Molecular iodine as a mild catalyst for cross-coupling of alkenes and alcohols

Introduction

Carbon–carbon (C–C) bond construction is one of the most vital approaches for the synthesis of complex natural backbones and bioactive compounds such as pharmaceuticals and agrochemicals. These reactions have encouraged scientific exploration from the very beginning and have revolutionised the field of organic chemistry, as has been proven in 2010, when R. F. Heck, E. Negishi and A. Suzuki won the Nobel Prize in chemistry for Pd catalysed C–C coupling [1].

The expanding recognition of green chemistry and its 12 principles in academic and industrial community in the course of the recent decades has resulted in a movement toward a more sustainable world [2]. As alcohols and aromatic hydrocarbons are among the most abundant and commonly used feedstocks in industrial processes, their exploitation in C–C coupling reactions has become one of the significant approaches in direct functionalization of C–H bonds, avoiding the need for halogen-containing intermediates and thereby eliminating synthetic steps, which results in enhanced atom economy [3]. Alcohols are a highly attractive class of alkylating agents since they are inexpensive, they are usually easily derived from natural sources and the only side product resulting from associated coupling reaction is water [4]. On the other hand, the flexibility of the alkene functional group offers leverage for subsequent transformation into various other moieties [5]. Therefore, their dehydrative coupling is an attractive waste-free approach, resulting in generation of substituted alkenes as coupling products, which provides an excellent platform for further modification. In addition, some of these products, such as 4H-chromenes, also exhibit biological activities [6].

Previously described methods for dehydrative coupling of alcohols and alkenes can be divided into two basic approaches: metal catalysis and metal-free catalysis. The majority of known methods include the use of very effective transition metal catalysts, such as complexes of Pd, Cu, Ni, Fe, Ru and Lewis acids based on weak metals (e.g. Bi, In and Ga) [7], [8], [9], [10], [11], [12], [13], [14], [15], [16], [17]. However, toxicity of organometallic reagents and their removal to meet the strict specifications of pharmaceutical grade purity is a challenging and demanding process. Therefore, metal-free catalysis has lately gained interest, searching for greener alternatives in organic synthesis. Among the metal-free synthetic approaches, classic Brønsted [18], [19], [20] and Lewis acids [21], [22] are most commonly used, however, reactions catalysed by acid-functionalised ionic liquids [23], [24] and solid supported catalysts [25], [26] have also been reported. The aggressiveness of strong Brønsted acid/base catalysis can lead to poor reaction selectivity and the use of toxic halogenated solvents in numerous examples is not among the favourable practises due to the risks of carcinogenicity and ozone layer depletion. On the other hand, ionic liquids are considered a greener alternative to many commonly used solvents, yet they suffer from drawbacks such as toxicity, difficult preparation and high cost [27]. In fact, from the green-chemical standpoint, the best solvent for the reaction media is considered ‘no solvent’, hence solvent-free reaction conditions are preferable, if possible [28], [29], [30].

Iodine, as an easily manipulable solid, soluble in many organic solvents, has recently been recognised as a perspective catalyst in organic synthesis. Thirty thousand tons of iodine is produced annually and 16% of it is utilised in industrial catalysis [31]. Its catalytic activity is attributed to an attractive halogen bond between the iodine atoms and the substrate molecule [32], which is the consequence of a positive polarization of the iodine atoms (the σ hole) due to an anisotropic electron distribution [33], [34]. Unlike the heavy metals, iodine is an environmentally friendly and a relatively inexpensive element, which exhibits high catalytic activity even under solvent-free reaction conditions (SFRC) [35], [36], [37], [38], [39]. Aside from being a metal-free substance, it has several assets, such as broad catalytic potential, water-tolerance, lower price and environmental friendliness. It is also a weak oxidant and a weak electrophile. Therefore, it is not surprising that it is employed both as a stoichiometric reagent and as a catalyst [40], [41] in numerous organic transformations.

