W. C. BAUMAN
January, 1946 I N D U S T R I A L A N D E N G I N E E R I N G C H E M I S T R Y 47
S
IN C E the pioneering w ork of Adams and Holmes (2) on the synthesis of ion exchange resins, rapid progress has been made in the improvement of these resinous exchangers and their applica
tion to a variety of chemical processes (1 ). Syn
thesis of resins showing marked ion exchange prop
erties from a wide variety of raw materials has clarified the nature o f the exchange reaction and has made it possible to define the essential require
ments for a good exchange material: (a) a highly cross-linked structure to ensure insolubility; (6) a multiplicity of ion active groups— e.g., S 03H, COOII, OH groups for cationic exchange, and amine groups for anionic exchange. The breadth of application of the ion exchange process will de
pend to a large extent upon the degree o f success in increasing the density of these ion active groups with maintenance of sufficient cross linking for stability and insolubility.
The phenol-formaldehyde condensation produces an excellent cross-linked structure for ion exchange materials. Ion active groups, either cationic or anionic, may be introduced during the condensation reaction or, in some cases, by aftertreatment of a
standard phenol-formaldehyde resin (10). In the cation ex
change field, resins may be produced of all' three basic types—
sulfonic acid, carboxylic acid, and polyhydric phenol (7). The latter two types may find some specialty applications such as for buffer filters in the neutral pH range or in chemical processing where their high selectivity for hydrogen, calcium, and mag
nesium ions m ay be used to advantage. In the field of water softening, dealkalizing, and demineralizing, resins of the sulfonic acid type are preferable because of their high acidity and ready reversibility in the sodium-calcium exchange. These sulfonic acid resins are o f two types: nuclear sulfonic acid in which the SOsH groups are directly connected to the benzene ring, and m-sulfonic acid in which the active group is — CH2S 03H.
An improved ion exchange resin of the nuclear sulfonic acid type (Dowex-30) is the subject o f this paper. I t is a condensation product o f o- and p-phenolsulfonic acid with formaldehyde, as described by Wassenegger and Jaeger (10). Its composition is
O il
^ / V C H ,
-Flow in G a llo n s /S q . Ft./M in.
i I i I i i i
F i g u r e 1 . Pressure D rop in Dowriflow Operation in a 5 2 -In c li B ed o f N uclear S u lfo n ic A cid R esin
For m ost exchange applications, such as water softening and demineralization, only the sulfonic acid grouping is active.
However, the wet phenate form has shown promise in the removal of weak acids, such as II2S, C 0 2, and substituted phenols, from gases and oils. The weak acid is converted to the water-solublo sodium salt, which remains in the wet granules.
The nuclear sulfonic acid groups are very stable to hydrolysis at elevated temperatures in the presence o f either acid or alkali.
The resin has shown no color throw, no loss in capacity, and no breakdown of the particles when operated on either the' acid or
-X — X— X— X —
SO jH
with some of the S 0 3II groups replaced by a — CH 2— cross link.
P H Y S I C A L A N D C H E M IC A L P R O P E R T I E S
The resin is produced in the form o f hard black granules of irregular shape. A typical mesh analysis, as shipped in the wet sodium form, is: 69.8% on 20 mesh, 14.2% on 30 mesh, 8 .1 % on 40 mesh, 5 .0 % on 50 mesh, 2 .9 % through 50 mesh.
In its wet form it weighs 50 pounds per cubic foot and con
tains approximately 5 5 % moisture. The density of the wetted granules is 1.30 and o f the dried granules, 1.67. An operating bed of the resin contains about 3 8 % free space. The material is hard and shows little packing on downflow operation. The pressure drop on downflow operation o f a 52-inch bed at various flow rates is given in Figure 1. These results were obtained with a bed classified by thorough backwashing. The expansion o f a bed of this resin on backwashing increases with an increasing backwash rate in the following manner: 8.6% at 3 gallons per square foot per minute, 17.2% at 4 gallons, 31.0% at 5 gallons.
