July, 1960
939
SOTES
inches long and 3,/4 inch in diameter. Although there were many cracks, there were clear areas as large as a inch cube which were shown to be single crystal. The Pt tube was easily repaired for reuse. The cracking might best be minimized or eliminated by slowly cooling the boule in a uniform temperature zone after the last part has crystallized in the temperature gradient. The growth of mixed fluorides in a hydrothermal bomb has been illustrated by the preparation in a sealed vessel of Cu compounds which normally dissociate or form oxides in air. K. Knox3 has grown crystals of KCuF3 and K2CuF4 from the melt. These crystals were very small and the KCuF3 invariably twinned. By analogy with the KFKiFZ4and KF-MgFt systems, it is expected that KCuF3 and K2CuF4will exist in the KF-CuF2 system, the former congruently and the latter incongruently melting. An additional factor in the Cu Fystem is the fact that copper(I1) fluoride16and presumably potassium-copper-fluorides, have an appreciable dissociation pressure of fluorine at the melting point, so that in an open system loss of fluorine and reduction of the copper occur. The copper system should be several hundred degrees lower than the KF-NiF2 and KF-MgF2 systems. Because of oxide formation and the dissociation problem, the system was investigated in a sealed vessel. KHF, \vas used instead of K F because it is less hygroscopic, and has a melting point much below the dissociation temperature of CuF2. The melting point of KHF2 is 195", a t which temperature it begins to liberate HF. Because of the build up of pressure when heating over 200" a steel bomb7 with a platinum liner was used. This autoclave was designed for hydrothermal experimentation and was constructed to withstand pressures up to about 7000 p.s.i. A corrosion resistant steel was used which can be heated to 800" without vreep. Reagent grade CuF2,2H,0was dried in dry HF at 400". ;Ibore this temperature decomposition takes place. ;I, typical run was made as follows: The bomb, xt7hic3h has a capacity of 30 ml., was charged with 30 g. of liquid KHF, prepared by heating the powder in a platinum crucible. After solidification, 10 g. of dry CuF2mas added. A 0.005 inch thick platinum disc was placed over the top of the liner and the steel plunger was forced d o m on the disc by tightening the head in a vise. The bomb was heated in a muffle furnace to 500" for 16 hours. The temperature then was lowered at a rate of 3" per hour to 200", at which temperature the bomb mas removed from the furnace. This mixture of 20 mole % CuF2and SO mole ?& KHF? yielded single crystal plates of KCuF3, about 0.5 cm.2. They could be separated by dissolving the K F matrix in warm HzO whirh had no effect on the KCuF3. Single crystal plate. of K2CuF4were grown using the 3ame procedure but changing the mole ratio to 139, CuF, and S7% KHF,. These crystals were the same size and habit as the KCuF3 crystals. In this case, the flux could not be removed with HsO be(3) K. Knox. J Chem P h y s , SO, 991 (1959). ( 4 ) G. Wagner and D. Balz. Z Elektrochem., 66, 576 (1952). ( 5 ) R . C. DeVries and R Roy, J A m . Chem. SOL., 1 6 , 2481 (1953). ( 6 ) H. v. Wartenberg, 2. anorg. Chem., 241, 381 (193'JI4 ( 7 ) G . W. Morey, Am. M h t 91, 1121 (1957).
cause the K2CuF4 hydrolyzed as evidenced by a change in color from clear colorless to opaque blue; however, the crystals could be separated mechanically without difficulty. The crystals were show1 to be KCuF3 and KZCuF4, respectively, by X-ray powder and single crystal pictures. Finally, single crystals of MnF,, ZnFz and KRInFa have been grown by zone melting in an inert atmosphere. -4significant reduction in the impurity content of the crystals, as well as quantitative impurity doping of the crystal, were accomplished by this method. The author is indebted to K. Knox for helpful discussions on KCuF3 and K2CuF4and J. W. Nielsen and E. Dearborn for useful discussions on thcx Stockbarger technique.
