A (19)F nuclear magnetic resonance study of the conjugate Bronsted-Lewis superacid HSO3F-SbF5. Part 1

1869

19

A F nuclear magnetic resonance study of

the conjugate Brönsted–Lewis superacid

HSO₃F–SbF₅. Part 1

Dingliang Zhang, Markus Heubes, Gerhard Hägele, and Friedhelm Aubke

Abstract: The Brönsted–Lewis superacid HSO₃F–SbF₅or “magic acid” is re-investigated by modern ¹⁹F NMR methods

over a wide concentration range. The system is found to be considerably more complex than had been assumed

previously. A total of 13 different anions are identified of which only five have previously been identified in magic

acid. With increasing SbF₅contents the concentration of monomeric anions like [SbF₆]^–, [SbF₅(SO₃F)]^–, cis- and trans-

[SbF₄(SO₃F)₂]^–, and mer-[SbF₃(SO₃F)₃]^–gradually decreases. Except for [Sb₂F₁₁]^–, which is present in very small

concentrations only, the formation of oligomers involves exclusively m-fluorosulfato bridges. In addition to donor

(SO₃F)^–and acceptor (SbF₅) complex formation to give [SbF₅(SO₃F)]^–and possibly ligand redistribution, the solvolysis

of SbF₅or SbF₄(SO₃F) in HSO₃F appears to be the principal formation reaction for polyfluorosulfatofluoroantimonate(V)

anions. In glass (NMR tubes) the solvolysis product HF is converted to the oxonium ion [H₃O]⁺, which has previously

1

been identified by H NMR and structurally characterized as [H₃O][Sb₂F₁₁] by us.

Key words: magic acid, conjugate superacid, fluorosulfuric acid, ¹⁹F NMR spectra.

Résumé : On a réétudié l’ « acide magique » (superacide de Brönsted–Lewis), HSO₃F–SbF₃en faisant appel à la

RMN du ¹⁹F et en opérant sur une grande plage de concentrations. On a trouvé que le système est beaucoup plus

complexe que ce qui avait été envisagé auparavant. On a identifié un total de 13 anions différents, dont seulement cinq

avaient été identifiés antérieurement dans l’acide magique. Avec une augmentation de la concentration de SbF₅, il y a

diminution de la concentration des anions monomères, du type [SbF₆]^–, [SbF₅(SO₃F)]^–, cis- et trans-[SbF₄(SO₃F)₂]^–et

mer-[SbF₃(SO₃F)₃]^–. À l’exception du [Sb₂F₁₁]^–dont la concentration est très faible, la formation d’oligomères

implique exclusivement des ponts m-fluorosulfato. En plus de la formation de complexes provenant du donneur (SO₃F)^–

et de l’accepteur (SbF₅) conduisant au [SbF₅(SO₃F)]^–ou de la possibilité d’une redistribution de ligands, il semble que

la solvolyse du SbF₅ou du SbF₄(SO₃F) dans le HSO₃F soit la réaction principale de formation des anions

polyfluorosulfatofluoroantimonantes(V). En présence de verre (tubes de RMN), le HF produit par solvolyse se

1

transforme en ion oxonium [H₃O]⁺qui a été identifié antérieurement par RMN du H et pour lequel on a caractérisé la

structure, [H₃O][Sb₂F₁₁].

Mots clés : acide magique, superacide conjugué, acide fluorosulfurique, spectre RMN du ¹⁹F.

[Traduit par la Rédaction] Zhang et al. 1886

Introduction

number of substrates (1–4). The proton acidity of magic acid

is only exceeded by that of the related conjugate superacid

HF–SbF₅(1, 2). This is altogether not surprising because the

constituent Brönsted acids fluorosulfuric acid (5) and anhy-

drous hydrogen fluoride (6) are the strongest simple protonic

acids with identical Hammett (7) acidity function values of

-H₀= 15.1 (1, 2, 8), while the Lewis superacid SbF₅is

ranked as the strongest molecular Lewis acid (9).

The conjugate Brönsted–Lewis superacid HSO₃F–SbF₅or

“magic acid” (1) has found in the past extensive use as reac-

tion medium for the generation and stabilization of a wide

range of highly electrophilic, long-lived cations. Well docu-

mented examples extend from inorganic main group cations

(1–3) to a large number of carbocations (1, 4). The reasons

for the enormous versatility of magic acid are its outstanding

solvent characteristics, in particular its ionizing and pro-

tonating abilities, which allow the dissolution of a large

For magic acid Hammett acidity function studies (7) are

reported (10–13) for concentrations up to 80 mol% SbF₅. A

comparison between HF–SbF₅and HSO₃F–SbF₅(14, 15)

Received March 8, 1999.

This paper is dedicated to Professor Ted Schaefer in recognition of his outstanding contributions to Canadian chemistry.

D. Zhang and F. Aubke.¹Department of Chemistry, The University of British Columbia, 2036 Main Mall, Vancouver, BC V6T

1Z1, Canada.

M. Heubes and G. Hägele.¹Institut für Anorganische Chemie und Strukturchemie I, Heinich-Heine-Universität Düsseldorf,

Universitätsstr.1, D-40225 Düsseldorf, Germany.

¹Authors to whom correspondence may be addressed. FA: Telephone: (604) 822-3817. Fax: (604) 822-2847.

e-mail: aubke@chem.ubc.ca; GH: Telephone: ++49-211-8112288. Fax: ++49-211-8113085. e-mail: haegele@uni-duesseldorf.de

Can. J. Chem. 77: 1869–1886 (1999)

1870

Can. J. Chem. Vol. 77, 1999

suggests that at 0.5 mol % SbF₅, the former is more acidic

by a factor of approximately 10⁴(15). It is not immediately

obvious why this should be, since the constituent Brönsted

acids have identical -H₀values (1, 2, 9) and SbF₅is common

to both superacid systems.

Repeated NMR measurements spanning a period of several

months showed the stable, equilibrated state of NMR sam-

ples discussed in this paper.

¹⁹F NMR spectroscopy

To probe the composition of HSO₃F–SbF₅, several ¹⁹F

NMR studies (16–19) have been reported between 1965 and

1978. Because of the limited resolution and sensitivity of

NMR spectra at that time, accurate assignments of the sig-

nals have been difficult and some disagreement is found

among the reported studies (16–19). There is however gen-

eral agreement that very complex equilibria are encountered,

which will be discussed subsequently. The situation is much

simpler in the HF–SbF₅system, where in addition to [SbF₆]^–

oligomeric cis-fluorine bridged anions like [Sb₂F₁₁]^–,

[Sb₃F₁₆]^–, etc. are present at high SbF₅concentrations and

detected by ¹⁹F NMR (20). Since these early studies the res-

olution, sensitivity, and evaluation of NMR spectra have all

significantly improved by the introduction of pulse spectros-

copy and Fourier transformation in one- and two-

dimensional techniques (21). We have applied state of the art

NMR methods to reinvestigate the HSO₃F–SbF₅system with

two objectives:

Both 1D and 2D ¹⁹F NMR spectra were acquired at low

temperatures (see text) using a Bruker AM 200 NMR spec-

trometer equipped with a Bruker variable temperature unit

B-VT 1000. For each 1D ¹⁹F NMR spectrum, 64K memory

size was used for data storage (TD = 64K) and 1024 scans

were acquired (NS = 1024) for each Free Induction Decay

(FID). No zero-filling was applied to the FIDs before Fou-

rier transformation. For ¹⁹F–¹⁹F COSY NMR, a standard

program, COSY.AUR, was used with the following pulse se-

quence:

[1]

Relaxation time t₁– 90° pulse

– evolution time t – 45° pulse – FID

TD = 2 K and NS = 32 were used and 512 experiments were

measured for every 2D spectrum. For 2D J-resolved NMR, a

standard program, JRES.AUR, was used and the following

pulse sequence was implemented:

(i) to obtain more precise structural information on the an-

ionic species present and

(ii) to suggest plausible pathways of formation of these

anions from the interaction of HSO₃F with SbF₅.

