High-Pressure Methane Sorption Isotherms of Black Shales from The Netherlands

High-Pressure Methane Sorption Isotherms of Black Shales from TheNetherlandsM. Gasparik,*,† A. Ghanizadeh,† P. Bertier,‡ Y. Gensterblum,† S. Bouw,§ and Bernhard M. Krooss†

†Institute of Geology and Geochemistry of Petroleum and Coal, RWTH Aachen University, Lochnerstrasse 4-20, D-52056 Aachen,Germany‡Clay and Interface Mineralogy, RWTH Aachen University, Bunsenstrasse 8, D-52072 Aachen, Germany§OMV Exploration and Production GmbH, Trabrennstrasse 6-8, 1210 Vienna, Austria

*S Supporting Information

ABSTRACT: High-pressure methane sorption isotherms were measured on one Paleozoic and five Mesozoic shales, consideredas targets for shale gas exploration in The Netherlands. The samples varied in mineralogy, organic richness, and thermal maturity.Four of the samples were clay-rich (total clay content 60−71 wt %), one contained equal amounts of clays and quartz (36 wt %and 33 wt %, respectively) and one was a marl sample (clays 34 wt %, carbonates 49 wt %). The total organic carbon contents(TOC) ranged from <1 wt % to 10.5 wt %, and the thermal maturity, as inferred from Rock-Eval analysis, from immature toovermature. Excess (Gibbs) sorption isotherms for methane were measured at 65 °C on dry samples up to 25 MPa. Themaximum excess sorption capacities within this pressure range varied from 0.05 to 0.3 mmol/g (1.1−6.8 m3 STP/t). Nocorrelation of excess sorption capacity with TOC was found. Low-TOC, clay-rich shales had comparable or even higher methanesorption capacities per unit rock mass (mmol/g) than organic-rich shales, and a positive correlation was found between themaximum Langmuir capacity (nL) and the clay content. This observation supports the notion that clay minerals can contributesignificantly to the sorption capacity of shales. Furthermore, we demonstrate that significant errors in TOC-normalized sorptioncapacities may result from the uncertainties in TOC contents, especially at low TOC values. A comparison between theimmature and the overmature sample (both organic-rich with equal clay contents) did not show any enhancement of thesorption capacity with thermal maturity. However, the excess sorption isotherm of the overmature sample had a distinctmaximum, while no maximum was observed for the immature sample in the experimental pressure range. A Langmuir-typeabsolute sorption function, with a term taking the volume of the adsorbed phase explicitly into account, gave a goodrepresentation of the measured excess sorption isotherms. The three-parameter fit yielded the Langmuir parameters (nL and pL)and a nominal density value for the adsorbed phase (ρads). Two-parameter fits of nL and pL using different fixed values of ρads arediscussed.

■ INTRODUCTION

The growing interest in shale gas as an important unconven-tional energy resource during the past decades has led toincreased exploration and production activities and hasstimulated considerable research interest. Gas shales havebeen recognized as natural gas reservoirs with enormous gaspotential in the United States and Canada.1−4 As of 2012,Europe is still lacking experience with shale gas production, andthe development is further complicated by public acceptanceand environmental concerns. Nevertheless, considerable effortsare being undertaken by exploration programs in industry andinternational research programs (such as the GASH Project,http://www.gas-shales.org) to expand the inventory on Euro-pean shale gas resources and to improve our understanding ofthese complex reservoirs and their economic potential inEurope.One of the challenging tasks in the assessment of the

economic potential of a shale gas reservoir is the estimation ofthe amount of contained gas (gas-in-place, GIP). The gasstored in shale is composed of (1) “free” compressed gas inpores and fractures and (2) sorbed gas associated with organicand mineral constituents of the rock. The principal control onthe amount of free gas is the porosity. The sorbed gas capacity,

which includes the gas adsorbed on solid surfaces and the gasdissolved in pore fluids (water or hydrocarbon fluids), is acomplex function of chemical and pore-structural character ofthe rock matrix, composition, and properties of fluid phases andreservoir conditions. The contribution of the sorbed gas to thetotal GIP has been postulated to be as much as 50−60% insome cases.1 However, the role of sorbed gas is still poorlyconstrained and only few experimental data exist on gassorption of shales.3,5−8 The experience from coal-bed-methane(CBM) is not directly applicable to gas shales,7 and there is aneed for reliable sorption isotherms at high-pressure and high-temperature conditions.Organic matter (OM) present in these rocks is generally

thought to be the principal contributor to the sorption capacityin shales. In industrial practice, total organic carbon content(TOC) is used as a proxy to determine intervals with high gascontent. In the literature, several authors have found a positivecorrelation between the methane sorption capacity and TOCcontent of organic-rich shales.4,7,8 However, clay minerals have

