Article pubs.acs.org/JPCC
Iodide-Induced Organothiol Desorption and Photochemical Reaction, Gold Nanoparticle (AuNP) Fusion, and SERS Signal Reduction in Organothiol-Containing AuNP Aggregates Ganganath S. Perera,† Allen LaCour,† Yadong Zhou,‡ Kate L. Henderson,† Shengli Zou,‡ Felio Perez,§ Joseph P. Emerson,† and Dongmao Zhang*,† †
Department of Chemistry, Mississippi State University, Mississippi State, Mississippi 39762, United States Department of Chemistry, University of Central Florida, Orlando, Florida 32816, United States § Integrated Microscopy Center, University of Memphis, Memphis, Tennessee 38152, United States ‡
S Supporting Information *
ABSTRACT: Gold nanoparticles (AuNPs) have been used extensively as surface-enhanced Raman spectroscopic (SERS) substrates for their large SERS enhancements and widely believed chemical stability. Presented is the finding that iodide can rapidly reduce the SERS intensity of the ligands, including organothiols adsorbed on plasmonic AuNPs through both iodide-induced ligand desorption and AuNP fusion. The organothiols trapped inside the fused AuNPs have negligible SERS activities. Multiple photochemical processes were involved when organothiol-containing AuNP aggregates were treated with KI under photoillumination. The photocatalytically produced I3− reacts with both organothiol and AuNPs. Chloride and bromide also induce partial organothiol displacement and the fusion of the as-synthesized AuNPs, but neither of the two halides has detectable effects on the morphology and Raman signals of the organothiol-containing AuNP aggregates. The insight provided in this work should be important for the understanding of interfacial interactions of plasmonic AuNPs and their SERS applications.
■
INTRODUCTION Surface-enhanced Raman spectroscopy (SERS) has been used extensively to study molecular adsorption, desorption, displacement, and reaction on plasmonic gold and silver nanoparticle (AuNP and AgNP) surfaces.1−4 Plasmonic nanoparticles (NPs) can modify Raman signal of molecular adsorbates through three competing processes that include electromagnetic enhancement,5−7 chemical enhancement,8−10 and nanoparticle (NP) inner filter attenuation.11,12 While the electromagnetic enhancement strengthens the Raman signal of the molecules adsorbed onto the NP surfaces, the NP inner filter effect reduces the analyte Raman and SERS signal by absorbing and scattering the incident and Raman photons. The nature and contribution of chemical enhancement are less understood. While it bears the term of “enhancement” in its name by convention, the chemical interactions between the molecular adsorbates and SERS substrate should be able to enhance or reduce the Raman signal of the surface adsorbates depending on its structural perturbations imposed by SERS substrates. Indeed, recent reports showed that the resonance enhancement factor of Rhodamine 6G on nanoparticle is far less than that in water.13 While significant progresses have been made on the mechanistic understanding of SERS signal phenomena and on the SERS substrate fabrications, limited information is available on the postsynthesis morphological change of the SERS substrate and its potential effect on the Raman activity of the molecules on NPs. For example, it has been demonstrated that © XXXX American Chemical Society
iodide can eliminate SERS signal of adventitious molecules adsorbed onto AuNPs by ligand displacement.14 However, the applicability of this observation to general SERS-active molecules is unclear. In addition, the potential impact of iodide-induced AuNP structural modification on the ligand SERS activity has, to our knowledge, not been explored. Such a study is especially relevant for the ligand-containing AuNPs given the literature reports that iodide can induce fusion of assynthesized AuNPs.15−17 Reported herein is our finding that iodide can drastically reduce the SERS signal of the molecules including organothiols adsorbed onto AuNPs. This is because iodide can both induce ligand desorption and extensive AuNP fusion. The model ligands used in this work include adenine, aliphatic, and aromatic organothiols (Figure 1). These model ligands are chosen because (1) they differ significantly in their Raman activities and their binding affinities to AuNPs and (2) most of the ligands have been used in our recent study of NaBH4induced organothiol desorption and desulfurization on AuNPs.1,2 Therefore, this set of ligand molecules not only enables the evaluation of the general effect of iodide treatment on the structure and properties of the ligand-containing AuNPs Received: December 6, 2014 Revised: January 21, 2015
A
DOI: 10.1021/jp512168z J. Phys. Chem. C XXXX, XXX, XXX−XXX
Article
The Journal of Physical Chemistry C
