Formic acid production by simultaneous hydrothermal CO2 reduction and conversion of glucose and its derivatives

Journal of the Taiwan Institute of Chemical Engineers 139 (2022) 104504 Available online 30 August 2022 1876-1070/© 2022 The Authors. Published by Elsevier B.V. on behalf of Taiwan Institute of Chemical Engineers. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/). Formic acid production by simultaneous hydrothermal CO 2 reduction and conversion of glucose and its derivatives María And´ erez-Fern´ andez a , Sergio Ferrero b , Joao P.S. Queiroz c , Eduardo P´ erez d , Celedonio M. ´ Alvarez b , ´ Angel Martín a , * , M. Dolores Bermejo a a Presstech, Pressure Technologies Group, BioEcoUva Research Institute, Department of Chemical Engineering and Environmental Technology, Universidad de Valladolid, Spain b GIR MIOMeT IU CINQUIMA/Química inorg´ anica, Facultad de Ciencias, Universidad de Valladolid, Spain c Department of Chemical Engineering, Federal University of Sao Carlos, Brazil d Universidad Complutense de Madrid, Spain ARTICLE INFO Keywords: NaHCO 3 Subcritical water Biomass conversion Acetic acid Lactic acid ABSTRACT Background: The hydrothermal reduction of CO 2 using organic molecules such as alcohols can produce renewable fuels and platform chemicals, such as formic acid (FA). If the process is performed using glucose as reductant, the yield is especially high, because FA is selectively produced both by reduction of CO 2 and by degradation of glucose degradation in alkaline hydrothermal media. Methods: This article analyzes the origin of formic acid using NaH 13 CO 3 as CO 2 source, assisted by HPLC and 13 CNMR to study the origin of FA. Significant findings: Results show that two reactions pathways take place: the first one, at short reaction times, consists on the decomposition of glucose into formic acid and other by-products, achieving low yield to FA13 C, whereas longer reaction times lead to a higher yield to FA13 C. Similarly, temperature plays an important role, being 300◦C the optimal. Further reactions were performed using the main by-products of previous reactions (acetic acid, lactic acid, glycolaldehyde and glyceraldehyde) as reductants to understand the reaction mechanisms. Results indicate that the reduction pathway of NaHCO 3 take place by oxidation of the by-products rather than by the oxidation of glucose itself, needing long reaction times to achieve significant high ratios of FA13 C/ total FA. 1. Introduction Due to the utilization of fossil sources as fuels and the CO 2 levels in the atmosphere related to the anthropogenic activity, several studies have proposed different processes for both CO 2 mitigation and fossil fuels replacement, at the same time that new energy sources are developed, such as biological, photochemical and electrochemical processes [1]. One approach is the replacement of fossil fuels by renewable sources for the production of fuels and platform chemicals. Biomass (especially lignocellulosic biomass, which is composed of cellulose, hemicellulose and lignin) is a promising alternative due to these characteristics, as it is a renewable, sustainable, inexpensive and world wide spread feedstock. In the last years, different processes have been proposed to fractionate and obtain value-added products from this raw material, such as biological, enzymatic, thermochemical and hydrothermal processes [2–4]. On the other hand, the carbon capture, utilization and conversion represents an opportunity to reduce the concentration of CO 2 in the atmosphere, at the same time that this gas could be valorized by means of its utilization or conversion [5–8]. Besides the existing industrial applications of CO 2 , further conversion processes that involves renewable energies have been developed, such as photochemical, electrochemical and biological methods [9–12]. Recently, hydrothermal processes have also been found suitable to transform CO 2 into useful products such as formic acid (FA) or methanol, regarded as Liquid Organic Hydrogen Carriers (LOHC), that can be dehydrogenated in order to obtain H 2 or be directly used as feedstock for power cells, playing an important role in the hydrogen economy [13–18]. Hydrothermal method uses water at high temperature and pressure (above 100◦C and saturation pressure) In these processes, water can play * Corresponding author. E-mail address: [email protected] (´ A. Martín). Contents lists available at ScienceDirect Journal of the Taiwan Institute of Chemical Engineers journal homepage: www.journals.elsevier.com/journal-of-the-taiwan-institute-of-chemical-engineers https://doi.org/10.1016/j.jtice.2022.104504 Received 16 June 2022; Received in revised form 31 July 2022; Accepted 19 August 2022 Journal of the Taiwan Institute of Chemical Engineers 139 (2022) 104504 2 different roles in the reaction, such as green solvent, hydrogen donor and catalyst. In this way, the use of gaseous hydrogen is avoided, resulting in a safer process with lower reliance in fossil fuels (for H 2 production) [13–15]. Previous studies have shown the possibility to reduce CO 2 under hydrothermal conditions using different metals (Fe, Mn, Al or Zn) or organic compounds as reductants, being FA the main product from CO 2 reduction. Besides as LOHC, FA can be used in different industries, such as textile, chemical, agricultural and pharmaceutical industries [13,19–27]. In many of these studies, the best results are obtained when CO 2 is in form of bicarbonate, which is the main product of capturing CO 2 using NaOH solutions [28–30]. In addition, hydrothermal processes are receiving attention not only for CO 2 conversion, but also for biomass valorization, obtaining value-added chemicals by its fractionation and conversion [27,31–34]. In fact, formic acid can be also produced by means low-temperature hydrothermal processes from biomass [35–39]. As biomass is a world-wide spread, sustainable and inexpensive feedstock (sometimes even considered as a residue), the production of commodity chemicals using the hydrothermal technology