Iodine mediated coupling of activated aromatics, such as indole derivatives, with E-1,3-diphenyl-2-propen-1-ol has previously been reported [42]. However, to the best of our knowledge, iodine catalysed direct cross-coupling of alcohols and olefins has not been published so far. In our continuous pursuit of environmentally friendlier chemical approaches to organic synthesis [43], [44], [45], [46], [47], we present a metal-free method for direct dehydrative cross-coupling of alcohols and alkenes using molecular iodine as a mild Lewis acid catalyst under solvent-free reaction conditions. The efficiencies of previously used catalysts relative to iodine for cross-coupling of some chosen phenyl ethenes with benzyl alcohols are listed in Table 1. From the collected data it is evident that reaction yield in case of iodine use is comparable with the best yields of cross-coupling with other catalysts.

Table 1:

Comparison of catalyst efficiency for cross-coupling of phenyl ethenes and benzyl alcohols.

Entry	X	R1	R2	Catalyst	Reaction conditionsa	Yield (%)	Ref.
1	H	Ph	C≡CPh	Cu(OTf)2	DCE, reflux, 10 min	94	[5]
2	H	Ph	Ph	FeCl3·6H2O/TsOH	DCM, 45°C, 2.5 h	92	[15]
3	H	Ph	C≡CPh	FeCl3·6H2O	MeCN, 80°C, 0.5 h	81	[16]
4	p-Me	H	Ph	(CF3CO)2O/Pd(OAc)2/PPh3	DMF, 100°C, 39 h	56	[17]
5				HCl	MeCOOH (anhydrous), reflux, 1.25 h	30	[18]
6	H	Ph	Ph	TfOH	DBE, 60°C, 4–24 h	98	[19]
7	H	Ph	Ph	[BsOdP][OTf]b	DCM, 80°C, 3 h	91	[23]
8	H	H	Ph	[NMP]+HSO4−c	80°C, 3 h	90	[24]
9	H	H	Ph	SiO2/policresulen	DBE, 80°C, 5 h	92	[25]
10	H	Ph	Ph	NaHSO4/SiO2	DCE, 60°C, 4 h	90	[26]
11	H	Ph	Ph	I2	Neat, 70°C, 1 h	95	This work

1. ^aDCE, 1,2-dichloroethane; DCM, dichloromethane; DMF, dimethylformamide; DBE, 1,2-dibromoethane. ^b[BsOdP][OTf], 1octadecyl-1-(4-sulfobutyl)pyrrolidin-1-ium trifluoromethanesulfonate. ^c[NMP]⁺HSO₄⁻, N-methyl-2-pyrrolidone hydrogen sulfate.

Results and discussion

Substitution of hydroxyl group in alcohols is a common strategy for the synthesis of bioactive substances in pharmaceutical industry, which is achieved by activation of alcohol molecule. As already mentioned, this approach is attractive from the green chemical perspective, as it should furnish water as the only by-product [48]. However, the substitution of hydroxyl group is a challenging transformation, as hydroxide is a poor leaving group, thereby demanding an additional activation. Such direct activation can be achieved through S_N1 reactions on benzylic, allylic and propargylic alcohols by the use of Brønsted or Lewis acids [49].

With this in mind, the coupling of benzhydrol 1a and 1,1-diphenylethylene 2a has been chosen as a model reaction to employ iodine as a catalyst in direct cross-coupling of alkenes and alcohols and to study the effects of different reaction conditions. Initially, the effect of different solvents on conversion of 1,1-diphenylethylene was examined. Beside the solvents listed (Table 2, entries 1–6), also MeOH, EtOH, i-PrOH, Me₂CO, EtOAc and MeCN have been tested as reaction mediums; however, they all reacted with 2a and were therefore inappropriate for further studies. Quantitative conversion was achieved, when Solkane^® (1,1,1,3,3-pentafluorobutane), n-hexane or no solvent were used (Table 2, entries 4–6). As the selectivity of the reaction was approximately the same in all three cases, further examinations were carried out under neat reaction conditions. The catalyst amount was varied and a considerable increase in conversion was observed, when the amount of iodine was changed from 3 mol% to 5 mol%. The complete conversion was achieved at 7 mol% of the catalyst and no significant improvement in selectivity was detected with additional raise of the catalyst loading (Table 2, entries 6–10).

