The Natural Product Lepidiline A as an N-Heterocyclic Carbene Ligand Precursor in Complexes of the Type [Ir(cod)(NHC)PPh₃)]X: Synthesis, Characterisation, and Application in Hydrogen Isotope Exchange Catalysis

Abstract

A range of iridium(I) complexes of the type [Ir(cod)(NHC)PPh₃)]X are reported, where the N-heterocyclic carbene (NHC) is derived from the naturally-occurring imidaozlium salt, Lepidiline A (1,3-dibenzyl-4,5-dimethylimidazolium chloride). A range of complexes were prepared, with a number of NHC ligands and counter-ions, and various steric and electronic parameters of these complexes were evaluated. The activity of the [Ir(cod)(NHC)PPh₃)]X complexes in hydrogen isotope exchange reactions was then studied, and compared to established iridium(I) complexes.

1. Introduction

Isotopically labelled compounds are of importance in a range of scientific disciplines. In particular, the isotopes of hydrogen, deuterium, and tritium find application in several key areas. Within the physical sciences, deuterated compounds are routinely used in the elucidation of reaction mechanisms [1]. In the life sciences, deuterated and tritiated compounds are used to determine the pharmacokinetic properties of bioactive molecules [2,3,4]. More recently, active pharmaceutical ingredients themselves have begun to feature deuterium, exploiting the effect of relative C-H/C-D bond strengths to deliver drug molecules with improved metabolic profiles [5,6].

Amongst the synthetic methods available for incorporation of deuterium or tritium into a molecule, hydrogen isotope exchange (HIE) stands out due to its simplicity and generality [7,8,9]. A particularly powerful variant of this process is directed HIE, whereby a Lewis basic group within the molecule directs a metal catalyst to an adjacent site, and a subsequent C-H activation results in a metallocyclic intermediate which facilitates HIE at this adjacent position (Scheme 1).

Iridium(I) complexes are amongst the most effective metal catalysts for directed HIE [10] and, in recent years, we have introduced a suite of highly active Ir(I) catalysts bearing a combination of bulky phosphine and N-heterocyclic carbene ligands (Figure 1). These catalysts are compatible with a broad variety of directing groups, enabling HIE on a wide range of substrates [11,12,13,14,15,16,17,18].

These benchmark complexes 1–2 feature three distinct phosphine ligands. However, our early studies [11,12], which focused on identifying the most effective catalyst system, evaluated a broader range of phosphines based on their steric and electronic properties. In contrast, although we have optimised the NHC ligand for some specific HIE use cases [19,20], there remains scope for a more systematic study of the steric and electronic parameters of this ligand component. The natural product Lepidiline A 3·HCl is an imidazolium chloride salt, featuring two N-benzyl substituents, and methyl groups at the heterocycle’s 4- and 5-positions [21] (Figure 2). We envisaged that this naturally-occurring imidazolium salt could serve as a vehicle to begin to explore the effect of steric and electronic changes in the NHC on the activity of our Ir(I) NHC/phosphine catalysts [22].

Herein we report a study of the complexes 1a, 6a and 7, of the type [Ir(cod)(NHC)PPh₃)]PF₆ (cod = 1,5-cyclooctadiene), where the NHC ligand is IMes 4, IBn 5, or IBn^Me 3, derived from Lepidiline A 3·HCl (Figure 3). These complexes are then used to explore the effect of steric and electronic variance within the NHC through comparison of (a) N-benzyl vs. N-mesityl substituents; and (b) 4,5-H₂ vs. 4,5-Me₂ substitution, on the reactivity in HIE processes. The effect of the counter on the activity of the IBn^Me complexes is also investigated.

2. Results and Discussion

To initiate our studies, we set out to prepare quantities of the desired [Ir(cod)(NHC)PPh₃)]PF₆ complexes 1a, 6a and 7. Complex 1a was prepared by a modification of previously reported conditions [11,23]. The IBn-containing complex 7 was prepared [23] by using an analogous procedure from the previously reported complex [Ir(cod)(IBn)Cl] [24]. Lastly, the imidazolium chloride salt 3·HCl, representing the natural product Lepidiline A, was targeted to deliver complex 6a.

