- 1 Introduction and scope
- 2 Notes
- 3 Photophysics of Cr(
iii ) complexes - 4 Homoleptic polypyridyl complexes of Cr(
iii ) - 5 Heteroleptic polypyridyl complexes of Cr(
iii ) - 6 Acetylide complexes of Cr(
iii ) - 7 Macrocyclic bis-cyano complexes of Cr(
iii ) - 8 Photoluminescent Cr(0) complexes
- 9 Cr(
iii ) complexes in photoredox catalysis - 10 Concluding remarks
Recent advances in chromium coordination chemistry: luminescent materials and photocatalysis
-
Published:02 Nov 2020
-
Special Collection: 2020 ebook collection
P. A. Scattergood, in Organometallic Chemistry: Volume 43, ed. N. J. Patmore and P. I. P. Elliott, The Royal Society of Chemistry, 2020, vol. 43, pp. 1-34.
Download citation file:
With the design and development of sustainable photofunctional materials based upon Earth-abundant elements being of growing interest, coordination complexes of Cr(iii), with their fascinating and long appreciated photophysical and photochemical properties, have returned to prominence. This review charts the most recent developments towards new photoluminescent complexes of Cr(iii), discussing not only the synthetic strategies giving access to increasingly more complicated molecular architectures but also highlighting how an improved photophysical understanding of the electronic excited state underpins the design of impressively efficient Cr(iii)-based phosphors. This article concludes with a survey of the first applications of Cr(iii) complexes within visible light-driven catalysis.
1 Introduction and scope
Photoactive transition metal complexes have been of enormous interest over the last half century, finding use in a diverse array of applications ranging from light-emitting devices and new display technologies,1–3 to optical sensing,4 biological imaging,5–7 artificial photosynthesis8,9 and solar catalysis.10,11 Key to the success of these complexes is the facile ability to tune their rich photophysical properties through not only judicious choice of metal centre but also through careful design of the surrounding ligand architecture, giving chemists access to a vast array of photofunctional materials. However, whilst many of these examples show great promise within light-driven systems and for the capture and conversion of solar energy, they are frequently based upon rare and precious transition metal cations such as Ru(ii), Ir(iii), Re(i) and Os(ii). As society strives to become ever more sustainable, it is clear that such complexes are not economically viable. Consequently, attention has begun to shift towards complexes of the first row transition metal elements which are several orders of magnitude more abundant within the Earth's crust. Whilst such a shift may at first glance appear trivial, coordination chemistry of the first row transition metals presents many fresh challenges, particularly to the synthetic chemist owing to underdeveloped synthetic strategies, the regular occurrence of facile ligand exchange reactions and frequently encountered poor photo-stability.
Whilst photoactive complexes based upon abundant metals such as Cu(i)12–14 and Fe(ii)15–18 have been the subject of notable attention, this review focuses on the coordination chemistry of Cr(iii). Although complexes of Cr(iii) have long been of interest as molecules with rich photophysical properties and fascinating redox behaviour, research into these materials has recently undergone somewhat of a renaissance and returned to prominence within the scientific literature. This review charts and summarises these most recent advances, describing developments principally over the last decade towards new photoluminescent chromium-based complexes in addition to their applications within sustainable photo-redox catalysis.
2 Notes
For simplicity, the micro-symmetry of the hexa-coordinate complexes described within this review is approximated to octahedral and is thus described using the Oh point-group classification. It is noted that a lowering of the symmetry can, in some instances, lead to further splitting of excited states. Unless otherwise stated, quoted wavelengths of photoluminescence are those corresponding to emission from an excited state of 2Eg(Oh) parentage.
3 Photophysics of Cr(iii) complexes
In understanding the photophysical properties of Cr(iii) coordination complexes it is valuable to first consider both the nature of states and the processes which may occur within the electronically excited state. As both the photophysics and photochemistry of Cr(iii) complexes are well understood, the following descriptions serve only to summarise the key concepts and fundamentals, with more detailed accounts having been previously documented elsewhere.19–24
Unlike many well-known photo-active transition metal complexes, such as the ubiquitous [Ru(bpy)3]2+, which display photoluminescence as a result of charge transfer transitions (e.g. metal-to-ligand (MLCT) or ligand-to-metal (LMCT)), the photophysics of Cr(iii) complexes is dominated by the population of metal-centred (MC) (or ligand-field) excited states, which when accompanied by a ‘spin–flip’ can give rise to particularly long-lived luminescent excited states. In this review, all considered Cr(iii) complexes are hexa-coordinate with a d3 (t2g3, eg0) ground state electronic configuration. For simplicity, the coordination geometry for many of these examples can be approximated as being octahedral, thus the key electronic excited states can be sufficiently described through a d3 Tanabe-Sugano diagram derived for a complex with Oh symmetry (Fig. 1a). Thus the electronic ground state for a typical Cr(iii) complex is of quartet multiplicity (4A2g), with photoexcitation resulting in a spin-allowed transition to populate a Frank-Condon state corresponding to a vibrationally hot level of the 4T2g excited state. As this electronic transition represents the population of a metal–ligand anti-bonding orbital the excitation is accompanied by significant structural distortion, hence the displacement of the 4T2g excited-state surface relative to the ground state (Fig. 1c). From the simplified Tanabe-Sugano diagram it may be seen that, beyond a critical ligand-field strength, the identity of the lowest lying excited state switches from 4T2g to 2Eg. Thus, in the presence of strong-field donors, the initially populated 4T2g state may undergo inter-system crossing (ISC) to the thermally equilibrated 2Eg and 2T1g states. As these doublet states have essentially ground-state geometry and bonding, their excited state surfaces are ‘nested’ with that of the ground state, leading to relatively inefficient non-radiative deactivation and the occurrence of long-lived and narrow ‘line-like’ photoluminescence. It is pertinent to note that the energy of the 4T2g state can be readily modulated by ligand-field strength, whilst the energy of the phosphorescent doublet states remain largely invariant. This observation is key to rationalising the design of many efficient phosphorescent Cr(iii)-centred complexes, where maximising the ligand-field strength of the coordinated ligands not only ensures that the 4T2g/2Eg crossing point is surpassed, but also that the energy gap between the 2Eg and 4T2g states is sufficient to avoid, or certainly minimise, detrimental back inter-system crossing (BISC). Consequently, efforts to this end are prevalent within the various molecular design strategies featured throughout this review towards achieving efficient luminescent materials based on Cr(iii).
