- Email: [email protected]
Conjugate Brönsted–Lewis superacids in fluorosulfuric acid—Hammett acidity function and electrical conductivity studies of Ta(V), Nb(V) and Sb(V) Lewis acids
ELSEVIER
Journal 01’Fluorine C‘hemlstry XY ( 199X)
I 17-l 23
Conjugate Brijnsted-Lewis superacids in fluorosulfuric acid-Hammett acidity function and electrical conductivity studies of Ta( V),Nb( V) and Sb( V) Lewis acids Walter V. Cicha, Dingliang
Zhang,
Friedhelm
Aubke *
Abstract Conjugate Ta( SO,F),
BrBnsted-Lewis
range of Lewis
protonic
solutions
in HOS,F
HF. All conjugate 19%
Elsevier
acidities
acid. Electrical
Science
mol%
in HSO,F
acid are studied by two techniques.
of Lewis
acid are reported.
equal or higher to thae
conductivities
over the complete superacids
in fluoroaulfuric
range of O-3.3
acid concentration
conjugate
8
superacids
in the concentration
of the neat Lewk
concentration are found
2HA=H2A:,,,,+A,,,,,,
(1)
will be, in the simplest case, affected by the addition of Y in two ways: (i) the acidium ion H,A+ concentration will be increased, and (ii) the nucleophilicity of the base ion A will be reduced by complexation to Y to give the less basic, complex anion [ AY 1 : (3)
+IAYl,,,,,
author. Current
address: II-VI
Inc. Saxonburg, P.A.
lhO.56. USA.
’ Brlinstcd 100% H,SO,.
wperac~ds are delined ah protonic acid\ of higher acidity chat] See the work of Gilkpie
and Peel
[ 3.51,Gillr\pie
et al.
[ 61
and Gillespie 17 I.
2Levi\ qxracids I seethe work
AICIx
are defined as stronger electron pair acceptor\ than of Olah et al.
11.2I )_Note:
the term i\ rrstrictcd to
molecular Lcwi\ acids and does not include atomic Lewis acids H 0022-l 13Y/YX/$lY.O0 PI/S0022-I
is commonly
ionizing
function
studies on HSO,F-
SO,F),
shows in this
assumed to be the strongest
,,, /z = 3 or 4, M = Nh or Ta, and of their
to conductivities
and strongly
acidity
system HSO,F-Ta(
of SbF,
in HSO,F
and in anhydrous
over the entire concentration
range.
S.A. All rights reserved.
Among the very strong protonating and ionizing solvents. the highest proton acidities are encountered in conjugate Briinsted-Lewis acids [l-3]. These solvent systems are composed of a strong protonic or Briinsted acid-r superacid-as solvent,’ HA, and a molecular Lewis superacid.’ Y as solute. The self-ionization orproton-transferofHA according to:
* Corresponding
which
acids of the type MF,,( SOIF),
conducting
1. Introduction
~HA+Y,,,,,=HJA:,,,,
of HSO,FF3SO,-SbF,,
range are studied and compared to be highly
Hammett
The conjugate
or M”
8: IYYX Elsevier Science S.A. &II rtghb rewrved.
139(98)00097-Y
’
The Hammett acidity function, -H,, [S], is commonly used as a measure of acidity and procedures have been developed to apply -H,, determinations to conjugate superacids [ 4-6). A critical evaluation of acidity determinations of highly acidic conjugate systems with -H,, values of 2 20 has appeared [ 91. Of the simple protonic superacids of the type HA, fluorosulfuric acid HSO,F [ 7, IO. I I ] and anhydrous HF [ I2 1 have emerged as the strongest proton donors with identical -H,, values of 15. I [ d-6.13]. For Lewis superacids, a number of methods have been employed to arrive at a ranking according to their acceptor ability [2,14 1. It is. however. generally agreed that Sb( V)fluoride, SbF,, ranks as the strongest Lewis superacid of those compared [ 2, I4 1. It should hence follow that the conjugate superacids HF-SbF, and HSO.>-SbF, or ‘Magic Acid’ 1I ,2,15,16] should display the very highest acidities. and estimates of - H,,up to 30 (or about IS orders of magnitude more acidic than HF or HSO,F) have been made [ I ,2 ] ; however, accurate measurements beyond - H,,, values 01‘23 are experimentally very demanding and open to criticism 191. Of the protonic superacids, fluorosulfuric acid, because of its large liquid range ( - 89 to 167.7”C), its compatibility with glass, the availability of purification methods [ 171 and its outstanding solvent characteristics I7.10,1 1 ] offers clear
advantages over anhydrous HF both in its applications and in the use of physical measurements. In particular. electrical conductivity measurements [ 17,18], ‘“FNMR studies at variable temperatures [ 7, IO, I I ] and Hammett function studies [ 4,6.5 ] are more easily performed in this solvent than in HF. This is best illustrated by the -H,, values of HF. which is not directly measured, but rather obtained by interpolation methods [ I3 I. The interpolated value of - H,, = IS. I differs dramatically from the measured value of - I I [ I ,2 1. On the other hand, the ‘heteroleptic’ [ IS] conjugate superacid HSO?F-SbF, or ‘Magic Acid’ suffers from considerable complexity because of F vs. SO,F redistribution, apparent in the reported “‘F NMR spectra [ I&19]. There has hence been some effort spent on the development of ‘homoleptic’ 1IS] conjugate superacids in HSO,F which would involve binary fluorosulfates as Lewis acids. Four such conJugate superacid systems, HSO,F-Au( SO,F)., [ 20.3 I 1, HSO,FPt(SO,F), [22], and HSO,F-M(SO,F)s. M=Nb or Ta I23 1. have been developed and studied by various methods, among them conductometry [ 20-23 1, Marc recently, the potential system HSO?F-Sb(S03F), has come into focus with the isolation and characterization of Cs[Sb( SOF),] [ 21 la. In addition a number of well defined fluoride-Buorosulfates of Nb(V) and Ta(V) with the composition MF,,( SO,F),_ ,,, with II = 3 or 4 [ 241, which are found to be miscible with HSO,F at any concentration. In contra
1-log[BH+
]/[ B]
(3)
the indicator base ratio BH+ /B is determined spectrophotometrically as discussed previously 14-6 1. The noble-metal systems HSO,F-Au( SO,F), [ 20,2 I ] and HSO,F-Pt( SO,F )j [ 22 ] are unsuited for this purpose because both are intensely coloured. possibly on account of ligand-to-metal charge transfer transitions 120-22 1, Hence. the homoleptic system HSO,F-Ta( S03F)5 [ 23 1 is chosen. with the Lewis acid Ta(SO,F), prepared in situ by the oxidation of Ta by bis( fluorosulfuryl)peroxide. S?O,F, [ 2829 I. The conjugate system HS03F-Nb( SO,F), appears to be a weaker protonic acid as judged by a conductometric study in HSO,F up to concentrations of -0.04mol kgg [23]. This study is expected to increase our insights into novel and useful conjugate superacid systems of Sb( V). Ta( V) and Nb( V) Lewis acids. The conductivity measurements over the entire concentration range of the HA-Y systems move from the traditional dilute solution studies (up to 0.05
mol kg ’ ) to ranges that are used in practical applications superacids [ I-3 1,
of
2. Experimental 2. I. Chrrnii~uls The Lewis acids Ta(SO,F), [23], TaF,(SO,F), [24], TaF,( S03F) [ 24 I. NbF,( SO,F) [ 24 I. and NbF,( SOiF) [ 241 were synthesised according to published methods. The fluorides TaF,, NbF,, and SbF, were obtained from AtoChem North America ( formerly Ozark-Mahoning) The formertwo were of a quoted purity of 99% and were used without further purification. Sb( V) fluoride was purified by repeated distillation. The pure product, recognizable by its high viscosity, was obtained by a linal trap-to-trap distillation. Details on the purilication and manipulation of SbF, were published recently [ 30 1. Fluorosulfuric acid of technical grade ( Orange County Chemicals) is purilied by double distillation at atmospheric pressure as described previously [ 171. Bis( Huorosulfuryl)peroxide. S206FZ. was obtained by the catalytic fluorination ( AgF,) ofSO,. An improved procedure has recently been published [ 29 1. Ta and Nb metal powders were obtained from Matthey and Johnson. The indicator bases 2,4-dinitrofluorobenzene (DNBF) (Matheson, Coleman Bell) and 2.4,6-trinitrotoluene (TNT) (Eastman Organic Chemicals) were obtainedcommercially. Both were recrystallized from methanol and dried in vacua over P,O,,,. They and their monoprotonated cations, DNFBH+ and TNTH +. were employed in this study. 2.2. Mrusurrnwnts Hammett acidity function measurements followed published precedents [ 6.5 1. Our equipment to measure electrical conductivities [ 20,2 I 1, the conductivity cell design [ 17. I8 1, cell calibration procedures [ 17. I&20,2 I 1, and the constanttemperature bath [ 20,2 I ] operating at 25.O”C have all been described. Conductivity measurements on Lewis acidHSO,F solutions were started with measurements of the neat Lewis acids. HSO,F was gradually added from a graduated, pre-weighed burette inside a drybox (Vacuum Atmosphere, Hawthorne, CA) model DL-001-S-G filled with dry N,. Measurements were made about 30 min after mixing. 2.3. Ultrtr~~iolet/~~isiblroptic~~llc~rlls and rvpipnrnt To allow various manipulations of solutions without exposing them to the atmosphere, l-mm quartz Spectrosil precision optical cells were attached via a Pyrex bridge to a 25ml round-bottom flask fitted with a B I9 cone. In addition. the apparatus was fitted with a Kontes Teflon stem stopcock and a sidearm attached to a B IO cone. A matching adaptor consisting of a Kontes Teflon stem stopcock between a B IO cone and a B I9 socket was also provided. Sample solutions
were usually loaded into the solvent-containing flask in the drybox. mixed thoroughly and then transferred into the optical cell chamber by tilting the apparatus. Reproducibility was tested by repeating the above mixing procedure a few times between readings. On some occasions, it was adequate to use IO-mm Spectrosil precision optical cells, fitted with Teflon plugs and sealed with Teflon tape. Electronic spectra were recorded on a Hewlett Packard single-cell mode array spectrophotometer. Model X452A. incorporating HP Vectra computer hardware and a HP Think Jet printer. Software was available for internal sample referencing.
