有機物﹑礦物及水在深埋或隱沒沉積物中交互作用之實驗研究(2/3)

行政院國家科學委員會專題研究計畫 期中進度報告

有機物､礦物及水在深埋或隱沒沉積物中交互作用之實驗研

究(2/3)

計畫類別： 個別型計畫 計畫編號： NSC93-2116-M-002-003- 執行期間： 93 年 08 月 01 日至 94 年 07 月 31 日 執行單位： 國立臺灣大學地質科學系暨研究所 計畫主持人： 黃武良 計畫參與人員： 賴士禾、張英如、莫慧偵 報告類型： 精簡報告 報告附件： 出席國際會議研究心得報告及發表論文 處理方式： 本計畫可公開查詢

中 華 民 國 94 年 7 月 8 日

行政院國家科學委員會補助專題研究計畫期中報告

※※※※※※※※※※※※※※※※※※※※※※※※※※

※ ※

※

※

※ ※

有機物、礦物及水在深埋或隱沒沉積物中交互作用

之實驗研究

※※※※※※※※※※※※※※※※※※※※※※※※※※

計畫類別：ｖ

個別型計畫 □整合型計畫

計畫編號：NSC 93－2116－M－002－003

執行期間：

93 年 8 月 1 日至 94 年 7 月 31 日

計畫主持人：黃武良([email protected])

計畫參與人員： 賴士禾、張英如、莫慧偵

本成果報告包括以下應繳交之附件：

□赴國外出差或研習心得報告一份

□赴大陸地區出差或研習心得報告一份

ｖ

出席國際學術會議心得報告及會議發表講稿一份

□國際合作研究計畫國外研究報告書一份

執行單位：國立台灣大學地質科學系

中 華 民 國

94 年 7 月 8 日

行政院國家科學委員會專題研究計畫期中報告

有機物、礦物及水在深埋或隱沒沉積物中交互作用之實驗研究

計畫編號：NSC 93-2116-M-002-003

執行期限：93 年 8 月 1 日至 94 年 7 月 31 日

主持人：黃武良 國立台灣大學地質科學系

計畫參與人員：賴士禾、張英如、莫慧偵 國立台灣大學地質科學系

The geochemical evolution of carbonaceous matter as a function of

temperature, pressure, and other geological factors is essential to our

understanding of the fate of organic matter, through a series of geological

processes, in burial or subducted sediments. Previous studies focused mostly on

the effect of temperatures and pressures on organic reactions with little concern

on the influence of other geological factors. This three-year research project

covers a series of experimental programs, which will be conducted successively

to elucidate several important issues including the effect of mineral-water

environments, the effect of type of organic matter, and development of new

experimental technique.

Our research in the first year (2003-2004) focuses on the influence of

oxidation-reduction condition on the generation of methane and carbon dioxide

from organic matter in diagenetic environments. The results have been reported

in the 2004 Gordon Research Conference (Organic Geochemistry) and

documented in a paper submitted to a SCI journal “Organic geochemistry”

(currently in revision; see attachment I).

Our research in the second year (2004-2005) focuses on the visual behavior

of organic matter at high temperature and pressures, and the Fourier transform

infrared spectroscopy (FTIR) studies of organic matter. The results show that the

DAC-pyrolysis technique is useful for identifying the oil vs. gas prone organic

matter. The FTIR has been confirmed to be a promising tool for quantifying the

extent of organic reactions and determine the change of chemical functionality

of the organic matter during transformation, although we are still unable to

conduct the in situ measurement of IR spectroscopy at high temperature and

pressure. These techniques have been applied to a variety of organic matter

worldwide, including individual macerals separated from humic coals. The

results have been reported in 2004 Annual Meeting of Geological Society of

China, and an international conference (ISATT) in July 2005, and have been

documented in a manuscript (see attachment II) to be submitted to a SCI journal

“International Journal of Coal Geology”. In addition, we have investigated the

thermal stability of a carbonaceous matter, fullerene (C

60

), under geological

conditions. The occurrences of nature fullerene have been reported in sediments

from K-T and P-T boundary, in black shale and in organic rich rock of

greenschist facies. Its stability or kinetics may be used as a potential

geothermometer. In this research year, we have conducted preliminary

experiments and developed technique to quantify the transformation of C

60

to

amorphous carbon under hydrothermal conditions.

In the third year (2005-2006) of this research project, we shall focus on the

kinetics of the transformation of C

60

to amorphous carbon. In addition, we shall

conducted experimental study on the effect of sulfur radical on the kinetics of

organic transformation.

Papers derived from researches supported by NSC were published or to be

published in SCI journals during the research period (2003-2005) including:

1. Huang, W.L., 2003, The nucleation and growth of polycrystalline quartz:

Pressure effect from 0.05 to 3 GPa. European Journal Mineralogy 15, 843-853

(SCI). Impact factor: 1.335.

2. Lin, S.J., Huang, W.L., 2004. Polycrystalline calcite to aragonite

transformation kinetics: experiments in synthetic systems. Contributions to

Mineralogy and Petrology 147, 604-614. (SCI), impact factor: 2.433

3. Cheng, A.L., W.L., Huang, 2004. Selective adsorption of hydrocarbon gases

on clay and organic matter. Organic Geochemistry 35, 413-423. (SCI). impact

factor: 1.756.

4. Shen, P.Y., Hwang, S.L., Chu, H.T., Yui, T.F., Pan, C.N., Huang, W.L., 2005.

On the transformation pathways of α-PbO2-type TiO2 at the twin boundary of

rutile bicrystals and the origin of rutile bicrystals .European Journal of

Mineralogy. (SCI) (in press).

5. Su, K.W., Huang, W.L., 2005. Generation of hydrocarbon gases and CO

2

from a humic coal: Experimental study on the effect of water, minerals and

transition-metals. Organic Geochemistry. (in revision)

附件一

Manuscript Submitted to

Organic Geochemistry (in revision)

Generation of hydrocarbon gases and CO

2

from a humic coal:

Experimental study on the effect of water, minerals and transition-metals

Kwan-Hwa Sua and Wuu-Liang Huanga*

a

Department of Geosciences, National Taiwan University, Taipei, Taiwan

Abstract

The yields and composition of hydrocarbon gases and CO2 generated from a coal were

experimentally determined under different litho logical conditions simulated by adding a variety of minerals and transition metals, including hematite, siderite, pyrrhotite, gypsum, Fe, Ni and V in the coal sample using confined hydrous pyrolysis technique. The experiments were conducted at 340 and 360 °C for time durations ranging from 24 to 240 hours. The results reveal that the studied mineral-water environments have a small, but measurable influence on both the yields and composition of the generated gases. The difference in total hydrocarbon gas yields in different mineral environments at the same maturity can be as high as 15 mol. %. The hydrocarbon gas yields, in general, increase slightly with decreasing oxygen activity or increasing hydrogen fugacity of the experiments; this trend becomes more pronounced at higher maturity. Sulfide (pyrrhotite) tends to inhibit slightly the production of hydrocarbon gases while sulfate (gypsum) does not. Carbonate (siderite) or the generated CO2

exerts an adverse influence on the production of hydrocarbon gases. The effect of mineral environments on the variations of CO2 yields is more pronounced than that of hydrocarbon

gases at the studied conditions. CO2 production is highest in carbonate and lowest in metallic

iron or nickel environments.

