Benzylic deuteration of alkylnitroaromatics via amine‐base catalysed exchange with deuterium oxide

This paper describes the deuterium‐labelling of alkylnitroaromatics by base‐catalysed exchange with deuterium oxide. As the alkyl protons alpha to the aromatic ring are the most acidic sites in the molecule, regioselective hydrogen isotope exchange at this benzylic location leads to a regiospecifically deuterated product. The exchange labelling takes place in good yields and with high atom% abundance in the presence of an appropriate nitrogen base. Alkylated 2,4‐dinitrobenzenes deuterate at room temperature under catalysis by triethylamine, whilst alkylated 2‐nitro‐ or 4‐nitrobenzenes and related mono‐nitroaromatics require higher temperatures and catalysis by 1,5‐diazobicyclo[4.3.0]non‐5‐ene (DBN). The labelling reactions require an inert gas atmosphere, but otherwise are simple and high yielding with no obvious byproducts. Those compounds in which the benzylic protons are in an ortho‐orientation with respect to the nitro group label somewhat more slowly than the analogues where there is a para relationship. In addition, higher alkyl homologues undergo benzylic deuteration at slower rates than methyl.

This paper describes the deuterium-labelling of alkylnitroaromatics by basecatalysed exchange with deuterium oxide. As the alkyl protons alpha to the aromatic ring are the most acidic sites in the molecule, regioselective hydrogen isotope exchange at this benzylic location leads to a regiospecifically deuterated product. The exchange labelling takes place in good yields and with high atom% abundance in the presence of an appropriate nitrogen base. Alkylated 2,4-dinitrobenzenes deuterate at room temperature under catalysis by triethylamine, whilst alkylated 2-nitro-or 4-nitrobenzenes and related mononitroaromatics require higher temperatures and catalysis by 1,5-diazobicyclo [4.3.0]non-5-ene (DBN). The labelling reactions require an inert gas atmosphere, but otherwise are simple and high yielding with no obvious byproducts. Those compounds in which the benzylic protons are in an orthoorientation with respect to the nitro group label somewhat more slowly than the analogues where there is a para relationship. In addition, higher alkyl homologues undergo benzylic deuteration at slower rates than methyl.

K E Y W O R D S
alkylnitroaromatics, amine-base-catalysis, deuteration, DNT, HIE, isotope exchange, nitrotoluenes 1 | INTRODUCTION Nitrated alkylaromatics are an important class of compounds and are utilised as intermediates in many highvolume industrial processes. Typical markets for nitrotoluene feedstocks include pigments, dyestuffs, imaging chemicals, agrochemicals, pharmaceuticals and plastics. 1 In addition, alkylated nitroaromatics form the basis of many conventional explosives and energetic plasticisers, which are toxic and environmentally hazardous 2 and which therefore require analysis [3][4][5] and remediation methodologies. 6 Methods for labelling alkylnitroaromatics are therefore of interest. This paper describes the regiospecific benzylic deuteration of alkylnitroaromatics to provide labelled compounds for use in studies of the reactivity of alkyl-2, 4-dinitrobenzenes under oxidising and nitrating conditions. 7 To provide a general labelling method, we decided to approach the deuteration of this class of compounds via hydrogen isotope exchange methodology. 8,9 Isotopic exchange labelling of benzylic positions in conjunction with metal catalysts, sometimes with good specificity in particular cases, has been known for well over 50 years and continues to be of use. [10][11][12][13] Of current interest are nanoparticle catalysts that have been utilised very effectively and with good regiospecificity. 14,15 Alkalimetal amide superbase catalysis has also been used recently. 16 Unfortunately, the above approaches generally use hydrogen, deuterium or tritium gas as the initiator or isotope donor and are therefore inappropriate for the labelling of nitro-compounds, which are easily reduced species. For the labelling of alkylnitroaromatics in the benzylic positions, 17-22 therefore, we decided to avoid reductive conditions and to utilise deuterium oxide as the donor. In addition, we exploited the differences in acidity between their aromatic and benzylic protons to provide the required labelling specificity.
In alkylnitroaromatics, the benzylic protons are rendered particularly acidic by the presence of electron withdrawing nitro-substituents, 23 which can stabilise benzylic carbanions by resonance and inductive effects. However, studies of the interaction of 2,4,6-trinitrotoluene (pKa 12 in DMSO/methanol) with ethoxide or hydroxide have shown the competing formation of an addition product. 24 In this reaction, attack of the nucleophile at the position meta to the methyl group, or on the aromatic carbon bearing the methyl group, leads to anionic σ-Meisenheimer species and renders the target methyl protons in these species unavailable for deuterium exchange. The analogous situation is summarised in Figure 1 for 2,4-dinitrotoluene, a dinitro-compound, which has been studied in most detail with respect to this behaviour. 25 In this case it was shown that significant concentrations of potential Meisenheimer adducts (e.g., C) are avoided provided that the basic catalyst is an amine and the nitroaromatic is only weakly acidic. Under these conditions a small equilibrium concentration of the methylene carbanion (B) seems to be present. This renders the desired isotopic exchange likely, provided that a suitably acidic deuterium donor (e.g., D 2 O) is present. The striking blue colour of the reaction in the case of the dinitro-compound was ascribed to the carbanion, though highly coloured complexes of 2,4,6-trinitrotoluene with amines have also been proposed in which the interaction ranges from simple complex formation between the amine and the nitro-substituents to the fully σ-bonded Meisenheimer complexes. 26,27 In the light of the above considerations, bearing in mind the inability of sodium hydroxide and deuterium oxide to facilitate deuteration in a previous investigation, 25 we studied the efficacy of amine basecatalysed isotope exchange between mononitro-and dinitro-alkylaromatics and deuterium oxide using the substrates in Figure 2.

