Organometallic Chemistry within the Structured Environment Provided by the Macrocyclic Cores of Carbaporphyrins and Related Systems

Abstract

The unique environment within the core of carbaporphyrinoid systems provides a platform to explore unusual organometallic chemistry. The ability of these structures to form stable organometallic derivatives was first demonstrated for N-confused porphyrins but many other carbaporphyrin-type systems were subsequently shown to exhibit similar or complementary properties. Metalation commonly occurs with catalytically active transition metal cations and the resulting derivatives exhibit widely different physical, chemical and spectroscopic properties and range from strongly aromatic to nonaromatic and antiaromatic species. Metalation may trigger unusual, highly selective, oxidation reactions. Alkyl group migration has been observed within the cavity of metalated carbaporphyrins, and in some cases ring contraction of the carbocyclic subunit takes place. Over the past thirty years, studies in this area have led to multiple synthetic routes to carbaporphyrinoid ligands and remarkable organometallic chemistry has been reported. An overview of this important area is presented.

1. Introduction

Porphyrins are extraordinarily effective ligands that form coordination complexes to virtually every metal or metalloid element [1]. Although this versatility may be diminished upon core modification, porphyrin analogues still give remarkably diverse metalated derivatives [2]. When one or more of the nitrogen atoms within the porphyrin core are replaced by carbon atoms, the resulting carbaporphyrins [3] commonly form organometallic derivatives with late transition metals, including catalytically important cations such as nickel(II), palladium(II), platinum(II), silver(III) and gold(III) [4]. In these systems, the metal cation is constrained within a highly ordered coordination sphere, and this can lead to unusual reactivity, including selective oxidation reactions. The best known porphyrinoids of this type are the so-called N-confused porphyrins (NCP, 1) [5,6,7,8,9,10,11,12], and these can easily be prepared by a one-pot procedure from pyrrole and benzaldehyde [13,14]. However, many other intriguing carbaporphyrinoid systems such as carbaporphyrins 2 and 3 [15], azuliporphyrins 4 [16], benziporphyrins 5 [17,18], oxybenziporphyrins 6 [19], and tropiporphyrins 7 [20] have been reported (Figure 1) and these exhibit diverse structural and spectroscopic properties, unusual reactivity, and varying degrees of aromatic, nonaromatic and antiaromatic characteristics. Carbaporphyrinoids have attracted widespread interest and have been the subject of a number of reviews [19,20,21,22,23,24,25,26,27,28,29,30]. This article focuses on the formation and reactivity of metalated carbaporphyrinoids that have carbon-metal bonds within 16-atom macrocyclic cavities. Methods used to prepare these fascinating ligands will be briefly discussed and the reactivity of different families of carbaporphyrinoids will be presented. N-Confused porphyrins are included in these discussions but will be covered in less depth as this area has been covered in some detail elsewhere [5,6,7,8,9,10,11,12]. Contracted and expanded systems are also briefly discussed.

2. Synthetic Routes to Carbaporphyrinoid Systems

Although N-confused porphyrins (NCPs) were first reported in 1994 [31,32], speculations on the formation of porphyrin analogues with inverted pyrrole units were made over 50 years earlier. Calvin and coworkers speculated that by-products formed in the Rothemund reaction could correspond to isomers of meso-tetraphenylporphyrin 8 (TPP, Scheme 1) in which one or two of the pyrrole rings have been inverted, although some more bizarre suggestions were also made [33]. Subsequently, the major by-product in the synthesis of TPP from pyrrole and benzaldehyde was shown to be the related chlorin 9 [34], and the reported UV-vis spectra for Calvin’s porphyrin-like fractions do not resemble the spectra later obtained for NCP [31,32]. Pauling also speculated about the existence of this type of porphyrin isomer around the same time, although these musings were not published until they were discovered among his papers in 2011 [35]. These prescient speculations had been all but forgotten when Latos-Grażyński and Furuta, in 1994, independently isolated modest yields of NCPs from reactions of aromatic aldehydes with pyrrole under acid-catalyzed conditions. Lindsey, Geier and coworkers later developed an efficient synthesis of tetraphenyl NCP with yields of up to 40% using methanesulfonic acid as the catalyst (Scheme 2) [13,14]. Early investigation by Latos-Grażyński, Furuta and others demonstrated the propensity of this system to form organometallic derivatives [5,6,7,8].

The ‘3 + 1’ variant of the MacDonald condensation provides an effective route to porphyrin analogues such as carbaporphyrins [36,37]. This approach was first used by Johnson to prepare 21-oxa- (10a), 21-thia- (10b), 21,23-dioxa- (10c), 21,23-dithia- (10d) and 21-oxa-23-thiaporphyrins (10e) (Scheme 3) [38]. The strategy involved reacting tripyrrane 11a and related species 11b,c with furan or thiophene dialdehydes 12 in the presence of HBr. Although this approach gave the first examples of core modified porphyrins, no further applications of the ‘3 + 1’ route were made for nearly 25 years. This was due in part to tripyrranes being relatively inaccessible at that time, although an efficient route to these intermediates was subsequently reported by Sessler in 1987 [39]. The ‘3 + 1’ strategy provided access to previously unknown heteroporphyrins, but thiaporphyrin 10b was isolated as an isomeric mixture [38]. Johnson noted that even at 60 MHz, proton NMR spectroscopy showed that additional peaks were present, and this was attributed to “traces of other isomers formed by cleavage recombination reactions” [38]. Recently, syntheses of oxa-, thia- and selenaporphyrins from tripyrranes 13 and 14 and heterocyclic dialdehydes 12a–c were reported using trifluoroacetic acid (TFA) as the catalyst, and this methodology afforded pure heteroporphyrins 15a–c and N-methylheteroporphyrins 16a–c that were free from isomeric impurities (Scheme 3) [40]. In the mid-1990’s, the ‘3 + 1’ route was applied to the synthesis of porphyrins and b-annulated porphyrins [36,41,42,43,44,45,46,47]. Using trifluoroacetic acid as a catalyst, isomerically pure porphyrin products were generally obtained [46], although exceptions have been noted [48]. In addition, two groups independently utilized this approach to prepare carbaporphyrinoid systems. Berlin and Breitmaier prepared a benziporphyrin 17 by reacting isophthalaldehyde with tripyrrane 13 in the presence of HBr in acetic acid [49], while Lash reported the synthesis of oxybenziporphyrin 18 [19] by condensing 4-formylsalicylaldehyde with 13 in the presence of TFA, followed by oxidation with 2,3-dichloro-5,6-dicyano-1,4-benzoquinone (DDQ) (Scheme 3). Benziporphyrin 17 was mistakenly reported to have multiple tautomeric forms that failed to interconvert on the NMR timescale [49], but the results were subsequently shown to be due to the presence of isomeric impurities [15,50,51]. This problem persisted in early attempts to prepare carbaporphyrins using the HBr-catalyzed conditions [52,53]. However, pure carbaporphyrins were obtained when TFA was used as a catalyst, and these conditions have allowed the synthesis of structurally diverse carbaporphyrinoid structures (Scheme 4) [15,19,40,50,54,55,56,57,58,59,60,61,62,63,64,65,66,67,68,69,70,71,72,73,74,75]. Tripyrrane analogues have also been used to prepare porphyrin analogues, including heterotripyrranes [76,77], azulitripyrranes [61,78,79,80], benzitripyrranes [81,82,83,84], and pyritripyrranes [85] (Scheme 5). Although less commonly used, MacDonald ‘2 + 2’ syntheses of carbaporphyrinoids have been reported, including meso-unsubstituted NCP 19 [86], azuliporphyrin 20 [87], and neo-confused porphyrins such as 21 and 22 [88,89,90] (Scheme 6).

One pot syntheses of meso-tetraarylbenziporphyrins [91,92] and azuliporphyrins [93,94,95] have also been reported (Scheme 7). Benzenedicarbinols 23 react with pyrrole and aromatic aldehydes in the presence of BF₃.Et₂O to give, following an oxidation step, benziporphyrins 24 [91]. This approach was also used to prepare p-benziporphyrins 25 from 1,4-benzenedicarbinols, albeit in low yield [96]. Azulene 26 favors electrophilic substitution at the 1,3-positions, which are analogous to the α-positions of pyrroles, and this characteristic has been utilized in the preparation of calix[4]azulenes 27 [97,98], azulitripyrranes 28 [78], and meso-tetraarylazuliporphyrins 29 [93,94,95] (Scheme 7). Reaction of azulene or 6-substituted azulenes, with three equivalents of pyrrole and four equivalents of an aryl aldehyde in the presence of BF₃.Et₂O in chloroform, followed by oxidation with DDQ, gave azuliporphyrins 29 in up to 20% yield [93,94,95]. Given that the reaction requires the selective formation of eight covalent bonds between a 1:3:4 mixture of three different reagents, the outcome of the chemistry is remarkable.

Stepwise routes to meso-substituted carbaporphyrinoids are also known (Scheme 8). For example, benzenedicarbinols 23 react with excess pyrrole and BF₃.Et₂O to afford benzitripyrranes 30 and these condense with heterocyclic dicarbinols 31 to afford a series of heterobenziporphyrins 32⁸³ Dimethoxythiabenziporphyrins 33⁸⁴ and inverted pyriporphyrins 34 [85,99], where a pyridine subunit has been incorporated with the nitrogen orientated towards the periphery of the macrocycle (N-confused pyriporphyrins), were prepared similarly. In an innovative application of this strategy, a ferrocene-embedded tripyrrane analogue 35 was used to generate tetraphenylcarbaporphyrin 36 [100]. The ferrocene unit acts as a protected cyclopentadienyl moiety and spontaneously demetalates to give the macrocyclic product. Tripyrrane analogues have been widely applied to the synthesis of expanded porphyrinoid systems [101].

An alternative route to carbaporphyrins and their heteroanalogues from carbatripyrrins has been developed (Scheme 9) [102]. Carbatripyrrin 37a and oxacarbatripyrrin 37b can be prepared in three steps from technical grade indene. Condensation with pyrrole and furan dialdehydes gave moderate yields of macrocyclic products 38 [102,103] and related carbaporphyrins with fused phenanthrene [102], acenaphthylene, pyrene and chrysene units [104] were also obtained. In addition, dioxacarbaporphyrin 39 was generated when dioxocarbatripyrrin 40 was reacted with pyrroledicarbaldehyde 41 [105]. Unfortunately, many of carbaporphyrins prepared by this strategy are poorly soluble due to the absence of substituents. This problem can be overcome by reacting 37a and 37b with furan, thiophene, selenophene and tellurophene dicarbinols 31a–d in the presence of boron trifluoride etherate, followed by oxidation with DDQ, to give a series of relatively soluble diphenyl-heterocarbaporphyrins 42a–h, including the first examples of porphyrin analogues with four different elements within the macrocyclic core [102,103,106]. Telluracarbaporphyrin 42d proved to be prone to air oxidation and afforded the hydroxytellurophene derivative 43.

A cyclopentadiene analogue of the tripyrranes 44 similarly reacted with a thiophene dicarbinol to give low yields of heterocarbachlorin 45a (Scheme 10) [107,108]. Oxidation with DDQ afforded the related thiacarbaporphyrin 46a in 25% yield together with the quinone addition product 47. Very recently, related oxacarbachlorins 45b and oxacarbaporphyrins 46b were prepared in a similar fashion [109].

Finally, it may also be possible to convert specific carbaporphyrin-like systems into other classes of carbaporphyrinoids. The best-known strategy of this type involves oxidative ring contraction of azuliporphyrins 4 to give benzocarbaporphyrins (Scheme 11) [110]. Reaction of azuliporphyrin 4 with tert-butyl hydroperoxide in the presence of base initiates nuclophilic attack from a peroxide anion, and subsequent Cope rearrangement and elimination affords mixtures of benzocarbaporphyrins 48 and related aldehydes 49a,b [61]. Tetrarylazuliporphyrins 29 similarly give the related meso-substituted benzocarbaporphyrins 50 [93,94,95]. The same strategy has been applied to the synthesis of carbachlorins 51 from azulichlorins 52 [111] and carbatriphyrin(1.2.1)s 53 from azulitriphyrins 54 [112]. Ring contraction reactions triggered by metalation are discussed in later sections of this review.

3. Organometallic Chemistry of N-Confused Porphyrins

N-Confused porphyrins are particularly proficient at forming organometallic derivatives and can act as dianionic or trianionic ligands (Figure 2). As the confused nitrogen, i.e., the external nitrogen atom, bears a proton in tautomer 1B, replacement of the two internal protons facilitates the incorporation of transition metal dications. However, replacement of all three protons in tautomer 1A would provide a suitable environment for trications. Although tautomer 1A is slightly more stable than 1B and is favored in nonpolar solvents, 1B is sufficiently close in energy to be accessible and in fact it is the primary species in polar aprotic solvents such as DMF and DMSO [113]. The earliest report on the formation of an organometallic derivative for NCP 1 (Ar = p-tolyl) involved reaction with nickel(II) chloride to give a nickel(II) complex 55 in which one of the NH protons had been relocated onto the external nitrogen (Scheme 12) [32]. As is the case for tautomer 1B, this structure is cross-conjugated and exhibits greatly reduced aromatic characteristics. Reaction of 55 with methyl iodide gave the C-methylated nickel(II) complex 56 together with a dialkylated nickel(III) species 57 (Scheme 12) [114]. Complex 56 can be considered to be derived from an dianionic ligand corresponding to tautomer 1c (Figure 2). Metal complexes formally derived from tautomer 1B have been reported for Pd(II) [115], Pt(II) [116,117,118], Cu(II) [119,120], Mn(III) [121], Co(II) [122], Rh(IV) [123,124] and Mo(II) [125]. Reaction of 3-ethoxy-NCP 58 with NiCl₂ gave unstable nickel(II) complex 59 and this gradually air oxidized to give oxygen bridged nickel(III) complex 60 together with nickel(II) carbaporpholactam 61 [126]. When intermediate 59 was oxidized with (bis(trifluoroacetoxy)iodobenzene (PIFA) in chloroform-ethanol, nickel(III) 3,21-diethoxyNCP 62 was generated.

Copper(II) and copper(III) complexes of NCP have been reported (Scheme 13) [119,120]. Reaction of NCP 1 or 21-methylNCP 63 with copper(II) acetate afforded copper(II) complexes 64a and 64b, respectively. Tetrakis(pentafluorophenyl) NCP 1c also generated copper(II) complex 64c but this was readily converted into the corresponding copper(III) complex 65 upon treatment with 1.5 equivalents of DDQ [120]. However, this species is somewhat unstable and solutions in CHCl₃ or CH₂Cl₂, 65 gradually converted back into 64c.

