LUND UNIVERSITY

PO Box 117221 00 Lund+46 46-222 00 00

Two-color Trapped-particle Optical Microscopy

Malmqvist, Lars; Hertz, H. M

Published in:Optics Letters

DOI:10.1364/OL.19.000853

1994

Link to publication

Citation for published version (APA):Malmqvist, L., & Hertz, H. M. (1994). Two-color Trapped-particle Optical Microscopy. Optics Letters, 19(12),853-855. https://doi.org/10.1364/OL.19.000853

General rightsCopyright and moral rights for the publications made accessible in the public portal are retained by the authorsand/or other copyright owners and it is a condition of accessing publications that users recognise and abide by thelegal requirements associated with these rights.

• Users may download and print one copy of any publication from the public portal for the purpose of private studyor research. • You may not further distribute the material or use it for any profit-making activity or commercial gain • You may freely distribute the URL identifying the publication in the public portalTake down policyIf you believe that this document breaches copyright please contact us providing details, and we will removeaccess to the work immediately and investigate your claim.

June 15, 1994 / Vol. 19, No. 12 / OPTICS LETTERS 853

Two-color trapped-particle optical microscopy

L. Malmqvist and H. M. Hertz

Department of Physics, P.O. Box 118, S-221 00 Lund, Sweden

Received January 4, 1994

We demonstrate a method for nonintrusive scanned near-field optical microscopy. The microscope utilizes anoptical trap to position accurately a 50-100-nm-diameter lithium niobate particle. The infrared trapping beamis frequency doubled in the particle, resulting in a visible microscopic optical probe. By separation of the trapping

and detection wavelengths, objects that are transparent in the infrared (e.g., biological) may be positioned close

to the particle, resulting in high resolution. The current experimental resolution is limited to approximately500 nm by the properties of the test objects. The theoretical resolution is less than 100 nm.

Scanned near-field optical microscopes' (SNOM's)have demonstrated imaging with an order-of-magnitude better resolution than the 200-300-nmconventional diffraction limit by scanning an opticalprobe in immediate proximity to the studied object.The SNOM has been used primarily to study drysurfaces, but biophysical applications are receiv-ing increasing attention.2 Current SNOM probesinclude etched apertures, protrusions,3 fluorescenttips,4 and pulled fiber tips.2 Because the position-ing of these probes requires mechanical access tothe object, samples with intervening membranesor rough surfaces (e.g., biological) are not alwaysaccessible for such studies. Trapped-particle opticalmicroscopy5 (TPOM) eliminates this restriction byusing a nonintrusive optical trap6'7 to position andscan the microscopic light source. In addition toits nonintrusive character, the elastic nature of theTPOM probe permits probe-sample collisions with-out damage. Should a probe particle be accidentallylost, it is easily replaced by a new one since manyparticles are available for injection into the trap. Inthis Letter we describe an increase in the resolutionof TPOM by separating the trapping and detectionwavelengths, using a microscopic lithium niobatecrystal as the optical probe.'

In the first TPOM experiments5 scattered lightfrom 290-nm dielectric particles trapped by an argon-ion laser (A = 514 nm) was used for the imaging.However, the resolution was limited to approximately2 ,m since the minimum particle-object distancewas a few micrometers. At smaller distances, thescattering by the object is significant compared withthe scattering by the particle, resulting in a lowsignal-to-noise ratio. We eliminate this by trappinga nonlinear crystal particle with an IR beam andusing the emitted visible frequency-doubled light forthe microscopy (two-color TPOM). By detection ofonly the visible light, the scattering problem is cir-cumvented, and smaller particle-object distancesmay be used, resulting in higher resolution. Fur-thermore, the IR trapping wavelength exhibits alow absorption in most biological materials, therebyminimizing thermally induced turbulence and dam-age to the studied object. It has been shown thatapproximately 100 mW of trapping power may be

used without damaging biological cells.7 Finally,the second-harmonic power shows no sign of thebleaching known to occur in many fluorescent probes.