2.1. Synthesis of Lepidiline A

Surprisingly, to our knowledge, no preparation of this naturally-occurring material has yet been reported. Our synthetic route is summarised in Scheme 2. Chlorination of the commercially available hydroxymethyl imidazolium salt 8 in neat thionyl chloride was easily achieved in excellent yield on a multi-gram scale to deliver 9. Reduction and basification of 9 yielded the free 4,5-dimethylimidazole 10 with good levels of efficiency. Subsequently, double-alkylation of 10 under basic conditions could be achieved using benzyl chloride to give Lepidiline A 3·HCl with good effectiveness, or with benzyl bromide to give 3·HBr in excellent yield. Imidazolium bromide 3·HBr could then undergo salt metathesis to very efficiently afford 3·HPF₆ and 3·HBAr^F.

Following the successful preparation of Lepidiline A 3·HCl and a range of its counterion derivatives, X-ray quality crystals of 3·HCl were grown and the structure solved to reveal a molecular footprint in direct agreement with the spectroscopic data and the reported natural product structure. The crystal structure was found to be a solvate containing one molecule of water per cation (Figure 4, left). The same cationic motif was found in the X-ray structure of 3·HPF₆. This latter structure has two crystallographically independent salt units per asymmetric unit (Z’ = 2). Both cations in this structure adopt anti conformations of the benzyl rings with respect to the heteroaromatic ring (Figure 4, right). This is in contrast to the syn conformation found for 3·HCl and in the Ir complexes described below.

2.2. Preparation of Iridium Complexes Based on Lepidiline A 3

With a range of imidazolium salt precursors in hand, formation of the corresponding iridium complexes was investigated. Based on work by Wang and Lin [25], and Liu et al. [26], the [(cod)Ir(NHC)Cl] complexes can be prepared via a transmetallation procedure from the corresponding silver bis-carbene complex. Employing the bromide salt 3·HBr, mixing with silver(I) oxide and sodium iodide in dichloromethane gave silver complex 11 in high yield (Scheme 3). This intermediate could be taken forward into a transmetallation procedure with [(cod)IrCl]₂ 13 to provide the desired chloro-carbene complex 12 in excellent yield following a short reaction time. Alternatively, it was also found that isolation of 11 could be circumvented via direct addition of [(cod)IrCl]₂ 13 following in situ formation of the silver complex 11, delivering 12 in a further improved overall yield. With this high-yielding method in place, X-ray quality crystals of 12 were grown, confirming the proposed structure (Figure 5).

With quantities of the NHC chloride complex 12 secured, treatment with triphenylphosphine, followed by silver hexafluorophosphate, delivered target NHC-phosphine complex 6a in a good 75% yield (

Table 1. Preparation of [(cod)Ir(IBn^Me)PPh₃]X complexes 6a–6d.

Entry	Complex	X	Yield
1	6a	PF6	75
2	6b	BF4	87
3	6c	SbF6	81
4	6d	OTf	88

, entry 1). In order to probe the effects of varying the counter-ion in these novel complexes, derivatives 6b–6d were also prepared in an analogous manner and with excellent isolated yields (Table 1, entries 2−4).

We have previously shown that iridium NHC/phosphine complexes bearing a BAr^F (tetrakis [3,5-bis(trifluoromethyl)phenyl]borate) counter-ion have a broader solvent applicability than the corresponding PF₆ complexes [13]. Based on this, BAr^F-containing complex 6e was prepared in a one-pot procedure from [(cod)IrCl]₂ 13 in a good 82% yield (Scheme 4). Complexes 6a–6e were all characterized by X-ray crystallography [23].

