THE: J O U R N A L
OF
PHYSICAL CHEMISTRY Registered i n U . 13. Patent Ofice
@ Copyright, 1964, by the American Chemical Society ~
~~~~~~~
VOLUME 68, NUMBER 5 MAY 15, 19614
The H.igh Temperature Vaporization Properties of Boron Carbide and the Heat of Sublimation of Boron'
by Harry E. Robson and Paul W. Gilles Department
of
Chemistry, University of Kansas, Lawrence, Kansas
(Received December 19, 1963)
--
-
The vaporization behavior of boron carbide solid solutions is shown to be the preferential loss of gaseous boron regardless of the sample composition. An invariant system consisting of a graphite crucible and ab carbon-saturated boron carbide sample was used to measure the vapor pressure over the temperature range 2184-2522OK. The Ihudsen technique employing condensation targets was used, and boron was specifically assayed by a coulometric titration of the mannitol complex. The derived third-law heats of sublimation in kcal./mole of boron are AH"o = 136.8 f 0.1 and AH0298 = 138.0 f 0.1 for boron carbide and AHoo = 133.8 f 0.7 and AHOzss = 135.0 f 0.7 for boron.
Introduction The high melting and boiling points, the great hardness, and the chemical inertness of the refractory borides cause them to be a unique group of substances. These very properties, however, make their thermodynamic properties difficult to obtain except by vaporization studies, which themselves must employ a reliable value for the heat of sublimation of boron. Elemental boron is so reactive a t high temperatures that values of its vapor pressure! and heat of sublimation obtained froim experimental studies with the element are subject to great errors arising from reactions between the sample and the other parts of the system. One of the purposes of the present work is to obtain the heat of sublimation of boron from a vaporization study of boron carbide. Searcy and I\iIyer~,~ Chupka,a Schissel and Williams, Akishin, Nikitin, and Gorokhovls Priselkov, Sapozhnikov, and Tseplyaeva,6 Alcock and Grieveson,' Verhaegen, Stafford,Ackerman, and Drowart18Verhaegen and
D r o ~ a r tPaule , ~ and Margrave, lo and Hildenbrand and Hall" all have sought to obtain the sublimation energy of boron, but the results are in disagreement. ~
(1) Abstracted in part from the Ph.D. Thesis of H. E. Robson, University of Kansas, 1958. (2) A. W. Searcy and C. E. Myers, J . Phus. Chem., 61, 957 (19517).
(3) W. A. Chupka. Argonne National Laboratory Report ANL-5667, 1957, p. 75. (4) P.Schissel and W. Williams, B d l . A m . Phys. SOC.,4 , 139 (1969). (5) P. A. Akishin, 0. T. Nikitin, and L. N. Gorokhov, Dokl. Akad. Naiik SSSR, 129, 1075 (1950). (6) Yu. A. Priselkov, Yu. A. Sapozhnikov, and A. V. Tseplyaeva, Izv. Akad. Nazrk SSSR, Otd. Tekhn. iVazLk Met. i Toplico, 134 (1960). (7) C. Alcock and P. Grieveson, "Thermodynamics of Nuclear Materials," International Atomic Energy Agency, Vienna, Austria, 1962, pp. 571, 572. (8) G . Verhaegen, F. E. Stafford, M. Ackerman, and J. Drowart, Nature, 193, 1280 (1962). (9) G. Verhaegen and J. Drowart, J . Chem. Phys., 37, 1367 (1962). (10) R. C. Paule and J . L. Margrave, J . Phus. Chem., 6 7 , 1368 (1963). ( 1 1 ) D. L. Hildenbrand and W. F. Hall, ibid., 68, 989 (1964).
