ISSN 0361-5219, Solid Fuel Chemistry, 2008, Vol. 42, No. 6, pp. 349–353. © Allerton Press, Inc., 2008. Original Russian Text © Yu.A. Noskova, V.A. Kazakov, M.A. Perederii, 2008, published in Khimiya Tverdogo Topliva, 2008, No. 6, pp. 29–34.
Adsorption Method for the Recovery of Hydrocarbons from Natural Gas and Associated Petroleum Gas Yu. A. Noskova, V. A. Kazakov, and M. A. Perederii Institute for Fossil Fuels, Leninskii pr. 29, Moscow, 119991, Russia e-mail:
[email protected] Received February 13, 2008
Abstract—Carbon sorbents were prepared based on various raw materials, and their sorption capacities for gasoline fraction were studied under static and dynamic conditions. A pilot batch of the most efficient sorbent was tested in a pilot plant under conditions similar to the operating parameters of commercial natural gas–gasoline processing plants. The dynamic adsorption capacity of a carbon sorbent was higher than this characteristic of imported silica gel, which is commonly used for natural gas stripping. DOI: 10.3103/S0361521908060049
The use of carbon sorbents in the sorption recovery of gasoline fraction (stripping) from natural gas and associated petroleum gas for the production of stable natural-gas gasoline and the long distillate of light hydrocarbons is of considerable current interest [1–3]. It is well known that large amounts of gasoline fraction are lost in the course of the production, transportation, processing, and storage of hydrocarbon materials. In the Russian Federation, only about 40% of associated petroleum gas (~12 billion m3) is processed to obtain raw materials for petroleum chemical plants and liquefied gas for human needs; another 40% is combusted at thermal power plants without processing, and the other 20% is burned in flares at fields. The loss of profit from a billion cubic meters of unprocessed associated petroleum gas is equivalent to the loss of the mass of commodities to the amount of $270 million; in this case, budget losses are about $35 million [4, 5]. In addition to the business factor, the ecological factor is also responsible for the necessity of solving the problem of hydrocarbon losses. As a result of the combustion of associated petroleum gas, ~400000 t of harmful substances (carbon oxides, nitrogen oxides, hydrocarbons, and soot) has been released into the atmosphere per annum. Petroleum refineries and petrochemical plants released more than 1.1 million tons of hydrocarbons per annum [6]. Even greater losses occur in the storage and use of gasoline fractions. A number of methods are used for the recovery of gasoline fractions from gas atmospheres. In this case, the adsorption stripping of hydrocarbon gases with the use of carbon adsorbents is the most efficient method. The adsorption refining of natural gas is commonly used worldwide for the manufacture of gasoline fractions and the preparation of gas for transportation; more than 200 stripping plants with high degrees of gasoline recovery are in operation [7, 8]. Natural gas–gasoline
processing plants based on pressure swing adsorption (PSA) are most efficient. The main difference of this technique from traditional adsorption separation processes is that the duration of an adsorption stage lasts a few minutes and the heating of a sorbent is not used at the stage of regeneration [9, 10]. The operating experience of closed-cycle plants shows that short-cycle stripping plants provide almost complete recovery of hexane and polar hydrocarbons. In this case, the total recovery of high-molecular-weight hydrocarbons is as high as 70% and depends on the initial gas composition and the duration of an adsorption cycle; as a rule, it increases with shortening this cycle [11]. Various solid absorbents, such as active carbon or carbon adsorbents, silica gel, zeolite, and activated alumina gel, are used in the adsorption recovery of gasoline fractions from gas [12]. Active carbons exhibit a maximum adsorption capacity for hydrocarbon vapors; they are superior to all the other adsorbents in the recovery of light hydrocarbons, such as propane and butane. Carbon sorbents, which are hydrophobic, do not adsorb water vapor from gases, and they do not simultaneously dry gases as, for example, mineral sorbents. Carbon sorbents are much superior to mineral sorbents in terms of selectivity and, finally, the efficiency of gas stripping. Carbon sorbents are efficient absorbents for gasoline fractions from not only natural and associated petroleum gases but also any gas mixtures. Thus, crushed active carbon from Rochester is used in the adsorbers of automobiles equipped with gasoline vapor trapping systems in the United States. The adsorbent used in the gasoline vapor trapping systems should exhibit sufficiently high mechanical strength and capacity for gasoline vapor; it should be easy to regenerate and should not adsorbed moisture from air. Under these conditions, the gasoline vapor trapping systems
349
350
NOSKOVA et al.
