Intricacies of ligand coordination in tricarbonylchromium(0) complexes with ortho- and para-fluoro­biphenyls

Solid-state geometries of [(η⁶-C₆H₅)(2-F-C₆H₄)]Cr(CO)₃ and [(η⁶-C₆H₅)(4-F-C₆H₄)]Cr(CO)₃ indicate exclusive coordination of the metal center to the more electron-rich arene ring. Unlike the structures of the free ligands, the chromium complexes exhibit no crystallographic symmetry and no positional disorder.

Abstract

The steric and electronic factors that influence which of the two rings of a substituted biphenyl ligand coordinates to chromium are of interest and it has been suggested that haptotropic rearrangements within these mol­ecules may be limited if the arene–arene dihedral angle is too large. Two tricarbonylchromium(0) complexes and their respective free ligands have been characterized by single-crystal X-ray diffraction. In the solid state, tricarbon­yl[(1′,2′,3′,4′,5′,6′-η)-2-fluoro-1,1′-biphen­yl]chromium(0), [Cr(C₁₂H₉F)(CO)₃], (I), exists as the more stable isomer with the non­halogenated arene ring ligated to the metal center. Similarly, tricarbon­yl[(1′,2′,3′,4′,5′,6′-η)-4-fluoro-1,1′-biphen­yl]chromium(0) crys­tallizes as the more stable isomer with the phenyl ring bonded to the Cr⁰ center. The arene–arene dihedral angles in these complexes are 55.77 (4) and 52.4 (5)°, respectively. Structural features of these complexes are compared to those of the DFT-optimized geometries of ten tricarbon­yl[(η⁶-C₆H₅)(4-F-C₆H₄)]chromium model complexes. The solid-state structures of the free ligands 2-fluoro-1,1′-biphenyl and 4-fluoro-1,1′-biphenyl, both C₁₂H₉F, exhibit arene–arene dihedral angles of 54.83 (7) and 0.71 (8)°, respectively. The mol­ecules of the free ligands occupy crystallographic twofold axes and exhibit positional disorder. Weak inter­molecular C—H⋯F inter­actions are observed in all four structures.

Introduction  

In our ongoing study of the syntheses, structures and haptotropic rearrangements of substituted (biphen­yl)tri­carbonyl­chromium compounds, we have become especially inter­ested in steric and electronic factors that influence which of the two rings of the substituted biphenyl ligand coordinates to chromium. We have also focused our attention on the arene–arene dihedral angles in these complexes, since it has been suggested that haptotropic rearrangements within these mol­ecules may be limited if the arene–arene dihedral angle is too large. A study describing a series of substituted (biphen­yl)tri­carbonyl­chromium compounds has suggested that a necessary condition for inter-ring haptotropic rearrangement is an arene–arene dihedral angle less than 22°, as determined either by calculation or by single-crystal X-ray crystallography (Oprunenko, 2000 ▸). Inter-ring steric hindrance effected by a nonhydrogen substituent in the ortho position has been shown to twist the arene rings out of coplanarity, in some cases, to a great enough degree to prevent the metal from migrating between rings.

