Nanoparticle-enabled integration of air capture and conversion of CO 2 †

Integrating air capture and conversion of CO 2 is key to realizing energy sustainability. However, current integration approaches require high temperature and pressure, making them energy intensive. Here, we demonstrate a nanoparticle (NP) catalysis approach for the hydrogenation of alkyl carbonate, an intermediate obtained from the CO 2 capture process, to formate, achieving one-pot air capture and conversion of CO 2 under ambient conditions. The capture is realized in an ethylene glycol (EG) solution of KOH (EG – KOH) at room temperature, where CO 2 is selectively converted into HO – CH 2 CH 2 – O – COOK ( ∼ 100% conversion). This carbonate is then hydrogenated using ammonia borane (under ambient pressure and at 50 °C) to formate (HCOOK) (>90% yield) in the presence of a stable Pd NP catalyst with EG being regenerated. Atomistic simulations suggest that the CO 2 absorption process in the EG – KOH solution is energetically stable, and the catalyst surface provides the reaction site to break the C – O bond in the – O – COOK structure, enabling the hydrogenation of the alkyl carbonate to formate and the regeneration of EG. Our study provides a promising NP-catalysis approach for air capture and conversion of CO 2 into value-added chemicals/fuels under ambient conditions.


Introduction
2][3][4][5] The key to the success of this approach is the energy-efficient integration of the capture/conversion processes, 6,7 a goal that has been extremely challenging to achieve.9][20][21][22][23] To realize the energy-efficient integration of CO 2 capture and conversion, a stable and recyclable catalytic system should be present, which can activate and convert the alkyl carbonate into a valuable chemical or fuel under ambient conditions.
Nanoparticles (NPs) have been studied extensively for catalytic applications.Herein, we report a Pd nanoparticle (NP)-enabled approach for air capture and conversion of CO 2 under ambient conditions, as summarized in Scheme 1.The key idea of our approach is first to capture CO 2 in an ethylene glycol (EG) solution of KOH to convert CO 2 into alkyl carbonate and then to reduce this alkyl carbonate in situ to formate via Pd NP-catalyzed transfer hydrogenation of ammonia borane (NH 3 BH 3 , AB) with EG being regenerated.Previous studies have shown that pure CO 2 may be hydrogenated by a metal borohydride or AB in the presence of a special pincer complex as a catalyst, [24][25][26][27][28] but such a hydrogenation reaction is not suitable for the desired one-step air capture and conversion process.We found that Pd NPs were active in catalyzing the AB hydrogenation of alkyl carbonate to formate under ambient pressure.The reaction led to full conversion (100%) of alkyl carbonate and high yield (92%) formation of formate in a 6 h reaction period at 50 °C.The process can be readily applied to integrate air capture and conversion of CO 2 into formate.The Pd NPs are stable in the one-pot capture/conversion process and can be reused for the next capture and conversion step.This NP-catalysis enabled one-pot approach demonstrates a potential energy-efficient integration of air capture and conversion of CO 2 into value-added chemicals/fuels.

Experimental procedures
Pd nanoparticle (NP) synthesis 5 nm Pd NPs were synthesized by modifying a previously reported method. 290.150 g of Pd(acac) 2 and oleylamine (15 mL) were added into a three-necked flask and then degassed with nitrogen at 60 °C for 30 min.0.300 g of borane tert-butylamine was dispersed in 4 mL of oleylamine by vortexing and shaking.The temperature of the Pd precursor solution was lowered to and maintained at 40 °C.The solution of borane-tert-butylamine was injected into the reaction flask quickly, and within one minute the color of the solution changed to dark brown-black.The temperature was raised to 90 °C at a ramp rate of 3-5 °C min −1 .The solution was maintained at 90 °C for 1 h.The NPs were precipitated by adding 100 mL of ethanol and separated by centrifugation at 9000 rpm for 8 min.The NPs were then washed twice with ∼75 mL of 1 : 9 hexane : ethanol (9000 rpm for 8 min).The NPs were redispersed in hexane for storage.The subsequent catalyst loading and the activation step are presented in the ESI.† Direct air capture of CO 2 and in situ hydrogenation reaction 5 mmol of KOH was dissolved in 10 mL of EG.A water aspirator was used to create a partial vacuum in the reaction system so that ambient air (∼413 ppm CO 2 ) could continuously bubble through the capture solution.The flow rate was controlled to be 0.2 L min −1 .The solution pH was monitored during the capture process to determine the CO 2 saturation point.After 48 hours of bubbling, the solution volume increased by about 3 mL due to moisture in the air.The capture solution was then purged with N 2 to remove gas impurities.Next, 10 mg of Pd/C catalysts and 100 mg of ammonia borane were added to the solution, which was sealed using a balloon.The reaction mixture was maintained at 50 °C for 24 hours.Both gas and liquid products were subjected to GC and NMR analysis, respectively.
