5858
J . Phys. Chem. 1988, 92, 5858-5860
Influence of Cation Coordination Numbers on Transport Properties of Ionic Liquid Mixtures Anselmo Elias Department of Chemistry, Lisbon University, R . Escola Politecnica, I294 Lisbon Codex, Portugal
and
C. Austen Angell*
Department of Chemistry, Purdue University, West Lafayette, Indiana 47907 (Received: June 28, 1988)
We test a conjecture, based on spectacular differences between FeCI3-basedand CrC13-basedsolutions, that transport properties in molten salt solutions are conditioned by the coordination number that the highest z/r cation adopts in the solution. To do this we study a case in which change of anion type dictates a change of cation coordination number from 4 to 6. We use additions of fluoride ion to a molten chloride solution containing Co2+in one case and Ni2+in a second, and compare the ionic conductivities with those of isostoichiometricsoltktions containing only chloride anions. In the presence of fluoride both solutions show lower conductivities than in the all-chloride equivalent, consistent with formation of some MCls4- sites, and the difference is greater for Ni(I1) solutions, consistent with the greater preference of the ds ion for octahedral sites.
Introduction
characteristics of molten salt mixtures.
In an earlier publication' we demonstrated that the effect of complex ion-forming second components on the electrical conductivity and viscosity of a binary solution with a common first component was strongly influenced by the geometry of the complex ion being formed. In brief, it was found that addition of FeC13 to a monovalent cation chloride leads to conductivity increases and viscosity decreases due to the formation of FeC1, species while addition of CrCI3 as second component produces the opposite effects due to the formation of octahedrally coordinated CrCls3-. In this case the coordination states which are responsible for the contrasting transport behavior are driven by the electronic characteristics of the cations in the added second components, in particular by the octahedral site stabilization energy in the case of the d3 Cr3+ ion.* It is of interest to see whether similar effects on the transport properties of binary or pseudobinary mixtures can be induced by the formation of octahedral coordination states when the driving force to octahedral coordination has a different origin. For instance, in a melt containing small as well as large anions an added cation which has the appropriate size could choose to coordinate sixfold with the smaller anion rather than fourfold with the larger anion. To explore this idea we have compared the effects on the electrical conductivity of separate addition of two divalent chlorides, NiC12 and CoCl2, to a common first component which contains both chloride and fluoride ions. Here the electronically driven preference of the d8 Ni(I1) ion for six-coordination could be manifested in the presence of fluoride ion, giving NiF,"- species (or a mixed ligand species), while Co(I1) would remain four-coordinated by chloride as tetrahedral CoCI,*-. Ni(I1) and Co(I1) are always observed to be tetrahedrally coordinated in basic pure chloride melts, even in the glassy ~ t a t e ,while ~ ? ~ both are found in octahedral coordination in pure fluoride melts.s.6 We choose conductivity measurements over viscosity measurements because of the relative simplicity and precision of the measurement. The present paper reports the results of such investigations and seems to confirm the conjecture that cation coordination states may be of predominant importance in deciding the transport
Experimental Section
(1) Angell, C. A.; Elias, A. J . Phys. Chem. 1983, 87, 4704. (2) Cotton, F. A.; Wilkinson, G. Advanced Inorganic Chemistry; Interscience: New York, 1972. (3) Gruen, D. M.; McBeth, R. L. Pure Appl. Chew. 1963, 6, 23. (4) Angell, C. A.; Abkemeier, M. L. Inorg. Chem. 1973, Z2, 1462. ( 5 ) Boston, C. R.; Liu, C. H.; Smith, G. P. Inorg. Chem. 1968, 7, 1938. (6) Wong, J.; Angell, C. A. Glass: Structure by Spectroscopy; Marcel Dekker: New York, 1976; Chapter 6 .
