Supporting information
Crystallographic Information File (CIF) https://doi.org/10.1107/S0108270105037200/jz1769sup1.cif | |
Structure factor file (CIF format) https://doi.org/10.1107/S0108270105037200/jz1769Isup2.hkl |
CCDC reference: 296354
Please give brief details of synthesis, including quantities of reagents and reaction conditions, or a reference to such a description.
H atoms were placed in calculated positions and were treated as riding atoms, with C—H distances of 0.98 (CH3), 0.99 (CH2) and 0.95 Å (CH); methyl groups were allowed to rotate freely.
Room-temperature ionic liquids are attracting significant interest owing to their interesting chemical characteristics and potentially useful solvent properties (Rogers & Seddon, 2002, 2003; Wasserscheid & Welton, 2002). Since ionic liquids from the widely used imidazolium salts can retain some definite local `structure' despite their homogeneous appearance, they potentially offer a unique opportunity to obtain direct and quantitative structural information of the fluid state (Bowron et al., 2003).
Although crystalline halide salts of 1-methyl-3-butylimidazolium (BMIM+) have been characterized (Holbrey, Reichert, Nieuwenhuyzen, Johnston et al., 2003; Saha et al., 2003), they are structurally compromised for the direct comparison of the liquid and solid structures by the presence of substantial hydrogen-bonding forces. By contrast, imidazolium salts of the weakly coordinating hexafluorophosphate (PF6-) are less prone to participate in such extra hydrogen-bonding interactions and are thus more likely to lead to local interionic interactions that resemble those in liquid-state structures. Indeed, recent small-angle X-ray scattering (SAXS) studies of BMIM+·PF6-, (I), have provided an interesting semi-quantitative view of the structural inhomogeneity that can persist in ionic liquids (Billard et al., 2003). Two peaks corresponding to the interplanar imidazolium distance of 4.4 Å and centroid-to-centroid distance of 6.3 Å were observed, and the coherence distance of 15 Å was attributed to the extent of local order in the fluid state – beyond this limited resolution, no definitive structural information could be gleaned. Furthermore, the surface-layering effect has also been studied by X-ray reflectivity (XR), and several possible structures of the BMIM+·PF6- surface have been proposed (Solutskin et al., 2005).
We now report the successful crystallization of (I) for a definitive X-ray crystallographic structure, in which a liquid sample of (I) was initially supercooled at 243 K and crystalline material was obtained by shock-induced crystallization. Many alternative attempts at the crystallization of this highly hygroscopic and low-melting salt have been unsuccessful because its melting point of 284 K is severely depressed by the trace presence of water and by other organic solvents. The entire bulk of the sample crystallizes spontaneously. Given this fact, we assume that the structure of such a crystalline material is very much akin to the structure of this ionic liquid in its fluid state.
The structure of (I) is shown in Fig. 1. The planar imidazolium ring is highly delocalized, as evidenced by the C—N and C—C bond distances; the N1—C2 and N2—C2 bonds of 1.3226 (18) and 1.3259 (18) Å, respectively, are shorter than N1—C3 and N2—C4 bonds of 1.3751 (17) and 1.3781 (17) Å, respectively, and the C3—C4 bond length is 1.346 (2) Å. Atoms C1 and C5 lie in the plane of the imidazolium ring (mean deviation is 0.0153 Å). The butyl chain, in which all four C atoms lie in a plane (mean deviation is 0.0121 Å), propagates in the all-trans conformation. This chain is twisted with respect to the C5—N2 bond, as indicated by the N2—C5—C6—C7 torsion angle of -60.80 (17)°. The butyl group makes an interplanar angle of 82.3 (s.u.?)° with the plane of the imidazolium ring.
It is interesting to compare the cation packing of (I) with that of its short- and long-chain analogs. The series of 1-methyl-3-alkylimidazolium salts are arbitrarily categorized as either short-chained, the alkyl being methyl (Holbrey, Reichert, Nieuwenhuyzen, Sheppard et al., 2003) or ethyl (Fuller et al., 1994), or long-chained, the alkyl being dodecyl (Gordon et al., 1998) or tetradecyl (De Roche et al., 2003), all with melting points above room temperature and readily crystallized. X-ray crystallographic studies show that structures in short-chain ionic liquids are dominated by electrostatic cation/anion interactions that lead to salts with high lattice energy and higher melting points. In long-chain ionic liquids, the significant hydrophobic alkyl chain/chain interactions govern their higher melting points. The n-butyl analog (I) is the exception, the intermediate length of the alkyl chain disrupting the strong cation/anion interactions but being insufficient for chain–chain effects to be dominant. As such, (I) is the only structurally characterized imidazolium hexafluorophosphate that remains liquid at room temperature.
