W
H E N Adams and Holmes (1) announced in 1936 th a t phenol-form aldehyde resins exhibited ion exchange properties, a new era of exchange adsorbents was opened.
These sy nthetic ion exchange resinous m aterials have been de
scribed by several authors (3, 4, 7-11) and m ay be grouped into two m ajor types. T he first, cation active, will exchange or adsorb substances which possess a positive ionic charge or which, owing to their electronic configuration, can form positive ions by the addition of a proton such as the su b stitu ted am m onium group.
T he second ty p e will adsorb or exchange acidic substances.
N eu tral m aterials adjusted to the zw itter ion condition by pH or other controls are n o t adsorbed or exchanged; substances of greater basicity or acidity will be affected proportionately over and above those of lesser reactivity. T he gel stru c tu re of the resins gives them m any of th e properties of a highly porous sur- face-active adsorbent. Resinous exchangers are insoluble in w ater an d all common solvents. T hey operate w ith a high specific reaction ra te which, however, varies w ith the subject ions. T hey exhibit different adsorptions for various ions. T his principle m ay be the basis for m any separations by the simple ad ju stm en t of contact tim es of solution w ith resin. T he resins are effective in relatively large particle size so th a t the flow of the solu
tion m ay be rapid and unobstructed by any packing of th e adsorb
ent. U nder these conditions commercial application is feasible.
Consideration of these properties of the resinous exchangers suggests th eir ad ap tatio n to th e T sw ett chrom atographic adsorp
tion technique which has found widespread application {12, 13, 14). D ifferential adsorptions have already been reported by M yers, Eastes, an d U rq u h a rt (11) who observed th a t sulfuric acid was bound in preference to hydrochloric, and copper was exchanged in preference to zinc. T h e chemical n atu re of these adsorbents offers advantages over such adsorbents as sugar, pre
cipitated lime, carbon, alum ina, fuller’s earth, etc.
A D SO R P T IO N O F T H IA M IN E
Two of the m ore fam iliar vitam ins, thiam ine and riboflavin, were selected for prelim inary experim ent; th e resin was Amber- lite IR-100. T hiam ine and riboflavin were determ ined q u a n tita tively by th e m ethods of Hennessy and Cerecedo (5)' and Conner and S trau b (2), using a K le tt fluorim eter (6). For comparison several parallel experim ents were ru n w ith typical nonresinous adsorbents. T he initial experim ents were perform ed on solu
tions prepared from crystalline thiam ine hydrochloride and riboflavin.
T h e adsorption of thiam ine by the sodium form of IR-100 and by other m aterials, including Am berlite IR-4, an acid-active res
inous exchanger, was investigated first (Table I). T he diam eters of th e modified T sw ett columns were approxim ately 1.5 cm.
T he synthetic zeolite and th e resinous cation exchanger are ex
cellent adsorbents of th e v ita m in ; alum ina is unsatisfactory. Ad
sorption by Am berlite IR -4 was n ot anticipated and, for prac
tical purposes, none was noted.
The principles of chromatographic adsorption and base exchange have been applied to the recovery of natural products through the use of synthetic ion exchange resins. In particular, the adsorption, separation, and concentration of thiamine by the acid-regenerated form of the resin have been studied. A n adsorption column tech
nique, utilizing resin with a particle size near 0 .4 0 mm. rather than the usual finely powdered material, completely removes thiamine, for which the resin has a large capacity, in the presence of riboflavin.
Since riboflavin is not adsorbed under the conditions in which thiamine is completely removed, the separation is satisfactory. The vitamin is eluted in excellent yields by passage of strong mineral acid through the column, although other reagents may be used to advantage in specific instances. These results are achieved with percolation rates equivalent to as much as 5 gallons per square foot per minute with beds no greater than 0.5 foot in depth. The practicability of the process has been demonstrated by the recovery of thiamine from rice bran extracts. The resin stability, particularly in strong acids, makes possible its application where such pH con
ditions are necessary to protect products from alkaline degradation.
T he recovery of the v itam in was considered next. A hot 25%
solution of sodium chloride gave a 76% recovery from th e zeolite using a volume of solution ten tim es the volume of th e zeolite, b u t th e recovery from IR -100-N a was only about 31% in a corresponding volume. T his indicated th a t th e vitam in is bound more tenaciously by tKe active groups of the resinous exchanger th a n by those of the zeolite. However, since the resinous m aterials m ay be regenerated w ith acid, th e possibility existed of elution w ith th e acid.
