Ind. Eng. Chem. Res. 2003, 42, 1517-1521
1517
Formation Kinetics of Propane Hydrates Carlo Giavarini,* Filippo Maccioni, and Maria Laura Santarelli Chemical Engineering Department, University of Rome “La Sapienza”, Via Eudossiana 18, 00184 Rome, Italy
Propane hydrates are rarely considered by scientists. Despite the narrow borders of the formation region, they can form during storage and transportation of the liquefied petroleum gases. Therefore, it is important to know the induction time for solid hydrate formation. The present paper has considered the formation both from melting ice (at 1 °C and 4 bar) and from water (at 2 °C and 3.6-4.8 bar) in a stirred vessel. The formation from ice was quite instantaneous, while the production from water took about 15 h to start (and about 3 days to be completed) and depended on the pressure: in fact, it was slower at low pressure. All hydrates contained a high amount of ice (75-80%). The modulated differential scanning calorimetry was used for hydrate characterization: the reversing (heat-capacity) curves permitted one to quickly distinguish between hydrate and ice, also allowing a semiquantitative evaluation of the hydrate content. Introduction and Motivation The problem of gas hydrate formation is well-known in the petroleum industry, which takes care to prevent pipe plugging by solid hydrates.1-3 Gas hydrates (or clathrates) are icelike structures composed of a host lattice consisting of hydrogen-bonded water molecules that can host small gaseous molecules (CH4, C2H6, C3H8, CO2, etc.). Most studied are methane hydrates, both because of the gas pipe-plugging problems and because methane hydrates are expected to develop as a new energy resource.4,5 Propane hydrates are rarely considered by scientists. The proceedings of the Fourth International Conference on Gas Hydrates6 report only one paper dealing with the first ever measurements of CH4 + C3H8 hydrate composition as a function of pressure, temperature, and vapor composition.7 The hydrates were prepared by condensing the gases on powdered ice; characterization was performed by 13C NMR spectroscopy. The paper is important in hydrate modeling for natural gas hydrate prediction programs. Another recent paper8 deals with the use of additives to enable C3H8 hydrate existence in a lower pressure range and up to higher temperatures; the equilibrium hydrate-liquid water-vapor (H-Lw-V) has been studied carefully. However, for storage, transportation, and use of liquefied petroleum gases (LPG), mainly composed of propane and butane, it is important to know the conditions and kinetics of hydrate formation in the presence of humidity or ice; in other words, it is useful to know the induction time for solid hydrate formation from the gas phase present in storing vessels and general equipment. To avoid valve plugging and other problems, because hydrate formation from butane is quite difficult,9 this paper presents an experimental study limited to formation kinetics of propane hydrate from ice and from water. A modulated differential scanning calorimetric (MDSC) technique was used at atmospheric pressure for the C3H8 hydrate characterization; it was the first time the modulated calorimetry was considered to study the gas * To whom correspondence should be addressed. Tel: +390644585565. Fax: +390644585416. E-mail: carlo.giavarini@ uniroma1.it.
Figure 1. Propane’s gas hydrate phase behavior in the system H2O + C3H8,8 where Lw is the liquid water, H is the hydrate, V is the vapor, and I is the ice.
