In situ Si doping of heteroepitaxially grown c-BN thin films at different temperatures

Abstract

Phase pure cubic boron nitride (c-BN) films have been epitaxially grown on (001) diamond substrates at 420 °C, 600 °C and 900 °C. Si was added during film growth at these different deposition temperatures in order to achieve in situ n-type doping. The Si concentration increases as the growth temperature decreases. The effects of the different deposition temperatures and the silicon concentration on the transport properties of the c-BN epitaxial films have been systematically investigated. The results demonstrate that the film doped at 420 °C has lower resistivity than the ones doped at elevated temperatures. The temperature dependence of the electron transport properties and their corresponding theoretical fittings reveal that the activation energy of Si doped films is about 0.3 eV and the increased Si concentration could improve the compensation from the deep-level acceptors.

---

1. Introduction

As the simplest of the group-III nitrides, cubic boron nitride (c-BN) has attracted worldwide attention for over half a century.¹ Besides the superhardness and excellent chemical inertness, an approximate 6 eV wide band gap, the possibility for both n- and p-type doping, high thermal conductivity and wide optical transmittance over a large spectral range from UV to visible make it not only applicable as a hard, protective coating and a material for tools, but also a very promising candidate in high temperature, high power and high frequency microelectronic and optoelectronic industrial applications. In fact, a c-BN junction diode has been fabricated by growing Si-doped c-BN on a Be-doped seed crystal under high pressure and high temperature (HPHT), which exhibited good rectification at 530 °C.² Unfortunately, the small crystal size of c-BN prepared by HPHT limits its application in large-area electronic devices. Thus it is indispensable to develop a controlled growth of high quality c-BN films in an epitaxial manner as well as with the appropriate electronic properties for the purpose of realistic industrial application.

However, c-BN films have mostly been of poor structural quality due to the applied energetic ion bombardment during deposition, which is necessary for the formation of the cubic phase. Consequently, the majority of these films are a mixture of phases (amorphous BN (a-BN), turbostratic (t-BN), and c-BN) containing high densities of defects and grain boundaries, severely prohibiting the electronic applications. Significant progress in the synthesis of high quality c-BN films has been achieved, of which the advent of epitaxial c-BN films on appropriate substrates marked a milestone.^3,4 These epitaxial films are robust, phase pure and composed of single crystals with large area, which opens an exciting opportunity for the electronic application of c-BN films. A recent report demonstrated a way to grow single-crystal c-BN films by molecular beam epitaxy (MBE), possibly allowing precise control of thickness and doping in the near future.⁵ Further attempts on the effective doping of these films towards p- and n-type conductivity have been made including ion implantation and in situ doping respectively.^6–9

Ion implantation can effectively turn the intrinsic insulating c-BN films into semiconducting or even semi-metallic films.⁸ However, ion energies and radiation doses during the implantation process need to be precisely controlled because the defects are inevitably accompanied by radiation damage, which will probably lead to the eventual destruction of the c-BN cubic phase and to a transformation into a mechanically soft hexagonal BN (h-BN) phase. Consequently, the high ionized impurity scattering will result in a low Hall mobility far from compatible with industrial levels.¹⁰ On the other hand, in situ doping can generally introduce the desired dopants into the films without over disruption of the matrix structure. However, the incorporation of Si into the c-BN film during deposition was restricted by segregation of Si towards the surface at high temperatures according to a previous report.⁷ The effective donor concentration of the resultant n-type film is very limited due to the high compensation ratio at elevated temperatures. Moreover, the behavior of the impurities in the host matrix of c-BN is considerably complicated, which is mainly reflected by two or more conduction mechanisms dominating in different temperature regions.^11–13

Recently, doping of c-BN films has been extensively investigated including theoretical and experimental efforts.^14–18 However, doping of epitaxial c-BN films has been rarely presented. Hence the purpose of the present study is to experimentally investigate the role that doping temperature plays on the electron transport properties of in situ Si doping of c-BN epitaxial films. c-BN films have been epitaxially grown on (001) diamonds at different temperatures following the previous procedure and Si has been incorporated into c-BN films during film deposition. Temperature dependent resistivity and Hall measurements have been taken to systematically investigate the transport behavior of these doped films.