10 min. The total number of time-resolved spectra for each sample ranges from 1000 to 5000. The integration time for each spectral acquisition is 4 s, and the samples were kept inside the sampling cell during the entire ∼3 to ∼15 h experimental period. In contrast, the time-dependent UV−vis measurement was conducted every few hours after the KI addition, and the samples were kept in dark to minimize the light exposure. 6-TG Displacement from Gold Foil. 150 μL of 10 μM 6TG was added to a 1 mL Eppendorf tube that contained the gold foil. The sample left to sit overnight for complete ligand adsorption. The amount of 6-TG adsorbed was quantified by HPLC-UV−vis determination of the amount of 6-TG remains free in solution. After removal of the excess 6-TG, the 6-TGtreated gold foil was immersed overnight in a freshly prepared 0.1 M KI solution. The amount of 6-TG displaced to the supernatant was quantified with HPLC-UV−vis. SERS Spectral Acquisition. As-synthesized AuNPs were mixed with equal volume of 6-TG, 6-MP, 2-MBI, p-MBT, ET, or adenine with specified concentrations. The ligand/AuNP mixtures were incubated overnight and allowed to aggregate. Excess ligands were removed before KI addition. SERS spectra were acquired with the AuNP aggregates deposited onto the stainless steel substrates. The spectral integration time varies from 1 s for the time-resolved SERS measurement to ∼5 min in the static state SERS measurements. The excitation laser wavelength is 632.8 nm, and the laser intensity before entering the sample was 1.3 mW. TEM and SEM Images. TEM measurements were acquired using a JEOL 2100, and all the SEM images were taken using a JEOL 6500F. TEM images were taken with Cu grids covered with a Formvar carbon film at an accelerating voltage of 200 kV. SEM images were acquired on silicon wafers at an accelerating voltage of 5 kV. The AuNP aggregates were washed to remove excess organothiols or KI before deposit onto the silicon wafers for the SEM measurements. XPS Analysis. Four gold foils each having a dimension of ∼0.5 cm × ∼1 cm were used in XPS measurements. One is used as a control for estimation of the sulfur and iodine content in the blank. The other gold foils were used to study sulfur and iodine content after the gold foil was treated with saturated 6TG solution alone, 0.1 M KI alone, and sequential 6-TG and KI treatment. The gold foils were thoroughly washed with 18 MΩ cm−1 Nanopure water and then dried with N2 gas before XPS measurements. XPS spectral acquisition was conducted with a Thermo Scientific K-Alpha XPS system equipped with a monochromatic X-ray source at 1486.6 eV, corresponding to the Al Kα line, with a spot size of 400 μm2. Photoelectrons were collected from a takeoff angle of 90° relative to the sample surface. Measurements were done in the Constant Analyzer Energy mode. The survey spectra were taken at a pass energy of 200 eV, while the high-resolution core level spectra were taken at a 50 eV pass energy, an energy step size of 0.1 eV and using an average of 20 scans. The XPS data acquisition and analysis were performed using the “Avantage v5.932” software provided with the instrument. The XPS analysis was carried out on six representative areas on each of the four gold films. The C 1s, I 3d, S 2p, and Au 4f signal intensities were determined by fitting their respective peaks with a mixture of Lorentzian and Gaussian curves. The relative iodine and sulfur contents in the gold foil control, gold foil treated with 6-TG or KI alone, and gold foil treated
Figure 1. Molecular structures of model ligands.
but also allows us to compare and contrast the NaBH4 versus iodide ligand desorption from the AuNPs.
■
EXPERIMENTAL SECTION Materials and Equipment. All chemicals and the gold foil (2.5 cm × 2.5 cm × 0.127 mm) were purchased from SigmaAldrich and used as received. UV−vis spectra were taken using an Olis HP 8452 diode array spectrophotometer. The SERS spectra were acquired using the LabRam ARAMIS confocal Raman microscope system with a 633 nm HeNe Raman excitation laser. HPLC separation, detection, and quantification of ligand desorption were performed using a Dionex 3000 series HPLC with a UV−vis detector in series with a Brüker micro-ToFQII ESI-MS in positive mode. Separation of KI from the ligands was performed using a reversed-phase Dionex C18 column with a mobile phase of 5% acetic acid:acetonitrile (94:6 v/v). Data analysis was performed using Brüker HystarPP and DataAnalysis program software associated with the Brüker MS system. AuNP Synthesis. AuNPs were prepared using the citratereduction method.18 In brief, HAuCl4·3H2O (0.0415 g) was added to 100 mL of 18 MΩ cm−1 Nanopure water, and the solution was brought to boil. Then 10 mL of 1% trisodium citrate dehydrate was added, and the mixture was kept boiling for ∼20 min while stirring. The average diameter of the AuNPs was determined to be ∼13 nm. Ligand Displacement from AuNPs. Known concentrations of model ligands were mixed with equal volume of assynthesized AuNPs, and the ligand/AuNP mixtures were incubated overnight to allow the AuNP aggregates to settle to the bottom of the sample cuvettes. Adsorbed ligands were quantified, and the excess ligands were removed by washing the AuNP aggregates with 18 MΩ cm−1 Nanopure water. Upon addition of freshly prepared KI, the (ligand/AuNP)/KI mixture was kept in dark to prevent KI photoconversion to I2.19−21 The amount of ligand desorbed was determined periodically with either HPLC-UV−vis (adenine) or direct UV−vis quantification (2-MBI, 6-MP, and 6-TG) with the supernatant of the ligand desorption solution at a predefined time intervals. KI concentration dependence on ligand desorption was measured for 6-TG, 6-MP, 2-MBI, and adenine. Time-resolved SERS measurements were conducted for all the model ligands before and after KI addition. Time-Resolved UV−Vis Acquisition. The effect of light exposure on the KI-induced organothiol desorption, and photochemical reactions were studied using time-resolved UV−vis measurements. The excitation source used in the time-resolved UV−vis measurements is a 30 W deuterium lamp. The time interval between two consecutive spectra is 1− B
