Synthesis of ferrocenyl aryl ethers via Cu(I)/phosphine catalyst systems

1 Introduction

Ferrocenes are an important class of organometallic compounds for synthesis [1–7]. Current interest in ferrocenyl ethers lies in particular in the possibility for their use as organometallic ligands in homogeneous catalysts [8–13]. The synthetic availability of heteroatom-substituted ferrocenes is still limited because the routes for the introduction of heteroatoms at benzenes are not suitable for ferrocenes [12, 13]. Pertinent examples of successful syntheses are the amino- or hydroxyferrocenes functionalized via nucleophilic substitution with alkyl halides [14–17] or the synthesis of ferrocenyl alkyl ethers via a trialkylsilyl-protected hydroxyferrocene [18, 19].

Presently, only very few ferrocenyl aryl ethers are known. The first synthesis of such compounds date back to the 1960s, when Rausch [20] and Nefedova [21], employed Ullmann-type coupling resulting in product yields of ferrocenyl phenyl ether of 10–25%.

2 Results and discussion

In a report by an der Heiden et al. [1], the high-yield synthesis of several ferrocenyl aryl ethers was described starting from iodo- and 1,1′-diiodoferrocene and phenols, mediated by Cu(I)/2,2,6,6-tetramethylheptane-3,5-dione (TMHD) or CuI/1,10-phenanthroline catalytic systems. The Cu(I)/TMHD system showed only half the activity when using brominated ferrocenes, and a much higher catalyst loading was required. The promising results obtained in this study with a CuI/phosphine system in toluene were not further evaluated. As a consequence, we used a CuI/phosphine system in toluene, a similar system to the Venkataraman catalyst (CuBr(PPh₃)₃) [22, 23], to further evaluate the potential of this coupling reaction of iodoferrocene and 1,1′-dibromoferrocene with phenols.

The results for this reaction are collected in Table 1. Entries 1 and 2 differ in a carbon–iodine vs. carbon–bromine bond, while the same catalyst was used. The more sophisticated reaction using 1,1′-dibromoferrocene resulted in a low yield (15%) compared with the results obtained with iodoferrocene [1]. A change to a less bulky alkylphosphine ligand did slightly improve the yield (up to 27%, Table 1; entry 3). In the present investigation, we have extended our previous studies and make use of different carbon–halogen bond strengths in the variation of product formation. According to the previous publication, we changed the ligand system back to arylphosphines, which provided a reasonable improvement in yield using Cs₂CO₃ as the base (up to 86%, Table 1; entries 4–7). A comparison of the two different bases Cs₂CO₃ and K₃PO₄ showed an additional improvement from 83 to 97% (Table 1; entries 7–8). Nearly identical results were also obtained for iodoferrocene in the previous work by an der Heiden et. al., where 85 and 99% were obtained, respectively [1]. The influence of the ligand system and the solvent does not play any specific role, contrary to previously published results using iodoferrocene and the NMP/TMHD/toluene catalytic systems for less sterically hindered phenols (substitution in meta or para position) (comparison of Tables 1–3 in reference [1] with the present Table 1 entries 9 and 10). A dramatic decrease in yield was only observed for ortho-substituted phenols, regardless of whether Cs₂CO₃ or K₃PO₄ was used as the base (Table 1, entries 11–13).

Table 1

Screening for the coupling of ferrocene halides and phenols in toluene ^a.

Entry	R	Ferrocene	Ligand	Base	Yieldb (%)	Compound
1	4-t-Bu	FcIc	(1-Ad)2PBn 2 eq.	Cs2CO3	84 [1]	1
2	4-t-Bu	1,1′-FcBr2d	(1-Ad)2PBn 2 eq.	Cs2CO3	15	2
3	4-t-Bu	1,1′-FcBr2d	Cy2PBn 2 eq.	Cs2CO3	27	2
4	4-t-Bu	1,1′-FcBr2d	dppm 1 eq.	Cs2CO3	86	2
5	4-t-Bu	1,1′-FcBr2d	PPh2(o-i-propylphenyl) 2 eq.	Cs2CO3	86	2
6	4-t-Bu	1,1′-FcBr2	P(o-tolyl)3 2 eq.	Cs2CO3	64	2
7	4-t-Bu	1,1′-FcBr2d	PPh3 2.5 eq.	Cs2CO3	83	2
8	4-t-Bu	1,1′-FcBr2d	PPh3 2.5 eq.	K3PO4	97	2
9	4-Cl	FcIc	PPh3 3 eq.	Cs2CO3	96	3
10	3-t-Bu	FcIc	PPh3 3 eq.	Cs2CO3	94	4
11	2-t-Bu, 4-Me	FcIc,e	PPh3 3 eq.	K3PO4	27	5
12	2,4-t-Bu2	FcIc,e	PPh3 3 eq.	K3PO4	23	6
13	2,4,6-Me3	FcIc	PPh3 3 eq.	Cs2CO3	37	7
14	2,4,6-Me3	1,1′-FcBr2d	PPh3 3 eq.	Cs2CO3	23	8
15	3,5-Me2	1,1′-FcBr2d	PPh3 3 eq.	Cs2CO3	54	9
16	2-t-Bu, 4-OMe	FcIc	PPh3 3 eq.	Cs2CO3	(24)	10
17	2-NO2	1,1′-FcBr2d	PPh3 3 eq.	K3PO4	<5 (none)	–
18	2-NO2	FcIc	PPh3 3 eq.	Cs2CO3	<5 (none)	–
19	2-NO2, 4-OMe	FcIc	(1-Ad)2PBn 2 eq.	Cs2CO3	<5 (none)	–
20	2-NO2, 4-OMe	FcIc	(1-Ad)2PBn 2 eq.	DBU	<5 (none)	–
21	2-NO2, 4-OMe	FcIc	Cy2PBn 2 eq.	KOt-Bu	<5 (none)	–
22	4-t-Bu	1,1′-FcBr2f	PPh3 3 eq.	Cs2CO3	(43)g	11
23	2,4-Me2	11c	PPh3 3 eq.	Cs2CO3	(56)	12

^aConditions: ferrocene (0.125 mmol), CuI (5 mol.%), ligand (10–15 mol.% according to CuI), base (0.25 mmol), T=110°C, toluene. Samples were taken after 26 h; ^byields determined by GC; in parentheses the isolated yield is given; ^cphenol (0.25 mmol); ^dphenol (0.35 mmol); ^ereaction time was 60 h instead of 26 h; ^fphenol (0.125 mmol); ^gresults in ∼62% yield from theory.

