Mechanochemical Synthesis of Chromium(III) Complexes Containing Bidentate PN and Tridentate P-NH-P and P-NH-P′ Ligands

Abstract

Chromium(III) complexes bearing bidentate {NH₂(CH₂)₂PPh₂: PN, (S,S)-[NH₂(CHPh)₂PPh₂]: P’N} and tridentate [Ph₂P(CH₂)₂N(H)(CH₂)₂PPh₂: P-NH-P, (S,S)-(iPr)₂PCH₂CH₂N(H)CH(Ph)CH(Ph)PPh₂: P-NH-P′] ligands have been synthesized using a mechanochemical approach. The complexes {cis-[Cr(PN)Cl₂]Cl (1), cis-[Cr(P’N)Cl₂]Cl (2), mer-Cr(P-NH-P)Cl₃ (3), and mer-Cr(P-NH-P′)Cl₃ (4)} were obtained in high yield (95–97%) via the grinding of the respective ligands andthe solid Cr(III) ion precursor [CrCl₃(THF)₃] with the aid of a pestle and mortar, followed by recrystallization in acetonitrile. The isolated complexes are high spin. A single-crystal X-ray diffraction study of 2 revealed a cationic chromium complex with two P’N ligands in a cis configuration with P′ trans to P′ with chloride as the counteranion. The X-ray study of 4 shows a neutral Cr(III) complex with the P-NH-P′ ligand in a mer configuration. The difference in molecular structures and bulkiness of the ligands influence the electronic, magnetic, and electrochemical properties of the complexes as exhibited by the bathochromic shifts in the electronic absorption peaks of the complexes and the relative increase in the magnetic moment of 3 (4.19 μ_β) and 4 (4.15 μ_β) above the spin only value (3.88 μ_β) for a d³ electronic configuration. Complexes 1–4 were found to be inactive in the hydrogenation of an aldimine [(E)-1-(4-fluorophenyl)-N-phenylmethanimine] under a variety of activating conditions. The addition of magnesium and trimethylsilyl chloride in THF did cause hydrogenation at room temperature, but this occurred even in the absence of the chromium complex. The hydrogen in the amine product came from the THF solvent in this novel reaction, as determined by deuterium incorporation into the product when deuterated THF was used.

Introduction

The industrial demand for cheaper and less toxic alternative metals in the design of catalysts with higher efficiency and selectivity for essential fine chemicals in the flavor, perfumery, agricultural, and pharmaceutical industries is on the increase.¹⁻⁷ This has led to tremendous interest in the chemistry of first transition series elements which are more abundant and cheaper than the second and third transition series elements commonly given higher priority in catalytic design.^2,3,7−9 Among the first transition series elements, chromium has shown great potential in catalysis, especially when coordinated to aminophosphine (PN) and aminodiphosphine (PNP) ligands for the production of hex-1-ene and octa-1ene via trimerization and tetramerization of ethene.¹⁰⁻¹⁵ The combination of hard (N) and soft (P) donors of these ligands is suitable to form stable complexes with Cr(III). The presence of hydrogen on the amino nitrogen, the varying of the substituents on the phosphine(s), and the introduction of asymmetric carbon provide endless possibilities in the synthesis of such metal complexes.¹⁶⁻¹⁹ Complexes of chromium in several oxidation states have been synthesized including those of Cr(0) with a bidentate PPh₂CH₂CH₂NH₂ ligand,²⁰ Cr(III) with bidentate Ph₂PNR’PPh₂,²¹ 2,6-R¹-4-R²-C₆H₂–N=CH–C₆H₄–2-PPh₂¹³ or Ph₂PNMeCH₂CH₂C₅H₄N²² ligands, and Cr(II) with tridentate [tBu₂PCHC₅H₃NCHPtBu₂]²³ ligands, and their structures and catalytic properties were investigated. However, there is not a report on homochiral tridentate P-NH-P′ Cr complex with asymmetric carbon groups. Similarly, there are few reports of PN Cr complexes.^13,20 It is therefore not surprising that applications of such Cr complexes have not been given much consideration. Unlike Cr, chiral PN and unsymmetrical P-NH-P′ complexes of Ru,²⁴ Fe,²⁵⁻³¹ Mn,^32,33 and Ir³⁴ are known, and they have been explored in diverse applications.

