• PHASE FORMATION AND MORPHOLOGY IN NICKEL AND NICKEL/GOLD CONTACTS TO GALLIUM NITRIDE
  • ABSTRACT
  • INTRODUCTION
  • EXPERIMENTAL PROCEDURE
  • RESULTS
    • Phase formation in Ni/GaN
    • Morphology of Ni films on GaN
    • Morphology of Ni/Au contacts to GaN
  • CONCLUSIONS
  • ACKNOWLEDGMENTS
  • REFERENCES

PHASE FORMATION AND MORPHOLOGY IN NICKEL AND NICKEL/GOLD CONTACTS TO GALLIUM NITRIDE

ABSTRACT

Metallurgical reactions between Ni and GaN have been explored at temperatures between 400 and 900 °C in N₂, Ar, and forming gas. A trend of increasing Ga content in the reacted films was observed with increasing temperature. The reactions are consistent with the thermodynamics of the Ni-Ga-N system. Changes in the film morphology on annealing were also examined. Metal island formation and a corresponding deep, non-uniform metal penetration into GaN were observed at high temperatures. The relevance of the observed nature of phase formation and morphology in these thin films to electrical properties of Ni/GaN and Au/Ni/GaN contacts is also considered.

Keywords: Morphology, Ni contacts, Ni/Au contacts, GaN
PACS:

INTRODUCTION

The III-V nitrides are receiving considerable attention for optoelectronics in the blue and UV wavelengths as well as for their potential for high temperature electronics [1]. Therefore, ohmic contacts and Schottky barriers to GaN are of considerable interest. Since Ni has been used in Ni/Au ohmic contacts to p-type GaN [2, 3] and has also been examined as a Schottky barrier to n-type GaN [4], an understanding of the metallurgy of the Ni/GaN system is clearly needed.
Bermudez et al. [5] examined the growth of thin Ni films on GaN and noted that a pronounced interfacial reaction occurred upon annealing above 600 °C in vacuum. Guo et al. [4] studied the thermal stability of Ni Schottky contacts on n-type GaN. According to these investigators, Ga₄Ni₃ was identified by XRD along with Ni in the as-deposited film. However, no intermixing between Ni and Ga was observed in the corresponding SIMS spectra. On subsequent annealing in a N₂ atmosphere at 200 °C and 400 °C, nickel nitrides were identified by XRD. The reported reaction between Ni and GaN at room temperature is surprising from a kinetic standpoint and also given that thermodynamic estimates [6] suggest that Ni could actually be thermodynamically stable on GaN at room temperature. Thus, an improved understanding of the metallurgy of the reactions between Ni thin films and GaN is clearly desirable.
Changes in film morphology of M/GaN thin films (M= Pt, Pd, Ni) on annealing have been investigated using Rutherford backscattering spectroscopy (RBS) and scanning electron microscopy by Duxstad et al. [7]. These authors observed that for Ni on GaN, islands began to form above 700 °C. However, the effect of this metal non-uniformity on the uniformity of the metal/GaN reaction front have not been reported. If the metal in localized regions has reacted through the GaN layer in a device, the device could be rendered inoperable. In addition, it has not been reported whether this island formation is seen upon annealing Ni/Au films on GaN. This system is important for ohmic contacts to p-GaN [2, 3]. Anneal temperatures of 400 °C and 600 °C have been reported for these contacts [8, 9].
In this study, metallurgical reactions between Ni and GaN have been explored at temperatures between 400 °C and 900 °C in N₂, Ar, and forming gas. Glancing angle x-ray diffraction and Auger depth profiling were employed to determine the extent of interdiffusion between Ni and GaN and identify the phases that form upon annealing. Scanning electron microscopy (SEM) of the annealed films was performed, followed by atomic force microscopy (AFM) of the GaN surface after the metal contact was etched away.

