Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy 136 (2015) 1593–1599
Contents lists available at ScienceDirect
Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy journal homepage: www.elsevier.com/locate/saa
Surface-enhanced Raman scattering of perchlorate on cationic-modified silver nanofilms – Effect of inorganic anions Jumin Hao a,b, Mei-Juan Han a,⇑, Xiaoguang Meng a, Wayne Weimer b, Qingwu K. Wang b b
Center for Environmental Systems, Stevens Institute of Technology, Hoboken, NJ 07030, USA Agiltron Inc., 15 Presidential Way, Woburn, MA 01801, USA
h i g h l i g h t s
g r a p h i c a l a b s t r a c t
Cationic thiol modified SERS
a r t i c l e
i n f o
Article history: Received 9 July 2014 Received in revised form 12 October 2014 Accepted 16 October 2014 Available online 24 October 2014 Keywords: Surface-enhanced Raman scattering (SERS) Perchlorate detection Inorganic anions Cationic thiol Silver nanofilm Competitive adsorption
930 cm-1 7000
5000 4000
100000
SERS spectra
6000
Int
substrates were fabricated for selective detection of ClO4. Effects of common anions on ClO4 SERS spectra were investigated and analyzed by modeling. Competitive interaction mechanisms of the anions with the SERS substrate were discussed. Quantitative SERS detection of ClO4 in the presence of common anions was demonstrated. The study will help developing a robust SERS substrate for aqueous ClO4 detection in matrix.
Calibration curve
ClO 4- peak height
a
1042 cm-1 975 cm-1
3000 2000 1000
10000
1000
100
1200
1000
10
800
100
1000
10000
ClO4- concentration (uM)
Raman shift / cm-1
a b s t r a c t Surface-enhanced Raman scattering (SERS) has emerged as one of the most sensitive spectroscopic analysis methods for the detection of environmental contaminants in water, including perchlorate (ClO 4 ). However, as with other commonly used analytical techniques, analysis of realistic environmental samples by SERS presents a challenge due to complex chemical components coexisting in the samples. In this work, we investigated the influence of inorganic anions (particularly oxyanions) on SERS spectra of ClO 4 using a cationic thiol modified silver nanofilm substrate (Cys–Ag/rCu). The results show that the anions present in the samples did not shift the ClO4 characteristic band positions, but did decrease signal intensities due to their competitive binding with the –NH+3 groups of cationic thiol molecules immobilized on the substrates. The pH changes caused by both the dissociation of H2PO 4 and the hydrolysis of HCO3 may also play a non-negligible role. The selectivity of the Cys–Ag/rCu substrate towards these anions was 2 determined to be in the following order: ClO 4 > SO4 > HCO3 , NO3 > Cl > H2PO4 , indicating preferential adsorption of ClO 4 ions. In the solutions with multiple anions present, the ClO4 SERS spectra were affected simultaneously by all the coexisting anions. Calibration curves with very good linear relationships were successfully obtained, demonstrating the great potential of quantitative detection of aqueous ClO 4 in the matrix. Ó 2014 Elsevier B.V. All rights reserved.
Introduction
⇑ Corresponding author. Tel.: +1 201 216 8013; fax: +1 201 216 8303. E-mail address: [email protected] (M.-J. Han). http://dx.doi.org/10.1016/j.saa.2014.10.052 1386-1425/Ó 2014 Elsevier B.V. All rights reserved.
Perchlorate (ClO 4 ) is a widely used component of solid fuel missile and rocket propellants, explosives, and pyrotechnics [1]. It has recently emerged as a widespread environmental contaminant
1594
J. Hao et al. / Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy 136 (2015) 1593–1599
found in foods, milks, drinking water, groundwater, and soils in many countries including the United States [1–4]. ClO 4 is believed to disrupt the thyroid function by inhibiting the uptake of iodide, affecting the production of thyroid hormone and possibly causing mental retardation in fetuses and infants [4,5]. The U.S. Environmental Protection Agency (EPA) is developing a proposed national primary drinking water regulation for ClO 4 , and has established an Interim Drinking Water Health Advisory limit of 15 lg/L (ppb) [6]. This concentration limit will definitely have a significant impact on the management and remediation of perchlorate-contaminated sites and public water systems, and will require technological innovations for sensitive, rapid, and inexpensive field analyzers. Currently used techniques for ClO 4 analyses include mainly ion chromatography (IC), electrospray ionization mass spectrometry (ESI-MS), IC–MS and ion-selective electrodes (ISE) [7]. However, these methods either require large, expensive equipment, timeconsuming and complex sample preparation, well-trained users, or are associated with issues of sensitivity and selectivity, making them not suitable for field applications. Therefore, there is a critical need for the development of a rapid, simple and cost-effective portable chemical analyzer for on-site identification and quantification of trace-level ClO 4 in environmental water. Surface-enhanced Raman scattering (SERS), as one of the most sensitive spectroscopic analysis methods, has been extensively investigated for chemical and biological sensing [8–17] including the possibility of detecting single molecules [18–20]. In particular, the combination of the SERS technique with commercially available portable/handheld Raman systems has shown great potential to meet the criteria of field assays of environmental pollutants including ClO 4 ions [21–26]. It is well known that only the molecules which adsorb chemically or physically onto nanostructured substrates of noble metals (mostly Ag or Au) produce a significant SERS effect. The SERS detection of trace level ClO 4 in water appears to be very difficult due to its high water solubility and extremely weak sorption tendency on these metals [27,28]. To solve this problem, two approaches have been used. One is to measure SERS spectra using samples air-dried on planar SERS substrates to deposit perchlorate onto nanostructured Ag or Au surfaces [22,29]. Another approach has been developed to chemically modify the SERS substrates, trapping and concentrating ClO 4 anions onto Ag or Au surfaces [7,30–36]. The modification reagents commonly bear positively-charged groups such as –NH+3, –NH(CH3)+2 and – N(CH3)+3. For example, we recently developed a cysteamine hydrochloride (Cys) modified Ag nanofilm deposited on a roughened Cu foil (Cys–Ag/rCu) for rapid and sensitive SERS analysis of aqueous ClO 4 without the need for drying. The Cys (HSCH2CH2NH2 HCl) is a cationic thiol. The Cys molecules formed a self-assembled monolayer (SAM) on the surface of the Ag/rCu nanofilm, and their protonated amino groups –NH+3 make the substrate surface positively charged, which promotes the adsorption and enrichments of the negatively charged ClO 4 ions. Thus, the SERS signal of ClO4 can be enhanced dramatically, enabling sensitive SERS measurement to be accomplished immediately once the aqueous ClO 4 samples are applied onto the Cys–Ag/rCu film. However, as with other commonly used techniques, analysis of realistic environmental samples by SERS presents a challenge because of complex chemical components coexisting in the samples that interfere with the analysis [30,36,37]. Moreover, many of the coexisting components are present in higher concentrations than the analytes of interest, making the detection more complicated. Therefore, it is very important to account for the interactions of both the interferents and the analytes with the SERS substrates as well as the effect of the interferents on the SERS detection of the analytes. In this work, we systematically investigated the SERS spectra of ClO 4 solutions in the presence of common inorganic 2 anions (especially oxyanions), typically including Cl, NO 3 , SO4 ,
HCO 3 and H2PO4 using the positively-charged Cys–Ag/rCu substrates and further analyzed the effects of these inorganic anions on the SERS detection of ClO 4 . Also, the interaction mechanisms were examined in order to gain a better understanding on the effect of the anions on the ClO 4 SERS spectra and aid in development and optimization of the sensitive and selective SERS substrates for the ClO 4 detection in environmental water.
