Open Access
Add to BibSonomyAbstract
Surface second harmonic generation (s-SHG) spectroscopy is a powerful tool to investigate layers or adsorbates on surfaces with high sensitivity. For this nonlinear technique, sophisticated reference methods are needed to properly treat the measured raw data. We present an easy-to-implement reference measurement method for s-SHG spectroscopy for surface layers or adsorbates. It directly allows for extracting reference-corrected s-SHG spectra from raw data. SHG from thin slabs of BK7 and MgO in the spectral range from 450 to 900 nm (fundamental beam) is used to obtain the reference spectrum. The method includes the experimental determination of the dispersive properties of the optical setup over the relevant spectral range. The accuracy of the presented procedure is demonstrated by applying the method to the study of a thin molecular film of 1, 1′-Bi-2-naphthol (Binol) supported on a BK7 substrate.
© 2013 Optical Society of America
1. INTRODUCTION
Optical spectroscopy is a powerful technique to study the electronic and geometric properties of molecules in the gas phase. Many industrial applications such as catalysis, however, require molecules to be supported on a surface. Interaction between the molecule of interest and the support material is known to play a decisive role in the overall performance of the system [1–7], and often its behavior cannot be deduced from available gas phase data. Consequently, it is of great importance to apply optical surface spectroscopic methods to such systems for direct measurements. The relatively large penetration depth of photons into support materials hinders the unambiguous study of surface adsorbates or coatings by common linear spectroscopic methods due to the bulk contribution. Coherent surface second harmonic generation (s-SHG) is dipole forbidden in media with inversion symmetry but is always allowed at surfaces and interfaces where the inversion symmetry is broken normal to the surface. Therefore, s-SHG can be used to investigate surfaces with high sensitivity and with negligible contributions from the bulk if applied to proper media as has been shown previously [8–17].
One crucial step in spectroscopy, when applied over a wide wavelength range, is the correction of the measured data in order to obtain the pristine spectrum of the sample. This is necessary since for most of the common spectroscopic setups, the obtained raw data is a combination of the spectral response of the sample and the spectral properties of the setup (in the following termed “spectrometer function”). The latter are given by the dispersive properties of, e.g., the light source, the detector’s sensitivity, and the optical elements within the beam path, etc. All these lead to an additional wavelength dependency of the measured SH intensity. In experiments using linear spectroscopic techniques, such as light transmission measurements, the correction of the measured data is commonly performed by carrying out two measurements, one of the sample of interest and a second of a suitable reference sample. Subsequently, the final spectrum is obtained by a simple data treatment using the two measured data sets. In s-SHG spectroscopy, the measurement of a reference sample is not trivial. The nonlinear technique is nearly background free, meaning that the SH conversion efficiency in absence of an SHG active resonance is very low. This in combination with the restriction of the SH generation to the closer interface region of a substrate with inversion symmetry, leads to very weak SH intensities, which makes their detection rather challenging. As a consequence of these difficulties, common s-SHG reference techniques employ relative methods (e.g., Maker fringe [18–20] and wedge techniques [18,21,22]) where noncentro-symmetric reference samples are used in order to generate larger SH intensities. By using relative techniques, it is still possible to deliver absolute values, when the sample signal is compared to a reference signal that is typically collected from alpha quartz or ammonium dihydrogen phosphate. Although these materials are commonly used as reference materials, their nonlinear properties have been subject to considerable discussions in the past [23,24], which makes their use as reference rather questionable in terms of the accuracy of the obtained corrected s-SHG spectrum. The major drawback of all reference methods mentioned above is the necessity of an additional beam line (split out of the main beam) for the reference measurements that further complicates the setup and reduces the reliability of the obtained reference data.
Therefore it is favorable to use the very same substrate for the reference measurement as is used as support material for the sample of interest (for the sake of comparability of the two measurements). Consequently, a highly sensitive apparatus is required to detect the weak SH signals in such reference measurements of centro-symmetric substrates. In order to extract the contribution of the spectrometer from the measured reference spectrum, it is crucial to know the exact nonlinear spectral response of the reference sample. Here, several spectral features that only appear in nonlinear spectra of such substrates must be considered as shown in the following sections. The other major challenge toward absolute SHG measurements is the quadratic power dependence of the SH intensity with the consequence that changes in the laser intensity over the applied wavelength range (as are present for most of the commercial available tunable laser sources) have a huge impact on the intensity of the generated SH signal. All the above-mentioned issues further complicate the extraction of a nonlinear spectrum from a measurement using a reference sample. Clearly, a sophisticated and reliable data treatment method is highly needed.