*Corresponding author: Tel.: +886-2-33662929; fax: +886-2-83692853

1.INTRODUCTION

Accurate characterization of the oil generation potential of source rocks is essential to assess hydrocarbon accumulation in a petroleum system. The amounts and types of organic matter in the source rocks, in addition to temperature and time, determine the yields and composition of the generated hydrocarbons. The roles of these factors have been routinely evaluated in predicting hydrocarbon occurrence in petroleum systems (Tissot et al., 1987; Waples, 1994). However, failure of a prediction, for example the occurrence of hydrocarbon gas in nature at maturity much lower than conventional thought, is not uncommonly encountered, particularly in coal beds (Muscio et al., 1993; Rowe and Muhlenbachs, 1999; Ramaswamy, 2003). Other factors which may play important roles in hydrocarbon generation and its associated organic maturation include the catalytic effects of transition elements, organic-sulfur, minerals or aqueous reactions in the source rocks (Goldstein, 1983; Mango, 1992b; 1997; Helgeson et al. 1993; Lewan 1997; 1998; Seewald 1997; 2001; 2003; Butala et al., 2000; Behar et al. 2003). Mango (1992; 1996) experimentally demonstrated that the generation rate of hydrocarbon gases and their methane contents in the presence of metal-catalysts, are significantly higher than those by conventional pyrolysis technique but close to levels observed in nature. The effect of organic–inorganic interactions, in particular, the availability of hydrogen, on the generation of hydrocarbons in sedimentary basins has been reviewed comprehensively by Seewald (2003).

Among these factors, the role of minerals or aqueous reactions in the source rock in hydrocarbon generation and its associated organic maturation has been increasingly recognized (Helgeson et al. 1993; Lewan 1997; Berndt, et al., 1996; Seewald 1997; 2003; Behar et al. 2003). The high surface area of minerals or clays associated with the organic matter may catalyze the organic maturation (Johns, 1979; 1982); whereas the mineral-water reactions may significantly change the oxygen, sulfur and carbon dioxide activity, and indirectly affect the organic maturation within the sources rocks (Helgeson et al. 1993; Berndt et al., 1996; Seewald, 1997; 2003). Most previous studies on the catalytic experiments were conducted under dry conditions (Brook, 1952; Eisma and Jurg, 1964; Hosfield and Douglas, 1980; Goldstein, 1983; Huizinga et al., 1987; Mango, 1992; 1997), which rarely occurred within natural source rock environments. As the role of water in the generation of petroleum has been increasingly recognized (Hoering, 1984; Siskin and Katritzky, 1991; Lewan, 1997; Seewald 1997), the pyrolysis experiments in the presence of water are believed to more accurately simulate the oil/gas generation and expulsion from source rocks (Barth et al., 1994; Lewan 1997; Behar et al., 1991; 1997; Seewald et al., 1998; Ritter, et al., 1995;

Schimmelmann, et al., 1999). In addition, water is believed to affect the catalytic efficiency; for instance the efficiency of acidic-type catalysts (e.g. clays) can be significantly lowered in the presence of water (Goldstein, 1983; Tannenbaum et al., 1985). The effect is attributed to the difference in reaction mechanism of gas generation: a carbonium-ion mechanism in the absence of water in contrast to a free radical mechanism in the presence of water (Eisma and Jurg, 1969). The extents and mechanism of the catalytic effect of mineral matrices, particularly in the presence of water, however, are still not fully understood.

The present study considers potential natural catalysts which could significantly change the gas generation rate and its composition. We have experimentally measured the yields and composition of gases generated from a humic coal in the presence of water with a variety of minerals or metals, including hematite, siderite, pyrrhotite, gypsum, Fe, Ni and V. Our observations focus on the yields, rates of hydrocarbons and carbon dioxide, the gas compositions, and the isomer distributions.

2. Experimental procedures

2.1 Starting materials

The coal sample was collected by from the coal seam in Shan-Fu-Chi sandstone from the middle Miocene age located at Ming-Ted Dam profile outcropping on the west frank of the Chuhuangkeng anticline in northwestern Taiwan. Petrogtraphic study shows that the coal contains mainly desmocollinite (about 75 %) implying a terrestrial high plant origin. The VR (0.35 %) measured from telecolinite in the sample indicates that it is an immature coal. The geochemical parameters measured using Rock-Eval pyrolysis show that the total organic carbon (TOC), hydrogen index (HI), oxygen index (OI) and Tmax, respectively, are 58 %, 273 mg HC/g TOC, 22 mg CO2/g TOC, and 410 °C. The coal sample was ground in a

nitrogen atmosphere to an average size around 40 μm. The powdered samples without special treatments or kerogen isolation, were used throughout the experiments. The minerals and metals (thereafter named as mineral matrices) added to the system including hematite (Fe2O3),

siderite (FeCO3), pyrrhotite (Fe1-xS), gypsum (CaSO4.2H2O), sylvite (KCl), iorn (Fe), nickel

(Ni) and vanadium (V).

The experiments were conducted under three confined pyrolysis conditions: confined hydrous, confined non-hydrous and confined dry conditions. Confined pyrolysis was conducted in a closed system using flexure containers, which can be deformed to adjust the sample pressure to equal the externally applied pressure (Monthioux, et al., 1985; Behar, et al., 1991; Landais et al., 1994; Freund et al., 1994). In confined hydrous pyrolysis, water and the solid sample were loaded and sealed within a gold capsule, whereas in confined non-hydrous pyrolysis, no water was added to the sealed capsule (Behar et al., 2003). Although there was no added water in the confined non-hydrous pyrolysis, water may be generated during the early stage of coal pyrolysis in the system (Huang, 1996; Behar et al., 2003). Thus, in

confined dry pyrolysis, KCl salt was added along with the coal sample in the gold capsule in order to minimize the activity of water in the system (Newton, 2002) since the salt could dissolve into the water generated during pyrolysis.

All experiments with mineral matrices were conducted under confined hydrous pyrolysis with mass proportion of coal: mineral: water around 3: 1: 1. Powdered coal (about 10 mg) sample with water and mineral matrices water was loaded and electric arc sealed in gold capsules (0.5 mm OD x 15 mm length). For comparison, all coal samples without added mineral matrices were performed in the presence or absence of water. The sample capsules were bound together using platinum wire and loaded into a Parr pressure vessel partially filled with water. The vessel was heated isothermally at 340 and 360 °C for run durations of 24, 72, 240, and 720 hours. In order to prevent the capsules from bursting due to overpressure

generated during the gas generation within the capsule, an external argon pressure of 50 bars was applied for runs at 360 °C. The pressure within the gold capsules during the experiments was close to the saturated water vapor pressure for runs at 340 °C and 50 bars above the vapor pressure at 360 °C.

2.3. Analysis of gases and solid pyrolysates

After experiments each cleaned sample capsule was loaded into a glass vial with a gas-tight septum. The free space in the vial was accurately measured with a correction for the volume of solid gold. The capsule was punched with several holes using a stainless needle through the septum in order to release gases into the vial space from the capsule. The analysis was then performed by extracting 0.25 ml of the gas from the vial using a micro-syringe and

injecting it into the inlet of a Gas-chromatograph (GC). The yields of individual gases under different experimental conditions were quantified using a gas chromatograph with FID detector for hydrocarbon gases (HC) and TCD detector for CO2 in a HP6890

Gas-chromatograph. The capillary column had 0.53 I.D. and 15 m length (HP-Plot Al2O3).

Chromatographic grade (99.9995%) helium gas was used as carrier gas. The calibration curves for hydrocarbon gas components (C1 to C6) were made frequently to quantify the

absolute amount of each gas component. The details of analytical procedure were as reported in Cheng and Huang (2004).

The solid pyrolysates were removed from the sample capsule for VR, infrared spectroscopic, x-ray diffraction and scanning microscopic analyses. The VR was measured under an x20 objective lens with oil immersion using a petrographic microscope. The reported reflectance (%Ro) was averaged from about 100 measurements of each sample with very small standard deviation around + 0.015. The chemical functionality, particularly aliphatics and aromatics of the solid pyrolysates analyzed using Bomem model DA8.3 Fourier transformed infrared spectroscopy (FTIR). The samples were prepared in KBr pellets containing 0.5% of solid pyrolyzed residues.

3. EXPERIMENTAL

RESULTS

3.1. GENERAL REMARKS

Experimental results provide information about the amounts and the composition of hydrocarbon gases (HC gas), and carbon dioxide generated from the coal with different mineral matrices as a function of maturity (Tables 1 to 8). The results show that water activity, the redox conditions and the mineral matrix (sulfide, sulfate and carbonate) have some influence on the gas generation. The absolute and relative yields of individual hydrocarbon gas, methane (C1), ethane (C2), propane (C3), normal butane (n-C4), iso-butane (i-C4), pentane

(C5), and hexane (C6) and CO2 under different experimental conditions have been quantified

and compared. The analysis of liquid hydrocarbons (C7+) was not performed due to the

insufficient amounts of pyrolyzed residue samples. The thermal maturity of the experimental conditions was determined mainly by the VR measured from the solid residues (Fig 1), which are slightly higher than those calculated from the experimental temperature-time pairs using Easy%Ro (Sweeney and Burnham, 1990).

