Probing Molecular Junctions Using Surface Plasmon Resonance

NANO LETTERS

Probing Molecular Junctions Using Surface Plasmon Resonance Spectroscopy

2006 Vol. 6, No. 12 2797-2803

Ken T. Shimizu, Ragip A. Pala, Jason D. Fabbri, Mark L. Brongersma, and Nicholas A. Melosh* Department of Materials Science and Engineering, Stanford UniVersity, Stanford, California 94305 Received August 11, 2006; Revised Manuscript Received October 5, 2006

ABSTRACT The optical absorption spectra of nanometer-thick organic films and molecular monolayers sandwiched between two metal contacts have been measured successfully using surface plasmon resonance spectroscopy (SPRS). The electric field within metal−insulator (organic)-metal (MIM) cross-bar junctions created by surface plasmon-polaritons excited on the metal surface allows sensitive measurement of molecular optical properties. Specifically, this spectroscopic technique extracts the real and imaginary indices of the organic layer for each wavelength of interest. The SPRS sensitivity was calculated for several device architectures, metals, and layer thicknesses to optimize the organic film absorptivity measurements. Distinct optical absorption features were clearly observed for R6G layers as thin as a single molecular monolayer between two metal electrodes. This method also enables dynamic measurement of molecular conformation inside metallic junctions, as shown by following the optical switching of a thin spiropyran/polymer film upon exposure to UV light. Finally, optical and electrical measurements can be made simultaneously to study the effect of electrical bias and current on molecular conformation, which may have significant impact in areas such as molecular and organic electronics.

Understanding molecular configuration and behavior at the interface with metallic contacts is becoming increasingly important as more organic devices are being explored in applications for field-effect transistors,1 photovoltaics,2 lightemitting devices,3 and molecular electronics.4,5 These devices rely upon electron or hole injection from the metal and subsequent transport through the organic species. It is thus important to realize that charging effects, electric fields, and redox chemistry can lead to extensive molecular reconfiguration and thereby impact device operation. This is especially true in molecular electronics, where nanoscale junctions of molecular monolayers are placed between metal electrodes to manipulate electron flow (e.g., conjugated molecular wires, saturated alkane molecular resistors, molecular diodes), and have been characterized using primarily current-voltage measurements.6,7 Voltage-induced switching has been reported in oligophenylene ethylenes (OPE’s),7 catananes, and rotaxane devices,8 yet also unexpectedly within rather passive molecules such as alkanethiols.9 Recently, the switching behavior of these nonfunctional molecules has been implicated as metal filament formation10 or a bonding-debonding mechanism,11 yet the actual mechanisms are still not understood. The controversy highlights the uncertainty of determining molecular junction switching * Corresponding author. E-mail: [email protected]. 10.1021/nl061893h CCC: $33.50 Published on Web 10/31/2006

© 2006 American Chemical Society

mechanisms based upon current-voltage traces alone. Simultaneous observation of the optical absorption spectrum of the organic layer during electrical operation is particularly appealing because optical absorption features are highly sensitive to the electronic configuration and energy levels of the molecule. However, absorption measurements have been stymied by the highly absorptive/reflective nature of the electrical contacts. Although metal contacts are used ubiquitously in electrical connections, they need not limit optical characterization of the underlying organic film. Previous studies have addressed molecular optical properties in organic or molecular junctions after top-contact electrode deposition using Raman and IR spectroscopy.12,13 However, in those studies optical and electrical measurements could not be performed simultaneously because of differing geometric constraints between the electrical and optical devices. Though losses due to metal absorption will always be present, local enhancement of the electromagnetic field at a metallic interface using surface plasmons may be sufficient to enable measurement of the molecular properties. Surface plasmon-polaritons (SPP) are collective electron density waves propagating at the surface of a metal, which create an evanescent transverse magnetic wave whose intensity decays exponentially from the surface. Recently, these waves have attracted significant attention for their

Figure 1. (a) Schematic of the Kretchmann configuration. Light is coupled into the surface plasmon-polariton (SPP) through an SF10 hemispherical prism. At a critical angle, the light is resonantly converted into SPP that evanescently decays into the metal and air layers. (b) Typical reflectance trace and model fit using the Fresnel equations. (c) Calculated electric field strength profile for surface plasmon modes of the metal-insulator-metal (MIM) structure. The MIM “slow” modes inside the organic layer are inaccessible coupling via the Kretchmann configuration because of the much higher wavevectors.