could result in an inexpensive and green method, solving in this way the conundrum caused by the shortage of fossil fuels and excessive levels of CO 2 in the atmosphere. The reduction potential of different organic molecules has been previously investigated. C3 alcohol molecules such as isopropanol and glycerol were used as reductants obtaining significant yields to formic acid and concomitant production of acetone and lactic acid as oxidized products respectively [13,25]. In a previous report testing several biomass model compounds, the highest yield to formic acid was achieved using glucose as and sodium bicarbonate (NaHCO 3 ) as CO 2 source, being of high interest as glucose is the main product of the hydrothermal degradation of cellulose. As by-products, acetic and lactic among others were obtained. These latter compounds, however, are rather hydrolysis products (redox neutral reaction) since the oxidation number is the same than the starting glucose and they are also obtained when no CO 2 is present. Further exploratory investigations in the topic, using glycerol, algae and polyvinyl chloride, examined the origin of formic acid by using sodium bicarbonate 13 C (NaH 13 CO 3 ) [40–42]. The main purpose of using NaH 13 CO 3 is to detect the origin of formic acid by 13 C-NMR techniques in combination with HPLC analyses. Yang et al. tested algae and its main constituents as possible reductants of NaH 13 CO3 at 300◦C during 2 h, showing that formic acid is produced (in different proportions) from both sodium bicarbonate and the organic molecules. In the case of glucose, it was observed that a 47.9% of the total formic acid was produced by NaH 13 CO 3 reduction (without further optimization of reaction conditions), indicating a concomitant production of formic acid from both reagents under hydrothermal conditions [40]. It is then proven that bicarbonate, C(IV) undergoes a reduction to formic acid, C (II), but the mechanism is still unclear because no oxidation by-products have been detected in significant amounts. In this work, NaH 13 CO 3 is used as CO 2 source to study in detail the mechanistic effects of CO 2 reduction with glucose (a representative monosaccharide of carbohydrates and component of lignocellulosic biomass). The origin of FA is determined using marked NaH 13 CO 3 as inorganic carbon source, varying the operational conditions to determine the plausible mechanisms through which glucose hydrolyses and/ or reduces bicarbonate. To do this, not only glucose was used as reductant as cellulose, but also its hydrolysis products, such as lactic acid, acetic acid, glyceraldehyde and glycolaldehyde. 2. Materials and methods 2.1. Materials As reagents, NaH 13 CO 3 (99%, 98 atom % 13 C), D-(+)-glucose (99%), glyceraldehyde (90%) and glycolaldehyde dimer (99%) were used as reductants and standards, and they were obtained from Sigma Aldrich. For HPLC standards, the following chemicals where used: glycerol (99.5%, Spain), n-propanol (>99.7%), pyruvaldehyde (40%), 5-HMF (99%), furfural (99%), fructose (99%), rahmnose (99%), resorcinol (99%), glucuronic acid (>98%), oxalic acid (≥99%), acrylic acid (99%), d-glyceric acid calcium salt (99%), catechol (99%) and guaiacol (>99%) were bought to Sigma Aldrich. Ethanol (EtOH, 99.5%), acetone (99.5%), isopropanol (iPrOH, 99.9%), formaldehyde stabilized in MeOH (37-38% w/w) and propanoic acid (99%) were acquired from Panreac. Formic acid (98%) and galacturonic acid (≥97%) were purchased from Fluka. Methanol (MeOH, 99.99%) was obtained from Fisher Scientific. Ethylenglycol (≥99.5%) was obtained from Merck. All the chemicals were used without further treatment. For 13 C-NMR analysis, acetic acid-113 C (99%, 99 atom % 13 C) and formic acid13 C (95 wt% in H 2 O, 99 atom % 13 C) were purchased from Sigma Aldrich. Deuterated solvent (D 2 O, 99.5% D atom) were purchased from Cortecnet 2.2. Experimental procedure Solutions in MilliQ water (ultrapure water) were prepared using a concentration 0.50 M of NaH 13 CO 3 , and 0.05 M of glucose, unless otherwise indicated. Experiments were performed in batch reactors, described elsewhere [28]. These reactors consist of tubes with a length of 12 cm; o.d. of ½” and 1.24 mm of thickness, made of SS 316 stainless steel with an internal volume of 15.6 mL. The solution was loaded in the reactor up to the 50% of the volume, and the reactor was sealed. Afterwards, the reactor was introduced in a preheated fluidized alumina bed, which allows a fast heating of the solution, reaching the temperature set point inside the reactor (300◦C) in less than 3 minutes. Considering the introduction of the reactors in the fluidized alumina as the initial time (t =0 min), after the desired reaction time, the reactor was cooled down using an ice bath. Once cooled, the reactor was opened, and liquid products were collected. Reactions were performed at least twice to ensure reproducibility. Reactors were weighted before and after the reaction to check the correct sealing. 2.3. Analytical procedure Liquid samples were collected and filtered through a 0.22 µm Nylon filter. Two different techniques were used to analyse and quantify the products distribution: high performance liquid chromatography (HPLC) and nuclear magnetic resonance (NMR). HPLC analyses were carried out in a Waters Alliance separation module e2695, using a Rezex ROAOrganic Acid H+(8%) LC column (300 ×7.8mm, Phenomenex) and a refractive index detector. The column temperature was kept at 40◦C and the detector temperature at 30◦C. The mobile phase was a 25 mM H 2 SO 4 solution in MilliQ water, with a flow rate 0.5 mL/min under isocratic conditions. Quantitative 13 C{ 1 H}-NMR analyses were performed using a 500 MHz Agilent instrument equipped with OneNMR probe in the Laboratory of Instrumental Techniques (LTI) Research Facilities, University of Valladolid. For the quantification of 13 C products, two different analyses of the sample were performed: one sample without further treatment and another adding acetic acid-113 C (CH 3 13 COOH) as internal standard. In all cases, 600 µL