Mid-infrared Mid-infrared Microspectroscopy of Difficult...

20 Spectroscopy 19(2) February 2004 www.spectroscopyonl ine.com

espite recent advances inFourier transform-infrared(FT-IR) microspectroscopy andmicro-attenuated total

reflectance (ATR), for many types ofsamples there remains a need for evenhigher spatial resolution, or for anondestructive method of recordinglocalized mid-IR spectra from as-received samples without the risk ofdamage or alteration by preparationtechniques. A promising technique isthe recently developed method of near-field IR microspectroscopy using aphotothermal probe. We discussprogress in the development of thistechnique and give examples of how itcan be applied to the spectroscopiccharacterization of real-world samples.

Following a brief survey of severallimitations of the more commonly usedmid-IR microspectroscopy samplingtechniques, this article briefly discussesthe variants of near-field scanning probemethods, in which lateral spatial resolu-

DD

Mid-infrared

The authors discuss

progress in near-field IR

microspectroscopy using a

photothermal probe and

show how it can be applied

to the spectroscopic

characterization of

real-world samples.

Mid-infrared Microspectroscopy ofDifficult Samples Using Near-FieldPhotothermal Microspectroscopy

Azzedine Hammiche,Laurent Bozec, MattJ. German, John M.Chalmers, Neil J.Everall, GrahamPoulter, MikeReading, Dave B.Grandy, Francis L.Martin, and HubertM. Pollock

tion is not subject to the diffraction limit,and which have been adapted to yieldspectroscopic data in the mid-IR region.We then focus our discussions on thephotothermal technique, and describe itsuse in characterizing and mapping thecomposition of a range of differentsamples. The goal, which has not yet beenachieved, is true IR imaging, with spatialresolution at the tens of nanometerslevel; at present, the spectral detail issensitive to changes as small as 2 µm inthe positioning of the probe.

Existing Mid-IR MicrospectroscopicTechniques (1)Methods that combine microscopy withspectroscopy have been an importantstep forward in the analysis of complexinhomogeneous samples because anyparticular feature identifiable in amicroscope image can be selected forlocalized spectroscopic analysis. Amongthe microoptical techniques, FT-IRmicrospectroscopy has a prime positionwithin many industrial laboratories.Modes of operation include transmis-sion, ATR mode, or transflectance mode.

In FT-IR transmission microscopy,for a continuous sample such as anorganic polymer film, the sample thick-ness typically must be 10 µm or less.This should enable a satisfactory finger-print spectrum showing all the peakmaxima in the spectrum to be recorded,in which the strongest bands havetransmission minima in the region of10–20%. The sample also should be ofuniform thickness, and preferablynonscattering or at least minimally scat-

PH

OTO

CR

ED

IT

www.spectroscopyonl ine.com22 Spectroscopy 19(2) February 2004

Mid-infrared

tering. Given these conditions, such afilm commonly features interferencefringes, appearing as a superimposedsinusoidal pattern in the spectrum. Torecord transmission spectra that arefree from other optical artifacts, acommon practice is to flatten samplessuch as fibers or powder particles so

that the material being examined has auniform thickness. This often is done ina compression cell. Such treatmentscan, of course, alter the properties suchas crystallinity, polymorphic form, ormolecular orientation.

In internal reflection infrared (ATR)spectroscopy, the surface of a sample isplaced into intimate contact with ahigher refractive index, infrared trans-parent internal reflection element, and

its spectrum is recorded with infraredradiation that is incident through thereflection element at an angle greaterthan the critical angle. Compared withtransmission spectra, the relative inten-sity of bands in ATR spectra increaseswith increasing wavelength. Anomalousdispersion sometimes can give rise to

distorted band shapes, with bandsappearing to have a somewhat firstderivative–like appearance. The effect isrelated to the change of refractive indexthrough an absorption band. Distor-tions are most apparent for strongbands and when operating at lowerangles of incidence. A recently devel-oped and convenient sampling tech-nique, single-bounce micro-ATR, hasbecome widely used. In this technique a

small region of the sample is locatedand pressed onto the face of a smallATR element, commonly diamond.

In the transflectance mode, a thinlayer of a sample is placed onto a reflect-ing substrate such as a gold mirror. Themid-IR radiation passes through thesample and is reflected back by the

mirror surface, so that the radiationeffectively has passed through the sampletwice. These transflectance spectra havesuperimposed on them a weaker specu-lar reflectance spectrum, which has theappearance of a first-derivative spec-trum. Spectra also can be recorded inspecular (front-surface) reflection modefrom suitable samples (that is, smooth,optically thick materials) but the result-ant spectra are heavily distorted and

Circle 25

PHOTOTHERMAL MICROSCOPY PROVIDES A NONDESTRUCTIVE MICROPROBE

APPROACH TO MID-IR SPECTROSCOPY FOR A WIDE RANGE OF SAMPLE TYPES, WITH LITTLE OR NO NEED FOR SAMPLE PREPARATION.

must be treated (usually with aKramers–Kronig transform) to separatethe refractive index and absorptionindex components of the spectra.