In this contribution we demonstrate an alternative in-line reference measurement method, by applying s-SHG spectroscopy to thin commercially available BK7 glass substrates (Marienfelder, 130 μm thick) in transmission mode, which previously have been used as support material in related systems [25–27]. Earlier reports on nonlinear properties of BK7 are mainly concerned with evaluation of different components of the second-order susceptibility of the material and not much with its wavelength dependence [28,29]. Unlike traditional reference measurement methods, the determination of spectrometer function is included in the method presented here. In an experiment with a thin film of Binol on BK7 substrates, the accuracy of the presented method is demonstrated. Furthermore, the method directly gives precise values for the refractive index of the substrate even in the vicinity of the absorption edge, which are also close to the absorption edge, as well as accurate values of the phase angle, , indicating possible phase-retardation effects that can be caused by resonances in the sample of interest.
In Section 2 a simple analytical expression for s-SHG from a thin slab is introduced. Section 3 describes the experimental apparatus, followed by Section 4, where the results are presented and discussed. In Section 5 our conclusions are given.
2. s-SHG FROM A THIN AMORPHOUS SUBSTRATE IN TRANSMISSION GEOMETRY
At any surface/interface there is no inversion symmetry normal to the surface/interface, and consequently, for materials with inversion symmetry, including amorphous materials, surfaces are the only source of second harmonic generation and thus the method becomes surface specific if the contribution of higher-order terms from the bulk is neglected. Therefore, the sample must be irradiated by -polarized light. In the case of s-SHG measurements using a thin plan-parallel slab, however, several interference effects contribute to the total observed SHG signal. Furthermore, no optical arrangement possesses a constant transmission and sensitivity over a wide range of wavelengths. All these effects must be considered in order to analyze the measured spectra. In this section we give a simple analytical expression for the overall SH signal generated from a thin slab with bulk inversion symmetry in a general case. Figure 1 shows a sketch of the SHG process from a thin plan-parallel slab. An intense laser beam with intensity , strikes on the substrate and travels through it, undergoing multiple reflections that cause interference effects. At both surfaces, frequency doubling occurs and a detector will observe a superposition of these two contributions. The SH beam intensity, , observed at the detector can be expressed as
where is the spectrometer function, is the intensity of the fundamental beam inside the substrate, is the wavelength-dependent relative SHG conversion efficiency of the surface, is the transparency function of the substrate, and is the characteristic time delay related to phase shifts and path differences between the SH beams generated at the two surfaces. The latter is given by Here, is the thickness of the substrate, is the incidence angle of the fundamental beam, is the refractive index at frequency , and is the speed of light in vacuum. The last term in Eq. (1) describes the modulation of the SH beam due to the superposition of the SH beams generated at the two surfaces of the slab. Note that the intensity of the fundamental beam at the harmonic generating interface, , is also angular dependent and Fresnel’s relations and etaloning of the fundamental beam inside the slab govern its angular dependence. The SHG efficiency can be calculated as where is the vacuum permittivity, is the second-order susceptibility of the substrate material, which is treated as a scalar based on the classical model of the driven anharmonic oscillator [30], and is Fresnel’s relation for reflection of -polarized light at the SH wavelength. Using Eqs. (1)–(3) one can determine the angular dependence of the observed SHG signal at any given wavelength by varying the incidence angle and keeping the photon frequency constant. In addition, keeping the incidence angle constant and varying , delivers the wavelength-dependent SHG signal and thus the SHG spectrum. It should be mentioned here that the spectrometer function needs to be determined experimentally, which is the main topic of this contribution. In order to extract from a measured spectrum, however, all the wavelength-dependent quantities of the three other terms in Eq. (1) must be known, especially the transparency function of the substrate as well as its second-order susceptibility.
Fig. 1. llustration of second harmonic generation from a thin plan-parallel slab. Solid curves represent the fundamental laser beam and dashed curves represent the SHG beam. (a) Both the fundamental and the SHG beam leave the substrate at the same angle. (b) For wavelengths where the substrate is transparent to the SHG beam, SH generated at the first interface contributes to the total observed SHG signal. (c) For wavelengths where the substrate is opaque to the SHG beam, THE SHG beam from the first interface is absorbed by the substrate and does not contribute to the total observed SHG signal.
