[CONTRIBUTION FROM
1559
THESECOND IONIZATION CONSTANT OF HYDROGEK SELENIDE
April 5, 1958
RADIATION LABORATORY AND DEPARTMENT OF
CHEMISTRY, UNIVERSITY OF CALIFORNIA, BERKELEY]
The Second Ionization Constant of Hydrogen Selenide BY ROBERTH. WOOD^ RECEIVED AUGUST12, 1957 The hydrolysis constant of HzSe has been calculated from the following measurements: (a) the titration curve of HpSe with potassium hydroxide, (b) the change in pH on addition of H&e to 0.2 M KOH at 0.5", and (c) the solubility of Xi32Se as a function of the concentration of OH-. The results are (a) K H > lo-*, ( b ) for 0.5" and an ionic strength of 0.2, K E ~= , ~ 1.1 2~ 0.4, and (c) for 22" and an ionic strength of 4.2, K H ' . ~= 0.25 f 0.09. From an estimate of the activity coefficients of the various ions and the temperature coefficient of KP,the second ionization constant of HzSe (for 22" and zero ionic strength Kto = 10-16.0* 0.6) has been calculated. Three methods for the analysis of alkaline solutions of HpSe nave been developed. These methods involve ( a ) weighing the selenium as the metal, ( b ) precipitating AgZSe, and (c) reduction of Ip t o I- by the Se'.
This investigation was undertaken to clear up the discrepancy in the values for the second ionization constant of hydrogen selenide. H. Hagisawa reported a value (K2= lo-") from a titration curve of H2Se with sodium hydroxide using a glass electrode.2 Lingane and Niedrach measured the polarographic waves of H,Se a t various values of pH and found for hydrogen selenide K , < 10-14.3
Experimental Preparation of H2Se.-The H2Se used in these experiments was prepared by adding water to aluminum selenide. The aluminum selenide was prepared by the method of G. R. Waitkins and R. Shutt4 using Fisher certified reagent grade selenium. The manufacturer's analysis gave the sulfur content as 0.05%. The apparatus for preparing the hydrogen selenide consisted of: (1) a reaction vessel in which oxygen-free water could be added to aluminum selenide; (2) a gas washer containing water; and (3) a reaction vessel containing either sodium or potassium hydroxide. T o free the system of oxygen the generator and washer were evacuated and filled with nitrogen and then the reaction vessel was flushed with a steady stream of nitrogen. The nitrogen was freed from all traces of oxygen by passing it over copper a t about 300". Analysis of (HSeSe-).-Three methods for the analysis of solutions of HSe- and Se- were developed. In the first method, the solution to be analyzed is allowed t o stand in air until all of the selenium has been oxidized to either selenium metal or selenite. The oxidation to selenite is completed with nitric acid and the selenite reduced t o selenium according t o the method of Treadwell and except that the selenium was converted to the grey form by heating to 50 to 60" instead of by boiling the solution. Both Se04" and SeOa- are reduced to Se by this method. Table I gives the results of several trial analyses in which pure selenium was used as the starting material. The
+
final volume is the volume of the solution from .which the selenium is precipitated. The results show that the anal>-sis is accurate to 0.2% if the final volume is kept a': about 50 ml. per gram of selenium. The second method of analysis consists of adding a measured volume of the selenide solution to a solution of Ia- (prepared from 1 0 3 - and I - ) in acetic acid and potassium iodide. The selenium is oxidized to the metal by the Is-. The excess 1 3 - is titrated with SZOS-,with starch as the indicator. The acid is added to reduce the possibility of reaction of the 1,- with alkaline Se- t o produce SeOs" instead of selenium metal. In the third method, of analysis, a known volume of solution is added to excess AgNOa in an acetic acid solution. The excess Ag+ is titrated with standard KSCN, with a silver electrode as an end-point indicator. The acetic acid is used to suppress the formation of 4gpO and its subsequent coprecipitation. Both these methods of analysis were used by McAmis and Felsing for the analysis of H2Se solutions, except that McAmis and Felsing weighed the AgpSe precipitate, instead of titrating the excess Ag+.6 Their results agreed to within 0.275 when aqueous HpSe was being analyzed. Table I1 gives the results of several experiments in which the second and third analyses were checked against each other. TABLE I1
ASALYSIS
OF
(HSe-
2 h' acetic acid, ml.
HaO, ml.
10 5 5 10 5
0 5 5 5 5
4 Se-) (XaOH 0.02035
108 -,
N
ml .
