Nitrogen heterocycles are privileged structures in pharmaceutical chemistry and an enormous effort is continuously being made to improve their synthesis. Among the numerous possible synthetic approaches, the reductive cyclization of nitroarenes and nitroalkenes by carbon monoxide, catalyzed by transition metal complexes, appears to have some highly desirable features: (1) nitroarenes are usually the entry point for the introduction of nitrogen-containing groups on the aromatic ring and nitroalkenes can also be often prepared easily, e.g., by an Henry reaction; (2) carbon monoxide is cheap with respect to virtually any other reducing agent except for dihydrogen, which however affords anilines and not heterocyclic compounds in most cases; (3) the only stoichiometric byproduct is CO₂, which spontaneously separates from the reaction products at the end of the reaction, thus simplifying the work-up. This is a clear advantage with respect to reactions employing phosphites or phosphines as reductants (e.g., the classical Cadogan reactions), whose oxidized product usually needs a chromatographic purification to be completely eliminated; (4) selectivities in the desired heterocycles are often very high and almost quantitative in several cases; (5) low catalyst loading is possible, up to 0.1 mol % or even lesser in some cases. Given these features, it may appear surprising that such a synthetic approach has not become widespread in synthetic organic chemistry laboratories or even in industrial practice. The main reasons for this are clearly technical: performing these reactions requires the use of high-pressure equipment and pressurized CO lines. The latter, in particular, are not present in the overwhelming majority of chemical laboratories. The problem is also common to other carbonylation reactions and, in the last decade, different solid or liquid substances able to liberate CO under the reaction conditions have been developed. The field has been reviewed several times. However, several of these so-called CO surrogates are quite expensive, highly toxic, or require the use of a two-chamber reactor to be employed. Several years ago, we started to investigate the use of CO surrogates in the field of reductive cyclization reactions of organic nitro compounds and selected formate esters as reagents because they are cheap, non- or little-toxic and because the stoichiometric byproduct, an alcohol or phenol, is unlikely to interfere with the reaction course. In this account, our results in the field are summarized. For the sake of completeness, it should be mentioned that other groups have also employed Mo(CO)₆, Co₂(CO)₈, and a triformate ester as CO surrogates for related reductive cyclization reactions of nitroarenes to give N-heterocycles.

• The initial activation of the nitro compound, at least when late transition metal catalysts are employed, is always an electron transfer from the metal to the nitro group. For this reason, low-valent metal complexes need to be used. However, due to the high sensitivity of the latter, metal complexes in higher oxidation states are often used as precatalysts, which are reduced by CO under the operating conditions. By the same token, nitroarene and to a lesser extent nitroalkenes are suitable substrates, but nitroalkanes have higher oxidation potentials and are unreactive in these systems.
• Palladium, ruthenium and rhodium compounds have all been employed as catalysts, but the best results have been obtained by the use of palladium and in the last decade the other two metals have only rarely been used.
• Phosphines have been used as ligands for palladium in many cases, but it has been shown that they are oxidized to phosphine oxides during the reaction. Since we aim at developing a catalytic system that may also be applied at an industrial level, we prefer to avoid using them. No successful use of N-heterocyclic carbenes as ligands in this field has ever been reported. The best ligands in terms of activity and stability of the catalytic system are phenanthroline and its substituted derivatives.
• Aryl formates can be decomposed to CO and phenols even by weak organic bases. Alkyl formates are cheaper, but they are activated only by very strong bases, which would not be compatible with most reactions. Alternatively, they can be decomposed by the action of a ruthenium-based catalytic system.
• When using CO surrogates, the features of the vessel in which the reaction is performed are important for the success of the reaction and for safety reasons. We have discussed the pros and cons of different kinds of “pressure tubes” in a previous paper, thus we will not do it here again.