single-molecule vibrational spectroscopy

7/28/2019 Single-Molecule Vibrational Spectroscopy

1/4

Single-Molecule Vibrational Spectroscopyand Microscopy

B. C. Stipe, M. A. Rezaei, W. Ho*

Vibrational spectra for a single molecule adsorbed on a solid surface have been obtainedwith a scanning tunneling microscope (STM). Inelastic electron tunneling spectra for anisolated acetylene (C

2H2) molecule adsorbed on the copper (100) surface showed an

increase in the tunneling conductance at 358 millivolts, resulting from excitation of theC-H stretch mode. An isotopic shift to 266 millivolts was observed for deuteratedacetylene (C

2D2). Vibrational microscopy from spatial imaging of the inelastic tunneling

channels yielded additional data to further distinguish and characterize the two isotopes.Single-molecule vibrational analysis should lead to better understanding and control ofsurface chemistry at the atomic level.

Vibrational spectroscopy is a powerful toolfor the analysis of molecules adsorbed onsurfaces. Knowledge of the active vibration-al modes of a molecule, as well as the vi -brational energies, can lead to an under-

standing of its adsorption site, orientation,and changes in bonding upon adsorption.This understanding is necessary for the elu-cidation of surface phenomena and fortechnologically important processes suchas heterogeneous catalysis and epitaxialgrowth. Vibrational spectra are also finger-prints of adsorbed molecules and can beused for chemical identification.

In 1966 it was discovered that vibration-al spectra can be obtained from moleculesadsorbed at the buried metal-oxide inter-face of a metal-oxide-metal tunneling junc-tion (1). In that experiment, the tunnelingcurrent I was measured as a function of

voltage V across the junction. Small, sharpincreases in the ac tunneling conductance,dI/dV, were observed when the energy ofthe tunneling electrons reached the energyof a vibrational mode for molecules in thejunction. This increase is the result of elec-trons losing their energies to the vibrationalmode, giving rise to an inelastic tunnelingchannel, which is forbidden when tunnel-ing electrons have energies below the quan-tized vibrational energy. In the experiment,a peak at each vibrational energy was ob-served in d2I/dV2. This method, known asinelastic electron tunneling spectroscopy

(IETS), has been applied to a wide range ofsystems and has led to a better understand-ing of molecules in the adsorbed state (2).However, IETS has a major drawback: Mol-ecules are buried within the junction in acomplex environment that is difficult tocharacterize. For this reason, IETS is not asbroadly applicable as electron energy loss

spectroscopy (EELS) and infrared spectros-copy, which can be performed on exposed,well-ordered surfaces in ultrahigh vacuum(UHV). However, IETS is about 10 timesas sensitive as these other techniques (3); as

few as 109

molecules are needed within thejunction to give a useful spectrum (4).Here, we used a UHV scanning tunnel-

ing microscope (STM) to extend the IETSmethod to a single molecule. In this ar-rangement, the metal-oxide-metal tunneljunction is replaced by the STM tunneljunction: a sharp metal tip, a vacuum gapof several angstroms, and a surface withthe adsorbed molecules. The STM is ca-pable of imaging surfaces with atomic res-olution, and therefore the moleculesbonding environment with respect to anordered surface can be determined precise-ly. By probing individual molecules, it

should be possible to explore how vibra-tional properties are affected by neighbor-ing coadsorbed species or surface defects.It is widely believed that such local effectsplay a critical role in surface chemistry.The combination of atomic resolution andvibrational spectroscopy also allows thecreation of atomic-scale spatial images ofthe inelastic tunneling channel for eachvibrational mode, in a manner similar tothat used to map out the electronic den-sity of states with the STM (5). Thus,individual adsorbed molecules can beidentified by their vibrational spectra and

inelastic images. In contrast, identifica-

tion and characterization by electronicspectroscopy is problematic for two rea-sons: considerable broadening and shiftingof the electronic energy levels occurs uponadsorption, and the adsorbed moleculesspectrum becomes convolved with theSTM tips electronic spectrum.

