Using the acetylacetonates of zinc and aluminium for the Metalorganic Chemical Vapour Deposition of aluminium doped zinc oxide films

Materials Science in Semiconductor Processing 39 (2015) 467–475

Contents lists available at ScienceDirect

Materials Science in Semiconductor Processing journal homepage: www.elsevier.com/locate/mssp

Using the acetylacetonates of zinc and aluminium for the Metalorganic Chemical Vapour Deposition of aluminium doped zinc oxide films Abdelkader Nebatti a,n, Christian Pflitsch a, Benjamin Curdts a, Burak Atakan a,b a b

Thermodynamics, IVG, Mechanical Engineering, University of Duisburg Essen, Campus Duisburg, Lotharstr.1, D-47057, Germany CeNIDE, Center for Nanointegration Duisburg-Essen, Germany

a r t i c l e i n f o

Keywords: Zinc oxide Metalorganic Chemcical Vapour Deposition Zinc acetylacetonate Aluminium acetylacetonate Transparent conducting oxide TCO

abstract Metalorganic Chemical Vapour Deposition is a promising method for the growth of thin aluminium doped zinc oxide films (ZnO:Al), a material with potential application as transparent conducting oxide (TCO), e.g. for the use as front electrode in solar cells. For the low-cost deposition, the choice of the precursors is extremely important. Here we present the deposition of quite homogeneous films from the acetylacetonates of zinc and aluminium that are rather cheap, commercially available and easy to handle. A usermade CVD-reactor activating the deposition process by the light of halogen lamps was used for film deposition. Well-ordered films with an aluminium content between 0 and 8% were grown on borosilicate glass and Si(100). On both types of substrate, the films are crystalline and show a preferred orientation along the (002)-direction. The 0.3 to 0.5 μm thick films are highly transparent in the visible region. The best films show a low electric resistivity between 2.4 and 8 mΩ cm. & 2015 Published by Elsevier Ltd.

1. Introduction Zinc oxide (ZnO) is a transparent low-cost and nontoxic material with various interesting properties e.g. a wide direct band gap (3.37 eV) and a large binding energy (60 meV [1,2]. Due to the similarity of its optical and electronic properties to those of gallium nitride it is discussed as a possible alternative to GaN-based applications, including the use in blue or UV-LED devices. In contrast to the preparation of GaN, the precursors for the preparation of ZnO are easy to handle and less toxic. Doping the ZnO with group III elements such as gallium n Corresponding author at: Thermodynamics, IVG, Mechanical Engineering, University of Duisburg Essen, Campus Duisburg, Lotharstr.1, D-47057, Germany. E-mail address: [email protected] (A. Nebatti).

http://dx.doi.org/10.1016/j.mssp.2015.05.053 1369-8001/& 2015 Published by Elsevier Ltd.

[3], boron [4], and aluminium [5] results in an improved electrical conductivity without degrading the optical transmission in the visible spectral range. Therefore, Mdoped ZnO (M¼ Ga, B, Al) is discussed as a transparent conducting oxide (TCO) with potential applications as e.g. front electrodes in solar cells devices [6,7], or electrical gas sensors [8]. Furthermore, ZnO doping is achieved by replacing Zn2 þ atom with atoms of elements of higher valence such as Pb4 þ , Sn4 þ , In3 þ , and Al3 þ . It is worthy to mention that among the entire group III, Al is a cheap, abundant, non-toxic material and can be an ideal candidate for dope ZnO [9]. Due to its potentially cheaper production, ZnO:M (M ¼Al, Ga, B) are possible alternative materials for indium-tin-oxide (ITO) which is usually used in such devices. Keeping the potential market of these applications in mind, it is obvious that the development of low-cost and