Tetraphenyl NCP 1a reacted with silver(I) trifluoroacetate to give silver(III) complex 65 (Scheme 14) [127]. NCP acts as a trianionic ligand in this case and the silver(I) cation is transformed into the Ag(III) complex. This is believed to occur via a disproportionation reaction: 3 Ag⁺ → 2 Ag³⁺ + 2 Ag° [128]. The macrocycle retains a porphyrin-like 18π electron delocalization pathway and 65 exhibits strongly diatropic properties. The related gold(III) complex 66 cannot be obtained directly from 1, but instead it is necessary to initially carry out a monobromination with N-bromosuccinimide to form 21-bromoNCP 67 and subsequent reaction with 3.3 equiv of AuCl.SMe₂ then gives 66 [129]. Gold(III) complex 66 exhibited unique luminescent properties at room temperature. Silver(III) NCPs have modified reactivity that enables unusual structural transformations to occur (Scheme 15). For instance, reaction of 65a with dimethylamine results in oxidative addition to afford 21-dimethyl NCP 66 [130]. Oxidation with one equivalent of DDQ produced internally bridged NCPs 67a and 67b. Reaction of 1a with potassium diphenylphosphide afforded 21-diphenylphosphanyl-NCP 68 (Scheme 15) [131]. Oxidation of 68 with DDQ generated the related diphenylphosphoryl-NCP 69 but attempts to metalate this compound with silver(I) acetate led to elimination of the phosphoryl unit and conversion back to silver(III) NCP 65a. Thiophosphorylation was also observed when 68 was reacted with elemental sulfur (S₈)¹³¹. Another intriguing transformation occurs when silver(III) NCP 65a is treated with lithium hydroxide and methyl or ethyl iodide [132]. Under these conditions, cleavage of the confused ring, together with demetalation, occurs to produce a novel ethynyl-linked triphyrin 70. This system, named porphyriyne, has a porphyrin-like UV-vis spectrum with a Soret band at 418 nm, and the proton NMR spectrum demonstrates the presence of a strong diamagnetic ring current [132].

As noted above, NCPs can act as dianionic or trianionic ligands and afford metal complexes A–C (Figure 3) formally related to tautomers 1A–C (Figure 2), respectively. NCPs give rise to diverse coordination complexes. In addition to silver(III) and gold(III) complexes, type B complexes include Co(III) [122], Rh(III) [123,124], and Sb(V) [133] derivatives. Many examples of type D complexes, which lack direct carbon-metal bonds, have been reported including zinc [134], manganese(II) and iron(II) derivatives [135,136,137]. The manganese(II) and iron(II) derivatives react with molecular oxygen to produce internally oxo-bridged complexes of type E. NCPs can also coordinate at the external nitrogen and this facilitates the formation of diverse structures. For instance when NCP 1b was reacted with palladium(II) acetate in refluxing toluene, aryl-bridged dimers 71a and 71b were formed, together with the organometallic palladium(II) complex 72 (Scheme 16) [115]. The remarkable coordination chemistry of NCPs has been widely reviewed elsewhere and more detailed descriptions fall outside of the scope of the current review.

Although less well studied, metalated derivatives of meso-unsubstituted NCPs have also been investigated [56,62]. Much of the initial work on related carbaporphyrinoid systems was carried out on meso-unsubstituted structures and for this reason complexes of meso-unsubstituted NCPs allow valuable comparisons to be made. NCP 73 reacted with nickel(II) acetate in DMF at 145 °C to give the corresponding nickel(II) complex 74 (Scheme 17) [56]. This complex exhibited greatly reduced aromatic character, but when TFA was added to a solution of 74 in CDCl₃ a strong diamagnetic ring current was generated. An internal CH resonance was observed at −4.93 ppm, while the external meso-protons gave rise to four singlets between 9.37 and 10.01 ppm. The new species was identified as a C-protonated nickel complex 74 with an 18π electron delocalization pathway. Protonation is reversible, but cation 74 slowly demetalated in the presence of TFA to give protonated 73. Similar C-protonation was subsequently reported for copper(II) and nickel(II) complexes of tetraphenyl-NCP [119]. N-Methyl and N-phenyl NCPs 75 also gave nickel(II) and palladium(II) complexes, 76 and 77, respectively, and addition of TFA to solutions of these organometallic derivatives similarly afforded aromatic C-protonated species 76H⁺ and 77H⁺, respectively (Scheme 18) [62]. Higher concentrations of TFA were required to protonate palladium complexes 77 compared to the 76, but the resulting palladium cations 77H⁺ proved to be far more stable under acidic conditions [62]. NCPs 75 reacted with silver(I) acetate to give silver(III) carbaporpholactams 78 and these proved to be highly diatropic compounds (Scheme 18). The proton NMR spectrum for a solution of 78a in CDCl₃ showed the meso-protons as four downfield 1H singlets between 9.10 and 9.86 ppm. NCP 75a also reacted with gold(III) acetate to give a low yield of the related gold(III) complex 79 [62].

4. X-Confused Heteroporphyrins

O-confused oxaporphyrins and S-confused thiaporphyrins, collectively known as X-confused heteroporphyrins, have similar structures to NCPs but possess inverted furan or thiophene units in place of the confused pyrrole moiety (Figure 4) [138]. Although X-confused heteroporphyrins are cross-conjugated and only weakly aromatic, dihydro-O-confused oxaporphyrins are chlorin analogues that possess macrocyclic aromaticity due to the presence of 18π electron delocalization pathways. Pyrrole-appended O-confused porphyrinoid 80 reacted with nickel(II) chloride or palladium(II) chloride in the presence of anhydrous potassium carbonate to afford the corresponding metal complexes 81a and 81b, respectively (Scheme 19) [139]. O-Confused porphyrinoid 80 exhibits macrocyclic aromaticity and proton NMR spectroscopy shows that it possesses a strong diatropic ring current. However, nickel(II) and palladium(II) complexes 81 have substantially reduced diatropicity due to the furan unit introducing a cross-conjugated element. When 80 was reacted with silver(I) acetate in acetonitrile, a fully aromatic silver(III) complex 82a was formed but when ethanol was added to the reaction mixture, the related ethoxy-derivative 82b was generated [139]. Addition of TFA facilitated elimination of ethanol to give the highly diatropic cation 83. The aromatic properties exhibited by this species can be attributed to electron-donation from the pyrrole substituent (resonance structure 83′). Porphyrinoid 80 also reacted copper(II) acetate in refluxing THF to give copper(III) complex 84 in quantitative yield (Scheme 19) [140]. This organometallic complex also gave a proton NMR spectrum that was consistent with an aromatic macrocycle, although the downfield shifts to the external protons were reduced compared to silver(III) complex 82a. In the presence of oxygen, 84 was initially converted into copper(II) complex 85 but further exposure to O₂ resulted in oxidative cleavage to give tripyrrinone complex 86 (Scheme 19). Addition of bromine to 85 generated an aromatic cation 87 that was analogous to silver(III) cation 83. When copper(III) complex 84 was treated with hydrogen peroxide in the presence of KOH, an oxygen atom was inserted into the macrocyclic core to give 88, and this could be demetalated with hydrochloric acid to afford hydroxyporphyrinoid 89.

Ethoxy-O-confused porphyrinoid 90 reacted with silver(I) acetate to give the silver(III) organometallic complex 91, and subsequent addition of TFA led to elimination of ethanol to afford cationic complex 92 (Scheme 20) [141]. This species was unstable and slowly converted into the carbaporpholactone 93, most likely due to nucleophilic attack from water followed by oxidation. As silver(I) is lost during this process, silver(III) complex 93 may be responsible for the observed oxidation. Porphyrinoid lactone 93 could be remetalated with silver(I) acetate to give the silver(III) complex 94. Reaction of 94 with methylamine or dimethylamine resulted in demetalation and the formation of amino-derivatives 95 [130], while treatment with sodium diphenylphosphide afforded the phosphine derivative 96 [142]. This derivative was oxidized with DDQ to produce the corresponding phosphonate 97, and further reaction with copper(II) acetate in the presence of air afforded a nonaromatic copper(II) complex 98. However, treatment of 97 with silver(I) acetate led to loss of the phosphonyl group and regeneration of the silver(III) complex 94 [142].

S-Confused thiaporphyrin 99 acts as a monoanionic ligand when reacted with cadmium(II) chloride in chloroform or zinc(II) chloride in THF to give coordination complexes 100a and 100b possessing axial chlorides (Scheme 21) [143]. Coupling between the protons and carbon-13 nuclei of the thiophene unit and the NMR active cadmium isotopes (¹¹¹Cd and ¹¹³Cd) suggest that there is a strong agostic interaction between the metal and thiophene units despite the absence of a formal carbon-metal bond. This interpretation is supported by the X-ray crystallographic data. Organometallic nickel(II) and palladium(II) complexes 101 were also obtained from 99, demonstrating that S-confused thioporphyrins can also act as dianionic ligands [25].

5. Organometallic Chemistry of True Carbaporphyrins

Carbaporphyrins retain the porphyrin framework but replace one of the nitrogens with a carbon atom. In early studies, the term “true carbaporphyrins” was introduced [22,144] to differentiate structures such as 102–110 (Figure 5) [15,58,67,71,72,102,104] from other carbaporphyrinoid systems including N-confused porphyrins and azuliporphyrins. This definition includes ring fused structures such as 103–110 in much the same way as benzoporphyrin would be considered to be a “true porphyrin” [145,146]. Much later, another author considered only 102 to be a true carbaporphyrin [108] but given that the original definition predates this by at least a dozen years, we suggest that our definition is more appropriate. In any case, the reactivity, aromaticity, and spectroscopic properties of carbaporphyrins are no more affected by ring fusion of this type than they are by introducing electron-withdrawing substituents [69]. Early investigations into the metalation of carbaporphyrins were performed on benzocarbaporphyrins 111 (Scheme 22) due in part to the relative accessibility of these porphyrinoids. Initially, attempts were made to react first row transition metal cations, including Ni(II), Cu(II) and Co(II), were unsuccessful, although 111 was shown to undergo a regioselective oxidation in the presence of 500 equivalents of iron(III) chloride in alcohol solvents to give ketal derivatives 112 (Scheme 22) [147,148]. Ketals 112 gave intense long wavelength absorptions and proved to be effective agents in the treatment of leishmaniasis [149,150]. Subsequently, reaction of benzocarbaporphyrins at room temperature with silver(I) acetate generated silver(III) complexes 113 in excellent yields [144,151]. Silver(III) carbaporphyrin complexes retain highly diatropic characteristics and the proton NMR spectra for 112a showed the resonances for the meso-protons downfield near 10 ppm. The UV-vis spectra for these stable organometallic derivatives were also porphyrin-like and gave a strong Soret band at 437 nm [144,151]. The X-ray crystal structure for benzocarbaporphyrin 111b showed that the indene unit was tilted by approximately 15° relative to the mean macrocyclic plane, but when the silver(III) cation replaces the three inner hydrogens complex 112b takes on a near planar conformation. meso-Unsubstututed benzocarbaporphyrin 111b also reacted with gold(III) acetate to give low yields of the corresponding gold(III) complex 113 (Scheme 22) [151]. Reaction of 111a with [Rh(CO)₂Cl]₂ generated rhodium(I) complex 114 and this underwent oxidation in refluxing pyridine to afford rhodium(III) carbaporphyrin complex 115a [152]. Prior to this study, rhodium(III) N-confused porphyrins had been prepared in the same way [124]. When 111a was heated with [Ir(COD)Cl]₂, a closely related iridium(III) complex 115b was generated [152]. Interestingly, iridium(III) complexes of NCPs are not currently known.

Related carbaporphyrins undergo similar metalation reactions (Scheme 23) [153]. Naphthocarbaporphyrin 117 reacted with silver(I) acetate to give 118 [71], while treatment with [Rh(CO)₂Cl]₂ afforded rhodium(I) complex 119 [154]. As seen for the benzocarbaporphyrin series, heating 119 with pyridine afforded the corresponding rhodium(III) complex 120. Carbaporphyrin diester 121 reacted with AgOAc to produce silver(III) complex 122a [69], while metalation with Au(OAc)₃ afforded an excellent yield of gold(III) complex 122b [69]. The electron-withdrawing ester moieties appear to stabilize the macrocycle and this inhibits oxidation reactions. Reaction of 121 with [Rh(CO)₂Cl]₂ gave rhodium(I) complex 123 and this was converted into the related rhodium(III) derivative 124 in refluxing pyridine [152]. Carbachlorin 125, which is protected from oxidation on the carbocyclic ring by the presence of a gem-diester unit, formed silver(III) complex 126 with AgOAc [69]. Carbachlorins 125 and 127 also gave rhodium(I) complexes 128 and 129, respectively, with [Rh(CO)₂Cl]₂ [154]. Although the NMR spectra showed that single products had been formed, in both cases two different structures 128a,b and 129a,b were consistent with the available data (Scheme 24). X-ray crystal structures could not be obtained, and it was not possible to determine which specific isomers had been formed. Finally, carbachlorin 130 was shown to react with 3.5 equiv of AgOAc to give silver(III) carbachlorin 131. However, the reaction occurred relatively slowly compared to carbaporphyrins [67]. When a larger excess of AgOAc was used, a low yield of impure carbaporphyrin complex 132 was isolated.

meso-Tetraarylbenzocarbaporphyrins 50 also reacted with silver(I) acetate to give the silver(III) complexes 133 (Scheme 24) [151]. Reaction with gold(III) acetate in refluxing pyridine gave much better results for this series, affording gold(III) derivatives 134 in 67–83% yield [151]. The presence of meso-substituents appears to protect the macrocycle from oxidative degradation. Reaction of benzocarbaporphyrins 50 with Re₂(CO)₁₀ and potassium carbonate in refluxing 1,2,4-trichlorobenzene gave oxorhenium(V) complex 135 and oxygen-bridged rhenium(VII) complex 136 [155]. The structures of these complexes were confirmed by X-ray crystallography. The formation of such unusual derivatives indicates that benzocarbaporphyrins may prove to have untapped potential in the formation of organometallic complexes.

