Singlet Oxygen Generation by Cyclometalated Complexes and Applications

Abstract

While cyclometalated complexes have been extensively studied for optoelectronic applications, these compounds also represent a relatively new class of photosensitizers for the production of singlet oxygen. Thus far, singlet oxygen generation from cyclometalated Ir and Pt complexes has been studied in detail. In this review, photophysical data for singlet oxygen generation from these complexes is presented, and the mechanism of ¹O₂ generation is discussed, including evidence for singlet oxygen generation via an electron transfer mechanism for some of cyclometalated Ir complexes. The period from the first report of singlet oxygen generation by a cyclometalated Ir complex in 2002 through August 2013 is covered in this review. This new class of singlet oxygen photosensitizers may prove to be rather versatile due to the ease of substitution of ancillary ligands without loss of activity. Several cyclometalated complexes have been tethered to zeolites, polystyrene, or quantum dots. Applications for photooxygenation of organic molecules, including “traditional” singlet oxygen reactions (ene reaction, [4+2] and [2+2] cycloadditions) as well as oxidative coupling of amines are presented. Potential biomedical applications are also reviewed.

INTRODUCTION

Some Sources of Singlet Oxygen – a Brief Overview

Singlet oxygen (¹Δ_g), the lowest excited state of the dioxygen molecule, can be produced thermally or photochemically (1–5). The former method involves loss of dioxygen in its singlet state from various precursors such as phosphite ozonites or various peroxides while the latter method involves either photochemically induced loss of ¹O₂ from a peroxide or, much more commonly, excitation of a photosensitizer followed by energy transfer to ground state (triplet) dioxygen. Singlet oxygen is widely used for oxyfunctionalizations of organic molecules. It is also involved in a variety of biological and biomedical processes: It is the major cytotoxic species in photodynamic therapy (PDT) (6) and some photoactive antiviral drugs (7). It may also be involved in apoptosis (8), and the chemistry of catalytic antibodies (9–10). Given these very diverse contexts and applications, it is not surprising that a number of different methods for thermal and photochemical singlet oxygen generation have been developed, and, despite the apparent abundance of thermal and photochemical sources of singlet oxygen, this has remained an active area of research.

For thermal singlet oxygen generation, the most important group of precursors are arguably aromatic endoperoxides, several hundred of which have been prepared and found to release ¹O₂ upon heating (11). Among the aromatic endoperoxides, anthracene derivatives have been most thoroughly investigated. Unfortunately, not all of these endoperoxides release ¹O₂ in high yield. In a classic paper, Turro and co-workers showed that the activation barrier for release of dioxygen from 9,10-anthracene endoperoxides via a retro [4+2] cycloaddition is ca. 38 kcal/mol (12–13). Since the excitation energy of singlet oxygen is 22.5 kcal/mol, dioxygen could be released either in the singlet or triplet state. Turro et al. demonstrated that 9,10-endoperoxides may undergo stepwise release of dioxygen, leading to a biradical intermediate which undergoes intersystem crossing (ISC) to a triplet biradical which then loses ³O₂ rather than ¹O₂. On the other hand, the retro-[4+2] cycloaddition in 1,4-endoperoxides is a concerted process leading only to singlet oxygen and the parent anthracene derivative. In general, attaching various substituents to a parent 1,4-anthracene derivatives allows for control over solubility, stability, and thus the temperature and yield of singlet oxygen release during the cycloreversion of the corresponding anthracene endoperoxides, making these compounds the most common thermal source of singlet oxygen in the laboratory (11).

In addition to these thermal sources, singlet oxygen can be produced photochemically by various dyes, including porphyrins and related compounds by photosensitization. Just as aromatic endoperoxides can be fine-tuned to affect the parameters of thermal singlet oxygen release, porphyrins and porphyrin derivatives as well as phthalocyanines can be fine-tuned to control the solubility, excitation wavelength, and singlet oxygen quantum yield. Several reviews have been published on singlet oxygen generation by these compounds (14–16). One drawback of these sensitizers is that synthetic modification to attach them to other materials or biomolecules is usually not trivial; aggregation of porphyrin-type sensitizers in solution is also a common problem. As an alternative, during the past decade, a number of boron-dipyrromethene (BODIPY)-based photosensitizers for production of singlet oxygen have been developed. This new class of sensitizers has recently been reviewed (17,18).

Yet another major class of singlet oxygen photosensitizers are various metal complexes. Numerous diimine transition metal complexes derived from Ru(II), Os(II), and Pt(II) as well as Ir(III) are singlet oxygen sensitizers (19–26). Perhaps the most promising group of metal-based singlet oxygen sensitizers are cyclometalated complexes of late transition metals. Just like porphyrins and phthalocyanins, the photophysical parameters of these compounds can be systematically manipulated by changes of the electronic properties of the complex (27). In addition, unlike other sensitizers, cyclometalated complexes often possess ancillary ligands that can easily be replaced, thereby allowing facile attachment of the sensitizer to other molecules, including nanomaterials or biomolecules such as DNA bases or proteins. This review will focus on singlet oxygen generation by cyclometalated complexes, and various applications of this relatively new class of sensitizers. All cyclometalated complexes which have thus far been evaluated as singlet oxygen sensitizers are either octahedral d⁶ Ir(III) complexes or square planar d⁸ Pt(II) complexes.

Photophysical Properties of Cyclometalated Complexes

During the past two decades, cyclometalated complexes have been primarily studied for applications in optoelectronic materials, especially organic light emitting devices (OLEDs) (28–31). The emissive properties of the excited states of all these complexes include high luminescence efficiencies and long excited state lifetimes. Cyclometalating ligands are strong-field ligands as the metal-carbon bond of the cyclometalating ligand enhances splitting of the d orbitals of the metal. Upon excitation of a cyclometalated complex, this large splitting of the metal d orbitals leads to preferential formation of metal-to-ligand charge transfer (MLCT) states instead of metal-centered excited states. The large extended π system of the cyclometalating ligand may also undergo excitation of an electron to a π* orbital thereby generating a purely ligand-based excited state. The high Z metal center promotes rapid ISC to a triplet π-π* ligand-based excited state. Consistent with this analyses, a computational study by Hay has shown that the LUMO and LUMO+1 are indeed π* orbitals located on the ligands (32). As the ligand-based MLCT and π-π* states are close in energy, there appears to be significant overlap between these two excited states, leading to a mixed MLCT-³[π-π*] state. In general, this mixed triplet excited state is the long-lived (usually in the microsecond range) emissive state of these complexes (33–36), although inter-ligand energy transfer from the ³MLCT state to an ancillary ligand (which thereby becomes the emitting center) may occur in some cases (37).

Ancillary ligands can be used to fine-tune the emission wavelength. For example, Yersin et al. have studied the photophysical properties of two bis-cyclometalated Ir complexes bearing identical cyclometalating ligands (2-(4′,6′-difluorophenyl)pyridine) (38). These complexes differ only in that one complex has an acetonylacetonato ancillary ligand, while the second complex has an N,O-bound picolinato ligand. The nitrogen atom possesses higher ligand field strength than the oxygen atom, thereby lowering the energy of the three occupied d-orbitals of the t_2g level (HOMO). This increases the d-π* energy gap, thereby causing a blue shift in the emission, as the energy of the MLCT transition is raised (39). The effect of electron-withdrawing substituents on the cyclometalating ligand is somewhat harder to predict. While such substituents lower the energy of the LUMO (i.e. the π* orbital on the cyclometalating ligand), they also withdraw electron density from the metal-based HOMO (i.e. the t_2g manifold) thereby raising its energy level. Usually the latter effect predominates (39).