With this knowledge in hand, the less reactive 1,1′-dibromoferrocene was tested in entry 14, resulting in another decrease in yield (24% by GC) compared with the iodoferrocene reaction (37%). If the steric hindrance is reduced by a meta-substitution, with 1,1′-dibromoferrocene, a moderate yield could also be obtained (Table 1 entry 15). The reaction of iodoferrocene with a deactivated phenol such as 4-methoxyphenol resulted in a product yield of only 24% after column chromatography (see Table 1 entry 16). Reactions with phenols containing the strong electron-withdrawing substituent -NO₂ failed completely (Table 1 entries 17–21), even if the phosphine or the ferrocene species were modified. Similar unsuccessful results were obtained in previous work [1].

Considering the obtained yields for these reactions, we found a suitable variation in starting material and side-product rates. For example, for entries 2 and 3, we obtained between 54 and 68% bromoferrocene and 1,1′-dibromoferrocene, ∼9% 1,1″-biferrocene, and 6–7% ferrocene after the given reaction time. For entry 4, we found 1% bromoferrocene, 3% 1,1′-dibromoferrocene, ∼3% 1,1″-biferrocene, and 7% ferrocene. For entry 17, where 1,1′-dibromoferrocene was used, we found a 3:1 ratio of bromoferrocene: 1,1′-dibromoferrocene after 26 h. In contrast to entries 18–21, where the more reactive iodoferrocene was used, we obtained after the 26-h reaction time mainly unreacted material (∼50–60%) and 1,1″-biferrocene and ferrocene in a ratio of ∼2:3.

If only one equivalent of phenol was used in the reaction with 1,1′-dibromoferrocene (Table 1, entry 22), a yield of 43% was obtained for product 11. In an additional attempt to prepare a mixed phenoxyferrocene, compound 11 was reacted with 2,4-dimethylphenol to obtain compound 12 in 56% yield after workup by column chromatography.

Crystals of 1,1′-di(4-tert-butylphenoxy)ferrocene (2) could be obtained by slow evaporation of the solvent (CDCl₃), and their structure has been determined (Figs. 1 and 2). The result has provided data for the first structurally authenticated ferrocenyl diaryl ether complex. Selected bond lengths (in Å) and angles (in deg) are presented in the caption of Fig. 1.

Fig. 1

diamond [24] plot of compound 2 in the solid state, showing 50% probability displacement ellipsoids and the atom numbering scheme. H atoms have been omitted for clarity. D1–Fe1 1.6549(4), D2–Fe1 1.6565(4), C1–O1 1.412(3), C6–O2 1.397(3), C11–O1 1.398(3), C21–O2 1.396(3), O1–C1–C6–O2 ∼74.12° (D1=centroid C1–C5, D2=centroid C6–C10).

Fig. 2

diamond [24] plot of compound 2, showing the conformation of the two Cp rings. H atoms have been omitted for clarity.

The two Cp rings are rotated by 74.1(2)° against each other, with the two ether groups pointing in opposite directions out of the Cp planes to reduce steric hindrance. The molecules are connected by a single intermolecular hydrogen bond C9–H9···O2 with (H9···O2 2.51(3), H9–C9 0.97(4), C9···O2 3.398(4) Å). All other intermolecular hydrogen bonds establishing the three-dimensional framework are longer than 2.76(4) Å (H5···C12 in C5–H5···C12; H5–C5 0.95(4) Å, C5···C12 3.632(5) Å).

3 Conclusions

We present herein a simple high-yield synthesis of a variety of different ferrocenyl aryl ethers by an Ullmann-type coupling protocol. Using the more inexpensive bromo- and 1,1′-dibromo-ferrocenyl precursors instead of their iodo analogues, we have developed a more cost-efficient method for the synthesis of ferrocenyl aryl ethers. For the first time, a mixed 1,1′-diferrocenyl aryl ether could be prepared via a simple synthetic protocol.

4 Experimental section

4.3 Single-crystal X-ray structure determination of compound 2

Crystal data and details of the structure determination are presented in Table 2. Single crystals suitable for the X-ray diffraction study were grown from chloroform. A clear brown prism (0.30×0.20×0.10 mm³) was stored under perfluorinated ether and transferred into a Lindemann capillary for data collection. Data were corrected for Lorentz, polarization, and, arising from the scaling procedure, for latent decay and absorption effects [28]. The structure was solved by a combination of Direct Methods and difference Fourier syntheses [29]. All nonhydrogen atoms were refined with anisotropic displacement parameters. All hydrogen atoms were found in the final difference Fourier map and allowed to refine freely with isotropic displacement parameters. Full-matrix least-squares refinements with 434 parameters were carried out by minimizing ∑w(F_o²–F_c²)² with the shelxl-97 [30] weighting scheme and stopped at shift/err<0.001.

CCDC 1017347 contains the supplementary crystallographic data for this paper. These data can be obtained free of charge from the Cambridge Crystallographic Data Centre via www.ccdc.cam.ac.uk/data_request/cif.