Prelim inary experim ents revealed th a t thiam ine is removed from solution ju st as satisfactorily by the acid as by the sodium form of the resin; however, the type of adsorption was such th a t good recovery is n ot possible w ith the conventional concentra
tions (2-7 %) of acid. Since these resinous m aterials are n o t acid sensitive, still greater concentrations were investigated.
So r p t i o n b v Am b e r l i t e IR-100-H. A 25-cm. column of Am berlite IR-100-H (average particle size 0.40 mm.) was pre
pared (bed volume 41.6 m l.), and a solution of 5 p.p.m . thiam ine hydrochloride was passed downflow a t a rate 28 ml. per m inute (equivalent to about 4 gallons per square foot per m inute). A to tal of 5 liters containing 25 mg. of thiam ine was passed through the bed. T here was no detectable q u an tity of vitam in in the effluent. After a brief rinse, a solution of 18% hydrochloric acid a t room tem perature was passed downflow a t a flow rate of 10 ml.
(equivalent to 1.5 gallon per square foot) per m inute u n til tw enty 50-ml. samples of effluent were collected. T he samples were analyzed for thiam ine. Figure 1 gives the recovery data.
There appears to be no tru ly chemical exchange, for on an equivalent basis a considerable excess of hydrochloric acid is necessary to elute the vitam in. Using this concentration of acid, 631
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 Vol. 37, No. 7
VOL. ACID PER VOL. IR-IOO-H
4.8 9.6 144 192 2 4 .0
Figure 1. Recovery of Thiamine from Amberlite IR-100-H by 1 8 % Hydrochloric A c id
ab o u t 50% of the vitam in is recovered in th e first ten volumes of acid regenerant, each equal to the bed volume; more th a n 75%
is recovered if tw enty-five volumes are used.
A portion of synthetic zeolite was packed in a sim ilar column an d pretreated w ith 25% sodium chloride solution, and upon it were adsorbed 25 mg. of vitam in B*. T he vitam in was eluted w ith 25% sodium chloride; the same general ty p e of recovery w ith ab o u t the sam e degree of efficiency was obtained as w ith 18% hydrochloric acid on IR -100-H (Figure 2).
T he use of even stronger acid to improve the characteristics of th e recovery process was contem plated, since no deleterious
ef-VOL. SALT PER ef-VOL. ZEOLITE 4.87 9.74 14.6 19.4 2 4 3
>-<z Id>
oO Ui Q:
Table I. Adsorption of Thiamine by Various Adsorbents [2000 ml. of t h i a m i n e (1 p . p .m .) so lu ti o n p as sed a t a flow r a t e of 20—35 ml.
p er min.J
B ed v olum e, ml.
B ed d e p th , cm . P a r tic le size, mesh T h ia m in e rem o v al
A c tiv a te d A lu m in a 4 5 . 6 2 5 . 4 80- 2 0 0
N one
S y n th e tic Zeolite
4 0 . 6 2 5 . 4 3 0 -5 0 C o m p lete
A m b er lit e I R - 1 0 0 - N a
4 1 . 2 2 5 . 4 3 0 - 5 0 C o m p l e te
A m b e r l ite IR -4 4 1 . 1 2 5 . 4 3 0 - 5 0 Tr a c e
feet was noted upon th e resin by co n tact w ith 18% hydrochloric acid for a considerable tim e during repeated runs. T o t h a t end th e usual ty p e colum n was packed w ith 41.0 ml. of A m berlite IR -100-H over w hich were passed 8 liters of a 250 p.p.m . thiam ine solution a t a ra te of 35 ml. (equivalent to 5 gallons per square foot) per m inute. Since no detectable a m o u n t of th e v itam in could be discerned in th e effluent even a t th is high concentration an d flow rate, it was evident t h a t 2 gram s of v itam in h a d been adsorbed. T h e recovery w as carried o u t w ith 37% hydro
chloric acid a t a ra te of 7 ml. per m in u te (equivalent to 1 gal
lon per square foot per m in u te); tw e n ty 50-ml. sam ples w ere col
lected.
T he characteristics of th e recovery curve are in d icated in F igure 3 w hich shows t h a t th e elution is m o st ra p id in th is system , for the peak is reached a fte r only tw o volum es of eluent, each equal to the volum e of the adsorbent, are passed. I t will be seen th a t 60% of the vitam in is recovered in th e first five volumes;
if tw en ty volumes or m ore are passed, th e recovery efficiency approaches 100%.