3. EXPERIMENTAL DETAILS
The spectroscopic setup used for these studies is originally designed for investigating optical properties of adsorbed molecules and therefore includes, among others, a molecular evaporator. The setup is described in detail elsewhere [27,31]. Here, a short description and a schematic view of the optical arrangement (Fig. 2) are given for completeness. A midband OPO laser (GWU, premiScan ULD/400) is used as probe laser. The sample is mounted on a four-dimensional manipulator () inside an ultrahigh vacuum chamber. The probe laser light is focused onto the substrate by means of a lens () down to a spot of about . The light leaving the sample, containing both the fundamental and second harmonic wavelengths, is then collimated using an lens. The separation of the fundamental and the generated second harmonic is achieved first by spatial dispersion through a pair of Pellin–Broca prisms mounted on motorized rotatable stages keeping the SH light path constant. Further spectral filtering is performed through a monochromator (LOT-Oriel, 300) before the second harmonic light is detected by a photomultiplier tube (Hamamatsu, H9305-03). A beam splitter is positioned before the first lens, guiding a small fraction of the fundamental laser light to a photodiode for monitoring the fundamental laser intensity. During the measurement, the SH beam intensity and the fundamental laser power are recorded for each laser pulse, simultaneously, also permitting a correction of the measured SH intensity data for the fluctuations of the fundamental laser pulses. Here, the quadratic dependency of the SH intensity on the fundamental laser power must be considered.
Fig. 2. Schematic view of the optical arrangement. BS, beam splitter; lenses; PBP, Pellin–Broca prism; BD, beam dump; PD, photodiode; MC, monochromator; and PMT, photomultiplier tube.
The s-SHG spectrum is obtained by recording the SH light intensity generated at the sample’s surface, as the probe laser wavelength is scanned. Angle dependence measurements are performed at specific wavelengths by rotating the sample, employing the rotational manipulator. In order to reduce noise in the spectra, 100 laser pulses are recorded at each wavelength/angle.
4. RESULTS AND DISCUSSION
A. Angular Dependence
Figure 3 shows the angular dependence of the s-SHG signal of a BK7 glass substrate at two representative fundamental wavelengths. Also shown are the calculated s-SHG signals using Eqs. (1)–(3). Two points deserve extra attention here: first, the very fact that the s-SHG signal from the thin BK7 substrate is recorded over a relatively extended wavelength range is noteworthy. This confirms the high sensitivity of the spectrometer presented here. Although SHG from glass has been reported in the literature for single wavelengths [28,29], the observed signal has been related to the magnetic effects from the bulk to a high extent. Second, the absence of SHG signal at normal incidence indicates that there are no measurable contributions from the bulk and thus justifies the theoretical representation given here. It should be mentioned here that at a fixed wavelength the second-order susceptibility in Eq. (3) is a constant and the angular-dependent part of the susceptibility is accounted for by the sine term in the equation. At 520 nm, the substrate is opaque to frequency-doubled light (260 nm) and the SH generated at the first interface will not contribute to the overall signal (). The fundamental beam, however, undergoes etaloning and causes interference leading to the sharp peaks that can be observed in the measured spectrum at high values of . The good quantitative agreement between calculation and measurement strongly suggests the appositeness of the considered assumptions. Especially, all prominent features of the spectrum are well described by the calculation, which further confirms that their origins are correctly identified. At 640 nm, BK7 is transparent to the second harmonic wave () and s-SHG from the first surface can pass through the substrate’s bulk to interfere with the s-SH beam generated at the second surface leading to the additional observed modulation. Furthermore, all features that were observed at 520 nm are also present at 640 nm. Here again a good match to experimentally measured data is achieved by using the set of equations introduced above, confirming the correctness of the theoretical formalism. This good agreement also supports the suitability of taking as a scalar variable throughout the analytical representation. The Brewster’s angle ( for BK7 in the visible range) is indicated in Fig. 3. As expected, the modulation of the signal caused by the interference of the fundamental beam is absent in the vicinity of this angle, since -polarized light is fully transmitted at Brewster’s angle. However, at higher angles occasionally larger SH signals can be achieved, and the strong interference pattern causes unnecessary complication. Using Brewster’s angle geometry, a large generated SH signal and the absence of the fundamental beam interference are achieved. Consequently, all the following wavelength scans are carried out in this geometry.