The experimental results generally show that the cumulative yields of total hydrocarbon gases increase consistently with increasing vitrinite maturity thermally matured at different temperature-time pairs. The generation rate is nearly constant in the early stage but slightly increases at vitrinite maturity above about 1.5 %Ro (Fig. 2). This trend is similar for most studied conditions containing a variety of mineral matrices (Fig. 2). Similarly, methane or each individual wet gas yield increases progressively with increasing maturity and their relative yields are in inverse proportion to their molecular weights (Fig. 3). The proportion of each individual gas varies only slightly in experiments containing the different mineral matrices. The cumulative yields of methane relative to wet gases in terms of C1/(C1+C2+C3)

ratios (Fig. 4) or C1/(C1 to C6) percentages (Fig. 5), decrease with increasing maturity up to

Ro = 1.42 %, then increase at higher maturity for runs without mineral matrices and with KCl, Fe, and FeS (Fig. 4A). It is noteworthy that the reverse of the trends for experiments containing gypsum, hematite, siderite, nickel and vanadium occurs at vitrinite maturity about 0.1 %Ro higher than other runs (Fig. 4B). In general, a similar trend has also been observed in most runs of this study, although the variation of the ratios is slightly different with different minerals. The i-C4/n-C4 ratios in experiments with coal and water without mineral

matrix show a progressive decrease ranging from 1.1 to 0.62 with increasing experimental maturity from 1.15 to 1.9 %Ro (Fig. 6); the change of this ratio appears to be less pronounced at maturity higher than 1.6 %Ro. Similar trends but with wide variation of the ratios have been found in experiments with water and different mineral matrices (Fig. 7).

The effect of water or water activity has been investigated here in three different pyrolysis methods using the coal sample without a mineral matrix. The cumulative yields of total HC gas from the coal at confined dry hydrous condition are only slightly higher than that at confined non-hydrous and confined hydrous conditions (Fig. 8A). It appears that very low water activity in the system tends to slightly enhance the gas generation. The difference in HC gas yields generated from confined hydrous and confined non-hydrous experiments is very small, and may be negligible if experimental uncertainty is considered (Fig. 8A). The cumulative yields of methane relative to wet HC gases as a function of maturity vary only very slightly with water activity. The i-C4/n-C4 ratios in the gas generated in the maturity

range of this study, however, are consistently lower in confined non-hydrous experiments than other two methods although the differences are small (Fig. 6).

The experiments show that the variation of hydrocarbon gas yields in the presence of different studied mineral matrix and water, although small, is experimentally significant (Fig.

9). The influence of mineral matrices appears more noticeable at higher maturity; an effect that can be as high as 15 % of total HC gas yield at maturity close to Ro = 1.9 %. The coal sample in experiments with vanadium at 360 °C for 720 hours generated about 15 % more HC gas than in experiments with siderite. The mineral effects on the enhancement of the total HC gas generation increase generally in the order of Fe2O3 < FeS < FeCO3 < Ni < added water <

CaSO4 < no added water < Fe < dry (KCl) < V (Fig. 9).

The results from two series of experiments containing sulfur show that the one with gypsum yields total HC gas similar to that in pure water but the one with pyrrhotite yields slightly less total HC gas than that with sulfate over the range of maturities studied (Figs. 10A and 10B). The effect of carbonates on gas generation was investigated by a series of experiments with siderite and water. The results when compared with other iron-bearing mineral matrices show that the HC gas yields in the presence of siderite are higher than hematite, but lower than in the presence of iron or magnetite.

3.3. Carbon dioxide generation over a wide range of maturity

The results from runs without minerals show that most CO2 was generated at maturity

lower than Ro = 1.15 %. The cumulative yields increase slightly at the maturity ranging from 1.15 to 1.65 %Ro, then increase at a higher rate at higher maturity (Fig. 11). Similar to the HC gas yields, the difference in CO2 yields between confined hydrous and non hydrous pyrolysis

is negligible whereas the CO2 yield is slightly lower in confined dry pyrolysis. The addition

of minerals, pyrrhotite, sulfate, or hematite in the system shows no significant influence on the CO2 generation (Fig. 12). In contrast, the CO2 yield is about 50 % and 30 %, respectively,

lower in the presence of iron and nickel and about 10 to 50 % higher in the presence of

siderite. The CO2 yields, in general, increase in order from Fe < Ni < dry (KCl) < V < gypsum

< pyrrhotite < no added water < added water< hematite < siderite (Fig 12). This order implies that with a higher oxidation state in the system, there is higher CO2 yield. Note that the

experiments in the presence of siderite produced less CO2 at low maturity but significantly

more CO2 at high maturity compared to samples in pure water. The CO2/(CO2 + HC gas) mol

ratioin total gas in confined hydrous pyrolysis decreases progressively with increasing maturity from higher than total HC gas yields at maturity lower than 1.6 %Ro to lower at higher maturity. Similar trends were found for experiments under different conditions (Fig. 13). The CO2/ (CO2 + HC gas) mol ratios vary considerably with mineral matrices (Fig. 14).

The effect of mineral matrices on CO2 generation, in general, is more pronounced than on

total HC gas (Fig. 15A and 15B).

3.4. Characterization of solid pyrolysates

The measured reflectances of vitrinite-like materials in the solid pyrolysates in confined hydrous pyrolysis experiments with no added mineral matrix have been used to indicate the maturity of artificial maturation of the studied coal (Fig. 1). The reflectances were compared with those selectively measured in experiments containing gypsum, hematite and vanadium (Fig. 16). The results generally show that the addition of these mineral matirces tends to suppress very slightly, even insignificantly, the vitrinite reflectance (VR) relative to that with water only. At maturity lower than about 1.4%Ro, vanadium suppresses the reflectance most of among these mineral matrices whereas at higher maturity the effects of these mineral matrices vary significantly but with no systematic trend. A dramatic suppression of VR was observed at maturity from 1.9 to 1.72%Ro in experiments with hematite. In addition, a significant suppression of VR in the system containing gypsum has also been observed at high maturity (Fig. 16).

The measured VRs in the solid pyrolysates from confined hydrous pyrolysis and confined non-hydrous pyrolysis experiments with no added mineral matrix were correlated with the results measured by FTIR. The results show that the peak intensity of absorption band ratios of 1500 - 1600 cm-1 and 2900 - 3000 cm-1, which approximately represent the relative amount of aromatics to aliphatic functional groups, respectively, increases with increasing maturity (Fig. 17). It appears that the change of the measured functional groups with %Ro is more pronounced in samples with water than those with no water added (Fig. 17). The variation of the absorption band ratio in experiments with different mineral matrices is small except for pyrrhotite and siderite (Fig. 18). It is noteworthy that there is an inverse correlation between the absorption band ratios (aromatics / aliphatics ratio) and the total hydrocarbon yields in the presence of all the mineral matrices except siderite (Fig. 18).

4. Discussion

The yields and composition of laboratory-generated hydrocarbon gas are comparable to those previously reported. Figure 19 compares the methane yield artificially pyrolyzed from a variety of coals as a function of maturity. The methane yield from our studied coal is similar to those derived from Illinois coal # 6 (Vorres, 1990) and Morwell coal (Behar et al., 1995) but slightly lower than that from Mahakam coal (Behar et al., 1995). Calvert lignite generated much lower methane than other coals (Behar et al., 2003). The comparison reveals that the methane yields depend highly on the type and composition of coals.

