1. Introduction
Half-sandwich iron(II) complexes of the type [Fe(η
5-Cp’)X] (Cp’ = C
5H
5 or substituted cyclopentadienyl, X = Cl – I) are useful as highly reactive cyclopentadienyliron(II) transfer reagents, which, due to their thermal lability, are usually generated in situ at low temperatures for immediate use [
1]. Seminal work was published already in 1985 by Kölle, who described the generation of [Fe(η
5-Cp*)Br] (Cp* = C
5Me
5) from LiCp* and [FeBr
2(DME)] in THF at −80 °C [
2]. The corresponding chlorido complex [Fe(η
5-Cp*)Cl] is particularly popular as a Cp*Fe
+ source [
3,
4,
5,
6,
7,
8,
9,
10]. In contrast to the thermal lability of this compound, its TMEDA chelate [Fe(η
5-Cp*)Cl(TMEDA)] is perfectly stable at room temperature [
4] and the same holds true for the closely related TMEDA complexes [Fe(η
5-C
5Me
4Et)Cl(TMEDA)] [
11] and [Fe(η
5-Cp*)Br(TMEDA)] [
12]. Similar to these TMEDA containing
N,
N-chelates,
C,
N-chelates [Fe(η
5-Cp*)X(NHC
N)] (X = Cl, I) containing N-heterocyclic carbenes functionalised with an
N-donor moiety (NHC
N) have also been described [
14,
15,
16]; unchelated analogues [Fe(η
5-Cp*)Cl(NHC)] proved sufficiently stable for isolation with the standard NHC IMes and the bulkier 1,3-diisopropyl-4,5-dimethylimidazolin-ylidene [
17,
18,
19]. Stabilisation by external donors is not necessary for isolation when extremely bulky Cp’ ligands [
20,
21,
22] are applied, leading to “self-stabilised” halido-bridged dimers [{Fe(η
5-Cp’)(μ-X)}
2] according to single-crystal X-ray diffraction (XRD) (Cp’ = C
5iPr
5, X = Br; Cp’ = C
5H
iPr
4, X = Br, I; Cp’ = C
5H
2-1,2,4-
tBu
3, X = Br, I; Cp’ = C
5(
p-C
6H
4Et)
5, X = Br) [
23,
24,
25,
26]. Manners and Walter independently found that [{Fe(η
5-C
5H
2-1,2,4-
tBu
3)(μ-I)}
2] undergoes heterolytic cleavage in toluene, affording [Fe(η
5-C
5H
2-1,2,4-
tBu
3)(C
7H
8)]
+ and [Fe(η
5-C
5H
2-1,2,4-
tBu
3)I
2]
− [
27,
28]. In the same vein, deaggregation of [{Fe(η
5-C
5H
2-1,2,4-
tBu
3)(μ-I)}
2] was achieved by reaction with NR
4I (R = Et,
nBu), giving rise to the formation of NR
4[Fe(η
5-C
5H
2-1,2,4-
tBu
3)I
2] [
27,
28]. The only other closely related compound is [Fe(η
5-C
5H
2-1,2,4-
tBu
3)(μ-Br)
2Na(DME)
2], which Sitzmann had obtained by serendipity and in trace amounts only in the preparation of [{Fe(η
5-C
5H
2-1,2,4-
tBu
3)(μ-Br)}
2] from [FeBr
2(DME)] and the corresponding sodium cyclopentadienide in DME [
29]. This dinuclear complex might be viewed as contact ion pair [Na(DME)
2][Fe(η
5-C
5H
2-1,2,4-
tBu
3)Br
2], thus exhibiting, cum grano salis, the [Fe(η
5-C
5H
2-1,2,4-
tBu
3)Br
2]
− anion. In view of the mature state of half-sandwich iron(II) chemistry [
1], the paucity of compounds containing simple anions of the type [Fe(η
5-Cp’)X
2]
− is quite surprising. Together with the enormous popularity of the Cp* ligand [
22], this prompted us to address the synthesis of compounds containing [Fe(η
5-Cp*)X
2]
− (X = Cl – I).
2. Results and Discussion
The synthesis of our target compounds (
Scheme 1) was inspired by the work of Manners and Walter mentioned above.
