Absorption and Humidification. Unit Operations Review

also in the Slavic countries, with chief emphasis, of course, on Soviet Russia. Some original Japanese reports were consulted for this review; however...
0 downloads 0 Views 467KB Size
an

( X / E q Unit Operations Review

Absorption

and Humidification

by Max Leva, Consulting Chemical Engineer, Pittsburgh, Pa. Chin-Yung Wen, West Virginia University, Morgantown, W . Vu.

N e w expressions for HTU and HETP mark progress in design correlation and calculation of tower capacity

DURING

the past year the number of studies in English and other western languages has been, as in previous years, considerable. However, a n increase in activity has been noted in Japan and also in the Slavic countries, with chief emphasis, of course, on Soviet Russia. Some original Japanese reports were consulted for this review; however, Russian studies could only be covered by way of condensed abstracts.

N e w Equipment A new type of absorber was described (4), intended for dissolving hydrogen chloride or hydrogen fluoride in water. T h e new unit resembles a Poly-Bloc heat exchanger but is of somewhat different shape. Formation of a stable liquid film is claimed when the unit is operated a t very low liquid rates. A new tower for absorbing nitrogen oxides in nitric acid manufacture (7) is composed of horizontal trays constructed such that a portion of the acid is held in place. Through contact of the gas with this acid, reoxidation of nitrogen oxide is said to be accelerated.

liquid Distribution and Wetted Packing Area Liquid flow in ring-packed beds was investigated by Greenaway (25). T h e columns (3- and 4-inch diameter) had “walls” of wire mesh, precluding gas flow. T h e study evaluated distribution of flow through the inside and over the outside of the rings. Flow through the ring interior decreased from 29 to 12% as ring sizes increased from 0.5- io 1.25-inch nominal size. T h e liquid flow mechanism was also studied by Sheridan and Donald ( 6 1 ) . Isolated pairs of equal and unequal spheres were used either unconfined or held in a vessel. Flow passed from one sphere to the other in accord with certain requirements of sphere diameter and water flow rate. When spheres were confined, the wall was denuded of liquid. Collecting zones specified in earlier research were too large to indicate this. Liquid distribution in wood and carbon grids was studied by Mullin (47). T h e extent of wetting with water

340

depended on rate as well as slot diameter and pitch. T h e earlier proposed idea of minimum effective liquid rate (MELR) was generally confirmed. Air flow (about 1800 pounds/square foothour) seemed to have little eRect on liquid spread. Davidson and coworkers ( 7 7, 72) investigated liquid hold-up and effect on film coefficients in packed towers. T h e preliminary study involved absorption of carbon dioxide into water over a string of balls, noting hold-up and capacity data. I n a simplified model for packed beds, individual packing elements provide a multitude of vertical and inclined surfaces of characteristic lengths. Literature capacity data could be predicted by typical correlations. A column containing a row of ceramic disks was used (50) for ascertaining the effect of ammonia absorption on capacity data and wetted area. A minim u m wetted area was provided when the inert gas contained about 40% ammonia. Wetted area did, however, increase with increasing ammonia content in the liquid, but absorption coefficients decreased pointing to possible variations in film flow characteristics.

Falling Film, Wetted Wall, and Disk Column Data Operating a column comprising 27 disks, Warner (67) demonstrated the importance of gas-liquid relative velocity on absorption coefficient and how this may mask the effect of diffusivity if not accounted for. For the tests, diffusivity varied about eightfold. A disk column was used (65) to absorb iodine vapor in water and various salt solutions. Depending on the liquid chosen, liquidfilm resistance varied significantly. Gas velocity has a n effect on carbon dioxide absorption by water in a disk column only a t low gas rates (44). At least two separate liquid effects were noted. A relation was also derived (4.5) for physical absorption in a disk column, based on the Higbie theory. Theoretically estimated absorption rates for several temperatures were lower than actual values; this was attributed to turbulence in the liquid film passing over the disks. Ammonia was absorbed

