Polymer-Coated Gold Island Films as Localized Plasmon Transducers

14530

J. Phys. Chem. B 2008, 112, 14530–14538

Polymer-Coated Gold Island Films as Localized Plasmon Transducers for Gas Sensing† Tanya Karakouz, Alexander Vaskevich,* and Israel Rubinstein* Department of Materials and Interfaces, Weizamnn Institute of Science, RehoVot 76100, Israel ReceiVed: June 1, 2008; ReVised Manuscript ReceiVed: August 7, 2008

Ultrathin (typically e10-nm thick) gold island films evaporated on transparent substrates show a prominent localized surface plasmon (SP) extinction in the visible-to-NIR range. Changes in the dielectric properties of the contacting medium influence the SP absorption band, providing a scheme for optical sensing based on refractive index change. In the present work, the gas sensing capability of gold island based localized surface plasmon resonance (LSPR) transducers was explored using polymeric coatings as the active interface. LSPR transducers were fabricated by spin-coating of polystyrene (PS) or polystyrene sulfonic acid, sodium salt (PSS) onto 5-nm-thick (nominal thickness) gold island films evaporated on silanized glass and annealed. Detailed characterization of the transducers was carried out using high-resolution scanning electron microscopy, cross-sectional transmission electron microscopy, and in situ atomic force microscopy under controlled atmosphere. The hydrophobic PS film exhibits swelling and significant thickness increase upon exposure to chloroform vapor and little or no change in water vapor, whereas the hydrophilic PSS film shows the opposite behavior when exposed to the same vapors. Polymer swelling upon absorption of vapors of good solvents shows a net effect of lowering the effective refractive index in the vicinity of the gold islands, manifested as a characteristic decrease of the SP band intensity and a blue shift of the band maximum. The response, measured for four different vapors, is fast (∼15 s) and reversible. It is shown that gold island systems coated with polymeric films can be applied to vapor recognition in an array configuration. Introduction There is substantial current interest in noble metal clusters such as colloidal particles or evaporated metal island films, deriving from their unique optical, electronic, and chemical properties, differing from those of the bulk material.1,2 In recent years, metallic nanostructures have been employed in various applications, including optical devices,1,3 near-field scanning optical microscopy,1 surface-enhanced spectroscopies,1,4-6 as well as chemical1,3,6,7 and biological1,3,6,8 sensing. The absorbance of an array of metallic nanoparticles or nanoislands depends on various parameters, in particular on the average particle size, shape, and spacing, as well as on the effective refractive index of the surrounding medium. The latter has been widely exploited for sensing applications.1,9-12 A substantial amount of work has been published on the detection and quantification of chemical and biological molecules that bind to metallic clusters (either directly or via receptor layers), using the induced change in the SP resonance absorption band.1,6-8,13-16 We have extensively studied the morphological and optical properties of gold island films prepared by thermal evaporation on transparent substrates, as well as their potential application as transmission localized surface plasmon resonance (LSPR) transducers for chemical17-20 and biological21,22 sensing. We recently demonstrated stabilization of nanostructured gold film transducers on glass.23,24 Various systems for sensing of volatile organic compounds have been reported, including nanoparticles modified with organic molecules,25-30 polymers,31-37 polymer-conductive nanoparticle composites,38-42 and organic compound thin films.43-45 Polymeric † Part of the “Janos H. Fendler Memorial Issue”. * Corresponding authors. E-mail: [email protected]; [email protected].