983
HARRY E. ROBSON AND PAUL W. GILLES
984
In the present work, boron carbide, saturated with carbon and contained in graphite Knudsen cells, was vaporized under high vacuum; the sublimate was collected and analyzed chemically; the dissociation pressure and the heat of sublimation of boron from boron carbide were measured; and the heat of sublimation of elemental boron was calculated. Graphite crucibles were used to provide a fixed composition for the samples so that meaningful pressures could be obtained. Although the measurements were made several years ago, uncertainty has existed until the present concerning the nature of the gaseous species. Verhaegen, Staff ord, Ackerman, and Drowarts have recently shown mass spectrometrically that monatomic boron is the predominant species arising from the vaporization of boron from graphite Knudsen cells although BC2 and B& constitute about 5 and 1% of the vapor, respectively. The boron-carbon system contains only one firmly established intermediate phase which is usually described as B&, but it shows extensive solid solution on the boron-rich side of this composition. The crystals exhibit the space group D5~,-R>mand our measurements give hexagonal parameters, a. = 5.601 f 0.001 and co = 12.072 f 0.002 8.12-la
Experimental Materials. Some samples of carbon-saturated boron carbide were obtained through the courtesy of Dr. Gordon Fiiilay of the Norton Co. in the form of 12.5-mm. rods. Other samples of various compositions were synthesized from mixtures of previously outgassed amorphous boron and graphite. The boron was purchased from Fairrnount Chemical Co., which supplied the analysis 99.10% boron, 0.40% Fe, and 0.30% C ; the spectrographic graphite powder with negligible impurities was purchased from the National Carbon Co. These samples were prepared by sintering the mixed powders in vacuo for 0.1 to 2 hr. a t temperatures between 1600 and 2300’ in graphite crucibles about 20 mm. in diameter and 25 mm. tall which had been machined from spectroscopic grade graphite rod. The pressure during preparation was always less than mm. and usually below mm. These rather high pressures are attributable to gases present in the graphite. Apparatus. The apparatus for effusion studies is shown in Fig. 1. For the preparations, the coolant reservoir and target receiver sections of the apparatus were replaced by a simple cylindrical tube. The crucible assembly is shown in Fig. 2. The spectrographic graphite inner crucible, of about 16 mm. inside diameter and 16 mm. inside height, was contained in a closeThe Journal of Physical Chemistrg
L I Q U I D NITROGEN RESERVOIR
IONIZATION
TARGET MAGAZINE
ROD TRANSFER TRACK WATER JACKET
COOLING WATER I N HUTTER TEWERATURE WINDOW
S E A L T O PUMP ASSEMBLY
Figure 1. High vacuum effusion apparatus for vapor pressure measurements.
fitting tantalum outer crucible. The two crucibles sat on a tripod formed from tantalum wire. Tungsten shields on top of the crucible served to keep the lid a t least as warm as the walls of the crucible. The thinedged orifice was defined by the tantalum lid, and its size was measured carefully with a comparator both before and after a seriesof experiments. The ratio of inside surface area to orifice area was 200 or larger. Platinum targets were contained in individual cassettes stacked in the target magazine which was cooled by liquid nitrogen. Vapor escaping through the orifice in the lid of the crucible was condensed either on the walls of the apparatus or on the platinum targets. After one platinum target had been exposed to the molecular beam it could be ejected from its position and stored in the receiver on the left. As an exposed target was removed, the next target fell into receiving position: (12) F. Laves, Nachr. Ges. Wiss. Goettingen, Math-physik. K l . , Fachgruppe I V , 57, 1 (1934). (13) G. S. Zhdanov and N. G. Sevast’yanov. Russ. J . Phys. Chem., 17, 326 (1943). (14) H. K. Clark and J. L. Hoard, J . A m . Chem. Soc., 65, 2115 (1943). (15) F. L. Glaser, D. Moskowitz, and B. Post, J . Appl. Phys., 24, 731 (1953). (16) R. D. Allen, J . A m . Chem. Soc., 7 5 , 3582 (1953). (17) G. 8. Zhdanov, N. N. Zhuravlev, and L. S. Zevin, Dokl. Akad. Nauk SSSR, 92, 767 (1953). (18) G. S. Zhdanov, G. A . Meerson, N. N . Zhuravlev, and G . V. Samsonov, Zh. Fiz. K h i m . , 2 8 , 1076 (1954).
HIGHTEMPERATURE
VAPORIZATION PROPERTIES O F
BORONCARBIDE
-.
