1406
INDUSTRIAL AND ENGINEERING CHEMISTRY
(27) Maillard, L.C., Ann. ch(m., [9] 5, 258(1916). (28) Nelson, E. K., and Browne, C. A., J . Am. Chem. Soc., 51, 830 (1929). (29) Neuberg, C . , and Oertel, W., Biochem. Z., 55,496 (1913). (30) Neuberg, C., and Rewald, B., Ibid.,71, 143 (1915). Ber., 63,3068 (1930). (31) Neuberg, C., and Scheuer, M., (32) Neuberg, C., and Straws, E., Archiv. Biochem., 7, 226 (1945). (33) Pinkus, G., Ber., 31,31 (1898). (34) Sattler, L., in “Advances in Carbohydrate Chemistry,” Vol. 3, pp. 113-28, New York, Academic Press, 1948. (35) Sattler, L., and Zerban, F. W., IND. ENG.CHEZI.,37, 1133 (1945). (36) Sattler, L., and Zerban, F. W., Science, 108, 207 (1948). (37) Sattler, L., and Zerban, F. W., Sugar, 39, 28 (December 1944). (38) Ibid., 42, 26 (December 1947). (39) Sharefkin, D. M., private communication. (40) Sigurdsson, S.,Ph.D. dissertation, Munich (1944). (41) Singh, B., Dean, G. R., and Cantor, S.M., J . A m . Chem. SOC., 70, 517 (1948). (42) Speck, J, C., Jr., private communication.
Vol. 41, No. 7
(43) Thornton, B. J., and Speck, J. C., Jr. Presented before the Division of Sugar Chemistry and Technology at the 112th iMeeting of the AMERICAN CHEMICAL SOCIETY, New York, N. Y . , 1947. (44) Troje, E., Deut. Zuckerind., 60, 101 (1935). (45) Tsang,,,S.hl., Wood, E. H., and Johnson, J. R., “Organic Synthesls, Vol. 24, P. 75, New York, John Wiley & Sons,1944. (46) Weast, C . -4., and Mackinney, G., IXD. ENG.C H E M . 33, , 1408 (1941) (47) Winter, H., 2. V e r . Riibenzuckerind., 44, old series 1049 (1894). (48) Wolfrom, M. L., Schuetz, R. D., and Cavalieri, L. F., J. Am. Chem. Soc., 70, 514 (1948). (49) ZemplBn, G., Ber., 60, 1555 (1927). (50) Zerbrtn, F. Fv., and Patt,ier, L., IXD.ENG.C H E M . , ~1180 ~, (1942), and Sattler, L., Sugar, 42,44 (February 1947). (51) Zerban, F. W., (52) Zuman, P., L i s t y Cukrouar., 62, 81, 89, 97 (1946). I
RECEIVED July 8, 1948. Presented before the Division of Sugar Chemijtr> and Technology a t the 113th Jloeting of the AafERIoAN CHBMICAL SocImY, Chicago, 111.
egeneration of a Cation Exchange HUGO P. KORTSCHAK1,ROBERT F. GILL, J R . ~ ,AND JOHN H. PAYhTE1 Pacijic Chemical and Fertilizer C o m p a n y , Honolulu, T . If. A t moderate flow rates, upflow regeneration of Dowex 30 cation exchange resin is only slightly less efficient than downflow. Upflow regeneration has the advantage of permitting a high initial removal in the exhaust phase, without the use of the large excess of acid required to complete regcneration. Variation of the flow rate between 0.5 and 4 gallons per square foot per minute has little effect on downflow regeneration. With upflow regeneration, efficiency decreases when noticeable lifting of the bed occurs, between 1 and 2 gallons per square foot per minute. Variation of the concentration of sulfuric acid regenerant has little effect on total capacity for sodium chloride, but the more dilute arid results in better maximum removal of sodium in the subsequent exhaustion phase of the cycle. Sulfuric acid is as efficient as h j drochloric acid for regenerating a resin in the sodium form. A resin in the calcium form is apparently less efficiently regenerated by sulfuric acid; a major part of the reduction in capacity is due to subsequent exhaustion by the calcium sulfate which is precipitated during the regeneration and dissolved by the following rinse water.
