Pierścińska D et al (2012) Electrical and optical characterisation of mid-IR GaAs/AlGaAs

quantum cascade lasers. Proc SPIE 8432:84321S. doi:10.1117/12.922020


http://dx.doi.org/10.1063/1.2962984

http://dx.doi.org/10.1117/12.808227

http://dx.doi.org/10.2478/s11772-012-0029-7

http://dx.doi.org/10.1109/irmmw-THz.2011.6105118

http://dx.doi.org/10.1117/12.922020


Abstract Recent advances in the development of compact sensors based on mid- infrared continuous wave (CW), thermoelectrically cooled (TEC) and room temperature operated quantum cascade lasers (QCLs) for the detection, quantifi ca-tion and monitoring of trace gas species and their applications in environmental and industrial process analysis will be reported. These sensors employ a 2 f wavelength modulation (WM) technique based on quartz enhanced photoacoustic spectroscopy (QEPAS) that achieves detection sensitivity at the ppbv and sub ppbv concentration levels. The merits of QEPAS include an ultra-compact, rugged sensing module, with wide dynamic range and immunity to environmental acoustic noise. QCLs are convenient QEPAS excitation sources that permit the targeting of strong fundamental rotational-vibrational transitions which are one to two orders of magnitude more intense in the mid-infrared than overtone transitions in the near infrared spectral region.

Keywords Laser spectroscopy • Quartz enhanced photoacoustic spectroscopy • Wavelength modulation spectroscopy • Quantum cascade lasers • Trace gas detection • Carbon monoxide • Nitric oxide

Chapter 21 Mid-infrared Laser Based Gas Sensor Technologies for Environmental Monitoring, Medical Diagnostics, Industrial and Security Applications

Frank K. Tittel , Rafał Lewicki , Mohammad Jahjah , Briana Foxworth , Yufei Ma , Lei Dong , Robert Griffi n , Karol Krzempek , Przemyslaw Stefanski , and Jan Tarka

F. K. Tittel (*) • R. Lewicki • M. Jahjah • B. Foxworth • Y. Ma Department of Electrical and Computer Engineering , Rice University , 6100 Main Street , Houston , TX 77005 , USA e-mail: [email protected]

L. Dong • R. Griffi n Department of Civil and Environment Engineering , Rice University , 6100 Main Street , Houston , TX 77005 , USA

K. Krzempek • P. Stefanski • J. Tarka Laser and Fiber Electronics Group , Wroclaw University of Technology , Wybrzeze Wyspianskiego 27 , 50-370 Wroclaw , Poland

154

21.1 Introduction

The development of compact trace gas sensors, in particular based on quantum cascade (QC) lasers, permit the targeting of strong fundamental rotational-vibrational transitions in the mid-infrared that are one to two orders of magnitude more intense than overtone transitions in the near infrared. The architecture and performance of three sensitive, selective and real-time gas sensor systems based on mid-infrared semiconductor lasers will be described [ 1 – 4 ]. A QEPAS based sensor capable of ppbv level detection of CO, a major air pollutant, was developed. We used a 4.61 μm high power CW DFB QCL that emits a maximum optical power of more than 1 W in a continuous-wave (CW) operating mode [ 3 , 5 , 6 ]. For the R(6) CO line, located at 2,169.2 cm −1 , noise-equivalent sensitivity (NES, 1σ) of 2 ppbv was achieved at atmospheric pressure with a 1 s data acquisition time. Furthermore, a high performance (>100 mW) 5.26 μm CW TEC DFB-QCL (mounted in a high heat load (HHL) package) based QEPAS sensor for atmospheric NO detection will be reported [ 7 , 8 ]. A 1σ minimum detection limit of 3 ppbv was achieved for a sampling time of 1 s using interference free NO absorption line located at 1,900.08 cm −1 .

21.2 Quartz Enhanced Photo-Acoustic Spectroscopy (QEPAS)

In this work we selected quartz-enhanced photoacoustic spectroscopy (QEPAS), a gas-sensing technique fi rst reported in 2002 by our Rice University Laser Science Group [ 9 , 10 ]. The QEPAS sensor technology allows performing sensitive trace gas measurements in gas samples of a few mm 3 in volume. QEPAS employs readily available quartz tuning forks (QTFs) as sharply resonant acoustic transducers, instead of broadband electric microphones used in conventional photoacoustic spectroscopy sensor systems. The QTF is a piezo-electric element, capable of detecting weak acoustic waves generated when the modulated optical radiation interacts with a trace gas. The mechanical deformation of the QTF due to interac-tion with the acoustic waves results in the generation of electrical charges on its electrode pairs deposited on the prongs of the QTF. In order to further enhance the QEPAS signal, a so-called micro-resonator (mR) can be added to the QTF sensor architecture. The mR typically consists of two metal tubes [ 10 , 11 ]. The QTF is positioned between the tubes to probe the acoustic waves excited in the gas contained inside the mR. To date, such a confi guration has been used in most reported QEPAS based gas sensors [ 12 , 13 ]. A recent optimization study revealed that for a 32 kHz QTF, two 4.4 mm-long tubes with 0.5–0.6 mm inner diameter yields the highest QEPAS signal-to-noise ratio (SNR) [ 11 ]. However to simplify optical alignment process and eliminate any potential optical fringes the 4 mm long tubes with 0.8 mm inner diameter are commonly used for QEPAS experiments in the mid- infrared region.

F.K. Tittel et al.

155

21.3 QEPAS Based ppb-Level Detection of Carbon Monoxide

21.3.1 Motivation for CO Detection

Carbon monoxide (CO), one of the major air pollutants in the United States, is mainly produced and released into the atmosphere by a variety of incomplete combustion activities, including the burning of natural gas, fossil fuel, and other carbon containing fuels. CO has an important impact on the atmospheric chemistry through its reaction with hydroxyl (OH) for troposphere ozone formation and also can affect the concentration level of greenhouse gases [ 13 – 15 ]. Furthermore, CO even at low concentration levels is dangerous to human life and therefore must be accurately and precisely monitored in real time.

21.3.2 CW DFB-QCL Based QEPAS Sensor System

A schematic of the QEPAS based CO sensor platform is shown in Fig. 21.1 . As an excitation source a 4.61 μm high power, continuous wave (CW), distributed feed-back quantum cascade laser (DFB-QCL) operating at 10 °C from Northwestern University [ 2 , 5 , 6 ] was employed. An external water cooling system was used to remove the heat dissipation from the hot surface of a TEC mounted in a commercial QCL housing (ILX Lightwave Model LDM-4872). The DFB-QCL beam is collimated

Fig. 21.1 Schematic confi guration of a high power 4.61 μm CW RT TEC DFB-QCL based QEPAS system for CO and N 2 O ppbv level detection. Pc L plano-convex lens, Ph pinhole, QTF quartz tuning fork, mR acoustic micro-resonator, RC reference cell

21 Mid-infrared Laser Based Gas Sensor Technologies…

156

using a black diamond antirefl ection coated aspheric lens optimized for 3–5 μm spectral region with a 1.7 mm effective focal length (Lightpath model 390037-IR3). In order to further improve the QCL beam quality two additional 50 mm and 40 mm focal length plano-convex CaF 2 lenses, L 1 and L 2 , and a pinhole with diameter of 200 μm as a spatial fi lter were used. The second lens L 2 was used to direct the laser beam through the mR and the gap between prongs of the QTF, located inside an acoustic detection module (ADM), with a transmission effi ciency of >93 %. A ZnSe wedged window acting as a beam splitter (BS) is placed after the ADM to refl ect ~20 % of the DFB-QCL beam into a gas reference channel. The rest of the high power CW DFB-QCL beam is delivered to an optical power meter (Ophir model 3A-SH) and used for alignment verifi cation of the QEPAS system. A 3 f reference channel signal is employed for locking of the QCL laser frequency to the peak of absorption line of the target analyte. In the reference channel the QCL beam is detected by a pyroelec-tric detector (InfraTec model LIE-332f-63) after passing through a reference gas cell. For precise and accurate CO concentration measurements, a 5 cm long reference gas cell fi lled with a 500 ppm CO:N 2 mixture at 150 Torr pressure (fabricated by Wavelength References, Inc) was used.

To improve the CO vibrational-translational relaxation processes an external humidifi er was added at the inlet to the ADM of the QEPAS sensor system. In this case the addition of a 2.6 % H 2 O vapor concentration to the target trace gas mixture acts as an effective catalyst and results in higher detected amplitude for CO and N 2 O. A needle valve and fl ow meter (Brandt Instruments, Inc., Type 520) were used to set and monitor the gas fl ow through the QEPAS sensor system at a constant rate of 140 ml/min. A pressure controller (MKS Instruments, Inc., Type 649) and a vacuum pump were employed to control and maintain the pressure in the sensor system. The DFB-QCL current and temperature were set and controlled by an ILX Lightwave current source (model LDX 3220) and a Wavelength Electronics temperature controller (model MPT10000), respectively. For sensitive CO concentration measurements wavelength modulation spectroscopy (WMS) with 2nd harmonic detection [ 10 , 11 ] was utilized. Modulation of the laser current was performed by applying a sinusoidal dither to the direct current ramp of the DFB-QCL at half of the QTF resonance frequency ( f = f 0 /2 ~ 16.3 kHz). The piezoelectric signal generated by the QTF was detected by a low noise trans-impedance amplifi er with a 10 MΩ feedback resistor and converted into a voltage. Subsequently this voltage was transferred to a custom built control electronics unit (CEU). The CEU provides the following three functions: (1) measurement of the QTF parameters, i.e. the quality factor Q, dynamic resistance R, and the resonant frequency f 0 ; (2) modulation of the laser current at the frequency f = f 0 /2; and (3) measurements of the 2 f and 3 f harmonic components generated by the QTF.

21.3.3 Performance of a 4.6 μm CW TEC DFB-QCL

The optical power emitted by the DFB-QCL operating at 1,250 mA current and 10 °C temperature is 987 mW in the CW operating mode (see Fig. 21.2a ).

F.K. Tittel et al.

157

The experimentally determined temperature and current tuning coeffi cients are −0.16 cm −1 /°C and −0.0065 cm −1 /mA, respectively. This DFB-QCL can be current tuned to target the R(5) and R(6) absorption lines of the ν 1 CO fundamental band at 2,165.6 cm −1 and 2,169.2 cm −1 , respectively (see Fig. 21.2b ).

