betaFeedback

WorksPhysics and Astronomy

Hydrogen spillover and storage on graphene with single-site Ti catalysts

Abstract

Hydrogen spillover and storage for single-site metal catalysts, including single-atom catalysts (SACs) and single nanocluster catalysts, have been elucidated for various supports but remain poorly understood for inert carbon supports. Here, we use synchrotron radiation-based methods to investigate the role of single-site Ti catalysts on graphene for hydrogen spillover and storage. Our in-situ angle-resolved photoemission spectra results demonstrate a bandgap opening and the X-ray absorption spectra reveal the formation of C-H bonds, both indicating the partial graphene hydrogenation. With increasing Ti deposition and H2 exposure, the Ti atoms tend to aggregate to form nanocluster catalysts and yield 13.5% sp3-hybridized carbon atoms corresponding to a hydrogen-storage capacity of 1.11 wt% (excluding the weight of the Ti nanoclusters [1]). Our results demonstrate how a simple spillover process at Ti SACs can lead to covalent hydrogen bonding on graphene, thereby providing a strategy for a rational design of carbon-supported single-site catalysts.

#Contributed equally

*Correspondence and requests for materials should be addressed to C.-H. Chen and C.-L Wu

(E-mail: [email protected] and [email protected])

Chemical catalytic reactions involving hydrogen, which are ubiquitous in the chemical industry, require hydrogen molecules (H2H_2) to be dissociated into HH atoms at a catalytic site. The hydrogen spillover phenomenon was first observed in 1964 in the heterogeneous catalysis system WO3/PtWO_3/Pt [2]. Many subsequent experimental results on hydrogen spillover revealed that H2H_2 could be easily dissociated on a metal catalyst (although the hydrogen-hydrogen bond in H2H_2 is strong and therefore unlikely to break under normal reaction conditions), followed by migration of the HH atoms to an adjacent support surface (such as reduced metal oxides) and chemisorption there. When considering the optimization of a catalyst, common strategies to achieve the desired activity and selectivity consist in adjusting the crystallographic structure [3-4], particle size [5-10], and geometry of the catalyst [11-14]. Still, synthesizing heterogeneous catalysts with isolated active sites having efficiency at an atomic level with a clarified mechanism remains a critical challenge. Various support materials also affect catalyst optimization significantly. Some examples indicate that strong interactions with the support might enhance the catalytic performance of transition-metal catalysts via charge-transfer effects [15]. Several groups have successfully prepared transition-metal-based single-atom catalysts (SACs) to minimize the amount of metal required to catalyze the hydrogen spillover efficiently, but the catalytic activity still has much room for improvement [16,17].

Since the first report on hydrogen spillover and storage in carbon nanotubes by Dillon et al. [18], carbon supports represent an obvious alternative to metal oxides to exploit hydrogen spillover for storage, even though carbon atoms do not participate directly in the reaction as catalyst, because of their higher inertness relative to conventional oxide supports. Achieving a fundamental understanding of carbon materials as catalyst support is a challenge due to their structural and chemical complexity, which might explain the difficulties in obtaining reproducible adsorption capacities from various studies with volumetric and gravimetric measurements [19,20]. The migration of spillover-dissociated hydrogen atoms is kinetically unfavorable on graphene or graphite [21], which means that hydrogen storage on graphene supports is restricted to small distances around the catalytic sites. How many hydrogen atoms could absorb is thus an important question, which must be addressed for practical applications of H-storage. Inspired by this objective, based on our previous work showing that single Ti atoms adsorb on graphene at the energetically favorite hollow sites [22], and the structural variation upon Ti deposition on graphene and its hydrogen uptake studied by scanning tunneling microscope [23-26], we utilized Ti atomic deposition with a view to mimic SACs on graphene. Unlike previous studies, which used dissociated hydrogen sources (as shown in the Supporting Information Table S1), here we demonstrate that atomically dispersed Ti atoms on crystalline graphene can produce an efficient spillover for hydrogen, which is then chemisorbed around the anchored Ti SACs. The evaluated H storage capacity exceeds that of Ti nanocluster catalysts at a similar Ti loading.

