ANALYTICAL CHEMISTRY
1558 (3) Dinsmore, H. L., and Smith, D. C., Ibid., 20, 11 (1948). (4) Field, J. E., Woodford, D. E., and Gehman, S. D., J. Applied Phys., 17,386 (1946). ( 5 ) Fred, M., and Putscher, R. E., ANAL.CHEW,21,900 (1949). (6) Gore and Johnson, P h y s . Rev., 68, 1283 (1945). ( i ) Lipkin, M. R., Joffecker, UT.A., Martin, C. C., and Ledley, R. E., ANAL.CHEM.,20,130 (1948).
(8) Rlair, B. J., J . Research S a t l . Bur. S t a n d ~ d s34, , 436 (1945). (9) Katl. Bur. Standards, Circ. C461 (1947). (10) Rasmussen, R. S.,Brattain, R. R., and Zucco, P. S.,J . Chem. Phys., 15,135 (1947). (11) Thompson, H. W., and Toikington, P., Trans. Faraday SOC.,41, 246 (1945).
RECEIVED for review August 2 , 1951. Accepted July 18. 1952.
Laboratory Evaluation of Granular Carbons W. R. FETZER, E. K. CROSBY, AND C. E. ENGEL Clinton Food Znc., Clinton, Iowa Although bone char is w-idely used as a refining agent, there are few published methods for evaluation of its decolorizing pow-er. Rlost of the methods are long and not too adaptable for routine procedures. None makes allowance for differences in decolorizing pow-er through variation in mesh distribution. In a newmethod for determining the color adsorption power of bone char, the material is first ground to pass 200-mesh and the decolorizing power is obtained in terms of a reference powdered carbon. The mesh distribution of the char is determined. By means of factors for each fraction, the over-all absorptive power is calculated. Other factors are given for reducing these batch process figures to a percolation basis. The method provides a means of evaluating bone char in terms of pow-dered carbon, which is of importance when bothrefiningagents areused together. I t is believed that the procedure can be applied to other granular adsorbents which are commercially available.
T
HE wet milling of corn has increased substantially during
the past 15 years, but this expansion has been achieved with little capital outlay for new buildings. Three plants have replaced bone char refining by powdered vegetable carbon systems. One refiner has gained capacity by using powdered carbon for corn sirup and retaining bone char refining for dextrose. I n other refineries powdered carbon is used as a topping agent, for it is believed that a combination of refining agents produces better liquors. Two plants have retained their bone char plants intact. The use of two refining agents has raised the question of rating one in terms of the other, in order to obtain the relative efficiency for manufacturing and also for accounting practices. This research was undertaken to provide a method for evaluating practices involving the use of powdered carbon and bone char, and also a means of evaluating the neiv synthetic granular carbons. The procedure adopted for establishing this evaluation was as follows : Decolorization efficiencies of granular adsorbents when reduced to fine powders and their relative efficiencies in this form when compared to commercial powdered carbons. Effect of mesh size of a granular adsorbent on the decolorizing efficiency as carried out in a single contact or batch process. Effect of percolation or multiple contact on the decolorizing efficiency. Laboratory evaluation of the decolorizing efficiencies of some process bone chars and comparison of their relative efficiencies in a percolation system to commercial powdered carbons in a batch process. HISTORICAL
I t was logical for the corn sirup and dextrose industry to draw heavily on the technology of the cane refiners in the construction of its plants. Bone char was adopted as the refining agent, and char filters with a capacity of 33,000 or 75,000 pounds were installed with dimensions and flow rates as used in cane sugar refining. However, one difference between the refining operations of the two processes became evident very early. Corn sirup liquors are much more viscous than sucrose liquors, and this results in excessive back-pressure in the char filters. This was
