Deamination of 1-Aminoalkylphosphonic Acids: Reaction Intermediates and Selectivity

Deamination of 1-aminoalkylphosphonic acids in the reaction with HNO2 (generated “in situ” from NaNO2) yields a mixture of substitution products (1-hydroxyalkylphosphonic acids), elimination products (vinylphosphonic acid derivatives), rearrangement and substitution products (2-hydroxylkylphosphonic acids) as well as H3PO4. The variety of formed reaction products suggests that 1-phosphonoalkylium ions may be intermediates in such deamination reactions.

While the deamination of 2-aminoalkanoic acids as substrates has been described i a great number of articles, reactions of structurally similar 1-aminoalkylphosphonic acid are scarcely described in the literature. In 1950 [49] and 1954 [50], Kabachnik and Medved described the analytical applications of the deamination reaction of ami nomethylphosphonic acid and 1-amino-1-phenylethylphosphonic acid with nitrous ox ides in which hydroxymethylphosphonic acid and 1-hydroxy-1-phenylethylphosphoni acids were formed respectively (Scheme 1b). Much later, reactions of related aminoalky lidene-1,1-diphosphonic acids with nitrous acid were described by Blum and Worms [51 53]. The authors concluded that carbenium ions with two phosphonic groups are formed and the reaction products are hydroxyalkylidene-1,1-diphosphonic acids, chloroalkyli dene-1,1-diphosphonic acids, and derivatives of vinylphosphonic acid (Scheme 1c). It i worth mentioning that 1-phosphonoalkylium ions 9, which may be intermediates in th deamination reaction of 1-aminoalkylphosphonic acids 1, have also not been extensivel studied in the literature. Only theoretical calculations for the simplest phosphonomethyl ium ion (9h) (which exist in the cyclic form 10h) have been described by Pasto (Schem Scheme 1. General presentation of deamination reaction carried out on 2-aminoalkanoic acids and their phosphorus analogs [34][35][36][37][49][50][51][52][53]. While the deamination of 2-aminoalkanoic acids as substrates has been described in a great number of articles, reactions of structurally similar 1-aminoalkylphosphonic acids are scarcely described in the literature. In 1950 [49] and 1954 [50], Kabachnik and Medved described the analytical applications of the deamination reaction of aminomethylphosphonic acid and 1-amino-1-phenylethylphosphonic acid with nitrous oxides in which hydroxymethylphosphonic acid and 1-hydroxy-1-phenylethylphosphonic acids were formed respectively (Scheme 1b). Much later, reactions of related aminoalkylidene-1,1-diphosphonic acids with nitrous acid were described by Blum and Worms [51][52][53]. The authors concluded that carbenium ions with two phosphonic groups are formed and the reaction products are hydroxyalkylidene-1,1-diphosphonic acids, chloroalkylidene-1,1-diphosphonic acids, and derivatives of vinylphosphonic acid (Scheme 1c). It is worth mentioning that 1phosphonoalkylium ions 9, which may be intermediates in the deamination reaction of 1-aminoalkylphosphonic acids 1, have also not been extensively studied in the literature. Only theoretical calculations for the simplest phosphonomethylium ion (9h) (which exist in the cyclic form 10h) have been described by Pasto (Scheme 2a) [54]. On the other hand, Creary et al. studied the formation of carbenium ions substituted with phosphonic ester group 11 in the solvolysis reactions of mesylates 12 [55][56][57][58]. Experiments on the α-deuterium isotope effect proved that intermediates have an open form 11 and that no cyclic compounds 13 are formed (Scheme 2b). Intrigued by the very scarce literature reports on the deamination of 1-aminoalkylphosphonic acids, and interested in revealing the reactivity of the 1-phosphonoalkylium ions, possible intermediates in deamination of 1-aminoalkylphosphonic acids, we decided to study this interesting reaction in greater detail (Scheme 1d).
Molecules 2022, 27, x FOR PEER REVIEW 3 2a) [54]. On the other hand, Creary et al. studied the formation of carbenium ions su tuted with phosphonic ester group 11 in the solvolysis reactions of mesylates 12 [55 Experiments on the α-deuterium isotope effect proved that intermediates have an o form 11 and that no cyclic compounds 13 are formed (Scheme 2b). Intrigued by the scarce literature reports on the deamination of 1-aminoalkylphosphonic acids, and i ested in revealing the reactivity of the 1-phosphonoalkylium ions, possible intermed in deamination of 1-aminoalkylphosphonic acids, we decided to study this interestin action in greater detail (Scheme 1d).

Scheme 2.