We report here the solid-state structures of the isomeric com­pounds tricarbon­yl[(1′,2′,3′,4′,5′,6′-η)-2-fluoro-1,1′-bi­phenyl]chromium(0), [(η⁶-C₆H₅)(2-F-C₆H₄)]Cr(CO)₃, (I), and tri­car­bon­yl[(1′,2′,3′,4′,5′,6′-η)-4-fluoro-1,1′-biphen­yl]chromium(0), [(η⁶-C₆H₅)(4-F-C₆H₄)]Cr(CO)₃, (II), and of their respective free ligands 2-fluoro-1,1′-biphenyl or ortho-F-C₆H₄-Ph, (III), and 4-fluoro-1,1′-biphenyl or para-F-C₆H₄-Ph, (IV). Our intention in preparing the metal complexes was first to explore how the steric and electronic effects of a fluorine substituent on the biphenyl ligand would influence ring selectivity of chromium coordination. Regarding steric effects, it seemed reasonable to expect that during the syntheses of (biphen­yl)tri­carbonyl­chromium compounds, steric hindrance imparted by multiple substituents on one of the biphenyl rings might decrease the propensity of that ring to coordinate to chromium. We wondered if the small size of fluorine might reduce this effect. Regarding electronic effects, it was believed that a ring that bears a highly electronegative substituent would be electron-poor and less likely to coordinate to chromium. Although fluorine is highly electronegative, there are cases where it is known to act as a π-donor. Some examples of this are indicated by the ortho/para-directing nature of fluorine in electrophilic aromatic substitution (Carey & Giuliano, 2014 ▸), the −0.03 value of the σ⁻ parameter used for fluorine in the Hammett equation (Ritchie, 1964 ▸), and fluorine π-donor characteristics that have been described in structures such as (η⁶-C₆H₅F)Cr(CO)₃ (Zeller et al., 2004 ▸).

Experimental  

Synthesis and crystallization  

Compounds (I) and (II) were synthesized according to the method of Mahaffy & Pauson (1990 ▸) by heating a butyl ether/tetra­hydro­furan solution of hexa­carbonyl­chromium for 24 h with 2-fluoro­biphenyl in the case of (I) and with 4-fluoro­biphenyl in the case of (II). Details of the synthesis, purification, and spectroscopic characterization are provided below. Recrystallization from a 3:1 (v/v) hexa­ne–ether solution of (I) or (II) led to the formation of single crystals suitable for crystallographic analysis. The free ligands ortho-F-C₆H₄-Ph, (III), and para-F-C₆H₄-Ph, (IV), were purchased from Acros Organics and MilliporeSigma; their crystals were grown by sublimation in a round-bottomed flask. NMR spectra were obtained on a Bruker AC300 spectrometer (¹H NMR at 300 MHz and ¹³C NMR at 75 MHz) and a Bruker Avance III 400 spectrometer (¹H NMR at 400 MHz and ¹³C NMR at 100 MHz). Solution IR spectra were recorded on a MIDAC Prospect-IR PRS-102 using CsI solution cells.

Preparation of [(η⁶-C₆H₅)(2-F-C₆H₄)]Cr(CO)₃, (I)  

A mixture of hexa­carbonyl­chromium (0.510 g, 2.32 mmol) and 2-fluoro­biphenyl (0.390 g, 2.26 mmol) in butyl ether (30 ml) and tetra­hydro­furan (3 ml) was degassed and heated under reflux for 24 h under a nitro­gen atmosphere. Tetra­hydro­furan, butyl ether, and the remaining unreacted hexa­carbonyl­chromium were removed by vacuum distillation. Purification of the remaining yellow residue by column chromatography [silica gel, 2:1 (v/v) hexane–diethyl ether] gave a yellow band from which (I) was isolated as a yellow air-stable solid (yield 0.188 g, 27.0%; m.p. 417–420 K). ¹H NMR (CDCl₃, 300 MHz): δ 7.08–7.52 (4H, m, ar­yl), 5.88 (2H, d, ³ J _HH = 6.1 Hz, metal-bound ar­yl), 5.61 (3H, m, metal-bound ar­yl). ¹³C{¹H} NMR (CDCl₃, 75 MHz): δ 218.1 (CO), 161.0 (¹ J _CF = 250.1 Hz), 130.9 (² J _CF = 21.7 Hz), 130.7 (³ J _CF = 9.2 Hz), 130.4 (³ J _CF = 9.2 Hz), 124.9 (⁴ J _CF = 3.9 Hz), 116.5 (² J _CF = 21.7 Hz), 104.9, 94.7, 92.3, 91.5. IR (CH₂Cl₂, CsI): 1970, 1895 cm⁻¹.

