On the XeF+/H2O system: Synthesis and characterization of the xenon(II) oxide fluoride cation, FXeOXeFXeF+

Published on Web 08/26/2009

On the XeF⁺/H₂O System: Synthesis and Characterization of

the Xenon(II) Oxide Fluoride Cation, FXeOXeFXeF⁺

Michael Gerken,^†,|Matthew D. Moran,^†He´le`ne P. A. Mercier,^†Bernard E. Pointner,^†

Gary J. Schrobilgen,*^,†Berthold Hoge,^‡,⊥Karl O. Christe,*^,‡and Jerry A. Boatz^§

Department of Chemistry, McMaster UniVersity, Hamilton ON L8S 4M1, Canada, Loker

Hydrocarbon Research Institute and Department of Chemistry, UniVersity of Southern

California, Los Angeles, California 90089, and Air Force Research Laboratory, Edwards Air

Force Base, California 93524

Received June 19, 2009; E-mail: schrobil@mcmaster.ca; kchriste@usc.edu

Abstract: The reported synthesis of the H₂OF⁺cation as a product of the oxidative ﬂuorination of H₂O by

[XeF][PnF₆] (Pn ) As, Sb) in HF solution has been reinvestigated. The system exhibits complex equilibria,

producing two new Xe(II) compounds, [Xe₃OF₃][PnF₆] and [H₃O][PnF₆]·2XeF₂, refuting the original claim

for the synthesis of the H₂OF⁺cation. Both compounds have been isolated and characterized by vibrational

spectroscopy and single-crystal X-ray diffraction. The X-ray crystal structures of the [Xe₃OF₃][PnF₆] salts

contain the Z-shaped FXeOXeFXeF⁺cation, which represents the ﬁrst example of an isolated Xe(II) oxide

ﬂuoride. The crystal structure of the [H₃O][AsF₆]·2XeF₂adduct contains XeF₂molecules that interact with

the H₃O⁺cations. The vibrational assignments for the Xe₃OF₃cation have been made with the aid of

+

quantum-chemical calculations and were conﬁrmed by ¹⁸O-enrichment, and the assignments for

[H₃O][AsF₆]·2XeF₂were conﬁrmed by ²D- and ¹⁸O-enrichment. Quantum-chemical calculations have also

been carried out for H₃O⁺·nXeF₂(n ) 1-4) and have been used to interpret the X-ray crystal structure

and vibrational spectra of [H₃O][AsF₆]·2XeF₂. The energy-minimized geometries and vibrational frequencies

for HOF and H₂OF⁺have been calculated, further disproving the original report of the H₂OF⁺cation. Both

+

FXeOH and FXeOH₂have also been computed and are viable intermediates in the proposed equilibria

between XeF⁺and H₂O that lead to the Xe₃OF₃cation.

+

ﬂuorous acid, HOF, a well-known unstable compound,^8,9and

its protonated form, the H₂OF⁺cation,^10,11if indeed it should

Introduction

Chlorine oxide pentaﬂuoride, ClOF₅, would be the highest

performing earth-storable, liquid rocket-propellant oxidizer.

Although the synthesis of ClOF₅had been claimed in 1972,¹

this claim was subsequently refuted.²Continued efforts to

prepare this compound by oxidative ﬂuorination of substrates

exist. It was claimed that the synthesis of the latter cation could

be accomplished by the oxidative ﬂuorination of H₂O with

[XeF][PnF₆] (Pn ) As, Sb) in HF solution at -60 °C.^10,11The

resulting product, formulated as [H₂OF][PnF₆], was reported to

1

be pale red and was characterized by infrared and H and ¹⁹F

3-5

or ClOF₄^6,7have failed because the oxygen

-

such as ClOF₃

NMR spectroscopy. It was of particular interest because it

ligand is more easily oxidized than the central chlorine atom.

This failure prompted a search for alternate synthetic approaches,

such as the reaction of ClF₅with an oxidative oxygenator.

Potential candidates for oxidative oxygenators include hypo-

10,11

+

allegedly oxygenated ClF₃to ClOF₂

.

Despite the remarkable oxygenating power attributed to the

alleged H₂OF⁺cation,^10,11no further reports concerning H₂OF⁺

have materialized since its reported discovery, except for two

Master’s theses from the same laboratory,^12,13which failed to

substantiate the remarkable oxidizing power attributed by the

earlier authors to their product.

This paper describes attempts to reproduce the syntheses of

H₂OF⁺salts and their reactions with chlorine ﬂuorides.^10,11

Although these attempts failed and proved that the previous

^†McMaster University.

^‡University of Southern California.

^§Edwards Air Force Base.

^|Present address: Department of Chemistry and Biochemistry, The

University of Lethbridge, Lethbridge AB T1K 3M4, Canada.

^⊥Present address: Fakulta¨t fu¨r Chemie, Universita¨t Bielefeld, Postfach

10 01 31, D-33501 Bielefeld, Germany.

(1) Zuechner, K.; Glemser, O. Angew. Chem., Int. Ed. Engl. 1972, 11,

1094–1095.

(8) Appelman, E. H. Acc. Chem. Res. 1973, 6, 113–117.

(2) Christe, K. O.; Schack, C. J. AdV. Inorg. Chem. Radiochem. 1976,

18, 319.

(9) Appelman, E. H.; Dunkelberg, O.; Kol, M. J. Fluorine Chem. 1992,

56, 199–213.

(3) Christe, K. O.; Pilipovich, D.; Lindahl, C. B.; Schack, C. J.; Wilson,

R. D. Inorg. Chem. 1972, 11, 2189–2192.

(10) Minkwitz, R.; Nowicki, G. Angew. Chem., Int. Ed. Engl. 1990, 29,

688–689.

(4) Pilipovich, D.; Rogers, H. H.; Wilson, R. D. Inorg. Chem. 1972, 11,

2192–2195.

(11) Nowicki, G. Ph.D. Thesis, Universita¨t Dortmund, Dortmund, Germany,

1992.

(5) Christe, K. O.; Curtis, E. C. Inorg. Chem. 1972, 11, 2196–2201.

(6) Christe, K. O.; Schack, C. J.; Pilipovich, D. Inorg. Chem. 1972, 11,

2205–2208.

(12) Schwenke, S. Diplomarbeit, Universita¨t Dortmund, Dortmund, Ger-

many, 1990.

(13) Zoermer, M. Diplomarbeit, Universita¨t Dortmund, Dortmund, Ger-

many, 1991.

(7) Christe, K. O.; Curtis, E. C. Inorg. Chem. 1972, 11, 2209–2211.

9

13474 J. AM. CHEM. SOC. 2009, 131, 13474–13489

10.1021/ja905060v CCC: $40.75  2009 American Chemical Society

Xenon(II) Oxide Fluoride Cation, FXeOXeFXeF⁺

A R T I C L E S

claims were false, they have led to the isolation and character-

ization of the ﬁrst genuine example of a Xe(II) oxide ﬂuoride,

FXeOXeFXeF⁺(hereafter referred to as Xe₃OF₃⁺), as well as

the adduct, [H₃O][AsF₆]·2XeF₂. Neutral xenon oxide ﬂuorides

of Xe(IV) (XeOF₂),^14-16Xe(VI) (XeOF₄,¹⁷XeO₂F₂¹⁸), and

Xe(VIII) (XeO₃F₂,¹⁹XeO₂F₄²⁰) are known, and the xenon oxide

[XeF][AsF₆] + HF + H₂O h XeF₂+ [H₃O][AsF₆]

[XeF][AsF₆] + XeF₂h [Xe₂F₃][AsF₆]

(1)

(2)

(i) XeF⁺/H₂O in HF at -78 °C. The [XeF][AsF₆]/H₂O and

XeF₂/[H₃O][AsF₆] systems showed no tendency to eliminate

xenon gas at -78 °C over periods of up to 8 months and were

initially investigated at this temperature because the reaction

rates were sufﬁciently slow to permit low-temperature Raman

spectra of the reaction intermediates to be recorded at various

intervals. Moreover, equimolar reaction mixtures of [XeF][AsF₆]

and H₂O were soluble in HF at initial concentrations of ca. 0.5

M and were stable with respect to Xe(II) reduction up to -30

°C, decomposing above this temperature with xenon gas

evolution (δ(¹²⁹Xe), -5308.3 ppm; xenon dissolved in HF at

-30 °C). This behavior contrasts with the original report of

the oxidative ﬂuorination of H₂O by XeF⁺at -60 °C to yield

H₂OF⁺and Xe gas.¹⁰Reaction rates were strongly dependent

on the concentration and the degree of initial mixing. Typically,

at equimolar XeF⁺/H₂O concentrations at -78 °C in HF,

[H₃O][AsF₆]·2XeF₂was initially formed within ca. 24 h at 0.25

M and within ca. 2 h at 2 M, followed by the formation of

[Xe₂F₃][AsF₆] as the major species in the solid precipitate within

ca. 48 h at 0.25 M and within ca. 2 h at 2 M. Upon standing

for 0.5-8 days at -78 °C, deep red-orange, crystalline

[Xe₃OF₃][AsF₆] was formed as the major product. The highest

initial concentration of XeF⁺/H₂O (2 M) provided the most rapid

synthesis of [Xe₃OF₃][AsF₆].

21-23

XeO₂F⁺,^21,22,24,25

+

ﬂuoride cations of Xe(VI) (XeOF₃

,

µ-F(XeO₂F)₂^24,25) have been synthesized by ﬂuoride ion

abstraction from the parent neutral oxide ﬂuoride using the

strong ﬂuoride acceptors AsF₅or SbF₅. Prior to this work, no

systematic study of the hydrolytic behavior of Xe(II) ﬂuoride

species had been performed, though it has been noted that

aqueous solutions of XeF₂are stable for short periods of time,²⁶

and no examples of neutral or ionic Xe(II) oxide ﬂuorides had

been isolated and reported.

+

Results and Discussion

Syntheses and Properties. To elucidate the true products of

the reaction of H₂O with XeF⁺, NMR spectroscopy was used

to characterize HF solutions of H₂O and [XeF][AsF₆], as well

as HF and BrF₅solutions of XeF₂and [H₃O][AsF₆]. Reactions

of HOF with HF/AsF₅and HOSO₂F/SbF₅were also studied to

establish whether HOF can be protonated in superacidic media.

Solid products of the reactions of XeF₂with H₃O⁺were studied

by low-temperature Raman spectroscopy in situ under HF

solvent and at various stages during controlled warm-up of solid

reaction mixtures toward room temperature. Finally, reaction

conditions for the reliable syntheses of [Xe₃OF₃][PnF₆] (Pn )

As, Sb) salts in anhydrous HF have been parametrized.

(ii) Intermediacy of FXeOH and H₂O---XeF⁺. To account for

the formation of Xe₃OF₃⁺, the intermediate FXeOH is necessary,

and it is reasonable to assume that it is initially formed by the

hydrolysis of XeF₂and/or XeF⁺(eqs 3 and 4). The FXeOH

(a) [FXeOXeFXeF][AsF₆]. Equimolar solutions of [XeF][AsF₆]

and H₂O (ca. 1 M each) in HF at -78 °C were found to be in

equilibrium with [H₃O][AsF₆] and XeF₂(eq 1; see NMR

spectroscopy). At a higher XeF⁺concentration, trigonal

[Xe₂F₃][AsF₆] crystallized from a HF solution that was initially

composed of [XeF][AsF₆] and H₂O in a 3:1 molar ratio (ca.

2.8 M in XeF⁺) at -30 °C (eq 2). When larger excesses of

H₂O were used (i.e., 2:1 and 4:1 molar ratios of H₂O/

[XeF][AsF₆]) at -30 °C, the right-hand side of eq 1 was favored.

Beyond the establishment of these equilibria, the complex

reactions that take place among XeF₂, [H₃O][AsF₆], [XeF]-

[AsF₆], and H₂O in HF solution were extensively investigated

by periodically monitoring the low-temperature Raman spectra

of the resulting solid mixtures under frozen HF solvent and by

+

intermediate then reacts with Xe₂F₃to form [Xe₃OF₃][AsF₆]

(eq 5), which is stable in HF solvent.

XeF₂+ H₂O h FXeOH + HF

(3)

(4)

(5)

XeF⁺+ H₂O h H₂O---XeF⁺h FXeOH + H⁺

FXeOH + FXeFXeF⁺f FXeOXeFXeF⁺+ HF

To underscore the sensitivity of the synthesis of

[Xe₃OF₃][AsF₆] to reaction conditions and the intermediacies

of FXeOH and/or H₂O---XeF⁺, a dilute solution of XeF₂(0.20

M) and [H₃O][AsF₆] (0.20 M) was allowed to equilibrate at

-64 °C for 12 h. Although the resulting clear, colorless solution

did not show any signs of xenon evolution during this time

period, [Xe₃OF₃][AsF₆] also did not crystallize during this time

interval. Rapid removal of the solvent at -64 °C resulted in a

mixture of [H₃O][AsF₆] · 2XeF₂, [XeF][AsF₆], and [Xe₂F₃]-

[AsF₆]. This behavior is in accordance with our general

observation that H₃O⁺ion is quite resistant to oxidative attack

by strong oxidizers and does not readily participate in hydrolysis

reactions. Thus, it appears that free H₂O is needed to form

FXeOH (eqs 3 and 4). Rapid removal of HF from the

equilibrium mixture apparently did not allow sufﬁcient time to

establish a signiﬁcant equilibrium concentration of H₂O (eq 1)

single-crystal X-ray diffraction and were shown to yield three

+

cationic xenon species, XeF⁺, Xe₂F₃⁺, and Xe₃OF₃

.

(14) Ogden, J. S.; Turner, J. J. Chem. Commun. 1966, 693–694.

(15) Gillespie, R. J.; Schrobilgen, G. J. J. Chem. Soc., Chem. Commun.

1977, 595–597.

(16) Brock, D. S.; Bilir, V.; Mercier, H. P. A.; Schrobilgen, G. J. J. Am.

Chem. Soc. 2007, 129, 3598–3611.

(17) Smith, D. F. Science 1963, 140, 899–900.

(18) Huston, J. L.; Willett, R. D.; Peterson, S. W. J. Chem. Phys. 1973,

59, 453–459.

(19) Huston, J. L.; Claassen, H. H. J. Chem. Phys. 1971, 55, 1505–1507.

(20) Huston, J. L. J. Am. Chem. Soc. 1971, 93, 5255–5256.

(21) Gillespie, R. J.; Schrobilgen, G. J. Inorg. Chem. 1974, 13, 2370–2374.

(22) Gillespie, R. J.; Landa, B.; Schrobilgen, G. J. Inorg. Chem. 1976, 15,

1256–1263.

+

that would be needed to form FXeOH and Xe₃OF₃(eq 5).

