THE JOURNAL OF INDUSTRIAL AND ENGINEERING CHEMISTRY
146
Vol. 14, No. 2
Electrolytic Recovery of Magnesium from Salt Works Residue’ By K. S. Boynton and Verne Langford with J. F. G. Hicks DEPARTMENT OF CHBMISTRY, STANFORD UNIVERSITY,STANFORD, CALIFORNIA
In view of the fact that the raw materials and power necemry for the product,ion of magnesium a t a compamtively low cost are to be found on t.he Pacific Coast, the problem of producing the met.al from these raw materials was undertaken.
Raw MATERIALS In the experimental work crude hydrated magnesium chloride, commonly called “bittern,” was used as the source of magnesium. Large quant,ities of this bittern form the residues from the recovery of salt from sea water. The composition is as follows: Ferric and alumini Calcium chloride, MgCln.GHz0.. . . . Sodium and potassium chlorides.
................
.............
..............................
Per cent 0.05 Trace 0.59
94.42 4.32 0.59
Commercial ammonium chloride and sodium chloride were used as dehydrating agents. Dehydration of t~hemagnesium chloride by heating it; alone always resulted in the formation of basic salts, and the dehydrated mass was worthless for elect’rolytic purposes. The mixture finally selected contained: Parts
PURIFICATION OF THE MAGNESIUM Metallic magnesium was deposited a t the cathode surface in small molten globules or “shot,” which slowly increased in aize and were finally detached by the convection currents set up in the bath. Theqe currents distributed the globules throughout the bath, the specific gravity of which was very nearly that of the globules. This accounts for the fact that the globules could be skimmed from the surface of the bath or scraped from the bottom of the cell during the same run. At the end of each run the fusion was poured into a mold, cooled and crushed, and the metal was sieved out. The “buttons” of magnesium varied in size from minute spheres to globules weighing over half a gram. To free the metal from occluded flux, carbon particles, oxides, etc., some dehydrated magnesium chloride was fused in a carbon crucible, fitted with an air-tight lid containing two holes to permit ingress and egress of illuminating gas. The magnesium shot were thrown into the fused mass and the lid clamped on tightly. The degree of purification of the metal is shown by the following table: Insoluble Iron Aluminium
$2,
The following advantages are claimed for it : I-Reduction of hydrolysis to a minimum. 2-Minimum amounts of sodium and ammonium chlorides. 3-A maximum current efficiency. 4-A maximum concentration of magnesium in the bath, for electrolysis 5-Lowest viscosity of ‘any bath tried. 6-Higher percentage yield. 7-Relatively low melting point.
DEHYDRATION OF CRUDE MAGNESIUM CHLORIDE This mixture was dehydrated by heating gradually to 200’ C. in a revolving st,eel cylinder, through which air was forced for the more rapid removal of steam and small quantities of hydrochloric acid gas. 4 few flint pebbles in the cylinder served t o keep the mixture broken up. The anhydrous chlorides were heated in iron pots until ammonium chloride was entirely driven off, and tjhe residue had melted. This fusion was electrolyzed.
ELECTROLYSIS Two types of electrolytic cell were used: In one, both electrodes were carbon; in the other, the anode was carbon and t.he cathode iron. Two patterns of .the carbon-iron cell, one the’ordinarg cylindrical pattern, and the other V-shaped, were found to be satisfactory. The carbon-carbon cell yielded the best results from a laboratory standpoint, but was considered impractical for work on a large scale. Electrolyses were conducted in an atmosphere of illuminating gas. Obviously, any ot’her cheap reducing or neutral atmosphere would answer the same purpose. It was possible to produce inetal a t a minimum voltage of 4,but the average for eleven runs was 6.2. As only a small generator was available for these experimental runs, t’he average amperage was only 100. The averagc faraday efficiency for eleven runs with the iron cylindrical cell was 30 per cent, and the average time was 1.7 hrs. After readily reproducible conditions were obtained, it was a comparatively simple m a t h to predict the yield of a given run to within 25 per cent of the actual yield. 1 Received
August 18, 1921.
