Photocontrol of catalysis in CuAAC reactions by air stable Cu(I) complexes of phenylazopyrazole-incorporated ligands

Abstract

Metal complexes containing organic photoswitches are capable of modulating the steric and electronic environment around the metal center through photoisomerization, enabling their use in photoswitchable catalysis. Herein, we design a new class of photoswitchable tripodal tetradentate ligands L1–L3 that can readily form air-stable Cu(I) complexes (C1-PF₆, C1-BF₄, C2, C3). The design strategy integrates flexible spacers and phenylazopyrazole units in the same ligand framework that ensures efficient photoisomerization and sustained stability of the photoswitched state. The complexes were screened for catalyzing the CuAAC reaction between alkynes and azides and C1-PF₆ was identified as a catalyst capable of exerting temporal control over the reaction through photoisomerization. Based on the optimized conditions and the substrate scope, the ZZZ (photoswitched) form of complex C1-PF₆ exhibits substantially improved catalytic performance compared to its EEE (native) form. In this article, we describe detailed experimental and computational investigations aimed at understanding how photoisomerization regulates the catalytic activity of Cu(I) complexes of arylazopyrazole-based tripodal tetradentate ligands.

---

Introduction

Due to its non-invasive nature and high spatiotemporal precision, light is the optimal external stimulus for controlling catalytic reactions. Photocontrol over catalytic processes can be achieved by incorporating organic photochromes within the catalyst frameworks.^1–6 The resultant photoresponsive catalysts reversibly control the steric/electronic environment at the catalytically active center by light, thereby influencing the reactivity and/or selectivity of catalytic reactions.^7–9 Despite significant advancements in organic photoswitchable catalysts,^1–6 designing metal-containing analogues presents unique challenges.^7–18 In such complexes, photoisomerization behavior and stability of the photoswitched states are often suppressed due to metal coordination.⁹ In addition, competitive photo-triggered processes such as decoordination of the metal center or energy transfer processes add further complexity.⁹

The limited number of examples available in the literature for azobenzene-functionalized metal-based photoswitchable catalysts^9–18 highlights the scope for further exploration in this field, especially with regards to ligand frameworks connected to multiple azobenzene units. If the photoisomerization of the attached azobenzene units is synergistic, such moieties can maximize the impact of photoisomerization-induced steric and electronic changes around the metal center, improving the extent of photo-control over the catalytic process. Choosing the correct ligand design and the photochrome is crucial to ensure maximum photoisomerization and better photo-control of the properties of the coordinated metal center. Integrating photoswitches like arylazopyrazole^19,20 in the ligand frameworks ensures coordination to the metal center (through the N atom) and enables efficient bidirectional photoisomerization, which is necessary to design better photoswitchable catalysts.^21,22 All these factors are influential in warranting maximum light-induced modulation of catalytic activity.^7,9

Here, we report the design and synthesis of three new arylazopyrazole incorporated tripodal tetradentate photoactive ligands L1–L3 and their Cu(I) complexes C1-PF₆, C1-BF₄, C2, and C3 (Fig. 1). The complexes were air-stable and retained the photoresponsive behavior of the free ligands. We explored the catalytic properties of these complexes in Copper-catalyzed Azide Alkyne Cycloaddition (CuAAC) reactions. Traditionally, light has been utilized to activate catalysts or generate reactive intermediates in Click reactions.^23,24 Conversely, our hypothesis was centered around the Cu(I) coordinated tris(pyrazolylethyl)amine framework acting as an active catalyst in the reaction. The photoisomerization of the attached azo unit induced a distinct change in the steric or electronic environment, leading to variation in the reaction yield. The light-triggered temporal control over the CuAAC reactions was demonstrated by detailed spectroscopic studies. Density Functional Theory (DFT) calculations were also performed to gain further insights into mechanistic details. The calculations proved that the temporal control originated from the considerable differences in the steric/electronic environment of the EEE and ZZZ forms of the catalyst. Notably, our recent effort in exercising photocontrol over CuAAC reaction by photoisomerization-triggered changes in solubility was one of the first examples of photoswitchable catalysis in CuAAC reaction.³⁴ The current study extends this approach and explores the feasibility of light-induced control over catalytic activity in CuAAC reactions through photo-triggered modulation of steric and/or electronic effects.

Fig. 1 Ligands L1–L3 and complexes C1-PF₆, C1-BF₄, C2 and C3.

Results and discussion

Design and synthesis of ligands and complexes

Ideally, a few key aspects must be considered for designing catalysts that can maximize photo-triggered changes in catalytic activity. (a) The attached organic photochrome should be capable of exhibiting efficient and quantitative bidirectional photoisomerization; (b) the photoswitched states need to be stable with sufficiently longer half-lives; (c) catalyst has to be designed in a way that the photoisomerization can bring maximum changes to the reactivity/selectivity.

All the aforementioned points were considered while designing our target ligands. Tris(pyrazolylethyl)amine core was selected based on its exceptional capability to bind with and stabilize the Cu(I) center,^25–27 aside from serving as the framework for the formation of ligands (L1–L3) comprising three arylazopyrazole units. These photoswitchable units, positioned near the catalytically active Cu(I) center, were introduced to effectively modulate the electronic and steric factors through photoisomerization. Flexible –CH₂ spacers were added within the ligand frameworks to facilitate effective forward and reverse photoisomerization (of both ligands and complexes), in the solution phase or solid state. Phenylazopyrazoles were chosen as photoswitchable units due to their ability to ensure effective photoisomerization and maintain excellent stability in photoswitched states. Altering the substitutions at the 3,5-positions of the pyrazole ring (H, Me, and Et) enabled variations in photoisomerization and coordination behavior in the ligands and complexes. These variations also allowed us to introduce subtle differences in steric crowding near the coordination site, thus allowing for fine-tuning of catalytic behavior in the resultant complexes.