solved by using a coarser bone char, usually mesh sizes of 4 X 8 or 6 X 12. Corn sirup liquors are refined on the acid side, usually in the pH range of 4.8 to 5.5. The acidic liquors leach calcium sulfate from the revivified char. This causes gypsum haze in the finished corn sirup, which in extreme cases gives corn sirup a “milky” appearance. Acid “tempering” of the revivified char largely eliminates this trouble. Acid tempering of the char also removes some of the calcium phosphate. This, with the normal tendency of a bone char to increase in carbon upon revivification, has resulted in process chars with a higher carbon content than those usually found in the cane sugar industry. The literature on the evaluation of the efficiency of bone char is not extensive and most of the published data are concerned with the efficiency of char revivification. The oldest and most used test for the efficiency of char revivification is the caustic test ( 6 ) , wherein char is added to sodium hydroxide solution, boiled, and filtered, and the color is determined on the filtrate. The amount of color is often measured by the Lovibond Tintometer, or a locally prepared standard. Color indicates incomplete removal of impurities in revivification; a colorless or only slightly colored solution indicates well revivified char. Baus ( 2 ) proposed the use of permanent standards of acidified bromothymol blue for the comparison of these tests. Other tests of efficiency of revivification are the pH of water extracts of char suggested by Wayne ( 7 ) and the gas evolution tests, more recently described by Babcock ( 1 ) . Perhaps the most widely used test for the evaluation of the efficiency of bone char is the “batch test.” The test is based on the amount of color and ash adsorbed by the unknown bone char in comprtrison to that adsorbed by a standard or reference char from a reference or standard cane sugar sirup a t apparent equilibrium, and employing manual agitation. When it is necessary to employ a new char or a new sugar sirup for the standard, each is converted to the original standards by means of a factor, which establishes continuity for comparative data. Knorvles (6) published an improved method in which the char contained in a woven wire basket is slowly lowered and raised mechanically in a standard sugar sirup, thereby obtaining greater precision.
1559
V O L U M E 24, NO. 10, O C T O B E R 1 9 5 2
a BDH Lovibond Tintonieter, equipped with a neutral wedge for brightness control, employing the 52 caramel series slides. Standard Sugar Solution. A dextrose refinery processes two types of sugar li uors, first and second sugars. The former is largely the sugar8ydrolyzate obtained from the starch; the latter, the liquor obtained from reconversion of first greens or the mother liquor obtained by the centrifugation of the dextrose obtained from the first sugar liquor, The amount of color in the second sugar liquor is usually four to five times the amount occurring in the first sugar liquors. The standard sugar solution m-as prepared in 5-gallon quantities from first sugar liquor. This amount was obtained from the refinery, adjusted to pH 5.0, refiltered in the laboratory, employing Filter Cel on a Buchner funnel, and maintained under refrigeration. The specifications on the liquor a t this stage were: Baume (commercial) PH Color units. 5-inch cell basis 81
za
-
mi 0 I
20
30
40
50
60
7 0 B O 90
t
Figure 1. Typical Adsorption Isotherm
Since this research was completed, a complete discussion of existing methods for testing bone char and recommended procedures has been published ( 3 ) . BASIS FOR COiMPARISON
The determination of the relative decolorizing efficiencies of powdered carbons is a quick laboratory procedure, for these carbons are fine powders, a t least 97% passing 200 mesh. The determination of the equilibrated color values requires only 20 to 30 minutes. Because decolorization of corn sugar liquors follows the Freundlich equation of
= KC1/n,it is a relatively easy
matter to determine the n eight of carbon to effect any degree of decolorization. The average over-all removal of color from liquors entering the factory refining system is 807, based on powdered carbon per 100 pounds of dry substance sugar solids. ilccordingly, it has been the practice in laboratory evaluation to u5e 80Yc as the criterion and to evaluate all carbons on this basis. A single powdered carbon was chosen as the reference carbon. In order to standardize the test, a large quantity of a single lot was obtained and storpd in a number of containers. Samples of commercial pondered carbons, virgin bone char, synthetic granular carbons, and process bone char from cane sugar refineries and from corn sirup and dextrose industries were obtained. The following data were obtained:
lr0
Apparent weight per cubic foot (all samples) Combustibles as "carbon" (all samples) ilsh (all samples) Relative decolorizing efficiency in terms of reference carbon, all samples of granular carbons having been ground to pass 200-mesh (all samples) Granular char carbons-screen analysis METHODS
