Recent developments of iron pincer complexes for catalytic applications

Iron catalysis is attractive for organic synthesis because iron is inexpensive, abundant, and non-toxic. To control the activity and stability of an iron center, a large number of iron pincer complexes have been synthesized. Many such complexes exhibit excellent catalytic activity in a number of important organic reactions such as hydrogenation, hydrosilylation, dehydrogenation

by a pyridinyl or diarylamine group leads to two major classes of variants of the original pincer ligands. In addition to their strong chelating ability and structural rigidity, the advantage of pincer ligands is their diversity. By modifying the central and side donors as well as the ligand backbone, it is possible to synthesize an almost endless number of ligands with varying electronic and steric properties. 5 This diversity is very attractive for homogeneous catalysis where systematic studies of metal-ligand combinations are desired.
In this review, iron complexes of pincer ligands based on the following frameworks are discussed: 2,6-disubstituted pyridine, 1,3-disubstituted benzene, N,N-diarylamine, isoindoline and bis( phosphinoethyl)amine. Only catalysis by pre-formed iron complexes, but not in situ generated iron species, will be presented. The examples are selected to give a representative, but not comprehensive overview of the developments in the field. Mostly symmetrically substituted pincer ligands with a C 2 or C 2v symmetry are included. Unsymmetrically substituted ligands and terpyridine-type ligands are not treated. Pyridine diimines (PDI) might be considered as pincer ligands, and their iron complexes are very active for polymerization, [16][17][18][19] hydrogenation, and other alkene, alkyne, and ketone functionalization reactions. Several excellent reviews covering these iron complexes have been published recently. 20,21 Therefore, they will not be reviewed here.

Neutral pyridine-based PNP ligand systems and their applications in iron catalysis
The first iron PNP complexes were synthesized by Dahlhoff and Nelson in 1971 by reacting 2,6-bis(diphenylphosphinomethyl)pyridine with FeX 2 (X = Cl, Br, I, NCS). The resulting complexes were in a high spin configuration and exhibited a 5-coordinate geometry with a slightly distorted square pyramidal structure. 6 This ligand framework became popular, and in addition to the variance of the substituents on the phosphorus donors, 5 the linker (L) that connects the phosphorus donors to the central pyridine ring could be modified to include not only CR 2 , 6,22-31 but also NR 5,32-39 and O (Fig. 2). 40,41 To differentiate between the different linkers, these pincer ligands are abbreviated as "P L N L P", where L = C, N, O, to represent CR 2 , NR, and O, respectively.
Several groups studied the influence of linkers on the electronic density of the metal centre using the IR frequencies of M-CO bonds as a probe (Fig. 3). 26,34,40 The CO stretching frequency of [(PNP)FeX 2 (CO)] (X = Cl, Br) decreases when the  linker is changed from O to NH and to CH 2 . This result suggests that the electron donating character of the corresponding PNP ligands follows a similar order. An interesting feature of these PNP pincer complexes is their metal-ligand cooperation. The slightly acidic CH 2 and NH linkers are easily deprotonated under basic conditions, which causes a reversible dearomatization of the central pyridine ring. 42 Further details will be discussed below.
2.1 Iron(II) P N N N P pincer complexes for the selective formation of 3-hydroxyacrylates from benzaldehydes and ethyl diazoacetates Aldehydes react with ethyl diazoacetates (EDA) in the presence of Lewis acids (e.g. BF 3 , ZnCl 2 , AlCl 3 , SnCl 2 , GeCl 2 ) to form β-keto esters. 43 It was previously shown that the Lewis acid [(η 5 -Cp)Fe(CO) 2 THF](BF 4 ) catalyzed the reaction of benzaldehydes with EDA to give β-hydroxy-2-aryl acrylates. 44,45 More recently Kirchner and co-workers employed cis-[( iPr P N N N P)Fe(CO)(CH 3 CN) 2 ](X) 2 (5-X; X = BF 4 − , BArF − ) for the reaction of p-anisaldehyde with EDA ( Fig. 4). 33 These reactions produced selectively a hydroxyl acrylate (A, >80%) and only trace amounts of a β-keto ester (B). The two counter ions, BF 4 − and BArF − , gave similar results. A tentative mechanistic proposal is given in Fig. 5. The CO ligand exhibits a stronger trans effect than the nitrogen of the pyridine ring. Therefore the acetonitrile opposite to the carbonyl will be cleaved first. The dissociation of acetonitrile and successive coordination of an aldehyde gives intermediate 6.