of sample was transferred to a NRM tube and 50 µL of D 2 O for internal lock reference were added. Samples analysed with CH 3 13 COOH as the internal standard were prepared as follows: 2.5 mL of filtered sample were spiked with 5 µL of CH 3 13 COOH, achieving a concentration of 34 mM CH 3 13 COOH, and then the mix was shaken in a vortex for 20 s to ensure the correct mixing. The acquisition parameters for quantitative measurements were: 25 ◦C, 70 s relaxation delay between transients, 45◦pulse width, and spectral width of 31250.0 Hz, a total of 32 transients and 2.044 s acquisition time. In order to obtain a decoupled spectrum without NOE enhancements, the inverse-gated proton decoupling method was used. All the studies of 13 C{ 1 H} experiments were recorded on 500 MHz Agilent instruments equipped with OneNMR probe, and 13 C { 1 H} chemical shifts (δ) were reported in part M. And´ erez-Fern´ andez et al. Journal of the Taiwan Institute of Chemical Engineers 139 (2022) 104504 3 per million (ppm) and referenced to tetramethylsilane (TMS). The final spectra were manipulated and processed using Mestrelab Research software (MNova 12.0) (Santiago de Compostela, Spain). Prior to NMR analysis, 13 C{ 1 H} data obtained were subsequently apodized with line broadening window function of 1 Hz and a value of 64K zero-filling was applied to improve the digital resolution of the final spectrum. Yields reported in this work are calculated with respect to the organic reductant, which as indicated in Section 2.2 is the limiting reagent. Yields with respect to the inorganic carbon source NaHCO 3 can be obtained from the reported yields considering the NaHCO 3 :organic molar ratio, e.g. dividing by 10 for a concentration 0.50 M of NaH 13 CO 3 and 0.05 M of glucose. The yield to total formic acid (total FA) and the rest of by-products was calculated as follows (Eq. 1): Yproduct,HPLC =CProduct,f,HPLC Corganic,i ×100 (Eq. 1) Where C product,f,HPLC is the molar concentration of the different products after the reaction, determined by HPLC, and C organic,i is the initial molar concentration of the reductant. The yield to formic acid obtained from NaH 13 CO 3 was determined as shown in Eq. (2): YFA,NMR =CFA,f,NMR Corganic,i x100 (Eq. 2) Where C FA,f,NMR is the molar concentration of formic acid 13 C after the reaction, determined by 13 C-NMR analysis, and C organic,i is the initial molar concentration of the reductant. The comparative results between HPLC and NMR techniques are displayed in the Supporting Information (Section A1). All the experimental data presented in this work have a relative standard deviation (RSD) lower than 10%. 3. Results and discussion In order to test the validity of the RMN method, a primary set of reactions were performed using NaH 12 CO 3 and NaH 13 CO 3 as CO 2 sources and glucose as reductant at 300◦C for 120 min (Fig. S1). It can be observed that, in the reaction using NaH 12 CO 3 , no signals in the 13 C { 1 H}-NMR spectrum are detected, as expected for the low natural abundance of 13 C(1.1%) in the 13 C-NMR studies recording under quantitative conditions. However, reactions with NaH 13 CO 3 resulted in the production of formic acid 13 C (FA13 C), which is observed at 173 ppm at reasonable acquisition times in the quantitative 13 C-NMR spectra. This demonstrates that CO 2 can be transformed into FA using glucose (or one of its hydrolysis products) as reductant. The 13 C{ 1 H} -NMR spectra of experiments carried out using NaH 13 CO 3 as CO 2 source showed other significant peaks: The intense peak observed at 163 ppm is assigned to the unreacted NaH 13 CO 3 that remained in the solution after the reaction. Due to the equilibrium of the unreacted NaH 13 CO 3 in water, another weak signal is observed at 123 ppm, attributed to dissolved 13 CO 2 [43,44]. 3.1. Effect of time in the origin to formic acid The effect of time on NaH 13 CO 3 reduction using glucose as reductant was tested, as depicted in Fig. 1. As observed in this Fig., the yield to FA shows a complex behavior: it is high at short reaction times (10 min), decreases at intermediate times of 10-120 min and increases again after 120 min. Correspondingly, the yield to FA13 C is very low at short reaction times at increases in the time interval 10 min – 90 min. The highest yield to total FA (47.5%) was achieved after 10 min of reaction, but also the lowest yield to FA13 C (5.9%). Therefore, at short reaction times, a rapid FA production is mainly due to glucose decomposition. However, as the reaction proceeds the proportion of FA13 C increases, suggesting that the FA initially formed is being partially replaced by FA from bicarbonate. To explain these observations, it is necessary to assume the existence of reversible reactions in the reaction mechanism. As the yield to total FA decreased at reaction times between 30 and 90 min, FA must partially decompose under the reaction conditions. FA13 C concentration increases to a plateau after 30 min, as well as the proportion FA13 C/FA. At 120 min the reaction seems to undergo into a new stage, as total FA yield increases again. Besides FA, other products were identified after the reaction (Fig. S2), which were not detected in the 13 C-NMR spectra, indicating that were produced by the glucose degradation as they are not enriched in 13 C. No or negligible amount of glucose or fructose were detected in any of the samples. The main by-products obtained were acetic acid and lactic acid. The yield former one remains nearly constant along the whole reaction Lactic acid formation is also fast, reaching a maximum yield of 43% but starting from 120 min, it decreases, apparently being converted into glycolaldehyde (and probably FA). Acetic acid yield achieves values of ca. 40% in the first minutes of the reaction and remains almost constant throughout time, which suggests that is produced at the very early stages of the reaction by non-reversible reactions. In the case of lactic acid, its concentration increases very fast in the first 30 min of reaction to achieve a maximum yield of 43.1% at 90 min of reaction, decreasing afterwards, and apparently being converted into glycolaldehyde (plus probably FA). Glyceraldehyde was also detected by HPLC analysis, and its yield decreases with time to very small proportions. 