The advent of focal plane arraydetectors has ensured that globalimaging by FT-IR microspectroscopyhas become an extremely valuable tech-nique for the characterization of inho-mogeneous samples. Nevertheless, the

restrictions imposed by sample prepa-ration requirements remain. Further-more, there is the major problem ofdiffraction, which effectively imposeslimits on the lateral spatial resolution:any sample area examined that is of theorder of the interrogating wavelengths(approximately 10 µm) is likely to sufferfrom diffraction effects, which give riseto spectral distortion and incursion of

spectral features attributable to materialoutside of the area being examined.Recent developments include FT-IRmicrospectroscopy using eithersynchrotron radiation as the infraredsource, with its high brilliance, lowdivergence and small effective sourcesize, or tunable free-electron lasers.These enable spectral information to berecorded at significantly improvedspatial resolution. Despite these devel-opments, extending infrared investiga-tions to below the diffraction limitrequires more specialized approaches,using, for example, one of the varioustypes of scanning probes.

Existing Scanning ProbeApproaches to VibrationalSpectroscopyThe generic term “scanning probemicroscopy” (not to be confused withscanning electron microscopy) includesparticular methods such as atomic forcemicroscopy, scanning tunnelingmicroscopy, and many other variants.The key feature of such scanning probetechniques is that they are not subject tothe diffraction limit of spatial resolution.In traditional variants of microscopy, thespatial resolution depends upon wave-length, as determined by the relevantclassical theory (more precisely, far-fieldoptics, where the resolution is limited bydiffraction). In contrast, the scanningprobe methods in general are near-fieldtechniques, and the spatial resolution isdominated by other factors, such asprobe size. Until recently, most attemptsto apply scanning probe technology tolocalized spectroscopy have employedUV/Vis absorption or photolumines-cence to provide the signal that deter-mines the contrast in the microscopeimage. The past few years have seenefforts to harness the analytical power ofIR spectroscopy and Raman scattering.Östershultze and colleagues (2), Dragneaand Leone (3), and Pollock and Smith(4) provide reviews of these topics. Inscanning near-field optical microscopy,to achieve the near-field criterion, thelight source must be smaller than theoptical wavelength. Examples of such asource are an aperture such as the tip of atapered optical fiber, or a minute object

www.spectroscopyonl ine.com

Mid-infrared

24 Spectroscopy 19(2) February 2004Circle 27


Mid-infrared

that acts as a light-scattering source (alsoknown as apertureless scanning near-field optical microscopy). These twoapproaches can achieve spatial resolutionin the range of 10–100 nm.

For mid-IR scanning near-fieldoptical microscopy, low signal levels poseproblems. There are difficulties inpreparing high-throughput fibers for usein the mid-IR, especially at wavelengthsgreater than approximately 4 µm. Inaddition, photomultiplier tubes oravalanche photodiodes as detectors donot operate in the mid-IR region of thespectrum, and typical photoconductivedetectors for mid-IR wavelengths areseveral orders of magnitude less sensitivethan they are in the visible. Severalgroups have concluded that once again,the apertureless probe offers a moreattractive approach (5–7). A cantilevered

metal wire probe or a metallized atomicforce microscope probe, acting as a scat-terer of the signal from the sample, isplaced at a distance from the sample thatis less than the wavelength of the light(near-field scattering). The scatteredsignal is detected using far-field optics.In addition, recent technical develop-ments have allowed sub-wavelengthRaman mapping using near-field probesto be achieved (8–9). A Raman imagingspectrometer is integrated with a scan-ning near-field optical microscopy basedon either an extruded optical fiber or ascattering center such as a colloidalnoble metal surface.

Near-Field Photothermal IR Microspectroscopy (PTMS)At the heart of this technique (10) is aspecial version of the type of probe used

in atomic force microscopy (AFM): thescanning thermal probe. This is a minia-ture thermometer, and originally wasdeveloped to serve the version of AFMknown as scanning thermal microscopy(11). As in scanning thermal microscopy,the photothermal instrument generallyuses an “active” thermal probe as aheater as well as a thermometer (Figure1). It can give multiple images of thesample whose contrast reveals surfacetopography as well as sub-surface detailthat results from local variations in thethermal conductivity of the sample.Again, as in other scanning probemethods, these images are not subject tothe diffraction limit of spatial resolution.

