In situ photo-on-demand phosgenation reactions with chloroform for syntheses of polycarbonates and polyurethanes

Phosgene is an important carbonyl source for industrial production of polycarbonates (PCs) and polyurethanes (PUs). However, since it is highly toxic, alternative compounds and/or new phosgenation reactions have been explored for safety reasons. Given this background, we found a novel photochemical reaction enabling the synthesis of phosgene from chloroform. Subsequently, we developed new phosgenation reactions and reaction systems, and the key objective was “safe application” to organic synthesis. This focus review reports our recent use of in situ photo-on-demand phosgenations of alcohols and amines in synthesizing PC, PU, and their precursors, such as chloroformates, carbonate esters, and diisocyanates, in batch reaction systems, which are preferable for laboratory or small-scale industrial syntheses. We believe that the present reactions have advantages over conventional phosgenation reactions, especially in terms of safety and environmental impacts, and are expected to make positive contributions to practical organic syntheses in both academia and industry. We have developed innovative new phosgenation reactions and their special reaction systems with the key objective of “safe application” to organic synthesis. This focus review summarizes our recent studies on in situ photo-on-demand phosgenation reactions of alcohols and amines for synthesizing polycarbonates, polyurethanes, and their precursors such as chloroformates, carbonate esters, and diisocyanates, in batch reaction systems, which are preferable for laboratory or small-scale industrial syntheses.


Introduction
Phosgene has the chemical formula COCl 2 , is an important C1 building block in organic synthesis and is used as a raw material for the syntheses of polycarbonates (PCs) and polyurethanes (PUs) [1][2][3][4][5][6][7][8][9][10]. However, its laboratory use is often restricted due to its extremely high toxicity [11]. Industrially, it is used in large quantities worldwide with strict legal controls of (1) the storage and transport of phosgene; (2) the use of large quantities of toxic substances (gases) as raw materials; and (3) the disposal of waste, including chlorinated byproducts. The global market for phosgene has grown continuously with the increased production of PCs and diisocyanates such as toluene diisocyanate (TDI) and methylene diphenyl diisocyanate (MDI), which are precursors in the production of PUs [4][5][6][7][8][9][10], along with the production of pharmaceuticals, pesticides, and dyes. Phosgene is generally manufactured on-site and ondemand, where and when it is needed, and only in the quantities needed. Currently, in large-scale industrial processes, it is produced from carbon monoxide (CO) and chlorine (Cl 2 ) gas at 50-150°C with a carbon catalyst [ Fig. 1, reaction (a)] [12,13]. This method of production has remained essentially unchanged since approximately 1920 and is now an established method for phosgene production [2]. However, the method has clear associated safety risks due to the high toxicity of CO and the toxicity/ corrosiveness of Cl 2 , and their intense exothermic reaction requires temperature control. Furthermore, given global climate change and the consequent strong demand and need for carbon neutrality, innovative phosgene production methods have attracted social attention [14].
To reduce the safety risk, in situ production of phosgene via decomposition of triphosgene (BTC), which is a solid at room temperature and soluble in organic solvents, with organic bases [ Fig. 1, reaction (b)] is mainly used for relatively small-scale chemical syntheses in both academia and industry [15][16][17]. However, Cotarca and coworkers recently reported on the risks of BTC, which also has high toxicity and easily reaches toxic concentrations due to its high vapor pressure, and warned against unrestricted use [18]. Furthermore, the use of organic bases in BTC phosgenation reactions produces the hydrochloride salt in solution owing to the presence of HCl generated from the reaction of phosgene and the substrate. This can lead to mechanical system issues, such as difficulty in stirring the batch tanks and clogging the flow systems. Furthermore, the hydrochloride salt generated in the reaction must be removed through an additional purification process. Although alterations to conventional phosgene syntheses and alternative nonphosgene methods have been studied [19,20], they have only marginally replaced the current methods due to their higher costs and higher environmental impacts.