3.2. Effect of reaction temperature on yield to formic acid Fig. 2 depicts the yield to total FA and FA13 C, and the ratio FA13 C/ total FA as a function of reaction temperature, observing that formic acid was produced at all the temperatures tested. However, at 200◦C, all the FA produced was a result of glucose decomposition, as no FA13 C was detected by 13 C-NMR analysis. As temperature increases, both the total yield to FA and the proportion FA13 C/total FA increases, the latter significantly. At 325◦C however, the total yield to FA decreases again. Despite at this temperature the ratio FA13 C/total FA was the highest at the temperatures tested, reaching a ratio FA13 C/total FA of 55.0%, this improvement is counterbalanced by the decrease in yield. This behaviour is consistent with a set of equilibrium reaction either being shifted to FA13 C (bicarbonate reduction) by effect of temperature of by speeding up their equilibrium kinetics. However, if the temperature is too high, decomposition reaction of FA to CO 2 and H 2 start getting importance [45]. The main by-products were acetic acid and lactic acid followed by glyceraldehyde and glycolaldehyde at all temperatures (Fig. S3). Fig. 1. Effect of reaction time on the yield to total FA and FA13 C. Legend: (⬛): yield to total FA; (⬜): yield to FA13 C; (◆) ratio FA13 C. General reaction conditions: 0.50 M NaH 13 CO 3 ; 0.05 M glucose; T:300◦C, t: 120 min. M. And´ erez-Fern´ andez et al. Journal of the Taiwan Institute of Chemical Engineers 139 (2022) 104504 4 However, the trend of each by-product differs. Acetic and lactic acids yield slightly increased when the reaction temperature was higher, showing the results with lactic acid a peak at 250◦C that can be attributed to experimental uncertainty. Interestingly, the yield to glyceraldehyde followed a trend opposite to total FA production. The yield to glycolaldehyde is greatly affected by temperature, when temperature is higher than 300◦C it decreases ca. six-fold. 3.3. Effect of initial concentration of glucose on yield to formic acid The influence of the initial concentration of glucose was tested for values between 0.025 and 0.5 M (Fig. 3). At low glucose concentrations of 0.025 M – 0.05 M, the performance of the reaction does not significantly vary. However, at higher concentrations of 0.05 M – 0.5 M, the higher the initial concentration of glucose, the lower the yield to total formic acid. The proportion FA13 C/total FA also decreases with the initial concentration of glucose but at a lower rate. Similar to the results presented in previous sections, the main by-products were lactic acid and acetic acid, followed by glyceraldehyde and glycolaldehyde (Fig. S4). In all cases, the higher the initial concentration of glucose resulted in the lower yield to by-products, reaching the maximum yield to these products using an initial concentration of glucose of 0.025 M and following a similar trend than the yield to formic acid. These results suggest that an excess of glucose has a detrimental effect on the production of formic acid 3.4. Reduction of NaH 13 CO 3 using degradation products from glucose In order to obtain a better understanding of the reaction mechanism and the substance that acts as direct reductant of NaH 13 CO 3 , experiments were carried out using major by-products of glucose degradation, namely acetic acid, lactic acid, glycolaldehyde and glyceraldehyde. In the case of acetic acid, no formic acid was observing neither in HPLC nor 13 C-NMR analysis, detecting only unreacted acetic acid and NaH 13 CO 3 in the analyses. This molecule is stable enough to be produced after a non-reversible reaction of hydrolysis. On the other hand, formic acid was obtained when using lactic acid, glycolaldehyde and glyceraldehyde as reductant (Fig. 4). Employing the by-products as reductants, the maximum yield to total FA (26.1%) was obtained using glyceraldehyde after 10 min of reaction, followed by glycolaldehyde (21.8%, 60 min) and lactic acid (13.8%, 180 min). Similar to glucose, longer reaction times led to a higher yield to FA13 C in all cases. It was also observed that the formic acid produced by the degradation of organic molecules followed a different production pathway with lactic acid than with glycolaldehyde and glyceraldehyde. The formation of FA from the lactic acid is slower than from the other two. While glycolaldehyde and glyceraldehyde are converted almost totally at all reaction times, ca. 82% of initial lactic acid remained unconverted after 180 min of reaction. It should be noted that the yield to FA reaches a maximum to decrease later with reaction time when glycolaldehyde and glyceraldehyde are used as reductants. Lactic acid then is a relatively stable hydrolysis product as it has been detected as a typical product from biomass hydrolysis, however, it can further proceed to FA at lower extent. The evolution of by-products as a function of reaction time for the three different organic model compounds are shown in the Fig. S5. In the case of lactic acid, acetic acid and glyceraldehyde were detected as the main by-products, indicating a related production mechanism. When using glycolaldehyde or glyceraldehyde, the main by-products obtained were also lactic acid and acetic acid. They are produced in high yield at the first moments of the reaction and its proportion does not change much at longer times. It should be noted that lactic acid (a C3 molecule) was produced from glycolaldehyde, a C2 aldose. The production lactic acid from glycolaldehyde in alkaline hydrothermal reaction has been reported by Kishida et al. [46]. Glycolaldehyde seems to be an important intermediate in the hydrothermal glucose conversion. Similar mechanisms may be happening in this research, as the media is also alkaline caused by NaHCO 3 . Glyceraldehyde (an intermediate