RuII , IrIII and OsII mesoionic carbene complexes : e ffi cient catalysts for transfer hydrogenation of selected functionalities † ‡

Pyridine-appended triazolylidene donors have been recently used as ligands in various homogeneous catalytic processes. We present here a new pyrimidine substituted triazolium salt which was prepared and used in the coordination to Ru and Os. The new triazolium salt and the obtained complexes were characterized by multinuclear NMR spectroscopy. The molecular composition of the mentioned compounds was confirmed by positive ion electrospray ionization (ESI+) mass spectra. The new pyrimidyl containing complexes, as well as the related pyridyl-triazolylidene containing complexes, were applied in transfer hydrogenation reactions of carbonyls, alkenes, imines and nitroarenes. The pyrimidyl containing complexes reveal an over-all better activity in comparison to their pyridine bearing analogues. The studies of electronic effects of the ligands, as well as mechanistic insights for the reduction of nitrobenzene with three selected precatalysts are presented.


Introduction
The transfer hydrogenation reaction promoted by metalcomplexes is currently one of the most commonly investigated homogeneous hydrogenation processes. 1This reaction is highly appealing for several reasons.It does not require hazardous gaseous hydrogen and pressurized reactors, and it employs cheap and sustainable solvents as the source of hydrogen, to name just a few.The wide range of functional groups that can be reduced by transfer hydrogenation is noteworthy.These include carbonyl, olefin, imine, nitrile, and nitro compounds.The need to develop simple and sustainable reaction conditions that require only low catalyst loadings is currently emerging in developing new catalysts and catalytic systems.
In the last two decades N-heterocyclic carbenes (NHCs) have been developed as a powerful class of ligands.Despite being used in nearly every field of modern chemical sciences, without a doubt NHCs have most dramatically affected organometallic chemistry and catalysis.Owing to their strong σdonation ability, air and moisture stability, high structural variability, as well as low toxicity, they have become viable alternatives to widely used phosphines in many catalytic processes.Today, a vast number of transformations are known to proceed only with NHC-based catalysts. 21,2,3-Triazol-5ylidenes are a prominent subclass of NHCs possessing a mesoionic carbene (MIC) structure.Due to a lower heteroatom stabilization and mesoionic character, MICs are among one of the best σ-donors among the NHCs. 3 The highly efficient and modular nature of click reaction that is generally used to access 1,2,3-triazol-5-ylidenes makes the design and preparation of diverse libraries of these ligands extremely easy. 4Complexes bearing triazolylidenes are increasingly used in homogeneous metal-based catalysis, as well as for photophysical and biochemical purposes.4d Many catalytic systems of this type have been studied, showing comparable or even better efficiency in comparison to the majority of their NHC analogues. 5Metal complexes of MIC ligands have been used as catalysts among others in olefin metathesis, 6 oxidations, 7 reductions, 8 click reactions, 9 and C-C/C-N bond formation reactions. 10ecently, three series of complexes consisting of pyridine appended triazolylidenes as ligands with ruthenium(II), iridium(III) and osmium(II) were reported and they were successfully used in transfer hydrogenation of selected carbonyl substrates (Scheme 1). 11All complexes have shown remarkable activity even at very low catalyst loadings with several different substrates investigated.The ruthenium(II) complexes were also used in oxidation of primary and secondary alcohols.7d Encouraged by these results, we were prompted to test the activity of these complexes also for transfer hydrogenation of substrates bearing other functional groups.In particular, we were interested in different carbonyl, olefin, imine, and nitroarene substrates.
In addition to the pyridine-appended MICs, we decided to design novel MICs that are functionalized with a pyrimidine ring.The starting 1-(2-pyrimidyl)-functionalized triazolium salt was prepared in excellent yields and purity and used to construct the corresponding Ru(II) and Os(II) complexes (Scheme 1).The coordination was achieved by a transmetallation protocol and the corresponding complexes were characterized by 1 H, 13 C and 15 N NMR spectroscopy and high resolution mass spectrometry.Unfortunately, the preparation of the Ir(III) analogue failed.The two pyrimidine-appended MIC complexes of Ru(II) and Os(II) and the previously reported 7d,11 pyridineappended MICs of Ru(II), Ir(III), and Os(II), shown in Scheme 1, were used in transfer hydrogenation of the above mentioned functionalities.This is also the first report on pyrimidineappended triazolium salts, the corresponding MICs and their transition metal complexes.Some mechanistic insights into the reduction of nitrobenzene with the three selected complexes are presented.

