Energy & Fuels 1999, 13, 401-410
401
Hydrous Pyrolysis of Polycyclic Aromatic Hydrocarbons and Implications for the Origin of PAH in Hydrothermal Petroleum Thomas M. McCollom,†,‡,§ Bernd R. T. Simoneit,*,‡ and Everett L. Shock† GEOPIG, Earth and Planetary Sciences, Washington University, St. Louis, Missouri 63130, and Petroleum and Environmental Geochemistry Group, College of Oceanic and Atmospheric Sciences, Oregon State University, Corvallis, Oregon 97331 Received April 21, 1998. Revised Manuscript Received November 23, 1998
Polycyclic aromatic hydrocarbons (PAH) are found at high concentrations in thermally altered organic matter and hydrothermally generated petroleum from sediment-covered seafloor hydrothermal systems. To better understand the factors controlling the occurrence of PAH in thermally altered environments, the reactivities of two PAH, phenanthrene and anthracene, were investigated in hydrothermal experiments. The compounds were heated with water at 330 °C in sealed reaction vessels for durations ranging from 1 to 17 days. Iron oxide and sulfide minerals, formic acid, or sodium formate were included in some experiments to vary conditions within the reaction vessel. Phenanthrene was unreactive both in water alone and in the presence of minerals for up to 17 days, while anthracene was partially hydrogenated (5-10%) to di- and tetrahydroanthracene. In the presence of 6-21 vol % formic acid, both phenanthrene and anthracene reacted extensively to form hydrogenated and minor methylated derivatives, with the degree of hydrogenation and methylation increasing with the amount of formic acid. Phenanthrene was slightly hydrogenated in sodium formate solutions. The hydrogenation reactions could be readily reversed; heating a mixture of polysaturated phenanthrenes resulted in extensive dehydrogenation (aromatization) after 3 days at 330 °C. While the experiments demonstrate that reaction pathways for the hydrogenation of PAH under hydrothermal conditions exist, the reactions apparently require higher concentrations of H2 than are typical of geologic settings. The experiments provide additional evidence that PAH may be generated in hydrothermal systems from progressive aromatization and dealkylation of biologically derived polycyclic precursors such as steroids and terpenoids. Furthermore, the results indicate that PAH initially present in sediments or formed within hydrothermal systems are resistant to further thermal degradation during hydrothermal alteration.
Introduction Hydrous pyrolysis has become a frequently used technique to examine chemical processes taking place during oil generation and maturation of organic matter in the presence of water.1,2 In hydrous pyrolysis experiments, hydrothermal conditions are employed under the presumption that high temperatures will allow processes to be examined on laboratory time scales that occur at lower temperatures but over much longer time periods in geologic settings such as sedimentary basins. There are many natural settings in which similar heating and rapid thermal maturation of organic matter occurs. For instance, circulation of hydrothermal fluids through organic-rich sediments at sediment-covered seafloor spreading centers, such as in Guaymas Basin in the Gulf of California or in Middle Valley on the East †
Washington University. Oregon State University. Present address: MS #4, Woods Hole Oceanographic Institution, Woods Hole, MA 02540. (1) Lewan, M. D. In Organic Geochemistry: Principles and Applications; Engel, M. H.; Macko, S. A., Eds.; Plenum Press: New York, 1993; pp 419-442. (2) Lewan, M. D. Geochim. Cosmochim. Acta 1997, 61, 3691-3723. ‡ §
Pacific Rise, results in hydrothermal alteration of organic matter and the generation of hydrothermal petroleum. 3-7 Sediment-covered hydrothermal systems may be thought of as “natural hydrous pyrolysis experiments”, and examination of thermal maturation of organic matter in these systems may help to elucidate reactions taking place in more conventional oil reservoirs as well as increase the understanding of the chemical reactions taking place during hydrous pyrolysis of sediments in the laboratory. Unsubstituted polycyclic aromatic hydrocarbons (PAH) occur at high concentrations in thermally altered sediments and hydrothermal petroleum relative to their (3) Simoneit, B. R. T. The Role of Heat in the Development of Energy and Mineral Resources in the Northern Basin and Range Province, Geothermal Res. Council, Spec. Rpt. 13, Davis, CA, 1983; pp 215241. (4) Simoneit, B. R. T. Can. J. Earth Sci., 1985, 22, 1919-1929. (5) Simoneit, B. R. T. In Organic Matter in Hydrothermal Systemss Petroleum Generation, Migration and Biogeochemistry; Simoneit, B. R. T., Ed.; Appl. Geochem. 1990, 5, 3-15. (6) Simoneit, B. R. T. In Organic Geochemistry: Principles and Applications; Engel M. H.; Macko S. A., Eds.; Plenum Press: New York, 1993; pp 397-418. (7) Gieskes, J. M.; Simoneit, B. R. T.; Brown, T.; Shaw T.; Wang, Y.-C.; Magenheim, A. Can. Mineral. 1988, 26, 589-602.
