Two-dimensional (2D) materials such as graphene are known to have specific electronic properties and a large surface area due to their 2D nature. As such, they are expected to be applicable in a wide range of fields.^1–3 Among 2D materials, those related to boron differ in terms of their polymorphs; theoretical calculations suggest that various stable structures may exist.^4,5 These include 2D compounds such as boron chalcogenides,⁶ boron oxides,^7,8 boron hydrides,^9,10 and boron phosphides.¹¹ Various stable 2D boron (borophene) sheets have been observed in experimental synthesis after the deposition of boron atoms on Ag(111) surfaces.¹² Borophene layers on metal surfaces and borophene powders have also been subsequently reported in several studies.^13–15

Hydrogen boride (HB) sheets were reported as the first experimentally synthesized hydrogenated borophene.¹⁶ HB sheets are composed of hydrogen and boron in a 1:1 stoichiometric ratio. HB sheets are terminated with protons on negatively charged hexagonal boron sheets and have a large H₂ content (8.5 wt %), unique electronic properties,^17,18 catalytic properties as solid acid catalysts,¹⁹ and a light-responsive hydrogen release function.²⁰ This study reports on HB sheets as reductants for metal ions, a new property of HB sheets. We have estimated the redox potential of HB sheets and demonstrated the facile formation of HB sheet nanocomposites with metal nanoparticles by considering the characteristics of HB sheets.

HB sheets were prepared using a previously reported ion-exchange method.^16–20 MgB₂ powder (1.0 g, 99%, Rare Metallic Co., Ltd., Tokyo, Japan) in acetonitrile (300 mL, 99.5%, Wako Pure Chemical Industries Ltd., Osaka, Japan) was added into a mixture of an ion-exchange resin (60 mL, Amberlite IR120B hydrogen, Organo Corp., Tokyo, Japan) and acetonitrile (200 mL) in a Schlenk flask under a nitrogen atmosphere. The process was sensitive to water because of the hydrolysis reaction of MgB₂.²¹ As such, water was carefully removed beforehand. This mixture was stirred with a magnetic stirrer at 400 rpm for 2 d at room temperature (c.a. 293 K). The supernatant was then kept for 1 d at 255 K to physically separate the byproduct B(OH)₃. Dried HB sheets were prepared by heating the resulting liquid at 343 K while pumping with a liquid nitrogen trap. We rigorously carried out the characterization of the product using X-ray photoelectron spectroscopy (XPS, JPS 9010 TR, JEOL, Ltd. Japan) to confirm the absence of Mg, the presence of negatively charged B, and the absence of oxidized B.^16–20 Ultraviolet–visible (UV-vis, Ocean Optics, DH-2000-BAL) spectroscopy was carried out to evaluate the suitability of the HB sheets as reductants and estimate the redox potential of the HB sheets. Noble metal or transition metal salts were dissolved in acetonitrile at a concentration of 0.005 or 0.0025 mol/L, after which HB sheets (powder) were added to the solution to a maximum concentration of 0.05 mol/L. Acetonitrile rather than water was used to prevent the reaction of HB sheets with water, despite HB sheets very weakly reacting with water (details of the interaction between HB sheets and water to be reported in future work). The preparation of the solution was carried out in a glove box to prevent oxidation of the HB sheets. The UV-vis measurement was then performed under a home-made black box, after mixing at room temperature (c.a. 293 K). Although HB sheets are only responsive to intense UV irradiation,²⁰ the sample was irradiated with UV light only for the time it took to measure a spectrum (approximately 1 min for each measurement) to prevent any unnecessary damage. The metal salts used in this work are palladium(II) bis(acetylacetonate) (Pd(C₅H₇O₂)₂, assay (TLC) minimum 97%, FUJIFILM Wako Pure Chemical Corporation), diamminedinitritoplatinum(II) (Pt(NO₂)₂(NH₃)₂, assay (Pt) minimum 60%, Soekawa Chemicals), nickel(II) bis(acetylacetonate) (Ni(C₅H₇O₂)₂, 98%, MERK), cobalt(II) bis(acetylacetonate) (Co(C₅H₇O₂, 99%, ALDRICH), zinc(II) bis(acetylacetonate) (Zn(C₅H₇O₂)xH₂O, 98%, STREM CHEMICALS Inc), and copper(II) acetate (Cu(CH₃COO)₂, 98%, Aldrich).