COSCN.
0.0200 N AgiiOs, ml.
15.01 15.01 35.01 15,Ol 15.01
:=
0.6 ni)
Selenium mole X 108
0.1096 ,1099
,1096 ,1093 ,1096
A pipet with a stopcock attached to two side arms was used to transfer the selenium solutions. The solution was drawn up past the stopcock into one arm of the pipet. -After TABLE I the tip of the pipet was wiped off, the solution in the lower SELENI'CY AXALYSISNo. 1 part of the pipet was drained using the other side arm to let in air. The pipet was rinsed by first drawing up and reW t . of Se taken, Final vol., g. ml . To Recovered turning some of the solution into which it had been drained, and then drawing up distilled water and adding this to the 0.8748 50 99.9 reactants. The volume of the pipet was found by calibra.SO04 50 100.05 tion with mercury to be 2.158 ml. A similar pipet with a ,9019 50 100.25 volume of 6.079 ml. was also constructed for use where the 1.1553 200 99.8 concentration of selenium was small. The Agf titrations a t first gave results that were about 0.8178 400 99.1 1% high. It was found that, after standing for 1 or 2 days, (1) University of Delaware, Newark, Delaware. Abstracted in the solution contained more Ag+. This effect undoubtedly is part from a thesis, presented by R o b e r t H. Wood, University of due t o some coprecipitation of Ag,O. KO further Ag+ was California, 1957, University of California Radiation Laboratory U n - formed after another week of standing. The final results show that the two methods agree to within 0.24,. When classified Document, UCRL-3751. the 1 2 titration was checked in 1 M KOH solution i t was (2) H. Hagisawa, Bd1. Insl. Phys. Chem. Research, Japan, 98, No. found that a sevenfold excess of IZ gave results that were 7 1034. t o lZyo high. A solution of HzSe in 3.7 M NaOH, when (3) J. J. Lingane a n d L. W. Niedrach, T H I S JOURNAL, 70, 4118 titrated with 13-, gave answers that increased severalfold (1948). as the concentration of 1 3 - was increased and concentration (4) G. R. Waitkins a n d R . Shutt, "Inorganic Syntheses," edited by W. C. Fernelius, Vol. 11, McGraw-Hill Book Co., New York, N. Y., of acetic acid was decreased. However, the Ag+ titrations gave reproducible values, which were about 20y0 lower than 1946, p. 184. the lowest value obtained with a n Ip titration. The varia( 5 ) F. P. Treadwell a n d W. T. Hall, "Analytical Chemistry," 9th tion of the 1 2 results is undoubtedly due to the oxidation of Edition, John Wiley and Sons, lnc., New York, N. Y.,1955, Vol. 2 , Se-to SeOs by the 12. p. 107.
ROBERTH. W O O D
1560
pH Measurements.-All PH measurements were made with a Beckman Model GS p H meter with a “blue” glass electrode and a calomel, saturated KCl reference electrode with an asbestos fiber junction. The electrodes were standardized in Beckman buffers. Titration Curve.-Hydrogen selenide was added to 50.0 ml. of water, and this was titrated with 0.141 N KOH. Glass Electrode Measurements.-In this experiment the change in the pH of KOH solutions after addition of H2Se was measured. The potassium hydroxide solution was standardized with hydrochloric acid that had previously been standardized against both silver nitrate and potassium hydrogen phthalate. To remove the Cot’ from the solution.some Ba(0H)Z was added. All the experiments were carried out at 0.5 i 0.5”. After the measurements of the hydrolysis constant, the glass electrode was checked by measuring the p H change on adding standard hydrochloric acid to the potassium hydroxide solution. Solubility of NazSe.-Hydrogen selenide was added to various mixtures of 3.72 m sodium hydroxide and 4.32 nz acetic acid until a precipitate of XmSe was obtained. The concentration of selenium in solution was determined by the AgZSe method. The solution was sucked up through a medium sintered-glass filter and into the pipet. I n all cases the solutions were checked at least a week after being titrated to see if any AgzO had dissolved. The silver ion formed in this period was never more than 1% of the total. When an estimate of the amount of sodium selenide precipitated was needed, the solution was stirred rapidly and an unfiltered aliquot taken out and analyzed for selenium. The difference between this analysis and the analysis of the filtered solution gives the amount of NazSe precipitated per unit volume of solution. The PH measurements were made after withdrawing a sample of the solution and placing it in a beaker which was constantly swept with nitrogen. The sodium hydroxide solution was analyzed for OH- arid Cot’ by titrating with standard hydrochloric acid to the phenolphthalein and the brom cresol green end-points. The sodium acetate was analyzed by drying aliquots of the solution to constant weight a t 140”.