INDUSTRIAL AND ENGINEERING CHEMISTRY

in water and weak ammonium hydroxide (27) by passing the liquid down the wall of a tube (0.5-inch diameter, 0.5 to 4 feet long) contained in a n atmosphere of the gas. Owing to the presence of ripples, absorption data were about five times higher than predicted by penetration theory. Other experiments with \vetted wall columns \rere made by Hikita (29). First investigated was the effect of wetting agent on carbon dioxide absorption; earlier findings of Pigford were essentially substantiated. I n a n extended study (32), a wetted wall column was used with a variety of gases and solvents to obtain a general correlation of liquid phase mass transfer coefficients. Finally (31), the dissolution rates of iron and zinc from a pipe wall into dilute sulfuric acid and aqueous iodine-potassium iodide were noted. Data compared well with theoretically calculated values, assuming unsteady state of diffusion and parabolic velocity distribution in the film. Carbon dioxide absorption by jets and falling films of caustic solutions was studied ( 4 9 ) . Data agreed with theory. Nitrogen dioxide was absorbed in a !vetted wall column (13). Flow was concurrent, and gas flow in the core was laminar. A model stipulating chemical equilibrium in the gas phase between nitrogen dioxide and tetroxide, solution of nitrogen tetroxide a t the interface, and a simultaneous pseudo first-order reaction bet~veennitrogen tetroxide and water permitted data analysis. A new column (43) consists of two telescoped glass tubes, with a n annulus between them. T h e inner tube presumably carries the liquid phase, as a film can be rotated. In effect this is a wetted wall column permitting variation in vapor flow requirement without affecting throughput. Very low hold-up, pressure drop, and H E T P were reported.

Design CorrelationTower Capacity Data For gas-film controlled operations Whitt (68) has proposed correlating H T U data bv

an where d is the nominal ring size, L the irrigating mass velocity, and ,u the liquid viscosity. T h e fact that the expression is dimensionally homogeneous is noteworthy; it is similar to that proposed earlier by Hands and Whitt. I n another proposal ( 7 4 :

( +

”>

HETPl = HETP? 1 0.7 log -2 D2 where D1 and D Zare respective column diameters. An additional proposal is: HETP‘ = HETPz

4;-

where dl and d i are nominal Raschig ring sizes. Formulas were verified by data covering Raschig rings of 15- and 25-mm. diameter and columns of 150and 400-mm. diameter. T h e formulas apply to 70% flooding. Systems considered were benzene-methylcyclohexane and toluene-methylcyclohexane. Simultaneous heat and mass transfer in liquid-gas contactors was considered (46), and a method of calculation was suggested by employing enthalpy difference of the gas as driving force. T h e change of condition of the gas through the tower may be followed, and the required absorption is ascertained.

Cooling Towers, Psychrometry, Humidity Working with a 6-feet-high, rectangular, grid-packed tower of 1 square foot cross-section, liquid film resistance to heat transfer was not negligible (66). For analysis of heat and mass transfer the Mickley enthalpy method was proposed. Heat transfer data were empirically correlated in terms of flows and tower height. O n the basis of water cooling data pertaining to flow through a tower stacked with 4-inch ceramic rings, Cribb (10) also concluded that there is a liquid film resistance. However it was of such an order of magnitude that it may be neglected for all practical purposes in direct contact cooler design. Equations were presented ( I ) for the psychrometric analysis of complex airvapor mixtures. The proposed formulas assume special importance in solvent recovery.

Sieve Plates, Grids, Foams Absorption of ammonia by ammoniacal copper solution was studied (55) on a single sieve plate. As ammonia content in the solution increased, absorption coefficient decreased. The effect of foam height on a sieve plate during absorption of carbon dioxide was also investigated (56, 57). Again a single plate column was used with 3 N sodium hydroxide as absorbent. Expressions for foam height, absorption coefficient, and plate efficiency were given. A pressure drop study on sieve plates in relation to foam height at low gas

rates was made by Yokota (69). An empirical expression for total pressure drop in terms of individual component pressure drops (because of flow through dry plates), liquid head on plate, and liquid-metal surface tension was suggested. Oxygen desorption from water was studied (76) in a grid tray tower. Grids were constructed of 3/le-inch bar stock with slit widths ranging from ‘ / a to 3/8 inch. Free area was 13 to 23.5%. Carbon dioxide absorption by water was studied (28) in a 10-inch diameter steel column with a single bubble cap plate which permitted design variations in the number of caps. Transfer units achieved were correlated in terms of various dimensionless numbers, being functions of stream properties and plate and tower parameters.

Agitated and Mechanically Operated Vessels Removal of nitrogen oxides from gas mixtures was examined (54) with an agitated vessel into the base of which the gas was conducted through a perforated dispersing disk. This was more effective than a typical bubble cap or wetted wall unit. High removal efficiency was believed to be due to chemical interactions not facilitated in other equipment. Using Higbie’s penetration and static film model. Lightfoot (42) solved differential equations expressing absorption rate of a sparingly soluble gas in a n agitated tank with a simultaneous irreversible first-order reaction. Mass transfer in gas-liquid systems with and without mechanical agitation was studied by Calderbank (5). When introducing gas-bubble swarms into agitated liquid the mass transfer coefficient was independent of bubble size and velocity in the liquid. A decrease of bubble size causes a proportional decrease in slip velocity, so that effective exposure time remains unaltered.

