- Email: [email protected]
post-annealing process to improve the electron field emission properties of ultrananocrystalline diamond films
Diamond & Related Materials 24 (2012) 188–194
Contents lists available at SciVerse ScienceDirect
Diamond & Related Materials journal homepage: www.elsevier.com/locate/diamond
The induction of a graphite-like phase by Fe-coating/post-annealing process to improve the electron field emission properties of ultrananocrystalline diamond films☆ Kuang-Yau Teng a, Pin-Chang Huang b, Wen-Ching Shih b,⁎, I.-Nan Lin a,⁎ a b
Department of Physics, Tamkang University, Tamsui, New-Taipei 251, Taiwan, ROC Graduate Institute in Electro-Optical Engineering, Tatung University, Taipei 104, Taiwan, ROC
a r t i c l e
i n f o
Available online 30 January 2012 Keywords: Fe-coating Post-annealing Electron field emission Ultrananocrystalline diamond
a b s t r a c t The electron field emission (EFE) process for ultrananocrystalline diamond (UNCD) films was tremendously enhanced by Fe-coating and post-annealing processes. The extent of enhancement changes with the granular structure of the UNCD films and the post-annealing conditions (temperature and atmosphere). The best EFE properties are obtained by post-annealing the films at 900 °C in an H2 environment for 5 min. The EFE behavior of the films can be turned on at E0 = 1.28 V/μm, attaining a large EFE current density of 772 μA/cm2 at an applied field of 8.8 V/μm. Microstructural analysis indicates that the mechanism for the improvement in the EFE process is the formation of graphene-like phase (a-few-layer graphite) with good crystallinity, surrounding the Fe (or Fe3C) nanoclusters. Presumably, the nanographites were formed via the reaction of Fe-clusters with diamond films, viz. the Fe-clusters dissolved the carbons in the diamond grains and the re-precipitated them on the surface of the other side of clusters, a process similar to the growth of carbon nanotubes via Fe clusters as catalyst. © 2012 Elsevier B.V. All rights reserved.
1. Introduction Diamond films possess many desirable physical and chemical properties [1–3] and have been the focus of intensive research since the successful synthesis of diamonds in the low pressure and low temperature chemical vapor deposition (CVD) process [4]. Due to the negative electron affinity (NEA) [5,6] characteristics of the reconstructed (100) surface of diamond films, the diamond is considered to have great potential for applications as electron field emitters [7]. Generally, a good electron field emitter requires a sufficient supply of electrons from the back contact of materials, effective transport of electrons through the films and efficient emission from the film surface. The large electronic band gap (5.45 eV) of diamond films hinders their electron field emission (EFE) behavior tremendously due to the lack of conducting electrons required for field emission. Doping the diamond films with boron or nitrogen species provides abundant interband energy levels, which have been observed to markedly enhance both the supply of electrons and facilitate the transport of electrons and hence improve the EFE properties of the materials [8–12]. However, the EFE properties for these diamonds are still not satisfactory due to the fact that most of the emitting surfaces do not
☆ Presented at NDNC 2011, the 5th International Conference on New Diamond and Nano Carbons, Suzhou, China. ⁎ Corresponding authors. E-mail addresses: [email protected] (W.-C. Shih), [email protected] (I.-N. Lin). 0925-9635/$ – see front matter © 2012 Elsevier B.V. All rights reserved. doi:10.1016/j.diamond.2012.01.013
possess NEA characteristics because they are not re-constructed (100) or H-terminated surfaces. Modification on the surface characteristics of diamond to enhance the EFE process has thus been the main focus of research [5,13–15]. Among various approaches, a thin metallic coating was observed to influence the EFE properties of diamond films prominently [14]. Post-treatment in an environment containing Fe species was also observed to tremendously improve the EFE properties of microcrystalline diamond [15,16], but the related effect on the ultrananocrystalline diamond (UNCD) films is not clear. In this paper, the effect of Fe-coating and post-annealing processes on the surface characteristics and the EFE behavior of UNCD films was systematically examined. Transmission electron microscopy (TEM) was used to investigate the microstructure of the films, and the possible mechanism is discussed based on the observations. 2. Experimentals The diamond films were grown on a p-type silicon substrate by a microwave plasma enhanced CVD (MPE-CVD) process. The substrates were first thoroughly cleaned by rinsing the Si wafer sequentially in water-diluted hydrogen peroxide/ammonium hydroxide and hydrogen peroxide/hydrochloric acid solution. The cleaned Si-substrates were then ultrasonicated in a methanol solution containing nanosized diamond powders and Ti-powders (b32.5 nm) for 45 min. The substrates were ultrasonicated again in methanol to remove the nano-particles, which were possibly adhered on the Si-substrates. The UNCD films were grown in a CH4/H2/Ar = 4/0/196 or 4/12/184
K.-Y. Teng et al. / Diamond & Related Materials 24 (2012) 188–194
sccm plasma, excited by a 1200 W (2.45 GHz) microwave with a 150 torr total pressure for 1 h. Thus obtained diamond films were designated as UNCDI and UNCDII films, respectively. The UNCD films were then coated with a thin layer of Fe by a DC sputtering process for 1 min to a thickness around 5 nm. They were then thermally post-annealed in an NH3 (or H2) atmosphere (100 sccm) for 5 min with heating and cooling rates of 15 °C/min. These films were designated as (Fe/UNCDI)an and (Fe/UNCDII)an films, respectively. The morphology and structure of the films were investigated using scanning electron microscopy (SEM, Joel JSM-6500 F) and Raman spectroscopy (Renishaw inVia Raman microscopes), respectively. The detailed microstructure was examined using transmission electron microscopy (TEM, Joel 2100). The EFE properties of the diamond films were measured using a parallel plate setup, in which the cathode-to-anode distance was set by a fixed spacer (125 μm) and the current–voltage (I–V) characteristics were acquired by a Keithley 2410 at 10 -6 torr. The EFE properties were analyzed by the Fowler–Nordheim (F–N) model [17], and the turn-on field was designated as the interception of the lines extrapolated from the highfield and low-field segments of the F-N plots. 3. Results and discussion Fig. 1(a) and (b) show the SEM morphology of UNCDI and UNCDII films, respectively, indicating that the UNCD films are conformally deposited on the Si-substrate. There is no pinholes or voids on the UNCD films. Both of the UNCD films contain ultra-small grains with equi-axised geometry. However, detail examination using TEM reveals a subtle difference in granular structure of the two films. Fig. 1(c) shows that the UNCDI films grown using CH4/0%H2/Ar plasma contain equi-axised diamond grains about 5 nm in size, whereas Fig. 1(d) indicates that the UNCDII films grown using hydrogen containing plasma (CH4/6%H2/Ar) contain diamond rods about 100 nm in size.