Ligand synthesis and coordination
In contrast to earlier reports of unselective methylation of pyridyl-triazoles, 12 some of us have recently reported a selec-tive route to generate only N Triazole methylated pyridyl-triazolium salts. 13Some of these salts have already been successfully used to generate ruthenium(II), osmium(II) and iridium(III) complexes.7d,11 Inspired by these reports, we designed and prepared an analogous derivative, that has a pyrimidine ring attached directly to the triazole nitrogen atom N1 (Scheme 2).The starting triazole 1 was prepared by copper-catalyzed cycloaddition of phenylacetylene with 2-azidopyrimidine (tetrazolo [1,5-a]pyrimidine) as reported previously. 14The methylation into 2 was achieved by mixing the triazole with MeOTf in dry dichloromethane at 0 °C for 24 h (Scheme 2).Triazolium salt 2 was isolated as a triflate salt in 91% yield.In analogy to 1-(2-pyridyl)-1,2,3-triazoles, 13 1-(2-pyrimidyl)-1,2,3-triazole 1 could be selectively alkylated with MeOTf at triazole nitrogen N3 with no need of pyrimidine ring protection.
Next, the coordination ability of triazolium salt 2 towards Ru(II), Ir(III) and Os(II) was examined.To this end, we chose the already established transmetalation protocol via the intermediate carbenic silver(I) complex, shown in Scheme 3. 15 The triazolium salt 2 was stirred with Ag 2 O and in dry acetonitrile at room temperature for 4 days (Scheme 3).The formation of complex Ag-2 was suggested by the disappearance of the triazolium proton H-5 from the 1 H NMR spectral analysis in DMSO-d 6 .Due to the general instability of such silver complexes, we were unable to obtain more experimental data for its structure.Subsequently, Ag-2 was subjected to transmetalation without previous isolation employing [Ru( p-cym)Cl 2 ] 2 as a ruthenium source (Scheme 3).The reaction progress was monitored by TLC analysis.The pure complex Ru pyrim was isolated as a triflate salt after 2 hours by simple three-step filtration-solvent evaporation-crystallization workup in 85% yield (Table 1).For the preparation of Os pyrim , which involved transmetalation with the Os II precursor [Os( p-cym)Cl 2 ] 2 the reaction mixture had to be stirred for a prolonged time of 3 days.The complex Os pyrim was isolated in the pure form as a PF 6 salt in a moderate yield of 53% (Table 1).Unfortunately, we were unable to prepare the Ir III analogue.After the addition of [IrCp*Cl 2 ] 2 to the reaction mixture of the triazolium salt 2 and Ag 2 O in acetonitrile, no pure product could be isolated.Instead a mixture of products was isolated.The use of a base such as Cs 2 CO 3 or heating of the reaction mixture did not lead to any improvement.Even though we could identify the product in the mass spectrum of the crude reaction mixture (see the ESI ‡), attempts to purify it led to its decomposition.Although it is not clear at this stage as to why the product decomposes at the purification stage, we think that it does not survive either the filtration through Celite, or the salt metathesis with KPF 6 that is necessary for its purification.