In light of this, we developed a photo-on-demand synthesis of phosgene from chloroform (CHCl 3 ) [ Fig. 1, reaction (c)] [21,22]. The photochemical oxidation of CHCl 3 to COCl 2 occurs efficiently upon irradiation with a 20 W low-pressure mercury lamp (LPML), which emits 184.9 and 253.7 nm UV light, with O 2 bubbling at room temperature. In the early stage of this study, we synthesized a variety of organic chemicals, including polymers, with the gaseous photooxidation products of CHCl 3 by using a gastransport reaction system (Fig. 2, system [I]) [23,24]. However, this reaction system had a potential risk of leakage, such as the case of reaction (a) in Fig. 1. To ensure safe use of this reaction, we then developed "in situ" photo-ondemand phosgenation reactions, which required the design and construction of batch and flow reaction systems [25][26][27][28][29][30][31][32][33][34]. We have successfully applied this photo-ondemand phosgenation reaction on a practical scale for most common phosgenation reactions, such as the syntheses of chloroformates [25,26], carbonate esters [28,29], PCs [28], isocyanates [30], Vilsmeier reagents [27], acyl chlorides [25], and α-amino acid N-carboxyanhydrides (NCAs) [32] (Fig. 3). This review article highlights the reactions of alcohols and amines used in synthesizing PC, PU, and their precursors, and the examples are limited to those using the batch reaction system. These reactions have the potential to replace the conventional reactions and exhibit superior safety and low energy consumption; additionally, chloroform is a common organic solvent and is readily available as a raw material for a wide variety of reactions conducted on various scales.

Mechanism for the photochemical oxidation of CHCl 3
The photochemical oxidation of CHCl 3 may proceed through a radical chain mechanism initiated by photolytic cleavage of a C-Cl bond [path (a) in Fig. 4]. The eliminated Cl • reacts with CHCl 3 to give Cl 3 C • and HCl. The resulting Cl 3 C • initiates a radical chain reaction with O 2 to give COCl 2 with elimination of a Cl • [path (b) in Fig. 4] [22,35]. This reaction mechanism was supported by the observation that the yield of the byproduct hexachloroethane (C 2 Cl 6 ) increased with decreasing oxygen concentrations in the reaction system. It is also known that COCl 2 decomposes under UV irradiation to give CO, CO 2 , and Cl 2 . The generated Cl 2 most likely served as an additional source of Cl • , which accelerated the photochemical oxidation of CHCl 3 . In general, alcohols such as ethanol are used as stabilizers to inhibit the decomposition of CHCl 3 [36] and may serve as radical scavengers to regenerate CHCl 3 from the halomethane radicals. However, when the concentration of the halomethane radicals exceeded the threshold alcohol concentrations in both the liquid and vapor phases, the photooxidation of CHCl 3 supported phosgenation of the alcohol. This enabled the development of the in situ photo-on-demand phosgenation reactions described in this review article.

Experimental setup of a batch-type photoreaction system
Low-pressure mercury lamps generally exhibit low electric power consumption and generate UV light with wavelengths of 184.9 and 253.7 nm, which match the electronic absorption bands of CHCl 3 arising from σ-σ* and/or n-σ* transitions [37]. The lamp (20 W, ø24 mm × 120 mm) exhibited a 254 nm illuminance of 6.2-9.0 mW/cm 2 at 5 mm from the lamp and was inserted into the CHCl 3 solution in a quartz glass jacket (ø28 mm × 150 mm) fixed in the center of a cylindrical flask (ø42 mm × 170 mm) equipped with an alcohol/water-cooled condenser (Fig. 2, system [II]). The photochemical reactions were conducted in this reaction system while a steady flow of O 2 (0.1-0.5 L/ min) was bubbled through CHCl 3 with or without the substrate for the phosgenation reaction and with stirring of the sample solution at various temperatures. The reactions were demonstrated with a closed system, but the exhausted gas containing unreacted COCl 2 and the generated HCl Base-free in situ photo-on-demand synthesis of chloroformate and one-pot syntheses of carbonate esters and carbamates Chloroformate plays important roles in organic synthesis and is generally prepared from COCl 2 and an alcohol ( Fig. 5) [38]. In our pioneering study of in situ photo-ondemand phosgenation reactions, we reported that chloroformate was produced in high yield upon photoirradiation of a CHCl 3 solution containing an alcohol with O 2 bubbling [25,26]. This approach provided one-pot syntheses of unsymmetric carbonates and carbamates via subsequent addition of alcohols or amines, respectively. When CHCl 3 solutions containing 20 mmol of a primary alkyl alcohol (n = 0, 1, 2, 3, or 6) in the above photochemical reaction system were exposed to UV light at  In situ photo-on-demand phosgenation reactions with CHCl 3 reported by the author's group 30°C, the corresponding chloroformates were obtained as the major products in 7-93% yields. (Fig. 5a). Formates and carbonate esters were also obtained as minor products with 5% yields. The chloroformate yield clearly decreased with decreasing alcohol chain length because the alkyl alcohols with shorter chain lengths evaporated more easily, which slowed the photochemical conversion of CHCl 3 to COCl 2 in the gas phase. In support of this proposed mechanism, the reaction was decelerated dramatically when the reaction temperature was raised to 50°C, which may have vaporized both the alcohol and COCl 2 in the CHCl 3 . Longer irradiation times also tended to decrease the product yields, most likely due to photodecomposition of the products. With the concentration of alcohol, the temperature, and the irradiation time optimized, triethylene glycol monomethyl ether (TEGM) was converted to the corresponding chloroformate in 89% yield. Although aryl alcohols, which are generally less nucleophilic than alkyl alcohols, showed no notable reaction, aryl-substituted aliphatic alcohols such as benzyl alcohol and 9-fluorenylmethanol provided the corresponding chloroformates.