The effects of mineral matrices revealed in the present study, is small relative to those resulted from the variation of coal types. The relatively small mineral effect on the organic reaction is inconsistent with the results from previously experiments conducted in the dry system (Goldstein, 1983; Mango, 1996). Mango (1996) experimentally demonstrated that the HC gas can be generated at very low thermal stress in the presence of source rocks containing transition metals. Our results, however, suggest that the large catalytic enhancement on the cracking of organic matter containing transition metals (Fe, Ni and V) previously reported has not been observed in our experiments with water during confined pyrolysis. Although the retardation mechanism of water on the catalytic effect is still not clear, these results are consistent with prior studies which predicted that water may significantly reduce the catalytic efficiency of an acid-type mineral surface (Eisma and Jurg, 1969). The loss of efficiency of mineral catalysts in the experiments may also be attributed to the wet ability of the mineral surfaces, which minimizes the contact between organic matter and mineral matrix. The prediction of hydrocarbon gas yields in a petroleum system, therefore, should consider more the coal type than the lithology or mineral matrices associated with the coal.

The mole percentages of methane in the HC gas ranged from 62% to 68% observed in this study, which are consistent with those artificially matured at similar maturities using Mahakam coal in the confined closed system pyrolysis (Michels et al., 2002). These are lower than the methane proportion in HC gas accumulated in natural gas reservoirs (Mango, 1992; 2000; Michels et al., 2002), but higher than those found in source rock cuttings (Snowdon, 2001). The artificial maturation in semi-open system conducted by Michels et al. (2002), however, shows higher proportions of methane (90 to 97%) which are similar to those in reservoir.

The high initial methane proportion and the decrease of the proportion with maturity (Fig. 4) for both the studied coal and the Mahakam coal imply that methane (C1) is the dominant

only in middle stages. The reverse trend in the late stage, which shows an increase of the methane proportion at higher maturity (>1.42 %Ro), is probably attributed to the secondary cracking of wet gases (C2 to C4) or liquid hydrocarbons (C5+) to form methane, although this

does not exclude the likelihood that the reverse of the trend is simply due to the primary cracking of the solid coal. Artificial maturation of Mahakam coal show similar reverse trends at 1.25%Ro and 1.1%Ro, respectively in both closed and semi-open system (Michels et al., 2002). The reverse trends for experiments containing gypsum, hematite, siderite, nickel and vanadium, which occurs at about 1.52 %Ro (Fig. 4B), implies that these minerals or metals may retard the secondary cracking relative to those without mineral matrices or with iron or pyrrhotite. The increase of the C1/ (C1 to C6) ratio and the absolute amount of C1 at high

maturity may be attributed more to the secondary cracking of liquid hydrocarbons (C5+) or

primary cracking of solid residue and less to the wet gases (C2 to C4) since the thermal

stability of wet gases is higher than the long-chain hydrocarbons or the methyl-bearing aromatics groups in the liquid hydrocarbons (C5+) (Laidler et al., 1962; McNeil and BeMent,

1996). This interpretation is consistent with the semi-open confined pyrolysis data of Michels et al. (2002) who suggested that the high methane proportion at high maturity may not be attributed to the secondary cracking of wet gases since they were stepwise released out of the system during pyrolysis.

The interpretation of the secondary cracking of liquid hydrocarbons has been supported from data on butane isomers. The initial high i-C4/n-C4 ratios (about 1.2) in the HC gas

generated at low maturity imply that coal releases i-C4 at rates faster than n-C4. The

progressive decrease of the ratio from 1.2 to 0.6 as experimental maturity increases to about 1.9 %Ro (Fig. 7) indicates either the decrease of i-C4 release rate or increase of n-C4

generation rate with maturity. He’roux et al. (1979) reported that primary cracking of solid organic matter tends to release more iso-butane isomer than normal-butane (with the isomer ratio >0.9) and vice versa (with the ratio < 0.8) during the secondary cracking of oils. The observed decrease of the ratio with maturity in our experiments may partially contribute to the progressive cracking of early formed oil. The decline of the ratio across the boundary (0.8 to 0.9) was observed at maturity around 1.45 %Ro, suggesting significant secondary cracking of liquid hydrocarbons may occur around this maturity. This maturity is consistent with that (1.42%Ro) observed based on the C1/(C1+C2+C3) ratio (Fig. 4A). This general trend is similar

for most conditions in a variety of mineral matrices, although the isomer ratios show significant variation under different conditions.

4.2. Role of water

The HC gas yield in a system with very low water activity (confined dry condition) is about 5 % higher than those in the presence of water (confined non-hydrous and hydrous conditions). This effect, although small, is significant because similar effects were consistently observed in experiments with different maturity. The small retardation by water on the light hydrocarbon generation from coal suggests that most hydrocarbon gases were decomposed from coal molecules rather than secondarily cracked from long-chain hydrocarbons (Lewan, 1997; Seewald, 2003). The water effect, however, is less than that found by Behar et al. (1997; 2003), who showed 13.9 % higher HC gas yield from a lignite in the absence of water (open nonhydrous condition) than in the presence of water (closed nonhydrous condition). Our results, which show no significant difference in gas yields between confined nonhydrous and hydrous conditions are also not consistent with that of Behar et al. (2003), who showed that experiments at closed non-hydrous condition generates more HC gas than at hydrous pyrolysis conditions. This discrepancy may be attributed to different experimental setups. Our confined pyrolysis was performed using flexure containers, which can be deformed to adjust the sample pressure to equal the externally applied pressure whereas Behar et al. (2003) conducted closed pyrolysis in rigid autoclaves. Free water is believed to be present also in the experimental systems under both confined non-hydrous and closed non-hydrous conditions because of the generation of water from organic samples in the early-stage of pyrolysis (Huang, 1996; Behar et al., 2003). The similarity in HC yields generated in our experiments in both confined non-hydrous and hydrous pyrolysis is not unexpected since water pressure is a similar result from the same external pressure. In contrast, the difference in HC gas yields in Behar’s experiments was attributed to different water vapor pressures in closed non-hydrous and closed hydrous pyrolysis (Behar et al., 2003). These results imply that water pressure plays a significant role in the HC gas generation (Hill et al., 1994).

The water activity also influences the dryness of the generative gas. The significantly lower proportion of methane in the gas generated at very low water activity than in the presence of water has been observed at maturity lower than 1.6%Ro (Fig. 4A). It appears that water may enhance the release of methyl groups from the coal. However, the effect is reversed at higher maturity, where the secondary cracking of the wet gas prevails, implying

that water may retard the cracking of wet gases to methane. The i-C4/n-C4 ratios are

consistently higher in runs with excess water, implying that water may enhance preferentially the release of branch alkanes relative to straight chain alkanes from the coal.

4.3. Effect of reduction/oxidation (redox) condition

Our results show that the effect of water on hydrocarbon gas generation is small without added mineral matrices. Water may be involved in the organic reactions if it proceeds hydrolytic disproportional reactions by dissociating into hydrogen and oxygen via mineral-water reactions (Seewald 1997). The present study demonstrates from four series of experiments with different oxygen fugacity in systems where oxygen fugacities in the system which are buffered by minerals at the experimental conditions were calculated (Bethke 1996) or estimated from previously determined values (Ulmer & Barnes 1987). The oxygen fugacity buffered in the system decreases in the order of V-V2O3, Fe-Fe3O4, Ni-NiO and Fe3O4-Fe2O3

under the experimental conditions. The results show an increase of total HC gas or methane yields when the oxygen fugacity is decreased or the hydrogen fugacity is increased (Fig. 20A), and the trend is more pronounced at higher maturity (Fig. 20B). The present experiments confirm that the redox conditions indeed influence the gas generation, although the effect is small. The redox potential or hydrogen activity in an aqueous system, therefore, appears to influence HC gas yields in such as a way that excess HC gases may be produced at reducing conditions by reacting excess hydrogen generated from water with organic carbon in the system. However, the present study show that the enhancement of gas yield in the presence of transition metals, Fe, Ni, and V was not as pronounced as previous thought (Mango, 1996; Butala et al., 2000). The discrepancy is probably attributed to the suppression of catalytic effect by water.