2.3. Steric and Electronic Characterisation of the Iridium NHC/Phosphine Complexes

Having successfully prepared a range of Ir complexes containing the Lepidiline A-derived NHC 3, attention now turned to evaluating this novel ligand sphere versus the alternative ligand sets in PPh₃/IMes complex 1a and PPh₃/IBn complex 7. The electronic properties of ligands in a specific complex are often characterised by carbonylation of the complex, whereby the ν_CO stretching frequency reports on the electron-donating nature of the other ligands on the metal [27]. Recent work on Ir(I) dicarbonyl complexes has revealed that the configuration of the carbonyl ligands, though preferentially cis, is dictated by the ligand size, with larger combinations driving a trans dicarbonyl arrangement [28]. This structural fluidity limits the use of electronic parameters derived from such complexes [28]. However, the synthesis and X-ray characterisation of the carbonlyated analogues of our NHC/phosphine complexes could instead be used to indicate a steric threshold—the ligand size at which Ir(I) carbonyl complexes switch from a cis- to a trans-configuration, and this could be correlated with the well-known steric parameter, percentage buried volume, %V_bur, [29] of the ligands.

Accordingly, we investigated the carbonylation of complexes 1a, 6a, and 7 (Scheme 5). Direct exposure of complexes 6a and 7 to an atmosphere of CO resulted in complexes 14 and 15, where both the NHC/phosphine pair and the two CO ligands are each cis-orientated, as shown by the X-ray crystal structures [23]. The combined ligand %V_bur values (Σ(%V_bur)) were similar for both complexes, as determined by density functional theory (DFT) calculations (

Table 2. Properties of Ir-carbonyl complexes 14–16.

Entry	Complex	N-Substituent	R	Yield	%Vbur (PPh3) 2	%Vbur (NHC) 2	Σ (%Vbur)	θ (°) 1
1	14	Bn	H	49%	26.9	28.9	55.8	93.2
2	15	Bn	Me	50%	27.0	29.2	56.2	89.4 3
								92.6 3
3	16	Mes	H	68%	27.9	32.6	60.5	160.0

¹ θ represents the angle OC-Ir-CO. ² Representative %V_bur values are calculated from DFT-optimized [(NHC)(PPh₃(Ir(H)₂(DCM)₂]⁺ complex cations. Full details are presented in the SI. ³ Carbonyl complex 15 crystallized with two different conformations.

) [23]. As perhaps expected, the enhanced bulk of the NHC ligand in 1a resulted in a complex, 16, with a trans-arrangement of the NHC/P pair and of the two CO ligands, indicating the size and geometry of this complex to be quite distinct from the IBn and IBn^Me analogues 14 and 15. Based on these parameters, there is little steric difference evident on moving from the IBn to the IBn^Me ligand, with the latter appearing to be a very slightly bulkier system, based on the %Vbur values.

Although these dicarbonyl complexes permitted only a steric, and not electronic, evaluation, literature examples of Ir(I) chloro-carbonyl complexes have demonstrated that the combined influence of various ligand partnerships on the metal’s π-donating ability can be measured via IR spectroscopy, based on the carbonyl stretching frequency, ν_CO [26,30]. Furthermore, the effect of the ligands on the σ-donating nature of the complex can be measured by ¹H nuclear magnetic resonance (NMR) spectroscopic evaluation of the corresponding dihydride complexes [31]. The electronic characteristics of the IMes, IBn, and IBn^Me ligands in our complexes were thus explored by preparing the chloro-carbonyl complexes, 17–19 (

Table 3. Electronic properties of Ir Cl/CO and dihydride complexes.

Entry	R	R’	Cl/CO Complex	ν(CO)DCM	Dihydride Complex	δH
1	Bn	H	17	1950 cm−1	20	−21.35 ppm
2	Bn	Me	18	1948 cm−1	21	−21.40 ppm
3	Mes	H	19	1942 cm−1	22	−21.58 ppm

). Additionally, from the NHC/PPh₃ complexes, in situ generation of the dihydride complexes 20–22, in CD₃CN [23], provided insight into the σ-donating properties (Table 3). Based on both electronic parameters, the IBn^Me/PPh₃ combination was found to be slightly more electron-rich than IBn/PPh_3,, and more significantly less electron-rich than IMes/PPh₃.