10.1021/ef980089i CCC: $18.00 © 1999 American Chemical Society Published on Web 01/15/1999
402 Energy & Fuels, Vol. 13, No. 2, 1999
abundance in organic matter from sedimentary basins and typical crude oils, with concentrations comparable to those found in highly polluted soils and industrial areas.8,9 In addition, the relative concentrations of PAH increase with depth in the altered sediments and in proximity to intruded igneous sills while the proportion of substituted to unsubstituted PAH decreases (e.g., methylphenanthrenes to phenanthrene).9 These trends in PAH abundance are interpreted to reflect increasing thermal maturation of the organic matter in altered sediments. The amounts of PAH observed in hydrothermal sediments suggests that most of the PAHs originate in situ during alteration rather than representing a thermally resistant fraction of the initial sedimentary organic matter whose relative concentration increases due to decomposition of less resistant components during thermal maturation. PAH may be generated within hydrothermal systems from either of two processes: (1) transformation of polycyclic, biologically generated precursor compounds such as steroids or terpenoids to PAH through progressive aromatization and dealkylation reactions (e.g., Figure 1) or (2) pyrolytic fragmentation of kerogen or other organic compounds followed by reforming to PAH. To examine the processes responsible for the observed concentrations of PAH in hydrothermal systems, we have initiated an experimental study on the reactivity of polycyclic compounds under hydrothermal conditions and report here the initial results on the reactivities of two PAHs, phenanthrene and anthracene. The present study was primarily concerned with two issues. First, we sought to examine the thermal stability and reactivity of PAH in hydrothermal environments. Previous hydrous pyrolysis studies have found phenanthrene to be unreactive in water at high temperatures when heated alone or in the presence of inorganic salts, CaCO3, or montmorillonite.10,11 However, other studies have shown that phenanthrene as well as anthracene may be extensively hydrogenated in the presence of H2 or other hydrogen donors.12-16 These results suggest the possibility that PAH may be partially hydrogenated within H2-rich hydrothermal systems, but hydrogenation has only been reported for anhydrous conditions with the PAH dissolved in organic solvents, and the effect of water on hydrogenation reactions remained uncertain. With increasing temperature, the dielectric constant of water decreases and the solvent properties of water more closely resemble those of organic solvents. Thus, it is expected that hydrogenation reactions behave similarly in hot water as in organic solvents, unless water is involved in the reaction directly. In addition, experiments with other organic compounds under hydrothermal conditions have shown that reactions be(8) Kawka, O. E.; Simoneit, B. R. T. Appl. Geochem. 1990, 5, 1727. (9) Kawka, O. E.; Simoneit, B. R. T. Org. Geochem. 1994, 22, 947978. (10) Siskin, M.; Katritzky, A. R.; Balasubramanian, M. Energy Fuels 1991, 5, 770-771. (11) Siskin, M.; Katritzky, A. R.; Balasubramanian, M. Fuel 1993, 72, 1435-1444. (12) Benbenek, S.; Fedorynska, E.; Winiarek, P. Fuel 1994, 73, 1348-1353. (13) Ross, D. S.; Blessing, J. E. Fuel 1979, 58, 433-437. (14) Ross, D. S.; Blessing, J. E. Fuel 1979, 58, 438-442. (15) Yang, S.; Stock, L. M. Energy Fuels 1996, 10, 516-517. (16) Yang, S.; Stock, L. M. Energy Fuels 1996, 10, 1181-1186.
