Preparation of Potassium Nitrate from Solid Potassium Chloride and

on a method for its preparation from solid potassium chloride and nitrogen peroxide, the cheapest forms of potash and nitric nitrogen. Mehring, Ross, ...
0 downloads 0 Views 511KB Size
Vol. 23, No. 12

INDUSTRIAL AND ENGINEERING CHEMISTRY

1410

Literature Cited ( 1 ) Rerthelot, Comfit. rend., 62, 905 (1866); 63, 788,834 (1867). (2) Davidson, J. IND. END.CHEM.,10, 901 (1918). (3) Davis, I b i d . , Anal. Ed., 1, 61 (1929). ( 4 ) Davis, Crandall, and Higbee, Ibid., Anal. Ed., 3, 108 (1931). (5) Frolich, Simard, and White, IND. ENG.CHEM.,22, 240 (1930).

(6) Hague and Wheeler, J . Chem. SOC.,1923, 378. (7) Htrslam and Chappell, IND.END.CHEW.,17, 402 (1925). (8) Hurd and Spence, J . Am. Chem. Soc., 61,3561 (1929). (9) Ipatiev, Ber., 44,2978, 2987 (1911); 46, 1748 (1913). (IO) Lebeau and Damiens, Ann. chim., 8, 221 (1917). (11) Schneider, Doctor's Thesis, Mass. Inst. Tech., 1931. (12) Thiele, Ann., 308, 337 (1899).

Preparation of Potassium N.itrate from Solid Potassium Chloride and Nitrogen Peroxide' Colin W. Whittaker, Frank 0. Lundstrom, and Albert R. Merz FERTILIZER AND FIXEDNITROGEN INVESTIGATIONS, BUREAUOF CHEMISTRY AND SOILS,DEPARTMENT OF AORICULTURE, WASHINGTON, D. C.

The direct conversion of solid potassium chloride to solid potassium nitrate by the reaction of the chloride with gaseous nitrogen peroxide is suggested for the preparation of potassium nitrate. Thermodynamic calculations indicate that the free energy change is favorable a t ordi-

MONG the more promising salts for use in concentrated fertilizer mixtures is potassium nitrate. This salt contains two fertilizing elements, is much less hygroscopic than other fertilizer nitrates, and has recognized merits as a fertilizer. The present paper describes preliminary work on a method for its preparation from solid potassium chloride and nitrogen peroxide, the cheapest forms of potash and nitric nitrogen. Mehring, Ross, and Merz (17) have shown that, when nitrogen peroxide reacts with a saturated potassium chloride solution, hydrogen chloride is evolved, and a solution containing potassium nitrate and nitric acid is formed. When, however, in cyclic operation, the acid concentration reaches a certain value, nitrosyl chloride is formed and comes off with the hydrogen chloride. Now if nitrogen peroxide could be made to react with solid potassium chloride, there is presented the possibility of converting the solid chloride directly into the solid nitrate and of obtaining all the chlorine as nitrosyl chloride according to the reaction:

nary temperatures and that the reaction is exothermic. Experiments are described in which it was found that the reaction takes place a t ordinary temperatures in the presence of a small amount of water with good yield. Nitrosyl chloride is produced simultaneously.

A

2N02

(9)

+ KCI

(8)

= KNOs

(a)

+ NOCl

(g)

Thermodynamic calculations indicate a free-energy change favorable to the occurrence of the foregoing reaction at ordinary temperatures, and the experiments described later show that in the presence of a little moisture this reaction actually takes place. Thermodynamic Considerations

In the following calculations 4.182 joules have been taken as equivalent to 1 calorie. To obtain the free-energy change of the reaction, KClG)

+ 2NO:k)

= KNOa(,)

+ NOCIk)

the free energy of formation of KCl(,) was first determined from the free energies of Na(.), K(,), KCl(.), and NaCI(,) by means of the free energy of formation of NaCl(.). Using the data presented in International Critical Tables (W), i t is found that AFftcl(.) = -97,361. The free energy of formation of KN03 in infinite dilution by the reaction between KCI(.) and NOz(,) is now easily determined. 1 Received

July 2, 1931.