| RESULTS AND DISCUSSION
We began our studies with 2,4-dinitrotoluene (1). In this case, the benzylic protons are quite acidic (pKa 15.3) due to extensive resonance stabilisation of the benzylic carbanion by the two nitro-substituents. Deuterium labelling of this compound had been noted previously 25 when sodium hydroxide solutions were treated with deuterosulphuric acid. Interestingly, though, no deuterium exchange was noted without the addition of deuterosulphuric acid, implying that the benzylic carbanion itself was inert to exchange under treatment with a strong Bronsted base. However, we found that a simple tertiary amine base was efficient in catalysing the exchange of 2,4-dinitrotoluene, provided that oxygen was excluded. If this was not the case, both the degree of deuteration and the yield were reduced.
When carried out under nitrogen or argon, the labelling of 2,4-dinitrotoluene proceeded smoothly at room temperature in tetrahydrofuran, and also in N,Ndimethylformamide. Thus, [methyl-D 3 ]2,4-dinitrotoluene was prepared easily, regiospecifically, and with high isotopic enrichment, under catalysis by triethylamine (pKa 10.75 in water) without any of the possible complications due to the formation of the Meisenheimer complex or of inhibition of the exchange by stoichiometric deprotonation to the carbanion. Figure 3 shows a typical timecourse for the labelling reaction.
Moreover, throughout, the labelling reaction followed the behaviour expected for a simple statistical H/D exchange 28 (Figure 4 shows the data at 71.5% D; also see Data S1, section 1 for other examples of statistical labelling of alkylnitroaromatics).
However, the choice of the base catalyst did prove important, as the degree of deuteration achieved showed F I G U R E 1 Potential species present when 2.4-dinitrotoluene is treated with base no simple correlation with base pKa (Data S1, section 2). In practice, both triethylamine and N-methylpyrrolidine (pKas ca. 10.8) proved very effective with dinitro-systems, provided, of course, the benzylic protons of the alkyl substituent were in an ortho or para relationship with the nitro-groups. Figure 5 shows the relative rates of deuteration of three 2,4-dinitroalkylbenzenes under competitive conditions (see Data S1, section 3 for a typical ion chromatogram). Clearly the rates for the higher alkyl substituents are slower. This could be due to the increased inductive (electron donating) effect of the extended alkyl group on the acidity of the benzylic protons (destabilisation of the carbanion): conversely it could be due to a steric effect, or to both these causes.