Received: March 6, 2012Revised: July 4, 2012Published: July 5, 2012

Article

pubs.acs.org/EF

© 2012 American Chemical Society 4995 dx.doi.org/10.1021/ef300405g | Energy Fuels 2012, 26, 4995−5004

http://www.gas-shales.org

pubs.acs.org/EF

also been recognized to participate in the sorption process inclay-rich shales.4,7,9,10 Schettler et al. (1991) postulated thatTOC controls the methane sorption capacity in high-TOCshales, whereas comparable sorption capacities can be found inlow-TOC shales, which the authors attributed to the sorptiveproperties of clay minerals, in particular illite.9 Appreciablemethane sorption capacities, comparable to those of organic-rich shales, were measured on pure illite7,9,10 and montmor-illonite.7 However, Ross and Bustin7 argued that thecontribution of clay minerals to the sorption capacity in shalesis irrelevant in the presence of moisture due to the high affinityof water to the clay surfaces that block the access of gasmolecules to the sorption sites.Other factors affecting methane sorption capacity of shales

are (1) organic matter type and (2) thermal maturity, (3) watersaturation and (4) temperature. Ross and Bustin4,7 suggestedthat the thermal maturation of the organic matter in shalespositively affects their sorption capacity. The authors attributedthis to the increase in the microporosity and/or decrease in theheterogeneity of pore surface in the process of the trans-formation of the organic matter. However, their finding hasonly a qualitative character, and they could not find any distinctrelationship between the maturity and the methane sorptioncapacity.They also investigated the influence of moisture on the

methane sorption in shales by comparing the sorptioncapacities of moisture-equilibrated samples (moisture equilibra-tion at 97% relative humidity) to those of dry samples. Theyfound a 40% decrease in sorption capacity in the moisture-equilibrated samples.This paper summarizes the results from a study on methane

sorption capacity for selected black shales from The Nether-lands that are being considered as targets for shale gasexploration. Excess sorption isotherms at 65 °C and up to 25MPa were measured on five samples of Mesozoic age and onePaleozoic sample in the dry state. The samples contain variousamounts of type II organic matter and cover a representativerange of mineralogical compositions. The main goals of thisstudy were (1) to assess the maximum sorption capacity ofthese shales for methane, (2) to study the role of organic andmineral constituents and (3) thermal maturity on sorptioncapacity, and (4) to validate the applicability of Langmuirisotherms in representing the experimental high-pressuresorption data.

■ SAMPLESThe studied samples originate from the Roer Valley Graben,the Central Netherlands Basin and the West Netherlands Basin(Figure 1). All but one sample were collected from corematerial. Information on well name, depth, and stratigraphicunit is given in Table 1. The Sleen 2 sample comes from theWinterswijk quarry in the eastern part of The Netherlands.Five samples were taken from Mesozoic shale formations

two from the Sleen (Sleen 1, 2) and the Aalburg (Aalburg 1, 2),and one from the Posidonia shale formation. The Paleozoicsample is from the Geverik member (early-mid Namurian age).For detailed information on the stratigraphy, geology, andpetroleum potential of these shales, the reader is referred to ref11.

■ METHODSTOC and Rock-Eval Analysis. Total organic carbon (TOC) and

Rock-Eval analysis using a Rock-Eval 6 (Vinci Technologies)

instrument was carried out at Energie Beheer Nederland B.V.(EBN). The Rock-Eval pyrolysis is an established method forcharacterizing the type and thermal maturity of organic matter insedimentary rocks, as well as their petroleum generation potential.13 Asample is subjected to programmed heating in an inert (He)atmosphere to determine the amount of volatile gas and liquidhydrocarbons (S1 peak) and the amount of nonvolatile hydrocarbonsand oxygen-containing organic compounds released during thermalcracking of the remaining organic matter in the rock (recorded as S2and S3 peaks, respectively). The temperature (Tmax), at which S2 ismaximum, is a measure of thermal maturity of the source rock. Thehydrogen index (HI = 100 × S2/TOC) and oxygen index (OI = 100 ×S3/TOC) are usually plotted in an HI vs OI diagram, from which thetype and the maturation stage of the organic matter can be inferred.

Additional TOC analysis on powdered samples using a LECO RC-412 multiphase carbon/hydrogen/moisture determinator was carriedout at the RWTH Aachen University and compared with the reportedTOC data from the Rock-Eval 6 analysis. The discrepancies betweenthe measured and reported TOC contents and the implications forusing sorption data based on TOC are discussed below. For the sakeof consistency, all sorption data are normalized to the TOC valuesmeasured on the LECO instrument.

X-ray Diffraction Analysis (XRD). Bulk mineralogical composi-tions were derived from the X-ray diffraction patterns measured onrandomly oriented powder preparates. Rock samples were crushedmanually in a mortar and milled subsequently with a McCroneMicronising mill for 15 min to ensure uniform crystallite sizes. Milling

Figure 1. Map showing major sedimentary basins in the Netherlandsand sampling locations. (Map modified with permission from ref 12.Copyright 2003, Netherlands Journal of Geosciences' Foundation.)

Table 1. Provenance of the Studied Samples and TheirLithostratigraphic Origina

sample well/outcrop depth (m) formation

Aalburg 1 AST-01 1609 ATALAalburg 2 HAG-02 2609 ATALSleen 1 HAG-02 2658 ATRTSleen 2 Winterswijk quarry ATRTGeverik WSK-01 4238 DCGEPosidonia LOZ-01 2491 ATPO

aATAL = Aalburg fm., ATRT = Sleen fm., ATPO = Posidonia shalefm., Geverik = Geverik member.

