Adsorption of Riboflavin by Lactose

ABRAHAM LEVITON. Bureau of Dairy Industry, U. S. Department of Agriculture, Washington, D. C. IN. THE present com- mercial process for the manufacture...
3 downloads 0 Views 596KB Size
Adsorption of Riboflavin by Lactose INFLUENCE OF TEMPERATURE ABRAHAM LEVITON Bureau of Dairy Industry, U. S. Department of Agriculture? Washington, D. C . comThis work follows a study previously reported ( I ) on the inThe resulting stock solution had a PH of 4.6 and confluence of concentration. A knowledge of the influence of mercial process for the manufacture of lactose, temperature furnishes the most adequate basis for the t$$lo*1,",$?3~~& ' ~ grain curd casein whey is production of adsorbates (milk sugar containing adsorbed $0 15-ml. samples, lactose neutralized and boiled to riboflavin) as by-products in the manufacture of the milk and synthetic riboflavin were sugar of commerce. To utilize the productive capacity of added t o Yield solution^ of coagulate proteins and insoluble salts. The filtrate existing milk sugar plants for the manufacture of adsorb~ o ; ? ' ~ is then concentrated under ates, it is necessary to establish conditions for the conc.ontaining these solutions vacuum to approximately trolled preparation of these adsorbates. The purpose of and shielded from light by foil were rotated in a thermo70% solids, and the sugar this paper is to furnish data to operators for the establishstat, the temperature Of which ment of these conditions. The range of concentrations is permitted to crystallize was mainkained conatant under conditions facilitating and temperatures studied include those which would be within 0.2 . Temperatureu the establishment of the encountered in the manufacture of adsorbates from grain ranged from 5 O to 63" C. proper grain size. Under curd casein whey and are applicable to a wide range of Riboflavin determinations operating conditions. Although the data are considered were made fluorimetrically by these conditions even if all a Duboscq nephelometer with of the riboflavin present in primarily from the standpoint of their practical applicathe following whey were adsorbed, a weak tion, a number of problems of both theoretical and pracGeneral Electric type H3, 85-watt, mercury vapor lamp; adsorbate would be obtained, tical interest are considered. Corning glass filters No. 511 inasmuch as the ratio be(to limit wave length band tween the quantity of lactose of incident light) and No. 352 crystallizing from whey and the quantity ot riboflavin in (to limit wave length band of fluorescent light). The content.; of the various tubes were filtered rapidly (within 5 seconds), and the whey is large. Actually all of the riboflavin is never adsorbed. Furthermore, experiments in this laboratory indicate that the lactose Was washed with 70% The uantity Of adsorbate recovered was weighed directly. The re&s checked quantity adsorbed may prove variable. well with polarimetric determinations of lactose. Blanks run at T o obtain concentrated adsorbates in good yields and under the higher tem eratures indicated that no significant quantity of controlled conditions, it is necessary to modify the present comliboflavin was Post. mercial process in such a way that th grystallization operation is EFFECT O F COMPLETE CRYSTALLIZATION conducted in two steps: in the first only insignificant quantities, Adsorption results under conditions of complete crystallization and in the second the greater quantity of riboflavin would be adrZregiven in Tables I and 11 and Figure Riboflavin consorbed. The phenomenon UilderWg the c ~ n d u c tof the first centration is expressed as the number of micrograms contained step lies in the existence of a critical concentration of riboflavin initially in ml, of solution just saturated with lactose at the below which adsorption does not occur. This critical concentratemperature indicated. Table I concentrations are referred to tion is a function both of lactose concentration and temperature. just saturated with lactose at c. In Table 11 conIf crystallization of lactose can be carried out under such condicentrations are expressed in terms of solutions saturated at 5 ~ , tions that thc critical concentration is not exceeded until a conand at the temperatures indicated. Crystallization in these siderable portion of lactose hm crystallized, then the first step experiments proceeded to the point at u,hich solutions were satuof the crystallization operation may be effected. rated with lactose. T o ascertain the limits of lactose and riboflavin concentration in Figure indicate that, as the concentration of The under which this first step in the crystallization operation can be ribaflavin increases, only traces of riboflavin are adsorbed until a carried out, it is necessary to generalize and extend considerably critical concentration is exceeded. ~h~ degree of adsorption then the data previously recorded (I), to apply not O ~ to Y Cheddar rises sharply and increases linearly at 50 and 2 8 O c. writh incheese \rhey (2) but also to grain curd casein whey, and not only creasing concentration. Even at 50.8' C. the increase appears at one temperature, 5" c., but Over a range Of temperatures (8). to be linear, but the data have not been sufficiently extended to The critical concentration increases markedly with temperature; decide this point. There are then, two definite stages in the consequently, by choosing a temperature Sufficiently high, it is adsorption process it relates to the influence of riboflavin conpossiblc to crystallize a considerable portion of lactose free of centration; in the first stage traces of riboflavin are a& riboflavin. Then by lolqrering the temperature it becomes POSsorbed, and in the second, the degree of adsorption increases with sible to crysta~lizeanother batch of lactose containing adsorbed increasing concentration, linearly, at 5 0 and 28" C. and perhaps riboflavin in good yields. Tables are given from which the even at higher temperatures. At high temperatures the transioperator may choose a t his dimretion operating conditions with tion between the two stages is gradual than at lower I espect to concentration and temperature. These conditions peratures. would then detcrmine the concentration and the yield of the w&s stressed previous~y (f). The empirical The linear adsorbate. formula, EXPERlMENTAL TECHNIQUE a = (c - 2.5)/s (12 where c = initial riboflavin concentration, -y/ml. The technique was essentially the same as that previously rea = degree of adsorption, -y/gram ported (f): A stock solution was prepared from mother liquor s = excess lactose at 5" C,, grams/ml. obtained in the manufacture of milk sugar from grain curd 2 5 = minimum riboflavin concentration a t 5" below casein whey. This liquor was stirred at 5" C. t o remove excess which adsorption doe.: not occur, -y/ml. lactose and treated with decolorizing carbon to remove riboflavin.