2. Experimental

2.1. Deposition

All the c-BN films studied in the present work were heteroepitaxially grown on HPHT-synthesized Ib-type (001) diamond single crystals (3 × 3 × 1 mm³) by ion beam assisted deposition (IBAD) using two separate ion sources (dual beam technique). Source I delivers a 1.3 keV Ar⁺ ion beam used for sputtering the water cooled boron target. During the deposition of the B atoms onto the substrate, the growing film is bombarded with a mixture of 280 eV N₂⁺ and Ar⁺ ions provided by source II. As has been reported previously,¹⁹ epitaxial c-BN growth can be obtained under optimized bombarding conditions if diamond single crystals ((001)-oriented, size 3 × 3 × 1 mm³) held at 900 °C are employed as substrates. Complementary to the already published procedure for preparation, we have investigated the possibility of adding small concentrations of another element like Si during the c-BN film growth in order to test the corresponding doping behavior. For the purpose of n-type doping, a small strip (1.5 × 60 mm²) of an undoped Si wafer was inserted through a linear feedthrough allowing the exposure of different amounts of its area to the primary ion beam. In this way, the concentration of the Si in the deposited c-BN films can be varied thereby controlling the flux of the additionally sputtered Si atoms. In order to maximize the Si concentration in the c-BN films, the Si flux has been precisely controlled.

2.2. Characterization

The obtained films were characterized by FTIR, AES, XPS and EELS. FTIR spectra of all the films show a single narrow peak attributed to c-BN through a 400–4000 cm⁻¹ range, indicating a 100% cubic content. AES shows a 1 : 1 stoichiometry of B : N, while EELS gives a primary plasmon energy of 31 eV, indicating a rather good quality. All the films used in this work are around 300 nm thick confirmed by Talystep. For the purpose of electrical measurements, Au/Cr was fabricated on the top of c-BN films as electrodes by Pulsed Laser Ablation (PLD) through a four-point mask, exhibiting perfect Ohmic characteristics (for more detail, please see ref. 20). Four 30 μm diameter gold wires were bonded onto these contacts and connected with measurement housing, while all the undoped diamond substrates, about 1 mm thick, act as an insulator between the c-BN films and the housing. All the electrical transport measurements were carried out from room temperature to about 500 °C using the van der Pauw method.

3. Results and discussion

As reported previously,^7,9 the introduction of Si will be restricted to 200 ppm at 900 °C due to the surface segregation effect. As a consequence, Si might be preferably incorporated at lower deposition temperature regions where the segregation effect is not dominating. In order to study the influence of the doping temperature, a series of Si doped c-BN films were prepared on top of (001) diamond substrates at 420 °C, 600 °C, and 900 °C. The corresponding Si concentrations in c-BN films grown at 420 °C, 600 °C and 900 °C are 465, 300 and 200 ppm respectively, determined by time of flight secondary ion mass spectrometry (ToF-SIMS) (see ESI†). As expected, the Si concentration increases while the growth temperature decreases. The incorporation of more Si atoms would benefit the electronic transport properties of the c-BN films, on the other hand it may probably disturb the cubic structure, which may be replaced by a hexagonal phase.

FTIR measurements were performed to evaluate the volume fraction of the cubic phase by subtracting the background due to the diamond substrate from the transmission spectrum and the compressive stress by estimating the TO-position of c-BN.²¹ The corresponding FTIR spectra have been presented in Fig. 1. Clearly the FTIR spectrum is dominated by only one peak, which is characteristic of the transverse-optical (TO) mode of the cubic phase.²² The absence of characteristic h-BN peaks strongly indicates a 100% cubic phase without any h-BN interface, although some of the films have been grown at lower deposition temperatures on diamond substrates. To further demonstrate the temperature effect on the film quality, the cubic TO mode peak position and full width at half maximum (FWHM) are plotted as a function of deposition temperature, as shown in Fig. 2. It is obvious that as the deposition temperature increases from 420 °C to 900 °C the c-BN peak position shifts from 1097 cm⁻¹ to 1075 cm⁻¹, suggesting that the built-in compressive stress decreases due to the elevated temperatures.²³ On the other hand, as the deposition temperature is increased, the FWHM value decreases from 190 cm⁻¹ to 110 cm⁻¹, indicating an improved crystalline quality. For the Si doped c-BN films at 900 °C with 200 ppm Si concentration the cubic phase can be preserved simultaneously with no indication of the h-BN interface layer. However, it is noteworthy to mention that for the other doped films the unique cubic phase remains stable even though the Si concentration is significantly increased and the deposition temperature is decreased. Thus, it is indispensable to search for an optimal point between Si concentration and the film quality.