ability to manipulate the flow of light at the nanoscale.14,15 Because of the subwavelength confinement of the electromagnetic wave, the field intensity at the metal surface can be hundreds of times higher than the incident radiation. Information about the local environment can be derived from the efficiency of coupling incident light with a particular electromagnetic wavevector, kˆ ) 2πnˆ /λ, into SPPs. This process is extremely sensitive to the real (n) and imaginary (k) indices of refraction of materials near or in contact with the metal, where nˆ ) n + ik. Surface plasmon resonance (SPR) sensing is a widely used technique that exploits this sensitivity to obtain the real part of the refractive index of thin films on top of a metallic surface, for instance to measure protein binding constants.16 However, this technique is not limited to only metal-insulator layer (MI) geometries because the Fresnel calculations can predict the surface plasmon excitation efficiency for any given number of layers and layer indices. The optical constants for these layers are then determined from the best fit to the experimentally measured plasmon coupling curve as a function of incident wavevector. Experimentally, surface plasmon resonance experiments are usually performed by measuring the specularly reflected light from the metal surface over a range of incident wavevectors. In the Kretchmann configuration, a prism is used to increase the wavevector of the incident light in order to match the higher wavevectors necessary to excite surface plasmons. In this geometry, shown in Figure 1a, the incident wavevector parallel to the metal surface is varied by changing the incident angle θ, such that kˆ inc ) 2πnprism sin θ/λ. When this incident wavevector matches that of the SPP mode, coupling occurs, and its efficiency is measured by the reflection from the metal surface. A significant decrease in the amount of reflected light occurs at the resonance angle where the incident light has coupled into the SPP mode. A reflectivity curve (black squares) and an accompanying Fresnel fit (red line) for an Al/organic/Au junction at 650 nm are shown in Figure 1b. Although traditionally the angular location of the reflection minimum, θmin, is used to derive the refractive index of the material layers, for many metals and wavelengths the dip is relatively broad (e.g., at 450 nm for Au electrodes), such that determining the minimum accurately is difficult. Fortunately, the entire curve contains information about the real and imaginary components of the layers, even if a sharp dip or minimum is not 2798

observed. Analysis of the SPR curve shape reveals that the location of the minimum reflectivity depends primarily on the real refractive indices, whereas the depth and width of the minima depend upon the imaginary components. The optical extinction coefficient, R, for a particular layer can then be calculated from k R)

4π k λ

(1)

where λ is the wavelength of the incident light. In this Letter, we demonstrate the utility of determining SPR curves for optical analysis of monomolecular and ultrathin film organic layers sandwiched between two metal electrodes. This technique can be applied as a straightforward platform that can perform in situ, noninvasive optical investigation into the state of the organic molecules inside a solid-state device. The optical absorption spectra are obtained by performing the SPR measurement at every wavelength of interest and fitting the reflectivity curves to obtain the imaginary index of the organic species at that wavelength. Surface plasmon resonance spectra were collected in the Kretchmann configuration by rotating an SF10 glass hemispherical prism (n ) 1.72) relative to the incoming light source and measuring the intensity of the reflected light.17,18 The main detector was rotated with respect to the optical axis at a 2θ angle. Because the index calculations fit the absolute magnitudes of reflectivity curves, a reference detector was used to monitor any intensity fluctuations in the light source (an Oriel 1 kW Xenon arc lamp mated to a 300 mm monochromator). The beam is collimated before striking the prism to reduce angular dispersion, which tends to broaden reflectivity features; however, this procedure also limited the minimum spot size to ∼1 mm2. The SF10 glass substrate was index matched to the hemispherical prism by a thin layer of immersion oil. The real (n) and imaginary (k) components of the refractive index for the organic layer were found from a least-squares fitting routine to the reflectivity curves based upon a fivelayer stack of materials: prism/bottom electrode/organic layer/top electrode/air. Each layer was described by a thickness t and complex index (n and k) value. The theoretical reflectivity curves were calculated from the Fresnel equations from these layers based upon Hansen’s matrix formulation.19 Nano Lett., Vol. 6, No. 12, 2006