N EVALUATING a cation exchange resin for a particular use, a thorough study of t,he regeneration requirements is of prime importance since the chief operating cost, is for rcgenerating chemicals. Thompson and Roberts (4)pointed out the necessity for data OIL amounts of regenerant, strength of regenerant, and flow rates in engineering considerations, and gave break-through capacity figures for jix resins. Other data on regeneration are scattered through the literature. Kuniii (3)gave an extensive bibliography of this literature. The study described here gives the regeneration requirements for the commercial cat,ion exchanger Uowex 30, a resin of the phenol-formaldehyde type containing nuclear sulfonic acid groups, describcd by Bauman ( 1 ) h-en- Dowex 30 was wetscreened to remove particles sn~allerthan 50 mesh. The resin was supported in a Pyrex tube of 35-mm. outside diameter, and Present address, Experiment Station of the Hawaiian Sugar Planters’ bssociation, Honolulu, T. H. 2 PxsonE address, Havaiian Pineapple C o m p n n s , Honoiiilu. T. H. 1
9
the regenerated, rinsed, backwashed, and drained volume was adjusted t o 300 cc. The resin was completely exhausted with sodium chloride solution (0.5 gram per 100 ml.), run downflow a t 25 ml. per minute. A 500-ml. rime a t 50 ml. per minute and a 1500-ml. backwash a t 250 ml. per minute foIIowed. All water used vias purified by ion exchange. The regeneration was carried out with sulfuric acid under variations which followY. A 500-ml. rinse conipleted the regeneration cycle. The regeneration effluent was collected a t 100-ml. intervals and titrated against 0.1 S sodium hydroxide solution. EFFECT OF AMOUNT OF REGENEIRAA-T
The effect of the amount of regenerant used ma.s determined in a eeries of runs using 0.17 N sulfuric acid a t a flow rate of 25 ml. per minute. Both d o w n f l o ~ and upflow trials iTere made. Table I gives data for six runs. iit this flow ratc, upflow regeneration is only slightly less efficient than downflow. (Efficiency is the percentage of total acid used which is absorbed by the resin.) Since t’hc ion exchange reaction is reversible, ResSa
+ ~ / z H ~ SeO ~ResH + i/nKa2SOa
with K for the reaction as written approximately 1, it is evident that high capacity can be att,ained only with a large excess of rcgenerant. When the resin is only partially regenerated, downflow regeneration leaves the lower portion of the resin in the sodium form. As t,he sodium chloride solution runs through the bed, the acid
TABLEI. EFFECTO F
h & I O U S T O F REGEXERAST O N CAP+4CITY A K D EFFICIENCY
I12SOa To!. of
0.1; .v HnSOi, Liters 0.5 1.0 2.0 3.0 4.0 6.0
Absorbed,
Equivalent Up Down 0,084 0.153
0.239
0.286 0.305
0.303
0.084 0.150 0,232 0,277 0.298 0.304
Effioiency of Regeneration,
% Down
UP
58 90 70 66 45 30
98 88
68
54 44 30
M a x . Na Removal,
% ~Down 44
62
84 94 96
99
CP 7. 8_
89 58 QY 98 99
I N D U S T R‘IA L A N D E N G I N E E R I N G C H E M I S T R Y
July 1949
produced by the reaction in the upper portion of the column tends to regenerate this lower part, with consequent apparent “leakage.” As a result, removal is poor during the first part of the run (Figure 1A). After upflow regeneration this does not occur (Figure I B ) , and removal of K a + from the solution is best a t the beginning of the run. (The slight rise shown may be ascribed t o dilution effects due t o the water in the bed at the beginning of the run.) For this reason, nearly complete removal of N a + from the solution can be obtained with a partially regenerated resin, whereas with downflow regeneration a more nearly complete, and therefore less efficient, regeneration is required. Table I shows that the maximum removal figure is a more sensitive indicator of complete downflow regeneration than is the determination of the total capacity. EFFECT O F FLOW RATE
Four liters of 0.17 N sulfuric acid were used for each regeneration in a series of four runs in which the flow rate was varied from 25 to 200 ml. per minute. Since the cross-sectional area of the resin column was approximately 0.01 square foot and the volume of resin nearly 0.01 cubic foot, flow rates may be converted approximately to gallons per square foot per minute or to gallons per cubic foot per minute by multiplying by 0.025. Table 11 gives the data. RATEO N CAPACITY AND EFFICIEXCY TABLE 11. EFFECTOF FLOW Flow Rate, Ml./Min. 25 50 100 200
H%SO4 Absorbed, Equivalent Down Up 0.305 0.298 0.311 0.299 0.283 0:307 0.286
Efflciency of Regeneration,
%
,
Down 45 45
..