Complex 10-PF 6 gave a 20% yield of the product while complex 10-BF 4 gave 88% yield. These results are in strong contrast to previous results obtained using 5-BF 4 and 5-BArF as catalysts, where the counter ion showed no influence. 33 The reaction mechanism of the catalysis with 10-BF 4 was investigated by DFT/B3LYP computations. It was found that the formation of ethyl 3-hydroxy-2-arylacrylates had a lower energy barrier compared to the formation of the β-keto ester. Hydrogen bonds between the acidic N-H of the ligand, the BF 4 anion and the EDA (N-H⋯F-BF 2 -F⋯EDA) seemed to play an important role. 46

Iron(II) PNP pincer complexes for catalytic hydrogenation
Chirik and co-workers reported that the dinitrogen cis-dihydride complex [( iPr P C N C P)Fe(H) 2 (N 2 )] (11, Fig. 6) catalyzed the hydrogenation of simple acyclic and cyclic alkenes. 24 The hydrogenation of 1-hexene under 4 bar of H 2 was achieved using a 0.3 mol% catalyst loading in three hours and with a conversion of more than 98%. The conversion for the hydrogenation of cyclohexene was only 10%. Milstein and coworkers developed a new iron pincer complex [( iPr P C N C P)FeH (CO)Br] (12, Fig. 6). 26 The complex was active for the hydrogenation of ketones under mild conditions. The reactions typically proceeded in an ethanolic solution with 0.05 mol% of 12 under 4 bar of hydrogen pressure and temperatures of 26-28°C. A maximum TON of 1880 was reached. A wide range of aromatic and aliphatic ketones could be hydrogenated. The catalysis was less selective for enone substrates such as E-4phenylbut-3-en-2-one and cyclohex-2-enone, where the CvC double bonds were preferably hydrogenated. Interestingly, for benzaldehyde only 36% of benzylic alcohol could be isolated after the hydrogenation. The low yield might be due to catalyst poisoning by benzoic acid, which was formed in small quantities via a Cannizzaro reaction. By lowering the catalyst loading to 0.025 mol%, increasing the loading of KOtBu (base) to 0.625% and running the reaction in an ethanol/NEt 3 (2 : 1) solution at 40°C under 30 bar H 2 , the yield for the hydrogenation of benzaldehyde could be increased to up to 99% and the TON was up to 4000. This modified protocol was applied for the hydrogenation of various aromatic and aliphatic aldehydes. 31 For the hydrogenation of ketones catalyzed by 12, the reactions were best run in ethanol; in THF or neat acetophenone no conversion was observed. Stoichiometric reactions revealed that the bridging methylene group of 12 was deprotonated by KOtBu to form the complex [( iPr P C N C P -H )FeH(CO)] (13, Fig. 7) containing a dearomatized pyridine ligand. It was proposed  that the 5-coordinated 16-electron species 13 might be stabilized by the reversible addition of ethanol to afford the 6-coordinate species 13′. In the proposed mechanism, the coordination of ketone to 13 followed by insertion into the iron-hydride bond gave the alkoxide complex 14. The coordination of dihydrogen to 14 then gave intermediate 15.

Inorganic Chemistry Frontiers Review
Heterolytic cleavage of the dihydrogen in 15 then generated either the hydrido alkoxide complex 16 or its dearomatized form 16′.
