Scanning electron microscope imaging of grain structure and phase distribution within gas-hydrate-bearing intervals from JAPEX/JNOC/GSC et al. Mallik 5L-38: what can we learn from comparisons with laboratory-synthesized samples?

Cryogenic scanning electron microscopy provides an excellent means for examining phase distributions, grain morphologies, and grain contacts within gas-hydrate-bearing samples of both natural and laboratory origin. Here, recovered samples from JAPEX/JNOC/GSC et al. Mallik 5L-38 are imaged, and comparisons are made between the geometrical arrangement and appearance of the gas hydrate ± ice phase and arrangements, grain contacts, and textures displayed by pure gas hydrate and gas hydrate–sediment aggregates grown and tested in the laboratory. Many gas hydrate ± ice sections in Mallik samples occur as smooth, dense material surrounding isolated macropores, whereas other closely juxtaposed sections display highly mesoporous textures. The possible origins of the mesoporosity are evaluated, and likewise the origin of ice mixed with the gas hydrate, both of which influence interpretations of physical-property measurements. Lastly, the authors image lab-synthesized methane hydrate+quartz aggregates using phase proportions comparable to those displayed by Mallik material from sand intervals, and discuss the implications of strength measurements made on the synthetic samples. United States Geological Survey, 345 Middlefield Road, MS/ 977, Menlo Park, California 94025 U.S.A. U.C. Lawrence Livermore National Laboratory, P.O. Box 808, Livermore, California 94550 U.S.A. 1 Résumé : La microscopie électronique à balayage avec platine cryogénique est une excellente méthode pour examiner la répartition des phases, la morphologie des grains et la nature des joints intergranulaires dans des échantillons naturels ou synthétiques renfermant des hydrates de gaz. Au moyen de cette technique, nous avons obtenu des images d’échantillons prélevés dans le puits de recherche sur la production d’hydrates de gaz JAPEX/JNOC/GSC et al. Mallik 5L-38, et nous avons comparé les agencements géométriques, les joints intergranulaires et les textures de la phase hydrates de gaz ± glace dans ces échantillons à ces caractéristiques dans des échantillons produits et éprouvés en laboratoire, constitués soit d’hydrates de gaz purs ou d’agrégats d’hydrates de gaz et de sédiments. Dans un grand nombre des tranches d’échantillons provenant du puits Mallik, la phase hydrates de gaz ± glace montre une texture dense et unie avec des macropores isolés, alors que dans des tranches voisines cette phase a une texture fortement mésoporeuse. Nous avons considéré les origines possibles des mésopores et celles de la glace mêlée aux hydrates de gaz, deux facteurs qui influencent l’interprétation des mesures de propriétés physiques. Nous avons également obtenu des images d’agrégats synthétiques contenant de l’hydrate de méthane et du quartz en proportions comparables à celles des échantillons prélevés dans des intervalles de sable du puits Mallik. De plus, nous discutons des implications de mesures de résistance effectuées sur les échantillons synthétiques. INTRODUCTION AND BACKGROUND One of the challenges of investigating both natural and laboratory-made gas hydrate involves careful evaluation of their grain and pore structures, characteristics that can be useful guides to in situ initial growth conditions as well as to the effects of changes in environmental conditions including those during recovery, handling, transport, and/or storage. The amount, geometrical distribution, fabric, and morphologies of gas hydrate in nature not only influence properties of sediments or formations in which they occur, but can also affect physical-property measurements made on recovered samples. Scanning electron microscopy (SEM) offers significant potential for obtaining such textural information due to its versatility in detection capabilities, its potential resolution, and its large depth of focus. When applied to gas hydrate, however, there are a number of technical challenges: avoiding condensation of atmospheric water on samples during cold transfer, maintaining the gas hydrate sample material at conditions that avoid spontaneous decomposition or significant sublimation under vacuum, and either avoiding electron beam damage of the imaging area or learning to properly identify it when it does occur. Distinguishing handling-induced surface artifacts from the intrinsic sample surface morphology can also be difficult, as well as distinguishing gas hydrate from ice. Use of SEM for imaging gas hydrate has only recently been reported in the literature. W. Kuhs, D. Staykova, A. Klapproth, G. Genov, and coworkers (Kuhs et al., 2000; Staykova et al., 2003; Klapproth et al., 2003; Genov et