4 Homoleptic polypyridyl complexes of Cr(iii)
The photophysical properties of homoleptic polypyridyl complexes of Cr(iii) have long been of interest, where efforts have principally focused on complexes of α,α′-diimines such as 2,2′-bipyridine (bpy), 1,10-phenanthroline (phen) or substituted derivatives thereof.21–23,25 The parent complex of the family [Cr(bpy)3]3+ (1a) (Fig. 2) is luminescent in aqueous fluid solution (summarised photophysical data for selected complexes are presented in Table 1), displaying two emission bands at 694 nm and 728 nm which are typical for this class of complex.21,25 The latter, narrow band is the more intense and assigned to luminescence from the 2Eg excited state, whereas the former is only discernible as a high-energy shoulder, attributed to the slightly higher lying 2T1g state. As the energy difference between these two doublet states (690 cm−1)25 is within the range of thermal energy, conversion between these two states is rapid, resulting in the establishment of an equilibrium distribution and identical excited state lifetimes of 63 μs in deaerated 1M HCl(aq) solution.21 Exchange of 2,2′-bipyridine ligands with 1,10-phenanthroline (2a) leads to a near identical photoluminescence maximum, although the lifetime of emission increases 4-fold (τem=270 μs), behaviour ascribed to the vibrationally more rigid phen donors diminishing the rate of non-radiative decay.21 Upon chemical substitution of either bpy (1b–1c) or phen (2b–2c) the wavelength of emission is largely unperturbed, consistent with luminescence deriving from the 2Eg state, although notable enhancements in excited state lifetime are observed (Table 1). Further, the homoleptic bis-terdentate complex [Cr(tpy)2]3+ (tpy=2,2′:6′,2″-terpyridine)(3) is only weakly emissive relative to 1a, with a significantly shortened lifetime and red-shifted luminescence maximum (λem=769 nm, τem=140 ns). The origin of the shift in emission and stabilisation of 2Eg excited states is not clear, although has been tentatively rationalised through a reduction in spin-pairing effects within individual metal-based orbitals as a result of greater tg electron delocalisation over the extended aromatic system of the tpy ligands.21,26,27
Complex . | λem (2Eg)/nm . | τem/μs . | Ref. . | Complex . | λem (2Eg)/nm . | τem/μs . | Ref. . |
---|---|---|---|---|---|---|---|
1a | 728a | 63a | 21 | 7 | 775b | 899b | 34 |
1b | 731a | 230a | 21 | 8a | 782b | 770b | 39 |
1c | 742a | 140a | 21 | 8b | 782b | 1100b | 39 |
1d | 733a | 7.7a | 27 | 9 | 747c | 1200c | 46 |
2a | 728a | 270a | 21 | 10 | 689d | 0.004d | 47 |
2b | 734a | 340a | 21 | 11 | 671d | 180d | 47 |
2c | 743a | 370a | 21 | 12 | 676d | 0.0001d | 47 |
3 | 770b | 0.14b | 26 | 13 | 667d | 1.2d | 47 |
6a | 785b | 0.30b | 26 | 14 | 689d,e | <0.01 | 48 |
6b | 788b | 0.40b | 26 | 15 | 669c | 235c | 49 |
6c | 785b | 0.28b | 26 | 16 | Not observed | 50 | |
6d | 796b | 0.60b | 26 | 17 | 740b | 19b | 50 |
6e | Not observed | 26 | 18 | 748f | 4500f | 51 |
Complex . | λem (2Eg)/nm . | τem/μs . | Ref. . | Complex . | λem (2Eg)/nm . | τem/μs . | Ref. . |
---|---|---|---|---|---|---|---|
1a | 728a | 63a | 21 | 7 | 775b | 899b | 34 |
1b | 731a | 230a | 21 | 8a | 782b | 770b | 39 |
1c | 742a | 140a | 21 | 8b | 782b | 1100b | 39 |
1d | 733a | 7.7a | 27 | 9 | 747c | 1200c | 46 |
2a | 728a | 270a | 21 | 10 | 689d | 0.004d | 47 |
2b | 734a | 340a | 21 | 11 | 671d | 180d | 47 |
2c | 743a | 370a | 21 | 12 | 676d | 0.0001d | 47 |
3 | 770b | 0.14b | 26 | 13 | 667d | 1.2d | 47 |
6a | 785b | 0.30b | 26 | 14 | 689d,e | <0.01 | 48 |
6b | 788b | 0.40b | 26 | 15 | 669c | 235c | 49 |
6c | 785b | 0.28b | 26 | 16 | Not observed | 50 | |
6d | 796b | 0.60b | 26 | 17 | 740b | 19b | 50 |
6e | Not observed | 26 | 18 | 748f | 4500f | 51 |
Deaearted 1M HCl(aq).
Deaearted MeCN.
Deaearted H2O.
DMSO/HSO3CF3.
77 K.
Deaerated D2O/DClO4.
Alongside several photophysical investigations, many early reports identified the strongly photooxidising behaviour and rich electrochemistry of simple Cr(iii) polypyridyls, although exact assignment of the nature of redox events and the corresponding electronic structures remained unclear and even contradictory. This was addressed in 2011 by Wieghardt and co-workers who examined the electron transfer series [Cr(tBubpy)3]n+ (n=3, 2, 1, 0) (4a–4d) (tBubpy=4,4′-di-tert-butyl-2,2′-bipyridine) through a comprehensive combined experimental and theoretical investigation.28 Their findings reveal that successive one-electron reduction occurs exclusively at the ligands, forming bpy-centred π-radical anions (bpy−˙). Here, the spins of the one-electron reduced bpy−˙ ligands are antiferromagnetically coupled to those of the Cr(iii)-centre (d3, S=3/2), accounting for the S=3/2, 1, ½ and 0 ground states as determined by magnetic susceptibility measurements for 4a–4d respectively. Importantly, the central metal ion remains as Cr(iii) throughout the series, revealing that the dicationic species [Cr(tBubpy)3]2+ cannot be described as having a Cr(ii) d4 low-spin configuration as has been typically and widely documented.29,30 The same co-workers also returned to examine the closely related series [Cr(tpy)2]n+ (n=3,2,1,0) (5a–5d) again convincingly demonstrating that sequential one-electron reductions are ligand-localised and that the metal ion can be described as Cr(iii) throughout.31 Interestingly, further in-situ reduction permits the species [Cr(tpy)2]− (5e) to be examined, where the two coordinated dianionic tpy ligands exist in two different spin configurations, either the singlet (tpy2−)2− or triplet (tpy˙˙)2− states, highlighting the rich and often under-appreciated redox chemistry of this ubiquitous ligand and the hitherto unknown electronic structures of its complexes with Cr(iii).
