Received 17 January 2015; accepted 2 May 2015; published 7 May 2015

ABSTRACT

The synthesis and structure of hexagonal boron carbonitride (h-BCN) film on polycrystalline diamond surface were reported. Polycrystalline diamond and/or diamond-like carbon were first fabricated on Si (100) and then diamond like carbon was used as substrate. The deposition was performed by radio frequency plasma enhanced chemical vapor deposition. In order to reduce the content of nitrogen void defects, the deposition was performed at the high temperature of 950˚C under the working pressure of 2.6 Pa. The typical sample with atomic composition of B₃₁C₃₇N₂₆O₆ in the h-BCN lattice was characterized by X-ray photoelectron spectroscopy. The fine structure of the film was studied by near-edge X-ray absorption fine structure (NEXAFS) measurements. The B K-edge and N K-edge of NEXAFS spectra revealed that the synthesized h-BCN film had the ideal honeycomb-like BN₃ configuration without nitrogen void defects.

Keywords:

h-BCN, Defect, Chemical Vapor Deposition, XPS, NEXAFS

1. Introduction

Hexagonal boron carbonitride (h-BCN) compound has been focusing as essential nanomaterials for its various applications especially in nanoelectronics due to its semiconducting property. The h-BCN could be used in photoluminescent devices, solar cells, sensor, high temperature nanotransistor, lightweight electrical conductors, or high temperature lubricants etc. [1] [2] . The efforts of preparing h-BCN have been initiated since 1980s [3] -[6] , and till to date numerous efforts have been aimed to synthesize single phase of h-BCN using different methods [7] -[11] . However, most of those compounds are amorphous and/or mixtures of pure BN and graphite domains. In fact, the most challenging obstacle for preparing crystalline h-BCN is phase separation and presence of different defect fine structures. Indeed, single phase of crystalline h-BCN with nitrogen void defect free structure is not achieved yet. For the synthesis of h-BCN film, a plasma enhanced chemical vapor deposition (PE-CVD) method is suitable, since the PE-CVD provides unique advantages over other techniques such as freedom to choose the substrate material and a uniform deposition over relatively large area, and controllable deposition parameter to prepare BCN of various compositions and thickness. Furthermore, plasma plays a very active role for film deposition because it makes ions and radicals.

In the previous paper, we report that we have successfully synthesized h-BCN on Si (100) [12] [13] . The h-BCN basal planes perpendicularly standing up orientation with respect to silicon surface were obtained. Also we synthesized h-BCN on Ni (111) and polycrystalline Ti. Randomly oriented h-BCN films were found with respect to nickel and titanium substrates. However, the h-BCN films deposited on these substrates have some nitrogen void defects, which were revealed by the measurements of the near-edge X-ray absorption fine structure (NEXAFS) [12] -[14] . In order to reduce the number of nitrogen void defects in h-BCN, we consider that it would be better to apply higher temperature during the deposition. Also it is better to use a substrate with similar lattice constant with that of h-BCN. In this paper, our succeeding research to obtain h-BCN without nitrogen void defects by applying high temperature in choosing diamond as substrate is presented.

2. Experimental Method

Hexagonal boron carbonitride (h-BCN) films were synthesized on polycrystalline diamond surface by radio frequency plasma enhanced chemical vapor deposition (RF-PECVD) using tris-dimethylaminoborane (TDMAB) as precursor. The experimental method is similar as we reported in the reference [12] . In this study, diamond and/or diamond-like carbon were first developed on Si (100) wafer and then it was used as substrate. The wafer (1 × 1 cm²) was cleaned by ultrasonic cleaner in acetone bath followed by hydrofluoric acid (50%) then the as- cleaned wafer was etched with diamond powder in ethanol for 25 min. After etching, the wafer was washed three times in acetone followed by in ethanol. X-ray diffraction (XRD) (Rigaku diffractometer) using CuK_a radiation (30 kV, 20 mA) was taken for the as-cleaned wafer in order to confirm nonappearance of diamond powder on the wafer surface. Then the wafer was installed in the reaction chamber where the inner pressure was set at ~2.6 Pa. The substrate was heated by RF plasma at around 950˚C and then CH₄ plasma was flown into the chamber for 30 min at flow rate of 10 sccm. The XRD confirmed that diamond and/or diamond-like carbon were formed on the wafer due to the flounce of CH₄ plasma. Finally, BCN films were synthesized in-situ by the flounce of plasma of TDMAB on the diamond surface.

The synthesis was done with varying temperature from 950˚C to 1050˚C. Since the films compositions were almost identical, only one typical result has been discussed. The films were characterized by X-ray photoelectron spectroscopy (XPS), near-edge X-ray absorption fine structure (NEXAFS) and field emission scanning electron microscopy (FE-SEM) (JSM-6700 FSS, JEOL), respectively. The XPS and NEXAFS measurements were performed at the Beam Line 27A and the Beam Line 11A of the KEK photon factory, Tsukuba, Japan. The XP spectra were taken by using synchrotron radiation of energy hn = 2.2 keV. The NEXAFS spectra were measured using 300 l/mm laminar gratings in an UHV chamber (~1.7 ´ 10⁻⁶ Pa) with total electron yield mode by recording sample current. Details of XPS and NEXAFS have been described here [15] .

3. Results and Discussion

Figure 1 show high resolution FE-SEM surface morphologies for the typical sample prepared on polycrystalline diamond surface. It can be seen that the crystals are almost homogeneously distributed at micrometer size all over the surface. High-magnification image (Figure 1(b)) also confirms the homogeneity of the crystals. However, FE-SEM image presents only surface morphologies in micrometer scale. Therefore, composition and structures are studied by spectroscopic methods.