Table 2:

The effect of solvent and catalyst loading on the cross-coupling of 1a with 2a.^a

Entry	Solvent	I2 (mol%)	Conversion (%)b	Selectivityb3a/4a/5a/6a/7a
1	Water	7	23	41:41:18:0:0
2	2-MeTHF	7	40	54:35:11:0:0
3	TFE	7	99	86:3:2:9:0
4	Solkane®	7	100	79:5:3:13:0
5	n-hexane	7	100	81:6:4:9:0
6	Neat	7	100	81:6:0:13:0
7	Neat	0	0	/
8	Neat	3	57	66:28:6:0:0
9	Neat	5	98	80:8:0:12:0
10	Neat	9	100	80:8:0:12:0

1. ^aReaction conditions: 1a (1.0 mmol), 2a (1.0 mmol), I₂ (0–9 mol%), solvent (1 mL), 70°C, 24 h. ^bDetermined by ¹H NMR spectroscopy with respect to the 1,1-diphenylethylene 1a.

Furthermore, the effect of temperature on the reaction system was investigated. Although quantitative conversion has been obtained already at 40°C, the selectivity of the reaction improved, when the temperature was raised to 70°C, increasing the percentage of the desired coupling product prop-1-ene-1,1,3,3-tetrayltetrabenzene 3a in the reaction mixture from 67% to 81% (Table 3). In order to reduce the time needed for efficient conversion of the starting material, reaction time screening was subsequently performed and the results showed quantitative conversion already after 1 h without the loss of selectivity. The side products 6a and 7a are the consequence of 1,1-diphenylethylene cyclodimerisation and dimerisation, respectively, since phenyl-substituted alkenes are usually very sensitive to reaction conditions and often undergo polymerisation. As previously demonstrated by our group, when 2-phenyl propene 8 is heated in the presence of 5 mol% of iodine for 15 min at 70°C, complete conversion to indane derivative 9 is achieved (Fig. 1). The authors concluded that firstly dimeric alkene 10 is formed, which is furtherly cyclised to indane derivative [50].

Fig. 1:

Dimerisation and subsequent cyclisation of alkene 8 to indane derivative 9.

Inspired by these facts, we increased the amount of alcohol 2a from 1.0 to 1.1 equiv. with respect to alkene with the aim to move the equilibrium toward the desired coupling product 3a. The result was rewarding, as 1,1-diphenylethylene 1a was quantitatively converted into polysubstituted olefin 3a, while the remaining excessive amount of alcohol was transformed into dimeric ether 4a.

Encouraged by these promising results, we examined the scope of the reaction system by applying the obtained optimal reaction conditions on cross-coupling reaction of 1,1-diphenylethylene with different alcohols. The method proved efficient on secondary benzyl alcohols, as depicted in Table 4. The coupling reaction with benzyl alcohols bearing electron-donating groups was finished in 1 h (Table 4, entries 2 and 3), while the electron-withdrawing groups (p-F and p-NO₂) bonded to the phenyl ring of the alcohols resulted in longer reaction time of 6 h (Table 4, entries 4 and 5). In cases where benzyl position of the alcohol was substituted with alkyl group (Table 4, entries 6 and 7), the reaction was considerably slower relative to the cases where it was substituted with allyl or propargyl group (Table 4, entries 8 and 9).

Table 4:

Synthesis of polysubstituted olefins 3a–i catalysed by molecular iodine.^a

Entry	Product	Structure	Time (h)	Yield (%)b
1	3a		1	95
2	3b		1	97
3	3c		1	85
4	3d		6	92
5	3e		6	90
6	3f		21	95
7	3g		24	87
8	3h		3	96
9	3i		2	97

1. ^aReaction conditions: all reactions were carried out under neat reaction conditions at 70°C using 1,1-diphenylethylene 1a (1.0 mmol), alcohol (1.1 mmol) and I₂ (0.07 mol). ^bIn all cases the conversion of the starting material was >95% and the yields of the pure products were calculated based on 1a.