A well-defined single-crystalline graphene support is suitable as a platform to investigate hydrogen spillover from a single-site Ti catalyst, which dissociates H2_2 molecules into H atoms which then migrate to the support graphene (as the receptor of H atoms, schematic illustration in Figure 1).

Schematic diagram of the spillover effect of Ti on graphene

Figure 1. Schematic diagram of the spillover effect of Ti on graphene, which involves dissociation of molecular hydrogen near 300 K, migration of hydrogen atoms on epitaxial graphene and their chemisorption.

Hydrogen atoms migrating to the support layer, assisted by the single-site Ti catalyst, can provide substantial storage of atomic hydrogen on graphene in terms of storage quantity. Graphene used in this work was epitaxially grown via thermal decomposition in a resistively heated cold-wall reactor (Aixtron HT-BM) on a hydrogen-etched 4H-SiC (0001) substrate. To achieve a sufficient monolayer uniformity, we grew pristine crystalline graphene in an Ar atmosphere at 1600 K and 780 mbar for 10 min. Single-site catalysts were formed by depositing a sub-monolayer amount of Ti atoms (0.03 to 0.5 ML), calibrated with a quartz crystal microbalance (QCM). The Ti coverage is defined with respect to hexagonal close-packed bulk Ti [23]. Angle-resolved photoemission spectra (ARPES) show a Ti adatom-induced renormalization of the band structure for low Ti coverage (~0.03 ML), as shown in Figure 2. This was attributed to a strong Ti 3d3d and C 2pz2p_z orbital hybridization of isolated Ti-atoms at the graphene hollow sites [22]. After exposing the sample to 4.5 L of H2_2 near 300 K (1 L, exposure to 1×10−61 \times10^{-6} Torr for 1 s), the significant ARPES spectral contrast in Figure 2(A) shows a band gap opening of 280 meV in the graphene. This condition indicates that hydrogen atoms chemisorbed on graphene, which is evidence for the dissociation of gaseous H2H_{2} molecules and hydrogen spillover by isolated Ti SACs. Moreover, without Ti, neither band structure modification nor renormalization was observed in the ARPES spectrum of graphene after H2H_{2} exposure, as shown in Supporting Information Fig. S1. Thus, the Ti catalysis facilitates the dissociation of molecular H2H_{2} and the hydrogenation of graphene near 300 K in vacuum.

ARPES measurements of pristine, Ti-doped, and hydrogenated graphene, with bandgap versus hydrogen coverage

Figure 2. ARPES measurements performed on pristine epitaxial graphene, after Ti deposition, and after a subsequent dose of H2_2 molecules (H2_2 = 4.5 L) near 300 K. (A) Band structure of pristine and Ti-doped graphene for 0.03 ML. The sample was then exposed to a 4.5 L dose of H2_2. The data were taken along direction K→Γ\mathrm{K}\to\Gamma in the vicinity of the K-point. White dashed lines indicate the energy dispersion fitted by the momentum dispersion curves (MDC). (B) Data points from previously performed DFT calculations [27,28,29] were used to obtain the extent of graphene hydrogenation from the measured band gap induced by H-C bonds. The dependence of the bandgap of graphene on H coverage is ΔEg=A2η2+A1η+A0\Delta E_g = A_2\eta^2 + A_1\eta+ A_0, with η=H/C\eta= H/C, A2=2.49A_2 = 2.49 eV, A1=1.47A_1 = 1.47 eV, and A0=9.67×10−2A_0 = 9.67 \times10^{-2} eV. The hydrogen coverage η\eta corresponding to graphene band gap Eg=0.28E_g=0.28 eV is obtained via two different approaches, with a lower limit of 7.3 % and an upper limit of 16.2 % using a fit of DFT+PBE and DFT+GGA data, respectively. Insets show angle-integrated curves of energy distribution (EDC) in the vicinity of the K point taken before (red) and after (blue) 4.5 L H2_2 dosage for graphene with 0.03 ML Ti.

