Oxidation of Phosphorus to a Pentavalent Form by Carbon Dioxide

ide-carbon dioxide-phosphorus pentoxide- phosphorus tetroxide system is such thatl. ' * with equal molal quantities of the pp&- oxide and tetroxide pr...
3 downloads 0 Views 1MB Size
Oxidation of Phosphorus to a Pentavalent Form b-y Carbon Dioxide P. H. EMMETT' AND J. F. SHULTZ

Equilibria in the PhosphorusCarbon-Oxygen System

Bureau of Chemistry and Soils, U. S. Department of Agricplture, M'ashington, D. C.

The equilibrium in the carbon inonoxide-carbon dioxide-phosphorus pentoxidephosphorus tetroxide system is such thatl with equal molal quantities of the pp&oxide and tetroxide present the ratio of carbon monoxide to dioxide is about 1 to 2. In the presence of phosphate rock the oxidation of phosphorus by carbon dioxide to the pentavalent form in the above temperature range is complete; the product formed is calcium metaphosphate.

In the temperature range 800' to 1,200' C. a gaseous mixture of phosphorus (P4) and carbon dioxide containing more than 11.1 per cent phosphorus vapor quickly reaches an equilibrium corresponding to the conversion of about 80 per cent of the carbon dioxide to carbon monoxide. If less than 11.1 per cent phosphorus is present, practically all elementary phosphorus is absent at equilibrium in this temperature range, a mixture of phosphorus pentoxide and phosphorus tetroxide being obtained.

D

f 5c = 5CO -k 3CaSiOa 4-'/zPa

It has been customary (3) either to burn this phosphorus-carbon monoxide mixture in excess air to produce phosphorus pentoxide and carbon dioxide or first to separate the phosphorus from the carbon monoxide by condensation and then burn tho former to phosphorus pentoxide. By the latter procedure ten volumes of carbon monoxide become available for each volume of phosphorus vapor (PJ produced. It seemed worth while to investigate the possibility of oxidizing phosphorus with carbon dioxide instead of air. If this could be done, one might conserve the original carbon monoxide in the mixture coming from the phosphorus furnace and produce a n additional ten volumes of carbon monoxide for each volume of phosphorus vapor oxidized to the pentavalent form. Accordingly a program of research was undertaken having as its primary object the determination of equilibria in the phosphorus-carbon-oxygen system and a study of the conditions influencing the rates of the various reactions that might be involved.

Flowmeters for obtaining any desired combination of flow of pure dry nitrogen, carbon monoxide, and carbon dioxide are not shown. The insulation on the tubes between the phosphorus bath and reaction tube A was heated sufficiently to prevent condensation of phosphorus as yellow phosphorus. The tubes were also kept entirely fihielded from light to prevent or retard conversion of yellow phosphorus vapor to red phosphorus. The means adopted for obtaining a sample of the oxides of phosphorus formed is also illustrated. A porcelain tube, B , of E / * inch internal diameter, was ground at one end to fit into a ground seat in U-tube C. Another ground connection provided a tight fit between the U-tube and the 1-inch (i. d.) porcelain reaction tube, A . When a sample of oxides was desired, the flow of gas through the furnace was momentarily halted and tube B was raised into position. Its u per end extended to a point about 2 inches below the center orfurnace F. The gas flow was then resumed; the oxides of phosphorus condensed in B whereas the carbon monoxide-carbon dioxide mixture passed on through the U-tube and into a sampling bulb of 25-cc. capacity. The carbon monoxide-carbon dioxide mixture was analyzed on a standard Bone and Wheeler gas analysis apparatus. The U-tube was kept immersed in a carbon dioxide-alcohol bath at -78" C. during a run to condense any volatile products that might be coming through the furnace. In some experiments broken porcelain was inserted into reaction tube A by wedging a perforated porcelain plate (Gooch filter plate) crosswise in the tube. It was possible with a little practice to insert such supporting plates with comparative ease.

Apparatus and Materials I n the course of this work the apparatus underwent a continuous evolution until its form and materials of construction were suitable for the measurements desired. Preliminary experiments were carried out in a quartz reaction bulb packed with broken quartz, but it was found that all phosphorus pent1

*

oxide produced in the oxidation of phosphorus with carbon dioxide was dissolved in the quartz in the temperature range above 800" C. in which it was necessary to work to obtain reasonable reaction rates. Porcelain tubes were adopted as a standard material of construction in the high-temperature experiments. A flow system was used in the equilibrium measurements. Phosphorus vapor of a known partial pressure was obtained by passing nitrogen or carbon dioxide through a thermostated bulb of molten phosphorus. The form of the apparatus used is shown in Figure 1:

URING the last few years a considerable amount of research and development work has been carried on relative to the production of elementary phosphorus by the reduction of phosphate rock with carbon in high-temperature electric furnaces by the reaction: Cas(POd)2f

'

Present address. Johns Hopkins University, Baltimore, Md.

