Mixed Noble‐Gas Compounds of Krypton(II) and Xenon(VI); [F₅Xe(FKrF)AsF₆] and [F₅Xe(FKrF)₂AsF₆]

Abstract

The coordination chemistry of KrF₂ has been limited in contrast with that of XeF₂, which exhibits a far richer coordination chemistry with main‐group and transition‐metal cations. In the present work, reactions of [XeF₅][AsF₆] with KrF₂ in anhydrous HF solvent afforded [F₅Xe(FKrF)AsF₆] and [F₅Xe(FKrF)₂AsF₆], the first mixed krypton/xenon compounds. X‐ray crystal structures and Raman spectra show the KrF₂ ligands and [AsF₆]⁻ anions are F‐coordinated to the xenon atoms of the [XeF₅]⁺ cations. Quantum‐chemical calculations are consistent with essentially noncovalent ligand−xenon bonds that may be described in terms of σ‐hole bonding. These complexes significantly extend the XeF₂–KrF₂ analogy and the limited chemistry of krypton by introducing a new class of coordination compound in which KrF₂ functions as a ligand that coordinates to xenon(VI). The HF solvates, [F₅Xe(FH)AsF₆] and [F₅Xe(FH)SbF₆], are also characterized in this study and they provide rare examples of HF coordinated to xenon(VI).

KrF₂ and [XeF₅][AsF₆] react in anhydrous HF to form [F₅Xe(FKrF)AsF₆] and [F₅Xe(FKrF)₂AsF₆]. The complexes were structurally characterized by low‐temperature single‐crystal X‐ray diffraction and Raman spectroscopy. The KrF₂ ligands interact with xenon(VI) through Xe‐ ‐ ‐FKrF secondary bonds that are shown by computational studies to be noncovalent, electrostatic, σ‐hole interactions. The complexes provide unique examples of mixed noble‐gas compounds.

Introduction

Krypton reactivity was discovered ^[1] soon after the landmark synthesis of the first true noble‐gas compound, Xe[PtF₆]. ^[2] Although the precise formulation of Xe[PtF₆] remains unproven, it is likely a salt or a mixture of [XeF]⁺ salts.[ ³ , ⁴ ] Although the discoveries of xenon and krypton chemical reactivities occurred nearly 60 years ago and within a year of one another, their chemistries never became entwined to form a single compound that contains both chemically bound noble gases. In contrast with xenon, which exhibits formal oxidation states in its isolated compounds of 0, +$1/2$ , +2, +4, +6, and +8, krypton only exhibits the +2 oxidation state and a far more limited chemistry. The only binary krypton fluoride that can be synthesized in macroscopic and synthetically useful amounts is KrF₂,[ ⁵ , ⁶ , ⁷ ] from which all other krypton compounds have been derived.[ ⁵ , ⁶ , ⁷ , ⁸ , ⁹ , ¹⁰ , ¹¹ , ¹² ]

Prior studies have explored the ligating properties of KrF₂ and have provided several KrF₂ adducts that were structurally characterized by low‐temperature (LT) single‐crystal X‐ray diffraction (SCXRD) and Raman spectroscopy. The latter include complexes with a main‐group Br^V oxyfluoride cation, [F₂OBr(FKrF)₂AsF₆], ^[8] a neutral covalent transition‐metal Hg^II compound, Hg(OTeF₅)₂⋅1.5KrF₂, ^[9] a transition‐metal cation, Hg²⁺, [Hg(FKrF)₈][AsF₆]₂, ^[11] and a main‐group metal cation, Mg²⁺, [Mg(FKrF)₄(AsF₆)₂]. ^[10] Most recently, the KrF₂ adducts of the weak fluoride‐ion acceptor, CrOF₄, have been reported and structurally characterized for KrF₂⋅nCrOF₄ (n=1, 2). ^[12] The xenon analogues, [F₂OBr(FXeF)₂AsF₆], ^[13] Hg(OTeF₅)₂⋅1.5 XeF₂, ^[9] XeF₂⋅nCrOF₄ (n=1, 2), ^[12] and [Mg(FXeF)₄(AsF₆)₂] ^[14] have also been synthesized and structurally characterized by SCXRD and Raman spectroscopy. The linear, centrosymmetric (D _∞h) NgF₂ (Ng=Kr, Xe) molecules ^[15] distort upon coordination to a fluoride‐ion acceptor (A) to form a Ng−F_b‐ ‐ ‐A bridge in which the Ng−F_b bond is elongated and the terminal Ng−F_t bond is contracted relative to free NgF₂. The extent to which distortion and polarization of the NgF₂ ligand occurs, and thus the extent to which the positive charge on Ng is enhanced, depends on the Lewis acidity of the fluoride‐ion acceptor.[ ⁵ , ⁶ , ¹⁶ ] Interactions with the strongly Lewis acidic pnictogen pentafluorides, PnF₅, result in the formation of strongly ion‐paired [NgF][PnF₆][ ⁶ , ¹⁵ ] salts in which the [NgF]⁺ cations and [PnF₆]⁻ anions interact by means of Ng‐ ‐ ‐F_b−Pn bridges. The electrophilicities of [NgF]⁺ and coordinated NgF₂ ligands relative to free NgF₂ are manifested by marked increases in their oxidative fluorinating abilities. ^[5] In the case of KrF₂, the number of suitable Lewis acids that can coordinate to KrF₂ and withstand its extraordinary oxidative fluorinating strength is very limited.

Two criteria must therefore be met for the formation of a mixed xenon/krypton adduct: (1) The fluorobasicity of KrF₂ must closely balance the Lewis acidity of the xenon substrate, i.e., a Lewis acid that is too weak will be unable to coordinate KrF₂, whereas a Lewis acid that is too strong will abstract F⁻ to form a more electrophilic and strongly oxidizing [KrF]⁺ salt. (2) The fluoride‐ion acceptor must be sufficiently resistant to attack by the potent oxidative fluorinator, KrF₂. The Lewis acidic [XeF₅]⁺ cation meets these criteria by virtue of its net positive charge and the high formal oxidation state of xenon (+6).

