Pathways and Kinetics for Autocatalytic Reduction of CO2 into Formic Acid with Fe under Hydrothermal Conditions.

The utilization of CO2, as a cheap and abundant carbon source to produce useful chemicals or fuels, has been regarded as one of the promising ways to reduce CO2 emissions and minimize the green-house effect. Previous studies have demonstrated that CO2 (or HCO3 -) can be efficiently reduced to formic acid with metal Fe under hydrothermal conditions without additional hydrogen and any catalyst. However, the pathways and kinetics of the autocatalytic CO2 reduction remain unknown. In the present work, the reaction kinetics were carefully investigated according to the proposed reaction pathways, and a phenomenological kinetic model was developed for the first time. The results showed that the hydrothermal conversion of HCO3 - into formic acid with Fe can be expressed as the first-order reaction, and the activation energy of HCO3 - is 28 kJ/mol under hydrothermal conditions.

Introduction

Due to the increasing demand for energy and consumption of fossil fuels, the level of CO2 atmosphere has raised at a faster rate, which leads to a series of problems in the environment and ecological balance.[1−3] In recent years, a great deal of focus has been expended to reduce the CO2 concentration in the atmosphere, such as photosynthesis, electrochemical and biochemical technology, and so on.[4−7] Among these methods, artificial photosynthesis is regarded as one of the most promising methods for solar energy technologies.[8] However, there are still many challenges in the direct conversion of CO2 using solar energy, such as the low conversion efficiency and product selectivity. Recently, the catalytic reduction of CO2 with hydrogen has attracted increasing attention due to its commercial feasibility.[9−11] Nevertheless, high purity hydrogen and noble metal catalysts (Ir, Ru, Rh, etc.) are usually needed in the reduction process, leading to high energy consumption and cost.[12,13] Therefore, the development of an alternative method for the feasible reduction of CO2 is highly desirable.

Hydrothermal chemistry has played an important role in the formation of fossil fuels and origin of life in the earth’s crust and deep-sea hydrothermal vents.[14−16] The abiotic synthesis of organics suggests that highly efficient dissociation of H2O and the subsequent reduction of CO2 into organics could be achieved with metals under hydrothermal conditions. In the abiotic synthesis of organics, the generally inferred pathway involves the reduction of CO2 dissolved in water that accompanies the hydrothermal alteration of minerals, in which a primary role for the minerals is to generate H2 through the reducing conditions as the reaction of ferrous Fe-bearing minerals with water.[17,18] Recently, Jin’s group has developed a new strategy for the hydrothermal reduction of CO2 with various zero-valent metals, such as Fe, Zn, and Mn, and it was found that formic acid was the main product from reduction of CO2.[19−22] As an important chemical, formate can serve as the raw material for the environmentally friendly road de-icer.[23] Furthermore, the dehydrogenation of formic acid can proceed easily under mild conditions.[24] Therefore, as an excellent hydrogen storage carrier, formic acid can play an important role in the future context of a hydrogen energy economic picture.

Although previous research has demonstrated the potential of the autocatalytic reduction of CO2 into formic acid with zero-valent metal Fe under hydrothermal conditions, few studies have been focused on investigating the comprehensive pathways and reaction kinetics in CO2 reduction. In this work, the detection and distribution for all products from the autocatalytic hydrothermal reduction of CO2 were conducted, and then, based on these results, a possible reaction network and a quantitative model for the kinetics of hydrothermal carbon dioxide reduction were developed. Considering that HCO3– is the product of CO2 captured from waste streams by basic solution, NaHCO3 was used as the CO2 source. Simultaneously, the application of NaHCO3 can also simplify the experimental procedure and ensure the accuracy of carbon amount.

Results and Discussion

Product Distribution

First, a series of experiments were conducted to investigate the distribution of products from the reduction of NaHCO3 in water with Fe as a reductant. From Figure , it was shown that formic acid was the main liquid product. A little amount of acetic acid was also detected after 600 s of reaction time. With the increase of the temperature, the conversion of HCO3– significantly increased from 30 to 50 mol %, similarly with the trend for the yield of formic acid. However, the yields of gas products remained steady without significant change.

