Intrinsic limitations to the doping of wide-gap semiconductors

https://doi.org/10.1016/S0921-4526(01)00417-3Get rights and content

Abstract

Doping limits in semiconductors are discussed in terms of the amphoteric defect model (ADM). It is shown that the maximum free electron or hole concentration that can be achieved by doping is an intrinsic property of a given semiconductor and is fully determined by the location of the semiconductor band edges with respect to a common energy reference, the Fermi level stabilization energy. The ADM provides a simple phenomenological rule that explains experimentally observed trends in free carrier saturation in a variety of semiconductor materials and their alloys. The predictions of a large enhancement of the maximum electron concentration in III–N–V alloys have been recently confirmed by experiment.

Introduction

It has been realized early on that many of the large variety of semiconductor materials are difficult to dope. The problem has been especially severe in wide-bandgap semiconductors where in many instances n- or p-type doping cannot be achieved at all, significantly limiting the range of applications of these materials [1], [2], [3]. These doping limitations have become even more important in new, emerging device technologies that put stringent demands on nanoscale control of electronic and structural properties of semiconductor materials. Such devices require the preparation of small size structures with very high doping levels and abrupt doping and composition profiles. The limits of the maximum doping levels and the question of the stability of dopant and compositional profiles are becoming key issues for many of the new devices.

The past several years have witnessed spectacular progress in the development of a new generation of short wavelength optoelectronic devices based on group III nitrides [4], [5], [6] and wide-gap II–VI semiconductors [7], [8]. In both cases this progress was made possible through the discovery of more efficient ways to activate acceptor impurities in these material systems. Despite this progress, the high resistance of p-type layers is still a major hurdle in the development of the devices requiring high current injection levels.

There have been numerous attempts to understand the maximum doping limits in semiconductors. Most of these were aimed at explaining limitations imposed on a specific dopant in a specific semiconductor. Thus, it has been argued that in the case of amphoteric impurities in III–V compounds, doping is limited by the impurities occupying both acceptor and donor sites, compensating each other. Redistribution of impurities can also lead to limitations of the maximum doping level in the materials with impurity diffusion strongly depending on the Fermi energy [3]. Formation of new stable solid phases involving dopant atoms can be a severe limitation in achieving high doping levels. This limitation depends on the chemical identity of the dopants and the host lattice elements and may be critical in the cases where only a limited number of potential dopants is available [9].

Passivation of donor and acceptor impurities by highly mobile impurities is another major mechanism limiting the electrical activity of dopants. Hydrogen, lithium and copper are known to passivate intentionally introduced dopants in semiconductors. Hydrogen has been an especially extensively studied impurity as it is a commonly used element in most semiconductor processing techniques and in all the growth techniques involving metalorganic precursors [10]. In some cases hydrogen can be removed during a post-growth annealing. Magnesium doped p-type GaN is frequently obtained by thermal annealing of MOCVD grown, hydrogen passivated films [11], [12]. However in other instances, as in the case of N doped ZnSe, hydrogen is too tightly bound to the N acceptors and cannot be removed by a thermal annealing [13].

Over the last few years a considerable effort has been directed towards overcoming the doping limits. For example it has been proposed that one can enhance incorporation of electrically active centers by co-doping with donors and acceptors. It has been argued, based on theoretical calculations that because of the reductions of the lattice relaxation and Madelung energies formation energies of proper donor acceptor complexes can be lower than the formation energy of isolated dopant species [13]. Some preliminary experimental results indicate that indeed the co-doping method has produced p-type ZnO that cannot be achieved by any other method [14]. Further studies are needed to fully understand the issues of poor reproducibility of the results obtained by the co-doping method.

In this paper, the saturation of the free carrier concentration in semiconductors obtained through doping will be discussed in terms of the amphoteric defect model (ADM). In recent years, the model has been successfully applied to numerous doping related phenomena in semiconductors. It has been used to explain doping induced suppression of dislocation formation [15] as well as impurity segregation [16], [17] and interdiffusion [18] in semiconductor superlattices. We will show that the ADM provides a simple phenomenological rule capable of predicting trends in the doping behavior of a large variety of semiconductor systems.

Section snippets

Amphoteric defect model

All point defects and dopants can be divided into two classes: delocalized, shallow dopants and highly localized defects and dopants. Shallow hydrogenic donors and acceptors belong to the first class. Their wave functions are delocalized and formed mostly out of the states close to the conduction band minimum or the valence band maximum. As a result the energy levels of these dopants are intimately associated with the respective band edges, conduction band for donors and the valence band for

Maximum doping limits in GaAs

Numerous electronic and optoelectronic applications have made GaAs one of the most extensively studied compound semiconductor. It has been realized very early that it is rather easy to dope GaAs with acceptors. Very high concentrations, in excess of 1020 cm−3, can be readily obtained by doping with group II atoms [34]. Even higher concentrations close to 1021 cm−3 were obtained by doping with carbon [35]. On the other hand n-type doping is much more difficult to achieve. The doping becomes less

Group III-Nitrides

Recent years have witnessed an unprecedented growth of interest in the Group III-Nitrides as a new distinct class of III–V compounds with strongly ionic bonds, smaller lattice constants and large band gaps. These materials form the foundation of a new technology for short wavelength optoelectronics [4] and high power, high-speed electronic devices [44]. Group III-Nitrides have by now been studied for many years. As with many other wide-gap materials the main impediment for practical

