Metal to semiconductor transition in lightly Al doped ZnO thin films grown by sequential pulsed laser deposition

Amit K. Das*1, P. Misra1, R. S. Ajimsha1, D. M. Phase2 and L. M. Kukreja1 1

Laser Materials Processsing Division, Raja Ramanna Center for Advanced Technology,

Indore – 452 013 2

UGC-DAE Council for Scientific Research, Indore – 452 001

*Corresponding author e-mail- [email protected]

Abstract Metal to semiconductor transition (MST) was observed in the temperature dependent resitivity measurements in lightly Al doped ZnO (AZO) thin films with different Al concentration in the range from 0 to ~ 0.5 % grown by sequential pulsed laser deposition. MST transition temperature in the AZO films was found to decrease with increasing Al concentration. The ~ 0.5 % doped AZO film showed metallic behavior at all the temperature range without any MST. The MST and the associated variation of MST transition temperature with Al concentration was explained on the basis of competition between carrier activation behavior and various scattering mechanisms.

ZnO is an well known wide bandgap II-VI semiconductor with a direct bandgap of ~ 3.3 eV at room temperature1. As grown ZnO thin films grown by various techniques are generally found to be n-type conductive with resistivity ~ 10-1 ohm-cm due to the presence of native donor defects such as oxygen vacancies and Zn interstitials1. It is possible to further reduce the resistivity of ZnO thin films by intentionally doping ZnO with various n-type dopants such as Al, In, B, Ga and Si2-6. Such transparent and conducting thin films are very good candidate for application as transparent conducting electrodes in displays, solar cells, surface acoustic wave based device applications etc. Out of all the n-type dopants in ZnO, Al is highly suitable because of its easy availability, low cost, ease of doping and superior properties. Although there are numerous reports on Al doping of ZnO thin films to enhance its conductivity, studies on the metal to semiconductor transition (MST) in the temperature dependent resistivity, observed generally in such doped semiconductors5,7 are scanty. It is well known that as grown pure ZnO films are semiconducting with decreasing resistivity (ρ) on increasing temperature (dρ/dT –ve) and when doped with small amount of (~ 2%) n-type dopant, the films become metallic and show increasing resistivity with increasing temperature (dρ/dT +ve)5. Hence it is expected that the temperature dependent resistivity of lightly Al doped ZnO (AZO) thin films (with < 2% Al) should show MST. The metal to semiconductor transition and associated transition temperature is expected to show dependence on the Al doping concentration and can be probed in detail by studying temperature dependent variation of resistivity on a number of ZnO films doped with very small amount of Al (< 2%). Generally in PLD it is difficult to grow AZO films with such low concentrations because of the difficulty involved in making the corresponding pellets with homogeneous

Al distribution. This problem may be circumvented by using sequential PLD scheme6,8 where two independent targets of ZnO and Al2O3 are used to grow ZnO films with different Al concentration in precise manner. This method was earlier successfully utilized to grow Cd doped and Si doped ZnO thin films6,8. In this report we have grown highly c-axis oriented AZO thin films with different Al concentrations in the range of 0 to1% using sequential ablation of ZnO and Al2O3 pellets and studied their temperature dependent resistivity as a function of Al dopant concentration. Conspicuous metal to semiconductor transition was observed in these films and the corresponding transition temperature decreased with increasing Al concentration. This variation of MST transition temperature was explained using theoretical models of carrier activation and scattering. Details of these studies are presented in this paper.

Al doped ZnO thin films were deposited on c-axis sapphire substrates by sequential pulsed laser deposition technique wherein sintered ZnO and Al2O3 pellets were ablated sequentially many number of times to obtain films of desired thickness.