The possibility of vibrational spectrosco-py with the STM was apparent soon after itsinvention. Theoretical estimates of themaximum changes in the ac conductanceacross the vibrational excitations ranged

from 1 to 10% (69). An experiment prob-ing the ac conductance over a cluster ofsorbic molecules on graphite at 4 K reportedextremely large (up to 1000%) jumps,which were attributed to characteristic vi-brations of the molecules (10). However,no reproducible spectra were obtained be-cause of diffusion events (molecules jump-ing) during the experiment. It is also possi-ble that these diffusion events were respon-sible for some of the spectral features seen.In another experiment, a microscopic tun-nel junction was created by adjusting thepressure between two crossed wires coatedwith a thin film (11). Two peaks wereobserved in d2I/dV2 at bias voltages of 359and 173 mV; these were assigned to theC-H stretch and bending vibrations of hy-drocarbon contaminants, respectively. Theseresults demonstrated that IETS could beachieved in a microscopic tunneling junc-tion. Vibrational spectroscopy with theSTM has proved difficult because of the

extreme mechanical stability necessary toobserve small changes in tunneling conduc-tance. For example, the physics of tunnel-ing dictates a tunneling gap stability of0.01 to keep the conductance stable towithin 2%.

Details of our homemade STM and ex-perimental setup have been described else-where (12). The STM tip and Cu(100)sample were prepared and cooled to 8 K(13). The surface was dosed with acetyleneuntil a coverage of 0.001 monolayer wasobserved on the surface with the STM. Ifdesired, the surface was subsequently dosedwith deuterated acetylene to the same cov-

erage. Although the identity of the isotopeswas known by comparing images before andafter each dose, they were not distinguish-able in the constant-current STM images.

A definitive determination of the ad-sorption site and orientation of acetyleneon Cu(100) has not been reported to date,although the C-C bond is known to beparallel to the surface (14). Considerablerehybridization of the molecule occurs uponadsorption, leading to a lengthening of theC-C bond (14) and a concomitant tilting ofthe C-H bond away from the surface (15).The atomically resolved STM images (Fig.

1) indicate that the molecule is adsorbed onthe hollow sites of the surface, with theplane of the molecule across the diagonal ofthe square formed by four Cu atoms. This isconsistent with one of two proposed adsorp-tion sites, as determined from EELS data(16), and with the proposed site for acety-lene on Ni(100) (17). There are two equiv-alent adsorption sites rotated 90 from eachother. The reversible rotation of the mole-cule between these two orientations can beinduced by the tunneling current. There-fore, it is desirable to position the tip over

Laboratoryof Atomicand Solid State Physics andCenterfor Materials Research, Cornell University, Ithaca, NY14853, USA.

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

SCIENCE VOL. 280 12 JUNE 1998 www.sciencemag.org1732


2/4

the center of the adsorbed molecule so thatthe rotational motion does not affect thetunneling current in the recording of thevibrational spectra.

In IETS, changes in the ac conductance, dI/dV, appear at characteristic tunnel-ing voltages (Vv, defined as vibrationalvoltages) such that eVv h, where e is thecharge of the electron, h is the Planck con-

stant, and is the vibrational frequency.Vibrational spectra were obtained with theuse of an iterative tracking scheme to posi-tion the tungsten tip at the center of themolecule, with lateral and vertical resolu-tions of better than 0.1 and 0.01 , respec-tively. This was done by locating a localminimum of the tip height with the STMsfeedback loop turned on to maintain con-stant tunneling current. For sharp tips, thisminimum did not occur at the center of themolecule; offsets were added to position thetip in these cases. Tracking of the symmetrypoint of acetylene was carried out at a sam-

ple bias of 250 mV and tunneling current of1 nA. Feedback was then turned off. Thetip remained stationary while the vibration-al spectrum was recorded. A sinusoidalmodulation at 200 Hz and 2 to 20 mV rmsamplitude was added to the dc sample biasvoltage. The tunneling current was fed into

a lock-in amplifier to determine the firstand second harmonics of the modulationfrequency (these are proportional to dI/dVand d2I/dV2, respectively). These signalswere recorded as the sample bias voltagewas swept from 0 to 500 mV and back downto 0 mV in 2 min (18). After each trian-gular sweep of the bias voltage, feedbackwas turned on to allow automatic tracking

of the molecule before the next sweep.Spectra were averaged after each pass, al-though the C-H (and C-D) stretch peak ind2I/dV2 was usually distinguishable from thebackground after a single sweep. The noiselevel of the tunneling current during theseexperiments was measured to be 25 fAHz1/2 at 1 nA, which is close to the theo-retical shot noise limit of 18 fA Hz1/2 atthis current. The I-V curves were measuredseparately.

