Hydrolytic Behavior of Metal Ions. VI. Ultracentrifugation of Zirconium

VI. Ultracentrifugation of Zirconium(IV) and. Hafnium (IV); Effect of Acidity on the Degree of Polymerization12. By James S. Johnson and. Kurt A.Kraus...
0 downloads 0 Views 863KB Size
Aug. 20, 1956

ULTRACENTRIFUGATION OF Zn(1V) [CONTRIBUTION FROM

THE

AND

Hf (IV)

3937

OAK RIDGENATIONAL LABORATORY, CHEMISTRY DIVISION]

Hydrolytic Behavior of Metal Ions. VI, Ultracentrifugation of Zirconium(1V) and Hafnium(IV); Effect of Acidity on the Degree of Polymerization1b2 BY JAMES S. JOHNSON

AND

KURTA.

KRAuS

RECEIVED JANUARY 12, 1956 Polymerization of zirconium and hafnium (ca. 0.05 M ) has been investigated by equilibrium ultracentrifugation. Hf (IV) a t this concentration is probably polymerized in 5 M HC1. There is definite polymerization in 3 M HC1. In the range 0 j-2 M HCl, Hf(1V) is probably either a trimer or tetramer and the degree of polymerization is relatively independent of Hf(IV) and HC1 concentrations. At acidities 0.2 M and less, Hf(1V) is more highly polymerized and polydisperse. Polymerization of Zr(1V) is similar. The principal difference is that Zr(1L’) is more highly aggregated than Hf(I\.’) at the lower acidities. Centrifugation of solutions containing varying concentrations of Zr( 11’) and of supporting electrolyte in 1 Jf HC1 indicate that the polymers, like those of hafnium, are trimeric or tetramericat this acidity and carry charges of G Q . plus one per monomer unit. The same degree of polymerization was found for Zr(1Y) in 1 -11HClOd Decrease of temperature from ca. 25’ to ca. 5” markedly lowers the degree of polymerization in 0.08 M HC1, in contrast to the small effect observed in 1 M HCl. Aging of ZrOCll solutions, or heating, increased the degree of polymerization signihcantly but did not produce very large polymers.

The hypothesis that Zr(1V) polymerizes on hydrolysis is of long standingj3but descriptions of the process inferred from different methods are somewhat contradictory. Workers in the field seem to have abandoned the idea4 that a monomeric ZrO++ ion predominates in aqueous solutions over a wide range of acidity. Polymerization of Zr(1V) is suggested by many types of results, including slow conductivity changes5e6; freezing point depressions and potentiometric studies7; and titrations of ZrCl4 with base.8 From diffusion measurements, Jander and Jahrg concluded that Zr(1V) in perchlorate solution forms low molecular weight polymers and that the degree of polymerization is unchanged over a wide range of acidity. Small polymers were also postulated from measurements of hydrolysis of ZrC1410in acidic solutions and from the rate of Zr(IV) uptake by ion-exchange resin from nitric acid solutions.l1 On the other hand, Connick and Reas12measured the extraction of Zr(1V) from one to two molar HC104 into benzene-thenoyltrifluoroacetone (TTA) solutions, and analyzed their results by the continuous polymerization hypothesis of Sill&. lS Their (1) This document is based on work performed for the U.S. Atomic Energy Commission at the Oak Ridge National Laboratory. Previous paper: J. S. Johnson, K. A. Kraus and R. W. Holmberg, THISJOURNAL. 7 8 , 26 (1956). (2) Presented in part, 126th ACS Meeting, New York, September, 1954, Abstracts, p. 60-R, preliminary report, K. A. Kraus and J. S. Johnson, ibid.,76. 5769 (1953). (3) W. B. Blumenthal, Ind. Eng. Chcm., 46,528 (1954). This paper reviews zirconium literature and has a more complete set of references than presented here. Earlier literature is reviewed by F. P. Venable, “Zirconium and Its Compounds,” ACS Monograph Series, Chemical Catalog Co.. Inc., New York. 1922. (4) See e.&, (a) E. Chauvenet, A n n . chin., 18, 59 (1920); (b) W. M. Latimer, “The Oxidation States of the Elements and Their Potentials in Aqueous Solutions,” Second Edition, Prentice-Hall, Inc., New York, N . Y . , 1952, p. 271. (5) R . Ruer, 2. anorg. Chem., 48, 282 (1905). (6) F. P. Venable and D. H. Jackson, THIS JOURNAL,42, 2531 (1920). (7) M. Adolf and W. Pauli, Kolloid-Z., 29, 173 (1921). (8) H. T. S. Britton, J . Chcm. Soc., 127, 2120 (1926). (9) G. Jander and K. F. Jahr, Kolloid-Bcih., 48, 295 (1936). (10) K. A. Kraus and S. Y.Tyree, Chem. Div. Quarterly Report ORNL-499, September 30, 1949, p. 26. (11) B. A. J. Lister and L. A. McDonald, J . Chcm.’Soc., 4315 (1952). (12) R . E. Connick and W. H. Reas, THISJOURNAL,78, 1171 (1951). (13) (a) F. Grandr and L. G. Silldn. Acta Chcm. Scond., 1, 631 (1947); (b) L. G. Sillbn, ibid., 8, 299 (1954); (e) 8, 318 (1954).