Magnesium
Crude Magnesium Per cent Trace 0.74 Trace Trace 0.02 99,24
Refined Magnesium Per cent None 0.42
Trace Trace Trace
99.58
RECOVERY OF VOLATILIZED AMMONIVM CHLORIDE Some experiments were made on the possibility of rccovering the ammonium chloride by condensation. Small amounts of moisture very markedly lowered the efficiency of this condensation process, although better than 50 per cent recoveries were easily attained. ESTIMATE OF COSTOF PRODUCIKG MAGKESIUM ON SAN FRllNCISCO BAY From the analysis of crude “bittern” it will be seen that, theoretically, 8.88 tons (2000 lbs.) of this raw material are required to produce 1 ton of magnesium. Allowing 2.5 per cent wastage to maintain the purity of the electrolyte, 11.10 tons of “bittern,” 1.64 tons of salt, and 1.11 tons of ammonium chloride will be required to produce a ton of metal. DEHYDRATION-since hydrated magnesium chloride contains 53.5 per cent of water, the weight of water in 11.10 tons of raw material would be 5.94 tons. One barrel of fuel oil will evaporate 1.3 tons of water in dehydrating alum. As the “bittern” loses its water a t a lower temperature than alum, the above ratio seems to be a safe estimate. At this rate 4.57 bbls., or, to allow 10 per cent for removal of additional moisture of the mixture, 5.02 bbls. of fuel oil would be required for the dehydration of sufficient, material to produce one ton of magnesium. TIME AND POWER NECESSARY TO DEPOSIT OKE TOK OF MAGNESIUM AT A FARADAY EFFICIEKCY OF 30 PER CENT-1 fara-
day =96, 500 coulombs =0.0264 lb. of magnesium.
2000 -
0.0264
= 75,757.57 faradays = 7,310,605,505 coulombs per ton of
magnesium. Therefore, to produce 1 ton of magnesium 24;368,685,010 coulombs would be needed. 24J368f685’010 = 6,769,079 ampere hours t o produce a ton 3600 6759 = 9401.4 ampere months 720 (hrs. in 1 mo.)
Feb., 1922
T H E JOURNAL OF INDUSTRIAL A N D ENGINEERING CHEMISTRY
147
ILLUMINATING GAS-For a cell of sufficient size to produce 9401.4 X 6 (volts) = 56.41 kw.-mo. Allowing 25 per cent (14.10) for interruptions, 70.51 kw.- a ton of magnesium in 720 hrs., 50,000 cu. ft. of illuminating gas would be required to form a reducing atmosphere above mo. would be required for the production of 2000 Ibs. FREIGHT AND HAKDLING OF RAW MATF,RIAIB--T~~ p h t t the electrolyte. would be within a few miles of the source of the raw material, COST O F PRODUCING 1 TONOW MAGNESIUM hence freight and handling would be a comparatively small Crude MgCIz.6HzO Il.ltonsat$30.00 $ 333.00 hTHiCl 11.1 tons a t $60.30 ........... . .. .. .. .. .. .. .. .. 67.00 item of expenqe. For each ton of metal produced, 18.85 Commercial Commercial NaCl 1.64 tons a t $7.00 . . 11-48 Oil for dehydration 5.02 bbls. a t $1.80 . . . . . . . . . . 9.04 tons of material would need to be freighted and handled. Illuminating gas 50,000 cu. ft. a t $0.50 . . . . , . . . . . 25.00 LABoR-Six men, including a chemist, would handle the Electric power 70.51 kw.-mo. a t $4.00. . . . . . . . . . 282.04 Oil for heating cell 5 tons a t $10.00 proposed plant. Anode carbon 200 Ibs. a t $0.04 .. .. .. .. .. .. .. .. .. .. 50.00 8.00 and handling of raw OIL FOR HEATIKG csLI.-The crude oil for outside heating Freight materials 18.85tons a t $5.00 . . . . . . . . . . . . 94.26 6 men, including a chemist . . . . . . 1000.00 of the cell would be required, a t most, during one-fourth of Labor the run. For this heating 5 tons would be required. TOTAL .............. $1879.81 CARBON AIioDEs--,kbout 200 lbs. of carbon anodes would 182000 79.81 = $0.939 per Ib. to produce a ton of magnesium be consumed per ton of metal.
Loss of Carbon Dioxide from Dough as a n Index of Flour Strength”” By C. H. Bailey and Mildred Weigley DIVISIONOF AGRICULTURAL BIOCHEMISTRY, MINNESOTA AGRICULTURAL EXPSRIMENT STATION,UNIVERSITY FARM,ST. PAUL, MI”.
