Volume 55, Number 6, 2001 APPLIED SPECTROSCOPY 6550003-7028 / 01 / 5506-0655$2.00 / 0q 2001 Society for Applied Spectroscopy

accelerated paper

In Situ Infrared Technique for Studying Adsorption ontoParticulate Silica Surfaces from Aqueous Solutions

BRIAN J. NINNESS, DOUG W. BOUSFIELD, and CARL P. TRIPP*Laboratory for Surface Science and Technology (LASST) (B.J.N., C.P.T.), Department of Chemical Engineering (B.J.N., D.W.B.),and Department of Chemistry (C.P.T.), University of Maine, Orono, Maine 04469

An in situ infrared technique is described which allows the detectionof adsorbed surface species on metal oxide particles in an aqueousenvironment. The technique involves � rst formulating a ‘‘coating’’comprised of high-surface-area silica particles and a polymericbinder in a suitable solvent. The resulting coating is applied to thesurface of an internal re� ection element and mounted in a � ow-through attenuated total re� ection (ATR) apparatus. The techniqueis demonstrated with a ZnSe element coated with fumed silica par-ticles in a polyethylene (PE) matrix. Access of the silica surface inthe matrix to adsorbates was evaluated by comparing the gas-phasereaction of silanes on silica/PE-coated CsI windows in transmissionwith silica/PE-coated ZnSe in an ATR evacuable cell. It is shownthat the PE weakly perturbs about 25% of the surface hydroxylgroups, and that all surface groups are available for reaction withsilanes. The silica/PE is inde� nitely stable in an aqueous environ-ment and has advantages of at least 2 orders higher sensitivity anda wider spectral range over studies using oxidized silicon wafers.The usefulness of this technique for studying adsorption on metaloxide surfaces is demonstrated with the reaction of succinic anhy-dride on an aminosilanized silica surface. This reaction sequence isa common method used to prepare glass surfaces in the attachmentof probe oligonucliotides for microarray biochip technology.

Index Headings: Silica particles; Attenuated total re� ection; ATR;Aqueous adsorption.

INTRODUCTION

Surface-modi� ed metal oxide particles have numerousapplications in colloid and surface science-related tech-nologies.1 One of the principal tools used to interrogatethe surface chemistry and surface reactions on metal ox-ide particles is infrared spectroscopy.2,3 Typically, themetal oxide particles have high surface areas (25–400m 2 /g), which affords easy detection of bands due to sur-face groups as well as characteristic bands identifying thenature of the adsorbed species. While there have beennumerous in situ infrared studies of gaseous reactions andthose conducted from nonaqueous solutions, very few in-frared studies have been reported for aqueous-based ad-sorption on metal oxide particles. Surface studies in waterare dif� cult because water is a strong absorber of infraredlight.

Received 1 February 2001; accepted 8 March 2001.* Author to whom correspondence should be sent.

The most common approach used to overcome thestrong absorption of infrared radiation by water is atten-uated total re� ection (ATR) spectroscopy. The � nite pen-etration of the evanescent wave into the aqueous solutionis such that spectral transparency is obtained over muchof the infrared region. Furthermore, by working at diluteconcentrations, one can ensure that the spectral contri-butions due to excess adsorbate in solution are small incomparison to the spectral features due to adsorbed spe-cies.4 However, the use of ATR for surface studies re-quires that the internal re� ection element (IRE) be usedfor the surface/water interface. Unless the surface of in-terest is available as an IRE,5–8 the IRE must be modi� edin some way to mimic the desired metal oxide surface.

Adsorption onto silica has been one of the most widelystudied systems by ATR methods. A common approachis to establish a silica layer by preparing an oxide layerdirectly on a silicon IRE.9–11 This technique has the ad-vantage of preparing an oxide surface that is stable in anaqueous environment and is amenable to orientation stud-ies using polarized light.10 However, the bands due toadsorbed species are usually weak in intensity, owing tothe low surface area of the oxidized IRE. Typically onlythe most intense bands of the adsorbate are detected. Fur-thermore, the spectral range of a silicon IRE is limited tothe region above 1550 cm21 and many characteristicbands lie below this cutoff. Sputtering a thin � lm of sil-icon onto a ZnSe IRE can circumvent this limitation bytaking advantage of the expanded spectral range ofZnSe.12 This technique has proven bene� cial in accessingthe spectral window between 1550 and 1100 cm21.