hydrates. A number of techniques have been employed to obtain compositions of the gas hydrates; however, the successful application of these techniques is hampered by their inability to distinguish between ice and hydrate or by the necessity to have specially modified apparatuses to produce the hydrate inside the analytical apparatus. NMR and Raman spectroscopy are normally used for hydrate characterization, and the hydrates are formed directly inside the apparatus.7,10 Analytical calorimetry (DSC) has been used by a number of authors;1,11-13 because the gas hydrates are generally stable under pressure of the hydrate-forming gas, the calorimeters had to be modified for the study of samples under pressure.11 The problem has been overcome by using tetrahydrofuran or ethylene oxide hydrates, which are stable at room temperature and pressure.1,14 In our case, owing to the relatively higher stability of the propane hydrate, MDSC was used without any modification, at atmospheric pressure, to detect propane hydrates, in the presence of ice. Background Propane forms S II hydrate with 17 maximum hydration number; together with i-C4H9, it is at the limit of
10.1021/ie0207764 CCC: $25.00 © 2003 American Chemical Society Published on Web 02/26/2003
1518
Ind. Eng. Chem. Res., Vol. 42, No. 7, 2003
Figure 2. Main steps of propane hydrate formation from melting ice, with reference to the Sloan equilibrium curve.1,18 Table 1. Propane’s Characteristics propane air propylene water isobutane ethane carbon monoxide carbon dioxide n-butane unsaturated C4 sulfur
Figure 3. Propane hydrate formation from ice (20 g). The upper curve refers to the pressure drop; the lower curve refers to the temperature. Numbers on the curves refer to the steps indicated in Figure 2.
99.9000 vol % 10.0000 ppmvol 10.0000 ppmvol 3.0000 ppmvol 300.0000 ppmvol 400.0000 ppmvol 2.0000 ppmvol 5.0000 ppmvol 100.0000 ppmvol 50.0000 ppmvol 0.3000 ppmvol
the clathrate formation region.1,3,8,9 In the system H2O + C3H8, the pressure of the two quadruple points is relatively low: PQ1 ) 1.5 bar and PQ2 ) 6.0 bar, respectively.8 The temperature of Q2 is also relatively low (TQ2 ) 279 K), and the equilibrium line H-Lw-LC3H8 restricts the temperature region of C3H8 hydrate to a temperature which is only slightly higher than TQ2 (Figure 1). Although C3H8 hydrate formation is possible at relatively mild conditions of pressure and temperature, the narrow borders of the formation region restrict the possibility of its formation.
Figure 4. Weight losses with time of the propane hydrate (lower curve); the upper curve refers to ice taken as a reference.
MDSC Analytical Technique MDSC is a relatively new technique that overcomes a number of limitations of conventional DSC. Up to now it has been mostly used to study polymer transitions. The MDSC concept involves the imposition of a sinusoidal wave on the normally linear heating ramp so that portions of each cycle are at different heating and cooling rates, although the general overall trend is a linear change in average temperature.16-21 The mode of modulation affords the thermal analyst an opportunity to study a physical or chemical change in greater detail. One of the major contributions of this technique and analysis is that the total heat flow rate can be separated into two additional signals following the general equation proposed by TA Instruments22 and by other authors.17,19,21
δQ dT ) Cp + f(T,t) δT dt The first term of the equation is directly correlated to the heat capacity of the material, while the second is a function of temperature and time characteristic of the tests. These terms can be plotted separately in two curves. One of these curves represents the component
Figure 5. Propane hydrate formation from water, followed by decomposition.
which is heating rate dependent (i.e., which is in phase with the modulated heating); it is frequently defined as the “thermodynamic” component and associated with a “reversing heat flow”. The second curve corresponds to the rate of heat flow that depends only on the absolute temperature (i.e., which is out of phase with the modulated heating): it is usually defined as a “kinetic” component and associated with a nonreversing heat flow. Events associated with polymers that are nonreversing during the period of the oscillation are molecular relaxations, cold crystallization, evaporation, thermoset cure, and decomposition. The change in Cp at the glass transition and the endotherm due to melting are, on the contrary, normally reversible.16,20-22 We must keep in mind, however, that these reversible processes may alter their appearance as the frequency
Ind. Eng. Chem. Res., Vol. 42, No. 7, 2003 1519
Figure 6. Influence of the gas pressure on the propane hydrate formation at 1.5 °C.