The supramolecular structure of imidazolium rings of (I) shown in Fig. 2 represents a hybrid of both short-chain and long-chain structures. Two types of imidazolium interplanar arrangements are found; in one set the butyl groups point into the space between the imidazolium rings (solid box) and in the other they point outwards (dashed box). The first motif is somewhat similar to those in the short-chain analogs, where cations arranged in discrete pairs. In (I), these pairs propagate continuously. The centers of corresponding imidazolium rings of these pairs are shifted by 4.31 (s.u.?) Å. The imidazolium–imidazolium interplanar distance of these pairs is 4.3378 (15) Å, with a centroid-to-centroid separation of 6.11 (s.u.?) Å. The pairs of imidazolium rings with butyl chains pointing outwards forms a set of planes similar to those found in the long-chain analogs. The alkyl chains from neighboring cations form interpenetrating regions along the crystallographic b axis with about 4 Å separation between the chains (Fig. 3), but this alkyl chain arrangement does not have a pronounced effect on the crystal structure of (I). Unlike in the corresponding dodecyl and tetradecyl hexafluorophosphates, the imidazolium rings in (I) are all parallel, with a 3.4618 (14) Å interplanar distance and a 4.86 (s.u.?) Å centroid-to-centroid separation. The rings are also shifted with respect to each other by 3.41 (s.u.?) Å. Because of these shifts, there are no close π–π interactions between the imidazolium rings. The hexafluorophosphate anions are positioned close to the imidazolium rings and lie in the channels between the imidazolium rings and butyl chains, as shown in Fig. 3.
The PF6- anion forms several H···F contacts that are less than the sum of the van der Waals radii of H (1.2 Å) and F atoms (1.5 Å). All H···F contacts with C—H···F angles greater than 90° that match this criterion are listed in Table 1. The shortest contact is the H2···F1 distance of 2.36 Å, which is the shortest hydrogen–fluorine distance so far found in linear alkyl imidazolium hexafluorophosphate ionic liquids (the corresponding distances for the methyl, ethyl, dodecyl and tetradecyl analogs are 2.635, 2.576, 2.578 and 2.471 Å, respectively). Nonetheless, it represents a very weak hydrogen bond according to published criteria (Hitchcock et al., 1993; Gordon et al., 1998; Desiraju, 2004).
With the detailed structural information of (I) in hand, it is now possible to draw conclusions about the structure of this ionic liquid in its fluid phase. Close examination of the crystal structure suggests that the previous structural assignment was only partially correct. The plane-to-plane and ring-to-ring solid-state distances of 4.3378 (15) Å and 6.11 (s.u.?) Å may be safely correlated with the 4.4 Å and 6.3 Å peaks observed by the SAXS method (Billard et al., 2003). Concomitantly, the coherence distance of 15 Å is better described by two repetitions of alternating 4.3378 (15) and 3.4618 (14) Å imidazolium–imidazolium interplanar distances, giving a value of 15.6 Å (Fig. 2). It is also interesting to compare the electron density of the bulk BMIM+·PF6- liquid (with ρ= 0.42 e Å-3) and its surface (Solutskin et al., 2005) (with ρ = 0.50 e Å-3) to the electron density of the crystalline sample (with ρ = 0.48 e Å-3. It seems plausible to suppose that structural features of the crystalline sample are also retained in the surface of a liquid sample to an even greater degree than in the bulk liquid.
In summary, the detailed structure of the widely used and extensively studied room-temperature ionic liquid (I) has been obtained. Given the precise X-ray crystal data, we are now able to interpret the structural data obtained previously by small angle X-ray scattering and reflectivity on liquid samples and to elucidate the structure of the ionic liquid (I) to a greater degree of certainty. By comparing the electron densities of the surface layer and the crystalline sample, it is clear that the structural information obtained will greatly facilitate the elucidation of the structural features of this material not only in the bulk but also at the liquid surface.
