16588
J. Phys. Chem. C 2010, 114, 16588–16595
Charge-Transfer Enhancement Involved in the SERS of Adenine on Rh and Pd Demonstrated by Ultraviolet to Visible Laser Excitation Li Cui,†,‡ De-Yin Wu,‡ An Wang,‡ Bin Ren,*,‡ and Zhong-Qun Tian‡ Key Laboratory of Urban EnVironment and Health, Institute of Urban EnVironment, Chinese Academy of Sciences, Xiamen 361021, China, and State Key Laboratory of Physical Chemistry of Solid Surfaces and Department of Chemistry, College of Chemistry and Chemical Engineering, Xiamen UniVersity, Xiamen 361005, China ReceiVed: June 17, 2010; ReVised Manuscript ReceiVed: August 16, 2010
In an attempt to understand the single-molecule SERS of some small nonresonant molecules, such as adenine, it is inevitable to include the chemical enhancement mechanism to provide additional enhancement to the electromagnetic mechanism, although it may be much smaller than the electromagnetic field enhancement. We will report here the first experimental investigation of the charge-transfer (CT) enhancement of protonated adenine molecules on Rh and Pd by performing the potential-dependent SERS using one UV laser (325 nm) and two visible lasers (514.5 and 632.8 nm). A UV laser displays a significant role in the verification of the CT process due to its much larger photon energy and thus a much larger shift of potential of the maximum SERS intensity (Emax) than visible lasers. We find a well-discernible Emax and a linear relationship between Emax and the photon energy of the laser for adenine on both Rh and Pd surfaces. The Emax was found to shift positively with the increasing photon energy, which strongly indicates an electron transfer from the Ef of Rh and Pd to the lowest unoccupied orbital of adenine molecules. In addition, different contributions of CT enhancement to adenine Raman bands are also briefly discussed. By analyzing the wavelength-dependent intensity change and UV-vis absorption spectroscopy, we propose the contribution of preresonance Raman enhancement to the UV-SERS signal for the band at around 1330 cm-1. The present study demonstrates that the use of a UV laser opens a promising way to understand the enhancement mechanism, especially the chemical enhancement mechanism. Introduction Surface-enhanced Raman spectroscopy (SERS) has been experiencing a surge since the late 1990s following the observation of single-molecule SERS (SM-SERS).1-4 Thereafter, a lot of efforts have been devoted to understanding the mechanisms responsible for the single-molecule sensitivity. It becomes clear now that the huge SERS enhancement is contributed dominantly by the electromagnetic (EM) field enhancement.5 The molecule may experience a giant EM field enhancement when it resides in the narrow gap of two closely approaching nanoparticles (forming the so-called hot spot), and the enhancement factor at the hot spot can reach a value of as high as 1011.6 Early SM-SERS tended to use dye molecules, for example, rhodamine-6G or hemoglobin, with a very large Raman cross section as the probe molecules, and the obtained SM-SERS signal contained the contribution of the resonant or preresonant Raman effect.1,5 Recently, there are some SM-SERS reports on nonresonant small organic molecules, such as adenine and 4-mercaptopyridine.3,4,7 The solo contribution of the EM field to the giant enhancement is quite doubtful for these nonresonant molecules because the extremely strong electromagnetic fields that is necessary for the observation of SMSERS may be sufficiently strong to decompose molecules or even induce sintering of two nanoparticles.8-10 Therefore, it is inevitable that one has to consider other mechanisms that may make additional contributions to the total enhancement. * To whom correspondence should be addressed. E-mail: bren@ xmu.edu.cn. Fax: +86-592-2186979. † Chinese Academy of Sciences. ‡ Xiamen University.
Chemical enhancement mechanisms have also been proposed for interpreting the SERS feature that cannot be fully understood by the EM mechanism since the very early stage of SERS. Among them, the charge-transfer (CT) mechanism is the most important one. It involves the complex interactions among metal nanostructures, molecules, and photons, which add extra difficulties in exploration of the complex CT process. Many attempts have been devoted to understanding the CT processes.11-15 More and more experimental and theoretical evidence has been accumulated to confirm and even quantify the CT contribution on the total SERS and evaluate the selection rule of CT.11-15 Experimentally, various direct and indirect methods have been proposed to isolate the CT process and demonstrate the enhancement by the CT process. Haran and co-workers estimated the CT enhancement of 3-4 orders of magnitude for 4-mercaptopyridine (4MPY) on silver by comparing the Raman cross section of 4MPy adsorbed in silver hot spots with benzene in solution, a molecule with a similar structure to 4MPY.13 Campion and co-workers attributed the SERS signal of molecules on smooth metal surfaces to the contribution of a CT enhancement, by considering that a smooth surface cannot effectively couple with the EM field of the excitation light to generate the enhanced strong EM filed.16 Zhao and co-workers attributed the SERS enhancement from semiconductor nanoparticles to CT enhancement because the excitation line is far from the surface plasmon resonance frequency of semiconductors.17,18 Domke and Pettinger proposed the contribution of CT enhancement to the TERS signal of adenine adsorbed on Au(111) because the signal from adenine is much stronger than other DNA bases with very similar structures to adenine.19
10.1021/jp1055717 2010 American Chemical Society Published on Web 09/14/2010
