Realization of highly asymmetric hydrogenated graphene in the van der Waals confined space

ABSTRACT

The van der Waals (vdW) confined space provides a distinct environment from free space, enabling the production of two-dimensional Janus materials, like highly asymmetric hydrogenated graphene (AH-Gr). Here, we develop a vdW confined space assisted hydrogenation method to produce AH-Gr. The confined space between graphene and the substrate aggregates hydrogen radicals, making the bottom-side of graphene more prone to hydrogenation. The dense and homogeneous confined spaces between adjacent vdW crystals promote rapid and uniform distribution of carbon-hydrogen (C−H) bonds. The hydrogen-to-carbon atomic (H/C) ratios can be quantitatively controlled by adjusting the permeated proton dose. All AH-Gr, regardless of H/C ratios, remain vacancy-free. The spatial distributions of C−H bonds significantly influence the electrical and magnetic properties of AH-Gr. Asymmetric hydrogenation transforms graphene from a semi-metal to a semiconductor, suppresses the quantum Hall effect, and reduces the phase coherence length. This study provides new insights into the preparation and characteristics of hydrogenated graphene, broadening the applications of vdW confined space.

INTRODUCTION

Van der Waals (vdW) confined space, the empty space between adjacent vdW crystals spanning angstroms to nanometers [1], provides a fantastic nano-environment of causing various physical and chemical performances, including the vdW pressure [2,3], frictionless flow [4], untraditional phase transition [5,6], selective permeation [7,8], confined catalysis [9], etc. The narrowest interlayer space restricts the transport of most molecules and ions [10], but they can be expanded by intercalating alkalis and acid molecules under specific conditions [11,12]. The spaces between vdW crystals and the non-vdW substates are also regarded as vdW confined spaces [5], albeit less homogeneous but still prevalent and significantly influential. For example, water or oxygen molecules confined between graphene and SiO₂/Si can induce p-type doping of graphene [13].

Graphene, the first exfoliated vdW crystal, holds promise in electronic, photonic and spintronic applications [14]. However, its lack of an intrinsic bandgap restricts its use in transistors and sensors [15]. Bandgaps can be induced by breaking the graphene's structural symmetry through functionalization with various molecules [16,17], such as hydrogen [18–22], fluorine [23], organic radicals [16], et al. Among these, hydrogenation is widely employed due to its accessibility, reversibility and harmlessness [18,21]. Depending on the distribution of carbon-hydrogen (C−H) bonds, hydrogenated graphene can be classified into lowly asymmetric forms (LH-Gr), such as ideal double-side hydrogenated graphene, and highly asymmetric forms (AH-Gr), such as ideal single-side hydrogenated graphene [18,24]. AH-Gr exhibits distinct properties compared to intrinsic graphene [25], including an indirect bandgap [26,27], strong spin-orbit coupling [28,29], spin-triplet exciton [30] and ferromagnetism [31]. Particularly, graphone, a fully hydrogenated form of single-side hydrogenated graphene, is theoretically predicted to manifest many distinctive properties [29,31].

Currently, hydrogenation of graphene typically utilizes hydrogen radicals or protons as reactants, including H₂ plasma [18,21], hydrogen atom exposure [19,20,22,32] and hydrogen silsesquioxane dissociation [28]. AH-Gr films have been reported to be produced by treating graphene on SiO₂/Si (Gr/SiO₂) substrate with hydrogen plasma [18], generating active H radicals on epitaxial graphene under ultra-high vacuum (UHV) [32], or applying hydrogen silsesquioxane followed by electron beam irradiation [28]. These methods introduce magnetic moment asymmetry and significant spin-orbit coupling in hydrogenated graphene [28,32]. However, recent findings indicate that the graphene's hexagonal lattices are permeable to both protons and hydrogen molecules [33–36], complicating the construction of AH-Gr since both sides are exposed to active H radicals or protons. Therefore, achieving AH-Gr with a precise hydrogen-to-carbon atomic ratio (H/C) remains a significant challenge.

