Chemical Characterization of Crude Petroleum Using Nanospray

Dec 13, 2011 - Improving Hydrodenitrogenation Catalyst Performance through Analyzing ... High-Speed Tandem Mass Spectrometric in Situ Imaging by ...
1 downloads 0 Views 4MB Size
Article pubs.acs.org/ac

Chemical Characterization of Crude Petroleum Using Nanospray Desorption Electrospray Ionization Coupled with High-Resolution Mass Spectrometry Peter A. Eckert,† Patrick J. Roach,‡ Alexander Laskin,† and Julia Laskin*,‡ †

William R. Wiley Environmental and Molecular Sciences Laboratory Chemical and Materials Sciences Division, Pacific Northwest National Laboratory, P.O. Box 999, K8-88, Richland, Washington 99352, United States



S Supporting Information *

ABSTRACT: Nanospray desorption electrospray ionization (nano-DESI) combined with high-resolution mass spectrometry was used for the first time for the analysis of the polar constituents of liquid petroleum crude oil samples. The analysis was performed in both positive and negative ionization modes using three solvents, one of which (acetonitrile/toluene mixture) is commonly used in petroleomics studies while two other polar solvents (acetonitrile/water and methanol/water mixtures) are generally not compatible with petroleum characterization using mass spectrometry. The results demonstrate that nano-DESI analysis efficiently ionizes petroleum constituents soluble in a particular solvent. When acetonitrile/toluene is used as a solvent, nano-DESI generates electrospray-like spectra. In contrast, strikingly different spectra were obtained using acetonitrile/water and methanol/water. Comparison with the literature data indicates that these solvents selectively extract water-soluble constituents of the crude oil. Water-soluble compounds are predominantly observed as sodium adducts in nano-DESI spectra indicating that addition of sodium to the solvent may be a viable approach for efficient ionization of water-soluble crude oil constituents. Nano-DESI enables rapid screening of different classes of compounds in crude oil samples based on their solubility in solvents that are rarely used for petroleum characterization providing better coverage of the crude oil composition as compared to electrospray ionization (ESI). It also enables rapid characterization of water-soluble components of petroleum samples that is difficult to perform using traditional approaches.

C

Chemical analysis of crude oil samples using high-resolution MS (HR-MS) relies on soft ionization techniques that produce analyte ions without significant fragmentation.1 Electrospray ionization (ESI)18 has been widely used for the analysis of solvent extracts of petroleum. ESI typically generates [M + H]+ or [M + Na]+ ions in the positive mode and [M − H]− species in the negative mode and is particularly sensitive to acidic and basic compounds. Atmospheric pressure photoionization (APPI)19 has demonstrated its potential to ionize a broader range of compounds in petroleum samples.10,11,20,21 For example, nonpolar aromatic compounds containing pyrrolic nitrogen that are not observed in the positive mode ESI spectra yield abundant M+• ions in the positive mode APPI.11 Other ionization techniques including field desorption ionization (FD),22,23 atmospheric pressure chemical ionization (APCI),24 laser-induced acoustic desorption (LIAD)25,26 followed by gentle chemical ionization ionization, atmospheric pressure

rude petroleum oil is a complex mixture of hydrocarbons containing a small fraction of heteroatomic organic molecules and trace amounts of organometallics.1−4 Despite extensive studies focused on characterization of molecular composition of petroleum, it remains enigmatic on a molecular level due to its high complexity. Ultrahigh resolution mass spectrometry (MS) enabled assignment of the elemental composition of a variety of polar and some nonpolar components of petroleum crude oil with a level of detail that was not previously attainable.1,5 Specifically, a large number of sulfur-, nitrogen-, and oxygen-containing organic compounds have been resolved and identified.6−17 Identification of heteroatomic organic molecules in crude oil is crucial to understanding and mitigating the environmental effects of energy production from oil combustion. High mass accuracy, mass resolving power, and dynamic range are the key features of ultrahigh resolution mass spectrometry that created the field of petroleomics.1,2 Petroleomics is focused on obtaining a detailed understanding of the effect of the composition and reactivity of the individual constituents of crude oil on the physical and chemical properties of this complex mixture. © 2011 American Chemical Society

