Extraction of oil from oil shale by new, more environmentally acceptable methods

Prior work on oil shales has concentrated on retorting one or a small group of oil shales. For this study, oil shales from different localities and of different types were sourced so that conclusions would of general significance. The aim was then to investigate high temperature, high pressure reactions and the products of the oil shales under a wide range of conditions, with a view to finding mild conditions that would give good conversions with minimum environmental effects. The reactivity was to be correlated with the chemical structure of the oil shales if possible. The organic part of some of the oil shales, the kerogen, was isolated by treatment with NaOH-HCl, safer than the usual HCl-HF method, and the structure and reactions of the kerogens were compared with those of the original oil shales. The structure and reactions of the oil shales were also compared with those of an algal coal, torbanite, an oil shale with low inorganic content. The oil shales and their kerogens were analysed by methods including ash yield, elemental analysis, solubility in organic solvents at room temperature, FTIR, X-ray diffraction (XRD), energy dispersive X-ray fluorescence (EDX), inductively coupled plasma mass spectroscopy (ICP-MS), solid state 13C NMR, theromgravimetric analysis (TGA), X-ray photoelectron spectroscopy (XPS) and pyrolysis gas chromatography mass spectrometry (GC-MS). The products of oil shale and kerogen reactions were investigated by elemental analysis, 1H and 13C liquid NMR, GC-MS, multidimensional and two-dimensional gas chromatography coupled with flame ionization detectors (FID), flame photometric detectors (FPD) and mass spectrometry. Reactions were carried out in batch autoclaves (27 ml) under a range of temperatures (355-425oC) for various times (1, 5 h) and under various atmospheres (H2 or N2, 3 MPa cold). In some reactions potential catalysts were added, including Fe, Cu, Ni, Co, Mo, Ni/Mo and Co/Mo. The El-Lajjun (Jordan) oil shale was almost completely converted to oil (hexane soluble) and asphaltene (CH2Cl2-soluble, hexane insoluble) at 355oC after 1 h under H2 or N2 with small CO2 production. In contrast the torbanite required 1 h at 425oC or 5 h at 390oC to give high oil yield. The difference in reactivity may be associated with the presence of a large number of relatively weak C-S bonds in the El-Lajjun oil shale. The marine oil shales (Jordanian and Australian) all showed high atomic H/C ratio, S and calcite contents and low carbon aromaticity. A large range of trace elements was present. The lacustrine oil shale (Colorado) also had a high atomic H/C ratio, low carbon aromaticity, but a much lower S content and relatively high N content. The carbonate was ankerite rather than calcite. In all cases conditions could be found such that most of the organic matter could be converted to oil and asphaltene with little production of CO2. The conversions were not in general correlated with elemental analysis of the oil shales, the amount of volatile materials produced by TGA, the oil yield determined by Fischer Assay or the amount of bitumen obtained by solvent extraction at room temperature. The oil shales fell into distinct classes based on the reactivity, but these were not related to locality. Julia Creek, Ma’an and Colorado oil shales tended to have lower reactivity than El-Lajjun, Sultani, Attrat and Yarmouk. The effect of substituting H2 for N2, and the oil to asphaltene ratio varied considerably between the oil shales but generally at high temperature less asphaltene was produced. In all cases Mo based catalysts were the most effective but the pattern of catalyst activity varied considerably between the oil shales. In all cases the CH2Cl2-soluble products were of high atomic H/C ratio, and contained little aromatic hydrogen. The 1H NMR parameters, like the pattern of catalyst activity, clearly distinguished the Ma’an, Yarmouk and Julia Creek products from the others. The GC-MS indicated a clear distinction between the marine oil shales, with less prominent aliphatic hydrocarbon peaks and high abundance of S-compounds and the lacustrine Colorado oil shale and torbanite, with more prominent aliphatic hydrocarbon peaks, no S-compounds and some N-compounds present. In comparing the reactivity of different oil shales it is difficult to distinguish the effect of the mineral components on the organic matter. Therefore kerogens were isolated from five oil shales by the NaOH-HCl method, a rarely used method which has safety advantages over the usual HCl-HF method. Elemental analysis, ATR-FTIR and 13C NMR indicated that, provided the treatment was not repeated too often, the isolated kerogen was similar in structure and composition to the organic part of the original oil shale. El-Lajjun kerogen tended to be less reactive than the original oil shale, whereas the others tended to be more reactive. This is possibly because primary reaction products from the El-Lajjun kerogen tended to repolymerise more easily than those from the other oil shales. Elemental analysis of the products indicated that part of the S and N in some of the original oil shales was fixed by the mineral matter. 1H NMR indicated clear differences in some cases between the structure of the CH2Cl2-solubles from the oil shales and those from the kerogens, possibly due to catalytic effects of the mineral matter. The GC-MS of the CH2Cl2-solubles were similar for the oil shales and kerogens. Batch autoclave reactions and solvent workup are probably not a realistic model for commercial oil shale production. Therefore ‘flow-through’ experiments were carried out in which N2 at atmospheric pressure was passed over the oil shale in an autoclave heated to 425oC and the volatile product swept out and condensed. There was no general correlation between the amount or percentage of condensate and the amount of volatiles obtained by TGA, the room temperature CH2Cl2 extract yield, the oil fraction of the extract, or the Fischer Assay oil yield. The liquid 13C NMR and 1H NMR indicated that the aromatic carbon in the condensate was heavily substituted. There was no correlation between the aromaticity of the original oil shale and that of the condensate. The GC-MS of the condensate gave a more complicated chromatogram than that of the total CH2Cl2-solubles from batch autoclave reactions. A detailed multidimensional and multidetector GC analysis was carried out for a selected CH2Cl2-soluble product obtained from El-Lajjun oil shale under N2. Multidimensional GC revealed a large number of S-compounds including thiophenes, benzothiophenes and small amounts of dibenzothiophenes and benzonaphthothiophenes. In addition, a range of aliphatic alkanes and cycloalkanes, ethers, polar single ring aromatic compounds and small amounts of polycyclic aromatics were also identified. Some of these compound classes were not uniquely observable by conventional 1D GC. The total number of distinct compounds was very large (ca. > 1000). The product classes found did not necessarily give reliable information concerning functional groups in the original oil shale/kerogen, so that a detailed solid state 13C NMR and XPS study of the oil shales and kerogens was undertaken. The 13C NMR indicated no significant difference between oil shale and kerogen except for Julia Creek, where the carbon aromaticity of the kerogen was significantly reduced possibly due to oxidation. The lacustrine (type I) Colorado oil shale was less aromatic and showed longer aliphatic chain length than the marine oil shales (type II) and Rock-Eval analysis indicated it was more mature. All the oil shales showed a large average cluster ring size for the admittedly small fraction of aromatic carbon. The elemental analysis deduced from the XPS indicated differences, particularly for the oil shales, between the surface and the bulk concentration of O, S and N. The sulfur species in all the kerogens included aliphatic, aromatic and sulphate species, indicating that transformations occurred during high temperature reactions to form, almost exclusively, thiophenic compounds. Nitrogen species detected included pyridinic, pyrrolic, and a limited amount of quaternary nitrogen. The work reported indicates that the simple correlations often used to define product yields on the basis of oil shale characteristics are inadequate when a wide range of oil shales is considered. The distinction between the marine and the lacustrine oil shales is chemically important, but marine oil shales from different locations can resemble each other more closely than a group of marine oil shales of the same age and location. By suitable choice of gas, temperature and time, much higher yields of useful products, as defined by CH2Cl2 solubility, can be obtained from any oil shale than by conventional Fischer Assay with minimal production of CO2.