Experimental Materials Copper foils (Cu, 10 cm 10 cm in area, 0.25 mm in thickness, 99.99%) were purchased from Goodfellow (Oakdale, PA, USA). Sodium perchlorate (NaClO4 H2O) was purchased from EM Science (Cherry Hill, NJ, USA). Silver nitrate (AgNO3) and cysteamine hydrochloride were obtained from Fisher Scientific (Fair Lawn, NJ, USA) and Sigma–Aldrich (Milwaukee, WI, USA), respectively. All other chemicals including sodium chloride (NaCl), sodium nitrate (NaNO3), sodium sulfate (Na2SO4), sodium bicarbonate (NaHCO3) and sodium phosphate monobasic (NaH2PO4) were analytical grade and purchased from Sigma–Aldrich (Milwaukee, WI, USA) or Fisher Scientific and used as received. Deionized (DI) water with a resistivity of 18.2 MX cm (Millipore Milli-Q System) was used throughout the experiments. Preparation of Cys–Ag/rCu substrates The Cu foils were cut into pieces of 10 cm 2.5 cm and washed consecutively with acetone and ethanol to remove organic impurities. To obtain fresh and roughened Cu surfaces, the cleaned Cu foils were etched by immersion in a 1:1 (V:V) nitric acid (HNO3) solution for 20 s. After rinsing thoroughly with DI water, the roughened Cu (rCu) foils were kept in a 1% sulfuric acid (H2SO4) solution to prevent the rCu foil from undergoing excessive oxidation. The deposition of the Ag nanofilm and the subsequent Cys modification followed an optimized procedure reported in our previous work [34]. Briefly, a piece of rCu foil was immersed into a 50 mL aqueous solution containing 2.5 mM AgNO3 for 5 min at room temperature. After rinsing with copious amounts of DI water and dried under a stream of compressed nitrogen gas, the resulting Ag/rCu film was incubated in a freshly prepared 5 mM Cys solution for 3 min to allow the formation of the SAM on the Ag surface. The Cys–Ag/rCu substrate was obtained after DI rinsing and N2 drying, and kept in an enclosed container prior to the SERS measurements. Instruments and methods The deposition of the Ag nanofilm on the roughened Cu surface was confirmed by images (Fig. 1S in Supplementary Data) obtained with a field-emission scanning electron microscope (FESEM) (LEO 982, LEO Electron Microscopy Inc., Thornwood, NY) operated at an accelerating voltage of 7 kV and working distance (WD) of 5 or 6 mm. The Raman spectra were collected in high resolution mode on a Thermo Nicolet Almega XR Dispersive Raman Spectrometer (Thermo Fisher Scientific Inc., USA) equipped with a CCD detector, an optical microscope and a digital camera, and a 780 nm laser line with a maximum power of 32 mW (16 mW was applied in the experiments). All measurements were conducted using a 180° backscattering geometry. A 10X microscope objective was used, providing a laser spot size of 3.1 lm. Aqueous ClO 4 samples in a concentration range of 1–2000 lM were prepared by diluting a stock solution of 0.1 M with DI water. 2 Solutions of Cl, NO 3 , SO4 , HCO3 and H2PO4 were prepared by diluting their sodium salt solutions of 0.1 M with DI water, respec-
1595
J. Hao et al. / Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy 136 (2015) 1593–1599
tively. The SERS solutions of ClO 4 containing each of these five anions were prepared by mixing equal volumes (500 lL for each) of ClO 4 solutions and the corresponding anion solutions to obtain desired concentrations of both ClO 4 and coexisting anion, respectively. The SERS solutions of ClO 4 containing all the five anions (multiple anions solutions) were prepared by mixing equal volumes (200 lL for each) of five anion solutions followed by adding 10 lL of ClO 4 solutions (1% total volume) with different concentrations to obtain the desired concentrations of each anion. The pH values of all the solutions were measured without any adjustments. For SERS analysis, 100 lL of SERS solution was transferred onto the Cys–Ag/rCu substrate and restricted to a 10 mm 10 mm (100 mm2) area using a special sampling module. The sampling module consists of a stage and a holder with a 10 mm 10 mm square hole. Prior to the water solution addition, the Cys–Ag/rCu substrate was fixed and sealed with a robber ring between the stage and the holder. The data acquisition time was 3 s per scan and 10 scans were used for each spectrum collection. For reliable and reproducible SERS measurements, five spectra were collected at five different positions for each sample (Figs. 2S and 3S in Supplementary Data), and an averaged spectrum was obtained.
914 cm1 around the position of the ClO SERS band 4 (930 cm1), no obvious Raman peak appears at 930 cm1, eliminating the possibility of a false positive response resulting from the background spectrum of the substrate. However, the nearby 914 cm1 SERS band of Cys may interfere with SERS detection of low concentrations of ClO 4. Effects of individual coexisting anions on ClO 4 SERS spectra In order to individually examine the effect of each coexisting anion on the ClO 4 SERS spectra and compare the effects of these anions, five series of SERS solutions containing a constant concentration (50 lM) of ClO 4 and varying concentrations (0–5000 lM) of each coexisting anion were prepared. SERS spectra in the range of 700–1200 cm1 were obtained from these solutions and are shown in Fig. 1B–F. From Fig. 1B, we can see that the intensity of ClO 4 peak decreased steadily as the HCO3 concentration increased, while the ClO4 peak position was unchanged. The effects of other 2 four tested anions NO 3 , H2PO4 , SO4 and Cl on ClO4 SERS spectra were observed similarly, as shown in Fig. 1C–F. These results indicate that all the tested anions could interact with the substrate and + were competitive with ClO 4 ions for binding to the –NH3 groups of the Cys molecules. Meanwhile, we noticed that when the NO 3 concentration P50 lM and SO2 4 P10 lM, two new peaks appeared at 1042 and 975 cm1 in the SERS spectra, and the intensities of both 2 peaks increased as the concentrations of NO 3 and SO4 increased, respectively, as shown in Fig. 1C and E. These two bands could be 1 1 attributed to NO ) and SO2 ) ions, respec3 (1042 cm 4 (975 cm tively [7]. Although previous investigations indicated that HCO 3 and H2PO 4 exhibited the characteristic Raman or SERS bands [15,30,43], none of them were observed in Fig. 1B and D even at high concentrations up to 5000 lM, implying very weak interaction between these two anions and the –NH+3 groups of Cys molecules. The above results indicate that SO2 and NO 4 3 may have stronger interaction with the –NH+3 groups of Cys molecules and exert higher influence on the SERS spectrum of ClO 4 than other anions such as HCO anion is a single 3 and H2PO4 . Since the Cl atom species, it does not exhibit a Raman active mode and no SERS peak appeared in spite of its concentration, as shown in Fig. 1F. It is interesting that the 975 cm1 peak of the background spectrum almost disappeared when 50 lM ClO 4 was applied on the
Results and discussion Formation of Cys SAM and background spectrum of the substrate The formation of the Cys SAM on the Ag/rCu nanofilm was confirmed by its SERS spectrum (i.e. the background spectrum of the Cys–Ag/rCu substrate). Fig. 1A illustrates a typical Cys SERS spectrum in the range of 500–1250 cm1 collected in DI water, showing the characteristic SERS bands of Cys. A very strong Raman band appears at 635 cm1, which can be assigned to the C–S stretching vibration mode of a gauche-conformation of the Cys molecule; while a weak band observed at 715 cm1 can be assigned to the trans-conformation [34,38–42]. This indicates that both conformers coexist in the Cys SAM on the substrate, and the gauche one predominates. The characteristic bands of other vibrational modes such as C–C and C–N stretching vibrations and the CH2 stretching vibrations were also observed (not shown in the spectrum). We note that although there are two peaks at 975 and
635
Int
7000
4000
6000 5000
975 914 828
4000
715
B
4000
3000
800
600
1200
4000
4000
Int
Int
3000 2000
1200
3000
Raman shift / cm-1
930
E
800
1000
Raman shift / cm-1
930
D
800
1000
Raman Shift /cm-1
4000 975
930
F
3500
Int
1000
2000
0
2000 1200
1042
1000
2500
3000
930
C
3000
3500
Int
A
8000
Int
930
4500
9000
3000
2000 2500
1000 1200
1000 1000
800
Raman shift / cm-1
1200
2000 1000
800
Raman shift / cm-1
1200
1000
800
Raman shift / cm-1
Fig. 1. (A) Typical background spectrum of the Cys–Ag/rCu substrate (in the DI water) prepared by incubating Ag/rCu nanofilms in 5 mM Cys solution for 3 min. (B–F) SERS 2 spectra of 50 lM ClO 4 solutions with the presence of different concentrations of (B) HCO3 , (C) NO3 , (D) H2PO4 , (E) SO4 and (F) Cl , respectively, on Cys–Ag/rCu substrates. The concentrations of every coexisting anion were 0, 10, 50, 100, 500, 1000 and 5000 lM (from top to bottom). The spectra were shifted vertically for clarity but the relative intensity was kept unchanged.
J. Hao et al. / Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy 136 (2015) 1593–1599
substrates with or without the presence of the other anions (Fig. 1B–F) unless high concentrations of SO2 coexisted in the 4 sample solutions (Fig. 1E). This observation implies that the 975 cm1 peak of the background spectrum could be assigned to the adsorbed SO2 4 ions from the 1% H2SO4 solution (see the above Experimental Section), which were not removed completely during the subsequent Ag nanofilm deposition, Cys modification and washing. In order to further investigate the effects of the coexisting inorganic anions, we conducted the following calculations. First, normalized ClO 4 peak heights (IN) were calculated using the following equation by dividing each ClO 4 peak height (I) in the presence of each coexisting anion by that without the presence of the coexisting anion (I0), respectively:
Ij;C;N ¼ Ij;C =Ij;0
ð1Þ
where Ij,C,N is the normalized ClO 4 peak heights in the presence of C
lM coexisting anion j, and Ij,C and Ij,0 are the ClO4 peak heights in the presence of C lM and in the absence (0 lM) of coexisting anion j, respectively. Here, C = 0–5000 (lM); when C = 0 (lM), Ij,C = Ij,0, which is the case in the absence of any coexisting anion. The lower value of IN represents a higher effect of the coexisting anion on ClO 4 SERS intensity or stronger interaction of the coexisting anion with the substrate. Then we plotted the IN against the concentration of each coexisting anion, respectively, to obtain relationships between them as shown in Fig. 2A. A rapid decrease of the normalized ClO 4 peak heights followed by a gradually slowing decrease was
Chloride Bicarbonate Nitrate dihydrogen phosphate Sulfate
1
0.8
-
Normalized ClO 4 peak heigh t
1.2
0.6
0.4
0.2
930cm -1 7000
0
6000
0
5000
1042 cm-1 975 cm-1
4000 3000 2000 1000 1200
500
1500
2500
3500
4500
5500
Coexisting anion concentration (µM)
Chloride Bicarbonate Nitrate dihydrogen phosphate Sulfate Linear (Chloride) Linear (Bicarbonate) Linear (Nitrate) Linear (dihydrogen phosphate) Linear (Sulfate)
1
0.8
-
Normalized ClO 4 peak heigh t
1.2
0.6
0.4
0.2
0 0.5
1
1.5
2
2.5
3
3.5
4
Log [coexisting anion concentration (µM)] Fig. 2. Normalized ClO 4 peak height plotted against (A) the concentration of each coexisting anion, and (B) log[coexisting anion concentration (lM)].
