Palladium-catalyzed α-arylation of benzylic phosphonates.

A new synthetic route to access diarylmethyl phosphonates is presented. The transformation enables the introduction of aromatic groups on benzylic phosphonates via a deprotonative cross-coupling process (DCCP). The Pd(OAc)2/CataCXium A-based catalyst afforded a reaction between benzyl diisopropyl phosphonate derivatives and aryl bromides in good to excellent isolated yields (64-92%).

Phosphonates and their derivatives are a class of organophosphorus compounds that exhibit a wide range of applications in medicinal[1] and agricultural chemistry.[2] They are also flame retardants, metal extractants,[3] and reagents in the Horner–Wadsworth–Emmons (HWE) reaction.[4] Given the interest in phosphonates, and their applications as reagents to prepare a host of useful molecules (Figure 1),[5] there is a significant demand for their synthesis.

Figure 1

Examples of interesting molecules obtained from dialkyl (diarylmethyl)phosphonates.

Despite the widespread use of dialkyl (diarylmethyl)phosphonates, only a few methods for their synthesis have been reported. Classical preparations include Michaelis–Arbuzov or Michaelis–Becker reactions,[6] which are dependent on the limited availability of the starting diarylmethyl halides. Recently, a Friedel–Crafts type reaction has been developed by Chakravarty et al. starting from diethyl (hydroxy(aryl)methyl)phosphonates to obtain diethyl (diarylmethyl)phosphonates (Figure 2A).[7] The products were obtained in good yields, although the scope suffers from the usual limitations of Friedel–Crafts reactions. To date, some examples of α-arylation of activated phosphonates have been reported.[8] They involved the deprotonation of relatively acidic methylene C–H’s (pKa around 17 in DMSO) sandwiched between phosphonates and a second electron-withdrawing group (Figure 2B–D). One example of palladium-catalyzed α-arylation of dimethyl methylphosphonate in 70% yield has been reported by Hagadorn and Hlavinka using Zn(tmp)2 as the base (Figure 2E).[9] This approach employs an irreversible deprotonation, rather than a reversible deprotonation employed in our work.[10] In this Letter, we disclose the first general palladium-catalyzed direct α-arylation of dialkyl benzyl phosphonates (Figure 2F). These substrates (pKa = 27.6 in DMSO) are around 10 orders of magnitude less acidic than those in Figure 2B–D, making their arylation considerably more challenging.

Figure 2

α-Arylation and Friedel–Crafts type reaction of phosphonates.

Our group is interested in the functionalization of weakly acidic sp3 C–H bonds via deprotonative cross-coupling processes (DCCP). Substrates employed to date include diarylmethanes, sulfoxides, sulfones, amides,[11]’[12] and most recently phosphine oxides.[13] Encouraged by the utility of phosphonates, and the lack of general methods to α-arylate weakly acidic members of these compounds, we set out to develop the arylation of benzylic phosphonates. Based on our previous observations in the arylation of phosphine oxides,[13] our investigations were initiated using six bases [LiOt-Bu, NaOt-Bu, KOt-Bu, LiN(SiMe3)2, NaN(SiMe3)2, and KN(SiMe3)2], two palladium sources [Pd(OAc)2 and Pd(dba)2], two ligands [CataCXium A (L1)[14] and Xantphos (L2),[15] Figure 3], and four solvents [CPME (cyclopentyl methyl ether), 1,4-dioxane, THF, and DME], using microscale (10 μmol) High-Throughput Experimentation (HTE) techniques (see Supporting Information).

Figure 3

Structures of CataCXium A (L1) and Xantphos (L2).

From this microscale screen, two hits were obtained with NaOt-Bu, Pd(OAc)2 and CataCXium A (L1, Figure 3) or Xantphos (L2, Figure 3) in CPME at 80 °C. On laboratory scale (0.2 mmol), these two systems led to the desired arylation product in 85% and 84% yield, respectively (Table 1, entries 1–2). Changing the concentration from 0.1 to 0.2 M resulted in an increase in the yields to 99% and 98%, respectively (entries 3–4). To differentiate these two catalysts, reactions with an electron-rich and -poor aryl bromide were tested. While CataCXium A (L1) generated the desired products in 99% yield with 4-methoxy bromobenzene and with 4-fluoro bromobenzene (entries 5 and 7), the use of Xantphos (L2) resulted in a drop in the yield to 79% and 88%, respectively (entries 6 and 8). Based on these results, CataCXium A (L1) was chosen as the more suitable ligand for this reaction. Interestingly, Xantphos proved to be a better ligand for the α-arylation of phosphine oxides.[13] Note that attempts to reduce the amount of base or the temperature resulted in a drop in the yield (entries 9 and 10).

