1056
INDUSTRIAL AND ENGINEERING CHEMISTRY
(21) Jacobs, C . J., and Parks, G. S., Ibid., 56, 1513 (1934). (22) Jessup, R. S., J . Besearch N a t l . Bur. Standards, 20, 589 (1938). (23) Johnson, W.H., Prosen, E. J., and Rossini, F. D., Ibid., 35, 141 (1945). (24) Kemp, J. D., and Egan, C . J., J . Am. Chem. SOC.,60 1521 (1938). (25) Kennedy, W. J., Shomate, C . I%., and Parks, G. S., I b i d . , 1507 (1938). (26) Kistiakowsky, G. B., Ruhoff, J. R., Smith, H. A , , and Vaughn, W.E., Ibid., 57, 876 (1935). (27) Ibid., 58, 137 (1936). (28) Ibid., 58, 146 (1936). (29) Linnett, J. V., and Avery, W.H., J . Chem. Phys., 6 , 686 (1938). (30) Messerly, G. H., and Kennedy, R.PI.,J . Am. Chem. Soc., 62, 2988 (1940). (31) Moore, G. E., Renquist, M . L., and Parks, G. S.,Ibid., 62, 1505 (1940). (32) Parks, G. S.,Chem. Revs., 27, 75 (1940). (33) Parks, G. S.,and Huffman, H. M., J . Am. Chem. Soc., 52, 4381 (1930). (34) Parks, G. S.,Todd, 5. S.,and Shomate, C. H., Ibid., 58, 2505 (1936). 135) Pitzer. K. 5..I b i d . . 62. 1224 11940). (36) Pitser: K. S.,J. Chem. P h y s . : 8 , 711 (1940). (37) Pitaer, K. S.,and Gwinn, \Y.D., Ihid., 10, 428 (1942). J . Am. C h e m . Soc., 63, 2419 (38) Pitzer, K. S.,and Scott, D. K,, (1941). (39) Ibid., 65, 803 (1943)a
Vol. 41, No. 5
(40) Prosen, E. J., Gilmont, R., and Rossini, F. D., J . Research Natl. Bur. Standaids, 34, 65 (1945). (41) Prosen, E. J., Jessup, R. S., and Rossini, I?. D., I b i d . , 33, 447 (1944). (42) Prosen, E. J., and Rossini, F. D., I h i d . , 34, 163,263 (1945). (43) Rossini, F. D., Ibid.,22, 407 (1939). (44) Iiuehrwein, K. A . , and Huffman, H. M., J . Am. C h e m . SOC.,65, 1620 (1943). (45) Schumann, 8. C., Aston, J. G., and Sagenkahn, M.L., I b i d . , 64, 1039 (1942). (46) Scott, R. B., Ferguson, TV. J., and Brickaedde, F. G . , J . nesearch Natl. Bur. Standards, 33, 1 (1944). (47) Scott, R. B., Meyers, C. H., Rands, R. lo.,Brickwcdde, F. C., and Bekkedahl, X., I b i d . , 35, 39 (1945). (45) Sherman, J., and Ewell, R.B., J . P h y s . Chem., 46, 641 (1942). (49) Souders. M., Jr., Matthew, C . S., and Hurd, C . O . , IXD. ENG. CHnXf., 41, 1037 (1949). (50) Stitt, F., J.Chem. Phys., 8, 56 (1940). (51) Todd, S.S.,and Parks, G. S., J . Am. Chem. Soc.. 58, 134 (1936). (52) Voge, H. H., and May, N. C . , l b i d . , 68, 550 (1946). (53) Vvederskii, A. A., Vinnikova, S.G., Zharkova, V. R.,and Funduiler, B. M., J . Gen. Chetn. (U.S.S.R.),3, 718 (1933). (54) Willingham, C. B., Taylor, W. J., Pignocco, J. M., and Rossini, F. D., J . Research S a t l . Bur. Standards, 35, 219 (1945). (55) Yost, D. &I.,Oshorne, D. W.,and Garner, C . S., J . Am,. Chwa. SOC.,63, 3493 (1941). RECEIVED Deoomber 29, 1947.
creen Packing for Fractionatin Columns H. J. JQHN AND
c. E. REHBERG
Eastern Regional Research Laboratory, Philadelphia 18, P a .
An efficient, inexpensive wire-screen packing is described
CONSTRUCTION O F THE COLUivIN
which is easy to fabricate. It requires no welding, grinding, precise fitting, or preassembly of parts, and ordinary glass or metal tubing is sufficiently uniform in bore for use with the packing. Fractionating efficiency, liquid holdup, and pressure drop are given for three columns 0.625,1, and 2 inches in inside diameter. These operating characteristics compare favorably with those of other types of screen-packed columns, which require much more skill, precision, and expense in fabrication.