Preparation of [(η⁶-C₆H₅)(2-F-C₆H₄)]Cr(CO)₃, (II)  

A mixture of hexa­carbonyl­chromium (0.495 g, 2.25 mmol) and 4-fluoro­biphenyl (0.385 g, 2.24 mmol) in butyl ether (30 ml) and tetra­hydro­furan (3 ml) was degassed and heated under reflux for 24 h under a nitro­gen atmosphere. Tetra­hydro­furan, butyl ether, and the remaining unreacted hexa­carbonyl­chromium were removed by vacuum distillation. Purification of the remaining yellow residue by column chromatography [silica gel, 2:1 (v/v) hexa­ne/diethyl ether] gave a yellow band from which (II) was isolated as a yellow air-stable solid (yield 0.151 g, 21.9%; m.p. 386–389 K). ¹H NMR (CDCl₃, 400 MHz): δ 7.48 (dd, ³ J _HH = 7.8, ⁴ J _HF = 5.3 Hz, 2H, ar­yl), 7.11 (dd, ³ J _HF = 8.3, ³ J _HH = 7.8 Hz, 2H, ar­yl), 5.63 (2H, d, ³ J _HH = 6.1 Hz, metal-bound ar­yl), 5.50 (2H, tr, ³ J _HH = 6.1 Hz, metal-bound ar­yl), 5.34 (1H, tr, ³ J _HH = 6.1 Hz, metal-bound ar­yl). ¹³C{¹H} NMR (CDCl₃, 100 MHz): δ 232.6 (CO), 163.2 (¹ J _CF = 249.2 Hz), 132.6 (⁴ J _CF = 2.8 Hz), 128.9 (³ J _CF = 8.5 Hz), 115.9 (² J _CF = 21.9 Hz), 109.6, 92.7, 91.9, 91.3. IR (CH₂Cl₂, CsI): 1972, 1900 cm⁻¹.

Refinement  

Crystal data, data collection and structure refinement details are summarized in Table 1 ▸. The mol­ecule of ortho-F-C₆H₄-Ph, (III), resides on a twofold axis and the F atom is thereby disordered equally over two positions, as is the corresponding H atom. The mol­ecule of para-F-C₆H₄-Ph, (IV), also resides on a twofold axis that contains the arene–arene bond. The F atom is disordered over the two para positions, with the major component contribution being 78.9 (2)%. The C1—F1 bond length to the major component is 1.3280 (17) Å, whereas the C8—F2 bond length to the minor component is suspiciously short at 1.243 (4) Å, despite an applied restraint of 1.36 (1) Å based on a search of the Cambridge Structural Database (CSD; Groom et al., 2016 ▸). An unrestrained refinement produces an even shorter distance of ∼1.22 Å. To ensure this was not a consequence of a systematic error, a second data set was collected on a different crystal from a different batch but the results were essentially identical, both in terms of disorder percentage and the length of the C8—F8 bond. A mass spectrometry experiment on a Thermo Q Exactive^TM Plus mass spectrometer showed a peak at M _r = 172, corresponding exactly to C₁₂H₉F, thereby eliminating the possibility of compositional contamination due to impurity. The origin of the abnormally short C—F bond remains unclear, and one may question the reliability of an inter­atomic distance determination to an atom that represents a peak of ∼1.8 e Å⁻³.

Table 1. Experimental details.