Figure 1 shows the results of the UV-vis measurements before and after the addition of HB sheets to the acetonitrile solution containing Pt(NO₂)₂(NH₃)₂, or Ni(C₅H₇O₂)₂. In both cases, visible light was absorbed, indicating the reduction of metal ions in the presence of HB sheets. In Pt(NO₂)₂(NH₃)₂, absorption was not observed before the addition of the HB sheets (Figure 1a). This was consistent with the transparent color of the solution (Figure 1c). The absorbance increased immediately after adding the HB sheets, mainly because of the absorption by the sheets (Figure S1). Importantly, absorbance gradually increased with time following the addition of the HB sheets. Specifically, absorbance components at c.a. 370 and 440 nm increased with time, as shown by the subtracted spectra in Figure 1b. The color of the solution consistently turned a dark yellow (Figure 1c). The increased absorbance component may be ascribed to absorbance by the metal nanoparticles of Pt.²² The results suggest that Pt²⁺ cations are formed from Pt(NO₂)₂(NH₃)₂ followed by a reduction to Pt⁰ because of the presence of HB sheets. The resultant Pt metal atoms may form Pt metal nanoparticles on HB sheets. For Ni(C₅H₇O₂)₂, the absorbance component at c.a. 420 nm increased with the time following the addition of HB sheets and the spectrum shape became constant after 1 h (Figure 1d). Based on the literature, the increased absorbance component may be ascribed to the absorbance of the Ni metal nanoparticles.²³ Thus, Ni metal nanoparticles seem to form on HB sheets as was the case with Pt(NO₂)₂(NH₃)₂. Concomitantly, the color of the solution changed from transparent to dark yellow (Figure 1f). A similar result was also obtained for Pd(C₅H₇O₂)₂ (Figure S2).

Figure 1. UV-vis measurements of an acetonitrile solution containing Pt(NO₂)₂(NH₃)₂, or Ni(C₅H₇O₂)₂ (0.005 mol/L) before and after the addition of HB sheets (0.05 mol/L). (a, d) spectrum change over time, where the arrow in d indicates the signal-saturated (meaningless) region, (b, e) subtracted spectrum of HB addition, where the spectrum after 1 min was subtracted from each spectrum and the arrows in e indicate the unsaturated signal (meaningful) region. (c, f) photo of acetonitrile solution.

In contrast, there was no evident change in the UV-vis spectrum following the addition of the HB sheets for Zn(C₅H₇O₂)₂ and Co(C₅H₇O₂)₂ as shown in Figure 2. These results indicate that Zn and Co ions are not reduced by the presence of HB sheets; a sharp contrast to the Ni, Pt, and Pd ions (Figures 1 and S2).

Figure 2. UV-vis measurements of an acetonitrile solution of (a) Zn(C₅H₇O₂)₂, and (b) Co(C₅H₇O₂)₂ (0.005 mol/L) before and after the addition of HB sheets (0.05 mol/L).

Despite the pH in the acetonitrile solution being unclear, we can make a rough estimate of the redox potential for HB = (HB)⁺ + e⁻ by considering which metal ions are reduced by HB sheets. Ni ions were reduced (Figure 1d) while Co ions were not (Figure 2b) in the presence of HB sheets. Based on previously reported redox potentials for various metals (Table S1), those of Ni and Co are −0.277 and −0.257 V, respectively, versus the standard hydrogen electrode (SHE). This indicates that the redox potential of the HB sheets is between −0.277 and −0.257 V versus SHE. This redox potential is between that of NaBH₄ (at least below −0.44 V versus SHE)²⁴ and Mg-deficient hydroxyl-functionalized boron nanosheets (c.a. 0.16 V versus SHE).²⁴

The estimated redox potential indicates that HB sheets can reduce metal ions with redox potentials larger than −0.257 V versus SHE, as shown in Table S1 (Ni²⁺, V³⁺, Mo³⁺, Sn²⁺, Pb²⁺, Cu²⁺, Cu⁺, Fe³⁺, Ag⁺, Pd²⁺, Pt²⁺). However, HB sheets cannot reduce metal ions with redox potential smaller than −0.277 V versus SHE, as shown in Table S1 (Zn²⁺, Fe²⁺, Co²⁺). The redox potential also indicates that CO₂ may be converted to HCOOH. Also, CO₃²⁻ can be reduced to form HCHO in the presence of HB sheets at least thermodynamically (see Table S1). Thus, HB sheets are attractive in applications as catalysts for CO₂ conversion.