CHaRACTER OF FLOURS EMPLOYED 1K INVES’IIGA‘I’ION Flour strength studies conducted during the past quarter century have usually employed as the ultimate criterion of Two flours were employed, a “strong” flour milled from strength the comparative physical properties of yeast- hard spring wheat, and a “weak” flour milled a t Spokane, leavened loaves of bread produced on baking. I-Iuniphries Washington, from typical soft wheat of that region. The and Biffin3 suggest that “a strong wheat is one which yields composition, and the baking qualities of the two flours, flour capable of making large, well-piled loaves.” Quality determined by the method previously described by of the loaves is thus determined by their size or volume, and are shown in Table I. The strong flour contained a materially the texture and other related properties of the crumb. Bali- higher percentage of crude protein, and on baking gave a loaf ing tests, however, are not exact procedures which yield of larger volume and better texture than did the weak flour. uniform results when the same materials are employed. The differences in these respects were considerable. As the Judgment comes into play to a considerable extent, and flours mere of about the same grade, as indicated by the ash experts sometimes differ in their opinion of the relative merits content and color score, they were well suited to a study of of different flours which are being compared. Such tests properties rcsponsible for variations in baking Strength. do not of necessity indicate the reasons for variations that are TABLE I-ANALYSES AND BAKINGTESTSOF STRONG AND WEAKFLOURS observed, or the methods for effecting desired improvements. ExpanFor these reasons efforts have been directed toward developCrude Volume simMois- Protein of eter ing more exact methods for testing the properties of flour ture (NX5.7) Ash Loaf Color Tex- Test MARK Percent Percent Percent Cc. Score ture Cc. that are of significance in this connection. The work reweakflour .... .. 10.88 8.00 0.52 1200 97 85 520 ported in this paper represents an attempt in this general B878, Composite,strongflour 8.95 12.00 0.56 1580 98 100 910 direction. That the two flours also varied widely in colloidal properPROPERTIES THAT CONSTITUTE STREWTII ties was established by the procedure developed by Gortner A consideration of the factors involved in the production and Sharp,5 in which the change in viscosity of ii mixture of “large well-piled” loaves suggested to one of the authorsi of water and on the addition of normal lactic acid solution is that “the strength of flour is determined by the ratio between traced by a iLIacR/Iichael viscosimeter. Table I1 gives the the rate ol production of carbon dioxide in and the rate results of this testUEQuite evidently there are decided differof loss of carbon dioxide from, the fermenting mass of dough.” ences in the properties of these two flours which are responsible The absolute rate of carbon dioxide production in doughs for variations in the viscosity of such preparations. The made from different flours varies, to be sure, through fairly I N MACMICHAEL DEGREESOF FLOUR AND WATER wide limits. When known, the rate can be varied in the TAB& 11-VISCOSITY MIXTURESOF WEAKAND STRONG FLOURS desired direction, however, by adjustments in the formula, (Mixture contained equivalent of 25 g. dry matter f 100 cc. water) Normal Lactic Weak Flour Strong Flour and particularly in the proportions of yeast, yeast accelerators, Acid Added (B878) (Composite) cc. Degrees MacMichael Degrees MacMichael salt, fermentable sugars, and diastase preparations. The 68 0.0 32 retention of gas is much more difficult to bring under control, 0.5 73 32 1.0 104 38 since it is apparently determined in large measure by the 1.6 135 43 percenhage and physical properties of the gluten proteins in 161 2.0 50 183 2.5 53 the flour. Hence, in testing flours, the ratio between pro196 3.0 56 212 4.0 60 duction nnd loss of carbon dioxide apparently becomes of 22 1 5.0 61 practical significance, and this was the point attacked in 222 6.0 62 222 8.0 62 the present investigation. 21 9 10.0 fil Presented before the Division of Biological Chemistry a t the 62nd Meeting of the American Chemical Society, New York, N. Y ,September 6 to 10, 1921. Published with the approval of the Director as Paper No 274, Journal Series, Minnesota Agricultural Experiment Station. * “The Improvement of English Wheat,” J. Agu Sci , 2 (1907), 1 4 C. H. Bailey, THIS JOURNAL, 5 (1916), 208.
III6 “The Physicochemical Properties of Strong and Weak Flours. Viscosity as a Measure of Hydration Capacity and the Relation of the Hydrogen-Ion Concentration to Imbibition in the Different Acids.” Presented before the.Division of Physical and Inorganic Chemistry a t the 62nd Meeting, A. C . S., New York, N. Y., September 6 to 10, 1921. 8 Determinations made by Paul F. Sharp.
