October, 1932
I K D U S T R I A I, A N D E N G I N E E R I N G C H E M I S T R Y
material was first cooked and pressed, as 1s done in the conimercial process of obtaining the oil. A certain amount of soluble protein is lost in this process. It may be presumed that this lost protein is of high digestibility and perhaps also of high biological value. Further, the cooking and pressing may possibly tend to lover the value of the remainder of the protein. To what extent these factors rnay be responsible for the differences noted between the two kinds of meal dried in the same way cannot be stated. In view of the differences in raw material and treatment prior t o drying, as set forth above, these data cannot be considered to furnish evidence that haddock proteins are superior to inenhaden proteins per se. However, since the cooking and pressing process, with the resultant loss of s o h ble protein, is generally employed in the menhaden industry a t present, the data have a practical significanc%ewith respect to the products used in livestock feeding. From the standpoint of the menhaden industry they serve to strengthen the contention of Harrison ( 2 ) that procedures should be adopted to lessen or eliminate the losses of water-qoluble material. Daniel and McCollum ( 1 ) have reported from growth studies t h a t vacuum-dried white fish meal appears to contain a protein of similar quality to vacuum-dried menhaden meal. The white meal was a commercial product consisting of cod and haddock, whereas the menhaden was an experimental product prepared by the Bureau of Fisheries. Since these products differed both in ram material and in the methods of preparation from those used in this laboratory, the results from the two experiments cannot be considered as contradictory. However, they serve to emphasize the ronclusion that general statements cannot he made
1171
regarding the relative value of white and menhaden iiieala, since this value is dependent both upon the nature of the raw material and the methods of processing. The results here presented suggest that the superior nutritive value of the protein of vacuum-dried haddock meal in comparison with flame-dried menhaden meal, as previously found in this laboratory (4, ?), is due both to the method of drying and to the nature of the material dried. They confirm the findings of the nitrogen-partition studies of Invaldsen (d) that high temperatures have a detrimental effect upon the quality of fish proteins. A deleterious effect of heat upon cereal proteins has been shown by Morgan ( 6 ) . AI1 of these results point to the practical importance of further study of the influence of the temperatures used in various manufacturing and home processes involved in the preparation of protein foods, both human and animal.
LITERATURE CITED (1) Daniel, E. P., and McCollum, E. V., U. S. Bur. Fisheries. Iitwestiaational Reut. 2 (1931).
(2) Harrison, R. W.,*Ihid.,' 1 (1931). (3) Invaldsen, T., Can. Chem. J f e t . , 13, 97 (1929). (4) Maynard, L. A , Bender, R. C., and McCay, C. LI., J . Agi'. Research, 44, 591 (1932). ( 5 ) Mitchell, H. H., J . Bid. Chene., 58, 673 (1924). (6) Morgan, A. F., Ihid., 90, 771 (1931). (7) Schneider, R. H., J . d g r . Research, 44, 723 (1932). RECEIVEDMay 31, 1932. Presented before the Division of Biological Chemistry a t the 83rd Meeting of t h e ilmerican Chemical Society, New Orleans, La., March 28 to April 1, 1932. This investigation wae supported in part b y a felloashipbgrant from The Birdseye Laboratories, Glourester, Mass.
