Article pubs.acs.org/JPCC
Analysis of the Gold/Polymer Electrolyte Membrane Interface by Polarization-Modulated ATR-FTIR Spectroscopy Keiji Kunimatsu,*,† Kenji Miyatake,†,‡ Shigehito Deki,† Hiroyuki Uchida,†,‡ and Masahiro Watanabe*,† †
Fuel Cell Nanomaterials Center, and ‡Clean Energy Research Center, University of Yamanashi, 4 Takeda, Kofu 400-8510, Japan S Supporting Information *
ABSTRACT: We developed a new FTIR system with two polarizers in its optics in order to conduct polarization-modulated measurements. Polarization characteristics were examined for the Kretschmann polarizationmodulated attenuated total reflectance (ATR) configuration by the use of gold-sputtered films of 10−100 nm thickness on Ge and ZnSe prisms. The marked increase of the polarization characteristics for Au film thicknesses below 30−40 nm is closely associated with a large reflectivity decrease of the p-polarized radiation. A cast film of sulfonated block poly(arylene ether sulfone ketone) membrane was formed on the Au film, and the interfacial spectra were acquired by the use of the ATR FTIR system. The interfacial spectra resemble those of the ATR spectra of the bulk membrane but exhibited strong dependence of the intensity and line shape of the vibrational modes on the Au thickness. The dependence is closely associated with a change of the polarization characteristics of the interface. Electromagnetic as well as chemical effects were concluded to be responsible for the band anomalies and enhancement.
1. INTRODUCTION Electrochemical reactions in polymer electrolyte fuel cells (PEFCs) take place at three-phase boundaries, which consist of the anode or cathode catalyst, polymer electrolyte, and reactant gases. Improving the fuel cell performance requires fundamental knowledge of the interactions between the catalyst and polymer electrolyte. For PEFCs using proton exchange membranes (PEMs) the interaction includes specific adsorption of sulfonic acid groups contained in the pendant side chains of perfluorosulfonic acid (PFSA) membranes such as Nafion or in the hydrophilic units of aromatic (nonfluorinated) hydrocarbon polymer membranes. The hydrophobic unit of the latter type membrane could also be adsorbed on the anode or cathode catalyst, which may lead to a lowering of the fuel cell performance. The catalyst/membrane interaction and electrochemical reactions at the interface can be investigated by an in situ vibrational spectroscopic method by the use of a fuel cell-type spectroelectrochemical cell that allows Kretschmann mode ATR measurements to be carried out in the absence of an electrolyte solution.1,2 The oxygen reduction reaction (ORR) was investigated at the Pt/Nafion interface1 and the interface between carbon-supported Pt and Pt3Co nanoparticle catalysts and Nafion membrane 2 by applying the subtractively normalized interfacial FTIR spectroscopy (SNIFTIRS) scheme.3 In the latter, the background potential was set to 0.1 V vs the reversible hydrogen electrode (RHE) to probe the changes in the interfacial spectra at higher potentials. We found an intensity increase of the νs(SO3) band of the sulfonic acid groups, as well as an intensity increase of the ν(OCO) band of © 2015 American Chemical Society
the ether groups, which were accompanied by intensity decreases of the ν(OH) and δ(HOH) bands of water in the membrane. These results were interpreted in terms of the adsorption of sulfonic acid groups at higher potentials, which led to a depletion of hydrated protons and water from the catalyst−membrane interface at higher potentials. The SNIFTIRS study mentioned above suggested a need to acquire absolute spectra at constant potentials to conduct a more quantitative analysis of the interaction between the catalyst and the membrane. However, acquisition of the absolute spectra was not possible, because it was not feasible to prepare a membrane-free catalyst surface in situ to serve as a spectral background. Double-modulation FTIR combined with a photoelastic modulator and a fixed polarizer could be useful for acquiring such absolute spectra at electrochemical interfaces, but this method has been developed mostly for the external infrared reflection absorption (IRAS) mode.4 Hatta et al. developed a polarization-modulated method applicable to IRAS as well as the Kretschmann ATR configuration by the use of a grating spectrometer.5 When applying the Kretschmann ATR configuration in FTIR or grating spectrometers combined with the polarization modulation technique, however, it is important to discuss the polarization characteristics associated with the ATR mode, as the latter gives rise to a spectral background of the polarization-modulated spectra. Received: May 13, 2015 Revised: June 30, 2015 Published: July 1, 2015 16754
DOI: 10.1021/acs.jpcc.5b04622 J. Phys. Chem. C 2015, 119, 16754−16761
Article
The Journal of Physical Chemistry C