The hydrogenation reaction could also be performed under air without N 2 purging, yielding a similar reaction result.
All other electrochemical reaction procedures, characterization methods and computational details are given in the ESI.†

Results and discussion
To study the activation of alkyl carbonate, we first reacted pure CO 2 with an ethylene glycol (EG) solution of KOH to form HO-CH 2 CH 2 -O-COOK.This alkaline chemistry leading to efficient CO 2 capture has been well-known. 7,12,13In this study, we bubbled CO 2 from its gas cylinder into the EG solution of KOH at room temperature.In the alkaline EG solution, EG can react with KOH to form a glycoxide anion and water, quickly reaching equilibrium.Upon exposure to CO 2 at room temperature, this equilibrium is shifted to the right, as CO 2 reacts with the glycoxide to form alkyl carbonate (Fig. 1A).This CO 2 absorption process reaches its saturation point once KOH is consumed.We monitored solution pH and product formation to determine the endpoint of the capture process as well as the CO 2 capture capacity.As shown in Fig. 1B, the initial pH of the EG solution of KOH (5 mmol KOH dissolved in 10 mL EG) is around 10.7.Upon CO 2 bubbling, the pH gradually decreased and reached a plateau of 7.8 after 40 min.The CO 2 capture capacity of this KOH solution at saturation was measured using the weight change of the solution before and after capture, which was determined to be 220 mg gravimetrically.The capture efficiency can reach nearly 100%, which corresponds to 1 mol CO 2 per mol KOH (Fig. 1C).The alkyl carbonate formed in the EG solution is stable after the capture process and can be quantitatively characterized by both 1 H and 13 C nuclear magnetic resonance (NMR) spectroscopy, which is consistent with the weight change (Fig. S1 †).In this capture process, the amount of KOH is the limiting factor, which determines the maximum amount of CO 2 the system can capture.
The alkyl carbonate could not be activated and reduced by a common hydride agent (AB or sodium borohydride) under atmospheric pressure and at temperatures ranging from room temperature to 80 °C we screened.However, when we studied NP catalysis for transfer hydrogenation of AB, we found Pd NPs to be more active in catalyzing the hydrogenation of alkyl carbonate to formate at 50 °C.We prepared 5 nm Pd NPs via solution phase reduction of Pd(acac) 2 with borane t-butylamine in oleylamine. 29Fig. S2A † shows the transmission electron microscopy (TEM) image of the 5 nm Pd NPs obtained from the synthesis.The NPs were then loaded onto carbon support, followed by an acetic acid wash to remove the surfactant oleylamine as reported. 30The TEM image showed that the Pd NPs are well-dispersed on the carbon support after this treatment, showing no sign of aggregation (Fig. S2B †).The Pd NPs have a typical fcc structure as characterized by X-ray diffraction (XRD) (Fig. S2C †).We studied the reduction of alkyl carbonate by AB in the capture solution in the presence of a Pd NP catalyst (Fig. 2A and ESI †).We used a balloon to seal the reaction vessel to ensure that the reaction proceeded under ambient pressure.The conversion of alkyl carbonate was monitored by 1 H NMR using dimethylformamide as an internal standard in d 6 -DMSO solvent.A small amount of aliquot was taken for analysis at each hour time mark.The alkyl carbonate conversion increased with reaction time and reached 100% after 6 h of reaction (Fig. 2B).In the solution phase, formate was the only product characterized by both 1 H and 13 C NMR (Fig. S3 †).The time-dependent formate yield during the reaction is shown in Fig. 2C, reaching its maximum of 92% after showed that alkyl carbonate is stable in the reaction solution without AB and NPs, even under heating at 120 °C (Fig. S4 †).4][35] We further investigated the Pd-catalyzed hydrogenation reaction in the presence of NaBH 4 and obtained a similar formate yield.However, atmospheric H 2 was not active enough for the hydrogenation process as the formate yield was only 7% (Table S2 †).This indicates that under our current reaction conditions, a hydride intermediate, not a free H 2 molecule, is an active species in the Pd NP-catalyzed hydrogenation of alkyl carbonate to formate.