468

A. Nebatti et al. / Materials Science in Semiconductor Processing 39 (2015) 467–475

powerful deposition techniques is important. Until now, several strategies have been followed to deposit thin zinc oxide films, including wet chemical methods like sol–gel deposition [10] and spray pyrolysis [11]. Also physical vapour deposition, such as rf-magnetron sputtering [12] and molecular beam epitaxy (MBE), or chemical vapour deposition (CVD) have been used. A brief review of the latter technique is given in Ref. [13]. Among these techniques, metal-organic chemical vapour deposition (MOCVD) is a powerful method for the preparation of high quality films with high deposition rates. It was used in the present study. It is obvious that for the development of a low-cost deposition process, the choice of the precursor is most important. For the MOCVD growth of zinc oxide, different precursors have been reported in the literature: often diethyl zinc (DEZ) [14,15]and dimethyl zinc (DMZ) [16] are used, resulting in the growth of highly pure ZnO films. Less attention has been paid to the growth of ZnO-films from zinc acetate [17] and zinc acetylacetonate (Zn(acac)2) [18–20], because it is reported that these precursors tend to produce carbon contaminations. It is suspected that these contaminations influence the electric and optical properties of the grown films negatively. However, zinc acetate and Zn(acac)2 have the advantages of being commercially available and thus being rather cheap. Further, they are chemically stable in atmospheric air, non-toxic and easy to handle. This favours their use in an industrial coating process. Because of these reasons, we have studied the growth of zinc oxide films from zinc acetylacetonate in more detail. Aluminium acetylacetonate (Al(acac)3) was used as an aluminium source for the preparation of Al-doped ZnO films. Undoped ZnO and Al-doped ZnO films were deposited on silicon and borosilicate glass substrates by a photo assisted rapid thermal metal organic chemical vapour deposition (RTMOCVD) technique, where the deposition process is activated by the radiation of halogen lamps. On the one hand, our investigations were focused on the control of the deposition parameters, such as pressure, deposition temperature, and dopant (Al) concentration. On the other hand, electrical and optical properties, which are most important for the above mentioned applications, were extensively studied. Finally, it should be mentioned that to the best of our knowledge, no previous work was done on the deposition of doped zinc oxide from acetylacetonates of zinc and aluminium using a similar deposition process. 2. Experimental 2.1. Film deposition The experimental set-up used for film deposition is shown in Fig. 1. For further details see Refs. [21,22]. It is composed of a stainless steel cylindrical chamber of 200 mm height with two nozzles. The deposition nozzle has a 4 mm diameter hole and is positioned at an angle of 301 from the horizontal surface of the substrate holder. It's function is to introduce the reactants. The second nozzle is used to protect the window. A flow of inert gas (N2) is introduced through this nozzle in order to avoid the deposition on the window during the deposition process.

Fig. 1. Sketch of the used CVD-reactor: (1) Lamps, (2) substrate, (3) Nozzle, (4) CVD chamber, (5) Themocouple, (6) Exhaust, (7) Substrate holder, (8) Window protection nozzle, (9) N2 flow entrance for cooling the lamps.

The temperature of the substrate was monitored and controlled by using a K-type thermocouple inserted in the substrate holder and attached to the backside of the substrate. The substrate was heated with 10 halogen lamps and maintained at the desired temperature using a controller. For all experiments, the total pressure within the chamber was fixed at 200 mbar. During deposition of the un-doped zinc oxide films, zinc acetylacetonate (Zn(acac)2) was evaporated in a fluidized bed evaporator at a temperature of 110 1C (not shown in Fig. 1). The vapour was transported to the chamber with a flow ΦZn;N2 of 500 standard cubic centimetre (sccm) nitrogen as carrier gas, entering through a cylindrical nozzle of 4 mm diameter. Before reaching the orifice, 100 sccm oxygen was added for supporting the oxidation. Additional 200 sccm N2 (ΦN2 ) was added in order to increase the gas velocity. Finally, the gas mixture at the orifice consists of nitrogen, Zn(acac)2, and oxygen in excess. This gas mixture was then blown onto the substrate under an angle Θ¼301 relative to the surface, as shown in Fig. 1. The distance between the orifice and the substrate was approximately 70 mm. For the deposition of the Al-doped ZnO films, Al(acac)3 was evaporated in a separate fluidized bed evaporator and transported to the chamber with a carrier gas flow of nitrogen ΦAl;N2 which was varying for different Alconcentrations between 0 and 80 sccm. Before reaching the orifice, it was added to the above mentioned Zn (acac)2-vapour/N2/O2-mixture while the additional nitrogen flow ΦN2 was decreased by the amount of ΦAl;N2 . Thus, the total flow of the gas mixture was constant for all experiments. The substrates were either pieces of clean silicon wafers (Si (100)-orientation) type (n) or borosilicate glasses, both sized (1 cm  1 cm). Prior to deposition, these substrates were ultrasonically cleaned in ethanol for 20 min, rinsed in deionized water, and dried thereafter.

A. Nebatti et al. / Materials Science in Semiconductor Processing 39 (2015) 467–475

2.2. Film characterisation In our study, various techniques were employed to characterise the grown films: the structures were analysed by X-ray diffraction (XRD) using a Bruker-model D8 Advance X-ray diffractometer working with Cu radiation (λ ¼1.5406 Å) and grazing incidence optics. The angle of incidence was 21 in order to be sensitive to the near surface region. The surface morphology was analysed with a high-resolution scanning electron microscope (FEI ESEM Quanta 400) working with an acceleration voltage of 20 kV. The SEM was equipped with energy-dispersive Xray (EDX) analysis for the detection of the chemical composition of the films. The optical transmittance was studied with a (UV-2102 PC) UV–vis spectrometer. The electrical resistance was measured by four probe geometry. A commercial TGA/DTA (Bähr STA503) was used to perform thermogravimetric measurements on the zinc acetylacetonate precursor. 3. Results and discussion 3.1. Precursor analysis Both acetylacetonate precursors used to deposit Zn and Al are solid powders at room- temperature. In order to study their sublimation, DTA/TG measurements were performed under conditions similar to those in the fluidized bed evaporator during the deposition experiments. The evaporation of the Al(acac)3 is discussed elsewhere [23], and the DTA/TG measurements of the Zn(acac)2 are shown in Fig. 2. Zn(acac)2 was continuously heated in these experiments in a nitrogen gas atmosphere while the TG/DTA signals were recorded (Fig. 2). The heating rate was 1 1C/min.