The three new tris(pyrazolylethyl)amine-based photoswitchable ligands L1–L3 were synthesized as depicted in Scheme 1a. Ligand L1–3 were synthesized in good to excellent yields via nucleophilic substitution of 2,2′,2′′-trichlorotriethylamine hydrochloride (1) with the corresponding anions of phenylazopyrazole derivatives (AzPyz, Me-AzPyz, and Et-AzPyz), in situ generated using NaH. All three ligands were characterized with the help of ¹H, ¹³C-NMR, IR, HRMS, and UV–vis spectroscopic techniques (see section S2 in ESI† for details). The reactions of the copper salts such as Cu(CH₃CN)₄PF₆, Cu(CH₃CN)₄BF₄, Cu(ClO₄)₂·6H₂O, and CuI with the ligands L1–L3 were utilized to form the respective air-stable Cu(I) complexes C1-PF₆, C1-BF₄, C2 and C3 as shown in Scheme 1b (see section S2 in ESI† for details). The complexes were characterized by NMR, HRMS, and IR methods. Their solid-state structures were confirmed by SCXRD (see section S3 in ESI† for details), and the bulk purity of the samples was established by elemental analysis.

Scheme 1 Synthesis of (a) tris(pyrazolylethyl)amine-based photoswitchable ligands L1–L3, and (b) metal complexes C1-PF₆, C1-BF₄, C2, and C3.

Surprisingly, the ligand L2 formed an air-stable Cu(I) complex C2, from a Cu(II) precursor Cu(ClO₄)₂·6H₂O without the presence of any external reducing agents. Only a few ligands are known to auto-reduce Cu(II) precursors (in the absence of external reductants) to afford stable Cu(I) complexes.^28–30 This phenomenon of ligand-mediated reduction of the metal center is still elusive and the mechanism has not been fully explored. Based on available reports, the reduction of the Cu(II) center can be induced by oxidation of the solvent, and consequent stabilization of the Cu(I) center by the ligand. Alternatively, there is a possibility of in situ formation of Cu(II) complex, followed by reduction mediated by the water present in the medium (via OH˙ radical; through a homolytic rupture of the Cu²⁺-OH₂ bond). In section S4 of ESI,† we provide a detailed discussion and spectroscopic evidence towards understanding the possible mechanistic details of this process.

Crystal structure of complexes

Complex C1-PF₆ was crystallized from acetonitrile under open-air conditions and block-shaped crystals were obtained. C1-PF₆ was found to crystallize in the trigonal system with the P₃₁c space group. The solid-state structure of the complex revealed its monomeric nature where the L1 ligand coordinated to Cu from four nitrogens (N4, N4′, N4′′, and N5) forming a distorted tetrahedral geometry around the Cu center. The X-ray structure also confirmed the cationic nature of the Cu complex and PF₆⁻ is a non-interacting counter ion. Selected interatomic distances (Å) and bond angles (°) are provided in the caption for Fig. 2. X-ray quality crystals of C1-BF₄ were grown from slow evaporation of a saturated solution of the complex in acetonitrile. Complex C1-BF₄ was found to crystallize in the monoclinic system and P2₁/n space group. The geometry around the Cu center is distorted tetrahedral similar to the C1-PF₆ complex. The asymmetric unit also contains acetonitrile (MeCN) as the solvent of crystallization, which in Fig. 2 has been omitted for clarity. The block-shaped colorless crystals of C2 were grown from the saturated acetonitrile solution by slow evaporation at room temperature. The compound crystallized in the monoclinic system with the P2₁/c space group. The pyrazole ligand L2 coordinated to Cu(I) from four nitrogens (N16, N18, N19, and N23) and the geometry around the Cu center was distorted tetrahedral. The counter anion ClO₄⁻ did not show any interaction with the Cu center and indicated the ionic nature of the Cu complex. Important interatomic distances (Å) and bond angles (°) are given in Fig. 2. Complex C3 also crystallized in a monoclinic system with a C2/c space group. The ethyl-substituted pyrazole ligand forms a distorted tetrahedral geometry around the Cu center. The solid-state structure also confirmed the monomeric nature of the cationic Cu complex and the dimeric nature of (CuI₂)₂²⁻ counter anion.

Fig. 2 Single crystal X-ray structure of complexes C1-PF₆, C1-BF₄, C2, and C3. Thermal ellipsoids are set at 50% probability. All hydrogen atoms have been omitted for clarity. For C3, only one cationic center has been depicted. Selected interatomic distances (Å) and bond angles for C1-PF₆: Cu1–N4 = 1.980(4), Cu1–N5 = 2.175(5), P1–F1A = 1.605(1), P1–F1B = 1.600(1), P1–F2A = 1.572(1), P1–F2B = 1.612(1); N4–Cu1-N4 = 117.17(6), N4–Cu1–N5 = 99.79(1); for C1-BF₄: Cu1B–N5 = 2.169(3), Cu1B–N6 = 1.975(3), Cu1B–N10 = 2.165(7), Cu1B–N3 = 1.842(6), F3–B1 = 1.403(4); N6–Cu1B–N5 = 99.41(12), N6–Cu1B–N10 = 109.0(3), N10–Cu1B–N5 = 93.8(2), N3–Cu1B–N5 = 105.0(2), N3–Cu1B–N6 = 132.4(4); for C2: Cu2–N16 = 2.001(6), Cu2–N18 = 2.183(5), Cu2–N19 = 1.996(6), Cu2–N23 = 2.018(5); N16–Cu2–N18 = 98.2(2), N16–Cu2–N23 = 116.0(2), N19–Cu2–N16 = 123.6(2), N19–Cu2–N18 = 99.0(2), N19–Cu2–N23 = 113.6(2); for C3: Cu1–N3 = 1.979(6), Cu1–N5 = 2.219(6), Cu1–N7 = 2.018(6), Cu1–N11 = 2.018(6); N3–Cu1–N5 = 99.3(2), N3–Cu1–N7 = 122.0(2), N3–Cu1–N11 = 122.6(2), N7–Cu1–N5 = 95.9(2), N7–Cu1–N11 = 110.1(2), N11–Cu1–N5 = 97.4(2).

