Ion Association and Solvation in Dichloromethane of Tetrachloro- and Tetrabromoferrates(III) Compared with Simple Halides

Home , Ion association

S. Bait, G. du Chattel, W. de Kieviet, and A. Tieleman Department of Inorganic Chemistry Free University, De Lairessestraat 174, Amsterdam, The Netherlands

Z. Naturforsch. 33b, 745-749 (1978); received April 26, 1978 Ion Association, Dichloromethane, Tetrachloroferrate(III), Tetrabromoferrate(III), Hydrogen-bridge Formation Conductivity measurements are reported in dichloromethane for tetraethylammonium and tetraphenylarsonium salts of tetrachloro- and tetrabromoferrate(III) and halide ions (chiefly chloride and bromide). Analysis of the data in terms of the recent Fuoss equation has given the ion-pair association constants. The experimental values are close to the ones calculated from the electrostatic Fuoss-Eigen equation. The solvent acts as an acceptor by way of forming hydrogen-bonds to the simple halide ions and to a much less extent to the ferrates. Solvation of chloro- and bromoferrates(III) seems to be comparable, as deduced from the association constants and NMR paramagnetic line broadening. The phenyl groups of the tetraphenylarsonium ion seem to give ^-interaction with the solvent.

Introduction proved to be a very good starting point, because it has negligible donor-properties [1, 2], but is a In the study of kinetics and mechanism of co- remarkably good solvent for ion-paired complexes. ordination compounds the emphasis seems to be In addition the symmetrical nature of the anion shifting from aqueous to non-aqueous solvents. In makes it in combination with symmetrical cations this connection a (semi)quantitative knowledge of a model case for treating ion-association with cur- donor and acceptor properties [1] of the various rent conductivity theories. solvents is of crucial importance [2]. It has been recognised [2, 3] that the formation of outer-sphere As a preliminary for the kinetic work we report complexes is an essential part of the reaction route here the ion-association constants for tetraalkyl- for substitution reactions. This means that the ammonium and tetraphenylarsonium tetrahalofer- relevant outer-sphere association constants must fates(III) and halides and the ensuing conclusion be available for a complete elucidation of reaction about the differences in solvation between simple mechanisms. It has been shown [4, 5] that a purely and complex ions. electrostatic model [6, 7] is inadequate to account This study is of separate interest because it sheds for the differences in association constants in vari- light on the acceptor properties of the solvent di- ous solvents, but that specific nuoleophilic and chloromethane. Additional information on solvation electrophilic properties of the solvents have to be has been obtained from NMR studies. considered. Also a comparison of kinetic data for the substitution reactions of solvated nickel(II) Experimental ions with N-donor ligands has shed doubts on the Chemicals adequacy of the purely electrostatic model [8, 9]. Tetraethylammonium chloride, bromide and io- Substitution reactions that have been studied in dide, tetra-w-butylammonium chloride and tetraphenylarsonium chloride were obtained commer- this respect nearly always involve the solvent as an cially in reagent grade quality. The remaining non- active incoming or leaving ligand. It seemed there- metal salts were prepared by double exchange from fore of interest to study a system that is stable the appropriate compounds in anhydrous ethanol. enough to remain intact in various solvents. This The ferrates were prepared by adding equivalent requisite is met by the tetrahaloferrate(III) ions, amounts of the anhydrous iron(III) halides and the tetraethylammonium or tetraphenylarsonium hal- if solvents are chosen with comparatively weak ides in anhydrous ethanol [10]. Tetraethylammo- donor-properties. To this end dichloromethane nium chloride and bromide were recrystallized from an acetonitrile-ethyl acetate (1:2) mixture. The remaining compounds were recrystallized from Requests for reprints should be sent to Dr. S. Bait, anhydrous ethanol under dry nitrogen until the Scheikundig Laboratorium Vrije Universiteit, De analysis results were better than 1% relative for Lairessestraat 174, Amsterdam, The Netherlands. Fe and 2% relative for the halide.

