Experimental investigation of main controls to methane adsorption in clay-rich rocks

Applied Geochemistry 27 (2012) 2533–2545

Contents lists available at SciVerse ScienceDirect

Applied Geochemistry

journal homepage: www.elsevier .com/ locate/apgeochem

Experimental investigation of main controls to methane adsorptionin clay-rich rocks

Liming Ji a,b, Tongwei Zhang b,⇑, Kitty L. Milliken b, Junli Qu a, Xiaolong Zhang c

a Key Laboratory of Petroleum Resources Research, Institute of Geology and Geophysics, Chinese Academy of Sciences, Lanzhou 730000, Chinab Bureau of Economic Geology, Jackson School of Geosciences, The University of Texas at Austin, Austin, TX 78713, USAc The School of Earth Sciences, Lanzhou University, Lanzhou 730000, China

a r t i c l e i n f o

Article history:Received 14 May 2012Accepted 28 August 2012Available online 8 September 2012Editorial handling by R. Fuge

0883-2927/$ - see front matter � 2012 Elsevier Ltd. Ahttp://dx.doi.org/10.1016/j.apgeochem.2012.08.027

⇑ Corresponding author. Tel.: +1 512 232 1496; faxE-mail address: [email protected] (T

a b s t r a c t

In this study a series of CH4 adsorption experiments on clay-rich rocks were conducted at 35 �C, 50 �C and65 �C and at CH4 pressure up to 15 MPa under dry conditions. The clay-dominated rock samples used arefresh samples from quarries and mines. Samples are individually dominated by montmorillonite, kaolin-ite, illite, chlorite, and interstratified illite/smectite. The experimental results show that clay mineral typegreatly affects CH4 sorption capacity under the experimental conditions. In terms of relative CH4 sorptioncapacity: montmorillonite� illite/smectite mixed layer > kaolinite > chlorite > illite. Physisorption is thedominant process for CH4 absorption on clay minerals, as a result, there is a linear correlation betweenCH4 sorption capacity and BET surface area in these clay-mineral dominated rocks. The abundance ofmicro-mesopores in the size range of a few to a few 10 s of nanometers in montmorillonite clay andillite–smectite interstratified clay results in large BET surface area values for these mineral species.

A good linear relationship between the natural logarithm of Langmuir constant and the reciprocal oftemperature exists for clay-mineral dominated rocks, which provides a way to quantify the impact of claymineral type on gas adsorption capacity. Thermodynamic parameters, the heat of CH4 adsorption and thestandard entropy, are calculated based on this linear correlations. The heat of adsorption (q) and the stan-dard entropy (Dso) range from 9.4 to 16.6 kJ/mol and from �64.8 to �79.5 J/mol/K, respectively, valuesconsiderably smaller than those for CH4 adsorption on kerogens. Thus, it is expected that CH4 moleculesmay preferentially occupy surface sites on organic matter, in addition, the clay minerals are easilyblocked by water. As a consequence, organic-rich mudrocks possess a larger CH4 sorption capacity thanclay-dominated rocks lacking organic matter.

� 2012 Elsevier Ltd. All rights reserved.

1. Introduction

Organic-rich mudrocks have been the focus of renewed re-search during the past few years because of their emergence ashydrocarbon reservoirs (Montgomery et al., 2005; Jarvie et al.,2007; Loucks and Ruppel, 2007; Rowe et al., 2008; Ruppel and Lou-cks, 2008; Loucks et al., 2009). Shale gas is a typical unconventionalgas resource in which the mudrock is both the source of, and thereservoir for CH4, which is derived from the organic matter withinthe mudrock through biogenic and/or thermogenic processes (Hillet al., 2007; Strapoc et al., 2010). Methane may be stored in thenatural pores of the argillaceous source rock as free-gas and ad-sorbed onto the surfaces of kerogen and clay minerals, or dissolvedin bitumen (Curtis, 2002; Ross and Bustin, 2007). Among thesestorage sites, adsorbed gas accounts for between 20% and 85% ofthe total gas in five US shale formations (Curtis, 2002; Montgomeryet al., 2005). The adsorbed gas could be the major component of

ll rights reserved.

: +1 512 471 0140.. Zhang).

shale gas, and is generally up to 50% of the total gas in the Devo-nian shales in the Appalachian basin (Lu et al., 1995). Free gas isan important component for determining Gas-In-Place (GIP) inshale gas reservoirs, and adsorbed gas in some degree determinesthe longevity of shale gas producing well(s). In general, a high freeto adsorbed gas ratio corresponds to high GIP.

Organic matter and clay minerals are two major components forgas adsorption. The effect of organic matter content, kerogen type,and thermal maturation on gas adsorption has been extensivelyinvestigated (Zhang et al., 2012). In general, CH4 sorption capacityof organic-rich shales is positively correlated to total organic C con-tent; high vitrinite or inertinite shows stronger CH4 sorptioncapacity (Jarvie et al., 2007; Ross and Bustin, 2007; Chalmers andBustin, 2007a,b, 2008a,b). Previous studies have shown that theclay mineral composition and its micropore structure also affectthe gas sorption capacity of organic-rich shales (Aringhieri, 2004;Wang et al., 2004). Pores of 1–2 nm radius between crystal layersof clay minerals (mainly kaolinite, illite and smectite) provide theadsorption sites for CH4 and other gases due to the large surfacearea (Cheng and Huang, 2004). Ion exchange within the

2534 L. Ji et al. / Applied Geochemistry 27 (2012) 2533–2545

expandable structure of some clay minerals, for instance montmo-rillonite, may increase adsorption capacity (Aringhieri, 2004).

Ross and Bustin (2009) found that CH4 sorption capacity in-creases with increasing organic C content and micropore volumein the Devonian–Mississippian shale in Western Canada. They alsoobserved that and the mercury porosimetry porosity of clay-richshale is larger than silica-rich shales due to open porosity associ-ated with the aluminosilicate fraction. Furthermore, they deducedthat the gas sorption capacity of clay minerals mainly depends onthe types of clay and the micropores between irregular surfaces ofclay platelets (Ross and Bustin, 2009). Their work could not exclu-sively address the effect of clay minerals on gas adsorption becauseof the co-existence of organic matter.

The difference in the type and structure of clay minerals couldplay an important role in CH4 adsorption in clay-rich rocks, how-ever, the issue has not been addressed thoroughly in previous stud-ies. Clay mineral crystals can be constructed either by the doublelayer structure with an anhydrous Si–O tetrahedral layer (T) andwater–Al octahedral layer (O;TO,1:1), or by the three-layer struc-ture with two tetrahedral layers and an octahedral layer(TOT;2:1). Montmorillonite is a typical 2:1 type clay mineral, andthe unit cell consists of a water–Al layer and two Si–O layers, thewater–Al layer is sandwiched between Si–O layers, and each layeris not tightly bound together (Bergaya et al., 2006). Montmorillon-ite with interlayer water will freely expand and contract along thec axis during hydration and dehydration (Tsipursky and Drits,1984). Calcium and Mg can enter the interlayer structure of mont-morillonite and water can also enter and discharge. Montmorillon-ite has a large internal surface area, and shows a strong cation-exchange capacity. In contrast, kaolinite has the typical 1:1 typeclay structure, and the unit cell consists of a water–Al layer anda Si–O layer. Weak H-bonds are formed between the remainingH charges to balance O on the bottom of the upper water–Al layerand the O atoms exposed on the top of the lower Si–oxide layer,and connecting the upper and lower layers to form a solid latticestructure. Kaolinite has no active layer and inner surface area (Pal-omino and Santamarina, 2005), and it does not shrink or swell. Il-lite refers to the ordinary fine-grained muscovite, and has a three-tier structure similar to montmorillonite and no or very lowexpandability. In illite, Si atoms in the tetrahedral layer are re-placed by Al causing a negative charge imbalance that is compen-sated by K+ between the TOT layers. Weathering of crystal edgeswill release K ions. Illite may have a limited internal surface areaand thus a moderate adsorption capacity. Chlorite has a similarcrystal layer structure to illite, but the K ion layer is replaced byan additional octahedral layer containing Mg or another cation.Therefore, the sorption capacity and the affinity of the CH4 molec-ular for clay minerals will vary with different types of clay miner-als. Recent studies have also shown that diagenesis has erased thecontrol of primary texture on bulk rock properties in the BarnettShale, and primary detrital composition exerts its control on bulkrock properties only through its effect on the diagenetic pathway(Milliken et al., 2012).

