OLATIONDURING HETEROGENEOUS DEHYDRATION
Oct. 5, 1953
nickel(I1) and zirconyl chlorides whose compositions lie within the ranges studied have been demonstrated as intermediates in the dehydration reactions. New hydrates have been found whose previous lack of identification may be traced to either (l) Properties similar to those Of known hydrates Or (2> absence Of data On the composition range wherein they exist. The results are summarized in Table I. When hydration conditions for zirconyl chloride
4807
were adjusted so as to minimize uncertainties caused by hydrolysis, the results were analogous to those obtained by dehydration. Ac~ow~edgment.-To the Abbott Fund of Northwestern University and the Veterans Administration, the authors gratefully express their thanks for financial aid which helped to make this research possible. EVANSTON, ILLINOIS
[CONTRIBUTION FROM THE CHEMICAL LABORATORIES O F NORTHWESTERN UNIVERSITY]
Properties of Hydrates. 11. Olation during Heterogeneous Dehydrations ; Structural Formulas for Some Metal Dichloride Hydrates’ BY WILLIAMS. CASTOR, J R . , ~AND FRED BASOLO RECEIVED MAY23, 1953 Compounds in each of the solid systems MnClzccHzO, NiClz%HzOand ZrOClz.xHzOare characterized by the presence of coordinated hydroxyl groups and water molecules. As a consequence of the structural features deduced here, each of these systems reveals a peculiar dehydration behavior which can be interpreted as a series of olations. The number and kind of these condensations are determined by the number of hydroxyl groups and the geometry of the coordination. Each monomer first yields a diol dimer. The nickel and zirconyl complexes condense again, and the latter does so a third time, before olation is complete. Further dehydration then causes a decrease in coordination number from the symmetrical configuration first obtaining t o the more usual symmetrical arrangement for each metal ion. The nickel and manganese salts remain chloro complexes, but the zirconyl salt probably does not. A structure suggested for the latter can account for its hydrolysis to 8-coordinated zirconia upon thermal decomposition. The stronger metal-chlorine bonds of the other complexes favor decomposition t o anhydrous chlorides.
I n the f i s t paper of this series3 it was shown that the hydrates of manganese(I1) chloride are related by the reactions MnC12.4Hz0 = MnCl~.3l/zHzO= = MnCln.3Hz0 MnCl2.2Hz0 =
+ + + +
M ~ C ~ Z . ~ ~ / Zl / H z H~~O O MnClz.3HzO l/zH~O MnC12.2H20 HnO MnCl2.H20 HzO
Similarly, the dehydration of NiC12.6H20 proceeds first by a double loss of ‘/2H20; and dehydration of ZrOClz-9H20 gives first a twofold loss of ‘/2H20, then a fourfold loss of ‘/4H20, and then again a double loss of ‘/zHzO. This repeated occurrence of fractional hydrates is unusual, but perhaps only because other salts have not yet been scrutinized so closely. There is a certain regularity in the water losses, and it is the purpose of this paper to present an explanation of the phenomena involved. Evidence for Coordination of Hydroxyl Groups. -The nature of the metal-oxygen bonds in hydrates has been the subject of numerous investigations, 4-13 and many conflicting hypotheses have (1) Based primarily upon a part of the thesis submitted by William S. Castor, Jr., to the Graduate School of Northwestern University in September, 1950, in partial fulfillment of the requirements for the degree of Doctor of Philosophy. (2) American Cyanamid Company, Calco Chemical Division, Piney River, Va. (3) W. S. Castor, Jr., and F . Basolo, THISJ O U R N A L , 76, 4804 (1953). (4-13) The references are intended to be illustrative, not complete. (4) L. Pauling, “The Nature of the Chemical Bond,” Cornell University Press, Ithaca, N. Y., 1950, p. 114. (5) J. P. Hunt and H . Taube, J. Chem. Phys., 18,757 (1950). ( 8 ) H. L. Friedman, H. Taube and J. P. Hunt, ibid., 18,759 (1950). (7) J. P. Hunt, A. C. Rutenberg and H. Taube, THISJOURNAL, 74, 288 (1952). (8) H . J. Emel6us and J. S. Anderson, “Modern Aspects of In-
been advanced. Although there is room for argument about details, the various data all seem consistent with the view that the linkages are generally ionic with some “covalent character.”14 Since manganese(II), nickel(I1) and zirconium(1V) exhibit marked coordination tendencies, they might therefore be expected to show evidence for such “semi-covalent” bonding in their hydrates. Certain observations have, in fact, revealed the existence of coordinated hydroxyl groups in the systems of interest. Despite repeated references16J6 to the zirconyl ion as Zr=O++, its existence as an oxy monomer is extremely doubtful. Schmid” suggested pyro-linkages as an improvement, and these are known to be possible.l 8 Since a reasonable coordination for the zirconium ion must also be maintained, it can be concluded that the zirconyl ion is either an aquohydroxy monomer or some form of polymer. If organic Chemistry,” D. Van Nostrand Co., Inc., New York, N . Y., 1947, pp. 154ff. (9) C. A. Beevers and C. M. Schwartz, 2. Krisl., 91, 168 (1935). (10) L. Hackspill and A. P. Kieffer, Ann. chim., [IO], 14, 227 (1930). (11) A. W. Thomas, “Colloid Chemistry,” McGraw-Hill Book Co., Inc., New York, N . Y., 1934, pp. 141ff. (12) J. Louisfert, J . bhys. radium, 8, 45 (1947). (13) M. Prasad, S. S. Dharmatti, C. R. Kanckar and N . S. Biradar, J . Chcm. P h y s . , 17, 813 (1949). (14) 0. K. Rice, “Electronic Structure and Chemical Binding,” McGraw-Hill Book Co., Inc., New York, N . Y . , 1940, pp. 421-426. (15) F. P. Venable, “Zirconium and Its Compounds,” American Chemical Society Monograph Series, New York, N . Y . , 1921, p. 32. (18) W. B. Blumenthal, J. Chem. Education, 26, 472 (1949). (17) P. Schmid, Z. unorg. allgem. C h e m . , 167,369 (1927). (18) Consider, for example, the structures of the dioxides [R. W. G. Wyckoff, “Crystal Structures,” Interscience Publishers, Inc., New York, N. Y., 1948, Sec. I , Chap. IV, text pp. 3 4 1 and of the oxysulfide [J. D. McCullough, I,. Brewer and L. A. Bromley, A c f a C r y s f . , 1, 287 (1048)l.