Group II–VI Semiconductors

Wide-gap group II–VI semiconductors are the group of materials that exhibit the most severe limitations on doping. Indeed, it is this family of materials for which the problem of doping has been recognized first [2]. Early studies have shown that all wide-gap II–VI compounds show a propensity for either n- or p-type conductivity. As grown ZnO, ZnS, HgSe, CdSe and CdS show n-type conductivity and p-type doping is very difficult if not impossible to achieve in these compounds. On the other hand

Group III–N–V alloys

An excellent example for the predictive power of the ADM has been the recently discovered high activation efficiency of shallow donors in GaInNAs alloys. It has been shown several years ago that alloying of group III–V compounds with group III-nitrides leads to dramatic change of the electronic properties of the resulting group III–N–V alloys [60]. For example GaNAs with only 1% of N has its band gap reduced by 0.18 eV [61]. We have shown recently that the reduction of the band gap results from

Other wide-gap semiconductors

As has been shown above, the ADM works well in explaining the doping limitations in a number of compound semiconductor systems. The obvious question arises whether it can provide any guidance in evaluating doping limits in elemental semiconductors. Fig. 8 shows the band offsets for group IV materials with energy gaps ranging from zero-gap in grey tin (α-Sn) to 5.5 eV in diamond. Since both Si and Ge have relatively small energy gaps it is not surprising that there are no serious limitations for

Conclusions

It has been shown that native defects in a semiconductor crystal lattice exhibit amphoteric behavior. Depending on the location of the Fermi energy they can act either as acceptors or donors. The demarcation energy separating donor- from acceptor-like behavior plays an important role of the energy at which the Fermi level is stabilized in the presence of large concentrations of native defects. It also serves as a convenient energy reference to evaluate the Fermi energy dependent part of the

Acknowledgements

The author would like to thank Eugene Haller for a critical reading of the manuscript. This work was supported by the Director, Office of Science, Office of Basic Energy Sciences, Division of Materials Sciences, of the US Department of Energy under Contract No. DE-AC03-76SF00098.

References (75)

  • F.A. Kroger et al.

    Solid State Phys.

    (1956)
  • T. Yamada et al.

    J. Crystal Growth

    (1989)
  • Y.-F. Wu et al.

    Solid-State Electron.

    (1997)
  • Y. Sato et al.

    J. Crystal Growth

    (1994)
  • W. Walukiewicz

    J. Crystal Growth

    (1996)
  • S.O. Ferreira et al.

    J. Crystal Growth

    (1994)
  • R.L. Longini et al.

    Phys. Rev.

    (1956)
  • E.F. Schubert, Doping in III–V Semiconductors, Cambridge University Press, Cambridge,...
  • S. Nakamura

    J. Vac. Sci. Technol. A

    (1995)
  • I. Akasaki et al.

    Electron. Lett.

    (1996)
  • P. Perlin et al.

    Phys. Rev. B

    (1992)
  • R.M. Park et al.

    Appl. Phys. Lett.

    (1990)
  • A. Ishibashi, Proceedings of seventh International Conference on II–VI Comp. and Devices, Edinburgh, Scotland, UK,...
  • D.B. Laks et al.

    Phys. Rev. B

    (1992)
  • E.E. Haller

    Semicond. Sci. Technol.

    (1991)
  • I. Akasaki et al.

    Jpn. J. Appl. Phys.

    (1997)
  • S. Nakamura et al.

    Jpn. J. Appl. Phys.

    (1991)
  • T. Yamamoto et al.

    Jpn. J. Appl. Phys.

    (1999)
  • M. Joseph et al.

    Jpn. J. Appl. Phys.

    (1999)
  • W. Walukiewicz

    Phys. Rev. B

    (1989)
  • W. Walukiewicz

    Mater. Res. Soc. Symp. Proc.

    (1993)
  • W. Walukiewicz

    Inst. Phys. Conf. Ser. No

    (1994)
  • W. Walukiewicz

    Appl. Phys. Lett.

    (1989)
  • A. Zunger

    Ann. Rev. Mater. Sci.

    (1985)
  • J.M. Langer et al.

    Phys. Rev. Lett.

    (1985)
  • D.D. Nolte et al.

    Phys. Rev. Lett.

    (1987)
  • V.N. Brudnyi, V.A. Novikov, Fiz. Tekh. Poluprovodn. 19 (1985) 747 [Sov. Phys. Semicon. 19 (1985)...
  • V.N. Brudnyi, M.A. Krivov, A.I. Potapov, V.I. Shakhovostov, Fiz. Tekh. Poluprovodn 16 (1982) 39 [Sov. Phys. Semicon. 16...
  • J.W. Cleland et al.

    Phys. Rev.

    (1955)
  • V.N. Brudnyi, V.A. Novikov, Fiz. Tkh. Poluprovodn. 16 (1982) 1880 [Sov. Phys. Semicon. 16 (1983)...
  • N.P. Kekelidze, G.P. Kekelidze, Radiation Damage and Defects in semiconductors, Vol. 16, Inst. Phys. Conf. Series, IOP,...
  • T.V. Mashovets, R.Yu. Khansevarov, Fiz. Tverd. Tela 8 (1966) 1690 [Sov. Phys-Solid State 8 (1966)...
  • W. Walukiewicz

    Phys. Rev. B

    (1988)
  • W. Walukiewicz

    J. Vac. Sci. Technol. B

    (1988)
  • W. Walukiewicz

    J. Vac. Sci. Technol. B

    (1987)
  • G.A. Baraff et al.

    Phys. Rev. Lett.

    (1985)
  • S.B. Zhang et al.

    Phys. Rev. Lett.

    (1991)
  • Cited by (0)

    View full text