The roles of eqs 1 and 2 are also supported by the reaction

of H₂O with a 3-fold molar excess of [XeF][AsF₆] in HF. Even

after 7 months at -78 °C, no Xe₃OF₃⁺had formed. The white

solid that was obtained was shown by Raman spectroscopy to

be a mixture of [Xe₂F₃][AsF₆], [H₃O][AsF₆], and [XeF][AsF₆].

Single-crystal X-ray diffraction also established the presence

(23) Mercier, H. P. A.; Sanders, J. C. P.; Schrobilgen, G. J.; Tsai, S. S.

Inorg. Chem. 1993, 32, 386–393.

(24) Christe, K. O.; Wilson, W. W. Inorg. Chem. 1988, 27, 2714–2718.

(25) Pointner, B. E.; Suontamo, R. J.; Schrobilgen, G. J. Inorg. Chem. 2006,

45, 1517–1534.

(26) Holloway, J. H. J. Fluorine Chem. 1986, 33, 149–157.

9

J. AM. CHEM. SOC. VOL. 131, NO. 37, 2009 13475

A R T I C L E S

Gerken et al.

1

Table 1. Calculated and Experimental H and ¹⁹F Chemical Shifts

of [Xe₂F₃][AsF₆] in its trigonal modiﬁcation.²⁷The formation

of [Xe₃OF₃][AsF₆] was apparently prevented because H₂O is

suppressed and Xe₂F₃⁺and H₃O⁺formation are favored by eq

1 and 2 at high XeF⁺concentrations.

for H₂OF⁺, HOF, H₂NF₂⁺, and HNF₂

chemical shifts, ppm^a

δ(¹H)

exptl

δ(¹⁹F)

(b) Optimized Syntheses of [Xe₃OF₃][PnF₆] (Pn ) As, Sb).

The most efﬁcient syntheses of high-purity [Xe₃OF₃][PnF₆]

involved dissolution of near-equimolar amounts of [H₃O][PnF₆]

and XeF₂(up to ca. 20 mol % excess XeF₂) at ca. 0.2-3 M

H₃O⁺in HF at -50 °C, followed by rapid warming of the

solution to -35 °C for ca. 30 s to effect dissolution of XeF₂,

followed by immediate cooling to -50 °C. These reactions were

shown in the prior discussion to be very slow at -78 °C but

accelerated dramatically at -50 °C. After 5 min at -50 °C, a

deep red-orange voluminous precipitate of [Xe₃OF₃][PnF₆]

formed. Maintenance of the reaction mixture at -50 °C for an

additional 20-30 min ensured that the equilibrium had been

established. Unreacted XeF₂and [H₃O][PnF₆], as well as any

[Xe₂F₃][PnF₆] byproduct, were soluble and were decanted from

the precipitate at -50 °C. Although the amount of XeF₂relative

to [H₃O][PnF₆] was signiﬁcantly less than the ideal 3:1

stoichiometric ratio required by the overall reaction (eq 6), the

1:1 stoichiometry provided, through equilibrium 1, for the

simultaneous presence of XeF₂and H₂O that is required for

FXeOH formation. Decanting of the supernatant from the

precipitated product at -50 °C ensured that the product was

relatively free of unreacted starting materials and [Xe₂F₃][PnF₆].

Above -30 °C, [Xe₃OF₃][PnF₆] decomposed under HF with

Xe gas evolution, as conﬁrmed by ¹²⁹Xe NMR spectroscopy

(vide supra).

compd

solvent

temp, °C

calcd

exptl

calcd

b

HOF

CCl₄/CD₂Cl₂

CH₃CN^c

25

12.8

15.5

12.9

e

11.9 27.5

33.5

-8.5

16.0

e

SO₂ClF^d

-70

-60

g

AsF₅/HF

SbF₅/HOSO₂F

e

H₂OF⁺

HNF₂

H₂NF₂

HF

g

HF

(8.06)^f11.1 (-68.0)^f359.2

7.2^h

10.2 -6.0^h

-11.7

16.8

+

-40

14.2^h

10.7 11.6^h

^aChemical shifts were calculated by the GIAO-MBP2 method using

CCSD(T)/aug-cc-pVDZ optimized geometries. The calculated H and ¹⁹F

1

chemical shifts were obtained by subtracting the calculated isotropic

chemical shift from the calculated isotropic chemical shifts of TMS and

CFCl₃, respectively. ^b5:1 v/v.^{9 c}Reference 9. ^dThis work. ^eThis work;

neither HOF nor H₂OF⁺were observed, most as a result of ¹H and/or

¹⁹F exchange between HOF and HF/HOSO₂F. ^fIncorrect values reported

in ref 10. ^gNo solvent or temperature were provided in the original

work (ref 30). The ¹H and ¹⁹F chemical shifts reported in ref 30 were

recalculated and referenced to CFCl₃in ref 31. ^hReported in ref 31.

spectroscopy showed the formation of [XeF][SbF₆], [Xe₂F₃]-

[SbF₆], and [Xe₃OF₃][SbF₆]; the latter salt was also characterized

by a crystal structure determination (see X-ray Crystallography).

The [XeF][SbF₆] salt is likely formed by oxygen loss from an

unstable OXeF⁺intermediate (eq 12) which, in turn, reacts with

H₂O in HF to generate [Xe₂F₃][SbF₆] and [Xe₃OF₃][SbF₆] (eqs

1-5).

3XeF₂+ [H₃O][PnF₆] f [Xe₃OF₃][PnF₆] + 3HF (6)

[H₃O][SbF₆] + XeF₄h H₂O + [XeF₃][SbF₆] + HF

XeF⁺₃+ H₂O h H₂O· · ·XeF₃⁺h F₂XeOH⁺+ HF

F₂XeOH⁺f OXeF⁺+ HF

(9)

(10)

(11)

(12)

(c) [H₃O][AsF₆]·2XeF₂. Hydrogen-bond formation between the

F-ligands of XeF₂and [H₃O][AsF₆] yields [H₃O][AsF₆]·2XeF₂

(eq 7). The adduct was isolated from HF solution between -78

°C and the freezing point of HF as colorless to very pale orange

crystals (see X-ray Crystallography).

1

[OXeF][SbF₆] f [XeF][SbF₆] + /₂O₂

[H₃O][AsF₆] + 2XeF₂f [H₃O][AsF₆] · 2XeF₂

(7)

(e) NMR Spectroscopy. (i) Attempted Protonation of HOF.

A sample of HOF was prepared by the literature method,²⁹and

its ¹H and ¹⁹F NMR spectra were recorded in SO₂ClF solution

When HF-free solid [H₃O][AsF₆]·2XeF₂was allowed to

warm toward room temperature for very brief time periods,

followed by quenching to -130 °C, the composition progres-

sively shifted toward XeF⁺/H₂O, in accord with equilibrium 1.

Warming [H₃O][AsF₆]·2XeF₂for longer time periods to room

temperature resulted in a transient red-orange solid, presumably

[Xe₃OF₃][AsF₆], which rapidly melted with Xe, O₂, and HF

evolution, leaving behind residues of [XeF][AsF₆] and

[H₃O][AsF₆]. When solid [H₃O][AsF₆]·2XeF₂was allowed to

decompose in a sealed glass tube and the decomposition

products were cooled to -78 °C or below, an intense blue-

green product was obtained that was attributed to the reversible

formation of [Xe₂][AsF₆] from Xe and [XeF][AsF₆] (eq 8).²⁸

at -60 °C. The ¹⁹F (16.0 ppm, J(¹⁹F-¹H) ) 73 Hz) and H

(12.9 ppm, ²J(¹H-¹⁹F) ) 73 Hz) NMR parameters were in close

agreement with the published values (27.5, 12.8 ppm; 80-83

Hz, respectively) in 5:1 CCl₄/CD₂Cl₂solvent (Table 1). Addition

of a 3-fold molar excess of a 1:1 mixture of either AsF₅/HF or

SbF₅/HOSO₂F to SO₂ClF solutions of HOF did not provide any

evidence for the formation of H₂OF⁺by way of signiﬁcant

chemical shift or coupling constant changes (see Computational

Results).

2

1

(ii) Reaction of [XeF][AsF₆] with H₂O in HF. Equimolar

mixtures of [XeF][AsF₆] and H₂O (ca. 1 M each) in HF solution

at -78 °C were found to slowly react to form [H₃O][AsF₆] and

XeF₂(eq 1). Xenon diﬂuoride was identiﬁed at -75 °C by ¹⁹

F

3[XeF][AsF₆] + 3Xe h [Xe₂F₃][AsF₆] + 2[Xe₂][AsF₆] (8)

and ¹²⁹Xe NMR spectroscopy (δ(¹²⁹Xe), -1516.3 ppm; δ(¹⁹F),

(d) XeF₄/H₃O⁺System and Formation of [FXeOXeFXeF]-

[SbF₆]. In the pursuit of other novel xenon oxide ﬂuorides, the

reaction of XeF₄with [H₃O][SbF₆] in HF was explored as a

possible route to the OXeF⁺cation (eqs 9-11). Instead, Raman

-200.68 ppm, ¹J(¹²⁹Xe-¹⁹F), 5670 Hz). In addition to the XeF₂

1

and the HF solvent resonances (-193.66 ppm; ∆ν _/, 38 Hz), a

2

1

broad singlet at -68.5 ppm (∆ν _/, 1560 Hz) corresponding to

2

-

the AsF₆anion was observed in the ¹⁹F NMR spectrum. The

(27) Fir, B. A.; Gerken, M.; Pointner, B. E.; Mercier, H. P. A.; Dixon,

D. A.; Schrobilgen, G. J. J. Fluorine Chem. 2000, 105, 159–167.

(28) Stein, L.; Norris, J. R.; Downs, A. J.; Minihan, A. R. J. Chem. Soc.,

Chem. Commun. 1978, 502–504.

(29) Appelman, E. H.; Jache, A. W. J. Am. Chem. Soc. 1987, 109, 1754–

1757.

(30) Kennedy, A.; Colburn, C. B. J. Am. Chem. Soc. 1959, 81, 2906–2907.

(31) Christe, K. O. Inorg. Chem. 1975, 14, 2821–2824.

9

13476 J. AM. CHEM. SOC. VOL. 131, NO. 37, 2009

Xenon(II) Oxide Fluoride Cation, FXeOXeFXeF⁺

A R T I C L E S

¹⁷O NMR spectrum of a sample prepared from ¹⁷O-enriched

H₂O (¹⁶O, 35.4%; ¹⁷O, 21.9%; ¹⁸O, 42.7%) showed a singlet at

Scheme 1. Products Resulting from the Equilibrium between

XeF⁺/H₂O and XeF₂/H₃O⁺(Pn ) As, Sb)

+

17

1

-0.47 ppm (∆ν _/, 177 Hz), which is assigned to H₃O (δ( O),

2

10.2 ppm; SO₂solvent at -20 °C)³²and was broad, owing to

exchange with HF and residual H₂O, preventing the observation

32

1

of the J(¹⁷O-¹H) coupling (103.5 Hz).

(iii) Reactions of XeF₂with [H₃O][AsF₆] in BrF₅Solution. In

+

an attempt to observe either FXeOH or Xe₃OF₃by solution

¹⁹F NMR spectroscopy and to circumvent possible rapid

chemical exchange with the solvent, XeF₂was allowed to react

with [H₃O][AsF₆] in BrF₅at -55 °C. The initial ¹⁹F NMR

spectrum of a sample of XeF₂and [H₃O][AsF₆] in BrF₅consisted

of resonances corresponding to XeF₂(-183.25 ppm;

¹J(¹²⁹Xe-¹⁹F) ) 5737 Hz), AsF₆^-(-59.71 ppm), and BrF₅(F_ax,

2

273.22 ppm; F_eq, 135.73 ppm; J(¹⁹F-¹⁹F) ) 76 Hz). After

+

several minutes at -55 °C, resonances corresponding to Xe₂F₃

(F_t, -250.05 ppm, F_b, -183.48 ppm; J(¹⁹F-¹⁹F) ) 299 Hz,

2

¹J(¹²⁹Xe-¹⁹F_t) ) 6659 Hz; ¹²⁹Xe satellites of F_boverlapped with

the resonances of XeF₂and HF), HF (-192.33 ppm; ¹J(¹⁹F-¹H)

reported syntheses of [H₂OF][PnF₆] salts from the reactions of

H₂O with [XeF][PnF₆] in anhydrous HF nor the alleged

oxidative oxygenating reactions of such solutions toward

ClF₃^10,11could be duplicated. Rather, the XeF⁺/H₂O and XeF₂/

H₃O⁺systems were demonstrated to be very complex and to

involve, as a minimum, the reactions shown in Scheme 1. The

crystal structures of [H₃O][AsF₆]·2XeF₂and [FXeOXeFXeF]-

[PnF₆] (see X-ray Crystallography) have been key to under-

standing the true nature of the XeF⁺/H₂O and XeF₂/H₃O⁺

systems.

1

) 519 Hz), and BrO₂F (194.35 ppm, ∆ν _/) 1950 Hz) were

2

also observed and increased with time relative to the ¹⁹F

resonance of XeF₂. Equation 13 accounts for the resonances

observed in the NMR spectra. In contrast to the analogous

reaction in HF, where excess HF shifts eq 1 completely to the

side of XeF₂and H₃O⁺, XeF₂and H₃O⁺are partially converted

to XeF⁺, H₂O, and HF. The interaction of XeF₂and XeF⁺

subsequently yields the Xe₂F₃⁺cation, and H₂O reacts with BrF₅

solvent to form BrO₂F and HF. The ¹⁹F NMR signals for HF

In HF solution, mixtures of XeF⁺and H₂O were shown to

quickly establish an equilibrium with XeF₂and H₃O⁺. This

equilibrium was strongly inﬂuenced by the amount of HF

present. With a large excess of HF, the equilibrium is shifted

completely to the side of XeF₂and H₃O⁺. Hydrogen ﬂuoride

plays a 2-fold role in inﬂuencing this equilibrium. There is the

obvious mass effect expected from the Le Chatelier Principle;

however, the main driving force is the vastly different acidities

of monomeric and polymeric HF. Polymeric HF has been shown

to be a much stronger acid than monomeric HF.³⁵Therefore,

[H₃O][HF₂·nHF] can displace XeF₂from its XeF⁺salts, while

XeF₂can displace monomeric HF from its H₃O⁺salts. This

phenomenon has been previously observed in azide systems,

where polymeric, liquid HF can quantitatively displace HN₃

from ionic azides, while HN₃can displace monomeric HF from

solid ionic ﬂuoride salts.³⁵Reactions of both [XeF][PnF₆] with

H₂O, and XeF₂with [H₃O][PnF₆] in HF solution demonstrated

that the reaction proposed in Scheme 1 is a true equilibrium, as

it was established from both sides.

