UNITED STATES DEPARTMENT OF THE INTERIOR GEOLOGICAL SURVEY THE SOLUBILITY OF METHANE IN SODIUM CHLORIDE BRINES by K. A. McGee N. J. Susak A. J. Sutton J. L. Haas, Jr. Open-File Report 81-1294 This report is preliminary and has not been reviewed for conformity with U.S. Geological Survey editorial standards and stratigraphic nomenclature. Any use of trade names is for descriptive purposes only and does not imply endorsement by the USGS. 1981 ABSTRACT The solubility of methane in pure water and in solutions of 5%, 10% 15%, and 20% (by weight) sodium chloride was determined experimentally at various temperatures and pressures. Methane solubility in sodium chloride solutions is significantly less than it is in pure water because of the salting-out effect. For example, in a solution of 20 weight % NaCl, the solubility of methane is only 28% of that in pure water under the same temperature and pressure conditions. We report here an experimentally determined value of 0.129 ± 0.03 for the salting-out (Setschenow) constant for methane. Also presented is a set of equations for calculating the solubility of methane under conditions of pressure, temperature and salinity as great as 137.9 MPa (20,000 psi), 350°C (662°F) and 25 weight % sodium chloride, respectively when only the salting-out constant and the solubility of methane in pure water at the temperature and pressure of interest are known. INTRODUCTION In the northern Gulf of Mexico sedimentary basin, large regions are distinguished by highly saline pore waters at unusually high temperatures and pressures (Kharaka and others 1977). These regions, known as geopressured- geothermal zones, underlie vast areas of coastal Texas and Louisiana along the Continental Shelf and are known to contain dissolved natural gas, primarily methane (Papadopulos and others, 1975). Pressure, temperature, and salinity in these subsurface waters can reach nearly 124.1 MPa (18,000 psi), 200°C (392°F) and 275,000 mg/L, respectively (Kharaka and others, 1978). Knowledge of the chemical behavior of methane in aqueous solution under these conditions is therefore of the utmost importance if estimates of the energy content of the dissolved gas in these geopressured-geothermal reservoirs are to have maximum reliability. Existing data on the solubility of methane at conditions approximating those in the geopressured-geothermal zones are limited and cover rather narrow temperature, pressure, and salinity ranges. The present investigation was undertaken in order to gather enough experimental data to derive a set of equations that will accurately predict methane solubility in pure water and brines of as much as 25 weight % NaCl at temperatures as high as 350°C (662°F) and pressures as great as 137.9 MPa (20,000 psi). EXPERIMENTAL METHOD The experimental apparatus (figure 1) consists of two 6 inch-long gold-lined pressure vessels of known volume, each fitted with a titanium plunger and connected by a titanium capillary tube to a valve assembly. The valve assembly consists of two stainless steel miniature valves and a Setra Model 204 pressure transducer. The temperature of each pressure vessel was measured by a type T thermocouple and a digital voltmeter. Pressure was displayed by a Setra Model 300A digital pressure readout. Tn addition, both temperature and pressure were continuously monitored on a strip-chart recorder. For each run, the pressure vessels were placed inside a mechanical convection oven and evacuated overnight to 1.33 Pascals. The pressure transducer was then adjusted for the correct reading on the digital readout. A 150-cc stainless steel gas transfer bottle filled with Matheson Purity methane (99.99%) was weighed and then connected to the entry valve. The gas was admitted to the pressure vessel until the pressure was 0.37 MPa. (0.37 MPa represents the ideal gas pressure for 0.07 gm. CH4 and is used so that pressure will not exceed the limits of the pressure transducer after the temperature is increased.) The bottle was then removed and reweighed, and the amount of methane added was then calculated by the weight difference (generally about 0.0600 to 0.0800 gm). The brine solution was prepared from a weighed amount of reagent-grade sodium chloride and a weighed amount of deionized water which had been degassed by boiling under a vacuum for 3-4 hours and rapidly cooled. The sodium chloride solution was transferred to a graduated cylinder and weighed. The volume of solution needed was calculated from the volume needed to fill the pressure vessel at 300°C (572°F) and 68.9 MPa (10,000 psi) using the density