The formation of potassium carbonate (K 2 CO 3 ) and/or bicarbonate (KHCO 3 ) is inevitable when water and KOH are present in the capture solution.We further studied how the formation of K 2 CO 3 /KHCO 3 affects the AB transfer hydrogenation of alkyl carbonate.We found that both K 2 CO 3 and KHCO 3 could be dissolved in EG and further reduced by AB in the presence of Pd NPs to formate (87% and 91% yield, respectively) (Fig. 3).This finding is significant for the successful air capture and conversion of CO 2 , as it makes the hydrogenation of K 2 CO 3 /KHCO 3 possible without high energy input (inorganic carbonate is generally considered inactive and can only be reduced under harsh conditions 36,37 ).Under our reaction conditions, EG plays an important role in activating K 2 CO 3 /KHCO 3 by forming alkyl carbonate, as confirmed by NMR of the precursor solution containing K 2 CO 3 /KHCO 3 and EG (Fig. S5 †).This suggests that there exists an equilibrium between carbonate and alkyl carbonate.To further support this point, we performed K 2 CO 3 /KHCO 3 reduction in an aqueous solution without EG under the same conditions, and we could only detect a very small amount of formate (<6% yield).Therefore, we conclude that EG reacts with carbonates to form alkyl carbonate, reaching equilibrium.Under the Pd NP-catalyzed AB hydrogenation conditions, alkyl carbonate is reduced to formate, which shifts the equilibrium to the right until carbonates are consumed.This indicates that water in the EG/KOH solution only slows down the hydrogenation process, but does not affect the final reaction outcome, which is key to continuous air capture and conversion of CO 2 .
Our model study on pure CO 2 capture and conversion in EG-KOH solution as well as the EG role in activating K 2 CO 3 /  KHCO 3 motivated us to further explore CO 2 from air, where moisture is also ubiquitously present.The capture and conversion system and the related reactions are outlined in Scheme 1.A water aspirator was used as a pump to draw ambient air into the capture solution, in which ultra-dilute CO 2 (∼413 ppm) was captured by KOH in EG. 7 The solution pH was monitored throughout the capture process, and it took 48 h for the pH to stabilize (Fig. 4A).The 13 C NMR analysis of the capture product shows an additional peak of CO 3 2− (165.70 ppm) in addition to the presence of alkyl carbonate (Fig. 4B and C).The CO 3 2− species must be formed from reactions among water (from the air), KOH, and CO 2 .Based on the quantitative 13 C NMR analysis, the ratio between alkyl carbonate and K 2 CO 3 was about 1 : 1.When the air capture solution was hydrogenated by AB at 50 °C in the presence of the Pd NP catalyst for 24 h, all carbonates were converted into formate (91% yield).Compared with the result observed from the pure CO 2 capture and conversion demonstrated above, the hydrogenation of carbonates from air capture requires a longer reaction time due to the less efficient conversion of inorganic carbonate into alkyl carbonate in the presence of water.Another highlight of our reaction system is that the Pd NP-catalyzed transfer hydrogenation is O 2 tolerantthe reaction can proceed without the need to purge the reaction system with N 2 to remove O 2 .As a comparison, O 2 must be removed from the reaction system under conventional hydrogenation reaction conditions. 2,7,12Finally, the Pd NPs showed excellent stability under the catalytic transfer hydrogenation reaction conditions and maintained a formate yield of over 90% after 5 reaction cycles (Fig. 4D) (the same amounts of AB and KOH were added in each reaction cycle).The Pd NPs retain their crystal structure (Fig. S6A †) and NP size/dispersity (Fig. S6B †) as confirmed by the XRD and TEM analyses of the NPs before and after the catalytic reactions, indicating the long-term stability of the Pd NPs as an efficient catalyst for continuous air capture and conversion.