	(I)	(II)	(III)	(IV)
Crystal data
Chemical formula	[Cr(C12H9F)(CO)3]	[Cr(C12H9F)(CO)3]	C12H9F	C12H9F
M r	308.22	308.22	172.19	172.19
Crystal system, space group	Monoclinic, P21/c	Monoclinic, P21/c	Orthorhombic, F d d2	Orthorhombic, P b c n
Temperature (K)	100	120	100	100
a, b, c (Å)	7.0166 (3), 27.4725 (11), 7.1101 (3)	7.0616 (5), 28.3503 (19), 7.0571 (4)	12.9824 (9), 23.0588 (15), 5.7527 (4)	5.6462 (14), 20.739 (5), 7.292 (2)
α, β, γ (°)	90, 113.738 (1), 90	90, 114.324 (5), 90	90, 90, 90	90, 90, 90
V (Å3)	1254.61 (9)	1287.40 (15)	1722.1 (2)	853.9 (4)
Z	4	4	8	4
Radiation type	Mo Kα	Cu Kα	Cu Kα	Mo Kα
μ (mm−1)	0.93	7.48	0.74	0.09
Crystal size (mm)	0.4 × 0.34 × 0.14	0.66 × 0.31 × 0.16	0.56 × 0.14 × 0.02	0.39 × 0.11 × 0.03
Data collection
Diffractometer	Bruker SAINT CCD area detector	Bruker APEXII CCD	Bruker SMART APEXII	Bruker APEXII Quazar
Absorption correction	Multi-scan (SADABS; Krause et al., 2015 ▸)	Multi-scan (SADABS; Krause et al., 2015 ▸)	Multi-scan (SADABS; Krause et al., 2015 ▸)	Multi-scan (SADABS; Krause et al., 2015 ▸)
T min, T max	0.708, 0.881	0.083, 0.381	0.671, 0.754	0.706, 0.746
No. of measured, independent and observed [I > 2σ(I)] reflections	20790, 2571, 2421	19716, 2520, 2479	7085, 836, 825	19177, 1055, 899
R int	0.021	0.039	0.028	0.029
(sin θ/λ)max (Å−1)	0.625	0.618	0.620	0.667
Refinement
R[F 2 > 2σ(F 2)], wR(F 2), S	0.024, 0.068, 1.07	0.028, 0.077, 1.09	0.039, 0.107, 1.13	0.045, 0.143, 1.11
No. of reflections	2571	2520	836	1055
No. of parameters	181	181	64	68
No. of restraints	0	0	1	1
H-atom treatment	H-atom parameters constrained	H-atom parameters constrained	H-atom parameters constrained	H-atom parameters constrained
Δρmax, Δρmin (e Å−3)	0.39, −0.20	0.30, −0.39	0.36, −0.16	0.45, −0.16
Absolute structure	–	–	Flack x determined using 351 quotients [(I +) − (I −)]/[(I +) + (I −)] (Parsons et al., 2013 ▸)	–
Absolute structure parameter	–	–	−0.07 (17)	–

Computer programs: SMART (Bruker, 2003 ▸), APEX3 (Bruker, 2016 ▸), SAINT (Bruker, 2003 ▸), SAINT-Plus (Bruker, 2016 ▸), SHELXS97 (Sheldrick, 2008 ▸), SHELXT (Sheldrick, 2015a ▸), SHELXL2016 (Sheldrick, 2015b ▸) and OLEX2 (Dolomanov et al., 2009 ▸).

Results and discussion  

In the syntheses of both (I) and (II), careful examination of the crude product mixtures showed that the only (biphen­yl)tri­carbonyl­chromium products which formed had the metal atom coordinated to the nonfluorinated ring. Even when shorter 15–30 min reaction times were used, product yields were very low (<10%), but analysis of the crude product mixtures still showed coordination of chromium only to the nonfluorinated ring. This might be rationalized by the high electronegativity of fluorine rendering the fluorinated ring electron-poor, but perhaps also on the basis of steric hindrance, since a fluorine substituent on the coordinated ring would be expected to exhibit additional inter­actions with carbonyl ligands on chromium.