4808
w.s. CASTOR, JR.,
AND
FREDBASOLO
VOl. 75
monomeric forms exist, they should show one or fractional hydrates result from olation condensamore reactions characteristic of coordinated hy- tions. droxyl groups. One such reaction, a polymerizaFixing the Initial Coordination Numbers.-To tion, has been observed in chloridelo and perchlo- ascertain the coordination number of the metal ratez0solutions and carefully studied in the case atom in an aquohydroxy complex, we must rely of the perchlorate.21 Although the condensation upon an assumption which has been invoked frehas not yet been completely characterized, it has quently in the past.1° It may be stated as follows : previously been related to coordinated hydroxyl simple hydrate water will be lost more readily than groups in the zirconium species.22 the constitutional water of the aquohydroxy comThe observations and arguments of M e e r ~ e i n ~plex. ~ The statement is not axiomatic, for it improvide evidence for coordinated hydroxyl groups plies a thermodynamic criterion and kinetic considin the hydrates of nickel(I1) and manganese(I1) erations are neglected. However, upon the basis of chlorides. Concentrated solutions of these salts this reasonable assumption, it may be concluded that exhibit acidic hydrogens which cannot be attributed the olating species contain no uncoordinated water. to hydrolysis, but can be understood in terms of The half-hydrates characteristic of a primary olaH[M(OH) . . .] groupings. The effect disappears tion occur directly upon dehydration of either rapidly upon dilution. In each case the remainder MnClr4HzO or NiC12-6HzO. In the case of zirof the coordination sphere of the cation must be cony1 chloride, half-hydrates occur at two places in satisfied by anions and/or undissociated water the dehydration-during decomposition of ZrOC12. molecules. Since coordinated ligands present in 9Hz0 and of ZrOClz.7H~0. However, the 8-hyconcentrated solutions are probably also in evi- drate decomposes to lose 1/aH20 four times, the dence in the crowded environment of the highest behavior expected for olation condensation of a diol solid hydrates, the literature cited indicates that dimer, so the first olation begins with the decompothese compounds are aquohydroxy complexes. sition of the 9-hydrate. It appears, therefore, that Identification of the Olation Reaction.-A com- MnClv4Hz0, NiCI2.6Hz0 and ZrOCl2-9HzO are mon property of aquohydroxy complexes is their aquohydroxy monomers in which all the water is ability to undergo the olation reaction11v24 coordinated. Since both the manganese and nickel salts have II been shown to contain acidic hydrogenslZathe chlo/o\ H 2 0 2[R---M(OH)(H20)] +I = [R(H20)M i'vf(OH)R]f2x ride ions must be included in the coordination If the condensing species are identical, a second ola- sphere with the water. Because the monomers are stable in equilibrium with8 an aqueous solution, tion leads to the formation of a cis-diol bridge the alternative treatment would be equivalent to H H writing a non-existent hydrolysis. Noting that the /4 /o\M-R]+2x [R(HsO)M M(OH)R]+'" = [R--M IlsO dehydration behavior of zirconyl chloride reveals olations characteristic of three hydroxyl groups, it '0' H appears that ZrOC12.9HtO also contains a MeerA single hydroxyl group in the monomer results in wein-type acidic hydrogen. Hence, the same arguthe elimination of one H20/M in two steps of ment about chloride coordination should apply. Upon the basis of the foregoing discussion, i t is '/zHzO each. If the R groups in the above formulations were such that each metal atom in the diol possible to write preliminary structural formulas dimer still retained one coordinated, unshared hy- for the three monomers and examine the coordinadroxyl group and a t least one coordinated water tion numbers of the metal ions27 molecule, olation could proceed. However, since MnC1~4H20 = H IMn(HzO)a(OH)Cle] the condensing species would then be dimers, four NiC1~6H20 = H [Ni(H20)6(OH)C12] consecutive losses of l/4H20 per metal atom would ZrOC12.9H20 = H [Zr(H20).1(0H)~C12] occur. Additional hydroxyl groups would permit These notations are tentative because the argument still further condensation. Heretofore, the olation reaction has been studied has not yet been carried far enough to completely fix chiefly in solution,24but the concept has been ex- the number of acidic hydrogens. However, the tended to the hydrous oxides by Thomas, Graham one shown in each case above is the minimum, and and other^.",*^.