Nano Express | Open | Published:
Electrical characterization and nanoscale surface morphology of optimized Ti/Al/Ta/Au ohmic contact for AlGaN/GaN HEMT
Nanoscale Research Lettersvolume 7, Article number: 107 (2012)
Good ohmic contacts with low contact resistance, smooth surface morphology, and a well-defined edge profile are essential to ensure optimal device performances for the AlGaN/GaN high electron mobility transistors [HEMTs]. A tantalum [Ta] metal layer and an SiNx thin film were used for the first time as an effective diffusion barrier and encapsulation layer in the standard Ti/Al/metal/Au ohmic metallization scheme in order to obtain high quality ohmic contacts with a focus on the thickness of Ta and SiNx. It is found that the Ta thickness is the dominant factor affecting the contact resistance, while the SiNx thickness affects the surface morphology significantly. An optimized Ti/Al/Ta/Au ohmic contact including a 40-nm thick Ta barrier layer and a 50-nm thick SiNx encapsulation layer is preferred when compared with the other conventional ohmic contact stacks as it produces a low contact resistance of around 7.27 × 10-7 Ω·cm2 and an ultra-low nanoscale surface morphology with a root mean square deviation of around 10 nm. Results from the proposed study play an important role in obtaining excellent ohmic contact formation in the fabrication of AlGaN/GaN HEMTs.
AlGaN/GaN high electron mobility transistors [HEMTs] are a promising technology for high-frequency, high-temperature, and high-power electronic devices due to the fact that AlGaN/GaN HEMTs have high breakdown voltage, high sheet carrier density, and high saturation current [1–4]. The formation of low-resistance ohmic contacts is a key issue in the fabrication process of AlGaN/GaN HEMTs, which are also required to have a smooth surface morphology and well-defined edge acuity [5–7]. The conventional ohmic contact metal uses a Ti/Al-based Ti/Al/Ti/Au or Ti/Al/Ni/Au structure [8–11]. Ti is essential because it participates in the reaction at the interface with nitrides to form TiN or AlTi2N layers by high-temperature rapid thermal annealing [RTA], and nitride vacancies are simultaneously formed at the AlGaN/GaN surface. At the same time, the diffused Ti and Al reduce the native gallium oxide on the AlGaN/GaN surface . Au is applied as an outer layer to prevent the oxidation of the Ti/Al metals during the RTA process. In addition, a diffusion barrier layer (Ti or Ni) is applied to prevent or minimize the Au upper layer from diffusing downward. However, this scheme often exhibits a very bumpy surface morphology and a significant lateral overflow during alloying at high temperature due to the intermixing of the Au and Al, which forms a viscous AlAu4 phase at high annealing temperatures. In order to minimize this phenomenon, special Au diffusion barrier layers, such as Pt  and Mo , and various studies regarding such factors as Ti/Al metal thickness adjustment, RTA condition changes, ion implantation annealing, recess etching of the ohmic contact, and epi-layer optimization have been carried out [15–17]. Unfortunately, these methods cannot solve all the problems of low contact resistance, smooth surface morphology, and well-defined line edge issues simultaneously.
In this paper, we present a report on an optimized Ti/Al/Ta/Au ohmic contact metallic system with an application of SiNx encapsulation layer before the annealing process, which results in a significantly better ohmic contact behavior for the AlGaN/GaN HEMTs. We present the comparative electrical, morphological, and microstructural properties for different ohmic metallic stacks. It has been found that the proposed ohmic metallization scheme using Ti/Al/Ta/Au with an SiNx encapsulation layer can provide a reliable solution for AlGaN/GaN HEMT applications by supplying superior electrical, morphological, and microstructural properties.