Similar ring selectivity had been observed previously when 2-bromo­biphenyl was heated with hexa­carbonylchromium to give only the bromine analog of (I) (Czerwinski et al., 2003 ▸). Conversely, when a more strongly electron-donating amino substituent was used, heating 2-amino­biphenyl or 4-amino­biphenyl with hexa­carbonylchromium led to the formation of only amino–biphenyl compounds, where chromium had coordinated to the more electron-rich amino-substituted ring (Czerwinski et al., 2011 ▸). We also reported that use of amide-substituted 2-acetamido­biphenyl as the coordinating ligand led to the formation of a mixture of the isomer that has chromium coordinated to the substituted ring and the isomer that has chromium coordinated to the unsubstituted ring. These previous results suggest that both steric and electronic factors may play a role in determining which ring of the biphenyl ligand coordinates to chromium. In the case of the syntheses of fluorinated compounds (I) and (II), it appears that any π-electron donation from fluorine is outweighed either by its steric inter­action with carbonyl ligands or its high electronegativity, or both.

To corroborate our predictions, we performed density functional theory (DFT) calculations at the B3LYP/LANL2DZ level of theory on ten model [(C₆H₅)(F-C₆H₄)]Cr(CO)₃ complexes in order to determine the most stable isomers, identify which ring has the higher propensity to ligate to the metal center, and compare the arene–arene dihedral angles between the theoretically minimized complexes and experimentally determined geometries (GAUSSIAN09; Frisch et al., 2016 ▸). The theoretical results are shown in Fig. 1 ▸. The five isomers with the unsubstituted arene ring coordinated to the Cr center are more stable than com­plexes with the F-C₆H₄ ring bonded to the metal atom. Among the respective compounds with the meta-, ortho-, and para-substitution patterns, the complexes with the ligated arene ring are approximately 3 kcal mol⁻¹ more stable. The would-be equivalent ortho-substituted isomers differ in energy due to different positions of the F atom relative to the Cr(CO)₃ group. The same is true for the meta-substituted congeners. The ∼31 K difference in melting points of chemical isomers (I) and (II) indicates that the latter is by far more stable. This finding contrasts with the DFT gas-phase energy calculations for the two systems that are very similar and illustrates the effect of theoretical approximations for isolated mol­ecules.

Figure 1.

(a) The relative energies of the ten [(C₆H₅)(F-C₆H₄)]Cr(CO)₃ isomers in kcal mol⁻¹. The numbers correspond to the position of the lone F atom in the ligand; the F atoms are not shown. For example, [(η⁶-C₆H₅)(3-F-C₆H₄)]Cr(CO)₃, with the C—F vector pointing away from the carbonyl ligands, is the most stable (energy of 0.00 kcal mol⁻¹), while [(η⁶-2-F-C₆H₄)(C₆H₅)]Cr(CO)₃, with the C—F vector pointing away from the viewer, is the least stable, with a relative energy of 4.50 kcal mol⁻¹. (b) The torsion angles (°) between the six-membered rings. The numbers correspond to the position of the lone F atom in the ligand.

Our second major objective was to characterize the solid-state structures of (I) and (II) by single-crystal X-ray diffraction and to compare the experimentally observed geometries with the theoretical calculations. The solid-state structure of (I) corresponds to the most stable among the four possible conformations of the model [(C₆H₅)(ortho-F-C₆H₄)]Cr(CO)₃ system (Fig. 2 ▸ a). The Cr⁰ atom exhibits a typical three-legged piano-stool geometry with expected geometrical parameters (Table 2 ▸). The C—C bond lengths in the coordinated arene ring average 1.411 (9) Å, which is 0.022 Å longer than the average bond length of 1.389 (6) Å for C—C bonds within the noncoordinated fluorinated ring. The C—C bond lengths between the arene rings in (I)–(IV) are 1.4881 (19), 1.485 (2), 1.494 (4) and 1.4959 (19) Å, and are shorter in the Cr complexes. Inter­estingly, the difference in the arene–arene bond lengths in free ligand (IV) and the Cr complexes is statistically significant, whereas it is not statistically significant in the case of (III) due to the larger s.u. value.

Figure 2.

The mol­ecular structures of (a) (I) and (b) (II), shown with 50% probability displacement ellipsoids. H atoms have been omitted for clarity.