Viscosity of Corn Sirup W. B. BISHOPAND NEIL YOUNG,A. E. Staley Mfg. Co., Decatur, Ill. corn sirup will run from 41 to L'sing a falling-sphere z'isconieter f o r which 43 Per cent Purity. and Pipe lines, study of details of calibration are giz'en, the ?>iscosity evaporation data, and adapof corn sirup as sold (comrrzercial glucose) has tation of a product to particular DESCRIPTION O F APF'ARATUS been determined for rurious densities and temuses, a kllowledge of viscosity is a f a c t o r of prime importance. peratures. It has been shown that within the The falling-sphere.viscometer range incestigated corn sirup acts as a truly has been a p p l i e d successfully There have been few data published on the viscosity of coinl,iscoussolufion. The of per reducing by Ladenburg (41,Sheppard ( 6 ) , mercial corn sirup. Kashburii G i b s o n and J a c o b s (S), and sugars on viscosity has also been siudied. Bennett and Sees (1) for deand Shelton (8)give data on a termination of the viscosity of glucose sirup and mixtures of this with pure dextrose, but these are incomplete, for one teni- turpentine, nitrocellul~*c, sugar solutions, and molasses. Because of the transparency of corn sirup, this method seemed perature only, and of limited value. Commercial corn sirup as sold in bulk is manufactured t o well adapted. It requires simple equipment, and because of strict standards of moisture content and degree of eonversion, this will give reasonably accurate results without a great and the products of various companies are surprisingly uni- deal of special equipment or technic. The spheres used in this investigation \$ere ordinary steel form and comparable. The moisture content of the sirup to 3/3 inch according is specified by giving the Baume a t 100" F. This is measured ball bearings varying in diameter from a t 140' F. by means of a hydrometer (145 modulus), and cor- to the viscosity of the sirup, The tubes were ordinary gradurected to 100" F. by adding the arbitrary, although approxi- ates, and 250-cc., 500-cc., and one-liter graduates were calimately correct, figure of 1 Bauni6. Thus a sirup which has brated. The larger sizes were used to reduce the end and a density of 42" measured a t 140' F. is said to be 43' corn side corrections necessary when the equations are used instead sirup, Corn sirup, or "glucose," is marketed, unmixed, in of a standard liquid for calibration. The tubes were fitted concentrations of 42". 43", 44", and 45" Baum6, the 43" a t the top with metal caps or rubber stoppers holding a Baume predominating. The degree of conversion is specified starting tube of about 1.2 cni. inside diameter, which dipped as per cent purity, or per cent of the dry substance present below the surface of the sirup for several centimeters. .I which reacts as a reducing sugar to Fehling's solution, calcu- sufficient distance below the starting tube t o allow attainment lated as dextrose Except for special products, commercial of equilibrium velocity, the starting mark was located, and
I
K THE design of pumps
O
INDUSTRIAL I N D ENGINEERISG CHEMISTRY
1172
6 to 8 cm. from the bottom, the final mark. These were extended around the tube to eliminate parallax. The tubes were placed in a glass-walled constant-temperature bath which could be adjusted to bring them exactly vertical. For removing the balls after completion of the run without disturb-
VOI. ",
No. 10
For convenience in measuring viscosities over a wide range, and also to study the applicability of the equations, seven different sizes of balls were calibrated in the three different sized tubes. Ten balls of each size were weighed and found to be quite uniform, the largest deviation from the mean weight varying from 0.06 per cent in the case of the '/,-inch balls to 0.5 per cent in the case of the 3/]&- and l/&ch balls. An average density of 7.77 was found for the steel balls. The constants of the tubes were as follonw: VISCOMETER
Capacity: R, cm. L,cm. h. cm.
1000 cc. 3.05 20.2 38
.I
vI8COMETER B 500 cc. 2.40 18.9 33
Vl?rOMETER
c
260 c c .
1.78 20.5 3.5
Data on calculated values for K and those determined experimentally are given in Table I. The experimental data represent the average of several of each size sphere a t temperatures of 5O, lj', and 25' C. for the Stanolind oil. TABLEI. CALIBRATIOX OF VISCOMETERS NOMIKAL VISCOMETERA VISCOMETER B YISCOMETERC BALL Av. K K K K K K SIZE WT. RADIES calcd. exptl. calcd. exptl. calcd. exptl. Inch Grams Cm. 1s', 0,1273 0.1576 0 . 2 3 4 0.237 0.243 0,245 0,214 0,218 5/82 0.2533 0 . 1 9 8 2 0.358 0.368 0.370 0,372 0.324 0,330 8/16 0.4402 0 . 2 3 8 3 0 . 5 0 3 0 . 5 1 1 0 . 5 1 5 0 . 5 1 1 0.447 0,467 7/32 0 . 6 9 9 9 0.2781 0.665 0.675 0.678 0.678 0.572 0,622 1/1 1.0429 0.3177 0.843 0.855 0.855 0.816 0.730 0.764 6/1a 2.0396 0 . 3 9 7 2 1.246 1.266 1.250 1.220 1 . 0 5 2 1.042 8 1 s 3.5242 0.4767 1.702 1 . 7 0 8 1.692 1 . 4 8 5 1 . 4 0 2 1.240
.