Photoswitching and thermal stability of photoswitched states

The photoisomerization properties and the thermal stability of the photoswitched states of the ligands and the complexes were evaluated using UV-Vis and ¹H NMR spectroscopy (Table 1 and Fig. 3, see sections S5 and S6 in the ESI† for details). The UV spectra of solutions of μM concentrations of all ligands and metal complexes were recorded before and after irradiation with different wavelengths of light. The thermodynamically stable native EEE configuration in both ligands and complexes exhibited prevalent π–π* and weaker n–π* absorption bands. When irradiated at 340/365 nm, the π–π* bands exhibited a blue-shift, and the n–π* bands intensified, indicating the transition to the photoswitched state. Subsequent exposure of the photoswitched isomer to different wavelengths of light resulted in the partial or complete regeneration of the native EEE isomer (Table 1). The electronic spectral data for the native and photoswitched states of the ligands and the complexes are included in Table 1. As envisioned, the flexible spacers present in ligands and complexes ensured efficient photoswitching behavior in all three ligands (L1–L3) comparable with that of free photoswitches in the solution state. The ligands, L2 and L3, and the complex C3 displayed excellent photochromic properties accompanying color changes (yellow to orange) in the solution phase, while such changes were not prominent for the rest. We also explored the solid-state photoswitching of ligands L1 and L3 (L2 is semi-solid at room temperature) and all four complexes in the KBr medium using diffused reflectance spectroscopy. (Table S5-2 in the ESI†) The ligand L3 showed better photoisomerization properties in the solid-state compared to L1. We assume that the ethyl groups present in the 3,5-position of the arylazopyrazole unit increased the free volume and also, prevented π−π stacking, facilitating efficient switching of L3 in the bulk state.³¹ Compared to the free ligands, photoswitching of Cu-complexes in the solid-state was restricted (Table S5-3 in the ESI†). Specifically, the complexes C1-PF₆ and C1-BF₄ did not undergo photoisomerization in the solid-state, consistent with the trend (low isomerization conversion; 16%) observed for ligand L1. Notably, none of the complexes showed emissive properties. Metal-to-ligand charge transfer (MLCT) bands were not observed for any of the complexes. To ascertain the formation of individual photoisomers (EEE, EEZ, EZZ, and ZZZ) and quantify their exact composition in photostationary states (PSS) during forward and reverse photoisomerization steps, we investigated the photoswitching behavior of the ligands and the complexes with the help of ¹H NMR spectroscopy (see section S6 in the ESI†). In this regard, solutions (of mM concentration; in [D₆]DMSO) of all ligands and Cu(I) complexes were made to undergo forward and reverse isomerization steps by appropriate wavelengths of light. Among the ligands, only L1 showed the prevalent presence of intermediate isomers in the PSS after forward and reverse isomerization. However, for all Cu(I) complexes, the PSS composition after forward and reverse photoisomerization steps displayed a predominant presence of intermediate (EEZ and EZZ) isomers. The overlapping nature of the signals of the intermediate (EZZ and EEZ) photoisomers in C1-PF₆ complicated the quantification of individual photoisomers (see Table 1 and section S6 in the ESI†). In addition, the thermal reverse isomerization solutions (of μM concentration in DMSO) of ligands and complexes were also evaluated with the help of UV-Vis spectroscopy (see Table 1 and section S5 in the ESI†). The thermal reverse isomerization rates of the ligands and complexes were measured at elevated temperatures. In each case, a first-order rate constant was estimated for the formation of the EEE isomer. Using the Arrhenius plot, we calculated the rates of thermal reverse isomerization for all the ligands and complexes at room temperature by extrapolation (Table 1 and see section S5 in the ESI†). Our investigations revealed that the ligand L1 (rate extrapolated to 25 °C ≈ 1.1 × 10^–6 s⁻¹) displayed better stability for the photoswitched state compared to the other two ligands L2 and L3 (rate extrapolated to 25 °C ≈ 3.9 × 10^–4 s⁻¹ and 1.4 × 10^–4 s⁻¹ at 25 °C for L2 and L3, respectively). Interestingly, the thermal stability of the photoswitched states of the Cu(I) complexes was comparable to that of their respective free ligands. Among the four complexes, C1-PF₆, and C1-BF₄ showed the highest thermal stability. Since the photoisomerization of the three photoswitchable units was independent, this trend can be attributed to the stability of photoswitched states of individual photochromes. Usually, the substitutions at the 3,5-position of the pyrazole ring critically influence the stability of the photoswitched states of arylazopyrazole photoswitches. This is likely due to the stabilizing C–H⋯π interaction leading to a T-shape geometry of the Z-isomer.²⁰ In this case, such enhanced stability is best witnessed in the case of phenylazopyrazole-incorporated ligand L1 and its Cu(I) complexes C1-PF₆ and C1-BF₄.

Fig. 3 Solution phase photoisomerization of C1-PF₆ and demonstration of photo-triggered temporal modulation of catalytic activity of C1-PF₆. (a) Forward and reverse photoisomerization (DMSO, 20 μM) of C1-PF₆, blue trace: before irradiation, maroon trace: after irradiation at 365 nm, green trace: after irradiation at 280 nm [inset: photoswitching stability of C1-PF₆ in DMSO over five consecutive cycles {Absorbance at λ_max of π–π* (328 nm); followed by reversibly switching in the forward (at λ = 365 nm) and the reverse (at λ = 280 nm) isomerization steps}]; (b) Possible photoisomers of the Cu(I) complex C1-PF₆. (c) The catalytic activity of EEE-C1-PF₆, EEE-C1-PF₆@365 nm, (and EEE-C1-PF₆@405 nm) in CuAAC reaction (in DMSO) between PhAc and BzAz at (d) 60 °C, and (e) 40 °C.

Table 1 Electronic UV-Vis spectral data depicting the absorption features of native, photoswitched states, and the isomerization conversions in the forward and reverse photoisomerization steps from ¹H NMR spectroscopic studies

S. no.	Compound	Electronic spectral data^a				Irradiation wavelength^b	PSS composition^d			Thermal reverse isomerization^e k (s^−1)
		Before switching		After photoswitching
		λ max /π−π* (ε^c)	λ max /n–π*	λ max /π−π*	λ max /n–π*		EEE(%)	EEZ +EZZ (%)	ZZZ (%)
a Estimated from UV-Vis absorption spectroscopy for solutions in the concentration range 10–20 μM in DMSO. b In nm. c In Lmol^−1cm^−1. d PSS% of photoisomers estimated from ^1H NMR spectroscopy ([D6]DMSO, concentration range = 5–8 mM). Normal font has been utilized for the forward isomerization, while bold letters have been used for reverse isomerization. e Estimated from UV-Vis absorption spectroscopy (DMSO, concentration range = 10–40 μM) (for details, see sections S5 and S6 in the ESI†).
1	L1	330 (57 761 ± 1358)	416	278	404	340		7 (EEZ) + 13 (EZZ)	80	1.1 × 10^−6
						280	14	38 (EEZ) + 36 (EZZ)	12
2	L2	340 (61 146 ± 2876)	X	307	441	365	—	—	>98	3.9 × 10^−4
						490	>98	—	—
3	L3	340 (41 796 ± 1501)	X	294	445	365	—	—	>98	1.4 × 10^−4
						490	80	-	20
4	C1-PF6	328 (45 778 ± 2613)	410	279	396	340	—	24 (EZZ)	76	1.0 × 10^−6
						405	—	87 (EZZ + EEZ)	13
5	C1-BF4	329 (51 020 ± 2108)	416	278	408	340	—	—	>98	2.8 × 10^−6
						280	9	36 (EEZ) + 38 (EZZ)	17
6	C2	332 (49 590 ± 3040)	x	289	440	365	—	7 (EEZ) + 39 (EZZ)	54	1.3 × 10^−5
						490	75	19 (EEZ) + 6 (EZZ)	-
7	C3	330 (36 485 ± 1472)	x	279	445	365	—	7 (EEZ) + 32 (EZZ)	61	3.6 × 10^−5
						505	42	40 (EEZ) + 15 (EZZ)	3