TUNGSTEN TOP SHIELDS ANTALUM CRUCIBLE GRAPHITE INNER CRUCIBLS
STAND-UP
I
J
Figure 2. Crucible assembly.
A magnetically operated shutter was opened and closed to define the time during which the target was exposed to the molecular beam. The platinum targets were 31.8 mm. in diameter and 0.13 mm. in thickness and were held in the stainless steel cassettes by stainless steel springs. The distance between the orifice and target was measured with a cathetometer. Definition of the beam was established by a collimator which also supported the cassettes. Its diameter was measured with a micrometer. An auxiliary experiment demonstrated that there was no reflection from the targets. Before being introduced into the system, the targets were cleaned in hot chromic acid, washed in distilled water, dried in a 100’ oven, and heated in air for about 10 min. a t 100OO in a separate induction heater. Heating was accomplished by high-frequency induction from a 25-kw. 450-kc. General Electric induction heater. The shutter a t the bottom protected the lightdeflecting prism from becoming coated with vapor. The pumping system consisted-of an oil diffusion pump, a mechanical pump, a thermocouple gage, and a Philips gage. The pumps and gages are below the apparatus shown in Fig. 1. Temperature iWeasurement. Because temperature errors would lead to errors in the final results, considerable care was taken in the measurement of the temperature
985
in the effusion experiments. A Leeds and Korthrup disappearing filament optical pyrometer was sighted through a calibrated window and prism into a blackbody hole in the bottom of the crucible. This pyrometer was compared with a similar one which had been standlardized a t the U. S. National Bureau of Standards. Both pyrometers had been stand?rdized at Argonne National Laboratory. Auxiliary experiments consisting of alternate measurements of the temperature in the blackbody hole and through the orifice established that no correction need be applied to the blackbody observation to obtain the corresponding orifice temperature. The calibration of the window and prism was made both before and after a series of experiments, but no significant change was found. Procedure. After the crucible assembly and target chamber had been properly placed, the system was pumped for 16 hr. or more without refrigerant in either the pump baffles or the top dewar. During this timle the pressure fell to about 10-6 mm. Liquid nitrogen was added to the dewar and a COz-acetone slush to the pump baffles. A warm-up period of several hours was allowed to keep the pressure below mm. When the pressure had fallen to 5 X lop6, exposures were begun by opening the shutter. Temperatures weire measured every 10 min. during exposure. After appropriate exposure time had elapsed, the shutter was closed, the target was ejected, and the procedure was repeated. At the completion of a series of measurlements, the targets were removed and assayed for boron. Analyses. Combustion analyses of boron carbide samples were kindly performed by Prof. Paul Arthur and Dr. Raymond Annino a t Oklahoma State University. Details are given in Dr. Annino’s dis~ertation.’~ The assay of boron collected on the platinum targets during the vaporization experiment was performed coulometrically, The targets were inductively heated for 1 min. as they rested in a clean 30-ml. silica beakier with the boron-containing surface down. Then 2 ml. of 1.1 N NaOH was added to the beaker to dissolve the resulting Bz03. That this procedure removed all the boron was substantiated by spark spectrographic analysis. After the target had been rinsed with distilled water, the solution was acidified with HC1 and was analyzed coulometrically between a platinum cathode and silver anode in the following manner. When a pH of 6.0 was attained, 5 ml. of saturated mannitol solution was added and the titration continued to a pH of 6.3. This value was determined by preliminary studies to be equivalent to a titration to the break in the pH curve between 7.0 and 7.5 followed by a blank correction. (19) R. Annino, P h . D . Thesis, Oklahoma State University, 1956.
Volume 68,Xumber 6 M a y , 1964
HARRY E. ROBSON AND PAUL W. GILLES
986
Eight successive experiments a t the same temperature, but for different lengths of time, show that the boron assay was proport,ional to the exposure time over a fourfold range and that the target weight gains, though less reliable, were approximately proportional to the boron assay.