Tahlc I The Hammett acidities of Ta( SO,F) i in HSO ,F at 20°C
~TatSO,l~)il.mol%
-H,,
Indicator
0
IS.07
DNFB, TNT
0.055
IS.55
DNFB, TNT
0. I s4
16.07
TNT
0.31X
I h.73
TNT, DNFBH
0.9 13
I8.03
DNFBH
+
I.3
IX.36
DNFBH
+ . TNTH +
I .x0
1X.5X
TNTH +
2.1 I
IX.71
TNTH +
3.37
18.91
TNTH +
+
3. Results and discussion
High concentrations of Lewis acid are not suitable for study due to various experimental restrictions. among them problems encountered when trying to quickly dissolve more Ta metal powder in the S,O,F,/HSOjF mixtures. The reaction times needed are too long, leading to contamination from a slow leakage of air into the reactor or from trace amounts 01 grease dissolved in the media. It is found that only reactions of less than about 5 days duration give reproducible H,, values. The Hammett acidity values for Ta( SO,F), in HSO,F at select concentrations studied are given in Table 1. The plot ot -H,, vs. mol% Lewis acid is shown in Fig. I fol Ta( SO,F) 5, and for the two strong Lewis acids SbF, ( ‘Magic Acid’ ) and SbF?( SO,F) i [ 4-6 1. The principal feature of the plot is that beyond a concentration of about I mol%, Ta( SO,F), appears to be at least as strong as SbFI( SO,F) 1. The second striking l’eature worth noting is that compared to tither SbF, or SbF,( SO,F) + the rate of - Hii increase ij considerably less for Ta( SO,F), in the O-l mol% range, whereas beyond this concentration. it is equal or even greater than for the other two solutions. Both features of the Ta( SO?F), acidity can be explained. Its unexpectedly high value at concentrations beyond - I mol% ( - 0. I m) has already been predicted by the conductance results 123 1. which revealed the oligomeric nature ot this system in addition to a IO-fold increusc of its acidic dissociation constant with a similar increase in concentration ( from 0.01 to 0. I m ). By extrapolation of the conductivity results. the acidic dissociation constant. K,,. for -ra( SO+‘), should be of the order of2 X IO ’ m at I mol% and I X IO ’ m at 5 mol%. From the previously estimated concentration ofH2S01F in 100% HSO,F and its - H,, value, the idealired Eqs. (4) and I 5) can be used to estimate K,, for Ta( SO+‘) i in HSO,F at any given concentration. [ H$O,F ’ 1 is equal to 1Ta( SOiF).] in the latter equation. -H,,=log[H,SO.,F+ K.=
j+lX.79
IH,SO,F+ IlTa(SO,F), ITa(SO,F)jl
I
1
0
I
I
I
2
3
4
MOLE % LEWIS ACID Fig. I. Hammett acidity of TaCSO,F),.
61 in HSOrF
ShF, [W3]
and SbF?( SO,F,
i 14
at ambient temperature.
(4) I
mol kg
’
(5)
Fig 2. Dependcncc conccntralion
of the acidic dissociation
constant.
in HSO,F at ambient temprt-ature.
K,,. on Tut SO,F),
A plot of K,, vs. Ta( SO,F), concentration is shown in Fig. 2. K, increases steeply at greater than - 1 mol%, up to a value Of - 5 m at the maximum concentration, indicating virtually complete dissociation of the acid. Furthermore. this value is an order of magnitude greater than the Kc, value of - 0. I m predicted at this concentration from the conductivity studies. The K, vs. concentration curve is deceiving. however, since the rate of K, increase at the lower concentrations is hidden by the scale of the plot. For this reason, a plot of InK, vs. concentration is also shown in Fig. 2. which indicates that the greatest ioguritlzmic~ rate of’K, imwtrse is at concentrations of less than about I mol%. Following this ‘critical point’, the rate quickly decreases and InK, approaches a constant value. Extrapolation of the InK<, plot to infinite dilution leads to a very approximate K:, value of 8X IO ’ mol kg-‘. which is about an order of magnitude less than that estimated from the conductivity measurements. This suggests a large dependence of the acidic dissociation constant shown in Eq. (1) on concentration. which in turn implies that it is not a very accurate representation of the system’s acidity, as was already indicated from the conductivity measurements [ 23 I. The increase in magnitude of this system’s acidic dissociation constant with concentration can be partially attributed to formation of stronger polymeric acids at higher concentrations, as suggested for the SbF, systems 14-6 1. The slope difference between the three systems’ -H,, vs. concentration curves (Fig. I ) at < I mol% Lewis acidconcentration reflects the lower initial K, value of the Ta system. The formation of H, [Ta( SO,F), +, ] ( with .x> I i type acids (and/or oligomeric analogs) in solution at higher Ta( SO,F), concentrations is not inconceivable. since Cs2[ Ta( SO,F),] is isolable. This could result in 2 or even 3 mol of H,SO,F ’ forming per mole Ta( SO,F), upon acidic dissociation. leading to an approximate two- or three-fold increase in the acidity expected from simple acidic disxociation, and thus further contributing to the magnitude of the -H,, values at higher concentrations. Species of this type are not known to exist in either the SbF, or SbF,( SO ,F ) 1 systems. However. it must be stressed that the conductometric
SW,
titration results do not provide any evidence for such polybasic acids in the neutral range. Previous acidity studies with the HSO,F-SbF, [4,6,5, I8 1 systems have shown that acidity increases steadily with the number of moles of SO, added. but a maximum of only 3 mol SOi could be inserted into the Sh-F or As-F bonds. Hence, the presence of an unprecedented five fluorosulfate groups per metal center may also be partly responsible for the high acidity of Ta(SO?F),. Since there are strong indications from comparative electrical conductivity studies [ 23 1 that HSO,F-Ta( SO,F) 4 is a considerably stronger acid than HSO,F-Nb( SO,F),, acidity function studies of the latter system were not undertaken.