Transfer hydrogenation catalysis
After having the two complexes Ru pyrim and Os pyrim prepared, we examined their activity in transfer hydrogenation catalysis.In this context we also tested some Ru II , Ir III and Os II complexes from our previous reports, which were already used for carbonyl group reduction. 11As the model reaction we examined the reduction of benzaldehyde to benzyl alcohol.After testing at room temperature, 60 °C and 80 °C, we observed that the best conversions are obtained at 100 °C.Hence, all transfer hydrogenation reactions (other than nitrobenzene) were carried out at 100 °C.The results for benzaldehyde are presented in Table 2.
The reduction of benzaldehyde under the applied reaction conditions proceeds smoothly with all tested complexes.Applying 0.5 mol% of the precatalysts, as well as with only 0.01 mol%, the reduction occurred with full conversion of the starting material, reflecting in a TON of about 10 000.This is comparable to known ruthenium(II)-NHC complexes 16 and in line with the activity of their pyridyl-MIC analogues. 11To investigate further the catalytic performance of the Ru pyrim and Os pyrim , acetophenone was used as a substrate.Ketones in general are known to be more challenging substrates than aldehydes for the reduction.The results are outlined in Table 3.
Whereas the reduction of acetophenone with the ruthenium complex Ru pyrim at the precatalyst loadings of 0.5 mol% gave full conversion of acetophenone to 1-phenylethanol, the osmium complex Os pyrim gave only 70% of the reduced substrate.By using 0.01 mol%, the pyrimidine analogue Ru pyrim has shown the best activity amongst all tested precatalysts.With 0.5 mol% of the Ru 2 precatalyst, we were able to isolate the product in 85% yield (conversion >99%).Prompted by very good overall results with acetophenone we decided to continue with a more sterically hindered substrate, benzophenone.The results are collected in Table 4.
Benzophenone was fully converted into diphenylmethanol with the Ru pyrim precatalyst at 0.5 mol%.The same complex gave an excellent 93% conversion also at 0.1 mol% precatalyst loading, which is the best performance for all tested compounds.With 0.1 mol% of the Ru 1 precatalyst, we were able to isolate the product in 60% yield (conversion 69%, Table 4).On the other hand Os pyrim gave only 72% conversion by employing  0.5 mol% of the precatalyst.This result is in line with other Os(II) complexes Os 1-3 .To finish, we conducted the reduction of cyclohexanone as a representative of an aliphatic substrate.The results are summarized in Table 5.
The two pyrimidyl-MIC complexes Ru pyrim and Os pyrim have shown the best activity for the transfer hydrogenation of cyclohexanone into cyclohexanol.Employing either 0.5 mol% or 0.1 mol% of the precatalyst resulted in an excellent overall activity.The Ru pyrim gave a very good 78% conversion also by using only 0.05 mol% of the precatalyst.Good results of the complexes Ru pyrim and Os pyrim in the transfer hydrogenation of the carbonyl group prompted us to test other functional groups.We started with olefin substrates: cyclooctene, methylstyrene and trans-stilbene.The results for the reduction of cyclooctene are presented in Table 6.
For the reduction of cyclooctene, we used similar reaction conditions as mentioned above.A mixture of cyclooctene, the selected precatalyst and KOH in iPrOH was stirred under an inert atmosphere at 100 °C for 24 h.By using 1.0 mol% of ruthenium-based precatalysts we observed full conversions into cyclooctane with all four tested complexes.Also with the iridium and osmium analogues, the conversions were very good.Scaling down to 0.5 mol% of Ru(II) complexes, we could still reach excellent conversions, better than those described for already known similar systems. 17The best performance was seen with Ru 2 , bearing the 4-methoxyphenyl donor group at the triazole, which is in line with the results for the reduction of carbonyls.Excellent performance in the reduction of cyclooctene encouraged us to try other such substrates.We took trans-β-methylstyrene and applied the same reaction conditions as for cyclooctene.The results are summarized in Table 7.
In comparison to cyclooctene, trans-β-methylstyrene proved to be a more challenging substrate.With the exception of the Os(II) representatives, the complexes still performed very well.Nonetheless, our catalysts have shown similar or better performance in this reduction than the previously reported examples. 17The ligand effect trends were similar to the previous observations.For the Ru(II) complexes the electrondonating substituents at the triazole ring worked better, whereas the Ir(III) complexes preferred electron withdrawing groups. 11Furthermore, we investigated the transfer hydrogenation of trans-stilbene, a sterically more hindered alkene.The results from Table 8 confirm the reduction of trans-stilbene into 1,2-diphenylethane as more demanding for our catalytic system, which has been also reported for other such systems. 17  To continue we took an imine substrate and selected N-benzylideneaniline. 18 We employed both K 2 CO 3 as well as KOH as bases and stirred the reaction mixture for 18 h at 100 °C.Both bases delivered similar conversions.The results are collected in Table 9.
With 1.0 mol% precatalyst loading the ruthenium-based precatalysts converted the N-benzylideneaniline into N-benzylaniline fully.Also by using 0.5 mol% the same complexes gave an excellent overall conversion.The Ru pyrim performed best, but also the Ru 2 showed excellent activity, confirming the ligand effects for the Ru(II) complex series.On the other hand the Ir(III) and Os(II) precatalysts were less active for the reduction of this type of substrates.The activity of our complexes is comparable to the previously reported ruthenium catalyst in the transfer hydrogenation of N-benzylideneaniline. 18 We also tested the reduction of cyclohexenone as an example of an α,β-unsaturated carbonyl.Whereas conversions were high for all tested precatalysts (Table 10), only Os 2 displayed relatively high selectivity towards selective reduction of the CvC bond over the CvO bond.
To conclude, we tested the transfer hydrogenation of the nitro group.Recently, an efficient transfer hydrogenation catalytic system for the reduction of nitroarenes into anilines was reported and was based on Ru(II) and Ir(III) MIC precatalysts.8a We carried out an initial screening with nitrobenzene as a substrate, applying 1.5 mol% of the precatalysts and 0.5 equiv. of KOH in iPrOH at 80 °C.8a Azobenzene and azoxybenzene were expected as the side products.The results are presented in Table 11.The screening showed a catalyst dependent conversion, as well as product distribution.In contrast to the previously tested functional groups, the Ir(III) precatalysts were the most active of all the tested complexes, affording aniline as the major product.The Ru(II) complexes gave good conversion into aniline and azoxybenzene with only minute amounts of azobenzene.Once again the pyrimidine analogue Ru pyrim showed a General reaction conditions: trans-β-methylstyrene (0.5 mmol), precatalyst, KOH (0.085 mmol), iPrOH (4 mL), 100 °C for 24 h.
superior activity.The Os(II) complexes performed similar to the Ru(II) analogues, giving slightly higher conversions.The best performance was seen for Os 2 , where aniline was the main product.As was previously observed, 8a when employing 5.0 mol% of the precatalysts, we observed a drastic improvement in catalytic activity.The best results were seen with all four Ru(II) complexes, giving full conversion of nitrobenzene into aniline.The Ir(III) analogues performed very similarly.On the other hand, the Os(II) complexes were less active, except for Os pyrim , which yielded aniline quantitatively.

