Radical hydroxymethylation of alkyl iodides using formaldehyde as a C1 synthon

Radical hydroxymethylation using formaldehyde as a C1 synthon is challenging due to the reversible and endothermic nature of the addition process. Here we report a strategy that couples alkyl iodide building blocks with formaldehyde through the use of photocatalysis and a phosphine additive. Halogen-atom transfer (XAT) from α-aminoalkyl radicals is leveraged to convert the iodide into the corresponding open-shell species, while its following addition to formaldehyde is rendered irreversible by trapping the transient O-radical with PPh3. This event delivers a phosphoranyl radical that re-generates the alkyl radical and provides the hydroxymethylated product.

Purification by flash column chromatography on silica gel gave 32 (6. were sequentially added. The tube was place in front of the blue LEDs (approx. 10 cm) and the lights were switched on. The mixture was stirred under continuous irradiation for 16 h whilst being cooled by a fan to give an internal temperature between 25-30 °C. The lights were switched off and the tube was opened. The mixture was diluted with brine (2 mL) and EtOAc (2 mL), then 1,3-dintrobenzene (1 mL, 0.05 M solution in EtOAc) was added as an internal standard and the mixture was vigorously shaken. The layers were separated and the aqueous layer was extracted with EtOAc (x 2). The combined organic layers were dried (MgSO4), filtered and evaporated. The crude was solubilised in CDCl3 and analysed by 1 H NMR spectroscopy. To further improve the efficiency of the process we have used the statistical software Ellistat for DoE. We investigated the effects of varying equivalents of amine, PPh3, HCHO and H2O leading us to the conditions reported in Scheme S1. Scheme S1.
Further screening was performed using the conditions described below (Scheme S2) according to GP2, which is detailed in Table S2.
The crude was solubalised in CDCl3 and analysed by 1 H NMR spectroscopy.  The mixture was diluted with brine (2 mL) and EtOAc (2 mL), then 1,3-dintrobenzene (1 mL, 0.05 M solution in EtOAc) was added as an internal standard and the mixture was vigorously shaken. The layers were separated, and the aqueous layer was extracted with EtOAc (x 2). The combined organic layers were dried (MgSO4), filtered and evaporated. The crude was solubilised in CDCl3 and analysed by 1 H NMR spectroscopy.
The tube was place in front of the blue LEDs (approx. 10 cm) and the lights were switched on.
The mixture was stirred under continuous irradiation for 16 h whilst being cooled by a fan to give an internal temperature between 25-30 °C. The lights were switched off and the tube was opened. The mixture was diluted with brine (2 mL) and EtOAc (2 mL), then 1,3-dintrobenzene (1 mL, 0.05 M solution in EtOAc) was added as an internal standard and the mixture was vigorously shaken. The layers were separated and the aqueous layer was extracted with EtOAc (x 2). The combined organic layers were dried (MgSO4), filtered and evaporated. The crude was solubilised in CDCl3 and analysed by 1 H NMR spectroscopy. Table S5. 16

Stern-Volmer Quenching Studies
Stern-Volmer experiments were carried out monitoring the emission intensity of argondegassed solutions of 4CzIPN (3 x 10 -5 M solution in CH3CN) containing variable amounts of the quencher in dry acetonitrile. The reported excited-state lifetime for 4CzIPN in CH3CN (1.4 s) 9 was used for kq calculations (see Table S6). These experiments show the i-Pr2NEt quenches *4CzIPN at faster rates than any other reagent ( Figure S2 and Table S6).

Ruling Out the Formation of Electron Donor-Acceptor (EDA) Complexes
To rule out the formation of EDA complexes between the alkyl iodide and the amine that might be absorbing in the visible region, we have performed UV/Vis absorption spectroscopy studies ( Figure S3). These studies demonstrate that there is not EDA complexation between the amine and the alkyl iodide.

Evidences Supporting XAT by the Phosphoranyl Radical
In order to obtain supporting evidences on the ability of phosphoranyl radicals to sustain a chain-propagation based on XAT, we evaluated the hydroxymethylation of 1 in the absence of amines as well as any other possible reductant (e.g. *4CzIPN or 4CzIPN •-).
We speculated that by treatment of 1, HCHO and PPh3 with (t-BuO)2, the phosphoranyl radical Under these conditions we monitored the reaction by NMR to detect formation of MeI or acetone, formed from β-fragmentation of the phosphoranyl radical X. Neither of these sideproducts were detected, ruling out the possibility that Me• was generated and acted as XAT agent/initiator. solubilised in CDCl3 and analysed by 1 H NMR spectroscopy.

Quantum Yield () Determination
The quantum yield () of the photochemical hydroxymethylation reaction of 1 was determined at 50 ºC following procedures described in literature (Scheme S6). 10 Elevated temperatures were required to accelerate the reaction for quantum yield determination, as it was not possible to record an accurate quantum yield at room temperature due to the reaction progressing slowly.
The degassed reaction tube was irradiated using blue LEDs plates (λmax = 444 nm) and product yield was determined by 1 H NMR spectroscopy analysis. The photon flux of the blue LEDs used was determined by standard ferrioxalate actinometry. 11

Scheme S6.
Reactions where a radical chain propogations is present are typically expected to provide a  > 1. In our case we have observed that the hydroxymethylation reaction displays a signifcant induction time that might account for the low  observed ( Figure S4).

Hydroxymethylation of Alkyl Bromide 30
We have evaluated the reactivity of alkyl bromide 30 under our standard conditions since aminoalkyl radical-mediated XAT is feasible. 12 However, the desired product 2 was obtained in low yield with remaining 30 accounting for the remaining mass balance (Scheme S7).

Scheme S7.
Despite considerable efforts aimed at optimising this reactivity changing all reaction parameters we did not succeed in engaging this class of derivatives in higher yield.
According to our proposed mechanism we speculated that there might have been an issue with one of the two XAT steps: either the one mediated with the -aminoalkyl radical or the one mediated by the phosphoranyl radical. Various alkylamines were screened to evaluate their impact in the reactivity (Table S7). Table S7.

SI-27
As shown in Table S7, this led to no improvement in the reaction yield. Interestingly, we detected the hydroxymethylation of some amines by mass spectrometry analysis of the reaction crudes. This suggests that in the case of the alkyl bromides were XAT is slower, the nucleophilic -aminoalkyl radical can trap HCHO and undergo PPh3-mediated hydroxymethylation. To obtain further evidences on this reactivity we have performed DFT studies to determine the reaction parameters for the addition step. As shown in Scheme S8, the reaction ofan-aminoalkyl radical derived from Et3N should undergo a feasible addition to HCHO.

Scheme S8.
Depending on the structure of the amine, mono-or tri-hydroxymethylation was observed. When irreversible waves were obtained the potentials were estimated at half the maximum current, as previously described by Nicewicz. 14

Computational Methods
Density functional theory (DFT) 30 calculations were performed using Gaussian 09 (revision E.01) 31  index, B3LYP functional. 32b, 32d, 36 was used and the geometry of studied radical was optimized at the UB3LYP/6-311+G(d,p) level of theory, followed by frequency calculation at the same level. 37 The computed Hirshfeld charges on the radicals were also calculated at the same level of theory. 38 All stationary points were characterized as minima based on normal vibrational mode analysis. Thermal corrections were computed from unscaled frequencies, assuming a standard state of 298.15 K and 1 atm. SI-38