4.4 Effect of sulfur

It appears that the presence of sulfur as calcium sulfate in the system has little effect on the HC gas generation (Fig. 10), in contrast to the conventional thought that organic sulfur or sulfur radicals may accelerate organic maturation at lower thermal stress (Baskin and Peters, 1992; Lewan, 1997). These experiments demonstrate that gypsum minerals dehydrate mostly into anhydrite and that there are negligible sulfur products (e.g. H2S), implying that the

thermal reduction of calcium sulfate by organic matter does not take place in the experiments at high maturity (1.9%Ro) or at high temperature (360 °C). This result contradicts the common observation that the thermochemical sulfate reduction (TSR) can proceed in evaporate-bearing formations at similar or lower maturity. It appears that the high thermal stability and low aqueous solubility of calcium sulfate in the laboratory system generates insufficient amounts of sulfur radicals required for initiation of the catalytic breaking of C-C bonds (Lewan, 1998). The enhancement of TSR reactions in nature, on the other hand, is probably attributed to other catalytic effects, for instance, the high initial H2S pressure

(Machel, 2001). The slight or negligible effects of pyrrhotite on the generation of HC gas yield is surprising because the presence of FeS in the system tends to create conditions more reducing and richer in sulfur radicals. Both of these two conditions should favor HC formation.

4.5. CO

2

generation

That the CO2/CH4 ratiodecreases progressively with increasing maturity is mainly

attributed to the continuous generation of hydrocarbon gases as the rate of CO2 generation

tends to decline. This observation is consistent with the previous thought that the generation of CO2 from type III organic sources takes place at an earlier stage of organic maturation than

methane during laboratory pyrolysis and natural catagenesis (Hunt, 1996). The CO2 yields

from runs (with KCl) at very low water activity are slightly lower than that in the presence of water, suggesting that water contributes oxygen to the formation of carbon dioxide.

These results suggest that the mineral-water reactions have been involved in CO2

generation in these experiments, either by releasing the CO2 from siderite or by consuming

oxygen within organic matter. The later process may compete with the formation of oxides from Fe or Ni. The results show that significant decomposition of siderite into CO2 and iron

oxides (magnetite and hematite) has taken place under the experimental conditions. The decomposition of siderite in the experiments was confirmed by the high CO2 yield in a high

maturity experiment (%Ro = 1.90) containing about double amounts of siderite. By contrast, the presence of reducing agents such as Fe, V and pyrrhotite seems to retard the formation of CO2. These results suggest that the mineral-water reactions have been involved in CO2

4.6. Vitrinite reflectance and infrared absorption of solid pyrolysates

Thermal maturity for each set of experimental conditions was approximately represented by the reflectance (%Ro) measured from the vitrinite-like materials in solid run products conducted under confined hydrous pyrolysis. The measured VRs are slightly higher than those calculated from the corresponding temperature and time of experiments using the Easy%Ro program (Sweeney and Burnham (1990) (Fig. 1). This difference is not unexpected since the calculated VR was calibrated against natural samples at temperatures much lower than laboratory temperatures. For application of the experimental data to the prediction of the gas yields and composition in nature, we recommend using the calculated VR because the measured vitrinite was matured at high temperature. The kinetic extrapolation to natural conditions can not be done due to the lack of sufficient data in the study.

Based on limited types of the studied mineral matrices, the very slight suppression of VR in the run products in experiments containing gypsum, hematite and vanadium suggests that the inorganic environments in shales relative to coals may not be an important factor affecting the vitrinite maturation. However, the observed suppression of VR in the presence of gypsum and hematite at high maturity may imply that a high oxic environment tends to suppress the maturation of vitrinite. If this is the case, the slight suppression of VR in the presence of reducing environment, vanadium, at maturity below 1.43%Ro is hard to explain. Further studies are required to determine the effects of mineral matrices on the variation of vitrinite reflectance under similar thermal stress.

The correlation of the measured VR with the aromatics/aliphatics ratio measured by FTIR is not unexpected since the increase of VR is attributed to the aromatization of the organic matter (Rouxhet et al., 1978). Similar observations have been previously reported (Rouxhet et al., 1978; Landais, 1995 Riesser et al., 1984). Landais (1995) found a nice correlation for coals. The good correlation is consistent with most previous studies on the aromatization of organic matter during maturation (Riesser et al., 1984; Benkhedda et al., 1992; Machnikowska et al., 2001). However, micro-FTIR data on vitrinite maceral show significant scattering from the trend (Lin and Ritz, 1993). The poor correlation has also been observed in type-III organic matter (Ganz and Kalkreuth, 1987). The results from our

experiments demonstrates that the maturation of a single source of starting vitrinite may generate a more consistent trend. The observed scattering of the data from the trend or the

poor correlation of the trend is probably attributed to the composition variation of starting vitrinite particles in the sample.

The inverse correlation between the amount of HC gas generated and the

aromatics/aliphatics group ratio (Fig. 18) may be coincidental but that dose not exclude a possible unknown mechanism which relates to the catalysis of minerals. It appears that the generation of HC gases may retard the aromatization of organic matter. Since more total hydrocarbon gas yields were found in reducing environments, the retardation of aromatization may be due to the higher reducing environment.

5. Conclusions

This study experimentally demonstrated that the mineral-water reactions can influence the yields and composition of hydrocarbon gases and CO2 generated from a humic coal. The

proposed mineral-water reactions which may influence the gas generation include the oxidation-reduction between metal and oxide, sulfate-sulfide conversion, and carbonate decomposition. The extent of influence, although small, depends on the mineral type or mineral-water reactions. The decrease of water or oxygen activity tends to enhance slightly the total hydrocarbon gas yield while the presence of carbonate (siderite), carbon dioxide, and sulfide (pyrrhotite) tends to inhibit slightly the production of hydrocarbon gases. To date, the results based on limited data from some representative minerals and transition metals show a maximum variation of about 15 % for hydrocarbon gas yields during aqueous reactions at maturity equivalent to vitrinite reflectance of about 1.95%. The effect of transition metals is not as pronounced as previously reported at dry condition. The effect of mineral environments on the variations of CO2 yields is more pronounced than that of hydrocarbon gases at the

studied conditions. CO2 production is highest in carbonate and lowest in metallic iron or

nickel environments. This study confirms the previous thought that mineral-water reactions may play a role in natural gas generation in a petroleum system but this study shows that the influence of mineral matrices is small relative to that imposed by the compositional variation of organic matter.

Acknowledgements

This research was supported by the Earth Sciences Sections, National Sciences Council of ROC, NSC Grant 92-2116-M002-018 to W. L. Huang. The authors would like to

acknowledge Mr. Jun-Chin Shen of EDRI, Chinese Petroleum Corporation for supplying the starting coal samples and performing the Rock-Eval pyrolysis and vitirinite reflectance measurement. Thanks are extended to Drs. C. L. Kuo J. N. Weng and T. Y. Yang for reviewing on early version of the manuscript.

References

Aranovich, L.Ya. Newton, R.C., 1996. H2O activity in concentrated NaCl solutions at high

pressures and temperatures measured by the brucite-periclase equilibrium. Contributions to Mineralogy and Petrology 125, 200-212.

Barth, T., Schmidt, B.J., Nielsen, S.B., 1996. Do kinetic parameters from open pyrolysis describe petroleum generation by simulated maturation? Bulletin of Canadian Petroleum Geology 44, 446-457.

Baskin, D.K., Peters, K.E., 1992. Early generation characteristics of a sulfur-rich Monterey kerogen. Amer. Assoc. Petrol. Geol. Bull. 76, 1-13.

Behar, F., Kressmann, S., Rudkiewicz, J.L., 1991. Experimental simulation in a confined system and kinetic modeling of kerogen and oil cracking. Organic Geochemistry 19, 173-189.

Behar, F., Lewan, M.D., Lorant, F., Vandenbroucke, M., 2003. Comparison of artificial maturation of lignite in hydrous and nonhydrous conditions. Organic Geochemistry 34, 575-600.

Behar, F., Vandenbroucke, M., Teermann, S.C., Hatcher, P.G., Leblond, C., Lerat, O., 1995. Experimental simulation of gas generation from coals and marine kerogen. Chemical Geology 126, 247-260.