In this study, organic-matter-free clay-mineral dominated rocksare used to investigate the specific effect of various clay mineralson CH4 adsorption. The goal of this study is to (1) understand theimpact of clay mineral composition on CH4 sorption capacity; (2)quantify the relationship between CH4 sorption capacity andclay-mineral dominated rock properties, i.e. BET specific surfacearea; and (3) determine the thermodynamic parameters of CH4

adsorption on different types of clay minerals. A series of CH4

adsorption experiments on rocks dominated by montmorillonite,kaolinite, illite, chlorite, and mixed-layer illite/smectite were con-ducted at 35 �C, 50 �C and 65 �C and CH4 pressure of up to 15 MPaunder dry conditions. BET surface area and SEM image analyseswere also conducted. The experimental results provide insight into

the effect of clay minerals on gas adsorption and provide an ap-proach for estimating the adsorbed gas attributed to clay mineralsunder reservoir pressures and temperatures.

2. Samples and experimental methods

2.1. Samples

The clay-dominated rock samples used in this study are all freshsamples from quarries and mines. Samples are individually domi-nated by montmorillonite, kaolinite, illite, chlorite, and interstrat-ified illite/smectite (Table 1, Fig. 1). All samples were crushed tofour particle sizes, 20–50 mesh (270–830 lm), 50–100 mesh(150–270 lm), 100–270 mesh (53–150 lm), and larger than270 mesh (<53 lm). Jurassic-age montmorillonite clay was ob-tained from a bentonite mine in Santai County, Sichuan Provincein China. There, a pink, gray, and gray–green block bentonite layerwith laminated and banded structure is exposed between finesandstone layers. Samples PRT-1 and PRT-5 contain 65.4% and78.7% montmorillonite, 14% and 26% quartz, 3.1% and 4% calcite,and 1.4% and 2.3% K-feldspar and plagioclase, respectively. Kaolin-ite clay was collected from a kaolinite quarry in Fugu County, Sha-anxi Province in China, developed in Carboniferous–Permian coal-bearing strata. GLT-4 contains 95% kaolinite, and 5% quartz. Illiteclay was collected from Ouhai, Zhejiang Province, from UpperJurassic pyroclastic rocks in which illite has developed by hydro-thermal-alteration. YLS-3 contains 99% illite and 1% pyrite. Chloriteclay was collected from a jade mine in Laizhou County, ShandongProvince in China, where hydrothermal chlorite layers are inter-spersed within a variety of lithologies of poorly-known age. LNS-3 contains 92.3% chlorite. A sample of interstratified illite/smectite(I/S ratio of 70:30) obtained from the Clay Minerals Society (ISCz-1)is from a sedimentary stratum in Slovakia, and contains 44.5% I/S,50.5% quartz. For comparison purposes, samples of siltstone andquartzite (one each) are used in the study. The siltstone (FS) iscomposed of 81.6% quartz, 11.3% plagioclase, 1.9–3.3% illite andchlorite. Quartzite (QT) is pure quartz.

The basic petrology of the four studied samples (excluding thestandard clay material) was assessed using a conventional petro-graphic light microscope. Polished thin sections (25 lm thickness)were prepared using a low-viscosity surface impregnation (bySpectrum Petrographics). The kaolinite sample was visibly layeredand displays other fabrics (grain ghosts) related to its sedimentaryorigin. The composition of this sample is nearly 100% kaolinitehowever and thus, the replacement of the original sediment isnearly complete, as is typical in many fireclays. Despite its likelyorigin as a replacement of volcanic ash, the montmorillonite dis-plays no textural evidence of shard ghosts. The metamorphic sam-ples (illite and chlorite) have a greater density and hardness thanthe sedimentary samples and contain minor non-clay minerals inthe form of porphyroblasts. In summary, it is important to notethat whereas all of the samples are materials dominated by partic-ular clay minerals, some of the samples are not sedimentary andthe clay crystals are not all clay-size (Fig. 2). Thus, this study seeksto examine the impact of clay mineralogy on gas absorption some-what apart from the effect of crystal size.

2.2. Experimental methods

2.2.1. CH4 gas sorption isotherms under high pressureThe CH4 sorption isotherm, which is the relationship between

the amount of adsorbed CH4 and gas pressure at a constant tem-perature, was measured in the Gas Geochemistry Laboratory ofthe Bureau of Economic Geology. Isothermal adsorption experi-ments were conducted at 35 �C, 50 �C, and 65 �C, respectively,

Table 1Lithology and XRD mineralogical compositions of the studied samples.

SampleID

Lithology Quartz Plagioclase K-feldspar

Calcite Dolomite Steatite Pyrite Kaolinite Montmorillonite Illite Chlorite I–Smixedlayer

QT Light brown massive quartzite 100FS Yellow-green massive siltstone 81.63 11.29 1.87 3.31 1.89YLS-3 Light gray–green massive

microcrystalline, illite clay rockswith a wax-like chlorite

1 99

LNS-3 Gray–green massive aphaniticchlorite, clay rocks

0.65 1.75 5.31 92.29

GLT-4 Gray, khaki massive kaolinite clayrocks with a smooth feel

4.99 95.01

I–S Light gray–green silty illite/smectite interstratified clay rocks

50.47 4.13 0.94 44.46

PRT-1 Light gray–green layered smectiteclay rocks, friable, shiny and with asmooth feel

26.09 1.45 4.04 2.98 65.44

PRT-5 Light pink block smectite clayrocks, with a small hardness, shinyand a smooth feel

14.25 2.3 1.67 3.12 78.65

Mineral content is in wt.%.

L. Ji et al. / Applied Geochemistry 27 (2012) 2533–2545 2535

and CH4 (of ultra-high purity, 99.999%) was used as the adsorbategas. The range of experimental pressure was 0–14.5 MPa. All sam-ples were heated in an activation furnace at a constant tempera-ture of 250 �C for 4 h with He as a carrier gas before beingsubjected to the gas adsorption test.

The experimental setup for pure gas adsorption basically con-sists of a reference cell for gas storage and a sample cell connectedby a 1/1600 tube and a three-port Valco valve. The volumes of thereference cell and the sample cell are 2.38 mL and 7.15 mL, respec-tively. In the experimental setup, reference cell, sample cell, pres-sure transducer (Series PA-33X/80801/5000PSIS, provided byKeller America), and Valco valves (VICI) are placed in the HP5890GC oven box as a thermostat. A thermal couple was used for tem-perature measurement, and the accuracy is within 0.1 �C. Variationof temperature and pressure was monitored and recorded by com-puter. The detailed experimental construction is described byZhang et al. (2012).

The pretreated samples were placed in the sample cell. A 2-lminline filter was used to prevent mineral particles from enteringthe valves. A system-leak test was performed after the loaded sam-ple cell was connected to the experimental setup. The adsorptionsystem, including both reference and sample cells, was first pres-surized with He gas to 15 MPa, and the pressure change was thenmonitored for 2 h at a constant temperature (e.g., 35 �C). Anacceptable leakage rate is 6.89 � 10�4 MPa/h. The experimentalsetup was evacuated with a vacuum pump (XDS10 Edwards) afterthe leak test was completed. The void volume of the sample cellwas determined by He expansion. Five void volume points weremeasured to ensure accuracy from a He pressure of 0.69–15 MPa.The void volume measurements at 35 �C, 50 �C and 65 �C were con-ducted individually and applied in the calculation of the corre-sponding CH4 adsorption isotherms. This was found to be anappropriate way to correct the systematic error in void volumecaused by changing the system temperature.The experimental set-up was evacuated, and then CH4 was introduced into the referencecell. During isotherm measurement, a certain amount of CH4 wascharged into the reference cell. After equilibrium was reached, asassessed by monitoring pressure variations of <6.9 � 10�4 MPa,within 5 min, the valve between the reference cell and the samplecell was opened, and the gas was expanded into the sample cell. Bymeasuring the pressures before and after expansion, gas molardensities at different stages were calculated using an appropriateequation of state (EOS), and the amount of gas adsorbed at one

pressure level could be determined. The isotherm was obtainedby repeating these procedures until the measurement at the high-est desired gas pressure was achieved. The amount of adsorbed gas(excess sorption or Gibbs excess sorption), Cads(P) in mmol/g units,at pressure, P, is given by Lu et al. (1995) and the equations to cal-culate the adsorbed gas at any given pressure are presented inZhang et al. (2012). The thermophysical properties database ofthe US National Institute of Standards and Technology (NIST) wasused for calculating the density of He and CH4 at any given exper-imental temperature and pressure conditions on the basis of theequations of state of McCarty and Arp (1990) and Setzmann et al.(1991).