Received: October 21, 2011 Accepted: December 13, 2011 Published: December 13, 2011 1517

dx.doi.org/10.1021/ac202801g | Anal. Chem. 2012, 84, 1517−1525

Analytical Chemistry

Article

laser ionization (APLI),27,28 easy ambient sonic spray ionization (EASI),29 and discharge-induced desorption electrospray ionization (DESI)30 have been used to generate ions of petroleum constituents that are not readily ionized using ESI and APPI. Mass spectrometry analysis typically requires extraction of crude oil samples into a suitable solvent (toluene, dichloromethane, hot carbon disulfide, etc.). In both ESI and APPI, the crude oil sample is first dissolved in toluene and subsequently diluted with acetonitrile for analysis. In this study we demonstrate that online liquid microextraction followed by nanospray ionization enables analysis of petroleum samples without any sample preparation using solvents that are generally not compatible with crude oil. Specifically, we report high-resolution mass spectra obtained using methanol/water and acetonitrile/water mixtures and compare them with mass spectra obtained using acetonitrile/toluene that is traditionally used in petroleomics studies. The experiments were performed using nanospray desorption electrospray ionization (nanoDESI) recently developed in our laboratory for ambient pressure ionization of analytes deposited onto substrates.31 In nano-DESI, the analyte is desorbed into a liquid bridge formed between two fused-silica capillaries and ionized through a selfaspirating nanospray. Nano-DESI has been used for sensitive analysis of complex mixtures including atmospheric organic aerosols collected on substrates32 and tissue samples.33 Here we demonstrate the utility of nano-DESI for the analysis of liquid samples using petroleum crude oil as a model system.

(1) a 70:30 mixture of acetonitrile/toluene, (2) a 50:50 (v/v) acetonitrile/water mixture, and (3) 50:50 (v/v) methanol/ water mixture. In each experiment, mass spectra were acquired in both positive and negative ion modes. Data Analysis. Mass spectral features with a minimum signal-to-noise ratio of 5 were extracted from the averaged mass spectra of the background and the sample using Decon2LS software developed at Pacific Northwest National Laboratory (PNNL) (http://ncrr.pnl.gov/software/). Data processing was performed using a suite of Microsoft Excel macros developed in our group.34 First, the peaks corresponding to 13C isotopes were removed from the list by examining m/z values separated by the exact mass difference between the mass of 13C and 12C and taking into account the relative abundance of the peaks Next, the background and signal peaks lists were aligned and the background peaks were removed from the list. However, peaks occurring with a minimum of 2.5 times greater intensity in the signal than in the background were retained. Finally, the peaks were grouped using the first- and second-order mass defect analysis described elsewhere.34 This analysis is an extension of the widely used Kendrick transformation35,36 into multiple dimensions. In this study, the CH2-based first-order transformation was followed by the H2-based second-order transformation, which enabled clustering of the two-dimensional homologous series of peaks separated by the number of CH2 and H2 units into distinctive groups. Identification of one member in each group uniquely identifies all other members of the group. Our previous study34 demonstrated that this approach significantly reduces the complexity and simplifies the analysis of high-resolution mass spectra of complex mixtures. Furthermore, it helps identify higher molecular weight compounds that cannot be unambiguously assigned using Kendrick transformation. Elemental formulas were assigned to one peak in each group using MIDAS molecular formula calculator (http://magnet.fsu. edu/∼midas/). Formula assignments were performed using the following constraints that are typically used in the analysis of high-resolution MS spectra of crude oil: C ≤ 100, H ≤ 100, N ≤ 4, O ≤ 8, S ≤ 4, Na ≤ 1.37 Neutral formulas were obtained by removing sodium or proton from the ionic formulas of [M + Na]+ and [M + H]+ ions, respectively, and adding a proton to the formulas of [M − H]− species. The double-bond equivalent (DBE) of the neutral precursor was calculated from the assigned formulas using eq 1:



EXPERIMENTAL SECTION Mass Spectrometry. A sample of the NIST standard reference material of heavy sweet petroleum (NIST standard 2722) was used in this study. A 5 μL drop of the petroleum sample was placed on a glass slide for analysis, and a drop of solvent, of the same size, was deposited adjacent to and in contact with the petroleum. Analysis was performed with an LTQ/Orbitrap mass spectrometer (Thermo Electron, Bremen, Germany) equipped with a nano-DESI source described in detail elsewhere.31 Briefly, the nano-DESI probe was assembled using 193 μm o.d./100 μm i.d. fused-silica capillaries with polyimide coating (Polymicro Technologies LLC, Phoenix, AZ). The primary capillary supplies the solvent to the analysis region; the solvent from the drop formed on the glass slide is transferred into the nanospray capillary that fills up by the capillary action and empties by nanoelectrospray. The high voltage is applied to the stainless steel union that holds the primary capillary. In the experiments described in this study, the solvent was supplied at 0.3−0.9 μL/min flow rate to maintain a stable solvent drop on the surface. The nanospray capillary was positioned 0.5−1 mm away from the heated capillary of the mass spectrometer inlet. The spray voltage was optimized to obtain stable signal and was typically in the range of 2−4 kV. Mass spectra were obtained both in the positive and negative ion modes with a resolving power of 100 000 m/Δm at m/z 400. The instrument was calibrated using a standard mixture of caffeine, MRFA, and Ultramark 1621 (calibration mix MSCAL 5, Sigma-Aldrich, Inc.). In a typical experiment, background signal was recorded for 1−1.5 min prior to analysis by placing the nano-DESI probe on a glass slide away from the sample. Subsequently the sample was moved such that the probe capillaries were placed in the solvent droplet adjacent to the oil drop and mass spectra were acquired for 2−5 min. Three solvents were used in this study:

DBE = c − h /2 + n/2 + 1

(1)

where c, h, and n correspond to the number of carbon, hydrogen, and nitrogen atoms in the neutral formula, respectively. Oxygen and sulfur are divalent and, therefore, do not contribute to the DBE. Higher valences of sulfur were not considered in this study. It should be noted that eq 1 provides accurate values of DBE only if the assigned compound does not contain sulfur, where a higher than calculated DBE may occur.38



RESULTS AND DISCUSSION Nano-DESI analysis of liquid crude oil samples was performed using 70:30 (v/v) acetonitrile/toluene (AcN/Tol), 50:50 (v/v) acetonitrile/water (AcN/H2O), and 50:50 (v/v) methanol/ water (MeOH/H2O) as solvents. The first solvent, AcN/Tol, is commonly used in mass spectrometry characterization of petroleum samples, whereas the more polar AcN/H2O and MeOH/H2O mixtures are rarely used in petroleomics studies 1518

dx.doi.org/10.1021/ac202801g | Anal. Chem. 2012, 84, 1517−1525

Analytical Chemistry

Article

Figure 1. Positive mode nano-DESI spectra of the NIST heavy sweet petroleum obtained using mixtures of AcN/Tol (a), AcN/H2O (b), and MeOH/H2O (c) as solvents. The corresponding formula assignments are summarized in the pie charts shown on the right-hand side of each spectrum.

MeOH/H2O, whereas strikingly different mass spectra were obtained using AcN/Tol as a solvent. Positive Mode. Peak assignments performed using the procedure described earlier revealed that the positive mode AcN/Tol spectrum was dominated by the protonated CxHyN1 (N1) species. Similar spectrum was obtained using ESI/MS (data not shown). The N1 class of compounds containing pyridinic nitrogen is typically observed as the most abundant species in both ESI and APPI spectra of crude petroleum.2,6,11,12,15 The mass-to-charge ratios of the N1 molecules observed using nano-DESI are in the range of m/z 186−900. The DBE values of the N1 molecules observed in this study (range, 4−23) and the carbon numbers (range, 12−65) are in excellent agreement with the values reported by other groups.6,11,12,15 Other classes of nitrogen-containing com-

because of the poor solubility of the crude oil in these solvents. Figures 1 and 2 show high-resolution mass spectra obtained using the three solvents in the positive and negative modes, respectively, along with the summary of the corresponding formula assignments. The most abundant peaks and their assignments are summarized in Tables S1−S4 of the Supporting Information. Figure S1 of the Supporting Information shows a zoom-in view of a small portion of the positive AcN/Tol spectrum. Tables 1 and 2 list carbon numbers and DBE values for all assigned classes of compounds in the spectra. Figures 3 and 4 show plots of DBE versus the carbon number for the major classes of compounds observed in AcN/Tol and MeOH/H2O spectra. Similar spectra were obtained in both ionization modes using AcN/H2O and 1519

dx.doi.org/10.1021/ac202801g | Anal. Chem. 2012, 84, 1517−1525

Analytical Chemistry

Article

Figure 2. Negative mode nano-DESI spectra of the NIST heavy sweet petroleum obtained using mixtures of AcN/Tol (a), AcN/H2O (b), and MeOH/H2O (c) as solvents. The corresponding formula assignments are summarized in the pie charts shown on the right-hand side of each spectrum.

pounds observed in the positive AcN/Tol nano-DESI spectrum include CxHyNS (NS) and CxHyNOS (NOS) predominately cationized on sodium, protonated CxHyNO (NO), a small number of protonated CxHyN2 (N2) and CxHyNOx (x = 3−6) molecules, sodiated CxHySO3 (SO3), and both protonated and sodiated CxHyO3 (O3) species. NS and NOS compounds comprise 22% and 16% of the observed peaks in the AcN/Tol spectrum, respectively. These compounds are observed in the mass range of m/z 400−1,000 and have significantly lower values of DBE than the N1 species (0−14 for NS and 0−16 for NOS molecules). However, whereas NS species are predominantly observed within a fairly narrow range of carbon numbers of 31−56 with most species found in the range of 40−50, NOS compounds span a much broader range of carbon numbers of 23−64. Similarly, NO