+

and Xe₂F₃broadened with time, presumably because inter-

+

molecular exchange between HF and Xe₂F₃increases with

increasing concentration.

BrF₅

2XeF₂+ 2[H₃O][AsF₆] + BrF₅

8 [Xe₂F₃][AsF₆] +

BrO₂F + 6HF (13)

(f) Reaction of XeF⁺/H₂O with ClF₃in HF Solvent. The prior

claim^10,11that solutions of [XeF][AsF₆] and H₂O in HF

+

oxygenate ClF₃to ClOF₂was re-examined. Under nearly

identical experimental conditions, an equimolar solution of

[XeF][AsF₆] and H₂O in HF solvent was initially allowed to

react at -64 °C for 12 h, followed by the addition of an

equimolar amount of ClF₃at -196 °C. Warming of the reaction

mixture for 2 min to -78 °C yielded a red suspension that

subsequently turned white. In contrast with the earlier claim

+

that ClOF₂was formed, only [ClO₂][AsF₆] was isolated. The

TheformationofXe₂F₃⁺fromXeF⁺andXeF₂iswell-known.^27,36

The co-crystallization of [H₃O][AsF₆] and XeF₂as the hydrogen-

bonded adduct, [H₃O][AsF₆]·2XeF₂, is not surprising, because

XeF₂is known to form hydrogen bonds with the strong acid,

HNO₃.³⁷

present ﬁndings are reminiscent of the reaction of H₂O with

[ClF₂][AsF₆] in HF at -78 °C,³³in which a red-brown color

was initially observed and [ClO₂][AsF₆] and unreacted

[ClF₂][AsF₆] were recovered. The formation of [ClO₂][AsF₆]

in the XeF⁺/H₂O system is not surprising because ClF₃is known

to hydrolyze rapidly to unstable FClO, which dismutates to ClF

and FClO₂. In the presence of AsF₅, FClO₂forms stable

[ClO₂][AsF₆], while the thermally unstable [Cl₂F][AsF₆] byprod-

uct (dissociation vapor pressure ca. 4864 Torr at 25 °C)³⁴was

pumped away under the given conditions.

The remaining products identiﬁed in the XeF⁺/H₂O system

stem from the components of the equilibrium reacting with each

other, as summarized in Scheme 1. The only component of

Scheme 1 that was not directly observed was FXeOH. Its

formation as an intermediate, however, accounts for the forma-

tion of the catenated FXeOXeFXeF⁺cation according to eq 5,

(g) Overview of the XeF⁺/H₂O and XeF₂/H₃O⁺Systems

and Their Reactions. In the present work, neither the previously

(35) Christe, K. O.; Wilson, W. W.; Bau, R.; Bunte, S. W. J. Am. Chem.

Soc. 1992, 114, 3411–3414.

(32) Olah, G. A.; Berrier, A. L.; Prakash, G. K. S. J. Am. Chem. Soc. 1982,

104, 2373–2376.

(36) Sladky, F.; Bulliner, P. A.; Bartlett, N. J. Chem. Soc. A. 1969, 14,

2179–2188.

(33) Christe, K. O. Inorg. Chem. 1972, 11, 1220–1222.

(34) Christe, K. O.; Sawodny, W. Inorg. Chem. 1969, 8, 212–219.

(37) Moran, M. D. Ph.D. Thesis, McMaster University, Hamilton, Ontario,

Canada, 2007.

9

J. AM. CHEM. SOC. VOL. 131, NO. 37, 2009 13477

A R T I C L E S

Gerken et al.

+

Table 2. Summary of Crystal Data and Reﬁnement Results for

[Xe₃OF₃][PnF₆] (Pn ) As, Sb) and [H₃O][AsF₆]·2XeF₂

Table 3. Experimental Geometrical Parameters for the Xe₃OF₃

Cation in [Xe₃OF₃][PnF₆] (Pn ) As, Sb) and Calculated

+

Geometrical Parameters for the Xe₃OF₃Cation

[Xe₃OF₃][AsF₆]

[Xe₃OF₃][SbF₆]

[H₃O][AsF₆]·2XeF₂

AsF₆^-salt^a,b

SbF₆^-salt^b,c

space group

chemical formula

a (Å)

b (Å)

c (Å)

Pc (No. 13) P2₁/c (No. 14) I4/mcm (No. 140)

bond lengths (Å)

1.992(6) Xe(1)-F(1)

2.101(10) Xe(1)-O/F(2)

1.919(9) Xe(2)-O/F(2)

2.502(10) Xe(2)---F/O(2A)

F₉OAsXe₃

6.711(1)

9.115(2)

8.674(2)

96.85(3)

526.8(2)

2

F₉OSbXe₃

6.687(1)

9.306(2)

8.822(2)

96.784(5)

545.1(2)

2

H₃F₁₀OAsXe₂

8.714(4)

12.988(9)

90

Xe(1)-F(1)

1.975(6)

2.085(6)

1.908(6)

2.513(6)

2.085(6)

1.975(6)

Xe(1)-O/F(1)

Xe(2)-O/F(1)

Xe(2)---F/O(2)

Xe(3)-F/O(2)

Xe(3)-F(3)

ꢀ (deg)

2.104(9)

1.977(6)

Xe(1A)-F/O(2A)

Xe(1A)-F(1A)

V (Å³)

986.4(9)

4

Z (molecules/unit cell)

mol wt (g mol^-1

)

655.82

4.135

702.65

4.281

546.54

3.680

bond angles (deg)

calcd density (g cm^-3

)

F(1)-Xe(1)-O/F(1)

Xe(1)-O/F(1)-Xe(2)

177.4(5)

123.5(6)

F(1)-Xe(1)-O/F(2)

Xe(1)-O/F(2)-Xe(2)

O/F(2)-Xe(2)---F/O(2A)

178.6(3)

124.6(3)

177.9(1)

T (°C)

-116

-125

-173

µ (mm^-1

λ (Å)

)

12.81

11.790

0.71073

0.0419

0.0896

10.32

O/F(1)-Xe(2)---F/O(2) 178.3(5)

Xe(2)---F/O(2)-Xe(3)

F/O(2)-Xe(3)-F(3)

123.6(6)

178.2(5)

Xe(2)---F/O(2A)-Xe(1A) 124.6(3)

F/O(2A)-Xe(1A)-F(1A) 178.6(3)

0.71073

0.0460

0.1311

0.71073

0.0208

0.0544

a

R₁

b

wR₂

calcd^d

PBE1PBE

^aR₁is deﬁned as Σ||F_o| - |F_c||/Σ|F_o| for I > 2σ(I). ^bwR₂is deﬁned as

1

2

[Σ[w(F_o- F_c²)²]/Σw(F_o)²] ^/for I > 2σ(I).

SVWN

BP86

B3LYP

B3PW91

MP2

2

bond lengths (Å)

Xe₁-F₁

Xe₁-O₁

O₁-Xe₂

Xe₂---F₂

F₂-Xe₃

Xe₃-F₃

1.965 2.004

1.948

2.158

1.955

2.405

2.108

1.926

1.974

2.189

1.984

2.440

2.133

1.950

1.960

2.169

1.968

2.418

2.118

1.936

1.958

2.151

1.952

2.378

2.111

1.925

which can arise through either partial hydrolysis of XeF₂(eq

3) or deprotonation of the XeF⁺/H₂O donor-acceptor adduct

(eq 4) as the ﬁrst step in this sequence of reactions. The fact

that FXeOH was not directly observed is not surprising, because

compounds containing OH and F ligands attached to the same

central atom are highly reactive and can undergo very facile

intramolecular HF elimination.³⁸

2.139 2.198

1.972 2.014

2.316 2.414

2.119 2.156

1.931 1.971

bond angles (deg)

175.2 174.5

F₁-Xe₁-O₁

175.8

120.8

177.6

170.0

179.7

175.3

121.7

177.4

171.6

179.7

175.5

121.2

177.1

179.9

179.8

176.5

117.1

177.1

177.2

179.9

The formation of the FXeOXeFXeF⁺cation, the red-orange

species erroneously attributed to H₂OF⁺,^10,11was unforeseen.

It constitutes the ﬁrst oxide ﬂuoride of Xe(II) and, along with

[H₃O][AsF₆]·2XeF₂, forms the basis for much of the remaining

structural and theoretical discussion.

X-ray Crystallography. Details of the data collection param-

eters and other crystallographic information for [Xe₃OF₃][PnF₆]

(Pn ) As, Sb) and [H₃O][AsF₆]·2XeF₂are given in Table 2.

Experimental and calculated bond lengths and angles in

[Xe₃OF₃][PnF₆] are provided in Table 3.

Xe₁-O₁-Xe₂119.3 120.4

O₁-Xe₂---F₂

179.6 174.0

Xe₂---F₂-Xe₃151.4 160.9

F₂-Xe₃-F₃178.8 179.0

^aThe As-F distances and F-As-F angles range from 1.64(2) to

1.89(2) Å and 85.8(8)° to 109.2(9)°, respectively. ^bThe complete list of

bond lengths and bond angles can be found in Supporting Information

(cif ﬁles). ^cThe Sb-F distances and F-Sb-F angles range from

1.846(14) to 1.904(11)

Å and 87.5(6)° to 91.6(5)°, respectively.

^daug-cc-pVTZ(-PP).

troscopy and Computational Results). The bridging oxygen and

ﬂuorine positions are disordered, which causes the bridging

Xe(2) position to be split between a longer Xe(2)---F/O and a

shorter Xe(2)-O/F bond. In the case of the SbF₆salt, the

disorder results from a crystallographic inversion center, whereas

(a) [Xe₃OF₃][PnF₆], (Pn ) As, Sb). Single-crystal X-ray

diffraction revealed that the As and Sb salts of [Xe₃OF₃][PnF₆]

crystallize in the monoclinic space groups Pc and P2₁/c,

respectively. Both structures consisted of well-separated but

-

+

-

positionally disordered Xe₃OF₃cations and PnF₆anions

(Figure 1a). The experimental and calculated (SVWN/aug-cc-

-

in the AsF₆salt, the bridging Xe(2) and Xe(2A) atoms are

situated on two crystallographically independent sites. Attempts

to resolve the O/F positions into two components were unsuc-

cessful. Because the O/F occupancy factors are 50/50, extrapo-

lation methods^39,40could not be applied to extract the individual

Xe-O/F bond lengths from these data sets. Thus, quantum-

chemical calculations were relied upon to obtain remaining

structural details (see Computational Results). The two bonding

models, (a) FXeOXe⁺---FXeF and (b) FXeFXe²⁺---OXeF^-, are

not distinguishable solely on the basis of diffraction data.

However, model a, which can be rationalized as the addition

of XeF₂to a FXeOXe⁺fragment, was chosen because model b

would result in an unfavorable charge distribution giving rise

to negatively and double positively charged fragments. The

+

pVTZ(-PP)) geometries of the Xe₃OF₃cation (also see

Computational Results) are depicted in Figure 1b and show

-

generally good agreement. The PnF₆anions are packed in

chains along the a-axis and are located in channels that are

formed by the Xe₃OF₃⁺cations (Figures S1 and S2, Supporting

Information). Substitution of the AsF₆^-anion with SbF₆^-results

in widening of the anion channels and signiﬁcant increases in

the lengths of the b- and c-axes. Because the packing along the

a-axis in both [Xe₃OF₃][PnF₆] salts is dominated by the cations,

the observation of a shorter a-axis in [Xe₃OF₃][SbF₆] is most

likely a thermal effect because the data for the SbF₆^-salt were

acquired at a lower temperature.

+

-

The Xe₃OF₃cations in both PnF₆salts are disordered but

must contain three xenon atoms, two terminal ﬂuorine atoms,

one bridging ﬂuorine atom, and one bridging oxygen atom. The

presence of an oxygen atom in the cation is unambiguously

established by the observed ¹⁸O isotopic shifts in the vibrational

spectra and quantum-chemical calculations (see Raman Spec-

+

Xe₃OF₃cation is nearly planar with a mean deviation from

the molecular plane of ( 0.017 Å. Quantum-chemical calcula-

tions conﬁrmed the assignment to model a, as well as the near-

planarity of the cation (mean deviation, (0.012 Å; see

Computational Results).

(39) Vij, A.; Zhang, X.; Christe, K. O. Inorg. Chem. 2001, 40, 416–417.

(40) Lork, E.; Mews, R.; Viets, D.; Watson, P. G.; Borrmann, H.; Vij, A.;

Boatz, J. A.; Christe, K. O. Inorg. Chem. 2001, 40, 1303–1311.

(38) Christe, K. O.; Hegge, J.; Hoge, B.; Haiges, R. Angew. Chem., Int.

Ed. 2007, 46, 6155–6158.

9

13478 J. AM. CHEM. SOC. VOL. 131, NO. 37, 2009

Xenon(II) Oxide Fluoride Cation, FXeOXeFXeF⁺

A R T I C L E S

Figure 2. Packing diagram of [H₃O][AsF₆]·2XeF₂along the c-axis showing

four O · · ·F contacts. Thermal ellipsoids given at the 50% probability level.

F(3)---O(1) distance is shorter than the F---O distance deter-

mined for XeF₂·HNO₃(2.690(1) Å),³⁷in agreement with the

higher positive charge on H₃O⁺, and is within the sum of the

van der Waals radii for O and F (2.99 Å).⁴²

Figure 1. (a) Disordered structure of [Xe₃OF₃][SbF₆]. Thermal ellipsoids

are given at the 50% probability level. (b) One of the two orientations

obtained from the experimental crystal structure, and the calculated geometry

The AsF₆^-anions are located on special positions (4/m). The

thermal ellipsoid of F(2) in the AsF₆^-anion is elongated in the

a,b-plane, indicating some residual rotational motion along

the coincident F(1)-As(1)-F(1A)- and c-axes. The AsF₆^-anion

exhibits a tetragonal distortion from octahedral symmetry with

the As(1)-F(1) bond (1.722(3) Å) longer than the As(1)-F(2)

bond (1.695(3) Å). While this result is in agreement with the

lowering of the anion symmetry observed by Raman spectros-

copy, the local symmetry of the anion (D_4h) is overestimated

as a result of the high crystal symmetry (D_4h) and, like XeF₂,

shows vibrational bands consistent with a lower symmetry (vide

infra).

+

(SVWN/aug-cc-pVTZ(-PP)) for the Xe₃OF₃cation.

The only bond lengths that are unaffected by the disorder

are those of the terminal Xe-F bonds [As, 1.977(6) and 1.992(6)

Å; Sb, 1.975(6) Å], which are comparable to the Xe-F bond

length in XeF₂(2.00(1) Å)⁴¹and are longer than the terminal

Xe-F bonds in Xe₂F₃⁺(1.929(6) Å, 1.908(7), and 1.908(6) Å).²⁷

The bond angle around the xenon atom also is not affected by

the disorder, with the F-Xe-O/F angles [As, 177.4(5)° and

178.2(5)°; Sb, 178.6(3)°] deviating only slightly from linearity.

