Selective Esterification of Phosphonic Acids

Here, we report straightforward and selective synthetic procedures for mono- and diesterification of phosphonic acids. A series of alkoxy group donors were studied and triethyl orthoacetate was found to be the best reagent as well as a solvent for the performed transformations. An important temperature effect on the reaction course was discovered. Depending on the reaction temperature, mono- or diethyl esters of phosphonic acid were obtained exclusively with decent yields. The substrate scope of the proposed methodology was verified on aromatic as well as aliphatic phosphonic acids. The designed method can be successfully applied for small- and large-scale experiments without significant loss of selectivity or reaction yield. Several devoted experiments were performed to give insight into the reaction mechanism. At 30 °C, monoesters are formed via an intermediate (1,1-diethoxyethyl ester of phosphonic acid). At higher temperatures, similar intermediate forms give diesters or stable and detectable pyrophosphonates which were also consumed to give diesters. 31P NMR spectroscopy was used to assign the structure of pyrophosphonate as well as to monitor the reaction course. No need for additional reagents and good accessibility and straightforward purification are the important aspects of the developed protocols.


Introduction
Organophosphonates are a relatively diverse and common group of compounds, which feature the presence of a stable C-P bond. In nature, they are present in microorganisms as well as in metazoans. Furthermore, they constitute a considerable part of phosphorus compounds found in oceans [1]. It is also considered that organophosphonates could play an important role at the early stages of the development of life on Earth when the oxygen concentration in the atmosphere was negligible [2]. Currently, phosphonates have a large number of applications in agriculture [3] and medicine [4,5], as well as markers of chemical warfare agents (CWAs), and many others [6].
An alternative route for target compounds is direct esterification or transesterification of the phosphonic acids. The esterification can be achieved in the presence of chlorinated silica gel [21], by exploitation of Garegg-Samuelsson conditions [22], dicyclohexylcarbodiimide mediated esterification [23] or O-alkylation [24], whereas alkyl hydrogen drogen phosphonates can be received from the diesters by partial hydrolysis [20,25] or via transesterification with silyl halides and selective cleavage by protic solvents [26]. Monoesters can also be obtained directly from phosphonic acids upon selective monoesterification. For successful syntheses, phenylarsonic acid [27,28], carbodiimides [29,30], solid-phase synthesis [31] and ionic liquids [24] were used. Unfortunately, none of the mentioned methods can be performed under mild conditions, with high yields and good selectivity towards monoesters.
Recently, we have found that esterification of carboxylic acid proceeds efficiently upon the application of orthoesters [32], carbonates [33], orthocarbonates [34], acetals and ketals [35] as alkoxy group donors. The studies on the reaction mechanism proved that dialkoxy carboxylate is a key intermediate that is subsequently and irreversibly transformed into ethyl acetate which is a driving force for the reaction [32]. Since phosphonic acids are the isosteres of carboxylic acids, the same principle is proposed for the esterification of phosphonic acids (1) using corresponding alkoxy group donors (2). The esterification of phosphonic acids is more complicated since during the reaction two possible products, mono-(3) and diesters (4), can be formed simultaneously (Scheme 1). The straightforward synthesis of diester 4 is obvious [36][37][38][39] while the selective formation of monoesters is very tricky. The application of alkoxy group donors in monoesterification of phosphonic acids has never been reported before. Scheme 1. The general concept of esterification of phosphonic acids.