21.4 Experimental Details

21.4.1 Selection of CO and N 2 O Spectrum Absorption Line

Quantitative measurements of CO in the ν 1 fundamental rotational-vibrational band were reported previously by several research groups [ 13 , 15 – 17 ]. In Ref. [ 13 ] CO detection was performed using the R(8) CO absorption line located at 2,176.3 cm −1 by employing a Daylight Solutions, Inc external cavity quantum cascade laser (EC-QCL) based QEPAS sensor system. This CO sensor was operating at a reduced gas pressure of 100 Torr in order to avoid partial overlap with neighboring nitrous oxide (N 2 O) line. In Ref. [ 16 ] the authors targeted the R(12) CO absorption line located at 2,190.0 cm −1 which has a small spectral overlap N 2 O. However the R(12) line does not have a high absorption line intensity compared to other CO lines of the R branch [ 18 ]. In this work the R(6) CO absorption line located at 2,169.2 cm −1 was selected in order to measure the CO concentration with high accuracy and detection sensitivity.

To assess potential interferences from other atmospheric species, HITRAN based spectra of CO, N 2 O, and H 2 O absorption lines near 2,169 cm −1 were simulated at atmospheric pressure (760 Torr) and depicted in Fig. 21.3a . From simulated 2nd harmonic absorption spectra (Fig. 21.3b ) it is clear that the R(6) CO line is free from spectral interference and can be used effectively in wavelength modulation spectroscopy of CO. Furthermore, the R(6) is one of the strongest line in the CO v 1 vibrational band with a 1.54 times higher absorption line intensity than for the R(12) line. For N 2 O concentration measurements, an interference-free P(41) N 2 O

Fig. 21.2 ( a ) LIV curve of a 4.61 μm RT, CW, DFB-QCL from Center for Quantum Devices, Northwestern University; ( b ) DFB-QCL current tuning at different operating temperatures


158

absorption line located at 2,169.6 cm −1 was selected at a gas pressure of 100 Torr. The optical power measured after the ADM was 400 mW near 2,169 cm −1 for the CW DFB-QCL operating at 10 °C. The high QCL power helps to improve the QEPAS signal (S 0 ), which is proportional to: S 0 ~ (α·P·Q)/ f 0 where α is the absorption coeffi cient, P is the optical power, Q is the quality factor of the resonator and f 0 is the resonant frequency.

21.4.2 Line Locking for Continuous Monitoring of CO and N 2 O Concentration Levels

Continuous monitoring of CO and N 2 O concentration levels and the evaluation of the long term sensor performance of the QEPAS based sensor system was performed in the line locking mode, where the CW DFB-QCL frequency is kept at the center of the targeted absorption line. For line-locked measurements of the CO concentration at atmospheric pressure the modulation depth decreased from an optimum value of 50 mA to 40 mA because the 3 f reference signal shape for the QEPAS sensor operating at 760 Torr is pressure broadened. The sealed CO reference cell was fi lled at a total pressure of 150 Torr.

Fig. 21.3 HITRAN based simulation spectra of CO and N 2 O at a temperature of 296 K, a standard atmospheric pressure, an optical path length of 1 cm for 200 ppbv CO, 2 % H 2 O and 300 ppbv N 2 O, respectively. ( a ): absorption spectra; ( b ): Second harmonic absorption spectra

F.K. Tittel et al.

159

To verify the linear response of the mid-infrared QEPAS based CO sensor platform a calibration mixture of 5 ppmv CO:N 2 containing a constant 2.6 % concentration of water vapor was diluted six times down to 50 ppbv CO concentration levels (Fig. 21.4a ). The data acquisition time for these measurements was set to 1 s.

The measured QEPAS signal amplitude as a function of CO concentration, is plotted in Fig. 21.4b . The calculated R-square value, which represents how well the regression line approximates real data points, is equal to ~0.999 after a linear fi tting procedure. This implies that the sensor system exhibits a good linearity response to monitored CO concentration levels. However, due to the decrease of the modulation depth to 40 mA, the measured signal amplitude of moisturized 5 ppmv CO: N 2 mixture is 22 % lower compared to the line scanning mode experiments if a 50 mA modulation depth is used. Based on the data in Fig. 21.4a , the calculated MDL is 1.9 ppbv which is in good agreement with the MDL value that was calculated for the QEPAS sensor operated in the scanning mode. The re-calculated NNEA coeffi cient in this case is 2.04 × 10 −8 cm −1 W/√Hz.

To evaluate the N 2 O QEPAS sensor performance similar measurements were carried out by targeting the P(41) N 2 O absorption line using a certifi ed mixture of 1.8 ppmv N 2 O:N 2 . The optimum signal level was obtained when the gas pres-sure and modulation depth were set to 100 Torr and 20 mA, respectively. The addition of a 2.6 % H 2 O concentration to the analyzed N 2 O:N 2 mixture resulted in a fivefold enhancement of QEPAS signal amplitude and resulted in a MDL of 23 ppbv. The corresponding NNEA coeffi cient was found to be 2.91 × 10 −9 cm −1 W/√Hz.

To investigate the long term stability and precision of the CO QEPAS sensor an Allan deviation analysis was performed by passing pure N 2 through the sensor system. From the Allan deviation plot shown in Fig. 21.5 the optimum averaging time for the CO sensor is found to be 500 s, which results in a MDL of 280 pptv. A similar Allan deviation analysis was also performed for the N 2 O QEPAS sensor

40

x104a bx104

30

20

10

0

40

30

20

10

017:00

Modulation depth=40 mAP=760 Torr

Linear fity=7459+83x R-square=0.99873

Gas with 2.6% H2O

17:20

CO

QE

PA

S S

igna

l (C

nts)

CO

QE

PA

S S

igna

l (C

nts)

17:40

Time (HH:MM) CO concentration (ppm)

0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.018:00 18:20

N2N2

2500 ppb

5000 ppb

1000 ppb

500 ppb250 ppb

100 ppb50 ppb

Fig. 21.4 ( a ) QEPAS signal amplitude recorded in the line locking mode as the CO concentration is varied at atmospheric pressure and a modulation depth of 40 mA. ( b ): QEPAS signals amplitude averaged from Fig. 21.4a as a function of CO concentration. 1cnt = 6.67 × 10 −16 A


160

when the laser wavelength was locked to the P(41) N 2 O absorption line. In this case, after averaging the acquired data for 500 s a MDL of 4 ppbv was achieved for N 2 O measurements.

For ambient CO and N 2 O concentration measurements using a line-locking operational mode, an inlet tube of the QEPAS sensor was placed outside the laboratory and atmospheric air was pumped into the sensor. The results of continuous measure-ments of atmospheric CO and N 2 O concentration levels for a 5 h period are shown in Fig. 21.6a, b , respectively. The highest CO concentration spikes are caused by cigarette smoke whereas all other less intense spikes, recorded on top of the CO atmospheric background of ~130 ppbv, are due to automobile emissions. The mean atmospheric concentration of N 2 O was calculated to be 350 ppbv when using the P(41) N 2 O line at 2,169.6 cm −1 . Due to a long atmospheric residence time, the N 2 O concentration is well mixed in the lower atmosphere and therefore its atmospheric concentration level is relatively stable as can be seen from Fig. 21.6b .

21.5 QEPAS Based ppb-Level Detection of Nitric Oxide (NO) Detection

21.5.1 Motivation for NO Detection

The capability of detecting and quantifying nitric oxide (NO) at ppbv concentration levels has an important impact in diverse fields of applications including environmental monitoring, industrial process control and medical diagnostics. The major sources of NO emission into the atmosphere are associated with industrial combustion processes as well as automobile, truck, aircraft and marine transport emis-sions. Long term, continuous, reliable NO concentration measurements in ambient

Fig. 21.5 Allan deviation plot for time series measurements of pure N 2 for the QEPAS based CO sensor system

F.K. Tittel et al.

161

air are important because of NO’s role in the depletion of earth’s ozone layer and in the formation of acid rains and smog [ 14 , 18 ].

21.5.1.1 CW TEC DFB-QCL Based QEPAS NO Sensor System

The QEPAS sensor for NO detection is similar to Fig. 21.1 . In order to target the optimum H 2 O and CO 2 interference-free NO doublet absorption lines centered at 1,900.08 cm −1 a 5.26 μm CW HHL packaged TEC DFB-QCL was used as an excitation source (see Fig. 21.7 ). The DFB-QCL emits ~100 mW optical power at an operating temperature and current of 22 °C and 890 mA, respectively. Similar to the CO QEPAS sensor the DFB-QCL current and temperature were set and controlled by a control electronics unit (CEU), which is also employed to modulate the laser current, to lock the laser frequency to the selected absorption line, and to measure the current generated by QTF in response to the photoacoustic signal. During the NO sensor evaluation test all the measurements were performed at a gas pressure of 250 Torr and modulation depth of 5 mA [ 7 ].

Similar to the CO QEPAS sensor the DFB-QCL current and temperature were set and controlled by a CEU, which is also employed to modulate the laser current, to lock the laser frequency to the selected absorption line, and to measure the current generated by QTF in response to the photoacoustic signal. During the NO

Fig. 21.6 Continuous monitoring of atmospheric CO and N 2 O concentration levels from an air sampled on Rice University campus, Houston, TX, USA (Latitude and longitude are: 29° 43′ N/95° 23′ W). ( a ): CO concentration measurements; ( b ): N 2 O concentration measurements


162

sensor evaluation test all the measurements were performed at a gas pressure of 250 Torr and modulation depth of 5 mA [ 7 ].

A photograph of a completed compact, autonomous DFB-QCL based WMS QEPAS NO platform enclosed in a 12.3 × 5.3 × 5.1 in. aluminum enclosure is shown in Fig. 21.8 .

The 2 f QEPAS signal amplitude when the DFB-QCL frequency is tuned across and locked to the NO doublet absorption line at 1,900.08 cm −1 is depicted in Fig. 21.9a, b , respectively. For a 95 ppbv of NO in N 2 calibrated mixture and 2.5 % water vapor concentration the calculated noise-equivalent (1σ) concentration of NO with a 1 s averaging time is 3 ppbv at gas pressure of 250 Torr. The corresponding absorption coeffi cient normalized to the detection bandwidth and optical power is 6.2 × 10 −9 cm −1 W/Hz 1/2 .