The general routes adopted in the synthesis of these complexes often involve the use of solvents, which are often toxic and expensive. However, environmental safety concerns have given credence to the concept of green synthesis that minimizes or deliberately avoids the use of solvents. Mechanochemical techniques (dry grinding and liquid assisted grinding) are green routes to synthesize complexes without the use of solvent or with a few drops of solvent especially when the byproducts can be easily removed. This usually avoids the multiple purification and drying processes and also eliminates the generation of waste solvents.²⁰ These methods have been previously highlighted.³⁵ No reports of such a synthesis route have been adopted in the synthesis of metal complexes of chiral PN, P-NH-P, and P-NH-P′ ligands.

There are several protocols for the synthesis of chiral PN. However, the protocol developed by Guo et al.(36) provides a less complicated route with higher yield and purity. The unsymmetrical P-NH-P′ ligand can be derived from the chiral PN ligand via a condensation reaction with α-disubstituted phosphine acetaldehyde, as previously described by our group.³⁷ The tridentate P-NH-P′ pincer ligands are known to adopt different coordination modes around metal centers. In an octahedral coordinate geometry, the P-NH-P′ ligand can adopt a mer or fac coordination mode depending on the bulkiness of the substituents on the phosphines. The mer mode is thermodynamically more stable and more prevalent with bulky substituents.^30,32,38,39 Few examples of chromium complexes are known to adopt fac mode on the coordination of Cr(0) to tridentate PPh₂CH₂CH₂N(C₂H₅)CH₂CH₂PPh₂⁴⁰ and Cr(II) to PPh₂CH₂Si(CH₃)₂NCH₂Si(CH₃)₂PPh₂^41,42 ligand. There are several examples of chromium complexes that adopt mer mode with tridentate P-NH-P ethylene spacer groups,^14,43 PNP with phenylene backbone,^44,45 PNP disubstituted pyrrole backbone,⁴⁶ PNP disubstituted pyridine backbone,⁴⁷⁻⁵² PNP methylene-dimethylsilane backbone,^53,54 PNP disubstituted piperidine backbone,⁵⁵ PNP disubstituted triazine backbone,⁵⁶ and PNP substituted carbazole backbone.⁵⁷

In the past, we have studied different Fe and Mn complexes of achiral and chiral PNNP⁵⁸⁻⁶⁵ and unsymmetrical P-NH-P’^{33,38,39,66,67} ligands, which have been demonstrated as efficient catalysts in asymmetric transfer hydrogenation and asymmetric pressure hydrogenation for the synthesis of chiral alcohols with high demand in the perfumery and pharmaceutical industries (Figure 1). However, a dearth of reports on the Cr complexes of these ligands prompted us to investigate the possible interactions and coordination modes that may exist between the Cr(III) ion and these ligands. Moreover, we also explore the mechanochemical technique as a green route for the synthesis of the complexes. Herein, we describe the synthesis, characterization, electrochemical properties, and crystal structures of Cr(III) complexes of chiral PN and unsymmetrical P-NH-P′ ligands.

Figure 1.

Representative of previously reported PNNP, P-NH-P′, and (PN)₂ complexes by our group and the new P-NH-P′ and PN chromium complexes in this work (see refs (17, 33, 38 and 39) for the structures shown).

Results and Discussion

The preparations of the ligands [2-(diphenylphosphino)ethylamine (PN), (1S,2S)-2-(diphenylphosphino)-1,2-diphenylethylamine (P’N), Ph₂P(CH₂)₂N(H)(CH₂)₂PPh₂ (P-NH-P), and (S,S)-(iPr)₂PCH₂CH₂N(H)CH(Ph)CH(Ph)PPh₂ (P-NH-P′)] have been previously reported by our group^38,39 and others,^14,43 and the details are provided in the Supporting Information. These pincer ligands were coordinated to Cr(III) using CrCl₃(THF)₃ as the Cr(III) precursor in an equimolar stoichiometric ratio via a mechanochemical synthesis. The synthesized complexes are listed in Figure 1. All attempts to synthesize CrLCl₃(THF) (L: PN or P’N) by varying the stoichiometric ratios resulted in [CrL₂Cl₂]⁺ (1 and 2). This reflects the strong affinity of the ligands for the Cr(III) ion. The complexes were dissolved in acetonitrile and left to crystallize at room temperature. Single crystals suitable for X-ray crystallographic study were obtained for complexes 2 and 4. Complex 3 has been previously synthesized in THF, and the crystal structure was reported by McGuinness et al.(43) However, the use of a solvent was avoided in the method of synthesis adopted in this study. The similarity of the obtained product to the previously reported complex⁴³ shows that the same mechanism of interaction occurs between the metal salt and the ligands, with or without the use of solvent. This shows the viability of the adopted mechanochemical synthesis.^35,68,69 The report of McGuinness et al.(43) entails the crystal structure, mass spectra, and elemental analyses of complex 3. We provide more detailed characterization and electrochemical properties of complex 3 in this study for comparative analysis with the newly synthesized complexes 1, 2, and 4.

