Colorimetric Determination of Hydrogen Dissolved in Water SIR: Stahlberg and Isoniemi (8) reported that hydrogen concentrations in water can be measured by adding an excess of alkaline permanganate to the water solution and allowing the reduction of permanganate by molecular hydrogen to proceed a t ambient temperatures. The absorbances of the reduction product were measured in a spectrophotometer a t a wave length of 410 to 420 mp a t 2 hours and again a t 6 hours. The absorbance of the reduced permanganate was proportional to hydrogen concentration. An investigation had been made in this laboratory to apply the Stahlberg Isoniemi method to the determination of hydrogen dissolved in waters used to test nuclear reactor components (1). The main objectives were to shorten the time required and to extend the range of the method to concentrations of hydrogen above those normally encountered a t ambient temperature and pressure. Preliminary qualitative investigation revealed that alkali appeared to have no significant function and that heat speeded the hydrogen-permanganate reaction. A buffer of monosodium hydrogen phosphate was added to stabilize p H values of specimen waters.
Figure 1 . Absorbance of potassium permanganate 40 rnl. of saturated hydrogen water
_ _ _ _ _ 20 ml. of saturated hydrogen water
.........N o hydrogen
the specimen waters on the standard curve, multiplying the concentration found by the dilution factor made in the injection syringe. EXPERIMENTAL RESULTS
50
lP
0
180
190
IO0
Pull 10 ml. of saturated hydrogen water into a 50-ml. syringe, followed by enough degassed water to fill the syringe METHOD to the 50-ml. mark. Inject the diluted Reagents. POTASSIUM PERMANGA-hydrogen solution into the first bottle with the needle tip immersed and fill NATE SOLUTION. Dissolve 4.80 grams the bottle to the lip with degassed of potassium permanganate and 55.2 water. Restopper the bottle without grams of monosodium hydrogen phosincluding air bubbles. Repeat this prophate (0.1 HzO) in 4 liters of deionized cedure with 20-, 30-, 40-, and 50-ml. water. Allow to stand overnight. aliquots of hydrogen water. Add a last Filter through a medium-porosity bottle to the series containing the fritted-glass Buchner funnel into a glasspermanganate reagent diluted to volume stoppered 4liter bottle. Discard the with degassed water but with no hydrofilter. gen water added. SATURATED HYDROGENWSTER. Add bottles to the standard series Bubble electrolytic grade cylinder hycontaining aliquots of the specimen drogen through deionized water for a t waters diluted in the injecting syringe least 1 hour. Determine the ambient with degassed water to the concentratemperature and pressure. Read the tion range 0 to 17 cc. of hydrogen per concentration of dissolved hydrogen kilogram of water. from tables of published data ( 2 ) . Heat all bottles for 25 minutes in a Continue bubbling during performance constant temperature bath held a t of the procedure. 75' C. and cool rapidly. Determine the DEGASSED WATER. Maintain several absorbances of the contents of each liters of deionized water a t the boiling bottle a t 350 mp in a spectrophotometer point. Withdraw degassed water from that has been adjusted to zero absorbthe bottom of the container through a ance on the contents of the bottle that cooling coil. contained no hydrogen. Procedure. Place about 60 ml. of degassed water in each of a series of A plot of the absorbances us. the 125-ml. reagent bottles. Add exconcentrations of the standards is a actly 10.0 ml. of potassium permanstraight line. The unknown concentraganate solution t o each bottle and tions are read from the absorbances of restopper the bottle. 1598 *
ANALYTICAL CHEMISTRY
Two known concentrations of dissolved hydrogen and a blank were treated as described. Their absorbances were determined a t 5-mp intervals between 320 and 700 mp on a spectrophotometer successively zeroed on deionized water. Figure l indicates the choice of 350 mp for sensitive and linear measurement of the hydrogenpermanganate reaction product. Groups of three determinations of identical known hydrogen concentrations were made as described but heated a t 75" C. for measured periods of time between 15 minutes and 1 hour. Examination of the resulting average absorbances revealed no increase after 20 minutes of heat treatment. A 25-minute heating period was therefore adopted to ensure complete reaction. Groups of 10 samples containing identical concentrations of dissolved hydrogen were treated as described, but each group was heated for 45 minutes a t selected temperatures between 50" and 95' C. Figure 2 indicates that the smallest per cent error resulted from a heat treatment a t 70" C. but consideration of the desirable 25minute characteristic led to a decision to use a 75" C. temperature. During this investigation an accidental breakdown in the carbon filter used to ensure absence of organic material in the supply of deionized water resulted in immediate indication of organic impurity, with a large
increase in absorbance in all blank determinations. The method described was used for determinations of hydrogen content in pure water systems adjusted to contain between 5 and 50 cc. of hydrogen per kilogram of water on a daily routine basis. Standard deviation from the mean of multide determinations on the same system was *50j,. No comparisons were made to other methods for hydrogen dissolved in water.