In this study, we demonstrate the simple formation of nanocomposites of HB sheets with Cu metal nanoparticles using the reducing property of HB sheets. Such nanocomposites are very important and useful for various applications, including catalysts.^1,25 Figure 3a shows the UV-vis results for acetonitrile solution containing Cu(CH₃COO)₂ after it was mixed with HB sheets. The absorbance increased with time, indicating the reduction of Cu²⁺ and the formation of Cu nanoparticles on the HB sheets. A distinct peak at c.a. 400 nm and a weak peak at c.a. 600 nm may be ascribed to absorption in electronic transitions from d-band to the sp-band of Cu²⁶ and absorption by the surface plasmon resonance of Cu nanoparticles.²⁷ We note that there was no absorbance peak at c.a. 670 nm from the beginning of the mixing in contrast to the case of the spectrum of Cu(CH₃COO)₂ in acetonitrile (Figure S3). This indicates that all Cu²⁺ ions were reduced immediately by HB after mixing. That is, the gradual change of the spectrum in Figure 3a is representative of the gradual growth of Cu nanoparticles on HB sheets. The sample dispersion in the solution was indicated by the Tindal effect, as shown in the inset of Figure 3a. Figure 3b presents the transmission electron microscopy (TEM, JEM-2010F 2010F, JEOL Ltd., Japan, 200 kV) image of the nanocomposites. Highly dispersive nanoparticles were observed on the HB sheets as dark spots, approximately 2 nm in size. Energy dispersive X-ray spectroscopy (EDX) indicated that these were Cu nanoparticles. These results clearly demonstrate that nanocomposites of HB sheets with Cu nanoparticles (2.0 nm in diameter) may form spontaneously simply through the addition of HB sheets and Cu ions to an acetonitrile solution. We have also successfully produced a similar nanocomposite using a methanol solution (Figure S4).

Figure 3. Formation of HB sheet nanocomposites with Cu metal nanoparticles. (a) UV-vis measurements of an acetonitrile solution of Cu(CH₃COO)₂ (0.0025 mol/L) after mixing with HB sheets (0.05 mol/L). (b) TEM and (c) XPS results of the product after the mixture (over 24 h immersion).

In the above process (i.e. the addition of HB sheets to the acetonitrile solution containing Cu(CH₃COO)₂), we did not observe any obvious gas production from the system. Hydrogen gas was detected only in amounts 2–3 orders of magnitude smaller than the used HB amount and no other gas (such as B₂H₆) was detected in our gas analysis of the head space of the experimental system. Thus, the following reactions are considered to have occurred.

To verify this mechanism, we conducted XPS analysis for the nanocomposites as shown in Figure 3c. We observed a metallic Cu 2p 3/2 peak at 932.8 eV, while two peaks were observed in the B 1s region. The B 1s peak at 188.0 eV corresponds to that of the HB sheets (negatively charged boron),^16–20 while the peak at 192.5 eV may be considered an oxidized boron peak.²¹ These results are consistent with eq 1, i.e., metallic Cu and oxidized B were detected in addition to the negatively charged B. Quantitatively, the Cu amount was estimated as 7.25 ± 2.50 at % based on the XPS peak areas and sensitivity (B 1s: 2.10, and Cu 2p3/2: 60.5 or Cu 2p1/2: 31.1). This is similar to the prepared 5 at %, indicating that all prepared Cu²⁺ ions were reduced as shown by UV-vis spectrum (Figures 3a and S3). The oxidized boron peak dominated as much as 46% of B 1s peaks, much larger than expected 5%. This may be attributable to other subsequent reaction of boron. Using methanol, we could prepare the nanocomposites of Cu and HB without any oxidized boron (Figure S5). Importantly, the metallic states of Cu were clearly shown in all prepared nanocomposites (XPS analysis for the cases of Ni an Co also illustrate this consistency (Figure S6)). If other species with a reactive redox potential are present, a part of the HB sheets may be oxidized in other ways together with a reduction of the Cu²⁺ cations during the formation of the nanocomposites (see for example Figure S5). As the HB sheets are electronically conductive¹⁸ (though they have an optical band gap due to the selection rule),^16–18,20 electron transfer may occur in the HB sheets. Thus, both the oxidation reaction of a part of the HB sheets and the reduction reaction of metal ions may occur concurrently on the same HB sheets at different sites, with electron transfer in the HB sheet. The simple preparation of nanocomposites using such a method has also been reported for nanocomposites using graphene,²⁷ layered CaSi₂,²⁸ polysilane,²⁹ and Mg-deficient hydroxyl-functionalized boron nanosheets.²⁴ As reported in nanoparticles on graphene,³⁰ we may control the size and density of nanoparticles by controlling the initial prepared amount of metal ions and HB. A smaller amount of metal ions with respect to the HB sheets leads to smaller sized nanoparticles. Figure 3a demonstrates that the reduction occurred immediately while growth occurred slowly. Therefore, in this particular case, the control of the immersion time may also regulate the size of the nanoparticles. The other preparation conditions such as temperature, and type of solution may also be important in controlling the size and density of nanoparticles on HB sheets.

In summary, we found that HB sheets may be used as reductants for several metal ions. The redox potential of the HB sheets was estimated on the basis of UV-vis measurements; it was between −0.277 and −0.257 V versus SHE. We then demonstrated a facile formation of nanocomposites of the HB sheets with Cu metal nanoparticles using the reduction property, as schematically summarized in Figure 4.

Figure 4. Schematic image of the facile nanocomposites formation of HB sheets with metal nanoparticles.

T. Kondo

M. Miyauchi

S. Ito