800
1000 -1
Raman shift / cm ClO 4
Fig. 3. SERS spectra of (50 lM) in the presence of the five coexisting anions 2 Cl, HCO 3 , NO3 , H2PO4 and SO4 with equal concentration for each as a function of total concentration on the Cys–Ag/rCu substrates. The concentration of aqueous ClO 4 was kept constant at 50 lM, and total concentration of the five coexisting anions were 0, 5, 50, 250, 500, 2500, 3750, 5000 and 10,000 lM (from top to bottom). The inset shows an expansion of the ClO 4 SERS band region for the range of the high concentrations of the coexisting anions. The spectra were shifted vertically for clarity but the relative intensity was kept unchanged.
observed as the concentrations of the coexisting anions increased from 0 to 5000 lM, indicating that the sensitivity of Cys–Ag/rCu substrates to the ClO 4 ion was depressed by the coexisting anions due to the competitive adsorption. It can be deduced that the SERS band of ClO 4 (50 lM) would be quenched if the coexisting anion concentration was further elevated. Even so, 50 lM ClO 4 could still be detected in the presence of the coexisting anions with two orders of magnitude higher concentrations, revealing a higher selectivity of the substrate towards ClO 4 ion than the other anions tested here. More interestingly, the curves in Fig. 2A are considerably different, suggesting the effects of these coexisting anions on ClO 4 SERS intensity have significant differences. Evidently, SO2 exerts the 4 most significant effect while the H2PO 4 exerts the least among the five tested anions. To more distinctly illustrate the observation, the normalized ClO 4 peak heights were plotted as a function of log[coexisting anion concentration (lM)], as shown in Fig. 2B. The plots (as symbols) were fit by linear regression using Microsoft Excel regression routines. The corresponding fitting lines are also shown in Fig. 2B. The fitting equation is of the form:
Ij;C;N ¼ ax þ b ¼ a log½C þ b 0 -500
0
Int
1596
ð2Þ
where Ij,C,N and C have same meanings as those in Eq. (1), and a and b are fitting parameters, representing the slope of the fitting line and intercept at the y-axis, respectively. The values of a and b are listed in Table 1 together with regression correlation coefficient (R2) values. The high R2 values in Table 1 demonstrate good linear relationships between the normalized ClO 4 peak heights and log [C] in the existing anion concentration range from 0 to 5000 lM. This indicates the potential of quantitative evaluation regarding the effects of coexisting anions on ClO 4 SERS intensity. From Fig. 2B, we can see that the four fitting straight lines, except for the case of HCO 3 , did not cross each other, suggesting the magnitude of the effect of the different coexisting anions on ClO 4 SERS fol lowed a constant order of SO2 4 > NO3 > Cl > H2PO4 for the concentration range tested here. Obviously, this order mainly resulted from the different interaction strengths of the existing anions with the –NH+3 groups of Cys molecules. The stronger the interaction is, the higher the effect is. It has been found that the –NH+3 group has a stronger interaction with those anions which possess a higher charge or a smaller solvated radius or greater polarizability. Mosier-Boss and Lieberman have calculated the ion-pair formation constants between the –NH+3 group and various anions, 2 yielding the following order: ClO (1620) > NO 4 (6150) > SO4 3
1597
J. Hao et al. / Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy 136 (2015) 1593–1599 Table 1 Parameters obtained from the linear fitting of the normalized ClO 4 peak heights against the values of log[coexisting anion concentration] using Eq. (2). Coexisting anion HCO 3
NO 3
H2PO 4
SO2 4
All 5 anions
0.217 1.057 0.979
0.267 1.112 0.966
0.195 0.924 0.966
0.200 1.101 0.948
0.158 0.597 0.936
0.279 1.111 0.976
Effects of multiple coexisting anions on ClO 4 SERS ClO 4
Fig. 3 shows the SERS spectra in the presence of all five 2 coexisting anions Cl , HCO with equal 3 , NO3 , H2PO4 and SO4 concentration for each as a function of total concentration, obtained in spiked water samples on Cys–Ag/rCu substrates. The SERS solutions contained a constant concentration (50 lM) of ClO 4 and varying concentrations (0–10,000 lM) of coexisting anions (1–2000 lM for each). Similarly, the intensity of ClO 4 SERS band decreased steadily as the total concentration of the coexisting anions increased. When the total concentration of the coexisting anions P2500 lM (500 lM for each), the ClO 4 SERS band became a shoulder the background band of 914 cm1. When the total concentration was raised to 10,000 lM (2000 lM for each), the ClO 4 SERS band has been quenched completely. Since NO3 and 2 SO4 as two of five matrix anions simultaneously existed in the
solutions, we could see that their characteristic SERS bands at 1042 and 975 cm1 appeared when the total concentration of the coexisting anions P250 lM (50 lM of NO 3 ) and P50 lM (10 lM of SO2 4 ), respectively. This is in agreement with the above observations. Whereas, the intensities of these two bands especially the NO 3 band appeared to be weaker compared with those in the case of individual coexisting anion (see Fig. 1) due to the competitive of other coexisting anions. This indicated that the competitive binding for the –NH+3 groups of Cys molecules existed simultaneously among all the anions in the solution. The plots of the normalized ClO 4 peak height against the total concentration of the five coexisting anions (lM) and the log[total concentration of five coexisting anions (lM)] are shown in Fig. 4A and B, respectively. The similar profiles to those in the cases of the individual coexisting anions (see Fig. 2) were observed and the good linear relationships were obtained (indicated by a high R2 value as also listed in Table 1).
1.2
1
0.8
-
Normalized ClO 4 peak height
(382) > Cl (146) > H2PO 4 (no interaction) [30]. These results agree well with our observations. In fact, this order also represents the selectivity of the Cys–Ag/rCu substrate towards these anions including ClO 4. We notice that Mosier-Boss’ calculation indicated there was no interaction between the –NH+3 group and H2PO 4 . However, our results indicated the H2PO 4 did have a non-negligible effect on the ClO 4 SERS as described above. Dilute NaH2PO4 water solution exhibits a weak acidic property due to a weak dissociation of + 2 H2PO 4 to HPO4 and H (pKa2 = 7.20), but the H2PO4 species predominates. The dissociation resulted in more acidity of the solution. In our previous study, it was found maximum SERS signal of perchlorate appeared at pH 6–7; departing from this pH range the SERS signal decreased greatly [34]. The more acidity possibly changed the nanostructure of the Ag film unfavorably. On the other hand, the resulting HPO2 4 ions possess higher charge, which may have an increased interaction with –NH+3 groups than H2PO 4 . Both factors mentioned above possibly lead to the non-negligible effect of H2PO 4 on ClO4 SERS intensity which was observed in our study. Another interesting observation is that the fitting line for HCO 3 shown in Fig. 2B exhibited a particular profile. It has a cross connection with the fitting line for NO 3 at the concentration of 500 lM. In the range of lower concentrations, the effect of HCO 3 on concentration ClO4 SERS intensity appeared to be smaller than that of NO3 ; while in the range of higher concentrations, the effect of HCO 3 became greater than that of NO3 . NaHCO3 dissolution into water will increase the pH value of the solution due to HCO (pKb2 = 7.62). The 3 hydrolysis yielding H2CO3 and OH higher the concentration is, the higher the pH is, but the prevalent species is still HCO 3 . When the HCO3 concentration was higher than 1000 lM, the pH of the solution was determined to be >8. As mentioned above, our previous study has shown that the ClO 4 SERS intensity decreased as pH increased when pH > 7 [34]. This is because deprotonation of the –NH+3 groups is increased as the pH is increased, which diminished the affinity of the substrates for ClO 4 . Therefore, the effect of HCO3 on the ClO4 SERS intensity would increase faster and become greater than that of NO 3 as their concentrations increase. Based on the above results and discussion, the selectivity of the substrate towards the anions including ClO 4 2 can simply be concluded: ClO 4 > SO4 > HCO3 , NO3 > Cl > H2PO4 .
0.6
0.4
0.2
0 -500
1000
2500
4000
5500
7000
8500
10000
Total concentration of 5 coexisting anion (µM) 1.2
1
y = -0.279x +1.1114 R 2 = 0.9761
0.8
-
a b R2
Cl
Normalized ClO 4 peak height
Fitting parameter
0.6
0.4
0.2
0 0.5
1
1.5
2
2.5
3
3.5
4
Log [Total concentration of 5 anions (µM)] Fig. 4. Normalized ClO 4 peak height plotted against (A) the total concentration of five coexisting anions, and (B) log[total concentration of 5 anions]. The normalized peak heights were calculated using Eq. (1), and the fitting line was obtained by fitting the data with Eq. (2).
1598
J. Hao et al. / Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy 136 (2015) 1593–1599
930cm -1 12000 10000
Int
8000 6000 4000
940
920
900
2000 -0 1200
800
1000 -1
Raman shift / cm ClO 4
Fig. 5. SERS spectra of in the presence of five coexisting anions Cl, HCO 3, 2 NO with equal concentration for each as a function of ClO 3 , H2PO4 and SO4 4 concentration, obtained in spiked water samples on Cys–Ag/rCu substrates. The total concentration of the five coexisting anions was kept constant at 500 lM, and the concentration of aqueous ClO 4 were 0, 10, 25, 50, 100, 250, 500, 1000 and 2000 lM (from bottom to top). The spectra were shifted vertically for clarity but the relative intensity was kept unchanged.