CT Enhancement in the SERS of Adenine on Rh and Pd It has been proposed that the most reliable method to study CT enhancement experimentally is to examine the dependence of SERS spectra on electrochemical potentials and photon energy of excitation lines.20-22 CT is a resonance Raman-like process, which can be tuned into and out of resonance by changing either the Fermi level (Ef) of the metal or the photon energy of the excitation line (hν). The Raman intensity will reach the maximum in the resonance state where the energy gap between Ef and the molecule electronic level (Em) matches the photon energy (Ei ≈ |Ef - Em|).20-22 In an electrochemical environment, Ef can be easily tuned by changing the applied potential, so the strictest way to prove CT experimentally is to seek the potential at which the intensity reaches maximum (Emax) and then examine whether Emax varies linearly with the photon energy of the excitation line.21,23-28 If Emax shifts more positively as hν increases (a positive slope), the process corresponds to a charge transfer from Ef of the metal electrode to the lowest unoccupied orbital of the molecules. Contrarily, if Emax shifts negatively with hν (a negative slope), the charge transfer is from the highest occupied orbital of the molecule to the Ef of the metal.29 Although this method has been successfully employed in many molecule/metal systems, such as Au(Ag)/pyridine25-28,30,31 and Ag/CN- (SCN-),21,24,32 there are still many systems that are difficult to prove the presence of CT via this way.13,33 The main reason is that the electrochemical window is too narrow to allow the observation of Emax or that the change of Emax with the change of the photon energy is very small. It is well known that the SERS intensity is not only influenced by the CT process, but also by other factors, such as the coverage of molecules on the metal surface, the adsorption orientation,34-36 and the EM enhancement,35,37-39 which may be significantly influenced by the electrode potential, especially at extremely negative or positive potentials. If the influence of these latter factors is more significant than the CT process, it will be difficult to observe the resonance-shaped I-E curve. The surface coverage of many molecules usually decreases at very negative potentials, especially in the potential region of the hydrogen evolution.40-42 Furthermore, Emax for some molecules, such as mercaptopyridine, adenine, pyridine, and pyrazine, appears at a very negative potential or even in the potential region of hydrogen evolution with visible laser excitations.13,43-46 The SERS intensity will inevitably be influenced by factors other than CT, which makes the clear identification of Emax difficult. Such a phenomenon is quite common and can be found in many previous studies.13,30,31 For example, Haran et al. observed very broad I-E curves for 4MPy on Ag islands, which makes it difficult to identify the shifting direction of Emax at three excitation lines of 488, 532, and 633 nm.13 An identical Emax has been reported for 647.1 and 1064 nm excitations in the Au/pyridine system.30 A more extreme case is that when the properties of the metal substrate (for example, Pd) changes at the potentials of hydrogen evolution and the discussion of the SERS intensity becomes very risky.33 The above problems can be well-solved if a shorter excitation wavelength, such as a UV laser, is employed. A UV laser has a much higher photon energy than visible ones. When it is used for investigating systems exhibiting a metal-to-molecule charge transfer, that is, Emax shifts positively with the increasing photon energy, it will result in a large positive shift of Emax. By this way, one can effectively avoid the problem associated with the unfavorable effects of negative potentials and possibly obtain a more discernible Emax. The CT process in more metal/molecule systems may thus be clearly demonstrated by using a UV laser.
J. Phys. Chem. C, Vol. 114, No. 39, 2010 16589 Despite the great advantages of a UV laser in the study of the charge-transfer process over visible lasers, there is no relevant report by now. The main reason is due to the difficulties in obtaining observable SERS spectra with the UV laser excitation (UV-SERS) on many metals, including the three typical SERS metals of Au, Ag, and Cu, that show extremely strong enhancement in the visible and near-infrared region. In recent years, after the first report of UV-SERS on Rh and Ru electrode surfaces in 2003,47 other metals, such as Co (2006), Al (2007), and, more recently, Pt and Pd (2008 and 2009), have been successfully demonstrated to be UV-SERS-active.43,48-51 UV-SERS spectra of various molecules, such as adenine, pyridine, crystal violet, and SCN-, have also been obtained on these metals.43,47-51 Among them, adenine is an important DNA base and has been extensively used as a probe in various SERS studies owing to the very strong SERS signal.3,4,19,52 Both SM-SERS and TERS spectra have been obtained on this nonresonant small organic molecule with visible exctiations.3,4,19,52 Although charge-transfer enhancement of adenine on metals has been predicted to exist,4,19 there is still no direct and convincing experimental evidence to support it. In this paper, we will present the first experimental investigation of the CT enhancement of adenine by probing the potentialdependent SERS spectra excited with one UV laser (325 nm) and two visible lasers (514.5 and 632.8 nm). The advantage and indispensable role of UV laser excitation in CT studies will be shown. Persuasive proof to demonstrate the CT contribution to SERS of adenine on Rh and Pd will be presented. The chargetransfer direction, charge-transfer efficiency, and different contribution of CT enhancement to adenine Raman bands will also