The

concentrations of Al in the films were varied by varying the ablation pulse ratio of Al2O3 and ZnO pellets. In one sequence Al2O3 pellet was ablated for one pulse and the number of pulses for the ZnO pellet were varied from 500 to 12 to obtain different concentrations of Al in the films. The growth was carried out using a KrF excimer laser (248 nm, 20 ns and 10 Hz repetition rate) at a fluence of ~ 0.8 J/cm2. Prior to deposition of the AZO films a high temperature ZnO buffer layer of thickness ~ 30 nm were grown on the sapphire substrate at 750oC. The AZO films were then grown over this buffer layer at 400oC. The concentrations of Al in the films were estimated from the ablation duration

ratios of Al2O3 and ZnO pellets, corrected for their individual growth ratios and from SIMS measurements. Variation of Al concentration with depth in the films was also studied using TOF SIMS measurements. TOF SIMS was performed using Bi+ primary ion beam and Cs+ sputtering beam. Presence and chemical state of Al in the films were evaluated by XPS measurements. The XPS studies were carried out using a photoelectron spectrometer at a base pressure better than 5 × 10−10 Torr. Al Kα radiation was employed for recording the spectra, with the source operated at an emission current of ~10 mA and an anode voltage of ~10 kV. A concentric hemispherical energy analyzer with 50 eV pass energy giving an overall resolution of ~0.8 eV was used. Au 4f7/2 at 84.7 eV served as an external reference. To correct the shifts in binding energies of core levels due to the charging effect, the graphitic C 1s peak at 284.7 eV was used as an internal reference. The chemical species were identified through the binding energies, which were determined by fitting the spectral line shapes with combination of Lorentzian and Gaussian functions. For electrical measurements In contact was made on the films. The ohmic character of the contacts was confirmed from I-V measurements. Electrical measurements of the samples were carried out in the 4-probe van der Pauw geometry using a lock-in amplifier.

Figure 1 shows the XPS spectrum of Al 2p peak in the film with Al concentration of ~0.50 %. The XPS peaks corresponding to Zn 2p and O1s were conspicuously observed at their respective binding energies (figure not shown). The Al 2p spectrum can be deconvoluted into two component peaks at ~73.8 eV and ~71.8 eV. The peak at 73.8 eV can be assigned to Al at Zn substitutional sites in an oxygen deficient ZnO matrix9 and

the lower energy peak at 71.8 eV can be assigned to metallic Al in the films which probably exist at the grain boundaries9. Efforts to detect Al by XPS for samples with Al concentrations lower than 0.50 % were not successful because of low concentration and low values of ionization cross section of Al9. Inset of figure 1 shows the typical SIMS depth profile of the AZO thin film with Al concentration of 0.05 % where variation of Al, Zn and O are shown as function of film thickness. It can be seen that the concentration of Al is nearly constant with film thickness upto depth of ~ 200 nm. Beyond ~200 nm the Al concentration gradually rises perhaps because of diffusion of Al from sapphire substrate in to growing ZnO film and then saturates at higher value corresponding to Al content in sapphire10. Although it is inappropriate to quantify Al concentration through these measurements, however it was observed that the ratio of SIMS counts corresponding to Zn to Al species in the films with different Al concentration showed nearly similar trend as that of ratio of ablation durations of individual targets. Normal ω-2θ and ω-rocking curves (figures not shown here) of the films showed highly c-axis oriented growth with good crystalline quality. Transmission spectra of the films showed that all the films were ~ 85 % transparent in the visible spectral range with a sharp band edge in the UV region. The band edge shifted from ~ 3.28 eV for undoped ZnO to ~ 3.53 eV for ~ 0.50 % Al doped ZnO. This blue shift is due to the filling of the conduction band of ZnO with the electrons donated by Al donors11,12. However the blue shift may be somewhat neutralized by the competing narrowing of the bandgap due to many body effects13. The total blue shift was explained by a combination of band filling effect and bandgap narrowing effect6.

The results of room temperature Hall measurements are shown in figure 2. As can be seen, with increasing Al concentration from 0 to ~ 0.50 % the resistivity of the films decreased from ~ 4.8x10-2 ohm-cm to ~ 8.8x10-4 ohm-cm. The minimum resistivity obtained in this study was ~ 2.4x10-4 ohm-cm for Al concentration of ~ 2.0 % (point not shown in the figure). Carrier concentration increased from 2.0x1019 cm-3 for 0.05 % AZO film to 1.1x1021 cm-3 for 0.50 % AZO film as seen in the inset of figure 2. Here it can be noted that the electron concentration could be controlled precisely in small steps due to the low doping level achievable using sequential PLD technique. The increase in electron concentration is primarily due to the donation of electrons from the Al donors to conduction band. The mobility is found to increase from ~ 28 cm2 V-1 s-1 for undoped ZnO to ~ 58 cm2 V-1 s-1 for ~ 0.50 % of Al doped films. The increase in mobility can be attributed to the increasing suppression of grain boundary scattering4 of free carriers at higher carrier concentration.