2.2.2. Langmuir equation and theoretical fitting of experimentalsorption isotherms

Adsorption is the accumulation of molecules on the surfaces ofa material (adsorbent). This process creates a film of the adsorbateon the adsorbent’s surface and is a consequence of surface energy(Gregg and Sing, 1982). The adsorption process is generally classi-fied as physisorption (characteristic of weak van der Waals force) orchemisorption (characteristic of covalent bonding). Adsorption isusually described through isotherms, i.e., the amount of adsorbedgas on the adsorbent as a function of its pressure at constant tem-perature. The quantity adsorbed is nearly always normalized bythe mass of the adsorbent to allow comparison of different materi-als. The Langmuir equation or Langmuir isotherm is an isothermthat relates the coverage or adsorption of gas molecules on a solidsurface to gas pressure at a fixed temperature. Assuming that gasadsorption follows the monolayer adsorbates theory, which statesthat clay-dominated shale as a sorbent offers many nearly energet-ically homogenous adsorption sites (Gregg and Sing, 1982; Kellerand Staudt, 2005), the ratio of the amount of adsorbed gas (C) ata given pressure (P) to the Langmuir maximum amount of ad-sorbed gas (Cmax) on the sorbent is equivalent to the fractional sur-face coverage. Therefore, CH4 adsorption on organic-rich shale canbe expressed by the following Langmuir equation:

C ¼ CmaxK � P

1þ K � Pð1Þ

where P is pressure in MPa, K is Langmuir constant in 1/MPa. TheLangmuir pressure constant (PL, in MPa), which is a reciprocal ofthe Langmuir constant, represents the pressure at which

Fig. 1. Photomicrographs of clay rock samples under transmitted light. (A, B) Kaolinized silty mudstone, Sedimentary laminations are apparent and slightly offset by a soft-sediment fault (red arrows). All grains in this rock, including quartz, are massively kaolinite-replaced. Yellow outline indicates the outline of one of the larger such ‘‘ghostgrains’’. (A) Plane-polarized light. (B) Cross-polarized light. Sample No.: GLT-4. (C, D) Montmorillonitic bentonite. No obvious evidence of volcanic ash (e.g., shard ghosts) isvisible. Silt-size grains of a variety of minerals (oxides, silicates, and carbonate; blue arrows) may be the remains of crystal ash. In cross-polar view note the unusually largecrystal size of the clay (10 s of lm). (A) Plane-polarized light. (B) Cross-polarized light Sample No.: PRT-5. (D, E) Illitized rock. Massive illitization has left little evidence of theprotolith. Sample No.: YLS-3. (F, G) Chloritized schist. Similar to the kaolinized sample above, chlorite has massively replaced the precursor rock. Yellow outlines denoteformer porphyroblasts (possible plagioclase grains) that have not been entirely chloritized. Sample LNS-3. (For interpretation of the references to color in this figure legend,the reader is referred to the web version of this article.)

2536 L. Ji et al. / Applied Geochemistry 27 (2012) 2533–2545

gas-storage capacity equals one-half of maximum gas adsorptioncapacity—an important parameter for evaluating the feasibility ofgas desorption under reservoir pressure. The Langmuir coefficientcan be obtained with Eq. (1) using adsorption isotherms. The

temperature dependence of the Langmuir coefficient is describedby Eq. (2) (Xia et al., 2006; Xia and Tang, 2012).

K � po ¼ expq

RTþ Dso

R

� �or ln K ¼ q

RTþ Dso

R� ln po ð2Þ

Fig. 2. SEM images of clay crystals. Note that all 4 of the samples, both those of sedimentary/diagenetic (A, B) and metamorphic (C, D) origins, contain many or mostly crystalsthat are larger than the 2 lm upper limit of clay-size. (A) Kaolinite booklets. Coarser kaolinite crystals have nucleated on detrital micas (m) but much of kaolinite is of clay-size within the matrix. Sample GLT-4. (B) Montmorillonite crystals. Silt particles include monocrystaline silicates (s) as well as lithic fragments (L). Sample PRT-5. (C) Illitecrystals. Sample YLS-3. (D) Chlorite crystals. Sample LNS-3.

L. Ji et al. / Applied Geochemistry 27 (2012) 2533–2545 2537

where the Langmuir coefficient K is a function of adsorption ther-modynamic parameters; po is 0.1 MPa as standard atmosphericpressure; q is the heat of adsorption; Dso is standard entropy ofadsorption, and R is the gas constant with the value of8.3145 JK�1 mol�1.

The Langmuir maximum adsorbed amount (Cmax) at differenttemperatures is constant and depends only on the availability ofthe surface area of adsorbents, and the Langmuir constant (K, theratio of adsorption to desorption rate) is a function of temperature.Given a fixed Langmuir maximum for a specific clay-mineral dom-inated rock, the Langmuir coefficient (K) at 35 �C, 50 �C, and 65 �Ccan be determined with a least-squares method to fit the experi-mentally measured CH4 sorption isotherms and the calculatedones with Eq. (1). Then the heat of adsorption and the standard en-tropy can be derived with Eq. (2) from the Langmuir coefficient Kusing a plot of lnK versus 1/T with several different temperaturepoints. The slope of the line in the plot is q/R, and the interceptis (Dso/R – ln0.1). So a solution can be obtained for q and Dso ofCH4 sorption on a clay-mineral-dominated rock.

2.2.3. Specific surface area from N2 sorptionThe BET specific surface areas of the samples were determined

using a Beckman Coulter SA3100 surface area analyzer. The BETsurface area is calculated from a multilayer adsorption theorywhich is based upon the assumption that the first layer of mole-cules adsorbed on the surface involves adsorbate–adsorbent ener-gies, and subsequent layers of molecules adsorbed involve theenergies of vaporization of the adsorbate–adsorbate interaction.

The BET equation should produce a straight line plot, the linearform of which is most often represented as:

1

VaPoPs� 1

� � ¼ 1Vm � C

þ C � 1Vm � C

� �� Ps

Poð3Þ

where Va and Vm (ml/g rock) are the volume adsorbed and the vol-ume equivalent to monolayer adsorption, respectively. Po and Ps arethe saturation pressure and sample pressure, respectively, and C is aconstant related to the enthalpy of gas adsorption. The specified rel-ative pressure range (usually 0.05–0.34) is chosen and the isothermdata is used to calculate the BET function. Vm and C are determinedfrom the slope (C � 1)/VmC and intercept (1/VmC) of a straight linewhich is plotted on the left side of Eq. (3) against relative pressure(Ps/Po). The BET surface area in (m2/g) is then determined from thefollowing expression:

SBET ¼Vm � Na � Am

Mvð4Þ

where SBET is the BET surface area, Na is Avogadros number, Am is thecross sectional area occupied by each adsorbate molecule, which isassumed to be 0.162 nm2 for nitrogen BET determination, and Mv isthe gram molecular volume (22414 mL). Samples with masses from0.5 to 1 g were outgassing at 200 �C for 2 h. The sorption measure-ments were performed at �196 �C (77 K) with N2 as the adsorbategas. The amount of gas adsorbed (Va) at relative pressures (Ps/Po)from 0.05 to 0.34 was used to compute the specific areas.

2538 L. Ji et al. / Applied Geochemistry 27 (2012) 2533–2545

3. Results

3.1. CH4 sorption isotherms on clay-mineral-dominated rocks at 50�C

A series of CH4 sorption isotherms on clay-mineral-dominatedrocks were measured at pressures up to 14.5 MPa and at 50 �C.For comparison, CH4 sorption isotherms on siltstone and quartzitewere measured under the same conditions. Fig. 3 and Table 2 showthe comparison of CH4 adsorbed on different clay-mineral-domi-nated rocks at 50 �C. The particle size of the samples in this com-parison studies was <53 lm (passing a 270 mesh sieve).

