MRI Visualization of SpontaneousMethane Production From Hydrates inSandstone Core Plugs When Exposed
to CO2A. Graue and B. Kvamme, U. of Bergen; B.A. Baldwin, Green Country Petrophysics LLC; J. Stevens and
J. Howard, ConocoPhillips; E. Aspenes, G. Ersland, and J. Husebø, U. of Bergen; and D. Zornes, ConocoPhillips
SummaryMagnetic resonance imaging (MRI) of core samples in laboratoryexperiments showed that CO2 storage in gas hydrates formed inporous rock resulted in the spontaneous production of methanewith no associated water production. The exposure of methanehydrate in the pores to liquid CO2 resulted in methane productionfrom the hydrate that suggested the exchange of methane mol-ecules with CO2 molecules within the hydrate without the additionor subtraction of significant amounts of heat. Thermodynamicsimulations based on Phase Field Theory were in agreement withthese results and predicted similar methane production rates thatwere observed in several experiments. MRI-based 3D visualiza-tions of the formation of hydrates in the porous rock and themethane production improved the interpretation of the experi-ments. The sequestration of an important greenhouse gas whilesimultaneously producing the freed natural gas offers access to thesignificant amounts of energy bound in natural gas hydrates andalso offers an attractive potential for CO2 storage. The potentialdanger associated with catastrophic dissociation of hydrate struc-tures in nature and the corresponding collapse of geological for-mations is reduced because of the increased thermodynamic sta-bility of the CO2 hydrate relative to the natural gas hydrate.
IntroductionThe replacement of methane in natural gas hydrates with CO2
presents an attractive scenario of providing a source of abundantnatural gas while establishing a thermodynamically more stablehydrate accumulation. Natural gas hydrates represent an enormouspotential energy source as the total energy corresponding to naturalgas entrapped in hydrate reservoirs is estimated to be more thantwice the energy of all known energy sources of coal, oil, and gas(Sloan 2003). Thermodynamic stability of the hydrate is sensitiveto local temperature and pressure, but all components in the hy-drate have to be in equilibrium with the surroundings if the hydrateis to be thermodynamically stable. Natural gas hydrate accumula-tions are therefore rarely in a state of complete stability in a strictthermodynamic sense. Typically, the hydrate associated with fine-grain sediments is trapped between low-permeability layers thatkeep the system in a state of very slow dynamics. One concern ofhydrate dissociation, especially near the surface of either subma-rine or permafrost-associated deposits, is the potential for the re-lease of methane to the water column or atmosphere. Methanerepresents an environmental concern because it is a more aggres-sive (∼25 times) greenhouse gas than CO2. A more serious concernis related to the stability of these hydrate formations and its impacton the surrounding sediments. Changes in local conditions of tem-perature, pressure, or surrounding fluids can change the dynamicsof the system and lead to catastrophic dissociation of the hydrates
and consequent sediment instability. The Storegga mudslide inoffshore Norway was created by several catastrophic hydrate dis-sociations. The largest of these was estimated to have occurred7,000 years ago and was believed to have created a massive tsu-nami (Dawson et al. 1988). The replacement of natural gas hydratewith CO2 hydrate has the potential to increase the stability ofhydrate-saturated sediments under near-surface conditions. Hydro-carbon exploitation in hydrate-bearing regions has the additionalchallenge to drilling operations of controlling heat production fromdrilling and its potential risk of local hydrate dissociation (Yaku-shev and Collett 1992).
The molar volume of hydrate is 25–30% greater than the vol-ume of liquid water under the same temperature-pressure condi-tions. Any production scenario for natural gas hydrate that in-volves significant dissociation of the hydrate (e.g., pressure deple-tion) has to account for the release of significant amounts of waterthat in turn affects the local mechanical stress on the reservoirformation. In the worst case, this would lead to local collapse ofthe surrounding formation. Natural gas production by CO2 ex-change and sequestration benefits from the observation that thereis little or no associated liquid water production during this pro-cess. Production of gas by hydrate dissociation can produce largevolumes of associated water, and can create a significant environ-mental problem that would severely limit the economic potential.