The crucial strategy for producing AH-Gr lies in designing differentiated chemical environments on each side of the graphene. Actually, graphene on substrate fulfills this criterion, i.e. free space on the upper side and confined space on the bottom. The vdW confined space between graphene and substrate serves as a nano-sized reactive chamber that influences internal chemical reactions [5,9], enhancing the stability of active sites and improving reaction activities [37]. In this study, we propose a vdW confined space assisted hydrogenation method for precise AH-Gr production. The permeated protons accumulate and recombine into H radicals between graphene and substate, resulting in a higher hydrogen concentration that facilitates hydrogenation at the bottom side. By controlling the stabilization conditions of C−H bonds, the bottom side of graphene can be controllably hydrogenated, while the upper side remains almost non-hydrogenated. Comprehensive characterizations confirm the highly asymmetreic feature and the recoverability of graphene lattices. The degree of hydrogenation in AH-Gr can be precisely controlled by adjusting the dose of permeated protons. The spatial distributions of hydrogen atoms significantly influence electrical and magnetic properties, allowing AH-Gr transformation from semi-metal to semiconductor even at low ppm hydrogenation levels.

RESULTS AND DISCUSSION

AH-Gr in the vdW confined space

Figure 1a illustrates the hydrogenation schematics of Gr/SiO₂ via weak hydrogen plasma. At low temperature (LT), active H radicals bond to the upper side of graphene, while generated protons will permeate the graphene and recombine into H radicals, forming C−H bonds on the bottom side, leading to LH-Gr (scenario I). Noting that, the hydrogenation processes are hugely affected by temperature and concentration of H radicals [18,35,36]. As temperature increases, the formed C−H bonds are easily broken, particularly those on the upper side exposed to the free space. Once hydrogenation gets underway at moderate temperature (MT), C−H bonds on the upper side tend to break while those on the bottom side remain relatively stable because of the sub-nanometer confined space and the accumulated high-concentration of H radicals, leading to formation of AH-Gr (scenario II). If graphene is free-standing, both sides remain non-hydrogenated under MT (scenario III), serving as criteria for the hydrogenation temperature range. At high temperature (HT), C−H bonds on the bottom side of graphene also become unstable, preventing hydrogenation (scenario IV). Therefore, the crucial factors for producing AH-Gr in the confined space should be the hydrogenation temperature range under nondestructive plasma conditions.

Raman spectroscopy is used to characterize C−H bonds in graphene, and the D to G peak intensity ratios (I_D/I_G) serve as indicator of non-sp² bonds [38,39]. Figure 1b and c show Raman spectra of free-standing graphene on Cu grid and Gr/SiO₂, respectively, after undergoing hydrogenation processes at different temperatures. At LT, both exhibit high I_D/I_G ratios. The reversibility of I_D/I_G indicates that the D peaks originate from the formed C−H bonds (Fig. S1a and b). As temperature rises, the I_D/I_G ratios are both reduced, indicating that hydrogenation is suppressed by the elevated temperature. Notably, the D peak in free-standing graphene completely disappears at 350°C, but it still exists in Gr/SiO₂. Additionally, with prolonged treatment, the D peak in Gr/SiO₂ becomes more pronounced, while the free-standing graphene remains at a state of basically no-D peak (Fig. 1d and Fig. S1c). The absence of a D peak indicates that free-standing graphene is not hydrogenated. Both sides of the free-standing graphene are exposed to free space, there are no C−H bonds existing on either side. In contrast, the upper side of Gr/SiO₂ also has free space and should be free of C−H bonds, hence the D peaks should mainly originate from the C−H bonds on the bottom side, i.e. AH-Gr. As temperature exceeds 425°C, D peaks are negligible, indicating no C−H bonds formed. Figure 1e summarizes the normalized I_D/I_G ratios of free-standing graphene and Gr/SiO₂ hydrogenated at different temperatures (details in Fig. S1d and e). After undergoing the hydrogenation process at 350–425°C, only Gr/SiO₂ can be hydrogenated indicated by the distinguishable I_D, while free-standing graphene shows almost no C−H bonds. Consequently, we define the temperature ranges of hydrogenation as LT of 25–350°C for LH-Gr, where the asymmetry of hydrogenated graphene gradually increases with the increase in temperature, MT of 350–425°C and HT of >425°C for AH-Gr and non-hydrogenation, respectively.