DOI: 10.1021/jp512168z J. Phys. Chem. C XXXX, XXX, XXX−XXX
Article
The Journal of Physical Chemistry C sequentially with 6-TG and KI were estimated by using gold signal intensity as the internal reference. ICP-MS Quantification of 6-TG in KI-Treated AuNP Aggregates. After UV−vis quantification of the 6-TG displaced by iodide from the AuNPs, the (AuNPs/6-TG)/KI mixture was split into two equal volume fractions. Only the bottom fraction contained AuNP aggregates. An equal amount of freshly prepared aqua regia solution was added into both top and bottom fractions of the ligand displacement solution to digest the AuNPs and convert the 6-TG to sulfonate. The sulfur content in the top and bottom fractions was then quantified with a PerkinElmer ELAN DRC II ICP-MS instrument equipped with a dynamic reaction cell. The difference in the sulfur content in the top and bottom fraction represents the amount of undisplaced 6-TG that likely trapped inside the fused AuNPs. Computational Simulation. Computational simulations were conducted to determine the SERS enhancement of molecules adsorbed on and trapped inside the fused AuNPs. The SERS enhancement factors of the molecules adsorbed on the surface of aggregated AuNPs were calculated using a similar strategy discussed in the previous paper.11 The coordinates of aggregated particles were generated by Monte Carlo method; the enhanced electric fields at both the excitation and Raman photon wavelengths outside the particle surface were calculated using the T-matrix method.22 Molecules were randomly arranged on metal nanoparticle surfaces using the Monte Carlo method, and the calculations were repeated over 1 million times to reduce the oscillation due to the randomness in the Monte Carlo method. For molecules embedded inside the fused particles, the electric fields inside the particle at both the excitation and Raman photon wavelengths were calculated using the generalized Mie theory.23 The enhancement factor of the Raman signal was calculated by multiplying the enhancement factor of the electric field |E|2 at both the excitation and Raman photon wavelengths.
Figure 2. Representative time-resolved or time-dependent SERS spectra of (A) (AuNPs/6-TG), (B) (AuNPs/6-MP), (C) (AuNPs/2MBI), (D) (AuNPs/p-MBT), (E) (AuNPs/ET), and (F) (AuNPs/ adenine).The nominal concentrations of AuNPs, 6-TG, 6-MP, 2-MBI, p-MBT, ET, adenine, and KI are 10 nM, 92 μM, 90 μM, 85 μM, 94 μM, 84 μM, 92 μM, and 1 M, respectively.
Figure 3. UV−vis detection of iodide-induced ligand desorption from the ligand-containing AuNP aggregates. The compositions of the samples in the figure are the same as that used in corresponding samples in Figure 2. (A) Time-dependent UV−vis spectra taken before and after 1 M KI addition into 6-TG-containing AuNP aggregates. The photograph (inset I) shows that AuNPs are fully aggregated and precipitated due to the ligand adsorption. Inset II is the time course of percentage 6-TG displaced. (B) KI concentration dependence of the percentage ligand desorption from AuNP aggregates. The time-dependent UV−vis spectra and the time course of the 6-MP and 2-MBI desorption are shown in Figure S1 of the Supporting Information.
■
RESULTS AND DISCUSSION Time-dependent SERS spectra (Figure 2) show that ligand SERS intensities decrease immediately upon the KI addition into ligand-containing AuNPs. The speed of the disappearance of the ligand SERS signal depends critically on the ligand structure. It takes ∼30 min for iodide to completely remove the 2-MBI SERS feature (Figure 2C), while no detectable adenine SERS feature within the first 5 s of the iodide addition (Figure 2F), the instrument dead time for our time-resolved SERS spectral acquisition. Iodide eliminating the SERS signal of chemicals adsorbed onto AuNPs has been reported before, and it was attributed to the complete ligand displacement.14 However, our ligand desorption experiments (Figure 3), conducted with the samples identical to that used in the time-resolved SERS acquisition (Figure 2), showed that only can adenine be completely displaced from the AuNP surfaces. The organothiol ligands have only been partially displaced from the ligand-containing AuNP aggregates despite of the complete absence of ligand SERS signal in the KI-treated samples (Figure 2). The highest percentage of organothiols that can be displaced from the AuNPs is below 60% for all three organothiols tested with the time-dependent UV−vis measurements (Figure 3B). This result indicates that the absence of the ligand SERS signal should not be taken as a marker of complete ligand desorption.
