Centrifugation. - Industrial & Engineering Chemistry (ACS Publications)

---

/antes 0. M a l o n e y , head o f the Department of Chemical Engineering and director of the Research Foundation a t the University o f Kansas, was born in St. Joseph, M o . , A p r i l 29, 7975. H e received his B.S. in chemical engineering in 1936 from the University of Illinois and his M.S. and Ph.D. in the same field in 7939 and 1 9 4 1 , respectively, from the Pennsylvania State College. From 1941 until December, 1945, M a l o n e y was employed b y €. 1. du Pont d e Nemours & Company, Inc., in the Engineering and Explosives Departments. U n t i l his transfer into the Explosives Department in 7943, he was responsible for following developments and acting in an advisory capacity on filtration and centrifugation. W h i l e in the Explosives Department he was stationed at t h e U n i versity of Chicago. H e assumed his present position in December, 1945. H e is a member of the American Chemical Society and of the American lnstitote of Chemical Engineers.

CENTRIFUGATION James

c

0.Maloney

EKTRIFUGAL equipment has undergone considerable development during the war years. The ultracentrifuge has ceased to be a laboratory curiosity and has been developed on a pilot plant scale for the separation of U235 by centrifugal force. Centrifuges operated by remote control have been used in the isolation of plutonium. The preparation of blood plasma, the concentration of rubber latex, and the separation of penicillin solvents are only a few of the many new uses of centrifugal equipment. Unfortunately, basic design information for the selection and scaling up of centrifuges continues to be absent from the literature. Of the recognized unit operations, this one has been the subject of few articles. With the increased application of centrifuges, it is important that there be developed for the chemical engineer sufficient information to permit him t o make a preliminary selection of equipment in this field. The application of the ultracentrifuge to the separation of the uranium isotopes as described by Smyth (f7)is undoubtedly one of the major developments in the field in the last few years. Alt,hough the specific details of the development are classified information, the general features of such an operation are prcsented by Smyth (f7) and Beams ( 2 ) . While the theoretical effectiveness of centrifuges for separating isotopes has been recognized for many years, only within the ten-year period before the war was equipment developed capable of achieving appreciable isotopic separations. The separation is obtained. as a result of centrifugal force which causes the heavier isotope t o concentrate a t the periphery of the centrifuge and the lighter one near the center. The principal advantage of such a method for separation i s that its separating efficiency depends upon the difference in niolecular weights of the two isotopes. The principal disadvantage of the method is that extremely high centrifugal forces must he developed in the machine, and this gives rise to all of the problems attendant on high-speed machinery. The equation developed by Mulliken (If) relates the ratio of the equilibrium concentration of the heavy isotope a t the periphery of a centrifuge to the concentration of the heavy isotope a t the center:

where S = separation factor a t equilibrium KO = concentration of heavy isotope at periphery K = concentration of heavy isotope at center Mz = molecular weight of heavy isotope or its compound M I = molecular weight of light isotope or its compound v = peripheral velocity of tube, cm./sec. R = gas constant, 8.3 X 104 T = absolute temperature, O K. 24