This reaction enables the preparation of CHCl 3 solutions containing chloroformates, and thus, these solutions are available for one-pot syntheses of unsymmetric carbonate esters and carbamates (Figs. 5b and 5c). An as-prepared CHCl 3 solution of n-hexyl chloroformate was stirred at 30-70°C for 1-3 h to remove the HCl, and the unreacted COCl 2 dissolved in the sample solution. Since the carbonyl carbon in chloroformate is less electrophilicity than that of COCl 2 , the second substitution reaction with an alkyl alcohol to form the carbonate ester occurs slowly when the sample solution is heated. For example, when 1.5 equiv. of 1-hexanol was added into a CHCl 3 solution containing chloroformate, and the sample solution was stirred for 15 h at 90°C to evaporate the CHCl 3 solvent from the system, the corresponding symmetric carbonate was obtained in 95% yield (isolated amount and yield: 1.29 g and 56%, respectively). With a similar procedure, benzyl alcohol and TEGM provided unsymmetrical carbonates in 87% and 50% yields, respectively. 2-Propanol, a secondary alcohol, did not react without a base, but the corresponding unsymmetrical carbonate was obtained in 60% yield upon addition of pyridine. (-)-Menthol, with a boiling point higher than that of 2-propanol, underwent the reaction without a base and at a higher temperature to give the corresponding carbonate in 54% yield. Although phenol, an In situ photo-on-demand synthesis of (a) chloroformates and one-pot syntheses of (b) symmetric and unsymmetric carbonates and (c) carbamates aryl alcohol which is less nucleophilic than alkyl alcohols, did not react with chloroformate even at elevated temperatures, it underwent the reaction after triethyl amine (TEA) was added and provided the asymmetric hexyl phenyl carbonate in 56% yield. With similar one-pot procedures, biobased nonionic amphiphiles were synthesized from CHCl 3 solutions of TEGM. The TEGM underwent an initial photochemical conversion in CHCl 3 to form the corresponding chloroformate, and subsequent addition of citronellol or geraniol to the sample solution and heating provided the corresponding amphiphilic carbonates in 69% and 70% yields, respectively.
The introduction of N-protecting groups to amines is an important application of chloroformates [39]. One-pot syntheses of carbamates were also achieved upon addition of an amine to the prepared CHCl 3 solution of chloroformate (Fig. 5c). For example, the photochemical conversion of 1-hexanol to the corresponding chloroformate in CHCl 3 and a subsequent reaction with aniline under reflux generated the carbamate in 70% yield based on the alcohol (isolated amount and yield: 1.13 g and 48%, respectively) with the elimination of HCl. With a similar procedure, an amphiphilic carbamate was also synthesized from TEGM in 56% yield. Cbz-and Fmoc-protection of cyclohexylamine were then achieved with this one-pot procedure.
Base-promoted in situ photo-on-demand syntheses of carbonate esters and polycarbonates COCl 2 and chloroformates undergo faster condensation reactions with alcohols in the presence of a base, which serves as a catalyst as well as an HCl scavenger. Since organic bases such as TEA and pyridine absorb UV light [40,41], difficulties were anticipated for in situ photo-ondemand syntheses of carbonate esters with CHCl 3 solutions containing a mixture of alcohol and organic base, both of which inhibit the photooxidation of CHCl 3 . In fact, the amount of COCl 2 generated by photooxidation of CHCl 3 (30 mL, 0.37 mol) decreased to 46% when 0.03 mol (8%) of pyridine was added. However, the reactions occurred with aryl alcohols and haloalkyl alcohols to give the corresponding carbonate esters (Fig. 6). This reaction provides convenient in situ photo-on-demand syntheses of carbonate esters in high yields on a gram scale [28].