[2]

Relaxation time t₁– 90° pulse – evolution time t

– 180° pulse – evolution time t – FID

We present here a ¹⁹F NMR study of the HSO₃F–SF₅

superacid system over a wide concentration range with SbF₅

molar fractions x between 0.001 and 0.795. The techniques

employed include high resolution ¹⁹F 1D- and 2D NMR

methods of the type ¹⁹F–¹⁹F COSY and ¹⁹F J resolved spec-

tra at 181 MHz.

TD = 4 K along the F₁axis and TD = 128 along the F₂axis

were used, and 64 experiments were measured for each 2D

spectrum. The FIDs for 1D NMR were processed using the

Bruker Win-NMR software on an IBM-compatible PC and

the FIDs for 2D NMR were processed with the program

UXNMR running on a UNIX workstation (ASPECT).

Experimental

Results and discussion

Preparation of NMR samples

An already published (22) observation, made at the outset

of this study, will be briefly summarized. All samples pre-

pared as described in the experimental section are found

to exhibit in the ¹⁹F NMR spectrum a resonance line at

The purity of freshly distilled fluorosulfuric acid was

checked by conductometry and by NMR. Purified antimony

pentafluoride was transferred in vacuo into a 50 mL vial fit-

ted with a Teflon stopcock. The amount of the antimony

pentafluoride transferred was determined by difference

weighing. Inside a dry box, freshly distilled fluorosulfuric

acid was pipetted into the reactor. The amount of fluoro-

sulfuric acid was also determined by difference weighing.

The mixture was then stirred magnetically for ca. 15 min.

About 0.4 mL of the solution was pipetted into an NMR

tube. A sealed capillary tube containing a solution of 5%

(w/w) CFCl₃in CD₃COCD₃was inserted into the NMR tube

in order to provide chemical shift references and a field lock.

The NMR tube was attached to a B10–B10 adapter fitted

with a Kontes Teflon stem stopcock and then taken out of

the dry box. The NMR tube was evacuated after it was

cooled to –95 ~ –100°C in an ethanol – dry ice slurry bath.

The sample was degassed by three freeze–pump–thaw cy-

cles. The NMR tube was then flame-sealed under vacuum.

1

~–162 ppm, which is attributed to dissolved SiF₄. The H

NMR spectrum, in addition to the main resonance due to ex-

change between HSO₃F and the acidium ion H₂SO₃F⁺, a

second single resonance between ~9.5 and ~8.0 ppm, which

is assigned to the oxonium ion H₃O⁺. The intensity of both

resonances increases with increasing SbF₅contents. From

NMR tubes with a SbF₅content of 20 mol% or higher a

white crystalline solid is obtained, which is identified by vi-

brational spectroscopy as [H₃O][Sb₂F₁₁] and structurally

characterized by single crystal X-ray diffraction. It is con-

cluded, that magic acid contains variable amounts of HF,

which will react with glass to give H₃O⁺. Evidence for this

view is obtained in a recently completed ¹⁹F NMR study²of

magic acid in tubes made from fluoroplastics, where the

SiF₄resonance is absent in the ¹⁹F NMR spectrum and the

H₃O⁺resonance is of exceedingly low intensity. The com-

monly held view (1), that HSO₃F–SbF₅, just like pure

HSO₃F and SbF₅will not attack glass must be seriously

questioned.

1

Samples were checked immediately after preparation by H-

NMR in Vancouver and then shipped to Düsseldorf for ¹⁹F-

NMR studies. All samples were kept at room temperature.

²J.H. Kühn-Velten, M. Bodenbinder, G. Hägele, and F. Aubke. To be submitted.

Zhang et al.

1871

Fig. 1. ¹⁹F NMR spectrum of the HSO₃F–SbF₅system with x_SbF= 0.342 at 213 K.

5

singlets due to trans-[SbF₄(SO₃F)₂]^–units are also observed.

Therefore, they argue that [SbF₆]^–and [Sb₂F₁₁]- are formed

more likely through ligand redistribution such as

1D and 2D ¹⁹F–¹⁹F COSY spectroscopy

The first ¹⁹F NMR study of the HSO₃F–SbF₅system is re-

ported by Thompson et al. (16). Due to the limited sensitivity

and resolution, only two partially overlapped AX₄spin sys-

tems are observed from the spectrum of a 1.7 mol kg^–1solu-

[9]

2[SbF₅(SO₃F)]^–W [SbF₄(SO₃F)₂]^–+ [SbF₆]^–

tion of SbF₅in HSO₃F (x_SbF= 0.15). They are assigned to

5

monomeric [SbF₅(SO₃F)]^–and the SO₃F-bridged dimer

[Sb₂F₁₀(SO₃F)]^–, and the following mechanism for their for-

mation is proposed:

rather than by decomposition of HSO₃F as suggested in

reactions [6]–[8]. More recently, Brunel et al. (19) have

carried out ¹⁹F NMR studies of the HSO₃F–SbF₅and

HSO₃CF₃–SbF₅systems, using sulfuryl difluoride (SO₂F₂)

as an internal reference. They conclude that the cis-F-bridged

H[Sb_nF_5n(SO₃F)] (n = 0–5) species are the main species in

high SbF₅concentration ranges and that the SO₃F-bridged

dimer [Sb₂F₁₀(m-SO₃F)]^–is present in low concentration only

[3]

[4]

SbF₅+ HSO₃F W H[SbF₅(SO₃F)]

H[SbF₅(SO₃F)] + HSO₃F W H₂SO₃F⁺

+ [SbF₅(SO₃F)]^–

in systems with low x_SbF. This conclusion is contradictory

5

to previous studies (16, 18), which indicate the 1:1 and 2:1

complexes H[SbF₅(SO₃F)] and H[Sb₂F₁₀(m-SO₃F)] are the

principal species.

[5]

2H[SbF₅(SO₃F)] W H₂SO₃F⁺+ [Sb₂F₁₀(SO₃F)]^–

Commeyras and Olah (17) extend the ¹⁹F NMR investigation

of the HSO₃F–SbF₅system to a wider SbF₅/HSO₃F ratio

range of 0~1.4 (x_SbF= 0~0.58) and support the earlier

assignments (16). In addition, they observe signals due to the

monomer [SbF₆]^–(–120 ppm) and F-bridged dimer [Sb₂F₁₁]-

(–115 ppm) as well as two unassigned signals at –122 and

–124 ppm. To explain the formation of the perfluoroanti-

monate(V) species, the existence of polysulfuric acids is

proposed according to

In summary, previous ¹⁹F NMR studies of the HSO₃F–

SbF₅system differ in many aspects: the SbF₅concentrations

chosen, the use of diluents, the spectra observed, and the

spectral interpretations. The approach in the present study is

5

to start with the HSO₃F–SbF₅system with x_SbFas low as

5

possible, provided that reasonable spectra can be acquired.

The monomeric antimony(V) species are most likely present

at low x_SbF. The mole fraction of SbF₅is gradually in-

5

creased to observe oligomerization and the changes in rela-

tive concentrations of species. To avoid any possible

interference, no diluent was added into the HSO₃F–SbF₅

system in this study.

[6]

[7]

[8]

2SbF₅+ 2HSO₃F W H₂SO₃F⁺+ [Sb₂F₁₁]^–+ SO₃

SO₃+ HSO₃F W HS₂O₆F

SO₃+ HS₂O₆F W HS₃O₉F

General features of 1D ¹⁹F NMR spectra of the

HSO₃F–SbF₅system

This proposal is subsequently disputed by Dean and

Gillespie (18). To decrease the viscosity of the system in or-

der to improve the resolution of the spectrum, these authors

use sulfuryl chlorofluoride (SO₂ClF) as a diluent for the

HSO₃F–SbF₅system with the SbF₅:HSO₃F ratio at 0.92:1.00

(x_SbF= 0.48). Besides the signals observed previously, two

A₂X₂⁵spin systems due to cis-[SbF₄(SO₃F)₂]^–units and two

As an example, the ¹⁹F NMR spectrum of an HSO₃F–SbF₅

system with x_SbF= 0.35 is shown in Fig. 1.