---

Photocontrol of catalytic activity

Symmetrical tertiary amine ligands, along with their complexes, are known for catalyzing several organic transformations,³² including CuAAC reactions.³³ In this regard, all four Cu(I) complexes (C1-PF₆, C1-BF₄, C2, and C3) were investigated for their catalytic efficiency in CuAAC reactions. The switchable –N=N–Ph moiety in these complexes were installed as photo-handles to tune the catalytic activity by light. It was hypothesized that the presence of three –N=N–Ph units in the vicinity of the Cu(I) center would hinder the catalytic activity of the catalyst in its native EEE state (Fig. 3). The photoisomerization of the catalyst to the ZZZ-isomeric state or even to the intermediate EZZ and EEZ isomeric states was expected to alter the steric crowding and/or electronic influence, thereby modulating the catalytic activity by light.

With an aim to validate this hypothesis, we first carried out CuAAC reactions between phenyl acetylene (PhAc) and benzyl azide (BzAz) using 1 mol% of the native (EEE) isomers of the catalysts (C1-PF₆, C1-BF₄, C2, and C3) in DMSO at 60 °C (For details about the reaction conditions, see Section S7A–B in the ESI†). Both C1-PF₆ and C1-BF₄ yielded the same amount (81–82%) indicating no influence of counter ion (entries 1 and 2 in Table 2). Only trace amounts of the product 1-benzyl-4-phenyltriazole (Tz) could be isolated in reaction catalyzed by C2, whereas a maximum product was formed in C3 catalyzed reaction (95%) under similar conditions (entries 3 and 4 in Table 2). While all four Cu(I) complexes (C1-PF₆, C1-BF₄, C2, and C3) had three –N=N – units attached in the vicinity of Cu(I) center, the presence of methyl groups in the 3,5-positions of the pyrazole rings created further steric constraints for the reaction catalyzed by EEE-C2 (entry 3 in Table 2). Conversely, the excellent catalytic activity of C3 evidently included a substantial contribution from the counter anion (CuI₂)₂²⁻ (entry 4 in Table 2). To harness maximum impact of light control on catalytic activity, it is crucial to maintain a reaction rate that is neither excessively rapid nor sluggish. The rest of the optimization steps were carried out using the catalyst C1-PF₆ (entries 6–26, Table 2).

Table 2 Optimization table for the CuAAC reaction between phenylacetylene (PhAc) and benzyl azide (BzAz) using Cu(I) complexes C1-PF₆, C1-BF₄, C2, and C3

Entry	Catalyst	Catalyst mol%	Solvent	Temperature (°C)	Time (min)	Isolated yield (%)
N.R. = no reaction.
1	(EEE)-C1-PF6	1	DMSO	60	300	82
2	(EEE)-C1-BF4	1	DMSO	60	300	81
3	(EEE)-C2	1	DMSO	60	300	Trace
4	(EEE)-C3	1	DMSO	60	300	95
5	—	—	DMSO	25	300	N.R
6	(EEE)-C1-PF6	0.5	DMSO	60	300	75
7	(EEE)-C1-PF6	1	DMSO	60	300	82
8	(EEE)-C1-PF6	2	DMSO	60	300	81
9	(EEE)-C1-PF6	3	DMSO	60	300	82
10	(EEE)-C1-PF6	1	ACN	60	300	83
11	(EEE)-C1-PF6	1	DMF	60	300	80
12	(EEE)-C1-PF6	1	MeOH	60	300	72
13	(EEE)-C1-PF6	1	DMSO	50	300	62
14	(EEE)-C1-PF6	1	DMSO	40	300	20
15	(EEE)-C1-PF6	1	DMSO	22	300	Trace
16	(EEE)-C1-PF6	1	DMSO	60	15	7
17	(EEE)-C1-PF6	1	DMSO	60	30	20
18	(EEE)-C1-PF6	1	DMSO	60	60	35
19	(EEE)-C1-PF6	1	DMSO	60	90	50
20	(EEE)-C1-PF6	1	DMSO	60	120	64
21	(EEE)-C1-PF6	1	DMSO	60	180	71
22	(EEE)-C1-PF6	1	DMSO	60	240	77
23	(EEE)-C1-PF6	1	DMSO	60	300	82
24	(EEE)-C1-PF6	1	DMSO	60	360	90
25	(EEE)-C1-PF6	1	DMSO	60	420	92
26	(EEE)-C1-PF6	1	DMSO	60	480	95
27	C1-PF6@365 nm	1	DMSO	60	300	98

---

The catalyst loading of C1-PF₆ was varied from 0.5 to 3 mol% and the reactions were executed at 60 °C in DMSO (entries 6–9, Table 2). Since there was no remarkable difference between 1, 2 and 3 mol%, we restricted the catalyst loading to 1 mol% (entry 7, Table 2). Afterwards, the screening was performed for the effect of solvents using DMSO, DMF, MeOH and MeCN (entries 7, 10–12 in Table 2). Although the yields were similar for MeCN as that of DMSO, we performed the rest of the experiments in DMSO.³⁵ Furthermore, the reactions were also carried out at different temperatures like 22, 40, and 50 °C. As expected, the yields were maximum at 60 °C (entries 7, 13–15 in Table 2). After choosing the solvent, we paid our attention to the reaction time in DMSO at 60 °C. Based on the isolated yields at various times such as 15, 30, 60, 90, 120, 180, 240, 300, 360, and 480 minutes, we observed an increase in yield from 7% to 95% (entries 16–26 in Table 2).

