878
T H E J O U R N A L OF 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 .
paratus which consists of a cooling coil with water spray and a small tubular cooler, which cools the liquor coming from the treatment ring of the ammonia still, thus cooling i t for use in the first scrubber. The impurities given off by this liquor in the ammonia still saturate the concentrated ammonia in the condenser of the ammonia still, while the excess impurities pass off into the air after being" washed of all traces of ammonia in the catch-pot. The effects of washing the crude gas with this liquor are shown by the preceding data. All of the cyanogen in the crude gas is eliminated by the liquor wash. On a n average, we found the liquor wash absorbed Prom 0.3 t o 0.5 per cent. of carbon dioxide from the gas. My opinion is this absorption would be greatly increased with a liquor wash containing a higher percentage of ammonia and with a longer time of contact. The temperature of the liquor wash t o the first scrubber must not be higher than 30' C., neither should the temperature of the gas to the scrubber be higher than 30' C. The best results are obtained when the two temperatures are held a t about 2 5 O C. This is the temperature of the gas to the scrubbers that y e meet with in ordinary coal gas practice. If gas liquor is substituted for the liquor wash there is practically no increase in hydrogen sulphide in the gas liquor or decrease of hydrogen sulphide in the gas. Since our scrubbing apparatus was limited to one fresh-water scrubber consisting of eight bell-shaped compartments, I divided the scrubber using the four lower compartments for the liquor wash and the four top ones for fresh water. The results would have been much more satisfactory if we had one scrubber for the liquor wash and one for the fresh water, then we could have used much stronger liquor and would have obtained much higher elimination of impurities per gallon of liquor wash used, without fear of losing any ammonia from insufficient washing. Another very disagreeable feature of our operation was the uneven make of gas, the amount of gas washed for one hour differing as much as 50 per cent. from another hour of the same day, with little regulation of the liquor wash t o meet the increase or decrease in the gas. Considering all the disadvantages mentioned above, still nowhere in the process was there the least possibil&y of ammonia loss, also the average yield of ammonia for the six months exceeded the average yield for the six months previous t o the installation of the process. No difficulty was experienced in the operation of the ammonia still. The total amount of steam required t o operate the ammonia still was not appreciably increased because the cooling water for the ammonia vapors was practically eliminated. During the month previous to installing the process, 2 7 per cent. of the coal used €or gas-making was a high sulphur coal. The purifying capacity consisted of three boxes, holding a total of 6000 bushels of shavings and iron oxide; during this month we found it impossible t o change the oxide in the boxes rapidly
Dec.,
1912
enough t o purify the gas from hydrogen sulphide, even using new oxide and introducing some air into the gas before the purifiers. For the six months after installing the liquor purification process, the oxide in the boxes was not changed, no labor was expended on purification, the average monthly output of gas for the first three months increased 1 2 per cent. and for the last three months decreased 2 0 per cent. (partly due to coal shortage), the percentage of high sulphur coal used was increased t o 50 per cent. of the total, and a t the end of this period we found all boxes clean. A very small quantity of air was drawn into the gas before the exhausters, approximately 0.5 per cent.; the quantity of nitro-' gen in this air introduced into the gas was more than offset by the quantity of carbon dioxide absorbed from the gas by the liquor wash. The gas issuing from the first purifying box in the series would a t times show a trace of hydrogen sulphide, but on changing the flow of the gas, so as t o remove this box from being first t o receive the gas, it would immediately clean up. The gas has gone through the boxes for a month a t a time without the gas from the first box showing a trace of hydrogen sulphide with lead acetate paper. During the severe winter weather, the wash liquor in the pipe line t o the scrubber became frozen, consequently for several hours the liquor wash was discontinued. The three boxes went foul even though a short time before they were all perfectly clean. When the wash liquor was again turned on the three boxes gradually cleaned up, proving t h a t b y using this liquor wash with a very small quantity of air, the foul boxes could again be made clean, without the necessity of first passing the gas through a clean box. The actual saving obtained by using this process is very large, even for a small plant, both in iabor cost and using high sulphur coal; the process is simpler than oxide purification, elastic, independent of one grade of coal, without danger of fire or explosions, does away with any installation of new purifying boxes, and saves valuable space. The gas business being a world-wide industry, consequently any improvement in the manufacture of gas is of world-wide interest. With the modern development of uses for gas, any improvement in its manufacture is a direct aid t o the progress of the world. GENEVA,N. Y.