The spectrum ⁵may be divided into two regions: the F(S)

region (d CFCl₃= 40~45 ppm), where the signals due to

fluorines of SO₃F groups are observed, and the F(Sb).region

(d CFCl₃= –80~–150 ppm) where signals due to fluorines

1872

Can. J. Chem. Vol. 77, 1999

bonded to Sb(V) are observed. In addition, there is a sharp

fluorosulfato-antimonate(V) species can be extracted from

the signals in the F(S) region, and no assignment of reso-

nances in this region will be made.

singlet at ca. –162 ppm, which is attributed to SiF_4(solv)

.

Apparently, there is a trace amount of HF that reacts with

SiO₂, the main component of glass, to produce SiF₄accord-

ing to

In contrast, the F(Sb) region (d CFCl₃= –80~–150 ppm)

(see Fig. 3) provides a great deal of structural information.

The observed splitting of the signals results from the cou-

pling between non-equivalent fluorines bonded to the same

antimony atom. The complex spectral feature of this region

indicates complicated equilibria of various chemical species

in the system. The overlap of signals is sometimes so severe

that an analysis based solely on traditional 1D ¹⁹F NMR

spectra remains ambiguous. Even with the assistance of 2D

NMR methods, the spectral analysis is occasionally incon-

clusive.

[10] 4HF + SiO₂® 2H₂O + SiF₄

The other product, H₂O, is a strong base in the system and is

protonated to form the oxonium ion H₃O⁺, which is observ-

able using ¹H NMR spectroscopy. Its salt with weakly

nucleophilic [Sb₂F₁₁]^–as the counter anion has been isolated

as discussed (22). The other possible route for the produc-

tion of H₂O, dehydration of HSO₃F according to

Significant signal broadening begins to be observed in the

[11] 2HSO₃F W H₂O + S₂O₅F₂

spectra of the HSO₃F–SbF₅system with x_SbF³ 0.500. The

5

is unlikely because the signal due to S₂O₅F₂at 47~48.5 ppm

(23, 24) is not observed in the ¹⁹F NMR spectra of all sam-

ples studied by us here. Evidence for reaction [10] also co-

mes from the visual observation of etching of NMR tubes

and capillary tubes after several weeks.

The mechanism for the generation of HF in magic acid is

not obvious. The observation of progressively increasing

spectra in the F(Sb) region of the systems with x_SbF= 0.649

5

and 0.795 are dominated by broad peaks. Such signal broad-

ening precludes any detailed analysis in terms of individual

species. Therefore, the assignment of signals discussed in

the following sections will focus on the F(Sb) region in the

spectra of solutions with low SbF₅content (0.000 999 £ x_SbF

5

³ 0.342). For the convenience, the following notation system

for the assignable signals in the F(Sb) region is used. The

signals in the F(Sb) region are labeled according to their

multiplicity: S for singlet, D for doublet, T for triplet, Q for

quartet, q for quintet, s for sextet, etc. For signals of the

same multiplicity, a number is attached in the order of their

appearance from low field to high field.

amounts of SiF₄with increasing x_SbFin the system suggests

5

that the trace amounts of HF may be introduced as an impu-

rity in SbF₅. However, as discussed in the following sec-

tions, partial solvolysis of SbF₅in HSO₃F according to

[12] SbF₅+ HSO₃F W SbF₄(SO₃F) + HF

The increasing viscosity of the system, as the SbF₅con-

tent increases, may not be the only factor responsible for the

broadness of signals, because a number of relatively sharp

lines are also observed as well in the F(Sb) region. The

oligomeric structures of (SbF₅)_mand possibly H[Sb_mF_5m(SO₃F)]

or H[Sb_mF_5m+1], as well as ligand exchange, may contribute

to the broadness of the signals. The sharp lines are probably

from some relatively small complex anions. Among these

sharp lines, the most intense signals labeled T3 and T6 can

be assigned to cis-[SbF₄(SO₃F)₂]^–, S8 to [SbF₆]^–, and Q1 to

[Sb₂F₁₁]^–.

[13] SbF₄(SO₃F) + HSO₃F W SbF₃(SO₃F)₂+ HF

is a much more likely source for HF.

This proposal is based on the assumption that SbF₅or

SbF₄(SO₃F) are more readily solvolyzed by HSO₃F than

coordinatively saturated anions like [SbF₆]^–and [SbF₅(SO₃F)]^–.

Since SiF₄is also observed by ¹⁹F NMR spectroscopy for

the neat solvent HSO₃F, it is also possible that some HF may

be formed by the self dissociation of HSO₃F (K < 3 × 10^–7):

[14] HSO₃F W SO₃+ HF

As reported previously (16–19), the proton exchange be-

tween the Sb(V) species and HSO₃F appears to be rapid and

only averaged signals of both complex acids and their conju-

gate anions are observed in the ¹⁹F NMR spectra. For conve-

nience, the anionic formulation is used to represent both

forms in the discussion. The assignment of segments

(monomeric Sb(V) moieties), SbF_n(SO₃F)_5–nm-(SO₃F), and

m-(SO₃F)SbF_n(SO₃F)_4–nm-(SO₃F), in the oligomeric species

will be discussed below. To distinguish the segments from

monomeric species (both complex acids and conjugate an-

ions), the segments are formulated as neutral fragments.

(see ref. 25) or may be present as an impurity in HSO₃F. It is

not at all uncommon to observe SiF₄as well as BF₃in the

gas phase of freshly distilled HSO₃F or introduced with

HSO₃F as an impurity.

Figure 2 illustrates the ¹⁹F NMR spectra of pure HSO₃F

and HSO₃F–SbF₅systems with various x_SbFvalues in the

5

F(S) region (d CFCl₃= 40~45 ppm).

In this region, all signals are singlets. For the solvent and

solutions with low SbF₅content, intense signals due to

HSO₃F dominate the region. A main peak due to F-³²SO₃H

and a weak satellite peak due to F-³⁴SO₃H can be observed

(26, 27). For example, in the 1D ¹⁹F NMR spectrum of

fluorosulfuric acid (see Fig. 2), the very intense peak at

41.98 ppm and a weak satellite peak at 41.94 ppm are ob-

served. The isotope shift is 0.044 ppm, in agreement with

the previous studies (26, 27). For less intense F(S) signals,

the satellite peak (F-³⁴SO₃) is not observed. Long-range cou-

plings of the types ¹⁹F-S-O-Sb-¹⁹F and ¹⁹F-S-O-Sb-O-S-¹⁹F

are not observed for any samples in this study. Conse-

quently, little information about the structures of the fluoro-

Assignment of monomeric species

In the ¹⁹F NMR spectra (F(Sb) region) of the HSO₃F–SbF₅

system with x_SbF= 0.000 999 (see Fig. 3a), there are a num-

5

ber of multiplets in addition to two singlets. Apparently, sev-

eral chemical species are present in the system. The possible

–

monomers of the type [SbF_n(SO₃F)_6–n

Fig. 4.

]

are illustrated in

In these species, antimony(V) is assumed to be octahe-

drally coordinated by fluoro and fluorosulfato ligands.

Zhang et al.

1873

Fig. 2. Low temperature ¹⁹F NMR spectra from the F(S) region of the HSO₃F–SbF₅system.