0022-3654188 l2092-5858$01.50/0 , I
,
The introduction of fluoride ions into any experimental system leads to serious difficulties when the experiments are conducted in glass vessels. For this reason, it has been necessary in the present project to develop techniques which avoid contact of the melts with glass containers. In a separate publication7 one of us has described the design and performance of a Teflon-based cell which has proved very satisfactory for this type of investigation. That publication also contained conductivity data for the primary solution of the present work, viz., a mixed halide melt (a-methylpyridinium (picolinium) cations with F and C1- anions) of composition a-picHF.4(a-picHCl). The conductivity of this solution proves to be slightly lower at all temperatures than that of pure a-picHC1 presumably because of the greater charge density on the fluoride ion and the consequent increase in Coulomb cohesion. In the present work no new techniques needed to be developed. Binary solutions containing varying fractions of NiC12 or CoClz were prepared in Teflon vessels in a drybox atmosphere and transferred to the conductivity cell for subsequent measurement. Conductivities were determined with a Wayne Curve B641 autobalance universal bridge operating at a fixed frequency of 1592 Hz. The cell constant was 27.91 cm-' and the resistance being measured varied between 0.5365 and 5.875 k!2 over the temperature range explored. Under these circumstances there is no problem with frequency dependence of the measured conductivity.8 Glass transition temperatures utilized in the subsequent interpretation of results were determined by using a simple differential thermal analysis apparatus employing copper-constantan thermocouples, the output of which was recorded on a Houston instruments XY recorder. Typical traces are shown in the inset to Figure 1.
Results and Discussion In Figure 1 we compare the conductivity of the a-picHF.4(apicHC1) solution with that of the pure a-picHC1 reported in ref 7. We note the strong non-Arrhenius character of the conductivity in each case, consistent with the approach to the glassy state at the lower temperature Tg = 230 K (a-picHC1) and 233 (apicHF-4-a-picHCl)-see Figure 1 insert. In Figure 2 we show Arrhenius plots for the three binary solutions of CoCl, dissolved in the fluoride-chloride melt in Ar(7) Elias, A. Rev. Sci. Instrum. 1988, 88, 339. (8) Jaffe, I. S.; Van Artsdalen, E. R. J . Phys. Chem. 1956, 60, 1125.
0 1988 American Chemical Society
The Journal of Physical Chemistry, Vol. 92, No. 21, 1988 5859
Letters
- I
I 300.
Y \
I250'
1
0
a-picHCI
o
a-picHF. 4a -picHCI I
200 J
0
2.2
2.4
2.6
5
10
Figure 1. Arrhenius plot of electrical conductivity of a-picolinium
chloride (solid circles) and the mixed fluoride-chloride solution, apicHF.4(a-picHCl). Inset: DTA traces showing the glass transition and definition of TBfor each liquid.
.
20
2.8
x coc12
103K/T
3
15
Figure 3. Glass transition temperatures TB,and VTF equation parameter To(.), derived for constant E , = 650 K. The values of A , parameter are
also shown.
-
0.00 % COCI, 7.38 %
6 -
E,
13.97 %
y
q/
V
N CI2
b COC12
T 130'c
N
0-
5 -
10
P
0
E
4 4 1
3
5
2.4
2.5
2.6
l o 3K /
2.7
3
2.8
T
Figure 2. Arrhenius plot of electrical conductivity for binary solutions of CoC1, dissolved in a-picHF.4(a-picHCI).
rhenius form. As in the case of the all-chloride system cypicHC1-CoC12 reported separatelyg the composition dependence of the conductivity is very weak. This is consistent with the behavior of TKplotted in Figure 3, since TKmirrors the behavior of the ideal glass transition temperature Towhich is the controlling parameter in the Vogel-Tammann-Fulcher equation where A,, B,, and To are constants. The values for A and To obtained for the assumed condition B, = constant (=650 K) are displayed in Figure 3 where it can be seen that the To parameter essentially parallels the values of the glass transition temperature. An alternative parametrization in which Tois kept at a constant would result in an almost constant value of A , fraction, 0.93 TK, and a weakly decreasing value of B,. The range and precision of the data are not enough to determine by least-squares best fitting which of these alternatives is most appropriate to the present system, though the latter would be consistent with concept of the preexponent having a value related to a barrier-crossing attempt frequency,' which should vary only weakly with composition. Of greater importance to the present project is the composition dependence of the conductivity, in particular the comparison of the earlier results for the all-chloride systems and the results for the present systems in which sixfold-coordination of the divalent cation by fluoride ions is a possibility. This is shown in Figure (9)Elias, A.; Angell, C.A. J . Chem. Eng. Data 1988, 32, 1.'
