Nov., I917
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
SERIESXI11 (Nos. 105--107)--Here 4 cc. of formaldehyde were used with large amounts of sugar. T h e mirrors were good but not sufficient advantage was gained t o warrant t h e use of more formaldehyde.
Series and
No.
XI11 105 106 107 XIV 108 109 110 111 112 113 114 115
EXPERIMENTS WITH DIFFERENT ALCOHOLS 3 Cc. M / 5 AgNOs Solution Used in Each Experiment Molar 0.8% SILVER Alcohol Cane Formal- DEPOSITED Used Sugar dehyde Weight Per ALCOHOL Cc. Cc. Cc. Gram cent 4 Methyl, 807,. . . . . . 1 . 5 1.5 0.0087 1 3 . 6 1 . 0. . 1 . 5 4 ~ 0.0094 14.5 0.0100 15.4 1.5 0.5 0.0189 1 7 . 5 Ethyl ............ 1 . 0 0.0 0.0218 2 0 . 2 1.0 1.0 0.0194 17.9 Propyl.. . . . . , 1.0 0.0 0.0211 19.5 1.0 1.0 Amyl ........ . 1.0 0.0157 14.5 0.0 0.0194 17.9 1.0 1.0 0.0176 1 6 . 3 Ethylene Glycol.. 1 .O 2 0.0 0.0207 1 9 . 1 2 1.0 1.0
. .. . . .. .. .
T!me in Min. 40 40 40 40 120 35 120 33 104 30 94
SERIESX I V (Nos. 108-1 I 5)-Other alcohols were introduced in t h e place of methyl alcohol. Apparently they all work t h e same. It is of interest t o note t h a t here there was also a n increase in deposit where sugar was used in connection with t h e various alcohols. The time of reaction was increased so much that a higher deposit resulted, with a mixture of sugar and alcohol, t h a n with either when used alone with formaldehyde. GEXERAL COKCLCSIONS
I n hot processes cane sugar is t h e most satisfactory reducing agent. Other sugars give good results but no better t h a n cane sugar. The tartrates are not as satisfactory as the sugars. A rapid cold process was developed. It requires a batch of the following proportions used as indicated: 2 0 cc. of 0 . 2 molar silver nitrate solution are mixed with 0.5 cc. of 80 per cent methyl alcohol; t o this 0 . j cc. of 40 per cent formaldehyde is added and t h e whole mixed thoroughly. A slow cold process was developed which is based upon the following principles: ( I ) Low concentration of the reducing agent (formaldehyde) gives t h e best results. This helps lengthen the time of reaction. ( 2 ) Low concentration of silver nitrate gives high deposition percentages. (3) Sugar added t o the solution in sufficient amounts controls the action of t h e formaldehyde in such a way as t o produce even mirrors. It also lengthens the time of action. (4) Alcohol added t o the solution results in a high deposition efficiency. By grouping these principles t h e following batch results : I 6 . 5 cc. of 0 . 0 3 7 molar silver solution I . o cc. of I . ooo molar cane sugar solution 0 . 5 cc. of 80 per cent methyl alcohol 2 . o cc. of 0.8 per cent formaldehyde solution. Total, 2 0 cc. After 40 minutes, over 2 0 per cent of t h e total silver deposits as mirror in each case, and if left longer a heavier mirror results. The cost of materials is not over one cent per square foot, or eleven cents per square meter of surface silvered. The mirror is very firmly attached t o t h e glass for two reasons: (I) The temperature of t h e glass and solution is t h e
I035
same: as a result deposition is more uniform than t h a t which occurs at higher, more difficultly controllable temperatures. ( 2 ) Since deposition takes place a t room temperature, differences in coefficient of expansion do not cause t h e newly formed mirror t o be loosened from t h e glass as may be t h e case where hot processes are employed. UNIVERSITY OF, PITTSBURGH PITTSBURGH, PENNSYLVANIA
THE RECOVERY OF WATER-SOLUBLE POTASH AS A BY-PRODUCT IN THE CEMENT INDUSTRY' By WILLIAM H.