Figure 1 shows the experimental arrangement fortwo-color TPOM. The optical trap consists of a me-chanically chopped cw TEMOO Nd:YAG laser beam(A = 1.064 tum) focused by a N.A. = 1.25, 10OX wa-ter immersion objective into a water cell. In most ofthe experiments described below, the average powerat the focus was approximately 50 mW. A lithiumniobate particle is trapped just below the focus, andA = 532 nm frequency-doubled light is emitted. Thislight is collected with a N.A. = 0.12 microscope ob-jective to an optical fiber, and through interferenceand IR-blocking filters it is detected by a low-noisephotomultiplier tube (PMT) and a lock-in amplifier.The IR laser light scattered perpendicularly to thebeam is monitored by an IR video camera. The testobjects are mounted in the water cell, which may bepositioned accurately with a piezoelectric stage. Thedata acquisition and the piezoelectric stage are con-trolled by a personal computer.

We produce the lithium niobate particles bygrinding a crystal in a mortar. The particles arethen mixed with water. After a few days of sed-imentation the top layer of the water-particlemixture is used for the trapping. With electronmicroscopy the size range of the used particles wasdetermined to be 50-100 nm in diameter. The

CHOPPER

1.06 gum

Fig. 1 mef

532 rim

TEST OBJECT

FILTERS IL |COMPUTERI

pMT " Lock-in|

Fig. 1. Experimental arrangement for two-color TPOM.

0146-9592/94/120853-03$6.00/0 © 1994 Optical Society of America

854 OPTICS LETTERS / Vol. 19, No. 12 / June 15, 1994

lithium niobate particle-water mixture is injectedinto the trap with a 100-Am-diameter syringe di-rected toward the focus. With a reasonably highparticle concentration in the injected fluid, the proba-bility of trapping a particle is high. Still the residualconcentration of particles in the water cell is very low,resulting in a small probability of collecting manyparticles in the trap during the experiment. Rou-tinely a particle is trapped for several minutes. Theemitted visible intensity from the particles may varyby 2 orders of magnitude for different particles. Thisis consistent with the size range determined abovewith the assumption that the intensity is proportionalto the sixth power of the radius.9 Over the sizerange, we typically detect 0.01-1 pW of A = 532 nmradiation from a trapped lithium niobate particle.Most measurements discussed below were performedin the 0.1-1-pW regime. One may achieve higherpower by Q switching the trapping laser beam witha high repetition rate.10 The signal-to-noise ratiois typically 20-30 for the 100-ms averaging used inthe measurements, but in some instances the signal-to-noise ratio deteriorates and the power fluctuates.This might be due to rotation of the particle in thetrap, with the fluctuations reflecting the large dif-ference in the coefficients in the nonlinear opticaltensor of lithium niobate.1"

The method is experimentally demonstrated ontest objects that exhibit significant absorption at vis-ible wavelengths and low absorption in the IR. Thelow absorption is necessary so that we can avoidthermally induced turbulence that is due to absorp-tion of the trapping laser beam. The objects alsoshow resemblence to absorption properties of dyedbiological objects. The first object used was litho-graphically produced dyed photoresist patterns onmicroscope glass slides. In separate spectrometermeasurements the absorption of the approximately0.8-,um-thick film was determined to be 70% atA = 532 nm and a few percent at A = 1064 nm.Figure 2 shows a gray-scale image of a 6-,um-wideline with 0.2 ,m X 0.2 Aum pixels. The total data-acquisition time for the 90 X 30 pixel image wasapproximately 5 min, illustrating the stability of themethod. However, the resolution is low because ofthe large film thickness and repulsive surface forces,which result in a too-large particle-object distance.Furthermore, the resolution is limited by the edgesharpness of the test object, which was determinedby electron microscopy to be 0.7-1 ,um (10-90%).