Crystal Structure Analysis

The single-crystal X-ray diffraction study reveals that complex 2 crystallized in the orthorhombic space group P2₁2₁2₁ (Table S1). Two P’N ligands are coordinated to the Cr(III) via the N (amino) and P (phosphino) binding sites. The P binding sites occupy the axial position in a trans-conformation, while the two N sites are in the equatorial region in a cis-conformation (Figure 2). Similarly, two chlorides also coordinate to Cr(III) in a cis-conformation to complete the octahedral geometry. The coordinated P’N ligands form two five-membered rings with the Cr(III) center. The bond lengths around the Cr(III) center follow the covalent radii [P (1.11 Å), Cl (0.99 Å), and N (0.71 Å)] of the coordinated atoms in the order Cr–P > Cr–Cl > Cr–N. The bond angles 174.88° [N(2)–Cr–Cl(1)], 175.57° [N(1)–Cr–Cl(2)], and 159.40° [P(1)–Cr–P(2)] are less than 180 expected for an ideal octahedral complex, which reflect the distortion in complex 2 along its linear planes and also account for the twist of the P’N ligands as they coordinate to the Cr(III) center. Similar angular deviations from 90° are observed in the bond angles 79.52° [(N(1)–Cr–P(1)], 87.76° [N(2)–Cr–P(1)], and 97.34° [Cl(1)–Cr–P(1)] of the coordinated groups between the axial and equatorial planes, which also reflects the observed distortion in the geometry of the complex.

Figure 2.

X-ray single-crystal structure of complex 2. The hydrogen atoms of the phenyl groups have been omitted for the sake of clarity. The atomic symbols and the corresponding numbering of the elements in the structures are indicated for carbon (C), nitrogen (N), phosphorus (P), and chromium (Cr). Selected bond lengths (Å) and angles (deg) of complex 2: Cr(1)–N(1) 2.086(3), Cr(1)–N(2) 2.104(3), Cr(1)–Cl(1) 2.3012(9), Cr(1)–Cl(2) 2.3053(9), Cr(1)–P(1) 2.4815(13), Cr(1)–P(2) 2.5030(13), N(1)–Cr(1)–N(2) 95.02(10), N(1)–Cr(1)–Cl(1) 86.32(8), N(2)–Cr(1)–Cl(1), 174.88(10), N(1)–Cr(1)–Cl(2) 175.57(10), N(2)–Cr(1)–Cl(2) 84.54(8), Cl(1)–Cr(1)–Cl(2) 94.50(3), P(1)–Cr(1)–P(2) 159.40(3), N(1)–Cr(1)–P(1) 79.52(10), N(2)–Cr(1)–P(1) 87.76(10), Cl(1)–Cr(1)–P(1) 97.34(5), Cl(2)–Cr(1)–P(1) 96.05(5), N(1)–Cr(1)–P(2) 85.96(10), N(2)–Cr(1)–P(2) 78.98(10), Cl(1)–Cr(1)–P(2) 96.21(5), and Cl(2)–Cr(1)–P(2) 98.27(5).

Unlike complex 2, complex 4 with a tridentate P-NH-P′ ligand crystallized in the triclinic space group P2₁ (Table S1). The ligand coordinates via the P and N moieties of the phosphine and amino moieties in a meridional mode. The coordination of three Cl^– ions to the Cr(III) center completed the octahedral geometry. The Cl^– ions are dispersed along the Meridional plane transversely positioned to the P-NH-P′ ligand (Figure 3). This arrangement minimizes the geometric constraints and enhances the thermodynamic stability of the complex in comparison to the facial mode.^70,71 The P-NH-P′ ligand forms two fused five-membered bicyclic rings with the Cr(III) center, with the C(2) and C(3) occupying an endo position to the N of the amino moiety.³² The bond lengths Cr–Cl, Cr–N, and Cr–P are longer than those of complex 2 with the P’N ligand. This reflects the steric hindrance imposed by the bulkiness of the P-NH-P′,^33,38 which induces distortion in the octahedral geometry with elongated axial bond lengths 2.5252(9) Å [Cr–P(1)] and 2.4734(9) Å [Cr–P(2)] in comparison with the bond lengths along the equatorial plane. The bond angles also deviate from 90 and 180°, which also indicates distortion in the octahedral geometry of complex 4. This can be attributed to the steric influence of the P-NH-P′ ligand due to its bulkiness. Complex 4 has a strong structural similarity with previously reported complex 3(43) as indicated by the closeness of the bond angles: P(1)–Cr–P(2) (3: 163.10°, 4: 163.14°), N–Cr–Cl(2) (3: 176.74°, 4: 177.00°), N–Cr–P(1) (3: 81.08°, 4: 81.85°), and N–Cr–P(2) (3: 82.07°, 4: 82.18°).

