Overall shapes of volume-pressure plots vary considerably (Fig. 4). Many of the differences between volume-pressure plots can be understood by considering individual PCS cores and varying experimental parameters. For example, the significant increase in pressure with no change in volume for Core 164-995B-10P (Fig. 4J) undoubtedly was caused by warming of the PCS for 300 min with the PCS valve closed (Dickens et al., Chap. 43, this volume). We have identified many of these experimental artifacts in Table 1, Table 2, Table 3, and Table 4 and in the notes for Fig. 4.
However, despite all differences in experimental parameters for individual cores, there are two general profiles for volume-pressure plots. Most cores give a concave downwards volume-pressure plot as exemplified by Core 164-995A-45P (Fig. 4E). After a threshold pressure is surpassed, there is a large drop in pressure with a disproportionally small increase in CH4 volume. Subsequent drops in pressure then release greater volumes of CH4 per drop in pressure. The threshold pressure (or second threshold pressure) for these cores (Table 2) is less than 2.9 MPa—the pressure of the CH4 gas-CH4 hydrate-water equilibrium curve at 0ºC (Fig. 2).
In contrast, Cores 164-996A-7P (Fig. 4K), 164-996D-7P (Fig. 4L), 164-997B-49P (Fig. 4P), 164-997A-55P (Fig. 4Q), and 164-997B-10P (Fig. 4R) give volume-pressure plots with portions that are horizontal or concave upwards. In these cases, after the threshold pressure is surpassed, a small (or negligible) drop in pressure leads to a disproportionally large release of CH4 volume. The threshold pressure for these cores (Table 2) is at or above 2.9 MPa.
Based on these observations, we suggest the difference in overall volume-pressure profiles distinguishes PCS cores that contained only dissolved CH4 (e.g., 164-Cores 995A-45P and 164-995A-60P) from cores that contained CH4 hydrate and dissolved CH4 (Cores 164-996A-7P, 164-996D-7P, 164-997B-49P, 164-997A-55P, and 164-997B-10P). This suggestion is broadly consistent with the theory outlined by Hunt (1979) and Kvenvolden et al. (1983). However, we give a fundamental qualifier: our interpretation is for the start of experimental conditions when the PCS is placed in an ice bath and not at in situ conditions in the sediment column, which is at temperatures from 3ºC at the sediment-water interface to greater than 25ºC at depth (Paull, Matsumoto, Wallace, et al., 1996).
The emphasis between experimental and in situ conditions can be understood by comparing observed and "expected" pressure-volume plots, and by considering how the PCS operates. Several PCS cores suspected of containing gas hydrate at in situ conditions (Table 1) do not give volume-pressure plots similar to expectations for cores with gas hydrate (e.g., Core 164-995A-45P; Fig. 4E). Moreover, Core 164-997B-10P rendered a volume-pressure plot somewhat similar to that expected for a core with gas hydrate (Fig. 4R), although there is no evidence that this core actually contained gas hydrate at in situ conditions (Table 1).
The problem is twofold. First, in situ temperatures for PCS cores are significantly greater than 0ºC because of the geotherm. Thus, there is the possibility that a core recovered with free CH4 gas and dissolved CH4 at in situ pressure and temperature conditions (i.e., no gas hydrate) could form CH4 hydrate upon cooling to 0ºC.
Second, the PCS contains two chambers and substantial quantities of borehole water. There is an inner chamber that contains a sediment core and borehole water to a total volume of 1385 mL. There is also an outer chamber that contains 2615 mL of borehole water. Because the two chambers are connected inside of the PCS (Pettigrew, 1992), and because borehole water contains negligible quantities of CH4 (Paull, Matsumoto, Wallace, et al., 1996), a core recovered at high CH4 concentration at in situ conditions equilibrates with borehole water of low CH4 concentration before gas release experiments are initiated. The dilution resulting from this mixing is significant (at least 70%). For example, a 1385-cm3 core with 60% porosity and a CH4 molality of 0.20 mol/kg at in situ conditions will have a CH4 molality of 0.05 mol/kg before the first release of gas from the PCS if the core completely equilibrates with the large volume of CH4-poor borehole water that is present within the PCS (note that kilograms here and elsewhere refers to the entire CH4-water system, including solid hydrate but excluding sediment).
