Heavy Petroleum Composition. 5. Compositional ... - ACS Publications

Dec 28, 2012 - †National High Magnetic Field Laboratory, and ‡Future Fuels ... resonance mass spectrometry (FT-ICR MS) provided compositional data...
1 downloads 0 Views 3MB Size
Article pubs.acs.org/EF

Heavy Petroleum Composition. 5. Compositional and Structural Continuum of Petroleum Revealed David C. Podgorski,†,‡ Yuri E. Corilo,† Leonard Nyadong,† Vladislav V. Lobodin,†,‡ Benjamin J. Bythell,† Winston K. Robbins,‡,∥ Amy M. McKenna,†,‡ Alan G. Marshall,†,§ and Ryan P. Rodgers*,†,‡,§ †

National High Magnetic Field Laboratory, and ‡Future Fuels Institute, Florida State University, 1800 East Paul Dirac Drive, Tallahassee, Florida 32310-4005, United States § Department of Chemistry and Biochemistry, Florida State University, 95 Chieftain Way, Tallahassee, Florida 32306-4390, United States S Supporting Information *

ABSTRACT: Twenty-five years ago, Boduszynski et al. conducted a comprehensive study of heavy oil composition and concluded that crude oil composition increases gradually and continuously with regard to aromaticity, molecular weight, and heteroatom content from the light distillates to non-distillables (the Boduszynski continuum model). Previous exhaustive characterization of heavy vacuum gas oil by Fourier transform ion cyclotron resonance mass spectrometry (FT-ICR MS) provided compositional data that strongly supports the continuum model. However, when the molecular formulas obtained by FT-ICR MS for the distillates and asphaltenes from the same parent crude oil are plotted as double bond equivalents (DBE) versus carbon number, a gap appears between the compositional space of “asphaltenes” and “maltenes”, in contradiction to the Boduszynski−Altgelt model. Here, a heavy distillate cut (atmospheric equivalent boiling point of 523−593 °C) is fractionated according to the number of aromatic rings by HPLC-2. The C7-deasphalted whole oil (C7-DAO), its pentane soluble/insoluble fractions, and each of their ring number fractions are comprehensively characterized by atmospheric pressure photoionization (APPI) FT-ICR MS and tandem mass spectrometry (MS/MS). The HPLC-2 fractions from both the C5-soluble and C5insoluble C7-DAO represent a gradual and continuous progression that fills the compositional “gap” in carbon number and aromaticity between asphaltenes and maltenes as a function of the increasing aromatic ring number, as predicted by Boduszynski. MS/MS results indicate that each ring number fraction comprises both island and archipelago structural motifs. FT-ICR MS reveals a continuum in carbon number and aromaticity. The C5-insoluble C7-DAO components have a similar structure but with higher-order fused ring core structures and are composed of a higher proportion of archipelago structures than the C5-soluble C7-DAO components. Thus, fractionation by the aromatic ring number of “maltenic” and “asphaltenic” species from the C7solubles from a high boiling distillate validates the compositional continuum of petroleum components, and MS/MS exposes the aromatic building blocks of “maltenic” and “asphaltenic” species (structural continuum) that comprise island and archipelago structural motifs.



INTRODUCTION In the final installment of this five-part series, we address the compositional “gap” exposed by the comparison of the molecular level information obtained from the exhaustive characterization of distillate fractions up to and including heavy vacuum gas oil to their corresponding heptane-insoluble asphaltenes. In doing so, we address the structural moieties (island and archipelago) of asphaltenes and maltenes and rationalize the absence of species that lie in compositional space [defined as a plot of double bond equivalents (DBE = number of rings plus double bonds involving carbon) versus carbon number] above previously observed (more aromatic) maltenes but below (most aromatic) asphaltenic species that lie just below the polycyclic aromatic hydrocarbon (PAH) planar limit line.1,2 Asphaltenes are operationally defined by their insolubility in excess paraffin, designated by the n-alkane used for their precipitation (i.e., pentane or heptane). There is general agreement as to asphaltene molecular weight,3−14 number of fused rings per PAHs in asphaltene monomers, asphaltene aggregation number, critical nanoaggregation concentration (CNAC)15−17 (also addressed in part 3 of this © 2012 American Chemical Society

series), and asphaltene cluster size. However, despite many reports to the contrary,14,18−28 the specific structural elements (i.e., number and type of PAH building blocks) of asphaltenes and high boiling species (vacuum bottoms) remain unknown.29 Literature reports that strongly support a predominately island structural model for asphaltenes (and vacuum bottoms) rely heavily on a fundamental property of asphaltenes, selfassociation. In parts 3 and 4 of this five-part series, we postulate that aggregation restricts the remaining continuum of asphlatenes to those that contain island-type structures, because those island-type species ionize as nanoaggregates whose masses exceed the upper mass limit of most mass analyzers. However, time-of-flight mass spectrometers can readily detect asphaltene aggregates (4−25 kDa) and confirm that, at concentrations above the CNAC, a portion of the asphaltenes in a whole crude oil is directly observable as nanoaggregates. The aggregation explains their absence in mass spectra below Received: October 26, 2012 Revised: December 22, 2012 Published: December 28, 2012 1268

