Article pubs.acs.org/ac
Impact of Different Ionization Methods on the Molecular Assignments of Asphaltenes by FT-ICR Mass Spectrometry Andras Gaspar, Elio Zellermann, Sami Lababidi, Jennifer Reece, and Wolfgang Schrader* Max-Planck-Institut für Kohlenforschung, Kaiser-Wilhelm-Platz 1, D-45470 Mülheim an der Ruhr, Germany ABSTRACT: Over the years, ultrahigh resolution mass spectrometry has successfully illustrated the extreme complexity of crude oil and related solubility or polarity based fractions on a molecular level. However, the applied ionization technique greatly influences the outcome and may provide misleading information. In this work, we investigate the atmospheric pressure laser ionization (APLI) technique coupled with Fourier transform ion cyclotron resonance mass spectrometer to analyze the asphaltene fraction of a crude oil. These results were compared to data obtained by using other existing atmospheric pressure ionization methods. Furthermore elemental analysis and solid state NMR were used to obtain the bulk characteristics of the asphaltene sample. The results of the different ionization techniques were compared with the bulk properties in order to describe the potential discrimination effects of the ionization techniques that were observed. The results showed that APLI expands the range of the assigned molecules, while retaining information already observed with the generally used ion sources. aromatic structures are classified into “island” models (one aromatic core) or “archipelago” models (multiple aromatic centers bridged by alkyl chains) as proposed using timeresolved fluorescence depolarization7,14 and mass spectrometry.15,16 Another aspect that can influence the analysis is the aggregation and self-assembly of the asphaltenes, especially in toluene solution. Fluorescence quenching measurements indicated that the critical nanoaggregate concentration (CNAC) is on the order of 60 mg/L in toluene.17−19 The indeterminate complex composition of oil and its fractions sometimes makes it difficult to define and interpret results with existing analytical tools.20 However, methods like diffusion NMR, small-angle neutron scattering and various fluorescence measurements, describe average characteristics that can still support structure−function relations. Ultrahigh resolution mass spectrometry like FT-ICR MS, with its unsurpassed mass resolution and mass accuracy, enables a molecular level analysis of complex petroleum mixtures and asphaltenes that can be compared to bulk measurements.21,22 on the basis of accurate mass measurements, unique elemental compositions (CcHhNnOoSs) can be generated if the resolution23 and the mass accuracy24 is sufficiently high. Assigned molecular compositions enable classification according to heteroatom containing classes and the degree of aromaticity. Electrospray ionization (ESI) is especially efficient at generating acceptable results in the case of polar constituents with high heteroatom content. The combination of positive and negative ESI with FT-ICR MS for mass spectral analysis of asphaltenes was demonstrated by Klein et al.19 In their studies
D
espite the continual development of renewable energy sources, energy supplies will be dependent upon the availability of crude oil for at least the next two decades. As the remaining light crude oils diminish, previously unconventional resources1,2 will need to be upgraded into petroleum.3,4 Many of the problems associated with recovery, separation or processing of crude oils are related to the presence of high concentrations of asphaltenes.5 The asphaltene fraction is defined by its solubility in toluene and insolubility in normal paraffinic solvents (e.g., n-heptane). A commonly applied fractionation method uses the different solubilities to separate crude oil into saturates, aromatics, resins, and asphaltene (SARA) fractions,6 where asphaltenes are initially precipitated by using nonpolar solvents. Asphaltenes, as a heterogeneous mixture, are highly disperse in both molecular size and chemical composition, with a high content of heteroatoms (N, O, and S) and metals (i.e., V, Fe) causing them to exhibit polar characteristics.7 The asphaltene fraction is thought to be largely responsible for some adverse properties of oil, such as high viscosity, emulsion and polymer-forming tendencies. Furthermore, asphaltenes are known precursors of coke in acid catalysis and can inhibit catalysts. Therefore, an in-depth characterization of these constituents in oil is highly desirable.8−10 The spectrometric analysis of the asphaltene fraction remains a challenge because of molecular properties that result in limited solubility and aggregation.11,12 The solubility of asphaltenes is generally governed by the balance of the van der Waals attraction of single heteroatom-substituted polyaromatic hydrocarbons (PAH) versus the steric repulsion of alkane chains.13 As the van der Waals attraction increases (ratio of condensed structures), the solubility decreases and results in partial precipitation of the asphaltene. These condensed © 2012 American Chemical Society
Received: January 13, 2012 Accepted: May 17, 2012 Published: May 18, 2012 5257
