which is based on adding water by diffusion through a capillary t o a dry stream of gas. The results compared to calculated values were not satisfactory, because an all-glass diffusion cell was used without special temperature control. This method did appear promising; and under favorable conditions satisfactory operation is expected. The range can be adjusted from 0.1 to 1000 p.p.m. The authors recommend empirical calibration rather than calculation. Propylene. A stream containing 30% propylene in the presence of saturated hydrocarbon gases was monitored for a week without difficulty. R a t e r was measured down to 5 p.p.m. The background correction corresponded to the values in Table I and did not change over the period of measurement. The absorption efficiency was not checked, but t,he
efficiency could conceivably decrease with time because of polymer formation on the electrode surface. No indication of polymer formation was observed. Hydrogen. Recombination of hydrogen and oxygen t o produce water complicates the application to hydrogen gas. The use of the variable flow method for successful determination of moisture in hydrogen gas has been described ( 2 ) . Experience obtained here has substantiated this variable flow method. Background corrections obtained by various methods are shown in Table I. These values are the sum of IL and I R in Equation 1. The hydrogen gas was dried with Linde Molecular Sieve 5A. No response to flow rate was obtained with hydrogen dried in this way, even though the cell reading was very high. Linear response to flow rate was obtained with both cata-
lytic reformer recycle gas (76% hydrogen) and pure hydrogen a t flow rates below 200 cc. per minute. LITERATURE CITED
(1) Carlstrom, A. A., S encer, C. F., Johnson, J. F., ANAL. 32, 1056 (1960). ( 2 ) Consolidated Electronics Corp., 360
HEM.
Sierra Madre Villa, Pasadena, Calif., "CEC Recordings," XIV, No. IV,
p. 19: (3) Keidel, F. A,, APU'AL. CHEM.31, 2043 (1959). ( 4 ) .Marsh, R. F., private conimunica-
tion.
(5) McKelvev. J. M.. Hoelschcr.' H. E.. ANAL.CH~M. 29, 123 (1987). ( 6 ) Taylor, E. S., Refrig. Eng. 64, No. 7, 41 (1956). \
I
~~
FREDERICK BAUMANN California Research Corp. Richmond, Calif.
Highly Sensitive Low Temperature Fractionation Method for Separation and Measurement of Boron Hydrides SIR: A highly sensitive method capable of detecting quantities of the boron hydrides, B4H10 and B5H11, as small as 0.02 pmole was required for the analysis of the products of a very low conversion kinetic study of the photolysis of diborane ( 7 ) . Various analytical methods have been described in the literature for boron hydrides: infrared absorption ( 2 , 9, I O ) , neutron absorption (S), gas chromatography (b), low temperature fractionation (4, I I ) , and mass spectroscopy ( I ) . Neutron absorption was used to determine total Bl0 content of individual boron hydrides for use in isotopic exchange studies. Gas chromatography was used successfully for B ~ H Iand o B5H9, though not for BSHll because of the latter's instability. Minimum detectable amounts for mixtures were given only for the infrared absorption method, where they were about 0.5 pmole of B4HIO and B5Hl1,so that the method n-as too insensitive for much of our investigation. A method was needed where a minimum detectable amount in the presence of large amounts of B2H6 was as low as 5 x 10-4 mole per cent. It is necessary in certain studies with boron hydrides to proceed to very low per cent conversions to avoid complications due to the reactions of the products. The products were separated in a low temperature LeRoy fractionation column (8) and measured in a special gas buret. The principal components in the analytical scheme were the LeRoy column and a double-range gas buret capable of measuring up to 200 pmoles. Its collection arm was sealed, instead of containing the usual greased stop-
cock, to eliminate solubility of the boron hydrides in the grease under pressure. A measurement error of about 0.2 cm. in the gas buret corresponded to 0.001 pmole, and the smallest amount required by the experimentation was 0.02 pmole. A LeRoy column had previously been applied to the determination of B4H10and BsH9 (4) in considerably larger quantities. The procedure evolved after detailed vapor pressure measurements of the pure components and of mixtures: Hz from the photolysis was first separated from the boron hydrides at -196" C. The major portion of the BzH6was next removed a t -136" C. in a simple trapto-trap distillation. The boron hydrides were then collected in the LeRoy column a t -196' C., and the temperature was slowly raised to -154" C. After the pressure had become constant, the was removed by trapping a t -196" C. until a pressure of about 0.5 micron was achieved. This procedure was repeated for and BsH11, whose separation temperatures were -120" C. and -74" C., respectively. The most critical separation,
Table 1.