To reduce the degrees of freedom for fitting the organic optical constants, we determined the parameters of the metal layers independently. For instance, to measure the spectrum of a 13 nm Al/3 nm Al2O3/10 nm RG6-PMMA/ 18 nm Au MIM junction, the reflection curves for the 13 nm aluminum/3 nm Al2O3 and 18 nm gold layers were measured separately (either on part of the sample without the organic layer, or on a reference sample evaporated at the same time), for all wavelengths of interest. The actual metal indices differed slightly from evaporation to evaporation and also differed from published indices of the bulk metals.20 The empirically determined metal parameters were then fixed, and the reflectivity of the MIM junction was fit to obtain the optical constants of the organic layer at every wavelength. The error in the n and k values were obtained from the standard deviation of the fit for each parameter. Although surface plasmon spectroscopy of molecular films on a single bottom metal surface has been shown previously,17,21,22 there are substantial differences in optimizing this technique for the MIM geometry. Most obviously, the SPP mode profiles differ substantially for measuring the optical absorption of molecular films placed between two metal films. Figure 1c illustrates the mode profile along the direction normal to the substrate surface. Although a number of SPP modes are supported in a MIM device, for nanometerthick organic layers, the Kretchmann configuration can only couple into the “fast” mode. This mode has its highest intensity at the top metal/air interface but exhibits some overlap with the underlying dielectric layer and thus enables the probing of the optical parameters of the sandwiched organic material. Although a SPP mode exists that is highly localized in the organic (wavevector >2 × 105 cm-1 in Figure 1c), such “slow” MIM cavity modes are inaccessible by the Kretchmann configuration because the wavevector for the slow mode is too large to couple in with a prism.23 Nonetheless, a well-chosen top metal contact performs two vital functions: as an electrical contact and as a SPP coupler. There are two important parameters to maximize the SPRS response to the absorptivity of the organic layer: optimal metal thicknesses and the metal electrode materials. In choosing these, the objective was to minimize the decrease in the plasmon field strength due to joule heating of the metal and to maximize the resonant coupling of far-field light into surface plasmons. The SPRS sensitivity, defined here as ∆σ2/ ∆k, was used to quantify the change in reflectivity for a change in k in the organic, where ∆σ2 is the magnitude of the reflectivity changes summed over all angles: ∆σ2 )

∑θ |R(n,k + dk,θ) - R(n,k,θ)|2

(2)

Here R is the reflected intensity for a given stack of layers and θ is the angle of the incident light with respect to the metal surface. This metric is different from the SPR sensitivity defined as dθmin/dn, the change in the dip minimum per change in index due to the fact that SPRS uses the entire curve to fit the Fresnel model. This is particularly true for the cases in which significant reflectivity changes occur at angles other than the dip minimum. Nano Lett., Vol. 6, No. 12, 2006