Selected data for element fluoride fluorosulfates and for binary fluorides. both formed by group 5 and group I5 elements, are collected in Table 2. Data on two oxyfluorosulfates SeO( SO,F)I [ 36 1and CIO,SO,F [ 37 ] are added for purpose 01‘reference. The latter is found to act as a base in HSO,F to produce the 1CIO,] (,,,,, i + cation [ 371. The remaining system of group 5 and IS are Lewis acids and can produce conjugate superacids in HSO,F. As far as appearance is concerned, all compounds listed in Table 2 are viscous liquids. frequently of limited volatility. Evidence from vibrational spectra has been collected to suggest the presence of fluoride- or Ruorosulfate bridged oligomers at room temperature [ 2335.37) A heterolytic cleavage of those oligomers to give conducting ionic fragments is suggested. In all instances. where measurements over a limited temperature range have been carried out, the specific conductivitics are found to rise with increasing temperature. which is consistent with the proposed ionic dissociation as cause for the observed electrical conductivities. There is one exception: Sb( V) Ruoridc exhibits extremely low specific conductivities even at slightly elevated tetnperaturex and
<4x10
”
VF,
2.43x
NhFi
2.33 x IO
’
T‘lF,
1..5xx0
’
r‘+,(
IO ’
i
SO,FI
1.10x
IO
NhF,( SO,F,
5.x0x
10 ’
A+:(
5.x5x
IO
SO,F I.
i
TaF;c SO;F, I
l.2XXIO
NhE.CSO,l;i_
1.7hX IO ’
SCO( SOJ* (‘I(),(
%);I.
1 /
I..50 x IO 7.51 x IO
-
’
121
homolytic rather than heterolytic dissociation at elevated temperatures [ 26,271 seems to occur. Two other exceptions are TaF, and NbF,. which at room temperature are crystalline solids with ordered tetrameric structures [ 2.51. On heating beyond their respective melting points, conducting liquids form. In spite of the higher temperatures (92.6 and 93.8”C-all other measurements are taken at 25”C), an approximate order of decreasing electrical conductivities emerges: VF, > NbF,( SO,F)? > TaF,- NbF,(SO,F) > TaF,( SO,F) > (SO~F)~>ASF.~(SO~F)~, NbF, > TaF, >> SbF,. For any composition. Nb( V) compounds appear to show higher conductivities than the corresponding Ta( V) systems. Furthermore. conductivities appear to increase with substitution of fluoride by Huorosulfate. The actual values, measured at 25°C. are in the range of IO 5 to IO ‘R-’ cm-’ which corresponds to the specific conductivity of HSO,F ( I .0X.5X 1OpJ 0 ’ cm ’ ) [ IO. I7 I. where on account of the proton jump mechanism the selfionization ions H$O,F~’ and SOJFmmcontribute substantially to the measured value 1IO]. Anion jump. or in this case SOJF jump, appears to be a probable contributor to the conductivities of the element fluoride fluorosulfates, as the SO,F-group is well capable of bridge formation. The relatively high conductivity of VF, may be due to its linear fluorine-bridged polymeric structure in both the solid and the liquid state [ 381, which differs from that of NbF, and TaF, 125 1.
The specific conductivities of solutions of the Nb( V ) and Ta( V) fluoride-fluorosulfates of the type MF,,( SO,F) 5 ,,. M = Nb or Ta, II = 3 or 4. are shown in Fig. 3. plotted against the Lewis acid concentration in mol%. All four materials are completely miscible with HSOJF and the resulting solutions are more highly conducting than are the individual components [ 7.10. I I ] ( see also Table 2). While the electrical conductivity increases with increasing SO+content. the relative order of the specific conductivities is now different. The Ta( V) Lewis acids appear to give higher specific conductivities than their Nb( V) counterparts. The reverse had been observed for the neat compounds. As a consequence. the highest conductivities are observed for TaF,( SO?F),. For this system, only conductometric titrations at very low Lewis acid concentration ( - 0.05 mol kg ’ ) are possible with KSO,F as standard base. Also included in Fig. 3 are the conductivities of SbF, in HSO,F. Even though liquid SbF, is poorly conducting (see Table 2). its solutions are more conducting than those 01 TaF,( SO,F)? near the maximum by a factor of 3 to 4. One has to be careful however, when including data from the HSO,F-SbF, system. also known as Magic Acid [ I 21. A recent re-investigation of the system by “‘F NMR [ 39.191 has shown that SbF, is extensively solvolysed in HSO,F to produce HF in addition to various oligomcric St+F-SO+ containing anions. Hydrogen fluoride, which can cause enhanced conductivity. is also found to react with glass and
400
-7
300
g + “x .T ‘3 2 x 8 $
200
100
,m
f
0
0.20
0.40
0.60
0.80
1.oo
Mole fraction of Lewis acid, X, Fig. 3. Spwitic conducGvities trf’ MF, ,(SO;), CM=Nh.Ta: r=l,2) and ShF, in HSO;F at ambient temperature (inwr(: hpccific conductivity ofSbF, in HF al ambient temperature IX.27 1 1.
at high SbF5 concentrations ( > 30 mol%), crystalline [ H,O] [ Sb,F, , I-its structure has recently been reported by LIS [ 40]-forms over a period of several weeks. It hence seems, that processes in HS03F-SbF, are complex and will produce various conducting ions, among them H,O ’ , and a number of monomeric and oligomeric anions ( IS diffcrcnt species have been observed [ 39. I9 ] ) In the oligomeric anions, SO,F-bridging takes precedent over Fbridging I 39. I9 ] in departure of previous views [ 3 I .42 I. While the complexity in the HSO+SbF, system can at least qualitatively explain the higher conductivity in HSO,F than found for solutions of MF,,( SOF), ,,. M = Nb or Ta, II = 3 or 4. it is interesting to note in Fig. 3 that for all five Lewis acids, a maximum in electrical conductivity is observed at Lewis acid concentrations of about 20 mol%. The exact positions of the conductivity maxima are listed in Table 3 together with the maximum conductivities found in each system. The resulting plots are similar in shape for all five systems and are best described as rather asymmetric bell-shaped curves, with the conductivities trailing off at the high Lewis acid concentrations. As a consequence, all five solutions are highly conducting media over the whole Lewis acid concentration range. A differently shaped conductivity vs. SbF, concentration plot is reported for HF-SbF,. I29.30] shown as an insert into Fig. 3. The curve is considerably sharper with the maximum at 3.5 mol% SbF,, where the electrical conductivity is
Tahlc 3 Conductivity
ShF, TLIF~(SO;F)~ TaF,( SOiF) NhF;CSO,F), NhF,C SOIF)
4. Summary and conclusions maxima in HSO,F dt 25°C
9.3-19.7 17.7 77 1-13 __.--. 6
34 I.1 0.17
IO.4
036
22.3
0.3 I
-300X IO3 i1- ’ cm- ’ and thus larger than in the HSO,FSbFs Magic Acid by a factor of IO. In the high Lewis acid concentration range. electrical conductivity drops off sharply [ 16 1 and measurements become difficult [ 27 1. Considering the difference in reported ion mobilities ofthe acidium ions (A” for HISOIF+ is I35 1 IO 1 vs. 350 for H,F’ 127 1). it is expected that at very low SbF, concentrations where the proton jump mechanism is active [ lQ2h.27 I. HFSbF, should be more conducting than the HSO,F-ShF,, system. The absence of side reaction in HF-SbF, contributes as well to higher conductivities, while in HSOJ-ShF, the formation of Hz0 followed by protonation to give H,O -c [ 39.19.40 I will reduce the overall acidity of‘ the system and decrease the H,SO,F’ concentration. Finally. oligomerization of the Lewis acids becomes a competing process to the generation of acidium ions by removal of the base ions. F or SO,F-, Oligomerization via Huorosulfate bridges is more likely in the HSO,F-SbF, system than the formation of Fbridges in the HF-SbF? system. These three reasons may be given to explain the steep increase in conductivity at low to intermediate SbF, concentrations for the conjugate acid HFSbFs and the rather slow gradual increase in all superacid systems in HSO,F. At intermediate Lewis acid concentration. conductivity via the proton jump mechanism becomes less likely as the bulk solvent (HF and HSOIZ:) decreases and hydrogen bonding between the solvent as well as its self-ionization becomes less probable than hydrogen bonding between solvent and solute anions. Hence. the electrical conductivity goes to a maximum and starts to decline. In addition, the increased viscosity impedes the migration of conducting ions in the conductivity experiment. At very high Lewis acid concentrations, the MF,,( SOIF),-,,, M = Nb or Ta, rr=3 or 4. systems are good conductors for reasons discussed in Section 3.2. I. Solvolysis of SbF, in HSO,F appears to generate in addition to HF similar SbF,,(SO,F), ~,, 139.191 species. This, of course. is not possible in the conjugate HF-SbF, system. The suggested break-up ofoligomers into conducting species for the mixed fluoride-fluorosulfate appears to be aided by the gradual addition of HSO,F. Thus. it seems to us that at least two different mechanisms may be active over the entire concentration range, as discussed above.