4 for the four systems in question. We observe from Figure 4 that in both NiC12 and CoCl2 cases the conductivity in the fluoride-containing solution increases with divalent cation chloride additions less rapidly than in the corresponding all-chloride system. The difference is particularly pronounced in the NiC12-containing system which we noted at the outset should be the one in which six-coordination should be most favored. The differences are highlighted in the insert, part c, in which the value of Au, the difference between all-chloride and fluoride/chloride solutions is seen to be distinctly larger for the NiC12-containing solutions. This is fully consistent with the recognition that Ni2+, with its d8 electronic configuration and consequent octahedral site stabilization energy, will have a greater tendency to enter six-coordination than will Co2+. It is unfortunate that it was not possible to investigate solutions with higher fluoride ion contents due to the low solubility of the divalent fluorides. It will be desirable in future work to confirm, by appropriate spectroscopic studies, the assertion that the conductivity differences observed in this work are indeed associated with the preponderance of sixfold-coordination in the nickelcontaining over cobalt-containing solutions. For instance, thin-film visible spectroscopy could, in principle, use the electronic d-d transition spectra of the divalent cations to quantitatively define the relative concentrations of octahedral and tetrahedral coordination states of Co(I1) and Ni(I1) present in the solutions. In view of the much greater anion polarization (by high z/r next neighbors) needed to obtain Co(I1) in octahedral coordination in
5860
J . Phys. Chem. 1988, 92, 5860-5861
chloride melts,lOillsuch a study would be of considerable value in its own right. Assuming that spectroscopic studies fully confirm the conductance/structure relations that we seem to have established, the question will remain, how exactly does the presence of a large cation coordination number for the higher charged cation cause a generally smaller particle mobility than in a melt of the same composition in which the same cations have a lower coordination number. Our best explanation is that complexation effectively screens the charge on the cation so that the cohesion of the melt is determined by the Coulomb interaction between the lower charged cations and the complex anion. If the latter is of high coordination number than the residual negative charge on the complex is large (-4 for [CoCI6]) and so the cohesion is large. By contrast, if the coordination number is smaller then the residual negative charge is smaller (-2 for [CoCI4]) and the cohesion is correspondingly smaller. These effects were discussed in more detail in an earlier paperI2 in which the argument was supported (10) Oye, H.A.; Gruen, D. M. Inorg. Chem. 1965, 4 , 1173. (11) Angell, C. A,; Gruen, D. M. Inorg. Nucl. Chem. 1967, 29, 2243.
by results from computer simulation experiments.
Conclusions Comparison of conductivity changes on dissolution of divalent chlorides in solutions containing either all-chloride or mixed chloride-fluoride ligands are consistent with the notion that the conductivity and fluidity of solutions containing complex ions are greater if the complex ions formed are of low net charge than when they are of high net charge. The charge on such complexes is determined by the coordination number for fixed cation charges. Accordingly tetrahedral complexes favor high fluidities and conductances while octahedral or dodecahedral complexes produce higher viscosity, lower conductance, solutions. Acknowledgment. This work was supported by a grant from the Instituto Nacional de Investigacao Cientifice (INIC), and by the U S . National Science Foundation under Solid State Chemistry Grant DMR 8304887. (12) Angell, C. A.; Hodge, I. M.; Cheeseman, P. A. Proceedings of the International Conference on Molten Salts; Pensler, J. P., Ed.; The Electrochemical Society: Pennington, NJ, 1976; p 138.
An Evaluation of Water-Water Analytical Potentials in the Region of Low-Energy Trifurcated Structures M. Mezei* and J. J. Dannenberg Department of Chemistry and The Center for Study in Gene Structure and Function, Hunter College and The Graduate School, City University of New York, New York, New York 10021 (Received: July I , 1988)
Several pairwise additive potentials, frequently used to describe liquid water, and a cooperative potential are examined with respect to their abilities to adequately describe the trifurcated water dimer structures recently reported to be of energies comparable to the best linear structures. All pairwise additive potentials were in error of at least 3 kcal/mol while the cooperative model was within 1 kcal/mol of the best quantum-mechanical results. It is suggested that, while these pairwise additive potentials may be of considerable use for the description of aqueous water, they should be used with caution where interactions involving individual water molecules or pairs are important.