1896

15°C

I = 800 mA

a b

M575AL (Epi-down)

5.27 5.26 20

16

12

8

4

0

Spe

ctra

l Pow

er D

ensi

ty (

arb)

AR/HR; CW

Resolution is instrumentlimited @ 0.123 cm-1

25°C35°C

1898

Wavenumber (cm-1)Wavenumber [cm-1]1900 1902 1904

1895P=1 atm, L=1m

100 ppm of NO 5% of H2O 5% of CO2

1900

Abs

orpt

ion

[%]

1905

Fig. 21.7 ( a ) Emission spectra of 1,900 cm −1 TEC CW DFB QCL and ( b ) HITRAN simulated spectra of NO, H 2 O, and CO 2

Fig. 21.8 Compact Prototype NO sensor

F.K. Tittel et al.

163

Similar to CO QEPAS sensor a long term stability of the NO sensor platform was also investigated by using an Allan variance analysis. From the Allan deviation plot the optimum averaging time for compact prototype NO QEPAS sensor is 200 s, what corresponds to an improved NO minimum detectable concentration of ~0.3 ppbv. For the purpose of environmental monitoring, where sensor time response is not a critical parameter, a 200 s averaging time can be normally utilized to allow a detection limit of NO below 1 ppbv.

21.6 Summary and Future Directions

This work focused on recent advances in the development of sensors based on infrared semiconductor lasers for the detection, quantifi cation and monitoring of trace gas species and their applications in atmospheric chemistry and industrial process control. The development of compact trace gas sensors, in particular based on quantum cascade lasers permit the targeting of strong fundamental rotational- vibrational transitions in the mid-infrared, that are one to two orders of magnitude more intense than overtone transitions in the near infrared.

The architecture and performance of two sensitive, selective and real-time gas sensor systems based on mid-infrared semiconductor lasers was described. High detection sensitivity at ppbv and sub-ppbv concentration levels requires sensitivity enhancement schemes such as wavelength modulation and quartz-enhanced-photoacoustic spectroscopy. These spectroscopic methods can achieve minimum detectable absorption losses in the range from 10 −8 to 10 −11 cm −1 /√Hz.

A QEPAS based sensor capable of ppbv level detection of CO, a major air pollutant, was developed. A 4.61 μm high power CW DFB QCL that emits a maximum optical power of >1 W in a continuous-wave (CW) operating mode [ 15 ] was used. Noise-equivalent sensitivity (NES, 1σ) of 2 ppbv was achieved at atmospheric pressure

8

95 ppb NO + 2.5% H2O, P=250 Torr95 ppb NO+ 2.5% H2O, P=250 TorrNO reference cell

860 864 868 872

Laser Current (mA)876 880 884

11:15:00

0

2

4

6

8

QE

PAS

Am

plitu

de (

a.u.

)

ba

11:26:36 11:38:12

Time (HH:MM:SS)

DFB-QCL frequency

For 1900.08 cm-1

NO line a MDL(1σ)=3 ppb in 1 sec

MDL(1σ)=3 ppb in 1 sec

locked to ν=1900.08 cm-1

11:49:48 12:01:24

N2-100

-50

0

50

150

1006

4

2

QE

PAS

2f s

igna

l (a.

u.)

3f s

igna

l (a.

u.)

0

-2

-4

-6

Fig. 21.9 2 f QEPAS signal amplitude when QCL frequency is tuned across ( a ) and locked ( b ) to the NO doublet absorption line at 1,900.08 cm −1


164

with a 1 s. acquisition time targeting the R(6) CO line, located at 2,169.2 cm −1 . Furthermore, a QEPAS sensor for atmospheric NO detection using a high performance (>100 mW) 5.26 μm CW TEC DFB-QCL (mounted in a high heat load package) is reported. A 1σ minimum detection limit of 3 ppbv was achieved for a sampling time of 1 s. using an interference free NO absorption line located at 1,900.08 cm −1 [ 7 , 8 ].

New target analytes such as SO 2 , OCS, CH 2 O, HONO, H 2 O 2 , C 2 H 4 and other hydrocarbons are planned. Furthermore, monitoring of broadband absorbers such as acetone, acetone peroxide and UF 6 will be investigated. In addition, ultra-compact, low-cost robust sensor designs are being considered.

Acknowledgments The Rice University group acknowledges fi nancial support from a National Science Foundation (NSF) grant EEC-0540832 entitled “Mid-Infrared Technologies for Health and the Environment (MIRTHE)”, a NSF-ANR award for international collaboration in chemistry “Next generation of Compact Infrared Laser based Sensor for environmental monitoring (NexCILAS)” and grant C-0586 from the Robert Welch Foundation.

References

1. Faist J (2013) Quantum cascade lasers. Oxford University Press, Oxford. ISBN 13: 978-0198528241

2. Capasso F (2010) High-performance midinfrared quantum cascade lasers. SPIE Opt Eng 49:111102

3. Razeghi M, Bai Y, Slivkin S, Davish SR (2010) High-performance InP-based midinfrared quantum cascade lasers at Northwestern University. SPIE Opt Eng 49:111103-4

4. Lyakh A, Maulini R, Tsekoun AG, Patel CK (2010) Progress in high-performance quantum cascade lasers. SPIE Opt Eng 49:111105

5. Razeghi M (2009) High-performance InP-based Mid-IR quantum cascade lasers. IEEE J Sel Top Quantum Elect 15:941–951

6. Le QY, Bai Y, Bandyopadhyay N, Slivken S, Razeghi M (2010) Room-temperature continuous wave operation of distributed feedback quantum cascade lasers with watt-level power output. Appl Phys Lett 97:231119-1

7. Dong L, Spagnolo V, Lewicki R, Tittel FK (2011) Ppb-level detection of nitric oxide using an external cavity quantum cascade laser based QEPAS sensor. Opt Express 19:24037–24045

8. Spagnolo V, Kosterev AA, Dong L, Lewicki R, Tittel FK (2010) NO trace gas sensor based on quartz enhanced photoacoustic spectroscopy and external cavity quantum cascade laser. Appl Phys B 100:125–130

9. Kosterev AA, Bakhirkin YA, Curl RF, Tittel FK (2002) Quartz-enhanced photoacoustic spectroscopy. Opt Lett 27:1902–1904

10. Kosterev AA, Tittel FK, Serebryakov D, Malinovsky A, Morozov A (2005) Applications of quartz tuning fork in spectroscopic gas sensing. Rev Sci Instrum 76:043105

11. Curl RF, Capasso F, Gmachl C, Kosterev AA, McManus B, Lewicki R, Pusharsky M, Wysocki G, Tittel FK (2010) Quantum cascade lasers in chemical physics. Chem Phys Lett Frontiers Article 487:1–18

12. Dong L, Kosterev AA, Thomazy D, Tittel FK (2010) QEPAS spectrophones: design, optimization, and performance. Appl Phys B 100:627–635

13. Ma Y, Lewicki R, Razeghi M, Tittel FK (2013) QEPAS based ppb-level detection of CO and N2O using a high power CW DFB-QCL. Opt Express 21:1008–1019

14. Seinfeld JH, Pandis SN (1998) Atmospheric chemistry and physics: from air pollution to climate change. Wiley, New York

F.K. Tittel et al.

165

15. Ma Y, Lewicki R, Razeghi M, Tittel FK (2013) QEPAS based ppb-level detection of CO and N2O using a high power CW DFB-QCL. Opt Express 21:1008–1019

16. Li J, Parchatka U, Königstedt R, Fischer H (2012) Real-time measurements of atmospheric CO using a continuous-wave room temperature quantum cascade laser based spectrometer. Opt Express 20:7590–7601

17. Tao L, Sun K, Amir Khan M, Miller DJ, Zondlo MA (2012) Compact and portable open-path sensor for simultaneous measurements of atmospheric N2O and CO using a quantum cascade laser. Opt Express 20:28106–28118

18. Kasyutich VL, Holdsworth RJ, Martin PA (2008) Mid-infrared laser absorption spectrometers based upon all-diode laser difference frequency generation and a room temperature quantum cascade laser for the detection of CO, N2O and NO. Appl Phys B 92:271–279



Abstract Passive THz imaging devices have big potential for solution of the security problem. Nevertheless, one of the main problems, which take place on the way of using these devices, consists in modest image quality, produced by the majority of passive THz cameras. To change this situation, it is necessary to improve the engineering characteristics (resolution, sensitivity and so on) of the THz camera or to use computer processing of the image. In our opinion, the last issue is more preferable because it is more inexpensive. We demonstrate the improvement of the quality of the images captured by TS4 – a commercially available THz passive cam-era manufactured by ThruVision Systems Ltd. Our approach bases on application of original spatial fi lters and algorithms, developed for computer processing of images produced by the passive THz camera. The presented examples show the big potential for the detection of small hidden objects at long distances (6–10 m) and observation the difference in temperature on the human body, which is caused by different temperatures inside the body. Developed algorithms are used also for computer pro-cessing of the images captured by THz passive camera manufactured by various Companies. In all cases, we increase the quality of the THz images.

Keywords Passive THz imaging • Computer processing • Security screening • De-noising • Forbidden objects

Chapter 22 Computer Processing of Images Captured with a Commercially Available THz Camera at Long Distances

Vyacheslav A. Trofi mov , Vladislav V. Trofi mov , Norbert Palka , and Marcin Kowalski

V. A. Trofi mov (*) • V. V. Trofi mov Faculty of Computational Math and Cybernetics , Lomonosov Moscow State University , Moscow , Russia e-mail: [email protected]

N. Palka • M. Kowalski Institute of Optoelectronics , Military University of Technology , Warsaw , Poland e-mail: [email protected]

168

22.1 Introduction

Detection and identifi cation of hidden substances (explosive materials, drugs, etc.) is a very urgent problem nowadays. As it is well known, for solution of these problems, the electromagnetic wave of THz range has a big potential in this fi eld and many companies propose active or passive systems. Passive THz imaging because of its unique properties is one of the promising technologies for the detec-tion of objects hidden under clothes. However, the main diffi culty in practical realization of the THz imaging systems consists in low image quality in many cases due to low sensitivity and small number of pixels in detecting modules of cameras. Below we study effi ciency of the TS4 camera from ThruVision Systems Ltd. for the detection of objects. Let us notice that in our opinion this camera has good performance and image quality.