Figure 3.

X-ray single-crystal structure of complex 4. The hydrogen atoms of the phenyl and isopropyl groups have been omitted for the sake of clarity. The atomic symbols and the corresponding numbering of the elements in the structures are indicated for carbon (C), nitrogen (N), phosphorus (P), and chromium (Cr). Selected bond lengths (Å) and angles (deg) of complex 4: Cr(1A)–N(1A) 2.136(2), Cr(1A)–Cl(2A) 2.3032(8), Cr(1A)–Cl(3A) 2.3205(8), Cr(1A)–Cl(1A) 2.3345(8), Cr(1A)–P(2A) 2.4734(9), Cr(1A)–P(1A) 2.5252(9), N(1A)–Cr(1A)–Cl(2A) 177.00(7), N(1A)–Cr(1A)–Cl(3A) 86.05(8), Cl(2A)–Cr(1A)–Cl(3A) 91.47(3), N(1A)–Cr(1A)–Cl(1A) 84.80(8), Cl(2A)–Cr(1A)–Cl(1A) 97.72(3), Cl(3A)–Cr(1A)–Cl(1A) 170.67(3), N(1A)–Cr(1A)–P(2A) 82.18(7), Cl(2A)–Cr(1A)–P(2A) 99.49(3), Cl(3A)–Cr(1A)–P(2A) 89.22(3), Cl(1A)–Cr(1A)–P(2A) 87.70(3), N(1A)–Cr(1A)–P(1A) 81.85(7), Cl(2A)–Cr(1A)–P(1A) 96.70(3), Cl(3A)–Cr(1A)–P(1A) 95.06(3), Cl(1A)–Cr(1A)–P(1A) 85.49(3), and P(2A)–Cr(1A)–P(1A) 163.14(3).

Electronic Absorption Spectra

The spectra of the complexes (1–4) in DMSO display an intense prominent absorption in the UV region (304–354 nm) (Figure 4A). This corresponds to intra-ligand transition which is strongly influenced by the nature of the ligand.⁷² A considerable redshift and broadening were observed in the peak absorption wavelength on going from the cationic complexes 1 and 2 to the neutral complexes 3 and 4. The order of the spin-allowed ligand field transitions characteristic of a Cr(III) complex with d³ electronic configurations is ⁴A_2g → ⁴T_2g (v₁) < ⁴A_2g → ⁴T_1g (F) (v₂) < ⁴A_2g → ⁴T_1g (P) (v₃).^73,74 We use these octahedral field symmetry labels in our work for comparison with assignments for literature complexes, even though our complexes are distorted from this ideal geometry. The low energy transition B in the region of 578–586 nm, the ⁴A_2g → ⁴T_2g (v₁), has the energy of the ligand field splitting 10Dq (Table 1). Complexes 1 and 2 are thought to have similar cis geometries. However, the PN ligands in complex 1 have less hindered, more flexible backbones than the P’N ligands in complex 2, and this causes a slightly larger splitting and stabilization of complex 1 (Dq = 1712 cm^–1) compared to complex 2 (Dq = 1701 cm^–1). Similarly, the more flexible PNP ligand causes the larger splitting in complex 3 (Dq = 1730 cm^–1) compared to complex 4 (Dq = 1707 cm^–1) with the more rigid PNP’ ligand, even though the latter has a potentially more basic PiPr₂ donor group. These Dq values are intermediate between those observed for [CrCl₆]^3– (1360) and [Cr(en)₃]³⁺ (2188)^75,76 as might be expected. The absorption peak B (Figure 4) in the region 439–468 nm can be attributed to ⁴A_2g → ⁴T_1g (F) v₂ transition.^77,78 The spacing between the v₁ and v₂ transitions of the neutral complexes (4485 cm^–1 for 3 and 4303 cm^–1 for 4) is smaller than that for the cationic complexes (5656 cm^–1 for 1 and 5926 cm^–1 for 2). This spacing is even larger for [Cr(OH₂)₆]³⁺ and [Cr(NH₃)₆]³⁺ (>6000 cm^–1) due to an increase d electron repulsion as expressed by the Racah B parameter.⁷⁵ The transition of the highest energy [⁴A_2g → ⁴T_1g (P)] is obscured by the more intense charge transfer transitions and would be expected in this case at approximately 40,000 cm^–1 (12Dq + 15B + 9349) assuming the Racah parameter B is 680 cm^–1 and the interaction energy between the two ⁴T_1g states is 9349 cm^–1 as determined from the position of v₂.⁷⁹ The molar extinction coefficients (ε) observed for v₁ (60–301 M^–1 cm^–1) and v₂ (55–214 M^–1 cm^–1) (Table 1) are within the expected range for spin-allowed transitions.⁷⁴

Figure 4.

