THE PREPARATION OF FINELY. DWIDED FERRIC OXIDE Thais for the Degree of M. S. MICHTGAN STATE COLLEGE ,William Francis Taffeo, Jr. 1949. fiHiSNi "fl!W”!’“il'i‘imimiii’li”iii!" SUPPLEMENTARY MATERML IN BACK OF BOOK This is to certiig that the thesis entitled THE PRELPJ'JUATIJIJ Oi" FIJELY DIVIDED F'EfiiiIQ UZ‘LIDE presented bl] TilLiam Francis Taffee, Jr. has been acreptvd towards iuHilhm-nt oi the requirements for ‘ ItLASTERS deg“? in CHLgICAL EIGIL‘IEEHIJG K—w S. L — (W? Major liruicswr Date 1" ’ 4n" *f 0-169 In: ”‘1! -_ in! '_“.~.. <— u.-- ‘*.‘-—-— —-—.—-— .hL-‘-_-_.ni FLEELY CXIDE m'm 31'3" I‘ r‘fi‘? ”'I ' PigPAAAleL ‘J DLVLDED FERRIC U) wLLtlam FiAucz A M' A THESl Submitted to the School cf Graduate Studies of Hieaigan State College of Agriculture and Applied Selence in partial fulfillment of the requirements for the degree of EASTER CF SCIENCE Department of Chemical Engineering 191:9 TH ESIS ACKNCMLEDGEMENT The author wishes to express appreciation to Professor C. C. DeWitt for his guidance and assistance and to Professor J. W. Donnell and Mr. W. B. Clippinger. Thanks are due for their valuable cooperation. The equipment and supplies for this investigation were furnished by the Michigan Engineering EXperiment Station under Project No. 87. 97-. 5”“ ”:4 ‘1 "i 6 ) Push fr J . 5' "I" ’md TABLE CF CONTENTS Page InterUCtion OOOOOOOOOOOOOOIOOOOOOOI00.0.0000... 1 History ........................................ 3 Procedure ...................................... 10 Equipment ...................................... 11 Summary of Results ............................. 1h Individual Results ............................. 15 Discussion ..................................... 25 Uses of Ferric Oxide ........................... 3O Suggestions for Further Work ................... 32 Bj-b1iography OOOOOOOOOOOIOOOOOOOOOOOOOOOOOOOOOOO 3h INTRODUCTION The problem of waste recovery has long held the attention of many people. Business men are interested because it means more saleable goods. State and Federal government conservation authorities are interested because of the stream pollution abatement brought about by proper waste treatment. Lastly, the public may be interested if the by-product thus recovered means a new, better or less expensive consumer good. A particular problem of interest to all three groups is exemplified by the spent pickle liquor from steel fabricating mills. This is composed of ferric and ferrous sulfates and a small amount of sulfuric acid. Because of the acidity and dissolved solids the liquor cannot be dumped directly into streams or sewer systems. Here then, the problem is of concern to both pollution control and the businessman. The steel men have, at the insistence of conservation authorities, set about to find satisfactory methods for the disposal of this waste, which in l9h5 amounted to nearly 1.3 billion gallons. As yet no entirely satisfactory method for diaposal or recovery has been found. To be satisfactory the process should provide for the following: 1. Neutralization of the excess acid. 2. Removal of the dissolved solids. 3. Conversion of these recovered solids to a saleable form by an economically feasible process. h. Provide an effluent of nearly neutral character which may flow directly into streams or sewer systems. At present the processes used for the treatment of waste liquor fail to meet all of the above requirements. The method of treating with lime and recovering a building material from the sludge is not economically feasible except in areas when building materials are scarce or particularily eXpensive. Other methods for the recovery of ferrous sulfate and its eventual conversion to ferric oxide, are available. The ferrous sulfate may be sold as such, to a limited market, or it may be converted to ferric oxide by the calcination method of Fireman. A third method which meets all the conditions set forth is the precipita— tion of the ferric hydrogel and the conversion of this hydrogel to ferric oxide by means of heat and pressure. HISTORY This approach to the problem is not altogether new. Rather it is a combination of several ideas preposed many years ago but apparently neglected after their conception. It was recognized by Willians(b) early in 1802 that heat and pressure would bring about the dehydration of ferric hydroxide. By far the largest amount of work appears to have been done on ferrous iron salts. These processes are based on the precipitation of ferrous hydroxide and its oxidation to ferric hydrogel. The ferric hydrogel is then con~ verted to ferric oxide. There is no attempt on the part of this author to discuss in detail the exact mechanism of the conversion of the hydrogel to hematite (anhydrous Fe203) or goethite, Fe203 H20, (ferric oxide monohydrate) but rather to illustrate some of the factors controlling the rate of con- version and the physical properties of the product. Little work appears to have been done before the year 1800 on the pmecipitation methods of preparing ferric oxide pigments. Scheele