One of the challenges hindering the development and use of Cr(iii) polypryidyl complexes within light-driven applications is the very weak electronic absorption typically displayed by this class of complex within the visible region. The series of homoleptic bis-terpyridine-based complexes (6a–6e) (Fig. 2) reported by Lovaasen26 successfully exploit intra-ligand charge transfer (ILCT) transitions to funnel visible light excitation energy to the low-lying and emissive MC states. Functionalising arylterpyridine ligands with electron-donating groups positions the ligand-localised CT absorption within the visible region, lying below those bands associated with π–π* ligand-centred excitations. 6a–6e display enhanced absorption in the visible compared to [Cr(tpy)2]3+ (3) with 6d in particular exhibiting appreciable absorption at 450 nm (ε=11 900 M−1 cm−1) and 500 nm (5090 M−1 cm−1). Excitation into the ILCT band of 6a–6d results in sharp 2Eg-derived emission bands between 785–796 nm (Table 1), slightly red-shifted relative to 3 (λem=770 nm) with notably elongated lifetimes and in the case of 6d a four-fold increase in intensity.26 Despite 6e exhibiting intense low-energy absorption features (λabs=532 nm ε=56 500 M−1 cm−1, λabs=691 nm ε=13 600 M−1 cm−1) direct excitation into these bands does not result in luminescence. A recent report demonstrates that visible light irradiation of 6b populates MC excited states which can singly (in the case of PPh3) or doubly (for triethanolamine) oxidise sacrificial electron donors and thus function as a multi-electron reservoir for photoinduced charge accumulation,32 whilst the group of Nair have taken advantage of the photoluminescence of 6b and 6c in studying interactions with DNA, where quenching of luminescence upon intercalation is indicative of photo-oxidation of nucleobases and the cleavage of DNA.33
Although near-infrared (NIR) luminescence can be achieved with simple homoleptic Cr(iii) polypyridyl complexes, the quantum efficiencies are typically very low. This is rationalised through the small energy difference between the emitting 2Eg/2T1g states and the 4T2g level (Fig. 1a). Even for popular strong-field donors such as bpy and tpy this energy gap remains sufficiently small to facilitate efficient BISC, not only diminishing the efficiency and lifetime of luminescence but also leading to photosubstitution reactions owing to the lability of the 4T2g state. Indeed, the photoaquation of several Cr(iii) complexes, including [Cr(bpy)3]3+ 35 have been reported.19 Recently, Heinze and co-workers reported the synthesis of a Cr(iii) complex of the terdentate donor ddpd (N,N′-dimethyl-N,N′-dipyridine-2-yl-pyridine-2,6-diamine) [Cr(ddpd)2]3+ (7).34 Capable of forming 6-membered metal chelates with bond angles close to 90° and behaving as a moderate σ-donor but poor π-acceptor, the ddpd motif was envisaged as being the ideal strong-field ligand to enlarge the 4T2g/2Eg energy gap and thus realise efficient, photo-stable NIR emitters.34,36,37 Indeed, the so-called ‘molecular ruby’ 7 has a particularly large energy separation between 2Eg and 4T2g states of 7100 cm−1, with excitation into the LMCT absorption band at 435 nm yielding phosphorescence (λem=775 nm) (Fig. 3) whose quantum efficiency (Φem=12.1%) and lifetime (τem=899 μs) far surpass those seen with more traditional Cr(iii) polypyridyls. Further, 7 exhibits exceptional substitutional stability, being inert to ligand loss in basic aqueous solution upon prolonged irradiation over a period of 2.5 months. Through selective deuteration of methyl groups and by performing luminescence measurements in deuterated solvent, Heinze and co-workers have further demonstrated that non-radiative excited state decay pathways involving energy transfer to high frequency ligand and solvent oscillators can be restricted, raising the luminescence lifetime of 7 to 2.3 ms and the quantum yield of phosphorescence to a record level of 30.1%.38 Synthetic modification of the ddpd framework to replace N–Me groups with N–H functionality affords the terdentate ligand tpda and corresponding Cr(iii) complex 8a.39 In acetonitrile solutions, 8a remains emissive (λem=782 nm), although with a lower quantum yield (8.8%) and lifetime (770 μs) relative to 7 owing to multiphonon relaxation via the N–H oscillators. This is reduced upon deuteration, with the N–D functionalised complex 8b exhibiting a 1.4 fold increase in both the efficiency and lifetime of phosphorescence. Interestingly, deprotonation of 8a reduces the intensity of luminescence, which is fully quenched at pH 9.9 but recovers upon addition of acid, demonstrating potential application as a luminescent pH sensor.39 Indeed, the unprecedented photoluminescence of this ‘molecular ruby’ has lead to a plethora of studies over the last 2–3 years aimed at exploring potential practical applications. For example, dual emission from thermally equilibrated 2Eg and 2T1g states has been exploited, with the temperature dependence of the intensities of the two emission bands functioning as a ratiometric luminescent thermometer over the range 210–373 K,40 whilst external pressure applied to either aqueous solutions or solid samples of 7 and 8a results in red-shifted luminescence maxima from approximately 780 nm to 805 nm (28 kBar).41 Biological nanosensors designed to simultaneously measure temperature, oxygen and pH have incorporated 7 as the functional component in the optical sensing of the former two stimuli.42 The role of 7 as a water-soluble photosensitiser in photo dynamic therapy has been evaluated, although the photocytotoxicity is found to be poor despite a moderate quantum yield of 1O2 sensitisation of 61%.43 7 has been additionally trialled by van Slageren and co-workers as a potential molecular quantum bit (MQB) giving a record phase memory time at 7 K for a Cr(iii) complex of 8.4 μs.44 Very recent studies have attempted to resolve chiral 7 into its corresponding entantiomers.45 Despite the simplified pictorial depiction, 7 is not achiral, with the ddpd ligands adopting a strongly twisted conformation about the metal centre (see Fig. 4 for the molecular structure), giving rise to two distinct homo-chiral helicates (M, M and P, P) which exhibit considerable configurational stability to racemisation. Thus, Seitz and co-workers have successfully resolved 7 into its two enantiomers, with NIR circularly polarised luminescence measurements indicating high luminescence dissymmetry factors (glum) of up to 0.093.45 Similarly, the group of Piguet have synthesised 9 based upon the 2,6-di(quinolin-8-yl)pyridine ligand architecture.46 Displaying very long-lived luminescence in deaerated aqueous solution (λem=747 nm, Φem=5.2%, τ=1.2 ms), 9 adopts a non-planar helical conformation which can be chirally resolved, with the resultant enantiomers displaying exceptional luminescence dissymmetry (glum=0.2), higher than that achieved with any d-block metal complex.
The 1990s saw a small number of reports concerning N-donor cage or quasi-cage complexes of Cr(iii), principally designed to investigate the effects of trigonal distortion upon excited state lifetimes. Perkovic and Endicott compared the closely related 10 and 1147 (Fig. 5), finding that the phosphorescence lifetime of the former was 4.4×104 shorter, despite originating from a 2Eg state of near-identical energy. Likewise, complex 12 has a shorter luminescence lifetime by a factor of 104 than 13 which has an analogous coordination environment around the metal centre.52 Similarly, Sargeson noted that the cage complex 14 displays a very short (<10 ns) 2Eg state lifetime, which increases 2×104-fold upon expansion to the 6-membered chelate rings present in 15 (τ=235 μs).48,49 Broadly, it may be concluded from these studies that trigonal twisting plays a major role in excited state deactivation, with non-radiative relaxation being promoted by those ligands whose geometrical constraints permit distortion from an octahedral towards trigonal prismatic geometry. The theme of trigonal capped Cr(iii) complexes was revisited in 2012 by the group of Shores.50,53 16 shows improved absorption in the visible region over its tris-bidentate iminopyridine analogue 17 although is non-emissive, whereas 17 is weakly luminescent in deaerated acetonitrile (λem=740 nm, Φem=6×10−2%, τ=19 μs).50 Here the difference in luminescence behaviour is not attributed to trigonal twisting induced by ligand strain, owing to the flexibility of the ethyl bridges attached to the capping N-atom, but rather the unexpected involvement of the bridgehead N-atom in the frontier orbitals which induces the efficient population of a low-energy non-emissive ligand-based quartet state.