al., 2004) used SEM techniques with excellent success in imaging and identifying grain structures in CH4, CH4-N2, CO2, and Ar hydrate prepared from reaction of ice with gases or liquids, as well in imaging natural gas hydrate recovered from marine and subpermafrost settings (Kuhs et al., 2004). Their SEM investigations also revealed the remarkable development of ‘mesoporous’ gas hydrate (i.e. pervasive submicrometre-sized pore development) formed by reaction of gas with ice at temperatures near the ice point. The present authors have also reported on SEM investigations of gas hydrate samples from both laboratory and natural origin, including pure methane hydrate (±sediments) at various extents of reaction and recrystallization (Stern et al., 2004a), dissociation and dissolution textures (Stern et al., 2002, 2003, 2004a), compaction and deformation textures (Durham et al., 2003a), pure, porous CO2 and propane hydrate (Circone et al., 2003; Stern et al., 2004a), and natural gas hydrate from the Gulf of Mexico (Stern and Kirby, 2004; Stern et al., 2004a). The present authors have also imaged numerous gas-hydrate-bearing samples with known fractions of ice, as well as samples used in surface sublimation or partial decomposition tests (Stern et al., 2004a) to specifically address the question of distinguishing gas hydrate from ice. For the study of gas hydrate samples recovered from natural or otherwise remote settings, these challenges are amplified by such additional unknowns as the complex in situ environmental conditions controlling the original growth textures, and any effects of subsequent recrystallization, annealing, secondary growth, dissociation, dissolution, or chemical exchange processes. The extent of sample damage or alteration incurred during retrieval of the gas hydrate presents additional unknowns. Lacking a wider sampling archive and long-term experience in assessing these issues, it is presumed that most interpretations of SEM images of natural gas hydrate samples are still somewhat speculative. Nonetheless, useful information about grain structure, pore characteristics, phase composition, and phase distribution may still be gleaned from even exploratory investigations, particularly if natural gas hydrate can be compared to other materials with precisely known composition and histories. Here, a ‘first look’ at some natural gas-hydrate-bearing samples recovered from JAPEX/JNOC/GSC et al. Mallik 5L-38 is presented (Dallimore and Collett, 2005; see also Techmer et al., 2005), and the observed features are compared to those documented previously in lab-synthesized gas hydrate samples of known composition, grain structure, pressuretemperature (P-T) processing histories, and compaction or deformation history. These results are largely qualitative, as only limited compositional information on the natural gas hydrate can be collected at this time. Instead, preliminary interpretations drawn from comparisons with the SEM image archive are presented. Lastly, the geometrical arrangement of phases observed in several Mallik samples from sand intervals are compared to those of mixed-phase aggregates formed in the lab, and implications of such comparisons are discussed. Benefits of pursuing comparative studies of natural and synthetic samples are the relative low cost, ease, and reproducibility of creating gas-hydrate-bearing samples in the lab with well characterized phases in known concentrations and geometrical arrangements, and the ability to isolate or introduce additional complexities in a controlled manner. EXPERIMENTAL METHODS Samples from Mallik 5L-38 arrived in a liquid nitrogen ‘dry’ vapor shipper, in which each sample was individually sealed within a plastic container. A small amount of free liquid nitrogen was present in the base of the shipper upon arrival, indicating that samples remained stable during transit. SEM preparation and imaging procedures For SEM imaging, small sections of samples (~0.5 cm x 0.5 cm x 0.75 cm) were cleaved under liquid nitrogen from the bulk samples, transferred to a sample stage within an evacuated and prechilled (below 100 K) cryo-preparation and coating station (Gatan Alto Model 2100), which was in turn attached to a LEO 982 field emission SEM. While still in the preparation chamber, the section was again cleaved by cold blade to produce fresh fracture surfaces for imaging that were not contaminated by surface-water condensation. A few samples were coated with AuPd for 60 s using a non-heat-emitting sputter head, but most were imaged uncoated to avoid increased time under vacuum exposure. Samples, still under vacuum conditions, were then inserted directly through the back of the preparation chamber and onto an auxiliary cryo-imaging stage in the SEM column. Sample temperature was continuously