Quantitative SERS detection of ClO 4 in the presence of multiple coexisting anions Fig. 5 shows SERS spectra of ClO 4 at concentrations varying from 0 to 2000 lM in the presence of Cl, HCO 3 , NO3 , H2PO4 and
12000
SO2 with equal concentrations for each (total 500 lM) in the 4 spiked water samples on the Cys–Ag/rCu SERS substrates. From the figure, it is clear that a steady increase of the peak intensity 1 of the ClO is observed with increasing 4 SERS band at 930 cm ClO concentration in solutions. When ClO 4 4 concentration is lower 1 than 10 lM, the ClO could not be dis4 SERS band at 930 cm cerned. Thus, 10 lM was determined to be the limit of detection (LOD) in the presence of these anions. Compared with our previous result obtained from the ClO 4 solutions without coexisting anions [34], the LOD determined here in the presence of these anions appeared to be much higher. Besides the ClO 4 SERS band, the characteristic bands of both SO2 4 and NO3 were also observed in the spectra as expected, while their intensities were depressed as the ClO 4 concentration increased. To evaluate the potential of ClO 4 quantitative analysis in the presence of the coexisting anions, the peak height of ClO 4 Raman band was plotted as a function of the ClO 4 concentration (10– 2000 lM), as shown in Fig. 6A. A log–log plot was also conducted as shown in Fig. 6B. Clearly, from Fig. 6A, it can be seen that at low ClO 4 concentration range (6100 lM), the ClO4 peak height increased linearly with the increase of concentration from 10 to 100 lM, and an excellent linear fitting curve could be obtained as shown in the inset of Fig. 6A. At the high concentration, the response will level off as the adsorption sites (–NH+3) on the Cys– Ag/rCu substrate became fully occupied. It is of much interest that a log–log plot between the ClO 4 peak height and the concentration also yielded a good linear relationship over the whole concentration range of 10–2000 lM tested here (Fig. 6B). These results indicate that the calibration curves could be fabricated successfully for quantitative analysis of ClO 4 using Cys–Ag/rCu SERS substrates in the presence of multiple anions in solution.
8000
Conclusions
6000
The effects of the coexisting inorganic anions including Cl, 2 HCO 3 , NO3 , H2PO4 and SO4 on SERS of aqueous ClO4 have been examined using the Cys–Ag/rCu substrates in this work. All the anions tested here did not change the position of the ClO 4 characteristic SERS band but did contribute to a decreased intensity of + the ClO 4 SERS band due to their competitive binding with –NH3 groups of Cys molecules of the Cys–Ag/rCu substrates. Besides the competitive binding, both the dissociation of H2PO 4 and the hydrolysis of HCO 3 may play a non-negligible role. The selectivity of the Cys–Ag/rCu substrate towards these anions was concluded to be in 2 the following order: ClO 4 > SO4 > HCO3 , NO3 > Cl > H2PO4 . In the solution with multiple anions, the ClO SERS was affected simulta4 neously by all the coexisting anions, and the SERS spectra of the aqueous ClO 4 at different concentrations were analyzed. The two kinds of calibration curves have been successfully constructed in the two different concentration ranges, demonstrating the potential of quantitative detection of aqueous ClO 4 in the presence of these anions. The LOD of ClO 4 was found to be higher compared with our previous result obtained in the solutions without these anions. Further effort will be put on the improvement of the sensitivity and the selectivity of the SERS substrate towards ClO 4 ions and the validation of the methodology for the application to the drinking water and liquid foods.
-
ClO 4 peak height
10000
4000 y = 20.348x + 240.9 R 2 = 0.988
2000 0
0
0
500
20
1000
40
60
1500
80
100
1
2000
ClO 4 - concentration (µM)
10000
-
ClO 4 peak height
100000
y = 111.4x 0.6199 R 2 = 0.987
1000
100 10
100
1000
10000
ClO4- concentration (µM) Fig. 6. Concentration dependence of the ClO 4 peak heights in the presence of five 2 coexisting anions Cl, HCO with equal concentration for 3 , NO3 , H2PO4 and SO4 each: (A) The plot of the peak height of the 930 cm1 band against the ClO 4 concentration between 10 and 2000 lM. The inset is the plot for the range of 10–100 lM, and a calibration curve has been obtained by fitting the plot with a linear Eq. I = a C + b. (B) The log–log plot of the ClO 4 peak height against the ClO4 concentration between 10 and 2000 lM, which has been fitted by a power fit using Eq. I = r Cb. The SERS intensity of the background spectrum at 930 cm1 has been subtracted from all the peak heights.
Acknowledgements We thank Dr. Tsan-Liang Su from Center for Environmental Systems for his technical support in the facilities. We also thank Dr. Tsengming Chou from Laboratory for Multiscale Imaging (LMSI) for his assistance in FESEM measurements.
J. Hao et al. / Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy 136 (2015) 1593–1599
Appendix A. Supplementary material Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.saa.2014.10.052. References [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] [14] [15] [16] [17] [18] [19] [20]
P.B. Hatzinger, Environ. Sci. Technol. 39 (2005) 239A. C. Wang, L. Lippincott, X. Meng, Water Res. 42 (2008) 4619. C. Wang, L. Lippincott, X. Meng, J. Hazard. Mater. 153 (2008) 663. A.B. Kirk, Anal. Chim. Acta 567 (2006) 4. R. Renner, J. Environ. Monitor. 1 (1999) 37N. U.S. EPA, EPA 505-F-11-003, Washington, DC, 2012. B. Gu, C. Ruan, W. Wang, Appl. Spectrosc. 63 (2009) 98. K.C. Bantz, A.F. Meyer, N.J. Wittenberg, H. Im, O. Kurtulus, S.H. Lee, N.C. Lindquist, S.-H. Oh, C.L. Haynes, Phys. Chem. Chem. Phys. 13 (2011) 11551. J. Hao, M.-J. Han, Z. Xu, J. Li, X. Meng, Nanoscale Res. Lett. 6 (2011) 263. J. Du, C. Jing, J. Colloid Interface Sci. 358 (2011) 54. S.J. Lee, M. Moskovits, Nano Lett. 11 (2010) 145. R.A. Alvarez-Puebla, L.M. Liz-Marzán, Angew. Chem. Int. Ed. 51 (2012) 11214. Y.C. Cao, R. Jin, C.A. Mirkin, Science 297 (2002) 1536. M. Mulvihill, A. Tao, K. Benjauthrit, J. Arnold, P. Yang, Angew. Chem. Int. Ed. Engl. 47 (2008) 6456. M.-J. Han, J. Hao, Z. Xu, X. Meng, Anal. Chim. Acta 692 (2011) 96. L. He, T. Chen, T.P. Labuza, Food Chem. 148 (2014) 42. J. Zheng, L. He, Compr. Rev. Food Sci. Food Safe. 13 (2014) 317. K. Kneipp, Y. Wang, H. Kneipp, L.T. Perelman, I. Itzkan, R.R. Dasari, M.S. Feld, Phys. Rev. Lett. 78 (1997) 1667. S. Nie, S.R. Emory, Science 275 (1997) 1102. B. Sharma, R.R. Frontiera, A.-I. Henry, E. Ringe, R.P. Van Duyne, Mater. Today 15 (2012) 16.
1599
[21] R.A. Halvorson, P.J. Vikesland, Environ. Sci. Technol. 44 (2010) 7749. [22] N.A. Hatab, C.M. Rouleau, S.T. Retterer, G. Eres, P.B. Hatzinger, B. Gu, Analyst 136 (2011) 1697. [23] L.X. Quang, C. Lim, G.H. Seong, J. Choo, K.J. Do, S.-K. Yoo, Lab Chip 8 (2008) 2214. [24] A.I. Lopez-Lorente, B.M. Simonet, M. Valcarcel, Analyst 138 (2013) 2378. [25] C. Chen, J. Hao, L. Zhu, Y. Yao, X. Meng, W. Weimer, Q. Wang, J. Mater. Chem. A 1 (2013) 13496. [26] J. Zheng, S. Pang, T.P. Labuza, L. He, Talanta 129 (2014) 79. [27] R. Holze, Surf. Sci. 202 (1988) L612. [28] M.-J. Han, J. Hao, C. Christodoulatos, G.P. Korfiatis, L.-J. Wan, X. Meng, Anal. Chem. 79 (2007) 3615. [29] J. Hao, Z. Xu, M.-J. Han, S. Xu, X. Meng, Colloids Surf. A: Physicochem. Eng. Aspects 366 (2010) 163. [30] P.A. Mosier-Boss, S.H. Lieberman, Appl. Spectrosc. 57 (2003) 1129. [31] S. Tan, M. Erol, S. Sukhishvili, H. Du, Langmuir 24 (2008) 4765. [32] C. Ruan, W. Wang, B. Gu, Anal. Chim. Acta 567 (2006) 114. [33] W. Wang, C. Ruan, B. Gu, Anal. Chim. Acta 567 (2006) 121. [34] J. Hao, M.-J. Han, J. Li, X. Meng, J. Colloid Interface Sci. 377 (2012) 51. [35] J. Xiao, T. Zhang, R. Li, Y. Meng, W. Wen, Appl. Spectrosc. 66 (2012) 1027. [36] P.A. Mosier-Boss, M.D. Putnam, Spectrochim. Acta Part A: Mol. Biomol. Spectrosc. 133 (2014) 156. [37] F.C. Nascimento, C.E.A. Carneiro, H.D. Santana, D.A.M. Zaia, Spectrochim. Acta Part A: Mol. Biomol. Spectrosc. 118 (2014) 251. [38] A. Kudelski, W. Hill, Langmuir 15 (1999) 3162. [39] A. Michota, A. Kudelski, J. Bukowska, Langmuir 16 (2000) 10236. [40] A. Michota, A. Kudelski, J. Bukowska, Surf. Sci. 502–503 (2002) 214. [41] T. Goto, H. Watarai, Langmuir 26 (2010) 4848. [42] O.J. Glembocki, M. Gowda, S. Geng, S.M. Prokes, N.Y. Garces, J. Cushen, J.D. Caldwell, in: I.S. Mark (Ed.), SPIE, 2010, p. 77570I. [43] N. Wen, M.H. Brooker, J. Phys. Chem. 99 (1995) 359.
