Vacuum distillation technique for recovery alkali-metal reaction

Citation data is made available by participants in Crossref's Cited-by Linking ... Recovery and determination of crystallographic modifications of K3T...
0 downloads 0 Views 356KB Size
Vacuum Distillation Technique for Recovery of Alkali-Metal Reaction Products Randall F. Gahn Lewis Research Center, Cleveland, Ohio 44135 CENTRAL TO THE UNDERSTANDING of alkali-metal reactions, and in particular, the corrosion of materials by alkali metals, is the recovery and identification of reaction products. Basically, the problem of alkali-metal reaction-product recovery is t,wofold. First, the reaction product must be separated from excess alkali metal. Second, because alkali-metal reaction products generally react readily with water and/or carbon dioxide, the product must be protected from exposure to air prior to and during the separation and in subsequent handling. Several attempts at alkali-metal reaction-product recovery have been reported (1-3). In these recovery attempts, it is likely that reaction-product contamination occurred. Generally, in the studies cited, excess alkali metal was removed by treatment with alcohol or ammonia. Additionally, in some instances the residues were exposed to air. This report describes a procedure for the isolation and recovery of alkali-metal reaction products. The procedure was designed to minimize reaction-product contamination by carrying out all operations in high-vacuum or inert gas atmospheres. The separation of reaction products from excess alkali metal is accomplished by vacuum distillation to avoid the use of solvents or reactants. This procedure was applied to the recovery of products from the reactions of potassium and lithium with tantalum, niobium, vanadium, and oxygen. As an indication of the application and reproducibility of the procedure, two examples are detailed herein: the isolation of the reaction product from the potassium-tantalum-oxygen system, and the isolation of the reaction product from the potassium-oxygen system.

-

Movable rod

,-Transfer

chamber

I'I

Ball valve-,

1 Quick connector-

wrench-.'

-

Gate va tve

'

-

/-Sight port ,-Opening chamber

'

I-

I

cutter

11 A-

,.-O-ring Test

-Tube

1

Oxygen supply joint

.-

Disti I lat ion chamber

EXPERIMENTAL

Apparatus. The product separation apparatus is shown in Figure 1. The apparatus is fabricated of stainless steel except for the distillation chamber, which is made of glass, All operations within the apparatus are carried out under a pressure of 7 X 10- torr or less. Wilson seals are employed to permit manipulation in vacuum of the handling devices (shown in Figure l), which are the tube cutter, movable rod, Allen wrench, and capsule holder. Capsule temperature is measured at the base of the capsule holder with a Chromel-Alumel thermocouple. Procedure. POTASSIUM - TANTALUM - OXYGEN REACTIONPRODUCT RECOVERY.A sealed tantalum capsule, 1.27 cm in diameter by 4.45 cm long, containing potassium (-20 ppm oxygen) and a tantalum wire specimen that was doped with a known amount of oxygen (in the range of 2000 to 4000 ppm) is heated at 982" C for 100 hours in a vacuum furnace (4). Following the reaction, a metal wire loop is spot welded

(1) A. P. Litman, U.S. At. Energy Comm. Rept. No. ORNL-3751 (1965). (2) C. Tyzack, "Advances in Materials," Pergamon Press, Oxford, 1964, Pp 151-58. (3) R. E. Clary, S. S. Blecherman, and J. E. Corliss, Pratt and Whitney Aircraft, Report No. TIM-850 (1965). (4) C. W. Hickam, Jr., Nat. Aeron. Space Admin., Technical Note D-4213 (1967).

996

ANALYTICAL CHEMISTRY

Figure 1. Product separation apparatus

to the top of the capsule, and the capsule is secured by means of a set screw in the capsule holder of the product separation apparatus. The apparatus is evacuated, and the capsule is outgassed by heating between 150" and 200" C for hour. The capsule is raised to the opening chamber and positioned in the jaws of the tube cutter. By rotating the holder 150" in each direction, an opening is made in the capsule without completely severing the lid. The intact lid acts as a potassium vapor deflector during the potassium distillation process. The capsule is lowered into the distillation chamber and heated from 3 to 5 hours, to a maximum temperature of 325" C (5). The vaporized potassium condenses on the inside walls of the air-cooled distillation chamber. Following distillation, the capsule is raised to the opening chamber, loosened in ( 5 ) R. F. Gahn and L. Rosenblum, ANAL.CHEM., 38,1014 (1966).

the holder, and lifted into the transfer chamber by the use of the hooked, movable rod. The transfer chamber, under vacuum, is transferred to an argon glove box, where the reaction product is removed from the capsule. Samples for chemical analysis are transferred into air-tight weighing bottles; samples for X-ray diffraction analysis are sealed into 0.5-mm diameter glass X-ray diffraction tubes. POTASSIUM-OXYGEN REACTION-PRODUCT RECOVERY. In the initial stage of this procedure, the product separation apparatus is modified by replacing the transfer chamber with a potassium extruder chamber (6). A metal crucible, 1.27 cm in diameter by 4.45 cm long, is placed in the capsule holder. The apparatus is evacuated, and the crucible is outgassed. Next, the crucible is raised into the extruder chamber, filled with about 1 gram of potassium by extrusion, and lowered into the distillation chamber. The potassium in the crucible is heated to 125" C, and the distillation chamber is valved off from the rest of the apparatus. A known amount of oxygen is admitted and allowed to react with the molten potassium. After the uptake of oxygen, the distillation chamber is opened to the vacuum system, and the potassium is distilled by heating to 325' C. Next, the extruder chamber is replaced with the transfer chamber. The crucible is raised to the opening chamber, loosened in the holder, and lifted into the transfer chamber. The transfer chamber is taken to the argon glove box for the product recovery.

