A Process for the Utilization of Pyrite Ash Citation Bridgeford, Frank Robert (1922) A Process for the Utilization of Pyrite Ash. Bachelor's thesis, California Institute of Technology. doi:10.7907/P747-5S86. https://resolver.caltech.edu/CaltechETD:etd-04192004-160316 Abstract No Abstract. Item Type: Thesis (Bachelor's thesis) Subject Keywords: Chemical engineering Degree Grantor: California Institute of Technology Division: Chemistry and Chemical Engineering Major Option: Chemical Engineering Thesis Availability: Public (worldwide access) Research Advisor(s): Unknown, Unknown Thesis Committee: None, None Defense Date: 1 January 1922 Record Number: CaltechETD:etd-04192004-160316 Persistent URL: https://resolver.caltech.edu/CaltechETD:etd-04192004-160316 DOI: 10.7907/P747-5S86 Default Usage Policy: No commercial reproduction, distribution, display or performance rights in this work are provided. ID Code: 1415 Collection: CaltechTHESIS Deposited By: Imported from ETD-db Deposited On: 22 Apr 2004 Last Modified: 03 Oct 2019 23:08 Thesis Files Preview PDF (Bridgeford_fr_1922.pdf) - Final Version See Usage Policy. 2MB Repository Staff Only: item control page

Alles, Gordon A Ames,Panl R Beman;illerd <3: Bridgeford, Prank R. Burks, Gillies, Robert W Jesse C Knight,Alfred ¥ Reynolds, Maynard 5S Ritchie,Chéerles Fisher Vesper, Warner, Howard G Lester 0° THESSS 1922 A new acridine dye. The specific heat capacity of heavy pétroléum products. ™he Tiowig procese for caustic soda using trana as a rew materio?, A process for the utilization of pyrite ash. Seni-works development of a germanium extraction process. An attempt to dehydrogenate cottonseed oil. Perchloric acid from ammonium perchlorate and oxides of nitrogen. The utilization of a denosit of magnesite located nesr Porterville,California as a source of raw materiel for the manufacture of C.P. Epsom salts. An investigation on the production of sodium thiosulfate from trona. The preparation of carbon monosulphide and hydrocyanic acid by the use of activated carbon. The reaction between nitrogen and methane in the silent electric discharge. California Institute of Technology Pasadena, California CLASS OF 1922 Department of Chemistry Wndergraduate Theses of Gordon A.Alles Paul R.Ames Willard J.Beman rank R.Bridgeford Jesse C.Burks — Robert W.Gillies Alfred ™.KXnight Maynard 5.Reynolds Charles *isher Ritchie Howard G.Vesper Lester 0. Warner A PROCESS for the UTILIZATION OF PYRITE ASH ~0o0000~ Thesis by Frank R. Bridgeford. In partial fulfillment of the requirements for the degree of Bachelor of Science in Chemical Engineering. CALIFORNIA INSTITUTE OF TECHNOLOGY Pasadena, California. 1922. A PROCESS FOR THE UTILIZATION OF PYRITE ASH Introduction. The purpose of this investigation was to find a commercial process by which the iron could be recovered, as ferrous sulfate, from burnt pyrite. Pyrite, Fe So» is very widely distributed in na- ture. The largest deposit is at Rio Tinto, Spain, where it occurs associated with both gold and silver. The chief mines in the United States are in New Jersey, Penn- sylvania, New York, Massachusetts, Virginia, Georgia, Colorado, and California (1).* Practically the entire output is used in the manufacture of sulfuric acid. The total consumption in the United States, for 1915, was 1,358,758 long tons. By 1920 it had fallen to 607,606 jong tons, because of the increased use of sulfur (2). The burnt residue of ferric oxide, known as "blue-billy", has various uses. At the Stauffer Chemical Company's plant in Los Angeles, it was used solely as a road-bed material. Although the roads are at first firm and dry, they are soon pulverized by the traffic and in wet weather become very muddy. * Numbers refer to Bibliography at end of paper. The material is used as a catalyst to promote “the combination of sulfur dioxide and oxygen (3). It is also used to catalyze the ferrite process for the manufac- ture of sodium hydroxide (4). | A process has recently been developed in Italy, for. the preparation of ferric sulfate from pyrite ash and niter cake (5). The moistened cinders may also be used to absorb hydrochloric acid from salt-cake furnace gases, etc. (6). Paint pigments may be made from the ash elther directly or by first forming ferrous sulfate (7). Numerous methods are proposed to cut down the sulfur content of the ash so that it may be used in the manufacture of pig iron (8). There are two main processes for the manufacture of ferrous sulfate from pyrite ash. . In the first, known as the A. and P. Buisine process (9), the material is heated with 60° Be'. sulfuric acid (boiling point 190°C.) to somewhat less than 300° c. Although the resulting mass of ferric sulfate is not immediately soluble in water, it becomes soluble by hydration at somewhat higher tenmpera- ‘tures. The solution, after being reduced to ferrous sul- fate by digestion with iron, is evaporated to 27-28° Be’. (measured at 90°C.) and crystallized with pieces of scrap- iron suspended in the vat. In the other method devised by Pezzolato (10), “the pyrite ash is heated to a red heat in a closed vessel with enough powdered charcoal to reduce it te magnetic iron oxide. After concentrating with a magnetic separa- tor, it is treated with sulfuric acid until dissoived. From there the process is similar to the one described above. In both cases any copper present is separated out on the excess scrap-iron. Plan of Experimental Work. The logical method of utilizing burnt pyrite from a sulfuric acid plant would be to produce ferrous or ferric sulfate. The two methods, at present used for this purpose have been described above. Preliminary experiments with the ash showed it to be but slightly soluble in cold sulfuric acid of moder- ate concentrations. This partially explains why, in the Builsine process, a concentrated acid at a very high temp- erature is used. This method, besides requiring a con~ siderable initial outlay for equipment would also have a ‘large upkeep cost due to the high pressures used. _ rhe Pezzolato process appears to be unnecessar- ily complicated besides being difficult to carry out in practice. The magnetic separation removes the non-mag- 4. netic impurities from the iron. However those impuri- ies, which are soluble in sulfuric acid, may probably be separated by fractional erystallization. A preliminary experiment with a horse-shoe magnet showed that consider- able finely divided carbon, from the reduction, was carried mechanically with the large iron particles and « clean separation was not obtained. The material would need to be ground extremely fine to separate each iron particle from the adjacent gangue and this would greatly increase the cost of operation. The first step of this investigation was to determine the efficiency of the solution of the ash with