Mesh Distribution. The mesh distribution was obtained using 100 grams of the granular material in an End Shak shaker (Newark Wire Cloth Co.) employing National Bureau of Standards sieve series Nos. 8, 14, 20, 30, 40, 50, and 60. Apparent Specific Gravity. The granular or powdered carbons were passed through a glass funnel 4 inches in diameter with 0.5inch orifice, supported on a tripod 9 inches tall into a calibrated metal cup (743-mI. capacity). The distance from the top of the funnel to the top of the container was 145 mm. After the cup was filled to overflowing, the sample was leveled by means of a straightredge. The weight was obtained and the necessary calculations were made to place the data on a cubic foot basis. Volatiles or Carbon. One to 2 grams of sample contained in a porcelain crucible were ashed in a muffle a t 500" to 550' C. for 24 hours. The loss in weight is termed "volatiles" or carbon, Moisture. Ten grams of sample m r e dried in a forced-draft air oven a t 115" C. for 1 hour. Color Measurements. 811 color measurements were made in
28.0 to 3 0 . 0 ' 4 . 8 to 5 . 0 600 t o 700
The color was obtained on a 0.25-inch cell and the reading obtained was multiplied by 20 to place the readings on a 5-inch basis. The standard sugar liquor was darkened to the established color standard when necessary by the addition of a small amount of second sugar liquor. DECOLORIZATION T E S T S
Preparation of Carbon Samples. Powdered carbons were evaluated as received. Granular adsorbents and bone chars xere acid "tempered" in a manner simulating plant practice. The adsorbent mas leached with hydrochloric acid for 0.5 hour-27 grams of 20" Baume acid per gallon. The sample was then placed on a Buchner funnel and washed continuously iTith distilled n-ater until the effluent had a pH range of 5.0 to 5.5. The washed adsorbent was dried overnight a t 115" C. A quantity of each granular adsorbent was ground in a Coors mortar and pestle to pass 200 mesh. Apparatus. Constant temperature n-ater bath, 71" =t0.5' C. Quart AIason jars n-ith small holes through t,he lid to provide for the agitator. Definition of Analytical Terms. 807, DECOLORIZATI~S. 4 dextrose refinery usually operates a t an over-all decoloration of 80y0 from neutralized starch hydrolyzate to the final liquors entering the crystallizers. For this reason, it has been the custom to evaluate all adsorbents on this basis in the laboratory. This custom may ignore an adsorption isotherm, but the procedure has been firmly established as the beet way to evaluate an adsorbent for plant performance. EQUIVBLESCE.An adsorbent is said to be equivalent to the reference carbon, when a determined amount of the adsorbent removes as much color from 100 pounds of dry substance sugar as does 1 pound of reference carbon. REFEREXCE CARBOX(RC). A poxvdered activated carbon. VIRGINBONECHAR(VC). A reference bone char taken as a standard of granular carbon performance. granular carbon whose equivaCARBOXDER TEST(CT). lence factor is to be determined. EQUIVALEXCE FACTOR (F). The number of pounds of carbon under test which when used in a granular form in a percolation process is equivalent to reference carbon, REFERESCECARBOA\'.ALEE(RCV or RCT). The number of pounds of virgin bone char or carbon under test which, when pulverized to pass 200 mesh, is equivalent in a single batch contact of 2-hour duration to 1 pound of reference carbon in a single batch contact of 1-hour duration. REFEREXCE CARBOX VALVEGRASULAR(RCGT). The number of pounds of carbon under test which in original granular form is equivalent in a single batch contact time of 10-hour duration to 1 pound of reference carbon in a single batch contact of I-hour duration. EQUIVALENCE FACTOR V (FV). The number of pounds of 20 X 30 mesh virgin bone char which, in a percolation test, is equivalent to 1 pound of reference carbon in a batch test. PERCOLATION FACTOR (PF). Ratio of equivalence factor V to reference carbon value granular for 20/30 mesh (F4). (SF) s (F,, Ff, FI-F,). The number of pounds of SIZEF a c ~ o ~ the various sieve fractions of virgin bone char which, in a single batch contact of 10-hour duration are equivalent to reference carbon in a single batch contact of 1-hour duration. VIRGINBONE CHAR EQUIVALEST OF CARBONUNDER TEST (VCE). The number of pounds of carbon under test which
1560
ANALYTICAL CHEMISTRY
would be equivalent to equivalence factor if carbon under test possessed the same decolorizing capacity as virgin bone char but retained its own particle size distribution. Evaluation of Adsorbents as Fine Powders (RCV or RCT). Quantities of standard sugar liquor are weighed out, each containing 100 grams of dry substance sugar solids (approximately I46 ml,). Dry substance is determined according to the Filter Cel procedure of Cleland and Fetzer (4). Several different wights of the unknown ground to pass 200 mesh and 1.00 gram of reference carhon are neighed out for use in the above 1iq;or.