On the other hand, transmission IR studies on silicaparticles typically probe 3 orders of magnitude more sur-face area than oxidized silicon IRE and can access theentire IR spectral region.13 Given these advantages, theuse or development of methods to anchor high-surface-area silica particles to ATR crystals is highly desirable.One simple method is to directly deposit the high-sur-face-area metal oxide particles from a suspension onto aZnSe IRE. This technique has been used to prepare water-stable, high-surface-area TiO2 layers for the study of sur-face reactions in water.14 The technique simply involvesforming a suspension of suitable concentration of themetal oxide particles in a solvent. After thoroughly mix-

656 Volume 55, Number 6, 2001

ing and sonicating to achieve maximum dispersion, onedeposits a small amount of the suspension onto the sur-face of the IRE. After drying, and depending on the sur-face energetics of the system, a stable hydrated gel layerof the metal oxide can be formed. This deposition tech-nique has also been used with colloidal silica particles ona cylindrical ZnSe IRE in nonaqueous solutions.15 How-ever, attempts to prepare high-surface-area silica sub-strates in this manner that are stable in water have provenunsuccessful. The silica is removed from the crystal whenplaced in contact with � owing water. An alternativemethod is to prepare a metal oxide layer via sol-gel tech-niques. Silica � lms deposited by sol-gel techniques havethe advantage of allowing control over the � lm thickness,porosity, and chemical composition. These systems haveproven to exhibit much greater sensitivity for the detec-tion of trace analytes in water as compared to an uncoatedwaveguide.16 The problems associated with sol-gel tech-niques include the time-intensive preparation of the sol,the aging and controlled drying conditions, and the as-sociated adhesion and cracking of the � nal gel layers.17

On the basis of these limitations, it is clear that a uni-versal technique for anchoring any colloidal particle tothe surface of an IRE would expand the applicability ofsurface reaction studies by aqueous-phase ATR spectros-copy. Given the importance of metal-based oxides inaqueous colloidal dispersions and the heightened activityof biological adsorption processes (i.e., silica is used asa substrate in microarray biochip technology18), there isa clear need to develop a general infrared technique thatcan be used to follow adsorption from aqueous solutionson metal oxide particles.

In this report we present a new technique for preparinghigh-surface-area particulate layers on a ZnSe IRE forsubsequent studies of adsorption from aqueous solutions.Due to the aforementioned dif� culties in studying aque-ous-phase surface reactions on colloidal silica layers viaATR, this metal oxide is the focus of this initial study.The technique involves using a binder, polyethylene (PE),to anchor the colloidal silica particles to the IRE surface.PE is used because it has few bands (C–H modes) in themid-IR region and therefore is a good window material.A deuterated polyethylene (d-PE) can be substituted ifaccess to the C–H modes is required. The resulting ‘‘coat-ing’’ of silica particles and binder is extremely stable inwater and allows reaction with the silica surface to beinvestigated. To determine the accessibility of the silicasurface in this composite layer, we examine the gas-phasereaction of hexamethyldisilazane (HMDS) with silica.Comparison is made between results obtained for the gas-phase adsorption of HMDS measured for transmissionstudies for both pure silica � lms and silica/PE coatingssupported on a CsI window, as well as gas-phase ATRmeasurements of the silica/PE deposited onto a ZnSeIRE. The suitability of this technique for aqueous-phaseadsorption studies is then investigated by the reaction ofsuccinic anhydride with an aminosilane-treated silica sur-face. This is a common recipe used to prepare glass slidesfor attachment of probe oligonucliotides for use in mi-croarray biochip technology.18

EXPERIMENTALThe fumed silica used in this study was Aerosil 380

(Degussa) with a measured speci� c surface area of 375

m 2g21, as determined by BET nitrogen adsorption. Thepolyethylene was obtained from Aldrich and has a weightaverage molecular weight of 4000 with a polydispersity(Mw /Mn) of 2.3. Hexamethyldisilazane and succinic anhy-dride were used as received from Aldrich. The aminosi-lane, (3-aminopropyl)dimethylethoxysilane (APDMES),was purchased from United Chemical Technologies.HMDS and APDMES were transferred to gas bulbs anddegassed according to standard freeze/thaw methods.