W hile concentrated hydrochloric acid is preferable to sulfuric acid because of the density an d th e th erm al an d oxidative char
acteristics of th e latte r, th e recovery m ay be carried o u t ju st as readily w ith 32% sulfuric acid (Figure 4).
Of in terest in regard to possible, comm ercial ap p licatio n is the to ta l capacity of the resin for th e v itam in. T h ro u g h a 10-cra.
bed of A m berlite IR -100-H (bed volum e 18.7 m l.) w as passed a 250 p.p.m . solution of thiam ine hydrochloride a t a ra te of 35-50 ml. (equivalent to 5 -7 gallons per square foot) per m inute. This corresponds to a reten tio n tim e of 22-30 seconds, b u t even higher
VOL. ACID PER VOL. IR-I00-H
4 .88 9.76 14.65 19,50 24.40
<E UJ> oo
U I
£E
Figur« 2. Recovery of Thiamine from Synthetic Z e o lite by 2 5 % Sodium Chloride
Figure 3. Recovery of Thiamine from Amberlite IR -in o u b y 3 7 % Hydrochloric A c id w - n
July, 1945 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 633
Table il.
S am p le N o.
O riginal concn., p .p .m . E q u ilib r iu m concn.,
p . p . m .
C a p a c ity , m g ./g . d r y a d s o r b e n t
A d s o r p tio n o f Thiamine b y A m b e r lite IR -1 0 0 -H and Synthetic Z e o l it e (Freundlich A d so r p tio n Isotherms)
A m b erlite IR -1 0 0 -H
9 6 . 2 7 2 . 2 4 8 . 2 2 4 .1
0 0 0 0
G re a te r th a n 65 a t 100 p .p .m .
S y n th e tic Zeolite
9 2 . 6 7 2 .2 4 8 . 2 2 4 .1 4 3 . 5 2 0 . 2 1 7 .5 4 . 4 2 8 .7 28 2 1 6 .7 1 0 .7
A m b erlite IR -1 0 0 -H
1420 1064 710 355 990
291
822 470 181 164 162 118
rates of flow w ould be feasible if a deeper bed of resin were used (Figure 5).
A sm all concentration of the v ita m in w as noted in the effluent ju st as th e fifth liter h ad passed, a t which point the capacity is 68 mg. thiam ine per ml. of resin (4.2 pounds thiam ine per cubic foot) or 168 mg. per gram dry weight. A lthough the ra te a t which th e concentration of thiam ine in th e effluent increases as a function of the volume of solution passed is fairly rapid, even after 12 liters have been treated the column is still 90% efficient.
T his means th a t a ty p e of countercurrent operation employing a second column of exchanger to adsorb th a t vitam in which has passed th e first would considerably increase the practical ca
pacity of the resin.
T he am enability of the resin to sta tic experim ents is indicated in T able I I and Figure 6. These d a ta were obtained by stirring 0.200 gram of resin (or zeolite) in 100 ml. of solution for 3 hours and th e n determ ining the residual concentration of thiam ine in the solution.
T he binding tendency of A m berlite IR -100-H is so strong a t low concentrations th a t essentially all th e thiam ine is removed;
only a fte r a m uch more concentrated solution was substitu ted was a m easurable equilibrium concentration found. T he ca
pacity of around 160 mg. per gram is in good agreem ent w ith th a t from th e column technique.
ADSORPTION OF RIBOFLAVIN BY AMBERLITE ¡R-100-H Se p a r a t i o n o f Th i a m i n e f b o m Ri b o f l a v i n. Since th e rela
tiv e degree of basicity of these two compounds would be the
determ ining factor in any successful separation, it was necessary to establish conditions under which riboflavin is adsorbed by A m b e r l i t e I R - 1 0 0 - H E ig h t liters of a solution of 4.7 p.p.m . riboflavin were passed downflow over a bed (volume 40 ml., diam eter 1.5 cm.) of Am berlite IR -100-H a t a ra te of 7 ml. per m i n u t e . L e a k a g e o c curred throughout the ra n
an d h ad risen to a 50% value a t the end of the eighth liter. T he average leakage over th e entire ru n was 2 6 % F u rth e r explor
atory experim ents showed th a t under the sam e conditions the leakage from fuller’s e arth was never greater th a n 0.1 p.p.m., w hereas A m berlite IR -4, th e acid-active m aterial, failed to adsorb riboflavin.