The conversion from methane hydrate to a CO2 hydrate isthermodynamically favorable in terms of free energy differences,and the phase transition is coupled to corresponding processes ofmass and heat transport. The essential question is then if it ispossible to actually convert methane hydrate as found in sedimentsto CO2 hydrate. Experiments that formed natural gas hydrates inporous sandstone core plugs used MRI to monitor the dynamics ofhydrate formation and reformation. The paper emphasizes the ex-perimental procedures developed to form the initial natural gashydrates in sandstone pores and the subsequent exchange withCO2 while monitoring the dynamic process with 3D imaging on asub millimetre scale. The in-situ imaging illustrates the productionof methane from methane hydrate when exposed to liquid CO2
without any external heating.
Hydrate Formation, Dissociation,and ReformationNatural gas hydrate was formed under controlled laboratory con-ditions in porous rock while being monitored with MRI. Twooutcrop Bentheim sandstone core plugs, 9.4 cm long and 3.8 cmdiameter, were used as reservoir rock in this study. Grain densitywas 2.65 g/cm3. Porosity and fluid permeability of the core plugswere measured at 22% and 1.1 Darcy, respectively, for both plugs.The mineralogy for this sandstone is 99% quartz, with traceamounts of the clay mineral kaolinite. The sandstone was charac-terized by uniform pore geometry, with an average pore diameterof 125 microns.
The first experiment was designed to study hydrate growthpatterns in a whole core-plug and to compare MRI intensity signalsto pump volume data of consumed methane as a function of timeduring hydrate formation to evaluate reaction kinetics. The coreplug was initially placed horizontally in a coreholder in the MRI
This paper (SPE 118851) was prepared for presentation at the 2006 Offshore TechnologyConference (paper OTC 18087), Houston, 1–4 May, and revised for publication. Originalmanuscript received for review 7 February 2006. Revised manuscript received 31 May2007. Paper peer approved 7 November 2007.
and vacuum-evacuated. The coreholder was made of compositematerial to minimize influence on the imaging capabilities. A cool-ing system was constructed capable of holding the core at a se-lected, stable temperature within the range of 0–200°C, with anaccuracy of 0.1°C, for an extended period of time. The coolingfluid, Fluorinert (FC-84), contains no hydrogen and therefore isinvisible to MRI, and is also used as the confining fluid for pres-sure maintenance in the coreholder. To keep the volume of theconfining fluid small and the cooling bath and pumps at a safedistance from the magnet, a heat exchange system was used. Thetubing which circulated the high-pressure confining fluid was em-bedded in larger PVC tubing filled with the temperature-controlledcirculating propolene glycol, which was not piped through theMRI. A sketch of the experimental apparatus is shown in Fig. 1.Introduction of fluids to the core-plug was performed by pumpsconnected to inlet and outlet ports. The core plug was pressurizedto 81.6 bars with methane followed by brine injection from oneend. The waterflooded core reached a uniform saturation of 50%water and 50% methane. The system was maintained at 81.6 barsby controlling the methane pump. Injected/consumed methane asfunction of time was continuously recorded; any consumption ofmethane during hydrate formation was monitored by volumechanges. High-spatial-resolution fluid saturation imaging (0.86mm×1.88 mm×1.88 mm voxel size) was applied while loweringthe temperature from 23 to 3°C. The cooling (and heating) wasperformed by constant rate of confinement-fluid circulation, andlasted for a period of approximately 5 hours. The MRI signaloriginates from hydrogen in liquid water. When water moleculescrystallize (freeze) into a solid hydrate phase with methanetrapped, the hydrogen no longer provides measurable magneticresonance signals because of very fast relaxation. Dynamic 3Dimaging of hydrate formation is therefore obtained by observingthe lack of signal from the water phase.
After heating and cooling the system several times, the forma-tion and dissociation of hydrates were monitored as function oftime. Hydrate formation occurred as a frontal advancement fromthe methane injection end (Fig. 2). The first image (a) in the figureis obtained before cooling and reveals higher MRI intensities,identified as a more redish color. The yellow color representslower MRI intensities (i.e., lower water signal) which suggest that
hydrate formation occurred both as frontal movement and, to someextent, throughout the whole core. The dominating front growthmay also be caused by salinity effects. The result of increased saltconcentration in the unhydrated water may prevent the completeformation of hydrate (Makogan 1997) and form a high-salinitywater bank adjacent to the hydrate front. Similar experiments, witha more uniform water distribution, and lower salinities, showed nofront growth. Hydrate formation started after an induction periodbetween 0 and 30 hours after cooling. Experiments were carriedout at two different temperatures, first at 3°C, then again at 5°C, tobe certain that no ice was formed in the aqueous phase. As inearlier MRI experiments of bulk hydrate (Moudrakovski et al.2002; Baldwin et al. 2003), MRI was found to provide excellentresolution between the hydrate and its liquid/gas precursors andallowed the dynamic, spatial imaging of the formation and disso-ciation of hydrates in the porous sandstone under various injectionconditions. The decrease in MRI intensity during hydrate forma-tion corresponded very well to kinetic data of consumed methane.
