Hydrogen Peroxide-Induced Oxidation of Mixtures ... - ACS Publications

Nov 4, 2011 - Department of Chemistry and Biochemistry, Samford University, Birmingham, Alabama 35229-2236, United States. bS Supporting Information...
0 downloads 0 Views 1MB Size
ARTICLE pubs.acs.org/ac

Hydrogen Peroxide-Induced Oxidation of Mixtures of Alkanethiols and Their Quantitative Detection as Alkanesulfonates by Electrospray Ionization Mass Spectrometry Rachel Roberts,† Joshua A. Driver,‡ Danielle M. Brown,§ Sagar H. Amin,^ and Brian W. Gregory* Department of Chemistry and Biochemistry, Samford University, Birmingham, Alabama 35229-2236, United States

bS Supporting Information ABSTRACT: Finding optimal experimental conditions for generating stable negative ion electrospray ionization ion trap mass spectra (ESI-IT-MS) of alkanethiol-derived species is critical for quantitatively characterizing multicomponent alkanethiol-based self-assembled monolayers by this technique. Since alkanethiolates slowly oxidize in solution, purposeful oxidation of alkanethiols to their fully oxidized form (alkanesulfonates) is advantageous: sulfonates are chemically stable and have little affinity for covalent binding to metal surfaces. We have used ESI-IT-MS to characterize the products of H2O2 oxidation of simple n-alkanethiols in solution and have observed monomeric alkanesulfonate species as well as alkanesulfonic acid/alkanesulfonate adducts, yielding gas-phase dimers and trimers. MS intensities of both monomers and adducts exhibit a dependence on the ion transfer capillary temperature that is alkyl-chain-length-dependent and that appears to be correlated with CS bond cleavage. The trend in optimal capillary temperatures indicates that entropic effects lead to lower thermal decomposition temperatures for short-chain species relative to the longer-chain homologues. MS calibration data from alkanesulfonate mixtures are characterized by large linear dynamic ranges (106103 M) and detection limits influenced by their thermal decomposition. The high degree of precision in the calibration data should facilitate distinguishing among mixed SAMs having similar compositions.

O

ver the past few years, mass spectrometry (MS) has become an increasingly important characterization tool for surfaces modified by organosulfur-based self-assembled monolayers (SAMs). MS analyses of these films allow one in principle to examine such issues as (1) the long-term stability of SAMs and reactivity at the AuS interface; (2) the products generated from, as well as the efficiency of, surface-coupling reactions at functionalized SAMs; and (3) the film composition of mixed SAM films. Surface MS techniques are inherently destructive, since some portion of the film must be desorbed, ionized, and transported to the mass analyzer. To accomplish this task, various pulsed desorptionionization probes have been designed, the most popular approaches being high-energy primary ion bombardment (e.g., used in secondary ion mass spectrometry, SIMS), direct laser desorption ionization (LDI), and matrix-assisted laser desorptionionization (MALDI).1 Time-of-flight (TOF) mass analysis is particularly well-suited for surface MS studies because it is easily adaptable to pulsed desorptionionization techniques and because it exhibits high mass sensitivity. Many TOFMS studies of organosulfur SAMs have appeared using various desorptionionization probes, including TOF-SIMS,211 LDI-TOFMS,1216 and MALDITOFMS.1728 Pulsed sources have also been employed with other mass analyzers in studies of these films, including quadrupolebased SIMS,29,30 laser desorptionionization Fourier-transform MS (LDI-FT-MS),31,32 and laser desorptionionization ion trap MS (LDI-IT-MS).33 The general characteristics, strengths, and r 2011 American Chemical Society

weaknesses of various surface mass spectrometric methods have been discussed previously.1 Although qualitative information about SAM film composition is available using pulsed MS techniques, the extraction of quantitative information is generally quite difficult. Reliable quantification of the surface composition is dependent on the effects of several variables related to the pulsed source and to the particular SAM substrate being studied. For example, the most popular method for surface MS of molecular adsorbates—static SIMS—employs a high-energy, low-intensity, primary ion beam (130 keV, 10121013 ions/cm2) to desorb and ionize surface species. Despite the relatively low flux of primary ions, studies of SAM films by static SIMS are characterized by considerable adsorbate fragmentation as well as metalion and metalcluster attachment, dimerization, and proton transfer.1,7,11 Furthermore, the ion-to-neutral ratio in the sputtering process for static SIMS is normally quite low (106103) unless some method for postionization of the sputtered neutrals is introduced.1 SIMS sources also tend to enhance coupling between the desorption and ionization events, which leads to significant matrix effects, thereby making it difficult to determine the extent to which the local chemical environment influences both the sputter yield (defined as the number of secondary particles sputtered by one Received: September 9, 2011 Accepted: November 4, 2011 Published: November 04, 2011 9605

dx.doi.org/10.1021/ac202400h | Anal. Chem. 2011, 83, 9605–9613

Analytical Chemistry primary particle) and the ionization probability for the secondary particle. Direct LDI-MS tends to suffer from many of the same problems as static SIMS: (1) low ion yields; (2) extensive fragmentation at the laser power densities necessary to simultaneously desorb and ionize the adsorbate; and (3) observed fragmented product ions that are dependent on the properties of the laser source (e.g., beam energy, power density, laser pulse duration), as well as those of both the substrate and sample.1 For alkanethiolbased SAMs, extensive fragmentation has been observed in positive ion spectra produced by LDI-MS, whereas molecular ion formation occurs in negative ion mode only if the laser energy is greater than 3.68 eV (λ < 337 nm).14 The introduction of two-laser MS (L2MS), in which the desorption and ionization events are decoupled by using two separate subnanosecond-pulsed lasers, has helped alleviate some of these problems. Postionization of the desorbed neutrals is usually achieved nonselectively by single-photon ionization (SPI) using vacuum UV radiation or selectively using a multiphoton technique, such as two-photon (1 + 1) resonanceenhanced multiphoton ionization (REMPI).1 For example, studies of organothiol SAMs using SPI/L2MS have resulted in positive ion spectra dominated by gas-phase dimers with little or no fragmentation.14 In the same study, mixed SAMs consisting of 11-mercaptoundecane/11-mercaptoundecanol (1:5 solution ratio) were investigated, in which larger signal intensities were observed for the methyl-terminated disulfide dimer compared with that for the hydroxyl-terminated one. This observation is curious, given the equivalent chain lengths of both species, the higher mass of the hydroxyl-labeled component, the potential for hydrogen bonding between surface hydroxyls, and the significant predominance of the hydroxyl-labeled component in solution. All of these factors should lead one to predict a higher proportion of hydroxyl-terminated alkanethiol in the resulting mixed SAM. Such results suggest that the relative ionization cross sections and desorption efficiencies may be different for the two components and indicate that quantification of surface composition by SPI/ L2MS is still an unresolved problem, even for simple organic films. Since both SPI and REMPI require pulsed laser desorption to initially generate gas-phase neutrals, both techniques are also susceptible to laser power variability issues that complicate surface quantification with LDMS. MALDI-MS has gained increasing visibility over the past decade as a surface technique to study SAMs and other molecular films. Its use has often been directed at characterizing reactions at model biological surfaces. To study molecular films by MALDIMS, a thin layer of the laser-absorbing matrix compound is deposited directly onto the film surface from a solution that contains only the matrix material and internal standard in a volatile solvent, and only after reaction or adsorption of the analyte has occurred. Accurate quantification usually depends on the capacity of the matrix solution to desorb and solubilize the analyte within itself prior to crystallization.27 Despite the advantages of MALDI,25 quantification has primarily been limited by the irreproducibility in applying the analytematrix solution to the laser target, as well as the heterogeneity of the resulting crystalline deposit following solution evaporation.27 Surface MALDI intensities are strongly influenced by the binding strengths (physisorption vs chemisorption) of the adsorbates; these can lead to significant differences in desorption efficiency as well as ionization efficiency. Competitive ionization during laser desorption/ionization also limits quantification by MALDI;