Contents lists available at ScienceDirect
Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy journal homepage: www.elsevier.com/locate/saa
Surface-enhanced Raman scattering of perchlorate on cationic-modified silver nanofilms – Effect of inorganic anions Jumin Hao a,b, Mei-Juan Han a,⇑, Xiaoguang Meng a, Wayne Weimer b, Qingwu K. Wang b b
Center for Environmental Systems, Stevens Institute of Technology, Hoboken, NJ 07030, USA Agiltron Inc., 15 Presidential Way, Woburn, MA 01801, USA
h i g h l i g h t s
g r a p h i c a l a b s t r a c t
Cationic thiol modified SERS
a r t i c l e
i n f o
Article history: Received 9 July 2014 Received in revised form 12 October 2014 Accepted 16 October 2014 Available online 24 October 2014 Keywords: Surface-enhanced Raman scattering (SERS) Perchlorate detection Inorganic anions Cationic thiol Silver nanofilm Competitive adsorption
930 cm-1 7000
5000 4000
100000
SERS spectra
6000
Int
substrates were fabricated for selective detection of ClO4. Effects of common anions on ClO4 SERS spectra were investigated and analyzed by modeling. Competitive interaction mechanisms of the anions with the SERS substrate were discussed. Quantitative SERS detection of ClO4 in the presence of common anions was demonstrated. The study will help developing a robust SERS substrate for aqueous ClO4 detection in matrix.
Calibration curve
ClO 4- peak height
a
1042 cm-1 975 cm-1
3000 2000 1000
10000
1000
100
1200
1000
10
800
100
1000
10000
ClO4- concentration (uM)
Raman shift / cm-1
a b s t r a c t Surface-enhanced Raman scattering (SERS) has emerged as one of the most sensitive spectroscopic analysis methods for the detection of environmental contaminants in water, including perchlorate (ClO 4 ). However, as with other commonly used analytical techniques, analysis of realistic environmental samples by SERS presents a challenge due to complex chemical components coexisting in the samples. In this work, we investigated the influence of inorganic anions (particularly oxyanions) on SERS spectra of ClO 4 using a cationic thiol modified silver nanofilm substrate (Cys–Ag/rCu). The results show that the anions present in the samples did not shift the ClO4 characteristic band positions, but did decrease signal intensities due to their competitive binding with the –NH+3 groups of cationic thiol molecules immobilized on the substrates. The pH changes caused by both the dissociation of H2PO 4 and the hydrolysis of HCO3 may also play a non-negligible role. The selectivity of the Cys–Ag/rCu substrate towards these anions was 2 determined to be in the following order: ClO 4 > SO4 > HCO3 , NO3 > Cl > H2PO4 , indicating preferential adsorption of ClO 4 ions. In the solutions with multiple anions present, the ClO4 SERS spectra were affected simultaneously by all the coexisting anions. Calibration curves with very good linear relationships were successfully obtained, demonstrating the great potential of quantitative detection of aqueous ClO 4 in the matrix. Ó 2014 Elsevier B.V. All rights reserved.
Introduction
⇑ Corresponding author. Tel.: +1 201 216 8013; fax: +1 201 216 8303. E-mail address: [email protected] (M.-J. Han). http://dx.doi.org/10.1016/j.saa.2014.10.052 1386-1425/Ó 2014 Elsevier B.V. All rights reserved.
Perchlorate (ClO 4 ) is a widely used component of solid fuel missile and rocket propellants, explosives, and pyrotechnics [1]. It has recently emerged as a widespread environmental contaminant
1594
J. Hao et al. / Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy 136 (2015) 1593–1599
found in foods, milks, drinking water, groundwater, and soils in many countries including the United States [1–4]. ClO 4 is believed to disrupt the thyroid function by inhibiting the uptake of iodide, affecting the production of thyroid hormone and possibly causing mental retardation in fetuses and infants [4,5]. The U.S. Environmental Protection Agency (EPA) is developing a proposed national primary drinking water regulation for ClO 4 , and has established an Interim Drinking Water Health Advisory limit of 15 lg/L (ppb) [6]. This concentration limit will definitely have a significant impact on the management and remediation of perchlorate-contaminated sites and public water systems, and will require technological innovations for sensitive, rapid, and inexpensive field analyzers. Currently used techniques for ClO 4 analyses include mainly ion chromatography (IC), electrospray ionization mass spectrometry (ESI-MS), IC–MS and ion-selective electrodes (ISE) [7]. However, these methods either require large, expensive equipment, timeconsuming and complex sample preparation, well-trained users, or are associated with issues of sensitivity and selectivity, making them not suitable for field applications. Therefore, there is a critical need for the development of a rapid, simple and cost-effective portable chemical analyzer for on-site identification and quantification of trace-level ClO 4 in environmental water. Surface-enhanced Raman scattering (SERS), as one of the most sensitive spectroscopic analysis methods, has been extensively investigated for chemical and biological sensing [8–17] including the possibility of detecting single molecules [18–20]. In particular, the combination of the SERS technique with commercially available portable/handheld Raman systems has shown great potential to meet the criteria of field assays of environmental pollutants including ClO 4 ions [21–26]. It is well known that only the molecules which adsorb chemically or physically onto nanostructured substrates of noble metals (mostly Ag or Au) produce a significant SERS effect. The SERS detection of trace level ClO 4 in water appears to be very difficult due to its high water solubility and extremely weak sorption tendency on these metals [27,28]. To solve this problem, two approaches have been used. One is to measure SERS spectra using samples air-dried on planar SERS substrates to deposit perchlorate onto nanostructured Ag or Au surfaces [22,29]. Another approach has been developed to chemically modify the SERS substrates, trapping and concentrating ClO 4 anions onto Ag or Au surfaces [7,30–36]. The modification reagents commonly bear positively-charged groups such as –NH+3, –NH(CH3)+2 and – N(CH3)+3. For example, we recently developed a cysteamine hydrochloride (Cys) modified Ag nanofilm deposited on a roughened Cu foil (Cys–Ag/rCu) for rapid and sensitive SERS analysis of aqueous ClO 4 without the need for drying. The Cys (HSCH2CH2NH2 HCl) is a cationic thiol. The Cys molecules formed a self-assembled monolayer (SAM) on the surface of the Ag/rCu nanofilm, and their protonated amino groups –NH+3 make the substrate surface positively charged, which promotes the adsorption and enrichments of the negatively charged ClO 4 ions. Thus, the SERS signal of ClO4 can be enhanced dramatically, enabling sensitive SERS measurement to be accomplished immediately once the aqueous ClO 4 samples are applied onto the Cys–Ag/rCu film. However, as with other commonly used techniques, analysis of realistic environmental samples by SERS presents a challenge because of complex chemical components coexisting in the samples that interfere with the analysis [30,36,37]. Moreover, many of the coexisting components are present in higher concentrations than the analytes of interest, making the detection more complicated. Therefore, it is very important to account for the interactions of both the interferents and the analytes with the SERS substrates as well as the effect of the interferents on the SERS detection of the analytes. In this work, we systematically investigated the SERS spectra of ClO 4 solutions in the presence of common inorganic 2 anions (especially oxyanions), typically including Cl, NO 3 , SO4 ,
HCO 3 and H2PO4 using the positively-charged Cys–Ag/rCu substrates and further analyzed the effects of these inorganic anions on the SERS detection of ClO 4 . Also, the interaction mechanisms were examined in order to gain a better understanding on the effect of the anions on the ClO 4 SERS spectra and aid in development and optimization of the sensitive and selective SERS substrates for the ClO 4 detection in environmental water.