Table 2. Selected geometrical parameters (Å) of complexes (I) and (II) compared to literature data.

Compound	(I)	(II)	CSD complexes*
Cr1—CO (av)	1.846 (4)	1.8432 (15)	1.798–1.855
Cr1—C(Ph)	2.2056 (14)–2.2311 (14)	2.2061 (17)–2.2316 (15)	2.195–2.265
Cr1—Centroid	1.7067 (6)	1.7067 (7)	1.707–1.728

Note: (*) eight related complexes have been considered; their CSD (Groom et al., 2016 ▸) refcodes are COCRDP02 and MAJLAL (Guzei & Czerwinski, 2004 ▸), EMEYIE (Czerwinski et al., 2003 ▸), EXUFAF and EXUFIN (Czerwinski et al., 2011 ▸), FOZLEL (Mailvaganam et al., 1987 ▸), QIQXOF (Brydges et al., 2013 ▸), and ABISAI (Miles et al., 2016 ▸).

The arene–arene dihedral angle of 55.77 (4)° in (I) is substanti­ally larger than the theoretical value of 41.32° (Fig. 1 ▸ b), but is in good agreement with the corresponding dihedral angle of 54.83 (7)° between the arene rings in free 2-fluoro­biphenyl. The dihedral angle in free 2-fluoro­biphenyl is 45.82° in the DFT-minimized geometry at the B3LYP/6-311+G(d,p) level of theory (Frisch et al., 2016 ▸). The ligand conformation is stabilized by weak C—H⋯F inter­actions. Atom F1 is a bifurcated acceptor of two weak C—H donors from atoms C6 and C10 of two adjacent mol­ecules (Table 2 ▸). All four compounds exhibit C—H⋯F inter­actions that are necessarily weak due to the low acidity of the available Csp ²—H donor bonds and the weak acceptor nature of the Csp ²—F organic fluorines (Thalladi et al., 1998 ▸).

The crystal structure of (II) corresponds to the more stable isomer between the two possible isomers for the para-F-substituted biphenyl (Fig. 1 ▸ b), with the nonhalogenated arene ring ligating to the Cr center. Not surprisingly, the solid-state structure of (II) shows many similarities to that of (I). The geometrical parameters of this complex fall in the expected ranges (Table 2 ▸). The average C—C bond length in the coordinated arene ring is 1.411 (10) Å, which is identical to the value found for (I), and which is 0.02 Å longer that the average bond length of 1.39 (2) Å for C—C bonds within the noncoordinated fluorinated ring. The arene–arene dihedral angle of 52.4 (5)° is smaller than the corresponding angle of 55.77 (4)° in (I). This contrasts with the theoretical calculations that predicted a marginally larger angle for (I) than for (II), with values of 41.32 and 38.08°, respectively (Fig. 1 ▸ b). Similar to (I), compound (II) exhibits C—H⋯F contacts in which atom F1 participates in two hydrogen bonds (Table 3 ▸). These bonds are characterized by shorter donor–acceptor distances compared to (I), but less favorable C—H⋯F angles.

Table 3. C—H⋯F and C4—H4⋯π inter­actions in the structures of (I)–(IV).

π represents the centroid of the C1–C6 ring.

Compound	Inter­action	C—H⋯F (Å)	C⋯F (Å)	C—H⋯F (°)
(I)	C6—H6⋯F1i	2.66	3.5612 (17)	158.1
	C10—H10⋯F1ii	2.50	3.4013 (17)	158.1
(II)	C6—H6⋯F1iii	2.48	3.122 (2)	124.8
	C8—H8⋯F1iii	2.46	3.295 (2)	146.8
(III)	C3—H3⋯F1iv	2.61	3.539 (4)	165.3
	C6—H6⋯F1v	2.65	3.548 (4)	158.6
	C4—H4⋯πiv	2.80	3.60	142.1
(IV)	C2—H2⋯F2vi	2.67	3.291 (3)	123.8
	C7—H7⋯F1vii (in plane)	2.63	3.2529 (15)	123.3
	C7—H7⋯F1viii (between layers)	2.85	3.6113 (14)	137.5

Symmetry codes: (i) x, −y + , z − ; (ii) x + 1, −y + , z + ; (iii) x − 1, −y + , z − ; (iv) x + , −y + , z + ; (v) −x + 1, −y + 1, z − 1; (vi) −x + , y + , z; (vii) −x + , y − 0.5, z; (viii) −x + , −y + , z + .