Carbaporphyrins generally act as trianionic ligands. In an attempt to convert this system into a dianionic ligand, the introduction of N-alkyl substituents was investigated [156]. Reaction of benzocarbaporphyrin 111a with methyl or ethyl iodide in the presence of potassium carbonate gave a mixture on N- and C-substituted benzocarbaporphyrins 137 and 138, respectively (Scheme 25) [156]. The major products 137 were alkylated at position 22; no trace of alkylation at the 23-position to give 139 was observed. N-Alkyl carbaporphyrins 137 were heated with palladium(II) acetate in acetonitrile with the expectation that palladium(II) complexes 140 would be generated (Scheme 25). However, the metalation reaction occurred with concomitant alkyl group migration to give C-alkyl palladium(II) complexes 141. When the reaction was stopped after a few minutes, complexes 140a,b were observed but attempts to purify these derivatives by column chromatography were unsuccessful as partial conversion to 141 always took place. It was suggested that alkyl group migration could involve a [1,5] sigmatropic rearrangement [156], but a stepwise mechanism involving a transient Pd-alkyl species is now favored [157]. Palladium(II) complexes 141 retain strongly diatropic characteristics. In the proton NMR spectrum for 141a, the meso-protons gave two downfield 2H singlets at 9.56 and 10.27 ppm, while the internal methyl group afforded an upfield 3H resonance at −3.21 ppm [156]. In order to further examine this chemistry, 23-methylbenzocarbaporphyrin 139a was prepared from an N-methyltripyrrane [157]. Reaction of 139a with Pd(OAc)₂ in refluxing acetonitrile gave an N-methyl complex 142 that could be isolated and fully characterized. When the reaction mixture was heated under reflux for 16 h, the rearranged C-methyl derivative 141a was generated. N-Methyl-carbaporphyrin aldehyde 143 similarly reacted with Pd(OAc)₂ to give palladium(II) N-methyl complex 144 [40]. This compound could be isolated and characterized, but longer reaction times gave C-methyl derivative 145. Alkylation of carbaporphyrin diester 121 with methyl iodide and potassium carbonate in refluxing acetone afforded 21-methylcarbaporphyrin 146 [69]. Reaction with Pd(OAc)₂ generated palladium(II) complex 147, while treatment with Ni(OAc)₂ gave nickel(II) complex 148 [69]. N-Methyl carbachlorin 149 was also reacted with Pd(OAc)₂ and gave a low yield of the rearranged palladium(II) complex 150 [67]. It was proposed that this conversion involves a sequential metalation-oxidation-rearrangement process and yields were substantially improved when the oxidant FeCl₃ was present. Attempts to isolate the intermediary carbachlorin complex 151 were unsuccessful. Reactions of 22-, 21- and 23-methylcarbaporphyrins 137a, 138a and 139a with di-μ-chlorotetracarbonyldirhodium(I) were also investigated [157]. Conventional rhodium(I) complexes 152 were obtained for 137a and 138a, but the presence of a 23-methyl group in 139a blocks the formation of this type of structure. However, when 139a was heated with [Rh(CO)₂Cl]₂ in toluene, an usual rhodium(III) complex 153 was isolated in 31% yield. The identity of this structure was confirmed by X-ray crystallography. In this case, the methyl group has again migrated onto the internal carbon atom but is converted into a bridging methylene unit. Hence, alkyl group migration is not limited to palladium(II) carbaporphyrins. In the proton NMR spectrum for 153, the methylene bridge gave rise to a broadened doublet at −3.22 ppm while the meso-protons appeared downfield as two 2H singlets at 9.51 and 10.13 ppm. These data demonstrate that the macrocycle retains strongly aromatic properties and also shows that ¹⁰³Rh (100% natural abundance, I = ½) is coupling to the methylene unit.

Directly reacting benzocarbaporphyrins 111 with palladium(II) acetate or palladium(II) chloride primarily led to decomposition. However, tetraphenylcarbaporphyrin 36 has been shown to react with PdCl₂ to generate palladium complex 154 (Scheme 26) [100,158].

Carbaporphyrins favor tautomers such as 155 and 155B that have three hydrogens within the macrocyclic cavity, two of which are attached to nitrogen atoms. The aromatic character associated with carbaporphyrins can be attributed, at least in part, to the presence of the 18π electron circuit shown in bold for these structures (Figure 6). Less favored tautomers 155′ and 155B′ can be considered that possess internal methylene units, although these have not been observed experimentally. While these still have 18π electron delocalization pathways, benzocarbaporphyrin tautomer 155B′ can also introduce a 22π electron circuit that incorporates the fused benzo-unit. Density functional theory (DFT) calculations indicate that this delocalization pathway is favored [159,160]. Palladium(II) complexes of type 156 and 156B effectively freeze in place the conjugation pathways found in tautomers 155′ and 155B′ and this allows extended aromatic circuits to be probed. In order to further assess how ring fusion modifies the aromatic character of carbaporphyrins, syntheses of naphtho [2,3-b]-21-carbaporphyrin 157 and anthro[2,3-b]-21-carbaporphyrin 158 have been developed (Scheme 27) [71,160].

Naphthocarbaporphyrin 157 was prepared by reacting diformylbenzoindane 159 with tripyrrane 11 (R = Et) in the presence of TFA, followed by oxidation with DDQ (Scheme 27) [71]. Very recently, anthracene-fused carbaporphyrin 158 was synthesized from 11 (R = n-hexyl) and the related dialdehyde 160 [160]. Reaction of 157 and 158 with methyl iodide and potassium carbonate in refluxing acetone gave N-methylcarbaporphyrins 161 and 162, respectively, and subsequent metalation afforded, following an alkyl group migration, palladium(II) complexes 163 and 164 [71,160]. As anticipated, the spectroscopic data indicated that the aromatic conjugation pathway was extended through the fused rings facilitating 26π and 30π electron pathways. This analysis was supported by DFT calculations. The placement of the internal alkyl substituent necessitates a relocation of the π-delocalization pathway and thereby traps the structures in an arrangement that corresponds to tautomers 155N′ and 155A′. The global conjugation pathways all follow Hückel’s rule, but the ¹H NMR spectra indicate that the aromatic ring currents decrease as the size of the delocalization pathways increase (

Table 1. Selected proton NMR chemical shifts (ppm) for palladium(II) carbaporphyrins.

	150	141	163	164
5,20-H	10.42	10.27	9.85	9.54
10,15-H	10.00	9.56	9.13	8.84
7,18-Me	3.49	3.33	3.16	3.05
21-Me	−4.46	−3.21	−2.18	−1.45

). When considering palladium complexes 150, 141, 163 and 164, which correspond to the series shown in Figure 6, 156, 156B, 156N and 156A, the degree of deshielding to the external protons and shielding to the internal methyl groups decreases as the extent of π-conjugation increases. For instance, the resonance for the internal methyl substituent shifts downfield from −4.46 to −1.45 ppm as the size of the aromatic circuit increases, while the external meso-protons move upfield from values of 10.00 and 10.42 ppm in 150 to 8.84 and 9.54 ppm in 164. The extended conjugation also leads to substantial bathochromic shifts in the electronic absorption spectra. The longest wavelength absorption for benzo-complex 141a appears at 697 nm but this shifts to 772 nm in naphtho-derivative 163 and to 841 nm in anthracene-version 164 [71,160]. These insightful observations have been confirmed with NICS calculations and anisotropy of induced current density (AICD) plots [160].

Heterocarbaporphyrins have also been synthesized with furan, thiophene, selenophene and tellurophene rings replacing of pyrrole units [76,77,106,161]. Monoheterocarbaporphyrins act as dianionic ligands [77]. 23-Oxacarbaporphyrin 165 reacted with nickel(II) acetate, palladium(II) acetate or platinum(II) chloride in DMF to give the corresponding organometallic derivatives 166 (Scheme 28), although only low yields of the platinum(II) complex were isolated [76,77]. As expected, these metallo-derivatives retained strongly diatropic characteristics. 23-Thiacarbaporphyrin 167 similarly afforded palladium(II) complex 168, and 22-oxacarbaporphyrin 169 also reacted with palladium(II) acetate to give palladium(II) complex 170 [162]. X-ray diffraction analysis demonstrated that palladium(II) complexes 166b and 170 are both near planar. Addition of TFA to solutions of 168 led to the formation of an aromatic cation 168H⁺ (Scheme 28) [84]. Thiacarbachlorin 171 and thiacarbaporphyrin 172 have also been reported to give palladium(II) complexes 173 and 174, respectively [108].

Diphenyl oxa-, thia-, selena- and telluracarbaporphyrins 175a–d reacted with Pd(OAc)₂ in acetonitrile-chloroform to give a series of palladium(II) complexes 176a–d (Scheme 29) [106,161]. Oxa-, selena- and telluracarbaporphyrin complexes 176a,c,d were characterized by X-ray crystallography. The macrocyclic conformation for oxacarbaporphyrin complex 176a is essentially planar [161]. However, the selenophene ring in 176c is pivoted from the mean macrocyclic plane by 36.0° [161], while the tellurophene unit in 176d is twisted away by 49.2° [106]. Nevertheless, the proton NMR spectra for all four complexes showed that they retained highly diatropic characteristics, although the UV-vis absorptions were broadened and shifted to longer wavelengths as the size of the heteroatoms increased from S to Se to Te. It is remarkable that a metal complex can be formed when a core atom as large as tellurium is present. Oxacarbaporphyrin reacted with nickel(II) acetate in refluxing DMF to afford the corresponding nickel(II) complex 177a [161]. However, good results were only obtained when the reaction was performed under nitrogen. In the presence of air, oxidative demetalation occurred to give 21-oxocarbaporphyrinoid 178a [161]. When pure 177a was heated with DMF, the metalated product was converted into the carbonyl derivative and this demonstrates that oxidation only occurs following the introduction of nickel(II). Thiacarbaporphyrin 175b also reacted with nickel(II) acetate under nitrogen to give nickel complex 177b in 85% yield. In the presence of air, low yields of thiacarbaporphyrin oxidation product 178b were formed. Ketones 178 are tautomers of 21-hydroxyheterocarbaporphyrins 179 but are only weakly aromatic. The system has 20 and 24 π-electron pathways, structures 178^a and 178^b that could result in antiaromatic character, but dipolar contributors such as 178′ provide 18π electron circuits that promote a degree of aromatic character [161]. Recently, oxacarbaporphyrin 180 was reported to undergo air oxidation to form a similar nonaromatic keto-structure 181 [109]. In addition, oxacarbachlorin 182 was oxidized with silver(I) acetate to give a structurally related aromatic derivative 183; in this case, the competing antiaromatic pathways are no longer present [109].

6. Organometallic Chemistry of Azuliporphyrins

Azuliporphyrins 184 (Scheme 30) have substantially reduced diatropic character compared to carbaporphyrins due to the presence of a cross-conjugated azulene unit [16,53]. However, a degree of aromatic character is retained as the proton NMR spectra for meso-unsubstituted azuliporphyrins shows that the inner CH proton has been shifted upfield compared to azulene to approximately +3 ppm. Nevertheless, fully aromatic carbaporphyrins commonly give a resonance for this CH at −7 ppm. The intermediary aromatic properties of azuliporphyrins have been attributed to electron-donation from the seven-membered ring which can take on a degree of tropylium character [53,163]. Unlike carbaporphyrins, azuliporphyrins are dianionic organometallic ligands. Azuliporphyrins 184 have been shown to react with nickel(II) acetate, palladium(II) acetate or platinum(II) chloride to give good yields of the corresponding metal complexes 185a–c (Scheme 30) [164,165]. These stable organometallic derivatives have increased diatropicity and the meso-proton resonances are shifted further downfield than the values observed for free base azuliporphyrins. The largest effects are observed for palladium(II) complexes 185b, which are considered to be the most aromatic macrocycles for this series. The meso-protons for platinum complex 185c showed sidebands due to transannular coupling with ¹⁹⁵Pt (⁴J_Pt,H = 4.4–5.6 Hz). meso-Unsubstituted azuliporphyrins 184 could also be converted into iridium(III) complexes 186 [166] and rhodium(III) complexes 187 [167] (Scheme 30). Reaction of 184 with [IrClCOD]₂ in refluxing o- or p-xylene gave benzoyliridium(III) complexes 186, albeit in relatively low yields, although no reaction was observed in refluxing toluene. A tert-butyl substituted complex gave crystals that were suitable for X-ray diffraction analysis, and this confirmed the presence of an axial aroyl unit. The macrocycle proved to be near planar and the iridium coordination environment had a 5-coordinate square pyramidal geometry. It was suggested that an iridium(III) chloride macrocyclic complex was initially formed and that this reacted with the solvent to form a benzyl iridium(III) species [166]. The proton NMR spectra showed that the benzene protons were strongly shielded by the porphyrinoid system. Further oxidation with molecular oxygen presumably converts the axial ligand into the observed benzoyl units. Reaction of 184 with [Rh(CO)₂Cl]₂ in refluxing o-, m- or p-xylene gave rhodium(III) complexes 187 with axial benzyl ligands. Again, no reaction was observed in toluene, possibly due to the lower boiling point of this solvent. Oxidation of the coordinated methylene units was not observed for the rhodium series. X-ray crystal structures were obtained for the products from all three xylene isomers and these demonstrated that the porphyrinoid macrocycles were planar with the methylbenzyl ligands occupying an orthogonal binding site [167]. The proton NMR spectra showed the coordinated methylene resonances as upfield doublets near −1.9 ppm (²J_RhH = 2.6–3.8 Hz). A related rhodium(III) complex 188 with an axial CH₂C(O)CH₃ unit was obtained when the crude rhodium(III) intermediate 189 was reacted with acetone and basic alumina in toluene. Reaction of azuliporphyrins 184 with copper(II) salts led to an oxidative metalation to form copper(II) complexes 190 [168], possibly via a copper(II) organometallic intermediate 185 (M = Cu). The structure of 190 is nonplanar and the oxyazulenyl unit is pivoted 31.76° relative to the plane described by the core nitrogen atoms. Attempts to form cobalt complexes of 184 by reacting it with CoCl₂.6H₂O or Co₂(CO)₈ were unsuccessful and instead 21-oxyazuliporphyrin 191 was generated [168]. Although reactions of 184 with Cu(II) or cobalt reagents were carried out under nitrogen, trace amounts of molecule oxygen appeared to be responsible for the formation of these oxygenated products. Oxyazuliporphyrin 191 is the keto-tautomer of 21-hydroxyazuliporphyrin. X-ray crystallography conclusively demonstrated the identity of 191 and while a 24π electron pathway is present, proton NMR spectroscopy indicates that the system is essentially nonaromatic. Protonation of this system afforded an aromatic cation. Oxyazuliporphyrin 191 acts as a dianionic ligand and reacted with nickel(II) acetate or palladium(II) acetate to afford metal complexes 192a and 192b [168]. These derivatives are structurally equivalent to copper(II) complexes 190 and the X-ray crystal structures for palladium(II) complex 192b (R = t-Bu) was virtually superimposable with the structure obtained for copper(II) complex 190 (R = t-Bu). Reaction of 184 with silver(I) acetate led to the formation of silver(III) complexes of benzocarbaporphyrins (Scheme 30) [168]. Oxidative ring contraction of azuliporphyrins to benzocarbaporphyrins is well known and results in the formation of structures with unsubstituted benzo-units and related aldehydes [110]. The introduction of a formyl substituent at position 2¹ generally only occurs to a minor extent due to steric effects. Metalation of azuliporphyrin 184 to form a silver(III) complex triggers the ring contraction but the resulting regioselectivity is greatly altered. For 184 (R = H), 2¹-formyl derivative 193b is the major product, albeit in 20% yield, while two other products, 193a and 193c, were each isolated in <4% yield. tert-Butylazuliporphyrin 184 (R = t-Bu) gave 194a as the major product in 31% yield, but a greater than expected yield (12%) of sterically crowded benzocarbaporphyrin aldehyde 194b was also isolated. A mechanism was proposed to explain the observed results (Scheme 31). Formation of a silver(I) complex, followed by complexation of molecular oxygen, would give 195 and a subsequent internal redox reaction would produce silver(III) complex 196 with an axial peroxide ligand. Alternatively, silver(III) azuliporphyrin cation 197 might be formed initially, followed by formation of the peroxide derivative. The location of the axial peroxide unit facilitates nucleophilic attack at the nearby 2¹-position and subsequent Cope rearrangement and elimination of water gives the observed aldehyde product [168]. Examples of heteroazuliporphyrins with furan, thiophene or selenophene subunits have also been synthesized. Metalation of thiaazuliporphyrin 198 with palladium(II) acetate led to a similar ring contraction to form palladium(II) benzocarbaporphyrins 199 (Scheme 30) [168]. In these reactions, formation of crowded 2¹-formyl derivatives was not favored indicating that the process did not occur via the type of directed intramolecular nucleophilic attack proposed for silver derivatives.

