Previously, we successfully developed ammonium salt-accelerated hydrazinolysis of unactivated amides using ammonium iodide and hydrazine monohydrate under mild conditions (Shimizu et al., 2014). To evaluate this reaction using the FGE kit, we selected N-(4-(trifluoromethoxy)phenyl)-4-(trifluoromethyl)benzamide (1) as the substrate because yields of the corresponding products 4-(trifluoromethyl)benzohydrazide (2) and 4-(trifluoromethoxy)aniline (3) can be determined by ¹⁹F NMR analysis of the crude mixture without affecting ¹H-based contaminants such as additives in the FGE kit.

We first optimized reaction conditions for the hydrazinolysis of amide 1aa using General Procedure A (Table 1) (See Supplementary Material for detailed information). Under the standard conditions of our previous studies using 10 equiv of hydrazine hydrate and 1.0 equiv of ammonium iodide at 70 °C for 48 h, the reaction gave only 20% of benzohydrazide 2a (Entry 1). The reaction in the absence of solvent did not proceed (Entry 2). Increasing the reaction temperature to 100 °C increased the yield of 2a to 85% in 24 h (Entry 3). Aiming to accelerate the reaction, we performed the reaction using acidic solvent hexafluoroisopropanol (Zhang et al., 2018; Kabi et al., 2020; Bhattacharya et al., 2021) and trifluoroethanol (Entries 4 and 5) (Smithrud et al., 1990; Bautista et al., 1999; Bai et al., 2019); the use of trifluoroethanol led to desired product 2a in 93% yield. Using trifluoroethanol as the solvent in the absence of ammonium iodide decreased the reaction rate, however, clearly indicating that ammonium salt is important for achieving high reactivity (Entry 6) (Shimizu et al., 2012). On the other hand, the addition of 2 equiv of ammonium iodide decreased the yield of the product (Entry 7). NH₄OAc (Kumar et al., 2022) instead of NH₄I resulted in the lower yield of 2a (Entry 8). When hydrazine monohydrate was not added, the reaction did not proceed and amide 1aa was recovered quantitatively (Entry 9). Although increasing the amount of hydrazine from 10 to 20 equiv. improved the yield (Entry 10), the change was not significant (Entry 5). On the other hand, two equivalents of hydrazine gave 2a in 14% yield (Entry 11); therefore, 10 equiv was considered optimal. The concentration of the solvent also affected the reaction (Entries 12–14), and a 1 M solvent concentration was optimal (Entry 5). As shown in this table, the yields of benzohydrazide 2a and aniline 3a were nearly identical, so only the yields of 2a are listed in the following tables (See Table 1).

After determining the optimized reaction conditions (Table 1, Entry 5), we applied the FGE kit to the reaction. Additive compounds A0–A26 for our FGE kit contain the 4-chlorophenyl moiety as the parent backbone to facilitate ¹H NMR monitoring of the remaining additives (See Table 2). Because of its UV absorption, the 4-chlorophenyl moiety also enables HPLC analysis of the remaining additives. Furthermore, the isotopic distribution of the chlorine atoms facilitates the detection of chlorine-containing additive molecules as well as side products associated with the additives using mass spectroscopic analysis. We selected 26 functional groups found in many organic molecules, including amino acid residues such as the nitrogen-containing heterocycles imidazole A7 and indole A22.

Before starting the evaluation with the FGE kit, it was necessary to check whether the 4-chlorophenyl structure in the additive affects the reaction. Therefore, a control experiment was carried out by adding 1.0 equiv of additive A0 (1-butyl-4-chlorobenzene), which does not contain an additional functional group, to the amide bond cleavage reaction. Another objective of the control experiment with additive A0 was to check for reproducibility using the criterion of the standard deviation (σ) of the product yield (%) within 5 (σ ≤ 5) in five experiments (n = 5). Detailed experimental methods for using the FGE kit are reported in our previous work (Saito et al., 2023) (See Supplementary Material for detailed information).

Under optimized reaction conditions, the reaction was carried out five times with 1.0 equiv of additive A0 (Figure 1). The average yield decreased to 73% due to the decrease in the concentration caused by the addition of A0. Therefore, evaluation of the additive effect using A1–A26 is based on this yield. The standard deviation of the yield (%) was 0.97, suggesting that sufficient reproducibility could be obtained even under the conditions with additives. No decreases in the remaining additive (%) were observed (97%), and high reproducibility was obtained with a standard deviation of 2.52, indicating that the 4-chlorophenyl structure is tolerant to the reaction conditions (Supplementary Table S1).