ARTICLE

ionization suppression of the analyte is acutely sensitive to the level of sample contamination and is more problematic with more complex samples. The difficulties associated with accurate quantification of surface amounts (both absolute and relative) in SAMs by surface MALDI techniques have been noted elsewhere recently.23,24 The focus of the research described in this article is to examine the feasibility of using electrospray ionization ion trap mass spectrometry (ESI-IT-MS) as a quantitative tool for SAM films. The high mass sensitivity requirements in surface and thin film work often limit the choice of mass analyzer that is employed (e.g., TOF vs IT). Given the relatively high detection limits of many standard benchtop IT mass spectrometers (1091012 g range), it is therefore not surprising that IT-MS has seen limited use in such studies. The coupling of ESI to IT-MS tends to compound this problem, since low numbers of analyte species are typically produced upon desorption into the effluent (monolayer coverages fall into the range 1011109 mol/cm2). Accordingly, ESI has been utilized only infrequently as an ionization source for MS studies of SAMs and SAM-related applications.3436 Preconcentration of the analyte by desorption of multiple SAM films into the same small aliquot of solution (1 pmol/μL).34,36 Although this approach would raise the concentration above the detection limit, it significantly compromises one’s ability to evaluate sample-to-sample variation in any direct way. Despite the aforementioned limitations, ESI-IT-MS does have the sensitivity to be useful for investigations of SAMs. For example, full monolayer coverage of a n-decanethiol (C10H21SH) SAM on Au(111) is characterized by a mass density of 1.3 ng/mm2. Desorption of this SAM from a Au surface of reasonable area (e.g., 4 cm2) into a low-volume electrochemical cell (100 μL) should therefore yield a 5.2 ng/μL solution, which is well within the detection limits for IT-MS analysis. Thus, ESIIT-MS has the sensitivity to detect monolayer (and even submonolayer) quantities of materials if configured properly. More important is the fact that quantification by ESI-IT-MS is straightforward compared with the surface MS techniques described above. So long as desorption of the adsorbate leads to an electroactively stable solution product and to effluent concentrations that are large enough to be detected, successful quantification of SAMs (or any other surface film) by ESI-IT-MS should be realizable. The research presented here involves our initial investigations into the general behavior of some simple n-alkanethiols (CnH(2n+1)SH = CnSH) using ESI-IT-MS, the conditions under which optimal MS signals are obtained, and their quantification in negative ion mode. Negative ion detection was chosen because these compounds exhibit a proclivity to form anionic products in solution either before or upon oxidation: thiolates (RS), sulfinates (RSO2), or sulfonates (RSO3). Conversion of alkanethiols to their sulfonate forms is advantageous from a quantitative perspective, as the weak Au-sulfonate interaction and high electron affinity of the sulfonate moiety are expected to facilitate gas-phase molecular anion formation. Thus, all the quantitative work discussed later in this paper focuses on analyses of n-alkanethiols purposefully oxidized to sulfonates by excess H2O2. The approach described in this paper is currently being applied to single-component SAMs on Au and will be the subject of a future publication. It is expected that our method should be readily 9606

dx.doi.org/10.1021/ac202400h |Anal. Chem. 2011, 83, 9605–9613

Analytical Chemistry extrapolated to SAM systems that are more complex than the simple ones described here.

’ EXPERIMENTAL SECTION Reagents. All reagents in these investigations were used as received: butane-1-thiol (C4SH, Acros Organics, 99+%), pentane-1-thiol (C5SH, Aldrich, 98%), octane-1-thiol (C8SH, Aldrich, 98.5%), decane-1-thiol (C10SH, Aldrich, 96%), dodecane1-thiol (C12SH, Acros Organics, 98.5+%), octadecane-1-thiol (C18SH, Aldrich, 98%), ammonium acetate (98% SigmaAldrich), and hydrogen peroxide (30%, Sigma-Aldrich). Alkanethiol solutions were freshly prepared in standard volumetric glassware using HPLC grade ethanol (Aldrich, 99.5%) or methanol (Pharmco, 99.9%) as the solvent. Prior to use, all volumetric glassware were cleaned in 0.51.0% v/v aqueous Hellmanex solutions (Hellma Worldwide) and were extensively rinsed afterward with ultrapure water (resistivity = 18.2 MΩ-cm) provided by a Millipore Synergy water system. Initial experiments employed Proton Sponge (PS = 1,8-bis(dimethylamino)naphthalene; Aldrich, 99%) to form deprotonated species (alkanethiolates) for detection. PS is a relatively strong, MS-friendly base (pKa = 12.34) that lacks the ionization suppression characteristics of typical strong inorganic bases. Despite its utility as a proton scavenger, PS tended to tarnish the ion transfer capillary tube and other components within the electrospray region, where its presence was signaled by a reddishbrown stain. The persistence of PS on the mass spectrometer components made cleaning these parts difficult and time-consuming, and thus, cleaning had to be performed frequently. The use of PS was eliminated in later experiments when H2O2 was used to oxidize all alkanethiols to alkylsulfonates. Since sulfonic acids are strong acids (pKa = 02), no additional proton scavenger was required. Standard Solutions. MS calibration curves were created to examine the effect of solution buffering and internal standard (IS) concentration on the linear dynamic range of our method, as well as the practical detection limits for some simple alkanethiols (as alkanesulfonates). Another goal of these studies was to investigate the importance of matrix effects when mixtures of alkanethiols are considered. C8SH and C10SH were chosen for this initial quantitative work based on the expectation that the linear dynamic range and detection limit should be similar for both. Sets of standard solutions were designated as either “binary” or “ternary” in nature, depending upon the number of structurally distinct alkanethiol components present. Since the IS was also an alkanethiol (C5SH), this designation includes the IS, as well. Initial quantitative investigations of binary alkanethiol mixtures began with solutions containing only C8SH and the IS. Sets of seven standard solutions were prepared by serial dilution of a stock solution containing 1.02 mM C8SH in ethanol. The stock solution was prepared with a 5-fold excess of H2O2 to promote complete oxidation of both alkanethiols to sulfonates. Concentrations of H2O2 significantly higher than this were avoided, since this often resulted in plugging of the 0.005-in.-i.d. PEEK tubing downstream from the injection port, suggesting that the excess oxidant may degrade its polymeric lining. Alkanethiol oxidation was monitored mass spectrally and usually took 23 days for completion. Sequential dilution of the stock solution was performed only after the peaks for both alkanethiolates and alkanesulfinates were no longer observed. Various sets of serially diluted standards differed with respect to buffer concentration