Figure 2 shows XRD patterns for the typical BCN sample on diamond surface (top), diamond film (middle) and silicon wafer (bottom). The XRD patterns confirm nonexistence of diamond powder on the silicon wafer (bottom). The pattern for the as-deposited diamond film (middle) shows a prominent peak at 2q = 44.0˚ which

corresponds to the diffraction from the (111) plane in the diamond lattice [16] . It is clear from the figure that the diffraction peak corresponding to diamond phase or diamond-like carbon phase has been developed on the silicon wafer. The XRD pattern for as-deposited BCN film also shows a single peak at 2q = 44.0˚ with low intensity. But no peak observed at 2q = 26.3˚ in the pattern which previously suggested in formation of h-BCN lattice in the (002) plane [13] [17] . No peaks corresponding to development of other phases like graphite or amorphous carbon are observed. The peak at 2q = 44.0˚ for the sample pattern might be due to development of diamond on the wafer or it may come from the as-developed diamond film. From XRD analyses we could not really say the growth of the crystalline h-BCN phase due to the short-rang atomic order of the lattice.

Figure 3 shows wide scan XP spectrum of the typical sample shown in Figure 2. The composition B₃₁C₃₇N₂₆O₆ shown inside of the figure was estimated from I_B1s/s_B1s:I_C1s/s_C1s:I_N1s/s_N1s, where I is intensity of photoelectron peak and s is photoionization cross-section for the respective core-level indicated as subscript. From the wide scan XPS results it is confirmed that the BCN film is composed of B, C, and N atoms. The O1s peak corresponding to oxygen was found due to the surface contamination. However, the oxygen contamination does not affect the bulk structure of the BCN film, because the formation of the B-O bond was not observed in the B K-edge NEXAFS spectra which probe deeper region than that of XPS (discussed later). In order to confirm that the sample was really made up of B-C-N atomic hybrid rather than a mixture of BN and graphite binary phases, narrow scan B1s, C1s and N1s XPS spectra were measured.

Figure 4 shows the B1s, C1s and N1s narrow scan XPS spectra for the typical sample. It is seen that all the B1s (190.8 eV), C1s (284.5 eV) and N1s (398.0 eV) spectra found almost symmetrical, non-splitting and single Gaussian-like curve. Peak splitting around 192.0 eV in the B1s spectra, 287.5 eV in the C1s spectra and 401.0

eV in the N1s spectra would observed if there is an existence of formation of BN-like and graphite-like phases as observed in the references [10] [18] . Therefore, formation of separated binary phases such as BN and/or graphite was not obvious in the present sample. NEXAFS measurements were performed for further assessment of the nitrogen void fine structures.

Figure 5 shows B K-edge p* resonance NEXAFS spectra taken at normal (q = 90˚) and at grazing (q = 15˚) incidence angles of the X-ray for the typical sample. Since the intensity and shape of the s* resonance peak was unchanged with grazing incidence and hence less informative. Therefore, only p* resonance peak has been considered because several fine structures usually found when the spectra were taken at grazing incidence angle within the p* resonance area. It is seen that a single Gaussian-like peak B₂ has been observed for the sample at photon energy around 192.1 eV and intensity of the peak does not change at the grazing (q = 15˚) incidence, suggesting factual formation of hexagonal h-BCN hybrids with a similar configuration like-BN₃ in the honeycomb-like structure.

It has been reported that if a boron atom is bonded with three carbon atoms forming BC₃-like configuration, the p* resonance peak would appear around 191.0 eV (marked B₁ in Figure 5(a)) [19] , but no peak is seen in this energy region. Also if B-O₂C and B-ON₂ bonds are formed, the p* resonance peak would appear around 192.5 eV (marked B₃) and 193.0 eV (marked B₄), respectively [20] [21] , but scarcely any peaks are observed in these energy regions. Probing depth of NEXAFS in total electron yield is in the order of 10 nm [22] , which is fairly deeper than that of XPS. Therefore, it is confirmed that the oxygen contamination observed in the XPS wide scan (Figure 3) does not affect the bulk structure of the BCN film. In order to clarify the existence of the void structure, the NEXAFS spectra for BCN on diamond are compared with those for the samples deposited on HOPG and Ni substrates. All the nitrogen void fine structures B₁, B₃ and B₄ are clearly observed (Figure 5(b) and Figure 5(c)). Same findings were reported in the case of silicon and titanium substrates [12] [14] .

However, the spectra for the present sample show no fine structures other than the main peak B₂. The B K-edge NEXAFS analyses confirmed that the h-BCN layers composed of highly ordered configuration without nitrogen void and/or with few nitrogen void defects structures.

Figure 6 shows N K-edge resonance NEXAFS spectra for the typical sample. It has been observed that the p* as well as s* resonance peak of the N K-edge does not change with grazing incidence angles. Moreover, the spectral shape is very similar to each other and not fine structures usually found near 397.5 eV (marked N₁ in the Figure 6) due to nitrogen void are appearing [19] . This observation also confirms that the atomic arrangement around N atoms is very homogenous and similar to each other in comparison with those around the B atoms suggesting the ideal honeycomb h-BCN hybrid configuration.

4. Conclusion

We synthesized nano-crystalline hexagonal BCN films on polycrystalline diamond surface by RF-PECVD using TDMAB. The crystalline phase could not be detected by XRD due to short-range atomic order of the (002) lattice plane. However, the crystalline phase in the sample is confirmed by NEXAFS analyses. XPS suggests that the film is composed of B-C-N atomic hybrids. It is suggested from the B K-edge and N K-edge NEXAFS spectral analyses that the hexagonal honeycomb h-BCN lattice has an ideal structure without nitrogen void defects.

Acknowledgements

The authors are grateful to the staff of the KEK-PF to run the experiments (Proposal No. 2008G502).

References

NOTES

^*Corresponding author.