The technique for preparing the silica/PE coating wasas follows. The PVC is de� ned in the paints and coatingsindustry as the ratio of the pigment volume to the totalvolume of pigment and binder. Paints are formulated atlow PVC values, where the polymeric binder � lls theinterstitial voids between the pigment particles in thedried layer. This formulation gives the paint weatherabil-ity and mechanical strength. At the other extreme, papercoatings are formulated at high PVC values, where thereis not enough binder present to � ll the interstitial voids.Therefore, air voids are present in the � nal dried layer,and these contribute to the scattering of the incident light,which enhances the gloss and opacity of the coating. Forthis application the silica/PE coatings are formulated atthe highest possible PVC. This approach produces a sil-ica-based coating with the least amount of binder present,which remains stable when in contact with a � owingstream of water.

For the materials in this study we use a mass ratio ofsilica to PE of approximately 4:1. Typically, 10 mg ofpolyethylene is completely dissolved in 3 mL of toluene,heated to 105 8C, and thoroughly mixed. To this, 40 mgof the fumed silica is added and mixed vigorously for 30min. A 100 ml aliquot of this dispersion is withdrawn anddeposited onto a CsI support for transmission measure-ments or 200 mL is deposited onto both sides of a ZnSeIRE for ATR measurements. The solvent is then allowedto evaporate at ambient conditions. Solvent evaporationcauses shrinkage of the coating layer and produces in-dividual islands of coating on the substrate. Figure 1 de-picts a typical coating layer on a ZnSe IRE. The individ-ual islands are roughly 20 mm in size and are separatedby approximately 5 mm cracks. With the use of a styluspro� lometer, these islands of coating are found to beabout 5 mm in thickness.

The infrared cell and the thin � lm technique for trans-mission measurements are described in detail elsewhere.13

Gas-phase adsorption studies by ATR were performedwith the in-house designed cell shown in Fig. 2. The cellconsists of a two-piece quartz chamber with epoxiedNaCl windows on the ends of the chamber. The IRE ismounted in a metal holder that rests on a self-compen-sating metal base and is held in position in the chamberwith a spring clip. This setup allows easy alignment ofthe cell on top of a twin, parallel mirror re� ection at-tachment from Harrick, and the entire chamber is con-nected to a standard vacuum line. The aqueous-phase ex-periments were carried out with a standard ATR liquid� ow cell arrangement from Harrick. A peristaltic pumpwas used to � ow the liquid solution across the silica/PEsurface during the aqueous-phase experiments. All spec-tra were recorded on a Bomem MB-Series FT-IR (Fouriertransform infrared) instrument with a liquid N2-cooled

APPLIED SPECTROSCOPY 657

FIG. 1. Optical image of the silica/PE coating deposited onto a ZnSe IRE.

mercury cadmium telluride (MCT) detector. Typically100 scans were coadded at a resolution of 4 cm21.

RESULTS AND DISCUSSION

Transmission Spectra of Silica and Silica/PE Films.The thin-� lm transmission spectra of silica dispersed ona CsI window and of the silica embedded in the PE coat-ing evacuated at room temperature are shown in Fig. 3aand 3b, respectively. The strong bands appearing at 1090and 810 cm21 are Si–O bulk modes of silica, and the peakat 3747 cm21 is assigned to the isolated surface SisO–Hgroups.13 In Fig. 3b, the C–H modes of the polyethyleneappear at 2917, 2850, and 1464 cm21.