-

The PnF₆anions are disordered between two orientations

+

and exhibit Xe---F contacts to the Xe₃OF₃cation that are as

Raman Spectroscopy. (a) [Xe₃OF₃][PnF₆] (Pn ) As, Sb). The

Raman spectra of solid [Xe₃^16,18OF₃][AsF₆] and [Xe₃¹⁶OF₃]-

[SbF₆] were recorded at ca. -160 °C. The Raman frequencies

and intensities along with their assignments are provided in

Table 4. The natural abundance and ¹⁸O-enriched spectra of

[Xe₃OF₃][AsF₆] are given in Figure 3, and the natural abundance

spectrum of [Xe₃OF₃][SbF₆] is given in Figure 4.

short as 3.021-3.114 Å (As) and 2.980-3.087 Å (Sb), which

are signiﬁcantly less than the sum of the van der Waals radii of

xenon and ﬂuorine (3.63 Å).⁴²

(b) [H₃O][AsF₆]·2XeF₂. The compound [H₃O][AsF₆]·2XeF₂

crystallizes in the tetragonal space group I4/mcm and consists

of two XeF₂molecules, one H₃O⁺cation, and one AsF₆^-anion

(Figure 2). The oxygen atom of the H₃O⁺cation is located on

a high-symmetry position (m.mm), which precludes the observa-

tion of the hydrogen atoms in the electron density difference

map. The xenon atom of XeF₂is also located on a high-

symmetry position (..2/m), giving two equivalent ﬂuorine atoms,

preventing the observation of the asymmetry of XeF₂as made

evident by Raman spectroscopy (vide infra). The average

Xe(1)-F(3) bond length of 1.984(3) Å is similar to that found

in the crystal structures of XeF₂(2.00(1) Å)⁴¹and XeF₂·N₂O₄

(1.985(3) Å).³⁷

+

Assuming C_ssymmetry for the free Xe₃OF₃cation (see

X-ray Crystallography and Computational Results), a maximum

of 15 Raman- and infrared-active bands (11A′ + 4A′′) are

-

expected. The Raman spectra of the AsF₆and SbF₆salts,

however, reveal 19 (natural abundance AsF₆salt), 21 (¹⁸O-

-

enriched AsF₆salt), and 19 (natural abundance SbF₆salt)

+

bands that can be attributed to the Xe₃OF₃cation (Table 4).

-

The remaining 16 bands were assigned to the AsF₆anion in

the spectra of the ¹⁶O and ¹⁸O isotopomers, whereas ﬁve bands

were assigned to the SbF₆^-anion in the natural abundance salt.

The H₃O⁺cations and AsF₆^-anions pack in separate chains

parallel to the c-axis and alternating along the a- and b-axes

(Figure 2). The H₃O⁺cation exhibits four short F(3)---O(1)

(2.571(3) Å) contacts to ﬂuorine atoms belonging to four

crystallographically related XeF₂molecules, which indicate the

presence of hydrogen bonding between XeF₂and H₃O⁺. The

-

Overlaps between cation and AsF₆bands were evident from

changes in the relative intensities of these bands in the spectra

of the ¹⁶O- and ¹⁸O-substituted salts.

+

Factor-group analyses were carried out for the Xe₃OF₃

cation, and the AsF₆^-and SbF₆^-anions (symmetries and number

of bands for the SbF₆^-salt appear in square brackets when they

-

differ from that of the AsF₆salt). The factor-group analyses

(41) Argon, P. A.; Begun, G. M.; Levy, H. A.; Mason, A. A.; Jones, C. G.;

Smith, D. F. Science 1963, 139, 842–844.

correlating the free cation (C_s) and anion (O_h) symmetries to

their crystal site symmetries (C₁) [C_i] and to the unit cell

(42) Bondi, A. J. Phys. Chem. 1964, 68, 441–451.

9

J. AM. CHEM. SOC. VOL. 131, NO. 37, 2009 13479

A R T I C L E S

Gerken et al.

+

Table 4. Experimental Raman Vibrational Frequencies and Intensities for the Xe₃OF₃Cation in [Xe₃OF₃][AsF₆], [Xe₃¹⁸OF₃][AsF₆], and

+

[Xe₃OF₃][SbF₆] and Calculated Vibrational Assignments for Xe₃OF₃

^aValues in parentheses denote relative Raman intensities. Symbols denote the following: shoulder (sh), broad (br), and not observed (n.o.). ^bSome

additional bands which intensity proved to vary from sample to sample were observed at 497(12) in the ¹⁶O spectrum and at 503.9(19) and 544.6(31) in

the ¹⁸O spectrum and are tentatively assigned to XeF₂. ^cBands arising from [H₃O][SbF₆] were observed at 173(4), 281(13), 639(2), 672(6), and 681(2)

cm^-1

. .

An additional band arising from XeF₂was also observed at 497(22) cm^-1^dThe atom numbering scheme is as follows:

F₁-Xe₁-O-Xe₂---F₂-Xe₃-F₃⁺. Symbols denote stretch (ν), bend (δ), F_w(wag), F_t(twist), and F_r(rock). The abbreviations denote in-plane (ip) and

e

-

out-of-plane (oop). Only major contributions to the mode descriptions are provided. Mode coincident with a mode associated to the AsF₆anion.

^fAssignments are made under C₁crystal site symmetry (see Table S1) but are correlated to O_hsymmetry in this Table. ^gMode coincident with a mode

associated to the Xe₃OF₃cation. ^hMode coincident with a [H₃O][SbF₆] mode.

+

symmetries (C_s) [C_2h] are provided in Tables S1 and S2 and

show that both the cation and anion modes are split into Raman-

and infrared-active A′ and A′′ [Raman-active A_gand B_g,

infrared-active A_uand B_u] components under C_s[C_2h] unit cell

symmetry, giving the potential to observe a total of 60 and 34

Raman bands for the [Xe₃OF₃][AsF₆] and [Xe₃OF₃][SbF₆] salts,

respectively, if all factor-group splittings were to be observed.

Although not all of these splittings were observed, signiﬁcantly

more bands (35-37 and 24 bands, respectively) are observed

compared to the predicted numbers of 30 and 15, respectively,

if only site-symmetry lowering is operative. The additional

splittings indicate that there is some vibrational coupling within

9

13480 J. AM. CHEM. SOC. VOL. 131, NO. 37, 2009

Xenon(II) Oxide Fluoride Cation, FXeOXeFXeF⁺

A R T I C L E S

stretching mode, ν(Xe₁-O₁) + ν(Xe₂-O₁), with a small

contribution from ν(Xe₂-F₂) - ν(Xe₃-F₂). The bands at 511.9

cm^-1and 539.5, 555.7 cm^-1are assigned to ν(Xe₁-F₁) and

ν(Xe₃-F₃), respectively, with a small contribution from

ν(Xe₁-O₁) - ν(Xe₂-O₁) for the latter at the B3LYP and

B3PW91 levels of theory (Table 5), and show very small oxygen

isotopic dependencies, as expected for essentially pure Xe-F

stretching modes. The band at 479.5 cm^-1is also insensitive to

¹⁸O substitution.

The bands below 260 cm^-1are attributed to bending and

deformation modes. The weak bands at 206.3, 214.4 cm^-1are

the only bands that exhibit signiﬁcant isotopic dependencies

[-5.4, -6.3 cm^-1; calcd (all methods), -1.6 to 0.0, cm^-1] and

are assigned to δ(F₂Xe₃F₃)_oop+ δ(F₁Xe₁O₁)_oop

.

(b) [H₃O][AsF₆]·2XeF₂. The Raman and infrared frequencies

and intensities for [H₃^16/18O][AsF₆]·2XeF₂and [D₃O][AsF₆]·

2XeF₂, along with their assignments, are provided in Table 6,

and the spectrum of [H₃O][AsF₆]·2XeF₂is shown in Figure 5.

The Raman bands of [H₃O][AsF₆]·2XeF₂at 190(5), 470(11),

and 552(100) cm^-1are reminiscent of those observed for free

centrosymmetric XeF₂, which occur at 213.2 (ν₂(Π), infrared),

Figure 3. Raman spectra of [Xe₃OF₃][AsF₆] recorded at -160 °C using

1064-nm excitation for natural abundance (lower trace) and 98.6% ¹⁸O-

enriched (upper trace). Symbols denote FEP sample tube lines (*) and

instrumental artifact (†).

497 (ν₁(Σ_g), Raman), and 555 (ν₃(Σ_u), infrared) cm^-1.⁴¹While

the Raman activity may be explained by a distortion of XeF₂

with a simultaneous loss of the symmetry center, the high Raman

intensity of the 552 cm^-1band and the low intensity of the 470

cm^-1band suggest that the antisymmetric and the symmetric

stretching modes may not be vibrationally strongly coupled.

Similar Raman intensity and frequency patterns are also

observed for metal coordination complexes of XeF₂in which

+

XeF₂is terminally coordinated through one of its ﬂuorine ligands

43

to a metal center such as Cd²⁺

.

The high-frequency bands

above 800 cm^-1(Table 6) are also assigned to the H₃O⁺cation

and are comparable to those observed in the [H₃O][AsF₆] salt.⁴⁴

The disordered crystal structure of [H₃O][AsF₆]·2XeF₂shows

four nearest neighbor XeF₂molecules that are equivalently

hydrogen-bonded to each H₃O⁺cation through a ﬂuorine ligand

(see X-ray Crystallography). However, the structure is disor-

dered and the high crystal site symmetry of XeF₂(..2/m) does

not reveal the lower, precise symmetry of XeF₂that is apparent

from the Raman spectrum, i.e., each H-bonded XeF₂is asym-

metrically bridged to two H₃O⁺cations. For this reason, the

XeF₂modes were assigned under C_∞Vsymmetry, i.e., ν₃(Π),

190; ν₂(Σ⁺), 470; and ν₁(Σ⁺), 552 cm^-1instead of the D_∞h

symmetry suggested by the crystal structure. For similar reasons,

Figure 4. Raman spectra of [Xe₃OF₃][SbF₆] recorded at -160 °C using

1064-nm excitation. Symbols denote FEP sample tube lines (*) and

instrumental artifact (†). The spectrum also contains bands attributable to

[H₃O][SbF₆] and XeF₂.

the unit cell, but for the most part it is relatively weak and the

majority of the observed splittings are the result of site-symmetry

lowering.

+

The frequency assignments for Xe₃OF₃are based on

calculated frequencies and intensities using DFT methods and

are provided in Tables 4 and 5. The frequencies and mode

descriptions associated with the Xe₃OF₃⁺cation are very similar

in both salts; consequently, only the vibrational frequencies of

-

and based on the number of observed Raman bands, the AsF₆

anion modes were assigned under C_4Vsymmetry instead of its

higher crystal site symmetry, D_4h. Aside from three bands at

190, 470, and 552 cm^-1and those above 800 cm^-1(vide infra),

the remaining bands can be conﬁdently assigned to a distorted

-

the ¹⁶O- and ¹⁸O-substituted cations in their AsF₆salts are

considered in the ensuing discussion.

The majority of the Xe₃OF₃⁺vibrational modes are strongly

coupled. Although the predicted degree of coupling varies with

the level of theory, the predominant components in the mode

descriptions remain the same and are therefore used as the basis

for the discussion. Moreover, all observed frequency trends and

^16/18O isotopic frequency shifts are in accordance with the

calculations. The seven bands observed between 418 and 596

cm^-1in the ¹⁶O spectrum are assigned to ﬁve stretching modes,

where two of these bands exhibit a signiﬁcant low-frequency

isotopic shift upon ¹⁸O substitution. The highest frequency band

at 595.8 cm^-1shows a large isotopic shift and splitting (-27.0,

-31.4 cm^-1; calcd (all methods), -29 to -36 cm^-1) and is

assigned to the antisymmetric ν(Xe₁-O₁) - ν(Xe₂-O₁) stretch-

ing mode. The bands at 418.7, 429.8 cm^-1also show very large

isotopic shifts [-24.8 and -27.8 cm^-1, respectively; calcd (all

methods), -16 to -24 cm^-1] and are assigned to the symmetric

-

AsF₆anion under C_4Vsymmetry.

To further establish whether asymmetric coordination of XeF₂

to H₃O⁺accounts for the observed Raman spectrum of

[H₃O][AsF₆]·2XeF₂, four H₃O⁺·nXeF₂adducts (n ) 1-4) were

calculated at the PBE1PBE/aug-cc-pVTZ(-PP) level (Figure 6

and Table S5); the choice of n ) 4 for the maximum number

of XeF₂molecules coordinated to H₃O⁺was dictated by the

local environment of H₃O⁺in the crystal structure of

[H₃O][AsF₆]·2XeF₂(also see Computational Results), which

showed that the H₃O⁺cation was coordinated to four neighbor-

ing XeF₂molecules through ﬂuorine (Figure 2).

ˇ

(43) Tavcar, G.; Benkicˇ, P.; Zemva, B. Inorg. Chem. 2004, 43, 1452–1457.

(44) Christe, K. O.; Schack, C. J.; Wilson, R. D. Inorg. Chem. 1975, 14,

2224–2230.

9

J. AM. CHEM. SOC. VOL. 131, NO. 37, 2009 13481

A R T I C L E S

Table 5. Calculated^{a b}Vibrational Frequencies and Infrared and Raman Intensities for Xe₃OF₃

Gerken et al.