It is instructive to compare and contrast the KI- and NaBH4induced organothiol desorption. Kinetically, the ligand desorption rates induced by both iodide and NaBH4 are very fast. Maximum ligand desorption was observed within a few minutes of the KI and NaBH4 addition. Furthermore, both reagents can completely eliminate the SERS signal of the adsorbed organothiols. However, there are sharp differences between the KI- and NaBH4-induced organothiol desorptions. First, no organothiol can be readsorbed onto AuNPs in the iodide-induced organothiol desorption. The UV−vis intensity of the desorbed ligand remains constant in the supernatant of the KI-treated AuNP mixtures even with prolonged sample incubation (inset II in Figure 3A). This is in contrast to the NaBH4-induced organothiol desorption in which the ligand can be entirely readsorbed onto AuNPs. The latter is because of the poor stability of hydride in aqueous solution.1 Second, iodide can only cause partial organothiol displacement even when the C
DOI: 10.1021/jp512168z J. Phys. Chem. C XXXX, XXX, XXX−XXX
Article
The Journal of Physical Chemistry C KI concentration is as high as 1 M, but 0.1 M NaBH4 can induce a complete organothiol displacement.1 This result indicates that hydride is more effective than iodide in displacing organothiols, which may be indicative of that hydride has higher binding affinity to AuNPs than iodide. Third, the SERS activity of the NaBH4-treated organothiol-containing AuNPs can be fully recovered after organothiol readsorption,1 while no SERS signal can be observed in the KI-treated AuNP aggregates even though the organothiol on AuNPs is only partially displaced. The data in Figures 2A−C and 3B showed that there are at least 40% of the 6-TG, 6-MP, and 2-MBI initially adsorbed onto the AuNP aggregates escaped both UV−vis and SERS detections in the KI-treated samples. Several possible scenarios could be imagined for these unaccounted organothiols. First is that all the organothiols on AuNPs are displaced by iodide into the AuNP supernatant thereby not detected by SERS, but some of the displaced organothiols are degraded into chemicals that are not UV−vis active. Second, the unaccounted organothiols remained adsorbed onto the iodide-treated AuNP aggregates, but they are not SERS active. The first scenario is excluded on the basis of the control experiments that show these model organothiols are totally stable in KI containing liquid under the experimental conditions used for the ligand displacement study (Figure S2, Supporting Information). This indicates that the unaccounted organothiols are remained in the KI-treated AuNP aggregates. Indeed, ICP-MS measurement confirms that the amount of sulfur remains in the iodide-treated organothiolcontaining AuNP aggregates is approximately equal to the quantity difference between the organothiol initially adsorbed onto AuNPs and that displaced by iodide into the supernatant. The most likely reason for the absence of the Raman signal of the organothiols remained attached to the iodide-treated AuNP aggregates is the AuNP fusion (Figure 4). Without iodide, the organothiol-containing AuNP aggregates remain black for at least six months in solution. The grain size of the organothiol-containing AuNPs is very similar to that for the assynthesized AuNPs (Figure 4B,D). However, upon overnight incubation with KI, the colors of both as-synthesized and organothiol-containing AuNPs turn into golden color (Figure 4A). SEM measurements reveal that there are extensive AuNP fusion in the KI-treated AuNP aggregates. The individual fused AuNPs as large as 100 nm in diameter can be seen in the KItreated AuNP aggregates (Figure 4C,E and Figure S3, Supporting Information). The organothiol remained in the iodide-treated AuNP aggregates should be mostly trapped inside the fused AuNPs, but not on the NP surfaces accessible to iodide. Our HPLCUV−vis experiment shows that 6-TG adsorbed onto planar gold film can be completely displaced by iodide (Figure S4, Supporting Information). X-ray photoelectron spectroscopic (XPS) analysis confirms that sulfur on 6-TG treated planar AuNP film is almost completely displaced by iodide after overnight immersion of 6-TG treated gold film in 0.1 M KI solution (Figure S5, Supporting Information). This result indicates that undisplaced organothiols on the fused AuNPs should locate only in areas inaccessible to iodide. Consistent with this conclusion is the computational simulation, which strongly suggests that these undisplaced organothiols should be trapped inside, but not on, the surface of the fused AuNPs. Otherwise, the SERS activity of the remaining organothiols should be higher than that for the initially adsorbed organothiols. Indeed, the computed SERS
Figure 4. (A) Photographs of ligand/AuNP mixtures (top row) without and (bottom row) with KI treatment. Vial (a) is the AuNP control, and the ligands in vials (b) to (g) are 6-TG, 6-MP, 2-MBI, pMBT, ET, and adenine, respectively. The nominal concentrations of AuNPs, 6-TG, 6-MP, 2-MBI, ET, p-MBT, adenine, and KI are 10 nM, 92 μM, 90 μM, 85 μM, 90 μM, 84 μM, 92 μM, and 1 M. (B) TEM image of the as-synthesized AuNPs; the scale bar represents 25 nm. (C) Representative SEM images of AuNP aggregates formed by overnight incubation of colloidal AuNP with 1 M KI. (D, E) Representative SEM images obtained with organothiol-containing AuNPs without and with overnight KI incubation, respectively. The scale bars in the SEM images are 100 nm. Details of sample preparation and SEM data acquisition are shown in the Experimental Section.
