Conference-ISSS-7-PEEM and Micro PES Study of Graphene Growth on Ni ( 110 ) Substrate

Graphene growth on a carbon-doped Ni(110) substrate by surface segregation and surface precipitation has been investigated by photoelectron emission microscopy (PEEM) and micro photoelectron spectroscopy (μ-PES). Single-layer and multi-layer graphenes were found to grow on the substrate with anisotropic shape, which indicate the strong effect on the growth from the Ni(110) substrate. Most of the surface was covered by thick multi-layer graphene when the precipitation had been made with the cooling rate 0.8◦C/s, while small domains of single-layer graphene and a few-layer graphene grew on the bare Ni(110) substrate by the rapid cooling (4.1◦C/s). The layer dependent work functions have been measured by μ-PES. The work functions were 4.2, 4.4 and 4.6 eV for the single, double and > 3 layers graphene on the Ni(110) substrate, respectively. The μ-PES measurement with He-I reveals the shift of the final state peak characteristic to graphite or graphene depending on the number of layers. [DOI: 10.1380/ejssnt.2015.347]


INTRODUCTION
A number of studies about graphene were reported since graphene has many unique properties, such as carrier mobility, thermal conductivity, maximum current density, and young's modulus [1][2][3].Many kinds of graphene synthesis methods have been developed to fabricate large-scale graphenes and control their thickness [1,2,[4][5][6][7][8][9][10].Among them, epitaxial growth of graphene on transition metal using surface segregation promised to produce high-quality graphene without defects [9,10].There were a number of studies about graphene growth on Ni(111) [6][7][8][9][10][11][12], since graphene grows epitaxially due to the fairly good matching of lattice constants between graphene and Ni(111).Observations of epitaxial growth of graphene have been reported on Ni(111) by Photoelectron Emission Microscopy (PEEM) and Low Energy Electron Microscopy (LEEM) [11,12].The graphene normally grew with triangular and hexagonal shapes on the Ni(111) substrate [8].However, a few studies were reported for the growth on the Ni(110) substrate [13] comparing to a number of studies on the Ni(111) substrate.It is interesting to investigate the graphenes growth depending on the surface orientation.Here, we observed graphene growth on Ni(110) using the surface segregation by PEEM and µ-PES.It was found that the graphene precipitated with rectangle shapes on Ni(110).
It was reported for the graphenes on Ni(111), TaC(100) and SiC substrates that the work functions were different depending on the number of graphene layers [14,15].The single layer graphene usually showed the smallest work function.The work function increased with the increase of layer numbers, up to the work function of the graphite.
We have obtained the work functions of the single, double and triple layers of graphene, and found similar trend of the work function.The layer numbers of the graphene on Ni(110) could be determined by the work function.
We have measured µ-PES spectra for the graphene grown areas and the bare Ni(110) area with He-I.We observed a peak that is characteristic to the graphite in the secondary electron region [16,17].The peak is known as the σ-band peak of the final state structure, which corresponds to the in-plane anti-bonding state.The sigmaband peaks were also observed in the domains of few-layer graphenes, but the peak position was shifted to lower kinetic energy.It might suggest the strong interrelation between the graphene and the Ni(110).