The isolated hydroxyl groups are the principal adsorp-tion sites on silica. The 3747 cm21 surface band shifts to3690 cm21 when dispersed in CCl4 or cyclohexane,19

3590 cm21 in toluene, 3350 cm21 in alcohols, and 2700cm21 with amines.20 This shift represents weak physi-sorption between the hydroxyl groups on silica and CCl4

and cyclohexane, a slightly stronger interaction with thearomatic ring of toluene, and hydrogen bonding with al-cohols and amines. From a comparison of the relativeintensity of the band at 3747 cm21 for the spectra in Fig.

3a and 3b, it is concluded that most of the isolated SiOHgroups are not perturbed by the presence of the PE.

An expanded view of this high-frequency region ofFig. 3 is shown in Fig. 4. For the spectrum of PE/silicashown in curve 4b, there is an additional weak featurecentered at 3690 cm21. The assignment of this band isattributed to a weak interaction between a portion of thesurface OH groups of the silica and the PE (Si–OH · · · PE)that is on the same order of strength as that obtained forweakly physisorbed CCl4 or cyclohexane. By taking theratio of the 3747 cm21 integrated peak area (SiO–H) tothat of the 810 cm21 integrated peak area (Si–O bulkmode) we can evaluate the extent of interaction of PEwith the surface hydroxyl groups of the silica. For thespectrum of silica, this ratio is calculated to be 0.0587,while the silica/PE spectrum gives a calculated value of0.0427. This result shows that approximately 73% of thetotal available isolated surface hydroxyl groups are un-affected by the presence of the PE binder. From thesedata we conclude that the fumed silica particles are thenjust basically ‘‘spot-welded’’ together with the polyeth-ylene, as depicted in Fig. 5.

Reaction with HMDS. The gas-phase reaction of

658 Volume 55, Number 6, 2001

FIG. 2. ATR cell designed for gas-phase adsorption studies.

FIG. 3. (a) IR transmission spectrum of a thin � lm of silica spreadonto a CsI window. (b) IR transmission spectrum of a silica/PE coatingdeposited onto a CsI window.

HMDS was studied to ascertain the suitability of silica/PE � lms for IR adsorption experiments and to determinethe accessibility of the hydroxyl groups perturbed by thePE. Transmission experiments of the adsorption of gas-eous HMDS on the silica particles embedded in the PEmatrix were compared to the results obtained with thin� lms of silica dispersed on a CsI window and to ATRexperiments using a silica/PE � lm on a ZnSe IRE. HMDSreacts with all isolated SiOH groups (band at 3747 cm21)at room temperature 21 according to the following reac-tion:

2SiOH 1 (CH ) SiNHSi(CH )3 3 3 3

® 2 Si–O–Si(CH ) 1 NH3 3 3

Similar spectral features are obtained for the differencespectra for HMDS adsorbed on silica (Fig. 6a) and onsilica/PE (Fig. 6b) recorded in transmission as well as theATR spectrum on silica/PE (Fig. 6c). The differencespectra in Fig. 6 are all referenced against a backgroundof either the pure silica (Fig. 6a) or the silica/PE coating(Fig. 6b and 6c). Although we plot difference spectra,analysis of the single-beam spectra before and afterHMDS exposure shows that the decrease in the bands at3747 cm21 represents 100% reaction with the isolated

APPLIED SPECTROSCOPY 659

FIG. 4. (a) Expanded view of high-frequency region of Fig. 3a. (b)Expanded view of high-frequency region of Fig. 3b.

FIG. 5. Schematic diagram of the dried silica/PE coating depositedonto a ZnSe IRE.