compounds comprising approximately 6% of the peak assignments are observed in the mass range of m/z 400−900. This group of peaks is characterized by a narrower DBE distribution than the N1 species, with a maximum and minimum DBE of 20 and 8, respectively. The majority of the NO species fall in a fairly narrow range of carbon numbers of 30−43 and a range of DBE of 9−16. Similar classes of petroleum crude oil constituents are commonly observed in high-resolution ESI mass spectra.11,12,15,39,40 For example, N1, NO, NS, and NOS species were reported by Qian et al.6 in their study focused on the identification of nitrogen-containing aromatic compounds in crude oil extracts using positive mode ESI. It is reasonable to assume that, because petroleum is typically extracted using toluene or other nonpolar solvents, nano-DESI analysis 1520

dx.doi.org/10.1021/ac202801g | Anal. Chem. 2012, 84, 1517−1525

Analytical Chemistry

Article

Table 1. Range of Carbon Numbers and Double-Bond Equivalents (DBE) for Compounds Detected in the Positive Mode solvent AcN/Tol compound class

carbon no.

DBE

N1 N2 NO NOx NS NOS NOS2 NO2S NO3S O1 O2 O3 O5 OS O2S O3S O4S OS2 O2S2 O3S2 O3S3

12−65 24−30 28−65 51−60 31−56 23−64 42−55

4−23 5−6 8−20 6−8 0−14 0−16 1−9

34−50

3−15

44−47

2−5

AcN/H2O carbon no. 14−20 13−20

3−8

12−26

0−4

24−36

3−12

14−40 11−25 15−49 24−37 8−22 15−41 7−37 4−33 10−30 14−39 28−37 19−52

4−11 2−12 3−19 10−16 1−9 2−11 3−12 0−10 0−9 1−12 6−10 2−15

MeOH/H2O

DBE

carbon no.

DBE

5−10

14−20

5−6

12−36

6−18

13−30 20−28

3−11 4−10

11−21 10−15 7−37 7−22 13−20 16−30 14−30 11−35 11−39 11−30 18−41

2−9 2−6 2−18 1−8 2−9 1−8 0−9 2−11 1−18 0−10 0−8

Table 2. Range of Carbon Numbers and Double-Bond Equivalents (DBE) for Compounds Detected in the Negative Mode solvent AcN/Tol compound class

carbon no.

DBE

O1 O2 O3 O4 O5 O6 O2S O3S O4S O5S O6S O4S2 NO3 NO4

9−21 7−35 6−35 9−35 13−35 16−20 10−11 8−48 7−39 10−30 15−23 11−19 9−24 16−25

1−10 1−13 1−13 2−12 3−12 3−8 2 0−9 0−10 0−8 0−5 2−4 2−11 7−11

AcN/H2O carbon no.

MeOH/H2O

DBE

carbon no.

DBE

12−21 10−30 12−24 11−22

4−11 1−13 2−8 2−8

12−14 10−26 12−26 12−21

4−5 1−12 1−9 2−8

15−21 11−26 11−16 16−18

3−7 1−7 0−3 2

16−20 15−22 11−23 10−22 15−18

3−6 3−8 1−7 1−5 1−3

Figure 3. Plots of double-bond equivalents (DBE) vs carbon number for different classes of compounds detected using AcN/Tol in the positive mode (a) and negative mode (b and c) nano-DESI spectra of the NIST heavy sweet petroleum. The size of the dot is proportional to the logarithm of the ion abundance.

compounds was observed in AcN/H2O and MeOH/H2O spectra. Both spectra are dominated by CxHyOS (OS), CxHyO2S2 (O2S2) compounds cationized on sodium and either protonated or sodiated CxHyOz species. The OS compounds are observed in the mass range of m/z 180−350 and have DBE values in the range of 1−8 with mode and median values of 2 and 3, respectively. In contrast, the O2S2 compounds are characterized by a broader distribution of DBE (1−12) with mode and median values of 6 and 7, respectively, which are at least 2 times higher than the values characteristic of the OS

involves online liquid extraction followed by nanospray ionization of the more soluble polar components of the petroleum sample. Our results indicate that, when AcN/Tol is used as a solvent, nano-DESI generates spectra similar to the more traditionally used ionization approaches. However, nanoDESI provides an advantage for petroleum characterization in that no sample preparation is necessary prior to analysis. Although the nano-DESI spectrum obtained using AcN/Tol closely resembles ESI spectra of petroleum crude oil extracted into toluene or heptane, strikingly different distribution of 1521

