2. Synthetic Approaches to Polyphosphodiesters: An Overview
In a recent review,
[5] Iwasaki presented several important examples of the synthetic approaches to PCPAs. In this section, researchers have tried to enhance, refine and discuss alternative synthetic approaches to polyphosphodiesters. The synthesis of the most simple polyphosphodiesters, poly(ethylene phosphoric acid) (PEPA) and poly(1,3-propylene phosphoric acid) (1,3-PPPA), was reported by Penczek’ group back in 1976.
[21] To date, multiple approaches to polyphosphodiesters have been developed. The most evident synthetic pathway is based on the interaction of phosphoric acid with diols reviewed by Penczek et al. in 2015
[1] or on transesterification of dialkyl (or diaryl) phosphonates followed by oxidation of P–H bonds.
[22][23][24][25][26][27][28] Ring-opening polymerization (ROP) of strained cyclic phosphonates (containing P–H bonds) and phosphates, followed by post-modification (oxidation or hydrolysis/hydrolytic thermolysis, respectively) is another efficient pathway to polyphosphodiesters.
[29][30] Meanwhile, modern methods of the construction of hydrocarbon fragments of the PCPA backbone, i.e., metathesis polycondensation and polymerization,
[31][32][33][34][35] should not be dismissed (
Scheme 2). Note that the use of acyclic diene metathesis (ADMET) polycondensation in the synthesis of ‘precision polymers’ was the subject of review by Schulz and Wagener.
[36]
Scheme 2. General synthetic approaches to polyphosphodiesters.
3. Polycondensation and Related Methods
3.1. Reactions of H3PO4 with Diols and Polyols
Phosphoric acid H
3PO
4 is a relatively weak tribasic acid (pK
a1 = 2.15, pK
a2 = 7.09, pK
a3 = 12.32). With the transition to pyrophosphoric acid H
4P
2O
7, one can note a substantial increase of acidity (pK
a1 = 1.0, pK
a2 = 2.0) and, therefore, reactivity of H
4P
2O
7 in comparison with H
3PO
4. Poly(phosphoric acid) is a well-known ‘superacid’; however, its use in the synthesis of PCPAs is essentially restricted by the requirements of the hydrolytic stability of PCPAs that implies the absence of di-/oligophosphate fragments in the main polymer chain. In this way, successful synthesis of PPDEs was limited by the use of H
3PO
4 and H
4P
2O
7 in polycondensation with diols and polyols. This approach was developed mainly by Penczek and coll who studied direct condensation of H
3PO
4 with ethylene glycol.
[37][38][39] The following steps were detected during this reaction:
-
The reaction starts by the relatively slow dimerization of H3PO4 with a formation of H4P2O7 (and higher polyphosphoric acids) at 100 °C within 40 h, during this stage the water was removed either in the stream of neutral gas or azeotropically with heptane.
-
After the addition of EG at 100 °C, H4P2O7 transformed to H3PO4 immediately, and the first phosphorylation reaction within additional 80 h was the formation of HOCH2CH2OP(O)(OH)2 and (HOCH2CH2O)2P(O)OH, triesters were formed in minimal amounts.
-
Activation of the monophosphate esters (end groups) at any polymerization degree with H3PO4 proceeds via conversion of monoesters into pyrophosphoric acid esters –OCH2CH2OP(O)(OH)–OP(O)(OH)2 that represent reactive acidic sites.
-
The polycondensation product is mostly linear with a structure of PEPA –(OCH2CH2OP(O)(OH))n–.
-
Some branch points (triesters) are formed only at high temperature and prolonged polycondensation time.
The reaction resulted in the formation of relatively low molecular weight (MW) products, the maximum achieved degree of polymerization (
DPn) was 21 after 100 h at 150 °C even in the presence of Sc(CF
3SO
3)
3 as a catalyst. Polycondensation was also accompanied by the formation of ether bonds (di- and triethylene glycol fragments were detected), acetaldehyde and vinyl end-fragments.
[37][38] To avoid dehydration side processes during the reaction with H
3PO
4, Penczek et al. proposed the use of 2,2-dimethyl-1,3-propanediol; however, no polymers were obtained, and the main reaction product was 2-methylbutanal formed via methyl migration (
Scheme 3).
[37]
Scheme 3. Formation of 2-methylbutanal during the reaction of neopentyl glycol with H
3PO
4.
[37]
The reaction of H
3PO
4 with glycerol is a more complex process.