Possible structures of the 1-phosphonoalkylium ions known from the literature [54 Herein we present the results of our detailed study on the deamination reactio structurally diverse 1-aminoalkylphosphonic acids carried out with nitrous acid. The sented results show the potential application of this transformation in organic synth and shed light on the possible reaction mechanism and reaction intermediates.

Results
For our study, we selected a representative and structurally diverse palette of 1noalkylphosphonic acids ( Figure 1, 17 examples). The selected examples include p phorus analogs of such amino acids as alanine 1a, valine 1b, leucine 1d, glycine 1h, nylalanine 1g, and phenylglycine 1f. Herein we present the results of our detailed study on the deamination reaction of structurally diverse 1-aminoalkylphosphonic acids carried out with nitrous acid. The presented results show the potential application of this transformation in organic synthesis and shed light on the possible reaction mechanism and reaction intermediates.

Results
For our study, we selected a representative and structurally diverse palette of 1aminoalkylphosphonic acids ( Figure 1, 17 examples). The selected examples include phosphorus analogs of such amino acids as alanine 1a, valine 1b, leucine 1d, glycine 1h, phenylalanine 1g, and phenylglycine 1f.

Acids-Preliminary Experiments
We started our preliminary experiments using the conditions applied for the diazotization of 2-aminoalkanoic acids (NaNO 2 , 5M HCl) (Scheme 3) [59]. Preliminary experiments clearly showed that 1-aminoalkylphosphonic acids reacted with nitrous acid (HNO 2 ), generated in situ from sodium nitrite (NaNO 2 ), differently than the tested amino acids. The degree of conversion in the case of 1-aminoalkylphosphonic was slightly higher than in the case of classical amino acids. No other products than the ones depicted on Scheme 3 were observed and they were additionally accompanied by unreacted starting material. Under the examined conditions, no selectivity towards chloride ions was observed and 1-hydroxyalkylphosphonic acids were the main reaction products.
Moreover, in the case of amino acids 2, as expected, the main reaction products were 1-hydroxy or 1-chloroalkanoic acids, while in the case of 1-aminoalkylphosphonic 1 a greater number of reaction products, including rearrangement and fragmentation products, were observed (Scheme 3).

Diazotization of 2-Aminoalkanoic acids vs. 1-Aminoalkylphosphonic Acids-Preliminar Experiments
We started our preliminary experiments using the conditions applied for the d tization of 2-aminoalkanoic acids (NaNO2, 5M HCl) (Scheme 3) [59]. Preliminary ex ments clearly showed that 1-aminoalkylphosphonic acids reacted with nitrous (HNO2), generated in situ from sodium nitrite (NaNO2), differently than the tested am acids. The degree of conversion in the case of 1-aminoalkylphosphonic was slightly hi than in the case of classical amino acids. No other products than the ones depicte Scheme 3 were observed and they were additionally accompanied by unreacted sta material. Under the examined conditions, no selectivity towards chloride ions was served and 1-hydroxyalkylphosphonic acids were the main reaction products.
Moreover, in the case of amino acids 2, as expected, the main reaction products 1-hydroxy or 1-chloroalkanoic acids, while in the case of 1-aminoalkylphosphonic greater number of reaction products, including rearrangement and fragmentation p ucts, were observed (Scheme 3).
Due to the complex composition of post-reaction mixtures, we decided to modif original reaction conditions used for the diazotization of amino acids. Expecting to ob complex reaction mixtures, we wanted to focus first on generating the carbenium ions then observe their reactivity with just a limited number of nucleophiles to simplify analysis of the results. Based on the literature data describing the diazotization of am acids, we envisaged that the most important parameter is the initial pH of the reac mixture [60][61][62][63]. Lowering the pH should increase the concentration of the electrop nitrosating agent, but at the same time causes the protonation of the amino group in starting 1-aminoalkylphosphonic acids, which lowers the nucleophile concentration. ditionally, we have assumed that 1-aminoalkylphosphonic acids are strong enough a to generate the nitrosating agent in situ from sodium nitrite in water, therefore there need to use hydrochloric acid in the reaction. After this simplification, the only nu philes in the reaction mixture were nitrite ions and water. Molecules 2022, 27, x FOR PEER REVIEW 5 Scheme 3. Preliminary experiments on the diazotization of alanine (2a) and valine (2b) and corresponding phosphorus analogs 1a and 1b.