As shown in Fig. 3, CH4 sorption capacity varies significantly fordifferent clay types. The montmorillonite-based bentonite has thelargest CH4 sorption capacity of all tested samples, and the maxi-mum amount of adsorbed CH4 reached 0.33 mmol/g rock. Illite/smectite interstratified clay (I–S mixed-layer clay) and kaolinite,have significantly less CH4 sorption capacity; the maximumamount of adsorbed CH4 is 0.13 mmol/g rock for I–S mixed-layerclay, and 0.08 mmol/g rock for the kaolinite. Chlorite and illite havemaximum CH4 excess adsorbed values of 0.06 mmol/g rock and0.05 mmol/g rock, respectively. As expected, CH4 sorption capaci-ties for siltstone and quartzite are much lower than for all claymineral-dominated rocks, with maximum adsorbed CH4 of0.04 mmol/g rock and 0.02 mmol/g rock, respectively. The N2 BETmeasurements yielded BET surface area of 76.4 m2/g for the mont-morillonite, 30.8 m2/g for I–S mixed-layer clay, 15.3 m2/g for kao-linite, 11.7 m2/g for chlorite and 7.1 m2/g for illite. The sequenceof the decrease in CH4 sorption capacity for different clay-min-eral-dominated rocks is closely related to the surface area, andthe details will be discussed in Section 4.1.

A slightly higher CH4 sorption capacity for the siltstone versuspure quartz, on which CH4 sorption capacity is very low, might beattributed to the presence of a small amount of clay and micropo-rous feldspar. Therefore, in clay-mineral dominated rocks, eventhose containing significant quartz, should have a CH4 sorptioncapacity that relates mostly to their clay content. For example,the montmorillonite-rich clay and I–S mixed-layer clay contain78% and 44.5% clay (wt.%), respectively, so that the calculated CH4

sorption capacity based on the pure clay mineral is 0.42 mmol/gmontmorillonite and 0.29 mmol/g I–S mixed layer, respectively.Gas sorption capacity on montmorillonite is larger by a factor of1.3 than the I–S mixed-layer clay according to the results.

3.2. CH4 sorption isotherms on the samples of different particle sizes at50�C

To examine the effect of particle size on CH4 sorption capacity,the rock chips of chlorite clay (LNS-3) and montmorillonite clay

Fig. 3. CH4 sorption isotherms on clay-mineral dominated rocks at 50 �C.

(PRT-5) were crushed and then sieved into four different particlesize groups: 270–830 lm (20–50 mesh), 150–270 lm (50–100 mesh), 53–150 lm (100–270 mesh), and < 53 lm (>270 mesh).A series of CH4 adsorption isotherms on each aliquot of these clayswere measured with CH4 pressures up to 14.5 MPa at 50 �C, and theresults are summarized in Table 3. As shown in Fig. 4a and b, CH4

sorption capability generally increases with decreasing particlesizes. The finest particle size has the largest gas sorption capacityfor both of montmorillonite clay and chlorite clay.

It was found that the difference in CH4 sorption capacity formontmorillonite across different crushed particle sizes is small.The maximum amount of adsorbed CH4 on montmorillonite clayis about 0.27 mmol/g rock for the coarsest fraction, which takesmore than 85% of the total amount of adsorbed CH4 by the finestparticles. In other words, more than 85% of the surface is exposedwhen the montmorillonite fragments are 270–830 lm, suggestingthat the pores are dominated by the micropore system which arebetter connected in this sedimentary sample regardless of the dif-ference in prepared particle size. In contrast, a large difference inCH4 sorption capacity occurs in the chlorite clay between thecoarsest and finest fractions. The maximum amount of adsorbedCH4 is about 0.02 mmol/g rock for the coarsest fraction, whichamounts to less than 30% of the total amount of adsorbed CH4 onthe finest one. This adsorption behavior might suggest that thehydrothermally-formed chlorite is less-developed, or the absenceof a micropore system.

The measured BET surface area of studied samples is shown inTable 4. BET surface area increases from 66.5 m2/g to 76.4 m2/gfor the coarsest to finest montmorillonite fractions, and from2.9 m2/g to 11.7 m2/g for the chlorite fractions. The increase inBET surface area with the finer particle size relates to the greaterexposure of the internal surfaces of connected pores. The externalsurface area per unit mass of non-porous media can be calculatedaccording to the formula of 6/qL (herein L is the radius of the par-ticle size and q is the sample density). Given the average density of2.55 g/cm3 for clays and the average radius of 550, 210, 102 and27 lm for four different particle size groups, the calculated exter-nal surface area per unit mass of the four groups of different parti-cle size are 0.0021, 0.0056, 0.011 and 0.043 m2/g, respectively.Apparently, the external surface area of the finest sample groupis still far less than the measured BET surface area. Therefore, theconnected micropores in montmorillonite-rich and chlorite-richclays are the primary factor in the high BET surface area and CH4

sorption capacity.

3.3. CH4 sorption isotherms for clay-mineral dominated rocks at 35�C,50�C and 65�C

Methane sorption isotherms were measured on the five clay-rich samples at 35 �C, 50 �C and 65 �C under a CH4 equilibriumpressure of up to 14.5 MPa. Differences in specific surface areaand clay mineral type results in variation of CH4 sorption capacity,which is clearly a function of temperature and pressure (Fig. 5a–d).Measured CH4 sorption isotherms at different temperatures can bewell fitted by the Langmuir function (Fig. 5a–d). To achieve such afit, the three isotherms were fitted simultaneously with a constantLangmuir maximum adsorbed amount (Cmax) at different temper-atures and the different Langmuir constant K, which is a function oftemperature. Table 5 summarizes these two parameters obtainedfrom the experimental CH4 sorption isotherms on different clay-mineral-dominated rocks. The Langmuir maximum of CH4 sorptioncapacities for montmorillonite, I–S mixed-layer clay, illite, kaolin-ite, and chlorite are 0.38 mmol/g rock, 0.18 mmol/g rock,0.08 mmol/g rock, 0.12 mmol/g rock and 0.10 mmol/g rock, respec-tively (Table 5). For the illite, kaolinite and chlorite samples(all > 95% clay), the values of the Langmuir maximum CH4 sorption

Table 2Measured CH4 sorption capacity for various clay-mineral dominated rocks at different temperatures.

35.4 �C 50.4 �C 65.4 �C

P (MPa) CH4 (mmol/g rock) P (MPa) CH4 (mmol/g rock) P (MPa) CH4 (mmol/g rock)

Montmorillonite0.54 0.066 0.57 0.059 0.58 0.0481.19 0.119 1.22 0.106 1.24 0.0891.84 0.159 1.92 0.143 1.96 0.1232.56 0.193 2.66 0.174 2.71 0.1513.33 0.221 3.41 0.199 3.46 0.1754.08 0.242 4.19 0.220 4.22 0.1944.84 0.259 5.17 0.240 5.02 0.2105.61 0.273 6.07 0.255 5.85 0.2246.77 0.288 7.03 0.266 6.74 0.2367.90 0.296 8.01 0.275 7.71 0.2469.02 0.301 9.00 0.280 8.68 0.254

10.17 0.302 10.05 0.284 9.76 0.26011.38 0.300 11.24 0.285 10.83 0.263

Illite-semectite0.40 0.017 0.40 0.014 0.22 0.0070.90 0.033 0.88 0.027 0.59 0.0161.46 0.048 1.44 0.040 1.02 0.0262.15 0.063 2.06 0.052 1.55 0.0372.95 0.078 2.76 0.065 2.15 0.0473.94 0.093 3.57 0.076 2.86 0.0585.05 0.107 4.43 0.087 3.69 0.0696.25 0.118 5.40 0.097 4.57 0.0797.41 0.126 6.47 0.106 5.50 0.0888.80 0.132 7.61 0.113 6.54 0.096

10.11 0.134 8.82 0.119 7.65 0.10311.28 0.134 10.05 0.122 8.80 0.108

11.20 0.123 9.88 0.111

Illite0.61 0.008 0.50 0.006 0.52 0.0021.22 0.015 1.07 0.011 1.06 0.0081.89 0.021 1.68 0.017 1.72 0.0142.57 0.026 2.31 0.021 2.45 0.0193.27 0.031 2.99 0.025 3.14 0.0243.97 0.036 3.68 0.029 3.89 0.0284.70 0.040 4.40 0.033 4.67 0.0325.46 0.044 5.18 0.037 5.52 0.0366.28 0.047 6.01 0.039 6.42 0.0397.16 0.050 6.89 0.042 7.43 0.0428.13 0.051 7.78 0.044 8.53 0.0449.11 0.054 8.79 0.046 9.69 0.045

10.17 0.054 9.89 0.047 10.88 0.04511.32 0.055 11.11 0.047 11.86 0.04512.17 0.056 12.03 0.047 12.59 0.045