ACKNOWLEDGMENT
The authors are indebted to F. D. Bell of this laboratory for analytical assistance and to the U. S. Atomic Energy Commission and the Westinghouse Electric Corp. for permission to publish this work. LITERATURE CITED
(1) Easor, J. E., Jr.,
Harves, T. O., Conkhn, D. B., Office of Technical
Services, Dept. of Commerce, Bettis Tech. Rev. WAPD-BT-7,146 (1958). (2) “Handbook of Chemistry and Physics,” 39th ed., p. 1606, Chemical Rubber Publishing Co., Cleveland,Ohio, 1957-58. (3) StaNberg, K., Isoniemi, M., Suomen Xemistilehti 27B,86 (1954). RICHARD H. ROBINSON B. CONKLIN DWIGHT Atomic Power Division Westinghouse Electric Corp. Pittsburgh 30, Pa.
183.
Growth and Crystal Structure of Single Crysta Is of Pr(N03)3. 6H,O J. W. RICHARDSON, Q. W. CHOI,l F. VRLTN?, and J. M. HONIG2 Department of Chemistry, Purdue University, Lafayette, Ind.
I
of studies on the oxidation of nonstoichiometric praseodymia, it was necessary to obtain information on the properties of praseodymium nitrate. Because of lack of information on the growth and properties of this compound, a brief study was undertaken. N THE COURSE
Single crystals of the material were grown by dissolving high-purity Pr6011 in hydrochloric acid solution and evaporating the solution almost to dryness; the resulting chloride was treated with nitric acid and evaporated almost to dryness. To ensure complete elimination of chloride ion, this treatment was repeated three times on a steam bath. The residue was placed in a desiccator together with a 15% sodium hydroxide aqueous solution in another beaker. Because of the extreme hygroscopicity of praseodymium nitrate, this compound slowly dissolved in water withdrawn from the sodium hydroxide solution: the process was terminated as soon as complete solution of the nitrate .was achieved. A 40 to 45% aqueous sodium hydroxide solution was now substituted. As its vapor pressure was slightly less than that of the praseodymium nitrate solution, a slow transfer of water to the sodium hydroxide reservoir occurred. This resulted in a slow but steady growth of a single crystal of Pr(N03)s.6Hz0,with dimensions 2 x 1.5 X 0.5 cm. The crystal exhibited a marked spiral dislocation pattern on the top surface. This technique can be used whenever Present address, Department of Chemistry, Cornel1 University, Ithaca, N. Y. * Present address, M.I.T. Lincoln Laboratory, Lexington 73, Mass.
the substance to be crystallized is soluble in moderately volatile solvent and the solvents can be distilled in a similar manner, It has the following advantages: No special equipment is required; moderate temperature fluctuations can be tolerated because the aqueous vapor pressures of the praseodymium nitrate and sodium hydroxide solutions change in the same direction with temperature changes; and the vapor pressure of the desiccating solution increases as evaporation from the crystallizing system takes place, so that the rate of evaporation of water vapor is automatically reduced as the solution approaches the stage of nucleation. This makes it possible to grow a single crystal at a time. The latter effect can be enhanced by utilizing a small amount of desiccant solution relative to crystallizing solution. The rate of distillation can be adjusted easily by varying the surface area of the vessel containing the desiccant. Another set of single crystals was accidentally grown during attempts to produce the anhydrous material in large quantities, in the course of which the vacuum line was destroyed. Hydration then took place and the single crystals thus obtained measured 1 x 0.5 X 0.1 cm. One of these was cut up into small sections and a fragment was sealed in a thin-walled borosilicate glass capillary for x-ray diffraction studies. The crystal was mounted on a Weissenberg camera; x-ray scattering from nickel-filtered CuKa radiation was recorded photographically. Because of the reported thermal instability, air
was blown over the capillary during measurements. A serious difficulty arose in the process of aligning the crystal axis with the rotation axis. During the preliminary x-ray measurements, the crystal apparently changed from an initial to some other structure. I n the course of further alignment, the crystal returned to its initial state, although to a somewhat different orientation within the capillary. Considerable difficulty was encountered in preserving the specimen through the alignment stage. However, after the process was completed, 0-level Reissenberg (hkO) data were recorded. Following this, the crystal became disoriented, and another alignment process had to be attempted, with the same attendant difficulties as before. After a second alignment, a complete rotation photograph and a second set of Weissenberg 0-level photographs were again obtained. Thereupon , the crystal totally disintegrated. An analysis of the rotation photograph and the first set of 0-level Weissenberg data showed that the lattice is monoclinic and assumed primitive; the cell constants are:
*
QQ = 8.64 0.05 A. bo = 11.75 zt 0.05 A. co = 6.78 -f 0.05 A. 7 =
112 zt
3O.
Calculated density for two formula units per unit cell, 2.42 grams per cc. Observed density, 2.48 grams per cc. by displacement of ethyl acetate. The indicated uncertainties are rough estimates, Indexing of the Weissenberg data revealed that all hkO reflections are in VOL. 31, NO. 9, SEPTEMBER I959
1599