Using computer processing [ 1 – 4 ] we enhance the quality of the image captured by the camera for long distances (5–10 m) and small sizes of objects hidden on a person. One has to notice that there are many fi lters for processing ordinary photos, which have been developed in various papers for enhancing of the image quality. Unfortunately, directly applying these fi lters to THz images is ineffective usually because the quality of photos is many times better than achievable THz image quality and usually the THz image consists of several thousand (2,000–25,000) pixels. This number of pixels is many times less than the corresponding number of pixels in a photo. Therefore, approaches useful for processing a high-resolution photo may not be directly suitable for lower resolution THz images.

It should be emphasized that the THz images, obtained by TS4 camera, were carried out in the Military University of Technology (Warsaw, Poland).

22.2 Ceramic Knife and Metallic Plate Located at 6 m from THz Camera

In our studies, we considered the detection of various dangerous objects like a pis-tol, a bomb, a metallic plate, metallic discs with various thickness, a knife and oth-ers. As an example of the computer processing results, we present the knife hidden under the clothes on body of the person (Fig. 22.1 ). Original images produced by the passive THz camera, are shown in Fig. 22.1 . We can hardly notice that the shape of the knife on the left chest of the person are blurred and unclear. After computer processing of the raw image, the edges of the image become more well-defi ned. One can see a detail of the knife (handle of the knife) and clothes (pocket of the shirt) of the person. It should be noticed that the contour of the object will be more clear in all images after computer processing of them (Fig . 22.2 ).

The second example is shown in Fig. 22.3 . Here we see the THz image of enough big metallic plate. Computer processing of the images allow to enhance the image

V.A. Trofi mov et al.

169

Fig. 22.1 Original images of the knife hidden under shirt ( a ) and sweater ( b ) shown by the passive THz camera located at 6 m from the person. Inset – photo of the knife

Fig. 22.2 Images of the knife hidden under shirt ( a ) and sweater ( b ) after computer processing by various fi lters

22 Computer Processing of Images Captured with a Commercially Available THz…

170

quality. As a result, we see very clear the contour of the plate. Original image pro-duced by the passive THz camera does not allow to see this feature of the image. In certain image one can see a shirt collar near the person neck. This circumstance illustrates big potential of additional computer processing of the image, which is produced by such THz camera.

22.3 Small Object Located at 8 or 10 m from THz Camera

Our computer code allows to increase the number of pixels belonging to the image incoming from the passive THz imaging device practically without losses of the image quality. This option is very useful for the vision of small object at big distance because the enhancement of number of pixels belonging to the image of the object leads to enhancement of sizes of the object in the picture and helps to see it on the computer screen.

As an example in Figs. 22.4 and 22.5 the original images of a person (a) con-cealing the small object near the neck under blouse is shown. The image is produced by passive THz camera, located at 8 or 10 m from a person, corre-spondingly for Figs. 22.4 and 22.5 . Noise of THz camera infl uences in strong way on image of the object and we do not see the small object even on part of the image at increased number of pixels (Figs. 22.4 and 22.5b, c ). Situation

Fig. 22.3 Raw THz images of the big plate, hidden under shirt or sweater captured by the passive THz camera located at 6 m from the person. Inset – photo of the metallic plate. Other images cor-respond to computer processing images by various fi lters


171

Raw image ofperson concealingthe small object.

Part of the rawimage of a person atincreased number ofpixels in three times(b) and computerprocessing of thisimage (c)

Part of the image (d) atincreased number of pixels (e)and computer processing ofthis image (f).

object

Small objectunder blouse

Full de-nosing(a)

(b)

(c)

(d)

(f)

(e) (g)

Fig. 22.4 Raw image of a person ( a ) concealing the small object (diameter is equal to 2.5 cm, see ( g )) under blouse displayed by the passive THz camera located at 8 m distance. Part of the raw image of a person at increased number of pixels in three times ( b ) and computer processing of this image ( c ). The image of a person after removing the noise ( d ). Part of the image ( d ) at increased number of pixels ( e ) and computer processing of this image ( f )

Raw image ofperson concealingthe small object.

Part of the rawimage of a person atincreased number ofpixels in three times(b) and computerprocessing of thisimage (c)

Part of the image(d) at increased numberof pixels (e) andcomputer processingof this image (f).

object

Small objectunder blouse

Full de-noising

(a)

(b)

(c)

(d)

(e)

(f)

(g)

Fig. 22.5 Raw image of a person ( a ) concealing the small object (diameter is equal to 2.5 cm, see ( g )) under blouse displayed by the passive THz camera located at 10 m distance. Part of the raw image of a person at increased number of pixels in three times ( b ) and computer processing of this image ( c ). The image of a person after removing the noise ( d ). Part of the image ( d ) at increased number of pixels ( e ) and computer processing of this image ( f )


172

changes cardinally if we remove a noise from the image (Figs. 22.4d and 22.5d ). We see (but, no so clearly) a spot near the neck of a person. After increasing of a number of pixels, the blot is seen more pronounced (Figs. 22.4e and 22.5e ). More clear image of the object one can appear if we done computer processing of the part of the image (Figs. 22.4f and 22.5f ). It is very important that the simi-lar features occur for the image of a person located at 10 m from the THz camera (Fig. 22.5 ).

22.4 Image of Gun Located at 8 m from THz Camera

Figure 22.6 demonstrates new possibilities of computer processing. Using the dou-ble data transformation and some math operations for de-noising, one can see a pocket with small object and gun located in one’s belt. We see dark spot in this region. It is very important that the computer processing allows to see a pocket with some object. This feature demonstrates a possibility of used algorithms.

22.5 Image of Big Metallic Plate Located at 10 m from THz Camera

The next example deals with the metallic plate located in the pocket on the breast. After computer processing of the image and de-noising, we can see very clear

image of the plate. Additionally, well-defi nite contour of the belt is seen in the Fig. 22.7 . Obviously, images produced by the THz camera cannot give such details.

22.6 Conclusion

Computer processing of the THz images captured by the passive camera can be a very useful tool, which allows to improve visibility of the images and help fi nd hidden items. Using computer processing one may enhance the image quality and reduce a noise in the image. In some cases, it is possible to achieve full de-noising of the image. After applying the computer processing of the image, its quality enhances many times. We stress that the performance of developed computer code is enough high and does not restrict the performance of passive THz imaging device.

The obtained results demonstrate the high effi ciency of our approach for the detection of hidden objects and they are a very promising solution for the security problem. Nevertheless, developing the new spatial fi lter for computer processing of the THz image remains a modern problem at present time.


173

Fig. 22.6 Raw image of a person ( a – d : 2012/02/25 11_45_36; 39; 40; 41) concealing the small gun under sweater. Figures ( e – i ) correspond to double transformation of data format. Figures ( k – n ) cor-respond to fi rst removing of the noise. Figures ( o – p ) correspond to second removing of the noise


174

Acknowledgments Russian part of the team was partly supported by Russian Foundation for Basic Research (grant number 11-07-13139-Ofi -m-2011-RRW).

References

1. Trofi mov VA, Trofi mov VV, Chao Deng, Yuan-meng Zhao, Cunlin Zhang, Xin Zhang (2011) Possible way for increasing the quality of imaging from THz passive device. Proc SPIE 8189:81890L. doi: 10.1117/12.897900

2. Trofi mov VA, Trofi mov VV (2012) Real-time computer treatment of THz passive device images with the high image quality. Proc SPIE 8362:83620J. doi: 10.1117/12.919770

3. Trofi mov VA, Trofi mov VV (2012) Real-time computer processing of image from THz passive imaging device for improving of images. Proc SPIE 8544:85440J. doi: 10.1117/12.965546

4. Trofi mov VA, Trofi mov VV, Palka N, Kowalski M (2012) Increasing the quality of image of a commercially available passive THz camera due to computer processing of image. Proc SPIE 8546:85460G. doi: 10.1117/12.965539

Fourier transform using

Sequence of row images

sweaterObject; sizes:12 cm × 9 cm

Computer processed imagesCorrelation function using

BeltBelt

Pocket andobjectPocket and

object

a

c d

b

Fig. 22.7 Raw image of a person ( a – d : 2012/02/25 11_45_36; 39; 40; 41) concealing the small gun under 10 m without de-noising


http://dx.doi.org/10.1117/12.897900

http://dx.doi.org/10.1117/12.919770

http://dx.doi.org/10.1117/12.965546

http://dx.doi.org/10.1117/12.965539


Abstract Pulsed radars widely used in microwave and optical frequencies are not available in very interesting for applications millimeter-wave (3–0.3 mm/0.1–1 THz) spectral range due to lacking of a miniature, simple, high-power all - solid-state pulsed emitter. We suggest utilization of a phenomenon which we recently found in powerfully avalanching GaAs-based bipolar transistor structure and termed “ collapsing fi eld domains”. Here we explain the operating principle of the emitter and show an example of transmission sub-terahertz imaging. The pulsed imaging system uses a prototype of our new source operating in sub-nanosecond time domain in milliwatt range, and a commercial Schottky detector. The imaging traditionally utilizing pulse attenuation in the objects is compared with suggested here time- domain imaging that uses the propagation delay of the transmitted pulse. Advantages of each of those two operating modes are discussed. Presented here pulsed imaging radar is a prospective candidate for Detection of Explosives and CBRN in places of people crowding when utilization of x-ray is problematic. It can also be used in nondestructive tests when x-ray does not show suffi cient contrast, or a portable, not dangerous for humans and cost-effective system is required. Very interesting is a prospective of skin and breast cancer detection. Realization of the refl ection imaging regime using the same source is pending.

Keywords THz source • Pulsed radar • Millimeter-wave imaging • Avalanche switching • Negative differential mobility

Chapter 23 Transmission Subterahertz Imaging Utilizing Milliwatt-Range Nanosecond Pulses from Miniature, Collapsing-Domain-Based Avalanche Source

S. N. Vainshtein and J. T. Kostamovaara

S. N. Vainshtein (*) • J. T. Kostamovaara Electronics Laboratory, Department of Electrical Engineering , University of Oulu , Oulu , Finland e-mail: [email protected]

176

23.1 Introduction

Well known THz gap (0.1–10 THz) between radio and optical frequencies tends to narrow down to 0.1–1.5 THz due to success in quantum-cascade lasers (QCL) development. The remaining gap is of major importance for sub-THz imaging in security, quality control and medicine. For example, a reliable detection of arms, plastic explosives and drags hidden under clothing or in parcels and envelopes require realization of active terahertz (THz, T-ray) imaging at relatively large distances (from a few to more than ~30 m) [ 1 , 2 ]. It means that preferable is probing signal within the spectral band below ~0.5 THz [ 3 , 4 ] as water vapor absorption continues growing at higher frequencies. The same argument is also valid for THz imaging and spectral analyses [ 5 , 6 ] in medicine (skin cancer: surface proteins) [ 7 ] biology (bioparticle absorption signatures) [ 8 ], biochemistry (DNA absorption spectra) [ 9 ] and in various quality control tasks. A requirement of a reasonable spatial resolution limits frequency of the source from down and thus the spectral band of ~0.1–0.5 THz seems to be optimal for many tasks.