In solvent systems like the conjugate superacids discussed here, the corrosiveness of the systems and the ever present problem of contamination by water and other extrenuous materials make precise measurements rather difficult. The Lewis acids of the MF,,( SO,F), ~,, type used present additional problems. We have recently concluded [ 43 I that many of the viscous and oligomeric element fluoride tluorosulfates are phases which frequently are non-stoichiometric rather than well delined molecular compounds. The Nb and Ta compounds are all carefully purified by distillation [ 34 1 and their composition is contirmed by microanalysis and spectroscopy [ 241. Nevertheless, composition may vary within limits from one batch to the next. Hence, care must be taken when interpreting results of the conductivity measurements described here and only general trends can be recognized. For these reasons, we also have decided not to report tabulated conductivity data in this publication. The complete miscibility of the fluoride fluorosulfates with HSO,F allows measurements over the whole concentration range rather than only in the very low region ( up to molalities Of - 0.05 mol kg- ’ ). This is not possible for MF,, M = Nh or Ta. which show limited solubility in both HF and HSOIF 1 I-3 1, and consequently new applications are anticipated for the ternary Lewis acids discussed here. It is also important to obtain information on conjugate superacids at concentrations which are actually used in synthetic chemistry and in superacid catalysis [ I-31. It is clear now that in this region of equimolar amounts of HSO,F and Lewis acid, highly conducting media are encountered. which may favor the formation of reactive cations not just by protonation. The observed maxima in electrical conductivity in HSO,F suggest further work and applications at these concentrations of around 20 mol% Lewis acid ( see Table 3). Finally. we would like to draw some attention to conjugate superacids involving Ta( V) Lewis acids that are less likely to act as oxidizing agents than. for example. SbF, or AsF, [ I-31.
Acknowledgements Financial support by the Natural Sciences and Engineering Research Council of Canada is gratefully acknowledged. F.A. is thankful to the Alexander v. Humboldt Foundation for the award of a Research Prize. which provided the time to write up this long overdue account as a tribute to an old friend.
References 1I I G.A.Olah. G.K.S. Prahash, J. Sommer. Science 206 ( 1979) 13. 11I C;.A.Olah. G.K.S. Prahash. J. Sommer, Superacids. Wiley. New Yorh. 19X5. and ref\. therein.
131 T.A. O’Donnell. Superacids and Acidic Melts ah Inorganic Chemical Reaction Media. VCH P&l., New York. 1993, and rcf’s. therein.
(a)
141 R.J. Gillespir, T.E. Peel. Adv. Org. Chem. 9 t 1972) I. IS1 R.J. Gillespie, T.E. Peel, I. Am. Chem. Sot. 95 ( I973 i
W.V. Cicha. F. Aubke, J. Am. Chem. Sot. I I I
I I Y7 I )
65 (1961)
I71 R.J. Gillespie. Aoc. Chem. Rea. I ( 1968) 202. 181 L.P. Hammett, A.J. Deyrup. J. Am. Chem Sot. S-1 ( lY93) 272 I.
( lYX81
I Ed. ),
I71 and rel\.
Inorganic Sulphur Chemistry.
Elsevier. Amsterdam. 1968, p. 587.
1111 A.W. Jache. Adv. Inorg. Chem. Radiochem 16 ( 1974) 177. 1121 M.F.A. Dove. A.F. Clifford, in: J. Jander, H. Spandan, (’ c‘. Addison (Ed\.),
Chemistry
in Non-Aqueous
Ionizing
Solvent\.
Vol.
2. I
1131 R.J. Gillespie, J. Liang, J. Am. Chcm. Sot. I IO ( 19X8) 6037. P.L. FabrC. J. Dcvynh. B. Tremiflon, Chem. Rev. X2 f I YX2) SY I and IIJI ret’\. therein
( l9SS)
433.
1161 G.A. Olah. M.B. Comiaarow, CA. Chem Sot X7
( 196S)
Cupas, C.U. Pittmann. J Am.
2997.
I171 J. Barr. R.J. Gillespie. R.C. Thompson. fnorg. Chcm. 3 ( 1964) I I-l’l. R.A. Rothenhul-g. Inorp. 1181 R.C. Thompwn, J. Barr, R.J. Gillrqie. Chem. 3
IIYI
( 1965)
1641.
( 1966) I 170. ( 1957) 5 13.
D. Zhang, C. Wang, F. Mistry. B. Powell. F. Aubke. J. FluorincChem. 76
( 19Y6)
83.
C. Wang. S.C. Siu. G. Hwang, B. Bley, M. Bodenbinder, C. Bach. H. Willner. F. Aubke, Eur. J. Solid State Inorg. Chem. 33
( 1996)
917.
H.C. Clark, H.J. Emel&. J. Chem. Sot. ( 1957) 21. H.C. Clark. H.J. EmefCus. J. Chcm. SOL. (19%) 180. F. Fairhrother, W.C. Frith, A.A. Wooli’. J. Chem. Sot.
C1054)
103 I.
G.H. Grundy, A. Thompson. J. Chem. Sot.
H. fmoto, F. Auhke, J. Fluorine Chcm. IS ( IYXO) 59. H.A. Carter. PhD Thehi\. University of British Columbia. 1970. H.A. Carter. F. Aubke, Inorg. Nucl. Chem. Lctt. 5 ( I969)
990.
H.A. Carter. A.M.Qurcshi.F. Auhke.Chem. Commun. ( 1968) 1461. f:.A. Cotton, G. Wilkinson. Advanwd Inorganic Chemistry, 4th edn.. Wiley. New York, 19x0. H. Heuheb. Dipl. Thesis, Heinrich Heine Uniwrsitiit. Diisscldori’. Germany, 1994. 1). Zhang. S.J. Rettig. J. Trotter, I:. Auhkc. Inorg. Cllem. 3.S t fYY6)
couber. Canada. 1995.
G.A. Olah, A. Commeyras. J. Am. Chem. Sot. 91
Lee. F. Auhke. Inorg. Chem. IY
( 1965)
761.
6113.
Lee. F. Auhke. fnorg. Chem. IX
2124.
432X.
F.B. Dudley, G.H. Cady, J. Am. Chcm. Sot. 79
D.L. Zhang, PhD Thesi\. The Univerhiry of British Columbia. Vaw
1201 K.C. 1211 K.C.
(1984)
123.
FT.Fairhrother.
Viehweg Puhl., 197 I.
1151 A.A. Woolf. J. Chern. Sot.
( l98Y) ( 1992) 8 I,
3714.
R.J. Gillespie, K.C. MO.\\, J. Chern. Sot. A
therein.
1101 R.C. Thompson. in: G. Nicklea\
A.J. Edwards, J. Chem. Sot. (1964)
Chem.23
H.H. Hyman, L.A. Quatcrman, M. Kilpatric. J. Katz. J. Phys. Chem.
sox3.
191 R. Jest, I. Sommer, Rev. Chem. Intermed. Y
31.53: (h) K.C. Lee,F.Aubkc,Inorg.
D. Zhang, F. Aubke, J. Fluorine Chem 58
5173.
161 R.J. Gillespie. T.E. Peel. E.A. Robinson. J. Am Chcm. Sot. Y3
D. Zhang. S.J. Rettig. J. Trotter. F. Aubke, Inorg. Chcm. 34
( 1995)
( 1979) 389 t 1980) I I9
( lY6Y)
2929.