The catalytic cycle was closed by elimination of the alcohol and regeneration of 13. 26,47 The Milstein group further found that minor modifications of the [( iPr P C N C P)FeH(CO)Br] (12) catalyst led to base-free hydrogenation of ketones. 28 They found that [( iPr P C N C P)FeH (CO)(η 1 BH 4 )] (17, Fig. 6) catalyzed the hydrogenation of aceto-phenone under 4 bar of hydrogen at 40°C without the need for an additional base. Aromatic and aliphatic ketones could be readily reduced to the corresponding alcohols with yields of 53% to 99%. Interestingly, a related complex, [( iPr P C N C P)FeH-(η 2 BH 4 )] (18, Fig. 6), was not active. In aprotic solvents such as benzene-d 6 and toluene-d 8 , [( iPr P C N C P)FeH(CO)(η 1 BH 4 )] (17) loses BH 3 to give trans-[( iPr P C N C P)Fe(H) 2 (CO)], which is in equilibrium with cis-[( iPr P C N C P)Fe(H) 2 (CO)]. Neither complex reacts with acetophenone, excluding its participation in the catalytic cycle.
Kirchner and co-workers recently synthesized a series of [( iPr P N N N P)FeH(CO)L] n (19-L) complexes with L = Br, CH 3 CN, pyridine, PMe 3 , SCN − , CO and BH 4 − and n = 0 and +1. 39 The spacers between the phosphine and pyridine donors were NH   Interestingly, both 19a-Br and 19b-Br were active for the hydrogenation of aldehydes. Thus, 19b-Br is a chemoselective cata-lyst for the hydrogenation of aldehydes. The mechanism of hydrogenation catalyzed by 19a-Br was investigated by stoichiometric reactions and DFT computations. 39 The results suggested that 19a-Br reacted with one equivalent of a base to afford the deprotonated 5-coordinated complex [( iPr P N N N P -H )-FeH(CO)] (20; Fig. 9). An incoming ketone coordinated to the vacant coordination site on the complex and inserted into the iron-hydride bond to give the 5-coordinated alkoxide complex [( iPr    Complex 24-Br is a chemoselective catalyst for the hydrogenation of aldehydes. The scope of the reaction was probed for a range of different aliphatic and aromatic aldehydes, using 10 mol% of 24-Br in methanol under 8 bar hydrogen at room temperature (Table 1). Importantly, functional groups such as terminal and internal alkenes, α,β-unsaturated aldehydes and keto were tolerated. The addition of 10 mol% HCOONa promoted the reaction. The catalyst loading and the H 2 pressure could be reduced to 5 mol% and 4 bar, respectively (Table 2). While 24-CH 3 CN was not active for hydrogenation under the conditions shown in Table 1, it was active when HCOONa was used as an additive. 41 The promotion by HCOONa encouraged the study of 24-L for transfer hydrogenation of aldehydes using sodium formate as the hydrogen donor. Indeed, both catalysts were active and chemoselective for this reaction (Table 2). 41 Milstein and co-workers showed that trans-[( tBu P C N C P)Fe-(H) 2 CO] (25; Fig. 11) catalyzed the hydrogenation of sodium bicarbonate. 27 A TON of up to 788 and a TOF of up to 156 h −1 were obtained. The yields of sodium formate were around 40%. NMR and IR studies suggested that carbon dioxide inserted into one iron hydride bond to give [( tBu P C N C P)Fe(H)-(η 1 -OOCH)CO] (26). The structure of this compound was confirmed by X-ray analysis. The formate ligand in 26 was easily replaced by water to form [( tBu P C N C P)Fe(H)(H 2 O)CO] (27). A catalytic cycle was proposed (Fig. 12), in which the coordination of hydrogen to 27 gives [( tBu P C N C P)Fe(H)(H 2 )CO] (28). An incoming hydroxide deprotonated the hydrogen via heterolytic cleavage (29). This might involve the deprotonation of the PNP ligand forming the dearomatized complex 29′. Either way, after elimination of water the starting dihydride complex 25 was regenerated. 27 Complex 25 was also an active catalyst for the decomposition of formic acid, with a TON of up to 100 000. 29