The coordination behavior of the [XeF₅]⁺ cation in its salts is well documented for Xe‐ ‐ ‐F_b interactions between the [XeF₅]⁺ cations and their counteranions. ^[17] Examples in which [XeF₅]⁺ is coordinated to a XeF₂ ligand are known, for [F₅Xe(FXeF)XeF₅(AsF₆)₂], [F₅Xe(FXeF)AsF₆], and [F₅Xe(FXeF)₂AsF₆],[ ¹⁸ , ¹⁹ ] which were characterized by SCXRD and Raman spectroscopy, and [F₅Xe(FXeF)RuF₆], ^[20] which was characterized by Raman spectroscopy. In contrast, the cocrystal, [XeF₅][SbF₆]⋅XeOF₄, ^[21] exhibits no interactions between [XeF₅]⁺ and XeOF₄, in accordance with the low relative fluorobasicity of XeOF₄.

Results and Discussion

Syntheses

In the present work, the Lewis acidity of the [XeF₅]⁺ cation and the fluorobasic character of KrF₂ have been exploited for the syntheses of the first mixed noble‐gas (Kr/Xe) compounds that are isolable in macroscopic quantities. The products obtained from the LT reactions of [XeF₅][AsF₆] and KrF₂ in anhydrous HF (aHF) solvent and subsequent crystallizations at LT depended on the initial KrF₂:[XeF₅][AsF₆] molar ratio. The complex, [F₅Xe(FKrF)AsF₆] (1), was obtained by use of a 1.5:1 molar ratio of reactants, whereas a stoichiometric excess of KrF₂ (3.5:1 or 2.1:1) resulted in [F₅Xe(FKrF)₂AsF₆] (2). In an attempt to prepare the KrF₂ analogue of [F₅Xe(FXeF)XeF₅(AsF₆)₂], ^[19] a 1:1.9 molar ratio of reactants was used, which resulted in crystallization of 1 and [F₅Xe(FH)AsF₆] (3). Compound 3 was also isolated from an aHF solution of [XeF₅][AsF₆] upon removal of HF at LT. The synthesis of the antimony analogue [F₅Xe(FH)SbF₆] (4) is described in the Supporting Information. The syntheses of 1–3 are in accordance with the proposed Equilibria 1, 2, 3, 4, which are supported by LT SCXRD structure determinations of the adduct‐cation salts, [F₅Xe(FKrF)AsF₆] and [F₅Xe(FKrF)₂AsF₆], and the intermediate solvate, [F₅Xe(FH)AsF₆], as well as by LT Raman spectroscopy. Vibrational frequency assignments were aided by calculated frequencies and intensities obtained from DFT calculations (vide infra). It is apparent that HF also behaves as a weak ligand towards [XeF₅]⁺ in an HF solution and that KrF₂, a somewhat less fluorobasic ligand than XeF₂, ^[22] is sufficiently fluorobasic to displace HF to form [F₅Xe(FKrF)AsF₆] [Eq. 3].

$${\displaystyle \left[\mathrm{XeF}{}_5\right]\left[\mathrm{AsF}{}_6\right]+\mathrm{HF}\overrightarrow{\leftarrow }\left[\mathrm{F}{}_5\mathrm{Xe}\left(\mathrm{FH}\right)\mathrm{AsF}{}_6\right]}$$ (1)

$${\displaystyle \left[\mathrm{XeF}{}_5\right]\left[\mathrm{AsF}{}_6\right]+\mathrm{KrF}{}_2\overrightarrow{\leftarrow }\left[\mathrm{F}{}_5\mathrm{Xe}\left(\mathrm{FKrF}\right)\mathrm{AsF}{}_6\right]}$$ (2)

$${\displaystyle \mathrm{and}/\mathrm{or}\left[\mathrm{F}{}_5\mathrm{Xe}\left(\mathrm{FH}\right)\mathrm{AsF}{}_6\right]+\mathrm{KrF}{}_2\overrightarrow{\leftarrow }\left[\mathrm{F}{}_5\mathrm{Xe}\left(\mathrm{FKrF}\right)\mathrm{AsF}{}_6\right]+\mathrm{HF}}$$ (3)

$${\displaystyle \left[\mathrm{F}{}_5\mathrm{Xe}\left(\mathrm{FKrF}\right)\mathrm{AsF}{}_6\right]+\mathrm{KrF}{}_2\overrightarrow{\leftarrow }\left[\mathrm{F}{}_5\mathrm{Xe}\left(\mathrm{FKrF}\right){}_2\mathrm{AsF}{}_6\right]}$$ (4)

 

X‐ray Crystallography

Details of X‐ray data collection and crystallographic information pertaining to [F₅Xe(FKrF)AsF₆] (1), [F₅Xe(FKrF)₂AsF₆] (2), [F₅Xe(FH)AsF₆] (3), and [F₅Xe(FH)SbF₆] (4) are summarized in Table 1.

Table 1.

Summary of X‐ray crystal data and refinement results for [F₅Xe(FKrF)AsF₆] (1), [F₅Xe(FKrF)₂AsF₆] (2), [F₅Xe(FH)AsF₆] (3), and [F₅Xe(FH)SbF₆] (4).

Compound	1	2	3	4
Space group	P21/n	P21/c	P2/c	P21/c
a [Å]	9.03170(10)	9.3142(5)	12.2989(4)	6.3279(2)
b [Å]	9.7065(2)	8.0482(4)	6.4853(2)	15.2663(4)
c [Å]	12.1261(2)	16.1545(8)	10.6717(3)	8.8234(2)
β [o]	106.8920(10)	95.942(3)	106.317(2)	92.6730(10)
V [Å3]	1017.18(3)	1204.48(11)	816.91(4)	851.45(4)
Z	4	4	4	4
M W [g mol−1]	537.02	658.82	435.23	482.06
D calcd [g cm−3]	3.507	3.633	3.539	3.761
T [°C]	−173	−173	−173	−173
μ [mm−1]	11.101	13.062	8.411	7.319
R 1 [a]	0.0288	0.0306	0.0228	0.0195
wR 2 [b]	0.0561	0.0568	0.0452	0.0415

[a] R ₁=Σ∥F _o|−|F _c∥/Σ|F _o|. [b] wR ₂=[Σ(w(F _o ²−F _c ²)²)/Σ(w(F _o ²)²)]^1/2.

