Polysulfide-assisted urea synthesis from carbon monoxide and ammonia in water

Experimental results

Figure 1A shows the yield of urea, based on CO, at three different temperatures (i.e., 35 °C, 50 °C, and 65 °C). Urea formation was faster at higher temperatures, reaching 80% yield within 24 h at 65 °C, 48 h at 50 °C, and 168 h at 35 °C. Longer reaction duration led to further urea formation; 95–100% yield of urea was obtained after the 32-, 64-, and 216-h reaction at 65 °C, 50 °C, and 35 °C, respectively (Fig. 1A). Similar time and temperature dependencies were observed at an NH₃ concentration of 5.6 M (Fig. S5). CO₂ was formed as a by-product, but the yield was at most 0.8% under the examined reaction conditions. When no NH₄Cl was added as a starting material, the urea selectivity slightly decreased, while CO₂ formed in higher amount, compared with those in the presence of NH₄Cl under otherwise identical condition. For example, the 120-h reaction at 50 °C in the absence of NH₄Cl resulted in 94 ± 5% yield of urea and 2 ± 0.2% yield of CO₂. No urea was detected when either NH₃ or S⁰ was absent and when CO was replaced with CO₂. In the absence of NaHS, the reaction started with a sluggish urea-formation stage (green dotted line in Fig. 1A). This induction period was followed by an accelerated growth of the urea yield that eventually exceeded 90%. The mechanism underlying the sigmoidal-like growth is discussed later.

The initial amount of S⁰ had a strong impact on the concentration of H₂S by-product, as well as the yield of urea, as seen in the results obtained after the 120-h reaction at 50 °C (Fig. 1B). With an increase in the S⁰/CO molar ratio from 0 to 19.1, the H₂S gas concentration initially exhibited a rapid increase to approximately 6 Pa, followed by a decrease to a value less than 1 Pa at the S⁰/CO molar ratio of 4.8, and then reached a steady value (ca. 0.2 Pa) at higher S⁰/CO molar ratios. As for urea, the yield increased significantly when the S⁰/CO molar ratio was between 0 and 2 at the expense of CO (Fig. 1B).

The amount of S⁰ always decreased throughout the reaction; few or no S⁰ remained in the experiments with the initial S⁰/CO molar ratio less than 4.8 (Fig. S6). Meanwhile, the aqueous solutions exhibited a yellow color characteristic of PSs (Bedoya-Lora, Hankin & Kelsall, 2019); darker yellow was observed when larger amount of S⁰ was used (Fig. S6). These features indicate that S⁰ was reduced to PSs coupling with CO oxidation to urea (Eq. (1)) and CO₂ (Eq. (2))

(1) $$\begin{array}{c}CO+2N{\mathrm{H}}_3+n{\mathrm{S}}^0\to {\left(\mathrm{N}{\mathrm{H}}_2\right)}_{2}CO+2{\mathrm{H}}^{+}+{\mathrm{S}}_n^{2-}\end{array}$$

(2) $$\begin{array}{c}CO+{\mathrm{H}}_{2}O+n{\mathrm{S}}^0\to \mathrm{C}{\mathrm{O}}_2+2{\mathrm{H}}^{+}+{\mathrm{S}}_n^{2-}\end{array}$$

Indeed, the molar amount of PSs formed through the reaction exhibited a one-to-one correlation with the amount of CO consumed (Fig. S7).

Thermodynamic calculation

Thermodynamic calculation predicts that in the presence of S⁰, PSs are the dominant dissolved S species over HS⁻ and H₂S at the examined pH (10.3–10.5) (Fig. 2A). The presence of PSs with S⁰ thus suppress the accumulation of H₂S by-product in the gas phase, keeping the H₂S gas concentration below 0.5 Pa in our experimental setting (Fig. 2B). When no PS formation was assumed, two orders of magnitude higher values (13–21 Pa) were calculated for the steady-state H₂S gas concentration after the complete conversion of CO to urea or CO₂ (Fig. 2B). Because the mean chain length of PSs is 5.0 (i.e., n = 5 in S_n²⁻, Fig. 2A), the capability of PSs to suppress H₂S accumulation is expected to work effectively at S⁰/CO molar ratios higher than five. When an insufficient amount of S⁰ is used (e.g., S⁰/CO molar ratio = 2), in contrast, the S⁰-to-PSs reduction cannot fully extract the reduced sulfur (S⁰ → S²⁻), resulting in an elevated emission of H₂S gas (Fig. 1B). Thermodynamic calculations taking PSs into account thus explain our experimental results for H₂S. At the reaction temperatures of 35 °C, 50 °C, and 65 °C, the gas-phase H₂S are expected to reach 0.3–0.8, 1.4–3.0, and 4.7–9.8 Pa, respectively.