Monoanionic benzene-based PCP ligand systems
The synthesis and application of PCP iron complexes are less reported than their PNP counterparts. These complexes are mostly formed via cyclometalation reactions. To promote the iron-carbon bond formation, iron precursors in a low oxidation state or with basic ligands such as alkoxides and alkyl groups are used. [48][49][50] Creaser and Kaska reported the synthesis of [( Me P C C C P)Fe-(H)(dmpe)] (30, Fig. 13). 51 Treating the Me P C C C P ligand with hydrated ferrous chloride in an ethanolic solution resulted in the precipitation of a white polymeric compound [( Me P C C C P)-FeCl 2 ] n (A). This polymeric compound was converted to 30 in the presence of dmpe (1,2-bis(dimethylphosphino)ethane) and 30% sodium amalgam. The structure and composition of 30 were confirmed by NMR and IR. The insertion of a Fe ion into the C-H bond of benzene was proposed to take place when the Fe(II) ion was reduced to Fe(0).
Guan and co-workers synthesized [( R P O C O P)FeH(PMe 3 ) 2 ] (31-R; R = iPr, Ph; Fig. 15) 52,53 by cyclometalation of the R P O C O P ligand with Fe(0)(PMe 3 ) 4 . 31-iPr catalyzed the hydrosilylation of various aryl aldehydes by (EtO) 3 SiH. Aryl ketones were less reactive under similar reaction conditions. In order to gain some mechanistic insights, stoichiometric reactions were probed. Benzaldehyde did not react with 31-iPr, excluding its insertion into the Fe-H bond as a catalytic step. The reaction of 31-iPr with CO for 24 hours at room temperature gave quantitatively [( iPr P O C O P)FeH(PMe 3 )(CO)] (32), where the CO ligand was trans to the hydride. This result was expected because the hydride had a strong trans-effect, which made the trans PMe 3 more labile than the cis one. Complex 32 slowly isomerized to the thermodynamically more stable complex [( iPr P O -C O P)FeH(PMe 3 )(CO)] (32′) where CO was cis to the hydride, but the reaction took more than 7 days at 60°C. The reaction of deuterium-labelled C 6 H 5 CDO with a mixture of Ph 2 SiD 2 and 31-iPr (1 : 1 : 1) led to the formation of Ph 2 SiD(OCD 2 C 6 H 5 ) and Ph 2 Si(OCD 2 C 6 H 5 ) 2 in a ratio of 4 : 1. No deuterium incorporation in 31-iPr was detected. This result further excluded the involvement of the hydride in 31-iPr in the hydrosilylation reaction. Based on the above results Guan and co-workers proposed the following mechanism (Fig. 14). 52 The PMe 3 in 31-iPr trans to the hydride dissociated to give a vacant coordination site (33, Fig. 14). The isomerisation in which the remaining PMe 3 ligand moved to a position trans to the hydride was too slow to be catalytically relevant. An incoming carbonyl compound might coordinate to the open site at first before being reduced by silane (34, Fig. 14; cycle I). Alternatively, silane might first coordinate to the vacant site, get activated, and then reduce the carbonyl compound (35, Fig. 14; cycle II). 52 The details of these steps were not clear. Wang et al. reported a DFT study on the mechanism of this hydrosilylation reaction, favouring cycle I. They found that the carbonyl-coordinated compound underwent an isomerization to move PMe 3 to a position trans to the hydride. This was followed by the insertion of the carbonyl compound into the Fe-H bond, giving an iron alkoxide complex. The cycle was    closed by σ-bond metathesis with silane, yielding silyl ether and regenerating the iron hydride species. 