After fulfilling the reproducibility criteria of additive A0, we examined the additive effects of the other additives, A1–A26, in the FGE kit. Each additive was subjected to an F-test through duplicate experiments (n = 2). Only when the F-test indicated that the variance differed and the variability was too high, two additional experiments with the same additives were performed. The following symbols are used in the experimental results (Figure 2; Table 2). The blue plus sign (+) indicates a statistically significant increase effect. The green plus/minus sign (±) indicates no statistically significant effect, but a result that was comparable to that of the control experiment. The yellow dash sign (−) indicates a statistically significantly decrease effect. When there was a significant decrease of more than half compared with the control experiment, a red “X” was used. Significant differences were determined by a t-test.

The effects of 26 additives A1–A26 in the FGE kit were examined using General Procedure B (Table 2) (See Supplementary Material). Additives A2, A7–A9, A12, A19, A22, A23, and A26 with alcohol, imidazole, aryl bromide, iodide, silyl ether, thiol, indole, primary amine, thioester moieties had no effect on product yield nor remaining additives. It is particularly important that highly reactive thiol A19 and thioether A26, as well as N-unprotected imidazole A7 and indole A22, were tolerated. Additives A5, A6, A10, A11, A15, A18, A21 and A24 with aldehyde, primary amine, terminal alkene, alkyne, enone, nitrile, guanidine and ester moieties did not affect the product yields, but the remaining additives were decreased, suggesting that the functional groups of these additives reacted under the reaction conditions (vide infra). In the case of additives A3 and A14 with aryl chloride and epoxide moieties, the yield of the product slightly increased and the remaining additive was unchanged. On the other hand, additives A1 and A17 with carboxylic acid moieties provided significantly higher product yields than the control reaction with additive A0 although the remaining additives were decreased due to the formation of byproducts (vide infra). For additives A4, A13 and A16 with alkyl ketone, naphthol and aryl ketone moieties, both the yield of the product and the remaining additive were decreased. The addition of A20 and A25 with Bpin and Boc-protected amine moieties decreased the product yields but did not change the remaining additives.

The decrease in the remaining additives was mainly due to the formation of the corresponding byproducts by a competitive reaction between the additives and hydrazine. The byproducts that formed from the additives were either isolated from the crude mixture or, if difficult to isolate, synthesized by methods reported in the literature, and their structures were confirmed using NMR and mass spectrometry analysis (See Supplementary Material).

As expected, carboxyl groups in additives A1 and A17, the amide group in additive A6, and the ester group in additive A24 reacted with hydrazine to afford the corresponding acyl hydrazides such as B1.

The carbonyl groups in additive A4, A5, and A16 also reacted with hydrazine to form the corresponding hydrazones such as B4, B5, and B16 (Scheme 2). Additive A15 with an enone moiety and additive A18 with a nitrile moiety also reacted with hydrazine to give nitrogen-containing heterocycles pyrazoline B15 and tetrazine B18 in good yields. On the other hand, some unexpected reactions were also observed. Additive A21 with a guanidine moiety reacted with hydrazine to give amine B21.

In the reactions using additives A10 and A11 containing terminal alkene and alkyne, a partial reduction proceeded to produce alkane B10 and alkene A10 (Scheme 3). This type of reduction of olefins by hydrazine was reported by Imada et al., where hydrazine was in situ-oxidized to diimide and the reduction was promoted by Brønsted acids (Arakawa et al., 2017). Surprisingly, the phenolic hydroxyl group and Ar-Cl moiety in additive A13 were converted to amino groups to give naphthalene-1,4-diamine (B13). The discovery of these unexpected reactions is an advantage of using the FGE kit.

Functional group compatibility of a reaction is evaluated from two aspects: the effect of the functionalized additive on the product yield and the remaining additive. The two-dimensional plot in Figure 3 shows the degree of the functional group compatibility for each reaction at a glance. In this plot, dots positioned near additive A0 (red dot in the plot) indicate that the functional groups in the corresponding additive are tolerant in this reaction. On the other hand, dots father away from additive A0 suggest that the reaction is significantly (often negatively) affected by the presence of the functional groups in the additives. This plot clearly shows that under the conditions of this amide bond cleavage reaction, although some functional groups reacted (reducing the recovery of the additive), many functional groups did not inhibit the reaction itself (reducing the yield). In addition, there are many additives with positive effects, and additives A1 and A17 with a carboxyl group significantly increased the product yield.

Carboxylic acids are easy to handle and often inexpensive (Gooßen et al., 2008; Pichette Drapeau and Gooßen, 2016). Therefore, even if the amide bond cleavage reaction requires an equivalent amount of additive, the carboxylic acid addition conditions can be a useful reaction system.