On the other hand well known radar principle in its pulsed realization is widely used in both radio frequencies (e.g. centimetre wavelengths radars tradi-tional in military applications), and visible (VIS) or near-infrared (NIR) light (LIDARS, wavelength around ~1 μm). If a beam scanning or emitter/detector arrays are used, this allows the three-dimensional (3-D) refl ection imaging to be realized thanks to distance information provided by the time-of-fl ight technique. In case of semi- transparent objects for a particular electro-magnetic radiation one can realize transmission imaging by placing the same emitters and detectors from two sides of an object and measuring typically an attenuation of the radia-tion in the sample, and as will be seen below the pulsed time-of-fl ight technique can provide additional advantages.

Those informative techniques are not used , however, in sub-THz (millimetre- wave, 3–0.3 mm) range in which radiation penetrates through various dielectric objects as cloth, packages, various construction materials (unlike VIS and NIR), provides fairly good spatial resolution of ~1 mm (unlike microwave radars), and can be used for detection of narcotics and plastic explosives and give the contrast in some mate-rials where X-ray fails. (One principal importance with respect to X-rays is that THz radiation is not harmful for humans ). Furthermore, there is growing number of new applications of THz emission, e.g. THz contrast separating cancer and healthy tissue [ 10 – 13 ] both in vivo and in vitro, etc. The main problem is that miniature solid-state pulsed emitters in sub-THz range do not yet exist .

Fortunately, there are high-speed receivers available , which can be used for ranging with precision of ~1 mm (~6 ps in time interval measurement). Namely miniature, simple, and suitable for mm-wave radar Schottky detector exists and provides fairly good temporal response (sub-nanosecond and even picosecond range), but its sensitivity and noise-equivalent-power (NEP) require the detector to be used together with an emitter of at least milliwatt (better multi-milliwatt) range in peak power. Thus an appropriate temporal resolution of the millimetre-wave

S.N. Vainshtein and J.T. Kostamovaara

177

radar (imager) comparable with spatial resolution in lateral direction is possible provided sub-nanosecond (sub-ns) pulse were generated by miniature, simple in construction, highly effective and potentially very low-cost emitter operating at room temperature . Listed requirements refl ect the fact that future real-time imaging systems should generate images with a high frame rate, and must use a matrix (or at least a line) of array of both emitters and receivers. Despite signifi cant efforts within last two decades to fi ll the THz gap, none of the currently existing emitters satisfi es the listed requirements, which does not allow millimetre –wave radar to be properly realised.

We suggest an approach to the problem, which has chances to start a new era in active sub-THz imaging. It utilizes (i) unique , pulsed, high-speed, miniature sub- THz emitter that provides means to realize (ii) transmission imaging employing transmission delay contrast supplementing traditional attenuation measurements, which provides synergetic effect of using both methods together (see imaging examples given below in this chapter) and (iii) 3–D refl ection (radar) imaging which allows to see hidden objects under clothing and packing.

Brief history of the phenomenon our new source is based on is as follows. It was shown in 1955 [ 14 ] that a bipolar junction transistor (BJT) can be switched from high-impedance to low-impedance state by an application of a triggering pulse to the base, and remain in the low-impedance state with switched-off base current due to positive feedback between avalanche hole generation in the collector region and electron injection from the emitter. This regime, termed avalanche switching, allows relatively slow (≥100 ns) switching to be realized from the maximum voltage of collector-base breakdown down to ~ twice lower voltage and with the maximum current below ~1A. Very soon much faster (several nanosecond) switching down to much lower voltages and with the maximum currents up to several dozen amperes was experimentally found and termed “second breakdown”, while the physical reason for that phenomenon remained open till 1970th. Most popular analytical interpretation of the phenomenon in 1970th [ 15 ] remained unchanged until physics-based numerical modelling [ 16 – 19 ] in 2000th has shown that artifi cial 3-D assumptions of the model [ 15 ] are not needed for the phenomenon interpretation. Further detailed investigations have shown fairly peculiar 3-D transient behaviour [ 20 – 22 ] of fast avalanche switching that contradicts prepositions of the model [ 15 ], and this fi nally ruined old interpretation of the “second breakdown” which remained unchanged within more than four decades. Most important for us property of high-current, fast avalanche switching (“second breakdown”) [ 16 – 22 ] is that this mode forms fairly dense electron-hole (e-h) plasma in the collector region of a Si BJT, which co-exists with a domain of high electric fi eld (slightly exceeding the ionization threshold). In a Si BJT this stationary high-fi eld domain is situated in the boundary between the collector and sub-collector, while the situation changes drastically and becomes very peculiar in case of a GaAs BJT.

Instead of a single stationary fi eld domain (as in Si BJT), in a GaAs transistor multiple, moving fi eld domains appear which amplitude exceeds the ionization threshold by a factor of 2–3 [ 23 , 24 ] (see Fig. 23.1 ). This huge amplitude (up to 0.6 MV/cm) causes monstrous impact ionization in the domains and very fast

23 Transmission Subterahertz Imaging Utilizing…

178

growth in the density of electron-hole (e-h) plasma surrounding the domains. This results, in turn, in drastic domain shrinkage (“collapsing” [ 25 ]) with corresponding growth in the amplitude and further increase in the ionization rate. Very powerful impact generation of e-h plasma and very fast domain shrinkage cause unique superfast switching in GaAs avalanche transistors [ 26 , 27 ] and apparently in GaAs thyristors [ 28 ]. Even more exciting for applications is copious emission in sub-THz range [ 25 ] caused by picosecond-range current oscillations, which result from nucleation and annihilation of multiple collapsing domains and certain plasma waves. Despite the observed phenomenon that we term “collapsing fi eld domains” has similarities with Gunn effect [ 29 ], and fundamental physical reason [ 30 ] (negative differential mobility, NDM, and GaAs) is also similar, there are drastic differences between two phenomena. First, in Gunn effect only a single domain takes place, with an amplitude that is much smaller that the ionization threshold,

2.5 3.0 3.5 4.0 4.50

50

100

2.5 3.0 3.5 4.0 4.50

50

100

150

200

250

300

experiment

load

cu

rren

t, A

time, ns

elec

tric

fi

eld

x1

0- 5

V/c

m -5 0 5 10 15 20 25 30 350

2

4

6

-5 0 5 10 15 20 25 30 350

2

4

6

-5 0 5 10 15 20 25 30 350

2

4

6

-5 0 5 10 15 20 25 30 350

2

4

6

con

tact

con

tact

t=3.2 ns

n0 collector

n+

colle

cto

r

p b

ase

n+

emit

ter

t=3.287 ns

t=3.447 ns

t=3.647 ns

longitudinal (along the channels)coordinate, mm

simulations

experiment

colle

cto

r vo

ltag

e , V

simulations

a

b

Fig. 23.1 ( a ) experimental and simulated voltage and current temporal profi les during superfast avalanche switching of a GaAs bipolar junction transistor; ( b ) electric fi eld profi les corresponding to different instants of the switching transient in ( a ). Note that the ionization threshold E i is about 2 × 10 5 V/cm, while the maximum domain amplitudes in the fi gure are a factor of ~3 larger. The fi gure represents data of Ref. [ 24 ]


179

width of a few microns, only one sort of carriers (electrons) is presented and no impact ionization manifests itself. In contrary, in our case multiple fi eld domains are generated, which number varies from a few to ~20 during the transient; the ionization threshold is exceeded several times and causes extremely powerful impact ionization; both electrons and holes are presented forming plasma between the domains; drastic domain shrinkage causes very fast voltage reduction across the structure and even complete domain disappearance [ 27 ]. As for the fundamental physical reason, the collapsing domain effect requires not simply presence of NDM [ 30 ] (as Gunn effect), but much more hard condition must be satisfi ed [ 31 ]. Namely, NDM must take place well above the ionization threshold . We have proved this condition so far [ 31 ] only for GaAs, in which NDM takes place up to at least 600 kV/cm, while for other III-V materials the question is still open.

Reported in Ref. [ 25 ] peak power emitted in sub-THz range by a sample used in those fi rst experiments was around ~0.1 mW. That structure and chip design was certainly not optimized as a sub-THz source, and purpose of the experiment [ 25 ] was only a demonstration of the emission in the support of the collapsing domain concept. Here we report on first steps towards development of an optimized sub- THz (mm-wave) pulsed emitter, and demonstrate several examples of pos-sible applications. No detailed information will be given here on the device design except the claim that the emission principle is based on the collapsing domains described above. Instead we discuss briefl y the results of a characterization of one of the most promising emitter prototypes available at the moment in our laboratory, describe some details of the measurement methodology, and show several imaging examples some of which are relevant to Detection of Explosives and CBRN (as an alternative to, e.g., [ 32 , 33 ]).

23.2 Emitter Characterization

Shown in Fig. 23.2 are the current pulse across the emitter and its sub-THz response measured by Schottky diode-based Quasi-Optical Detector (QOD, Virginia Diodes Inc.).

The peak power measured using Golay cell (see Fig. 23.3 ) was ranging between 3 and 4.2 mW, which correspond to extremely high power density generated in the active region of the device ~1 GW/cm 3 (~1 mW/μm 3 ). This resolves an important trade-of typical of solid-state microwave/mm-wave emitters: increased area of the active region (aiming at increasing the power) increases also the capacitance that shunts high-frequency signal. Then, sharp leading edge allows picosecond precision of time-interval detection to be realized providing an attractive option of transmission imaging with propagation delay contrast . Spectral characterization (see Fig. 23.4 ) of the emission was performed using a set of specially developed, manufactured and characterized band-pass fi lters.