Data collection: SMART (Bruker, 2003); cell refinement: SAINT (Bruker, 2003); data reduction: SAINT and SADABS (Bruker, 2003); program(s) used to solve structure: SHELXTL (Bruker, 2003); program(s) used to refine structure: SHELXTL; molecular graphics: XP (Bruker, 1999); software used to prepare material for publication: SHELXTL and XCIF (Bruker, 1999).
C8H15N2+·PF6− | Z = 2 |
Mr = 284.19 | F(000) = 292 |
Triclinic, P1 | Dx = 1.560 Mg m−3 |
Hall symbol: -P 1 | Melting point: 6C K |
a = 8.7549 (12) Å | Mo Kα radiation, λ = 0.71073 Å |
b = 8.9042 (12) Å | Cell parameters from 9189 reflections |
c = 9.0128 (13) Å | θ = 2.5–30.0° |
α = 95.810 (3)° | µ = 0.28 mm−1 |
β = 114.955 (2)° | T = 173 K |
γ = 103.061 (3)° | Plate, colorless |
V = 604.96 (15) Å3 | 0.50 × 0.18 × 0.01 mm |
Bruker SMART diffractometer | 3486 independent reflections |
Radiation source: fine-focus sealed tube | 2840 reflections with I > 2σ(I) |
Graphite monochromator | Rint = 0.020 |
ω scans | θmax = 30.0°, θmin = 2.4° |
Absorption correction: multi-scan (SADABS; Bruker, 2003) | h = −12→12 |
Tmin = 0.889, Tmax = 1.000 | k = −12→12 |
9189 measured reflections | l = −12→12 |
Refinement on F2 | Primary atom site location: structure-invariant direct methods |
Least-squares matrix: full | Secondary atom site location: difference Fourier map |
R[F2 > 2σ(F2)] = 0.043 | Hydrogen site location: inferred from neighbouring sites |
wR(F2) = 0.125 | H-atom parameters constrained |
S = 1.05 | w = 1/[σ2(Fo2) + (0.0749P)2 + 0.0074P] where P = (Fo2 + 2Fc2)/3 |
3487 reflections | (Δ/σ)max = 0.001 |
156 parameters | Δρmax = 0.41 e Å−3 |
0 restraints | Δρmin = −0.27 e Å−3 |
C8H15N2+·PF6− | γ = 103.061 (3)° |
Mr = 284.19 | V = 604.96 (15) Å3 |
Triclinic, P1 | Z = 2 |
a = 8.7549 (12) Å | Mo Kα radiation |
b = 8.9042 (12) Å | µ = 0.28 mm−1 |
c = 9.0128 (13) Å | T = 173 K |
α = 95.810 (3)° | 0.50 × 0.18 × 0.01 mm |
β = 114.955 (2)° |
Bruker SMART diffractometer | 3486 independent reflections |
Absorption correction: multi-scan (SADABS; Bruker, 2003) | 2840 reflections with I > 2σ(I) |
Tmin = 0.889, Tmax = 1.000 | Rint = 0.020 |
9189 measured reflections |
R[F2 > 2σ(F2)] = 0.043 | 0 restraints |
wR(F2) = 0.125 | H-atom parameters constrained |
S = 1.05 | Δρmax = 0.41 e Å−3 |
3487 reflections | Δρmin = −0.27 e Å−3 |
156 parameters |
x | y | z | Uiso*/Ueq | ||
