XBphos-Rh: a halogen-bond assembled supramolecular catalyst.

The use of halogen bonding as a tool to construct a catalyst backbone is reported. Specifically, pyridyl- and iodotetrafluoroaryl-substituted phosphines were assembled in the presence of a rhodium(i) precursor to form the corresponding halogen-bonded complex XBphos-Rh. The presence of fluorine substituents at the iodo-containing supramolecular motif was not necessary for halogen bonding to occur due to the template effect exerted by the rhodium center during formation of the halogen-bonded complex. The halogen-bonded supramolecular complexes were successfully tested in the catalytic hydroboration of terminal alkynes.

Introduction

The evolution of metal-based homogeneous catalysis has run in parallel with the development of structurally diverse ligands.1 Ligand design allows the behaviour of the metal centre to be influenced and its catalytic activity and selectivity to be modified.

Supramolecular chemistry has emerged as a more efficient tool for the construction of (enantiopure) ligand backbones of metal complexes when compared with standard covalent chemistry. Supramolecular strategies mainly rely on the self-assembly of building blocks that contain both the complementary motifs required for the assembly, and the binding groups necessary for the desired catalysis event. This approach has been successfully employed for preparing catalysts by assembling supramolecularly complementary building blocks through hydrogen-bonding, metal-ligand, or ionic interactions.2 However, the use of halogen bonding for constructing the backbone of a catalyst remains, to the best of our knowledge, unexplored.3

The directionality and strength of halogen bonding,4 together with a greater tolerance to solvent polarity changes,5 prompted us to design and develop new halogen-bonded metal catalysts (see Scheme 1). We speculated that the halogen bonding-mediated assembly of two appropriate building blocks, each bearing a ligating group, could lead to the formation of a metal complex due to the chelation effect that should be exhibited in the presence of the metal centre. With regard to the ligating groups, we focused our attention on phosphino groups due to their pre-eminence in homogeneous catalysis6 and their tolerance to functionalisation.7 With respect to the halogen bonding motifs, we envisaged that the use of a fluorinated iodoarene and a pyridine group could be suitable due to the reported complementarity of both supramolecular motifs.8 Concerning the metal centre, we focused on rhodium as its complexes are catalytically active in pivotal organic transformations. In the discussion that follows, we report our results in the preparation and full characterisation of the first examples of halogen-bonded rhodium(i) complexes and their catalytic performance in alkyne hydroborations.

Scheme 1

Supramolecular halogen-bonded catalysts (A: halogen bond acceptor; X: halogen bond donor; M: metal centre).

Results and discussion

Examination of CPK models revealed that the relative arrangements of the P-ligating groups and the halogen bond acceptor and donor motifs in 2-pyridyldiphenylphosphine (1) and 2-iodo-3,4,5,6-tetrafluorophenyldiphenylphosphine (2), respectively, were adequate for the formation of the target halogen-bonded Rh(i) complex (see Scheme 2). Building blocks 1 and 2 were successfully prepared in good yields in a multigram scale by slight modification of well-established synthetic protocols.9 We envisaged that the use of [Rh(Cl)(CO)2]2 along with a halide scavenger would render a putative [Rh(CO)2]X intermediate that should react with phosphines 1 and 2. After some experimentation, halogen-bonded rhodium(i) complexes were obtained by first reacting [Rh(Cl)(CO)2]2 and NaBArF,10 and then adding ligands 1 and 2 as indicated in Scheme 2.11 The structure of the complex was confirmed by standard spectroscopic studies. The 31P{1H} NMR spectrum showed two complex multiplets, from which a high value 2JP–P coupling constant (i.e. 276 Hz; unequivocally assigned by 2D 31P{1H} J-resolved spectroscopy; Fig. SI 52 ESI†) was calculated. According to the literature reviewed, this high value 2JP–P coupling constant indicates a trans-spanning bisphosphine coordinated to the rhodium centre.12 We initially expected a dicarbonyl complex, but ESI-MS analysis identified complex [Rh(CO)(1)(2)]BArF (4), referred to as XBphos-Rh13 in the discussion that follows.

Scheme 2

Synthesis of halogen-bonded rhodium(i) chelates.

X-ray analysis confirmed the proposed halogen-bonded structure of the rhodium complex derived from 1 and 2 (see Scheme 2 for the structure of XBphos-Rh and Fig. 1A for the crystal structure). The nitrogen and iodine are aligned with a N–I distance of 2.757(8) Å and with a N···I–C bond angle of 169.4(5)°. These structural parameters are in agreement with those reported for halogen-bonded complexes involving pyridine and fluorinated iodoaryl moieties.14 Interestingly, the iodine appears to be coordinated to the rhodium centre (Rh–I distance = 2.5535(7) Å). Thus, the tridentate coordination of the halogen-bonded ligands in square planar XBphos-Rh resembles that observed in pincer-type complexes,15 and therefore this complex can be formally considered as the first reported P–I–P pincer-type complex.

