(Distinctive Doe)
09-05-02 10:48
No 353072
      Aromatic Allylation via Diazotization
(Rated as: excellent)

This is very cool!!
To bad it only works well on electron deficient aromatics.
Is there an easy way to convert aromatic nitro substituents into other things?

Aromatic Allylation via Diazotization: Metal-Free C-C Bond Formation
Fredrik Ek, Oskar Axelsson, Lars-Göran Wistrand, and Torbjörn Frejd
J. Org. Chem., 67 (18), 6376 -6381, 2002.

A new method for the synthesis of allyl aromatic compounds not involving any metal-containing reagent or catalyst has been developed. Arylamines substituted with a large number of different substituents were converted via diazotizative deamination with tert-butyl nitrite in allyl bromide and acetonitrile to the corresponding allyl aromatic compounds. The allylation reaction was found to be suitable for larger scale synthesis due to short reaction times, a nonextractive workup, and robustness toward moisture, air, and type of solvent.

The reactions presented in Table 1 were performed on a 3 mmol scale. The arylamine (solid) was added in portions to a dry and oxygen-free acetonitrile solution of tert-butyl nitrite and allyl bromide during approximately 20 min, while the temperature was maintained as shown in Table 1. 26 Warning! A too rapid addition of the arylamine to the reaction mixture may result in an uncontrolled evolution of heat and nitrogen gas. A reaction time of 60 min was employed, although in several cases complete conversion was achieved already ca. 5 min after the final addition of the arylamines. An extractive workup was not necessary, in contrast to the methods using metal-containing reagents and catalysts. Subsequent removal of volatile compounds from the reaction mixture at reduced pressure gave the crude product, which was then purified by methods outlined in the Experimental Section. A slightly modified procedure was used for the synthesis of 1h on a multigram scale (for details see the Experimental Section).

As seen in Table 1, allylated products were isolated in yields ranging from 0 to 85% on the basis of the corresponding arylamines.28 In accordance with the typical behavior of radical reactions, a large number of functionalities were tolerated.13,14 Electron-deficient arylamines generally had a higher reactivity and gave higher yield of the allylated products. A similar electronic effect on reactivity has been reported in radical addition reactions involving allyl, vinyl, and aryl substrates and also in radical abstraction of hydrogen or halogen atoms.21,25,29,30 The common explanation for this behavior is that electron-withdrawing groups convert the phenyl radical (a probable intermediate in the mechanism of the allylation reaction, which will be discussed later) from a relatively nonpolar to an electrophilic radical.21,30-32

Similar to the behavior of other radical addition reactions, steric hindrance has an adverse effect on the outcome of the allylation reaction.33-35 When 1o was subjected to the standard conditions no evolution of nitrogen gas could be detected. Even when the temperature was increased to >60 C, the corresponding allylated product could not be isolated. Also, 4p was more unreactive than 4h. According to HPLC analysis, only 20% of 4p was consumed after 1 h when the allylation reaction was performed at 60 C. However, extended reaction time (18 h) and addition of extra tert-butyl nitrite (8 equiv was added in portions during 4 h) gave almost complete conversion of 4p, although the isolated yield of 1p was only 49%. Similarly, the steric bulk of the benzoyl substituent in 4d is probably the reason for the moderate yield of the allylated product 1d.

Electron-withdrawing substituents not only facilitated the allylation reaction but also made the corresponding phenyl radical more selective for the addition to allyl bromide relative to abstraction of a bromine atom, the most important side reaction. As seen in Table 1, the more electron-deficient arylamines generally gave a higher ratio of allylation versus bromination. Additionally, steric hindrance appeared to increase the tendency of bromination as exemplified in entries 4 and 16 (Table 1).

Finally, a method for the large-scale (0.41 mol) synthesis of 1h was developed. A smaller amount of acetonitrile was used in order to limit the total volume of the reaction mixture. The addition time of the arylamine had to be prolonged because of the exothermicity of the reaction. An increased yield compared to the same reaction at a 3 mmol scale was noticed. Filtration of the crude product dissolved in toluene through two pads of neutral alumina gave a product that was pure enough to be crystallized from isooctane. Thus, chromatography, which would not be acceptable in a larger scale procedure, was avoided.

Changes in the reaction conditions such as concentration, moisture levels, and temperature only moderately affected the yield (66-79%) and the selectivity (20-25:1).26 The reaction proceeded even in allyl bromide as solvent with minor changes in the outcome (yield 70%, selectivity 15:1). Unexpectedly, the replacement of tert-butyl nitrite with isoamyl nitrite lowered the yield with 11%. A possible explanation could be the greater tendency of homolytic cleavage of the secondary carbon-hydrogen bonds in the isoamyl unit.36 Further, to see if the phenyl radical was able to select between allylic substrates with and without a suitable terminal leaving group, equal amounts allyl bromide and allyl acetate were employed. (OAc is considered to be a poor leaving group in radical reactions due to the strong C-O bond).20 The yield dropped from 74% to 45%. When only allyl acetate was used in the reaction the allyl product could not be detected. Instead, a multitude of nonidentified compounds was formed according to HPLC. These experiments indicated that the phenyl radical attacks both allylic components indiscriminately and that only attack on the bromide leads to allylation.