[39][40][41] This reaction was conducted at 100 °C with azeotropic water removal (heptane) or under reduced pressure. The rate of esterification and the product ratios depended on the reagent ratios. So, for example, for a H
3PO
4/glycerol ratio of 1:1 the conversion of H
3PO
4 reached 90% after 35 h, whereas at a H
3PO
4/glycerol ratio of 1:2 even after 140 h only 80% conversion was detected, and the ratio of 2:1 led to monoester as a main product. Five- and six-membered cyclic esters were detected in the reaction mixtures in minor amounts. At a 1:1 H
3PO
4/glycerol ratio, cross-linking was observed. The degree of polymerization of soluble products was limited by dealkylation, leading to the formation of di- and oligo-glycerol units, incorporated into the product structure. Polycondensation of diglycerol (HOCH
2CH(OH)CH
2)
2O) with H
3PO
4 resulted in the formation of highly branched gels.
[41] The prospects of the further use of these polymers still remains unclear due to the unpredictability of their microstructure and hydrolytic behavior.
In conclusion, it should be mentioned that the reaction of H
3PO
4 with ethylene carbonate, first described by Munoz et al.
[42] and reproduced by Imoto and coll,
[43] resulted in low-MW PEPA with an unknown structure. Additionally, note that the reaction of H
3PO
4 with oxirans results in a formation of triester species
[44][45] and therefore cannot be considered as a method of the synthesis of PPDEs.
3.2. The Reaction of Dichlorophosphates with Diols
Glycolysis of PET with a formation of bis(2-hydroxyethyl)phthalate is the most efficient method of chemical recycling of this polymer.
[46][47] The reaction of bis(2-hydroxyethyl)phthalate with Cl
2P(O)OR (R = Me, Et) resulted in the formation of copolymers, further treatment by terephthaloyl chloride and NaI/acetone allowed for a copolymer containing >P(O)–OH fragments to be obtained (
Scheme 4).
[48] However, the current trends in developing actual synthetic approaches to biodegradable materials imply the abandonment of chlorine-containing reagents, and therefore dichlorophosphates are not currently used in the synthesis of polyphosphodiesters.
Scheme 4. Synthesis of the phosphate-containing analog of PET.
[48]
3.3. Reaction of Dialkyl (or Diaryl) Phosphonates with Diols and Post-Modification
Since polymers with –O–P(O)H–O– fragments can be easily and almost quantitatively oxidized to corresponding poly(phosphodiesters) containing –O–P(O)(OH)–O– fragments,
[21][22][49] polycondensation of dialkyl phosphonates (RO)
2P(O)H with diols can be considered as a prospective method of the synthesis of polyphosphodiesters. However, when using propane-1,3-diol, a six-membered cyclic phosphonate is formed at elevated temperatures, and further low-temperature ROP is needed for the synthesis of PPDE.
[50] In addition, Penczek and coll have proposed that for the successful synthesis of high-MW polymer the alcohol ROH has to be removed as fast as possible.
[51]
Relatively high-MW poly(alkylene phosphonates) (
Mn = 9.3–28 kDa) were obtained by the reaction of (MeO)
2P(O)H with HO–(CH
2)
n–OH (
n = 5–10, 12).
[22] Polytransesterification of dimethyl phosphonate (MeO)
2P(O)H and poly(ethylene glycol)s with
Mn 200 Da (PEG200) and 600 Da (PEG600) resulted in copolymers with
Mn = 3.5 and 7.1 kDa, respectively;
[23][24] similar results were obtained using PEG400, transesterification was conducted within 5 h at 135 °C under atmospheric pressure, and then 4 h at 160 °C plus an additional 15 min at 185 °C in vacuo (1 Torr), degree of polymerization (
DPn) was 28.
[25] The reaction of H(OCH
2CH
2)
13O H with (MeO)
2P(O)H also resulted in the formation of the polymer (
Mn = 13.5 kDa).
[26] Poly(1,2-propylene glycol) (PPG)-based oligo(alkylene phosphonate)s with
DPn 12, 6 and 5 were synthesized with the use of PPG400, PPG1200 and PPG2000, respectively.
[27] Triblock copolymers mPEG750-
b-[(P(O)(H)O(CH
2)
6]
17-
b-mPEG750 and mPEG2000-
b-[(P(O)(H)O(CH
2)
6]
17-
b-mPEG2000 were obtained by polycondensation of (MeO)
2P(O)H with HO–(CH
2)
6–OH (4 h at 80 °C and then 9 h at 140 °C/1 Torr, 0.05 mol% Na to form the catalyst), followed by the reaction with mPEG (140 °C/1 Torr).
[52] To achieve high molecular weights of the polycondensation products, Penczek and coll proposed the use of diphenyl phosphonate in reaction with diols.
[53] The reaction was conducted at 140 °C with the elimination of the phenol, and PPDEs with
Mn up to 40 kDa were obtained (
Scheme 5).