Electronic absorption spectra of complexes 1–4 in DMSO 1 mg/mL; [1.6 (1), 1.1 (2), 1.7 (3), and 1.5 (4) mM].

Table 1. Electronic Absorption Peak Wavelength (nm), the Corresponding Wavenumber (cm^–1), Molar Extinction Coefficient (ε; M^–1 cm^–1), and Crystal Field Splitting Energy [Dq (cm^–1)] of 1–4.

	π–π*	ε (π–π*)	4A2g → 4T1g (v2)	ε (v2)	4A2g → 4T2g (v1)	ε (v1)	Dq
1	304 (32,895)	1415	439 (22,779)	60	584 (17,123)	55	1712
2	311 (32,154)	2022	436 (22,936)	90	588 (17,007)	111	1701
3	321 (31,153)	1316	459 (21,786)	72	578 (17,301)	78	1730
4	331 (30,211)	1496	468 (21,368)	301	586 (17,065)	214	1707
	354 (28,249) sh	1362

FTIR Vibrational Spectroscopy

The vibrational bands for v(N–H) stretching in the region 3171–3298 cm^–1 of the spectra of films of the complexes (1–4) indicate that the amino protons were retained after the coordination of the ligands to the Cr(III) center.^14,74 Hence, the amino ligands remained neutral, and the Cr(III) oxidation state was preserved. The vibrational bands in the region 400–500 cm^–1 are characteristics of v(Cr–Cl^–), which accounts for the chlorides in the coordination sphere.^74,77 It is intriguing to note that the influence of the difference in symmetric nature of the ligands becomes obvious as the primary amino protons in complex 2 display three vibrational bands (3201, 3242, and 3298 cm^–1) in contrast to complex 1, where a single band (3173 cm^–1) is indicated for its primary amino protons. This difference is, however, not observed in complexes 3 and 4 with tridentate ligands, which contain secondary amino protons with vibrational bands at 3171 and 3286 cm^–1, respectively.

Magnetic Susceptibility and Electron Paramagnetic Resonance Spectroscopy

The effective magnetic moments (μ_eff) of complexes 1–4 were determined by the Evans’ method.⁸⁰ The μ_eff for the complexes fall within the expected range of spin only magnetic moment (μ_s.o = 3.88 μ_β) for a d³ electronic configuration.^80,81 The X-band electron paramagnetic resonance (EPR) spectra of complexes 1–4 as powders (see the Supporting Information) were obtained at room temperature and display the broad band characteristics of Cr(III) species with a d³ electronic configuration.⁸² The broad peaks hide any hyperfine couplings to the ³¹P and ⁵³Cr nuclei.^21,83

Electrochemical Study

The cyclic voltammogram (CV) of complex 1 (Figure 5) shows a complex series of redox events including Cr^III/II and Cr^II/I reduction processes with cathodic peak potentials (E_pc) at −0.44 and −1.42 V, and anodic peak potentials (E_pa) at −0.32 and −0.77 V. Reduction of substitution inert Cr^III produces labile Cr^II, resulting in irreversible electrochemical behavior. The E_pa at −0.12 V indicates an irreversible oxidation of Cl^– with a surge in current as the product of the oxidation (Cl₂ gas) desorbed from the surface of the electrode.^84,85 The redox activity associated with Cl^– is equally observed by measuring the CV of tetrabutyl ammonium chloride (TBA⁺ Cl^–) in DMF and MeCN, where a similar increase in current is indicated in the CVs at −0.18 and −0.09 V in DMF and MeCN, respectively (Figure S6). The high potential suggests the substitution of dissociated Cl^– with a solvent molecule (DMF) from reduced chromium species (eqs 1, 2).

1

2

Figure 5.

CVs of complexes 1–4 in 0.1 M [ⁿBu₄N][BF₄]/DMF.

