Local electrical properties of hydrogenated (001) and (111) surfaces of single crystalline diamond

Local electrical properties of hydrogenated (001) and (111) surfaces of single crystalline diamond

Vacuum 83 (2009) 1118–1122 Contents lists available at ScienceDirect Vacuum journal homepage: www.elsevier.com/locate/vacuum Local electrical prope...

390KB Sizes 0 Downloads 7 Views

Vacuum 83 (2009) 1118–1122

Contents lists available at ScienceDirect

Vacuum journal homepage: www.elsevier.com/locate/vacuum

Local electrical properties of hydrogenated (001) and (111) surfaces of single crystalline diamond Y.L. Li, Z.L. Wang, Q. Wang, X.X. Xia, J.J. Li*, C.Z. Gu** Beijing National Laboratory for Condensed Matter Physics, Institute of Physics, Chinese Academy of Sciences, Beijing 100080, China

a r t i c l e i n f o

a b s t r a c t

Article history: Received 3 February 2009 Accepted 4 February 2009

The hydrogenation of Ib-type single crystalline diamond with grain size of several tens of micrometers, synthesized by high-pressure and high-temperature (HPHT) sintering, was carried out by hydrogen plasma treatment in a hot-filament chemical vapor deposition (HFCVD) system. After exposure to air, the surface conductivity of (001) and (111) facets of HPHT single crystalline diamond was measured. The influences of hydrogenation duration, temperature and gas pressure on the surface conductivity of (001) and (111) facets have been investigated. The measurement results show that the variation of hydrogenation conditions has a noticeable effect on the surface conductivity of single crystalline diamond, which is closely related to the formation of a chemical vapor deposition (CVD) regrowth layer on the facets induced by hydrogen plasma treatment process. In addition, (001) surface exhibits higher electrical conductivity than (111) surface, which is mainly attributed to less nitrogen concentration on the (001) surface than on the (111) surface. Ó 2009 Elsevier Ltd. All rights reserved.

Keywords: Diamond Regrowth Surface conductivity

1. Introduction Diamond is known for its excellent properties, such as its high electron and hole mobilities, high breakdown field strength and large band gap (5.5 eV), and hence is an ideal building material for electronic devices with better performance. Furthermore, p-type surface conductivity in undoped diamond, which is observed on the H-terminated diamond surfaces, is a unique feature superior to all other semiconductors and insulators. The earliest observation of this feature was reported by Landstress et al. in 1989, who found that natural diamond crystals and CVD diamond films showed a fairly high surface conductivity due to the presence of hydrogen after hydrogenation using a plasma chemical vapor deposition technique [1,2]. Since then, great efforts have been spent to find the applications of this property for different electronic devices, such as diamond-based Schottky barrier diodes and metal–semiconductor field-effect transistors [3–5]. The HPHT diamond is a commercially available material. It is well known that the substitutive nitrogen centers, which are the main impurities of HPHT diamond, are present in different concentrations within different growth sectors [6]. Whether the incorporation of

* Corresponding author. ** Corresponding author. E-mail addresses: [email protected] (J.J. Li), [email protected] (C.Z. Gu). 0042-207X/$ – see front matter Ó 2009 Elsevier Ltd. All rights reserved. doi:10.1016/j.vacuum.2009.02.006

the nitrogen centers has induced surface conductivity in HPHT diamond is still debatable. F. Maier et al. confirm that synthetic HPHT diamond crystals of type Ib do not show the surface conductivity [7], while Ristein et al. reported that very weak conductive layer was found in the HPHT diamond sectors where higher nitrogen concentrations were present [8], based on the reason that the substitutive nitrogen centers are not only the dominant defects in HPHT diamond crystals but also act as a compensating donor for the surface acceptors, which suppress the hole accumulation. F. Maier et al. have proposed an electrochemical model in which a redox reaction (2H3O þ 2e ¼ H2 þ 2H2O) in an adsorbed water layer provides the electron sink for the subsurface hole-accumulation layer, and which also demonstrated that hydrogenated diamond surface presents substantial conductivity, though there was another report to suggest that the hydrogenation is only a partial requirement for surface conductivity and that exposure to air is also essential [9]. But V. Chakrapani [10] et al. proposed a different electrochemical model, indicating that in the air a more likely electrochemical reaction is the oxygen redox coupling (4H3O þ 4e þ O2 ¼ 6H2O). Therefore, the proper mechanism of hydrogenation and the induced surface conductivity are still not fully understood and need to be further explored. In this work, the effect of the different hydrogen processes on (001) and (111) surface conductivity of HPHT single crystalline diamond has been investigated. We found that different hydrogenation conditions such as hydrogenation duration, temperature