coatings, showing rapid, reversible, and reproducible response to different vapors associated with swelling and morphological changes, present convenient gas sensing active layers for LSPR transducers. The variety of available polymers and the possibility to incorporate additional components (molecules, functionalized nanoparticles, etc.) in polymers can provide substantial specificity using an array configuration. In the present work, the possibility of using polymer-coated gold island films on glass as LSPR transducers for vapor recognition is demonstrated. Polystyrene (PS) and polystyrene sulfonic acid, sodium salt (PSS) were selected as model coatings presenting two different media: a nonpolar hydrophobic polymer (PS) and a hydrophilic ionomer (PSS). The probe vapor analytes were chosen as good solvents for PS (chloroform, toluene) and for PSS (methanol, water), showing orthogonal behavior toward the two polymers. The optical response of the two polymercoated transducers is shown to correlate directly with the hydrophobicity of the polymer coating and the gas analyte. In situ atomic force microscopy (AFM) imaging and profilometry indicate that the LSPR response is dominated by the effect of polymer swelling. Experimental Section Chemicals. Chloroform (BioLab), toluene (Frutarom), ethanol (Baker analyzed, J.T. Baker), methanol (anhydrous, Mallinckrodt chemicals), sulfuric acid (95-98%, BioLab), hydrogen peroxide (30%, Frutarom), ammonium hydroxide (Frutarom), 3-aminopropyl trimethoxysilane (APTS) (Aldrich), polystyrene (PS) (average MW 250 000, Acros), polystyrene sulfonic acid, sodium salt (PSS) (completely sulfonated, average MW 70 000, Polysciences), and gold (99.99%, Holland-Moran) were used as received. Water was triply distilled. The inert gas used was household nitrogen (from liquid N2).

10.1021/jp804829t CCC: $40.75  2008 American Chemical Society Published on Web 09/20/2008

Au Island Films as Plasmon Transducers for Gas Sensing

Figure 1. Schematic presentation of polymer-coated gold island film preparation: (1) silanization of the glass substrate, (2) gold evaporation, (3) thermal annealing, and (4) spin coating with a polymer layer. Sizes of the various components are not to scale.

Figure 2. Schematic presentation of the experimental setup: (F) gas mixing flask, (I) injection septum, (M) manometer, (Q) measuring cuvette, (S) spectrophotometer, (V1) vapor valve, (V2) vacuum valve, (V3) air valve, (V4) N2 valve, (H) heating mantle.

Transducer Preparation. Microscope cover-glass slides (Menzel-Glaser, No. 3) were cut to 22 × 9 mm2 and cleaned by immersion in freshly prepared hot piranha solution (1:3 H2O2/ H2SO4) (Caution: Piranha solution reacts Violently with organic materials and should be handled with extreme care) for 1 h followed by rinsing with triply distilled water, immersion in “RCA” solution (1:1:5 H2O2/NH3/H2O) for 1 h at 70 °C followed by rinsing with triply distilled water, and rinsing in methanol three times in an ultrasonic bath (Cole-Parmer 8890). The slides were then immersed in 10% APTS solution in methanol overnight, rinsed with methanol and twice with ethanol in the ultrasonic bath, and dried under a nitrogen stream (step 1 in Figure 1). The silanized slides were mounted in a cryo-HV evaporator (Key High Vacuum) equipped with a Maxtek TM-100 thickness monitor for evaporation of ultrathin gold films. Homogeneous gold deposition was obtained by moderate rotation of the substrate plate. Gold (nominal thickness, 5 nm) was resistively evaporated from a tungsten boat at (1-3) × 10-6 torr and a deposition rate of 0.005-0.006 nm s-1 (determined by measuring the evaporation time of 0.1 nm of gold using a stopwatch) (step 2 in Figure 1). Postdeposition annealing of gold-coated slides was carried out in air at 200 °C for 20 h in an oven (Ney

J. Phys. Chem. B, Vol. 112, No. 46, 2008 14531

Figure 3. Cross-sectional TEM (A, C) and HRSEM (B, D) images of 5-nm (nominal thickness) gold island films on silanized glass. (A, B) As evaporated. (C, D) After annealing 20 h at 200 °C. The lines in A and C are drawn as a guide to the eye.