Dieses Werk wurde im Jahr 2013 vom Verlag Zeitschrift für Naturforschung This work has been digitalized and published in 2013 by Verlag Zeitschrift in Zusammenarbeit mit der Max-Planck-Gesellschaft zur Förderung der für Naturforschung in cooperation with the Max Planck Society for the Wissenschaften e.V. digitalisiert und unter folgender Lizenz veröffentlicht: Advancement of Science under a Creative Commons Attribution Creative Commons Namensnennung 4.0 Lizenz. 4.0 International License. 746 S. Bait et al. • Ion Association and Solvation in Dichloromethane

Dichloromethane (Merck) was further purified The tetraethylammonium salts of the mixed by treatment with activated carbon and dried over ligand complexes FeCl4_nBrn_ have been prepared molecularsieves. The solvent, purified in this way, and their spectra recorded in acetonitrile [13]. We had a water content of 0.004 wt% (Karl-Fischer titration) and a specific electric conductivity better prepared solutions of the mixed ferrates in di- than 10~9 Ü"1 cm-1 (limit of measuring equipment). chloromethane by adding together solutions of the parent compounds in the calculated ratio. The Conductivity cell and apparatus reaction was complete within the mixing time. The The conductivity cell was based on the Kraus- spectra of the mixed compounds agree with those Erlenmeyer [11] using circular platinum coated glass discs of ca. 7 cm2 area, 0.3 cm apart. The cell published for acetonitrile as a solvent [13]. From constant was determined following the procedure mol-ratio curves at different wavenumbers it could described by Fuoss and coworkers [12]; a value of be inferred that only minor disproportionation 0.050 ± 0.001 cm-1 was obtained. The conductivity takes place. was measured with a Wayne-Kerr autobalance uni- versal bridge B641 provided with a Rohde and Schwarz SUB BN 4083 RC generator. A tempera- Ion-pair association constants from conductivity ture constancy better than 0.01 °C was obtained measurements by immersion of the cell in a thermostatted bath The conductivity data were analyzed as de- filled with light mineral oil. The temperature was measured inside the cell with a platinum resistance scribed by Fuoss [14]. The bulk of the conductivity thermometer. data belong to the region under the limit of 3.2 x 10-7 D3 M (2.3 X 10-4 M for the present case), Procedure for conductivity measurements set by Fuoss. Values of the dielectric constant D The experiments were performed in the dark to were taken from published Tables [15]. The viscosity prevent photochemical decomposition of the solvent was calculated [16] from the expression r\ = C.exp and the ensuing halide exchange and attack on the platinum electrodes. Solutions in the range (0.2 to (ß/kT) and rj values at 15 °C and 30 °C [17]. 5.0) X 10-4 M (mol dm-3) were made up from As expected [18, 19] generally more than one stock solutions of (2-5) X 10"3 M. A complete least-squares minimum was found, of which the one series for one compound at one temperature closest to the Bjerrum distance (for the present case consisted of 14 measurements of the equivalent conductivity as a function of the concentration 3100 pm) was selected. Fortunately, the effect of (the complete set of data is available on request). choosing the other minimum on the value of the After 1 h the temperature inside the cell and the thermodynamic association constant KA is small, resistance reading were constant. Variation of the only a few percent. Much more pronounced differ- measuring frequency from 5000 to 10000 Hz did not influence the measured resistance within the ences in KA were found when other conductivity experimental error ( < 0.5%). equations were used. For instance, the older Fuoss- Onsager-Skinner equation [20] gave significantly Spectra lower KA values, but also a pronouncedly poorer fit, X H NMR spectra were recorded on a Varian exemplified by twice the value of the standard EM 360 Spectrometer. In view of the band width analysis, care was taken to avoid saturation. UV deviation of the fit to the Fuoss equation. Fig. 1 spectra were recorded on a Beckman Acta M IV contains a comparison between experimental and Spectrophotometer using 1 cm quartz cells. calculated values using the two equations for a representative case (tetraethylammonium tetra- Results bromoferrate at 25 °C). The significant deviation at Solvolysis and identity of ferrates (III) the limit of high concentration may be indicative of The identity of the species in solution was incipient triple-ion formation. checked by recording ultra-violet and visible spec- It is immediately clear that the older equation tra. The UV spectra of the ferrates do not show has the effect of flattening the curve. This fact, in evidence of interaction with the anhydrous solvent addition to the poorer fit, made us prefer the KA (0.004 wt% water). The spectrum is constant over values from the more recent Fuoss equation. The the concentration range 10~5 to 10~3 M and the results of this analysis in terms of equivalent position and molar absorbances of the bands, at conductances at infinite dilution (Ao), KA values 363 and 313 nm (FeCLr) and at 472 and 391 nm (also referring to infinite dilution) and the Guerney (FeBr4~), are independent of the counter ions. co-sphere radius R, together with the standard 747 S. Bait et al. • Ion Association and Solvation in Dichloromethane deviation between observed and calculated molar ings at two other temperatures (5.5 °C and 16.5 °C). conductivities are listed in Table I. As the KA values as a function of the temperature For three compounds in Table I the temperature did not obey the van't Hoff equation, no further dependence of KA was determined by taking read- analysis of the temperature dependence was at- tempted. Results of conductivity measurements for the 100 mixed complexes are also in Table I.