THE ASSOCIATION EQUILIBRIUM I K THE METHYL BROMIDE-ALUMINUM BROMIDE SYSTEM. ESTIMATED BOKDING STRENGTHS OF ALURIINUIJI BROMIDEADDITIOK JIOLECULES WITH METHYL BROMIDE, P E N T E S E AKD B E S Z E S E BY D. G. WALKER Research and Detelopment Dzizaion, Humble OaZ & Refining Compnnii Baytozcn, Texas Receaied January 28. 1960
Brown and Wallace1 have reported an excellent and thoroughgoing study of the addition compounds of aluminum halides with alkyl halides. They obtained experimental vapor pressure-composition data for the system methyl bromide-aluminum bromide at -80, -64.4, -45.8. -31.3 and 0' and concluded that the aluminum bromide in solution was present in the form of an additional molecule CH3Br:A1Br3. A redetermination of the vapor pressure-composition curve for methyl bromide-aluminum bromide a t 0' was made in the course of some other work. In the homogeneous liquid phase, excellent agreement was found with the results of Brown and Wallace. Also, in agreement with their results, the appearance of a solid phase was found to occur a t a CH3Br/AlBr3 liquid-phase ratio of 1.19 and at a system pressure of 235 mm. However, a welldefined pressure plateau was obtained between CH Br/A1Br3 ratios of 1.19 to zero whereas Brown and Wallace found no such pressure plateau in this region corresponding to compound formation in the solid phase. They concluded that solid solution phenomena must be present. Possibly they did not allow adequate time for the system to equilibrat? between withdrawals of methyl bromide. It seemed that if aluminum bromide was the solid precipitate, the system might be simple enough that considerable information of interest could br gained by a further study. Results and Discussion I n Fig. 1 is plotted the system vapor pressure versus composition obtained experimentally at 0'. If one awumes that only the addition mole{-
(1) H. C. Brown and W. J. Wallace, THISJOURNAL, 76,16279
(iwa).
SOTES
940
, 5OC) 1
400
- 1 1
c
-
z
I
a
I I
100'
-
I I l
-A'--------
0
1
2
LJp-2-
-
3 4 5 6 7 6 Composition Mol Ratio of CH3BrlQIEr3.
i ..
9
I
I
I
I
1 I
In Table I1 the calculated composition at four temperatures of the invariant liquid phase, saturated with AlzBr6, is given. The corresponding ideal solubility of &Br6 as calculated from
(h -
1180 -
2.7012 X
is shown in the last column of Table 11. The general agreement of this value with that found from the calculated equilibrium constants for (1) is an independent confirmation of the fact that the three molecular species of (1) must form a nearly ideal solution.
94
500 i l
I
have very nearly the same intermolecular attraction in the liquid phase as the two parent species and by no means should be thought of as an ionic or even a very polar compound. Data on the vapor pressure-composition of methyl bromide-aluminum bromide mixtures similar to that of Fig. 1 were obtained a t three other temperatures. The composition at which &Bra was observed to precipitate as well as the plateau pressure of the invariant liquid region was used to calculate an equilibrium constant for (1) at each experimental temperature (Table I).
m
Fig. 1.-Vapor pressure of the methyl bromide-aluminum bromide system a t 0'. P C H ~a tB 0" ~ = 659.5 mm. Experimental: -, theoretical assuming: (a) only C H a F and CH3Br:AlBr3 (non-volatile) are present; ( b ) Raoult s law is obeyed. - Theoretical assuming: (a) equilibrium = (AlzBrd1/2(CH3Br) 0.216; (b) AlzBre solubility = [CH3Br:A1Br3] 0.073 mole fraction; (c) Raoult's law.
r
Vol. 64
0Pent-2-ene
TABLE I EXPERIMENTAL DATAON C H 3 B r - A ~ u r n ~BROMIDE u~ MIXTURES EQUILIBRIUM CONSTANT AS A FUNCTION OF TEMPERATURE Vanor ----
I
100
O
I 50
I
I
I 10
0
Mol R a t i o Pentene/Al2Brg. Fig. 2.-Molecular weight of aluminum bromide in olefin solution (experimental work of Fairbrother and Field6). Calculated assuming: (1) only the equilibrium olefin-AlBr3 F! 1/z.412Br6 olefin exists; (2) Kepuil = 3 at 0'.