Synthesis of Cyclometalated Iridium(III) Complexes

Octahedral Ir(III) complexes can possess one, two or three cyclometalating ligands. Biscyclometalated complexes of the type Ir(C^N)₂L₂ [C^N= cyclometalating ligand, L=ancillary ligand) are most common, but mono and tris-cyclometalated Ir(III) complexes (Ir(C^N)₃) can be prepared as well. Monocyclometalated Ir(III) complexes are comparatively rare, and have not yet been investigated as sensitizers for the production of singlet oxygen. Most Ir(III) complexes that have been investigated as singlet oxygen sensitizers are biscyclometalated complexes; in addition to the two cyclometalating ligands, they posses either two singly bound ancillary ligands or one chelating ancillary ligand. The synthesis of such biscyclometalated Ir(III) complexes is rather facile, and is usually accomplished by simply refluxing IrCl₃ in the presence of four equivalents of the cyclometalating ligand (33). This leads to a chloro-bridged dimer with two cyclometalating ligands on each Ir atom. The dimer can then be cleaved with an ancillary ligand. A wide range of compounds can be used to cleave the chloride-bridged dimer, ranging from 2,4-pentadione in the presence of base to bipyridine to amino acids which can be N,O-bound. Monoamines can also be used to cleave the dimer, in which case each of the monomers retains one of the two bridging chloride atoms. The latter reaction allows for facile linking of these sensitizers with spacers to biomolecules or nanomaterials (33, 39). The synthesis of tris-cyclometalated Ir(III) complex is also rather straightforward, as these compounds can be prepared from biscyclometalated precursors via displacement of an ancillary ligand. For example, for complexes containing 2-phenylpyridine as the cyclometatlating ligand, displacement of the acetonylacetate (acac) ligand from the biscyclometalated complex with another cyclometalating ligand and reflux at high temperature (>200°C) leads to the facial (fac) isomer of the tris-cyclometalated complex. The meridional (mer) isomer can be obtained by simply decreasing the reflux temperature after addition of the 2-phenylpyridine to the bis- cyclometalated complex to 140 °C. Alternatively, refluxing Ir(acac)₃ with three equiv. of the cyclometalating ligand represents a direct pathway to the fac isomer of the tris-cyclometalated complex (33, 40–42). The various routes to bis-and tris-cyclometalated complexes bearing 2-phenylpyridine ligands are shown in Scheme 1 below.

Scheme 1.

Recently, Swager and co-workers have developed another simple route to a wide range of cyclometalated Ir(III) complexes using a Cu(I) trazolide intermediate for the transmetalation step (43). This simple procedure allows a one-pot procedure for ligand synthesis and cyclometalation, thereby again allowing facile preparation of a wide range of these complexes with different ancillary ligands and substituents on the cyclometalating ligands. Overall, from a synthetic point of view, cyclometalated complexes are a class of singlet oxygen sensitizers that are quite easy to prepare and fine-tune both for different absorption profiles and for different applications. Literally hundreds of different complexes have been prepared during the past decade, and a review of the detailed synthesis and characterization of each of these complexes would be beyond the scope of this paper. We will focus on complexes for which photosensitized generation of singlet oxygen has been experimentally demonstrated.

SINGLET OXYGEN GENERATION BY CYCLOMETALATED COMPLEXES

Iridium(III) Complexes as Sensitizers - Mechanistic Aspects

In 2002, the groups of Thompson and Selke reported that several Ir(III)biscyclometalated complexes are excellent photosensitizers for the production of singlet oxygen (44). This was followed by a full paper in 2007 which explored the mechanism of singlet oxygen generation by these complexes (45). The structures of the Ir complexes (1–10) are depicted in Scheme 1 below. They all possess two cyclometalating ligands, and several different ancillary ligands, including N,O- bound glycine (Complex 5). Several different cyclometalating ligands were employed in this study, namely 2-phenylpyridine (ppy); 1-naphtylpyridine (1np); 2-phenylquinoline (pq); 2-(1-naphthyl)benzothiazole (bsn); 2-phenylbenzothiazole (bt); 2-phenylbenzo-oxazole (bo); and (2-pyridyl)benzothiophene) (btp).

Table 1 gives an overview of the photophysical data for all cyclometalated Ir(III) complexes that have been investigated for the production of singlet oxygen to date (August 2013). Many of these complexes produce singlet oxygen in high yield, with quantum yields approaching unity in some cases.

Table 1.

Singlet oxygen generation and related photophysical data for Ir cyclometallated complexes.

Compound	Solvent	Emission lifetime (degassed) (τ, μs)	Emission lifetime (aerated) (τ, μs)	Emission quenching by 3O2 (kq, 109 M−1 s−1)	Fraction of T1 quenched by 3O2 leading to formation of 1O2 (fT,Δ)	1O2 quantum yield (ΦΔ)	1O2 quenching rate (kT, 106 M−1s−1)	E1/2 (Sen+/*) (V)a	Ref.
[(bsn)2Ir(acac)] 1	C6H6					0.59 ± 0.07	6.3 ± 0.2		44
	2MeTHF	1.8							33
[(bsn)2Ir(dpm)] 2	C6H6	1.49	0.165	2.9 ± 0.1	0.73	0.60 ± 0.06	4.0 ± 0.3	−1.06	44, 45
[(pq)2Ir(acac)] 3	C6H6	1.5	0.071	7.2 ± 0.3	0.65	0.62 ± 0.05	1.0 ± 0.2	−1.28	44, 45
	2MeTHF	2							33
[(bt)2Ir(acac)] 4	C6H6	1.41	0.085	5.9 ± 0.6	0.92	0.86 ± 0.07	0.5 ± 0.2	−1.27	44, 45
	2MeTHF	1.8							33
[(bsn)2Ir(gly)] 5	C6H6					0.54 ± 0.02	2.1 ± 0.5		44
[(bt)2Ir(py)Cl] 6	C6H6					0.95 ± 0.09	None detected		44
[(ppy)2Ir(acac)] 7	toluene	1.43	0.026	23	0.92	0.90 ± 0.05	None detected	−1.63	45
	2MeTHF	1.6							33
[(bo)2Ir(acac)] 8	toluene	1.31	0.076	7.1	0.81	0.76 ± 0.06	0.18 ± 0.03	−1.31	45
	2MeTHF	1.1							33
[(1np)2Ir(acac)] 9	toluene	3.45	0.104	5.2	0.97	0.76 ±0.05	None detected	−1.35	45
[(btp)2Ir(acac)] 10	toluene	5.66	0.079	6.8	0.68	0.72 ± 0.06	None detected	−1.25	45
	2MeTHF	5.8							33
[Ir(dpyx)(phbpy)]+ 11	CH3CN	0.12	0.049	6.4		0.4 ±0.05			51
[Ir(ppy)2(bpy)]+ 12	CH3CN					0.5 ±0.05			51
[Ir(ppy)2(bpy)]+ 12	THF	0.18							51
[(ppy)2Ir(bpy)]+ 12	CH2Cl2: MeOH (9 : 1)	0.34		43	0.98	0.97			54
[Ir(dppy)(tpy-CO2Et)]+ 13	CH3CN					0.6 ±0.05			51
[(pip)2Ir(acac)] 14	CH2Cl2: MeOH (9 : 1)					0.26			52
[fac-Ir(ppy)3] 15	CH2Cl2: MeOH (9 : 1)	1.9		46	0.5	0.5			52
	toluene	1.6	0.036						65
[(ppy)2Ir(phen)]+ 16	CH2Cl2: MeOH (9 : 1)	0.88		33	0.93	0.93			52
[(ppy)2Ir(dmbpy)]+ 17	CH2Cl2: MeOH (9:1)	0.78		27	0.78	0.78			52
	CH3CN	1.2							66
[Ir(deatpy)3] 18	DMSO/ H2O						None detected		53
[Ir(deatpy)3-H3]3+ 19	DMSO/ H2O/HCl	1.5					+		53
[Ir(btpy)2(biS)]+ 20	CH2Cl2						+b	None detected	54
[Ir(btpy)2(biSe)]+ 21	CH2Cl2						++b	None detected	54
[Ir(Mebib)(ppy)]+ on SBA-15 22	CH3CN						+++c		56
[Ir(Mebib)(ppy)]+ on MCM-48 23	CH3CN						++c		56
[Ir(Mebib)(ppy)]+ on MCM-41 24	CH3CN						+c		56
[(piq)2Ir(pypzS2)]–CdSe/ZnS 25	MeOH						0.87		57
[(ppy)2Ir(bpy-CONH-PEG)]+ 26	CH2Cl2	0.38							58
	CH3CN	0.21							58
	DMSO		0.17				0.24		58
[(ppy)2Ir(bpy-CONH-Et)]+ 27	CH2Cl2	0.40							58
	CH3CN	0.23							58
	DMSO		0.16				0.38		58
[(pppy)2Ir(bpy-CONH-PEG)]+ 28	CH2Cl2	0.49							58
	CH3CN	0.34							58
	DMSO		0.24				0.38		58
[(pppy)2Ir(bpy-CONH-Et)]+ 29	CH2Cl2	0.50							58
	CH3CN	0.33							58
	DMSO		0.21				0.50		58
[(pba)2Ir(bpy-CONH-PEG)]+ 30	CH2Cl2	4.26							58
	CH3CN	2.86							58
	DMSO		0.54				0.58		58
[(pba)2Ir(bpy-CONH-Et)]+ 31	CH2Cl2	2.86							58
	CH3CN	3.72							58
	DMSO		0.54				0.79		58
[(pq)2Ir(bpy-CONH-PEG)]+ 32	CH2Cl2	2.08(10%) 0.78(90%)							58
	CH3CN	2.16(15%) 0.54(85%)							58
	DMSO		0.31				0.51		58
[(pq)2Ir(bpy-CONH-Et)]+ 33	CH2Cl2	0.67							58
	CH3CN	0.54							58
	DMSO		0.29				0.58		58
[(bsn)2Ir(bpy-CONH-PEG)]+ 34	CH2Cl2	4.1							58
	CH3CN	3.76							58
	DMSO		0.80				0.69		58
[(bsn)2Ir(bpy-CONH-Et)]+ 35	CH2Cl2	4.26							58
	CH3CN	3.72							58
	DMSO		0.78				0.83		58
[(ppy)2Ir(deac)] 36	CH3CN 2-Propanol 4:1	75.5					+d		62
[(ppy)2Ir(pdeac)] 37	CH3CN 2-Propanol 4:1	73.6					+d		62
[(ppy)2Ir(bpy-PBI)]+[PF6]− 38	MeOH	22.3	0.133				0.91		64
[(ppy)2Ir(bpy-aPBI)]+[PF6]− 39	CH2Cl2/ MeOH	8.7	0.177				0.15		64
[(ppy)2Ir(pBodipy)]+[PF6]− 40	MeOH	23.7					0.52		63
[(ppy)2Ir(2Bodipy)]+[PF6]− 41	MeOH	87.2					0.97		63