ROSS A N D
ALBERTR . MERZ
When a potash silicate like feldspar is ignited a t a sufficiently high temperature, sintering and finally fusion take place without any change in the chemical composition of the mineral. With the addition of lime, however, an entirely different result is obtained. A calcium aluminium silicate is then formed and t h e potash is set free or volatilized in amount depending on t h e proportion of lime taken, on t h e temperature and t h e time of ignition. I n t h e case of small scale experiments i t has been shown t h a t at a temperature and with a proportion of lime corresponding t o t h a t used in Portland cement manufacture complete volatilization of the potash may be brought about on igniting for about one hour. Since potash silicates occur in t h e raw materials used in cement manufacture i t is t o be expected from these observations t h a t loss of potash will take place in t h e burning of cement. T h a t some potash is lost has long been known, but there seems t o have been no quantitative data on the subject previous t o t h e results obtained by R. K. Meade in 1905 from analyses of the raw material fed into the kiln, the clinkers coming out and the coal used as fuel. The results of three separate tests with t h e material from one plant showed t h a t the potash volatilized varied from 46 t o 5 2 per cent of t h e total present. PRELIMINARY EXPERIMENTS
With a view t o securing more definite information in this direction, an investigation was undertaken over five years ago t o ascertain t h e percentage of potash occurring in t h e raw materials used in t h e different cement plants of this country and also t h e percentage of t h e total potash t h a t is volatilized in each case. Representative samples of raw mix and ground clinker were collected from the different cement plants in the United States and in Canada with a view t o analyzing each sample for potash. K i t h t h e data thus obtained and knowing t h e ratio between the raw mix and t h e cement produced, and t h e output of t h e latter, it was thought possible t o calculate for each plant t h e approximate quantity of potash t h a t escapes daily from the kilns. After partly completing this work i t had t o be abandoned for a time. A few months ago the problem was again taken u p and entirely new samples were collected. The work has now been completed and the results obtained form t h e subject matter of this paper. 1 Presented before the Fertilizer Division at the 55th Meeting of the American Chemical Society, Boston. September 10 to 13. 1917.
1036
T H E J O U R N A L O F 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 Vol. g, No. A N A L Y S E S OF S A M P L E S
The total number of cement mills from which samples were received number 113, of which 104 are in this country and g in Canada. The number of plants in this country which supplied samples is only four less t h a n t h e total number which operated in 1915, or nine less than the total number said to be in operation in 1916. The results obtained in the analyses of the samples show t h a t the potash occurring in the raw mix used by the different plants from which samples were received varies from 0.20 t o 1.16 per cent. T h e raw materials used in the mills of the Universal Portland Cement Company, and consisting of blast furnace slag a n d a high-grade limestone, contain considerably less potash than the average for the other cement plants of the country. I n a communication received from the manager of one of the large cement plants in the Lehigh Valley district, the statement was made t h a t the raw cement material used in this district contains less t h a n 0.3 per cent. We find, however, t h a t in no case does t h e raw mix used in the plants of this district run as low as 0.3 per cent, but t h a t some of these plants make use of cement material almost as rich in potash as may be found in any cement plant in the country. I n general, it may be stated, however, t h a t high potash-bearing raw material is not characteristic of any particular part of the country. Knowing the percentage of potash in the raw mix €or each individual plant the total potash in the daily run of raw mix is readily calculated. I n a corresponding way is obtained the potash in the daily production of cement. The difference between the two values gives the amount of potash volatilized. This is found t o vary in different plants from 0.35 t o 5.34 lbs. of KzO per barrel of cement produced. The average for all the plants, when consideration is tzken of the variation in output of cement of the different plants, amounts to 1.94 lbs. per barrel of cement. Omitting the plants of the Universal Portland Cement Company, the corresponding average for the remainder of the g g plants of this country from which samples were received amounts t o 2.09 lbs. The average for the nine Canadian plants is 1.95 lbs. POTASH ESCAPING F R O M CEMENT PLANTS