189
Fig. 2(a) shows the EFE properties of the UNCDI and UNCDII films with the inset showing the F–N plots, from which the turn-on field was deduced. This figure indicates that the UNCDI films can be turned on at 2.2 V/μm, attaining EFE current density of Je=9.0 μA/cm2 at an applied field of 8.8 V/μm, whereas the UNCDII films can be turned on at 3.9 V/μm, achieving Je = 5 μA/cm 2 at the same applied field (8.8 V/μm). The addition of H2 (6%) into the films results in slight degradation on the EFE properties. The coating a thin layer of Fe (~ 5 nm) almost completely suppressed the EFE properties of the films (not shown). Raman spectra shows in Fig. 2(b) reveals that both films contain typical Raman resonance peaks of the UNCD films. The Raman resonance peaks are very broad due to the smallness in the diamond grains [18]. Generally, the Raman spectra contain υ1-band (1140 cm − 1) and υ2-band (1480 cm − 1), which represents transpolyacetylene along the grain boundaries [19], and the D*-band (1350 cm − 1) and G-band (1580 cm − 1), which represents disordered carbon and graphitic phase contained in the UNCD films [20]. The Raman characteristics were not changed due to Fecoating (not shown). Post-annealing processes after Fe-coating markedly improved the EFE properties for the UNCD films. Fig. 3 indicates that the extent of enhancement varies with the granular structure of the UNCD films and the parameters (temperature and atmosphere) used for postannealing the samples. To more clearly illustrate the Fe-coating/ post-annealing effect on improving the EFE properties of these films, the important EFE parameters, the turn-on field (E0) and EFE current density (Je, at an applied field of 8.8 V/μm) were extracted from these J-E curves and were plotted in Fig. 4 against the postannealing temperature (Ta). Fig. 4(a) and (b) show, respectively, the variation of E0- and Je- values with respect to Ta, where the E0 and Je of the starting materials (UNCDI and UNCDII films) were plotted as dotted lines to facilitate the comparison. Generally, in the range of post-annealing temperature of interest, the turn-on field (E0) decreases, whereas the EFE current density (Je) increases,
Fig. 1. (a, b) The SEM micrograph and (c, d) the TEM micrograph of the as-grown UNCD films, which were grown in CH4/H2/Ar plasma with H2 = 0%, UNCDI (a, c) or H2 = 6%, UNCDII (b, d).
190
K.-Y. Teng et al. / Diamond & Related Materials 24 (2012) 188–194
Fig. 2. The (a) EFE properties and (b) Raman spectra of the as-grown UNCD films. The inset in (a) shows the corresponding Fowler–Nordheim plots.
monotonically with the post-annealing temperature for the samples post-annealed in H2-atmosphere (solid symbols, Fig. 4(a) and (b)). In contract, the E0 increases and the Je decreases monotonically with the post-annealing temperature for the samples post-annealed in NH3 atmosphere (open symbols, Fig. 4). Moreover, post-annealing in H2-environment is more efficient in enhancing the EFE process for the UNCD films, as compared with that in NH3-environment. The best EFE behavior attainable for UNCDI films are (E0)UNCDI = 2.0 V/μm and (Je)UNCDI = 120 μA/cm 2 at 8.8 V/μm, whereas those obtainable for UNCDII films are (E0)UNCDII = 1.28 V/μm and (Je)UNCDII = 772 μA/ cm2 at 8.8 V/μm, when they were Fe-coated/post-annealed at 900 °C (H2). The UNCDII films attain better EFE properties than the UNCDI films after Fe-coating/post-annealing process. These EFE properties were listed in Table 1. The Fe-coating/post-annealing processes insignificantly change the Raman structure of the UNCD films, regardless of the granular structure of the starting films and the post-annealing parameters (temperature and atmosphere). Fig. 5 indicates that all the films contain very broad υ1- & υ2-bands and D⁎- & G-bands Raman resonance peaks, which, again, is due to the smallness in grain size for the films. The Fe-coating/post-annealing processes insignificantly alter the Raman characteristics for the UNCD films. However, these processes markedly modify the surface morphology of the films. Fig. 6(a) and (b) show the SEM micrographs of the UNCDI films that were Fecoated/post-annealed at 900 °C(H2) and 850 °C(NH3), respectively.
