Photocatalytic N-Formylation of CO₂ with Amines Catalyzed by Diethyltriamine Pentaacetic Acid

Abstract

In the present work, inexpensive and commercially available diethyltriamine pentaacetic acid (DTPA) was used as an initiator to catalyze the N-formylation reaction of CO₂ with amines via the construction of C-N bonds in the presence of xanthone as the photosensitizer and PhSiH₃ as the reducing agent. After a systematic study of various factors, the optimal conditions for the photocatalytic reaction were obtained: 2.5 mmol of amine, 2.5 mmol of PhSiH₃, 10 mol% of DTPA, 20 mol% of xanthone, 1 mL of dimethylsulfoxide (DMSO), atmospheric pressure, and 35 W UV lamp irradiation for 48 h. Under the optimal conditions, the catalyst system afforded high performance for the N-formylation of amines (primary and secondary amines) and CO₂, and the yields of the N-formylated products of dialkylamines were above 70%. Further studies exhibit that the catalytic system has a wide scope of substrate applications. For various alicyclic secondary amines, heterocyclic secondary amines, aliphatic primary amines, and aromatic primary amines, the corresponding N-formylation products can be obtained efficiently. In addition, the catalyst can be recycled by simple precipitation and filtration. After five cycles of recycling, there was no significant change in the catalytic and structural properties of DTPA. Finally, a possible reaction mechanism is proposed.

1. Introduction

With the development of industrial societies, the amount of carbon dioxide in the atmosphere has increased year over year. Carbon dioxide, which accounts for about 82% of all greenhouse gases, is thought to be the main cause of global warming. However, climate change due to global warming could be one of the major threats to modern civilization. Therefore, the reduction of CO₂ emissions and the utilization of CO₂ has become an urgent issue [1].

Utilization of CO₂, or, better to say, “chemical fixation of CO₂”, can be obtained via a simple reaction between CO₂ and various compounds, such as epoxides, amines, and so forth. The N-formylation reaction of amines with CO₂ is one of the attractive routes to high value-added formamide chemicals. Formamide is widely used in the synthesis of chemicals. It is not only a key intermediate in the synthesis of isocyanates [2,3], drugs [4,5,6], and pesticides [7] but also a solvent for organic synthesis reactions such as N,N-dimethylformamide or N,N-dimethylacetamide [8,9,10]. Traditionally, the synthesis of formamide generally requires the use of CO, which has a high degree of toxicity, as a carbonylation reagent to react with the amine. In contrast, the use of CO₂ with amines to generate formamide or methylamine via N-formylation and N-methylation reactions is a green, safe, and more attractive production strategy. However, the schemes reported so far use either homogeneous catalysts [11,12,13], heterogeneous catalysis [14,15], or higher reaction temperatures [16,17]. Therefore, N-formylation of amines and CO₂ using visible light under mild reaction conditions is a promising but less explored approach. However, although many catalyst systems have been adopted successfully to catalyze the N-formylation between CO₂ and amines under traditional chemothermal conditions, few successful works under photocatalyzed conditions have been reported [18,19].

In 2017, Han’s group used glycine betaine, the carboxylate ammonium salt, as a catalyst for the reduction reaction of CO₂ and amine to generate formamide, amine acetal, and methylamine [20]. In 2018, Yu’s team used Ir(ppy)₂(dtbbpy) PF₆ as a photosensitizer and benzyl quaternary ammonium salt as a substrate to realize the coupling of CO₂ with benzyl to form the corresponding benzyl formate without needing external addition of a single-electron donor [21]. This work inspired our interest in the area of photocatalyzed N-formylation of amines and CO₂. In most reported photocatalysis, expensive metal compounds and photosensitizers were used. So, in the present work, cheap and commercially available diethyltriamine pentaacetic acid (DTPA, a good electron-donating reagent with 3 N) was used as a non-metal organic photocatalyst to catalyze the N-formylation reaction of CO₂ with amines in the presence of relative cheap xanthone as the photosensitizer.