There remains to be determined the free-energy change in bringing KNOJ in infinite dilution to the solid state or (since the activity of a solute in its saturated solution equals that of the solid solute) to saturation. The activity of potassium nitrate at the eutectic point (270.1 K.) is obtained from freezing-point data by means of the formula (16):

inwhichj = 1 where

Y

-2 UX m

= number of ions formed

X = molal freezing-point lowering at infinite dilution by non-dissociating molecule ( = 1.858)

and

= freezing-point lowering rn = molality a = activity of solute

The value of the f i s t integral of Equation 8 up to m is given by -jdlogrn

=

= 0.01

- 2.303a -(0.01)"

where a and p = 0.565 and 0.427, respectively (15).

and from m = 0.01 to m = 1.247 (the molality a t the eutectic point) by graphical integration. The value of the second integral is obtained by graphical integration also. By substitution of the value of a at saturation (m = 1.247), from Table I, in the equation A F = RT In a ( 1 3 ) it is found that Kf

+ NOS-

= KNOa(,); AFar0.1

-912

(94

INDUSTRIAL A N D ENGINEERING CHEMISTRY

December, 1931

On the assumption that A H for this change of state is constant [ -8678 calories (5) ] over the small range of temperature, integration of the equation

gives AF2w.1 = -107. On the assumption that the molal heat capacities are constant [ - 17.20 (18) and 21.64 (3) I, solution of the equation AF

I

AH,, - ACpT In T

1411

used, it was found that the salt contained only 1.7 per cent of potassium nitrate on the assumption that all the nitrogen was present as nitrate. I n a second experiment the anhydrous salt was replaced by potassium chloride that contained 2.4 per cent of moisture. About 20 cc. of liquid nitrogen peroxide was vaporized and allowed t o pass through the potassium chloride in the course of 3 hours. After the tower had been flushed out with oxygen to remove the nitrogen-containing gases, the salt that remained in the tower had, after drying at 120' C., a nitrogen content corresponding to 93.23 per cent of potassium nitrate.

+ r T (12) .

gives AF2es.l = -77. The true value lies somewhere between these two values. Assuming it to be midway ( l l ) , K+

+ NOS- = KNO*(.);

AF2sa.l

-92

(9b)

whence, by combination of (7) with (9b) KCl(.,)

+ 2NO*(,) = KNOa(8)

Since N ~ 0 4= 2N02;

Also, KCI(m)

AF298.1

NOCl(,); AFros.1 = -5623 = 1127 ( 9 )

+ NzO4(g) = KNOs(a) + NOCl(g); AFzg8.1 = -4496

(10)

(11) (12)

The free-energy differences of Equations 10 and 12, both being negative values of several thousand calories, indicate that potassium chloride w ill react with nitrogen peroxide a t ordinary temperatures to form potassium nitrate and nitrosyl chloride. Furthermore, calculations of the heats of reaction, using the data of the International Critical Tables (4), give values of AHB1.l as -16,691 and -3683 calories, respectively, for Equations 10 and 12; so the reaction should be exothermic. Experimental

The general procedure consisted in passing nitrogen peroxide gas through a tower packed with potassium chloride crystals and determining the nature of the products of the reaction. The nitrogen peroxide, prepared by the oxidation of ammonia with oxygen, was kindly supplied by J. Y. Yee of this laboratory. It contained some nitric oxide and probably also nitrous and nitric acids, and was purified by triple distillation. I n each distillation, oxygen was bubbled through the liquid nitrogen peroxide and passed with the gaseous peroxide through a phosphorus pentoxide drying tube and then through the condenser where the nitrogen peroxide was condensed. The potassium chloride and other reagents used were c. P. Effect of Moisture on t h e Reaction

A tube, containing the purified liquid nitrogen peroxide, was sealed to the tower containing potassium chloride crystals that had been dried a t 120' C., and a trap for condensing the effluent gas was sealed to the outlet of the tower. Dry oxygen was then passed through the reaction tower and the trap for about 3 hours to remove any moisture that might have gotten into the apparatus when the seals were made. Twenty-five cubic centimeters of liquid nitrogen peroxide (a large excess) were allowed to evaporate through the potassium chloride tower over a period of 24 hours. The reaction tower was then flushed out with dry oxygen to remove any nitrogen-containing gases, and the nitrogen content of the salt therein was determined by the Devarda alloy method. AIthough a large excess of nitrogen peroxide had been

Figure 1-Apparatus for D e t e r m i n i n g N a t u r e of G a s e o u s Reaction Products