of modulation increases to a point that the process cannot follow the temperature cycle. Experimental Section To investigate the formation of C3H8 hydrate, a 0.2 L stirred vessel similar to that used by Stern et al.15 was used. A cooling jacket permitted one to control the temperature at 2 ( 0.2 °C in the water-C3H8 tests and at 1 ( 0.5 °C during the tests with ice. The pressure was varied in the 3.6-4.8 bar range for the experiments with water and fixed at 4.0 bar for the experiments with ice. The pressure and temperature were continuously monitored and registered. Propane with 99.9000% purity was used; the main impurities are reported in Table 1. Stirring was applied for the experiments both with water (400 rpm) and with ice (300 rpm). During the tests with water, stirring was necessary to facilitate the contact with gaseous propane; in the case of the ice, stirring was applied to have a more homogeneous reaction. The same experimental procedure was used for all experiments. A sample of ice (about 20 g) was taken from a freezer at -14 °C; the diameter of the ice granules was 2 mm on average. The ice was put in a stirred vessel (Parring Instruments Co.); some pressure-vacuum cycles were carried out with propane to
eliminate the air entrapped in the vessel. Finally, propane was introduced under pressure, and stirring was started. For the experiments with water, about 20 g of distilled water was fed to the stirred vessel. The temperature was controlled by a water-NaCl solution cooled by a cryostat. The temperature and pressure were monitored respectively with a PT100 sensor and with an Ellison standard sensor. The elaboration of the experimental data was carried out by a Virtual-Bench Logger (National Instruments). Hydrate formation was detected by the pressure drop. To roughly check the amount of hydrate formed during the process, a very simple method was applied. After formation, the hydrate (about 20 g) was overcooled in a pressurized vessel to -5 °C and immediately transferred to an analytical balance to measure the weight losses; another sample (about 5-15 mg) was analyzed by MDSC (TA Instruments model 2920). Taking into account the influence of the water vapor pressure, the weight losses were a measure of the gas entrapped in the hydrate. This value was in agreement with the amount of propane calculated from the pressure drop. The MDSC tests were carried out at a 5 °C/min heating rate, with a modulation amplitude of (0.5 °C every 60 s. The temperature range between -20 and +20 °C was considered with the modulation. A nitrogen purge gas was used during the tests (50 mL/min). Results and Discussion Hydrate Formation and Decomposition. Figure 2 shows in detail the main steps of the propane hydrate formation process with reference to the Sloan equilibrium curve Q1-Q2.1,23 Figure 3 reports respectively the pressure and temperature curves against time for the propane-melting ice reaction. The hydrate formation is followed by a slight increase of the temperature and a sensible drop of the pressure, which is correlated with the propane formation kinetics. The pressure drop of 2.6 bar corresponds to about 20% conversion of the ice (20 g originally) into hydrate. The weight loss of about 4.8%, reported in Figure 4, roughly confirms the composition of the reactor product. Therefore, the hydrate
Figure 7. MDSC curves of pure ice: (a) total heat flow; (b) kinetic component; (c) heat-capacity (reversing) component.
1520
Ind. Eng. Chem. Res., Vol. 42, No. 7, 2003
Figure 8. MDSC curves of propane hydrate: (a) total heat flow; (b) kinetic component; (c) heat-capacity (reversing) component.
Figure 9. Reversing MDSC curves of hydrates with different ice contents: (a) 20% hydrate, (b) 8%, (c) 7% (analytical balance).
formation from melting ice (at least under these experimental conditions) is practically instantaneous, but the hydrate contains a noticeable amount of ice. The problem of hydrates purity is seldom mentioned by the literature; however, especially when the formation region is quite narrow, hydrates normally contain variable amounts of ice. Contrary to the formation from melting ice, the propane hydrate formation from water was, under the adopted experimental conditions, quite slow. Figure 5 shows the trend of pressure and temperature with time, during hydrate formation (starting from 40 g of water). In this case the amount of hydrate is higher (about 25%), but the final product still contains a great amount of ice. The induction time is extended and directly depends on the starting pressure (Figure 6). In fact, it ranges from 13.3 h at 4.8 bar of pressure to 18.0 h at 3.6 bar of pressure. Figure 5 also shows the decomposition of the hydrate because of an induced temperature increase. After decomposition, the pressure in the reactor reaches the original level.