It should be noted these samples are the ones which exhibit the best EFE properties among the Fe-coated/post-annealed samples. The microstructure of the UNCDI films has been completely changed. The equi-axised granular structure was no longer observable. There formed network-like surface structure with spherical particulates about 20–30 nm size (white dots) sparsely distributed on the surface of the samples. Similar granular structure was obtained for UNCDII for those which post-annealed at 900 °C(H2) and 850 °C(NH3) (Fig. 6(c) and (d)). The population of spherical particulates are more abundant for the UNCD films post-annealed in H2-environment (Fig. 6(a) and (c)), as compared with those post-annealed in NH3-environment (Fig. 6(b) and (d)). The phenomenon that Fe-coating/post-annealing processes can enhance the EFE properties of the diamond films has also been observed for microcrystalline diamond (MCD) films [15,16]. It was proposed that the Fe-coating on MCD films formed nano-sized Feclusters in the heating up stage of the post-annealing process [15,16]. The nano-sized Fe-clusters catalytically dissociated the carbon in sp 3-bonded diamond lattices at post-annealing temperature and then re-precipitated it out at the opposite side of Fe-clusters to form a-few-layer graphite. The Fe3C particulates were formed when the carbon contained in the Fe-clusters were frozen inside the Feclusters when the samples were cooled fast in a rate. The same phase transformation mechanism was presumed to occur in the Fe-coating/ post-annealing of UNCD films. The white spherical particulates observed in Fig. 6 are thus presumed to be the Fe3C particulates and the network-like surface structure is presumed to be the graphene-like phase formed by dissolution and re-precipitation process. To understand the possible formation process for the microstructure, the Fe-coated/post-annealing samples were examined using TEM. Figs. 7(a) and 8(a) show the typical bright field TEM micrograph of the UNCD films Fe-coated/post-annealed in H2(900 °C) and NH3(850 °C) atmosphere, respectively. Both films contain complicated microstructure. Selected area electron diffraction patterns (inset, Fig. 7(a) and (b)) indicate that the films consist of 3 diffraction rings that correspond to diamond lattices, which are designated as d111, d220, and d311, 2 faint diffraction rings nearby the d111 ring that correspond to Fe3C nano-particles and 1 diffuse ring in the center that corresponds to graphitic (or amorphous carbon) materials. To more clearly illustrate the phase constituents for the films, a dark field image was composed by superimposing the TEM dark field images taken from the diffraction spots corresponding to diamond, Fe3C and nanographite phase. Figs. 7(b) and 8(b) show the composed dark field images corresponding to Figs. 7(a) and 8(a), respectively, indicating that the Fe3C particulates, about 10–20 nm in size, and the nano-graphite (or amorphous carbon), are distributed evenly all over the samples. More detailed analysis using high resolution TEM indicates that the presence of Fe3C particulates (area 2, FT2, Fig. 7(c)) in the H2 (900 °C) post-annealed UNCD films, besides the diamond grains (area 1, FT1, Fig. 7(c)). Moreover, there forms graphite (or amorphous) phase on the other side of Fe3C particulates, against the diamond grains. The presence of graphite phase is even more clearly illustrated in Fig. 8(c), which correspond to the UNCD films postannealed in NH3-environment. Nano-sized clusters (a few nm in size, area 2) are observed in the vicinity of Fe3C particulates (area 1, FT1). The FT image (FT2) clearly illustrates that the nano-clusters in area 2 is graphite. Actually, the nano-graphite phase is observable surrounding the Fe3C particulates. Although these micrographs cannot clearly illustrate that the nano-graphitic phase is formed near every Fe3C particles due to the small abundance of these phases, it is believed that the formation process of graphene-like phase occurred to all of the Fe-clusters in the Fe-coated/post-annealed UNCD films. Restated, the Fe-coating form nano-clusters in the beginning stage of post-annealing process. The Fe-clusters then reacted with diamond films, dissolving the carbon species in the diamond and reprecipitating them out to form nano-graphite.
K.-Y. Teng et al. / Diamond & Related Materials 24 (2012) 188–194
191
Fig. 3. (a,b) The EFE properties of the UNCDI films, which grown in CH4(1%)/H2(0%)/Ar(99%) plasma and post-annealed at various temperature in H2(a) and NH3(b) atmosphere; (c, d) Those of UNCDII films, which were grown in CH4(1%)/H2(6%)/Ar(93%) plasma and post-annealed in H2(c) and NH3(d) atmosphere.
However, how the post-annealing in H2-environment leads to better EFE properties for the UNCD films than the post-annealing in NH3-environment is still not clear. SEM examinations indicate that post-annealing in H2-environment results in more abundant Fe3C particulates (cf. Fig. 6). Such a result implies that the reaction between diamond and Fe-clusters is somehow retarded in NH3-environment. The diffusion of carbon into Fe-clusters is thus hindered. However, more systematic studies in this phenomenon are required to elucidate the genuine mechanism. Moreover, while the EFE properties of these Fe-coated/post-annealed UNCD films are still inferior to the nanocarbon films [21–24], the silent feature of these films is that the
nano-graphite phase were precipitated out from the Fe-clusters. The nano-graphite clusters adhere very well with the underlying UNCD films and the substrates, while the nano-carbon films grown directly on Si-substrates usually suffer from poor adhesion. Therefore, the Fe-coating/post-annealing process are more robust in device applications. 4. Conclusions The Fe-layer coated on UNCD films reacted with diamond as soon as the post-annealing temperature reached 800 °C, resulted in the
Fig. 4. Variation of (a) turn-on field, E0, and (b) EFE current density, Je, of the Fe-coating/post-annealed UNCD films, which were grown in CH4/H2/Ar plasma with H2 = 0 or 6% and were post-annealed in H2 or NH3 atmosphere.