I

N T H E present

744

~

~

August, 1944

INDUSTRIAL AND ENGINEERING CHEMISTRY

derived to apply to the solutions studied previously, did not apply generally to the solutions of the present investigation. At 5" C. a minimum critical concentration of 2.5 was found, which agrees favorably with the value previously obtained. However, the measured values of residual riboflavin, unlike the values established previously, were not equal to the minimum concentration nor independent of the concentration of riboflavin. Formula 1may be generalized toapply, withincertain temperaturc limits, to all of the solutions studied. Thus,

-

a = k - (c-cJ/(s st) (2) where k = a constant ct = critical concentration below which adsorption does not occur at temperature t s = initial lactose concentration, grams/ml. st = solubility of lartose a t f

From data obtained here and previously reported, it appears that a t low supersaturation levels with respect to lactose there is a range of lactose concentration within which k in Equation 2 may be taken as unity. Within this range of concentration, ut 5" and 28" C. and probably within the range 5-28.0" C., k does not differ significantly from unity. For highly supersaturated. solutions k is lesa than unity. This point is brought out strikingly in curves 1, ,'l 2, and 2' of Figure 1. The lines representing adsorption at 28" from solutions containing, respectively, 0.56 and 0.20 gram of excess lactose per ml., converge to a point on the horizontal axis representing 7.5 micrograms per ml. The respective values of k deduced from the slopes of these lines and from the quantities of lactose recovered are 0.84 and 1.0. Similarly, the values of k deduced from the slopes of curves 1 and 1' (representing adsorption a t 5" from solutions containing 0.63 and 0.20 gram excess lactase) are 0.88 and 1.0, respectively. The value of k cannot exceed unity as long as conditions are such that c, remains invariant; that is, the measured or extrapolated values of ct, the critical concentration, can be less than but cannot exceed the measured value of residual riboflavin concentration for any given initial value of c. From a practical standpoint the observation that k can be less than but cannot exceed unity indicates that the equation, a(s,