of the proposed mechanism) was not detected as by-product in any case, indicating that this is a very reactive intermediate. It is also noticing that higher yields to acetic and lactic acid were obtained using glyceraldehyde, compared to those achieved with glycolaldehyde. The ratio FA13 C / total FA of these reactions are shown in Table 1. In all cases, the same trend can be observed: longer reaction times resulted in a higher ratio; that is, there is an increasing replacement of FA generated directly from the reductant by FA from bicarbonate. Again, this strongly suggest a mechanism of reversible reactions. In the case of lactic acid, the ratio FA13 C/ total FA increases faster for reaction times up to 60 min. compared to the aldoses, probably because LA reacts more slowly to FA. As reaction time is longer, a major proportion of FA13 C is produced out of the total FA detected by HPLC analysis, reaching the maximum ratio FA13 C /total FA of 76.5% and 80.5% after 180 min of reaction using glycolaldehyde and glyceraldehyde as reductant. Summarizing, these results indicate that: a) the intermediate products tested behave very similarly than glucose (except acetic acid, which is stable and lactic acid which reacts more slowly); b) the organic molecules quickly hydrolyse to produce FA (and other by-products); c) reduction of NaHCO 3 to formic acid requires of longer reaction times; d) the FA formed from reduction replaces that formed by hydrolysis by a mechanism of reversible reactions. Fig. 2. Effect of temperature on the yield to total FA and FA13 C. Legend: (⬛): yield to total FA; (⬜): yield to FA13 C; (◆) ratio FA13 C; (◆): ratio FA13 C. General reaction conditions: 0.50 M NaH 13 CO 3 ; 0.05 M glucose; t: 120 min. Fig. 3. Effect of initial glucose concentration on the yield to total FA and FA13 C. Legend: (⬛): yield to total FA; (⬜): yield to FA13 C; (◆) ratio FA13 C; (◆): ratio FA13 C. General reaction conditions: 0.50 M NaH 13 CO 3 ; T:300◦C, t: 120 min. M. And´ erez-Fern´ andez et al. Journal of the Taiwan Institute of Chemical Engineers 139 (2022) 104504 5 3.5. Proposed mechanism The results described in the previous sections have been used to elucidate a mechanism of the reaction, based on the mechanism proposed in literature about the hydrothermal conversion of glucose and other organic compounds [47,48]. Fig. 5 depicts the plausible routes of glucose conversion under hydrothermal conditions (black) and the proposed mechanism for the obtention of formic acid by simultaneous conversion of CO 2 and glucose/glucose derivatives (green). Under hydrothermal conditions, glucose may isomerize into fructose or be dehydrated to anhydroglucose [33]. The addition of alkali to the hydrothermal media results the in isomerisation of glucose into fructose and inhibiting its dehydration into anhydroglucose [37,49]. Both glucose and fructose can further undergo a retro-aldol condensation, resulting in three molecules of glycolaldehyde and two molecules of glyceraldehyde, respectively [33,50]. Glyceraldehyde may undergo reversible isomerization to dihydroxyacetone. Both can be dehydrated to pyruvaldehyde, which further can be converted into lactic acid by benzylic acid rearrangement [51,52]. Glyceraldehyde can also undergo reverse aldol condensation, obtaining glycolaldehyde, acetic acid and formaldehyde [53–56]. Besides its conversion into acetic acid, glycolaldehyde can reach significant yields to lactic acid under alkaline conditions, involving a series of different reactions, such as, aldol condensation, dehydration, keto-enol tautomerization, reverse aldol condensation and benzylic acid rearrangement, being glyceraldehyde one of the reaction intermediates [46,57]. Formaldehyde is obtained from the degradation of glyceraldehyde. Thus, the different reactions aforementioned about the glucose hydrothermal degradation and the results obtained in the previous sections about the determination of formic acid’s origin could contribute to deduce a plausible reaction mechanism. According to the data obtained, part of the formic acid produced in the reaction and detected by HPLC techniques resulted from Fig. 4. Effect of reaction time on yield to total FA and FA13 C using lactic acid (a), glycolaldehyde (b) and glyceraldehyde (c) as reductant. Symbols: (▴): total yield to formic acid, determined by HPLC; (△): yield to formic acid 13 C, determined by 13 C-NMR analysis. Reaction conditions: 0.05 M organic model compound, 0.50 M NaH 13 CO 3 , 50% filling, T: 300◦C. Table 1 Ratio FA13 C / total FA (%) as a function of reaction time using lactic acid, glycolaldehyde and glyceraldehyde as reductants. Reaction conditions: 0.05 M reductant, 0.50 M NaH 13 CO 3 , 50% filling, T: 300◦C % FA13 C /total FA t (min) Lactic ac. Glycolaldehyde Glyceraldehyde 10 35.5 25.8 27.4 60 58.2 34.5 34.7 120 65.6 69.6 62.9 180 74.2 76.5 80.5 Fig. 5. Proposed mechanism for NaHCO 3 reduction using glucose as reductant. Black data were obtained from previous studies and green ones have been proposed using the data obtained in this work. M. And´ erez-Fern´ andez et al. Journal of the Taiwan Institute of Chemical Engineers 139 (2022) 104504 6 glucose and glucose derivatives degradation, especially at short reaction times (10-60 min), as the different aforementioned reactions take place within minutes, obtaining also glycolaldehyde and glyceraldehyde as by-products [35–39]. On the other hand, the formic acid produced from NaH 13 CO 3 reduction, which is detected by 13 C-NMR analysis, requires of longer times (>60 min), being reduced by the by-products (especially lactic acid, glycolaldehyde and glyceraldehyde) produced by the degradation of glucose in the first minutes of reaction, obtaining mainly acetic acid as by-product. Although reduction of bicarbonate is a fact, none of the C2 or C3 intermediates mentioned so far seems to be the direct reductant of the C (IV) species because it has not been detected the corresponding oxidized molecule. Instead of that, we propose that formaldehyde, which can be produced by degradation of