Fig. 1 FTIR spectra for c-BN films grown on diamond substrates at 420 °C, 600 °C and 900 °C respectively.

Fig. 2 c-BN peak position (red) and FWHM (blue) as a function of deposition temperature.

The electronic transport properties have been thoroughly investigated by temperature dependent sheet resistance and Hall effect measurements. The sheet resistance as a function of temperature was measured for the Si doped c-BN films grown at 420 °C, 600 °C and 900 °C, respectively, using the van der Pauw method from 300 K to 800 K, as shown in Fig. 3. Sheet resistances for these three samples are logarithmically plotted as a function of 1/k_BT (k_B Boltzmann constant), from which the activation energy can be extracted. Since Si prefers to segregate to the film surface at high temperatures, only sheet resistance values can be shown in Fig. 3 because of the lack of knowledge regarding the segregation layer at different temperatures. It can be clearly seen that the films grown at higher deposition temperatures, which, in turn, contain lower impurity concentrations as estimated from ToF-SIMS, exhibit a higher resistance as expected. The film doped at 900 °C exhibits a sheet resistance value of 2 × 10¹⁰ Ω at room temperature, while the film doped at 420 °C has a lower value of 10⁹ Ω, implying a higher Si doping efficiency at lower temperatures. However, two different activation energies are observed within the present temperature range, indicating some complex conduction mechanism. For doped semiconductors, the resistivity is often represented as a sum of terms; ρ⁻¹ = σ = σ₁ exp(−E₁/k_BT) + σ₂ exp(−E₂/k_BT), where these two terms correspond to two different conductions with activation energies of E₁ and E₂. E₁ and E₂ are the approximated impurity ionization energies. The data were fitted over a limited range of temperatures, assuming that mobility changes are inconsequential. In the present resistance fitting, the dominant conduction mechanism shifts from one impurity ionization with an activation energy in the range of 0.17–0.26 eV at low temperature to another impurity ionization with an activation energy in the range of 0.45–0.89 eV at high temperature. Furthermore, the activation energy reduces as the Si impurity concentration increases from 200 ppm to 465 ppm within both temperature regions.

Fig. 3 Temperature dependence of sheet resistance for silicon doped c-BN films grown at 420 °C (open squares), 600 °C (open circles) and 900 °C (open triangles) on top of diamond substrates. The solid lines are fittings.

Unfortunately, based exclusively on resistance measurements this phenomenon cannot be unravelled. Therefore, the conduction mechanism resulting in two activation energies will be discussed in detail combined with Hall effect measurements. Nevertheless, from Fig. 3, a clear tendency of lower resistances resulting from more impurities can be seen, suggesting Si was successfully incorporated into the film, lowering the film resistance at least near the surface. In the following section, Hall measurements will reveal more reliable activation energies and provide more information in detail.