interpretation indicated a mixture of species, considerably more highly aggregated and more polydisperse than postulated by the earlier workers. Later, however, spectrophotometric studies of Zr(1V)-TTA complexes, carried out in the same laboratory, were interpreted on the basis of zirconium trimers and tetramers under these conditions.l 4 Larsen and Wang16 recently reported results of cation exchange of mixed Zr(1V)-Hf (IV) solutions M ) in (total metal concentrations to 0.5-2.0 M HClOI. They concluded that in the aqueous phase Zr(1V) and Hf(1V) are present as Z r f 4 and Hf +4 ions a t their lowest metal concentration and that the species in the resin are hydrolyzed. I n chloride solutions, ultracentrifugations have indicated that in solutions 0.0075-0.1 M in Hf(1V) and 1 M in HC1, most of the Hf (IV) is found in one or a few species having a weight average degree of polymerization, N,, of 3 or 4.’ Other studies of the solution chemistry of hafnium are rare. It is generally believed that i t is similar to that of Zr(1V). The solubilities of Zr0 C l ~and HfOCl2 as a function of HCl concentration are indeed much alike.16 Larsen and GammillI7 found evidence in a study of precipitation of the hydroxides that hafnium is somewhat less hydrolyzed than zirconium. I n this paper a study of the effect of acidity on the polymerization of Zr(1V) and Hf(1V) in aqueous solutions, principally by equilibrium ultracentrifugation, is presented.

Experimental Details of centrifugation procedure have been described earlier.lI1* Briefly, solutions were centrifuged in a Specialized Instrument Corporation Model E ultracentrifuge. Observation of sedimentation is based on refractive index gradients with radius resulting from concentration gradients induced by the centrifugal field. The quantity used for interpretation is Z*,a number proportional to the refractive index gradient attributable to the presence of the polymeric solute. The solutions were contained in cells of approximately 12 mm. thickness in the direction of the optical path. The (14) A. J. Zielen. U.S. A.E.C., UCRL-2268, July, 1963. (15) E. M. Lacsen and P. Wang. THISJOURNAL, 76, 6223 (1954). (16) 0. v. Revesy, Kgl. Danskc Vidcnskab. Selskab, Mat.-jys. Mcdd., 6, No. 7 , I (1925). (17) E. M. Larsen and A. M. Gammill. THIS JOURNAL,72, 3615 (1950). (18) J. S. Johnson, K. A. Kraus and T . F. Young, i b i d . , 76, 1436 (1954).

n938

Vol. 7 s TABLE I L-LTRACENTRIFUCATION OF ZIRCONIUM(IV) A N D HAFSICM(II.) IN CHLORIDE SOLUTION

Expt. no. and conditions"

Medium

Hf(1V)

Initial molarity Zr(IV) HCI

-

>IC1

--,:e z = 6.3

7

x = 6.9

Temp., OC.