180

23.3 Imaging Examples

Several imaging examples utilizing the fi rst prototype of our original pulsed emitter are shown in Figs . 23.5 , 23.6 , 23.7 , and 23.8 . Figures 23.5 and 23.6 show unwanted effect of the diffraction in the traditional attenuation imaging, and new possibilities provided by the propagation delay imaging with picosecond-range temporal resolution.

0,0 0,5 1,0 1,5 2,0 2,50,0

0,3

0,6

0,9

1,2

0,0

0,1

0,2

0,3

0,4

curr

ent,

A

ampl

ified

(30

DB

) de

tect

or r

espo

nse,

V

transient time, ns

Schottky diode response:

Fig. 23.2 The current pulse across an emitter prototype and emitted pulse of sub-THz emission measured using 30 GHz oscilloscope in the output of a Schottky detector

transistor

circuitboard

GolaycellGolaycell

-10 0 10 20 30 40 50 60 700.0

0.1

0.2

0.3

0.4

aver

age

po

wer

, mW

(Go

lay

cell

resp

on

se)

time, ms

train of 1.1 ns pulseswith 100 KHz rate

Peak Power = 460 nW x 10 ms / 1.1 ns = 4.2 mW

Fig. 23.3 Peak power estimate using Golay cell and an elliptic mirror. The emission of the source collected from hemisphere is focused on the Golay cell; an average power of the train of nanosec-ond pulses measured with the calibrated cell allowed the peak power of the pulse to be calculated using known duty cycle and pulse duration (see Fig. 23.2 )


181

One can see from the images that the propagation delay imaging when supple-menting the attenuation transmission imaging can provide at least twice as reach information as the traditional attenuation imaging alone. This concerns removed diffraction effects (Figs. 23.5 and 23.6 ), improved spatial resolution (compare teeth in two images of Fig. 23.5 ), clear location of the hidden “narcotics” in Fig. 23.7 and clear visualization and even inhomogeneities in the “plastic explosives” which are hidden in the granulated sugar and cannot be detected even by using X-ray. Further advances of suggested here source may be found in a future from testing in vivo and in vitro breast and skin cancer, permanent controlling of the conveyer belts in mines (which destruction causes big troubles), real-time mail and passels security check-ing on a conveyer belt in a post offi ce, etc. etc. Some of the applications will require realization of the refl ection imaging that is currently under planning. Realistic is an application of 2-D arrays of our emitters in long-distance security check in the air-ports that requires very signifi cant problems to be solved, however. More realistic looking is sub-THz radars for installation on the police cars with a moderate distance of below 10 m. (New York police has recently got experimental mm-wave radars with imaging distance up to only 1 m).

23.4 Conclusion

We believe that suggested here mm-wave (sub-THz source) based on the collapsing domain phenomenon may open a new era in pulsed mm-wave radars for their appli-cations in large number of diagnostic and quality control tasks in medicine, security, industry and building. Despite research and development of this source is currently in the stage of its infancy, we can predict already now very signifi cant increase in

0.0 0.5 1.0 1.5 2.0 2.50.0

0.1

0.2

0.3

0.4

0.5

0.6

ampl

ified

(30

DB

) de

tect

or r

espo

nse,

V

transient time, ns

Filter75GHz

Filter90GHz

Filter110GHz

Filter130GHz

Filter155GHz

Filter250GHz

Filter400GHz

Monochromatic emission waveforms (75-400 GHz): Fig. 23.4 Monochromatic waveforms measured form the emitter prototype using QOD and a set of spectral fi lters


182

200 mm

X × 2, mm

Y , mm

attenuation, a.u.

X × 2 , mm

Y , mm

propagation delay, ns (the total range ~30 ps)

PHOTO

ATTENUATION

DELAY

0,0 0,5 1,0 1,5 2,0 2,5 3,00,0

0,2

0,4

0,6

0,8

1,0

1,2

curr

ent,

Aam

plifi

ed (

30 D

B)

dete

ctor

res

pose

, V

transient time, ns

AREA

PEAK

LEVEL (e.g. 15%)

PLATEAU

DELAY

350

350

50

50

40

40

30

30

20

20

10

10

0

0

0

−0.08

0.190.2

0.210.22

−0.06−0.04−0.02

300

300

250

250

200

200

150

150

100

100

50

50

0

0

Fig. 23.5 ( a ) parameters of the detected pulse used for imaging. ( b ) photo and two sub-THz transmission images of a comb. The upper is a “traditional” image utilizing the pulse amplitude (“peak”) in which blue level corresponds to completely transmitted, and red level to completely attenuated pulse (images utilizing “plateau” or “area” are similar). Total signal attenuation between the comb teeth and at the edges of the comb ( red level) is caused by the diffraction and creates diffi culties in the object interpretation. More clear and with better spatial resolution is the lower image , which utilizes the propagation delay. The difference between minimal ( blue ) and maximal ( red ) delay corresponds to ~25 ps, and from the picture contrast a temporal resolution of a few picoseconds can be assumed


183

60

0.16

PROPAGATION DELAY (~0-50ps)

ATTENUATION:

0.14

-1

-0.5

0

60

60

60

0.135

0.14

0.145

0.15

0.155

0.16

0.165

0.17

40

40

40

40

20

20

20

20

0 0

0 0

Fig. 23.6 Two photos, X-ray transmission image, and two sub-THz transmission images of a tap. The upper sub-THz image, together with attenuation of the radiation in thicker parts of the tap, displays an attenuation ( red ring ) near the edge of the object. This is caused by the diffraction and creates erroneous idea on the tap shape if it is hidden in a box, or is a hidden part of more complicated object. The lower image utilizing the propagation delay is free of this defect of the attenuation imaging. Total delay range between blue and red levels is about 25 ps


184

the peak power from a single emitter towards multi-milliwatt range, shift in the spectral peak from ~100 towards ~500 GHz, and very competitive manufacturing price of only a few € for a single emitter chip in serial production. Together with very high effi ciency of the electrical power conversion into sub-THz emission (~0.2 % or higher) this makes the source very promising for cost-effective compact systems utilizing 2-D arrays of novel collapsing-domain-based emitters and 2-D arrays of Schottky detectors for real-time 3-D imaging.

Acknowledgments We are thankful to TEKES selected this activity for support as a strategic project 354/31/2013 to Finnish Academy for the fi nancial support, to MNTC of University of Oulu for the engineering and Dr. Guoyong Duan for the software support of the experimental part of imaging tests.

propagation delay, max. 20 ps

attenuation, a.u.

hidden chalk powder

Fig. 23.7 “Hidden narcotics”: part of tobacco in the central cigarette is replaced with a “narcotic” (chalk powder). X-ray does not give any contrast at all, and even transmission imaging ( upper picture ) can recognize only boundaries of the objects due to diffraction effects, while the propaga-tion delay ( lower image ) detects “the drug” with the contrast corresponding to ~10 ps difference in the propagation delay


185

proparagion, ns (total range 25 ps)

soap

attenuation

140

-1

-0.5

0

120 100 80 60 40 20 0

140

1.721.71

80

60

4

2

1.715

1.72

1.725

1.73

1.735

1.74

1.74

120 100 80 60 40 20 0

80

60

40

20

0

Fig. 23.8 Plastic explosives have chemical composition and THz absorption spectra somewhat similar to sugar and soap. In the photo ( up / left ) there is a cell from the organic glass fi lled with granulated sugar in which several pieces of lump sugar and one bar of soap are hidden. X-ray image ( up / right ) hardly recognizes the lump sugar, and does not show the soap at all. In the attenu-ation sub-THz imaging the boundaries are well resolved thanks to the diffraction, but the most correct impression on the object is provided by the propagation delay image ( the lowest ). Moreover, in very homogeneously looking soap bar the propagation delay image shows certain structure with the delay difference ( yellow / dark red ) of about 10 ps


186

References

1. Duncan W et al (2008) An optical system for body imaging from a distance using near- terahertz frequencies. J Low Temp Phys 151:777–783

2. Cooper K et al (2008) Penetrating 3-D imaging at 4- and 25-m range using a submillimeter- wave radar. IEEE Trans Microw Theory Tech 56:2771–2778

3. Zandonella C (2003) Terahertz imaging: T-ray specs. Nature 424:721–722 4. Appleby R (2007) Standoff detection of weapons and contraband in the 100 GHz to 1 THz

region. IEEE Trans Anten Propag 55:2944–2956 5. Siegel P (2004) Terahertz technology in biology and medicine. IEEE Trans Microw Theory

Tech 52:2438–2447 6. Tonouchi M (2007) Cutting-edge terahertz technology. Nat Photonics 1:97–105 7. Wolbarst A, Hendee W (2006) Evolving and experimental technologies in medical imaging.

Radiology 238:16–39 8. Woolard D et al (2005) Terahertz frequency sensing and imaging: a time of reckoning future

application? Proc IEEE 93:1722–1743 9. Li X et al (2008) Terahertz absorption of DNA decamer duplex. J Phys Chem A 112:

12090–12096 10. Yu C, Fan S, Sun Y, Pickwell-Macpherson E (2012) The potential of terahertz imaging for

cancer diagnosis: a review of investigations to date. Quant Imaging Med Surg 2:33–45 11. Woodward R et al (2003) Terahertz pulse imaging of ex vivo basal cell carcinoma. J Invest

Dermatol 120(1):72–78 12. Fitzgerald AJ et al (2006) Terahertz pulsed imaging of human breast tumors. Radiology

239:533–540 13. Joseph C, Patel R, Neel V, Giles R, Yaroslavsky A (2012) Imaging of ex vivo nonmelanoma

skin cancers in the optical and terahertz spectral regions. J Biophotonics 3:1–10 14. Miller SL, Ebers JJ (1955) Alloyed junction avalanche transistors. Bell Syst Technical J

34(5):883–902 15. Hower PL, Reddi VGK (1970) Avalanche injection and second breakdown in transistors. IEEE

Trans Electron Device 17(4):320–335 16. Vainshtein SN, Yuferev VS, Kostamovaara JT (2002) Properties of the transient of avalanche

transistor switching at extreme current densities. IEEE Trans Electron Device 49(1):142–149 17. Vainshtein SN, Yuferev VS, Kostamovaara JT (2003) Avalanche transistor operation at

extreme currents: physical reasons for low residual voltages. Solid State Electron 47(8):1255–1263