If t h e average annual production of Portland cement in this country be now taken in round numbers as go,ooo,ooo bbls., and if the same average conditions be assumed t o prevail in the few plants from which no samples were received as was found on a n average for all other plants, then t h e total potash (KzO) escaping from all the cement plants in the United States as a t present operated will amount t o about 87,000 tons annually. The process which a t present is most generally considered in connection with the recovery of t h e potash t h a t escapes from cement kilns is t h a t of electrical precipitation. With present installations a recovery of approximately gg per cent of the dust has been secured. It happens, however, t h a t the finest portion of the dust escaping from any cement plant
II
contains relatively the highest percentage of potash and a gg per cent recovery of the dust therefore represents a somewhat less efficient recovery of the potash. In the plant of the Security Cement and Lime Company, where the Cottrell process has been installed for collecting the dust, it is estimated t h a t the potash recovered amounts t o go per cent of the total volatilized. If this value be accepted as the efficiency of recovery of the potash in the most economic installations, then the total recoverable potash in the cement plants of this country under present working conditions amounts t o about 78,000 tons annually. The potassium compounds occurring in cement dust may be divided into three groups as follows: (I) Those which are readily soluble in water; ( 2 ) those which are slowly soluble; and (3) those which are insoluble. The insoluble potash represents the combinations occurring in the original silicates of the raw mix which is carried over mechanically in the dust before being subjected t o a sufficiently high temperature to bring about decomposition. The form of combination which is slowly soluble in water is supposed to be due t o a recombination of t h e volatilized potash with the silicates in the dust. I n some plants where coal is used for burning, the extent t o which t h e potash occurs in this “recombined” form may be considerable, while in certain other plants where oil is used for fuel this combination of t h e potash is present in comparatively small amount. Thus, in a n analysis of a sample of dust from the oil-fired plant of the Riverside Portland Cement Company, t h e acid-insoluble potash was found t o amount t o 1.9 per cent of the total, t h e slowly soluble potash t o 6.5 per cent and t h e soluble potash t o 91.6 per cent. The corresponding values obtained for the dust from the plant of the Security Cement and Lime Company, where coal is used for burning, amounted, respectively, t o 4.4,36.0 and 59.6 per cent. I n the case of the dust from the Cementon plant of the Alpha Portland Cement Company, the proportion of slowly soluble potash is still greater and is sufficient t o greatly depreciate the commercial value of the dust as a fertilizer. B U R E A U OF SOILS E X P E R I M E N T S
I n experiments made some time ago in the Bureau of Soils, it was shown t h a t the greater part of the potash in feldspar may be made t o pass into solution by digesting with lime under a steam pressure of I O t o 1 5 atmospheres. I n cement dust, as it escapes from the kilns, the slowly soluble and insoluble potash is already associated with a considerable percentage of free lime, and it was consequently thought t h a t the greater part of these constituents might be recovered in readily soluble form by digesting the dust with steam alone under pressure. This has been found t o be the case, and with dust from the Security and Cementon plants i t has been found possible t o recover in this way, in soluble form and with little increase in pressure, about 95 per cent of the total potash present. I n carrying out these experiments 2 0 g. of the dust were placed in a small autoclave provided with a steam gauge for indicating pressure. Water was
Nov., 1917
T H E J O U R N A L OF I N D U S T R I A L A N D ENGINEERING C H E M I S T R Y
a d d e d to the dust in the autoclave and the latter then heated so as quickly t o give the desired pressure. This pressure was then maintained constant for varying lengths of time in the different experiments. The a m o u n t of water added was just sufficient to give a paste with the dust a t the end of each treatment. Digesting beyond one-haif hour was found to have little effect in increasing the amount of potash made soluble. Thus, after treating a t 5 0 lbs. pressure for one-half hour the water-soluble potash in the Security dust amounted t o 82 per cent of the total and only increased t o 91 per cent in 8 hours. At IOO lbs. pressure 90 per cent of the total became soluble in one-half hour a n d upwards of 95 per cent in 8 hours. With the Cementon dust similar results were obtained. With this dust, though, it was noticed t h a t prolonged treatment seemed to decrease rather t h a n increase the solubility of the potash. Thus, a t 50 lbs. pressure, 81 per cent of the total potash present became soluble in one-half hour, 87 per cent in 4 hours and 8 2 per cent in 16 hours. At I O O lbs. pressure a maximum solubility of 93 per cent was obtained in 2 hours. Experiments a t I j o and 2 0 0 lbs. pressure, which are expected t o yield still higher results, are now in progress. RECOVERABLE POTASH