glyceraldehyde and glycolaldehyde, would be the reductant of NaHCO 3 to formic acid. Interestingly, the oxidized product is also be formic acid, so the overall reaction would be a redox comproportionation of C(0), formaldehyde and C(IV), bicarbonate to yield C(II), formate. Formaldehyde is not detected by HPLC in our reaction products but that can be because it reacted effectively with water or because its free form is present in aqueous solutions in low quantities [58]. The 13 C-NMR analysis indicate that these reactions must be reversible. The detailed comproportionation mechanism cannot be determined with the experimental information available, but a plausible possibility is that it proceeds by H 2 as intermediate. H 2 can be released to solution to subsequently react with bicarbonate. The reduction of C(IV) with gas hydrogen has been widely reported [59]. Formaldehyde is known to thermally decompose into CO and H 2 at high temperatures. However, in aqueous media and moderate temperature, it can undergo decomposition via two mechanisms, a Cannizzaro reaction yielding formate and methanol or a dehydrogenation yielding formate and H 2 [58]. Both mechanisms are base-catalyzed, but since no methanol is detected in our tests, we can propose that in the latter the one that plays a role in this process. H 2 is partially soluble in high-temperature water so once released it can participate in the reaction bicarbonate (avoiding interphase mass transfer limitations). A proposed scheme of reaction is: H2CO +H2O→HCOOH +H2(1) HCOOH⇄DCO2+H2(2) The observations are consistent with this reaction mechanism, at the first stages of the reaction, FA is directly produced from formaldehyde by dehydrogenation (reaction 1). Reversible reaction (2) explains how the 13 C-FA replaces FA as time proceeds. The addition of reaction (1) and reaction (2) reversed results in the comproportionation reaction proposed (reaction (3)). H2CO +CO2+H2O→2HCOOH (3) Finally, in the last stages of the reaction or for high temperatures, decomposition of FA and formaldehyde into CO, CO 2 and H 2 happens to an extent enough to release these gases to the vapor phase thus decreasing FA total yield. In summary, simultaneous hydrolysis of glucose and oxidation of bicarbonate can be achieved and we propose that these reactions proceed by a mechanism composed by a set of reversible reactions. The products obtained in highest be obtained directly from glucose hydrolysis products or by a comproportionation reaction between these and C (IV) added to medium. The importance of FA has already been discussed. The generation of the other two by-products seems difficult to avoid, but that is not a big drawback as they have also interesting applications: lactic acid for example is the monomer of the renewable polymer polylactic acid; acetic acid is an important chemical used in many industrial processes such as terephthalic acid or cellulose acetate production. 4. Conclusions In this research, the hydrothermal reduction of NaHCO 3 using glucose as reductant was achieved, studying a wide range of operational conditions that contributed to understand the reaction mechanism, concerning the use of glucose as reductant. As glucose may also decompose into formic acid under hydrothermal conditions, NaH 13 CO 3 was used as CO 2 source, being thus possible to differentiate the formic acid produced by NaHCO 3 reduction from that obtained from glucose conversion by means of 13 C-NMR analysis. Two different pathways for the production of FA were observed: at short reaction times, the FA detected was produced by the degradation of glucose, whereas at long reaction times, FA was mainly produced by the reduction of NaHCO 3 under hydrothermal conditions, reaching the maximum yield of 20.7 % mol FA13 C/ mol glucose (yield to total FA of 41.1%) after 120 min. Further experiments using the by-products detected in reactions with glucose as reductants have contributed to shed a light on the plausible reaction routes, indicating that glucose degradation products are responsible of NaHCO 3 reduction to formic acid at long reaction times and that reaction mechanisms proceed through a set of equilibrium reactions in which formaldehyde is hypothesized to be a key reaction intermediate. Declaration of Competing Interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper. Acknowledgements Authors would like to acknowledge the contribution of Dr. Laura Quintana and laboratory technicians Isabel Rodríguez and Daniel Fern´ andez in the construction and commissioning of the experimental set-up. This work was supported by the Regional Government of Castilla y Le´ on and the EU-FEDER program (CLU-2019-04) and by the Ministry of Science and Universities through project RTI2018-097456-B-I00. C. M. A acknowledges the Spanish Ministry of Science, Innovation and Universities (MCIU) for funding the project number PGC2018-099470B-I00 Supplementary materials Supplementary material associated with this article can be found, in the online version, at doi:10.1016/j.jtice.2022.104504. References [1] Abas N, Kalair A, Khan N. Review of fossil fuels and future energy technologies. Futures 2015;69:31–49. https://doi.org/10.1016/j.futures.2015.03.003. [2] Song B, Lin R, Ho C, Wu H, Tsui T, Yu Y. Recent advances and challenges of interdisciplinary biomass valorization by integrating hydrothermal and biological techniques. Renew Sustain Energy Rev 2021;135:110370. https://doi.org/ 10.1016/j.rser.2020.110370. [3] Halim SA, Mohd NA, Razali NA. A comparative assessment of biofuel products from rice husk and oil palm empty fruit bunch obtained from conventional and microwave pyrolysis. J Taiwan Inst Chem Eng 2022;134:104305. https://doi.org/ 10.1016/j.jtice.2022.104305. [4] Nosakhare A, Okoye PU, Gunes B, Al HTL. Waste biomass valorization for the production of biofuels and valueadded products: A comprehensive review of thermochemical, biological and integrated processes. Process Saf Environ Prot 2022;159:323–44. https://doi.org/10.1016/j.psep.2021.12.049. [5] Hung YC, Hsieh CM, Machida H, Lin