McCollom et al.
Figure 1. Representative reaction for the formation of PAH from a biological precursor. Shown is a schematic reaction for the transformation of cholestane to phenanthrene. Individual steps in the scheme follow reactions studied experimentally by Mackenzie et al.19 and Abbott et al.20-22 The reaction pathway shown is just one of many possible pathways by which phenanthrene may form from biological precursors in hydrothermal systems. In this scenario, the transformation from precursor to phenanthrene occurs through a series of dehydrogenation and dealkylation reactions, leading ultimately to the unsubstituted, fully aromatic PAH.
tween compounds involving loss or gain of H2 were sensitive to the concentration of H2 present in the reaction vessel.17,18 To test whether hydrogenation of PAH may proceed under hydrous conditions and to examine the sensitivity of hydrogenation to H2 concentrations, phenanthrene and anthracene reactivities were examined in hydrothermal experiments with a range of H2 concentrations provided by mineral buffers and decomposition of formic acid (see below). The second issue we examined was whether hydrothermal conditions favored formation of unsubstituted PAH from substituted and polysaturated polycyclic analogues. Previous laboratory studies related to the evolution of biomarkers in sedimentary basins have indicated that heating of alkylpolycyclic compounds such as steroid hydrocarbons under hydrothermal conditions results in progressive dehydrogenation and aromatization of the cyclic rings accompanied by loss of substituent groups (i.e., dealkylation). For instance, heating C29-monoaromatic steroid hydrocarbon produces C28-triaromatic steroid hydrocarbons,19-21 and hydrous (17) Seewald, J. S. Nature 1994, 370, 285-287. (18) Seewald, J. S. Mater. Res. Soc. Symp. Proc. 1997, 432, 317331.
Pyrolysis of Polycyclic Aromatic Hydrocarbons
Energy & Fuels, Vol. 13, No. 2, 1999 403
pyrolysis of 5R(H)-cholestane produces numerous dealkylated and unsaturated polycyclic compounds that include dimethylperhydrophenanthrene and dimethyloctahydrophenanthrene.22 With continued thermal maturation, dealkylation and dehydrogenation of such polycyclic compounds should lead to formation of unsubstituted, fully aromatic PAH as end products, suggesting a reaction pathway for the generation of PAH from biological precursors (Figure 1). Therefore, we subjected methylated polyhydrophenanthrenes to hydrothermal conditions to test whether these compounds were likely to react to form PAH. In some of the experiments, iron oxide and sulfide minerals were included in the reaction vessels. The intent of including these minerals is two-fold: (1) to provide potential catalytic surfaces resembling those found in metal-rich hydrothermal systems and (2) to control the oxidation state (H2 concentration) of the experimental environment. The mineral assemblages hematite-magnetite (HM) and pyrite-pyrrhotitemagnetite (PPM) buffer the oxidation state to a fixed value by producing or absorbing hydrogen through the reactions17,18
4Fe3O4 + 2H2O ) 6Fe2O3 + 2H2 magnetite hematite
(1)
6FeS + 4H2O ) 3FeS2 + Fe3O4 + 4H2 (2) pyrrhotite pyrite magnetite which affect the oxygen concentration through the water disproportionation reaction
H2 + 0.5O2 ) H2O
(3)
At the temperature used in these experiments (330 °C), the presence of the hematite-magnetite mineral pair buffers the activity of H2(aq) (aH2) in the system to log aH2 ) -4.5, while the presence of the PPM assemblage produces greater reducing conditions (log aH2 ) -3.0) (calculated using SUPCRT92).23 Equivalent oxygen fugacities (fO2) are log fO2 ) -28.8 for HM and -31.7 for PPM. The use of minerals in the present experiments differs somewhat from previous experiments by other researchers in which minerals have been added only to examine their potential catalytic properties. In additional experiments, formic acid (or sodium formate) was added as a source of H2 and carbon. At the temperatures of the experiments, formic acid rapidly decomposes to form primarily H2 and carbon dioxide: 24-26
HCOOH w H2 + CO2 formic acid
(4)
(19) Mackenzie, A. S.; Hoffman, C. F.; Maxwell, J. R. Geochim. Cosmochim. Acta 1981, 45, 1345-1355. (20) Abbott, G. D.; Lewis, C. A.; Maxwell, J. R. Philos. Trans. R. Soc. London A 1985, 315, 107-122. (21) Abbott, G. D.; Lewis, C. A.; Maxwell, J. R. Nature 1985, 318, 651-653. (22) Abbott, G. D.; Bennett, B.; Petch, G. S. Geochim. Cosmochim. Acta 1995, 59, 2259-2264. (23) Johnson, J. W., Oelkers, E. H.; Helgeson, H. C. Comput. Geosci. 1992, 18, 899-947. (24) Elliot, D. C.; Sealock, L. J., Jr. Ind. Eng. Chem. Prod. Res. Dev. 1983, 22, 426-431. (25) Elliot, D. C.; Hallen, R. T.; Sealock, L. J., Jr. Ind. Eng. Chem. Prod. Res. Dev. 1983, 22, 431-435.