1105

106

INDUSTRIAL AND ENGINEERING CHEMISTRY

VOL. 31, NO. 1

idly. A total phosphorus determination by the usual gravimetric methods was too time consuming. After numerous procedure was adopted: The experiments the samples of mixed oxides were dissolved in water and made up tling the main reaction tube. to 100 cc. Aliquot portions were taken for determination of In some experiments a PhosPhorus Pentoxide tube, D,was intotal phosphorus and of trivalent phosphorus, respectively. serted at point E between the phosphorus bulb and reaction tube A to provide a certain partial pressure of phosphorus pentoxide. After one of the aliq'Jots had been heated at l35" c*in an It was found impossible, however, to provide known partial presautoclave for 1 hour, the total phosphorus was determined as sures of phosphorus pentoxide in this manner because of the wellorthophosphorous (H3POs)or orthophosphoric acid (H3P04) known way in which the pentoxide changes over from a high to a by direct titration with sodium hydroxide in a solution satulow vapor-pressure form ( 7 ) . Estimates of the amount of phosrated with sodium chloride. This time was sufficient to conphorus pentoxide being put into the reacting systemhad to be vert all of the metaphosphoric and pyrophosphoric acid to made, therefore, by periodic blank runs using nitrogen as a carrying gas. orthophosphoric acid. S n y phosphorus tetroxide present reacted with water to form orthophosphorous and orthophosphoric acids. Titration with standard sodium hydroxide to yield the disodium ocouple salts (phenolphthalein as indicator) thus gave a value for the total phosphorus present as phosphoric and phosphorous acids. The method of analysis used for determining trivalent phosphorus was that of Wolf and Sung (IO). The solution to be analyzed was placed in a 500-cc. flask fitted with a ground glassstopper and exactly neutralized.2 Then 50 cc. of a 0.2 molal sodium bicarbonate solution (saturated with carbon dioxide) and 50 cc. of a 0.1 N iodine solution were added. (An excess of at least 10 cc. of iodine above that needed for the oxidation should always be present.) After 45 to 60 minutes the excess iodine was determined by titration with a 0.1 N standard arsenious acid solution (containing sodium carbonate according to the procedure of Treadwell and Hall, 8A) using a 0.5 per cent starch solution as indicator. One cubic centimeter of 0.1 N iodine solution is equivalent to 4.103 mg. of orthophosphorous acid. The value obtained in this way for the trivalent phosphorus together with the total pentavaH K lent and trivalent phosphorus content obm tained by the sodium hydroxide titration Fig I made it possible to calculate the amount of phosphorus tetroxide and of phosphorus pentIn a given experiment the oxides of phosphorus collected in oxide in the sample of oxides. Separate experiments were tube B were washed into a beaker and dissolved without allowing made on several occasions to ascertain whether or not any any contact of the hot oxides with air. The method of analysis of the lower oxides were present. Phosphorus trioxide was used for determining the quantities of phosphorus pentoxide and shown to be unstable in the furnace a t 1,000" C. so that phos horus tetroxide is detailed below: Ye$ow phosphorus was dried in tube H by vaporizing moisture 2 No indicator can be used here as i t will later react with the iodine from it into a vacuum. It was then distilled at low pressure (apThus the quantity of alkali to neutralize the acids is determined on an proximately 100 mm.) in a stream of nitrogen (50 cc. per minute aliquot part, and that same quantity is added a t this point. at standard tem erature and pressure) into bulb J, and from J into bulb K. d n k nitrogen and carbon dioxide were used, the former being freed of oxygen by passage over c. P. copper and dried over phosphorus pentoxide in the usual fashion. The carCONVERSION OF CARBON DIOXIDETO TABLEI. PERCENTAGE bon dioxide was merely dried, for analysis showed that it conCARBON MONOXIDE IX THE PRESENCE O F EXCESSPHOSPHORUS. tained no more than 0.1 per cent noncondensable gas. The carbon monoxide used w&s prepared by the apparatus described by Partial Carbon p T m g of Pressure of Temp. Flow Flow as CO Thompson (8) and stored in steel cylinders at 1,000 pounds per orus Phosphorus of PxOs Furnace of of in Exit osp square inch pressure. It was freed of iron carbonyl by passage Bath Vapor Bulb Temp. COz CO Gas through liquid air. Phosphorus pentoxide for the equilibrium ~ m . C. a C. Cc./min. Cc./min. % c. runs was distilled in oxygen in an iron reaction tube at 800' C. Quartz Reaction Tube according to the method of Finch and Frazer (6). Phosphorus 10 .. 79.3 103 a 1007 194.0 trioxide was prepared by the method of Wolf and Schmager (11) 5 .. 79.0 103 100s 194.3 by the controlled combustion of phosphorus at 90 mm. pressure 5 .. 78.6 99 1095 192.8 1097 5 .... 78 1 193.4 101 D in a 75 per cent oxygen-25 per cent nitrogen mixture. The 10 78.4 101 a 1022 193.5 phosphorus trioxide was then converted to the tetroxide by heatPorcelain Reaction Tube ing in a closed tube at 250" C. for 48 hours (9). 0 *. 350 1003 .. 45 79.7 In the experiments on the reaction of phosphorus vapor and carbon dioxide with phosphate rock, the latter was supported on a porcelain plate in tube B instead of tube A . This facilitated changing from one sample of phosphate rock to another and made possible the removal of the product of reaction without disman-

0

0

(1

(I

Methods of Analysis Early in the experimental work it became evident that a suitable method would have to be devised for analyzing mixtures of phosphorus pentoxide and phosphorus tetroxide rap-

139.1 139 138.8 139

18:2 18.2 18 2 18.2

Not attached.

310 320 322 323 323

1007 1009 1008 1006 1005

..

15 15 20 75

45 35 35 30 42.5

80.3 78 1 77.9 73.3 83.4

JANUARY, 1939

INDUSTRIAL AND ENGINEERING CHEMISTRY

TABLE 11. EQUILIBRIUM DATAFOR Partial Pressure of Phosphorus R u n Flow Flow inEntering Av. C No. of COn of R'2 Gas as CO Cc./min. Cc.lmin Mm. %

THE

50 50 50 50 50

200 201 209 133 134

50 50 50 50 50

135 136 161 162 142

50 50 50 50 50

143 150 151 154 155

50 50 50 50 50

P as Pa08

P4On

P as

-Calcd. From Pi08

%

Mu.

Mg.

cc.

168 169 172 173 137

50 50 50 50 25

138 139 163 164 144

25 26 25 26 25

145 140 141 152 153

25 25 25 25 25

156 148 149 158 170

25 12.5 12.5 12.5 12.5

171 146 147 159 160

12.5 12.5 12.5 35 35

186 187 188 189 192

25 25 25 25

195 196 179 180 181

50 50 25 25 25

182 183 184 185 193

25 25 25 25 12.5

194 197 174 175 176

12.5 5 25 25 25

177 190 191

25 25 25

0

b

cc.

cc.