Based on the thermodynamic calculation (Fig. 2A), the urea synthesis reaction can be simplified as:

(3) $$\begin{array}{c}CO+2N{\mathrm{H}}_3+5{\mathrm{S}}^0\to {\left(\mathrm{N}{\mathrm{H}}_2\right)}_{2}CO+2{\mathrm{H}}^{+}+{\mathrm{S}}_5^{2-}\end{array}$$

On average, five atoms of S⁰ were converted to a single molecule of S₅²⁻ when one molecule of CO was consumed to form urea.

Kinetic characteristics

In water, PSs readily react with CO to form OCS (Kamyshny et al., 2003), a likely key intermediate in PS-assisted urea synthesis (Scheme 1). OCS was indeed observed in our experiments as a trace and transient gas species (Fig. S8). The formation rate of OCS is linearly proportional to the concentrations of CO and PSs (Kamyshny et al., 2003). Once formed, the OCS-to-urea conversion proceeds in competition with the OCS hydrolysis to CO₂, whose half-life is in the seconds-to-minutes range under the examined reaction conditions (e.g., 18 s at pH10.0 and 50 °C; Kamyshny et al., 2003). Because CO₂ was a minor product in our experiments although the rate of OCS hydrolysis is much higher than that observed for urea synthesis (Fig. 1A), OCS formation is expected to be the rate-determining step of the overall urea synthesis process.

In accordance with this kinetic consideration, the experimental results (Fig. 1A), together with the corresponding CO consumption, were well represented with the following second-order rate equation parameterizing the CO and PS concentrations (Figs. 3A–3C ):

(4) $$\begin{array}{c}{\displaystyle \frac{\mathrm{d}\left[\mathrm{u}\mathrm{r}\mathrm{e}\mathrm{a}\right]}{dt}=k\left[PSs\right]\left[CO\right]}\end{array}$$where [x] denotes the concentration of species x in water. Note that the urea-formation reaction (Eq. (1)) generates an equimolar amount of PSs while consuming an equimolar amount of CO. PSs and CO thus has the following kinetic relationships with urea in this reaction:

(5) $$\begin{array}{c}{\displaystyle \frac{\mathrm{d}\left[\mathrm{P}\mathrm{S}\mathrm{s}\right]}{dt}={\displaystyle \frac{\mathrm{d}\left[\mathrm{u}\mathrm{r}\mathrm{e}\mathrm{a}\right]}{dt}}}\end{array}$$

(6) $$\begin{array}{c}{\displaystyle \frac{\mathrm{d}{[\mathrm{C}\mathrm{O}]}_{Tot}}{dt}=-V_L{\displaystyle \frac{\mathrm{d}\left[\mathrm{u}\mathrm{r}\mathrm{e}\mathrm{a}\right]}{dt}}}\end{array}$$

In Eq. (6), the time derivative of the total amount of CO ([CO]_Tot) is correlated with that of urea concentration multiplied by the volume of sample solution (V_L), because CO is distributed in both the gas and aqueous phases. Thus, as long as competing reactions (e.g., Eq. (2)) are not significant, CO and PSs do not need to be monitored during the experiment for the use of Eq. (4). Considering the high OCS hydrolysis rate at high pH (Kamyshny et al., 2003), a moderately alkaline pH (e.g., pH 10–11) is preferable for such selective urea synthesis. It is also suggested from the adequate kinetic representation (Figs. 3A–3C) that the other intermediate species (i.e., thiocarbamate and isocyanate; Scheme 1) occupy a tiny fraction of the carbon compounds present in the course of the reaction.