+
cule, CH3Br:A1Br3,and CH3Br exist in the liquid phase, that only CH3Br has an appreciable volatility and that Raoult's law is obeyed, a vapor pressure curve corresponding to the dotted line in Fig. 1 is predicted. As may be seen, the calculated line is in very good agreement with experiment for dilute solutions of CH3Br:AlBr3 in CH3Br and falls surprisingly close to the experimental values even in more concentrated solutions. The pressure plateau, which exists between CH3Br:AIBrs ratios of 1.19 to zero, shows that solid aluminum bromide first precipitates at st composition ratio of 1.19 and correspondingly an invariant liquid phase of CH3Br:A1Br3, CH3Br and AlzBr6exist in equilibrium with solid AhBrs. These data may be used to calculate an equilibrium constant a t this temperature (Fig. 1). &Br6 and CH3Br have identical solubility parameters (6 = 9.3) and would be expected to form an ideal solution. The fact that the addition molecule CHsBr/AlBr3 (6 unknown) forms nearly ideal solutions with methyl bromide (see Fig. 1) is extremely interesting. This addition molecule must
T. 'C.
pressure CHaBr, mm.
Compn. CHsBr/ AlBrra
1.08 5.3 804 1.19 0 659.5 1.30 8.1 505 1.30 -23.9 233 A t which A12Br8precipitates.
-
EXDtl. plat'eau pressure, mm.
Calod.
K
for eq. 1
280 235 170 67
0.256 .216 .137 .076
TABLE I1 COMPOSITION IN LIQUIDPHASE T, OC.
N(CHrBr: AlBrr)
N(C1l:Br)
N(A1zBrd
Ideal N(A1sBrs)
5.3 0 -8.1 23.8
0.510 .538 .602 .678
0.348 ,356 ,337 .288
0.141 .lo6 .060 .033
0.089 .073 ,054 .028
The slope of the usual type plot of the logarithm of K (from Table I) vs. 1/T gives a value of 5.69 kcal./mole for the heat of reaction of (l), and an entropy of reaction of +17.7 e.u. Thus in the liquid phase Fischer and Rahlfs2 found in the gas phase
By neglecting any heat of solution of monomeric AlBra in the liquid phase of ( 2 ) , equations 2 and 3 can be combined to yield (2)
W. Fiacher and 0. Rahlfs. 2. anorg. ollgcm.
(1932).
Cham., 206, 37
NOTES
July, 1960
941
THE GROWTH OF BARIUM TITANATE (4)
The bonding strengt,h of CH3Br:AlBr3 is then g l 9 kcal. Other aluminum bromide solutions may be interpreted in a similar manner. Kespita13 by the use of dipole measurements
SINGLE CRYSTALS FROM MOLTEN BARIUM FLUORIDE BY R. C. LINARES Bell Telephone Laboratones, Inc., Murray Hall, New Jersey Recetved January 86,1960
Single crystals of barium titanate have been of interest for ferroelectric studies. Most of these found that an addition molecule, - -A1Br3, crystals have been grown from molten salts, and at moderate temperatures (1000-1250°), because Baexisted in small concentration in benzene solutions TiOa undergoes a change from the cubic to hexaof aluminum bromide. U l r i ~ h , by ~ cryoscopic gonal form a t 1460°.1 Some of the solvents sucmeasurements in very dilute solutions of aluminum cessfully used include: BaC12,2KF,3PbF2,4PbO4 bromide in benzene, obtained results consistent and Na2C03.h Of these, the KF process has been with the conclusions of Nespital. Both of these the most widely used ; however, potassium from studies are consistent with the equilibrium the solvent and platinum from the crucible enter the crystals as impurities. The purpose of this paper is to report the solubility of BaTiOI in BaFL and point out that BaTiOs crystals can be grown from BaF2 solutions. Although this method has for which K = 700 @ 20'. This, combined with some drawbacks, which will be discussed, crystals an estimated (Sackur-Tetrode equation) A S = can be grown free of potassium and platinum. +17.9 e.u., yields an estimated bond strength of Solubility Curve -An approximate phase diagram for tho
Q
the addition compound
0
- - -AlBrs of 14.6 kcal.