various temperatures, times of treatment, concentrations and amounts of acid. It was hoped that with this data an inexpensive procedure for the solution of the ash gould be worked out. it was desired to determine from an economic standpoint the relative efficiencies of the several processes and also to test out the method finally selected on a large scale. Analysis of the Pyrite Ash. The sample for this work was obtained from the Steuffer Chemical Company of Los Angeles. It contained material ranging from about 50-mesh up to lumps nearly Se nalf an inch in diameter. The finer portion of the sam- “ple appeared to ve well burnt to a reddish purple color but the larger lumps when broken up showed a core of un- burnt pyrite. In some of these a yellow substence, ap- parently free sulfur, could also be seen. The thorough- ly purnt pyrite could be pulverized readily but the larg~ er lumps, containing the unburnt portion, were hard. | Lunge states (11) that the pyrite cinders when properly burnt have the red color of ferric oxide, but if much copper is present they are black. They are light and porous, have no unburnt core in the larger pieces, and contain no fused slag, (the latter showing that the temperature of the combustion was correct). The sample used in this work fulfilled these requirements, except that it contained some unburnt material. | The sample was ground down to approximately 80~ mesh in a Braun Pulverizer, type UA4, driven at about 570 YePel. Moisture Content VYhen a sample was dried in an electric over at 108° c., it gave a distinct odor of sulfur dioxide. It reached a constant weight, after heating for two hours, showing a less corresponding to a moisture content of 6. 9,40%. Another sample, dried to a constant weight in a guifuric acid dessicator for 50 hours showed a loss of 2.15%. The higher value is undoubtedly due to the loss of sulfur. This action has been noted by Jones (12) who states thet hydrogen sulfide and thio-sulfide acid are formed. (lunge, however, (135) recommends determining the moisture in sulfur by heating a sample for several hours at 100°). The latter of the two results, giving a mois- ture content of 2.15% of the original semple, is taken as the correct value and all analyses reported in this paper are determined from the dessicator-dried sample and cor- rected by this value. Free Sulfur. The free sulfur will be assumed to be entirely crystalline. It was determined by extraction, for 114 hours, with c.p.carbon disulfide in a Soxhlet apparatus. The result computed from the loss in weight of the sample gave 0.88% free sulfur. ¥Yhen computed from the weight of the residue, after volatilizing the carbon disulfide, 1.02% was obtained. The mean of these two, giving 0.95% free sulfur in the original sample, is taken as correct. fTotel Sulfur. At first it was attempted to determine total sulfur by the Fresenius method (14). In this procedure the sample is fused with sodium carbonate and potassium nitrate in a nickel crucible. The analysis was attemp- ted three times but each time the crucible was eaten through before the charge could be completely fused. The difficulty appeared to be due to the high temperature re- quired to fiux the mixture. The sodium peroxide fusion method (15) was tried as it was known that this would give a readily fusi- ble mixture. Here the sample is fused with sodium perox- ide and sodium carbonate. The sulfate is extracted with dilute acid and determined as the barium salt. Two de- terminations gave 3.04 and 2.81% of total sulfur in the original material. A blank test showed no barium sulfate precipitate. — Mellor's fusion method (16) was used as a check. Here the sample is fused with a sodium carbonate and sodium nitrite mixture. The result obtained by this ‘method was 2.96%, giving an average for all three of 2.94% which is taken as the correct value. Determination of Iron. During the first part of the work the iron was determined by the permanganate method after reduction by stannous chloride (17), but later the electrometric di- “ehromate method was used (see below). Two analyses of the pyrite ash by the permanga- nate method gave 48.42 and 47.65% of iron in the original material. (Average value 48.03% iron). A definite end- point and consistent results were difficult to obtain by this method. The chief trouble seemed to be due to the lack of permanence in the end-point when titrating an ac- tual iron solution. Experience in-this analysis wouid probably remedy the difficulty but since the time was limited it was thought advisable to try another method. The electrometric titration of iron by potas- sium dichromate has been studied by Hostetter and Roberts and found to be very satisfactory (18). While the pro- cedure used here is essentially the same as described by them a few minor changes were made. A diagram of the apparatus is shown in Fig. 1. The titration nhalf-cell consisted of a 250 cc. beaker, surrounded by cotton and asbestos paper to provide heat insulation. The tempera- ture of the solution during the titration dropped from about 100° C- at the start to from 20° to 50° at the end, “WIRING DIAGRAM FOR ELECTROMFTRIC TITRATION ales Grif cx eH * Electra storage battery, £.0% volta. * Galvanometer. PO L 4, nated sensibility 1.06 . Titration cell. . Decade resistance box, 10,000 chms, PCL #3 . tt it tt " " PCL, 4 10. aepending on the time taken for the analysis. This be- came shorter as experience was gained and varied from two and a half hours for the first few titrations to about half an hour for the last ones made. The capillary at the end of the bridge from the calomel cell was flushed out at intervals from an elevated reservoir containing l-normal petassium chloride solution saturated with calo- mel. The sum of the two resistances in the decade boxes ‘was kept always at 10,000 ohms. By this arrangement the voltage of the titration cell is directly proportional to the resistance of one of the boxes. The voltage of the storage battery, during the two months it was used, did not drop more than 0.004 volt. The procedure used in the analysis is as fol- lows: The solution to be analyzed ismade up to 500 ec. in a standard volumetric flask. A 50 ec. portion is placed in the 250 ce. beaker and 100 ce. of 355.0 normal sulfuric acid added. The solution is allowed to stand for about ten minutes in the air or from three to five minutes in running tap water until the temperature has _ dropped to about 100° C. While this is being done, the electrode is ignited and the tip of the capillary bridge rinsed out. The sdlution is then put in place and stan- il. nous chloride run in until the voltage is lowered to about + 0.05 volt. The dichromate solution is then ad- ded in about 0.05 ec. portions and the e.m.f. read to the nearest 0.002 volt. Hand stirring