From the coupling patterns and relative intensities of the

signals, they can be readily assigned to some of the

monomeric species expected to be present. With the assis-

tance of 2D ¹⁹F–¹⁹F COSY NMR, the analysis becomes

more straightforward and definitive. In the 2D ¹⁹F–¹⁹F

COSY spectrum (F(Sb) region) of the HSO₃F–SbF₅system

1874

Can. J. Chem. Vol. 77, 1999

Fig. 3. ¹⁹F NMR spectra from the F(Sb) region of the HSO₃F–SbF₅system.

Zhang et al.

1875

Fig. 3 (continued).

1876

Can. J. Chem. Vol. 77, 1999

Fig. 3. (concluded).

with x_SbF= 0.000 999 (see Fig. 5), the cross-peaks D1T5

[SbF₆]^–, trans-[SbF₄(SO₃F)₂]^–, fac-[SbF₃(SO₃F)₃]^–, trans-

and cis-[SbF₂(SO₃F)₄]^–, and [SbF(SO₃F)₅]^–(see Fig. 4). In

previous ¹⁹F NMR studies of HSO₃F–SbF₅systems with

high SbF₅:HSO₃F ratios, Commeyras and Olah (17) assign

the signal at CFCl₃= –120 ppm to [SbF₆]^–. In the present

study, however, no singlet at –120 ppm is observed. Dean

and Gillespie (18) observe two singlets at –102~–104 ppm

in the spectrum of a sulfuryl chlorofluoride solution of

5

and T3T6 clearly indicate the correlation between doublet

D1 and triplet T5, and between triplet T3 and triplet T6.

Thus, we can unequivocally assign D1 and T5 (AX₂spin

system) to mer-[SbF₃(SO₃F)₃]^–and T3 and T6 (A₂X₂spin

system) to cis-[SbF₄(SO₃F)₂]^–.

The singlets in the F(Sb) region may be assigned to spe-

cies containing chemically equivalent fluorines, namely

Zhang et al.

1877

Fig. 4. Monomeric species of the type [SbF_n(SO₃F)_6–n]^–(n = 0–6)

Fig. 5. ¹⁹F–¹⁹F COSY NMR spectrum of the HSO₃F–SbF₅

and corresponding spin systems in the F(Sb) region.

system with x_SbF= 0.000 999.

5

All signals assigned to monomeric species are observable

in the HSO₃F–SbF₅system at least up to x_SbF= 0.489 except

5

D1 and T5, which is only barely observable for x_SbF

=

5

0.0500 and is absent in the systems with higher x_SbF. This

5

observation suggests that the monomeric mer-[SbF₃(SO₃F)₃]^–,

which has the highest SO₃F content among all species ob-

served in this study, exists only at very low SbF₅concentra-

tion. It has not been observed in the previous NMR studies

of HSO₃F–SbF₅systems with high HSO₃F:SbF₅ratios.

However, it has been reported by Thompson et al. (16) for

the HSO₃F–SbF₄(SO₃F) and HSO₃F–SbF₅–SO₃systems. It

was reported that, based on cryoscopic measurements, the

freshly distilled HSO₃F contains ~0.0006 mol kg^–1of SO₃

(25), which is comparable with the concentration of SbF₅in

the x_SbF= 0.000 999 system. Therefore, the insertion reac-

5

tion of SO₃with SbF₅in HSO₃F (16) may have contributed

to the formation of species with high SO₃F/F ratios.

The chemical shifts and coupling constants of signals as-

signed to monomeric species in the HSO₃F–SbF₅system

SbF₅:HSO₃F (0.92:1.00) recorded at –100°C and assign both

of them to trans-[SbF₄(SO₃F)₂] units. Brunel et al. (19) as-

sign a high-field singlet at d CFCl₃= –115.8 ppm to [SbF₆]^–

and the low-field singlet at d CFCl₃= –91.9 ppm to trans-

[SbF₄(SO₃F)₂]^–. Following these previous assignments, the

high-field sharp line S8 at –125.6 ppm is assigned to [SbF₆]^–

and the other singlet S7 at –105.0 ppm is assigned to trans-

[SbF₄(SO₃F)₂]^–.

with x_SbF= 0.009 84 are summarized in Table 1.

5

In spite of some variations in chemical shifts for the sig-

nals with increasing x_SbF, the following general correlation

between the ¹⁹F chemica⁵l shifts and their chemical environ-

ment is observed:

As x_SbFincreases from 0.000 999 to 0.009 82 (see Fig. 3b),

5

D2 and q4 become considerably more intense in the spec-

trum. The relative intensities and distinct splitting patterns in

the 1D ¹⁹F NMR spectrum and the cross-peak D2q4 in the

¹⁹F–¹⁹F COSY spectrum (see Fig. 6) clearly indicate that the

two signals represent an AX₄spin system.

This AX₄spin system can be assigned to monomeric

[SbF₅(SO₃F)]^–, which is in agreement with the previous

studies (16–19).

(i) For the fluorine trans to a fluorosulfato group, the

signal shifts to low field as cis fluorines are sequentially

replaced by fluorosulfato groups in fluoro-fluorosulfato-

antimonate(V).

(ii) Similarly, the signal of the fluorine cis to another

fluorine shifts to low field as cis fluorines are sequentially

replaced by fluorosulfato groups in fluoro-fluorosulfato-

antimonate(V).

1878

Can. J. Chem. Vol. 77, 1999

Fig. 6. ¹⁹F–¹⁹F COSY NMR spectrum of the HSO₃F–SbF₅

Fig. 7. (a) J-resolved ¹⁹F NMR spectrum of the HSO₃F–SbF₅

system with x_SbF= 0.009 84.

system with x_SbF= 0.0500. (b) ¹⁹F–¹⁹F COSY NMR spectrum

5

of the HSO₃F–SbF₅system with x_SbF= 0.0500.

5

(iii) In the same fluoro-fluorosulfato-antimonate(V), signals

due to fluorines trans to fluorine appear at lower fields than

those trans to fluorosulfate.

These trends agree in general with the empirical equation

established for substituted perfluoro complex anions of the

m–

type [EF_nL_6–n

(28)

]

(E = Ti, W, Sn; m = 2 and E = P; m = 1)

[15] D _F= pC + qT

In this equation, D _Fis the difference in chemical shift of a

signal between the fully fluorinated and partly substituted

species. C and T are constants for the particular substituent

and central atom. The quantities p and q are the number of

substituents cis and trans, respectively, to the fluorines

whose shifts are calculated. However, attempts to quantita-

tively fit the data with the equation were not successful. A

possible explanation is, that the data collected from this sys-

tem are the averaged chemical shifts of the complex acids

and their conjugate anions.

Based on integration of the signals, about 27% of the

Sb(V) species is estimated to be present as [SbF₅(SO₃F)]^–,

52% as [SbF₄(SO₃F)₂]^–, 4% as [SbF₃(SO₃F)₃]^–, and less than

1% as [SbF₆]^–in the system with x_SbF= 0.009 82. This re-

5

sult is not consistent with initial formation of [SbF₅(SO₃F)]^–,

by direct combination of SbF₅with HSO₃F (eqs. [3] and [4])

and subsequent ligand redistribution reactions according to

Zhang et al.

1879

Table 1. Assignment of monomeric species for a HSO₃F–SbF₅system with x_SbF= 0.009 82. Chemicals shifts d_F(ppm), coupling

5

constants J_FF(Hz).

cated by broken lines, it is likely that the first two peaks on

the left belong to a triplet T5 with one peak of the multiplet

buried in the intense D2, and the high-field peak is one half

of a doublet D4, again, with the other half buried in D2.

From the corresponding ¹⁹F–¹⁹F COSY spectrum (see Fig. 6),

D4 is correlated with a quintet q1, which is partially overlapped

with the intense q4, to form another AX₄spin system. It can

be assigned to a segment of the type [SbF₅-(m-SO₃F)^–].