Results and Discussion
For preliminary experiments, both butylphosphonic (1a) and phenylphosphonic (1b) acids were selected as model substrates. Initially, various potential alkoxy group donors were investigated in the monoesterification reaction (Table 1). Their structures are presented in Figure 1. The reactions were conducted in methyl tert-butyl ether (MTBE) at 30 °C for 24 h. The conversion of substrates and selectivity toward products were determined by 1 H and 31 P NMR analyses.
In the next step of our research, we tested the impact of the reaction medium using 2a as a donor (Table 2), because the preliminary used MTBE was selected arbitrarily. For butylphosphonic acid 1a, when the reaction was conducted in tetrahydrofuran, a slight drop in reaction yield was observed with respect to MTBE while in the case of phenylphosphonic acid 1b, the result was similar ( Table 2, entries 1 and 2). In dichloromethane, conversion was high for acid 1a and almost quantitative for acid 1b, with good selectivity towards monoesters 3a and 3b (Table 2, entry 3). In ethyl acetate and acetonitrile, the conversions were lower for both substrates ( Table 2, entries 4 and 5). Acetone was found to be an inappropriate solvent with very low substrate conversion ( Table 2, entry 6). The reaction proceeds effectively in toluene (Table 2, entry 7) but the highest conversion (95%) was observed when 2a was used as a reagent and solvent ( Table 2, entry 8). No reaction progress was observed in experiments conducted in dimethyl sulfoxide and dimethylformamide ( Table 2, entries 9 and 10). DMSO 0 0 0 0 0 0 10 DMF 0 0 0 0 0 0 1 Reaction conditions: phosphonic acid 1a-b (1 equiv., 0.05 mmol), triethyl orthoacetate (3 equiv., 0.15 mmol) in 1 mL of solvent at 30 • C for 24 h. 2 The substrate conversion and product yield determined by 31 P NMR. 3 One milliliter of triethyl orthoacetate.
The analysis of data from Table 2 concludes that the best medium for the studied transformation is triethyl orthoacetate 2a when used in large excess with respect to substrate (over 5 equiv.). After selection of the best solvent for reaction, we studied the impact of temperature on the esterification ( Table 3). Reactions of butylphosphonic acid 1a with orthoacetate 2a were performed at temperatures varying from 30 • C to 100 • C. The increase in the reaction temperature from 30 to 40 • C resulted in a higher conversion of the substrate (over 99%) and slightly influenced the yield of the product (Table 3, entries 1 and 2). At elevated temperatures, the conversion of the substrate became quantitative (Table 3, entries 2-8) and the selectivity towards diester increased gradually. The amount of monoester 3a in the reaction mixture dropped to 1% at 90 • C (Table 3, entry 7). At 100 • C, partial decomposition of the product was observed ( Table 3, entry 8). In conclusion, the considered data indicate that at higher temperatures, diester 4a is formed as a single product (Table 3, entries 7 and 8). The best conditions for diesterification of butylphosphonic acid 1a involved reaction in excess of orthoester 2a at 90 • C. An analogous esterification performed at 30 • C led to the selective formation of monoethyl butylphosphonate (3a).
With the optimized reaction conditions in hand, we proceeded to assess the generality and scope of the developed procedures. Table 4 summarizes the results of the preparation of several mono-and diesters from various phosphonic acids. It should be noted that for the experiments performed on a laboratory scale, the favourable amount of orthoacetate 2a was 15 equiv. for monoesterification and 30 equiv. for diesterification. 1 Monoesterification and diesterification were conducted according to Sections 3.5 and 3.6, respectively. Reactions were monitored by 31 P NMR and terminated after complete conversion of substrate. 2 The yield of isolated product. 3 The substrate readily polymerized at elevated temperatures. 4 No reaction, due to insolubility of substrate.
It is important to note that using a single reagent-triethyl orthoacetate-two types of products, mono-(3) and diethylesters (4), were obtained under mild reaction conditions in a reasonable time without the use of any toxic or polluting reagents. In almost all cases, products were obtained as single compounds upon evaporation of the reaction mixture to dryness. The developed procedure partially fulfills green chemistry principles and can be applied to a fair range of phosphonic acids.
The gain insight into the reaction mechanism, additional experiments were performed on one of the model substrates-benzylphosphonic acid 1b-with selected donor 2a at 30 • and 75 • C. Close inspection of 31 P NMR spectra of crude 3b product obtained in reaction at 30 • C did not show any additional signals from possible reaction intermediates (see Supplementary Materials Section S2.3). However, in the reaction performed at the higher temperature, an interesting phenomenon was observed. Formally, before the reaction had started, the spectrum of the substrate 1b was measured. It is at the top of Figure 2 and shows a single peak at 21.3 ppm. After the addition of triethyl orthoacetate and heating the reaction mixture for 1 h at 75 • C, the NMR spectra changed substantially. The signal associated with substrate 1b diminished almost to zero and four new signals appeared on the spectra. The main peak at 26.5 ppm comes from diester 4b. The amount of product 4b increased rapidly during this first hour and reached a maximum after 24 h. The small and broad signal at 23.3 ppm represents monoester 3b. The most significant peak of the 3b product appeared in the first spectrum (after an hour). Further measurements demonstrated that the amount of 3b dropped close to zero after 8 h. Surprisingly, two new signals at 18.65 and 18.77 ppm of similar intensity appeared. Close inspection of literature data [40] allowed us to assign those two signals to diastereomers 6a and 6b of pyrophosphonate 6 ( Figure 3). The appearance of compound 6 as a reaction intermediate is a valuable clue to the reaction mechanism.