192
K.-Y. Teng et al. / Diamond & Related Materials 24 (2012) 188–194
Table 1 The EFE performance of as-deposited UNCD films (UNCDI and UNCDII) and those after Fe-coating/post-annealing in H2 (UNCDIa & UNCDIIa) or NH3 (UNCDIb & UNCDIIb) enrivonment. Samples Deposition parameters CH4/H2/Ar (sccm) UNCDI UNCDIa UNCDIb UNCDII UNCDIIa UNCDIIb
4/0/196 4/0/196 4/0/196 4/12/184 4/12/184 4/12/184
Annealing conditions temperature (atmosphere) (°C)
E0a (V/ μm)
Je b (μA/ cm2)
– 900 850 – 900 850
2.2 2.0 1.44 3.9 1.28 0.8
9 120 32 5 772 166
(H2) (NH3) (H2) (NH3)
a E0: turn-on field designated as the interception of the lines extrapolated from the high field and low field segments in F–N plot. b Je: field emission current density measured at 8.8 V/μm.
carbon dissolution and re-precipitation processes, which markedly enhanced the EFE process for UNCD films. However, to achieve a proper microstructure and to enhance the EFE process, not only is a sufficiently high post-annealing temperature (850 °C) required but also a highly active reducing atmosphere (H2 or NH3) is needed. Microstructural analysis indicates that the mechanism for the improvement of the EFE process is the formation of nanographites surrounding the Fe3C (or Fe) nanoclusters, which were formed via the reaction of Fe-clusters with diamond films. Presumably, the nanographites were formed by the re-precipitation of carbon species, which were dissolved in the Fe-clusters. The best EFE properties achieved were a turn-on field of E0 = 1.28 V/μm with an EFE current density of Je = 772 μA/cm 2 at an applied field of 8.8 V/μm, which were obtained for the films post-annealed in the H2 environment at 900 °C for 5 min. The degree of enhancement on the EFE properties varies with the granular structure of UNCD films and the post-annealing parameters (temperature and atmosphere).
Fig. 5. The Raman spectroscopy of the (a) UNCDI and (b) UNCDII films that were post-annealed in H2 and NH3, annealing parameters are Ia: 800 °C(H2), Ib: 850 °C(H2), and Ic: 900 °C(H2); IIa: 850 °C(NH3), IIb: 900 °C(NH3), and IIc: 950 °C(NH3).
Fig. 6. The SEM micrograph of the (a,b) UNCDI and (c,d) UNCDII films that were post-annealed at 900 °C(H2) and 850 °C(NH3).
K.-Y. Teng et al. / Diamond & Related Materials 24 (2012) 188–194
Fig. 7. (a) Bright field, (b) dark field, and (c) structure image, TEM micrographs of the UNCD films, which were post-annealed in H2 atmosphere (900 °C).
193
Fig. 8. (a) Bright field, (b) dark field, and (c) structure image, TEM micrographs of the UNCD films, which were post-annealed in NH3 atmosphere (850 °C).
194
K.-Y. Teng et al. / Diamond & Related Materials 24 (2012) 188–194
Acknowledgements This work was sponsored by the Tatung University, and Tamkang University, Taiwan, R.O.C. The authors deeply appreciated their financial and technical support. References [1] [2] [3] [4] [5]
J.E. Field, The Properties of Diamonds, Academic, London, 1979. H. Liu, D.S. Dandy, Diamond Relat. Mater. 4 (1995) 1173. J.C. Angus, H.A. Will, W.S. Stanko, J. Appl. Phys. 39 (1968) 2915. B.V. Spitsyn, L.L. Bouilov, B.V. Derjaguin, J. Cryst. Growth 52 (1981) 219. J. van der Weide, Z. Zhang, P.K. Baumann, M.G. Wensell, J. Bernholc, R.J. Nemanich, Phys. Rev. B 50 (1994) 5803. [6] J.B. Cui, J. Ristein, L. Ley, Phys. Rev. Lett. 81 (1998) 429. [7] M.W. Geis, N.N. Efremow, J.D. Woodhouse, M.D. Mcaleese, M. Marchywka, D.G. Socker, J.F. Hochedez, IEEE Electron Device Lett. 12 (1991) 456. [8] A.T. Sowers, B.L. Ward, S.L. English, R.J. Nemanich, Jpn. J. Appl. Phys. 86 (1999) 3973.
[9] P.T. Joseph, N.H. Tai, H. Niu, U.A. Palnitkar, W.F. Pong, H.F. Cheng, I.N. Lin, Diamond Relat. Mater. 17 (2008) 1812. [10] Y.C. Lee, S.J. Lin, I.N. Lin, H.F. Cheng, J. Appl. Phys. 97 (2005) 054310–054311. [11] C.F. Shih, K.S. Liu, I.N. Lin, Diamond Relat. Mater. 9 (2000) 1591. [12] Y.H. Chen, C.T. Hu, I.N. Lin, Appl. Phys. Lett. 75 (1999) 2857. [13] V.I. Polyakov, N.M. Rossukanyi, A.I. Rukovishnikov, S.M. Pimenov, A.V. Karabutov, V.I. Konov, J. Appl. Phys. 84 (1998) 2882. [14] A. Lamouri, Y. Wang, G.T. Mearini, I.L. Krainsky, J.A. Dayton Jr., W. Mueller, J. Vac. Sci. Technol. B 14 (1996) 2046. [15] P.C. Huang, W.C. Shih, H.C. Chen, I.N. Lin, Jpn. J. Appl. Phys. 50 (2011) 08KE04-1. [16] P.C. Huang, W.C. Shih, H.C. Chen, I.N. Lin, J. Appl. Phys. 109 (2011) 084309. [17] R.H. Fowler, L. Nordheim, Proc. R. Soc. London Ser. A 119 (1928) 173. [18] D. Zhou, D.M. Gruen, L.C. Qin, T.G. McCauley, A.R. Krauss, J. Appl. Phys. 84 (1998) 1981. [19] A.C. Ferrari, J. Robertson, Phys. Rev. B 63 (2001) 121405. [20] R.G. Buckley, T.D. Moustakas, L. Ye, J. Varon, J. Appl. Phys. 66 (1989) 3595. [21] P.G. Collins, A. Zettl, Appl. Phys. Lett. 69 (1996) 1969. [22] W. Zhu, C. Bower, O. Zhou, G. Kochanski, S. Jin, Appl. Phys. Lett. 75 (1999) 873. [23] A.N. Obraztsov, A.P. Volkov, I. Pavlovsky, Diamond Relat. Mater. 9 (2000) 1190. [24] Y. Chen, D.T. Shaw, L. Guo, Appl. Phys. Lett. 76 (2000) 2469.