1 2

B-5 33.2 0 ,811 1.18 HCI 0 . 049 0 0 B-4 20.1 I1 1.28 1,53 HC1 ,049 3 *3* I 1,043 (I 29.0 3 0.88 1.14 HCl B-5 4 I1 .05Il 2 1.82 HCl B-4 26.9 32.2 56 ,048 1 1 HC1-Sac1 C-5 1.88 33.1 B-5 fib 1 1.25 ,047 1 H C1-SaC I 34.0 7 ,050 1 HC1-CSCI 1 0.89 B-5 8 0.52 :j:3 . 8 1 1 .88 HCI-CSCI B-4 9 :3:3 , I1 1 1.29 049 1 HCI-LiC1 B-,5 10 :33 . :3 1 1 35 ,023 1 HCI-LiCl B-3 1) 34,; 11 ,ll31) 1.27 B-5 1 HC1 26.7 12 0.50 0 5 1. 5 1 .9B H C1-LiCl B-4 3 0 , (1 13" 1 . $1 1.88 2.09 (148 .2 B -3 HC1-Sac1 1Ab 29.7 1,77 2.32 B-6 1 .9 H C1- SacI ,043 .2 27.2 3.71 4.59 15 1.9 H C1-LiC1 050 .08 .I-2 8.5 ,050 1.9 2.75 3 .39 HC1-LiC1 16 ,118 A-2 17 26.2 2,90 4 .85 H C1-LiC 1 1.9 C-4 ,051 . 08 0.9 1.9 2.20 2 06 18 H CI-LiC 1 ,051 . (18 B-4 (I 10.(1 12.2 28.1 19" 2 LiCl ,05 B-1 ) 2fY 20 3 1) 13 2 18 ,'3 LiCl .05 B-1 Bar angle: -4,35"; B, 45'; C, 55'. =\pproximate speed of rotation (r,p,m.): 1, 9253; 2, 17,980; 3, 20,410; 4, 23,15(!; *5, 27,690. Zr(1V) or Hf(1V) introduced as ZrClc or HfCla. All other experiments, introduced as oxychloride. e ExperiValues of s,' ment 19 on unheated ZrOClz solution; Experiment 20 on ZrOClz solution heated for one hour at 100". used to compute Fig. 1. e a: = 108 S: T / d . Singl-. values of CY: auply to average radius of sdutions ( x = 6.6 cm.) and indicate that SL was either constant or decreased slightly with 5 . J Equilibrium not attained. Y

cells were located a t a radius of about 65 nun. M o s t solutions attained equilibrium distribution in the field after about a week of centrifugation. I n n few cases, a t lower acidities, slow changes of Z* with time were observed after considerably longer periods. Apparently polymerization was still occurring in these solutions. In these cases, the results reported were computed from the photographs made at the termination of the experiments as though equilibrium had been attained. The degrees of polymerization obtained in this way probably represent fairly well the aggregation of the solute a t the end of the centrifugation. Zirconium( IY) and hafnium( IV) were usually introduced as the oxychlorides, though the tetrachlorides were used in a few cases. The hafnium compounds were essentially zirconium free and, along with the other chemicals used, have been discussed earlier Commercial ZrOCJz was used after recrystallization from 9 Af HCl. Zirconium tetrachloride was supplied to us by Dr. s. Y. Tyree of the Univeristy of S o r t h C a r ~ l i n a . ' ~Spectroscopic analysis of both compounds showed negligible impurities except for 1.4 to 2.5 weight per cent. hafnium.10 Metal-chloride mole ratios determined by gravimetric analysis of the oxychlorides were equal to the theoretical values within accuracy of the analyses. Densities of the solutions were determined with a ppcnometer and with gradient tubes.2l

Results and Discussion Solutions of Hf (IV) at stoichiometric acidities 0.08-5 M HC1 in excess of the oxychloride were studied by equilibrium ultracentrifugation. The (acid) range for zirconium was 0-3 M excess HC1. The initial concentration of Hf(1V) or Zr(1V) was about 0.05 M except for a series in 1 M excess HCl, for which the effect of concentration was studied. The effect of temperature on polymerization in the range 0-30' was investigated for solutions 0.08 M in (19) W. S. Hummers, S. Y. Tyree and Seymour Yolles, in " I n organic Syntheses," Vol. 4, McGraw-Hill Book Co., New York, N. Y., 1863, p. 121. (20) We are indebted t o C. Feldman and M. Murray of t h e O R N L Spectroscopic Laboratory for t h e analyses. (21) K. Linderstr$m-Lang and H , [.an%, C o n ~ b l ,r r n d . f r n v l o b . Cc,risb.r.q, 2 1 , 315 (1038).

HC1. .A solution of ZrOClz with no excess acid, which had been heated a t 100' for an hour, was also centrifuged in order to correlate the changes in chemistry observed by Ruer with possible differences in degree of polymerization. The conditions for the equilibrium centrifugations are summarized in Table I. The experiments with Hf(1V) in 0.5-2 M HC1 have been reported earlier,' but are included for comparison. Two centrifugations in 1 -1f HC104will be discussed (section 7) but are not included in Table I. -1few velocity centrifugations were also carried o u t to obtain information on polymerization of freshly prepared solutions. The new equilibrium centrifugation results for chloride solutions are summarized in Fig. 1, which is a plot of the difference (log ( Z * ' x ) - S:.x2) z's. x'. This difference should be independent of .x2 for an ideal monodisperse solute. The values of Si = d log (Z*/x),/d(x2) = Se/2.303 were selected to give the minimum range of ordinate in Fig. 1 and are listed in Table I. Only those centrifugations for which equilibrium was attained are included in the figure. The symbols and equations used in this paper have been discussed more fully earlier. l 1. Method of Interpretation. (a) Monodisperse Systems.-The polymers studied are probably charged, and the magnitude of the charge is unknown. Since equilibrium distribution is affected by charge as well as by molecular weight,?? even though centrifugations are carried out a t a high concentration of light supporting electrolyte, estimation of the degree of polymerization in principle requires knowledge of charge. For systems having an essentially constant weight average degree of polymerization, AVw,( i . e , , d In LVwd ( 9 ) = 01, and ( 2 2 ) 0. I.nrnm, A v k i o K i w i , . 1 4 i u ~ v < z l C a d