First, we screened various inorganic and organic Brønsted acids as additives using General Procedure B (See Supplementary Material). To evaluate the additive effects on the reaction rate, product yields were examined for 6 h during the course of the reaction. The use of inorganic acids C1–C3 significantly decelerated the reactions (Table 3). Next, we examined a wide variety of mono-carboxylic acids C4–C34 including A1 and A17 as additives to evaluate changes in the product yield. Overall, aliphatic or aromatic carboxylic acids were not very important, and the correlation between the electrical effects and reactivity was low. On the other hand, it is important to have a certain molecular size, and carboxylic acids, especially those with aromatic rings such as benzene rings, exhibited a good tendency. Ortho-substituted Benzoic acids C16–C18 and C26 were expected to have excellent effects because their carboxylic acid moieties did not react with hydrazine due to steric hindrance. They, however, did not significantly impact the product yields, suggesting that carboxylic acids without steric hindrance should be considered as additives to accelerate the reaction. The acceleration effect of heterocyclic rings is low (C29–C34). The presence of hydroxyl (C8 and C22) and amino (C7) groups in the vicinity of carboxylic acids was not desirable, and dicarboxylic and tricarboxylic acids (C34–C42) rather inhibited the reaction. Other types of acidic compounds, 4-toluenesulfonic acid (C41) and BINOL (C42), were also ineffective. Based on this screening study, we concluded that 4-butyl benzoic acid (C19) and 2-naphthoic acid (C27) were the most effective additives for the amide bond cleavage reaction.

As mentioned above, although several carboxylic acids had positive additive effects, the carboxyl groups reacted with hydrazine to form the corresponding hydrazide (Scheme 4). Therefore, control experiments between carboxylic acids and hydrazides were performed to determine whether the positive effect was due to the carboxylic acid or the in situ-formed hydrazide.

We selected benzoic acid (C14) and 4-chlorobenzoic acid (A17) as representative carboxylic acids, and the corresponding hydrazides were synthesized to study their effects. As shown in Table 4, the addition of hydrazides had lower effects (Entries 4 and 7) than carboxylic acids (Entries 2 and 5). Furthermore, the addition of equal amounts of carboxylic acid and hydrazide gave intermediate-level results (Entries 3 and 6). These results clearly suggest that carboxylic acids accelerated the amide bond cleavage reaction, not hydrazides.

Because carboxylic acids are gradually converted to less effective hydrazides under the reaction conditions, we next examined Lewis acids as additives. Some Lewis acids are reported to accelerate C–H/C–C/C–O/C–N bond cleavage reactions and improve reaction efficiencies (Kita et al., 2013; Baglia et al., 2016; Gao et al., 2021).

In fact, the addition of 1.0 equiv of Zn(OTf)₂ greatly accelerated the reaction, increasing the yield of 2a from 37% to 91% under the conditions shown in Supplementary Table S9. Furthermore, when the amount of Lewis acid was reduced to 0.1 equiv, the additive effect was maintained, although the yield was reduced to 66% (Table 5, Entry 9). In this case, no products other than the substrate 1aa and the target products 2a and 3a were observed.

We then screened various Lewis acid catalysts for the amide bond cleavage reaction using General Procedure B (Table 5) (See Supplementary Material). Although many Lewis acids showed no positive effects, a catalytic amount of Fe(OTf)₃ along with Zn(OTf)₂ efficiently accelerated the amide bond cleavage reactions and increased the yield of product 2a (Entry 15). Furthermore, in the case of Zn(OTf)₂, the reaction was accelerated more efficiently in the absence of ammonium iodide (Entry 10). On the other hand, in the case of Fe, no acceleration effect was observed under conditions without NH₄I (Entry 16).

Finally, with the optimized activation systems using carboxylic acids C19 and C27 (1.0 equiv) and Lewis acid catalysts Fe(OTf)₃ and Zn(OTf)₂ (0.1 equiv) in hand, the substrate generality of amide was examined using General Procedure B (Table 6) (See Supplementary Material).

We found that the additives also had a reaction-accelerating effect on other amide substrates. Again, to clarity the differences in the effects of these four systems, the results are shown for a reaction time of 6 h before the reaction is complete. Although the degree of effectiveness of each system varied depending on the substrates, in all cases the addition of 1.0 equiv of carboxylic acid was effective. In the case of 8-aminoquinoline amide 1be, a highly effective and frequently used directing group amide, the system using a zinc catalyst without the addition of NH₄I showed drastic acceleration effects, increasing the yield from 30% to 99% at 6 h. With acidic additives/catalysts, the reactions using amides 1bb–1bf were almost completed by prolonging the reaction time. The addition of carboxylic or Lewis acids to the amide bond cleavage reaction is expected to be useful for cleavage of less reactive amide bonds.