be analyzed. Finally, on the basis of the excitation linedependent relative intensity change of two main Raman bands, the enhancement mechanisms, including the preresonance enhancement, CT, and EM field involved in UV-SERS of adenine on metal, will be analyzed. Experimental Section The SERS spectra excited with the laser lines of 325 nm (He-Cd laser) and 514.5 nm (Ar+ ion laser) were recorded on a UV-vis R1000 Renishaw micro-Raman system equipped with a UV-enhanced CCD detector (Renishaw, U.K.). For 325 nm, a 15× fused silica UV objective with a working distance (WD) of 8.5 mm and a numerical aperture (NA) of 0.32 was used. For 514.5 nm, a 50× objective (Leica) with a WD of 8.5 mm and a NA of 0.55 was used. The laser powers on the sample were 7.1 and 6.9 mW for 325 and 514.5 nm, respectively. SERS spectra excited by a 632.8 nm He-Ne laser were acquired on a LabRam I confocal microscope Raman system (JY, France) with a 50× objective (Olympus) with a long WD of 8 mm and an NA of 0.55. In the above two Raman systems, the objectives were used to both direct the laser to the sample surface and collect the Raman signal in the backscattering geometry. UVSERS spectra presented in this paper have been subtracted with the solution spectrum using the water band at 1635 cm-1 as the internal standard. A spectroelectrochemical cell, with a Pt wire and a saturated calomel electrode (SCE) serving as the counter and the reference electrodes, was used for both the electrochemical and the in situ electrochemical (EC) SERS measurements, respectively. The solution used for EC-SERS measurement is 1 mM adenine + 0.1 M ClO4- (pH ) 2.8). Perchlorate ion (ClO4-) was used as the electrolyte due to its much weaker adsorbability on the electrode surface and the least influence on the charge-transfer
16590
J. Phys. Chem. C, Vol. 114, No. 39, 2010
Cui et al.
process. There are two reasons for choosing the solution pH value of 2.8. First, the value of pH 2.8 is lower than the pKa of adenine (3.8),53 so protonated adenine molecules will form. We found in our previous work that the protonated adenine shows much better thermal stability to the UV laser than its neutral form and thus ensures the acquisition of high-quality UV-SERS spectra.43,50 Second, compared with an even lower pH value, such as pH 1, the hydrogen evolution potential will shift to a much more negative potential, which is ca. -0.9 V at pH 2.8 and ca. -0.3 V at pH 1. Thus, pH 2.8 will ensure that the spectra will be obtained over a wide potential range, which greatly facilitates the observation of Emax, especially for two visible laser excitations. The details about the preparation of the roughened Rh electrode via electrochemical oxidation reduction cycles and synthesis of Au core-Pd shell (Au@Pd) nanoparticles can be found elsewhere.43,54 For Au@Pd nanoparticles, the Au cores have a diameter of ca. 60 nm and the thickness of the Pd shells is ca. 20 nm, which showed optimum UV-SERS signals among other Au@Pd nanoparticles.43 Results and Discussion EC-SERS Spectra of Adenine on Rh with Three Laser Excitations. Figure 1 shows the SERS spectra of adenine adsorbed on an electrochemically roughened rhodium electrode surface with 325 (UV), 514.5, and 632.8 nm (visible) laser excitations at the potentials shown in the figure. The spectra shown in Figure 1 are all acquired with a same potential changing sequence from -0.95 V to a positive potential of ∼0.3 V. Potentials more negative than -0.95 V will lead to a severe hydrogen evolution and thus interfere with the acquisition of SERS spectra. The two dominant Raman bands appear at 734 and 1321 cm-1. The band at 734 cm-1 is assigned to a ringbreathing mode. The band at 1321 cm-1 may have two contributions. The first one involves the C5-N7 and N1-C2 symmetric stretching coupled with the C2-H and C8-H symmetric in-plane bending. The second one involves the C2-H, C8-H, and N9-H in-plane bending with a minor contribution from the C6-N1, C8-N9, and N3-C4 stretches. Both of them belong to in-plane vibrational modes.55 It can be seen that the spectra intensity changes dramatically with both applied potentials and excitation lines, especially the band at ca. 1321 cm-1. Besides the intensity changes with the applied potentials, the shape of the band at ca.1321 cm-1 also changes from a narrow one at E g -0.6 V to a broad one composed of two peaks at 1310 and 1327 cm-1 at E < -0.6 V with 632.8 and 514.5 nm excitations. With the 325 nm excitation, the spectra acquired at potentials more negative than -0.4 V become very weak and thus are not shown here. To understand this change, cyclic voltammograms on a roughened Rh electrode in 0.1 M ClO4- at pH 2.8 with and without 1 mM adenine are acquired (Figure 2). Two pairs of redox peaks appear at about -0.4 and -0.9 V in both solutions, which may indicate that these reactions are irrelevant to adenine molecules. In the solution without adenine, the redox peaks (peaks 1 and 2) at -0.4 V should come from the oxidation and reduction of free H+ in the solution. Because the amount of free H+ at pH 2.8 is very limited, with the negative shift of the potential, H+ in the vicinity of the electrode surface will be consumed and the current will be under diffusion control. When the potential is moved to -0.9 V, redox peaks (peaks 3 and 4), which are related to water reduction and H2 oxidation, appear. In the solution containing adenine, the redox peaks (peaks 1
Figure 1. Potential-dependent SERS spectra from protonated adenine adsorbed on an ORC roughened Rh electrode surface with excitation wavelengths of (A) 632.8, (B) 514.5, and (C) 325 nm in 1 mM adenine + 0.1 M ClO4- solution at pH 2.8. The band marked with a star is from ClO4-. Arrows indicate the potential changing sequence.