It is worth noting that azulene-based pincer ligands have been reported that can bind divalent transition metal cations [169]. Azulene bis-thioamide 200 reacted with palladium chloride and lithium chloride in refluxing methanol to afford organometallic palladium(II) complex 201a (Scheme 32). Reaction of 200 with PtCl₂(PhCN)₂ in acetonitrile generated the related platinum complex 201b. These structures have similar features to metalated azuliporphyrins such as 185.

Metalation reactions for tetraarylazuliporphyrins 202 show some significant differences to the chemistry of meso-unsubstituted azuliporphyrins 184. However, they also react with Ni(OAc)₂, Pd(OAc)₂ and PtCl₂ to give similar organometallic derivatives 203 (Scheme 33) [165]. Furthermore, attempts to metalate 202 with copper(II) salts led to the formation of copper(II) oxyazuliporphyrins 204 [170,171]. X-ray crystallography showed that the structure was highly distorted and the azulene ring was tilted by almost 53° relative to the mean macrocyclic plane [170]. Reactions in the presence of ¹⁸O₂ demonstrated that the oxygen atom derives from molecular oxygen. A recent study demonstrated that copper(II) complex 205 can be isolated under strictly anaerobic conditions using a glove box [172]. Exposure to air then led to the formation of 204. However, even in the absence of O₂, conversion to 204 still takes place via inner core nucleophilic attack from water or hydroxide ions. Demetalation of 204 with 10% TFA-CHCl₃ gave 21-oxyazuliporphyrins [170,171]. As was the case for the meso-unsubstituted series, keto-tautomers 206 were favored over hydroxyazuliporphyrins 207. Metalation of 206 with Ni(OAc)₂, Pd(OAc)₂ or PtCl₂ gave nickel(II), palladium(II) and platinum(II) complexes 208 [171]. X-ray crystallography showed that the conformations of the Pd(II) and Pt(II) complexes were virtually identical to the structure obtained for copper(II) oxyazuliporphyrin 204 [171].

Ruthenium complexes of azuliporphyrins 202 have also been reported (Scheme 34) [173,174]. Reaction of 202 with one equivalent or less of Ru₃(CO)₁₂ gave ruthenium(II) complex 209. The proton NMR spectrum showed the external pyrrolic protons between 7.62 and 7.69 ppm, indicating that the macrocycle has a moderate aromatic ring current. Addition of excess Ru₃(CO)₁₂ led to the formation of cluster complex 210a, and related bimetallic derivatives 210b–d were obtained from nickel(II), palladium(II), and platinum(II) azuliporphyrins 203a–c. In the presence of air, 209 slowly oxidized to give ruthenium 21-oxyazuliporphyrin complex 211. This complex can also be prepared in 33% yield by reacting 21-oxyazuliporphyrin 206 with Ru₃(CO)₁₂. Interestingly, 211 can be converted back into 206 by reacting it with Ru₃(CO)₁₂ in refluxing chlorobenzene. Oxyazuliporphyrin complex 211 readily added a further ligand to form hexacoordinate ruthenium complexes 212, and a structure of this type that incorporated 1-butanol was characterized by X-ray crystallography. In the absence of a suitable ligand, a dimeric complex 213 could be isolated instead. Cyclic voltammetry showed that 209 underwent two reversible one-electron oxidations, and the first oxidation gave rise to an easily accessible radical cation 214. This type of oxidation was also accomplished with DDQ or bromine to give the π-radical species 214a,b (Scheme 34).

In contrast to meso-unsubstituted azuliporphyrins, tetraphenylazuliporphyrin 202 reacted with CoCl₂ or Co(OAc)₂ to give cobalt(II) azuliporphyrin 215 (Scheme 35) [172]. In addition, treatment of 202 with Co₂(CO)₈ afforded a transient π-allyl complex 216 that slowly converted into 215. When 215 was exposed to air, oxidation to cobalt(II) oxyazuliporphyrin 217 was observed. This species exists in equilibrium with dimer 218 but addition of ligands such as pyridine leads to the formation of hexacoordinate cobalt(II) complexes 219. Reaction of oxyazuliporphyrin 206 with cobalt(II) acetate also afforded 217. In pyridine solutions, 215 air oxidized to give cationic cobalt(III) complex 220 and the proton NMR spectrum for this structure indicated that the system had taken on significantly increased diatropicity. This can be attributed to the seven-membered ring taking on tropylium character while facilitating 18π electron delocalization pathways in the porphyrinoid ligand.

Tetraarylthiaazuliporphyrin 221 reacted with palladium(II) chloride to give palladium(II) thiaazuliporphyrin cation 222 (Scheme 36) [175]. The X-ray structure for this species showed that the thiophene ring was deflected from the macrocyclic plane by over 30°. The electron-deficient seven-membered ring underwent ring contractions in the presence of suitable nucleophiles to generate palladium(II) thiacarbaporphyrins 223. The best results were obtained when 222 was heated with palladium(II) acetate in DMF and ring-contraction products 223a and 223b were isolated in 12% and 51% yields, respectively. Thiaazuliporphyrin 221 also reacted with ruthenium reagents in chlorobenzene to give chlorocarbonylruthenium(II) complex 224. Addition of silver(I) trifluoroacetate in dichloromethane produced the related radical cation 225. However, when 221 was reacted with ruthenium reagents without carbonyl ligands, an oxidation reaction took place to give the paramagnetic ruthenium(III) complex 226.

Metalation of internally alkylated azuliporphyrins was investigated (Scheme 37) [176]. These alkyl derivatives proved to be unstable and were isolated as the corresponding hydrochloride salts. 21-Alkylazuliporphyrins 227a.2HCl and 227b.2HCl were reacted with palladium acetate in refluxing chloroform-acetonitrile. A poor yield of palladium(II) complex 185b (<10%) was isolated where the internal alkyl substituents had been lost. This may be due to nucleophilic displacement of the metalloporphyrinoid from the alkyl substituents in intermediate 228. A second aromatic product was noted but could not be identified. The main product from the reaction of 23-methylazuliporphyrin 229.2HCl with Pd(OAc)₂ was also a dealkylated palladium(II) azuliporphyrin complex 185b′ (45% yield) [176]. However, two minor products corresponding to palladium(II) benzocarbaporphyrins 230a,b were isolated in 5% and 2.4% yields, respectively. These products were profoundly modified by a combination of oxidative ring contractions and methyl group migration. It has been reported that palladium can induce the formation of peroxides from molecular oxygen [177] and it was proposed that the first step leading to the formation of 230a,b involves nucleophilic attack from a hydroperoxide ion onto intermediate 231 to give 232 (Scheme 37) [176]. Cope rearrangement would generate a six-membered ring while closing off a cyclopropane unit affording 233. Further elimination of water and CO would generate palladium(II) 23-methyl tert-butylbenzocarbaporphyrin 234a. This would be expected to undergo a methyl group migration to afford the observed product 230a. Alternatively, intermediate 233 could eliminate isobutylene and water to produce 234b and this would further rearrange to give aldehyde 230b.

In an attempt to prepare 6-methoxyazuliporphyrin 235, pyrrole dialdehyde 236 was condensed with azulitripyrrane 237 in the presence of TFA, followed by oxidation with FeCl₃ (Scheme 38) [178]. Unexpectedly, tropone-fused carbaporphyrin 238 was generated instead. Although the UV-vis spectrum for 238 appeared to be a hybrid of the electronic spectra for azuliporphyrins and carbaporphyrins, 238 was fully aromatic and behaved as a trianionic ligand. Specifically, 238 reacted with silver(I) acetate to give silver(III) complex 239.

7. Organometallic Chemistry of Benziporphyrins, Naphthiporphyrins and Related Systems

Benziporphyrins are porphyrin analogues which have a benzene ring that replaces one of the pyrrole units [17,18,179]. Naphthiporphyrins are similar systems incorporating naphthalene rings instead. Benziporphyrins can act as monoanionic or dianionic ligands [17]. meso-Unsubstituted benziporphyrin 240 reacted with nickel(II) acetate in refluxing DMF to give the nickel(II) complex 241a in 42% yield (Scheme 39) [65]. Palladium(II) complex 241b was also obtained in 35% yield by reacting 240 with palladium(II) acetate in refluxing acetonitrile. The proton NMR spectrum for 241a showed the meso-protons as two 2H singlets at 7.16 and 7.48 ppm, while palladium complex 241b gave these resonances at 7.35 and 7.72 ppm, indicating that metallobenziporphyrins have weakly aromatic properties that are enhanced for the palladium(II) complex [65]. The aromatic properties can be attributed to dipolar resonance structures such as 241′ that possess 18π electron pathways. Diphenylbenziporphyrins 242 gave good yields of palladium(II) complexes 243 when treated with palladium(II) acetate in refluxing acetonitrile [83]. Naphthiporphyrin 244 similarly reacted with palladium(II) acetate to generate metal complex 245 [65]. The X-ray crystal structure for this complex demonstrated that the porphyrinoid macrocycle was slightly saddled. An example of a related pyreniporphyrin 246 was prepared by reacting pyrene dialdehyde 247 with tripyrrane 13 in the presence of TFA, followed by oxidation with DDQ [68]. As expected, this benziporphyrin-like structure reacted with palladium(II) acetate to produce the related palladium(II) complex 248 [68].

Most of the organometallic chemistry of benziporphyrins has been carried out using tetra-arylbenziporphyrins 249 [17]. Reaction of 249 with PdCl₂ or Pd(OAc)₂ in refluxing acetonitrile or CHCl₃-CH₃CN gave the corresponding palladium(II) complexes 250 in 49–70% yield [91,92] (Scheme 40). Tetraphenylbenziporphyrin also reacted with platinum(II) chloride in refluxing benzonitrile to afford a 20% yield of platinum complex 251 [91]. Reaction of 249 with nickel(II) chloride initially gave chloronickel(II) complex 252 but this converted into the organometallic complex 253 [180,181]. The conversion of 252 into 253 can accelerated by adding anhydrous potassium carbonate, but 253 is converted back into 252 upon treatment with dry HCl in chloroform. When 249 was treated with silver(I) acetate, a regioselective oxidation occurred to afford the 21-acetoxy-derivative 254 [91], and in a related reaction 249 was shown to react with AgBF₄ in pyridine to give 22-pyridiniumyl derivative 255 (Scheme 40) [182]. It was proposed that a silver(III) benziporphyrin is initially formed, followed by reversible axial coordination of pyridine and reductive elimination of silver(I). When a solution of 255 in CDCl₃ was heated for 12 h, intramolecular cyclization occurred to give phlorin cation 256.

Dimethoxybenziporphyrins 257 and 258 also reacted with silver(I) acetate to give the 21-acetoxy derivatives 259 (Scheme 41) [183,184]. The proton NMR spectra for 259 (R = H) gave the acetate methyl resonance near 1.3 ppm, showing that this unit is shielded by the macrocyclic π-system. Protonation with TFA gave a dicationic species 259H₂²⁺ where the acetate resonance shifted further upfield to 0.5 ppm, while the external pyrrolic protons shifted downfield by 0.5–0.8 ppm, indicating that the macrocycle has taken on a significant diamagnetic ring current due to resonance contributors such as 259′H₂²⁺ [184]. The effects were much reduced when a methyl group was present between the two methoxy substituents because steric crowding prevents the OMe units from lying coplanar with the benzene ring, diminishing electron donation. Reaction of 257 and 258 with nickel(II) acetate in refluxing chloroform-methanol gave the nickel(II) complexes 260 in 74–81% yield, while treatment with palladium(II) acetate in refluxing acetonitrile afforded palladium derivatives 261 in 73–82% yield (Scheme 41) [183,184]. The proton NMR spectra for these compounds were consistent with moderately aromatic structures. For instance, nickel(II) complex 260 (R = H, Ar = Ph), gave the pyrrolic proton resonances comparatively downfield between 7.24 and 7.66 ppm. The effect was slightly larger for palladium(II) complex 261 (R = H, Ar = Ph) and in this case the pyrrole resonances appeared between 7.30 and 7.74 ppm, The X-ray crystal structure of nickel(II) complex 260 (R = H, Ar = Ph) showed that the macrocycle has a highly distorted saddle shaped geometry. The diatropic properties of the metal complexes were substantially reduced when a methyl group was placed between the two methoxy units.

