An Examination of the Exchange of Deuterium from Deuterium Oxide in Carbon Tetrachloride Solution with Hydrogen in Glass Using Infrared Spectrometry Paul K. Glasoe and Charles N. Bush Wittenberg University, Springfield, Ohio 45501
IN WORKING on the infrared spectra of solutions of D20 in CCl,, it was extremely difficult to get spectra which did not contain some evidence of the presence of hydrogen in the form of an absorption band at 3661 cm-l, due to the 0-H stretch of HOD. Such a solution showed strong‘bands for 0-D stretch at 2752 and 2643 cm-I and a slight shoulder at 2692 cm-l. This latter band is attributed to 0-D stretch in H O D and is prominently present in the spectrum of a solution of H 2 0in CC14containing a small amount of DzO,along with the strong 0-H stretch bands at 3706 and 3612 cm-l.(I). Even though the thoroughly dried CC1, used for the DzO solution showed no evidence of any absorption in the 0-H region, the addition of some D 2 0produced some HOD band at 3661 cm-’. This hydrogen could have been present as impurity in the stock DsO or could have been introduced in the process of transferring solutions to the infrared cell. However, after prolonged exposure of the cell to a solution of D20, it was possible to get a spectrum from a fresh sample of stock solution which showed no sign of absorption at 3661 cm-1. It seemed possible that the hydrogen was introduced by exchange of deuterium from the DzO with hydrogen from the glass surface. To test this, some ground glass was introduced into the cell containing a D20-CCl4 solution and the rate of formation of HOD was measured. EXPERIMENTAL
Carbon tetrachloride was purified by distillation from Drierite and was dried by storing it over Drierite. Extensive investigation showed that no foreign material was added to the CCll from the Drierite. The IR spectrum of this CC1, was checked against material specially purified and was identical. The spectrum of the dried CCll showed no absorbance in the 4000 to 3500 cm-’ range in a cell with a path length of 25 mm. The D20 was obtained from General Dynamics as >99.8z D 2 0 and was used without any treatment. Solutions of H20 or D 2 0 in CC1, were obtained by placing an aqueous layer on top of the dried CCI, and allowing it to come to equilibrium over a period of several days. The concentration of H 2 0in a saturated solution in CC1, at 25 ‘C mole per liter, and the concentration of the is 8.6 X saturated solution of D 2 0 is 7.5 X 10- mole per liter. The two-phase systems were kept in glass stoppered flasks in a dry box equipped for periodic circulation of the air through columns of molecular sieve and temperature control at 25 i 0.1 “C. When 25-mm cells which had been stored in the dry box were filled with dry CCl4, the spectrum showed no OH absorption, indicating a very low relative humidity in the dry box. The cells used were Barnes Engineering Model 902-0040, 25-mm path length with CaF2 windows. The body of the cell was borosilicate glass and was equipped with a ground glass stopper. The CC1, solutions were removed from under (1) P. Saumagne and M. L. Josien, Bull. SOC.Chim. Fr. Mem., Ser. 5, 813-19 (1958).
the aqueous layer by means of a glass syringe with a stainless steel hypodermic needle. All transfers of solutions to the cells were carried out in the dry box. The spectra were obtained on a Perkin-Elmer Model 237B, with an auxiliary recorder on which both the absorbance ordinate and the wave number scale could be expanded up to lox. The spectra of the D20 or H 2 0 solutions in cc14 were compensated spectra with dry CC14 in the reference beam. in the region 4000 cm- to 2500 cm-l, the transmittance of a 25-mm path length of CCl, is large enough that no problem of energy was encountered. A piece of ordinary “soft” glass tubing was ground up in a mortar and the very fine material was removed by sifting it through a 200-mesh sieve. The remaining material was crudely analyzed for particle size by sifting it through a series of standard sieves. The range of particle sizes was 0.04 cm to 0.015 cm with an average size of 0.028 cm. A 1.0-gram sample of the ground glass was placed in the infrared sample cell, and the cell was filled with a saturated solution of H 2 0 in CC14. The IR spectrum was obtained and found to be identical with that obtained in the absence of the glass. The ground glass was kept on the bottom of the cell so that it was not in the IR beam. This solution was left in contact with the glass in the cell for several hours. Then the solution was poured out of the cell, with care to keep the glass in the cell. The cell plus glass was rinsed 3 times with dry CC14 and filled with dry CC1+ The IR spectrum showed no absorption in the hydrogen region. The cell was emptied, still retaining the glass in the cell, and filled with a saturated solution of D20 in CC14. The rate of increase of the absorbance of the HOD band at 3661 cm-’ was measured. After equilibrium had been obtained, the cell was emptied, the cell and ground glass were rinsed thoroughly with dry CCl4 and filled with dry CCla. The spectrum was checked in the deuterium band region to make sure that all the D20had been rinsed out of the cell. The dry ccl4 was emptied out and the cell filled with a saturated solution of H 2 0 in CC14. The rate of increase in the absorbance of the deuterium band at 2692 cm-1 was followed. After the attainment of equilibrium, this solution was removed and the process repeated, always keeping the same glass sample in the cell. RESULTS AND DISCUSSION