angle of polarizer #2 upon a triggering signal from the FTIR. Switching of the angle can be completed within 0.2 s. We developed two options to control polarizer #2, which are associated with two polarization modulation measurements: one is to control the state of polarization of the infrared radiation by the triggering signal supplied to SP02 under the SNIFTIRS scheme. Usually SNIFTIRS is used to acquire the spectral difference between two potentials, but it can be applied to acquiring a spectral difference between the two polarization states, i.e., p- and s-polarized states, by use of the SP02 interface. Under the SNIFTIRS scheme, we can define the number of interferometer scans n for p- and s-polarized states and the total number of scan sets m. For this scheme, we have
The purpose of the present report is 2-fold: first, to develop a new polarization modulation scheme combined with an FTIR spectrometer applicable to the Kretschmann ATR mode by the use of two infrared polarizers, examining the polarization characteristics of the optical configuration, and, second, to investigate the gold/aromatic (nonfluorinated) hydrocarbon polymer membrane interface. For the first purpose, we examined the reflection characteristics of p- and s-polarized incident infrared radiation and evaluated the polarization characteristics defined by −log(Ip/Is), which would serve as a background for the polarization-modulated spectra for the Kretschmann ATR mode, where Ip and Is are the single-beam intensities of the reflected p- and s-polarized infrared radiation, respectively. The aromatic (nonfluorinated) hydrocarbon polymer membranes have been developed as lower cost, more environmentally compatible alternatives to the perfluorosulfonic acid (PFSA) membranes such as Nafion.6,7 We synthesized a series of poly(arylene ether sulfone ketone) multiblock copolymers having highly sulfonated hydrophilic blocks sequenced with hydrophobic groups with a sulfone−ketone structure8−10 and reported that poly(arylene ether sulfone)s containing sulfofluorenyl groups, SPE-bl-1 membranes, are highly proton conductive and durable for 5000 h in operating fuel cells.9 Our final goal is to investigate the interface between the SPE-bl-1 ionomer and the fuel cell catalyst by the polarization-modulated technique, but as a first step, we started with the investigation of the interface between the membrane and a sputtered Au film.
Ip = (∑ Ip , m)/(n × m)
(1)
m
Is = (∑ Is , m)/(n × m)
(2)
m
where Ip,m and Is,m are the sums of the p- and s-polarized light intensities for n scans at the mth scan set, whereas Ip and Is are the overall averages of the p- and s-polarized light intensities for n × m scans. We have polarization characteristic = − log(Ip/Is) = − log(p /s) (3)
The polarization characteristic will be simply denoted by −log(p/s) in this report. The polarization modulation measurement and subsequent data analysis were carried out under the SNIFTIRS scheme. We found alternatively, however, that −log(Ip/Is) can be obtained simply by measuring Ip and Is at p and s polarization, respectively, by controlling polarizer #2 manually. The quality of the spectra was as good as those obtained under the SNIFTIRS scheme, owing to the high stability and resolution of the FTIR spectrometer. The polarization modulation scheme mentioned above allows us to conduct a steady-state measurement at a given potential. However, we needed another scheme to conduct time-resolved measurements during potential cycling or hydration/dehydration of the fuel cell membranes. We found that such measurements can be successfully carried out by the use of the kinetic mode operation of the FTIR spectrometer. For this purpose, a single-beam spectrum was acquired at the s-polarized state at the beginning of the kinetic mode operation to obtain s(t = 0), which served as the background in the subsequent time-resolved measurements with p-polarized radiation. We have
2. EXPERIMENTAL SECTION 2.1. New Polarization Modulation Schemes Combined with an FTIR Spectrometer. We converted conventional 60° reflection optics with Al mirrors to ones suitable for the polarization-modulated measurements. First, we sputtered a gold film of ca. 200 nm thickness on all of the mirrors to mask the strong infrared absorption around 1210 cm−1 from the protective silica film usually attached to commercial Al mirrors. Second, we introduced two infrared polarizers, #1 and #2, in front of the sample surface, as shown in Figure 1. The angle of
−log(p/s)(t ) = −log(p(t )/s(t = 0)
(4)
2.2. Sample Preparation for PM-SNIFTIRS by the Kretschmann ATR Mode. We prepared ZnSe and Ge semicylindrical prisms of 2 × 2.5 cm with sputtered Au films of various thicknesses ranging from ca. 10 to 100 nm. The effect of the film thickness on the polarization characteristics associated with the Kretschmann ATR mode was investigated by the use of these samples. The prisms were polished successively with 1, 0.3, and 0.05 μm alumina suspensions and subjected to cleaning in an ultrasonic bath containing Milli-Q water and acetone. Au films were then prepared with a magnetron sputtering ion coater VD MSP-1S (Vacuum Devices, Japan) under a low-pressure air atmosphere at room
Figure 1. Optics for polarization modulation FTIR measurements with two infrared polarizers.