Methane Production by CO2 SequestrationThe second core plug was cut along the long axis, into two half-cylinders, in order to establish an open fracture. When assembledin the core holder, a polyethylene spacer, 4 mm wide, with twoopen and connected compartments was inserted in between the twohalf-cores. The high-molecular-weight polyethylene spacer wasmachined with channels in the vertical support to allow easy trans-port of water, methane, and CO2 through the frame. The spacer(Fig. 3) allowed both an increased surface area for methane toenter the core during hydrate formation, and created a volume tocollect methane for MRI measurement during the subsequent car-bon dioxide exchange process. The core was placed horizontally inthe MRI and saturated with pressurized methane. Water was al-lowed to imbibe into the matrix through the fracture until the watersaturation reached a predetermined value. The system was cooledto 4°C to allow for hydrate formation.
At the first timestep shown in Fig. 4, the core exhibits thesaturation distribution obtained after the spontaneous brine imbi-bition was terminated. The predominant signal source is water inthe matrix. Green color represents high MRI intensities, while bluecolor represents lower signal intensities. Methane in centre spacer
Fig. 1—Experimental apparatus.
147June 2008 SPE Journal
and the far side core face spacer reveal blue color owing to lowerdensity (less hydrogen). A limited, predetermined amount of waterwas injected into the fracture, which then imbibed into the twohalf-cores resulting in a 50% water saturation. Trapped methane inthe invaded pore space and free methane in the upper left part ofthe core halves were supported by available free methane from thefracture. Because the MRI signal responds to the water and freemethane, an absence of signal indicates that the water has trans-formed into hydrates. At 50 hours, the hydrate formation ceased inthis experiment; it is seen that at the plug end containing thedispersed, or decreasing, water saturation, only partial hydrate for-mation is observed. Methane is not detected within the porous rockby MRI in this experiment. Bulk methane is observed in the frac-ture during the hydrate formation because a methane source wasreadily available at the upstream end of the core plug. At the finaltimestep in Fig. 4, it is seen that the major part of the water hasbeen transformed into hydrate. The temperature was held constant
at 3.6°C (i.e., above freezing temperature for water to avoid iceformation).
When hydrate formation ceased, CO2 was injected into thefracture and allowed to interact with the hydrate. The subsequentkinetic interaction with the hydrate is shown in MRI images col-lected over time (Fig. 5). As can be seen from this figure, methanespontaneously was produced into the fracture. This was verifiedfrom GC analysis of the gases collected at the outlet sample port.As a consequence, the thermodynamic driving force for conversionof the hydrate is reduced correspondingly. A profile of the MRIintensity generated from the methane concentration along thelength of the fracture as a function of time gave in-situ informationon the amount of methane produced. The noticeable difference inthe change of the MRI intensity along the length of the fracturesuggested that the methane production continued over 4 days.When no increase in methane concentration could be observed, asecond CO2 flush of the fracture was performed. This procedurewas repeated 2–4 times to allow for additional methane produc-tion. GC samples were analyzed for every effluent sample. Fig. 6exhibits the in-situ methane production, measured by MRI, asfunction of time between each CO2 flush. It is noticeable from Fig.6 that the methane production rate for each of the three productionswas initially fairly high, then slowly decreased with time. Theproduction of methane ceased earlier for each subsequent CO2
flush. The latter may indicate that the amount of methane hydratecontacted by CO2 was reduced after each CO2 flush, while theformer is, as indicated previously, related to changes in chemicalpotential of CO2 and the kinetic effects of reduced thermodynamicdriving force on the methane production. Fig. 6 includes the ef-fluent methane production by material balance calculations fromthe three CO2 flushes. Total gas recovery after three CO2 flushesamounted to 50–85% of the gas originally in the hydrate, in ad-dition to the original free methane in the gas-filled pore space. Atthese experimental conditions, the density of pure methane is typi-cally in the gas-phase range. This is in contrast to pure CO2, whichis in the liquid region at the same conditions. The relative buoy-ancy of a separate CH4 phase relative to a separate CO2 phase isreduced because of mutual solubility of the two components,which for intermediate concentration ranges will lead to phaseseparation. Practically, this means that the escape of released CH4
may be limited by the kinetics of CH4 dissolution into the CO2
Fig. 3—Bentheim sandstone with polyethylene spacer with twoconnected compartments.