Energy & Fuels Article

dx.doi.org/10.1021/ef300405g | Energy Fuels 2012, 26, 4995−50044996

was done in ethanol to avoid dissolution of water-soluble componentsand strain damage to the samples. An internal standard (corundum, 20wt %) was added to improve the accuracy of the analysis. Mineralquantification was performed on diffraction patterns from randompowder preparates. Sample holders were prepared by means of a sidefilling method to minimize preferential orientation. The measurementswere done on a Huber MC9300 diffractometer using Co Kα-radiationproduced at 45 kV and 35 mA. During the measurement, the sample isilluminated through a fixed divergence slit (1.8 mm, 1.45°), a graphitemonochromator, and 58 mm, 0.3 mm spacing soller slits. Thediffracted beam is measured with a scintillation detector with countingtime of 20 s for each step of 0.02° 2θ. Diffractograms were recordedfrom 2° to 76° 2θ. Quantitative phase analysis was performed byRietveld refinement. BGMN software was used, with customized claymineral structure models.14 All reported mineral compositions relateto the crystalline content of the analyzed samples.High-Pressure Sorption Experiments. Sample Preparation.

The sorption measurements were performed on dry powderedsamples. The drying procedure consisted of drying overnight at 105°C under vacuum and additional drying was performed in situ in themeasuring cell of the sorption setup at 105 °C and vacuum for >4 h.This additional drying is necessary, as samples take up air moistureduring their placement into the measuring cell, which can influencetheir sorption capacities.Experimental Setup. Sorption measurements were performed on a

manometric setup at 65 °C and pressures of up to 25 MPa. The setup(Figure 2) consists of a stainless steel measuring cell (MC), two high-

pressure shut-off valves (V1 and V2) and a high-precision Tecsispressure transducer with an accuracy of 0.05% of the end value (25MPa). The components are connected by 1/16″ stainless steel tubing.The dead volume of the pressure transducer and the volume of thetubing between valves V1 and V2 are used as reference volume(“reference cell”, RC). A three-port valve (V3) switches between gassupply and a vacuum pump. Valves are operated by computer-controlled actuators. The cells, pressure transducer, and the valves areplaced in an air-bath, and temperature is kept constant within 0.1 °C.Temperature readings are taken from a high-precision Pt-100resistance temperature detector (RTD). The volumes of the referenceand the measuring cell (1.748 ± 0.002 cm3 and 11.35 ± 0.01 cm3,respectively) were determined by helium expansion using steelcylinders of known dimensions placed in the measuring cell as areference. The uncertainty of measured excess sorption increases withpressure and is estimated to be 0.02 mmol/g at the high-end pressurevalue of 25 MPa. The measuring procedure was described by Krooss etal. (2002).15 Sufficient time was allowed for gas to reach equilibrium inthe reference and in the measuring cell. For all samples, the gas uptakewas rapid, and no noticeable change in the pressure readings wasobserved after 15 min.Calculation of Excess Sorption. The sorption isotherms obtained

experimentally by any conventional technique of sorption measure-

ment are the so-called “excess” sorption isotherms. An overview of theconcept of excess sorption, also denoted Gibbs surface excess, is givenby Sircar (1999).16 The excess amount (expressed as excess mass,mads

excess normalized to sample weight) is calculated from the followingmass balance:

ρ= −m m p T V( , ) [g/g]gadsexcess

total void (1)

where mtotal denotes the total mass of methane transferred into themeasuring cell (MC) and Vvoid is the void volume of MC determinedby helium expansion at 65 °C. The underlying assumption is that theadsorption of He is negligible. He expansion tests at pressures up to 15MPa did not show any change in Vvoid with pressure (data are providedin the Supporting Information). The methane density ρg (p, T) as afunction of pressure and temperature was calculated using theequation of state by Setzman and Wagner.17

Parameterization of the Experimental Results. The measuredexcess sorption data were parametrized using an adapted Langmuirfunction, as described in Gensterblum et al. 2009:18

ρ

ρ

ρ

ρ=

+− = −

⎛⎝⎜⎜

⎞⎠⎟⎟

⎛⎝⎜⎜

⎞⎠⎟⎟

⎡⎣⎢

⎤⎦⎥

m mp

p p

p Tm

p T

gg

1( , )

1( , )

adsexcess

LL

g

adsadsabsolute g

ads

(2)

or expressed in molar quantities:

ρ

ρ

ρ

ρ=

+− = −

⎛⎝⎜⎜

⎞⎠⎟⎟

⎛⎝⎜⎜

⎞⎠⎟⎟

⎡⎣⎢

⎤⎦⎥

n np

p p

p Tn

p T1

( , )1

( , )

mmolg

adsexcess

LL

g

adsadsabsolute g

ads

(3)

where nadsexcess (mmol/g) denotes the excess sorbed amount of substance

at pressure p (MPa). pL (MPa) is the Langmuir pressure,corresponding to the pressure at which half of the (“monolayer”)sorption sites are occupied, mL (g/g) and nL (mmol/g) are themaximum Langmuir capacities (corresponding to the “Langmuirvolume”), denoting the amount adsorbed at full occupancy of the“Langmuir monolayer”. The second term on the right hand side(RHS) of eqs 2 and 3 expresses the relationship between the absoluteand the excess sorption as a function of a ratio of the bulk andadsorbed phase density (ρads).

19−21 The parameters nL (or mL), pL, andρads are fitted to the experimental excess sorption data using a least-squares minimization procedure.

