Carbon dioxide adsorption and conversion to methane and ethane on hydrogen boride sheets

Hydrogen boride (HB) sheets are metal-free two-dimensional materials comprising boron and hydrogen in a 1:1 stoichiometric ratio. In spite of the several advancements, the fundamental interactions between HB sheets and discrete molecules remain unclear. Here, we report the adsorption of CO2 and its conversion to CH4 and C2H6 using hydrogen-deficient HB sheets. Although fresh HB sheets did not adsorb CO2, hydrogen-deficient HB sheets reproducibly physisorbed CO2 at 297 K. The adsorption followed the Langmuir model with a saturation coverage of 2.4 × 10−4 mol g−1 and a heat of adsorption of approximately 20 kJ mol−1, which was supported by density functional theory calculations. When heated in a CO2 atmosphere, hydrogen-deficient HB began reacting with CO2 at 423 K. The detection of CH4 and C2H6 as CO2 reaction products in a moist atmosphere indicated that hydrogen-deficient HB promotes C–C coupling and CO2 conversion reactions. Our findings highlight the application potential of HB sheets as catalysts for CO2 conversion.

Supplementary Figure 2 | Estimation of chemisorbed CO 2 . a) CO2 pressure change at 297 K for HB sheets pre-heated at 473 K under vacuum for 1 h (first cycle). b) CO2 pressure change at 297 K for HB sheets after the first cycle, where instead of conducting the pre-heating process, the gas left in the system was simply evacuated (second cycle). The pressure change in the second cycle did not resemble that in the first cycle, which contradicted the case shown in Fig. 2. In this case, because there was already chemisorbed CO2 on the sample, the pressure change due to physisorption was small. The amount of chemisorbed CO2 was thus estimated as 1.44 × 10 −5 mol gHB −1 by subtraction of the observed physisorbed CO2 amount.
Supplementary Figure 3 | X-ray photoelectron spectra of HB sheets. B 1s spectra of (a) HB sheets after heating at 573 K in ultra-high vacuum (UHV), (b) HB sheets after heating at 573 K followed by exposure to air at 300 K and evacuated to UHV, and (c) as-prepared HB sheets. Compared to the as-prepared HB sheets, the main peak at 187.9 eV (corresponding to B − ) decreased from 80% to 49% after heating without air exposure (peak at 188.8 eV may correspond to the B at defects and/or edges). In addition, a peak appeared at 192 eV corresponding to an oxidized component. This indicates that 31% of the boron in the HB sheets changed its chemical state upon heating at 573 K, which corresponds to H2 release. After air exposure at 300 K and evacuation to UHV, most of the B − peak component shifted, indicating that the chemical nature of all boron species in the sample changed upon air exposure owing to the adsorption of molecules at the reactive sites. Some of the boron species may recover their original states after the desorpiton of these adsorbed species; however, some may fail to recover, thus causing the HB sheets to degrade. These results suggest that the hydrogen-deficient HB sheets prepared by heating at 573 K are chemically unstable and react readily with residual oxygen, CO2, and/or water in the system to form oxidized boron, which hinders the surface adsorption of CO2.

Supplementary Figure 4 | Potential energy surfaces for CO2 in the side-on configuration on a pristine HB sheet at various CO 2 -HB sheet distances (z) and molecular orientations (θ).
The adsorption configuration is shown on the left: z is defined by the difference between the z-coordinate of CO2 and the average z-coordinate of the surface H atoms of the HB sheet. In the calculations, the C atom of CO2 was placed on a 4 × 4 grid of the primitive surface unit cell of the HB sheet. The values of adsorption energy are shown in the bar on the left, where blue (or red) indicates the repulsive (or attractive) interaction between CO2 and the HB sheet. At a short distance (z = 2.033 Å), the repulsive interaction dominates as indicated by the blue iso-surfaces, whereas at larger distances, the interaction is attractive as indicated by the red iso-surface; however, there is no significant preference for the adsorption site or orientation.

Supplementary Figure 5 | Potential energy surfaces for CO2 adsorption in the side-on configuration on a H-vacant HB sheet at various CO 2 -HB sheet distances (z) and molecular orientations (θ).
The adsorption configuration is shown on the left: z is defined by the difference between the z-coordinate of CO2 and the average z-coordinate of the surface H atoms on HB. In the calculation, the C atom of CO2 was placed on a 4 × 4 grid of the primitive surface unit cell of the HB sheet. The values of the adsorption energy are shown in the bar on the left, where blue (or red) indicates the repulsive (or attractive) interaction of CO2 with the HB sheet. At a short distance (z = 2.028 Å), there is a preferential adsorption at the H vacancy site, which is indicated by the dark red iso-surfaces.

Supplementary Figure 6 | Potential energy surfaces for CO2 in the end-on configuration on pristine and H-vacant HB sheets (VH) at various CO2-HB sheet distances (z).
The adsorption configuration is shown on the left: z is defined by the difference between the z-coordinate of the bottom O atom and the average z-coordinate of the surface H atoms on HB. In the calculation, the C atom of CO2 was placed on a 4 × 4 grid of the primitive surface unit cell of HB. The values of the adsorption energy are shown in the bar on the left, where blue (or red) indicates repulsive (or attractive) interaction of CO2 with the HB sheet. The interaction of CO2 with the HB sheet is attractive but very weak regardless of the H vacancy, as indicated by the pale red iso-surfaces.

Supplementary Figure 7 | Reaction products of CO2 (10 cm 3 ) and HB sheets (100 mg) at 523 K under moist conditions (0.1 cm 3 H2O).
Typical examples of gas chromatographic analysis using a) a molecular sieve (5 Å) and b) Porapak Q columns. For comparison, the result while using He (no CO2) is also shown in panel a. We note that O2 and N2 were detected from the inevitable inclusion of air in the syringe during sampling in every case. However, we found that the O2/N2 intensity ratio was larger than that of air in the experiments with CO2 and H2O, while the experiment with Ar showed the same O2/N2 intensity ratio as that of air. Thus, we estimated the O2 intensity in the experiments with CO2 and H2O by subtracting the background O2 amount, as estimated from the detected N2 intensity and the O2/N2 ratio of air (measured beforehand). Thus, O2, CH4, C2H6, and CO were detected when the HB sheets were heated in CO2 and H2O (exemplary quantities are shown in Supplementary Table 1). We note here that the amount of O2 dissolved in the inputted H2O was three orders of magnitude smaller than the detected amount. We sometimes detected small quantities of C3H8, as shown in b. These signals did not appear upon heating in a He atmosphere; thus, these species were identified as reaction products of HB and CO2 at 523 K. As such, CO2 was converted to hydrocarbons, such as methane, ethane, and propane, by reacting with the HB sheets, because CO2 was the only carbon source in the system. Figure 8 | Reaction products of CO2 (10 cm 3 ) and HB sheets (100 mg) after 6 h at 523 K under moist conditions (0.1 cm 3 H 2 O). Gas chromatography analysis using a) a molecular sieve (5 Å) and b) Porapak Q columns. c) GC-MS analysis at mass number 30 (C2H6).

Supplementary
Supplementary Figure 9 | X-ray photoelectron spectroscopy survey scan of HB at 300 K, after heating at 523 K in CO 2 , and after heating at 873 K in CO 2 . Au sheet was used as the sample holder.