Contents lists available at SciVerse ScienceDirect
Diamond & Related Materials journal homepage: www.elsevier.com/locate/diamond
The induction of a graphite-like phase by Fe-coating/post-annealing process to improve the electron field emission properties of ultrananocrystalline diamond films☆ Kuang-Yau Teng a, Pin-Chang Huang b, Wen-Ching Shih b,⁎, I.-Nan Lin a,⁎ a b
Department of Physics, Tamkang University, Tamsui, New-Taipei 251, Taiwan, ROC Graduate Institute in Electro-Optical Engineering, Tatung University, Taipei 104, Taiwan, ROC
a r t i c l e
i n f o
Available online 30 January 2012 Keywords: Fe-coating Post-annealing Electron field emission Ultrananocrystalline diamond
a b s t r a c t The electron field emission (EFE) process for ultrananocrystalline diamond (UNCD) films was tremendously enhanced by Fe-coating and post-annealing processes. The extent of enhancement changes with the granular structure of the UNCD films and the post-annealing conditions (temperature and atmosphere). The best EFE properties are obtained by post-annealing the films at 900 °C in an H2 environment for 5 min. The EFE behavior of the films can be turned on at E0 = 1.28 V/μm, attaining a large EFE current density of 772 μA/cm2 at an applied field of 8.8 V/μm. Microstructural analysis indicates that the mechanism for the improvement in the EFE process is the formation of graphene-like phase (a-few-layer graphite) with good crystallinity, surrounding the Fe (or Fe3C) nanoclusters. Presumably, the nanographites were formed via the reaction of Fe-clusters with diamond films, viz. the Fe-clusters dissolved the carbons in the diamond grains and the re-precipitated them on the surface of the other side of clusters, a process similar to the growth of carbon nanotubes via Fe clusters as catalyst. © 2012 Elsevier B.V. All rights reserved.
1. Introduction Diamond films possess many desirable physical and chemical properties [1–3] and have been the focus of intensive research since the successful synthesis of diamonds in the low pressure and low temperature chemical vapor deposition (CVD) process [4]. Due to the negative electron affinity (NEA) [5,6] characteristics of the reconstructed (100) surface of diamond films, the diamond is considered to have great potential for applications as electron field emitters [7]. Generally, a good electron field emitter requires a sufficient supply of electrons from the back contact of materials, effective transport of electrons through the films and efficient emission from the film surface. The large electronic band gap (5.45 eV) of diamond films hinders their electron field emission (EFE) behavior tremendously due to the lack of conducting electrons required for field emission. Doping the diamond films with boron or nitrogen species provides abundant interband energy levels, which have been observed to markedly enhance both the supply of electrons and facilitate the transport of electrons and hence improve the EFE properties of the materials [8–12]. However, the EFE properties for these diamonds are still not satisfactory due to the fact that most of the emitting surfaces do not
☆ Presented at NDNC 2011, the 5th International Conference on New Diamond and Nano Carbons, Suzhou, China. ⁎ Corresponding authors. E-mail addresses: [email protected] (W.-C. Shih), [email protected] (I.-N. Lin). 0925-9635/$ – see front matter © 2012 Elsevier B.V. All rights reserved. doi:10.1016/j.diamond.2012.01.013
possess NEA characteristics because they are not re-constructed (100) or H-terminated surfaces. Modification on the surface characteristics of diamond to enhance the EFE process has thus been the main focus of research [5,13–15]. Among various approaches, a thin metallic coating was observed to influence the EFE properties of diamond films prominently [14]. Post-treatment in an environment containing Fe species was also observed to tremendously improve the EFE properties of microcrystalline diamond [15,16], but the related effect on the ultrananocrystalline diamond (UNCD) films is not clear. In this paper, the effect of Fe-coating and post-annealing processes on the surface characteristics and the EFE behavior of UNCD films was systematically examined. Transmission electron microscopy (TEM) was used to investigate the microstructure of the films, and the possible mechanism is discussed based on the observations. 2. Experimentals The diamond films were grown on a p-type silicon substrate by a microwave plasma enhanced CVD (MPE-CVD) process. The substrates were first thoroughly cleaned by rinsing the Si wafer sequentially in water-diluted hydrogen peroxide/ammonium hydroxide and hydrogen peroxide/hydrochloric acid solution. The cleaned Si-substrates were then ultrasonicated in a methanol solution containing nanosized diamond powders and Ti-powders (b32.5 nm) for 45 min. The substrates were ultrasonicated again in methanol to remove the nano-particles, which were possibly adhered on the Si-substrates. The UNCD films were grown in a CH4/H2/Ar = 4/0/196 or 4/12/184
K.-Y. Teng et al. / Diamond & Related Materials 24 (2012) 188–194
sccm plasma, excited by a 1200 W (2.45 GHz) microwave with a 150 torr total pressure for 1 h. Thus obtained diamond films were designated as UNCDI and UNCDII films, respectively. The UNCD films were then coated with a thin layer of Fe by a DC sputtering process for 1 min to a thickness around 5 nm. They were then thermally post-annealed in an NH3 (or H2) atmosphere (100 sccm) for 5 min with heating and cooling rates of 15 °C/min. These films were designated as (Fe/UNCDI)an and (Fe/UNCDII)an films, respectively. The morphology and structure of the films were investigated using scanning electron microscopy (SEM, Joel JSM-6500 F) and Raman spectroscopy (Renishaw inVia Raman microscopes), respectively. The detailed microstructure was examined using transmission electron microscopy (TEM, Joel 2100). The EFE properties of the diamond films were measured using a parallel plate setup, in which the cathode-to-anode distance was set by a fixed spacer (125 μm) and the current–voltage (I–V) characteristics were acquired by a Keithley 2410 at 10 -6 torr. The EFE properties were analyzed by the Fowler–Nordheim (F–N) model [17], and the turn-on field was designated as the interception of the lines extrapolated from the highfield and low-field segments of the F-N plots. 3. Results and discussion Fig. 1(a) and (b) show the SEM morphology of UNCDI and UNCDII films, respectively, indicating that the UNCD films are conformally deposited on the Si-substrate. There is no pinholes or voids on the UNCD films. Both of the UNCD films contain ultra-small grains with equi-axised geometry. However, detail examination using TEM reveals a subtle difference in granular structure of the two films. Fig. 1(c) shows that the UNCDI films grown using CH4/0%H2/Ar plasma contain equi-axised diamond grains about 5 nm in size, whereas Fig. 1(d) indicates that the UNCDII films grown using hydrogen containing plasma (CH4/6%H2/Ar) contain diamond rods about 100 nm in size.