2.5 0%

2.0 1.5

-20%

1.0 -40% 0.5

DTA/ mV

Relative mass loss

They were mounted on a thermally well isolated holder. Through an optical window in the top of the chamber, they were heated by the thermal radiation of an array of 10 tungsten halogen lamps, activating the chemical reaction of the precursors as shown in Fig. 1. These lamps with a variable electrical power up to a maximum of 5 kW are controlled electronically until a steady temperature of the substrate (in the following called the “deposition temperature”) is reached. Therefore, the temperature is measured with a thermocouple which is mounted at the back side of the substrate. All depositions were carried out at a total pressure of 200 mbar and 550 1C deposition temperature. It is obvious that the film deposition using one nozzle and a stagnation point flow geometry has a radial shape around the flow direction of the gas flow. Thus, the deposition area had a concentric shape of around 5 cm length and 2 cm with. However, for the electrical and optical measurements, it was necessary to create films rather homogeneous in thickness and composition. Therefore, film growth was only studied within an area of 1 cm  1 cm, nearby the centre of the stagnation point. The rest of the deposition area (5 cm  2 cm) was covered by a mask. Within the uncovered area the film deposition is nearly homogeneous concerning the thickness and the chemical composition as being proved experimentally.

469

0.0

-60%

-0.5 -80% -1.0 -1.5

-100% 25

50

75

100

125

150

175

200

225

Temperature (°C) Fig. 2. TGA/DTA analysis of the zinc acetylacetonate precursor: Zn(acac)2 was heated in nitrogen atmosphere of 1000 mbar at a constant heating rate of 1 1C/min.

The TG experiment for Zn(acac)2 (left Ordinate in Fig. 2) shows a two-step evaporation process with two rapid mass losses at around 63–75 1C and at 85–140 1C. Simultaneously, two endothermic peaks appear in the DTAsignal (right ordinate in Fig. 2), one between 63 1C and 75 1C and another one rather sharp peak at around 124 1C. The mass loss between 63 1C and 75 1C is presumably due to the evaporation of chemically combined water which typically evaporates at such low temperatures. At about 85 1C, observable evaporation of Zn(acac)2 starts, indicating the lowest possible evaporation temperature of the precursor. The relative mass loss rises with increasing temperature as the vapour pressure increases according to the Clausius–Clapayron equation. Therefore, the evaporation temperatures of the precursor should be more than 85 1C in the deposition experiments. It is also expected that the deposition rate increases with increasing evaporation temperature. However, the sharp endothermic peak which is observed in the DTA-signal at 124 1C (Fig. 2) indicates the melting of the substance. Therefore, the temperature should not exceed 124 1C if a fluidized bed evaporator is used as in the present study, because this type of evaporator needs powder as precursor source. As a consequence, we used 110 1C for the evaporation of the Zn(C5H7O2)2 in order to get high deposition rates and to beware of the problem of melting the precursor. The Al (acac)3 was evaporated at the same temperature, because 110 1C is also a useful evaporation temperature for aluminium acetylacetonate in CVD-experiments [23,24]. Using the same temperatures for both precursors has the advantage to need only one thermobath for controlling both precursor temperatures. This reduces the costs of the experiments. 3.2. Film structure and surface morphology The crystallisation behaviour of the grown films was reported by us in detail in an earlier publication: film growth starts at deposition temperatures around 400 1C and increases with increasing temperature. Highest deposition rates are observed at the highest possible temperature, which is around 550 1C for the glass substrate used. Higher deposition temperatures would destroy the glass substrates. However, no

Intensity (arb.units)

Intensity (arb.units)

470

A. Nebatti et al. / Materials Science in Semiconductor Processing 39 (2015) 467–475

(002)

(103) Glass

Silicon

(002)

(004)

Glass

Si (100)

Intensity (arb.units)

Silicon

30

(002)

borosilicate glass. The presented measurements which were performed under a fixed angle of incidence (Θ¼21) are sensitive to the near surface region but not useful for the detailed study of the preferred orientation. Thus, we also performed X-ray diffraction measurement under Θ2Θ-geometry (Fig. 3b); they show the “tendency” of the grown film to the (002)-orientation which is in good agreement with the results in the literature [26–28]. The SEM-image of the un-doped film (Fig. 4) supports this observation as most of the observed ZnO-grains show a hexagonal surface in the top view of the sample (Fig. 4a). The preferred orientation is seen even better in the cross section of the sample (Fig. 4b and c). The undoped ZnO thin films are rather smooth on a micrometer-scale for glass substrates (see SEM image: Fig. 4(d)) In contrast to the un-doped ZnO-films, the XRDpatterns of Al-doped ZnO (Fig. 3c) show all reflections which are expected for the hexagonal ZnO. Obviously, those films have a polycrystalline structure. However, the (002) and (103) reflections still dominate the patterns. Thus, the preferred (002) orientation of the crystallites is still present, but less distinctive. 3.3. Chemical composition