P1 | 0.18444 (5) | 0.79488 (4) | 0.44901 (4) | 0.02785 (12) | |
F1 | 0.35521 (12) | 0.93965 (11) | 0.57091 (12) | 0.0410 (2) | |
F2 | 0.29779 (13) | 0.67516 (11) | 0.50853 (13) | 0.0490 (3) | |
F3 | 0.12957 (14) | 0.78229 (15) | 0.59432 (13) | 0.0539 (3) | |
F4 | 0.07314 (13) | 0.91488 (12) | 0.38757 (13) | 0.0464 (3) | |
F5 | 0.24066 (15) | 0.81036 (14) | 0.30242 (13) | 0.0544 (3) | |
F6 | 0.01489 (13) | 0.65083 (12) | 0.32406 (13) | 0.0511 (3) | |
N1 | 0.75199 (15) | 0.72952 (14) | 0.65019 (14) | 0.0270 (3) | |
N2 | 0.78894 (16) | 0.83706 (14) | 0.89274 (15) | 0.0277 (3) | |
C1 | 0.6931 (2) | 0.6937 (2) | 0.46853 (18) | 0.0403 (4) | |
H1A | 0.5839 | 0.7216 | 0.4106 | 0.061* | |
H1B | 0.6715 | 0.5806 | 0.4294 | 0.061* | |
H1C | 0.7846 | 0.7551 | 0.4444 | 0.061* | |
C2 | 0.70855 (18) | 0.83227 (17) | 0.72967 (18) | 0.0282 (3) | |
H2 | 0.6322 | 0.8927 | 0.6783 | 0.034* | |
C3 | 0.86449 (19) | 0.66566 (18) | 0.76617 (19) | 0.0317 (3) | |
H3 | 0.9163 | 0.5890 | 0.7436 | 0.038* | |
C4 | 0.8872 (2) | 0.73249 (18) | 0.91775 (18) | 0.0323 (3) | |
H4 | 0.9578 | 0.7115 | 1.0224 | 0.039* | |
C5 | 0.7659 (2) | 0.92927 (17) | 1.02293 (19) | 0.0352 (3) | |
H5A | 0.7116 | 1.0112 | 0.9774 | 0.042* | |
H5B | 0.8828 | 0.9834 | 1.1198 | 0.042* | |
C6 | 0.6500 (2) | 0.82430 (17) | 1.08181 (17) | 0.0307 (3) | |
H6A | 0.7070 | 0.7448 | 1.1303 | 0.037* | |
H6B | 0.6420 | 0.8901 | 1.1721 | 0.037* | |
C7 | 0.4647 (2) | 0.73904 (19) | 0.94394 (18) | 0.0335 (3) | |
H7A | 0.4717 | 0.6697 | 0.8554 | 0.040* | |
H7B | 0.4083 | 0.8179 | 0.8927 | 0.040* | |
C8 | 0.3510 (2) | 0.6397 (2) | 1.0089 (2) | 0.0418 (4) | |
H8A | 0.4061 | 0.5612 | 1.0594 | 0.063* | |
H8B | 0.2336 | 0.5856 | 0.9156 | 0.063* | |
H8C | 0.3400 | 0.7084 | 1.0936 | 0.063* |
U11 | U22 | U33 | U12 | U13 | U23 | |
P1 | 0.0262 (2) | 0.0302 (2) | 0.0242 (2) | 0.01010 (15) | 0.00859 (15) | 0.00349 (14) |
F1 | 0.0314 (5) | 0.0347 (5) | 0.0449 (5) | 0.0065 (4) | 0.0112 (4) | −0.0037 (4) |
F2 | 0.0416 (6) | 0.0359 (5) | 0.0559 (6) | 0.0192 (4) | 0.0069 (5) | 0.0070 (4) |
F3 | 0.0496 (6) | 0.0787 (8) | 0.0408 (6) | 0.0161 (6) | 0.0276 (5) | 0.0193 (5) |
F4 | 0.0408 (6) | 0.0453 (6) | 0.0534 (6) | 0.0246 (5) | 0.0153 (5) | 0.0136 (5) |
F5 | 0.0623 (7) | 0.0740 (8) | 0.0387 (6) | 0.0241 (6) | 0.0323 (5) | 0.0107 (5) |