2. Materials and Methods

2.1. Reagents and Instruments

The following reagents are purchased from Aladdin Reagents Ltd. and used without further purification: diethylenetriaminepentaacetic acid (99% purity), xanthone (99% purity), dimethyl sulfoxide (99.7% purity), 2-ethylbutan-1-amine (99% purity), di-n-butylamine (99% purity), dihexylamine (99% purity), di-isopropylamine (99.5% purity), pyrrolidine (99% purity), piperidine (99% purity), 2-methylpiperidine (99% purity), polymer-bound (99% purity), thiomorpholine (98% purity), aniline (99.5% purity), 4-fluoroaniline (99% purity), 4-chloroaniline (99% purity), 4-bromoaniline (99% purity), 4-nitroaniline (99% purity), p-toluidine (99% purity), 4-ethylaniline (99% purity), 4-butylaniline (98% purity), 4-isopropylaniline (98% purity), 2-ethylaniline (98% purity), 2,6-dimethylbenzenamine (98% purity), 2,6-diethylaniline (98% purity), 2,6-diisopropylaniline (90% purity), and CO₂ (99.99% purity) with all reagents are purchased from Aladdin Reagents Ltd. and used without further purification. Reaction conversions and selectivity were determined with a Bruker Al-400 MHz NMR instrument (Bruker Technologies Ltd., Migdal HaEmek, Israel) and a LED UV lamp (Zhonglian UV Optical Factory) with a wavelength of 365 nm.

2.2. Preparation of Formamide

Weighed reagents and magnets were added to a 50 mL dry Schlenk flask, and carbon dioxide was used to displace the air in the reaction flask at least three times. Then, the reaction flask is connected to a balloon filled with carbon dioxide, thus ensuring that there is enough carbon dioxide to fully react with the substrate. Subsequently, a quantitative amount of liquid reagent was added into the reaction vial via syringe. The reaction was initiated by turning on the 35 W UV lamp (365 nm). When the reaction finished in set time, a small amount of the mixture was taken for NMR test and calculation of the conversion yield.

2.3. Catalyst Recycling Procedure

After each experiment, 25 mL of ethyl acetate was added to the reaction mixture to induce the precipitation, and then filtered and dried in vacuum for 2 h. Then, the collected catalyst was used for next run.

3. Results and Discussion

3.1. Optimization of Reaction Conditions

The formylation of N-methylaniline and CO₂ was chosen as the model reaction to screen the optimal conditions with DTPA as the catalyst, and xanthone as the photosensitizer (Figure 1).

3.1.1. Light Intensity Effect on the Reaction

First, the reaction of N-methylaniline and CO₂ was carried out without light, and almost no product was obtained even at 80 °C (entry 1,

Table 1. Effect of light intensity on the N-formylation reaction of N-methylaniline with CO₂.

Entry	LED (W)	Conversion (%) a,b	Selectivity (%) b (2a:2b)
1 c	-	trace	-
2	20	88	96:4
3	35	92	93:7
4	50	85	93:7
5	80	81	90:10
6	100	80	88:12

^a Reaction conditions: N-methylaniline (2.5 mmol), xanthone (20 mol%), DTPA (10 mol%), DMSO as solvent (1 mL), CO₂ (0.1 MPa), PhSiH₃ (2.5 mmol), rt, and 48 h. ^b The yield was determined by ¹H NMR spectra analysis using 1,3,5-trimethoxybenzene as an internal standard. ^c Without LED light, 0.1 MPa CO₂ pressure, 80 °C.

). Then, the wattage of the LED lamp was adjusted. When a 35 W UV lamp was used as the light source, the conversion of N-methylaniline reached 92%, and the N-formylated product was obtained with 93% selectivity (entry 3, Table 1). As the wattage continues to increase, both conversion rate and selectivity further decreases. The increase in wattage will lead to an increase in temperature (<10 °C), and higher temperature is favorable for the generation of N-methylation products, and this trend is also observed in the non-photocatalytic systems [22,23,24,25].

3.1.2. Screening of Amount of DTPA and Xanthone and Reaction Time

The amounts of DTPA and xanthone were calculated according to substrate N-methylaniline. As shown in

Table 2. Effect of DTPA/xanthone amount and reaction time on the N-formylation reaction of N-methylaniline with CO₂.