Beams ( 2 ) in 1940 reported the development of a continuous vapor feed centrifuge, 3 inches in diameter, 15 inchcs long, revolving a t a speed of 1100 revolutions per second, and generating a force of 180,000 times gravity for the separation of isotopes. Such a unit would be capable of obtaining a theoretical separation factor of 1.038 in separating the hexafluoride isotopes of uranium a t 65" C. Smyth ( 1 7 ) states that the units actually developed were 3 feet long. He reported that they were designed to obtain a countercurrent action in which the lighter material rose vertically through the center and the heavier mat,erial flowed downward a t the diameter of the tube. Such an arrangement obviously could permit achievement of more than one stage in a single unit. The units would be arranged both in parallel and series. Smyth (I?) reports that these unit,s were successfully developed through the combined efforts of Westinghouse Electric & Manufact'uring Company, St,andard Oil Development Company, and J. W. Beams a t the University of Virginia. The work, however, was limited to successful pilot plant operation. Hausen (7) made an extensive theoretical study on the expected behavior of centrifugal equipment employing countercurrent' operation for the separation of gases. Because of the limited capacity of such machines, he concluded that the principal field of application is in the separation of isotopes. In the operation of the separation plant at Hanford, Wash., for recovering plutonium, Smyth (fi') reports the use of centrifuges in the chemical separations work. Since these machines had to be essentially foolproof and operated by remote control, the design features will doubtless be of considerable interest as soon as the information is released. One of the major applications of the centrifuge to liquid-liquid scparations during the war has been in the purification of oil and fuel on naval vessels. Although the application is not new, the number of machines employed is impressive. Submarines are equipped with four of these large oil purifying units. In a submarine the proper buoyancy is maintained by filling the tanks with sea water as the oil is used. The oil must then be separated from the water before use. The removal of dirt and grit from chicle which is used in the manufacture of chewing gum is an interesting demonstration of the effectiveness of centrifuges i n separating solids from very viscous materials. I n this operation the viscosity is reduced by heating the frame and cover with steam. Machines have been constructed to concentrate virus. This operation is conducted under sterile conditions and a t 40,000 times gravity. It is believed that such concentrat,es will be of considerable assistance in studies on infantile paralysis.

INDUSTRIAL AND E N G I N E E R I N G C H E M I S T R Y

Vol. 38, No. 4

Riegel (14) presented a gencral description of centrifugals, together with some operating schedules and apparatus cost. A description of all the types of machines on the market was published in 1943 (1). The information on the fields of application, capacity, and details of the various units is as complete as is available, except for that obtainable from the trade literature a q d consulhation with manufacturers. The mechanical design and control of ultracentrifuges has been the subject of considerable study (1,6,6,10, 13, 19,20, $9). Particle size determinations have been made using the Sharples centrifuge (8). The theory of motion of material in the drum of a continuous centrifuge has been discussed mathematically (3). The application of automatic and continuous centrifuges to the separation of sugar has been reported (16,IS). Interest in the improvement of continuous centrifuges is shown by the patent literature (9, 19, 16, 18, 21). Significant contributions to the literature on centrifugal equipment are badly needed. While it is true that manufacturers of this equipment are experienced in the proper selection of centrifugals, the basic design iriformation is not available to the chemical engineering profession as a whole. Quantitative information must be available on this operation before an economic selection can be made between dryers, filters, and centrifuges. Studies are needed to develop theories for centrifugal settling and filtering based on experimental data. Such developments would be extremely useful in (a)interpreting small-scale tests, ( b ) determining the optimum operating conditions under which a centrifugc should be operated, ( c ) predicting the effects of necessary changes in operating conditions, and ( d ) reducin6 markedly the number of tests required to make a satisfactory design. Since the basic laws governing centrifugal settling and centrifugal filtering are different, different theoretical treatments are to be expected. Among the variables which should be considered in the efficient operation of centrifuges are centrifugal force, particle size, temperature of operation, viscosity, solids content of the feed, and time of spinning. There will undoubtedly be limitations to the application of any theory developed. It will have to be applied to the particular slurry under consideration, and the test information obtained must be on the actual process material and not on material prepared elsewhere. I n the application of centrifugals to filtering solids from liquids, the solids form a cake on the filter medium; i t is on the face of this gradually increasing cake that filtration takes place. This is precisely the condition ensuing in ordinary filtration, and it is logical to employ a similar approach in the development of a preliminary theory for this operation. The following method is auggested as a possible method of attack. The differential form of the filtration equation expressing the rate of flow of filtrate at a time, 8, when a volume of filtrate, V , has been collected, is:

AP

dV

where eV/A = cake resistance r,,, = filter medium resistance A = area P = pressure c = volume of cake-forming solids per unit volume of filtrate fi = viscosity of the filtrate a = specific cake resistance

method of scale-up. The area of filtration decreases as the centrifuge fills with solids according to the relation:

A = 2nhr, where h = height of centrifuge basket r, = radius to the inner edge of cake The average area of filtration may be taken as:

A.,

=

2n

(r. + r') h

de

A2P ~CYCV

(3)

By substituting experimental data in this equation, i t is possible to determine the specific resistance of the filter cake and obtain

January, 1946

(5)

where ro = radius of centrifuge The pressure also changes during the operation:

where P = pressure N = revolutions/unit time PL = density of liquid slurry gc = gravitational conversion factor ri = radius to edge of liquid r. radius to inner edge of cake =i

The relation between the cake thickness and the amount of filtrate is given by n(rO2- rc2)h= Vc

(7)

Substitutions may now be made in Equation 3 to give:

This equation woukd be used to determine the cake resistance from the filtration rates measured on a 12-inch-diameter machine by plotting dV/& against

The slope of the curve a t any given point is l/a. T h e effect of pressure on CY is determined by plotting the instantaneous value of CY against the corresponding pressure. With this information the filtration rate to be expected at other cake thicknesses, speed of rotation, and machine diameters could be calculated and compared with information obtained from actual tests. Additional tests would be required to determine the effect of viscosity of the liquid, particle size, concentration of solids in the feed, temperature, speed, and type of material on the filtration rate. The best method for washing the cake needs to be determined as well as the effect of the length of spinning time on the residual liquid content of the cake. The initial studies could be carried out using representative materials which are usually considered as forming compressible and noncompressible cakes. A familiar example of the compressible cake is aluminum hydroxide. Sand and calcium carbonate form noncompressible filter cakes. I n the settling of solids from liquids in centrifuges, the centrifugal force acts on the solids which have a density greater than the liquid and causes them to be deposited against the solid bowl. The action occurring is similar to gravity settling, and similar equations can be expected to apply. The laws of motion for sphericaI particles settling in a fluid depend upon the Reynolds number; in Wneral, however, the Stokes law applies to most centrifugal settling:

Any consistent set of units may be employed. I n general the resistance of the filter medium may be neglected and Equation 2 becomes :

dV= -

(4)

(9) where v. = velocity of settling of particle at radius r from axis of rotation a = acceleration due to centrifugal force p s = density of solid particle ' PL = density of liquid p = viscosity of liquid (Continued on page 37)

INDUSTRIAL AND ENGINEERING CHEMISTRY

25

CENTRIFUGATION Any consistent set of units may be empl a, i s equal to (2?rNr)*/r,where N is th

centrifuge. Test data are needed to application of the Stokes law to settling i centrifuge. Informative studies could be made on the effect of particle size, feed rate, vis centration of slurry on the tive sizes and densities wou As a result of such experimental studies, sound methods of scale-up will become av@ilable. These dethods will not, in general, furnish sufficient information to make a choice between filters and centrifuges. Cost data on sucrh items as investment, labor, operation, and maintenance are also required. When all of these facts become available, it is then possible economic selection. It would appear that the engineering schools in are in an excellent position to cooperate with the m of centrifugal equipment in developing a sound theoretical background for this unit operation.

recovery of acetpse and etbyl Jcohol from dilute aqueous solutions with solvents m c h methyl isobutyl kstone, and isoamyl alcohol of butylene gly001-from dilute fermentation liquors with methylvi acetate. The use of multistage extraction systems involving mixers requires a knowledge of the degree of mixing and the power requirements ch equipment for proper design. An important contribut this end was made recently by Miller and Mann (31). They studied the power requirements and degree of mixing for seven agitatolt designs for two-phase immiscible liquid mixtuies in unbamed tanks. I n addition to extensive design and power data, their work provides a correlation of the power required to agitate systems of two immiscible liquids with the operating oharaoteristics and design of the system in terms of a plot of a power function V.S. Reynolds number. They also developed the application of the results to the design of larger scale equipment. APPLICATIONS.Steady advances in the applications of solvent, extraction to industrially important separations have been made” in the expansion and improvement of commercial installations using the older extraction processes, such as in lubricating oii and vegetable oil refining; the introduction of several industrially