| Labelling of mono-nitroalkylaromatics
The benzylic protons of mono-nitro compounds are much less acidic than those of dinitro-systems (e.g., the pKa of 4-nitrotoluene is ca. 22 in DMSO containing 5% water 23 ). Nevertheless, mono-nitro systems with the alkyl substituent ortho or para to the nitro group can still be labelled smoothly using this approach, provided that a stronger base (e.g., 1,5-diazabicyclo[4.3.0]non-5-ene, DBN), a longer reaction time (overnight) and a higher temperature (95 C) are applied. Figure 6 shows a typical time course for 4-nitrotoluene deuteration.

| Factors affecting labelling ability
To determine the molecular factors affecting labelling propensity, the extent of deuteration of a range on alkylmononitrobenzenes was compared under identical conditions for a fixed reaction time ( Figure 7). Competitive reactions were also carried out that showed the same rank ordering in the %D achieved. The resulting data suggest the following issues should be considered when applying the methodology.
Clearly, the orientation of the benzylic protons relative to the nitro groups is important. Labelling requires the nitro group to be ortho or para to the benzyl group, as expected if resonance stabilisation of the benzylic anion is a significant parameter facilitating the exchange.
Additionally, other sterioelectronic effects are significant in the exchange. In the alkyldinitroaromatics discussed earlier ( Figure 5), the benzylic exchange at an ethyl substituent is less facile than that of a methyl group, whilst a propyl group reacts even more slowly. For the mono-nitro substrates a similar pattern is seen for 2-methyl, 2-ethyl and 2-propyl-nitrobenzene ( Figure 7). Moreover, the sluggish exchange of 2,6-dimethylnitrobenzene compared with other dimethylnitrobenzenes also suggests the steric effect is significant in this case.
Also of significance is the ortho versus para orientation of the alkylnitroaromatic. Figure 8 shows a comparison of such pairs. In all cases, the substrates in which the benzylic position is ortho to the nitro group labels less well than when the orientation is para. It is known that charge delocalisation by an ortho-nitro substituent is less effective than for the corresponding para substituent 26 making the ortho analogue less acidic by ca. 1.5 pK a units. 23 Bearing all the above behaviours in mind it seemed likely that steric inhibition of resonance was involved. Hence, the conformations of the substrates were examined using simple AM1 minimisation of the unsolvated ground states using Arguslab.4.0.1 or Spartan-14 packages. Figure 9 shows the relationship between the propensity for deuteration and the dihedral angle between the plane of the nitro group and the plane of the aromatic ring (see Data S1, section 4 for minimised conformations). Clearly the presence of bulky alkyl groups in an ortho-orientation has the effect of twisting the nitrogroup out of plane and hence reducing the potential for resonance stabilisation of the benzylic anion. Studies using more sophisticated modelling of carbanion and transition state energies are indicated to confirm this interpretation and to identify the various factors leading to the reactive conformations or reactive species.

| CONCLUSIONS
The benzylic protons of alkylnitroaromatics may be exchanged for deuterium from deuterium oxide under catalysis by amine bases. In the case of 2,4-dinitro-alkylaromatics, the exchange is facile in the presence of triethylamine, whereas mono-nitro systems require a higher temperature and a stronger base, such as 1,5-diazobicyclo[4.3.0]non-5-ene (DBN). Systems in which the benzylic protons are in an ortho-orientation with respect to the nitro group label somewhat more slowly than those that are in a para relationship. The labelling reactions require an inert gas atmosphere, but in CDCl 3 at 500 MHz or 300 MHz for 1 H and at 126 MHz or 100 MHz for 13 C. 2 H-NMR spectra were recorded in C 1 HCl 3 at 77 or 61 MHz. All spectra were analysed by Bruker TopSpin software. Raman spectra were measured on crystalline samples supported on glass microscope slides using a ThermoScientific DXR Raman Microscope at laser frequencies of either 532 or 780 nm. Infrared spectra were recorded for crystalline or neat liquid samples on a sapphire anvil using an Agilent Resolutions Pro spectrometer. GC-MS analysis was performed using 1 μl injections of solutions at ca. 1 mg/ml in dichloromethane using an Agilent 6890N gas chromatograph coupled to a 5973 mass-selective detector. The column was a Phenomenex ZB-5MS, 30 m Â 0.25 mm, used with a temperature profile: 50 C (held for 3 min), ramp at 10 C per minute to 250 C and held for 2 min. The carrier gas was helium at 1 ml/min.