ARTICLE

(0 vs 1.0 mM ammonium acetate) and IS concentration (0.1 vs 1.0 mM). For those solutions containing an IS, a requisite quantity of neat C5SH was added to each solution following the serial dilution process to bring the IS concentration to either 0.10 or 1.0 mM. To matrix-match the serially diluted solutions, the ethanolic diluting solution contained H2O2 at the same concentration as the stock solution. For buffered solutions, the diluting solution contained 1.0 mM ammonium acetate, as well. Mass spectra were acquired over the m/z range necessary to observe both alkanesulfonates simultaneously for each injection. It was determined from these studies that optimal calibration data were generated using 1.0 mM C5SH and 1.0 mM ammonium acetate. These conditions were subsequently employed for binary solutions of C10 SH + IS as well as for ternary mixtures of C8SH, C10SH, and IS. For the ternary mixtures, a more extensive serial dilution scheme for creating standard solutions was devised. Details on this approach can be found in the Supporting Information (Table S-1). ESI-IT-MS Instrumentation and Conditions. MS data were acquired using a ThermoFinnigan LCQ Deca XP MAX quadrupole ion trap mass spectrometer fitted with an orthogonal ESI source running at room temperature. The ESI source operated using nitrogen (NM30LA Nitrogen Generator, Peak Scientific) as the drying/carrier gas. Qualitative studies of the various anionic product species were performed in negative ion mode by direct infusion at flow rates of 540 μL min1. Mass spectra were typically acquired by coadding 100 individual scans taken under optimized conditions. Tuning and mass calibration of the ESI system was performed using a solution containing caffeine, the tetrapeptide MRFA (Met-Arg-Phe-Ala), and the perfluoroalkyl triazine Ultramark 1621, dissolved in a 1% acetic acid acetonitrilemethanolwater mixture. Calibration curve data were acquired by performing 810 direct injections from either binary or ternary sets of standard solutions, as discussed above. Sample introduction into the ESI source was accomplished using a stainless steel 20.0 μL sample injection loop attached to the Deca XP’s integrated Valco divert/ inject valve, with ethanol continuously supplied from a Finnegan Surveyor Plus LC system at 40 μL/min. Mass spectra were normally acquired over a predetermined m/z range that encompassed the nominal masses of pentane-1-sulfonate (C5SO3, m/z 151), octane-1-sulfonate (C8SO3, m/z 193), and decane-1sulfonate (C10SO3, m/z 221). At the flow rates employed, each injection took ∼35 min, and thus, 810 injections from each external standard solution could be easily performed in less than an hour. After data collection, the integrated signals for C8SO3, C10SO3, and C5SO3 for each injection were determined from plots of the signal intensity vs time at their nominal m/z values using Xcalibur software. For each injection, the integrated signal for C8SO3 (or C10SO3) was ratioed to that for the IS, then averaged and plotted as the log (ratioed signal) vs the log (alkanethiol concentration). Separate MS tune files for both C8SO3 and C10SO3 were employed during data acquisition so that the data set for each species was acquired under its own optimized conditions. Given their similarity in chain lengths, the optimized conditions for each tune file were nearly identical (Supporting Information, Table S-2).

’ RESULTS AND DISCUSSION General Observations of Alkanethiol-Derived Species. Given that alkanethiols are relatively weak acids, one should be 9607

dx.doi.org/10.1021/ac202400h |Anal. Chem. 2011, 83, 9605–9613

Analytical Chemistry

ARTICLE

Figure 1. Negative ion ESI mass spectrum of: (A) a freshly prepared n-alkanethiol mixture, consisting of C4SH, C8SH, C10SH, C12SH, and C18SH (1 mM each) and Proton Sponge (PS = 25 mM) in methanol; (B) the alkanethiol mixture after prolonged oxidation by dissolved oxygen, resulting in the appearance of both sulfinates and sulfonates; (C) the alkanethiol mixture following oxidation with 25 mM H2O2, which results in only alkylsulfonates; and (D) the alkanethiol mixture following oxidation with 25 mM H2O2, showing peaks associated with various alkanesulfonate dimeric adducts (underlined). See Table 1 for all m/z peak assignments.