dx.doi.org/10.1021/ef301737f | Energy Fuels 2013, 27, 1268−1276

Energy & Fuels



m/z 3000, even in samples that contain a high mass fraction of asphaltenic material. It is worth noting that most liquid-phase asphaltene analyses are performed at analyte concentrations well above the asphaltene CNAC. Moreover, laser and surface desorption/ ionization techniques [laser desorption ionization (LDI), matrix-assisted laser desorption ionization (MALDI) and laser-induced acoustic desorption (LIAD) post-ionization] do not escape this limitation even if the solution concentration is intentionally kept well below the asphaltene CNAC, because evaporation of the solvent prior to sample introduction increases the asphaltene concentration well above the CNAC (in the evaporating droplets) before the mass spectral analysis can begin. The subsequent aggregation of asphaltenes will limit all compositional information to those species that are most loosely in the nanoaggregate. Extensive solvent stripping of heptane asphaltenes followed by mass spectral analysis strongly suggests that these weakly held species are composed almost exclusively of highly aromatic, island-type asphaltenes.30 Therefore, it is not surprising that all mass spectral data to date (including our own) support an island structural model for unaggregated asphaltenes. The purported absence of archipelago architectures for maltenes and asphaltenes based on presumed instability over geologic time was recently suggested. However, we are not familiar with any data that suggests inherent instability for archipelago structures. In fact, stability could result from their increased tendency to self-associate relative to large, alkyldeficient asphaltenes that are readily identified (in their monomeric form) by mass spectral analysis.1 Although evidence for islandic unaggregated asphaltenes is abundant in the literature, there is ever growing evidence for archipelago structures as well.31−34 Specifically, ruthenium-ioncatalyzed oxidation (RICO) produces diacids consistent with some degree of archipelago structure in asphaltenes. In combination with pyrolysis gas chromatography−mass spectrometry (GC−MS) results35,36 and the wide boiling range of thermally cracked asphaltenes, evidence for archipelago-type structures of asphaltenes is hard to ignore. The Boduszynski “continuum” model was originally inferred from then state-ofthe-art sector-based mass spectrometers that lacked the requisite resolution to uniquely determine elemental compositions throughout the observable mass range. The “continuum” model is therefore even more remarkable, because it was based on data not yet fully available. Here, a heptane-deasphalted oil (C7-DAO) heavy distillate cut (AEBP 523−593 °C) obtained by DISTACT distillation37 is analyzed by ultrahigh-resolution Fourier transform ion cyclotron resonance mass spectrometry (FT-ICR MS) before and after ring number fractionation by HPLC-2. The DAO is further separated into C5-soluble and C5-insoluble fractions and analyzed by a combined chromatographic/mass spectrometric method. C5-soluble C7-DAO and C5-insoluble C7DAO are each further separated on the basis of the number of aromatic rings by HPLC-2 separation in an effort to resolve the compositional “gap”. Each fraction is exhaustively characterized by FT-ICR MS to determine the compositional space occupied by each HPLC-2 isolated ring number fraction and by tandem mass spectrometry (MS/MS) experiments to reveal the component aromatic core building blocks. Thus, the composition and structural components are provided for both the “maltenic” (C5-soluble C7-DAO) and “asphaltenic” (C5insoluble C7-DAO) fractions of heavy petroleum.

Article

EXPERIMENTAL SECTION

Sample Preparation. Solvents were high-performance liquid chromatography (HPLC)-grade (Sigma-Aldrich Chemical Co., St. Louis, MO). A heptane-deasphalted distillate cut (523−593 °C) from a North American heavy crude oil was fractionated on the basis of solubility in n-pentane according to the IP 143/90 method. After each sample was dried under a stream of nitrogen, the pentane-soluble fraction (“C5-maltenes”) was reconstituted in pentane and the C5insoluble material (“C5-asphaltenes”) was reconstituted in a 40:60 (v/ v) mixture of toluene/hexanes to a final concentration of 400 mg mL−1 for HPLC-2 separation. Seven HPLC-2 fractions were collected in preweighed vials and desolvated with dry nitrogen. Each fraction was diluted with toluene to a final concentration of 250 μg mL−1 for the atmospheric pressure photoionization (APPI) analysis. Mass recovery for the C5-insoluble material was approximately 77%. HPLC-2 Separation. Each separation was performed with a Waters e2695 Alliance separation module equipped with a photodiode array detector and a Waters fraction collector III (Waters Corp., Milford, MA). Two columns were used in tandem for all separations. The first column was an ES Industries RingSep 5 μm, 60 Å, 25 cm × 4.6 mm (ES Industries, West Berlin, NJ) followed by a Waters Spherisorb 5 μm, SCX in Ag form, 4.6 × 250 mm column (Waters Corp., Milford, MA). The method was similar to that previously reported by W. K. Robbins.38 APPI. A modified Thermo-Fisher APPI source (Thermo-Fisher Scientific, Inc., San Jose, CA) was coupled to an FT-ICR mass spectrometer through a custom-built interface.39,40 Each sample was directly infused at 50 μL min−1 through a 5 mL Hamilton gas-tight syringe by a syringe pump to the heated vapor region (300 °C) of the APPI source, where a N2 sheath gas (50 psi) facilitated nebulization. Gas-phase neutrals were photoionized by a 10 eV (120 nm) ultraviolet krypton lamp (Syagen Technology, Inc., Tustin, CA). Atmospheric Pressure Laser-Induced Acoustic Desorption Chemical Ionization (AP/LIAD-CI). The AP/LIAD-CI MS interface and conditions for the analysis of saturated hydrocarbons (HCs) have been previously described.41,42 Briefly, the ion source consists of a sample probe fitted at one end with a hollow stainless-steel disk to sandwich a sample-containing piece of Ti foil (12.7 μM). A stainlesssteel cone tapered to 4 mm inner diameter is mounted on that end of the foil assembly to confine a flow of reagent gas and LIAD-vaporized analytes, which are directed toward the MS inlet. AP/LIAD-CI is performed by irradiating the sample-coated Ti foil from the back side with high-energy laser pulses from a Q-switched Nd:YAG laser (532 nm, pulsed at 10 Hz, with a pulse energy ranging from 53 to 105 mJ and a pulse duration of ∼5 ns). The laser-induced acoustic waves cause desorption of neutral analytes from the opposite side of the Ti foil into a flow of O2 reagent gas generated by a corona discharge. The reagent gas serves as a carrier gas for the entrainment of LIAD-vaporized analytes, purges the MS inlet of atmospheric gases, and generates reactive species required for chemical ionization (see Figure S1 of the Supporting Information). 9.4 T FT-ICR MS. Each HPLC-2 fraction was analyzed with a custom-built FT-ICR mass spectrometer equipped with a passively shielded 9.4 T 220 mm room-temperature bore superconducting magnet (Oxford Instruments, Abingdon, U.K.) located at the National High Magnetic Field Laboratory in Tallahassee, FL.43 Time-domain transient signals were collected and processed by a modular ICR data acquisition system (MIDAS).44 Positive ions were accumulated externally for 5 ns for AP/LIAD-CI and 25−500 ms for APPI45 in the second radio frequency (rf)-only octopole and collisionally cooled with helium prior to transfer through a rf-only octopole to a sevensegment, open cylindrical cell with capacitively coupled excitation electrodes similar to the configuration by Tolmachev et al.46,47 Chirp excitation (∼1400−70 kHz at a sweep rate of 50 Hz μs−1 and 360 Vp−p amplitude) accelerated the ions to a detectable cyclotron radius. Approximately 50−150 (APPI) or 25 (AP/LIAD-CI) time-domain acquisitions were co-added, Hanning-apodized, and zero-filled once prior to fast Fourier transform and magnitude calculation. Frequency was converted to m/z by the quadrupolar electric trapping potential 1269