Vulcan 3-550). The heating rate was 5 °C min-1, and the annealed slides were left to cool in air to room temperature (step 3 in Figure 1). Thin polymeric films were spin-coated on the substrates from various solutions (0.33-4% w/w PS in toluene and 4% w/w PSS in water) using a commercial spin-coater (KW-4A, Chemat Technology). The spin-coating process was carried out in two steps (i.e., cleaning with the respective solvent (toluene or water) and coating with the chosen polymer). A consecutive two-speed procedure was used for the coating: deposition of the casting solution (0.2-0.5 mL) and spin up (500 rpm for 3 s), thinning, and evaporation (2500 rpm for 60 s) (step 4 in Figure 1). UV-Vis Spectroscopy. Transmission spectra were obtained with a Varian CARY 50 spectrophotometer operated at a wavelength resolution of 1 nm and an average acquisition time per point of 0.1 s, using air as baseline. The spectra were measured using a special slide holder ensuring a reproducible position of the sample.18 For gas sensing measurements a special static system was designed that allowed preparation of vapor mixtures with a variable analyte concentration, introduction into the spectrophotometer cell, and admittance of air or inert gas (Figure 2). The system consisted of two main parts: a vapor mixing flask (F) and a measuring cell (Q). The measuring procedure was as follows: The system was pumped down to low vacuum (∼1500 Pa) through the vacuum valve (V2). A measured volume of the liquid analyte (organic solvent or water) was injected into the vapor mixing flask through the injection septum (I) and evaporated using a heating mantle (H). The vapor pressure in the system was measured with a manometer (Wika) (M) at an accuracy of (340 Pa. Preparation of gas mixtures was performed in a similar manner by introducing a known amount of each substance into the mixing flask before evaporation. The measuring cell, isolated from the mixing flask by the vapor valve (V1), was aerated through the air valve (V3), followed by placing a sample substrate (polymer-coated gold island film) in the measuring quartz cuvette (Q). The cell was then pumped

14532 J. Phys. Chem. B, Vol. 112, No. 46, 2008

Karakouz et al.

TABLE 1: Average Morphological Parameters for 5-nm (Nominal Thickness) Gold Island Films Evaporated on Silanized Glassa annealedb

unannealed a

b

c

a

b

c

island dimensions (nm)

30 ( 9

19 ( 4

30 ( 7

24 ( 5

[1 - (b/a)2]0.5 [1 - (c/a)2]0.5 island surface density (µm-2) island area fraction (%)

13 ( 2 (12 ( 1)

22 ( 6 (22 ( 2)

0.8 0.9 1000 57

eccentricity

a

0.6 0.7 700 41

Island dimensions measured using HRSEM (a, b), cross-sectional TEM (c), and AFM (c, values in parentheses) images. b 20 h at 200 °C in

air.

Figure 4. (A) Thickness of polymer films spin-coated from solutions with different concentrations of the polymer, measured by AFM profilometry. Black squares: PS films. Red circles: PSS films. Full symbols: polymer deposited on 5-nm gold evaporated on silanized glass, annealed. Open symbols: polymer deposited on bare glass. (B) Normalized UV-vis extinction spectra of 5-nm gold island films evaporated on silanized glass, annealed, coated with PS films from solutions of different concentrations as in A.

TABLE 2: RMS Roughness Values Obtained from AFM Images of Typical Polymer Films on Bare Glass and on 5-nm Gold Island Film Evaporated on Glass, Annealeda PS concn (% w/w) rms roughness glass air (nm) CHCl3 Au island air film CHCl3

1

PSS

0

0.33

2

4

4

0.2 0.1 5

0.2 0.2 2

0.2 0.2 0.3 0.1 0.2 0.2 0.4 0.3 0.3

0.2 0.2 0.3

5

0.4

0.3 0.3 0.3

0.2

a Measurements were performed in air and in 0.5 Psat CHCl3 environments.

down to low vacuum, followed by careful transfer of analyte vapor from the mixing flask to the measuring cell (valve V1) until a desired analyte pressure (measured as a fraction of the maximal vapor pressure, Psat) was achieved. The mixing flask was then disconnected from the system by closing vapor valve

Figure 5. AFM images of PS films spin-coated from a 1% w/w PS solution in toluene on 5-nm gold island film on silanized glass, annealed, imaged sequentially (A) in air, (B) in water vapor (0.67 Psat), (C) in chloroform vapor (0.67 Psat), and (D) in N2. A typical cross section is shown below each image; note the different z-scales in the cross sections. The polymer thickness measured from the cross sections is 50, 50, 60, and 51 nm for A-D, respectively.