Y.L. Li et al. / Vacuum 83 (2009) 1118–1122

1119

and pressure had a noticeable effect on the surface conductivity of HPHT single crystalline diamond. Different conductivities for (001) and (111) surface at the same hydrogenation condition have been found and an attempt to analyze the phenomenon was made. It is expected that the present work may shed new light on the understanding of surface conductivity mechanism of hydrogenated single crystalline diamond. 2. Experimental Synthetic type Ib diamond single crystals formed by HPHT, with the size of several tens of micrometers, were used in the experiments. The diamond crystals were first ultrasonically cleaned in alcohol and acetone for about 10 min each. They were then placed in an HFCVD chamber where hydrogen plasma was generated by high energy electron bombardment of hydrogen molecules. The surfaces of diamond crystals exposed to hydrogen radicals were hydrogen terminated. The duration, temperature and pressure were varied during hydrogenation as the following: the durations were from 0 to 90 min at the fixed temperature of w500  C, and then the temperature and pressure were changed from 500 to 1200  C and from 20 to 66.66 mbar, respectively. Other experimental parameters in the hydrogenation process were: hydrogen flow at 100 sccm; filament temperature at w2200  C; bias voltage at 400–500 V. The hydrogenation temperature can be measured by a thermocouple placed on the substrate holder. At the end of each treatment, the samples were slowly cooled in hydrogen ambient to room temperature in order to ensure complete hydrogenation of the surface. The samples were then stored for several hours under normal atmospheric conditions before other measurements. Diamond samples with different degrees of hydrogenation were obtained in the above processes for further measurements. As-hydrogenated diamond samples were inspected by a scanning electron microscope (SEM), the variation of surface morphology and chemical bonding state were characterized by atomic force microscope (AFM), Fourier transform infrared (FTIR) spectroscopy. The surface conductivity and electrical characteristics of HPHT single crystalline diamond were measured at room temperature by a double-probe system mounted in the SEM (see Fig. 1) and Keithley 6517 electrometer. During measurements, the two probes with an apex radius of r ¼ 0.5 mm at center to center distance of D ¼ 10 mm were pressed onto (111) and (001) surfaces, respectively. The same distance was maintained for all the measured samples. The probes were made of tungsten filament sharpened by electrochemical etching method. 3. Results and discussion As shown in Fig. 1, the diamond exhibits a large rectangular (001) surface in the center of the platelet as well as triangle (111) surface on each corner. The current–voltage (I–V) measurements of the samples which have different degrees of hydrogenation were performed from 10 V to 10 V at room temperature and in vacuum with a base pressure of 6  102 mbar. The ohmic characteristics for each measurement were verified and the conductance was derived from linear regressions of the I–V plots. Fig. 2 shows typical I–V curves of (001) and (111) surfaces for the sample with 30 min hydrogenation. It can be seen clearly that the current at the (001) surface increases rapidly when voltage is increased from 10 to 10 V, whereas the current rises only slightly at the (111) surface, showing a better conductivity at the (001) surface than at the (111) surface. The hydrogenation conditions such as duration, temperature and pressure were varied to study their influences on the conductivity of (001) and (111) surfaces, as shown in Figs. 3–5. The

Fig. 1. SEM of diamond single crystal and two probes pressed onto the (a) (111) surface and (b) (001) surface, respectively, with same distance at the corresponding positions for each sample.