Similarly, an irreversible reduction at E_pc of −1.96 V can be attributed to the substitution of the PN ligand with DMF.^86,87 The CV of complex 2 also shows a two-step electron transfer processes with E_pc at −0.41 and −1.33 V and corresponding E_pa at −0.32 and −0.77 V. The rapid increase in current in the region where E > −0.15 V again suggests the contribution of the irreversible Cl^– ion oxidation/substitution with DMF. The CVs of the complexes with tridentate ligands (3 and 4) also reveal a two-step electron transfer characteristic of Cr ion. The influence of the Cl^– ion oxidation significantly increased the current in the region where E > −0.10 V (eq 3), and the involvement/substitution of the PNP and PNP’ ligands can be attributed to the E_pc at −2.10 and −2.11 V in the CV of complex 3 and 4, respectively.

3

The likely substitution of the Cl^– and the pincer ligands with solvent molecules in the electrochemical process is further investigated by replacing DMF with MeCN.⁷³ MeCN significantly shifts the redox potentials in the electrochemical process, especially for the potentials associated with the substitution of the Cl^– and pincer ligands (Figure S7).

Lack of Catalytic Activity of Complexes 1–4 in Hydrogenation Reactions

The catalytic activity of the chromium complexes (1–4) was investigated in the hydrogenation of the aldimine [(E)-1-(4-fluorophenyl)-N-phenylmethanimine] to the amine [N-(4-fluorobenzyl)aniline] using molecular H₂ under 25 bar in THF at 60 °C with a 5 mol % catalytic loading. The reaction was monitored by ¹⁹F NMR spectroscopy. Complexes 1–4 showed no observable activity. Further catalytic investigations were carried out using complexes 1–4 (1 mol %) in the asymmetric transfer hydrogenation of diphenylphosphinoyl imines to their respective amines analogues using KO^tBu (8 mol %) as a base and 2-PrOH as a solvent and as a reducing agent at 30 °C.^62,88,89 After 24 h, no conversion was noted using ³¹P NMR spectroscopy to monitor the reaction. We envisaged the possible activation of the complexes by the generation of their corresponding hydrides using LiHBEt₃ in THF at −30 to 28 °C to induce their catalytic activity in transfer hydrogenation.^38,90 The formation of hydrides was not successful. We also attempted to aid the hydride formation of the complexes by first generating the triflate analogues of the complexes 1–4 but the complexes showed no reaction in the presence of AgOTf in toluene. On the basis of these negative results, we did not proceed further on investigating the complexes as catalysts in transfer hydrogenation.

Following a recent report on the chromium complexes of cyclic (alkyl)(amino) carbene (CAAC)-catalyzed hydrogenation of alkynes for the selective generation of E- and Z- olefins with high yield in the presence of Mg and trimethylsilyl chloride (TMSCl),⁹¹ we further investigated the possibility that our complexes 1–4 are catalysts for the conversion of imine [(E)-1-(4-fluorophenyl)-N-phenylmethanimine] to amine [N-(4-fluorobenzyl)aniline] using H₂. However, our preliminary investigation showed that the hydrogenation to the amine proceeded in the presence of Mg and TMSCl with or without our complexes as catalysts (Table 2, entries 1–5). We wondered whether the hydrogenation proceeded via the in situ generation of a Grignard-like reagent from Mg and TMSCl. However, when only the Grignard reagent (PhCH₂MgBr) (entry 6) or activated Mg (entry 7) was used under H₂ in the absence of a chromium complex, lower conversion to the amine was observed. Furthermore, when we excluded H₂ (entry 8) and conducted the reaction under Ar at room temperature (entry 9), the hydrogenation still proceeded. This implicates the solvent (THF) as being the hydrogen source. We confirmed this hypothesis by conducting two separate experiments. In the first, the dideuterated amine PhNDCDHC₆H₄F was generated by the reaction of the imine with NaBD₄ in CD₃OD in order to obtain its ¹⁹F NMR and mass spectrum. Then, deuterated THF-D₈ (entry 10) was used as the solvent in the Mg/TMSCl reaction under the same conditions as in entry 9 but with a longer duration (48 h). The mass spectrum of the isolated crude amine product was consistent with partial conversion to the deuterated amine PhNDCDHC₆H₄F. However, the conversion was less than that of entry 9, presumably because of the significant kinetic isotope effect of breaking the carbon-deuterium bond of the THF-D₈ [BDE(C–H) 92 kcal/mol⁹². Thus, we suggest that the hydrogenation proceeds by magnesium reduction of the imine or TMS-iminium chloride with hydrogen atom abstraction from the THF. This side reaction was not reported for the Cr(CAAC) complexes that showed catalytic activity and selectivity in the hydrogenation of alkynes in the presence of Mg/TMSCl in THF.⁹¹

Table 2. Catalytic Study of Hydrogenation of an Imine (0.1 M) to an Amine.

entry	[Cr]	Mg (equiv)	TMSCl (equiv)	yield (%)
1	1	2	0.5	98
2	2	2	0.5	100
3	3	2	0.5	96
4	4	2	0.5	97
5		2	0.5	99
6a				32
7b				33
8c		2	0.5	99
9c,d		2	0.5	90
10c,d,e		2	0.5	20

^aGrignard reagent (PhCH₂MgCl).