,

+

assignment

(PBE1PBE)^c

SVWN

BP86

PBE1PBE

B3LYP

B3PW91

MP2

+

Xe₃¹⁶OF₃

604.3(351)[102] 557.5(256)[126] 612.4(55)[159] 580.2(2)[175]

560.8(181)[222] 522.1(356)[215] 581.4(120)[231] 549.8(139)[213] 566.2(129)[218] 556.4(35)[175] ν(Xe₁-F₁)

588.0(300)[92] 534.0(322)[20] 626.9(276)[33] 587.1(392)[3] 605.9(365)[8] 656.9(40)[122] ν(Xe₁-O₁) - ν(Xe₂-O₁)

597.8(6)[178]

599.6(127)[131] ν(Xe₃-F₃) + [ν(Xe₁-O₁) - ν(Xe₂-O₁)]_small

441.5(134)[506] 408.8(165)[476] 435.7(172)[509] 419.8(220)[589] 428.4(194)[558] 431.7(299)[955] [ν(Xe₃-F₂) - ν(Xe₂-F₂)] + [ν(Xe₁-O₁) +

ν(Xe₂-O₁)]

420.0(35)[76]

381.3(36)[97]

409.5(21)[350] 390.5(13)[216] 400.7(18)[260] 450.0(63)[43]

[ν(Xe₁-O₁) + ν(Xe₂-O₁)] + [ν(Xe₂-F₂) -

ν(Xe₃-F₂)]

δ(F₂Xe₃F₃)_ip

219.3(1)[14]

184.3(3)[11]

196.3(<1)[12] 186.9(<1)[11] 187.7(<1)[11] 191.1(<1)[12]

184.3(<0.1)[19] 169.0(<0.1)[19] 195.2(<0.1)[22] 185.6(<0.1)[21] 190.3(<0.1)[21] 195.1(<0.1)[26] δ(F₂Xe₃F₃)_oop+ δ(F₁Xe₁O₁)_oop

176.2(2)[1]

174.1(2)[2]

128.6(3)[8]

100.4(3)[<1]

99.9(3)[<1]

40.8(11)[<1]

19.6(<1)[<1]

18.7(<1)[<0.1]

162.8(2)[<0.1]

159.0(3)[<1]

112.1(6)[9]

90.2(<1)[<1]

78.1(6)[2]

45.1(2)[<1]

12.3(4)[<1]

3.2(<1)[<1]

186.3(2)[<]

178.8(4)[1]

109.9(4)[14]

96.9(<1)[<1]

92.9(3)[6]

175.8(2)[2]

169.9(4)[1]

105.8(4)[12]

92.2(<1)[<1]

87.8(3)[6]

181.2(2)[2]

174.0(4)[<0.1]

107.7(4)[11]

94.7(<1)[<1]

83.9(2)[7]

184.3(1)[<1]

181.0(9)[<1]

113.4(4)[14]

101.1(<1)[<1]

95.0(2)[6]

δ(F₂Xe₃F₃)_oop- δ(F₁Xe₁O₁)_oop

δ(F₁Xe₁O₁)_ip

F_r(F₁Xe₁O₁)_ip+ ν(Xe₂-F₂)

F_t(F₁Xe₁O₁)_oop- F_t(F₂Xe₃F₃)_oop

δ(Xe₂O₁Xe₁)_ip- F_r(F₂Xe₃F₃)_ip

δ(Xe₂O₁Xe₁)_ip+ F_r(F₂Xe₃F₃)_ip

F_t(F₁Xe₁O₁Xe₂)_oop- F_t(F₂Xe₃F₃)_oop

F_r(F₁Xe₁O₁Xe₂)_ip- F_r(F₂Xe₃F₃)_ip

43.1(3)[1]

42.6(4)[1]

43.9(3)[1]

51.1(2)[1]

5.5(<1)[<0.11]

15.0(<1)[<1]

13.2(1)[<0.1]

14.6(1)[<1]

19.0(<1)[<0.1]

11.4(2)[<0.1]

16.9(2)[<1]

29.4(2)[<0.1]

+

Xe₃¹⁸OF₃

604.3(238)[139] 558.1(209)[139] 615.8(171)[99] 584.1(181)[108] 600.7(173)[107] 609.4(107)[151] ν(Xe₃-F₃) + [ν(Xe₁-O₁) - ν(Xe₂-O₁)]_small

566.3(31)[48] 525.9(48)[112] 579.4(180)[281] 547.1(257)[262] 563.4(229)[272] 567.7(59)[197] ν(Xe₁-F₁)

552.5(505)[218] 503.9(603)[103] 595.2(70)[25] 557.7(61)[9] 575.5(65)[11] 632.5(43)[61] ν(Xe₁-O₁) - ν(Xe₂-O₁)

441.3(148)[544] 408.0(188)[522] 434.8(197)[694] 418.4(233)[686] 425.9(210)[686] 435.0(241)[993] [ν(Xe₃-F₂) - ν(Xe₂-F₂)] + [ν(Xe₁-O₁) +

ν(Xe₂-O₁)]

399.6(19)[39]

363.5(20)[52]

391.2(5)[168]

372.9(5)[119]

382.9(6)[135]

419.2(46)[31]

[ν(Xe₁-O₁) + ν(Xe₂-O₁)] + [ν(Xe₂-F₂) -

ν(Xe₃-F₂)]

δ(F₂Xe₃F₃)_ip

220.0(1)[14]

182.7(<0.1)[17] 167.6(<1)[17]

184.9(3)[11]

193.0(<1)[11]

195.2(<1)[20]

183.7(2)[4]

179.6(4)[<0.1]

110.1(4)[12]

96.1(<1)[<1]

87.3(2)[7]

44.8(3)[1]

18.7(<1)[<0.1]

11.0(2)[<0.1]

187.3(<1)[11]

185.3(<1)[19]

172.9(2)[5]

171.0(4)[1]

106.2(4)[12]

91.4(<1)[<1]

88.3(3)[6]

42.7(4)[1]

13.1(1)[<0.1]

14.6(1)[<1]

187.7(<1)[11]

189.7(<1)[18]

178.4(2)[4]

175.2(4)[<1]

108.2(4)[11]

94.1(<1)[<1]

85.1(2)[7]

43.3(3)[1]

18.7(<1)[<0.1]

10.4(2)[<1]

202.4(<1)[11]

205.1(<0.1)[21] δ(F₂Xe₃F₃)_oop+ δ(F₁Xe₁O₁)_oop

173.5(2)[1]

175.8(2)[3]

129.2(3)[9]

100.8(5)[<1]

99.5(<1)[<1]

41.1(12)[<1]

20.3(<1)[<1]

19.2(<1)[<0.1]

161.0(2)[1]

159.9(3)[<1]

112.5(5)[9]

89.3(<1)[<1]

78.7(6)[2]

45.2(2)[<1]

12.4(5)[<1]

3.1(<1)[<0.1]

190.2(1)[3]

185.7(6)[<1]

115.5(4)[13]

105.2(<1)[<1]

96.8(2)[6]

52.2(2)[<1]

31.8(1)[<0.1]

17.6(2)[<1]

δ(F₂Xe₃F₃)_oop- δ(F₁Xe₁O₁)_oop

δ(F₁Xe₁O₁)_ip

F_r(F₁Xe₁O₁)_ip+ ν(Xe₂-F₂)

F_t(F₁Xe₁O₁)_oop- F_t(F₂Xe₃F₃)_oop

δ(Xe₂O₁Xe₁)_ip- F_r(F₂Xe₃F₃)_ip

δ(Xe₂O₁Xe₁)_ip+ F_r(F₂Xe₃F₃)_ip

F_t(F₁Xe₁O₁Xe₂)_oop- F_t(F₂Xe₃F₃)_oop

F_r(F₁Xe₁O₁Xe₂)_ip- F_r(F₂Xe₃F₃)_ip

^aThe aug-cc-pVTZ(-PP) basis set was used. The aug-cc-pVDZ(-PP) values are reported in Table S3. ^bCalculated Raman [infrared] intensities, in

units of Å⁴amu^-1[km mol^-1] appear in parentheses [square brackets]. ^cThe atom numbering is as follows: F₁-Xe₁O-Xe₂---F₂-Xe₃-F₃. Symbols

denote stretch (ν), bend (δ), F_w(wag), F_t(twist), and F_r(rock). The abbreviations denote in-plane (ip) and out-of-plane (oop). Only major contributions

to the mode descriptions are provided.

Table 6. Observed Vibrational Frequencies^afor [H₃O][AsF₆]·2XeF₂, [D₃O][AsF₆]·2XeF₂, and [H₃¹⁸O][AsF₆]·2XeF₂and Their Assignments

^aFrequencies are in cm^-1^bValues in parentheses denote relative Raman intensities. ^cSymbols denote shoulder (sh) and broad (br). ^dSymbols denote

.

medium (m), weak (w), very strong (vs), and strong (s).

In H₃O⁺·XeF₂, the ν(XeF_term) and ν(XeF_bridge) (Figure 6)

coupling is mostly interligand, occurring among ν(XeF_term)-type

or ν(XeF_bridge)-type modes of coordinated XeF₂molecules. In

modes are coupled, whereas in H₃O⁺·nXeF₂(n ) 2-4), the

9

13482 J. AM. CHEM. SOC. VOL. 131, NO. 37, 2009

Xenon(II) Oxide Fluoride Cation, FXeOXeFXeF⁺

A R T I C L E S

B3LYP, B3PW1 and MPW1PW91) with the aug-cc-pVDZ-

(-PP) and aug-cc-pVTZ(-PP) basis sets.⁴⁵Møller-Plesset (MP2)

and coupled-cluster (CCSD(T)) methods gave a near-linear bond

angle for the Xe-F---Xe moiety, which was also observed for

Xe₂F₃⁺at the Hartree-Fock level.²⁷All optimizations resulted

in stationary points with all frequencies real (Tables 3 and 5

and Tables S3 and S4; also see Experimental Section). The

H₃O⁺·nXeF₂(n ) 1-4) adducts were optimized at the

PBE1PBE/aug-cc-pVTZ(-PP) level of theory. The H₂OF⁺cation

was optimized using the CCSD(T) method and its parent neutral

species, HOF, has been optimized using the SVWN, PBE1PBE,

MP2, and CCSD(T) methods. The calculated vibrational fre-

quencies and infrared intensities show that the infrared bands

previously attributed to the H₂OF⁺cation¹⁰are incompatible

with our predictions (Tables S7 and S8). The intermediates,

FXeOH and H₂OXeF⁺, proposed in eqs 3 and 4, were also

optimized using PBE1PBE and CCSD(T) methods (Table S9).

The geometry and vibrational frequencies of XeF₂were

calculated at various levels of theory to serve as a benchmark

for H₃O⁺·nXeF₂(Table S10). It was concluded that the

PBE1PBE/aug-cc-pVTZ(-PP) level gives results that are similar

to those of the “expensive” CCSD(T)/aug-cc-pVTZ level of

theory, and can thus be used to reliably compute the

H₃O⁺·nXeF₂systems dealt with in this work.

Figure 5. Raman spectrum of [H₃O][AsF₆]·2XeF₂recorded at -145 °C

using 1064-nm excitation. Symbols denote FEP sample tube lines (*) and

instrumental artifact (†).

all cases, there is some intramolecular coupling with the

associated ν(F_bridge---H_bridge) modes. As the number of XeF₂

coordinated molecules increases, the frequencies arising from

the ν(XeF_term) modes progressively shift to lower frequencies

(616, 617 cm^-1for n ) 2; 605, 605, 610 cm^-1for n ) 3; and

581, 599, 604, 608 cm^-1for n ) 4) and those involving

ν(XeF_bridge) modes shift to higher frequencies (423, 496 cm^-1

for n ) 2; 453, 455, 507 cm^-1for n ) 3; 456, 493, 510 cm^-1

for n ) 4). Overall, the frequency differences between the

ν(XeF_term) and ν(XeF_bridge) modes decrease as n increases. For

n ) 4, these differences are in good agreement with what is

observed experimentally (470 and 552 cm^-1). Moreover, the

calculated intensities account for the high intensity of the high-

frequency band of XeF₂.

The band at 190 cm^-1is assigned to a superposition of

F_termXeF_bridge-type bending modes and occurs in a region similar

to that of free XeF₂(calcd., 214 cm^-1). It is worth noting that

this mode is calculated at 214 cm^-1for free XeF₂and between

178 and 245 cm^-1for the H₃O⁺adducts. In support of these

ﬁndings, the infrared spectrum of [H₃O][AsF₆]·2XeF₂showed

bands having the same frequencies as the Raman bands but with

reversed relative intensities. The assignments of the 190, 470,

and 552 cm^-1bands to the XeF₂moiety of the adduct was

further substantiated by the Raman spectra of [D₃O][AsF₆] ·

2XeF₂and [H₃¹⁸O][AsF₆] · 2XeF₂, which exhibited only H/D

isotopic shifts for the bands above 800 cm^-1, conﬁrming their

assignments to H₃O⁺, whereas the XeF₂and anion bands were

unshifted (Table 6).

(a) Xe₃OF₃⁺. (i) Geometries. The minimum energy structure

+

of Xe₃OF₃was found to possess C_ssymmetry with all

frequencies real (Table 3 and Table S4). Because many of the

bond lengths and angles of the [Xe₃OF₃][PnF₆] (Pn ) As, Sb)

salts in their crystal structures cannot be commented on because

of the positional disorder, the calculated structure was relied

upon to provide the exact geometry. All methods and basis sets

used resulted in similar trends. The two terminal bond lengths

(Xe₁-F₁and Xe₃-F₃) are shorter than the bridge bond lengths

(Xe₃-F₂and Xe₂---F₂), in agreement with those observed for

the Xe₂F₃⁺cation.²⁷The short Xe₂-O₁bond and long Xe₂---F₂

contact support the FXeOXe⁺---FXeF model proposed for the

disordered crystal structures (see X-ray Crystallography). The

similar bond lengths calculated for Xe₁-O₁and Xe₃-F₂likely

account for why it was not possible to split the O/F positions

in the crystal structures. The angles about Xe₁, Xe₂, and Xe₃

were found to be slightly less than 180° at all levels of theory,

in agreement with the crystal structure and the structure of the

Xe₂F₃cation.²⁷While the range of the Xe₁-O-Xe₂angles

+

was relatively narrow (116.5 to 121.7°) at all levels of theory,

a wider range of values was obtained for the more deformable

Xe₂---F₂-Xe₃angle (142.7-179.9°). Because the structure is

disordered, no conclusive statement can be made about which

method best reproduces this angle, although DFT (SVWN and

BP86) was the only method which gave a signiﬁcantly bent

angle.

A second, more simplistic, model was used in which the

frequencies and infrared intensities for XeF₂were obtained using

the B3LYP/aug-cc-pVTZ(-PP) method with the constraint that

the two Xe-F bond lengths differed by 0.05, 0.10, and 0.15 Å

(Table S6). It was found that with increasing bond length

difference, the frequency differences between the two stretching

modes also signiﬁcantly increased and the higher frequency

mode became the most intense Raman band, in accordance with

our experimental observations. On the basis of the observed

frequency difference of 82 cm^-1between the experimental 470

and 552 cm^-1bands, a Xe-F bond length difference of 0.07 Å

was predicted for [H₃O][AsF₆]·2XeF₂; however, this cannot be

observed in the present structure due to disorder.

Attempts were also made to optimize the (FXe)₃O⁺isomer

of FXeOXeFXeF⁺starting from a planar D_3hstructure and from

a pyramidal C_3Vstructure with an initial Xe-O-Xe angle of

108.5°. In both cases the geometries failed to optimize but the

ﬁnal geometries were very close to planar.

(ii) Gas-Phase XeF⁺Afﬁnities. The XeF⁺afﬁnities of XeF₂,

+

HF, H₂O, and FXeOXeF that yield HFXeF⁺, H₂OXeF⁺, Xe₂F₃,

and Xe₃OF₃⁺, respectively, were calculated at the PBE1PBE/

(45) Basis sets were obtained from the Extensible Computational Chemistry

Environment Basis Set Database, version 2/25/04, as developed and

distributed by the Molecular Science Computing Facility, Environ-

mental and Molecular Science Laboratory, which is part of the Paciﬁc

Northwest Laboratory, P.O. Box 999, Richland, WA 99352.