In order to identify the conduction mechanism for Si dopants, silicon doped c-BN films grown at 900 °C, 600 °C and 420 °C on (001) diamonds were characterized by Hall measurements using the van der Pauw method from 300 K to 800 K. A clear negative Hall signal could be detected confirming the silicon doped films grown at 600 °C and 900 °C to be of n-type conduction, distinct from the previous report of hopping conductions that occurred in Si-containing c-BN polycrystalline films.¹¹ Unfortunately, it was difficult to detect the Hall voltage from the film grown at 420 °C due to the higher impurity concentrations. The temperature dependence of carrier concentrations for silicon doped c-BN films grown at 600 °C and 900 °C are shown in Fig. 4 as well as after H-plasma treatment. For the film doped at 600 °C, the carrier concentration is measured to be 3 × 10¹³ cm⁻³ at about 330 K. When adding silicon to the film at 900 °C, the carrier concentration decreased to 5 × 10¹² cm⁻³, which, again, is consistent with the previous conclusions from ToF-SIMS, the higher deposition temperature has lower impurity concentration. Since Si tended to segregate to the film surface when doping during the film growth, H-plasma (50 W, P_H2 = 10⁻² mbar) was used to remove the segregated surface layer. After H-plasma treatment for 20 min, even lower carrier concentrations have been revealed by Hall measurements. As the temperature increases, the electron concentration increases exponentially for all the samples. The exponential relationship between carrier concentration and 1/k_BT indicates that the conduction can be attributed to thermal activation of the carriers. For a more detailed understanding of the electrical transport, the effect of carrier compensation on donor concentration, N_D, due to unintentional acceptor concentration, N_A, originating from other impurities and/or defects should be taken into consideration. The general expression of carrier concentration for an n-type semiconductor is given by

where n is the electron concentration, g_D is the degeneration factor of donor stats, which equals 2, k_B is the Boltzmann constant, T is the temperature, E_D is the donor energy level, and N_C is the effective density of states in the conduction band. The number of equivalent minima in the conduction band is 6 for c-BN. The density of states effective mass, m^*_dos, for the electron is given by m^*_dos = (m^*_tm^*_tm^*_l)^1/3, where m^*_l and m^*_t are longitudinal and transverse electron effective masses, respectively. For this calculation of c-BN, we used m^*_l = 1.2m₀ and m^*_t = 0.26m₀, where m₀ is the free electron mass.²⁴

Fig. 4 Temperature dependence of carrier concentration of silicon doped c-BN films grown at 600 °C and 900 °C and after H-plasma treatment. Open squares and triangles show the films grown at 600 °C and 900 °C respectively and solid symbols show the corresponding films after H-plasma treatment. The solid lines show the simulation results.

From Fig. 4, the slopes of the logarithm of the carrier concentrations versus the reciprocal of k_BT decreased for the film from 600 °C, to 900 °C and after H-plasma treatment, which was consistent with the observation that more activated donors could induce a lower activation energy. The calculated results are well fitted with the experimental data over the entire temperature region. The simulations confirm the activation energy, E_D, to be 0.3 eV and 0.35 eV for the films doped at 600 °C and 900 °C, respectively. Furthermore, after H-plasma treatment, the activation energy increases to 0.4 eV for both of these two films, which is due to the decreasing carrier concentration resulting from H-plasma. The best-fit values and experimental results are listed in Table 1. Considering the n-type conduction, Si was the only possibility for this n-type behavior, consistent with the theoretical calculations.²⁵ The acceptor concentration as well as the donor concentration increases with Si doping concentration. The compensation ratio, N_A/(N_A + N_D), is almost the same at 49% for the film grown at 900 °C before and after H-plasma treatment, while it is slightly decreased from 49% for the film grown at 600 °C to 44% after H-plasma treatment. This suggests that the concentration of compensating acceptors is increased and the donor concentration increases proportionally. A possible candidate for the compensating acceptor is a substituted C in an N site, C_N, and/or the complex of C_N with an N vacancy, V_N–C_N.²⁶ The simulations confirm that the activation energy of 0.3–0.4 eV is in a good agreement with the previous experimental results for Si doped c-BN.^27,28 However, one may notice that the H-plasma treated samples exhibit higher activation energies than the as-prepared ones, which is due to the lower sheet carrier concentration following H-plasma treatment, although so far, a proton involved transport mechanism in c-BN is still unknown. Nevertheless, surface removal by H-plasma treatment is expected. In the lower temperature region, a deviation of the fitting lines from the experimental data can be observed, indicating the highly compensated situation. One may notice that the activation energies extrapolated from resistance and Hall measurements are not consistent, at least E_D is much higher for resistance measurements in the higher temperature region than the values calculated from Hall measurements. This is probably because the mobility in the higher temperature region could not be treated as linear behaviour, which, actually as we will show in the following, obeys the power law.