polarizer #1 was fixed at 45° with respect to the plane of incidence, so that p- and s-polarized light beams of equal intensity are produced by the second polarizer #2 by controlling its angle between 0° and 90° under the control of the FTIR spectrometer. SP01 is a unit to control the polarizer #1 angle to 45 ± δ°, whereas SP02 is an interface to control the 16755
DOI: 10.1021/acs.jpcc.5b04622 J. Phys. Chem. C 2015, 119, 16754−16761
Article
The Journal of Physical Chemistry C
Figure 2. Polarization characteristics of the Kretschmann ATR mode with (A) ZnSe and (B) Ge prisms as a function of Au thickness dAu.
Figure 3. (A) Dependence of polarization characteristics and (B) reflectivity of s- and p-polarized radiations on Au thickness for the Kretschmann ATR configuration with Ge and ZnSe prisms.
characteristics on ZnSe were nearly independent of wavenumber over a wide range; however, we obtained a strong, narrow band at 908 cm−1 and a broad band at 3450 cm−1 on Ge. The former can be assigned to the oxide on the Ge surface and the latter to the ν(OH) band of water associated with the oxide.11 Because these bands appeared after an Au film was sputtered on Ge, they can be interpreted as being due to a surface enhancement (SEIRA) effect12 by the Au film. The bands diminished for thicker Au films, which is consistent with the well-known characteristics of the SEIRA effect. The presence of the strong oxide band at 908 cm−1 obscures the spectral region around 1000 cm−1, which makes Ge unsuitable for use as an ATR prism to monitor spectra of hydrocarbonbased fuel cell membranes, which have important vibrational bands in this spectral region.13 The polarization characteristic at 2000 cm−1 is plotted as a function of Au film thickness in Figure 3A. We can see that the polarization characteristic exhibits a large increase below ca. 50 nm for both prisms. It should be noted, however, that thicknesses below 10 nm were not examined due to the limited sensitivity associated with the film thickness determination. We then examined the origin of the polarization characteristic presented in Figure 3A by investigating the individual reflectivity changes of the p- and s-polarized radiation upon total reflection at the ZnSe/Au/air and Ge/Au/air boundaries. For this purpose, single-beam intensities of the p- and spolarized radiation in the absence, I, and presence, IAu, of an Au film were measured. Figure 3B shows the reflectivity, IAu/I, of the p- and s-polarized radiation on ZnSe and Ge as a function of Au thickness. We note that the reflectivity of the p-polarized radiation dropped sharply whereas the s-polarized radiation decreased only slightly with decreasing Au thickness below ca. 50 nm. It is clearly demonstrated that the origin of the marked increase of the polarization characteristic with decreasing Au thickness shown in Figure 3A is the much larger reflectivity
temperature. The typical sputtering rate was ca. 1 nm/s. The nominal film thickness was estimated from the weight increase of the prism after sputtering. The weight was monitored with a laboratory balance with an accuracy of 0.1 mg, which is equivalent to a film thickness of ca. 10 nm for a prism of 2 × 2.5 cm size. Therefore, the accuracy of the quoted thickness is 10 nm in the present report. A cast film of sulfonated block poly(arylene ether sulfone ketone) membrane, to be called SPE-bl-1, was formed on the Au films sputtered on the ZnSe prisms. The thickness of the gold films was varied between 10 and 100 nm also in the presence of the cast membrane to examine the effect of film thickness on the polarizationmodulated spectra of the cast SPE-bl-1. Measured amounts of SPE-bl-1 solution in DMAc were delivered uniformly over the Au surface on ZnSe by the use of a glass syringe. The SPE-bl-1 solution was dried at 50 °C for 5 h and then at 80 °C in vacuum overnight. A cast film of 1 μm thickness was finally formed on the Au film. The cell was equipped with a gas inlet and outlet to supply humidified or dry N2 gas to humidify or dry the cast SPE-bl-1 film through the bulk SPE-bl-1 membrane. The ZnSe/Au/SPE-bl-1 film sample was mounted in a homemade IR reflection cell13 with an angle of incidence of 60°. A bulk film of 33 μm thick SPE-bl-1 membrane was placed below the sample so that the cast film and the bulk membrane were firmly in contact with each other.