^aE(Sen^+/0*) = E(Sen^+/0) − E_0-0.

^bNot quantified but [Ir(btpy)₂(biSe)]⁺ had 2.4 times larger ¹O₂ emission intensity than [Ir(btpy)₂(biS)]⁺.

^cNot quantified but relative singlet oxygen levels were SBA-15>MCM-48>MCM-41.

^dNot quantified, but efficient photooxidation of 1,5-dihydroxynaphthalene to 5-hydroxy 1,4-naphthalenedione (juglone).

As discussed above, cyclometalated complexes possess a long-lived triplet state of mixed MLCT-π-π* character. This triplet state is rapidly quenched by triplet (ground state) oxygen. The fraction of triplet quenched by ground state oxygen (P_T,O2) can easily be determined by comparison of the triplet lifetime in degassed samples vs. in air (equation 1)

$$P_{\mathrm{T},\mathrm{O}2}=1-({\mathrm{\tau }}_{\text{air}}/{\mathrm{\tau }}_{\mathrm{N}2})$$ (1)

where τ_air and τ_N2 are the triplet lifetimes in air and under nitrogen, respectively. P_T,O2 can also be expressed as the fraction of the rate of triplet quenching (k_q) by triplet oxygen relative to the intrinsic rate of decay of the triplet state under anaerobic conditions (i.e. τ_N2⁻¹) (equation 2).

$$P_{\mathrm{T},\mathrm{O}2}={\mathrm{k}}_{\mathrm{q}}[{{}^3\mathrm{O}}_2]/({{\mathrm{\tau }}_{\mathrm{N}2}}^{-1}+{\mathrm{k}}_{\mathrm{q}}[{{}^3\mathrm{O}}_2])$$ (2)

In the absence of any chemical reaction, the rate of decay of the triplet state is the sum of the rate constants for nonradiative decay and phosphorescence, and hence eq. (2) can be rewritten as

$$P_{\mathrm{T},\mathrm{O}2}={\mathrm{k}}_{\mathrm{q}}[{{}^3\mathrm{O}}_2]/({\mathrm{k}}_{\mathrm{q}}[{{}^3\mathrm{O}}_2]+{\mathrm{k}}_{\text{nr}}+{\mathrm{k}}_{\mathrm{p}})$$ (3)

where k_nr and k_p are the rate constant for nonradiative decay and phosphorescence, respectively. The overall quantum yield of singlet oxygen production (Φ_Δ) is given by equation (3).

$${\mathrm{\Phi }}_{\mathrm{\Delta }}={\mathrm{\Phi }}_{\mathrm{T}}{P}_{\mathrm{T},{\mathrm{O}}_2}{f}_{\mathrm{T}},$$ (4)

where Φ_T is the quantum yield of triplet formation and f_T,Δ is the fraction of triplet quenched by ground state oxygen leading to formation of singlet oxygen. In cyclometalated complexes, the value of Φ_T is usually unity, as the heavy metal of these complexes facilities formation of the triplet mixed MLCT-π-π* state. On the other hand, the value of the fraction of triplet quenched by ground state oxygen leading to formation of singlet oxygen f_T,Δ may vary considerably, and often are considerably less than unity. Triplet oxygen may quench the triplet excited MLCT-π-π* state of the cyclometalated Ir complex by three different mechanisms, namely energy transfer which directly leads to formation of singlet oxygen ¹O₂ (equation 5), or electron transfer leading to formation of superoxide anion (O₂^•−, equation 6), or simple physical deactivation (equation 7).

$$[\text{Ir}{(\mathrm{C}^\mathrm{N})}_2{\mathrm{L}}_2]\ast +{{}^3\mathrm{O}}_2\to {{}^1\mathrm{O}}_2+[\text{Ir}{(\mathrm{C}^\mathrm{N})}_2{\mathrm{L}}_2]$$ (5)

$$[\text{Ir}{(\mathrm{C}^\mathrm{N})}_2{\mathrm{L}}_2]\ast +{{}^3\mathrm{O}}_2\to {{\mathrm{O}}_2}^{•-}+{[\text{Ir}{(\mathrm{C}^\mathrm{N})}_2{\mathrm{L}}_2]}^{+}$$ (6)

$$[\text{Ir}{(\mathrm{C}^\mathrm{N})}_2{\mathrm{L}}_2]\ast +{{}^3\mathrm{O}}_2\to {{}^3\mathrm{O}}_2+[\text{Ir}{(\mathrm{C}^\mathrm{N})}_2{\mathrm{L}}_2]$$ (7)