EXPERIMENTAL PROCEDURE

Gallium nitride films (ATMI) with n = 10¹⁷ cm^-3 were grown by MOCVD on (0001) Al₂O₃ substrates. The GaN surface was degreased in methanol, etched in 1:1 HCl:DI, rinsed in DI, and dried in N₂ gas immediately before the sample was loaded into a vacuum deposition system. Ni films of 500 Å thickness were sputtered onto the GaN at a rate of 2.0 Å/s. For Ni/Au contacts, Ni film of 500 Å thickness was first sputtered onto GaN followed by a Au film of 1000 Å.
The Ni/GaN specimens were then annealed at various temperatures between 400 °C and 900 °C in N₂ (99.99 % purity), 95 % N₂+ 5 % H₂ (forming gas), or Ar (99.99 % purity). A few of the specimens were annealed in a conventional tube furnace while others were annealed in an AG 610 rapid thermal annealing (RTA) furnace, in each case under flowing gas atmospheres. In order to determine if more extensive reaction would occur during prolonged annealing, a few samples were sealed in quartz tubes under vacuum and annealed in a box furnace for a week.
Glancing angle XRD, using Cu K radiation, was performed with a Philips diffractometer. Auger electron spectroscopy (AES) was performed using a Kratos scanning Auger microprobe with a 3 keV electron beam. The sample was simultaneously argon ion sputter-etched to obtain a depth profile.
Scanning electron microscopy of the annealed films was performed using an ISI SX-40A SEM at an acceleration voltage of 25 kV. Elemental analysis was obtained by energy dispersive spectroscopy (EDS) of the individual regions. For the AFM study, GaN films supplied by Cree Research were used. Nickel and Ni/Au films were deposited on the GaN by sputtering and subsequently annealed at various temperatures. The annealed Ni/GaN and Au/Ni/GaN thin films were then etched with 3:1 HNO₃:HCl and 1:5 HNO₃:HCl to dissolve Au and Ni, respectively. Subsequently, AFM was performed on the GaN surface to characterize the uniformity of reaction between the metal films and GaN.

RESULTS

Phase formation in Ni/GaN

XRD of the as-deposited Ni film on GaN indicates the presence of polycrystalline Ni on GaN. Our result is in contrast to that of Guo et al. [4], who reported Ga₄Ni₃ to be present in the as-deposited Ni film on GaN. To investigate the extent of reaction between Ni and GaN at 400 °C in forming gas (95% N₂ + 5% H₂), a Ni/GaN film was annealed by RTA for 10 min. in this atmosphere. No reaction between Ni and GaN was detected by XRD. The conditions were chosen to simulate those used by Trexler et al. [9] for the anneal of Au/Ni/p-GaN ohmic contacts. To investigate whether a reaction could be detected after more prolonged annealing, a sample was then subjected to flowing N₂ gas for 17 hours at 400 °C. The interface between Ni and GaN is as sharp after annealing for 17 hours at 400 °C as it was in the profile of the as-deposited sample. The XRD scan also showed no indication of any new phase formation or of any alteration in the lattice parameter of Ni. Our results are consistent with a recent report by Ishikawa et al. [10] who did not observe a reaction between Ni and GaN when the contacts were annealed at 400 °C for 10 min.
Nickel/GaN films were annealed in a tube furnace for periods of 1 and 24 hours in flowing Ar and N₂ atmospheres at 600 °C. After annealing in flowing Ar, the diffraction pattern was consistent with a face-centered cubic phase. The lattice parameter varied from 3.53 to 3.55 Å as the annealing time was increased from 1 to 24 hours. In flowing N₂, the corresponding variation was from 3.52 to 3.55 Å. The observations of an increasing lattice parameter along with an increased Ga signal in the thin film (Auger depth profiles) are consistent with the significant solubility of Ga in face-centered cubic Ni (up to 15% Ga at 600 °C, with an increase in the lattice parameter to 3.552 Å [11, 12]).
The x-ray diffraction peaks observed after annealing at 600 °C were again observed after annealing for 1 hour in Ar at 750 °C, although the peaks were shifted, revealing an increase in the lattice parameter after annealing at 750 °C. Cubic lattice parameters of 3.56 Å and 3.58 Å were observed after annealing for 1 hour in Ar and N₂ respectively. Auger depth profiles indicate that the extent of Ga dissolution is greater at 750 °C than at 600 °C, and that significantly more Ga than N is present in the film. Thus, N₂ gas has been released to the annealing environment, even when this annealing environment was N₂ at 1 atm. The identity of the Ni-Ga phase present in the samples subjected to prolonged annealing at 750 °C is not completely clear. From the lattice parameter alone, the phase is either the disordered face-centered cubic solution of Ni and Ga, near the limit of Ga solubility, or the ordered Ni₃Ga phase.
The x-ray diffraction plot for a Ni/GaN sample annealed for 10 minutes in N₂ atmosphere indicates the presence of a cubic phase with a lattice parameter of 2.89 Å in both Ar and N₂ atmospheres, consistent with the B2 NiGa phase. A comparison of the Auger depth profiles at 750 and 900 °C also indicates that the Ga/Ni ratio in the reacted film is approximately three times higher at 900 °C as compared to that at 750 °C. The reaction products in the Ni/GaN system at various times and temperatures are shown in Table 1. As predicted by the calculations of Mohney et al. [6], the trend of Ga incorporation into the reacted film, accompanied by the release of N₂ gas, was observed in this experimental study.