Fig. 2—Dynamics of hydrate formation in sandstone.
148 June 2008 SPE Journal
phase, and eventually a phase split, depending on the relativeamounts of CH4 and CO2 at the local conditions of temperatureand pressure. This implies that the production rate is determined inthis case both by the reduced available methane hydrate and alsoby reduced thermodynamic driving force caused by the dilutionand phase splitting of the CO2 phase in the spacer compartment. Ina real reservoir with channels/fractures, the escape of CH4 willthus be a complex function of the local conversion kinetics and thelocal flow situation. Natural hydrate structures vary significantly.A typical pattern, as for instance in the Nankai Trough, is that thehydrate layers are separated by low-permeability clay layers. Fro-zen hydrate layers can be considered as more or less a stationarystructure, which is typically not thermodynamically stable in arigorous sense but rather trapped into a situation of slow dynamics.In addition to the presence of fractures in the sediment structure, aswell as in the trapping layers above, the success of a natural CH4
replacement scenario depends sensitively on the type of sedimentand corresponding flow properties. These issues, as well as theextension and connectivity of the hydrate layers, are key factors indetermining the actual value of a hydrate reservoir as a storage sitewith corresponding natural-gas production.
Presently, there is no available experimental information on theconversion mechanism for the reformation of methane hydrateover to CO2 hydrate for temperatures above 0°C. One hypothesisfor the conversion mechanism is that liquid CO2 molecules ap-proaching the methane hydrate surface induce a microscopic dis-sociation of the outer hydrate layer. The released liquid-like mol-ecule then subsequently forms hydrate with CO2. On a macro-
scopic scale, the induction time for hydrate formation on theinterface between water and liquid CO2 is essentially zero, and assuch, no free liquid water phase is expected on this scale. Thematerial balances related to the conversion experiments presentedin this paper, however, indicate that very little CH4 enters theconverted hydrate. This might be taken as an indication of signif-icant changes of the hydrogen-bonded network, and a correspond-ing microscopic-liquid-like transition step that allows the methaneto escape from the hydrate at these conditions of temperature andpressure. Phase Field Theory (PFT) simulations also support thepresence of a microscopic liquid film. The general PFT approachis described elsewhere (Kvamme et al. 2004; Granasy et al. 2004)and the corresponding thermodynamic properties of liquid water,CO2 hydrate, and CH4 hydrates were predicted using the approachof Kvamme and Tanaka (1995). PFT is similar to density func-tional theory, which formulates the phase-transition kinetics asfunction of the change in molecular structure. The main differenceis that PFT uses directly the formal statistical mechanical relation-ship between molecular structures in the two phases and the cor-responding free energies of the phases. In a simplified fashion,PFT can be described as an approach for sampling paths of mini-mum free energy progress under the constraint of mass and heattransport. In Fig. 6, we plot the results of an isothermal phase fieldsimulation of CH4 hydrate being contacted with CO2 liquid at thesame conditions as the experimental conditions. The phase transi-tion appears to be mass-transport-controlled in this case, and adiffusivity coefficient of CO2 equal to 1.7�10−9 m2/s appears torepresent the experimental data well. Diffusivity coefficient for
Fig. 4—Hydrate formation in the fractured core plug.
149June 2008 SPE Journal
Fig. 5—Methane production from the hydrate into the CO2-filled fracture.
Fig. 6—Methane production as function of time.
150 June 2008 SPE Journal
CO2 in seawater is reported to between 1.0�10−9 m2/s (0°C) and1.9�10−9 m2/s (24°C) (Panhuis et al. 1988). The heat conductivityof the CO2 phase is low and it is therefore likely that most of thenet released heat will be transported towards the rock matrix andthe hydrate structure and as such assist in the conversion process.This will be true in general, whether or not the above hypothesisis correct, because added heat to the hydrate structure will weakenthe hydrogen-bonded structure and reduce the barrier for conver-sion over to the more stable CO2 hydrate correspondingly. Buteven if the net released heat may be significant on a microscopiclevel, at the dissociating front the total effects of the heat releaseon the dissociation kinetics is more uncertain. The kinetic rate ofheat transport is at least two orders of magnitude larger than thecorresponding rates of mass transport. Estimates of the heat ca-pacity and heat conductivities of the rock matrix, and the hydratestructure in the pores, indicate that most of the net released heat islikely to disappear from the conversion front rapidly (Svandal andKvamme 2005).