54 The computed mechanism, however, was at odds with the experimental findings, which excluded the insertion of benzaldehyde into the Fe-hydride of 31-iPr as a relevant step. 52 Guan and co-workers also used 31-iPr for the decomposition of NH 3 BH 3 (AB) to give H 2 (Fig. 15). 55 The reactivity of 31-iPr could be increased by replacing the two PMe 3 ligands with PMe 2 Ph, giving the complex [( iPr P O C O P)Fe(H)(PMe 2 Ph) 2 ] (36). Complex 36, which showed an increased activity in the beginning of the reaction, was less stable than 31-iPr during the reaction. A further attempt was undertaken to increase the electron density of the metal center, by adding a methoxy group para to the Fe-C bond (37, Fig. 15). Complex 37 was indeed more active than 31-iPr and 36, and more stable than 36. 55 Recently Guan and co-workers prepared cationic complexes

Pyridine-and phenyl bis(oxazolinyl) pincer systems
Iron pyridine bis(oxazoline) ( pybox) complexes (Fig. 16) have been widely used in catalytic reactions. In the majority of the cases, these complexes were formed in situ by combining iron salts and pybox ligands. The reactions studied included cyclizations, 57-60 additions, 61-65 the kinetic resolution of racemic sulfoxides, 66 and hydrosilylation of aromatic ketones. 67 The active species were unclear, so they were not further discussed.
In a subsequent study, Nishiyama and co-workers showed that 51-R could be activated by the addition of 6 mol% Zn. The model substrate, 4-acetylbiphenyl, could be reduced in a 98% yield with a 65% ee. 79 But in this reaction the S-enantiomer was the major product, which is opposite to the reaction using the mixture of Fe(OAc) 2 with a ligand. The origin of the change in selectivity was unclear. The new reaction conditions were also general for the enantioselective hydrosilylation of ketones (Table 8; condition B).

2,5-Disubstituted pyrrolidinebased pincer ligands
The group of Gade developed a new class of pincer ligands based on bis(oxazolinylmethylidene)isoindolines (boxmi). Complex [(boxmi-Ph(S))FeCl(S n )] (54-Ph(S); S = solvent; n = 1 or 2) was an active catalyst for the azidation of cyclic β-keto esters and 1-indanones (Table 10). 82 The indanone derived tert butyl β-keto esters could be converted in high yields and high enantioselectivities (90-93%) were achieved for many substrates. It was shown that the bulky tert butyl substituent on the ester is crucial for higher enantioselectivities (compare entry 7 with entry 9). Substituted cyclopentenone derivatives could also be converted in high enantioselectivity. 82 Gade and co-workers then applied this catalytic system for the azidation of 3-aryloxindoles. The previously optimized conditions gave lower ee's. The selectivity could be increased by the in situ formation of the catalytically active species, which was generated by mixing 10 mol% Fe(OOCEt) 2 with 12 mol% 55-Ph(S) in diethylether at room temperature. Under the optimized reaction conditions the reactions of various 3-aryloxindoles were successful. The yields were between 85 and 90% and the ee were between 87 and 94% (Table 11). 82 Gade and co-workers then showed that [(boxmi-Ph(R))Fe-(OAc)] (56-Ph(R)) and [(boxmi-Ph(R)Fe((S)-OCHCH 3 Ph)] (57-Ph(R)) catalyzed the enantioselective hydrosilylation of aryl ketones (Fig. 22). 83 The latter complex was employed for the hydrosilylation of various substrates (Table 12).