F6 | 0.0351 (5) | 0.0396 (5) | 0.0504 (6) | 0.0042 (4) | 0.0005 (5) | −0.0039 (4) |
N1 | 0.0280 (6) | 0.0320 (6) | 0.0238 (6) | 0.0110 (5) | 0.0124 (5) | 0.0098 (5) |
N2 | 0.0307 (6) | 0.0281 (6) | 0.0265 (6) | 0.0086 (5) | 0.0149 (5) | 0.0079 (5) |
C1 | 0.0500 (10) | 0.0503 (10) | 0.0240 (7) | 0.0207 (8) | 0.0165 (7) | 0.0104 (7) |
C2 | 0.0280 (7) | 0.0313 (7) | 0.0308 (7) | 0.0127 (6) | 0.0154 (6) | 0.0128 (6) |
C3 | 0.0319 (7) | 0.0318 (7) | 0.0340 (8) | 0.0157 (6) | 0.0136 (6) | 0.0117 (6) |
C4 | 0.0331 (7) | 0.0337 (7) | 0.0289 (7) | 0.0135 (6) | 0.0100 (6) | 0.0130 (6) |
C5 | 0.0461 (9) | 0.0285 (7) | 0.0324 (8) | 0.0078 (6) | 0.0219 (7) | 0.0026 (6) |
C6 | 0.0362 (8) | 0.0320 (7) | 0.0238 (7) | 0.0095 (6) | 0.0144 (6) | 0.0039 (5) |
C7 | 0.0338 (8) | 0.0416 (8) | 0.0265 (7) | 0.0160 (7) | 0.0130 (6) | 0.0073 (6) |
C8 | 0.0324 (8) | 0.0503 (10) | 0.0421 (9) | 0.0098 (7) | 0.0189 (7) | 0.0060 (7) |
P1—F3 | 1.5795 (10) | C3—C4 | 1.346 (2) |
P1—F4 | 1.5916 (9) | C3—H3 | 0.9500 |
P1—F6 | 1.5979 (10) | C4—H4 | 0.9500 |
P1—F2 | 1.5980 (9) | C5—C6 | 1.520 (2) |
P1—F1 | 1.6000 (9) | C5—H5A | 0.9900 |
P1—F5 | 1.6005 (10) | C5—H5B | 0.9900 |
N1—C2 | 1.3226 (18) | C6—C7 | 1.517 (2) |
N1—C3 | 1.3751 (17) | C6—H6A | 0.9900 |
N1—C1 | 1.4682 (18) | C6—H6B | 0.9900 |
N2—C2 | 1.3259 (18) | C7—C8 | 1.522 (2) |
N2—C4 | 1.3781 (17) | C7—H7A | 0.9900 |
N2—C5 | 1.4738 (18) | C7—H7B | 0.9900 |
C1—H1A | 0.9800 | C8—H8A | 0.9800 |
C1—H1B | 0.9800 | C8—H8B | 0.9800 |
C1—H1C | 0.9800 | C8—H8C | 0.9800 |
C2—H2 | 0.9500 | ||
F3—P1—F4 | 90.17 (6) | C4—C3—N1 | 106.93 (13) |
F3—P1—F6 | 91.03 (6) | C4—C3—H3 | 126.5 |
F4—P1—F6 | 89.64 (6) | N1—C3—H3 | 126.5 |
F3—P1—F2 | 90.85 (6) | C3—C4—N2 | 107.19 (12) |
F4—P1—F2 | 178.96 (6) | C3—C4—H4 | 126.4 |
F6—P1—F2 | 90.55 (5) | N2—C4—H4 | 126.4 |
F3—P1—F1 | 90.18 (6) | N2—C5—C6 | 111.48 (12) |
F4—P1—F1 | 90.24 (5) | N2—C5—H5A | 109.3 |
F6—P1—F1 | 178.78 (6) | C6—C5—H5A | 109.3 |
F2—P1—F1 | 89.55 (5) | N2—C5—H5B | 109.3 |
F3—P1—F5 | 179.17 (6) | C6—C5—H5B | 109.3 |
F4—P1—F5 | 89.25 (6) | H5A—C5—H5B | 108.0 |
F6—P1—F5 | 89.56 (6) | C7—C6—C5 | 113.87 (12) |
F2—P1—F5 | 89.73 (6) | C7—C6—H6A | 108.8 |
F1—P1—F5 | 89.23 (6) | C5—C6—H6A | 108.8 |
C2—N1—C3 | 108.73 (12) | C7—C6—H6B | 108.8 |
C2—N1—C1 | 125.26 (12) | C5—C6—H6B | 108.8 |
C3—N1—C1 | 125.99 (13) | H6A—C6—H6B | 107.7 |
C2—N2—C4 | 108.31 (12) | C6—C7—C8 | 112.30 (12) |