At moderate temperature (T < temperature on the CH4 gas-CH4 hydrate-water equilibrium curve) and high pressure (P > pressure on the CH4 gas-CH4 hydrate-water equilibrium curve), the presence or absence of CH4 hydrate in a pressurized container will depend on CH4 concentration (Fig. 2). The theoretical concentration necessary to form hydrate in seawater at 0ºC is 0.052 mol/kg (Handa, 1990).
Methane concentration can be calculated from the amount of CH4 released from the container and the mass of the water inside of the container. We have made two calculations for CH4 concentration (Table 3). The first calculation is the pore-space CH4 concentration. This is the CH4 concentration in the sediment core prior to equilibration with surrounding borehole water. In making this calculation, we have used shipboard values for porosity and have assumed that all CH4 released from the PCS was originally in pore space. The latter assumption is consistent with the fact that most PCS cores did not release CH4 unless they contained sediment (Table 1, Table 2, and Table 3). The second calculation is the total CH4 concentration. This is the CH4 concentration in the PCS after equilibration with surrounding borehole water, calculated assuming that there is 4000-cm3 total volume (borehole water, sediment, and pore space) inside of the PCS.
All but two cores had sufficient quantities of CH4 to form CH4 hydrate in pore water at 0ºC and pressure greater than 2.9 MPa (Table 3). However, only six cores had sufficient quantities of CH4 to form CH4 hydrate in the PCS at 0ºC and pressure greater than 2.9 MPa after equilibration with borehole water (Table 3). Five of these six cores are those that gave volume-pressure plots interpreted as representing cores with CH4 hydrate. Volumes and pressures were released exceptionally fast from Core 164-997A-25P, and the volume-pressure plot for this particular core (Fig. 4N) cannot be used to make any interpretations regarding the presence or absence of CH4 hydrate.
Results of the first calculation--pore-space CH4 concentration (Table 3)—are of considerable interest and discussed elsewhere (Dickens et al., 1997). Pore-space CH4 concentrations are total CH4 amounts in pore space before borehole water dilution and release of pressure. Thus, these data represent in situ methane concentrations and can be used in conjunction with appropriate stability curves (Handa, 1990; Duan et al., 1992; Dickens and Quinby-Hunt, 1994, 1997; Tohidi et al., 1995) to estimate in situ quantities of gas hydrate and free gas (Dickens et al., 1997).
As gas is slowly released from a pressurized core under isothermal conditions, expected changes in concentration and pressure (i.e., the "path" of the core in concentration-pressure space) are entirely dictated by the initial gas concentration (Fig. 2). A "synthetic" volume-pressure plot, therefore, can be made for a pressurized core if the CH4 concentration and water mass are known, and if it is assumed that complete equilibrium is maintained at all times during degassing (i.e., gas is removed reversibly in an infinite number of steps).
Table 4 lists the expected threshold pressure, CH4 hydrate volume, and dissolved CH4 volume for all 29 PCS cores at 0ºC and 5ºC given the calculated CH4 concentration and water mass for each core. Using this information, we have superimposed synthetic volume-pressure plots on observed volume-pressure curves for 20 of the PCS cores (Fig. 4).
In general, at a given volume, pressures on the synthetic volume-pressure plot exceed observed pressures. This suggests nonequilibrium degassing, although the cause is unclear. During degassing of the PCS, internal pressure is maintained by gas exsolution of CH4 that is dissolved in water. The offset between synthetic and observed curves therefore suggests that insufficient gas has exsolved from water during degassing. Possible explanations include (1) water inside of the PCS is supersaturated with CH4 (perhaps because of absorption on sediment particles or because bubbles cannot nucleate in small pore space; B. Clennell, pers. comm., 1998); or (2) CH4 transfer between inner and outer chambers is slow such that CH4 concentrations are higher in the outer chamber. In any case, the offset reflects nonequilibrium conditions.