Fig. 3. Angular dependence of s-SHG intensity from a BK7 substrate at 520 and 640 nm. Gray curves show measured data and black curves show the simulated data using Eqs. (1)–(3). In both panels the arrow indicates the corresponding Brewster’s angle. Insets are included to demonstrate the good agreement between measurement and simulation.
B. Wavelength Dependence
In a further experiment the s-SHG spectrum of the BK7 glass substrate is measured at . The theoretical description of this experiment is also given by Eqs. (1)–(3). However, for wavelength scans at Brewster’s angle, the equation can be simplified. First, since no interferences of the fundamental beam occur at this angle, can be replaced by , which obviously has no angular dependence. Second, all the quantities that have no wavelength dependence can be summarized in the quantity , which possesses a constant value for wavelength scans performed at Brewster’s angle (note, due to the very little wavelength dependency of the Fresnel coefficients, they can be included in good approximation). Putting Eqs. (1)–(3) together and rearranging, one gets
Since the fundamental laser intensity, , is determined in the experiment simultaneously, the spectrum can easily be power corrected and the resulting spectrum has the theoretical form
The obtained power-corrected s-SHG spectrum of the BK7 substrate is presented in Fig. 4 [the interruption in the experimental data at 350 nm is due to the low intensity of the OPO at the signal-idler crossover wavelength (710 nm)]. It can be observed that above a fundamental wavelength of the interference between the two SH beams generated at the two surfaces of the substrate leads to the expected modulation of the measured SH signal, which is described by the interference term in Eq. (5). In agreement with the theoretical considerations presented above, the onset of the obtained modulation is located at the same position as the adsorption edge of the linear transmission spectrum of BK7 (the dashed curve in Fig. 4). The transmission spectrum is obtained using a commercial UV–Vis spectrometer and correcting the resulting curve for reflection losses. The so-treated UV–Vis spectrum consequently equals the substrate transparency function , which is used in the equations above.
Fig. 4. Laser power-corrected wavelength-dependent s-SHG intensity spectrum of a thin BK7 substrate is represented by a gray solid curve. The solid black curve is the result of the smoothing procedure. The dashed black curve shows the transmission curve of the same substrate at SH wavelengths (top axis).
The next step is the determination of the contribution of the spectrometer function to the measured spectrum. To this end, the oscillatory part of the spectrum must first be removed. The most practical way to do so is to apply a smoothing procedure to the obtained spectrum acting as a low-pass filter. The curve smoothing is analytically equivalent to averaging the cosine term in Eq. (5) to zero. Since both the spectrometer function and the second-order susceptibility of BK7 will not possess any sharp spectral features, smoothing does not have any effects on them. The result can be analytically expressed as
Using Eq. (6), the combined factor can be extracted after dividing the smoothed data by (using the function determined from the UV–Vis spectrum). The resulting curve is included in Fig. 4. In order to examine the contribution of the spectrometer function to the extracted curve, the wavelength dependency of the second-order susceptibility must be determined ( is nearly independent of the photon energy).
For nonabsorbing insulators, it is possible to derive a simple analytical expression for the second-order susceptibility based on the model of the driven anharmonic oscillator [30]:
where is the resonance frequency of the oscillator, is the frequency of the fundamental beam, and is a material specific constant. In the absence of a resonance close to the fundamental and the SH frequency ( and ), shows a smooth, slightly increasing curve toward smaller wavelengths without any pronounced features. A typical candidate of such a material is the MgO crystal, which is fully transparent down to almost 200 nm. To what extent the BK7 glass is similar to this description must be determined since the material is, in contrast to MgO, intransparent below 300 nm, which is within the experimental range of the current measurements. This is done by comparing the fundamental power-corrected s-SHG spectra of the BK7 glass substrate and the MgO crystal of comparable thickness (150 μm) measured using the same setup with both samples positioned at Brewster’s angle. Note, due to the cubic crystal structure of MgO, the second harmonic frequency is also here exclusively generated at the two surfaces. The result of the measurements is depicted in Fig. 5. It can be observed that the modulation of the SH signal from the MgO sample is present throughout the spectrum, which is caused by the transparency of MgO in the entire SH wavelength range. Note that the oscillation frequency of the modulation in SH intensity is here for small wavelengths so high that the peaks cannot be fully resolved. In the range above 600 nm of the fundamental wavelength, the spectra of BK7 and MgO show similar behavior and intensity; however, in order to compare the two curves in the lower wavelength range, the MgO spectrum must be treated. Therefore, the data from the MgO spectrum are smoothed to remove the oscillatory contribution, as was shown in the previous section for BK7. The resulting curve is included in Fig. 5. Mathematically, the smoothed MgO curve can be expressed by introducing () in Eq. (6). The resulting function has the following form:Fig. 5. Laser power-corrected wavelength-dependent s-SHG intensity spectrum of a MgO thin substrate is shown in solid gray. As expected, interference modulation is observed throughout the measured range, since MgO is transparent to all SHG wavelength covered in this measurement. The s-SHG intensity spectrum of BK7 is also shown for comparison in solid black. The dashed black curve represents the result of the smoothing procedure on MgO data.