A further 2019 study investigated the Cr(iii) coordination chemistry of the tripodal ligand 1,1,1-tris(pyrid-2-yl)ethane (tpe).51 Centrosymmetric complex [Cr(tpe)2]3+ 18 is strongly emissive at room temperature (λem=748 nm, Φ=3.2%) owing to particularly large ligand-field splitting and a slow rate of non-radiative relaxation of the 2Eg state. The presence of an inversion centre results in luminescence being both spin and Laporte-forbidden, with the broadened profile of the emission band indicating that vibronic coupling is required to facilitate this electronic transition. Consequently, in D2O/DClO4 room temperature solution, 18 displays the longest luminescence lifetime (τ=4500 μs) of any Cr(iii) emitter reported to date.51
5 Heteroleptic polypyridyl complexes of Cr(iii)
The first heteroleptic polypyridyl complex of Cr(iii) was reported in 2001, with Wheeler, Kane-Maguire and co-workers disclosing an eloquent synthetic route involving initial preparation of an intermediate complex of the type cis-[Cr(diimine)2(CF3SO3)2]+ 54 (Scheme 1). The weakly coordinating triflate ligands are readily displaced through reaction with a second diimine of choice when combined in a non-coordinating solvent. The key advantage of this route is the reliance on the use of Cr(iii) over Cr(ii) intermediates, with complexes of the latter being not only highly air sensitive but also prone to facile ligand scrambling. Indeed, synthetic attempts to form heteroleptic polypyridyl complexes via Cr(ii) precursor materials has been reported to lead to intractable mixtures of ligand scrambled homo- and heteroleptic complexes.27 The authors of the initial report were able to demonstrate the use of this route in the preparation of 19–22, all of which are found to be emissive in deaerated aqueous solution in the region of 730 nm (Table 2), characteristic of 2Eg→4A2g phosphorescence and showing only modest variance in emission energy and lifetime to the related homoleptic systems 1a & 2a. The group of Housecroft have utilised the synthetic route shown in Scheme 1 to prepare further heteroleptic complexes of Cr(iii)55 whilst Ronco and co-workers reported a range of bis-1,10-phenanthroline complexes of the form [Cr(phen)2L]+ (23a–23f), where L is a substituted 2,2′-bipyridine or 1,10-phenanthroline (Fig. 6).56 Whilst the wavelength of luminescence displayed by 23a–23f is largely insensitive to the identity of the second diimine ligand (L), the lifetime of phosphorescence varies dramatically, ranging from 0.21 μs for 23a to 317 μs for 23c, largely dictated by non-radiative deactivation mechanisms involving energy transfer to ligand-localised high frequency oscillators. Importantly, the ability to form heteroleptic complexes of Cr(iii) with substituted diimines allows for extensive scope in the tuning of redox potentials, vital for the development of these complexes for photocatalytic applications.
Complex . | λem (2Eg)/nm . | τem/μs . | Ref. . | Complex . | λem (2Eg)/nm . | τem/μs . | Ref. . |
---|---|---|---|---|---|---|---|
19 | 730a | 200a | 54 | 26b | 730d | Not reported | 58 |
20 | 730a | 112a | 54 | 27 | 721a | 7.6a | 60 |
21 | 734a | 280a | 54 | 28a | 733a | 132a | 61 |
22 | 737a | 450a | 54 | 28b | 732a | 154a | 61 |
23a | 730b | 0.21b | 56 | 28c | 733a | 198a | 61 |
23b | 729b | 208b | 56 | 28d | 732a | 170a | 61 |
23c | 732b | 317b | 56 | 29a | 726b | 214b | 62 |
23d | 730b | 259b | 56 | 29b | 726b | 259b | 62 |
23e | 730b | 91b | 56 | 29c | 726b | 177b | 62 |
23f | 728b | 50b | 56 | 29d | 726b | 17b | 62 |
24a | 732c | 87c | 27 | 29e | 743b | 23b | 62 |
24b | 742c | 108c | 27 | 30 | Not reported | 64 | |
24c | 734c | 47c | 27 | 31 | Not reported | 64 | |
24d | 733c | 7.7c | 27 | 32a | Not reported | 65 | |
25a | 728a | 56a | 57 | 32b | Not reported | 65 | |
25b | 728a | 125a | 57 | 33a | 771b | 1002b | 66 |
25c | 728a | 169a | 57 | 33b | 774b | 980b | 66 |
25d | 735a | 180a | 57 | 34 | 779e | 2.3, 40.3e | 67 |
26a | 730d | Not reported | 58 |
Complex . | λem (2Eg)/nm . | τem/μs . | Ref. . | Complex . | λem (2Eg)/nm . | τem/μs . | Ref. . |
---|---|---|---|---|---|---|---|
19 | 730a | 200a | 54 | 26b | 730d | Not reported | 58 |
20 | 730a | 112a | 54 | 27 | 721a | 7.6a | 60 |
21 | 734a | 280a | 54 | 28a | 733a | 132a | 61 |
22 | 737a | 450a | 54 | 28b | 732a | 154a | 61 |
23a | 730b | 0.21b | 56 | 28c | 733a | 198a | 61 |
23b | 729b | 208b | 56 | 28d | 732a | 170a | 61 |
23c | 732b | 317b | 56 | 29a | 726b | 214b | 62 |
23d | 730b | 259b | 56 | 29b | 726b | 259b | 62 |
23e | 730b | 91b | 56 | 29c | 726b | 177b | 62 |
23f | 728b | 50b | 56 | 29d | 726b | 17b | 62 |
24a | 732c | 87c | 27 | 29e | 743b | 23b | 62 |
24b | 742c | 108c | 27 | 30 | Not reported | 64 | |
24c | 734c | 47c | 27 | 31 | Not reported | 64 | |
24d | 733c | 7.7c | 27 | 32a | Not reported | 65 | |
25a | 728a | 56a | 57 | 32b | Not reported | 65 | |
25b | 728a | 125a | 57 | 33a | 771b | 1002b | 66 |
25c | 728a | 169a | 57 | 33b | 774b | 980b | 66 |
25d | 735a | 180a | 57 | 34 | 779e | 2.3, 40.3e | 67 |
26a | 730d | Not reported | 58 |
Deaerated H2O.
Deaerated MeCN.
Deaerated 1M HCl(aq).
Aerated H2O.
Powder at 5 K.
24a–24d were investigated by Shores et al. in 2010.27 Accessed via the synthetic protocol described above, the methyl-ester groups serve to protect carboxylic acid functionality, a well known anchoring group for attachment to metal oxide surfaces, essential for molecules which have the potential to inject positive holes into semiconductor substrates upon irradiation. Aside from electronic transitions within the UV region, complexes 24a–24d show appreciable absorption extending into the visible region, with 24b in particular exhibiting bands at 450 nm (ε=1960 M−1 cm−1) and 480 nm (ε=1490 M−1 cm−1). Excitation of 24a, 24c and 24d at 320 nm results in weak phosphorescence (Φ<0.6%) observed at approximately 733 nm, although interestingly red-shifted to λem=742 nm for the bathophenanthroline-containing complex 24b. This latter observation suggests a subtle lowering in energy of the 2Eg state, tentatively rationalised by the authors as being due to the effects of intraligand electronic delocalisation over the larger π-system reducing the energetic perturbation of individual metal orbitals upon spin-pairing within the doublet configuration. 24a–24d display rich electrochemistry, with the carboxyl ester functionality shifting the first reduction potentials to remarkably anodic potentials. With excited state reduction potentials in the region of +1.2 V vs Fc/Fc+, these systems are identified as particularly potent photooxidants.27
Also in 2010, Kane-Maguire and co-workers reported the synthesis of heteroleptic Cr(iii) complexes 25a–25d containing the well-known DNA intercalator fragment dipyridophenazine (dppz) (Fig. 7).57 All complexes were reported to display typical Cr(iii) phosphorescence in the region 728–735 nm, with lifetimes of 33–54 μs in aerated solution. Photoluminescence is heavily quenched upon the addition of calf thymus B-DNA, attributed to photo-induced electron transfer and the photooxidation of guanine bases. Indeed, excited state reduction potentials of 25a–25d (∼ 1.50 V vs SHE) are not only tuneable as a function of the identity of the ancillary diimine ligands but reveal the Cr(iii) complexes to be thermodynamically competent for guanine oxidation. Kelly, Quinn and co-workers examining 25b and then closely related 26a–26b, found that luminescence was almost completely quenched upon addition of DNA, with chemical substitution of the dppz moiety effecting the equilibrium binding constant but leaving the energy of the 2Eg emitting state unperturbed.58 Later, a combined ultrafast transient absorption (TA) and time-resolved infra-red (TRIR) study on 25b bound to defined sequence DNA found that the excited state of the intercalated complex was very rapidly quenched, with the absence of detectable electron transfer products suggesting that photooxidation of guanine is immediately followed by extremely rapid back electron transfer (<3 ps).59 These studies have been further supplemented by the investigation of the interaction of 27 with both mono- and poly-nucleotides.60 Based on the first report of tris-heteroleptic Cr(iii) complexes (28a–28d),61 the synthesis of which proceeds via the successive precursors [Cr(diimine)Cl3(dmf)] (dmf=N,N-dimethylformamide) and cis-[Cr(diimine)(diimine′)(CF3SO3)2]+, methylimidazole-containing 27 is luminescent at 721 nm in the absence of DNA,60 with photoaquation of the monodentate ligands offering scope for future investigations into the formation of covalent photoadducts with DNA.