STUDIES ON FISH OILS.' 111. Properties of Fish and Vegetable Oil Mixtures. By GEORGEF. WHITEAND ADRIANTHOMAS.
Received Aug. 14, 1912.
I n consequence of the continually increasing interest in the utilization of various fish oils for the most diverse commercial operations, their substitution where possible for the more valuable vegetable oils, and the lack of intimate and complete knowledge of their physical and chemical properties, i t has seemed desirable t o continue and extend the work which was carried out recently by one of us on fish oils of known origin. The uncertainty of the various tests t h a t have 1
For previous communications bearing on this subject see THIS 4, 106 and 267.
JOURNAL,
'
Dec., 1912
THE J O U R N A L OF I N D U S T R I A L A N D ENG1:VEERIA'G C H E M I S T R Y .
879
been applied to any,oil to detect possible adulteration by another of inferior value has been due rather t o the want of a full conception of the behavior of fish oils under all circumstances than to any inherent difficulty connected with the problem. A careful study of mixtures of fish oils with each other and with vegetable oils should yield results of value t o both analyst and manufacturer. I n this article there is presented the results of the analysis of dogfish liver (Mustelus canis), soya bean, linseed, china wood oils, and mixtures of these latter with the fish oil. The fish oil was extracted from livers by steam pressure by the method previously described by one of us.1 Mixtures were made up by weight. The viscosity, density, index of refraction, saponification number, acid number, and iodine number were then measured according to the description given below.
a t 30' to note a n y possible change due to chemical decomposition.
VISCOSITY -4XD F L U I D I T Y .
TABLEIII.-SOYA B E A NOIL.
I t has been shown t h a t the fluidities of fish oil mixtures are additive, and i t should be expected t h a t the same condition would obtain x i t h all oils except in those cases where there is association, decomposition, or some other exceptionally disturbing factor. As the viscosity of a n oil is of particular interest t o the paint manufacturer, and fish oils are largely used for outside paints or to adulterate oils for other purposes, this property has therefore been especially considered in our study of these oils. A viscosimeter capable of very accurate measurement gave satisfactory service in our previous work and a new one was made and calibrated according to the method already given. The constants obtained for this instrument are: 1 . 6 cm. Length of capillary (approx.). , . , . . . . . . . . . . . . . 'Volume of left limb . . . . . . . . . . . . . . . . . . . , . . . . . 2 . 1 3 1 0 cc. Volume of right l i m b . . . . . , . . . . Pressure correction (30°), left li Pressure correction (30°), right R a t i o of r4 t o 1 . . . . . . . . . . . . . .
.. .
. . . .. . . .....
- 0 . 3 0 cm. 0.00005835 0.00005822 0.00005837
The formula used for the calculation of the viscosity was
TABLEI.-DOGPISH OIL. Viscosity. Temp.
Right limb.
30' 50' 70 90 30'
0.4143 0.2113 0.1247 0.08046 0.4142
1 The Soya Bean and China wood oils were furnished u s b y hlarden,
Orth, and Hastings. of Boston & Amend, Sex%-York
The linseed 011 was obtained Crom Elmer
Average.
0.4140 0.2110 0.1247 0.08063 0.4142
0.4142 0.2112 0.1247 0.08055 0.4142
.
Fluidity. 2.414 4.735 8.020 12.41 2.414
TABLEII.-LIKSEED OIL. Viscosity. Temp.
Right limb.
Leit limb.