189
Fig. 2(a) shows the EFE properties of the UNCDI and UNCDII films with the inset showing the F–N plots, from which the turn-on field was deduced. This figure indicates that the UNCDI films can be turned on at 2.2 V/μm, attaining EFE current density of Je=9.0 μA/cm2 at an applied field of 8.8 V/μm, whereas the UNCDII films can be turned on at 3.9 V/μm, achieving Je = 5 μA/cm 2 at the same applied field (8.8 V/μm). The addition of H2 (6%) into the films results in slight degradation on the EFE properties. The coating a thin layer of Fe (~ 5 nm) almost completely suppressed the EFE properties of the films (not shown). Raman spectra shows in Fig. 2(b) reveals that both films contain typical Raman resonance peaks of the UNCD films. The Raman resonance peaks are very broad due to the smallness in the diamond grains [18]. Generally, the Raman spectra contain υ1-band (1140 cm − 1) and υ2-band (1480 cm − 1), which represents transpolyacetylene along the grain boundaries [19], and the D*-band (1350 cm − 1) and G-band (1580 cm − 1), which represents disordered carbon and graphitic phase contained in the UNCD films [20]. The Raman characteristics were not changed due to Fecoating (not shown). Post-annealing processes after Fe-coating markedly improved the EFE properties for the UNCD films. Fig. 3 indicates that the extent of enhancement varies with the granular structure of the UNCD films and the parameters (temperature and atmosphere) used for postannealing the samples. To more clearly illustrate the Fe-coating/ post-annealing effect on improving the EFE properties of these films, the important EFE parameters, the turn-on field (E0) and EFE current density (Je, at an applied field of 8.8 V/μm) were extracted from these J-E curves and were plotted in Fig. 4 against the postannealing temperature (Ta). Fig. 4(a) and (b) show, respectively, the variation of E0- and Je- values with respect to Ta, where the E0 and Je of the starting materials (UNCDI and UNCDII films) were plotted as dotted lines to facilitate the comparison. Generally, in the range of post-annealing temperature of interest, the turn-on field (E0) decreases, whereas the EFE current density (Je) increases,
Fig. 1. (a, b) The SEM micrograph and (c, d) the TEM micrograph of the as-grown UNCD films, which were grown in CH4/H2/Ar plasma with H2 = 0%, UNCDI (a, c) or H2 = 6%, UNCDII (b, d).
190
K.-Y. Teng et al. / Diamond & Related Materials 24 (2012) 188–194
Fig. 2. The (a) EFE properties and (b) Raman spectra of the as-grown UNCD films. The inset in (a) shows the corresponding Fowler–Nordheim plots.
monotonically with the post-annealing temperature for the samples post-annealed in H2-atmosphere (solid symbols, Fig. 4(a) and (b)). In contract, the E0 increases and the Je decreases monotonically with the post-annealing temperature for the samples post-annealed in NH3 atmosphere (open symbols, Fig. 4). Moreover, post-annealing in H2-environment is more efficient in enhancing the EFE process for the UNCD films, as compared with that in NH3-environment. The best EFE behavior attainable for UNCDI films are (E0)UNCDI = 2.0 V/μm and (Je)UNCDI = 120 μA/cm 2 at 8.8 V/μm, whereas those obtainable for UNCDII films are (E0)UNCDII = 1.28 V/μm and (Je)UNCDII = 772 μA/ cm2 at 8.8 V/μm, when they were Fe-coated/post-annealed at 900 °C (H2). The UNCDII films attain better EFE properties than the UNCDI films after Fe-coating/post-annealing process. These EFE properties were listed in Table 1. The Fe-coating/post-annealing processes insignificantly change the Raman structure of the UNCD films, regardless of the granular structure of the starting films and the post-annealing parameters (temperature and atmosphere). Fig. 5 indicates that all the films contain very broad υ1- & υ2-bands and D⁎- & G-bands Raman resonance peaks, which, again, is due to the smallness in grain size for the films. The Fe-coating/post-annealing processes insignificantly alter the Raman characteristics for the UNCD films. However, these processes markedly modify the surface morphology of the films. Fig. 6(a) and (b) show the SEM micrographs of the UNCDI films that were Fecoated/post-annealed at 900 °C(H2) and 850 °C(NH3), respectively.