The vapor pressure measurements of Fairbrother and Field" on solutions of aluminum bromide in cis-pent-2-ene a t 0' also can be interpreted as due to the existence of an addition molecule between an olefin and A1Br3 in equilibrium with AIzBrg and free olefin. I n Fig. 2 are plotted the apparent molecular weights of aluminum bromide in pentene solution as were experimentally determined by Fairbrother and Field. The dotted line represents calculated values of this same molecular weight function assuming that equilibrium (6) exists in the liquid phase Olefin---A1Br3
AllBraf olefin
(6)
and has a value of 3.0 a t 0'. Again combining an estimated A S = f17.5 kcal. with this value for the Kequilibriiim, a bond strength of 17.5 kcal. is estimated for the pentene-AlBr3 molecule. Experimental Apparatus.--..lll of the experiments were carried out with high vacuum apparatus and techniques in which the materials came in contact only with glass and mercury. Procedure.-A weighed amount of synthesized aluminum bromide was resublimed several times into a glass vessel under dry nitrogen. The vessel was attached to the vacuum system and evacuated. A measured amount of methyl bromide was condensed on the aluminum bromide and the vessel was agitated until solution was complete. Temperature control was by the use of appropriate slush baths. Pressure was read from a mercury manometer. The composition was varied by allowing fractional portions of the vapor to transfer to the vacuum system. Compositions reported as "CH~B~/AlBr3ratio" were corrected for the presence of CHaBr in the vapor space above the system. The vapor pressure of the methyl bromide recovered after a given experiment was identical t o that recorded before the experiment. The volume of gas recovered also was very nearly identical to that charged. (3) W. Nerrpital, Z . physak. Chsm., B16, 153 (1932).
(4) H. Ulrich. ibid., B18, 427 (1931). ( 5 ) F. Fairbrother and K. Field, J . Chem. SOC.,2614 (1956).
system BaFrBaTiOj was determined by a simple quenched melt technique. A charge of barium titanate and barium fluoride was heated in an electric furnace for four hours at a temperature 20' above the desired temperature. .4t the end of the time the temperature was dropped 20' and held for another two hours. Yext the temperature close to the crucible was determined with a Pt-Pt 10% Rh thermocouple probe, the crucible was removed from the furnace, and the melt was poured off and quenched in a platinum crucible Only those runs in which there was an excess of barium titanate in the crucible a t the time of pouring were used for the solubility determination The eutectic mixture was made by heating a mixtiire of 30 g. of BaTiOl and 60 g. of BaFZ to 1350' and cooling slowlv to 1200". The BaTi03 crystals were removed merhanicallv leaving the eutectic mixture and the melting point of this mixture was determined in a platinum wound micro furnace. The average of seven readings was taken as the melting point and was found to be 1260" with the maximum deviation being 14'. The composition of the melts was determined by analveiP for barium and titanium by an X-ray fluorescence techniqiie 6 The results are considered to be corrert to within 1% The molar composition was calculated on the assumption that barium and titanium were present onlv as RaF2 and RnTiOI Although X-ray pictures show only these two phases prcsent in low temperature runs, long runs (24 hr ) at high temperatures (1420') show a small amount of an iinidmtified third phase. Also if the weight per cent. of hariiim titanate and barium fluoride in these samples is ralculitcd, these total shghtlv more than 100% indicating the possible prewnrc of another phase. I t is possible that this nhmr is 't b?rium platinum oxide formed by slow decomposition of BnF2 to BaO which would attack the platinum rrucihle The proposed phase discrram for the svstem RaFn-RnTiO, is given in Fig. 1. This Rhows the solubility of RaTiOl to range from 16 5 mole yc a t t h e euteciir point of 1260' to 49 2 mole % at 1.393". Thp acciii-acv of thp ciirvc' has been confirmed bv observations made during crvstal growth bv pulling When using a melt of a givrn rompohition, crystallization began very near the temperature predicted by the solubility curve as drawn. The vapor pressure of BaF2 is quite low even a t elevatcti
(1) Phase Dirtgiarns for Ceraniists. 1956, p . 42. (2) Blattner, Matthias. Mera and Schemer, Helu. Clqrn. &!a, 20, 225 (1947) ; Ezperientia, 3, 4 (1947). (3) J. P. Remeika, J . A n . Chem. SOC.,76, 940 (1954). (4) J. W. Nielsen, personal communication. ( 5 ) Shoao Sawada, Choichiro Nomura and Shin'ichi Fugii, IZept. I n s t . Sci. Technol., Uniu. Tokyo, 6, 7 (19513. (6) T. C. Loomis, unpublished work.