1. Introduction

In order to obtain materials with novel properties different approaches have been proposed, among which the preparation of ternary, quaternary or alloys has proven to be useful for the study of novel systems. For example, the oxychalcogenides based on II-VI compounds BiCuOSe [1] and CuCdTeO [2–4] have been fabricated and studied due to their attractive thermoelectrical and charge transport properties, respectively. Chalcogenide II-VI materials are direct band gap semiconductors with high optical absorption coefficients [5]. In particular, ZnSe has attracted great interest over the last years due to its thermal, electrical and optical properties amenable to optoelectronic applications. Investigation of the properties of ZnSe-based materials is an important step in the search for new materials whose physical properties can be designed. ZnSe has a high optical transmission and a band gap energy near 2.7 eV. It has been grown and studied by different authors and techniques [6–12]. In this work we present the structural, electrical and optical characterization of films grown by rf sputtering using ZnSe-Cu₂O composite targets. This is a continuation of our previous work where films grown from CdTe-CuO composite targets were studied as a function of the [CuO]/[CdTe] ratio and substrate temperature [3,4]. The purpose of the present work was twofold: i) provide first information on the behavior of the CuZnSeO system, and ii) explore the Cu-O high concentration regime aimed towards the possible formation of a stoichiometric compound of the form Cu₂ZnSeO. For this case the atomic concentration of Cu₂O in the target should be 60% (i.e. in atomic percentages: 40% Cu, 20% O, 20% Se and 20% Zn) if one takes into consideration that sputtering is a technique that closely reproduces the target composition in the deposited films. Moreover, in order to learn about the behavior of the ZnSe-Cu₂O system, lower than 60 at.% Cu₂O concentrations in the target were also explored. Namely, the [Cu₂O]/[ZnSe] ratios used in this work were 12, 36, and 60 at.%. Hereafter, when the specification of the [Cu₂O]/[ZnSe] ratio is not relevant for the discussion, the films will be designated generically as CuZnSeO. It is shown below that the formation of the stoichiometric compound Cu₂ZnSeO was not achieved and that, instead, the films grown from ZnSe-Cu₂O composite targets yielded amorphous and polycrystalline ZnO-Cu_2-xSe composite films, depending upon the substrate temperature. The structural, optical and electrical properties of these films are discussed in terms of the [Cu₂O]/[ZnSe] ratio and of the substrate temperature.

ZnO is a II-VI direct band gap (~3.3 eV at room temperature) semiconductor material with potential applications in several fields, especially in optoelectronics. ZnO has found applications in the fabrication of green, blue-UV and white light emitting devices. It has a large exciton binding energy (~60 meV) and a relatively simple crystal growth technology that could eventually yield low-cost devices [13]. Other potential applications for ZnO include the realization of transparent transistors and its use as n-type window layer in photovoltaic structures. Indeed, Zang et al. demonstrated photovoltaic activity in ZnO/Cu₂O heterostructures [14] by using Cu₂O deposited by low temperature radical oxidation [15].

2. Experimental

Composite ZnO-Cu_2-xSe thin films were deposited by rf magnetron sputtering onto Corning glass substrates 2947 using ZnSe and Cu₂O powders. The targets were prepared by cold-pressing mixtures of ZnSe-Cu₂O powders to form a 2”-target. For this work the nominal concentrations of Cu₂O in the target were 12, 36 and 60%. The transformation of Cu₂O to CuO that occurs after its long exposure to ambient air was minimized by using the Cu₂O powder right after opening the ampoule (filled with inert gas) as received by the supplier. The targets were placed into the vacuum chamber immediately after their preparation. The substrate temperature (T_s) for different depositions was room temperature (RT, i.e. unintentional heating), 100, 200, 300 and 400 °C. During growth, the rf power was fixed to 50 watts with a deposition time of 120 minutes. The films were analyzed structurally in a DMax-2100 Rigaku X-ray diffractometer (Cu k_α = 1.5406 A) at grazing angle geometry. In order to determine the in-depth homogeneity in the films, measurements were taken at different incidence angles: 0.8, 1.5 and 3.0° of the X-ray beam. Transmittance and Reflectance spectra were obtained using a Cary-5000 UV-Vis-IR spectrometer and morphology measurements were obtained on an Atomic Force Microscope Nanoscope IV Dimension. Thickness measurements were carried out in a Sloan Dektak II profilometer. The resistivity values were obtained by the four-point probe method in a T600 Loresta-GP system and the hot-point probe experiments were carried out in a home-made standard experimental set up.