To evaluate better the method's potential resolu-tion, we needed an object with sharper edges. Forthis purpose free-standing silicon bridges were manu-factured. The 1-mm-long 10--tm-thick bridges wereproduced by anisotropic etching in KOH.10 Becauseof the 1.11-eV band gap of silicon, the test objectsabsorb strongly at visible wavelengths and a few per-cent at A = 1064 nm. The edge sharpness was deter-mined by electron microscopy to be <±50 nm. Toreduce repulsive surface forces (see below) we per-formed the experiments in a 0.01 M sodium chlo-ride solution. Figure 3 shows the recorded TPOMsignal from the silicon edge. The 10-90% sharp-ness is 0.5 ± 0.05 ,m. Assuming that the effective

light source is approximately Gaussian (see below),this corresponds to a resolution of approximately0.45 Am.

If the particle-object distance is very small, theresolution of TPOM is ultimately determined by theeffective size of the trapped particle, i.e., the particle'sphysical size convolved with its displacement that isdue to Brownian motion in the trap. In addition tothe effective size in the radial r direction, the dis-placement in the axial z direction has to be consid-ered since this influences the distance between theparticle and the object and, thus, the resolution. As-suming that the optical trap is a harmonic potentialwell, we may calculate the rms displacement from thescattering and gradient radiation forces by followingthe Gaussian beam model in Refs. 5 and 12. Typ-ically the rms displacement in the z direction for a75-nm-diameter lithium niobate particle trapped by100 mW of A = 1.06 ,m radiation is 40 nm. Hereojo = 350 nm is used in the model, in order to producefocal intensity gradients equivalent to those in ourstrongly truncated Gaussian beam. In the radial rdirection the rms displacement is a factor of 2 smallerowing to the stronger gradient forces in this direction.Convolving the physical size with the FWHM of therms displacement results in a Gaussian-like lightsource with an approximately 80-nm FWHM. Usinga A = 800 nm trapping beam, which also exhibits lowabsorption in biological material, results in a 25% re-duction in the FWHM since smaller particles may beused. The numbers clearly indicate the potential forsubdiffraction-limit imaging with the TPOM. Near-field enhancement may result in higher resolution.

Although measurement of edge sharpness is an in-sufficient test of resolution in the near-field region,2

the experimentally determined resolution is clearlyless than the theoretically calculated resolution andthe conventional diffraction limit. In the case ofthe dyed photoresist, the slope of test object and

I 10 gmFig. 2. Image of a dyed photoresist test object.

1

0.5+

ce

m

0

Fig. 3.edge.

2 4 6

DISTANCE (jim)

One-dimensional scan over an etched silicon

June 15, 1994 / Vol. 19, No. 12 / OPTICS LETTERS 855

large particle-object separation are the limiting fac-tors. For the silicon edge experiments, the resolu-tion is limited primarily by surface forces, result-ing in a too-large separation between the particleand the object. Because of the small magnitude ofthe radiation forces trapping the particle (typically10-12_10-13 N), these forces may be exceeded byattractive and repulsive surface forces13"4 at smallparticle-object distances. Repulsive double-layerelectrostatic forces may be of the order of the radia-tion forces up to a micrometer away from the surfacein pure water. However, these forces can largely beeliminated by use of a sodium chloride solution in-stead of pure water, which is also consistent with ex-periments on biological objects. With a 1.5 x 10-1 M(isotone) sodium chloride solution the Debye length13is 0.8 nm, resulting in negligible electrostatic forcesat distances larger than a few nanometers. Attrac-tive van der Waals forces may not be screened ina similar way. Because of the high refractive in-dex of the silicon bridge (n = 3.58) these forces areequal to or larger than the trapping forces up to ap-proximately 50 nm away from the surface and maydominate over the repulsive surface forces. In thiscase, the probe particle may be lost by adsorptionto the surface. For the oscillating particle not to beadsorbed,we estimate that the particle should be an-other five times the rms displacement away from thispoint. This results in a particle-object distance ofapproximately 250 nm in the experiments describedabove. At this distance the recorded FWHM of thesource should be approximately 500 nm,15 which isconsistent with our measurements.