+

Computational Results. The electronic structures of Xe₃OF₃

were optimized using DFT methods (SVWN, BP86, PBE1PBE,

9

J. AM. CHEM. SOC. VOL. 131, NO. 37, 2009 13483

A R T I C L E S

Gerken et al.

Figure 6. Calculted geometries for H₃O⁺·nXeF₂(n ) 1-4). The atoms labeled F_t, F_t′, F_t′′, F_t′′′ and F_b, F_b′, F_b′′, F_b′′′ are referred to in the discussion as F_term

and F_bridge, respectively.

Table 7. Calculated^aNatural Charges, Valencies, and Bond Orders for FXeOXeFXeF⁺

SVWN

BP86

PBE1PBE

B3LYP

B3PW91

MPW1PW91

natural charges [valencies]

F₁

Xe₁

O₁

Xe₂

F₂

Xe₃

F₃

-0.511 [0.362]

1.153 [0.707]

-0.855 [0.909]

1.095 [0.647]

-0.624 [0.341]

1.218 [0.599]

-0.477 [0.394]

-0.491 [0.341]

1.125 [0.656]

-0.802 [0.860]

1.063 [0.600]

-0.625 [0.303]

1.197 [0.564]

-0.466 [0.371]

-0.528 [0.358]

1.210 [0.659]

-0.887 [0.881]

1.129 [0.617]

-0.686 [0.293]

1.269 [0.565]

-0.508 [0.376]

-0.520 [0.343]

1.195 [0.639]

-0.859 [0.863]

1.109 [0.601]

-0.679 [0.280]

1.258 [0.543]

-0.503 [0.362]

-0.520 [0.353]

1.195 [0.654]

-0.868 [0.874]

1.116 [0.611]

-0.679 [0.292]

1.257 [0.561]

-0.501 [0.373]

-0.527 [0.356]

1.211 [0.655]

-0.884 [0.876]

1.126 [0.613]

-0.687 [0.292]

1.270 [0.562]

-0.508 [0.374]

bond order

0.345

0.321

0.545

0.077

Xe₁-F₁

Xe₁-O₁

O₁-Xe₂

Xe₂-F₂

F₂-Xe₃

Xe₃-F₃

0.352

0.358

0.537

0.109

0.214

0.383

0.331

0.330

0.516

0.085

0.202

0.360

0.332

0.316

0.534

0.073

0.191

0.349

0.341

0.321

0.539

0.078

0.198

0.360

0.343

0.319

0.542

0.076

0.198

0.361

0.199

0.363

^aaug-cc-pVTZ(-PP).

aug-cc-pVTZ(-PP) level of theory (Table S11). The standard

heats and Gibbs free energies (kJ mol^-1) for the reactions of

XeF⁺with HF (∆H°, -48.35; ∆G°, -81.45), H₂O (∆H°,

-118.75; ∆G°, -156.46), XeF₂(∆H°, -122.83; ∆G°, -155.97),

and FXeOXeF (∆H°, -206.74; ∆G°, -238.40) are all exother-

mic and spontaneous, with O(XeF)₂exhibiting the greatest

afﬁnity for XeF⁺. Xenon diﬂuoride and water have comparable

values, which are intermediate in the series, with HF exhibiting

the lowest afﬁnity.

(iii) Natural Bond Orbital (NBO) Analysis. The NBO analyses

(Table 7) were carried out for all DFT-optimized gas-phase

geometries and provide further support for the FXeOXe⁺---

FXeF model. Similar trends in calculated valencies, bond orders,

and charges were found at all levels, but only the PBE1PBE

values are considered in the ensuing discussion.

All of the calculated bond orders are signiﬁcantly less than

1 and are consistent with the highly polar natures of these bonds.

There is a strong covalent interaction between Xe₂and O₁

9

13484 J. AM. CHEM. SOC. VOL. 131, NO. 37, 2009

Xenon(II) Oxide Fluoride Cation, FXeOXeFXeF⁺

A R T I C L E S

H₂NF₂,

³¹and HNF₂.³⁰Again, it can be seen that the ¹⁹F value

+

(0.545) and a very weak bonding interaction between Xe₂and

F₂(0.077), which supports the choice of FXeOXe⁺---FXeF as

the dominant resonance structure (see X-ray Crystallography).

Moreover, the group charges, FXeOXe⁺(0.924), FXeF (0.075)

and FXeO^-(-0.205), ²⁺XeFXeF (1.204), also support the

former bonding description.

previously attributed¹⁰to H₂OF⁺is in marked disagreement with

calculated values. The reported value differs by almost 440 ppm

from the predicted value and is unlikely to arise from H₂OF⁺.

In contrast, agreement between calculated and observed values

+

for HOF, H₂NF₂, and HNF₂is ca. (6 ppm. This again supports

(b) H₃O⁺·nXeF₂(n ) 1-4). To support the vibrational

assignments for [H₃O][AsF₆] · 2XeF₂, four H₃O⁺· nXeF₂ad-

ducts (n ) 1-4) were calculated at the PBE1PBE/aug-cc-

pVTZ(-PP) level (Figure 6 and Table S5). The H₃O⁺·XeF₂and

H₃O⁺·2XeF₂adducts both optimized under C_ssymmetry,

whereas H₃O⁺·3XeF₂and H₃O⁺·4XeF₂both optimized under

C₁symmetry. Coordination of XeF₂to H₃O⁺gives rise to a

slight asymmetry in the Xe-F bond lengths, the longer bond

being that in which its ﬂuorine atom hydrogen bonds to the

H₃O⁺cation. The Xe-F_termbond lengths increase as the number

of coordinated XeF₂molecules increases, while the Xe-F_bridge

bond lengths decrease. Therefore, the asymmetry, which is

measured by the bond length difference, d(Xe-F_bridge) -

d(Xe-F_term), decreases as n increases. Correspondingly, the

the experimental evidence presented above that the previously

published report of the H₂OF⁺cation is erroneous. The broad

¹⁹F NMR signals at -57.8 and -68.0 ppm were attributed to

AsF₆^-and H₂OF⁺, respectively.¹⁰The latter resonance actually

arises from AsF₆^-, which occurs at -68.5 ppm (this work).

(d) FXeOH and H₂OXeF⁺. The geometries and harmonic

frequencies for FXe¹⁶OH, FXe¹⁸OH, FXe¹⁶OD, H₂¹⁶OXeF⁺,

H₂¹⁸OXeF⁺, and D₂¹⁶OXeF⁺were calculated using the PBE1PBE

and CCSD(T) (values in square brackets) methods and the aug-

cc-PVTZ(-PP) basis set, resulting in optimized C_sgeometries

having all frequencies real (Table S9). The calculations show

that FXeOH and H₂OXeF⁺are indeed viable species and could

therefore be possible intermediates in Scheme 1.

The calculated geometry of the H₂OXeF⁺cation suggests that

XeF⁺forms an oxygen-bonded, donor-acceptor adduct with

water, having an Xe---O bond length of 2.320 [2.344] Å, an

Xe-F bond length of 1.899 [1.905] Å, and an F-Xe-O angle

close to 180° (178.4 [178.3]°). The Xe-F bond length is

signiﬁcantly shorter than that of XeF₂(1.986 [1.991] Å, Table

S10; 2.00(1) Å⁴¹but longer than that of XeF⁺(1.84(1) Å in

[XeF][Sb₂F₁₁]⁴⁶). The near-linear F-Xe-O angle (177.9 [177.8]°)

of FXeOH is also near linear with Xe-O (2.038 [2.062] Å)

and Xe-F (2.029 [2.036] Å) bond lengths that are shorter and

longer, respectively, than in H₂OXeF⁺.

difference in average frequencies, ν(XeF_term)_av- ν(XeF_bridge)_av,

also decreases with increasing n. The F_bridge---H_bridgedistances

range from 1.033 Å (n ) 2) to 1.543 Å (n ) 4) and indicate

that the contacts between the XeF₂molecules and the H₃O⁺

cation are signiﬁcant. A second approach, which ﬁxed the Xe-F

distances of XeF₂by contracting one Xe-F bond and elongating

the other by the same amount, resulted in calculated frequencies

that showed a similar trend (Table S6).

(c) H₂OF⁺. A series of high-level theoretical calculations using

DFT and ab initio (MP2 and CCSD(T)) methods (Table S7)

were run for both HOF and DOF to serve as benchmarks. The

best agreement between observed and calculated geometries was

obtained at the CCSD(T)/aug-cc-pVTZ level. This level also

proved to be good for reproducing various trends among

frequencies as well as isotopic shifts and was consequently used

for the study of the H₂OF⁺cation. It is worth noting that the

intensity trends were reproduced at all levels.

Conclusion

The reactions of [XeF][PnF₆] (Pn ) As, Sb) with H₂O in HF

solution depend strongly on the conditions and result in the

formation of two new xenon species, [Xe₃OF₃][PnF₆] (Pn )

As, Sb) and [H₃O][AsF₆]·2XeF₂. The Xe₃OF₃⁺cation represents

the ﬁrst example of a Xe(II) oxide ﬂuoride, and its formation is

strong evidence for the formation of FXeOH as an unstable

intermediate in the hydrolysis of XeF₂. The [Xe₃OF₃][PnF₆] salts

have been characterized by Raman spectroscopy and single-

crystal X-ray diffraction, establishing the presence of the

(i) Vibrational Spectra of H₂OF⁺. The geometries and

vibrational frequencies of H₂¹⁶OF⁺, D₂¹⁶OF⁺, H₂¹⁸OF⁺, and

D₂¹⁸OF⁺were calculated at the CCSD(T)/aug-cc-pVTZ level

(Table S8) and do not match the reported spectra in several

key respects: (1) most of the intensities do not ﬁt; for example,

ν₃(A′) is predicted to be a strong band but is either observed

only as a weak band or is not observed; ν₄(A′) is predicted to

be weak but was reported as strong, and ν₆(A′′) is calculated to

have zero intensity and was reported to be of medium intensity;

(2) the observed and calculated isotopic shifts are very different

for ν₂(A′), ν₄(A′), and ν₅(A”) of D₂¹⁶OF⁺and ν₂(A′) and ν₄(A′)

of D₂¹⁸OF⁺; and (3) in the previously reported infrared spectra,

ν₅(A′′), which was assigned as ν_as(OX₂) (X ) H, D), was shown

to be a very intense band, yet it is not observed in the

experimental vibrational spectra of the alleged H₂¹⁸OF⁺and

D₂¹⁸OF⁺cations. In summary, there can be no doubt that the

previously reported spectra cannot arise from the H₂OF⁺cation.

The absence of the strong infrared bands around 470 and 552

cm^-1for the XeF₂moieties of [H₃O][AsF₆]·2XeF₂also rules

out the presence of this adduct in the previously reported

spectra.¹⁰

+

Z-shaped Xe₃OF₃cation. The [H₃O][AsF₆]·2XeF₂adduct and

its D- and ¹⁸O- isotopically enriched counterparts have also been

characterized by Raman and infrared spectroscopies and single-

crystal X-ray diffraction. The geometries and vibrational assign-

ments for the Xe₃^16/18OF₃⁺cation and the [H/D₃^16/18O]-

[AsF₆]·2XeF₂adduct are supported by quantum-chemical cal-

culations.

The present study failed to provide evidence for the presence

of the previously claimed H₂OF⁺cation in solutions of XeF⁺

and H₂O in HF, the oxygenation of ClF₃by these solutions,

and the protonation of HOF in HF/AsF₅or HOSO₂F/SbF₅. The

plausibility of the ¹⁹F chemical shift, previously attributed to

H₂OF⁺, is also rendered highly questionable by the results of

quantum-chemical calculations, which show very poor agree-

ment between the predicted and the previously reported values.

Experimental Section

1

(ii) H and ¹⁹F NMR Chemical Shifts for H₂OF⁺. Chemical

+

shifts were also calculated for H₂OF⁺, HOF, H₂NF₂, and HNF₂

Apparatus and Materials. All manipulations involving air-

sensitive materials were carried out under strictly anhydrous

by the GIAO method using CCSD(T)/aug-cc-pVDZ optimized

geometries. The results are summarized in Table 1 and are

compared with the reported values for H₂OF⁺,¹⁰HOF,⁹

(46) Bartlett, N. EndeaVour 1972, 31, 107–112.

9

J. AM. CHEM. SOC. VOL. 131, NO. 37, 2009 13485

A R T I C L E S

Gerken et al.

conditions as previously described.⁴⁷Volatile materials were

handled on vacuum lines constructed of nickel, stainless steel, and

FEP. Nonvolatile materials were handled in the atmosphere of a

drybox or a glovebag. Reaction vessels/Raman sample tubes and

NMR sample tubes were fabricated from 0.25-in. o.d. and 4-mm

o.d. FEP tubing, respectively, and outﬁtted with Kel-F or 316

stainless steel valves. All reaction vessels and sample tubes were

rigorously dried under dynamic vacuum prior to passivation with

ClF₃or 1 atm of F₂gas.

Commercially available D₂O (99.9% D, MSD Isotopes), enriched

H₂O (¹⁶O, 35.4%; ¹⁷O, 21.9%; ¹⁸O, 42.7%; Ofﬁce de Rayonnements

Ionisants, Saclay, France), H₂¹⁸O (Isotec, 98.6% ¹⁸O), DF, AsF₅

and BiF₅(Ozark-Mahoning) were used as received. Antimony

pentaﬂuoride, SbF₅, (Ozark-Mahoning Co.) and ClF₃(Matheson)

were puriﬁed by fractional condensation prior to use. Literature

methods were used for the puriﬁcation and/or syntheses of XeF₂,²³

XeF₄,⁴⁸[XeF][AsF₆],⁴⁹[H₃O][AsF₆],⁴⁴[H₃O][SbF₆],⁴⁴AsF₅,²³BrF₅

(Matheson and Ozark-Mahoning),⁵⁰and HF (Harshaw Chemi-

cals).⁵⁰

Synthesis and Attempted Protonation of HOF. The preparation

of HOF was carried out using a modiﬁcation of the procedure

published by Appelman and Jache.²⁹A recirculating loop was built,

consisting of three 0.5-in. o.d. FEP U-traps and a stainless-steel

bellows pump (Model MB-21, Metal Bellows Corp.). The ﬁrst

U-trap (the reactor) was maintained at -45 °C and was ﬁlled with

Teﬂon Raschig rings, wetted with 2 mL of distilled water. The

second U-trap was maintained at -78 °C to trap unreacted H₂O

and HF, while the third U-trap was maintained at -196 °C to trap

the desired HOF and any OF₂byproduct. The -196 °C trap could

be closed by Teﬂon valves and was connected to a pressure gauge,

which was protected by a Teﬂon diaphragm, a Teﬂon infrared cell

with BaF₂windows, and a third connector for withdrawing HOF

samples. Any OF₂byproduct could be taken off either as a fore

run at low temperatures or by pumping at -142 °C. In place of

neat F₂, a 4:1 mixture of N₂:F₂at a pressure of about 400 Torr was

used. The whole recirculating loop was connected to a stainless-

steel Teﬂon vacuum line to allow evacuation and the introduction

of other reagents.

product was characterized by low-temperature infrared and Raman

spectroscopy.