B4HIO from BsHll a t - 120" C., was performed a t the very low pressure of about 10 microns to obtain the maximum separation efficiency. After being collected a t - 196' C. separately, B4H1" and &HI1 were transferred to the gas buret using a Toepler pump and by direct expansion, respectively. About 250 analyses were made, and the amounts of these higher boron hydrides varied from 0.02 to 34 pmoles. Several experiments were designed to determine the effectiveness of the separation and possible loss due to internal reactions with the analytical system which contained greased stopcocks. By unusually extensive conditioning (6, 7 ) of the apparatus with and B5H11 and by keeping the pressure less than about 10 microns during gas transfers, this loss was minimized. Known mixtures from the separated and purified products of a photolysis of diborane were analyzed (Table I). Further indirect evidence for the accuracy and precision of the detection
Analysis of Known Mixtures of B4H10 and B6Hll
Kn?m
Analyzed mix. (Mmoles) 0.92 1.29 5.16 4.33
mix.
Expt. No.
Product
(pmoles)
ST2 ST2 ST1 ST1
B4H10a BsHii' B4HlO
0.86 1.32 5.18 4.46
BSHII
Diff. (mole)
$0.06 -0.04 -0.02 -0.13
70 $7
-3 -0.5 -3
These products were identified by means of their infrared absorption spectra (9).
VOL. 34, NO. 12, NOVEMBER 1962
0
1665
of very small amounts of BcHIO can be found in the two plots of BIHIOvs. time of photolysis for a diborane pressure of 0.08 cm. These are depicted elsewhere (6). The B,HI0 varied between 0.02 and 0.15 pmole, and the two curves n-ere straight lines which extrapolated to within 0.02 pmole of the origin. LITERATURE CITED
( 1 ) Bragg, J. K., hlcCarty, L. V., Korton, F. J., J. Am. Chem. SOC.73,2134 (1951).
(2) Burwasser, H., Pease, R. N., J . Phys. Chem. 60, 1589 (1956). (3) Hamlen, R. P., Koski, W. S., ANAL. CHEM.28, 1631 (1956). (4) Hirata, T., Gunning, H. E., J . Chem. Phys. 27, 477 (1957). (5) Kaufman, J. J., Todd, J. E., Koski, W7.S., ANAL.CHEM.29, 1032 (1957). (6) Kreye, W. C., Ph.D. thesis, Polytechnic Institute of Brooklyn, 1960. (7) Kreye, W. C., Marcus, R. A., J . Chem. Phys. 37, 419 (1962). (8) LeRoy, D. J., Can. J . Research, B28, 492 (1950). (9) McCarty, L. V., Smith, G. S., McDonald, R. S., ANAL.CHEM.26, 1027 (1954).
(10) Stenart, R. D., .Idler, R. G , 134th
Meeting, ACS, Chicago, September 1958.
(11) Stock, A , , “Hydrides of Boron and Silicon,” Cornell University Press. Yew
York. 1933.
W. C. I i m m R. A . .\liacus
Department of Chemistry Polytechnic Institute of Brooklyn Brooklyn 1, S. Y. Work supported by grants from the Allied Dye and Chemical Corp. and from the Office of Naval Research.