Further studies using a scanning tunnelling microscope (STM) show the morphology of AH-Gr with I_D/I_G ratios of ∼0.2 and ∼0.8, revealing perfect hexagonal lattices, with no upper-side C−H bonds, as shown in Fig. 1f and h. The undamaged lattices without C−H domains further confirm its highly asymmetric feature, and the D peaks should mainly derive from bottom-side C−H bonds. Figure S2 shows typical STM images of graphene hydrogenated at LT, showing numerous upper-side C−H domains, which is the same as the reported surface topography of hydrogenated graphene hydrogenated at LT [19,20,22]. The C−H bonds can be also completely removed by UHV annealing. Notably, the plasma utilized for hydrogenation is too weak to introduce atomic vacancies, as confirmed by the recovered lattices and our reported results [13,35,36]. Moreover, scanning tunnelling spectroscopy (STS) (Fig. 1g and i) show the changed band structure of AH-Gr, with opened bandgaps of 0–30 meV and 120–430 meV for I_D/I_G of ∼0.2 and ∼0.8, respectively. The STS spectrum of pristine graphene is shown as the blue line in Fig. 1g for comparison. Figure 1j summarizes the bandgap distributions from dozens of spectra (Fig. S3), indicating the changed band structures and probable derivative properties.

Furthermore, we compared the wettability of pristine graphene, LH-Gr and AH-Gr through wetting angle measurements, as shown in Fig. S4. For pristine graphene, the wetting angle is ∼86°. After hydrogenation into LH-Gr (I_D/I_G = 3.7), the angle decreases to ∼60°, due to increased surface energy [40]. In contrast, AH-Gr consistently exhibits lower wetting angles and higher surface energy compared to LH-Gr with similar I_D/I_G ratios. Additionally, we further investigated the transformation between LH-Gr and AH-Gr using ex situ wetting angle measurements. Starting with AH-Gr (I_D/I_G = 2.4), LT hydrogenation increased the wetting angle from 42° to 57° (I_D/I_G = 2.6), and then to 73° (I_D/I_G = 3.5), indicating higher hydrogenation and changes in C−H bond distribution (LH-Gr). The annealing process further transforms LH-Gr back towards AH-Gr, decreasing the wetting angle from 73° to 53° (I_D/I_G = 2.7). These results demonstrate that hydrogenation temperature can tune the spatial distribution of C−H bonds.

Utilization factor of permeated protons for AH-Gr

Next, we estimate the utilization factor of aggregated H radicals for hydrogenating AH-Gr. First, we fabricate a double-layer graphene film consisting of top ¹³C-graphene and bottom ¹²C-graphene, with distinguishable Raman G and D peaks. Subsequently, this double-layer graphene is exposed to H₂ plasma at MT. The protons permeate into the vdW confined space between ¹³C-graphene and ¹²C-graphene, re-combined H radicals will mainly bond to the bottom side of the top ¹³C-graphene and the upper side of bottom ¹²C-graphene (Fig. 2a). Meanwhile, a few protons permeate ¹²C-graphene and form ¹²C−H bonds at the bottom side. Their Raman spectra with different hydrogenation durations are plotted in Fig. 2b, and the isotopic ¹³C-graphene presents the redshifted ¹³G, ¹³D and ¹³2D peaks. Although the ¹³G intensity is 1/3 weaker than ¹²G, the ¹³D is still twofold stronger than ¹²D. Figure2c plots the relative ratio for ¹³C−H and ¹²C−H bonds, which is normalized by the ratio of I_13D/I_13G to I_12D/I_12G. We find that the concentration of ¹³C−H is ∼80%, representing the utilization factor for the bottom side hydrogenation of top AH-Gr.

Additionally, we fabricate a triple-layer film consisting of top ¹³C-graphene, middle ¹³C-graphene and bottom ¹²C-graphene. After MT hydrogenation for different durations, we use the ¹³D and ¹²D peaks to estimate the hydrogenation depth through permeating protons, as illustrated in Fig. 2d. Figure 2e displays the Raman spectra of the triple-layer film following varying hydrogenation durations. The weak intensity of ¹²D peak indicates minimal hydrogenation in the bottom ¹²C-graphene, with the concentration of ¹²C−H estimated to be <6.7% (Fig. 2f). The proton permeation depth here is a single or double-layer graphene lattice, consistent with the H₂ bubble encapsulated by single or double-layer graphene [36].

