Materials Research Bulletin 47 (2012) 1306–1309

Carrier concentration dependence of donor activation energy in n-type GaNepilayers grown on Si (1 1 1) by plasma-assisted MBE

Mahesh Kumar a,b, Thirumaleshwara N. Bhat a, Basanta Roul a,b, Mohana K. Rajpalke a,A.T. Kalghatgi b, S.B. Krupanidhi a,*a Materials Research Centre, Indian Institute of Science, Bangalore 560 012, Indiab Central Research Laboratory, Bharat Electronics, Bangalore 560 013, India

A R T I C L E I N F O

Article history:

Received 6 September 2011

Received in revised form 16 February 2012

Accepted 5 March 2012

Available online 13 March 2012

Keywords:

A. Nitrides

B. Epitaxial growth

D. Luminescence

A B S T R A C T

The n-type GaN layers were grown by plasma-assisted MBE and either intentionally doped with Si or

unintentionally doped. The optical characteristics of a donor level in Si-doped, GaN were studied in terms

of photoluminescence (PL) spectroscopy as a function of electron concentration. Temperature dependent

PL measurements allowed us to estimate the activation energy of a Si-related donor from temperature-

induced decay of PL intensity. PL peak positions, full width at half maximum of PL and activation energies

are found to be proportional to the cube root of carrier density. The involvement of donor levels is

supported by the temperature-dependent electron concentration measurements.

� 2012 Elsevier Ltd. All rights reserved.

Contents lists available at SciVerse ScienceDirect

Materials Research Bulletin

jo u rn al h om ep age: ww w.els evier .c o m/lo c ate /mat res b u

1. Introduction

Gallium nitride (GaN) and its related materials have beenwidely studied for their unique applications in optoelectronic andhigh temperature/high power electronic devices with relativelylow power consumption [1]. A number of studies have beenperformed successfully to grow GaN on different substrates such asAl2O3, SiC, GaAs and Si [2–5]. Silicon is considered to be one of themost promising candidates for the GaN epitaxy because of its manyadvantages such as high-quality, large size, low cost and a well-known existing device technology [6]. The origin of the residual n-type conductivity of undoped GaN layers grown on differentsubstrates, by various epitaxial techniques, is still relevant issues.To date, among all the reported buffer layers for GaN on Siheteroepitaxy, the AlN buffer layer approach yields the best resultsreported in the literature [7,8]. However, the mutual solubility of Aland Si is very high at the buffer-layer growth temperature (�820 8Cversus eutectic temperature 577 8C). Therefore, interdiffusion of Aland Si at the interface is severe, resulting in high unintentionaldoping levels in the epilayers and Si substrates [9]. To overcomethis serious drawback, it has been demonstrated that a siliconnitride buffer layer can be used for the low unintentionally dopedGaN growth [10–13]. The band-gap narrowing (BGN) effectinduced by many body Coulomb interactions is expected to playan important role in the optoelectronic properties of the nitrides

* Corresponding author. Tel.: +91 80 22932943; fax: +91 80 23607316.

E-mail address: sbk@mrc.iisc.ernet.in (S.B. Krupanidhi).

0025-5408/$ – see front matter � 2012 Elsevier Ltd. All rights reserved.

doi:10.1016/j.materresbull.2012.03.016

because of their large exciton binding energy and effective electronmass [14]. The BGN has been studied experimentally throughintentional doping, optical pumping and current injection [15–17].In addition, it has been reported that the potential fluctuation byrandomly distributed impurities significantly influences bothelectrical and optical properties of GaN [18,19]. Near-band edge(NBE) transition energy and Hall mobility of GaN:Si grown bymetal organic chemical vapor deposition (MOCVD), decreased withincreasing carrier concentration, indicating strong dependence ofthe photoluminescence (PL) transition energy on potentialfluctuation [20]. In this letter, we have studied the nature of theSi-related donor level in GaN epilayers on Si (1 1 1) substrate bystudying the luminescence properties as a function of carrierconcentration.

2. Experiments

The samples used for this study were grown by RF-MBE system(OMICRON) equipped with a radio frequency (RF) plasma sourceand the base pressure better than �1 � 10�10 mbar. The semi-insulating Si (1 1 1) substrates (resistivity > 3000 V cm) wereultrasonically degreased in isopropyl alcohol (IPA) for 10 min andboiled in trichloroethylene, acetone and methanol at 70 8C for5 min, respectively, followed by dipping in 5% HF to remove thesurface oxide. The substrates were outgassed at 900 8C for 1 h inultra-high vacuum. The samples were grown by nitridation–annealing–nitridation process, in which first the nitridation of thesubstrate was carried out at 530 8C for 30 min, followed byannealing at 900 8C for 30 min and again nitridation at 700 8C for

Fig. 2. NBE photoluminescence spectra of samples (a)–(e) taken at room

temperature. The peaks are at 3.4309, 3.4253, 3.4205, 3.4155 and 3.4123 eV,

respectively.