The epitaxial layers used in this work are grown using metal organic chemical vapor deposition on an Si (111) substrate, composed of a GaN buffer layer, followed by a 50-nm undoped GaN layer, a 3-nm undoped Al0.3Ga0.7N spacer, a 4.5 × 1018-cm-3 Si-doped Al0.3Ga0.7N layer, and capped by a 5-nm undoped Al0.3Ga0.7N layer. The mesa isolation was defined by inductively coupled plasma reactive ion etching [ICP-RIE] for 35 s using Cl2/BCl3 (6:1) mixing gas; the etching depth obtained was at least 200 nm. Various ohmic contact metallization structures were deposited by electron beam evaporation with a base pressure of 1 × 10-8 Torr and patterned by the lift-off process. Before evaporation, the surfaces of the samples were treated by a HCl/H2O = 1:10 solution and smooth O2/H2 plasma etching at 80°C for 60 s in order to remove the native oxide layer. The specific contact resistance [ρc] was measured through transmission line method measurements; the square (100 × 100 μm2) contacts were separated by 2, 4, 8, 16 and 32 μm. The surface morphology was characterized using a scanning electron microscope [SEM] and an atomic force microscope [AFM]. The root mean square [RMS] value for surface roughness of these metal contacts was measured from the analysis of a 10 × 10-μm2 area on the ohmic metal pattern. The exposed components of ohmic contact were detected using energy dispersive spectrometry [EDS] spectra. The schematic diagram of the AlGaN/GaN HEMT sample is shown in Figure 1.
Results and discussion
At the initial experiment, Ti/Al/Ta/Au ohmic metal stacks with different Ta thicknesses (20, 40, 60, 80 nm) were used and annealed in an N2-flowing ambient at various RTA temperatures for 30 s. The thickness of Ti, Al, and Au were fixed with 20, 80, and 100 nm, respectively. Figure 2 shows the typical behavior of the ρc values as a function of the annealing temperature for Ti/Al/Ta/Au ohmic contact systems with different Ta thicknesses. As shown in Figure 2, the ρc decreases as the annealing temperature increases from 830°C to 850°C, and the lowest value of 2.03 × 10-6 Ω·cm2 is obtained when the Ta thickness is 40 nm. Therefore, Ta thickness is the dominant factor affecting the ρc of Ti/Al/Ta/Au ohmic metal stack, and a Ta with 40-nm thickness is preferred for achieving ohmic contact with low contact resistance.
In order to evaluate the influence of the SiNx encapsulation layer on the performance of the ohmic contact, a divided thickness of 50, 100, 150, and 200 nm of SiNx was deposited by plasma-enhanced chemical vapor deposition [PECVD] using 320/9/2,000 sccm of SiH4 (5%) + He (95%)/NH3/N2 mixing gas, an RF power of 100 W, a chamber pressure of 1200 mTorr, and a chamber temperature of 250°C on the Ti/Al/Ta/Au (20/80/40/100 nm) ohmic metal. Figure 3 shows a visual surface morphology of the process control monitor pattern with different thicknesses of SiNx after the RTA process. As shown in Figure 3, the surface morphology decreases as the SiNx thickness increases from 50 to 200 nm. An encapsulation layer of SiNx with a 50-nm thickness is an effectual way to obtain ohmic contact with smooth surface morphology.
