dna detection using organic thin film transistors:...

DNA Detection Using Organic Thin Film Transistors:Physical Origins of Electrical Transduction Behavior

and Optimization of Sensitivity

Lakshmi JagannathanVivek Subramanian

Electrical Engineering and Computer SciencesUniversity of California at Berkeley

Technical Report No. UCB/EECS-2009-11

http://www.eecs.berkeley.edu/Pubs/TechRpts/2009/EECS-2009-11.html

January 22, 2009

Copyright 2009, by the author(s).All rights reserved.

Permission to make digital or hard copies of all or part of this work forpersonal or classroom use is granted without fee provided that copies arenot made or distributed for profit or commercial advantage and that copiesbear this notice and the full citation on the first page. To copy otherwise, torepublish, to post on servers or to redistribute to lists, requires prior specificpermission.

1

DNA Detection Using Organic Thin Film Transistors:

Physical Origins of Electrical Transduction Behavior

And Optimization of Sensitivity

by Lakshmi Jagannathan

Submitted to the Department of Electrical Engineering and Computer Sciences, University of

California at Berkeley, in partial satisfaction of the requirements for the degree of Master of

Science, Plan II.

Approval for the Report and Comprehensive Examination:

Committee:

Professor Vivek Subramanian

Research Advisor

Date:

* * * * * * *

Professor Ali Javey

Second Reader

Date:

2

Acknowledgements I am indebted to many people for directly and indirectly helping me with this thesis. I want to

thank my advisor, Vivek Subramanian, for his advice and support throughout. He has been a great

role model, and I have learned a lot working with him for the past few years. I would also like to

thank our entire EECS Organic Electronics group at Berkeley, for their advice, encouragement,

knowledge transfer, immediate willingness to help, and in general, just making me love to come to

the office everyday . I am grateful to Dr. Qintao Zhang, a graduate from our group, for the

knowledge transfer of the biosensor system, and his willingness to help to this day. I would also like

to thank Dr. Kanan Puntambekar, a postdoc graduate from our group, for helping me better

understand and explain different aspects of my project. I want to thank Daniel Forchheimer for his

help with the DNA microfluidics experiment, and his contribution of valuable ideas to my project. I

would like to express thanks to Professor Ali Javey for being a second reader for my thesis, and for

his encouragement during my first years in graduate school at Berkeley. I want to also acknowledge

Dr. Steve Ruzin and the staff at the College of Natural Resources, for their training and help with the

fluorescent microscopy system. I want to thank Ruth Gjerde and Pat Hernan at Berkeley for their

constant encouragement; they always believed in me more than I did! I would like to thank

Semiconductor Research Corporation (SRC) and Intel for sponsoring and funding me for graduate

school. I am grateful to Karen, Soundarya, Bharath, Kristen, Sridevi, Dhevi, and Shalini, for their

constant support and encouragement, and in particular for proofreading this document. I want to

also thank my entire youth group for their encouragement and help with everything. I would like to

thank all of my friends for being there for me when I needed them. Finally, I want to express my

heartfelt gratitude to my spiritual guru and guide, Sri Sathya Sai Baba, my entire family in Dallas,

and my family in San Jose for their love and incredible support, without which this truly would have

not been possible!

3

Table of Contents

Acknowledgements ......................................................................................................................... 2

Table of Contents ............................................................................................................................ 3

List of Figures ................................................................................................................................. 4

List of Tables .................................................................................................................................. 5

I. Introduction ............................................................................................................................... 6

A. Abstract .................................................................................................................................. 6

B. Motivation .............................................................................................................................. 6

C. Pentacene Organic Thin Film Transistors for Biosensing ................................................... 10

II. DNA Immobilization and Doping........................................................................................... 13

A. DNA Immobilization on Pentacene Film ............................................................................ 13

B. DNA and Pentacene Interaction .......................................................................................... 14

III. Investigation of Physical Origins of Electrical Transduction Behavior .................................. 19

A. Fluorescent Microscopy ....................................................................................................... 19

B. Time-of-Flight Secondary Ion Mass Spectroscopy (TOF-SIMS) ........................................ 23

i. Definition and Technical Capabilities .............................................................................. 23

ii. Experiment, Procedure, and Results ............................................................................... 25

IV. Optimization of DNA Immobilization and Sensor Sensitivity ............................................... 30

A. Pentacene Characterization Experiment .............................................................................. 30

i. Experimental Setup: Response Surface Design of Experiments (DOE) ........................... 31

ii. Experiment Protocol ......................................................................................................... 33

ii. Results ............................................................................................................................... 34

B. Optimization of DNA Immobilization and Sensor Sensitivity ............................................ 51

i. Experiment Protocol ......................................................................................................... 51

ii. Control Experiments ........................................................................................................ 52

iii. Results: Optimization of DNA Immobilization and Sensor Sensitivity........................... 53

V. Conclusion and Future Work .................................................................................................. 57

VI. Sources ................................................................................................................................... 58

APPENDIX ................................................................................................................................... 60

4

List of Figures

Figure 1. DNA is a building block for everything within a human being. ..................................... 7

Figure 2. Ideal DNA Detection System. ........................................................................................ 7

Figure 3. Process flow for formation of printed FETs. .................................................................. 9

Figure 4. Pentacene TFT Structure. ............................................................................................. 10

Figure 5a. Double stranded DNA chemical structure ......................................................................

Figure 5b. Pentacene (5-benzene ring) .......................................................................................... 14

Figure 6. Current-Voltage characteristics from an experiment show a positive Vt shift and a

higher hole current after DNA immobilization. ............................................................................ 15

Figure 7. Surface potential measurement of DNA molecules on a pentacene film. ..................... 16

Figure 8. Representative SCLC curves of original pentacene-based transistors and DNA ......... 18

Figure 9. Fluorescent Microscope System.. ................................................................................. 20

Figure 10. Axioimager M1 Fluorescent Microscope Set-up at the College of Natural Resources

(CNR) in Berkeley. ....................................................................................................................... 20

Figure 11. Fluorescent density images captured using the digital CCD camera in the fluorescent

microscope system. ....................................................................................................................... 21

Figure 12. Central Composite Design Methodology. .................................................................. 31

Figure 13. A set of 9 points was measured for each of the 17 experiments in the DOE to study

variation within the wafer and wafer-to-wafer variation. ............................................................. 33

Figure 14. Pentacene TFT structure. ............................................................................................ 34

Figure 15. Idsat (electrical characteristics) for all 17 experiments. ............................................... 35

Figure 16. Explanation of the t-test mean diamonds used for statistical purposes. ..................... 35

Figure 17. Leverage Plots relating input and output parameters using the measurements

performed.. .................................................................................................................................... 37

Figure 18. Relationship between the input parameters and Idsat. .................................................. 38

Figure 19. Schematic diagram depicting the phases and grain orientation in various pentacene

film. ............................................................................................................................................... 38

Figure 20. Interaction Profiler Plot for Idsat. ................................................................................. 40

Figure 22. Id- Vd characteristics from 2 different experiments with the lowest and highest

performing samples. ...................................................................................................................... 42

Figure 23. Relationship between the input parameters and mean roughness. ............................. 42

Figure 24. Interaction Profiler Plot for Mean Roughness. ........................................................... 44

Figure 25. AFM Images of samples from 2 different experiments with the lowest and highest

surface roughness parameters. ..................................................................................................... 45

Figure 26. Relationship between the input parameters and grain size. ........................................ 46

Figure 27. Interaction Profiler Plot for Grain Size. ...................................................................... 48

Figure 28. AFM images of samples from 2 different experiments consisting of the smallest and

biggest grain sizes ......................................................................................................................... 48

Figure 29. Relationship between the input parameters and evaporation rate and coverage area. 49

Figure 31. Control Experiments performed to neutralize buffer solution and DI water effects on

pentacene TFTs. ............................................................................................................................ 52

Figure 32. Relationship between input parameters and Idsat ratios. .............................................. 53

Figure 33. AFM Image of terrace-like formation of pentacene thin film on SiO2 substrate. ....... 54

Figure 35. Mass Spectra of positive ions of interest for samples 1 to 3; Mass to Charge Ratio: 0-

200................................................................................................................................................. 60

5

Figure 36. Mass Spectra of positive ions of interest for samples 1 to 3; Mass to Charge Ratio:

100-300 ......................................................................................................................................... 61

Figure 37. Mass Spectra of positive ions of interest for samples 1 to 3; Mass to Charge Ratio:

300-700 ......................................................................................................................................... 62

Figure 38. Mass Spectra of negative ions of interest for samples 1 to 3; Mass to Charge Ratio: 0-

100................................................................................................................................................. 63

Figure 39. Mass Spectra of negative ions of interest for samples 1 to 3; Mass to Charge Ratio: 0-

140................................................................................................................................................. 64

Figure 40. Mass Spectra of negative ions of interest for samples 1 to 3; Mass to Charge Ratio:

100-400 ......................................................................................................................................... 65

List of Tables

Table 1. Normalized positive ions of interest .............................................................................. 27

Table 2. Normalized negative ions of interest ............................................................................. 27

Table 3. Input and Outputs for Pentacene Characterization Experiment. .................................... 30

Table 4. JMP Surface Response DOE for Pentacene Characterization Experiment. .................. 32

6

I. Introduction

A. Abstract

Advancements in research and technology are happening now more than ever in the

biotechnology industry. Better detection and treatment methods are constantly being sought by

researchers. Collaboration among different fields of science and technology has brought us one

step closer towards achieving this goal. The vision of achieving a personalized system of disease

detection and treatment has triggered the ultimate purpose of my project: to be able to detect

genetic diseases using electrical means of sensing and detection. Previous work has shown the

potential of organic (pentacene) thin film transistors (OTFT) for DNA detection by showing

different electrical performance shifts in response to single and double stranded DNA[5]. The

goal of this thesis is to present two aspects of using OTFTs for genetic disease detection, namely

the physical origins of the observed electrical shifts, and the characterization of the pentacene

surface to allow optimization of the same for DNA immobilization and sensor sensitivity.

B. Motivation

Using OTFTs for DNA (Deoxyribonucleic acid) microarray technology paves the path for an

ideal DNA detection system. DNA detection is important to detect mutations that cause genetic

diseases. A genetic disorder is a disease that is caused by an abnormality in an individual's DNA.

Abnormalities can range from a small mutation in a single gene to the addition or subtraction of

an entire chromosome or set of chromosomes.

7

Figure 1. DNA is a building block for everything within a human being [1].

An ideal detection system would consist of an ultra low-cost, disposable DNA detection chip,

which can be analyzed with a handheld device (e.g. PDA), using a quick and easy testing

protocol. Moreover, the results/feedback from the chip would be available within a few hours.

Figure 2. An Ideal DNA Detection System. (Image: Courtesy of Dr.Qintao Zhang, Graduate from EECS Organics Electronics Group, UC Berkeley)

Gene: Portion of an organism‟s DNA with coding and non-coding

sequences.

Chromosome: formed from a single DNA molecule that contains many genes.

8

Use of organic transistors facilitates the possibility of an ultra low-cost, disposable DNA chip

through printed electronics. Ultra-low cost printed electronics have gained a great deal of

interest over the past few years because of their promise to greatly reduce the cost of many

electronic applications. They seek to reduce the manufacturing cost of electronics with less

expensive, all-additive printing methods that conventional silicon manufacturing cannot

replicate. There are several applications for low-cost printed electronics including radio

frequency identification (RFID) tags, electronic sensors, displays, smart cards, packaging, and

printed circuit boards (PCB). Based on performance requirements for the above applications,

suitable electronic materials for printing are typically examined, characterized, and implemented.

For example, soluble gold nanoparticle ink for metallization, printable organic dielectrics

including polyimide and PVP, and a high performance organic semiconductor such as a

pentacene precursor semiconductor ink have all been used to make a fully printed organic thin

film transistor [2]. All of these materials have plastic substrate compatible activation

temperatures (<200°C).

Printed electronics using organic semiconductors have many advantages over typical silicon

processes. Although silicon processing uses many of the ideas in printing to achieve low costs

and rapid manufacturing, certain limitations make it difficult to reduce costs further. Some of the

high costs associated with silicon processing include energy because of the high thermal budget

and requirement of ultra-clean processes, and the capital and process cost expenditures associated

with pattern transfer or photolithography due to the extra time and material (>90% material

wasted overall) needed by the lithography processing steps [2]. The processing steps for printed

electronics have a low thermal budget, allowing for a variety of low-cost and flexible substrates,

and additive processing, through which material is only deposited where needed, reducing the

overall consumption of materials. Figure 3 depicts the process flow for printed OTFTs.

9

Figure 3. Process flow for formation of printed FETs. (Image: Courtesy of Dr.Vivek Subramanian, UC Berkeley)

Printed OTFTs are particularly suitable for sensor applications because of the ability to print

multiple active layers on the same substrate and the bottom gated structure. The ability to print

multiple layers on the same substrate is very valuable for biosensor applications because it gives

the DNA microarray technology (explained in the following section) another dimension. The

bottom gated structure allows for the channel layer to be exposed. This allows for easy and non-

destructive inclusion of analytes, bio-molecules/proteins, etc., as the final „layer.‟ Moreover, the

open channel layer allows for easy electrical measurements before and after inclusion of the

objects that are to be sensed. DNA sensitivity is measured using the saturation current before

and after the immobilization of DNA. Having this layer open is critical for this step in the

experiment to take place.

In this work, evaporated pentacene transistors were used instead. As we will discuss later in

the thesis, the morphology of the pentacene surface is critical to the immobilization of the DNA

and the sensitivity of the sensor. By evaporating pentacene instead of printing it, we are able to

control this surface structure very precisely. For the purposes of the work covered in this thesis,

the control of this surface was very important for optimization. Therefore, evaporated pentacene

10

transistors were used. Translating from evaporated to printed transistors, in terms of the

fabrication processes and output characteristics, has previously been mastered by our group and

will be implemented in the future work of this project.

C. Pentacene Organic Thin Film Transistors for Biosensing

The pentacene TFT bottom gated structure used throughout this project is shown in Figure 4

below.

Figure 4. Pentacene TFT Structure.

The thickness of the pentacene can be varied as necessary, but for our purposes, it ranges from

10-30nm. The steps for fabricating the transistor and including DNA on the pentacene surface

will be discussed further in chapter IV of the thesis.

Pentacene TFTs are potentially useful for biosensors because of the advantages they provide

over current state of the art fluorescent detection technology. The current DNA microarray/chip

technology involves a meticulous process. For example, to determine whether an individual

possesses a mutation for BRCA1 or BRCA2, genes whose mutations are known to cause as many

as 60% of all cases of hereditary breast and ovarian cancers [3], a scientist first obtains a sample

of DNA from the patient's blood as well as a control sample, one that does not contain a mutation

in either gene. The researcher then denatures the DNA in the samples, a process that separates

100nm SiO2

Source

(100nm)

Drain

(100nm)

Gate (n++ Si)

Pentacene

11

the two complementary strands of DNA into single-stranded molecules. The long strands of

DNA are then cut into smaller, more manageable fragments. Each fragment is then labeled by

attaching a fluorescent dye. The individual's DNA is labeled with green dye and the control

(normal) DNA is labeled with red dye. Both sets of labeled DNA are then inserted into the chip

and allowed to hybridize (bind) to the synthetic BRCA1 or BRCA2 DNA on the chip.

If the individual does not have a mutation for the gene, both the red and green samples will

bind to the sequences on the chip. If the individual does possess a mutation, the individual's

DNA will not bind properly in the region where the mutation is located. The scientist can then

examine this area more closely to confirm that a mutation is present [3].

With our pentacene TFT sensor, we propose to do the same mutation detection electrically.

Instead of using fluorescent laser technology to correlate the extracted fluorescent signals with

mutations, we will detect the mutations by analyzing conductance and threshold voltage shifts.

Our technology has several advantages compared to the current optical detection technology:

1) Higher Sensitivity: As many previous researchers have already shown, electrical detection

is known to be more sensitive than optical detection [4]. If the detector is more sensitive,

needless to say, the diseases can be detected earlier (important for early cancer detection,

for example) and much more accurately.

2) Label free method: The DNA used with our proposed technology does not need to be

tagged with fluorescent molecules. This means that the expensive fluorescent laser

detection system can also be eliminated. The whole process becomes simpler and cheaper.