able to easily generate the singly charged alkanethiolate anions by the addition of a reasonably strong base. Proton Sponge (PS = 1,8-bis(dimethylamino)naphthalene, FW = 214.31) was initially chosen for this purpose because of its high pKa (∼12.34). The addition of PS to unbuffered solution mixtures of alkanethiols of various chain lengths was performed using a PS/alkanethiol ratio of 5:1 in either ethanol or methanol. The deprotonation of butane-1-thiol is shown below as an example of the expected process: CH3 ðCH2 Þ3 SH þ PS f CH3 ðCH2 Þ3 S þ PS-Hþ Figure 1A displays a negative ion ESI mass spectrum of a mixture of five n-alkanethiols (C4SH + C8SH + C10SH + C12SH + C18SH) that have been converted into alkanethiolates by PS. The nominal m/z ratios for each of the alkanethiolates present in this solution are shown in Table 1. Positive ion spectra of such solutions showed evidence for protonation of the PS, as an (M + 1) peak for PS at m/z = 215 was observed (Supporting Information, Figure S-1). Once deprotonated, however, alkanethiolates are readily oxidized by dissolved oxygen to alkanesulfinates (RSO2) and alkanesulfonates (RSO3), of which the latter are the most stable. Figure 1B displays a negative ion ESI mass spectrum taken from another solution having the same composition as that shown in Figure 1A, except that this data was acquired more than

one day following its preparation. Note that in this solution, the sulfinate and sulfonate features have become significantly more intense than the thiolate peaks (see Table 1 for m/z assignments). Eventually both the thiolate and sulfinate species disappear over the course of 34 days if the solution is not deoxygenated, leaving only the sulfonate derivatives. This general observation concerning the reaction of alkanethiolates with dissolved oxygen has been noted previously.34 From a quantitative standpoint, it is preferable to ensure that all the alkanethiol headgroups exist in only one chemical form; otherwise, one is forced to quantitatively account for three chemically distinct forms for each alkanethiol. The conversion of both the thiolates (1) and sulfinates (2) to the most stable form (sulfonates, 3) can be readily achieved using a MS-friendly oxidizer, such as H2O2: CH3 ðCH2 Þ3 S 1 CH3 ðCH2 Þ3 SO2  2

þ

3H2 O2

f

þ

H2 O2

f

CH3 ðCH2 Þ3 SO3  3 CH3 ðCH2 Þ3 SO3  3

þ

3H2 O

þ

H2 O

Figure 1C displays a typical negative ion ESI mass spectrum obtained following this procedure, in which the oxidant was allowed to react with the alkanethiol mixture for two days prior to data collection. Note the complete absence of thiolates and sulfinates following H2O2 oxidation. These results therefore 9608

dx.doi.org/10.1021/ac202400h |Anal. Chem. 2011, 83, 9605–9613

Analytical Chemistry

ARTICLE

Table 1. Nominal m/z Values of Alkanethiol-Derived Monomers and Alkanesulfonate Adductsa Monomers alkanethiolate

CnS

alkanesulfinate alkanesulfonate

CnSO2 CnSO3

C4SH

C8SH

C10SH

C12SH

C18SH

89

145

173

201

285

121 137

177 193

205 221

233 249

317 333

Dimeric Adducts

CnSO3

+

C18SO3

+

H+

=

Adduct

C18

333

+

333

+

1

=

667

C12

249

+

333

+

1

=

583

C10

221

+

333

+

1

=

555

C8

193

+

333

+

1

=

527

C4

137

+

333

+

1

=

471

CnSO3

+

C12SO3

+

H+

=

Adduct

C12

249

+

249

+

1

=

499

C10

221

+

249

+

1

=

471

C8

193

+

249

+

1

=

443

C4

137

+

249

+

1

=

387

CnSO3

+

C10SO3

+

H+

=

Adduct

C10

221

+

221

+

1

=

443

C8

193

+

221

+

1

=

415

C4

137

+

221

+

1

=

359

CnSO3

+

C8SO3

+

H+

=

Adduct

C8

193

+

193

+

1

=

387

C4

137

+

193

+

1

=

331

CnSO3

+

C4SO3

+

H+

=

Adduct

137

+

137

+

1

=

275

C4

Other Ions Observed

a

PS + H+

215

PS + Me+

229

PS PS + Me

214 229

PS + Et

243

PS = Proton SpongeTM; Me = methyl; Et = ethyl

indicate that alkanethiols can be easily converted into alkanesulfonates by oxidation with excess H2O2, which should simplify any subsequent quantitative analyses. Observation of Alkanesulfonate-Based Anionic Adducts. Our initial MS studies of unbuffered alkanesulfonate mixtures also provided evidence for the formation of anionic adducts of neutral alkanesulfonic acids with other alkanesulfonates to form gas-phase dimers, trimers, etc. Figure 1D displays a negative ion ESI mass spectrum in the m/z 250650 range of a solution containing a mixture of the five n-alkanesulfonates (C4SO3 + C8SO3 + C10SO3 + C12SO3 + C18SO3), showing peaks that can be attributed to the formation of dimeric anionic adducts (CnSO3 + CmSO3 + H+), in which the masses of the adducts are underlined; Table 1 lists the MS assignments for these species. Observations were also made at higher m/z of trimeric anionic adducts (i.e., CnSO3 + CmSO3 + ClSO3 + 2H+) as well as of mixed sulfonatesulfinate adducts from solutions that were only partly oxidized and contained both species (data not

shown). In general, MS intensities for the dimeric adducts were significantly weaker (e10%) than those for monomers, and trimeric adducts were significantly weaker than those for the dimers. Later quantitative studies of alkanesulfonate mixtures buffered with 1.0 mM ammonium acetate have produced little or no evidence of gas-phase adduct formation, however. Even though the pKa’s of alkanesulfonic acids are low enough to expect their near complete ionization at neutral pH’s, their appearance as anionic adducts only in the unbuffered solutions indicates that significant changes in solution pH may be occurring during the electrospray process. In solutions buffered with ammonium acetate, the extremely weak basicity of the resulting alkanesulfonate species is not sufficient to effectively compete for protons with acetate ion, which is the stronger base; hence, the absence of adduct formation. Our observations from solutions lacking such a competitive base suggest that solution pH may be an important factor in contributing to adduct formation during the electrospray process. However, the relationship between solution pH and analyte response in negative ion mode is not clearly understood, and various studies have evidenced trends that oppose simple expectations.37 For example, ionization suppression of deprotonated species has been found to occur at high pH in negative ion mode and that in some cases enhancement of the ESI response for deprotonated species occurs even under acidic conditions. Explanations for such behavior are diverse, ranging from proton transfer reactions that occur in the gas-phase but not within the droplets themselves to changes in the droplet pH during solvent evaporation.38 Although the observation of gas-phase cationic adducts is wellknown (e.g., Na+ and K+ adducts of many neutral biomolecules), the observation of anionic adducts in mass spectrometry is less common, although not unusual.3944 This difference has been attributed both to weaker thermodynamic binding constants between the components of anionic adducts and to an increase in the rate of charge recombination via electrical discharge effects near the metal capillary.40,45 ESI-IT-MS intensities of the dimeric adducts observed in this work were usually an order of magnitude smaller than those for the monomers. This is fortunate, since the presence of adduct species complicates quantitative analyses of the monomer species. When adducts are present, accurate quantification of a given alkanesulfonate requires that one account for the intensity of the monomeric species as well as every adduct species containing that particular alkanesulfonate. This leads to a more laborious quantification scheme, since the intensities of several different species containing the same alkanesulfonate must be tracked and tallied. In addition, direct determination of the gasphase adduct concentrations is difficult, and there is no straightforward means by which their intensities can be used to gauge their concentrations relative to the monomers. Furthermore, it can be expected that larger gas-phase concentrations of adducts will tend to reduce peak intensities for the monomers and, thus, lead to poorer detection limits for those species. Consequently, approaches that can reduce the rate of alkanesulfonate adduct formation in the gas phase are preferred for quantitative work. Instrumental Factors Affecting ESI-IT-MS Peak Intensities. To provide accurate quantitative information of alkanesulfonate concentrations, an understanding of the effect of several instrumental factors on the MS peak intensities is required. In this regard, we have found their intensities exhibit a profound dependence on both the ion transfer capillary temperature (Tcap) and ESI source voltage (Vs). Initial studies of the effect of Vs 9609