evidence of hydrogen adsorption on the surface of HPHT diamond induced by hydrogenation process is provided by local FTIR spectrum of hydrogenated HPHT diamond, as shown in Fig. 6. This infrared absorption spectrum presents four main peaks around 2851, 2897, 2918,and 2952 cm1, which are well known as arising from different types of carbon–hydrogen bonds such as C–H2(2850 cm1, and 2920 cm1) and C–H(2900 cm1), 1 1 C–H3(2880 cm , and 2960 cm ) [12]. Although it is difficult to precisely obtain the complete integrate intensity of each peak which strongly depends on the peak fitting parameters, the total bonded hydrogen content is proportional to the total area of the CH stretching since the intensity of local vibration caused by hydrogen

1120

Y.L. Li et al. / Vacuum 83 (2009) 1118–1122

Fig. 2. The I–V curves of (001) and (111) surface of single crystal diamond grains at room temperature in vacuum with a base pressure of 102 mbar. The constant experimental parameters: 900  C temperature, 53.33 mbar pressure and hydrogenation duration of 60 min.

Fig. 4. Effect of hydrogenation temperatures on surface conductivity curves of (001) and (111) surfaces. The constant experimental parameters: 53.33 mbar pressure and hydrogenation duration of 80 min.

incorporation into the diamond host lattice. In addition, the surface conductivity of single crystal diamond hydrogenated at different conditions can be determined by a linear regression from its I–V characteristics. The measured conductance G is related to the twodimensional surface conductivity s2D by s2D ¼ Gp/ln (D/r  1) [11]. With the probe tip radius of r ¼ 0.5 mm and center distance of D ¼ 10 mm, the geometric factor g ¼ ln (D/r  1) is 0.94. Hence the two-dimensional surface conductivity is approximately equal to the conductance in these plots. Fig. 3 shows the change of surface conductance with hydrogenation duration from 0 to 100 min in 10 min interval. It is apparent that long hydrogenation improves the surface conductivity. However, the surface conductivity becomes saturated if the hydrogenation is longer than 80 min. The result also revealed that the conductivity of (111) surface is always less than that of (001) surface within the 80 min of hydrogenated duration. Fig. 4 shows the conductivity of (001) and (111) surfaces as a function of hydrogenation temperature. The surface conductivity reaches its maximum when the hydrogenation temperature is at around 900  C, and then decreases with further increase of the

temperature. This indicates that the hydrogenation at around 900  C is most effective. The influences of gas pressure from 20 to 53.33 mbar on the surface conductivity of single diamond grain at 900  C are shown in Fig. 5. It has the similar effect in the hydrogenation duration. The conductivity of (001) and (111) surfaces become stable when the pressure is above 53.33 mbar. Further increase of pressure no longer has effect on the surface conductivity. Therefore, the optimum conditions in our hydrogenation experiments are: 900  C temperature, 53.33 mbar pressure and hydrogenation duration of 80 min. Under these conditions maximum improvement on surface conductivity can be achieved. The results obtained in our experiments are clearly different from the previously reported results [7,8]. This phenomenon is unlikely caused by more hydrogen adsorption on the surface of HPHT diamond with the evolution of hydrogenation process, because nitrogen as a deep donor in the HPHT diamond can suppress surface conductivity. A possible cause is that a CVD diamond regrowth occurs during the long period of hydrogenation. The carbon residues removed from HPHT diamond surface in the

Fig. 3. The dependence of surface conductivity curves of (001) and (111) surfaces on different hydrogenation durations. The constant experimental parameters: 900  C temperature, and 53.33 mbar pressure.

Fig. 5. The surface conductivity as a function of hydrogenation pressures. The constant experimental parameters: 900  C temperature and hydrogenation duration of 80 min.

Y.L. Li et al. / Vacuum 83 (2009) 1118–1122

Fig. 6. The infrared absorbance spectrum of HPHT diamond after sufficient hydrogenated treatment.