45
Max. Na Removal, 70
Up 44 43 41 41
Down 96 95
Up
94
94
99 99 97
1407
The efficiency of removal, however, is noticeably affected by higher flow rates in both cases (Figure 2). With downflow regeneration, this is limited to poorer removal at the beginning of a run; with upflow regeneration, removal was poorer throughout the run, but not appreciably so until a flow rate of 100 ml. per minute was reached. This caused a noticeable lifting of the bed; 200 ml. per minute increased the bed volume by about 30%. EFFECT OF CONCENTRATION O F REGENERANT
Regeneration was studied with a volume of sulfuric acid which contained 0.68 equivalent a t a flow rate of 50 ml. per minute. The strength of the acjd varied from 0.17 to 1.71 normal. Table 111shows the effect of concentration. There appears to be little change in total capacity as the concentration of sulfuric acid increases. TABLE111. EFFECTOF CONCENTRATION OF REGENERAXT ON CAPACITY AND EFFICIENCY Normality of
Has01 0.17 0.53
1.71
HaSOa Absorbed, E q u i v a lUp ent Down 0.311 0.299 0.304 0.291 0.298 0.292
Efficiency of Regeneration, Down%
Up
45 44 44
43 42 43
Max. Na Removal, Down% 95 93
91
UP
99 98 96
The efficiency of sodium removal is slightly less a t higher concentration (Figure 3). This is especially evident after downflow regeneration, as would be expected since the regeneration is not quite complete. The more concentrated acid regenerates the upper portion of the resin column more completely and leaves less acid available for the regeneration of the lower part. REGENERATION IN SODIUM AND CALCIUM FORMS
These data show that increasing the flow rate has no appreciable effect on the capacity after downflow regeneration. The apparatus k e d did not permit the application of the higher pressures which would be needed to increase the flow rate above 200 ml. per minute. A t some point, of course, an effect would become apparent. Higher flow rates give a slightly lower capacity with upflow regeneration, the maximum difference found being about 5%.
T o determine the difference between sulfuric acid and hydrochloric acid for regenerating the resin in the sodium and calcium states, a separate resin sample was used. It was similar t o that previously described except that the volume, not so accurately measured, was somewhat greater. Regeneration was downflow throughout and was carried out by running the amount of solution containing one equivalent of acid at 25 ml. per minute, followed by 200 ml. of water. The salt solution, 0.50 N sodium chloride
INDUSTRIAL AND ENGINEERING CHEMISTRY
1408
A.
Figure 3.
