Acessibilidade / Reportar erro

Carbon doping of InAlAs layers grown by metalorganic vapor phase epitaxy

Abstract

Carbon doped AlInAs layers grown by low pressure metalorganic vapor phase epitaxy are investigated as a function of growth temperature. Photoluminescence spectra show a significant drop in peak intensity as the growth temperature is reduced, due to the incorporation of non-radiative defects. It is shown that the C doping is not related to the deterioration of the optical properties of the layers. On the other hand, both the net hole concentration and the conductivity increase as the growth temperature is reduced due to a more efficient C incorporation and a lower incorporation of the passivating H atoms.


a32v322a

Carbon Doping of InAlAs Layers Grown by Metalorganic Vapor Phase Epitaxy

M. L. P. Ribeiro, B. Yavich** with the A. F. Ioffe Physico-Technical Institute, Saint-Petersburg, Russia. with the A. F. Ioffe Physico-Technical Institute, Saint-Petersburg, Russia. , C. V. B. Tribuzy, and P. L. Souza

Laboratório de Semicondutores, Centro de Estudos em Telecomunicações,

Pontifícia Universidade Católica do Rio de Janeiro,

Rua Marquês de São Vicente 225, Rio de Janeiro, 22453-900 RJ, Brazil

Received on 23 April, 2001

Carbon doped AlInAs layers grown by low pressure metalorganic vapor phase epitaxy are investigated as a function of growth temperature. Photoluminescence spectra show a significant drop in peak intensity as the growth temperature is reduced, due to the incorporation of non-radiative defects. It is shown that the C doping is not related to the deterioration of the optical properties of the layers. On the other hand, both the net hole concentration and the conductivity increase as the growth temperature is reduced due to a more efficient C incorporation and a lower incorporation of the passivating H atoms.

I Introduction

InAlAs doped with acceptors has many applications in electronic and optoelectronic devices. The p-type layer of a pin amplitude modulator structure, when InGaAs/InAlAs multiple quantum well (MQW) form the intrinsic region, is a p-type InAlAs layer. Normally, Zn is used as the p-type dopant and gives rise to excellent quality layers and high controllable doping levels. The disadvantage of this impurity is its large diffusivity [1]. Large diffusion coefficients imply in including a thick intrinsic undoped InAlAs layer between the MQW region and the doped layer to avoid Zn diffusion into the MQW region. This leads to a thicker total intrinsic region and therefore larger applied reverse voltages. Moreover, if one intends to use p-type delta-doped layers in the MQW structures [2] one should avoid Zn because the atomic diffusion will give rise to broad doping profiles. The alternative would be to use C, which is an acceptor by substituting As atoms.

Carbon has recently received much of attention as an intentional p-type dopant for InAlAs grown by metalorganic vapor phase epitaxy (MOVPE) due to its low diffusion coefficient [3] and high achievable doping level (over 1019 cm -3), which is attractive for fabricating a range of high-speed electronic and optoeletronic devices [4]. However, to obtain high doping levels one should use low growth temperatures (Tg), which, on the other hand, are undesirable because they favor defect incorporation. Therefore, knowledge of the limitations of C doping of InAlAs is fundamental for device design.

In this work we present results of a systematic investigation of transport and optical properties of C doped InAlAs layers grown at different temperatures. It is shown that for growth temperatures at which the transport properties of the layers are adequate, the optical characteristics are unappropriate.

II Experimental details

The lattice-matched InAlAs epitaxial layers were grown in a horizontal MOVPE Aixtron 200 reactor at 100 mbar at temperatures between 530ºC and 650ºC. Hydrogen was used as the carrier gas. The precursors are trimethylgallium (TMGa), trimethylaluminium (TMAl), arsine (AsH3), tetrabromide (CBr4) for Ga, Al, As and C respectively. Hall measurements were carried out using a Bio-Rad HL 5500 fixed magnetic field system in the standard Van der Pauw geometry at 300 K. X-ray measurements were made with a Bede Scientific QC2a x-ray diffractometer to determine the alloy composition. Secondary ion mass spectrometry (SIMS) was used to evaluate the C atomic concentration, [C]. The photoluminescence (PL) measurements were performed with the 514 nm line of an Ar + laser for excitation. The signal was dispersed by a 250 mm monochromator and detected by a nitrogen cooled Ge photodetector.

III Results and discussion

C doped InAlAs layers were grown with Tg varying between 530ºC and 650ºC. The growth temperature normally used for InAlAs is around 630 ºC. However, we find that if such high temperatures are used, no matter how high the IV/V flux ratio is, the layers are n-type and not p-type. Three different factors may contribute to this effect. First, C is an amphoteric dopant and its probability of occupying a site III or V is most likely temperature dependent. However, such an effect should be very weak [5] Second, it is well known that low Tgleads to a significant incorporation of defects, which can act as hole traps or compensating species. Finally, H atoms are incorporated with the C atoms, passivating them. Heat-treatments can be used to remove the H atoms and re-establish the C electrical activity.

Fig. 1 shows the two dimensional net free charge concentration, [c] as a function of Tgfor samples grown using the same CBr4 flux. It is clear that the p doping level increases as Tgis reduced. One can also see in Fig. 1 that for Tg¿ 575ºC, the samples are n-type. After growth they were subjected to a heat-treatment for 15 minutes under forming gas flow. In all cases, the free hole concentration, [p], increased with the annealing temperature, as shown in Figs 2a and 2b for samples grown at 650ºC and 530ºC, respectively, in agreement with results published by other authors[6, 7].