NMR line widths 90 The excess line width compared to the pure solvent was determined as a function of the concentration of the tetraethylammonium salts of tetra- _ 80 chloro- and tetrabromoferrate(III). For comparison "o E one concentration of tetraphenylarsonium tetra- -u 70 chloroferrate was measured. Results are given in E Table II. Under the concentrations used, ion- -OC

X 60 Table II. Line broadening of H NMR signal of CH2CI2.

Compound 103 x conc. Excess line width ./.a -1 50 mol kg Hz

Et4NFeCl4 20.0 9.00 9.78 4.80 40 4.89 2.40 0.5 10 15 2.0 25 Et4NFeCl3Br 8.00 2.70

lOVconc (mol. I4)"2 Et4NFeBr4 4.07 2.25 3.26 1.65 Fig. 1. Equivalent conductivity of tetraethylammo- 2.44 1.20 nium tetrabromoferrate in dichloromethane as a function of the concentration. The experimental data (C6H5)4AsFeCl4 20.0 11.0 are compared with calculated least-squares lines obtained from a fit to the Fuoss-Onsager-Skinner ( ) a Ratio: IO-2 X excess line width/conc. and the Fuoss equation ( ). (in Hz kg mol-1).

Table I. Parameters obtained from the analysis Compound A0 10~3 X KA IO-2 X R a* of conductivity data using the Fuoss-1975 cm2 ß-1 eq-1 mol-1 dm3 pm theorya. Et4NCl 122 ± I 92 ±3 22 0.17 Et4NBr 132 ± 3 80 ±5 31 0.32 Et4NI 118 ± 5 42 ±5 30 1.0 Et4NC104 146 ± 4 54 ±4 30 0.53 B114NCI 104 ±2 41 ±3 29 0.40 Et4NFeCl4 146 ± 5 24 ±3 29 1.6 Et4NFeCl4c 129 ±2 19 ±1 28 0.65 Et4NFeCl4d 119 ±2 17 ± 27 0.76 Et4NFeCl3Br 144 ± 2 22 29 0.83 21 ± Et4NFeCl2Br2 141 ± 2 ± 28 0.70 Et4NFeClBr3 139 ±2 19 ± 29 0.78 EtjNFeBn 141 ±2 20 ± 27 0.88 c ± Et4NFeBr4 123 ±2 14 ± 27 0.81 Et4NFeBr4d 112 ± 1 13 26 0.68 (C6H5)4ASC1 105 ± 1 2.3 ±0.1 22 0.61 111 ± I 2.8 ±0.1 36 0.17 (CeH5)4AsBr a At 25.00 °C unless otherwise indicated (c and d). 117 ± I 2.9 ± 0.2 23 0.93 (CeHsl^ClC^ For definition of Ao, KA and R see the text. 116 ± 1 3.0 ± 0.2 23 0.93 (C6H4)4AsFeCl4 b Average standard deviation between calculated 0 107 ± I 2.6 ± 0.1 25 0.61 (C 6H5 )4 AsFeCU and experimental equivalent conductivities. 98 ± 1 2.4 ± 0.1 22 0.64 (CeHsJ^FeCl^ c Measured at 16.50 °C and 114 ± 1 3.0 ±0.1 0.80 (C6H5)4AsFeBr4 22 d at 5.50 °C. 748 S. Bait et al. • Ion Association and Solvation in Dichloromethane association is complete by 80-90% for the tetra- The haloferrates show the same picture to a less ethylammonium salts and 70% for the tetraphenyl- extent: The stability order Et4NFeCLi > Et4NFeBr4 arsonium salt. The NMR results consequently refer is only levelled and not reversed in the tetraphenyl- to the associated forms. At lower concentrations no arsonium case: (CeHs^sFeCU~(C6H5)4AsFeBr4. sufficiently accurate values could be obtained. This result means that specific solvation effects as The line broadening effect of the paramagnetic hydrogen-bridge formation are much less important ions on the XH NMR signal of CH2CI2 is insensitive in the complex ferrates, a conclusion that agrees to the nature of the counter ion, as the figures in with the NMR results. Table II will make clear. A similar effect has been It is interesting to note that the mixed complexes found in NMR studies of solvation of simple anions Et4NFeCl4-nBr» show a nearly linear increase in KA in other solvents [21]. on going from Et4NFeCl4 to Et4NFeBr4. This fact The excess line width Av due to the interaction means that specific ion-solvent dipole effects do not with the paramagnetic ion is given by the simplified play a significant role. form [22] of the McConnell equation [23] If the conclusion is accepted that ion-association