was used through- out, being stopped for about two seconds before each reading was taken. It was found that the e.m.f. at the ‘stannous chloride end-point rises from ten to twenty min- utes before reaching a constant value. The average in- crease at this end-point is about 0.28 volt. The second or ferrous end-point comes to a constant value in less than one minute with a rise of about 0.16 volt. It was noticed that when hydrochloric acid was used instead of sulfuric acid in making up the solution the stannous chloride end-point passes through zero, while with sulfuric acid the curve is entirely positive. In all, over fifty analyses were made by this method and, while at no time could an analysis be made as rapidly as with the permanganate method, the results ob- tained check with one another and are regarded as much more reliable. In every case in which an indistinct . end-point was obtained, it was found that the platinum electrode in the titrating solution was at fault. If the glass near the tip is cracked the results are ex- tremely variable. 4000 a Electromerric nee Titration Curve. 3000 % 2500 > — Ww S hy /500 /000 PQ /ron Content———> 0500 0000 oO / #2@ 38 4 5§ 6 7T 8 9 105 CO Ae Cr, 0, | 13. A representative set of curves is shown in ; Figure 2. An idea of the accuracy of this method of analysis may be obtained from the standardization data of the dichromate solution. The normality of the solu- tion when computed from the amount of c.p. potassium di- chromate used in its preparation was 0.12598. The nor- mality found by an electrometric titration with standard- ization iron wire was 0.12592, the difference being .05%. The latter value is taken as the correct one because it includes all errors inherent in the method. It is prob- able that some of the analyses, to be given below, are not this accurate, for in a few the solution was added in 0.10 cc. portions at the end-points. The maximum error due to this cause depends on the total iron con-~ tent and in no case is greater than about 1%. The amount of iron in the pyrite ash, deter- mined electrometrically and calculated on the basis of the original sample was 48.22%. This agrees fairly well with the mean permanganate value (48.03%) and will be used (as the correct value) rather than a mean. The sample for the electrometric titration was dissolved in hydrofluoric acid while the one for the permanganate was dissolved in hydrochloric. The agreement shows that 14. all of the iron in the pyrite ash is soluble in hydro- ehloric acid. Silica. The residue from the hydrochloric acid treat- ment was a sandy white and probably pure silica; of the other impurities found in the qualitative analysis: 5S, Cu, Al, Zu (7), and Ca, all but the first have soluble chlorides. While the free sulfur is insoluble, it may be destroyed by ignition. The residue from each of the two analyses was filtered, washed, ignited,and weighed. The weights give as the per cent of silica in the origi- nal sample the values 23.679% and 23.682% with a mean of 23.68%. The state in which the combined sulfur is pres- ent in the sample is unknown. From the large amount of unburnt pyrite in the center of the lumps, it is reason- able to assume that most of it is present in that form, although some may be in the form of sulfate, either of iron or the other metals... Assuming that all of the com- bined sulfur is present as pyrite and that the remainder of the iron occurs as ferric oxide, the following com- puted results are obtained (based on the original sample): 15. Moisture 2.15% Free. 'S 0.95% Fe So 3.72% Feo Os 66.53% Si 0, 23.68% Undetermined: Cu, Al Zu (7?) and Ca oxides or sulfates 2.97% 100 .00% * Theorstical Behavior with Sulfuric Acid. Ferric oxide is "attacked by acids with diffi- culty, the more so the higher it has been heated. Hydro- chloric acid is the best solvent, in which it is more quickly soluble by long digestion at a gentile heat than by boiling. % % 4% It is most easily soluble in six- teen parts of a mixture of eight parts sulfuric acid and three parts water." (19) This latter is equivalent to 6.5 times the theoretical quantity of 25.4 normal acid. Friend states that nitric acid does not attack the ig- nited oxide (20). No direct theremo-chemical data for the solu- tion of ferric oxide in sulfuric acid could be found. 16. The following values are given by Landolt-Bornstein (21) for the heats of formation of the compounds involved: Fey 0g, solid +195,600. cal. Hp S04 16.6 Hp0 #208 ,600 (6=-normat ) Feo (SO,)z, solution +650,500. HO, liquid +68,150. Therefore in the equation Feo Oz (solid) + 3Hy SO, (solution) == Fep (80,)3 (solution) # 3H,0 (liquid)y 4H, 4H = = 35,550 cal., and the reaction is exothermic. It would be expected, therefore, that a small increase in temperature although increasing the rate would decrease the solubility. If the temperature is raised above the boiling point of the solution, both the speed of the re~- action would increase and the equilibrium would be shif- ted forward, because water, a product of the reaction, will then be partially removed. If the temperature is increased beyond 317° C., the pressure remaining at ate mospheric, the sulfuric acid completely boils and the re- action stops. Further increase beyond 530° Cc. reverses the reaction by decomposing the ferric suifate to ferric oxide (22). The amount of pyrite ash dissolved by a given solution of acid depends on both the concentration and the total amount of hydrogen ion available. Sinee the {fonization decreases with increasing acid concentration and since for low concentrations the nydrogen ion is al- so low, it would seem that there should be a concentra- tion most efficient for dissolving the ash. Tne greater the amount of acid present the larger should be the per cent of the total iron soluble in a given time. The magnitude of these effects cannot be determined theoret- cally f the concentrated solutions which are to be used but must be ascertained by experiment. Since the rate of the reaction depends on both the hydrogen and the sulfate ion concentration, it might be expected that it would fall off rapidly as the amount of ferric sulfate increases. Since, moreover, the ion- ization constant for a concentrated soiution is much smaller than for a dilute, a given amount of ferric sul- fate in solution in each would cause a greater decrease in the rate in the first case than in the latter. eo 18. Solubility Experiments on the Original Material. All of the experiments reported below were made in one of two ways: A 2.50 gram sample was weighed out to the near- est 0.01 gram and placed in a thin 6-inch test-tube to which the proper quantity and normality of acid was then added. The test-tube was placed in a 50-cc. boiling flask, nearly filled with concentrated sulfuric acid, and