Kaolinite0.63 0.009 0.60 0.006 0.64 0.0051.26 0.019 1.25 0.015 1.29 0.0121.92 0.028 1.94 0.023 1.94 0.0192.63 0.037 2.65 0.031 2.62 0.0253.36 0.044 3.34 0.037 3.33 0.0314.12 0.051 4.06 0.044 4.08 0.0374.87 0.058 4.77 0.049 4.85 0.0425.72 0.063 5.49 0.054 5.69 0.0476.60 0.069 6.22 0.058 6.58 0.0517.51 0.073 7.03 0.062 7.56 0.0558.45 0.077 7.94 0.066 8.55 0.0589.44 0.080 8.84 0.069 9.27 0.060

10.46 0.082 9.79 0.071 10.18 0.06111.57 0.083 10.74 0.073 11.20 0.063

11.71 0.074

Chlorite0.62 0.007 0.66 0.005 0.60 0.0051.30 0.016 1.35 0.013 1.28 0.0132.02 0.024 2.08 0.021 1.98 0.0192.78 0.031 2.90 0.028 2.76 0.0263.60 0.038 3.80 0.034 3.62 0.0324.50 0.045 4.81 0.041 4.50 0.0375.49 0.050 5.95 0.047 5.49 0.042

(continued on next page)

L. Ji et al. / Applied Geochemistry 27 (2012) 2533–2545 2539

Table 2 (continued)

35.4 �C 50.4 �C 65.4 �C

P (MPa) CH4 (mmol/g rock) P (MPa) CH4 (mmol/g rock) P (MPa) CH4 (mmol/g rock)

6.53 0.055 7.16 0.052 6.57 0.0477.57 0.059 8.46 0.056 7.72 0.0508.73 0.062 9.78 0.058 9.04 0.0549.94 0.063 11.11 0.059 10.30 0.055

11.24 0.064 12.08 0.060 11.51 0.056

Note: Unit conversion factor: 1 mmol/g = 711.24 scf/ton.

Table 3Measured CH4 sorption capacity for montmorillonite and chlorite of various particlesize at 50 �C.

Particle size Montmorillonite Chlorite

P (MPa) CH4 (mmol/g rock) P (MPa) CH4 (mmol/g rock)

20–50 Mesh0.56 0.052 0.51 0.0021.21 0.096 1.20 0.0051.91 0.132 1.93 0.0072.64 0.161 2.58 0.0083.39 0.185 3.41 0.0104.18 0.205 4.34 0.0114.97 0.221 5.42 0.0135.78 0.234 6.59 0.0146.70 0.245 7.90 0.0157.66 0.254 9.30 0.0158.64 0.260 10.61 0.0169.65 0.264 11.93 0.02

10.75 0.266 13.43 0.0211.86 0.265

50–100 Mesh0.57 0.059 0.76 0.0041.22 0.106 1.60 0.0071.92 0.144 2.58 0.0112.66 0.174 3.59 0.0133.41 0.199 4.66 0.0164.19 0.220 5.76 0.0175.17 0.240 6.82 0.0196.07 0.255 7.95 0.0207.03 0.266 9.10 0.0218.01 0.275 10.32 0.0219.00 0.280 11.51 0.021

10.05 0.28411.24 0.285

100–270 Mesh0.57 0.057 0.74 0.0021.18 0.102 1.49 0.0061.81 0.140 2.27 0.0092.53 0.174 3.10 0.0123.33 0.206 4.00 0.0144.31 0.235 4.95 0.0175.42 0.260 6.00 0.0196.58 0.279 7.12 0.0207.88 0.292 8.25 0.0219.21 0.299 9.40 0.022

10.66 0.300 10.62 0.02311.72 0.298 11.86 0.023

>270 Mesh (see Table 2)0.55 0.060 0.66 0.0051.17 0.107 1.35 0.0131.84 0.147 2.08 0.0212.55 0.180 2.90 0.0283.32 0.208 3.80 0.0344.11 0.232 4.81 0.0414.95 0.252 5.95 0.0475.88 0.270 7.16 0.0526.93 0.284 8.46 0.0567.97 0.294 9.78 0.0589.11 0.301 11.11 0.059

10.24 0.30511.43 0.305

Note: Unit conversion factor: 1 mmol/g = 711.24 scf/ton.

Fig. 4. CH4 sorption isotherms on montmorillonite and chlorite dominated rockswith different particle size at 50 �C.

2540 L. Ji et al. / Applied Geochemistry 27 (2012) 2533–2545

capacities are close to those normalized to the mass of clay min-eral. The concentration of clay minerals in the montmorilloniteclay and illite/smectite interstratified clay samples is about 78.6%and 44.5%, respectively. The Langmuir maximum of CH4 adsorptioncapacity normalized to clay minerals is 0.48 mmol/g for montmo-rillonite and 0.41 mmol/g for I–S mixed layer, respectively. Mont-morillonite clay has a higher sorption capacity than the I–Smixed layer and others, whereas illite shows the lowest sorptioncapacity. The sequence in Langmuir maxima of CH4 sorption capac-ity with clay mineral types is: montmorillonite (0.48 mmol/g) > I–Smixed layer (0.41 mmol/g) > kaolinite (0.12 mmol/g) > chlorite(0.10 mmol/g) > illite (0.08 mmol/g) (Table 5). In contrast, theLangmuir constant decreases with clay mineral type in the follow-ing sequence: montmorillonite > I–S mixed layer > illite > kaolin-ite = chlorite (Table 5).

4. Discussion

4.1. Effect of clay mineral type on CH4 sorption

Methane sorption is closely related to the type of clay mineral.Fig. 7 shows a linear correlation between the Langmuir maximum

Table 4BET surface area (m2/g) of the studied samples.

Sample ID Lithology 20–50 Mesh 50–100 Mesh 100–270 Mesh >270 Mesh

QT Light brown massive quartzite 1.32 1.70 1.91 3.85FS Yellow-green massive siltstone 8.48YLS-3 Light gray–green massive microcrystalline, illite clay rocks with a wax-like chlorite 2.26 3.21 4.00 7.12LNS-3 Gray–green massive aphanitic chlorite, clay rocks 2.88 4.40 5.38 11.74GLT-4 Gray, khaki massive kaolinite clay rocks with a smooth feel 11.47 12.41 13.38 15.28I–S Light gray–green silty illite/smectite interstratified clay rocks 29.88 31.31 30.79 30.83PRT-5 Light pink block smectite clay rocks, with a small hardness, shiny and a smooth feel 66.48 71.50 74.11 76.41

Fig. 5. CH4 sorption isotherms on clay-mineral dominated rocks at 35 �C, 50 �C and 65 �C. Points are the measured amount of adsorbed CH4, and lines are calculated ones ofadsorbed CH4 based on the Langmuir equation.

Table 5Langmuir fitting results of methane adsorption for the different clay-mineral-dominated rocks.

Montmorillonite I–S mixed layer Illite Kaolinite Chlorite

Langmuir maximum (mmol/g rock) 0.380 0.182 0.079 0.120 0.103Langmuir constant (1/MPa)

35.4 �C 0.451 0.269 0.211 0.179 0.16550.4 �C 0.343 0.203 0.181 0.134 0.13265.4 �C 0.254 0.163 0.148 0.105 0.119

Note: Unit conversion factor: 1 mmol/g = 711.24 scf./ton.

L. Ji et al. / Applied Geochemistry 27 (2012) 2533–2545 2541

for CH4 sorption on different clay-mineral-dominated rocks andthe BET surface area. This indicates CH4 sorption on clay-mineraldominated rocks is controlled by the specific surface area. The fol-lowing linear regression with a Pearson R2 value of 0.92 is obtainedbased on the experimental observations:

CCH4ðadsÞ ¼ 0:0053SBET ð5Þ

where CCH4ðadsÞ is the Langmuir maximum of adsorbed CH4 in mmol/g rock, and SBET is the measured BET surface area in m2/g rock, andthe unit for the coefficient is in mmol/m2. Given the cross sectionalarea of 0.113 nm2 for CH4 and 0.162 nm2 for N2 in the considerationof the diameter difference in CH4 and N2 molecules, the theoreticalcalculated coefficient should be 0.007 mmol/m2. The experimen-tally-obtained coefficient is 0.0053 mmol/m2, which is slightly