The addition of N
nPr
4Cl (1 equiv.) to [Fe(η
5-Cp*)Cl], generated in situ from LiCp* and FeCl
2 in THF at low temperatures, afforded a green solution. LiCl was precipitated by addition of toluene and subsequently removed by filtration. Storing of the filtrate at −40 °C afforded N
nPr
4[Fe(η
5-Cp*)Cl
2] as green crystals in 60% yield. The use of N
nPr
4Br instead of N
nPr
4Cl furnished N
nPr
4[Fe(η
5-Cp*)BrCl] in 39% yield. Both compounds were structurally characterised by XRD. Their molecular structures are shown in
Figure 1 and
Figure 2 and pertinent metric parameters are collected in
Table 1. Not surprisingly, the [Fe(η
5-Cp*)BrCl]
− anion exhibits a disorder of the halogen atoms.
Our attempts to prepare N
nPr
4[Fe(η
5-Cp*)Br
2] in an analogous way from Kölle’s compound [Fe(η
5-Cp*)Br] and N
nPr
4Br furnished the product in 33% yield, but invariably afforded crystals whose structural investigation by XRD was fraught with problems due to severe cation disorder. Our best result is shown in
Figure S1 in the Supporting Information. Although bond lengths and angles are given only for the heavy atoms in
Table 1, these data should be treated with particular caution in the case of N
nPr
4[Fe(η
5-Cp*)Br
2], where they are not taken into consideration for our discussion. The problems encountered with the tetra-
n-propylammonium cation prompted us to use the tetraphenylphosphonium cation instead. The preparation of PPh
4[Fe(η
5-Cp*)X
2] (X = Cl, Br, I) by addition of PPh
4X (1 equiv.) to [Fe(η
5-Cp*)X] (prepared in situ from FeX
2 and KCp*) turned out to be straightforward, although the isolated yields were unsatisfactorily poor (21% at most), probably due to the much lower solubility of PPh
4X in comparison to N
nPr
4X. In contrast to the synthesis of N
nPr
4[Fe(η
5-Cp*)X
2], a trend towards even lower yields was observed when LiCp* was used instead of KCp*. The product was obtained as crystals suitable for XRD in each case and no disorder problems were encountered, as anticipated. The molecular structures of PPh
4[Fe(η
5-Cp*)X
2] are shown in
Figure 3 (X = Cl),
Figure 4 (X = Br), and
Figure 5 (X = I).
The compounds listed in
Table 1 exhibit very similar iron–cyclopentadienyl ring centroid distances between 1.96 and 1.99 Å, which is much larger than the corresponding distances in the ferrocenes [Fe(η
5-Cp*)
2] (1.65 Å) [
30] and [Fe(η
5-C
5H
2-1,2,4-
tBu
3)
2] (1.72 Å) [
31] and marginally larger than those in the open-shell half-sandwich complexes [Fe(η
5-Cp*){N(SiMe
3)
2}] (1.90 Å) [
32], [Fe(η
5-C
5iPr
5){N(SiMe
3)
2}] (1.92 Å) [
13] and [{Fe(η
5-C
5H
2-1,2,4-
tBu
3)(μ-X)}
2] (1.92 and 1.93 Å for X = Br and I, respectively) [
24,
25]. The differences in the Fe–X bond lengths observed for X = Cl, Br and I are in accord with the different radii of the halogen atoms. A particularly good agreement is achieved with Pauling’s tetrahedral covalent radii, which reflect a convolution of covalent and dative bonding, the values being 0.99, 1.11 and 1.28 Å for Cl, Br and I, respectively [
33]. Not surprisingly, the X–Fe–X angles of the Cp* complexes are wider (by ca. 5 °) than those of the congeners containing the bulkier C
5H
2-1,2,4-
tBu
3 ligand, whose comparatively less symmetric nature may be the reason for the significant difference of the two Fe–I bond lengths (Δ
d 0.09 Å) in the anion of N
nBu
4[Fe(η
5-C
5H
2-1,2,4-
tBu
3)I
2]. The tetraalkylammonium cations are engaged in CH···X contacts compatible with weak hydrogen bonds (indicated as dotted lines in
Figure 1, not shown for the disordered species in
Figure 2 and
Figure 3).[
34,
35] The contacts of the two halogen atoms are almost equidistant in each case (CH···Cl 2.67 and 2.73 Å for N
nPr
4[Fe(η
5-Cp*)Cl
2], CH···I 3.10 and 3.14 Å for N
nBu
4[Fe(η
5-C
5H
2-1,2,4-
tBu
3)I
2]). The PPh
4+ cations interact with the [Fe(η
5-Cp*)X
2]
− anions through phenyl CH···X contacts (2.76–2.95, 2.96 and 3.09–3.15 Å for X = Cl, Br and I, respectively; not shown in
Figure 3,
Figure 4 and
Figure 5). In addition, the
para-H atom of a phenyl ring points towards the centre of the Cp* ligand (phenyl CH···C 2.53–2.74 Å, phenyl CH···Cp* ring centroid 2.28–2.38 Å shown as dotted lines in
Figure 3,
Figure 4 and
Figure 5), indicating a CH···π interaction [
36,
37] similar to that in the T-shaped benzene dimer [
38,
39,
40,
41,
42,
43], for which a CH···C
6H
6 ring centroid distance of 2.25 Å was computed recently [
44].