In geologic systems, formic acid may form as an intermediate in the conversion of carbon monoxide and water to CO2 and H2.27,28
CO + H2O S HCOOH S H2 + CO2 formic acid
(5)
This reaction is known in chemical industry as the water-gas shift reaction. Experimental Section The hydrothermal experiments were carried out in sealed, 316 stainless steel vessels constructed of Sno-Trik highpressure couplings using methods similar to those described by Leif29 and Leif and Simoneit.30 The internal capacity of the vessels was approximately 200 µL. The vessels were charged with phenanthrene or anthracene, deionized water, and, when appropriate, minerals, formic acid, or sodium formate. The reaction vessels were sealed and placed immediately in an oven for the duration of the experiment. All experiments reported here were performed at 330 ( 3 °C. During heating, the fluid would have expanded to fill the entire reaction vessel volume with a fluid density corresponding to the amount of charge added in a particular experiment. All samples were prepared with sufficient water added to ensure that a single, liquid phase would occupy the entire internal volume of the vessel at the final temperature.1 It was not possible with our apparatus to control or monitor the pressure inside the reaction vessels, however, based on the amount of water added, the pressures would be in the range of 600-1000 bar, which is well within the tolerances of the reaction vessels. Upon removal from the oven, the vessels were allowed to cool to room temperature by standing in the lab. The vessels were generally opened the day following removal from the oven and in all cases within 24 h. The reaction vessels were cooled on ice before opening to induce condensation/precipitation of reaction products, which were frequently evident as a white globular precipitate floating on the water. After opening, each sample was transferred by micropipet to a glass vial and the reaction vessel rinsed 4-5 times with dichloromethane (DCM), which was added to the glass vial to yield a total of approximately 1 mL of DCM extract. In experiments where formic acid or sodium formate was added to the reaction vessel, residual pressure within the vessel (due to CO2 and other volatiles generated during the experiment) sometimes caused a portion of the sample to be expelled onto the internal walls of the reactor vessel during opening. In these cases, the internal walls were also rinsed with DCM and the rinse added to the total extract. We were unable to capture or measure volatile products of the experiments. Analysis of the products was performed by gas chromatography (GC). Peak identifications were confirmed by known standards and by gas chromatography-mass spectrometry (GC-MS) analysis of a few select samples. The GC analyses were carried out on a Hewlett-Packard 5890 gas chromatograph using a 25 m × 0.2 mm i.d. fused silica capillary column coated with DB-5 (J and W, Inc.). The GC-MS analyses were conducted on a Hewlett-Packard model 5973 MSD quadrupole mass spectrometer coupled to a Hewlett-Packard 6890 GC. The GC-MS instrument was used with linear scanning (50-600 da, 1.2 s/decade) and in the electron impact ionization (70 eV) mode. Sample introduction was by on-column injection onto a (26) Maitlis, P. M.; Long, H. C.; Quyoum, R.; Turner, M. L.; Wang, Z. Chem. Commun. 1996, 1-8. (27) Horvath, I. T.; Siskin, M. Energy Fuels 1991, 5, 932-933. (28) Giggenbach, W. F. Geochim. Cosmochim. Acta 1997, 61, 37633785. (29) Leif, R. N. Ph.D. Thesis, Oregon State University, 1993. (30) Leif, R. N.; Simoneit, B. R. T. Origins Life Evol. Biosphere 1995, 25, 417-429.