After optimizing the conditions, we followed the reactions with the photoswitched state of the catalyst C1-PF₆ (entry 27 in Table 2, section S7 in the ESI†). Interestingly, the reaction was faster with an isolated yield of 98% product after 300 minutes. To gain further insights on the effect of photoswitching of the catalyst, we followed the reaction progress with the help of NMR spectroscopy (Fig. 3 and Fig. S7.1–S7.15 in the ESI†).³⁶ We observed that the isolated yields in all cases were comparable to the conversions followed by NMR spectroscopy. The time evolution profiles based on the product conversions showed a significant difference in the reaction rates for the EEE-C1-PF₆ (native state) and ZZZ-C1-PF₆ (after irradiation at 365 nm) catalyzed reactions (Fig. 3d and e). We repeated the process at four different temperatures (60, 50, 40 and 22 °C) and observed a similar trend (Fig. S7.1–S7.10 of the ESI†). At 40 °C, we checked the reversibility of the catalytic activity by irradiating the catalyst solution at 405 nm. Although the rate of the reaction was slower than that of the reaction catalyzed by the photoswitched state of the catalyst, it remained faster than the catalyst in its native state (Fig. 3e and Fig. S7.8 in the ESI†). Such trends are correlated with the moderate photoisomerization conversions of C1-PF₆ after irradiation at 405 nm (or the predominant presence of mixture of the photoisomers, 87% EEZ + EZZ and 13% ZZZ) (Fig. S6.2 in the ESI†). It is noteworthy to mention that a reaction carried out with 1 mol% ligand L1 and 1 mol% Cu(CH₃CN)₄PF₆ allowing in situ formation of Cu(I) complex in the reaction medium also demonstrated similar differences in catalytic activity for the native and photoswitched isomer (Fig. S7.11 and S7.12 in the ESI†).

To understand the scope of the catalyst, additional substrates were also screened for CuAAC reactions with native and photoswitched isomers of the catalyst C1-PF₆. We screened a total of eight combinations of azides and alkynes differing in electronic properties under optimized conditions (Table 3 and section S7A–B in the ESI†). Electron-deficient and electron-rich azides were reacted with phenylacetylene, TBA-substituted phenyl acetylene, and/or propargyl alcohol. A similar trend was observed in the isolated yields of products Tz1-4, Tz6-8. The rate of the CuAAC reaction increased when the photoswitched isomer of C1-PF₆ was used as a catalyst (Fig. 4 and Fig. S6.16–S6.26 in the ESI†). The time evolution profiles of the reactions of 4-methylphenyl azide (TA) with PhAc, (Fig. S7.16, S7.17 and Fig. S7.18 in the ESI†) and 4-tert-butylphenylacetylene (TBA) with BzAz (Fig. S7.19, S7.20 and Fig. S7.21 in the ESI†) are depicted in Fig. 4i–ii. Based on the reaction between phenylacetylene and substituted benzyl azides, we observed that the electron-rich azides showed good reactivity, while electron-deficient azides were less reactive or unreactive. However, photoisomerization of the catalyst improved the yields in all cases. The extension of substrate scope strengthens our hypothesis, although the possible role of electronic influence cannot be excluded. These results indeed confirmed that the temporal control of click chemistry can be extended to other substrates.

Table 3 Substrate scope for the CuAAC reactions

S. no.	Alkyne	Azide	Temperature (°C)	Catalyst	Yield^a (%)
a isolated yields (conditions: 0.5 mmol alkyne, 0.4 mmol azide, 1 ml DMSO, 1 mol% catalyst). b Reaction continued for 420 minutes. c Reaction continued for 540 minutes; for C1-PF6@365 nm, the catalyst stock solution was irradiated at 365 nm.
1^b	PhAc	TA	22	C1-PF6	50
				C1-PF6@365 nm	67
2^b	TBA	BzAz	25	C1-PF6	58
				C1-PF6@365 nm	90
3^c	PhAc	ClA	60	C1-PF6	8
				C1-PF6@365 nm	15
4^c	PhAc	OMeA	60	C1-PF6	30
				C1-PF6@365 nm	70
5^c	PhAc	NO2A	60	C1-PF6	NA
				C1-PF6@365 nm	NA
6^c	PhAc	BrA	60	C1-PF6	10
				C1-PF6@365 nm	15
7^c	PrAc	ClA	60	C1-PF6	5
				C1-PF6@365 nm	10
8^c	PrAc	OMeA	60	C1-PF6	15
				C1-PF6@365 nm	30

---

Fig. 4 Catalytic activity of EEE-C1-PF₆, EEE-C1-PF₆@365 nm, in CuAAC reaction (in DMSO, at 25 °C) between (i) TA and PhAc; (c) TBA and BzAz. (0.004 mM Stock solution of the catalyst in DMSO was used for those reactions with and without irradiation).

Although maximum reactions were screened with the native and photoswitched isomer of C1-PF₆, we also screened the effect of photoswitching in CuAAC reactions using the other catalysts C2 and C3 (Fig. S7.13–S7.15 in the ESI†). As indicated before, we obtained contrasting product yields with C2 and C3, and observed only minimal differences in the catalytic activities of those catalysts before and after photoisomerization. The reaction rate was too slow for the reaction catalyzed by the native EEE form of the catalyst C2. Only a marginal improvement in catalytic activity was observed after photoisomerization of EEE-C2 to ZZZ-C2 (Fig. S7.13 and S7.14 in the ESI†). Whereas for C3, the minimum deviation in the reactivity of the catalyst could be attributed to the additional reactivity of the anion CuI₄²⁻ (Fig. S7.15 in the ESI†).

Overall, the experimental studies provided compelling evidence for phototriggered changes in catalytic activity for C1-PF₆. The study was extended by exploring additional substrates and the observed trends were consistent, emphasizing the applicability of the strategy. We have performed a few control experiments and spectroscopic studies to ensure that the observed changes were indeed due to photoirradiation (see section S7A-B of the ESI†). Other Cu(I) complexes (C2 and C3) also showed similar, yet less effective photo-triggered modulation of catalytic activity. Indeed, the novelty of the catalyst lies not in its ability to catalyze the reaction, but rather in its ability to exert photo-induced control over the CuAAC reaction.