Table 1Room-temperature electronic properties of n-type GaN on Si (1 1 1).

Sample name Electron concentration (cm�3) Carrier mobility (cm2/V s)

(a) 7.25 � 1017 371.9

(b) 2.69 � 1018 247.3

(c) 5.65 � 1018 185.6

(d) 1.09 � 1019 78.8

(e) 1.60 � 1019 48.4

M. Kumar et al. / Materials Research Bulletin 47 (2012) 1306–1309 1307

30 min [21]. After nitridation, a low-temperature GaN buffer layerof 20 nm was grown at 500 8C, where the Gallium beam equivalentpressure was kept 5.6 � 10�7 mbar. Afterwards, 225 nm thick,undoped (sample (a)) and Si-doped (sample (b)–(e)) GaN epilayerswere grown on the buffer layer at 700 8C. Nitrogen flow rate andplasma power were kept at 0.5 sccm and 350 W, respectively fornitridation, buffer layer and GaN growth. The structural charac-terization of the samples was carried out by HRXRD. The PL spectrawere recorded in the temperature range of 5–300 K using a closedcycle optical cryostat and He–Cd laser of 325 nm excitationwavelength with a maximum input power of 30 mW. The Halleffect measurements were conducted in the temperature rangefrom 80 to 300 K at 0.5 T of magnetic field. Samples of5 mm � 5 mm size were cut from the wafers and metal dots werevacuum evaporated in the four corners to obtain electrical contactsin the Van der Pauw geometry.

3. Results and discussion

Fig. 1 shows the HRXRD 2u–v scans of the GaN films grown onSi (1 1 1) substrate. From the figures it can be seen that except thesubstrate peak, only a strong (0 0 0 2) GaN diffracted peak at2u = 34.598 and a weak (0 0 0 4) peak at 2u = 738 are present,indicating the epitaxial GaN thin film to be highly oriented alongthe [0 0 0 1] direction of the wurtzite GaN. The transportsmeasurements on the epi-GaN films were done using a Hallmobility setup. Hall Effect measurements revealed a strong n-typeconductivity, with donor concentrations of 7.25 � 1017 (sample(a), undoped), 2.69 � 1018 (sample (b), Si-doped), 5.65 � 1018

(sample (c), Si-doped), 1.09 � 1019 (sample (d), Si-doped) and1.60 � 1019 cm�3 (sample (e), Si-doped) at room temperature asshown in Table 1. The estimated electron mobilities are in goodagreement with reported values [22].

The investigations of the luminescence properties of GaN layerswere studied at room temperature by recording the photolumi-nescence spectra and are shown in Fig. 2. From the figure it can beseen that the PL spectra of all samples are dominated by NBE band,which generally contains the band-to-band transition as well asthe transition from neutral donor level to the valence band (DBE).Since donor levels form a band in the forbidden gap, the redshift ofthe NBE peak with increasing carrier concentration (i.e., higherdoping levels) is consistent with the DBE emission and mayindicate the broadening of the Si-related donor band as well BGNeffect [15]. Fig. 3 represents the PL peak positions of the sample as afunction of the cube root of carrier density, and in fact manifeststhe BGN effect due to Si doping. BGN effect can be evaluated by an

Fig. 1. HRXRD 2u–v scans of GaN on Si (1 1 1) substrate.

empirical relation, as reported by Lee et al. [23],

DEG ¼ EGð0Þ � EGðnÞ ¼ Kn1=3 (1)

where EG(0) is the band-gap energy of pure sample, EG(n) the band-gap energy depending on doping density, and n the electrondensity. The n1/3 dependence of DEG resembles the prevailingexchange contribution of electron–electron interaction [24]. Asshown in Fig. 3, our experimental data are fitted well by therelation shown in Eq. (1). The BGN coefficients (K) evaluated byfitting were estimated to be � �1.15 � 10�8 eV cm.

Fig. 3. Peak position of NBE transition as a function of n1/3, showing clear band-gap

narrowing effect. The solid lines are from the least square fit. The error in PL peak

position is ��1 meV.

Fig. 4. Full width at half maximum of the NBE transition as a function of carrier

concentration. The solid lines are from the least square fit. The error in FWHM of PL

is ��4 meV.

