JOURNAL OF GEOPHYSICAL RESEARCH, VOL. 108, NO. B4, 2182, doi:10.1029/2002JB001872, 2003

2. Apparatus and Techniques

[5] We deform samples in the standard mode of rock mechanics, called conventional triaxial testing, in which a cylindrical sample is confined on the outside by gas at pressure P that is sufficiently high to suppress fracturing and microfracturing, thus keeping the deformation strictly in the ductile field and assuring that the rheology measured is an intrinsic material property. Samples are deformed in compression by the application of an axial stress sigma , called the differential stress, which is sufficiently high to cause the sample to yield and permanently shorten at a strain rate varepsilon defined as the increment of length change per unit length per increment of time. Sample volume remains constant, so that as the sample undergoes permanent shortening, it also expands radially.

[6] The steady state rheology of most crystalline materials can be described by a mechanical equation of state or constitutive relationship [Poirier, 1985] of the form:

where R is the gas constant and A, n, E*, and V* are material-specific flow parameters, called the preexponential constant, the stress exponent, the activation energy, and the activation volume, respectively. Our objective is to characterize the rheology of methane hydrate, i.e., to identify the parameters in equation (1). Steady state means that varepsilon does not change in an increment of strain if sigma , T, and P do not change. Steady state cannot be assumed, but must be confirmed by experiment. On the basis of experience with ice and other oxides in high-temperature creep of fully dense, mono-mineralic materials, we anticipate that steady state deformation is reached at strains of a few percent after the first application of, or change in, sigma .

[7] The experiments described here are creep tests, wherein sigma is held constant and varepsilon is the dependent variable. Most of the samples were tested under multiple sets of conditions, that is, sigma , T, or P were stepped to new values after steady state, or at least the appearance of steady state, was achieved in the previous step.

Figure 2. Scanning electron microscope (SEM) images of fracture surfaces through methane hydrate starting material (a and b) and after hydrostatic compaction and deformation (c; sample 459). Although the relict texture of the granular ice reactant is apparent in Figure 2a, the hydrate product is found to be densely recrystallized to 20–40 µm grains surrounding open pores (Figure 2b). Samples after testing (Figure 2c) are fully dense, with no obvious evidence of residual porosity or of an ice contaminant phase. Samples for SEM observation were prepared and imaged under vacuum at temperatures below -160°C, at 1–2 kV.

Figure 3. Photographs of three samples (a) before compaction under hydrostatic pressure, sample 461, and (b and c) after hydrostatic pressurization to 100 MPa at 280 K, samples 459 and 457, respectively, and (d) cross-sectional sketch of the assembly. Photographs and sketch are at approximately the same scale. End cap, polycrystalline methane hydrate sample, porous disc, ZrO₂ spacer, and narrow neck of the force gage are surrounded by a 0.5-mm-indium tube that is sealed to metal column parts at its very top and bottom. High gas pressure outside the jacket squeezes all porosity out of the sample, which begins with a length of about 63 mm (Figure 3a), but shortens and narrows under confining pressure. The volume collapse from Figure 3a to Figure 3b is typical (about 30%); the much greater volume collapse of 457 in Figure 3c is anomalous. Figure 3c shows the entire column assembly of sample plus force gage. The narrow, thin-walled portion of the force gage (see Figure 3d) is the gage length, whose elastic distortion is a direct measure of the axial force on the sample. The wide portion at the top is the pressure vessel plug. Note that the main pressure seal is made at the stepped portion of the plug, so that the force measured by the gage is not affected by tractions at the pressure seal. The narrow tube emerging from the top of the plug is the pore pressure line, which reaches the porous disc and sample through an axial hole in the ZrO₂ spacer.