LITERATURE CITED

(1) Anonymous, Chsm. & Met. Eng,, 50, 1 (2) Beams, J. W..“Science in Progress”, Yale Univ. Press, 1940. (3) Bellin, M. I., and Libina, B. I,, Khim. M NO. 7,3-9. (4) Dietz, T. J., J . Franklin Inst., 236,451(6) Dietz, T. J., and Kishbaugh, T. V., Ibid., (6) Faucher, G . H., Oliphant, S. C.+and Bo ENO.CEXOM., ANAL.ED., 1 (7) Hausen, H.,2. Ver. deut, Ing (8)Hosking, J. S.,J. Council Sci. Ind. Research. 17,234 (1944 (9) Howe, A. F.,U. S. Patent 2,37Q,353(Feb. 27, 1946). (10) Moore, D.H., Rm.Sci. Instrumenla, 14,296-7 (1943). ( 1 1 ) Mulliken, R. S., J . Am. Chem. SOC,,44,1033-8 (1922). (12) Neuman, J. J., U, 9.Patent 2,341,230(Teb. 8.1944). (13) Oliphant, 5. C., Houssiere, Am. Inst. Mining Met. Engrs., Tech. Pub. 1539 (1942). (14) Riegel, E.K.,“Chemical Machinery”, New York, Reinhold Pub. Gorp., 1944. (16) Sanchez,F.G., Proc. 14th Assoc. Sugar Tech. Cuba, 19 (16) Sanchez, F.G., U. S. Patent 2,335,794(Nov. 30, 194 “Atomia Energy for Military Purposes”, Prince(17) Smyth, H. D,, ton, N. J., Princeton Univ. Press, 1945. U. 8.Patent 2,368,876 (Feb. (18) Terrados, E.P. Y,, (19) Terrill, H. M.,J. Franklin I @ . , 237,73-6 (1944) (20) Veksler, G.M., K h h . Mmhinostroenie, 1940,No. -8. (21) Vilter, E.F., U. S. Patent 2,360,465,(Oct, 17, 1944). (22) Wilsmann, W., Chem. A p p . , 29, 186-90, (23) YarmolinskiX, M. B., Sa&har,1940, No.

and separation of come use of liquid propane and employed. Other substitution of liquiddistillation, evaooratiofi to dryness, and extraction of the dry residue, in the production of vitamins (21) and similar pharmaceuticals. I n this field salting-out procedures have been superseded by more efficient and eeonomical extraction methods, Solvent extraction is also an important operation as an alternate to adsmptbn in the concentration and purification of penicillin (83s). cids and glyceride esters solvent action is a p treatment of oil

ean with furfural and other solvents ging) and unsaturated (drying) commercial operation. to the treatment of other ch ~LIcottonseed. linseed. and vegetable oils (2.3.13.18. ,,. coconut, and to the separation ‘and refining of free fatty acids derived from them. Several commercial processes making use of some type of extraction with organic solvents, such as methanol, pane, furfural, and acetone, have also been developed. o separate high-molecular-weight fatty acids such &s ic. Little information, is available, but the work ffixson (13) and Hirson and Bockelmann (19) gives insight in some cases into their technical bases. Smith and Funk (39) showed that liquid-liquid solvent extraction with mixed organic solvents, such aa various glycol mixtures with methanol, can be commercially feasible for recovering aromatic components from hydrocarbon mixtures such as gasoline, although the recovery and purity of the aromatic extract is limited. Solvent extraction has been applied to the recovery of itaconic acid from fermentation liquors with organia solvents r

SOLVENT EXTRACTION CONTINUED FROM PAQS 27

Re

-

modified Reynolds number density nscosity Subscripts = core and wall fluids, c and w, respectively =

p ~.r=

In theory these should bo general correlations, The results of a number of pilot plant studies of specific extraction processes have been recently reported (19,84, 36, 89) involving the refining of crude long-chain fatty acids by liquid propane extraction, extraction of aromatics from gaaolines with mixed solvents composed of methanol and ethylene glycol ; January, 1946

I

INDUSTRIAL A N D ENQINEERING CHEMISTRY

37