The BK7 spectrum in the lower wavelength range (below 600 nm of fundamental wavelength) can be described by Eq. (6) by simply setting () yielding the following expression:
Comparing the BK7 spectrum with the smoothed curve of the MgO spectrum, it can be observed that all spectral features that are present in the BK7 spectrum also appear in the latter. This means that all statements that were given regarding the second-order susceptibility of MgO also hold for the BK7 substrate. The absence of a peak in close to the bulk absorption edge further confirms that there are no bulk contributions in the recorded s-SHG signal. It also suggests that the thin interface layer that contributes to the generation of the second harmonic beam behaves optically very differently from the bulk. It can be stated that all features that appear in the s-SHG spectra of MgO and BK7, apart from the interference modulations, are highly dominated by the spectrometer function and that its spectral behavior is well described by the extracted curve in Fig. 4. It is noteworthy that common surface treatments, such as argon sputtering or exposure to ambient atmosphere did not have any measurable effect on the appearance of the s-SHG signal recorded for substrates used in this study. If the nonlinear response of the surface is modified by any treatment, the presented method will still be suitable as it takes the status quo of the substrate as the reference.
C. Reference Measurements
In the previous sections, the nonlinear spectroscopic properties of the BK7 substrate are shown and described theoretically. In order to use this knowledge for the development of a reference technique, further theoretical considerations are needed. Therefore, we suppose that on the backside of the BK7 substrate there is a coated layer of a specific material with an unknown second-order susceptibility (representing a typical sample for our investigations). The generation of the second harmonic radiation now has three sources, the front and the back sides of the substrate as well as the coated material. For simplification we assume that the material is randomly ordered on the surface leading to the same symmetry conditions for the susceptibility tensors and presented in the previous sections (and thus the values are equal). In order to find an analytical expression for the measured SH intensity from such a sample, the different SH contributions must be superimposed, as was done already for the substrate only. The following equation is found:
where the angle, , describes a possible phase shift between the SH contribution of the coated material and that of the backside of the substrate. Measuring the s-SHG spectrum of the pure BK7 substrate as reference and subtracting the result from the measurement of the sample yields, after applying the smoothing procedure, the following expression:Now Eq. (6) can be used to remove the contribution of the spectrometer function, and one gets
For the usual case that (valid if resonances in the coated material are probed), the equation can be further simplified:
The term on the right side of Eq. (13) represents the spectral response of the coated material normalized to the spectral response of the BK7 substrate. Since the latter can be regarded as nearly constant over the applied wavelength range (see the previous section), the term is approximately proportional to , which is the quantity of interest when performing s-SHG measurements of a coated sample. Equation (13) describes the in-line reference measurement procedure that is presented in this contribution. An s-SHG experiment using this procedure consequently consists of the measurement of the bare substrate and subsequently the measurement of the sample of interest.