30' 500 70' 90' 30
0.3312 0.1757 0.1071 0.07103 0.3312
0.3311 0.1757 0.1069 0 . 0 7 1 19 0.3316
Average. 0.3312 0.1757 0.1070 0.07111 0.3314
Fluidity. 3.019 5.692 9 346 14.06 3.017
Viscosity. Temp.
Right limb.
Leit limb.
30' 500
0.4067 0.2060 0,1207 0.07807 0.4069
0.4058 0.2065 0.1209 0,07836 0.4067
70' 90 30'
Average. 0 ,4063 0.2063 0.1208 0.07822 0.4068
Fluidity. 2.461 4,847 8.278 12.79 2.458
TABLEIV.-CHINA WOODOIL. Viscosity. Temp. 30' 350 500
70' 90' 350
Right limb. , . .
1.591 0 . 7796 0.3699 0 2126 1.933
Left limb.
-4verage.
Fluidity.
...
1.859 1.589
0.5379 0.6293 1.282 2.697 4.702 0.5139
1.587 0.7805 0.3717 0.2128 1.957
0.7so1 0.3708 0.2127 1.946
TABLEV.-DOGFISH LIVER-LIXSEED OIL MIXTURES. 25 P e r cent. Linseed. Viscosity.
Temp.
Right limb.
30' 50' 70 a 90 '
0.3954 0.2027 0.1204 0.07784
30' 70' 900
0.3712 0.1930 0.1150 0.07554
30' 500 70' 90'
0.3555 0.1866 0.1123 0.07425
Left limb. 0.3943 0.2018 0.1201 0.07798
Average. 0.3949 0.2023 0.1203 0.07791
Fluidity. 2.533 4.943 8.313 12.84
5 0 P e r cent. Linseed.
500
where q is the viscosity, r the radius, and I the length of the capillary, v the volume of liquid passing through the capillary, d the density of the liquid, p the pressure, and t the time of flow. The method of calibration and manipulation has been fully described in previous papers. I n the following tables are presented the results of the viscosity measurements, which were taken a t 30°, s o o , 7 0 ° and 9 0 ° , for the right and left limbs, the average of the two, and the reciprocal of this latter, which is the fluidity. The china wood oil was measured a t 3 5 O and the values a t 30' extrapolated from the curve, as the oil was so viscous a t the lower temperature t h a t the time necessary for a determination was very long. The pure oils were cooled after measurement a t 90 O and the viscosity redetermined
Left limb.
0.3712 0.1925 0.1151 0.07561
0.3712 0.1928 0.1151 0.07558
2.694 4.943 8.688 12.23
7 5 P e r cent. Linseed. 0.3556 0.1866 0.1124 0.07427
0.3556 0.1866 0.1124 0.07426
2.812 5.354 8.897 13.47
TABLE;VI.-DOGFISH LIVER-SOYA BEANOIL MIXTURES. 25 P e r cent. Soya Bean. 30
500 700 90' 30' 500 70' 90 30'
500
ioo 90'
0.4121 0.2121 0.1234 0.05952
0.4125 0.4123 0.2094 0.2108 0.1227 0.1231 0.07934 0.07941 5 0 P e r cent. Soya Bean. 0.4099 0.4130 0.4115 0.2074 0.2082 0.2078 0.1215 0.1227 0,1221 0.7930 0.07954 0.07942 75 Per cent. Soya Bean. 0.4074 0.4071 0.4073 0.2072 0,2075 0,2074 0.1210 0.1214 0.1212 0.07852 0.07865 0.07859
2.425 4 ,744 8.123 125.9 2.430 4.813 8.190 12.59 2.455 4.822 8.251 13.72
T H E JOURlZiAL OF I l Y D U S T R I A L AiVD E,VGI.\;EERISG
880
TABLE VII.-DOGFISH LIVER-CHINA WOOD OIL MIXTURES. 50 Per cent. China Wood Oil.
Viscosity. A
F
Temp.
Right limb.
Left Limb.
30
....
....