tables of Potter and Brown (1977). This volume was transferred to the pressure vessel by means of a screw pump pressure 3 generator. The graduated cylinder was then reweighed, and the weight of solution added was determined by difference. The oven was then turned on and the temperature was increased in steps, the resulting pressures being recorded at each step. The runs were allowed to equilibrate for a minimum of 3 hours (usually longer) at each temperature interval. The temperature was increased in increments until the volume of the liquid expanded to fill the vessels completely. This is shown graphically by line AB in figure 2. The temperature increases were then continued in small steps to determine the P-T curve for the liquid at constant volume (ignoring dv/dt for the pressure vessel) and composition until the pressure limits of the equipment were reached (nominally 68.9 MPa [10,000 psi]). This is shown graphically by line RC in figure ?. As a check, the temperature was then decreased in small increments, with the resulting pressure being recorded at each interval, until room temperature was again reached. Lines AB and BC in figure 2 depict two intersecting curves representing two-phase (liquid-vapor) and one-phase (liquid) conditions respectively. By using a curve-fitting routine on the experimental data, the function (dp/dtXiV ) can be derived for each of the curves. Those functions are in turn used to calculate the intersection temperature and pressure for the two curves. These results, along with the fixed values for volume and composition, are used to calculate the methane solubility. The experimental results are shown in table 1. METHANE SOLUBILITY IN PURE WATER The solubility of methane in pure water has been measured previously by various investigators (Culberson and McKetta, 1951; Huffy and others, 1961: O'Sullivan and Smith 1970: Sultanov and others, 1972: and Price 1979). The data from those studies and the present study were evaluated in several different ways in order to derive the simplest algebraic function to describe the data. Of the correlations tried we found that only the plots of log Q versus XCH were linear for several isotherms over the entire pressure ' 4 range. Q is defined here by the equation, 0 = __4 (1) PCH 4 where XCH is the molal concentration of methane in water and p^ is 4 ' 4 derived by subtracting the vapor pressure of pure water at the temperature of interest from the total pressure. The logarithm of Q is written as log 0 = log XCH - log pCH (2) 4 4 The data from the present investigation and most of the previously mentioned investigations are in good agreement. By using these data each isotherm was fit to the function, log 0 = a + bx^ (3) 4 The intercept (a) and the slope (b) were derived by using the technique of Haas (1978). These two vriables were then each fitted to polynomials in temperature (°C) the reciprocal of the variances at each isotherm being used to weight the data. The data of Price (1979) were not included in the calculation of a and b because when his data were used, plots of log Q versus the molality of methane did not yield linear results. The following equations were derived for the intercept and the slope of equation 3: a = -4.3345 x 102f2 + 35.7727^ - 3.56262 + 2.19921 x 10-st2 - 2.68155 x 10-8t3 (4) b = 40.3932t-"1 - 4.04655 + 2.85046 x 10'?t - 7.08934 x 10-^ + 6.14288 x 10-8t3 (5) where t is in degrees Celsius (°C). The estimated partial pressure of methane in the system is the difference between the total pressure and the partial pressure of water at the temperature under consideration. PCH = Ptotal - PH 0 (6) 4 2 Combining equations 2, 3, and 6 and rearranging terms results in the expression lo9(Ptotal - PH 0) = -a - b*CH + Io9 XCH (7) 2 44 The partial pressure of water can be derived from data in Haas (1976). When those values for p^ Q are used along with the total pressure, the solubility 2 of methane in pure water can be computed from equation 7. Values for the solubility of methane in water calculated in this manner are plotted in figure 3. METHANE SOLUBILITY IN SODIUM CHLORIDE SOLUTIONS Several investigators in the past have pointed out that the solubility of nonpolar gases in aqueous electrolyte solutions decreases with increasing electrolyte concentration because of the salting-out effect (Morrison and Billett, 1952 Namiot, 1961). The relationship between the solubility of a gas like methane in pure water versus its solubility in an electrolyte solution can be quantitatively expressed as a Setschenow type-equation, log yCH = "log XCH -ASI (8) 4 4 