After fixing the thickness of the barrier layer Ta in the Ti/Al/Ta/Au ohmic contact stack and the thickness of the encapsulation layer SiNx, four different ohmic metallization schemes having the same metal thickness (20/80/40/100 nm) were prepared for comparison: sample 1 was Ti/Al/Ti/Au, sample 2 was Ti/Al/Ni/Au, sample 3 was Ti/Al/Ta/Au, and sample 4 was the optimized Ti/Al/Ta/Au with a 50-nm SiNx encapsulation layer. The as-deposited samples were then introduced into an RTA system that ranged from 830°C to 870°C for 30 s under an N2 atmosphere. After the RTA processing, the SiNx encapsulation layer was etched out using a dry/wet etching method in order to avoid damage to the epitaxy materials. First, 40 nm of the layer was dry etched by the low-damaged ICP-RIE, and then the remaining 10 nm was wet etched by a 1:6 buffer oxide etch solution. It has been found that PECVD-deposited SiNx can be successfully removed without damaging the contact metal and the epi-layer after a high-temperature RTA process. Figure 4 shows the typical behavior of the ρc as a function of the annealing temperature for the four ohmic contact systems. As shown in Figure 4, a better contact resistance is detected when the Ta is used as an effective diffusion barrier, and the lowest ρc value of 7.27 × 10-7 Ω·cm2 is obtained in sample 4 at an annealing temperature of 850°C and an annealing time of 30 s. It is noted that after annealing, samples 3 and 4 with the Ta diffusion barrier have significantly less metal inter-diffusion when compared with samples 1 and 2. A very limited in-diffusion of Au and out-diffusion of Ti and Al occurred through the Ta barrier. All of these factors resulted in a low contact resistance. SEM was used to characterize the film smoothness and edge acuity. The Ti/Al/Ti/Au ohmic contact developed a very rough surface morphology and considerably deteriorated edge acuity. Semispherical bulges possessing various sizes were found to be distributed randomly on the surface, some of which had cracks and breaks (see Figure 5a). The Ti/Al/Ni/Au had a slightly superior surface smoothness, but the edge of the contact metal became considerably wider after annealing at 850°C for 30 s (see Figure 5b) when compared with the Ti/Al/Ti/Au contact.
We posit that the Ti-Al or Ni-Al alloy aggregation in some local areas is the reason for the poor surface formation, and these precipitated Ti-Al or Ni-Al droplets are surrounded by Au-Al alloys. The surface morphology and edge acuity were improved in the Ti/Al/Ta/Au-based ohmic metallization as can be observed in Figure 5c,d. This could be due to the fact that the high melting point of metal Ta is used as a diffusion-impervious layer to suppress the inter-diffusion and reaction between the Au and Al layers in the contact stack which are the main causes of poor surface smoothness and edge acuity.
AFM was performed on the fabricated samples in order to check the surface roughness. Figure 6 shows the comparison of the surface morphology among the four sets of samples annealed at 850°C for 30 s. Consentaneous surface morphology information between the SEM and AFM can be obtained. A continuous increase in the metal roughness can be seen upon moving from sample 1 (see Figure 6a) to sample 4 (see Figure 6d). The RMS value of the surface roughness for the optimized Ti/Al/Ta/Au contact was only around 10 nm, which is significantly lower than the previously reported data of over 40 nm for Ti/Al/Ni/Au contacts [10, 17], and 36 nm for Ti/Al/Mo/Au . The ohmic contact performances of the Ti/Al/metal/Au schemes differ and are dependent on the nature of the metal barrier layer used. The fragmentation of the metal barrier layer into bulges has a strong implication on the nature of lateral flow. These bulges could act as physical impediments for the lateral flow of the Al-Au solid solution. The scheme with Ta is the most stable in terms of intermetallic reactions and surface morphology. Correlation between low contact resistance and low RMS value has been observed. Pretorius et al.  have presented a thermodynamic interpretation of the possible interactions that can take place between Al and many of the metal barrier layers. The effective concentration of the elements at the growth interface is taken to be that of the lowest temperature eutectic of the binary systems, and predictions were made about the effective heat of formation of various binaries: TiAl3 of -2.96 kJ/mol, NiAl3 of -5.32 kJ/mol, MoAl12 of -3.9 kJ/mol, and TaAl3 of -2.40 kJ/mol. The effective heat of formation was shown to be successful in correctly predicting the first and most thermodynamically feasible phases to form in metal-Al binaries. It can be clearly seen that Ta-Al alloy formation has among the lowest effective heats of formation. EDS experiments were conducted for all of the samples with the distance from AlGaN/GaN layers to the alloyed ohmic contact surface of 80 nm (Al area). Comparing the EDS analysis, a continuous decrease in the atom percentage of Al can be seen upon moving from sample 1 (see Figure 7a) to sample 4 (see Figure 7d); it is obvious that a good deal of the Al in sample 4 diffuses to the AlGaN and SiNx surface to create N vacancies and simultaneously reacts with the Ti and Au to confirm the formation of TiAl and AuAl phases. The concentration of N vacancies led to heavy doping and, hence, reduction in contact resistance . In addition, significant intermetallic diffusion and intermixing occurred due to the fact that in-diffusion of the Au towards the Al interface is detected with the atom percentages of 35.96 (sample 1), 34.44 (sample 2), and 33.3 (sample 3) rather than that of 26.11 found in sample 4, which is a possible reason for the improvement of the surface morphology and edge line definition in sample 4. The observed experiments strongly suggest that the optimized Ti/Al/Ta/Au ohmic scheme combines the potential of low contact resistance with superior surface morphology and edge acuity.