T he adsorption was th en checked in a static system where ex
cessive contact tim e could be allowed. F reundlich adsorption isotherm s were ru n using 0.200 gram of resin (or fuller’s earth) stirred w ith 100 mL of riboflavin solutions of various initial con
centrations for 3 hours. T he results are shown in T able III.
Therefore, th e equilibrium to ta l capacity of IR -100-H is equal to th a t of fuller’s earth, although in a dynam ic system w ith lim ited contact tim e th e riboflavin is bound m uch more
effec-Figure 5 . A d s o r p tio n of Thi
am ine b y A m b e r lite IR-10Û-H
Table 111. A d s o r p tio n of Riboflavin b y A m b e r lite 1R -100-H and Fuller's Earth (Freundlich A d s o r p tio n Isotherms)
A m b erlite IR -1 0 0 -H F u lle r ’s E a r t h
S a m p le N o . 1 2 3 4 1 2 3 4
O riginal concn.,
p .p .m . 5 0 .0 3 7 . 5 2 5 . 0 1 2 .5 5 0 . 0 3 7 . 5 2 5 . 0 1 2 .5 E q u ilib riu m concn.,
p .p .m . 2 8 . 3 2 1 .9 1 5 .6 7 . 3 2 7 . 6 1 7 .2 1 3 .0 5 . 7 C a p a c ity , m g ./g .
d ry a d s o r b e n t 1 2 .9 9 . 3 5 . 6 3 . 1 1 2 .3 1 1 .2 6 . 6 3 . 7
Figure
VOL ACID PER VOL IR-IOO-H
4 . R eco v ery ©f Thiamine from A m b e r lite 1R-10Q-H b y 3 2 % Sulfuric A c i d
tZ a
tiim
100 1000
R E S ID U A L THIAMIN C O N C E N T IO N - PP.M . 8 Figure 6 . A d so rp tio n Isotherm of Thiamine b y A m b erlite
IR -1 0 0 -H and Synthetic Z e o l it e
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 Vol. 37, No. 7
tively by the fuller’s earth . I t is also evident th a t th e affinity of riboflavin for IR -100-H is m uch less pronounced th a n for thiam ine. Riboflavin is likewise liberated reversibly to strong acid, b u t th e concentration of th is acid need n o t be so great as for thiam ine recovery.
F rom these d a ta a separation of riboflavin from thiam ine should be possible if th e contact tim e of solution w ith resin is properly adjusted. A bed of IR -100-H (bed volum e 41.2 ml., diam eter 1.5 cm.) was packed, and through it were drained 5 liters of a solution composed of 150 p.p.m . thiam ine hydrochloride and 10 p.p.m . of riboflavin a t a ra te of 35 ml. per m inute. T here was no leakage of thiam ine, although th roughout th e last 60%
of th e ru n complete leakage of riboflavin was noted (Table IV).
Table IV. Separation of Riboflavin from Thiamine by Amberlite IR-100-H
L iters of L eak ag e, P . P .M . L iters of L eak ag e, P .P .M . S o lu tio n R iboflavin T h ia m in e S o lu tio n R ib o flav in T h iam in e
1 6 . 0 0 . 0 6 9 . 9 0 . 0
2 8 . 2 0 . 0 7 1 0 . 2 0 . 0
3 9 . 0 0 . 0 8 1 0 . 6 0 . 0
4 9 . 5 0 . 0 9 1 0 .3 0 . 0
5 9 . 7 0 . 0
T he ap p aren t initial adsorption of riboflavin is probably due to dilution of the influent by the w ater on the resin a t the beginning of the ru n and to a small degree of adsorption of the v an der W aals type. This m eans of separation is b o th controllable and practical. Since it is possible to replace one nitrogeneous sub
stance adsorbed upon IR-100-H by another more alkaline sub
stance, an additional m eans of elution is available and has been used in a few instances.