enhancement factor increases from 2.4 × 105 to 4.7 × 105 when the average AuNP diameter changes from 13 to 50 nm. Apparently, if the undisplaced organothiol molecular adsorbates remain on the surface of the fused AuNPs, iodide treatment should not have significant effect on the organothiol Raman signal. This is because the AuNP fusion from 13 to 50 nm almost doubles SERS enhancement factor, adequate to compensate for the lost Raman contribution due to the partially displaced organothiols initially attached to the 13 nm AuNPs. However, if the molecules are located inside the AuNPs, the AuNP quenches instead of enhancing the Raman signal of the trapped molecules. Figure 5 shows the calculated attenuation factors qd,x and qd,s for the electrical field |E|2 at the excitation and Raman photon wavelengths, respectively, and Raman enhancement factor (gd = qd,x × qd,s) experienced by a molecule sitting inside a 50 nm AuNP. The Raman enhancement factor increases monotonically when the molecule moves from the AuNP center toward the AuNP surface. However, the values of these Raman enhancement factors are all smaller than 1 as long as the probe molecule is inside the AuNP, indicating that AuNP quenches Raman signal of the trapped molecules. Mechanistically, this quenching can be understood by the optical damping imposed by the metallic overlayer on the incident excitation and Raman photons (Figure 5). The presence of KI unvariably reduces the ligand SERS intensities, even though completely eliminate the organothiol SERS signal requires KI in large excess (Figure 6A,C). It is noted that unlike the experimental data shown in Figures 2−4 where all the excess ligands were removed before KI addition, the excess adenine and 6-TG were kept in the samples used in D
DOI: 10.1021/jp512168z J. Phys. Chem. C XXXX, XXX, XXX−XXX
Article
The Journal of Physical Chemistry C
effect on the Raman signal of the organothiols remaining attached onto AuNP aggregates (Figure 7B). Chloride, bromide, and iodide inducing the fusion of as-synthesized AuNPs have been reported before.15−17 However, the effect of these halides on the organothiol-containing AuNPs has, to our knowledge, not been available in the literature. The fact that all three halides can induce fusion of the as-synthesized AuNPs, but only iodide induced significant fusion of the organothiolcontaining AuNPs, strongly suggests that organothiols have to be mostly removed before that significant AuNP fusion occurs. The observed drastic difference among these halides in inducing the organothiol desorption and AuNP fusion in the organothiol-containing AuNPs is consistent with reports that bonding energy of gold with iodide is stronger than that with bromide and chloride.24,25 Light exposure has a drastic effect on the KI interactions with the organothiol-containing AuNP aggregations (Figure 8 and Figure S6, Supporting Information). An immediate 6-TG desorption was observed upon the KI addition (Figure 8A). However, unlike the desorbed 6-TG that is stable in solution for at least several days in the time-dependent UV−vis measurements in Figure 3A, the peak UV−vis intensity of the iodide-treated 6-TG-containing aggregates monotonically decreases and the peak absorption wavelength monotonically blue-shifted as a function of spectral acquisition time until they both reaches a constant (Figure 8A). This result indicates the light exposure from the time-resolved UV−vis measurement is strong enough to trigger photochemical reactions in the AuNP/ organothiol/KI mixtures. It is noted that light exposure in the time-dependent and time-resolved UV−vis measurements are drastically different, despite the fact that these spectra were taken with the same instrument. Fewer than 10 UV−vis measurements were taken in the time-dependent UV−vis study, while 1000−5000 UV−vis measurements were conducted in the time-resolved UV−vis study for the effect of KI on the organothiol-containing AuNP aggregates. There are most likely multiple chemical and photochemical processes involved in the photoilluminated organothiol/AuNP/ KI mixtures. Upon light illumination, the I− can photocatalytically converted to I2 that subsequently combines with I− forming I3−.19−21 The production of I3− under the timeresolved UV−vis measurement conditions was confirmed with a control experiment. The latter shows that the characteristic I3− UV−vis peak at 352 nm monotonically increases as timeresolved UV−vis acquisition proceeds for the KI solution (Figure 8B). Figure 8C shows that 6-TG/KI mixture is not stable under photoillumination either. This is because the 6-TG UV−vis absorption monotonically decreases. Since control experiments showed that all these model ligands are stable under the time-resolved UV−vis measurement conditions (Figure S7, Supporting Information), the most reasonable explanation to the photochemical reactions observed in 6-TG/ KI, 6-MP/KI, and 2-MBI/KI solutions is a two-step reaction mechanism. KI is first photochemically converted to I3−, and the latter subsequently reacts with the organothiols, producing reaction products that has different UV−vis feature as that for intact organothiols. The time-resolved UV−vis spectra in Figure 8A, inset II, are similar to the time-resolved UV−vis spectra obtained with 6TG/I2 solution (Figure 8D), but not that with 6-TG/KI solution (Figure 8C). This is evident from the fact that the UV−vis blue-shift was observed only in the spectra obtained with 6-TG/I2 and AuNP/6-TG/KI solutions, but not in the 6-
Figure 5. Computationally modeled electric field |E|2 damping factor qd,x, qd,s at the excitation and Raman photon wavelengths of 632.8 and 670 nm and the Raman enhancement factor gd = qd,x × qd,s for molecules encapsulated inside an AuNP of 50 nm in diameter. d represents the distance of the molecule from the AuNP center.