(100) (101)

(103) (102)

(112)

Glass

Silicon

35

40

45

50

55

60

65

70

75

2 Theta ( degree) Fig. 3. XRD-patterns of (a, b) undoped ZnO, and (c) aluminium doped ZnO films, which were deposited at 550 1C on Si(100) and borosilicate glass. The measurements in (a) and (c) were performed under a fixed angle of incidence, Θ¼21, while the detector was moved (detector-scans). (b) shows the Bragg scans of the un-doped samples. The reflexes of ZnO are indicated.

mask was used in our earlier studies and therefore these films were quite inhomogeneous concerning the thickness, in fact too inhomogeneous for an electrical and optical analysis. In the following, we report on the analysis of quite homogeneous films, which were grown during 90 min deposition at 550 1C within a mask of 1 cm  1 cm: The grown films are highly transparent. Their thickness was estimated from the density [25]of ZnO (5.6 g/cm3) and the change in weight during deposition to be around 0.5 mm. Fig. 3a shows the XRD-patterns of the un-doped ZnO-films which were deposited on the glass and the silicon substrates. Those of Al-doped ZnO films (Al-concentration: 1.7%) are shown in Fig. 3b. Fig. 4 shows four SEM-image of an un-doped ZnO-film deposited on Si(100) and borosilicate glass As presented in Fig. 3a, two reflexes are observed for the un-doped ZnO-films on both samples, a strong reflex at 2Θ¼34.461 for plane (002) and a less strong one at 2Θ¼62.901 for plane (103). Both indicate the presence of the hexagonal structure of ZnO. Several of the ZnO-reflections are missing, indicating a preferred orientation of the films on the substrates, both silicon and

The chemical compositions of the films were studied by energy dispersive x-ray fluorescence, EDX (Fig. 5a and b). The presented measurements were performed on 3 μm thick films which were grown after 6 h deposition, because only such thick films can reliably be analysed by quantitative EDX. Two EDX-spectra typical for the grown films are given in Fig. 5a: one of an un-doped ZnO on Si(100) and another one of an Al-doped film. They clearly show the presence of zinc, oxygen, and (the upper spectrum only) aluminium. Also some amount of carbon was detected for both samples. As Si (from the substrate) is not present in the spectra, the sample acts like a semi-infinite half-space, which enables the quantitative analysis of the chemical compositions of the films. For this analysis, the EDX-setup was calibrated with a commercially available ZnO-powder (Sigma Aldrige, purity499%): it turned out that EDX always reveals carbon contaminations around 5–10%, even for the pure ZnO-powder. Obviously, the carbon content cannot be analysed quantitatively by EDX and was therefore neglected in the following. For the un-doped ZnO-film (lower spectrum in Fig. 5a), the software of the calibrated EDX- setup reveals the presence of  47% Zn,  53% O, as expected for ZnO within the precision of the measurement. For the Al-doped film (upper spectrum in Fig. 5a), the EDX-measurements reveal the presence of 46% Zn,  52% O, and 2% Al, clearly confirming the doping of the sample. The Al/Zn concentration within the films was analysed as a function of the gas flow ΦAl;N2 of nitrogen carrier gas through the aluminium acetylacetonate evaporator, while the nitrogen carrier gas flow through the zinc acetylacetonate evaporator (ΦZn;N2 was kept constant at 500 sccm. Fig. 5b shows the molar ratio n[Al]/n[Zn] of the Al and Zn-atoms within the films (calculated from the EDX spectra) as a function of the ratio of the gas flows used (ΦAl;N2 /ΦZn;N2 ). Each data point is

A. Nebatti et al. / Materials Science in Semiconductor Processing 39 (2015) 467–475

471

Fig. 4. SEM images: (a) top view, (b, c) cross section of a 0.5 mm thick ZnO film grown on Si(100).