The absorbances of the HOD band at 3661 cm-l after various times of exposure of the “hydrogenated” ground glass to a solution of DzO in CCI, and those for the HOD band at 2692 cm-’ in the solution of H 2 0in CC1, are given in Figure 1. The concentrations of HOD to which these absorbances correspond were read from calibration curves made in this laboratory on solutions containing a large excess of either H 2 0 or D20, in which it can be considered that all of the H or D in smaller amount is in the form of HOD. The absorbances being measured are very small and are subject to a considerable experimental error. Duplicate spectra on the same solution gave absorbances with an average deviation of 10.004. This is equivalent to an ANALYTICAL CHEMISTRY, VOL. 44, NO. 4, APRIL 1972
833
5
4
*2
3
X
P
*
0
I:
a
6
2
I
0
T I M E , MINUTES
Figure 1. Rate of formation of HOD in the exchange between D 2 0 and hydrogen in glass and between HzO and deuterium in glass 0 First exchange between D20and glass
A X
*
Exchange between H 2 0 and the glass equilibrated with DzO Exchange between DzO and the rehydrogenated glass Exchange between H20and the redeuterated glass
average deviation in the concentration of HOD of around -10.5 x 1 0 - 4 ~ . A large part of the exchange occurs relatively rapidly and this is followed by a slow exchange over a long period of time. It seems to take a longer time to attain equilibrium in the DzO solution than in the HzO solution. One significant fact shown by these results is that the increase in the amount of glass surface has significantly increased the amount of deuterium exchange. When only the glass walls of the cell were available for exchange, the concentration of HOD was around 2 X 10-4M compared to 5 X 10-4M with the ground glass. This increase is well beyond the experimental error. When a second exchange of deuterium for hydrogen (or vice versa) is carried out on the same sample of glass, the amount of exchange is the same within the experimental error. It is also very evident that deuterium is adsorbed on the glass surface to a greater degree than hydrogen. The amount of HOD formed by exchange of deuterium for hydrogen on the glass is appreciably greater than that formed in the reverse process. An approximate calculation of the surface area of the glass particles was made from the size distribution given from the sieving operation. A value of 0.8 X 10l8 square Angstroms was obtained. The number of molecules of HOD put into the solution in the cell as a result of the exchange with the DzO was 4.8 X lola. On the basis of an area of one square Angstrom per water molecule, this calculation suggests that a good deal of the deuterium is exchanging with hydrogen
834
ANALYTICAL CHEMISTRY, VOL. 44, NO. 4, APRIL 1972
under the surface of the glass. The very slow exchange which occurs for the last 40-50z of the exchanged deuterium may be attributable to such penetration. More accurate determinations of the surface area are necessary to draw any firm conclusions on this question. The greater adsorption of deuterium compared to hydrogen on the glass surface suggests a possible explanation for the deuterium effect shown by glass electrodes. It has been shown (2) that when a glass electrode which has been standardized on a pH meter to read pH in a water solution is placed in a solution of a deuteroacid, the reading on the pH meter is lower than the p D of the solution by 0.40 pH unit. That is, the potential of the glass electrode indicates a greater acidity in the deuterium solution than in a water solution at an equivalent concentration of hydrogen ions. The increased adsorption of the deuterium ions on the glass surface could account for this enhanced acid ion effect. This increased adsorption of deuterium is probably related to the weaker acid character of an SiO-D bond compared to an SiO-H bond, which is the usual characteristic of deuteroacids.
RECEIVED for review July 9, 1971. Accepted October 29, 1971. This research was supported by the Petroleum Research Fund, administered by the American Chemical Society. (2) P. K. Glasoe and F. A. Long, J. Phys. Clzem., 64,188 (1960).