D. Brunet. A. Germain. A. Commeyras, Nowo. J. Chim. 2 27S.
( lY7R)
Journal 01’Fluorine C‘hemlstry XY ( 199X)
I 17-l 23
Conjugate Brijnsted-Lewis superacids in fluorosulfuric acid-Hammett acidity function and electrical conductivity studies of Ta( V),Nb( V) and Sb( V) Lewis acids Walter V. Cicha, Dingliang
Zhang,
Friedhelm
Aubke *
Abstract Conjugate Ta( SO,F),
BrBnsted-Lewis
range of Lewis
protonic
solutions
in HOS,F
HF. All conjugate 19%
Elsevier
acidities
acid. Electrical
Science
mol%
in HSO,F
acid are studied by two techniques.
of Lewis
acid are reported.
equal or higher to thae
conductivities
over the complete superacids
in fluoroaulfuric
range of O-3.3
acid concentration
conjugate
8
superacids
in the concentration
of the neat Lewk
concentration are found
2HA=H2A:,,,,+A,,,,,,
(1)
will be, in the simplest case, affected by the addition of Y in two ways: (i) the acidium ion H,A+ concentration will be increased, and (ii) the nucleophilicity of the base ion A will be reduced by complexation to Y to give the less basic, complex anion [ AY 1 : (3)
+IAYl,,,,,
author. Current
address: II-VI
Inc. Saxonburg, P.A.
lhO.56. USA.
’ Brlinstcd 100% H,SO,.
wperac~ds are delined ah protonic acid\ of higher acidity chat] See the work of Gilkpie
and Peel
[ 3.51,Gillr\pie
et al.
[ 61
and Gillespie 17 I.
2Levi\ qxracids I seethe work
AICIx
are defined as stronger electron pair acceptor\ than of Olah et al.
11.2I )_Note:
the term i\ rrstrictcd to
molecular Lcwi\ acids and does not include atomic Lewis acids H 0022-l 13Y/YX/$lY.O0 PI/S0022-I
is commonly
ionizing
function
studies on HSO,F-
SO,F),
shows in this
assumed to be the strongest
,,, /z = 3 or 4, M = Nh or Ta, and of their
to conductivities
and strongly
acidity
system HSO,F-Ta(
of SbF,
in HSO,F
and in anhydrous
over the entire concentration
range.
S.A. All rights reserved.
Among the very strong protonating and ionizing solvents. the highest proton acidities are encountered in conjugate Briinsted-Lewis acids [l-3]. These solvent systems are composed of a strong protonic or Briinsted acid-r superacid-as solvent,’ HA, and a molecular Lewis superacid.’ Y as solute. The self-ionization orproton-transferofHA according to:
* Corresponding
which
acids of the type MF,,( SOIF),
conducting
1. Introduction
~HA+Y,,,,,=HJA:,,,,
of HSO,FF3SO,-SbF,,
range are studied and compared to be highly
Hammett
The conjugate
or M”
8: IYYX Elsevier Science S.A. &II rtghb rewrved.
139(98)00097-Y
’
The Hammett acidity function, -H,, [S], is commonly used as a measure of acidity and procedures have been developed to apply -H,, determinations to conjugate superacids [ 4-6). A critical evaluation of acidity determinations of highly acidic conjugate systems with -H,, values of 2 20 has appeared [ 91. Of the simple protonic superacids of the type HA, fluorosulfuric acid HSO,F [ 7, IO. I I ] and anhydrous HF [ I2 1 have emerged as the strongest proton donors with identical -H,, values of 15. I [ d-6.13]. For Lewis superacids, a number of methods have been employed to arrive at a ranking according to their acceptor ability [2,14 1. It is. however. generally agreed that Sb( V)fluoride, SbF,, ranks as the strongest Lewis superacid of those compared [ 2, I4 1. It should hence follow that the conjugate superacids HF-SbF, and HSO.>-SbF, or ‘Magic Acid’ 1I ,2,15,16] should display the very highest acidities. and estimates of - H,,up to 30 (or about IS orders of magnitude more acidic than HF or HSO,F) have been made [ I ,2 ] ; however, accurate measurements beyond - H,,, values 01‘23 are experimentally very demanding and open to criticism 191. Of the protonic superacids, fluorosulfuric acid, because of its large liquid range ( - 89 to 167.7”C), its compatibility with glass, the availability of purification methods [ 171 and its outstanding solvent characteristics I7.10,1 1 ] offers clear
advantages over anhydrous HF both in its applications and in the use of physical measurements. In particular. electrical conductivity measurements [ 17,18], ‘“FNMR studies at variable temperatures [ 7, IO, I I ] and Hammett function studies [ 4,6.5 ] are more easily performed in this solvent than in HF. This is best illustrated by the -H,, values of HF. which is not directly measured, but rather obtained by interpolation methods [ I3 I. The interpolated value of - H,, = IS. I differs dramatically from the measured value of - I I [ I ,2 1. On the other hand, the ‘heteroleptic’ [ IS] conjugate superacid HSO?F-SbF, or ‘Magic Acid’ suffers from considerable complexity because of F vs. SO,F redistribution, apparent in the reported “‘F NMR spectra [ I&19]. There has hence been some effort spent on the development of ‘homoleptic’ 1IS] conjugate superacids in HSO,F which would involve binary fluorosulfates as Lewis acids. Four such conJugate superacid systems, HSO,F-Au( SO,F)., [ 20.3 I 1, HSO,FPt(SO,F), [22], and HSO,F-M(SO,F)s. M=Nb or Ta I23 1. have been developed and studied by various methods, among them conductometry [ 20-23 1, Marc recently, the potential system HSO?F-Sb(S03F), has come into focus with the isolation and characterization of Cs[Sb( SOF),] [ 21 la. In addition a number of well defined fluoride-Buorosulfates of Nb(V) and Ta(V) with the composition MF,,( SO,F),_ ,,, with II = 3 or 4 [ 241, which are found to be miscible with HSO,F at any concentration. In contra
1-log[BH+
]/[ B]
(3)
the indicator base ratio BH+ /B is determined spectrophotometrically as discussed previously 14-6 1. The noble-metal systems HSO,F-Au( SO,F), [ 20,2 I ] and HSO,F-Pt( SO,F )j [ 22 ] are unsuited for this purpose because both are intensely coloured. possibly on account of ligand-to-metal charge transfer transitions 120-22 1, Hence. the homoleptic system HSO,F-Ta( S03F)5 [ 23 1 is chosen. with the Lewis acid Ta(SO,F), prepared in situ by the oxidation of Ta by bis( fluorosulfuryl)peroxide. S?O,F, [ 2829 I. The conjugate system HS03F-Nb( SO,F), appears to be a weaker protonic acid as judged by a conductometric study in HSO,F up to concentrations of -0.04mol kgg [23]. This study is expected to increase our insights into novel and useful conjugate superacid systems of Sb( V). Ta( V) and Nb( V) Lewis acids. The conductivity measurements over the entire concentration range of the HA-Y systems move from the traditional dilute solution studies (up to 0.05
mol kg ’ ) to ranges that are used in practical applications superacids [ I-3 1,
of
2. Experimental 2. I. Chrrnii~uls The Lewis acids Ta(SO,F), [23], TaF,(SO,F), [24], TaF,( S03F) [ 24 I. NbF,( SO,F) [ 24 I. and NbF,( SOiF) [ 241 were synthesised according to published methods. The fluorides TaF,, NbF,, and SbF, were obtained from AtoChem North America ( formerly Ozark-Mahoning) The formertwo were of a quoted purity of 99% and were used without further purification. Sb( V) fluoride was purified by repeated distillation. The pure product, recognizable by its high viscosity, was obtained by a linal trap-to-trap distillation. Details on the purilication and manipulation of SbF, were published recently [ 30 1. Fluorosulfuric acid of technical grade ( Orange County Chemicals) is purilied by double distillation at atmospheric pressure as described previously [ 171. Bis( Huorosulfuryl)peroxide. S206FZ. was obtained by the catalytic fluorination ( AgF,) ofSO,. An improved procedure has recently been published [ 29 1. Ta and Nb metal powders were obtained from Matthey and Johnson. The indicator bases 2,4-dinitrofluorobenzene (DNBF) (Matheson, Coleman Bell) and 2.4,6-trinitrotoluene (TNT) (Eastman Organic Chemicals) were obtainedcommercially. Both were recrystallized from methanol and dried in vacua over P,O,,,. They and their monoprotonated cations, DNFBH+ and TNTH +. were employed in this study. 2.2. Mrusurrnwnts Hammett acidity function measurements followed published precedents [ 6.5 1. Our equipment to measure electrical conductivities [ 20,2 I 1, the conductivity cell design [ 17. I8 1, cell calibration procedures [ 17. I&20,2 I 1, and the constanttemperature bath [ 20,2 I ] operating at 25.O”C have all been described. Conductivity measurements on Lewis acidHSO,F solutions were started with measurements of the neat Lewis acids. HSO,F was gradually added from a graduated, pre-weighed burette inside a drybox (Vacuum Atmosphere, Hawthorne, CA) model DL-001-S-G filled with dry N,. Measurements were made about 30 min after mixing. 2.3. Ultrtr~~iolet/~~isiblroptic~~llc~rlls and rvpipnrnt To allow various manipulations of solutions without exposing them to the atmosphere, l-mm quartz Spectrosil precision optical cells were attached via a Pyrex bridge to a 25ml round-bottom flask fitted with a B I9 cone. In addition. the apparatus was fitted with a Kontes Teflon stem stopcock and a sidearm attached to a B IO cone. A matching adaptor consisting of a Kontes Teflon stem stopcock between a B IO cone and a B I9 socket was also provided. Sample solutions
were usually loaded into the solvent-containing flask in the drybox. mixed thoroughly and then transferred into the optical cell chamber by tilting the apparatus. Reproducibility was tested by repeating the above mixing procedure a few times between readings. On some occasions, it was adequate to use IO-mm Spectrosil precision optical cells, fitted with Teflon plugs and sealed with Teflon tape. Electronic spectra were recorded on a Hewlett Packard single-cell mode array spectrophotometer. Model X452A. incorporating HP Vectra computer hardware and a HP Think Jet printer. Software was available for internal sample referencing.