Thus, given its ability to providespectral data (as explained below), theinstrument in principle can be operatedin either of two modes: the mappingmode, in which IR spectra are obtainedfrom individual regions selected fromthe scanning thermal microscopyimage; and the IR imaging mode, inwhich an image is obtained whosecontrast is determined by the localconcentration of material that absorbsin a given band of the spectrum. Todate, data from the IR imaging modeare as yet unpublished.

A related development of scanningthermal microscopy is micro-thermalanalysis (12, 13), in which the AFMimages are used to select individualregions of the sample, which then arefingerprinted or analyzed by localizedthermomechanometry, calorimetry, oranalytical pyrolysis. In all cases the samethermal probe is used as with thephotothermal technique; thus a singleinstrument could be used to probe arange of material properties, includingchemical, morphological and physicalproperties, from precisely the same areaof the sample. The ability to record IRdata that can be related directly tocomplementary images and microther-mal analyses would be a boon toresearchers, and has been shown to bepossible for composite samples ofacrylic resin–polyethylene terephthalate(14). Thermal conductivity mapping,localized thermomechanical analysis,and IR spectra were obtained from indi-vidual regions. Other advantages of the

Wollastonwire

100 �mPlatinumfilament

MirrorFigure 1. Wollaston

thermal probe (28).

The temperature is

sensed by the apex of a

V-shaped 5-mm diame-

ter platinum–10%

rhodium wire (after ref-

erence 29, copyright

2001, with permission

from Elsevier Science).

See also references 26

and 30.

Figure 2. Photother-

mal interferogram. The

maximum signal-to-

noise ratio achieved to

date was obtained

using the experimental

probe being developed

by T.A. Instruments

(30), as in the single-

pass background inter-

ferogram shown here

(no coaddition was

used).

26 Spectroscopy 19(2) February 2004


Mid-infrared

photothermal technique include thepotential for circumventing diffractionproblems associated with FT-IRmicroscopy, as mentioned earlier, andhence obtaining better lateral spatialresolution information. Here we showthat even without particularly highspatial resolution, PTMS can be viewedas a convenient, nondestructive, direct(maybe even noncontact) samplingtechnique that allows for directsampling of materials of awkwardgeometry, such as fibers or hair.Another potential application is single-particle analysis without the need ofsample preparation. Clearly, there aresituations when the advantage of doingno sample preparation on a sample is ofhigh value to avoid effects such as pres-sure-induced polymorphic changes.

Instrumentation. If a sample absorbsIR radiation it heats up. When anintensity-modulated beam supplied bya conventional FT-IR spectrometer isused, the measured temperature of thesample fluctuates accordingly. Thistemperature is measured, as a functionof time, by the scanning thermal probein passive mode, in which it acts simplyas a thermometer. Its output signal isamplified and fed into the externalinput of the same spectrometer, thusproviding an interferogram (Figure 2),which replaces the interferogramnormally obtained by means of directdetection of the IR radiation transmit-ted or reflected by the sample. TheFourier transform algorithm isperformed on this interferogram afterdigitization, in the normal way, to givethe spectrum.

To perform this procedure as a func-tion of location on the sample surface,the probe is mounted on a simple posi-tioning system, or, if images arerequired, in an atomic force micro-scope. In either case an appropriateoptical interface is required to directthe spectrometer’s external infraredbeam onto the sample and to increasethe flux at the sample surface andincrease the signal-to-noise ratio. TheIR beam is focused to a spot 2 mm indiameter. Probe and sample surface arebrought into contact at the focal point.As an alternative, we have constructed a

dedicated AFM designed to fit directlyinside the sample chamber of most FT-IR spectrometers.

Proof of concept. Proof of concept ofPTMS has been detailed in earlier publi-cations (10, 15). These describe how, fora number of polymers, prominentabsorption peaks are resolved clearly andcorrelate well with those contained in aconventional ATR spectrum obtainedfrom the same material, and that theoverall shape of the main peaks or bandsreflects the fact that the spectrum is aconvolution of different contributionsfrom both optical and thermal proper-ties. Furthermore, tests on a modelbilayer system have confirmed the feasi-bility of subsurface detection of poly-mers; a “nanosampling” version of thetechnique is possible (14), in which thesample material is deliberately made tocontaminate the tip, which then is liftedaway so that the contaminant can beanalyzed by PTMS. A complementarytechnique also has been developed thatuses tunable monochromatic radiation.An optical parametric generator is usedas the infrared source. Experiments havebeen performed using a synchrotronradiation source giving greatly improvedspectral contrast (15).