2.4. Evaluation of Iridium Complex Catalytic Activity in HIE

We next chose to focus on the relationship between the IBn/PPh₃ and IBn^Me/PPh₃ complex pair in terms of catalyst activity in HIE reactions. The two precatalysts 6a and 7 were compared via kinetic analysis of the simple reference labelling reaction with acetophenone 23 (

Table 4. Evaluation of complexes 6a and 7 in a benchmark HIE reaction.

Entry	Complex	NHC	kobs
1	6a	IBnMe	0.0224 s−1
2	7	IBn	0.0145 s−1

). Catalyst 6a, derived from Lepidiline A, was clearly faster reacting, displaying a pseudo first order rate constant, k_obs(6a) = 0.0224 s⁻¹, versus k_obs(7) = 0.0145 s⁻¹ for IBn/PPh₃ catalyst 7 [23].

This result was supported by DFT modelling of the turnover-limiting [12] C-H activation step (24→26, Scheme 6). Based on our electronic characterization of these complexes, the greater reactivity of catalyst 6a over 7 can be rationalised on the basis of an increased π-donating ability of the Ir centre, and not on any appreciable steric difference between the two catalysts 6a and 7.

The variation in reactivity between the IBn^Me/PPh₃ ligand sphere in 6a and IMes/PPh₃ in 1a was most interestingly drawn out in studying the counterion dependence on catalyst reactivity (

Table 5. Evaluation of counter-ion effects in IBn^Me/PPh₃ complexes 6.

Entry	Substrate	%D
		6a, X = PF6	6b, X = BF4	6c, X = SbF6	6d, X = OTf	6e, X = BArF
1	23	96	96	96	79	97
2	27	91	87	93	29	97
3	28	98	87	93	43	99
4	29	82	89	90	85	87
5	30	5	4	5	4	16

). Even at a relatively high and unoptimised catalyst loading of 5 mol%, larger and more weakly coordinating substrates 23, 27, and 28 perform relatively poorly when catalyst 6d, bearing the OTf counter-ion, is used in the HIE reaction. Alternatively, the more strongly binding substrate, 29, is relatively unaffected, and the order of reactivity generally reflects that previously observed when evaluating counter-ion effects with the IMes/PPh₃ combination [13]. Presumably, the smaller size of the IBn^Me/PPh₃ system suffers from competitive coordination of the triflate to the metal centre [32]. Most curiously, the catalyst series 6 is ineffective at labelling the weakly coordinating substrate, nitrobenzene 30, whereas the flagship IMes/PPh₃ system 1a labels to ~95% D under similar conditions [11]. This may be due to differences in available intermolecular (for example π-π) interactions of the substrate to the different catalyst systems upon binding. Nonetheless, such effects have the potential to be exploited favourably to induce directing group chemoselectivity in labelling multifunctional molecules.

3. Materials and Methods

See the Supplementary Information for details of all experimental and computational procedures, as well as all crystallographic data.

4. Conclusions

In summary, the first synthesis of the simple imidazolium-based natural product, Lepidiline A, 3·HCl, has been reported, along with a range of other counter-ion derivatives. This has been followed by the development of novel HIE catalysts and the use of a silver-NHC transmetallation approach to HIE catalyst synthesis. Detailed and combined experimental/computational analysis of the resulting iridium complexes 6 and 7 has led to quantitative insight into the parameters responsible for observed differences in the resulting catalytic reactivity of 6a versus the less substituted analogue, 7, and existing flagship HIE catalyst, 1a. Overall, dimethyl substitution on the backbone of the novel NHC was found to impart an electronic influence with a minimal impact on the steric environment. The reactivity differences between the catalysts studied are now being exploited in labelling regioselectivity studies, which will be reported in due course.