D. Performance of the Proposed Method
In order to test the reliability of the proposed method, an s-SHG experiment is performed using a thin surface layer of 1, 1’-Bi-2-naphthol (Binol) as a sample. The sample is prepared by evaporation of Binol onto a clean BK7 substrate under UHV conditions. Subsequently, the s-SHG spectra of the sample and the bare BK7 substrate are measured followed by the data treatment presented in Eq. (13). Additionally, a linear UV spectrum of a thick Binol layer coated onto a quartz glass substrate (which in contrast to the BK7 substrate is transparent in the UV region) is recorded for comparison using a standard UV–Vis spectrometer. The s-SHG spectrum of a molecule is in general very similar to the linear spectrum if the SHG resonance enhancement originates from a resonance of the molecule at the second harmonic frequency and if there is no considerable absorption at the fundamental frequency. This rule holds whenever all the peaks that appear in the linear spectrum belong to resonances that are SHG active. Slight changes in peak height may be present, which can be attributed to different transition probabilities between the linear and the nonlinear cases.
Binol is fully transparent in the visible region but possesses several absorption bands in the near UV. Consequently, Binol is a suitable molecule in order to test the performance of the presented reference method. In Fig. 6 the results of the experiment are depicted. In the upper panel the uncorrected raw data of the s-SHG measurement of the Binol sample are presented. Considerable SHG enhancement can be clearly observed. In the panel in the middle, the s-SHG spectrum is shown after correction for the fundamental laser power. It can be observed that some features that are pronounced in the uncorrected spectrum clearly originate from changes in the fundamental laser power as they vanish after the correction. This demonstrates the necessity of the correction. In the lower panel, the s-SHG spectrum of Binol after the complete data treatment using Eq. (13) is shown. Also shown is the linear UV spectrum of the Binol. By comparing the linear UV spectrum with the three s-SHG spectra, it can clearly be observed that the fully corrected s-SHG spectrum fits the linear spectrum where the other two untreated s-SHG spectra possess features that apparently do not represent the true nonlinear response of the Binol layer. Here, spectral effects that are caused by changes in the fundamental laser power and the spectral properties of the spectrometer function lead to a distortion of the real s-SHG spectrum. The fully corrected spectrum, however, is very similar to the linear spectrum, confirming that the reference method presented here yields realistic results on the pure s-SHG spectrum of a coated material (surface adsorbate). The slight differences in the peak positions between the corrected s-SHG spectrum and the linear UV–Vis spectrum that can be observed in Fig. 6 might originate from the different thickness of the probed Binol layer using the two methods. Due to its surface sensitivity, the s-SHG spectrum represents the spectral response of those Binol molecules that are located at the surface, whereas the UV–Vis technique probes the absorption properties of the entire Binol layer. In order to guarantee for a well-resolved UV–Vis spectrum, a comparatively thick Binol layer had to be used for the UV–Vis measurement with the result that the obtained spectrum mainly represents the spectral response of bulk Binol.
Fig. 6. Data treatment procedure is demonstrated using Binol coated BK7 as a test system. The upper panel shows the raw data before any data treatment. The middle panel shows the same data after laser power correction. In the lower panel the black solid curve shows the data after full data treatment including spectrometer function correction. Gray circles indicate single data points. The dashed black curve represents the linear absorption spectrum of Binol on quartz glass measured by a commercial UV–Vis spectrometer.
The use of a thin transparent substrate for s-SHG measurements in transmission geometry allows for extracting important supplementary information from the measurements. The peak positions of the SH interference modulation in the s-SHG spectrum of the bare substrate in Fig. 4 are very sensitive to the values of the refractive indices , as is given by Eq. (2). Accordingly, very precise values for the refractive index of the substrate can be derived from the spectrum. In particular, the values close to the absorption edge, which are normally not easily accessible, can be determined. A very important quantity that can also be determined from these measurements is the phase angle [see Eq. (11)]. It describes the phase of the SH oscillation generated by the coated material with respect to the phase of the SH contribution of the backside of the substrate. This phase information and its dependency on the wavelength (e.g., in the region of resonance transitions) can be easily extracted from the measurement by analyzing the oscillatory part of the s-SHG spectrum.
5. CONCLUSIONS
In this paper a simple analytical description for s-SHG from thin plan parallel optical materials is given. Based on the formalism presented above, the nonlinear response of thin BK7 and MgO substrates is explained and compared to s-SHG measurements. Based on the results, an experimental procedure is developed that allows for correcting the measured s-SHG raw data from any material coated onto a substrate in order to obtain the pure s-SHG spectrum of the coating. This reference method allows any transparent optical material of centro-symmetric structure to be used as reference for s-SHG measurements. The reliability of this reference method is tested by performing an s-SHG measurement of a thin layer of Binol coated onto a BK7 glass substrate. Comparison of the obtained results with a linear UV spectrum of a thick Binol layer coated onto a quartz glass substrate using a commercial UV–Vis spectrometer confirms the strength of the method. The comparison clearly reveals the necessity and the accuracy of the presented reference measurement method. Furthermore, it is mentioned that by analyzing the oscillatory part of the s-SHG spectrum, which is caused by interference effects of the different generated SH beams, phase shifts induced by adsorbed molecules can be extracted.