3. RESULTS AND DISCUSSION 3.1. Polarization Characteristics at ZnSe, Ge/Air Interfaces in the Presence of Au-Sputtered Films. We present changes of the −log(p/s) with thickness of the sputtered Au film observed on ZnSe and Ge prisms in Figure 2A and 2B, respectively. The −log(p/s) values were close to zero in the absence of Au films for both prisms. The polarization characteristics were highest for the 10 nm thickness and decreased for larger thicknesses. The polarization 16756
DOI: 10.1021/acs.jpcc.5b04622 J. Phys. Chem. C 2015, 119, 16754−16761
Article
The Journal of Physical Chemistry C
Figure 4. Change of reflectivity of p-polarized radiation with Au film thickness for (A) ZnSe and (B) Ge prisms.
decrease of the p-polarized radiation compared to that of the spolarized radiation shown in Figure 3B. The reflectivity data for the p-polarized radiation in this figure are those obtained at 2000 cm−1, but similar results were observed throughout the spectral region investigated, as the reflectivity decreased essentially in parallel with decreasing Au film thickness. The results are shown in Figure 4 for the ZnSe and Ge prisms. The spectral regions shown extended to 7000 and 5000 cm−1 for ZnSe and Ge, respectively, due to the optical cutoff characteristics of these materials at higher wavenumbers. The optical properties of thin films of metals such as gold and silver have been studied extensively due to their specific nature and applications.12,14−18 It has been known that gold thin films exhibit an optical transmittance minimum at ca. 600 nm, which undergoes a red shift and broadening to finally diminish with increasing film thickness. This change is associated with a morphological transition from a discontinuous island structure to a quasi-continuous film structure around a nominal thickness of 10−15 nm.17,18 The minimum is explained by an excitation of localized surface plasmon− polaritons (SPP) by the incident light, and the tail of the absorption extends well into the mid-infrared region12,19 due to the dipole coupling between the islands.20 The absorption suggests that the collective electron resonance is excited even in the mid-infrared region to give rise to an enhanced local electromagnetic field.12 The local electromagnetic excitation is essentially polarized along the surface normal on the gold islands.21 The reflectivities of the s- and p-polarized radiation of the Kretschmann ATR configuration were calculated by Osawa et al.22 and compared with experiments by Suzuki et al.23 at 1800 cm−1 for a 20 nm thick Au film deposited on Ge. The calculation was based on the Fresnel formula combined with the effective medium approximation (EMA) model for the inhomogeneous structure of the deposit, and fair agreement between the calculated and measured reflectivities was reported. The detailed dependence of the reflectivity upon Au thickness was not presented,23 but our results in Figure 3B are in qualitative agreement with their results. They showed also that the lower reflectivity of p-polarized radiation is associated with the higher infrared absorption intensity of the molecules adsorbed on the Au deposit. The reflectivity change of the p-polarized radiation with gold thickness, presented in Figures 3B and 4, can be well explained by the optical properties of gold thin films that have been established thus far. 3.2. Polarization-Modulated ATR-FTIR Spectra of the Cast Film of a Fuel Cell Membrane. Figure 5 shows the
Figure 5. Polarization-modulated ATR-FTIR spectra of the cast SPEbl-1 film on a 10 nm thick Au-sputtered film for hydrated and dry states of the film: (A) high- and (B) low-wavenumber regions.
polarization-modulated ATR-FTIR spectra of the cast film on a 10 nm thick Au-sputtered film for hydrated and dry states of the film, observed, respectively, after 2 h hydration with humidified N2 gas and dehydration with dry N2 gas overnight. Development of the ν(OH) band of water in the cast film is obvious upon hydration in Figure 5A, whereas the bands associated with the vibrational modes of the cast membrane superimposed upon the δ(HOH) band around 1600 cm−1 are evident in Figure 5B. Assignment of the bands in Figure 5 was conducted based on that of the bulk membrane13 and are listed in Table1, which compares the bulk and adsorbed bands on the Au (10 nm) film. Although the peak positions were quite close between the two membranes, the bands for the adsorbed membranes were much broader, most likely due to their interaction with the gold film. For the δ(HOH) bands, we obtained two peak positions; Table 1. Band Assignments of SPE-bl-1: Bulk and Adsorbed on Au(10 nm), and Band Shapes of the Vibrational Modes of the Latter
16757
mode
bulk SPE (cm−1)
SPE/Au(10 nm) (cm−1)
band shape
v(OH) δ(HOH) CO CC CC ring vas(COC) vas(SO2) vs(SO2) vas(SO3) vs(SO3)
3430 1710/1637 1720 1656 1585 1488 1240 1150 1105 1205 1030
3380 1680/1622 1750 1650 1579 1475 1230 1140 1100 not clearly observed 1034
bipolar normal normal normal normal normal normal normal normal inverted
DOI: 10.1021/acs.jpcc.5b04622 J. Phys. Chem. C 2015, 119, 16754−16761
Article
The Journal of Physical Chemistry C
Figure 6. Change of the polarization-modulated ATR spectra of the cast SPE-bl-1 membrane on Au(10 nm) referred to its dry state during its hydration. The color scheme for spectra in regions A and B is the same.