Results Characterization of Vaporization Processes. Four qualitative experiments demonstrated that boron is preferentially vaporized even from carbon-saturated boron carbide, and that the vaporization reaction is
N o other phase was detected in the sample by X-ray diffraction, and spectrographic analysis did not indicate any impurities above the trace level. The results of 39 measurements are given in Table I in which the first column gives the series and exposure numbers, the second column gives the time interval of collection of the effusing vapor, the third gives the mass of boron in micrograms collected on the target as obtained by the analysis, and the fourth gives the corrected temperature of the sample. The fifth column gives the pressure and the sixth, its logarithm, calculated from the Knudsen equation
+
0.25B&(~)= 0.25C(~) B(g)
(1) In the first experiment, a Xorton Company boron carbide cylinder, 12.5 mm. in diameter and 12.5 mm. long, was heated to about 2200' for about 30 min. by radiation from an inductively heated tantalum susceptor. A tantalum lid on the tantalum susceptor gained weight and an X-ray pattern on the surface revealed substantial amounts of TaBz and smaller amounts of TaC. The sample exhibited a black coating which was too thin for X-ray identification. The second experiment consisted of a succession of heatings of a similar boron carbide cylinder followed by X-ray examination of the boron carbide surface which had been polished. The strong graphite X-ray diffraction line was observed to appear and to increase in intensity with progressive vaporization. Since there was no other source of carbon in the system, graphite appeared as the result of reaction 1. The third observation was made on a specimen of synthetic boron carbide from which 6% of the sample had been vaporized during a Knudsen experiment. An X-ray examination of the top surface showed a significant increase in the graphite content. The fourth experiment consisted of a preferential vaporization of a carbon-deficient safiple contained in a tantalum crucible that had been previously thoroughly borided by heating it with boron. During the course of the vaporization, the lattice parameters showed changes characteristic of an increase in carbon content. Some doubt concerning the nature of the vaporization process was indicated early in the investigation by the presence of boron carbide in an orifice following some vaporization experiments. The diffusion of carbon from the close inner graphite crucible probably is the source of the carbon needed to form the boron carbide, yet some carbon may have come directly from the sample by vapor transport as the molecules BZC and BC2. Vapor Pressure Measurements. Ten sets of effusion measurements were performed with synthetic boron carbide containing small amounts of excess graphite. The Journal of Physical Chemistry
p (atm.)
=
m(T/M)'/'/44.33Gat
(2)
in which m is the mass of boron, T is the temperature, M is the molecular weight of boron, a is the area of the orifice, t is the time duration of the exposure, and G is the geometry factor calculated in the usual manner from the dimensions of the apparatus. The pressure is calculated on the basis that only B(g) is important. The last two columns contain the change in the free-energy function and AHo, for reaction 1 from the equation AHoo = T [ ( - R In p ) - (AFO - A H o o ) / T ] (3)
The necessary free-energy functions for B(g) and C(s) were obtained from N.B.S. Circular 50O2O and values for boron carbide from a linear extrapolation of the results of Evans, Wagman, and ProsenZ1 according to the expression
6,
=
26.8
f
0.2
+ (2.0
f
0.2) X 10-'T
(4)
Except for the experiments of run 8 which are low in pressure by about a factor of about 1.8, the results from the different sets are in good agreement with one another. The pressure does not appear to vary appreciably with orifice area. Run 8 was performed with a previously unused graphite liner whereas all the others employed preconditioned graphite liners which had been exposed to boron vapor for long periods of time. The discrepancy between the results of run 8 and the others is attributed to undersaturation, and the results are not included in the final analysis. The results from run 2 , which are about 20% low, are subject to uncertainty because the targets were (20) F. D. Rossini, D. D. Wagman, W. H. Evans, S. Levine, a n d I. Jaffe, "Selected Values of Thermodynamic Properties," U. 5. National Bureau of Standards, Circular 500, Series 111, U. S.Govt. Printing Office, Washington, D. C., 1952. (21) W. H. Evans, D. D. Wagman, and E . J. Prosen, U. S.National Bureau of Standards, Report No. 4943, U. S.Govt. Printing Office, Washington, D. C., 1956.
