Chemo- and regioselective oxygenation of C(sp3)–H bonds in aliphatic alcohols using a covalently bound directing activator and atmospheric oxygen

Aerobic, site-selective C(sp3)–H oxygenation using a novel N-oxyl radical directing activator (chemically reactive directing group) is described.


The procedure for gram-scale oxidation and the removal of the directing activator
There is difference from the typical procedure in the equivalent of the reagents.
With A, benzylic C-H oxygenation produced the corresponding phenyl ketone product in only 8% yield.
A more challenging methylene C(sp 3 )-H oxygenation did not proceed at all. In both cases, the efficiency of the C-H oxygenation was much lower than that of N-hydroxyphthalimide itself. Because the reactivity of N-oxyl radicals is in accord with their electron-deficiency, we next designed B bearing a CF 3 group at the α-carbon of the nitrogen atom to enhance the reactivity of the DA. As a result, B promoted simple methylene C(sp 3 )-H oxygenation to produce γ-oxo product in 67% yield. Although the α-C(sp 3 )-H bond adjacent to the ether oxygen atom is the innate reactive site, the less reactive γ-C(sp 3 )-H was predominantly oxygenated. Under the same conditions, benzylic C(sp 3 )-H oxygenation also proceeded well (85% yield). Introducing an additional CF 3 substituent at the benzene ring (C), however, did not further improve the yield. Additional structural modifications (D, E, F, or G) were not successful, resulting in complex mixtures and/or decomposition of the DA during the reaction. Thus, B proved to be the best DA.

Secondary C(sp 3 )-H oxygenation
Solvents, metal salts, and additives were screened for higher yields of secondary C(sp 3 )-H oxygenation.
Tested additives such as some acids, bases, molecular sieves, and reductants all brought no improvements.
The screening of the solvents for 1-pentanol oxygenation is shown below. As shown in Table S2, TFE (2,2,2-trifluoroethanol) and HFIP (1,1,1,3,3,3-hexafluoroisopropanol) accelerated the reaction to almost the same degree and produced almost the same yields, so the cheaper TFE was employed as the best solvent. S16 Then, metal salts were screened using B-anchored 1-butanol instead of B-anchored 1-pentanol because B-anchored 1-butanol gave only one regioisomer and made the analysis easier.
As shown in Table S3, the combination of Co(OAc) 2 and Mn(OAc) 3 ·2H 2 O produced the best result. The role of Mn(OAc) 3 ·2H 2 O is not clear, but we suppose that the main role of Mn(OAc) 3 ·2H 2 O is to trap reversibly reactive intermediates such as peroxy radicals and alkoxy radicals to suppress unfavorable side reactions, considering the fact that the reaction rate became lower by adding Mn(OAc) 3 ·2H 2 O than when Co(OAc) 2 alone was used as a metal source even though Mn(OAc) 3 ·2H 2 O itself was able to catalyze the reaction.
The 1-butanol oxygenation was conducted at 30, 40, or 50 °C. The reaction didn't complete at 30 °C, and became slightly complicated at 50 °C compared to that at 40 °C. Therefore, 40 °C was revealed to be the best reaction temperature.

Tertiary C(sp 3 )-H oxygenation
Although 67% NMR yield was obtained for 1-butanol oxygenation under condition A, the reaction became complicated and only 20% NMR yield was obtained for 3-methyl-1-butanol oxygenation (1h) under the same conditions. It was probably because condition A was too harsh for tertiary C(sp 3 )-H oxygenation, so the reaction conditions had to be refined for higher yields. The goal was accomplished by reducing the amount of the metal salts. Lower reaction temperatures and different solvents were not effective. The best reaction conditions were these: Co(OAc) 2 (1 mol%), Me 2 S (1.2 eq), O 2 (1 atm), TFE (0.1 M), 40 °C.