2542 L. Ji et al. / Applied Geochemistry 27 (2012) 2533–2545

smaller than the theoretical one. The main reason for the differenceis possibly that the Langmuir maximum (CCH4ðadsÞ) from CH4 sorp-tion isotherms is based on the monolayer adsorption theory, how-ever, BET surface area (SBET) is based on the multilayer adsorptiontheory. Nevertheless, A comparative result in this study from CH4

sorption isotherms up to the pressure of 15 MPa and the tempera-ture of 65 �C and N2 sorption isotherms up to a pressure of0.1 MPa and a temperature of �196 �C provides a way to correlateCH4 gas sorption capacity and pore characterization of clay-min-eral-dominated rocks and organic-rich mudstones. In addition, theSBET value of metamorphic illite in this study was approximatelythree times smaller than the SBET values of 20.5 and 17.5 m2/g foundfor the Cambrian shale illites IMt-1 and IMt-2 from Montana, USA(Dogan et al., 2007), probably a consequence of the large crystal sizein the metamorphic rocks. The SBET value of the montmorillonitewas well within the wide range of SBET values from 50 to130 m2 g�1 found for Ca/Mg montmorillonites (Van Olphen and Fri-piat, 1979; Kaufhold et al., 2010). The specific surface area of min-erals is an essential parameter to quantify interaction processes atthe gas–solid interface. Swelling clays such as montmorillonitesadditionally exhibit internal surfaces in their interlayer spaces.Although these internal surfaces bear permanent negative chargeslike the basal surfaces, they have only limited accessibility for CH4

molecules. In clays, pores in the range of 2–50 nm in diameter aremainly interparticle pores produced by the microstructuralarrangement of the clay mineral particles and have a minor effecton gas adsorption surface area (Kaufhold et al., 2010), althoughintraparticle pores within organic matter and dissolved feldsparand carbonate grains have also been documented in this size range(Loucks et al., 2009; Schieber et al., 2010). Pores < 2 nm in diametermay result from intraparticle space in quasi crystalline overlap re-gions and from the partly N2-accessible interlayer space at particleedge surfaces of swelling clay minerals (Aylmore and Quirk, 1967;Rutherford et al., 1997; Michot and Villieras, 2006; Kaufholdet al., 2010).

The clay mineral composition also affects the affinity of CH4

molecules for the sorbent surface, and larger values of the Lang-muir constant (or lower Langmuir pressure) indicate strongeraffinity of the gas for the sorbent. A least-squares fit was appliedto the experimentally measured CH4 adsorption isotherms andthe calculated ones based on the Langmuir function for determin-ing Langmuir constants at 35 �C, 50 �C and 65 �C. For the samples

Fig. 6. The plot of natural logarithms of the Langmuir constant (lnK) versus the recipregressions to the measured points with R-squared values larger than 0.97.

examined in this study, the value of the Langmuir constant varieswith clay mineral type: montmorillonite clay exhibits the largestLangmuir constant, kaolinite and chlorite clay exhibit the lowest(Fig. 6) and I–S mixed layer and illite are intermediate. A clear lin-ear relationship exists between the natural logarithm of the Lang-muir constant (K) and the reciprocal of temperature (1/T) (Fig. 6).The linear relation is defined by the experiments on the impactof clay mineral type on CH4 adsorption. Empirical equations (6)–(9) were derived from this study:

lnðKÞ ¼ 1999=T � 7:26 montmorillonite ð6Þ

lnðKÞ ¼ 1752=T � 6:99 I-S mixed-layer clay ð7Þ

lnðKÞ ¼ 1240=T � 5:56 illite ð8Þ

lnðKÞ ¼ 1135=T � 5:49 kaolinite and chlorite ð9Þ

where T is temperature in degrees Kelvin and K is the Langmuir con-stant in units of 1/MPa. Note that the above, empirical linear regres-sions were obtained from a limited number of clay-mineraldominated rock samples. Differences in chemical structure of clayminerals, depositional environment, and moisture content may al-ter these relationships (Krooss et al., 2002; Busch et al., 2003,2004, 2006; Ross and Bustin, 2007, 2008, 2009; Fathi and Akkutlu,2009). The results show, however, that the effects of varying claymineral composition and temperature on CH4 sorption can be quan-tified by determining the Langmuir constant at differenttemperatures.

In this study, the effect of the presence of moisture on gasadsorption by clay minerals was not investigated quantitatively.The issue needs to be addressed in a future study because the pres-ence of moisture can greatly reduce gas-sorption capacity. Thehypothesis is that moisture mainly occupies surface sites of hydro-philic clay minerals possibly restricting access of CH4 to active sites(see further discussion below).

4.2. Effect of clay minerals on thermodynamic parameters for CH4

adsorption

The heat of adsorption (q) and the standard entropy of adsorp-tion (Dso) are two important thermodynamic parameters whichcan be used to describe the temperature dependence of the

rocal of temperature (1/T) for clay-mineral dominated rocks. The lines are linear

Fig. 7. Linear correlation between the Langmuir maximum of adsorbed CH4 andBET surface area for clay-mineral dominated rocks.

L. Ji et al. / Applied Geochemistry 27 (2012) 2533–2545 2543

Langmuir constant. Both of these parameters were calculated forclay-mineral dominated rocks according to Eq. (2) with the exper-imentally determined Langmuir constants at three different tem-peratures and are listed in Table 6. The heat of adsorptiondetermined for montmorillonite, I–S mixed layer clay, illite, kaolin-ite and chlorite is 16.6, 14.6, 10.3, 9.6 and 9.4 kJ/mol, respectively,and these values are obviously smaller than the previously re-ported quantities for kerogen (21.9–28 kJ/mol) and organic-richshales (15.1–18.4 kJ/mol) (Zhang et al., 2012). The observed trendof decreasing heat of adsorption for montmorillonite to chlorite re-flects the difference in the affinity of CH4 for different types of clayminerals. The differences in the heats of adsorption determined fordifferent types of clay minerals are likely related to differences inthe structures and pores present in the clays. As discussed above,compared to illite, kaolinite and chlorite, montmorillonite has rel-atively more and smaller pores and internal surface areas that mayenhance CH4 adsorption. Myers (2004) has shown that the stan-dard enthalpy (the absolute value is equivalent to the heat ofadsorption) of CH4 molecules on the zeolites of different pore-sizeincreases strongly with decreasing pore size due to hindered trans-lational and rotational motion for the CH4 molecule. The entropy ofadsorption reflects the restricted mobility of adsorbed molecules(Xia et al., 2006). The entropy of physisorption should be close tothe loss of translational entropy from three-dimensional free gasto two-dimensional adsorbates, and this entropy loss is �82 to�87 J/mol/K at temperatures of 273–423 K for CH4 molecules(Xia and Tang, 2012). The Dso for CH4 adsorption on the clay min-erals studied is �64.8 to �79.5 J/mol/K, and the absolute values ofentropy for clay minerals are relatively small.

The difference in the thermodynamic parameters for CH4

adsorption on kerogen, organic-rich shales and organic-lean clayminerals (Table 6) provides a way to quantitatively estimate theindividual contribution of organic matter and clay minerals toCH4 adsorption under shale-gas reservoir temperature and pres-sure conditions. A large CH4 sorption capacity is always associatedwith high TOC content, and an empirical linear correlation be-tween the CH4 sorption capacity and TOC content was observedby Zhang et al. (2012) and was used to calculate CH4 sorptioncapacity at a given TOC content for a shale gas reservoir. Methanesorption capacity for organic-lean clay minerals can be estimatedby Eq. (5).

The values of the heat of energy and standard entropy for or-ganic-rich shales and clay minerals in Table 6 are used to calculatethe Langmuir pressure for given reservoir rock properties andtemperature conditions. In fact, the extrapolation to the reservoir

Table 6Thermodynamic parameters of CH4 adsorption on different materials.

Adsorbent q(kJ mol�1)

Ds(J mol�1 K�1)

Montmorillonite clay 16.6 �79.5I–S mixed layer 14.6 �77.2Illite clay 10.3 �65.3Kaolinite clay 9.6 �65.3Chlorite clay 9.4 �64.8Paleozoic black shale (TOC = 5.8%, Ro = 2.4%)

moisture-equilibrated (wwater = 2.7%)26.2a �110.8a

Dry 23.5a �100.0a

Blakely #1 shale (TOC = 6.6%, Ro = 2.01%) 18.4b �101.6b

Lee C-5-1 shale (TOC = 7.9%, Ro = 0.58%) 14.0b �76.0b

Woodford shale (TOC = 17.2%, Ro = 0.58%) 13.8b �76.3b

Activated carbon 16–20c �62.8d

Zeolite 15.7c �55.9d

a Calculated from the data of Gaspark et al. (2012).b Zhang et al. (2012).c Cavenati et al. (2004).d Xia and Tang (2012).

temperature and pressure conditions to predict the sorption capac-ity is a complicated task and one should consider the properties oforganic matter, mineral compositions, moisture effect and poretype. Here it is simply assumed that the thermodynamic parame-ters for CH4 adsorption on dry samples might apply to real reser-voir high temperature condition. For example, assuming thetemperature of a shale gas reservoir is 120 �C, the value of theLangmuir constant for clay-rich rocks and organic-rich shales iscalculated based on Eq. (2) with the thermodynamic parametersin Table 6. The estimated gas sorption isotherms on clay-domi-nated rocks and organic-rich shales at 120 �C are shown in Fig. 8a.