Computational studies

The Density Functional Theory (DFT) approach was used to carry out geometry optimization and elucidate reaction mechanisms (for more details, please see section S8 of the ESI†). The calculations were specifically performed for ligand L1 and the Cu(I) complex C1-PF₆ due to their relevance in the previously mentioned photo-triggered modulation of catalytic activity. First, we compared the photoisomerization behavior obtained from spectroscopic studies with calculations (see section S8 of the ESI†). Generally, we observed good agreement between the experimental and computed values for the λ_max of π−π* and n−π* bands, although shifts were observed in some cases (see section S8 of the ESI†). Next, we focused on elucidating the reaction mechanism for the catalytic process and investigating the variations in catalytic activity between the EEE- and ZZZ isomers of C1-PF₆. The mechanism for the Cu(I) catalyzed click reaction may involve a mononuclear³⁸ (Scheme 2) or a dinuclear Cu(I) species^37,38 promoting the reaction. Depending on the ligand architecture and reaction conditions, experimental and theoretical studies support both mechanisms.^37–39

Scheme 2 Possible reaction pathways for the mononuclear mechanism of the native and photoswitched C1-PF₆.

The mechanism involving the mononuclear Cu(I) species is shown in Scheme 2. In this mechanism, the Cu(I) catalyst coordinates with the deprotonated alkyne in the first step, leading to the formation of int1. In the next step, benzyl azide coordinates to Cu(I), leading to the formation of int2. The proximity of the alkyne and the azide is expected to trigger the cyclization step, leading to the formation of triazole ring in int3. In the next step, the catalyst regeneration is expected with the release of the triazole product Tz. To investigate the differences in catalytic activity between EEE-C1-PF₆ and ZZZ-C1-PF₆, all of the intermediates were separately optimized for both isomers. The optimized structures of the catalyst revealed EEE-C1-PF₆ is stable by 122 kJ mol⁻¹ with respect to the ZZZ-C1-PF₆ isomer. Such large destabilization after photoisomerization could indeed be attributed to the adoption of non-conjugated sterically crowded structure for all three –N=N–Ph units in the ZZZ-C1-PF₆. The formation of the int1 is found to be exothermic for both isomers, with the ZZZ analogue being more stable compared to that of EEE by 7.5 kJ mol⁻¹ (Fig. 5; the energy values of the intermediates were normalized with respect to EEE- and ZZZ-C1-PF₆). At the int1 stage, the Cu(I) centre adopts a T-shaped geometry (see Fig. S8.6 and Fig. S8.7†).³⁵

Fig. 5 Computed energy profile for mononuclear pathways (the optimized structures at B3LYP/def2TZVP level of theory are included; energies are expressed in kJ mol⁻¹ and significant internuclear distances in Å are added).

The formation of T-shaped geometry is ascertained to the strong σ and π-overlap of the alkyne ligand, leading to strong destabilisation of the d_x²−y² orbital (see Fig. S8.6, Fig. S8.7 and Fig. S8.9†). When phenyl acetylene approaches the metal center at int1, two of the coordinated nitrogen atoms (tertiary amine N and one of the pyrazole N) of the ligand dissociate, enabling the establishment of a three-coordinate geometry near Cu(I) for both EEE and ZZZ isomers. This was concluded based on the internuclear distance (see Fig. S8.6 and Fig. S8.7 in the ESI†). Previous literature reports⁴⁰ have showed that this type of partial ligand dissociation is facilitated due to the chelating nature of the ligand. The Cu–C(alkyne) distance in the int1 is found to be relatively shorter for the ZZZ analogue compared to that of EEE, suggesting stronger binding of the substrate in the former (1.866 Å vs. 1.870 Å) (see Fig. S8.6 and Fig. S8.7 in the ESI†). The higher stability of the ZZZ isomer compared to EEE in int1 is due to six additional C–H⋯π interactions (see Fig. S8.8†) present in the former. The formation of int2 from int1 is computed to be endothermic for both ZZZ and EEE isomers (9.9 and 10.9 kJ mol⁻¹, respectively) (Fig. 5). As the T-shaped geometry of int1 is found to be strongly stabilized, the anti-bonding d_x²−y² orbitals of the Cu(I) mixes strongly with the π-orbital of the alkyne that disfavours any additional coordination to form a square planar geometry. Therefore, the phenyl azide could only be weakly bound along the axial direction, which can be understood from the Cu⋯N distance of 3.345 and 3.465 Å for ZZZ and EEE isomers, respectively (see Fig. S8.6 and Fig. S8.7 in the ESI†). Further, several C–H⋯π interactions were found to anchor the azide at the vicinity of the Cu(I) centre for both isomers (see Fig. S8.8 in the ESI†). At int2, the ZZZ isomer is found to be more stable (8.5 kJ mol⁻¹) than the EEE isomer. The endothermicity of the int2 is due to additional strain brought by the phenyl azide, and it is positing slightly weakening the Cu–C(alkyne) distance. At this intermediate, the nitrogen atom of the azide, possessing a significant positive charge, and the carbon atom of the alkyne, possessing a significant negative charge, come in close proximity, favouring cyclization in the forthcoming step (C(12)⋯N(107) distances are 3.362 Å vs. 3.595 Å for EEE and ZZZ isomers, respectively) (see Fig. S8.6 and Fig. S8.7 in the ESI†). The formation of int3 is found to be exothermic, and the geometry around Cu centre is found to be tetrahedral for EEE while trigonal planar for ZZZ isomer. A less sterically hindered positioning of the triazole ring formed facilitates coordination of an additional pyrazole ring that was dissociated earlier for EEE isomer. For the ZZZ isomer, however, the steric hinderance is significant, leading to the formation of a trigonal planar geometry. In int3, EEE isomer is stable by about 6.3 kJ mol⁻¹. Further, from int2 to int3 expected to undergo six-membered transition states and given the short Cu–C(C) bond, approaching this acetylenic carbon is expected to have a significant energy penalty, and this is likely to be different for the EEE-C1-PF₆ and ZZZ-C1-PF₆ isomers, rationalising the observed differences. Once again in the mononuclear mechanism, ZZZ isomer of C1-PF₆ is higher in energy compared to the corresponding EEE isomer. However, upon complexation with reactants, the following factors influences the approaches of the reactant molecules towards the catalytic site and binding: (i) strong intra ligand C–H⋯π interactions especially additional C–H⋯π interactions occurring due to changes in the ligand environment upon photoisomerization; (ii) coordination strength of the ligands to the Cu(I) and its influence in facilitating the donor ability (see Fig. S8.8†).