When the photoreactions were conducted by adding 3.5 equiv. of pyridine to CHCl 3 solutions containing ethanol (EtOH) or 2-propanol (2-PrOH), neither the carbonate ester nor the chloroformate was obtained (data not shown). Both pyridine and alcohol inhibit the oxidative photodecomposition of CHCl 3 by absorbing UV light and/or trapping the radical species generated from CHCl 3 .
Although phenol, whose nucleophilicity is lower than those of alkyl alcohols, also showed no reaction in the absence of a base, the phosgenation reaction occurred to give diphenyl carbonate (DPC) in >99% yield (isolated amount: 1.1 g) in the presence of 5 equiv. of pyridine. Phenol is relatively more acidic (pK a = 10.0) than alkyl alcohols (pK a = 15-18) [42], and pyridine forms a weak acid-base complex with phenol with an association constant of K = 33 M -1 in CHCl 3 solution at 20°C. This interaction may decelerate radical trapping and accelerate the nucleophilic additions of phenol to COCl 2 and chloroformate. 2-Naphthol, which has a larger π-conjugated structure, and 4-methoxyphenol, which contains an electron-donating methoxy group, have lower and higher pK a values, respectively, (9.51 and 10.05, respectively) than phenol and provided the corresponding carbonate esters in >99% yields [43,44]. 4-Fluorophenol, with an electron-withdrawing F on the phenol ring, provided the corresponding carbonate esters in 93% yield. Furthermore, 4-nitrophenol and 4-cyanophenol contain electronwithdrawing substituents and were converted to the Fig. 6 In situ photo-on-demand syntheses of carbonate esters with CHCl 3 solutions containing alcohols and pyridine corresponding carbonate esters in 39% and 62% yields, respectively. Even with pentachlorophenol and pentafluorophenol, which have lower pK a values of 4.96 and 5.53 [45,46], respectively, the reactions occurred with high yields.
These base-promoted in situ photo-on-demand syntheses of carbonate esters were also available for the syntheses of PCs from diols (Fig. 7). Bisphenol A (BPA) provided a quantitative yield for a conventional PC (isolated amount: 2.7 g) with an average molecular weight of M w = 52,000, M n = 23,000, and PDI = 2.26. Bisphenol AF (BPAF), which contains electron-withdrawing trifluoromethyl groups, provided a PC with M w = 14,000, M n = 6,400, and PDI = 2.19. The lower molecular weight may be ascribed to the lower nucleophilicity of BPAF relative to that of BPA. 2,2,3,3,4,4-Hexafluoro-1,5-pentanediol, a fluoroalkyl diol, afforded the corresponding PC in 59% yield with M w = 41,300, M n = 24,400, and PDI = 1.69. The low solubility of the fluoroalkyl PC in organic solvents reduced the isolated yield.
By preparing the diisocyanates in CHCl 3 solution, onepot syntheses of biscarbamates, which serve as blocked isocyanates, were achieved (Fig. 9). When EtOH or HFIP was added to a CHCl 3 solution containing MDI, which was prepared with the two-step procedures described above without and with added pyridine, the corresponding biscarbamates were obtained in 78 and 71% yield, respectively. HDI, which is less reactive than MDI, also reacted with HFIP to give the corresponding biscarbamate with a lower yield of 37%. Fluorinated HDI (8FHDI), which is more reactive and unstable in air, was then prepared from a CHCl 3 solution containing the HCl salt of 2,2,3,3,4,4,5,5octafluorohexane-1,6-diamine (8FHDA·2HCl). The one-pot synthesis proceeded after addition of HFIP to the sample solution and afforded the corresponding biscarbamate in 37% yield.