Experimental Materials Copper foils (Cu, 10 cm 10 cm in area, 0.25 mm in thickness, 99.99%) were purchased from Goodfellow (Oakdale, PA, USA). Sodium perchlorate (NaClO4 H2O) was purchased from EM Science (Cherry Hill, NJ, USA). Silver nitrate (AgNO3) and cysteamine hydrochloride were obtained from Fisher Scientific (Fair Lawn, NJ, USA) and Sigma–Aldrich (Milwaukee, WI, USA), respectively. All other chemicals including sodium chloride (NaCl), sodium nitrate (NaNO3), sodium sulfate (Na2SO4), sodium bicarbonate (NaHCO3) and sodium phosphate monobasic (NaH2PO4) were analytical grade and purchased from Sigma–Aldrich (Milwaukee, WI, USA) or Fisher Scientific and used as received. Deionized (DI) water with a resistivity of 18.2 MX cm (Millipore Milli-Q System) was used throughout the experiments. Preparation of Cys–Ag/rCu substrates The Cu foils were cut into pieces of 10 cm 2.5 cm and washed consecutively with acetone and ethanol to remove organic impurities. To obtain fresh and roughened Cu surfaces, the cleaned Cu foils were etched by immersion in a 1:1 (V:V) nitric acid (HNO3) solution for 20 s. After rinsing thoroughly with DI water, the roughened Cu (rCu) foils were kept in a 1% sulfuric acid (H2SO4) solution to prevent the rCu foil from undergoing excessive oxidation. The deposition of the Ag nanofilm and the subsequent Cys modification followed an optimized procedure reported in our previous work [34]. Briefly, a piece of rCu foil was immersed into a 50 mL aqueous solution containing 2.5 mM AgNO3 for 5 min at room temperature. After rinsing with copious amounts of DI water and dried under a stream of compressed nitrogen gas, the resulting Ag/rCu film was incubated in a freshly prepared 5 mM Cys solution for 3 min to allow the formation of the SAM on the Ag surface. The Cys–Ag/rCu substrate was obtained after DI rinsing and N2 drying, and kept in an enclosed container prior to the SERS measurements. Instruments and methods The deposition of the Ag nanofilm on the roughened Cu surface was confirmed by images (Fig. 1S in Supplementary Data) obtained with a field-emission scanning electron microscope (FESEM) (LEO 982, LEO Electron Microscopy Inc., Thornwood, NY) operated at an accelerating voltage of 7 kV and working distance (WD) of 5 or 6 mm. The Raman spectra were collected in high resolution mode on a Thermo Nicolet Almega XR Dispersive Raman Spectrometer (Thermo Fisher Scientific Inc., USA) equipped with a CCD detector, an optical microscope and a digital camera, and a 780 nm laser line with a maximum power of 32 mW (16 mW was applied in the experiments). All measurements were conducted using a 180° backscattering geometry. A 10X microscope objective was used, providing a laser spot size of 3.1 lm. Aqueous ClO 4 samples in a concentration range of 1–2000 lM were prepared by diluting a stock solution of 0.1 M with DI water. 2 Solutions of Cl, NO 3 , SO4 , HCO3 and H2PO4 were prepared by diluting their sodium salt solutions of 0.1 M with DI water, respec-
1595
J. Hao et al. / Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy 136 (2015) 1593–1599
tively. The SERS solutions of ClO 4 containing each of these five anions were prepared by mixing equal volumes (500 lL for each) of ClO 4 solutions and the corresponding anion solutions to obtain desired concentrations of both ClO 4 and coexisting anion, respectively. The SERS solutions of ClO 4 containing all the five anions (multiple anions solutions) were prepared by mixing equal volumes (200 lL for each) of five anion solutions followed by adding 10 lL of ClO 4 solutions (1% total volume) with different concentrations to obtain the desired concentrations of each anion. The pH values of all the solutions were measured without any adjustments. For SERS analysis, 100 lL of SERS solution was transferred onto the Cys–Ag/rCu substrate and restricted to a 10 mm 10 mm (100 mm2) area using a special sampling module. The sampling module consists of a stage and a holder with a 10 mm 10 mm square hole. Prior to the water solution addition, the Cys–Ag/rCu substrate was fixed and sealed with a robber ring between the stage and the holder. The data acquisition time was 3 s per scan and 10 scans were used for each spectrum collection. For reliable and reproducible SERS measurements, five spectra were collected at five different positions for each sample (Figs. 2S and 3S in Supplementary Data), and an averaged spectrum was obtained.
914 cm1 around the position of the ClO SERS band 4 (930 cm1), no obvious Raman peak appears at 930 cm1, eliminating the possibility of a false positive response resulting from the background spectrum of the substrate. However, the nearby 914 cm1 SERS band of Cys may interfere with SERS detection of low concentrations of ClO 4. Effects of individual coexisting anions on ClO 4 SERS spectra In order to individually examine the effect of each coexisting anion on the ClO 4 SERS spectra and compare the effects of these anions, five series of SERS solutions containing a constant concentration (50 lM) of ClO 4 and varying concentrations (0–5000 lM) of each coexisting anion were prepared. SERS spectra in the range of 700–1200 cm1 were obtained from these solutions and are shown in Fig. 1B–F. From Fig. 1B, we can see that the intensity of ClO 4 peak decreased steadily as the HCO3 concentration increased, while the ClO4 peak position was unchanged. The effects of other 2 four tested anions NO 3 , H2PO4 , SO4 and Cl on ClO4 SERS spectra were observed similarly, as shown in Fig. 1C–F. These results indicate that all the tested anions could interact with the substrate and + were competitive with ClO 4 ions for binding to the –NH3 groups of the Cys molecules. Meanwhile, we noticed that when the NO 3 concentration P50 lM and SO2 4 P10 lM, two new peaks appeared at 1042 and 975 cm1 in the SERS spectra, and the intensities of both 2 peaks increased as the concentrations of NO 3 and SO4 increased, respectively, as shown in Fig. 1C and E. These two bands could be 1 1 attributed to NO ) and SO2 ) ions, respec3 (1042 cm 4 (975 cm tively [7]. Although previous investigations indicated that HCO 3 and H2PO 4 exhibited the characteristic Raman or SERS bands [15,30,43], none of them were observed in Fig. 1B and D even at high concentrations up to 5000 lM, implying very weak interaction between these two anions and the –NH+3 groups of Cys molecules. The above results indicate that SO2 and NO 4 3 may have stronger interaction with the –NH+3 groups of Cys molecules and exert higher influence on the SERS spectrum of ClO 4 than other anions such as HCO anion is a single 3 and H2PO4 . Since the Cl atom species, it does not exhibit a Raman active mode and no SERS peak appeared in spite of its concentration, as shown in Fig. 1F. It is interesting that the 975 cm1 peak of the background spectrum almost disappeared when 50 lM ClO 4 was applied on the
Results and discussion Formation of Cys SAM and background spectrum of the substrate The formation of the Cys SAM on the Ag/rCu nanofilm was confirmed by its SERS spectrum (i.e. the background spectrum of the Cys–Ag/rCu substrate). Fig. 1A illustrates a typical Cys SERS spectrum in the range of 500–1250 cm1 collected in DI water, showing the characteristic SERS bands of Cys. A very strong Raman band appears at 635 cm1, which can be assigned to the C–S stretching vibration mode of a gauche-conformation of the Cys molecule; while a weak band observed at 715 cm1 can be assigned to the trans-conformation [34,38–42]. This indicates that both conformers coexist in the Cys SAM on the substrate, and the gauche one predominates. The characteristic bands of other vibrational modes such as C–C and C–N stretching vibrations and the CH2 stretching vibrations were also observed (not shown in the spectrum). We note that although there are two peaks at 975 and
635
Int
7000
4000
6000 5000
975 914 828
4000
715
B
4000
3000
800
600
1200
4000
4000
Int
Int
3000 2000
1200
3000
Raman shift / cm-1
930
E
800
1000
Raman shift / cm-1
930
D
800
1000
Raman Shift /cm-1
4000 975
930
F
3500
Int
1000
2000
0
2000 1200
1042
1000
2500
3000
930
C
3000
3500
Int
A
8000
Int
930
4500
9000
3000
2000 2500
1000 1200
1000 1000
800
Raman shift / cm-1
1200
2000 1000
800
Raman shift / cm-1
1200
1000
800
Raman shift / cm-1
Fig. 1. (A) Typical background spectrum of the Cys–Ag/rCu substrate (in the DI water) prepared by incubating Ag/rCu nanofilms in 5 mM Cys solution for 3 min. (B–F) SERS 2 spectra of 50 lM ClO 4 solutions with the presence of different concentrations of (B) HCO3 , (C) NO3 , (D) H2PO4 , (E) SO4 and (F) Cl , respectively, on Cys–Ag/rCu substrates. The concentrations of every coexisting anion were 0, 10, 50, 100, 500, 1000 and 5000 lM (from top to bottom). The spectra were shifted vertically for clarity but the relative intensity was kept unchanged.
J. Hao et al. / Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy 136 (2015) 1593–1599
substrates with or without the presence of the other anions (Fig. 1B–F) unless high concentrations of SO2 coexisted in the 4 sample solutions (Fig. 1E). This observation implies that the 975 cm1 peak of the background spectrum could be assigned to the adsorbed SO2 4 ions from the 1% H2SO4 solution (see the above Experimental Section), which were not removed completely during the subsequent Ag nanofilm deposition, Cys modification and washing. In order to further investigate the effects of the coexisting inorganic anions, we conducted the following calculations. First, normalized ClO 4 peak heights (IN) were calculated using the following equation by dividing each ClO 4 peak height (I) in the presence of each coexisting anion by that without the presence of the coexisting anion (I0), respectively:
Ij;C;N ¼ Ij;C =Ij;0
ð1Þ
where Ij,C,N is the normalized ClO 4 peak heights in the presence of C
lM coexisting anion j, and Ij,C and Ij,0 are the ClO4 peak heights in the presence of C lM and in the absence (0 lM) of coexisting anion j, respectively. Here, C = 0–5000 (lM); when C = 0 (lM), Ij,C = Ij,0, which is the case in the absence of any coexisting anion. The lower value of IN represents a higher effect of the coexisting anion on ClO 4 SERS intensity or stronger interaction of the coexisting anion with the substrate. Then we plotted the IN against the concentration of each coexisting anion, respectively, to obtain relationships between them as shown in Fig. 2A. A rapid decrease of the normalized ClO 4 peak heights followed by a gradually slowing decrease was
Chloride Bicarbonate Nitrate dihydrogen phosphate Sulfate
1
0.8
-
Normalized ClO 4 peak heigh t
1.2
0.6
0.4
0.2
930cm -1 7000
0
6000
0
5000
1042 cm-1 975 cm-1
4000 3000 2000 1000 1200
500
1500
2500
3500
4500
5500
Coexisting anion concentration (µM)
Chloride Bicarbonate Nitrate dihydrogen phosphate Sulfate Linear (Chloride) Linear (Bicarbonate) Linear (Nitrate) Linear (dihydrogen phosphate) Linear (Sulfate)
1
0.8
-
Normalized ClO 4 peak heigh t
1.2
0.6
0.4
0.2
0 0.5
1
1.5
2
2.5
3
3.5
4
Log [coexisting anion concentration (µM)] Fig. 2. Normalized ClO 4 peak height plotted against (A) the concentration of each coexisting anion, and (B) log[coexisting anion concentration (lM)].