HIGHTEMPERATURE VAPORIZATION PROPERTIES OF BORON CARBIDE
987
-. Table I : Vapor Pressure Measurements on Boron Carbide"
0.25BpC(s) = 0.25C(gr) Serieaexposure
2-13 2-14 2-15 2-16 2-17 3-1 3-2 3-7 3-8 3-10 3-11 3-12 4-13 4-14 5-15 5-16 5-17 5-18 5-19 5-20 5-21 5-22 5-23 7- 1 7-2 7-3 7-4 7-5 7- 6 8-1 8-2 8-3 8-4 10-7 10-8 10-9 10-10
rg.
t , sec.
7%
606 1202 2401 4806 2399 1203 2400 1204 2404 9601 4800 2405 1204 2401 1231 1200 1200 2400 3601 4817 3601 2401 1200 3600 7200 3610 1.800 1800 1800 6000 3600 2400 2400 2400 3600 6004 3604
11.3 16.5 33.0 62.0 30.4 9.4 18.7 8.4 18.4 80.5 41.9 20.3 23.8 48.4 28.1 23.5 21.7 48.5 68.9 87.3 65.6 46.0 21.9 22.1 3.2 57.8 49.6 87.5 177.8 14.7 20.2 39.3 68.4 19.5 18.4 15.5 25.5
T. OKIc.
2303 2302 2308 2306 2300 2299 2296 2294 2297 2300 2299 2298 2392 2389 2390 2395 2392 2391 2392 2392 2396 2396 2395 2308 2199 2376 2421 2464 2522 2270 2343 2405 2483 2283 2236 2184 2261
+ B(g)
P, 10-0 atm.
6.31 4.64 4.65 4.37 4.28 5.88 5.86 5.25 5.76 6.31 6.57 6.35 15.52 15.82 18.20 15.63 14.42 16.11 15.26 14.45 14.10 15.29 14.56 4.79 0.34 12.67 22.02 39.18 80.55 1.75 4.07 12.05 21.30 2.90 1.80 0,900 2.51
-(AF'
log P
-5.200 -5.334 -5.332 -5.359 -5.368 -5.230 -5.232 -5.280 -5.240 -5.200 -5.182 -5.197 -4.809 -4.801 -4.740 -4.806 -4.841 -4.793 -4.816 -4.840 -4.837 -4.815 -4.837 -5 319 -6.469 -4,897 -4.657 -4.407 -4.094 -5.757 -5.390 -4.919 -4.672 - 5.538 -5.745 -6 046 -5.600
- AH"~)/T, e.u.
35.103 35.102 35.103 35.102 35.103 35.103 35.101 35.102 35.102 35.102 35,102 35.102 35.102 35,102 35.103 35.102 35. io2 35.1'04 35.102 35.102 35,102 35.102 35.102 35,102 35.103 35.102 35.101 35.099 35.093 35.102 35.102 35.102 35.099 35.103 35.102 35.101 35.101
AH kcal.
135.6 137.0 137.3 137.5 137.2 135.7 135.6 136.0 135.7 135.5 135.2 135.3 136.5 136,3 135.7 136.7 136.9 136.3 136.6 136.9 137.1 136.8 137.0 137.2 142.3 136.6 136.5 136.1 135.7 139.4 140.0 138.5 140.2 137.9 137.2 137.0 137.2
0 The orifice areas in cm.2, orifice to collimator distance in om., collimator radius in cm., and geometry factors were, respectively: run 2: 0.04722, 0.28, 1.346, 0.02061; run 3: 0.02307, 9.69, 1.346, 0.01893; run 4: 0.02307, 9.85, 1.353, 0.01852; run 5: 0.02312, 9:94, 1.353, 0.01819; run 7: 0.02285, 9.86, 1.353, 0.01848; run 8: 0.02141, 9.16, 1.353, 0.02135; run 10: 0.04490, 9.36, 1.353, 0.02047.