We have observed that the product distribution in these reactions depended stro Due to the complex composition of post-reaction mixtures, we decided to modify the original reaction conditions used for the diazotization of amino acids. Expecting to obtain complex reaction mixtures, we wanted to focus first on generating the carbenium ions and then observe their reactivity with just a limited number of nucleophiles to simplify the analysis of the results. Based on the literature data describing the diazotization of amino acids, we envisaged that the most important parameter is the initial pH of the reaction mixture [60][61][62][63]. Lowering the pH should increase the concentration of the electrophilic nitrosating agent, but at the same time causes the protonation of the amino group in the starting 1-aminoalkylphosphonic acids, which lowers the nucleophile concentration. Additionally, we have assumed that 1-aminoalkylphosphonic acids are strong enough acids to generate the nitrosating agent in situ from sodium nitrite in water, therefore there is no need to use hydrochloric acid in the reaction. After this simplification, the only nucleophiles in the reaction mixture were nitrite ions and water.
We have observed that the product distribution in these reactions depended strongly on the structure of the starting 1-aminoalkylphosphonic acid 1, therefore the reaction results are outlined in Tables 1-4, according to the structure of the substrates used.
To avoid the formation of secondary products, the crude post-reaction mixtures were analyzed directly by NMR spectroscopy without isolation of the reaction products, and thus the results are given in the form of conversion. In all cases, the structures of reaction products were confirmed by NMR spectroscopy (especially 31 P NMR and 1 H NMR) by the addition of known reference compounds (synthesized separately) or by analysis and comparison of the NMR spectra of the crude reaction mixture with spectra of products known from the literature (see Supplementary Materials for more details).
Substitution was generally the main reaction for most of the investigated 1-aminoalkylphosphonic acids 1 (Tables 1 and 2), especially for those that do not have protons in the β-position (1f, 1n, 1h). For example, in the reaction of amino(phenyl)methylphosphonic acid (1f) the conversion of substrate to hydroxy(phenyl)phosphonic acid (5f) was 97% ( Table 2, entry 1).
In turn, elimination was the major reaction for 1-aminoalkylphosphonic acids 1q, 1l, and 1i which have bulky substituents (Table 3). For substrates 1l and 1i, two isomers of vinylphosphonic acid derivatives were formed: 7l, 7 l for 1l and 7i, 7 i for 1i. We assume that in this case the steric hindrance impedes the access of nucleophiles and, as a result, the elimination reaction is favored.
Furthermore, for substrates 1j, 1b, and 1k that have β-position migrating groups, the major reaction product was phosphoric acid (H 3 PO 4 ), accompanied by various amounts of substitution products 5 and rearrangement products 5 .
While direct substitution on the diazonium group in 1-phosphonoalkenediazonium salts 8 cannot be excluded (Scheme 4a), the complex composition of the post-reaction mixtures suggests that 1-phosphonoaklylium ions 9 may be intermediates in the diazotization reaction of 1-aminoalkylphosphonic acids 1. This assumption is supported by the fact that all typical products of carbenium ion reactions, especially rearrangement products 5 , were observed simultaneously in the crude post-reaction mixtures (Scheme 4b). It has to be mentioned that the accepted mechanism of deamination of analogous aliphatic 2-aminoalkanoic acids assumes the presence of α-lactones as reaction intermediates (Scheme 1a). As postulated, their formation is the reason for the high enantioselectivity of these reactions. By analogy, in the reaction of 1-aminoalkylphosphonic acids similar cyclic intermediates, namely 2-hydroxy-2-oxa-1,2-oxaphosphiranes 10, could also be postulated (Scheme 4c). However, there is no experimental information about intermediate 10 described thus far in the literature. In addition, our results indicate that the formation of 10 is unlikely. For example, the reaction products of 3-amino-3-phosphonopropanoic acid (1e) with nitrous acid may be explained by the assumption that 1-phosphonoalkylium ion 9e is formed (Scheme 5). The 3-hydroxy-3-phosphonopropanoic acid (5e) is formed in the reaction of nucleophile (water) addition to 1-phosphonoalkylium ion 9e, while (E)-3-phosphonoacrylic acid (7e) is formed as the result of proton elimination from 9e. Carbenium ion 9e also undergoes fragmentation and as a result, vinylphosphonic acid (7a) and carbon dioxide are formed. Table 1. Reaction of HNO 2 with 1-aminoalkylphosphonic acids 1 that are stabilized by substituents in 1-position or those that cannot rearrange a .
To avoid the formation of secondary products, the crude post-reaction mixtures were analyzed directly by NMR spectroscopy without isolation of the reaction products, and thus the results are given in the form of conversion. In all cases, the structures of reaction products were confirmed by NMR spectroscopy (especially 31 P NMR and 1 H NMR) by the addition of known reference compounds (synthesized separately) or by analysis and comparison of the NMR spectra of the crude reaction mixture with spectra of products known from the literature (see Supplementary Materials for more details).