800
1000 -1
Raman shift / cm ClO 4
Fig. 3. SERS spectra of (50 lM) in the presence of the five coexisting anions 2 Cl, HCO 3 , NO3 , H2PO4 and SO4 with equal concentration for each as a function of total concentration on the Cys–Ag/rCu substrates. The concentration of aqueous ClO 4 was kept constant at 50 lM, and total concentration of the five coexisting anions were 0, 5, 50, 250, 500, 2500, 3750, 5000 and 10,000 lM (from top to bottom). The inset shows an expansion of the ClO 4 SERS band region for the range of the high concentrations of the coexisting anions. The spectra were shifted vertically for clarity but the relative intensity was kept unchanged.
observed as the concentrations of the coexisting anions increased from 0 to 5000 lM, indicating that the sensitivity of Cys–Ag/rCu substrates to the ClO 4 ion was depressed by the coexisting anions due to the competitive adsorption. It can be deduced that the SERS band of ClO 4 (50 lM) would be quenched if the coexisting anion concentration was further elevated. Even so, 50 lM ClO 4 could still be detected in the presence of the coexisting anions with two orders of magnitude higher concentrations, revealing a higher selectivity of the substrate towards ClO 4 ion than the other anions tested here. More interestingly, the curves in Fig. 2A are considerably different, suggesting the effects of these coexisting anions on ClO 4 SERS intensity have significant differences. Evidently, SO2 exerts the 4 most significant effect while the H2PO 4 exerts the least among the five tested anions. To more distinctly illustrate the observation, the normalized ClO 4 peak heights were plotted as a function of log[coexisting anion concentration (lM)], as shown in Fig. 2B. The plots (as symbols) were fit by linear regression using Microsoft Excel regression routines. The corresponding fitting lines are also shown in Fig. 2B. The fitting equation is of the form:
Ij;C;N ¼ ax þ b ¼ a log½C þ b 0 -500
0
Int
1596
ð2Þ
where Ij,C,N and C have same meanings as those in Eq. (1), and a and b are fitting parameters, representing the slope of the fitting line and intercept at the y-axis, respectively. The values of a and b are listed in Table 1 together with regression correlation coefficient (R2) values. The high R2 values in Table 1 demonstrate good linear relationships between the normalized ClO 4 peak heights and log [C] in the existing anion concentration range from 0 to 5000 lM. This indicates the potential of quantitative evaluation regarding the effects of coexisting anions on ClO 4 SERS intensity. From Fig. 2B, we can see that the four fitting straight lines, except for the case of HCO 3 , did not cross each other, suggesting the magnitude of the effect of the different coexisting anions on ClO 4 SERS fol lowed a constant order of SO2 4 > NO3 > Cl > H2PO4 for the concentration range tested here. Obviously, this order mainly resulted from the different interaction strengths of the existing anions with the –NH+3 groups of Cys molecules. The stronger the interaction is, the higher the effect is. It has been found that the –NH+3 group has a stronger interaction with those anions which possess a higher charge or a smaller solvated radius or greater polarizability. Mosier-Boss and Lieberman have calculated the ion-pair formation constants between the –NH+3 group and various anions, 2 yielding the following order: ClO (1620) > NO 4 (6150) > SO4 3
1597
J. Hao et al. / Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy 136 (2015) 1593–1599 Table 1 Parameters obtained from the linear fitting of the normalized ClO 4 peak heights against the values of log[coexisting anion concentration] using Eq. (2). Coexisting anion HCO 3
NO 3
H2PO 4
SO2 4
All 5 anions
0.217 1.057 0.979
0.267 1.112 0.966
0.195 0.924 0.966
0.200 1.101 0.948
0.158 0.597 0.936
0.279 1.111 0.976
Effects of multiple coexisting anions on ClO 4 SERS ClO 4
Fig. 3 shows the SERS spectra in the presence of all five 2 coexisting anions Cl , HCO with equal 3 , NO3 , H2PO4 and SO4 concentration for each as a function of total concentration, obtained in spiked water samples on Cys–Ag/rCu substrates. The SERS solutions contained a constant concentration (50 lM) of ClO 4 and varying concentrations (0–10,000 lM) of coexisting anions (1–2000 lM for each). Similarly, the intensity of ClO 4 SERS band decreased steadily as the total concentration of the coexisting anions increased. When the total concentration of the coexisting anions P2500 lM (500 lM for each), the ClO 4 SERS band became a shoulder the background band of 914 cm1. When the total concentration was raised to 10,000 lM (2000 lM for each), the ClO 4 SERS band has been quenched completely. Since NO3 and 2 SO4 as two of five matrix anions simultaneously existed in the
solutions, we could see that their characteristic SERS bands at 1042 and 975 cm1 appeared when the total concentration of the coexisting anions P250 lM (50 lM of NO 3 ) and P50 lM (10 lM of SO2 4 ), respectively. This is in agreement with the above observations. Whereas, the intensities of these two bands especially the NO 3 band appeared to be weaker compared with those in the case of individual coexisting anion (see Fig. 1) due to the competitive of other coexisting anions. This indicated that the competitive binding for the –NH+3 groups of Cys molecules existed simultaneously among all the anions in the solution. The plots of the normalized ClO 4 peak height against the total concentration of the five coexisting anions (lM) and the log[total concentration of five coexisting anions (lM)] are shown in Fig. 4A and B, respectively. The similar profiles to those in the cases of the individual coexisting anions (see Fig. 2) were observed and the good linear relationships were obtained (indicated by a high R2 value as also listed in Table 1).
1.2
1
0.8
-
Normalized ClO 4 peak height
(382) > Cl (146) > H2PO 4 (no interaction) [30]. These results agree well with our observations. In fact, this order also represents the selectivity of the Cys–Ag/rCu substrate towards these anions including ClO 4. We notice that Mosier-Boss’ calculation indicated there was no interaction between the –NH+3 group and H2PO 4 . However, our results indicated the H2PO 4 did have a non-negligible effect on the ClO 4 SERS as described above. Dilute NaH2PO4 water solution exhibits a weak acidic property due to a weak dissociation of + 2 H2PO 4 to HPO4 and H (pKa2 = 7.20), but the H2PO4 species predominates. The dissociation resulted in more acidity of the solution. In our previous study, it was found maximum SERS signal of perchlorate appeared at pH 6–7; departing from this pH range the SERS signal decreased greatly [34]. The more acidity possibly changed the nanostructure of the Ag film unfavorably. On the other hand, the resulting HPO2 4 ions possess higher charge, which may have an increased interaction with –NH+3 groups than H2PO 4 . Both factors mentioned above possibly lead to the non-negligible effect of H2PO 4 on ClO4 SERS intensity which was observed in our study. Another interesting observation is that the fitting line for HCO 3 shown in Fig. 2B exhibited a particular profile. It has a cross connection with the fitting line for NO 3 at the concentration of 500 lM. In the range of lower concentrations, the effect of HCO 3 on concentration ClO4 SERS intensity appeared to be smaller than that of NO3 ; while in the range of higher concentrations, the effect of HCO 3 became greater than that of NO3 . NaHCO3 dissolution into water will increase the pH value of the solution due to HCO (pKb2 = 7.62). The 3 hydrolysis yielding H2CO3 and OH higher the concentration is, the higher the pH is, but the prevalent species is still HCO 3 . When the HCO3 concentration was higher than 1000 lM, the pH of the solution was determined to be >8. As mentioned above, our previous study has shown that the ClO 4 SERS intensity decreased as pH increased when pH > 7 [34]. This is because deprotonation of the –NH+3 groups is increased as the pH is increased, which diminished the affinity of the substrates for ClO 4 . Therefore, the effect of HCO3 on the ClO4 SERS intensity would increase faster and become greater than that of NO 3 as their concentrations increase. Based on the above results and discussion, the selectivity of the substrate towards the anions including ClO 4 2 can simply be concluded: ClO 4 > SO4 > HCO3 , NO3 > Cl > H2PO4 .
0.6
0.4
0.2
0 -500
1000
2500
4000
5500
7000
8500
10000
Total concentration of 5 coexisting anion (µM) 1.2
1
y = -0.279x +1.1114 R 2 = 0.9761
0.8
-
a b R2
Cl
Normalized ClO 4 peak height
Fitting parameter
0.6
0.4
0.2
0 0.5
1
1.5
2
2.5
3
3.5
4
Log [Total concentration of 5 anions (µM)] Fig. 4. Normalized ClO 4 peak height plotted against (A) the total concentration of five coexisting anions, and (B) log[total concentration of 5 anions]. The normalized peak heights were calculated using Eq. (1), and the fitting line was obtained by fitting the data with Eq. (2).
1598
J. Hao et al. / Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy 136 (2015) 1593–1599
930cm -1 12000 10000
Int
8000 6000 4000
940
920
900
2000 -0 1200
800
1000 -1
Raman shift / cm ClO 4
Fig. 5. SERS spectra of in the presence of five coexisting anions Cl, HCO 3, 2 NO with equal concentration for each as a function of ClO 3 , H2PO4 and SO4 4 concentration, obtained in spiked water samples on Cys–Ag/rCu substrates. The total concentration of the five coexisting anions was kept constant at 500 lM, and the concentration of aqueous ClO 4 were 0, 10, 25, 50, 100, 250, 500, 1000 and 2000 lM (from bottom to top). The spectra were shifted vertically for clarity but the relative intensity was kept unchanged.