The electronic structure of the anion of N
nBu
4[Fe(η
5-C
5H
2-1,2,4-
tBu
3)I
2] has been scrutinised by SQUID magnetometry, EPR spectroscopy and ab initio Complete Active Space Self Consistent Field-Spin Orbit calculations, which revealed a high-spin d
6 iron(II) centre with a strongly anisotropic
S = 2 ground state [
28]. This in-depth study by Manners makes an analogous investigation of our closely related compounds dispensable. The paramagnetic nature of their [Fe(η
5-Cp*)X
2]
− anions is clearly evident from the NMR spectra. The Cp* ligand gives rise to a
1H NMR signal at
δ ≈ 200 ppm. This may be compared with the data reported for the substituted cyclopentadienyl ligands of [{Fe(η
5-C
5iPr
5)(μ-Br)}
2] in C
6D
6 [
δ(
1H) = 95.7 (CH
Me2), 11.3 (C
HMe
2), and −117.3 ppm (CH
Me2)] [
23] and of N
nBu
4[Fe(η
5-C
5H
2-1,2,4-
tBu
3)I
2] in THF-
d8 [
δ(
1H) = −20.3 and −31.4 ppm (2 ×
tBu)] [
28].
Kölle demonstrated the successful generation of the highly reactive compound [Fe(η
5-Cp*)Br] by trapping reaction with carbon monoxide at −80 °C, which furnished the diamagnetic carbonyl complex [Fe(η
5-Cp*)Br(CO)
2] in 59% yield [
2]. In the same vein, Walter obtained [Fe(η
5-C
5H
2-1,2,4-
tBu
3)I(CO)
2] by carbonylation of the “self-stabilised” halido-bridged dimer [{Fe(η
5-C
5H
2-1,2,4-
tBu
3)(μ-I)}
2] with CO at room temperature in 80% yield [
25]. We have studied the carbonylation of our target compounds exemplarily with N
nPr
4[Fe(η
5-Cp*)Cl
2] and observed an essentially quantitative reaction with CO under the same mild conditions, affording the well-known carbonyl complex [Fe(η
5-Cp*)Cl(CO)
2] [
45,
46]. The crystal structure of this compound reported in 1988 had been determined at room temperature [
46], which prompted us to redetermine the structure at 100 K (see the Supporting Information).
3. Materials and Methods
Experimental Details. All reactions were performed in an inert atmosphere (argon or dinitrogen) by using standard Schlenk techniques or a conventional glovebox. Solvents were dried with a commercial Solvent Purification System (M. Braun, MB SPS 7), degassed and stored over 3 Å molecular sieves under inert atmosphere. Starting materials were procured from standard commercial sources and used as received. LiCp* and KCp* were synthesised by deprotonation of pentamethylcyclopentadiene in n-hexane with n-butyllithium and potassium metal, respectively, and isolated by filtration or centrifugation. NMR spectra were recorded with a Varian MR-400 and Varian NMRS-500 spectrometers operating at 400 and 500 MHz, respectively, for 1H. Elemental analyses were carried out with a HEKAtech Euro EA-CHNS elemental analyser at the Institute of Chemistry, University of Kassel, Germany.