Reaction of benziporphyrins 249 with zinc chloride, cadmium chloride, or mercury salts gave metal complexes 262 (Scheme 42) [180]. Similar nickel(II), zinc, and cadmium derivatives 263 were synthesized from benziporphodimethene 264. Evidence for agostic interactions with the internal C−H bond for complexes 262 was provided by X-ray crystallography. Reaction of benziporphyrin 249 with iron(II) bromide and lutidine in refluxing THF under an inert atmosphere afforded a high- spin iron(II) complex 262d in 71% yield (Scheme 42) [181]. In contrast, under anaerobic conditions, copper(II) chloride reacted with 249 to give a dimeric copper complex 265, where an oxidative chlorination had occurred on the internal carbon atoms [181]. X-ray crystallography showed that the copper(II) benziporphyrin units were connected by a [Cu₂Cl₄]²⁻ cluster. Dimeric silver(I) complexes of benziporphodimethenes have also been prepared [185], and a benziporphodimethene has been reported to be a selective zinc cation fluorescence switch-on sensor [186]. Furthermore, a zinc benziporphodimethene has been used as a building block to construct multidimensional nanostructure arrays [187].

Organometallic complexes for structurally related benzene-containing macrocycles have been reported. Triazamacrocycles 266 reacted with copper(II) salts to give copper(III) organometallic complexes 267 (Scheme 43) [188,189,190]. The reaction involves a disproportionation to form Cu^III and Cu^I. The copper(III) can be displaced by various nucleophiles and methanol reacts to give methoxy derivative 268, while 2-pyridone produces adduct 269. Tetraazacalix[1]arene[3]pyridines 270 exhibit similar reactivity [191,192,193,194]. Treatment of 270 with copper(II) perchlorate gave copper(III) complex 271, and further reaction with a variety of nucleophiles afforded substitution products 272.

Reaction of tetraphenylbenziporphyrin 249 with [Rh(CO)₂Cl]₂ gave a six-coordinate rhodium(III) complex 273 (Scheme 44) [195]. When a solution of 273 in dichloromethane was absorbed onto a silica column and left for 12 h, aromatic rhodium complex 274 was generated. The formyl unit in 274 was shown to be derived from dichloromethane and this transformation involves the intermediacy of 2-chlorovinyl derivative 275. Porphyrinoid 274 possesses a rhodacyclopropane unit and proved to be rather unstable. When 273 or 274 were absorbed onto basic alumina, a mixture of rhodium(III) carbaporphyrins 276a–c was formed in a combined yield of 25%. The proton NMR spectrum for 276a showed the cyclopentadiene protons downfield near 9.4 ppm, while the rhodacyclopropane proton appeared upfield at −3.5 ppm, confirming the highly diatropic nature of this system. These remarkable results demonstrate that the benzene ring in benziporphyrins can undergo a ring contraction to afford a cyclopentadiene unit. Unfortunately, the low yields of 276a–c obtained from 249 makes this approach impractical for synthesizing metallo-carbaporphyrins.

Regular benziporphyrins, sometimes called meta-benziporphyrins, have the same 16-atom core as true carbaporphyrins. An isomeric system, para-benziporphyrin 277, has a slightly expanded core due to the presence of a para-phenylene unit [96]. This system exhibits global aromatic character that has been attributed to 18π electron delocalization pathways shown in bold for resonance contributors 277′ and 277″ (Scheme 45). Nevertheless, the p-phenylene unit is strongly pivoted away from the mean macrocyclic plane and rapidly undergoes a teeter-tottering motion that switches the CH=CH units back and forth between the interior and exterior of the structure. p-Benziporphyrins 277 reacted with cadmium(II), zinc(II), and nickel(II) chlorides to give metal complexes 278a–c [180]. Although these are only coordinated to the three core nitrogen atoms, structural evidence for η²-interactions with the phenylene unit was provided. A related anthriporphyrin 279 reacted with palladium(II) chloride in acetonitrile to yield chloropalladium complex 280. Subsequent treatment with potassium carbonate in the presence of air resulted in oxidative ring opening to form the anthracene-appended tripyrrolic palladium(II) complex 281 [196].

p-Benziporphyrins undergo ring contractions with transition metal cations to give palladium, rhodium and gold carbaporphyrin complexes. Reaction of 277 with PdCl₂ in acetonitrile gave chloropalladium derivative 282 (Scheme 46) [197]. As is the case for other p-benziporphyrin complexes, the metal cation is involved in an η²-interaction with the arene subunit. This interaction freezes the phenylene unit in place, and the proton NMR spectrum showed the arene protons as two 2H singlets at 1.40 and 9.04 ppm, demonstrating that the complex has strongly diatropic characteristics. Treatment of 282 with potassium carbonate resulted in a rearrangement to produce palladium(II) carbaporphyrins 154 and 283 in a 1:3.5 ratio [197]. The reaction was monitored by proton NMR spectroscopy, and initial anti-addition of hydroxide and palladium to the six-membered ring was observed to give of intermediate 284 (Scheme 46). Subsequent β-elimination of H₂ affords ketone 285. Ring contraction involving a 1,2-hydride shift produces 283, while extrusion of CO generates 154. The proton NMR spectrum of formyl derivative 283 gave an upfield resonance for the aldehyde proton at 2.43 ppm, confirming the strongly aromatic properties of this complex. Reduction of 282 with sodium borohydride afforded the cyclohexadiene-palladium complex 286a, while sodium borodeuteride stereoselectively yielded the related deuterated product 286b. Nucleophilic addition of sodium ethoxide to palladium complex 282 gave ethoxy derivative 287, but unlike 284, this complex did not further react to give carbaporphyrin complexes. Structurally related 1,4-naphthiporphyrin 288 similarly reacted with palladium(II) chloride to afford chloropalladium complex 289 [198]. Both 288 and 289 had folded conformations where the naphthalene unit was placed over the porphyrinoid cavity (Scheme 46). When 289 was treated with potassium carbonate, ring contraction to give palladium(II) benzocarbaporphyrin 290a in 21% yield together with small amounts of 290b was observed. Although the palladium(II) cation in 290a is placed in a square-planar coordination environment, X-ray crystallography shows that the macrocycle has a curved geometry.

Reaction of 277 or 288 with sodium tetrachloroaurate also induced ring contractions (Scheme 47) [199]. When p-benziporphyrins 277 were treated with Na[AuCl₄]·2H₂O and potassium carbonate, gold(III) complexes 291 were generated in 10–14% yield. Addition of acid led to a reversible protonation on the internal carbon to give cations 291H⁺, a phenomenon that has also been reported for N-confused porphyrins [56,62,119]. Reaction of 277 with Na[AuCl₄]·2H₂O in methanol afforded nonaromatic dimethoxy derivatives 292 rather than the ring contraction products. 1.4-Naphthiporphyrin 288 also reacted with sodium tetrachloroaurate to give gold(III) benzocarbaporphyrin 134c in 32% yield (Scheme 47) [199]. Similar gold(III) complexes had previously been prepared directly from tetraarylbenzocarbaporphyrins (Scheme 24) [151].

Reaction of di-μ-chlorotetracarbonyldirhodium(I) with p-benziporphyrin 277 in toluene gave rhodium(III) complex 293 (Scheme 48) [200]. X-ray crystallographic analysis of 293 showed that the rhodium cation was close to C21–C22, suggesting the presence of a Rh−η² interaction with the phenylene unit. The proton NMR spectrum showed the resonances for the inner and outer phenylene protons at 0.78 and 8.97 ppm, respectively, and this is consistent with a strong aromatic ring current. Furthermore, in the carbon-13 NMR spectrum the C21-C22 signal was split into a doublet due to coupling with ¹⁰³Rh (I = ¹/₂). When 293 was vigorously stirred under reflux with K₂CO₃ and a small amount of water in p-xylene, ring contraction of the arene unit was observed to give rhodium(III) carbaporphyrin 294 in 46% yield. The complex contained a rhodacyclopropane unit and closely resembles a rhodium(III) complex prepared from a 23-methylcarbaporphyrin [157]. The macrocycle has a fully aromatic 18π-electron delocalization pathway, and the proton NMR spectrum showed the bridged methylene unit upfield at −3.58 ppm, while the external cyclopentadiene protons were shifted downfield to 9.14 ppm. When a solution of 294 in dichloromethane saturated with anhydrous HCl was placed on basic alumina, oxidation to oxycarbaporphyrin complex 295 was observed. The appearance of a carbonyl ligand was unexpected but may have originated from the methylene bridge in 294. Reaction of 295 with NaOMe, followed by addition of 1-pentene, generated a related π-coordinated complex 296. Addition of hydrochloric acid to 294 opened up the the methylene bridge to give C-methyl rhodium(III) carbaporphyrin cation 297. This process was reversible and in the presence of water the cation reverted to 294 [200].

The phenylene unit in p-benziporphyrins undergoes some remarkable rearrangements to give carbaporphyrin organometallic complexes. However, p-benziporphyrins can only be synthesized in low yields and this strategy cannot be used to prepare significant quantities of carbaporphyrin derivatives. An alternative route to gold(III) carbaporphyrins was devised using 22-methylbenziporphyrins 297 [201,202]. Benziporphyrin 297 was obtained in a respectable 23% yield from dicarbinol 298 (Scheme 49) and was used to access novel carbaporphyrin derivatives. Treatment of 297 with Na[AuCl₄].2H₂O in dichloromethane gave a quantitative yield of chemically unstable gold(III) dication 299. Attempts to purify 299 by column chromatography on alumina or silica resulted in the formation of gold(III) carbaporphyrins 300 and 301. Interestingly, ring contractions mediated by alumina favored the formation of 300, whereas silica showed a preference for aldehyde 301. A quantitative yield of 300 and 301 was obtained when 297 was refluxed in benzene with sodium tetrachloroaurate, followed by chromatography on silica. Under basic conditions (triethylamine or potassium carbonate), 301 was converted into cross-conjugated ketone 302. This porphyrinoid retains some aromatic character, presumably due to dipolar canonical forms such as 302′ that retain access to 18π electron delocalization pathways. Protonation with TFA or HCl resulted in the reversible formation of a cationic species 303 that showed greatly enhanced diatropic character. In particular, the proton NMR spectrum for this species gave an upfield resonance for the internal methyl substituent at −3.82 ppm, while the external pyrrolic hydrogens gave downfield peaks between 8.80 and 8.98 ppm. On the basis of deuterium exchange studies, it was proposed that the major protonated species 303 was in equilibrium with ketocarbachlorin 304 [201].

Oxa-, thia-, selena- and tellurabenziporphyrins have been prepared and some metalation studies have been performed on these systems (Scheme 50). Selenabenziporphyrin 305a was reported to react with palladium(II) chloride to give cationic organopalladium complex 306 [203] but the corresponding tellurabenziporphyrin 305b afforded palladium(II) coordination complex 307 [204]. Reaction of dimethoxythiabenziporphyrin 308 initially gave palladium(II) derivative 309 but subsequent displacement of a methyl group afforded the aromatic palladium(II) oxybenziporphyrin derivative 310 [84]. The complex was reversibly protonated by TFA on the carbonyl oxygen to give cation 310H⁺.

Lee and coworkers have investigated the synthesis and reactivity of benziporphyrins with exocyclic double bonds (Scheme 51). Benziporphyrins 311 reacted with palladium(II) chloride in refluxing acetonitrile to give palladium(II) derivatives 312 [205]. Addition of TFA to 312b (Ar = C₆F₅), but not 312a (Ar = p-tolyl), led to cleavage of the carbon-metal bond to form trifluoroacetate complex 313. Reaction of 311b with nickel(II) chloride in acetonitrile gave a similar chloronickel benziporphyrin complex 314, while silver(I) nitrate reacted with 311b to afford, following addition of 2,6-lutidine, silver(I) complex 315. 1,3-Dipolar cycloadditions of benziporphyrin complex 311b with an azomethine ylide derived from N-methylglycine and paraformaldehyde in toluene were investigated [206]. At 80 °C, monoadduct 316 was generated in 37% yield. However, at 90 °C or higher, the bis-adduct 317 was obtained in 39% yield as a diastereomeric mixture. Oxidation of 316 with DDQ afforded pyrroloporphyrinoid 318.

Related systems such as inverted pyriporphyrins and nitrogen bridged porphyrinoids have also been investigated. N-confused pyriporphyrins 319 and 320, where a pyridine ring has been inserted into the porphyrin framework so that the nitrogen faces outwards, are known and may be viewed as azabenziporphyrins (Figure 7). Importantly, the coordination cavities of N-confused pyriporphyrins are essentially the same as those found in benziporphyrins. Reaction of pyriporphyrin 321 with FeBr₂ and collidine in THF gave iron(II) complexes 322 and 323 (Scheme 52) [207,208]. Treatment of 322 with bromine gave the corresponding iron(III) complex, while exposure to oxygen generated five-coordinate iron(III) complex 324. Pyriporphyrin 325 reacted with palladium(II) acetate in refluxing acetonitrile to give palladium(II) complex 326 [85]. Addition of TFA afforded the externally protonated cation 326H⁺. As expected, the proton NMR spectra for 326 and 326H⁺ were consistent with nonaromatic porphyrinoids.

Phthalocyanines are a widely investigated group of porphyrin-like structures that have nitrogen bridges instead of the methine carbons found in porphyrins (Figure 8). Benziphthalocyanines 327 and 328 were discovered 70 years ago [209,210,211,212,213], but their potential to be organometallic ligands was only explored relatively recently. Dibenziphthalocyanine 327, also known as dicarbahemiporphyrazine, reacted with silver(I) acetate to give silver(I) complex 329 (Scheme 53) [214]. Reaction of 327 with copper(II) and copper(I) salts gave more complicated results and afforded a copper(I) complex 330 with a pyridiniumyl group attached to a benzene ring [215]. It was proposed that copper(II)-assisted elimination of hydride resulted in the formation of the Cu(I) center. The chemistry resembles the formation of pyridiniumyl substituted porphyrinoid 255 from the reaction of benziporphyrin with silver(I) acetate and pyridine (Scheme 40) [182]. Attempts to recrystallize 330 with dichloromethane in the presence of air resulted in demetalation to form 331. Dibenziphthalocyanine 327 formed coordination complexes with lithium, manganese, cobalt, and iron [215,216], and also afforded nickel(II) organometallic derivatives [217,218]. Reaction of 327 with Ni(COD)₂ gave nickel(II) complex 332, but upon exposure to molecular oxygen, a nonplanar phenolate complex 333 was generated. Benziphthalocyanine 328 readily formed cobalt and nickel(II) organometallic derivatives (Scheme 54). Reaction of 328 with Co₂(CO)₈ in pyridine, followed by recrystallization under anaerobic conditions, generated pyridine derivative 334, and exposure to air afforded the corresponding six-coordinate cobalt(III) complex 335 [218]. However, when 328 was reacted with cobalt(II) acetate in DMF under aerobic conditions and the product was crystallized from p-xylene-pyridine, partial oxidation produced cobalt(III) complex 336 [218]. Benziphthalocyanine 328 also reacted with Ni(COD)₂ in DMF-methanol to afford nickel(II) complex 337. In order to form a neutral complex with Ni²⁺, it is necessary to transfer two protons onto the bridging nitrogens; relocation of protons onto bridging nitrogens is a common feature for complexes derived from 327 and 328.

