Article pubs.acs.org/ac
Vacuum Ultraviolet Detector for Gas Chromatography Kevin A. Schug,*,† Ian Sawicki,† Doug D. Carlton, Jr.,† Hui Fan,† Harold M. McNair,‡ John P. Nimmo,† Peter Kroll,† Jonathan Smuts,§ Phillip Walsh,§ and Dale Harrison§ †
Department of Chemistry & Biochemistry, The University of Texas at Arlington, 700 Planetarium Place, Box 19065, Arlington, Texas 76019, United States ‡ Department of Chemistry, Virginia Tech, Blacksburg, Virginia 24061, United States § VUV Analytics, Inc., Austin, Texas 78717, United States S Supporting Information *
ABSTRACT: Analytical performance characteristics of a new vacuum ultraviolet (VUV) detector for gas chromatography (GC) are reported. GC-VUV was applied to hydrocarbons, fixed gases, polyaromatic hydrocarbons, fatty acids, pesticides, drugs, and estrogens. Applications were chosen to feature the sensitivity and universal detection capabilities of the VUV detector, especially for cases where mass spectrometry performance has been limited. Virtually all chemical species absorb and have unique gas phase absorption cross sections in the approximately 120−240 nm wavelength range monitored. Spectra are presented, along with the ability to use software for deconvolution of overlapping signals. Some comparisons with experimental synchrotron data and computed theoretical spectra show good agreement, although more work is needed on appropriate computational methods to match the simultaneous broadband electronic and vibronic excitation initiated by the deuterium lamp. Quantitative analysis is governed by Beer−Lambert Law relationships. Mass on-column detection limits reported for representatives of different classes of analytes ranged from 15 (benzene) to 246 pg (water). Linear range measured at peak absorption for benzene was 3−4 orders of magnitude. Importantly, where absorption cross sections are known for analytes, the VUV detector is capable of absolute determination (without calibration) of the number of molecules present in the flow cell in the absence of chemical interferences. This study sets the stage for application of GC-VUV technology across a wide breadth of research areas.
G
synchrotron facilities. High intensity sources were needed to overcome significant background absorption. In fact, essentially every chemical compound absorbs strongly in the VUV spectral range (115−185 nm). Photons in this regime are able to probe the excitation of atomic species and virtually any chemical bond, especially σ → σ* and short wavelength high probability π → π*, which cannot typically be probed in traditional UV/vis absorption spectroscopy. Prior efforts aimed at the development and integration of UV and VUV absorption detection with GC have been limited.1−4 A far UV detector (FUVD) was first combined with capillary GC in 1987 by Zlatkis and co-workers.1 It was limited to probing essentially a single wavelength absorption at 122 nm (10.2 eV), which is sufficient for detection of a wide variety of chemical species including alkanes, but provides no qualitative information. A more energetic photoionization detector (PID) (106 nm; 11.7 eV) was also evaluated in that study. Originally commercialized in 1976, the higher energy PID showed less interference from compounds, such as water compared to the 10.2 eV FUVD. Additional work was performed on FUVD and
as chromatography (GC) is a mature and highly reliable technique for the analysis of complex mixtures of volatile and semivolatile compounds. Myriad associated column chemistries, injection techniques, and detectors provide for good selectivity and sensitivity to address a multitude of problems in environmental, energy, clinical, food, and pharmaceutical research. GC combined with mass spectrometry detection (GC-MS) currently provides the greatest breadth of performance in terms of universal detection of chemical compounds, sensitivity for trace quantitative analysis, and capability for qualitative identification of chemical constituents. Many official standardized methods rely on GC-MS technology. Even so, some limitations exist where isomeric, isobaric, small, or labile chemical compounds are desired to be analyzed. Where the mass spectrometer fails to differentiate some species (e.g., meta- and para-xylenes and some polycyclic aromatic hydrocarbons), special (and potentially, expensive) column chemistries must be used to achieve chromatographic resolution for precise speciation. To address some known limitations of GC-MS and to offer new capabilities and efficiencies for the application of GC, a benchtop vacuum ultraviolet detector (VUV) has been recently developed and tested. Historically, the measurement of VUV absorption spectra has been restricted to bright source © 2014 American Chemical Society
Received: May 16, 2014 Accepted: July 31, 2014 Published: July 31, 2014 8329
dx.doi.org/10.1021/ac5018343 | Anal. Chem. 2014, 86, 8329−8335
Analytical Chemistry
Article
Figure 1. Schematic (not to scale) of the GC-VUV instrument. The dimensions of the VGA-100 are 13 in. (w) × 30 in. (d) × 17 in. (h). Flow cell volume is ∼80 μL. The path length is 10 cm.
