190
Energy & Fuels 1995,9, 190-191
Unraveling the Kinetics of Petroleum Destruction by Using l,2J3C Isotopically Labeled Dopants Alan K. Burnham,* Hugh R. Gregg, and Robert L. Braun Lawrence Livermore National Laboratory, Livermore, California 94551 Received May 31, 1994 The stability of crude oil in the subsurface is of continuing interest as the search for new supplies pushes new f r ~ n t i e r s . l -Beyond ~ mere stability, some investigators are developing kinetic models that would predict oil composition as well as o c ~ u r r e n c e . ~Two -~ problems arise when deriving these compositional models: (1) it is difficult to unravel the reactions that simultaneously create and destroy a particular component in a complex mixture, and (2) the reactions of a given component by itself are not necessarily the same as in a complex matrix. A new isotopic method involving a doubleJ3C label detected by gas chromatographymass spectrometry (GC-MS) is presented here for unraveling these reactions. The method is demonstrated by measuring the intrinsic destruction kinetics of labeled n-hexadecane in ordinary n-hexadecane and in three crude oils of substantially different composition: a typical North Sea oil, a high-paraffin oil, and a high-sulfur oil. The unexpected conclusion is that the kinetics of hexadecane destruction are nearly equal in these three dissimilar oil matrices but only 60% as fast as in neat hexadecane. The use of isotopic labels in kinetic studies is common (e.g., Hoerin8) but investigations of skeletal reactions usually use a single labeled atom. A double-13Cisotopic label is much easier to follow in an oil matrix by conventional GC-MS. The natural abundance of 13Cis 0.011, so the normal probability of finding two 13C atoms in a small mass spectrometric fragment is very small: 0.000 12 for a ethyl radical (mlz = 31) and 0.000 36 for a propyl radical (mlz = 45). This low natural abundance makes it possible to follow both the disappearance and the appearance of doubly labeled species at very low dopant levels. A 1 wt % dopant of 1,2-13C2-nhexadecane is easily detected in a m l z = 31 chromatogram with a signal-to-noise ratio of several hundred. Even though the added 13C is only 10% of the total 13C in the oil, it represents 99% of all mass 31 ethyl groups in the oil. Three dissimilar oils were provided by industrial sponsors of this work: a North Sea oil (38.4" API gravity, 0.2 wt % S, 46% saturates), a high-paraffin oil (42.5" API gravity, 0.03 wt % S, 71% saturates), and a high-sulfur oil (25.4"API gravity, 3.3 wt % S, 28% (1) Saxby, J. D. Nature 1984,308, 177-179. (2) Quigley, T. M.; Mackenzie, A. S. Nature 1988, 333, 549-552. (3) Mango, F. D. Nature 1991, 352, 146-148. (4) Horsfield, B.; Schenk, H. J.; Mills, N.;Welte, D. H. Org. Geochem. 1992,19, 161-172. (5) Sweeney, J. J.; Burnham, A. IC; Braun, R. L. Bull. Am. Assoc. Petr. Geol. 1987, 71, 967-985. (6) Behar, F.; Ungerer, P.; Kressmann, S.; Rudkiewicz, J. L. Rev. Instrum. Fr. Petr. 1991, 46, 151-183. (7)Braun, R. L.; Burnham, A. K. Org. Geochem. 1992, 19, 191204. _. .. (8) Behar, F.; Kressmann, S.; Rudkiewicz, J. L.; Vandenbrouke, M. Org. Geochem. 1992,19, 173-189. (9) Hoering, T. C. Org. Geochem. 1984,5, 267-278.
saturates). The oils were doped with 1.4, 2.5, and 0.9 wt % of 1,2-13C2-n-hexadecane, respectively, which
roughly correspond t o the amount of natural n-hexadecane in each oil. The neat hexadecane was doped with 2.3 wt % of the labeled hexadecane. Approximately 0.055 wt % of 13C-naphthalenewas added to all samples as an internal (unreactive) standard before pyrolysis. Both isotopic materials were obtained from Isotech, Inc., Miamisburg, OH. Two microliters of oil solution and glass beads were added to a bent capillary tube,1° the tube sealed with a torch, and heated in a GC oven at 344 "C for 8-64 days. The oven temperature was stable to within 1 "C, and the absolute temperature is known to f l "C. The void volume of a filled tube is 20 pL. After pyrolysis, the tube was broken inside a Teflon tube and washed with methylene chloride. Naphthalene-d8 was added as a second internal standard during this wash step in order to check the reactivity of the 13C-naphthalene. All samples were analyzed at least twice by GC-MS (Finnegan TSQ700). The ratio of the 13Clnaphthalene-dspeak areas did not change conclusively with pyrolysis time within the observed 30% standard deviation, although the ratios in the 64 days samples did seem to be on the low side. Some alkylated 13C-naphthalenes were observed. Consequently, the ratios of any given compound to the 13C-naphthalene can be used as an absolute measure of destruction or creation, with the caveat that the 64-day ratios may be high by as much as 30%. Rate constants for the destruction of labeled nhexadecane were determined from corrected mlz 129 P3C-naphthalene) and 228 (1,2-13Cz-n-hexadecane)areas as shown in Figure 1. The decomposition rate is fastest in neat n-hexadecane and it is close to that observed earlier for neat n-hexadecanein autoclaves.11J2 Even though the differences among the oils are not clearly outside of experimental precision, the rate in the paraffinic oil is closest to that of neat hexadecane. Increased sulfur content in kerogen is thought to lower the temperature required for bitumen generation,13J4but it is not obvious whether the breakdown of hydrocarbons would be accelerated by a radical pool modified by sulfur species. H2S does catalyze hexadecane cracking in the gas phase.15 Our results indicate that oil matrices of varied composition all appear to stabilize the normal alkanes relative to the neat alkane. This is consistent with Baskin and Peters:14 the kinetic ~~~