Figure 2 compares the SPRS sensitivities for different metals and metal thicknesses at 450 and 630 nm wavelengths. The sensitivity is plotted as a function of the top (x axis) and bottom (y axis) metal thicknesses, with a fixed 10-nmthick organic layer. These calculations show that, as expected, thinner metals improve the sensitivity to the organic layer; however, different metals have different magnitude effects at different wavelengths. Figure 2a and b demonstrates that, regardless of the wavelength, highly absorptive aluminum films diminish the sensitivity as thickness is increased. Gold-organic-gold devices, Figure 2c and d, show a significant sensitivity difference between 450 and 630 nm because of the strong absorption of Au below 550 nm. Although the gold-organic film-gold geometry provides high sensitivity, the dramatic increase in absorptivity below 550 nm complicates obtaining a flat optical response over the entire visible range. Another alternative is to use a bottom aluminum electrode and a top gold electrode as shown in Figure 2e and f. This metal combination results in a flatter spectral response for obtaining unbiased optical absorption profiles compared to a gold/organic/gold device. Therefore, we have adopted ∼13 nm aluminum bottom electrodes and ∼18 nm gold top electrodes in the devices reported here. A mixed gold and aluminum device also offers other advantages: the aluminum provides a stable, protective oxide for minimizing electrical shorting, and the top gold film can be nondestructively lifted on through a polymer-assisted liftoff (PALO) process24 to minimize molecular film damage during fabrication. Silver metal was not chosen despite the long SPP propagation distance and sharp resonance curves because of pragmatic issues regarding continued surface oxidation during the measurement process that changed the optical parameters. The MIM devices were prepared with spin-cast polymer/ molecular films or Langmuir-Blodgett monolayers deposited onto evaporated bottom electrodes. Electrodes were usually 1-3 mm wide, prepared by e-beam evaporation through a shadow mask onto polished SF10 glass substrates after cleaning in a piranha solution (30% H2O2, 70% H2SO4) at 90 °C, and rinsing thoroughly with ultrapure deionized water. Dye-doped polymer films were made by dissolving Rhodamine 6G or spiropyran (6,8-Dibromo-1′,3′-dihydro-1′,3′,3′-trimethylspiro[2H-1-benzopyran-2,2′-(2H)-indole] in trace methanol and adding to (0.25-1 wt %) PMMA in chlorobenzene solution (MicroChem 495C). The polymer films were cast by spin-coating at 3000 RPM for 30 s, resulting in (10-40 nm) thick films depending upon the initial PMMA concentration. The film thicknesses were confirmed by SPR and atomic force microscopy (AFM). Top metal contacts were deposited either through shadow masked e-beam evaporation, or soft-deposited using the PALO technique to avoid electrical shorts. Having optimized the device geometry, Figure 3a, d, and g shows the reflectivity curves and best fit for a 10-nmthick R6G/PMMA film spin-cast on (i) 18 nm Au, (ii) 13 nm Al, and (iii) sandwiched between 13 nm Al and 18 nm Au films for form an MIM junction. The weight percent of R6G in PMMA was 10%, and a homogeneous distribution 2799

Figure 2. Sensitivity plot of the SPR reflection measurement inside a 10-nm-thick organic layer (n ) 1.45) for varying bottom (x axis) and top (y axis) metal thicknesses. Al (bottom) and Al (top) at (a) 450 nm and (b) 630 nm. Au (bottom) and Au (top) at (c) 450 nm and (d) 630 nm. Al (bottom) and Au (top) at (e) 450 nm and (f) 630 nm.

of dye molecules was assumed. Comparison of the three reflectivity curves confirm that the SPR dip magnitude decreases for the more absorptive Al metal and that significant broadening of the reflectivity dip occurs at shorter wavelengths for Au films. The Fresnel model of the ideal stack fits the measured reflectivity curves within χ2 ≈ 1 over the entire angular range. The real and imaginary components of the organic layer refractive index, and thus molecular absorptivity, determined from the reflection curves are shown in Figure 3b and c. The R6G dye on Au shows two absorption peaks at 525 and 550 nm, whereas the real part of the index has a maximum at 560 nm. An absorption spectrum of a higher concentration R6G/PMMA control sample in the absence of any metal layers was obtained from a standard, normal incidence transmission measurement and is shown for comparison (green line). The spectrum shows excellent agreement between the two, with a very slight broadening on Au. Significantly, the real and imaginary indices show a distinct Kramers-Kronig relationship, even though these values were determined independently from the reflectance fits. The same R6G/PMMA film spun on aluminum electrodes is shown in Figure 3e and f. Comparing the spectra between Au and Al bottom layers (Figure 3c and f), the peak at 550 nm on Al has decreased in intensity (from k ) 0.12 to 0.10) relative 2800