Table 1. Treatment conditions and quality characteristics of carbonization and activation products Treatment conditions T, °C
τ, min
Combustion loss, α, %
DKM-k DKM-s
800 800
115 45
0 34.5
0.55 0.36
DSO-k DSO-s
800 800
110 60
0 38.5
DDB-k DDB-s
700 700
100 60
DBU-k DBU-s
750 750
GGU-k GGU-s
800 800
Sample
γa, g/cm3
Π, %
Pore volume, cm3/g VΣ
Ws
Vh
Vmacro
93.7 84.7
0.58 1.00
023 0.52
–
0.35 0.48
0.57 0.35
97.9 90.1
0.51 1.03
0.20 0.57
– 0.55
0.31 0.46
0 40.2
0.25 0.15
96.0 75.7
0.83 1.63
0.07 0.35
–
0.76 1.28
110 75
0 36.1
0.72 0.46
89.4 77.6
0.22 0.85
0.08 0.55
– 0.50
0.14 0.04
120 90
0 41.1
0.73 0.43
96.0 92.2
0.29 0.97
0.17 0.54
– 0.52
0.12 0.43
Note: DKM, DSO, DDB, and DBU refer to the crushed samples from olive pits, walnut shells, birch woods, and brown coal, respectively; GGU refers to the granulated samples from gas coal; k and s refer to carbonization products and sorbents, respectively; T is the treatment (carbonization or activation) temperature; τ is the process duration; α is the degree of carbon loss upon activation; γa is the apparent density; Π is the abrasion strength; VΣ is the total pore volume; Ws is the volume of sorbing pores on a benzene basis; Vh is the capacity for heptane; and Vmacro is the macropore volume.
fix 90–95% hydrocarbon vapors released in the operation of an automobile [13]. Comparative tests of American active carbon from Rochester and a number of Russian active carbons were performed at the NPO Neorganika; these tests demonstrated the advantages of the American active carbon in all of the parameters other than strength [14]. An analysis of published data and patents demonstrated that efficient domestic hydrophobic sorbents for the stripping of natural and associated petroleum gases and the trapping of gasoline vapor with high absorption capacity, strength, and regenerability are currently not available. The aim of this study was to prepare carbon sorbents based on various raw materials and to test their absorption capacity for gasoline hydrocarbons under static and dynamic conditions. The following raw materials for the synthesis of carbon sorbents were studied: Kansk-Achinsk brown coals from the 2B group, 2G gas coals from the Kuznetsk Basin, birch woods, walnut shells, and olive pits. Spherical granulated carbon sorbents were prepared based on the gas coals; the other raw materials were converted into crushed carbon sorbents (with irregularly shaped particles) [15–18]. Raw materials for the preparation of crushed carbon sorbents were reduced to a particle size of ≤5.0 mm and sieved to separate a working fraction; gas coal was reduced to powder with a particle size of <100 µm, and the pulverized coal was granulated on a plate granulator with a 2% aqueous solution of sulfite waste liquor. The raw materials thus prepared were dried at 105 ± 5°C to a residual moisture content of <10 wt % and subjected
to carbonization and activation in an electrically heated quartz reactor with automatically regulated rate of heating and final treatment temperature. The fixed bed of a material was blown from bottom to top with an inert gas (nitrogen) during carbonization or with a mixture of an inert gas and water vapor (activating agent) in a volume ratio of 1 : 1 during activation. Carbonization was performed under conditions found by thermogravimetric studies at a temperature required for the complete degradation of raw materials. Next, the carbonization products were activated at activation temperature and time required for the production of carbon sorbents with the highest structure and strength characteristics. The strength, apparent density, and total pore volume (moisture capacity) were determined using standard procedures [19]. The volumes of sorbing pores (total micropores and mesopores) were measured with methane, benzene, and heptane using a gravimetric desiccator method with the saturation of the pore volume of a sorbent sample with adsorbate vapor at a constant temperature [12]. The macropore volume was calculated by difference between the total pore volume and the sorbing pore volume, and the combustion loss upon activation was found from a change in the apparent density on going from carbonization products to carbon sorbents. Table 1 summarizes carbonization and activation conditions and the quality characteristics of the resulting carbon sorbents. All of the carbon sorbent samples prepared exhibited high sorption space volume Ws and capacity for heptane Vh; in this case, the crushed sorbent based on birch wood was characterized by the lowest value of Ws and the highest macropore volume at an insufficiently SOLID FUEL CHEMISTRY