[16] 2[SbF₅(SO₃F)]^–W [SbF₄(SO₃F)₂]^–+ [SbF₆]^–

[17] 2[SbF₄(SO₃F)₂]^–W [SbF₃(SO₃F)₃]^–

+ [SbF₅(SO₃F)]^–

The insertion reaction of residual SO₃and SbF₅in HSO₃F

may contribute partly to the formation of species with a high

SO₃F contents in the system with x_SbF= 0.000 999. For the

In the 1D NMR spectrum of the HSO₃F–SbF₅system with

x_SbF= 0.009 82 (see Fig. 3b), the signals at –97.5~–101 ppm

are ⁵apparently two triplets, T1 and T2, with close chemical

shifts and slightly different coupling constants. This is much

clearer in the 1D ¹⁹F NMR spectrum (see Fig. 3c) and ¹⁹F J-

resolved spectrum (F(Sb) region) of the HSO₃F–SbF₅sys-

5

systems with x_SbF³ 0.009 82, such a contribution seems too

5

small, when the extremely low concentration of SO₃is con-

sidered. Solvolysis of SbF₅in HSO₃F as shown in eqs. [12]

and [13] appears to have occurred. Otherwise, a higher con-

centration of [SbF₆]^–would have been anticipated.

tem with x_SbF= 0.0500 (see Fig. 7a).

5

Assignment of segments in the oligomeric species

Although oligomerization is expected as the SbF₅content

increases, it is still surprising that it occurs in the system

with an SbF₅concentration as low as x_SbF= 0.009 82. It is

In the 1D spectrum (see Fig. 3c), the signals in the region

–97~–101 ppm are resolved into three triplets T1, T2, and

T3. The ¹⁹F–¹⁹F COSY spectrum (F(Sb) region) (see Fig. 7b)

indicates T1 and T2 are correlated with Q2 and Q5, Q3, and

Q4, respectively. In the 1D ¹⁹F NMR spectrum, the quartets

Q2 and Q3 are partially overlapped to form a “sextet” with

the area ratio of 1:3:4:4:3:1 in the region –115~–118.5 ppm

(see Inset 1, Fig. 3c). The other two quartets, Q4 and Q5,

form another “sextet” at –123.5~–127.5 ppm (see Inset 2,

Fig. 3c), which is overlapped with the readily recognizable

singlet S8 at –125.8 ppm assigned to [SbF₆]^–. Careful in-

spection of the frequencies and area ratios of the peaks sug-

noted from the spectrum of the system wit⁵h x_SbF= 0.009 82

5

(see Fig. 3b) that three small peaks just downfield of D2 do

not belong to the same multiplet. The separation from the

central peak is 100 and 50 Hz for the lower field and higher

field peaks, respectively. Moreover, the intensity of the peak

at higher field to the central peak is approximately twice the

intensity expected for a triplet. As illustrated in Inset 2 in

Fig. 3b, where the expected but unobserved peaks are indi-

1880

Can. J. Chem. Vol. 77, 1999

Fig. 8. ¹⁹F–¹⁹F COSY NMR spectrum of the HSO₃F–SbF₅

fac-[SbF₃(SO₃F)₃], [SbF₂(SO₃F)₄], and [SbF(SO₃F)₅], is less

likely because:

system with x_SbF= 0.194.

5

(i) the relative intensity of S4, S5, and S6 increase as the

SbF₅concentration increases, which is contrary to what is

expected for the monomers with a high SO₃F/F ratio. As the

x_SbFincreases from 0.000 999 to 0.009 82, in fact, the sig-

5

nals D1 and T5 due to mer-[SbF₃(SO₃F)₃]^–are barely ob-

servable, and the relative intensities of S7 due to trans-

[SbF₄(SO₃F)₂]^–and T3T6 due to cis-[SbF₄(SO₃F)₂]^–are re-

duced significantly; and

(ii) singlets S4, S5, and S6 cannot be observed even in the

spectrum of the HSO₃F–SbF₅system with x_SbF= 0.000 999,

which has a high HSO₃F–SbF₅mole ratio. Si⁵nglets S1, S2,

and S3 are observed only in the HSO₃F–SbF₅system with

low HSO₃F–SbF₅mole ratios (x_SbF³ 0.194).

For the same reasons, assignment⁵of these singlets to seg-

ments with a high SO₃F content, such as trans-[(SO₃F)₃SbF₂-

(m-SO₃F)-] and [(SO₃F)₄SbF-(m-SO₃F)-], seems implausible.

Two triplets, T4 and T7, emerge in the ¹⁹F NMR spectrum

of the system with x_SbF= 0.0989 (the inset in Fig. 3d) and

5

higher. Parallel to triplets T3 and T6, which are attributed to

the monomeric species cis-[SbF₄(SO₃F)₂]^–, T4 and T7 are

also correlated to each other as seen from the ¹⁹F–¹⁹F COSY

spectrum (see Fig. 8).

They compose another A₂X₂spin system, which is better

attributed to a segment, cis-[-(m-SO₃F)-SbF₄-(m-SO₃F)-],

linking two identical moieties.

A glimpse at the signals at –109~–111 ppm in spectra of

the systems with x_SbF= 0.0989 and 0.194 may give the im-

5

pression that they are three overlapped doublets. However,

the separation between the two peaks of lowest intensity var-

ies from 55 Hz for x_SbF= 0.0989–72 Hz for x_SbF= 0.194.

5

Both these separations are lower than all the coupling con-

stants found for terminal fluorines in this particular system.

Analysis of the intensities of the signals suggests that they

may be composed of four doublets. In the spectrum of the

gests that each of them results from A or M of two AMX₂

spin systems T1Q2Q5 and T2Q3Q4, respectively. These two

AMX₂systems can be assigned to the segments of the type

cis-[(SO₃F)SbF₄-(m-SO₃F)-], or of the type cis-[-(m-SO₃F)-

SbF₄-(m¢-SO₃F)-]with two different terminal segments on

each side.

HSO₃F–SbF₅system with x_SbF= 0.194, for example, these

5

signals were analyzed by deconvolution into the following

2

four doublets: D2 (109.6 ppm, J_FF= 99 Hz), D3

It is also evident from the J-resolved spectrum of the

x_SbF= 0.0500 system (see Fig. 7a) that the signals at

–13⁵4~–139 ppm are three overlapped quintets, q1, q2, and

q4. In the ¹⁹F–¹⁹F COSY spectrum, the cross peaks D2q4,

D4q2, and D5q1 indicate that q4, q2, and q1 are coupled

with three partially overlapped doublets D2, D4, and D5, re-

spectively, to form three AX₄systems. One of the AX₄sys-

tems, D2q4, has been assigned to monomeric [SbF₅(SO₃F)]^–.

The other two may be assigned to a segment of the type

[SbF₅(-SO₃F)-] with different segments on the other side of

the bridging fluorosulfato group.

(110.1 ppm, ²J_FF= 100 Hz), D4 (110.3 ppm, ²J_FF= 100 Hz),

2

and D5 (110.9 ppm, J_FF= 100 Hz) (see Inset 2, Fig. 3e).

From the ¹⁹F–¹⁹F COSY spectrum (see Figs. 7 and 9), these

doublets are correlated to the quintets q1, q2, q3, and q4

at –134~–140 ppm, respectively. Signal q3, which is corre-

lated to the weakest of the doublets, D3, is so weak that it

cannot be clearly observed in the spectrum of the x_SbF

=

5

0.0989 system. As shown in Inset 3 of Fig. 3e, only one

peak remains unassignable to either q1, q2, or q4 and there-

fore is assigned to q3 while the rest of q3 is buried in the in-

tense signal due to q4. Among four AX₄spin systems, D2q4

has been assigned to the monomer [SbF₅(SO₃F)]^–. The other

three AX₄systems can be attributed to fragments of the type

[SbF₅-(m-SO₃F)-].

A quartet, Q1, with the area ratio of 1:1:1:1 appears at

~–116.4 ppm in the spectra of the systems with x_SbF³ 0.342.