waiting in 1777 mentions the preparation by precipitation. In 1802 J. L. iiilliamsslé) in England, secured satisfactory precipitates by dissolving ockues and pyrites in nitric acid and adding an excess of ammonia. Lord R055694) at about the same time, prepared polishing rouge by a combination of precipitation and calcination. The pattern set by Rosse has been faithfully followed throughout the £Vea113. Caloination has been the mainstay of the preparation of ferric oxides . The raw materials varied from pyrites through mill scale and pickle liquo1~ tnu;in.all cases the last process was calcination. The caloination processes devised by Fireman and Penniman — Zoph are still used by the Nagnetic Pigment Company. As late as 1919 Marks(ll) recommended the use of calcination, as an alternative in a process which would appear to be the forerunner of the work done regarding the effect of pressure on the oxidation of ferrous iron salt solutions. The idea of using increased pressures to hasten the oxidation of (11) ferrous salt solutions was first patented by Marks. After Marks came (5) DuFair who treated ferrous sulfate solutions with calcium chloride and removed the resulting calcium sulfate. After the removal of the sulfate ion, he was left with a solution of ferrous chloride rather than the ferrous sulfate with which he began. This ferrous chloride solution was then pre- cipitated with calcium carbonate and the mass agitated with air, under pressure to precipitate the iron in its ferric state. DuFair does not mention the fact, but it would appear possible to recover the calcium chloride formed in the precipitation step and recycle it to the first treat- ment step. Muller(12) in 1938 patented a method for preparing ferric oxide by heating ferric hydroxide to "above 100 deg. C. at raised pressure, in the presence of dissolved ferrous salt". Uebler and Muller<20) also patented a process for making red pigments by heating ferric oxide in a closed vessel to "130 deg. C. or more in the presence of an excess of water". In this case the pressure would approach hO psia. . . 1 In the same V91“ J. W- Ayers( ), in 1939, patented a process for pro- dUCirlg red oxide pigments by treating a ferrous salt solution with sodium carbcxnate and heating the resulting slurry to a temperature in the range of 5 50 to 175 deg. C. at pressures above atmospheric. The same Mr..AyersS2) in 1939, sucessfully prepared black oxide pigments by adding an excess of sodium hydroxide to a hot ferrous sulfate solution and heating the result- ing slurry to between 220 deg. and 290 deg. F. under a pressure of hO to 100 psia., at the same time supplying air to the mixture. The black pig- ment resulting is undoubtedly a mixture of ferric and ferrous oxides, and if the air blow and heating were continued for a sufficiently long time, a red pigment would result, the shade of which would depend upon the amount of sodium hydroxide added in excess of the stoichiometric quantity. This same reasoning, of the effect of pH on pigment color, may have occurred to Ayers or his coaworkers, for in l9h0 the C. K. Williams 30‘21) patented in Germany a process specifying the Specific gravity of both the iron salt solution and the alkali to be used and the temperature at which they are to be mixed. The resulting slurry is then heated to 105 to 150 deg. C. at a pressure of 2.8 to 7 atmospheres while passing an oxidizing gas through the solution. However, the black pigment was not as finely divided as it should have been, for a grinding step was necessary. This patent may be considered as an extension of Ayers Canadian patent number 379,225S2) In these two patents by Ayers is the first indication of any recognition of the effect of pH on the precipitation and subsequent conver- sion of the hydroxide gel to ferric oxide, or as Specified by the patents, ferroso-ferric oxide mixtures. Another path pursued by many individuals was the use of a catalyst of some kind to hasten the conversion of the ferrous iron to ferric iron or 'to Speed up the conversion of the hydroxide gel to the oxide. One of the 6 earliest pieces of work recorded in this field was that of Obladen(13) who prepared pigments "ranging from yellow to red" by heating an aqueous sus— pension or a paste of ferric hydroxide gel. The gel was heated, under pressure to "a temperature above 100 deg. C.", using boric acid or the salts thereof as a catalyst. Bruno Uebler(l9) suggested that anhydrous ferric hydroxide be added to an aqueous suspension of ferric hydroxide gel and the mixture heated to above 100 deg. C. under pressure. He states that red pigments may thus be obtained. An I. G. Farbenindustrie A. G. patent in 1928 claims the manufacture of a finely divided ferric oxide from the hydrate, by heating an alkaline solution at a temperature above 100 deg. C. Here again no thought is given to pH. In 1931 Stahl Chemie G.M.B.H.