■ RESULTSTOC and Rock-Eval. The results of the TOC and Rock-

Eval analysis are listed in Table 2. Based on the measured TOCvalues, two samples (Aalburg 1, Sleen 1) can be regarded asorganic-lean shales, having TOC values of less than 1.2%. Therest of the samples are organic- rich shales with TOC in therange from 5% to 10%. The Rock-Eval results were plotted in apseudo-van-Krevelen (HI vs OI), and a HI vs Tmax diagram(Figure 3). The data indicate that the samples Aalburg 1 andSleen 1 are in the oil window, while samples Aalburg 2, Sleen 2,and Posidonia are immature based on their Tmax values.However, considering that the Aalburg 2 and the Sleen 1samples come from the same well within a depth interval of 50m, no significant difference in their thermal maturities is to beexpected. Furthermore, the samples Sleen 1 and Aalburg 1 ploton the maturation pathway for type III kerogen, whereas in theliterature, these shales have been reported to contain type IIkerogen.11 The interpretation of the Rock-Eval data iscomplicated by the very low TOC contents of the Aalburg 1and the Sleen 1 sample and a highly active mineral matrix (as

Figure 2. Flow scheme of the manometric high-pressure sorptionsetup. MC and RC refer to the measuring and the reference cell,respectively.



suggested by the high clay mineral content for the Aalburg andSleen samples, Table 3).22−24

The results for the Geverik sample show that the rock hasreached a postmature stage. This is also supported by the verylow hydrocarbon yields (S1 and S2 peaks) despite its highTOC content (6.1%). In the literature, the Geverik Membershale has been described as type II marine source rock.25

The Posidonia sample is the least mature within this samplesuite, evidenced by its low Tmax and high HI values. It has beencharacterized as organic-rich type II source rock with HI of upto 800 mg HC/g TOC.11

Bulk Mineralogy. The mineralogical composition of thestudied samples is listed in Table 3. The samples from the Sleenand Aalburg formations have mineralogical compositions verytypical of shales. Clay contents range between 60 and 71 wt %.The Geverik sample is more silty, having a quartz content of 36wt %. The Posidonia sample is a marl with a total carbonatecontent of 48 wt %.In all samples, the clay mineralogy is dominated by a long-

range ordered illite−smectite mixed layer phase (R3), identifiedfrom broad asymmetric 10 Å reflections in the XRD patterns.This type of mixed layer clay is very typical for diageneticallymature shales. In the diffraction patterns of samples Aalburg 1and Sleen 1, sharp 14 Å reflections of chlorite were identified.The pattern of the Sleen 2 sample displays a very broaddiffraction peak from about 16 to 12 Å. This wide d-spacingrange is typical for smectites formed in soils or meteoricweathering environments. The same sample contains an R3ordered mixed layer Illite-smectite clay. Both, the mixed layerand the soil smectite cannot have coexisted in the sample atburial depths greater than about 3 km. Therefore, the latter isconsidered to be an alteration product of the former, due toweathering at surface. Sample Sleen 2 is the only one amongthe studied samples that was taken from an outcrop.The effect of drying and exposure to atmosphere on the pore

structure of the studied core samples is considered to benegligible, since none of them contains significant amounts ofswelling clays. In all samples, traces of gypsum were detected.These are likely an alteration product of pyrite, due to exposureto the atmosphere and light. This alteration is not considered toaffect the sorption characteristics.

Methane Excess Sorption Isotherms. The measuredexcess sorption isotherms at 65 °C for all six samples are shownin Figure 4. The sorption capacities as well as the shapes ofisotherms vary among the samples. Except for one sample, allexcess isotherms have a maximum. This maximum occurs atpressures between 14 and 16 MPa for the clay-rich Aalburg andSleen samples and at a pressure of about 6 MPa for the Geveriksample. The excess isotherm for the Posidonia sample has nomaximum in the experimental pressure range. The maxima inexcess sorption and the corresponding pressures for eachsample are summarized in Table 4. The sorption capacitieswithin the 0−25 MPa pressure range show a rather continuousvariation from 0.05 mmol/g (Geverik) to 0.14 mmol/g (Sleen1), while the sorption capacity of 0.3 mmol/g of the outcropsample (Sleen 2) is considerably higher and distinct from theothers.

Reproducibility of Excess Sorption Measurements. Thereproducibility of excess sorption isotherm measurements wastested on two samples by duplicate runs. The results are shownin Figure 5 (Sleen 1) and Figure 6 (Posidonia) anddemonstrate a very good reproducibility. The slight offset ofthe first isotherm for the Posidonia sample is likely to be due toT

able

2.Summaryof

TOC,Rock-Eval,andXRD

Analyses

TOC

Rock-Evalanalysis

XRD

analysis

sample

measureda

(wt%)

reported

(wt%)

S1(m

g/g)

S2(m

g/g)

S3(m

g/g)

Tmax

(°C)

hydrogen

index

(mgHC/g

TOC)

oxygen

index

(mgC

O2/gT

OC)

productio

nindex

quartz

+feldspars

(wt%)

phyllosilicates

(wt%)

carbonates

(wt%)