The attractive forces are reduced a factor of -30compared with silicon for low-refractive-index objectssuch as biological samples (average refractive index= 1.37). Combining this method with a sodium chlo-ride solution to reduce the long-range double-layer re-pulsion, we expect to reach particle-object distancesof tens of nanometers, resulting in subdiffraction-limited resolution. The lower attractive forcesresult in a thin (a few nanometers) remaining repul-sive potential barrier at the surface,'3 which shouldprevent the adsorption of the probe particle. Thebarrier also would make it possible to scan the probeparticle in the vicinity of rough surfaces without lossof the particle. Other potential factors that increasethe effective particle-object distance are mechanicalvibrations and thermally induced turbulence; care-ful engineering and test objects similar to biologicalobjects with low IR absorption, respectively, shouldmake these factors a lesser problem.

The experiments were performed without probe-sample distance regulation, and the distances derivedabove were determined from the edge response as theprobe was moved closer to the sample. Clearly, in apractical microscope the probe-sample distance mustbe controlled. In tip-based SNOM's this is oftenachieved with shear-force systems,'6 which are notapplicable to the nonintrusive probe. We are cur-rently developing a capacitive distance control sys-

tem, which will permit well-defined TPOM measure-ments on flat or slightly rough surfaces.

In summary, we have demonstrated a method forperforming nonintrusive scanning near-field opticalmicroscopy with an optically trapped probe. Thisresearch is motivated primarily by the possibilityof performing nonintrusive optical near-field stud-ies of living biological material with high resolution.The experimental resolution is significantly improvedover that in earlier research and is limited mainly bythe surface forces that are due to the high refractiveindex of the silicon in the test objects. Our futureresearch will focus on increasing the resolution by us-ing low-refractive-index objects and obtaining higherpower by pulsing the trapping laser beam.

The authors gratefully thank Lars Rosengren andKarin Ljungberg for the silicon bridges, KristinaGeorgsson for her help with the electron microscopyand the photolithography, Walter G. Hertlein forpreparing the dyed photoresist, Fredrik Laurell forthe lithium niobate, and Hakan Wennerstrdm andLars Rymell for stimulating discussions. This re-search was financed by the Swedish Natural ScienceResearch Council, the Swedish Board for Technicaland Industrial Development, the Carl Trygger Foun-dation, and the Crafoord Foundation.

References

1. D. W. Pohl, in Advances in Optical and Electron Mi-croscopy, C. J. R. Sheppard and T. Mulvey, eds. (Aca-demic, London, 1991), Vol. 12, pp. 243-312.

2. E. Betzig and J. K. Trautman, Science 257, 189 (1992).3. U. Ch. Fischer and D. W. Pohl, Phys. Rev. Lett. 62,

458 (1989).4. A. Lewis and K. Lieberman, Nature (London) 354, 214

(1991).5. L. Malmqvist and H. M. Hertz, Opt. Commun. 94, 19

(1992).6. A. Ashkin, J. M. Dziedzic, J. E. Bjorkholm, and S. Chu,

Opt. Lett. 11, 288 (1986).7. S. M. Block, in Noninvasive Techniques in Cell Biol-

ogy, J. K. Foskett and S. Grinstein, eds. (Wiley, NewYork, 1990), pp. 375-402.

8. S. Sato and H. Inaba, Electron. Lett. 28, 286 (1992).9. M. Kerker, The Scattering of Light (Academic, New

York, 1969), Chap. 3, p. 33.10. H. M. Hertz, L. Malmqvist, L. Rosengren, and

K. Ljungberg, "Optically trapped nonlinear particlesas probes for scanning near field optical microscopy,"Ultramicroscopy (to be published).

11. A. Yariv and P. Yeh, Optical Waves in Crystals (Wiley,New York, 1984), Chap. 12, p. 513.

12. S. Chandrasekhar, Rev. Mod. Phys. 15, 1 (1943).13. J. Israelachvili, Intermolecular and Surface Forces

(Academic, San Diego, Calif., 1992), Chaps. 11 and 12,pp. 176-259.

14. W. A. Ducker, T. J. Senden, and R. M. Pashley, Nature(London) 353, 239 (1991).

15. Y. Leviatan, J. Appl. Phys. 60, 1577 (1986).16. E. Betzig, P. L. Finn, and J. S. Wiener, Appl. Phys.

Lett. 60, 2484 (1992).