To quantitatively understand the spillover effect on graphene, we monitored the band gaps of graphene before and after exposure to H2H_{2} using energy-dispersion curves (EDC) extracted from ARPES (insets of Figure 2 (B)). Calculations [27-29] yielded valuable information on the origin of the graphene band-structure modulations induced by CC–HH bonds, from fully hydrogenated insulating graphene sheets to partially hydrogenated graphene displaying semiconducting properties, with band gaps depending on the H coverage of graphene. The band gap value can thus serve to quantify the number of hydrogen atoms incorporated in partially hydrogenated graphene. Based on the DFT-based theoretical results shown in Figure 2(B) [27-29], a quadratic relation ΔEg=A2η2+A1η+A0\Delta E_g = A_2\eta^2 + A_1\eta+ A_0 is obtained. η\eta is the absorbed hydrogen coverage on graphene (H/C value, i.e., the H storage capacity of graphene), A2=2.49A_2 = 2.49 eV, A1=1.47A_1 = 1.47 eV, and A0=9.67×10−2A_0 = 9.67 \times10^{-2} eV. The band-gap opening (ΔEg\Delta E_g) increases from about 0 to 1.5 eV as η\eta increases from 0 to 50 %. The H-storage capacity η\eta of partially hydrogenated graphene with band gap 280 meV is estimated to be about 10.6 % (0.88 wt%).

As indicated in Figure 2(B), this value is in between a lower limit of 7.3 % (0.6 wt%) obtained with the DFT+PBE approach and an upper limit of 16.2 % (1.33 wt%) with the DFT+GGA approach [27-29].

Previous theoretical and experimental work clearly showed that charge transfer between the catalyst and the support generates a noticeable impact on the electrocatalytic activity of the catalyst [30,31].

As we have concluded in our previous work [22], the anchored Ti SACs contributes a pronounced electron transfer from Ti to the graphene support (0.45 e−e^{-} per Ti atom), which leads to an n-type doping of graphene. Thus, we attribute the high storage activity observed here to the charge transfer between Ti atoms and graphene, which provides a more efficient and safer method for using SACs on graphene.

To explore the nature of the formation of C−H bonds accompanying hydrogenation, we recorded X-ray absorption spectroscopy (XAS) data on Ti-decorated graphene after exposure to H2H_2, following the same sample preparation procedure as for the ARPES experiment. We recorded the XAS at an X-ray beam incident angle of 0° relative to the surface normal. The C−H* resonance region is presented in Figure 3, for Ti-doped (~0.03 ML) and H2H_2-dosed (~ 4.5 and 49.5 L) samples. We plot the spectral differences, i.e. spectra from which the spectrum of pristine graphene is subtracted. A prominent spectral feature is clearly visible centered at 287.5 eV, which is attributed to the C−H* resonance [32,33]. Its intensity increases significantly after H2H_2 dosage, while the Ti-decorated graphene (without hydrogen dosing) shows no sign of C–H bonds. This result clearly indicates the formation of C–H bond after exposing the Ti SAC-loaded graphene to hydrogen, implying the spillover of hydrogen onto the graphene support. The full-range carbon KK-edge spectra are shown in the Supporting Information Section 4.

C $K$-edge X-ray absorption spectra showing difference spectra for Ti-doped, 4.5 L H$_2$-dosed, and 49.5 L H$_2$-dosed samples

Figure 3. C–H* resonance region of the Carbon KK-edge X-ray absorption spectra for Ti-doped (black curve), for 4.5 L H2_2-dosed (red curve), and for 49.5 L H2_2-dosed (blue curve) samples. We show difference spectra, i.e., spectra minus the spectrum of pristine graphene. The shaded area indicates the position of the C–H* resonance. See Supporting Information § 6 for full spectra.