Packed Tower Studies T h e dependence between liquid-side mass transfer cocfficient and physical properties in packed towers was studied (57) using carbon dioxide absorbed in ethyl alcohol, methyl alcohol, benzene, and other solvents. Capacity data were roughly inversely proportional to the square root of the Schmidt number. Dependency results between capacity data and liquid rate indicated a critical point a t which dependence changes, which was different for water than for the remaining solvents. The effect of packed height on liquidphase H T U was investigated (30). Height of the column packed with Raschig rings varied from 0.05 to 1.50 meters. Earlier workers had reported that kL ah-113 or kL =h-O.l9, where h is packed height. In this study, kL did

q v Unit Operations Review

not depend on bed height if initial liquid distribution was uniform. With initial uneven distribution a n effect of packed height on k L was noted. Mixing characteristics of a packed tower and effect on carbon dioxide absorption by water were considered (53). Results indicated that absorption proceeds rapidly and that with short columns the error caused by longitudinal mixing affects determination of liquid film capacity coefficient. I n an investigation (52) of carbon dioxide and hydrogen absorption by water in packed towers, mass transfer coefficients (liquid side) were correlated by a n expression based on the two-film theory as well as penetration theory. Absorption of ammonia, sulfur dioxide, carbon dioxide, ethyl alcohol, and acetone from air by water was studied (35) using a small ring-packed tower. Overall absorption coefficients were expressed in terms of physical properties, flows, and tower parameters. A packed bed was considered (77) as a series of perfect mixers, and a factor was derived for correcting the log-mean driving potential in plug-flow equations for fluid particle heat and mass transfer. A tower of about 23-cm. diameter packed 55 cm. high with 16-mm. Raschig rings was employed in air-water experiments by Schrader (59). Pressure drops and heat and mass transfer were measured in countercurrent flow. For all aspects, three different ranges were found which corresponded with each other; the higher the pressure drop, the higher were mass and heat transfer. One study (63) dealt with assessing the effect of solute concentration on gasphase mass transfer rate. A short, 4inch-diameter column, packed with 0.5inch Raschig rings, was employed. I t was suggested that previously reported mass transfer correlations be corrected by including the quantity ( P B M / p T ) 2 1 3 , where P B M is mean partial pressure of inert gas in the gas phase and P T is total system pressure.

Interfacial and Other Fundamental Studies

It has been commonly assumed that when applying either film or penetration theory, phase equilibrium exists a t the gas-liquid interface. T h e validity of this assumption has been tested by various researchers, but conflicting data have resulted because of lack of understanding of interfacial fluid dynamics and inaccurate experimental techniques. Scriven and Pigford (60) measured absorption rates of carbon dioxide into a short laminar water jet, noting the change in average concentration. Excellent agreement was reported with predictions based on an unsteady state VOL. 52, NO. 4

APRIL 1960

341

in

Unit Operations Review

diffusion theory, assuming interfacial equilibrium. They concluded that equilibrium prevails a t a newly formed carbon dioxide-water interface. Small quantities of surface active materials reduced the absorption rate greatly. Similar interfacial resistance studies were reported (8, 58), when using carbon dioxide and water, as well as oxygen and water. Laminar jets were passed through the pure gases. Average absorption rates were compared with rates calculated from Fick’s l a w for unsteady state diffusion into a liquid when in rodlike flow. I n contrast to Scriven and Pigford, reported values were lower than those estimated according to theory and no interfacial resistance. Hence, there must have been some interfacial resistance, small enough to be neglected a n d justify the assumption of interfacial equilibrium in most gas absorbers. Absorption of carbon dioxide into quiescent water was examined (26) using a n intermediate technique. With pure water as solvent there was n o measurable resistance to solution a t the interface. However, with surface active agents present there was a measurable resistance. A new theory was proposed