Aalburg

11.02

1.15

0.31

1.33

0.65

437

116

570.19

2371

1Aalburg

25.04

4.38

2.04

21.61

0.92

428

493

210.09

1460

15Sleen1

0.82

0.45

0.04

0.17

0.7

443

38156

0.19

1963

12Sleen2

4.83

2.72

0.07

7.75

0.49

425

285

180.01

1767

6Geverik

6.10

7.47

0.18

0.11

0.58

nd1

80.62

4733

3Po

sidonia

10.49

7.91

4.4

54.1

1.52

426

684

190.08

934

49aMeasuredwith

LECO

RC-412

Analyzer.



removal of small amounts of residual moisture after theexposure to high-pressure gas and subsequent evacuation.Parameterization of Excess Sorption Data. The measured

excess sorption data were parametrized using a three-parameterLangmuir-based excess sorption function (eq 3). The maximumLangmuir sorption capacities (nL, in mmol/g), the Langmuirpressures (pL, in MPa), and the adsorbed phase densities (ρads,in kg/m3) are summarized in Table 4. In addition to the three-parameter fit, an alternative fitting procedure was used byassuming a fixed value of 423 kg/m3 for the adsorbed phasedensity. This value corresponds to the liquid phase density ofmethane at the normal boiling point.26 Figure 7 shows excesssorption isotherms and the Langmuir fits for both fittingscenarios. Although the three-parameter fit provides a more

accurate representation of the experimental data, reasonable fitsfor most samples were also obtained with ρads fixed. For theGeverik sample, which has the lowest sorption capacity and avery prominent maximum in excess isotherm, the two-parameter fitting method gave unsatisfactory results.

■ DISCUSSION

Effect of Organic and Inorganic Constituents on theMethane Sorption Capacity. The methane sorptioncapacities were compared to the total organic content (TOC)of the studied shales. It was found that the methane sorptioncapacities (normalized to sample weight) of the high-TOCsamples (Geverik and Posidonia) were lower than those of the

Figure 3. HI vs OI (left) and HI vs Tmax (right) plots showing maturation pathwaysfor kerogen types I, II and III. The data obtained by the Rock-Eval analysis are represented by full black circles. The size of the circles is proportional to TOC (in wt %).

Table 3. Bulk Mineralogical Composition of Studied Samples from XRD Analysis (in wt %)

sample

phase Aalburg 1 Aalburg 2 Sleen 1 Sleen 2 Geverik Posidonia

quartz 17.1 12.5 15.4 13.2 36.3 8.0orthoclase 0.6 0.0 1.4 2.5 0.7 0.6albite 5.0 1.9 2.4 2.0 9.6 0.5illite-smectite 60.2 52.2 53.8 65.1 33.4 27.9kaolinite 4.7 5.0 9.6 2.1 0.0 6.3chorite 6.5 2.3 0.0 0.0 0.0 0.0micas 2.3 1.7 2.8 2.9 3.6 1.6calcite 0.6 14.1 0.0 0.2 2.9 46.1dolomite 0.3 1.4 0.0 5.7 0.0 2.6ankerite 0.0 0.0 9.7 0.0 0.0 0.0siderite 0.0 0.0 2.8 0.0 0.0 0.0gypsum 0.0 1.2 0.0 1.5 1.4 2.7anhydrite 0.9 1.6 1.8 0.0 0.4 0.0pyrite 0.5 5.5 0.1 4.2 11.1 3.3anatase 0.6 0.0 0.0 0.5 0.0 0.0apatite 0.6 0.6 0.2 0.1 0.6 0.4



low-TOC samples. Figure 8 shows a plot of the maximumLangmuir capacity versus TOC (in wt %). Clearly, nocorrelation exists between the TOC content and the sorptioncapacity for these samples. This indicates that the sorptioncapacity of these shales is not controlled by TOC contentalone. Clay minerals can provide additional surface area for theadsorption of the gas molecules and, depending on theirtexture, will contribute the overall sorption capacity. Thisnotion is supported by our observation that the sorptioncapacity for methane increases with the total clay content(Figure 9). The prominent sorption capacity of the outcropsample (Sleen 2) can be explained by the presence of smectiteas a weathering product of mixed layer illite−smectite (I−S).Such an alteration of I−S to smectite at near surface conditionshave been reported to result in an increase in specific surfacearea.27 An important practical implication for shale gasexploration is that laboratory sorption data determined onsamples collected at outcrops or coming from shallowexploration wells may be misleading and should be used withcaution.It is interesting to note that the methane sorption capacities

measured on dry samples show a very good correlation with thehygroscopic moisture content determined initially on the “asreceived” samples (Figure 10). The moisture content wasobtained from the weight difference before and after drying.This implies that methane and water molecules share the samesorption sites and in the presence of both, competitiveadsorption will take place.21

Effect of Thermal Maturity on the Methane SorptionCapacity. Thermal maturity has been reported to affect thesorption capacity of shales due to textural changes in organicmatter. An increase in sorption capacity with maturity observedby Ross and Bustin4,7 was attributed to the creation of organicmatter-hosted microporosity. We compared the sorptioncapacities of the immature Posidonia and the overmature

Figure 4. Excess sorption isotherms for methane (mmol/g vs MPa)measured on dry samples at 65 °C.