sparked in the emission apparatus before the chemicisl analysis was performed. The quantities of boron in runs 3, 4, and exposures 5-15 and 5-16, though in good agreement with others, represent the sum of two or more titrations. The final treatment uses the sixteen data from runs 5-17 to 5-23, 7 (except for 7-2), and 10. The linear second-law values for the enthalpy and entropy of reaction 1 at 2300'K. are AH = 146.8 f 1.7 kcal./mole of boron and AS = 39.2 f 0.7 e.u./ mole of boron. From the same set of sixteen data,
the average third-law value for reaction 1 is AHo, = 136.8 k 0.1 kcal./mole of boron and = 138.0 f 0.1 kcal./mole of boron. We take the third-law result to be the more reliable. To obtain the heat of sublimation of boron, one needs the heat of formation of BIG. A value of -12.2 f 2.2 was obtained by Smith, Dworkin, and Van Artsdalen.22 The heat of sublimation of boron calou~~
(22) D. Smith, A. 9.Dworkin, and E. R. Van Artsdalen, J . A m . Chem. Soc., 7 7 , 2664 (1955).
Volume 68, Number 5
M a y , 396Q
988
lated from the third-law result and this value of 3.0 per mole of boron is 133.8 f 0.7 kcal./mole at 0°K. and 135.0 f 0.7 a t 298°K.
Discussion The present work firmly establishes the vaporization behavior of boron carbide to be the preferential loss of boron to the vapor phase regardless of the composition of the sample and gives its vapor pressure from 2184 to 2522°K. This work is the only vapor pressure endeavor in which boron is collected and specifically assayed. The interpretation of the vapor pressure measurements on the basis of reaction 1 is justified by the proven nature of the vaporization reaction and by the absence of gaseous molecules of great importance. The few per cent of BC2 and BzC8and of B29 in the vapor are of negligible importance in the third-law calculation of the heat of sublimation of boron. A discrepancy exists between the second-law and the third-law values for the heat of sublimation, the former being higher by about 10 kcal. A slight trend with temperature of the third-law heats amounting to about 1.2 kcal. also exists and is consistent with the difference between the second- and the thirdlaw values. That is, the second-law plot is too steep; the second-law entropy is too high; and the high temperature points give third-law heats slightly lower than the low temperature points. Several possible errors or conditions might have contributed to the discrepancy. The temperature error required to take the trend out of the third-law values and to make the secondand third-law values agree is about 30" more a t one end of the range than a t the other, but such a progressive error of this magnitude is not likely in view of the care taken in the experiments and the calibrations. Errors in the estimated free energy functions which could amount to 1 e.u. would cause an error in the third-law enthalpy of sublimation of about 2.3 kcal., but could not remove the trend and would not, of course, alter the second-law result. Errors in the boron analysis would cause errors in the derived pressures and are most likely to occur in the low temperature region. Some difficulty with the target analyses was encountered, but errors in the blank corrections or the titers large enough to force agreement between the second- and the third-law results are not likely. The onset of another vaporization reaction a t the highest temperatures could cause the discrepancies. The proportions of Bz(g), BZC(g), and BC2(g), however, seem too small to be of importance. The Journal of Physical Chemistry
HARRYE. ROBSON AND PAUL W. GILLES
The crucibles used for these experiments had surface to orifice area ratios of about 200 and 400. No appreciable effect attributable to orifice size was found, and hence the pressures are close to the equilibrium values. The points of run 10 were taken with the large orifice and appear to be slightly low, but only by 7-10% instead of the factor of two required for an extremely low vaporization coefficient. Though in the correct direction, undersaturation, if present, cannot remove much of the discrepancy. Finally, an enhanced, abnormal transport 6f boron a t the highest temperatures, possibly caused by reaction with the residual gases in the system, could account for the discrepancies. Such an increased transport would amount to about 35% at the highest temperature and progressively less a t lower temperatures. Xone of the foregoing can be selected with certainty as the source of the discrepancies, but the last seems the most probable. Most of them would have a greater effect on the second-law values than on the third-law values, and hence we have chosen the average of the latter results for the sixteen best points. The heat of sublimation of boron a t 298°K. of 135.0 obtained in the present work clearly depends on the heat of formation of B4C, a quantity which has been measured but once. Our result for the heat of sublimation is in excellent agreement with the value of 135.7 obtained from torsion effusion vapor pressure measurements on boron carbide a t about the same temperatures as ours by Hildenbrand and Hall" and the value of 136.9 obtained by Paule and Margravelo from Langmuir vaporization measurements on boron a t lower temperatures. It is in good agreement with the value of 139 obtained by Searcy and ;Llyers2 from Knudsen measurements on boron. The agreement among these different types of experiments indicates the lack of serious errors. Except for the results of Akishin, et u Z . , ~ whose range includes practically all the values, the results obtained by mass spectrometric studies are lower by several kcal. and are in serious disagreement with the direct vapor pressure measurements. The most likely source of error in the mass spectrometric studies is the temperature measurement. The temperature in a crucible heated by electron bombardment in a mass spectrometer may not be well defined and is surely considerably harder to measure than in a typical effusion vaporization study. Buchler and Berkowita-h.lattu~k~~ have demonstrated the extreme care which must be taken (23) A. Buchler and J. B. Berkowitz-Mattuck, J . Chenz. Phys.. 39, 286 (1963).