- st)

= c

-

Cb

(3)

may be employed to calculate the maximum quantity, a ( s - St), of riboflavin which, given the temperature and the initial riboflavin concentration, can be adsorbed. This maximum quantity is independent of the initial lactose concentration. The relation between adsorption on growing crystals and temperature has received little notice in the past. The decrease in the degree of adsorption on growing lactose crystals with increasing temperature might be reasonably anticipated from the known behavior of insoluble adsorbents. However, the decrease attending adsorption on the growing crystals can be attributed, under conditions in which: k is equal to unity, only to the decrease with increasing temperature in the value of the critical concentration below which adsorption does not occur'. Under these conditions, once adsorption begins, the rate at which the degree of adsorption changes with riboflavin concentration becomes independent of temperature. The manner in which the critical concentration decreases with increasing temperature is shown in Table 11. The critical values corresponding at the various temperatures to the values for the degree of supersaturation are given in the third and fourth columns. The critical concentration is a function of temperature and is greater, the higher the temperature. Because it was deemed desirable to establish these critical concentrations with a minimum of extrapolation, the concentration range studied did not extend much beyond the highest evaluated critical concentration. At relatively low temperatures this concentration range included a sufficient number of points to determine the relation between the degree of ndsorption and riboflavin concentration.

745

At higher temperatures, however, only the critical concentration values could be ascertained $atisfactorily. Ascertained definitely at all temperatures WRS the region of transition between the first and second stages of adsorption. Because of the more gradual nature of the transition at 50.8-63" C., the critical values of the minimum were less clearly defined than at lower temperatures. The minimum values established s t these temperatures were those a t which the degree of adsorption fell below one microgram per gram. EFFECT O F VARIATION IN DEGREE OF CRYSTALLIZATION

Tables I and I1 show in detail the progress of adsorption at various temperatures as a function of tho degree of crystallization. The previous paper ( 1 ) indicated that the critical riboflavin concentration is a function of the instantaneous concentration of lactose, and that this critical concentration decreases as the degree of supersaturation with respect to lactose decreases, and approaches a final minimum value as the point of saturation is approached. These critical values were not established over a sufficiently wide range of temperatures and lactose concentration to satisfy practical requirements. One purpose of the present study wm the tabulation of data necessary to establish these values. The tables are drawn up to emphasize the practical significance of the phenomenon. From a practical standpoint, the residual riboflavin recoverable following the first stage of lactose crystallization is important, These values are given in the last column of Table I as the maximum percentage of riboflavin initially present which can be recovered a t 5" C., under conditions of crystallization such that k is equal to unity. These values were calculated by the formula: (''

- 2*5) x 100 C,

e:

max. % riboflavin recoverable

where c, = residual riboflavin concentration. Consider, for example, EI solution containing 32.4 micrograms of riboflavin per ml. and 0.77 gram or more of lactose per ml. The last three columns of Table IB indicate that, if crystallization proceeds to the point at which 0.45 gram of excess lactose remains

d

5.0 5.0 2- 28.0 2!.28.0 3- 50.8 4- 57.5 5- 63.0 I; I-

RIBOFLAVIN CONCENTRATION,

MICROCRAMS PER MILLILITER Figure 1. Relation between Degree of Absorption and Riboflavin Concentration at Various Temperature Level6

Vol. 36, Na. B

INDUSTRIAL AND ENGINEERING CHEMISTRY

746

TABLE I. ADSORPTIONAS LACTOSE CRYSTALLIZES AT VARIOUSTEMPERATURES, YIELDS,AND DEQREPJOF ADSORPTIONON RESIDUALLACTOSEAT 5 C. .Bc .