Vol. 42
No. 6
2008
ADSORPTION METHOD FOR THE RECOVERY OF HYDROCARBONS
351
Table 2. Pore structure parameters of carbon sorbents Sorbent sample DKM DSO DDB DBU GGU
Pore structure parameter SBET, m2/g
Smeso, m2/g
W0, cm3/g
Ws, cm3/g
E0, kJ/m
x0, nm
960 1250 830 900 1000
145 110 40 390 320
0.36 0.50 0.33 0.42 0.45
0.52 0.57 0.35 0.55 0.54
20.2 25.2 23.9 21.1 18.5
0.61 0.45 0.50 0.56 0.76
Note: SBET is the BET specific surface area, Smeso is the specific surface areas of mesopores, W0 is the limiting adsorption in micropores, Ws is the limiting volume of sorption space, E0 is the characteristic energy of adsorption, and x0 is the radius (halfwidth) of micropores.
high strength. The structure parameters of the test sorbents (Table 2) were calculated from the adsorption isotherms of nitrogen on these samples; these isotherms were measured on a Gravimat-4303 automated vacuum gravimetric unit at 77 K [20]. As can be seen in Table 2, all of the carbon sorbents prepared from various raw materials under different treatment conditions exhibited developed pore structures, which mainly consisted of micropores. The sorbent based on wood was purely microporous; the carbon sorbents based on plant raw materials (olive pits and walnut shells) were characterized by high specific surface areas of micropores (in the latter case, this characteristic was higher) with a developed mesoporosity; the sorbents from brown and gas coals exhibited developed structures of both of the pore types. The sorbents were tested in the process of gasoline vapor adsorption; in this case, the kinetics of saturation of carbon sorbent samples for a day was determined by a desiccator method using the precision weighing of a sorbent at regular adsorption time intervals at a constant temperature. Figure 1 shows the kinetic curves of gasoline adsorption on the carbon sorbents. An analysis of the experimental data suggests a direct relationship of the adsorption capacity of carbon sorbents for gasoline with the absorption capacity for heptane, which is given in Table 1. In turn, the absorption capacity for heptane directly depended on the limiting sorption space volume of a given carbon sorbent. It can be seen that three carbon sorbents with the greatest volumes of sorbing pores—DSO, GGU, and DBU—are characterized by the highest gasoline adsorption properties. In this case, the shapes of kinetic curves for GGU and DBU were identical; this fact may be explained by a similarity of their structure parameters. These sorbent samples were tested in the heptane adsorption–desorption process. The process of heptane adsorption from a flow of a methane–heptane mixture with a specified heptane content, which was passed through a sorbent bed at a flow rate of ~20 ml/min at room temperature, was monitored by on-line gas chromatography on a Kristall-4000 chromatograph. The completion of the adsorption process was determined chromatographically by detecting SOLID FUEL CHEMISTRY
Vol. 42
No. 6
2008
a breakthrough of heptane. The desorption of heptane from the carbon adsorbent was performed under temperature-programmed conditions at a heating rate of 2 K/min over a range of 20–290°C in a flow of argon; the completion of the process was detected by chromatography. The desorption process was considered complete at a heptane content of the test mixture of no higher than 0.01% on an initial basis. Table 3 and Figs. 2 and 3 show the conditions of adsorption–desorption processes and the results of tests. The above data indicate that the DBU and GGU carbon adsorbents exhibited similar adsorption capacities for heptane. However, a considerable difference was observed in important performance characteristics, which are directly related to the energy intensity of the process, such as the rate of adsorption and the degree of desorption; these characteristics were better in the spherically shaped granulated GGU sorbent. The larger scale tests of a prepared pilot batch of the spherical GGU sorbent were performed in a pilot Capacity for gasoline, cm3/g 0.4 1 2
0.3 3 4 0.2
5
0.1
0
4
8
12
16
20
24 Time, h
Fig. 1. Kinetic curves of gasoline adsorption on carbon sorbents: (1) DSO, (2) GGU, (3) DBU, (4) DKM, and (5) DDB.