* More Ga compared to the Ni-Ga solid solution at 600 °C
Table 1. Reaction products in Ni/GaN contacts.

Morphology of Ni films on GaN

Our SEM studies indicate that the thin film coverage on GaN is almost uniform at 600 °C and 750 °C, with Ni-rich non-uniformities present over approximately 5% of the surface of the contact annealed at 750 °C. However, upon annealing at 900 °C, the film exhibits considerable non-uniformity whereupon Ni islands form to reduce the overall interfacial energy of the system. Since this process occurs by surface diffusion, it is faster at high temperatures. In addition, at high temperatures, a greater amount of GaN is consumed by reaction. (More Ga-rich nickel gallides are formed.) This two-fold process of enhanced island formation along with increased extent of reaction could cause metal gallides formed beneath the islands to spike deeply into the GaN. To determine the extent of metal penetration into GaN, the annealed surfaces were etched and examined using AFM (Figure 1). Atomic force microscopy of the surface of an etched contact that was not annealed is also shown for comparison (Figure 2). A surface roughness of 100 Å was observed. After etching the metal from the Ni(500 Å)/GaN sample annealed at 800 °C, pits as deep as 800 Å can be observed. In other words, metal in the islands reacts with as much as 800 Å of GaN. This would certainly exert an influence in the electrical properties of the Ni/GaN interface.

Figure 1. AFM of the GaN surface revealed after a Ni/GaN contact annealed at 800 °C for 1 min. in nitrogen was etched away.

Figure 2. AFM of the GaN surface of an etched Ni/GaN contact that was not annealed.

Morphology of Ni/Au contacts to GaN

Film morphology upon annealing the Ni(500 Å)/Au(1000 Å) contact on GaN was also examined. No islands were observed to form after short anneals at 400 °C and 600 °C in N₂. The onset of island formation was observed by SEM for a 1 minute anneal at 700 °C in N₂. On etching the contact and examining the surface by AFM, pits as deep as 5000 Å were observed. On increasing the anneal temperature to 800 °C, large islands were observed to form by SEM (Figure 3). The underlying GaN between the metal islands was clearly visible by optical microscopy. After etching the contact and examining the surface by AFM, pits as deep as 2000 Å were observed (Figure 4). In other regions of the contact, pits as deep as 5000 Å were observed. Furthermore, the density of pits on the GaN surface observed after the 800 °C anneal is more than an order of magnitude higher than the pit density observed after the 700 °C anneal. In the Nichia InGaN-AlGaN-GaN double heterojunction LED [1] and in the InGaN single quantum well (SQW) LD [13], the p-GaN layer has a thickness of 3000 Å and 5000 Å, respectively. A typical contact to p-GaN is Ni/Au [2-3, 13]. Our study indicates that for these p-layer thicknesses, a Ni(500 Å)/Au(1000 Å) contact could react through the entire p-GaN layer in localized regions when annealed at temperatures higher than 700 °C, rendering the device inoperable. Hence, anneals of Ni/Au contacts on GaN at temperatures as high as 700 °C may not be suitable for many device structures.