FIG. 6. Difference spectra after exposure to HMDS for 10 min, fol-lowed by evacuation for 5 min. (a) Fumed silica on a CsI window intransmission mode. (b) Silica/PE coating on a CsI window in transmis-sion mode. (c) Silica/PE coating on a ZnSe IRE in ATR mode.

surface hydroxyl groups. The appearance of a lower in-tensity of the negative band at 3747 cm21 in the ATRspectrum is due to the wavelength penetration depth de-pendence of the evanescent wave.22 Both Fig. 6b and 6chave a negative band at 3690 cm21, and this featureshows that the surface hydroxyl groups, which were ini-tially perturbed by the PE binder, are able to participatein the reaction with HMDS. The assignments of the otherbands have been reported.21 The presence of a band inthe 1060 cm21 region is indicative of chemisorption ofthe HMDS with the silica surface through formation ofSis–O–Si linkage. The slight difference in the shape ofthe band at 1060 cm21 in the transmission spectra (Fig.6a and 6b) as compared to the ATR spectrum (Fig. 6c)is due to the effect of penetration depth. Slight changesin the refractive index of the region containing the strongSi–O–Si bulk modes will have an effect on band intensityin ATR spectra. All other positive bands are various C–H modes. The similarity of the spectra in Fig. 6 showsthat the presence of the polyethylene binder does not hin-der the detection of bands due to adsorbed species or theirreaction with the surface hydroxyl groups for either trans-mission or ATR studies.

Gas-Phase Reaction of APDMES Treated Silica.The main purpose of using the PE as a matrix in the ATRexperiments is to provide a means for anchoring the silicaparticles to the crystal in order to prevent their removalin a � owing stream of water. We have found no loss ofintensity in the bands due to silica when a silica/PE coat-ing on a ZnSe crystal is mounted in an ATR � ow-throughcell and subjected to several hours of � owing water at arate of 10 mL/min. To demonstrate the suitability of thisarrangement for adsorption studies from aqueous solu-tion, we examined the reaction of succinic acid with anaminosilanized silica surface.

The anticipated reaction is shown in the ‘‘ideal’’ armin Fig. 7. The reaction of the APDMES with the silica/PE surface was conducted in the gas phase according tothe same procedures outlined for the HMDS reaction.The spectrum obtained after reaction with APDMES isshown in Fig. 8 and is similar to that recently reported.23

In brief, the negative SiOH bands at 3747 and 973 cm21,the absence of the strong Si–O–C modes at 1118 cm21

for the gaseous molecule (not shown), and the appearanceof an Si–O–Si band at 1050 cm21 are clear evidence ofa surface Si–O–Si bond formed from a reaction of a sur-face SiOH group with the ethoxy group of APDMES.Furthermore, the amine group is not freely dangling fromthe surface but is instead hydrogen bonded to surfaceSiOH groups. The N–H bending mode appears at 1596cm21 instead of 1622 cm21 for the free amine. Thus, amore realistic picture for the adsorbed APDMES isshown in the ‘‘our work’’ arm and not the ‘‘ideal’’ armin Fig. 7. While the APDMES reaction was performeddirectly on the silica/PE ATR � lm, it is noted that thistreatment could have been performed ex situ on the silicaparticle and then dispersed in the PE solution. In thisway, a surface-modi� ed metal oxide particle could beprepared under more severe reaction conditions (high re-action temperatures, organic solvents, etc.) than could betolerated by the PE binder. For example, in our case, thesilica powder could be pretreated with APDMES in a gas

660 Volume 55, Number 6, 2001

FIG. 7. Ideal and proposed reaction scheme for the attachment of succinic acid to an APDMES-treated silica surface.

FIG. 9. ATR aqueous solution spectra. (a) A 6 3 1022 molar succinicacid solution. (b) A 1:1 mixture of APDMES and succinic acid. (c) A1.4:1 mixture of APDMES and succinic acid.

FIG. 8. ATR gas-phase difference spectrum of the silica/PE coatingafter exposure to APDMES for 10 min, followed by evacuation for 5min.

phase at 200 8C, then dispersed in the PE coating for lateraqueous-based adsorption measurements.