Five cores (Cores 164-995A-52P, 164-995A-60P, 164-995B-7P, 164-997A-18P, and 164-997A-55P) exhibit synthetic volume-pressure plots that are similar to observed volume-pressure plots (Fig 4. (G, H, I, M, Q). It is unclear, however, why gas released from these particular cores is close to that predicted for equilibrium degassing in contrast to other cores. There are no obvious parameters, including average CH4 release rate, that distinguish these five cores (Table 1 and Table 2), although we note that they generally have higher inferred pore-water CH4 concentrations in the inner chamber (Table 3).
Pressures on synthetic volume-pressure plots are less than observed pressures at a given volume for three cores (Cores 164-995A-18P, 164-995A-36P, and 164-995A-48P; Fig 4 (B, D, F). The likely reason for these atypical plots is that the total volume of gas was not recovered for these three cores (Table 2).
A simple synthetic volume-pressure plot cannot be generated for Core 164-997B-10P (Fig 4R) because this core was not maintained at isothermal conditions. Temperatures inside of the PCS equilibrate with the ice bath at 0ºC after about 60 min (Paull, Matsumoto, Wallace, et al., 1996, p. 123). Core 164-997B-10P must have been at a temperature greater than 0ºC when gas was first released from this core after only 39 min (Table 1). An appropriate synthetic volume-pressure plot for Core 164-997B-10P would have pressures at a given volume higher than expected for a core at 0ºC (Fig 4R).
Time-pressure plots for degassing of PCS cores at Sites 994, 995, and 996 have been presented by Paull, Matsumoto, Wallace, et al. (1996). Time-pressure plots for degassing of PCS cores at Site 997 can be constructed from data given by Dickens et al. (Chap. 43, this volume). Plots for Cores 164-997A-18P, 164-997A-49P, and 164-997B-15P are shown in Fig. 5. All cores analyzed during Leg 164 have time-pressure plots with intervals where pressure increased over time after gas was released from the PCS and the valve to the container was closed (Fig. 5). However, Core 164-997A-18P probably did not have hydrate at in situ conditions (although see Egeberg and Dickens, 1999), and Core 164-997B-15P probably did not have hydrate at the start of experimental conditions (Table 3). Based on these data and observations, we conclude that "sawtooth characteristics" on time-pressure plots cannot be used to discriminate cores with CH4 hydrate. Our explanation is that gas coming out of supersaturated water will also give rise to a pressure increase after gas is released from a pressurized container.
Kvenvolden et al. (1983) presented time-pressure plots for three pressurized cores recovered at DSDP Site 533. All three of these plots showed "sawtooth characteristics" similar to those described above, in which pressure repeatedly increased over time after gas was released from the container and the valve to the container was closed. Although Kvenvolden et al. (1983) attributed this pattern to dissociation of CH4 hydrate inside of the pressurized container, they recognized the possibility that the pattern could be caused by "inefficient transfer of gas coming out of solution."
An important observation made during Leg 164 is that sediment recovery using the XCB coring technique (Paull, Matsumoto, Wallace, et al., 1996, p. 314) was roughly inversely proportional to gas distribution directly determined from PCS experiments (Table 3; Dickens et al., 1997) and inferred from downhole logging, pore-water Cl- concentrations, and vertical seismic profiling (Holbrook et al., 1996; Paull, Matsumoto, Wallace, et al., 1996; Egeberg and Dickens, 1999). Presumably, large quantities of gas "blow" significant quantities of sediment out of the checkvalve at the top of the XCB when pressure is decreased quickly during core retrieval. This phenomenon can be dramatic. No sediment was recovered by the XCB from the critical depth interval of interest surrounding the bottom simulating reflector (BSR) at Site 997 (Paull, Matsumoto, Wallace, et al., 1996).
Information presented here suggests that significant volumes of gas are released from sediment pore water only after a specific threshold pressure is surpassed. The maximum threshold pressure for a sediment core depends largely on temperature and lies on the CH4 gas-CH4 hydrate-water equilibrium curve (Fig. 1, Fig. 2). With knowledge of the hydrotherm, it may be possible to estimate the depth in the water column where gas is rapidly released from sediment pore water as a core is being carried to the surface inside the drillstring. Slow wireline recovery across this water depth may significantly increase core recovery, because it would allow gas to escape more slowly with consequently less core disruption.