352
NOSKOVA et al.
Table 3. Process conditions and the results of heptane adsorption–desorption on carbon sorbents Carbon sorbent sample
Process conditions and results
DBU
GGU
Adsorption: temperature, °C CH4–C7 flow rate, ml/min linear velocity, cm/min average rate of C7 adsorption, ml min–1 g–1 Dynamic capacity, g C7/(g sorbent)
22–24 17 44 0.1 0.21
22–24 18 46 0.2 0.23
Desorption: heating rate, K/min temperature range, °C argon flow rate, ml/min Degree of desorption, wt %
2 20–285 17–20 ~91
2 20–290 17–20 ~98
adsorption plant at VNIIGAZ under conditions similar to the operating parameters of commercial natural gas driers and cleaners: a pressure of 5 MPa, an adsorption temperature of 20–28°C, and a linear gas flow rate of 0.06 m/s. The pilot plant consisted of a unit for the preparation and compression of an initial gas mixture, an adsorption unit of two adsorbers with an inner diameter of 50 mm and a height of 3000 mm with external electric heating, and a unit for gas sampling and analysis. n-Heptane was used as a model hydrocarbon, which was admixed to natural gas in a concentration of 3.0– 3.5 g/m3; the gas mixture was compressed to a pressure of 5.0 MPa and supplied to the adsorber from bottom to top. The concentrations of n-heptane in the gas at the adsorber inlet and outlet were determined on an Agilent 3000A gas chromatograph with a thermal conductivity detector and an OV-1 column (8 m × 0.15 mm × 2.0 µm). C7, ml 60
The carbon sorbent was regenerated with dried gas at 200°C for 3 h. The regeneration cycle including heating and cooling was about 6 h. The breakthrough concentration of n-heptane in gas stripping was 10% on an initial basis. The dynamic adsorption capacity (wt %) of the carbon sorbent was determined from the equation A = τbQC/G, where τb is the time of the protective action of the sorbent until a breakthrough, h; Q is the flow rate of the initial gas mixture, m3/h; C is the amount of the component absorbed by the sorbent, g/m3; and G is the weight of the regenerated adsorbent, g. Table 4 summarizes the results of the tests. For comparison, the best imported silica gel KC-Trockenperlen H, which is widely used in gas–gasoline processing, and the hydrophobic spherical sorbent were tested under the same conditions. The dynamic adsorption C7, ml 60
2 50
50
1
1 2
40
40
30
30
20
20
10 10 0 0
100
200
300
400
50
100
150
200
250
300
Time, min
500 Time, min
Fig. 2. Kinetics of heptane adsorption and desorption on the DBU carbon adsorbent: (1) adsorption and (2) desorption.