In addition to the spectra obtained overthe symmetry point of a molecule, spectrawere also recorded over a clean area of the

surface by displacing the tip by

20 .Contributions from the electronic spectrumof the tip and substrate can be minimized bysubtracting the clean spectra from the mo-lecular spectra. The I-V curves for a singleC2H2 molecule and the clean surface (Fig.2A) show the expected linear dependence

for metallic junctions. An increase of 4.2%in dI/dV at the vibrational voltage can beseen in the difference spectrum (Fig. 2B).The d2I/dV2 spectrum taken over the mol-ecule reveals a distinct peak at 358 mV (Fig.2C). In addition to enhancing the signal-to-noise ratio, the long signal average timeof 10 hours over the symmetry point of asingle molecule at 2 mV modulation (line 3

in Fig. 2C) demonstrates the stability of theexperiment. Here, a different tip was usedthat gave an increase of 3.3% in the acconductance. The size of the peak wasfound to be tip dependent, with sharper tips(as determined from the resolution in theSTM images) giving larger vibrationalpeaks. A change in ac conductance across

low highlow high

Fig. 1. (A) STM imageof a C

2H2

molecule onthe Cu(100) surface at 8K. Acetylene appears asa dumbbell-shaped de-pression on the surfacewith a maximum depth

of 0.23 . The stableclean metal tips neces-sary to perform IETSrarely yielded atomic res-olution of the Cu(100)surface. The imagedarea, 25 by 25 , wasscanned at a samplebias of 100 mV and tun-neling current of 10 nA.(B) The molecule in (A)wastransferred to the tipby means of a voltagepulse (0.6 V, 100 nA,1 s), and the same areawasscanned.The atom-

ic-resolution imagehas acorrugation of 0.009 .This corrugation is sensi-tive to the nature of thetip and the tunneling pa-rameters. Copper atomspacing is 2.55 . The image was scanned at a sample bias of 10 mV and tunneling current of 10 nA.(C) The atoms in (B) were fitted to a lattice. The lattice is shown on top of the image in (A). (D) Schematicdrawing showing side and top views of the molecules orientation and suggested adsorption site. Theadsorption site determination assumes that the transfer of the C

2H2

molecule to the tip did not changetheposition of thetips outermostatom. Thedashed line shows theoutline of theSTM image shape.Thedumbbell-shaped depression in STMimages mayresult from bonding to the Cu atoms perpendicularto the C-C axis, reducing the local density of states for tunneling. This would cause the axis of thedumbbell shape to be perpendicular to the plane of the molecule (rather than parallel).

Fig. 2. (A) I-Vcurves recorded with the STM tipdirectly over the center of molecule (1) and overthe bare copper surface (2). The difference curve(1 2) shows little structure. Each scan took10 s. (B) dI/dV from the lock-in amplifier record-ed directly over the center of the molecule (1)and over the bare surface (2). The differencespectrum (1 2) shows a sharp increase at asample bias of 358 mV (arrow). Spectra are theaverage of 25 scans of 2 min each. A samplebias modulation of 5 mV was used. (C) d2I/dV2

recorded at the same time as the data in (B)directly over the molecule (1) and over the baresurface (2). The difference spectrum shows apeak at 358 mV. The area of the peak gives theconductance change, (/ 4.2%). A dif-ference spectrum taken with 2 mV modulationand a different tip (3) shows a smaller peak(/ 3.3%). The data in (3) are the result of anaverage of 279 scans of 2 min each (10 hourstotal scanning time directly over the molecule).Data in (B) were calibrated by numerical integra-tion and comparison with the data in (A). Data in(C) were calibrated by numerical integration andcomparison with the calibrated data from (B).

REPORTS

www.sciencemag.org SCIENCE VOL. 280 12 JUNE 1998 1733


3/4

the C-H peak of 11.8% was measured forone sharp tip. The peak was observed forevery tip used, and spectra were reproduc-ible for a given tip. The peak position re-mained the same to within 1 mV evenwhen the tunneling current was increasedby a factor of 10. Spectra measured withnegative bias voltages showed the expectednegative peak at 358 mV. Therefore, no

electric fielddependent peak shifts havebeen measured.

The full width at half maximum(FWHM) of the d2I/dV2 vibrational peak isgiven by V [(1.7Vm)

2 (5.4kT/e)2

W2]1/2, where Vm is the modulation volt-age, k is the Boltzmann constant, T is tem-perature, and W is the intrinsic width (2).Typical values of the FWHM for the 358-mV peak obtained at modulation voltagesof 20, 10, 5, and 2 mV are 34, 17, 12, and 10mV, respectively. Comparison with theabove formula suggests an intrinsic width of

8 mV, larger than is typically seen inmacroscopic IETS.