Aside from intercalation studies with biological materials, heteroleptic polypyridyl complexes of Cr(iii) have more recently been examined as tunable optical sensitisers. A 2018 report from Doistau et al.62 describes the synthesis and photophysical properties of 29a–29e, highlighting that the lifetime of 2Eg-derived luminescence can be tuned through judicious incorporation of low-energy vibrational modes and additional rotational degrees of freedom within the ligand structure, the latter being evident for 29e where the shorter lifetime is attributed to enhanced crossing from 2Eg to 4A2g states as a result of the anharmonic misalignment of excited state potential energy surfaces. Further, synthetic routes to asymmetric heteroleptic complexes have also allowed photoluminescent Cr(iii) complexes to feature within extended polymetallic architectures, with a 2020 report describing the potential for Cr(iii)-centred fragments to function as very long-lived NIR sensitisers within light-converting polymetallic devices.63
In addition to those heteroleptic complexes of Cr(iii) featuring didentate ligands, the first examples containing two different terdentate ligands were reported by Housecroft and Constable in 2014.64 Relying upon an analogous synthetic methodology to that described in Scheme 1, 30–31 (Fig. 8) are accessed through a [Cr(tpy)(OTf)3] precursor, with the resultant complexes displaying intense electronic absorption bands within the UV-region. Absorption bands are shifted into the visible region for complexes 32a–32b,65 with intra-ligand charge transfer bands appearing at 507 and 523 nm respectively, somewhat stabilised relative to transitions occurring in the free-ligand as a result of ligation to Cr(iii). Surprisingly, whilst 30–31 and [Cr(tpy)2]3+ (3) are thermally stable in aqueous solution at pH 6.36 and below, these complexes are labile in the presence of base. The authors demonstrate that the likely photoproducts are Cr(ligand)(solvent)3 in nature, with at least one of the terdentate ligands being retained, going further in noting that ligand dissociation also occurs in methanolic solutions of 30–31 in the presence of added [Bu4N][F], leading to the formation of [Cr(ligand)(F)3] species.64
Piguet and co-workers have applied the aforementioned synthetic strategy towards heteroleptic bis(terdentate) complexes of Cr(iii) alongside the design features of the ‘molecular ruby’ 7, combining the ddpd ligand with tpy-derivatives in complexes 33a–33b (Fig. 8).66 Noting the acid sensitivity of some desirable ligands, this synthetic work eloquently proceeds via the route illustrated in Scheme 2, where the weaker Cr-Br bonds are more readily cleaved by AgOTf than those in CrCl3 under microwave irradiation, giving access to useful [Cr(L)(OTf)3] precursors whilst avoiding the use of strongly acidic reaction media. Here, the inclusion of a ddpd ligand partially compensates for the unfavourable distortion imparted by the 5-membered chelate rings of tpy, giving complexes which are photoluminescent (λem≈771 nm, Φ=0.06–0.14%) with impressive lifetimes of τ≈1 ms. As supramolecular chemistry is perhaps more developed for terpyridine-derivatives, these results are of significance for the incorporation of light-emitting Cr(iii) fragments within expanded photo-functional molecular architectures. Indeed, heteroleptic terdentate ‘back-to-back’ Cr(iii) dyads such as 34 have been synthesised as simple proof of concept models for the investigation of inter-metallic communication within multi-metallic systems.67
6 Acetylide complexes of Cr(iii)
The first structurally characterised acetylide complex of chromium was reported by Berben and Long in 2002.68 Synthesised through the addition of Me3SiCCLi to [(Me3tacn)Cr(CF3SO3)3] (Me3tacn=N,N′,N″-trimethyl-1,4,7-triazacyclononane) with later deprotection of the silyl group, this preparative route to 35 paved the way towards a multitude of acetylide complexes of Cr(iii). With bis trans-alkynyl complexes being of growing interest towards the development of molecular wires, several structures, such as 36 and 37, emerged as promising linear connector synthons for the construction of extended delocalised rigid-rod systems.69–72 Further, Shores et al. have utilised Cr-acetylide building blocks in the construction of dinitrogen-bridged bi-metallic dimers such as 38.73 Whilst these studies reveal a range of fascinating redox and magnetic properties, the photophysical attributes of these systems are rarely reported. On the other hand, trans bis-acetylide complexes of Cr(iii) featuring tetraaza macrocycles are of significant photophysical interest.74,75
trans-Cr(cyclam)bis(acetylide) (cyclam=1,4,8,11-tetraazacyclotetradecane) complexes may be prepared through the combination of [Cr(cyclam)(OTf)2]+ with lithiated acetylides,76 the synthetic methodology utilised by Wagenknecht and co-workers in the formation of 39a–c (Fig. 9).77 These early examples display ligand-field electronic absorption bands between 330–450 nm and are phosphorescent in aerated acetonitrile solutions (λem=748 nm) with lifetimes of 5–6 μs, notably increasing to around 30 μs in deaerated H2O. Whilst it may be reasonably expected that such complexes would display typical 2Eg-derived emission, the low-energy broad and featureless emission profiles indicate that these systems are in fact 2T1g emitters, an observation that warrants further explanation.