3) Disposable/Portable/Ultra-low cost: Adding to the previous point, the elimination of the

fluorescent laser detection system makes it cheaper and easily accessible. Also, since

12

OTFTs can potentially be printed on plastic substrates, the disposability and low-cost

factors are further made easier. More importantly, it is impossible to make a disposable

DNA detection chip unless it is done with an electrical or a label-free method.

4) Faster Process: The method of hybridization, which takes 12-24 hours, poses a bottleneck

for the current DNA microarray technology process. Amongst other steps that could

potentially save time in our proposed technology, the possibility of pulse-enhanced

hybridization [5], which could potentially reduce the hybridization time to milliseconds,

makes our proposed electrical detection process much faster.

Having thus motivated and introduced the topic, the next few sections will discuss the sensor

system further. Section II describes the immobilization and doping mechanisms of DNA on

pentacene. Section III discusses the methods that have been used and will be used to correlate

the amount of DNA on the surface with the electrical shift observed. Finally, Section IV will

present the effect of pentacene film formation on DNA detection behavior, and optimization of

the pentacene surface in terms of morphology and topology for highest DNA immobilization, and

more importantly, highest sensitivity for DNA detection.

13

II. DNA Immobilization and Doping

A. DNA Immobilization on Pentacene Film

Previously, it has been shown that pentacene can be used to detect DNA [5]. Through

physical absorptive immobilization, DNA immobilizes on pentacene film. The pentacene film is

hydrophobic, and DNA is diluted in saline-sodium citrate (SSC) buffer solution. The 20x

solution of buffer consists of 3M sodium chloride and 300 mM trisodium citrate. The buffer

solution is further diluted in water, as it is usually done in practice, to give a 2x concentration

solution. The buffer can be considered salt water for analytical purposes.

The hydrophobic interaction immobilization (interactions between ssDNA molecules and

crevices on hydrophobic substrate surface) was a very common method in the early stage of DNA

microarray development and has been studied very well [6]. Using this method, we can explain

the pentacene film interaction with DNA. When the DNA in buffer solution is pipetted onto the

pentacene surface, the DNA segregates to hydrophobic „holes‟ on the surface. During this

process, the salt water, in which the DNA is diluted, gets pushed out, „physically‟ immobilizing

the DNA in these crevices.

Initial hypotheses led us to believe that roughness and grain boundaries in particular

immobilize more DNA, and would give the highest sensitivity. Recent experiments, which will

be presented in later sections of the thesis, have shown that in addition to the morphological

features, the overall sensitivity is more strongly dependent on how much DNA is able to diffuse

to the part of the pentacene film which contains the channel. The channel is contained within the

first few monolayers of pentacene film.

14

B. DNA and Pentacene Interaction

The sensing of DNA is possible because of DNA interaction with pentacene. The DNA and

pentacene chemical structures are shown in Figure 5a and 5b below:

Figure 5a. Double stranded DNA chemical structure [7]. Figure 5b. Pentacene (5-benzene ring)

chemical structure [8].

Pentacene TFTs are p-type accumulation devices (negative gate voltage and negative

current). After DNA immobilization on surface, the hole current of the TFT increases.

Consequently, the threshold voltage (Vt) becomes more positive (i.e. shifts to the right). The

result from one of the experiments, shown in Figure 6 below, delineates the sensor characteristics

observed after DNA interaction with pentacene.

Ion Ratio is defined as:

Ion Ratio = onmobilizatiBeforeDNAI

onmobilizatiAfterDNAntDrainCurreI

d

d

Im

Im)(

The bigger the Ion Ratio, the higher the sensitivity of the sensor. The DNA used in the

experiment below and all of the experiments/results that will be presented in the thesis, is a 125-

base pair designed single strand DNA sequence, synthesized by Biosynthesis, Inc.

Pentacene (5 benzene-ring) chemical structure

Benzene Structure

Single Strand DNA chemical structure

15

Figure 6. Current-Voltage characteristics from an experiment show a positive Vt shift and a higher hole current after

DNA immobilization.

Two possible explanations of the interaction of DNA and pentacene will be discussed below.

One is the direct doping of DNA on pentacene. The net-effect of the DNA on the pentacene

surface is electron withdrawing, therefore increasing the hole current of the transistor. Although

DNA is known to be negatively charged because of the phosphate groups on its backbone, the

possible interaction of the two chemical structures shown in Figure 5 might be causing the shift

in hole current. More specifically, areas of interaction include the aromatic structures in base

pairs (such as adenine) of the DNA, and the inherent aromaticity of the 5 benzene-ring pentacene

structure. The result of this electron withdrawing interaction is supported and explained below

using the Kelvin Probe Microscopy technique.

Another explanation uses the trap filling caused by the electrical interactions with DNA to

improve the overall performance of the transistor, resulting in the shift seen in Figure 6. Using

SCLC (Space Charge Limited Current) measurement technique, Zhang et al are able to support

this hypothesis. A short summary of Dr. Zhang‟s work on these explanations is discussed below.

Id (drain current) before DNA immobilization Id (drain current) after DNA immobilization

Positive Vt shift

Ion Ratio

16

Further details on the doping mechanisms of DNA can be found in his thesis, OTFT Based DNA

Detection System.

Dr. Zhang performed Atomic Force Microscopy (AFM) measurements to verify the DNA

doping of pentacene. Figure 7 below shows an AFM image of a DNA molecule along with

potential measurements done through Kelvin probe microscopy [5].

a)

b) Figure 7. Surface potential measurement of DNA molecules on a pentacene film. a) An Atomic Force Microscopy

(AFM) image, where the single bright thread is an immobilized DNA molecule. b) A surface potential image matching the same area of the AFM image on the top. It is clear that the DNA molecule has negative potential, which

is possibly caused by electron-withdrawing from the bottom pentacene film [5].

Kelvin Probe Force Microscopy (KPFM) is a scanned probe method where the potential

offset between a probe tip and a surface can be measured using the same principle as a

macroscopic Kelvin probe. The cantilever in the AFM is a reference electrode that forms a

17

capacitor with the surface, over which it is scanned laterally at a constant separation. With KPFM

the workfunction of surfaces can be observed at atomic or molecular scales. The work function

relates to many surface phenomena, including catalytic activity, reconstruction of surfaces,

doping and band-bending of semiconductors, charge trapping in dielectrics and corrosion [12].

Unlike workfunction mismatch induced surface potential seen in solid thin film study though, the

surface potential of macro bio-molecules is caused by dipoles. DNA hybridization or antigen-

antibody reactions have been proved to increase the amplitude of potential differences [11]. The

bright thread seen in Figure 7a (shown by the green and red arrows on the image) represents

DNA. The KPFM results in Figure 7b show that at the location of the DNA, the potential

becomes negative indicating that DNA, as a result of the immobilization and interaction with the

pentacene surface, has become more negatively charged. Given these results of the KPFM

measurements, Dr. Zhang is able to conclude that the DNA molecule has negative potential,

possibly caused by electron-withdrawing from the bottom pentacene film [5].

A space charge limited current (SCLC) measurement explains the doping mechanism as well.

SCLC is a result of the space charge region that forms within a conductive semiconductor

material where the charge carriers have diffused away, or have been forced away by an electric

field. Space charge regions have been shown to exist in pentacene films with gold electrodes in

previous studies [29]. The space charge region is characterized by generation-recombination

centers, steep impurity gradients, and rapidly changing populations of holes and electrons. This

space charge region limits the current carried by the majority carriers because of the presence

traps within this region. The electrical impact of DNA, with its negative charge, is to fill these

traps, facilitating the flow of hole current.

An SCLC setup measures two-terminal devices. A voltage from 0 to 100V is swept between

the source/drain with a floating gate electrode, and the current flowing into the drain is measured.

18

A typical SCLC curve is plotted on a log-log scale and has more than two different slopes. The

slope=1 part of the curve is called ohmic regime and the slope≥ 2 part is the trap-filling regime.

As the electrical field is increased to a certain level, the concentration of the injected carriers

from the electrode overwhelms the intrinsic carrier concentration, thus starting to fill traps in the

films. At a slope very close to 2, all traps been filled. This trap-free regime, slope equal 2, is also

called SCLC regime [5]. The SCLC curve, measured by Dr. Zhang, is shown in Figure 8 below.

Figure 8. Representative SCLC curves of original pentacene-based transistors and DNA immobilized transistors [5].

SCLC measurements before and after DNA immobilization shown above illustrate the doping

mechanism of DNA molecules on the pentacene film. In the ohmic regime, the magnitude of current

increased, relating to the free carrier concentration increase or mobility increase, because of

immobilized DNA segments. In the trap-filling regime, the decrease in slope from 3.4 as measured

in the original pentacene transistor to 2.7 in the DNA immobilized pentacene transistors indicates

that deep traps have been filled by DNA molecules. Figure 8 shows that the trap-filling action in

hole-rich pentacene film has actually increased the hole concentration. This is because trapped

electrons in the pentacene film have been released by DNA segments, thereby facilitating the flow of

hole current [5, 6].

19

III. Investigation of Physical Origins of Electrical Transduction

Behavior

An important step in making a viable DNA sensor is to show that what is being detected is

just the DNA, and to eliminate any other noise that might affect the sensitivity. One way to do

this is to correlate the amount of DNA on the surface with the electrical shift that is seen after

DNA immobilization. A few methods were used to obtain this data, namely atomic force

microscopy (AFM), fluorescent microscopy, and time-of-flight secondary ion mass spectroscopy

(TOF-SIMS). Useful results and insight were derived from the latter two techniques and are

discussed in this section.

A. Fluorescent Microscopy

Fluorescent microscopy is a light microscope used to study properties of organic or inorganic

substances using the phenomena of fluorescence and phosphorescence instead of, or in addition

to, reflection and absorption. A component of interest in the specimen is specifically labeled

with a fluorescent molecule called a fluorophore (such as green fluorescent protein (GFP),

fluorescein).

The ssDNA used in this experiment was tagged with carboxyfluorescein molecule on the

5‟end of the DNA strand. Carboxyfluorescein is a fluorescent dye with an excitation and

emission of 492/517 nm, respectively. In other words, the DNA with the fluorescent molecule is

illuminated with light of wavelength of 492nm which is absorbed by the fluorophores. It then

emits a light of wavelength of 517nm, which is in turn passed by an emission filter specific to the

emitted wavelength. The illumination light is separated from the much weaker emitted

fluorescence through the use of the emission filter. Typical components of a fluorescence

microscope are the light source (xenon arc lamp or mercury-vapor lamp), the excitation filter, the

20

dichroic mirror (or dichromatic beamsplitter), and the emission filter [14]. Figure 9 below

delineates the fluorescent microscopy system.

Figure 9. Fluorescent Microscope System [13].

A Zeiss AxioImager M1 fluorescence microscope system from the College of Natural

Resources, Berkeley, was used in this experiment. The microscope contains the Photometrics

Quantix digital CCD camera that allows visualization of fluorescence with high spatial resolution

(1200X1300 pixels) and high bit depth (12-bit gray scale). A second camera, QImaging 5MPix

MicroPublisher, captures color images. The set-up is shown in Figure 10 below.

Figure 10. Axioimager M1 Fluorescent Microscope Set-up at the College of Natural Resources (CNR) in Berkeley.

The CNR website (microsopy.berkeley.edu) provides further details on the mechanism and

capabilities of the fluorescent microscope system shown above. Experimental results/images

21

obtained from fluorescent density measurements are shown below. The area shown below is just

that of the channel area, i.e. 110um*1mm (horizontal orientation), where the DNA is pipetted.

a)

b)

Figure 11. Fluorescent density images captured using the digital CCD camera in the fluorescent microscope system. The green dots in images a) and b), which are to represent green-fluorescent tagged DNA, are from 2 substrates with

different pentacene thicknesses and concentrations of DNA.

The green dots seen above represent the green-fluorescent tagged DNA. The substrate in

11a, given the pentacene evaporation conditions that were used, is expected to have a rougher

surface (referring to results shown in Section IV) than that of 11b. Correspondingly, more DNA

can be seen immobilized in 11a. The quantitative measurement that was used to make this

conclusion was fluorescent density measurements. Fluorescent density or intensity

measurements were made using the camera, and its in-built segmentation software tool that adds

up the intensity of different segments of green fluorescence within the imaged channel area.

Before trying to correlate the these results with electrical results, control experiments were done

to see if the fluorescent tag had any effects on the electrical characteristics. No-tag DNA and

fluorescent tagged DNA were pipetted and treated with the same experimental protocol. Results

showed that the fluorescent tag, by itself, causes no additional shift to the electrical

characteristics.

channel area where DNA is pipetted

Pentacene TFT structure on wafer

22

While trying to correlate the electrical and visual results, some potential problems were

discovered. After doing some control experiments on the pentacene surface, it was observed

that the pentacene substrate also fluoresces, contributing its own fluorescent signal to the

measurements above. This is to be expected of an organic semiconductor such as pentacene.

Fluorescence occurs when a molecule, atom or nanostructure relaxes to its ground state after

being electrically excited. State S0 is called the ground state of the fluorophore (fluorescent

molecule) and S1 is its first (electronically) excited state [14].

A molecule in its excited state, S1, can relax by various competing pathways. It can undergo

non-radiative relaxation in which the excitation energy is dissipated as heat (vibrations) to the

solvent. Excited organic molecules can also relax via conversion to a triplet state which may

subsequently relax via phosphorescence or by a secondary non-radiative relaxation step.

Relaxation of an S1 state can also occur through interaction with a second molecule through

fluorescence quenching. Finally, molecules that are excited through light absorption can transfer

energy to a second 'sensitized' molecule, which is converted to its excited state and can then

fluoresce [14]. The latter case is likely the cause of the fluorescent „noise‟ that we are seeing

with the measurements above. It is not easy to separate this from the actual signal unless the

DNA is tagged with a fluorescent molecule which fluoresces at a wavelength that does not excite

the pentacene molecule. Alternatives that are being considered will be discussed in the last

paragraph of this section.

The second problem with fluorescent microscopy using the digital CCD camera is the

resolution of the camera. The Max XY resolution of this fluorescent microscope is about 200nm.

With the resolution of the available CCD camera, 100s of molecules needs to be clumped

together in the 200nm space for the system to detect the fluorescent signal. As discovered

23

through AFM, the distribution of DNA on the surface is random. In other words, the DNA

density in different scan areas is not expected to be consistent. Given the resolution of the digital

CCD camera, the random distribution makes it difficult to accurately capture the amount of DNA

on the surface.

To keep the pentacene substrate from fluorescing, different control experiments with a range

of emission filters were tried on the surface. We were able to conclude that pentacene does not

fluoresce at blue light wavelengths (450-495nm [15]) or below. In addition, to get as close as

possible to the resolution of the Single Molecule Spectroscopy (SMS) [16], which would allow

measurement of a single fluorescent molecule or a single strand of DNA, the College of Natural

Resources at Berkeley has recently obtained a new EMCCD (electron-multiplying charge

coupled device) camera, which is highly sensitive and super bright. It is much more sensitive

than the camera that has been used so far.

Since results with the current system are promising, future experiments for correlation will be

performed with blue fluorescent tagged DNA and the new EMCCD camera.

B. Time-of-Flight Secondary Ion Mass Spectroscopy (TOF-SIMS)

This section covers the use of TOF-SIMS technique for correlation and the results obtains from

the experiment.

i. Definition and Technical Capabilities

Time-of-Flight Secondary Ion Mass Spectrometry (TOF-SIMS) is a surface analytical

technique that focuses a pulsed beam of primary ions onto a sample surface, producing secondary

ions in a sputtering process. Analyzing these secondary ions provides information about the

24

molecular and elemental species present on the surface. For example, if there were organic

contaminants, such as oils adsorbed on the surface, TOF-SIMS would reveal this information,

whereas other techniques may not. Since TOF-SIMS is a survey technique, all the elements in

the periodic table, including H, are detected. Moreover, TOF-SIMS can provide mass spectral

information, image information in the XY dimension across a sample, and also depth profile

information on the Z dimension into a sample [17].

The surface sensitivity of TOF-SIMS makes it a good first pass at problem solving. Other

techniques can be used to obtain in depth information. The imaging capabilities of TOF-SIMS

can provide elemental and molecular information from defects and particles on the micron scale.

TOF-SIMS can also be used for depth profiling and compliments dynamic SIMS. The

advantages for TOF-SIMS for profiling are its small areas capabilities and also its ability to do

survey depth profiles [17].