to the shoulder. In other studies focusing on Rhodamine 6G,25 changes in the ratio of peak intensities were attributed to aggregation effects of R6G molecules such as dimer formation. This trend continues for all devices with Al as the bottom electrode, but not where Al is deposited as the top electrode, suggesting that Al changes the molecular aggregation during spin-casting and sample preparation. The organic indices measured on an MIM device of 10 nm R6G/PMMA between a 13 nm aluminum bottom electrode and 18 nm gold top electrode are shown in Figure 3h and i. Although the signal-to-noise ratio has decreased because of the increased metal absorption, the spectral features are quite distinct and agree very well with the MI devices. The spectrum shows roughly equal intensities at 550 and 520 nm, implying increased dimer formation, similar to the Al bottom electrode alone. The peak width decreased from 85 to 80 nm, the cause of which is not known at this point. One important aspect of the SPRS technique is that the absolute magnitude of the absorptivity is constant irrespective of the absorptivity of the total junction. The k index for all devices was roughly constant between 0.09 and 0.1, further reinforcing that the reflectivity fitting measures the actual molecular absorptivity accurately. For a 10 vol % dilution of R6G film, k ) 0.10 corresponds to a molar absorptivity of 9.5 × 104 M-1cm-1, in good agreement with Nano Lett., Vol. 6, No. 12, 2006

Figure 3. (a) SPR reflection curves for 10 nm R6G/PMMA film on 18 nm of Au and the corresponding (b) real and (c) imaginary part of the refractive index derived from part a. A linear absorption spectrum of a R6G/PMMA film taken with a UV-vis spectrometer (solid line). (d) SPR reflection curves for the same film on 15 nm of Al and the corresponding (e) real and (f) imaginary part of the refractive index. (g) SPR reflection curves for the same film sandwiched between 15 nm of Al and 19 nm of Au and the corresponding (e) real and (f) imaginary part of the refractive index.

the measured solution absorptivity of 7.0 × 104 M-1cm-1. Standard transmission absorption measurements of the same MIM device showed no discernible features due to the organic because of the overwhelming absorption from the metal layers. Although polymer/dye films provide an excellent proof of principle, we are further interested in determining the optical properties of extremely thin organic films for application such as molecular electronics. To attain organic films with thicknesses that are comparable to molecular electronic systems (e.g., 1-3 nm), we prepared monolayers of dimyristoylphosphatidic acid, DMPA, and R6G by the LangmuirBlodgett (LB) technique.21 One mg/mL of DMPA and 2 mg/ mL R6G were dissolved in freshly distilled chloroform, and deionized water was used for the subphase. The LB film was compressed to 20 mN/m before it was deposited onto the metal electrodes. Because the DMPA head group is negatively charged, the positively charged R6G molecules assemble electrostatically on the LB film as illustrated in Figure 4a. The thickness of the LB film was determined to be 3 nm by SPR, consistent with previous reports.21 Nearedge X-ray absorption fine structure (NEXAFS) measurements showed that R6G molecules were located near the DMPA phopholipid headgroup and next to the bottom metal surface.26 Even for monolayer-thick films, molecular absorptivity could be clearly measured for both the MI and MIM geometries. Figure 4b shows the angular resolved reflection data and the corresponding theoretical fits, which again fit to better than χ2 ≈ 1. The optical constants for the R6G monolayer on top of a 15 nm Al bottom electrode are shown in Figure 4c and d. Unlike the spin-cast R6G/polymer composites on Al (Figure 3f), the absorption spectrum of Nano Lett., Vol. 6, No. 12, 2006

the LB film has a very weak shoulder at 500 nm and a sharp peak at 540 nm, consistent with no dye aggregation. The peak maximum is also blue-shifted 10 nm with respect to the polmer/R6G on Au bottom electrodes (Figure 3c), though it is consistent with the relatively broad peak on Al bottom electrodes. The peak maximum imaginary index increased to k ) 0.45, as expected for the higher concentration of molecules in the monolayer compared to the dispersion in PMMA, and again a clear Kromers-Kronig relationship is observed between the real and imaginary indices. Figure 4f shows the SPR derived absorption profile from a 15 nm Al-R6G-DMPA 18 nm Au MIM device. Again, the absorption maximum was located at 540 nm, and the molecular k ) 0.43 was equivalent to the MI device, even though the fits were not constrained in any manner. The signal-to-noise ratio decreased from 10:1 to 5:1 between the MI and MIM devices, yet the peak shapes are still clear. Interestingly, comparing the absorption spectrum of the MIM monolayer to the monolayer without the top contact shows a broadening of the absorption peak width. This may be an effect of a Stark effect resulting from the asymmetric metal contacts. Assuming a ∼1 V potential due to the work function difference between gold and aluminum across a 6 nm gap (3 nm of Al2O3 and 3 nm of the organic), an electricdipole Stark shift of ∼2 nm is expected from DFT calculations using Gaussian 03 Software.27 The peak width broadening observed here is ∼5 nm fwhm, which is larger than our calculations, but this may be attributed to additional peak broadening due to electronic hybridization between the molecule and the metal contacts. A significant obstacle for making simultaneous optical and electrical measurements on monolayer-thick films is the 2801