and 2) at -0.4 V should come from the reduction and oxidation of H+ in both the electrolyte solution and the protonated adenine molecules. The consumption and production of H+ will inevitably change the pH value of solution, especially for those H+ close to the electrode surface and thus further drive the
CT Enhancement in the SERS of Adenine on Rh and Pd
J. Phys. Chem. C, Vol. 114, No. 39, 2010 16591
Figure 2. Cyclic voltammograms of a Rh electrode in 0.1 M ClO4solution at pH 2.8 with (solid line) and without adenine (dotted line).
Figure 3. SERS spectra of adenine adsorbed on a roughened Rh electrode surface with 632.8 nm excitation under different pH and potential conditions.
protonation and deprotonation of adenine molecules (see the following reaction).
adenineH+ T adenine + H+
(1)
When the potential is positively moved from -0.95 V, oxidation of H(0) (peak 1) commences at about -0.6 V and produces H+, which will result in a lowering of the pH value and protonation of adenine. This is consistent with the SERS results in which transformation of the band shape also occurs at E g -0.6 V. On the basis of this result, we can deduce that it is neutral adenine on Rh at E < -0.6 V, whereas it is protonated adenine on Rh at E g -0.6 V. To further confirm this result, SERS spectra of adenine on Rh at -0.95 V in a neutral solution of 1 mM adenine + 0.1 M NaClO4 (pH 7) were acquired with the 632.8 nm excitation (Figure 3, spectrum b). It can be seen that spectrum b is similar to the spectra acquired at -0.95 V in pH 2.8 solution (Figure 3, spectrum a, extracted from Figure 1A). The band above 1300 cm-1 is also composed of two peaks at 1310 and 1327 cm-1 and shows the same bandwidth. In addition, the SERS spectrum from pH 2.8 at -0.3 V (Figure 3, spectrum c, extracted from Figure 1A) is similar to the spectra from pH 1 solution acquired at -0.3 V (Figure 3, spectrum d), showing a same narrow band at 1321 cm-1. It should be mentioned here that -0.3 V is the most negative potential at which SERS spectra can be acquired at pH 1 solution due to the severe hydrogen evolution. This
Figure 4. Potential-dependent integrated intensity of the band at 734 cm-1 (A) and bands at 1321 and 1340 cm-1 (B) for 632.8, 514.5, and 325 nm excitations on the Rh surface. The intensities are normalized against the maximum values at each excitation. To avoid the error induced by deconvoluting the two overlapped peaks at 1321 and 1340 cm-1, the integrated intensity was thus obtained from both of them. (C) Plot of the incident photon energy hν versus Emax from (B).
confirms again that neutral adenine molecules are adsorbed on Rh at E < -0.6 V and protonated adenine molecules are adsorbed on Rh at E g -0.6 V. A similar phenomenon was also observed with 514.5 nm excitations. Although the above change seems to complicate this system, it will not influence the study of the charge-transfer enhancement of protonated adenine molecules on Rh because all the conditions are the same except the excitation lines. In addition, Emax at all of the three excitation lines are more positive than -0.6 V, which corresponds to the protonated adenine adsorbed on the Rh electrode. Analysis of the Charge-Transfer Effect in Protonated Adenine/Rh Systems. The normalized integrated intensities of the band at 734 cm-1 and the two bands at 1321 and 1340 cm-1 are plotted in Figure 4A,B as a function of applied potentials for 632.8 nm (1.96 eV), 514.5 nm (2.41 eV), and 325 nm (3.80
16592
J. Phys. Chem. C, Vol. 114, No. 39, 2010
eV) excitations. To avoid the error induced by deconvoluting the two overlapped peaks at 1321 and 1340 cm-1, the integrated intensity was thus obtained from both of them. The intensities are normalized against the maximum intensity observed in the investigated potential range for each excitation. Compared with the 734 cm-1 band, the 1321 and 1340 cm-1 bands show a much better resonance-shaped I-E profile, and Emax can be obviously observed at -0.5 V for the 632.8 nm excitation and shifts to a more positive potential at -0.42 V for the 514.5 nm excitation and 0.05 V for the 325 nm excitation. Figure 4C plots the laser photon energy hν versus Emax of the 1321 and 1340 cm-1 bands. A linear relationship with a positive gradient of 3.24 eV/V is obtained. This result can be well interpreted in terms of the CT mechanism and the direction of the CT is from the Ef of Rh to the lowest unoccupied orbital of protonated adenine molecules. However, the band at 734 cm-1 does not yield a maximum intensity as clear as that of 1321 + 1340 cm-1 bands with the change of the potential in the wide potential range, especially with 632.8 and 514.5 nm excitations (Figure 4A). From the I-E curves at 632.8 and 514.5 nm excitations, it will be very risky to judge whether a CT effect exists. The obvious Emax shift with the 325 nm excitation from those of the two visible laser excitations confirms the presence of metal-to-molecule CT. The indispensable role of UV laser excitation in CT studies is thus clearly demonstrated. Here, one question