IS O L A T IO N O F TH IA M IN E FR O M RICE B R A N EXTRACT
T h e practicability of th e procedure was tested by the extrac
tion of thiam ine from a rice bran extract. Using a flow ra te of about 50 ml. (equivalent to 7 gallons per square foot) per m inute w ith a shallow bed (12 cm.), complete leakage of riboflavin was noted while about 60% of the to ta l thiam ine in th e solution was adsorbed. I t seems probable th a t only the free vitam in was ad
sorbed, b u t acid or enzym atic hydrolysis of th e cocarboxylase (pyrophosphate ester) or other complexes converts the vitam in
t o a fo rm a m e n a b le to a d s o rp tio n b y A m b e rlite IR -1 0 0 -H . R e co v ery c h a ra c te ris tic s of a n a c id re g e n e ra tio n o f th e re sin a r e sim ila r to th o se d e sc rib e d p re v io u sly . T h e v ita m in s c a n b e s u b s e q u e n tly release d fo r u se b y c h em ical m e a n s a lre a d y w ell know n.
A P PLIC A TIO N S
T h e a p p lic a tio n o f th is m e th o d m a y p ro ceed in tw o d ire c tio n s - a s a n a n a ly tic a l to o l o r in th e re c o v e ry of th ia m in e a n d rib o fla v in fro m w a ste so lu tio n s. A s a n a n a ly tic a l to o l th e p u r ity o f th e e lu a te is s a tis fa c to ry fo r th io c h ro m e a n a ly sis. N o in te r fe r in g su b sta n c e s h a v e b e e n e n c o u n te re d . I f d e sire d , th ia m in e m a y b e re a d ily c ry sta lliz e d fro m th e e lu a te w ith o u t e x te n siv e f u r t h e r p u rific a tio n . As a pro cess to o l to re c o v e r th ia m in e fro m n a t u r a l sources, th e re sin offers a d is tin c t p o te n tia l a d v a n ta g e o v er z eo lite a s a n a d s o rb e n t; n a m e ly , e lu tio n c a n b e acco m p lish ed w ith a v o la tile a c id in ste a d o f a s a lt w h ic h c a n n o t b e rem o v ed re a d ily fro m th e e lu a te . T h e v o la tile a c id c a n b e re c o v e re d a n d u se d re p e a te d ly .
LITERATURE CITED
(1) A dam s, B . A., an d H olm es, E . L ., J . Soc. Chem. I n d .} 54, 1—6T (1935); B rit. P a te n ts 450,308-9 (Ju n e 13, 1936), 474,361 (N ov. 25, 1937); F re n ch P a te n ts 796,796-7 (A pril 15, 1936):
U. S. P a te n ts 2,104,501 (Ja n . 4, 1938), 2,151,883 (M arch 28, 1938), 2,191,853 (Feb. 27, 1940).
(2) C onner an d S tra u b , I n d . E n g . Ch e m., A n a l . E d . , 13, 380 (1941).
(3) G addis, S., J . Chem. Education, 19, 327 (1942).
(4) H arrisson, J . W . E ., M yers, R . J., an d H e rr, D . S., J . A m . P harm . Assoc., 32, 121 (1943).
(5) H ennessy and Cerecedo, J . A m . Chem. Soc., 61, 179 (1939).
(6) K avanagh, I n d . E n g . Chem ., A n a l . E d . , 13, 108 (1941).
(7) M yers, F . J., I n d . E n g . Ch em., 35, 858 (1943).
(8) M yers, F . J ., Proc. 2nd A n n . Water Conf., E ng. Soc. Western Pa\, 1941, 133.
(9 ) M yers, R . J., and E astes, J . W ., In d. En g. Ch e m., 3 3 , 1203 (1941).
(10) M yers, R . J., E astes, J. W ., and M yers, F . J., Ib id ., 33, 697 (1941).
(11) M yers, R . J., E astes, J. W . an d U rq u h a rt, J. D ., Ibid., 33, 1270 (1941).
(12) S train , H . H ., “ C hrom atographic A dsorption A nalysis” , New Y ork, Interscience P ublishers, 1942.
(1 3 ) S t r a i n , H . H . , In d. En g. Ch e m., An a l. Ed., 1 4, 2 4 5 ( 1 9 4 2 ) . (14) Zechm eister an d Cholnoky, “ D ie chrom atographische
Adsorp-tionsm ethode. G m ndlagen, M ethodik, A nw endungen” , 2nd ed., Berlin, Ju liu s Springer, 1938.
Pb e s e n t e dbefo re th e D iv isio n of B iological C h e m is tr y a t t h e 1 0 6 th M eetin g of th e Am e r i c a n Ch e m i c a l So c i e t yin P itt s b u r g h , P a .
Amberlite Plant Unit to Produce Process Water Vitamin A ssay after Amberlite Separation