Sec. 8.5] INDIRECT METHODS FOR MEASURING DEUTERIUM 269 



in the position of the first one, and the process repeated for several redistilla- 

 tions [73]. 



It is imperative that all glass with which the water comes in contact be 

 scrupulously cleaned. The recommended procedure is to wash all the 

 articles concerned with hot solutions of trisodium phosphate, rinse with 

 distilled water, treat with boiling concentrated nitric acid, rinse about ten 

 times with freshly distilled water, and finally dry the articles in an electric 

 oven free from organic materials. 



8.4. Refractive Index. For the sodium D line, the refractive index n of a 

 mixture of deuterium oxide and water at 20°C lies between that of pure 

 deuterium oxide, 1.32828, and that of ordinary water, 1.33300, and is a 

 linear function of the mole fraction. In order to use this property in the 

 estimation of the deuterium content of a sample, one must determine its 

 value very exactly. The simpler types of refractometers are not suitable 

 for such determinations, but the interferometer can give adequate precision. 



As its name implies, the interferometer is' designed to utilize the inter- 

 ference of light waves. The difference between the refractive indices of 

 two samples is found from the shift in the position of interference bands. 

 Instruments are available commercially with which an accuracy of ± 1 X 10 -7 

 can be obtained, and with certain modifications, the accuracy can be extended 

 to + 1 X 10 -8 [46]. Because it is a differential instrument, the temperature 

 requirements are not nearly so exacting as in other instruments; no special 

 temperature control is required for an accuracy of + 1 X 10 -7 . However, 

 errors due to impurities in the sample may be great. There are a number of 

 indeterminate errors so that the measurements of An are not absolute and 

 calibration is necessary. 



The Columbia University group has used a Zeiss interferometer calibrated 

 to read directly in per cent of deuterium [13,14,45,47]. They find that the 

 deuterium content of 0.4 cc of water can easily be determined to an accuracy 

 of 0.02 atom per cent. 



The calibration of the instrument is accomplished by making determina- 

 tions on samples of known composition. This is a tedious task, but when 

 done provides a rapid and satisfactory method of analysis. 



8.5. Pycnometer. The pycnometer method usually serves as the primary 

 standard for other methods of deuterium analysis. Because it is tedious — 

 one determination may take an entire day — and because it requires large 

 samples (5 to 50 ml), it is not suited to routine determinations. 



The methods commonly used for accuracies of 1 part in 10 ,; are usually 

 modifications of the differential method described in detail by Washburn and 

 Smith [1,2]. In these methods twin silica pycnometers of the type illustrated 

 in Fig. 55 are employed, having a bulb volume of 50 cc and a capillary diam- 

 eter of 0.1 cm. The capillary stem of each has a thin reference mark A, 