The resin contains phenolic groups and nuclear sulfonic acid groups, both of which possess ion exchange activity. However, the phenolic groups become active only at pH values above 9.5, the conversion to the phenate salt being very slow below pH 11.5.
U U Z N N flU li
F igure 2 . pH T itra tio n Curves D esignation*
X — X— X— X — X— X— K resin --- — --- A resin --- C resin
— --- --- -- R resin
--- --- D ecation ized greensand
* 5 gram s o f resin ex cep t fo r greensand.
A ctiv e G rou p
—SO,H
— CHiSOtH
— COOH OH (20 g.>
48 I N D U S T R I A L A N D E N G I N E E R I N G C H E M I S T R Y V ol. 38, No. 1
T itra te d C a p a c ity » 3 2 , 0 0 0 G ra in s C a C O j/
Cu. Ft.
C C. 2 N- NoOH
sodium cycle below 9.5 pH at temperatures up to 212° F. Some softening of the particles and loss in capacity are obtained in sodium cycle operation at combined high pH and high tempera
tures. However, one unit has operated in the hydrogen cycle on water of 10.0-10.5 pH at 150-180° F. for over 100 cycles without any breakdown or loss in capacity.
The resin, because of its highly cross-linked structure, is in
soluble in all organic or inorganic liquids and is infusible. De
composition of the dry acid form starts at about 200° C. As with other organic exchange agents, the resin is decomposed by strong oxidizing agents, such as chlorine, bromine, HCrCL, etc.
Weaker oxidizing agents, such as oxygen, sulfuric acid, Fe'1"4" 1', show no detrimental effect. Even nitric acid is without appreci
able effect at room temperature and below 10% concentration.
IO N E X C H A N G E P R O P E R T IE S
Resins of the nuclear sulfonic type show the highest acidity of all ion exchange resins tested. Single-pass contacting of a 6- foot bed with strong sodium chloride solution has produced as high as 10% hydrochloric acid solutions. This property ensures
Greens;
0 0.5
L b s . H C L P e r K ilo g r a m C a C O j R e m o v e d Figure 5. O perating C apacity o f Io n E xchange
M a te ria l in Acid Cycle w ith IIC l Regeneration
O 0.5 1.0
L b s . NaC L p e r K ilo g r a m C a C 03 Rem oved
;ure 4. O perating C apacity o f Io n Exch an ge M a terials in S o d iu m Cycle
January, 1946 I N D U S T R I A L A N D E N G I N E E R I N G C H E M I S T R Y 49
L b s . H2SO4 P er K ilo g r a m C a C 03 R e m o ve d
Figure 6. O peratin g C apacity o f Ion E xchan ge ¡M aterials in Acid Cycle w ith IIíSO« R egeneration
Botli sodium chloride and liydroehloric acid regeneration of this nuclear sulfonic acid resin give an operating capacity closely approaching the total available capacity at high regenerant dosages. This feature appears to be common for most exchange materials o f tho resinous gel type (8), and indicates a high rate of diffusion of ions through tho gel structure. The exchange speed in the N a-H exchange in 0.01 N chloride solution is shown in Figure 7 for a batch system with rapid agitation as determined by conductivity. The equilibrium state is attained in about 2 minutes. The exchange rate is much lower for the aluminum silicates, either natural or synthetic. Natural greensand has a cation content equivalent to about 100,000 grains o f calcium carbonate per cubic foot o f which only about 3 % is available for exchange. Synthetic aluminum silicates have been produced of lower density and higher water content than greensand, so that about 4 0 % of their total capacity o f 25,000 grains of calcium carbonate per cubic foot is available in the softening cycle (4).