can be set as high as 100 Hz, making measurements compatible with fast GC applications. Following residence in the flow cell, the analyte zone is swept out through the exit vent. Absorption data are sent to the data station for processing. The objective of this work was to show how VUV spectroscopy provides a uniquely versatile, universal detector to support highly sensitive quantitative and qualitative GC analysis. A breadth of applications are highlighted including the analysis of fixed gases, alkanes, polyaromatic hydrocarbons, pesticides, steroid hormones, drug compounds, and fatty acids; applications particularly troublesome for GC-MS are emphasized. Limits of detection determined experimentally for a variety of common compounds are shown. Also shown is the potential for complementary computational deconvolution and spectral matching tools to enhance experimental measurements.
PID by Driscoll and co-workers in the late 1980s and early 1990s.2 Broadband UV absorption instrumentation capable of quantitative and qualitative analysis was reported in the 2000s by Lagesson and co-workers.3,4 Full scan spectra (168−330 nm) of eluted peaks could be recorded by the GC-UV system, allowing many heteroatom-containing and conjugated molecules to be identified based on unique gas phase absorption spectra. However, this system could not provide universal detection and transmittance of light in the shorter wavelengths of this range (e.g., 168−180 nm) was quite limited. Figure 1 provides a general schematic of the instrumental arrangement and operation principles of the GC-VUV instrument. Chemical compounds eluting from the gas chromatograph enter a heated transfer line (300 °C, typically), which incorporates a length of uncoated deactivated glass capillary. The VUV detector owns the transfer line and can be connected to any standard gas chromatograph through a punch-out in the GC oven casing. At the end of the transfer line, a makeup flow of carrier gas is introduced. This makeup flow can be used to alter the residence time of the sample zone in the detector cell. The analyte stream then reaches the 10 cm path length (80 μL volume) flow cell. Critical to the ability to collect high quality VUV (and some UV) absorption data between 115−240 nm, is the use of specially coated reflective optics and a back-thinned charged coupled device (CCD) light path monitor to simultaneously assess absorption features across the spectrum for peaks eluting from the GC column. At the beginning of the run, a dark noise reading is taken for background subtraction. As with conventional liquid phase UV, the VUV detector is mass-sensitive; this means that the detector response is proportional to the amount of analyte present per unit time. Thus, some advantages from signal averaging can be obtained by altering the residence time of the analyte in the flow cell. As seen below, gas phase absorption spectra are much more highly featured than typical liquid phase UV absorption spectra. Data acquisition rate for the GC-VUV
■
EXPERIMENTAL SECTION Two VGA-100 VUV detectors (VUV Analytics, Inc., Austin TX), one coupled to a Shimadzu GC-2010 gas chromatograph (Shimadzu Scientific Instruments, Inc., Columbia MD) and one coupled with a PerkinElmer Autosystem XL gas chromatograph (PerkinElmer, Inc., Waltham MA), were used to collect the data presented. A variety of columns and operational conditions were used in these systems, the full details of which are listed in the Supporting Information. Columns included a SGE BPX-5 (14 m × 250 μm i.d. × 0.25 μm df) (SGE Analytical Science, Ringwood, Australia), a Restek Rxi-1 ms (30 m × 250 μm i.d. × 0.25 μm df) (Restek Corporation, Bellefonte PA), and a Phenomenex ZB-5 (15 m x 250 μm i.d. x 0.5 μm df) (Phenomenex, Inc., Torrance CA). Column flow rates (helium carrier gas for the Shimadzu GC and nitrogen carrier gas for the PerkinElmer GC) ranging from 0.7−1.0 mL/min, split ratios ranging from 20:1−250:1, and both isothermal and temperature programming oven conditions were used. Injection volumes delivered by autosamplers on both instruments ranged from 0.1−0.5 μL. The VUV detector and accompanying 8330
dx.doi.org/10.1021/ac5018343 | Anal. Chem. 2014, 86, 8329−8335
Analytical Chemistry
Article
Figure 2. GC-VUV analysis of gasoline. A chromatographic separation of gasoline (A) reveals absorption signals for a variety of chemical compounds. Select spectra for ethanol (B), some alkanes (C), and some aromatic compounds (D) are shown. GC-FID and GC-MS chromatograms of this gasoline sample can be found in the Supporting Information.