V1, and a UV-vis spectrum was taken. Multiple spectral measurements could be carried out without replacing the substrate transducer. The analyte vapor pressure was increased and decreased in a single run, allowing study at different pressures as well as assessment of the repeatability of the optical response, without sample change. Increase of the analyte pressure was achieved by transfer of additional amount of vapor from the mixing flask to the measuring cell, while connection to the vacuum system allowed controlled decrease of the analyte pressure. In this study, 2-9 experiments were performed with each polymer film in different vapors. The influence of the total pressure on the LSPR spectral response was tested by adding an inert gas (air or N2) to the measurement cell containing a certain amount of analyte vapor. It was found that the presence of an inert gas does not change the optical response, and therefore most spectroscopic experiments were performed without adding an inert gas. The kinetics of transducer response were estimated by fast (of the order of 1 s) filling/evacuation of the measurement cell while continuously monitoring the absorbance at a single wavelength chosen near the SP maximum. Note that the optical response to a stepwise change in analyte pressure was an order of magnitude slower than the time needed

Au Island Films as Plasmon Transducers for Gas Sensing

J. Phys. Chem. B, Vol. 112, No. 46, 2008 14533 transmission electron microscope operating at 120 kV, equipped with a charge-coupled device camera (2kx2k, Gatan Ultrascan 1000). Samples for imaging were embedded in a phenol-based M-Bond 610 epoxy resin (Ted Pella, Inc.) according to a procedure described previously46 with some modifications. The gold island film was glued with several drops of the resin onto a premade 4 × 8 mm2 block of the epoxy (the resin solution was cast in a rubber mold (Structure Probe, Inc.) and cured 3 days at 80 °C). The glass substrate was removed by breaking it off the gold film or by dissolving it in diluted (∼5%) HF solution. The backside of the gold was covered with several drops of the resin solution. The epoxy resin curing was performed at 80 °C for 3 days. The embedded samples were then sectioned into thin (∼50 nm) slices using a diamond knife (Micro Star 45°) and a Leica ULTRACUT UCT Ultramicrotome. The samples were sliced perpendicular to the surface of the sample and placed on carbon/collodion-coated copper grids. Results and Discussion

Figure 6. Change in polymer thickness for PS (A) and PSS (B) coatings under exposure to chloroform and water vapors, measured by in situ AFM. Black full symbols: polymer deposited on 5-nm gold evaporated on silanized glass, annealed. Red open symbols: polymer deposited on bare glass.

for filling/evacuation of the system. All spectral measurements were carried out at a constant temperature of 22 ( 1 °C. Characterization Methods. Atomic Force Microscopy. AFM measurements were carried out in air as well as in situ in chloroform or water vapor environments at room temperature (22 ( 1 °C) using a Molecular Imaging (MI) PicoScan instrument operated in the acoustic AC (AAC) or contact mode. The cantilevers used were NSC12, CSC12, and NSC36 series of ultrasharp silicon (MikroMasch) with a resonant frequency of 70-200 kHz and an average radius of e10 nm. In AFM experiments, a chloroform or water vapor environment with variable concentration of the analyte vapor in an inert gas (N2) was obtained using the dynamic mixing system. A flow of N2 saturated with the tested vapor (by passing through a bubbler with a desired solvent) was mixed with a flow of pure N2 in various proportions (0:1, 1:1, and 2:1). An MI gastight glass environmental chamber was fitted to the AFM head to maintain the sample in a controlled environment. AFM examinations included morphology studies of films at different stages of preparation, as well as evaluation of the thickness. Film thicknesses were obtained by scratching a small “window” in the film using a high scanning force in the contact mode, removing the gold islands and polymer film in the window area. The film thickness was determined as the height difference between the bottom of the window and the mean height of the undamaged film, measured in the AAC mode. Calculation of the rms roughness was done for a scan window of 1 × 1 µm2. High-Resolution Scanning Electron Microscopy. HRSEM images were obtained using a SUPRA 55VP LEO highresolution SEM with a cold field-emission electron source using the in-lens SE detector. The measurements were carried out at an applied voltage of 2 kV and a working distance of 4 mm. Cross-Sectional Transmission Electron Microscopy. Crosssectional TEM imaging was performed with a Philips CM-120