Results Titration.-The titration curve is given in Fig. 1. Some selenium (formed by oxidation of H2Se
12 0.6
204
r-----T
- T
,
-
[
Fig. 1.-Titration
10-4)6J together with the titration curve is K , = 2X This value is much too large because of the formation of polyselenide. The formation of polyselenides is quite likely responsible for the value ( K , = lo-”) reported by Hagisawa from a titration curve of H2See2 Glass Electrode Measurements.-The data for the addition of H2Se to 0.2 m KOH are given in Table 111. In the calibration run the nieasuretl pH change is slightly lower than the calculated change. The pH changes measured in the H2Se experiments have been adjusted for this effect. The adjusted changes in PH are given in the column labeled “Calculated pH.” The reasonableness of this assumption is indicated by the observation that in the calibration of the glass electrode a t higher concentrations of potassium hydroxide, the measured @H change is increasingly smaller than the calculated $JH change. The concentration of selenium was measured by precipitating and weighing the selenium as the metal (see section on analysisfor details). The hydrolysis constant (KHO.~) for the selenide ion a t an ionic strength of 0.20 has been calculated using the equations Se‘
+ H 2 0 --+ HSe- + O R -
where K E is the hydrolysis constant in terms of concentration a t an ionic strength of and y is the mean activity coefficient at this concentration. The calculations were made as follows: from the PH change, the change in the concentration of potassium hydroxide was calculated on the assumption that the activity coefficients of H + and OHremain constant. This assumption is based on the fact that the ionic strength varies very little during the addition of the H2Se. From the change in concentration of the potassium hydroxide, the moles of H + neutralized can be calculated From this and the total concentration of selenium the concentrations of HSe- and Se- were calculated. When attempts were made to use higher concentrations of potassium hydroxide a white precipitate of Base was formed (KOH containing no Ba+f did not give the precipitate). Because the glass electrode does not work nearly as well in more alkaline solutions, no measurements were made a t higher concentrations of OH-. Solubility of Na2Se.-The data on the solubility of Na2Se as a function of the hydroxide ion concentration are given in Table IV. The calculation of the final concentration of OH- was made in one of two ways. I n experiments 1, 2 and 3, where the concentration o f NaOH was fairly high, the concentration of NaOH was calculated by the formula
-7
1
t
1I
Vol. 80
,/’ c/ ,-
of HtSe with 0.141 M KOH.
by traces of 02)was present in the titration mixture. This selenium originally was present as a grey metal. 22213,s. - cs.When the pH of the solution was raised above S, C O H - f i n i l = C O H - a t a r t - c ( € f S e - + Ee-)-V the selenium began to dissolve to form polyselenides where COS-indicates the concentration of the hyaccording to the reaction droxide ion, etc. The number of moles of sodium nSe -I- HSe- ---+H + Se-c, + 1) selenide precipitated is given by M N ~ and ~ s the ~ This reaction is undoubtedly the cause of the dis- volume of solution is given by L‘. The Cse“ has continuity in the titration curve a t pH 8 (see Fig. 1). (6) A. J \IcAmis and W. A. Felsing, THISJ O U R N ~ L , 47, 2636 The second ionization constant of H,Se calcu- (IW5). (7) hl d’I-Ilasko. J . Chem Phyr.. 20, 107 (1923). lated from the first ionization constant (& = 2.3 X
+
THESECOND IONIZATION CONSTANT OF HYDROGEN SELENIDE
April 5 , 1955
1561
TABLE I11 GLASSELECTRODE EXPERIMENTS Ionic strength = 0.21 * 0.01 Expt. no.
1 2
Calibration
KOH at start A pH (mole X 108) measured
4.26 6.40 6.52
0.416 ,314 .296
A pH calcd.
KOH final (mole X 108)
0.439 =k 0.01 .331 f .1 .313
1.55 f 0.05 3.03 3.46
HSe added (mole X 10')
HSe' calcd. (mole X 10')
Se- calcd. (mole X 109
Val. ml.
2.52 3.22
2.33 f 0.05 2.97
0.19 f 0.05 .25
20.4 32.4
..
.........
.........
KHO.21
1.0 f 0.3 1.3f0.5
..
.......
Weighted av.
1.1 f 0.4
TABLE IV
SOLUBILITY OF Na2Sevs. COHExpt. no.