Behar, F., Vandenbroucke, M., Tang, Y., Marquis, F., 1997. Thermal cracking of kerogen in open and closed systems: determination of kinetic parameters and stoichiometric coefficients for oil and gas generation. Organic Geochemistry 26, 321-339.

Benkledda, Z., Landais, P., Kister, J., Dereppe, J.M., Monthioux, M., 1992. Spectroscopic analysis of aromatic hydrocarbons extracted from naturally and artificially matured coals. Energy & Fuels 6, 166-172.

Berndt, M.E., Allen, D.E., Seyfried, W.E., Jr., 1996. Reduction of CO2 during

serpentinization of olivine at 300 degrees C and 500 bar. Geology 24, 351-354.

Bethke, C.M., 1996. Geochemical Reaction Modeling, Concepts and Application. Oxford University Press, New York.

Brooks, B.T., 1952. Evidence of catalytic action in petroleum formation. Indust. Eng. Chem. 44, 2570-2577.

Butala, S.J., Medina, J.C., Taylor, T.Q., Andrus, D.B. Bartholomew, C.H.,Lee,M. L., 2000. Mechanism and kinetics of reactions leading to natural gas formation during coal

maturation. Energy Fuels 14, 235-259.

Cheng, A.L., Huang, W.L., 2004. Selective adsorption of hydrocarbon gases on clays and organic matter. Organic Geochemistry 35, 413-423.

Eisma, E., Jurg, J.W., 1969. Fundamental aspects of the generation petroleum. In: Eglinton, G., Murphy, M.T.J.(eds.), Organic geochemistry; methods and results. Springer-Verlag, New York, pp. 676-698.

Ganz, H., Kalkreuth, W., 1987. Application of infrared spectroscopy to the classification of kerogen-types and the evaluation of source rock and oil shale potentials. Fuel 66, 708-711. Goldstein, T.P., 1983. Geocatalytic reactions in formation and maturation of petroleum.

Amer. Assoc. Petrol. Geol. Bull. 67, 152-159.

Helgeson, H.C., Knox, A.M., Owens, C.E., Shock, E.L., 1993. Petroleum, oil field waters, and authigenic mineral assemblages: Are they in metastable equilibrium in hydrocarbon reservoirs? Geochimica et Cosmochimica Acta 57, 3295-3339.

Heroux, Y., Chagnon, A., Bertrand, R., 1979. Compilation and correlation of major thermal maturation indicators. AAPG Bulletin 63, 2128-2144.

Hill, R.J., Jenden, P.D., Tang, Y.C., Teerman, S.C., Kaplan, I.R., 1994. Vitrinite Reflectance as a Maturity Parameter: Applications and Limitations. In: Mukhopadhyay, P.K., Dow, W.G. (Eds.), American Chemical Society Symp. Ser. 570, pp. 161-193.

Hoering, T.C., 1984. Thermal reactions of kerogen with water, heavy water and pure organic substances. Organic Geochemistry 5, 267-278.

Horsfield, B., Douglas, A.G., 1980. The influence of minerals on the pyrolysis of kerogens. Geochimica et Cosmochimica Acta 44, 1119-1132.

Huang, W.L., 1996. Experimental study of vitrinite maturation: effects of temperature, pressure, time, water and hydrogen index. Organic Geochemistry 24, 233-241.

Huizinga, B.J., Tannenbaum, E., Kaplan, I.R., 1987. The role of minerals in the thermal alteration of organic matter — III. Generation of bitumen in laboratory experiments. Organic Geochemistry 11, 591-604.

Hunt, J.M., 1996. Petroleum Geochemistry and Geology, 2nd Edtion. Freeman, New York.. Johns, W.D., 1979. Clay mineral catalysis and petroleum generation. Ann. Rev. Earth Planet.

Sci. 7, 183-198.

Johns, W.D., 1982. Clay mineral catalysis and petroleum generation . AAPG bulletin 66, 1445.

Jurg, J.W., Eisma, E., 1964. Petroleum hydrocarbons; generation from fatty acid. Science 144, 1451.

Laidler, K.J., Sagert, N.H., Wojciechowske, B.W., 1962. Kinetics and mechanisms of the thermal decomposition of propane. Proc. R. Soc. Lond. 270, 242-253.

Landais, P., 1995. Statistical determination of geochemical parameters of coal and kerogen macerals from transmission micro-infrared spectroscopy data. Organic Geochemistry 23, 711-720.

Landais, P.L., Michels, R., Elie, M., 1994. Are time and temperature the only constraints to the simulation of organic matter maturation? Organic Geochemistry 22, 617-630.

Lewan, M.D., 1997. Experiments on the role of water in petroleum formation. Geochimica et Cosmochimica Acta 61, 3691-3723.

Lewan, M.D., 1998. Sulphur-radical control on petroleum formation rates. Nature 391, 164-166.

Lin, R., Ritz, G.P., 1993. Studying individual macerals using i.r. microspectroscopy, and implications on oil versus gas/condensate proneness and "low-rank" generation. Organic Geochemistry 20, 695-706.

Machel, H.G., 2001. Perspectives on investigations of diagenesis for hydrocarbon exploration and exploitation in deep basins. Schriftenreihe der Deutschen Geologischen Gesellschaft 13, 4-7.

Machnikowska, H., Krzton, A., Machnikoski, J., 2002. The characterization of coal macerals by diffuse reflectance infrared spectroscopy. Fuel 81, 245-252.

Mango, F.D., 1992. Methane concentrations in natural gas; the genetic implications. Organic Geochemistry 32, 1283-1287.

Mango, F.D., 1992. Transition metal catalysis in the generation of petroleum and natural gas. Geochimica et Cosmochimica Acta 56, 553-555.

Mango, F.D., 1996. Transition metal catalysis in the generation of natural gas. Organic Geochemistry 24, 977-984.

Mango, F.D., 1997. The catalytic decomposition of petroleum into natural gas. Geochimica et Cosmochimica Acta 61, 5347-5350.

Mango, F.D., 1997. The light hydrocarbons in petroleum; a critical review. Organic Geochemistry 26, 417-440.

Mango, F.D., 2000. The origin of light hydrocarbons. Geochimica et Cosmochimica Acta 64, 1265-1277.

McNeil, R.I., BeMent, W.O., 1996. Thermal stabilities of hydrocarbons: laboratory criteria and field examples. Energy Fuels 10, 60-67.

Monthioux, M., Landais, P., Monin, J-C, 1985. Comparison between natural and artificial maturation series of humic coals from the Mahakam delta, Indonesia. Organic Geochem. 8, 275-292.

Muscio, G.P.A. Horfield, Welte, D.H., 1994. Occurrence of thermogenic gas in the immature zone – implications from the Bakken in-source reservoir system. Organic Geochemistry 22, 461-476.

Price, L.C., Schoell, M., 1995. Constraints on the origins of hydrocarbon gas from compositions of gases at their site of origin. Nature 378, 368-371.

Ramaswamy, G., 2003. A field evidence for mineral-catalyzed formation of gas during coal maturation. Oil and Gas Journal 100-38, 32-36.

Raymond, M., et al., 2002. Understanding of reservoir gas compositions in a natural case using stepwise semi-open artificial maturation. Marine and Petroleum Geology 19, 589-599.

Riesser, B., Starsinic, M., Squires, E., Davis, A., Painter, P.C., 1984. Determination of aromatic and aliphatic CH group in coal by FTIR. II. Studies of coals and vitrinite concentrates. Fuel 63, 1253-1261.

Ritter, U., Myhr, M.B., Vinge, T., Aareskjold, K., 1995. Experimental heating and kinetic models of source rocks: Comparison of different methods. Organic Geochemistry 23, 1-9. Rouxhet, P.G., Robin, P.L., 1978. Infrared study of the evolution of kerogens of different

origins during catagenesis and pyrolysis. Fuel 57, 533-540.

Rowe, D., Muhlenbachs, A., 1999. Low temperature thermal generation of hydrocarbon gases in shallow shales. Nature 398, 61-63.