10

Zn

O Al

C

ZnO:Al

Zn

Si 0

1

2 8 10 Energy (keV)

Al/ Zn- Concentation in % (measured by EDX)

Intensity (arb.u)

Zn

9 8 7 6 5 4 3 2 1

ZnO

12

0 0.00

0.04

0.08

0.12

0.16

φ Αl /φ Zn

Fig. 5. EDX-analysis of the grown films: (a) shows the typical EDX spectra of one un- doped and one Al-doped ZnO-film which was prepared by using a flows of ΦAl;N2 ¼70 sccm and ΦZn;N2 ¼500 sccm N2 carrier gas through the acetylacetonates evaporators. (b) shows the Al/Zn-concentration within the films as function of the flow rates ΦAl;N2 and ΦZn;N2 of the carrier gases through the evaporators. The Al/Zn-ratio (ordinate) was calculated from the peak intensities of the EDX-patterns (not shown). Each data point is the average of three EDX measurements. The error is the standard deviation.

the average of three EDX measurements and the error bars are the standard deviations. It is easily seen from Fig. 5b, that the aluminium concentration within the films increases linearly with the carrier gas flow ΦAl;N2 through the aluminium acetylacetonate evaporator. However, this result is not surprising as the amount of precursor vapour leaving the evaporator is proportional to the

used carrier gas flow as long as the gas mixture within the evaporator is saturated. As the evaporation temperature of the aluminium acetylacetonate is high, 110 1C, and the carrier gas flows ΦAl;N2 are rather low (o80 sccm) it is expected that the mixture is saturated. The ratio n[Al]/n[Zn] as a function of the ratio of the used carrier gas flows (ΦAl;N2 /ΦZn;N2 ) was fitted by a line

472

A. Nebatti et al. / Materials Science in Semiconductor Processing 39 (2015) 467–475

Table 1 Electrical and optical properties of the grown films. Al-Flux (sscm) n(Al)⧸n(Zn) (%) ρ (mΩ cm)b Eg (eV)c

a

0

30

60

90

0 8.2 7 0.4 3.34

2.6 4.2 7 0.4 3.42

5.1 2.487 0.40 3.50

8.6 7.170.7 3.57

a

Calculated from the carrier gas flows and Eq. (1). Measured by four point method. Calculated from the band edges (Fig. 9) by following the procedure of Gümüs et al. [47]. b c

Resistivity (mOhm.cm)

9 8 7 6 5 4 3 2 0 2 4 6 8 10 n(Al)/n(Zn)-concentration in %(measured by EDX) Fig. 6. Electrical resistivity of the 0.5 μm thick films on the borosilicate glass as a function of the dopant (aluminium) content. The data are the averages of each 5 measurement on two different samples for each Alconcentration. The errors are the standard deviation.

(dashed line in Fig. 5b): n½Al n½Zn

0:4280

ΦAl;N2 ΦZn;N2

ð1Þ

In the following, this relation is used for determine the dopant content within the films. 3.4. Electrical properties The electric resistivity of several ZnO:Al films of about 0.3 to 0.5 mm thickness were measured as a function of the dopant concentration by the four point method. Figure shows the results for four different aluminium concentrations n[Al]⧸n[Zn] between 0 and 8.6%, which were deposited for 90 min at 550 1C on borosilicate glasses by varying the carrier gas flows. The resistivity was calculated from the measured arial resistance where the thicknesses was obtained by the weighing the sample with a high precision balance before and after deposition. The results are listed in Table 1. The presented data are the averages of five measurements each on two different samples, which were prepared for each dopant concentration. All measurements were performed by using a current of 100 mA. The error bars represent the standard deviation. For all Al-concentrations, one observes a relatively

low resistivity between 2.4 and 8 mΩ cm (Fig. 6), confirming the usefulness of the developed films as transparent conducting oxides (TCOs). It is remarkable that as the Al concentration in ZnO films increased, the electrical resistivity decreases and reaches value of 2.48 mΩ cm for a 5.1% Al concentration. Qiao et al. [29] and Aktaruzzaman et al. [30] have reported similar resistivity value for Al-doped ZnO containing 1–2%. A brief review of the electric properties of such films is given in reference [31]. Their, the lowest values of the resistivity for such films are reported to be in the range of 1400 to 0.2 mΩ cm. In order to compare our results with other literatures results, Miamia et al. [32], they deposited un-doped and Al doped ZnO by AP-MOCVD. The lowest resistivity of 4.6 mΩ cm for undoped film, however, for doped film the resistivity augmented to 7 mΩ cm. Therefore, the effect of doping on electrical properties was unsatisfactory. This result is not similar to our films, where the Al doping modifies the resistivity, which is the aim of doping. With reference to Lee et al. work [33], they developed undoped and Aldoped ZnO by MOCVD using Ultrasonic atomisation. Undoped ZnO the resistivity around  3.5 Ω cm, the minimum resistivity of Al-doped ZnO was 0.6 mΩ cm at Al doping concentration of 6 at%. A resistivity of 2.48 mΩ cm is observed in our film for 5% Al concentration. This value is comparable to those obtained by Lee et al. for similar Al concentrations. The electrical properties of ZnO are primarily dominated by electrons generated from O2  vacancy and Zn interstitial atoms [34]. The decrease in resistivity with Al doping is due to the contribution from Al3 þ ions on substitutional sites of Zn2 þ ions and Al interstitial atoms, as well as from oxygen vacancies and Zn interstitial atoms [35]. But such Al atoms in the films do not always contribute as dopant. When the Al content in ZnO films increases above a certain critical concentration, it segregates into the grain boundaries. These segregates Al atoms do not activate as dopant and hence do not contribute in increasing the electron concentration [36]. They rather create defects which, in turn, enhance the resistivity. This is the reason for the observed increase in the resistivity of Al-doped ZnO films beyond an Al concentration of 5.2%. Minami et al. reported that when ZnO film is doped with Al by sputtering method using ZnO þAl2O3 target, electron concentration increases with increasing amount of Al2O3, and decreases when the amount of Al2O3 in the target is over 4%. Park et al.[35] and Oh [5] et al. and Tang et al. [37] observed a similar behaviour of the resistivity versus Al concentration in ZnO film (as is observed in the present case (see Fig. 6). 3.5. Optical properties Optical and further electrical properties such as band gap were calculated from the optical transmittance spectra of the films on borosilicate glass substrates. For all samples, the optical transmittance spectra were measured at room temperature between 300 nm and 800 nm with a commercial UV–vis-spectrometer (see Fig. 7). It is seen that the transmittance of the films improves with an increase of the aluminium concentration. The oscillations