where y^-i is the molal solubility of methane in the electrolyte solution, 4 XQI-I is the molal solubility of methane in pure water at the same temperature 4 and pressure As is the Setschenow (or salting-out) constant, and I is the ionic strength of the solution. Although the true ionic strength of the electrolyte solution will be less at higher temperatures owing to increased association of NaCl precise evaluations of the influence of temperature of ionic strength are beyond the scope of this report. Therefore I, is arbitrarily set to equal the molality of sodium chloride. To evaluate the Setschenow constant, equation 8 is rearranged to read /XCH \ , AS = W 4 A/I YCH 1 \ 4// Thus, for experimental runs in aqueous sodium chloride solutions, the molality of methane and the molality of NaCl are determined from the starting conditions The temperature and pressure at the intersection point (B on figure 2) are determined from the experiment, and a value for the solubility of methane in pure water at this same intersection temperature and pressure is then used for XQH in equation 9. 4 The value for the Setschenow constant determined from this investigation is 0.129 + 0.030. The major source of uncertainty for As is the determination of the pressure value for the intersection point (R) on the liquid-vapor curve. As line BC (figure 2) is relatively steep, small temperature changes result in large pressure changes, so that it is difficult to resolve the pressure precisely. The calculated value for As is also sensitive to the value used for the solubility of methane in pure water in equation 9. However, that value, as mentioned previously is based upon the present investigation in combination with several previous investigations, all in good agreement, so that the error contribution from XCH is considered 4 insignificant. Finally, we tested the variation of the Setschenow constant as a function of temperature, weight % NaCl , molality of methane, and pressure, using multiple-regression analysis, and we found that As could not be correlated with changes in any of these variables over the range of experimental conditions. Indeed it is probably too simplistic an approach to try to correlate As with any one variable because of the complex influence of many factors on intermolecular interactions within the sodium chloride solutions. Once As is known the solubility of methane in aqueous salt solutions at a variety of temperature and pressure conditions can be calculated from equation 8, when only the solubility of methane in pure water under the same conditions is known. A routine based on this equation for calculating methane solubility in pure water or NaCl brines that make use of a programmable hand calculator has been described previously (Susak and McGee, 1980). RESULTS Smoothed values for methane solubility are presented in several different ways. Figures 4 through 7 are semi logarithmic plots of methane solubility as a function of temperature at several fixed salinity values. Pressures of 6.9 MPa (1,000 psi) to 137.9 MAp (20,000 psi) are shown as contours. These figures show that increasing pressure results in an increase in methane solubility. Increasing the temperature also increases methane solubility except at low pressures. At low total pressures, the vapor pressure of water increases significantly as temperature increases, resulting in a corresponding decrease in the partial pressure of methane. As the vapor pressure of water approaches the total pressure in the system, the solubility of the gas approaches zero. Figures 8 and 9 are semi logarithmic plots showing methane solubility as a function of temperature at fixed pressures of 68.Q MPa (10,000 psi) and 137.9 MPa (20,000 psi), respectively. Salinity values from 0% to 20% sodium chloride are shown as contours. Roth these plots clearly illustrate the effect of the salting-out process on methane solubility. As the concentration of sodium chloride electrolyte is increased, the solubility of methane is reduced. Published values for the solubility of methane in sodium chloride solutions at temperatures of geologic interest are scarce. O'Sullivan and Smith (1970) reported data on the solubility and partial molar volume of methane in salt solutions at isotherms of 51.5°C, 102.5°C, and 125.0°C over the pressure range 10.1 to 60.8 MPa (1465 to 8818 psi). Figure 10 shows a comparison between the smoothed solubility values of the present study and the experimental data points (recalculated to molality) of O'Sullivan and Smith at 125°C. Methane molality is plotted as a function of pressure 9