A classic scheme for singlet oxygen generation by energy transfer (equation 5) developed originally by Gijzeman et al. (46–47) correlates the values of the rate constant for quenching of triplet excited states by triplet oxygen (k_q, equation (2)) with values of f_T,Δ, i.e. the fraction of triplet quenched by ground state oxygen leading to formation of singlet oxygen. A triplet excited sensitizer and triplet oxygen reversibly form an encounter complex of (T₁³Σ) spin configuration. There are four unpaired electrons, and hence nine different spin configurations for this encounter complex, but only the singlet [¹(T₁³Σ)] state of the encounter complex can dissociate to form singlet oxygen and ground-state sensitizer. Quenching rates of 1/9 of the diffusion-controlled rate limit (k_diff) thus imply exclusive quenching of the triplet excited sensitizer by energy transfer and a value of unity for f_T,Δ. Larger values for kq should lead to values of and f_T,Δ of less than unity, as quenching from other states besides the singlet [¹(T₁³Σ)] state of the encounter complex must occur. Empirically, quenching rates near 1/9 k_diff and very high singlet oxygen quantum yields have been observed for many different sensitizers, including some metal complexes, i.e. many metalloporphyrins (48–49). However, this scheme had to be subsequently modified to account for cases in which the f_T,Δ was near unity while the value of k_q was considerably larger than 1/9 k_diff. Formation of charge-transfer intermediates from the initial triplet excited sensitizer-triplet oxygen encounter complexes may take place in these cases. Singlet oxygen is then formed from these charge transfer complexes. Thus, for example, several Re(I), and Ir(III) bearing bipyridine (bpy) ligands have k_q values near 10¹⁰ M⁻¹sec⁻¹ (i.e. much larger than 1/9 k_diff) but produce singlet oxygen with quantum yields near unity which implies values of f_T,Δ near unity (21–22).

All of the biscyclometalated Ir complexes studied by Djurovich et al. (1–10) produce singlet oxygen in moderate to high yield (44–45). Singlet oxygen quantum yields were determined directly from the time-resolved ¹O₂ near infrared (NIR) emission signal. The authors also reported very large values of triplet oxygen quenching by the triplet excited complexes, in one case (Complex 7) over an order of magnitude larger than 1/9 k_diff (45). They suggested that that such high rates may be indicative of quenching by electron transfer (equation 6), in addition to an energy transfer mechanism (equation 5) leading to formation of superoxide anion and a sensitizer radical cation. Singlet oxygen would then be formed from back-electron transfer. Support for this hypothesis was derived from the fact that the very large observed values of k_q at the diffusion-controlled rate limit, which is typical for an electron transfer process and that the excited states of some tris-cyclometalated Ir(III) complexes are in fact quenched by triplet oxygen via an electron-transfer process (50). Further evidence for this hypothesis was derived from the observation that for (C^N)₂Ir(acac) complexes 2–4 and 7–10 there exists a correlation between the decrease in the values of f_T,Δ and the reduction potential of the triplet excited state of these complexes (E_1/2 (Sen^+/*) (41). As can be seen for complexes 2–4 and 7–10 in Table 1, singlet oxygen quantum yields (and values of f_T,Δ) decrease with decreasing excited state reduction potential. If singlet oxygen generation by cyclometataled Ir(III) complexes occurs at least partly by an electron transfer/back electron transfer mechanism, it follows that the singlet oxygen quantum yield should decrease as an unfavorable excited state reduction potential diminishes formation of such electron transfer or charge transfer intermediates.

Several other groups have determined singlet oxygen quantum yields for cyclometalated iridium complexes. In 2008, Williams and co-workers reported singlet oxygen generation by two interesting Ir(III)complexes possessing two terdentate ligands of the type Ir(N^C^N)(N^N^C, 11) and Ir(C^N^C)(N^N^N, 13), as well as the biscyclometalated complex 12 which has an ancillary bipyridine ligand (Scheme 3) (51). Time-resolved measurements of singlet oxygen luminescence (excitation wavelength at 355 nm) gave singlet oxygen quantum yields of 0.4–0.6. The reason for the somewhat lower quantum yields (compared to complexes 1–10) are not known. Murata and coworkers subsequently studied singlet oxygen generation by a number of neutral and cationic cyclometalated complexes (12 and 14–17, Scheme 2) (52). This includes cationic complexes 12, 16 and 17 as well as the tris-cyclometalated complex fac(Ir(ppy)₃ (15). Singlet oxygen quantum yields were measured by a chemical method, namely trapping of ¹O₂ by 1,5-dihydroxynaphthalene. They also determined values for quenching of the triplet state of these complexes by triplet oxygen, and obtained values near the diffusion-controlled rate limit. However, only the cationic complexes - except complex 17 - gave very high singlet oxygen yields, while those for the neutral complexes were lower. The reduced quantum yield of 17 was attributed to a steric blocking effect of the methyl groups on the bipyridine moiety. For the neutral complexes, the authors suggested that while electron transfer from the excited complex to triplet oxygen (eq.(6) is extremely rapid, the back electron transfer required to produce singlet oxygen by the electron transfer mechanism is unfavorable (52).

Scheme 3.

Scheme 2.

Aoki and co-workers prepared tris-cyclometalated Ir complexes bearing amino-substituted 2-phenyl pyridine ligands (18–19, Scheme 3) (53). They reported that the neutral complex 18 does not produce any singlet oxygen while the triprotonated complex 19 does, although the quantum yield is not known at this time. Singlet oxygen production was monitored by a chemical trap, i.e. 1,3-diphenylisobenzofuran (DPBF). Singlet oxygen production by this system was turned on by addition of HCl and turned off by addition of 1,8-diazabicyclo[5.4.0]undec-7-ene (DBU).

Son and co-workers prepared two interesting biscyclometalated complexes bearing an S-bound dithione (20) and a Se-bound di-selenone (21, Scheme 5; (btpy = bis(2-(2+-benzothienyl)pyridine; biS = bis(imidazoline thione; biSe = bis(imidazoline selone), 20) (54). The authors were able to detect singlet oxygen by its near-infrared emission signal. While quantum yields were not determined, the authors noted that the singlet oxygen emission intensity from the selenone complex 21 was 2.4 times that of the thione complex 20. The compounds were used for photooxidative couplings of benzylamines to imines.

Scheme 5.

Two groups (Chi and Yamashita) evaluated singlet oxygen production of cyclometalated Ir(III) complexes tethered to either zeolites (22–24) (55–56) or quantum dots (57) (Scheme 6). Singlet oxygen quantum yields were not determined in most cases except for a system in which the Ir(III) biscyclometalated complex [(piq)₂Ir(pypzS₂), 25] is attached the surface of a CdSe/ZnS quantum dot (QD) (57). No significant Förster resonance energy transfer (FRET) was observed from the QD to the Ir complex. Singlet oxygen was thus produced by direct excitation of the complex. Singlet oxygen production of the Ir-CdSe/ZnS QD hybrid material was assessed by monitoring of the singlet oxygen NIR emission in aerated MeOH solution. It appears that the presence of the QD does not significantly reduce production of ¹O₂ by this system. Using bis(triisobutylsiloxy) silicon-2,3-naphthalocyanine (SiINC) as a reference, the singlet oxygen quantum yield Φ_Δ was estimated to be 0.87 (57).

Scheme 6.

Lo and co-workers prepared a series of poly(ethylene glycol) (PEG)-modified Ir complexes and their free analogues (26–35) for potential use in photodynamic therapy (Scheme 7) (58). The PEG-Ir complexes are [Ir(ppy)₂(bpy-CONH-PEG)](PF₆) (26), [Ir(pppy)₂(bpy-CONH-PEG)](PF₆) (28), [Ir(pq)₂(bpy-CONH-PEG)](PF₆) (30), [Ir(bsn)₂(bpy-CONH-PEG)](PF₆), (32) and [Ir(pba)₂(bpy-CONH-PEG)](PF₆) 34), while complexes 27, 29, 31, 33, and 35 are the corresponding free analogues (bpy-CONH-PEG = 4-(N-(2-(ω-methoxypoly-(1-oxapropyl))ethyl)aminocarbonyl)-4′-methyl-2,2′-bipyridine; ppy = 2-phenylpyridine; pppy = 2-((1,1′-biphenyl)-4-yl)pyridine; pq = 2-phenylquinoline; bsn = 2-(1-naphthyl)benzothiazole; pba = 4-(2-pyridyl)benzaldehyde). The PEG-free compounds have (bpy-CONH-Et) in place of (bpy-CONH-PEG).