Obsvd. COa

cc.

c.

Oxidation Approach without Added Phosphorus Pentoxide

.. .. .. ..

55.05 51.75 56.10 25.15 25.50

44.95 48.25 43.90 74.85 74.50

718.4 660.0 830.4 271.0 259.6

266.6 315.5 265.5 361.0 375.4

1037 953 1200 391 375

481 569 480 652 678

1518 1522 1680 1043 1053

1650 1553 1720 1132 1147

1004 1003 1004 1010 1003

0 45 0.51 0.43 0.44 0.40

..

18.9 19.4 20.7 18.8 18.8

28.50 25.80 28.15 17.35 17.20

74.60 74.20 71.85 82.65 82.80

269.6 270.6 215.4 116.8 107.4

414.9 399.4 231.6 215.4 212.8

389 391 311 169 155

749 721 418 389 384

1138 1112 729 558 539

1147 1230 845 564 509

1004 1012 1000 1003 998

0.52 0.51 0.38 0.39 0.41

18.8 18.8 21.2 21.2 9.2

17.00 14.65 26.90 26.90 8.70

83.00 85.35 73.10 73.10 91.30

120.4 130.6 212.0 224.8 36.9

204.1 203.4 210.5 220.0 123.1

174 189 306 325 53

369 367 380 397 222

543 556 686 722 275

510 439 807 808 261

1013 1005 1016 1007 1008

0.35 0.27 0.36 0.36 0.32

9.3 9.2 9.2 9.4 9.4

10.35 12.35 12.45 12.35 12.50

89.65 87.65 87.55 87.65 87.50

48.0 19.3 75.2 84.2 97.2

124.0 45.9 202.8 213.8 256.8

69 28 109 121 140

224 83 366 386 463

293 111 475 507 603

311 526 736 741 751

1017 1020 1009 1002 1007

0.30 0.33 0.38 0.36 0.38

9.5 9.5 10.9 10.9 18.8

13.10 13.10 14.40 14.55 28.30

86.90 86,QO 85.60 85.45 71.70

137.8 135.4 82.6 91.8 217.6

276.7 298.6 150.6 157.6 142.4

199 196 119 134 314

499 539 272 285 257

698 735 391 419 571

787 784 431 432 424

1019 1005 1004 1007 1010

0.30 0.33 0.31 0.29 0.26

19.0 19.0 21.5 21.5 9.2

30.00 30.10 45.10 45.15 11.95

70.00 69.90 54.90 54.85 88.05

223.6 206.8 373.2 372,s 32.3

118.6 145.4 73.8 76.2 104.2

323 298 539 110 47

214 262 133 673 188

537 560 672 783 235

450 451 676 677 269

1009 1027 1012 1005

0.23 0.30 0.16 0.17 0.44

9.2 9.0 9.0 9.4 9.4

11.90 18.45 18.65 21.50 21.55

88.10 81.55 81.35 78.50 78.45

51.2 77.6 65.2 125,O 160.3

135.3 102.4 115.3 121,3 167.5

74 112 94 180 232

244 185 208 219 302

318 297 302 399 534

356 277 280 564 647

1016 1008 1001 1006 1008

0 36 0 30 0.40 0.27 0.29

0.40

9.3 9.3 9.3 9.3 10.4

21.75 10.85 10.95 12.20 14.60

78.25 89.15 89.05 87.80 85.40

159.0 30.3 39.8 20.7 28.5

176.6 50.2 72.2 53.9 50.3

230 44 57 30 41

319 91 130 97 91

549 135 187 127 132

652 163 164 251 219

1021 1004 1010 1009 1011

0.30 0.20 0.22 0.36 0.30

0.31

10.4

85,55 79.95 80.15 67.30 67.35

36.7 110.0 98.8 272.4 262.6

57.3 109.0 105.1 95.8 95.6

53 159 143 393 379

103 197 190 155 154

158 356 333 548 533

216 301 296 589 588

1017 1013 1000 1018 1014

0.26 0.25 0.26 0.17 0.18

0.27

2.75

9.0 9.3 9.3

14.45 20.05 19.85 32.70 32.65

0:26

3:01

o:is

2:97

39.9 39.9 39.6 39.6 2O.Ob

61.75 54.70 57.25 54.50 30.20

1008 1005 1004 1014 1002

0.61 0.56 0.51 0.53 0.50

20.0b 20.06 19.2 19.5 19.5

29.60 28.50 33.20 38.55 38.75

1008 1010 1012 1008 1013

0.51 0.51 0.54 0.49 0.53

20.1 20.1 19.2 19.2 20.Ob

39.00 37.90 37.55 37.10 39.30

1014 1018 1012 1008

0.49 0.56 0.53 0.52 0.68

20.0b 20.3b 11.4 11.4 11.4

36.75 36.70 18.10 17.95 17.55

1014 1016 997 1001 1008

0.57 0.55 0.49 0.46 0.46

11.4 9.3 9.3

16.90 18.05 18.10

1004 1010 1006

0.46 0.50 0.49

..

..

60

Yield of COFrom Paoio Total

40.0 41.9 43.2 18.9 18.9

..

.

SYSTEM CARBON DIOXIDE-CARBON MONOXIDE-PHOSPHORUS TETROXIDE-PHOSPHORUS PENTOXIDE

Av. C as COz

A.

202 203 205 198 199

107

.. .. .. .. .. .. .. .. .. .... ..

.. ..

.. 25 25 25 25 25

.. ..

25 25 25 25 25

... . .. ..

1215 12.5 15 15

.. ..

.. .. .. , .

.. .. .. .. .. .. .. .. ..

..

.. ..

.. ..