Entry	DTPA (mol%)	Xanthone (mol%)	Yield(%) a,b	Selectivity (%) b (2a:2b)
1	5	20	48	85:15
2	10	20	92	93:7
3	20	20	93	92:8
4	10	10	51	90:10
5	10	5	43	90:10
6 c	10	20	37	93:7
7 d	10	20	94	78:12

^a Reaction conditions: N-methylaniline (2.5 mmol), DMSO as solvent (1 mL), CO₂ (0.1 MPa), PhSiH₃ (2.5 mmol), rt, and 48 h. ^b The yields were determined by ¹H NMR spectra analysis using 1,3,5-trimethoxybenzene as an internal standard. ^c 24 h. ^d 72h.

, when DTPA dosage increased from 5 mol% to 10 mol% while keeping the xanthone amount at 20 mol%, the conversion of substrate increased sharply from 48% to 92%, and a much better selectivity of 93% for the N-formylated product was obtained (entries 1 and 2, Table 2). However, a further increase of DTPA amount to 20 mol% did not result in a significant change in yield and selectivity (entry 3, Table 2). When the xanthone amount decreased from 20 mol% to 10 mol% with a DTPA amount of 10 mol%, the conversion of substrate decreased dramatically from 92% to 51%, and the selectivity for N-formylated product also dropped (entry 4, Table 2). A further decrease in the xanthone amount to 5 mol% resulted in a lower yield. A longer reaction time than 48 h did not result in a significant increase in yield but much worse selectivity (entry 7, Table 2), and a shorter reaction time of 24 h resulted in a dramatic decrease in yield.

3.1.3. Screening of Solvents

The polarity of the solvent also has a strong influence on this reaction system (Figure 2). When the polar DMSO and DMF were used as solvents, the conversion of the reactants reached 92% and 93%, respectively. However, when DMF was used as a solvent, the selectivity of the N-formylated product decreased to 68% dramatically. When slightly less polar acetonitrile was used as the solvent, the conversion decreased to 64%, and the selectivity of the N-formylation product was 82%. When the solvent was changed to the less polar THF, the conversion of N-methylaniline was further reduced to 58%, and the selectivity of the formylated product was only 54%. Overall, the polarity of the solvent has a greater effect on the reaction, with greater polarity favoring the N-formylation reaction [26,27].

3.1.4. Screening of Reducing Agents and Their Dosages

In the formylation of amines with carbon dioxide, the choice of a reducing agent plays an important role in improving the yield and selectivity (

Table 3. Effect of reducing agents and their amounts on the N-formylation of N-methylaniline with CO₂.

Entry	Hydrosilane	Dosage (mmol)	Conversion (%) a,b	Selectivity (%) b (2a:2b)
1	PhSiH3	2.5	92	100:0
2	Ph2SiH2	2.5	50	100:0
3	Ph3SiH	2.5	26	100:0
4	Me2PhSiH	2.5	trace	-
5	Et3SiH	2.5	24	100:0
6 c	PhSiH3	1.25	48	85:15
7 d	PhSiH3	3.75	93	91:9

^a Reaction conditions: N-methylaniline (2.5 mmol), xanthone (20 mol%), DTPA (10 mol%), DMSO (1 mL), CO₂ (0.1 MPa), PhSiH₃ (2.5 mmol), rt, and 48 h. ^b The yields were determined by ¹H NMR spectra analysis using 1,3,5-trimethoxybenzene as an internal standard. ^c PhSiH₃ (1.25 mmol) 48 h. ^d PhSiH₃ (3.75 mmol).

). In the screening of phenylhydrosilanes, the selectivity of the N-formylated product reached 100% as the number of phenyls increased. However, there was a significant decrease in the conversion of N-methylaniline from 92% (entry 1, Table 3) to 50% (entry 2, Table 3) and 26% (entry 3, Table 3). Overall, the fewer the number of hydrogens in the reductant hydrosilane, the bigger the steric hindrance of the substituent and the lower the conversion of the substrate.