35 70'
0.6926 0.3895 0.2085
909
0.1258
0.6942 0.3875 0.2080 0.1259
500
7
Average. 0,7950 0.6934 0 3885 0,2083 0.1259
Fluidity. 1.258 1.442 2.574 4.801 7.943
The viscosity and fluidity are pictured graphically in Figs. 1-3. According t o Bingham's fluidity hypothesis,I the fluidities of liquid mixtures should be
additive when there is no association or action between the components. This is true when the fluiditytemperature curve is linear. I t may be seen from Fig. 2 t h a t the fluidities of the dogfish, linseed and soya bean oils are nearly linear functions of the temperature and their curves have the same slope. Fig. 3 shows that mixtures of these oils have fluidi-
CHEMISTRY.
Dec., 1912
I t s fluidity curve does not have the same slope a s t h a t of the other oils, and also its fish oil mixtures do not give additive fluidities. The data in Table IV shows t h a t it changes in composition on being heated t o 90' since its viscosity a t 35 O increases from I . 589 t o I .946. The values for the other oils remain constant. I t is interesting t o note how much more viscous
the china wood oil is than the other three oils, t h i s property being a good test of its purity, since a little adulteration b y other oils of viscosity approximating the others would lower the viscosity of the china wood oil markedly. The viscosity curves of the mixtures are given for 30' in1Fig. I . As has been shown before, the linear
"'e
;S
50 75 40Campoai Cion
NO
Dec.,
1912
T H E J O C R S A L OF I S D L - S T R I A L A.VD EA\'\TGIAYEERI.'l'GC H E M I S T R Y .
88 I
position of which changes on heating. Linseed and dogfish liver oil mixtures have additive densities a t 30' while the other mixtures clo not. These data were of value to us, principally in our calculation of the viscosities.
viscosity nor fluidity curves for the china wood oil mixtures are linear. DESSITY.
There is little to be said in regard to the data furnished by specific gravity determinations. These were made by use of the Ostwald pycnometer. The re-
I N D E X OF REFR.iCTIOS.
The refractive indices were measured by the use of a microscope, the apparent and true depths of a column of liquid being compared. Fig. 6 shows t h a t the index of refraction is not a good criterion of the composition of linseed and soya bean oils when being tested for the presence of fish oil, but it should be valuable in the case of china wood oil, the refractive
sults are given in Table VI11 and shown by graphs 4-j. Fig. 4 shows that the density is a linear in'Figs. TABLEVIII.-DESSITIES 30'.
35 '.
OF
OILS. 500.
700.
Dogfish l i v e r . . . . . . . . . . . 0 . 9 2 2 8 . . . . 0.9175 0 . 9 1 3 1 Linseed. . . . . . . . . . . . . . . . 0.9251 . . . . 0.9219 0.9183 Soya b e a n . . . . . . . . . . . . . . 0 . 9 1 9 1 ... 0.9149 0.9095 China wood. . . . . . . . . . . . 0 . 9 4 1 1 0 . 9 3 8 9 0 . 9 3 2 4 0.9279 Dogfish liver-25 per cent. linseed. . . . . . . . . . . . 0 9237 ... 0 9158 0 . 9 1 2 0 Dogfish liver-50 per cent. .... 0 . 9 1 2 4 0.9051 linseed. . . . . . . . . . . . . . 0.9235 Dogfish liver-75 per cent. . . . . 0.9169 0.9100 linseed. . . . . . . . . . . . . . . 0 . 9 2 4 5 Dogfish liver-25 per cent. soya b e a n . . . . . . . . . . . . 0.9200 . . . . 0.9158 0.9111 Dogfish liver-50 per cent. soya b e a n . . . . . . . . . . . 0.9200 .... 0.9149 0.9076 Dogfish liver-75 per cent. soya b e a n . . . . . . . . . . . . 0 . 9 1 9 1 . . . . 0 . 9 1 4 1 0.9076 Dogfish liver-25 per cent. china w o o d . , . . . . . . . . 0.9268 0.9237 0.9149 0 . 9 1 0 3 Dogfish liver-50 per cent. china wood.. . . . . . . . . 0 . 9 3 0 0 0 . 9 2 8 9 0.9255 0.9209 Dogfish liver-75 per cent. china w o o d . , . . . . . . . . 0 . 9 3 4 8 0.9322 0 . 9 2 4 5 0.9230