Scheme 5. Polycondensation of diphenyl phosphonate with diols.
[53]
To obtain PPDEs, PEG200- and PEG1000-based poly(alkylene phosphonate)s were oxidized by N
2O
4 in CH
2Cl
2.
[23] The same reagent was also used for the oxidation of block copolymers mPEG750-
b-[(P(O)(H)O(CH
2)
6]
17-
b-mPEG750 and mPEG2000-
b-[(P(O)(H)O(CH
2)
6]
17-
b-mPEG2000 in CH
2Cl
2 at −10 °C
[52] and poly(1,2-propylene glycol)-based poly(alkylene phosphonate)s.
[27]
Chlorination of poly(alkylene phosphonate)s at 0 °C resulted in the formation of poly(alkylene chlorophosphate)s that can be easily hydrolyzed with a formation of PPDEs
[53] (
Scheme 6a) or transformed into alkoxy-
[22] and amino-derivatives
[54] (
Scheme 6b). The degree of chlorination of poly(alkylene phosphonate)s can be varied when using trichloroisocyanuric acid as the chlorination reagent; the quantitative yield of the corresponding PPDE was confirmed by NMR monitoring of the hydrolysis of MeO[P(O)(Cl)O(CH
2CH
2O)
9]
28H in MeCN (full conversion after 15 min at 20 °C).
[25]
Scheme 6. Chlorination of poly(alkylene phosphonate)s followed by the: (
a) hydrolysis;
[53] or (
b) reaction with alcohols
[22] and amino acid esters.
[54]
Penczek and coll.
[55][56] have shown that the direction and selectivity of the hydrolysis of poly(alkylene amidophosphate)s depend on the pH value and the structure of the substituents in a nitrogen atom. When studying model amidophosphates, preferential cleavage of the P–O bond was detected at alkaline conditions, whereas at acidic conditions (MeO)
2P(O)OH was the main reaction product (
Scheme 7a). Poly(1,3-propylene amidophosphate)s demonstrated similar chemical behavior (
Scheme 7b) except for an O-ethyl-GlyGly derivative that formed 1,3-PPPA in both acidic and alkaline conditions. At pH~8 and 37 °C the P–NH bond was hydrolyzed 3–4 times faster than the P–O bond in the main chain.
[55]
Scheme 7. (
a) Hydrolysis of model amidophosphates; (
b) Acidic hydrolysis of poly(1,3-propylene amidophosphate)s.
[56]
Another method of the transformation of poly(alkylene phosphonate)s to poly(alkylene phosphate)s uses the Atherton–Todd reaction.
[24] In particular, this reaction was used in the synthesis of PPDEs containing (OCH
2CH
2)
13 spacers between phosphate groups.
[26] In conclusion of this section, one should refer to the successful synthesis of the polymers containing –OP(O)(H)O–(CH
2)
x– units (x = 10, 17, 21, 46) with
Mn 11–25 kDa by the reaction of the corresponding diols with dimethyl phosphonates.
[57] These polymers were not transformed to PPDEs, there was only one step to polyethylene mimicking polymers containing phosphate fragments in the main chain (note that similar polymers were nevertheless obtained by Wurm and coll. with the use of the ADMET approach, see Section 2.4.).
3.4. Polycondensation of (ω-Hydroxyalkyl)phosphonic Acids
In 2020,
[58] Penczek and coll. have shown that hydroxymethyl phosphonic acid can act as a catalyst and initiator of the ROP of ε-caprolactone (εCL) with the formation of εCL oligomers containing reactive groups on both ends of the macromolecule. Very recently they demonstrated that these oligomers can be subjected to polycondensation at 100–110 °C with a formation of PPDEs (
Mn up to 25 kDa) (
Scheme 8) with mostly linear microstructure (
31P NMR data).
[59]
Scheme 8. Synthesis and polycondensation of (HO)P(O)CH
2O(εCL)
nH.
[59]
4. ROP of Cyclic Phosphorus-Containing Monomers and Post-Modification
4.1. Synthesis of Cyclic Phosphorus-Containing Monomers
The key stage of the preparation of both cyclic phosphonates and cyclic phosphates is a reaction of diols with PCl
3 resulting in cyclic chlorophosphites
[60] that can be hydrolyzed with the formation of cyclic phosphonates (
Scheme 9a) or oxidized to chlorophosphates with subsequent substitution of Cl atom by alkoxy fragment that results in cyclic phosphates (
Scheme 9b). In some cases, the synthesis of cyclic phosphates is based on reverse reaction sequence, i.e., substitution of Cl in chlorophosphite followed by oxidation (
Scheme 9c).
[61] Cyclic phosphonates can also be synthesized by the reaction of diols with dialkyl phosphonates (
Scheme 9d).