As has been described earlier within this review, Cr(iii) complexes whose micro-symmetry can be approximated as octahedral (Oh) possess two emissive doublet states, 2Eg and 2T1g. However, as quadrate splitting increases and symmetry lowers towards D4h, it can be seen that the 2Eg (D4h) state of 2T1g origin (described hereafter as 2T1g (Oh)) drops to a lower energy than those states derived from a 2Eg (Oh) level (described hereafter as 2Eg(Oh)) (Fig. 10).20,78–80 Thus, the identity of the lowest lying luminescent state is heavily dependent upon the magnitude of quadrate splitting, which in turn is influenced by the difference in π-bonding characteristics between axial and equatorial ligands. This is especially applicable to the bis-alkynyl tetraaza Cr(iii) systems described here, where the N-atoms of the macrocycle act as σ-donors and the apical alkynyls tend, although not always, to behave as both σ- and π-donors. Usefully, emission from 2Eg(Oh)-derived states tends to be structured, falling within the region 650–710 nm, whereas 2T1g(Oh)-derived luminescence is identified by red-shifted broad and featureless emission profiles.78
Studies of Cr(iii)(cyclam)-containing complexes were expanded by the same workers in 2011, revisiting 39a–d whilst also reporting on the new series of trans bis-acetylides 40a–d (Fig. 9).82 Interestingly, in addition to cis/trans isomers being separable on the basis of solubility, the employed one-pot synthetic strategy requires the addition of a 2-fold excess of nBuLi in order to suppress a competing hydroamination reaction between the arylacetylene and cyclam N–H functionality. The authors also note the requirement to employ lithium diisopropylamide as a base in the synthesis of 40c, although this complex could not be obtained with satisfactory purity to permit further photophysical analysis. 40a–b display typical ligand-field absorptions between 320–550 nm, with modest molar extinction coefficients (ε≈170–700 M−1 dm3) suggesting a degree of ‘intensity-stealing’83 from proximal charge transfer bands. Additionally, the extended aromaticity in 40a–b leads to the onset of electronic absorption being red-shifted to approximately 550 nm when compared to 39a. Further, comparison of cis-39a vs trans-39a reveals excitations of very similar energy, although being twice as intense in the former due to the lack of a centre of symmetry. 40a–b are both luminescent in the region of 750 nm, displaying broad, featureless 2T1g (Oh) emission bands (Fig. 11a), with lifetimes similarly being unperturbed by the identity of the arylacetylide moiety, measured as 3 μs in aerated acetonitrile solution. A near-identical emission profile observed for 40e reveals that even perfluorinated arylacetylides retain sufficient π-donor character that the 2T1g(Oh)-derived excited state remains the lowest lying.78 Conversely, 40d containing comparatively poor π-donor cyclohexyl-alkynyl ligands displays notably different luminescence behaviour; with a blue-shifted emission maximum (λem=727 nm) (Fig. 11b) and fine-structuring being indicative of additional contributions from the 2Eg (Oh) excited state, perhaps further evidenced by the observation of two different lifetimes at 77 K.82 Deuteration of the cyclam N–H groups in 39a elongates luminescence lifetime approximately 4-fold, whereas deuteration upon the aryl ring has little effect, demonstrating the role of cyclam-based N–H oscillators in mediating non-radiative deactivation of the excited state within these systems.82 The trifluoropropynyl moiety has been investigated as an electronic surrogate for closely related cyano ligands (vide infra).84 40f displays near-identical ligand-field absorption bands to trans-[Cr(cyclam)(CN)2]+ (48a) at 420 and 340 nm, although the emission for 40f is weak, with clear fine structuring centred around 725 nm suggesting an emitting state of 2Eg (Oh) character. This is in stark contrast to 39a–40b described above, indicating that these electron-deficient alkynyl ligands no longer act as π-donors but rather moderate π-acceptors. Of further note is an additional paper from the group of Wagenknecht85 which finds that changing the reaction solvent from THF to Et2O raises the proportion of cis isomer formed from 14 to 48%. Thus, cis-40f is also found to be a 2Eg(Oh)-emitter, although with loss of centrosymmetry giving rise to a dominant 0–0 transition at λem=738 nm.
The synthetic versatility afforded by alkynyl functionality has been exploited by several workers to further expand the Cr trans bis-alkynyl core, introducing redox active units such as ferrocene for example95 or through the assembly of extended non-linear π-conjugated frameworks such as geminal-diethynylethenes (gem-DEE).91 Comparing 40g and 41a, the greater electron density introduced by the extended π-system is manifest in enhanced electronic absorption between 340–500 nm, whilst the improved π-donor capabilities red-shift the luminescence maximum of 41a (λem=750 nm) over that of 40g (λem=736 nm). Contrastingly, both 40h and 41b are not emissive in solution nor at 77 K in a solvent glass, behaviour ascribed to the occurrence of energy-transfer quenching of the Cr(iii)-centred doublet excited state by the ferrocene moieties.91
The scope of Cr(iii) bis-alkynyl complexes has recently been expanded by the group of Ren, employing the DMC tetraaza ligand (DMC=5,12-dimethyl-1,4,8,11-tetraazacyclotetradecane) in complexes 42a–c and 43 (Fig. 12).86 Interestingly, the preparation of 42a–b heavily favours the trans isomer regardless of whether cis- or trans-[Cr(DMC)Cl2]+ starting materials are employed, whereas the use of butadiynyl ligands (42c and 43) is more stereospecific, requiring employment of starting materials with the desired stereochemistry. In other instances, the conformation of the macrocycle itself ultimately dictates the geometry around the metal centre. 5,5,7,12,12,14-hexamethyl-1,4,8,11-tetraazacyclotetradecane (HMC) exhibits rac and meso diasteroisomers, with addition of lithiated alkynes to cis/rac-[Cr(HMC)Cl2]+ giving cis products (44a–b) whereas the use of trans/meso-precursors leads to exclusive formation of trans species 45a–b.87 Electronic absorbance spectra for these Cr(iii) cyclam derivatives reveal transitions 10–30 times more intense than their dichloride precursors, owing to strong mixing with proximal alkynyl-based charge transfer transitions. Remarkably, the d-d absorption envelope is highly structured, with vibronic progressions for 42c, 43, 44b and 45b falling within the range of C≡C stretching modes, potentially indicating coupling of these oscillators to the Cr centre.86,87 The absorbance profile for 43 is quite similar to that of its counterpart 42c, although marginally shifted to lower energy and being more intense, the latter feature consistent with the lack of an inversion centre. Similarly to 40h described above, 42b is essentially non-emissive, whereas 42a, 42c, 43 and HMC complexes 44a–45b display fairly broad, weak (Φ<0.15%) emission bands between 746 and 777 nm (Table 3), allowing them to be categorised as 2T1g (Oh)-emitters. 43 displays a shorter luminescence lifetime (τ=160 μs) at 77 K than its structural isomer 42c (τ=358 μs), consistent with 45b and 44b where the trans-isomer (τ=455 μs) exhibits an enhanced lifetime over the cis analogue (τ=129 μs). These observations are tentatively ascribed to increased geometric strain within cis systems and consequent operation of non-radiative decay pathways.86,87
Complex . | λem/nma . | τem/μs . | Ref. . | Complex . | λem/nma . | τem/μs . | Ref. . |
---|---|---|---|---|---|---|---|
39a | 748b | 225b | 77 | 42c | 755b | 113b | 86 |
39b | 749b | 259b | 77 | 43 | 777b | 68b | 86 |
39c | 748b | 203b | 77 | 44a | 764b | 212e | 87 |
39d | Not reported | 82 | 44b | 771b | 129e | 87 | |
40a | 749b | 250b | 82 | 45a | 746b | 469e | 87 |
40b | 750b | 238b | 82 | 45b | 746b | 455e | 87 |
40c | Not reported | 82 | 46 | 747b | 218b | 88 | |
40d | 727b | 331b | 82 | 47 | 777b | 97e | 88 |
40e | 747c | Not reported | 78 | 48a | 720f | 335f | 89 |
40f | 725d | 460d | 84 | 48b | 720f | 1500f | 89 |
40g | 736b | 351b | 91 | 49 | 735g | 0.24g | 90 |
40h | Not observed | 91 | 50 | 724h | 23g | 92 | |
41a | 750b | 185b | 91 | 51 | 705h | 24g | 93 |
41b | Not observed | 91 | 52 | 712i | 190i | 94 | |
42a | 750b | 95b | 86 | 53 | 721h | 147g | 93 |
42b | Not observed | 86 |
Complex . | λem/nma . | τem/μs . | Ref. . | Complex . | λem/nma . | τem/μs . | Ref. . |
---|---|---|---|---|---|---|---|
39a | 748b | 225b | 77 | 42c | 755b | 113b | 86 |
39b | 749b | 259b | 77 | 43 | 777b | 68b | 86 |
39c | 748b | 203b | 77 | 44a | 764b | 212e | 87 |
39d | Not reported | 82 | 44b | 771b | 129e | 87 | |
40a | 749b | 250b | 82 | 45a | 746b | 469e | 87 |
40b | 750b | 238b | 82 | 45b | 746b | 455e | 87 |
40c | Not reported | 82 | 46 | 747b | 218b | 88 | |
40d | 727b | 331b | 82 | 47 | 777b | 97e | 88 |
40e | 747c | Not reported | 78 | 48a | 720f | 335f | 89 |
40f | 725d | 460d | 84 | 48b | 720f | 1500f | 89 |
40g | 736b | 351b | 91 | 49 | 735g | 0.24g | 90 |
40h | Not observed | 91 | 50 | 724h | 23g | 92 | |
41a | 750b | 185b | 91 | 51 | 705h | 24g | 93 |
41b | Not observed | 91 | 52 | 712i | 190i | 94 | |
42a | 750b | 95b | 86 | 53 | 721h | 147g | 93 |
42b | Not observed | 86 |
Maximum of lowest energy emission band.