The calculated 31 P NMR yield of pyrophosphonate 6 reached 34% after 1 h and its amount remained at a similar level for the next 5 h and then slowly decreased to zero. The detailed data from the 31 P NMR spectra analysis are summarized in Table S1 (SI). To sum up, the NMR spectra recorded during the reaction course showed that the amount of diester 4b was continuously increasing with time while amounts of 3b and 6 reached a kind of plateau and then were diminished to zero after 24 h. As mentioned above, at a lower reaction temperature (30 • C) signals from pyrophosphonate 6 were not observed and a higher reaction temperature was required to transform monoester 3b into pyrophosphonate 6 and finally into diester 4b. The graphical representation of the results described above is shown in Figure 4. The calculated 31 P NMR yield of pyrophosphonate 6 reached 34% after 1 h and its amount remained at a similar level for the next 5 h and then slowly decreased to zero. The detailed data from the 31 P NMR spectra analysis are summarized in Table S1 (SI). To sum up, the NMR spectra recorded during the reaction course showed that the amount of diester 4b was continuously increasing with time while amounts of 3b and 6 reached a kind of plateau and then were diminished to zero after 24 h. As mentioned above, at a lower reaction temperature (30 °C) signals from pyrophosphonate 6 were not observed and a higher reaction temperature was required to transform monoester 3b into pyrophosphonate 6 and finally into diester 4b. The graphical representation of the results described above is shown in Figure 4. The calculated 31 P NMR yield of pyrophosphonate 6 reached 34% after 1 h and its amount remained at a similar level for the next 5 h and then slowly decreased to zero. The detailed data from the 31 P NMR spectra analysis are summarized in Table S1 (SI). To sum up, the NMR spectra recorded during the reaction course showed that the amount of diester 4b was continuously increasing with time while amounts of 3b and 6 reached a kind of plateau and then were diminished to zero after 24 h. As mentioned above, at a lower reaction temperature (30 °C) signals from pyrophosphonate 6 were not observed and a higher reaction temperature was required to transform monoester 3b into pyrophosphonate 6 and finally into diester 4b. The graphical representation of the results described above is shown in Figure 4.  The analysis of obtained data provided some insight into the reaction mechanism (Scheme 2). The first step is the reaction between triethyl orthoacetate (2a) and benzylphosphonic acid (1b) which leads to the formation of an unstable intermediate A. It readily decomposes to monoester 3b by releasing one molecule of ethanol and ethyl ac- The analysis of obtained data provided some insight into the reaction mechanism (Scheme 2). The first step is the reaction between triethyl orthoacetate (2a) and benzylphosphonic acid (1b) which leads to the formation of an unstable intermediate A. It readily decomposes to monoester 3b by releasing one molecule of ethanol and ethyl acetate in a two-step process, at room temperature. The intermediate A can be formed only from selected substrates used as alkoxy group donors, depicted in Figure 1. This mechanism allowed us to explain why only monophosphonates were formed from orthoesters 2a and 2b, diethoxymethyl acetate 2g and tetramethyl orthocarbonate 2h. At higher temperatures, monoester 3b reacts with a second molecule of triethyl orthoacetate to give intermediate B,  The analysis of obtained data provided some insight into the reaction mechanism (Scheme 2). The first step is the reaction between triethyl orthoacetate (2a) and benzylphosphonic acid (1b) which leads to the formation of an unstable intermediate A. It readily decomposes to monoester 3b by releasing one molecule of ethanol and ethyl acetate in a two-step process, at room temperature. The intermediate A can be formed only from selected substrates used as alkoxy group donors, depicted in Figure 1. This mechanism allowed us to explain why only monophosphonates were formed from orthoesters 2a and 2b, diethoxymethyl acetate 2g and tetramethyl orthocarbonate 2h. At higher temperatures, monoester 3b reacts with a second molecule of triethyl orthoacetate to give intermediate B, similar to A. This intermediate can react with the second molecule of monoester 3b to give compound C which is readily transformed into pyrophosphonate 6 and its structure was ambiguously assigned by kinetic experiments. Compound 6 reacts with the third molecule of orthoester 2a to give the final product diester 4b and intermediate B. According to this mechanism, for the formation of one molecule of diester 4, three molecules of orthoester 2a are required. This readily explains why the larger amount of triethyl orthoacetate favors double esterification with respect to monoesterification.