C2—N2—C5 | 125.63 (12) | C6—C7—H7A | 109.1 |
C4—N2—C5 | 125.90 (12) | C8—C7—H7A | 109.1 |
N1—C1—H1A | 109.5 | C6—C7—H7B | 109.1 |
N1—C1—H1B | 109.5 | C8—C7—H7B | 109.1 |
H1A—C1—H1B | 109.5 | H7A—C7—H7B | 107.9 |
N1—C1—H1C | 109.5 | C7—C8—H8A | 109.5 |
H1A—C1—H1C | 109.5 | C7—C8—H8B | 109.5 |
H1B—C1—H1C | 109.5 | H8A—C8—H8B | 109.5 |
N1—C2—N2 | 108.83 (12) | C7—C8—H8C | 109.5 |
N1—C2—H2 | 125.6 | H8A—C8—H8C | 109.5 |
N2—C2—H2 | 125.6 | H8B—C8—H8C | 109.5 |
C3—N1—C2—N2 | 0.01 (16) | C2—N2—C4—C3 | 0.28 (17) |
C1—N1—C2—N2 | −178.40 (14) | C5—N2—C4—C3 | 175.93 (13) |
C4—N2—C2—N1 | −0.18 (16) | C2—N2—C5—C6 | 103.77 (16) |
C5—N2—C2—N1 | −175.85 (12) | C4—N2—C5—C6 | −71.15 (19) |
C2—N1—C3—C4 | 0.17 (17) | N2—C5—C6—C7 | −60.80 (17) |
C1—N1—C3—C4 | 178.56 (15) | C5—C6—C7—C8 | −178.02 (13) |
N1—C3—C4—N2 | −0.27 (17) |
D—H···A | D—H | H···A | D···A | D—H···A |
C2—H2···F1 | 0.95 | 2.36 | 3.2264 (17) | 152 |
C2—H2···F5i | 0.95 | 2.58 | 3.1721 (19) | 121 |
C3—H3···F6ii | 0.95 | 2.43 | 3.3653 (19) | 167 |
C4—H4···F5iii | 0.95 | 2.56 | 3.4008 (18) | 147 |
C5—H5B···F4iii | 0.99 | 2.52 | 3.2917 (19) | 135 |
C1—H1B···F2ii | 0.98 | 2.45 | 3.329 (2) | 150 |
C8—H8C···F5iv | 0.98 | 2.55 | 3.499 (2) | 163 |
Symmetry codes: (i) −x+1, −y+2, −z+1; (ii) −x+1, −y+1, −z+1; (iii) x+1, y, z+1; (iv) x, y, z+1. |
Experimental details
Crystal data | |
Chemical formula | C8H15N2+·PF6− |
Mr | 284.19 |
Crystal system, space group | Triclinic, P1 |
Temperature (K) | 173 |
a, b, c (Å) | 8.7549 (12), 8.9042 (12), 9.0128 (13) |
α, β, γ (°) | 95.810 (3), 114.955 (2), 103.061 (3) |
V (Å3) | 604.96 (15) |
Z | 2 |
Radiation type | Mo Kα |
µ (mm−1) | 0.28 |
Crystal size (mm) | 0.50 × 0.18 × 0.01 |
Data collection | |
Diffractometer | Bruker SMART |
Absorption correction | Multi-scan (SADABS; Bruker, 2003) |
Tmin, Tmax | 0.889, 1.000 |
No. of measured, independent and observed [I > 2σ(I)] reflections | 9189, 3486, 2840 |
Rint | 0.020 |
(sin θ/λ)max (Å−1) | 0.703 |
Refinement | |
R[F2 > 2σ(F2)], wR(F2), S | 0.043, 0.125, 1.05 |
No. of reflections | 3487 |
No. of parameters | 156 |
H-atom treatment | H-atom parameters constrained |
Δρmax, Δρmin (e Å−3) | 0.41, −0.27 |
Computer programs: SMART (Bruker, 2003), SAINT (Bruker, 2003), SAINT and SADABS (Bruker, 2003), SHELXTL (Bruker, 2003), XP (Bruker, 1999), SHELXTL and XCIF (Bruker, 1999).