3? 4%

44 8

5.4 10.8

18.2

21.6

32.4

43.2

5.4 10.8

16.2

21.6

32.4

43.2

0.23 0.50 0.83 0.25 0.30 0.53 0.63 0.17 0.26 0.34 0.53 0.83 0.09 0.27 0.31 0.53 0.63 0.13 0.19 0.26 0.31 0.53 0.63 0.11 0.22 0.28 0.30 0.53 0.63 0.56 0.24 0.33 0.40 0.45 0.58 0.19 0.30 0.38 0.44 0.58 0.23 0.31 0.38 0.44 0.56 0.20 0.30 0.38 0.45 0.56 0.18 0.31 0.37 0.45 0.56

A. 0.10 1.5 3.2 0.18 0.8 12.5 11.8 0.40 2.0 13.4 21.3 18.4 0.15 12.5 27.0 31.3 26.3 0.80 22.0 37.5 49.0 52.5 43.5 0.80 33.5 68.4 70.0 68.3 55.8 B. 0.15 0.1 0.3 0.5 1.4 4.0 0.08 0.3 3.0 9.4 12.5 0.25 0.50 12.0 15.6 20.0 0.75 18.0 25.0 32.5 34.4 2.8 32.5 55.8 51.0 49.4

Adsorption at 5O C. 0.023 5.4 0.40 0.75 4.7 0.13 2.0 3.4 0.00 0.05 10.8 0.38 0.24 10.6 0.33 6.6 4.2 0.10 7.3 3.5 0.00 0.07 18.1 0.46 0.52 15.7 0.37 4.6 11.8 0.29 11.3 4.9 0.10 11.6 4.6 0.00 0.01 21.6 0.54 3.4 18.2 0.38 8.4 13.2 0.32 18.6 5.0 0.10 18.6 5.0 0.00 0.10 32.3 0.42 32.0 22.6 9.8 15.2 17.2 27.8 4.6 27.4 5.0 0.08 43.1 0.53 7.1 0.41 36.1 19.1 0.35 24.1 21.0 22.2 0.33 35.1 8.1 0.10 35.1 8.1 0.00 Adsorption at 28O C. 0.08 5.3 0.07 0.02 10.8 0.39 0.30 0.1 10.7 0.2 10.8 0.23 0.6 10.2 0.18 2.2 8.6 0.07 0.02 18.2 0.44 0.1 18.1 0.33 1.1 15.1 0.25 4.1 12.1 0.19 7.0 9.2 0.07 0.06 21.5 0.40 0.16 21.4 0.32 4.8 17.0 0.25 6.9 14.7 0.19 11.2 10.4 0.07 0.15 32.2 0.43 5.4 27.0 0.33 0.25 9.5 22.9 0.18 14.6 17.8 19.3 13.1 0.07 0.50 42.7 0.45 10.1 33.1 0.32 20.7 22.5 0.28 23.0 20.2 0.18 27.7 15.5 0.07

7.3 16.9

53.7 40.7

5.4

21.8 24.5 17.0

78.8 76.0 15.7

10.8

29.6 35.1 31.4 24.0

84.0 81.5 68.1 14.8

16.2 21.6

35.4 43.6 33.6 25.0

88.3 72.7 49.6 11.6

32.4

59.6 67.1 54.3 45.9 21.0

92.0 91 .o 62.0 45.3 6.5

78.8 82.0 81.7 59.7 66.0

94.0 77.8 50.0 45.6 13.0

40.0 21.3 27.3 35.2 42.7 82.2 31.2 41.2 50.4 50.5 95.7 47.5 59.0 58.0 64.2 113.0 69.0 74.3 81.6 85.0 152.0 89.3 95.6 76.9 98.3 188.0

51.8 78.8 75.9 75.0 71.3 56.4 84.5 84.0 77.8 59.2 41.3 87.8 87.5 87.1 56.4 36.8 91.6 75.6 63.0 47.2 32.8 93.0 70.8 46.3 40.9 30.1

...

... ...

...

...

...

.. ..

..

..