An optimized Ti/Al/Ta/Au metallization method has been successfully realized that yields improved contact characteristics that are superior to those of the conventional Ti/Al/Ti/Au or Ti/Al/Ni/Au scheme for AlGaN/GaN HEMTs. The effect of the thickness of Ta and encapsulated SiNx film for the proposed Ti/Al/Ta/Au contact has been studied. We have shown that the Ti/Al/Ta/Au metallization scheme encapsulated by a thin SiNx film before the annealing process leads to an excellent ohmic formation that not only obtains a low contact resistance, but also has a smooth morphology and favorable edge line definition. This optimized ohmic contact is most suitable in meeting the requirements of performance improvement for high-power AlGaN/GaN HEMT applications.
Camarchia V, Guerrieri SD, Pirola M, Teppati V, Ferrero A, Ghione G, Peroni M, Romanini P, Lanzieri C, Lavanga S, Serino A, Limiti E, Mariucci L: Fabrication and nonlinear characterization of GaN HEMTs on SiC and sapphire for high-power applications. Int J RF Microwave Comput Aided Eng 2006, 16: 70–80. 10.1002/mmce.20132
Boulay S, Touati S, Sar AA, Hoel V, Gaquiere C, De Jaeger JC, Joblot S, Cordier Y, Semond F, Massies J: AlGaN/GaN HEMTs on a (001)-oriented silicon substrate based on 100-nm SiN recessed gate technology for microwave power amplification. IEEE Trans Electron Devices 2007, 54: 2843–2848.
Faqir M, Verzellesi G, Meneghesso G, Zanoni E, Fantini F: Investigation of high-electric-field degradation effects in AlGaN/GaN HEMTs. IEEE Trans Electron Devices 2008, 55: 1592–1602.
Kim SC, Kim CK, Kim ES: Depth-of-focus and resolution-enhanced three-dimensional integral imaging with non-uniform lenslets and intermediate-view reconstruction technique. 3D Res 2011, 2: 1–9.
Lau WS, Tan JBH, Singh BP: Formation of ohmic contact in AlGaN/GaN HEMT structure at 500°C by ohmic contact recess etching. Microelectron Reliab 2009, 49: 558–561. 10.1016/j.microrel.2009.02.010
Piazza M, Dua C, Oualli M, Morvan E, Carisetti D: Degradation of TiAlNiAu as ohmic contact metal for GaN HEMTs. Microelectron Reliab 2009, 49: 1222–1225. 10.1016/j.microrel.2009.06.043
Gong RM, Wang JY, Liu SH, Dong ZH, Yu M, Wen CP, Cai Y, Zhang BS: Analysis of surface roughness in Ti/Al/Ni/Au ohmic contact to AlGaN/GaN high electron mobility transistors. Appl Phys Lett 2010, 97: 062115. 10.1063/1.3479928
Wang DF, Feng SW, Lu C, Motayed A, Jah M, Mohammad SN, Jones KA, Salamanca-Riba L: Low-resistance Ti/Al/Ti/Au multilayer ohmic contact to n-GaN. J Appl Phys 2001, 88: 6214–6217.