9.0

B. Oxidation Approach with Added Phosphorus Pentoxide 38.25 678.4 255.6 980 461 1441 1390 45.30 563.2 258.8 814 467 1281 1231 42.75 495.8 190.1 716 343 1059 1000 45.50 444.4 195.9 642 354 996 954 69.80 224.0 260.8 373 323 696 905 70.40 252.0 306.5 364 551 915 887 71.50 234.8 300.7 339 542 881 856 66.80 594.0 651.0 858 1175 2033 664 61.45 364.0 285.0 526 514 1040 866 61.25 328.0 274.5 474 496 970 871 61.00 309.8 235.2 447 425 872 878 62.10 324.0 301.5 468 544 1012 994 62.45 272.8 238.7 394 431 825 845 62.90 282.4 248.1 408 448 856 834 60.70 109.1 114.4 158 206 364 442 64.25 172.7 168.8 249 304 553 551 64.30 181.4 171.3 262 309 571 551 81.90 117.4 257.8 169 465 634 408 82.05 134.8 283.0 195 511 706 403 82.45 136.8 296.2 198 535 733 394 83.10 178.6 406.9 258 734 992 380 81.95 162.3 104 71.7 293 397 406 81.90 79.8 178.6 115 322 437 407

Observed CO is product of total flow during run and average fractional CO content of the exit gas. Temperature of phosphorus taken with a thermometer: 2 ' C . added for stem correction.

1010

1000

7.05 0:46

..

.. .. .. .. .. .. ..

i:05

..

... .

.. 0141

... .. ..

i:az

..

..

..

.. ..

22:38

.. ..

.. .. 0:33 I

.

.. ..

.. . t

.. 3:35

.. *. ..

3.99

..

0:22

..

.. .. .. I

.

..

..

16:75

... .

.. ..

4.92 I .

..

..

, .

6.6

0 : 56

..

s:9

0:51

7:s

..

..

.. .. .. , .

0153

..

0.62 0.55

.. ..

..

..

.. ..

i:s

.. .. .. ..

4.2

0:47

6:3

.. .. ..

INDUSTRIAL AND ENGINEERING CHEMISTRY

108

even with considerable inlet quantities of this oxide none could be detected in the exit gases. The aqueoussolution of the phosphorus oxides gave negative tests for mono- and divalent phosphorus by the method of Wolf and Jung (IO). Apparently the tetroxide and the pentoxide constituted the only oxides of phosphorus present in the exit gases under the experimental condition of the present work.

Fractional Conversion of Carbon Dioxide to Carbon Monoxide in Presence of Excess Phosphorus

VOL. 31. NO. 1

Equilibrium in the Oxidation of Phosphorus with Excess Carbon Dioxide When a porcelain tube was substituted for a quartz reaction tube, the exit carbon monoxide content never corresponded to the complete conversion of the phosphorus to the pentavalent form but t o a conversion intermediate between phosphorus pentoxide and phosphorus tetroxide. It appeared probable, therefore, that for the reaction, P&

+ 2co2 = P4010+ 2 c o

(2)