Figure 4. (a) Cartoon drawing of an LB monolayer of R6G and DMPA between top and bottom electrodes. (b) SPR reflection curves for the film in part a between 15 nm Al and 18 nm of gold where the top Au electrode was deposited gently via the PALO process. (c) The real and (d) imaginary part of the refractive index derived from a 3-nm-thick R6G/DMPA LB film on an Al electrode. (e) I-V trace for the Al/R6G/DMPA/Au device described in part b. (f) The imaginary component of the refractive index of the LB film inside the same device. The plot from part d is overlaid (dashed red) for comparison.

propensity for electrical shorts or molecular damage upon deposition of the top contact.28,29 This is particularly true for large-area devices, such as the 3 × 3 mm junctions reported here. To avoid this problem, we have recently designed a scheme for soft top metal deposition using a liftoff, float-on assembly utilizing a hydrophobic polymer backing.24 This method eliminates damage due to direct metal evaporation on the molecules, and produces the molecularly flat, large-area surfaces required for spectroscopic investigation. Figure 4e shows the current-voltage behavior across the R6G/DMPA monolayer MIM device measured optically in Figure 4f, an area of 9 mm2. The I-V trace shows nonshorting behavior with a current density of 2.3e-8 A/cm2, indicating electron tunneling. Fitting the I-V data to a Simmons tunneling model assuming a 3 nm Al2O3 layer gives a tunneling barrier height of 0.8 eV and a β value of 0.71 Å-1, quite consistent with β ) 0.7-0.8 Å-1 measured for insulating alkane molecules.30,31 The ability to monitor dynamic changes in optical absorption of organic layers inside metal junctions is a significant advantage of performing an in situ diagnosis. Here we demonstrate that SPRS can follow dynamic processes within the molecular junction by monitoring dynamic switching of spiropyran, a photochromic molecule that undergoes a ring2802

Figure 5. (a) Ring opening (closing) reaction of the spiropyran molecule when irradiated by UV light (visible light or heat). (b) Filled circles: changes in the SPR reflection curve for 600 nm light from an aluminum/spiropyran/pmma film/gold junction at the resonance angle after sequential exposure to UV light (blue arrow) or heat (red arrow). Open circles: SPR response from a control sample without spiropyran molecules. (c) The real and (d) imaginary components of the refractive index values for the dye-doped polymer layer before (triangle) and after (circle) UV illumination measured on an enclosed Al(15 nm)/spiropyran/PMMA film (30 nm)/Au (15 nm) junction. Solid line and dashed line show UVvis absorption spectra taken for spiropyran in the solution phase before and after UV irradiation.

opening reaction. The spiropyran is thermodynamically stable in its transparent, ring-closed state, Figure 5a, but upon exposure to UV light the ring opens and the molecule becomes strongly absorbing. Heat or visible light will convert the ring-open state back to the original transparent ring-closed conformation.32,33 SPRS measurement of an Al-spiropyran film-Au MIM with 30 wt % spiropyran dispersed in PMMA film shows a reversible change in optical absorption as a function of external UV light or heat activation. Figure 5d shows a significant increase in the k values after application of UV light (blue curve). The solution-phase absorption is shown in black for comparison, again showing that the essential features can be detected within an MIM junction. The process is followed dynamically by monitoring the reflectivity at 600 nm at an angle of 39° at 30 s intervals for five scans followed by either a 3 min UV exposure or 2 min exposure to heat (70 °C). The change in reflectivity at the absorption maximum, Figure 5b, shows that k changes as a function of time. Measurements at a fixed wavelength are more rapid than a full angular scan and only limited by the speed and sensitivity of the detector. Because the full absorption spectrum and angular reflection curve of the dye was measured prior to the reflectivity measurements, there is no ambiguity as to the origin of the changes that occurred inside this device. The behavior of the spiropyran molecule within the MIM junction could be followed during repeated cycling of UV/heat exposure and is a promising method for monitoring other dynamic processes inside MIM devices as well. A control experiment with no spiropyran molecules in the polymer showed no changes under identical conditions. Nano Lett., Vol. 6, No. 12, 2006