ST, Shimoyama Y. Modeling of phase separation solvent for CO2 capture using COSMO-SAC model. J Taiwan Inst Chem Eng 2022;135:104362. https://doi.org/10.1016/j.jtice.2022.104362. [6] Liu J, Wong DSH, Chen DS. Energy-saving performance of the process modifications for carbon capture by diluted aqueous ammonia. J Taiwan Inst Chem Eng 2022;130:103966. https://doi.org/10.1016/j.jtice.2021.06.060. [7] Aresta M, Dibenedetto A. Utilisation of CO 2 as a chemical feedstock: Opportunities and challenges. J Chem Soc Dalt Trans 2007:2975–92. https://doi.org/10.1039/ b700658f. [8] Leung DYC, Caramanna G, Maroto-Valer MM. An overview of current status of carbon dioxide capture and storage technologies. Renew Sustain Energy Rev 2014; 39:426–43. https://doi.org/10.1016/j.rser.2014.07.093. M. And´ erez-Fern´ andez et al. Journal of the Taiwan Institute of Chemical Engineers 139 (2022) 104504 7 [9] Anwar MN, Fayyaz A, Sohail NF, Khokhar MF, Baqar M, Yasar A, et al. CO 2 utilization: Turning greenhouse gas into fuels and valuable products. J Environ Manage 2020;260:110059. https://doi.org/10.1016/j.jenvman.2019.110059. [10] Yaashikaa PR, Senthil Kumar P, Varjani SJ, Saravanan A. A review on photochemical, biochemical and electrochemical transformation of CO 2 into valueadded products. J CO2 Util 2019;33:131–47. https://doi.org/10.1016/j. jcou.2019.05.017. [11] Ko TE, Hosseini S, Tseng CM, Tsai JE, Wang WH, Li YY. Tetrafunctional electrocatalyst for oxygen reduction, oxygen evolution, hydrogen evolution and carbon dioxide reduction reactions. J Taiwan Inst Chem Eng 2022;136:104397. https://doi.org/10.1016/j.jtice.2022.104397. [12] Nazeri MT, Siamak J, Maryam R, Shaabani A. A facile and green synthesis of cobalt phtalocyanine-conjugated multiwall carbon nanotube by the Ugi reaction: as an efficient CO 2 fixation catalyst. J Taiwan Inst Chem Eng 2022;136:104428. https:// doi.org/10.1016/j.jtice.2022.104428. [13] Shen Z, Zhang Y, Jin F. From NaHCO3 into formate and from isopropanol into acetone: Hydrogen-transfer reduction of NaHCO3 with isopropanol in hightemperature water. Green Chem 2011;13:820. https://doi.org/10.1039/ c0gc00627k. [14] Jin F, Zeng X, Liu J, Jin Y, Wang L, Zhong H, et al. Highly efficient and autocatalytic H 2 O dissociation for CO 2 reduction into formic acid with zinc. Sci Rep 2014;4:1–8. https://doi.org/10.1038/srep04503. [15] Gomez LQ, Shehab AK, Al-Shathr A, Ingram W, Konstantinova M, Cumming D. H2 -free synthesis of aromatic, cyclic and linear oxygenates from CO2. ChemSusChem 2020;13:647–58. https://doi.org/10.1002/cssc.201902340. [16] Enthaler S. Carbon dioxide\-the hydrogen-storage material of the future? ChemSusChem 2008;1:801–4. https://doi.org/10.1002/cssc.200800101. [17] Eppinger J, Huang K-WW. Formic acid as a hydrogen energy carrier. ACS Energy Lett 2017;2:188–95. https://doi.org/10.1021/acsenergylett.6b00574. [18] Bulushev DA, Ross JRH. Towards sustainable production of formic acid. ChemSusChem 2018;11:821–36. https://doi.org/10.1002/cssc.201702075. [19] Berndt ME, Allen DE, Seyfried WE. Reduction of CO 2 during serpentinization of olivine at 300◦C and 500 bar. Geology 1996;24:351–4. https://doi.org/10.1130/ 0091-7613(1996)024<0351:ROCDSO>2.3.CO;2. [20] Huo Z, Hu M, Zeng X, Yun J, Jin F. Catalytic reduction of carbon dioxide into methanol over copper under hydrothermal conditions. Catal Today 2012;194: 25–9. https://doi.org/10.1016/j.cattod.2012.06.013. [21] Jin F, Gao Y, Jin Y, Zhang Y, Cao J, Wei Z, et al. High-yield reduction of carbon dioxide into formic acid by zero-valent metal/metal oxide redox cycles. Energy Environ Sci 2011;4:881. https://doi.org/10.1039/c0ee00661k. [22] Roman-Gonzalez D, Moro A, Burgoa F, P´ erez E, Nieto-M´ arquez A, Martín ´ A, Bermejo MD. Hydrothermal CO 2 conversion using zinc as reductant: Batch reaction, modeling and parametric analysisof the process. J Supercrit Fluids 2018; 140:320–8. https://doi.org/10.1016/j.supflu.2018.07.003. [23] Ueda H, Sawaki Y, Maruyama S. Reactions between olivine and CO 2 -rich seawater at 300◦C: Implications for H 2 generation and CO 2 sequestration on the early Earth. Geosci Front 2017;8:387–96. https://doi.org/10.1016/j.gsf.2016.10.002. [24] Zeng X, Hatakeyama M, Ogata K, Liu J, Wang Y, Gao Q, et al. New insights into highly efficient reduction of CO 2 to formic acid by using zinc under mild hydrothermal conditions: A joint experimental and theoretical study. Phys Chem Chem Phys 2014;16:19836–40. https://doi.org/10.1039/c4cp03388d. [25] Shen Z, Zhang Y, Jin F. The alcohol-mediated reduction of CO 2 and NaHCO 3 into formate: a hydrogen transfer reduction of NaHCO 3 with glycerine under alkaline hydrothermal conditions. RSC Adv 2012;2:797–801. https://doi.org/10.1039/ C1RA00886B. [26] Yang Y, Zhong H, Yao G, He R, Jin B, Jin F. Hydrothermal reduction of NaHCO 3 into formate with hexanehexol. Catal Today 2018;318:10–4. https://doi.org/ 10.1016/j.cattod.2017.09.005. [27] Su J, Yang L, Yang X, Lu M, Luo B, Lin H. Simultaneously converting carbonate/ bicarbonate and biomass to value-added carboxylic acid salts by aqueous-phase hydrogen transfer. ACS Sustain Chem Eng 2015;3:195–203. https://doi.org/ 10.1021/sc5007158. [28] And´ erez-Fern´ andez M, P´ erez E, Martín A, Bermejo MD. Hydrothermal CO 2 reduction using biomass derivatives as reductants. J Supercrit Fluids 2018;133: 658–64. https://doi.org/10.1016/j.supflu.2017.10.010. [29] Lackner KS. Capture of carbon dioxide from ambient air. Eur Phys J Spec Top 2009; 176:93–106. https://doi.org/10.1140/epjst/e2009-01150-3. [30] Park S, Song K, Jo H. Laboratory-scale experiment on a novel mineralization-based method of CO2 capture using alkaline solution. Energy 2017;124:589–98. https:// doi.org/10.1016/j.energy.2017.02.100. [31] Demirel Y, Matzen M, Winters C, Gao X. Capturing and using CO 2 as feedstock with chemical looping and hydrothermal technologies. Int J Energy Res 2015;39: 1011–47. https://doi.org/10.1002/er.3277. [32] Yao G, Duo J, Jin B, Zhong H, Lyu L, Ma Z. Highly-efficient and autocatalytic reduction of NaHCO 3 into formate by in situ hydrogen from water splitting