Transducer Preparation and Characterization. The morphology of gold island films on glass was studied by HRSEM, cross-sectional TEM, and AFM. HRSEM imaging shows the lateral island dimensions, island distribution, and projection, while information on the vertical island dimensions and shape is provided by cross-sectional TEM imaging and AFM (the latter not shown here; see, for example, ref 47). HRSEM and cross-sectional TEM images of unannealed and annealed 5-nm (nominal thickness) gold island films are presented in Figure 3. The cross-sectional TEM image of the unannealed sample (ca. 50-nm-thick section) shows more than one row of islands, while that of the annealed film generally shows one row, demonstrating the increase in the island average size and separation upon annealing. The odd-shaped islands in the unannealed film become larger, more rounded, and better defined in shape after annealing. The annealing leads to an increase in the average island size as well as the average separation between islands. The results in Figure 3 show that before annealing the islands generally have a large contact area with the substrate (relative to the island size), attributed to the good adhesion provided by the APTS layer on the glass. The island shape is quite complex, reflecting concomitant accumulation of gold on the substrate surface and coalescence of primary nucleated islands during evaporation. Despite some radiation heating during the evaporation at a low rate (0.005-0.006 nm s-1), island coalescence and reshaping is minor. On the other hand, the prolonged postdeposition annealing at 200 °C promotes extensive island coalescence, the island shape becomes more rounded and the relative contact area with the substrate diminishes substantially. The latter is attributed to the effect of surface tension at the higher temperature, assisted by thermal damage to the APTS adhesion layer. Quantitative measurements of average island size and distribution were carried out manually on HRSEM (lateral dimensions) and cross-sectional TEM (vertical dimension) images of an ensemble of 100 islands. Values of the major (length) and minor (width) lateral dimensions of the islands (denoted axes a and b, respectively) and of the height (denoted axis c) were analyzed statistically. The gold coverage was estimated by multiplying the average island top-view projection (A ) πab/ 4) by the number of islands per unit substrate area. The average height of the gold island films was also measured by AFM as described in the Experimental Section. The combined analyzed data are presented in Table 1.

14534 J. Phys. Chem. B, Vol. 112, No. 46, 2008

Karakouz et al.

Figure 7. UV-vis extinction spectra of (A) 55-nm PS film and (B) 85-nm PSS film on 5-nm gold island films evaporated on silanized glass, annealed, in different vapor environments. Arrows indicate the direction of change of the SP band intensity with increasing analyte pressure.

Figure 8. Difference spectra corresponding to the extinction spectra of the PS- and PSS-coated transducers in Figure 7, obtained by subtracting the spectra for P/Psat ) 0. The dotted lines drawn through the maxima of the difference spectra indicate the PIC.

The increase in island size after annealing is manifested primarily as an increase in the average island height. The eccentricity of the islands, a quantification of the average deviation from circularity, is smaller after annealing (i.e., the particles become more rounded). The fractional surface coverage of annealed gold islands is smaller than that of unannealed films. This reflects the coalescence of adjacent islands upon annealing, increasing the average island volume as well as the average separation between islands. The average height values of gold islands obtained by AFM and by cross-sectional TEM are identical within the standard deviation of the measurements.

Polymer films were deposited by spin-coating on 5-nm (nominal thickness) gold island films evaporated on silanized glass and annealed. The rms roughness and the thickness of the films were measured by AFM (see Experimental Section). The thickness of the polymer film increases approximately linearly with its concentration in the casting solution (Figure 4A). The average thickness of the PS film increases from ca. 20 to 275 nm on both bare and gold island covered glass upon increasing the polymer concentration in the casting solution to 4.0% w/w (Figure 4A). For the PSS films, the average thicknesses are 73 and 87 nm on bare glass and gold island film substrates, respectively, for 4.0% w/w polymer concentration.

Au Island Films as Plasmon Transducers for Gas Sensing

J. Phys. Chem. B, Vol. 112, No. 46, 2008 14535

Figure 9. Change in the SP extinction intensity (PIC) and shift of the wavelength of the SP maximum, for (A) 55-nm PS film and (B) 85-nm PSS film on 5-nm gold island films evaporated on silanized glass, annealed, for various analyte pressures. The PIC is normalized to the maximum intensity of the spectrum in air.