Examples of Data Obtainable By PTMSFigures 3–9 show different types of datathat can be obtained using PTMS.Except where otherwise indicated infigure captions, the following equip-ment and conditions were used: Thesample was positioned using a simplexyz translation stage. AFM imaging wasaccomplished using a Veeco (Wood-bury, NY) Explorer scanning probemicroscope. A Wollaston-type thermalprobe (Figure 1) positioned at the IRfocal point was used as the detector.The spectrometer was a Bruker (Biller-ica, MA) Vector 22 FTIR. We used acustom-built optical interface andpreamplifier obtained from Specac Ltd.(Smyra, GA) and a Stanford Research(Sunnyvale, CA) model SR650 filteramplifier with the gain set to 100 dB.The mirror speed was 2.2 kHz, thespectral resolution was 16 cm-1, and thenumber of co-additions was 2000.

Abso

rptio

n (p

hoto

ther

mal

sig

nal)

Background

Sample + background

Sample

Background

Sample + background

Sample

Point 1

Point 2

4000 3500 3000 2500 2000 1500 1000 500

4000 3500 3000 2500 2000 1500 1000 500

Wavenumber (cm-1)

Wavenumber (cm-1)

Abso

rptio

n (p

hoto

ther

mal

sig

nal)

150 �m

150 �m0 �m

75 �m

75 �m

1

2

Figure 3. Analysis of a composite polymer

sample consisting of a thin film of polymer X

sandwiched between two slabs of polymer Y.

Top: topographic AFM image. Bottom: The

spectra labeled “Sample + background” are

derived from data obtained at points shown on

the image as 1 and 2. Subtraction of the back-

ground spectrum then gives the spectra that

characterize these regions of the sample. Exper-

imental parameters are described in the text.



Mid-infrared

First, various model samples havebeen used to assess the performance ofthe technique. A sandwich of two poly-mers is shown in Figure 3, where thejunction is seen in cross-section,together with IR spectra obtained fromregions on opposite sides of the junc-tion. If we choose a region of polymerA that is closer to the junction thansome minimum distance, typicallyapproximately 5–10 µm, then as a resultof thermal diffusion, the spectrumobtained will be contaminated by peakscharacteristic of material B. Therefore,with this probe and modulationfrequency, the lateral resolution, asdetermined by thermal diffusion, issimilar to the diffraction limit of aconventional FT-IR microscope. As isdescribed elsewhere (16), the datasubsequently might be processed by aweighted subtraction of the B spec-trum, if this is already known, to isolatethe spectrum of A. If not, the nanosam-pling technique, which avoids thethermal contamination problem, can beused to provide the B spectrum.

On a more applied theme, of obviousinterest is the question of the smallestquantity of a given material that willyield useful spectral data. We examineda powder sample consisting of poly-styrene spheres 5 µm in diameter:Figure 4 shows a recognizable spectrumof polystyrene obtained from a clusterof, at most, two of these spheres.Another example of data obtained froma powder sample is described in refer-ence 15. Spectra have been obtainedfrom samples of even smaller volumeby means of nanosampling, as shown inFigure 5. We conclude that if a particleis either isolated or surrounded by IR-transparent material, extremely smallamounts of material can be analyzed.

Earlier, we outlined how the probe ispositioned onto a selected region of thesample using the AFM procedure,which includes the use of force feed-back to control the final approach ofprobe to surface. Alternatively, anoptical microscope can be incorporatedinto the instrumentation, and accuratepositioning still is possible without theuse of force feedback, as shown inFigure 6. This shows an image of the

side of a human hair, together with thecorresponding IR spectrum. (As yet, thesetup is not well purged and hencesome intrusion of water vapor bandsinto the spectrum is evident.) In thethermal contrast image (scanningthermal microscopy mode), the probe isoperated as a heat source whosetemperature is maintained constant asit is scanned. This results in themapping of heat flow out of the probe,and the contrast mainly is due tothermal conductivity variation (that ofthe hair and that of the surroundingair). We see that the probe can be usedin a simple scanning system to obtainimages of the sample surface, thuspermitting positioning of the probewithout the need for a complex opticaldetection scheme for topographymapping. The figure shows also anotherexample of data obtained from asample with awkward geometry, namelya copper filament coated with a poly-meric material.