dx.doi.org/10.1021/ef301737f | Energy Fuels 2013, 27, 1268−1276

Energy & Fuels

Article

approximation.48,49 Spectra were internally calibrated from an extended homologous alkylation series (compounds that differ in elemental composition by integer multiples of CH2) of high relative abundance prior to peak detection [>6σ baseline root-mean-square (rms) noise] and automated elemental composition assignment. Mass segments (25 Da wide) were isolated by a mass-resolving quadrupole prior to external ion accumulation and subsequently transferred to the FT-ICR cell for mass measurement and dissociation. Unit nominal mass segments were first isolated with a mass-resolving quadrupole (2 Da window) prior to transfer to the FT-ICR cell and narrowed to a 1 Da window by stored waveform inverse Fourier transform (SWIFT) excitation50,51 in the ICR cell before dissociation. MS/MS was performed to structurally characterize each HPLC-2 fraction. Specifically, infrared multiphoton dissociation52 (IRMPD; λ = 10.6 μm, 40 W, 10−700 ms irradiation, Synrad CO2 laser, Mukilteo, WA) was performed to fragment a mass-isolated segment of the broadband mass spectrum from the onset of fragmentation (10−50 ms) for increasing time periods of irradiation to complete dissociation (500− 700 ms).



Figure 1. Composite isoabundance-contoured plot of DBE versus carbon number for five distillate cuts and asphaltenes from Middle Eastern heavy crude oil, displayed as a single image.53 Note the compositional “gap” between “maltene” and “asphaltene” space, in contradiction to the Boduszynski continuum model. For reference, the planar PAH boundary (orange dashed line) defines the highest possible DBE for any given carbon number.2

RESULTS AND DISCUSSION The high-boiling fraction selected for analysis facilitates reconciliation of the compositional “gap” in a number of important ways. First, the fraction consists of material operationally defined as vacuum bottoms (asphaltenes and maltenes), those species that boil at temperatures higher than heavy vacuum gas oil (HVGO), and thus, contains species of higher, broader carbon number range and a higher, broader range of aromaticity (DBE = number of rings + double bonds involving carbon) than lower boiling fractions. Second, because it is a true boiling cut defined by a lower and upper boiling temperature, the compositional progression from less aromatic to more aromatic species within the boiling cut should obey the Boduszynski continuum theory. Simply, the highest carbon number species should exhibit the lowest DBE, and the lowest carbon number species should comprise the most aromatic (highest DBE)/most polar species. Thus, fractionation of the high boiling distillate by the number of aromatic rings should “walk” the compositional coverage of each ring number fraction along a diagonal line from high carbon number and low aromaticity (DBE) to lower carbon number and higher DBE (red dashed line in Figure 1).53 Here, we present data acquired by positive-ion APPI and AP/LIAD-CI to characterize the high boiling nonpolar species. However, APPI does ionize a subfraction of polar species as well. Finally, although compositionally complex, the distillate fraction can be analyzed at sufficient resolution and mass accuracy to ensure greater than 97% compositional coverage based on known, closely spaced mass differences of lower boiling, compositionally similar materials. Thus, we are confident that we address the largest possible fraction of the petroleome. Ring Separation of C5-Soluble C7-DAO. A fundamental principle of the Boduszynski model is that the addition of an aromatic ring must result in a concurrent loss of carbon to reside in the same distillation cut. Visualization of this progression is provided by Figure 2 that shows isoabundancecontoured images of DBE versus carbon number for the HC class for each ring-separated (1, 2, 3, 4, and 5+ ring) fraction of C5-soluble C7-DAO. As predicted by Boduszynski, an increase in aromaticity (DBE) and decrease in carbon number accompanies the addition of aromatic rings for each HPLC-2 fraction, progressing from the 1-ring fraction (top, left in Figure 2) to the 5+-ring fraction (middle, bottom in Figure 2). Consistent with a continuum, each ring fraction overlaps the previous in compositional space. The composite result

Figure 2. Isoabundance-contoured plots of DBE versus carbon number for members of the HC class for the HPLC-2 fractions (1, 2, 3, 4, and 5+ rings) from C5-soluble C7-insoluble deasphalted oil (C5-soluble C7-DAO). A red dashed oval is drawn along the 10% relative abundance contour line to highlight the compositional space boundary comprising 90% of the relative abundance for each ring number fraction. The trend in carbon number and aromaticity (DBE) is gradual and continuous through the ring progression, as predicted by Boduszynski. The compositional space occupied by the 3-, 4-, and 5+-ring fractions fills the compositional “gap” in the petroleome and, thus, completes the continuum.