A. Nebatti et al. / Materials Science in Semiconductor Processing 39 (2015) 467–475

in the spectra are caused by optical interferences on the smooth surface. A sharp absorption edge is also observed

in the spectra, this confirms again the crystal quality because in amorphous films no sharp absorption edges are observed [38,39]. In Fig. 7, it is seen that the coated samples still have a high transmittance in the visible range, which is more than 80%. As noted above the transmittance spectra within the UV–vis region ( o550 nm, Fig. 7 can be used to calculate the band gap of the grown films. It is seen that the band edge of the un-doped ZnO film is around 380 nm. With increasing aluminium content the band gap increases due to the doping which is known as Burstein–Moss shift resulting from an increase of the carrier concentration. Thus, the band edge shifts to lower wavelength [ 355 nm for n(Al)/n(Zn) ¼0.086], which is in good agreement with the literature. To further study the optical properties of the films, the optical band gap were calculated by using the Tauc model [40], and the Mott model [41]. The optical absorption coefficient α can be expressed as [42]

100

Transmittance(%)

80 60 un-doped n(Al)/n(Zn)=0.026 n(Al)/n(Zn)=0.051 n(Al)/n(Zn)=0.086

40 20 0 300

350

400

450

500

550

600

650

700

750

473

800

wavelength(nm) Fig. 7. Optical transmittance spectra (a) of  0.5 μm thick ZnO:Al films on borosilicate glass for different dopant concentration. A full range optical transmittance sprectrum of Al-doped (5.1%) ZnO film is shown.

I ¼ I 0 expð αdÞ

ð2Þ

12 10 8 6 4

0 2.0

2.5

3.0

3.5

4.0 2.0

2.5

3.0

3.5

4.0

2.5

3.0

3.5

4.0

16

2

α (× 10

10

-2 Cm )

2

14 12 10 8 6 4 2 0 2.0

2.5

3.0

3.5

Photon energy (eV)

4.0

2.0

Photon energy (eV)

Fig. 8. Plot of α2(absorption) versus photon energy hν for (a) undoped, (b) n[Al]⧸n[Zn] ¼0.026,(c) n[Al]⧸n[Zn] ¼0.051,(d) n[Al]⧸n[Zn] ¼0.086.

Band Gap (eV)

474

A. Nebatti et al. / Materials Science in Semiconductor Processing 39 (2015) 467–475

3.56

by light radiation from halogen lamps. The main results of the present process can be summarised as follows:

3.52

 With an accurate selection of deposition conditions it

3.48 3.44 3.40

 

3.36 3.32 0

2

4

6

8

n(Al)/n(Zn)-Concentration in %



Fig. 9. Band gap as function of different dopant concentrations.

where I is the intensity of transmitted light, I0 is the intensity of incident light, and d is the film thickness. As the transmittance (T) is defined as I⧸I0, the above equation (2) can be used to determine the value of α. In direct transition semiconductor, α and Eg are related by [43],
1=2 ð3Þ α ¼ h ν  Eg where h is Planck's constant, ν is the frequency of the incident photon energy and Eg is the optical band gap. The Fig. 8 is the plot of α2 versus hν for doped ZnO films with different Al concentration. The linear dependence of α2 to hν indicates that Al-doped ZnO films are direct transition type semiconductor. The photon energy at the point where α2 ¼0 is Eg and is obtained by the extrapolation method. The optical band gap values obtained are summarised in Table 1. It was reported that the optical band gap of zinc oxide thin films developed using MOCVD lies between 3.25 and 4.06 eV depending on the various parameters, such as deposition temperature, doping, etc. [37,42–44]. Tang et al. [45] and Fragal et al. [46] reported the relationship between optical band gap and Al doping concentration of the ZnO thin films. They observed that in un-doped ZnO films, the optical band gap appeared at 3.25 eV while it increased to 3.47 eV with Al doping. A similar behaviour was observed in our films, where the Eg value increase from 3.34 eV (for un-doped film) to 3.57 eV (for high Aldoping concentrations) (see Table 1). Fig. 9 shows the dependence of band gap of Al doped ZnO films of similar thickness 432 nm as a function of Al concentration. It can be seen that the band gap with increases linearly with Al content. They were fitted using a linear function, which is   Eg ðeVÞ ¼ 3:34 þ2:6 n Al =n½Zn