In our standard method, a large amount (15 equiv) of allyl bromide was used. The yield of the desired allylation product decreased considerably at lower than 7.5 equiv of allyl bromide. However, most of the excess of allyl bromide could be recovered by distillation directly from the reaction mixture; the distilled allyl bromide contained acetonitrile and traces of tert-butyl nitrite. Thus, approximately 10-11 equiv of allyl bromide could be isolated and reused after the reaction, indicating that 4-5 equiv was consumed. We also noted that while the yield of the allylation product was quite dependent on the amount of allyl bromide the yield of the side product 5h was only marginally effected (±1%).

In agreement with the general behavior of radical reactions, we found only small variations in the product distribution when changing the solvent. Acetonitrile (74%, 25:1) and nitromethane (75%, 20:1) were found to be the best solvents, although both acetone and dimethoxyethane could be applied with minor alteration of the yield and selectivity (69%, 20:1). DMSO and tetrachloromethane gave somewhat lower yield (55-60%) and showed a poorer selectivity (10:1). THF gave the same yield of the allylated product as acetone but resulted in the lowest ratio of allylation versus bromination (8:1).

The mechanism of the aprotic diazotization of arylamines is not yet fully understood. However, we believe that the mechanism for the Gomberg-Bachmann (GB) reaction, proposed by Rüchhardt et al.,37 is probably also applicable to aprotic diazotization of arylamines, including the allylation reaction (Scheme 2). We have been able to isolate a reactive precipitate from the reaction mixture formed during the addition of 4h. It slowly dissolved when the addition of 4h was completed.38 Analysis of the isolated polar precipitate (Warning! The precipitate violently decomposed upon heating), with TOF HRMS (APCI-) (acetonitrile) and 1H NMR (DMSO-d6) suggested that the material was similar to 10, thus supporting the GB mechanism. This material is likely to be in a pH-dependent equilibrium with the corresponding diazonium salt 9 and the diazotate 11.39,40 Furthermore, addition of the precipitate to allyl bromide in acetonitrile resulted in the evolution of nitrogen gas and formation of 1h and the byproducts normally found in the allylation reaction. Similar to the mechanism postulated for the GB reaction, 9 and 11 then form the diazotic anhydride (12) which, in turn, decomposes and gives nitrogen gas, phenyl radical (14), and the long-lived diazotyl radical (13).41,42 The diazotyl radical probably abstracts a hydrogen atom or an electron from the reaction mixture, thus regenerating the diazotic acid/diazotate system (9-11). There is also a possibility that the diazotyl radical (13) may form a dimer, which then could decompose into nitrogen, oxygen, and two phenyl radicals.

General Procedure for the Allylation.
The arylamine (3.0 mmol) was added during 20 min to a solution of tert-butyl nitrite (535 L, 4.5 mmol) and allyl bromide (3.9 mL, 45.0 mmol) in dry and degassed CH3CN (3 mL) under argon atmosphere while maintaining the specified temperature (Table 1). At the end of the addition of arylamine, extra tert-butyl nitrite (180 L, 1.5 mmol) was added. The reaction mixture was then stirred at a temperature specified in Table 1 for 1 h. The volatile material in the reaction mixture was then removed at reduced pressure.

Allyl-2-chloro-4-nitrobenzene (1a).
Column chromatography (heptane-ethyl acetate 97:3) followed by preparative HPLC (gradient solution: 50:50 to 70:30 for 30 min and then 95:5 (CH3CN/H2O)) gave 467 mg (79%) of 1a as a pale yellow oil.

Allyl-4-nitro-2-trifluoromethylbenzene (1b).
Column chromatography (heptane-ethyl acetate 49:1) gave 587 mg (85%) of 1b as a pale yellow oil

Allyl-2-cyano-4-nitrobenzene (1c).
Column chromatography (heptane-ethyl acetate 23:2) gave 367 mg (65%) of 1c as a pale yellow oil.