Table 1

Optimization of α-Arylation of Benzyl Diisopropyl Phosphonate with Bromobenzene, 4-Methoxy Bromobenzene, and 4-Fluoro Bromobenzene

entry	R	ligand	base (equiv)	temp (°C)	concn (mol/L)	yielda (%)
1	H	L1	3	80	0.1	85%
2	H	L2	3	80	0.1	84%
3	H	L1	3	80	0.2	99%
4	H	L2	3	80	0.2	98%
5	OMe	L1	3	80	0.2	99%
6	OMe	L2	3	80	0.2	79%
7	F	L1	3	80	0.2	99%
8	F	L2	3	80	0.2	88%
9	H	L1	2	80	0.2	70%
10	H	L1	3	50	0.2	24%

Yields determined by 1H NMR of the crude reaction mixtures using CH2Br2 as the internal standard.

The optimized reaction conditions in Table 1 were then used to determine the substrate scope of aryl bromides 2a–j (Scheme 1). Bromobenzene underwent a reaction to give 3a in 91% isolated yield. It is noteworthy that an 88% yield was obtained when the reaction was carried out with 1 g of phosphonate 1a and bromobenzene, indicating that the catalyst and conditions are scalable. Reaction with 4-tert-butyl bromobenzene afforded compound 3b in 82% yield. No difficulty was noted with electron-donating groups, such as 4-methoxy bromobenzene and 4-bromo-N,N-dimethylaniline, which resulted in the formation of products 3c and 3d in 86% and 82% yield, respectively. Reaction with electron-withdrawing groups also proceeded well. With 4-fluoro bromobenzene, 3e was isolated in 82% yield. Similar reactivity was observed with 4-chloro bromobenzene, although a longer reaction time was required to reach complete conversion. Sterically hindered 2-methyl bromobenzene and 1-bromonaphthalene afforded products 3g and 3h both in 89% isolated yield. Finally, introduction of heteroaromatics 5-bromobenzofuran and 5-bromo-N-methyl indole were achieved in 80% and 71% yield, respectively.

Scheme 1

Substrate Scope of Aryl Bromides 2a–j in Pd-Catalyzed α-Arylation of Diisopropyl Benzyl Phosphonate 1a

1a: 3.90 mmol, 1.0 g. 32 h reaction time.

The substrate scope of the diisopropyl benzyl phosphonate was then studied (Scheme 2). The benzyl group in 1a was replaced with 4-methoxybenzyl (1b), 4-fluorobenzyl (1c), 2-methylbenzyl (1d), and 3-methylpyridyl (1e) groups. In each case, the substrate was coupled with neutral, electron-rich, and electron-poor aryl bromides (bromobenzene, 4-methoxy bromobenzene, and 4-fluoro bromobenzene). Phosphonate 1b, possessing a 4-methoxybenzyl group, afforded coupled products 3c, 3k, and 3l in 60–82% yield. It is noteworthy that a longer reaction time and/or a higher temperature was required for these substrates, probably due to the lower acidity of the benzylic protons of 1b. Compounds 3e, 3l, and 3m were obtained from 4-fluoro benzyl phosphonate (1c) in 77–84% yield. Sterically hindered 2-methyl benzyl phosphonate (1d) afforded 3g, 3n, and 3o in 73–92% yield. A longer reaction time and a higher temperature were, however, needed to reach complete conversion. Finally, coupling products 3p–r were obtained from 3-pyridyl containing phosphonate 1e in 67–84% yield after 24 h or 48 h of reaction time.

Scheme 2

Substrate Scope of Diisopropyl Benzyl Phosphonate in the Pd-Catalyzed α-Arylation with Aryl Bromides

32 h reaction time. 24 h reaction time, 110 °C. 24 h reaction time. 48 h reaction time.

Replacement of the isopropyl moieties of 1a has also been considered. Ethoxy groups are among the most commonly employed substituents on the phosphorus atom in phosphonate chemistry. We, therefore, applied our α-arylation conditions to diethyl benzylphosphonate 1f (Scheme 3). Unfortunately, the coupling did not proceed smoothly. Only a 30% yield of the product was obtained, along with the formation of a byproduct, tert-butyl ethyl benzylphosphonate BP1 (∼ 20% yield), and some degradation of the starting material. Changing the base did not solve the problem, as degradation was observed with all bases employed. Only the nature of the major byproduct formed differed: tert-butyl ethyl benzylphosphonate BP1 when tert-butoxide bases (Li, Na, K) were employed and diethyl (1-phenylpropyl)phosphonate BP2 when the MN(SiMe3)2 (M = Li, Na, K) bases were used. Formation of BP2 presumably also gives rise to BP3, but this charged species was not isolated. After extensive optimization, slow addition of sodium tert-butoxide to the reaction media (0.1 mL/h, see Supporting Information for details) avoided the degradation of the diethyl benzylphosphonate 1f and decreased the formation of the byproduct BP1. The product 3s was isolated in 64% yield.

Scheme 3

α-Arylation of Diethyl Benzylphosphonate 1f with Bromobenzene

In summary, we have developed the first Pd-catalyzed α-arylation of benzylic phosphonates with aryl bromides. Despite the perceived similarity between phosphine oxides and phosphonates, different catalysts gave the best results for each substrate class. For phosphonates, the combination of Pd(OAc)2, CataCXium A, and NaOt-Bu in CPME enabled access to useful diarylmethyl phosphonates in good to excellent yields through a deprotonative cross-coupling process.