1 2 3 4 5
NaOH concn. at start,
Na+ concn. at start, m
NarSe concn., m
3.72 1.40 0.370 .I30 .095
3.72 4.13 4.31 4.31 4.31
0.02 f 0.01 .01 f ,005 .004 .05 f .01 .02 f .01
m
9H
.......... .......... .......... 12.46 zt 0.10 11.47 f .05
been calculated by first assuming that i t is zero and then calculating the solubility product of NazSe. From the solubility product, the concentration of Se- is calculated and the correction is applied iteratively. I n experiments 4 and 5 , where the concentration of OH- is quite low, glasselectrode measurements were used for the calculation of the concentration of OH-. The correction for the effect of N a f on the electrodes (0.07 to 0.10 pH) was taken from a graph published by Beckman Instruments, I n c 8 The activity coefficient of sodium hydroxide in this solution was assumed to be the same as that of sodium hydroxide a t the same ionic strength as given by Robinson and stoke^.^ The activity of the water in this solution was assumed to be the same as in 4.2 m NaOH; ;.e., 0.85. The results of the calculations are given in Table IV. I n experiment no. 1 (in 3.72 rn NaOH) the analysis was about 1.4 times as high as the value expected on the basis of the other data. This is probably due either to some coprecipitation of AgzO with the Ag8e or to the formation of Sez-. This latter explanation is made more probable by the fact that the solution was noticeably yellow, whereas in the other experiments the yellow color, if present, was much lower in intensity. This experiment was discarded in calculating K,, and KH. The dependence of the solubility is given by the equations Kap = ( C R +)'Cge~
Na+ concn. final, m
(HSe-+ se-) concn.,
(Na+)l (HSeSe-) concn., m
3.68 4.11 4.30 4.21 4.27
0.00524 ,00283 ,00396 .01538 ,0590
0.0709 .0473 .0732 ,272 1.07
m
+
.O22fOo.0O6 .0os f .002
w. 11(CxB+)2(Cse-
CHSe-) ( COH-)
The intercepts of a straight line drawn through ) (l/Ksp), these points should give ( l / K s p K ~ and respectively. The two straight lines in Fig 2 are my estimates of the limits of accuracy with which
r-
-..-.~7
I
__
'I
, "s
2
Fig. 2.-Plot
(8) Beckman Bulletln 100-D. Beckman Instruments. Inc., La Habre, California. (9) R. A. Robinson and R. H. Stokes, "Electrolyte Solutions," Academic Press, New York. PIT. Y.,1965.
3.67 1.87 0.369
+
I/(CN~+)*(CHS.-
(
In these equations we assume that if the ionic
OH- concn. final, m
strength is constant the activity coefficients are also constant. Figure 2 is a plot of
or
or
+
c tc
e-+
se=cod-.l
of data om Na&e solubility as a function of hydroxide concentration.
the points determine a straight line. These lines yield Ksp4-'=0.040 f 0.002 and = 0.85 f 0.09. Theioniestrengthof thesolutionsis4.2 f 0.1. Interpretation and Discussion From the two experimental values of the hydrolysis constant (KH'.~= 1.1 f 0.4 a t 0.5' and
1562
C. D. MILLER,R . C. MILLERAXD IV. ROGERS, JR.
Vol. '30
KH"' = 0.25 =t 0.09 at 22") two valuesof K ~ ~ a tTable V gives the PKs (where $K = -log K ) of 22" can be calculated from the usual thermody- several of the acids analogous to H,Se. namic equations and the assumptions that (a) the Assuming a simple electrostatic model, one obactivity coefficients of NaHSe and KazSe in these tains the equations solutions are the same as those of NaOH and NanpK, 3 3 = SO4, respectively, at the ionic strength of the solu7 r tions, and (b) Latimer's estimates of the entropies of HSe- and Se=, namely, 22 and 0 e.u., respecPA', = -c tively, are correct. These calculations give, from the measurements at 0.5", K? = 10-14.5*n.4 where C is a constant, z1 and z2 are the charges on a t 22", where the limits of error come from the h i - the ions. and r is the bond length of the atoms. On its of error of the original measurement and my es- the basis of this model one would expect that pK1, timate of the reliability of the above assumptions. PK, and PK, - ~ Kwould I increase from the heavier The measurement a t 22' gives Kn0= 10-15.9* O.'. members of the series to the lighter members. If both measurements are taken into account with This does not occur. The discrepancy is shown their estimated limits of error, the best value is most distinctly by pK, - pK1, where the value is probably K = 10-1j." 0 . 6 , lower for H,S than for either H2Teor H2Se. Therefore, on the basis of this model, one would conclude TABLE Tthat the value of K 1for H,S is anomalous. The asXCID COSSTAXTS OF H2.1 signment of partially covalent character to the Compound ~ K I @Kz ZIG-~KI Ref. bonds does nothing to help resolve this anomaly. HzO 15.7 Acknowledgment.-I wish to express my ap5.9 11,12 7.0 12.9 HzS preciation to Professor Richard E. Pornell and 3.6 15 HzSe 11.4 ti, 7 to Professor Robert E. Connick for their many help2.6 11 H2Te 8 4 7,3 ___-ful suggestions in connection with this work, which (10) \V, 52. Latimer, "Oxidation Potentials," Pret:ticv-€€dl, S r i v was performed under the auspices of the U. S. York, S.Y . , 1953. Xtoniic Energy Commission. (11) H. Kubli, Helu. Chzm. A c t a , 29, 1962 (194G).