(bottom, right in Figure 2) superimposed on the compositional data obtained from the unfractionated distillate reveals an uninterrupted continuum in carbon number and aromaticity from maltenic compositional space up and across the compositional “gap” (blue dashed lines) that terminates at the upper bound of the “gap” compositional space. The species that comprise the compositional space above the upper bound of the compositional “gap” but below the PAH planar limit (orange dashed line) are readily observed in the heptane asphaltene fraction (Figure 1) but have been removed from this (DAO) distillate cut prior to analysis. Thus, the compositional 1270

dx.doi.org/10.1021/ef301737f | Energy Fuels 2013, 27, 1268−1276

Energy & Fuels

Article

“gap” in the petroleome is artificial, caused by a combination of preferential aggregation (discussed later) and the high boiling point (in the vacuum bottoms) of species that occupy this compositional space. Species that comprise the gap can now be readily observed in high-boiling, vacuum bottom distillates and their HPLC-2 ring-separated samples. MS/MS of Ring-Separated C5-Soluble C7-DAO. Infrared multiphoton dissociation (IRMPD) was performed for each HPLC-2 fraction to determine the fundamental structural motif (island or archipelago) and catalog the aromatic building blocks of the high-boiling maltenes. The irradiation period was incremented from 10 to 500 ms (Figure 3). We can thus

highest relative abundance exhibit lower DBE relative to the isolated precursors from each class. Finally, after 500 ms of irradiation, the precursor ions are completely dealkylated to their aromatic cores. The DBE distribution for the fully dealkylated cores extends to lower DBE values than that of the mass-isolated species. Thus, archipelago structures are clearly present. The upper bound for the fully dealkylated cores matches that of the mass-isolated species and, therefore, confirms the presence of islandic structural motifs as well. Thus, isolation and fragmentation of the 2-ring fraction isolated by HPLC-2 (the minimum required for archipelago structures) results in the loss of a 0 and 1 ring and suggests that both island and archipelago structural motifs are present. Figure 4 shows a broadband mass spectrum (top), quadrupole mass-isolated segment (middle), and complete dealkylation products (bottom) for the C5-soluble C7-DAO ring fractions. The top row of images recapitulates the gradual and continuous progression through carbon number and DBE compositional space that results from the presence of additional aromatic rings. The mass-isolated segments were isoabundancescaled to highlight the most abundant species (in red) in each of the ring fractions. Aromaticity (reflected by DBE) increases and carbon number decreases in progressing from 1 to 5+ ring fractions, as expected. A blue dashed line denotes the highest relative abundance for the DBE distribution for each massisolated segment (middle in Figure 4), and a red dashed line denotes the completely dealkylated fragments (bottom in Figure 4). Note that the lowest DBE limit for the completely dealkylated species remains constant for all ring number fractions, an obvious byproduct of structural diversity (but lower DBE limit) for archipelago structures throughout the continuum. The 2-, 3-, 4-, and 5+-ring fractions exhibit a progressive increase in the DBE range for the aromatic cores as the ring number increases. Although the 5+-ring fraction displays the widest range of aromatic cores, the difference between the most abundant DBE values of the mass-isolated fraction and its fragments is a minimum and does not vary with the ring fraction, most likely because of the additional presence of polar (islandic) species in this fraction, to be addressed in future work. In proceeding from 1-ring to 5+-ring fractions, the structures become an increasingly heterogeneous mixture of island and archipelago motifs. Isolation/Fragmentation of Components Spanning a Single Nominal Mass Segment. Although the isolation of ions of a single mass from a mixture of tens of thousands of closely spaced masses is not currently possible, it is feasible to isolate and fragment ions spanning a single nominal mass range, to yield aromatic core building blocks. Therefore, we isolated ions of a single nominal mass (m/z 632) from the 4-ring maltene fraction by SWIFT excitation after quadrupole mass filtering (left in Figure 5) and irradiated them for periods of 100 ms (middle in Figure 5) and 500 ms (right in Figure 5). The mass-isolated species are shown in the left panel and highlighted (red boxes) after 100 and 500 ms of laser irradiation. The dealkylation patterns (horizontal streaks) from island precursors are clearly evident at DBE = 22 for the HC class and DBE = 24 and 17 for the S1 class. However, the highest DBE HC and S1 precursors are in low abundance (HC) or completely absent from the fragments. For both classes, the dealkylated DBE distributions are equal to or below their mass-isolated precursors. Thus, archipelago and island structures are both clearly present. Remarkably, the fragmentation of fewer than 10 highly abundant species of a single

Figure 3. DBE versus carbon number images for a quadrupole-isolated mass spectral segment for the 2-ring C5-soluble C7-DAO fraction (the lowest potentially archipelago ring fraction) before (left) and after (middle and right) IRMPD fragmentation. The mass-isolated segment is highlighted by the red rectangle and superimposed on the middle and right panels to facilitate visual comparison. The fragments produced after 50 ms of irradiation exhibit lower carbon number and DBE for both the HC (top) and S1 (bottom) classes, clear evidence for archipelago structural motifs. After 500 ms, the mass-isolated ions dissociate to core structures discussed later. The species of highest relative abundance in the completely dealkylated panels correspond to island (blue dashed line) and archipelago (located 3, 4, and 5 DBE below the dashed line) structures.