(18) procured the following patent: "A red material of a remarkable coloring power is prepared by heating, under pressure, iron rust or the hydroxides of iron poor in water. The material used must be moistened with a small quantity of solution of ferric salts or hydrochloric or sulfuric acid". Obviously the reference is to the preparation of finely divided ferric oxide. However no control is exercised over the amount of acid or acid salt added. It is presumed, on the basis of the work reported in this investiga- tion, that no effective color control of the product was possible. A Swiss patent, issued to I. Hunyady in 1938(9), discloses a method for the recovery of both iron and aluminum oxides from bauxite. This is accom- Iilished according to the patent, simply by adding ammonium sulfate to the nthture and heating in a closed vessel. 7 Renkwitzslh) in 1938, patented another method for catalyzing the oxi- dation and conversion. His idea involves the use of an iron-oxygen compound of the structure of iron rust, heated under pressure. This work would seem to follow, very closely, the pattern set by Uebler<19). Riskin, Neroslovskaya and PugaohevaSlS) also in 1938, suggested the use of iron shavings in ferrous sulfate or ferrous chloride solutions to catalyze the oxidation which was done under a pressure of 5 atmOSpheres and at a temperature of 120 deg. C. (12) Muller , in December 1939, patented a process for preparing red ferric oxide in which a ferrous chloride solution was treated with calcium hydroxide to precipitate ferrous hydroxide. The ferrous hydroxide was sub- sequently converted to ferric hydroxide by blowing in air. The slurry, still moist with ferrous chloride solution, was heated to 132 deg. C. for one hour in an autoclave. Muller may have had a specific dosage of calcium hydroxide in mind, however, he failed to mention any specific dosage, thus had no pH control. Wurzschmitt and Beuther(22) in 1928 patented a process very similar to that of Muller except that they utilized ferrous sulfate solutions precipi- tated with sodium hydroxide. Sierp(17) in 1938 suggested the treatment of Spent pickle liquor, which has been neutralized and clarified, with either sodium carbonate or ammonia. Theeresulting precipitate is oxidized and the ferric hydrogel placed in an autoclave at a pressure of seven atmOSpheres. If a yellow pigment is desired, Crepaz(3) suggests the oxidation of ferznsus hydroxide hydrogel, with an atmOSphere of oxygen. The oxidation is 8 to be conducted first at "ordinary" temperatures then at the temperature of boiling under 2 to 3 atmOSpheres. The ferrous hydroxide hydrogel may be prepared by precipitating a ferrous salt solution with sodium carbonate, sodium hydroxide or calcium hydroxide. However, again there is no mention of pH control. Upon careful consideration of the work cited a generalized pattern, which all follow, begins to take form. The pattern is this: 1. Precipitation of a ferrous hydroxide hydrogel. 2. Oxidation of tfliS hydrogel to its ferric state. 3. Conversion, with or without a catalyst, of the ferric hydrogel by heat and pressure, to the desired form of ferric hydroxide. This pattern is followed by all the investigators cited. These men showed little interest in reproducing their colors from bath to batch. Perhaps this is one reason why none of these methods, though they are basically sound, are in use today and why there is no record of any pilot plant or full—scale plant work having been done on them. From previous work in this Laboratory there was evidence that pH of the ferrous salt solutions had a profound effect on the velocity of the oxidation reaction. However, in all of the available literature but one ‘Vague reference is made to the pH of the solutions. If pH is important in the oxidation step, why should it not be impor- tant too in the conversion of the iron hydrogel to the finely divided iron. oxifike? Since pH exerts an influence on the conversion product of a ge1(6) in 8£Ldition to controlling the rate and character of precipitation and agglcxneration, it seemed worthwhile to investigate the production of finely diVidEai iron oxide under adequate control conditions. 9 This objective suggested the outline for the work to be accomplished in the present investigation. Since pH had previously been neglected, a study of the effect of pH of the slurry was made with Special reference to: 1. Time required to convert the hydrogel to the oxide. 2. Temperature requirements. 3. Effect on color, everything else being constant. h. Effect on other physical properties (e.g. particle size, ability to mix with oil, drying characteristics, etc.). lO PdOCEDU The procedure followed in these tests has been standardized in order that the results might be compared. The method is as follows: 1. Dissolve an amount of iron salt equivalent to 8.75 pounds of FeCl3 in a 30 gallon crock. 2. Prepare a solution containing approximately 53 more than the calculated stoichiometric amount of haCH. 3. Add the NaOH and iron salt alternately to the mixing tank. After the solutions are in the tank, turn on the mixer and add sufficient water to make ho gallons. Kix for five minutes and test pH. Ad- just pH if necessary by use of base or acid. b. Pump out mixing tank into system and close system. 