Hydrogenation of Secondary Amides using Phosphane Oxide and Frustrated Lewis Pair Catalysis

Abstract The metal‐free catalytic hydrogenation of secondary carboxylic acid amides is developed. The reduction is realized by two new catalytic reactions. First, the amide is converted into the imidoyl chloride by triphosgene (CO(OCCl3)2) using novel phosphorus(V) catalysts. Second, the in situ generated imidoyl chlorides are hydrogenated in high yields by an FLP‐catalyst. Mechanistic and quantum mechanical calculations support an autoinduced catalytic cycle for the hydrogenation with chloride acting as unusual Lewis base for FLP‐mediated H2‐activation.


Substrate Synthesis
N-Methylbenzamide (1d) and N-Phenylbenzamide (1f) were commercially available and used in further reactions without purification. N-Isopropyl-4-bromobenzamide (1a), N-Isopropylbenz-  and N-Isopropyl-4-nitrobenzamide (1l) were prepared by the general method. NMR and mass spectral data were in good agreement with the literature.

General Method for Synthesis of N-Substituted Carboxamides [SI4]
The primary amine (11.0 mmol, 1.10 equiv.) and triethylamine (12.5 mmol, 1.25 equiv.) were dissolved in DCM (12 mL, 0.83 M) and placed in a 20 mL crimp seal glass vial. The reaction mixture was cooled in an ice/water bath and the acyl chloride (

Synthesis of N-Isopropyl-4-(phenylethynyl)benzamide (1i) [SI5]
N-Isopropyl-4-bromobenzamide (1a) was prepared according to the general procedure and was used for the next step without further purification. Tetrakis(triphenylphosphine)palladium (116 mg, 100 μmol, 1.00 mol%) and copper(I) iodide ( aqueous NH4Cl and extracted twice with DCM. The combined organic phases were washed with sat. aqueous NaCl and then dried over MgSO4. After removing the volatiles under reduced pressure, the crude product was purified by column chromatography (silica, CH/EA 5/1). Further purification was achieved by recrystallization from ethyl acetate. The product was obtained as a white solid. Yield: 29% (770 mg, 2.92 mmol).

Synthesis of N-Benzoyl-2-aminoethyl cinnamate (1j) [SI6]
N-(2-Hydroxyethyl)benzamide was prepared according to the general procedure starting from 2-aminoethanol and benzoyl chloride and was used for the next step without further purification.
After full conversion (TCL monitoring, 20 h) aqueous HCl (1 M) was added. The reaction mixture was extracted three times with ethyl acetate, and the combined organic phases were washed with aqueous HCl (1 M), sat. aqueous Na2CO3 and water. After drying over MgSO4 the volatiles were removed under reduced pressure. The crude product was purified by column chromatography (silica, CH/EA 2/1), which yielded the product as a light brown solid. Yield: 53% (1.56 g, 5.29 mmol).
The process was monitored by 1 H-NMR spectroscopy. The volatiles were removed afterwards, and the product was used without further purification for the catalyst screening (see section 7.1).

Optimization of Imidoyl Chloride Formation
In a glovebox, N-Isopropyl-4-bromobenzamide (1a) (12.1 mg, 50.0 μmol, 1.00 equiv.), the reagent (20 μmol -500 μmol, 0.4 equiv. -10 equiv.) and, if necessary, the catalyst (10 μmol, 20 mol%) were dissolved in 0.6 mL CDCl3 and transferred to a J. YOUNG NMR tube. The sample was heated on a shaking plate to 70 °C or 90 °C, and the process was monitored by 1 H-NMR spectroscopy. In a glovebox, a borane catalyst (20 µmol, 20 mol%) and N-Isopropylbenzimidoyl chloride (2b) (100 µmol, 1.00 equiv., synthesized according to section 5) were dissolved in 0.6 mL CDCl3 and transferred to a J. YOUNG NMR tube with Teflon tap. The sample was then frozen in liquid nitrogen, the headspace was evacuated, and the sample was charged with hydrogen at -196 °C. After sealing and thawing, the hydrogen pressure inside the sample reached approximately 4 bar. The sample was then heated on a shaking plate to ensure hydrogen exchange. After the given time (see Table S2) the NMR tube was cooled to room temperature, and the crude reaction mixture was analyzed by 1 H-NMR spectroscopy.

bar H2:
In a glovebox, B(2,3,6-F3-C6H2)3 (2.0 µmol, 2.0 mol%) and N-Isopropylbenzimidoyl chloride (2b) (100 µmol, 1.00 equiv.) were dissolved in 0.6 mL CDCl3. The sample was transferred to a Millireactor, charged with 80 bar hydrogen, and heated in an oil bath. After the given time (see Table S2), the crude reaction mixture was transferred to an NMR tube and analyzed by 1 H-NMR spectroscopy. a) according to Gutmann-Beckett with B(C6F5)3 referenced to 100% Lewis acidity, b) determined by 1 H-NMR spectroscopy; for yield calculation, signals of starting material before the reaction and of the product after the reaction were integrated and normalized against the silicone grease signal as an internal standard (example given below).
calculation example (