[F ₅ Xe(FKrF)AsF ₆ ] (1) and [F ₅ Xe(FKrF) ₂ AsF ₆ ] (2). The [XeF₅]⁺ cations are coordinated to four F atoms to give xenon coordination numbers, CN_Xe=5 + 4: one secondary bond from a KrF₂ ligand and three from three [AsF₆]⁻ anions in [F₅Xe(FKrF)AsF₆] (1) (Figure 1 a; Supporting Information, Figures S1 and S2), whereas in the case of [F₅Xe(FKrF)₂AsF₆] (2), two secondary bonds are from two KrF₂ ligands and two are from two [AsF₆]⁻ anions (Figure 1 b; Figures S3 and S4). The coordination spheres of the [XeF₅]⁺ cations in 1 and 2 are similar to their known xenon analogues, [F₅Xe(FXeF)AsF₆], ^[19] and [F₅Xe(FXeF)₂AsF₆]. ^[19] In the latter cases, the longer Xe‐ ‐ ‐F_As secondary bonds (1:1, 2.59, 3.03, and 3.15 Å; 1:2, 2.95 and 3.57 Å) are shorter than or equal to the sum of the Xe and F van der Waals radii (3.63, ^[23] 3.52 ^[24] Å). Although the crystal structure of [F₅Xe(FKrF)AsF₆] is isotypic with its xenon analogue, [F₅Xe(FKrF)₂AsF₆] is not.

Figure 1.

The X‐ray crystal structure of a) [F₅Xe(FKrF)AsF₆] (1) and b) [F₅Xe(FKrF)₂AsF₆] (2) where the coordination environments of the Xe atom are expanded to include symmetry‐generated atoms (symmetry codes: (1) (i) $1/2$ −x, y–$1/2$ , $1/2$ −z; (ii) 1−x, 1−y, 1−z; (2) (i) x, y+1, z). Thermal ellipsoids are drawn at the 50 % probability level.

The trajectories of the four Xe‐ ‐ ‐F secondary bonds in 1 and 2 avoid the Xe−F_eq bond pair and valence electron lone pair (VELP) domains of the square‐pyramidal [XeF₅]⁺ cation, where the lone pair lies on the pseudo C ₄‐axis and is trans to the F_ax atom of [XeF₅]⁺. The [AsF₆]⁻ anions of 1 are mer‐coordinated to three different [XeF₅]⁺ cations by means of asymmetric secondary Xe‐ ‐ ‐F_As bonds, where the cis‐Xe‐ ‐ ‐F_As bond is notably shorter (2.5944(10) Å) than the trans‐Xe‐ ‐ ‐F_As bonds (2.9147(10), 3.0572(11) Å). The three Xe‐ ‐ ‐F_As secondary bonds result in the layered structure depicted in Figure S1b. The [AsF₆]⁻ anions of 2 are asymmetrically trans‐coordinated to two [XeF₅]⁺ cations (Xe‐ ‐ ‐F_As, 2.812(2), 3.124(2) Å), which form chains that run parallel to the b‐axis of the unit cell (Figures S3b and S4). The KrF₂ ligands coordinate to [XeF₅]⁺ by means of secondary Xe‐ ‐ ‐F_b bonds that are shorter ((1) 2.5139(9) Å; (2) 2.550(2), 2.576(2) Å) than the secondary Xe‐ ‐ ‐F_As bonds of the coordinated [AsF₆]⁻ anions (Table 2; Supporting Information, Tables S1 and S2). The Kr−F_t and Kr−F_b bond asymmetry is somewhat more pronounced in 1 (1.8393(12), 1.9367(9) Å) than in 2 (1.845(2), 1.927(2) Å and 1.851(2), 1.917(2) Å), which is attributed to stronger and shorter Xe‐ ‐ ‐F_b interactions in 1 than in 2. Similar Kr−F_t and Kr−F_b bond length asymmetries are observed in the crystal structures of [FO₂Br(FKrF)₂AsF₆] (1.840(5) and 1.847(4) Å, 1.943(4) and 1.933(4) Å), ^[8] [Mg(FKrF)₄(AsF₆)₂] (1.817(2)–1.821(2) Å, 1.965(1)– 1.979(1) Å), ^[10] [Hg(FKrF)₈]²⁺ (1.822(1)–1.852(1) Å, 1.933(1)–1.957(1) Å), ^[11] and KrF₂⋅CrOF₄ (1.8489(9) and 1.9279(9) Å). ^[12] Regardless of their Kr−F bond asymmetries, the average Kr−F_t/b bond lengths ((1) 1.888(2) Å; (2) 1.886(2) and 1.884(2) Å) are comparable to the Kr−F bond lengths of α‐KrF₂ (1.894(5) Å) ^[15] and symmetrically bridged KrF₂ in KrF₂⋅2CrOF4 (1.8881(6) Å). ^[12] The F−Kr−F bond angles are essentially linear ((1) 178.49(6)°; (2) 178.47(8) and 179.40(7)°), whereas the Kr−F_b‐ ‐ ‐Xe angles are bent ((1) 133.24(5)°; (2) 137.40(8) and 141.80(7)°), as observed in all other KrF₂ adducts.[ ⁸ , ⁹ , ¹⁰ , ¹¹ , ¹² ] The Kr−F_b‐ ‐ ‐Xe angles of 1 and 2 are similar to the Kr−F_b‐ ‐ ‐Br angles of [FO₂Br(FKrF)₂AsF₆] (132.1(2) and 139.9(2)°) ^[8] and are intermediate with respect to the range of Kr−F_b‐ ‐ ‐Mg angles observed for [Mg(FKrF)₄(AsF₆)₂] (121.84(7)–144.43(8)°). ^[10]

Table 2.

Selected experimental bond lengths for [F₅Xe(FKrF)AsF₆] (1) and [F₅Xe(FKrF)₂AsF₆] (2); and calculated^[a] bond lengths and Wiberg bond indices (WBI) for [F₅Xe(FKrF)(AsF₆)₃]²⁻ (1′) and [F₅Xe(FKrF)₂(AsF₆)₂]⁻ (2′).