| Reagents
Substrates were purchased from Merck (Sigma-Aldrich Company Ltd), The Old Brickyard, New Road, Gillingham, Dorset, SP8 4XT United Kingdom, or were prepared by nitration of the appropriate aromatic using nitric and sulphuric acids as specified below. Deuterium oxide (99.9 atom%D) and N,N-dimethylformamide (DMF, <0.005% water content) were also obtained from Merck (address above). General solvents and reagents were obtained from regular chemical supply houses and were used as received.

| Determination of the percentage deuteration (isotopic abundance) of labelled compounds by mass spectrometry
The isotopic abundance of labelled nitroaromatics was determined by analysis of the integrated values of the whole GC-MS peak of the labelled compound to avoid any inaccuracies arising from isotopic fractionation. The raw values from the molecular ion clusters of the labelled compound reported by the Agilent MassLynx software were then corrected for the presence of natural abundance isotopes using IsoPat 29 and NAIC 30 software prior to calculation of the isotopic abundance.

| Preparation of deuteration substrates by nitration
Substrates not commercially available were prepared as below.
Nitration of ethyl-4-nitrobenzene to yield ethyl-2,4-dinitrobenzene Ethyl 4-nitrobenzene (5.0 g, 33.1 mmol) was pre-washed with sodium sulphite solution (20 ml, 10% w/v) and water (2Â 120 ml) before nitration by slow addition to a stirred nitrating mixture prepared from 70% nitric acid (12.8 g, 0.14 mol, d 1.42 g/ml) and concentrated sulphuric acid (19.2 g, 0.20 mol, d 1.84 g/ml) in an ice bath ensuring that the temperature did not rise above 20 C. The reaction was then left to stir at room temperature overnight. The resulting mixture was then transferred into a separating funnel and left standing for an hour to allow full separation of the phases. The organic layer was separated, washed with water (2Â, 10 ml) and diluted with dichloromethane (15 ml). The dichloromethane solution was washed with sodium sulphite solution (10 ml, 5% w/v) and then stirred with potassium carbonate solution (10 ml, 5% w/v) for 30 min at ambient temperature. The organic layer was separated and dried overnight over anhydrous sodium sulphate. The solution was filtered and the solvent removed by evaporation under reduced pressure to leave ethyl 2,4-dinitrobenzene as a yellow oil, 4. Nitration of propylbenzene to yield propyl-2-nitrobenzene, propyl-4-nitrobenzene and propyl-2,4-dinitrobenzene Propylbenzene (1.8 ml, 13 mmol) and concentrated sulphuric acid (3 ml) were stirred and cooled in ice and a mixture of concentrated nitric acid (0.5 ml) and concentrated sulphuric acid (1 ml) were added dropwise over 12 min. The reaction was allowed to attain room temperature with stirring for a further 21 min before being quenched with ice (25 g). Isolation of the products was carried out by partition of the reaction mixture between water (5 ml) and dichloromethane (5 ml). The dichloromethane extract was then washed with water (four portions of 5 ml) and dried. Removal of the dichloromethane yielded a mixture containing 2-propynitrobenzene, 4-propylnitrobenzne and 1-propyl-2,4-dinitrobenzene. A portion of this mixture was separated into its components by column chromatography using a silicagel column eluted with hexane containing increasing quantities of dichloromethane (from 10% to 40% v/v). The products were analysed by GC-MS and 1 H-NMR. Nitration of tetralin to yield 5-nitrotetralin and 6-nitrotetralin Tetralin (1,2,3,4-tetrahydronaphthalene, 2 mmol) was dissolved in dichloromethane (10 ml) and concentrated sulphuric acid (324 μl) was added. Then two portions of nitric acid (81 μl each time) were added over 5 min. After stirring for a further 2 h, the reaction mixture was neutralised by the addition of saturated sodium bicarbonate solution. After discarding the aqueous layer the dichloromethane layer was washed twice with water (10 ml), dried and evaporated. NMR analysis showed the product to be a mixture of 5-and 6-nitrotetralin. The products were partially separated using a column chromatography (silicagel, increasing quantities of dichloromethane in hexane from 10% to 50% v/v). The two products were then further purified by preparative thin layer chromatography (silicagel, 50% dichloromethane in hexane) and stripped from the stationary phase using 50% ethyl acetate in dichloromethane before characterisation by GC-MS and 1 H-NMR.   1 mmol) was dissolved in dry tetrahydrofuran (0.5 ml) and D 2 O (100 μl) added. The reaction tube was sealed with a septum and thoroughly flushed with nitrogen. Triethylamine (100 μl) was then injected via a syringe, resulting in the deep blue colour of the benzylic anion. This colour changed slowly to green over the subsequent 18 h. After this period 37% DCl in D 2 O (100 μl) was injected to terminate the reaction. Extraction of the labelled product was carried out by partition between dichloromethane (10 ml) and water (10 ml