Figure 7. Change of the ATR spectra of a bulk SPE-bl-1 membrane during its hydration. Spectra are referred to its dry state, and the color scheme in wavenumber regions A and B is the same.
between the spectra of the bulk and cast membranes: for the cast membrane, first, the ν(OH) band is broader and distorted, with a bipolar shape;24 second, there is no positive-going band around 1200 cm−1 to be assigned to νas(SO3) vibration in Figure 6B; and third, for the νs(SO3) band, we obtained only a downward band at 1034 cm−1, whose direction is opposite to that seen for the bulk membrane. The absence of the νas(SO3) band in the spectra of the cast film can be interpreted in terms of the surface selection rule applied to the νas(SO3) vibration on Au, which has its dipole moment change parallel with respect to the Au surface, whereas the νs(SO3) vibration, which has its dipole moment change perpendicular to the surface, is clearly seen in the polarization-modulated spectra. The reason for the opposite band direction of the νs(SO3) vibration is likely to be associated with the morphology of the nanostructured Au-sputtered film. This will be discussed later in more detail. The comparison of the bulk membrane and the one cast on Au(10 nm) has both the revealed similarity and the specific nature of the latter in terms of the band shape, which are listed in the last column of Table 1. We found that the intensities of these bands depend strongly on the thickness of the Au-sputtered film. This is demonstrated in Figure 8, which shows the marked decreases of the ν(OH) and νs(SO3) band intensities for Au thicknesses above ca. 20 nm. The areas of these bands, A(OH) and A(SO3), at full membrane hydration are plotted in Figure 9. The marked increase of the band area for Au thicknesses smaller than 30−40 nm can be interpreted in terms of the surface enhancement (SEIRA) effect on nanoislands of Au particles.12 Alternatively, a possible change of the penetration depth of the p-polarized radiation with a change of the refractive index of the gold film
the higher and lower wavenumbers denote, respectively, the vibrations of the hydrated protons and bulk-like water in the membrane.13 In Table 1, the νas(SO3) band of the adsorbed membrane was missing, and for the νs(SO3) band position, 1034 cm−1 was obtained, while no upward band is seen at this wavenumber in Figure 4B, where the two bands correspond to the asymmetric and symmetric SO3 vibrations, respectively. Their assignment is supported by a comparison of the spectral difference between the dry and the hydrated states for the cast and bulk membranes, as shown below. First, we present the changes of the polarization-modulated ATR spectra referred to its dry state observed for the cast membrane/Au(10 nm) interface in Figure 6. We compare the spectra with similar ones for the bulk membrane presented in Figure 7, observed with the same infrared reflection cell but without an Au film on ZnSe. The measurement was conducted by use of unpolarized infrared radiation. The development of the ν(OH) band around 3430 cm−1 for water in the membrane is clearly evident during the course of hydration. Below 2000 cm−1, however, development of the δ(HOH) band of water is superimposed on the negative-going bands associated with the band intensity decrease of the phenylene ring vibration modes at 1584 and 1485 cm−1. The decrease of these band intensities during hydration is due to the dilution effect.13 Further below 1400 cm−1, we obtained another negative-going band at 1235 cm−1, assigned to νas(COC), asymmetric COC vibration, and two positive-going bands at 1205 and 1028 cm−1, assigned, respectively, to νas(SO3) and νs(SO3). These bands increased in intensity as the dissociation of the sulfonic acid groups proceeded during membrane hydration. Comparison of Figures 6 and 7 reveals three major differences 16758
DOI: 10.1021/acs.jpcc.5b04622 J. Phys. Chem. C 2015, 119, 16754−16761
Article
The Journal of Physical Chemistry C
common to all of the vibrational modes, with different band shapes and enhancement factors for a given Au film thickness. In order to examine further the origin of the surface enhancement and band anomalies, we compared the dependence of A(OH) and the polarization characteristics, −log(p/s), on Au film thickness by plotting them on the same axis, as shown in Figure 10. We found a very similar dependence of
Figure 8. Changes of the (A) ν(OH) and (B) νs(SO3) bands of the cast SPE-bl-1 membrane with Au thickness dAu.
Figure 10. Dependence of ν(OH) band intensity and polarization characteristics of the SPE-bl-1/Au interface at 2000 cm−1.