It should be noted these samples are the ones which exhibit the best EFE properties among the Fe-coated/post-annealed samples. The microstructure of the UNCDI films has been completely changed. The equi-axised granular structure was no longer observable. There formed network-like surface structure with spherical particulates about 20–30 nm size (white dots) sparsely distributed on the surface of the samples. Similar granular structure was obtained for UNCDII for those which post-annealed at 900 °C(H2) and 850 °C(NH3) (Fig. 6(c) and (d)). The population of spherical particulates are more abundant for the UNCD films post-annealed in H2-environment (Fig. 6(a) and (c)), as compared with those post-annealed in NH3-environment (Fig. 6(b) and (d)). The phenomenon that Fe-coating/post-annealing processes can enhance the EFE properties of the diamond films has also been observed for microcrystalline diamond (MCD) films [15,16]. It was proposed that the Fe-coating on MCD films formed nano-sized Feclusters in the heating up stage of the post-annealing process [15,16]. The nano-sized Fe-clusters catalytically dissociated the carbon in sp 3-bonded diamond lattices at post-annealing temperature and then re-precipitated it out at the opposite side of Fe-clusters to form a-few-layer graphite. The Fe3C particulates were formed when the carbon contained in the Fe-clusters were frozen inside the Feclusters when the samples were cooled fast in a rate. The same phase transformation mechanism was presumed to occur in the Fe-coating/ post-annealing of UNCD films. The white spherical particulates observed in Fig. 6 are thus presumed to be the Fe3C particulates and the network-like surface structure is presumed to be the graphene-like phase formed by dissolution and re-precipitation process. To understand the possible formation process for the microstructure, the Fe-coated/post-annealing samples were examined using TEM. Figs. 7(a) and 8(a) show the typical bright field TEM micrograph of the UNCD films Fe-coated/post-annealed in H2(900 °C) and NH3(850 °C) atmosphere, respectively. Both films contain complicated microstructure. Selected area electron diffraction patterns (inset, Fig. 7(a) and (b)) indicate that the films consist of 3 diffraction rings that correspond to diamond lattices, which are designated as d111, d220, and d311, 2 faint diffraction rings nearby the d111 ring that correspond to Fe3C nano-particles and 1 diffuse ring in the center that corresponds to graphitic (or amorphous carbon) materials. To more clearly illustrate the phase constituents for the films, a dark field image was composed by superimposing the TEM dark field images taken from the diffraction spots corresponding to diamond, Fe3C and nanographite phase. Figs. 7(b) and 8(b) show the composed dark field images corresponding to Figs. 7(a) and 8(a), respectively, indicating that the Fe3C particulates, about 10–20 nm in size, and the nano-graphite (or amorphous carbon), are distributed evenly all over the samples. More detailed analysis using high resolution TEM indicates that the presence of Fe3C particulates (area 2, FT2, Fig. 7(c)) in the H2 (900 °C) post-annealed UNCD films, besides the diamond grains (area 1, FT1, Fig. 7(c)). Moreover, there forms graphite (or amorphous) phase on the other side of Fe3C particulates, against the diamond grains. The presence of graphite phase is even more clearly illustrated in Fig. 8(c), which correspond to the UNCD films postannealed in NH3-environment. Nano-sized clusters (a few nm in size, area 2) are observed in the vicinity of Fe3C particulates (area 1, FT1). The FT image (FT2) clearly illustrates that the nano-clusters in area 2 is graphite. Actually, the nano-graphite phase is observable surrounding the Fe3C particulates. Although these micrographs cannot clearly illustrate that the nano-graphitic phase is formed near every Fe3C particles due to the small abundance of these phases, it is believed that the formation process of graphene-like phase occurred to all of the Fe-clusters in the Fe-coated/post-annealed UNCD films. Restated, the Fe-coating form nano-clusters in the beginning stage of post-annealing process. The Fe-clusters then reacted with diamond films, dissolving the carbon species in the diamond and reprecipitating them out to form nano-graphite.
K.-Y. Teng et al. / Diamond & Related Materials 24 (2012) 188–194
191
Fig. 3. (a,b) The EFE properties of the UNCDI films, which grown in CH4(1%)/H2(0%)/Ar(99%) plasma and post-annealed at various temperature in H2(a) and NH3(b) atmosphere; (c, d) Those of UNCDII films, which were grown in CH4(1%)/H2(6%)/Ar(93%) plasma and post-annealed in H2(c) and NH3(d) atmosphere.
However, how the post-annealing in H2-environment leads to better EFE properties for the UNCD films than the post-annealing in NH3-environment is still not clear. SEM examinations indicate that post-annealing in H2-environment results in more abundant Fe3C particulates (cf. Fig. 6). Such a result implies that the reaction between diamond and Fe-clusters is somehow retarded in NH3-environment. The diffusion of carbon into Fe-clusters is thus hindered. However, more systematic studies in this phenomenon are required to elucidate the genuine mechanism. Moreover, while the EFE properties of these Fe-coated/post-annealed UNCD films are still inferior to the nanocarbon films [21–24], the silent feature of these films is that the
nano-graphite phase were precipitated out from the Fe-clusters. The nano-graphite clusters adhere very well with the underlying UNCD films and the substrates, while the nano-carbon films grown directly on Si-substrates usually suffer from poor adhesion. Therefore, the Fe-coating/post-annealing process are more robust in device applications. 4. Conclusions The Fe-layer coated on UNCD films reacted with diamond as soon as the post-annealing temperature reached 800 °C, resulted in the
Fig. 4. Variation of (a) turn-on field, E0, and (b) EFE current density, Je, of the Fe-coating/post-annealed UNCD films, which were grown in CH4/H2/Ar plasma with H2 = 0 or 6% and were post-annealed in H2 or NH3 atmosphere.