transfer line were thermostated at 300 °C. The detector scan time was varied between 100−400 ms (determined to be optimal for providing minimal noise and good peak coverage). Chemicals for analysis were acquired from various sources. 1Naphthol, 2-naphthol, benzene, captan, octane, nicotine, 17αestradiol, and 17β-estradiol were obtained from Sigma-Aldrich (St. Louis, MO USA). para-Xylene and ortho-xylene were from EMD ChemicalsMerck KGaA (Darmstadt, Germany). Toluene and meta-xylene were purchased from J.T. Baker (Phillipsburg, NJ). Methanol and cyclohexane were from Fischer Chemical (Guangzhou, China). Oleic acid was from Fluka (Buchs, Switzerland) and elaidic acid was from Alfa-Aesar (Ward Hill, MA). Iso-octane was from Riedel de Haen (Seelze, Germany), hexane was from Mallinckrodt (Phillipsburg, NJ), and acetonitrile was from Burdick and Jackson (Muskegon, MI). TMSI trimethylsilylation derivatization reagent was purchased from Supelco (Bellefonte, PA). Unleaded gasoline (87 octane) was obtained from a local gasoline filling station in Arlington, TX. Limits of detection (LOD) were determined for a subset of the chemicals listed above. Detector signals were obtained by integrating absorption across a specific spectral range. A series of standards were prepared to create calibration curves and one standard was prepared at approximately one to five times the estimated limit of detection. The latter standard was analyzed seven times and the standard deviation for the signal computed. The absolute LOD values (reported as mass on-column) were calculated as three times the standard deviation divided by the slope of the calibration curve. We performed ab initio computations using the Gaussian 095 program package. Ground-state structures were optimized to the second-order Møller−Plesset (MP2) level of theory using Dunning’s aug-cc-pvdz basis set.6 Electronic transition energies
and oscillator strengths of the molecules were calculated using the ZINDO/S method7,8 as implemented in Gaussian 09. The absorption spectra were computed without further parameter adjustments, and no adjustment of the overlap weighting factors was included.
■
RESULTS AND DISCUSSION The ability to visualize and deconvolve chemical compounds separated by GC depends on the characteristics of the detector. GC analyses utilizing detectors that cannot provide any qualitative information (e.g., flame ionization detection) will rely on the reproducible migration of mixture components through the column under a fixed set of method conditions. While it may be difficult to identify unknown compounds, some of these detectors provide excellent performance in terms of quantitative capabilities and ruggedness. Other detectors, such as the electron capture detector or the nitrogen phosphorus detector, provide unique specificity to detect and determine certain chemical compounds. This can be useful for discriminating target analytes from high background signals, but this selectivity comes with the cost of reducing the application breadth of the detector. MS is currently the most broadly applicable GC detector. MS can provide high performance quantitative and qualitative information. However, it is not fail safe, especially when considering its use for monitoring some important classes of bioactive compounds. In some cases, isomeric polycyclic aromatic hydrocarbons cannot be adequately distinguished chromatographically or based on generated mass spectra.9,10 In other cases, some pesticides are highly labile and decompose during injection, separation, or detection.11,12 N-Trihalomethylthio fungicides are examples of pesticides that readily decompose during electron ionization to yield poor mass 8331
dx.doi.org/10.1021/ac5018343 | Anal. Chem. 2014, 86, 8329−8335
Analytical Chemistry
Article
Figure 3. Absorption spectra (A) for a variety of small molecules. Measured spectra show good correlation with synchrotron data (B), as shown for methanol. Spectra for isomers are unique, as shown in (C) for xylenes and naphthols. This, in conjunction with appropriate software allows for deconvolution of overlapping signals in a chromatogram as shown in (D) for m- and p-xylenes.