[8] High-purity samples of polycrystalline methane hydrate were synthesized by statically reacting granular seed ice and pressurized methane gas in cylindrical molds with an inner diameter of 25 mm [Stern et al., 1996, 1998, 2000]. The approximately 0.2 mm grain-size seed ice was prepared by crushing and sieving bubble-free ice that was grown from triple-distilled water. Observation of the resulting hydrate by scanning electron microscopy (SEM) shows that the final material consists of 0.25-mm-diameter clusters of dense, smaller grains, typically 20–40 µm, shown in Figure 2. As synthesized, the resulting hydrate has a porosity of about 29% and a hydrate number of 5.89 ± 0.1 [Stern et al., 2000]. For deformation, samples were cut to a length of about 63 mm and sealed in tight-fitting tubes of indium metal (0.5-mm-wall thickness) between hard steel end caps (Figure 3). One sample (number 459) was prepared by disaggregating and pulverizing pieces of synthesized hydrate with mortar and pestle, then packing the powder with the impacts of light hammer blows directly into the indium tube. Sample cutting and jacketing was done by hand in a vented glove box at temperatures that varied from 77 to roughly 120 K. Once sealed, the samples were stored at 77 K until testing.

[9] The upper end cap of the sample (Figures 3c and 3d) is the termination of a single piece of steel that serves several functions. It is the static base against which the sample is compressed by a piston moving upward through the lower seal of the pressure vessel; it is the containment for the top seal of the pressure vessel; and it is the sample force gage. The elastic strain of the gage length (Figures 3c and 3d) is directly proportional to sigma , hence it is truly an “internal” force gage, although the elastic strain of the gage is measured by a transducer sitting outside the pressurized volume. Joined to a through hole in the upper end cap is a length of small-diameter, high-pressure tubing, which allows chemical communication to the sample (Figure 3d). For the experiments described here, we used this conduit to control the methane pressure, P_CH₄, in the sample. To ensure that sample material did not intrude into this pore pressure line and that P_CH₄ was communicated evenly across the top of the sample, we placed a porous metal disc directly between the sample and end cap (Figure 3d).

Figure 4. A portion of the phase diagram for methane plus water [Sloan, 1998]. The cross symbols show test conditions.

[10]   We have found that no matter how careful the sample handling, the absence of free water cannot be guaranteed. We have therefore adopted a strategy of melting any contaminant water phase and squeezing it from the sample before beginning the first deformation step. Porosity is also eliminated prior to the start of deformation, and the creation of new porosity is suppressed throughout the experiment. We achieved these conditions by the following procedure:
   1.  Transfer the sample assembly as quickly as possible from storage at 77 K to the pressure vessel at about 175 K, expeditiously close the vessel, and proceed with pressurization.
   2.  At T approximately 175 K pressurize hydrostatically to 100 MPa in steps of roughly 5–10 MPa, measuring sample length at every step (from column displacement at the point where the internal force gage registers piston contact) and for the last three samples, recording the gas evolved from the sample at every step. P_CH₄ was held at room pressure throughout this phase, which lasted anywhere from 1.5 to 5 h.
   3.  Still at T approximately 175 K depressurize, remove the sample assembly, quench to 77 K, observe, measure, and photograph, return the sample assembly to the vessel, and repressurize to roughly 50 MPa. This phase was skipped as unnecessary for the last two samples, after the first four showed that compaction was always uniform.
   4.  At P approximately 50 MPa apply P_CH4 10 MPa and warm the pressure vessel to 280 K, making occasional measurements of sample length. The edge of the methane hydrate stability field at 280 K is at P_CH4 = 5 MPa (Figure 4). Once temperature has reached 280 K, increase P to 100 MPa to achieve final compaction. Measure sample length (again, with piston contact), which becomes the starting length for the purpose of calculating ductile strain. This entire procedure lasted from a few hours to nearly a day.
   5.  Conduct the deformation tests. These tests typically took a few days to 2 weeks to complete unless jacket failure ended the run prematurely.
   6.  Lower T to 175 K or below as quickly as possible holding P < 30 MPa and P_CH4 approximately 10 MPa Depressurize and move the sample assembly to storage at 77 K.

Citation: Durham, W. B., S. H. Kirby, L. A. Stern, and W. Zhang, The strength and rheology of methane clathrate hydrate, J. Geophys. Res., 108(B4), 2182, doi:10.1029/2002JB001872, 2003.