Doivnflow r e g e n e r a t i o n
Upflow r e g e n e r a t i o n
Sodium Removed Following Regeneration with 0.68 Equivalent of Sulfuric Acid a t Concentrations of 0.17, 0.53, and 0.71 N and a Flow Rate of 50 311. per Minute
or calcium chloride, was then run a t 25 ml. per minute until 2000 ml. had been collccted, and this was titrated to give the capacity. The exhaustion phase was followed by a 500-ml. rinse and 1500ml. baclcwash, the latter at 250 nil. per minute. The results of runs where the regenerant concentration varied from 2.0 to 0.1 X follow in Table IV.
T.4BLE
B.
Vol. 41, No. 7
O F SULFURIC AKD HYDROCHLORIC ACID ON Isr. EFFECT REGENERATION I N S O D I U M AND CALCIUM FORMS
Capaaity of Resin, Equivalent With HC1 With HzSOa Kormality of NaCl CaCh NaCl CaCla Regenerating exhausted exhausted exhausted exhausted Acid 0.362 0.095 2 .O 0.362 0.321 0.356 0,157a 0.5 0.359 o2 s 0.339 0.130 0.1 0 347 0.164 a Successive determinations gave 0.147, 0.159, 0.157 (following run with 2 N sulfuric acid).
Differences in capacity are sniall for the various regenerations, when the resin is exhausted with sodium chloride. The results are strikingly different with calcium chloride. The effect of the acid concentration is large in the case of hydrochloric acid. The 2 N hydrochloric gives almost 90% of the capacity reached when the resin is exhausted with sodium chloride, whereas 0.1 N hydro-. chloric gives less than 50% of this capacity. This is in agreement with the derivation of Bauman and Eichhorn ( 2 ) based on the Donnan equilibrium, which predicts regeneration efficiencies independent of acid concentration when a resin is exhausted by a univalent cation, and efficiencies increasing with concentration when the cation is polyvalent. Regeneration of the resin in the calcium form with sulfuric acid introduces the added complication of calcium sulfate precipitation. This so clogged the bed that i t was necessary to admit, about 25 ml. of backwash water several times during the runs with 2 N and 0.5 N sulfuric acid to make it possible to maintairp
July 1949
INDUSTRIAL AND ENGINEERING CHEMISTRY
the flow rate. It is noteworthy that no visible precipitation occurred with 0.1 N sulfuric acid, and the capacity was nearly equal t o that attained after regeneration with hydrochloric acid of the same concentration. The first exhaustion with calcium chloride, which followed a 2 N sulfuric acid regeneration of resin in the sodium form, gave a capacity of 0.365; this confirmed the fact that the low capacities found are due t o incomplete regeneration and not to poor removal of calcium by the fully regenerated resin. It was assumed that the resin was more completely regenerated by sulfuric acid than the data indicate, but subsequently exhausted when the rinse water dissolved the precipitated calcium sulfate. A series of runs was made with 0.5 N sulfuric acid t o test this hypothesis. Regeneration was interrupted after 500
1409
ml. had been run, and the precipitated calcium sulfate washed out of the resin bed with 2000 ml. of water before regeneration was completed with the remaining 1500 ml. of acid. This raised the reproducible capacity to 0.203 equivalent, an indication that the assumption is correct. LITERATURE CITED
(1) Bauman, W. C., IND. ENG.CHEM.,38,46 (1946). (2) Bauman, W. C., and Eichhorn, J., J. Am. Chem. Soc., 69, 2830
(1947). (3) Kunin, Robert, IND.ENG.CHEM.,40,41 (1948). (4) Thompson, R. B., and Roberts, E. J., Chem. Eng. Progress, 43,97 (1947). RECEIVED April 27, 1948. This investigation was financed by the Hawaiian Sugar Planters' Association,
Maintenance of Heat Transfer bv Cooling Water Treatment J -
W. A. HALL AND ROY F. WESTON The Atlantic Refining Company, Philadelphia 1 , Pa.
'
A simple method is described by which exchanger fouling can be evaluated in terms of throughput capacity. Results of a study of plant equipment are presented which demonstrate %he applicability and usefulness of the method. The fouling encountered was caused by slime deposits from the cooling water. Bacteriological, chemical, physical, and fouling data are given for two systems, both with and without water treatment. Water treatment consisted of phenolic-type compounds in one case and chlorination in the other.