404 Energy & Fuels, Vol. 13, No. 2, 1999
McCollom et al.
Table 1. Initial Compositions for Hydrothermal Experiments Carried Out at 330 °C with Phenanthrene as the Starting Compound experiment phenanthrene alone
duration (days)
phenanthrene (mg)
H2O (µL)
3
0.8-1.3
150
hematite/magnetite (HM)
17
pyrite-pyrrhotite-magnetite (PPM)
17
6 vol % HCOOH 13 vol % HCOOH 21 vol % HCOOH
3 3 3
17 mg 34 mg 49 mg
3 3 3
Phenanthrene with Minerals 0.8 130 0.8
130
Phenanthrene + Formic Acid 0.8 140 0.8 130 0.8 120
other reactants none 15 mg of hematite 20 mg of magnetite 40 mg of PPMa 10 µL of 88% formic acid 20 µL of 88% formic acid 30 µL of 88% formic acid
Phenanthrene + Sodium Formate 0.8 140 17.1 mg of HCOONa 0.8 130 34.4 mg of HCOONa 0.9 120 48.7 mg of HCOONa
Time series- Phenanthrene + 13 vol % Formic Acid 1 0.8 130 20 µL of 88% formic acid 3 0.8 130 20 µL of 88% formic acid 8 0.8 130 20 µL of 88% formic acid 14 0.8 130 20 µL of 88% formic acid
1 day 3 days 8 days 14 days mix alone with HM
3 3
Reversal Experiments 140 140
with PPM
3
140
hydrogenated/methylated mixb hydrogenated/methylated mixb+ 15 mg each of hematite and magnetite hydrogenated/methylated mixb+ 30 mg of PPMa
PPM mix is composed of 7 g of pyrrhotite, 5 g of pyrite, and 5 g of magnetite. b Hydrogenated/methylated phenanthrene mix contained approximately 52% phenanthrene, 21% dihydrophenanthrenes, 11% tetrahydrophenanthrenes, 13% octahydrophenanthrenes, and 0.5% perhydrophenanthrenes, with 1.6% methyl- and 0.4% dimethylphenanthrenes and hydrogenated analogues. a
Table 2. Initial Compositions for Hydrothermal Experiments Conducted at 330 °C with Anthracene as the Starting Compound experiment anthracene alone
a
duration (days)
anthracene (mg)
H2O (µL)
3
0.7-1.1
150
none
other reactants
hematite/magnetite (HM)
Anthracene with Minerals 17 0.7
130
pyrite-pyrrhotite-magnetite (PPM)
17
130
15 mg of hematite 20 mg of magnetite 40 mg of PPMa
6 vol % HCOOH 21 vol % HCOOH
Anthracene + Formic Acid 3 0.7-1.1 3 0.9-1.1
140 120
10 µL of 88% formic acid 30 µL of 88% formic acid
0.9
PPM mix is composed of 7 g of pyrrhotite, 5 g of pyrite, and 5 g of magnetite.
DB-1 capillary column (30 m × 0.25 mm i.d.). The GC and GC-MS operating conditions were as follows: isothermal at 65 °C for 2 min (GC only), temperature programmed from 65 to 310 °C at 4 °C per min, isothermal hold at 310 °C for 3060 min, and using helium as carrier gas. The injector temperature was 290 °C, and the GC detector was at 325 °C. Mass spectrometric data were acquired and processed using the Chemstation data system, and compounds were identified by GC and GC-MS comparison with authentic standards and characterized mixtures. Quantification of reaction products was performed by establishing GC response factors for phenanthrene and anthracene by injection of solutions of known concentration and assuming that hydrogenated and methylated derivatives of these compounds had the same response factors as the parent compound. This appears to be a reasonable approximation since response factors determined from injection of known amounts of perhydrophenanthrene and 1-methylphenanthrene varied by