Experiments involving [XeF][SbF₆] were carried out in an fashion

analogous to those involving [XeF][AsF₆]. In one case, the low-

temperature Raman spectrum of a sample of [H₃O][SbF₆] and XeF₂,

obtained from an HF solution of [XeF][SbF₆] and H₂O at -64 °C,

was recorded with a relatively high laser power of 1.6 W and

revealed some additional features. In addition to the main products,

[H₃O][SbF₆] and XeF₂, the Raman spectrum showed a small amount

of a new compound with a relatively strong band at 452 cm^-1

.

Furthermore, careful temperature cycling in the laser beam between

-130 and -50 °C resulted in complete conversion of the XeF₂to

[XeF][SbF₆] and [Xe₂F₃][SbF₆], probably arising from local heating

effects. After warming brieﬂy to room temperature, the ﬁnal

products were [XeF][SbF₆], [Xe₂F₃][SbF₆], and [H₃O][SbF₆].

Synthesis of [Xe₃OF₃][AsF₆]. (a) Reaction of [XeF][AsF₆]

and H₂O in HF. In a typical experiment, 8.0 µL of H₂O (0.44

mmol) was syringed into a 0.25-in. o.d. FEP reaction tube equipped

with a Kel-F valve inside a well-purged dry nitrogen-ﬁlled glovebag.

Approximately 0.4 mL of HF was condensed into the tube at -196

°C, and the water was mixed with the HF at room temperature.

After the FEP tube was transferred into a drybox at room

temperature, 0.1549 g (0.4566 mmol) of [XeF][AsF₆] was added

to the frozen H₂O/HF mixture followed by warming to -78 °C

outside the drybox. The sample was maintained at -78 °C.

The reaction rate was found to be strongly dependent on the

concentration, reaction temperature, and the degree of initial mixing.

Without initial agitation, small amounts of [XeF][AsF₆] were

detected in the Raman spectrum of the precipitate at -78 °C after

ca. 1 week. After thorough mixing at an initial concentration of

1.77 M H₂O, [H₃O][AsF₆]·2XeF₂and [Xe₂F₃][AsF₆] were formed

after only 2 h with [XeF][AsF₆] being completely reacted at -78

°C, and a signiﬁcant amount of [Xe₃OF₃][AsF₆] was present after

12 h. At H₂O concentrations of 0.26 M, [H₃O][AsF₆]·2XeF₂was

formed within approximately 1 day, while [Xe₂F₃][AsF₆] was the

major species in the solid precipitate after approximately 2 days at

-78 °C. Upon standing for a further 5 days at -78 °C,

[Xe₃OF₃][AsF₆] became the major species.

A mixture of 1 mmol of pure HOF and 0.4 mL of liquid

SO₂ClF was prepared in the HOF U-trap and condensed into a

4-mm FEP NMR tube at -196 °C, followed by heat sealing of

the FEP tube. In a second experiment, a 3-fold excess of a 1:1

mixture of HF and AsF₅was added to the NMR sample, which

was maintained and recorded at -78 °C. In a third experiment,

an excess of a 1:1 mixture of HOSO₂F and SbF₅was ﬁrst

introduced into the NMR tube, and the HOF/SO₂ClF mixture

was condensed onto it at -196 °C.

Reactions of [XeF][PnF₆] (Pn ) As, Sb) and H₂O in HF

Solution. In a typical experiment, H₂O (18.0 µL, 1.00 mmol) was

syringed into a 0.75-in. o.d. FEP reactor. The ampoule was closed

with a stainless-steel valve, cooled to -196 °C, and evacuated.

Approximately 5 mL of anhydrous HF was condensed into the

reactor at -196 °C, and the mixture was homogenized at room

temperature. The reactor was transferred to a drybox and cooled

to -196 °C, and [XeF][AsF₆] (0.339 g, 1.00 mmol) was added.

The cold reactor was evacuated on the vacuum line, allowed to

warm to -64 °C, kept at this temperature for 12 h, and checked

for xenon evolution by monitoring the pressure above the liquid

phase with a pressure gauge. No gas evolution was observed, and

a clear colorless solution of [H₃O][AsF₆] and XeF₂resulted. The

HF solvent was pumped off at -64 °C, leaving behind a white

solid residue (0.360 g; the weight calculated for 1.00 mmol of a

1:1 molar mixture of [H₃O][AsF₆] and XeF₂was 0.377 g). The

An NMR sample was prepared by loading, inside the drybox,

0.20849 g (0.6146 mmol) of [XeF][AsF₆] into a 10-mm o.d. FEP

tube fused to a piece of 0.25-in. o.d. FEP tubing connected to a

Kel-F valve. Approximately 2.35 mL of HF was condensed onto

the [XeF][AsF₆] at -196 °C. The [XeF][AsF₆] was dissolved upon

warming the mixture to room temperature. Inside the drybox, 11.5

µL (0.60 mmol) of H₂¹⁷O was syringed onto the frozen [XeF][AsF₆]/

HF mixture and transferred outside the drybox where the sample

was heat sealed and stored at -196 °C until characterized by NMR

spectroscopy.

(b) Reaction of [H₃O][PnF₆] (Pn ) As, Sb) with XeF₂. (i)

In HF Solution. The syntheses of high-purity [Xe₃OF₃][PnF₆]

involved dissolution at -50 °C of near-equimolar amounts of

[H₃O][PnF₆] and XeF₂(up to ca. 20 mol % excess XeF₂) at ca.

0.2-3 M H₃O⁺in a 0.25-in. FEP reactor which had a side arm

fused to it. The solution was rapidly warmed to -35 °C for ca.

30 s and immediately cooled to -50 °C. After 5 min at -50 °C,

a voluminous deep red-orange precipitate of [Xe₃OF₃][PnF₆]

formed. The reaction mixture was maintained at -50 °C for an

additional 20-30 min to ensure the reaction was complete.

Unreacted XeF₂and/or [H₃O][PnF₆], as well as [Xe₂F₃][PnF₆]

byproduct, were soluble and were decanted from the settled

precipitate at -50 °C into the side arm of the reactor at -78 °C.

The product, [Xe₃OF₃][PnF₆], decomposed under HF above -30

°C with Xe gas evolution.

+

The critical effect of initial reaction temperature on Xe₃OF₃

(47) Casteel, W. J., Jr.; Kolb, P.; LeBlond, N.; Mercier, H. P. A.;

Schrobilgen, G. J. Inorg. Chem. 1996, 35, 929–942.

(48) Chernick, C. L.; Malm, J. G. Inorg. Synth. 1966, 8, 254–258.

(49) Gillespie, R. J.; Landa, B. Inorg. Chem. 1973, 12, 1383–1389.

(50) Emara, A. A. A.; Schrobilgen, G. J. Inorg. Chem. 1992, 31, 1323–

1332.

cation formation was demonstrated by a related synthesis at a lower

temperature (-64 °C). A solution of freshly prepared [H₃O][AsF₆]

(1.00 mmol; ca. 0.2 M) in anhydrous HF (ca. 5 mL) was cooled to

-196 °C and XeF₂(1.00 mmol; ca. 0.2 M) was added in the

drybox., the mixture was warmed to -64 °C for 12 h. The resulting

9

13486 J. AM. CHEM. SOC. VOL. 131, NO. 37, 2009

Xenon(II) Oxide Fluoride Cation, FXeOXeFXeF⁺

A R T I C L E S

clear, colorless solution did not show any signs of xenon evolution

during this time period, nor did [Xe₃OF₃][AsF₆] crystallize from

solution. The solvent was pumped off at -64 °C, leaving behind

a solid white residue, which was identiﬁed by low-temperature

°C) followed by brief warming to 0 °C outside the drybox to

solubilize XeF₄in the HF solvent. The sample was cooled to -78

°C and maintained at that temperature for several months. Inspection

of the sample revealed deep red-orange, needle-shaped crystals

above a white crystalline precipitate in a colorless supernatant. The

white material was shown to be [Xe₂F₃][SbF₆] by recording the

Raman spectrum under frozen HF. Crystals were isolated by

decanting the supernatant into the side arm of the reactor as

described above. The crystal used for the X-ray structure determi-

nation had the dimensions 0.10 × 0.10 × 0.06 mm³.

Raman spectroscopy as

[XeF][AsF₆], and [Xe₂F₃][AsF₆].

a mixture of [H₃O][AsF₆] · 2XeF₂,

(ii) As a Solid Mixture. A 1:1 molar solid mixture of

[H₃O][AsF₆] (0.1626 g, 0.782 mmol) and XeF₂(0.1324 g, 0.782

mmol) were slowly warmed in a dynamic vacuum to room

temperature. The volatile products were trapped at -196 °C and

identiﬁed by infrared spectroscopy as HF and AsF₅. During the

warm-up, the color of the sample changed from white to red-orange

and back to white. The white solid residue (0.186 g; weight

calculated for 0.2065 mmol of [Xe₂F₃][AsF₆] and 0.413 mmol of

[H₃O][AsF₆], 0.190 g) was identiﬁed by Raman spectroscopy as a

mixture of [Xe₂F₃][AsF₆] and [H₃O][AsF₆].

Reaction of [XeF][AsF₆]/H₂O with ClF₃. An equimolar solution

of [XeF][AsF₆] and H₂O in HF solvent was reacted for 12 h at

-64 °C, and an equimolar amount of ClF₃was added at -196 °C.

The mixture was warmed to -78 °C. After about 2 min, a red

suspension formed that eventually turned white. The mixture was

allowed to slowly warm to room temperature, and the volatile

products were pumped off at room temperature, leaving behind a

white solid residue that was shown by infrared spectroscopy to

contain [ClO₂][AsF₆] but no [ClOF₂][AsF₆].

Reaction of XeF₂and [H₃O][AsF₆] in BrF₅Solution. Inside

the drybox, 0.0419 g (0.2015 mmol) of [H₃O][AsF₆] and 0.0372 g

(0.2197 mmol) of XeF₂were loaded at low temperature (ca. -150

°C) into a 4-mm o.d. FEP tube connected to a Kel-F valve.

Approximately 0.25 mL of BrF₅was condensed into the tube at

-196 °C, and the FEP tube was heat-sealed under dynamic vacuum.

Reaction of XeF₄and [H₃O][SbF₆] in HF Solution. An NMR

sample was prepared using a 4-mm o.d. FEP tube equipped with a

Kel-F valve. Inside the drybox, 0.09110 g (0.3576 mmol) of

[H₃O][SbF₆] and 0.07908 g (0.3815 mmol) of XeF₄were loaded

into the FEP tube at low temperature (ca. -140 °C), followed by

condensation of ca. 0.2 mL HF onto the solid mixture. After sealing,

the sample was warmed to 0 °C, resulting in apparent incomplete

dissolution of the solid white reaction mixture and gas evolution.

The Raman spectrum of the solid under frozen HF showed bands

assigned to the reduction products, [XeF][SbF₆] and [Xe₂F₃][SbF₆].

Crystal Growth. Single crystals suitable for X-ray structure

determinations were grown from HF solvent inside a dry stream

of cooled nitrogen gas as previously described. With the exception

of [Xe₃OF₃][SbF₆], the 0.25-in. o.d. vessels used for crystal growth

were equipped with side arms to permit decantation of the

supernatant once crystal growth was completed.

(c) [H₃O][AsF₆]·2XeF₂. Inside the drybox, 8.0 µL (0.44 mmol)

of H₂O was syringed into a 0.25-in o.d. FEP T-shaped reactor ﬁtted

with a Kel-F valve. The reactor was then removed to a metal

vacuum manifold where ca. 0.5 mL of anhydrous HF was distilled

onto the sample. The reactor and contents were allowed to warm

to ambient temperature and mixed. The reactor, back pressured to

ca. 1 atm with dry N₂gas at -78 °C, was then transferred to the

drybox where 0.16178 g (0.47695 mmol) of [XeF][AsF₆] was

transferred into the reactor that was frozen at ca. -140 °C. The

sample was then immediately removed from the drybox, allowed

to warm to -68 °C, and mixed for a period of 12 min, resulting in

the formation of a colorless precipitate. The vessel was brieﬂy

warmed above -68 °C and mixed before being inserted into the

cold ﬂow (ca. -70 °C) of the crystal growing apparatus. Crystals

were grown slowly with the temperature being decreased at a rate

of 0.5 °C min^-1to ca. -85 °C, after which the HF solvent was

decanted into the side arm of the reactor, which was heat-sealed

off under vacuum and removed. The residual HF solvent was then

removed under vacuum at ca. -85 °C. The crystal used for the

X-ray structure determination had the dimensions 0.14 × 0.14 ×

0.14 mm³.

X-ray Structure Determinations of [Xe₃OF₃][PnF₆] (Pn )

As, Sb) and [H₃O][AsF₆]·2XeF₂. Crystals of [Xe₃OF₃][PnF₆] (Pn

) As, Sb) and [H₃O][AsF₆]·2XeF₂were mounted under a ﬂow of

cold nitrogen as previously described,⁵¹and data were collected at

-173, -110, and -145 °C, respectively.

(a) Collection and Reduction of X-ray Data. Data were

collected using a P4 Siemens diffractometer, equipped with a

Siemens SMART 1K charge-coupled device (CCD) area detector

(using the program SMART)⁵²and a rotating anode using graphite-

monochromated Mo KR radiation (λ ) 0.71073 Å). The crystal-

to-detector distances were 5.0120, 4.9840, and 4.9870 cm for

2XeF₂·[H₃O][AsF₆], [Xe₃OF₃][AsF₆], and [Xe₃OF₃][SbF₆], respec-

tively, and the data collection was carried out in a 512 × 512 pixel

mode using 2 × 2 pixel binning. Complete spheres of data were

collected to better than 0.8 Å resolution. Processing was carried

out by using the program SAINT,⁵³which applied Lorentz and

polarization corrections to three-dimensionally integrated diffraction

spots. The program SADABS⁵⁴was used for the scaling of

diffraction data, the application of a decay correction, and an

empirical absorption correction based on redundant reﬂections.