II. EXPERIMENTAL
A Ni(110) single crystal 10×4×2 mm 3 was used as the substrate.Carbon was doped into the Ni crystal by solidstate diffusion method [10], where the Ni crystal was heated at 850 • C for 204 hours in high purity graphite powder under a vacuum condition.The carbon concentration was ∼0.44 at.% for the present substrate.After being mechanically polished to a mirror finish with fine alumina powders and degreased ultrasonically in acetone, the carbon-doped Ni(110) mounted on a Mo holder was loaded into a UHV chamber of PEEM (Omicron IS-PEEM).Three UV sources, i.e. a Hg lamp (∼4.9 eV), a Xe lamp (∼6.7 eV) and a He discharged lamp (He-I/He-II), were available to perform PEEM observation in the present system.The whole energy range of the photoelectrons and the secondary electrons were used to obtain the PEEM images.Since the PEEM column was equipped with an electron spectrometer, the selected area spectroscopy (µ-PES) was also available.
The C-doped Ni crystal was heated by an electron bombardment from back side up to 900 • C, where the temperature was measured by infrared and optical pyrometers.When the Ni crystal was heating around 900 • C, graphenes and carbon atoms at the surface were dissolved into the bulk.Thus we can repeatedly start the graphene growth from the bare Ni(110) surface heating at 900 • C.After keeping 900 • C for ∼10 min., we decreased the temperature of the Ni crystal to grow the graphenes on the bare Ni(110) surface.PEEM measurements were performed during growth at heating sample and after growth at RT. Selected-area electron spectroscopy was performed after the growth to investigate the work function and electronic structures of the domains observed by PEEM.The energy of the photoelectron spectra was referred to the Fermi level (E F ) determined from a Fermi edge of photoemission spectra by He-I from the Ni(110) surface.

III. RESULTS AND DISCUSSIONS
PEEM images with the Hg lamp during a cooling process of the Ni(110) are shown in Fig. 1.The emission current of the sample heating was suddenly decreased from that for 900 • C to that for 675 • C at t = 0 s.Fig. 1a, b and  able to consider the light gray domains and dark gray domains as bi-layer graphene and tri-layer graphene, respectively.The stacking structure of the graphene layers on the Ni(110) surface was observed in Fig. 2a.
In order to confirm the graphene growth, we measured the selected area PES for these domains, i.e. the singlelayer (1L), bi-layer (2L), tri-layer graphene (3L) and bare Ni surface.The cut off region of the spectra for each domain measured by He-I are shown in Fig. 3.The horizontal axis is the kinetic energy from E F so that the energy of the cut off position corresponds to the work function of each domain.As seen in the figure, the work functions vary between the domains.The work functions of 1L, 2L, 3L and bare Ni areas were 4.2, 4.4, 4.6 and 5.0 eV, respectively.The reported work functions for a clean Ni(110) surface and a graphite are 5.04 [18] and 4.6 eV [14,19], respectively.Therefore, the present value for the bare Ni(110) surface is consistent with the previous value.It is also reasonable that the triple-layer graphene have the similar work function to that of the graphite.The trend of decrease in work function as decreasing the number of layers is also consistent with the previous study on the TaC and SiC substrates [14,15].Thus, the single, double and triple or more layers of graphenes can be distinguished by their work function on the Ni(110) surface.
For the slowly cooled sample in Fig. 2b, the surface was covered more complex domains.In order to assign these domains, we took the PEEM image of the same area with different UV sources.PEEM images shown in Fig. 4a,  b and c  sified into two types by their contrast in Fig. 4b and c.We name the black domains that were also black in Fig. 4b as "BB", and that were light gray in Fig. 4b as "BL"."BL" were observed as dark gray contrast in Fig. 4c, and indicated in the map (Fig. 4d) as blue color.In addition to the black domains, there were light gray "LG" and dark gray "DG" domains in Fig. 4a.We performed the selected area PES with Hg lamp for each domain, as shown in Fig. 5.The horizontal axis is the kinetic energy from E F so that the cut off energy corresponds to the work function of the domain.As expected from the contrast in Fig. 4a, the intensity was high for "LG" and "DG" and low for "BB" and "BL".However, the cut off energies are almost same for "LG", "DG" and "BB", and are close to the work function of the graphite (∼4.6 eV) [14,19].Since a smaller work function was expected for few layers graphene, the result indicated the graphite or thicker multi-layer graphene cover the most of the surface when the surface was prepared by the slow cooling.Although origins of the different contrast between "LG", "DG" and "BB" may be related to the thickness and quality of the graphene or graphite, we could not understand the reason Volume 13 (2015) Kadowaki, et al.