may be raised in respect to different behaviors of these dominant Raman bands: why the Emax of bands at 1321 and 1340 cm-1 is more discernible than the 734 cm-1 band. As we know, the SERS intensity can be influenced not only by the CT process but also by other factors, such as coverage, adsorption orientation,34,56 and even the possible variation of the EM enhancement with potentials.35,37-39 A clearer resonance shape of bands at 1321 and 1340 cm-1 may predict that they are more enhanced via the CT process than the band at 734 cm-1. On the basis of the energetic consideration, one might expect the slope to be (1 eV/V. However, the d(hν)/dEmax slope obtained by the linear fitting is 3.24 eV/V, which is much larger than 1 eV/V. Such a phenomenon was very common and has been observed in many other systems with charge-transfer enhancements, such as Ag/pyridine, Ag/pyrazine, Ag/benzene, Ag/imidazole, Ag/CN-, Au/pyridine, Co/pyridine, etc.21,24-28,31,57 The gradient was found to depend on both the chemical nature and the concentration of electrolyte ions and the adsorbate.27,28,45 Furtak et al. found the increase of the slope with the increasing concentration of Cl-.27 Otto et al. found that the slopes follow the order of Br- > Cl- > SO42- for the same adsorbed molecule and is different for benzene, pyridine, and pyrazine in the same electrolyte.45 The Lombardi group and Rubim groups found that the slopes are different for a series of substituted pyridines in the same electrolyte.28,58 It may be due to the adsorption of electrolyte ions or molecules on the surface, which results in a difference in the local potential at the metal surface from the applied potential. The change in the local potential causing the photon energy shift may be larger than changes in the externally applied potentials and thus leads to a slope larger than 1 eV/V.27,45 Otero et al. provided a model and expressions to address this phenomenon, accompanied by a relatively detailed discussion about the influence of the electron affinity of adsorbates and the nature of metals on the CT process.59 This explanation could also be adapted to the adenine/Rh system. As the ClO4- ion with a very weak adsorbability was used here as the electrolyte, the adenine molecule should be the dominant species, leading to the larger slope in the adenine/Rh system.
Cui et al.
Figure 5. Potential-dependent integrated intensity of the band at 734 cm-1 (A) and the two bands at 1321 and 1340 cm-1 (B) with 632.8, 514.5, and 325 nm excitations on a palladium electrode. The intensities are normalized with the maximum values for each excitation line. (C) Plot of the laser photon energy hν versus Emax from (B).
Analysis of the Charge-Transfer Effect in the Protonated Adenine/Pd System. The above phenomenon is not limited to the Rh surface. For comparison, we also studied the CT enhancement of adenine on a Pd surface. Figure 5A,B shows the normalized integrated intensity of the band at 734 cm-1 and bands at 1321 and 1340 cm-1 versus the applied potential with 632.8, 514.5, and 325 nm excitations. Similar to that in adenine/ Rh systems, the intensity of the 734 cm-1 band changes slightly in a wide potential range with both 632.8 and 514.5 nm excitations. The broad feature makes it almost impossible to obtain an exact Emax and to determine the CT process. However, an obviously positive shift of the I-E curve with the UV laser excitation supports the metal-to-molecule electron-transfer enhancement of the 734 cm-1 band. The normalized intensity of the summed intensity of the two bands at 1321 and 1340 cm-1 shows a resonance-shaped I-E profile, and the Emax obviously shifts positively with the excitation energy. Figure 5C plots the excitation photon energy versus the Emax. The linear fitting of the data produces a positive gradient (d(hν)/dEmax) of 4.53 eV/V, proving the presence of electron-transfer enhancement for adenine adsorbed on a Pd
CT Enhancement in the SERS of Adenine on Rh and Pd
J. Phys. Chem. C, Vol. 114, No. 39, 2010 16593
SCHEME 1: Energy Diagrams of Metals, Adenine, and Adenine Adsorbed on Metals and the LUMO of Protonated Adenine
surface and that the transfer direction is from Pd to adenine, which is the same as that of adenine on Rh. We notice that the slope of d(hν)/dEmax (4.53) in adenine/Pd is different from that in the adenine/Rh system (3.24). This may be a result of both the electron-donating ability of metals and the different adsorption configurations of adenine on Rh and Pd (therefore, the different interfacial structures). However, these assumptions still need more delicated evidence. According to our calculations (B3LYP/6-311+G**[C, N, and H atoms] under the polarized continuum model) and the literature, the most stable configuration for neutral adenine adsorption takes place at the 1-N position. The protonated adenine is adsorbed through the N7 or N3 position on Rh and Pd surfaces. This results in lowering of the energy level of the lone paired orbital in nitrogen atoms. The LUMO level of protonated adenine is located