L-x 7%WIRE G A U Z E
Figure 2. Decolorizing .4pparatirs Used for Bone Char Fractions
The standard sugar liquor, in a beaker and under agitation, is heated rapidly by means of a Fishel burnei to i 1 " C . and transferred to jars in the water bath. The decolorizing agents are then added. Although tests have s h o m that the color equilibrium is reached within 30 minutes, the test v a s continued for 1 hour for powdered carbons and 2 hours f o ~the ground granular materials (200 mesh). The progress of the decoloiization can be followed ~ J Yremoving 25-ml. portions, filtering, and obtaining the color removed. The liquor is filtered through an asbestos mat on a small Buchner funnel. After cooling, the pH is o!>tained and, if necessary, adjusted to 4 8 to 5.0. The colors are read on the Tintometer. The data fioni the reference carbon are used, if neceesarv, to adjuEt the unknon 11 samples to an 80% decolorization basis.
PERCOLATION T E S T S
Apparatus. The laboratory percolation apparatus which \\a; used is shown in Figure 3 and is self-explanatory. The dimer!sions and ratio of diameter t o height are essentially those of the factoryfilters. A bed depth of 10.8 em. represented a single filter: 21.7 cm. represented t x o filters in tandem. Bone Char. Virgin bone char which had been acid tempered as described above with a mesh size of 20/30 (F,) was used in all tests, as it was believed that. this size would fill the filter with le?. possibility of pocketing and provide a more uniform flox through the char. The weight of char was 60 grams to fill a single filter and 120 grams for tw-o-filter operation. Determination of Equivalence Factor for 20/30 Virgin Bone Char. Distilled water (100 ml.) was run into place for the single filter test and heated to 71" C. by means of xater circulated through the jacket. Sixty grams of char were fed into the filter in small increments t o minimize the formation of air pockets. The standard sugar liquor was fed into the filter a t a rate which was estimated to produce 80% decolorization in the composite filtrate in 10 hours, the time used in the batch experiments to produce apparent equilibrium. The flow rate for a single filter was approximately 126 ml. per hour (275 ml. for two in tandem). The first 100 ml. (200 ml. for two in tandem) was not used iri making the composite sample, as it was largely water. However. the amount of dry substance cont,ained was determined and used in the final calculations. Samples were first collected in 100-nil. portions, the color was determined, and the sam le reserved for the composite. Then the samples were taken fiourly and the color was determined on these hourly samples and on the accuniulated composite. These data vere sufficient to estimate the tinit. and to make minor adjustments in flox rate to complete the run in 10 hour9 and obtain the 80% decolorization. DECOLORIZATION O F GRANULAR ADSORBENTS AS FINE POWDERS
T h r decolorization efficiencies of all adsorbents were determine 1 as fine powders by the experimental procedures n o t e i above, anti the results are given in Table I. The decolorizing power of the commercial pondered carbons is largely the same, even though there are large differences in the apparent specific gravity and the carbon contents. There are a190 notshle differences in the
-
RESERV3IR
PINCH CLAMP
The method used in establishing the quantity of unknonn material (CT) t o an 80% basis is shoan in Figure l.