Mechanistic studies
Recently, a mechanistic study for a similar catalytic system was reported 8a where a two-pathway mechanism was postulated based on two previous literature reports. 19The mechanism is outlined in Scheme 4. The metal complex is converted into hydrides "[M]H 2 " in basic iPrOH.Next, nitrosobenzene is formed, which can be reduced into aniline through two different pathways.Route A proceeds through N-phenylhydroxyamine giving aniline directly.In contrast, path B involves an azoxybenzene intermediate, which gets converted to azobenzene before being reduced to aniline.To determine the reaction pathways for each of the complex families, we ran a time-dependent screening of the product formation.With the chosen precatalysts, we carried out the reduction of nitrobenzene under catalytic conditions with 1.5 mol% loading and analysed aliquots over time.The resulting data are represented in Fig. 1.With ruthenium complex Ru 1 54% conversion into aniline was observed after 32 h at 80 °C.While the amount of nitrobenzene decreased over time, the amounts of azoxybenzene and aniline increased.The formation of aniline, as well as azoxybenzene correlated well with the consumption of the starting nitrobenzene.Only minor amounts of azobenzene were traced after 32 h.The results point to pathway B. The iridium complex Ir 3 gave 84% conversion to aniline after 24 h with the amount of azoxybenzene less than 18%.The consumption of nitrobenzene correlates very well with the formation of aniline.These results indicate more preference for path A. With Os 2 74% conversion into aniline was observed with no azobenzene as a side product.The amount of aniline correlates very well with the consumption of nitrobenzene.The conversion into azoxybenzene was below 8% and quite constant after 24 h.This indicates the preference for pathway A.
For the transfer hydrogenation of the nitro functional group we chose four additional substrates: 2-, 3-and 4-nitroanisole, as well as 1-chloro-2-nitrobenzene.We selected Os 2 and Ir 3 as the precatalysts and conducted the transfer hydrogen-  ation with 1.5 mol% precatalyst loading.Each of the four substrates was stirred with 0.5 equiv. of KOH in iPrOH at 80 °C for 24 h (Table 12).2-Nitroanisole was converted with both tested complexes into o-anisidine as the only product.The conversions were 65% and 54%, respectively.Both complexes catalysed the reduction of 3-nitroanisole into m-anisidine very well.Ir 3 gave full conversion into the amino product.The reduction of 4-nitroanisole into p-anisidine proceeded with both complexes, however, giving lower conversions.By using the complex Ir 3 , 1-chloro-2-nitrobenzene was fully converted into 2-chloroaniline.In contrast, Os 2 gave only 58% of the aniline derivative after 24 h.Overall, the Ir(III) complex Ir 3 was more active in the transfer hydrogenation catalysis of the selected nitro substrates.Surprisingly no azo-or azoxy-side products were detected.For the reduction of cyclohexene and for the reduction of nitrobenzene, we performed the Hg-poisoning test by Hg during the catalysis.The addition of Hg had no effect on the yield or the rate of formation of the catalytic products.Hence, we believe that catalysis in the present case is likely homogeneous.