Good yields and selectivity were achieved for many substituted acetophenones. Different substituents had almost no influence on the outcome of the reaction. Changing the alkyl substituent of acetophenone to a more sterically demanding isopropyl group decreased the ee to 73% (entry 8). The use of diaryl ketones decreased the selectivity (entries 15 and 16). An inner sphere mechanism was proposed. Activation of the precatalyst gave an iron alkoxy complex, which undergoes σ-bond metathesis with silane to give the silyl ether product and the iron hydride species. The prochiral ketone coordinates to the hydride complex and subsequently inserts into the metal hydride bond to regenerate the iron alkoxy complex. 83

Hydrogenation of esters and nitriles
The groups of Guan and Beller independently reported the hydrogenation of esters to alcohols catalyzed by 62-iPr. 53,86,87 Both aliphatic and aromatic esters were hydrogenated in moderate to high yields. Aromatic esters showed higher reactivity compared to the aliphatic esters. For a substrate containing both CvC and CvO groups, methyl cinnamate, hydrogenation of both groups was obtained. Guan and co-workers then reported the hydrogenation of CE-1270, a methyl ester derived from coconut oil. 86 The substrate consisted of methyl laurate (C12, 73%), methyl myristate (C14, 26%) and trace amounts of C10 and C16 methyl esters (∼1%). The hydrogenation was carried out under neat conditions under 52 bar hydrogen pressure at 135°C, using 1 mol% 62-iPr as the catalyst. The reaction gave a full conversion after three hours, giving a combined GC yield of 99% for fatty alcohols. In an upscale experiment (100 g), the yield was only 25%, suggesting catalyst degradation. Lowering the temperature to 115°C increased the yield to 40%. 86 Complex 62-iPr was then further employed in the hydrogenation of neat coconut oil, without transforming it to the corresponding fatty acid methyl esters. A yield of 12% was reached after 23 hours. 92 Beller and co-workers reported the hydrogenation of nitriles and dinitriles using 62-iPr as the catalyst. 91 The catalyst showed good chemoselectivity, tolerating various functional groups (Table 13). a All reported yields were of the isolated products. b The absolute configuration was determined by a crystal structure of an 1,2,3-triazole which was obtained by click reaction of the azide compound, entry 3 with (3-bromophenyl)acetylene. The configuration was confirmed as the R-enantiomer.  Aryl nitriles with electron-donating groups could be reduced at a lower temperature than aryl nitriles with electronwithdrawing groups. The method could be employed for the hydrogenation of more challenging aliphatic nitriles.

Acceptorless dehydrogenation of alcohols, formic acid and N-heterocycles
Beller and co-workers showed that 62-iPr and [Fe( iPr P (CH 2 ) 2 -NH (CH2)2 P)H(CO)Br] (64-iPr; Fig. 24), were active catalysts for the dehydrogenation of methanol to give hydrogen and CO 2 in the presence of a base (8 M KOH) at 91°C. The TON was up to 10 000 after 46 hours. In order to improve the stability of 62-iPr in the reaction, an additional ligand (up to 5 equivalents) was added. The life time of 62-iPr could be improved from 43-66 h to 5 days. 3 Hazari and co-workers 95 showed that complex [Fe( R P (CH2)2 N (CH2)2 P)H(CO)] 65-R (R = iPr, cy; Fig. 24) catalyzed the base-free dehydrogenation of methanol in water. Methyl formate was obtained as a side product, suggesting the incomplete dehydrogenation of methanol. Following the reaction by NMR revealed the formation of [Fe( cy P (CH2)2 NH (CH2)2 P)-H(CO)(OOCH)] (66-cy; Fig. 24), which took the catalyst away from the catalytic cycle. In the presence of a Lewis acid the accumulation of 66-cy could be prevented. Thus, Lewis acids were used to promote the dehydrogenation of methanol. A TON of up to 51 000 could be reached. 95 In an analogous study, LiBF 4 was used to promote the dehydrogenation of formic acid catalyzed by 65-R and 66-R (R = iPr, cy). They obtained the best results using 0.0001 mol% of 66-iPr in a 2.2% solution of formic acid in dioxane at 80°C. 10 mol% of LiBF 4 was used as the additive. The reaction finished after 9.5 hours giving a TON of 983 642 with a TOF of 196 728 h −1 . 89   The groups of Schneider and Jones further reported the acceptorless dehydrogenation of different primary and secondary alcohols using 62-iPr as the catalyst. 90 The formed H 2 was removed from the reaction medium by a steady flow of N 2 , in order to shift the reaction equilibrium. Secondary benzylic alcohols (entries 1-8) were oxidized to the corresponding acetophenone derivatives in good to excellent yields (65-92%).