Isotopic substitution is a powerful toolfor vibrational analysis. The C-H stretch at358 mV was observed to shift to 266 mV forC2D2 (Fig. 3). These values are in closeagreement with those obtained by EELS:357, 360, and 362 meV for the C-H stretchof C2H2 on Cu(100) (16), Cu(110) (19),

and Cu(111) (20), respectively; 272 and270 meV for the C-D stretch of C2D2 onCu(110) (19) and Cu(111) (20), respec-tively. Changes in ac tunneling conduc-tance were typically 30% lower for theC-D stretch relative to the C-H stretch.Other vibrational modes of adsorbed acety-lene, such as the bending modes, were notobserved. The prominence of the C-Hstretch mode has also been observed in themacroscopic (1, 2) and microscopic (11)IETS measurements.

The ability to spectroscopically identifymolecules with the STM makes it possible

to implement chemically sensitive micros-

copy. Vibrational imaging of the adsorbedmolecule was obtained while scanning thetip in constant-current mode to record thesurface topography. At each data point,the feedback was turned off and the biasmodulation turned on to record dI/dV andd2I/dV2. This procedure resulted in threeimages of the same area. As expected, nocontrast was observed in a constant-cur-rent image of both acetylene isotopes (Fig.4A). When the dc bias voltage was fixedat 358 mV, only one of the two moleculeswas revealed in the image constructedfrom the d2I/dV2 signal (Fig. 4B). By

changing the dc bias voltage to 266 mV,the other molecule was imaged (Fig. 4C).Two small identical depressions observedat 311 mV (Fig. 4D) were attributed to the

change in the electronic density of stateson the sites of the two molecules. Thespatial extent of the inelastic tunnelingchannels appears to be at least as narrowas the constant-current molecular image.

For sharp tips, values of / of up to12% and 9% have been measured for theC-H and C-D stretch peaks, respectively.This is considerably larger than the values

obtained in conventional IETS. The dipolescattering approximation usually used tomodel IETS spectra yielded theoretical es-timates of/ in STM-IETS on the orderof 1% (6, 7). It has been suggested, howev-er, that the dipole scattering approximationmay not be valid for the molecular-sizedtunneling gaps in the STM, and that reso-nance and impact scattering may lead toenhancement in / by a factor of 10(8, 9). The size of the observed peaks, thespatial extent of the inelastic channels, andthe strong tip dependence are consistentwith the resonance scattering mechanism.

The relatively large intrinsic peak widthmay also be a result of the resonant tunnel-ing mechanism.

Vibrational spectroscopy at the single-molecule level opens new avenues of re-search, both experimental and theoretical.The extension of our work to other func-tional groups should allow the characteriza-tion of small molecules in more complexenvironments, as well as the analysis oflarger molecules of chemical and biologicalinterest. It may soon be possible to usesingle-molecule vibrational spectroscopyand microscopy to determine the identityand arrangement of functional groups with-

in a single molecule and to study its chem-ical transformation.

REFERENCES AND NOTES___________________________

1. R. C. Jaklevic and J. Lambe, Phys. Rev. Lett. 17,1139 (1966).

2. P. K. Hansma, Phys. Rep. 30C, 145 (1977); W. H.Weinberg, Annu. Rev. Phys. Chem. 29, 115 (1978);T. Wolfram, Ed., Inelastic Electron Tunneling Spec-troscopy(Springer-Verlag, Berlin, 1978); P. K. Hans-ma, Ed., Tunneling Spectroscopy (Plenum, NewYork, 1982).

3. H. Ibach and D. L. Mills, Electron Energy Loss Spec-troscopy(Academic Press, New York, 1982).

4. R. M. Kroeker and P. K. Hansma, Surf. Sci. 67, 362(1977).

5. R. J. Hammers, R. M. Tromp, J. E. Demuth, Phys.Rev. Lett. 56, 1972 (1986).

6. G. Binnig, N. Garcia, H. Rohrer, Phys. Rev. B 32,1336 (1985).

7. B. N.J. Persson and J.E. Demuth, SolidState Com-mun. 57, 769 (1986).

8. B.N. J.Persson and A.Baratoff, Phys.Rev. Lett. 59,339 (1987).

9. M. A. Gata and P. R. Antoniewicz, Phys. Rev. B 47,13797 (1993).

10. D. P. E. Smith, M. D. Kirk, C. F. Quate, J. Chem.Phys. 86, 6034 (1987).

11. S. Gregory, Phys. Rev. Lett. 64, 689 (1990).12. B. C. Stipe, M. A. Rezaei, W. Ho, J. Chem. Phys.

107, 6443 (1997). The variable-temperature STMand the sample are in thermal equilibrium inside adouble copper housing cooled by a continuously