TOF-SIMS is known to be one of the most sensitive surface analytical techniques, with a

detection limit of 107- 10

10 atoms/cm

2 sub-monolayer and a depth resolution of 2-3nm. It

provides specific molecular information on thin (sub-monolayer) organic films/contaminants

with an excellent detection limit (ppm) for most elements. The TOF-SIMS technique was used

instead of SIMS because TOF-SIMS is able to perform the SIMS analysis on organic thin films

and more specifically soft substrates. It is a non-destructive process tailored to work with soft

substrates such as pentacene. TOF-SIMS has a few limitations. The samples that need to be

analyzed must be vacuum compatible. In addition, this technique can be too surface sensitive

with sample packaging; prior handling may also impact quality of results. Despite the

limitations, the TOF-SIMS technique is extremely sensitive, reliable, and very useful in many

applications.

25

ii. Experiment, Procedure, and Results

The TOF-SIMS analysis was performed in the Evans Analytical Group labs in Sunnyvale,

CA. The TOF-SIMS analysis was performed to get a visual proof of DNA on surface and to see

if it is possible to correlate these results with the electrical measurements.

More specifically, the purpose of this analysis was to determine the relative levels of DNA on

three samples: Sample #1- Control: Pentacene substrate with no buffer or DNA but which has

gone through the same experimental protocol (fabrication and storage) as the other two; Sample

#2-Buffer: Pentacene substrate with just pure buffer and NO DNA. The buffer, as discussed

earlier is saline-sodium citrate solution; Sample #3-DNA in buffer: DNA was pipetted on the

channel of the transistor. Note that all 3 samples were taken from the same substrate; each of the

samples went through the same cleaning, evaporation, and storage procedures. The relative

levels of DNA were assessed by monitoring the levels of sodium (Na) and phosphate (POx) ions.

Experiment:

Data was obtained using a gallium liquid metal ion gun (LMIG) primary ion source (Ion

Potential: 12 kV for +ions and 18kV for –ions; DC Ion Current: 2nA). The instrument was

operated in an ion microprobe mode in which the bunched, pulsed primary ion beam was rastered

across the sample's surface. Three positive and three negative ion spectra were acquired from

each sample in order to investigate the reproducibility of the data. Acquisition of multiple spectra

serves to highlight possible chemical heterogeneity across the sample‟s surface. In each case the

analytical area was 80m × 80m. Spectra were acquired by specifying a particular area of the

sample as a “region of interest” (ROI). This allowed data to be acquired only from the channel,

eliminating signal from the surrounding gold lines.

The elements of interest were POx (for DNA) and Na (for buffer solution).

26

Summary of Results:

The highest levels of POx and Na were observed on sample #3. These species were observed

at significantly lower levels from sample #2. No POx was observed from sample #1 which had

the lowest levels of Na.

Other species observed from all samples included the pentacene substrate (C22H14),

Polydimethylsiloxane (PDMS), which is a signal observed because of the hexamethyldisilazane

(HMDS) layer in the dielectric/pentacene interface, and F, Fluorine.

Discussion of Results:

The results are presented as mass spectra, which are displayed as the number of secondary

ions detected (Y-axis) versus the mass-to-charge (m/z) ratio of the ions (X-axis). The ion counts

are shown on linear intensity scales, and probable empirical formulae for a number of peaks are

identified in the figures. The multiple spectra obtained from each sample were generally very

similar so only one representative spectrum from each sample in each polarity is included in this

report. Mass Spectra results can be found in the appendix of this report.

The results are also presented as tables of normalized intensities. These tables can be used to

compare the level of a given species between samples; however, they cannot be used to compare

the levels of different species either on the same sample or between samples. This is because

different species have different secondary ion yields so TOF-SIMS has different sensitivities for

different species. The tables show the mean of three measurements and also the standard

deviations.

27

1, control 2, buffer 3, (DNA + buffer)

m/z formula mean mean mean

Elements

23 Na 16.1 6.0 569 146 4790 580

28 Si 3670 180 3610 160 587 38

Pentacene

278 C22H14 105 11 40.4 7.7 55.0 6.2

PDMS

73 C3H9Si 408 31 317 42 95.9 12.0

147 C5H15Si2O 16.5 5.6 73.1 13.0 15.6 1.7

Table 1. Normalized positive ions of interest (normalized relative to total ion counts ×10000)

Table 1 above presents the normalized intensities of the positive ions of interest. Elemental

species observed from all three samples included Na, Si, and F. The levels of Na varied

significantly among the samples with the highest levels observed from sample #3 and the lowest

levels on Sample #1. The highest levels are expected from sample #3 because the DNA

molecules themselves will have the buffer solution (Na) attached to them. Sample #1 most likely

has Na because of contaminants from the environment. Comparable levels of Si were observed

on samples #1 and #2 while lower levels were found on #3. The lower levels in sample #3 are

possibly due to the strong signal exhibited by the Na in this sample. This can also be observed in

the mass spectra included in the appendix of this report.

1, control 2, buffer 3, DNA + buffer)

m/z formula mean mean mean

47 PO nd - 0.337 0.018 0.551 0.150

63 PO2 nd - 0.295 0.022 2.91 0.81

79 PO3 nd - nd - 2.45 0.73

“nd” indicates species not detected on that sample Table 2. Normalized negative ions of interest

(normalized relative to total ion counts ×10000)

28

Table 2 above presents the normalized intensities of the negative ions of interest. The focus was

on the POx ions to determine the presence of DNA on the surface and the relative amounts of

DNA on the different samples. Phosphate ions (POx) were observed in negative polarity only on

samples #2, and #3. The levels of POx on sample #3 were significantly higher than those

observed from sample #2. The PO3¯ ion in particular was observed only on sample #3.

Phosphate ions were not clearly observed in the mass spectra (shown in appendix) due to the

many strong hydrocarbons in the spectra.

Many ions due to HMDS (shown as PDMS in the mass spectra and tables) were observed

from all samples. Several hydrocarbon ions were also observed from all samples, and these are

shown in the mass spectra included in the appendix. Representative low mass positive ions

include C2H3, C3H5, and C4H7. Representative negative ions include C2H, C4H, and C6H.

Species such as these have multiple possible sources. They may be adsorbed low mass molecular

species or they may be fragment ions from other higher mass molecular species. The parent

molecule of the pentacene substrate (C22H14, m/z 278) was clearly observed on all samples. This

indicates that the coverage of adsorbed species (e.g. DNA and HMDS) is discontinuous and/or

less than a monolayer thick.

TOF-SIMS analysis has shown the presence of phosphate groups, particular PO2 and PO3

(from DNA) and sodium (from buffer solution) in highest concentrations in sample #3, the

sample with DNA immobilized in the channel. The control sample (sample #1) shows small

traces of sodium and the buffer sample (sample #2) shows small traces of PO possibly due to

environment contribution and/or contamination. Since Sample #3 has shown significantly higher

amounts of the species of interest, the presence of DNA on the pentacene surface has been

confirmed.

29

Fluorescent Microscopy and TOF-SIMS have both shown useful results to ascertain the

physical origins of electrical transduction behavior. TOF-SIMS results have confirmed the

presence of DNA on pentacene surface while modified fluorescent microscopy measurements

have shown potential for use in directly correlating the amount of DNA immobilized on the

surface with the shift of the electrical performance characteristics.

30

IV. Optimization of DNA Immobilization and Sensor Sensitivity

DNA immobilization has been previously discussed in Section II of this report. Physical

absorptive immobilization, the mechanism through which DNA immobilizes on pentacene,

highlights the importance of the topology of the pentacene film surface for immobilization of

DNA and the sensitivity of the sensor. DNA is known to segregate to topological features on

pentacene surface. We are able to exploit the control of pentacene evaporation conditions to tune

pentacene film morphology to maximize sensitivity. In this section, we demonstrate DNA

detection using optimized films.

The optimization experiment is split up into two parts. The first is a set of experiments that

characterize the pentacene (vary the pentacene morphology) by varying the input parameters.

The second integrates the DNA into the experiments, thereby arriving at the best surface and

evaporation conditions for highest sensor sensitivity.

A. Pentacene Characterization Experiment

By carefully controlling input parameters to set different evaporation conditions, we analyze

pentacene morphological and electrical characteristics. The input parameters that were varied

and the outputs that were extracted are shown in Table 3 below.

Input Variables Output Characteristics

Thickness of Pentacene film

Input Current (Temperature of the crucible)

Substrate Temperature

Idsat (Saturation Current)

Mean Roughness

Grain Size

Evaporation Rate

Coverage Area

Table 3. Input and Outputs for Pentacene Characterization Experiment.

31

i. Experimental Setup: Response Surface Design of Experiments (DOE)

Because the goal of this part of the experiment was to relate the input and output parameters,

the experiment needed to be tailored to ultimately result in providing these relationships. To

perform this comprehensive analysis, design of experiments (DOE) set-up was used. In

particular, a surface response DOE was used because of the importance of gaining a thorough

understanding of the output characteristics for optimizing DNA immobilization.

The DOE set-up was configured using JMP, a statistical software. The DOE platform in JMP

is an environment for describing the factors, responses and other specifications, creating a

designed experiment. JMP allows the design of different types of DOE including screening

design and response surface design. The former is to get an overall understanding while the

latter is gain an in-depth understanding of the relationships between input and output parameters.

Response surface designs are useful for modeling a curved surface (quadratic) to continuous

factors. If a minimum or maximum response exists inside the factor region, a response surface

model can pinpoint it. Three distinct values for each factor are necessary to fit a quadratic

function, so the standard two-level designs (such as screening design) cannot fit curved surfaces.

The most popular response surface design is the central composite design, illustrated by Figure

12 below [19].

Figure 12. Central Composite Design Methodology [18].

32

CCD combines a two-level fractional factorial (used in screening designs) and two other

kinds of points, namely center points, for which all the factor values are at the zero (or midrange)

value and axial (or star) points, for which all but one factor are set at zero (midrange) and that

one factor is set at outer (axial) values. There are two main types of central composite designs,

namely uniform precision and orthogonal designs. The properties of these central composite

designs relate to the number of center points in the design and to the axial values. For orthogonal

designs, the number of center points is chosen so that the second order parameter estimates are

minimally correlated with the other parameter estimates. Uniform precision, which was used for

the following DOE, means that the number of center points is chosen so that the prediction

variance at the center is approximately the same as at the design vertices [19].

Using a surface response design and more specifically, a uniform precision CCD set up, the

DOE shown in Table 4 below was followed. A set of 17 evaporation conditions was designed

for the experiment.

Table 4. JMP Surface Response DOE for Pentacene Characterization Experiment.

33

The input variables are shown in columns 3, 4, and 5. The output characteristics are in columns

6-10. For the DNA immobilization part of this experiment, 2 or 3 more columns will be added,

namely Idsat ratio, Threshold Voltage (Vt) shift, and mobility shift.

A set of 9 points was measured within each of these experiments in order to study the

variation within wafer and wafer to wafer variation. The measurement setup used for each of the

17 experiments is shown in Figure 13 below.

Figure 13. A set of 9 points was measured for each of the 17 experiments in the DOE to study variation within the

wafer and wafer-to-wafer variation.

ii. Experiment Protocol

The experimental procedure and specifications for the pentacene characterization experiment

is discussed below:

1) N++ silicon as substrate

2) 1000Å wet oxidation for SiO2 (using Tystar oxidation furnace)

3) HMDS on SiO2 surface (to optimize the interface between hydrophobic (photoresist) and

hydrophilic (SiO2) surfaces)

4) Photoresist spinning followed by lithography to define source and drain contacts

5) Evaporation of 1000Å of gold for source and drain contacts

6) Lift-off for 45 minutes in acetone

11 33

22

44

55 66 77

88

99

34

7) Sonication/Cleaning of Substrate- another 45 minutes in acetone followed by 45 minutes

in IPA (isopropyl alcohol)

8) HMDS on surface (to optimize interface between SiO2 (hydrophilic) and

pentacene(hydrophobic))

9) Evaporation of Pentacene using the specified input parameters on the DOE (the pentacene

evaporation was performed under vacuum with a chamber pressure of ~2.2 x10-7

Torr)

Figure 14. Pentacene TFT structure.

The sonication/cleaning step (step 7) is especially important for sensor sensitivity since this

establishes the interface between the dielectric and the semiconductor. DNA immobilizes and

interacts very close to this interface to cause the shift in the transistor characteristics.

ii. Results

Results: Within Wafer and Wafer to Wafer Variation

The results shown in Figure 15 below highlight the variation within each experiment and

across the different experiments. In addition, the wide range of electrical characteristics across

the DOE further emphasizes the importance of controlling these evaporation conditions to

determine the optimum immobilization for highest sensor sensitivity. The x axis of the graph

corresponds with the experiments on the DOE in Table 4.

100nm SiO2

Source

(100nm)

Drain

(100nm)

Gate (n++ Si)

Pentacene

35

The output shown here is saturation current (Idsat). The variation within each wafer and from

wafer to wafer will have a direct correlation with the other output characteristics. If there‟s more

variation (bigger standard deviation) within one experiment, then this is a result of lack of

uniformity of the pentacene surface in that experiment. This relates back to the morphological

characteristics. For example, higher input current or thicker pentacene, depending on which

factor dominates, likely causes non-uniform morphology within a wafer, increasing the variation

in electrical performance. This will be further discussed in the next part of this section.

Figure 15. Idsat (electrical characteristics) for all 17 experiments. These results display the different performance

characteristics across the DOE as well as the variation across the experiments.

The green diamonds seen in Figure 15 do not represent standard deviation. A means diamond

illustrates a sample mean and 95% confidence interval, as shown in Figure 16 below.

Figure 16. Explanation of the t-test mean diamonds used for statistical purposes [19].

36

The line across each diamond represents the group mean. The vertical span of each diamond

represents the 95% confidence interval for each group. Overlap marks are drawn above and

below the group mean. For groups with equal sample sizes, overlapping marks indicate that the

two group means are not significantly different at the 95% confidence level. The confidence

interval computation assumes that variances are equal across observations. Therefore, the height

of the confidence interval (diamond) is proportional to the reciprocal of the square root of the

number of observations in the group [19].

The horizontal extent of each group along the x-axis (the horizontal size of the diamond) is

proportional to the sample size of each level of the x variable. It follows that the narrower

diamonds are usually the taller ones because fewer data points yield a less precise estimate of the

group mean [19]. The t-test/anova analysis shown in Figure 15 also shows the sameness data.

Between two experiments, if the overlap/sameness value (measured by t-test) is <5%, then we

can be „confident‟ that they are statistically different.

The variation can be characterized by the standard deviation value for each of the

experiments. Experiment 9 had the highest standard deviation value of 0.000022, and the

corresponding variation can be seen clearly in Figure 15. Experiment 14 and Experiment 6, had

the lowest variation with a standard deviation of 0.000001.

Results: Relating Input and Output Parameters

1. Saturation Current (Idsat)

Figure 18 below shows the effect that the three input parameters (Input Current, Thickness of

Pentacene, and Substrate Temperature) have on Saturation current (Idsat). Saturation current and

other electrical characteristics were measured using the Agilent HP4156 probe station purged in

37

nitrogen. The measurements were done under nitrogen, as is usually the practice when measuring

pentacene transistors, because pentacene is oxygen-sensitive.

Figures 17 and 19 are extracted by JMP using the 153 experimental/electrical measurements

(17 experiments * 9 measurements for each) performed. Using leverage plots, some of which are

shown in Figure 17 below, and relating the input parameters to outputs, JMP is able to provide a

relationship between the input and output characteristics, which are shown in Figures 17 and 19.

Figure 17. Leverage Plots relating input and output parameters using the measurements performed. Some parameters show a cause-effect relationship while others do not affect the saturation current.

Figure 18 below shows how the three input parameters independently affect the saturation

current.

38

Figure 18. Relationship between the input parameters and Idsat.

As mentioned earlier, the pentacene TFT device is a p-type accumulation device; the Y-axis in

the figure above is negative current representing hole current. The figure above shows that as the

thickness of the pentacene increases, the hole current (magnitude) also increases. This result can

be justified using a schematic of the known growth modes of pentacene shown in Figure 19. The

stacking of the pentacene and surface coverage causes this increase in electrical characteristics.

Figure 19 below shows the schematic diagram of the pentacene throughout a range of

thicknesses, and its effect on the electrical characteristics is further explained below.

Figure 19. Schematic diagram depicting the phases and grain orientation in various pentacene film [20].