at -2.03 eV with respect to the vacuum level. The LUMO energy levels of adsorbed adenine on two metal surfaces were estimated to be at -1.55 eV for Rh and -2.06 eV for Pd, which may be a result of the different interaction between adenine and the two metals. The electron work functions of bulk Rh and Pd are 4.98 and 5.22 eV, respectively.60 Meanwhile, their potentials of zero charge are at -0.004 and 0.1 V versus SHE.61 On the basis of the data mentioned above, we can estimate the LUMO orbital of adenine (adsorbed adenine) was about ∼2.95 eV (3.43) and ∼3.09 eV (3.06) higher than the Fermi levels of Rh and Pd at the potentials of zero charge. Energy diagrams of the metals, adenine, adsorbed adenine, and the LUMO of protonated adenine molecules are presented in Scheme 1. This supports the existence of a photondriven charge-transfer mechanism for adenine adsorbed on the two metal electrode surfaces. Analysis of the Excitation-Line-Dependent Relative Intensity Change and Contribution of Preresonance Enhancement to UV-SERS. The other notable difference for the three excitation lines is the relative intensity of the bands at 1321 and 1340 cm-1 to the band at 734 cm-1, that is, I(1321+1340)/I734. Figure 6A compares the SERS spectra acquired at -0.1 V with three excitation lines. It can be clearly seen that the intensity ratio increases with the photon energy of the excitation line. In addition, it is interesting to find that the intensity ratios, that is, I1337/I718 for adenine dissolved in 70% HClO4 (Figure 6B) and I1330/I721 for adenine solid powders (Figure 6C), display the same changing trend with the excitation photon energy hν. A high concentration of 70% HClO4 solution was used for the following two reasons. First, the solubility of adenine can be greatly
improved, which ensures the acquisition of Raman spectra of adenine solution with all of the three excitation lines; second, adenine is protonated. The quantum efficiency of CCD at different wavelengths was also considered; however, their influence on the relative intensity change is very small and even negligible. As we know, the resonance Raman (RR) effect can selectively enhance certain vibrational modes and this enhancement is wavelength-dependent.62 If the RR effect exists, we may expect the same changing trend for the intensity ratio in SERS and normal Raman spectra. To verify this assumption, UV-vis absorption spectra from a solution of 10 mM adenine + 0.1 M ClO4- with a pH value of 2.8 were obtained (see Figure 7). It can be obviously seen that, although the absorption is peaked at about 240 nm, the tail of this absorption does extend to 325 nm, which indicates that pre-RR enhancement may exist with the 325 nm excitation. A more persuasive way is to acquire the wavelength-dependent Raman spectra of adenine. For this purpose, it is necessary to use an internal standard to remove the factor originating from the instrument response. Here, we use HClO4 as the internal standard for the following two reasons. First, HClO4 has no absorption at above 200 nm, which can be seen in Figure 7. Second, in our experiment, adenine was dissolved in HClO4, and therefore, we can obtain the Raman signal from adenine and HClO4 simultaneously from the same volume of solution at each excitation, which is not achievable by other calibrating standards. Figure 6D shows the Raman spectra from adenine dissolved in HClO4 with three excitations. The intensity of each spectrum is normalized with that of ClO4at 933 cm-1. The intensity of adenine bands with 325 nm excitation is obviously higher than that with the other two visible laser excitations, especially the band at 1337 cm-1. This evidence convincingly confirms that the pre-RR effect contributes to both SERS and NRS with the UV excitation. Furthermore, one may infer that the 1330 cm-1 band is selectively enhanced by pre-RR and thus gives rise to a larger intensity ratio than that with two visible laser excitations. We did not observe absorption in the UV-vis spectra at two visible wavelengths, but we still observe some difference in the intensity ratio with 514.5 and 632.8 nm excitations. It could be easily understood because the pre-RR is more sensitive than the UV-vis absorption, resulting in that the selective enhancement by pre-RR is observed. In fact, Maria et al. showed that the first singlet excited state of protonated adenine has an excited energy of about 4.10 eV,63
16594
J. Phys. Chem. C, Vol. 114, No. 39, 2010
Cui et al.
Figure 6. (A) SERS spectra of the protonated adenine adsorbed on a Rh electrode surface, extracted from Figure 1. (B) Raman spectra of adenine solution dissolved in 70% HClO4, whose Raman bands have been subtracted in order to make the adenine bands obvious. (C) Raman spectra of adenine solid powders. (D) Raman spectra of adenine solution dissolved in 70% HClO4 normalized with the band at 933 cm-1 from ClO4- with 325, 514.5, and 632.8 nm excitations.
Figure 7. UV-vis absorption spectra from a solution of 10 mM adenine + 0.1 M ClO4, pH 2.8, and 0.1 M HClO4.