Fig. 3. Kinetics of heptane adsorption and desorption on the GGU carbon adsorbent: (1) adsorption and (2) desorption. SOLID FUEL CHEMISTRY
Vol. 42
No. 6
2008
ADSORPTION METHOD FOR THE RECOVERY OF HYDROCARBONS
353
Table 4. Conditions and results of natural gas stripping on the spherical GGU sorbent Sorbent weight, g
Gas flow rate, m3/h
650
34.51
Amount of absorbed heptane (ρ = 0.72 g/cm3) cm3
g
210
151.2
capacity of the silica gel for n-heptane vapor was 5.7 g/(100 g), which is lower than this characteristic of the carbon sorbent by a factor of 4. Thus, the spherical GGU sorbent can be used in the stripping of natural gas and associated petroleum gas. REFERENCES 1. Afanas’ev, A.I., Afanas’ev, Yu.V., Bekirov, T.M., et al., Tekhnologii pererabotki prirodnogo gaza i kondensata: Spravochnik (Technologies for Natural Gas and Condensate Processing), Moscow: Nedra, 2002. 2. Klimenko, A.P., Szhizhennye uglevodorodnye gazy (Liquefied Hydrocarbon Gases), Moscow: Nedra, 1974. 3. Berlin, M.A., Gorschenkov, V.G., and Volkov, N.P., Pererabotka neftyanykh i prirodnykh gazov (Oil Gas and Natural Gas Processing), Moscow: Nedra, 1981. 4. Baraz, V.I., Dobycha neftyanogo gaza (Oil Gas Production), Moscow: Nedra, 1983. 5. Rizaev, R.G., Guseinov, Ch.S., and Sheinin, V.E., Neft. Khoz., 1994, nos. 11–12, p. 37. 6. Shimkovich, V.V., Neftepererabatyvayushchaya i neftekhimicheskaya promyshlennost’, Ser. Okhrana okruzhayushchei sredy (Oil Refining and Petrochemical Industry, Ser. Environmental protection), Moscow: TsNIITEneftekhim, 1996, issue 2. 7. Braginskii, O.B. and Shlikhter, E.B., Perspektivy khimicheskoi pererabotki prirodnogo i poputnogo gazov (Prospects for the Chemical Processing of Natural and Associated Gases), Moscow: TsNIITEneftekhim, 1991.
SOLID FUEL CHEMISTRY
Vol. 42
No. 6
2008
Time to breakthrough, h 2.0
Dynamic adsorption capacity for n-heptane g/(100 g)
g/(100 ml)
23.3
10.0
8. Chebotarev, V.V. and Khafizov, A.R., Adsorbtsionnoe razdelenie skvazhinnoi produktsii (Adsorption Separation of Well Products), Ufa: Izd. UGNTU, 1999. 9. RF Patent 2108363, 1998, Byull. Izobret., 2000, no. 23. 10. Breshchenko, E.M., Neft. Khoz., 1992, no. 1, p. 39. 11. Shakhov, A.D., Cand. Sci. (Eng.) Dissertation, Moscow: Gubkin Univ. of Oil and Gas, 2001. 12. Kel’tsev, N.V., Osnovy adsorbtsionnoi tekhniki (Basics of Adsorption Instrumentation), Moscow: Khimiya, 1984. 13. Zashchita atmosfery ot promyshlennykh zagryaznenii: Spravochnik, (Protection of the Atmosphere from Industrial Pollution: A Handbook), Kalvert, S.I., Ed., Moscow: Metallurgiya, 1988, part 1. 14. Mukhin, V.M., Tarasov, A.V., and Klushin, V.N., Aktivnye ugli Rossii (Russian Active Carbons), 2000. 15. Perederii, M.A. and Surinova, S.I., USSR Inventor’s Certificate no. 1528729, 1987, Byull. Izobret., 1989. no. 46. 16. Perederii, M.A., Sirotin, P.A., and Kazakov, V.A., Khim. Tverd. Topl. (Moscow), 2002, no. 6, p. 19. 17. Karaseva, M.S., Noskova, Yu.A., and Perederii, M.A., Khim. Tverd. Topl. (Moscow), 2006, no. 5, p. 50. 18. Perederii, M.A. and Noskova, Yu.A., Khim. Tverd. Topl. (Moscow), 2008, no. 4, p. 20. 19. Kolyshkin, D.A. and Mikhailova, K.K., Aktivnye ugli. Spravochnik (Active Carbons: A Handbook), Leningrad: Khimiya, 1972. 20. Voloshchuk, A.M., Khokhlova, G.P., and Kryazhev, Yu.G., Khim. Tverd. Topl. (Moscow), 2005, no. 4, p. 85.
