The first difficulty encountered was in obtaining a boric-acid solution with the same salt effect on the indicators as the borax solution. Starting from the conclusions of Palitzsch as to the salt value of boric acid, we finally developed the solution given. We have calibrated these tubes with sea-water about 50 times and in all parts of their range. Our tubes are labeled according to column 4 of the table, and include a set of o-cresolsulfophthalein tubes from P. 7.45 to 8.30 and a set of thymolsulfophthalein tubes from P. 7.9 to 9.05. The tubes deposit borax crystals when kept long at a low temperature and these crystals should be dissolved by warming the tubes before using them.
No special precautions are necessary in making the solutions. Reagent borax and boric acid needs no recrystallization, but com mercial products must be recrystallized by dissolving in hot distilled water and filtering and cooling, then removing the mother liquor as much as possible and rinsing the crystals in ice-cold distilled water. The boric acid may be dried in a desiccator, but must not be heated above 70° to hasten the drying. Borax crystals effloresce in air on the surface, and powdery crystals must be discarded. It is usually sufficient to discard the surface layer of crystals from the bottle. The Na content of the borax may be titrated with 0.1 n HC1, using methyl red or dibrom-o-cresolsulfophthalein as indicator. It was found, how ever, that the eye was more capable of detecting efflorescence of the crystal before it was dissolved than by the use of methyl red afterwards. The other indicator is more sensitive and a flask containing the equiva lent quantity of boric acid and NaC1 may be used to get the end-point. The titration does not, however, distinguish between impurity and loss of water.
The borax solution may be prepared from boric acid and CO! free NaOH, but it takes more time to do this unless it is done at the same time that the solutions for determining the non-volatile buffer are prepared. Reagent NaOH contains more carbonate than is apparent on the surface of the sticks and must be purified. For this purpose, a saturated solution in distilled water is prepared (100 grams NaOH 100 c.c. and after thorough mixing it is sealed in a number of nonsol test-tubes by fusing the glass. These are placed upright in a thermos bottle (Dewar flask) or wrapped in cotton wool, in order to retard convection currents that stir up the precipitate. In 2 or 3 days the solution in the middle of each tube will be clear, the insoluble carbonate having settled, except for a little on the sides and the surface film. The only way to prevent a trace on the surface film is to com pletely fill the tubes, which can only be approximated. Owing to the
high viscosity of the solution it can not be measured with delivery pipettes, and it is very convenient to have pipettes whose capacity is known (which can be determined by weighing when empty and when full of water and multiplying the difference by the density of water at that temperature). A tube is opened by cutting off the sealed end, the pipette tip is plunged to the center, and the sample quickly drawn. The meniscus is adjusted, the outside is wiped off, and the NaOH washed out of the pipette into a beaker with distilled water and titrated. If the opened test-tube is to be used again it must be quickly sealed with a lump of soft paraffin and inclosed in an air-tight container or compartment, preferably in a thermos bottle, for the disturbed carbonate to settle again; but it is advisable to use a tube only once. The solution of boric acid is prepared in a volumetric flask with distilled water freshly boiled in a metal container, but enough space is left to admit the NaOH, which is run in before the flask is filled to the mark. It is usually necessary to adjust the NaOH concentration by successive titration and dilutions with freshly boiled distilled water before it can be measured with sufficient accuracy. All solutions should be sealed or provided with soda-lime tubes. The sealed tubes of saturated NaOH gradually become contaminated by the solution of glass. This process is slower in the borax solution, but it should be kept in the best resistance glass available.
Our boric acid was recrystallized several times in quartz beakers and the reagent borax was titrated, to make sure of the proper Na content. We standardized our HCl solution with reagent that had been heated to 270° to 300° (short of sintering) by Dr. F. B. Kingsbury, using methyl orange as indicator and a solution for standard color of the end-point of the indicator. The weights used were standardized and volumetric apparatus certified and used at constant temperature. All of these precautions are probably not necessary, however, to duplicate our results. Owing to the high concentration of the buffers larger absolute errors are permissible than in Sorensen's buffer mixtures. The boric acid can be titrated in borax with NaOH, using phenolphthalein as indicator, provided large quantities of neutral glycerol are added to prevent hydrolysis. 1 m NaOH titrates 1 m (Treadwell-Hall). We consider this titration superfluous, provided the crystals do not appear to have lost water, and especially if the Na content is correct. It is probably less expensive to obtain pure borax than to titrate it.