The schematic diagram shown above has been produced as a result of GIXD (grazing incidence

X-ray diffraction) characterization done by Dr. Sandra Fritz [20]. The schematic diagram shows

-8e-5

-6e-5

-4e-5

-2e-5

0e+0

2e-5

Idsat

Act

ua

l

-0.00008-0.00005 -0.00002 0 .00002

Idsat Predicted P<.0001 RSq=0.41

RMSE=1.4e-5

Actual by Predicted Plot

RSquare

RSquare Adj

Root Mean Square Error

Mean of Response

Observati ons (or Sum Wgts)

0.409977

0.369748

0.000014

-0.00002

142

Summary of Fit

Model

Error

C. Total

Source

9

132

141

DF

1.8854e-8

2.71339e-8

4.59879e-8

Sum of Squares

2.0949e-9

2.056e-10

Mean Square

10.1911

F Ratio

<.0001

Prob > F

Analysis of Variance

Lack Of Fit

Pure Error

Total Error

Source

5

127

132

DF

1.1973e-8

1.51609e-8

2.71339e-8

Sum of Squares

2.3946e-9

1.194e-10

Mean Square

20.0591

F Ratio

<.0001

Prob > F

0.6703

Max RSq

Lack Of Fit

Intercept

Thi ckness of Pentacene(140.54,259.46)&RS

Substrate Temperature(34.122,60.878)&RS

Input Current(4.951,5.249)&RS

Thi ckness of Pentacene(140.54,259.46)*Substrate Temperature(34.122,60.878)

Thi ckness of Pentacene(140.54,259.46)*Input Current(4.951,5.249)

Substrate Temperature(34.122,60.878)*Input Current(4.951,5.249)

Thi ckness of Pentacene(140.54,259.46)*Thickness of Pentacene(140.54,259.46)

Substrate Temperature(34.122,60.878)*Substrate Temperature(34.122,60.878)

Input Current(4.951,5.249)*Input Current(4.951,5.249)

Term

-0.000014

-0.000008

-0.000003

-0.000007

-0.000007

2.3039e-7

0.0000011

-0.000001

-0.000001

-0.000004

Estimate

0.000003

0.000001

0.000001

0.000001

0.000002

0.000002

0.000002

0.000002

0.000002

0.000002

Std Error

-4.05

-5.89

-2.56

-5.24

-4.03

0.14

0.66

-0.74

-0.89

-2.25

t Rati o

<.0001

<.0001

0.0116

<.0001

<.0001

0.8918

0.5088

0.4591

0.3742

0.0263

Prob>|t|

Parameter Estimates










Source

1

1

1

1

1

1

1

1

1

Nparm

1

1

1

1

1

1

1

1

1

DF

7.12763e-9

1.34661e-9

5.65133e-9

3.33073e-9

3.8216e-12

9.0222e-11

1.1331e-10

1.6346e-10

1.03827e-9

Sum of Squares

34.6742

6.5509

27.4924

16.2032

0.0186

0.4389

0.5512

0.7952

5.0510

F Ratio

<.0001

0.0116

<.0001

<.0001

0.8918

0.5088

0.4591

0.3742

0.0263

Prob > F

Effect Tests

-0.00005

-0.00004

-0.00003

-0.00002

-0.00001

0

0.00001

0.00002

0.00003

0.00004

Idsat

Resid

ua

l

-0.00008-0.00005 -0.00002 0 .00002

Idsat Predicted

Residual by Predicted Plot

Whole Model

-0.00008

-0.00006

-0.00004

-0.00002

0

0.00002

Idsat

Le

vera

ge

Resid

ua

ls

-1.5 -1.0 -0.5 .0 .5 1.0 1.5

Thi ckness of

Pentacene(140.54,259.46)&RS Leverage,

P<.0001

Leverage Plot

Thickness of Pentacene(140.54,259.46)&RS

-0.00008

-0.00006

-0.00004

-0.00002

0

0.00002

Idsat

Le

vera

ge

Resid

ua

ls

-1.5 -1.0 -0.5 .0 .5 1.0 1.5

Substrate

Temperature(34.122,60.878)&RS

Leverage, P=0.0116

Leverage Plot


-0.00008

-0.00006

-0.00004

-0.00002

0

0.00002

Idsat

Le

vera

ge

Resid

ua

ls

-2.0 -1.5 -1.0 -0.5 .0 .5 1.0 1.5 2.0

Input Current(4.951,5.249)&RS Leverage,

P<.0001

Leverage Plot


-0.00008

-0.00006

-0.00004

-0.00002

0

0.00002

Idsat

Le

vera

ge

Resid

ua

ls

-0.000026 -0.000021 -0.000017 -0.000013

Thi ckness of

Pentacene(140.54,259.46)*Substrate

Temperature(34. Leverage, P<.00

Leverage Plot

Thickness of Pentacene(140.54,259.46)*Substrate Temperature(34.122,60.878)

-0.00008

-0.00006

-0.00004

-0.00002

0

0.00002

Idsat

Le

vera

ge

Resid

ua

ls

-0.00002 -0.000019-0.000019-0.000018

Thi ckness of

Pentacene(140.54,259.46)*Input

Current(4.951,5.249 Leverage, P=0.8

Leverage Plot

Thickness of Pentacene(140.54,259.46)*Input Current(4.951,5.249)

-0.00008

-0.00006

-0.00004

-0.00002

0

0.00002

Idsat

Le

vera

ge

Resid

ua

ls

-0.00002 -0.000019 -0.000018

Substrate

Temperature(34.122,60.878)*Input

Current(4.951,5.249) Leverage, P=0.5

Leverage Plot


-0.00008

-0.00006

-0.00004

-0.00002

0

0.00002

Idsat

Le

vera

ge

Resid

ua

ls

-0.000021 -0.00002 -0.000019-0.000018

Thi ckness of

Pentacene(140.54,259.46)*Thi ckness of

Pentacene(14 Leverage, P=0.4

Leverage Plot

Thickness of Pentacene(140.54,259.46)*Thickness of Pentacene(140.54,259.46)

-0.00008

-0.00006

-0.00004

-0.00002

0

0.00002

Idsat

Le

vera

ge

Resid

ua

ls

-0.000021 -0.00002 -0.000018-0.000017

Substrate

Temperature(34.122,60.878)*Substrate

Temperature(34.1 Leverage, P=0.3

Leverage Plot


-0.00008

-0.00006

-0.00004

-0.00002

0

0.00002

Idsat

Le

vera

ge

Resid

ua

ls

-0.000025-0.000021-0.000018-0.000015

Input Current(4.951,5.249)*Input


Leverage Plot


Thi ckness of Pentacene(140.54,259.46)

Substrate Temperature(34.122,60.878)

Input Current(4.951,5.249)

-0.000001

.

.

-0.000007

-0.000001

.

2.3039e-7

0.0000011

-0.000004

-0.000008

-0.000003

-0.000007

Coef

Thi ckness of Pentacene(140.54,259.46)Substrate Temperature(34.122,60.878)Input Current(4.951,5.249) Idsat




Variabl e

186.39104

33.056779

4.9301363

Crit ical Val ue

Soluti on i s a SaddlePoint

Predi cted Value at Soluti on-0.000007

Solution

Response Surface

-9e-5

-7e-5

-5e-5

-3e-5

-1e-5

1e-5

3e-5

Idsat

-9e-5

-7e-5

-5e-5

-3e-5

-1e-5

1e-5

3e-5

Idsat

-9e-5

-7e-5

-5e-5

-3e-5

-1e-5

1e-5

3e-5

Idsat

Thi ckness of Pentacene

34.122

60.878

4.951

5.249

150 200 250

140.54

259.46


4.951

5.249

40 50 60

140.54

259.46

34.12260.878

Input Current

5 5.1 5.2 5.3

Thick

ne

ss o

f Pe

nta

cene

Sub

stra

te T

em

pe

ratu

reIn

put C

urre

nt

Interaction Profiles

Idsat

0.00002

-7.5e-5

-0.00001


140

.54

259

.46

200


34.1

22

60.8

78

47.5

Input Current

4.9

51

5.2

49

5.1

Prediction Profiler

Response Idsat

39

that after about 400 Å, the pentacene starts exhibiting „bulk-like‟ characteristics. 200-400 Å

represents the transition. The increase in current in our experiment is mainly due to the stacking

of the monolayers and uniform coverage till about 200 Å or so. In our experiment, the

thicknesses were varied from 100-300 Å . Throughout this range, the stacking helps increase the

mobility and conductivity. The coverage starts getting worse after about 400 Å .

The results in Figure 18 also show that as the substrate temperature increases, the saturation

current increases. As the substrate temperature increases, the grain size increases due to

thermodynamic factors such as sticking coefficient and diffusion of the pentacene grains during

evaporation. As many groups have shown, as grain size increases, the mobility increases. Since

electrical conductivity is limited by charge transfer across grain boundaries the formation of

larger crystallites leads to enhanced conductivities [23]. Mobility increase leads to the current

increase that we see in our experiment.

Increase in input current, which directly relates to the temperature of the crucible holding the

pentacene, interestingly increases the overall saturation current as well. An interplay of multiple

factors is most likely causing this shift.

The interplay among the different factors is shown in the interaction plot shown in Figure 20.

When the lines within an interaction box are not parallel, it indicates that the two inputs

corresponding to the „interaction‟ box are dependent on each other in causing an effect on the

output, which in this case in Idsat. Note that Figure 18 delineates how the inputs are

independently affecting Idsat while Figure 20 presents the dependent relationships. In Figure 20

below, the two factors that are interacting in affecting Idsat are thickness of pentacene and

substrate temperature. These are circled in the figure below.

40

Figure 20. Interaction Profiler Plot for Idsat.

In Figure 18, the values of the inputs were set at midpoint to get the general trend. The

interaction plot reveals that since thickness of pentacene and substrate temperature are

interacting, if the thickness of pentacene for example changed from the midpoint value of 200 Å

to 100 Å , then the way the substrate temperature affects the Idsat will be different (i.e. different

slope). An illustration of this is shown in Figure 21 below. When tuning the output

characteristics, these interactions need to be taken into consideration as well.

-8e-5

-6e-5

-4e-5

-2e-5

0e+0

2e-5

Idsat

Act

ua

l

-0.00008-0.00005 -0.00002 0 .00002


RMSE=1.4e-5


RSquare

RSquare Adj


Mean of Response


0.409977

0.369748

0.000014

-0.00002

142

Summary of Fit

Model

Error

C. Total

Source

9

132

141

DF

1.8854e-8

2.71339e-8

4.59879e-8

Sum of Squares

2.0949e-9

2.056e-10

Mean Square

10.1911

F Ratio

<.0001

Prob > F


Lack Of Fit

Pure Error

Total Error

Source

5

127

132

DF

1.1973e-8

1.51609e-8

2.71339e-8

Sum of Squares

2.3946e-9

1.194e-10

Mean Square

20.0591

F Ratio

<.0001

Prob > F

0.6703

Max RSq

Lack Of Fit

Intercept










Term

-0.000014

-0.000008

-0.000003

-0.000007

-0.000007

2.3039e-7

0.0000011

-0.000001

-0.000001

-0.000004

Estimate

0.000003

0.000001

0.000001

0.000001

0.000002

0.000002

0.000002

0.000002

0.000002

0.000002

Std Error

-4.05

-5.89

-2.56

-5.24

-4.03

0.14

0.66

-0.74

-0.89

-2.25

t Rati o

<.0001

<.0001

0.0116

<.0001

<.0001

0.8918

0.5088

0.4591

0.3742

0.0263

Prob>|t|

Parameter Estimates










Source

1

1

1

1

1

1

1

1

1

Nparm

1

1

1

1

1

1

1

1

1

DF

7.12763e-9

1.34661e-9

5.65133e-9

3.33073e-9

3.8216e-12

9.0222e-11

1.1331e-10

1.6346e-10

1.03827e-9

Sum of Squares

34.6742

6.5509

27.4924

16.2032

0.0186

0.4389

0.5512

0.7952

5.0510

F Ratio

<.0001

0.0116

<.0001

<.0001

0.8918

0.5088

0.4591

0.3742

0.0263

Prob > F

Effect Tests

-0.00005

-0.00004

-0.00003

-0.00002

-0.00001

0

0.00001

0.00002

0.00003

0.00004

Idsat

Resid

ua

l

-0.00008-0.00005 -0.00002 0 .00002

Idsat Predicted


Whole Model

-0.00008

-0.00006

-0.00004

-0.00002

0

0.00002

Idsat

Le

vera

ge

Resid

ua

ls

-1.5 -1.0 -0.5 .0 .5 1.0 1.5

Thi ckness of


P<.0001

Leverage Plot


-0.00008

-0.00006

-0.00004

-0.00002

0

0.00002

Idsat

Le

vera

ge

Resid

ua

ls

-1.5 -1.0 -0.5 .0 .5 1.0 1.5

Substrate


Leverage, P=0.0116

Leverage Plot


-0.00008

-0.00006

-0.00004

-0.00002

0

0.00002

Idsat

Le

vera

ge

Resid

ua

ls

-2.0 -1.5 -1.0 -0.5 .0 .5 1.0 1.5 2.0


P<.0001

Leverage Plot


-0.00008

-0.00006

-0.00004

-0.00002

0

0.00002

Idsat

Le

vera

ge

Resid

ua

ls

-0.000026 -0.000021 -0.000017 -0.000013

Thi ckness of



Leverage Plot


-0.00008

-0.00006

-0.00004

-0.00002

0

0.00002

Idsat

Le

vera

ge

Resid

ua

ls

-0.00002 -0.000019-0.000019-0.000018

Thi ckness of



Leverage Plot


-0.00008

-0.00006

-0.00004

-0.00002

0

0.00002

Idsat

Le

vera

ge

Resid

ua

ls

-0.00002 -0.000019 -0.000018

Substrate



Leverage Plot


-0.00008

-0.00006

-0.00004

-0.00002

0

0.00002

Idsat

Le

vera

ge

Resid

ua

ls

-0.000021 -0.00002 -0.000019-0.000018

Thi ckness of



Leverage Plot


-0.00008

-0.00006

-0.00004

-0.00002

0

0.00002

Idsat

Le

vera

ge

Resid

ua

ls

-0.000021 -0.00002 -0.000018-0.000017

Substrate



Leverage Plot


-0.00008

-0.00006

-0.00004

-0.00002

0

0.00002

Idsat

Le

vera

ge

Resid

ua

ls

-0.000025-0.000021-0.000018-0.000015



Leverage Plot





-0.000001

.

.

-0.000007

-0.000001

.

2.3039e-7

0.0000011

-0.000004

-0.000008

-0.000003

-0.000007

Coef





Variabl e

186.39104

33.056779

4.9301363

Crit ical Val ue



Solution

Response Surface

-9e-5

-7e-5

-5e-5

-3e-5

-1e-5

1e-5

3e-5Id

sat

-9e-5

-7e-5

-5e-5

-3e-5

-1e-5

1e-5

3e-5

Idsat

-9e-5

-7e-5

-5e-5

-3e-5

-1e-5

1e-5

3e-5

Idsat


34.122

60.878

4.951

5.249

150 200 250

140.54

259.46


4.951

5.249

40 50 60

140.54

259.46

34.12260.878

Input Current

5 5.1 5.2 5.3

Thick

ne

ss o

f Pe

nta

cene

Sub

stra

te T

em

pe

ratu

reIn

put C

urre

nt


Idsat

0.00002

-7.5e-5

-0.00001


140

.54

259

.46

200


34.1

22

60.8

78

47.5

Input Current

4.9

51

5.2

49

5.1

Prediction Profiler

Response Idsat

41

Idsa

t

0.00002

-7.5e-5

-7.43e-6


140

.54

259

.46

141


34.

122

60.