dx.doi.org/10.1021/ac202400h |Anal. Chem. 2011, 83, 9605–9613

Analytical Chemistry

ARTICLE

Figure 2. Effect of Tcap on the absolute intensities of various alkanesulfonate (A) monomers and (B) dimeric adducts generated by negative ion ESI-IT-MS. Nominal m/z assignments are indicated in the legends. For clarity, only homodimers are shown. Heterodimers exhibit similar behavior and are displayed in Figure S-3 of the Supporting Information. ESI spray voltage = 4.5 kV; flow rate = 30 μL/min. (C) Plot showing dependence of Tcap on mass for both the monomers and dimeric adducts. For the latter data, the outlier (m/z 499) is not included in the regression. 

(Supporting Information, Figure S-2). All subsequent work therefore employed relatively high source voltages (4.04.5 kV). Figure 2A shows the effect of Tcap on monomer peak intensities. Note that this effect is dependent on alkyl chain length such that each chain length exhibits an optimal capillary temperature (Tcap,opt) which shifts to higher temperature as the chain length is increased. In fact, within the range of Tcap’s allowed by the instrument (e400 C), Tcap,opt for octadecane-1-sulfonate is not attained; thus most of the C18SO3 may be depositing within the capillary or near its inlet. Note that as Tcap is initially raised above 150 C, all alkanesulfonate intensities increase together, which likely indicates an increase in their rate of desorption and transfer through the heated capillary to the ion trap. In fact, in the temperature range 150 C < Tcap < ∼240 C, monomer intensities scale inversely with alkyl chain length such that shorter chain species exhibit greater MS intensities than the longer chain homologues. Such behavior is inconsistent with arguments based on surface activity effects, since higher intensities would then be expected for the longer chain species. In addition, we are operating with a room-temperature ESI source, and it is the ion transfer capillary that is heated. Thus, the temperature behavior we observe is related to the gas-phase processes occurring in the heated capillary and not to surface activity effects at the ESI droplets. At Tcap > 240 C, the MS intensity for C4SO3 drops below that for C8SO3, C10SO3, and C12SO3, whereas at Tcap > 270 C, the MS intensity for C8SO3 drops below that for C10SO3 and C12SO3. This observation indicates that some other mechanism is reducing the alkanesulfonate MS intensities at higher temperature; otherwise, the intensity of each species would be expected to approach a relatively constant value as the capillary temperature is raised. Furthermore, the decrease in monomer intensities at temperatures higher than their respective Tcap,opt values appears to correlate with an increase in the intensity of the sulfur trioxide radical anion ( 3 SO3, m/z 80). This species presumably arises from homolytic cleavage of the CS bond, with the effect that shorter chain alkanesulfonates generate larger MS intensities for 3 SO3 than do the longer chain homologues at high capillary temperatures. These observations are consistent with previous studies of alkanethiol films which show evidence for CS bond cleavage46 and that upon heating, the rate of cleavage is enhanced.47,48 Although the CS bond strength is not expected to be significantly different between the alkanesulfonates studied here, it is expected that entropic effects should lead to lower thermal decomposition temperatures for short-chain species relative to the longer-chain homologues. In other words, shorter chain alkanesulfonates have fewer vibrational degrees of freedom (VDOF) available for distributing thermal energy, and therefore, a larger fraction of that energy ends up localized in the CS bond, leading to greater rates of thermal decomposition for those species. This expectation is borne out by the trend in Tcap,opt’s presented in Figure 2A. Furthermore, all alkanesulfonate mass spectra also show the presence of bisulfate ion (HSO4, m/z 97) (e.g., see Figure 1C). Bisulfate may result from the second-order decay reaction of 3 SO3, which produces sulfur trioxide and sulfite ion (SO32, m/z 40):49 3 SO3



on the mixture of five alkanesulfonates (C4SO3 + C8SO3 + C10SO3 + C12SO3 + C18SO3, 1.0 mM each) showed the same general effect on the all the alkansulfonates examined: peak intensities increased with increasing Vs, up to a maximum at ∼4 kV



þ 3 SO3  f SO3 þ SO3 2

both of which are not experimentally accessible, since one species is neutral and the other is outside the accessible m/z range of our mass spectrometer. Subsequent oxidation of sulfite to sulfate by 9610