| 1-[Ethyl-1 0 -D]ethyl-2,4-dinitrobenzene
1-Ethyl-2,4-dinitrobenzene (1.25 g, 6.4 mmol) was dissolved in DMF (42 ml) and deuterium oxide (8.3 ml) in a round-bottomed flask sealed under nitrogen via a rubber septum. Triethylamine (8 ml, 5.81 g, 57.4 mmol) was then added via a syringe and the solution was stirred at ambient temperature for 24 h. The reaction was then neutralised with 37% DCl in D 2 O and partitioned between water (60 ml) and dichloromethane (60 ml). The organic layer was washed with water (3Â, 60 ml) and dried overnight over anhydrous sodium sulphate. After cooling, the labelling reaction was terminated by the addition of 37% DCl in D 2 O (1 ml). The reaction mixture was then partitioned between dichloromethane (10 ml) and 4N hydrochloric acid (6 ml) and the dichloromethane layer separated and washed four times with water (8 ml each time) before being dried over anhydrous magnesium sulphate, filtered, and the filtrate evaporated at 70 degrees to constant weight to yield .0]non-5-ene, 0.4 ml) was then added and the tube heated at 95 degrees for 18 h. The labelling was terminated by the addition of 37%. DCl in D2O (0.5 ml). The reaction mixture was then partitioned between dichloromethane (10 ml) and 4N hydrochloric acid (7 ml)  (Further spectroscopic data for all the above labelled compounds are provided in Data S1, sections 5 and 6). 4.4 | Comparative deuteration of 2,4-dinitrotoluene, ethyl-2,4-dinitrobenzene and propyl-2,4-dinitrobenzene The three substrates (0.1 mmol each) in tetrahydrofuran (1.0 ml) and deuterium oxide (200 μl) were placed in a septum-capped tube and sealed under nitrogen. Triethylamine (200 μl) was then added via a polypropylene syringe and the tube shaken. Samples (ca. 100 μl) were removed via a gas-tight syringe at 1, 2, 5, 10, 20, 60, 120, 180 and 1100 min and added to 37% deuterochloric acid in deuterium oxide (100 μl). The resulting mixture was partitioned between dichloromethane (5 ml) and water (10 ml), the dichloromethane layer separated, washed with water (2 x 5 ml), and dried. An aliquot of the dichloromethane extract was then analysed by GC-MS to determine the percentage deuteration of each of the substrates at each time point.

| Protocol for comparison of the extent of deuteration of alkylmononitroaromatics
The substrates (0.1 mmol) in N,N-dimethylformamide (500 μl) and deuterium oxide (100 μl) were placed in thick-walled septum-capped pressure tubes and sealed under nitrogen. Next, 1,5-diazabicyclo[4.3.0]non-5-ene, DBN, (200 μl), was injected into each tube via a syringe. The reaction tubes were heated at 94 C for 4 h using a heating block. The reactions were run under conditions where equilibrium had not been achieved so that comparisons of the degree of deuteration could be made. After cooling, the tubes were decapped and 37% deuterochloric acid in deuterium oxide (100 μl) was added. The resulting solution was partitioned between water (5 ml) and deuterochloroform (1.5 ml). The deuterochloroform phase was then washed with water (4Â, 5 ml), separated and dried, before confirmation of identity by TLC and GC retention time and determination of the percentage deuteration by GC-MS.

DATA AVAILABILITY STATEMENT
The data that support the findings of this study are available from the corresponding author upon reasonable request.