Scheme 7.

Emission data taken for all ten compounds showed that the PEG-Ir compounds were generally similar to their corresponding PEG-free compounds giving green to red emissions and long triplet lifetimes (58). Singlet oxygen production was determined by monitoring the consumption of 1,3-diphenylisobenzofuran (DPBF) at 418 nm in an aerated DMSO solution. Quantum yields were calculated using methylene blue as a reference. A limiting factor in singlet oxygen production by some of these complexes (i.e. 26 and 27) may be a relatively small value for k_q such that the term k_q[³O₂] in equation (2) becomes small enough for other decay processes of the triplet excited complex (i.e. non-radiative decay) to be competitive with triplet oxygen quenching. Complexes with the longest emission lifetimes (34 and 35) had the highest singlet oxygen quantum yields, indicating that longer triplet lifetimes result in higher quantum yields of singlet oxygen. Steric factors were also found to be important, as quantum yields were slightly higher in the PEG-free complexes most likely due to the inability of oxygen to freely diffuse into and out of the PEG complex, making it more difficult to both form singlet oxygen, and for the singlet oxygen produced to then reach the DPBF (58).

Recently several cyclometalated iridium complexes with dramatically enhanced visible absorption have been reported (59–64). While conventional cyclometalated irdium complexes typically have moderately strong absorption in the visible region (typically, ε < 10⁴ M⁻¹cm⁻¹), the systems developed by Zhao et al. consist of a cyclometalated Ir complex to which an organic fluorophore is attached which increases visible absorption by at least an order of magnitude. These complexes all possess two cyclometalating 2-phenylpyridine (ppy) ligands, and an ancillary bipyrdine (bpy) ligand to which the fluorophore is attached. Several different fluorophores have been utilized, namely diethylcoumarin and phenyldiethylcoumarin ([ppy)₂Ir(deac)]⁺ [PF₆]⁻ (36) (59, 62) and [(ppy)₂Ir(pdeac)]⁺[PF₆]⁻ (37)), perylenebisimide and an amino-substituted derivative ([(ppy)₂Ir(bpy-PBI]⁺[PF₆]⁻(38) (61, 64) and [(ppy)₂Ir(bpy-aPBI)]⁺[PF₆]⁻ (39), and boron-dipyrromethene (BODIPY) (63). Both the PBI and BODIPY ligands were attached via an acetylene linker to the bpy ligand of (ppy)₂Ir(bpy). Two different modes of attachment were used for the BODIPY moiety: attachment via the meso-phenyl group to afford [(ppy)₂Ir(pBodipy)]⁺[PF₆]⁻ (40) and attachment via the 2-position of the core π-system giving [(ppy)₂Ir(2Bodipy)⁺[PF₆]⁻ (41) (Scheme 8).

Scheme 8.

All of the complexes 36 – 41 have considerably longer emission lifetimes than conventional cyclometalated Ir complexes (columns 3 and 4 in Table 1). Thus, the perylenebisimide complex 38 has triplet lifetime of 22.3 μs, and a singlet oxygen quantum yield Φ_Δ of 0.91. Addition of the amino group caused a red shift in absorption from 546 nm in compound 38 to 675 nm in the amino-substituted derivative 39. Unfortunately, the amino-substituted complex 39 has a shorter triplet state lifetime (8.7 μs),s), and a significantly reduced singlet oxygen quantum yield (Φ_Δ = 0.15). For the BODIPY complexes, conjugation of the BODIPY with the core (i.e. 2Bodipy, 41) provides more efficient population of the excited triplet state, and thus more efficient singlet oxygen generation compared to pBodipy (40). Complex 40 has a triplet lifetimes of 23.7 μss and a singlet oxygen quantum yield of 0.52. Complex 41 has a red-shifted maximum absorbance at 534 nm (compared to 504 nm for complex 40), an exceptionally long triplet lifetime of 87.2 μss and a singlet oxygen quantum yield of 0.97. All of these quantum yields were calculated using 1,3 diphenylisobenzofuran (DPBF) as a singlet oxygen scavenger and Rose Bengal as a standard.

Cyclometalated Platinum(II) Complexes as Singlet Oxygen Sensitizers

The second group of cyclometalated complexes that has been studied for the production of singlet oxygen are square-planar d⁸ Pt(II) complexes. Similar to the octahedral Ir(III) complexes, these compounds can be prepared from the corresponding dichloride bridged dimers bearing a cyclometalating ligand on each Pt atom by reaction with a variety of compounds such as acetonyl actetonate to bipyridines, quinolines, quinoline thiols as well as many mondentate ligands ranging from triphenylphosphine to pyridine. Tridentate cyclometalating ligands such as C-deprotonated 6-phenyl-2,2′-bipyridine [C^N^N]⁻ have also been used to prepare cationic complexes of the type [Pt(C^N^N)R¹]ⁿ⁺ (67–72). The photophysical properties of these complexes are quite similar to their Ir counterparts. However, in many cases, the lowest excited state is a triplet intra-ligand charge-transfer (ILCT) state formed when the HOMO is located on the cyclometalating ligand and the LUMO on an ancillary ligand such as a quinoline. The triplet ILCT state of such complexes is the emissive state with a long lifetime in the microsecond range (73). Mixed triplet MLCT and ILCT states have also been reported (70, 74–75). Recent theoretical investigation found that for Pt complexes with a cyclometalating 2-phenylpyridine ligand and a quinoline ligand, the energy gap between the ground state and the triplet excited state indeed exceeds the excitation energy of singlet oxygen (22.5 kcal/mol), indicating that such complexes can indeed produce ¹O₂ (76). The first report of singlet oxygen production by Pt cyclometalated complexes was by Weinstein and co-workers in 2006 (77). Singlet oxygen quantum yields were determined directly from the time-resolved ¹O₂ near infrared (NIR) emission signal, using phenanelone (78) as a reference compound. The cyclometalated complexes studied by Weinstein et al. (42–50) are depicted in Scheme 9. Singlet oxygen quantum yields for these compounds range from 0.5–0.9, both in toluene and methylene chloride. The authors also determined the quenching rates of the triplet excited states by triplet oxygen. Unlike for some of the Ir(III) complexes discussed above (especially 7, 8, 10), these were found to be approximately 1/9 of k_diff, indicating that most likely singlet oxygen production by these complexes occurs only by energy transfer.

Scheme 9.

Subsequently, Djurovich et al. reported singlet oxygen generation from six different Pt(II) cyclometalated complexes (51–56) (79) with quantum yields near unity in all cases (45). The cyclometalating ligands were the same as those used for Ir complexes 1–10. Measurements were accomplished via excitation at 355 nm and monitoring the ¹O₂ (NIR) emission signal. Again, quenching rates of the triplet excited states of these complexes by triplet oxygen were found to be approximately 1/9 of k_diff, and, since Φ_Δ is unity, f_T,Δ must be unity as well, indicating that most likely singlet oxygen production occurs by energy transfer. Singlet oxygen quenching rates (k_T) by Pt complexes 51–56 were also determined, and found to be somewhat higher (6.9 × 10⁶ – 2.2 × 10⁷ M⁻¹ sec⁻¹) than those of the analogous Ir complexes, but still moderately low, and lower than sensitizers such as tetraphenyl porphyrine (k_T = 4.4 × 10⁷ M⁻¹ sec⁻¹) (80). The authors also noted that steric bulk on the ancillary ligands does not appear to affect the singlet oxygen quantum yield (45).

Lai et al. reported singlet oxygen quantum yields for a series of 2-(2′-thienyl)pyridyl platinum(II) complexes bearing various ancillary ligands such as Cl, PPh₃ and pyridine (57–63, Thpy=2-(2′-thienyl)pyridyl) (81). The lowest excited state of complexes 57–63 is a triplet intra-ligand charge-transfer (ILCT) state, as is common for cyclometalated Pt complexes.