dx.doi.org/10.1021/ac202400h |Anal. Chem. 2011, 83, 9605–9613

Analytical Chemistry the excess H2O2 present in solution, followed by protonation, may lead to the bisulfate MS signal that is observed: SO3 2 þ H2 O2 f SO4 2 þ H2 O SO4 2 þ H2 O f HSO4  þ OH The effect of thermal decomposition on the alkanesulfonate calibration data is discussed below. Figure 2B displays the ESI-IT-MS intensities of dimeric adducts as a function of Tcap taken at high Vs (4.5 kV) from a mixture of the same five alkanesulfonates studied previously, in which only data for the homodimers is shown (2CnSO3 + H+, Table 1). Heterodimer behavior is similar to that for the homodimers, but is not displayed to preserve clarity in presentation of the data (see the Supporting Information, Figure S-3 for a plot of both species). Like the monomers, all adducts evidence a Tcap,opt that shifts to higher values with increasing adduct mass. Greater MS intensities are observed for adducts in the middle of the mass range (m/z 387, 443, 471; Supporting Information Figure S-3 ), probably because these species arise from more than one possible combination of monomers (Table 1). The temperature behavior of both hetero- and homodimers is more clearly seen in Figure 2C, which displays Tcap,opt’s for all adducts as a function of their m/z values and in comparison with those observed for the monomers. Note that adduct Tcap,opt’s increase linearly with adduct mass and that all homodimers exhibit Tcap,opt’s that are 1545 C lower than their corresponding monomers. This latter observation suggests that the decreasing MS intensities observed for the homodimers above their respective Tcap,opt’s may be preferentially driven by intermolecular dissociative processes rather than by thermal decomposition via CS bond cleavage. In addition, all heterodimers exhibit Tcap,opt’s that fall between those for the monomers from which they are constituted. For example, the thermal stabilities of C4SO3-containing heterodimers are larger than the homodimer and increase with the size of the adduct partner (Figure 2C, Table 1). This is reasonable, given that larger adducts have more VDOF available to distribute thermal energy. However, it has been difficult to ascertain from these data the contribution(s) of various mechanisms (e.g., thermally driven dissociation vs thermal decomposition) to the observed decrease in adduct intensities above their respective Tcap,opt’s. This is primarily because adduct intensities were significantly weaker than the monomers, which made it difficult to assess the extent to which adduct loss increased monomer intensities. Furthermore, more than one possible combination of monomers gave rise to the same adduct m/z for several species, thereby complicating intensity analyses. Future experiments are planned to examine in more detail the thermal behavior of individual adducts. Mass Spectral Calibration Data: General Considerations. Quantitative measurements in ESI-IT-MS are often hampered by ionization suppression of the analytes, which can be particularly problematic when the sample contains other polar or ionic compounds that compete for charge during the processes of desolvation and ion extraction into the instrument. Ionization suppression leads not only to a loss of absolute sensitivity for the analyte of interest, but also to a reduction in the accuracy and precision of the measured signal. For quantitative measurements of binary mixed SAMs, which is the ultimate goal of this work, the effects of ionization suppression are expected to be small for two primary reasons. First, relatively simple solution matrixes are

ARTICLE

being used that contain only ESI-compatible solvents and buffers. Second, all chemisorbed alkanethiols will be converted directly to an ionic form (alkanesulfonates) upon removal from the surface. So long as the total mass at one time in the source is not so large that the compounds cannot be reproducibly desolvated and extracted into the instrument, ionization suppression is not expected to be significant. Internal standard calibration methods have been used here since they have been shown to lead to linear dynamic ranges, extending over 24 orders-ofmagnitude change in the concentration of anions.50,51 In our work, an odd chain length species (pentane-1-thiol, C5SH) was chosen as the IS to avoid the expense of using isotopically labeled compounds. Like the other alkanethiols, pentane-1-sulfonate is produced by the excess H2O2 and is identified and quantified separately from the other alkanesulfonates. Our quantitative investigations have focused on either binary or ternary alkanesulfonate mixtures, in which binary mixtures are those consisting of a single alkanesulfonate + IS and ternary mixtures are those consisting of two different alkanesulfonates + IS. Thus, the binary mixture data should be useful for quantifying single-component SAMs by ESI-IT-MS, whereas the ternary mixture data would be useful for quantifying binary mixed SAMs. Initially, the effects of IS concentration (1.0 vs 0.10 mM) and of a volatile ESI-compatible buffer (ammonium acetate) on the linear dynamic range were examined using the binary mixtures. Optimal conditions (IS concentration, buffer) in the binary systems were subsequently employed for the ternary mixtures. In addition, all quantitative data were obtained with a 20.0 μL injection loop using an injection process that optimized signal reproducibility; the integrated areas for each injection were determined and plotted for each alkanesulfonate (see Supporting Information, Figure S-4). High precision should allow us to distinguish among solutions whose alkanesulfonate compositions only slightly differ. Binary Mixtures of Alkanesulfonates. ESI-IT-MS data from standard solutions containing binary mixtures of a single alkanesulfonate + IS has yielded high quality calibration data. Although these standards are not representative of the solution compositions that result from the oxidative removal of binary mixed SAMs into solution, they were instructive in that they allowed us to examine the effects of varying both the IS concentration and buffer concentration. The data shown below were acquired under optimal conditions of IS and buffer concentration (1.0 mM each). Calibration data for C8SO3 were measured from a serially diluted set of buffered ethanolic C8SH standard solutions containing 1.0 mM C5SH as an IS (Supporting Information, Figure S-5). Concentrations below ∼ 1.0 μM could not be reproducibly measured using our injection technique; therefore, concentrations near this value were taken to be representative of the practical detection limits for this compound by our method. The linear dynamic range of the technique covered at least 3 orders of magnitude in alkanethiol concentration (1.0 μM to 1.0 mM), and the measurement variation was quite small. Linear regression of binary mixtures containing C8SO3 yielded the following best-fit parameters: y = 1.0618x + 8.9292, R2 = 0.9976. An analogous calibration plot was generated for binary mixtures containing C10SO3, which were created using the same serial dilution scheme, IS concentration, buffer concentration, and MS experimental conditions (Supporting Information, Figure S-5). The linear dynamic range and quality of fit are comparable to that for C8 SO 3 , demonstrating that this approach can easily 9611

dx.doi.org/10.1021/ac202400h |Anal. Chem. 2011, 83, 9605–9613

Analytical Chemistry be extended to other chain length alkanethiols (for C10SO3, y = 1.0715x + 9.1383, R2 = 0.9949). Note that there is a near 1:1 correspondence between the ratio of the integrated signals and the alkanethiol concentration. Despite the similarities in the slopes of the two curves, the ratio S C10SO 3/S IS is slightly larger than SC8SO 3/SIS over the entire concentration range, resulting in a slightly higher y-intercept for the C10SO3 data. Given that these data were acquired at Tcap = 350 C (which is close to Tcap,opt for both species) and that thermal decomposition at these high temperatures appears to proceed via CS bond cleavage, it is likely that this difference in y-intercepts is related to slight differences in the thermal stabilities of the two compounds. Figure 2A shows that C10SO3 exhibits larger absolute MS intensities than C8SO3 at this capillary temperature, even if injected from solutions that contain equimolar quantities. Ternary Mixtures of Alkanesulfonates. MS calibration data obtained from ternary mixtures of alkanesulfonates have also been examined, as these most closely replicate the solution compositions expected when binary component SAMs are oxidatively stripped into a solution containing C5SH as an IS. Initial investigations of ternary solutions have focused on mixtures of similar chain length, CH3-terminated alkanethiols, since the composition of binary SAMs formed from these components is expected to closely mirror the solutions from which they deposit. Calibration data for C8SO3 and C10SO3, respectively, were measured from a single, serially diluted set of ethanolic standard solutions containing both C8SH and C10SH, in addition to 1.19 mM C5SH as an IS (Supporting Information, Figure S-6). Concentrations below 2.0 μM could not be reproducibly measured for either alkanesulfonate, leading to slightly shorter dynamic ranges than those for the binary mixtures. Linear regression of the ternary mixtures yielded the following best-fit parameters: for C8SO3, y = 1.1964x + 9.2109 (R2 = 0.9954); for C10SO3, y = 1.2914x + 10.3778 (R2 = 0.9980). Note that these parameters are larger than those obtained for the corresponding species in the binary mixtures. In particular, the calibration sensitivities associated with the ternary mixtures are 1320% larger than their respective values for the binary mixtures, which is not an insignificant difference. Despite these differences, the quality of the calibration data is largely the same compared to the binary systems. Therefore, the addition of a third alkanesulfonate into binary mixtures of these species does appear to have a significant, measurable effect on their calibration data. Understanding these effects will be of great importance if ESI-IT-MS is to be utilized properly to quantify mixed SAM films.