Methane sorption capacity on a mineral matrix is greatlydependent on the type of clay minerals, which can be quantifiedwith BET surface area. The CH4 sorption capacity of organic-richshales with TOC content of 6.6% (LeeC-5-1) to 7.9% (Blakely #1)is even less than that of I–S mixed layer clay with a SBET value of30.8 m2/g rock. In contrast, CH4 sorption capacity of the Woodfordorganic-rich shale with a TOC content of 17.6% is obviously greaterthan that of I–S mixed layer clay, suggesting that TOC content is acritical factor in gas adsorption. In addition, high-maturity organic-rich shale (Blakely #1) and porous montmorillonite-dominatedrock give low values of Langmuir pressure (Fig. 8b), however,immature organic-rich shales (Lee C-5-1 Barnett shale and Wood-ford Shale) and other clay-mineral dominated rocks exhibit highLangmuir pressure values. The low Langmuir pressure for high-ma-ture organic-rich shales and montmorrillonite clay possibly resultsfrom abundant development of very small pores (Loucks and Rup-pel, 2007; Myers, 2004).

Gas sorption capacity on organic-rich shales can be greatly re-duced in the presence of moisture (Ross and Bustin, 2009; Gasparket al., 2012), and the reduction in sorption capacity for a moisture-equilibrated sample compared to a dry sample is around 60% (Gas-park et al., 2012). Gaspark et al. (2012) measured CH4 sorption iso-therms for a Paleozoic black shale (TOC = 5.8% and Ro = 2.4%) at45 �C, 65 �C and 75 �C under dry and moisture-equilibrated condi-tions, and the values of Langmuir pressure were determined withthe same approach as reported in this paper. In Table 6, the valuesof q and Dso of CH4 adsorption on dry and moisture-equilibratedPaleozoic black shale are calculated from their reported Langmuirpressure values at three different temperatures. The heat ofadsorption and standard entropy on the moisture-equilibratedsample is 26.2 kJ/mol and �110.8 J/mol K, which is slightly largerthan the values (23.5 kJ/mol and �100.0 J/mol K) of the dry sample.The calculated Langmuir pressure for the moisture-equilibratedsample at 120 �C is 20.5 MPa, which is significantly larger than that(12.8 MPa) for a dry sample at the same temperature. As a result,

(a)

(b)

Fig. 8. Predicted CH4 sorption isotherms (a) and the values of Langmuir pressure (b)for organic-rich shales and clay-mineral dominated rocks at a reservoir temperatureof 120 �C.

2544 L. Ji et al. / Applied Geochemistry 27 (2012) 2533–2545

CH4 desorption from the moisture-equilibrated organic-rich shaleoccurs more readily than from the dry one, and requires less of apressure drop to release adsorbed gas in shale gas production.These comparisons, along with the present work on CH4 adsorptionon kerogen and clays, clearly indicates that adsorption on themoisture-equilibrated sample is mainly due to adsorption on ker-ogen, as the clay is blocked; adsorption on the dry sample is con-tributed by adsorption on both kerogen and clay.

The presence of moisture not only affects the Langmuir pressurefor gas desorption, but also greatly reduces the sorption capacity.For example, measurements of gas-sorption capacity of Devo-nian–Mississippian shales in the West Canada Sedimentary Basin(WCSB) under dry and moisture-equilibrated conditions (Rossand Bustin, 2009) showed that sorption capacity under dry condi-tions is substantially greater than under moisture-equilibratedconditions. And the offset of about 0.07 mmol/g rock betweenmeasured CH4 sorption capacity at dry versus moisture-equili-brated conditions might result from a large exposure of surfacearea of clays and then significant increase of gas-sorption capacityunder dry conditions. The hypothesis is that moisture mainlyoccupies the surface sites of hydrophilic clay minerals, and thedistribution of hydrophobic sorption sites throughout the networkof shale-gas reservoirs would play a major role in the selective

sorption of CH4 and moisture. This issue needs to be addressed fur-ther in future research.

5. Conclusions

The studies of CH4 sorption characteristics of clay-mineral dom-inated rocks reveal important relationships between clay mineraltype and gas adsorption. Significant conclusions are as follows:

1. CH4 sorption capacities of different clay minerals are quite dif-ferent, and CH4 sorption capacity decreases in the order mont-morillonite� I–S mixed layer > kaolinite > chlorite > illite. Thesorption capacity of metamorphic chlorite and illite is obviouslysmaller than the capacity of clays from sedimentary rocks.

2. BET surface area of clay minerals correlates with the CH4 sorp-tion capacity for various types of clay minerals, and the linearcorrelation indicates that physisorption is a dominant processfor CH4 sorption on clay minerals.

3. CH4 adsorption on clay minerals can be described by the Lang-muir monolayer theory. The linear relationship between thelnK and the reciprocal of temperature allows determining ther-modynamic parameters: The heat of adsorption (q) and thestandard entropy (Dso) range from 9.4 to 16.6 kJ/mol and from�64.8 to �79.5 J/mol/K, respectively, which are smaller thanthe values for CH4 adsorption on kerogens.

4. The affinity of CH4 molecules to organic-rich shales is strongerthan CH4 affinity to the most common clay minerals. As a result,it is expected that CH4 molecules may preferentially occupy thesurface sites of organic matter because clay minerals are moreeasily blocked by water.

5. Experimentally-derived thermodynamic parameters can beused for estimating Langmuir pressure and CH4 sorption capac-ity under reservoir temperature and pressure conditions.

Acknowledgments

This research is jointly supported by The Bureau’s Mudrock Sys-tems Research Laboratory consortium, the Major State Basic Re-search Development Program of China (Grant No. 2012CB214704,2012CB214701), the National Science and Technology Major Pro-jects of the Ministry of Science and Technology of China (GrantNo. 2011ZX05008-002-22), Chinese Natural Science Foundation(Grant No. 20100559) and the Jackson School of Geosciences, TheUniversity of Texas at Austin. Thanks to two reviewers Drs. XinyuXia, Bernhard Krooss, and Executive Editor Prof. Ron Fuge for pro-vided constructive suggestions that significantly improved thequality of the manuscript. Publication is authorized by the Direc-tor, Bureau of Economic Geology.

References

Aringhieri, R., 2004. Nanoporosity characteristics of some natural clay minerals andsoils. Clays Clay Min. 52, 700–704.

Aylmore, L.A.G., Quirk, J.P., 1967. The micropore size distribution of clay mineralsystems. J. Soil Sci. 18, 1–17.

Bergaya, F., Theng, B.K.G., Lagaly, G., 2006. Handbook of Clay Science. Developmentsin Clay Science 1. Amsterdam, Elsevier, London.

Busch, A., Gensterblum, Y., Krooss, B.M., 2003. Methane and CO2 sorption anddesorption measurements on dry Argonne premium coals: pure componentsand mixtures. Int. J. Coal Geol. 55, 205–224.

Busch, A., Gensterblum, Y., Krooss, B.M., Littke, R., 2004. Methane and carbondioxide adsorption-diffusion experiments on coal: upscaling and modeling. Int.J. Coal Geol. 60, 151–168.

Busch, A., Gensterblum, Y., Krooss, B.M., Siemons, N., 2006. Investigation of high-pressure selective adsorption/desorption behaviour of CO2 and CH4 on coals: anexperimental study. Int. J. Coal Geol. 66, 53–68.

Cavenati, S., Grande, C.A., Rodrigues, A.E., 2004. Adsorption equilibrium of methane,carbon dioxide, and nitrogen on zeolite 13X at high pressures. J. Chem. Eng.Data 49, 1095–1101.

L. Ji et al. / Applied Geochemistry 27 (2012) 2533–2545 2545

Chalmers, G.R.L., Bustin, R.M., 2007a. On the effects of petrographic composition oncoalbed methane sorption. Int. J. Coal Geol. 69, 288–304.