Quantification of H/C ratios in AH-Gr

We utilize exfoliated hBN as a denser substrate to create homogeneous vdW confined spaces, where recombined H radicals are assumed to efficiently form C−H bonds [35,36,41,42]. Figure 3a illustrates the schematic of AH-Gr on hBN, the distance of vdW confined space is typically 0.34–0.65 nm for various 2D materials [10,33,43,44]. After transferring monolayer graphene onto few-layer hBN [45], we perform hydrogenation, Gr/hBN exhibits a significantly enhanced hydrogenation effect compared to Gr/SiO₂, and with an extended temperature threshold for AH-Gr up to 460°C (Fig. S5).

To quantitatively estimate the H/C ratios of AH-Gr, we establish an analogy approach. Given that hBN and graphene exhibit lower permeability to protons compared to amorphous SiO₂ [33], ∼80% of the permeated protons are utilized for generating AH-Gr. This allows us to calculate the C−H ratios via the dose of the permeated protons following the equation H/C = Γ_P  × τ/n_C, where Γ_P is the permeating rate of proton through graphene and is estimated to be ∼940 s⁻¹μm⁻² under the same plasma condition [36], τ is the proton permeation time, and n_C is the carbon atom density in graphene of 3.82 × 10¹⁹ m⁻².

Figure 3b and c plot the typical Raman spectra of Gr/hBN after different permeation times and the extracted I_D/I_G versus calculated H/C ratios, respectively. After 5 s of hydrogenation, the I_D/I_G ratio reaches 0.6, corresponding to an H/C ratio of 98 ppm, with the average distance between neighboring C−H bonds (L_H) being ∼16.3 nm. This distance aligns closely with Raman analysis [38], but is slightly larger than that established by e-beam irradiated HSQ [28]. We then compare the correlation between H/C ratios and hydrogenation duration for Gr/hBN and Gr/SiO₂. Notably, AH-Gr on both substrates with the same H/C ratios exhibit identical I_D/I_G, confirmed by transferring the same AH-Gr films to SiO₂/Si and hBN (Fig. S6). This indicates that Raman I_D/I_G for AH-Gr only derives from H/C ratios, and are irrelevant to the substrate. Figure 3d shows that the average hydrogenation rate is ∼20 ppm/s for Gr/hBN, significantly higher than the 0.2 ppm/s for Gr/SiO₂. The two orders magnitude difference indicating a narrower space of Gr/hBN with denser H radicals. The higher I_D/I_G ratio for AH-Gr/SiO₂ requires a longer hydrogenation duration (Fig. 1e). Additionally, the vdW confined space enhances the homogeneity of AH-Gr, as confirmed in Fig. S7. Figure 3e plots the I_D/I_G spatial distributions of AH-Gr/SiO₂ and AH-Gr/hBN. Although both exhibit similar H/C ratios with an average I_D/I_G ratio of 0.67, the distribution for Gr/SiO₂ is notably broader, with ∼90% of counts between 0.52 and 0.79, whereas Gr/hBN is more homogeneous with ∼90% of counts between 0.55 and 0.76.

X-ray photoelectron spectroscopy (XPS) is employed to analyze the elemental composition and chemical states. Figure 3f shows the XPS spectra of AH-Gr with H/C ratio of 160 ppm, with distinct subpeaks for different components. The C−H bonds (sp³-1) peak is located at 285.0 eV, C−C bonds (sp²) of graphene at 284.5 eV. Through peak differentiation, the C−H (sp³-1) component is extracted to estimate the H/C ratio. The XPS spectra of as-transfer graphene are shown in Fig. S8, where no C−H (sp³-1) peak is visible. Additional component peaks, like C−C (sp³-2) at 285.9 eV, C−O at 286.7 eV and O=C−O at 288.9 eV correspond to the quaternary carbon, methoxy group and carboxyl group, respectively, which may originate from the residual polymethyl methacrylate (PMMA) [46].