Equations (4)–(6) also enabled us to reproduce the experimental results obtained without NaHS as the starting material (Fig. S9). However, this simulation requires the initial PS concentration to be a non-zero value (0.18 mM; see the caption of Fig. S9 for the other parameters). This low PS concentration is likely derived from the hydrolysis of S⁰ (4S⁰ + 4H₂O → 3HS^– + SO₄²⁻ + 5H⁺) (Ellis & Giggenbach, 1971) followed by the formation of PSs from S⁰ and HS⁻ ((n−1)S⁰ + HS⁻ → S_n²⁻ + H⁺).

The determined rate constants k at three reaction temperatures (35 °C, 50 °C, and 65 °C) exhibited a linear trend in the Arrhenius plot (Fig. 3D). Extrapolation of the regression line to lower temperatures led to values roughly consistent with the reported rate constants for OCS formation from CO and PSs (diamond symbols in Fig. 3D) (Kamyshny et al., 2003).

The activation energy (Ea = 67.8 kJ mol⁻¹) and pre-exponential factor (ln(A) = 21.6) obtained from the Arrhenius plot (Fig. 3D) enables us to predict the urea yield as a function of time in a range of temperature. For example, 80% yield of urea is calculated to be achieved within 98, 47, 23, 12, and 6 h at 40 °C, 50 °C, 60 °C, 70 °C, and 80 °C, respectively (Fig. 3E). Such estimations, together with the thermodynamic calculation for H₂S exemplified above, should be an important basis for realizing urea manufacturing with maximum productivity while minimizing the emission of H₂S gas.

Versatility of the PS-assisted reaction mechanism

Several techniques are available for the isolation and purification of urea from aqueous solution, such as forward osmosis (Ray, Perreault & Boyer, 2019), the use of adsorbents (e.g., activated carbons, zeolites, ion-exchange resins, and silica) (Urbanczyk, Sowa & Simka, 2016), and selective dissolution in certain solvents (e.g., ethanol) and subsequent recrystallization (Marepula, Courtney & Randall, 2021). Methods have also been developed for NH₃ recovery from wastewaters (Kuntke et al., 2018; Kim, Moreno-Jimenez & Efstathiadis, 2021). PSs is oxidized by O₂ to S⁰ (Steudel, Holdt & Nagorka, 1986; Kleinjan, de Keizer & Janssen, 2005), which may be used for the next round of urea production. Urea can be isolated from dissolved inorganic species by utilizing its higher solubility in organic solvents (Marepula, Courtney & Randall, 2021). When ethanol is considered as a solvent for urea isolation from our product solution, the optimum purity of 88.2 mol% is expected based on the reported solubility data of urea and ammonium, sodium, and sulfate salts (Lee & Lahti, 1972; Toro, Dobrosz-Gomez & Garcia, 2014; Galvao et al., 2021). Experimental test of the isolation process and its further development must be a crucial next step for urea commercialization. It is also noteworthy that in addition to urea, NH₃, S⁰, and SO₄²⁻ (a by-product of PS oxidation) are well-used fertilizer components. Thus, not only as an isolated form, urea as a mixture component with the other N and S species could also be useful in agriculture (Wagenfeld et al., 2019).

The versatility of sulfur in water demonstrated in this study should also be beneficial to the production of various urea derivatives serving as intermediates for fine chemicals, pharmaceuticals, cosmetics and pesticides (Franz & Applegath, 1961; Franz et al., 1961a, 1961b; Mizuno et al., 2006, 2007; Mizuno, Nakai & Mihara, 2009; Peng, Li & Xia, 2006); thus, possessing wide applicability in material manufacturing. The reaction mechanism may also have played a role in prebiotic chemical processes in primordial hydrothermal vent environments rich in sulfur, CO and NH₃ that eventually led to the origin of life (Li et al., 2017; Li, Kitadai & Nakamura, 2018; Nakashima et al., 2018; Lee et al., 2021; Kitadai, Kameya & Fujishima, 2017; Kitadai et al., 2018, 2019, 2021).