A(OH) and −log(p/s). In other words, the dependence of A(OH) on Au thickness is closely associated with the polarization characteristics of the interface. The origin of the increase of the polarization characteristics with decreasing Au film thickness was the reflectivity decrease of the p-polarized radiation, as shown in Figure 3B in the absence of the SPE-bl-1 cast film. We found that this was true also in the presence of the cast film. The absorbed p-polarized radiation produces a strong electric field perpendicular to the Au surface, which would lead to the surface enhancement effect, as observed below 30−40 nm Au thickness. The data in Figure 10 are direct evidence showing the electromagnetic effect in the enhancement.12 However, the different band shapes and enhancement factors for the different vibrational modes suggest that the band anomalies and enhancements are specific to each vibrational mode. The specific interaction between each vibrational mode and the Au surface would give rise to chemical effects in addition to the electromagnetic field enhancement.30 3.3. Morphology of the Au-Sputtered Films. Lastly, we examined the surface morphology of the Au-sputtered films by SEM, EDX, and AFM measurements to investigate its role in the enhancement and band anomalies found at the SPE-bl-1/ Au film interface. For this purpose we prepared two samples with nominal Au thicknesses of 15 and 52 nm on 1 mm thick, 4 × 4 mm ZnSe plates. The cross-sectional TEM images of the Au-sputtered films were observed by cutting the samples with a focused ion beam (FIB). It was found that the surface was already covered by a continuous layer of gold at a film thickness of 15 nm. The roughness of the surface revealed by the topographic AFM image was on the order of less than ca. 5 nm. At a nominal thickness of 52 nm, we can observe the surface to include nanoparticles smaller than ca. 100 nm in diameter distributed randomly on top of a continuous gold layer, as shown in Figure 11. The height of a typical nanoparticle was 27 nm, and the thickness of the layer was 34 nm, as shown by a TEM image of a cross section of the Au film, presented in Figure 12. It is apparent from Figure 12 that the nanoparticles were produced after the underlying continuous gold layer was completed. This suggests that the increased thickness of the
Figure 9. Dependence of ν(OH) and νs(SO3) band area on the Au thickness.
could contribute to such a band intensity change with the film thickness. However, it has been recognized that the presence of a thin metal film such as gold and silver on the ATR crystal in contact with the sample material, characterized by the Kretschmann configuration, reduces the penetration depth to only a few nanometers from the thin metal film surface.12 This is totally different from the usual ATR configuration, for which the penetration depth is on the order of micrometers. In addition, the polarization modulation method adopted in the present report senses the surface region, where the s-polarized infrared light intensity is essentially zero. Therefore, the refractive index change of the Au film with varying thickness would not lead to a change of the penetration depth in the measurements. The above interpretation of our method is supported by the fact that all of the vibrational frequencies and band widths in the observed interfacial spectra are different from those of the bulk membrane measured by the usual ATR method. We estimated the enhancement factor around the 10 nm Au thickness by referring to the band intensity at 73 nm by assuming it corresponds to the bulk Au surface. The enhancement factors thus estimated are 15 (ν(OH)), 7 (ring), and 60 (νs(SO3)). From these results, we can conclude that the band shape and the enhancement factors depend on the vibrational mode of the membrane. In general, the band shape change is accompanied by an intensity enhancement of the vibrational mode of molecules adsorbed on nanostructured metal surfaces. This is a well-known phenomenon for CO adsorbed on surfaces of Pt25−28 and Fe29 films. In that work, the analysis of the band anomalies combined with the surface enhancement has been done based on effective medium theory, EMT, with increased volume fraction of metal particles on the surface and the percolation threshold of the nanostructured surface. In our present results, such geometrical parameters are 16759
DOI: 10.1021/acs.jpcc.5b04622 J. Phys. Chem. C 2015, 119, 16754−16761
Article
The Journal of Physical Chemistry C
continuous layer at a film thickness of 15 nm. However, there was surface roughness on the order of less than ca. 5 nm. It is concluded that this surface is responsible for the observed enhancement and band anomalies.
■
ASSOCIATED CONTENT
S Supporting Information *
TEM/EDX data of the cross-section of the 15 nm thick Au film, and AFM image of its surface. The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.jpcc.5b04622.
■
AUTHOR INFORMATION
Corresponding Authors
*E-mail:
[email protected]. *E-mail:
[email protected]. Notes
The authors declare no competing financial interest.
■
ACKNOWLEDGMENTS We are grateful to the New Energy and Industrial Technology Development Organization (NEDO) of Japan, which supported the Research on Nanotechnology for High-Performance Fuel Cells (HiPer-FC) project, under which this work was conducted. We are also grateful to Dr. M. Hara and Ms. T. Gomyo of the University of Yamanashi for their help in conducting AFM and SEM, TEM, and EDX measurements, respectively.