Tahlc I The Hammett acidities of Ta( SO,F) i in HSO ,F at 20°C
~TatSO,l~)il.mol%
-H,,
Indicator
0
IS.07
DNFB, TNT
0.055
IS.55
DNFB, TNT
0. I s4
16.07
TNT
0.31X
I h.73
TNT, DNFBH
0.9 13
I8.03
DNFBH
+
I.3
IX.36
DNFBH
+ . TNTH +
I .x0
1X.5X
TNTH +
2.1 I
IX.71
TNTH +
3.37
18.91
TNTH +
+
3. Results and discussion
High concentrations of Lewis acid are not suitable for study due to various experimental restrictions. among them problems encountered when trying to quickly dissolve more Ta metal powder in the S,O,F,/HSOjF mixtures. The reaction times needed are too long, leading to contamination from a slow leakage of air into the reactor or from trace amounts 01 grease dissolved in the media. It is found that only reactions of less than about 5 days duration give reproducible H,, values. The Hammett acidity values for Ta( SO,F), in HSO,F at select concentrations studied are given in Table 1. The plot ot -H,, vs. mol% Lewis acid is shown in Fig. I fol Ta( SO,F) 5, and for the two strong Lewis acids SbF, ( ‘Magic Acid’ ) and SbF?( SO,F) i [ 4-6 1. The principal feature of the plot is that beyond a concentration of about I mol%, Ta( SO,F), appears to be at least as strong as SbFI( SO,F) 1. The second striking l’eature worth noting is that compared to tither SbF, or SbF,( SO,F) + the rate of - Hii increase ij considerably less for Ta( SO,F), in the O-l mol% range, whereas beyond this concentration. it is equal or even greater than for the other two solutions. Both features of the Ta( SO?F), acidity can be explained. Its unexpectedly high value at concentrations beyond - I mol% ( - 0. I m) has already been predicted by the conductance results 123 1. which revealed the oligomeric nature ot this system in addition to a IO-fold increusc of its acidic dissociation constant with a similar increase in concentration ( from 0.01 to 0. I m ). By extrapolation of the conductivity results. the acidic dissociation constant. K,,. for -ra( SO+‘), should be of the order of2 X IO ’ m at I mol% and I X IO ’ m at 5 mol%. From the previously estimated concentration ofH2S01F in 100% HSO,F and its - H,, value, the idealired Eqs. (4) and I 5) can be used to estimate K,, for Ta( SO+‘) i in HSO,F at any given concentration. [ H$O,F ’ 1 is equal to 1Ta( SOiF).] in the latter equation. -H,,=log[H,SO.,F+ K.=
j+lX.79
IH,SO,F+ IlTa(SO,F), ITa(SO,F)jl
I
1
0
I
I
I
2
3
4
MOLE % LEWIS ACID Fig. I. Hammett acidity of TaCSO,F),.
61 in HSOrF
ShF, [W3]
and SbF?( SO,F,
i 14
at ambient temperature.
(4) I
mol kg
’
(5)
Fig 2. Dependcncc conccntralion
of the acidic dissociation
constant.
in HSO,F at ambient temprt-ature.
K,,. on Tut SO,F),
A plot of K,, vs. Ta( SO,F), concentration is shown in Fig. 2. K, increases steeply at greater than - 1 mol%, up to a value Of - 5 m at the maximum concentration, indicating virtually complete dissociation of the acid. Furthermore. this value is an order of magnitude greater than the Kc, value of - 0. I m predicted at this concentration from the conductivity studies. The K, vs. concentration curve is deceiving. however, since the rate of K, increase at the lower concentrations is hidden by the scale of the plot. For this reason, a plot of InK, vs. concentration is also shown in Fig. 2. which indicates that the greatest ioguritlzmic~ rate of’K, imwtrse is at concentrations of less than about I mol%. Following this ‘critical point’, the rate quickly decreases and InK, approaches a constant value. Extrapolation of the InK<, plot to infinite dilution leads to a very approximate K:, value of 8X IO ’ mol kg-‘. which is about an order of magnitude less than that estimated from the conductivity measurements. This suggests a large dependence of the acidic dissociation constant shown in Eq. (1) on concentration. which in turn implies that it is not a very accurate representation of the system’s acidity, as was already indicated from the conductivity measurements [ 23 I. The increase in magnitude of this system’s acidic dissociation constant with concentration can be partially attributed to formation of stronger polymeric acids at higher concentrations, as suggested for the SbF, systems 14-6 1. The slope difference between the three systems’ -H,, vs. concentration curves (Fig. I ) at < I mol% Lewis acidconcentration reflects the lower initial K, value of the Ta system. The formation of H, [Ta( SO,F), +, ] ( with .x> I i type acids (and/or oligomeric analogs) in solution at higher Ta( SO,F), concentrations is not inconceivable. since Cs2[ Ta( SO,F),] is isolable. This could result in 2 or even 3 mol of H,SO,F ’ forming per mole Ta( SO,F), upon acidic dissociation. leading to an approximate two- or three-fold increase in the acidity expected from simple acidic disxociation, and thus further contributing to the magnitude of the -H,, values at higher concentrations. Species of this type are not known to exist in either the SbF, or SbF,( SO ,F ) 1 systems. However. it must be stressed that the conductometric
SW,
titration results do not provide any evidence for such polybasic acids in the neutral range. Previous acidity studies with the HSO,F-SbF, [4,6,5, I8 1 systems have shown that acidity increases steadily with the number of moles of SO, added. but a maximum of only 3 mol SOi could be inserted into the Sh-F or As-F bonds. Hence, the presence of an unprecedented five fluorosulfate groups per metal center may also be partly responsible for the high acidity of Ta(SO?F),. Since there are strong indications from comparative electrical conductivity studies [ 23 1 that HSO,F-Ta( SO,F) 4 is a considerably stronger acid than HSO,F-Nb( SO,F),, acidity function studies of the latter system were not undertaken.