To understand why Pd NPs are efficient for the AB hydrogenation of alkyl carbonate, we used density functional theory (DFT) calculations to further investigate the reaction mechanism on Pd surfaces.We suggest a 5-step mechanism, starting with the incorporation of CO 2 into a solution of KOH in EG, as shown in Fig. 1.Given that the Pd NPs are nano-sized polyhedra (∼5 nm) (Fig. S2 †), which means that they consist of many step and terrace sites, both the planar (111) and the stepped (211) crystal facets of Pd were examined using DFT calculations, as shown in Fig. S9.† This comparison shows a more facile route on the (211) step edge than on the (111) terrace, and we highlight the (211) results herein.
The overall suggested reaction mechanism and the corresponding Gibbs free energies are summarized in Fig. 5, where "*" represents an adsorbed species on the Pd surface.State 1 → state 2 represents the formation of an EG-CO 2 complex under the pH 7.8 environment; we considered the pH effect in state 1 protons are involved in the reaction for forming the EG-CO 2 complex. 38As expected from the experiments in which alkyl carbonate intermediates were formed immediately when CO 2 was purged, the reaction from state 1 to state 2 proceeds spontaneously with a downhill energy path.While the reaction from state 1 to state 2 takes place without a Pd catalyst, after state 2, the reaction takes place on the surface of the Pd catalyst.Accordingly, each Gibbs energy of activation (reaction energy barrier) from state 3 was calculated to investigate the rate-determining step.T1 is an abbreviation of transition state 1, and a total of 3 transition states (T1, T2, and T3) and each Gibbs energy of activation (ΔG ‡ T1 ; ΔG ‡ T2 ; and ΔG ‡ T3 , respectively) were calculated by the dynamic nudged elastic band (DyNEB) method. 39We modeled the process of H 2 formation by AB as a process in which H* is adsorbed to Pd from H 2 (state 2 → state 3), and this process is a thermodynamically easy reaction with a downhill energy profile.Considering the consistency in the calculation environment, we started the reaction with an H 2 gas molecule represented by a black color bar, while also showing the reaction energy for generating H 2 from NH 3 BH 3 in a gray color bar to account for the use of NH 3 BH 3 in the actual experiments.The theoretical Gibbs energy of activation ΔG ‡ theoretical ¼ 0:88 eV was calculated as the free energy change between state 3 (the lowest-energy state) and transition state 2 (the highest-energy barrier), which suggests that the rate-determining step of this reaction is the hydrogenation and release of EG (HOCH 2 CH 2 OCOO − + H* → HOCH 2 CH 2 OH).Considering that the experimental turnover frequency (TOF) is 0.059 s −1 , we can estimate that ΔG ‡ exp (Gibbs energy for the experimental activation) is around 0.95 eV from the Eyring equation: where κ is the transmission coefficient, which was assumed to be 1 in this study, ΔG ‡ is the Gibbs energy of activation, k B is Boltzmann's constant, T is the temperature, h is Planck's constant, and R is the gas constant.
Although the DFT-calculated barriers were close to the experimental barrier, we wanted to understand if the deviation was reasonable.We predicted errors from the ensemble-based Bayesian error estimation functional (BEEF) (Fig. 5, see the blue and red gradient box), 40 where the calculated ΔG ‡ theoretical was in the range between 0.80 eV and 0.95 eV, as shown in Fig. 5. Given the error in any experimental measurement, we consider the experimental and calculated barriers to be indistinguishable.
After the breakage of the intermediate product and adsorption on the Pd surface (state 4), a hydrogenation reaction proceeds to produce EG and HCOOH.State 4 → state 5 represents the production of recycled EG, and state 5 → state 6 is an HCOOH production step.As a result, it was confirmed that the CO 2 absorption process in the solution of KOH in EG is energetically stable (state 1 → state 2), and the major role of the Pd catalyst is to break the C-O bond of the alkyl carbonate intermediate (state 3 → state 4), as well as to provide a reservoir of hydride for product release.