(b) Solution and Reﬁnement of the Structures. The structure

determinations were performed as previously described using the

program SHELXTL.⁵⁵The program XPREP⁵⁵was used to conﬁrm

the unit cell dimensions and the crystal lattices. Solutions were

obtained using direct methods that located the Xe and As or Sb

atoms in the structures of [H₃O][AsF₆]·2XeF₂, [Xe₃OF₃][AsF₆], and

[Xe₃OF₃][SbF₆]. The positions of all ﬂuorine and oxygen atoms

were revealed in successive difference Fourier syntheses. The

(a) [Xe₃OF₃][AsF₆]. Inside a well-purged glovebag, 11.0 µL

(0.61 mmol) of H₂O was syringed into a 0.25-in. o.d. FEP tube

equipped with a Kel-F valve that had a 0.25-in. o.d. FEP side arm

fused to it. Approximately 0.5 mL of HF was condensed into the

tube at -196 °C, and the contents were mixed at room temperature.

After transferring the reaction tube into a drybox at room temper-

ature, 0.2082 g (0.6138 mmol) of [XeF][AsF₆] was added to the

frozen H₂O/HF mixture followed by warming to -78 °C outside

the drybox. After maintaining the sample at -78 °C for ca. 3 weeks,

red-orange dendrimeric clusters of crystals appeared above an

orange-brown precipitate. The crystals were isolated by decanting

the supernatant into the side arm of the reactor, which was cooled

to -196 °C and subsequently heat-sealed off, followed by removal

of the residual HF solvent from the crystals under dynamic vacuum

at -78 °C. The crystal used in the X-ray structure determination

had the dimensions 0.14 × 0.12 × 0.10 mm³.

(51) Gerken, M.; Dixon, D. A.; Schrobilgen, G. J. Inorg. Chem. 2000, 39,

4244–4255.

(52) SMART, release 5.054; Siemens Energy and Automation Inc.: Madison,

WI, 1999.

(53) SAINT+, release 6.01; Siemens Energy and Automation Inc.: Madison,

(b) [Xe₃OF₃][SbF₆]; Reaction of XeF₄with [H₃O][SbF₆]. Inside

the drybox, 0.07226 g (0.2836 mmol) of [H₃O][SbF₆] was

transferred to a 0.25-in. o.d. FEP tube ﬁtted with a Whitey ORM2

stainless steel valve. Approximately 0.5 mL of HF was condensed

into the tube at -78 °C. Inside the drybox, 0.0671 g (0.3122 mmol)

of XeF₄was added to the frozen [H₃O][SbF₆]/HF mixture (ca. -140

WI, 1999.

(54) Sheldrick, G. M. SADABS (Siemens Area Detector Absorption

Corrections), Version 2.10; Siemens Analytical X-ray Instruments, Inc.:

Madison, WI, 2004.

(55) Sheldrick, G. M. SHELXTL, release 5.1; Siemens Analytical X-ray

Instruments, Inc.: Madison, WI, 1998.

9

J. AM. CHEM. SOC. VOL. 131, NO. 37, 2009 13487

A R T I C L E S

Gerken et al.

PLATON program⁵⁶was used to check for possible alternative

space groups. Despite the presence of multiple disorders in each

structure, the ﬁnal reﬁnements are quite acceptable.

using a Bruker 5-mm ¹H/¹³C/¹⁹F/¹³C QNP probe. Samples in 4-mm

FEP tubes were heat sealed under dynamic vacuum at -196 °C

and inserted into a thin wall precision glass 5-mm NMR tube

(Wilmad).

The ¹⁹F, ¹²⁹Xe, and ¹⁷O NMR spectra of a sample of [XeF][AsF₆]

and ¹⁷O-enriched H₂O (HF solvent at -75 °C) were acquired at

470.477, 138.867, and 67.811 MHz in 64 K, 32 K, and 32 K

memories with spectral settings of 150, 100, and 100 kHz, yielding

acquisition times of 0.218, 0.164, and 0.164 s and data point

resolutions of 2.29, 3.05, and 3.05 Hz/data point, respectively. The

number of transients accumulated for the ¹⁹F, ¹²⁹Xe, and ¹⁷O NMR

spectra was 721, 363, and 5253 using a pulse width of 2.5, 10, and

10 µs, respectively.

The reaction of [H₃O][AsF₆] and XeF₂in BrF₅solvent was

monitored by ¹⁹F NMR spectroscopy, with spectra acquired at

282.409 MHz (AC-300), followed by an acquisition at 470.477

(DRX-500) when the reaction was deemed to be complete (values

in parentheses), in 32 (64) K memory with a spectral setting of

100 (100) kHz yielding an acquisition time of 0.164 (0.328) s and

data point resolution of 6.104 (1.53) Hz/data point using a pulse

width of 4.0 (2.5) µs. The number of transients accumulated was

4100 (1000).

The ¹⁷O, ¹⁹F, and ¹²⁹Xe NMR spectra were referenced externally

at 30 °C to samples of neat H₂O, CFCl₃, and XeOF₄, respectively.

The chemical shift convention used is that positive (negative) sign

indicates a chemical shift to high (low) frequency of the reference

compound.

Raman Spectroscopy. Raman spectra were obtained directly

in the FEP reaction vessel using two different Raman spectrometers:

(a) a Cary Model 83GT using the 488-nm line of a Lexel Model

95 Ar⁺laser and (b) a Bruker RFS 100/S FT Raman spectrometer

using 1064-nm excitation as previously described.⁴⁸As previously

described, low-temperature spectra on the Cary and the Bruker

spectrometers were recorded using a macrochamber⁵⁷and a Bruker

low-temperature accessory,⁵¹respectively.

Infrared Spectroscopy. Infrared spectra were recorded in the

range 300-4000 cm^-1on a Midac Model M FTIR spectrometer.

Spectra of solids were obtained by using dry powders pressed

between AgCl windows in an Econo press (Barnes Engineering

Co.). For the low-temperature spectra, the cold sample was placed

between cold AgCl windows inside the drybox. The windows were

mounted in a liquid-nitrogen-cooled copper block mated with an

O-ring ﬂange to an evacuable glass cell equipped with outer CsI

windows.⁵⁸

Computational Methods. Electronic structure calculations were

carried out using DFT (SVWN, BP86, B3LYP, B3PW91, PBE1PBE

and MPW1PW91) and ab initio (MP2 and CCSD(T)) methods using

the program Gaussian 03 (version C.02). The aug-cc-pVDZ, and

aug-cc-pVTZ⁴⁵basis sets as implemented in the Gaussian program

were utilized for all elements except Xe, for which the semirela-

tivistic small core pseudopotential basis sets aug-cc-pVDZ-PP and

aug-cc-pVTZ-PP were used. The combined use of aug-cc-pVnZ

and aug-cc-pVnZ-PP basis sets is indicated as aug-cc-pVnZ(-PP).

All DFT and MP2 calculations were performed at McMaster

University, and the CCSD(T) calculations were performed at Edwards

Air Force Base. The program GaussView was used to visualize the

vibrational displacements that form the basis of the vibrational mode

descriptions given in Tables 4, 5, S3, and S5-S8). Nuclear magnetic

resonance shielding tensors were calculated by the GIAO-MBP2

method using CCSD(T)/aug-cc-pVDZ optimized geometries.

(i) [Xe₃OF₃][PnF₆]. The original cells were conﬁrmed, and the

lattice exceptions were unambiguous, showing the lattice to be

monoclinic primitive. An examination of the E-statistics gave a

value that was intermediate between the expected values for a

centrosymmetric and a noncentrosymmetric space group; however,

this parameter was considered with caution because of the large

number of heavy atoms in the structure. The presence of a c-glide

was evident from the systematic absences, but the 0k0 reﬂections

(k ) odd) were either weak or absent; hence a ﬁrst choice for the

space group was P2₁/c. Although a reasonable solution could be

obtained in this space group for both salts, comparison of the Raman

-

spectra of the AsF₆and SbF₆salts and the results of a factor-

group analysis using P2₁/c revealed an inconsistency in the case

-

of the AsF₆salt. To see all the modes observed experimentally,

the site symmetry could not be C_ias implied by P2₁/c because the

bands originating from the ungerade modes (T_1uand T_2u) were

observed. At this point, the tolerance (i.e., minimum I/σ(I) GAP)

was changed before the space group was selected; this gave rise to

new options for the choice of space group. The structure of the

AsF₆^-salt was resolved using the space group Pc, and the factor-

group analysis was redone accordingly, and accounted for the factor-

group components associated with the T_1uand T_2ubands (Tables

S1 and S2; also see Raman Spectroscopy).

In the structure of [Xe₃OF₃][SbF₆], the Xe(2) atom was located

on a general position which resulted in a positional disorder, i.e.,

Xe(2) gives rise to Xe(2A) by symmetry. In the structure of

[Xe₃OF₃][AsF₆], Xe(2) and Xe(2A) are not related by symmetry.

Electroneutrality required the presence of an oxygen bridge in the

Xe₃OF₃⁺cation which, in conjunction with the split Xe(2) position,

results in two possible disorder models (see Results and Discussion).

-

The PnF₆anions are disordered between two orientations. The

ﬁnal reﬁnements were obtained by using data that had been

corrected for absorption and by introducing anisotropic thermal

parameters for all the atoms. During the ﬁnal stages of the

reﬁnement, all reﬂections with F²< -2σ(F²) were suppressed,

extinction parameters were used, and weighting factors recom-

mended by the reﬁnement program were introduced.

(ii) [H₃O][AsF₆]·2XeF₂. The atom at the special position (m.mm)

was deﬁned as the oxygen of the H₃O⁺cation, and as a

consequence, the disordered positions of hydrogen atoms about the

oxygen atom could not be found in the difference map. The possible

interpretations of the crystal structure of [H₃O][AsF₆]·2XeF₂as

2HOXeF ·[H₃O][AsF₆] or [H₂OXeF]₂[F][AsF₆] were ruled out

because of the absence of ¹⁸O and D isotopic shifts for the 552 and

470 cm^-1bands in the experimental Raman spectra of

[H₃¹⁸O][AsF₆]·2XeF₂and [D₃O][AsF₆]·2XeF₂. Furthermore, the

lack of agreement between the observed and calculated harmonic

frequencies for the H₂OXeF⁺cation and HOXeF eliminate these

species as possible constituents of the solid. In addition, failure to

observe the ¹⁸O isotopic shifts, predicted computationally, also ruled

out this species (see Computational Results).

Nuclear Magnetic Resonance Spectroscopy. Fluorine-19, ¹²⁹Xe,

and ¹⁷O NMR spectra were recorded unlocked (ﬁeld drift <0.1 Hz

h^-1) on a Bruker DRX-500 spectrometer equipped with an 11.744-T

cryomagnet and a Bruker 10 mm broadband probe. The NMR probe

was cooled using a nitrogen ﬂow and variable-temperature controller

(BV-T 3000). For NMR spectroscopic characterization, samples

were loaded into 9-mm FEP tubes heat-fused to a ca. 5-cm length

of 0.25-in. FEP, which in turn was connected to a Kel-F valve,

subsequently heat-sealed under dynamic vacuum at -196 °C, and

inserted into a thin wall precision glass 10-mm NMR tube

(Wilmad). Fluorine-19 NMR spectra acquired while monitoring the

reaction of XeF₂with [H₃O][AsF₆] in BrF₅solvent were recorded

unlocked ((ﬁeld drift <0.1 Hz h^-1) on a Bruker AC-300 spectrometer

Acknowledgment. We thank the Natural Sciences and Engi-

neering Research Council of Canada for support in the form of a

Discovery Grant (G.J.S.) and for the award of postgraduate

scholarships (M.G. and M.D.M.). Furthermore, we thank McMaster

(57) Miller, F. A.; Harney, B. M. Appl. Spectrosc. 1970, 24, 291–292.

(58) Loos, K. R.; Campanile, V. R.; Goetschel, T. C. Spectrochim. Acta,

Part A 1970, 26A, 365.

(56) Farrugia, L. J. Appl. Crystallogr. 1999, 32, 837–838.

9

13488 J. AM. CHEM. SOC. VOL. 131, NO. 37, 2009

Xenon(II) Oxide Fluoride Cation, FXeOXeFXeF⁺

A R T I C L E S

University for the award of James A. Morrison Memorial Scholar-

ships and the Ontario Ministry of Education and Training for the

award of Ontario Graduate Scholarships (M.G. and B.E.P.) and

for the award of a Richard Fuller Memorial Scholarship (M.G.).

The authors also thank David S. Brock for improving the synthesis

and Raman spectrum of [Xe₃OF₃][SbF₆]. The quantum-chemical

calculations carried out at McMaster were made possible by the

facilities of the Shared Hierarchical Academic Research Computing

Network (SHARCNET: www.sharcnet.ca). The work at USC was

supported by the National Science Foundation under Grant Nos.

CHE-9985833 and CHE-0456343, the Air Force Ofﬁce of Scientiﬁc

Research, and the Ofﬁce of Naval Research. We also thank the

Deutsche Forschungsgemeinschaft for a Stipend (B.H.), Dr. J.

Sheehy, Edwards Air Force Base, for some of the quantum-chemical

calculations, and Prof. George A. Olah for his steady support.

[Xe₃OF₃][PnF₆] (Pn ) As, Sb) (Tables S1 and S2); calculated

vibrational frequencies and geometrical parameters for Xe₃OF₃

+

using the aug-cc-pVDZ(-PP) basis set (Tables S3 and S4);

calculated vibrational frequencies for [H₃O][AsF₆]·nXeF₂(n )

1-4) (Table S5); calculated geometries, frequencies and infrared

intensities for symmetric and asymmetric XeF₂(Table S6);

calculated vibrational frequencies and geometrical parameters

for HOF (Table S7), H₂OF⁺(Table S8), and FXeOH and for

H₂OXeF⁺(Table S9); benchmark computational studies for

XeF₂(Table S10); calculated XeF⁺afﬁnities (Table S11);

absolute energies for all computed species (Table S12); X-ray

crystallographic ﬁles in CIF format for the structure determina-

tions of [Xe₃OF₃][PnF₆] (Pn ) As, Sb) and [H₃O][AsF₆]·2XeF₂.

This material is available free of charge via the Internet at http://

pubs.acs.org.

Supporting Information Available: Views of [Xe₃OF₃][PnF₆]

(Pn ) As, Sb) (Figures S1 and S2); factor-group analyses for

JA905060V

9

J. AM. CHEM. SOC. VOL. 131, NO. 37, 2009 13489

Article Doi

On the XeF+/H2O system: Synthesis and characterization of the xenon(II) oxide fluoride cation, FXeOXeFXeF+

DOI: 10.1021/ja905060v

Source and publish data:

Authors:

Article abstract of DOI:10.1021/ja905060v

Full text of DOI:10.1021/ja905060v

Products guided by the article

R&D Labs maybe for 223246-95-5

Relevant to this article

Hot Product