Conclusions
The scope of the pyridine-functionalized triazolylidene complexes of ruthenium, iridium and osmium as precatalysts for transfer hydrogenation was investigated.We have also presented the synthesis of the 1-(2-pyrimidyl)-functionalized triazolium salt and used it as a precursor for the carbenic ligand for coordination to ruthenium(II) and osmium(II).This is the first report of a pyrimidine-appended triazolium salt and also the first examples of ruthenium and osmium complexes with such MIC ligands.All tested complexes are very efficient ( pre-) catalysts for the reduction of several functional groups such as carbonyls, alkenes, imines and also nitroarenes.We demonstrated that in general the additional nitrogen atom of the pyrimidine ring is beneficial in terms of catalytic activity.The two pyrimidine-functionalized complexes have shown a better conversion in comparison to pyridine analogues, except for the reduction of CvC double bond substrates.The ruthenium and osmium complexes having triazolylidene ligands substituted with an electron-donating group performed better, whereas for the iridium complexes the effects were the opposite: complexes with electron-withdrawing groups are superior catalysts.In the final part we have shown some mechanistic studies for the reduction of nitrobenzene with three selected precatalysts, one for each metal (ruthenium, iridium and osmium).The Cp*-Ir-containing pre-catalysts presented here deliver better conversions for the conversion of CvO groups and comparable conversion for the reduction of CvC compared to the reported Cp*-Ir complexes with NHC ligands.2l  a General reaction conditions: nitro substrate (0.5 mmol), precatalyst (1.5 mol%), KOH (0.21 mmol), iPrOH (4 mL), 80 °C for 24 h.
b Conversion in % determined by GC using hexadecane as an internal standard.
On the other hand, the Cp*-Ir-containing complexes deliver more aniline through the reduction of nitrobenzene compared to related complexes.8a These results are important to understand the behaviour of the complexes during the catalysis.The results also show the potential of the tested complexes with chelating pyridyl-and pyrimidyl-triazolylidenes in the transfer hydrogenation of several functionalities, such as the carbonyl group, CvC double bond, imine and also the nitro group.The newly synthesized pyrimidyl-triazolylidenes might be useful ligands for generating both homo-and heterodinuclear complexes.