8. Oxybenziporphyrins, Oxynaphthiporphyrins and Related Systems

Oxybenziporphyrins 338 are the favored keto-tautomers of 2-hydroxybenziporphyrins 339 and can act as dianionic or a trianionic ligands [19]. Oxybenziporphyrins 338 reacts with silver(I) acetate to give the silver(III) complexes 340a,b (Scheme 55) [65,153]. Closely related oxynaphthiporphyrins 341 reacted in the same way to give silver(III) organometallic derivatives 342. Attempts to prepare gold(III) complexes were far less successful but a 7% yield of 340c was obtained by reacting 338b with gold(III) acetate [65]. These complexes retain the aromatic characteristics associated with oxybenzi- and oxynaphthiporphyrins. When 338a was reacted with one equivalent of palladium(II) chloride in the presence of potassium carbonate, an aromatic anion 343 was generated (Scheme 56) [219]. Although this species might be expected to be a cross-conjugated phenolate anion 343′, the proton NMR spectrum showed the meso-protons downfield between 9.08 and 10.37 ppm, values that are consistent with a strongly diatropic macrocycle incorporating an 18π electron delocalization pathway. Addition of one equivalent of TFA converted 343 to palladium(II) hydroxybenziporphyrin 344. The diatropic character of 344 was much reduced compared because dipolar canonical forms such as 344′ are less favorable. A related platinum complex 345 was prepared by reacting 338 with PtCl₂ and KOAc in mixtures of DMF and acetic acid (Scheme 56) [220]. Anionic palladium(II) complex 343 is an ambident nucleophile that can react on the oxygen or the inner carbon atom [219]. Treatment of 343 with acetic anhydride and pyridine afforded acetate 346a, while reaction with p-toluenesulfonyl chloride gave the related p-toluenesulfonate 346b (Scheme 56). However, reaction of 343 with methyl iodide generated C-methylated product 347a, although treatment with n-butyl iodide afforded a mixture of the C-alkylated derivative 347b and the O-alkylation product 346c (Scheme 56). Unexpectedly, C-alkylation products 347a,b were highly diatropic, possibly due to dipolar resonance contributors such as 347′ [219]. The proton NMR spectrum for 347a gave four downfield 1H singlets for the meso-protons at 9.20, 9.22, 9.25 and 10.40 ppm, while the internal methyl group produced an upfield 3H singlet at −2.00 ppm. Protonation with TFA generated the aromatic cation 347H⁺.

Tetraaryloxybenziporphyrins 348 were synthesized by reacting phenolic dicarbinol 349 with pyrrole and aromatic aldehydes in the presence of boron trifluoride etherate, followed by oxidation with DDQ [221]. Reaction of 348 with silver(I) acetate in pyridine gave the related silver(III) complexes 350, while treatment with gold(III) acetate generated the gold(III) derivatives 351 (Scheme 57). Gold(III) complexes could only be obtained in low yields for meso-unsubstituted oxybenziporphyrins but the presence of meso-substituents protects these porphyrinoids from oxidative degradation and yields of 67–83% were obtained for 351 [221]. Both silver(III) and gold(III) oxybenziporphyrins exhibited strongly diatropic properties, although the aromatic character was slightly enhanced for the gold complexes.

Further oxidized benziporphyrin systems have also been prepared [57,81,82] and diketone 352 acted as a trianionic ligand, reacting with silver(I) acetate at room temperature to give the silver(III) complex 353 in 87% yield (Scheme 58) [81,82]. 24-Methyl-oxybenziporphyrin 354 is a dianionic ligand and gives a fully aromatic palladium(II) complex 355 [40]. Hetero-oxybenziporphyrins have also been prepared and these are also dianionic ligands. Oxa-oxybenziporphyrin 356 reacted with palladium(II) chloride in benzonitrile to give palladium(II) complex 357 (Scheme 58) [222]. Thiaoxybenziporphyrin 358 reacted under milder conditions with palladium(II) acetate in refluxing chloroform-acetonitrile to give a similar palladium(II) complex 359 [84]. Complexes 355, 357 and 359 all retained fully aromatic characteristics.

Phthalocyanine analogues of oxybenziporphyrins have also been prepared [223,224]. Bis-resorcinol containing macrocycles 360 of this type were converted into organopalladium complexes (Scheme 59). Specifically, metalation of 360a and 360b with bis(dibenzylideneacetone)palladium(0) gave palladium(II) complexes 361a and 361b in 51% and 79% yields, respectively [225]. The X-ray structure indicated that they retain the ligand’s bisquinoidal structure, and the palladium(II) complexes appeared to have zwitterionic characteristics.

9. Miscellaneous Monocarbaporphyrinoids

Tropiporphyrins 362 are trianionic ligands and react with silver(I) acetate and DBU in refluxing pyridine to give the silver(III) complexes 363 (Scheme 60) [22]. Although these derivatives are diatropic in character, the macrocycle is quite distorted. A single crystal X-ray diffraction analysis for 363b showed that the tripyrrolic component was somewhat ruffled, but the cycloheptatriene ring was severely twisted. Reaction of tropiporphyrin with palladium(II) acetate primarily led to decomposition [226]. However, when 362a was reacted with palladium(II) acetate in dichloromethane in the presence of potassium carbonate at 5 °C, two benziporphyrin products, 364a,b, were isolated in a combined yield of 19% [226]. Although ring contractions of azuliporphyrins and benziporphyrins had been observed previously, this result was unprecedented. Reaction of 362a with methyl iodide and potassium carbonate in refluxing acetone afforded 24-methyl tropiporphyrin 365. When 365 was reacted with palladium(II) acetate, palladium(II) tropiporphyrin 366 was obtained in 48% yield and rearranged products were not observed [226]. However, it was important to limit the reaction time to 5 min at room temperature to avoid extensive decomposition. In addition, 25-methyltropiporphyrin 367 was converted to the corresponding palladium complex 368 in 43% yield under the same conditions [40]. Mechanisms for the ring contractions were proposed (Scheme 61) [226]. Addition of Pd²⁺ to the seven-membered ring of 362a, possibly involving the initial formation of a π complex, followed by nucleophilic addition of hydroxide, would give 369, and subsequent elimination of palladium(0) will lead to hydroxy-derivative 370. Cope rearrangement and ring opening of the resulting cyclopropane unit will produce dihydrobenziporphyrin aldehyde 371 and subsequent oxidation and metalation would then afford 364b. Alternatively, loss of a proton from 370 gives hydroxycycloheptatriene 372, and following a tautomerization step, tropone 373 will be generated. Cope rearrangement can give rise to cyclopropanone 374, and following extrusion of CO, oxidation and metalation, 364a will be formed.

Carbaporphyrinoids 375 with pyrazole subunits reacted with nickel(II) acetate and palladium(II) acetate to give the organometallic derivatives 376a,b, respectively (Scheme 62) [63,64]. Pyrazoloporphyrins 375 and their metalated derivatives are cross-conjugated and are only weakly aromatic [64]. However, the proton NMR spectra for these structures indicate a slight increase in diatropicity for the metalated structures, particularly for palladium(II) complexes 376b. However, addition of TFA to solution of 376a or 376b gave monocations 376aH⁺ and 376bH⁺ that showed virtually no aromatic character. The weak aromatic properties for free base 375 and metal complexes 376a,b were attributed to dipolar resonance contributors such as 375′, 376a′ and 376b′ that possess 18π electron delocalization pathways, but protonated complexes 376aH⁺ and 376bH⁺ do not favor these canonical forms because it is necessary to place two positive charges next to one another (see structures 376a′H⁺ and 376b′H⁺) [64].

In N-confused porphyrins, the “confused” pyrrole ring has been connected at the 2.4-positions instead of the usual 2.5-positions (Figure 9) [88,89,90]. Neo-confused porphyrins (neo-CPs) are a more recent addition to the porphyrin isomer family in which one of the pyrrole units is connected at the 1.3-positions so that a nitrogen is directly linked to a bridging methine carbon [159,227]. This system has a 17-atom 18π electron delocalization pathway and possesses an internal CH. Benzo-neo-confused porphyrin 377 was shown to react with nickel(II) and palladium(II) acetate in acetonitrile to give the corresponding organometallic derivatives 378a,b, respectively (Scheme 62) [88,90]. Similarly, neo-CP methyl ester 379a gave stable nickel(II) and palladium(II) complexes 380a and 381a [90]. The X-ray crystal structures of 378a,b, 380a and 381a showed that all four complexes are essentially planar. Proton NMR spectroscopy suggests that there is a slight increase in diatropic character for the metal complexes compared to the parent carbaporphyrinoid ligand. However, these structures have reduced aromatic ring currents compared to regular porphyrins, carbaporphyrins or N-confused porphyrins. Neo-CPs 379b,c with phenyl or bromo-substituents instead of an electron-withdrawing ester moiety were also prepared but these were somewhat unstable [228]. However, phenyl neo-CP 379b could also be converted into the related nickel(II) and palladium(II) complexes, 380b and 381b. Another recent addition to carbaporphyrinoid systems are quinoliziniporphyrins 382 [75]. This system has intermediary global aromatic character and the upfield shift of the internal proton resonance to between 3.0 and 3.5 ppm in their proton NMR spectra is similar to the results obtained for azuliporphyrins 184. The UV-vis spectrum for 382 is also surprisingly similar to spectra obtained for 184 as well. The aromatic character associated with 382 can be ascribed to dipolar resonance contributors such as 382′ or hybrid species 382^h with an 18π electron circuit due to the presence of an anionic [17]annulene substructure. Oxyquinoliziniporphyrins 382 are dianionic ligands and reacted with nickel(II) and palladium(II) acetate to give the related metalated derivatives 383a and 383b, respectively [75].

Ring-fused thiabenziporphyrins 384 were prepared by condensing thienylnaphthalene dicarbinols 385 with dipyrrylmethane 386 in the presence of boron trifluoride etherate, followed by oxidation with DDQ (Scheme 63) [229]. The resulting “meso-fused carbaporphyrins” 384 were isolated in 10–15% yield. This system has 18π- and 22π-electron pathways (384A–C), which give meso-fused carbaporphyrins moderate diatropic character. The proton NMR spectra for 384 showed that the external naphthalene, pyrrole, and thiophene protons in the range of 7.3–9.5 ppm, and broad upfield peaks at 3.72 and 4.33 ppm were assigned to the internal CH and NH protons. Reaction of 384b with palladium(II) acetate afforded a palladium(II) complex 387 in 69% yield. Single-crystal X-ray diffraction showed that the thiophene unit was tilted relative to the mean macrocyclic plane by 38.19°. A π-extended anthracene-embedded porphyrinoid 388 was also reported and this gave the related palladium(II) complex 389 [230]. meso-Fused anthriporphyrin reacted with diethylacetylenedicarboxylate in 1,2-dichloromethane to give Diels-Alder adduct 390 in 46% yield and this also reacted with palladium(II) acetate to give organometallic derivative 391. A related, more flexible, system 392 called allyliporphyrin has been prepared and this also gave excellent yields of the corresponding palladium(II) complex 393 (Scheme 63) [231]. Porphyrinoids 394 with intermediary structures between 384 and 392 have been described [232]. As is the case for 392, benzo-fused allyliporphyrin 394a is in equilibrium with alternative conformations, or tautomers, in particular the alternative aromatic species 394a′. The latter structure was favored in the solid state. The presence of a nitrogen in pyrido-fused structure 394b relieves steric interactions with the adjacent vinylene hydrogen atom, and this results in enhanced diatropic characteristics. Reaction of 394a with palladium(II) acetate afforded the corresponding palladium(II) complex 395, thereby locking the aromatic conformation in place.

An interlinked thianaphthiporphyrin dimer 396 was prepared from bis-naphthibilane 397 in 2% yield (Scheme 64) [233]. Although cross-conjugated, proton NMR spectroscopy indicates that the system is weakly diatropic. Anisotropy of induced current density (AICD) plots indicate that this is primarily due to the 22 and 34π electron pathways shown in bold for structures 396′ and 396″. Dimer 396 reacted with palladium(II) acetate at room temperature to give bis-palladium complex 398 in 97% yield. A weakly antiaromatic heterobenziporphyrin system with embedded fluorene units was prepared using the same methodology (Scheme 64) [234]. Tripyrrane analogue 399 underwent BF₃.Et₂O catalyzed condensation with thiophene, selenophene or tellurophene dicarbinols to give, following oxidation with DDQ, indeno-heterobenziporphyrins 400. The proton NMR spectra for 400a–c showed that these porphyrinoids have paratropic ring currents resulting in downfield shifts to the internal CH resonances and small upfield shifts to the external protons. Thiabenziporphyrin 400a gave the inner CH resonance at 12.03 ppm. Potential π-electron circuits with 20 and 24π electron pathways, as shown in structures 400′, 400″ and 400‴, may be responsible for this effect. Similar considerations may also be responsible for the reduced aromatic character of indenoporphyrins [235,236]. When reacted with palladium(II) chloride, thia- and selenabenziporphyrins 400b and 400c gave cationic organopalladium derivatives 401 [234].