3. Results

3.1 Structural characterization

The X-ray diffraction patterns of CuZnSeO films for different concentrations of Cu₂O in the target and various substrate temperatures are shown in Fig. 1. In the case of [Cu₂O] = 12 at.%, Fig. 1a, the patterns show that for substrate temperatures lower than 200 °C the films are amorphous. When T_s = 200 °C the films start crystallizing yielding weak broad peaks of cubic Cu_2-xSe (berzelianite, JCPDS No. 06-0680,) label c in Fig. 1. The observed peaks correspond to the growth of the c(111) and c(220) planes. At 300°C and above, phase segregation continued so that the films were a mixture of hexagonal zinc oxide (zincite, JCPDS No. 36-1451, label h in Fig. 1) and Cu_2-xSe. In the case of films with [Cu₂O] = 36 and 60 at.%, Figs. 1(b) and 1(c), the overall characteristics of the diffraction patterns are similar to the 12% case. Some differences, however, are the appearance at 2θ = 53° of the c(533) peak when Ts≥200 °C for [Cu₂O] = 36%, Fig. 1(b), and of the o(111) and o(510) peaks of orthorhombic Cu₂Se (bellidoite, JCPDS No. 37-1187, labeled o in Fig. 1) at 13° and 40° respectively, for films prepared from the target with [Cu₂O] = 60 at.% and T_s = 400°C, Fig. 1(c). It is worth mentioning that the highest concentration of Cu₂O in the target (60 at.%) induced the formation of Cu₂Se instead of the copper-deficient Cu_2-xSe phase (i.e. with Cu vacancy sites randomly occupied) observable for [Cu₂O] = 12 and 36 at.%. This is reasonable since more copper atoms were available for the highest concentration of Cu₂O in the target (60 at.%), thus avoiding the formation of Cu vacancies. In all cases the ZnO peaks were observable from T_s = 300 °C, that is, from this temperature on X-ray diffraction results clearly show that segregation occurred and that ZnO and a form of copper selenide (Cu_2-xSe or Cu₂Se) coexist. In what follows Cu_2-xSe may be used as a general form to describe copper selenide phases, including Cu₂Se (for which x = 0). The context should clarify when this is the case.

 

Fig. 1 X-ray diffraction patterns of the films grown at different substrate temperatures for different concentrations of Cu₂O in the target: a) 12 at.%, b) 36 at.% and c) 60 at.%.

Download Full Size | PDF

From these results it may be stated that sputtering of the composite ZnSe/Cu₂O targets produced dissociation of the ZnSe and Cu₂O molecules. The resulting species formed other compounds (ZnO and Cu_2-xSe) more favorable energetically according to the specific conditions of temperature encountered on the substrate (and at the growth chamber pressure). The single consideration of the magnitudes of the bond strengths (D°) may not quite suffice to explain the resulting compounds. Indeed, the formation of Cu-Se bonds seem favored since for these bonds D° = 251 KJ mol⁻¹ [16] and for Cu-O D° = 269 ± 20.9 KJ mol⁻¹. For the zinc compounds D° is substantially smaller: D° = 159.4 and 170.7 ± 25.9 KJ mol⁻¹ [16] for the Zn-O and Zn-Se bonds, respectively. In other words, it may seem that according to the above values Cu-Se bonds are not strongly favored over Cu-O. It must be taken into account, nonetheless, that the quoted values of D° were obtained for a temperature of 298 K and at 1 atmosphere of pressure [16]. It is plausible to hypothesize, thus, that for the substrate temperatures used in this work and the pressure in the chamber during growth (~5 mTorr) the formation of Cu-Se was indeed significantly favored, and that the remaining Zn and O atoms bonded together as a consequence. The formation of Cu_2-xSe phases at lower temperatures than ZnO supports this hypothesis, Fig. 1. That is, the nucleation processes of the crystallites forming the films at T_s and at 5 mTorr, energetically favors the creation of Cu-Se bonds, and, therefore, the growth of Cu_2-xSe crystallites. Unfortunately, data of D° for temperatures higher than 298 K are not available to confirm the higher value for Cu-Se bonds than for the rest of the involved bonds.