Figure 1

Distribution of products from hydrothermal reduction of HCO3– (2 mmol NaHCO3, 12 mmol Fe, and 600 s).

The analysis of the gas samples by gas chromatography/thermal conductivity detection (GC/TCD) showed that H2, CO2 and a trace amount of CO were produced at a reaction retention time of 600 s. In Table , it is obviously seen that the hydrogen was the main product in gas products, which was mainly from the decomposition of H2O. Only a trace amount of CO was produced after the reaction. With the increasing temperature, the ratio of carbon dioxide was first decreased and then increased, which means that the decomposition of products such as formic acid and acetic acid appeared at higher temperatures.

Table 1

Gas Product Distribution from Hydrothermal Reduction of HCO3– at Different Temperatures (2 mmol NaHCO3, 12 mmol Fe, and 600 s)

temperature (°C)	H2 (wt %)	CO2 (wt %)	CO (wt %)
250	90	9.1	0.9
300	91	7.9	1.1
350	91	8.4	0.6

Effect of Reaction Time and Temperature

A series of experiments were carried out to investigate the effect of residual time and temperature on hydrothermal conversion of HCO3– by varying the time from 0 to 600 s and the reaction temperature from 250 to 350 °C with the same amount of Fe (12 mmol), respectively. As shown in Figure a, it was suggested that the yield of formic acid has shown a rapid increase with the increase of reaction time in 180 s at 250 and 300 °C, while the yield of formic acid increased linearly further with the reaction time over 300 s in all test temperatures. For the temperature at 350 °C, the yield of formic acid increased rapidly at all set reaction times. However, the trend of acetic acid yield was in contrast to that of formic acid. In Figure b, it was observed that the yield of acetic acid increased first from 0 to 600 s and dropped a little in 600 s at 350 °C. With regard to the decomposition of acetic acid, the decarboxylation pathway existed under hydrothermal conditions.[25] It has also been reported that acetic acid decomposed preferentially into CO2 and H2 at a temperature of 325 °C and a pressure of 350 bars.[26] The possible reason for the decreasing yield of acetic acid is that the decarboxylation of acetic acid dominated gradually and exceeded the formation rate of acetic acid with increasing temperature.

Figure 2

Effect of the reaction time and temperature on the yield of (a) formic acid and (b) acetic acid (2 mmol NaHCO3, 12 mmol Fe and 600 s).

Effect of the Fe Amount

To further examine the effect of the hydrogen amount on the yield of product distribution, the effect of the initial Fe amount was also investigated at 350 °C. In Figure , it is shown that the yield of formic acid, acetic acid, and gas products all improved in the set reaction time (600 s) with the increase of the amount of Fe from 2 to 12 mmol. The high formic acid yield can be obtained without adding any other catalysts, which may be caused by several factors. First, being a closed system, when the H2 amount was increased, the total gas amount and pressure increased; this caused the increase of H2 partial pressure and the solubility of H2 in the liquid phase. If the H2 concentration was increased, the reaction rate of HCOO– also increased as per Le Chatelier’s principle. The same effect can be achieved by increasing the pressure by reducing the empty volume of the reactor, which was shown by Roman-Gonzalez et al.[27] The second reason may be that Fe3O4 formed in hydrothermal conditions acted as a catalyst, which corresponds to the previous study.[28] In Section , it is also suggested that with the increase of temperature, Fe improved the formation of formic acid. It is possible that higher temperatures are favorable for the formation of Fe3O4.

Figure 3

Effect of Fe amount on the hydrothermal reduction of HCO3– at 350 °C for 600 s (■ formic acid, blue ▲ gas products, and red ● acetic acid).

To test this assumption, the X-ray diffraction (XRD) patterns of the solid residues obtained at different temperatures are shown in Figure . In our previous research, we indicated that Fe first reacts with CO2 and H2O to form FeCO3, which then loses CO2 to form Fe3O4. In addition, the existence of HCO3– also accelerated the Fe oxidation in water to produce hydrogen. Simultaneously, Fe3O4 is reduced in situ, leading to the formation of more active sites on the surface of Fe3O4–.[28] The formed hydrogen and HCO3– are activated on the Fe3O4– surface. This suggested that the more the Fe is oxidized to Fe3O4 under hydrothermal conditions, the more the surface of Fe3O4– and the amount of H2 could be acquired, improving the reduction of HCO3–.