The first equilibrium experiments on the oxidation of an equilibrium was being obtained. Accordingly, the method described in the present paper was devised for obtaining samphosphorus with carbon dioxide were made a t about 1,000" C. ples of the phosphorus oxides a t 1,000"C. as well as samples using a quartz reaction tube. It was found that the phosof the gas for carbon monoxide and carbon dioxide analysis. phorus oxidation yielded an amount of carbon monoxide corThe results obtained are shown in Table IIA for partial presresponding to complete oxidation of the phosphorus to the sures of phosphorus corresponding approximately to 10, 20, pentavalent form. It was discovered, however, that the and 40 mm. Table IIB shows some results obtained when a quartz was reacting with the phosphorus pentoxide as fast as known partial pressure of phosphorus and an unknown partial i t was formed. Accordingly, in the quartz apparatus under pressure of phosphorus pentoxide were being introduced with the conditions leading to complete oxidation of the phosknown carbon dioxide or carbon dioxide-nitrogen gas mixphorus by carbon dioxide (to the pentavalent form), free phostures. phorus pentoxide was not obtained, but a reaction product Table I11 presents various experimental attempts to obtain consisting of quartz and phosphorus pentoxide. an equilibrium approach to reaction 2 from the reduction I n experiments in which an excess of phosphorus was presside. I n all of them small flows of carbon monoxide ranging ent, it appeared that an equilibrium was definitely obtained from 1 to 5 cc. per minute in a carrying stream of 24 to 20 cc. even in a quartz apparatus. Table I shows the results of a of nitrogen per minute were passed into the reaction tube tonumber of experiments in which carbon dioxide-phosphorus gether with an unknown amount of phosphorus pentoxide. mixtures containing ratios smaller than 8 to 1 were passed Analysis of the oxides of phosphorus formed were carried out through a quartz reaction tube. The exit carbon monoxidein the manner described above. The carbon dioxide-carbon carbon dioxide gas generally contained about 80 per cent carmonoxide analysis was obtained by a flow method. After bon monoxide. Experiments are also listed in Table I that passage through the U-tube a t -78" C., the exit gases were were performed in the porcelain apparatus with a known mixsent first through a drying tube to remove any water vapor ture of phosphorus, carbon monoxide, and carbon dioxide tothat might be present, then through ascarite to catch the cargether with an unknown amount of phosphorus pentoxide. bondioxide, then through a hot cupric oxide tube followed by a The exit gas composition clearly shows that as long as excess drying tube and an ascarite tube. Most of the results appear phosphorus is present, the exit carbon monoxide-carbon dioxide ratio tends towards 80 to 20. When, for example, 70 per cent carbon ~onOxide-30Per cent carbon dioxide is TABLE111. EQUILIBRIUM DATAFOR THE SYSTEM CARBON DIOXIDE-CARBON MONadded, the conversion raises the carbon OXIDE-PHOSPHORUS TETROXIDE-PHOSPHORUS PENTOXIDE (REDUCTION APPROACH) monoxide content to about 78 per cent; Av. Av. Temp. Time P4O*,*=CO Ft Flow Flow $ Temp. p;$, %g $6; F4g (m) (m) when, on the other hand the entering gas is 85 per cent or more carbon co.1 cc:/ monoxide, the exit percentage of carrnzn. mzn. % % C. Mg. Mg. Min. 1009 5 20 361.6 60.4 90 107 328 67.05 32.95 0.83 bon monoxide is lower than the entering. 1013 482.2 482.6 2.5 22.5 58.10 75 108 354 41.90 0.72 1009 148.8 It appears from these and other data 24 64.0 90 110 331 36.95 63.05 0.38 1 1011 408.0 249.2 24 70.15 111 357 29.85 0.33 1 120 that the first step of the oxidation of 1006 22.5 169.2 54.5 112 1.11 353 66.20 33.80 2.5 90 1009 283.2 10.3 113 379 70.50 120 29.50 0.43 2 . 5 22.5 phosphorus by carbon dioxide must 1015 128.0 30.4 114 24 1 120 377 58.80 41.20 0.69 1008 115 120.4 37.1 24 57,25 42.75 1 proceed to some other oxide than the 115 378 0.74 1012 128.3 79.15 208.7 24 292 20.85 105 1 11s 0.65 pentoxide; otherwise the same equilib1005 118.3 218.1 24 1 7 . 4 0 82.60 75 124 1 337 0.28 3 2 . 6 4 9 . 6 122 7 7 . 2 5 1007 24 2 2 . 7 5 347 128 1 0 .36 rium percentage carbon monoxide in the exit gases could not be obtained in both a quartz and porcelain tube inasmuch as entirely reasonable and confirm the equilibrium values obphosphorus pentoxide has been shown to react rapidly and tained from the oxidation side. The percentage carbon monpractically completely a t 1,000" C. with a quartz reaction oxide in the first run on each day was, in general, lower than tube. It seems probable that the oxide formed is phosphorus that for the other runs and was viewed with suspicion. Ustetroxide (P408) and that the over-all reaction is: ually several determinations of carbon dioxide and carbon Pq 8C02 8CO PdOs (1) monoxide were made for each sample of phosphorus oxides. Of these gas analyses only the first on each day was omitted. (co)*(~~os) If this is true, the equilibrium constant K, = (C02)8(P4) In calculating the equilibrium constants in the last column of Table 111,averages of the gas analyses for a given oxide were will be about 6 X lo4 for, in the experiments in which an taken. exit ratio of carbon monoxide to carbon dioxide of 4 to 1 was obtained, the partial pressures of phosphorus tetroxide and Reaction between Phosphorus-Carbon Dioxide phosphorus would have been roughly equal as calculated by and Phosphate Rock the known entering phosphorus content and the amount of The above experiments disclosed that an equilibrium excarbon monoxide formed. In the experiments listed in Table I isted in the oxidation of phosphorus with carbon dioxide such no attempt was made to analyze the mixture of phosphorus that a mixture of phosphorus tetroxide and phosphorus pentand phosphorus oxides obtained.

$6

,/iaoe

P;,

ROufn

.

+

+

JANUARY, 1939

INDUSTRIAL AND ENGINEERING CHEMISTRY

oxide rather than pure phosphorus pentoxide was always obtained as a product; it became apparent, therefore, that if this method of oxidizing phosphorus was to be of practical value, some means had to be found for removing the phosphorus tetroxide from the mixture of oxides. Separation of the two oxides by distillation appears to be entirely impracticable because their vapor pressures are nearly equal and be-

TABLEIV. REACTION BETWEEN PHOSPHORUB, CARBON DIOXIDE, AND FLORIDA PHOSPHATE ROCK^ Time from Start Min.

Oxide No.

Partial Pressure of Flow Phosphorus COz Vapors Cc./min. Mm.

420 ... 445 ... 570 660 720 755 ... 803 833 201 863 0 Analyses before run, 67.9 per cent.

...

... ... ... ...

50 50 50 50 50 50 50 50 50

PiOs

19.2 19.2 19.2 19.2 19.2 19.2 19.2 19.2 19.2 = 31.1 per cent;

CO in

Exit Gas

Remarks

%

28.4

Phosphate rock in tube B

28.3 27.3 26.1

Porcelain tube B empty 25.8 analyses after run, PzO,

-

cause both oxides tend to go over readily to low vapor-pressure forms. Gas-phase oxidation of the tetroxide to pentoxide is possible but proceeds rather slowly up to 500" C., as some of the experiments cited below show. Oxidation of phosphorous acids in solution a t high pressure b y oxygen or steam in the presence of a catalyst is possible but entails some difficult and expensive procedures. It was finally decided to try the oxidation of phosphorus by carbon dioxide in the presence of phosphate rock in the hope that the latter, like quartz, would preferentially remove the phosphorus pentoxide from the equilibrium of reaction 2 and thus enable the carbon dioxide to oxidi~ecompletely all of the lower valence phosphorus to the pentavalent form. The results are shown in Table IV. The phosphate rock reacted completely to form the metaphosphate, the product having practically an identical analysis to that reported for the product of reaction of phosphorus pentoxide and phosphate rock by Curtis, Copson, and Abrams @). The increase in the percentage of carbon monoxide in the exit gas when the phosphate rock is present clearly confirms the additional oxidation of the lower oxide of phosphorus by carbon dioxide that takes place. No trace of trivalent phosphorus could be detected by analysis in the glassy metaphosphate product of the reaction.