In order to determine the in-depth structure and homogeneity, X-ray diffraction measurements were carried out to probe different depths by varying the angle of incidence (α) of the beam. Three angles were used: 0.8, 1.5 and 3.0° measured with respect to the film surface. It was found that all the diffraction patterns were basically the same with no significant differences in peaks positions or relative intensities. As an example, Fig. 2 shows the patterns for the sample with the largest [Cu₂O] in the target and the highest T_s, that is, for the sample where in-depth inhomogeneities could be more likely to form. The only significant difference in the patterns in Fig. 2 is the observance of a broad band with maximum around 24° for the pattern measured at an angle of incidence α = 3°. This broad feature is produced by the substrate (glass), which indicates that the X-ray beam is probing the entire film thickness for this angle of incidence. Similar results were found for the samples with [Cu₂O] = 12 and 36 at% in the target. These results show that the formation of zinc oxide or copper selenide does not occur preferentially at any particular region of the films, neither close to the film surface, nor near the substrate. In this respect it may be stated that the zinc oxide and copper selenide regions are homogeneously distributed throughout the volume of the films.

 

Fig. 2 X-ray diffraction patterns of the films grown with [Cu₂O] = 60 at.% in the target at 400°C for different angles of incidence (α). “o” stands for orthorhombic and “h” for hexagonal phases.

Download Full Size | PDF

3.2 Optical properties

Figure 3 shows the transmission spectra for different substrate temperatures and Cu₂O contents in the target. In the case of 12% of Cu₂O, Fig. 3(a), at low substrate temperatures the values of transmission are around 80% for RT and 65% for 100°C. As the substrate temperature is increased, the transmission values drop and present a transmission band between 450 and 1050 nm, approximately. This behavior is correlated with the X-ray diffraction data. Indeed, the more transmitting films are those with amorphous characteristics, i.e. those where no segregation into copper selenide/zinc oxide aggregates has occurred. For substrate temperatures of 200°C and above, the transmission spectra are dominated by the optical properties of Cu_2-xSe since ZnO is transparent from ca. 376 nm (E_g = 3.3 eV) [13], although absorption processes from defects in the ZnO crystallites cannot be discarded. This defect-related absorption is particularly evident for the amorphous films. The transmission window observed in the spectra of Fig. 3 are in good agreement with the optical transmission data of sputtered copper selenide films reported recently by H. Ahn and Y. Um [17]. Analogous observations can be made to the spectra corresponding to the films grown with [Cu₂O] = 60 at.% in the target, Fig. 3(c). However, for the films grown with [Cu₂O] = 36 at.%, Fig. 3(b), the transmission values above ~900 nm are significantly reduced in comparison with the those of the other Cu₂O concentrations, presenting values as low as 30%. In other words, the optical transmission of the low-temperature amorphous phases presents different behavior. This could be due to dissimilarities in the local atomic arrangement generated by the differences in the relative concentrations of the four chemical elements, which, in addition is also affected by the substrate temperature during growth. A simple comparison of the electronic response, probed through light absorption, can be assessed by considering the films spectral response to light transmission normalized by the film thickness. This normalized transmission could yield useful information only if the light dispersion at the films surface is similar, that is to say, if the samples surface roughness is of the same order of magnitude, which actually occurs for the amorphous films as shown below (v.g. the RMS roughness of the films grown at room temperature ranged between 1.3 and 5.5 nm). In this manner, the normalized transmittances (τ) become independent of the film thickness and may be thought of as a measure of the electronic response to the incident photons (photon absorption processes) per thickness unit length.