Figure 6. SERS spectra obtained with a series of (A) (AuNP/ adenine)/KI and (C) (AuNP/6-TG)/KI mixtures in which the two components in the parentheses were mixed overnight before addition of KI of different concentrations. (B) and (D) are the adenine and 6TG SERS intensity as a function of the KI concentrations. The nominal AuNP and ligand concentrations were 4.3 nM and 16.7 μM, respectively, in all the samples. The concentrations of KI are 0, 0.05, 0.5, 5, 50, and 500 mM for spectra (a) to (f) in parts A and C, respectively. SERS spectra were taken after overnight of the KI addition. The dashed lines in (B) and (D) represent the SERS intensity of adenine and 6-TG controls.
Figures 6A and 6C, respectively. The threshold KI concentration for complete SERS signal reduction (>99%) in Figure 6B,D samples is 5 mM for adenine and 500 mM for 6-TG. This experimental obervation is consistent with the time-resolved SERS study shown in Figure 2F, and it provides further evidence that adenine can be more readily displaced by iodide. Excess chloride and bromide can also partially displace 6-TG from the AuNP surfaces. However, the amount of the 6-TG can be displaced by chloride and bromide is 25 and 16 times, respectively, smaller than that of iodide (Figure 7C). Moreover, both chloride and bromide can induce extensive fusion of the as-synthesized AuNPs (Figure 7A), but neither chloride nor bromide induces significant AuNP fusion in the organothiolcontaining AuNP aggregates (Figure 7A) or has a detectable E
DOI: 10.1021/jp512168z J. Phys. Chem. C XXXX, XXX, XXX−XXX
Article
The Journal of Physical Chemistry C
Figure 7. Comparison of the 6-TG desorption from the aggregated AuNPs induced by Cl−, Br−, and I−. (A) Photograph of the (top) as-synthesized AuNP mixed with (a) KCl, (b) KBr, and (c) KI, (bottom) 6-TG-containing AuNPs after overnight incubation with (a) KCl, (b) KBr, and (c) KI. (B) SERS spectra obtained with (a) 6-TG-containing AuNP aggregates and after treated with (b) KCl, (c) KBr, and (d) KI. (C) UV−vis spectra of the 6-TG displaced by (a) KCl, (b) KBr, and (c) KI. Spectrum (d) is the 6-TG-containing AuNP control that shows no 6-TG in the supernatant without halide treatment. The nominal concentrations of AuNPs, KCl, KBr, KI, and 6-TG are 10 nM, 1 M, 1 M, 1 M, and 80 μM, respectively.
Attempts to characterize the chemical compositions of the photochemical reaction products in the (organothiol/AuNP)/ KI samples after their time-resolved UV−vis measurements were unsuccessful. Preliminary mass spectrometric measurements indicate that the chemical composition is highly complicated. This result is not surprising giving the complexity of the photochemical processes. Besides the aforementioned organothiol reaction with photochemically produced I3−, other factors can also contribute to the UV−vis spectral change observed in the time-resolved UV−vis samples. One possible factor is that the desorbed organothiol chelates the gold ion etched off the AuNP by I3−. Such chelating can not only modify the UV−vis spectral feature of both organothiol and the gold ion.
■
CONCLUSIONS In summary, iodide can completely displace organothiols from the planar gold surface, but the iodide-induced organothiol displacement from AuNPs is complicated by the iodide-induced AuNP fusion. Only can up to 60% of the organothiols be displaced from the AuNPs, and the remaining organothiols likely trapped inside the fused AuNPs with essentially no detectable Raman signal. Chloride and bromide can also induce partial organothiol displacement but have no appreciable effect on the SERS activity of the remaining organothiols. Light exposure further complicates the KI-induced organothiol desorption and AuNP fusion process. Photochemically generated I3− can react both with the desorbed organothiols and with AuNPs, converting organothiols to chemicals less UV−vis active material. The insight provided in this work should be important for understanding the chemical and photochemical ligand interactions with plamonic AuNPs and their SERS applications.