M. Kumar et al. / Materials Research Bulletin 47 (2012) 1306–13091308

Fig. 4 shows that the full width half maxima (FWHM) of NBEemission of PL became broadened with increasing carrierconcentration. Randomly distributed impurities unavoidably giverise to potential fluctuation or tail states of band edges, resulting inbroadening of the luminescence line. From the data we haveobtained, an empirical relation was fitted for FWHM of Si-dopedGaN as shown in Eq. (2).

DEðnÞ ðeVÞ ¼ 3:22 � 10�8 n1=3 (2)

where DE(n) is the FWHM of PL and n the electron density.The intensity of NBE luminescence was also monitored as a

function of temperature. It was observed that the intensity decays

Fig. 5. Arrhenius plot showing the Peak PL intensities of NBE transition as a function

of temperature. Activation energy was calculated by fitting Eq. (3).

with sample temperature T, in agreement with the followingexpression [25],

IðTÞ ¼ Io

½1 þ Co expð�Ea=kTÞ þ C1expð�Eloc=kTÞ� (3)

where I(T) is the PL intensity at temperature T, Co, C1 and Io areconstants, k is Boltzmann’s constant and Ea and Eloc are theactivation energy and localization energy in the high and lowtemperature regime, respectively. Fig. 5 shows the Arrhenius plotsof the peak PL intensities for the NBE transition related to PLemission as a function of temperature. The activation energies Ea

were obtained from the slopes of Arrhenius plot shown in Fig. 5. Incase of a DBE transition, Ea is related to the ionization energy ofdonor. Lower the value of the activation energy, the more likely isthe ionization of the donor by a conduction band hole (D0–e ! D+).Since an ionized level does not participate in recombination via theDBE route, the rate of these transitions (i.e., the intensity of theluminescence) decreases with Ea at any given temperature.Conversely, for a constant Ea, the intensity decays with increasingtemperature as more and more donors are ionized.

From Fig. 5 it can be seen that the activation energy shows asystematic dependence on the carrier concentration. The values ofEa are 27.5 � 0.20, 26.2 � 0.22, 24.5 � 0.25, 23.2 � 0.18 and22.1 � 0.16 meV were obtained for samples with carrier concentra-tions of 7.25 � 1017, 2.69 � 1018, 5.65 � 1018, 1.09 � 1019 and1.60 � 1019 cm�3, respectively. These values are in reasonableagreement with the ionization energy of Si donors predicted byJayapalan et al. [26] to have a value in the range 22–28 meV. Thephenomenon of variation of the dopant activation energy with carrierconcentration in semiconductors has been attributed to a number ofcauses. Among these are the formation of the band-tail states thatextend into the forbidden gap, the broadening of the donor band inthe gap, and the reduction of binding energy due to Coulombinteraction between the electrons in the conduction band and theionized donor states [27]. Furthermore, the decay of activation energywith carrier density, n, follows a common pattern observedpreviously in other semiconductors [28] and is described by anequation of the type

EaðNþD Þ ¼ Eað0Þ � aðNþD Þ1=3

(4)

where NþD is the concentration of ionized donors, Ea(0) is theionization energy at very low doping levels, and a is a constantaccounting for geometrical factors as well as for the properties ofthe material. Fig. 6 demonstrates that Eq. (4) provides a reasonablefit to the experimentally obtained activation energies under the

Fig. 6. Decrease of activation energy as a function of ionized donor concentration.

Fig. 7. Temperature dependence of electron concentration in sample (e).

M. Kumar et al. / Materials Research Bulletin 47 (2012) 1306–1309 1309

approximation that NþD � N�A ¼ n, where N�A is the density ofionized shallow acceptors. Due to compensation, the n-typeconductivity is determined by the difference between theconcentrations of ionized donors and acceptors. The value of awas found to be equal to �3.4 � 10�6 meV cm. Fig. 7 shows thecarrier concentration, determined from the Hall coefficient, as afunction of temperature for the sample (e). It can be seen from thefigure that as the carriers start freezing out at lower temperatures,the carrier concentration decreases with temperature and alsosupports the involvement of donor electrons.

4. Conclusions

The n-type GaN layers were grown by plasma-assistedmolecular beam epitaxy and either intentionally doped with Sior unintentionally doped. The variable-temperature photolumi-nescence studies of Si-doped n-type GaN allowed us to estimatethe activation energy of the Si-related donor in the range of 21–28 meV. The activation energy was found to be strongly dependentupon the electron concentration. A redshift of photoluminescencepeak with increasing carrier concentration was observed. The

involvement of donor levels is supported by the temperature-dependent electron concentration measurements.

References

[1] J.W. Yang, A. Lunev, G. Simin, A. Chitnis, M. Shatalov, M. Asif Khan, J.E. VanNostrand, R. Gaska, Appl. Phys. Lett. 76 (2000) 273.