The quantum yields for singlet oxygen production were obtained by time-resolved laser measurements of NIR signal intensity of ¹O₂ at an excitation wavelength of 355 nm in deuterated acetonitrile and dichloromethane (Table 2) using C₆₀ (Φ_Δ = 1) (82) as a reference. The cationic complexes 58 and 59 had a notably higher singlet oxygen quantum yield compared to their neutral counterparts. The triplet excited states of these cationic compounds also have a considerably longer lifetime, indicating a smaller value for the non-radiative decay rate constant, and hence a larger value for the term P_T,O2 in equations (2) and (3). The authors also investigated two binuclear Pt complexes linked via a 5,17-bis(diphenylphosphino)-25,26,27,28-tetra-n-butoxycalix[4]arene group (62 and 63) (83). They found that neither the large calixarene nor the second metal center significantly decrease the singlet oxygen quantum yield (81).

Table 2.

Singlet oxygen generation and related photophysical data for Pt cyclometallated complexes.

Compound	Solvent	Emission lifetime (degassed) (τ, μs)	Emision lifetime (aerated) (τ, μs)	Emission quenching by 3O2 (kq, 109 M−1 s−1)	Fraction of triplet quenched by 3O2 leading to formation of 1O2 (fT,Δ)	1O2 quantum yield (ΦΔ)	1O2 quenching rate (kT, 106 M−1 s−1)	Ref.
[(thpy)Pt(QOH)] 42	CH2Cl2	5.3	0.26			0.82		77
	Toluene					0.82
[(ppy)Pt(QOH)] 43	CH2Cl2	4.3	0.26			0.80		77
	Toluene					0.90
[(dba)Pt(QOH)] 44	CH2Cl2	4.4	0.23			0.80		77
	Toluene					0.84
[(thpy)Pt(QOCl)] 45	CH2Cl2	2.4	0.3			0.81		77
	Toluene					0.85
[(thpy)Pt(QOMe)] 46	CH2Cl2	1.8	0.26			0.59		77
	Toluene					0.67
[(ppy)Pt(QOMe)] 47	CH2Cl2	1.3	0.1			0.51		77
	Toluene					0.65
[(dba)Pt(QOMe)] 48	CH2Cl2	0.6	0.17			0.51		77
	Toluene					0.54
[(thpy)Pt(QS)] 49	CH2Cl2	1.6	0.21			0.64		77
	Toluene					0.86
[(ppy)Pt(QS)] 50	CH2Cl2	1.6	0.21			0.68		77
	Toluene					0.69
[(ppy)Pt(acac)] 51	Toluene	2.9	0.155	3.1	1.03			45
	MeOD					0.98 ± 0.12	15 ± 4
[(ppy)Pt(dpm)] 52	Toluene	2.52	0.15	3.3	1.01			45
	MeOD					0.95 ± 0.09	22 ± 6
[(bt)Pt(dpm)] 53	Toluene	5.65	0.272	1.9	1.05			45
	MeOD					1.03 ± 0.10	7.7 ± 1.5
[(pq)Pt(dpm)] 54	Toluene	7.08	0.333	1.8	1.01			45
	MeOD					0.96 ± 0.06	20.8 ± 6
[(btp)Pt(acac)] 55	Toluene	8.85	0.179	1.9	1.02			45
	MeOD					1.02 ± 0.10	6.9 ± 1.2
[(btp)Pt(dpm)] 56	Toluene	9.87	0.177	1.9	1.02			45
	MeOD					1.04 ± 0.05	7.0 ± 1.1
[Pt(Thpy)(PPh3)Cl] 57	CH3CN	6.8	< 0.1			0.44 ± 0.02	14.0 ± 0.4	81
	CD2Cl2					0.12 ± 0.03
[Pt(Thpy)(PPh3)CH3CN]ClO4 58	CH3CN	26.2	< 0.1			0.75 ± 0.04	12.6 ± 0.5	81
	CD2Cl2					0.27 ± 0.05
[Pt(Thpy)(PPh3)py]ClO4 59	CH3CN	6.73	0.26			0.82 ± 0.06	14.6 ± 0.4	81
	CD2Cl2					0.55 ± 0.05
[Pt(Thpy)(HThpy)Cl] 60	CH3CN	12	< 0.1			0.56 ± 0.04	36.4 ± 0.4	81
	CD2Cl2					0.31 ± 0.03
[Pt(Thpy)(Hthpy)py]ClO4 61	CH3CN	24	0.33	554	0.38	0.95 ± 0.05	55.4 ± 0.3	81
	CD2Cl2					0.27 ± 0.03
[(PtThpyCl)2Calix(OnBu)] 62	CH3CN	1.38	< 0.1	559	0.021	0.42 ± 0.03	9.1 ± 1.1	81
	CD2Cl2					0.15 ± 0.03
[{PtThpy(CH3CN)}2Calix(On Bu)](ClO4)2 63	CH3CN	25.2	0.35	556	0.2	0.81 ± 0.06	16.9 ± 0.2	81
	CD2Cl2					0.26 ± 0.02
(acac)Pt(deac) 64	MeOH	31.4				0.64		85
(acac)Pt(coumarin) 65	CH2Cl2	2.7				0.76		85
Pt(dpbpy)SBA-15 66	CH3CN					0.40		86
PolyPt(pbp)L 67	CH2Cl2	0.3				0.79		87

Zhao et al. used ligands with strong absorption in the visible range to prepare a new class of cyclometalated Pt complexes (59, 84–85). They synthesized a 7-diethylaminocoumarin-benzothiazole (deac) Pt complex (acac)Pt(deac) (64), and a non-aminated coumarin-benzothiazole containing compound, (acac)Pt(coumarin) (65). In both cases, the coumarin-benzothiazole moiety was directly cyclometalated to the metal center (Scheme 12), in contrast with the cyclometalated Ir complexes 36 and 37 prepared by the same group in which a coumarin moiety was attached to an ancillary bipyridine ligand (Scheme 8).

Scheme 12.

While the triplet lifetime of 64 is much longer than that of 65 (31.4 μs vs. 2.7 μs), both compounds 64 and 65 are good singlet oxygen sensitizers with Φ_Δ = 0.64 and 0.76, respectively. In addition, complex 64 has unusually strong absorption in the visible range (λ_max = 495 nm, ε= X 10⁴ M⁻¹cm⁻¹). Singlet oxygen quantum yields were calculated using DPBF as singlet oxygen scavenger and Rose Bengal and TPP as standards (85).

Wu and co-workers have found that cyclometalated Pt complexes attached to zeolites (66) (86) or polystyrene (67) (87) still produce singlet oxygen in moderate to high yield, i.e. Φ_Δ= 0.40 and 0.79, respectively (Scheme 13). Singlet oxygen production from 66 and 67 was assessed by monitoring formation of 2,2,6,6-tetramethylpiperidine oxide (TMPO) upon irradiation of 66 or 67 in the presence of 2,2,6,6-tetramethylpiperidine (TMP).

Scheme 13.

Singlet oxygen quantum yields and related photophysical properties for all of the Pt complexes (42–67) are summarized in Table 2 below.

APPLICATIONS OF CYCLOMETALATED COMPLEXES AS SINGLET OXYGEN SENSITIZERS

Oxidation of Organic Compounds with Singlet Oxygen Generated by Cyclometalated Complexes

Several groups have used cyclometalated sensitizers to carry out a wide variety of photooxidations of organic molecules, including all of the traditional singlet oxygen reactions (ene reaction, [4+2] addition and [2+2] addition) (1,4), as well as oxidative coupling of amines. Some of these investigations attempted to evaluate the stability and efficiency of these sensitizers compared to traditional singlet oxygen sensitizers, while other have attached these sensitizers to various materials in order to develop new nanoreactors for oxidation of organic molecules.