Fig. 1

Crystal structures of XBphos-Rh (A) and 5 (B). Hydrogen atoms and the BArF unit have been omitted for the sake of clarity. Colour scheme: C: black, P: orange, Rh: green, F: green, N: blue, I: purple, O: red. Atomic displacement ellipsoids are drawn at a 50% probability.

Given the favourable template effect exerted by the rhodium centre on the coordinated atoms in XBphos-Rh, we envisaged that fluorine-free compound 3 could also be used as the halogen bond donor. Thus, complex 5 could be prepared in good yields following a synthetic protocol analogous to that employed for XBphos-Rh (Scheme 2). Crystals from 5 suitable for X-ray analysis confirmed the proposed structure for the rhodium complex derived from 1 and 3 (Scheme 2 and Fig. 1B).

The analysis of the solid-state structures revealed a longer N–I distance for 5 than for XBphos-Rh (2.84(1) Å for 5 and 2.757(8) Å for 4), which indicates a stronger halogen bond interaction in the fluorine-containing complex.

The formation of a halogen bonding interaction between the pyridine and iodoarene motifs is driven by an enthalpy gain but penalised by an entropic unfavourable contribution arising from the geometry constraints for effective bonding (i.e. linear arrangement of the N···I–C atoms). Overall, the free energy associated with a single interaction is low.16 In order to gain a deeper insight into the halogen bonding event, full level DFT calculations were carried out. Geometries and stabilities of [Rh(CO)(1)(2)]+ and [Rh(CO)(1)(3)]+ and those of the starting building blocks were computed with Gaussian 09 (ref. 17) with the TPSS18-D3 19 density functional employing a medium-sized triple-zeta (def2TZVP20) basis set, which is a good compromise between the computational cost and the reliability in describing halogen bonding.21 Selected geometrical parameters and computed binding free energies are summarised in Table 1. To aid comparison, the supramolecular interaction between 4MePy and IPFB (see footnote a in Table 1 for the abbreviations) was also computed and was found to be a slightly endergonic process with a calculated binding constant value being in agreement with that experimentally measured (i.e.; Kexp = 1 ± 1 M–1 in toluene in ref. 16 and Kcalc = 0.03 M–1 from the predicted ΔΔG value in entry 1 in Table 1).

Table 1

Selected geometrical parameters for halogen-bonded complexes (TPSS-D3/def2TZVP, 298 K, toluene) and computed binding free energies

Entry	Complex	d N–I (Å)	N···I–C bond angle (°)	ΔΔG (kcal· mol–1)
1	4MePy·IPFB a	2.714	179.4	+2.03 b
2 c	[Rh(CO)(1)(2)]+	2.709 (2.757(8))	175.2 (169.4(5))	–112.83 d
3 c	[Rh(CO)(1)(3)]+	2.827 (2.84(1))	176.2 (171.2(7))	–111.97 d

4MePy·IPFB = complex between 4-methylpyridine and iodopentafluorobenzene.

ΔΔG = ΔG4MePy·IPFB – ΔG4MePy – ΔGIPFB.

X-ray values in parentheses.

ΔΔG = ΔG[Rh(CO)()(or)]+ + ΔGCO – ΔG[Rh(CO)+ – ΔG – ΔGor.

The computed N–I distances were in agreement with the experimental values and their formations were calculated to be highly exergonic processes. Computational studies confirmed a slightly higher stability for [Rh(CO)(1)(2)]+ than for [Rh(CO)(1)(3)]+ (entries 2 and 3, Table 1). Electrostatic potential surfaces were calculated at the level of theory described above (see Fig. 2 for 2 and [Rh(CO)(1)(2)]+ and ESI† for all information). The maximum values of the σ-hole in 2 (Fig. 2) and fluorine-free derivative 3 (ESI†) are 25.1 and 12.5 kcal.mol–1, respectively. A lower σ-hole value indicates a lower donor character of the halogen atom to the supramolecular bond. This observation derived from electrostatic potentials is in agreement with previously discussed X-ray data (i.e. a longer N–I bond for 5 than for XBphos-Rh).

Fig. 2

Electrostatic potential surfaces at an isovalue of 0.001 a.u.