ð4Þ

4. Summary and conclusion The objective of this work was to deposit and characterise un-doped and Al-doped ZnO thin films were deposited by home built CVD-reactor, which is a new user-made CVD-reactor activating the deposition process





was possible to deposit high quality ZnO films on both substrates, silicon-(100)-oriented and borosilicate glass, where the x-ray diffraction (XRD) show only (002) and (004) peaks indicating a strong preferred orientation. Furthermore, the SEM analysis substantiate the observation from the XRD. Deposition does not take place below 400 C. Homogeneous undoped and Al-doped ZnO films of 0.3 to 0.5 mm thickness were deposited for the first time from acetylacetonate of zinc and aluminium using a similar deposition process. The films are highly transparent in the visible region. The Al-doped ZnO films have a relatively low resistance, around 2.48 mΩ cm. Thus, these films are quite promising for the application as transparent conducting oxides (TCO). The electrical properties like band gap, which is dependent on the aluminium content, can be easily controlled by varying the gas flows through the evaporators. Precise aluminium doping was possible as proved by EDX. Both acetylacetonates are promising candidates as precursors for the deposition of un-doped and Al-doped ZnO films. They are commercially available, quite cheap, non-toxic and easy to handle, supporting the development of a low-cost industrial deposition process.

Acknowlegements Financial support by the University of Duisburg-Essen and the Land Nordrhein-Westfalen is gratefully acknowledged. A. Nebatti thanks the DAAD (Deutscher Akedemischer Austauschdienst) for his fellowship. References [1] C. Jagadish, S.J. Pearton, Zinc Oxide Bulk, Thin Films and Nanostructures: Processing, Properties, and Applications, Elsevier, 2011. [2] Ü. Özgür, Y.I. Alivov, C. Liu, A. Teke, M. Reshchikov, S. Doğan, V. Avrutin, S.-J. Cho, H. Morkoc, J. Appl. Phys. 98 (2005) 041301. [3] Y.-C. Huang, Z.-Y. Li, H.-h. Chen, W.-Y. Uen, S.-M. Lan, S.-M. Liao, Y.-H. Huang, C.-T. Ku, M.-C. Chen, T.-N. Yang, Thin Solid Films 517 (2009) 5537–5542. [4] X. Chen, B. Xu, J. Xue, Y. Zhao, C. Wei, J. Sun, Y. Wang, X. Zhang, X. Geng, Thin Solid Films 515 (2007) 3753–3759. [5] B.-Y. Oh, M.-C. Jeong, W. Lee, J.-M. Myoung, J. Cryst. Growth 274 (2005) 453–457. [6] R.W. Birkmire, E. Eser, Annu. Rev. Mater. Sci. 27 (1997) 625–653. [7] W. Jeong, S. Kim, G. Park, Thin Solid Films 506 (2006) 180–183. [8] S. Roy, S. Basu, Bull. Mater. Sci. 25 (2002) 513–515. [9] A. Kelchtermans, K. Elen, K. Schellens, B. Conings, H. Damm, H.-G. Boyen, J. D’Haen, P. Adriaensens, A. Hardy, M.K. Van Bael, RSC Adv. 3 (2013) 15254–15262. [10] G.T. Delgado, C.Z. Romero, S.M. Hernández, R.C. Pérez, O.Z. Angel, Sol. Energy Mater. Sol. Cells 93 (2009) 55–59. [11] J.-H. Lee, S.-Y. Lee, B.-O. Park, Mater. Sci. Eng: B 127 (2006) 267–271. [12] K.-H. Lee, N.-I. Cho, E.-J. Yun, H. Nam, Appl. Surf. Sci. 256 (2010) 4241–4245. [13] T. Gerfin, K.H. Dahmen, CVD of Nonmetals, pp. 151–191.