Table 4. Maxima in Excess Sorption Isotherms and Langmuir Fitting Parameters Obtained by Three-Parameter Fit and by Two-Parameter Fit Assuming Liquid Methane Density (423 kg/m3) for the Adsorbed Phase

maximum in excesssorption andcorresponding

pressure Langmuir parameters for three-parameter fit Langmuir parameters with ρads fixed

sampleTOC (meas.)

(wt %) (mmol/g)P

(MPa)nL

(mmol/g)pL

(MPa)ρads

(kg/m3) Δnno. of datapoints

nL(mmol/g)

pL[MPa]

ρads(kg/m3)

Δn

Aalburg 1 1.02 0.11 ∼16 0.31 15.6 327 4.8 × 10−5 61 0.24 10.5 423 2.2 × 10−4

Aalburg 2 5.04 0.08 ∼16 0.22 13.2 295 9.1 × 10−5 44 0.13 5.8 423 5.2 × 10−4

Sleen 1 0.82 0.14 ∼15 0.33 10.3 312 8.5 × 10−5 32 0.24 6.3 423 5.9 × 10−4

Sleen 2 4.83 0.31 ∼15 0.72 9.6 332 2.7 × 10−4 27 0.56 6.5 423 1.3 × 10−3

Geverik 6.10 0.05 ∼7 0.10 3.0 195 2.1 × 10−4 49 0.06 0.6 423 1.1 × 10−3

Posidonia 10.49 0.11 a 0.20 10.1 (608) 9.4 × 10−5 31 0.26 14.5 423 2.2 × 10−4

ano maximum observed in the pressure range 0−25 MPa.

Figure 5. Reproducibility of the excess isotherms (65 °C) for theSleen 1 sample.

Figure 6. Reproducibility of the excess isotherms (65 °C) for thePosidonia sample.



Geverik sample (both high in TOC and with equal total claycontents) and found no enhancement of sorption capacity pergram rock or gram TOC for the overmature sample. However,a distinct difference was observed in the shape of the excessisotherms. The excess sorption isotherm of the Geverik samplehad a pronounced maximum at around 6 MPa, whereas theisotherm of the Posidonia sample had no maximum in themeasured pressure range. The Aalburg and the Sleen samples,

for which the maturity implied from the Rock-Eval data ishigher than for the Posidonia sample, all have maxima in themethane excess sorption isotherms between 14 and 16 MPa.The occurrence of maxima in excess isotherm has been

reported in the literature for different gases at supercriticalconditions.20,28,29 Whether a maximum in excess isotherm for aparticular gas occurs depends on (1) the proximity of themeasuring temperature to the critical temperature of the gas,

Figure 7. Excess sorption isotherms with fitted three-parameter sorption function (eq 3) and two-parameter fits assuming liquid methane density(423 kg/m3) for the adsorbed phase.



(2) the pressure range at which sorption is measured, (3) thedefinition of the void volume, and (4) the pore size and poresize distribution of the porous material.30,31

The different shapes of the isotherms in this study reflectdifferent pore characters of the shales. Rother et al. (2012)31

showed that the maximum in excess sorption for CO2 on silicaoccurred at lower pressures for smaller pores. Hence, it isconceivable, by comparing the isotherms of the Geverik and thePosidonia samples, that the overmature Geverik samplecontains a larger proportion of micropores in the organicmatter. However, in case of the clay-rich samples and, inparticular, for the low-TOC Aalburg 1 and Sleen 1 samples, theshape of the isotherms is most likely to result from the porecharacter of the clay minerals. These observations show that theoverall sorptive potential and the shape of excess sorptionisotherm of shales is a rather complex function of organic andinorganic composition and maturity (and hence, the pore-structural character) of shales. Further investigations involving amultivariate statistical approach on a larger data set arenecessary to elucidate the role of each of these variables inthe gas sorption process in shales.

Uncertainties Due to Normalization of Sorption Datato TOC. For the applications in reservoir simulation programs,sorption data are commonly reported along with TOCcontents, or normalized to TOC. Here, we demonstrate thatthis approach can result in significant errors in the gas contentestimates due to the uncertainties in measured TOC values.The comparison of the measured and the reported TOC data(Table 2) shows discrepancies of up to 40%. These areattributed to differences in analytical procedures, samplepreparation (e.g., acid treatment), or sample homogenizationissues. The effect of these discrepancies on the calculated TOC-normalized sorption data is demonstrated in Figure 11. The

horizontal error bars show the differences between themeasured and the reported TOC values for individual samples.The vertical error bars show the difference between the TOC-normalized maximum Langmuir capacities (nL) based on thetwo TOC data sets. It is evident that uncertainties in TOC canlead to significant errors in calculated sorbed gas contents forlow-TOC samples, no matter how accurate the sorption dataare.

Representation of the Excess Sorption Data and theDensity of Adsorbed Phase. The observation of maxima in

Figure 8. Maximum Langmuir sorption capacity (nL, mmol/g) fromboth fitting scenarios vs TOC content (wt %).

Figure 9. Relationship between maximum Langmuir sorption capacity(nL, mmol/g), TOC content (wt %) and clay content (wt %).

Figure 10. Relationship between maximum Langmuir capacity (nL,mmol/g) and the hygroscopic moisture content (in wt %) of the as-received samples.