To seek the optimal H storage capacity, we recorded in situ ambient pressure X-ray photoemission spectra (APXPS) with a larger amount of Ti catalyst loading and under a larger H2H_2 exposure (1 mbar), to monitor the C–H bonding structure including the underlying substrate. This is useful to clarify their roles in the progressively hydrogenated Ti-decorated graphene under a nearly ambient gaseous pressure. Figure 4 shows the synchrotron-radiation excited C 1s core-level spectra recorded in situ near 300 K taken from pristine graphene, Ti-deposited graphene (Ti ~ 0.5 ML), and Ti-deposited graphene under H2H_2 exposure for one hour (Ti + H2H_2). By analyzing the measured C 1s core-level spectra with a Doniach-Sunjic line shape and comparing them to reference spectra of nonhydrogenated crystalline graphene grown on a SiC substrate, both measured here and from literature, the amount of hydrogen stored in our graphene samples can be precisely determined. Details of the fitting results of the C 1s APXPS spectra are listed in Table S3 in the Supporting Information Section 6. The C 1s core-level spectra of pristine and Ti-deposited graphene are well fitted with one graphene. sp3\mathrm{sp}^{3} bonding peak (Gr at 284.4 eV), one SiC\mathrm{SiC} bulk component peak (B at 283.3 eV), and two interfacial components S1 and S2 related to localized reconstruction bonds, they represent a (63×63)R30∘(6\sqrt{3} \times6\sqrt{3})\mathrm{R}30^\circ reconstructed SiC(0001)\mathrm{SiC}(0001) surface formed underneath the epitaxial graphene as a buffer layer [34]. A Shirley background was included in the fitting procedure. The interfacial state S1 at 284.2 eV can be identified with the covalent bonding between the buffer layer and SiC\mathrm{SiC} substrate, and S2 at 285.0 eV is due to the sp2\mathrm{sp}^{2}-bonded C atoms within the buffer layer. These two interfacial states S1 and S2 have a constant area ratio of ∼1/2\sim1/2 [34,35]. They would not be affected by covering the sample with Ti catalysts and H atoms, which is consistent with the XPS results of epitaxial graphene on SiC\mathrm{SiC} (0001) substrate [35-37]. With Ti deposition and hydrogenation, the C 1s core level shows a progressive attenuation due to the Ti metal covering and chemisorption of atomic hydrogen. Hydrogen bonding to graphene implies that the H atom induces an out-of-plane C−H bond (Gr1, hydrogenated C atom) and can be detected in the C 1s spectrum at a higher binding energy position of 285.0 eV than sp2\mathrm{sp}^{2} graphene (G) at 284.4 eV, which is consistent with the previous result of hydrogen adsorption in graphene [38]. Furthermore, another component appears in the spectrum (Gr2), which is due to non-hydrogenated C atoms adjacent to C−H bonds, as illustrated in Figure 4. Therefore, the line shape of Gr changes gradually upon hydrogenation, and two shoulder peaks appear with respect to Gr, which are 0.6 eV higher (Gr1) and 0.2 eV lower (Gr2) in binding energy [39-42]. Thus, the Gr1 peak area is highly reliable for quantitative determination of the hydrogen content, using η(%)=Gr1/(Gr+Gr1+Gr2)\eta(\%) = \mathrm{Gr1}/(\mathrm{Gr}+\mathrm{Gr1}+\mathrm{Gr2}). We found here hydrogenated graphene with a H coverage of about η∼13.5±0.5\eta\sim13.5 \pm0.5 % (or 1.11±0.041.11 \pm0.04 wt%), which is larger than previous reports about atomic hydrogenation of crystalline graphene (η∼8.7\eta\sim8.7 to 9 %) [41,43].

Image not extracted

Figure 4. From left to right: APXPS C 1s1s spectra of pristine, Ti-deposited (0.5 ML), and Ti + H2_2 exposed graphene (H2_2 1 mbar for 1 hour). Spectra were deconvoluted using Gaussian-broadened Doniach-Sunjic (DS) functions including Shirley background. Gr, B, and S1, S2 indicate C 1s1s chemical environments for graphene, SiC bulk, and interfacial reconstruction (at the buffer layer interface), respectively. Component Gr1 represents hydrogenated C atoms (C–H bond) of graphene. Gr2 represents not-hydrogenated C atoms next to C–H bonds, as illustrated in the inset. From the integrated area of the relative constituents of the graphene C 1s1s C–H components (Gr, Gr1, Gr2), we estimate η∼13.5\eta\sim13.5 % hydrogen coverage.

peakPosition (eV)Intensity (a.u.)Area
PristineGr284.456141.780760
PristineB283.457071.470301
PristineS1284.79098.924696
PristineS2285.232528.546114
TiGr284.536700.052768
TiB283.337000.647492
TiS1284.85982.116239
TiS2285.122000.231193
Ti+H2\mathrm{H}_2Gr284.420000.022187
Ti+H2\mathrm{H}_2B283.327600.134834
Ti+H2\mathrm{H}_2S1284.74500.111563
Ti+H2\mathrm{H}_2S2285.116499.923302
Ti+H2\mathrm{H}_2Gr1285.051005378
Ti+H2\mathrm{H}_2Gr2284.11500012189

Table S3. Summary of fitting peak position, relative intensity, and relative area of every peak of the CC 1s APXPS spectra (pristine, Ti, and Ti+H2\mathrm{H}_2) after baseline subtraction.