monitor fragmentation from the mass-isolated species (left in Figure 3) to their partially (middle in Figure 3) and completely (right in Figure 3) dealkylated cores. The mass-isolated precursor ions are highlighted in a red box and superimposed on the dealkylation panels. The DBE (horizontal blue dashed line) corresponding to the highest relative abundance HC class (top) and S1 class (bottom) facilitates a comparison of the mass-isolated precursors relative to their partial (middle) and completely dealkylated (right) fragments. The intermediate fragmentation patterns for both classes show an immediate decrease in the carbon number and DBE, characteristic of the loss of an aromatic moiety upon fragmentation (archipelago structures). Precursor ions range from DBE ≈ 7 to 24 for the HC class and from DBE ≈ 6 to 20 for the S1 class. After 50 ms of irradiation, the products for the HC class range from DBE ≈ 3 to 24, whereas the DBE distribution for the S1 class remains relatively unchanged. Furthermore, prior to irradiation, the species of highest relative abundance are located at DBE ≈ 13 (HC) and 11 (S1). After irradiation, the fragments with the 1271

dx.doi.org/10.1021/ef301737f | Energy Fuels 2013, 27, 1268−1276

Energy & Fuels

Article

Figure 4. Composite (DBE versus carbon number images) for the HC species identified for each ring number fraction (top row), quadrupole massisolated segments (middle row), and IRMPD fragments of the mass-isolated fragments (bottom row), exposing the aromatic core structures of the C5-soluble C7-DAO fraction. The DBE value for the species of highest relative abundance for the mass-isolated segment (middle row) for each ring number fraction is highlighted by a blue dashed line superimposed on the IRMPD fragments for visual comparison.

Core Analysis of Ring-Separated C5-Soluble C7-DAO. Figure 6 (top) shows aromatic core analysis for the 1−5+-ring C5-soluble C7-DAO fractions after complete dissociation by IRMPD. Ruiz-Morales proposed that asphaltenes are comprised mostly of fused six-membered aromatic rings,54 with a maximum of 10 fused rings. Ruiz-Morales provided optimized ring structures representative of those thought to exist in petroleum. A comparison of those structures (top in Figure 6) to those consistent with complete dissociation of the 1-, 2-, 3-, 4-, and 5+-ring HPLC-2 fractions enables tentative assignment of all of the abundant ring structures directly detected by mass spectrometry. The addition of one aromatic ring to the core structures of each HPLC-2 fraction is a result of the loss of hydrogen and the opening/closing of cycloalkane rings after IRMPD (see Figure S1 of the Supporting Information). Note that higher-order fused-ring cores are added (e.g., the 4-ring HPLC-2 fraction contains the same cores as fractions 1−3 in addition to higher-order fused cores not present in fractions 1− 3) through the addition of an aromatic ring for each successive HPLC-2 fraction. Furthermore, as the HPLC-2 fractions increase in aromaticity, the most abundant core subsequently progresses to higher-order fused rings but the lower DBE limit remains constant in each fraction. Figure 6 (bottom) shows proposed structures for the most abundant cores in each ringseparated C5-insoluble C7-DAO fraction after IRMPD. As stated previously, while the lower DBE limit for the cores remains constant, higher-order fused-ring cores are more abundant and extend to higher DBE than C5-soluble C7DAO. However, the most abundant cores for C5-insoluble C7DAO fractions 1−3 are identical to those identified in fractions 1−3 from C5-soluble C7-DAO Figure 6 (top), most likely because of entrainment of C5-soluble C7-DAO during C7 flocculation. The differences in gravimetric yield and more

Figure 5. Composite summary (HC and S1 species) from isolation of a single nominal mass (m/z 632) from the 4-ring C5-soluble C7-DAO fraction (left) after 100 ms (middle) and 500 ms (right) IR laser irradiation. A red rectangle highlights the mass-isolated species in the middle and right panels for visual comparison. Archipelago and island structures are clearly evident, because fragmentation of species of a single nominal mass is sufficient to reproduce a continuum of possible core structures.

common nominal mass suffices to regenerate the entire dealkylated aromatic ring distribution previously reported from a 25 Da precursor ion mass range (Figure 4). The wealth of elemental compositions that form a continuum in DBE and carbon number following dissociation of ions of a single nominal mass attests to the enormous structural diversity of heavy petroleum constituents. 1272

dx.doi.org/10.1021/ef301737f | Energy Fuels 2013, 27, 1268−1276

Energy & Fuels

Article

Figure 6. DBE versus carbon number images for the HC class for the ring-separated C5-soluble C7-DAO (top) and C5-insoluble C7-DAO (bottom) fractions after 500 ms IRMPD. Complete dealkylation of the mass-isolated species facilitates the assignment of proposed structures for the most abundant cores for each fraction.54 As the fractions progress in ring number, higher-order fused-ring cores are added. The 4- and 5+-ring fractions of the C5-insoluble C7-DAO fraction exhibit higher DBE and carbon number than the corresponding C5-soluble C7-DAO fraction.

detailed analysis of maltenes and asphaltenes isolated by aromatic ring number fractionation will be pursued in subsequent publications. However, we can address here the differences between C5-soluble C7-DAO and C5-insoluble C7DAO that elute in the same ring number fraction and occupy the same compositional space. Ring Separation and MS/MS Characterization of C5Insoluble C7-DAO. The deasphalted distillate fraction (C7DAO), previously isolated by heptane solubility, was subsequently dissolved in 40-fold volumetric excess of pentane to yield C5-insoluble C7-DAO. The procedure isolates “asphaltenic” species insoluble in pentane but soluble in heptane and, thus, defines the most “asphaltenic” components of the original “maltenes”. Because of solubility constraints, the initial conditions for the HPLC-2 procedure (that employs 100% hexane) were modified to allow for isolation of the 1-, 2-, 3-, 4-, and 5+-ring fractions from C5-insoluble C7-DAO. Alternatively, a similar fraction could be obtained from the original C5 asphaltenes by extraction with heptane (data not shown). Figure 7 shows HC class DBE versus carbon number images for each of the 1−5+-ring fractions obtained by the HPLC-2 procedure. The images are quite similar to those for C5-soluble C7-DAO (Figure 2), except that the C5-insoluble C7-DAO 5+-ring fraction exhibits wider DBE and carbon number distributions and penetrates deeper into the compositional space bound by the “gap” (dashed blue line) on the low