The amount of reducing agent also significantly affects the yield of this reaction. When 1.25 mmol of PhSiH₃ was used, the conversion was only 48% (entry 5, Table 3). When 3.75 mmol of PhSiH₃ was used in the reaction, there was no significant increase in the conversion, while the selectivity for the formylated product decreased slightly (entry 6, Table 3). Therefore, 2.5 mmol of PhSiH₃ was chosen as the optimal amount of reducing agent.

In summary, the optimal photocatalytic conditions were screened as 20 mol% xanthone, 10 mol% DTPA, 35 W, 48 h, and 2.5 mmol PhSiH₃.

3.2. Studies on the Suitability of N-Formylated Substrates for Secondary Amines

The application scope of catalytic systems has been widely tested under optimal photocatalytic conditions. As shown in

Table 4. N-formylation of secondary amines.

Entry	Substrate	Product	Yield (%) a,b
1			88.5
2			79.4
3			77.0
4			98.0
5			45.2
6			92.6
7			88.5
8			72.9
9			96.5
10			95.2
11			92.6
12			90.1

^a Reaction conditions: Substrate (2.5 mmol), xanthone (20 mol%), DTPA (10 mol%), DMSO (1 mL), CO₂ (0.1 MPa), PhSiH₃ (2.5 mmol), rt, 35 W, and 48 h. ^b The yields were determined by ¹H NMR spectra analysis using 1,3,5-trimethoxybenzene as an internal standard.

, for most of the secondary amines, the corresponding formamide products were obtained in good to excellent yields (70–95%). Good yields of 88.5%, 79.4%, and 77.0% were also obtained when dialkylamines, such as di-n-propylamine, di-n-butylamine, di-n-hexylamine, and di-isopropylamine, were used as substrates for the reaction, respectively. The yield decreases gradually with the increase in carbon chain length, probably due to the increase in viscosity and steric hindrance. In most cases of cyclic secondary amines, hexagonal ring compounds of piperidines with CO₂ formed corresponding N-formylated products in good to excellent yields of 72.9–92.6%. However, when pentacyclic pyrrolidine was used as substrates, a yield of only 45.2% was obtained, probably due to their larger ring tension. Heterocyclic amines, such as morpholine and thiomorpholine with electron-rich atoms O or S, can easily react with CO₂ to yield N-formylated products with excellent yields (>95%). Higher-molecular-weight secondary amine compounds with aromatic phenyl group can also react with CO₂ in high yields (>90%) when catalyzed by the present photocatalytic system.

3.3. Studies on the Suitability of N-Formylated Substrates for Primary Amines

3.3.1. Investigation of Aromatic Primary Amine Substrate Suitability

Aromatic primary amines were also explored under the optimal reaction conditions. As can be seen from

Table 5. Aromatic primary amine N-formylation reactions.

Entry	Substrate	Product	Yield (%) a,b
1			82.6
2			79.4
3			75.8
4			73.5
5			53.5
6			69.0
7			53.5
8			39.4
9			48.8
10			94.3
11			91.7
12			86.2
13			87.7

^a Reaction conditions: Substrate (2.5 mmol), xanthone (20 mol%), DTPA (10 mol%), DMSO (1 mL), CO₂ (0.1 MPa), PhSiH₃ (2.5 mmol), rt, 35 W, and 48 h. ^b The yields were determined by ¹H NMR spectra analysis using 1,3,5-trimethoxybenzene as an internal standard.

, N-formylation for most of the aromatic primary amines and CO₂ showed good selectivity: more than 99% of the products are monosubstituted formamide products, and almost no double-substituted or methylamine products were produced. Aniline containing halogen atoms (e.g., -F, -Br, -Cl) in the para position gave moderate yields of 79.4%, 75.8%, and 73.5%, respectively. When p-nitroaniline was used as the substrate, the yield of the product was only 53.5% due to the strong electron-withdrawing effect. Primary amines with electron-donating groups such as methyl, ethyl, n-butyl, and isopropyl in the para position of aniline have progressively lower yields as the carbon chain grows. In the case of ethylaniline, the corresponding yield of o-ethylaniline was much higher than that of para-ethylaniline, probably indicating the existence of an intermolecular steric hindrance of substrates to N-formylation. However, for ortho-substituted anilines, such as 2,6-dimethylaniline, 2,6-diethylaniline, and 2,6-diisopropylaniline, which have larger steric hindrance, spatial potential barrier effect was not so notable, and the efficiency of the N-formylation reaction in yield ranges from 86.2 to 91.7%.