In conclusion, SPRS is capable of measuring molecular optical properties inside a MIM junction for nanometer-thick films and even molecular monolayers. Although the top metal contact occludes normal transmissive spectroscopy, it is possible to utilize it to couple a localized SPP that can sensitively probe the optical properties of dielectric material sandwiched between the top and bottom contacts. Absorption spectra of polymer-dispersed R6G molecules within MIM junctions were very similar to solution-based spectra, whereas LB monolayers show a distinct spectral change because of the lack of molecular aggregation. The SPRS results were quite robust and gave consistent molar absorptivities that were nearly the same as in solution. Electrical tunneling measurements could also be made simultaneously on the same devices without shorting using a soft-deposition technique for the top electrode. Finally, dynamic molecular optical properties within MIM junctions were measured by switching the configuration and optical absorption of spiropyran photochromic dyes. These demonstrations open the possibility of measuring a wide array of metal-molecule devices under voltage bias, and correlating their electrical and optical behaviors. This technique thus may have significant impact in the field of organic and molecular electronics. Acknowledgment. We thank Prof. T. Willey for assistance in the NEXAFS data acquisition, Jason Komadina for Gaussian assistance, and Prof. R. Zia for thoughtful discussions. This work was supported by the NSF career award DMR 0449385 and DOE project DE-AC02-76SF00515. Supporting Information Available: Calculation of refractive index values. This material is available free of charge via the Internet at http://pubs.acs.org. References (1) Halik, M.; Klauk, H.; Zschieschang, U. et al. AdV. Mater. 2003, 15, 917. (2) Spanggaard H.; Krebs, F. C. Sol. Energy Mater. Sol. Cells 2004, 83, 125. (3) Friend, R. H.; Gymer, R. W.; Holmes, A. B. et al. Nature 1999, 397, 121. (4) Pease, A. R.; Jeppesen, J. O.; Stoddart, J. F. et al. Acc. Chem. Res. 2001, 34, 433. (5) Joachim, C.; Gimzewski, J. K.; Aviram, A. Nature 2000, 408, 541. (6) Luo, Y.; Collier, C. P.; Jeppesen, J. O. et al. Chemphyschem 2002, 3, 519. (7) Cai, L. T.; Cabassi, M. A.; Yoon, H.; Cabarcos, O. M.; McGuiness, C. L.; Flatt, A. K.; Allara, D. L.; Tour, J. M.; Mayer, T. S. Nano Lett. 2005, 5, 2365. (8) Collier, C. P.; Jeppesen, J. O.; Luo, Y.; Perkins, J.; Wong, E. W.; Heath, J. R.; Stoddart, J. F. J. Am. Chem. Soc. 2001, 123, 12632. (9) Stewart, D. R.; Ohlberg, D. A. A.; Beck, P. A.; Chen, Y.; Williams, R. S.; Jeppesen, J. O.; Nielsen, K. A.; Stoddart, J. F. Nano Lett. 2004, 4, 133.