corresponding to the excitation wavelength about 302 nm. We demonstrate here that the vibrational mode with a frequency of 1337 cm-1 is sensitive to structural change due to the electron transition to the LUMO. In our previous work, we have demonstrated that the localized surface plasmon resonance of Au@Pd nanoparticles and Pd nanovoids mainly accounts for the UV-SERS enhancement.43,50 We conclude in this work that CT (charge-transfer) enhancement contributes to the UV-SERS in the adenine/Pd or Rh systems in addition to the EM (electromagnetic) and pre-RR (preresonance Raman) enhancement. Conclusion In this work, we performed a systematic experimental investigation of charge-transfer enhancement of protonated adenine molecules adsorbed on Rh and Pd, mainly through
acquisition of potential-dependent SERS spectra excited with a UV laser and two visible lasers. The use of a UV laser has demonstrated a much larger shift of Emax than that shown by two visible lasers. CT enhancements of the two dominant Raman bands of adenine at ca. 1330 and 734 cm-1 are investigated. The band at 1330 cm-1 shows an obvious Emax that shifts linearly with the photon energy of the excitation lines, providing sufficient proof for the existence of CT enhancement. The positive gradient of d(hν)/dEmax of this band indicates a metal-to-molecule electron transfer. Moreover, by analyzing the excitation-wavelength-dependent intensity change of both solution and surface species, we revealed that preresonance enhancement is involved in the selective enhancement of the 1330 cm-1 band in UV-SERS. Therefore, in the UV-SERS of adenine on both Rh and Pd, three kinds of enhancement mechanisms are involved, that is, chargetransfer, electromagnetic field, and preresonance Raman effects. This paper convincingly demonstrates the advantage of using a UV laser in studying the complex CT mechanisms, especially for those systems with a very small Emax shift. UV-SERS may also become an important tool for studying the enhancement mechanism involved in SM-SERS and may also offer a promising way for studying the correlation between the chargetransfer process related to molecular conductance in the field of molecular electronics with the charge-transfer process in SERS. Acknowledgment. This work was supported by the Natural Science Foundation of China (Nos. 20825313, 20827003, 20973143, and 20903076) and MOST of China (2009CB930703).
CT Enhancement in the SERS of Adenine on Rh and Pd References and Notes (1) Nie, S. M.; Emory, S. R. Science 1997, 275, 1102. (2) Kneipp, K.; Kneipp, H.; Kartha, V. B.; Manoharan, R.; Deinum, G.; Itzkan, I.; Dasari, R. R.; Feld, M. S. Phys. ReV. E 1998, 57, R6281. (3) Blackie, E. J.; Le Ru, E. C.; Etchegoin, P. G. J. Am. Chem. Soc. 2009, 131, 14466. (4) Futamata, M. Faraday Discuss. 2006, 132, 45. (5) Xu, H.; Bjerneld, E. J.; Kall, M.; Borjesson, L. Phys. ReV. Lett. 1999, 83, 4357. (6) Xu, H. X.; Aizpurua, J.; Kall, M.; Apell, P. Phys. ReV. E 2000, 62, 4318. (7) Wang, Z. J.; Rothberg, L. J. J. Phys. Chem. B 2005, 109, 3387. (8) Otto, A.; Bruckbauer, A.; Chen, Y. X. J. Mol. Struct. 2003, 661, 501. (9) Otto, A. J. Raman Spectrosc. 2005, 36, 497. (10) Lombardi, J. R.; Birke, R. L. J. Chem. Phys. 2007, 126, 244709. (11) Zhao, L.; Jensen, L.; Schatz, G. C. J. Am. Chem. Soc. 2006, 128, 2911. (12) Zhao, L. L.; Jensen, L.; Schatz, G. C. Nano Lett. 2006, 6, 1229. (13) Shegai, T.; Vaskevich, A.; Rubinstein, I.; Haran, G. J. Am. Chem. Soc. 2009, 131, 14390. (14) Creighton, J. A. Surf. Sci. 1986, 173, 665. (15) Centeno, S. P.; Lopez-Tocon, I.; Arenas, J. F.; Soto, J.; Otero, J. C. J. Phys. Chem. B 2006, 110, 14916. (16) Campion, A.; Ivanecky, J. E.; Child, C. M.; Foster, M. J. Am. Chem. Soc. 1995, 117, 11807. (17) Wang, Y. F.; Ruan, W. D.; Zhang, J. H.; Yang, B.; Xu, W. Q.; Zhao, B.; Lombardi, J. R. J. Raman Spectrosc. 2009, 40, 1072. (18) Yang, L. B.; Jiang, X.; Ruan, W. D.; Zhao, B.; Xu, W. Q.; Lombardi, J. R. J. Phys. Chem. C 2008, 112, 20095. (19) Domke, K. F.; Zhang, D.; Pettinger, B. J. Am. Chem. Soc. 2007, 129, 6708. (20) Gersten, J. I.; Birke, R. L.; Lombardi, J. R. Phys. ReV. Lett. 1979, 43, 147. (21) Otto, A.; Billmann, J.; Eickmans, J.; Erturk, U.; Pettenkofer, C. Surf. Sci. 1984, 138, 319. (22) Ueba, H. Surf. Sci. 1983, 131, 347. (23) Osawa, M.; Matsuda, N.; Yoshii, K.; Uchida, I. J. Phys. Chem. 1994, 98, 12702. (24) Billmann, J.; Otto, A. Surf. Sci. 1983, 138, 1. (25) Otto, A. J. Electron Spectrosc. Relat. Phenom. 1983, 29, 329. (26) Billmann, J.; Otto, A. Solid State Commun. 1982, 44, 105. (27) Furtak, T. E.; Macomber, S. H. Chem. Phys. Lett. 1983, 95, 328. (28) Lombardi, J. R.; Birke, R. L.; Sanchez, L. A.; Bernard, I.; Sun, S. C. Chem. Phys. Lett. 1984, 104, 240. (29) Lombardi, J. R.; Birke, R. L.; Lu, T. H.; Xu, J. J. Chem. Phys. 1986, 84, 4174. (30) Chase, B.; Parkinson, B. J. Phys. Chem. 1991, 95, 7810. (31) Kudelski, A.; Bukowska, J. Chem. Phys. Lett. 1994, 222, 555. (32) Furtak, T. E.; Roy, D. Phys. ReV. Lett. 1983, 50, 1301. (33) Liu, Z.; Yang, Z. L.; Cui, L.; Ren, B.; Tian, Z. Q. J. Phys. Chem. C 2006, 111, 1770. (34) Birke, R. L.; Lombardi, J. R. In Spectroelectrochemistry: Theory and Practice; Gale, R. J., Ed.; Plenum: New York, 1988.