Figure 11. FE-SEM image of the Au-sputtered film surface (nominal thickness 52 nm) on ZnSe.
■
REFERENCES
(1) Kunimatsu, K.; Yoda, T.; Tryk, D. A.; Uchida, H.; Watanabe, M. In-situ ATR-FTIR Study of Oxygen Reduction at the Pt/Nafion Interface. Phys. Chem. Chem. Phys. 2010, 12, 621−629. (2) Hanawa, H.; Kunimatsu, K.; Watanabe, M.; Uchida, H. In Situ ATR-FTIR Analysis of the Structure of Nafion−Pt/C and Nafion− Pt3Co/C Interfaces in Fuel Cell. J. Phys. Chem. C 2012, 116, 21401− 21406. (3) Davidson, T.; Pons, S.; Bewick, A.; Schmidt, P. Vibrational Spectroscopy of the Electrode/Electrolyte Interface. Use of Fourier Transform Infrared Spectroscopy. J. Electroanal. Chem. Interfacial Electrochem. 1981, 125, 237−241. (4) Golden, W. G. In Fourier Transform Infrared Spectroscopy; Ferraro, J. R., Basile, L. J. , Eds.; Academic Press: New York, 1985; Vol. 4, p 315. (5) Hatta, A.; Wadayama, T.; Suëtaka, W. A Polarization Modulation Infrared Reflection Technique Applied to Study of Thin Films on Metal and Semiconductor Surfaces. Anal. Sci. 1985, 1, 403−408. (6) Kim, Yu S.; Pivovar, B. S. Moving Beyond Mass-Based Parameters for Conductivity Analysis of Sulfonated Polymers. Annu. Rev. Chem. Biomol. Eng. 2010, 1, 123−148. (7) Park, C. H.; Lee, C. H.; Guiver, M. D.; Lee, Y. M. Sulfonated Hydrocarbon Membranes for Medium-Temperature and LowHumidity Proton Exchange Membrane Fuel Cells (PEMFCs). Prog. Polym. Sci. 2011, 36, 1443−1498. (8) Bae, B.; Miyatake, K.; Watanabe, M. Synthesis and Properties of Sulfonated Block Copolymers Having Fluorenyl Groups for Fuel Cell Applications. ACS Appl. Mater. Interfaces 2009, 1, 1279−1286. (9) Bae, B.; Yoda, T.; Miyatake, K.; Uchida, H.; Watanabe, M. Proton-Conductive Aromatic Ionomers Containing Highly Sulfonated Blocks for High-Temperature-Operable Fuel Cells. Angew. Chem., Int. Ed. 2010, 49, 317−320. (10) Bae, B.; Miyatake, K.; Watanabe, M. Sulfonated Poly(arylene ether sulfone ketone) Multiblock Copolymers with Highly Sulfonated Block. Synthesis and Properties. Macromolecules 2010, 43, 2684−2691.
Figure 12. TEM image of a cross-section of the Au film (nominal thickness 52 nm) cut by FIB.
underlying layer was responsible for the reduced enhancement, but the presence of the nanoparticles was not.