90 '. 0.9082 0.9141 0.9063 0.9260 0.9090 0.8986
0.9051 0.9082 0.9055 0.9046 0.9074 0.9188 0.9204
index of which is considerably greater than that of the other oils. This has been brought out by
function of the temperature over the range at which we worked, except with the china wood oil, the com-
TABLEI X I n d e x of refraction r Oil a t 25 '. .......................... 1,4801 Linseed. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.4805 1.4837 S o w b e a n. . . . . . . . . . . . . . . . . . . . . . . . . . . . . China w o o d . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.5560 Dogfish-25 per cent. linseed.. . . . . . . . . . . . . . 1.4948 50 per cent. linseed. . . . . . . . . . . . . . . . . 1.4872 75 per c e n t . linseed.. . . . . . . . . . . . . 1 . 4 8 i O 25 per c e n t . soya b e a n . . . . . . . . . . . . . 1 4 i 9 0 50 per cent. soya b e a n . . . . . . . . . . . . . . 1 . 4 8 1 9 75 per cent. soya b e a n . . . . . . . . . . . . 1.48.34 25 per cent. china wood. . . . . . . . . . 1.4997 50 per cent. china wood . . . . . . . . . . 1 . 5 1 2 0 7 5 per c e n t . china wood . . . . . . . . . . 1 . 5 3 6 3
Acid number.
Saponification number.
Iodine number. ____-A__-_
1.
2.
2.91 3.98 3.08 1.47 3.06 3.41 3 75 2.85 2.94 3.00 2.60 2.14 1.78
2.94 4.01 3.08 1.49 3.06 3.43 3.68 2.85 2.92 3.03 2.72 2.19 1.76
Av. 2.93 4.00 3.08 1.48 3.06 3 42 3 72 2.85 2.93 3.02 2.66 2.17 1.77
1. 191.3 194.0 196.7 190 1 191.4 192 0 191 0 193 5 195.6 196 4 193 2 18i 6 188.5
2. 191 2 195.3 196.4 190 2 191 6 192.4 194 9 193.5 195.9 196.1 193.0 187.9 188.9
riv. 191.3 194.7 196.5 190.2 191.5 192.2 193 0 193.5 195.8 196.3 193.1 187.8 188 7
1. 122.7 152.1 130.5 (124 4 ? ) 129.2 137.9 144.3 123.3 123 1 126.9 128 0 133 2 135.6
2.
Av.
120.3 152 i 125 1 137 1
121.5 152.4 127.8 137.1 129.2 138.5 144.7 123.2 122 A 127.3 127.2 132 3 136.1
.
.
139.1 145.1 123.1 122 1 127 7 126.4 131.4 136.5
882
T H E J O L - R N A L OF I-\-DCSTRIdL
the work of Wise1 in a recent study of
this
oil.
A C I D , S A P O N I F I C A T I O N , Ah-D I O D I N E K U L I B E R S .
The results of the measurements of the acid, saponification and iodine numbers are presented in Table I X and Figs. 7-9. The first two numbers represent the milligrams of potassium hydroxide necessary for one gram of oil, while the iodine numbers were optained by Hubl's method, being per cent. iodine absorbed by the oil. The acid numbers are seen to be generally additive for the mixtures, while the saponification numbers offer no general behavior. The iodine numbers are additive for the linseed and soya bean mixtures, the china wood mixtures, however, giving a hyperbolic curve. As the iodine number of the linseed is so much higher than t h a t of the dogfish liver oil, this, as is well known, furnishes good evidence of the purity of the former. The numbers for the other vegetable *oils are so near those of the fish oil t h a t great value cannot be placed in these tests. Reference may be made t o the work of McIlhineya for a method of examining china wood oil. SUMMARY.