Electrical and magnetic properties of AH-Gr with different H/C ratios

To investigate the impact of hydrogenated graphene with varying C−H bond distributions, we perform electrical transport measurements on films hydrogenated at different temperatures, all with an identical Raman I_D/I_G ratio of ∼0.5. Figure 4a and Fig. S9a plot gate voltage (V_g) dependent four-probe longitudinal resistances (R_xx). All the electrical transport curves remain bipolar, the maximum R_xx at charge neutrality point (CNP) increasing with hydrogenation temperature, with values of 7.5, 14, 31,104 and 107 kΩ for 25,150, 200,300 and 370°C, respectively. This indicates a transformation of samples from LH-Gr to AH-Gr. Notably, graphene films hydrogenated at 300°C have similar resistance to AH-Gr prepared at 370°C, indicating that hydrogenation at 300°C approximates AH-Gr.

Figure 4b shows that R_xx of AH-Gr increases with H/C ratios. At H/C ratio of 175 ppm, R_xx is 216 kΩ at CNP, about 20-fold higher than pristine graphene. This substantial increase should be attributed to the C−H scattering sites and the enhanced inversion asymmetry. Comparatively, for AH-Gr with 34 ppm H/C ratios, R_xx demonstrates a slight increase of 15%, indicating weak hydrogenation, where partial graphene exhibits the properties of a semiconductor. Figure 4c presents R_xx statistics of AH-Gr with various H/C ratios at for CNP (green circles), moderate doping (yellow circles, V_CNP+30 V) and heavy doping (red circles, V_CNP+50 V). Generally, R_xx at CNP is ∼2 orders of magnitude higher than under both moderate doping and heavy doping. Moreover, for an H/C ratio of 474 ppm, R_xx at CNP reaches 3.4 MΩ, ∼300 times higher than pristine graphene. A significant trend shows that R_xx rapidly increases from 0 to 40 ppm H/C ratio, indicating a weak C/H region with coexisting semi-metal and semiconductor regions. Furthermore, higher ratios indicate bandgap opening in all carbon lattices.

Next, we calculate the carrier mobility of AH-Gr with different H/G ratios, summarized in Fig. 4d. Carrier mobilities decrease with increasing H/C ratio, consistent with the reduced conductivity. It drops sharply from ∼12 000 to ∼2000 cm²V⁻¹s⁻¹ within the weak hydrogenation degree, then gradually declines. The AH-Gr with an H/C ratio of ∼174 ppm (STS-measured bandgap of 240 meV) maintains high mobility ∼1000 cm²V⁻¹s⁻¹, while with ∼480 ppm, mobility decreases to ∼40 cm²V⁻¹s⁻¹. Notably, the average distance between the neighboring C−H bonds (L_H) is calculated to be ∼21 nm for AH-Gr with an H/C ratio of 40 ppm, which may be the sparsest distance of hydrogen distribution that can open a bandgap for surrounding carbon lattices.

Figure 4e shows the temperature-dependent resistance (R-T) measurements of AH-Gr. For AH-Gr with weak H/C ratio of 6 ppm, R_xx at CNP drops from 110 to 1.5 K, then increases from 110 to 295 K, similar to pristine graphene (Fig. S9b–e). For higher H/C ratios (54–278 ppm), R_xx decreases with rising temperature, indicating typical semiconductor behavior. We further fit their bandgaps from the Arrhenius equation R_xx = exp[E_g/(2k_BT)], where E_g is the bandgap, k_B is the Boltzmann constant and T is the temperature [21], as plotted in Fig. 4f, marked by blue dots, with the dashed line being a guideline to the trend. The fitted E_g is ∼1 order magnitude smaller than the STS value, likely due to the thermally excited carriers and substrate-induced pseudo-bandgap [48]. By comparing LH-Gr hydrogenated at 200°C, as redrawn in Fig. 4f by green dots, AH-Gr opens a larger bandgap while they have the same hydrogenation concentration. The R-T measurements of LH-Gr are shown in Fig. S9f.