Materials and physical methods
The reagents and solvents were used as obtained from commercial sources (Sigma-Aldrich, Fluka, Alfa Aesar).Acetonitrile was dried over 3 Å molecular sieves, and dichloromethane over 4 Å molecular sieves.For methylation reactions dry glassware and solvents were used.Triazole 1 was prepared as described in the literature. 14NMR spectra were recorded with a Bruker Avance 300, Bruker Avance III 500 (Ljubljana) and Jeol ECS 400 (Berlin) spectrometers.Proton and carbon spectra were referenced to Si(CH 3 ) 4 as the internal standard.Some 13 C chemical shifts were determined relative to the 13 C signal of the solvent DMSO-d 6 (39.5 ppm). 19F NMR and 31 P NMR spectra were referenced to CCl 3 F and 85% phosphoric acid, respectively, as external standards at δ = 0.The nitrogen chemical shifts were extracted from 1 H- 15 N HMBC spectra with 20 Hz digital resolution in the indirect dimension. 15N chemical shifts are rounded to integer numbers because of the digital resolution limits of the experiment.The reported 15 N chemical shifts were extracted from 1 H- 15 N HMBC spectra (with 20 Hz digital resolution in the indirect dimension) with respect to external 90% CH 3 NO 2 in CDCl 3 and converted to the δ 15 N (liq.NH 3 ) = 0 ppm scale using the relationship: δ 15 N (CH 3 NO 2 ) = δ 15 N (liq.NH 3 ) + 380.5 ppm.Chemical shifts are given on the δ scale ( ppm).Coupling constants ( J) are given in hertz.The multiplicities are indicated as follows: s (singlet), d (doublet), t (triplet), q (quartet), sept (septet), m (multiplet) and br (broadened).Assignments of proton, carbon and nitrogen resonances were performed by standard 2D NMR techniques ( 1 H-1 H COSY, 1 H- 13 C HSQC, 1 H-13 C HMBC, and 1 H-15 N HMBC).The numbering used for the assignment of NMR signals is as follows: 1,2,3-triazole ring, simple figures; pyridine ring, primed figures; phenyl ring, double-primed figures (C-1″ is attached to the 1,2,3-triazole ring).An Agilent 6224 time-of-flight (TOF) mass spectrometer equipped with a double orthogonal electrospray source at atmospheric pressure ionization (ESI) coupled to an Agilent 1260 HLPC and an Agilent 6210 ESI-TOF spectrometer were used for recording HRMS spectra.Elemental analysis was performed with a Perkin-Elmer 2400 Series II CHNS/O Analyzer.IR spectra were obtained with a Perkin-Elmer Spectrum 100, equipped with a Specac Golden Gate Diamond ATR as a solid sample support.
Melting points were determined on a Kofler micro hot stage instrument.The reactions were monitored by TLC on TLC-CARD silica gel, 220-440 mesh. 1 H NMR spectra of benzyl alcohol, 1-phenylethanol, diphenylmethanol, cyclohexanol and cyclooctane were compared with those of the authentic samples (Tables 2-6).Propylbenzene 20 (Table 7), 1,2-diphenylethane 21 (Table 8) and N-benzylaniline 22 (Table 9) were identified by comparison of their 1 H NMR spectral data with the literature reports.GC-MS analysis was performed on a Varian Saturn 2100C (column: Varian factory four capillary column VF-5 ms; temperature: 50 to 250 °C, heating rate 20 K min −1 ).The calibration was done as follows: we collected the GC analyses of all the products, using pure samples.We obtained ret.times and also factors for calculating the areas of the GC peaks into conversions, percent yields.Ret.times (min):

Synthesis of the ligand
Pyrimidyl-triazole 1 in dry CH 2 Cl 2 was purged with dry argon and stirred on an ice-bath at 0 °C for a few minutes.MeOTf was added via syringe and the reaction mixture was stirred for 30 min at 0 °C and then 24 h at room temperature.The reaction mixture was concentrated using a rotary evaporator and the residue was column chromatographed on silica gel using a mixture of CH 2 Cl 2 and MeOH (10 : 1) as the eluent (R f ≈ 0.3).Re-crystallization from EtOAc-light petroleum afforded analytically pure triazolium salts 2OTf.Quantities used in the above procedure, analytical and spectral data of 2OTf are given below.