^bActivated Mg.

^cUnder Ar.

^d22 °C.

^eTHF-D₈ was used as the solvent.

Conclusions

We have demonstrated the preparation of new Cr(III) complexes in high-yield (95–97%) with a solventless mechanochemical approach and studied the electronic, magnetic, and electrochemical properties of these complexes. The structural analysis revealed that the bidentate P’N ligands form a cis complex with trans Cr–P bonds. The P-NH-P′ ligand coordinates in a mer arrangement about the Cr(III) center with three mer Cl^– ions. In the case of complex 1, no suitable crystals were obtained for the X-ray diffraction study. However, we envisage that it will adopt a structural configuration similar to that of complex 2 based on similarity in properties just as complex 4 shows strong structural resemblance to the previously reported complex 3. The structural differences in the ligands induce variations in the electronic, magnetic, and electrochemical properties of the complexes, which influence their stability. The only conditions discovered for the hydrogenation of an aldimine did not involve the chromium complexes but instead involved the novel reduction of the imine by a Mg/TMSCl mixture under Ar at room temperature with hydrogen abstraction from the THF solvent.

Experimental Section

General Experimental Details

All the procedures and modifications reported in this work were conducted under a N₂ of Ar atmosphere using standard Schlenk line and glovebox facilities, unless otherwise stated. The reagents were used as received from Sigma-Aldrich, Alfa Aesar, and ACROS Organics. All solvents were predegassed and dried based on standard procedures before the commencement of each experiment. The ¹H (500 MHz) NMR spectra were recorded on an Agilent DD2 500 MHz. The FTIR spectra were acquired on a Bruker Alpha spectrometer equipped with an ATR platinum-diamond attachment. X-ray crystallographic data for 2 and 4 were collected on a Bruker Kappa APEX-DUO diffractometer equipped with a PHOTON II CMOS detector and were measured using a combination of ϕ scans and ω scans. The data were processed using APEX3 and SAINT. Mass spectra were obtained with a JEOL AccuTOF Plus 4G equipped with a direct analysis in real time ion source. The absorption spectra were measured by using DMSO solutions of the complexes (1–4). For the Evans’ method,⁸⁰ the NMR frequency was 500 MHz, and the reference solvent used for 1 was C₄H₈O/C₄D₈O, while CH₃CN/CD₃CN was used for 2–4. The X-band EPR spectra were obtained by using a Bruker CW X-band ECS-EMXplus EPR spectrometer at 9.5 GHz. The electrochemical study was conducted using a three-electrode setup with a platinum wire as a working electrode, a tungsten auxiliary electrode, and a Ag/AgCl reference electrode. The electrolyte was a DMF solution with 0.1 M [ⁿBu₄N][BF₄]. All potentials were referenced to the ferrocenium/ferrocene (Fc⁺/Fc) couple (0.72 V versus the standard hydrogen electrode, SHE).⁶⁷ All characterizations were conducted using powder products of the complexes, except for the X-ray diffraction studies using crystals obtained from MeCN.

Mechanochemical Synthesis

[Cr(PN)₂Cl₂]Cl (1)

Purple CrCl₃(THF)₃ (0.24 g, 0.65 mmol) and white PN (0.30 g, 1.3 mmol) were ground with a pestle and mortar in a glovebox for 5 min to a dark blue powder and left overnight at room temperature. A solid product was obtained. Yield: 97% (0.39 g, 0.63 mmol). HRMS: calcd for C₂₈H₃₅Cl₃CrN₂P₂, 619.90 ([M]⁺); found, (ESI⁺): 619.06. EA: % Anal. Calcd (found) for C₂₈H₃₅Cl₃CrN₂P₂: 619.90 g/mol: C, 54.25 (54.19); H, 5.69 (5.62); N, 4.52 (4.49). FT-IR (ATR, v/cm^–1): 3173 (m, NH₂), 3053, (m, CH), 2949 (m, CH), 2924 (m, CH), 2870 (m, CH), 459 (s, Cr–Cl), 437 (s, Cr–Cl). μ_eff: 3.31 μ_β.