D—H···A | D—H | H···A | D···A | D—H···A |
C2—H2···F1 | 0.95 | 2.36 | 3.2264 (17) | 152 |
C2—H2···F5i | 0.95 | 2.58 | 3.1721 (19) | 121 |
C3—H3···F6ii | 0.95 | 2.43 | 3.3653 (19) | 167 |
C4—H4···F5iii | 0.95 | 2.56 | 3.4008 (18) | 147 |
C5—H5B···F4iii | 0.99 | 2.52 | 3.2917 (19) | 135 |
C1—H1B···F2ii | 0.98 | 2.45 | 3.329 (2) | 150 |
C8—H8C···F5iv | 0.98 | 2.55 | 3.499 (2) | 163 |
Symmetry codes: (i) −x+1, −y+2, −z+1; (ii) −x+1, −y+1, −z+1; (iii) x+1, y, z+1; (iv) x, y, z+1. |
Room-temperature ionic liquids are attracting significant interest owing to their interesting chemical characteristics and potentially useful solvent properties (Rogers & Seddon, 2002, 2003; Wasserscheid & Welton, 2002). Since ionic liquids from the widely used imidazolium salts can retain some definite local `structure' despite their homogeneous appearance, they potentially offer a unique opportunity to obtain direct and quantitative structural information of the fluid state (Bowron et al., 2003).
Although crystalline halide salts of 1-methyl-3-butylimidazolium (BMIM+) have been characterized (Holbrey, Reichert, Nieuwenhuyzen, Johnston et al., 2003; Saha et al., 2003), they are structurally compromised for the direct comparison of the liquid and solid structures by the presence of substantial hydrogen-bonding forces. By contrast, imidazolium salts of the weakly coordinating hexafluorophosphate (PF6-) are less prone to participate in such extra hydrogen-bonding interactions and are thus more likely to lead to local interionic interactions that resemble those in liquid-state structures. Indeed, recent small-angle X-ray scattering (SAXS) studies of BMIM+·PF6-, (I), have provided an interesting semi-quantitative view of the structural inhomogeneity that can persist in ionic liquids (Billard et al., 2003). Two peaks corresponding to the interplanar imidazolium distance of 4.4 Å and centroid-to-centroid distance of 6.3 Å were observed, and the coherence distance of 15 Å was attributed to the extent of local order in the fluid state – beyond this limited resolution, no definitive structural information could be gleaned. Furthermore, the surface-layering effect has also been studied by X-ray reflectivity (XR), and several possible structures of the BMIM+·PF6- surface have been proposed (Solutskin et al., 2005).
We now report the successful crystallization of (I) for a definitive X-ray crystallographic structure, in which a liquid sample of (I) was initially supercooled at 243 K and crystalline material was obtained by shock-induced crystallization. Many alternative attempts at the crystallization of this highly hygroscopic and low-melting salt have been unsuccessful because its melting point of 284 K is severely depressed by the trace presence of water and by other organic solvents. The entire bulk of the sample crystallizes spontaneously. Given this fact, we assume that the structure of such a crystalline material is very much akin to the structure of this ionic liquid in its fluid state.
The structure of (I) is shown in Fig. 1. The planar imidazolium ring is highly delocalized, as evidenced by the C—N and C—C bond distances; the N1—C2 and N2—C2 bonds of 1.3226 (18) and 1.3259 (18) Å, respectively, are shorter than N1—C3 and N2—C4 bonds of 1.3751 (17) and 1.3781 (17) Å, respectively, and the C3—C4 bond length is 1.346 (2) Å. Atoms C1 and C5 lie in the plane of the imidazolium ring (mean deviation is 0.0153 Å). The butyl chain, in which all four C atoms lie in a plane (mean deviation is 0.0121 Å), propagates in the all-trans conformation. This chain is twisted with respect to the C5—N2 bond, as indicated by the N2—C5—C6—C7 torsion angle of -60.80 (17)°. The butyl group makes an interplanar angle of 82.3 (s.u.?)° with the plane of the imidazolium ring.
It is interesting to compare the cation packing of (I) with that of its short- and long-chain analogs. The series of 1-methyl-3-alkylimidazolium salts are arbitrarily categorized as either short-chained, the alkyl being methyl (Holbrey, Reichert, Nieuwenhuyzen, Sheppard et al., 2003) or ethyl (Fuller et al., 1994), or long-chained, the alkyl being dodecyl (Gordon et al., 1998) or tetradecyl (De Roche et al., 2003), all with melting points above room temperature and readily crystallized. X-ray crystallographic studies show that structures in short-chain ionic liquids are dominated by electrostatic cation/anion interactions that lead to salts with high lattice energy and higher melting points. In long-chain ionic liquids, the significant hydrophobic alkyl chain/chain interactions govern their higher melting points. The n-butyl analog (I) is the exception, the intermediate length of the alkyl chain disrupting the strong cation/anion interactions but being insufficient for chain–chain effects to be dominant. As such, (I) is the only structurally characterized imidazolium hexafluorophosphate that remains liquid at room temperature.