71-Av = T2-1 = PM/T2M in dichloromethane is influenced by special solvation in which T2 is the observed relaxation time, PM is effects only to a limited extent, a simple electro- the probability that a solvent molecule is in the static model must give good quantitative results for paramagnetic environment of the solvation shell, the association constant KA. Therefore we calculated and T2M is the relaxation time characteristic of this KA using the Fuoss-Eigen [6, 7] equation. This shell. If it is assumed that for a given solvent calculation uses for the halide ions the Pauling U molecule T2M is the same for all haloferrates, as ionic radii; for the ammonium ions the SPT com- demonstrated for the interaction of a series of pressibility radii [25], agreeing with the thermo- chromium(III) complexes with anions [24], the chemical radii [26], and for (CeHs^sFeCU the equality of the slope Av versus concentration of distance of closest approach in the crystal [27, 28]. In the same way the distance in (CeHs^sI [29] was ferrate means that interaction of the solvent with obtained, which, together with the ionic radius of I-, both complex ions, FeCLr and FeBr4_, is compa- gave the cationic radius. This value (419 pm) agrees rable. with the one based on the Van der Waals volume Discussion (425 pm) [30]. No usable information for the bromo- To start with, the simple halide anions will ferrate ion was found in the literature. be considered. The tetraethylammonium salts The results together with the radii used for the show a decrease of the extent of ion-association calculations are in Table III.

in the series Et4NCl > Et4NBr > Et4NI, as expected from an elementary electrostatic model. Table III. Calculated (Fuoss-Eigen equation) values of the association constant KA and comparison with The order is reversed in the tetraphenylarsonium experimental values. series that shows an increase according to Compound r+ + r- 10-3 x K (calc) Ka (exp) (C6H5)4ASC1 < (C6H5)4AsBr. A -1 3 pm mol dm KA (calc) The tetraethylammonium series shows that hydrogen-bridge formation between solvent and Et4NCl 342 + 181 58 1.6 Et4NBr 342 + 195 46 1.7 halide ion is not extensive in dichloromethane, as Et4NI 342 + 216 33 1.3 this phenomenon would reverse the stability order Et4NC104 342 + 236 25 2.2 [5]. On the other hand when only the Born solvation Bu4NCl 437 + 181 15 2.7 Et4NFeCl4 342 + 263 18 1.3 is considered substitution of Et4N+ by the larger (C6H5)4ASC1 419 + 181 19 0.12 (C6H5)4As+ ion would be expected to introduce a (C6H5)4AsBr 419 + 195 16 0.18 levelling of the stability order, but not the observed (C6H5)4AsFeCl4 419 + 263 8 0.38 reversed order, that can only be explained from specific hydrogen-bridge formation [5]. For the tetraethylammonium salts the agreement Consequently hydrogen bond formation plays a between calculated and experimental values is (seemingly small) role in the solvation of the halide satisfactory in view of the fact that the calculated ions in dichloromethane. KA values are probably accurate to within a factor 749 S. Bait et al. • Ion Association and Solvation in Dichloromethane