878

47.5

Input Current4

.951

5.2

495.1

-8e-5

-6e-5

-4e-5

-2e-5

0e+0

2e-5

Idsat

Act

ua

l

-0.00008-0.00005 -0.00002 0 .00002


RMSE=1.4e-5


RSquare

RSquare Adj


Mean of Response


0.409977

0.369748

0.000014

-0.00002

142

Summary of Fit

Model

Error

C. Total

Source

9

132

141

DF

1.8854e-8

2.71339e-8

4.59879e-8

Sum of Squares

2.0949e-9

2.056e-10

Mean Square

10.1911

F Ratio

<.0001

Prob > F


Lack Of Fit

Pure Error

Total Error

Source

5

127

132

DF

1.1973e-8

1.51609e-8

2.71339e-8

Sum of Squares

2.3946e-9

1.194e-10

Mean Square

20.0591

F Ratio

<.0001

Prob > F

0.6703

Max RSq

Lack Of Fit

Intercept










Term

-0.000014

-0.000008

-0.000003

-0.000007

-0.000007

2.3039e-7

0.0000011

-0.000001

-0.000001

-0.000004

Estimate

0.000003

0.000001

0.000001

0.000001

0.000002

0.000002

0.000002

0.000002

0.000002

0.000002

Std Error

-4.05

-5.89

-2.56

-5.24

-4.03

0.14

0.66

-0.74

-0.89

-2.25

t Rati o

<.0001

<.0001

0.0116

<.0001

<.0001

0.8918

0.5088

0.4591

0.3742

0.0263

Prob>|t|

Parameter Estimates










Source

1

1

1

1

1

1

1

1

1

Nparm

1

1

1

1

1

1

1

1

1

DF

7.12763e-9

1.34661e-9

5.65133e-9

3.33073e-9

3.8216e-12

9.0222e-11

1.1331e-10

1.6346e-10

1.03827e-9

Sum of Squares

34.6742

6.5509

27.4924

16.2032

0.0186

0.4389

0.5512

0.7952

5.0510

F Ratio

<.0001

0.0116

<.0001

<.0001

0.8918

0.5088

0.4591

0.3742

0.0263

Prob > F

Effect Tests

-0.00005

-0.00004

-0.00003

-0.00002

-0.00001

0

0.00001

0.00002

0.00003

0.00004

Idsat

Resid

ua

l

-0.00008-0.00005 -0.00002 0 .00002

Idsat Predicted


Whole Model

-0.00008

-0.00006

-0.00004

-0.00002

0

0.00002

Idsat

Le

vera

ge

Resid

ua

ls

-1.5 -1.0 -0.5 .0 .5 1.0 1.5

Thi ckness of


P<.0001

Leverage Plot


-0.00008

-0.00006

-0.00004

-0.00002

0

0.00002

Idsat

Le

vera

ge

Resid

ua

ls

-1.5 -1.0 -0.5 .0 .5 1.0 1.5

Substrate


Leverage, P=0.0116

Leverage Plot


-0.00008

-0.00006

-0.00004

-0.00002

0

0.00002

Idsat

Le

vera

ge

Resid

ua

ls

-2.0 -1.5 -1.0 -0.5 .0 .5 1.0 1.5 2.0


P<.0001

Leverage Plot


-0.00008

-0.00006

-0.00004

-0.00002

0

0.00002

Idsat

Le

vera

ge

Resid

ua

ls

-0.000026 -0.000021 -0.000017 -0.000013

Thi ckness of



Leverage Plot


-0.00008

-0.00006

-0.00004

-0.00002

0

0.00002

Idsat

Le

vera

ge

Resid

ua

ls

-0.00002 -0.000019-0.000019-0.000018

Thi ckness of



Leverage Plot


-0.00008

-0.00006

-0.00004

-0.00002

0

0.00002

Idsat

Le

vera

ge

Resid

ua

ls

-0.00002 -0.000019 -0.000018

Substrate



Leverage Plot


-0.00008

-0.00006

-0.00004

-0.00002

0

0.00002

Idsat

Le

vera

ge

Resid

ua

ls

-0.000021 -0.00002 -0.000019-0.000018

Thi ckness of



Leverage Plot


-0.00008

-0.00006

-0.00004

-0.00002

0

0.00002

Idsat

Le

vera

ge

Resid

ua

ls

-0.000021 -0.00002 -0.000018-0.000017

Substrate



Leverage Plot


-0.00008

-0.00006

-0.00004

-0.00002

0

0.00002

Idsat

Le

vera

ge

Resid

ua

ls

-0.000025-0.000021-0.000018-0.000015



Leverage Plot





-0.000001

.

.

-0.000007

-0.000001

.

2.3039e-7

0.0000011

-0.000004

-0.000008

-0.000003

-0.000007

Coef





Variabl e

186.39104

33.056779

4.9301363

Crit ical Val ue



Solution

Response Surface

-9e-5

-7e-5

-5e-5

-3e-5

-1e-5

1e-5

3e-5

Idsat

-9e-5

-7e-5

-5e-5

-3e-5

-1e-5

1e-5

3e-5

Idsat

-9e-5

-7e-5

-5e-5

-3e-5

-1e-5

1e-5

3e-5

Idsat


34.122

60.878

4.951

5.249

150 200 250

140.54

259.46


4.951

5.249

40 50 60

140.54

259.46

34.12260.878

Input Current

5 5.1 5.2 5.3

Thick

ne

ss o

f Pe

nta

cene

Sub

stra

te T

em

pe

ratu

reIn

put C

urre

nt


Idsat

0.00002

-7.5e-5

-0.00001


140

.54

259

.46

200


34.1

22

60.8

78

47.5

Input Current

4.9

51

5.2

49

5.1

Prediction Profiler

Response Idsat

a)

b)

Note that in Figure 21b, the thickness of pentacene has been set at 141 Å (i.e. the middle red line

has been moved). In this case the substrate temperature has the opposite effect on Idsat as

compared to Figure 21a. The reason for the interaction here can be supported by the

explanations that have been already provided on how the thickness of pentacene affects the

stacking and consequently the saturation current.

Figure 22 shows the wide range of electrical characteristics obtained by varying the

evaporation conditions. The saturation current varied from about ~2 A in the lowest

performing sample to ~60 A on the highest performing sample. Figure 22 below shows the

characteristics from two different experiments, representing the lowest and highest performance.

Figure 21. Illustration of the significance of the interaction plot.

42

Figure 22. Id- Vd characteristics from 2 different experiments with the lowest(left) and highest(right) performing samples.

To summarize, the three input factors affect the Idsat in the following way. Increase in

thickness of pentacene causes an increase in the hole current because of the increase in film

stacking and coverage explained using the schematic diagram. The increase in substrate

temperature leads to bigger grain sizes, which consequently causes the enhanced conductivity.

Finally, the input current increase also enhances the performance because of a combination of an

increase in grain size and interplay with the other input factors.

2. Mean Roughness of Pentacene Film Surface

Using a similar system of leverage plots as shown in Figure 17, the relationship between the

input factors and mean roughness has been extracted and shown in Figure 23. Mean roughness

was measured using Atomic Force Microscopy (AFM).

Figure 23. Relationship between the input parameters and mean roughness.

2

2.5

3

3.5

4

4.5

5

5.5

6

6.5

Me

an

Ro

ug

hn

ess

Act

ua

l

2.0 2.5 3.0 3.5 4.0 4.5 5.0 5.5 6.0 6.5

Mean Roughness Predi cted P<.0001

RSq=0.75 RMSE=0.4609


RSquare

RSquare Adj


Mean of Response


0.746553

0.729272

0.46086

4.243979

142

Summary of Fit

Model

Error

C. Total

Source

9

132

141

DF

82.58211

28.03579

110.61790

Sum of Squares

9.17579

0.21239

Mean Square

43.2021

F Ratio

<.0001

Prob > F


Lack Of Fit

Pure Error

Total Error

Source

5

127

132

DF

9.611127

18.424668

28.035795

Sum of Squares

1.92223

0.14508

Mean Square

13.2498

F Ratio

<.0001

Prob > F

0.8334

Max RSq

Lack Of Fit

Intercept










Term

4.2663565

0.7709168

-0.231355

-0.041969

-0.080681

0.1065694

-0.139403

-0.024359

-0.006368

0.0063131

Estimate

0.108316

0.041569

0.041569

0.042661

0.054313

0.054313

0.054313

0.050573

0.050573

0.051564

Std Error

39.39

18.55

-5.57

-0.98

-1.49

1.96

-2.57

-0.48

-0.13

0.12

t Rati o

<.0001

<.0001

<.0001

0.3270

0.1398

0.0519

0.0114

0.6308

0.9000

0.9027

Prob>|t|

Parameter Estimates










Source

1

1

1

1

1

1

1

1

1

Nparm

1

1

1

1

1

1

1

1

1

DF

73.047982

6.578877

0.205558

0.468673

0.817707

1.399186

0.049276

0.003367

0.003184

Sum of Squares

343.9294

30.9751

0.9678

2.2066

3.8500

6.5877

0.2320

0.0159

0.0150

F Ratio

<.0001

<.0001

0.3270

0.1398

0.0519

0.0114

0.6308

0.9000

0.9027

Prob > F

Effect Tests

-1.5

-1.0

-0.5

0.0

0.5

1.0

1.5

Me

an

Ro

ug

hn

ess

Re

sid

ual

2.0 2.5 3.0 3.5 4.0 4.5 5.0 5.5 6.0 6.5

Mean Roughness Predi cted


Whole Model

2.0

2.5

3.0

3.5

4.0

4.5

5.0

5.5

6.0

6.5

Me

an

Ro

ug

hn

ess

Leve

rage

Re

sid

uals

-1.5 -1.0 -0.5 .0 .5 1.0 1.5

Thi ckness of


P<.0001

Leverage Plot


2.0

2.5

3.0

3.5

4.0

4.5

5.0

5.5

6.0

6.5

Me

an

Ro

ug

hn

ess

Leve

rage

Re

sid

uals

-1.5 -1.0 -0.5 .0 .5 1.0 1.5

Substrate


Leverage, P<.0001

Leverage Plot


2.0

2.5

3.0

3.5

4.0

4.5

5.0

5.5

6.0

6.5

Me

an

Ro

ug

hn

ess

Leve

rage

Re

sid

uals

-2.0 -1.5 -1.0 -0.5 .0 .5 1.0 1.5 2.0


P=0.3270

Leverage Plot


2.0

2.5

3.0

3.5

4.0

4.5

5.0

5.5

6.0

6.5

Me

an

Ro

ug

hn

ess

Leve

rage

Re

sid

uals

4.164.184.204.224.244.264.284.304.32

Thi ckness of


Temperature(34. Leverage, P=0.1

Leverage Plot


2.0

2.5

3.0

3.5

4.0

4.5

5.0

5.5

6.0

6.5

Me

an

Ro

ug

hn

ess

Leve

rage

Re

sid

uals

4.15 4.20 4.25 4.30 4.35

Thi ckness of



Leverage Plot


2.0

2.5

3.0

3.5

4.0

4.5

5.0

5.5

6.0

6.5

Me

an

Ro

ug

hn

ess

Leve

rage

Re

sid

uals

4.10 4.15 4.20 4.25 4.30 4.35 4.40

Substrate



Leverage Plot


2.0

2.5

3.0

3.5

4.0

4.5

5.0

5.5

6.0

6.5

Me

an

Ro

ug

hn

ess

Leve

rage

Re

sid

uals

4.21 4.22 4.23 4.24 4.25 4.26 4.27 4.284.29

Thi ckness of



Leverage Plot


2.0

2.5

3.0

3.5

4.0

4.5

5.0

5.5

6.0

6.5

Me

an

Ro

ug

hn

ess

Leve

rage

Re

sid

uals

4.22 4.23 4.24 4.25 4.26 4.27

Substrate



Leverage Plot


2.0

2.5

3.0

3.5

4.0

4.5

5.0

5.5

6.0

6.5

Me

an

Ro

ug

hn

ess

Leve

rage

Re

sid

uals

4.21 4.22 4.23 4.24 4.25 4.26 4.27



Leverage Plot


Me

an

Ro

ug

hn

ess

6.127

2.208

4.266357


140

.54

259

.46

200


34.1

22

60.8

78

47.5

Input Current

4.9

51

5.2

49

5.1

2

2.5

3

3.5

4

4.5

5

5.5

6

6.5

Me

an

Ro

ug

hn

ess

2

2.5

3

3.5

4

4.5

5

5.5

6

6.5

Me

an

Ro

ug

hn

ess

2

2.5

3

3.5

4

4.5

5

5.5

6

6.5

Me

an

Ro

ug

hn

ess


34.122

60.878

4.9515.249

150 200 250

140.54

259.46


4.951

5.249

40 50 60

140.54

259.46

34.122

60.878

Input Current

5 5.1 5.2 5.3

Thick

ne

ss o

f Pe

nta

cene

Sub

stra

te T

em

pe

ratu

reIn

put C

urre

nt


Prediction Profiler

Response Mean Roughness

Id-Vd

-2.50E-06

-2.00E-06

-1.50E-06

-1.00E-06

-5.00E-07

0.00E+00

5.00E-07

-50.00 -40.00 -30.00 -20.00 -10.00 0.00 ID

ID2

ID3

ID4

ID5

Id-Vd

-7.00E-05

-6.00E-05

-5.00E-05

-4.00E-05

-3.00E-05

-2.00E-05

-1.00E-05

0.00E+00

1.00E-05

-50.00 -40.00 -30.00 -20.00 -10.00 0.00ID

ID2

ID3

ID4

ID5

43

As the thickness of pentacene increases, the film roughness increases because the higher the

number of monolayers, the rougher the film usually tends to be. This is further confirmed when

looking at the schematic diagram in Figure 19. As the thickness of the pentacene film increases,

the stacking becomes misaligned, especially after about 160 Å or so. Despite the increase in

roughness, the current has still gone up with higher thickness possibly because other factors that

affect the current such as better transport within a monolayer dominated.

As the substrate temperature increases, the roughness decreases. When evaporation of

pentacene is done at higher substrate temperatures, the formation of the pentacene crystals on the

surface is more ordered, thereby reducing the overall roughness. The effect of substrate

temperature on the ordering of thin films of pentacene has been analyzed previously [24].

The input current does not independently affect the mean roughness according to Figure 23

above, but it does interact with both substrate temperature and to a lesser extent, thickness of

pentacene as shown in Figure 24 below. Although at a pentacene thickness of 200 Å and a

substrate temperature of 47.5 C , the input current does not affect the mean roughness, at other

values of thicknesses and substrate temperatures, it does. Depending on what thickness and

substrate temperature turns out to be ideal for DNA immobilization and overall sensitivity, the

input current will have to be tuned accordingly.

44

Figure 24. Interaction Profiler Plot for Mean Roughness.

As previously discussed, the topological features on the pentacene film are critical for

DNA immobilization and sensor sensitivity. By doing AFM on these samples, surface

roughness data on the samples were extracted. Figure 25 shows AFM images of samples from 2

different experiments, representing surfaces with the lowest (2.208nm) and highest (6.127nm)

mean roughness. The wide range of the surface roughness data further emphasizes the need and

importance of controlling the input parameters to get optimum morphological conditions for

highest sensitivity.