We also tested if the direct air capture system can be coupled with electrocatalysis to reduce alkyl carbonate without the use of AB.In our preliminary studies, Pd NPs active in catalyzing the electrochemical reduction of alkyl carbonate to formate, but the reduction at the catalyst surface produced at too low a concentration to detect.Instead of pursuing the reduction to a liquid product, we tested the reduction to a gaseous product that can be easily separated from the reaction system and characterized.Since Au nanowires are known to be active to electro-catalyze CO 2 to CO, 41 we prepared 2 nm ultra-thin Au nanowires (Fig. S7 †) and studied their catalysis for the electrochemical reduction of alkyl carbonate in the capture solution (EG/H 2 O v/v ∼ 5/1) (Fig. S8 †).At −1.4 V (vs.Ag/AgCl), the CO faradaic efficiency reached 21% with H 2 as the side product.An important point to note is that carbonate/bicarbonate cannot be reduced under common electrochemical CO 2 reduction conditions, 42 and some studies suggest that the electrochemical conversion of carbonates from direct air capture solution could be attributed to the acidification of the carbonates to CO 2 at the bipolar membrane interface. 43We are currently working on demonstrating nanocatalysts for efficient electro-reduction of alkyl carbonate as a new approach for the integration of direct air capture and conversion of CO 2 into value-added chemicals.

Conclusions
This report demonstrates an NP-catalyzed hydrogenation of carbonates to realize the integration of air capture and conversion of CO 2 into formate.The key concept is to fast capture CO 2 from the air via a strong alkaline solution of EG-KOH, converting the CO 2 into alkyl carbonate HOCH 2 CH 2 O-COOK.This carbonate is stable in the capture medium and can be characterized and quantified.More importantly, the alkyl carbonate can be further activated by Pd NPs and then hydrogenated by AB (100% conversion) to formate (>90% yield) under ambient pressure and 50 °C with EG being regenerated.The formation of carbonate/bicarbonate salts from the air (moisture) capture process only affects the reaction kinetics, not the overall reaction outcome due to the established equilibrium among carbonates, EG, and alkyl carbonate.Without EG, these carbonates cannot be easily hydrogenated, demonstrating the important role of EG in promoting the hydrogenation of carbonates.DFT calculations suggest that the (211) step edge exposure on the Pd NP surface is especially active for promoting the carbonate hydrogenation reaction.The Gibbs free energy diagram shows a 5-step reaction mechanism for the CO 2 capture and conversion process, and the rate-determining step is the hydrogenation and release of EG.After incorporating the solvation effect, the calculated theoretical activation barriers match well with the experimental ones.Our studies provide a promising approach for the continuous air capture of CO 2 and subsequent formate synthesis.Our preliminary tests also show that it is possible to use an NP catalyst to activate the alkyl carbonate electrochemically.We are working to develop a robust electrocatalyst system to selectively convert this alkyl carbonate into value-added chemicals/fuels and to realize energy-efficient integration of direct air capture and conversion of CO 2 for green and sustainable energy applications.

Fig. 2 (
Fig. 2 (A) Integrated CO 2 capture and conversion scheme.(B) Time dependent alkyl carbonate hydrogenation reaction at 50 °C showing carbonate conversion.(C) Temperature effect on the alkyl carbonate hydrogenation showing formate yield.Reaction conditions: solutions from the capture step were directly used for hydrogenation.Alkyl carbonate: 2.0 mmol, ammonia borane (AB): 1.0 mmol, Pd nanocatalyst: 0.1 mol%, and ambient pressure.Formate yields were determined by 1 H NMR, and the average value was reported based on three independent experiments.

Fig. 5
Fig. 5 The schemes at the top of the graph show the reaction process, with the red, gray, white, and cyan atoms in the schemes representing O, C, H, and Pd, respectively.Bottom shows the free energy diagram for formic acid (HCOOH) formation by EG and CO 2 at 50 °C and 1 atm.The y-axis represents the Gibbs free energy (G, eV), and the Gibbs free energy of activation ΔG ‡ Tx is shown (red color) inside the graph.Free energy diagram with error estimation (blue gradient box for each state) obtained from the BEEF ensembles.The free energy required for H 2 production from NH 3 BH 3 is shown in gray for state 1 and state 2.