J. Phys. Chem. C, Vol. 114, No. 39, 2010 16595 (35) Wu, D. Y.; Li, J. F.; Ren, B.; Tian, Z. Q. Chem. Soc. ReV. 2008, 37, 1025. (36) Brolo, A. G.; Irish, D. E.; Szymanski, G.; Lipkowski, J. Langmuir 1998, 14, 517. (37) Lundqvist, B. I.; Gunnarsson, O.; Hjelmberg, H.; Norskov, J. K. Surf. Sci. 1979, 89, 196. (38) Gersten, J.; Nitzan, A. J. Chem. Phys. 1980, 73, 3023. (39) Ali, A. H.; Foss, C. A. J. Electrochem. Soc. 1999, 146, 628. (40) Prado, C.; Prieto, F.; Rueda, M.; Feliu, J.; Aldaz, A. Electrochim. Acta 2007, 52, 3168. (41) Stolberg, L.; Lipkowski, J.; Irish, D. E. J. Electroanal. Chem. 1991, 300, 563. (42) Brolo, A. G.; Irish, D. E.; Lipkowski, J. J. Phys. Chem. B 1997, 101, 3906. (43) Cui, L.; Wang, A.; Wu, D. Y.; Ren, B.; Tian, Z. Q. J. Phys. Chem. C 2008, 112, 17618. (44) Xiao, Y. J.; Chen, Y. F.; Gao, X. X. Spectrochim. Acta, Part A 1999, 55, 1209. (45) Thietke, J.; Billmann, J.; Otto, A. Jerusalem Symp. Quantum Chem. Biochem. 1984, 17, 345. (46) Huang, Q. J.; Lin, X. F.; Yang, Z. L.; Hu, J. W.; Tian, Z. Q. J. Electroanal. Chem. 2004, 563, 121. (47) Ren, B.; Lin, X. F.; Yang, Z. L.; Liu, G. K.; Aroca, R. F.; Mao, B. W.; Tian, Z. Q. J. Am. Chem. Soc. 2003, 125, 9598. (48) Doerfer, T.; Schmitt, M.; Popp, J. J. Raman Spectrosc. 2007, 38, 1379. (49) Tian, Z. Q.; Yang, Z. L.; Ren, B.; Wu, D. Y. Top. Appl. Phys. 2006, 103, 125. (50) Cui, L.; Mahajan, S.; Cole, R. M.; Soares, B.; Bartlett, P. N.; Baumberg, J. J.; Hayward, I. P.; Ren, B.; Russell, A. E.; Tian, Z. Q. Phys. Chem. Chem. Phys. 2009, 11, 1023. (51) Lin, X. F.; Ren, B.; Yang, Z. L.; Liu, G. K.; Tian, Z. Q. J. Raman Spectrosc. 2005, 36, 606. (52) Watanabe, H.; Ishida, Y.; Hayazawa, N.; Inouye, Y.; Kawata, S. Phys. ReV. B 2004, 69, 155418. (53) Puppels, G. J.; Otto, C.; Greve, J.; Robert-Nicoud, M.; Arndt-Jovin, D. J.; Jovin, T. M. Biochemistry 1994, 33, 3386. (54) Ren, B.; Lin, X. F.; Yan, J. W.; Mao, B. W.; Tian, Z. Q. J. Phys. Chem. B 2003, 107, 899. (55) Giese, B.; McNaughton, D. J. Phys. Chem. B 2002, 106, 101. (56) Tian, Z. Q.; Ren, B.; Wu, D. Y. J. Phys. Chem. B 2002, 106, 9463. (57) Xie, Y.; Wu, D. Y.; Liu, G. K.; Huang, Z. F.; Ren, B.; Yan, J. W.; Yang, Z. L.; Tian, Z. Q. J. Electroanal. Chem. 2003, 554-555, 417. (58) Rubim, J. C.; Corio, P.; Ribeiro, M. C. C.; Matz, M. J. Phys. Chem. 1995, 99, 15765. (59) Arenas, J. F.; Fernandez, D. J.; Soto, J.; Lopez-Tocon, I.; Otero, J. C. J. Phys. Chem. B 2003, 107, 13143. (60) Lide, D. R., Ed. CRC Handbook of Chemistry and Physics; CRC Press: Boca Raton, FL, 1987. (61) Bard, A. J., Ed. Encyclopedia of Electrochemistry of the Elements; Marcel Dekker Inc: New York, 1982. (62) Smith, E.; Dent, G. Modern Raman Spectroscopy - A Practical Approach; John Wiley & Sons: Chichester, U.K., 2005. (63) Marian, C.; Nolting, D.; Weinkauf, R. Phys. Chem. Chem. Phys. 2005, 7, 3306.
JP1055717