Deaerated MeCN.
MeCN.
Deaerated H2O.
4 : 1 EtOH/MeOH glass at 77 K.
5×10−3 M HNO3(aq).
0.01M HCl(aq).
DMSO/H2O at 77 K.
DMSO/H2O.
Further Cr(iii)-centred complexes of HMC have been prepared which feature 1-ethynylnaphthalene ligands.88 Both 46 and 47 display red-shifted d-d electronic absorption bands relative to 44a and 45a due to the increased aromaticity of the aryl substituent. Luminescence from 46 and 47 is predominantly 2T1g(Oh) in nature (λem=747 and 777 nm respectively), although an additional small high-energy shoulder (λem=727 nm) observed for 46 at 77 K indicates partial dual emission from a 2Eg(Oh) state, similar behaviour to that seen for 42a. Despite employing a well-known fluorophore, naphthalene-based emission is not detected for these complexes, indicating efficient quenching by the metal centre.
7 Macrocyclic bis-cyano complexes of Cr(iii)
The preparation of trans-[Cr(cyclam)(CN)2]+ (48a) was reported in 1983 and unlike several contemporary complexes (e.g. trans-[Cr(NH3)4(CN)2]+ and trans-[Cr(en)2(CN)2]+) was found to be completely inert to photosubstitution.89 48a displays intense phosphorescence at 720 nm with a lifetime of 335 μs, elongated 5-fold upon selective deuteration of the cyclam (48b) (see Table 3 for summarised photophysical data). Further, deprotonation/protonation of cyclam N–H groups allows the phosphorescence to be quenched and then fully restored. 48a and 48b featured in a later 2002 study,79 where flash photolysis was used to probe energy-transfer dynamics between the longer-lived deuterated species and its proteo analogue, essentially representing self-exchange energy transfer.
Bis-cyano Cr(iii) complexes featuring cyclam derivatives have continued to be of further interest, with complexes 49 and 50 containing constrained macrocycles 1,4-C2-cyclam and 1,11-C3-cyclam respectively being reported (Fig. 13).90,92,96 Synthesis of these systems proceeds via the usual [Cr(cyclam′)Cl2]+ precursors, which along with the resultant cyano complexes are produced with exclusively trans geometry. Whilst X-ray crystal structures show the coordination sphere bond lengths to be consistent, the strain imparted by the additional macrocycle linkages leads to distortion of bond angles away from idealised values of 90° and 180° as found in 48a, with the CrN4 plane being additionally distorted in 49.92 Perhaps surprisingly, whilst 48a is photoinert, both 49 and 50 display moderate rates of photoaquation, with loss of axial CN− ligands. Electronic absorption spectra of 48a–50 exhibit ligand-field transitions of very similar energy, although marginally shifted to lower energy for 50 and then 49 owing to the weaker ligand-field caused by the aforementioned structural distortions hindering efficient donor-metal orbital overlap. Remarkably, the molar absorptivity of these transitions clearly increases in the order 48a<50<49, concomitant with the progressive deviation from octahedral symmetry, perfectly illustrating the consequences of relaxation of the Laporte rule. 49 and 50 are luminescent at 735 and 724 nm respectively, although they are much weaker emitters than 48a with correspondingly shorter lifetimes (τ=0.24 μs for 49 and 23 μs for 50). Unlike 48a, these emission profiles are dominated by intense 0–0 transitions, further representing the lowering in symmetry. Curiously, 49 and 50 show no enhancement of luminescence lifetime upon N–H deuteration when in aqueous solution, although deuterated 50 exhibits a 4-fold enhancement in phosphorescence lifetime when in a frozen glass.92
51–53 have also been investigated as luminescent bis-cyano complexes of Cr(iii), featuring very subtle structural modifications of the macrocyclic motif (Fig. 13).93 The progressively increasing ligand-field strength imposed by the expanded macrocycles [16]aneN4, [15]aneN494 and then isocyclam is evident in electronic absorption spectra where the energetic ordering of the d-d transitions follows the trend 51<52<53. These systems are categorised as 2Eg(Oh) emitters, with largely similar peak positions centred around 720 nm (Table 3). The lower degree of centrosymmetry around the coordination sphere in 53 imparted by isocyclam is represented by an emission spectrum at 77 K dominated by the 0–0 transition, whereas the highly symmetrical macrocyle within 51 gives rise to clear vibronic structuring. In room temperature aqueous solution, 53 displays a shorter lifetime (τ=147 μs) than the parent complex 48a (τ=335 μs), although this is elongated compared to 51 (τ=24 μs). Once more, with non-radiative decay of emissive 2Eg(Oh) excited states involving tunnelling to the 4A2g(Oh) ground state, deuteration of the high-frequency N–H oscillators in 53 extends the lifetime in aqueous solution almost 3-fold.93
8 Photoluminescent Cr(0) complexes
Although this review is primarily concerned with photoactive complexes of Cr(iii), it is worth briefly noting a particularly impressive example of a recently reported photoluminescent Cr(0) complex. Despite 77K emission from Cr(0) species being noted over 40 years ago,97 it was only in 2017 that the group of Wenger made use of bulky chelating diisocyanide ligands in obtaining 54 (Fig. 14).98–100 This zero-valent complex is orange/red in colour, exhibiting 1MLCT electronic absorption bands between 400–600 nm and displaying luminescence in room temperature THF solution at 630 nm with a quantum yield of 10−5 and lifetime of 2.2 ns.98 Significantly, the lifetime of the 3MLCT excited state is at least an order of magnitude longer than currently achievable with isoelectronic complexes of the Earth-abundant metal Fe(ii), sufficiently long to allow energy transfer to acceptors and subsequent photon upconversion. Further, 54 is a particularly potent photoreductant, with an excited state oxidation potential of −2.43 V vs Fc/Fc+.98
9 Cr(iii) complexes in photoredox catalysis
Over the last decade photoredox catalysis has emerged as an increasingly popular synthetic tool for the mediation of a variety of useful organic transformations. Whilst commonly employed photocatalysts rely upon the use of precious metals such as iridium and ruthenium, attention has recently shifted towards complexes of Earth-abundant elements.101–104
Although Cr(iii) complexes have long been recognised for their photooxidising capabilities, their first use as photoredox catalysts in organic synthesis was reported in 2015 by Ferreira, Shores and Stevenson.105 Simple tris Cr(iii) polypyridyls such as 1d, 2a & 24a (Ered*=+1.45 to +1.84 V vs SCE) were found to catalyse the [4+2] dimerisation of 1,3-cyclohexadiene, generally giving better yields than the related but less powerful photooxidant [Ru(bpz)3]2+ (bpz=2,2′-bipyrazine) (Ered*=+1.45 V vs SCE) under the same conditions. The bathophenanthroline-containing complex 2c, with its slightly lower oxidising power (+1.40 v vs SCE) was identified as being ideal for facilitating cross-cycloadditions between dienes and electron-rich dieneophiles, selectivity oxidising the latter. Indeed, 2c successfully catalyses the cycloaddition of trans-anethole with isoprene with an excellent yield of 88% under near-UV irradiation in nitromethane. Interestingly the authors of this initial report noted the requirement for the presence of oxygen in achieving catalytically activity. This was later rationalised through a detailed mechanistic study, where the dual role of oxygen as both an electron and energy transfer shuttle was revealed.106 As illustrated in the proposed mechanism (Scheme 3), 3O2, or rather subsequently formed 1O2 and superoxide, play a critical role in either regenerating the Cr(iii) catalyst or reducing the radical cationic cycloaddition product. This is a mechanistically distinct picture from that found in similar reactions catalysed by [Ru(bpz)3]2+ where a radical chain propagation occurs, a consequence here of the exceptionally long-lived 2Eg excited state of the Cr(iii) catalyst (2c, τ=13 μs under reaction conditions) permitting the build-up of an appreciable concentration of 1O2 and thus initialising the oxygen-mediated photocatalytic cycle. Further finer details of this complicated mechanism have been resolved computationally by Dang et al. who suggest the role of a quintet intermediate complex [CrL3-O2]2+ in the final electron transfer step to the radical cationic cycloadduct.107