To gain additional insights, the dinuclear mechanism has also been investigated (for more details, please see section S8 of the ESI†). However, due to large energy penalty in the formation of intermediate (endothermic) suggests that the reaction is unlikely to proceed via the dinuclear mechanism. Moreover, our computational analyses unveil that in the dinuclear mechanism, the ZZZ-C1-PF₆ is predicted to be thermodynamically less favorable when compared to the EEE-C1-PF₆ catalyst. This finding contradicts experimental observations, leading to the exclusion of this mechanism both theoretically and in comparison, to experimental data.

The mononuclear mechanism strongly supports steric differences between the EEE-C1-PF₆ and ZZZ-C1-PF₆ being the main driving force in the observed differences in catalytic activity of the two isomers in CuAAC reactions. Despite ZZZ-C1-PF₆ being of higher energy in nature, the intermediates were stabilized by strong additional intra ligand C–H⋯π interactions (Fig S8.8 in the ESI†). This facilitated the higher catalytic activity in ZZZ-C1-PF₆, explaining the observed trend, though further evidences for electronic effects and/or Lewis basicity⁴¹ of the ligand are elusive.

Conclusions

In summary, this study demonstrates how the intricate relationship between ligand design and photoisomerization influences catalytic performance and sheds light on nuanced details of the rational design of future light-responsive catalysts. The catalytic studies, focusing on Cu(I)-catalyzed azide–alkyne cycloaddition (CuAAC) reactions with different substrates, provided compelling evidence for light-induced temporal control over catalytic activity. The DFT calculations provided valuable insights into factors governing photo-triggered modulation of catalytic activity and the stability of distinct isomeric forms. Our comprehensive mechanistic study supports the notion that a mononuclear mechanism is operational here, as opposed to a dinuclear mechanism. Importantly, this computational evidence aligns with and corroborates the general consensus derived from experimental findings. This research significantly contributes to the burgeoning field of photoswitchable catalysis, introducing a promising avenue for the development of light-responsive metal-based catalysts with tunable and precisely controllable catalytic activity.

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

S. V. and G. R. thank DST Indo-Czech Republic Bilateral Scientific Research Cooperation, Ministry of Science and Technology, New Delhi, India (DST/INT/CZ/P-17/2019). S. V. thanks the Science and Engineering Research Board (SERB), New Delhi for the financial support (CRG/2023/003861). G.R. would like to thank SERB for funding (SB/SJF/2019-20/12; CRG/2022/001697). We are thankful to IISER Mohali for financial support, the departmental and central research facilities and other instruments (XRD, NMR, and HRMS including DST-FIST 400 MHz NMR facility, SR/FIST/CS-II/2019/94(C) TPN No. 32545). D. G., A. K. G and S. K. T thank MHRD, and D. C. thanks UGC. A. thanks UGC and PMRF for the research fellowship. We are also thankful to Dr Angshuman Roy Choudhury (IISER Mohali) for helpful discussions.