The interesting structural features of XBphos-Rh and 5 (Table 2),22 prompted us to investigate reactivity features as catalysts that could complement existing catalytic tools. These complexes were tested in the hydroboration of terminal alkynes towards vinylboronic acid derivatives, as the outcome of these reactions is heavily influenced by the electronic and steric properties of the ligand.23 Methods for synthesizing vinylboronic acid derivatives, which are important synthetic intermediates, by an atom-economical hydroboration reaction24 are scarce in the literature.23,25

Table 2

Selected bond distances (Å) and angles (°)

Entry	Complex	P1–P2	Rh–CO	P1–Rh–P2
1 b	[Rh(CO)(xant)]BF4	4.506(2)	1.798(5)	164.42(4)
2	XBphos-Rh	4.618(2)	1.872(8)	177.09(7)
3	5	4.632(9)	1.848(9)	178.6(3)

X-ray values.

xant = Xantphos; see ref. 22a.

To aid comparison of the results obtained with XBphos-Rh and 5, we included the monodentate [Rh(CO)(PPh3)3]BArF (6) complex in our catalytic studies.26 A summary of the results obtained is shown in Table 3. Both XBphos-Rh and 5 were active hydroboration catalysts, and E-isomers 8 were preferentially formed in all cases. Although XBphos-Rh is the most selective hydroboration catalyst for phenylacetylene 7a (88% overall yield for HBpin and 51% for HBcat; entries 1 and 4 in Table 3), hydroboration selectivities for 1-octyne 7b using XBphos-Rh are slightly lower than those with 6 (entries 7–12, Table 3). It is important to mention that in all cases selectivity towards branched derivatives is higher with the halogen-bonded catalysts than with the background catalyst 6.26 For example, a ten-fold increase in yield was obtained for the branched product 10a,pin when XBphos-Rh was used (entries 1 and 3, Table 3). It is also noteworthy that, as far as we are aware, XBphos-Rh provides the highest reported yield for the branched derivative 10b,cat (44% branched product with respect to all hydroboration products, entry 10, Table 3). Regarding electronic effects of the product distribution in the hydroboration of aryl-substituted acetylenes, higher ratios of the branched product were obtained for electron-deficient arene 7c than for electron rich derivative 7d (entries 13 and 14, Table 3).

Table 3

Rh-catalysed hydroboration of alkynes

Entry	Alkyne, (R2O)2BH b	Cat.	8 + 9 + 10 yield%	Ratio 8 : 9 : 10
1 c	7a, HBpin	XBphos-Rh	88	78 : 2 : 20 d
2	7a, HBpin	5	36	72 : 5 : 23
3	7a, HBpin	6	54	94 : 3 : 3
4	7a, HBcat	XBphos-Rh	51	82 : 2 : 16
5	7a, HBcat	5	49	82 : 2 : 16
6	7a, HBcat	6	50	86 : 9 : 5
7	7b, HBpin	XBphos-Rh	49	79 : 4 : 17
8	7b, HBpin	5	50	79 : 4 : 17
9	7b, HBpin	6	59	93 : 2 : 5
10	7b, HBcat	XBphos-Rh	50	52 : 4 : 44
11	7b, HBcat	5	51	66 : 4 : 30
12	7b, HBcat	6	58	86 : 4 : 10
13 c	7c, HBpin	XBphos-Rh	58	71 : 2 : 27
14 c	7d, HBpin	XBphos-Rh	74	83 : 3 : 14

Reactions were performed in CD2Cl2 (0.5 M) under N2 and reacted for 24 h at room temperature. Yields were determined by 1H NMR using 1,2,4,5-tetramethylbenzene as the internal standard. See ESI† for details.

(R2O)2BH: HBpin = pinacolborane, HBcat = catecholborane.

2 equiv. of (R2O)2BH.

Results obtained with bidentate complex [Rh(CO)(Xantphos)]BArF as catalyst in the reaction of 7a and HBpin are as follows: 69% overall yield (ratio 8 : 9 : 10 = 97 : 2 : 1).

Conclusions

In summary, we have successfully prepared and fully characterised the first halogen-bonded supramolecular rhodium(i) complexes. To the best of our knowledge, these are the first examples where halogen bonding is the driving force for the assembly of the catalyst backbone. Furthermore, we have demonstrated that the presence of fluorine groups at the iodo-containing supramolecular motif is not necessary for efficient halogen bonding due to the favourable template effect exerted by the rhodium centre. The two new catalysts are active in the hydroboration of terminal alkynes, with the catalytic activity of XBphos-Rh rivalling that of the highest performing catalysts. Moreover, this catalyst favours enhanced ratios of the branched alkenyl boronic acid products. Ongoing investigations in our group include mechanistic studies on the outcome of the reaction,27 application of this system to other catalytically relevant transformations, extension of the application of this halogen bonding approach to other transition metals, and studies into alternative ligand designs.

Conflicts of interest

There are no conflicts to declare.

Click here for additional data file.

Click here for additional data file.