Scheme 2.
A plausible mechanism for esterification of phosphonic acid 1b with orthoester 2a.

General
Commercially available reagents were used without additional purification. Phosphonic acids were purchased or prepared according to reported literature procedures. The water and hexane mixtures were previously distilled. Other solvents (analytical grade) were used without extra drying and purification. Reactions were performed in dry laboratory glassware under an air atmosphere (otherwise noted) using magnetic stirrers. Solvents and volatile reagents were evaporated under reduced pressure. Merck (Darmstadt, Germany) silica gel plates 60 F254 were used for TLC analysis.
The majority of obtained compounds did not need additional purification after reaction workup. Those requiring isolation were separated via distillation under reduced pressure on Kugelrohr apparatus. Melting points were measured on a Boetius hot plate microscope (Nagema, Dresden, Germany). The phosphorous nuclear magnetic resonance ( 31 P{1H} NMR) spectra of analyzed compounds, dissolved in DMSO-d 6 , D 2 O or CDCl 3 , were recorded with a Bruker spectrometer (162.0 MHz) at 30 • C and using 85% H 3 PO 4 as external calibration. 1 H and 13 C NMR spectra were also recorded in DMSO-d 6 , D 2 O or CDCl 3 (30 • C) on the same Bruker spectrometer at 400 and 101 MHz, respectively. Chemical shifts were reported in ppm and referred to residual deuterated solvent signal; coupling constants (J) were noted in Hz. Low-resolution mass spectra were recorded on the API365i API 3000 spectrometer, and the ESI technique was used to analyte ionization. High-resolution mass spectra were recorded on the Maldi SYNAPT G2-S HDMS (Waters, Milford, MA, USA) apparatus with a QqTOF analyzer.

General Procedure for Esterification of Phosphonic Acid Using Various Alkoxy Group Donors
To a solution of butylphosphonic acid 1a (1 equiv., 7 mg, 0.05 mmol) or benzylphosphonic acid 1b (1 equiv., 9 mg, 0.05 mmol) in MTBE (1 mL), the respective alkoxy group donor 2 (3 equiv.) was added and stirred at 30 • C for 24 h. The solvent and other volatile compounds were evaporated under reduced pressure. The crude product was dissolved in 0.6 mL DMSO-d 6 and the 31 P NMR spectrum was recorded. Subsequently, the conversion of substrate 1 and yields of products 3 and 4 were determined on the basis of relative peak integrals.

General Procedure for Esterification of Phosphonic Acids Using Triethyl Orthoacetate in Different Solvents
To a solution of butylphosphonic acid 1a (1 equiv., 7 mg, 0.05 mmol) or benzylphosphonic acid 1b (1 equiv., 9 mg, 0.05 mmol), 1 mL of respective solvent triethyl orthoacetate 2a (3 equiv., 25 mg, 0.15 mmol) was added and stirred at 30 • C for 24 h. The solvent and other volatile substances were evaporated under reduced pressure. Obtained crude was dissolved in 0.6 mL DMSO-d 6 and the spectrum was recorded. The conversion of substrate 1 and yield of products 3 and 4 were determined based on 31 P NMR spectrum analysis.

The Influence of Temperature on the Selectivity of Esterification Reaction of Butylphosphonic acid with Triethyl Orthoacetate
Butylphosphonic 1a (1 equiv., 14 mg, 0.1 mmol)) in 1 mL orthoacetate 2a was heated at the temperature indicated in Table 3 for 24 h. After that, the reaction mixture was cooled to room temperature and the excess 2a, as well as volatile byproducts, was evaporated to dryness. Received crude was dissolved in DMSO-d 6 and the prepared sample was used for the 31 P NMR measurements, based on which the conversion of 1a and product 3a and 4a yields were determined.