Synthesis of complexes
Ruthenium complex Ru pyrim .A mixture of a triazolium salt 2OTf and Ag 2 O in acetonitrile was purged with argon and stirred in the absence of light at room temperature for 4 days.After the addition of [Ru(η 6 -p-cym)Cl 2 ] 2 the stirring was continued for 2 hours.The reaction mixture was filtered through Celite.The solvent was removed in vacuo and the residue was crystallized from chloroform (1-2 mL) with slow addition of hexane to give the pure Ru pyrim .For quantities used, and analytical and spectral data of Ru pyrim PF 6 , see below.Osmium complex Os pyrim .Triazolium salt 2OTf was mixed with Ag 2 O and KCl and stirred in dry acetonitrile under a nitrogen atmosphere (in the absence of air), in the absence of light, at room temperature for 4 days.Later [Os( p-cym)Cl 2 ] 2 was added and the mixture was stirred for an additional 3 days under a nitrogen atmosphere, in the absence of light at room temperature.The reaction mixture was filtered over Celite and washed with dichloromethane.All the solvents were removed in vacuo.The residue was dissolved in methanol (10 mL), KPF 6 was added (10 equiv.)and the mixture was stirred for 10 minutes.Then water was added (100 mL).An orange precipitate was collected after 10 minutes of stirring by filtration and dried at room temperature.Pure Os pyrim PF 6 was obtained and fully characterized.For quantities used, and analytical and spectral data of Os pyrim PF 6 , see below.General procedure for the catalysed transfer hydrogenation Carbonyl group.The selected substrate (0.5 mmol), KOH (0.085 mmol), iso-propanol (4 mL) and the chosen precatalyst were mixed in a Schlenk tube under a nitrogen atmosphere.The reaction mixtures were stirred at 100 °C for 3 hours.The solvent was then removed in vacuo and the residue was analysed by 1 H NMR spectroscopy, using hexamethylbenzene as an internal standard, to identify products and determine conversions indicated in Tables 2-5.
Olefin group.The selected substrate (0.5 mmol), KOH (0.085 mmol), iso-propanol (4 mL) and the chosen precatalyst were mixed in a Schlenk tube under a nitrogen atmosphere.The reaction mixtures were stirred at 100 °C for 24 hours.The solvent was then removed in vacuo and the residue was analysed by 1 H NMR spectroscopy, using hexamethylbenzene (for trans-stilbene) or 4-bromobenzaldehyde (for trans-β-methyl-styrene and cyclooctene) as an internal standard, to identify products and determine conversions indicated in Tables 6-8.
Imine group.N-Benzylideneaniline (0.5 mmol), K 2 CO 3 (0.025 mmol), iso-propanol (4 mL) and the chosen precatalyst were mixed in a Schlenk tube under a nitrogen atmosphere.The reaction mixtures were stirred at 100 °C for 18 hours.The solvent was then removed in vacuo and the residue was analysed by 1 H NMR spectroscopy, using hexamethylbenzene as an internal standard, to identify products and determine conversions indicated in Table 9.
Nitro group.The selected substrate (0.5 mmol), KOH (0.21 mmol), iso-propanol (4 mL) and the chosen precatalyst were mixed in a Schlenk tube under a nitrogen atmosphere.The reaction mixtures were stirred at 80 °C for 24 hours.The crude reaction mixture was diluted with ethyl acetate and subjected to GC to identify products and determine conversion indicated in Tables 10 and 11.Hexadecane was used as an internal standard.

Table 1
The synthesis of Ru pyrim and Os pyrim complexes

Table 6
The reduction of cyclooctene a

Table 7
The transfer hydrogenation of trans-β-methylstyrene a

Table 9
The transfer hydrogenation of N-benzylideneaniline a

Table 11
The transfer hydrogenation of nitrobenzene a

Table 12
The transfer hydrogenation reduction of four selected nitroarenes a