	1	1′	1′	2	2′	2′
	Bond lengths [Å]		WBI	Bond lengths [Å]		WBI
Xe−Fax	1.8067(11)	1.919	0.550	1.813(2)	1.911	0.566
Xe−Feq	1.8394(12)	1.910	0.596	1.8371(14)	1.912	0.595
	1.8404(12)	1.910	0.595	1.8418(14)	1.903	0.607
	1.8455(13)	1.909	0.601	1.8449(14)	1.909	0.598
	1.8462(12)	1.910	0.600	1.8482(14)	1.925	0.597
Xe‐ ‐ ‐Fb	2.5139(9)	2.500	0.102	2.550(2) 2.576(2)	2.784 2.626	0.037 0.069
Xe‐ ‐ ‐FAs	2.5944(10) 2.9147(10) 3.0572(11)	2.468 2.802 2.828	0.098 0.031 0.028	2.812(2) 3.124(2)	2.480 2.563	0.096 0.068
Kr−Ft	1.8393(12)	1.860	0.610	1.845(2) 1.851(2)	1.862 1.874	0.603 0.576
Kr−Fb	1.9367(9)	1.939	0.450	1.917(2) 1.927(2)	1.909 1.930	0.502 0.464

[a] APFD/aVDZ(‐PP)(Kr, Xe, As)/aVDZ(F).

[F ₅ Xe(FH)AsF ₆ ] (3). The asymmetric unit in the crystal structure of 3 is comprised of two [XeF₅]⁺ cations located on special positions, and an [AsF₆]⁻ anion and an HF molecule located on general positions (Figure 2; Figures S5 and S6). The Xe^VI atoms have CN_Xe=5 + 4, where one [XeF₅]⁺ cation has four Xe‐ ‐ ‐F_As secondary bonds originating from the coordination of two pairs of symmetry‐related [AsF₆]⁻ anions (2.647(2), 3.058(2) Å) whereas the other [XeF₅]⁺ cation interacts with two symmetry‐related [AsF₆]⁻ anions through two secondary Xe‐ ‐ ‐F_As bonds (2.930(2) Å) and with two symmetry‐related HF ligands through two short secondary Xe‐ ‐ ‐F_H bonds (2.656(2) Å) (Table S3). Each HF molecule is also H‐bonded to two neighboring [AsF₆]⁻ anions with F_H⋅⋅⋅F_As distances of 2.545(2) Å, where the As−F bond of the H‐bonded F ligand (1.746(2) Å) is the second longest As−F bond of the [AsF₆]⁻ anion. Each [AsF₆]⁻ anion also coordinates to three [XeF₅]⁺ cations in a mer‐arrangement where the As−F bonds of the interacting fluorine atoms are slightly elongated (1.719(2), 1.727(2), 1.749(2) Å) with respect to the two non‐interacting axial As−F bonds (1.701(2), 1.705(2) Å). The secondary Xe‐ ‐ ‐F_As and Xe‐ ‐ ‐F_H bonds result in corrugated layers that are parallel to the ac‐plane and are stacked along the b‐axis of the unit cell.

Figure 2.

The crystal structure of [F₅Xe(FH)AsF₆] (3). The coordination environment of Xe1 is expanded to include symmetry‐generated atoms (symmetry codes: (i) −x, y, $1/2$ −z). Thermal ellipsoids are drawn at the 50 % probability level; hydrogen atoms are shown as spheres of arbitrary radius. The coordination environment of Xe2 is shown in Figure S5a.

The Xe‐ ‐ ‐F_H secondary bonds of [F₅Xe(FH)PnF₆] (As, 2.656(2) Å; Sb, 2.6501(10) Å) are similar, but are significantly greater than those of [F₃Xe(FH)Sb₂F₁₁] (2.462(2) Å) ^[25] and [FXe(FH)Sb₂F₁₁] (2.359(4) Å), ^[26] in accordance with the lower Lewis acidity of [XeF₅]⁺ relative to [XeF₃]⁺ and [XeF]⁺.[ ¹⁰ , ¹³ , ¹⁹ ]

A brief description of the crystal structure of the nonisotypic antimony analogue (4) is provided in the Supporting Information along with associated X‐ray data (Figures S7 and S8, Table S3). The LT Raman spectrum of [F₅Xe(FH)AsF₆] (3) was also acquired (Figure S9, Table S4).

Raman Spectroscopy

The LT solid‐state Raman spectra of [F₅Xe(FKrF)AsF₆] (1) and [F₅Xe(FKrF)₂AsF₆] (2) are depicted in Figure 3. Vibrational assignments for 1 were initially made by comparison with the calculated frequencies and assignments of gas‐phase [F₅Xe(FKrF)AsF₆] (1′′) (Table S5). Although this model accounts for the majority of experimental frequencies and intensities, several differences occur for modes that mainly involve [AsF₆]⁻ anion displacements. This is expected because the coordination sphere of [XeF₅]⁺ in the gas‐phase [F₅Xe(FKrF)AsF₆] model (1′′, CN_Xe=5 + 3; Figure S10) differs from that of the solid‐state structure 1 (CN_Xe=5 + 4). An alternative gas‐phase model, [F₅Xe(FKrF)(AsF₆)₃]²⁻ (1′, CN_Xe=5 + 4, Figure 4 a), addresses these differences and better reproduces the xenon coordination environment of [XeF₅]⁺ by coordination of two additional [AsF₆]⁻ anions to the [XeF₅]⁺ cation. The gas‐phase [F₅Xe(FKrF)₂(AsF₆)₂]⁻ model (2′, Figure 4 b), which well reproduces the coordination environment of the [XeF₅]⁺ cation in 2, was used to aid in the assignment of the Raman spectrum of 2. The vibrational assignments of [XeF₅]⁺ and [AsF₆]⁻ in 1 and 2 were also aided by comparisons with those of [XeF₅][AsF₆], ^[27] [XeF₅][BF₄], ^[28] [XeF₅][fac‐OsO₃F₃], ^[29] [XeF₅][μ‐F(OsO₃F₂)₂], ^[29] [XeF₅]₂[Cr₂O₂F₈], ^[30] [XeF₅][Xe₂F₁₁][CrOF₅]⋅2CrOF₄, ^[30] [XeF₅] [M₂O₂F₉] (M=Mo, W), ^[17] [F₂OBr(FKrF)₂AsF₆], ^[8] and [Mg(FKrF)₄(AsF₆)₂]. ^[10] The experimental and calculated frequencies, their detailed assignments, and mode descriptions for 1 and 2 are provided in Tables S5 and S6, respectively. The vibrational frequencies and intensities of the gas‐phase KrF₂ molecule were also calculated (Table S7) in order to estimate the degree to which the calculated frequencies of coordinated KrF₂ are over‐ or underestimated in 1′, 1′′, and 2′. The experimental vibrational frequencies and their trends are well reproduced by the calculated frequencies, with the exception of their ν(Kr−F) stretching frequencies, which are overestimated.