Substitution was generally the main reaction for most of the investigated 1-aminoalkylphosphonic acids 1 (Tables 1 and 2), especially for those that do not have protons in the β-position (1f, 1n, 1h). For example, in the reaction of amino(phenyl)methylphosphonic acid (1f) the conversion of substrate to hydroxy(phenyl)phosphonic acid (5f) was 97% ( Table 2, entry 1).         COOH H 4% a Reaction conditions: 1-aminoalkylphosphonic acid (1.0 mmol), NaNO2 (2.0 mmol), evolution of N2 occurs, 21 °C, and NMR analysis of crude reaction mixture; b Conversions calculated based on 31 P NMR (recorded in D2O) of the crude reaction mixture.     In turn, elimination was the major reaction for 1-aminoalkylphosphonic acids 1q, 1l, and 1i which have bulky substituents (Table 3). For substrates 1l and 1i, two isomers of vinylphosphonic acid derivatives were formed: 7l, 7′l for 1l and 7i, 7′i for 1i. We assume that in this case the steric hindrance impedes the access of nucleophiles and, as a result, the elimination reaction is favored.
Furthermore, for substrates 1j, 1b, and 1k that have -position migrating groups, the major reaction product was phosphoric acid (H3PO4), accompanied by various amounts of substitution products 5 and rearrangement products 5′. In turn, elimination was the major reaction for 1-aminoalkylphosphonic acids 1q, 1l, and 1i which have bulky substituents (Table 3). For substrates 1l and 1i, two isomers of vinylphosphonic acid derivatives were formed: 7l, 7′l for 1l and 7i, 7′i for 1i. We assume that in this case the steric hindrance impedes the access of nucleophiles and, as a result, the elimination reaction is favored.
Furthermore, for substrates 1j, 1b, and 1k that have -position migrating groups, the major reaction product was phosphoric acid (H3PO4), accompanied by various amounts of substitution products 5 and rearrangement products 5′.   In turn, elimination was the major reaction for 1-aminoalkylphosphonic acids 1q, 1l, and 1i which have bulky substituents (Table 3). For substrates 1l and 1i, two isomers of vinylphosphonic acid derivatives were formed: 7l, 7′l for 1l and 7i, 7′i for 1i. We assume that in this case the steric hindrance impedes the access of nucleophiles and, as a result, the elimination reaction is favored.
Furthermore, for substrates 1j, 1b, and 1k that have -position migrating groups, the major reaction product was phosphoric acid (H3PO4), accompanied by various amounts of substitution products 5 and rearrangement products 5′.  In turn, elimination was the major reaction for 1-aminoalkylphosphonic acids 1q, 1l, and 1i which have bulky substituents (Table 3). For substrates 1l and 1i, two isomers of vinylphosphonic acid derivatives were formed: 7l, 7′l for 1l and 7i, 7′i for 1i. We assume that in this case the steric hindrance impedes the access of nucleophiles and, as a result, the elimination reaction is favored.
Furthermore, for substrates 1j, 1b, and 1k that have -position migrating groups, the major reaction product was phosphoric acid (H3PO4), accompanied by various amounts of substitution products 5 and rearrangement products 5′.  In turn, elimination was the major reaction for 1-aminoalkylphosphonic acids 1q, 1l, and 1i which have bulky substituents (Table 3). For substrates 1l and 1i, two isomers of vinylphosphonic acid derivatives were formed: 7l, 7′l for 1l and 7i, 7′i for 1i. We assume that in this case the steric hindrance impedes the access of nucleophiles and, as a result, the elimination reaction is favored.
Furthermore, for substrates 1j, 1b, and 1k that have -position migrating groups, the major reaction product was phosphoric acid (H3PO4), accompanied by various amounts of substitution products 5 and rearrangement products 5′.  In turn, elimination was the major reaction for 1-aminoalkylphosphonic acids 1q, 1l, and 1i which have bulky substituents (Table 3). For substrates 1l and 1i, two isomers of vinylphosphonic acid derivatives were formed: 7l, 7′l for 1l and 7i, 7′i for 1i. We assume that in this case the steric hindrance impedes the access of nucleophiles and, as a result, the elimination reaction is favored.