Nano Lett., Vol. 6, No. 12, 2006

(10) Lau, C. N.; Stewart, D. R.; Williams, R. S.; Bockrath, M. Nano Lett. 2004, 4, 569. (11) Ramachandran, G. K.; Hopson, T. J.; Rawlett, A. M. et al. Science 2003, 300, 1413. (12) McCreery, R.; Dieringer, J.; Solak, A. O.; Snyder, B.; Nowak, A. M.; McGovern, W. R.; DuVall, S. J. Am. Chem. Soc. 2003, 125, 10748. (13) DeIonno, E.; Tseng, H. R.; Harvey, D. D.; Stoddart, J. F.; Heath, J. R. J. Phys. Chem. B 2006, 110, 7609. (14) Barnes, W. L.: Dereux, A.; Ebbesen, T. W. Nature 2003, 424, 824. (15) Zia, R.; Schuller, J.; Brongersma, M. L. Mater. Today 2006, 9, 20 (16) Raether, H. Surface Plasmons on Smooth and Rough Surfaces and on Gratings. Springer Tracts in Modern Physics; Springer-Verlag: New York, 1988; Vol. 111. (17) Pockrand, I.; Swalen, J. D.; Gordon, J. G. et al. Surf. Sci. 1978, 74, 237. (18) Knoll, W. Annu. ReV. Phys. Chem. 1998, 49, 569. (19) Hansen, W. N. J. Opt. Soc. Am. 1968, 58, 380. (20) Handbook of Optical Constants of Solids; Palik, E. D, Ed.; Academic Press Inc: Orlando, FL, 1985. (21) Pockrand, I.; Swalen, J. D.; Santo, R. et al. J. Chem. Phys. 1978, 69, 4001. (22) Pockrand, I.; Swalen, J. D.; Gordon, J. G. et al. J. Chem. Phys. 1979, 70, 3401. (23) Economou, E. N. Phys. ReV. 1969, 182, 539. (24) Shimizu, K. T.; Fabbri, J. D.; Jelincic, J. J. et al. AdV. Mater. 2006, 18, 1499. (25) Arbeloa, F. L.; Gonzalez, I. L.; Ojeda, P. R. et al. J. Chem. Soc., Faraday Trans. 2 1982, 78, 989. (26) Unpublished work. (27) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb, M. A.; Cheeseman, J. R.; Montgomery, J. A., Jr.; Vreven, T.; Kudin, K. N.; Burant, J. C.; Millam, J. M.; Iyengar, S. S.; Tomasi, J.; Barone, V.; Mennucci, B.; Cossi, M.; Scalmani, G.; Rega, N.; Petersson, G. A.; Nakatsuji, H.; Hada, M.; Ehara, M.; Toyota, K.; Fukuda, R.; Hasegawa, J.; Ishida, M.; Nakajima, T.; Honda, Y.; Kitao, O.; Nakai, H.; Klene, M.; Li, X.; Knox, J. E.; Hratchian, H. P.; Cross, J. B.; Bakken, V.; Adamo, C.; Jaramillo, J.; Gomperts, R.; Stratmann, R. E.; Yazyev, O.; Austin, A. J.; Cammi, R.; Pomelli, C.; Ochterski, J. W.; Ayala, P. Y.; Morokuma, K.; Voth, G. A.; Salvador, P.; Dannenberg, J. J.; Zakrzewski, V. G.; Dapprich, S.; Daniels, A. D.; Strain, M. C.; Farkas, O.; Malick, D. K.; Rabuck, A. D.; Raghavachari, K.; Foresman, J. B.; Ortiz, J. V.; Cui, Q.; Baboul, A. G.; Clifford, S.; Cioslowski, J.; Stefanov, B. B.; Liu, G.; Liashenko, A.; Piskorz, P.; Komaromi, I.; Martin, R. L.; Fox, D. J.; Keith, T.; AlLaham, M. A.; Peng, C. Y.; Nanayakkara, A.; Challacombe, M.; Gill, P. M. W.; Johnson, B.; Chen, W.; Wong, M. W.; Gonzalez, C.; Pople, J. A. Gaussian 03, revision C.02; Gaussian, Inc.: Wallingford, CT, 2004. (28) Allara, D. L.; Dunbar, T. D.; Weiss, P. S. et al. In Molecular Electronics: Science and Technology; Aviram, A., Ratner, M., Eds.; New York Academy of Sciences: New York, 1998; Vol. 852, p 349. (29) Herdt, G. C.; Jung, D. R.; Czanderna, A. W. Prog. Surf. Sci. 1995, 50, 103. (30) Salomon, A.; Cahen, D.; Lindsay, S. et al. AdV. Mater. 2003, 15, 1881. (31) Wang, W. Y.; Lee, T.; Reed, M. A. Physica E (Amsterdam, Neth.) 2003, 19, 117. (32) Krysanov S. A.; Alfimov, M. V. Chem. Phys. Lett. 1982, 91, 77. (33) Richert R.; Bassler, H. Chem. Phys. Lett. 1985, 116, 302

NL061893H

2803