This review focuses on the metalation of carbaporphyrinoids that share the same 16-atom core arrangement as porphyrins and true carbaporphyrins. However, other closely related systems have some bearing as these may give insightful complementary results. Vacataporphyrins 402 are a case in point [237]. Although this system no longer has the 16-atom core it essentially shares the porphyrin framework minus one of the core atoms. Vacataporphyrins, or deazaporphyrins, were prepared by heating telluraporphyrins 403 with hydrochloric acid at 180 °C (Scheme 65). Vacataporphyrins have similar 18π electron delocalization pathways to regular porphyrins and are strongly aromatic. They react with cadmium(II), nickel(II) and zinc chloride to afford coordination complexes 404a–c [238] and treatment with Pd(PhCN)₂Cl₂ gave a similar palladium complex 404d. Exposure to light led to the formation of a carbon-palladium bond and generated aromatic complex 405a [239]. Reaction with methyl iodide in the presence of AgBF₄ produced the C-methylated cationic palladium(II) complex 405b, and upon heating with methanol deprotonation resulted in the formation of 405c. The proton NMR spectrum for 405b was consistent with a paratropic system due to the conformation facilitating Möbius-type antiaromaticity. Complex 405c is aromatic but this species can be converted back into 405b by treatment with fluoroboric acid.

10. Dicarbaporphyrinoid Systems

Dicarbaporphyrinoid systems have two of the nitrogens within a porphyrin-type cavity replaced by carbon atoms. The first example of this class, 21,23-dicarbaporphyrin 406, was reported in 1999 [240], but many other examples (e.g., 407–413) have been reported over the last 23 years (Figure 10) [78,84,87,241,242,243,244,245,246,247,248,249]. Many of these porphyrinoids are less stable than monocarbaporphyrins and the metalation reactions for these systems has not been explored to the same extent. Nevertheless, interesting examples of metalated dicarbaporphyrinoids have been reported. cis-Doubly N-confused porphyrin (cis-N₂CP, 414) reacted with silver(I) acetate in 10% pyridine-chloroform to give silver(III) complex 415a, and copper(II) acetate reacted similarly to afford the copper(III) derivative 415b (Scheme 66) [250,251,252,253]. The reaction of cis-N₂CP 414 with palladium(II) acetate in refluxing toluene gave a more unusual result, affording a palladium(II) species 416 that had undergone arylation onto an internal carbon atom [254]. A mixture of meta- and para-tolyl isomers were observed in a ratio of 2:1. When N-fused porphyrinoid 417 was treated with potassium hydroxide in ethanol or methanol, ring opening produced alkoxy-substituted trans-N₂CPs 418a,b in 53% and 31% yields, respectively (Scheme 66) [255]. Proton NMR spectra demonstrated that the trans-N₂CP system is highly diatropic, and 418a showed the pyrrolic protons downfield between 8.44 and 8.55 ppm, while the inner CH resonances were observed upfield at −4.34 and −4.36 ppm and the NHs appeared at −2.73 and −3.21 ppm. trans-N₂CPs 418 reacted with copper(II) acetate to give good yields of copper(III) organometallic derivatives 419. Reaction of 417 with five equivalents of thiophenol gave doubly N-confused isophlorin 420 in 11% yield (Scheme 66) [256]. In the presence of air, or on standing over alumina, oxidation took place to give trans-N₂CP dithioether 421. Isophlorin 420 reacted with copper(II) acetate under aerobic conditions to give copper(III) complex 422.

adj-Diazuliporphyrins 423 were isolated as monoprotonated forms as the corresponding free bases were unstable [257]. Reaction with palladium(II) acetate in refluxing acetonitrile gave palladium(II) complex 424 in 26% yield (Scheme 67). This polar organometallic complex can be represented as a series of dipolar or tetrapolar canonical forms. The X-ray crystal structure for 424 revealed that the porphyrinoid skeleton was slightly saddled. The proton NMR spectrum of 424 showed the meso-protons downfield at 7.9 (1H), 8.8 (2H) and 10.0 (1H) ppm, suggesting that this derivative is weakly aromatic. The aromatic properties of 424 can be attributed to resonance contributors such as 424′ that incorporate 18π electron delocalization pathways [257]. Reaction of adj-dicarbaporphyrin 425 with palladium(II) acetate resulted in the formation of a remarkable tripalladium sandwich complex 426 (Scheme 67) [245]. The X-ray crystal structure showed that the complex consisted of a palladium(IV) metallocene-type structure with η⁵-coordination to two palladium(II) dicarbaporphyrin dianions. The individual porphyrinoid units are planar and lie parallel to one another. The remarkable stability of this palladium(IV) derivative shows that dicarbaporphyrinoid systems are capable of stabilizing unusual oxidation states. Reaction of 425 with [Rh(CO)₂Cl]₂ gave rhodium(I) complex 427 [154]. The related dicarbachlorin 428 similarly afforded rhodium(I) complex 429 [154,258].

Carbaazuliporphyrins 430 [87] and carbabenziporphyrins 431 [246] both have three hydrogens within the macrocyclic core and are potentially trianionic ligands. However, attempts to prepare silver(III) derivatives 432 or 433 were unsuccessful (Scheme 68). Reaction of 430 with silver(I) acetate in methanol or ethanol instead selectively afforded nonaromatic oxidation products 434a,b [87]. These dialkoxy derivatives were isolated as single diastereomers, although the precise stereochemical outcome was not determined. Carbabenziporphyrins 431 also gave selective oxidation reactions with silver(I) acetate in methanol-dichloromethane, and nonaromatic products 435 with two meso-methoxy substituents were isolated [246]. Again, the reactions were stereospecific in that only one diastereomer could be identified. A minor hydroxy-derivative 436 was also identified, presumably arising due to the presence of trace amounts of water.

In a recent paper, a dicarbaporphyrinoid system incorporating N-heterocyclic carbenes was described (Scheme 69) [259]. In much the same way as porphyrins act as dianionic ligands, carbenaporphyrins 437 have a similar core arrangement that can potentially behave in the same way, albeit while generating organometallic derivatives. Copper-catalyzed alkyne-azide cycloaddition of a 1.8-diazidocarbazole 438 with a 1.8-diethynylcarbazole 439, an example of a double click reaction, gave macrocycle 440 in 52% yield. Alkylation with Meerwein’s reagent quantitatively generated dication 441 as a bis(tetrafluoroborate) salt. Deprotonation of 441 with four equivalents of lithium hexamethyldisilazide gave a dilithium complex 442 that is equivalent to the target structure 437. Transmetalation with scandium trichloride in THF gave a scandium carbenaporphyrin complex 443 that could be characterized by X-ray crystallography. Treatment of 443 with lithium cyclopentadienide afforded the corresponding cyclopentadienyl complex 444 as an orange solid [259]. Although the versatility of carbenaporphyrin ligands has yet to be demonstrated, this system has the potential to further extend the applications of dicarbaporphyrinoid systems.

11. Tri- and Tetracarbaporphyrins

In principle, replacement of three or four of a porphyrin’s nitrogens with carbons would give tri- and tetracarbaporphyrins (Figure 11) [23,260,261,262]. However, these types of bridged annulene structure are presently unknown, although their significance has been discussed for many years [260]. The theoretical importance of tetracarbaporphyrin (quatyrin) 445 was appreciated by Vogel, who used this structure are a starting point when planning the synthesis of tetraheteroporphyrin dications and porphyrin isomers [263,264,265,266,267]. Unfortunately, attempts to synthesize quatyrin and related structures such as 446 and 447 have so far been unsuccessful [268]. Tri- and tetracarbaporphyrins 445–447 have been assessed using DFT and NICS calculations and the results show that quatyrin is planar and strongly aromatic [261]. However, it is worth noting that dicarbaporphyrins are much less stable than monocarbaporphyrins, and stability issues may plague further work in this area. Other structures with a porphyrin-type framework and four internal carbons can be considered and there has been some success in synthesizing macrocycles of this type. However, this possibility has not yet been applied to N-confused porphyrinoids. Doubly N-confused calix[4]pyrroles have been reported, but attempts to prepare quadruply N-confused calixphyrin 448 were unsuccessful [269].

The carbon skeleton for quatyrin is present in calix[4]azulenes 449, which can be prepared by reacting azulenes with paraformaldehyde in the presence of florisil (Scheme 70) [97,98], and these macrocycles show interesting supramolecular interactions [270,271,272]. As noted earlier, azulene favors electrophilic substitution at the 1,3-positions and can substitute for pyrrole in the construction of porphyrinoid macrocycles. Treatment of 449b with triphenylcarbenium hexafluorophosphate afforded a partially conjugated dication 450 that corresponds to a dihydroquatyrin [98]. This species was dark blue in solution and gave a strong absorption at 616 nm in its UV-vis spectrum. Oxidation of tetraarylcalix[4]azulenes 451 with DDQ in the presence of tetrafluoroboric acid gave the tetraazuliporphyrin tetracations 452 [273]. Although these tetracations might be considered to be didehydroquatyrins, DFT calculations demonstrate that they have severely distorted conformations and the macrocycles are not fully conjugated. A similar triazuli-thiaporphyrinoid 453 has also been reported [274]. Condensation of azulene with a thiophene dicarbinol and p-tolualdehyde in the presence of boron trifluoride etherate gave 454 as a diastereomeric mixture. Oxidation with DDQ and addition of 20 equivalents of HBF₄.Et₂O gave the nonaromatic tetracation 453. Tetraazuliporphyrinoids have not been metalated in the core, although cluster complexes with the seven-membered rings have been reported [275].

Although tetracarbaporphyrin ligands are not presently known, cyclic N-heterocyclic carbenes readily form organometallic derivatives. Macrocycles constructed from four imidazolium subunits have been reported and these form metalated derivatives such as 455 and 456 (Figure 11) [276,277,278,279]. These systems have a similar coordination framework to carbaporphyrins. Tetraimidazolyl tetracation 457 reacted with gold(III) acetate to give gold(III) trication 458 [280], while reaction with copper(II) acetate afforded copper(III) complex 459 (Scheme 71) [281]. Reaction of the latter complex with copper(I) salts and acetic acid, followed by demetallation, gave imidazolone trication 460. Silver and gold cluster complexes of 457 were also reported. Reaction of 457 with bis[bis(trimethylsilyl)amido]iron(II) gave iron(II) complex 461a [282,283,284,285], while treatment with cobalt(II) chloride afforded 461b [286]. Both of these complexes initially reacted with O₂ to give dioxygen complexes. At room temperature, the iron system afforded a μ-oxo dimer, while the cobalt complex generated a μ-peroxy species. Metalation of the related tetrabenzo-ligand 462 [280,287] (Figure 11) and boron-bridged analogues [288,289,290,291] have also been investigated.

12. Contracted Carbaporphyrinoids

Carbaporphyrinoid systems with smaller rings are known, including azulitriphyrin[1.2.1]s 463 [112,292], carbaporphyrins[1.2.1] 464 [112], azulicorroles 465 [293,294], and ethynyl-linked azuliporphyrinoid 466a [295] (Figure 12). Some metalation studies have been performed on contracted systems, including the formation of ruthenium(II) complex 466b, although work in this area is still in its early stages. Technically, these systems fall outside of the primary focus for this review, but the reactivity of these structures is clearly relevant.

Corroles (Figure 12) are contracted porphyrins with only three bridging carbons. These have a more crowded cavity but retain aromatic character and act as trianionic ligands [296,297,298,299,300]. Furuta investigated the metalation of N-confused and neo-confused corroles (Scheme 72) [301,302]. N-confused bilane 467 was oxidatively cyclized with 3.3 equivalents of DDQ in acetonitrile to give N-confused corrole 468 in 5% yield. Isomeric bilane 469 similarly afforded N-confused corrole 470 in 18% yield but in this case a second corrole isomer 471 was generated in 1% yield [301]. Both N-confused corroles favored tautomers with an external NH. Neo-confused corrole 471, named norrole, has a direct link between a pyrrole nitrogen and an adjacent pyrrole unit [301]. Norrole exhibits some diatropic characteristics, and the proton NMR spectrum showed the inner CH as an upfield resonance at 1.21 ppm. N-confused corrole 468 reacted with copper(II) acetate to give copper(III) complex 472, and 470 similarly gave a related copper(III) derivative 473 when the reaction was carried out at 273 K. Although both of the copper(III) complexes are stable, 473 dimerizes in the presence of excess copper(II) acetate or in the presence of the oxidant magic blue to give 474. This complex is linked via the internal carbon atoms. Oxidative cyclization of bilane 475 incorporating an indole unit with chloranil and copper(II) acetate afforded copper(III) benzonorrole 476 in 68% yield [302]. Reductive demetalation with zinc-hydrochloric acid produced free-base benzonorrole 477 in 94% yield. The X-ray structure of the copper(III) complex showed that the tetrapyrrolic unit had a nearly planar conformation. Reaction of benzonorrole 477 with [Ir(COD)(OMe)]₂, 4-substituted pyridines, and potassium carbonate in toluene gave a series of near-infrared phosphorescent iridium(III) complexes 478 [303]. These derivatives have two axial pyridine ligands, but otherwise the macrocycle is near planar.

In an attempt to synthesize silaporphyrins, silole dicarbinol 479 was reacted with pyrrole and p-tolualdehyde in the presence of BF₃.Et₂O, and following oxidation with DDQ, two partially oxidized macrocycles 480 and 481 were isolated [304]. Further oxidation of 481 failed to give a fully conjugated silaporphyrin but instead afforded a low yield of nonaromatic carbacorrole 482. Reaction of 482 with silver(I) tetrafluoroborate or copper(II) acetate gave silver(III) complex 483 and copper(III) derivative 484, respectively (Scheme 73) [304]. Metal insertion was associated with tautomerization to give fully conjugated carbacorrole species. The proton NMR spectra for 483 and 484 showed that these complexes are strongly diatropic, and the protons on the internal tolyl substituents are shifted upfield. For example, solutions of silver(III) complex 483 in CDCl₃ at 180 K showed the o-tolyl protons at 4.46 ppm. Silver(III) complex 483 reacted with aqueous HCl in the presence of O₂ to give oxacorrole 485. During the course of this reaction, the interior benzylic unit and the silver(III) cation are lost.

Azulicorrole 465 was obtained in low yield by condensing azulene, pyrrole and 4-trifluoromethylbenzaldehyde in 10% TFA-CH₂Cl₂, followed by oxidation with DDQ (Scheme 74) [293]. Azulicorrole reacted with copper(II) acetate and gold(III) acetate to give metalated derivatives 486a,b in 89% and 32% yield, respectively [293]. The X-ray structure for 464 showed that the azulene ring was tilted ca. 40° relative to the remaining macrocyclic plane, but as might be expected copper(III) complex 486a was relatively planar. Bilane 487 with two terminal indole units was cyclized with 10 equivalents of copper(II) acetate to give copper(III) complex 488 together with 2.2′-biindole-linked macrocycle 489 in 12% and 18% yield, respectively [305]. The proton NMR spectrum of 488 indicates that there is an 18π electron delocalization pathway in the complex. Attempts to demetalate 488 with zinc dust in TFA-acetonitrile-CH₂Cl₂ to form the parent porphyrinoid were unsuccessful and resulted in the structure being converted back into bilane 487.