This method of treating cement dust should be applicabje whether it is desired simply t o increase the soluble potash in the dust or to bring about a subsequent separation of the potash salts present. It has not been demonstrated that the method would prove as efficient under commercial condisions as in the laboratory tests, but if such would prove to be the case, then the average percentage of soluble potash t h a t could be obtained in the dust from different plants should amount to from 90 to 95 per cent of the total. Assuming then t h a t it would be practical t o recover or render available, say, 90 per cent of the potash in t h e recoverable dust of cement plants, then the recoverable a n d available potash escaping from t h e cement plants of this country amounts approximately t o 70,000 tons annually. If, however, t h e dust from all plants losing less t h a n I lb. of potash per barrel of cement be considered as too low-grade for use as a source of fertilizer material and is omitted in t h e calculations, then the value for t h e available recoverable potash in the cement industry of this country will be in t h e neighborhood of about 67,000 tons annually. This estimate is based on the amount of potash lost under present working conditions. It is possible, however, t o bring about an increase in t h e amount lost by two different ways: ( I ) By increasing the percentage of potash in the raw mix. ( 2 ) By increasing by chemical or other means, the percentage of potash volatilized. At the plant of the Riverside Portland Cement Company, the potash in the raw mix has been increased from 0 . 2 per cent t o as much as 0.8 per cent. This was done by utilizing t h a t portion of the clay deposit highest in potash a n d by enriching t h e raw mix with the addition of feldspar. T h e proportion
I037
of silica t o alumina in feldspar is, however, higher than in the argillaceous material ordinarily used in cement manufacture, and an extended use of this mineral by itself would therefore not be practical in all plants, particularly in those plants where a siliceous cement is already obtained. By use of the proper proportion of iron ore, however, in connection with the feldspar, the latter may then be used even in plants where siliceous cement is ordinarily produced, while in t h e case of plants producing aluminiferous cement a certain proportion of feldspar should be possible without the addition of any oxide of iron. At the plant of t h e Security Cement and Lime Company, a n increase in t h e potash volatilized has been brought about by use of ordinary salt. This is effected by adding t h e salt in t h e solid state both to t h e raw mix and t o the coal used for fuel. The quality of the resulting cement is in no sense impaired by the addition" of the salt for i t is found t h a t the total alkalies in the clinker when salt is added are actually less t h a n when no salt is added. By this treatment the percentage of volatilized potash which a t this plant normally runs about 38 per cent is increased t o about 6 j per cent. I N C R E A S I N G P R O P O R T I O N OF P O T A S H V O L A T I L I Z E D
Other reagents may also be used for increasing the amount of potash carried over in the dust. Thus a n ingenious method devised by the Riverside Portland Cement Company makes use of calcium fluoride for this purpose. Volatilization of the potash then takes place largely in the f o r m of the fluoride and as i t escapes from the kilns i t becomes associated with the salts of calcium carried over mechanically in the dust. It thus happens t h a t when the dust is leached with hot water for the separation of the potash the latter passes into solution as the sulfate or carbonate while the fluoride remains as insoluble calcium fluoride. On using the residue for cement material the same fluoride is fed into the furnace again and may thus be repeatedly used in bringing about increased volatilization of the potash. By this cyclic process the volatilization a t this plant has been increased from 60 to 90 per cent. An increase in the proportion of potash volatilized may also be brought about by increasing the temperature of burning. This is illustrated in t h e analysis of a number of samples furnished us by the Tidewater Portland Cement Company and which represented variations in burning from t h a t which gave a very soft clinker t o t h a t sufficient t o give a very hard clinker. The results showed t h a t t h e proportion of potash volatilized in the different steps varied from 39 t o 6 2 per cent. According t o the analysis of the representative samples sent us from the different plants of the country the proportion of potash volatilized in the different plants varies from about 2 j t o 96 per cent. This variation is too great t o be attributed to differences in the temperature of burning alone, a n d it is probable t h a t a number of factors operate t o bring about this result. The average for dry process plants amounts t o 47 per cent. a n d for wet process plants to 46 per cent. Both averages are above the value for the vola-
1038
T H E J O U R N A L O F 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 Vol. 9, No.