Figure 10. Optical response (extinction intensity) of (A) 55-nm PS film and (B) 85-nm PSS film on 5-nm gold island film evaporated on silanized glass, annealed, to repetitive exposure of the transducer to various vapors (at 0.5 Psat), interrupted by evacuation. The extinction was measured at 580 nm.

The rms roughness (Table 2) provides information on the surface coverage by the polymer film. The thickness of the thinnest film spin-coated from 0.33% w/w PS in toluene is ca. 20 nm and the polymer film generally follows the morphology of the gold islands, while at increasing PS (or PSS) concentrations the roughness is smaller and remains constant (i.e., the polymer film is thick enough to eliminate the island-induced corrugation). This is verified by comparison with films coated similarly on bare glass substrates, showing a thicknessindependent rms roughness similar to that obtained with the thicker films on gold islands (Table 2). Comparison of the thickness (Figure 4A) and rms roughness (Table 2) of substrates with and without polymeric coatings furnishes insight into the organization of the polymer on metal island films. The thickness of the gold island film substrate (for bare gold, this is the average island height) does not changed after PS coating from the most dilute solution (0.33% w/w PS), while the spectroscopic data show a marked red shift of the SP band maximum (Figure 4B), confirming the formation of a PS

Figure 11. Optical response of 55-nm PS film on 5-nm gold island film evaporated on silanized glass, annealed, to increasing and decreasing chloroform vapor pressure. The PIC is normalized to the maximum intensity of the spectrum in air.

film on the surface. Additional evidence is given by the marked decrease in the rms roughness (Table 2) measured in air after PS coating, changing from 5 to 2 nm. Moreover, the rms roughness of this film changes from 2 nm in air to 0.4 nm in 0.5 Psat chloroform vapor and back to 2 nm after changing back to air. The polymer swelling upon exposure to chloroform vapor leads to smoothening of the surface, most prominently observed in the case of the 0.33% w/w PS film where the deposited polymer is accumulated mainly in the voids between the gold islands. For thicker polymer films, the rms roughness decreases to ca. 0.3 nm and remains essentially unchanged upon exposure to analyte vapor. As seen in Figure 4B, coating the gold islands with PS films of various thicknesses results in a red shift of the extinction maximum, as expected for an increase in the effective refractive index of the medium. In all cases, the thickness of the polymer film is higher than the decay length of the SP evanescent field in this system, and therefore all films can be approximated as bulk and are expected to display an SP band with a maximum at the same wavelength. The differences observed in the SP

14536 J. Phys. Chem. B, Vol. 112, No. 46, 2008

Karakouz et al.

Figure 12. (A) UV-vis extinction spectra of PS films of different thicknesses on 5-nm gold island film evaporated on silanized glass, annealed, under chloroform vapor. Arrows indicate the direction of change of the SP band intensity with increasing analyte pressure. (B) Change in the SP extinction intensity (PIC) and shift of the wavelength of the SP maximum extinction corresponding to the films in A, for various chloroform vapor pressures. The PIC is normalized to the maximum intensity of the spectrum in air.

Figure 13. Change in the SP extinction intensity and shift of the wavelength of the SP maximum extinction for a 55-nm PS film on 5-nm gold island film evaporated on silanized glass, annealed, and exposed to different vapor mixtures (indicated).

maximum for the different polymer thicknesses are attributed to (i) certain variations of the refractive index of PS with thickness at lower thicknesses, as discussed by Hu et al.,48 and (ii) the effect of interference at higher thicknesses, as shown below. In Situ AFM Measurements. The thickness and rms roughness of PS and PSS films on gold island film and on bare glass substrates were measured during exposure to chloroform or water vapor by in situ AFM, as described in the Experimental Section. The height of the mechanically formed window in the