Molecular Weight Determination on the High Vapor-Pressure Form of Phosphorus Tetroxide I n connection with an experimental study of the properties of phosphorus tetroxide West (9) reported some values for a molecular weight determination. His experiments were carried out on a sample of phosphorus tetroxide prepared from the trioxide. The measurements of West were made a t 1,400"C.hy a Victor Meyer type of apparatus. Experiments a t 900" C. had failed to yield any expulsion of gas from the apparatus, indicating that a t 900" C. the vapor pressure of the form of tetroxide employed was less than one atmosphere. West's results indicated a molecular weight of P801efor the tetroxide. I n the present work samples of tetroxide were prepared by the usual methods. Since one step in the preparation involves a sublimation at 200" C., it appeared possible to de-

109

termine the molecular weight of this low-temperature form of the tetroxide. The apparatus used was essentially a Victor Meyer apparatus arranged for operation a t 100-200 mm. rather than a t atmospheric pressure. The sample of tetroxide sealed in a capsule in vacuum was brought to the desired temperature (about 500") in the Victor Meyer apparatus and then broken with a magnetic plunger. The amount of tetroxide used was determined by weighing the filled capsule initially and subtracting the weight of the broken glass particles of the capsule a t the end of the run. It was also determined by titrating the tetroxide washed from the apparatus a t the end of each run. I n fourteen consecutive runs the apparent molecular weight of the tetroxide as judged from the weight of the tetroxide and from the titration (shown in parentheses) wasasfollows: 293 (272), 309 (263), 331 (286), 515 (271)) 721 (678), 409 (439), 470 (301), 412 (393), 322 (313), 447 (280), 402 (299), 399 (295), 468 (366), and 337 (306). Failure to collect all particles of the broken capsule, failure to allow time for complete vaporization of the sample, and failure to prevent diffusion of some of the tetroxide through the surrounding nitrogen atmosphere to the cold part of the Victor Meyer tube would all tend to cause the apparent molecular weights to be too high. If the formula were P8016, the molecular weight would be 504. It appears probable, therefore, that the low-temperature form of the tetroxide is P40s; unfortunately time did not permit sufficient work with the molecular weights to make them conclusive.

Kinetics of Phosphorus Oxidation by Carbon Dioxide A few experiments were carried out with a view to obtaining some knowledge as to the temperature coefficient of the oxidation of phosphorus by carbon dioxide, as to the dependence of the rate upon carbon dioxide and phosphorus partial pressures, and as to whether the reaction occurred in the gas phase or on the walls of the reaction vessel. The results obtained are in no sense comprehensive or final. Although much additional work should be done, the experiments are cited here because of the meagerness of experimental results in this field. The reaction rate experiments had to be carried out a t a temperature sufficiently low to yield only relatively small percentages of oxidation. Only in this way could one stay far enough away from equilibrium conditions to make the results significant from a reaction rate point of view. Temperatures in the neighborhood of 700" to 800" C. were used. The results are shown in Table v. The comparison of results with and without a packing of cracked porcelain in the reaction tube is also included.

TABLEV. DEPENDENCE OF REACTION RATEOF PHOSPHORUS AND CARBON DIOXIDE ON TEMPERATURE AND ON PARTIAL PRESSURES OF PHOSPHORUS AND CARBON DIOXIDE Partial Pressure of co Flow Flow Phosphorus Formed of Con of Nz Vaoor CO Der Min. 0 c.Cc./min. cc./min. Aim. % cc. Empty Porcelain Reaction Tube 19.3 850 20 5 9.2 1.84 19.3 15 8.1 847 10 0 81 19.3 10.8 850 10 1.08 15 20.1 10 12.7 1.27 15 85 1 8.4 0.84 15 10 9.0 852 13.0 21.0 1.30 15 850 10 8.3 0.83 15 9.2 10 852 10.6 15 19.3 10 1.06 850 4.0 19.3 821 20 5 0.79 4.7 19.3 0.47 821 15 10 2.5 21.5 1.25 808 50 .. ... Porcelain Reaction Tube Packed with Broken Porcelain 25 25 21.9 2.8 0.70 SO8 805 25 25 21.9 2.9 0.72 801 25 25 21.9 2.9 0.72

No. of Runs Furnace iiveraned Temo.

2 2 2 3 3 3 3 4 2 3 1 1 1 1

110

INDUSTRIAL AND ENGINEERING CHEMISTRY

Miscellaneous Observations I n the course of the experimental work numerous observations were made that are not entirely explainable, and yet as far as the authors know are not listed in the literature. A few of them will be recorded here. They merit much more extensive investigation. Phosphorus tetroxide was distilled back and forth in a stream of air through a furnace raised to higher and higher temperatures, the time of contact in each case being about 10 seconds. Oxidation in the absence of a catalyst began a t about 350" C. but did not become complete with short times of contact until about 500" C. was attained. The presence of platinized asbestos accelerated the oxidation greatly. I n connection with preliminary molecular weight determination of phosphorus tetroxide, it was noticed that when carefully purified tetroxide was heated quickly in the Victor Meyer apparatus, it invariably yielded an appreciable amount of brilliant red product. This was true even when carefully selected transparent crystals of the tetroxide were sealed in tubes and used for the determinations. It was first thought that traces of phosphorus were present and being converted to red phosphorus. This theory was disproved by carrying out partial oxidation of the tetroxide in oxygen before making the molecular weight determinations. Eventually it was established that the red product was formed by the reaction of traces of water vapor with the tetroxide. If the latter in the course of preparation was sealed into capsules under conditions that excluded all traces of water vapor, molecular weight determinations such as those presented above could be obtained without the formation of a red product. A few experiments were carried out with a mixture of phosphorus trioxide and carbon dioxide. The products of oxidation were again an equilibrium mixture of the tetroxide and the pentoxide, The tetroxide itself is unstable so that if it is passed through the reaction tube in a stream of nitrogen a t 1,000" C. it decomposes, yielding an orange red product. No analysis of the product was attempted. Experiments showed that mixtures of phosphorus pentoxide and phosphorus in nitrogen would react in the porcelain tube a t 1,000" C. Apparently the pentoxide is rapidly reduced to the tetroxide under such conditions.