Table 1 SIMS and Hall measurement results compared with the simulated parameters^a

Parameters	Simulation			SIMS		Hall
	ED (eV)	ND (cm^−3)	NA (cm^−3)	CSi (cm^−3)	CC (cm^−3)	n (cm^−3)
a ED: donor energy level; ND: donor concentration; NA: acceptor concentration; CSi: Si concentration; CC: C concentration; n: electron carrier concentration.
Film grown at 600 °C	0.3	4.2 × 10^16	4.0 × 10^16	5.0 × 10^19	1.0 × 10^21	10^13–10^15
After H-plasma treatment	0.4	1.0 × 10^17	8.0 × 10^16	—	—	10^13–10^16
Film grown at 900 °C	0.35	3.1 × 10^16	3.0 × 10^16	3.3 × 10^19	8.0 × 10^20	10^12–10^14
After H-plasma treatment	0.4	3.1 × 10^16	3.0 × 10^16	—	—	10^12–10^14

---

To further shed light on the electric properties of in situ doping of c-BN films, Hall mobility has been derived from the temperature dependent Hall measurements, as shown in Fig. 5. Hall mobility increased gradually with increasing temperature for both Si doped c-BN films and the values were 20 and 0.5 cm² V⁻¹ s⁻¹ at around 300 K, respectively. As a matter of fact, the film grown at 600 °C has higher mobility than that grown at 900 °C, which is normally related to the impurity and defect densities. Furthermore, it is important to note that Si doped c-BN exhibited dominant n-type semiconductor properties and nominally undoped films exhibited p-type properties.^7,9 The latter had a higher resistivity with lower sheet hole density, but larger Hall mobility, which indicates that adding silicon to the film resulted in a high defect density and thus high scattering centres. Unlike nominally undoped films which showed a negative temperature dependent mobility, a positive temperature dependence of Hall mobility is found for these silicon doped c-BN films, indicating that an additional scattering mechanism is occurring in the doped films. According to the fitting results, both of these Si doped films could be fitted to a T^1.5 law, suggesting an impurity ionization scattering effect. However, for the film grown at 900 °C, this T^1.5 fitting curve does not describe the data precisely at high temperatures between 600 K and 800 K. In this range, another T⁶-law delivers a reasonable description, which is still unknown.

Fig. 5 Temperature dependence of the Hall mobility of silicon doped c-BN films grown at 600 °C and 900 °C. The open symbols show the measured mobility; the solid lines show fitting by T^1.5, and the dotted line shows fitting by T⁶ in the high temperature range.

4. Conclusions

High quality c-BN films have been deposited on a (001) diamond substrate at 420 °C, 600 °C and 900 °C, while Si was added during the film growth in order to achieve in situ doping. Si concentration increases as the deposition temperature decreases, determined by both SIMS and temperature dependent Hall effect measurements. As the deposition temperature increases from 420 °C to 900 °C, the intrinsic stress of the epitaxial c-BN films will decrease as estimated by the characteristic IR peak shift and the film quality is improved, judged by the significant change of the peak width. A negative Hall signal was observed confirming n-type conduction for all the Si doped c-BN films. c-BN films doped at 420 °C have lower resistivity than the ones doped at elevated temperatures. The temperature dependence of the Hall effect measurements and theoretical fittings reveal that the increased Si concentration could improve the compensation from the deep-level acceptors.

Acknowledgements

This work was financially supported by Deutsche Forschungs gemeinschaft (DFG) within Zi 317/20-1.