,

1 7 A , paper 2 5 ( l W $ >

ULTRACENTRIFUGATION OF 2 n(1V)

Aug. 20, 1936

obeying certain other restrictions, simultaneous evaluation of N , and of z', the charge per monomer unit, is possible by carrying out centrifugations under a variety of conditions.'tZ3 The equations for computing N, and z' from experimental results were derived23on the assumption that charge z', activity coefficient (y+i), partial specific volume ( a i ) , and specific refractive index increment (bnldwi) are constant for each species, and that solution density ( p ) is also constant. The weight average degree of polymerization is given by the equation Nw =

S I Ai

1

21'cf

- -a_

(3 -

-7

AND

3939

Hf (IV)

zc

i

(1)

2~3Al

In this equation S = d In ci/d(x2); c = d In c3/d(xz); Ab = Ml(1 - flzp)w2/2RT; w is the angular velocity; R the gas constant; T the absolute temperature; c: the concentration of the polymer component (component 2 ) expressed as monomer; and cg the concentration of the supporting electrolyte (MX, component 3). The polymer component is defined as (PX, - ( z / 2 ) MX) where P+. is the polymeric cation and X- the anion. The polymer component has the molecular weight = NAJJ, the charge z = Nz', and the partial molal volume Vi appropriate to its definition. Equation 1 is in terms of S = d In c;/d(x2), rather than the experimentally obtained S, = d In (Z*/x)/d(x2). The relation between these quantities is given by the equation' 35

45

40

50

xz

Fig. 1.-Test

and the assumption that S, is equal to d In (dcil xdx)/d(x2) = S,. By equations 1 and 2 values of NL may be computed and plotted for the plausible range of z'. A series of centrifugations are carried out a t different The intersection of these curves or ratios of c;,'c3. the point where they approach closest to each other then gives N , (or N ) and z'. Equations 1 and 2 are for systems of the three components, water, supporting electrolyte and polymeric solute. If more than one supporting electrolyte is present and if the supporting electrolytes differ greatly in M (1 - D p ) , so that an average value is not adequate, more elaborate though similar equations must be used.' (b) Polydisperse Systems.-For those polydisperse systems where N , does not change greatly in any single experiment, equations 1 and 2 may be used directly without introducing significant errors. These systems are characterized by essential invariance of the difference log Z*/x - Sh2 as a function of x or by a downward curvature of this function (charge effect). However, some of the systems studied here showed upward curvature of this function, which implies considerably greater polydispersity. For these, d In Nw/d(x2) is not constant and equations 1 and 2 are not directly applicable. However, if the analog of equation 2 is (23) J S Johnson, K.A Krauo and G. Scatchard, J. Phrs. Chcm., 6P, 1034 (1954)

of constancy of Si

derived without neglect of d In -J7,Jd(x2),the equation

is obtained. A rather cumbersome iteration is required to obtain c: as a function of x and such effort does not appear justified in the present work, in the light of the limited experimental accuracy and the uncertainties in many of the assumptions. Thus, i t is unlikely that the individual polymeric species are hydrolyzed to the same degree, and the assumption that they all have constant properties, e.g., constant specific refractive index increments, is probably not as good for highly polydisperse solutes as for solutes distributed in a narrow range of molecular weights. Treatment of results for obviously polydisperse systems, therefore, will be by the less ambitious procedure of setting z' = 0. Lansing and KraemerZ4have shown that for uncharged polymer systems, for which species activity coefficients iji and (bn/bwi) are constant, the 2-average molecular weight Mz is directly obtained (24) W ,D. Lansing and E. 0.Kraemer, THISJOURNAL, 67, 1369 (1936).