Table I. Potassium-Tantalum-Oxygen Reaction-Product Chemical Analysis [Theoretical ratio of potassium to tantalum for K3Ta04,0.648.1

Capsule 1 2 30 40 5

(6) W. A. Dupraw, J. W. Graab, and R. F. Gahn, ANAL.CHEM., 36, 430 (1964). (7) A. Reisman, F. Holtzberg, M. Berkenblit, and M. Berry, J. Amer. Chem. SOC.,78, 4514 (1956). (8) E. Zintl, A. Harder, and B. Dauth, Ztschr. Elektrochem., 40,588 (1934).

6.62 7.97 6 . 15d 1.02d 4.70 4.30 4.73

4.00 5.50 4.25 0.675 3.06 2.80 3.00

Ratio of potassium to tantalum 0.604 0.690 0.691 0.662 0.651 0.651 0.634

Tantalum determined spectrophotometrically using pyrogallol. b Potassium determined by flame photometry. c Reaction product divided into two samples. d Average value of duplicate analysis on sample. a

hkl 111 200 220 31 1 222

RESULTS

A total of 70 alkali-metal reaction-product separations were made by the described procedure. Reaction products from the potassium-tantalum-oxygen, potassium-niobium-oxygen, lithium-tantalum-oxygen, lithium-niobium-oxygen, lithium-vanadium-oxygen, and potassium-oxygen systems were quantitatively isolated and identified by X-ray diffraction and chemical analysis. Two typical results are given as examples. Potassium-Tantalum-Oxygen Reaction-Product Recovery. The product recovered from the potassium-tantalum-oxygen reaction was potassium tantalate, K3TaO4,a crystalline, hygroscopic solid. The contents of five test capsules were chemically analyzed for potassium and tantalum, and the results are given in Table I. The average ratio of potassium to tantalum was calculated to be 0.655 d= 0.031 (Table I). The theoretical ratio of potassium to tantalum for the compound K3Ta04is 0.648. Identification of the product was further confirmed by a comparison of its X-ray diffraction pattern with that for K3Ta04cited in (7). The pertinent patterns are tabulated and discussed in (4). Potassium-Oxygen Reaction-Product Recovery. Two potassium-oxygen reaction tests were made. In the first test, X-ray diffraction analysis indicated that the reaction product was potassium monoxide, K 2 0 , by a comparison of its X-ray diffraction pattern with the reported pattern for KzO (8). These data are compared in Table 11. The second test, the reaction product was transferred to a weighing bottle in the argon glove box. After weighing, the reaction product was dissolved in water, titrated with standard acid, and analyzed for potassium by flame photometry. The chemical analysis results that follow indicate the product to be K 2 0 :

Weight of metal found, mg Tantaluma Potassiumb

400

331 420 422 511 440 531

1.866 1.613 1.480 1.443 1.316 1.241 1.140 1.091 1.074 l .019 0.9850 0.9717 0.9298 0.9036 0.8939 0.8614

m

1.850 1.604

S

1.434 1.309

m

S

vw m

1.136

W

m vw m+

mS

vw

m 1.069 W 620 mi1.016 W 533 vvw 622 m0.9683 vw 444 W 0.9274 vw 551 vvw 640 W0.8908 vw 642 m+ 0.8581 W 820 0.7780 mK20 prepared by reduction of mercuric oxide with potassium.

600

5

Table 11. Potassium-Oxygen ReactionProduct X-Ray Diffraction Analysis Reference 8a This work Interplanar Interplanar distance, d, distance, d, A (lo-* cm) Intensity5 8, (10-6 cm) Intensityb 3.74 W 3.72 m3.23 S 3.18 vs 2.29 vs 2.27 vs 1.953 W 1.943 vw

X-ray camera used had the radius of camera used in present work. For purposes of comparison, scale of intensity values used in (8) was converted to conform with scale in present work. Intensity code: s, strong; m, medium; w, weak; v, very.

Actual weight, mg K-0 reaction product. . . . . . . . . . . . . . . . . . . . . . . . . . K found by flame photometry. . . . . . . . . . . . . . . . . . K found by titration. . . . . . . . . . . . . . . . . . . . . . . . . . . Theoretical weight, mg K, if product is KzO.. . . . . . . . . . . . . . . . . . . . . . . . . . .