Selected data for element fluoride fluorosulfates and for binary fluorides. both formed by group 5 and group I5 elements, are collected in Table 2. Data on two oxyfluorosulfates SeO( SO,F)I [ 36 1and CIO,SO,F [ 37 ] are added for purpose 01‘reference. The latter is found to act as a base in HSO,F to produce the 1CIO,] (,,,,, i + cation [ 371. The remaining system of group 5 and IS are Lewis acids and can produce conjugate superacids in HSO,F. As far as appearance is concerned, all compounds listed in Table 2 are viscous liquids. frequently of limited volatility. Evidence from vibrational spectra has been collected to suggest the presence of fluoride- or Ruorosulfate bridged oligomers at room temperature [ 2335.37) A heterolytic cleavage of those oligomers to give conducting ionic fragments is suggested. In all instances. where measurements over a limited temperature range have been carried out, the specific conductivitics are found to rise with increasing temperature. which is consistent with the proposed ionic dissociation as cause for the observed electrical conductivities. There is one exception: Sb( V) Ruoridc exhibits extremely low specific conductivities even at slightly elevated tetnperaturex and
<4x10
”
VF,
2.43x
NhFi
2.33 x IO
’
T‘lF,
1..5xx0
’
r‘+,(
IO ’
i
SO,FI
1.10x
IO
NhF,( SO,F,
5.x0x
10 ’
A+:(
5.x5x
IO
SO,F I.
i
TaF;c SO;F, I
l.2XXIO
NhE.CSO,l;i_
1.7hX IO ’
SCO( SOJ* (‘I(),(
%);I.
1 /
I..50 x IO 7.51 x IO
-
’
121
homolytic rather than heterolytic dissociation at elevated temperatures [ 26,271 seems to occur. Two other exceptions are TaF, and NbF,. which at room temperature are crystalline solids with ordered tetrameric structures [ 2.51. On heating beyond their respective melting points, conducting liquids form. In spite of the higher temperatures (92.6 and 93.8”C-all other measurements are taken at 25”C), an approximate order of decreasing electrical conductivities emerges: VF, > NbF,( SO,F)? > TaF,- NbF,(SO,F) > TaF,( SO,F) > (SO~F)~>ASF.~(SO~F)~, NbF, > TaF, >> SbF,. For any composition. Nb( V) compounds appear to show higher conductivities than the corresponding Ta( V) systems. Furthermore. conductivities appear to increase with substitution of fluoride by Huorosulfate. The actual values, measured at 25°C. are in the range of IO 5 to IO ‘R-’ cm-’ which corresponds to the specific conductivity of HSO,F ( I .0X.5X 1OpJ 0 ’ cm ’ ) [ IO. I7 I. where on account of the proton jump mechanism the selfionization ions H$O,F~’ and SOJFmmcontribute substantially to the measured value 1IO]. Anion jump. or in this case SOJF jump, appears to be a probable contributor to the conductivities of the element fluoride fluorosulfates, as the SO,F-group is well capable of bridge formation. The relatively high conductivity of VF, may be due to its linear fluorine-bridged polymeric structure in both the solid and the liquid state [ 381, which differs from that of NbF, and TaF, 125 1.
The specific conductivities of solutions of the Nb( V ) and Ta( V) fluoride-fluorosulfates of the type MF,,( SO,F) 5 ,,. M = Nb or Ta, II = 3 or 4. are shown in Fig. 3. plotted against the Lewis acid concentration in mol%. All four materials are completely miscible with HSOJF and the resulting solutions are more highly conducting than are the individual components [ 7.10. I I ] ( see also Table 2). While the electrical conductivity increases with increasing SO+content. the relative order of the specific conductivities is now different. The Ta( V) Lewis acids appear to give higher specific conductivities than their Nb( V) counterparts. The reverse had been observed for the neat compounds. As a consequence. the highest conductivities are observed for TaF,( SO?F),. For this system, only conductometric titrations at very low Lewis acid concentration ( - 0.05 mol kg ’ ) are possible with KSO,F as standard base. Also included in Fig. 3 are the conductivities of SbF, in HSO,F. Even though liquid SbF, is poorly conducting (see Table 2). its solutions are more conducting than those 01 TaF,( SO,F)? near the maximum by a factor of 3 to 4. One has to be careful however, when including data from the HSO,F-SbF, system. also known as Magic Acid [ I 21. A recent re-investigation of the system by “‘F NMR [ 39.191 has shown that SbF, is extensively solvolysed in HSO,F to produce HF in addition to various oligomcric St+F-SO+ containing anions. Hydrogen fluoride, which can cause enhanced conductivity. is also found to react with glass and
400
-7
300
g + “x .T ‘3 2 x 8 $
200
100
,m
f
0
0.20
0.40
0.60
0.80
1.oo
Mole fraction of Lewis acid, X, Fig. 3. Spwitic conducGvities trf’ MF, ,(SO;), CM=Nh.Ta: r=l,2) and ShF, in HSO;F at ambient temperature (inwr(: hpccific conductivity ofSbF, in HF al ambient temperature IX.27 1 1.
at high SbF5 concentrations ( > 30 mol%), crystalline [ H,O] [ Sb,F, , I-its structure has recently been reported by LIS [ 40]-forms over a period of several weeks. It hence seems, that processes in HS03F-SbF, are complex and will produce various conducting ions, among them H,O ’ , and a number of monomeric and oligomeric anions ( IS diffcrcnt species have been observed [ 39. I9 ] ) In the oligomeric anions, SO,F-bridging takes precedent over Fbridging I 39. I9 ] in departure of previous views [ 3 I .42 I. While the complexity in the HSO+SbF, system can at least qualitatively explain the higher conductivity in HSO,F than found for solutions of MF,,( SOF), ,,. M = Nb or Ta, II = 3 or 4. it is interesting to note in Fig. 3 that for all five Lewis acids, a maximum in electrical conductivity is observed at Lewis acid concentrations of about 20 mol%. The exact positions of the conductivity maxima are listed in Table 3 together with the maximum conductivities found in each system. The resulting plots are similar in shape for all five systems and are best described as rather asymmetric bell-shaped curves, with the conductivities trailing off at the high Lewis acid concentrations. As a consequence, all five solutions are highly conducting media over the whole Lewis acid concentration range. A differently shaped conductivity vs. SbF, concentration plot is reported for HF-SbF,. I29.30] shown as an insert into Fig. 3. The curve is considerably sharper with the maximum at 3.5 mol% SbF,, where the electrical conductivity is
Tahlc 3 Conductivity
ShF, TLIF~(SO;F)~ TaF,( SOiF) NhF;CSO,F), NhF,C SOIF)
4. Summary and conclusions maxima in HSO,F dt 25°C
9.3-19.7 17.7 77 1-13 __.--. 6
34 I.1 0.17
IO.4
036
22.3
0.3 I
-300X IO3 i1- ’ cm- ’ and thus larger than in the HSO,FSbFs Magic Acid by a factor of IO. In the high Lewis acid concentration range. electrical conductivity drops off sharply [ 16 1 and measurements become difficult [ 27 1. Considering the difference in reported ion mobilities ofthe acidium ions (A” for HISOIF+ is I35 1 IO 1 vs. 350 for H,F’ 127 1). it is expected that at very low SbF, concentrations where the proton jump mechanism is active [ lQ2h.27 I. HFSbF, should be more conducting than the HSO,F-ShF,, system. The absence of side reaction in HF-SbF, contributes as well to higher conductivities, while in HSOJ-ShF, the formation of Hz0 followed by protonation to give H,O -c [ 39.19.40 I will reduce the overall acidity of‘ the system and decrease the H,SO,F’ concentration. Finally. oligomerization of the Lewis acids becomes a competing process to the generation of acidium ions by removal of the base ions. F or SO,F-, Oligomerization via Huorosulfate bridges is more likely in the HSO,F-SbF, system than the formation of Fbridges in the HF-SbF? system. These three reasons may be given to explain the steep increase in conductivity at low to intermediate SbF, concentrations for the conjugate acid HFSbFs and the rather slow gradual increase in all superacid systems in HSO,F. At intermediate Lewis acid concentration. conductivity via the proton jump mechanism becomes less likely as the bulk solvent (HF and HSOIZ:) decreases and hydrogen bonding between the solvent as well as its self-ionization becomes less probable than hydrogen bonding between solvent and solute anions. Hence. the electrical conductivity goes to a maximum and starts to decline. In addition, the increased viscosity impedes the migration of conducting ions in the conductivity experiment. At very high Lewis acid concentrations, the MF,,( SOIF),-,,, M = Nb or Ta, rr=3 or 4. systems are good conductors for reasons discussed in Section 3.2. I. Solvolysis of SbF, in HSO,F appears to generate in addition to HF similar SbF,,(SO,F), ~,, 139.191 species. This, of course. is not possible in the conjugate HF-SbF, system. The suggested break-up ofoligomers into conducting species for the mixed fluoride-fluorosulfate appears to be aided by the gradual addition of HSO,F. Thus. it seems to us that at least two different mechanisms may be active over the entire concentration range, as discussed above.