Magnetotransport measurements of AH-Gr films with varying C−H bonds monitor the evolution changes of the quantum Hall effect (QHE) [49]. Figure 5a and b plot R_xx and Hall conductivity (σ_xy) of AH-Gr films under out-of-plane magnetic field (B_┴) of 7 T at 1.5 K. The R_xx in as-transferred graphene film fall to zero, and σ_xy reach ± 2 e²/h plateaus near V_g = ±7 V, corresponding to Landau filling factors ν = ±2. Meanwhile, only AH-Gr with weak H/C ratio (34 ppm) shows zero R_xx, and σ_xy remains plateau shaped at ν = ±6. For other AH-Gr films, R_xx cannot reach zero and σ_xy plateaus mismatch filling factors, because of the hydrogen-induced broadening of Landau levels [49]. Additionally, the σ_xy of AH-Gr with H/C ratios ranging from 84–175 ppm show slight slopes near CNP, corresponding to the R_xy peaks (Fig. S9g). These dramatically changed peaks are expected to originate from the disruption of topological edge states induced by C−H bond-induced disorders at CNP [50]. Moreover, weak H/C ratio AH-Gr (34 ppm) still shows topological edge states, indicating interconnections between semimetal and semiconductor regions.

Magnetoresistance (MR) of AH-Gr at CNP with different H/C ratios is performed under weak B_┴ at 1.5 K. Figure 5c shows that MR decreases with enlarged B_┴, and the change rates of MR with B_┴, that is (ΔR)/R = [R(B_┴)—R(B_┴ = 0)/R(B_┴ = 0)]. Using weak localization (WL) theory [51], we fitted MR to obtain a phase coherence length (L_φ) of AH-Gr, with relationship to L_H plotted in Fig. 5d. L_φ of AH-Gr increases from 34 to 233 nm as L_H increases from 13 to 41 nm, and they have an approximate power function dependence of L_φ ∼ L_H^1.5. The reason is that C−H bonds can act as scattering sites causing charge carrier de-coherence [52].

CONCLUSIONS

In summary, we develop a confined space assisted production of AH-Gr via proton permeation at an MT of 350–425°C. LT hydrogenation leads to LH-Gr, while HT terminates hydrogenation. AH-Gr films are defect-free with opened bandgaps. HT annealing can remove all the formed C−H bonds, recovering their hexagonal lattices. The H/C ratios can be quantitatively controlled by calculating the permeated protons in the vdW confined space between graphene and hBN. AH-Gr on hBN shows a higher hydrogenation rate of 20 ppm/s, which is 2 orders of magnitude greater than on amorphous SiO₂. Due to larger structural asymmetry, AH-Gr exhibits higher resistance and larger bandgap than LH-Gr at similar hydrogenation degree. When H/C ratios exceed 40 ppm, AH-Gr changes from semi-metal to semiconductor. Magnetotransport measurements indicate that hydrogenation-induced disruption affects the QHE and reduces the L_φ. This study re-examines the production and features of hydrogenated graphene, expands the applications of vdW confined space, and presents a quantitative hydrogenation approach for AH-Gr with controlled spatial distribution.

FUNDING

This work was supported by the National Natural Science Foundation of China (52425203, 12104218, 11904163, 12374183 and 92165205), the Natural Science Foundation of Jiangsu Province (BK20240008), the National Key Research and Development Program of China (2021YFA1400403), the Innovation Program for Quantum Science and Technology (2021ZD0302800), the China National Postdoctoral Program for Innovative Talents (BX2021120), the Xiaomi Foundation, the Postdoctoral Fellowship Program of China Postdoctoral Science Foundation (CPSF) (GZC20231093), the Jiangsu Funding Program for Excellent Postdoctoral Talent (2023ZB553) and the Program A for outstanding PhD Candidates of Nanjing University (202401A04).

AUTHOR CONTRIBUTIONS

L.G. and G.Y. designed and supervised the experiments. X.H. performed hydrogenation, AFM and Raman. W.L. and X.H. performed the device fabrication and transport measurements. H.Z. and J.X. assisted in graphene growth and transfer, hydrogenation, AFM and Raman. K.W. and L.W. assisted in fabrication of exfoliated Gr/hBN heterostructures. L.Z. and S.L. performed the STM and STS. L.G., G.Y. and X.H. analyzed the data and wrote the manuscript, J.X., L.W. and S.L. revised it, and all authors commented on it.

Conflict of interest statement. None declared.

REFERENCES

Author notes