DBE side and the PAH limit (dashed orange line) on the upper side. The irradiation (IRMPD) period was systematically incremented to capture the composition of the precursor ions before and after complete dissociation to core structures of the ring-separated C5-insoluble C7-DAO. Figure 8 contains DBE versus carbon number images from MS/MS results for the C5-insoluble C7-DAO ring fractions (broadband mass spectrum at the top), the quadrupole massisolated precursors (middle), and complete dealkylation products (bottom). The compositional space spanned by each ring-separated C5-insoluble C7-DAO fraction is nearly identical to that covered by the corresponding ring-separated C5-soluble C7-DAO fraction Figure 4 (top). Furthermore, the progression in aromaticity and carbon number for the ringseparated fractions of C5-insoluble C7-DAO is gradual and continuous and follows the Boduszynski model. Thus, the model accurately describes the composition of non-distillable fractions of heavy petroleum. The fragmentation patterns (core structures) for each C5-insoluble C7-DAO fraction HPLC-2 Figure 8 (bottom) are nearly identical to those for the C5soluble C7-DAO fractions Figure 4 (bottom). The lower DBE limit remains constant, and higher-order fused-core structures result from the addition of an aromatic ring for each successive HPLC-2 fraction. Figure 9 shows the wide range of aromatic cores that comprise the building blocks for the 4-ring aromatic C51273

dx.doi.org/10.1021/ef301737f | Energy Fuels 2013, 27, 1268−1276

Energy & Fuels

Article

(fragmentation of an archipelago structure yields two or potentially more aromatic cores) are found at the mid- to upper portion of the diagonal core distribution. At the highest DBE values lie the islandic structures. Note that all species of high relative abundance can be assigned to cores calculated by Ruiz-Morales and that the maximum observed number of fused rings is 9 (in comparison to the Ruiz-Morales predicted maximum of 10).54 Finally, a comparison of the fragmentation products of massisolated C5-soluble C7-DAO (left in Figure 10) and C5insoluble C7-DAO (right in Figure 10) species exposes structural differences between the two 5+-ring fractions [C5 soluble (maltenic) and C5 insoluble (asphaltenic)] that occupy similar compositional space. Mass-isolated C5-insoluble C7DAO subjected to a short laser irradiation period (100 ms) yields product ions with a bimodal distribution along the PAH planar limit line (top, right in Figure 10), whereas the 100 ms dissociation products from C5-soluble C7-DAO yield a “comet” distribution composed of highly abundant high DBE islandic cores along with a low abundance “tail” of low DBE archipelago structures (top, left in Figure 10). Thus, although C5-soluble C7-DAO and C5-insoluble C7-DAO occupy identical compositional space, their structures are quite different. The same structural differences account for the 500 ms infrared dissociation products (bottom in Figure 10). The C5-insoluble C7-DAO cores extend to slightly higher DBE (i.e., larger aromatic cores) than C5-soluble C7-DAO. They also contain a greater relative abundance of low DBE archipelago fragments (highlighted in the red dashed oval). Thus, asphaltenes are composed of larger aromatic cores and a greater fraction of archipelago structures relative to the maltenes.

Figure 7. DBE versus carbon number images for the HC class for the HPLC-2 method ring fractions (1, 2, 3, 4, and 5+) from the C5insoluble C7-DAO fraction. A red dashed oval is drawn along the 10% relative abundance contour line to highlight the compositional space boundary comprising 90% of the relative abundance for each ring number fraction. The trend in carbon number and aromaticity (DBE) is gradual and continuous through the ring progression, as predicted by Boduszynski. The compositional space occupied by the 3-, 4-, and 5+-ring fractions fills the compositional “gap” in the petroleome and, thus, completes the continuum.

insoluble C7-DAO fraction after complete dissociation by IRMPD. Progression up the diagonal line (the PAH planar limit) identifies the small aromatic cores that were alkyl-linked (or cycloalkane) bridge structures (archipelago) prior to fragmentation. Their corresponding archipelago partners

Figure 8. Composite summary (DBE versus carbon number images) for the HC species identified in the ring number fractions (top row), quadrupole mass-isolated segments (middle row), and IRMPD fragments of the mass-isolated fragments (bottom row), exposing the aromatic core structures of the C5-insoluble C7-DAO fraction. The DBE value of highest relative abundance for the mass-isolated segment (middle row) for each ring number fraction is highlighted with a blue dashed line that is superimposed on the IRMPD fragments for visual comparison. As for the C5soluble C7-DAO fraction, fragmentation yields multiple simultaneous losses of carbon number and DBE, strongly suggesting archipelago structures. 1274

dx.doi.org/10.1021/ef301737f | Energy Fuels 2013, 27, 1268−1276

Energy & Fuels

Article

Figure 9. DBE versus carbon number image of the 4-ring C5-insoluble C7-DAO fraction after 500 ms of infrared irradiation, revealing the core structures of the mass-isolated species. Proposed structures54 reveal the structural diversity of core components within the 4-ring separated fraction.

by pentane) back into pentane. Surprisingly, every C7asphaltene pure ring number fraction (i.e., after separation of the various monomers) dissolved back into pentane (the solvent used to initially precipitate them). In the future, we shall extend to even higher boiling fractions, detailed mass balance accounting, and other solubility-drop routes to asphaltenes to further address asphaltene structure and solubility. A final issue is the nature and structure of the “bridges” between island cores in archipelago structures. Although the data presented here supports archipelago structures, much of the controversy between the island and archipelago structural theories could be resolved if archipelago structures are bridged by cycloalkane rather than linear alkane linkages. Thus, a single core could be rigid but potentially fragment upon thermal stress [see the cross-ring cleavage product (m/z 262.3) in Figure S1 of the Supporting Information]. To that end, a detailed characterization of alkyl- and cycloalkyl-substituted standards is currently underway.



ASSOCIATED CONTENT

S Supporting Information *

(+) AP/LIAD-CI FT-ICR MS DBE versus carbon number images for cholestane (top) and base oil (bottom) after 0 ms (left), 50 ms (middle), and 500 ms (right) infrared irradiation, showing the formation of an aromatic ring (DBE = 4) from cycloalkane species after complete dissociation (Figure S1). This material is available free of charge via the Internet at http://pubs.acs.org.