The hydrogen storage by spillover is a process of adsorption of atomic hydrogen, in which the numbers of catalytic sites and H adsorption sites determine the storage capacity. The density of SAC active sites, i.e., the hollow sites occupied by Ti atoms, were estimated based on the Ti coverage of 0.03 ML, which corresponds to 4.0×1013 cm−24.0 \times10^{13}\ \mathrm{cm}^{-2} for the Ti SAC density. How far H atoms can migrate from the Ti catalyst is dominated by the macroscopic diffusion coefficient (LDL_\mathrm{D}). According to a 2D-random-walk model for diffusion, LDL_\mathrm{D} can be expressed with a potential diagram in the classical limit. The diffusion length of H on graphene was estimated with this equation,

LD=(D0τ0)0.5exp⁡(Ead−Edif2kBT)(1)L_\mathrm{D} = (D_0\tau_0)^{0.5}\exp\left(\frac{E_\mathrm{ad}-E_\mathrm{dif}}{2k_\mathrm{B}T}\right) \tag*{(1)}

in which EadE_\mathrm{ad} and EdifE_\mathrm{dif} are the adsorption and diffusion energies of H on graphene, which are 78 KJ/mol and 91 KJ/mol, respectively [44]. D0D_0 and τ0\tau_0 are the pre-exponential factors of diffusion and residence period of C–H bonds on graphene. Their values are 3276 Å2^2/ps and 1.52 ps, respectively [45]. Near 300 K, the diffusion length LDL_\mathrm{D} is thus about 2.7 Å, which is only twice the C–C bond length, a=1.42a = 1.42 Å. As the Ti catalyst atoms are anchored on the energetically favorable graphene hollow sites, the spillover H atoms diffuse isotropically and are restricted to small distances around the Ti catalytic center on the graphene surface, as depicted in Figure 5 (A). This schematic illustration reveals that graphene hydrogenation occurs only on the first nearest-neighbor sublattices (i.e. one Ti surrounded by six H atoms). As shown in Supporting Information § 9, the H-storage capacity η\eta of Ti SACs with an amount 0.03 ML is estimated to be about 6.3 %. This value is closed to the low limit value (η∼7.3\eta\sim7.3 %) extracted from the partially hydrogenated graphene band gap and near that of most metal catalysts on various carbon supports, especially the reported atomic metal species on carbon supports under high pressure [20,46,47].

Image not extracted

Figure 5. (color). (A) Schematic diagram of single-atom to single-cluster catalyst formation. Low to high Ti deposition was explored to obtain well-isolated atoms or nanocluster catalysts on graphene. From the spillover of H atoms for an isolated Ti atom and Ti nanocluster catalyst, the 2D-diffusion length (LDL_{\mathrm{D}}) and nanocluster size (diameter RCR_{\mathrm{C}}) can be used to estimate the H-storage capacity for small and large Ti coverages. (B) Calculated H storage of Ti SACs and Ti nanoclusters with truncated bipyramid shape versus number NN of Ti atoms. (C) Schematic illustration of full H storage (η\eta=100%) on graphene with the help of Ti SACs through the formation of a Ti1H6\mathrm{Ti}_{1}\mathrm{H}_{6} structure.