Thermodynamical calculations indicate that the formation of potassium nitrite from potassium chloride and nitrogen peroxide is very unlikely. Moreover, tests of the salt produced by the reaction showed no nitrite to be present, although good tests were obtained for nitrate. The foregoing experiments showed that, although dry potassium chloride underwent but little change through the action of nitrogen peroxide, even after exposure of 24 hours, nearly complete conversion to potassium nitrate was accomplished in the mesence of a small amount of water within 3 hours. As indkated later, the reaction with the moist salt was quite rapid. To determine the nature of the gaseous products of the reaction, the procedure was R changed so that an excess of potassium chloride was used, and the gases escaping from the potassium chloride were sampled and e x a m i n e d . The apparatus used is illustrated in Figure 1. Liquid nitrogen peroxide, contained in a tube t o which was sealed the female part of the ground joint, J , was introduced into the generator, G, by completing the joint a t J and opening the stopcock on the generator and another on the nitrogen peroxide tube. The p e r o x i d e was boiled off by means of the n i c h r o m e heating element, H , and passed through the tower, T , packed with slightly moist potassium chloride crystals. The gases issuing from T passed through the bulb, B , and then through a trap immersed in carbon dioxide snow (not shown) where the nitrosyl chloride was condensed. When the gas flow had continued long enough to completely flush Figure 2-Meltlngout B and to establish a steady state through- Point Apparatus out the entire apparatus, B was isolated and the gases allowed to pass through the by-pass connections shown, for a time sufficient t o collect about 50 cc. of liquid nitrosyl chloride.

The course of the reaction upward through the tower can be followed visually by means of the reddish appearance of the salt below the advancing zone of reaction, due to the nitrogen peroxide in the interstices, and the straw-colored aspect above this zone, due to the nitrosyl chloride produced. The line of demarcation is quite distinct. The exothermic nature of the reaction is evidenced by a decided increase in the tempera-

INDUSTRIAL AND ENGINEERING CHEMISTRY

1412

ture of the tower wall a t the reaction zone. The small amount of water necessary for the reaction suggests that its function may be catalytic, perhaps by taking part in an intermediate reaction. The contents of B were sampled by attaching a weighed and evacuated 500-cc. gas balloon a t C. These operations were carried out before the zone of reaction had reached the top of the tower so that there would be no obvious contamination of the gaseous products with nitrogen peroxide. The gases issuing from the potassium chloride had the characteristic straw or amber color of nitrosyl chloride (when looking through a depth of about 1cm. or less) and solidified to a light blood-red solid which melted immediately on removal from the carbon dioxide snow to a cherry-red liquid. When a portion of this condensate was allowed to evaporate, it disappeared entirely before the ice, which collected on the exterior of the trap, began to melt; this would not have been the case had any appreciable amount of nitrogen peroxide been present. The melting point of the nitrosyl chloride was determined, using the apparatus shown in Figure 2. Ten cubic centimeters of nitrosyl chloride were introduced into the glass cylinder, C,of about 15 cc. capacity, equipped with a thermocouple well, W, of thin-walled tubing, and a glass stirrer, S,consisting of a ring sealed t o one end of a rod that extended out through the side tube, T. A flexible rubber sleeve, R, permitted the operation of the stirrer without admitting air or moisture to the apparatus. The apparatus was cooled by immersion in an acetone-carbon dioxide snow bath contained in a Dewar flask. The temperature of the bath was adjusted by adding small pieces of carbon dioxide snow. The temperature changes were followed with a carefully calibrated copper-constantan thermocouple of No. 30 wire, inserted in the well, W. The cold junction was maintained a t 0' C. by a n ice-water bath and the e. m. f . was read with a Leeds and Northrup Type K potentiometer and galvanometer. The thermocouple was calibrated by means of the boiling point of ammonia, and the subliming temperature of carbon dioxide snow, both corrected for the barometric pressure. The ice point was taken for the third point. Table I-Activity

0.01 0.02 0.05 0.10 0.20 0.30 0.50 1.00 1.247

2.303

1U) 0.03590 0.07074 0.17155 0.3314 0.6308 0.9144 1.441 2.56 3.005

3.590 3.537 3.431 3.314 3.154 3.048 2.882 2.56 2.41

0.0339 0.0482 0.0767 0.1082 0.1512 0.1798 0.2344 0.3111 0.3515

the gas contained 95.7 per cent nitrosyl chloride by weight. The remaining 4.3 per cent may be accounted for by permanent gases (chiefly air) in the sample. The chlorine determinations, when calculated to nitrosyl chloride, accounted for 97.3 per cent of the titratable gas in the sample when the latter was calculated to nitrosyl chloride. The nitrogen determinations, calculated in the same way, accounted for 97.4 per cent. The agreement was so good as to leave little doubt that the gas was principally nitrosyl chloride. The foregoing data are representative of those from several experiments, only one of which is given because of small deviations in procedure which rendered the others not strictly comparable. The other experiments, however, bear out in every way the conclusions reached in this paper. Velocity of the Reaction

The decided increase in temperature of the tower a t the reaction zone is evidence of a rapid reaction, since otherwise the heat of reaction would have been conducted away before any appreciable rise in temperature occurred. Moreover, the only visible effect of changing the rate of flow of nitrogen peroxide over rather wide limits (e. g., by doubling or tripling the amount of current passed through the heater used to vaporize the peroxide) was to cause the reaction zone to proceed #morerapidly up the tower. At no time was nitrogen peroxide observed escaping from the tower before the reaction zone reached the top. Copious red fumes of nitrogen peroxide did appear, however, the moment that the upper edge of the reaction zone reached the top of the potassium chloride in the tower. In fact, the main concern in most of the experiments was to prevent the reaction zone from reaching the top of the potassium chloride column too rapidly. Otherwise the gaseous reaction products would have been mixed with nitrogen peroxide, thus interfering with their determination.