Figure 4

XRD patterns of the solid residues obtained at different temperatures.

Kinetic Modeling for the Hydrothermal Reduction of HCO3–

Based on the above experimental data, the reaction network for hydrothermal reduction of HCO3– was investigated in the present work, which differs from that offered recently for isothermal hydrothermal reduction of bicarbonate concentration at subcritical temperatures.[29] In Figure , the reaction network is proposed, which includes a primary pathway which shows that the reversible reaction appeared between bicarbonate concentrate and formate, a secondary pathway through two stages of tandem reaction for the conversion of bicarbonate into acetate, and the final pathway that allows for gas formation from bicarbonate, formate, and acetate.

Figure 5

Reaction network for the hydrothermal reduction of CO2.

Mathematica 10.2 was employed to solve the system of ordinary differential equations and simultaneously estimate the Ai (Arrhenius pre-exponential factors) and Ei (activation energies) of hydrothermal HCO3– reduction by minimizing the sum of squared residuals (SSR) as eq .[30]

The first-order-rate law was postulated to describe the kinetics for each pathway, which corresponded with Chiang et al.[31] and then coupled them with the batch reactor design equation, as shown below in eqs –8. The subscripts on each mass fraction, x and k, referred to the yields of each product fraction and the rate constants for pathways (liquids = HCO3–, FA = formic acid, AA = acetic acid, G = gas). The temperature profiles of the proxy reactors for each set point temperature were fitted using power series models and incorporated directly into the model to give the reactor temperature as a function of time. Additionally, we have assumed that all reactions take place in a single fluid phase.

In Table , we describe the kinetics parameters and optimized Arrhenius parameters for the hydrothermal reduction of HCO3–. It was determined that the activation energy for formation of formic acid from HCO3– during hydrothermal conditions is about 28 kJ/mol, which is similar with the results from the range determined previously for decomposition of acids.[32] It was indicated that the kinetic constant for formic acid is significantly increasing with the temperature, which is consistent with the results from our experiments. The earlier kinetics modeling work on conventional hydrogenation of CO2 provided an opportunity for comparison of these activation energies (80 kJ/mol), which are higher than our results from hydrothermal reduction.[33] However, the activation energy of acetic acid from formic acid is about 72 kJ/mol, which is the rate-determining step to limit the formation of acetic acid. The activation energies for the gasification of acetic acid are also comparable to activation energies reported by Belsky et al. for the decarboxylation of acetic acid and its derivatives (71–178 kJ/mol).[34]

Table 2

Arrhenius Parameters for the Hydrothermal Reduction of HCO3–

rate constant	pathway	k250°C (min–1)	k300°C (min–1)	k350°C (min–1)	ln A	Ei (kJ/mol)
k1	liquids → formic acid	0.00817	0.0143	0.0230	1.63	28
k2	formic acid → solids	0.00290	0.00657	0.0131	3.58	41
k3	liquids → gas	0.00022	0.000859	0.00269	7.21	68
k4	formic acid → acetic acid	7.04 × 10–6	2.98 × 10–5	0.0001	4.69	72
k5	formic acid → gas	3.08 × 10–7	1.84 × 10–6	8.22 × 10–6	5.47	89
k6	acetic acid → gas	1.12 × 10–9	9.76 × 10–9	6.02 × 10–8	4.22	108

According to the determined the kinetic parameters based on the results of experiments, the correlation between the model calculations and the experimental data for the hydrothermal reduction of HCO3– at different temperatures and times are displayed in Figure . The model accurately describes the trends in the data and provides the species concentrations within experimental errors. From Figure a–c, it was also illustrated that the proposed reaction network and optimized Arrhenius parameters could capture the trends in the observed product yields for all three set point temperatures, which is also consistent with the first-order-rate law as postulated.

Figure 6

Experimental (points) and model calculated yields (continuous curves) for the hydrothermal reduction of HCO3– at set point temperatures (a) 250, (b) 300, and (c) 350 °C with different reaction times (0–600 s).