SIMS measurements were performed in order to determine [C] and to compare it with [c]. Table 1 shows the results of [C], [c] and the electrical activity efficiency of the C atoms for samples grown at three different temperatures. As Tg is reduced, not only the electrical activity increases but also the C atomic incorporation probability does. The electrical activity efficiency is defined as the ratio of [c] and [C]. The improved sticking coefficient for C at low Tg is attributed to the fact that AsH3 is only partially cracked at low temperatures favoring the incorporation of other species in the V sub-lattice. In fact, at low temperatures also oxygen is more effectively incorporated as a donor occupying sites in the V sub-lattice. [8]

Fig. 3 shows an increase in the conductivity as Tg is reduced due to the increase in [p]. One concludes that despite the more important incorporation of defects at low Tg, in particular of oxygen [8], the conductivity is improved. Thus, C doped InAlAs layers grown at temperatures of the order of 550ºC and subjected to a post-growth anneal are adequate, in terms of their transport properties, for application in electronic devices such as high speed transistors, as has already been discussed by other authors [4, 9]. However, for optical and optoelectronic devices it is crucial that good optical quality is also achieved.


We observe that as Tg decreases the optical quality of the layers gradually deteriorates. The near-edge optical transition decreases its intensity as Tgis reduced from 650 ºC and it is no longer observed when the growth temperature reaches 575ºC. This effect is independent of the presence of C impurities, since the same result is observed for the undoped reference samples. As shown in Fig. 1, only when Tg is lowered to 575 ºC the as-grown samples are p-type. In order to better understand this effect, one can think that lowering Tg, is the equivalent of decreasing the V/III ratio [10]. This is because at lower temperatures the cracking of the AsH3 is less efficient, as mentioned before. Hence, at low Tg, more defects are incorporated, as well as more C atoms and less passivating H atoms. As observed, the more efficient incorporation of C or its improved electrical activity are not related to the reduced PL intensity. Therefore, one concludes that some of the incorporated defects are very efficient non-radiative centers. The nature of such center is still under investigation.

The effect of the doping in the near-edge emission can be observed in Fig. 4, where the 30 K PL spectra for an undoped and a doped sample, both grown at 650ºC, are shown. In the case of the doped sample, the PL peak energy is 15 meV below that of the undoped reference sample. PL measurements as a function of excitation power (P) show that the peak energy blue-shifts as P increases for the doped sample, as expected from a donor-acceptor (D-A) type transition. Since this does not occur for the reference sample, one can conclude that the D-A transition is limited by the presence of C acceptors. The unintentionally incorporated donors compensate the acceptors. One should remember that these samples are n-type. The donors are most likely oxygen which are usually incorporated in Al containing material.


IV Conclusion

C doped InAlAs layers have been studied as a function of Tg. It was shown that the [p] and the conductivity increase as Tg decreases due to a more efficient C incorporation as well as a lower incorporation of passivating H atoms. Annealing the samples increases the electrical activity due to the removal of H atoms. It has been also observed that as Tg decreases, the intensity of the near-edge PL emission is reduced and for Tg < 575ºC, it is no longer observed. This effect is attributed to the incorporation of non-radiative defects at low Tg. It is suggested that TBAs, as As source, could be an alternative for AsH3 to obtain good optical quality C doped AlInAs, since its cracking temperature is lower that that of AsH3 [11], and consequently the low Tg would not lead to a significant incorporation of defects.

[2] P. L. Souza, C. V.-B. Tribuzy, S. M. Landi, M. P. Pires, R. Butendeich, A. C. Bittencourt, G. E. Marques and A. B. Henriques, Physica E to be published.

  • [1] K. Taeno and C. Amano, J. Crystal Growth 220, 393 (2000).
  • [3] T. H. Chiu and J. E. Cunningham, J. A. Dilzenberger, W. Y. Jan and S. N. G. Chu, J. Crystal Growth 111, 274 (1991).
  • [4] Hiroshi Ito and Tadao Ishibashi, Jpn. J. Appl. Phys 30, 944 (1991).
  • [5] C. R. Abernathy, S. J. Pearton, M. O. Manasreh, D. W. Fischer and D. N. Talwar, Appl. Phys. Lett. 57, 294 (1990).
  • [6] Abdallah Ougazzaden, Jay Holavanahalli, Michael Geva and Lawrence E. Smith, J. Crystal Growth 221, 66 (2000).
  • [7] Hiroshi Ito and Haruki Yokoyama, J. Crystal Growth 173, 315 (1997).
  • [8] S. Lea, F. Föhnsdorf, W. Stolz, R. Becker, A. Salzmann, A.Greiling, J. Crystal Growth 195, 98 (1998).
  • [9] T. Yamada, E. Tokumitsu, K. Saito, T. Akatsuka, M. Miyauchi, M. Konagai and K. Takahashi, J. Crystal Growth 95, 145 (1989).
  • [10] S. A. Stockman, A. W. Homsom and G. E. Stillmom, Appl. Phys. Lett. 60, 2903 (1992).
  • [11] P. Velling, Progress in Crystal Growth and Characterization of Materials 41, 85 (2000).
  • * with the A. F. Ioffe Physico-Technical Institute, Saint-Petersburg, Russia.
    with the A. F. Ioffe Physico-Technical Institute, Saint-Petersburg, Russia.
  • Publication Dates

    • Publication in this collection
      26 Nov 2002
    • Date of issue
      June 2002

    History

    • Received
      23 Apr 2001
    Sociedade Brasileira de Física Caixa Postal 66328, 05315-970 São Paulo SP - Brazil, Tel.: +55 11 3091-6922, Fax: (55 11) 3816-2063 - São Paulo - SP - Brazil
    E-mail: sbfisica@sbfisica.org.br