[62][63]
Scheme 9. Common synthetic approaches to cyclic phosphorus-containing monomers for ROP.
Hydrolysis of chlorophosphite was carried out in CH
2Cl
2 solution with a mixture of water and 1,2-dioxane (
Scheme 10). It was essential to use slightly less than the stoichiometric amount of water (0.8 equiv.), otherwise undesirable polymerization occurred.
[64]
Scheme 10. Synthesis of 4-methyl-2-oxo-2-hydro-1,3,2-dioxaphosphol.
[64]
The first systematic studies of the synthesis of five-membered cyclic phosphates (2-alkoxy-2-oxo-1,3,2-dioxaphospholanes,
Scheme 11), based on the reaction of cyclic chlorophosphates with ROH, were conducted by Penczek et al. back in the late 1970s.
[65][66] The synthesis of 2-chloro-2-oxo-1,3,2-dioxaphospholane was optimized recently by Becker and Wurm.
[67] 2-Chloro-1,3,2-dioxaphospholane was obtained with 67% isolated yield, and subsequent CoCl
2-catalyzed oxidation by dried air resulted in the obtaining of cyclic chlorophosphate that was separated by vacuum distillation, the yield was 70%. Additionally, note that the efficient continuous flow method of the end-to-end preparation of cyclic phosphate monomers with a semi-continuous modular flow platform was developed very recently by Monbaliu and coll.
[68]
Scheme 11. Synthesis of five-membered cyclophosphates, the yields on the last sage are given.
[66]
2-Methoxy-2-oxo-1,3,2-dioxaphospholane (methyl ethylene phosphate, MeOEP) contained, after distillation, an impurity of (MeO)
2P(O)OCH
2CH
2Cl, and the final purification involved treatment with an Na mirror. The reaction of cyclic chlorophosphates with alcohols has limitations on the substrate. Primary and secondary alcohols usually give satisfactory yields of cyclic phosphates,
[61][66] while
tert-butanol does not react in the right way due to the low reactivity of
tert-butanol at ambient conditions and low thermal stability of
tBuOEP.
The choice of the base is essential in the synthesis of cyclic phosphates by the reaction of chlorophosphates with alcohols. The presence of the traces of the ammonium salts complicates the separation of cyclic phosphates because of their acid-catalyzed polymerization. The use of lutidine was proposed in the first work devoted to the synthesis of ethylene phosphates,
[66] and it was this base that was used in the synthesis of unstable 2-benzyloxy-2-oxo-1,3,2-dioxaphospholane (benzyl ethylene phosphate, BnOEP).
[69]
Because of the low thermal stability of
tBuOEP and other
tert-butyl alkylene phosphates, alternative approaches to these valued monomers were developed. Nakamura et al. have used oxidation of cyclic phosphites by N
2O
4 (
Scheme 12a),
[70] and recently Nifant’ev et al. proposed a two-stage approach based on reaction of 2-chloro-1,3,2-dioxaphospholane with
tert-butanol followed by oxidation of 2-
tert-butyl-1,3,2-dioxaphospholane by 3-chloroperbenzoic acid (
mCPBA) (
Scheme 12b).
[61]
Scheme 12. Synthesis of
tert-butyl alkylene phosphates. (
a) oxidation by N
2O
4;
[70] (
b) oxidation by
mPCBA.
[61]
The synthesis of deoxyribose-based five-membered cyclic phosphonate stands somewhat apart from most other 1,3,2-dioxaphospholane derivatives, this compound was obtained by the reaction of methyl-2-deoxyribofuranose with P(NEt
2)
3.
[71]
4.2. ROP of Cyclic Phosphorus-Containing Monomers
ROP of cyclic phosphonates and phosphates (
Scheme 13a) represents the common strategy of the controlled synthesis of functional biodegradable polymers.
[29][30] This process is subject to the general thermodynamic rules for the ROP of cyclic monomers
[72] that predict high reactivity of more strained five-membered cycles
[66][73][74][75] and temperature-dependent reactivity of six-membered cycles.
[74][75] Different catalysts have been used successfully in controlled ROP of cyclic phosphonates and phosphates with the formation of polyphosphoesters (PPEs) (
Scheme 13b). The data on the synthesis of polymers suitable for post-modification to polyphosphodiesters are summarized in
Table 1.
Scheme 13. (a) ROP of cyclic phosphonates and phosphates; (b) Catalysts used in synthesis of polymers suitable for post-modification to polyphosphodiesters.
Table 1. Synthesis of polyphosphonates and polyphosphates suitable for post-modification with a formation of PCPAs. The structures of the catalysts are presented in Scheme 13b.