Schimmelmann, A., Lewan, M.D., Wintsch, R.P., 1999. D/H isotope ratios of kerogen

bitumen, oil, and water in hydrous pyrolysis of source rocks containing kerogen types I, II, IIS, and III. Geochimica et Cosmochimica Acta 63, 3751-3766.

Seewald, J.S., 1997. Mineral redox buffers and the stability of organic compounds under hydrothermal conditions. Mat. Res. Soc. Symp. Proc. 432, 317-331.

Seewald, J.S., 2001. Aqueous geochemistry of low molecular weight hydrocarbons at elevated temperatures and pressures: constraints from mineral buffered laboratory experiments. Geochimica et Cosmochimica Acta 65, 1641-1644.

Seewald, J.S., 2003. Organic-inorganic interactions in petroleum-producing sedimentary basins. Nature 426, 327-333.

Seewald, J.S., Benitez-Nelson, B.C., Whelan, J.K., 1998. Laboratory and theoretical constraints on the generation and composition of natural gas. Geochimica et Cosmochimica Acta 62, 1599-1617.

Siskin, M., Katritzky, A.R., 1991. Reactivity of organic compounds in hot water: Geochemical and technological implications. Science 254, 231-237.

Snowdon, L.R., 2001. Natural gas composition in a geological environment and the implications for the processes of generation and preservation. Organic Geochemistry 32, 913-931.

Sweeney, J.J., Burnham, A.K., 1990. Evaluation of a simple model of vitrinite reflectance based on chemical kinetics. Am. Assoc. Petroleum Geol. Bull. 74, 1559-1570.

Tannenbaum, E., Kaplan, I.R., 1985. Low-M hydrocarbons generated during hydrous and dry pyrolysis of kerogen. Nature 317, 708-709.

Tissot, B., Durand, B., Epitalie, J., Combaz, A., 1987. Thermal history of sedimentary basins, maturation indices, and kinetics of oil and gas generation. Am. Assoc. Petroleum Geologists Bulletin 58, 499-506.

Ulmer, G.C., Barnes, H.L., 1987. Hydrothermal Experimental Techniques. John Wiley & Sons, New York.

Vorres, K.S., 1990. The eight coals in the Argonne Premium coal sample program. Energy & Fuels 4, 420-426.

Waples, D. W., 1994. Modeling of the petroleum system-from source to trap. In: Magoon, L.B., et al. (Eds.), The petroleum system; from source to trap. American Association of Petroleum Geologists, Tulsa, pp. 307-322.

Figure Captions

Figure 1. The measured vitrinite reflectance (VR) correlated with those calculated using Easy%Ro (Sweeney and Burnham, 1990). The corresponding temperature- time pair of experiments for each measured VR is: 1.13% (340oC-24hr), 1.26 % (340 oC-72 hr), 1.34% (360 oC-24 hr), 1.43% (340 oC-240 hr), 1.52% (360o C-72 hr), 1.62% (360 oC-240 hr), 1.90% (360 oC-720 hr), where hr = hours. The measured VRs were derived from experiments with water but no mineral matrix.

Figure 2. Experimental data showing the cumulative yields of total hydrocarbon gas generated from the studied coal as a function of measured VR. Experiments were carried out at 340 oC (solid triangles) and 360 oC (solid circles) and with water and a variety of mineral matrices: A) KCl salt, B) coal only without added water and mineral matrix, C) added water but no mineral matrix, D) gypsum, E) pyrrhotite, F) vanadium, G) iron, H) nickel, I) hematite, J) siderite.

Figure 3. Methane or individual wet hydrocarbon gas yield increases progressively with

increasing maturity. Their relative yields are in inverse proportion to their molecular weights.

Figure 4. Experimental data showing the methane/(methane + ethane + propane) ratio i.e. C1/(C1+C2+C3) ratio, as a function of maturity, in terms of measured VR (%Ro): (A) runs

without mineral matrices and runs with FeS and Fe, (B) runs with CaSO4, Ni, Fe2O3, FeCO3,

and V.

Figure 5. Experimental data showing the percentage of individual gas in total HC gases (C1-C6) as a function of maturity (measured VR).

Figure 6. Experimental data showing the ratio of n-butane/ ( i-butane + n-butane), i.e.

nC4/(iC4 + nC4), as a function of maturity (measured VR). Symbols: Triangles = coal + added

water; squares = coal with no water added; diamonds = coal + KCl (very low water activity).

Figure 7. i-butane/n-butane ratios in experiments with water and different mineral matrices.

Figure 8. Experimental data showing the effects of water and maturity on (A) the HC gas yield, and (B) carbon dioxide yields. Symbols for maturity (measured %Ro) : solid rhombs

=1.13%; solid triangles = 1.26 % ; solid circles = 1.34% ; open triangles = 1.42%; open squares = 1.52% ; open circles = 1.62%; solid squares = 1.90%. The corresponding temperature and time pair of experiments for each VR see the legend of Fig. 1.

Figure 9. Experimental data compared among the total hydrocarbon gas yields pyrolyzed under a variety of conditions arranged in the order of increasing hydrocarbon yields. Maturity symbols see Fig. 8.

Figure 10. Experimental data showing the effect of sulfide, sulfates, carbonates on (A) the total hydrocarbon yields. Maturity symbols see Fig. 8, and (B) methane yields as a function of maturity.

Figure 11. Cumulative carbon dioxide yield as a function of measured VR under different conditions.

Figure 12. Cumulative carbon dioxide yield under different conditions at different maturity. Maturity symbols see Fig. 8.

Figure 13. Percentages of CO2 in the total gas generated from the coal as a function of

measured VR.

Figure 14. Experimental data showing the CO2/(CO2 + total HC gas) mol ratio under a variety

of conditions arranged in the order of increasing hydrocarbon yields. Maturity symbols see Fig. 8.

Figure 15. Experimental data showing the effect of siderite, sulfate or iron or magnetite on (A) the CO2 yields. Maturity symbols see Fig. 8, and (B) CO2/(CH4) ratio of the coal as a function

of maturity.

Figure 16 . Experimental data showing the effect of oxidation/reduction potential (hematite vs. vanadium) and sulfates (gypsum) on the VR measured for the solid products. Maturity

Figure 17. Experimental data showing the change of functionality (aromatic/alkane) ratio as a function of maturity. The intensity ratio of adsorption bands, 1500-1600 cm-1 and 2900-3000 cm-1 measured by FTIR. The ratio represents approximately the ratio of amounts of aromatic to saturates, respectively, in the solid residues. Symbols: triangle = with water; rhomb = no added water.

Figure 18. Experimental data showing the effect of mineral matrices on the

aromatics/saturates ratio in solid residue at 360 oC and 240 hours (measured %Ro = 1.62). The total hydrocarbon yields (triangles) at same conditions were presented for comparison.

Figure 19. Comparison of methane yields artificially generated from a variety of coals. Symbols: open circles = Illinois coal # 6 (Vorres, 1990); solid rhombs = Morwell coal (Behar et al., 1995); open squares = Mahakam coal (Behar et al., 1995); open triangles = Calvert lignite (Behar et al., 2003); solid triangles = Sahnfuchi coal (this study).

Figure 20. Experimental data showing the effect of oxidation/reduction potential under pyrolysis on the total hydrocarbon yields. Maturity symbols see Fig. 8.