FIGURE1. VISCOSITY-TEMPERATURE CURVESFOR CORN SIRUP ing the liquid or introducing air bubbles, a small electromagnet was used. CALIBRATIOX O F VISCOhlETER Ladenburg (4) gives a modification of the urell-known Stokes equation with corrections applying where the liquid is not infinite in extent, as in the relatively small tubes used here. The equation is:
( + 2.4 ;)(1 + 3.3 i )
9L 1
where 7
viscosity in poises
=
According to Ladenburg, Equation 1 holds when r l R is less than 0.08, whereas Sheppard showed the variation was not great for T ' R = 0.1. An examination of Table I shows that the calculated values for K are from 1 to 2 per cent in error for values of r/R below 0.12, somewhat erratic and from 4 to 8 per cent off for r / R between 0.12 and 0.16, and as high as 13 per cent for r / R between 0.2 and 0.25. Below 0.12 the variation is less than the experimental error. Previous experimenters (1, 7) have also confirmed the equation. Equation 2 with K experimentally determined should hold for any practical dimensions of spheres and tubes. Where calibrating liquids are available it is more convenient to use the smaller tubes, since less time is required for attainment of temperature equilibrium.
g = acceleration of gravity (980 cm. per s e t 2 )
r = radius of sphere in cm. S = .density of sphere in grams per cc. u = density of liquid in grams per cc. t = time of fall in seconds L = distance of fall in cm. R = radius of tube in cm. h = total height of liquid in cm.
DETERMISATIOS OE'1-ISCOSITY OF CORNSIRUP
As pointed out by Bennett and Kees ( I ) , this equation can be modified for a given viscometer and sphere to the form q =
K(S -
(2)
u)t
where K is the tube constant. K can be calculated from Equation 1 or can be determined experimentally for a given tube according to Equation 2 if a liquid of known viscosity is available. There is a scarcity of calibrating liquids for viscosities of 50 poises and over. The only oil sufficiently viscous and still transparent enough which could be secured was Stanolind No. 200. T-iscosity of this oil, as determined by the Bureau of Standards, together with its density is as follows: TEMPERATURE O
c.
5 15 25
VISCOSITY Poises 85.9 32.1 14.1
DENSITY Gram/cc. 0.881 0.872 0.866
The sirup, of known Baume and purity, was placed in the viscometer tube, covered, and allowed to stand in the bath a t 140' F. until free from air bubbles. The temperature of the bath was controlled to within 0.2' F. Following airbubble removal, the sirup was brought to the desired temperature (this requires several hours in the larger tubes), the starting tube put in place, and the various size balls dropped and timed with a stop watch. Spheres were selected giving times of fall of from 10 to 100 seconds. Figure 1 shows the data on commercial grades of corn sirup with the logarithm of viscosity plotted against temperature for the different Baume sirups. Those tested ran from 41.2 to 42.8 per cent purity. Each point represents the average of from 10 to 15 spheres dropped. From its composition it might be suspected that corn sirup is a plastic material. Disagreement between the viscosities as measured by small and large spheres should indicate this ( 6 ) . However, it was thought advisable to make a few check runs on a transpiration type viscometer to determine whether or not corn sirup was truly viscous and, at the same time, t o serye as a check on the falling-sphere viscometer. For these runs a consistometer similar to that described by Riilkley and
October, 1932
IXDUSTRIAL
AND ENGINEERING
CHEMISTRY
1173
Bitner ( 2 ) was used. The consistometer was calibrated both by actual measurement and by use of the Stanolind No. 200 oil. The curves secured are not reproduced here since they were all straight lines passing through zero, thus indicating that a t least within the range of accuracy and concentrations studied corn sirup was truly viscous. Table I1 shows comparative results from the two methods. The close checks secured on the two widely varying types of instrument also indicate viscous rather than plastic flow. TABLE 11. COMPARATIVE RESULTSWITH Two TYPES OF INSTRUMENT R A U M h AT
100° F.
TEMPERATURE SP. GR.
F.
100/60aF.
43 0
100 140
1.4216 1.4078
43.3
100 140
1.4258 1.4119
-VISCOSITYFaljing-sphere Tranepiration viscometer viscometer Poisea Poiaea 152 148 19.3 21.6
211 23.4
206 22.6
Table I11 shows the viscosities of corn sirups obtained from different manufacturers. The agreement is within about 5 per cent, and part of this variation is due t o difference in the purity of the sirups. TABLE111. VISCOSITIES OF CORX SIRUPSFROM DIFFERENT MILYUFACTURERS MANU- BAUMB AT looo F.