heated over a small flame. The temperature was regulated to the nearest degree using a thermometer, whose bulb was completely immersed in the acid, and so placed that it just touched the outer surface of the test tube. In near- ly all cases the solution was stirred by hand, with a glass rod, for about fifteen seconds per minute. (In the two exceptions noted below, the stirring was not possible because, due to either the high temperature or the small quantity of acid added, the mixture caked-up before the end of the run). After the sample had been treated in this manner for times ranging from 15 to 180 minutes, the test-tube was removed, the solution immediately filtered, and the residue thoroughly washed with boiling water. The filtrate was made up to 500-cc. in a volumetric flask and a 50-cc. sample taken for analysis. The permanganate method was used and, for the reasons given above, these 19. results, which are all underlined when quoted, are to be | regarded as somewhat less accurate than those obtained py the other method. The results are, if anything, too high so that the errors, which are perhaps as high as 2% in some cases, do not affect any of the conclusions to be drawn below. The second procedure used the electrometric di- chromate method of analysis described above. The maxi- mum error is perhaps about 1% with the average error much lower. Samples of 1.250 grams were weighed out to the nearest 0.005 gram and placed with the acid in small 50- ec. flat-bottomed flasks. All of the samples were kept at from 15° to 20° c. (normal room temperature), and the stirring effected by shaking the solution about once every three hours in the daytime. The solutions were filtered, washed and the sample taken as described above. The same amount of acid was added to both the reduced and the unreduced samples where the theoretical quantity specified was the same. (The error in this, based on an analysis of the reduced material, was about 0.1%. The decrease in weight due to the oxygen lost was almost exactly balanced by the weight of the carbon remaining). All of the results given below are expressed 20. on the basis of the per cent of the total iron, present at the start, dissolved by the acid treatment. The effect of the temperature of the acid on the amount dissolved. Samples, treated with 5.6 times the theoretical amount of 35.0 normsl sulfuric acid for 15 minutes, showed the Following results: Temperature Ratio of % in ° ¢, % Dissolved to abs. temp. 100° 34.8% .09 175 67.8 15 250 71.2 014 ‘These results show that the amount of the ash dissolved by suifuric acid is greatly increased by in- creasing the temperature. For high temperatures it seems to be roughly proportional to the absolute temper- ature but for lower ones the rate falls off. The second conciusion is shown more clearly by another experiment. A semple treated with 5.6 times the theoretical amount of 6<normal acia for 15 minutes at 90° C. showed 8.91% dissolved while one treated with 8.4 times the theoretical amount of 6=normal acid for ten days showed only 9.24% dissolved. It will be shown be- low that the amount of acid used has but little effect on the percentage dissolved so that it may be seen that in- creasing the temperature from 20° to 90° ¢c. is nearly 21. equivalent to decreasing the time of treatment from ten days to fifteen minutes. fwo sanples were treated with 5.6 times the theoretical amount cf 20e—normal acid for 15 minutes: Temperature % Dissolved 130° c. 72.4% 120° c. 65.4% : In the first cease all of the acid boiled away and left a pink colored residue of ferric sulfate and ferric oxide. In the second case the acid did not boil. These results show that the amount cf soluble material formed is some- what increased by allowing the water contained in the acid, or formed during the reaction, to evaporate. The effect of the concentration of the acid. Sanples were treated with various coneentrations of acid at temperatures ranging from 8° to 25° below the boiling point of the acid. The time of treatment was 15 minutes, using 5.6 times the theoretical amount of acid. Concentration Temperature B.P. % Dissclived Ratio of of of % to Treatment. Acid. Cone. én 95° 106° 8.91% 0.247 10 n 105° 113° 20.4% 0.204 16 n 115° 127° 52.9% 0.235 20 n 120° 150° 65.4% 0.164 35 on —- BB0° 281° «712% 0.058 22. These results represent about the highest temperatures “end therefore the maximum-solubilities which can be at- tained without boiling the acid. If the last result be neglected the solubility can be approximately determined by the empirical formula: % dissolved = 0.2 x (conc. of acid)” It would have been very interesting to compare the per cent dissolved with the concentration of the acid, the temperature remaining constant, but because of the lack of time this was not possible. The effect of the amount of acid. The solu-~ bility of the ash in 35.0 normel acid at 250° Cc. for 15 minutes was determined: Amount of acid. % Dissolved. 2.8x theoretical 58.2% 5 66x " 71.2% 8.4x " 72.8% The amount of the acid used is seen to have comparatively littie effect on the per cent dissolved. These results ere not complete enough to form the basis for any empiri- cal relation but they indicate that the increased yield from large amounts of acid is not large enough to compen- sate for the increased manufacturing cost. Effect of changing both the concentration and Zoe the amount of acid, keeping the volume constant. It has - peen found by previous experiments that the higher the concentration of acid the larger the amount of solubie iron formed per unit weight of pyrite ash. However, the higher the concentration of the sulfate ion the smaller the amount of ferric sulfate which can go into solution. A coating of ferric sulfate may form on the surface of the particles in the more concentrated solutions and pro- tect them against further action by the acic. But in- creasing the sulfate ion concentration automatically de- creases the hydrogen ion concentration. If this decrease is appreciable, as it probably would be in a concentrated solution, it would diminish the rate of the reaction as explained above. It might be expected, therefore, that the solu- bility in a concentrated solution at first would be much greater than for a dilute soluticn. As the reaction proceeds, however, and ferric sulfate is formed and ion- ized, the reaction in the concentrated solution should slow down, for the two reasons given above. The reac- tion in a dilute solution, due to the greater ionization constant, (and hence the greater hydrogen iron concentra- tion), would continue at a more uniform rate. To test this experimentally a series of runs were made at room temperature, extending over a period 24. of three weeks.. | Time in days: 1 3 vi 10 14 21 Acid used: © 8.4x theor.of G-n 4.56% 