*'
MESH FR4CTION TESTS
Preparation of Mesh Fractions of Bone Char (SF,, SF?, SF1 SF,,).Virgin bone char with a mesh size oi 4 X 12 was used. A large quantity was acid ',tempered" and washed as described above. This sample n a s then crushed and,'or ground to yield quantities of the following mesh sizes: Through
VEN
77
TO WATER EATH
f------
-.---
T
DISCHARGE TO 250 ML CYLINDER
35 CM.
4 o n 8 (Fi' 8 o n 14 (F2) 14 on 20 (FlI
2 6 CM
20 o n 30 (F4 30 o n 40 (Fa)
40 on 50 (F6J 50 on 60 ( F i j (RCV)
15 C M
200
Determination of Reference Carbon Value of Sieve Fractions of Granular Adsorbents (RCGT). The decolorizing procedure was essentially as described above. K i t h the smaller mesh sizes where the amounts of material were small, the stirrer vaq placed above the settled char and run to avoid mechanical disintegration of the char. Hollever, as the mesh size increased, and the amount of char increased to effect 80% decolorization, it was necessary to alter the bath procedure. A small mesh basket 2 inches in diameter n-as inserted in the jar and the char n a s placed outside the basket. The top of the char surface was covered with glass wool to contain the char within its enclosure, and the stirring was effected in the well formed by the basket (Figure 2). Tests were arbitrarilj- run for 10 hours to attain apparent equilibrium.
SINGLE
i
-_-FROM WATER BATH 160. F.
Figure 3.
CM. PUMP
7.:.2
I
T
LABORATORY BONE FILTER
Decolorizing -4pparatus Used in Percolation Tests
V O L U M E 2 4 , NO. 10, O C T O B E R 1 9 5 2
1561
amounts of soluble ash and iron contents, although no data are given for these items. There are very wide differences in the process chars from the corn sirup and dextrose refineries. The data would seem to indicate that the higher the carbon content the more efficient the char. The cane sugar process chars, with their lower carbon contents, appear to decolorize no better than the dextrose process chars when tested in reference dextrose liquors a t a pH of 5 . However, the cane sugar process chars havc been used previously in neutral slightly alkaline sucrose liquors and the procedure of “acid tempering” and subsequent use in acid dextrose liquors is a marked departure from the normal practice. A complete comparison would have required the use of both dextrose and sucrosc chars, as fine powders, in a reference sucrose liquor as employed in the batch test, which was deemed outside the scope of this particular paper. However, corn sirup and devtrose char h was routinely controlled for years a t 8OP0 of the decolorizing power of virgin char, based on tests ohtained from a batch test using a reference sucrose sirup.
Data on the loss in decolorizing efficiency with increased mesh size were obtained experimentally by using mesh fractions of virgin bone char in a single contact or batch process described under Methods. Although the technique employed can be criticized from the standpoint that the fractions, as prepared, do not represent the bone fractions which make up bone char, the authors believe that the data obtained are sufficiently valuable from the practical standpoint that the criticism may be overlooked. 50.5
EFFECT O F MESH SIZES ON DECOLORIZIh-G EFFICIENCY
The above data evaluate all adsorbents on a common basisi.e., fine powders, passing 200 mesh. Under these conditions, the equilibrated adsorption values are readily obtained and a table of relative efficiencies in terms of a standard carbon has been presented. However, these data do not answer the question of the loss in efficiency as the mesh size is increased, or, in more practical terms, the relative efficiency of a mesh fraction of bone char. For as the particle size is increased, physical considerations of structure would indicate that it would take longer to reach a n apparent equilibrium and that this apparent equilibrium would require more of the granular adsorbrnt to attain the same degree of decolorization in the same time interval. These considerations have long been known to the practical refinery man who seeks t o employ as much of the finer mesh fractions as is consistent with good flow rates and small pressure heads in the filteis. Some idea of their solution of this problem can be gained from the mesh distribution of process bone chars given in Table I1 and shown graphically in Figures 4 and 5. The mesh distribution of the corn sirup and dextrose bone chars A, B, and C, which are used on corn sirup, indicates the necessity for a coarser char to handle the more viscous corn sirup liquors. The D char is used exclusively for dextrose liquors and the mesh distril)ution is similar to the process chars usFd for cane sugar liquors. Table I.
Decolorizing Efficiencies of Adsorbents
Adsorbent Activated carbons
h
B
c
D Tirgin bone char
.I B
c Corn sirup a n d dextrose process bone char A B