NnPr4[Fe(η5-Cp*)Cl2]: A Schlenk tube charged with LiCp* (176 mg, 1.24 mmol) and FeCl2 (156 mg, 1.23 mmol) was cooled to −60 °C. THF (3 mL) cooled to the same temperature was added. The stirred mixture was allowed to warm up to −20 °C. NnPr4Cl (275 mg, 1.24 mmol) was added. The stirred mixture was allowed to warm up to ambient temperature and was subsequently filtered through a Celite pad. Toluene (ca. 3 mL) was slowly added to the green filtrate until formation of an essentially colourless precipitate was observed. Insoluble material was removed by filtration through a Celite pad. Storing of the filtrate at −40 °C afforded the product as green crystals, which were separated from the yellow mother liquor, washed with n-hexane (5 mL) and dried under vacuum. Yield 327 mg (60%). Elemental Analysis for C22H43NCl2Fe (448.34 g/mol): Calculated (%): C 58.94, H 9.67, N 3.12. Found (%): 58.18, H 9.37, N 3.21. 1H NMR (400 MHz, THF-d8): δ 194.4 (15H, s, ν½ = 379 Hz, Cp*), 16.8 (8H, s, ν½ = 270 Hz, (CH2)2CH3), 9.9 (8H, s, ν½ = 217 Hz, (CH2)2CH3), 1.3 (12H, s, ν½ = 267 Hz, (CH2)2CH3).
NnPr4[Fe(η5-Cp*)BrCl]: This compound was obtained by a procedure analogous to that described above for NnPr4[Fe(η5-Cp*)Cl2] by using LiCp* (130 mg, 0.91 mmol), FeCl2 (116 mg, 0.92 mmol) and NnPr4Br (245 mg, 0.92 mmol) in THF (3 mL). Yield 175 mg (39%). Elemental Analysis for C22H43NBrClFe (492.79 g/mol): Calculated (%): C 53.62, H 8.80, N 2.84. Found (%): C 54.24, H 8.81, N 2.48. 1H NMR (400 MHz, THF-d8): δ 206.5 (15H, s, ν½ = 2860 Hz, Cp*), 25.8 (8H, s, ν½ = 501 Hz, (CH2)2CH3), 15.1 (8H, s, ν½ = 353 Hz, (CH2)2CH3), 2.7 (12H, s, ν½ = 649 Hz, (CH2)2CH3).
NnPr4[Fe(η5-Cp*)Br2]: This compound was obtained by a procedure analogous to that described above for NnPr4[Fe(η5-Cp*)Cl2] by using LiCp* (65 mg, 0.46 mmol), FeBr2 (99 mg, 0.46 mmol) and NnPr4Br (122 mg, 0.46 mmol) in THF (1.5 mL). Yield 81 mg (33%). An analytical sample was obtained by recrystallization from benzene. Elemental Analysis for C22H43NBr2Fe·½C6H6 (576.29 g/mol): Calculated (%): C 52.10, H 8.05, N 2.43. Found (%): C 52.18, H 8.24, N 1.76. 1H NMR (400 MHz, THF-d8): δ 203.5 (15H, s, ν½ = 561 Hz, Cp*), 16.7 (8H, s, ν½ = 312 Hz, (CH2)2CH3), 10.7 (8H, s, ν½ = 241 Hz, (CH2)2CH3), 2.23 (12H, s, ν½ = 194 Hz, (CH2)2CH3).
PPh4[Fe(η5-Cp*)Cl2]: A Schlenk tube charged with KCp* (40 mg, 0.23 mmol) and FeCl2 (29 mg, 0.23 mmol) was cooled to −60 °C. THF (0.5 mL) cooled to the same temperature was added. The stirred mixture was allowed to warm up to −20 °C. PPh4Cl (86 mg, 0.23 mmol) was added. The stirred mixture was allowed to warm up to ambient temperature and was subsequently filtered through a Celite pad. The yellow filtrate was carefully layered with n-hexane, resulting in the slow formation of yellow crystals, which were separated from the mother liquor, washed with n-hexane (2 mL) and dried under vacuum. Yield 8 mg (6%). In view of the unsatisfactorily low yield, elemental analysis was not performed for this compound. 1H NMR (500 MHz, THF-d8): δ 188.2 (15H, s, ν½ = 311 Hz, Cp*), 13.2 (8H, ν½ = 95 Hz, Ph), 10.7 (8H, ν½ = 98 Hz, Ph), 10.2 (4H, ν½ = 80 Hz, Ph).