Furthermore, for substrates 1j, 1b, and 1k that have -position migrating groups, the major reaction product was phosphoric acid (H3PO4), accompanied by various amounts of substitution products 5 and rearrangement products 5′.  In turn, elimination was the major reaction for 1-aminoalkylphosphonic acids 1q, 1l, and 1i which have bulky substituents (Table 3). For substrates 1l and 1i, two isomers of vinylphosphonic acid derivatives were formed: 7l, 7′l for 1l and 7i, 7′i for 1i. We assume that in this case the steric hindrance impedes the access of nucleophiles and, as a result, the elimination reaction is favored.
Furthermore, for substrates 1j, 1b, and 1k that have -position migrating groups, the major reaction product was phosphoric acid (H3PO4), accompanied by various amounts of substitution products 5 and rearrangement products 5′.  In turn, elimination was the major reaction for 1-aminoalkylphosphonic acids 1q, 1l, and 1i which have bulky substituents (Table 3). For substrates 1l and 1i, two isomers of vinylphosphonic acid derivatives were formed: 7l, 7′l for 1l and 7i, 7′i for 1i. We assume that in this case the steric hindrance impedes the access of nucleophiles and, as a result, the elimination reaction is favored.
Furthermore, for substrates 1j, 1b, and 1k that have -position migrating groups, the major reaction product was phosphoric acid (H3PO4), accompanied by various amounts of substitution products 5 and rearrangement products 5′.  In turn, elimination was the major reaction for 1-aminoalkylphosphonic acids 1q, 1l, and 1i which have bulky substituents (Table 3). For substrates 1l and 1i, two isomers of vinylphosphonic acid derivatives were formed: 7l, 7′l for 1l and 7i, 7′i for 1i. We assume that in this case the steric hindrance impedes the access of nucleophiles and, as a result, the elimination reaction is favored.
Furthermore, for substrates 1j, 1b, and 1k that have -position migrating groups, the major reaction product was phosphoric acid (H3PO4), accompanied by various amounts of substitution products 5 and rearrangement products 5′.   In turn, elimination was the major reaction for 1-aminoalkylphosphonic acids 1q, 1l, and 1i which have bulky substituents (Table 3). For substrates 1l and 1i, two isomers of vinylphosphonic acid derivatives were formed: 7l, 7′l for 1l and 7i, 7′i for 1i. We assume that in this case the steric hindrance impedes the access of nucleophiles and, as a result, the elimination reaction is favored.
Furthermore, for substrates 1j, 1b, and 1k that have -position migrating groups, the major reaction product was phosphoric acid (H3PO4), accompanied by various amounts of substitution products 5 and rearrangement products 5′.  In turn, elimination was the major reaction for 1-aminoalkylphosphonic acids 1q, 1l, and 1i which have bulky substituents (Table 3). For substrates 1l and 1i, two isomers of vinylphosphonic acid derivatives were formed: 7l, 7′l for 1l and 7i, 7′i for 1i. We assume that in this case the steric hindrance impedes the access of nucleophiles and, as a result, the elimination reaction is favored.
Furthermore, for substrates 1j, 1b, and 1k that have -position migrating groups, the major reaction product was phosphoric acid (H3PO4), accompanied by various amounts of substitution products 5 and rearrangement products 5′.  In turn, elimination was the major reaction for 1-aminoalkylphosphonic acids 1q, 1l, and 1i which have bulky substituents (Table 3). For substrates 1l and 1i, two isomers of vinylphosphonic acid derivatives were formed: 7l, 7′l for 1l and 7i, 7′i for 1i. We assume that in this case the steric hindrance impedes the access of nucleophiles and, as a result, the elimination reaction is favored.
Furthermore, for substrates 1j, 1b, and 1k that have -position migrating groups, the major reaction product was phosphoric acid (H3PO4), accompanied by various amounts of substitution products 5 and rearrangement products 5′.  In turn, elimination was the major reaction for 1-aminoalkylphosphonic acids 1q, 1l, and 1i which have bulky substituents (Table 3). For substrates 1l and 1i, two isomers of vinylphosphonic acid derivatives were formed: 7l, 7′l for 1l and 7i, 7′i for 1i. We assume that in this case the steric hindrance impedes the access of nucleophiles and, as a result, the elimination reaction is favored.