Additional Publications on Absorption and Humidification Subject Ref. Floating-bed scrubber with bed of (6) mobile spheres Mechanical absorbers with vertical or horizontal shaft Nomograph for estimating packed tower height in gas absorption Absorption of CO from synthesis gas by Cu in solutions Equations for predicting distribution of component in multicomponent mixture Psychrometric data for air-ethyl acetate Determination of humidity Viscosities of air-H*0 mixtures for 1 to 2 atm. and 0-350” C. Absorption of NHa by H20 0 2 desorption from H?O Evaluation of mass transfer on perforated plates Evaluation of mass transfer on grid plates Effect of foam layer on bubble absorption of SO2 Liquid mixing over bubble cap plates and sieve plates Effect of gas and liquid flow rates on absorption of COz and HzO in bubble columns Absorption of SO2 and CSz in diphenylmethane hydrocarbons Design calculations of packed absorption and rectification towers Absorption from gas to liquid by passing Nn containing MeOH up an aqueous column of 10% MeOH Analysis of absorption rate in horizontal mechanical absorbers Review of mechanical absorption equipment Gas purification processes

342

(62) for mass transfer with a simultaneous fast irreversible reaction. Data observed with benzoic acid and water indicate a direct proportionality between Stanton number and friction factor a t high Prandtl or Schmidt numbers, not in agreement with the result predicted from analogy among heat, mass, and momentum transfer.

Miscellaneous Studies and Reviews Studies of mass transfer to liquid drops, suspended in certain gas atmospheres, were made (20, 27) in a wind tunnel. Systems included carbon dioxide and water, as well as water vapor into glycols and a n amine. Mass transfer i n gas lifts was investigated (22). Carbon dioxide was desorbed from an aqueous solution using air.

literature Cited (1) Alyavdin, N. A., Nauch. Doklady Vysshei Shkoly, Khim. i Khim. Tekhnol. 1958, pp. 180-4. (2) Blumer, U. S. Patent 2,848,306 (.4ug. 19, 1958). (3) Bretsznajder, S., Kawecki, W., others, PrzemysE. Chem. 35, 564-5 (1956). (4) Brit. Chem. Eng. 4,27 (1959). (5) Calderbank, P. H., Trans. Znst. Chem. Engrs. (London) 37, 173-85 (1959). (6) Chem. Eng. 6 6 , 106 (Dec. 14, 1959). (7) Chem. Zng. Tech. 31, 61 (1959). (8) Chiang, S. H., Toor, H. L., A.I.Ch.E. Journal 5 , 165-8 (1959). (9) Ciborowski, J., Wronski, S., Chem. Stosowana 2, 147-52 (1958). (10) Cribb, G., Brit. Chem. Eng. 4, 264 11959’1. ~ ~ . . - , (11) Davidson, J. F., Trans. Inst. Chem. Engrs. (London) 37, 131-6 (1959). (12) Davidson, J. F., Cullen, E. J., others, Zbid.,37, 122-30 11959). (13) Dekker, W. A,; Snoeck, E., Kramers, H., Chem. Eng. Sci.11, 61 (1959). (14) Delzenne, A . . Gdnie chim. 82. No. 3. 53-6 (1959): ’ (15) Egalon, R., Nuez, J., Ibid., 82, No. 5. 117 119591. (16j Ellis; S. R. M., Khan, A4.S., Birmingham Univ. Chem. Engr. 9, 61-72 (1958). (17) Epstein, N., Can. J. Chem. Eng. 36, 210 (1958). (18) Fan, L.-T., Petrol. Rejner 37, No. 9, 346 (1958). (19) Ganz, S. N., Lokshin, M. A., Zhur. Priklad Khim.32, 1477-83 (1959). (20) Garner, F. H., Kendrick, P., Trans. Inst. Chem. Engrs. (London) 37, 155-61 (1959). (21) Garner, F. H., Lane, J. J., Zbid., 37, 162-72 (1959). (22) Gasyuk, G. N., Bol’shakov, A. G., others, Zhur. Priklad. Khim 31, 1019-25 (19581. -(23) Gd’perin, N. I., Matveev, I. G., Bil’shau, K. V., Zhur. Priklad Khim. 31,1323-32 (1958). (24) Gilbert, T. J., Chem. Eng. Sci. I O , 243 (1959). (25) Greenaway, D., Ibid.,I O , 197 (1959). (26) Harvey, E. A., Smith, W., Ibid., 10,274 (1959). (27) Haselden, G. G., Malaty, S. A., Trans. Inst. Chem. Engrs. (London) 37, 13746 11959). (28) Hayakawa, T., Fujita, S., Kagaku Kogaku 23, 25 (1959).