[2] S.S. Tripathy, .J. Chua, P. Chen, Z.L. Miao, J. Appl. Phys. 92 (2002) 3503.[3] H. Nagano, Z. Qin, A. Jia, Y. Kato, M. Kobayashi, A. Yoshikawa, K. Takahashi, J. Cryst.

Growth 189/190 (1998) 265.[4] W.Y. Uen, Z.Y. Li, S.M. Lan, S.M. Li, J. Cryst. Growth 280 (2005) 335.[5] Y. Nakada, I. Aksenov, H. Okumura, Appl. Phys. Lett. 73 (1998) 827.[6] A. Dadgar, F. Schulze, M. Wienecke, A. Gadanecz, J. Blasing, P. Veit, T. Hempel, A.

Diez, J. Christen, A. Krost, New J. Phys. 9 (2007) 389.[7] K. Cheng, M. Leys, S. Degroote, B.V. Daele, S. Boeykens, J. Derluyn, M. Germain, G.V.

Tendeloo, G.E.J. Borghs, J. Electron. Mater. 35 (2006) 59.[8] A. Dadgar, M. Poschenrieder, J. Blasing, K. Fehse, A. Diez, A. Krost, Appl. Phys. Lett.

80 (2002) 3670.[9] E. Calleja, M.A. Sanchez-Garcıa, D. Basak, F.J. Sanchez, F. Calle, P. Youinou, E.

Munoz, J.J. Serrano, J.M. Blanco, C. Villar, T. Laine, J. Oila, K. Saarinen, Hautojarvi,C.H. Molloy, D.J. Somerford, I. Harrison, Phys. Rev. B 58 (1998) 1550.

[10] C.-L. Wu, J.C. Wang, M.H. Chan, T.T. Chen, S. Gwo, Appl. Phys. Lett. 83 (2003) 4530.[11] J. Huang, Z. Ye, L. Wang, J. Yuan, B. Zhao, H. Lu, Solid-State Electron. 46 (2002)

1231.[12] S. Sakai, T. Wang, Y. Morishima, Y. Naoi, J. Cryst. Growth 221 (2003) 334.[13] H.-M. Lee, C.T. Kuo, H.W. Shiu, C.H. Chen, S. Gwo, Appl. Phys. Lett. 95 (2009)

222104.[14] W.W. Chow, P.M. Smowton, P. Blood, A. Girndt, F. Jahnke, S.W. Koch, Appl. Phys.

Lett. 71 (1997) 157.[15] X. Zhang, S.-J. Chua, W. Liu, K.-B. Chong, Appl. Phys. Lett. 72 (1998) 1890.[16] S. Shokhovets, K. Kohler, O. Ambacher, G. Gobsch, Phys. Rev. B 79 (2009) 045201.[17] H.P.D. Schenk, S.I. Borenstain, A. Berezin, A. Schon, E. Cheifetz, S. Khatsevich, D.H.

Rich, J. Appl. Phys. 103 (2008) 103502.[18] E.F. Schubert, I.D. Goepfert, W. Grieshaber, J.M. Redwing, Appl. Phys. Lett. 71

(1997) 921.[19] T. Deguchi, A. Shikanai, K. Torii, T. Sota, S. Chichibu, S. Nakamura, Appl. Phys. Lett.

72 (1998) 3329.[20] E. Oh, H. Park, Y. Park, Appl. Phys. Lett. 72 (1998) 1848.[21] M. Kumar, B. Roul, T.N. Bhat, M.K. Rajpalke, P. Misra, L.M. Kukreja, N. Sinha, A.T.

Kalghatgi, S.B. Krupanidhi, Mater. Res. Bull. 45 (2010) 1581.[22] F. Schwierz, Solid-State Electron. 49 (2005) 889.[23] I.-H. Lee, J.J. Lee, P. Kung, F.J. Sanchez, M. Razeghi, Appl. Phys. Lett. 74 (1999) 102.[24] S.C. Jain, J.M. McGregor, D.J. Roulston, J. Appl. Phys. 68 (1990) 3747.[25] G. Davies, Phys. Rep. 176 (1989) 83.[26] J. Jayapalan, B.J. Skrommea, R.P. Vaudo, V.M. Phanse, Appl. Phys. Lett. 73 (1998)

1188.[27] O.L. Tirpak, L. Chernyak, Y.L. Wang, F. Ren, S.J. Pearton, K. Gartsman, Y. Feldman,

Appl. Phys. Lett. 90 (2007) 172111.[28] O.L. Tirpak, W.V. Schoenfeld, L. Chernyak, F.X. Xiu, J.L. Liu, S. Jang, F. Ren, S.J.

Pearton, A. Osinsky, P. Chow, Appl. Phys. Lett. 88 (2006) 202110.