..

per ml. of solution and if the solution is then filtered, 91% of the riboflavin originally present is recoverable at 5' C. in the form of an adsorbate containing 68.6 micrograms of riboflavin per gram. If, however, crystallization proceeds to the point at which 0.33 gram of excess lactose remains per ml. of solution, then only 75% of the riboflavin originally present is recoverable, although in slightly higher concentration. The sharp drop in yield occurs because at approximately 0.45 gram of excess lactose per ml. (Table 11)the minimum concentration of riboflavin below which adsorption does not occur is equal to 32.4 micrograms per gram; that is, only when the concentration of lactose becomes less than 0.45 gram per ml. does adsorption on lactose take place from solutions containing 32.4 micrograms of riboflavin per ml. Once adsorption begins, it proceeds rather rapidly and approximately linearly, only to fall off in the late stages of the crystallization process. Adsorption from highly supersaturated solutionB proceeds in three phases-a lag phase in which crystallization takes place

43.2

21.6 32.4

43.2

82.4 43.2 52.0 74.5

0.32 0.38 0.43 0.31 0.46 0.43 0.36 0.37 0.43 0.13 0.24 0.35 0.38 0.43 0.08 0.23 0.35 0.43 0.01 0.13 0.23 0.39 0.43 0.30 0.35 0.19 0.22 0.32 0.35 0.13 0.24 0.32 0.35 0.27 0.27 0.27 0.27

C. Adsorption at 50. So C. 0.05 0.02 5.4 0.31 0.05 0.02 5.4 0.27 0.05 0.02 5.4 0.20 0.05 0.02 10.8 0.32 0.075 0.03 10.8 0.27 0.075 0.03 10.8 0.20 0.12 0.04 16.2 0.27 0.12 0.04 16.2 0.26 0.12 0.05 16.1 0.20 0.07 0.01 21.6 0.50 0.09 0.02 21.6 0.39 0.12 0.04 21.6 0.28 0.60 0.23 21.4 0.25 0.65 0.24 21.4 0.20 0.22 0.01 32.4 0.57 1.7 0.39 32.0 0.40 7.5 2.6 29.8 0.28 6.3 2.7 29.7 0.20 0.52 0.05 43.2 0.62 3.8 0.47 42.7 0.50 8.0 1.8 41.4 0.40 17.5 6.8 36.4 0.24 17.5 7.5 35.7 0.20 D . Adsorption at 57.5O C. 0.13 0.04 21.6 0.33 0.20 0.07 21.5 0.28 0.1 0.02 32.4 0.44 0.1 0.02 32.4 0.41 0.5 0.18 32.2 0.31 0.7 0.25 32.1 0.28 0.4 0.05 43.1 0.50 2.5 0.60 42.6 0.39 5.0 1.80 41.6 0.31 5.0 1.75 41.4 0.28 Adaorption a t 83O C. 0.3 0.08 32.3 '0.36 0.3 0.08 43.1 0.36 1.8 0.49 51.5 0.38 16.0 4.3 70.2 0.36

.Solubility of lactose at ' 5 C., tration, 0.77 gram/ml.

- 0.14 gram/ml.;

9.4 10.7 14,5 25.9 30.7 41.5 50.6 52.5 68.0 38.2 49.0 88.3 75.7 94.5 52.4 73.7 97.5 138.0 65.7 80.4 94.7 141.0 166.0

53.7 53.7 53.7 76.8 78.8 76.8 84.6 84.6 84.0 88.8 88.3 88.3 87.4 87.4 92.4 91.2 85.0 84.7 94.2 93.1 90.1 78.4