7.30% 7.41% 9.24% 9.46% 10.70% 4,2x theor.of 5-n 5.65 5.76 5.69 6.73 8.32 10.60 1.4x theor.of len 5.99 4.22 4.79 5.59 6.38 17.41 These results which are plotted in Fig. 3, are in accordance with this view. The initial reaction rate in each case (or the slopes when t= 0) should be some- what proportional to the concentration of the acid. This may be seen to be roughly true by an examination of the CUPVES. Afterward, the slope for the 6-normal solution steadily decreases, while the other slopes remain approx- imately constant, showing that only the ionization con- stant for the former is sufficiently small to retard the solution. In general, for a short time of treatment, (up to about one week) a better yield is obtained, at room temperature, by using a large amount of concentrated acid. For times of treatment between one and three weeks, it seems that 3-normal acid is the best to use. In ail cases, judging from the curves, it is undesirable to use dilute acid at room temperature. % of total iron aisso/ved. 12 /0 /00 90 80 70 60 30 40 — 20 — Solubility of Pyrite Agh in Julturie Acid af £0°C. Unr@agucea. i 3 7 10 4 zo Days. 2 Frequced. Bo 42° Co» |4» Austng 84 x theoretical amount of Days. bia 266 The effect of the time of acid treatment. For conditicns at room temperature this may be shown by the curves from the preceding section. From them it may be seen that if the more concentrated acid is to be used it would perhaps be more economical to treat the ash for two short time intervals than for one long one (assum ing that it is not due to a surface effect). For the other conditions of acidity there would be, apparently, no difference. The results for the lenormal solution may be ex- pressed by the empirical equation, % dissolved = 3.82 + 0.17% x time in days. Although this can only yield approximate resuits when 6x~ trapolated beyond the experimentel data, it is probable that the time for complete solution, so calculated, will be the minimun value. This is found to be 566 days. The similar equation for the 3-normal solution % dissolved = 3.25 + 0.35 x time in days, which gives 276 days as the time for complete solution. The results for the 6-normal soluticn are not sufficient to form the basis for either an empirical equation or an estimate of the time for totel solution. A set cf experiments carried out with 5.6 times 27. the theoretical quantity of 35.0 normal acid at 250° C. gave the following results: Tine of treatment. % Dissolved. 5 minutes 68.9% These results also show that the reaction rate decreases markedly with the time, just aa noticed above. It is probably due to the small ionization constant of the acid and the high sulfate ion concentration. Summary . The above experiments on the solu- bility of the original material in sulfuric acid show, that in order to obtain any appreciable solution it is necessary to use fairly concentrated acid, (20-normal or above), at a temperature close to its boiling point. There is no benefit to be derived from treating the acid for a long time at this temperature, in fact it would seem better to treat it for two shorter time intervals with fresh portions of acid. The total quantity of acid need not be high, probably not over the theoretical amount. 28. Solubility of the Reduced Material. The effect of the reduction. Sinee the solu~ tion of the original material in sulfuric acid was found so difficult, a few preliminary experiments were made to determine whether the solubility could be increased by first reducing the material to either magnetic oxide, ferrous oxide, or iron, by heating with finely powdered wood charcoal as suggested by the Pezzolato process. The results. quoted below are grouped in sets, in each of which, usually, all of the experimental con- ditions, except one, are kept uniform. Any change in the solubility of the ash will, therefore, be due to the effect of this variable. To avoid unnecessary repeti- tion, these experimental conditions and results are col- lected in Table II, to which frequent reference will be made. By comparing runs 14 and 15, it is seen that, where the subsequent acid treatment is for 17 hours at room temperature, reducing the ash increases the amount dissolved from 5.95% to 36.3%. By comparing 11 with 17 it is seen that, where the acid treatment is at 105° C. for one hour, reduction increases the solubility from 44.4% to 95.8%. These results show thet the solubility is markedly inereased by first reducing the material. 29. The effect of the method of reduction was next studied. Both porcelain and iron crucibles heated to a bright red heat by a bunsen burner and a fireclay cruci- ble heated to about 1050° CG. in a crucible furnace were used. All of the crucibles had covers of the same ma~ terial. After the reduction the mixtures were treated with 5.6 times the theoretical amount of 6-normal acid for the time given. Method. Time of Time of Acid @ Dissolved. Reduction. treatment. 18. Iron Crucible 1 hour 1 hour 80.6% 19. Porcelain " 1 hour 1/2 hour 73.8 20. " " l hour | 2 hours 73.9 23. Fireclay " 1/2 hour Sample entirely reduced and fused in Crucible. It was noticed that the iron crucible had a tendency to scalis off during the heating. This was probably the cause of the high result in comparison with the mean of the two values for the porcelain crucible. For this reason those values quoted below which were ob- tained by reduction in iron crucibles may be regarded as somewhat unreliable. it was seen that although the mix- ture was heated for only one-half the time in the cruci- ble furnace the reduction was much more complete. The effect of various reduction mixtures on 50. the solubility was shown by the following set of experi- ments: Ratio of pyrite ash % Dissolved. to carbon. 24. 2:1 65.3% 28. 13.531 73.4 25. 20:1 62.9 27. 40:1 57.3 The ratio 13.5:1 is the theoretical mixture re- guired to give ferrous oxide and carbon monoxide as the products (regarding the samples as 100% Fe,0,). The re~ sults show that this is approximately the composition giving the best results under the conditions stated. It could be seen that this reduced material contained some unconsumed charcoal but not in a large amount, The pro- duct was a duli black and the reduction to ferrous oxide appeared to be fairly complete. The low yield for the mixtures with a larger proportion of carbon was undoubtedly due to the incom- plete reduction of the material. These samples were very crisp and contained, apparently, but little uncon- sumed carbon. Each contained a considerable quantity, {estimated at about 5% of the total weight), of a very hard metallic material, either iron or magnetic oxide, present in small particles about 1/32" in diameter. ol. The run made up in the ratio of 2 to 1 yielded a product in which a large amount of unreduced carbon could be seen. The material when cold and puiverized was a dark brownish black showing that the reduction was not entirely complete. The effect of the time of the reduction was studied with the results given below. For a porcelain crucible at bright red heat: Time of reduction. Color of Product. % Dissolved 20. 