polymeric film was measured by cross section analysis of the hole in an inert atmosphere, under analyte vapor, and back in an inert atmosphere. An example of such a measurement sequence for a PS film is presented in Figure 5. The thickness of the hydrophobic PS film remains unchanged in water vapor but increases in chloroform vapor because of polymer swelling. On the other hand, the hydrophilic PSS film swells under water vapor and shows little or no response to chloroform vapor. The change in polymer thickness is reversible (i.e., after re-exposure to inert atmosphere the original film thickness is recovered; Figures 5 and 6). Figure 6 summarizes the thickness variations in PS (Figure 6A) and PSS (Figure 6B) films during exposure to different environments. In both cases, vapors of good solvents induce polymer film swelling by up to ∼20% of the initial thickness. PS films swell under chloroform vapor (0.67 Psat) by up to 10 nm (18%), whereas water vapor causes no change in thickness. PSS shows the opposite behavior, namely, water vapor (0.67 Psat) induces expansion of the film by up to 20 nm (23%), whereas chloroform causes no thickness change or even some shrinking of the hydrophilic polymer in the hydrophobic vapor. In both cases, the change is fully reversible and the thickness returns to its original value after exposure to inert gas. Variations in the thickness of the polymer films coated on bare glass substrates are generally similar to those on the gold island films, except in the case of PS film on glass under chloroform vapor, where the hole in the polymer film partially closes and the thickness measurements are not reliable. LSPR Spectroscopy in Analyte Vapors. The optical response of polymer-coated (PS and PSS) gold island films to analyte vapors (i.e., chloroform, toluene, water and methanol) was tested. Transducers were prepared according to the procedure described in the Experimental Section. Transmission

Au Island Films as Plasmon Transducers for Gas Sensing UV-vis spectra were measured in situ during exposure to various relative pressures of analyte vapors in the apparatus shown in Figure 2. The results in Figure 7 reveal a distinctly different optical response of PS- and PSS-coated transducers. Chloroform and toluene induce a marked decrease in the SP extinction and a blue shift of the extinction maximum of PScoated transducers, while methanol and water vapors have no influence on the spectrum (Figure 7A). PSS-coated transducers show the opposite response to the same analyte vapors. A blue shift and a decrease of the SP extinction is a characteristic response to a decrease in the refractive index of the contacting medium, attributed in the present case to swelling of the polymer films upon exposure to vapors of good solvents. The difference spectra presented in Figure 8, derived from the data in Figure 7 by subtracting the spectra in air, emphasize the response to the different vapors, showing a maximum (previously denoted by us as the plasmon intensity change, PIC19,20,49) shifted to longer wavelengths compared to the position of the LSPR band maxima, in agreement with our previous results. The combined results, presented as wavelength shift and normalized PIC vs analyte vapor pressure, are summarized in Figure 9. The PS film is most influenced by chloroform vapor (ca. 20% decrease in the relative SP intensity and a wavelength shift of ∼10 nm under 0.8 Psat), while the response of the PSS film is maximal under water vapor (ca. 24% decrease in the relative SP intensity and a wavelength shift of ∼15 nm under 0.8 Psat). The optical response of LSPR transducers to analyte binding decreases exponentially with distance from the metal island surface, showing typical decay lengths of 10-30 nm.16,22,50,51 The thicknesses of polymer films used in this work (>50 nm) exceed considerably the decay length, and hence the sensor response can be evaluated using the refractive index sensitivity (measured with bulk liquids52) and assuming an isotropic refractive index of the polymer film at all stages. The refractive index sensitivity of the specific island system used (5-nm gold island film on silanized glass, annealed 20 h at 200 °C) measured per refractive index unit (RIU) is ca. 80 nm/RIU (SP wavelength shift) and 0.64 auPIC/RIU (intensity change).52 On the basis of these values, for the 55-nm PS film under 0.8 Psat chloroform vapor, 10-nm SP wavelength shift (Figure 9A) and 0.09 auPIC change (Figure 8A) translate to a refractive index decrease of 0.13 and 0.14 RIU, respectively. Similarly, for the 85-nm PSS film under 0.8 Psat water vapor, 15-nm SP wavelength shift and 0.10 auPIC change translate to a refractive index decrease of 0.19 and 0.16 RIU, respectively. The spread in the values may be because the refractive index sensitivity is determined using a homogeneous contacting phase (bulk liquids), whereas the polymer and analyte coating introduces a complex phase with inhomogeneous properties. The lower limit of detection for gaseous analytes under the present (not optimized) experimental conditions is ca. 0.05 Psat. The detection limit was determined as the vapor pressure for which the optical response under exposure to analyte vapor was