76.9

57.9 68.2 67.9 73.0 96.9 108.0 81.2 103.0 128.0 139.0

88.3 88.0 92.3 92.3 91.7 91.4 93.8 92.7 90.5 90.0

82.8 114.0 138.0 200.0

92.0

93.8 94.3 90.8

initial laetose oonoen-

without significant adsorption, an adsorption phase in which crystalliiation is attended by rapid adsorption, and a saturation phase in which crystallization approaches completion with practically no further adsorption. At high temperatures the transition between the first and second phases is not so sharp as the transition at lower temperatures. The approximate location of the points of transition at 5' and 28" for given initial values of c is given in TableII. Thew points represent the lactose concentrations at which lactose will. just begin to adsorb from a solution containing c micrograms, riboflavin per ml. These points were located by extrapolation of the points on the adsorption phase of curves plotted (from the data of Table 1.4 and B) of the relation between the quantity of riboflavin adsorbed and the quantity of lactose crystalliied at the various levels of initial riboflavin concentration indicated in t h e last column of Table 11. These points were looated only approximately but with sufficient accuracy for practical purposes. DISCUSSION

The data presented are intended to apply principally to grain curd casein whey proceesed for the removal of proteins and calcium salts as in the manufacture of milk sugar. There is no reason to assume, however, that the data are not equally valid or at least approximately valid for various types of cheese whey, The relation, U k(C-2.5)/(8-80)

INDUSTRIAL AND ENGINEERING CHEMISTRY

Auqwt, 1944

applies a t 5' C. equally well for the solutions described here and for solutions originating from Cheddar cheese whey (I). The stock solution used previously (I), exclusive of ita lactose content, possessed approximately the same concentration of total solids and the same pH as the one employed in the present study. Although the differencesin composition probably existed, they were insufficient to affect the general relation between adsorption and concentration. Of practical significance was the difference in the rate of crystallization. It was.previously reported that crystalliqation under the condition of seeding and stirring employed was complete in 3 to 5 weeks. In the present investigation crystalliiation under comparable conditions of concentration was complete within 3 days.

RIBOFLAVIN CONCENTRATION VALUESAT TAB^ 11. CRITICAL VARIOUS LACTOSE LEVELS AND TEMPERATURES Supersatn.

TZme 6.0

28.0

with Reapeat to Lpotolea, Gmm/Ml. 0.0 0.20 0.30 0.36 0.40 0.60 0.66 0.0 0.20 0.26 0.30 0.45 0.60 0.0 0.0 0.0

Critioal Concn. of Riboflavin@, r/ML 2.6 6.4 10.8 16.2 21.6 32.4 43.2 7.5 10.3 15.5 20.7 31 .O 41.3 ~. 19.0 27.0 37.6

Critioal Conon. of Riboflavinb, y/Ml. 2.6 6.4 10.8 16.2 21.6 32.4 43.2 7.8 10.8 16.2 21.6 32.4 43.2 21.8 32.9 48.9

50.8 67.6 63.0 a Referred to solution saturated with lactose at temperature indicated in b t column. a Referred to aolution saturated with lactose at 6* C.

The value of k may be dependent on the rate of crystallization, and this may account for the lower values obtained at the higher lactose levels. If this condition obtains, maximum yield can be obtained only when the rate of crystallization is sufficiently low. The yield values given in the last column of Table I are idealized and are results of calculations based upon the aasumption that it is possible to realize conditions of crystallization in which k is equal to unity. Although this is not always the case, in the preparation of concentrated adsoybates, conditions are met under which k is equal to unity. The values in Table I1 relating critical riboflavin concentration with lactose concentration obtain when the initial lactose concentration is equal to 0.77 gram per ml. It is not possible without further experimentation to say whether these values are independent of the initial lactose concentration. The data plotted in Figure 1 indicate that, for equilibrated lactose solutions, the critical concentration is independent of initial lactose concentration: consequently it might be reasonable to assume that for supersaturated solutions the critical values a t various lactose levels are ala0 independent of the initial lactose concentration. I n the manufacture of milk sugar from grain curd casein whey, operating conditions we such that the initial lactose concentration exceeds 0.77 gram per ml. On theoretical grounds it is plausible to m u m e that the degree of adsorption is a function of reiative diffusion rates, and consequently that increases in lactose concentration a t the same riboflavin level should result in a lowering in adsorption values. Such a lowering might be reflected in a lowering of critical values; on the other hand, it might be reflected merely in a decrease in the rate of adsorption, once adsorption begins. The maximum values cited in the next to the last column of Table I for the degree of adsorption were calculated by the formula,

a

(C

747

- 2 . 5 ) / ( ~- 0.14)