Murata and co-workers used the Ir cyclometalated sensitizers 7 (Scheme 2), 12, and 14–17 (Scheme 3) to study the oxidation of 1,5-dihydroxynaphthalene (68) to 5-hydroxy-1,4-naphthalenedione (69) (Juglone) (52). The authors noticed reduced yields of the oxidation product 69 after moderately long (15 minutes or more) irradiation of sensitizers 7 and 14, while the other sensitizers gave constant yields of the oxidation product regardless of irradiation time. This was attributed to photochemical loss of the acetyl acetonate (acac) ligand, leading to decomposition of the complex. In contrast, the cationic complexes 12, 16–17 (which do not have the ancillary acac ligand) and the tris(cyclometalated complex 15 did not undergo any photochemical decomposition. The cyclometalated Ir complexes with organic fluorophores 36–41 and the related Pt coumarin complexes 64 and 65 have also been used oxidize 1,5-dihydroxynaphthalene (68) to 5-hydroxy-1,4-naphthalenedione (69) (Juglone). Efficiencies of conversion of 68 to 69 were higher than for conventional biscyclometalated complexes, and generally correlated with the singlet oxygen quantum yields (52).

An interesting application of the dithione and diselone Ir complexes (Scheme 5) for photooxidative coupling of benzylamines to imines was developed by Son et al. (54) Both [(btpy)₂Ir(biS)⁺] (btpy = bis(2-(2′-benzothienyl)pyridine; biS = bis(imidazoline thione), 20) and [(btpy)₂Ir(biSe)⁺] (biSe = bis(imidazoline selone), 21) were able to efficiently catalyze the oxidative coupling of a number of benzylamine and dervatives 70–76 (Scheme 14). These coupling reactions were carried out at room temperature with visible light from a conventional LED with maximum emission at 460 nm, close to the absorption maxima of complexes 20 and 21. The Ir-dithione complex 20 resulted in 45% conversion of 0.1 mol benzylamine to imine in 5 hours while the Ir-diselone complex 21 led to 75% conversion under the same conditions and, at 0.25 mol % of benzylamine, 94% conversion was achieved. Replacement of the dithione or diselone ancillary ligands with N-hetercyclic dicarbene ligands led to poor yields. Further reactions were undertaken with the diselone sensitizer 21 and various benzylamine derivatives, i.e. 4-methoxybenzylamine (71, 99% yield), 4-methylbenzylamine (72, 95% yield), 4-chlorobenzylamine (73, 89% yield), 4-flourobenzylamine (74, 76% yield), 1-phenylethylamine (75, 61% yield), and (2-thienyl)methylamine (76, 60% yield).

Scheme 14.

Oxidative coupling reactions of this type may follow either a Type I or Type II photooxidation pathway. In the Type I pathway, the excited sensitizer transfers an electron to triplet oxygen forming a superoxide radical and sensitizer radical cation (88). Electron transfer from the amine to the oxidized sensitizer leads to a radical cation amine which can then react with the superoxide resulting in formation of an imine. Alternatively, a singlet oxygen pathway (Type II) can also lead to oxidation of the amine to the imine (89). The calculated excited state energy levels of 21 suggest that electron transfer from the benzylamine to oxidized 21 is unlikely, thus disfavoring the Type I photoredox pathway. Furthermore, the singlet oxygen emission signal from the diselone complex 21 was 2.4 times high than that of the dithione complex 20, consistent with a singlet oxygen pathway, as 21 leads to a considerably higher yields in the coupling reaction (54).

Substitution of ancillary ligands of cyclometalated complexes can be used to tether these complexes to a variety of materials. Thus, Yamashita and co-workers anchored the cyclometalated complex [Ir(bis(N-methylbenzimidazolyl)pyridine)(phenylpyridine)Cl] ([Ir(Mebib)(ppy)Cl], 77) to three different mesoporous silica materials (SBA-15, MCM-41 and MCM48) modified with (3-aminopropyl)trimethoxysilane (APTMS) giving [Ir(Mebib)(ppy)⁺]/APTMS-SBA-15 (22) [Ir(Mebib)(ppy)⁺]/APTMS-MCM-41 (23), and [Ir(Mebib)(ppy)⁺]/APTMS-MCM-48 (24) (56). In all three cases the pore diameter (2.0–7.1 nm) of the APTMS-modified mesoporous silica was sufficiently large to allow uniform distribution of the Ir complex throughout the channels (Ir: 0.4 wt %). The free Ir complex showed an intense phosphorescence band at 540 nm in degassed CH₃CN at 77 K which was drastically reduced at room temperature. This same reduction was not observed in any of the anchored Ir-mesoporous silica compounds. The intensity of the emission for the three compounds decreased in the order Ir-SBA-15 > Ir-MCM-48 > Ir-MCM-41. Compounds 22–24 were used for the oxidation of 1-naphthol (78) to 1,4-naphthoquinone (79) and trans-stilbene (80) to benzaldehyde (81) – two classic singlet oxygen reactions. For both reactions, the photocatalytic activity trend followed the emission intensity trend (i.e. Ir-SBA-15 > Ir-MCM-48 > Ir-MCM-41). To rule out the possibility of other reactive oxygen species playing a role, the reaction was separately run with the addition of either isopropyl alcohol (a hydroxy radical quencher) or superoxide dismutase and little effect catalytic activity was observed (56).

Two other catalytic systems using using cyclometalated complexes as photosensitizers anchored to mesoporous silica (66) or polystyrene (67) were developed by Wu and co-workers (Scheme 16) (86–87). The Pt(II) 4,6-diphenyl-2,2′-bipyridine complex ([Pt(dpbpy)Cl], 82) was attached to (3-aminopropyl)triethoxysilane (APTES)-modified mesoporous silica (SBA-15) by substitution of the chloride of the Pt complex with the amino group of the APTES resulting in the Pt complex anchored to mesoporous silica (66), analogous to the procedure for the Ir systems 22–24 (56, 86).

Scheme 16.

Using 66 as a singlet oxygen sensitizer, Wu et al. carried out a variety of photooxygenations (Scheme 17), namely an ene reaction with (1S,5S)-(−)-α-Pinene (83) leading to the allylic hydroperoxide 89, [2+2] additions with trans-stilbene (84) and 1,2-diphenyl-1,2-dimethoxyethylene (85) leading to the dioxetane cleavage products 90 and 91, as well as [4+2] additions with 1,3-cyclopentadiene (86) and 1,3-cyclohexadiene (87). The endoperoxides resulting from the latter two reactions were reduced with thiourea to the corresponding diols 92 and 93. Lastly, the photooxidation of 7-dehydrocholesterol (88) with 66 lead to both an allylic hyroperoxide (ene product, 94) and an endoperoxide ([4+2] product, 95) with a ratio of 3:1. All of these reactions proceeded with nearly quantitative yields. The high yields may be due to the large local olefin concentration within the pores of the mesoporous silica which was found to be eight times larger than in homogenous solution. The mesoporous silica pores function both as a heterogenous support for the Pt compound and nano-reactor for photooxidation. The Pt-mesoporous silica complex was quite stable; even after ten runs, no significant loss of reactivity nor leaching of the Pt ligand were observed. The complex could also be recovered readily through simple filtration (86).

Scheme 17.

A subsequent study by Wu and co-workers reported a cyclometalatd Pt complex attached to a polystyrene matrix (87). A precatalyst (MonoPt(pbp)Cl, 96) bearing a styrene molecule attached to the cyclometalating ligand 6-phenyl-2,2′ bipyridine-4-carboxylate (pbp) was polymerized with styrene to generate a polymer-support cyclometalated Pt(II) complex. Subsequent replacement of the ancillary chloride ligand with phenylacetylene led to the polymer Poly(Pt(pbp)L, where L = phenylacetylene, 67) (Scheme 16).