Absolute configuration of DA-anchored (+)-menthol (1k)
X-ray analysis of less polar diastereomer of O-(4-nitrobenzyl)-1k was conducted to determine the absolute configuration of DA-anchored (+)-menthol. The configuration of the hemiaminal center was revealed to be (R), which accords with that of reactive isomer in aerobic C-H oxygenation.
Crystallization procedure: less polar diastereomer of O-(4-nitrobenzyl)-1k was dissolved in MeCN and the solvent was gradually evaporated under air at room temperature. After a few days a needle crystal was obtained.
The supposed reason for the difference in the reactivity is depicted in the figure below. S18 According to molecular modeling, there is no steric repulsion during the abstraction of the red hydrogen in (R)-1k and it proceeds smoothly; whereas in the case of (S)-1k there is a steric repulsion between CF 3 and the axial α-hydrogen of the oxygen atom in menthol that makes it difficult for the N-oxyl radical to access the red hydrogen.

Comparison of reaction speed in benzylic CH oxygenation
The chemical yield after 30 min was compared for --, and -CH oxygenation. Reaction speed of -C-H oxygenation is significantly slower than and -C-H oxygenation. The difference between and -C-H oxygenation is, however, not large. This low / selectivity is a current limitation of our method.

Synthesis of substrate 3 Synthetic procedure from S5 to S6
To a round bottom flask were added S5 5 (2.15 g, 8.3 mmol), p-acetylaniline (1.07 g, 7.9 mmol) and pyridine (42 mL, 0.2 M). The mixture was stirred at reflux for 18 h. Pyridine was removed by evaporation.
The obtained crude mixture was purified by silica gel column chromatography (EtOAc/hexane = 2/1) to afford S6 with inseparable byproducts as a colorless sticky liquid (2.98 g). This mixture was used for next step.

Synthetic procedure from S7 to S8
To solution of ketone S7 (288 mg, 0.588 mmol) in CH 2 Cl 2 (2.94 mL, 0.2 M) were added triethylamine (0.590 mL, 4.23 mmol) and TBSOTf (0.494 mL, 2.82 mmol). The mixture was stirred at room temperature for 40 min. After ketone was consumed, which was checked on TLC, pentane and H 2 O were added. The organic layer was separated and washed with 10% aqueous CuSO 4 solution and 5%NaHCO 3 aq. solution, and then dried over Na 2 SO 4 . Filtration followed by evaporating the solvent at room temperature afforded the crude mixture containing the corresponding enol silyl ether.

Oxygenation of 3 (ultra-remote position-selective aerobic C-H oxygenation)
To a test tube were added 3 (30.0 mg, 41.7 mol) and TFE (0.42 mL, 0.1 M). Me 2 S (6.71 μL, 91.8 mol) was added after replacing the air inside the tube with O 2 (1 atm, balloon), and finally Co(OAc) 2 (0.2 M solution in DMSO, 2.1 μL, 0.42 mol) was added. The mixture was stirred at room temperature. After 1.5 h, Co(OAc) 2 (0.2 M solution in DMSO, 2.1 μL, 0.42 mol) and Me 2 S (6.71 μL, 91.8 mol) were added. After 3 was consumed, which was checked on TLC, TFE was removed by evaporation and water was added. The aqueous layer was extracted by EtOAc three times. The combined organic layer was dried over Na 2 SO 4 .

Structural anaylsis of oxygenated products
After aerobic oxygenation of 3, 1 H NMR chart of the crude mixture was carefully analyzed. No triplet peak was observed around 3.00-3.50 ppm, which corresponds to the -methylene protons of phenyl ketone 4' (marked in red in 4' of the above scheme), while a singlet peak at 2.58 ppm was observed, which corresponds to the -methyl protons of phenyl ketone of 4 (marked in red in 4). For comparison, oxidized S23 products of model substrate S11 afforded -methylene protons of phenyl ketone S12' observed at 3.35 ppm as triplet (marked in red in S12' of the scheme below), and α-methyl protons of ketone S12 observed at 2.58 ppm as singlet (marked in red in S12). This fact indicates that regioisomer 4' did not generate at all.
Thus, we concluded that the oxygenation of 3 occurred exclusively at the remote benzylic position.
The intermolecular oxygenation of model compound S11 produced a 2 : 1 regiomixture of S12 and S12' at 60 °C. This result, combined with the contrasting and exclusive remote-regioselectivity in oxygenation of 3, further supported the notion that oxygenation of 3 proceeded in an intramolecular manner controlled by the "long-arm linker" directing activator.