Heavy-walled Pyrex tubing or Pyrex pipe was used in most cases because of the internal pressure on the tube produced by the stress in the screen elements and retaining rings. The wire screen was stainless steel, Phosphor bronze, brass, or copper, the last two being less satisfactory because of their poor elasticity and tendency to relax under mild heat. For maximum efficiency the screen elements should maintain close contact with the column wall by virtue of firm and unrelaxing pressure owing to the elasticity of the metal. No study of mesh size as related to
r
CENT years, several efficient laboratory fractionating columns having wire-screen packing have been described (1-4). Although these columns have high efficiency (as indicated by low height equivalent t o a theoretical plate), their widespread use in routine work is severely limited by their high cost, Tvhich is due to the skill and precision required in their construction. This paper describes a form of wire-screen packing which is inexpensive to fabricate and yet is highly efficient in use. The individual screen elements are not welded or otherwise united; this makes possible the use of metal screen which would be difficult or impossible to spot weld. Furthermore, because the screen elements are individually packed into the column without any preassembly, and as placed in the column, they are under stress, they expand to fit the tube snugly despite variations in tube diameter and deviations 'rom a true circular cross section. This prevents leakage of liquid past the elements and permits the use of ordinary tube or pipe instead of precision-bore tube.
SPLIT R I N G
(2
Figure 1. Component Parts and Assembled Section of Fractionating Column
May 1949
INDUSTRIAL AND ENGINEERING CHEMISTRY
1057
TABLE I. SPECIFICATIONS OF COLUMNS Column No. Type tube
2
1 Pyrex combustion
Inside diameter, mm. 17 Packed length, 12 inches 76 Number of screen disks Diameter of disks 19 mm. Metal used Phosphor bronze Mesh of screen 60 X BO Diameter wire, 0.0075 inahes Size stock for rings, l / l 8 X '/ai inches
3
Pyrex pipe
Pyrex medium wall
25.4 9.5
50.8
11
48
22 57.2
28.6
Stainless steel 80 x 80 0.0055
1/18
X
Phosphor bronze 60 X 60 0.0075
l/sa
I/(
x
1/16
pot had a side inlet with a 10/30 ground-glass joint through which a tube extended This inlet tube was used to withdraw samples for analysis and also for connection t o a manometer for measurement of pressure drop through the column. The still head used was sometimes of the Rossini type 6) and sometimes of a simpler total condensation variable take-o type having a holdup of 0.2 t o 0.5 ml. All connections between pot, column still head, and thermometer were standard taper ground-glass joints.
I
DETERMINATION O F OPERATING CHARACTERISTICS
Figure 2.
Performance of Column 1
efficiency was made, but this is thought to be relatively unimport a n t as long as the mesh is fine enough for the liquid t o fill the interstices, thus forming a continuous film over the surface of the screen. The screen used in the present work was 60 by 60 or 80 by 80 mesh. Figure 1 shows the component parts and a n assembled section of packing. Circular screen disks are c u t out with a punch press. Then a ortion of a circle is c u t from the edge of each disk to a depth o f a b o u t one third the diameter of the disk, This portion which was cut out serves a s the vapor passage in the assembled packing and may be varied in size. Increasing the radius, or depth, of the cut inoreases the size of the vapor passage, thus increasing the capacity of the column. At the same time, however, it reduces the area of screen surfMe and hence slightly reduces the efficiency of the packing. The diameter of the screen disks should be 5 t o 10% greater than the diameter of the tube in which they are assembled. Then, when the disk is forced into the tube, i t will be bowed and will remain under stress, because of the elasticity of the metal. T F s stress serves the double purpose of holding the disk snugly in &lace and making it shape itself closely t o the wall of the tube. hus a tight fit is obtained despite variations in the diameter of the tube or departures from a true circular cross section in the tube. This close contact between the edges of the disks and the wall of the tube is extremely im ortant; otherwise liquid reflux would leak through and channel t o w n the side of the tube. The assembled unit, or cell, shown in Figure 1 consists of two disks bowed jn opposite directions (with their concave sides together) and having their vapor openings on opposite sides of the column. I n order t o be doubly sure that a tight seal is obtained and t h a t the cells are held firmly in place, split rings similar t o piston rings are placed above and below each cell. These rings are made of stainless steel, Phosphor bronze, or other elastic material which retains its elasticity at the temperatures which will be encountered in the use of the column. The assembled packing consists of a series of such cells placed so t h a t each touches those above and below, thus providing a continuous screen surface over which the reflux flows and a zig-zag open pathway up through which the vapor passes.
Of the numerou: eo umns constructed and evaluated in one or more respects, three were chosen for more complete testing because their size and design best fitted them for general use in the usual types of laboratory fractionating problems. Table I shows the pertinent data on their construction and Figures 2, 3, and 4 show their operating characteristics. The test mixture used was n-heptane and methylcyclohexane, and analyses were made by refractive index, the data quoted by Ward (6)being used. Efficiency tests were all made under total reflux. Pressure drop through the co!umns was measured by a U-tube manometer con-
16
* - THEORETlCAL
1
0-
I
PLATES PRESSURE DROP
I/ I
2
E 6 a i 05
b 0
P 3'
ASSEMBLY OF STILLS
For determination of operating characteristics, the columns were enclosed in double glass jackets, the inner jacket having a single winding of Nichrome ribbon. Thermometers were placed near the top and the bottom of the columns inside the inner jacket. Heat input, controlled by a Variac, was adjusted so that the thermometers near the top of the jacket read the same as the one in the vapor at the still head. The still pot was a round-bottomed flask heated with a Glas-Col heater controlled by a Variac. The
4
500
Figure 3.