One-pot syntheses of polyurethanes (PU) were achieved by adding diols instead of monoalcohols to the diisocyanate CHCl 3 solutions. When an equimolar amounts of 1,6-hexanediol (HD) relative to the diamines were added to CHCl 3 solutions containing MDI or 8FHDI, the corresponding PUs [poly(MDI-HD) or poly(8FHDI-HD)] were obtained in 50% and 91% yields (isolated amounts: 0.92 and 0.81 g, respectively) with average molecular weights of M w = 4100, M n = 2200, PDI = 1.86 and M w = 4000, M n = 2400, PDI = 1.67, respectively. Their relatively low average molecular weights may have originated from the poor solubility of the formed PUs. Then, polypropylene glycol (PPG) with an average molecular weight of 400 increased the solubility of the PU and provided poly(MDI-PPG) and poly(8FHDI-PPG) in 89 and 86% yields (isolated amounts: Fig. 8 Two-step procedure used to synthesize diisocyanates with the in situ photo-on-demand synthesis of COCl 2 from CHCl 3 and subsequent addition of diamines a Fig. 9 One-pot syntheses of biscarbamates and polyurethanes from diamines through photochemical conversions to the corresponding diisocyanates 2.91 and 1.18 g, respectively) with M w = 6200, M n = 3600, PDI = 1.72, and M w = 8800, M n = 3400, PDI = 2.59, respectively. The average molecular weight of the PU formed in this one-pot synthesis can be controlled by estimating the amount of the diisocyanate formed in the sample solution via spectroscopic and/or HPLC analyses. Cl 2 -promoted photochemical oxidation of CHCl 3 with visible light and application to one-pot organic syntheses Compared with the conventional phosgenation reactions run with CO/Cl 2 or triphosgene, the photo-on-demand phosgenation reactions described above for CHCl 3 are safe, convenient, and inexpensive, but the use of the LPML causes unfavorable side reactions arising from photodecomposition of both reagents and products by the high-energy UV-C light. The use of mercury lamps has recently been avoided due to their large environmental impacts and associated health hazards [47]. This background motivated us to develop a novel photo-on-demand photocatalytic reaction using lower-energy visible light, which enables the use of light-emitting diodes (LEDs) and sunlight as light sources instead of LPMLs. We found that bubbling O 2 through a CHCl 3 solution containing~2% Cl 2 caused photooxidation with the white LED light [31]. For example, when 30 mL of CHCl 3 subjected to O 2 /Cl 2 bubbling was exposed to white light from a 9 W LED at 20°C for 1.5 h, a 12.5% conversion to COCl 2 resulted. The added Cl 2 may have served as a visible light-responsive initiator for the radical chain reaction of CHCl 3 and O 2 (Fig. 4) [48].
This photochemical reaction using visible light provided one-pot syntheses of chloroformates, carbonate esters, and isocyanates (Fig. 10). However, it is important to note that the reaction should be performed with a two-step procedure since some of the alkyl alcohols and amines used as reactants react with Cl 2 . Using this method, hexyl chloroformate was synthesized quantitatively from a CHCl 3 solution (50 mL) containing 1-hexanol (30 mmol) without the need for an organic base. Dibutyl carbonate (DBC), DPC, and BHFC were also synthesized in 95, 84, and 70% yields via base-catalyzed condensation reactions. Furthermore, 1-isocyanatohexane, isocyanatobenzene, and (3-isocyanatopropyl)trimethoxysilane were synthesized in >99, 73, and 56% yields, respectively.

Conclusion
We have focused this review on our recent studies of in situ photo-on-demand phosgenation reactions of alcohols and amines, which were used to synthesize PCs, PUs, and their precursors such as chloroformates, carbonate esters, and diisocyanates in batch reactions. Given the current global warming problem, sustainable synthetic methods for polymer production are being actively studied. This photochemical reaction efficiently converts CHCl 3 to COCl 2 , which is extremely reactive toward nucleophiles, and this enables in situ syntheses of many organic chemicals and polymers while reducing energy consumption and waste generation. We believe that the present in situ photo-ondemand phosgenation reaction has advantages over conventional phosgenation reactions, especially in terms of safety and environmental impact. When combined with the flow photo-on-demand phosgenation system reported recently by our group [34], which enables scale-up of chemical processes, our present findings are expected to provide practical polymer syntheses of use in both academia and industry. Our group is currently developing suitable facilities for the safe use of this photo-on-demand phosgenation reaction on a larger scale and is constructing a reaction library describing the syntheses of various organic chemicals, including polymers, for eventual commercial use. R&D (A-STEP), Seeds development type from Japan Science and Technology Agency (JST). The author is highly grateful to the collaborators on the experiments of photo-on-demand organic syntheses conducted at Kobe University: Prof. Kazuo Eda, Dr. Fengying Liang, Namin Dai, Masaki Yanai, Ryo Muranaka, Yue Liu, Yuka Hashimoto, Yuto Suzuki, Itsuumi Okada, and Yuki Kuwahara.
Funding Open access funding provided by Kobe University.

Compliance with ethical standards
Conflict of interest The author declares no competing interests.
Publisher's note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons. org/licenses/by/4.0/.