Effect of Additives on Yield of the FLP-Catalyzed Hydrogenation of Imidoyl Chloride
In a glovebox, the imidoyl chloride 2a (13.0 mg, 50.0 μmol, 1.00 equiv.), B(2,3,6-F3-C6H2)3 (4.0 mg, 10 µmol, 20 mol%) and an additive (10 µmol, 0.2 equiv.) were dissolved in 0.6 mL CDCl3 and transferred to a J. YOUNG NMR tube with Teflon tap. The sample was then frozen in liquid nitrogen, the headspace was evacuated, and the sample was charged with hydrogen at -196 °C. After sealing and thawing, the hydrogen pressure inside the sample reached approximately 4 bar. The sample was then heated on a shaking plate to ensure hydrogen exchange. After 20 h, the crude reaction mixture was analyzed by 1 H-NMR spectroscopy. The effect of phosphine oxides, activation agents or amide on the FLP-catalyzed hydrogenation was shown. The phosphine oxides 3c and 3d had no interference with the borane in the FLP-catalyzed hydrogenation (see Table S3), but the conversion of amide to imidoyl chloride is more efficient with the phosphine oxide 3d.
CHCl3. The sample was charged with hydrogen (80 bar), and the stainless-steel high-pressure reactor was then heated to 70 °C or 90 °C in an oil bath. After 20 h or 40 h, the reactor was cooled to room temperature and depressurized. The product mixture was diluted with DCM and washed with aq.
ammonia and sat. aq. Na2CO3. The phases were separated, and the aqueous phase was extracted three times with DCM. The combined organic phases were loaded with a small portion of silica and evaporated to dryness. After column chromatography, the amine was analyzed by NMR spectroscopy and mass spectrometry.

N-Isopropyl-(4-bromobenzyl)amine (5a)
According to the general procedure 7.3, the activation was performed at 90 °C for 5 h, and the reduction was performed with 5 mol%      amine (5k) According to the general procedure 7.3, the activation was performed at 90 °C for 1 h, and the reduction was performed with 5 mol%
Adduct formation of phosphine oxides and B(2,3,6-F3C6H2)3 was observed. The broad signals in the 19 F-and 31 P-NMR spectrum indicate, that the phosphine oxide 3d has the weakest interaction with the borane. These results are in accordance with 7.2, since 3a and 3b had a significant impact on the FLP-catalyzed reduction of an imidoyl chloride, whereas 3c and 3d did not. 11 B-NMR: 19 F-NMR: 31 P-NMR:
A broadening of the signals in the 19 F-NMR spectrum and a change of chemical shifts in the 11 B-and 19 F-NMR spectrum was observed. This indicates that the chloride is reversibly bound to B(2,3,6-F3-C6H2)3. (4.0 mg, 10 µmol, 10 mol%) and hexamethylbenzene (HMB) (1.0 mg, 6.2 µmol, 6.2 mol%), as internal standard, were dissolved in 0.6 mL CDCl3 and transferred to a J. Young NMR tube with Teflon tap. An equal sample without the hydrochloride was prepared as a reference. The samples were then frozen in liquid nitrogen, the headspace was evacuated, and the J. Young NMR tubes was charged with hydrogen at -196 °C. After sealing and thawing, the hydrogen pressure inside the J.
Young NMR tubes reached approximately 4 bar. The samples were then heated on a shaking plate to ensure gas exchange. Every 30 minutes (later: every hour) the NMR tubes were cooled to room temperature and analyzed by 1 H-NMR spectroscopy. The product concentration was determined by integration against the internal standard. Figure S1: Product formation over time for 10 mol% B(2,3,6-F3-C6H2)3 (4e), 20 mol% ammonium chloride 5b and 4 bar H2 at 70 °C (red squares). Reference without ammonium chloride 5b (blue diamonds).

Hydrogenation of Imidoyl Chloride 2b Without Added Ammonium Hydrochloride
In a glovebox, N-Isopropylbenzimidoyl chloride (2b) (18.2 mg, 100 µmol, 1.00 equiv.), B(2,3,6-F3-C6H2)3 (4e) (4.0 mg, 10 µmol, 10 mol%) and hexamethylbenzene (HMB) (1.0 mg, 6.2 µmol, 6.2 mol%), as internal standard, were dissolved in 0.6 mL CDCl3 and transferred to a J. Young NMR tube with Teflon tap. The sample was then frozen in liquid nitrogen, the headspace was evacuated, and the J. Young NMR tube was charged with hydrogen at -196 °C. After sealing and thawing, the hydrogen pressure inside the J. Young NMR tube reached approximately 4 bar. The sample was then transferred to a pre-heated NMR spectrometer and placed inside the spectrometer at 60 °C. The spectrometer then proceeded to measure a 1 H-NMR spectrum every 1200 seconds for 16 h. The 1 H-NMR spectra were then analyzed, and the product concentration was determined by integration against the internal standard.  The D3 S11 (incorporated in the PBEh-3c composite method) and D4 S12-14 London dispersion correction schemes applying Becke-Johnson (BJ) damping S15,S16 and including Axilrod-Teller-Muto (ATM) S17,S18 type three-body contributions to the total London dispersion energy were applied. For a review on this topic see Ref. S19 Ro-vibrational corrections to obtain free energies were obtained from a modified rigid rotor harmonic oscillator statistical treatment S20 (T = 70.0 °C, 1 atm pressure) based on harmonic frequencies calculated at the geometry optimization level (PBEh-3c(COSMO(CHCl3))). To avoid errors in the harmonic approximation, frequencies with wave numbers below 100 cm -1 were treated partially as rigid rotors. S20 Gas phase single point energies were calculated at PW6B95-D4/def2-QZVP level applying the m5 numerical quadrature grid. Final Gibbs free energies were obtained by summing the gas phase single point energy E, the dispersion correction EDisp., the ro-vibrational correction GRRHO and the solvation correction δGsolv (Eq. S1).