Furthermore, for substrates 1j, 1b, and 1k that have -position migrating groups, the major reaction product was phosphoric acid (H3PO4), accompanied by various amounts of substitution products 5 and rearrangement products 5′.    3 Me H 8% a Reaction conditions: 1-aminoalkylphosphonic acid (1.0 mmol), NaNO2 (2.0 mmol), evolution of N2 occurs, 21 °C , and NMR analysis of crude reaction mixture; b Conversions calculated based on 31 P NMR (recorded in D2O) of the crude reaction mixture.    While direct substitution on the diazonium group in 1-phosphonoalkenediazonium salts 8 cannot be excluded (Scheme 4a), the complex composition of the post-reaction mixtures suggests that 1-phosphonoaklylium ions 9 may be intermediates in the diazotization reaction of 1-aminoalkylphosphonic acids 1. This assumption is supported by the fact that all typical products of carbenium ion reactions, especially rearrangement products 5′, were observed simultaneously in the crude post-reaction mixtures (Scheme 4b). It has to be mentioned that the accepted mechanism of deamination of analogous aliphatic 2-aminoalkanoic acids assumes the presence of α-lactones as reaction intermediates (Scheme 1a). As postulated, their formation is the reason for the high enantioselectivity of these reactions. By analogy, in the reaction of 1-aminoalkylphosphonic acids similar cyclic intermediates, namely 2-hydroxy-2-oxa-1,2-oxaphosphiranes 10, could also be postulated (Scheme 4c). However, there is no experimental information about intermediate 10 described thus far in the literature. In addition, our results indicate that the formation of 10 is unlikely. For example, the reaction products of 3-amino-3-phosphonopropanoic acid (1e) with nitrous acid may be explained by the assumption that 1-phosphonoalkylium ion 9e is formed (Scheme 5). The 3-hydroxy-3-phosphonopropanoic acid (5e) is formed in the reaction of nucleophile (water) addition to 1-phosphonoalkylium ion 9e, while (E)-3-phosphonoacrylic acid (7e) is formed as the result of proton elimination from 9e. Carbenium ion 9e also undergoes fragmentation and as a result, vinylphosphonic acid (7a) and carbon dioxide are formed. An interesting example illustrating the complexity of the deamination reaction of 1-aminoalkylphosphonic acid 1 is the reaction of 1-amino-2-phenylethylphosphonic acid (1g) with HNO 2 (Scheme 6). Among the expected products of substitution 5g, elimination 7g, and phosphoric acid, in the post-reaction mixture, the rearranged 2-hydroxy-1phenylethylphosphonic acid (5 g) was identified. Considering the formation of carbenium ion 9g we expected the rearrangement of this carbenium ion to 9 g, which should be more stable due to the stabilizing effect of the phenyl group. Subsequent addition of nucleophile (H 2 O) to both carbenium ions should lead to the corresponding hydroxyalkylphosphonic acids 5g and 5 g respectively (Scheme 6). However, analysis of the NMR spectra of the crude reaction mixture revealed that the second product of the reaction is not the 5 g but 5"g (see Supplementary Materials for more details).
Formation of 2-hydroxy-1-phenylphosphonic acid (5 g), as well as unrearranged 5g and phosphoric acid may be explained by the formation of cyclic intermediate 9 g (Scheme 6a). The nucleophilic attack of water on the less crowded side (pink arrow on Scheme 6a) of intermediate 9 g yields rearranged 2-hydroxyalkylphosphonic acid 5 g, while fragmentation of 9 g (Scheme 6b) yields styrene and metaphosphoric acid which hydrolyses to phosphoric acid. Finally, when examining the reactivity of 1-aminoalkylphosphonic acids 1 in a deamination reaction with HNO 2 , in every reaction we have always observed the presence of various amounts of phosphoric acid (H 3 PO 4 ). We postulate that the formation of H 3 PO 4 could be explained by two reaction mechanisms which depend on the structure of the used 1-aminoalkylphosphonic acids 1 (Schemes 7 and 8). According to the first reaction mechanism (Scheme 7a), if the structure of the formed 1-phosphonoalkylium ion 9 enables its rearrangement to the more stable 2-phosphonoalkylium ion 9 (compounds  1j, 1b, 1k, 1g), then ion 9 can further undergo fragmentation with cleavage of the C-P bond resulting in the formation of alkene and metaphosphoric acid (that undergo hydrolysis to phosphoric acid in the presence of water). A similar mechanism was proposed by Mastalerz and Richtarski for the deamination of 2-aminoethylphosphonic acid and related compounds, where the main reaction products were ethylene and phosphoric acid (Scheme 7b) [64][65][66].  