18. Vainshtein S, Yuferev V, Kostamovaara J (2003) Nondestructive current localization upon high-current nanosecond switching of an avalanche transistor. IEEE Trans Electron Device 50(9):1988–1990

19. Levinshtein M, Kostamovaara J, Vainshtein S (2005) Breakdown phenomena in semiconduc-tors and semiconductor devices, Selected topics in electronics and systems 36. World Scientifi c Publishing Co. Pte. Ltd., Singapore

20. Duan G, Vainshtein S, Kostamovaara J (2008) Lateral current confi nement determines silicon ava-lanche transistor operation in short-pulsing mode. IEEE Trans Electron Device 55(5):1229–1236

21. Duan G, Vainshtein S, Kostamovaara J (2012) Three-dimensional peculiarities in an avalanche transistor provide a broadened range of amplitudes and durations of the generated pulses. Appl Phys Lett 101:173506

22. Duan G, Vainshtein S, Kostamovaara J (2012) Turn-on spread determines the size of the switching region in an avalanche transistor. Appl Phys Lett 100:193505

23. Vainshtein S, Yuferev V, Kostamovaara J (2005) Analyses of the picosecond range transient in a high-power switch based on a bipolar GaAs transistor structure. IEEE Trans Electron Device 52(12):2760–2768

24. Vainshtein S, Yuferev V, Kostamovaara J (2005) Ultra-high fi eld multiple Gunn domains as the physical reason for superfast (picosecond range) switching of a bipolar GaAs transistor. J Appl Phys 97(2):024502/1–024502/9


187

25. Vainshtein S, Kostamovaara J, Yuferev V, Knap W, Fatimy A, Diakonova N (2007) Terahertz emission from collapsing fi eld domains during switching of a gallium arsenide bipolar transistor. Phys Rev Lett 99(17):176601

26. Vainshtein S, Kostamovaara J, Sveshnikov Y, Gurevich S, Kulagina M, Yuferev V, Shestak L, Sverdlov M (2004) Superfast high-current switching of a GaAs avalanche transistor. Electron Lett 40(1):85–86

27. Vainshtein S, Yuferev V, Kostamovaara J, Kulagina M, Moilanen H (2010) Signifi cant effect of emitter area on the effi ciency, stability and reliability of picosecond switching in a GaAs bipolar transistor structure. IEEE Trans Electron Device 57(4):733–741

28. Vainshtein SN, Kilpela A, Kostamovaara J, Myllyla R, Starobinets S, Zhilyaev J (1994) Multistreamer regime of GaAs thyristor switching. IEEE Trans Electron Device ED-41(8):1444–1450

29. Gunn JB (1963) Microwave oscillation of current in III-V semiconductors. Solid State Commun 1:88

30. Kroemer H (1964) Theory of the Gunn effect. Proc IEEE 52(12):1736–1736 31. Vainshtein S, Yuferev V, Palankovski V, Ong DS, Kostamovaara J (2008) Negative differential

mobility in GaAs at ultra-high fi elds: comparison between an experiment and simulations. Appl Phys Lett 92(6):062114

32. Schmielau T, Pereira MF Jr (2009) Nonequilibrium many body theory for quantum transport in terahertz quantum cascade lasers. Appl Phys Lett 95:231111

33. Pereira MF Jr (2008) Intervalence transverse-electric mode terahertz lasing without population inversion. Phys Rev B 78(24):245305-1–245305-5



Abstract In the present paper we report results of explosives vapor composition studying with the THz spectrometers based on CSR effects. Identifi ed gases (NO, N 2 O, NH 3 , acetone) hold a promise to be “markers” of explosives. We also present a prototype of compact explosives detector based on the THz spectrometer.

Keywords THz spectroscopy • Explosives detector • Gas-markers • Superlattice multiplier

24.1 Introduction

The general threat of terrorist attack, especially in public places, has stimulated extensive researches and development of technology for rapid detection of explosives that yielded several commercial handheld and stand-alone devices. The majority of technique used for remote and local detection employs registration of explosives vapors accumulated near surfaces and then diffused in ambient air. However the saturated vapor pressure of many explosives is extremely low (see Table 24.1 ) and requires high sensitivity of a detector (ppt-ppb).

Commonly used technique employs ion mobility spectrometry, gas chromatography and mass spectrometry.

The ion mobility spectrometry (IMS) is the cornerstone of many handheld and stand-alone detection devices used today [ 1 , 2 ]. This technology involves drawing a gaseous sample into a reaction chamber using an air pump. Then the gas molecules are ionized and passed through a weak electrical fi eld toward an ion detector.

Chapter 24 Sub-THz Spectroscopy for Security Related Gas Detection

V. Vaks , E. Domracheva , E. Sobakinskaya , and M. Chernyaeva

V. Vaks (*) • E. Domracheva • E. Sobakinskaya • M. Chernyaeva Department of Terahertz Spectrometry , IPM RAS, Lobachevsky State University of Nizhny Novgorod , Nizhny Novgorod , Russia e-mail: [email protected]

190

Gases are identifi ed according to the time it takes to traverse the distance to the detector and the amount of electrical charge detected. The world market of explosives detectors offers also other IMS technique: handheld detectors (M86 and M90, APD 2000, JUNO, VaporTracer, QS-H150) and stand-alone detectors (M8A1, Centurion II, IONSCAN, ITEMISER). General advantages of the IMS detectors include quite high sensitivity (usually 10 −13 g/cm 3 ), on-line analysis and easy-to-use exploitation. Although presence of different interferences in real world conditions and uncontrolled ionization process reduce selectivity and result in frequent false alarm. Besides the IMS devices prone to infl uence of ambient conditions: humidity, dustiness and temperature changes.

The gas chromatography (GS) is based on a physical separation of gas mixtures components [ 3 , 4 ]. The separation is obtained after a mixture (with carrier gas) had passed through a sorbent in chromatographic column. Dependence of mixture components velocities on adsorption-desorption process results in different arrival times which are used for gas identifi cation. The commercial GS detectors are widespread and realized in several models: EGIS 3000, EVD-8000, VIXEN, EKHO, E-5000, VIPER, Edelveis and Eho-M. Apparent advantages of gas chromatographs include high sensitivity, wide range of detectable substances, self-descriptiveness. The reason for skeptical opinion about “in fi eld” applications of the GS involves a need of a balloon with high purity gas carrier (nitrogen or argon), that makes the device operation dependent on presence of the gas carrier.

Explosive detectors based on mass-spectrometry (MS) are highly sensitive yet are not widely used in inspection process. The reason is operation complexity that requires high-qualifi ed user, moreover MS device are quite expensive.

To conclude, despite the world market of explosive detectors offers a variety of technique, the problem of explosive detection is not yet solved. The reason is that none of the present commercial devices meets all the requirements of an optimal detector for “in fi eld” use: compact, easy-to-use, fast, high selective and reliable. The efforts to meet the above requirements are focused on both improvement of the present explosives detectors [ 5 , 6 ] and development of the new ones [ 7 , 8 ]. The later involves, fi rst of all, application of spectroscopic methods in the IR, microwave, and THz ranges.

The specialists, working in the IR range, rest their hopes upon laser spectroscopy, that stimulated by proliferation of cheap commercial lasers with high power and frequency tuning [ 7 , 8 ]. Currently, the laser spectroscopy is used mainly to track

Table 24.1 Saturated vapor concentration of some explosives

Explosives Saturated vapor concentration, ppt, 293 K (substance volume fraction on 10 12 parts of air)

Saturated vapor concentration, ppt, 400 K

EGDN 1·10 8 Decomposition at 387 K NG 5,8·10 5 5,1·10 9 DNT 5,6·10 4 – TNT 9,4·10 3 4,7·10 8 PETN 18 2,8·10 7 RDX 6 2,1·10 6

V. Vaks et al.

191

chemical agents’ clouds that already have been identifi ed. In the technique, two close frequencies are used. One frequency is selected to match a peak absorption line of chemical agent of interest, while the second nearby frequency is selected to a region of low absorption for the same agent. Comparison of the return signals at these two frequencies gives a direct measure of the concentration of the gas, while the time of return is used to determine the distance from the observers. In general, the received return signal is contaminated by various types of noises from atmospheric background that leads to poor signal to noise ratio and requires special statistical averaging technique.

Currently, the only spectroscopic method which provides both the best spectral resolution (i.e. selectivity) and sensitivity which approaches the theoretical limit is THz spectroscopy based on the effect of coherent spontaneous radiation (CSR) of gas molecules. In brief, the CSR phenomenon can be outlined in the following way. Interaction of electromagnetic radiation with absorbing molecules results in induction of macroscopic dipole in the gas. After the radiation impact is over, absorbed energy is coherently reradiated at the frequency of molecular transition. Among advantages of the technique based on CSR there are, fi rst of all, high spectral resolution which is ensured by high stable radiation sources, and high sensitivity. The spectrometers also provide real-time measurements and opportunity of in situ studies. The above characteristics facilitate application of this technique for various problems [ 9 – 11 ], including security systems.

In the present paper we report results of explosives vapour composition studying with the THz spectrometers based on CSR effects.

24.2 Experimental Set Up

For studying the subTHz spectra of the explosives vapor we used a spectrometer with phase manipulation of the probe radiation [ 9 , 12 , 13 ] (Fig. 24.1 ).

The main blocks of the spectrometer involve:

– radiation source (a synthesizer of the THz range frequency that includes a backward wave oscillator (BWO), BWO power supply; BWO phase locked loop (PLL) on a reference frequency synthesizer; a device that provides phase manipulation of the probe radiation);

– measuring tract consisting of a gas cell and a pressure meter; – receiver, which includes a detector on quantum semiconductor superlattices

(SL), a preamplifi er, and a digital storage; – control unit.

The PLL system controls the BWO frequency and phase shift. The spectrometer can work both in the time domain and frequency domain regimes. Its sensitivity is close to the theoretical limit (at l = 1 m, T meas = 1 with a sensitivity of 5 × 10 −10 cm −1 ). The accuracy of measuring absorption line intensity (without preliminary calibration) is ≤5 %.

24 Sub-THz Spectroscopy for Security Related Gas Detection

192

One of the most important problems in adaption of the spectrometer for wide range measurements involves incorporating various BWOs in the spectrometer’s radiation source without changing the setup’s confi guration. This problem was solved by using harmonic mixers and multipliers based on superlattice structures (SL) instead of the traditionally used instruments based on Schottky diodes. The set of frequency synthesizers with 37–55 harmonics is designed [ 11 ] that enables us to move easily from the microwave range (117–178 GHz) to the THz range.