spectral signals. In contrast, VUV detection has the promise to meet or exceed the capabilities of MS. At the very least, it can be considered a powerful complementary and orthogonal detection technique to improve confidence in selective determination of volatile and semivolatile chemical compounds. Often, in cases where MS fails to distinguish between two closely related compounds, VUV can be used instead. Many examples of VUV spectra for different chemical compounds and classes were measured and are presented herein as part of the regular article or given in the Supporting Information. Figure 2 shows the separation and detection of components in gasoline using a ZB-5 column on the GC-2010 hyphenated with the VUV detector (for comparison, GC-FID and GC-MS chromatograms for the gasoline sample are given in Supporting Information). Shown associated with the averaged full wavelength range acquisition signals across the chromatogram are the absorption spectra and identity of several components in the mixture. The rich features of these gas phase absorption spectra are apparent. Spectra taken in the gas phase are not subject to the extensive broadening commonly observed for UV absorption spectra in solution. Where available, spectra acquired from synchrotron sources (even if only partially covering the VUV spectral range) closely match those acquired by the VUV detector (see Figure 3). Unfortunately, a comprehensive database of VUV absorption spectra for chemical compounds is not yet available. However, recorded synchrotron spectra provide absolute cross-section (i.e., absorptivity across a range of wavelengths) information, and
these can be used to enhance quantitative analysis, as discussed below. Software to support data acquisition and analysis on the GCVUV includes capabilities to compare and deconvolve closely related spectra. Selected spectra in Figure 3A show highly featured VUV absorption spectra recorded for a set of simple small molecules. Even where carrier gases used will have their own absorption background, this signal was background subtracted by zeroing the detector at the start of the run. Figure 3B demonstrates the consistency of spectra recorded by the VUV detector with that taken from published synchrotron data. Importantly, the VUV spectra for closely related species are unique. This is shown in Figures 3C and in the Supporting Information. Some chemical compounds, such as naphthols, xylenes, and cis-/trans-fatty acids are virtually indistinguishable based on electron ionization mass spectral profiles. However, at short wavelengths and in the gas phase, these isomeric compounds are distinguishable. As seen in Figure 3D, for meta- and para-xylene, which are difficult to resolve chromatographically, a single observed chromatographic peak could be deconvolved into the additive contributions from the two coeluting isomers. The growth of VUV spectral databases for users to search and match will be key to the successful expansion of GC-VUV into various application areas. Such availability would provide functionality similar to that which libraries of electron ionization mass spectra provide to support GC-MS. Computation of theoretical VUV/UV spectra is possible using a variety of methods ranging from semiempirical 8332
dx.doi.org/10.1021/ac5018343 | Anal. Chem. 2014, 86, 8329−8335
Analytical Chemistry
Article
Figure 4. Computed electronic spectra for xylene isomers.