ACKNOWLEDGMENTS
This work has been supported by the Deutsche Forschungsgemeinschaft (DFG) and the European Research Council (ERC, Project No. 246645-ASC3) through an Advanced Research Grant. A. K. thanks the Royal Society for his Newton International Fellowship (NIF).
REFERENCES
1. M. G. Mason, “Electronic-structure of supported small metal-clusters,” Phys. Rev. B 27, 748–762 (1983). [CrossRef]
2. H. Hovel, S. Fritz, A. Hilger, U. Kreibig, and M. Vollmer, “Width of cluster plasmon resonances—bulk dielectric functions and chemical interface damping,” Phys. Rev. B 48, 18178–18188 (1993). [CrossRef]
3. G. Jacobs, T. K. Das, Y. Q. Zhang, J. L. Li, G. Racoillet, and B. H. Davis, “Fischer-Tropsch synthesis: support, loading, and promoter effects on the reducibility of cobalt catalysts,” Appl. Catal. A 233, 263–281 (2002). [CrossRef]
4. C. Harding, V. Habibpour, S. Kunz, A. N.-S. Farnbacher, U. Heiz, B. Yoon, and U. Landman, “Control and manipulation of gold nanocatalysis: effects of metal oxide support thickness and composition,” J. Am. Chem. Soc. 131, 538–548 (2009). [CrossRef]
5. M. E. Vaida, T. M. Bernhardt, C. Barth, F. Esch, U. Heiz, and U. Landman, “Ultrathin magnesia films as support for molecules and metal clusters: tuning reactivity by thickness and composition,” Phys. Stat. Sol. B 247, 1001–1015 (2010). [CrossRef]
6. J. Lu, P. Serna, C. Aydin, N. D. Browning, and B. C. Gates, “Supported molecular iridium catalysts: resolving effects of metal nuclearity and supports as ligands,” J. Am. Chem. Soc. 133, 16186–16195 (2011). [CrossRef]
7. S. V. Ong and S. N. Khanna, “Origin of oxidation and support-induced structural changes in Pd(4) clusters supported on TiO(2),” J. Phys. Chem. C 115, 20217–20224 (2011). [CrossRef]
8. Y. R. Shen, “Surface 2nd harmonic-generation—a new technique for surface studies,” Ann. Rev. Mater. Sci. 16, 69–86 (1986). [CrossRef]
9. M. Asscher and Z. Rosenzweig, “Adsorbate interaction—an optical 2nd harminic-generation study,” J. Vac. Sci. Technol. A 9, 1913–1918 (1991). [CrossRef]
10. M. Buck, F. Eisert, J. Fischer, M. Grunze, and F. Trager, “Investigation of self-organizing thiol films by optical 2nd-harmonic generation and x-ray photoelectron-spectroscopy,” Appl. Phys. A 53, 552–556 (1991). [CrossRef]
11. A. Galeckas, M. Petrauskas, M. Willander, and Q. Wahab, “Optical 2nd harmonic-generation in reflection from silicon-carbide films” Surf. Interface Anal. 18, 71–72 (1992). [CrossRef]
12. T. Petrallimallow, T. M. Wong, J. D. Byers, H. I. Yee, and J. M. Hicks, “Circular-dichroism spectroscopy at interfaces—a surface 2nd harmonic-generation study,” J. Phys. Chem. 97, 1383–1388 (1993). [CrossRef]
13. R. M. Corn and D. A. Higgins, “Optical 2nd-harmonic generation as probe of surface-chemistry,” Chem. Rev. 94, 107–125 (1994). [CrossRef]
14. Y. R. Shen, “Nonlinear-optical studies of polymer interfaces,” Int. J. Nonlinear Opt. Phys. 3, 459–468 (1994). [CrossRef]
15. J. F. McGilp, “Optical characterization of semiconductor surfaces and interfaces,” Prog. Surf. Sci. 49, 1–106 (1995). [CrossRef]
16. X. Zhuang and Y. R. Shen, “The application of nonlinear optics to the study of polymers at interfaces,” Trends Polym. Sci. 4, 258–264 (1996).