Aqueous-Phase Reaction of Succinic Acid with anAPDMES Treated Silica. To aid in the assignment ofthe bands due to the surface adsorbed species, we � rstundertook a series of solution-phase experiments in theabsence of the silica/PE layer. The difference spectra inFig. 9 are the results after subtraction of pure water fromthe solution spectra. Figure 9a shows the spectrum of a6 3 1022 molar solution of succinic anhydride. The suc-cinic anhydride hydrolyses to a dicarboxylic acid, and thebands at 1780 and 1720 cm21 are due to the carbonyl

group, while the band at 1408 cm21 has been assigned tothe deformation frequency of a methylene group adjacentto a carbonyl group.24 Figure 9b shows the resulting spec-trum of a 1:1 mixture of APDMES and succinic acid.The formation of an amide linkage is indicated by theappearance of the amide I band at 1640 cm21 and theamide II band at 1555 cm21. This result is accompaniedby a decrease in the 1720 and 1780 cm21 bands due tothe decrease in the amount of free carbonyl groups of theacid. The positive band at 1258 cm21 has been previouslyassigned to the Si–CH3 bending mode of the APDMES.A growth in the band at 1400 cm21 along with a shift

APPLIED SPECTROSCOPY 661

FIG. 10. ATR difference spectra after exposure of the APDMES-treat-ed silica/PE coating to a 1023 molar solution of succinic acid. (a) Fiftypercent APDMES coverage. (b) Full APDMES coverage.

towards lower frequencies indicates that this band is acombination of the methylene deformation mode of thediacid and the amide linkage. Figure 9c is the result ofadding an excess of the aminosilane to the succinic acid;this solution contained an APDMES-to-acid ratio of 1.4:1. All traces of the free carboxylic acid have disappeared,accompanied by an increase in the amount of amide link-ages. The contribution of the amide linkage to the 1400cm21 is seen by an increase in the absolute intensity, ac-companied by a further shift to a lower frequency.

Figure 10 shows the results of two experiments wherea 1023 molar solution of succinic acid was allowed to� ow over the APDMES-treated silica/PE coating. Thespectra in Fig. 10 are referenced against the silica/PEcoating in pure water and differ from the spectra obtainedat high-solution concentrations and in the absence of thesilica/PE layer (Fig. 9b and 9c) by the absence of thestrong amide I band. The absence of the amide I band inthe spectra of Fig. 10 is due to masking by the water-bending mode at 1638 cm21. The deposition of the silica/PE coating causes an increase in the refractive index ofthe interfacial region, as compared to the solution sys-tems of Fig. 9. The result is an increase in the depth ofpenetration of the evanescent wave. The water-bendingmode for the ZnSe/solution system of Fig. 9 has an ab-sorbance of 1.3, whereas for the silica/PE system the ab-sorbance is increased to 1.9.

The spectra in Fig. 10 appeared immediately with con-tact of the succinic acid and did not change with time,indicating that the reaction proceeded rapidly. In Fig. 10a,the appearance of the amide II band at 1555 cm21 indi-cates reaction of the succinic acid with the amine groupof the APDMES. There is also evidence of free carbox-ylic acid, as indicated by the band at 1720 cm21. How-ever, in comparing Fig. 10b to 10a, we � nd that the 1720cm21 band due to the free acid is much weaker in inten-sity, while the absolute intensity of the amide II band hasincreased. The difference between the two samples is thatthe amount of initially attached APDMES on silica used

in the experiment to generate Fig. 10b was twice that forthe silica sample used in Fig. 10a.

These results clearly indicate that the experimentalconditions associated with the gas-phase reaction ofAPDMES with the silica surface have a profound effecton the structure of the functionalized surface. Speci� cal-ly, it is shown that the initial APDMES surface densityhas an effect on the number of acid functionalities dan-gling out from the surface. At the higher APDMES den-sity there is a greater percentage of diacid molecules re-acting difunctionally to form amide linkages with twoadsorbed APDMES species. A more realistic picture ofthe reaction of succinic acid with the APDMES-treatedsilica is shown in the ‘‘our work’’ arm of Fig. 7. Therelative number of free carboxylic groups is importantbecause these sites are used to attach amino-terminatedoligonucliotides. Thus, a variation in carboxylic acidgroups will translate to a variation in number of attachedoligonucliotide probe molecules.