2

2.5

3

3.5

4

4.5

5

5.5

6

6.5

Me

an

Ro

ug

hn

ess

Act

ua

l

2.0 2.5 3.0 3.5 4.0 4.5 5.0 5.5 6.0 6.5

Mean Roughness Predi cted P<.0001

RSq=0.75 RMSE=0.4609


RSquare

RSquare Adj


Mean of Response


0.746553

0.729272

0.46086

4.243979

142

Summary of Fit

Model

Error

C. Total

Source

9

132

141

DF

82.58211

28.03579

110.61790

Sum of Squares

9.17579

0.21239

Mean Square

43.2021

F Ratio

<.0001

Prob > F


Lack Of Fit

Pure Error

Total Error

Source

5

127

132

DF

9.611127

18.424668

28.035795

Sum of Squares

1.92223

0.14508

Mean Square

13.2498

F Ratio

<.0001

Prob > F

0.8334

Max RSq

Lack Of Fit

Intercept










Term

4.2663565

0.7709168

-0.231355

-0.041969

-0.080681

0.1065694

-0.139403

-0.024359

-0.006368

0.0063131

Estimate

0.108316

0.041569

0.041569

0.042661

0.054313

0.054313

0.054313

0.050573

0.050573

0.051564

Std Error

39.39

18.55

-5.57

-0.98

-1.49

1.96

-2.57

-0.48

-0.13

0.12

t Rati o

<.0001

<.0001

<.0001

0.3270

0.1398

0.0519

0.0114

0.6308

0.9000

0.9027

Prob>|t|

Parameter Estimates










Source

1

1

1

1

1

1

1

1

1

Nparm

1

1

1

1

1

1

1

1

1

DF

73.047982

6.578877

0.205558

0.468673

0.817707

1.399186

0.049276

0.003367

0.003184

Sum of Squares

343.9294

30.9751

0.9678

2.2066

3.8500

6.5877

0.2320

0.0159

0.0150

F Ratio

<.0001

<.0001

0.3270

0.1398

0.0519

0.0114

0.6308

0.9000

0.9027

Prob > F

Effect Tests

-1.5

-1.0

-0.5

0.0

0.5

1.0

1.5

Me

an

Ro

ug

hn

ess

Re

sid

ual

2.0 2.5 3.0 3.5 4.0 4.5 5.0 5.5 6.0 6.5

Mean Roughness Predi cted


Whole Model

2.0

2.5

3.0

3.5

4.0

4.5

5.0

5.5

6.0

6.5

Me

an

Ro

ug

hn

ess

Leve

rage

Re

sid

uals

-1.5 -1.0 -0.5 .0 .5 1.0 1.5

Thi ckness of


P<.0001

Leverage Plot


2.0

2.5

3.0

3.5

4.0

4.5

5.0

5.5

6.0

6.5

Me

an

Ro

ug

hn

ess

Leve

rage

Re

sid

uals

-1.5 -1.0 -0.5 .0 .5 1.0 1.5

Substrate


Leverage, P<.0001

Leverage Plot


2.0

2.5

3.0

3.5

4.0

4.5

5.0

5.5

6.0

6.5

Me

an

Ro

ug

hn

ess

Leve

rage

Re

sid

uals

-2.0 -1.5 -1.0 -0.5 .0 .5 1.0 1.5 2.0


P=0.3270

Leverage Plot


2.0

2.5

3.0

3.5

4.0

4.5

5.0

5.5

6.0

6.5

Me

an

Ro

ug

hn

ess

Leve

rage

Re

sid

uals

4.164.184.204.224.244.264.284.304.32

Thi ckness of



Leverage Plot


2.0

2.5

3.0

3.5

4.0

4.5

5.0

5.5

6.0

6.5

Me

an

Ro

ug

hn

ess

Leve

rage

Re

sid

uals

4.15 4.20 4.25 4.30 4.35

Thi ckness of



Leverage Plot


2.0

2.5

3.0

3.5

4.0

4.5

5.0

5.5

6.0

6.5

Me

an

Ro

ug

hn

ess

Leve

rage

Re

sid

uals

4.10 4.15 4.20 4.25 4.30 4.35 4.40

Substrate



Leverage Plot


2.0

2.5

3.0

3.5

4.0

4.5

5.0

5.5

6.0

6.5

Me

an

Ro

ug

hn

ess

Leve

rage

Re

sid

uals

4.21 4.22 4.23 4.24 4.25 4.26 4.27 4.284.29

Thi ckness of



Leverage Plot


2.0

2.5

3.0

3.5

4.0

4.5

5.0

5.5

6.0

6.5

Me

an

Ro

ug

hn

ess

Leve

rage

Re

sid

uals

4.22 4.23 4.24 4.25 4.26 4.27

Substrate



Leverage Plot


2.0

2.5

3.0

3.5

4.0

4.5

5.0

5.5

6.0

6.5

Me

an

Ro

ug

hn

ess

Leve

rage

Re

sid

uals

4.21 4.22 4.23 4.24 4.25 4.26 4.27



Leverage Plot





-0.024359

.

.

-0.080681

-0.006368

.

0.1065694

-0.139403

0.0063131

0.7709168

-0.231355

-0.041969

Coef

Thi ckness of Pentacene(140.54,259.46)Substrate Temperature(34.122,60.878)Input Current(4.951,5.249)Mean Roughness




Variabl e

416.52362

75.914608

4.509781

Crit ical Val ue


Crit ical values outsi de data range

Predi cted Value at Soluti on5.5074304

Solution

Response Surface

2

2.5

3

3.5

4

4.5

5

5.5

6

6.5

Me

an

Ro

ug

hn

ess

2

2.5

3

3.5

4

4.5

5

5.5

6

6.5

Me

an

Ro

ug

hn

ess

2

2.5

3

3.5

4

4.5

5

5.5

6

6.5

Me

an

Ro

ug

hn

ess


34.122

60.878

4.9515.249

150 200 250

140.54

259.46


4.9515.249

40 50 60

140.54

259.46

34.122

60.878

Input Current

5 5.1 5.2 5.3

Thick

ne

ss o

f Pe

nta

cene

Sub

stra

te T

em

pe

ratu

reIn

put C

urre

nt


Response Mean Roughness

45

a)

b)

Figure 25. AFM images of samples from 2 different experiments with the lowest (a) and highest (b) surface roughness parameters. The lowest extracted roughness value was 2.208nm while the highest was 6.127nm.

3. Pentacene Grain Size

Grain boundaries are important topological features where DNA can immobilize.

Therefore, grain size is an important parameter to characterize. Atomic Force Microscopy

(AFM) was used to measure grain size. After extracting AFM images from tapping mode scans,

the grain size was measured by boxing the area around an average-size grain in a particular scan.

46

Keeping the systematic variation in consideration, this measurement technique is valid to make a

general comparison across the different experiments. Figure 26 below, a derivation from the

leverage plots discussed earlier, shows how the input parameters affect the grain size.

Figure 26. Relationship between the input parameters and grain size.

To begin with, the thickness of pentacene, as expected, does not affect the grain size. This is true

in Figure 26 above as well as in the interactive profile shown below in Figure 27. As substrate

temperature increases, the grain size also increases (which in turn leads to better electrical

performance). The increase in grain size can be attributed to Oswaldt Ripening effects [25].

Oswaldt Ripening is an observed phenomenon in solid (or liquid) solutions which describes the

evolution of an inhomogeneous structure over time. When a phase precipitates out of a solid,

energetic factors will cause large precipitates to grow, drawing material from the smaller

precipitates, which shrink. With higher substrate temperature, this thermodynamically-driven

spontaneous process occurs because larger particles are more energetically favored than smaller

particles. This stems from the fact that molecules on the surface of a particle are energetically

less stable than the ones already well ordered and packed in the interior. Large particles, with

their lower surface to volume ratio, results in a lower energy state (and have a lower surface

0

1

2

3

4

5

6

7

8

9

Gra

in S

ize A

ctu

al

0 1 2 3 4 5 6 7 8 9

Grain Size Predicted P<.0001 RSq=0.61

RMSE=1.0864


RSquare

RSquare Adj


Mean of Response


0.608112

0.581392

1.086427

2.150019

142

Summary of Fit

Model

Error

C. Total

Source

9

132

141

DF

241.76662

155.80280

397.56942

Sum of Squares

26.8630

1.1803

Mean Square

22.7590

F Ratio

<.0001

Prob > F


Lack Of Fit

Pure Error

Total Error

Source

5

127

132

DF

116.38324

39.41955

155.80280

Sum of Squares

23.2766

0.3104

Mean Square

74.9916

F Ratio

<.0001

Prob > F

0.9008

Max RSq

Lack Of Fit

Intercept










Term

1.0581313

-0.170463

0.8965621

-0.355127

0.0342315

0.157945

0.4270159

-0.030844

0.2693819

1.0824579

Estimate

0.255342

0.097995

0.097995

0.100568

0.128037

0.128037

0.128037

0.119221

0.119221

0.121556

Std Error

4.14

-1.74

9.15

-3.53

0.27

1.23

3.34

-0.26

2.26

8.90

t Rati o

<.0001

0.0843

<.0001

0.0006

0.7896

0.2195

0.0011

0.7963

0.0255

<.0001

Prob>|t|

Parameter Estimates










Source

1

1

1

1

1

1

1

1

1

Nparm

1

1

1

1

1

1

1

1

1

DF

3.571505

98.799308

14.717925

0.084369

1.796158

13.128664

0.079002

6.026082

93.598258

Sum of Squares

3.0259

83.7052

12.4694

0.0715

1.5217

11.1229

0.0669

5.1054

79.2988

F Ratio

0.0843

<.0001

0.0006

0.7896

0.2195

0.0011

0.7963

0.0255

<.0001

Prob > F

Effect Tests

-2

-1

0

1

2

3

4

Gra

in S

ize R

esi

du

al

0 1 2 3 4 5 6 7 8 9

Grain Size Predicted


Whole Model

0

1

2

3

4

5

6

7

8

9

Gra

in S

ize L

eve

rag

e R

esi

du

als

-1.5 -1.0 -0.5 .0 .5 1.0 1.5

Thi ckness of


P=0.0843

Leverage Plot


-1

0

1

2

3

4

5

6

7

8

9

Gra

in S

ize L

eve

rag

e R

esi

du

als

-1.5 -1.0 -0.5 .0 .5 1.0 1.5

Substrate


Leverage, P<.0001

Leverage Plot


0

1

2

3

4

5

6

7

8

9

Gra

in S

ize L

eve

rag

e R

esi

du

als

-2.0 -1.5 -1.0 -0.5 .0 .5 1.0 1.5 2.0


P=0.0006

Leverage Plot


0

1

2

3

4

5

6

7

8

9

Gra

in S

ize L

eve

rag

e R

esi

du

als

2.12 2.13 2.14 2.15 2.16 2.17 2.18

Thi ckness of



Leverage Plot


0

1

2

3

4

5

6

7

8

9

Gra

in S

ize L

eve

rag

e R

esi

du

als

2.00 2.05 2.10 2.15 2.20 2.25 2.30

Thi ckness of



Leverage Plot


0

1

2

3

4

5

6

7

8

9

Gra

in S

ize L

eve

rag

e R

esi

du

als

1.7 1.8 1.9 2.0 2.1 2.2 2.3 2.4 2.5 2.6

Substrate



Leverage Plot


0

1

2

3

4

5

6

7

8

9

Gra

in S

ize L

eve

rag

e R

esi

du

als

2.112.122.132.142.152.162.172.182.19

Thi ckness of



Leverage Plot


0

1

2

3

4

5

6

7

8

9

Gra

in S

ize L

eve

rag

e R

esi

du

als

1.7 1.8 1.9 2.0 2.1 2.2 2.3 2.4 2.5

Substrate



Leverage Plot


0

1

2

3

4

5

6

7

8

9

Gra

in S

ize L

eve

rag

e R

esi

du

als

.5 1.0 1.5 2.0 2.5 3.0 3.5


Current(4.951,5.249) Leverage, P<.0001

Leverage Plot


Gra

in S

ize

8.514

0.007

1.051445


140

.54

259

.46

200


34.1

22

60.8

78

47.4

Input Current

4.9

51

5.2

49

5.1

0

1

2

3

4

5

6

7

8

9

Gra

in S

ize

0

1

2

3

4

5

6

7

8

9

Gra

in S

ize

0

1

2

3

4

5

6

7

8

9

Gra

in S

ize


34.122

60.878

4.9515.249

150 200 250

140.54259.46


4.9515.249

40 50 60

140.54259.46

34.122

60.878

Input Current

5 5.1 5.2 5.3

Thick

ne

ss o

f Pe

nta

cene

Sub

stra

te T

em

pe

ratu

reIn

put C

urre

nt


Prediction Profiler

Response Grain Size

47

energy). As the system tries to lower its overall energy, molecules on the surface of a small

(energetically unfavorable) particle will tend to diffuse through solution and add to the surface of

larger particle. Therefore, the smaller particles continue to shrink, while larger particles continue

to grow [27], leading to the larger grain size with higher substrate temperatures.

The input current shows an interesting non-linear relationship with grain size. The expected

relationship here would be that as input current decreases, the grain size would increase linearly

because of the way the grains are formed. When the pentacene crystals hit the substrate as they

are being evaporated, they grow till they hit another crystal around them. With lower

evaporation rate, the crystals have „time‟ to grow bigger on the surface. This linear relationship

is not seen because of two possible reasons. One is that there are more factors/variables

interacting that we are not aware of and/or we will see in the interaction plot below. The other

reason could be the way the grain size was measured. Since the grain size was measured as an

average measurement from sample to sample in AFM, a more precise way of measuring the grain

size may need to be implemented to get a clearer relationship between input current and grain

size.

Figure 27 addresses the interaction among the different parameters. The main interaction

happens between substrate temperature and input current, which might help better explain the

non-linear relationship being seen between input current and grain size.

48

Figure 27. Interaction Profiler Plot for Grain Size.

Again, since grain size and grain boundaries are an important part of sensitivity of our sensor,

the range of our grain size from 8.514 m2

down to 0.314 m2, illustrated in Figure 28,

emphasizes the importance of tuning the morphology to suit the necessary conditions for DNA

interaction. This will be discussed in depth in the following sections.

Figure 28. AFM images of samples from 2 different experiments consisting of the smallest grain size (left) of 0.314

m 2 and the biggest grain size (right) of 8.514 m

2.

0

1

2

3

4

5

6

7

8

9

Gra

in S

ize A

ctu

al

0 1 2 3 4 5 6 7 8 9

Grain Size Predicted P<.0001 RSq=0.61

RMSE=1.0864


RSquare

RSquare Adj


Mean of Response


0.608112

0.581392

1.086427

2.150019

142

Summary of Fit

Model

Error

C. Total

Source

9

132

141

DF

241.76662

155.80280

397.56942

Sum of Squares

26.8630

1.1803

Mean Square

22.7590

F Ratio

<.0001

Prob > F


Lack Of Fit

Pure Error

Total Error

Source

5

127

132

DF

116.38324

39.41955

155.80280

Sum of Squares

23.2766

0.3104

Mean Square

74.9916

F Ratio

<.0001

Prob > F

0.9008

Max RSq

Lack Of Fit

Intercept










Term

1.0581313

-0.170463

0.8965621

-0.355127

0.0342315

0.157945

0.4270159

-0.030844

0.2693819

1.0824579

Estimate

0.255342

0.097995

0.097995

0.100568

0.128037

0.128037

0.128037

0.119221

0.119221

0.121556

Std Error

4.14

-1.74

9.15

-3.53

0.27

1.23

3.34

-0.26

2.26

8.90

t Rati o

<.0001

0.0843

<.0001

0.0006

0.7896

0.2195

0.0011

0.7963

0.0255

<.0001

Prob>|t|

Parameter Estimates










Source

1

1

1

1

1

1

1

1

1

Nparm

1

1

1

1

1

1

1

1

1

DF

3.571505

98.799308

14.717925

0.084369

1.796158

13.128664

0.079002

6.026082

93.598258

Sum of Squares

3.0259

83.7052

12.4694

0.0715

1.5217

11.1229

0.0669

5.1054

79.2988

F Ratio

0.0843

<.0001

0.0006

0.7896

0.2195

0.0011

0.7963

0.0255

<.0001

Prob > F

Effect Tests

-2

-1

0

1

2

3

4

Gra

in S

ize R

esi

du

al

0 1 2 3 4 5 6 7 8 9

Grain Size Predicted


Whole Model

0

1

2

3

4

5

6

7

8

9

Gra

in S

ize L

eve

rag

e R

esi

du

als

-1.5 -1.0 -0.5 .0 .5 1.0 1.5

Thi ckness of


P=0.0843

Leverage Plot


-1

0

1

2

3

4

5

6

7

8

9

Gra

in S

ize L

eve

rag

e R

esi

du

als

-1.5 -1.0 -0.5 .0 .5 1.0 1.5

Substrate


Leverage, P<.0001

Leverage Plot


0

1

2

3

4

5

6

7

8

9

Gra

in S

ize L

eve

rag

e R

esi

du

als

-2.0 -1.5 -1.0 -0.5 .0 .5 1.0 1.5 2.0


P=0.0006

Leverage Plot


0

1

2

3

4

5

6

7

8

9

Gra

in S

ize L

eve

rag

e R

esi

du

als

2.12 2.13 2.14 2.15 2.16 2.17 2.18

Thi ckness of



Leverage Plot


0

1

2

3

4

5

6

7

8

9

Gra

in S

ize L

eve

rag

e R

esi

du

als

2.00 2.05 2.10 2.15 2.20 2.25 2.30

Thi ckness of



Leverage Plot


0

1

2

3

4

5

6

7

8

9

Gra

in S

ize L

eve

rag

e R

esi

du

als

1.7 1.8 1.9 2.0 2.1 2.2 2.3 2.4 2.5 2.6

Substrate



Leverage Plot


0

1

2

3

4

5

6

7

8

9

Gra

in S

ize L

eve

rag

e R

esi

du

als

2.112.122.132.142.152.162.172.182.19

Thi ckness of



Leverage Plot


0

1

2

3

4

5

6

7

8

9

Gra

in S

ize L

eve

rag

e R

esi

du

als

1.7 1.8 1.9 2.0 2.1 2.2 2.3 2.4 2.5

Substrate



Leverage Plot


0

1

2

3

4

5

6

7

8

9

Gra

in S

ize L

eve

rag

e R

esi

du

als

.5 1.0 1.5 2.0 2.5 3.0 3.5


Current(4.951,5.249) Leverage, P<.0001

Leverage Plot


Gra

in S

ize

8.514

0.007

1.051445


140

.54

259

.46

200


34.1

22

60.8

78

47.4

Input Current

4.9

51

5.2

49

5.1

0

1

2

3

4

5

6

7

8

9

Gra

in S

ize

0

1

2

3

4

5

6

7

8

9

Gra

in S

ize

0

1

2

3

4

5

6

7

8

9

Gra

in S

ize


34.122

60.878

4.9515.249

150 200 250

140.54259.46


4.9515.249

40 50 60

140.54259.46

34.122

60.878

Input Current

5 5.1 5.2 5.3

Thick

ne

ss o

f Pe

nta

cene

Sub

stra

te T

em

pe

ratu

reIn

put C

urre

nt


Prediction Profiler

Response Grain Size

5 m 5 m

5 m 5 m

49

4 & 5. Evaporation Rate and Coverage Area

The last two output parameters that were analyzed were evaporation rate and coverage area.