5.
Selected area PES spectra by the Hg lamp for domains "BL", "BB", "DG" and "LG" (see text) of the slowly cooled sample.
of the different contrast by the present study.The work function of "BL" domains was ∼5.3 eV, which is slightly larger than that of Ni(110) surface.The high work function suggests that there were no graphene at the surface of "BL".Further investigations including a chemical analysis should be required to assign these domains.
Selected area spectra by He-I for 1L, 2L, 3L graphene and the bare Ni area of the rapidly cooled sample as well as the spectra for "DG" and "LG" domains of the slowly cooled sample are shown in Fig. 6.In the spectra of "DG" and "LG", a characteristic peak "A", known as the final state peak of the graphite, appears at the secondary electron region.The peak originated from the unoccupied in-plane σ-bond of graphite [16,17].Thus the peak also supports the thick multi-layer graphene or graphite formation on the slowly cooled sample.There are only very weak intensity around the valence region of the spectra from "DG" and "LG".
We could see valence structures and the Fermi edge in the spectrum from the bare Ni(110) area.The valence structures were almost intact for spectra of 1L, 2L, 3L, while the intensity was slightly decreased from the bare Ni area.It indicates the electronic structures of the Ni surface are preserved under the graphene growth.
When the spectra were closely inspected in the secondary electron region, one may notice the small bump observed around the secondary slope for 1L, 2L and 3L spectra.There was no bump in the spectrum from the bare Ni.These bump should be assigned to the final state peak derived from the σ-band of the graphite ("A"), since the in-plane σ-bond also exist in the graphene.The positions of the bump were at ∼7.1, ∼7.6 and ∼7.7 eV above E F for 1L, 2L and 3L, respectively, and shifted from the peak A of the "LG" (7.8 eV) and "DG" (7.9 eV).The peak shifted ∼0.7 eV to the lower energy side for the single layer graphene.The large shift of the graphene elec- tronic states was reported for the graphene on the Ni(111) and was attributed to the hybridization [20].The present shift indicates strong hybridization of the electronic states between the graphene and the Ni substrate.

IV. CONCLUSIONS
We have investigated the graphene growth on the Ni(110) substrate by PEEM and selected area PES.The graphene was grown on the substrate by the surface segregation and precipitation method.Although only the thick multi-layer graphene grew on the slowly cooled sample, few layers graphenes were observed on the rapidly cooled sample.The work functions of the several domains were measured.The work function increased from the single layer to multi-layer graphenes.The selected area photoemission by He-I reveals the shift of the unoccupied σ-band of the graphene.Strong interaction between the graphene and the Ni(110) substrate was indicated.

1 .
PEEM images of the Ni(110) surface during the cooling process from 900 • C: (a), (b) and (c) are the images after 56 s (718 • C), 84 s (695 • C) and 126 s (682 • C) from the start of the cooling, respectively.

FIG. 2 .
FIG. 2. PEEM images of the Ni(110) surface at RT: (a) after rapid cooling at 4.1 • C/s and (b) after slow cooling at 0.8 • C/s (see text).Intensity scales are shown on the right side.
FIG.3.Secondary cut off region of selected area PES spectra for 1L, 2L and 3L graphenes and the bare Ni(110) area on the rapidly cooled sample (see Fig.1a).
FIG. 4. (a), (b) and (c) are PEEM images from the same area of the slow cooled sample excited by Hg lamp, Xe lamp and He-I, respectively.Intensity scales are shown on the right side.Typical domain contrasts are labeled, i.e.BL, BB, DG and LG, in (a) (see text).(d) Domain-boundary map created from (a).Thick multi-layer graphene covered the surface except blue domains (see text).

FIG. 6 .
FIG.6.Selected area PES spectra by He-I for bare Ni(110) area, 1L, 2L and 3L graphenes on the rapidly cooled sample.Spectra for "DG" and "LG"domains on the slowly cooled sample are also shown.The upper horizontal axis indicated the kinetic energy from the EF.