Figure 8. Time-resolved UV−vis study of the 6-TG desorption and reaction under light exposure. (A) The time course of the peak UV− vis absorbance in the time-resolved UV−vis spectra obtained after KI addition into 6-TG-containing AuNP aggregates. Inset I shows the representative UV−vis spectra corresponding the time course in which the supernatant UV−vis absorbance monotonically increases. Inset II shows the representative time-resolved UV−vis spectra in which the peak UV−vis decrease and the peak wavelength blue-shifted. (B) Time-resolved UV−vis spectra of KI solution. The appearance of peak at 350 nm indicates the photoactivated I3− production. (C) Timeresolved UV−vis spectra of KI and 6-TG mixture. The spectral change is likely due to combination of I2 production in combined with chemical reaction between I2 and 6-TG. (D) Time-resolved UV−vis spectra of 6-TG in I2 solution. The inset shows the blue-shift and the reduction of the UV−vis peak of 6-TG possibly due to the chemical reaction between I2 and 6-TG. Similar results were observed with 2MBI and 6-MP (Figure S6, Supporting Information). The nominal concentrations of AuNPs, 6-TG, 6-MP, 2-MBI, KI, and I2 are 10 nM, 92 μM, 90 μM, 85 μM, 1 M, and 0.5 M, respectively.
■
ASSOCIATED CONTENT
S Supporting Information *
Time-dependent UV−vis spectra and time course of the KIinduced 6-MP and 2-MBI desorption from AuNPs, HPLC-MS and UV−vis characterization of KI-induced ligand displacement from AuNPs, high-resolution SEM image of KI-induced fused AuNPs, quantification of the desorbed amount of 6-TG from gold foil by HPLC-UV−vis analysis, XPS analysis of 6-TG adsorption onto gold foil, time-resolved UV−vis spectral study of 6-MP and 2-MBI desorption and reactions under light illumination, time-resolved UV−vis spectra of 6-TG, 6-MP, and 2-MBI. This material is available free of charge via the Internet at http://pubs.acs.org.
TG/KI solution. This observation indicates the presence of AuNP disrupted the photochemical production of I3− or the I3− reaction with 6-TG. The latter is not surprising since I3− is a known etchant of gold.26−28 I3− reaction with 6-TG displaced to solution can be complicated with the completing I3− reaction with AuNPs and even surface-adsorbed 6-TGs. Unfortunately, it is currently impossible to decipher the interplay of the I3− production and I3− reactions with the organothiols and AuNPs. F
DOI: 10.1021/jp512168z J. Phys. Chem. C XXXX, XXX, XXX−XXX
Article
The Journal of Physical Chemistry C
■
(15) Zhang, Z.; Li, H.; Zhang, F.; Wu, Y.; Guo, Z.; Zhou, L.; Li, J. Investigation of Halide-Induced Aggregation of Au Nanoparticles into Spongelike Gold. Langmuir 2014, 30, 2648−2659. (16) Liu, Y.; Liu, L.; Guo, R. Br-Induced Facile Fabrication of Spongelike Gold/Amino Acid Nanocomposites and Their Applications in Surface-Enhanced Raman Scattering. Langmuir 2010, 26, 13479−13485. (17) Cheng, W.; Dong, S.; Wang, E. Iodine-Induced GoldNanoparticle Fusion/Fragmentation/Aggregation and Iodine-Linked Nanostructured Assemblies on a Glass Substrate. Angew. Chem., Int. Ed. 2003, 42, 449−452. (18) Frens, G. Controlled Nucleation for the Regulation of the Particle Size in Monodisperse Gold Suspensions. Nature 1973, 241, 20−22. (19) Boschloo, G.; Hagfeldt, A. Characteristics of the Iodide/ Triiodide Redox Mediator in Dye-Sensitized Solar Cells. Acc. Chem. Res. 2009, 42, 1819−1826. (20) Gardner, J. M.; Abrahamsson, M.; Farnum, B. H.; Meyer, G. J. Visible Light Generation of Iodine Atoms and I−I Bonds: Sensitized I− Oxidation and I3− Photodissociation. J. Am. Chem. Soc. 2009, 131, 16206−16214. (21) Rowley, J. G.; Farnum, B. H.; Ardo, S.; Meyer, G. J. Iodide Chemistry in Dye-Sensitized Solar Cells: Making and Breaking I−I Bonds for Solar Energy Conversion. J. Phys. Chem. Lett. 2010, 1, 3132−3140. (22) Mackowski, D. W. Calculation of Total Cross Sections of Multiple-Sphere Clusters. J. Opt. Soc. Am. A 1994, 11, 2851−2861. (23) Bohren, C. F.; Huffman, D. R. Absorption and Scattering of Light by Small Particles; John Wiley & Sons Inc.: New York, 1983. (24) Brown, J. R.; Schwerdtfeger, P.; Schröder, D.; Schwarz, H. Experimental and Theoretical Studies of Diatomic Gold Halides. J. Am. Soc. Mass Spectrom. 2002, 13, 485−492. (25) Dubois, L. H.; Nuzzo, R. G. Synthesis, Structure, and Properties of Model Organic Surfaces. Annu. Rev. Phys. Chem. 1992, 43, 437−463. (26) Davis, A.; Tran, T.; Young, D. R. Solution Chemistry of Iodide Leaching of Gold. Hydrometallurgy 1993, 32, 143−159. (27) Pal, T.; Jana, N. R.; Sau, T. K. Nucleophile Induced Dissolution of Gold. Corros. Sci. 1997, 39, 981−986. (28) Wang, H.-x.; Sun, C.-b.; Li, S.-y.; Fu, P.-f.; Song, Y.-g.; Li, L.; Xie, W.-q. Study on Gold Concentrate Leaching by Iodine-Iodide. Int. J. Miner. Metall. Mater. 2013, 20, 323−328.