CONCLUSION

The results presented here clearly demonstrate the abil-ity of the described technique to follow the surface chem-ical reactions that occur from aqueous solutions on thesurface of silica particles. These results would not be pos-sible without the higher surface area and extended rangeadvantages of using silica powders over traditional meth-ods using oxidized silicon wafers.

While we have shown how this technique can be usedto study adsorption on silica particles, it is noted that italso can be used for studies of infrared adsorption ontoany small particulate matter. It is also noted that our stud-ies to date have focused on relatively small molecule ad-sorption studies. At this time, it is not known whetherthe presence of the PE impedes the access of larger poly-meric molecules, and work in this area is currently un-derway. On a � nal note, the PE layer can be used totemperatures of 50 8C, which is suf� cient for most aque-ous-based processes. Slightly higher temperatures arepossible with the use of higher-molecular-weight PE.

ACKNOWLEDGMENTS

This work was supported by the Paper Surface Science Consortiumand the Department of the Navy, Naval Surface Warfare Center, Dohl-gren Division, Grant #N00178-1-9002.

1. R. K. Iler, The Chemistry of Silica (John Wiley and Sons, NewYork, 1979).

2. M. L. Hair, Infrared Spectroscopy in Surface Chemistry (MarcelDekker, New York, 1967).

3. A. V. Kiselev and V. I. Lygin, Infrared Spectra of Surface Com-pounds (John Wiley and Sons, New York, 1975).

4. H. E. Johnson and S. Granick, Macromolecules 23, 3367 (1990).5. R. P. Sperline, S. Muralidharan, and H. Freiser, Langmuir 3, 198

(1987).6. J. J. Kellar, W. M. Cross, and J. D. Miller, Appl. Spectrosc. 43,

1456 (1989).7. J. Drelich, W.-H. Jang, and J. D. Miller, Langmuir 13, 1345 (1997).8. M. L. Free and J. D. Miller, Langmuir 13, 4377 (1997).9. D. Neivandt, M. L. Gee, C. P. Tripp, and M. L. Hair, Langmuir 13,

2519 (1997).10. D. J. Neivandt, M. L. Gee, M. L. Hair, and C. P. Tripp, J. Phys.

Chem. B 102, 5107 (1998).11. C. E. Giacomelli, M. G. E. G. Bremer, and W. Norde, J. Colloid

Interface Sci. 220, 13 (1999).

662 Volume 55, Number 6, 2001

12. R. P. Sperline, J. S. Jeon, and S. Raghavan, Appl. Spectrosc. 49,1178 (1995).

13. C. P. Tripp and M. L. Hair, Langmuir 7, 923 (1991).14. S. J. Hug and B. Sulzberger, Langmuir 10, 3587 (1994).15. P. E. Poston, D. Rivera, R. Uibel, and J. M. Harris, Appl. Spectrosc.

52, 1391 (1998).16. L. Han, T. M. Niemczyk, Y. Lu, and G. P. Lopez, Appl. Spectrosc.

52, 119 (1998).17. K. Kajihara, K. Nakanishi, K. Tanaka, K. Hirao, and N. Soga, J.

Am. Ceram. Soc. 81, 2670 (1998).18. M. Schena, Microarray Biochip Technology, (Eaton Publishing,

Natick, Massachusetts, 2000).

19. C. P. Tripp and M. L. Hair, Langmuir 11, 149 (1995).20. G. Curthoys, V. Y. Davydov, A. V. Kiselev, S. A. Kiselev, and B.

V. Kuznetsov, J. Colloid Interface Sci. 48, 58 (1974).21. J. R. Combes, L. D. White, and C. P. Tripp, Langmuir 15, 7870

(1999).22. N. J. Harrick, Internal Re� ection Spectroscopy (John Wiley and

Sons, New York, 1987).23. L. D. White and C. P. Tripp, J. Colloid Interface Sci. 232, 400

(2000).24. L. J. Bellamy, The Infra-red Spectra of Complex Molecules (Chap-

man and Hall, Thetford, UK, 1975).