3.3.2. Study of the Suitability of Aliphatic Primary Amine Substrates

Subsequently, aliphatic primary amines were explored. As shown in

Table 6. N-formylation reaction of aliphatic primary amines.

Entry	Substrate	Product	Yield (%) a,b
1			90.1
2			87.0
3			85.5
4			50.9
5			40.3
6			38.0
7			53.5
8			46.1
9			98.0
10			48.8

^a Reaction conditions: Substrate (2.5 mmol), xanthone (20 mol%), DTPA (10 mol%), DMSO (1 mL), CO₂ (0.1 MPa), PhSiH₃ (2.5 mmol), rt, 35 W, and 48 h. ^b The yields were determined by ¹H NMR spectra analysis using 1,3,5-trimethoxybenzene as an internal standard.

, aliphatic primary amines showed higher reactivity than aromatic primary amines. Primary amines with different alkyl chains, e.g., n-butylamine, n-hexylamine, and n-octylamine, provided N-formylated products with excellent yields (85–90%). It was also found that the product yield decreased with the growth of the carbon chain. This phenomenon is consistent with the data reported by Zhang et al. [28] using nano Pd/C catalyst at 105 °C for 24 h to generate N-formylation. The yields in N-formylation of tert-butylamine and cyclohexylamine with larger steric hindrance were only 50.9% and 40.3%, respectively, after 48 h of reaction. Substituted benzylamines, with both electron-donating (-OCH₃) and electron-withdrawing (-F) groups, have higher yields compared to benzylamines (38.0%). Extension of the carbon chain between the benzene ring and the amino group resulted in excellent yield of N-formylated compound (98.0%). When using alkyne propylamine as a substrate, the alkyne group was not reduced even in the presence of the reducing agent PhSiH₃, and N-formylation occurred only on the amino group during the whole reaction period; similar reports have been presented in the literature [29].

3.4. Catalyst Recycling

In order to test the stability of the catalyst, catalyst recycling experiments were conducted under optimal reaction conditions (Figure 3). After each cycle, ethyl acetate (25 mL) was added to the reaction mixture to induce precipitation of the reactants; then, after filtration and drying in a vacuum for 2 h, the catalyst was used repeatedly. The results show that there was no significant loss of DTPA activity after five cycles of recycling.

In addition, comparing the IR spectra (Figure 4) of DTPA after five cycles of recycling with that of fresh DTPA finds that the structure of DTPA was still maintained, indicating good stability and recyclability of DTPA.

3.5. CO₂ Photocatalytic Reaction Mechanism

Based on previous literature reports [30,31,32,33], a possible catalytic mechanism was proposed in Figure 5. Firstly, the xanthone acts as a photocatalyst (PC) to form the excited state of PC* under the irradiation of light. PC* undergoes single electron transfer (SET) with DTPA to form a DTPA radical cation. Reaction of DTPA⁺ with phenylsilane forms phenylsilyl radicals and releases H⁺. Meanwhile, CO₂ reacts with aniline to form the carbamate anion (I). Then, carbamate anion reacts with phenylsilyl radicals to form radical anions (II). Subsequent decomposition of radical anions can generate the carbamoyl radical (III). Carbamoyl radical can undergo a hydrogen atom transfer process, which ultimately produces the desired product formamide.

4. Conclusions

Commercially available DTPA as a catalyst and relatively inexpensive xanthone as a photocatalyst have been used to form a photocatalytic system to promote the photocatalytic N-formylation of amines with CO₂ efficiently. Under the optimal conditions, N-formylated products were obtained in moderate to good yields with high selectivity. In addition, the catalyst exhibited high stability and good recyclability via simple precipitation and filtration. No significant changes in the catalytic properties and structure of DTPA were observed after five cycles of recycling. The present photocatalytic catalytic system does not require high reaction temperature and CO₂ pressure, manifesting an environmentally friend and non-toxic catalytic system with potential application for the industrial synthesis of formamide.