von Barner and Bjerrum
2-
NbOCl5 has been identified in hydrochloric acid,7 in
various nonaqueous solvents,8,9 and in CsCl-NbOCl3 melts
with CsCl in excess.10 NbOCl4 ions formed in the latter
-
melts10 when less CsCl was added, and the presence of this
complex was also indicated in acetonitrile8,9 and dioxane9
solutions. Concerning Ta(V), oxochloro species of types sim-
ilar to the ones for Nb(V) formed in CsCl-TaOCl3 melts.11
Because the equilibrium constants for the chloro complex
formation of Nb(V) in NaCl-AlCl3 melts have not previous-
ly been determined, we decided to investigate this especially
at low temperature (i.e., 175 °C). However, the main focus
of the present work is to clarify which oxochloro complexes,
niobium or tantalum, will form in NaCl-AlCl3 melts.
Experimental Section
AlCl3 was prepared at 700 °C by reaction between molten
aluminum (99.999% from Atomergic Chemetals Corp.) and dry
hydrogen chloride gas. Anhydrous NaCl was made by treating an
analytical-grade material in the molten state with dry HCl gas.
Details about the preparation of these salts, which made up the
solvent, have been given in a previous paper.12
AlOCl was applied as the oxide source. It was synthesized by
the reaction between V2O5 and AlCl3. The latter chemical was added
in an amount approximately 10 times greater than that of V2O5.
After reaction, VOCl3 and the excess of AlCl3 were distilled off at
300 °C. The final AlOCl had a chloride content of 44.97%
(determined by Volhard titration) compared to a calculated content
of 45.0%. NbCl5 was made by the reaction of metallic niobium
(99.99% from Schuchardt) with chlorine gas (99.9% from Fluka)
at 600 °C. The product was purified by 5 times successive
sublimation under chlorine (0.33 atm). TaCl5 (99.9% from Fluka)
was similarly sublimed at 200 °C.
The chlorine-chloride concentration cells for the potentiometric
measurements have been described in detail elsewhere.13 The
apparatus and optical cells applied in the spectrophotometric
experiments were of the same type as that in our previous work.3
The Raman spectra were recorded with the same instrumentation
as was previously used14 and were obtained from melts contained
in Pyrex ampules sealed under argon.
Figure 1. UV-vis gas-phase spectrum of NbCl5 at 300 °C.
From our previous work3 on Ta(V) in NaCl-AlCl3, we
know from UV-vis spectra that oxide addition completely
-
converts TaCl6 to an oxochloro complex, and as we shall
see later, both Raman and UV-vis spectroscopic measure-
ments support this assumption also for niobium.
The measurements at 175 °C were not corrected for
vaporization of niobium or tantalum species into the gas
phase over the melt. In the previous work,3 this situation
with TaCl5 dissolved in NaCl-AlCl3 melts was considered.
For potentiometric measurements performed in a similar pCl-
region as that in the present work, it was concluded that no
correction for vaporization of TaCl5 was needed at 175 °C.
Concerning niobium, we have previously2 measured the gas-
phase spectrum of pure NbCl5 at 300 °C. Unfortunately,
during the calculation of the molar absorptivities, the
measured spectrum was multiplied with a wrong factor. Thus,
the intensities on the spectrum in ref 2, Figure 4A, became
to high by a factor of 1.822. The correct spectrum for gaseous
NbCl5 at 300 °C is given in Figure 1. As can be seen,
maximum absorbance occurs at 35 × 103 cm-1 (286 nm).
At this wavelength, only a slightly visible shoulder appeared
on the vapor-phase spectrum above a NaCl-AlCl3 melt with
NbCl5 added (average chloride coordination number of
approximately 5.5) at 175 °C. This implies for our experi-
ments that less than 1% of the added NbCl5 will be present
in the vapor phase at this temperature. Like in the case for
KCl-AlCl3 melts,2 the vaporization of NbCl5 from acidic
NaCl-AlCl3 melts at 300 °C was not negligible. For the
latter solvent, we estimate, from a spectrophotometric
measurement, the vapor-liquid distribution coefficient for
NbCl5 to be 0.03 at 300 °C.
The program used for testing of the different model equilibria
was similar to the one applied before.13
Definitions and Assumptions
In the calculations, it has been assumed that AlOCl reacts
quantitatively with the pentachlorides of niobium and
tantalum, i.e., according to
MCl5 + AlOCl f MOCl3 + AlCl3
where M ) Nb and Ta.
(1)
(7) Griffith, W. P.; Wickins, T. D. J. Chem. Soc. A 1967, 675.
(8) Furlani, C.; Zinato, E. Z. Anorg. Allg. Chem. 1967, 351, 210.
(9) Miranda, C.; da Silveira, M.; Vernois, J. J. Inorg. Nucl. Chem. 1970,
32, 839.
(10) Rosenkilde, C.; Voyiatzis, G. A.; Jensen, V. R.; Ystenes, M.; Østvold,
T. Inorg. Chem. 1995, 34, 4360.
(11) Rosenkilde, C.; Voyiatzis, G. A.; Østwold, T. Acta Chem. Scand. 1995,
49, 405.
(12) Hjuler, H. A.; Mahan, A.; von Barner, J. H.; Bjerrum, N. J. Inorg.
Chem. 1982, 21, 402.
(13) Zachariassen, K.; Berg, R. W.; Bjerrum, N. J.; von Barner, J. H. J.
Electrochem. Soc. 1987, 134, 1153.
(14) Bjerrum, N. J.; Berg, R. W.; Christensen, E.; Kerridge, D. H.; von
Barner, J. H. Anal. Chem. 1995, 34, 2129.
Our AlCl3 often contained a minor impurity of AlOCl.
This amount of AlOCl can be calculated from a potentio-
metric measurement on the equimolar NaCl-AlCl3 solvent
as described previously.13 Such a measurement was always
performed before the addition of the solutes, and the AlCl3
amounts referred to in this work are the corrected ones. The
calculations of the average coordination numbers and the
equilibrium constants are performed with assumptions for
the solvent and the electrochemical cell similar to those we
used before.3 Because we assume a total reaction between
9848 Inorganic Chemistry, Vol. 44, No. 26, 2005