Chalmers, G.R.L., Bustin, R.M., 2007b. The organic matter distribution and methanecapacity of the Lower Cretaceous strata of northeastern British Columbia,Canada. Int. J. Coal Geol. 70, 223–239.

Chalmers, G.R.L., Bustin, R.M., 2008a. Lower Cretaceous gas shales in northeasternBritish Columbia, Part I: geological controls on methane sorption capacity. Bull.Can. Petrol. Geol. 56, 1–21.

Chalmers, G.R.L., Bustin, R.M., 2008b. Lower Cretaceous gas shales in northeasternBritish Columbia, Part II: evaluation of regional potential gas resources. Bull.Can. Petrol. Geol. 56, 22–61.

Cheng, A.L., Huang, W.L., 2004. Selective adsorption of hydrocarbon gases on claysand organic matter. Org. Geochem. 35, 413–423.

Curtis, J.B., 2002. Fractured shale-gas systems. Am. Assoc. Petrol. Geol. Bull. 86,1921–1938.

Dogan, M., Dogan, A.U., Yesilyurt, F.I., Alaygut, D., Buckner, I., Wurster, D.E., 2007.Baseline studies of the Clay Minerals Society special clays: specific surface areaby the Brunauer Emmett Teller (BET) method. Clays Clay Min. 55, 534–541.

Fathi, E., Akkutlu, I.Y., 2009. Matrix heterogeneity effects on gas transport andadsorption in coal bed and shale gas reservoirs. J. Trans. Porous Media 80, 281–304.

Gaspark, M., Ghanizadeh, A., Gensterblum, Y., Krooss, B.M., Littke, R., 2012. Themethane storage capacity of black shale. In: 3rd EAGE Shale Workshop,Barcelona, pp. 186–188 (Abstract).

Gregg, S.J., Sing, K.S.W., 1982. Adsorption, Surface Area and Porosity, second ed.Academic Press, New York.

Hill, R.J., Jarvie, D.M., Zumberge, J., Henry, M., Pollastro, R.M., 2007. Oil and gasgeochemistry and petroleum systems of the Fort Worth Basin. Am. Assoc.Petrol. Geol. Bull. 91, 445–473.

Jarvie, D.M., Hill, R.J., Ruble, T.E., Pollastro, R.M., 2007. Unconventional shale-gassystems: the mississippian barnett shale of north-central Texas as one modelfor thermogenic shale-gas assessment. Am. Assoc. Petrol. Geol. Bull. 91, 475–499.

Kaufhold, S., Dohrmann, R., Klinkenberg, M., Siegesmund, S., Ufer, K., 2010. N2-BETspecific surface area of bentonites. J. Colloid Interface. Sci. 349, 275–282.

Keller, J.U., Staudt, R., 2005. Gas Adsorption Equilibra: Experimental Methods andAdsorptive Isotherms. Springer Science, New York.

Krooss, B.M., van Bergen, F., Gensterblum, Y., Siemons, N., Pagnier, H.J.M., David, P.,2002. High pressure CH4 and carbon dioxide adsorption on dry and moistureequilibrated Pennsylvanian coals. Int. J. Coal Geol. 51, 69–92.

Loucks, R.G., Ruppel, S.C., 2007. Mississippian Barnett shale: lithofacies anddepositional setting of a deep-water shale-gas succession in the Fort WorthBasin. Texas. Am. Assoc. Petrol. Geol. Bull. 91, 579–601.

Loucks, R.G., Reed, R.M., Ruppel, S.C., Jarvie, D.M., 2009. Morphology, genesis, anddistribution of nanometer-scale pores in siliceous mudstones of theMississippian Barnett Shale. J. Sed. Res. 79, 848–861.

Lu, X.C., Li, F.C., Watson, A.T., 1995. Adsorption measurements in Devonian shales.Fuel 74, 599–603.

McCarty, R.D., Arp, V.D., 1990. A new wide range equation of state for helium. Adv.Cryog. Eng. 35, 1465–1475.

Michot, L.J., Villieras, F., 2006. Surface area and porosity. In: Bergaya, F., Theng,B.K.G., Lagaly, G. (Eds.), Handbook of Clay Science, vol. 1. Elsevier, Amsterdam,pp. 965–978.

Milliken, K.L., Esch, W.L., Reed, R.M., Zhang, T., 2012. Grain assemblages and strongdiagenetic overprinting in siliceous mudrocks, Barnett Shale (Mississippian),Fort Worth Basin, Texas. Am. Assoc. Petrol. Geol. Bull. 96, 1553–1578.

Montgomery, S.L., Jarvie, D.M., Bowker, K.A., Pallastro, R.M., 2005. MississippianBarnett Shale, Fort Worth basin, north-central Texas: gas-shale play with multi-trillion cubic foot potential. Am. Assoc. Petrol. Geol. Bull. 89, 155–175.

Myers, A.L., 2004. Characterization of nanopores by standard enthalpy and entropyof adsorption of probe molecules. Colloids Surf. A 241, 9–14.

Palomino, A.M., Santamarina, J.C., 2005. Fabric map for kaolinite: effects of pH andionic concentration on behavior. Clays Clay Min. 53, 211–223.

Ross, D.J.K., Bustin, R.M., 2007. Shale gas potential of the Lower Jurassic GordondaleMember, northeastern British Columbia, Canada. Bull. Can. Petrol. Geol. 55, 51–75.

Ross, D.J.K., Bustin, R.M., 2008. Characterizing the shale gas resource potential ofDevonian–Mississippian strata in the Western Canada sedimentary basin:application of an integrated formation evaluation. Am. Assoc. Petrol. Geol. Bull.92, 87–125.

Ross, D.J.K., Bustin, R.M., 2009. The importance of shale composition and porestructure upon gas storage potential of shale gas reservoirs. Mar. Petrol. Geol.26, 916–927.

Rowe, H.D., Loucks, R.G., Ruppel, S.C., Rimmer, S.M., 2008. Mississippian BarnettFormation, Fort Worth Basin, Texas: bulk geochemical inferences and Mo-TOCconstraints on the severity of hydrographic restriction. Chem. Geol. 257, 16–25.

Ruppel, S.C., Loucks, R.G., 2008. Black mudrocks: lessons and questions from theMississippian Barnett Shale in the southern Midcontinent. Sediment. Rec. 6, 4–8.

Rutherford, D.W., Chiou, C.T., Eberl, D.D., 1997. Effects of exchanged cation on themicroporosity of montmorillonite. Clays Clay Min. 45, 534–543.

Schieber, J., Southard, J.B., Schimmelmann, A., 2010. Lenticular shale fabricsresulting from intermittent erosion of muddy sediments – comparingobservations from flume experiments to the rock record. J. Sed. Res. 80, 119–128.

Setzmann, U., Wagner, W., Pruss, A., 1991. A new equation of state and tables ofthermodynamic properties for methane covering the range from the meltingline to 625 K at Pressures up to 1000 MPa. J. Phys. Chem. Ref. Data 20, 1061–1151.

Strapoc, D., Mastalerz, M., Schimmelmann, A., Drobniak, A., Hasenmueller, N.R.,2010. Geochemical constraints on the origin and volume of gas in the NewAlbany Shale (Devonian–Mississippian), eastern Illinois Basin. Am. Assoc.Petrol. Geol. Bull. 94, 1713–1740.

Tsipursky, S.I., Drits, V.A., 1984. The distribution of octahedral cations in the 2:1layers of dioctahedral smectites studied by obliquetexture electron diffraction.Clay Min. 19, 177–193.

Van Olphen, H., Fripiat, J.J., 1979. Data Handbook for Clay Minerals and OtherNonmetallic Minerals. Pergamon Press, New York.

Wang, C.C., Juang, L.C., Lee, C.K., Hsu, T.C., Lee, J.F., Chao, H.P., 2004. Effects ofexchanged surfactant cations on the pore structure and adsorptioncharacteristics of montmorillonite. J. Colloid Interface Sci. 280, 27–35.

Xia, X., Tang, Y., 2012. Isotope fractionation of methane during natural gas flow withcoupled diffusion and adsorption/desorption. Geochim. Cosmochim. Acta 77,489–503.

Xia, X., Litvinov, S., Muhler, M., 2006. A consistent approach to adsorptionthermodynamics on heterogeneous surfaces using different empirical energydistribution model. Langmuir 22, 8063–8070.

Zhang, T., Ellis, G.E., Ruppel, S.C., Milliken, K.L., Yang, R., 2012. Effect of organic-matter type and thermal maturity on methane adsorption in shale-gas systems.Org. Geochem. 47, 120–131.