1000 I500 REFLUX RATE.ML PER HR.
2000
Performance of Column 2
0
1QS8
I N D U S T R I A L A N D E N G I N E E R I N G CHEMISTRY
Vol. 41, No. 5
of the difficulty of maintaining steady conditions for a period sufficient t u give reliable and reproducible results. The maximum rate of boilup that could be obtained from a 5-liter flask heated with a Glas-Col heater was 7500 1111 per hour. This was not, sufficient t o flood Column 3 (Table I and Figure 4 ) ; hence the characteristics of this column are shown in Figure 4 only up to this boilup rate and not t o its flood point,. The behavior of Column 3 a t the boilup rate of 7500 nil. per hour indicated that its maximum capacity was a t least 8 or 9 liters per hour.
Any attempt to extrapolate the operating characteristics as shown in Figures 2 to 4 to colunins of other diameters is complicated by several factors: The number of disks per foot of column was not constant; the percentage of the area of the disks cut out t o form the vapor path was not exactly const,ant; and the ratio of the diameter of t,he disks t o t h a t of the tubing \?-as not coilstarit in the three columns described. However, a few qualitative statements about the effect of increasing the diameter on operating characteristics seem justified. The number of theoretical plates per foot is markedly reduced a t larger diameters, bot,h because of fewer disks per foot being used and because of increasing difficulty of maintaining close contact between disks and tubing. Pressure drop per foot of column decreases, primarily because of the presence of fewer disks per foot. The liquid holdup increases somewhat. The greater area of wet surfaces accounts for this, but the effect is diniinished bj- lhe presence of fewer disks per foot.
L a
w
W v)
4a -I
2 W B:
8 I-
Like other types of screen-packed columns, it is essential that the packing bc thoroughly and uniformly wet before beginning a distillation or test run. This is accomplished easily by a short period of rapid boilup under total reflux. T h e n wet, a uniform continuous film of liquid covers all screen surfaces. This film remains intact so long as reflux is continucd, no mat,tcr how low the boilup rate. Maximum plate efficiency was obtained a t very low boilup rat'es (barcly cnough to keep the pa,cltingwet,). LITERATURE CITED
netted to the still pot and open to the atmosphere on t,he other arm. Holdup was measured by the niethod recommended by Ward; heptane and biphenyl ether were used and analyses were made by refmct.ive index. The rate of boilup and t,he capacity of the columns were determined by operating with total take-off for a brief measured interval (0.5 or 1.0 minute) and measuring the volume of distillate. Thus, the rates were measured a t the still head and not at, the pot, though there should be little difference in rates a t the two points. S o tests were run beyond t,he point of incipient flooding because
(1) Bragg, L. B., I K D . E X G . CHEM., ~ " I L . ED.,11, 283 (1939). (2) Grisn-old, John, Morris, J. K., and Van Berg, C. F.,IXD.ENQ. CHEY.,36, 1119 (1944). (3) Lecky, H. S., and Ewcll, R. H., ISD.ENC.CHIIM.,AXAI..ED.,12, 644 (1940). (4) Stedman, D. F., X a t l . Petroleiim S e w s . 29, 11-125-6. 11-128 (Aug. 25, 1937). ( 5 ) Ward, C. C., U . S.Bur. Mines, Tech. Paper, 600 (1939). ( 6 ) Willingham, C . B.. and Rossini. E. D., J . Reseaich .Vall. Bur. Szandards, 37, 15 (1916). R E C E I V E .4pril D 3, 1948. The nienrion of comiiierciril products does not iiiiplp t h a t they are endorsed 01' recommended by the Dcpartment of Agriculture over others of a similar nature not mentioned.
L. L. AICKIXNEY AND W. F. SOLLARS' .Vorthern Regional Research Laboratory, Peoria, I l l .
I
N COKXECTIOX with studies of the nonenzymatic browning
on soybean protein, it was desired to inveatigate the extraction of protein from solvent-extracted soybean meal with sulfurous acid. Although i t was known that sulfurous acid could be used t o peptize the protein (6, Q), quantitative information had not been reported. Also, it had been claimed ( 6 ) t h a t soybean protein, extracted with sulfurous acid, exhibited a superior light color. Soybeans contain about 9% sugars ( l a ) . Only a small amount 1
Present address, Purdue University, Lafayette, Ind.
of reducing sugars are present (0.1 to 0.2%) ( 7 ) . It is possible, therefore, that part of the bromn color observed in soybean protein obtained from hexane-extracted meal Tvithout the use of sulfur dioxide is caused by the well-known "browning reaction." In order t o prevent the formation of this color, sulfur dioxide is used t o precipitate the protein from the meal extract ( I O , I S ) . Sodium sulfite is often used along with alkali in extracting the protein from the meal (6,8 ) . Another method involves removing sugars by leaching the solvent-extracted meal with aqueous sulfurous acid a t the isoelectric point of the protein (9, 13).