1 hour black 73.9% 22. 2 nours ~ reddish=black 13.5% Ke 5 hours a reddish-purple x Xx Since the 1 hour sample was black, the reddish- black color of the two hour sample showed that it had evi- dently been reduced and later partially reoxidized. The 3 hour sample was not analyzed as it was apparently com- pletely reoxidized to ferric oxide. | Reduction in a fireclay crucible at about 1050° Cot Time of Reduction. % Dissolved. 23 & 29. 1/2 hour 55.4% 26. 1/4 hour 78.9% The sample for the 1/2 hour treatment was com- pletely fused in the bottom of the erucible. On the See assumption that it was principally metallic iron, an analysis was made on an equivalent weight of cleaned and polished iron wire. It had a diameter of about 1/32 inch and was cut in pieces about 3/4 inch long. It is probable that the lower result obtained with it is due to the larger particles used and not wholly to the change in the time of the reduction. Since the iron was quite pure the decreased solubility might also be due to its "“pbassivity." While the above experiment gives inconclusive results for the relative solubilities it does at least snow that the time of the reduction should be less than 1/2 hour. Additional experiments should have been made at shorter time intervals. However, even if this had been done it would throw no light on the best conditions for large scale operation, for the efficiency of the re- duction is so dependent upon the character and the tem- perature of the furnace that what might be best for a small seale laboratory process would probably not be at all suitable for commercial practice. the effect of the amount of acid on the solu- bility of the reduced material is shown by a set of ex- periments with the following results: SS Avnount of acid. % Dissolved 28. 1.4 times theoretical 73.4% 26. 4.2 " " 78.9% 30. 8.4 " ft | 52.3% On this basis, the maximum solubility is ob- tained with about 4 times the theoretical amount of acid. However, since the amount dissolved by about the theo- retical quantity is so little lower, it should be pref- erred for economy and to prevent the evolution of hydro- gen when the ferric sulfate is reduced. A study of the effect of the concentration of the acid yielded the following results: . Concentration. % Dissolved. 4%/Concentration. 31. 3-normal 37.9% — 12.6 26. 6=normal 78.9% 13.1 It is seen that the pere-cent dissolved is greatly increased by increasing the concentration of the acid and that for treatment at room temperatures it is roughly proportional to the concentration. The normal- ity of the acid which may be efficiently used is limited by the volume of the solid material which is to be dis- solved. The smallest convenient amount of acid for 1.25 grams of the reduced material is about 5 ec. If only the theoretical amount of acid is to be used, this limits the D4 concentration to about 6-normal. The effect of the temperature of the acid treatment is shown rather poorly by the two results fol- lowing s Temperature. Time of treatment. % Dissolved. 21. 20° ¢. 2 hours TE 0% v7. 105° Cc. 1 hour 95.8% It is seen that, for 10-normal acic, despite the doubled time of treatment at room temperature, the other at 105° CG. dissolves about one-third more. An inerease of the temperature of the treat- ment with 6-normal acid from 20° to 90° G. is approxi- mately equivalent to increasing the time of the treatment from:1/2 hour to 24 hours as shown by the following data: Time of Temp. of Time of Temp. of Amt.of % Dissolved reduction reduction acid acid acid treatment 19. l how ved heat 1/2 hour 90°. 5.6x 73.8% theoretical 24. 1/4 hour at 1050°C. 24 hours 20°C. 4.2x 65.3% theoretical It is seen that raising the temperature of the acid greatly accelerates the solution of the pyrite ash, but it should be noted that varying the concentration of the acid changes both the temperature and the time required for a given quantity to dissolve. 5D « The effect of the time of the acid treatment on the amount dissolved is shown by the following re- sults at S0° C.: Time in hours. % Dissolved. 19. 1/2 hour 73.8% 20. 2 hours 73.9% This shows that the curve is very filet which is the same effect noticed above in the case of the un- reduced materiel. For room temperatures the effect cf the time of the acid treatment on the amount dissolved is shown by the following table. “the reduction mixture ratio was 15.35 to 1 and the time of the reduction 1/4 hour in the crucible furnace. Time indays: 1 3 ” 10 1422 Acid used: | 8.4 times theoretical of 6-normal 52.3% 70.3% 87.3% 87.4% 91.2% 99.6% 4.2 " theoretical of a S-normal 37.9 53.3 68.5 79.8 85.4 94.7 1.4 " theoretical of These results which are plotted in Fig. 3, are rather peculiar in that the curves for both the 6- and S- normal soluticns show a decreasing rate with respect to time while that for the l-nervmal is constant. This 36. verifies. the assumption used in the previous discussion, that the reaction in a concentrated acid solution slows down as the concentration of the sulfate ion increases because the ionization of the acid is depressed. Appar~ - entiy the ionization constant of the S-normal solution is sufficiently small to cause the rate to decrease where the solution is fairly concentrated in ferric sul- fate, as in this case, but where the sulfate ion concen- tration is smell, as was the case with the unreduced material, the rate remains uniform. Although these re- sults cannot be directly compared with those quoted im- mediately above, which were obtained at 90°, yet, & pos- sible reason for the flat portion of the curve occurring at a lower per-cent in the latter case may be due to the fact that, since the reaction is exothermic, the equilib- rium is displaced as the temperature rises. By extrapolation, which may be subject to con- siderable error, the minimum time necessary for complete solution may be determined in each case: 6-normal el days - 3-normal 29 days Llenormal 65 days the effect of changing both the concentration and the amount of the acid is also shown in the above 37. table. While specific conclusions cannot be drawn it may be seen that in general a large amount of concentra-~ ted acid is considerably more efficient for short time intervals than a small quantity of dilute acid. This shows that during the first period, even though the sol- ution of the material mey be somewhat retarded by the large sulfate ion concentration the amount the hydrogen ion is increased more than compensates for this. Since the cause of the retarding of the solu- tion does not seem to be wholly a surface phenomenon, it would apparently be also better in this case to treat the material for two short time intervals rather than for one long one. Summary . These experiments on the solubility .of the reduced material lead to the following conclusions: li. To secure rapid and complete solution the material must be reduced to the ferrous state. ee The best reduction mixture for treatment at 1050° C. is one part of carbon to approximately 13 parts by weight of pyrite ash. Oe The time of the reduction giving the best results varies with the method used. 4. The theoretical amount of acid should be used. 58. 