Predominance o f a surface reaction with greensand led early in
vestigators to treat the ion exchange reaction as one of adsorp
tion (9). W ork with the newer exchange resins has indicated clearly that the exchange is a true chemical exchange o f ions be
tween liquid and solid phases and is controlled by mass action
laws (8). %
The failure o f sulfuric acid to give a regenerant efficiency and operating capacity comparable to sodium chloride or hydrochloric acid is not yet fully understood. Concentrations of sulfuric acid higher than 2% give pronounced precipitation of calcium sulfate.
This limitation on concentration may cause lower regenerant efficiency because of the increased selectivity o f the resin for calcium and magnesium ions at low acid concentrations. It is also possible that the acidity of the second hydrogen in sulfuric acid is too low to regenerate effectively the strongly acid nuclear sulfonic acid groups in this resin.
L IF E T E S T S
A number of factors may contribute to the deterioration o f an ion exchange material during service. High temperature, oxidizing agents, high acidity, alkalinity, or brine concentration may destroy either the primary resin structure or the ion active groups. Rapid changes in salt concentration, excessive bed pack
ing, or attrition may split the particles into fines. Water tur
bidity or supersaturation m ay lead to a fouling of the resin particles. Bacterial growths m ay feed on the resin. Resistance to many of these deteriorating factors m ay be determined by laboratory tests; others can only be checked by field tests.
Two accelerated laboratory tests have been run on this im
proved nuclear sulfonic acid resin to determine its stability. In one o f them a 6-inch depth o f resin was operated in a 6-minute water-softening cyclo for 8500 cycles. In the second test in
dividual particles o f resin were carried through a 2-minute water- softening cycle for 40,000 cycles. In both tests saturated sodium chloride was used as a regenerant in order to give the resin parti
cles the maximum osmotic force shock— i.e., a rapid swelling and shrinking o f tho particles caused by osmotic force changes with changing salt concentration. Experience indicates that this swelling and shrinking is the usual cause o f particle splitting during operation of ion exchange resins. After each of the tests the nuclear sulfonic acid resin had shown no physical degradation or loss in exchange capacity.
Figure 7. Speed o f E xchange in I he Conversion o f N a Form to II Form a t 0.01 /V Chloride
C on ce n tration
This resin has been installed in a boiler feed water purification plant for three years. In this application it softens a lime- treated river water at 8 -9 pH and maximum temperature of 100° F. from 60-70 p.p.m. calcium carbonate hardness to less than 1 p.p.m. The resin has shown no volume loss, no increase in fines, and no loss o f total capacity during this period. After 2.5 years of operation the operating capacity had dropped 15-20%
because o f fouling o f the particles with ferric hydroxide and calcium carbonate. Acid treatment has been used to clean the resin and has restored the operating capacity to its initial value.
Tests also indicate that this resin could have been cleaned effectively by a backwash at 6 -8 gallons per square foot per minute, but insufficient head room was available in the com mercial units. The photograph on page 46 shows this installation.
A second installation has operated in the acid cycle for water demineralization for three years. This demineralization unit reduces the salt content of Midland city water from 200 to less than 2 p.p.m. Hydrochloric acid is used as regenerant. Full operating capacity has been maintained throughout this period.
An analysis of a resin sample taken after 20 months of service showed identical particle size and capacity as the original resin.
A third installation was made for experimental purposes in the softening cycle; it handled a surface water containing up to 100
50 I N D U S T R I A L A N D E N G I N E E R I N G C H E M I S T R Y V ol. 38, No. 1
p.p.m. turbidity at 140-150° F. Saturated sodium chloride was used as regenerant, also at about 150° F. Although the unit gradually lost operating capacity to the extent of about 2 5 % in 15 months because o f resin fouling, no degradation of the resin occurred. Treatment with 15% hydrochloric acid to remove the accumulation o f ferric hydroxide and complex silicates restored the original capacity.
An unexpected result in all o f these units has been a marked reduction in turbidity of the water being softened or demineral
ized. Water containing 5-10 p.p.m. turbidity which has passed through sand filters of 40-80 mesh will be reduced to 1-2 p.p.m.
turbidity in passing through a resin o f 12-30 mesh. Surface water o f 50-100 p.p.m. turbidity m ay be reduced to 15-20 p.p.m . The resin appears to coagulate the colloidal particles, possibly be
cause o f its high concentration of ionic charges.