AUTHOR INFORMATION
Corresponding Author
*E-mail:
[email protected] (D.Z.). Notes
The authors declare no competing financial interest.
■
ACKNOWLEDGMENTS This work was supported by an NSF CAREER Award (CHE 1151057), NSF Fund (EPS-0903787), and a seed grant provided to D.Z. from Agricultural Research Service, U.S. Department of Agriculture, under Project No. 5864022729. S.Z. thanks the National Science Foundation (NSF ECCS1238738) and the Office of Naval Research (ONR N00014-01-1118) for the support of the research.
■
REFERENCES
(1) Ansar, S. M.; Ameer, F. S.; Hu, W.; Zou, S.; Pittman, C. U.; Zhang, D. Removal of Molecular Adsorbates on Gold Nanoparticles Using Sodium Borohydride in Water. Nano Lett. 2013, 13, 1226− 1229. (2) Ansar, S. M.; Perera, G. S.; Ameer, F. S.; Zou, S.; Pittman, C. U.; Zhang, D. Desulfurization of Mercaptobenzimidazole and Thioguanine on Gold Nanoparticles Using Sodium Borohydride in Water at Room Temperature. J. Phys. Chem. C 2013, 117, 13722−13729. (3) Perera, G. S.; Ansar, S. M.; Hu, S.; Chen, M.; Zou, S.; Pittman, C. U.; Zhang, D. Ligand Desorption and Desulfurization on Silver Nanoparticles Using Sodium Borohydride in Water. J. Phys. Chem. C 2014, 118, 10509−10518. (4) Schlücker, S. Surface-Enhanced Raman Spectroscopy: Concepts and Chemical Applications. Angew. Chem., Int. Ed. 2014, 53, 4756− 4795. (5) Campion, A.; Kambhampati, P. Surface-Enhanced Raman Scattering. Chem. Soc. Rev. 1998, 27, 241−250. (6) Mock, J. J.; Norton, S. M.; Chen, S. Y.; Lazarides, A. A.; Smith, D. R. Electromagnetic Enhancement Effect Caused by Aggregation on SERS-Active Gold Nanoparticles. Plasmonics 2011, 6, 113−124. (7) Ausman, L. K.; Li, S.; Schatz, G. C. Structural Effects in the Electromagnetic Enhancement Mechanism of Surface-Enhanced Raman Scattering: Dipole Reradiation and Rectangular Symmetry Effects for Nanoparticle Arrays. J. Phys. Chem. C 2012, 116, 17318− 17327. (8) Moskovits, M. Persistent Misconceptions Regarding SERS. Phys. Chem. Chem. Phys. 2013, 15, 5301−5311. (9) Maitani, M. M.; Ohlberg, D. A. A.; Li, Z.; Allara, D. L.; Stewart, D. R.; Williams, R. S. Study of SERS Chemical Enhancement Factors Using Buffer Layer Assisted Growth of Metal Nanoparticles on SelfAssembled Monolayers. J. Am. Chem. Soc. 2009, 131, 6310−6311. (10) Doering, W. E.; Nie, S. Single-Molecule and Single-Nanoparticle SERS: Examining the Roles of Surface Active Sites and Chemical Enhancement. J. Phys. Chem. B 2001, 106, 311−317. (11) Ameer, F. S.; Ansar, S. M.; Hu, W.; Zou, S.; Zhang, D. Inner Filter Effect on Surface Enhanced Raman Spectroscopic Measurement. Anal. Chem. 2012, 84, 8437−8441. (12) Kloz, M.; Grondelle, R. v.; Kennis, J. T. M. Correction for the Time Dependent Inner Filter Effect Caused by Transient Absorption in Femtosecond Stimulated Raman Experiment. Chem. Phys. Lett. 2012, 544, 94−101. (13) Ameer, F. S.; Pittman, C. U.; Zhang, D. Quantification of Resonance Raman Enhancement Factors for Rhodamine 6G (R6G) in Water and on Gold and Silver Nanoparticles: Implications for SingleMolecule R6G SERS. J. Phys. Chem. C 2013, 117, 27096−27104. (14) Li, M.-D.; Cui, Y.; Gao, M.-X.; Luo, J.; Ren, B.; Tian, Z.-Q. Clean Substrates Prepared by Chemical Adsorption of Iodide Followed by Electrochemical Oxidation for Surface-Enhanced Raman Spectroscopic Study of Cell Membrane. Anal. Chem. 2008, 80, 5118− 5125. G
DOI: 10.1021/jp512168z J. Phys. Chem. C XXXX, XXX, XXX−XXX