17. T. Bornemann, A. Otto, W. Heuer, and H. Zacharias, “Second harmonic generation by cold-deposited silver films,” Surf. Sci. 420, 224–232 (1999). [CrossRef]
18. I. Shoji, T. Kondo, A. Kitamoto, M. Shirane, and R. Ito, “Absolute scale of second-order nonlinear-optical coefficients,” J. Opt. Soc. Am. B 14, 2268–2294 (1997). [CrossRef]
19. G. Dong, H. Tao, X. Xiao, C. Lin, X. Zhao, and S. Mao, “Mechanism of electron beam poled SHG in 0.95GeS2·0.05In2S3 chalcogenide glasses,” J. Phys. Chem. Solids 68, 158–161(2007). [CrossRef]
20. Y. Takahashi, Y. Benino, T. Fujiwara, and T. Komatsu, “Second harmonic generation in transparent surface crystallized glasses with stillwellite-type LaBGeO5,” J. Appl. Phys. 89, 5282–5287 (2001). [CrossRef]
21. H. Lotem, G. Koren, and Y. Yacoby, “Dispersion of nonlinear optical susceptibility in GAAS and GASB” Phys. Rev. B 9, 3532–3540 (1974). [CrossRef]
22. K. Pedersen, M. Schiek, J. Rafaelsen, and H. G. Rubahn, “Second-harmonic generation spectroscopy on organic nanofibers,” Appl. Phys. B 96, 821–826 (2009). [CrossRef]
23. D. A. Roberts, “Simplified characterization of uniaxial and biaxial nonlinear optical-crystal—a plea for standardization of nomenclature and conventions,” IEEE J. Quantum Electron. 28, 2057–2074 (1992). [CrossRef]
24. C. Bosshard, U. Gubler, P. Kaatz, W. Mazerant, and U. Meier, “Non-phase-matched optical third-harmonic generation in noncentrosymmetric media: cascaded second-order contributions for the calibration of third-order nonlinearities,” Phys. Rev. B 61, 10688–10701 (2000). [CrossRef]
25. J. M. Antonietti, M. Michalski, U. Heiz, H. Jones, K. H. Lim, N. Rösch, A. Del Vitto, and G. Pacchioni, “Optical absorption spectrum of gold atoms deposited on SiO2 from cavity ringdown spectroscopy,” Phys. Rev. Lett. 94, 213402 (2005). [CrossRef]
26. A. Del Vitto, G. Pacchioni, K. H. Lim, N. Rösch, J. M. Antonietti, M. Michalski, U. Heiz, and H. Jones, “Gold atoms and dimers on amorphous SiO2: calculation of optical properties and cavity ringdown spectroscopy measurements,” J. Phys. Chem. B 109, 19876–19884 (2005). [CrossRef]
27. A. Kartouzian, M. Thämer, T. Soini, J. Peter, P. Pitschi, S. Gilb, and U. Heiz, “Cavity ring-down spectrometer for measuring the optical response of supported size-selected clusters and surface defects in ultrahigh vacuum,” J. Appl. Phys. 104, 124313 (2008). [CrossRef]
28. F. J. Rodriguez, F. X. Wang, B. K. Canfield, S. Cattaneo, and M. Kauranen, “Multipolar tensor analysis of second-order nonlinear optical response of surface and bulk of glass,” Opt. Express 15, 8695–8701 (2007). [CrossRef]
29. F. J. Rodriguez, F. X. Wang, and M. Kauranen, “Calibration of the second-order nonlinear optical susceptibility of surface and bulk of glass,” Opt. Express 16, 8704–8710 (2008). [CrossRef]
30. P.-F. Brevet, Surface Second Harmonic Generation (PPUR, 1997).
31. M. Thämer, A. Kartouzian, P. Heister, S. Gerlach, M. Tschurl, U. Boesl, and U. Heiz, “Linear and nonlinear laser spectroscopy of surface adsorbates with sub-monolayer sensitivity,” J. Phys. Chem. C 116, 8642–8648 (2012). [CrossRef]