This clarification is advantageous for both boiler feed water and for municipal supplies.
L IT E R A T U R E C IT E D
(1) Adams, B. A. [to Ocean Salts (Products) Ltd.], Brit. Patents 536,266 and 541,450 (1941); Johnson, H. (to Norsk Hydro
Electric), Norwegian Patent 59,035 (193S); Findlay, D. M.
(to U.S. Rubber Co.), U. S. Patent 2,261,021 (1941).
(2) Adams, B. A., and Holmes, E. L., / . Soc. Chem. Ind., 54, 1-6T (1935); Brit. Patents 450.30S-9 (1936) and 474,361 (1937);
French Patents 796,796-7 (1936); U. S. Patents 2,104,501 (1938), 2,151,8S3 (1939), and 2,191,853 (1940).
(3) Akeroyd, E. I„ and Broughton, G., J. Phys. Chem., 42, 343
(7) Griessbach, R., “ Ueber die Herstallung und Anwendung neuerer Austausch adsorbienten, inbesonders auf Harzbasis” , Verlag Chemie, Berlin, 1939; abstd. in Angew. Chem., 52, 215-19
havior o f massecuites should develop from a kinetic inter
pretation of the crystallization of sucrose from pure solutions and from synthetic sirups. V elocity data and mechanism for pure sucrose solutions have already been discussed in previous papers o f this series (24, 25) as well as the effects o f com m on electrolytes.
The present paper supplements the list of salts already considered {25) and presents additional data on the effects o f various sub
stances which are likely to be found in cane and beet juices. In most cases the concentrations o f impurities considered are greater than those usually found in ordinary final molasses.
A static technique was used for most of the results presented.
The procedure consists o f seeding the sirup, which has previously been adjusted to the proper conditions, and following the rate of adj ustment refractometrically. It has already been demonstrated {24 , 25) that this method gives essentially the same values for gives velocities lower than methods in which stirring is involved.
This same difference is suggested by the observation that devia
tions from monomolecularity with unstirred solutions are more pronounced as the purity decreases. These variations, however, are all beyond the three-quarter life period of reaction and are therefore of little significance in the evaluation of rate constants;
they m ay be important in ascertaining the mechanism of crys
tallization from impure solutions.
It is well recognized that the pH of a sirup has a tremendous ef
fect on the crystallization velocity o f sucrose {5 ,1 0 ,1 6 ). An acid condition causes a diminished rate to a variable extent on account of inversion. From pH 6 -8 the rates are reproducibly constant;
above 8 the values are again decreased variably. A develop
ment of color is associated with this change. A pH of 7-8 was therefore adopted as standard in these investigations, and was realized by the addition of acid or base w’hen necessary. This was preferred to the use of buffered solutions in order to avoid salt effects.
V E L O C IT Y O F C R Y S T A L L I Z A T I O N
The general behavior o f synthetic sirups is demonstrated with the results for raffinose shown in Figure 1. The concave curva
ture is significantly maintained on a semilog plot, as discussed later in connection with Figure 5. The same type o f result w'as obtained w'hen invert sugar and betaine were the impurities;
in no case, at 30° C. or higher, w'as an increase in velocity observed at small concentrations o f nonelectrolytes (7, 12). How'ever, at 16° C. and an initial sucrose concentration equivalent to sucrose/
(sucrose + water) = 0.685, relative velocities o f 1.30 in thé pres
ence o f 0.4 gram of raffinose per 100 grams of water, 1.21 in the presence of 1.0 grams betaine per 100 grams water, and 1.11 with 1.0 grams invert per 100 grams water, were obtained. A t higher concentrations of impurities the relative velocities of crystalliza
tion were diminished and at higher temperatures these maxima also disappeared. This general behavior is exactly what is