Fig. 3. Background difference d2I/dV2 spectra forC2H2

(1) and C2D2

(2), taken with the same STMtip, show peaks at 358 mV and 266 mV, respec-tively. The difference spectrum (1 2) yields amore complete background subtraction. For thistip, / 6.2% and 4.5% for C

2H2

and C2D2,

respectively.

Fig. 4. Spectroscopicspatial imaging of the in-elastic channels for C

2H2

and C2D2. (A) Regular

(constant current) STMimage of a C

2H2

mole-cule (left) and a C

2D2

molecule (right). Data arethe average of the STM

images recorded simul-taneously with the vibra-tional images. The im-aged area is 48 by 48

. d2I/dV2 images of thesame area recorded at(B) 358 mV, (C) 266 mV,and (D) 311 mV are theaverage of four scans of25 min each with a biasmodulation of 10 mV. Allimages were scanned at1 nA dc tunneling current. The symmetric, round appearance of theimages is attributable to therotationof the molecule between two equivalent orientations during the experiment.

SCIENCE VOL. 280 12 JUNE 1998 www.sciencemag.org1734


4/4

flowing liquid He cryostat. Thermal stability of theSTM and the sample to within 0.01 K was reached 4hours after initial cooling from room temperature.

13. Tips were made by etching polycrystalline tungstenwire in KOH. Tips were cleaned in UHV by repeatedcycles of sputtering from field emission in a Ne at-mosphere and annealing with an electron beamheater. Tips were sharpened just before the experi-ment by bringing theend of thetip incontact with theclean copper surface. Thus, the last atom on the tipis likely to be copper rather than tungsten. Cu(100)

was cleaned with repeated cycles of 1 keV Ne

sputtering and annealing at 500C until large defect-free regions of the surface were observed with theSTM.

14. D. Arvanitis, U. Dobler, L. Wenzel, K. Baberschke, J.Stohr, Surf. Sci. 178, 686 (1986).

15. X. F. Hu, C. J. Chen, J. C. Tang, ibid. 365, 319(1996).

16. Ts. S. Marinova and P. K. Stefanov, ibid. 191, 66(1987).

17. N. Sheppard,J. Electron Spectrosc. Relat. Phenom.38, 175 (1986).

18. Drifts were generally negligible during the time of

each voltage sweep. With the feedback off, driftrates of 0.001 min1 and 0.01 min1 weremeasured in the vertical and horizontal directions,respectively. By averagingspectra in the forwardandbackward sweep directions, the resulting small driftin the tunneling conductance was canceled out.

19. N. R. Avery, J. Am. Chem. Soc. 107, 6711 (1985).20. B. J. Bandy, M. A. Chesters, M. E. Pemble, G. S

McDougall, N. Sheppard, Surf. Sci. 139, 87 (1984).21. Supported by NSF grant DMR-9707195.

18 February 1998; accepted 15 April 1998

Combinatorial Electrochemistry: A HighlyParallel, Optical Screening Method for Discovery

of Better Electrocatalysts

Erik Reddington, Anthony Sapienza, Bogdan Gurau,Rameshkrishnan Viswanathan, S. Sarangapani,

Eugene S. Smotkin,* Thomas E. Mallouk*

Combinatorial screening of electrochemical catalysts by current-voltage methods can beunwieldy for large sample sizes. By converting the ions generated in an electrochemical

half-cell reaction to a fluorescence signal, the most active compositions in a largeelectrode array have been identified. A fluorescent acid-base indicator was used toimage high concentrations of hydrogen ions, which were generated in the electrooxi-dation of methanol. A 645-member electrode array containing five elements (platinum,ruthenium, osmium, iridium, and rhodium), 80 binary, 280 ternary, and 280 quaternarycombinations was screened to identify the most active regions of phase space. Sub-sequent zoom screens pinpointed several very active compositions, some in ternaryand quaternary regions that were bounded by rather inactive binaries. The best catalyst,platinum(44)/ruthenium(41)/osmium(10)/iridium(5) (numbers in parentheses are atomicpercent), was significantly more active than platinum(50)/ruthenium(50) in a direct meth-anol fuel cell operating at 60C, even though the latter catalyst had about twice thesurface area of the former.