FACTURER
SP. GR. 100/60° F .
PURITT % drg substance
VISCOSITY
looo F.
140'
Poises
Poises
F.
I
0
I
20
DETERMINATIONS The accuracy of the calibration of the viscometer is probably well within 3 per cent. Viscosity of Stanolind varies about 9 per cent per O C., and temperature control was about 0.1" C. The variation of 0.5 per cent in the weight of balls would cause only about 0.3 per cent error in the time of fall, and hence no correction was attempted. Some of the times recorded were as low as 10 seconds, hence were measured only to 2 per cent.
I
I 80
1
/oo
A survey of the data on corn sirup shows the following percentage variation in viscosity for unit variation in other variables: l o F.
ACCURACY O F
I
FIGURE 2. EFFECT OF PER CENT REDU~ING SUQARSON VISCOSITYOF CORNSIRUP
0.1' BB, 1% PurltY
To determine the effect of purity on viscosity, tests were made on several samples of different purity which were available. These included a special low-purity sirup of 29.8 per cent purity, special high-purity sirups of about 55 per cent, unrefined corn sugar liquor of 88 per cent, and a sirup made from crystallized dextrose of 100 per cent purity. Where Baume of these sirups differed slightly from 43", the viscosities as determined were corrected to 43' by using percentages obtained from Figure 1. The resulting curve is shown in Figure 2.
I
40 60 PURITY I N P E R C E N T
a.5 to 7 8 to 10 6 to 7
Since temperature was controlled only to 0.2"F. and gravity measured t o only 0.05' BB., whereas the samples varied about 1 per cent in purity, it would seem that the method of determining viscosity was closer than the control of other variables. The latter might have caused a variation of from 5 t.0 10 per cent for any individual reading. Results were checked within 4 per cent between the transpiration and falling-sphere viscometers with one exception.
LITERATURE CITED Bennett, A. N.,and Nees, A. R., IND.ELG.CHEM., 22,91 (1930). (2) Bulkley, R., and Bitner, F. G., J. Rheol., 1, 269 (1930). (3) Gibson, W. H., and Jacobs, L. M., J. Chem. SOC.,117,473 (1920). (4) Ladenburg, R., Ann. P h y s i k , 23, 9 (1907). (5) Phipps, H. E., Colloid S y m p o s i u m Monograph, 5 , 259 (1927). (6) Sheppard, S. E., J. IND.ENO.CHEM.,9, 523 (1917). (7) Speicher, J. K., and Pfeiffer, G. H., Colloid S y m p o s i u m Mono(1)
graph, 5, 267 (1927). E. W., and Shelton, G . R., Univ. Ill. Eng. Expt. Sta., Bull. 140, 24 (1924); International Critical Tables, 5 , 23 (1929).
(8) Washburn,
R E C ~ V EMDa y 12, 1932:
Arrangements have been perfected for giving INDUSTRIAL D. D. Berolzheimer, 50 East 41 St., Sew York, N. Y. (to whom ENGINEERIXG CHEMISTRY the exclusive right .of first publi- all inquiries should be addressed), who originated the Percolator cation .of the series of reproductions of paintings, etchings, Series of Prints, and who has much material for further publicaengravings, and prints of alchemical and historical interest, tion by us. formerly known as the "Percolator Series of Prints."' To meet a demand, photographic negatives have been made of Twenty such illustrations have appeared in our numbers, each illustration, usually from the original painting. Black and starting with that of August, 1931. A list of the first seventeen white photographic prints, glossy or matt, 8 by 10 inches in size, pictures n-ill be found in our March, 1932, issue (page 317), can be supplied at $1.50 each, and enlargements, 16 by 20 inches the three additional ones having appeared in April, May, and in size (matt only), at $4.00 each. Subscriptions will be accepted for the entire series at $1.25 per June, 1932. Each picture will be accompanied by a short note concerning print, t o begin with No. 1 of the series or with the current print. Remittance should accompany all orders. the original, the artist, etc. Permission t o reproduce these illustrations in other publication!: This reproduction is made possible through the cooperation of should be obtained from Mr. Berolzheimer, who will be glad to 1 See IND. ENQ. CHEM.,23, 966 (1931); News Edition, M a y 10, 1931, supply electrotypes of the half-tone cuts a t cost. p. 155. No. 21 of the series appears in this issue, page 12011 AND