Figure 10. DBE versus carbon number images for the 5+-ring C5soluble C7-DAO (left) and C5-insoluble C7-DAO (right) fractions after 100 ms (top) and 500 ms (bottom) of infrared irradiation.





FUTURE DIRECTIONS In a manuscript published in Asphalt Paving Technology more than 2 decades ago, Boduszynski argued that once separated into purely monomeric (unassociated) components, asphaltenes would dissolve back into the paraffin used in their initial precipitation.55 In other words, it is the asphaltene nanoaggregate that is insoluble in the paraffin and not the “asphaltene” monomer. We have therefore begun to test the solubility of various ring number fractions (initially precipitated

AUTHOR INFORMATION

Corresponding Author

*Telephone: +1-850-644-2398. Fax: +1-850-644-1366. E-mail: [email protected]. Present Address ∥

Consultant, 10 Pogy Lane, Brunswick, Maine 04011, United States. Notes

The authors declare no competing financial interest. 1275

dx.doi.org/10.1021/ef301737f | Energy Fuels 2013, 27, 1268−1276

Energy & Fuels



Article

(29) Strausz, O. P.; Mojelsky, T. W.; Lown, E. M. Fuel 1992, 71 (12), 1355−1363. (30) Ehrmann, B. M.; Robbins, W. K.; Rodgers, R. P.; Marshall, A. G. Asphaltene co-precipitate molecular progression as a function of soxhlet extraction period, illuminated by negative ESI FT-ICR MS. Proceedings of the 58th American Society for Mass Spectrometry Annual Conference on Mass Spectrometry and Allied Topics; Salt Lake City, UT, May 23−27, 2010. (31) Artok, L.; Murata, S.; Nomura, M.; Satoh, T. Energy Fuels 1998, 12 (2), 391−398. (32) Su, Y.; Artok, L.; Murata, S.; Nomura, M. Energy Fuels 1998, 12 (6), 1265−1271. (33) Qian, K. N.; Edwards, K. E.; Mennito, A. S.; Freund, H.; Saeger, R. B.; Hickey, K. J.; Francisco, M. A.; Yung, C.; Chawla, B.; Wu, C.; Kushnerick, J. D.; Olmstead, W. N. Anal. Chem. 2012, 84 (10), 4544− 4551. (34) Gray, M. R. Energy Fuels 2003, 17 (6), 1566−1569. (35) Artok, L.; Su, Y.; Hirose, Y.; Hosokawa, M.; Murata, S.; Nomura, M. Energy Fuels 1999, 13 (2), 287−296. (36) Karimi, A.; Qian, K. N.; Olmstead, W. N.; Freund, H.; Yung, C.; Gray, M. R. Energy Fuels 2011, 25 (8), 3581−3589. (37) Boduszynski, M. M. Energy Fuels 1988, 2 (5), 597−613. (38) Robbins, W. K. J. Chromatogr. Sci. 1998, 36 (9), 457−466. (39) Purcell, J. M.; Hendrickson, C. L.; Rodgers, R. P.; Marshall, A. G. Anal. Chem. 2006, 78 (16), 5906−5912. (40) Purcell, J. M.; Rodgers, R. P.; Hendrickson, C. L.; Marshall, A. G. J. Am. Soc. Mass Spectrom. 2007, 18, 1265−1273. (41) Nyadong, L.; McKenna, A. M.; Hendrickson, C. L.; Rodgers, R. P.; Marshall, A. G. Anal. Chem. 2011, 83 (5), 1616−1623. (42) Nyadong, L.; Quinn, J. P.; Hsu, S. C.; Hendrickson, C. L.; Rodgers, R. P.; Marshall, A. G. Anal. Chem. 2012, 84 (16), 7131− 7137. (43) Kaiser, N. K.; Quinn, J. P.; Blakney, G. T.; Hendrickson, C. L.; Marshall, A. G. J. Am. Soc. Mass Spectrom. 2011, 22 (8), 1343−1351. (44) Blakney, G. T.; Hendrickson, C. L.; Marshall, A. G. Int. J. Mass Spectrom. 2011, 306 (2−3), 246−252. (45) Senko, M. W.; Hendrickson, C. L.; Emmett, M. R.; Shi, S. D. H.; Marshall, A. G. J. Am. Soc. Mass Spectrom. 1997, 8 (9), 970−976. (46) Tolmachev, A. V.; Robinson, E. W.; Wu, S.; Kang, H.; Lourette, N. M.; Pasa-Tolic, L.; Smith, R. D. J. Am. Soc. Mass Spectrom. 2008, 19 (4), 586−597. (47) Kaiser, N. K.; Savory, J. J.; McKenna, A. M.; Quinn, J. P.; Hendrickson, C. L.; Marshall, A. G. Anal. Chem. 2011, 83 (17), 6907− 6910. (48) Ledford, E. B.; Rempel, D. L.; Gross, M. L. Anal. Chem. 1984, 56 (14), 2744−2748. (49) Shi, S. D. H.; Drader, J. J.; Freitas, M. A.; Hendrickson, C. L.; Marshall, A. G. Int. J. Mass Spectrom. 2000, 196, 591−598. (50) Marshall, A. G.; Wang, T. C. L.; Ricca, T. L. J. Am. Chem. Soc. 1985, 107 (26), 7893−7897. (51) Guan, S. H.; Marshall, A. G. Int. J. Mass Spectrom. 1996, 157, 5− 37. (52) Little, D. P.; Speir, J. P.; Senko, M. W.; Oconnor, P. B.; Mclafferty, F. W. Anal. Chem. 1994, 66 (18), 2809−2815. (53) McKenna, A. M.; Blakney, G. T.; Xian, F.; Glaser, P. B.; Rodgers, R. P.; Marshall, A. G. Energy Fuels 2010, 24, 2939−2946. (54) Ruiz-Morales, Y. J. Phys. Chem. A 2002, 106 (46), 11283− 11308. (55) Boduszynski, M. M.; McKay, J. F.; Latham, D. R. Asphalt Paving Technology 1980, 49, 123−143.