nylethylphosphonic acid (5′'g) was identified. Considering the formation of carbenium ion 9g we expected the rearrangement of this carbenium ion to 9′g, which should be more stable due to the stabilizing effect of the phenyl group. Subsequent addition of nucleophile (H2O) to both carbenium ions should lead to the corresponding hydroxyalkylphosphonic acids 5g and 5′g respectively (Scheme 6). However, analysis of the NMR spectra of the crude reaction mixture revealed that the second product of the reaction is not the 5′g but 5"g (see Supplementary Materials for more details). Formation of 2-hydroxy-1-phenylphosphonic acid (5"g), as well as unrearranged 5g and phosphoric acid may be explained by the formation of cyclic intermediate 9"g (Scheme 6a). The nucleophilic attack of water on the less crowded side (pink arrow on Scheme 6a) of intermediate 9"g yields rearranged 2-hydroxyalkylphosphonic acid 5"g, while fragmentation of 9"g (Scheme 6b) yields styrene and metaphosphoric acid which hydrolyses to phosphoric acid. Finally, when examining the reactivity of 1-aminoalkylphosphonic acids 1 in a deamination reaction with HNO2, in every reaction we have always observed the presence of various amounts of phosphoric acid (H3PO4). We postulate that the formation of H3PO4 could be explained by two reaction mechanisms which depend on the structure of the used 1-aminoalkylphosphonic acids 1 (Schemes 7 and 8). According to the first reaction mechanism (Scheme 7a), if the structure of the formed 1phosphonoalkylium ion 9 enables its rearrangement to the more stable 2-phosphonoalkylium ion 9′ (compounds 1j, 1b, 1k, 1g), then ion 9′ can further undergo fragmentation with cleavage of the C-P bond resulting in the formation of alkene and metaphosphoric acid The second reaction mechanism should explain the formation of H 3 PO 4 in the case where there is no possibility of rearrangement of the formed carbenium ion 9 (Scheme 8), especially for the reaction of substrates 1a, 1f, 1n, and 1h. By analogy to reactions of 2-aminoalkanoic acids with HNO 2 [67], the reaction of 1-phosphonoalkylium ion 9 with biphilic nitrite ion (NO 2 -) gives 1-nitroalkylphosphonic acid 14 (Scheme 8a) or nitrite ester of 1-hydroxyalkylphosphonic acids 15 (Scheme 8b). Compounds 14 and 15 may undergo secondary reactions which ultimately produce phosphoric acid. Scheme 7. Mechanism of H3PO4 formation in rearrangement and fragmentation reaction [64][65][66]. The second reaction mechanism should explain the formation of H3PO4 in the case where there is no possibility of rearrangement of the formed carbenium ion 9 (Scheme 8), especially for the reaction of substrates 1a, 1f, 1n, and 1h. By analogy to reactions of 2aminoalkanoic acids with HNO2 [67], the reaction of 1-phosphonoalkylium ion 9 with biphilic nitrite ion (NO2 -) gives 1-nitroalkylphosphonic acid 14 (Scheme 8a) or nitrite ester of 1-hydroxyalkylphosphonic acids 15 (Scheme 8b). Compounds 14 and 15 may undergo secondary reactions which ultimately produce phosphoric acid.

General Information
The 1 H, 13 C{ 1 H}, 31 P NMR, and DEPT-135 spectra were collected on a Jeol 400yh instrument (Jeol, Ltd., Tokio, Japan) (400 MHz for 1 H NMR, 162 MHz for 31 P NMR, and 100 MHz for 13 C NMR) and were processed with dedicated software (Delta 5.0.5). NMR experiments recorded in D 2 O were referenced to the respective residual 1 H signal of the solvent. Multiplicities were reported using the following abbreviations: s (singlet), d (doublet), t (triplet), q (quartet), and m (multiplet). The reported coupling constants (J) values were those observed from the splitting patterns in the spectrum and may not reflect the true coupling constant values. The composition of post-reaction mixtures (as the conversion of substrate to the given product) was calculated based on 31 P NMR (recorded in D 2 O) of the crude reaction mixture. Structural assignments of 5 g were made with additional information from gCOSY, gHSQC, and gHMBC experiments.

Reagents
Aminomethylphosphonic acid (1h) was obtained in the reaction of benzamide, formaldehyde, and phosphorous trichloride [68]. 3-Amino-3-phosphonopropanoic acid (1e) was synthesized from diethyl acetamidomethylenemalonate [69]. The remaining 1-aminoalkylphosphonic acids 1 were obtained in the reaction of an appropriate carbonyl compound with acetamide, acetyl chloride, and PCl 3 in acetic acid, using Soroka's protocol [70]. 1-Hydroxyalkylphosphonic acids 5, which were used as reference materials for confirmation of reaction products structures, were synthesized by dealkylation of diethyl 1-hydroxyalkylphosphonates, which were obtained in the reaction of triethyl phosphite with suitable aldehyde or ketone and hydrogen chloride [71].

