2.1. TDO as a reducing agent
Firstly, the reducing properties of TDO (Compound
1) on nitrobenzenes were explored to outline and screen all the mechanochemical parameters for this step. Considering the few existing techniques for reducing nitrobenzene with TDO [
76,
77], we had to lay the groundwork for a methodology with broader applicability. We set the mechanochemical procedure at a 1.0 mmol scale using nitrobenzene as a reference substrate. We milled TDO
1 (1.0 mmol), nitrobenzene
2a (1.0 mmol), and NaOH (1.0 mmol) inside a 10 mL stainless steel (SS) vessel equipped with two balls (Ø = 7 mm, 2.67 g) of the same material, running the reaction for 1.0 hour (
Scheme 3).
Unfortunately, we detected the only presence of the starting material through a GC-MS analysis (
Table 1, entry 1). Several process parameters have been investigated to overcome these failures, and the whole optimization process is summarized in
Table 1. To begin with, we raised the ratio of
1 and NaOH (
Table 1, entries 2 and 3), which allowed us to convert
2a into aniline
3a with a 21% yield (
Table 1, entry 3). Then, prompted by these results, we tried to increase both the reaction time and the reducing mixture equivalents (TDO and NaOH). In the first case, the yield was even lower in a 2-hour milling with a 5% yield (GC-analysis), probably due to further reactivity of the formed aniline with the redox intermediates (
Table 1, entry 4). In the latter one,
2a was consumed entirely, but
3a was obtained with only a 29% yield together with other process intermediates (
Table 1, entry 5). Lastly, we ran the reducing process at 70 °C to accelerate the kinetics of the reaction, but we only obtained the corresponding symmetric diazobenzene (PhN=NPh,
Table 1, entry 6).
So, after all these failed attempts, we considered using drops of different solvents to run a
Liquid
Assisted
Grinding (
LAG) [
78,
79,
80,
81,
82] in a 90-minute procedure. Consistently, less polar, like decane or toluene, and polar solvents, like acetone or isopropanol, did not permit a reasonable conversion rate (
Table 1, entries 7-10). Lastly, methanol and water were used, as in analogous solvent-based procedures. However, in this case, the ratio of solvent/reagents was drastically cut down compared to the already reported methodologies (LAG, η = 0.44 µL/mg). Contrasting what is known, methanol used under LAG conditions produced a complex mixture of aniline and nitrobenzene reduction process intermediates (
Table 1, entry 11). Water, instead, led to excellent yields of
3a (
Table 1, entry 12). Its amount, however, was found to be a critical parameter since the conversion rate dramatically dropped when η = 0.22 µL/mg (
Table 1, entry 13). Contrarily, a little increase in the reaction time of up to 2 hours allowed a quantitative conversion of
2a to
3a (
Table 1, entry 14). Concerning the bases, weaker ones like sodium carbonate and sodium bicarbonate did not allow a comparable result (
Table 1, entries 15 and 16), proving that NaOH plays a crucial role in the mechanochemical process - as already insight by Hawkes for similar reactions in solution [
83]. Lowering the NaOH amount negatively affected the reaction performance as well (
Table 1, entry 17).
With the optimized conditions in hand, we extended the entire procedure to other nitrobenzenes to validate this mechanochemical process. In the case of activated substrates
2b and
2c, the process smoothly proceeded to a complete conversion within 2 hours, while for more complex substrates like
2s, the process was completed only in 3 hours (
Scheme 4).
Reducing instead 2-nitroaniline
2d, many unpredicted outcomes showed up, as summarized in
Table 2. In this case, we synthesized the
o-phenylenediamine
3d with a 37% yield in 90 minutes without a LAG (
Table 2, entry 1). Such different behavior can be ascribed to the positive effects of an EDG. However, unexpectedly, prolonging the reaction time to 2 hours under neat grinding conditions resulted in the formation of the corresponding benzimidazole, albeit in low yields (
Table 2, entry 2). For the sake of completeness, we also tried to reduce in neat conditions other substrates having a comparable charge distribution (
Table 2, entries 3 and 4). With 2- nitro anisole, we obtained the corresponding aniline in a 54% isolated yield. At the same time, the employment of 2-nitro phenol resulted in a mixture of various unidentified products, likely generated by the high phenoxide reactivity.
To better understand several critical details of the process, we have to better focus on several points of the process. First, under LAG conditions, the presence of methanol resulted in the concurrent formation of
3d and various reaction intermediates. At the same time, water use was associated with obtaining the desired product with a nearly quantitative yield in 2 hours (
Table 2, entries 5 and 6). These different outcomes can be reconducted to the role covered by water as a better proton source. On the other hand, water, to some extent, also inhibits the final cyclization pathway that leads to benzimidazole formation. The reduction of the nitro group and the construction of the benzimidazole ring both consume TDO, with the latter being kinetically faster. As a result, any attempt to decrease the NaOH equivalents failed because of the high reactivity of the formed
o-phenylenediamine towards the TDO that was still present in the reaction medium (
Table 2, entry 7). Once these issues are focused on, we can draw conclusions based on the above and after a long, meticulous exploratory study. Six equivalents of NaOH promote the sluggish kinetics of the reduction reaction to the detriment of the cyclization reaction, resulting in a complete reduction of the nitro group (
Table 2, entry 6).

Table 2.
Optimization process for the reduction of 2d to 3d.
Table 2.
Optimization process for the reduction of 2d to 3d.
Entry |
TDO eq. |
Base eq. |
Reaction Time (h) |
Additivesb
|
Yieldsa
|
1 |
3 |
6 |
1.5 |
/ |
37% |
2c
|
3 |
6 |
2 |
/ |
45% |
3d
|
3 |
6 |
2 |
/ |
54% |
4e
|
3 |
6 |
2 |
/ |
Complex mixture |
5f
|
3 |
6 |
2 |
MeOH |
0% |
6 |
3 |
6 |
2 |
H2O |
98% |
7g
|
3 |
3 |
2 |
H2O |
1% |
Once we understood the reactivity of 2-nitroanilines, we also extended this process to other 2-nitroaniline derivatives (
Scheme 5). As a result,
o-phenylenediamines
3e and
3f were successfully synthesized in a 2-hour ongoing process, whereas the substrates
3g-j and
3o-p needed a longer reaction time of 3 hours. These outcomes were utterly in line with Hammett’s parameters and steric hindrance on the aromatic ring of the starting materials.