Most of the experimentally determined structural parameters for (I) and (II) are in good agreement with literature data (Table 2 ▸). However, the arene–arene dihedral angles in similar compounds span a wide range. This angle is 0° in μ-(η⁶,η⁶-biphen­yl)-bis­(tri­carbonyl­chromium) (Guzei & Czerwinski, 2004 ▸), 38.08° in (η⁶-4-amino­biphen­yl)tri­carbonyl­chromium (Czerwinski et al., 2011 ▸), 60.70° in (η⁶-2-bromo-1,1′-biphen­yl)tri­carbonyl­chromium (Czerwinski et al., 2003 ▸), 65.76° in (η⁶-2-amino­biphen­yl)tri­carbonyl­chromium (Czerwinski et al., 2011 ▸), and 23.55 (5)° in (η⁶-biphen­yl)tri­carbonyl­chromium (Guzei & Czerwinski, 2004 ▸).

The arene–arene dihedral angle of 52.4 (5)° in (II) is surprisingly large when considering that the fluorine substituent is in the para position of the noncoordinated ring. In this position, it would be expected not to have significant inter­actions with the carbonyl ligands or with the ortho H atoms of the coordinated ring. By comparison, the solid-state structure of the unsubstituted analog [(η⁶-C₆H₅)(C₆H₅)]Cr(CO)₃ exhibits a arene–arene dihedral angle of 23.55 (5)°, whereas its free ligand, i.e. biphenyl, is planar in the solid state (Charbonneau & Delugeard, 1977 ▸). The twisting that is effected by coordination of biphenyl to Cr(CO)₃ prompted us to examine the solid-state structures of the free ligands 2-fluoro­biphenyl, (III), and 4-fluoro­biphenyl, (IV). Whereas the range of observed arene–arene dihedral angles is large, the energy difference among various relative ring positions falls within a few kcal mol⁻¹ and our variable-temperature NMR studies show no evidence of a significant rotational barrier about the arene–arene bond.

Whereas the structure of (III) at 293 K has been reported previously (Rajnikant et al., 1995 ▸), we sought to redetermine the structure at 100 K for consistency with the conditions of our determinations of (I), (II), and (IV) (Fig. 3 ▸).

Figure 3.

The mol­ecular structures of (a) (III) and (b) the symmetry-independent part of (III). Both structures are shown with 50% probability displacement ellipsoids. The H atom on C2 and atom F1 are 50% occupied due to positional disorder over a twofold axis perpendicular to the arene–arene bond.

The structure of (III) at 100 K closely resembles the previously reported 293 K analog, but expectedly the bond lengths are slightly longer in the 100 K structure. In particular, the C2—F1 bond length of 1.331 (4) Å is in a reasonably close agreement with the previously determined value of 1.319 (4) Å, but slightly longer, as expected due to reduced librational effects at 100 K. Both values are noticeably shorter than the 1.36 Å average for related compounds in the Cambridge Structural Database (Groom et al., 2016 ▸). The C—F distances in (I), (II), and (IV) are 1.3622 (17), 1.355 (2), and 1.3280 (17) Å, respectively. The arene–arene dihedral angles in (III) [54.83 (7)° at 100 K and 54 (3)° at 293 K] are quite close to the analogous angle in (I), suggesting that ligation to Cr(CO)₃ imparts little additional steric hindrance and results in only a marginal increase in the observed dihedral angle.