In solvent systems like the conjugate superacids discussed here, the corrosiveness of the systems and the ever present problem of contamination by water and other extrenuous materials make precise measurements rather difficult. The Lewis acids of the MF,,( SO,F), ~,, type used present additional problems. We have recently concluded [ 43 I that many of the viscous and oligomeric element fluoride tluorosulfates are phases which frequently are non-stoichiometric rather than well delined molecular compounds. The Nb and Ta compounds are all carefully purified by distillation [ 34 1 and their composition is contirmed by microanalysis and spectroscopy [ 241. Nevertheless, composition may vary within limits from one batch to the next. Hence, care must be taken when interpreting results of the conductivity measurements described here and only general trends can be recognized. For these reasons, we also have decided not to report tabulated conductivity data in this publication. The complete miscibility of the fluoride fluorosulfates with HSO,F allows measurements over the whole concentration range rather than only in the very low region ( up to molalities Of - 0.05 mol kg- ’ ). This is not possible for MF,, M = Nh or Ta. which show limited solubility in both HF and HSOIF 1 I-3 1, and consequently new applications are anticipated for the ternary Lewis acids discussed here. It is also important to obtain information on conjugate superacids at concentrations which are actually used in synthetic chemistry and in superacid catalysis [ I-31. It is clear now that in this region of equimolar amounts of HSO,F and Lewis acid, highly conducting media are encountered. which may favor the formation of reactive cations not just by protonation. The observed maxima in electrical conductivity in HSO,F suggest further work and applications at these concentrations of around 20 mol% Lewis acid ( see Table 3). Finally. we would like to draw some attention to conjugate superacids involving Ta( V) Lewis acids that are less likely to act as oxidizing agents than. for example. SbF, or AsF, [ I-31.
Acknowledgements Financial support by the Natural Sciences and Engineering Research Council of Canada is gratefully acknowledged. F.A. is thankful to the Alexander v. Humboldt Foundation for the award of a Research Prize. which provided the time to write up this long overdue account as a tribute to an old friend.
References 1I I G.A.Olah. G.K.S. Prahash, J. Sommer. Science 206 ( 1979) 13. 11I C;.A.Olah. G.K.S. Prahash. J. Sommer, Superacids. Wiley. New Yorh. 19X5. and ref\. therein.
131 T.A. O’Donnell. Superacids and Acidic Melts ah Inorganic Chemical Reaction Media. VCH P&l., New York. 1993, and rcf’s. therein.
(a)
141 R.J. Gillespir, T.E. Peel. Adv. Org. Chem. 9 t 1972) I. IS1 R.J. Gillespie, T.E. Peel, I. Am. Chem. Sot. 95 ( I973 i
W.V. Cicha. F. Aubke, J. Am. Chem. Sot. I I I
I I Y7 I )
65 (1961)
I71 R.J. Gillespie. Aoc. Chem. Rea. I ( 1968) 202. 181 L.P. Hammett, A.J. Deyrup. J. Am. Chem Sot. S-1 ( lY93) 272 I.
( lYX81
I Ed. ),
I71 and rel\.
Inorganic Sulphur Chemistry.
Elsevier. Amsterdam. 1968, p. 587.
1111 A.W. Jache. Adv. Inorg. Chem. Radiochem 16 ( 1974) 177. 1121 M.F.A. Dove. A.F. Clifford, in: J. Jander, H. Spandan, (’ c‘. Addison (Ed\.),
Chemistry
in Non-Aqueous
Ionizing
Solvent\.
Vol.
2. I
1131 R.J. Gillespie, J. Liang, J. Am. Chcm. Sot. I IO ( 19X8) 6037. P.L. FabrC. J. Dcvynh. B. Tremiflon, Chem. Rev. X2 f I YX2) SY I and IIJI ret’\. therein
( l9SS)
433.
1161 G.A. Olah. M.B. Comiaarow, CA. Chem Sot X7
( 196S)
Cupas, C.U. Pittmann. J Am.
2997.
I171 J. Barr. R.J. Gillespie. R.C. Thompson. fnorg. Chcm. 3 ( 1964) I I-l’l. R.A. Rothenhul-g. Inorp. 1181 R.C. Thompwn, J. Barr, R.J. Gillrqie. Chem. 3
IIYI
( 1965)
1641.
( 1966) I 170. ( 1957) 5 13.
D. Zhang, C. Wang, F. Mistry. B. Powell. F. Aubke. J. FluorincChem. 76
( 19Y6)
83.
C. Wang. S.C. Siu. G. Hwang, B. Bley, M. Bodenbinder, C. Bach. H. Willner. F. Aubke, Eur. J. Solid State Inorg. Chem. 33
( 1996)
917.
H.C. Clark, H.J. Emel&. J. Chem. Sot. ( 1957) 21. H.C. Clark. H.J. EmefCus. J. Chcm. SOL. (19%) 180. F. Fairhrother, W.C. Frith, A.A. Wooli’. J. Chem. Sot.
C1054)
103 I.
G.H. Grundy, A. Thompson. J. Chem. Sot.
H. fmoto, F. Auhke, J. Fluorine Chcm. IS ( IYXO) 59. H.A. Carter. PhD Thehi\. University of British Columbia. 1970. H.A. Carter. F. Aubke, Inorg. Nucl. Chem. Lctt. 5 ( I969)
990.
H.A. Carter. A.M.Qurcshi.F. Auhke.Chem. Commun. ( 1968) 1461. f:.A. Cotton, G. Wilkinson. Advanwd Inorganic Chemistry, 4th edn.. Wiley. New York, 19x0. H. Heuheb. Dipl. Thesis, Heinrich Heine Uniwrsitiit. Diisscldori’. Germany, 1994. 1). Zhang. S.J. Rettig. J. Trotter, I:. Auhkc. Inorg. Cllem. 3.S t fYY6)
couber. Canada. 1995.
G.A. Olah, A. Commeyras. J. Am. Chem. Sot. 91
Lee. F. Auhke. Inorg. Chem. IY
( 1965)
761.
6113.
Lee. F. Auhke. fnorg. Chem. IX
2124.
432X.
F.B. Dudley, G.H. Cady, J. Am. Chcm. Sot. 79
D.L. Zhang, PhD Thesi\. The Univerhiry of British Columbia. Vaw
1201 K.C. 1211 K.C.
(1984)
123.
FT.Fairhrother.
Viehweg Puhl., 197 I.
1151 A.A. Woolf. J. Chern. Sot.
( l98Y) ( 1992) 8 I,
3714.
R.J. Gillespie, K.C. MO.\\, J. Chern. Sot. A
therein.
1101 R.C. Thompson. in: G. Nicklea\
A.J. Edwards, J. Chem. Sot. (1964)
Chem.23
H.H. Hyman, L.A. Quatcrman, M. Kilpatric. J. Katz. J. Phys. Chem.
sox3.
191 R. Jest, I. Sommer, Rev. Chem. Intermed. Y
31.53: (h) K.C. Lee,F.Aubkc,Inorg.
D. Zhang, F. Aubke, J. Fluorine Chem 58
5173.
161 R.J. Gillespie. T.E. Peel. E.A. Robinson. J. Am Chcm. Sot. Y3
D. Zhang. S.J. Rettig. J. Trotter. F. Aubke, Inorg. Chcm. 34
( 1995)
( 1979) 389 t 1980) I I9
( lY6Y)
2929.
D. Brunet. A. Germain. A. Commeyras, Nowo. J. Chim. 2 27S.
( lY7R)