5e The concentration should be as high as possible but it. cannot be greater than ebout 6-normal for practical reasons. 6. Raising the temperature greatly increases the initial rate of solution. 7. j%$rIt is better to heat for two short time intervals rather than one long one. 39 6 Thermo-Chemistry of the Acid Treatment The theoretical volume of 6-normal acid requir- ed to dissolve 1 kg. of the pyrite ash as ferric sulfate is 4.56 liters. The weight of the resulting solution is 5.635 kee From above, the heat liberated by the solu- tion of 1 mol. of ferric oxide to ferric sulfate is 33,550 cal. Therefore, assuming the specific heat of the solution as .8, the amount the temperature will raise, if there is no loss by conduction, will be 32°5, The ae- tual heat loss will depend on the physical characteristics of the treating tank and the time of treatment, and cannot be determined until these are known. Distillation and Crystallization After all of the iron has been gotten into sol~ ution the next step is the reduction of the ferris sul- fate to ferrous sulfate by digestion with scrap iron. The solubility of ferrous sulfate in water is shown in Fig. 4. The following results are calculated on the as-~ sumption that the solution is evaporated at 55° (the point of maximum solubility) until saturated, and crystallized at 20° Ce It is further assumed that all of the dis- solved iron is present in the ferric state. Ferrous uh yare s Joubiiy 6 uve. boos ry Yr Temperature “Co Initial charge of pyrite ash . Volune of G«normal acid used (1 times theoretical) Weight of scrap iron necessary ‘ater to evaporate before first erystallization Weight of Fe SO, 7H50 recovered in first crop % of total Fe SO, 7H,0 recovered in first crop Water to evaporate before second crystallization Weight of Fe 80,7 O recovered in second crop % of total Fe SO, Y7H,0 recovered in first two crops Water to evaporate before third crystallization Weight of Fe SO, '7H,0 recovered in third crop % of total Fe SO, 7H,0 recovered in first three crops Total yield for 1 kg. pyrite ash 41. _ 1.000 kg. 4.36 liters 0.246 ke. None 2249 ke. 67.6% 1.25 ke. 0.75 kg. 88.1% 0.453 keg. 0.28 keg. 95 66% 3.52 kg. 42. The heating with en excess of scrap iron, over the theoretical amount recommended above, should occur in an open tank, with the solution at 55° Cc. ‘The time necessary for the complete conversion must be determined by experiment. As soon as the reduction is complete, the solution should be at once transferred to a crystal- lizing tank and cooled tc 20° CG. It may be found ne~ cessary to add sane excess water to the liquor in the reduction pan to take up the ferrous sulfate which will crystallize out, but this should be avoided if possible. The subsequent concentrations should occur in a vacuum evaporator. The crop from the first crystallization should be quite pure but the others, especially the last, may contain small amounts of aluminum and zine sulfetes. Large Scale Laboratory Crystallization This portion of the work, while it yielded no important quantitative results, showed that 1. A more efficient vacuum than can be readi- ly obtained commercially is necessary in orcer to evapor- ate the solution at 55° c, It was found that toward the end of the concentration the liquor boiled at 75° under a vacuum of £7 inches. However, since the solution con- 43. _ tained an unknown. amount of free acid during the concen- tration, not too much reliance should be placed on this result. Landolt-Bornstein (23) quotes results which show that at 46°92 a saturated solution of ferrous sul- fate has a vapor pressure only 5.5 mm. lower than pure water at that temperature. 2. It is very advisable to use only the theo- retical amount of acid to eliminate the undesirable evo- lution of hydrogen during the reduction. 3. Experiments should be made on a large scale in order to minimize as far as possible mechanical losses. 4. The recovery of the copper from the excess serap iron will be difficult because of the small amount present. 5. Further experiments are necessary in order to determine the quality of the product obtained as time was not available to analyze eitker the crystels or the mother liquor. 44, Outline of a Suggested Plant Method A commercial method for the preparation of ferrous sulfate from pyrite ash might be as follows: The material should be pulverized until approx- imately 95% passes through a 70 mesh screen. Since none of the pieces are larger than 3/8 inch in diameter it may be run directly through any of the dry pulverizers of which probably some form of the steel ball mill would be most suitable. A smaller and intermittently operated mill could be used to pulverize the charcoal. It should be | carried to about the same degree of fineness recommended above for the ash. | The charge of the pulverized material and crushed charcoal cm be mixed in smali bins with a shov- el as needed. It should be made up in the proportion of about 7-8 pounds charcoal per 100 pounds of pyrite ash. Some form of a muffle furnace must be used for the reduction, as the material is so finely ground that if it were exposed to the gases as would be necessary in a reverberatory or kiin furnace, part would be blown away. It would be best to use two furnaces heated by the same fire-box so that while the charge in one is being reduced the other may be emptied and recharged. It would prob- 45. ably be best to spread the charge only six or eight in- ches thick but this, as well as the time and temperature of the reduction, must be empirically determined by ac- tual runs with the furnace. The material should be dissolved in the theo- retical amount of 6-normal sulfuric acid at 100° ¢. The reaction should be complete in about a hour and a half. Tt would probably not be necessary to stir the mixture more than a few times, although this would de- pend entirely on the uniformity of the grinding. If there is a large amount of both large and small material it may pack. A large, hemispherical bottomed, enameled- steel kettle would probably be the best to use. It should be steam jacketed and have no cover. The solu- ‘tion and the undissolved