2-4 [31], even if only electrostatic forces are in aqueous solution, for which hydrogen bond operative. formation of the solvent with the jr-electron cloud In contrast to the tetraethylammonium salts the of the phenyl groups has been forwarded as a tetraphenylarsonium salts present a remarkable probable cause [32]. A similar effect may operate in difference between calculated and experimental dichloromethane. An additional explanation may be association constants. The reason cannot be the that the arsonium salts form less stable solvent larger ionic radius of the tetraphenylarsonium ion, separated ion-pairs, whereas the ammonium salts making the specific solvation effect of the halide have the possibility of direct cation-anion inter- ions comparably more important, as the tetrabutyl- action in contact ion-pairs. It must however be ammonium ion that has a comparable value of the emphasized that a difference in solvation of the cationic radius, in the association constant of the associated ion-pairs was not detected in the UV chloride, falls within the range of the tetraethyl- spectra nor in the NMR band widths that are ammonium salts. similar for both systems. The reason for this different behaviour may be The enhancement of solvation or of the apparent located in specific solvation of the tetraphenyl- radius of the tetraphenylarsonium ion has the result arsonium ion in dichloromethane, a phenomenon of overemphasizing the effect of the acceptor- that is expected to stabilize the free ion and sub- interaction of the solvent with the halide ions that sequently decrease KA. Exceptionally strong solva- reverses the stability order of halide association tion of the tetraphenylarsonium ion has its parallel compared to the alkylammonium salts.

[1] V. Gutmann, Chimia 31, 1 (1977) and references L17] Reference [15], Table F-41. therein. [18] P. Beronius, Acta Chem. Scand. A 30, 115 (1976). [2] V. Gutmann and R. Schmid, Coord. Chem. Rev. [19] P. Beronius and A. Brändström, Acta Chem. 12, 263 (1974). Scand. A 30, 687 (1976). [3] C. H. Langford and J. P. K. Tong, Acc. Chem. [20] R. M. Fuoss, L. Onsager, and J. F. Skinner, J. Res. 10, 258 (1977). Phys. Chem. 69, 2581 (1965). [4] U. Mayer and V. Gutmann, Adv. Inorg. Chem. [21] J. P. K. Tong, C. H. Langford, and Th. R. Radiochem. 17, 189 (1975). Stengle, Can. J. Chem. 52, 1721 (1974). [5] U. Mayer, Coord. Chem. Rev. 21, 159 (1976). [22] S. Behrendt, C. H. Langford, and L. S. Frankel, [6] R. M. Fuoss, J. Am. Chem. Soc. 80, 5059 (1958). J. Am. Chem. Soc. 91, 2236 (1969). [7] M. Eigen, Z. Phys. Chem. (Frankfurt a. Main) 1, [23] H. M. McConnell, J. Chem. Phys. 28, 430 (1958). 176 (1954). [24] Th. R. Stengle and C. H. Langford, J. Phys. [8] J. F. Coetzee and C. G. Karakatsanis, Inorg. Chem. 69, 3299 (1965). Chem. 15, 3112 (1976). [25] N. Desrosiers and J. E. Desnoyers, Can. J. Chem. [9] M. Cusumano, Inorg. Chim. Acta 25, 207 (1977). 54, 3800 (1976). [10] N. S. Gill, J. Chem. Soc. 1961, 3512. [26] T. C. Waddington, Adv. Inorg. Chem. Radio- [11] H. M. Dagget, E. J. Blair, and C. A. Kraus, chem. 1, 157 (1959). J. Am. Chem. Soc. 73, 799 (1951). [27] B. Zaslow and R. E. Rundle, J. Phys. Chem. 61, [12] J. E. Lind, J. J. Zwolenik, and R. M. Fuoss, 490 (1957). J. Am. Chem. Soc. 81, 1557 (1959). [28] F. A. Cotton and C. A. Murillo, Inorg. Chem. 14, [13] C. A. Clausen III and M. L. Good, Inorg. Chem. 9, 2467 (1975). 220 (1970). [29] R. C. L. Mooney, J. Am. Chem. Soc. 62, 2955 [14] R. M. Fuoss, J. Phys. Chem. 79, 525 (1975). (1940). [15] Handbook of Chemistry and Physics, 51st ed., [30] M. R. J. Dack, K. J. Bird, and A. J. Parker, Aust. The Chemical Rubber Co., Cleveland 1970, J. Chem. 28, 955 (1975). Table E-62. [31] R. G. Wilkins, Acc. Chem. Res. 3, 408 (1970). [16] G. Kortüm, Lehrbuch der Elektrochemie, Verlag [32] G. Kalfoglou and L. H. Bowen, J. Phys. Chem. Chemie, Weinheim 1966. 73, 2728 (1969).