Coefficients of Potassium Nitrate in Solutions j

m

Vol. 23, No. 12

0.0147 0.0209 0.0333 0,0470 0.0657 0.0781 0.0975 0.1351 0.1526

The temperatures corresponding to twelve e. m. f. readings, made in determining the melting point of the nitrosyl chloride, varied between -61.3' and -61.8" C., and averaged -61.5' C. This is in exact agreement with the carefully determined value of Trautz and Gerwig (19). It should be noted that this nitrosyl chloride had not been subjected to any purification but was condensed out immediately after leaving the potassium chloride tower. The gas sample in the balloon (0.5737 gram of gas) was absorbed in a known amount of 0.3 N sodium hydroxide, and the excess of sodium hydroxide was titrated with 0.5 N sulfuric acid, using phenolphthalein as an indicator. The titrated liquor was made up to 500 cc., and chlorine was determined gravimetrically in a 100-cc. aliquot. Nitrogen was determined by the Devarda alloy method in another 100-cc. aliquot. Assuming that the acidity of the gas was due to nitrosyl chloride, and that the absorption in alkali took place according to the reaction NOCl 4- 2NaOH +NaNOn 4-NaCl

r j d l o gm J"

0.0243 0.0367 0.0612 0.0888 0.1278 0.1569 0.2015 , 0.2810 0.3128

0.000125

.... ....

uu

0.0001 0.0001 0.0003 0.0004 0.0006 0.0010 0.0011

f dl

al/l

- log m 0.0390 0.0576 0.0944 0.1357 0.1932 0.2346 0.2984 0.4151 0.4643

m al/2

0.914 0.876 0.805 0.732 0.641 0.583 0.503 0.385 0.343

It was found, however, that further conversion occurs after the reaction zone proper has passed on. Thus the nitrate content of the converted salt was always higher a t the bottom of the tower than a t the top. I n a typical case, in which the flow of nitrogen peroxide was stopped when the reaction zone reached the top of the salt column, the salt from the upper part was 45 per cent potassium nitrate, as calculated from its nitrogen content, while that from the lower part was 83 per cent nitrate. That no intermediate products were produced was evident from an analysis of the 45 per cent potassium nitrate material which showed no other salt but potassium chloride present. The most complete conversion was obtained by continuing the nitrogen peroxide flow for a time after the reaction zone had disappeared. This was done in the experiment described in which the treated salt contained 93 per cent potassium nitrate. This delay in conversion is probably not due to slowness of the reaction but to the time required for the nitrogen peroxide to penetrate to the interior of the potassium chloride crystals.

INDUSTRIAL A N D ENGINEERING CHEMISTRY

December, 1931

Literature Cited International Critical Tables, Vol. IV, p. 259, McGraw-Hills I b i d . , Vol. V, pp. 88-91; Vol. VII, p. 303. I b i d . , Vol. V., p. 101. I b i d . , Vol. V., pp. 178-9, 203-4. I b i d . , Vol. V, p. 204. I b i d . , Vol. V I I , p. 238. I b i d . , Vol. VII, p. 239. I b i d . , Vol. V I I , p. 240. (9)I b i d . , Vol. VII, p. 241. (1) (2) (3) (4) (5) (6) (7) (8)

1926’

(10) (11) (12) (13) (14) (15) (IS) (17) (18) (19)

1413

I b i d . , Vol. VII, p. 306. Lewis and Randall, “Thermodynamics,” p. 108, McGraw-Hill, 1923. Lewis and Randall, I b i d . , p. 173. Lewis and Randall, I b i d . , p. 255. Lewis and Randall, I b i d . , p. 298. Lewis and Randall, I b i d . , p. 344. Lewis and Randall, I b i d . , p. 347. Mehring, Ross, and Merz, IND. END.CAEM.,21, 379 (1929). Randall and Rossini, J . A m . Chem. Soc., 61,335 (1929). Trautz and Gerwig, 2. unorg. ullgem. Chem., 134, 409 (1924).