-.
I . The fluidities of fish and vegetable oil mixtures are additive, except where there is decomposition of either component on heating. 2 . The viscosities of these mixtures are also additive when the values for the components approximate each other closely. 3. The fluidities of vegetable oils are nearly linear functions of the temperature. It has been previously shown that this holds true for various fish oils. 4. The fluidity of china wood oil, a s it is unusually low, is a good test of its purity. 5. The specific gravities of the mixtures are approximately additive and vary linearly with the temperature. The density of china wood oil is very high. 6. The index of refraction, acid number, and saponification number of the mixtures do not allow any general conclusions. The acid number of linseed oil is high, and t h a t of china wood oil low, compared with t h a t of dogfish liver oil. The refractive index of china wood oil is lowered by the introduction of small amounts of fish oil. 7 . The low iodine number of the dogfish liver oil makes the presence of this oil in vegetable oils known by marked lowering of their iodine values.
u. s. BUREAUO F FISHERIES, \vOODS
HOLE,M A S S . ; CLARK COLLEGE,WORCESTER, MASS., RICHMOND COLLEGE,RICHMOND, VA.
THE RELATION OF THE REFRACTIVE INDEX OF SODABARIUM AND SODA-LIME GLASSES TO THEIR CHEMICAL COMPOSITION. BY
A S D E.YGI.\-EERI-\-G
and by Lorentzz and Loren23 as wz-I
Ibid., 4, 496 (1912).
I
X K'. w " S 2 d This specific refractive power (K and K') represents a relation between the refractive index and the density of a substance and is nearly, but not quite, constant over a large range of temperature. I t is also dependent on both the chemical composition and the constitution, or structure, of the compound. The product of the specific refractive power and the molecular weight is known as the molecular refractive power and has been found, for many organic compounds, t o be the sum of its several atomic refractive powers modified b y the structure or atomic arrangement within the molecule. These relationships have been studied to only a limited extent in connection with silicates. Owing to our lack of knowledge of the compounds which are formed in a glass or of their atomic structure, the derivation of atomic refractive powers for the glass forming oxides is impossible. Investigations have, therefore, been limited t o the study of mixtures of pure silicates. Larsen4 has shown that the specific refractive power of mixtures of calcium and magnesium silicate is a nearly linear function of the composition, calculated in percentages by weight, and the writer has shown5 t h a t this is also true, within limits, of soda-lime glasses. Neither the refractive index nor the specific refractive power of mixtures is necessarily a linear function of the composition. I n addition t o the changes in volume which may occur on mixing two liquids, there may be a similar change in the specific refractive power which may or may not be parallel with, or proportional to, the change in volume. Schutt6 and Pulfrich' have derived some rather complex formulae t o express these relations, both of which show excellent agreement with the experimental data. Schutt's formula is a s follows, A = -X I
k
- G ( a +-)'
in which K is the specific refractive power of the mixture, x k is the sum of the specific refractive powers of the components calculated by percentages b y volume, c is the percentage change in volume on mixing, and ( a b p ) is a factor dependent on a similar change in the specific refractive power. Pulfrich's formula is analogous t o the above and is as follows,
+
I ( = ' )
Received July 1, 1912.
THISJOURNAL, 4, 497 (1912).
Dec., 1 9 1 2
Nearly all of these expressions are based upon a constant known as the specific refractive power which has been defined by Gladstone and Dale1 as
EDWINWARDTILLOTSON, JR.
Many attempts have been made to discover the relation which exists between the refractive index and t h e chemical composition of substances, and many formulae have been derived t o express this relation.
CHEMISTRY.
I-c
' P h i l . Trans., 537 (1863). W k d . Amn.., 9, 641 (1880). 3 I b i d . , 11, 70 (1880). 4 A m . J . Sci., 28, 263 (1909). THISJOURNAL, 4, 246 (1912). 6 Z.Dhys. Chem., 9, 349 (1892). 7 Ibid., 4, 561 (1889).
*
= x k ,