These results are consistent with our expectations related to the impact of electron donating/accepting moieties and conjugation on stabilisation of the carbocation, formed from alcohol molecule by dehydroxylation. To confirm the synthetic value of the presented methodology, the coupling reaction of 1a and 2a was performed also on 20 mmol scale (3.6 g of alkene 1a and 4.0 g of 2a) and 95% of the pure coupling product was isolated after 3 h.

In order to get clearer insight into the course of the reaction pathway, some control experiments were performed. Under the typical reaction conditions (Table 4) after 1 h benzhydrol 2a was quantitatively converted into dimeric ether 4a and when the latter was subsequently used as a starting material instead of 2a in the coupling reaction with 1,1-diphenylethylene 1a, the complete conversion to the polysubstituted olefin 3a was detected after 3 h (Fig. 2). These results strongly suggest the involvement of the dimeric ether 4a as the intermediate in the reaction pathway. The formation of 4a could be explained by the reaction between a molecule of alcohol and a stable benzylic carbocation, derived from another alcohol molecule. However, the direct reaction between carbocation and alkene molecule cannot be completely excluded. The proposed reaction pathway is illustrated in Fig. 3. To confirm the involvement of the intermediate 11, the reaction between asymmetric alkene 1-methoxy-4-(1-phenylvinyl)benzene 1b and benzhydrol 2a was conducted, leading to 100% conversion of the 1b to the E- (3j) and Z- (3j′) isomers of the corresponding coupling product in ratio 1:1 (Fig. 4). The planar molecular geometry of the carbocation intermediate 11a as the consequence of sp² hybridization on the carbon atom bearing the positive charge allows deprotonation from either of the planar sides of the molecule with the same probability, therefore yielding E- and Z- isomer in equal ratios.

Fig. 2:

Control experiments.

Fig. 3:

Plausible reaction pathway for the direct dehydrative coupling.

Fig. 4:

The reaction of 1-methoxy-4-(1-phenylvinyl)benzene 1b with benzhydrol 2a.

In addition, the dehydrative cross-coupling reaction of benzhydrol with styrene as one of the most interesting and widely available phenyl-substituted alkene was examined. Unfortunately, beside the selective formation of (E)-prop-2-ene-1,1,3-triyltribenzene, a considerable amount of polymeric material was formed. Similarly, α-alkyl substituted styrene derivatives were also found inappropriate alkene starting materials due to the diversity of possible deprotonation positions in the carbocation intermediate 11 during the C-step of the reaction pathway.

Due to the low cost and ready availability of some alcohols as starting materials relative to alkenes and in behalf of the well-known proneness to elimination reaction of tertiary alcohols in the presence of Brønsted/Lewis acids, the possibility of the use of tertiary benzyl alcohols as potential substitutes for the alkenes as starting compounds was investigated. The equimolar mixture of 1,1-diphenylethanol 12 and 1-phenylethanol 2f in the presence of I₂ (7 mol%) was heated for 24 h, resulting in formation of polysubstituted olefin 3f in 65% yield (Fig. 5).

Fig. 5:

The reaction between tertiary and secondary benzyl alcohol 12 and 2f.

Conclusion

In summary, a convenient method for direct dehydrative coupling of various benzylic alcohols with phenyl-substituted alkenes was developed, using iodine as an environmentally friendly metal-free Lewis acid catalyst under solvent-free reaction conditions. The reaction is assumed to proceed via carbocation intermediate, furnishing polysubstituted olefins in high yields. The scale-up procedure was conducted, verifying the synthetic applicability of the presented method. Finally, the possibility of direct coupling between tertiary and secondary benzyl alcohols was examined.

Experimental section

Supporting information

The copies of ¹H, ¹³C and ¹⁹F NMR of the purified products and other experimental procedures are available online.