References

1. R. H. Wentorf Jr, J. Chem. Phys., 1957, 26, 956.
2. O. Mishima, K. Era, J. Tanaka and S. Yamaoka, Appl. Phys. Lett., 1988, 53, 962.
3. W. J. Zhang, I. Bello, Y. Lifshitz, K. M. Chan, X. M. Meng, Y. Wu, C. Y. Chan and S. T. Lee, Adv. Mater., 2004, 16, 1405.
4. X. W. Zhang, H.-G. Boyen, N. Deyneka, P. Ziemann, F. Banhart and M. Schreck, Nat. Mater., 2003, 2, 312.
5. K. Hirama, Y. Taniyasu, S. Karimoto, Y. Krockenberger and H. Yamamoto, Appl. Phys. Lett., 2014, 104, 092113.
6. B. He, W. J. Zhang, Y. S. Zou, Y. M. Chong, Q. Ye, A. L. Ji, Y. Yang, I. Bello, S. T. Lee and G. H. Chen, Appl. Phys. Lett., 2008, 92, 102108.
7. H. Yin, H.-G. Boyen, P. Ziemann, B. Dohuard, L. Houssiau, F. Renaux, M. Hecq and C. BIttencourt, Diamond Relat. Mater., 2008, 17, 276.
8. H. Yin, I. Pongrac and P. Ziemann, J. Appl. Phys., 2008, 104, 023703.
9. H. Yin and P. Ziemann, Appl. Phys. Lett., 2014, 104, 252111.
10. T. W. H. Oates, L. Ryves, F. A. Burgmann, B. Abendroth, M. M. M. Bilek, D. R. McKenzie and D. G. McCulloch, Diamond Relat. Mater., 2005, 14, 1395.
11. C. Ronning, A. D. Banks, B. L. McCarson, R. Schlesser, Z. Sitar, R. F. Davis, B. L. Ward and R. J. Nemanich, J. Appl. Phys., 1998, 84, 5046.
12. J. Ying, X. W. Zhang, Z. G. Yin, H. R. Tan, S. G. Zhang and Y. M. Fan, J. Appl. Phys., 2011, 109, 023716.
13. X. W. Zhang, Thin Solid Films, 2013, 544, 2–12.
14. A. Pallas and K. Larsson, J. Phys. Chem. C, 2014, 118, 10061–10075.
15. Y. B. Li, T. Y. Cheng, X. Wang, H. X. Jiang, H. S. Yang and K. Nose, J. Appl. Phys., 2014, 116, 043507.
16. T. A. Abtew, W. W. Gao, X. Gao, Y. Y. Sun, S. B. Zhang and P. H. Zhang, Phys. Rev. Lett., 2014, 113, 136401.
17. K. Teii, T. Hori, Y. Mizusako and S. Matsumoto, ACS Appl. Mater. Interfaces, 2013, 5, 2535–2539.
18. H. Murata, T. Taniguchi, S. Hishita, T. Yamamoto, F. Oba and I. Tanaka, J. Appl. Phys., 2013, 114, 233502.
19. X. W. Zhang, H.-G. Boyen, P. Ziemann and F. Banhart, Appl. Phys. A: Mater. Sci. Process., 2005, 80, 735.
20. H. Yin, H. G. Boyen, X. W. Zhang and P. Ziemann, Diamond Relat. Mater., 2007, 16, 46.
21. S. Fahy, Phys. Rev. B: Condens. Matter Mater. Phys., 1995, 51, 12873.
22. P. J. Gielisse, S. S. Mitra, J. N. Plendl, R. D. Griffis, L. C. Mansur, R. Marshall and E. A. Pascoe, Phys. Rev., 1967, 155, 1039.
23. P. Ziemann, H.-G. Boyen, N. Deyneka, D. Schwertberger and P. Widmayer, Adv. Solid State Phys., 2000, 40, 423.
24. Y. N. Xu and W. Y. Ching, Phys. Rev. B: Condens. Matter Mater. Phys., 1991, 44, 7787.
25. V. A. Gubanov, E. A. Pentaleri, C. Y. Fong and B. M. Klein, Phys. Rev. B: Condens. Matter Mater. Phys., 1997, 56, 13077.
26. L. E. Ramos, J. Furthmueller, L. M. R. Scolfaro, J. R. Leite and F. Bechstedt, Phys. Rev. B: Condens. Matter Mater. Phys., 2002, 66, 75209.
27. O. Mishima, J. Tanaka, S. Yamaoka and O. Fukunaga, Science, 1987, 238, 181.
28. X. Li, S. Feng, X. Liu, L. Hou, Y. Gao, Q. Wang, N. Liu, H. Zhang, Z. Chen, J. Zheng and G. Jia, Appl. Surf. Sci., 2014, 308, 31.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c5ra01064k

---