The supramolecular structure of imidazolium rings of (I) shown in Fig. 2 represents a hybrid of both short-chain and long-chain structures. Two types of imidazolium interplanar arrangements are found; in one set the butyl groups point into the space between the imidazolium rings (solid box) and in the other they point outwards (dashed box). The first motif is somewhat similar to those in the short-chain analogs, where cations arranged in discrete pairs. In (I), these pairs propagate continuously. The centers of corresponding imidazolium rings of these pairs are shifted by 4.31 (s.u.?) Å. The imidazolium–imidazolium interplanar distance of these pairs is 4.3378 (15) Å, with a centroid-to-centroid separation of 6.11 (s.u.?) Å. The pairs of imidazolium rings with butyl chains pointing outwards forms a set of planes similar to those found in the long-chain analogs. The alkyl chains from neighboring cations form interpenetrating regions along the crystallographic b axis with about 4 Å separation between the chains (Fig. 3), but this alkyl chain arrangement does not have a pronounced effect on the crystal structure of (I). Unlike in the corresponding dodecyl and tetradecyl hexafluorophosphates, the imidazolium rings in (I) are all parallel, with a 3.4618 (14) Å interplanar distance and a 4.86 (s.u.?) Å centroid-to-centroid separation. The rings are also shifted with respect to each other by 3.41 (s.u.?) Å. Because of these shifts, there are no close π–π interactions between the imidazolium rings. The hexafluorophosphate anions are positioned close to the imidazolium rings and lie in the channels between the imidazolium rings and butyl chains, as shown in Fig. 3.
The PF6- anion forms several H···F contacts that are less than the sum of the van der Waals radii of H (1.2 Å) and F atoms (1.5 Å). All H···F contacts with C—H···F angles greater than 90° that match this criterion are listed in Table 1. The shortest contact is the H2···F1 distance of 2.36 Å, which is the shortest hydrogen–fluorine distance so far found in linear alkyl imidazolium hexafluorophosphate ionic liquids (the corresponding distances for the methyl, ethyl, dodecyl and tetradecyl analogs are 2.635, 2.576, 2.578 and 2.471 Å, respectively). Nonetheless, it represents a very weak hydrogen bond according to published criteria (Hitchcock et al., 1993; Gordon et al., 1998; Desiraju, 2004).
With the detailed structural information of (I) in hand, it is now possible to draw conclusions about the structure of this ionic liquid in its fluid phase. Close examination of the crystal structure suggests that the previous structural assignment was only partially correct. The plane-to-plane and ring-to-ring solid-state distances of 4.3378 (15) Å and 6.11 (s.u.?) Å may be safely correlated with the 4.4 Å and 6.3 Å peaks observed by the SAXS method (Billard et al., 2003). Concomitantly, the coherence distance of 15 Å is better described by two repetitions of alternating 4.3378 (15) and 3.4618 (14) Å imidazolium–imidazolium interplanar distances, giving a value of 15.6 Å (Fig. 2). It is also interesting to compare the electron density of the bulk BMIM+·PF6- liquid (with ρ= 0.42 e Å-3) and its surface (Solutskin et al., 2005) (with ρ = 0.50 e Å-3) to the electron density of the crystalline sample (with ρ = 0.48 e Å-3. It seems plausible to suppose that structural features of the crystalline sample are also retained in the surface of a liquid sample to an even greater degree than in the bulk liquid.
In summary, the detailed structure of the widely used and extensively studied room-temperature ionic liquid (I) has been obtained. Given the precise X-ray crystal data, we are now able to interpret the structural data obtained previously by small angle X-ray scattering and reflectivity on liquid samples and to elucidate the structure of the ionic liquid (I) to a greater degree of certainty. By comparing the electron densities of the surface layer and the crystalline sample, it is clear that the structural information obtained will greatly facilitate the elucidation of the structural features of this material not only in the bulk but also at the liquid surface.