dx.doi.org/10.1021/ac300133p | Anal. Chem. 2012, 84, 5257−5267
Analytical Chemistry
Article
Table 1. Calculated Mean Values of the Observed m/z, DBE, C Atom Number, and DBE/C, Derived from the Assigned Formulas Obtained with the Different Ionization Methods (Data Are Only Based on m/z Values) ESI H-ESI APCI APPI APLI
mean m/z
mean DBE
mean C
mean DBE/C
av O/C
av N/C
av S/C
528.1 565.7 572.7 559.5 524.6
16 17 14 13 15
38 41 41 39 37
0.44 0.45 0.36 0.35 0.41
0.0011 0.0034 0.0032 0.0037 0.0042
0.0389 0.0328 0.0108 0.0132 0.0137
0.0011 0.0027 0.0069 0.0165 0.0138
triggered the investigation of its potential for the analysis of asphaltene samples given their presumably aromatic structures and heteroatom moieties. When trying to solve the structural aspects of asphaltenes it is important to understand their behavior within the analytical system. Prior to any understanding that is solely based on mass spectrometric approach, understanding the key ionization parameters of the different ionization methods is crucial. While for a known molecule, based on its structure, the most suitable ionization method can be used, the ionization and analysis of extreme intricate samples, such as asphaltenes, can provide misleading information. These samples contain a wide range of constituents that can be readily ionized in solution or show no response for the selected ionization and eventually compete during the ionization process. Besides optimizing the solvent content to minimize the matrix effect, the solvent has to promote also the ionization process like for instance in case of photoionization where toluene serves as dopant. Since all ionization methods emphasize different compounds present in crude oil fractions like it was shown with vacuum gas oils, there should be even more differences when using such unconventional samples as crude oil asphaltenes. Therefore, we compared data from the relatively new ionizatizion method APLI with those from “mainstream” ionization sources.
the authors were able to determine the differences between asphaltenes, collected from C7 precipitation and live oil depressurization experiments, and asphaltenes from different geographic locations.8,22 Here, we must point out that the authors applied a sample concentration of 1 mg/mL, which exceeds the given CNAC threshold for asphaltenes in toluene, and they were able to detect, with ESI, constituents up to a DBE value of 37. Atmospheric pressure chemical ionization (APCI) is also suitable for the analysis of asphaltenes and its model compounds, especially if the measurement conditions are selected carefully.25 APCI was used, in combination with collision-activated dissociation (CAD),15 to describe the preferred fragmentation pathways for the island and archipelago structural models. Furthermore, by analysis of the asphaltene sample, the dominant loss of alkyl chains of varying lengths (between one and twelve carbon, mainly methyl groups followed by ethyl, propyl, butyl, etc.) was observed, indicating the existence of “island”-type structures. Although ESI and APCI are able to yield abundant ions from several polar species, purely hydrocarbon cycloalkanes and aromatic species remain inaccessible. Atmospheric pressure photo ionization (APPI) can positively charge these compounds, producing radical cations [M]+• and protonated molecules [M+H]+, and efficiently ionize polycyclic aromatic hydrocarbons with or without heteroatoms present.26,27 Using APPI, the changes in hydrocarbon and sulfur classes during deep hydrotreatment processes were observed.24 Additionally, APPI was also successfully used to identify and catalog diverse structures of vanadyl porphyrins.25,28 In parallel, the recent development of atmospheric-pressure laser ionization (APLI) enabled sensitive and selective analysis of polyaromatic compounds, with and without heteroatoms.29,30 APLI is based on resonant or near-resonant twophoton ionization of aromatic ring systems at 248 nm.31 Constapel et al.29 reported a significant enhancement of the sensitivity for low polarity analytes (benzo[a]pyrene, anthracene, fluorene, fluorantrene) over APCI measurements. They also detailed significant differences between the ionization mechanisms and efficiencies of APPI and APLI. The generally observed low ion yield of APPI is due to the limited photon flux and series of side reactions. To overcome this problem, addition of directly ionizable compound (dopant) can improve sensitivity but also may render the analysis more difficult because of adduct formation. Further problems occur during the photoionization due to the strong absorption of the matrix, and its O2, H2O moieties, causing oxidation of the analytes.27 Kersten et al.31 described the effect of neutral radicals generated by APPI through the oxidation of the pyrene radical cation, while APLI denoted exclusively the parent radical cations. Furthermore, the performance of APLI in the analysis of vacuum gas oil was also investigated and the outcome showed good response for nonpolar aromatic hydrocarbons.30 The described properties of this relatively new ionization technique
■