The practical conclusion is that, under our experimental conditions, propane hydrate formation is relatively slow in the presence of humidity (water vapor) in the storing equipment but becomes instantaneous if ice is present. The formed hydrate contains massive quantities of ice. Hydrate Characterization by MDSC. Traditional DSC curves (continuous line in Figures 7 and 8) do not show appreciable differences between the curves of ice and propane hydrate. However, when the MDSC curves are split into reversing heat flow and nonreversing (kinetic) components, different behaviors can be noticed between the reversing curves of ice and hydrate: the onset of endothermic ice fusion (Figure 7) starts at 0.82 °C with a peak temperature at 4.14 °C (dash-dotted line). In the presence of the propane hydrate, a new reversing effect appears (Figure 8) with the onset at -4.34 °C and a peak temperature at -1.47 °C; the ice fusion is superimposed at 3.05 °C. The first peak is due to the endothermic hydrate decomposition. This is in agreement with the stability range of our propane hydrates (containing 75-80% ice), which are
Ind. Eng. Chem. Res., Vol. 42, No. 7, 2003 1521
stable, at atmospheric pressure, up to about -8 °C.24 Owing to the MDSC heating rate (5 °C/min), the decomposition (hydrate) and melting (ice) temperatures are shifted at higher temperatures, with respect to the equilibrium temperatures. The presence of the endothermic propane hydrate peak in the so-called “reversing curve” can be explained by the fact that melting, lattice destruction, and consequent decomposition are practically simultaneous processes. In Figure 8 the kinetic curve is in this case similar to the curve of the total heat flow given by traditional DSC. The extension of the reversible heat flow peak for the hydrate depends on the hydrate contents in the mixture of hydrate-ice. Figure 9 shows the reversing (heatcapacity) curves of three samples with different hydrate contents: the loss of propane shifts the peak to higher temperatures because of the increased amount of ice in the sample. MDSC is, therefore, a good technique to check the presence of propane hydrate in an ice sample. However, the calculation of the hydrate percentage from the MDSC curve cannot be accurate because of the superimposition of the hydrate and ice peaks. In our case, the apparatus software gave the following results for the sample of Figure 9: (a) 20%, (b) 10.9%, and (c) 8.7%. They are generally higher than those estimated with the analytical balance, which were respectively (a) 20%, (b) 8%, and (c) 7%. Conclusions Propane hydrate formation is possible at mild conditions of pressure and temperature; however, the narrow borders of the existence region restrict the possibility of formation. Under the experimental conditions adopted in this work, the formation from ice was instantaneous; on the contrary, the formation from water was a long process, which took some days to be completed. The induction time directly depended on the pressure. All produced hydrates contained a considerable amount of ice (75-80%). Owing to the relatively high stability of the propane hydrates, it was possible to roughly determine their gas content by using a simple analytical balance. The characterization was performed by using an unmodified modulated analytical calorimeter: the reversing or heat-capacity curves of the hydrate were different from the ice curves, in both the shape and position of the peaks; they permitted a semiquantitative evaluation of the hydrate content. Acknowledgment The authors thank Prof. E. D. Sloan for useful suggestions. Literature Cited (1) Sloan, E. D., Jr. Clathrate Hydrates of Natural Gases; Marcel Dekker: New York, 1998. (2) Holder, G. D., Bishnoi, P. R., Eds. Gas Hydrates, challenges for the future; Annals of the New York Academy Sciences: New York, 2000. (3) Makogon, Y. F. Hydrates of hydrocarbons; Penwell Books: Tulsa, OK, 1997.