Figure 3. SEM of Au(1000 Å)/Ni(500 Å)/GaN contact after annealing at 800 °C for 1 minute in nitrogen.

Figure 4. AFM of an annealed Au(1000 Å)/Ni(500 Å)/GaN contact after etching the metal away.

CONCLUSIONS

In this study, metallurgical reactions between Ni and GaN have been explored at various temperatures between 400 and 900 °C in N₂, Ar, and forming gas. A trend of increasing Ga content in the reacted films was observed with increasing temperature, while the formation of nickel nitrides was not observed. The observed reactions are consistent with the thermodynamics of the Ni-Ga-N system. Changes in the film morphology upon annealing were also examined. Metal island formation and corresponding deep, non-uniform penetration of metal into GaN were observed at high temperatures. These non-uniformities in film morphology could be significant in determining the electrical properties of Au/Ni/p-GaN and Ni/n-GaN contacts after high temperature annealing.

ACKNOWLEDGMENTS

This work was supported by DARPA (Anis Husain) through AFOSR grant F49620-95-1-0516.

REFERENCES

1. H. Morkoc and S. N. Mohammad, Science 267, p. 51 (1995).
2. S. Nakamura, M. Senoh, and T. Mukai, Appl. Phys. Lett. 62, p. 2390 (1993).
3. S. Nakamura, T. Mukai, and M. Senoh, Appl. Phys. Lett. 64, p. 2557 (1994).
4. J. D. Guo, F. M. Pan, M. S. Feng, R. J. Guo, P. F. Chou, and C. Y. Chang, J. Appl. Phys. 80, p. 1623 (1996).
5. V. M. Bermudez, R. Kaplan, M. A. Khan and J. N. Kuznia, Phys. Rev. B. 48, p. 2436 (1993).
6. S. E. Mohney and X. Lin, J. Electron. Mater. 25, p. 811 (1996).
7. K. J. Duxstad, E. E. Haller, W. R. Imler, L. T. Romano, K. M. Yu, F. A. Ponce, and D. A. Steigerwald, in III-V Nitrides, F. A. Ponce, T. D. Moustakas, I. Akasaki, and B. A. Monemar (Eds.) (Mater. Res. Soc. Symp. Proc. 449, 1997).
8. J. T. Trexler, S. J. Pearton, P. H. Holloway, M. G. Mier, K. R. Evans and R. F. Karlicek, ibid.
9. J. T. Trexler, S. J. Miller, P. H. Holloway, and M. A. Khan, in Gallium Nitride and Related Materials, F. A. Ponce, R. D. Dupuis, S. Nakamura, and J. A. Edmond (Eds.) (Mater. Res. Soc. Symp. Proc. 395, 1996), p. 819.
10. H. Ishikawa, S. Kobayashi, Y. Koide, S. Yamasaki, S. Nagai, J. Umezaki, M. Koike, and M. Murakami, J. Appl. Phys. 81 (3) p. 1315 (1997).
11. P. Feschotte and P. Eggimann, J. Less-Common. Met. 63, p. 5 (1979).
12. S. Y. Lee and P. Nash in Phase Diagrams of Binary-Nickel Alloys, P. Nash (Ed.), ASTM International, Materials Park, OH, 1991, p. 133.
13. S. Nakamura, in Gallium Nitride and Related Materials, F. A. Ponce, R. D. Dupuis, S. Nakamura, and J. A. Edmond (Eds.) (Mater. Res. Soc. Symp. Proc. 395, 1996), p. 879.

© 1997 The Materials Research Society