Combinatorial synthesis and analysis havebecome commonly used tools in bioorganicchemistry (111). Although the combina-torial (or multiple sample) approach to thediscovery of inorganic materials has existedfor more than two decades (12), new mate-rials (13, 14) and sensors (15) only recentlyhave been discovered by this advanced Ed-isonian technique. There are several exam-ples of combinatorial searches that haveconfirmed the properties of known materi-als (16, 17) and catalysts (18). There are,however, no previous reports of superiorcatalysts that have been identified by com-

binatorial chemistry. Combinatorial search-

es for catalysts are often limited not bysynthesis but by the lack of methods thatcan simultaneously screen many composi-

tions. Recently, significant progress hasbeen made in the thermographic screeningof catalysts (18, 19). We describe here ageneral combinatorial screening method forelectrode materials and its application tothe problem of anode catalysis in directmethanol fuel cells (DMFCs). By screeningcombinations of five elements (Pt, Ru, Os,Ir, and Rh), we found several good catalystsin unexpected regions of composition spaceand identified a quaternary catalyst withsignificantly higher activity than the bestpreviously known catalyst (a binary Pt-Rualloy).

Because methanol is a renewable liquidfuel, DMFCs present some distinct poten-tial advantages over combustion enginesand hydrogen fuel cells for transportationand remote power applications. They aresilent and nonpolluting, and they use aneasily distributed, high-energy-density fuel(20). One of the factors that limits theirdevelopment and use is that known anodeand cathode electrocatalysts produce usefulcurrent densities only at high overpoten-tials. The anode catalyst is particularlyproblematic, because it must perform a ki-

netically demanding reaction, the six-elec-tron (6e) oxidation of methanol

CH3OH H2O 3 CO2 6H 6e(1)

Despite nearly three decades of research andoptimization, the best known anode electro-catalysts are a binary Pt-Ru alloy (21) and arecently discovered Pt-Ru-Os ternary (22).The performance of these alloy catalysts issignificantly enhanced relative to Pt alone.The bimetallic alloy is thought to work by a

bifunctional or ligand effect mechanism, inwhich Ru activates water, and Pt activatesthe methanol C-H bond and also binds theCO intermediate in the reaction (23, 24).Although binary alloys of Pt with many oth-er oxophilic elements have been studied,none is more active than high-surface-areaPt-Ru. No predictive model for investigatingternary or quaternary combinations hasemerged and, in fact, little is known aboutthe phase equilibria of ternary and quaterna-ry Pt alloys.

Although electrocatalysts are typicallytested by measuring current as a function ofpotential, this approach becomes increas-

ingly unwieldy as the number of samplesincreases. Successful uses of the combinato-rial methodfor example, massively paral-lel screening for biochemical affinity (19),organic host-guest interactions (10, 11),and inorganic phosphorescence (13)of-ten use optical (absorption or emission)detection. Optical detection methods arefast regardless of the complexity of the ar-ray, are simple to implement, and allow oneto ignore the uninteresting majority ofphase space. This method can be adapted toelectrochemical screening by recognizingthat all half-cell reactions (for example,

reaction 1) involve an imbalance of ions.Therefore, a fluorescent indicator that de-tects the presence or absence of ions (H inthe case of the DMFC anode) in the diffu-sion layer images the activity of the array.

This idea is illustrated in a small ternaryPt-Os-Rh array, prepared by manually pipet-ting appropriate metal salts and aqueous

NaBH4 onto Toray carbon paper (Fig. 1).The carbon substrate is electronically con-ducting but not catalytic for the oxidation ofmethanol. The array electrode serves as theworking electrode in a single-compartment,

T. E. Mallouk, E. Reddington, A. Sapienza, Departmentof Chemistry, The Pennsylvania State University, Univer-sity Park, PA 16802, USA.E. S. Smotkin, B. Gurau, R. Viswanathan, Department ofChemical and Environmental Engineering, Illinois Instituteof Technology, Chicago, IL 60616, USA.S. Sarangapani, ICET, Inc., Norwood, MA 02062, USA.

*To whom correspondence should be addressed. E-mail:[email protected] (E.S.S.) or [email protected] (T.E.M.).

REPORTS

www.sciencemag.org SCIENCE VOL. 280 12 JUNE 1998 1735

single-molecule vibrational spectroscopy

Documents