We also compared the yield of each product from experiment and the yield data from the predicted model in Figure . It displays a plot that compares the experimental and predicted product yields, in which the circles represent data from temperatures with long reaction time (0–20 min), triangles represent data from the amount of 14 mmol Fe, and squares represent data from 275 °C. It has been shown that the predicted data are almost below the diagonal. Though our perfection may not be fitted together to predict the product yields from reduction of CO2, it was a true prediction to an extrapolation to a reaction regime with limited parametrization.

Figure 7

Parity plot for product fraction yields from fast reduction of CO2. Circles represent data from the temperature with a long reaction time (0–20 min). Triangles represent data from the amount of 14 mmol Fe. Squares represent data from 275 °C.

Experimental Section

Materials

Zero-valence metal Fe powders (325-mesh size) were obtained from Aladdin Chemical Reagent, and NaHCO3 (used as a CO2 source) was obtained from Sinopharm Chemical Reagent Co., Ltd. All other reagents were commercially available from Sinopharm Chemical Reagent Co. Ltd. Deionized water was used throughout the study. The reactions were conducted by 316 stainless-steel Swagelok tube fittings with an internal volume of approximately 5.7 mL.

Reactors and Hydrothermal Reduction Procedure

In a typical experimental procedure, the desired amount of NaHCO3 (2 mmol), Fe powder, and deionized water was loaded in a batch reactor to occupy 35% of the total reactor volume. After loading, the reactor was immersed in a salt bath, which has been preheated at the set point temperatures of 250, 300, and 350 °C. It took the time at which the reactor reached its isothermal temperature as t = 0. While the reactors remained in the salt bath for an additional 0–600 s, the reactors were removed from the salt bath and cooled in a cold-water bath to quench the reactions. Then, the gas products were initially collected into a TCD for analysis, and the liquid and solid samples were separated through the filter membrane (0.22 μm filter film) for analysis, respectively. Liquid samples were analyzed by high-performance liquid chromatography (HPLC) (Agilent Technologies 1200 system), GC–flame ionization detector/mass spectrometry (Agilent 6890 gas chromatographs with a mass spectrometric), and a total organic carbon (TOC) analyzer (Shimadzu TOC 5000A). Solid residues were analyzed by XRD (Bruker D8 Advance X-ray diffractometer). All the experiments were conducted three times to force the influence of systematic errors.

Quantitative estimation of formic acid and acetic acid was based on the average value from the HPLC analysis, which was analyzed by the HPLC analyzer equipped with two Shodex RSpak KC-811 columns in series and a refractive index detector. The mobile solvent of the HPLC was HClO4 (2 mmol/L) on a flow rate of 1.0 mL/min and holding 30 min. The yield of acids was defined as the percentage of formic acid or acetic acid and the initial NaHCO3 on a carbon basis as follows in eqs , 7where CFA, CAA, and CS are the amounts of carbon in formic acid, acetic acid, and the initial NaHCO3 added to the reactors.

For gas products, a 15 ft stainless steel column, as Carboxen 1000 (Supelco), was employed to separate each component in the gas phase sample, which used Ar as the carrier gas (15 mL/min). The programmed temperature of the column was first held at 40 °C for 5 min, followed by heating at a rate of 20 °C/min until it reached 220 °C and held for 10 min, which has a total of 24 min for the run time. To quantify the amount of gas, the high-pressure valve of the reactor was connected to a sampling valve attached to an Agilent Technologies model 6890N GC equipped with a TCD. The calibration curve relates the mole fraction, yi, and the peak area for each component. The molar yield, ni, of each component was subsequently calculated from the mole fractions of each compound detected in the gas chromatograph using eq , which was determined by the moles of N2 from air in the reactor using the ideal gas law.

Conclusions

In this work, a kinetic study was conducted for the first time to investigate the distribution of products from the autocatalytic hydrothermal reduction of HCO3– using zero-valent metal Fe, and the pathways were proposed. The rate constants, activation energy, and frequency factors were calculated according to the Arrhenius equation, which is consistent with the first-order-rate law as postulated. The activation energy for the formation of formic acid from HCO3– reduction is about 28 kJ/mol, which is much lower than that in the earlier kinetic modeling work on conventional hydrogenation of CO2. The present study is helpful for providing a promising perspective to show the pathways and phenomenological kinetics of hydrothermal reduction of carbon dioxide.