[Cr(P’N)₂Cl₂]Cl (2)

CrCl₃(THF)₃ (0.24 g, 0.65 mmol) and white P’N (0.50 g, 1.3 mmol) were ground with a pestle and mortar in a glovebox for 5 min to a dark blue powder and left overnight at room temperature. A solid product was obtained, which was dissolved in acetonitrile and left to recrystallize at room temperature for use in the X-ray diffraction study. Yield: 95% (0.57 g, 0.62 mmol). HRMS: calcd for C₅₂H₅₁Cl₃CrN₂P₂, 922.20 ([M]⁺); found, (ESI⁺): 922.01. EA: % Anal. Calcd (found) for C₅₂H₅₁Cl₃CrN₂P₂: 922.20 g/mol: C, 67.57 (67.54); H, 5.56 (5.57); N, 3.03 (3.07). FT-IR (ATR, v/cm^–1): 3298 (w, NH₂), 3242 (w, NH₂), 3201 (w, NH₂), 3062, (w, CH), 2980 (w, CH), 459 (m, Cr–Cl), 432 (m, Cr–Cl). μ_eff: 3.23 μ_β.

[Cr(P-NH-P)Cl₃] (3)

CrCl₃(THF)₃ (0.48 g, 1.3 mmol) and white P-NH-P (0.57 g, 1.3 mmol) were ground with a pestle and mortar in a glovebox for 5 min to a dark blue powder and left overnight at room temperature. A solid product was obtained. Yield: 96% (0.75 g, 1.25 mmol). HRMS: calcd for C₂₈H₂₉Cl₃CrNP₂, 598.02 ([M]⁺); found, (ESI⁺): 598.04. EA: % Anal. Calcd (found) for with Mol wt for C₂₈H₂₉Cl₃CrNP₂: 598.02 g/mol: C, 56.07 (56.12); H, 4.87 (4.9); N, 2.34 (2.36). FT-IR (ATR, v/cm^–1): 3171 (m, NH), 3053 (w, CH), 2950 (w, CH), 2924, (w, CH), 2870 (w, CH), 458 (m, Cr–Cl), 436 (m, Cr–Cl). μ_eff 4.19 μ_β.

[Cr(P-NH-P′)Cl₃] (4)

CrCl₃(THF)₃ (0.48 g, 1.3 mmol) and P-NH-P’ (0.77 g, 1.3 mmol) were mixed with a pestle and mortar in a glovebox for 5 min and left overnight at room temperature. A dark blue solid product was obtained. Crystals suitable for X-ray diffraction were obtained by recrystallization from acetonitrile at room temperature. Yield: 95% (0.85 g, 1.24 mmol). HRMS: calcd for C₃₄H₄₁Cl₃CrNP₂, 647.15 ([M – Cl]⁺); found, (ESI⁺): 647.21. EA: % Anal. Calcd for C₃₄H₄₁Cl₃CrNP₂: 683.97 g/mol: C, 59.41 (59.47); H, 6.26 (6.26); N, 1.96 (1.95). FT-IR (ATR, v/cm^–1): 3286 (m, NH), 3056 (w, CH), 3030 (w, CH), 2918, 497 (m, Cr–Cl), 477 (m, Cr–Cl), 456 (w, Cr–Cl), 433 (w, Cr–Cl). μ_eff: 4.15 μ_β.

General Procedure for Catalytic Study

The catalytic study was conducted based on a modified procedure.⁹¹ Briefly, the imine (0.2 mmol, 40 mg), the synthesized complex (5 mol %), Mg (10 mg), TMSCl (13 μL), and THF (2 mL) were stirred continuously in a vial under an argon atmosphere. The reaction mixture was transferred to a 50 mL stainless steel Parr hydrogenation reactor at constant temperatures (aided with an oil bath) and pressures (23 atm of H_2(g)) with continuous stirring. The crude product was extracted in ethyl acetate (3 × 5 mL) after quenching the reaction with aqueous HCl (1 M, 2 mL). This was followed by drying the product in ethyl acetate over anhydrous Na₂SO₄ before removing the solvent in vacuo. The conversion was determined from the ¹⁹F NMR signal of the crude product (−115.6 ppm, 400 MHz, CDCl₃) relative to that of the starting imine (−108.02 ppm, 400 MHz, CDCl₃).

Supporting Information Available

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acsomega.4c02076.

• Synthesis of the ligands and complexes, FTIR spectra, selected X-ray diffraction information, CVs, synthesis of the imine, spectra corresponding to the entries of Table 2, and synthesis of the dideuterated amine (PDF)

• Crystal structure data for complex 2 (CIF)

• Crystal structure data for complex 4 (CIF)

The authors declare no competing financial interest.