Just as 66, this system has absorption maxima in the vis range (459 and 488 nm). The authors ran the same photooxygenation reactions as for 66, with identical results. Catalyst loading was higher for 67 than for 66 (2 mol% vs 0.8 mol%). Due to its incorporation into the polystyrene matrix, the complex was found to be soluble in many solvents including CH₂Cl₂, CHCl₃, DMF, and ethyl acetate. It could also be easily precipitated with polar solvents such as MeOH allowing for recycling. The polymer appears to be resilient to degradation as there was no observed decrease in activity over five photooxidation cycles with dimethoxystilbene (87).

Potential Biomedical Applications of Cyclometalated Sensitizers – Photodynamic Therapy

Several groups have investigated the photocytotoxicity of cyclometalated complexes. Lai et al. found that the Pt complexes 59 and 61 (which had the highest singlet oxygen quantum yields among the seven complexes 57–63 studied by these authors) exhibit photocytotoxicity towards HeLa cancer cells (human cervical epitheliod carcinoma). Substantial uptake of these complexes at the nucleus and mitochondria was observed (81).

Lo and co-workers studied a series of poly(ethylene glycol) (PEG)-modified cyclometalated Ir complexes (26, 28, 30, 32, 34) and their free analogues (27, 29, 31, 33, 35) for use in photodynamic therapy (Scheme 7) (58). Attachment to a 5 kDa PEG chain helps the Ir compound’s water-solubility and prevents unfavorable interactions with cellular components such that the Ir complex remains in the cell without causing cell damage until exposure to light, whereupon production of singlet oxygen induces cell death. Less polar cyclometalating ligands are more lipophilic and thus more readily taken up by the cell, with complexes 34 and 35 having the highest uptake. Interestingly, cellular uptake by the PEG compounds was found to be slower than for the PEG-free compounds. As tumors are known to accumulate compounds with slow intake times faster than normal cells, this could potentially be a very useful phenomenon for photodynamic therapy. The intake was found to be concentration dependent and localized in the mitochondria.

The cytotoxicity of all ten compounds (26–35) was tested in HeLa cells by 3-(4,5-dimethyl-2-thiazolyl)-2,5-diphenyltetrazolium bromide (MTT) assay. All PEG compounds were found to be non-cytotoxic in the dark (IC₅₀ > 300 μM) with IC₅₀ values dropping to 3.4–23.2 μM upon irradiation positively correlating with singlet oxygen quantum yields. One exception was the pba-PEG compound 30 which showed very low cytotoxicity both in the dark and upon irradiation despite a high singlet oxygen quantum yield. This could be due to the polar cyclometallating ligands and the PEG chain making the compound too hydrophilic to allow for substantial cellular uptake.

Potential Biomedical Applications of Cyclometalated Sensitizers – Imaging and Singlet Oxygen Generation

Chen et al. have attached a cyclometalated Ir compounds to Fe₃O₄/SiO₂ core/shell nanoparticles(90) and CdSe/ZnS quantum dots (Complex 25, Scheme 6) (57). This allows for both luminescence imaging using the core and singlet oxygen generation by the attached Ir complexes which makes them potential candidates for use in photodynamic therapy. The Fe₃O₄/SiO₂ core/shell nanoparticles can additionally function as a contrasting agent for MRI. While their singlet oxygen production was not directly evaluated, the cyclometalated Ir/Fe₃O₄/SiO₂ core/shell nanoparticles were indeed able to photoinduce cell death of HeLa cells (90).

Chen et al. have also demonstrated that it is possible to maintain both the imaging capabilities and the photosensitizing ability for a quantum dot-photosensitizer system (57). The aerated emission spectra of the QD-sensitizer system 25 showed a peak at 590 nm. This peak was assigned to the CdSe/ZnS emission due to its similarity in character and position to the known emission of TOPO-capped CdSe/ZnS QDs. In degassed MeOH, however, the peak was overlapped by the phosphorescence emission of the Ir complex (which was almost completely quenched in the presence of oxygen). In aerated MeOH, the CdSe/ZnS emission quantum yield was 0.4 and thus sufficient for imaging. No significant Förster resonance energy transfer (FRET) was observed from the QD to the Ir complex, perhaps due to the chain length between the QD and the complex. Singlet oxygen was thus produced by direct excitation of the complex. Singlet oxygen production in the Ir-CdSe/ZnS QD was assessed by monitoring the ¹O₂ NIR emission at 1270 nm in aerated MeOH solution. The disappearance of the 1270 nm emission in a degassed sample, and the large solvent isotope dependence of the emission lifetime (25 μs in CH₃OH and ~240 μs in CD₃OD) provided further evidence for singlet oxygen generation. Under aerobic conditions, the QD-cyclometalated complex hybrid material 25 can thus be used for imaging from the QD emission and singlet oxygen generation by direct excitation of the Ir complex (57).

Aoki and co-workers have synthesized an Ir complex with pH-dependent luminescence and ¹O₂ generation (Scheme 4) (53). This allows the complex to be used both for selective imaging as well as singlet oxygen production and induced cell death. The compound, fac-Ir(2-(5′-N,N-diethylamino-4′-tolyl)pryidine)₃, (fac-Ir(deatpy)₃, 18) contains three amino groups which become protonated under acidic conditions to yield [(fac-Ir(deatpy)₃-H₃]³⁺, 19). Upon protonation of the amino groups, the emission shifts from orange (554 nm) to bright green (494 nm). Both the blue shift and intensity change of the compound begin as the pH is lowered below 7.4 with intensity drastically increasing below pH 7. This change was also found to be reversible. The repeated addition of acid (HCl) or base (DBU (1,8-diazabicyclo[5.4.0]undec-7-ene)) could be used to “turn on/off” the green emission. DFT calculations also showed an increase in the HOMO-LUMO gap upon protonation of 18 to 19 consistent with the observed blue shift. The pH range of this switching suggests applicability to imaging in biological systems and thus staining of live cells (HeLa-S3) was attempted. Clear luminescent images were obtained through flourescent microscopy. The compound was found to luminesce in the lysosomes (an organelle known to have a pH below 5.5) (91). Singlet oxygen production by 18 and 19 was compared to Methylene Blue using 1,3-diphenylisobenzofuran (DPBF) as a ¹O₂ trap. While the deprotonated state of the compound had no effect on DPBF consumption (identical to the blank), the protonated state notably increased DPBF consumption, suggesting the production of ¹O₂. The authors noted that singlet oxygen production appeared to be lower than for Methylene Blue. The authors investigated whether the amount of singlet oxygen produced by 19 was sufficient to cause cell death. HeLa-S3 cells were incubated with 19 and irradiated at 377 nm or 470 nm for 30 minutes. In both cases membrane swelling was noticeable after 5 minutes, continued throughout the 30 minutes and was not seen in control experiments. In particular it was found that the compound was inducing necrosis-like cell death, most likely due to massive breakdown of the lysosome (the site of singlet oxygen production) as evidenced by apoptosis- and necrosis-specific staining (53).

Scheme 4.

CONCLUSIONS

Cyclometalated Ir and Pt complexes represent a relatively new but versatile class of singlet oxygen sensitizers. Singlet oxygen quantum yields are generally very high. In some cases, both energy transfer and electron transfer/back electron transfer may be involved in the generation of ¹O₂. Replacement of ancillary ligands has led to several systems in which the sensitizer is attached to a zeolite, or polystyrene polymer or quantum dot. These systems can be used for oxyfunctionalizations of olefins, and potentially for biomedical applications. Literally hundreds of cyclometalated complexes have been prepared in the last decade. For many of these complexes, singlet oxygen production has not yet been evaluated. Given the generally high singlet quantum yields for those complexes that have actually been investigated, it seems prudent to consider that whenever cyclometalated complexes are irradiated under aerobic conditions, there is a high likelihood of photosensitized generation of singlet oxygen. The attachment of fluorophores to cyclometalated complexes represents a route to singlet oxygen sensitizers with very strong absorption in the visible range which may in turn lead to more efficient oxyfunctionalizations of organic molecules with singlet oxygen.

Scheme 10.

Scheme 11.

Scheme 15.