(10) Horsfield, B.; Disko, U.; Leistner, F. Geol. Rund. 1989,78,361374. (11) Ford, T. J. Znd. Eng. Chem. Fundam. 1986,25, 240-243. (12) Jackson, IC J.; Burnham, A. K.; Braun, R. L; Knauss, K. G. Prepr. Pap.-Am. Chem. Soc., Diu. Fuel Chem. 1992, 37, (4), 16141620. (13) Hunt, J. M.; Lewan, M. D.; Hennet, R. J-C. Bull. Am. Assoc. Pet. Geol. 1991, 75, 795-807. (14) Baskin, D. IC; Peters, K. A. Bull. Am. Assoc. Pet. Geol. 1992, 76, 1-13. ~~~
0887-0624/95/2509-0190$09.00/0 0 1995 American Chemical Society
Communications
Energy &Fuels, Vol. 9, No. 1, 1995 191
1- /1
3 mk=29
North Sea oil
I North Sea oil, uncracked
0.019 day"
0.010day'
20
60
40
0
2
0
4
0
6
0
1 I mlzy 29
0.1
North Sea oil, 64 days at 344%
r
-2 0 4 0 6 0
0
0.01
0
time, days
20
40
60
time, days
Figure 1. Determination of rate constants for destruction of 1,2J3C-n-hexadecanein neat hexadecane and three crude oils of different composition. The reaction occurred at 344 "C in sealed glass capillary tubes. The standard deviations of the rate constants were about 0.001 day-'.
8
0.01
20
40 time, days
60
I
Figure 2. Amount of unlabeled hexadecane in the oil samples as a function of time. The creation of hexadecane from other compounds is at least as important as the destruction of hexadecane in the early stages of pyrolysis.
effects of sulfur are primarily in the early oil-generation phase. Instead, the dominant effect appears to be that other oil components provide a more labile source of hydrogen than n-hexadecane, thereby reducing the chain transfer reactions attacking hexadecane in the oil matrix relative to neat hexadecane. Unlike the labeled n-hexadecane, the amount of unlabeled n-hexadecane in each oil sample gives the net result of simultaneous hexadecane destruction and creation reactions. Figure 2 shows that creation reactions are important in all oils. The rates of creation and destruction are roughly equal for the first 16 days in the high-paraffin oil, after which the destruction reactions dominate. Creation reactions dominate the first 30 days for the North Sea and high-sulfur oils, after which the destruction reactions dominate. The relative creation rate is fastest in the high-sulfur oil and slowest (15) Rebick, C. In Frontiers of Free Radical Chemistry; Pryor, W. A., Ed.; Academic Press: New York, 1980; pp 117-137.
Retention time
Figure 3. GC-MSchromatograms of m l z 29 and 31 showing the conversion of labeledhexadecane.The region shown ranges from n-& t o n-Ca. Pristane and phytane are evident following n-C17 and n-Cls, respectively, in the uncracked oil. The m l z 31 chromatograms have been multiplied 16 times t o show minor peaks, and peaks are clearly evident for normal hydrocarbons less than C16 in the cracked oil. A minor deuterated naphthalene peak occurs just before n-Clz. in the high-paraffin oil, inversely related to the normal paraffin content of the oil. This implies that the creation of additional comes largely from nonparaffins, e.g., dealkylation of aromatics and polars. Isoprenoidfnormal hydrocarbon ratios in the oil decreased with time, as expected. For example, both the pristanelnaphthalene and pristaneln-heptadecane ratiosin the North Sea and high-paraffin oils decreased exponentially by about a factor of 10 during the 64 days. Assuming that creation reactions are minor for pristane, cracking of pristane is about 3 times faster than n-hexadecane. A complete chemical reaction network requires the nature of the products as well as the creation and destruction rate constants. Pyrolysis of neat n-hexadecane under high pressure forms hydrocarbons both larger and smaller than C ~ S . ~ The ' J ~ appearance of mlz 31 in the chromatogram of the cracked oils gives an indication of the reaction products of n-hexadecane pyrolysis. mlz 29 and 31 chromatograms of the CloCZOregion of the North Sea oil are shown in Figure 3. Most of the labeled ethyl groups are found in molecules smaller than CIS.
Acknowledgment. This work was performed under the auspices of the US.Department of Energy by the Lawrence Livermore National Laboratory under Contract No. W-7405-ENG-48 an cosupportedby DOE Basic Energy Sciences and industrial sponsors (ARCO, Exxon, Mobil, Conoco, Unocal, Norsk Hydro, Saga Petroleum, Statoil, IKU, AGIP, and Amoco). Eric Micheal of Conoco and Nigel Mills of Saga Petroleum provided the oil samples.