Structural Characteristics of Apatite-Like Substances and Composition of Phosphate Rock and Bone as Determined from Microscopical and X-Ray Diffraction Examinations‘ S. B. Hendricks, W. L. Hill, K. D. Jacob, and M. E. Jefferson BUREAUOF CHEMISTRY AND SOILS, DEPARTMENT OF AGRICULTURE, WASHINGTON, D . C.

Animal bone, free from organic matter, is a carbonate American continental phosphate rock consists essentially of submicrocrystalline fluorapatite, C:al~F2(PO4)6, apatite, Ca~oCos(Poa)s.HzO, isomorphous with fluorcan be prepared by which contains some excess fluorine and a small amount apatite. Oxyapatite, Cal~O(P04)6, of sodium. The compounds Calo(OH)z(PO,),, Cas(H2O)Z- heating hydroxyapatite or bone to constant weight at The apatite-like diffraction pattern of tri(P04)6, and CaloCOs(P04)e.H20 form extensive series of solid 900” C. solutions with fluorapatite, some members of which occur calcium phosphate hydrate is destroyed if the comin geologically recent rocks. The phosphate rock of Curacao pound is heated to constant weight at 900’ C. These Island is essentially a hydrate of tricalcium phosphate, various compounds in natural or artificial mixtures Cas(H20)n(PO&. Hydroxyfluorapatite, Calo(F,0H)~(P04)e,can be identified by means of their x-ray diffraction in which only about half of OH- is replaced by F-, occurs patterns. on Christmas, Nauru, and Ocean Islands.

HE phosphate rock deposits of the world have usually been described as consisting of various amorphous and microcrystalline apatite-like minerals. Lacroix (25) considered certain French phosphorites to consist essentially of colloidal tricalcium phosphate, collophane, crystalline staffelite (francolite), and crystalline dahllite (podolite). Schaller (58) concluded that phosphorites consist of the following compounds in solid solutions:

T

Dahllite Francolite Collophane Fluorapatite Hydroxyapatite

9Cs0~3P~Oa~CaCOs~HzO 9Ca0.3PzOrCa(F1, COa).H?O 9Ca0.3PzOsCaCOs~HzO xHzO SCa0.3PzOs~CaFz 9CaO.3PzOvCa(OH)r

+

Rogers (35) writes: “The principal constituents of phosphorites or so-called phosphate rocks and also of fossil bones is an amorphous substance with properties sufficiently characteristic to be entitled to recognition as a distinct mineral. I t may be called ‘collophane.’ ” He gives as the approximate formula of collophane: 3Ca3(PO4)~1-2Ca(C03,Fz,S04,0).xHz0

A number of poorly defined minerals have been described as occurring in phosphate rock. Some of them (17) are: Voelkerite Dahllite Podolite Staffelite A fibrous form Collophanite

1

CaloO(POd8

3Cas(POi)z.2CaCOa.HzO 3Cas(PO4)zCaCOz

~C~S(PO~)Z.C~COXC~F~.HZO

of francolite (3Cas(P01),)~.(2CaF?C03) y 2 H z 0 (also known‘as-pyroclasite nauruite, pyroguanite. sombrerite, monite, fluocollophanite: and lewistonite)

Received July 27, 1931.

The results of chemical analysis indicate that bone is either carbonate apatite (Ca10C03(P04)6),hydroxyapatite (Ca,,. (OH)2(P04)6), and calcium carbonate; or tricalcium phosphate and calcium carbonate (11, 13). Fossil bone has been described as collophane (35). Since the compounds just mentioned are closely related chemically, it is difficult to identify them in naturally occurring sub-microcrystalline materials. The small particle sizes of the non-phosphatic impurities render difficult the interpretation of chemical analyses. This makes it practically impossible to determine from chemical and microscopic analyses alone the correct distribution of fluorine and carbon dioxide between the phosphate minerals and possible fluoride and carbonate impurities. The use of inaccurate methods for the determination of fluorine (22) and the difficulty in accurately determining combined water in phosphate rocks has frequently, indeed usually, led to incorrect conclusions concerning the nature of the compounds present. The authors have used microscopical, analytical, and x-ray diffraction methods for determining the nature of the com. pounds present in phosphate materials; the application of the three methods usually leads to an unambiguous answer. Microscopical examination is of great value for the identification of minute quantities of crystalline substances, but the finely divided character of many of the compounds present in phosphate rock limit the applicability of this method. Chemical analyses are open to the objections mentioned previously. X-ray powder diffraction photographs are of some value for the identification of impurities present in amounts greater than about 2 per cent. They are of far