A. Nebatti et al. / Materials Science in Semiconductor Processing 39 (2015) 467–475

[14] Y. Wang, H. Wang, S. Li, S. Zhou, Y. Hang, J. Xu, J. Ye, S. Gu, R. Zhang, J. Cryst. Growth 284 (2005) 319–323. [15] D.C. Kim, B.H. Kong, S.O. Jun, H.K. Cho, D.J. Park, J.Y. Lee, Thin Solid Films 516 (2008) 5562–5566. [16] A. Dadgar, N. Oleynik, D. Forster, S. Deiter, H. Witek, J. Bläsing, F. Bertram, A. Krtschil, A. Diez, J. Christen, J. Cryst. Growth 267 (2004) 140–144. [17] J. Kim, H. Marzouk, P. Reucroft, C. Hamrin, Thin Solid Films 217 (1992) 133–137. [18] Y. Kashiwaba, K. Sugawara, K. Haga, H. Watanabe, B. Zhang, Y. Segawa, Thin Solid Films 411 (2002) 87–90. [19] M. Ogawa, Y. Natsume, T. Hirayama, H. Sakata, J. Mater. Sci. Lett. 9 (1990) 1351–1353. [20] K. Kamata, J. Nishino, S. Ohshio, K. Maruyama, M. Ohtuki, J. Am. Ceram. Soc. 77 (1994) 505–508. [21] A.E. Nebatti, C. Pflitsch, C. Eckert, B. Atakan, ECS Trans. 25 (2009) 459–465. [22] A.E.Muhsin, urn:nbn:de:hbz:464-20070713-111156-3, in: Thermodynamic Department, Duisburg-Essen, Duisburg, 2007. [23] M.A. Siddiqi, R.A. Siddiqui, B. Atakan, Surf. Coat. Technol. 201 (2007) 9055–9059. [24] C. Pflitsch, D. Viefhaus, U. Bergmann, B. Atakan, Thin Solid Films 515 (2007) 3653–3660. [25] M. Ohring, Materials Science of Thin Films, Academic Press, 2001. [26] X. Yan, Z. Li, R. Chen, W. Gao, Cryst. Growth Des. 8 (2008) 2406–2410 . [27] R. Khandelwal, A.P. Singh, A. Kapoor, S. Grigorescu, P. Miglietta, N.E. Stankova, A. Perrone, Opt. Laser Technol. 40 (2008) 247–251. [28] P. Prepelita, R. Medianu, B. Sbarcea, F. Garoi, M. Filipescu, Appl. Surf. Sci. 256 (2010) 1807–1811. [29] Z. Qiao, C. Agashe, D. Mergel, Thin Solid Films 496 (2006) 520–525.

475

[30] A. Aktaruzzaman, G. Sharma, L. Malhotra, Thin Solid Films 198 (1991) 67–74. [31] K. Ellmer, J. Phys. D: Appl. Phys. 34 (2001) 3097. [32] T. Minami, H. Sonohara, S. Takata, H. Sato, Jpn. J. Appl. Phys. 33 (1994) L743. [33] C. Lee, D. Kim, J. Electroceram. 33 (2014) 12–16. [34] Y. Igasaki, H. Saito, J. Appl. Phys. 70 (1991) 3613–3619. [35] K.C. Park, D.Y. Ma, K.H. Kim, Thin Solid Films 305 (1997) 201–209. [36] Y. Liu, J. Lian, Appl. Surf. Sci. 253 (2007) 3727–3730. [37] W. Tang, D. Cameron, Thin Solid Films 238 (1994) 83–87. [38] Y. Wang, S. Lau, H. Lee, S. Yu, B. Tay, X. Zhang, K. Tse, H. Hng, J. Appl. Phys. 94 (2003) 1597–1604. [39] R. Al Asmar, J. Atanas, M. Ajaka, Y. Zaatar, G. Ferblantier, J. Sauvajol, J. Jabbour, S. Juillaget, A. Foucaran, J. Cryst. Growth 279 (2005) 394–402. [40] J. Tauc, in: J. Tauc (Ed.), Amorphous and Liquid Semiconductors, Plenum, London, 1974, p. 159. [41] N. Mott, E. Davis, R. Street, Philos. Mag. 32 (1975) 961–996. [42] S. Tan, B. Chen, X. Sun, X. Hu, X. Zhang, S. Chua, J. Cryst. Growth 281 (2005) 571–576. [43] E. Ziegler, A. Heinrich, H. Oppermann, G. Stöver, Phys. Status Solidi A 66 (1981) 635–648. [44] B.E. Sernelius, K.-F. Berggren, Z.-C. Jin, I. Hamberg, C. Granqvist, Phys. Rev. B 37 (1988) 10244. [45] S.T. Tan, B.J. Chen, X.W. Sun, W.J. Fan, H.S. Kwok, X.H. Zhang, S. J. Chua, J. Phys. Appl. 98 (2005) 013505. [46] M.E. Fragala, G. Malandrino, M.M. Giangregorio, M. Losurdo, G. Bruno, S. Lettieri, L.S. Amato, P. Maddalena, Chem. Vap. Depos. 15 (2009) 327–333. [47] C. Gumus, O. Ozkendir, H. Kavak, Y. Ufuktepe, J. Optoelectron. Adv. Mater. 8 (2006) 299.