Chen J, Ivey DG, Bardwell J, Liu Y, Tang H, Webb JB: Microstructural analysis of Ti/Al/Ti/Au ohmic contacts to n-AlGaN/GaN. J Vac Sci Technol Part A Vac Surf Films 2002, 20: 1004–1010. 10.1116/1.1472428
Papanicolaou NA, Rao MV, Mittereder J, Anderson WT: Reliable Ti/Al and Ti/Al/Ni/Au ohmic contacts to n-type GaN formed by vacuum annealing. J Vac Sci Technol B Microelectron Nanometer Struct 2001, 19: 261–267. 10.1116/1.1331291
Roccaforte F, Iucolano F, Alberti A, Giannazzo F, Puglisi V, Bongiorno C, Di Franco S, Raineri V: Microstructure and current transport in Ti/Al/Ni/Au ohmic contacts to n-type AlGaN epilayers grown on Si(111). Superlattices Microstruct 2006, 40: 373–379. 10.1016/j.spmi.2006.06.017
Ruvimov S, Liliental-Weber Z, Washburn J, Duxstad KJ, Haller EE, Fan ZF, Mohammad SN, Kim W, Botchkarev AE, Morkoc H: Microstructure of Ti/Al and Ti/Al/Ni/Au ohmic contacts for n-GaN. Appl Phys Lett 1996, 69: 1556–1558. 10.1063/1.117060
Lee CT, Kao HW, Hwang FT: Effect of Pt barrier on thermal stability of Ti/Al/Pt/Au in ohmic contact with Si-implanted n-type GaN layers. J Electron Mater 2001, 30: 861–865. 10.1007/s11664-001-0072-5
Mohammed FM, Wang L, Koo HJ, Adesida I: Si-induced enhancement of ohmic performance of Ti/Al/Mo/Au metallisation for AlGaN/GaN HEMTs. Electron Lett 2005, 41: 984–985. 10.1049/el:20051849
Gillespie J, Crespo A, Fitch R, Jessen G, Via G: AlGaN/GaN ohmic contact resistance variations across epitaxial suppliers. Solid State Electron 2005, 49: 670–672. 10.1016/j.sse.2004.12.011
Recht F, McCarthy L, Rajan S, Chakraborty A, Poblenz C, Corrion A, Speck JS, Mishra UK: Nonalloyed ohmic contacts in AlGaN/GaN HEMTs by ion implantation with reduced activation annealing temperature. IEEE Electron Device Lett 2006, 27: 205–207.
Kim KH, Jeon CM, Oh SH, Lee JL, Park CG, Lee JH, Lee KS, Koo YM: Investigation of Ta/Ti/Al/Ni/Au ohmic contact to AlGaN/GaN heterostructure field-effect transistor. J Vac Sci Technol B Microelectron Nanometer Struct 2005, 23: 322–326. 10.1116/1.1856479
Mohammed FM, Wang L, Selvanathan D, Hu H, Adesida I: Ohmic contact formation mechanism of Ta/Al/Mo/Au and Ti/Al/Mo/Au metallizations on AlGaN/GaN HEMTs. J Vac Sci Technol B Microelectron Nanometer Struct 2005, 23: 2330–2335. 10.1116/1.2101691
Pretorius R, Vredenberg AM, Saris FW, Dereus R: Prediction of phase formation sequence and phase stability in binary metal-aluminum thin-film systems using the effective heat of formation rule. J Appl Phys 1991, 70: 3636–3646. 10.1063/1.349211
Chaturvedi N, Zeimer U, Wurfl J, Trankle G: Mechanism of ohmic contact formation in AlGaN/GaN high electron mobility transistors. Semicond Sci Technol 2006, 21: 175–179. 10.1088/0268-1242/21/2/014
This research was supported by the National Research Foundation of Korea (NRF) grant funded by the Korean government (MEST) (no. 2011-0030819). This work was also supported by the Research Grant of Kwangwoon University in 2012. We would like to thank Nano ENS, Inc. for their help in process development and Dr. Won-Sang Lee for the fabrication support of this program.
The authors declare that they have no competing interests.
CW conceived of the study, carried out the ohmic contact experiment, and drafted the manuscript. N-YK participated in the conception, guided the study, and revised the manuscript. All authors read and approved the final manuscript.