material should be drained off, through a large outlet in the bottom, to a tank on the floor below. The thermal effect attending the reaction has been discussed above. The settling tank, to which the mixture is transferred from the reaction pan, should be made of wood and lined with sheet lead. After settling the solution is decanted through a Duriron pump and pipe- line to a steam-jacketsd kettle, similar to the one used 46. in the preceding step. The residue may be washed once or twice by decantation and then pumped out as waste. In the second treatment in the open kettle the solution is digested at about 55° Cc. with ean excess of serap iron. When the reduction is complete the solu- tion is dropped into the crystallizing tanks. These may be made either of wood lined with lead, monel metal, or enameled steel. Pieces of scrap- iron should be suspended in some manner in the solution. Yhen the crystallization is complete the mother liquor is drained off through a hole in the lower end and sent to the evaporators. These should be single effect high-vacuun pans. The solution should be completely reduced at this stage but to provide against any possible solution due to reoxidation it would be best to use enameled steel. The solution is evaporated at as near 55° CG. ag possible, If the boiling point is considerably higher it should be concentrated until the water is just suffi- eient to form a saturated solution at 55° c. When this is complete, the vacuum is shut off and the solution di- gested with stirring at that temperature until all of the ferrous sulfate has re-dissolved. It is then crys-~ tallized as described above. aT. The erystals are removed from the crystalliz- ing pan and dried first with a centrifuge and then by stacking in a warm room. Suggested Future Work. This work which had for its purpose the deter- mination of a commercial method for the preparation of ferrous sulfate from pyrite ash has been almost entirely confined to a laboratory study of the conditions influ- encing the solution of the material in sulfuric acid. Because of the lack of time very little work was done on the crystallization of Fe SO, 7H50 from a given dilute solution of ferric sulfate. Further experiments should be made to deter- mine the following: i. The solubility of various materials such as Duriron, monel metal, etc. in solutions of ferric sul- fate. 2. The conditions of temperatures, pressure, and time for carrying out the concentration. 3. The most efficient method for the conver- sion of ferric to ferrous sulfate with regard to the amount and kind of serap-iron, its position in the crys- tallizing tank, whether or not it is necessary to have 43. the lron present during the evaporation, etc. 4, The amount crystallized as a function of the time of crystallization. 5. The number of crops of erystals which can be obtained. | 6. The purity of the crystals in each crop. 7. The regulation necessary in order to ob- tain the correct size of crystals. 8. The best method of drying the crystals. Finally, in order to put the process on a com~ mercial basis it would be necessary to 9. Work out completely on the best data ob- tainable the approximate manufacturing cost per pound of Finished product. LO. Determine the state of the market for fer- “‘rous sulfate and compare the selling price with the ap- proximate manufacturing cost. il. Locate and design a plant specifying the type, material, size and cost of the equipment required. 12. Determine the approximate income on the basis of the invested capital. 49. Summary A method has been devised for the preparation of ferrous sulfate from pyrite ash. The material is first reduced by carbon and then treated with the theo- retical quantity of 6-normal sulfuric acid at 100° Cc. The evaporation and crystallization of the resulting solution are discussed theoretically. -o0000= Suniary of Results for UNREDUCED SAMPLE 50. Number Or 12 11 Time (min. ) Gi Temp . oC. 250 130° 100 Cone. Acid 35 .Orn 550 _ Amount Acid 2.8xt 56 55.0 sate ee errr ey Oi aa tg re en 35.0 ren nd 5.6: 250 105 60 115 55.0 ae emer aerate geek dent a aT yew » 566 5.6 10.0 aba ome ite 15.0 Ferrou Fe 10.1% 4.84 Total 7 £Ee 72.8 4,84 68.9 44.4 58.2% 8.4x6n 4.56% 7430 7.421 9.24 9.46 10.70 0254 TABLE II. Sumeary of Results for EREDUCEI vooee + 7 1M, Sh et, i eof ow E E + ar a RK we Wt ital — ; i ' : t i 4 , rage fone , 4 2 20.5: 2 e@* 2 : B 265.55; 2 20 13.55] 40 13.35 Netal- i : : I flic Fel Carbon 1 Riobeoa PF Pil 1iol L, od 1/41 1 Lipo x a : é fa 4 * 3 a A t * ‘ ‘* Ff 4 i : 4 fi Time : ; i ; ; . 1 ' Fs 4 a q 5 > Fr oi ” Red. 1 br. fa) a 1 1 3 : 1 : 1 E 2 20.5 {0026 0.425 0.425 0-25 0.25: x 3 : i i : : : j 4 i ‘ 7 beskl? ; 1 x 1 2 2. 2) wy | Re | 24) 24 | 245 24 24 ; ' a . GO + 90 $90 } 20 ; 90; & 3 20 | 20) 20 4 20, 20. BO t i ‘ 7 a 4 : 3 Z ra a : i Oo at . in fo : : i ER i ; iO H - i g i : a ~ a & : 4 5 4 2 > —t £ 3 e} i Hf x x x X gle ow x Xi ox x Xs ox oer: ; O. ei 3 i 86.9 95.8 /80 6 ]75.8 f73.9 [72.0 413.5: 55.3 62.9; 78.9,57,3 73.4:55.4 * x é Z i PI * Porcelain crucible used in the reduction. ae it we Pire-olay tt it " tt “TS TABLE II. (Cont.) Reduction Ratio of 13.35 to 1: TF i Days | 8.4x6n | 4.2x3n | 1.4x1n {4 on «0 ia:) r cA ~ ° <9) SA , 27.7% scien (ines Ue 3 70.3 | 83.3 | 31.7 | 7 $87.3 . 68.5 | 45.6 10 87.4 © 79.8 | 48.0 / 14 (91.2 | 85.4 | 58.8 94.7 , 72.9 (o>) 21 99. | 22 (80-1 | 444 | 36.8 i(Perrous): : 1. oe 3s 5e 6. 7. 16. Mineralo oH ca ® Bibliography. ogy. Kraus and Hunt... . . . Page 209 The Mineral Industry during 1920,Rousch .Page 645 Textbook of Inorganic Chemistry. Friend. Vol. 9, Part 2... .. . « .Page 120 Sulfuric Acid and Alkali. Lunge 1917. Vol. 2, Part 2... .. . . .Page 821 Chemical Abstracts. 1921. 15:13879. Lunge. Lunge. Lunge. Lunge. Vol. 1, Part 3... .... . «Page 1455 Vol.l, Part 3, ... +... + + Page 1456 Also Chemical Abstracts 19135. 7:685. Vol. 1, Part 3, ...... . Page 1456-60 Also Chemical Abdstracts 1921, 15:733, and Utilization of Waste Products, Koller... . . . . . «Page 316 Vol. 1, Part 3, .. . ». « Page 1455 Also Comptes Rend. 1892, 65:51 Chemical Abstracts, 1910, 4:2554 Lunge. Journal. Lunge. Vol. 1, Part 1....... « «Page 445 Society of Chemical Industry i912. . .Page 815 Vol. 1, Part 1... 4.464. . . Page 53 Analytical Chemistry. Treadwell and Hall. 4th Ed. Yol. 2, ... . Page 3557 Analytical Chemistry. Treadwell and Hall. 4th Ed. Vol. 2... . .Page 848 Quantitative Analysis. Mellor... . .Page 617. o4. 17. Treadwell and Hall. Vol. 2... . . « Page 619 18. Journal American Chemical Society 41:13387. 19. A Dictionary of Chemical Solubilities. Comey . .. . .Page 200. 20. Friend. Vol. 9, Part 2, ..... . « «Page 119. el. Landolt - Bornstein. 1912. Physicalische Tabellen. 22. Metallurgical and Chemical Engineering 1912. 10:172. Coe Landolt - Bornstein. 1912. ..... . .Page 425.