Evaporation rate was calculated using the data from the crystal monitor within the pentacene

evaporator. The coverage area was calculated from the images scanned using AFM. According

to the relationships seen in Figure 29 below, evaporation rate is affected only by the input

current, and the coverage area does not show any strong relationships with the input factors.

Figure 29. Relationship between the input parameters and evaporation rate and coverage area.

The higher the input current, the higher the temperature of the crucible holding the pentacene

for evaporation. This in turn will cause a higher evaporation rate. For coverage area, none of the

input factors show a statistically significant effect since the slopes are all smaller than the error

bars. This is most likely due to some noise from the data extracted. One way to improve this

0

5

10

15

20

Eva

p R

ate

Actu

al

0 5 10 15 20

Evap Rate Predicted P=0.0836 RSq=0.83

RMSE=2.9697


Intercept










Term

5.0650119

1.4729952

0.5291968

3.4397451

1.219993

2.067977

0.8805256

-0.239633

-0.472243

0.5363697

Estimate

2.093768

0.803606

0.803606

0.803606

1.049961

1.049961

1.049961

0.975699

0.975699

0.975699

Std Error

2.42

1.83

0.66

4.28

1.16

1.97

0.84

-0.25

-0.48

0.55

t Rati o

0.0519

0.1165

0.5346

0.0052

0.2894

0.0964

0.4338

0.8142

0.6455

0.6024

Prob>|t|

Parameter Estimates










Source

1

1

1

1

1

1

1

1

1

Nparm

1

1

1

1

1

1

1

1

1

DF

29.63148

3.82459

161.58580

11.90706

34.21223

6.20260

0.53198

2.06602

2.66522

Sum of Squares

3.3598

0.4337

18.3217

1.3501

3.8792

0.7033

0.0603

0.2343

0.3022

F Ratio

0.1165

0.5346

0.0052

0.2894

0.0964

0.4338

0.8142

0.6455

0.6024

Prob > F

Effect Tests

-4

-3

-2

-1

0

1

2

3

4

Eva

p R

ate

Re

sid

ua

l

0 5 10 15 20

Evap Rate Predicted


Whole Model

0

5

10

15

20

Eva

p R

ate

Le

vera

ge R

esid

ua

ls

-1.5 -1.0 -0.5 .0 .5 1.0 1.5

Thi ckness of


P=0.1165

Leverage Plot


0

5

10

15

20

Eva

p R

ate

Le

vera

ge R

esid

ua

ls

-1.5 -1.0 -0.5 .0 .5 1.0 1.5

Substrate


Leverage, P=0.5346

Leverage Plot


0

5

10

15

20

Eva

p R

ate

Le

vera

ge R

esid

ua

ls

-1.5 -1.0 -0.5 .0 .5 1.0 1.5


P=0.0052

Leverage Plot


0

5

10

15

20

Eva

p R

ate

Le

vera

ge R

esid

ua

ls

3.5 4.0 4.5 5.0 5.5 6.0

Thi ckness of



Leverage Plot


0

5

10

15

20

Eva

p R

ate

Le

vera

ge R

esid

ua

ls

3 4 5 6 7

Thi ckness of



Leverage Plot


0

5

10

15

20

Eva

p R

ate

Le

vera

ge R

esid

ua

ls

4.0 4.5 5.0 5.5 6.0

Substrate



Leverage Plot


0

5

10

15

20

Eva

p R

ate

Le

vera

ge R

esid

ua

ls

4.6 4.7 4.8 4.9 5.0 5.1 5.2 5.3

Thi ckness of



Leverage Plot


0

5

10

15

20

Eva

p R

ate

Le

vera

ge R

esid

ua

ls

4.3 4.5 4.7 4.9 5.1 5.3 5.5 5.7

Substrate



Leverage Plot


0

5

10

15

20

Eva

p R

ate

Le

vera

ge R

esid

ua

ls

4.0 4.2 4.4 4.6 4.8 5.0 5.2 5.4 5.6 5.8



Leverage Plot


Response Evap Rate

0.91

0.92

0.93

0.94

0.95

0.96

0.97

0.98

0.99

1

Cove

rage

Are

a A

ctu

al

.91 .92 .93 .94 .95 .96 .97 .98 .991.00

Coverage Area Predi cted P=0.7170

RSq=0.50 RMSE=0.0251


Intercept










Term

0.9881251

-0.00136

-0.007235

-0.001655

0.0056331

-0.015849

0.007974

-0.003956

-0.001823

-0.004648

Estimate

0.017694

0.006791

0.006791

0.006791

0.008873

0.008873

0.008873

0.008245

0.008245

0.008245

Std Error

55.85

-0.20

-1.07

-0.24

0.63

-1.79

0.90

-0.48

-0.22

-0.56

t Rati o

<.0001

0.8479

0.3277

0.8156

0.5489

0.1243

0.4034

0.6483

0.8323

0.5934

Prob>|t|

Parameter Estimates










Source

1

1

1

1

1

1

1

1

1

Nparm

1

1

1

1

1

1

1

1

1

DF

0.00002524

0.00071484

0.00003740

0.00025386

0.00200961

0.00050867

0.00014501

0.00003080

0.00020016

Sum of Squares

0.0401

1.1350

0.0594

0.4031

3.1908

0.8077

0.2302

0.0489

0.3178

F Ratio

0.8479

0.3277

0.8156

0.5489

0.1243

0.4034

0.6483

0.8323

0.5934

Prob > F

Effect Tests

-0.03

-0.02

-0.01

0.00

0.01

0.02

Cove

rage

Are

a R

esid

ua

l

.91 .92 .93 .94 .95 .96 .97 .98 .991.00

Coverage Area Predi cted


Whole Model

0.91

0.92

0.93

0.94

0.95

0.96

0.97

0.98

0.99

1.00

1.01

Cove

rage

Are

a L

evera

ge

Resid

ua

ls

-1.5 -1.0 -0.5 .0 .5 1.0 1.5

Thi ckness of


P=0.8479

Leverage Plot


0.91

0.92

0.93

0.94

0.95

0.96

0.97

0.98

0.99

1.00

1.01

Cove

rage

Are

a L

evera

ge

Resid

ua

ls

-1.5 -1.0 -0.5 .0 .5 1.0 1.5

Substrate


Leverage, P=0.3277

Leverage Plot


0.91

0.92

0.93

0.94

0.95

0.96

0.97

0.98

0.99

1.00

1.01

Cove

rage

Are

a L

evera

ge

Resid

ua

ls

-1.5 -1.0 -0.5 .0 .5 1.0 1.5


P=0.8156

Leverage Plot


0.91

0.92

0.93

0.94

0.95

0.96

0.97

0.98

0.99

1.00

1.01

Cove

rage

Are

a L

evera

ge

Resid

ua

ls

.973 .975 .977 .979 .981 .983 .985

Thi ckness of



Leverage Plot


0.91

0.92

0.93

0.94

0.95

0.96

0.97

0.98

0.99

1.00

1.01

Cove

rage

Are

a L

evera

ge

Resid

ua

ls

.965 .970 .975 .980 .985 .990 .995

Thi ckness of



Leverage Plot


0.91

0.92

0.93

0.94

0.95

0.96

0.97

0.98

0.99

1.00

1.01

Cove

rage

Are

a L

evera

ge

Resid

ua

ls

.970.972.974.976.978.980.982.984.986

Substrate



Leverage Plot


0.91

0.92

0.93

0.94

0.95

0.96

0.97

0.98

0.99

1.00

1.01

Cove

rage

Are

a L

evera

ge

Resid

ua

ls

.974 .976 .978 .980 .982 .984 .986

Thi ckness of



Leverage Plot


0.91

0.92

0.93

0.94

0.95

0.96

0.97

0.98

0.99

1.00

Cove

rage

Are

a L

evera

ge

Resid

ua

ls

.977 .978 .979 .980 .981 .982

Substrate



Leverage Plot


0.91

0.92

0.93

0.94

0.95

0.96

0.97

0.98

0.99

1.00

1.01

Cove

rage

Are

a L

evera

ge

Resid

ua

ls

.973 .975 .977 .979 .981 .983 .985 .987



Leverage Plot


Response Coverage Area

Eva

p R

ate

18.3905

-2.0682

5.065012

Cove

rage

Are

a

1.03241

0.91491

0.988125


140

.54

259

.46

200


34.1

22

60.8

78

47.5

Input Current

4.9

51

5.2

49

5.1

0

5

10

15

20

Eva

p R

ate

0

5

10

15

20

Eva

p R

ate

0

5

10

15

20

Eva

p R

ate


34.12260.878

4.951

5.249

150 200 250

140.54

259.46


4.951

5.249

40 50 60

140.54

259.46

34.12260.878

Input Current

5 5.1 5.2 5.3

Thick

ne

ss o

f Pe

nta

cene

Sub

stra

te T

em

pe

ratu

reIn

put C

urre

nt

0.91

0.94

0.96

0.99

Cove

rage

Are

a

0.91

0.94

0.96

0.99

Cove

rage

Are

a

0.91

0.94

0.96

0.99

Cove

rage

Are

a


34.12260.878

4.951

5.249

150 200 250

140.54259.46


4.9515.249

40 50 60

140.54

259.46

34.12260.878

Input Current

5 5.1 5.2 5.3

Thick

ne

ss o

f Pe

nta

cene

Sub

stra

te T

em

pe

ratu

reIn

put C

urre

nt


Prediction Profiler

Least Squares Fit

50

noise would be to use a more precise and statistically accurate tool to measure coverage area.

The purpose of using a contrast tool to obtain coverage area data was to get a general idea of the

trend across all of the experiments in the DOE.

Similar to the figures shown previously, Figure 30 shows the lowest and highest coverage

area from samples of 2 different experiments. The evaporation rate, depending on the input

current, ranged from 0.838 Å /min to 18.391 Å /min.

The left image with the lowest coverage area exposes a lot of the dielectric surface. The stacking

of the image on the right is better, and coverage area is much higher since the dielectric surface is

completely covered by the pentacene crystals.

Sections 1-5 presented thus far have highlighted the importance of the relationship of the

different input parameters and their effect on the morphological and electrical characteristics of

pentacene TFTs. The following section takes this a step further and integrates the DNA

immobilization experiment into the data obtained so far, to arrive at the optimum conditions for

highest sensitivity.

5 m

5 m

5 m

5 m

Figure 30. AFM Images of samples from 2 different experiments that resulted in the lowest (left) and highest (right) coverage area. The yellow area in the images represents the pentacene crystals, and the dark red area is the SiO2 (dielectric) surface.

51

B. Optimization of DNA Immobilization and Sensor Sensitivity

The purpose of this part of the experiment is to determine pentacene evaporation conditions

that facilitate optimum DNA immobilization and highest sensitivity.

i. Experiment Protocol

The experiment protocol that was followed is enlisted below. For each of the experiments in the

DOE, the following steps are followed:

1) Pre-measurements of electrical characteristics of transistors on substrate (153 transistors from

all of the 17 experiments) using the probe station (same set up as described on page 37)

2) Pipetting of 1.5 L of DNA (125 Base-Pairs single strand DNA sequence synthesized

by BioSynthesis, Inc) on channels of pre-measured transistors; the concentration of DNA

used in this experiment was 0.5 g/ L

3) Immobilization of DNA on pentacene surface (in air) until buffer solution is dry

4) Washing of sample for 1-2 minutes with DI Water

5) Drying of substrate (gently)

6) Storage of transistors for a few hours in nitrogen (to neutralized any buffer solution and DI

water effects)

7) Post-Measurements of electrical characteristics

8) Calculation of sensitivity of particular substrate using Ion ratio and threshold voltage shifts

The storage of the sample in nitrogen (step 6) before the post-measurement of electrical

characteristics is important in order to neutralize the effects that buffer and DI water have on the

pentacene surface. This was concluded based on control experiments performed using different

52

protocols on the similar substrates. The control experiment performed and the results obtained

are discussed in the following section.

ii. Control Experiments

Since pentacene is moisture sensitive, control experiments had to be performed to make sure

that the effects of buffer solution and water on the pentacene transistors do not confound the

DNA immobilization results. Dr. Qintao Zhang had previous shown results from control

experiments he had performed using buffer solution and DI water on pentacene. The set of

control experiments done this time around was to figure out what step in the process actually

neutralized the effects. A few protocols were followed, and the one presented in the previous

section produced the desired results. The results from the control experiments performed using

this protocol is shown in Figure 31 below. Idsat ratio of 1 means that the electrical characteristics

have not been affected.

Idsat

ratio

s

0.5

1

1.5

2

2.5

3

3.5

4

buffer di water ssDNA

Type of Experiment Figure 31. Control Experiments performed to neutralize buffer solution and DI water effects on pentacene TFTs.

53

The storage in nitrogen neutralized the buffer solution and DI water effects. Further experiments

need to be performed to understand why the storage in nitrogen helps in this process so that the

storage time and atmosphere can be optimized.

iii. Results: Optimization of DNA Immobilization and Sensor Sensitivity

This section presents the results from integrating the pentacene characterization experiment

with DNA. To review, the Idsat ratio is defined as follows:

Idsat Ratio = onmobilizatiBeforeDNAI

onmobilizatiAfterDNAntDrainCurreI

d

d

Im

Im)(

As with previous experiments, JMP leverages the 153 points that were measured before and

after DNA immobilization (for Idsat Ratio), to produce the graphs and relationships in Figure 32

below.

Figure 32. Relationship between input parameters and Idsat ratios.