Figure 3.

The Raman spectra of a) [F₅Xe(FKrF)AsF₆] (1) and b) [F₅Xe(FKrF)₂AsF₆] (2) recorded at −144 and −161 °C, respectively, using 1064‐nm excitation. The spectrum of (2) also shows bands due to (1), which are indicated by bullets (•) (Table S6, footnote c). Symbols denote FEP sample tube bands (*) and an instrumental artifact (†).

Figure 4.

Calculated geometries [APFD/aVDZ(‐PP)(Kr, Xe, As)/aVDZ(F)] for a) [F₅Xe(FKrF)(AsF₆)₃]²⁻ (1′) and b) [F₅Xe(FKrF)₂(AsF₆)₂]⁻ (2′).

[F ₅ Xe(FKrF)AsF ₆ ] (1). Loss of the center of symmetry upon coordination of a fluorine atom of KrF₂ to Xe of [XeF₅]⁺ results in the appearance of two distinct stretching bands in the Raman spectrum that are assigned to ν(Kr−F_b) and ν(Kr−F_t). The calculated vibrational displacements of 1, show no significant intraligand coupling between the Kr−F_b and Kr−F_t stretching modes of coordinated KrF₂ ligands (Table S5). This contrasts with the KrF₂ ligands of KrF₂⋅CrOF₄, ^[12] which exhibit intraligand coupling between the ν(Kr−F_b) and ν(Kr−F_t) modes.

The most intense band in the Raman spectrum of 1 at 454 cm⁻¹ (calcd, 491 cm⁻¹) is assigned to the ν(Kr−F_b) stretching mode. As predicted, the ν(Kr−F_t) stretching band corresponding to the shorter Kr−F_t bond occurs at higher frequency, 533 cm⁻¹ (calcd, 585 cm⁻¹). The experimental frequencies of ν(Kr−F_b) and ν(Kr−F_t) bracket that of free KrF₂ (464 cm⁻¹), and are comparable to those of [F₂OBr(FKrF)₂AsF₆] (443/472 and 533/549 cm⁻¹). ^[8] The observed frequencies are in accordance with the experimental Kr−F bond length trend (Tables 2 and S1), with a similar trend observed for [F₅Xe(FXeF)AsF₆] (433 and 559 cm⁻¹). ^[19]

The degeneracy of the ν₂(Π_u) bending mode of free KrF₂ is removed upon coordination, which results in Raman‐active δ(F_tKrF_b)_i.p. and δ(F_tKrF_b)_o.o.p. modes that bend in‐plane and out‐of‐plane with respect to the XeF_bKrF_t‐plane. The calculated out‐of‐plane bend couples with the two ρ_w(F_eqXeF_eq) wagging modes of [XeF₅]⁺, whereas the in‐plane bend is not coupled. Both bands are predicted to have low relative Raman intensities and were observed as weak bands at 294 and 255 cm⁻¹ (calcd, 289 and 264 cm⁻¹), respectively. Both bands are shifted to higher frequency relative to ν₂(Π_u) of free KrF₂ (232.6 cm⁻¹), ^[31] but have frequencies that are comparable to the corresponding modes of [F₂OBr(FKrF)₂AsF₆] (301 and 254/266 cm⁻¹). ^[8] The bands assigned to the ρ_r(F_tKrF_b) rocking and ρ_t(F_tKrF_b) torsional modes are predicted at 152 and 136 cm⁻¹ and were observed at 143 and 130 cm⁻¹, respectively. Interestingly, and similar to the δ(F_tKrF_b) bending modes, the out‐of‐plane torsional mode, ρ_t(F_tKrF_b), also couples with the ρ_w(F_eqXeF_eq) wagging modes of the cation, whereas the in‐plane rocking mode, ρ_r(F_tKrF_b), does not couple. The δ(XeF_bKr) and δ(XeF_bAAs) bends are predicted at very low frequencies, 60 and 71/74 cm⁻¹, respectively, but could not be observed.

[F ₅ Xe(FKrF) ₂ AsF ₆ ]. Coordination of a second KrF₂ ligand to [XeF₅]⁺ results in additional splitting on the Kr−F stretching bands of the KrF₂ ligands that are due to intra‐ and interligand couplings. The bands at 543 and 564/567 cm⁻¹ (calcd, 585, 598 cm⁻¹) are respectively assigned to modes that are predominantly coupled ν(Kr−F_t) stretching modes, {[ν(Kr₁−F₁₂) − ν(Kr₂−F₁₄)] − [ν(Kr₁−F₁₃) − ν(Kr₂−F₁₅)]_small} and {[ν(Kr₁−F₁₂) + ν(Kr₂−F₁₄)] − [ν(Kr₁−F₁₃) + ν(Kr₂−F₁₅)]_small} (Table S6). Similar couplings, which are exclusively interligand couplings, also occur in other KrF₂ adducts that contain more than one NgF₂ ligand.[ ⁸ , ⁹ , ¹⁰ , ¹¹ ] The bands at 466 and 472/474 cm⁻¹ (calcd, 507, 517 cm⁻¹) are assigned to modes that are predominantly coupled ν(Kr−F_b) stretching modes, [ν(Kr₁−F₁₃) + ν(Kr₁−F₁₂)_small] and [ν(Kr₂−F₁₅) + ν(Kr₂−F₁₄)_small], respectively. Interestingly, and unlike adducts which contain more than one NgF₂ ligand, there are no interligand couplings among the ν(Kr−F_b) stretching modes. The room‐temperature Raman spectrum of the xenon analogue, [F₅Xe(FXeF)₂AsF₆], also displays split ν(Xe−F_b) (420/438, 479 cm⁻¹) and ν(Xe−F_t) (542, 550 cm⁻¹) bands which are likely due to vibrational mode coupling. ^[19]