PPh4[Fe(η5-Cp*)Br2]: This compound was obtained by a procedure analogous to that described above for PPh4[Fe(η5-Cp*)Cl2] by using KCp* (40 mg, 0.23 mmol), FeBr2 (50 mg, 0.23 mmol) and PPh4Br (96 mg, 0.23 mmol) in THF (0.5 mL). Yield 12 mg (8%). In view of the unsatisfactorily low yield, elemental analysis was not performed for this compound. 1H NMR (500 MHz, THF-d8): δ 193.7 (15H, s, ν½ = 647 Hz, Cp*), 11.3 (8H, ν½ = 162 Hz, Ph), 9.2 (8H, ν½ = 194 Hz, Ph), 8.6 (4H, ν½ = 188 Hz, Ph).
PPh4[Fe(η5-Cp*)I2]: This compound was obtained by a procedure analogous to that described above for PPh4[Fe(η5-Cp*)Cl2] by using KCp* (40 mg, 0.23 mmol), FeI2 (71 mg, 0.23 mmol) and PPh4I (107 mg, 0.23 mmol) in THF (0.5 mL). Yield 38 mg (21%). In view of the unsatisfactorily low yield, elemental analysis was not performed for this compound. 1H NMR (500 MHz, THF-d8): δ 209.9 (15H, s, ν½ = 5.34 Hz, Cp*), 10.1 (8H, ν½ = 55 Hz, Ph), 9.1 (8H, ν½ = 60 Hz, Ph), 8.7 (4H, ν½ = 55 Hz, Ph).
[Fe(η5-Cp*)Cl(CO)2]: A solution of N
nPr
4[Fe(η
5-Cp*)Cl
2] (40 mg, 0.09 mmol) in THF (2 mL) was subjected to an atmospheric pressure of CO, which led to an immediate colour change from green to red. The solution was stirred for 10 min. Volatile components were removed under vacuum. Benzene (0.7 mL) was added to the residue. Insoluble material was removed by filtration through a Celite pad. Slow evaporation of the filtrate afforded the product as red crystals. Yield 23 mg (92%). Spectroscopic data were found to be in good agreement with published values [
45,
46].
X-Ray Crystallography: For all data collections a single crystal was mounted on a micro-mount and all geometric and intensity data were taken from this sample by
ω-scans at 100(2) K. Data collections were carried out either on a Stoe StadiVari diffractometer equipped with a 4-circle goniometer and a DECTRIS Pilatus 200K detector (for N
nPr
4[Fe(η
5-Cp*)Cl
2], N
nPr
4[Fe(η
5-Cp*)Br
2] and PPh
4[Fe(η
5-Cp*)Cl
2]) or on a Stoe IPDS2 diffractometer equipped with a 2-circle goniometer and an area detector (for N
nPr
4[Fe(η
5-Cp*)BrCl], PPh
4[Fe(η
5-Cp*)Br
2], PPh
4[Fe(η
5-Cp*)I
2] and [Fe(η
5-Cp*)Cl(CO)
2]). The data sets were corrected for absorption (by multi scans), Lorentz and polarisation effects. The structures were solved by direct methods (SHELXT 2014/7) [
47] and refined using alternating cycles of least-squares refinements against
F2 (SHELXL2014/7) [
47]. H atoms were included to the models in calculated positions with the 1.2 fold isotropic displacement parameter of their bonding partner. Experimental details for each diffraction experiment are given in
Table S1 (Supplementary Materials). CCDC 2300615–2300621 contain supplementary crystallographic data for this paper. These data can be obtained free of charge from The Cambridge Crystallographic Data Centre,
www.ccdc.cam.uk/structures.