c
BERKELEY, CALIFORNIA
(12) N. Konopick and 0 . Leberl, .\fa?zatsh. Cheiii., 80, 781 (1949).
[COSTRIHUTIOS FROM THE I)F,PARTLIENT O F CHEMISTRY,
Phosphine Oxides. BY C A R O L I N E
D.
V.
TEMPLE USIVERSITT]
Intra- and Intermolecular Association1
h I I L L E R I E a 3ROBEKT C L A Y &!ILLER3 AND W I L L I A M
ROGERS, JR.
RECEIVED SEPTEMRER 27, 1957 >P---c< -111
iiitrarnolecular hydrogen bond of the type
4 9 '. i
0
has beeti found in certain a-hydroxy phosphoryl compounds.
1% The strength of this bond has been related to the negativity of the phosphoryl oxygen. n'hen the phosphoryl is extremely polarized, as in the phosphine oxides, an interinolcculnr association is also observed. This is probably a dipole-dipole interaction of these phosphoryl groups.
The spectra of a number of phosphoryl compounds have appeared recently in the literature, and various assignments have been made, both empirically and theoretically. The effect of electronegativity of the substituents on the phosphoryl stretching frequency has been shown,.1,5and the existence of hydrogen bonding phenomena involving the phosphoryl group has been d e m ~ n s t r a t e d . ~Anal,~ ( 1 ) Presented a t t h e Meeting of t h e American Chemical Society For t h e fourth paper in this series, aee R. C Miller, C. D . Miller. W. RoRera and L. A. Hamilton, TIIISJ O U R N A L , 79, 424 11967). ( 2 ) Abstracted in p a r t from t h e dissertation submitted by C . D . M. to rhe Temple University Graduate Council in partial fulfillment of the requirements for t h e degree of Doctor of Philosophy-. (3) Experimental Station, E . I. d u P o n t d e N e m o u r s Co., R'ilrnington, Delaaare. (4) L . W.Daasch and D . C. Smith, A n a l . C h e w . , 23, 853 (1951). ( 5 ) J . V . Bell. J . Heisler, H. Tannenbaum and J. Goldenson, THIS TOVRNAI, 76, 5183 (1954). (13)G. h1. Kosolapoff and J. F. McCullough, i b i d . , 73, 5392 (1951). (7) E. Halpern, J . B o ~ i c k ,H. Finegold and J . Goldenson, ibid.,77, 4472 (1955) at S f w York, N. Y., September lY57
ogous relationships involving the carbonyl group have received somewhat more attention * and, with the exception of the steric differences of the tetrahedral phosphorus atom of the phosphoryl from the planar carbon atom of the carbonyl, there exists a iiotable similarity. Recent work a t this Laboratory has resulted in the synthesis of several new classes of organophosphorus compounds, whose spectra have also been recorded, from which certain information concerning their structure and properties may be deduced. Chronologically the first, and structurally the simplest, of these are the disubstituted phosphine oxides, KJ'(0)H. Prior t o their general synthesis,8 such compounds were tentatively classed as trivalent phosphinous acids, R*P-OHg. However, the presence of strong P-H absorption a t 2335 cm.-' ( 8 ) R. H . Williams and L. A . Hamilton, ibid., 74, 5418 (1952). (9) G. >,I. Kosolapoff, "Organophosphorus Compounds," John W l e y and Sons, Inc., X e w York, N. Y., 1950.