Fig. 7. AFM images of (001) facet of HPHT single crystalline diamond before (a) and after (b) CVD epitaxial regrowth induced by hot-filament hydrogen plasma treatment during hydrogenation process.

1121

plasma can act as a carbon source. Fig. 7 shows the AFM images of (001) facet of HPHT single crystalline diamond at the same area before and after hydrogenation. Before hydrogenation, some scratches appear on the surface with a surface roughness of w19 nm at a scanning area of 5 mm  5 mm. After hydrogenation, most scratches were overgrown, the surface became smoother and the roughness was decreased to w7 nm, which is a proof of the formation of epitaxial diamond layer. The result is also in agreement with the that from M. Stammler et al. [13], who reported the growth of the epitaxial diamond films on the (001) facets of HPHT single crystalline diamond by HFCVD plasma treatment process. Besides, local Raman spectrum of (001) growth sectors HPHT single crystalline diamond after hydrogenation showed only a sharp peak at w1332 cm1 with a full width at half maximum (FWHM) of w2.3 cm1 and without any other peaks, indicating a good local quality of this CVD regrowth layer. This epitaxial diamond layer may explain the increased surface conductivity of HPHT single crystal diamond after hydrogenation. For the growth sectors of nitrogen-incorporated HPHT single crystalline diamond, this fresh diamond layer can drop the total surface built-in potential, and the epitaxial layer thickness is related to the surface acceptor concentration [14]. An estimated w15 nm thickness of the epitaxial layer is required for the surface conductivity to occur after exposure to air, which induced the minimum surface acceptor concentration needed to over compensate the nitrogen donors and inject additional holes into valence band to establish the surface conductivity. In addition, the re-grown layers were observed to show a p-type surface conductivity, which determined that the conductivity is indeed from the surface transfer doping. Therefore, this CVD epitaxial layer with the right thickness is believed to be the root cause of enhanced surface conductivity because of the long duration and high pressure of hydrogenation. Our results also showed that the (001) surface has higher conductivity irrespective of the hydrogenation conditions. This may be due to the difference in nitrogen concentration on different facets. It has been proven that the (111) facets have more nitrogen than (001) facets [6]. The presence of substitutional nitrogen, which acts as a deep and compensating donor for the surface acceptors at lower energies, suppresses hole accumulation and hence cannot contribute to surface conductivity. Higher concentration of incorporated nitrogen in the facets will have stronger suppressing effect on surface conductivity, which is also consistent with the results of Ristein et al. [8]. However, as the hydrogenation process proceeds, epitaxial diamond layer starts to form over both (001) and (111) facets. The difference in nitrogen suppression of surface conductivity between (001) and (111) facets becomes lesser with the growing diamond over layer, and eventually both facets have the same surface conductivity, as in the plots shown in Figs. 3 and 5. As for the enhancement of surface conductivity upon exposure to air, different models have been proposed to explain the phenomenon. Maier et al. [9] suggested that after exposure to atmosphere, diamond surfaces are naturally covered by a thin layer of water, which can provide an electron exchange system to the water layer governed by the hydrogen redox reaction: 2H3 Oþ þ 2e 5H2 þ 2H2 O. Electron can transfer from the valence band to solvated hydronium ions in an aqueous layer and hence holes are accumulated in surface layer of diamond films. The driven force of the reaction is the difference of the chemical potential of electrons between the liquid phase and diamond. If the chemical potential is lower than the Fermi level in diamond, electrons can flow from diamond to the water layer and therefore reduce H3Oþ to H2 and H2O. Besides, V. Chakrapani et al. [10] proposed a different electrochemical model, which suggests that the oxygen redox couple (O2 þ 4H3Oþ þ 4e ¼ 6H2O) is the real cause of enhancement of surface conductivity. It is described in detail