Vibrational coupling between the KrF₂ ligands results in two out‐of‐plane, δ(F_tKrF_b)_o.o.p. (calcd, 246, 247 cm⁻¹), and two in‐plane, δ(F_tKrF_b)_i.p. (calcd, 275, 280 cm⁻¹), bends which occur at 251 (o.o.p.) and 273/278 (i.p.) cm⁻¹. The bands at 145 and 110 cm⁻¹, are assigned to the in‐plane ρ_r(F_tKrF_b) rocking mode (calcd, 136 cm⁻¹) and the out‐of‐plane ρ_t(F_tKrF_b) torsional mode (calcd, 119 cm⁻¹), respectively.

Computational Results

The gas‐phase geometries of [F₅Xe(FKrF)AsF₆] (1′′) (Figure S10), the hypothetical model anions, [F₅Xe(FKrF)(AsF₆)₃]²⁻ (1′) and [F₅Xe(FKrF)₂(AsF₆)₂]⁻ (2′) (Figure 4; Figure S11), KrF₂, [XeF₅]⁺, and IF₅ were optimized with all frequencies real at the APFD/aVDZ(‐PP)(Xe, As, Kr)/aVDZ(F) level of theory (Tables S1, S2, S5–S8). The crystallographic coordinates were used as the starting geometries for the geometry optimizations. A limitation of the gas‐phase structural models used for 1′ and 2′ is the isolated nature of the ion‐pairs, which contrast with the extended (layer and chain) structures observed in the crystal structures of 1 and 2. However, both models reproduce the coordination environment of xenon and therefore proved useful for the assignment of the Raman spectra and provided insights into the secondary bonding interactions among [XeF₅]⁺ and coordinated KrF₂ and [AsF₆]⁻.

Calculated Geometries

[F ₅ Xe(FKrF)AsF ₆ ] (1′′) and [F ₅ Xe(FKrF)(AsF ₆ ) ₃ ] ²⁻ (1 ′). The [F₅Xe(FKrF)AsF₆] (1′′) ion‐pair was initially calculated, but resulted in twisting of the [AsF₆]⁻ anion such that it coordinated in a bidentate fashion through two cis‐fluorine ligands to the Xe atom to give CN_Xe=5 + 3 (Figure S10). In contrast, the [F₅Xe(FKrF)(AsF₆)₃]²⁻ model (1′) reproduced the experimental Xe coordination sphere (CN_Xe=5 + 4) and better reproduced the Xe‐ ‐ ‐F_b interactions and their avoidance of the Xe VELP and Xe−F_eq bond pair domains.

The calculated Kr−F_t bond length (1.860 Å) is shorter than the Kr−F_b bond length (1.939 Å), as observed in the crystal structure (1.8393(12) and 1.9367(9) Å), and the average calculated Kr−F bond length (1.900 Å) is very similar to the calculated (1.889 Å) and experimental (1.894(5) Å) ^[15] bond lengths of free KrF₂. The calculated F_t−Kr−F_b bond angle (176.7°) is in good agreement with the experimental value (178.49(6)°), whereas the Kr−F_b‐ ‐ ‐Xe bond angle (121.2°) is significantly smaller than the experimental value (133.24(5)°). The difference between the calculated and experimental Kr−F_b‐ ‐ ‐Xe angles is likely due to crystal packing and the deformability of this angle (δ(XeF_bKr), 60 cm⁻¹).

The calculated Xe‐ ‐ ‐F_b/As contact distances (2.500, 2.468, 2.802, and 2.828 Å) are underestimated relative to their experimental values (2.5139(9), 2.5944(10), 2.9147(10), and 3.0572(11) Å), but reproduce the alternation of their long and short Xe‐ ‐ ‐F secondary bonds in 1. The shorter calculated contact distances are accompanied by large F_ax−Xe‐ ‐ ‐F_b/As contact angles (142.1 and 141.1°) whereas long contact distances are accompanied by smaller F_ax−Xe‐ ‐ ‐F_b/As contact angles (128.9 and 128.5°), in very good agreement with the corresponding angles in 1 (143.54(6) and 146.48(6)°; 129.90(6) and 124.67(6)°).

[F ₅ Xe(FKrF) ₂ (AsF ₆ ) ₂ ] ⁻ (2′). The calculated geometrical parameters of the KrF₂ ligands reproduce the experimental values and trends in 2, i.e., the shorter Kr−F_t bond (calcd, 1.862 Å; exptl, 1.845(2) Å) is accompanied by a longer Kr−F_b bond (calcd, 1.930 Å; exptl, 1.927(2) Å) for one KrF₂ ligand, and a longer Kr−F_t bond (calcd, 1.874 Å; exptl, 1.851(2) Å) is accompanied by a shorter Kr−F_b bond (calcd, 1.909 Å; exptl, 1.917(2) Å) for the other KrF₂ ligand. The near‐linear F_t−Kr−F_b angles (178.47(8) and 179.40(7)°) of 2 are also reproduced (177.60 and 176.98°), but the Xe‐ ‐ ‐F_b−Kr angles (137.40(8) and 141.80(7)°) are significantly underestimated (123.99 and 124.71°). The difference is likely attributable to the absence of a secondary Xe‐ ‐ ‐F_As bond in 2′ that is trans to the bridging As−F bond in 2. This results in bridging As−F (1.7285(14) and 1.7433(14) Å) and Xe‐ ‐ ‐F_As (2.812(2) and 3.124(2) Å) bond lengths that are overestimated (1.834 and 1.847 Å) and underestimated (2.480 and 2.563 Å), respectively, and F_ax−Xe‐ ‐ ‐F_As angles (126.57(6) and 132.26(6)°) that are overestimated (139.81 and 140.53°). The Xe‐ ‐ ‐F_b bond lengths (2.550(2) and 2.576(2) Å) are overestimated (2.784 and 2.626 Å) and the F_ax−Xe‐ ‐ ‐F_b angles (139.59(7) and 140.35(7)°) are underestimated (128.26 and 130.55°). As observed in 2, there are two groups of alternating short and long calculated Xe‐ ‐ ‐F_b/As secondary bonds whose domains avoid the Xe VELP and Xe−F_eq bond pair domains.