(4) Giavarini, C.; Silla, R. Gli idrati degli idrocarburi gassosi. Riv. Combust. 1998, 52, 280. (5) Giavarini, C. Metano dagli idrati. Chim. Ind. (Milan) 2002, 84, 47. (6) Mori, Y. H., Ed. Proceedings of the Fourth International Conference on Gas Hydrates, Yokohama, Japan, 2002. (7) Kini, R. A.; Dec, S. F.; Sloan, E.D., Jr. Measurement of CH4 + C3H8 Hydrate Composition via 13C NMR Spectroscopy. Proceedings of the Fourth International Conference on Gas Hydrates, Yokohama, Japan, 2002. (8) Moojer, M. M.; Peters, C. J.; De Swaan Arons, J. Gas Hydrate Phase Equilibria for Propane in the Presence of Additive Components. Fluid Phase Equilib. 2002, 193, 245. (9) Ripmeester, J. A. Hydrate ResearchsFrom Correlations to a Knowledge-Based Discipline. In Gas Hydrates, challenges for the future; Holder, G. D., Bishnoi, P. R., Eds.; Annals of the New York Academy of Sciences: New York, 2000. (10) Sum, A. K.; Burrus, R. C.; Sloan, E. D., Jr. Measurement of Clathrate Hydrate Properties via Raman Spectroscopy. J. Phys. Chem. B 1997, 101, 7371. (11) Handa, Y. P. Composition, Enthalpies of Dissociation, and Heat Capacities in the Range 85 to 270 K for Clathrate Hydrates of Methane, Ethane, and Propane, and Enthalpy of Dissociation of Isobutane Hydrates, as Determined by Heat-Flow Calorimeter. J. Chem. Thermodyn. 1986, 18, 915. (12) Dalmazzone, C.; Herzhaft, B.; Dalmazzone, D. Characterization of Hydrate Formation in Drilling Muds Using Differential Scanning Calorimetry (DSC). Proceedings of the Fourth International Conference on Gas Hydrates, Yokohama, Japan, 2002. (13) Dalmazzone, D.; Kharrat, M.; Fouconnier, B.; Clausse, D. Thermodynamic Study of Several Hydrates in Dispersed Systems Using Differential Scanning Calorimetry. Proceedings of the Fourth International Conference on Gas Hydrates, Yokohama, Japan, 2002 (14) Leaist, D. G.; Murray, J. J.; Post, M. L.; Davidson, D. W. Enthalpies of Decomposition and Heat Capacities of Ethylene Oxide and Tetrahydrofuran Hydrates. Can. J. Phys. Chem. 1982, 86, 4175. (15) Stern, L. A.; Kirby, S. H.; Durham, W. B. Peculiarities of Methane Clathrate Hydrate Formation and Solid-State Deformation, Including Possible Superheating of Water Ice. Science 1996, 273, 1843. (16) Sauerbrunn, S.; Crowe, B.; Reading, M. Modulated differential scanning calorimetry. Am. Lab. 1992, 24 (12), 44 (17) Reading, M. Modulated differential scanning calorimetry a new way forward in materials characterization. Trends Polym. Sci. 1993, 1 (8), 248. (18) Wunderlich, B.; Jin, Y. M.; Boller, A. Mathematical description of differential scanning calorimetry based on periodic temperature modulation. Thermochim. Acta 1994, 238, 277. (19) Reading, M.; Wilson, R.; Pollock, H. M. Modulated differential scanning calorimetry: theory, pratice and applications. Proceedings of the North American Thermal Analytical Society Conference, Toronto, Canada, 1994. (20) Gallagher, P. K. Thermoanalytical Instrumentation, Techniques, and Methodology. In Thermal characterization of Polymeric materials, 2nd ed.; Turi, E. A., Ed.; Academic Press: San Diego, 1997. (21) Riga, A. T., Judovits, L., Eds. Materials Characterization by Dynamic and Modulated Thermal Analytical Techniques; ASTM STP 1402; ASTM: Baltimore, MD, 2001. (22) TA Instruments. DSC 2920 operator’s manual appendix C: modulated DSCTM Option, 1998. (23) Sloan, E. D., Jr. CSMHYD.EXE software. Clathrate hydrates of natural gas; Marcel Dekker: New York, 1998. (24) PVT Simular software package, version 11; Calsep: Lyngby, Denmark, 2000.
Received for review September 30, 2002 Revised manuscript received January 10, 2003 Accepted January 22, 2003 IE0207764