It can be assigned to the eight fluorines cis to the⁵bridging

fluorine in [Sb₂F₁₁]^–. As sketched in Inset 2 of Fig. 3f, the

signal is split into two lines with a 1:1 ratio by a terminal

fluorine (²J_FF= 102 Hz) and then further split by the bridg-

ing fluorine (²J_FF= 60 Hz). Only the most intense portion of

the nonet N1, due to the bridging fluorine, is barely observ-

In the ¹⁹F NMR spectrum of the HSO₃F–SbF₅system

with x_SbF= 0.009 82 (see Fig. 3b), the weak signals at –

5

103.50 ppm are recognized as three singlets, and based on

similarities in chemical shifts, are possibly the same singlets

as S4, S5, and S6 observed for the x_SbF= 0.0500 system

(see Fig. 3c). They can be assigned to se⁵gments of the type

trans-[(SO₃F)-SbF₄-(m-SO₃F)-]. At high x_SbF, there are ad-

5

ditional three singlets (S1, S2, and S3) at lower field (see In-

set 1, Figs. 3e and 3f), which may be better assigned to

segments of the type trans-[-(m-SO₃F)-SbF₄-(m¢-SO₃F)-] in

trimers or higher oligomeric species. The assignment of

these singlets to monomers with high SO₃F/F ratios, such as

Zhang et al.

1881

Fig. 9. ¹⁹F–¹⁹F COSY NMR spectrum of the HSO₃F–SbF₅system with x_SbF= 0.342.

5

able at –91.0 ppm (see Inset 1, Fig. 3f). The quintet q5, due

to the two terminal fluorines, falls into the q1q2q3q4 region.

The correlation between N1, Q1, and q5, however, is very

clear from the corresponding ¹⁹F–¹⁹F COSY spectrum (see

Fig. 9).

The assignments discussed above are summarized in Ta-

ble 2. Some variations in chemical shifts are observed for

gin to appear. The relative concentration of oligomeric spe-

cies is expected to increase while the relative concentration

of monomeric antimony(V) species decreases accordingly.

Among the monomeric species observed, [SbF₃(SO₃F)₃]^–

has the highest SO₃F content. Therefore, the dimeric species

would likely be limited to the types [Sb₂F_11–n(SO₃F)_n] (n =

0–3), formed by the following dimerizations by addition of

SbF₅to the monomeric species:

the signals as x_SbFchanges. The data listed in Table 2 are

5

measured from the spectrum of the HSO₃F–SbF₅system

[18] [SbF₆]^–+ SbF₅W [Sb₂F₁₁]^–

with x_SbF= 0.0989 except signals due to trans-[-(m-SO₃F)-

5

SbF₄-(m¢-SO₃F)-] and the dimer [Sb₂F₁₁]^–, which can only be

[19] [SbF₅(SO₃F)]^–+ SbF₅W [Sb₂F₁₀(SO₃F)]^–

[20] [SbF₄(SO₃F)₂]^–+ SbF₅W [Sb₂F₉(SO₃F)₂]^–

[21] [SbF₃(SO₃F)₃]^–+ SbF₅W [Sb₂F₈(SO₃F)₃]^–

clearly observed for HSO₃F–SbF₅systems with x_SbF³ 0.194

5

and for systems with x_SbF³ 0.342, respectively.

5

With the assignments proposed above, it is possible to

estimate the relative concentrations of different segments in

the HSO₃F–SbF₅system with x £ 0.342. The results are pre-

sented in Table 3. Due to the signal overlap and possible in-

accurate baseline settings over the wide frequency range,

integration errors inevitably occur to some extent. The errors

are small for intense, narrow peaks and large for weak,

broad peaks. The estimation of concentration of segments is

not feasible for the systems with x_SbF³ 0.489, since integra-

tion of the individual signals from ⁵overlapped, broadened

peaks is not reliable.

Dimeric SO₃F-containing species are unlikely to be linked

by F-bridges. Studies of ternary fluoride fluorosulfates based

on vibrational spectra indicate that fluorosulfate always

takes precedence over fluoride as a bridging ligand (29–33).

Moreover, F-bridged dimers would have spin systems of

three or more different types of fluorines, which should re-

sult in more complicated ¹⁹F NMR spectra. In this study,

simple splitting patterns are observed, which can be inter-

preted either in terms of monomeric species or in terms of

segments linked by SO₃F-bridges. Most spin systems ob-

served consist of one or two different types of fluorines.

Two AMX₂spin systems are satisfactorily assigned to seg-

Possible oligomeric species in the HSO₃F–SbF₅system

As the concentration of the Lewis acid SbF₅increases, it

is expected that oligomeric antimony(V) species should be-

1882

Can. J. Chem. Vol. 77, 1999

Table 2. Assignment of ¹⁹F NMR signals (F(Sb) region) to segments in oligomeric fluoro-fluorosulfato-antimonate(V) species in the

HSO₃F–SbF₅system with x_SbF= 0.0989. Chemical shifts d_F[ppm], coupling constants J_FF[Hz].

5

ments of the types cis-[SbF₄(SO₃F)-(m-SO₃F)-] or cis-[-(m-

SO₃F)-SbF₄-(m¢-SO₃F)-]. Coupling constants observed for

most multiplets (98~120 Hz) point to terminal rather than

bridging fluorines. The only exceptions are N1 and Q1 for

which smaller coupling constants (60 Hz) are observed. In-

deed, N1 and Q1 are due to the F-bridged perfluoro anion

[Sb₂F₁₁]^–, where no SO₃F^–competes with F^–as the bridg-

ing ligand. The small coupling constant (60 Hz) results from

the coupling between the bridging fluorine and cis-terminal-

fluorines (equatorial fluorines).

Since the fluorine on the bridging SO₃F^–does not couple

with other fluorines, information about connectivity between

different segments of SO₃F-bridged oligomers is unavail-

able. Only for the F-bridged dimer [Sb₂F₁₁]^–is the connec-

tivity manifested in a unique nonet due to the coupling of the

bridging fluorine to eight equivalent terminal cis-fluorines. In

the spectrum of the system with x_SbF= 0.342 (see Fig. 3f),

5

only a portion of the characteristic nonet N1 centered at d

CFCl₃= –91.0 ppm can be observed because of its low in-

tensity, but the correlated Q1 at d CFCl₃= –116.4 ppm due

to eight cis-fluorines is clearly observed.

Illustrated in Fig. 10 are all possible dimeric isomers of

the types [Sb₂F_11–n(SO₃F)_n] (n = 0–3).

Among the four AX₄spin systems, D2q4 has been as-

signed to the monomer [SbF₅(SO₃F)]^–. D5q1 is tentatively

assigned to the dimer [Sb₂F₁₀(m-SO₃F)]^–and D4q2 to the

segment [SbF₅-(m-SO₃F)-] in both cis- and trans-[SbF₅-(m-

SO₃F)-SbF₄(SO₃F)]^–. It is noted that the appearance of the

AX₄spin system D3q3 is accompanied by the emergence of

the A₂X₂spin system T4T7 and three singlets S1, S2, and

S3. T4T7 can be assigned to a segment cis-[(-(m-SO₃F)-

SbF₄-(m-SO₃F)-] linking two identical moieties at each end.

S1, S2, and S3 are assigned to segments of the type trans-

[-(m-SO₃F)-SbF₄-(m-SO₃F)-], which link either identical or

different moieties at each end. The coincident appearance of

these spin systems suggests the formation of various trimeric

The last three isomers of the type [Sb₂F₈(SO₃F)₃]^–have

two terminal fluorosulfates on the same antimony. Each of

the three isomers has an AX₂spin system. According to

eq. [20], these isomers should have formed from the direct

combination of mer-[SbF₃(SO₃F)₃]^–and SbF₅. However, sig-

nals due to AX₂spin systems other than those assigned to

the monomer mer-[SbF₃(SO₃F)₃]^–, are not observed in the

F(Sb) region. This suggests they may have rearranged to

other isomers. Thus, the possible dimeric Sb(V) species are

limited to [Sb₂F₁₁]^–, [Sb₂F₁₀(m-SO₃F)]^–, cis- and trans-

[Sb₂F₉(SO₃F)₂]^–, and cis,cis-, cis,trans-, and trans,trans-

[SbF₄(SO₃F)-(m-SO₃F)-SbF₄(SO₃F)]^–.