Turning to other practical applica-tions, there must be a wide range ofsamples from which useful spectraldata can be obtained from tests thatrequire no precise lateral positioning.Here, no microscope is needed,although clearly it still is necessary tohave a means of controlling the verticalapproach of the probe. Moreover, thesample need not consist of solid mate-rial. We have shown elsewhere (15) thateven though water absorbs strongly inthe mid-IR, a solution of a surfactantin microdroplet form gives a recogniza-ble spectrum. There is of course noobstacle to analyzing a small quantityof solution once it has dried, such as acontaminant on a wipe. In one experi-ment, when a drop of a solution of abinary polymer blend in toluene wasspilled on the laboratory bench andmopped up with paper tissue, weobtained spectra consistent with that of polystyrene.

There is increasing interest in thepotential application of FT-IR spec-troscopy as a diagnostic tool fordiseases such as cancer (17, 18). Therationale for this approach is that vibra-tional spectra from normal cells willprovide a particular biochemical finger-

Sample + background

Background

Sample

PS spectrum taken from Bruker database

4000 3500 3000 2500 2000 1500 1000 500

Wavenumber (cm-1)

50�m

Figure 4. Analysis of microspheres from a poly-

styrene powder sample. The particles, which had

nominal diameters of 5 µm, were spread and

embedded into the surface of a rock salt window.

The AFM image was obtained using a standard

Si3N4 probe. Spectra are labeled as in Figure 3.

For comparison, the fourth spectrum (poly-

styrene reference) is taken from the database of

Bruker UK. Experimental parameters as given in

the text, but with 4000 coadditions.



Mid-infrared

print and that, in comparison, analysisof cells from diseased tissue will giverise to spectral differences. It also hasbeen hypothesized that FT-IR spectral-derived estimations of intracellularmetabolites such as phosphate, glyco-gen, RNA, and DNA could be gainfullyemployed to differentiate normalversus malignant cells (17, 18). Ofcourse, the potential applications ofFT-IR microspectroscopy are wideranging and include studies into peri-cellular proteolysis (19), antioxidant-induced changes in oxidized DNA (20),and the structure and conformation ofproteins (21).

This leads to questions regardingsensitivity requirements: whether arelatively insensitive probe is requiredthat gives an averaged spectrum of themultiplicity of intracellular workings,or whether even greater sensitivity isrequired to isolate specific chemicalinteractions. An analysis of populationsof cells from normal versus malignanttissues, while clearly giving rise to spec-tral differences (18), could be a some-what crude approach because of dilu-tion effects. For instance, a biopsysample of neoplastic tissue mightcontain only 2% malignant cells andalso, normal adjacent tissue could wellstill contain a few transformed cells.Hence relatively small spectral changesheavily dependent on chance would begenerated. A better approach is toanalyze single cells by FT-IR spec-troscopy and then examine them retro-spectively for characteristic morpho-logical changes (17). Such an approachwould result in biochemical finger-prints based upon vibrational spectracharacteristic of normal cells of partic-ular origin: spectral deviations thenwould point to the presence of abnor-mal cells. This approach bypasses prob-lems associated with the examinationof cell populations that might containcells in different phases of growth (cellcycle) or death (apoptosis). However, itdoes not bypass the needle-in-a-haystack problem; that is, the fact thata cell can be looked upon as a worldunto itself with reactions and interac-tions continually ongoing in an intra-cellular galaxy of microenvironments.

0 100 200

4000 3600 3200 2800 2400 2000 12001600 400800

4000 3600 3200 2800 2400 2000 12001600 400800

Inte

nsity

Inte

nsity

Vertical position of sensor

Power

Wavenumber (cm-1)

Wavenumber (cm-1)

(a)

(b)

(c)

(d) (e) (f)Temperature (oC)

Figure 5. FT-IR nanosampling of paracetamol: (a) nanosample (1.6 kHz, 5 min, 8 cm-1 spectral resolution);

(b) standard contact spectrum (1.6 kHz, 15 min, 8 cm-1 spectral resolution). (c) The micro-thermal analysis

data are consistent with the known melting point of paracetamol. (d) The clean probe. (e) After micro-ther-

mal analysis, the molten material has resolidified. The mass of material analyzed can be estimated to lie in

the picogram to femtogram range. (f) A larger volume, in the nanogram range.



Mid-infraredIn

tens

ity

Wavenumber (cm-1)4000 3500 3000 2500 2000 1500 1000 500

Hair

Inte

nsity

Wavenumber (cm-1)4000 3500 3000 2500 2000 1500 1000 500

Wire coating

(c)

150 �m

150 �m0 �m

75 �m

75 �m

150 �m

150 �m0 �m

75 �m

75 �m

(a) (b)

Circle 33

Figure 6. Analysis of samples with awkward geometry. Left spectrum: Fine strand of human hair, 120 µm in diameter; (a) topography image, obtained from the photode-

tector feedback signal, without the use of force feedback; (b) thermal contrast image, probe operated as a heat source whose temperature is maintained constant as it is

scanned. This results in the mapping of heat flow out of the probe. The contrast is due mainly to thermal conductivity variation. Right spectrum: a free-standing chromel

wire (125 mm diameter) coated with a 3 µm thick polyimide film for insulation; (c) a schematic showing the PTMS probe in contact with the surface of the wire coating.