Figure 3 shows the results of the temperature dependent resistivity measurements of the films in the range 30 K to 300 K. The pure ZnO film shows typical semiconducting behavior where the resistivity increases with increasing temperature (figure not shown). However on doping with 0.05 % Al, the temperature dependent resistivity showed a metal to semiconductor transition (MST) at ~ 270 K as shown in fig 3(a) indicating that the films are metallic at temperature higher than ~ 270 K and semiconducting on the other side. In this context metallic behavior means the resistivity increases with increasing temperature and semiconducting behavior implies that the resistivity decreases with increasing temperature. The MST transition temperature was found to decrease from

~ 270 K to, ~ 50 K on increasing Al concentration from ~ 0.05 to ~ 0.25 %. It is explicitly shown in figure 4. The MST transition eventually vanished for the films with Al concentration of ~ 0.50 % and the film were completely metallic in the measurement temperature range. Although MST in the temperature dependent resistivity was earlier reported by Bhosle et al for Ga doped ZnO thin films5, however in that case the Ga concentration was much higher (~ 2-5%). Also the MST temperature was increasing with increasing Ga concentration. They attributed the MST to disorder induced weak localization of electrons which increased with increasing disorder thereby increasing transition temperature. The above stated weak localization model is not applicable in our case as our samples are lightly doped with Al (less than 0.5 %). Further the MST temperature in our case was found to decrease with increasing Al concentration. It our views the MST in lightly doped ZnO is not very well established and understood. However we found that it can be explained by considering the competition between the carrier activation behavior and scattering mechanisms at different temperatures.

It is well known fact that in general the metallic behavior is dominated by carrier scattering and the semiconducting behavior is dominated by carrier activation. Since we observe a combination of metallic and semiconducting behavior in the samples it may be possible to explain the observed MST in our sample by a combined mechanism of carrier scattering and activation. The resistivity ρ is given by

ρ=

1 neµ

(1)

Where n is the carrier concentration, e is the electronic charge and µ is the mobility. So the temperature dependence of ρ is decided by the temperature dependences of µ as well

as that of n. Since ZnO thin films with electron concentration higher than ~ 1x1019 cm-3 is degenerate14, the functional form of n for Al doped ZnO films can be written as

n = n0 (1 + n1 exp(−φ / kT ) )

(2)

Here we have assumed that the Al dopant atoms in the ZnO matrix are completely ionized and their concentration is given by temperature independent n0. We attribute the second term to be due to the activation of some native donors in ZnO with the activation energy of φ. T is the absolute temperature and k is the Boltzmann’s constant and n1 is a constant which is a measure of relative contribution of carrier activation from the native defects. The electron mobility is determined by various scattering mechanisms. In polycrystalline n-type doped ZnO thin films the primary mechanisms of electron scattering are grain boundary scattering, ionized impurity scattering and lattice vibration scattering and the total mobility is given by15 1 / µ = 1 / µ g + 1 / µi + 1/ µ ph

(3)

Where µg, µi and µph are the contribution to the mobility due to grain boundary scattering, ionized impurity scattering and lattice vibration scattering respectively. Grain boundary scattering arises due to the presence of potential barriers for electrons at the grain boundaries. Electrons can either be thermionically emitted across the grain boundary or tunneled through it. Out of this two mechanisms which one will be dominant is decided by the so called tunneling parameter E00 given by Crowell et al15,16, 1/ 2

E00 = 18.5 ×10

−12

 n   *  eV mε 

(4)