dx.doi.org/10.1021/ac202801g | Anal. Chem. 2012, 84, 1517−1525

Analytical Chemistry

Article

gradients between the capillary and the skimmer is shown in Figure S2 of the Supporting Information. A significant fraction of the O2S2 compounds is strongly suppressed in the 70 V spectrum. We attribute these compounds to noncovalent dimers of the OS species. However, ca. 15% of O2S2 molecules are very stable and are not significantly affected by the presence of the field gradient in the ionization region. These compounds with carbon numbers in the range of 14−22 and fairly low DBE values of 1−6 most likely correspond to genuine O2S2 species. In addition to sulfur-containing compounds, AcN/H2O and MeOH/H2O spectra contain oxygenated, CxHyO1−6, species. The most abundant groups of oxygenated organic molecules in these spectra correspond to CxHyO2 (O2), CxHyO4 (O4), and CxHyO5, (O5); the O2−4 molecules are cationized on sodium, whereas the more oxygenated O5−6 compounds are observed as protonated molecules. Interestingly Ox and OxS1−2 compounds were the most abundant species reported by Stanford et al.41 in their study of water-soluble constituents of crude oil using ESI of solvent extracts. However, whereas the OxS1−2 compounds were observed in that study in both positive and negative mode, the Ox compounds were detected only in the negative mode. Other classes of compounds observed by Stanford et al.41 using positive mode ESI and detected in this study by the positive mode nano-DESI include O1−4S, O1−3S2, NO0−2, NO2S, and NS. The striking similarity between the nano-DESI spectra obtained using AcN/H2O and MeOH/H2O and the comparison with the literature indicate that these two solvents selectively ionized water-soluble components of the crude oil sample. The small differences between the AcN/H2O and MeOH/H2O spectra most likely reflect differences in the solubility of the crude oil constituents in AcN and MeOH. Alternatively, some differences could originate from solvent− analyte reactions. For example, it has been demonstrated that oxygenated organic molecules may be converted into acetals and hemiacetals through condensation reactions with methanol.42 This process may contribute to additional peaks in the spectrum when methanol is used as a solvent. Obviously, the composition of the water-soluble fraction of petroleum is very different from the composition of the fraction extracted into AcN/Tol. Our results demonstrate that nano-DESI analysis using different solvents may be used for rapid surveying of petroleum samples for the presence and the composition of water-soluble constituents, which is critical for predicting and mitigating environmental effects of the oil spills. It should be noted that OxS and OxS2 compounds have been detected by several groups using derivatization of petroleum crude oil samples with methyl or phenyl groups followed by ESI-MS analysis.5,9,10,13,14,28,43 Alternatively, these compounds have been detected using APPI of underivatized samples.10 Such chemical derivatization approaches enable ionization of fairly nonpolar sulfur-containing molecules for mass spectrometry characterization. For example, Liu et al. observed O1−4S and O1−3S2 compounds primarily in the asphaltene fraction of the crude petroleum.13,14 On the basis of the low DBE values, they proposed that the OS compounds are most likely sulfoxides. Interestingly, the most abundant OS compounds observed in this study, C9H16OS, C9H18OS, C10H18OS, C10H20OS, C11H20OS, C12H20OS, C12H22OS, and C13H24OS, have very low DBE values in the range of 1−3. Although we cannot determine the structures of the OS compounds observed in this study based on their molecular formulas

Figure 4. Plots of double-bond equivalents (DBE) vs carbon number for different classes of compounds detected using MeOH/H2O in the positive mode (a and b) and negative mode (c) nano-DESI spectra of the NIST heavy sweet petroleum. The size of the dot is proportional to the logarithm of the ion abundance.

compounds. In-source fragmentation experiments were performed to distinguish between genuine O2S2 compounds and dimers of the OS species. In these experiments, ions are accelerated in the relatively high-pressure region by applying a potential gradient between the heated capillary and the skimmer located in the first differentially pumped region of the ion source of the instrument. Collisions of ions with the background gas induce fragmentation, which is particularly efficient for chemically labile species. Comparison between AcN/H2O nano-DESI spectra obtained using 0 and 70 V field 1522

dx.doi.org/10.1021/ac202801g | Anal. Chem. 2012, 84, 1517−1525

Analytical Chemistry

Article

adding the exact mass of hydrogen for the positive and negative mode data, respectively. Clearly, lower-mass species are observed in the negative mode. Our results demonstrate that high-resolution mass spectrometry analysis using different solvents and different modes of ionization provides better coverage of the constituents present in the crude oil sample. Nano-DESI enables rapid screening of different solvents including those that are not commonly used for the analysis of petroleum crude oil. The major difference between the AcN/H2O and MeOH/ H2O is the presence of a significantly larger fraction (12%) of O1 compounds in the AcN/H2O spectrum as compared to only 4% in the MeOH/H2O spectrum. These compounds observed as deprotonated species have the carbon numbers in the range of 12−21 and DBE values in the range of 4−11. In addition, highly oxidized molecules are more dominant in the MeOH/ H2O spectrum. For example, a broader distribution of the O5 and O4S compounds was observed using MeOH/H2O, whereas the O2 and O3S compounds were better represented in the AcN/H2O spectrum. It is important to note that, although oxygenated organic molecules were observed in all three solvents, differences exist in the distribution of these compounds extracted into AcN/Tol and the more polar solvents. Figure 6 compares the distribution of the O2 species