Zhang et al.

1883

Fig. 10. Dimeric species of the type [Sb₂F_11–n(SO₃F)_n]^–(n = 0–3) and corresponding spin systems in the F (Sb) region.

fluoro-fluorosulfato-antimonate(V) species, e.g., cis- and

trans-[SbF₅-(m-SO₃F)-SbF₄-(m-SO₃F)-SbF₅]^–. However, it must

be pointed out that the possibility of the existence of higher

oligomeric species cannot be excluded, particularly in the

systems with high SbF₅/HSO₃F ratios.

The most intense signal of the four singlets at low field

(–100~–105 ppm), S7, has been assigned to the monomeric

[SbF₄(SO₃F)₂]^–. The remaining three singlets are assigned to

the fragments of the type trans-[SbF₄(SO₃F)-(m-SO₃F)-]. S4

is tentatively assigned to trans-[SbF₅-(m-SO₃F)-SbF₄-(SO₃F)]^–,

S5 to cis, trans-[SbF₄(SO₃F)-(m-SO₃F)-SbF₄(SO₃F)], and S6

to trans, trans-[SbF₄(SO₃F)-(m-SO₃F)-SbF₄(SO₃F)].

dimers cis,cis- and cis,trans-[SbF₄(SO₃F)-(m-SO₃F)-SbF₄(SO₃F)]^–

or vice versa. The analysis of relative concentrations of seg-

ments favors the assignment of T2Q3Q4 to the dimer cis-

[SbF₅-(m-SO₃F)-SbF₄(SO₃F)]^–and T1Q2Q5 to the dimers

cis,cis- and cis,trans-[SbF₄(SO₃F)-(m-SO₃F)-SbF₄(SO₃F)]^–.

It is entirely possible that, at high x_SbF, terminal fluoro-

sulfato groups in the fragments of the typ⁵e cis-[SbF₄(SO₃F)-

(m-SO₃F)-] may transform into bridging fluorosulfato groups

to form fragments of the type cis-[-(m-SO₃F)-SbF₄-(m¢-SO₃F)-],

which then link two non-equivalent terminal segments to

form trimers or even higher oligomers. However, there is no

definitive evidence for such transformations in the ¹⁹F NMR

spectra. The variation in chemical shifts of the signals may

also be attributed to the shifts of the equilibria between com-

plex acids and their conjugate anions and to the change of

the medium.

Of the two AMX₂spin systems, T2Q3Q4 and T1Q2Q5,

either can be assigned to the segment cis-[SbF₄(SO₃F)-(m-

SO₃F)-] in the dimer cis-[SbF₅-(m-SO₃F)-SbF₄(SO₃F)]^–and

the other to the segment cis-[SbF₄(SO₃F)-(m-SO₃F)-] in the

1884

Can. J. Chem. Vol. 77, 1999

Table 3. Relative concentrations of monomeric species and segments in oligomeric species in the HSO₃F–SbF₅system.

Molar fraction of SbF₅, x_SbF

5

0.000999

0.00982

0.0500

0.0989

0.194

0.342

Species

Signals

Relative concentration, Sb(V) mol %

Monomeric species

[SbF₆] ^–

S8

2

15

48

21

14

<1

27

39

13

4

<1

36

20

5

<1

42

12

3

<1

45

8

1

0

<1

37

7

<1

0

[SbF₅(SO₃F)]^–

D2q4

T3T6

S7

cis-[SbF₄(SO₃F)₂] ^–

trans-[SbF₄(SO₃F)₂] ^–

mer-[SbF₃(SO₃F)₃] ^–

D1T5

2

0

Segments in oligomeric species

[SbF₅-(m-F)]

Q1

0

3

0

6

0

1

<1

0

10

7

0

8

0

3

10

12

<1

4

10

0

6

7

15

4

5

<1

15

4

17

8

D3q3

D4q2

D5q1

T4T7

T1Q2Q5

T2Q3Q4

S1

S2

S3

S4

S5

[SbF₅(m-SO₃F)]

cis-[SbF₄(m-SO₃F)₂]

cis-[SbF₄(SO₃F)(m-SO₃F)] or

cis-[SbF₄(m-SO₃F)(m¢-SO₃F)]

2

6

8

<1

1

<1

1

trans-[SbF₄(SO₃F)(m-SO₃F)]

or

trans-[SbF₄(m-SO₃F)(m¢-SO₃F)]

0

<1

1

<1

2

S6

2

Table 4. Proposed principal species and their relative concentrations in HSO₃F–SbF₅system.

Molar fraction of SbF₅, x_SbF

5

0.000999 0.00982 0.0500 0.0989 0.194 0.342

Sb(V) mol %

Constituent species

[SbF₆] ^–

2

15

48

21

14

100

0

<1

20

39

13

4

76^a

0

3

11

<1

5

<1

36

20

5

2

63

0

<1

42

12

3

0

57

0

12

<1

20

2

3

2

<1

45

9

1

0

54

0

15

1

14

1

5

<1

37

7

<1

0

44

<1

17

1

10

1

[SbF₅(SO₃F)]^–

cis-[SbF₄(SO₃F)₂]^–

trans-[SbF₄(SO₃F)₂]^–

mer-[SbF₃(SO₃F)₃]^–

Total of monomeric species

[Sb₂F₁₁]^–

[SbF₅-(m-SO₃F)-SbF₅]^–

trans-[SbF₅-(m-SO₃F)-SbF₄(SO₃F)]^–

cis-[SbF₅-(m-SO₃F)-SbF₄(SO₃F)]^–

trans,trans-[SbF₄(SO₃F)-(m-SO₃F)-SbF₄(SO₃F)]^–

cis,cis-[SbF₄(SO₃F)-(m-SO₃F)-SbF₄(SO₃F)]^–

cis,trans-[SbF₄(SO₃F)-(m-SO₃F)-SbF₄(SO₃F)]^–

Total of dimeric species

7

<1

17

2

6

4

36

0

36

0

2

<1

36

10

46

<1

31

24

55

24^a

0

39

4

43

cis-[SbF₅-(m-SO₃F)-SbF₄-(m-SO₃F)-SbF₅]^–, other trimers, and higher oligomers

Total of oligomeric species

0

24^a

^aFrom T2Q3Q4 + S4 = 7, it is expected that ca. 7% Sb(V) species may exist as the fragment [SbF₅-(m-SO₃F)-] in cis- and trans-[Sb₂F₁₀(SO₃F)]^–. The

signals D4q2 due to the -SbF₅moiety (an AX₄spin system) is overlapped with D2q4 due to another AX₄spin system by coincidence.

Based on these assignments, it is possible to estimate the

relative concentrations of the proposed oligomeric species

by analysis of the relative concentration of segments. The re-

Such an estimation is not feasible for HSO₃F–SbF₅sys-

tems with x_SbF³ 0.342 because it is impossible to obtain in-

5

tegral values for each of the overlapped, broad signals.

However, it can be roughly estimated that about 67% of

Sb(V) species exist as [SbF₅]- moieties, 25% as cis-[SbF₄]-

sults for HSO₃F–SbF₅systems with x_SbF£ 0.342 are listed

5

in Table 4.

Zhang et al.

1885

moieties, 4.5% as trans-[SbF₄]- moieties, and 3.5% as [Sb₂F₁₁]^–

Acknow ledgements

in the x_SbF= 0.489 system.

5

Thanks are due to the Fonds der Chemischen Industrie e.