The spectrum is consistent with that of polyimide, together with some surface contaminant material.


Mid-infrared

Figure 7 shows PTMS spectra we haveobtained from control andbenzo[a]pyrene-treated prostate (PC-3)cells. In accordance with previouslypublished studies (17, 18), characteris-tic vibrational spectra were obtained.(Nulling of the carbon dioxide absorp-tion band at 2300 cm-1 was used toperform a weighted backgroundsubtraction. Note that airborne watervapor bands at ~4000–3500 cm-1 havealso been eliminated from the spectrumon this occasion using this procedure.)Spectral differences in cells treated withbenzo[a]pyrene as compared withcontrol cells were observed. However,the question remains as to whetherthese spectral differences are just theconsequence of an analysis of twodifferent cells.

To illustrate the difficulty in isolatingintracellular effects let us examine whatoccurs in this simple experiment.Benzo[a]pyrene is a lipophilic carcino-gen that will readily cross cellmembranes. Dependent upon a cell’s

4000 3500 3000 2500 2000 1500 1000

Wavenumber (cm-1)

(a)

4000 3500 3000 2500 2000 1500 1000 500

Wavenumber (cm-1)

(b)Control PC-3 cells

B[a]P-treated PC-3 cellS

Figure 7. Characterization of malignant cells. The spectra show differences induced in whole cells following 24-h treatment with a chemical carcinogen, benzo[a]pyrene. Immortalized prostate (PC-3) cells were seeded as 3-mL aliquots (~1x104 cells) into 30-mm petri dishes containing glass coverslips at 37 °C and 5% carbon dioxide. The cells were allowed 24 h to attach to the coverslips and then were treated with 1 µM benzo[a]pyrene (added as a stock solution in dimethyl sulfoxide, <1% v/v) for a further 24 h. The medium then was aspirated and the cells were washed with phosphate buffered saline before being fixed with 70% ethanol. Samples then were dehydrated under vacuum overnight before analysis. (a) Normalized spectrum of control PC-3 cells (blue line) and benzo[a]pyrene-treated PC-3 cells (red line); (b) the result of subtracting the spectrum from carcinogen-treated cells from the spectrum obtained for control PC-3 cells. Some bands of interest include the amide I peak at 1654 cm-1, the amide II peak at 1545 cm-1, the DNA peak at 1020 cm-1, the RNA peak at 1121 cm-1, the glucose peak at 1030 cm-1, the phosphate peak at 1080 cm-1, and the phospholipids peak at 2924 cm-1.

Circle 35


Mid-infrared

ability to metabolically activate thiscarcinogen, benzo[a]pyrene isconverted to a reactive electrophile thatthen binds covalently to nucleophilicsites in DNA to form bulky chemicalDNA adducts. PC-3 cells contain thismetabolic machinery. A human cellgenome contains 3.12 billion bases andthe concentration of benzo[a]pyreneused here results in 2–15 adducts per108 nucleotides in prostate cells (22).While such adducts give rise to struc-tural alterations, these are going to berelatively localized and the possibility ofisolating such effects extremely random.To differentiate between a control andbenzo[a]pyrene-treated cell, the averageof how many spectral readings wouldbe needed? Probably thousands, at thevery least, would be required.

To address this question we exam-ined spectral changes in DNA isolatedfrom cells treated with picomolarconcentrations of the organochlorinepesticide, lindane. Such agents areendocrine disruptors and increasinglyare associated with the induction oflow-dose effects (23). In this experi-ment isolated DNA was denatured intotwo single strands. Spectral vibrationsbetween 1700 cm-1 and 1350 cm-1 areattributed to alterations in base struc-tures associated with the induction ofchanges in conformational properties;these are markedly apparent in Figure8. With increasing sensitivity of themethodology a correspondinglyincreasing simplicity of matrix isrequired: future experiments willinvolve the spectral characterization ofcarcinogen-modified oligodeoxynu-cleotide adducts (24).

It is likely that fine-scale PTMSmeasurements will provide key data forthe study of a far larger range ofsamples than can be outlined here,ranging from a single powder particleof an illicit drug or a fragment of paint,to part of a painting in situ, the lubri-cant on a ball-bearing, or the variationin surface crystallinity along a stressedpolymer molding. When necessary,spectra can be obtained without theneed for direct contact between probeand sample (Figure 9).