increasing molecular size (computations for napthols). The ability to accurately associate computed and measured absorption features with particular molecular features and functional groups will be a key next step in utilizing the detector for assigning the identities of various analytes. Quantitative analysis by VUV spectroscopy follows well established Beer−Lambert Law principles. A set of common volatile and semivolatile analytes was chosen to demonstrate the limits of detection for their determination by GC-VUV. These are given in Table 1. It is clear that from a quantitative
approaches and density functional theory calculations to sophisticated post-Hartree−Fock methods, as these are implemented in software packages such as Gaussian09.13−15 However, the applicability of highly accurate calculations is limited by the size of the molecule considered, and in many cases inexpensive parametric approaches such as the ZINDO/S method provide sufficient accuracy and good insight. A key point is that theoretical spectra assume discrete transitions between discrete energy levels. This is consistent with a synchrotron radiation source that is scanned across the relevant wavelengths to measure absorption. We find that computed spectra for similar compounds, e.g. isomers, show distinct features, but these do not always match the experimental VUV spectra. This is partially because the VGA-100 uses a broadband source with CCD detection; simultaneous excitation of a variety of electronic and underlying vibronic modes would be expected to complicate the absorption profile of discrete transitions.16,17 While further adjustment of theoretical models to better model experimental spectra is possible, already now the calculated results do provide insight into the specific transitions probed by the absorption events. This correlation may be eventually integrated to help establish trends or discern systematic changes to similar types of molecules analyzed. In Figure 4, the comparison of measured and calculated VUV absorption spectra for the xylene isomers is shown. One can compare the theoretical spectra with those measured and displayed in Figure 3C. Immediately, a large difference in the energy of the measured transitions is apparent; however, it is well-known that correction factors should be applied to the calculated values to bring them to the same scale as the experimental measurements.13 For the ZINDO calculations performed here, a correction factor of 0.8980 has been provided. Thus, multiplication of the theoretical transition wavelengths by this factor brings this more in line with the experimental values. Still, the ordering of values for the isomers is incorrect. In the VUV measurements, the apex of the most abundant transition proceeds in the order o-xylene (λmax = 187.5 nm) < p-xylene (λmax = 189 nm) ≈ m-xylene (λmax = 189 nm). The corrected calculated theoretical values give a different order: p-xylene (λmax = 187.7 nm) < o-xylene (λmax = 189.0 nm) < m-xylene (λmax = 189.5 nm). Even so, the intensity of a less intense transition approximately 35 nm higher than λmax shows more consistency with the experimental spectra. The shoulder for p-xylene is more pronounced (larger Δλ and larger oscillator strength) in both the theoretical and experimental spectra. Clearly more work is needed to understand what methods provide the best correlation between theoretical and experimental results. Some other computational results given in the Supporting Information reveal additional limitations, such as measured absorption range (alkane maximum absorption below 120 nm) and the potential for reduced accuracy with
Table 1. Determined Limits of Detection and Associated Wavelength Range Integrated for Selected Compounds water methanol benzene octane naphthol derivatized β-estradiol nicotine captan
LOD (pg on column)
integrated range (nm)
246 169 15 56 30 30 19 186
130−175 140−160 177−182 140−160 213−218 193−198 176−181 174−179
stand-point, the VUV detector provides a level of performance on par with or exceeding those of more established detectors. Water cannot be detected by FID, and it is very uncommon to monitor water by MS. A new and sensitive barrier ion discharge detector (Shimadzu Tracera; BID) provides for sensitive universal detection, but provides no qualitative information. N-trihalomethylthio antifungal pesticide compounds such as captan have been notoriously difficult to detect and differentiate from similar compounds, such as folpet, because of their extreme lability.12 For MS in the selected ion monitoring mode, it is not uncommon to reach low pg on-column detection limits. FID detection may have a bit more trouble seeing such small amounts. Linearity was determined for benzene as a model analyte, and was shown to be between three to 4 orders of magnitude. Of course, the dynamic range for analysis can be adjusted (especially in the high concentration range) by altering the monitored absorption wavelengths. While a rigorous assessment of accuracy and precision for determination of these compounds remains to be performed in the context of specific applications (the majority of methodological deficiencies in these analytical figures of merit will undoubtedly be attributable to sample preparation). Nevertheless, the reproducibility of response for several of the analytes in Table 1 injected at an amount near their limit of quantification on column was evaluated. Percent relative standard deviation (n = 7) were measured for benzene (62.2 pg, S/N = 12, 8.4% RSD), 8333
dx.doi.org/10.1021/ac5018343 | Anal. Chem. 2014, 86, 8329−8335
Analytical Chemistry
■
Article
CONCLUSION The VUV detector is a powerful new tool in the GC toolbox. It provides the potential to address several shortcomings of other detection technology. The ability to deconvolute multiple overlapping signals may provide for an enhanced dimension of analyte differentiation for which multidimensional GC is currently required. Whereas specialized columns are needed to be purchased to separate analytes that cannot be differentiated by MS, VUV absorption spectra are likely unique; this could limit the need and cost for specialized columns. A great deal more work is needed to establish the limits of this new technology, especially in conjunction with various sample preparation strategies to deal with complex real samples. Here, we showed the ability to speciate components in gasoline, where virtually all components are GC-amenable. It will be important to show the application of GC-VUV into other analytical realms, including gas analysis and environmental analysis to ensure its performance characteristics. Further, the establishment of spectral diagnostics, either by empirical measurements and the establishment of databases or through computational means will be important for VUV technology to mature to the level that GC-MS currently enjoys.