For studying explosives vapor composition two types of gas cells are developed: multimode waveguide (cross-section, 10 × 23 mm; length, 1 m) with fl uoroplast windows and gas cell made of a quartz tube ( d = 25 mm; length, 1 m). Results of the tests measurements demonstrate that the metal cell provides better sensitivity, but it is not convenient for work, since a noise signal originating from electrical interference is diffi cult to suppress. Thus, the quartz cell is used in all measurements. Apart from lower noise signal the latter provides chemical purity and inertness to aggressive media. The presence of water vapor with a rich and intense spectrum in the above ranges could reduce the sensitivity of the method and affect the results of the measurements. To reduce the concentration of water vapor, the cells are heated (to 470 K) for 48 h at constant evacuation. As a result, the concentration of water vapor is decreased by 100 times, to 10 −2 mol %.

High explosives (HE) samples (several tens of milligrams) are put in a small glass container joined to the gas cell by a vacuum valve. The cell and container are independently evacuated to a pressure 10 −4 mmHg. The container valve is then opened, and the HE vapor fi lls the cell up to 10 −1 mmHg. The gas cell is evacuated again to achieve 5 × 10 −2 mmHg, and this pressure is kept during all our measurements.

Coup.BWO Att.

Harmonicalmixer

Powersupply

Short pulseformer

PhaseLock/Switch

Control

ClockGenerator

ADC1 ADC2

Preamplifier

Low-passfilter

Digitalaverager

Dataacquisition and

averagingsystem

Referencefrequencysynthesizer

Fig. 24.1 The scheme of the THz spectrometer

V. Vaks et al.

193

The study of RDX and HMX requires heating of the container, since their saturated vapors pressure is much lower than that of TNT or NG. Heating to 340–360 K is performed using a magnetron at a frequency of 2.4 GHz and radiation power of no more than 100 W.

The fi rst stage of our studies involves spectral measurements in the range 117–178 GHz, because the spectrometer’s sensitivity is on maximum in this range. Moreover the microwave spectra database includes resonance frequencies of many substances that simplify identifi cation of gases. The studies are then continued in the THz range (up to 714 GHz), where the absorption lines of many materials are of maximum intensity.

24.3 Results

The structural formulae of HEs (TNT, NG, HMX and RDX), used in our studies, are presented in Fig. 24.2 .

Fig. 24.2 The structural formulae of HEs


194

The results of the experiments show that the vapors of all HEs contain nitrogen oxides NO and N 2 O. For some HEs, we also detect CO, NH 3 , and acetone. The absorption lines of CH 2 (OH)CHO and NO in the vapors of cyclotetramethilenetet-ranitramine (HMX) are presented in Fig. 24.3 .

In our opinion, the nitrogen oxides NO and N 2 O are not primary products of the HE natural decay. One of the primary products can be oxide NO 2 (see Fig. 24.3 ) [ 9 ]. Being a strong oxidizer, NO 2 can react with the HE components and admixture substances. As a result, less chemically active products of these reactions (NO, N 2 O, and NH 3 ) are observed in the HE vapors. Presence of aldehydes in HMX vapors can be attributed to HMX natural decomposition.

Being the vapor components of all HEs, nitrogen oxides NO and N 2 O can be considered indicators of explosives. However, these gases are also natural com-ponents of the atmosphere. Thus, a care should be taken when measuring con-centration of the NO and N 2 O. To ensure a reliable detection, an explosives detector must also provide registration of other gas-markers, which are normally absent in atmosphere. For our explosives detector prototype we have chosen NO, NH 3 , acetone.

The prototype is based on the spectrometer with phase switching mode and operates at the most intensive absorption lines of the chosen gas-markers. In the radiation source of the spectrometer [ 11 , 14 ] we use the Gunn-diode generator (97.5–117.5 GHz) together with the SL multiplier (Fig. 24.4 ) that provides a working range 0.5–2.5 THz. The main problem in the construction of the radiation source is to obtain phase-switching in the interval (0, π ) for an arbitrary harmonic number. That is why the cornerstone of the spectrometers design involves a system of phase stabilization of the oscillator frequency in phase switching mode for arbitrary harmonic number. Phase manipulation of the Gunn generator signal maximizes sensitivity of the method and is achieved by supplying ultra short impulses to feed the Gunn generator’s circuit. The impulses are obtained by differentiation of the modulation signal which is supplied as meander to primary winding of pulse trans-former through resistor.

Fig. 24.3 ( a ) The record of absorption line of CH 2 (OH)CHO at the frequency of 118,196 GHz; ( b ) The record of absorption line of NO at the frequency of 150,176 GHz

V. Vaks et al.

195

The prototype is able to distinguish absorption lines of NO, NH 3 , acetone in atmosphere. Its sensitivity is demonstrated to be hundreds ppb. The work is under way to increase the sensitivity and enlarge the gas-markers set to ensure reliable detection of explosives.

24.4 Conclusion

Application of THz spectrometers based on CSR effects for studying of explosives vapor compositions has resulted in registration of several gases, which can be used as gas-markers for remote explosives detection. Obtained information about vapor composition is used for development of compact explosives detector prototype. The prototype is based on the THz spectrometer on CSR effects, thus, providing high spectral resolution, on-line measurements and sensitivity about hundreds ppb. The work is under way to improve sensitivity of the prototype, so, it can be used for real life applications.

Acknowledgments This work is supported by Teradec 047.018.005, NATO.EAP.SFPP 984068, MegaGrant 11.G34.31.0066.

References

1. Roscioli KM, Davis E, Siems WF, Mariano A, Su W, Guharay SK, Hill HH (2011) Modular ion mobility spectrometer for explosives detection using corona ionization. Anal Chem 83(15):5965–5971. doi: 10.1021/ac200945k

2. Buryakov A (2011) Detection of explosives by ion mobility spectrometry. J Anal Chem 66(8):674–694. doi: 10.1134/S1061934811080077

Powersupply

PLL

HEMTamplifier

Harmonicmixer

DigitalLock in

Amplifier

Evacuated Opticswith Lens

Data Acquisition Unit

Detector(self-

heterodyne)

Harmonicmixer

Gunngenerator97.5-117.5

GHz

Referencesynthesizer8-12.5 GHz

Generator Harmonicson SL structures

PLL Unit

Fig. 24.4 The block-diagram of the explosives detector prototype


http://dx.doi.org/10.1021/ac200945k

http://dx.doi.org/10.1134/S1061934811080077

196

3. Collin OL, Niegel C, DeRhodes KE, McCord BR, Jackson GP (2006) Fast gas chromatography of explosive compounds using a pulsed-discharge electron capture detector. J Forensic Sci 51(4):815–818. doi: 10.1111/j.1556-4029.2006.00171.x

4. Bielecki Z, Janucki J, Kawalec A, Mikołajczyk J, Pałka N, Pasternak M, Pustelny T, Stacewicz T, Wojtas J (2012) Sensors and systems for the detection of explosive devices – an overview. Metrol Meas Syst 19(1):3–28. doi: 10.2478/v10178-012-0001-3

5. http://www.lifesafetysys.com/osb/itemdetails.cfm/ID/1374 6. http://www.lifesafetysys.com/osb/itemdetails.cfm/ID/1376 7. Hildenbrand J, Herbst J, Wöllenstein J, Lambrecht A (2009) Explosive detection using

infrared laser spectroscopy. Quantum sensing and nanophotonic devices, VI. In: Razeghi M, Sudharsanan R, Brown GJ (eds) Proceedings of SPIE, vol 7222, 72220B, San Jose, pp 1–12. doi: 10.1117/12.808976

8. Munson C, Gottfried J, De Lucia F, McNesby K, Miziolek A (2007) Laser-based detection methods of explosives. Elsevier, Amsterdam

9. Nabiev SS, Nadezhdinskii AI, Stavrovskii DB, Vaks VL, Domracheva EG, Pripolzin SI, Sobakinskaya EA, Chernyaeva MB (2011) Analysis of the products of the natural decay of high explosives by subterahertz and infrared Fourier spectroscopy. Russ J Phys Chem A 85(8):1404–1410

10. Vaks VL, Domracheva EG, Nikiforov SD, Sobakinskaya EA, Chernyaeva MB (2008) Application of microwave nonstationary spectroscopy for noninvasive medical diagnostics. Radiophys Quant Electron 51(6):493–498. doi: 10.1007/s11141-008-9049-z

11. Vaks V (2012) High-precise spectrometry of the terahertz frequency range: the methods, approaches and applications. J Infrared MillimTerahertz Waves 33:43–53. doi: 10.1007/s10762-011-9846-x

12. Vaks VL, Khodos VV, Spivak EV (1999) A nonstationary microwave spectrometer. Rev Sci Instrum 70(8):3447–3453. doi: 10.1063/1.1149935

13. Vaks VL, Brailovsky AB, Khodos VV (1999) Millimeter range spectrometer with phase switching – novel method for reaching of the top sensitivity. Infrared Millim Waves 20(5):883–896. doi: 10.1023/A:1021774217938

14. Vaks VL, Panin AN, Basov SA, Illyuk AV, Pripolzin SI, Paveliev DG, Koshurinov YI (2009) Nonstationary spectroscopy of the 1–2.5 THz frequency band with the use of solid-state devices. Radiophys Quant Electron 52:511–517. doi: 10.1007/s11141-009-9154-7

V. Vaks et al.

http://dx.doi.org/10.1111/j.1556-4029.2006.00171.x

http://dx.doi.org/10.2478/v10178-012-0001-3

http://www.lifesafetysys.com/osb/itemdetails.cfm/ID/1374

http://www.lifesafetysys.com/osb/itemdetails.cfm/ID/1376

http://dx.doi.org/10.1117/12.808976

http://dx.doi.org/10.1007/s11141-008-9049-z

http://dx.doi.org/10.1007/s10762-011-9846-x

http://dx.doi.org/10.1007/s10762-011-9846-x

http://dx.doi.org/10.1063/1.1149935

http://dx.doi.org/10.1023/A:1021774217938

http://dx.doi.org/10.1007/s11141-009-9154-7

[nato science for peace and security series b: physics and biophysics] terahertz and mid infrared...

Documents