References

1. V. Blanco, D. A. Leigh and V. Marcos, Artificial switchable catalysts, Chem. Soc. Rev., 2015, 44, 5341–5370.
2. R. Dorel and B. L. Feringa, Photoswitchable catalysis based on the isomerisation of double bonds, Chem. Commun., 2019, 55, 6477–6486.
3. G. Sobczak and V. Sashuk, Photoswitchable catalysis mediated by nanoparticles, ChemCatChem, 2021, 13, 506.
4. R. Liu, X. Zhang, F. Xia and Y. Dai, Azobenzene-based photoswitchable catalysts: State of the art and perspectives, J. Catal., 2022, 409, 33–40.
5. J. Li and S. Hecht, in Molecular Photoswitches: Chemistry, Properties, and Applications, ed. Z. L. Pianowski, Wiley–VCH GmbH, 2022, ch. 20, pp. 455–475.
6. G. C. Thaggard, J. Haimerl, R. A. Fischer, K. C. Park and N. B. Shustova, Traffic Lights for Catalysis: Stimuli–Responsive Molecular and Extended Catalytic Systems, Angew. Chem., Int. Ed., 2023, 62, e202302859.
7. Z. Freixa, Photoswitchable catalysis using organometallic complexes, Catal. Sci. Technol., 2020, 10, 3122.
8. O. Galangau, L. Norel and S. Rigaut, Metal complexes bearing photochromic ligands: photocontrol of functions and processes, Dalton Trans., 2021, 50, 17879–17891.
9. D. Gupta, A. K. Gaur, H. Kumar, S. Singh and S. Venkataramani, Light–Switchable Metal Complexes: Introducing Photoresponsive Behaviour Through Azoheteroarenes, ChemPhotoChem, 2023, 7, e202300068.
10. M. Li, P. Zhang and C. Chen, Light-controlled switchable ring opening polymerization, Macromolecules, 2019, 52, 5646–5651.
11. J. Ludwig, J. Helberg, H. Zipse and R. Herges, Azo-dimethylaminopyridine-functionalized Ni(II)-porphyrin as a photoswitchable nucleophilic catalyst, Beilstein J. Org. Chem., 2020, 16, 2119–2126.
12. W. Fu, Y. Pi, M. Gao, W. Wang, C. Li, R. Tan and D. Yin, Light-controlled cooperative catalysis of asymmetric sulfoxidation based on azobenzene-bridged chiral salen Ti IV catalysts, Chem. Commun., 2020, 56, 5993.
13. S. Park, S. Byun, H. Ryu, H. Hahm, J. Lee and S. Hong, Reversibly photoswitchable catalysts for olefin metathesis reactions, ACS Catal., 2021, 11, 13860–13865.
14. D. Peng and C. Chen, Photoresponsive palladium and nickel catalysts for ethylene polymerization and copolymerization, Angew. Chem., Int. Ed., 2021, 60, 22195–22200.
15. K. Nakamura, M. Kondo, C. G. Krishnan, S. Takizawa and H. Sasai, Azopyridine-based chiral oxazolines with rare-earth metals for photoswitchable catalysis, Chem. Commun., 2021, 57, 7414.
16. S. Kaler, P. McKeown, B. D. Ward and M. D. Jones, Aluminium(III) and zinc(II) complexes of azobenzene-containing ligands for ring-opening polymerisation of ε-caprolactone and rac-lactide, Inorg. Chem. Front., 2021, 8, 711–719.
17. J. Liu, E. O. Bobylev, B. de Bruin and J. N. H. Reek, A photoresponsive gold catalyst based on azobenzene-functionalized NHC ligands, Chem. Commun., 2023, 59, 8830–8833.
18. M. Marcon, S. Crespi, A. Pielmeier and B. König, A dinuclear copper(II) complex with photoswitchable catechol oxidation activity, Chem. Commun., 2023, 59, 948–951.
19. S. Devi, M. Saraswat, S. Grewal and S. Venkataramani, Evaluation of Substituent Effect in Z-Isomer Stability of Arylazo-1H-3,5-dimethylpyrazoles – Interplay of Steric, Electronic Effects and Hydrogen Bonding, J. Org. Chem., 2018, 83, 4307–4322.
20. J. Calbo, C. E. Weston, A. J. P. White, H. S. Rzepa, J. Contreras-García and M. J. Fuchter, Tuning azoheteroarene photoswitch performance through heteroaryl design, J. Am. Chem. Soc., 2017, 139, 1261–1274.
21. D. Gupta, A. K. Gaur, D. Chauhan, S. K. Thakur, V. Jeyapalan, S. Singh, G. Rajaraman and S. Venkataramani, Solid-state photochromic arylazopyrazole-based transition metal complexes, Inorg. Chem. Front., 2022, 9, 2315–2327.
22. D. Gupta, A. K. Gaur, S. K. Thakur, S. Singh and S. Venkataramani, Photoswitchable Copper(I) and Copper(II) Complexes of Phenylazo–3, 5−dimethylpyrazole Incorporated Ligands, ChemPhotoChem, 2023, 7, e202200338.
23. M. A. Tasdelen and Y. Yagci, Light–induced click reactions, Angew. Chem., Int. Ed., 2013, 52, 5930–5938.
24. G. S. Kumar and Q. Lin, Light-triggered click chemistry, Chem. Rev., 2021, 121, 6991–7031.
25. T. N. Sorrell and D. L. Jameson, Synthesis and characterization of sterically hindered CuN4 complexes of tripod ligands, Inorg. Chem., 1982, 21, 1014–1019.
26. T. N. Sorrell and A. S. Borovik, Absorption, emission, and photophysical properties of copper(I) pyrazole complexes and their carbonyl adducts, Inorg. Chem., 1987, 26, 1957–1964.
27. K. Fujisawa, S. Chiba, Y. Miyashita and K.-i. Okamoto, Copper complexes with neutral N4 tripodal ligands: Influence of the number of nitrogen donors on their structures, properties, and reactivity, Eur. J. Inorg. Chem., 2009, 3921–3934.
28. M. M. Turnbull, G. Pon and R. D. Willett, Autoreduction of copper(II) by a non-chelating amine ligand: Synthesis and X-Ray structure of tris (2, 3-dimethylpyrazine) dicopper(I) perchlorate, Polyhedron, 1991, 10, 1835–1838.
29. P. Levín, M. C. Ruiz, A. I. B. Romo, O. R. Nascimento, A. L. Di Virgilio, A. G. Oliver, A. P. Ayala, I. C. N. Diógenes, I. E. León and L. Lemus, Water-mediated reduction of [Cu(dmp)₂(CH₃ CN)]²⁺: Implications of the structure of a classical complex on its activity as an anticancer drug, Inorg. Chem. Front., 2021, 8, 3238.
30. S. Kumari, S. Muthuramalingam, A. K. Dhara, U. P. Singh, R. Mayilmurugan and K. Ghosh, Cu(I) complexes obtained via spontaneous reduction of Cu(II) complexes supported by designed bidentate ligands: Bioinspired Cu(I) based catalysts for aromatic hydroxylation, Dalton Trans., 2020, 49, 13829–13839.
31. A. Gonzalez, E. S. Kengmana, M. V. Fonseca and G. G. D. Han, Solid-state photoswitching molecules: structural design for isomerization in condensed phase, Mater. Today Adv., 2020, 6, 100058.
32. O. I. Afanasyev, E. A. Kuchuk, K. M. Muratov, G. L. Denisov and D. Chusov, Symmetrical tertiary amines: applications and synthetic approaches, Eur. J. Org. Chem., 2021, 543–586.
33. V. O. Rodionov, S. I. Presolski, D. D. Díaz, V. V. Fokin and M. G. Finn, Ligand-accelerated Cu-catalyzed azide− alkyne cycloaddition: A mechanistic report, J. Am. Chem. Soc., 2007, 129, 12705–12712.
34. D. Gupta, A. K. Gaur, R. Kaur, Ashish, N. Kaur and S. Venkataramani, Photoswitchable Azoheteroarene–Based Chelating Ligands: Light Modulation of Properties, Aqueous Solubility and Catalysis, Chem. – Eur. J., 2023, 29, e202301906.
35. DMSO was preferred for screening photo-triggered modulation of catalytic activities since the reaction had to be monitored at variable temperatures.
36. The progress of the CuAAC reaction was analyzed by ¹H NMR spectroscopy. An aliquot of (10 μL) reaction mixture under stirring was taken, mixed with 440 μl [D₆]DMSO and analyzed with the help of ¹H NMR.
37. B. Worrell, J. Malik and V. Fokin, Direct evidence of a dinuclear copper intermediate in Cu(I)-catalyzed azide-alkyne cycloadditions, Science, 2013, 340, 457–460.
38. J. Héron and D. Balcells, Concerted Cycloaddition Mechanism in the CuAAC Reaction Catalyzed by 1, 8-Naphthyridine Dicopper Complexes, ACS Catal., 2022, 12, 4744–4753.
39. L. Jin, D. R. Tolentino, M. Melaimi and G. Bertrand, Isolation of bis (copper) key intermediates in Cu-catalyzed azide-alkyne “click reaction”, Sci. Adv., 2015, 1, e1500304.
40. L. Chu, K. I. Hardcastle and C. E. MacBeth, Transition metal complexes supported by a neutral tetraamine ligand containing N, N-dimethylaniline units, Inorg. Chem., 2010, 49, 7521–7529.
41. L. Zhu, C. J. Brassard, X. Zhang, P. M. Guha and R. J. Clark, On the Mechanism of Copper(I)–Catalyzed Azide–Alkyne Cycloaddition, Chem. Rec., 2016, 16, 1501–1517.

---

Footnotes

† Electronic supplementary information (ESI) available. CCDC 2232645, 2232646, 2232647 and 2232648. For ESI and crystallographic data in CIF or other electronic format see DOI: https://doi.org/10.1039/d4qi00268g

‡ Contributed equally.

---