methanol (624.7 pg, S/N = 11, 9.2% RSD), naphthol (124.4 pg, S/N = 12, 7.8% RSD), and octane (279 pg, S/N = 15, 7.1% RSD). Representative chromatograms corresponding to these determinations are shown in the Supporting Information. With regard to stability, over a 1 h run under conditions equivalent to collecting Figure 2, the detector exhibits a drift of ±1.25 × 10−5 Abs/min. Aromatic and heteroatom-containing organic analytes clearly show the best VUV sensitivity. The functional units provide good chromophores for photon absorption. Octane shows intermediate sensitivity; alkanes have fewer features, as seen in Figure 2C, and the number of carbons correlates with absorption cross section in the low wavelength regime. In fact for alkanes, establishment of GC retention indices may be extremely valuable for deconvolution. Retention indices, in general, add specificity to GC determinations (i.e., retention time matching), and in combination with the VUV absorption spectra, they can be used to aid identification of compounds such as alkanes that have less distinct absorption features or limit searches to a more manageable subset of a library. Water and methanol are characterized by less intense absorption cross sections, and thus have poorer sensitivity. While captan might be expected to have fairly strong absorption based on its heteroatom content, there is still potential for its degradation in the GC column. β-Estradiol was derivatized to its trimethylsilyl form to improve its volatility. A good limit of detection was recorded, but it will be interesting to further investigate alternative derivatization agents that confer good chromophores to compounds of interest, especially for analytes that lack a great deal of functionality. A potentially useful characteristic of the technology is that if the VUV absorption cross-section for a chemical compound is known, the precise number of molecules in the detector can be determined based on the measured absorption signal. Based on terminology defined by Hulanicki, 18 GC-VUV can be considered a “pseudo-absolute” determination method in the absence of chemical interferences and losses from sample preparation. Conceptually, this means that if a model compound can be used to determine the instrument constant (degree of sample loss for fixed instrument conditions (e.g., split ratio), and assuming no time variation, prior to reaching the VUV detector), then the extent of absorption for an analyte can be used to directly determine the amount of analyte injected. This could prove to be extremely valuable for measurements designed to determine exceedances of action levels; however, for determinations from complex mixtures, sample losses because of sample preparation or instrumental errors prior to the detector are difficult to fully account, unless they are rigorously characterized prior to the analysis of interest. As previously mentioned, the VUV detector is mass-sensitive. Standard UV absorption spectrophotometers and the FID detector are also mass-sensitive. In spectroscopic absorption systems, multiple repeated measurements can be made on a sample in the flow cell without destroying the analyte. In the GC-VUV, a makeup flow of carrier gas incorporated postcolumn and predetector (see Figure 1) allows the user to optimize the residence time of analytes in the flow cell. Theoretically, an analyte could be made to reside in the flow cell for longer time to enhance signal averaging capability.
■
ASSOCIATED CONTENT
* Supporting Information S
Additional spectra, chromatograms, methodological details, and computational data. This material is available free of charge via the Internet at http://pubs.acs.org/.
■
AUTHOR INFORMATION
Corresponding Author
*E-mail:
[email protected]. Phone: 817-272-3541. Fax: 817-2723808. Notes
The authors declare the following competing financial interest(s): J.S., P.W., and D.H. are employees of VUV Analytics, Inc., the manufacturer of the instrumentation featured in this work.
■
ACKNOWLEDGMENTS The authors would like to thank Shimadzu Scientific Instruments, Inc. for the loan of a gas chromatograph to carry out this study.