Natural Bond Orbital (NBO) Analyses

The NBO analyses for [F₅Xe(FKrF)(AsF₆)₃]²⁻ (1′) and [F₅Xe(FKrF)₂(AsF₆)₂]⁻ (2′) (Table S9) show the total positive charges on the [XeF₅]⁺ cations of 1′ (0.776) and 2′ (0.771) are notably less than the net positive charge of the uncoordinated [XeF₅]⁺ cation, and are consistent with charge transfer from the KrF₂ ligand (1′, 0.092 and 2′, 0.092) and [AsF₆]⁻ (1′, −2.871 and 2′, −1.864). Charge transfer mainly affects the F_ax atoms of 1′ (−0.490) and 2′ (−0.475), which are significantly more negative relative to F_ax of free [XeF₅]⁺ (−0.384). In contrast the average NPA charges of the F_eq atoms of 1′ (−0.483) and 2′ (−0.486) are much closer to the F_eq charges of free [XeF₅]⁺ (−0.447). The F_b and F_t charges of KrF₂ in 1′ (−0.513, F_b and −0.453, F_t) and 2′ (−0.526, F_b and −0.456, F_t; −0.514, F_b and −0.475, F_t) bracket that of free KrF₂ (−0.492). The charge distribution is consistent with an axially distorted KrF₂ ligand in which partial removal of the bridging fluorine atom by the Lewis acidic [XeF₅]⁺ cation results in more KrF⁺ character and correspondingly shorter Kr−F_t and longer Kr−F_b bonds (Tables S1 and S2). The Kr−F_t and Kr−F_b Wiberg bond indices of 1′ (Kr−F_t, 0.610 and Kr−F_b, 0.450) and 2′ (Kr−F_t, 0.603 and Kr−F_b, 0.464; Kr−F_t, 0.576 and Kr−F_b, 0.502) bracket those of KrF₂ (0.551). The small Xe−F_b bond indices of 1′ (0.102) and 2′ (0.069 and 0.037), and the low degree of charge transfer from KrF₂ to [XeF₅]⁺ are consistent with predominantly electrostatic secondary bonding interactions between the Xe and the F_b atom(s) of the KrF₂ ligand(s) and the long experimental and calculated Xe‐ ‐ ‐F_b bonds observed in 1, 2, 1′, and 2′ (Tables S1 and S2). The NBO analyses of [AsF₆]⁻ in 1′ and 2′ show that the larger Xe‐ ‐ ‐F_As bond indices of 1′ (0.098) and 2′ (0.096) correspond to smaller As−F bond indices of 1′ (0.386) and 2′ (0.382). This is in accordance with the shorter Xe‐ ‐ ‐F_As bonds (calcd, 1′: 2.468 Å; 2′: 2.480 Å; exptl, 1: 2.5944(10) Å; 2: 2.812(2) Å) and correspondingly longer As−F bridge bonds (calcd, 1′: 1.844 Å; 2′: 1.847 Å; exptl, 1: 1.7559(10) Å; 2: 1.7433(14) Å).

Electron Localization Function (ELF) Analyses

The Electron Localization Function analyses[ ³² , ³³ ] were carried out for [F₅Xe(FKrF)(AsF₆)₃]²⁻, [F₅Xe(FKrF)₂(AsF₆)₂]⁻, KrF₂, [XeF₅]⁺, and isoelectronic IF₅. The abbreviations in the ensuing discussion denote electron localization function (η(r)); core basin (C(Ng), C(As)); monosynaptic valence basins (V(F) and V(Ng)); and f, a localization domain that is bounded by the isosurface, η(r)=f. The ELF isosurface plots for the aforementioned species at η(r)=0.55 are depicted in Figure 5 and Figure S12.

Figure 5.

ELF isosurface plots [APFD/aVDZ(‐PP)(Kr, Xe, As)/aVDZ(F)], η(r)=0.55, for a) [XeF₅]⁺, b) [F₅Xe(FKrF)(AsF₆)₃]²⁻, and c) [F₅Xe(FKrF)₂(AsF₆)₂]⁻. Color code: core basin (red); C(Ng), C(As); monosynaptic valence basin (blue); V(F), V(Ng).

The ELF analyses of [F₅Xe(FKrF)(AsF₆)₃]²⁻, [F₅Xe(FKrF)₂(AsF₆)₂]⁻, KrF₂, [XeF₅]⁺, and IF₅ display only monosynaptic Xe, Kr, As, F, and I valence basins in accordance with the polar‐covalent characters of their bonds. The toroidal shapes of the Kr valence basins result from the combination of the three valence electron lone‐pair (VELP) domains of Kr, with the atomic core electron basin (C(Kr)) lying at the center of the torus. The perturbations of the toroidal V(Kr) basin of [F₅Xe(FKrF)(AsF₆)₃]²⁻ and one of the toroidal V(Kr) basins of [F₅Xe(FKrF)₂(AsF₆)₂]⁻ arise from accommodation of the V(Kr) basins to their immediate environments (Figure 5). In both instances, the krypton valence basin torus of one KrF₂ ligand is flattened parallel to the KrF₂ molecular axis because the ligand is sandwiched between the fluorine valence basins of neighboring [AsF₆]⁻ ions (Figure 5). In contrast, the second KrF₂ ligand of [F₅Xe(FKrF)₂(AsF₆)₂]⁻ is less sterically congested which results in a V(Kr) basin that is essentially unperturbed, closely resembling the toroidal V(Kr) valence basin of free KrF₂ (Figure S12). Small perturbations of the toroidal V(Ng) basins (Ng=Kr, Xe) of NgF₂⋅CrOF₄ and NgF₂⋅2 CrOF₄ have also been noted and attributed to the asymmetries of their immediate environments. ^[12]

The valence basins, V(Xe) and V(I) of Xe and I in the isoelectronic [XeF₅]⁺ cation and IF₅, correspond to stereoactive electron lone‐pairs, where the [XeF₅]⁺ VELP (2.14 ų) is significantly contracted relative to that of IF₅ (3.08 ų), in accordance with the higher charge on Xe (3.17) of [XeF₅]⁺ relative to that of I (2.90) in IF₅ (also see MEPS analyses).