*^ The water losses described are additional acidity could not influence the deterthose found in the dehydration of MnCI2.4H20, mination of the coordination number. The initial coordination numbers deduced above NiC12.6Hz0 and ZrOC12.9Hz0.3 Since other evidence also points to formulation of these com- for the hln(II), Ni(I1) and Zr(1V) ions are 6, 8 and pounds as aquohydroxy complexes, it seems neces- 12, respectively. All are compatible with the exsary to conclude that the repeated occurrences of pectation of symmetrical configurations in predoniinantly ionic complexes, but the smaller symmetri(19) hl. Adolt and W . Pauli, Kotloid-Z., 29, 173 (1921). cal arrangements with coordination numbers 4, (i ( 2 0 ) G. Jander and K . 17, Jahr, Kolloid-Beihefle, 49, 295 (1935). and 8 might have been anticipated. The enhance(21) R. E. Connick and 1%'. €1. Reas, THISJOURNAL, 73, 1171 (1951). ment is very probably related to one previously (22) R. E.Connick and W. II. McVey, i b i d . , 71, 3182 (1949). observed. Pauling2*lists several cations whose co(23) H. Meerwein, Sitzber. Ges. Hefufder. Res. Naturw. Marburg,
+
+
64, 119 (1930): Chcin. Z r n l v . , 101, 11, 1!)62 (1830). ( 2 4 ) H J Eniri6us and J S. Anderson, ref. 8 . pp. 128-131. ( 2 5 ) K. P. N C H ~ C H ~ N (CHzCOO- t -OOCCH? /+
The differences between the third and fourth dissociation constants are too large in view of present methods of predicting dissociation constants of organic substances. It was hoped that the thermodynamic changes involved would help to explain this discrepancy. Experimental The method employed for the determination of thermodynamic dissociation constants consisted of the extrapolation t o zero ionic strength of the electromotive force of cells without liquid junction containing a buffer system consisting of appropriate ethylenediaminetetraacetate anions. The cells employed may be divided into two general classes I. Pt-€12, KzHzV(mi), K3HV( mz), KCl(m3). AgC1-Ag 11. Pt-Hz, KaHV(mi), KaV(mz), KCl(m3), AgCI-Ag where H4V represents ethylenediaminetetraacetic acid. Cell I, which contained the di- and tripotassium salts of ethylenediaminetetraacetic acid was used for the evaluation of k3, and cell 11,containing the tri- and tetrapotassium salts of ethylenediaminetetraacetic acid, was employed for the determination of k,. The electromotive forces of these cells were measured from 0 to 30". The values for 0' were evaluated by extrapolation of data taken a t slightly higher temperatures. Materials and Equipment.-The ethylenediaminetetraacetic acid was obtained through the courtesy of the Bersworth Chemical Company, Framingham, Massachusetts. It was further purified by two successive recrystallizations from water. The buffer solutions employed in cells of type I and I1 were made up by adding to the purified ethylenediaminetetraacetic acid the required amount of carbonatefree standard potassium hydroxide solution, prepared by the method of Schwarzenbach and Biedermax~n,~ and standardized against potassium acid phthalate in the usual manner. The potassium chloride used was freed from bromide contamination by recrystallization from water and 95% ethanol as outlined by Pinching and Bates.s The silver-silver chloride electrodes were prepared by the method of Shedlovsky and MacInnes,6 and the platinum-hydrogen electrodes were made according to the direction5 outlined in Weissberger.' Tank hydrogen was first passed through a deoxo purifier, a presaturator, and then through (4) G. Schwarzenbach and SV. Biedermann, I I e h . Chim. A r l u , 31, 331 (1948). (5) G. Pinching and R. Bates, J . Reseavch N u l l . Bur. Slnndovds,
87, 311 (1946).
(1) F. F. Carini and A. E. Martell, THIS JOURNAL, 74, 5746 (1952). (2) 0. Schwarzenbach and H. Ackermann, Hele. Chim. A d a , SO, 1798 (1942). ( 3 ) M . J. Cabell, A. E. R. E. Report C / R 813, Ministry of Supply, IIarwell, Berks., England, I q R I .
(6) T. Shedlovsky and D. A. MacInnes, THIS JOURNAL, 68, 1970 (1936). (7) A. Weissberger, "Physical Methods of Organic Chemistry," Second Bdirion, Vol. 11, Interscience Publishers, Inc., R e w York, N. Y., 1949, p, 1722,