alone, we tentatively assign them to sulfoxides based on the comparison with the literature.13,14 Efficient ionization of underivatized O x S and O x S 2 compounds observed in this study is mainly attributed to the presence of sodium that facilities the formation of [M + Na]+ ions. In our initial experiments, sodium ions originated from the glass slide. Thorough cleaning of the glass slide prior to petroleum analysis resulted in significant suppression of the peaks corresponding to the OxS and OxS2 compounds while addition of NaCl to the solvent at 10−50 μM concentration helped recover the signal. Obviously, the latter is a preferred approach for controlled and reproducible ionization of molecules having substantial sodium affinity. Negative Mode. Although the negative mode nano-DESI AcN/Tol spectrum is quite different from the spectra obtained using AcN/H2O and MeOH/H2O, there is a significant overlap between the classes of compounds observed using different solvents. Acidic oxygen- and sulfur-containing compounds dominate all three spectra. A broad distribution of peaks is observed in the AcN/Tol spectrum that contains a large number of O1−6 species with similar DBE values. However, although O2 and O3 molecules account for the majority of the features in the AcN/H2O and MeOH/H2O spectra, the oxygenated organics are fairly evenly distributed among the O2−5 species in the AcN/Tol spectrum. The O2 species constitute the most significant functional group in both AcN/H2O and MeOH/H2O spectra. This class of molecules is characterized by carbon numbers in the range of 10−30 and DBE values of 1−13. The most abundant peaks in this group have fairly low DBE values of 3−5. Similar classes of compounds were observed by other groups in the negative mode. A general consensus is that the Ox compounds correspond to naphthenic acids7,44−46 with one or more carboxylic functional groups or fatty acids.15 In addition to the oxygenated species, negative mode nanoDESI spectra contain a distribution of O2−6S compounds and a small fraction of NO3 and NO4 compounds. Although there is little overlap between the positive and negative mode spectra, O4S compounds are observed in both modes when MeOH/ H2O is used as a solvent. Figure 5 compares the distribution of

Figure 6. Distribution of the O2 compounds observed in the negative mode nano-DESI spectra using AcN/Tol (red) and MeOH/H2O (blue) as solvents. Data is shown for neutral species.

observed in the AcN/Tol and MeOH/H2O spectra. The major difference between the two solvents is observed in the lowermass range with more low-m/z compounds ionized using AcN/ Tol as compared to MeOH/H2O.



CONCLUSIONS

The recently developed nano-DESI technique was utilized for the first time for the analysis of liquid samples. We used petroleum crude oil as a model system to demonstrate the utility of the online liquid extraction employed by nano-DESI for rapid and sensitive analysis of a broad range of compounds in a complex liquid sample without any sample preparation. The result demonstrated that when AcN/Tol is used as a solvent, nano-DESI generates mass spectra that are similar to the spectra obtained using more traditionally used ESI and APPI techniques. A variety of nitrogen-containing species (e.g., N1, NO, NS, NOS, N2, etc.) were observed in the positive mode AcN/Tol nano-DESI spectrum, whereas the negative mode spectrum was dominated by the O1−6 and O1−6S

Figure 5. Distribution of the O2−6S compounds observed in the positive (red) and negative (blue) mode MeOH/H2O nano-DESI spectra. Data is shown for neutral species.

O4S species observed in the positive and negative mode. In this plot, the experimentally measured m/z values were converted into neutral masses by subtracting the exact mass of sodium and 1523