The protocol also tolerates a range of different functional groups (MeO, Me, NO 2 and halides). For substrates with electron withdrawing groups (entries 4-6) the catalyst loading could be lowered to 0.1 mol%. An aliphatic cyclohexanol (entry 10) was oxidized to cyclohexanone in a 64% yield. Primary alcohols and diols were also used as substrates. Benzylic alcohol formed benzyl benzoate exclusively (entry 9). The chemoselectivity of secondary over primary alcohol was tested using 1,2-propandiol and 1,3-butanediol (entries 13 and 14). In both cases the secondary hydroxyl group was oxidized before the primary group. The diols 1,2-benzenedimethanol and 1,5-pentanediol (entries 11 and 12) gave the corresponding lactones in 96 and 59% yield. Furthermore, it was shown that the reactions could be reversible, and the hydrogenation of ketones was achieved with 62-iPr as the catalyst (Table 14). 93 Beller and co-workers further studied the dehydrogenation of diol substrates to give lactones catalyzed by 62-iPr. 94 In contrast to the conditions described by Schneider and Jones. 90 (toluene, reflux, 1 mol% 62-iPr), their reactions took place in a Isolated yields at full conversion (primary amines were isolated as HCl-salts). b GC-yield with hexadecane as the internal standard. c 10% transesterification product (-COOiPr) product were observed. d Reaction was quenched after 20 min. tert-amyl alcohol at 150°C with 0.5 mol% 62-iPr and 10 mol% K 2 CO 3 . The conditions were applied to 18 different aromatic and aliphatic diols which could be converted to the corres-ponding 5-membered lactones in good to excellent yields. Also 6-membered and 7-membered lactones can be synthesized. The scope was further extended to 9 different α,ω-amino alcohols giving the corresponding lactams in good to excellent yields. Jones and co-workers applied complexes 62-iPr, 64-iPr and 65-iPr for the acceptorless dehydrogenation of saturated N-heterocycles to their aromatic counter parts and molecular H 2 . 88 The method was applied to a range of different 1,2,3,4-tetrahydroquinolines (Table 15, condition 1, entries 1-6). 2,6-Dimethylpiperidine and 2-methylindoline could be oxidized as well (entries 7 and 8). Partial dehydrogenation products were not observed. Remarkably, pyridine and quinoline products, which were potential ligands for iron, did not influence the reaction. It was observed that in the presence of 5 bar of H 2 62-iPr completely lost its catalytic activity for the dehydrogenation reaction, which suggested that this complex catalyzed the reverse reaction, the hydrogenation of unsaturated heterocycles in the presence of H 2 . This was indeed true, and con-  ditions were found for the hydrogenation of many such unsaturated heterocycles (Table 15, condition 2).

Summary and outlook
As described above, many new classes of iron pincer complexes have been recently synthesized and applied for organic synthesis. The reactivity of these complexes depends critically on the nature of the pincer ligands. The large number of successful applications underscores the advantages of pincer ligands, namely their diversity, modularity, and strong chelating ability. These examples also suggest that pincer ligands and iron catalysis are a perfect match for one another. While iron pincer complexes are often used for hydrogenation, hydrosilylation, and related dehydrogenation reactions, their applications in other reactions including epoxidation, aziridation, and C-C bond-forming reactions are emerging. Given the many benefits of iron catalysis in organic synthesis, it is clear that there is a huge amount of unexplored potential of iron pincer complexes in catalysis. Until now, enantioselective catalysis using chiral iron pincer complexes remains scarce. This state will likely change rapidly in the near future.