Deamination of 1-Aminoalkylphosphonic Acids 1 and 2-Aminoalkanoic Acids 2 in 5M HCl
The deamination experiments were conducted in a three-necked flask equipped with a reflux condenser, thermometer, dropping funnel, and magnetic stirrer, as described in the original protocol [40]. The solution of 1-aminoalkylphosphonic acid 1 or 2-aminoalkanoic acid 2 (10 mmol) in 5M HCl (65 mmol, 13 mL) was cooled in an ice/NaCl cooling bath to a temperature of −12 • C. Subsequently, 4 M NaNO 2 solution in water (16 mmol, 4.0 mL) was added dropwise for 2 min. The temperature of the reaction mixture was maintained under 0 • C for 5 h, and then at 25 • C for 12 h. The samples for 1 H and 31 P NMR spectra were prepared by diluting post-reaction mixtures (0.10 mL) in D 2 O (0.40 mL). The samples were re-measured after the addition of reference materials. The composition of the mixture was calculated based on the integration of signals on the 31 P NMR spectra (for phosphorous substrates) or on the 1 H NMR spectra (for 2-aminoalkanoic acids).

Deamination of 1-Aminoalkylphosphonic Acids 1 in Water
The deamination reactions of 1-aminoalkylphosphonic acids 1 were conducted in a round-bottom flask equipped with a magnetic stirrer and calibrated gas burette ( Figure S11 in Supplementary Materials). The flask was placed in a water bath at a temperature of about 20 • C. 1-Aminoalkylphosphonic acid 1 (3.0 mmol) was added to a 0.67 M solution of NaNO 2 (6.0 mmol, 9.0 mL). The solution or suspension was stirred by the means of a magnetic stirrer until the stoichiometric volume of gas was evolved, and additionally for 12 h. The 1 H and 31 P NMR spectra were recorded after that time and additionally after a few days. The composition of the mixture was calculated based on the integration of signals on the 31 P NMR spectra.

Conclusions
We have studied the deamination of 17 1-aminoalkylphosphonic acids 1 in the reaction with nitrous acid. We have postulated that 1-phosphonoalkylium ions 9 are plausible reactive intermediates in these reactions. Depending on the structure of 1-aminoalkylphosphonic acid 1 used, these ions 9 react with a nucleophile (H 2 O or NO 2 ), undergo elimination of protons, or a rearrangement/fragmentation reaction (Scheme 9). Furthermore, we explained the formation of the phosphoric acid (H 3 PO 4 ), present in every reaction mixture, through two mechanisms (Schemes 6 and 7). We have experimentally demonstrated that the selectivity of the reaction of 1-phosphonoalkylium ions 9 is not easy to control but, in some cases, the addition of nucleophile (H 2 O) is the major reaction and the starting 1-aminoalkylphosphonic acids 1 could be transformed into 1-hydroxyphosphonic acids 5 (Scheme 9). In turn, the derivatives of vinylphosphonic acid 7 resulting from proton elimination from 1-phosphonoalkylium ions 9 (Scheme 9) could be major products in the case of 1-aminoalkylphosphonic acids having a positive charge positioned at the tertiary carbon atom and surrounded by bulky substituents, such as compounds 1q, 1l, and 1i (Scheme 9, Table 3). Finally, if the generated 1-phosphonoalkylium ions 9 have migrating groups in the β-position, such as in compounds 9j, 9b, 9k, and 9g, they can further rearrange to more stable 2-phosphonoalkylium ions 9 and either react with a nucleophile to form 2hydroxyalkylphosphonic acid 5 or undergo fragmentation to alkene and H 3 PO 4 (Scheme 9, Table 4). Although the reported procedure of the deamination of 1-aminoalkylphosphonic 1 generally may have limited synthetic application, in specific cases, it may be an irreplaceable synthetic method leading to the desired products. migrating groups in the β-position, such as in compounds 9j, 9b, 9k, and 9g, they further rearrange to more stable 2-phosphonoalkylium ions 9′ and either react with a cleophile to form 2-hydroxyalkylphosphonic acid 5′ or undergo fragmentation to al and H3PO4 (Scheme 9, Table 4). Although the reported procedure of the deamination aminoalkylphosphonic 1 generally may have limited synthetic application, in spe cases, it may be an irreplaceable synthetic method leading to the desired products. Scheme 9. Possible transformations of 1-phosphonoalkylium ions 9 discussed in this study.

Conflicts of Interest:
The authors declare no conflict of interest. Scheme 9. Possible transformations of 1-phosphonoalkylium ions 9 discussed in this study.