dx.doi.org/10.1021/ac202801g | Anal. Chem. 2012, 84, 1517−1525

Analytical Chemistry

Article

(10) Purcell, J. M.; Juyal, P.; Kim, D. G.; Rodgers, R. P.; Hendrickson, C. L.; Marshall, A. G. Energy Fuels 2007, 21, 2869−2874. (11) Purcell, J. M.; Rodgers, R. P.; Hendrickson, C. L.; Marshall, A. G. J. Am. Soc. Mass Spectrom. 2007, 18, 1265−1273. (12) Shi, Q.; Xu, C.; Zhao, S.; Chung, K. H.; Zhang, Y.; Gao, W. Energy Fuels 2009, 24, 563−569. (13) Liu, P.; Shi, Q. A.; Chung, K. H.; Zhang, Y. H.; Pan, N.; Zhao, S. Q.; Xu, C. M. Energy Fuels 2010, 24, 5089−5096. (14) Liu, P.; Xu, C. M.; Shi, Q. A.; Pan, N.; Zhang, Y. H.; Zhao, S. Q.; Chung, K. H. Anal. Chem. 2010, 82, 6601−6606. (15) Shi, Q.; Hou, D. J.; Chung, K. H.; Xu, C. M.; Zhao, S. Q.; Zhang, Y. H. Energy Fuels 2010, 24, 2545−2553. (16) Shi, Q. A.; Pan, N.; Liu, P.; Chung, K. H.; Zhao, S. Q.; Zhang, Y. H.; Xu, C. M. Energy Fuels 2010, 24, 3014−3019. (17) Liu, P.; Shi, Q.; Pan, N.; Zhang, Y. H.; Chung, K. H.; Zhao, S. Q.; Xu, C. M. Energy Fuels 2011, 25, 3014−3020. (18) Fenn, J. B.; Mann, M.; Meng, C. K.; Wong, S. F.; Whitehouse, C. M. Science 1989, 246, 64−71. (19) Robb, D. B.; Covey, T. R.; Bruins, A. P. Anal. Chem. 2000, 72, 3653−3659. (20) Barrow, M. P.; Witt, M.; Headley, J. V.; Peru, K. M. Anal. Chem. 2010, 82, 3727−3735. (21) Purcell, J. M.; Hendrickson, C. L.; Rodgers, R. P.; Marshall, A. G. Anal. Chem. 2006, 78, 5906−5912. (22) Schaub, T. M.; Hendrickson, C. L.; Qian, K. N.; Quinn, J. P.; Marshall, A. G. Anal. Chem. 2003, 75, 2172−2176. (23) Smith, D. F.; Schaub, T. M.; Rodgers, R. P.; Hendrickson, C. L.; Marshall, A. G. Anal. Chem. 2008, 80, 7379−7382. (24) Kim, Y. H.; Kim, S. J. Am. Soc. Mass Spectrom. 2010, 21, 386− 392. (25) Crawford, K. E.; Campbell, J. L.; Fiddler, M. N.; Duan, P.; Qian, K.; Gorbaty, M. L.; Kenttamaa, H. I. Anal. Chem. 2005, 77, 7916− 7923. (26) Nyadong, L.; McKenna, A. M.; Hendrickson, C. L.; Rodgers, R. P.; Marshall, A. G. Anal. Chem. 2011, 83, 1616−1623. (27) Schmitt-Kopplin, P.; Englmann, M.; Rossello-Mora, R.; Schiewek, R.; Brockmann, K. J.; Benter, T.; Schmitz, O. J. Anal. Bioanal. Chem. 2008, 391, 2803−2809. (28) Schrader, W.; Panda, S. K.; Brockmann, K. J.; Benter, T. Analyst 2008, 133, 867−869. (29) Corilo, Y. E.; Vaz, B. G.; Simas, R. C.; Nascimento, H. D. L.; Klitzke, C. F.; Pereira, R. C. L.; Bastos, W. L.; Neto, E. V. S.; Rodgers, R. P.; Eberlin, M. N. Anal. Chem. 2010, 82, 3990−3996. (30) Wu, C. P.; Qian, K. N.; Nefliu, M.; Cooks, R. G. J. Am. Soc. Mass Spectrom. 2010, 21, 261−267. (31) Roach, P. J.; Laskin, J.; Laskin, A. Analyst 2010, 135, 2233− 2236. (32) Roach, P. J.; Laskin, J.; Laskin, A. Anal. Chem. 2010, 82, 7979− 7986. (33) Laskin, J.; Heath, B. S.; Roach, P. J.; Cazares, L.; Semmes, O. J. Anal. Chem. 2012, 84, 141−148. (34) Roach, P. J.; Laskin, J.; Laskin, A. Anal. Chem. 2011, 83, 4924− 4929. (35) Kendrick, E. Anal. Chem. 1963, 35, 2146−2154. (36) Hughey, C. A.; Hendrickson, C. L.; Rodgers, R. P.; Marshall, A. G.; Qian, K. N. Anal. Chem. 2001, 73, 4676−4681. (37) Kim, S.; Rodgers, R. P.; Marshall, A. G. Int. J. Mass Spectrom. 2006, 251, 260−265. (38) Badertscher, M.; Bischofberger, K.; Munk, M. E.; Pretsch, E. J. Chem. Inf. Comput. Sci. 2001, 41, 889−893. (39) Bae, E.; Na, J.-G.; Chung, S. H.; Kim, H. S.; Kim, S. Energy Fuels 2010, 24, 2563−2569. (40) Fernandez-Lima, F. A.; Becker, C.; McKenna, A. M.; Rodgers, R. P.; Marshall, A. G.; Russell, D. H. Anal. Chem. 2009, 81, 9941−9947. (41) Stanford, L. A.; Kim, S.; Klein, G. C.; Smith, D. F.; Rodgers, R. P.; Marshall, A. G. Environ. Sci. Technol. 2007, 41, 2696−2702. (42) Bateman, A. P.; Walser, M. L.; Desyaterik, Y.; Laskin, J.; Laskin, A.; Nizkorodov, S. A. Environ. Sci. Technol. 2008, 42, 7341−7346.