’ CONCLUSIONS The studies reported here have focused on finding optimal experimental conditions for generating stable negative ion ESI mass spectra of alkanethiol-derived species for quantitative purposes. Purposeful oxidation of alkanethiolates and alkanesulfinates to their fully oxidized form (alkanesulfonates) using H2O2 greatly simplifies quantification of these species, since the sulfonate headgroup is stable and has little affinity for covalent binding to metal surfaces. ESI-IT-MS has provided evidence for the formation of not only the monomeric alkanesulfonate product species, but also the anionic adducts of alkanesulfonates with alkanesulfonic acids, resulting in gas-phase dimers and trimers. The MS peak intensities for ESI-generated monomers

ARTICLE

exhibit a significant dependence on both the ESI source voltage and the ion transfer capillary temperature. The capillary temperature has a significant effect on monomer intensities such that each chain length exhibits an optimal capillary temperature that shifts to higher temperature as the chain length is increased. This behavior appears to be correlated with CS bond cleavage due to the appearance of both sulfur trioxide radical anion and bisulfate ion in the mass spectra. The trend in optimal capillary temperatures with increasing chain length indicates that entropic effects lead to lower thermal decomposition temperatures for shortchain species relative to the longer chain homologues. MS calibration data from binary and ternary mixtures of mediumchain-length alkanesulfonates are characterized by linear dynamic ranges extending from 103 to 106 M and by a high degree of precision. The detection limits for all alkanesulfonates appear to be dictated by their rates of thermal decomposition via CS bond cleavage. The quantification of both single- and multicomponent SAMs on Au by ESI-IT-MS using this approach will be the subject of a future publication.

’ ASSOCIATED CONTENT

bS

Supporting Information. Additional information as noted in text. This material is available free of charge via the Internet at http://pubs.acs.org.

’ AUTHOR INFORMATION Corresponding Author

*Fax: (205) 726-2479. E-mail: [email protected]. Present Addresses †

College of Medicine, University of South Alabama, Mobile, AL 36688 ‡ Department of Chemistry, University of Georgia, Athens, GA 30602 § Southview Middle School, Tuscaloosa, AL 35405 ^ McWhorter School of Pharmacy, Samford University, Birmingham, AL 35229

’ ACKNOWLEDGMENT This work was supported by a Major Research Instrumentation grant from the National Science Foundation (Grant No. 0619217). Acknowledgment is also made to the Donors of the American Chemical Society Petroleum Research Fund for partial support of this research. Support from both the Samford University Summer Undergraduate Research Program and the Alabama Power Research Fellowship Program, which is administered through the Samford University Fellows Program, is also gratefully acknowledged. The authors are indebted to Dr. Greg Gorman (McWhorter School of Pharmacy) and Drs. David Garza and Andrew Lampkins (Dept. of Chemistry & Biochemistry) for useful discussions during the evolution of this project. Student authors on this article are listed in chronological order with regard to their participation in this work. ’ REFERENCES (1) Hanley, L.; Kornienko, O.; Ada, E. T.; Fuoco, E.; Trevor, J. L. J. Mass Spectrom. 1999, 34, 705–723. (2) Hagenhoff, B.; Benninghoven, A.; Spinke, J.; Liley, M.; Knoll, W. Langmuir 1993, 9, 1622–1624. 9612