Where ε is the relative static dielectric constant, m* is the electron effective mass. For ZnO, ε = 8.65, m* = 0.28me 6, where me is the free electron mass. If E00 is larger than kT,

tunneling will be dominant. For n = 2x1019 cm-3, the value of E00 is ~ 53 meV which is greater than kT in the measurement temperature range. So it is expected that the current will be dominated by tunneling through the grain boundaries rather than thermionic emission15. Since tunneling is independent of temperature15,17, the mobility due to grain boundary scattering (µg) can also be taken temperature independent. In ZnO with an electron concentration of ~ 1x1019 cm-3, the Fermi level is expected to be in the conduction band (degenerate semiconductor)14. In such a case the ionized impurity scattering is also independent of temperature15. So mobility due to ionized impurity scattering (µi) is taken as constant. The lattice vibration scattering for degenerate semiconductor can be written similar to metals15 with µ ph ∝ T − p , where p is 1 if T is greater than Debye temperature18 and it ranges between 2 and 4 if T is less than Debye temperature19. Since for ZnO the Debye temperature is ~ 400 K which is much higher than our measurement temperatures, p is not likely to be 1. We take a value for p as 2 after Nistor et al20. Combining all the scattering mechanisms, the overall mobility will be given by, 1 / µ = C + DT 2 (5) C and D are constants. Combining equations (1), (2) and (5) we can write after simplification for resistivity ρ the following expression,

ρ=

A(1 + BT 2 ) (6) 1 + n1 exp(−φ / kT )

Where A, B, n1 are temperature independent constants and taken as fitting parameters. The fit of the equation (6) to the experimental data points are shown in the figure 2. As can be seen a very good fit with the experimental data is obtained with this equation. The

values of various fitting parameters for the samples are listed in Table 1. The value of the fitting parameter n1 gradually decreases with increasing Al concentration i.e., free electron concentration implying that the contribution of carrier activation is gradually decreasing with increasing free electron concentration. This is also supported by the decreasing values of the activation energy φ obtained from fitting of respective temperature dependent resistivity as shown in table 1. The decrease in the value of φ can be attributed to the increasing screening of the Coulomb potential of the donors with increasing electron concentration18 and/or band tailing arising due to heavy doping14.

The decrease in the MST temperature in lightly doped AZO thin films with increasing Al concentration can also be qualitatively explained by eq (6). As Al concentration increases i.e. electron concentration increases the activation energy of the native donors decreases. This is reflected by the decrease in the fitting parameters C and φ. Hence the contribution of carrier activation behavior in total resitivity decreases as compared to the contribution of corresponding carrier scatterings, thereby reducing the temperature at which the AZO transforms from metal to semiconductor (MST). In the AZO films with Al concentration at or above 0.50 %, the carrier activation behavior is completely absent and resistivity is totally dominated by carrier scatterings in the entire measurement temperature range typical of metallic behavior.

In conclusion, we have grown high quality lightly Al doped ZnO thin films with varying concentration of Al in the range of 0 to 0.50 % by sequential pulsed laser deposition technique and studied their electrical properties. Metal to semiconductor transition was

observed in the temperature dependent resitivity measurements in the samples with Al concentration below 0.50 %. MST transition temperature in AZO films was found to decrease with increasing Al concentration. The 0.5 % Al doped AZO film showed metallic behavior at all the temperature range without any MST. The MST and the associated variation of MST transition temperature with Al concentration was explained on the basis of competition between carrier activation behavior and various scattering mechanisms.

Acknowledgement: The authors thank Mr. R. Kumar and Dr. T. K. Sharma for their help in HRXRD measurements and Mr. A. Bose and Mr. S. C. Joshi for SIMS measurements. Authors also thanks Mr. A. Wadekar of UGC-DAE CSR, Indore for his help in XPS measurements.