Wavenumber (cm-1)

Ab

sorb

ance

(a)

Wavenumber (cm-1)

(b)

17001700 1600 1500 1400 1300

1.5

1.0

0.5

0.0 1600 1500 1400 1300

DNA from control MCF-7 cells

DNA from Lindane-treated MCF-7 cells

Figure 8. Damage to DNA induced by picomolar concentrations of lindane. The spectra show differences

induced in isolated single-stranded DNA following a 24-h treatment with lindane of immortalized estro-

gen-receptor positive mammary (MCF-7) cells. Cells were treated for 24 h with 10-11 M lindane (added as

a stock solution in dimethyl sulfoxide, < 1% v/v) before suspension in low-melting point agarose and

application onto glass coverslips. After agarose was allowed to set on a cold surface, the cells were lysed

at high pH to release individual cell genomes and denature double-stranded DNA into single strands. Fol-

lowing a DNA unwinding step, samples were dehydrated under vacuum overnight before analysis. (a)

Normalized spectrum of DNA isolated from control MCF-7 cells (blue line) and lindane-treated MCF-7

cells (red line); (b) the result of subtracting the spectrum from lindane-treated cell DNA from the spec-

trum obtained for control MCF-7 cell DNA. Spectral vibrations between 1750 cm-1 and 1350 cm-1 are

attributed to alterations in base structures, and comparisons of spectral properties in this region provide

insights into alterations in the base and backbone structure of DNA.

Contact

5 �m

10 �m

20 �m

50 �m

100 �m

4500 3750 3000 2250 7501500

Wavenumber (cm-1)

Figure 9. Noncontact spectra of polypropylene, illustrating heat transfer through air. The approximate dis-

tance of the probe from the sample surface is shown on the left.



Mid-infrared

Prospects For High Spatial Resolution in PTMSMuch already has been achieved withthe PTMS approach, even withoutachieving higher spatial resolution. Wehave seen that the technique provides anovel, nondestructive microprobeapproach to mid-IR spectroscopy, for awide range of sample geometries andphysical forms, and with little or noneed for sample preparation.

We anticipate even greater success inthe near future with improvements toour PTMS system. As outlined earlier(13), in other modes of imaging thesame type of thermal probe can achievesubmicrometer resolution. With PTMS,ultimately the spatial resolution will belimited simply by the size of the probeand the temperature distribution withinthe sample, rather than the infraredwavelength. Rather than diffractionbeing a limiting property to spatiallyresolving spectra, with PTMS thebalance between the optical absorptionand thermal diffusion lengths becomes

critical. Various additional factors willaffect the spatial resolution, includingthe size of the probe, the sharpness ofthe temperature distributions that resultfrom the absorption of infrared energyby the inhomogeneities that are to beidentified, and the finite heat capacity ofthe probe, which must perturb thetemperature distributions to be detected.Further improvement in spatial resolu-tion could be achieved by making use ofthe damped nature of thermal waves,such that the thermal diffusion length(25) can be made smaller than the probesize. In an as yet unpublished study wefind that PTMS at a spatial resolutionbetter than the diffraction limit may beachieved with the help of the smallermicromachined type of thermal probe(26), and elsewhere (27) we havediscussed the prospects of achievingbetter than 1-µm resolution.

AcknowledgmentsWe thank the U.K. Engineering andPhysical Sciences Research Council

(EPSRC) and TA Instruments, Inc.(New Castle, DE), for financial support,ICI plc Measurement Science Group(Redcar, England) for a studentship,Bruker UK (Coventry, England) for theloan of equipment, and J Laveissièrewho obtained the nanosampling results.

References1. J.M. Chalmers and P.R. Griffiths, Eds.,

Handbook of Vibrational Spectroscopy(John Wiley & Sons Ltd., Chichester, Eng-land, 2002).

2. E. Österschulze, M. Stopka, and R. Kass-ing, Microelectron. Eng. 24, 107 (1994).

3. Dragnea and S.R. Leone, InternationalReviews in Physical Chemistry 20, 59(2001).

4. H.M. Pollock and D.A. Smith, in Hand-book of Vibrational Spectroscopy (JohnWiley & Sons Ltd, Chichester, England),vol. 2, pp. 1472–1492.