These values may be exceeded in practice by dilution of the mother liquor in which the second stage of crystallization is effected. APPLICATION OF DATA

Pilot plant investigations on the application of the data to the preparation of concentrated adsorbates from whey and from related riboflavin-containing liquors will be reported separately. However, an example is given here which illustrates an operable process and the operations involved. The steps involving neutralization, heat coagulaEXAMPLE. tion of proteins and insoluble calcium salts, decantation, and filtration were identical with those employed commercially (4) in the manufacture of lactose from grain curd casein whey. Care was used to shield the hot liquor from light. The 400 pounds,of clarified whey obtained aa a result of these operations contained 6.0% solids, 4.8% lactose, and 1.0 microgram riboflavin per gram. This liquor was concentrated in a vacuum pan to contain 64.4% solids, 52.4% lactose, and 11.0 micrograms riboflavin per gram. Just before draining, the concentrate was brought to approximately 60" C. The concentrated liquor (36.5 pounds) waa mixed in a jacketed mixer equipped with agitators mounted concentrically and geared to rotate in opposite directions. The temperature of crystallization was maintained a p proximately a t 55-60" C. After 4 hours the warm mixture waa filtered by a centrifuge. The wet unwashed crude milk sugar (14.8 pounds) recovered in this manner contained 13.6 pounds solids, 12.6 pounds crystalline lactose, and 3.4 micrograms riboflavin per gram solids. Calculation showed that each gram of sugar had adsorbed 0.5 microgram riboflavin. The disposition of this wet crude is identical with the disposition of the wet crude obtained in normal practice (4). The residual liquor or sirup (21 pounds) flowing from the centrifuge contained 45.3% solids, 29.5% lactose, and 19 micrograms riboflavin per gram. It was well seeded and stirred at 5-10' C. for 24 hours; as a result, a second lot of lactose crystals containing adsorbed riboflavin crystallized. A 2.7-pound yield of wet crystals waa obtained by centrifuging. These wet crystals analyzed 93% lactose and were found to contain 112 micrograms riboflavin per gram lactose. Thus, starting with clarified whey containing 19.2 pounds lactose and 0.18 gram riboflavin, there were obtained 12.6 pounds of lactose t w crude lactose, 2.5 pounds of lactose as adsorbate, and 0.13 gram of adsorbed riboflavin, representingyieldsof 65% lactose, 13% adsorbate, and 72%riboflavin. PREPARATION OR REFINED ADSORBATE. The wet crude adsorbate was dissolved in water to yield a solution containing 2.5 pounds lactose per 8.25 pounds water; 25 ml. of concentrated hydrochloric acid, 3.5 grams of calcium oxide, and 15 grams of Filter-Cel were added. The mixture waa brought to boiling and filtered. The pH of the cooled filtrate was then adjusted with sodium hydroxide to 5.6. Crystallization a t 5-10' C. was permitted to take place under conditions of vigorous agitation and copious seeding for 24 hours, after which the formed crystals were separated and washed with a small quantity of cold water. Nine tenths pound' of air-dried crystals containing 280 micrograms riboflavin per gram were recovered, representing a yield of 36% lactose and 90% riboflavin. All first-run wash and mother liquors were combined with the second-run feed liquor entering the vacuum pan. However, the mother liquor belonging to the crude adsorbate was recycled only once. LITERATURE CITED (1)

(2) (3) (4)

Leviton, IND.ENQ.C H ~ M35,589-93 ., (1943). Leviton, U. S. Patent 2,116,931 (Mny 10, 1938); Leviton and Leighton, IND.ENQ.CHIPM., 30,1306-11 (1938). Leviton, U. 8. Patent Applioation 390,941 (April 29, 1941). Stringer, Food Indwrtriea, 11, 72-4, 2 6 2 5 , 1 9 0 (1939).