Introduction The correct description of water molecule interactions is of fundamental importance as it is essential to the critical study of biological systems. Both theoretical] and experimental2 studies have indicated that a linear hydrogen-bonding structure corresponds to a minimum on the potential energy surface for the dimer water dimer. A recent report3 has shown, however, that another dimer structure that forms trifurcated hydrogen bonds is within 0.2 kcal/mol of the best linear structure as calculated by using the MP4SDQ/6-3 1 1G** ab initio molecular orbital method: after optimization at the HF/6-31G* or MP2/6-31G* level. Several analytical potentials for water interactions have been developed in the past several years and used in computer simulations to describe liquid water. It is of significant interest to determine the extent to which they are able to describe this, and other, structures which have not been previously considered. In this note, we examine the water dimer potential energy surface (1) The following are representative of calculations using large basis sets and correction for electron correlation: (a) Frisch, M. J.; Pople, J. A,; Del Bene, J. E. J . Phys. Chem. 1985.89, 3664. (b) Frisch, M. J.; Del Bene, J. E.; Binkley, J. S.;Schaefer, H. F. 111 J . Chem. Phys. 1986,84, 2279. (c) Del Bene, J. E. J . Chem. Phys. 1987,86,2110. ( d ) Diercksen, G. H.F.; Kraemer, W. P.; Roos, B. 0. Theor. Chim. Acta 1975, 36, 249. See also ref 5-7. (2) Dyke, T. R.; Mack, K. M.; Muenter, J. S. J . Chem. Phys. 1977, 66, 498. (3) Dannenberg, J. J. J . Phys. Chem., in press. (4) GAUSSIAN-82; Binkley, J. S.; Frisch, M. J.; DeFrees, D. J.; Ragavachari, K.; Whiteside, R. A.; Wchlegel, H. B.; Fluder, E. M.; Pople, J. A .
0022-3654/88/2092-5860$0 1.50/0
in the region of these trifurcated and other structures. To this end, we compare several pairwise additive water-water potentials proposed in the literature5-12and frequently used in liquid studies, as well as the cooperative model of Campbell and Mezei'j with the ab initio results.
Calculations The calculations examined the six dimer structures considered in the ab initio study (see Figure 1 and Table I).) Structure I is the global minimum predicted by the AM1 semiempirical molecular orbital method;14 I1 was obtained by a partial optimization starting from 1 using the GAUSSIAN-82 program4 at the ( 5 ) Matsuoka, 0.;Clementi, E.; Yoshimine, M. J . Chem. Phys. 1976,64,
1351. (6) Clementi, E.; Habitz, P. J . Phys. Chem. 1983, 87, 2815. (7) Yoon, B. J.; Morokuma, K.; Davidson, E. R. J . Chem. Phys. 1985,83,
1223. (8) Stillinger, F. H.; Rahman, A. J . Chem. Phys. 1975, 60, 1545. (9) Berendsen, H.J. C.; Postma, J. M. P.; van Gunsteren, W. F.; Hermans, J. In Jerusalem Symposia on Quantum Chemistry and Biochemistry; Pullman, B., Ed.; Reidel: Dordrecht, Holland, 1981. (IO) Jorgensen, W. L. J . Chem. Phys. 1982, 77, 4156. (,l1) Jorgensen, W. L.; Chandrashekar, J.; Madura, J. D.; Impey, R.; Klein, M. L. J . Chem. Phys. 1983, 79, 926. (12) Jorgensen, W. L.; Madura, J. D. Mol. Phys. 1985, 56, 1381. (13) Campbell, E. S.; Mezei, M. J . Chem. Phys. 1977, 67, 2338. (14) Dewar, M. J. S.; Zoebisch, E. G.; Healy, E. G.; Stewart, J. J . P. J . A m . Chem. Sor. 1985, 107, 3902. (15) Zeiss, G. D ; Meath, W. J. Mol. Phys. 1977, 30, 161; 1977, 33, 1155.
0 1988 American Chemical Society