5. With pump running turn steam into heat exchanger. 6. Take samples every 15 minutes after start of run, using the bot- tom sampling tap. Centrifuge the sample to remove oxide from water suSpension. 7. Smear a thin film of oxide on glass plate and place in oven to dry. 8. After drying, rub the oxide with oil using a spatula and glass plate, to demonstrate color and mixing characteristics. 9. If no color difference is apparent between two samples taken 15 minutes apart, the run may oe considered complete at the time of the earliest sample. The pH determinations were made with a Beckmann pH meter and with Hydrion Ifiaper and are accurate to .25 pH units. ffiilaked lime was substituted in Runs 8 and.2 otherwise the procedure remained iderndcal. ll EQUIPMENT The equipment used in this work is the result of several years of experimentation in this Laboratory. When pumping the ferric hydroxide slurry through a heat exchanger was first attempted, pump failures were common and frequent. Heat exchangers of sufficient area and at the same time possessing a tube size small enough to maintain turbulent flow were unavailable. Previous work of this Laboratory had pointed to the necessity of hav- ing available some means of cooling the bearings and stuffing box of the pump used; all previously used pumps had failed due to hearing seizure and scoring of the shafts in the packing section. After more than a year of correspondence, two manufacturers agreed to furnish a pump which they felt would handle an alkaline or slightly acid ferric hydroxide slurry at temperatures up to 350 deg. F. and under a total pressure of 125 psig. Of these two, the pump manufactured by Dean Bros. Co. of Indianapolis, Indiana was offered at less than half the price of its competitor. This pump is a 1" by 1%" single stage centrifugal, having both water cooled stuffing box and ball bearing case, capable of delivering 25 g.p.m. under a total head of 50 feet. It has performed very satisfactorily for the period of this work, and shows no Sign of wear or other deterioration. The heat exchanger provided a second problem. A commercially avail- able exchanger of large area at low cost was desired. Again the problem ‘Was finally solved by an exchanger made by Bell and Gossett Co. of Morton Grove, Illinois. This exchanger, originally designed for hot water supply 'Work, is of the shell and tube type with four tube passes and U tube con- 12 struction, steam on the shell side. The tube bundles are approximately six feet long and the inlet and outlet connections 1:" diameter. Steam inlet is h" diameter, however in this work steam was supplied through a %" line. The exchanger has served admirably in its present condition and would appear to be the best exchanger yet tried in this service. Referring to the drawing of the equipment one can easily see that it is essentially a device for heating, batchawise, a quantity of liquid by pumping it through a shell and tube heat exchanger with 90 pound steam on the shell side. It becomes immediately evident that the pressure of this system, with all vents closed, will approach that of the steam plus the head developed by the pump. In the case of a concentrated slurry this combination may well reach 130 psig. at the pump discharge and 105 psig. in the tank as indicated in the data for Run number 2. Obviously then the requirements of heat and pressure are easily met in this apparatus since temperatures of 320 deg. F. were recorded. When full 90 psig. steam is available, the maximum is somewhat greater. Temperature control was attained by the use of an air operated valve in the steam line, the air supply to the valve being regulated by a propor— tioning, recorder-controller, supplied by the Foxboro Instrument Company, .Foxboro, Massachusetts. This combination gave excellent service, holding 'the temperature exactly at the predetermined level by throttling the steam s uppl y . The tank in the recirculation system was constructed of l/h inch boiler lilate. The ends were dished and all joints were welded. This tank has a caINacity of 100 gallons with 1% inch screwed inlet and outlet connections at "the top and bottom. In addition 1 inch sampling taps were installed. 13 These taps were welded into the side of the tank, on vertical centers, at distances of 6, 12, 2b and 36 inches from the bottom. A mixing tank, made from a 52 gallon steel drum with the head removed, was used. A 1% inch pipe coupling was welded in the bottom of this drum to allow its connection with the recirculating system through suitable piping. On this mixing tank was mounted an electric driven propellor type ' mixer, Hodel C h, manufactured by the Mixing Equipment Company, Rochester, New York. This mixer performed very satisfactorily3and its stainless steel shaft was not attacked by either the alkaline or the acid solutions in which it was used. HZMELHDWM Au Eaao