+

coz = Pzo6 f co 2Pzo4 f 2c02 = 2 c 0 f P4010 PzOa

VOL. 31, NO. 1 (3) (4)

The constants for reaction 3, together with some values for reaction 4, are included in Table 11. It is evident that the best constant from these equilibrium data corresponds to reaction 3. This is not definite proof that Pzo4 and PzO6 actually are present as the principal form of the oxides a t the exit of the heated zone of the furnace, but it is a t least an indication that such is the case. The usual methods of assuring that data correspond to true equilibrium are difficult to apply in the present instance. If, for example, one slows the rate of passage more and more, one would expect to obtain constant values of the equilibrium constant for rates of passage a t and below those necessary to permit the gas to equilibrate. I n the present experiments two factors made this procedure indecisive. I n the first place, the carbon monoxide balances calculated and shown in Table IIA indicate that when no P4010wasbeing admitted with the phosphorus, a considerable portion of the entering phosphorus was oxidized with the formation of carbon monoxide without the appearance of the corresponding amount of the oxides of phosphorus. Thus the carbon monoxide volumes equivalent to the oxides actually collected (Table IIA) are much smaller as a rule than the measured carbon monoxide. This factor appears to be the more important, the smaller the rate of gas passage. Hence, assuming that the pentoxide rather than the tetroxide is dissolved to some extent in the porcelain, the equilibrium constant might be lowered for the slow-flow experiments as a result of the removal of phosphorus pentoxide and failure of carbon monoxide-carbon dioxide ratio to readjust itself rapidly enough to the final phosphorus pentoxidetetroxide mixture. The second disturbing factor arises from the possibility that with longer times of contact the more highly polymerized products P408,PsOle,or P,Ol0may be present in increasing amounts. I n such an event the carbon monoxide-carbon dioxide mixture would strive t o adjust itself to conform to the constant for some of the reactions involving these higher polymers. Hence, we cannot be assured that final equilibrium is reached by merely slowing down the rate of gas passage until no further change in the equilibrium constant occurs. It appears that a better criterion for judging the attainment of equilibrium is the constancy of the '(equilibrium conDiscussion and Conclusions stahts" at a given total flow as the partial pressure of phosphorus in the entering gas is changed. According to this The equilibrium data on reaction 2 require further comcriterion, reaction 3 seems to represent the experimental data ment. It is well known that the molecular weight of gaseous best, as pointed out previously. phosphorus pentoxide corresponds to the formula P4010. The Even for reaction 1it is difficult to specify the reactants and molecular weight of the tetroxide, on the other hand, appears products with any certainty. The tetroxide may be present t o be PaOs for the low-temperature variety; it has been rein any of the polymers of POz as the gas mixture passes out of ported as PsO16 for the high-temperature variety. Furtherthe heated zone. The phosphorus is probably present as a more, in the extremely short time of contact used in the presmixture of P4, Pz,and P, though little is known about the rate ent experiments (usually 5 or 6 seconds), i t is possible that the of equilibration of such a mixture. Accordingly we can say first products of oxidation, PO2 and P&, might not have a merely that in the oxidation of phosphorus by carbon dioxide, chance to polymerize to P408or P8016,whichever of these we only about 80 per cent of the dioxide is converted to monoxide consider the stable molecular form a t 1,000" C. It is also true when the phosphorus is in excess, and the total oxidation that the pentoxide might well be present a t the end of a few corresponds t o about 50 per cent conversion of the phosphorus seconds time of contact as PZOsrather than as P 4 0 1 0 . It soon t o the tetroxide. Attention should be called to the fact that beca.me evident that the experimental data listed in Table I1 ( p 4 0 1 9 ) ( COZ)~ the 80 per cent conversion of carbon dioxide to carbon monwould not conform to the equilibrium constant (p40s)(co)2 oxide effected by phosphorus (when the ratio of COZto Ph is less than 8 to 1 initially) is in agreement with the result obcorresponding to reaction 2. Accordingly, a series of equatained by Britzke and Pestov (1). , or tions was written involving the molecules PO*,P z O ~P408, Attempts to approach the equilibrium constant from the rePsOl~,on the one hand, and the molecules PzOa or P4Ol0on the duction side were partly successful in the case of the tetroxother. A calculation of the equilibrium constants for two ide-pentoxide equilibrium studies. Here, too, it must be runs having different P4to COzratios for these reactions showed borne in mind, however, that in reduction one is probably that very poor constants were obtained for those instances starting with P&. The tetroxide formed is likely to be in in which either POz, P408,or P8OI6was assumed for the the form of P4O8or P801B.Accordingly, without some knowltetroxide. The remaining possible reactions are:

JANUARY, 1939

INDUSTRIAL AND ENGINEERING CHEMISTRY

edge of the equilibrium constants for the system P02P204P408-P8016and the system P4010-PaOa one cannot specify with certainty what "constant" for the reduction reaction should be compared t o the oxidation reaction. I n Table I11 the data for the reduction approach of equilibrium for reaction 2 are given. Values for *calculated from these data are in fair agreement with the Ks constants for the oxidation approach. In Table I11 the carbon dioxide obtained from the reduction approach exceeds in each case the amount equivalent to the phosphorus tetroxide formed. The explanation of this discrepancy is not known. I n some cases it can be attributed to the formation of free phosphorus. However, in the runs with less than 40 per cent carbon monoxide in the exit gas it does not seem possible that this can be the explanation. In view of the above facts it is difficult to specify an exact equilibrium constant for the system carbon dioxide-phosphorus tetroxide-carbon monoxide-phosphorus pentoxide. An exact determination can be made only if a reaction vessel can be used that does not dissolve appreciable quantities of the pentoxide and allows sufficient time of contact to assure getting the final form of the oxides of phosphorus into that corresponding to true equilibrium a t 1,000" C For practical purposes, however, one can safely say that in an equilibrium mixture, when the ratio of tetravalent phosphorus oxide t o pentavalent oxide is 1 to 1, the ratio of carbon dioxide to carbon monoxide will be about 2 to 1. The data in Table V relative to the kinetics of the oxidation of phosphorus are far from complete. Nevertheless, some definite indications were obtained as to the dependence of rate of the C02-P4reaction upon temperature, and upon the partial pressures of carbon dioxide and of phosphorus vapor. Numerous comparisons in Table V show the variation in the amount of carbon monoxide found when the carbon dioxide flow changed from 10 to 20 cc. per minute, the total flow of nitrogen and carbon dioxide and the partial pressure of phosphorus remaining constant. A comparison of eight runs a t 848-850' C. and eight runs at 820-825" C. shows that the rate increases between 1.65 and 1.72 times when the partial pressure of carbon dioxide is increased twofold. The rate thus appears to be proportional to (Pco,)",where n is between 0.7 and 0.8. A comparison of twelve runs in which the phosphorus partial pressure was varied indicates that a twofold increase in the partial pressure of phosphorus causes the rate t o increase about 1.51 times. The rate therefore appears to vary as the six-tenths power of the partial pressure of phosphorus. The temperature coefficient for the reaction can also be estimated from the data in Table V. From the runs employing 10 cc. carbon dioxide per minute, one finds by calculation from the usual exponential equation an energy of activation of 68,000 calories, and from the runs employing 20 cc. carbon dioxide per minute, an energy of activation of 76,000 calories. These values are necessarily approximate. They appear to be about the right order, however, to correspond to a bimolecular reaction with a rate that becomes appreciable only above 800" C. Runs to determine whether the oxidation of phosphorus by carbon dioxide was a homogeneous or wall reaction are almost entirely lacking. Three runs listed in Table V were performed when the porcelain tube was filled with five hundred pieces (4 to 6 mesh) to effect a five- to tenfold increase in surface area compared to the wall area of the heated portion of the empty tube. When due allowance is made for the slight differences in Pco,, PP,,and temperature in these runs compared to the runs with an empty porcelain tube, it appears that the rates of reaction in the packed and unpacked tubes are similar. This behavior would indicate that the first step in the oxidization of phosphorus by carbon dioxide is primarily a gas-phase oxidation rather than a surface reaction,

$11

though too few runs were made to warrant definite conclusions~ on this point. The results of the present work make it evident that it is impossible to oxidize phosphorus vapor to phosphorus pentoxide without the formation of considerable phosphorus tetroxide, unless we use prohibitively high ratios of COz to P4. This conclusion is contrary to the numerous patents (6) that have been obtained upon the oxidation of phosphorus to phosphorus pentoxide by carbon dioxide. It is in partial agreement, however, with the work of Britzke and Pestov ( I ) who concluded that the product of oxidation of P d with COz is P ~ Oin S the temperature range 800" to 1,300' C. The claim of these latter authors that with ratios of COzto P4as high as 22 t o 1 only P408is formed is a t variance with the present experimental results. The constant Ka of Table IIB would indicate that more than 50 per cent of the tetroxide formed would be converted to the pentoxide a t equilibrium a t about 1,000" C. It appears entirely practicable, however, to convert phosphorus vapor to pentavalent phosphorus in calcium metaphosphate by the proposed method (4) of causing phosphorus and carbon dioxide to impinge on phosphate rock. Theoretically only ninety volumes of carbon dioxide would have to be added t o each hundred volumes of a phosphoruscarbon monoxide mixture containing 91 per cent carbon monoxide and about 9 per cent phosphorus vapor in order to convert all of the phosphorus to the pentavalent form in the presence of phosphate rock. In practice it would be well to provide a safe margin by using an amount of carbon dioxide somewhat in excess of that theoretically required. The method of recovering the carbon monoxide from the exit gas would then merely entail removal of carbon dioxide by suitable scrubbing systems. The carbon dioxide for the oxidation could be regenerated by the water gas conversion reaction with carbon monoxide, hydrogen being obtained as a final product. The use of hydrogen so formed for ammonia synthesis or other commercial reaction is self-evident.

Acknowledgment The writers are indebted to T. H. Tremearne for the actuaI analysis of the calcium metaphosphate formed in the experiments.

Literature Cited (1) Britzke and Pestov, Trans. Sci. Inst. Fertilizers (U.S. S. R,), No. 59, 5-160 (1929) (summary in English). (2) Curtis, Copson, and Abrams, Chem. & Met. Eng., 44, 140-2 (1937). (3) Easterwood, Ihid., 40, 283-7 (1933) ; Curtis, Ibid., 42, 320-4, 488-91 (1935); Curtis and Miller, Ibid., 43, 408-12 (1936); Kirkpatrick, Ibid., 44, 643-50 (1937). (4) Emmett, P., U. 5. Patent 2,107,857 (1938). (5) Finch and Frazer, J. Chem. Soc., 129, 117-19 (1926). (6) Sigrist, J., French Patent 640,287 (1928); Voituron, E., German Patents 528,504,531,498,540,068 (1931). (7) Southard and Nelson, S. Am. Chem. Soc., 59, 911 (1937). (8) Thompson, IND.ENO.CHEW,21,389 (1929). (8A) Treadwell and Hall, "Analytical Chemistry," Vol. 11, p 555. (9) West, J. Chern. Soc., 81, 923 (1902). (IO) Wolf and Jung, 2.anorg. allgem. Chern., 201, 337 (1931 (11) Wolf and Schmager, Ber., 62, 780 (1929).

RECEIVED July 15, 1938. Presented before the Division of Industrial end Engineering Chemistry at the 96th Meeting of the American Chemica Society, Milwaukee, W s . , September 5 to 9, 1938.