Reference: 1

Ü. Özgür,Ya. I. Alivov, C. Liu, A. Teke,M. A. Reshchikov, S. Doğan,V. Avrutin, S.-

J. Cho, and H. Morkoç, J. Appl. Phys. 98, 041301 (2005). 2

T. Minami, H. Nanto and S. Takata, Jpn. J. Appl. Phys. 23, L280-L282 (1984).

3

A. Sarkar, S. Ghosh, S. Chaudhuri, A.K. Pal, Thin Solid Films, 204, 255-64 (1991).

4

J. Steinhauser, S. Faÿ, N. Oliveira, E. Vallat-Sauvain, and C. Ballif, Appl. Phys. Lett.

90, 142107 (2007). 5

V. Bhosle, A. Tiwari and J. Narayan, J. Appl. Phys. 100, 033713 (2006).

6

A. K. Das, P. Misra and L. M. Kukreja, J. Phys. D: Appl. Phys. 42, 165405 (2009).

7

Y. Furubayashi, T. Hitosugi, Y. Yamamoto, K. Inaba, G. Kinodo, Y.Hirose, T.

Shimada, and T. Hasegawa, Appl. Phys. Lett. 86, 252101 (2005). 8

P. Misra, P. K. Sahoo, P. Tripathi, V.N. Kulkarni, R. V. Nandedkar and L. M.

Kukreja, Appl. Phys. A 78, 37 (2004). 9

M. Chen, X. Wang, Y.H. Yu, Z.L. Pei, X.D. Bai, C. Sun, R.F. Huang, and

L.S. Wen, Appl. Surf. Sc. 158, 134 (2000). 10

Y.-G. Wang, N. Ohashi, Y. Wada, I. Sakaguchi, T. Ohgaki, and H. Haneda, J. Appl.

Phys. 100, 023524 (2006). 11

E. Burstein, Phys. Rev. 93, 632–3 (1954).

12

T. S. Moss, Proc. Phys. Soc. B 67, 775 (1954).

13

S. C. Jain and D. J. Roulston, Solid-State Electron. 34, 453–65 (1991).

14

J. G. Lu et al, J Appl. Phys. 101, 083705 (2007).

15

D. H. Zhang and H. L. Ma, Appl. Phys. A 62, 487-492 (1996).

16

C. R. Crowell, and V. Rideout, Solid-State Electron 12, 89 (1969).

17

S. Y. Myong, J. Steinhauser, R. Schlu¨chter, S. Fay, E. Vallat-Sauvain, A. Shah, C.

Ballif, and A. Ru¨fenacht, Solar Energy Materials & Solar Cells 91, 1269 – 1274 (2007). 18

N. F. Mott, Metal-Insulator Transition (Taylor and Francis, London, 1974).

19

X. D. Liu, E. Y. Jiang, and Z. Q. Li, J. Appl. Phys. 102, 073708 (2007).

20

M. Nistor, F. Gherendi, N. B. Mandache, C. Hebert, J. Perrière, and W. Seiler, J.

Appl. Phys. 106, 103710 (2009).

Figure Caption: Figure 1: XPS spectrum of Al 2p peak for the film with Al concentration of 0.50 %. The inset shows SIMS depth profile of the AZO film with Al concentration of 0.05 % where variation of concentration of Al (filled circles), Zn (filled triangles) and O (filled squares) are shown as function of film thickness. Figure 2: Variation of resistivity of the AZO films grown by sequential PLD as a function of Al doping concentrations. The inset shows the corresponding variation of electron concentration and mobility. Figure 3: Temperature dependent resistivity (filled circles) of the AZO films with Al doping concentrations of (a) 0.05 %, (b) 0.07 %, (c) 0.10 %, (d) 0.15 %, (e) 0.25 % and (f) 0.50 %. The red lines show the corresponding fitting of the data according to equation (6). Table 1: Values of different parameters of equation (6) obtained from fitting of the temperature dependent resistivity data. Al concentration

A

B

n1

(%)

φ (meV)

0.05

(7.560±.007)x10-3

(6.5±0.6)x10-7

0.34±.02

18.2±.5

0.07

(5.720±.003)x10-3

(1.18±.03)x10-6

0.29±.01

18.3±.3

0.10

(4.250±.003)x10-3

(1.70±.05)x10-6

0.23±.01

17.7±.5

0.15

(1.970±.007)x10-3

(1.82±.02)x10-6

0.15±.01

15.8±.3

0.25

(1.600±.001)x10-3

(2.12±.06)x10-6

0.10±.01

14.8±.9

0.50

(7.200±.005)x10-4

(3.00±.04)x10-6

0.10±.01

12.6±.7

Figure1:

Figure2:

Figure 3