Karl Fischer Reagent for Determination of and Differentiation between Trialkyl (Aryl) Organotin Hydroxides and Corresponding Oxides SIR: A rapid and satisfactory method has been developed for the quantitative determination of trialkyl and triaryl organotin oxides and hydroxides using the Karl Fischer reagent. It is extremely useful in differentiating beta-een members of this class of organotin compounds where both the oxide and the hydroxide exist in stable forms. The reagent has been employed in a number of cases and the determination is general for this class of compound. Mitchell and Smith (b), who discuss the applications and limitations of the Karl Fischer reagent a t length, state that the quantitative reaction of this reagent with inorganic oxides and hydroxides is fairly general in nature and, therefore, serves as the basis for a method of analysis. The following reactions were proposed by Mitchell and Smith (2), ignoring the salt-forming function of pyridine in the reagent to simplify formulation:
-
+ Ia+ SO2 + CHaOH ZnIl + HSOaCHa NaOH + Iz + SO2 + CHIOH KaI + HI + HS04CH3 ZnO
(R3Sn)ZO (1)
+
Table 1.
Gilman and Miller (1) referred to similar reactions with organic silanols and silanediols. While attempting to determine the water content of these materials using the Karl Fischer reagent, they found that consistently high results were obtained. Investigation led to the conclusion that not only was the water content being determined, but also that the silanol was reacting quantitatively with the reagent. The method was then extended by them to other organometallics-namely, triphenyl lead hydroxide, triphenyltin hydroxide, and phenyl boric oxide. We have found that the Karl Fischer reagent not only is effective in the quantitative determination of triaryl and trialkyltin hydroxides, but is also applicable to bis(trialky1tin)oxides and bis(triary1tin)oxides. Reactions analogous to those proposed by Mitchell and Smith (2) can be postulated.
(2)
+
+ IZ + SO2 + CHIOH
+ H S O ~ ~ H(3)S R3SnOH + 1, + SO2 + CHaOH R3SnI + HS04CHs HI (4) 2R3SnI
Comparative Analytical Data for Triphenyltin Hydroxide and Bis(tripheny1tin)oxide
Triphenyltin hydroxide Calcd. Found
...
32.37 58.69, 58.74 4.52; 4.45 118.5-120
4.90 1 .oo
4.83 0.984
32.36 58.90 4.39
HzO (apparent), Ip/Sn atom, moles%
(decomp.)
Bis(tripheny1tin)oxide Calcd. Found 33.15 60.37 4.22
...
33.10, 33.20 60.22, 60.17 4.40; 4 . 3 3 122-123.5
2.52 0.500
2.60, 2.62 0.506
The similarity between Equations 1 and 3 and 2 and 4 is apparent. Although in each reaction l mole of iodine is consumed for each mole of organotin compound, it is helpful to base the iodine consumption on the number of tin atoms. Thus, the RdnOH class of compounds consumes 1 mole of iodine for each tin atom; whereas with the (R3Sn)ZO type of compound, the ratio is 0.5. As the Karl Fischer procedure is already employed routinely by our Analytical Group for water determinations, a method of interpretation of data in the form normally reportedLe., as per cent HzO was worked out. A theoretical figure for apparent water was arrived at by employing the following formula:
% HzO (apparent) = moles of 1 2 er male of compound (theo.) X 18 ?mol. wt. of H20) X 100 mol. wt. of compound In the case of triphenyltin hydroxide: HzO (apparent)
=
1 X 1800 367 ~
0
ANALYTICAL CHEMISTRY
4.90%
while in the case of bis(tripheny1tin)oxide
The purity of the compound can then be determined by comparing the per cent of water found (apparent) with the theoretical value. Conversely, a value for per cent HzO can easily be converted to moles of Iz per atom of tin for compounds of the type RanOH as follows: Moles of 1 2 = mol. wt. of compound X % HIO (found) 1800
1666
=