will reproduce on synthetic culture media. These numbers indicate that probably 25% or more of the dry weight of the slime was bacterial cells. The remainder of the slime would consist of absorbed and adsorbed minerals and debris typical of the bacterial growth and the water supply. Table I1 gives typical chemical analyses of slimes. An accumulation of about 0.68 gram of this dry slime per square foot of surface area caused approximately a 33% loss in heat transfer capacity for the design conditions of one condenser. The nature of the slimes indicated that chemical treatment of the cooling water could control exchanger fouling. Consequently,
D
URING the war, when maximum throughput rates were needed for hydrocarbon processing equipment which had
long on-stream time, difficulty was experienced with loss in heat transfer capacity due to the fouling of water-cooled heat exchangers. The main units involved were condensers in which light hydrocarbons were being condensed on the shell side and cooling water was passing through the tubes. Inspection of the units after only two months of operation showed slimy accumulainch thick, on the water side of the tubes. Some tions, 1/16 to tubes had become clogged with this slime. The procedure described for evaluating a change in heat transfer capacity due t o fouling of the transfer surface has been used to study the commercial units just mentioned; it employs data normally available, and does not require the elaborate instrumentation or study needed for determining the actual scale resistance t o heat transfer. Results correlated so well and the required calculations were so short that this procedure ha? been established as a routine weekly check on cooling tower treatment. Enough bacteriological, chemical, and physical data are given here to demonstrate the applicability of the evaluation procedure. A preliminary study of the slime deposits indicated that they were bacterial in origin. Capsulated slime-forming organisms similar t o those previously described (3-7) were isolated from the slime. Table I shows the bacterial content of typical slimes from the plant cooling systems, as estimated by cultural methods. The estimated bacterial contents of 1 to 5 billion organisms per gram of dry slime represent only those living organisms which
TABLE I. BACTERIOLOQICAL CHARACTERISTICS OB SLIME FROM CONDENSER TUBESOB UNTREATED SYSTEMS
Source of Slime 850 unit Refrigeration plant Crude laboratory
Date in 1944 2/29 5/20 5/27
Total Bacteria Plate Count per Gram of Dry Slime (Millions) Sabouraud's W a k e m p ' s ThioNutrient dextrose fungi glycplate agar5 aparb agare mediumd 1300 ... 2930 426 40 5130 119 O.'i3 45
Solid culture medium of neutral reaction conducive t o the growth of a wide variety of heterotrophic bacteria. b Solid culture medium (PH about 5.5) conducive to the growth of yeast and molds. Solid culture medium (PH about 4.0) inhibitory t o most bacteria and conducive to mold growth. d Liquid medium for anaerobic and facultative anaerobic bacteria.
TABLE11. CHEMICAL CHARACTERISTICS OF SLIMEFROM CONDENSER TUBESOF UNTREATED SYSTEMS Compositiona, 9% 7 -
Source of Slime 851 unit 850 unit Refrigeration plant Crudelahoratory 3 farm units Biological filter a b
No. Samples Avgd. 3 1 1 1 16
l b
Loss
on ignition 31.8 34.8 32.3 33.8 35.7 16.1
Oven-dried sample. Sulfur-oxidizing biological filter.
Insol. in HCI 30.8 25.6 36.3 35.6 34.9 60.5
RzOs CaO MgO CuO 29.4 30.9 30.3 27.4 17.4 16.3
1 . 1 0 . 7 0.0 1.7 1.0 2.2 0 . 6 1.3 . , 0 . 6 1.6 0.9 0 . 9 7:1! 1.2 1.7 ...
.
Loss and undtd. 6 2 3.8 -0.8 1.0 -.6
4.2