The hypotheses of CO2 inducing a microscopic liquid film onthe methane hydrate surface and then subsequently forming a CO2
hydrate film is expected to be one of the rate-limiting factors in theconversion process. Knowledge of the kinetic rate of hydrate for-mation on the water/hydrate former interface is therefore importantalso for the conversion process.
In this context, it is important to keep in mind that the resolu-tion of the MRI sampling will not be able to capture the nucleationstage. Predictions using isothermal PFT indicate a critical radius(assuming spherical critical nucleus) for CO2 hydrate formingfrom saturated solution at 274°K and 150 bar of 2.4 nm (Kvammeet al. 2004).
Two immediate questions arise. The first question is, of course,the time for onset of visible hydrate growth, and the second ques-tion concerns the growth rate.
To address these questions, an experiment was designed where-in hydrate formation on the interface between bulk water andmethane was monitored in an open fracture between two solidhalf-cylinders made of polyethylene, separated by the spacershown in Fig. 3. Mounted vertically, half of the available spacervolume was filled with water and the remaining volume on top ofthe water was filled with methane and pressurized to 81.6 bars.The temperature was reduced from 23 to 3.6°C in approximately 4hours. Hydrate that nucleates and grows heterogeneously on aninterface between fluid hydrate formers and water will form a thinhydrate film (Sloan 2003; Kvamme 2000), as also observed inexperiments on in-situ deposition of CO2 at large depths (Koba-yashi et al. 2001). According to state of the art theoretical models(Kvamme et al. 2004; Granasy et al. 2004), this film will establishwithin a nanosecond scale and then act as a transport sealing,which reduces the transport rate of new molecules toward thegrowing surfaces. Actual transport mechanisms and correspondingtransport rates through a hydrate film are not known to any rea-sonable scientific precision, but some studies indicate that there isat least three orders of magnitude slower transport through hydratethan corresponding transport of the same molecules in liquefiedwater (Demurov et al. 2002; Broecker and Peng 1982). The thinhydrate film will, however, not be uniform in a microscopic sense,and will vary in thickness proportional to the number of originalnucleation sites, the thermodynamic driving force (free energydifference), and transport of mass and heat. Resulting local varia-tions in thickness, geometry, and crystal morphology result indifferences in free energies along the growing hydrate surface.Under the constraints of limited access to fresh molecules that cansupport further hydrate growth, the free energy gradients willeventually lead to consumption of less stable regions in favor ofgrowth of neighboring more-stable zones. This effect is parallel towhat is frequently observed during growth of hydrate at interfacesbetween aqueous phase and hydrate former phase. For the methanewater system, the experiments reported by Makogan (1997) illus-trate that smaller cores growing toward the methane phase mayeventually dissolve in favor of further growth of larger cores.Eventually, the thin zones of hydrate film may break and massivehydrate growth through the film is expected. The magnitude of the
thermodynamic driving forces will govern the size and distributionof break through channels. Large thermodynamic driving forcesare expected to give smaller channels or more needle-type hydratesinto the aqueous phase, as observed experimentally by Makogan(1997). These processes are typically governed by nanoscale ef-fects, and only the subsequent results of massive break through arevisible on sampled responses from the MRI-signal in Fig. 7. Here,growth of methane hydrate can be identified as black channelsgrowing downward in the liquid water phase, appearing white inthe images. Similar experiments with CO2 and water at the sameconditions of temperature and pressure did not show any progresstowards massive growth penetration of the thin hydrate film within340 hours and the resolution scale of the MRI experiment (200microns). We should keep in mind that the structure (Kvamme andKuznetsova 2005) and corresponding chemical potential of waterclose to a solid mineral structure is likely to be different thanchemical potential of water in the hydrate if the system is not in astate of complete equilibrium. With two components and threephases, the Gibbs phase rule leaves only one degree of freedom,and with two independent variables, pressure and temperature, thesystem cannot reach a state of complete three-phase equilibriumand will rather be governed by the requirement of minimum freeenergy. The chemical potential of hydrate close to a mineral wallmay therefore very well be lower than that of water in hydrate.Liquid film channels separating the hydrate from the mineral sur-faces provide extra channels of fluid transport and may explain thelong time lag before onset of massive hydrate growth in theseexperiments to the more rapid progress of hydrate conversion inporous media. Electron microscope (EMS) studies of natural hy-drate structures (Techmer et al. 2005) also provide evidence ofliquid water films separating the hydrate phase from the mineralsurfaces. Relative to the conversion of methane hydrate over toCO2 hydrate, such liquid films provide transport channels for CO2
and methane as well as increased conversion interface.