4. CONCLUSIONS Polarization-modulated spectra of the sulfonated block poly(arylene ether sulfone ketone) membrane on sputtered Au films were acquired, and the assignment of the vibrational modes was conducted based on a comparison with that of the bulk membrane. Spectral differences between the dry and the hydrated states revealed similarities and major differences between the bulk membrane and that adsorbed on the Au films. The peak frequencies of the major vibrational modes were similar between the bulk and the interfacial spectra, but the latter exhibited strong enhancements and anomalies of the lineshapes of the vibrational modes with decreasing Au thickness. The enhancement is closely associated with the decrease of reflectivity of the p-polarized radiation with decreasing Au thickness. The observed enhancement and band anomalies were specific to each vibrational mode, and electromagnetic as well as chemical effects were concluded to be responsible. The surface of the Au film was found to be a 16760
DOI: 10.1021/acs.jpcc.5b04622 J. Phys. Chem. C 2015, 119, 16754−16761
Article
The Journal of Physical Chemistry C (11) Litvinov, V. V.; Svensson, B. G.; Murin, L. I.; Lindström, J. L.; Markevich, V. P.; Peaker, A. R. Determination of Interstitial Oxygen Concentration in Germanium by Infrared Absorption. J. Appl. Phys. 2006, 100, 033525. (12) Osawa, M. Surface-Enhanced Infrared Absorption. In Near-Field Optics and Surface Plasmon Polaritons; Kawata, S., Ed.; Topics in Applied Physics; Springer: New York, 2001, Vol. 81, pp 163−187. (13) Kunimatsu, K.; Yagi, K.; Bae, B.; Miyatake, K.; Uchida, H.; Watanabe, M. ATR-FTIR Analysis of the State of Water in a Sulfonated Block Poly(arylene ether sulfone ketone) Membrane and Proton Conductivity Measurement During the Hydration/Dehydration Cycle. J. Phys. Chem. C 2013, 117, 3762−3771. (14) Grandqvist, C. G.; Hunderi, O. Optical Properties of Ultrathin Gold Particles. Phys. Rev. B 1977, 16, 3513−3534. (15) Dalacu, D.; Martinu, L. Optical Properties of Discontinuous Gold Films: Finite Size Effect. J. Opt. Soc. Am. B 2001, 18, 85−92. (16) Doron-Mor, I.; Barkay, Z.; Fillip-Granit, N.; Vaskevich, A.; Rubinstein, I. Ultrathin Gold Island Films on Silanized Glass. Morphology and Optical Properties. Chem. Mater. 2004, 16, 3476− 3483. (17) Axelevitch, A.; Apter, B.; Golan, G. Simulation and Experimental Investigation of Optical Transparency in Gold Island Films. Opt. Express 2013, 21, 4126−4138. (18) Siegel, J.; Lyutakov, O.; Rybka, V.; Kolská, Z.; Švorčík, V. Properties of Gold Nanostructures Sputtered on Glass. Nanoscale Res. Lett. 2011, 6, 96−105. (19) Osawa, M.; Ikeda, M. Surface-Enhanced Infrared Absorption of p-Nitrobenzoic Acid Deposited on Silver Island Films: Contribution of Electromagnetic and Chemical Mechanisms. J. Phys. Chem. 1991, 95, 9914−9919. (20) Yoshida, S.; Yamaguchi, T.; Kinbara, A. Optical Properties of Aggregated Silver Films. J. Opt. Soc. Am. 1971, 61, 62−69. (21) Gersten, J. I.; Nitzan, A. Photophysics and Photochemistry Near Surfaces and Small Particles. Surf. Sci. 1985, 158, 165−189. (22) Osawa, M.; Kuramitsu, M.; Hatta, A.; Suëtaka, W.; Seki, H. Electromagnetic Effect in Enhanced Infrared Absorption of Adsorbed Molecules on Thin Metal Films. Surf. Sci. 1986, 175, L787−L793. (23) Suzuki, Y.; Osawa, M.; Hatta, A.; Suëtaka, W. Mechanism of Absorption Enhancement in Infrared ATR Spectra Observed in the Kretschmann Configuration. Appl. Surf. Sci. 1988, 33/34, 875−881. (24) Fano, U. Effects of Configuration Interaction on Intensities and Phase Shifts. Phys. Rev. 1961, 124, 1866−1878. (25) Bjerke, A. E.; Griffiths, P. R.; Theiss, W. Surface-Enhanced Infrared Absorption of CO on Platinized Platinum. Anal. Chem. 1997, 71, 1967−1974. (26) Pecharromán, C.; Cuesta, A.; Gutiérrez, C. Calculation of Adsorption-Induced Differential External Reflectance Infrared Spectra of Particulate Metals Deposited on a Substrate. J. Electroanal. Chem. 2004, 563, 91−109. (27) Su, Z.; Sun, S.; Wu, C.; Gai, Z. Study of Anomalous Infrared Properties of Nanomaterials Through Effective Medium Theory. J. Chem. Phys. 2008, 129, 044707−1−044707−6. (28) Zhu, Y.; Uchida, H.; Watanabe, M. Oxidation of Carbon Monoxide at a Platinum Film Electrode Studied by Fourier Transform Infrared Spectroscopy with Attenuated Total Reflection Technique. Langmuir 1999, 19, 8757−8764. (29) Priebe, A.; Sinther, M.; Fahsold, G.; Pucci, A. The Correlation Between Film Thickness and Adsorbate Line Shape in Surface Enhanced Infrared Absorption. J. Chem. Phys. 2003, 119, 4887−4890. (30) Krauth, O.; Fahsold, G.; Pucci, A. Asymmetric Line Shapes and Surface Enhanced Infrared Absorption of CO Adsorbed on Thin Iron Films on MgO(001). J. Chem. Phys. 1999, 110, 3113−3117.
16761
DOI: 10.1021/acs.jpcc.5b04622 J. Phys. Chem. C 2015, 119, 16754−16761