5. R. Bachelot, P. Gleyzes, and A.C. Boccara,Opt. Lett. 20, 1924 (1995).

6. B. Knoll and F. Keilmann, Nature 399,134 (1999).

7. A. Piednoir, C. Licoppe, and F. Creuzet,Opt. Comms. 129, 414 (1996).

8. D.A. Smith, S. Webster, M. Ayad, S.D.

Circle 19


Mid-infrared

Evans, D. Fogherty, and D.N. Batchelder,Ultramicroscopy 61, 247 (1995).

9. Y. Narita, T.T. Ikeda, T. Saiki, S.Mononobe, and M. Ohtsu, Appl. Spec-trosc. 52, 1141 (1998).

10. A. Hammiche, H. M. Pollock, M. Reading,M. Claybourn, P. Turner, and K. Jewkes,Applied Spectroscopy 53, 810–815(1999).

11. A. Majumdar, Ann. Rev. Mater. Sci. 29,505 (1999).

12. D.M. Price, M. Reading, A. Hammiche,and H. M. Pollock, International Journalof Pharmaceutics 192, 85–96 (1999).

13. H.M. Pollock and A. Hammiche, J Phys D:Appl Phys 34, R23–R53 (2001).

14. M. Reading, D. Grandy, A. Hammiche, L.Bozec, and H.M. Pollock, VibrationalSpectroscopy 901, 257–260 (2002).

15. L. Bozec, A. Hammiche, H. M. Pollock, M.Conroy, J. M. Chalmers, N. J. Everall, andL.Turin, J Appl Phys 90, 5159–5165(2001).

16. L. Bozec, PhD Thesis, Lancaster Univer-sity (2003).

17. M. A. Cohenford and B. Rigas, Proc. Natl.Acad. Sci. USA 95(26), 15327–15332(1998).

18. S. Argov, J. Ramesh, A. Salman, I. Sinel-nikov, J. Goldstein, H. Guterman, and S.Mordechai, J Biomed. Opt. 7(2),

248–254 (2002).19. S. Federman, L.M. Miller and I. Sagi, Ma-

trix Biol. 21(7), 567–577 (2002).20. D.C. Malins, K.E. Hellstrom, K.M. Ander-

son, P.M. Johnson, and M.A. Vinson, ProcNatl Acad Sci USA 99(9), 5937-5941(2002).

21. C. Vigano, E. Goormaghtigh, and J.M.Ruysschaert, Chem. Phys. Lipids 122(1-2), 121–135 (2003).

22. F.L. Martin, K.J. Cole, G.H. Muir, G.G.Kooiman, J.A. Williams, R.A. Sherwood,P.L. Grover, and D.H. Phillips, ProstateCancer Prostatic Dis. 5(2), 96–104(2002).

23. E. Yared, T.J. McMillan, and F.L. Martin,Mutagenesis 17(4), 345–352 (2002).

24. K. Brown, C.A. Harvey, K.W. Turteltaub,and S.J. Shields, J Mass. Spectrom. 38(1),68–79 (2003).

25. D.P. Almond and P.M. Patel, Photother-mal Science and Techniques (Chapmanand Hall, London, 1996).

26. H. Zhou, A. Midha, G. Mills, S. Thoms,S.K. Murad, and J.M.R. Weaver, J Vac. Sci.Technol. B16, 54–58 (1998).

27. A. Hammiche, L. Bozec, M. Conroy, H.M.Pollock, G. Mills, J.M.R. Weaver, D.M.Price, M. Reading, D.J. Hourston, and M.Song, Journal of Vacuum Science &Technology B18(3), 1322–1332 (2000).

28. R.B. Dinwiddie, R.J. Pylkki and P.E. West,in Thermal Conductivity 22, T.W. Tong,Ed. (Technomics, Lancaster, USA, 1994),pp. 668–677.

29. P.G. Royall, D.G.M. Craig, and D.B.Grandy, Thermochimica Acta 380(2),165–173 (2001).

30. R. Dekhter, E. Khachatryan, Yu. Kokotov,and A. Lewis, Appl Phys. Lett. 77,4425–4427 (2000). ■

Azzedine Hammiche, LaurentBozec, Matt J. German, andHubert M. Pollock* Department of Physics, Lancaster University,EnglandJohn M. Chalmers School of Chemistry, University ofNottingham, and V S Consulting, EnglandNeil J. Everall Measurement Science Group, ICI plc, EnglandGraham PoulterSpecac Ltd, EnglandMike Reading and Dave B.GrandyIPTME, Loughborough University, EnglandFrancis L. MartinDepartment of Biological Sciences, LancasterUniversity, England.

*To whom all correspondence should be addressed.E-mail: [email protected].

Circle 8

Mid-infrared Mid-infrared Microspectroscopy of Difficult...

Documents

Transcript of Mid-infrared Mid-infrared Microspectroscopy of Difficult...