Dicarbacorroles with a biphenyl unit or a phenanthrene moiety have been reported [306,307,308]. A bilane-like intermediate 490 incorporating a biphenyl unit was cyclized with pentafluorobenzaldehyde and BF₃·Et₂O and, following oxidation with DDQ, dibenzicorrole 491 was obtained in 10% yield (Scheme 75) [306]. The X-ray crystal structure of 491 showed that the benzene rings were tilted by 19.52° and 20.06° relative to the mean macrocyclic plane. Dibenzicorrole 491 reacted with copper(II) acetate to give organometallic copper(III) complex 492a in 90% yield. Very recently, 491 was shown to react with [Rh(CO)₂Cl]₂ in dichloromethane-methanol to give rhodium(I) complex 492b [307]. However, when the reaction was performed in refluxing acetonitrile, a rhodium(III) organometallic derivative 492c was generated. A structurally related phenanthrene-containing system 493 was prepared in the same way (Scheme 75) [308,309]. The authors named this compound phenanthriporphyrin, but structurally the system is a dicarbacorrole. The proton NMR spectrum for 493 showed the external pyrrolic protons upfield as two 2H doublets at 5.24 and 5.59 ppm, while the inner CH protons were shifted downfield to 16.70 ppm. The external phenanthrene proton resonances were also relatively upfield, appearing at 5.94 and 6.94 ppm. These results are consistent with the macrocycle having a moderate paratropic ring current. The antiaromatic nature of 493 can be ascribed to the presence of 16π- and 20π-electron delocalization pathways shown in bold for resonance contributors such as 493a–d (Scheme 76) [309]. Phenanthriporphyrin 493 reacted with phosphorus trichloride and triethylamine, followed by treatment with methanol in the presence of air, to give phosphorus(V) complex 494 [309]. This species retains the antiaromatic characteristics of the parent ligand. Reaction of 493 with 1.6 equivalents of copper(II) acetate gave antiaromatic copper(III) complex 495 [310]. Regioselective photolytic cleavage of 495 in the presence of molecular oxygen gave copper(III) phenanthribilinone 496. When 493 was reacted with 7.8 equivalents of Cu(OAc)₂ in chloroform-methanol, a diastereoisomeric mixture of dimethoxy derivatives 497a,b was generated. Phenanthriquinone 498, which can be prepared by demethylation of 493 with boron tribromide or sulfuric acid [311], also reacted with copper(II) acetate to give copper(III) complex 499. Reaction with fluoroboric acid afforded BF₂ complex 500. Although 498 and 499 are nonaromatic, boron difluoride cation 500 has aromatic character. This can be rationalized as being due to canonical forms such as 500a–c with 14 or 18π electron circuits. Phenanthriporphyrin 493 reacted with Fe(CO)₅ and I₂ to give keto-derivative 501. It was proposed that this reaction involved the intermediacy of an iron(II) organometallic complex [312].

Porphyrinoids with two interlinked phenanthriporphyrin units have been described (Scheme 77) [313,314]. Diporphyrinoid 502 reacted with copper(II) acetate to give a bis-copper(III) complex 503. A monocopper(III) complex 504 could also be isolated and this reacted with Pd(PhCN)₂Cl₂ to give the mixed Cu^III-Pd^II complex 505 as a stable radical species. When 502 was reacted with Pd(PhCN)₂Cl₂, an unusual bis-palladium complex 506 was formed. Bis-porphyrioid 502 and bis-copper(III) complex 503 are antiaromatic, as judged by proton NMR spectroscopy, but the upfield shifts of the external protons are drastically reduced for dipalladium complex 506. Bond length analysis indicates that the complex has quinoidal character, indicating that the π-systems of the individual macrocycles are strongly interacting.

Dithiaethyneporphyrin 507 has an acetylene unit in place a pyrrole ring (Scheme 78) [315]. The four-carbon bridge facilitates conjugation, and this system can be represented as acetylene-linked structure 507, or the cumulene resonance contributor 507′, both having 18π electron delocalization pathways. Oxidation with silver acetate in an alcohol solvent gave nonaromatic alkoxyphlorins 508, while metalation with Ru₃(CO)₁₂ in refluxing chlorobenzene afforded ruthenium complex 509. A related monothiatriphyrin 510 (Scheme 78) is also aromatic, but protonation affords a nonaromatic cation 511 [316]. Oxidation of 510 with DDQ in the presence of fluoroboric acid gave an aromatic dication 512. Although both 510 and 512 are aromatic, the π-conjugation pathways are quite different. Porphyrinoid 510 acts as a dianionic ligand and reacts with copper(II) acetate to give a copper(II) complex 513a that has significant η²-interactions with the triple bond (Scheme 78). Similar complexes 513b,c were obtained when 510 was reacted with nickel(II) or palladium(II) acetate. Reduction of 513c with sodium borohydride gives an aromatic anion 514 in which the palladium(II) is directly bonded to a carbon atom.

Contracted carbaporphyrinoids can give rise to organophosphorus complexes such as 515 [317]. Reaction of telluraporphyrin 516 with phosphorus trichloride in triethylamine, followed by air oxidation, led to insertion of phosphorus and inversion of the tellurophene ring to give the carbaporphyrinoid complex 517 [318]. Oxidation with m-chloroperbenzoic acid (MCPBA) and reaction with water afforded a further oxidized nonaromatic product 518.

13. Expanded Carbaporphyrinoids

Expanded carbaporphyrinoid systems have also been investigated but these diverge a great deal from the systems discussed above and will not be covered in detail. Early examples of expanded carbaporphyrins are carbasapphyrins 519 and 520, and azulisapphyrin 521 (Figure 13) [319,320,321], but no metalation studies were conducted. It is worth noting that pentapyrrolic sapphyrins were the first expanded porphyrins to be discovered [322] and they continue to be widely investigated [323,324,325]. Dibenziamethyrin 522 was shown to react with Rh₂(CO)₄Cl₂ in benzene to give high yields of bis-rhodium(I) complex 523 and reactions with nickel(II) or palladium(II) acetylacetonate gave bis-nickel(II) complex 524a and bis-palladium(II) derivative 524b, respectively (Scheme 79) [326,327]. Similarly, dicarbaamethyrin 525 reacted with zinc acetate in methanol to give the bridged bis-zinc complex 526 [328]. However, organometallic derivatives for these systems have not been identified.

Expanded porphyrinoid systems often possess inverted heterocyclic rings that place CH units within the macrocyclic cavity, and this may allow the formation of organometallic derivatives. These structures are in essence carbaporphyrinoid-type systems, but only select examples will be presented. Hexaarylhexaphyrins 527 are particularly versatile organometallic ligands that often place four CH units within the macrocycle (Scheme 80) [329,330,331,332]. This provides two binding pockets that resemble dicarbaporphyrinoid structures. Hexaphyrin 527 reacted with NaAuCl₄ to give a mixture of the mono-gold(III) 528 (16%) and the bis-gold(III) complexes 529a (14%) [329]. Reduction of 528 or 529a with sodium borohydride gave the related antiaromatic [28]hexaphyrin complexes 530. This chemistry has been applied to preparation of mixed complexes with Ag(III)-Au(III), Cu(III)-Au(III), Rh(III)-Au(III), and Ir(III)-Au(III) (529b–e) [329,330,331,332,333]. Nickel(II), palladium(II) and platinum(II) complexes 531 were obtained by reacting [28]hexaphyrin(1.1.1.1.1.1) 532 with Ni(acac)₂, PdCl₂ or PtCl₂, respectively [334]. The chiral Mobius aromatic palladium(II) complex 531b has been resolved to give the individual enantiomers by using HPLC on a chiral stationary phase [335]. Treatment of 531b with tris(4-bromophenyl)aminium hexachloroantimonate induced a molecular topology change to give the Hückel aromatic complex 533a. Reaction of 533a with copper(II) acetate gave Pd(II)-Cu(III) [28]hexaphyrin complex 533b in 90% yield, while reaction with silver triflate in acetonitrile afforded Pd(II)-Ag(III) complex 533c in 93% yield [336]. Treatment of 533a with Pd(OCOCF₃)₂ generated the aromatic bis-palladium(II) complex 534, and this was readily deprotonated with tetrabutylammonium fluoride to produce the corresponding dianion (Scheme 80) [337].

Reaction of AuCl.SMe₂ with doubly N-confused hexaphyrin 535 gave the gold(III) complex 536, and further treatment with PtCl₂(PhCN)₂ afforded the mixed Pt(II)-Au(III) complex 537 (Scheme 81) [338]. In another intriguing study, reaction of palladium(II) acetate with dipyrihexaphyrin 538 gave three dipalladium complexes 538a,b and 539 (Scheme 82) [339]. Structure 539 is not an organometallic derivative but has an unusual interlocked structure with two pyricorrole-like components. Many examples of expanded porphyrins with m-phenylene or p-phenylene units have been described and these may also give organometallic derivatives. For example, dibenzihexaphyrin 540 reacted with palladium(II) chloride and potassium carbonate to give Möbius aromatic palladium(II) complex 541 (Scheme 83) [340]. Other examples include the Möbius aromatic palladium(II) porphyrinoids 542 and 543 (Figure 14) [341]. These examples illustrate some exciting examples of organometallic expanded porphyrinoids, but no attempt has been made to give comprehensive coverage of this area.

14. Related Systems

Many other closely related systems with porphyrin-like frameworks have been investigated. Metallocenoporphyrins such as 544–546 (Figure 15) incorporate ferrocene or ruthenocene units in place of a pyrrole ring [342,343]. Surprisingly, the metallocene units facilitate conjugation within these macrocycles and they exhibit a degree of aromatic, or in some cases antiaromatic, character. This shows that π-electron delocalization can be transferred through the d-orbitals of the metallocene component.

Another intriguing class of porphyrin-like structures have been prepared from 21,23-ditelluraporphyrins 547 (Scheme 84). Reaction of 547 with palladium(II) acetate and triethylamine gave derivative 548 where palladium(II) has replaced tellurium as one of the core atoms [344]. The new macrocycle can be viewed as a 21-pallada-23-telluraporphyrin, although the bonding interactions within the core are quite different from other porphyrinoids. X-ray crystallography shows that the Pd was covalently bound to only two neighboring atoms, nitrogen and tellurium, breaking the symmetry of the macrocycle. The proton NMR spectrum at 300 K appears to show a symmetrical aromatic structure, but many of the resonances split at 180 K. The results show that two equivalent structures, 548 and 548′, rapidly interconvert a room temperature via a symmetrical transition state 548^t. Reaction of 547 with Pt(PhCN)₂Cl₂ at 384 K gave an analogous platinatelluraporphyrin 549 [345]. Reduction with zinc amalgam in the presence of Cl₂, Br₂, MeI or allyl chloride gave a series of Pt(IV) addition products 550. When treated with sodium dithionite, 550 afforded the corresponding metallachlorin 551, but this could be oxidized back to 549 with DDQ. Reaction of 547 with [Rh(CO)₂Cl]₂ in refluxing toluene gave rhodatelluraporphyrin 552 and dirhodaporphyrin 553 [346]. The dirhodaporphyrin macrocycle was relatively planar and the rhodium atoms were linked via two chloride bridges. When treated with HCl, 552 was converted to the zwitterionic complex 554. Reaction of ditelluraporphyrin 547 with [Rh(CO)₂Cl]₂ in the presence of air led to loss of both tellurium atoms to afford oxarhodaporphyrin 555. Hence, remarkable new organometallic derivatives can be generated within porphyrin-like frameworks.

Another variation on the theme are structures with carbaporphyrin-like cavities that are built onto porphyrin macrocycles [347,348]. A case in point are the so-called porphyrin earrings (Scheme 85). Palladium-catalyzed Suzuki−Miyaura coupling of diboryltripyrrane 556 with nickel(II) dibromoporphyrins 557a or 557b gave porphyrin “earrings” 558a and 558b in 32% and 20% yields, respectively [347]. It was possible to install two “ears” onto a porphyrin by coupling tetrabromoporphyrin 559 with two equivalents of 556 and double-earring porphyrin 560 was generated in 8% yield [347]. It was necessary to introduce 3,5-didodecyloxyphenyl substituents to increase the solubility of these structures. Although the porphyrin earrings have curved geometries, the newly introduced cavities have the same core atoms as monocarbaporphyrinoid systems. Both 557a and 560 reacted with palladium(II) acetate to give the palladium(II) complexes 561 and 562, respectively, in >90% yield. A number of related porphyrins have been reported [349,350] that bind nickel(II) and palladium(II) within the appended carbaporphyrin-like loop, including structures 563–567 (Figure 16).

15. Conclusions

The 16-atom core of carbaporphyrins has a CNNN binding pocket that can facilitate the formation of metalated derivatives. Indeed, the ordered cavities found in these porphyrinoid structures provide an intriguing environment to probe organometallic processes. These systems form complexes with many of the late transition metals and can stabilize higher oxidation states. The ligands can be profoundly altered by introducing a multitude of different subunits. Pyrrole units can be replaced by furan, thiophene, selenophene or tellurophene. More importantly, the subunit that places a carbon atom within the cavity can be an inverted pyrrole, furan or thiophene, or cyclopentadiene, indene, azulene, cycloheptatriene, inverted pyridine, pyrazole, benzene, naphthalene, and so on. Furthermore, macrocycles with two internal carbons are also easily accessible. These structural changes not only affect metalation processes but also the spectroscopic and chemical reactions for these ligands. Carbaporphyrinoids may be fully aromatic but in some cases, they are nonaromatic or antiaromatic. Some expanded carbaporphyrinoids can even take on twisted conformations that lead to Möbius aromatic or antiaromatic structures. The unprecedented structural diversity of carbaporphyrinoid systems has led to the discovery of a remarkable wealth of coordination architectures and highly usual reactivity. The organometallic complexes also have value in the design of catalytic systems, including catalysts for cyclopropanation reactions [123,351] and CO₂ fixation [352]. In addition, medicinal applications of metallocarbaporphyrinoids as photosensitizers for photodynamic therapy have been noted [353,354]. This area continues to surprise and will no doubt lead to many further advances in the future.