EXPERIMENTAL SECTION Crude Oil, Asphaltene. Heavy crude oil for this study was obtained from a North-American source and stored under argon. The asphaltene was precipitated from the corresponding heavy crude oil using n-heptane (HPLC grade, Merck, Germany). To obtain the asphaltene, a SARA fractionation procedure was used.32 Here, 30 mL of n-heptane per gram of crude oil was added, and the mixture was refluxed for 2 h in a Soxhlett apparatus. The precipitated portion was filtered and dried under inert gas flow. The sample and filter were extracted with 300 mL toluene (HPLC-grade, Overlack, Germany) in a Soxhlett apparatus until no color changes were observed. The redissolved asphaltene fraction was evaporated under vacuum and dried under a continuous stream of nitrogen. The asphaltene content of the heavy crude oil was 9.9 wt %. Elemental Analysis. The C, H, N, and S contents of the asphaltene were determined in a Vario Elementar EL elemental analyzer using a TCD detector (Elementar Analysensysteme GmbH, Hanau, Germany). The oxygen content was estimated by the difference. Solid-State 13C NMR Analysis. Solid-state 13C CP-MAS NMR spectra were recorded on a Bruker Avance III 300 spectrometer, equipped with a double-bearing probe. The ZrO2 rotor (7 mm external diameter) was charged with the sample and sealed by a Kel-F inset. Optimal contact time for 13C CP was 1−2 ms. The spinning rate was 6.5 kHz. The external standard for 13C NMR was adamantane (δ CH2: 38.40 ppm, relative to TMS). To determine the relative amount of the 5258
dx.doi.org/10.1021/ac300133p | Anal. Chem. 2012, 84, 5257−5267
Analytical Chemistry
Article
Excel and Origin for data evaluation and preparation of the figures shown. Venn Diagrams, provided in Figure 5, were computed with Venn Diagram Plotter, version 1.4.3740.38143 (March 29, 2010; Kyle Littlefeld and Matthew Monroe, PNNL, Richland, WA, U.S.A.) from calculated data.38 The elemental compositions (C, H, N, O, S) for the lists of the assigned molecular formulas, derived from the different ionization methods were calculated as nonweighted averages using the equation
protonated aromatic carbons, TOSS and TOSS/NQS (Total Suppression Spectra/Non-Quaternal Suppression) methods were separately applied and the difference was calculated based on the spectra. The mole fraction of bridgehead carbons (Table 1, Row 8) and the average aromatic cluster size (Table 1, Row 9) were calculated based on the procedure described by Solum et al.33 Sample Preparation for FT-ICR MS Analysis. Asphaltene (1.5 mg) was dissolved in 1.5 mL toluene (HPLC-grade, Acros Organics). For the electrospray ionization (ESI) measurements, the stock solution (200 μL) was further diluted with methanol (UPLC-MS-grade, Biosolve, Netherland) to produce a final concentration of 100 μg/mL for the ESI and heated electrospray ionization (H-ESI) measurement. For the atmospheric pressure chemical ionization (APCI), atmospheric pressure photo ionization (APPI) and atmospheric pressure laser ionization (APLI) measurements only toluene was used for dilution to a final concentration of 100 μg/mL. FT-ICR MS Analysis. Mass analysis was performed on a 12 T LTQ FT-ICR MS (Thermo Fisher, Bremen, Germany) equipped with commercially available ESI/H-ESI/APCI/APPI sources. The data were collected and processed with the LTQ FT Ultra 2.5.5 (Thermo Fisher, Bremen, Germany) data acquisition system. The spectra were collected in positive mode using ESI, H-ESI, APCI, APPI, and APLI. For the ESI and HESI measurements, sample was infused at a flow rate of 2 μL/min, and the ions were generated from a microelectrospray source equipped with a metal-ESI needle. Typical ESI(+) conditions were as follows: needle voltage = 3.8 kV, sheath gas = 5 arbitrary units. The H-ESI source was operated at 60 °C and 4 kV. In case of the APPI, APCI and the APLI measurements the sample was infused with the flow rate of 20 μL/min and evaporated at 240 °C with a continuous sheath gas flow of 16 (arbitrary units). APCI current was set to 5 μA. For the APLI measurements pulsed laser radiation (50 Hz, 8 mJ) was obtained from a KrF* excimer laser (ATL Lasertechnik GmbH, Wermelskirchen, Germany), radiating at the wavelength of 248 nm. The sample was injected through the APCI nebulizer and the generated cloud was ionized with the unfocused laser beam positioned between the MS orifice and the exit of the ion source. In order to avoid the presence of multimers (aggregates), a gentle (35 V) source induced dissociation was also applied.13 The spectra were acquired with a low resolution linear ion trap (m/z 200−1500) and with a 12 T ultrahigh resolution FT-ICR MS with the mass range of m/z 200−1000 using the “spectral stitching” method.34,35 The observed resolution was approximately 650k (at m/z 400). The ion accumulation time was defined by the automatic gain control (AGC), which was set to 500k.36 Data Analysis. The mass spectra were externally calibrated and resulted in a mass accuracy of less than 1 ppm with an average error of 0.04, 0.4, 0.07, 0.01, and 0.12 ppm in ESI, HESI, APCI, APPI and APLI, respectively. The peak lists were converted to molecular formulas by Composer (Sierra Analytics, U.S.A.). The following chemical constraints were applied: Number of H unlimited, 0 < C