Furthermore, when the deposition of Ti reaches 0.5 ML, H-storage capacity η\eta increased non-linearly to 13.5 % according to the APXPS result. This indicates that for Ti deposition above 0.1 ML, the single Ti atoms are generally susceptible to aggregate to become nanoclusters on graphene, consistent with previous STM studies [23]. For Ti nanoclusters adopting a hexagonal close-packed cluster configuration (for details, see Supporting Information § 9) [48], the number of C−H bonding sites can be estimated to be within a concentric circular area from the nanocluster boundary to an outer diameter, which is larger by LDL_{\mathrm{D}}. This is shown in Figure 5 (A) for the case of a Ti nanocluster with Ti atom number N=10N=10. With increasing Ti coverage on graphene, the number of H atoms stored around each Ti nanocluster increases sub-linearly, while the H-storage performance per Ti atom decreases sharply, as shown in Figure 5 (B) (see also Supporting Information Table S2). Thus, due to the coalescence of Ti atoms into larger nanoclusters, the hydrogen storage capacity η\eta does not increase linearly with the Ti catalyst loading on graphene. Furthermore, to boost η\eta to 100 % on graphene, the full utilization of Ti SACs on graphene (6.3×10146.3 \times10^{14} sites cm−2\mathrm{cm}^{-2} for 0.48 ML) with a Ti1H6\mathrm{Ti}_{1}\mathrm{H}_{6} structure would need to form, as illustrated in Figure 5 (C). Even though the H-storage capacity at ambient temperature and in vacuum reported here is still below the DOE onboard vehicular target (~7.5 wt%), simultaneously enhancing the number of Ti SACs on graphene and high-pressure H2\mathrm{H}_{2} loading is anticipated to increase the performance for practical onboard hydrogen storage for vehicular fuel cells. Also, the H-storage capacity shown here is significantly larger than all reported results of high-pressure hydrogen-storage measurements on carbon nanotubes, graphene, and Pt-decorated graphene [1] [20, 46, 47].

In conclusion, we designed and implemented hydrogen spillover from single-atom to single-cluster catalysts by precisely controlling the Ti coverage on graphene. The atomic nature of C−H\mathrm{C}-\mathrm{H} bonds on graphene and the optimal amount of H storage capacity were spectrally characterized with ARPES and XAS. Our photoemission data were quantitatively explained by a short migration distance of H atoms diffusing at room temperature on graphene. Our finding provides a general rule that will enable a rational design of carbon-supported physisorbed metal catalysts for chemisorbed H-storage via spillover.

Supporting Information

This material is available free of charge via the internet at http://pubs.acs.org.

Details of experimental methods, crystal growth, and additional experimental results and discussion.

Acknowledgements

We thank Dr. Yeu-Kuang Hwu of Academia Sinica for providing a mu-metal chamber for the ARPES measurements and NSRRC beamline staff for skillful assistance during SR experiments. This work was funded by Taiwan Ministry of Science and Technology.

ToC Figure:

Image not extracted

Figure 6 (PDF p. 27).

Experimental methods

Crystal growth

Graphene was grown via thermal decomposition in a resistively heated cold-wall reactor (Aixtron HT-BM) on a hydrogen-etched 4H-SiC (0001) substrate. The growth was performed under Ar atmosphere, at 1600 K and 780 mbar for 10 min. The quality and homogeneity of the monolayer graphene was assessed with a combined analysis of Raman spectra and atomic-force microscopy.

Before ARPES, XAS, and APXPS measurements, the graphene samples were annealed at 900 K for more than 7 h under UHV conditions.

Details of measurements

ARPES was used to probe the electronic structure of pristine graphene and of Ti-decorated graphene, with and without the dosage of H2H_2 molecules. APXPS measurements were performed to determine the percentage of hydrogenation of Ti-decorated graphene as H2H_2 molecules adsorbed. After dosing the sample under continuous flow of H2H_2 for various time periods, the ARPES and APXPS data were measured after the chamber was pumped down to the base pressure. Analyzers (SPECS Phoibos 150 and SPECS Phoibos 150 NAP) were utilized to measure the ARPES and APXPS spectra, respectively.

Measurements were performed at Taiwan Light Source (TLS) beamlines 08A1 and 24A1 of the National Synchrotron Radiation Research Center (NSRRC) in Hsinchu, Taiwan. All measurements were performed near 300 K. Energy and momentum resolution are estimated to be better than 150 meV and ±0.005\pm0.005 A˚−1\mathring{\mathrm{A}}^{-1} for ARPES. Energy resolution is estimated to be better than 280 meV for APXPS.

The XPS spectra were deconvoluted with the software UNIFIT 2013. The peak area uncertainties were carefully analyzed according to R. Hesse et. al. [1].