3940

JAMES

S. JOHNSON

AND

from observations based on refractive index gradients. Thus

KURTA. KRAUS

VOl. 78

perimentally determined Vzrc~,= ca. 39 cc. (in 1 211 HC1-1 A 1 NaC1) by the equation VZZrOCI,

= V z m

- 2VECI + V"m0

(7)

with the assumption that YH~O= 18 cc. and PHCI = 19.9 cc. From this, z7zr~clr= 0.096 cc./g. was obtained, with 92 as the atomic weight of the metal Mz/ML = N z (5) (Zr adjusted for Hf content). Earlier, $HfOCl2 = where wi is the weight of species i in a given volume 0.06 cc./g. had been found. The degree of hyof solution. drolysis undoubtedly varies with acidity. HowIf the data were in terms of concentrations, rather ever, the effect on the computed degree of polymerthan refractive index gradients (;.e., if equation 4 ization is small. For example, if the monomer were in terms of S rather than Se), the weight aver- unit were Zr(OH)3Cl,rather than ZrOClz, then for ages, M , and N,, would be obtained. Weight aver- z' = 0, p = 1.05, Mi (1 - f i z p ) would be 12% lower ages defined by the equation and the computed degree of polymerization 1270 higher. The percentage error for Hf (11') would be less. Since details of hydrolysis are unknown, ZrOClz and HfOClz were assumed to be the monomer units in all computations. tend to be lower than Z-averages, since the latter The apparent molal volume of ZrCll in a s o h emphasize heavier species. If we compute an av- tion containing 0.0365 M ZrClc and 1 M NaCl was erage degree of polymerization Ne from the ob- 42 CC. For present purposes this is not much served slopes S, by equations 4 and 5 without con- greater than the value found in 1 M HCI-1 M Nasidering possible charge on the polymers, the com- C1, where polymerization is less. It appears that puted degree of polymerization will be equal to or errors from assuming Vz independent of acidity are lower than the Z-average. Actually, if the charge on not great. the polymer is large, N e might even become smaller 3. Polymerization of Hf(1V) a s a Function of than a weight average N,. A rough estimate of Acidity.-The results for Hf(1V) are presented in this effect was made for one of the more unfavor- Fig. 2 as a plot of Ne us. excess acidity (in excess of able cases, Experiment 19, Table I. If the poly- the composition ZrOClz). The range of N e given mer had z' = 1, the maximum plausible value, N e for each acidity was computed from S: a t radii 6.3 would be 45% lower than N,. At the other ex- and 6.9 cm., points approximately 3 mm. from the treme. for z' = 0, Ne would be equal to NZand 40% edges of the cell. Concentrations, c;, a t the radii Since it is unlikely that z' is selected, bracket the initial concentrations of the SOhigher than N,. either as high as one or as low as zero in this me- lution. Closed symbols are used, if S: either did not dium, Ne should not differ greatly from N , and in vary as a function of x or decreased slightly with inany event may be used for semi-quantitative es- creasing x, in the direction expected for a charge timation of aggregation and for comparison of Zr- effect. (IV) with HfrIlT)r Hf (IV) a t ca 0.05 M i s apparently somewhat poly2. Monomer Unit and Partial Specific Volume. merized even in 5 M HCI (Ne = 1.2 and 1.8). -Computation of degree of polymer'lzation requires Some doubt results from the fact that sedimentaevaluation of Mi and B z , for which an estimate of the tion of HCl a t this concentration is accompanied degree of hydrolysis n is needed. From solubility16 by appreciable increase in its activity coefficients and hydrolysis data,1° 2 n 2.3 was estimated with increasing radius. If the effect on the activfor the monomer unit for Hf(1V) in HCI. The ity coefficients of the polymers is in the same direcsame arguments apply to Zr(1V). For ZrOClp, the tion, the computed values of Ne should be too low. partial molal volume was computed from the ex- In 3 M HC1, Hf(1V) is polymerized and polydisperse ( N e = 2.0 and 2.4). I n the range 0.2 to 2 , -T----'5 I 1 1 1 HC1, N e is ca. 3 and essentially independent of HC1. This flatness suggests monodispersity and, as mentioned earlier, detailed study' of polymerization in 1 M HC1, with consideration of charge, indicated that the species was probably either a trimer or tetramer, and that z' = ca. 1. An indication of slight polydispersity in 0.2 M excess HCI became definite in 0.08 M acid, where Ne is also considerably greater. The curve is dotted in this range to indicate that equilibrium was not reached (centrifugation time ca. 10 days). Since S: was still increasing, though slowly, when the experiment was terminated, the values of N e = 5.7 and e 4 r r 7.1 given for this acidity are minimal. 4. Polymerization of Zr(1V) as a Function of Acidity.-The results for zirconium are similar to MOLARITY E X C E S S H C I those for hafnium, though fewer centrifugations Fig. 2.-Polymerization of Zr(1V) and Hf(I\') as a function were carried out. The curve (Fig. 2) was drawn of acidity. with the same general shape as that for Hf(1V).

<