Fig. 7—Methane hydrate growth in spacer.
151June 2008 SPE Journal
ConclusionsInjection of CO2 in natural gas hydrate reservoirs may offer stablelong-term storage of a greenhouse gas with the additional benefitof methane production without adding heat to the process and withno associated water production. The proposed sequestration pro-cess may offer large amounts of energy for the future.
CO2 hydrate is thermodynamically more stable than the origi-nal natural gas hydrate and, correspondingly, less sensitive to glob-al changes in temperature.
MRI specifically determined:• The dynamics of methane hydrate formation in sandstone.• Spontaneous conversion of methane hydrate to carbon dioxide
hydrate occurred when methane hydrate, in porous media, wasexposed to liquid carbon dioxide
• The conversion of methane hydrate into carbon dioxide hydratein sandstone takes place without adding heat to the system.
• No dissociation of hydrate to liquid water was observed withinthe scale of MRI resolution.
AcknowledgmentsSeveral of the authors are indebted for the support from the RoyalNorwegian Research Council. The authors acknowledge the per-mission from ConocoPhillips to publish this work.
ReferencesBaldwin, B.A., Moradi-Araghi, A., and Stevens, J.C. 2003. Monitoring
hydrate formation and dissociation in sandstone and bulk with mag-netic resonance imaging. Magnetic Resonance Imaging 21 (9): 1061–1069. DOI: 10.1016/j.mri.2003.05.001.
Broecker, W.S. and Peng, T.H. 1982. Tracers in the sea. Palisades, NewYork: Eldigio Press.
Dawson, A.G., Long. D., and Smith. D.E. 1988. The Storegga slides:Evidence from eastern Scotland for a possible tsunami. Marine Geol-ogy 82 (3–4): 271–276. DOI: 10.1016/0025-3227(88)90146-6.
Demurov, A., Radhakrishnan, R., and Trout, B.L. 2002. Computations ofdiffusivities in ice and CO2 clathrate hydrates via molecular dynamicsand Monte Carlo simulations. J. Chem. Phys. 116 (2): 702–709. DOI:10.1063/1.1425821.
Granasy, L., Pusztai, T., Tegze, G., Kuznetsova, T., and Kvamme, B. 2004.Phase field theory of hydrate nucleation: Formation of CO2 hydrate inaqueous solution. In Recent Advances in the Study of Gas Hydrates, ed.C. Taylor and J.T. Kiwan. New York City: Kluwer Academic/PlenumPublishers.
Kobayashi, I, Ito, Y., and Mori, Y.H. 2001. Microscopic observations ofclathrate-hydrate films formed at liquid/liquid interfaces. I. Morphol-ogy of hydrate films. Chem. Eng. Sci. 56 (14): 4331–4338. DOI:10.1016/S0009-2509(00)00544-3.
Kvamme, B. 2000. Phase transitions and stabilities in seawater containingcarbon dioxide. In Recent Advances in Marine Science and Technol-ogy, ed. N. Saxena, 279–289. Honolulu, Hawaii: Pacon International,.
Kvamme, B. and Kuznetsova, T. 2005. Investigation into stability andinterfacial properties of CO2 hydrate-aqueous fluid system. Interna-tional Conference of Computational Methods in Sciences and Engi-neering 2005, Loutraki, Greece, 21–26 October.
Kvamme, B. and Tanaka, H. 1995. Thermodynamic stability of hydratesfor ethane, ethylene, and carbon dioxide. J. Phys. Chem. 99 (18): 7114–7119. DOI: 10.1021/j100018a052.
Kvamme, B. et al. 2004. Kinetics of solid hydrate formation by carbondioxide: Phase field theory of hydrate nucleation and magnetic reso-nance imaging. Physical Chemistry Chemical Physics 6: 2327–2334.DOI: 10.1039/b311202k.
Makogan, Y.F. 1997. Hydrates of hydrocarbons. Tulsa: Pennwell Publish-ing.