V. for material support (G.H.), the Natural Sciences and En-

gineering Research Council of Canada for operating funds

(F.A.), and the NATO for a collaborative grant (G.H. and

F.A.).

Summary and conclusion

The results presented here depart from previous work by

Commeyras et al. (17, 19) in two important aspects:

References

(i) Concentrations of the monomeric anions cis- and trans-

[SbF₄(SO₃F)₂]^–are at any mole fraction of SbF₅consider-

ably higher than the concentration of [SbF₆]^–(see Table 3).

Hence, formation of the bis-fluorosulfato anions [SbF₄(SO₃F)₂]^–

solely by ligand redistribution of [SbF₅(SO₃F)]^–(eq. [9]) be-

comes improbable. Instead solvolysis of SbF₅or possibly of

SbF₄(SO₃F) in HSO₃F (eqs. [12] and [13]) and subsequent

complex formation with either SO₃F^–or F^–is much more

likely. The byproduct HF is rapidly converted by reaction

with glass to H₂O (eq. [12]), which functions as a base to

give the oxonium ion [H₃O]⁺. The [H₃O]⁺ion is identified

1. G.H. Olah, G.K.S. Prakash, and J. Sommer. In Superacids. J.

Wiley, New York. 1985, and refs. therein.

2. T.A. O’Donnell. In Superacids and acidic melts as inorganic

chemical reaction media. VCH Publishers, New York. 1993,

and refs. therein.

3. R.J. Gillespie. Acc. Chem. Res. 1, 202 (1968).

4. G.H. Olah and P. von R. Schleyer. In Carbonium ions. Vols. 1–

5. Wiley Interscience, New York. 1968, 1970, 1973, 1973, and

1976.

5. R.C. Thompson. In Inorganic sulphur chemistry. Edited by G.

Nickless. Elsevier, Amsterdam. 1968. p. 587.

6. M.F.A. Dove and A.F. Clifford. In Chemistry in non-aqueous

ionizing solvents. Edited by J. Jander, H. Spandau, and C.C.

Addison. Vieweg, Braunschweig/Wiesbaden. 1971, and

refs. therein.

7. L.P. Hammett and A.J. Deyrup. J. Am. Chem. Soc. 54, 2721

(1932).

8. R.J. Gillespie and J. Liang. J. Am. Chem. Soc. 110, 6037

(1988).

9. P.L. Fabré, J. Devynk, and B. Tremillon. Chem. Rev. 82, 591

(1982), and refs. therein.

10. R.J. Gillespie and P.E. Peel. J. Am. Chem. Soc. 95, 5173

(1973).

1

by H NMR and structurally characterized as [H₃O][Sb₂F₁₁]

(22).

(ii) In oligomeric anions or in segments of oligomers,

bridging by m-SO₃F groups, originally suggested by

Gillespie et al. (16, 18) for dilute solutions is now observed

even at high SbF₅concentrations. A m-fluoro group is only

found in [Sb₂F₁₁]^–. Oligomeric anions with fluorobridges

and terminal SO₃F-groups of the type [(SbF₅)_n(SO₃F)]^–(n =

2–4) all previously identified by Commeyras and co-workers

(19) are not detectable with modern NMR techniques.

Their postulated formation by reaction of cis-F-bridged

oligomeric (SbF₅)_n(n = 2–4) with the selfionization SO₃F^–

according to

11. R.J. Gillespie and P.E. Peel. Adv. Phys. Org. Chem. 9, 1 (1972).

12. V. Gold, K. Laali, K.P. Morris, and L.Z. Zdunek. J. Chem.

Soc. Chem. Commun. 769 (1981).

13. V. Gold, K. Laali, K.P. Morris, and L.Z. Zdunek. J. Chem.

Soc. Perkin Trans. 2, 859 (1985).

[22] (SbF₅)_n+ SO₃F^-¾¾¾¾®[(SbF₅)_n(SO₃F)] ^-

(solv)

HSO₃F

14. J. Sommer, P. Canivet, S. Schwartz, and P. Rimmelin. J. Am.

Chem. Soc. 100, 2576 (1978).

n = 2-4

15. J. Sommer, P. Canivet, S. Schwartz, and P. Rimmelin. Nouv. J.

Chim. 5, 45 (1981).

16. R.C. Thompson, J. Barr, R.J. Gillespie, J.B. Milne, and R.A.

Rothenbury. Inorg. Chem. 4, 1641 (1965).

17. A. Commeyras and G.A. Olah. J. Am. Chem. Soc. 91, 2929

(1969).

is not supported by our findings. The cis-F-bridges found in

liquid SbF₅do not persist in magic acid, when samples are

prepared as described in the experimental section. Our re-

sults are consistent with spectroscopic findings for the

ansolvo species SbF₄(SO₃F) (29, 30), SbF₃(SO₃F)₂(30), and

others (31–33) where in all instances only fluorosulfato

bridges are observed.

18. P.A.W. Dean and R.J. Gillespie. J. Am. Chem. Soc. 91, 7264

(1969).

19. D. Brunel, A. Germain, and A. Commeyras. Nouv. J. Chim. 2,

275 (1978).

20. J. Bacon, P.A.W. Dean, and R.J. Gillespie. Can. J. Chem. 48,

3413 (1970).

21. J.K.M Sanders and B.K. Hunter. Modern NMR spectroscopy:

a guide for chemists. 2nd ed. Oxford University Press, Oxford.

1993.

The evidence for solvolysis of the Lewis acids SbF₅and

SbF₄(SO₃F) (eqs. [12] and [13]) with HF and, via interaction

with glass, of H₂O as intermediates, presented here and else-

where (22), is reason for concern. Protonation of H₂O,

which is a base in HSO₃F, is expected to reduce the proton

acidity of magic acid drastically.

It is conceivable, that the difference in proton acidity be-

tween HF–SbF₅and HSO₃F–SbF₅noted by Sommer et al.

(14, 15) is due to water formed in these side reactions. In

spite of this limitation, the convenience of working in glass

outweighs in most cases the slight reduction in proton acidi-

ties (14, 15). To study the effect water has on the types and

concentrations of anions observable by ¹⁹F NMR in magic

acid we have recently completed an identical study in a glass

free environment²and will report shortly on the results.

22. D. Zhang, S.J. Rettig, J. Trotter, and F. Aubke. Inorg. Chem.

35, 6113 (1996).

23. R.J. Gillespie and E.A. Robinson. Can. J. Chem. 40, 675 (1962).

24. M. Lustig. Inorg. Chem. 4, 1828 (1965).

25. R.J. Gillespie, J.B. Milne, and R.C. Thompson. Inorg. Chem.

5, 468 (1966).

26. R.A. Stewart, S. Fujiwara, and F. Aubke. J. Chem. Phys. 49,

965 (1968).

27. R.J. Gillespie and J.W. Quail. J. Chem. Phys. 39, 2555 (1963).

1886

Can. J. Chem. Vol. 77, 1999

28. P.J. Chevrier and S. Brownstein. J. Inorg. Nucl. Chem. 42,

1397 (1980).

29. R.J. Gillespie and R.A. Rothenbury. Can. J. Chem. 42, 416

(1964).

31. H. Imoto and F. Aubke. J. Fluorine Chem. 15, 59 (1980).

32. D. Zhang and F. Aubke. J. Fluorine Chem. 58, 81 (1992).

33. H. Willner, F. Mistry, and F. Aubke. J. Fluorine Chem. 59, 333

(1992).

30. W.W. Wilson and F. Aubke. J. Fluorine Chem. 13, 431 (1979).

Article Doi

A (19)F nuclear magnetic resonance study of the conjugate Bronsted-Lewis superacid HSO3F-SbF5. Part 1

DOI: 10.1139/v99-182

Source and publish data:

Authors:

Article abstract of DOI:10.1139/v99-182

Full text of DOI:10.1139/v99-182

Products guided by the article

R&D Labs maybe for 286013-53-4

Relevant to this article

Hot Product