Notable differences occur between the Xe VELP distributions of the adduct‐cations, where the VELP volumes and shapes accommodate to the spaces provided by the neighboring V(F) basins of the KrF₂ ligands and [AsF₆]⁻ ions. The Xe VELPs of [F₅Xe(FKrF)(AsF₆)₃]²⁻ and [F₅Xe(FKrF)₂(AsF₆)₂]⁻ are sterically more congested in their nine‐ coordinate Xe environments, where the Xe VELPs are notably flattened and their volumes (0.44 and 0.46 ų, respectively) are significantly reduced with respect to those of [XeF₅]⁺ and IF₅ (vide supra). Similar steric influences on the Xe^VI VELP volume have been noted for the series, XeF₆ (C _3v), F₆XeNCCH₃, and F₆Xe(NCCH₃)₂. ^[34]

Molecular Electrostatic Potential Surface (MEPS) Analyses

The MEPS isosurfaces of [XeF₅]⁺ and isoelectronic IF₅ are depicted in Figure 6. Their isosurfaces have regions of high EP (Xe, 773 kJ mol⁻¹ and I, 228 kJ mol⁻¹), which are located trans to their F_ax atoms. The xenon atom is significantly more electrophilic than the iodine atom, and the MEPS maxima of the fluorine ligand isosurfaces of [XeF₅]⁺ are significantly more positive than those of IF₅ which have small negative values, in accordance with their NPA charges (Table S9).

Figure 6.

The molecular electrostatic potential surface (MEPS) contours calculated at the 0.001 e⋅bohr⁻³ isosurfaces of [XeF₅]⁺ and IF₅ and the top 5 % of the positive electrostatic potential range (bottom left). The extrema of selected electrostatic potentials are indicated by arrows. The optimized geometries and MEPS were calculated at the APFD/aVDZ(‐PP)(Xe, I)/ aVDZ(F) level of theory.

Examination of the top 5 % of the positive EP ranges in [XeF₅]⁺ and IF₅ (Figure 6) allowed the visualization of four regions of higher EP on the xenon and iodine MEPS isosurfaces (Xe, 798 kJ mol⁻¹ and I, 236 kJ mol⁻¹) that are located at the intersections of the F_eq isosurfaces. Similar regions have been reported for IF₅ and XeF₄. ^[35] These regions are symmetrically disposed with respect to the xenon and iodine VELPs, which are trans to F_ax atoms of [XeF₅]⁺ and IF₅. The experimental and calculated trajectories of the secondary Xe‐ ‐ ‐F_b and Xe‐ ‐ ‐F_As bonds in 1 (Figure S1), 2 (Figure S3), 1′ and 2′ (Figure S11) are staggered with respect to the F_eq atoms of [XeF₅]⁺ and avoid the xenon VELP, in accordance with the calculated positions of the four EP maxima on xenon. The crystal structure of XeF₂⋅IF₅ exhibits similar features, i.e., four I‐ ‐ ‐F_b secondary bonds whose trajectories are staggered with respect to the I−F_eq bond domains of IF₅ and avoid the VELP domain of iodine. ^[36] The electrostatic nature of the secondary Xe‐ ‐ ‐F_b bonds is also supported by the Wiberg bond indices obtained for 1′ and 2′ (Table S9), and may be ascribed to σ‐hole bonding. The MEPS of XeO₃ (C _3v) also show discrete regions of higher EP on the xenon MEPS isosurface, which were visualized by examination of the top 20 % of the xenon MEPS isosurface. ^[37] In contrast with [XeF₅]⁺ and IF₅, three regions of higher EP of XeO₃ are located trans to the highly electronegative oxygen atoms of the primary Xe−O bonds, a characteristic of σ‐hole bonding, ^[38] and have contact trajectories that are staggered with respect to these bonds.

Conclusion

The present study provides the first instances where both chemically bound krypton and xenon are present in the same compound. The [F₅Xe(FKrF)AsF₆] and [F₅Xe(FKrF)₂AsF₆] complexes have been isolated in macroscopic quantities and structurally characterized by X‐ray crystallography and Raman spectroscopy. Their syntheses, which significantly extend the limited chemistry of krypton and the XeF₂–KrF₂ analogy, provide a new class of coordination complex in which KrF₂ coordinates through a fluorine atom to Xe^VI of the [XeF₅]⁺ cation. The stabilities of these complexes are reliant on the Lewis acidity of [XeF₅]⁺ and its resistance to oxidation by the potent oxidative fluorinator, KrF₂. NBO, ELF, and MEPS analyses demonstrate that the bonding interactions between the fluorine bridge atom of KrF₂ and the Lewis acidic xenon atom are essentially noncovalent and may be ascribed to σ‐hole bonding. The HF solvates, [F₅Xe(FH)PnF₆] (Pn=As, Sb), also characterized in this study, provide rare examples of HF coordinated to Xe^VI.

Experimental Section

Cautionary statements relating to the safe handling of XeF₆, KrF₂, and [XeF₅]⁺ salts are provided in the Supporting Information. Details relating to the apparatus, starting materials, syntheses, low‐temperature crystal mounting, X‐ray data collection and refinement, Raman spectroscopy, and computational details are provided in the Supporting Information. Details of the crystal structure investigations may be obtained from the joint CCDC/FIZ Karlsruhe online deposition service by quoting the deposition numbers CSD 2000547 (1), 2000548 (2), 2000549 (3), and 2000550 (4).

Conflict of interest

The authors declare no conflict of interest.

Supporting information

As a service to our authors and readers, this journal provides supporting information supplied by the authors. Such materials are peer reviewed and may be re‐organized for online delivery, but are not copy‐edited or typeset. Technical support issues arising from supporting information (other than missing files) should be addressed to the authors.

Supplementary

Supplementary

M. Lozinšek, H. P. A. Mercier, G. J. Schrobilgen, Angew. Chem. Int. Ed. 2021, 60, 8149.