3.19 2.58 2.47 2.65

DISCUSSION

The results of the product-recovery tests serve to validate the techniques employed. The chemical and physical analyses indicate that the hygroscopic compounds K3Ta04and KzO were recovered in an uncontaminated condition. The reproducibility of the chemical analysis results for the potassium-tantalum-oxygen reaction product show that: the method used to open the capsule does not contaminate the VOL 40, NO. 6, MAY 1968

b

997

product, and potassium separation from the product was complete. A potential limitation of the separation method should be noted. The source of this limitation involves the time and temperature required for the distillation of the alkali metal and the changes in concentration that occur during distillation. Conceivably, certain reaction products that are initially formed under a given set of conditions might subsequently suffer alteration under the conditions of distillation. In the case of the potassium-oxygen reaction, there is a good probability that reaction-product alteration occurred during distillation. It was shown that potassium peroxide, Kz02, and potassium superoxide, KOz, are produced during the low-temperature and low-pressure oxidation of potassium (9). Heating of these oxides in the presence of excess potassium (as occurs during distillation) would lead to their conversion to the more stable monoxide, KzO (the product recovered in the tests reported herein). In the case of the potassium-tantalum-oxygen reaction, however, an alteration of the reaction product during distil-

lation is not expected because the distillation temperature is much lower than the reaction temperature. Support for the contention that no alteration of the reaction product had occurred was obtained in a separate test as follows. Following exposure to potassium at 982” C of a tantalum wire specimen containing 1150 ppm oxygen, the test capsule was opened, and the reaction product and the excess potassium were dissolved in butyl alcohol, rather than distilled. A tantalum analysis on the alcohol solution indicated 3.06 mg of tantalum present in the potassium. Analysis of the tantalum wire by X-ray diffraction indicated the complete depletion of oxygen; the lattice parameter of the 1150 ppm oxygen doped tantalum wire following exposure to the potassium was the same as for a pure tantalum wire specimen. If it is assumed that all of the depleted oxygen (1.10 mg) from the tantalum wire specimen is present as reaction product oxide, the tantalum to oxygen atom ratio of the product is 1 :4. This ratio corresponds to the tantalum-oxygen ratio for the compound K3Ta04.

(9) C. A. Kraus and E. F. Parmenter, J . Amer. Chem. Soc., 56, 2384 (1934).

RECEIVED for review December 11, 1967. Accepted February 29, 1968.

On the Conditions of Flash Pyrolysis of Polymers as Used in Pyrolysis-Gas Chromatography SIR: Pyrolysis-gas chromatography is a rapidly expanding and very promising method for ihe analysis of polymers ( I ) . Many applications have already been described, using widely different experimental conditions as shown by recent reviews (1,2). However, pyrolysis is not yet a sufficiently well understood phenomenon to allow determination of optimum conditions. This fact probably explains the large number of pyrolysis devices, the breadth of the temperature range used, and the lack of attention devoted to the effects of pyrolysis conditions. Most authors until now have assumed that pyrolysis of a sample takes place at the equilibrium temperature (usually 600” to 800” C) of the filament or the furnace used. Thus, careful attention is given to achieve reproducibility of this temperature; yet no attention is paid to the heating rate. It is the purpose of this letter to show that in most cases pyrolysis is completed at a temperature well below the equilibrium temperature of the heat source, and thus, that the heating rate is of paramount importance in flash pyrolysis. It should be noted that at the equilibrium temperature of the filament, the half-decomposition time of many polymers (Table I) is several orders of magnitude smaller than the time needed to reach this equilibrium temperature [usually 30 msec to 3 sec ( I ) ] . Admittedly, the data presented in Table I have been calculated by extrapolation from pyrolysis rate

(1) “Pyrolysis and Reaction Gas Chromatography,” G. Guiochon, Ed., Preston Abstracts, Evanston, Ill., to be published, 1968. (2) R. L. Levy, J. Gas Chromatog., 5 , 107 (1967).

998

ANALYTICAL CHEMISTRY

constants found in the literature ( 3 - 3 , far beyond the tem perature range at which these constants have been actually measured. Although the values of the half-decomposition time may be questioned, the results show that even with induction heating (@, which produces the fastest equilibration time, this time (about 30 msec) is longer at 600” C than the half-decomposition time for polytetrafluoro ethylene, which is one of the most stable industrial polymers (Table I). Thus, it seems quite improbable that decomposition actually takes place at the filament equilibrium temperature. Most, if not all, of the sample will be decomposed during the heating period. It is possible to determine whether this proposition is valid using a method derived to account for thermogravimetric results (5, 7, 8). As a first approximation, we shall assume that: the relationship between the rate constant and temperature can be extrapolated to a temperature range much higher than that at which the rate constant has been measured; the temperature of the heat source increases linearly with time during the heating period; and the temperatures of the heat source and of the polymer sample are always equal. Later we shall discuss the influence of deviations from these assumptions.

(3) S. L. Madorsky, “Thermal Degradation of Organic Polymers,” Wiley, New York, 1964, p. 61. (4) M. C. Anderson, J . Polymer Sci. C , 175 (1964). (5) F. Farre-Rius, and G. Guiochon, Bull. SOC. Chim. Frame 1965,455. (6) W. Simon, P. Kriemler, J. A. Voellmin, and H. Steiner, J. Gas Chromatog., 5, 5 3 (1967).