tilization of potash t h a t normally prevailed a t t h e plant of t h e Security Cement and Lime Company, but below the 6j per cent volatilization t h a t has now been secured a t this plant. If it be assumed t h a t t h e last-mentioned percentage volatilization could be attained by chemical or other means in all plants where a lower volatilization normally prevails, then it may be estimated t h a t t h e recoverable available potash in the cement plants of this country would be increased from 67,000 tons t o about I O O , O O O tons annually. This estimate represents a recovery according t o our analyses of only between jo a n d jj per cent of the total potash entering the kilns in the different plants. At the plant of the Security Cement a n d Lime Company, where no steps are yet taken t o make soluble the slowly soluble potash, the percentage of available potash recovered is somewhat less t h a n this, but a t t h e plant of t h e Riverside Portland Cement Company the recovery of soluble potash is already considerably greater. Improvements over present methods of recovering potash are t o be expected a n d it would therefore seem t h a t an estimated possible average recovery of about 50 per cent should not be considered excessive. Whether or not this percentage can be economically recovered in all plants remains t o be determined. I t may also be pointed out t h a t the weight t h a t can be attached t o these estimates is also dependent on the representative character of the samples analyzed. Analyses of a number of samples corresponding t o those which we have analyzed, but collected from six t o nine months previous to the time our samples were taken, have also been made by other chemists and particularly by the Western Precipitation Company. I n every case in which the results were compared there was found t o be a n agreement within t h e limits of experimental error. It is, of course, recognized t h a t the analysis of further samples collected from some other plants might not show the same close agreement, but from the observations which have been made it is thought t h a t as a whole the results obtained are representative. C O L L E C T I O N O F C E M E N T FLUE DUST
Cement flue dust is now being collected by the Cottrell process of electrical precipitation a t the plants of t h e Riverside Portland Cement Company and the Security Cement and Lime Company; at several plants of the Universal Portland Cement Company; a n d a t t h e Cementon Plant of the Alpha Portland Cement Company, At t h e Riverside plant the dust is being used for t h e preparation of concentrated potash salts, while a t the Security and Cementon plants it is disposed of directly for use as a fertilizer. T h e dust from t h e Universal plants, being too low grade t o serve as a source of potash, is simply returned t o t h e kilns for manufacture into cement. Installations of the Cottrell process are now being considered b y a number of other cement companies and operations have already been started on t h e installations a t the Coplay, Dexter and Ironton Portland Cement plants, a n d a t the plant of the Buffalo Potash a n d Cement Corporation, where a commercial installation
II
has been erected for t h e recovery of potash from feldspar by use of the latter in the manufacture of cement. Wet processes for precipitating cement dust are now in operation or are being installed for one or more kilns a t the plants of the California Portland Cement Company, the Santa Cruz Portland Cement Company, and the Sandusky Cement Company. SUMMARY
The constituents of t h e dust in all cement plants are much the same but may vary greatly in relative proportion. I n the case of freshly precipitated dust collected in the normal operation of a cement plant the soluble constituents consist for the most part of the sulfates, chlorides, sulfides and polysulfides of potash, soda and lime. Chlorides occur only in small amount. The percentages of sulfides and polysulfides in t h e dust from different plants may vary widely and in the case of dust from some oil- or gas-fired plants may be entirely absent. Free lime is found in all cement dust, and in t h e water extract of all samples analyzed, the basic radicles were found t o be in excess. I t follows, therefore, t h a t the potash in t h e water extract of freshly collected flue dust will be present t o a greater or less extent in t h e form of the hydroxide. As t h e dust becomes carbonated, either by chemical treatment or by exposure t o the air, the hydroxides present will change t o carbonates, and by interaction of the carbonates with calcium sulfate in the dust t h e water-soluble potash will be obtained principally in the form of the sulfate which is the form universally recognized as of most value for use as a fertilizer. BUREAU OF
SOILS
U.S. DEPARTMENT OF AGRICULTURE WASHINGTON, D.
c.
SOLVENT GASOLINE' By C. OLIN NORTH Received July 2, 1917
The sudden jump in the price of gasoline last year was of prime importance to the automobile owner, but if anything it affected the rubber companies more. Because of their high cost, gasoline had replaced t o a great extent all other solvents for use in rubber cements and doughs. A small rubber factory making I 500 tires a day will use approximately 12,000 gals. per month, while a large company will require close t o IO,OOO gals. per day. I n 1 9 1 j a good grade of solvent gasoline could be procured a t 1 5 t o 18 cents per gal., but in 1916 i t jumped to 2 6 to 30 cents per gal. Solvent gasoline is customarily of j o t o 7 2 " BC. gravity, though for certain purposes 80' BC. is required. It must be readily volatile and must be free from kerosene tails. A comparison of its rate of evaporation with those of some common solvents is as follows, when z cc. of each liquid were allowed t o evaporate from a metal surface 3 1 / 2 in. square, under similar conditions: Benzol (90%) ..... 14 min. Solvent Gasoline: Commerclal Naphtha, ,107 min. SOo BC. ... .4min. 7O0 BC... . .8 min. Turpentine. ,142 min. Toluol. . . , , , . .33 min. Distillation tests on representative samples of the three grades of gasoline, v i z . , Motor, Solvent, and Paint & Varnish Makers' Naphtha, resulted as follows: 1 Submitted as partial fulfillment of the requirements for the degree of Chemical Engineer, Carnegie Institute of Technology, Pittsburgh, Pa.