dx.doi.org/10.1021/ac202400h |Anal. Chem. 2011, 83, 9605–9613

Analytical Chemistry (3) Offord, D. A.; John, C. M.; Griffin, J. H. Langmuir 1994, 10, 761–766. (4) Offord, D. A.; John, C. M.; Linford, M. R.; Griffin, J. H. Langmuir 1994, 10, 883–889. (5) Beulen, M. W. J.; Huisman, B.-H.; van der Heijden, P. A.; van Veggel, F. C. J. M.; Simons, M. G.; Biemond, E. M. E. F.; de Lange, P. J.; Reinhoudt, D. N. Langmuir 1996, 12, 6170–6172. (6) Fisher, G. L.; Walker, A. V.; Hooper, A. E.; Tighe, T. B.; Bahnck, K. B.; Skriba, H. T.; Reinard, M. D.; Haynie, B. C.; Opila, R. L.; Winograd, N.; Allara, D. L. J. Am. Chem. Soc. 2002, 124, 5528–5541. (7) Wolf, K. V.; Cole, D. A.; Bernasek, S. L. Anal. Chem. 2002, 74, 5009–5016. (8) Sun, L.; Gardella, J. A., Jr. Langmuir 2002, 18, 9289–9295. (9) Arezki, B.; Delcorte, A.; Chami, A. C.; Garrison, B. J.; Bertrand, P. Nucl. Instrum. Methods Phys. Res., Sect. B 2003, 212, 369–375. (10) Arezki, B.; Delcorte, A.; Bertrand, P. Appl. Surf. Sci. 2004, 231/ 232, 122–126. (11) Sohn, S.; Schr€oder, M.; Lipinsky, D.; Arlinghaus, H. F. Surf. Interface Anal. 2004, 36, 1222–1226. (12) Trevor, J. L.; Hanley, L.; Lykke, K. R. Rapid Commun. Mass Spectrom. 1997, 11, 587–598. (13) Trevor, J. L.; Mencer, D. E.; Lykke, K. R.; Pellin, M. J.; Hanley, L. Anal. Chem. 1997, 69, 4331–4338. (14) Trevor, J. L.; Lykke, K. R.; Pellin, M. J.; Hanley, L. Langmuir 1998, 14, 1664–1673. (15) Hanley, L.; Zimmermann, R. Anal. Chem. 2009, 81, 4174–4128. (16) Gong, W.; Elitzin, V. I.; Janardhanam, S.; Wilkins, C. L.; Fritsch, I. J. Am. Chem. Soc. 2001, 123, 769–770. (17) Mouradian, S.; Nelson, C. M.; Smith, L. M. J. Am. Chem. Soc. 1996, 118, 8639–8645. (18) Fukuo, T.; Monjushiro, H.; Hong, H.; Haga, M.; Arakawa, R. Rapid Commun. Mass Spectrom. 2000, 14, 1301–1306. (19) Su, J.; Mrksich, M. Angew. Chem., Int. Ed. 2002, 41, 4715–4718. (20) Su, J.; Mrksich, M. Langmuir 2003, 19, 4867–4870. (21) Yeo, W.-S.; Mrksich, M. Langmuir 2006, 22, 10816–10820. (22) Patrie, S. M.; Mrksich, M. Anal. Chem. 2007, 79, 5878–5887. (23) Mrksich, M. ACS Nano 2008, 2, 7–18. (24) Ha, T. K.; Lee, T. G.; Song, N. W.; Moon, D. W.; Han, S. Y. Anal. Chem. 2008, 80, 8526–8531. (25) Vorm, O.; Roepstorff, P.; Mann, M. Anal. Chem. 1994, 66, 3281–3287. (26) Duncan, M. W.; Roder, H.; Hunsucker, S. W. Briefings Funct. Genomics Proteomics 2008, 7, 355–370. (27) Griesser, H.; Kingshott, P.; McArther, S. L.; McLean, K. L.; Kinsel, G. R.; Timmons, R. B. Biomaterials 2004, 25, 4861–4875. (28) McLean, K. M.; McArthur, S. L.; Chatelier, R. C.; Kingshott, P.; Griesser, H. J. Colloids Surf., B 2000, 17, 23–35. (29) Tarlov, M. J.; Newman, J. G. Langmuir 1992, 8, 1398–1405. (30) Cooper, E.; Leggett, G. J. Langmuir 1998, 14, 4795–4801. (31) Li, Y.; Huang, J.; McIver, R. T., Jr.; Hemminger, J. C. J. Am. Chem. Soc. 1992, 114, 2428–2432. (32) Scott, J. R.; Baker, L. S.; Everett, W. R.; Wilkins, C. L.; Fritsch, I. Anal. Chem. 1997, 69, 2636–2639. (33) Kornienko, O.; Ada, E. T.; Tinka, J.; Wijesundara, M. B. J.; Hanley, L. Anal. Chem. 1998, 70, 1208–1213. (34) McCarley, T. D.; McCarley, R. L. Anal. Chem. 1997, 69, 130–136. (35) Yang, W.; Zhang, R.; Willett, G. D.; Hibbert, D. B.; Gooding, J. J. Anal. Chem. 2003, 75, 6741–6744. (36) B€okman, C. F.; Zettersten, C.; Sj€oberg, P. J. R.; Nyholm, L. Anal. Chem. 2004, 76, 2017–2024. (37) Cech, N. B.; Enke, C. G. In Electrospray and MALDI Mass Spectrometry: Fundamentals, Instrumentation, Practicalities, and Biological Applications, 2nd ed.; Cole, R. B., Ed.; Wiley & Sons: Hoboken, NJ, 2010; Chapter 2 and references therein. (38) We have investigated these same solutions using an atmospheric pressure chemical ionization (APCI) source to generate ions. In APCI mode, proton transfer reactions tend to be the most common

ARTICLE

gas-phase processes occurring, leading to significant quantities of (M + H)+ ions. In negative ion APCI, we have observed significant quantities of H+-bound dimeric adducts. In fact, the dimer/monomer intensity ratios observed by APCI were often significantly larger than those observed by ESI. Under some instrumental conditions with APCI, the intensities of the dimeric adducts were observed to be as large as those for the monomers. This is an interesting result, since the observation of gas-phase anionic supramolecular complexes by APCI is generally not as widespread as that for ESI. Despite the large differences in dimer/ monomer intensity ratios observed between the two ionization sources, the relative intensities of the dimeric peaks were remarkably similar, regardless of the ionization source used. This result, along with the larger dimer intensities observed by APCI, suggests that adduct formation occurs primarily in the gas phase, where the effects of solution pH will tend to be less relevant. (39) Zhu, J.; Cole, R. B. J. Am. Soc. Mass Spectrom. 2000, 11, 932–941. (40) Cai, Y.; Cole, R. B. Anal. Chem. 2002, 74, 985–991. (41) Wu, Z.; Gao, W.; Phelps, M. A.; Wu, D.; Miller, D. D.; Dalton, J. T. Anal. Chem. 2004, 76, 839–847. (42) Thomas, M. C.; Mitchell, T. W.; Blanksby, S. J. J. Am. Soc. Mass Spectrom. 2005, 16, 926–939. (43) Song, L.; Wellman, A. D.; Yao, H.; Bartmess, J. E. J. Am. Soc. Mass Spectrom. 2007, 18, 1789–1798. (44) McEwen, C. N.; Larsen, B. S. J. Am. Soc. Mass Spectrom. 2009, 20, 1518–1521. (45) Kavallieratos, K.; Sabucedo, A. J.; Pau, A. T.; Rodriguez, J. M. J. Am. Soc. Mass Spectrom. 2005, 16, 1377–1383. (46) Zhong, C.-J.; Porter, M. D. J. Am. Chem. Soc. 1994, 116, 11616– 11617. (47) Cheng, L.; Bernasek, S. L.; Bocarsly, A. B.; Ramanarayanan, T. A. Chem. Mater. 1995, 7, 1807–1815. (48) Lim, H.; Carraro, C.; Maboudian, R.; Pruessner, M. W.; Ghodssi, R. Langmuir 2004, 20, 743–747. (49) Neta, P.; Huie, R. E. Environ. Health Perspect. 1985, 64, 209–217. (50) Barnett, D. A.; Horlick, G. J. Anal. At. Spectrom. 1997, 12, 497–501. (51) Li, Y.; George, E. J. Anal. Chem. 2005, 77, 4453–4458.

9613

dx.doi.org/10.1021/ac202400h |Anal. Chem. 2011, 83, 9605–9613