ACKNOWLEDGMENTS The work was supported by the National Science Foundation (NSF) Division of Materials Research through DMR-06-54118, the State of Florida, the Florida State University Future Fuels Institute, and the BP/The Gulf of Mexico Research Initiative.



REFERENCES

(1) Purcell, J. M.; Merdrignac, I.; Rodgers, R. P.; Marshall, A. G.; Gauthier, T.; Guibard, I. Energy Fuels 2010, 24, 2257−2265. (2) Hsu, C. S.; Lobodin, V. V.; Rodgers, R. P.; McKenna, A. M.; Marshall, A. G. Energy Fuels 2011, 25 (5), 2174−2178. (3) Hortal, A. R.; Martinez-Haya, B.; Lobato, M. D.; Pedrosa, J. M.; Lago, S. J. Mass Spectrom. 2006, 41 (7), 960−968. (4) Martinez-Haya, B.; Hortal, A. R.; Hurtado, P.; Lobato, M. D.; Pedrosa, J. M. J. Mass Spectrom. 2007, 42 (6), 701−713. (5) Hortal, A. R.; Hurtado, P.; Martinez-Haya, B.; Mullins, O. C. Energy Fuels 2007, 21 (5), 2863−2868. (6) Pomerantz, A. E.; Hammond, M. R.; Morrow, A. L.; Mullins, O. C.; Zare, R. N. J. Am. Chem. Soc. 2008, 130 (23), 7216−7217. (7) Pomerantz, A. E.; Hammond, M. R.; Morrow, A. L.; Mullins, O. C.; Zare, R. N. Energy Fuels 2009, 23, 1162−1168. (8) Klein, G. C.; Kim, S.; Rodgers, R. P.; Marshall, A. G. Energy Fuels 2006, 20 (5), 1965−1972. (9) Klein, G. C.; Kim, S.; Rodgers, R. P.; Marshall, A. G.; Yen, A. Energy Fuels 2006, 20 (5), 1973−1979. (10) Hughey, C. A.; Rodgers, R. P.; Marshall, A. G.; Qian, K. N.; Robbins, W. K. Org. Geochem. 2002, 33 (7), 743−759. (11) Pinkston, D. S.; Duan, P.; Gallardo, V. A.; Habicht, S. C.; Tan, X. L.; Qian, K. N.; Gray, M.; Mullen, K.; Kenttamaa, H. I. Energy Fuels 2009, 23, 5564−5570. (12) Cunico, R. L.; Sheu, E. Y.; Mullins, O. C. Pet. Sci. Technol. 2004, 22 (7−8), 787−798. (13) Qian, K. N.; Edwards, K. E.; Siskin, M.; Olmstead, W. N.; Mennito, A. S.; Dechert, G. J.; Hoosain, N. E. Energy Fuels 2007, 21 (2), 1042−1047. (14) Sabbah, H.; Morrow, A. L.; Pomerantz, A. E.; Zare, R. N. Energy Fuels 2011, 25 (4), 1597−1604. (15) Andreatta, G.; Bostrom, N.; Mullins, O. C. Langmuir 2005, 21 (7), 2728−2736. (16) Andreatta, G.; Goncalves, C. C.; Buffin, G.; Bostrom, N.; Quintella, C. M.; Arteaga-Larios, F.; Perez, E.; Mullins, O. C. Energy Fuels 2005, 19 (4), 1282−1289. (17) Goncalves, S.; Castillo, J.; Fernandez, A.; Hung, J. Fuel 2004, 83 (13), 1823−1828. (18) Groenzin, H.; Mullins, O. C. J. Phys. Chem. A 1999, 103 (50), 11237−11245. (19) Groenzin, H.; Mullins, O. C. Energy Fuels 2000, 14 (3), 677− 684. (20) Buenrostro-Gonzalez, E.; Groenzin, H.; Lira-Galeana, C.; Mullins, O. C. Energy Fuels 2001, 15 (4), 972−978. (21) Groenzin, H.; Mullins, O. C.; Eser, S.; Mathews, J.; Yang, M. G.; Jones, D. Energy Fuels 2003, 17 (2), 498−503. (22) Badre, S.; Goncalves, C. C.; Norinaga, K.; Gustavson, G.; Mullins, O. C. Fuel 2006, 85 (1), 1−11. (23) Buch, L.; Groenzin, H.; Buenrostro-Gonzalez, E.; Andersen, S. I.; Lira-Galeana, C.; Mullins, O. C. Fuel 2003, 82 (9), 1075−1084. (24) Andrews, A. B.; Guerra, R. E.; Mullins, O. C.; Sen, P. N. J. Phys. Chem. A 2006, 110 (26), 8093−8097. (25) Guerra, R. E.; Ladavac, K.; Andrews, A. B.; Mullins, O. C.; Sen, P. N. Fuel 2007, 86 (12−13), 2016−2020. (26) Schneider, M. H.; Andrews, A. B.; Mitra-Kirtley, S.; Mullins, O. C. Energy Fuels 2007, 21 (5), 2875−2882. (27) Wargadalam, V. J.; Norinaga, K.; Lino, M. Fuel 2002, 81 (11− 12), 1403−1407. (28) Borton, D.; Pinkston, D. S.; Hurt, M. R.; Tan, X. L.; Azyat, K.; Scherer, A.; Tykwinski, R.; Gray, M.; Qian, K. N.; Kenttamaa, H. I. Energy Fuels 2010, 24, 5548−5559. 1276

dx.doi.org/10.1021/ef301737f | Energy Fuels 2013, 27, 1268−1276