compounds. In contrast, both positive and negative mode spectra obtained using polar AcN/H2O and MeOH/H2O solvents were dominated by the Ox and OxSy compounds. Minimal overlap was observed between the neutral species detected in the different ionization modes indicating the complementary nature of the analysis performed in the positive and negative modes. Comparison with the literature data indicates that nano-DESI analysis using AcN/H2O and MeOH/H2O resulted in selective ionization of water-soluble components of the petroleum crude oil. These compounds were observed predominantly as sodium adducts, which could explain facile ionization of this class of molecules that is typically not observed in ESI spectra. The ability to detect water-soluble compounds in petroleum crude oil without sample preparation is important for understanding the environmental effects of oil spill accidents. Our results demonstrate that nano-DESI can be used for rapid surveying of the crude oil constituents based on their solubility in different solvents. This capability enables soft ionization of molecules that are difficult to characterize using traditional approaches.



ASSOCIATED CONTENT

S 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

*E-mail: [email protected].

ACKNOWLEDGMENTS The research presented here was performed at the W. R. Wiley Environmental Molecular Sciences Laboratory (EMSL)a national scientific user facility sponsored by the U.S. Department of Energy’s (U.S. DOE) Office of Biological and Environmental Research and located at the Pacific Northwest National Laboratory (PNNL). PNNL is operated by Battelle for the U.S. DOE. J.L. and P.J.R. acknowledge support from the Chemical Sciences Division, Office of Basic Energy Sciences; A.L. acknowledges support from the EMSL intramural research and development program EMSL. P.A.E. acknowledges support from the DOE Science Undergraduate Laboratory Internship (SULI) program at PNNL. P.A.E. is an undergraduate student from Trinity International University, IL.



REFERENCES

(1) Marshall, A. G.; Rodgers, R. P. Acc. Chem. Res. 2004, 37, 53−59. (2) Marshall, A. G.; Rodgers, R. P. Proc. Natl. Acad. Sci. U.S.A. 2008, 105, 18090−18095. (3) Rodgers, R. P.; McKenna, A. M. Anal. Chem. 2011, 83, 4665− 4687. (4) Panda, S. K.; Andersson, J. T.; Schrader, W. Anal. Bioanal. Chem. 2007, 389, 1329−1339. (5) Panda, S. K.; Andersson, J. T.; Schrader, W. Angew. Chem., Int. Ed. 2009, 48, 1788−1791. (6) Qian, K.; Rodgers, R. P.; Hendrickson, C. L.; Emmett, M. R.; Marshall, A. G. Energy Fuels 2001, 15, 492−498. (7) Headley, J. V.; Peru, K. M.; Barrow, M. P.; Derrick, P. J. Anal. Chem. 2007, 79, 6222−6229. (8) Hughey, C. A.; Rodgers, R. P.; Marshall, A. G.; Qian, K.; Robbins, W. K. Org. Geochem. 2002, 33, 743−759. (9) Muller, H.; Andersson, J. T.; Schrader, W. Anal. Chem. 2005, 77, 2536−2543. 1524

dx.doi.org/10.1021/ac202801g | Anal. Chem. 2012, 84, 1517−1525

Analytical Chemistry

Article

(43) Tomczyk, N. A.; Winans, R. E.; Shinn, J. H.; Robinson, R. C. Energy Fuels 2001, 15, 1498−1504. (44) Headley, J. V.; Peru, K. M.; Barrow, M. P. Mass Spectrom. Rev. 2009, 28, 121−134. (45) Barrow, M. P.; McDonnell, L. A.; Feng, X. D.; Walker, J.; Derrick, P. J. Anal. Chem. 2003, 75, 860−866. (46) Gabryelski, W.; Froese, K. L. Anal. Chem. 2003, 75, 4612−4623.

1525

dx.doi.org/10.1021/ac202801g | Anal. Chem. 2012, 84, 1517−1525