Quantitative SERS detection of ClO 4 in the presence of multiple coexisting anions Fig. 5 shows SERS spectra of ClO 4 at concentrations varying from 0 to 2000 lM in the presence of Cl, HCO 3 , NO3 , H2PO4 and
12000
SO2 with equal concentrations for each (total 500 lM) in the 4 spiked water samples on the Cys–Ag/rCu SERS substrates. From the figure, it is clear that a steady increase of the peak intensity 1 of the ClO is observed with increasing 4 SERS band at 930 cm ClO concentration in solutions. When ClO 4 4 concentration is lower 1 than 10 lM, the ClO could not be dis4 SERS band at 930 cm cerned. Thus, 10 lM was determined to be the limit of detection (LOD) in the presence of these anions. Compared with our previous result obtained from the ClO 4 solutions without coexisting anions [34], the LOD determined here in the presence of these anions appeared to be much higher. Besides the ClO 4 SERS band, the characteristic bands of both SO2 4 and NO3 were also observed in the spectra as expected, while their intensities were depressed as the ClO 4 concentration increased. To evaluate the potential of ClO 4 quantitative analysis in the presence of the coexisting anions, the peak height of ClO 4 Raman band was plotted as a function of the ClO 4 concentration (10– 2000 lM), as shown in Fig. 6A. A log–log plot was also conducted as shown in Fig. 6B. Clearly, from Fig. 6A, it can be seen that at low ClO 4 concentration range (6100 lM), the ClO4 peak height increased linearly with the increase of concentration from 10 to 100 lM, and an excellent linear fitting curve could be obtained as shown in the inset of Fig. 6A. At the high concentration, the response will level off as the adsorption sites (–NH+3) on the Cys– Ag/rCu substrate became fully occupied. It is of much interest that a log–log plot between the ClO 4 peak height and the concentration also yielded a good linear relationship over the whole concentration range of 10–2000 lM tested here (Fig. 6B). These results indicate that the calibration curves could be fabricated successfully for quantitative analysis of ClO 4 using Cys–Ag/rCu SERS substrates in the presence of multiple anions in solution.
8000
Conclusions
6000
The effects of the coexisting inorganic anions including Cl, 2 HCO 3 , NO3 , H2PO4 and SO4 on SERS of aqueous ClO4 have been examined using the Cys–Ag/rCu substrates in this work. All the anions tested here did not change the position of the ClO 4 characteristic SERS band but did contribute to a decreased intensity of + the ClO 4 SERS band due to their competitive binding with –NH3 groups of Cys molecules of the Cys–Ag/rCu substrates. Besides the competitive binding, both the dissociation of H2PO 4 and the hydrolysis of HCO 3 may play a non-negligible role. The selectivity of the Cys–Ag/rCu substrate towards these anions was concluded to be in 2 the following order: ClO 4 > SO4 > HCO3 , NO3 > Cl > H2PO4 . In the solution with multiple anions, the ClO SERS was affected simulta4 neously by all the coexisting anions, and the SERS spectra of the aqueous ClO 4 at different concentrations were analyzed. The two kinds of calibration curves have been successfully constructed in the two different concentration ranges, demonstrating the potential of quantitative detection of aqueous ClO 4 in the presence of these anions. The LOD of ClO 4 was found to be higher compared with our previous result obtained in the solutions without these anions. Further effort will be put on the improvement of the sensitivity and the selectivity of the SERS substrate towards ClO 4 ions and the validation of the methodology for the application to the drinking water and liquid foods.
-
ClO 4 peak height
10000
4000 y = 20.348x + 240.9 R 2 = 0.988
2000 0
0
0
500
20
1000
40
60
1500
80
100
1
2000
ClO 4 - concentration (µM)
10000
-
ClO 4 peak height
100000
y = 111.4x 0.6199 R 2 = 0.987
1000
100 10
100
1000
10000
ClO4- concentration (µM) Fig. 6. Concentration dependence of the ClO 4 peak heights in the presence of five 2 coexisting anions Cl, HCO with equal concentration for 3 , NO3 , H2PO4 and SO4 each: (A) The plot of the peak height of the 930 cm1 band against the ClO 4 concentration between 10 and 2000 lM. The inset is the plot for the range of 10–100 lM, and a calibration curve has been obtained by fitting the plot with a linear Eq. I = a C + b. (B) The log–log plot of the ClO 4 peak height against the ClO4 concentration between 10 and 2000 lM, which has been fitted by a power fit using Eq. I = r Cb. The SERS intensity of the background spectrum at 930 cm1 has been subtracted from all the peak heights.
Acknowledgements We thank Dr. Tsan-Liang Su from Center for Environmental Systems for his technical support in the facilities. We also thank Dr. Tsengming Chou from Laboratory for Multiscale Imaging (LMSI) for his assistance in FESEM measurements.
J. Hao et al. / Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy 136 (2015) 1593–1599
Appendix A. Supplementary material Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.saa.2014.10.052. References [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] [14] [15] [16] [17] [18] [19] [20]
P.B. Hatzinger, Environ. Sci. Technol. 39 (2005) 239A. C. Wang, L. Lippincott, X. Meng, Water Res. 42 (2008) 4619. C. Wang, L. Lippincott, X. Meng, J. Hazard. Mater. 153 (2008) 663. A.B. Kirk, Anal. Chim. Acta 567 (2006) 4. R. Renner, J. Environ. Monitor. 1 (1999) 37N. U.S. EPA, EPA 505-F-11-003, Washington, DC, 2012. B. Gu, C. Ruan, W. Wang, Appl. Spectrosc. 63 (2009) 98. K.C. Bantz, A.F. Meyer, N.J. Wittenberg, H. Im, O. Kurtulus, S.H. Lee, N.C. Lindquist, S.-H. Oh, C.L. Haynes, Phys. Chem. Chem. Phys. 13 (2011) 11551. J. Hao, M.-J. Han, Z. Xu, J. Li, X. Meng, Nanoscale Res. Lett. 6 (2011) 263. J. Du, C. Jing, J. Colloid Interface Sci. 358 (2011) 54. S.J. Lee, M. Moskovits, Nano Lett. 11 (2010) 145. R.A. Alvarez-Puebla, L.M. Liz-Marzán, Angew. Chem. Int. Ed. 51 (2012) 11214. Y.C. Cao, R. Jin, C.A. Mirkin, Science 297 (2002) 1536. M. Mulvihill, A. Tao, K. Benjauthrit, J. Arnold, P. Yang, Angew. Chem. Int. Ed. Engl. 47 (2008) 6456. M.-J. Han, J. Hao, Z. Xu, X. Meng, Anal. Chim. Acta 692 (2011) 96. L. He, T. Chen, T.P. Labuza, Food Chem. 148 (2014) 42. J. Zheng, L. He, Compr. Rev. Food Sci. Food Safe. 13 (2014) 317. K. Kneipp, Y. Wang, H. Kneipp, L.T. Perelman, I. Itzkan, R.R. Dasari, M.S. Feld, Phys. Rev. Lett. 78 (1997) 1667. S. Nie, S.R. Emory, Science 275 (1997) 1102. B. Sharma, R.R. Frontiera, A.-I. Henry, E. Ringe, R.P. Van Duyne, Mater. Today 15 (2012) 16.
1599
[21] R.A. Halvorson, P.J. Vikesland, Environ. Sci. Technol. 44 (2010) 7749. [22] N.A. Hatab, C.M. Rouleau, S.T. Retterer, G. Eres, P.B. Hatzinger, B. Gu, Analyst 136 (2011) 1697. [23] L.X. Quang, C. Lim, G.H. Seong, J. Choo, K.J. Do, S.-K. Yoo, Lab Chip 8 (2008) 2214. [24] A.I. Lopez-Lorente, B.M. Simonet, M. Valcarcel, Analyst 138 (2013) 2378. [25] C. Chen, J. Hao, L. Zhu, Y. Yao, X. Meng, W. Weimer, Q. Wang, J. Mater. Chem. A 1 (2013) 13496. [26] J. Zheng, S. Pang, T.P. Labuza, L. He, Talanta 129 (2014) 79. [27] R. Holze, Surf. Sci. 202 (1988) L612. [28] M.-J. Han, J. Hao, C. Christodoulatos, G.P. Korfiatis, L.-J. Wan, X. Meng, Anal. Chem. 79 (2007) 3615. [29] J. Hao, Z. Xu, M.-J. Han, S. Xu, X. Meng, Colloids Surf. A: Physicochem. Eng. Aspects 366 (2010) 163. [30] P.A. Mosier-Boss, S.H. Lieberman, Appl. Spectrosc. 57 (2003) 1129. [31] S. Tan, M. Erol, S. Sukhishvili, H. Du, Langmuir 24 (2008) 4765. [32] C. Ruan, W. Wang, B. Gu, Anal. Chim. Acta 567 (2006) 114. [33] W. Wang, C. Ruan, B. Gu, Anal. Chim. Acta 567 (2006) 121. [34] J. Hao, M.-J. Han, J. Li, X. Meng, J. Colloid Interface Sci. 377 (2012) 51. [35] J. Xiao, T. Zhang, R. Li, Y. Meng, W. Wen, Appl. Spectrosc. 66 (2012) 1027. [36] P.A. Mosier-Boss, M.D. Putnam, Spectrochim. Acta Part A: Mol. Biomol. Spectrosc. 133 (2014) 156. [37] F.C. Nascimento, C.E.A. Carneiro, H.D. Santana, D.A.M. Zaia, Spectrochim. Acta Part A: Mol. Biomol. Spectrosc. 118 (2014) 251. [38] A. Kudelski, W. Hill, Langmuir 15 (1999) 3162. [39] A. Michota, A. Kudelski, J. Bukowska, Langmuir 16 (2000) 10236. [40] A. Michota, A. Kudelski, J. Bukowska, Surf. Sci. 502–503 (2002) 214. [41] T. Goto, H. Watarai, Langmuir 26 (2010) 4848. [42] O.J. Glembocki, M. Gowda, S. Geng, S.M. Prokes, N.Y. Garces, J. Cushen, J.D. Caldwell, in: I.S. Mark (Ed.), SPIE, 2010, p. 77570I. [43] N. Wen, M.H. Brooker, J. Phys. Chem. 99 (1995) 359.