附件二 Manuscript to be submitted to

International Journal of Coal Geology

Generation and expulsion of petroleum from coal macerals visualized

in-situ during semi-open and closed system pyrolysis

F. J. Mo

a

, Wuu-Liang Huang

a,

*, Jack Machnikowski

b

a

Department of Geosciences, National Taiwan University, Taipei, Taiwan, ROC

b

Institute of Chemistry and Technology of Petroleum and Coal, Wroclaw

University of Technology, Gdanska 7/9, 50-344 Wroclaw, Poland

Abstract:

The oil and gas generative potential of lipinites, vitrinites and fusinites separated from coals of different maturity have been characterized using the diamond anvil cell pyrolysis technique in semi-open and closed systems. The visualization of samples in-situ during pyrolysis shows that lipinite generated large amounts of visible oil-like liquid along with some gas while fusinites generated no visible liquid and gas. The liquid (oil) from lipinite appears light in color and less viscous. In contrast, some non-perhydrous vitrinite

unexpectedly generated large amounts of visible liquid although the liquid is darker and thicker than that from lipinite, indicating that the oil potential of vitrinites is not limited to its perhydrous nature. The peak temperatures (Tpeak ) of maximum rate of liquid generation for

vitrinite (465 to 540 °C) are generally higher than those for lipinites (453 to 475 °C); both of which are within the range of Rock-Eval Tmax for most organic matter. The visual phenomena

and Tpeak of liquid generation in the closed system are essentially similar to those in a

semi-open system, whereas the gas generation appears more noticeable in a closed system. Since semi-open experiments are much easier to perform, the DAC pyrolysis may be

1. Introduction

Evaluation of oil generation potential of source rocks in a petroleum system is essential to the success of exploration. There are increasing evidences showing that some terrestrial coals, in addition to lacustrine and marine source rocks, may be an important source for oil (e.g. Durand and Paratte, 1983; Hunt, 1991; Law and Rice, 1993; Scott and Fleet, 1994; Wilkins and George, 2002). The oil potential of a coal depends mainly on its maceral constituents (Hunt, 1996). Lipinite (exinite) is the major component that contributes oil for many oil-prone coals (e.g. Indonesia and Malaysia coals, Land and Jones, 1987; Stankiewicz et al., 1996) whereas the perhydrous vitrinite contributes oil for vitrinite –rich but lipinite-poor coals (e.g. New Zealand, Newman et al., 1997; San Juan Basin, New Mexico, Clayton et al., 1991; Gipsland coals, Smith and Cook, 1984; Moore et al., 1992). However, it is still uncertain whether the perhydrous nature of vitrinite is the only source to account for the oil potential of vitrinite-rich coals. Other submaceral components of vitrinite may contribute the oil

generation potential.

While the distinction between oil-prone and gas-prone coals become a common practice in the source rock evaluation (Wilkins and George, 2002), the uncertainty about the expulsion or primary migration of oil from coals has caused an additional problem that hinders the accurate evaluation of petroleum potential. Most of techniques including pyrolysis-Gas chromatography (PY-GC; Bertrand, 1989; Powell et al., 1991), elemental analysis (Peper and Corvi, 1995), Rock-Eval pyrolysis (Baskin, 1997), maceral discrimination (Huang et al., 1997) , and spectroscopy (Miknis et al., 1981; 1996; Qin et al., 1991; Machnikowska et al., 2002), which used routinely in the measurements of the oil generation potential of source rocks or coals, are incapable of investigating the expulsion efficiency of oil from coals. The diamond anvil cell (DAC) pyrolysis (Huang, 1996; Weng et al., 2003) and hydrous pyrolysis technique (Lewan, 1997) among a few methods enable to study concurrently both the

generation and expulsion of oil from sources rocks. The visual capability of the DAC

technique provides the direct observations which enables to show how oil and gas generated and expulsed from a source rocks sample during pyrolysis. In this paper, we present the visual characteristics during pyrolysis of three major types of maceral using separated concentrates from 8 Polish coals. The use of high purity maceral concentrates permits the comparison of the oil and gas generation and expulsion of the individual macerals while minimizing the interference by the presence of other macerals.

2. Experimental methods

2.1 Starting materials

isolated from 8 coals of different ranks, from subbituminous to anthracite, containing 77.0 to 92.5% °C. The anthracite (sample C41) comes from Lower Silesian Basin, whereas the others come from Upper Silesian Basin. According to Polish classification, the coals represent the types with increasing rank in the order of flame coal (C31.2), gas coal (C33), gas-coking (C34), orthocoking (C35.2), semicoking (C37.1 and C37.2), semianthracite (C41), and anthracite (C42). The coals were separated into three major fractions namely vitrinite, lipinite and fusinite. Maceral separation was performed first by isolation of lithotypes by handpicking, and enriched by float-sink method. Lithotypes was then separated in centrifuge and examined by microscopic analysis for separation efficiency. The final separation products were dried in a vacuum oven at 60 °C. The detailed maceral separation and characteristics of the studied samples including proximate, ultimate analyses of coals and maceral structure revealed by diffuse reflectance infrared spectroscopy are reported in Machnikowska et al. (2002). The proximate, ultimate and petrographic, analyses of the studied macerals conducted by

Machnikowska et al. (2002) were presented in Tables 1 and 2. The atomic ratios for hydrogen carbon and oxygen were plotted on a van Krevan diagram (Fig. 1).

2.2 Diamond anvil cell pyrolysis technique

The diamond anvil cell (DAC) pyrolysis techniques (Huang, 1996; Huang and Otten, 1998; Weng et al., 2003) (Fig. 2) was used to simulate maturation of a variety of coal macerals and record the whole reaction process in-situ. The whole process of oil and gas generation and expulsion from macerlas can be visually observed. The qualitative and

semi-quantitative information has been obtained from the dynamic images. The variation and the yields of oil and gas can be quantified by measuring the volume in terms of the spreading area of the generated liquid phase during the pyrolysis (Weng et al., 2003).

The diamond anvil cell pyrolysis technique and sample configurations are similar to that described in Huang (1996) and Weng et al. (2003). A more detailed sketch of the cell

configuration was presented in Bassett et al. (1993) and Huang et al. (1994). A standard petrographic microscope with some modifications has been used for real time visualizing the sample. Real time video recording of experiments is accomplished with a high-resolution CCD (charge coupled device) color video camera and time lapse VCR (Fig 2).

Temperatures were measured using chromel-alumel thermocouples with their junctions closely contacting the pavilion faces of the diamond anvils; they are considered to be accurate to + 3°C. To prevent oxidation of the heating wires and diamonds at high temperatures, argon gas with 1 % hydrogen was circulated around the heating elements. The heating rates (10 ° C/min. up to 300°C, held for 5 min., then at 25 °C/min up to 550 °C), unless specified, are similar to the standard Rock-Eval pyrolysis (set at 300 °C for 5 min., then at 25 °C/min.).

The experiments were conducted under semi-open and closed system conditions. For the semi-open system, the sample with a size about 0.5 mm in diameter was loaded between diamond faces and pressed by the weight of upper diamond set. The anvil pressure ranges from ambient pressure at the sample margin to about 200 MPa at the center (Weng et al., 2003). The sample pressure may slightly increase with increasing the pyrolysis temperature due to the generation of liquid and gases; the pyrolysis experiment can be considered as semi-open system pyrolysis. It is likely that the newly generated volatile liquid or gas may escape from the system while the non-volatile liquid retains, and cracks to gas at higher temperatures, in the system during the pyrolysis. For closed-system experiments, the sample chamber of the DAC consists of a 500 μm diameter hole in a disk-shaped rhenium gasket (125 μm thick and 3 mm OD) sandwiched between two diamond anvils (Fig. 3). Three to four batches of each sample was loaded within the gasket chamber. The sample pressure may slightly increase with increasing the pyrolysis temperature due to the generation of liquid and gases.

2.3 Criteria for Interpretation of visual characters

The observations and characterization of organic samples during DAC pyrolysis focused on the followings:

(1) The presence or absence of oil and/or gas. Oil was considered to be present if the sample show softening, corner rounding and spreading of liquid lobes during pyrolysis. The presence of gas was recognized mostly as bubbles within the liquid phase or if the solid sample disintegrated or shrunk during the pyrolysis. However, we used “invisible” instead of absence if none of the above criteria was observed because its absence can not be certain.

(2) Oil yield in semi-open system during pyrolysis was estimated by the ratio of the maximum sample area (A) to the initial sample area (Ao); the initial sample area was

measured at 300 °C. We assigned four yield ranks from 1 to 4 corresponding to the ratios of A/Ao of less than 2, between 2 and 3, between 3 and 4, and larger than 4, respectively. For the

closed system, the estimation of yield is somewhat difficult. The oil yield was assigned as rank 1 if sample shows only corner rounding without spreading of liquid; it is assigned as rank 2 if liquid spreads around the sample but does not reach the chamber boundary, and as