The XAS spectra were recorded at TLS beamline 09A2. The incident angles of the linearly polarized undulator radiation were 0∘0^\circ and 45∘45^\circ with respect to the graphene surface normal. We recorded the sample current to obtain the XAS spectra in the total-electron-yield mode. All XAS spectra were recorded in on-the-fly scan mode to prevent radiation damage. The photon energy was calibrated against the 1s1s–π∗\pi^{*} resonance of HOPG at 285.5 eV.

A QCM was used to determine the Ti coverage, as a QCM has a narrow resonance that leads to highly stable oscillations and great accuracy, even in the case of minute coverages. To duplicate the deposition conditions, we placed the QCM in front of the Ti evaporator at the same distance, deposition angle and base pressure as the sample.

Band diagrams measured by ARPES on bilayer graphene (BLG) and H2H_{2}-exposed BLG (without Ti).

Image not extracted

Figure S1. ARPES (left) and MDC (right) measured on (A) pristine bilayer graphene and (B) after exposure to 4.5 L of molecular hydrogen. No Ti was deposited prior to hydrogenation. The results show that the addition of molecular hydrogen near 300 K does not alter the band structure of bilayer graphene (without Ti).

Summary of experimental hydrogen storage capacity in graphene

Table S1

H sourcesupportcatalystT (k)PressureCapacity (% and wt %)Ref.
H₂ plasmaExfoliated grapheneNoRT1 torr16.67 %1
H₂ plasmaCVD grapheneNoRT0.015 torr33 %2
H₂ plasmaCVD grapheneNoRT0.01 torr25 %3
cracked H₂CVD graphene on Ni (111)NoRT1.5×10−61.5 \times10^{-6} torr25 %4
cracked H₂CVD graphene on Ni (111)NoRT1.5×10−61.5 \times10^{-6} torr25.6 %5
cracked H₂CVD graphene on Ni (111)NoRT1.5×10−91.5 \times10^{-9} torr25 %6
H₂Exfoliated grapheneNoRT6 MPa0.2 wt %7
H₂Graphene powderNoRT10 bar0.72 wt %8
H₂Nitrogen-doped graphenePdRT4 MPa4.4 wt %9
H₂Nitrogen-doped graphenePd NPsRT2 MPa1.9 wt %10
H₂Nitrogen-doped graphenePd NPsRT4 MPa4.3 wt %11
H₂Crystalline graphene on SiC (0001)TiRT1 mbar (1 hr)13.5 %; 1.11 wt %This work

Table 1. Summary of experimental hydrogen storage capacity in graphene

References

  1. Luo, Z., Yu, T., Kim, K. J., Ni, Z., You, Y., Lim, S. Shen, Z. Wang, S., and Lin, J. Thickness-dependent reversible hydrogenation of graphene layers. ACS Nano 3, 1781 (2009)

  1. Gowda, P., Mohapatra, D. R., and Misra, A. Enhanced photoresponse in monolayer hydrogenated graphene photodetector. ACS Applied Materials & Interfaces 6, 16763 (2014)

  1. Son, J., Lee, S., Kim, S. J., Park, B. C., Lee, H. K., Kim, S., Kim, J. H., Hong, B. H., and Hong, J. (2016). Hydrogenated monolayer graphene with reversible and tunable wide band gap and its field-effect transistor. Nat. Commun. 7, 13261 (2016)

  1. Lizzit, D., Trioni, M. I., Bignardi, L., Lacovig, P., Lizzit, S., Martinazzo, R., and Larciprete, R. Dual-route hydrogenation of the graphene/Ni interface. ACS Nano 13, 1828 (2019)

Carbon KK-edge XAS spectra

Polarization-dependent carbon KK-edge XAS spectra taken with different soft X-ray incident angles relative to the surface normal are shown in Figure S2. For well-ordered systems, the polarization-dependent XAS can serve to probe the orientation of molecular orbitals [1, 2]. The resonance intensity associated with a particular molecular orbital is largest when the electric field of the incoming light is parallel to the direction of the orbital, and is minimized when the field is perpendicular to the orbital. Namely, normal incident soft X-rays (i.e., 0∘0^\circ with respect to the surface normal) are only sensitive to in-plane orbitals, whereas 45∘45^\circ incident soft X-ray are sensitive to both in-plane and out-of-plane orbitals. Since the π\pi orbital of the graphene sheet is perpendicular to the surface, only the spectra with 45∘45^\circ incidence can reveal the π∗\pi^* resonance.