The structure of (III) is stabilized by weak C—H⋯F inter­actions, forming a three-dimensional framework (Fig. 4 ▸ a). Atom F1 is a bifurcated acceptor of two H—C bonds from two adjacent mol­ecules (Table 3 ▸). Additionally, there is a C4—H4⋯π herringbone-type inter­action absent in the structure of (IV).

Figure 4.

The mol­ecular packing of the free ligands, showing (a) the three-dimensional framework in (III), viewed along [10], and (b) the layered structure of (IV), viewed along the b axis.

The solid-state structure of (IV) shows that the mol­ecule resides on a crystallographic twofold axis containing the C4—C5 bond (Fig. 5 ▸).

Figure 5.

(a) The mol­ecular structure of (IV) and (b) its symmetry-independent part. Both structures are shown with 50% probability displacement ellipsoids. In part (b), all disorder components are shown.

The C4—C5 bond length [1.4959 (19) Å] is shorter than a single C—C bond, but without substantial multiple-bond character. This bond length is in excellent agreement with the arene–arene distance of 1.494 (4) Å observed in (III). The arene–arene dihedral angle in (IV) of 0.71 (8)° suggests that the mol­ecule is essentially planar in the solid state.

In the crystal of (IV), mol­ecules form corrugated sheets stacked in the c direction (Fig. 4 ▸ b). Within the plane, the F atom in each of its positions forms two C—H⋯F inter­actions (Table 3 ▸). The corrugated layers are connected by additional weaker C7—H7⋯F1 inter­actions; thus, atom F1 participates in four weak C—H⋯F hydrogen bonds, i.e. two within the sheet plane and two to the layers above and below. Atom F2 participates in just two such nonclassical inter­actions, and this difference may rationalize why atom F1 is the major disorder component at 79.8 (2)%.

In light of the behavior of biphenyl, it is somewhat surprising that ligation of flat ligand (IV) to Cr(CO)₃ results in such a large dihedral angle of 52.4 (5)° in (II). This angle cannot be explained by π–π inter­actions, since they are absent in the structure of (II), nor can it be explained by inter­molecular C—H⋯F hydrogen-bonding inter­actions because they are observed in all structures.

Inter­estingly, both C ₂-symmetric structures of the free ligands (III) and (IV) exhibit positional disorder that is absent in (I) and (II). Thus, ligation of a fluoro­biphenyl to chromium effectively discriminates between the two rings and only one ordered product is observed in the solid state for (I) and (II).

Conclusions  

Compounds (I) and (II) exhibit relatively large arene–arene dihedral angles in the solid state. This angle in (I) may be understood in terms of inter-ring ortho-F⋯ortho-H steric hindrance that was shown to be present in the free ligand (III). The large dihedral angle in (II) is more difficult to rationalize, especially since the free ligand (IV) exhibits a flat structure that is consistent with a fluorine substituent in the para position. The variability in arene–arene dihedral angles in a series of substituted (biphen­yl)tri­carbonyl­chromium compounds may only represent small energetic differences, and might simply be attributed to packing forces in the solid state. Further studies of related compounds may be necessary to advance a complete explanation for this observation. The method of heating fluoro­biphenyl with hexa­carbonyl­chromium led only to products (I) and (II), where chromium is coordinated to the nonfluorinated ring, but efforts are currently underway to find alternative synthetic methods for the preparation of the isomers where chromium is coordinated to the fluorinated ring. Syntheses of these mol­ecules would allow their haptotropic rearrangements to be studied, and it is hoped that studies of their solid-state structures might further clarify how positional substitution of fluorine in the biphenyl ligand affects arene–arene dihedral angles. Unlike the solid-state structures of free ligands (III) and (IV), chromium complexes (I) and (II) exhibit no crystallographic symmetry and no positional disorder.

Supplementary Material

CCDC references: 1563759, 1563758, 1563757, 1563756