Figure 11. Uncertainties resulting from normalization of the sorptiondata to TOC content. Horizontal error bars correspond to thedifference between the measured and reported TOC contents. Verticalerror bars indicate the difference between TOC-normalized maximumLangmuir capacities (nL, mmol/g TOC) based on the two TOC datasets.



methane excess isotherms of shales at high pressures hasimportant implications for the choice of parametrization of theexperimental data. All commonly used empirical models basedon either “monomolecular layer” (e.g., Langmuir) or “pore-filling” (e.g., Dubinin−Radushkevich) mechanisms rely on thetransformation from excess sorption to absolute sorption fordescribing high-pressure sorption data at supercritical con-ditions. Such transformations require assumptions about eitherthe density or the volume of the adsorbed phase.32 In thisstudy, we used an adapted three-parameter Langmuir fittingfunction, which takes explicitly into account the density ofadsorbed phase (ρads). Representative fits were obtained thatreproduce also the maxima in the excess sorption isotherms(Figure 7). As a first-order estimation, fitting parameters nL, pL,and ρads were allowed to vary to obtain the best-fit to themeasured data. Adsorbed phase densities thus obtained were inthe range of 300−330 kg/m3 for the clay-rich samples, androughly 200 kg/m3 and 600 kg/m3 for the Geverik and thePosidonia samples, respectively. In order to ascribe a physicalmeaning to ρads, various estimates of the adsorbed density wereproposed in the literature, for example, liquid phase density(423 kg/m3), “van der Waals density” (373 kg/m3) or criticaldensity (see ref 32 and the references therein). We tested thefitting performance by assuming a fixed value for ρadsrepresenting the liquid phase density and obtained reasonablefits except for the Geverik sample (Figure 7). The high value of600 kg/m3 by best-fit for the Posidonia sample is a fittingartifact due to the fact that its excess isotherm has no maximumin the studied pressure range. Hence, an estimation of theadsorbed phase density from the declining part of the isotherm(as explained in ref 28) is not possible. It should be also noted,however, that using ρads = 423 kg/m3 provided an equally goodfit as using the best-fit approach. On the other hand, the lowvalue of the best-fit ρads and the unsatisfactory fit with ρads =423 kg/m3 for the Geverik sample indicate that ρads obtainedfor this sample should be used at best as a fitting parameter.Moreover, accurate estimates of ρads would have to consider theeffect of helium adsorption and differences in position of theGibbs interface for helium and methane to define the“accessible” void volume for methane.30,33,34

■ CONCLUSIONSHigh-pressure sorption isotherms for methane were measuredon six representative shales considered as targets for shale gasexploration in The Netherlands. The various TOC contentsand mineralogical compositions, as well as a range of thermalmaturities, enabled us to investigate the effect of theseparameters on the sorption capacity. The conclusions fromthis study can be summarized as follows:

• The maximum excess sorption capacities within thepressure range 0−25 MPa, measured at 65 °C on drysamples range from 0.05 to 0.3 mmol/g (1.1−6.8 m3

STP/t).• No correlation of the sorption capacity with TOC

content was found.• Clay minerals can contribute significantly to the overall

sorption capacity of shales. Moreover, weathering of clayminerals at surface conditions may render samplescollected from outcrops or shallow wells unrepresenta-tive for shale gas exploration.

• Maxima in the excess isotherms were observed for all butthe least-mature sample. The shape of the excess

isotherm is controlled by the thermal maturity and claymineralogy.

• A three-parameter Langmuir-type function that takesinto account the density of the sorbed phase was foundto represent adequately the measured excess sorptiondata.

• Significant uncertainties in TOC-normalized excesssorption data may result from uncertainties in measuredTOC contents, especially for low-TOC shales.

The data and observations made in this study contribute tothe still insufficient amount of published data on gas sorption inshales. The quality of sorption data has to be assured byinterlaboratory round robin studies. For applications in shalegas exploration, the experimental conditions for sorptionmeasurements have to be extended to pressures and temper-atures representative of reservoir conditions. Moreover, theinvestigation of the role of moisture in gas storage capacity ofshales is of great importance. These topics are the subject ofour ongoing research.

■ UNITS CONVERSIONThe conversion of sorption capacities from mmol/g to m3

STP/t follows the definition of standard temperature (273.15 Kor 0 °C) and standard pressure (105 Pa) − STP by IUPAC(http://goldbook.iupac.org/S06036.html). The conversion fac-tor for methane used here is as follows:

=1 mmol/g 22.71 (m STP)/t3

■ ASSOCIATED CONTENT*S Supporting InformationAll measured sorption data for methane, including the data forvoid volume determination with helium and the X-raydiffraction patterns from the XRD analysis. This material isavailable free of charge via the Internet at http://pubs.acs.org.

■ AUTHOR INFORMATIONCorresponding Author*Telephone: +49-214-80-982-91. E-mail: [email protected] authors declare no competing financial interest.

■ ACKNOWLEDGMENTSEnergie Beheer Nederland B.V. (EBN), Utrecht, is acknowl-edged for providing the sample material as well as TOC, Rock-Eval data. Dr. Karl-Heinz Wolf of T. U. Delft and fouranonymous reviewers are acknowledged for providing criticalcomments and helping to improve the manuscript.