1122

Y.L. Li et al. / Vacuum 83 (2009) 1118–1122

by the equation: 4(ehþ)dia þ O2air þ 2H2Oair þ 4CO2air ¼ 4 hþdia þ 4HCO3film. This equation indicates the formation of a space charge layer of holes, hþdia, in the diamond and compensating anions in the adsorbed film. The type of anion will depend on the nature of the environment. But this model of the oxygen redox couple is lacking for enough experimental evidences to support the relative proposition, and contrarily the model of the hydrogen redox reaction gives experimental evidences which support that electron transfer from diamond to the H3Oþ/(H2O þ H2) redox couple account for the holeaccumulation layer. Even now, the later model may be a helpful supplement of the former model because the evidence for any one electrochemical couple is not conclusive, and hence more works are still needed. In our experiment, a CVD diamond layer was formed on diamond single crystal surface during the hydrogenation process with a changed temperature and pressure, and then exposure to air results in the absorption of a water layer on the surface before measuring surface conductivity. Therefore, a hydrogen redox reaction occurs due to the absorbed water layer and generates a space charge layer of holes in the diamond and compensating anions in the adsorbed film, which helps to further enhance the surface conductivity. 4. Conclusions The electrical conductivity of (001) and (111) surfaces of hydrogenated HPHT single crystalline diamond has been investigated at different hydrogenation conditions. The results show that the surface conductivity increases with hydrogenation duration, temperatures and pressures and they become saturated with further increase of these conditions. The optimum hydrogenation conditions for the enhancement of surface conductivity have been obtained as hydrogenation temperature of 900  C and hydrogenation pressure of 53.33 mbar and hydrogenation duration of 80 min. The effect of hydrogenation conditions on the surface conductivity may be attributed to the CVD regrowth of diamond layer by the epitaxy during hydrogenation process, induced by hot-filament hydrogen plasma treatment. It is also found that the conductivity of

(001) facet is higher than that of (111) facet of HPHT single crystalline diamond at the same hydrogenation condition, which is believed due to a weaker suppression effect on surface conductivity for (001) facet than (111) facet induced by less incorporated nitrogen in (001) surface than (111) surface. In addition, the hydrogen redox reaction model based on the absorbed water film may be more plausible to explain the electron transfer process on the hydrogenated diamond surface for its surface conductivity. Our results may provide a better understanding on the mechanism of surface conductivity in HPHT single crystalline diamond. Acknowledgements This work was supported by the National High Technology Development program of China (Grand No.2002AA325090), the National Natural Science Foundation of China (Grand No.50672121and 60671048) and the Project-sponsored by SRF for ROCS, SEM. References [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] [14]

Landstrass MI, Ravi KV. Appl Phys Lett 1989;55:975. Landstrass MI, Ravi KV. Appl Phys Lett 1989;55:1391. Kawarada H, Aoki M, Sasaki H, Tsugawa K. Diamond Relat Mater 1994;3:961. Itoh M, Kawarada H. Jpn J Appl Phys 1995;34:4677. Jackman RB, Looi HJ, Pang LYS, Whitfield MD, Foord JS. Carbon 1999;37:817. Burns RC, Davies GJ. In: Field JE, editor. The properties of natural and synthetic diamond. London: Academic Press; 1992. p. 395. Maier F, Riedel M, Mantel B, Ristein J, Ley L. Phys Rev Lett 2001;87:209706. Ristein J, Riedel M, Maier F, Mantel BF, Stammler M, Ley L. Phys Stat Sol 2001;186:249. Maier F, Riedel M, Mantel B, Ristein J, Ley L. Phys Rev Lett 2000;85:3472. Chakrapani V, Eaton SC, Aanderson AB, Tabib-Azar M, Angus JC. Electrochem Solid State Lett 2005;8(1):E4. Strobel P, Riedel M, Ristein J, Ley L. Nature 2004;430:439. Tang CJ, Neves AJ, Carmo MC. Appl Phys Lett 2005;86:223107. Stammler M, Eisenbeib H, Ristein J, Neubauer J, Gobbels M, Ley L. Diamond Relat Mater 2002;11:504. Ristein J, Riedel M, Stammler M, Mantel BF. Diamond Relat Mater 2002;11:359.