General Procedure for Selective Monoesterification of Phosphonic Acids 1 Using Triethyl Orthoacetate 2a Leading to Ethyl Hydrogen Phosphonates 3
Phosphonic acid 1 (1 mmol) and triethyl orthoacetate (15 mmol, 2.75 mL) were stirred at 30 • C and the substrate conversion was monitored by 31 P NMR. After the competition of the reaction, the orthoacetate excess and volatile by-product were evaporated off under vacuum. In several cases, the remaining product required no further purification. If necessary, the extraction procedure proposed by Campbell [41] was applied: the crude was diluted with diethyl ether (10 mL) and extracted twice with NaHCO 3 (5 mL, 5% w/v solution in water). The aqueous fractions were combined and washed with two portions of ethyl ether (2 × 5 mL). The aqueous phase was acidified with hydrogen chloride solution (1 M) to reach pH 2 and extracted three times with ethyl acetate (3 × 5 mL). Combined ethyl acetate fractions were dried over magnesium sulfate and pure products were received after solvent evaporation to dryness.
3.6. General Procedure for Esterification of Phosphonic Acids 1 Using Triethyl Orthoacetate Leading to Diethyl Phosphonic Esters 4 Phosphonic acid 1 (1 mmol) and triethyl orthoacetate (5.5 mL, 30 mmol) were stirred at 90 • C (or, for sparingly soluble substrates, at a higher temperature) for 24 h under a reflux condenser. The progress of the reaction was monitored by 31 P NMR. The excess of orthoester was evaporated under reduced pressure and the product-dialkyl phosphonate 4-was received in pure form. In some cases, the purity of diester 4 was not satisfactory and additional purification by bulb-to-bulb vacuum distillation with Kugelrohr apparatus or column chromatography was performed.

Ethyl Hydrogen Ethylphosphonate (3c)
The synthesis of 3c was conducted according to Section 3.5. Yield 89%; NMR data consistent with the previous report [44]; colorless oil; 1 H NMR (400 MHz, DMSO-d 6  The synthesis of 3f was conducted according to Section 3.5, with the only difference being that the reaction time was extended to 72 h. Yield 89%; NMR data in accordance with the previous report [46]; colorless thick oil; 1 H NMR (400 MHz, DMSO-d 6  The synthesis of 3g was conducted according to Section 3.5. Yield 56%; NMR data consistent with reported spectra [47]; yellowish oil. 1  Due to the poor solubility of the substrate 1j in triethyl orthoacetate and its low conversion, the reaction was carried out according to the modified Section 3.5. The reaction was conducted at 40 • C for 24 h. The excess of orthoester was then evaporated off, and 83% conversion was ascertained by 31 P NMR analysis, and the product was isolated via extraction. Yield: 29%; 1 H NMR is in accordance with foregoing literature data [49]; milky white solid; m.p. 144.1-144. utilized for direct esterification of substrates and only for triethyl or trimethyl orthoacetate, diethoxymethyl acetate and tetramethyl orthocarbonate were the expected products obtained. Triethyl orthoacetate (2a) was found to be the best reagent as well as a solvent for esterification reactions. In the present studies, we discovered the effect of temperature on selective esterification of phosphonic acids. For the reactions conducted, at 30 • C the dominant formation of monoethyl esters was observed, while at 90 • C diesters 4 were obtained in good to excellent yields. The generality and applicability of the new method were verified on thirteen aromatic and aliphatic phosphonic acids. For benzylphosphonic acid (1b), mono-and diesterification procedures were performed in 1 and 10 mmol scales without significant loss of selectivity or reaction yield.
Additional experiments were performed to give insight into the reaction mechanism. The first step in the reaction led to the formation of highly unstable intermediate A which readily transforms into monoester 3 at room temperature. At a higher temperature, a subsequent reaction starts giving diesters 4 via an important pyrophosphonate 6 intermediate which was characterised by 31 P NMR spectroscopy.
It is important to note that for both reactions, no additional reagents are required and crude products obtained after reaction are sufficiently pure for most applications.
Supplementary Materials: The following are available online. Table S1: Summarized data from the 31 P NMR measurements during the course of the esterification at elevated temperature, Table S2: Recorded 31 P NMR shifts for substrate phosphonic acids, Table S3: Recorded 31 P NMR shifts for obtained mono-and diesters, as well as 1 H, 13 C and 31 P NMR spectra for all isolated products.