Moudrakovski, I.L., Ratcliffe, C.I., and Ripmeister, J.A. 2002. Applicationof magnetic resonance microimaging (MRM) to monitor the formationof gas hydrate. Proc., Fourth International Conference on Gas Hy-drates, Yokohama, Japan, 19–23 May.
Panhuis, M.I.H., Patterson, C.H., and Lynden-Bell, R.M. 1988. A Molecu-lar Dyamics Study of Carbon Dioxide in Water: Diffusion, Structureand Thermodynamics. Molec. Phys. 94 (6): 963–972.
Sloan, E.D. Jr. 2003. Fundamental principles and applications of naturalgas hydrates. Nature 426: 353–359. DOI: 10.1038/nature02135.
Svandal, A. and Kvamme, B. 2005. The influence of diffusion on hydrategrowth. Proc., 1st International Conference on Diffusion in Solids andLiquids, University of Aveiro, Aveiro, Portugal, 6–8 July.
Techmer, K., Heinrichs, T., and. Kuhs, W.F. 2005. Cryo-electron micro-scopic studies on the structures and composition of Mallik gas-hydrate-bearing samples. In Scientific Results from the Mallik 2002 Gas Hy-drate Production Research Well Program, Mackenzie Delta, NorthwestTerritories, Canada, Bulletin 585, ed. S.R. Dallimore und T.S. Collett,Ottawa, Ontario, Canada: Geological Survey of Canada.
Yakushev, V.S. and Collett, T.S. 1992. Gas hydrates in arctic regions: Riskto drilling and production. Proc., 2nd International Offshore and PolarEngineering Conference, San Francisco, 14–19 June.
SI Metric Conversion Factors°F (°F–32)/1.8 � °Cin. × 2.54* E+00 � cm
*Conversion factor is exact.
Arne Graue is Professor of Reservoir Physics at the Departmentof Physics and Technology at the University of Bergen, Norway.His primary research interests are in fundamental studies ofmultiphase flow in porous media, emphasizing improved oilrecovery utilizing complementary in-situ imaging techniques.Bjørn Kvamme is Professor of Physics at the Department ofPhysics and Technology at the University of Bergen, Norway.His primary research interests are in thermodynamic modeling,simulations of molecular kinetics, hydrate technology, andCO2 sequestration. Bernard A. Baldwin retired from Phillips Pe-troleum Company in 2000. As president of Green CountryPetrophysics, LLC, he works part-time for ConocoPhillips in theMRI lab. His current efforts primarily involve measuring oil andwater saturations inside reservoir rock using MRI. His prior re-search included automotive lubrication, surface analysis withXPS, and UV-Visible-IR spectroscopy. Jim C. Stevens is a SeniorStaff Technician with ConocoPhillips. He is currently working inthe Flow Visualization Lab developing and analyzing experi-ments using MRI to show the formation and disassociation ofmethane hydrates in porous media. His previous work includesspecial core analysis and microbial enhanced oil recovery.James J. Howard is a principal scientist in the reservoir mecha-nisms laboratory at ConocoPhillips’ Bartlesville TechnologyCenter, where he works on a variety of pore-scale character-ization projects using NMR/MRI and other measurement tech-niques. Eirik Aspenes is a PhD graduate in reservoir physics atthe Department of Physics and Technology at the University ofBergen, Norway. He is currently employed as reservoir engi-neer by StatoilHydro. Geir Ersland is a third-year PhD student inthe Department of Physics and Technology at the University ofBergen, Norway. His research involves applications of 2Dnuclear tracer imaging (NTI) and 3D MRI techniques in studiesof hydrate phase transitions and multiphase flow in reservoirrocks. Jarle Husebø is a PhD student in reservoir physics at theDepartment of Physics and Technology at the University of Ber-gen, Norway. His experimental work is focused on natural gashydrate (NGH) phase transitions in porous medium as well aspossible production schemes for NGH. He is also developingMRI-analysis software for in-situ monitoring of water-gas-hydrate dynamics in reservoir rock at the ConocoPhillips Tech-nology Center, Bartlesville, Oklahoma. David Zornes is currentlyPetroleum Engineering Technology Advisor for the COP Reser-voir Performance Group, with responsibilities for long-term R&Dand joint industry projects. Zornes has served Phillips and Cono-coPhillips for near 32 years in a variety of positions, includingResearch Technical Manager, EOR Director, Reservoir Director,Research Supervisor, EOR Engineer, and Research Engineer.