Image not extracted

Figure S2. The full range XAS spectra of pristine graphene, Ti doped, low H₂ dosage (4.5L), and high H₂ dosage (49.5L) samples. The indicated angles are the angles of the incident soft X-ray beam relative to the surface normal.

Further, by taking the spectral difference, i.e., the 45° spectra minus the 0° spectra, we will effectively select the out-of-plane orbital alone. Figure S3 shows the difference spectra (45°− 0°) of the π∗\pi^{*} resonance region, for pristine, Ti-doped (~0.03 ML), and H₂-dosed (~ 4.5 to 49.5 L) graphene. A pronounced asymmetric π∗\pi^{*} resonance as a spectral feature is clearly visible at about 285.5 eV. Its intensity decreases significantly with increasing H₂ dosage. This result indicates the formation of C−H bonds that suppress the number of π\pi orbitals.

Carbon K-edge X-ray absorption spectra

Figure S3. Carbon K-edge X-ray absorption spectra measured with linearly polarized soft X-rays on pristine graphene, after Ti deposition (Ti ~ 0.03 ML), and after exposure to molecular hydrogen

H2H_2 (4.5 and 49.5 L). We recorded the XAS spectra at 0∘0^\circ and 45∘45^\circ relative to the surface normal (see Figure S2). The difference spectra (45∘45^\circ-0∘0^\circ) are presented here. The schematic diagram shows the resonance characteristics corresponding to a C-C π∗\pi^* orbital (∼285.5\sim285.5 eV) on graphene.

H storage capacity of Ti SACs and Ti nanoclusters of different size on graphene

a. For a single-site Ti atom catalyst,

0.03 ML=4.0×1013 cm−20.03\ \mathrm{ML} = 4.0 \times10^{13}\ \mathrm{cm}^{-2}
ηH=HC=6×(4.0×1013)3.8×1015×100%≈6.3%\eta_H = \frac{H}{C} = \frac{6 \times(4.0 \times10^{13})}{3.8 \times10^{15}} \times100\% \approx6.3\%

b. For a single-site Ti nanocluster catalyst,

Number N of Ti atom in nanoclusters of different size [1]Nanocrystal Radius RcR_\mathrm{c} (nm) [2]Diffusion area (nm2^2)H storage amountH storage per Ti atom
100.2510.229252.5
380.5020.655411.079
950.7531.081570.6
1921.0051.507730.38
3391.2561.933900.265
5471.5072.3591060.194
8261.7582.7851220.148
11872.0093.2111380.116
16402.263.6381540.094
21962.5114.0641710.078

Table S2. Details of calculation of hydrogen storage capacity

Hydrogen storage amount is calculated as H storage amount = Diffusion area × C–H site density

H storage per Ti atom is calculated as H storage per Ti atom =

Diffusion area×C–H site densityN\frac{\text{Diffusion area} \times\text{C--H site density}}{N}

Formation of C−HC-\mathrm{H} bonds on graphene/SiC and its identification using APXPS

From the integrated area of the Gr1 component, we can determine the hydrogen coverage (in %) of crystalline graphene with Ti nanocluster catalysts.

References

  1. [1]Martin, T. P. Shells of atoms. Phys. Reports 1996 4, 199-241.DOI
  2. [2]Duncan Lyngdoh, R. H., Schaefer III, H. F., and King, R. B. Metal–metal (MM) bond distances and bond orders in binuclear metal complexes of the first row transition metals titanium through zinc. Chem. Rev. 2018 24, 11626-11706.DOI

Paper details

vlnc:W_fwnzb2yz
Version
v1 · latest
Published
Fresh — added to the Hub within the last 30 days
Added
License
CC BY 4.0
Usage

views · 0

downloads · 0

A view counts once per browser per day, a download each time the PDF is requested.

  • Physics and Astronomy

Not recorded for this paper

View statement
Looking for related papers…
Contents