■ REFERENCES(1) Montgomery, S. L.; Jarvie, D. M.; Bowker, K. A.; Pallastro, R. M.AAPG Bull. 2005, 89, 155−175.(2) Jarvie, D. M.; Hill, R. J.; Ruble, T. E.; Pollastro, R. M. AAPG Bull.2007, 91, 475−499.(3) Chalmers, G. R. L.; Bustin, R. M. Bull. Can. Pet. Geol. 2008, 56(1), 22−61.(4) Ross, D. J. K.; Bustin, R. M. AAPG Bull. 2008, 92, 87−125.(5) Chalmers, G. R. L.; Bustin, R. M. Int. J. Coal Geol. 2007, 70, 223−239.(6) Chalmers, G. R. L.; Bustin, R. M. Bull. Can. Pet. Geol. 2008, 56(1), 1−21.(7) Ross, D. J. K.; Bustin, R. M. Mar. Pet. Geol. 2009, 26, 916−927.



http://goldbook.iupac.org/S06036.html

http://pubs.acs.org

mailto:[email protected]

mailto:[email protected]

(8) Weniger, P.; Kalkreuth, W.; Busch, A.; Krooss, B. M. Int. J. CoalGeol. 2010, 84, 190−205.(9) Schettler, P. D.; Parmely, C. R.; Juniata, C. SPE J. 1991, SPE-23422.(10) Lu, X.-C.; Li, F.-C.; Watson, A. T. SPE Form. Eval. 1995, SPE-26632.(11) Geology of the Netherlands; Wong, Th. E., Batjes, D. A. J., DeJager, J., Eds.; Royal Netherlands Academy of Arts and Sciences:Amsterdam, 2007.(12) De Jager, J. Neth. J. Geosci. 2003, 82, 339−349.(13) Espitalie, J.; Laporte, J. L.; Madec, M.; Marquis, F.; Leplat, P.;Paulet, J.; Boutefeu, A. Rev. Inst. Fr. Pet. 1977, 32, 23−42.(14) Ufer, K.; Stanjek, H.; Roth, G.; Dohrmann, R.; Kleeberg, R.;Kaufhold, S. Clays Clay Miner. 2008, 56, 272−282.(15) Krooss, B. M.; van Bergen, F.; Gensterblum, Y.; Siemons, N.;Pagnier, H. J. M.; David, P. Int. J. Coal Geol. 2002, 51, 69−92.(16) Sircar, S. Ind. Eng. Chem. Res. 1999, 38, 3670−3682.(17) Setzmann, U.; Wagner, W.; Pruss, A. J. Phys. Chem. Ref. Data1991, 20 (6), 1061−1151.(18) Gensterblum, Y.; van Hemert, P.; Billemont, P.; Busch, A.;Charriere, D.; Li, D.; Krooss, B.; de Weireld, G.; Prinz, D.; Wolf, K.-H.Carbon 2009, 47, 2958−2969.(19) Yee, D.; Seidle, J. P.; Hanson, W. B. Gas sorption on coal andmeasurement of gas content. In Hydrocarbons from Coal; Law, B. E.,Rice, D. D., Eds.; AAPG Studies in Geology 38, AAPG: Tulsa, OK,1993; pp 203−218.(20) Richard, M. A.; Benard, P.; Chahine, R. Adsorption 2009, 15,43−51.(21) Busch, A.; Gensterblum, Y. Int. J. Coal Geol. 2011, 87, 49−71.(22) Espitalie, J.; Madec, M.; Tissot, B. AAPG Bull. 1980, 64, 59−66.(23) Katz, B. J. Org. Geochem. 1983, 4, 195−199.(24) Peters, K. E. AAPG Bull. 1986, 70, 318−329.(25) Van Balen, R. T.; Van Bergen, F.; De Leeuw, C. S.; Pagnier, H. J.M.; Simmelink, H. J.; Van Wees, J. D. A. M.; Verweij, J. M. Geol.Mijnbouw 2000, 79, 29−44.(26) Poling, B. E.; Prausnitz, J. M.; O’Connell, J. P. PVT relationshipsof pure gases and liquids. In The Properties of Gases and Liquids;McGraw-Hill, 2001; Vol. 5, pp 4.36.(27) Sucha, V.; Srodon , J.; Clauer, N.; Elsass, F.; Eberl, D. D.; Kraus,I.; Madejova, J. Clay Miner. 2001, 36, 403−419.(28) Menon, P. G. Chem. Rev. 1968, 68, 277−294.(29) Ustinov, E. A.; Do, D. D.; Herbst, A.; Staudt, R.; Harting, P. J.Colloid Interface Sci. 2002, 250, 49−62.(30) Do, D. D.; Do, H. D. Carbon 2003, 41, 1777−1791.(31) Rother, G.; Krukowski, E. G.; Wallacher, D.; Grimm, N.;Bodnar, R. J.; Cole, D. R. J. Phys. Chem. 2012, 116, 917−922.(32) Murata, K.; El-Merraoui, M.; Kaneko, K. J. Chem. Phys. 2001,114 (9), 4196−4205.(33) Gumma, S.; Talu, O. Adsorption 2003, 9, 17−28.(34) Herrera, L.; Fan, Ch.; Do, D. D.; Nicholson, D. Adsorption 2011,17, 955−965.



Documents

High-Pressure Methane Sorption Isotherms of Black Shales from The Netherlands