192
K.-Y. Teng et al. / Diamond & Related Materials 24 (2012) 188–194
Table 1 The EFE performance of as-deposited UNCD films (UNCDI and UNCDII) and those after Fe-coating/post-annealing in H2 (UNCDIa & UNCDIIa) or NH3 (UNCDIb & UNCDIIb) enrivonment. Samples Deposition parameters CH4/H2/Ar (sccm) UNCDI UNCDIa UNCDIb UNCDII UNCDIIa UNCDIIb
4/0/196 4/0/196 4/0/196 4/12/184 4/12/184 4/12/184
Annealing conditions temperature (atmosphere) (°C)
E0a (V/ μm)
Je b (μA/ cm2)
– 900 850 – 900 850
2.2 2.0 1.44 3.9 1.28 0.8
9 120 32 5 772 166
(H2) (NH3) (H2) (NH3)
a E0: turn-on field designated as the interception of the lines extrapolated from the high field and low field segments in F–N plot. b Je: field emission current density measured at 8.8 V/μm.
carbon dissolution and re-precipitation processes, which markedly enhanced the EFE process for UNCD films. However, to achieve a proper microstructure and to enhance the EFE process, not only is a sufficiently high post-annealing temperature (850 °C) required but also a highly active reducing atmosphere (H2 or NH3) is needed. Microstructural analysis indicates that the mechanism for the improvement of the EFE process is the formation of nanographites surrounding the Fe3C (or Fe) nanoclusters, which were formed via the reaction of Fe-clusters with diamond films. Presumably, the nanographites were formed by the re-precipitation of carbon species, which were dissolved in the Fe-clusters. The best EFE properties achieved were a turn-on field of E0 = 1.28 V/μm with an EFE current density of Je = 772 μA/cm 2 at an applied field of 8.8 V/μm, which were obtained for the films post-annealed in the H2 environment at 900 °C for 5 min. The degree of enhancement on the EFE properties varies with the granular structure of UNCD films and the post-annealing parameters (temperature and atmosphere).
Fig. 5. The Raman spectroscopy of the (a) UNCDI and (b) UNCDII films that were post-annealed in H2 and NH3, annealing parameters are Ia: 800 °C(H2), Ib: 850 °C(H2), and Ic: 900 °C(H2); IIa: 850 °C(NH3), IIb: 900 °C(NH3), and IIc: 950 °C(NH3).
Fig. 6. The SEM micrograph of the (a,b) UNCDI and (c,d) UNCDII films that were post-annealed at 900 °C(H2) and 850 °C(NH3).
K.-Y. Teng et al. / Diamond & Related Materials 24 (2012) 188–194
Fig. 7. (a) Bright field, (b) dark field, and (c) structure image, TEM micrographs of the UNCD films, which were post-annealed in H2 atmosphere (900 °C).
193
Fig. 8. (a) Bright field, (b) dark field, and (c) structure image, TEM micrographs of the UNCD films, which were post-annealed in NH3 atmosphere (850 °C).
194
K.-Y. Teng et al. / Diamond & Related Materials 24 (2012) 188–194
Acknowledgements This work was sponsored by the Tatung University, and Tamkang University, Taiwan, R.O.C. The authors deeply appreciated their financial and technical support. References [1] [2] [3] [4] [5]
J.E. Field, The Properties of Diamonds, Academic, London, 1979. H. Liu, D.S. Dandy, Diamond Relat. Mater. 4 (1995) 1173. J.C. Angus, H.A. Will, W.S. Stanko, J. Appl. Phys. 39 (1968) 2915. B.V. Spitsyn, L.L. Bouilov, B.V. Derjaguin, J. Cryst. Growth 52 (1981) 219. J. van der Weide, Z. Zhang, P.K. Baumann, M.G. Wensell, J. Bernholc, R.J. Nemanich, Phys. Rev. B 50 (1994) 5803. [6] J.B. Cui, J. Ristein, L. Ley, Phys. Rev. Lett. 81 (1998) 429. [7] M.W. Geis, N.N. Efremow, J.D. Woodhouse, M.D. Mcaleese, M. Marchywka, D.G. Socker, J.F. Hochedez, IEEE Electron Device Lett. 12 (1991) 456. [8] A.T. Sowers, B.L. Ward, S.L. English, R.J. Nemanich, Jpn. J. Appl. Phys. 86 (1999) 3973.
[9] P.T. Joseph, N.H. Tai, H. Niu, U.A. Palnitkar, W.F. Pong, H.F. Cheng, I.N. Lin, Diamond Relat. Mater. 17 (2008) 1812. [10] Y.C. Lee, S.J. Lin, I.N. Lin, H.F. Cheng, J. Appl. Phys. 97 (2005) 054310–054311. [11] C.F. Shih, K.S. Liu, I.N. Lin, Diamond Relat. Mater. 9 (2000) 1591. [12] Y.H. Chen, C.T. Hu, I.N. Lin, Appl. Phys. Lett. 75 (1999) 2857. [13] V.I. Polyakov, N.M. Rossukanyi, A.I. Rukovishnikov, S.M. Pimenov, A.V. Karabutov, V.I. Konov, J. Appl. Phys. 84 (1998) 2882. [14] A. Lamouri, Y. Wang, G.T. Mearini, I.L. Krainsky, J.A. Dayton Jr., W. Mueller, J. Vac. Sci. Technol. B 14 (1996) 2046. [15] P.C. Huang, W.C. Shih, H.C. Chen, I.N. Lin, Jpn. J. Appl. Phys. 50 (2011) 08KE04-1. [16] P.C. Huang, W.C. Shih, H.C. Chen, I.N. Lin, J. Appl. Phys. 109 (2011) 084309. [17] R.H. Fowler, L. Nordheim, Proc. R. Soc. London Ser. A 119 (1928) 173. [18] D. Zhou, D.M. Gruen, L.C. Qin, T.G. McCauley, A.R. Krauss, J. Appl. Phys. 84 (1998) 1981. [19] A.C. Ferrari, J. Robertson, Phys. Rev. B 63 (2001) 121405. [20] R.G. Buckley, T.D. Moustakas, L. Ye, J. Varon, J. Appl. Phys. 66 (1989) 3595. [21] P.G. Collins, A. Zettl, Appl. Phys. Lett. 69 (1996) 1969. [22] W. Zhu, C. Bower, O. Zhou, G. Kochanski, S. Jin, Appl. Phys. Lett. 75 (1999) 873. [23] A.N. Obraztsov, A.P. Volkov, I. Pavlovsky, Diamond Relat. Mater. 9 (2000) 1190. [24] Y. Chen, D.T. Shaw, L. Guo, Appl. Phys. Lett. 76 (2000) 2469.