1

2

3

4

5

6

DN

A-I

dsa

t_R

atios A

ctua

l

1 2 3 4 5 6

DNA-Idsat_Rati os Predi cted P<.0001

RSq=0.56 RMSE=0.6127


RSquare

RSquare Adj


Mean of Response


0.563176

0.533617

0.612651

2.547461

143

Summary of Fit

Model

Error

C. Total

Source

9

133

142

DF

64.35996

49.92033

114.28029

Sum of Squares

7.15111

0.37534

Mean Square

19.0523

F Ratio

<.0001

Prob > F


Lack Of Fit

Pure Error

Total Error

Source

5

128

133

DF

13.668493

36.251841

49.920335

Sum of Squares

2.73370

0.28322

Mean Square

9.6523

F Ratio

<.0001

Prob > F

0.6828

Max RSq

Lack Of Fit

Intercept










Term

2.2922838

-0.382604

0.4274611

-0.252851

-0.360626

-0.066864

-0.1492

0.0255407

0.2273335

0.0428153

Estimate

0.143982

0.055503

0.055503

0.055503

0.072741

0.072741

0.072741

0.067126

0.067126

0.067126

Std Error

15.92

-6.89

7.70

-4.56

-4.96

-0.92

-2.05

0.38

3.39

0.64

t Rati o

<.0001

<.0001

<.0001

<.0001

<.0001

0.3597

0.0422

0.7042

0.0009

0.5247

Prob>|t|

Parameter Estimates










Source

1

1

1

1

1

1

1

1

1

Nparm

1

1

1

1

1

1

1

1

1

DF

17.835760

22.263095

7.789717

9.225263

0.317140

1.579068

0.054338

4.304920

0.152699

Sum of Squares

47.5188

59.3143

20.7537

24.5784

0.8449

4.2070

0.1448

11.4694

0.4068

F Ratio

<.0001

<.0001

<.0001

<.0001

0.3597

0.0422

0.7042

0.0009

0.5247

Prob > F

Effect Tests

-2.0

-1.5

-1.0

-0.5

0.0

0.5

1.0

1.5

2.0

DN

A-I

dsa

t_R

atios R

esi

du

al

1 2 3 4 5 6

DNA-Idsat_Rati os Predi cted


Whole Model

1

2

3

4

5

6

DN

A-I

dsa

t_R

atios L

eve

rage

Re

sid

uals

-1.5 -1.0 -0.5 .0 .5 1.0 1.5

Thi ckness of


P<.0001

Leverage Plot


1

2

3

4

5

6

DN

A-I

dsa

t_R

atios L

eve

rage

Re

sid

uals

-1.5 -1.0 -0.5 .0 .5 1.0 1.5

Substrate


Leverage, P<.0001

Leverage Plot


1

2

3

4

5

6

DN

A-I

dsa

t_R

atios L

eve

rage

Re

sid

uals

-1.5 -1.0 -0.5 .0 .5 1.0 1.5


P<.0001

Leverage Plot


1

2

3

4

5

6

DN

A-I

dsa

t_R

atios L

eve

rage

Re

sid

uals

2.1 2.2 2.3 2.4 2.5 2.6 2.7 2.8 2.9 3.0

Thi ckness of



Leverage Plot


1

2

3

4

5

6

DN

A-I

dsa

t_R

atios L

eve

rage

Re

sid

uals

2.462.482.502.522.542.562.582.602.62

Thi ckness of



Leverage Plot


1

2

3

4

5

6

DN

A-I

dsa

t_R

atios L

eve

rage

Re

sid

uals

2.40 2.45 2.50 2.55 2.60 2.65 2.70

Substrate



Leverage Plot


1

2

3

4

5

6

DN

A-I

dsa

t_R

atios L

eve

rage

Re

sid

uals

2.502.512.522.532.542.552.562.572.58

Thi ckness of



Leverage Plot


1

2

3

4

5

6

DN

A-I

dsa

t_R

atios L

eve

rage

Re

sid

uals

2.2 2.3 2.4 2.5 2.6 2.7 2.8

Substrate



Leverage Plot


1

2

3

4

5

6

DN

A-I

dsa

t_R

atios L

eve

rage

Re

sid

uals

2.48 2.50 2.52 2.54 2.56 2.58 2.60



Leverage Plot


DN

A-I

dsa

t_R

atios

5.90977

1.04594

2.292284


140

.54

259

.46

200


34.1

22

60.8

78

47.5

Input Current

4.9

51

5.2

49

5.1

1

2

3

4

5

6

DN

A-I

dsa

t_R

atios

1

2

3

4

5

6

DN

A-I

dsa

t_R

atios

1

2

3

4

5

6

DN

A-I

dsa

t_R

atios


34.12260.878

4.951

5.249

150 200 250

140.54

259.46


4.951

5.249

40 50 60

140.54

259.46

34.122

60.878

Input Current

5 5.1 5.2 5.3

Thick

ne

ss o

f Pe

nta

cene

Sub

stra

te T

em

pe

ratu

reIn

put C

urre

nt


Prediction Profiler

Response DNA-Idsat_Ratios

54

The input parameters have already been related to the morphological and electrical

characteristics. In this section, we relate the input parameters to sensor sensitivity. To begin

with, as the thickness of pentacene increases, the Idsat ratio or sensitivity goes down. Thinner

films allow the DNA to be immobilized in the channel part of the film and affect the

performance more. The channel is formed in the first few monolayers (monolayer = 15 Å ) of the

pentacene film. Although thicker pentacene might immobilize more DNA because of higher

surface roughness, the sensitivity will not be high because not all of the DNA will be

immobilized in the channel part of the pentacene film to actually cause an effect.

As substrate temperature increases and input current decreases, the sensitivity increases.

Higher substrate temperature and lower input current (lower deposition rate) lead to bigger grains

on the surface. At thin film phases, the pentacene film growth produces a terrace like structure

shown in an AFM image in Figure 33.

Figure 33. AFM Image of terrace-like formation of pentacene thin film on SiO2 substrate [27].

Because of the terrace formation, the bigger the grain size, the more exposure there is to the first

few monolayers of the film. In other words, bigger size allows for a wider area of exposure of

55

the first few monolayers of the film, where the channel forms. This is in contrast to smaller grain

size, where even with the terrace formation, the channel part of the film is not widely exposed.

Therefore, the bigger the grain size, the more area (in the channel part of the film) the DNA has

to immobilize. This agrees very well with the data seen in Figure 32. With lower input current

and higher substrate temperature, bigger grains are grown, allowing the DNA to more effectively

immobilize and affect the channel part of the film, thereby causing the highest electrical shift and

giving the highest sensitivity. The sensitivity values range from 1.0456 up to 5.909. Figure 34

takes this analysis a step further and shows how the different input parameters interact together to

affect the sensitivity.

1

2

3

4

5

6

DN

A-I

dsa

t_R

atios A

ctu

al

1 2 3 4 5 6

DNA-Idsat_Rati os Predi cted P<.0001

RSq=0.59 RMSE=0.6013


RSquare

RSquare Adj


Mean of Response


0.587339

0.548652

0.601328

2.553529

141

Summary of Fit

Model

Error

C. Total

Source

12

128

140

DF

65.87610

46.28418

112.16029

Sum of Squares

5.48968

0.36160

Mean Square

15.1818

F Ratio

<.0001

Prob > F


Intercept










Grain Size

Mean Roughness

Idsat

Term

2.5584276

-0.306932

0.4108072

-0.24085

-0.335803

-0.066425

-0.170855

0.0224374

0.2202644

0.0623213

0.0196847

-0.053072

4211.2101

Esti mate

0.577503

0.108987

0.071038

0.066476

0.076951

0.073853

0.074986

0.066335

0.067605

0.089048

0.053302

0.126044

3702.706

Std Error

4.43

-2.82

5.78

-3.62

-4.36

-0.90

-2.28

0.34

3.26

0.70

0.37

-0.42

1.14

t Rati o

<.0001

0.0056

<.0001

0.0004

<.0001

0.3701

0.0244

0.7357

0.0014

0.4853

0.7125

0.6744

0.2575

Prob>|t|

Parameter Estimates










Grain Size

Mean Roughness

Idsat

Source

1

1

1

1

1

1

1

1

1

1

1

1

Nparm

1

1

1

1

1

1

1

1

1

1

1

1

DF

2.867860

12.092395

4.746658

6.885906

0.292512

1.877241

0.041370

3.838481

0.177112

0.049317

0.064106

0.467733

Sum of Squares

7.9311

33.4418

13.1270

19.0431

0.8089

5.1916

0.1144

10.6154

0.4898

0.1364

0.1773

1.2935

F Ratio

0.0056

<.0001

0.0004

<.0001

0.3701

0.0244

0.7357

0.0014

0.4853

0.7125

0.6744

0.2575

Prob > F

Effect Tests

Conti nuous factors centered by mean, scaled by range/2

Intercept










Grain Size

Mean Roughness

Idsat

Term

2.2949018

-0.306932

0.4108072

-0.24085

-0.335803

-0.066425

-0.170855

0.0224374

0.2202644

0.0623213

0.0807066

-0.103994

0.1998661

Scaled Esti mate

0.153313

0.108987

0.071038

0.066476

0.076951

0.073853

0.074986

0.066335

0.067605

0.089048

0.218536

0.246984

0.175732

Std Error

14.97

-2.82

5.78

-3.62

-4.36

-0.90

-2.28

0.34

3.26

0.70

0.37

-0.42

1.14

t Rati o

<.0001

0.0056

<.0001

0.0004

<.0001

0.3701

0.0244

0.7357

0.0014

0.4853

0.7125

0.6744

0.2575

Prob>|t|

Scaled Estimates

DN

A-I

dsa

t_R

atios

5.90977

1.04594

2.294902


140

.54

259

.46

200


34.1

22

60.8

78

47.5

Input Current

4.9

51

5.2

49

5.1

Grain Size

0.3

14

044

8.5

14

2.15762

Mean Roughness

2.2

08

6.1

27

4.24957

Idsat-0.0

000

7

0.0

00

02

-1.9e-5

Prediction Profiler

1

2

3

4

5

6

DN

A-I

dsa

t_R

atios

1

2

3

4

5

6

DN

A-I

dsa

t_R

atios

1

2

3

4

5

6

DN

A-I

dsa

t_R

atios


34.12260.878

4.951

5.249

150 200 250

140.54

259.46


4.951

5.249

40 50 60

140.54

259.46

34.122

60.878

Input Current

5 5.1 5.2 5.3

Thic

kne

ss o

f Pe

nta

ce

ne

Sub

stra

te T

em

pe

ratu

reIn

put C

urre

nt


Response DNA-Idsat_Ratios

Figure 34. Interaction Profiler Plot for Idsat Ratio/Sensitivity.

56

We have studied impact of evaporation parameters on pentacene morphology, and tuned/

analyzed morphology to maximize the sensitivity of our OTFT platform. With thinner pentacene

films, higher substrate temperature, and lower input current, we are able to achieve highest

sensitivity. Results suggest that sensitivity of DNA sensor can be optimized through morphology

of the film to expose as much of channel part of the film as possible for effective DNA

immobilization.

57

V. Conclusion and Future Work

By finding reliable methods to relate the electrical transduction behavior to physical origins,

and by optimizing the pentacene film surface for highest DNA sensitivity, we have gotten closer

to our goal of making a viable DNA sensor using pentacene TFTs.

Section IV.B. discussed the experimental protocol used for the DNA optimization

experiment. The storage in nitrogen to neutralize effects of buffer solution and DI water was

followed as a result of the performed control experiments. In order to optimize the time and

atmosphere of storage in the protocol, further work needs to be done to understand the nitrogen

interaction with buffer solution and DI water and its effects on the pentacene surface. Also, to

make the DNA immobilization more stable and predictable, future work should focus on

chemically modifying the pentacene surface to covalently bind to the DNA. This way, the

orientation of the DNA is clear, and future experiments with hybridization will be more

controllable. Since the goal is to be able to use this DNA detection chip for genetic disease

detection, more work on hybridization detection needs to be performed. Particularly, more

research needs to be performed into detecting single point mutation, known as „single nucleotide

polymorphism‟ with a 20-30 base pair sequence as done in medical practice today.

58

VI. Sources

[1] Knowledge Systems Institute, Image of Biological Structure, Knowledge Systems Institute.

[Online]. Available: http://distancelearning.ksi.edu/demo/bio378/DNA_files/image023.jpg .

[Accessed: December 2, 2008].

[2] S. Molesa, “Ultra-Low-Cost printed electronics,” Ph.D. dissertation, University of California-

Berkeley, Berkeley, CA, USA, 2006.

[3] National Human Genome Research Institute, “DNA Microchip Technology,” National

Human Genome Research Institute. [Online]. Available: http://www.genome.gov/10000205.

[Accessed: November 23, 2008].

[4] M. Gabig-Ciminska, “Developing nucleic acid-based electrical detection systems,” Microbial

Cell Factories., Mar. 2006.

[5] Q. Zhang, “OTFT-Based DNA Detection System,” Ph.D. dissertation, University of

California-Berkeley, Berkeley, CA, USA, 2007.

[6] D. Gillespie, “A quantitative assay for DNA-RNA hybrids with DNA immobilized on a

membrane,” Journal of molecular biology., vol. 12, pp.829, 1965.

[7] Molecular Station, Molecular Biology Images, Molecular Station. [Online].

Available:http://www.molecularstation.com/molecular-biology-images/data/502/557px-

DNA_chemical_structure.png. [Accessed: November 23, 2008].

[8] Penn State Department of Chemistry, Image of Pentacene Structure, Penn State Elberly

College of Science. [Online]. Available: http://www.chem.psu.edu/images/

Pentacene_JPEG.jpg. [Accessed: December 7, 2008].

[9] “Atomic force microscope,” [Online]. Available: http://en.wikipedia.org/wiki/Atomic_force_

microscope. [Accessed: December 7, 2008].

[10] C. Bai, Scanning Tunneling Microscopy and Its Application. Springer, 1995.

[11] M. J. Heller, “DNA Microarray Technology: Devices, Systems, and Applications,” Annu. Rev.

Biomed. Eng., vol. 4,, pp.129, 2002.

[12] “Kelvin probe force microscope,” [Online]. Available: http://en.wikipedia.org/wiki/Kelvin_

probe_force_microscope. [Accessed: December 5, 2008].

[13] Science Education Resource Center at Carleton College, Image of Fluorescent Microscopy System, Science Education Resource Center at Carleton College. [Online}. Available: http://serc.carleton.edu/images/microbelife/research_methods/microscopy/fluorescent_filters.jpg. [Accessed: December 7, 2008].

[14] “Fluorescence”, [Online]. Available http://en.wikipedia.org/wiki/Fluorescence. [Accessed: November 23, 2008]

[15] Wikipedia, ”What is the wavelength of blue light?” Wiki Answers. [Online]. Available: http://wiki.answers.com/Q/What_is_the_wavelength_of_blue_light. [Accessed: December 8, 2008].

59

[16] Arizona State University, “Single Molecule Spectroscopy,” Arizona State University. [Online].

Available: http://www.public.asu.edu/~laserweb/woodbury/smf.html. [Accessed: December 11, 2008].

[17] Evans Analytical Group, “Time-of-Flight Secondary Ion Mass Spectrometry (TOF-SIMS),” Analytical Techniques. [Online]. Available: http://www.cea.com/techniques/analytical_ techniques/tof_sims.php. [Accessed: December 2, 2008].

[18] R. Plasun, “Optimization of VSLI Semiconductor Devices,” Ph.D. dissertation, Vienna University of Technology, Vienna, Austria, 1999.

[19] JMP Statistical Discovery Software, JMP 5.0.1a Software Manual, 1989.

[20] S. E. Fritz Vos, “Structure and Transport in Organic Semiconductor Thin Films,” Ph.D.

dissertation, University of Minnesota, USA, May 2006.

[21] G. Horowitz and M .E. Hajlaoui, “Mobility in polycrystalline oligothiophene field-effect

transistors dependent on grain size,” Advanced Materials, vol.12, no.14, pp. 1046-1050,

2000.

[22] D. Knipp, D.K. Murti, B. Krusor, R. Apte, L. Jiang, J.P. Lu, B.S. Ong, and R. Street,

“Photoconductivity of Pentacene Thin Film Transistors,” Materials Research Society

Symposium Proceedings, Vol. 665, pp. 207-212, 2002.

[23] A. J. Salih, S. P. Lau, J. M. Marshall, J. M. Maud, W. R. Bowen, N. Hilal, R. W. Lovitt, and

P. M. Williams, “Improved thin films of pentacene via pulsed laser deposition at elevated

substrate temperatures,” Appl. Phys. Letters, vol. 69, pp. 2231, 1996.

[24] T. Minakata, H. Imai, and M. Ozaki, “Electrical properties of highly ordered and

amorphous thin films of pentacene doped with iodine,” Journal of Applied Physics, vol. 72,

pp. 4178, 1992.

[25] J.W. Chang, H. Kim, J.K. Kim, B. K. Ju, J. Jang, and Y.H. Lee, “Structure and Morphology

of Vacuum-Evaporated Pentacene as a Function of the Substrate Temperature,” Journal of

the Korean Physical Society, Vol. 42, pp. S647-S651, February 2003.

[26] Wikipedia, “Ostwald ripening,” Wikipedia, [Online]. Available: http://en.wikipedia.org/wiki /Ostwald_ripening. [Accessed: December 8, 2008].

[27] I. P. M. Bouchoms, W. A. Schoonveld, J. Vrijmoeth, and T. M. Klapwijk, “Morphology

identification of the thin film phases of vacuum evaporated pentacene on SiO2 substrates,”

Synthetic Metals, Vol. 104, Issue 3, pp.175-178, July 1999.

[28] K. Puntambekar, “Characterization of Structural and Electrostatic Complexity in Pentacene

Thin Films By Scanning Probe Microscopy,” Ph.D. dissertation, University of Minnesota,

USA, October 2005.

[29] K.O Lee and T.T. Gan, “Space-Charge-Limited Currents in Evaporated Films of

Pentacene,” Physics Stat. Solutions, vol. 43, p. 565, 1977.

60

APPENDIX MASS SPECTRA RESULTS FROM TOF-SIMS Analysis

Figure 35. Mass Spectra of positive ions of interest for samples 1 to 3; Mass to Charge Ratio: 0-200

61


62


63

Figure 38. Mass Spectra of negative ions of interest for samples 1 to 3; Mass to Charge Ratio: 0-100

64


65


dna detection using organic thin film transistors:...

Documents