2-Allyl-5-nitrobenzophenone (1d).
Column chromatography (heptane-ethyl acetate 23:2) gave 370 mg (46%) of 1d as a pale yellow oil, which crystallized upon standing overnight in the refrigerator: mp 50-52 C

Allyl-2-bromo-5-nitrobenzene (1e).
Column chromatography (heptane-ethyl acetate 49:1) gave 510 mg of a pale yellow oil. The oil was dissolved in pentane, and the solution was cooled to -20 C upon which 1e precipitated as white crystals. Recrystallization (pentane, room temperature to -20 C) gave 420 mg (58%) of the title compound as white crystals: mp 43-44 C

Allyl-4-methoxy-3-nitrobenzene (1f).
The crude product was dissolved in 20 mL of ethyl acetate-heptane (1:2), and the solution was filtered through a pad of silica. The solvent was removed at reduced pressure. Column chromatography (heptane-ethyl acetate 97:3) of the residue gave 230 mg (40%) of 1f as a yellow oil.

Allyl-4-nitro-3-trifluoromethylbenzene (1g).
Column chromatography (heptane-ethyl acetate 97:3) gave 430 mg (62%) of 1g as a pale yellow oil.

Allyl-3,5-dinitrobenzene (1h).
Column chromatography (heptane-ethyl acetate 23:2) gave 505 mg (81%) of 1h as a yellow oil.

Allyl-3,5-dichlorobenzene (1i).
Column chromatography (pentane) followed by preparative HPLC (gradient solution: 50:50 to 80:20 for 30 min and then 95:5 (CH3CN/H2O) gave 267 mg (48%) of 1i as a colorless oil.

Ethyl 3-Allyl-5-nitrobenzoate (1j).
Column chromatography (heptane-ethyl acetate 47:3) gave 470 mg (67%) of 1j as a pale yellow oil, which crystallized upon standing overnight in the refrigerator; melting below +20 C.

4-Allylnitrobenzene (1k).
Column chromatography (heptane-ethyl acetate 23:2) followed by preparative HPLC (gradient solution: 50:50 for 10 min then 60:40 for 20 min (CH3CN/H2O) gave 267 mg (55%) of 1k as a pale yellow oil.

4-Allylbromobenzene (1l).
Column chromatography (pentane) gave 200 mg (34%) of 1l as a colorless oil.

4-Allyl-tert-butylbenzene (1m).
Column chromatography (heptane) gave 100 mg (19%) of 1m as a colorless oil.

Allyl-2,4-dinitrobenzene (1p).
In addition to the general procedure for the allylation, the reaction time was extended 18 h (60 C reaction temperature). During the first 4 h of this additional reaction time, extra tert-butyl nitrite (2.86 mL, 24 mmol) was added in portions to the reaction mixture. Column chromatography (heptane-ethyl acetate 23:2) followed by preparative HPLC (gradient solution: 60:40 to 70:30 for 20 min and then 95:5 (CH3CN/H2O) of the remaining residue gave 305 mg (49%) of 1p as a yellow oil.

Large-Scale Synthesis of Allyl-3,5-dinitrobenzene (1h).
 Neat 4h (75 g, 0.41 mol) was added in portions to a solution of tert-butyl nitrite (84.6 mL, 0.71 mol) and allyl bromide (530 mL, 6.15 mol) in CH3CN (25 mL), keeping the temperature between 11 and 15 C. Before the addition of the final 25% of 4h, more tert-butyl nitrite (21 mL, 0.18 mol) was added. The reaction mixture was then stirred at room temperature (23-25 C) for 1 h. Excess tert-butyl nitrite, allyl bromide, and CH3CN were distilled off from the reaction mixture at reduced pressure, and toluene (500 mL) was added to the orange-brown residue. The resulting mixture was filtered twice through alumina pads (10 × 10 cm), and the pads were washed with a total of 1.5 L of toluene. The toluene was distilled off at reduced pressure, and isooctane (200 mL) was added to the remaining pale yellow residue. The mixture was stirred at 60 C for 0.5 h (to extract the partly polymerized allyl bromide and 1,2,3-tribromopropane from the product) and then cooled to approximately -50 C. As soon as a white precipitate (partly polymerized allyl bromide and 1,2,3-tribromopropane) started to form, the solvent was decanted (including the precipitate) and to the remaining yellow oil another portion of isooctane (200 mL, 20 C) was added. The oily residue in isooctane was then stirred at -50 C with a spatula until crystals formed. In some cases, it was necessary to decant the isooctane phase and add fresh isooctane before the oil started to crystallize. The crystals (remaining in the flask) were then washed twice with isooctane (precooled to -50 C) at -50 C. Residual isooctane was removed at reduced pressure to give 78.9 g of 1h as a yellow oil containing less than 4% of 5h according to 1H NMR analysis. This corresponds to 89% yield of allyl-3,5-dinitrobenzene. In an alternative procedure, the crystals (after washing with isooctane at -50 C) were collected by filtration at -50 C and then washed once with isooctane (precooled to -50 C). The latter method gave the same yield of slightly purer 1h. Spectral data were identical to those of the reference sample from the small scale synthesis.

Those who give up essential liberties for temporary safety deserve neither liberty nor safety