DECOMPOSITION PRESSURE OF BORON CARBIDE
989
-
-.
to ensure that good temperature coefficients are obtained from mass spectrometric measurements. Another likely source of error in the mass spectrometric measurements is in the absolute pressure measurement. One possible interpretation, and the most reasonable one in the authors' view, is that the mass spectrometric measuriements are in error because of tempemture errors or pressure errors, that the vaporizatioin coefficient of boron is nearly unity, that the vaporiza,tion coefficient, of boron from boron carbide is greater than the value of about 0.005 which is the reciprocal of the surface to orifice ratio in the present work, and that it is perhaps about the 0.07 value that Hildenbrand and Hall" suggest. On the other hand, the
apparent agreement among the mass spectrometric values on one value and the agreement among the direct vapor pressure values for another value suggests that some unusual factor, as yet unrecognized, maty actually exist. Acknowledgment. The authors are pleased to acknowledge the support of the U. S. Atomic Energy Commission under Contract AT(l1-1)-1140 and its predecessor, AT(11-1)-83, Project Yo. 1; the American Oil Co. for a graduate fellowship held by H. E. Robsoin ; and the U. S. Yavy in its boron program. We a150 wish to thank Dr. Gordon Finlay of the Norton Company for some of the boron carbide samples and Prof. Paul Arthur and Dr. Raymond Annino for tlhe combustion analyses.
The Decomposition Pressure of Boron Carbide and the Heat of Sublimation of Boron'
by D. L. Hildenbrand and W. F. Hall Research Laboratories, Philco Corporation, Newport Beach, California
(Received December 19, 1968)
The boron decomposition pressure over boron carbide, B,C, has been measured by the torsion-effusion method in the range 2350 to 2615OK. A small "hole-size effect" indicates a condensation coefficient of rloughly 0.07 for B(g) on B4C(s). The heat of dissociation of B4C to B(g) and C(graphite:l at 298OK. has been derived as 138.7 f 1.2 kcal./mole of boron from a third-law analysis of the vaporization data. From this result, the heat of sublimation of boron a t 298OK. has been evaluated as 135.7 f 1.3kcal./mole. ,4n approximate value of 171 kcal./mole has been obtained for the heat of formation of BC2(g) a t 298OK.
In spite of a number of recent determinations, there remains a relatively large uncertainty in the heat of sublimation of elemental boron. This uncertainty affects all bond-energy calculations involving boron compounds and all equilibrium calculations involving gaseous boron. All of the determinations to date2--'0 are based on vapor pressure measure-
(1) This work was supported by the Advanced Research Projects Agency and Bureau of Naval Weapons under Contract NOW 610905-c. (2) A. W. Seamy and C . E. Myers, J . Phys. Chcm., 61, 957 (1957). (3) H . E. Rohson and P. W. Gilles, ibid., 68, 983 (1964). (4) 1E. G . Paule and J. L. Margrave, ibid., 6 7 , 1368 (1963). (5) Yu. A. Priselkov. Y . A. Saposhnikov. and A. V. Tseplyaeva, Izu. Akad. h7auk SSSR, Otd. Tekhn. Naitk, Met. i Toplivo, 134 (1960).
Volume 68, Number 5
M a y , 1964