INDUSTRIAL A N D ENGINEERING CHEMISTRY

\ -

(29) Hikita, Harno, Zbid., 23, 23 (1959). (30) Hikita, H., Kataoka, T., Kondo, H., Zbid., 23, 520 (1959). (31) Hikita, H., Nakanishi, K., Asai, S., Zbid., 23, 28 (1959). (32) Hikita, H., Nakanishi, K., Kataoka, T., Zbid., 23, 459 (1959). (33) Isomura, I., Mizumaki, T., Ibid., 23, 467 11959). (34) Johnson, A. I., Bowman, C. W., Can. J . Chem. Eng. 36, 253-61 (1958). (35) Kafarov, V. V., Trofimov, V. I., Zhur. Priklad Khim. 31, 1809-16 (1958). (36) Koch, R., Przemysl Chem. 35, 247-50 (1956). ( 3 j j ZbEh., 37, 766-9 (1958). (38) Kohl, A. L., Riesenfeld, F. C., Chem. Eng. 66,159-62 (June 15, 1959). (39) Kroopelin, H., Prott, E., Naturwissenschaften 45, 333-4’(19585. . (40) Kuz’minykh, I. N., Rodionov, A. I., Zhur. PrikladKhim. 32. 1279-85 11959). , , (41) Ibid., pp. 1489-95.‘ (42) Lightfoot, E. N., A.I.Ch.E. Journal 4, 499-500 119581. (43) Macleod, N., Matterson, K. J., Chem. Eng. Sci.I O , 254 (1759). (44) Mika, V., Chem. lzsty 52, 864-8 (19581 \ - - - - / -

(45) Mika, V., Collection Czechoslov. Chem. Communs. 24,2843-50 (1959). (46) Mizushina, T., Oishi, J., Hashimoto, N., Chem. Eng. Sci.I O , 31-6 (1959). (47) Mullin, J. W., Trans. Inst. Chem. Engrs. 37, 89-107 (1959). (48) Mustacchi, C., Passino, R., Riv. combustibili 12, 581-94 (1958). (49) Nijsing, R. A. T. O., Hendriksz, R . H., Kramers, H., Chem. Eng. Sci. 10, 88 (1959). (50) Norman, W. S., Solomon, B. K., Trans. Znst. Chem. Engrs. 37, 237-43 (1959). (51) Onda, K., Sada, E., Kagaku Kogaku 23, 220 (1959). (52) Onda, K., Sada, E., Murase, Y., A.I.Ch.E. Journal 5, 235 (1959). (53) Otake, I., Kunugita, E., Kawabe, A., Kagaku Kogaku 23, 81 (1959). (54) Peters, M. S., Koval, E. J.. IND. ENC. CHEM.51. 577-80 11959). (55) Pozin, M. E.‘, Kopyle;, B. A., Petrova, Zhur. Priklad Khim. 31, 1007-13 N. .4., 11958). (56) Pozin, M. E., Kopylev, B. A . , Tarat, E. Ya., Zhur. Priklad. Khim.32, 1011-16 (1959). (57) Ibid.,32,1004-10 (1959). (58) Raimondi. P.. Toor, H. L., A.Z.CI1.E. ‘ Journal 5, 86192’(1959j. (59) Schrader, H., Kiiltetechnik I O , 290-5 (1958). (60) Scriven, L. E., Pigford, R. L., A.I.Ch.E. Journal 4, 439-44 (1958). (61) Sheridan, M . B., Donald, M. B., Ind. Chemist 35, 487, (1959). (62) Sherwood, T. K., Ryan, J. M., Chem. Eng. Sci.11, 81 (1959). (63) Shulman, H . L., Delaney, L. J., A.I.Ch.E. Journal 5 , 290 (1959). (64) Solomakha, G. P., Matvozov, V. I., Trudy Moskov. Znst. Khim. Mashinostroeniya 13, 53-77 (1957). (65) Taylor, R. F., Chem. Eng. Sci. I O , 68 (1959). (66) Thomas, W. J., Houston, P., Brit. Chem. Eng. 4, 160, 217 (1959). (67) Warner, N. A,, Cliem. Eng. Sci. 11, 130 (1959). (68) Whitt, F. R., Brit. Chem. Eng. 5, 395 (1959). (69) Yokota, N., Kagaku Kogaku 23, 438 11959). (70) Zhavoronkov, N. M., Malyusov, V. A., Nauch. Doklady Vysshei Shkoly Khim. i Khim. Tekhnol 1958, pp. 185-92. \ - - - - , -