Under visible-light irradiation, 2c also catalyses the [4+2] cycloaddition between dienes and electron-deficient dieneophiles (Scheme 4).108 This is somewhat surprising given that enone 55 has an oxidation potential at least 0.6 V higher than the excited state reduction potential of 2c, yet the cycloaddition product is observed to form with a yield of 85%. Mechanistically, this clearly requires a different explanation to that described above. With direct oxidation of the enone, and indeed the diene being thermodynamically unfavourable, additional pathways were suggested (Scheme 4). Route A relies upon the photoexcitation of the dieneophile and subsequent [2+2] cycloaddition with the diene to form a vinylcyclobutane intermediate (56). In the presence of 2c under irradiation, 56 may undergo single-electron oxidation rearrangement to give the product (57). Route B involves in situ dimerisation of starting enone 55 to give a cyclobutane dimer 58. Photooxidation may then induce cycloreversion and trapping of the resultant radical cation (59) by the diene giving product 57, although control experiments suggest this route plays a minor role in the formation of 57. Finally, route C recognises that the photoexcited enone 60 is more easily oxidised than when in its ground state, with the resultant radical cation once more reacting with an equivalent of diene to give the cycloaddition product. Molecular oxygen again appears to be beneficial to photocatalytic performance, although not essential in this instance. The substrate scope of this reaction has been expanded to include a wide range of both substituted dienes and electron deficient chalcone-derivatives, all undergoing cycloaddition with good to excellent yields. Notably, these reactions proceed with reverse regioselectivity to that seen in conventional Diels–Alder cyloadditions.108
Cr complexes 1d, 2a, 2c & 24a have been found to catalyse the cyclopropantion of diazo compounds with electron rich alkenes under visible light irradiation (Scheme 5a), with 2c giving the highest yield of cyclopropane 63 (60%) under comparable conditions.109 Expanding the substrate scope, 2c is particularly suited to catalysing the [2+1] cycloaddition between 62 and a range of substituted stilbenes, in addition to successful examples involving the reaction of 61 with α-alkyl diazo esters. In these latter instances, it is notable that diazo dimerisation does not occur, nor are β-hydride elimination products detected.109 Further, 2c has been additionally exploited in the photocatalysed [3+2] cycloaddition of alkenes and vinyl diazo substrates to form functionalised cyclopentenes (Scheme 5b).110 The reaction mechanism appears to involve initial oxidation of trans-anethole by the photo-excited Cr catalyst, generating an electron deficient radical cation which then undergoes direct nucleophilic attack by the vinyl diazo reagent. High diasteroselectivities are reported for these transformations, with the cycloadducts being readily diversified though tolerance of a wide variety of substituted coupling partners under the employed reaction conditions.
Arai and Ohkuma have utilised [Cr(bpy)3]2+ (1a) to catalyse the aza-Diels–Alder cycloaddition reaction of N-arylimines with various alkenes under blue-light irradiation (Scheme 5c).111 N-arylimines bearing both electron-withdrawing and –donating groups upon the aryl moiety react smoothly with vinyl-pyrrolidinones over 3–10 hours to give the 1,2,3,4-tetrahydroquinoline derivatives in moderate to excellent yields with catalyst loadings as low as 0.1 mol%.
Moving away from more traditional Cr(iii) polypyridyls, a 2017 study from the group of Heinze has yet further exemplified the wide utility of the ‘molecular ruby’ complex 7 by demonstrating its use as a photocatalyst for C–H bond functionalisation of tertiary amines (Scheme 6).112 In this instance, a mechanism involving photo-induced electron transfer from the excited catalyst is ruled out, but rather the reaction proceeds via Dexter energy transfer from the long-lived 2Eg state to 3O2, efficiently generating singlet oxygen (Φ 1O2=61%). A subsequent charge transfer complex between 1O2 and the amine substrate may then lead to hydride transfer, giving hydroperoxide and an iminium ion, the latter being quenched by a cyanide source. Impressively, 7 can catalyse the formation of 71 in 89% yield after only 20 minutes irradiation time, being not only re-usable but also giving appreciable conversions (23%, 20 mins) when used at loadings as low as 0.01 mol%.112
10 Concluding remarks
As chemists look to develop sustainable photofunctional materials based upon Earth-abundant elements, it is unsurprising that considerable recent attention has been paid to new coordination complexes of Cr(iii). Whilst long appreciated for their rich electrochemical and magnetic properties, recent advances have seized upon the remarkable photophysical properties offered by complexes of Cr(iii), exploiting the ‘spin–flip’ which occurs in the electronic excited state to achieve particularly long-lived luminescence in the near-infrared spectral region from ligand-field excited states.
With many traditional examples of luminescent Cr(iii) complexes being restricted to those homoleptic systems of polypyridyl ligands, this review has highlighted efforts, mostly over the last decade, towards the use of novel ligand architectures whose molecular design is informed by an improved understanding of the nature of the electronic excited state, in addition to the discovery of new synthetic routes towards previously unknown heteroleptic complexes. Further valuable efforts have been spent in exploiting ligand-centred transitions in extending the electronic absorption profile into the visible region, whilst also gaining a vital understanding of the solution stability and photoreactivity of these systems. Thus, particularly impressive complexes typified by the so-called ‘molecular ruby’ have emerged whose photophysical attributes now closely rival those displayed by complexes of popular precious metals such as Ru(ii) and Ir(iii), lending themselves to use in a plethora of light-driven applications.
The potential utility of photo-active complexes of Cr(iii) is no better exemplified than by the reports, from only over the last 5 years, of the use of these materials as efficient photo-redox catalysts. Relying upon the most plentiful of energy sources, sunlight, these inexpensive catalysts open avenues towards new and transformative environmentally friendly organic synthesis.
Whilst the near future clearly offers researchers many exciting opportunities for the development of sustainable photofunctional materials based upon Earth-abundant elements, there remains significant challenges to overcome. These are principally concerned with achieving a greater degree of control over the synthesis of these new complexes, improving their long-term thermal- and photo-stability and most importantly, extending their electronic absorption profiles and photon-harvesting abilities to cover a greater portion of the solar emission spectrum.
However, whilst there clearly remains room for furthering our understanding of the chemical reactivity and photophysical properties of complexes of Cr(iii), the new and exciting photo-active examples detailed in this review highlight that Cr(iii) is indeed emerging as a credible alternative to the precious metals that we have relied upon over the last half century within photofunctional materials.