■
REFERENCES
(1) Middleditch, B. S.; Sung, N.-J.; Zlatkis, A.; Settembre, G. Chromatographia 1987, 23, 273−278. (2) Driscoll, J. N.; Duffy, M.; Pappas, S. J. Chromatogr. 1988, 441, 63−71. (3) Lagesson, V.; Lagesson-Adrasko, L.; Andrasko, J.; Baco, F. J. Chromatogr., A 2000, 867, 187−206. (4) Hatzinikolaou, D. G.; Lagesson, V.; Stavridou, A. J.; Pouli, A. E.; Lagesson-Andrasko, L.; Stavrides, J. C. Anal. Chem. 2006, 78, 4509− 4516. (5) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb, M. A.; Cheeseman, J. R.; Scalmani, G.; Barone, V.; Mennucci, B.; Petersson, G. A.; Nakatsuji, H.; Caricato, M.; Li, X.; Hratchian, H. P.; Izmaylov, A. F.; Bloino, J.; Zheng, G.; Sonnenberg, J. L.; Hada, M.; Ehara, M.; Toyota, K.; Fukuda, R.; Hasegawa, J.; Ishida, M.; Nakajima, T.; Honda, Y.; Kitao, O.; Nakai, H.; Vreven, T.; Montgomery, J. A., Jr.; Peralta, J. E.; Ogliaro, F.; Bearpark, M.; Heyd, J. J.; Brothers, E.; Kudin, K. N.; Staroverov, V. N.; Kobayashi, R.; Normand, J.; Raghavachari, K.; Rendell, A.; Burant, J. C.; Iyengar, S. S.; Tomasi, J.; Cossi, M.; Rega, N.; Millam, J. M.; Klene, M.; Knox, J. E.; Cross, J. B.; Bakken, V.;
8334
dx.doi.org/10.1021/ac5018343 | Anal. Chem. 2014, 86, 8329−8335
Analytical Chemistry
Article
Adamo, C.; Jaramillo, J.; Gomperts, R.; Stratmann, R. E.; Yazyev, O.; Austin, A. J.; Cammi, R.; Pomelli, C.; Ochterski, J. W.; Martin, R. L.; Morokuma, K.; Zakrzewski, V. G.; Voth, G. A.; Salvador, P.; Dannenberg, J. J.; Dapprich, S.; Daniels, A. D.; Farkas, O.; Foresman, J. B.; Ortiz, J. V.; Cioslowski, J.; Fox, D. J. Gaussian 09, revision D.01; Gaussian, Inc.: Wallingford, CT, 2009. (6) Dunning, T. H. J. Chem. Phys. 1989, 90, 1007. (7) Ridley, J. E.; Zerner, M. C. Theor. Chem. Acc. 1973, 32, 111−134. (8) Zerner, M. C. In Reviews of Computational Chemistry, Vol. 2; Lipkowitz, K. B., Boyd, D. B., Eds.; VCH Publishing: New York, 1991, 313−66. (9) Fatoki, O. S.; Ximba, B. J.; Opeolu, B. O. Fres. Environ. Bulletin 2011, 20, 2012−2020. (10) Law, R. J.; Biscaya, J. L. Mar. Pollut. Bull. 1994, 29, 235−241. (11) Lehotay, S. J.; Koesukwiwat, U.; van der Kamp, H.; Mol, H. G. J.; Leepipatpiboon, N. J. Agric. Food Chem. 2011, 59, 7544−7556. (12) Mastovska, K.; Lehotay, S. J. J. Chromatogr. A 2004, 1040, 259− 272. (13) Suendo, V.; Viridi, S. ITB J. Sci. 2012, 44, 93−112. (14) Kato, H.; Nunes, Y.; Duflot, D.; Limao-Vieira, P.; Tanaka, H. J. Phys. Chem. A 2011, 115, 2708−2718. (15) Nunes, Y.; Martins, G.; Mason, N. J.; Duflot, D.; Hoffman, S. V.; Delwiche, J.; Hubin-Franskin, M.-J.; Limao-Vieira, P. Phys. Chem. Chem. Phys. 2010, 12, 15734−15743. (16) Turner, D. B.; Wilk, K. E.; Curmi, P. M. G.; Scholes, G. D. J. Phys. Chem. Lett. 2011, 2, 1904−1911. (17) De Mello Donega, C.; Meijerink, A.; Blasse, G. J. Phys.: Condens. Matter 1992, 4, 8889. (18) Hulanicki, A. Pure Appl. Chem. 1995, 67, 1906−1911.
8335
dx.doi.org/10.1021/ac5018343 | Anal. Chem. 2014, 86, 8329−8335