with metal/metal oxide redox cycle. J Energy Chem 2017;26:881–90. https://doi.org/ 10.1016/j.jechem.2017.08.011. [33] Cantero DA, ´ Alvarez A, Bermejo MD, Cocero MJ. Transformation of glucose into added value compounds in a hydrothermal reaction media. J Supercrit Fluids 2015;98:204–10. https://doi.org/10.1016/j.supflu.2014.12.015. [34] And´ erez Fern´ andez M, Rissanen J, P´ erez Nebreda A, Xu C, Willf¨ or S, García Serna J. Hemicelluloses from stone pine, holm oak, and Norway spruce with subcritical water extraction −comparative study with characterization and kinetics. J Supercrit Fluids 2018;133:647–57. https://doi.org/10.1016/j. supflu.2017.07.001. [35] Kamm B. Production of platform chemicals and synthesis gas from biomass. Angew Chemie Int Ed 2007;46:5056–8. https://doi.org/10.1002/anie.200604514. [36] Jin F, Zhou Z, Kishita A, Enomoto H, Kishida H, Moriya T. A new hydrothermal process for producing acetic acid from biomass waste. Chem Eng Res Des 2007;85: 201–6. https://doi.org/10.1205/cherd06020. [37] Jin F, Enomoto H. Rapid and highly selective conversion of biomass into valueadded products in hydrothermal conditions : chemistry of acid /base-catalysed and oxidation reactions 2011:382–97. https://doi.org/10.1039/c004268d. [38] Kneˇ zevi´ c D, van Swaaij WPM, Kersten SRA. Hydrothermal conversion of biomass I, Glucose conversion. Ind Eng Chem Res 2009;48:4731–43. [39] Kruse A, Funke A, Titirici M. Hydrothermal conversion of biomass to fuels and energetic materials. Curr Opin Chem Biol 2013;17:515–21. https://doi.org/ 10.1016/j.cbpa.2013.05.004. [40] Yang Y, Zhong H, He R, Wang X, Cheng J, Yao G, et al. Synergetic conversion of microalgae and CO 2 into value-added chemicals under hydrothermal conditions. Green Chem 2019;21:1247–52. https://doi.org/10.1039/c8gc03645d. [41] Lu L, Zhong H, Wang T, Wu J, Jin F, Yoshioka T. A new strategy for CO 2 utilization with waste plastics: Conversion of hydrogen carbonate into formate using polyvinyl chloride in water. Green Chem 2020;22:352–8. https://doi.org/ 10.1039/c9gc02484k. [42] Shen Z, Gu M, Zhang M, Sang W, Zhou X, Zhang Y, et al. The mechanism for production of abiogenic formate from CO 2 and lactate from glycerine: Uncatalyzed transfer hydrogenation of CO 2 with glycerine under alkaline hydrothermal conditions. RSC Adv 2014;4:15256–63. https://doi.org/10.1039/c4ra00777h. [43] Moret S, Dyson PJ, Laurenczy G. Direct, in situ determination of pH and solute concentrations in formic acid dehydrogenation and CO 2 hydrogenation in pressurised aqueous solutions using 1H and 13C NMR spectroscopy. Dalt Trans 2013;42:4353–6. https://doi.org/10.1039/c3dt00081h. [44] Barzagli F, Lai S, Mani F. CO 2 capture by liquid solvents and their regeneration by thermal decomposition of the solid carbonated derivatives. Chem Eng Technol 2013;36:1847–52. https://doi.org/10.1002/ceat.201300225. [45] Yu J, Savage PE. Decomposition of formic acid under hydrothermal conditions. Ind Eng Chem Res 1998;37:2–10. https://doi.org/10.1021/ie970182e. [46] Kishida H, Jin F, Yan X, Moriya T, Enomoto H. Formation of lactic acid from glycolaldehyde by alkaline hydrothermal reaction. Carbohydr Res 2006;341: 2619–23. https://doi.org/10.1016/j.carres.2006.06.013. [47] Cantero DA, Dolores Bermejo M, Jos´ e Cocero M. Reaction engineering for process intensification of supercritical water biomass refining. J Supercrit Fluids 2015;96: 21–35. https://doi.org/10.1016/j.supflu.2014.07.003. [48] Cantero DA, Bermejo MD, Cocero MJ. Kinetic analysis of cellulose depolymerization reactions in near critical water. J Supercrit Fluids 2013;75: 48–57. https://doi.org/10.1016/j.supflu.2012.12.013. [49] Watanabe M, Aizawa Y, Iida T, Aida TM, Levy C, Sue K, et al. Glucose reactions with acid and base catalysts in hot compressed water at 473 K. Carbohydr Res 2005;340:1925–30. https://doi.org/10.1016/j.carres.2005.06.017. [50] Gao P, Li G, Yang F, Lv X, Fan H, Meng L. Preparation of lactic acid, formic acid and acetic acid from cotton cellulose by the alkaline pre-treatment and hydrothermal degradation. Ind Crop Prod 2013;48:61–7. https://doi.org/10.1016/j. indcrop.2013.04.002. [51] Kabyemela BM, Adschiri T, Malaluan RM, Arai K. Glucose and fructose decomposition in subcritical and supercritical water : Detailed reaction pathway, Mechanisms, and Kinetics 1999:2888–95. [52] Kishida H, Jin F, Zhou Z, Moriya T, Enomoto H. Conversion of glycerin into lactic acid by alkaline hydrothermal reaction. Chem Lett 2005;34:1560–1. https://doi. org/10.1246/cl.2005.1560. [53] Ramírez-L´ opez CA, Ochoa-G´ omez JR, Gil-Río S, G´ omez-Jim´ enez-Aberasturi O, Torrecilla-Soria J. Chemicals from biomass: Synthesis of lactic acid by alkaline hydrothermal conversion of sorbitol. J Chem Technol Biotechnol 2011;86:867–74. https://doi.org/10.1002/jctb.2602. [54] Mainil RI, Paksung N, Matsumura Y. Determination of retro-aldol reaction type for glyceraldehyde under hydrothermal conditions. J Supercrit Fluids 2019;143: 370–7. https://doi.org/10.1016/j.supflu.2018.09.013. [55] Srokol Z, Estrik A Van, Strik RCJ, Maschmeyer T, Peters JA. Hydrothermal upgrading of biomass to biofuel ; studies on some monosaccharide model compounds 2004;339:1717–26. https://doi.org/10.1016/j.carres.2004.04.018. [56] Wang G, Meng Y, Zhou J, Zhang L. Selective hydrothermal degradation of cellulose to formic acid in alkaline solutions. Cellulose 2018;25:5659–68. https://doi.org/ 10.1007/s10570-018-1979-9. [57] Xu LL, Huo ZB, Wang FW, Jin FM. Hydrothermal conversion of cellulose into organic acids by metal oxides. Adv Mater Res 2014;860–863:497–500. https://doi. org/10.4028/www.scientific.net/AMR.860-863.497. [58] Trincado M, Grützmacher H, Prechtl MHG. CO 2 -based hydrogen storage – Hydrogen generation from formaldehyde/water. Phys Sci Rev 2018;3. https://doi. org/10.1515/psr-2017-0013. [59] Takahashi H, Liu LH, Yashiro Y, Ioku K, Bignall G, Yamasaki N. CO 2 reduction using hydrothermal method for the selective formation of organic compounds. J Mater Sci 2006;41:1585–9. https://doi.org/10.1007/s10853-006-4649-5. M. And´ erez-Fern´ andez et al.