 

Fig. 3 Optical transmission spectra of the films grown at different substrate temperatures with [Cu₂O] in the target of a) 12 at.%, b) 36 at.% and c) 60 at.%.

Download Full Size | PDF

Figure 4 shows a plot of the transmittance (T) of Fig. 3 normalized by the film thickness in micrometers (d) for each case, i.e. the transmittance per thickness unit length (TTUL). It may be observed that the spectra corresponding to the amorphous films group into a typical behavior for each concentration of Cu₂O in the target. This indicates that even though all films are amorphous in nature with no significant differences in their X-ray patterns, the electronic properties may vary importantly depending on the relative concentration of the constituent elements in the films so that three well defined amorphous-like behaviors of τ are observed in Fig. 4, one for each concentration of Cu₂O in the target.

 

Fig. 4 Transmission per thickness unit length or TTUL (τ = T/d, i.e. transmittance per in-depth micron) of the films at different substrate temperatures with [Cu₂O] in the target of a) 12 at.%, b) 36 at.% and c) 60 at.%.

Download Full Size | PDF

The behavior of τ (Fig. 4) for the polycrystalline films grown at T_s≥200 °C is different from that of the amorphous samples. The copper-selenide-like bands in Figs. 4(a) and 4(b) do not group into a characteristic behavior as they closely do in Fig. 4(c). This could be explained with the help of the X-ray diffraction results, which show that the copper selenide peaks of films grown with [Cu₂O] = 12 and 36 at.% in the target correspond to berzelianite Cu_2-xSe, while the peaks of films grown with [Cu₂O] = 60 at.% correspond to a different phase of copper selenide: bellidoite Cu₂Se. In the former, berzelianite is a material that may present distinct densities of copper vacancies that give rise to different polymorphs and electronic properties that depend upon their particular density and distribution [18, 19]. In the case of [Cu₂O] = 60 at.%, since Cu₂Se has a well-defined stoichiometry, strong variations in the electronic properties due to vacancies are not expected. In this sense the electronic properties of the films with stoichiometric Cu₂Se aggregates (i.e. [Cu₂O] = 60 at.%) yield little variation in the TTUL spectra, circled zone in Fig. 4(c), in contrast to those containing copper vacancies distributions, Cu_2-xSe ([Cu₂O] = 12, 36 at.%), where the behavior of τ presents more varied features (circled regions in Figs. 4(a) and 4(b)) which can be reasonably ascribed to the existence of various vacancy densities and distributions.

The films surface morphology presented different characteristics depending upon the substrate temperature and target chemical composition. In Fig. 5 are presented atomic force microscope (AFM) images of the films grown at RT and 400°C for the three Cu₂O concentrations in the target. As expected, the surfaces of the films grown at RT are smoother than those of the polycrystalline films. Table 1 presents the RMS roughness for the lowest and highest substrate temperatures for each concentration of Cu₂O in the target. The lowest roughness (1.35 nm) was obtained for the film grown from the target with [Cu₂O] = 36 at.% at RT, while the largest value (95.32 nm) corresponded to the film grown with [Cu₂O] = 60 at.% at 400°C. From Table 1, it may be observed that, the RMS roughness is of the same order of magnitude for the amorphous films (T_s = RT). This fact was assumed for the analysis above carried out for the TTUL as a means to evaluate differences in the electronic response of the amorphous films. In Fig. 6 are presented representative (for the largest and smallest T_s and [Cu₂O]) scanning electron images of the films grown at room temperature and 400 °C for [Cu₂O] = 12 and 60 at.%. As it may be observed, the grain size and shape depend strongly on both the substrate temperature and Cu₂O concentration.

 

Fig. 5 Atomic force microscopy images of the films surface grown at room temperature and at 400°C for a) 12 at.%, b) 36 at.% and c) 60 at.% of Cu₂O in the target. Note the different vertical scales.

Download Full Size | PDF

Table 1. Root mean square (RMS) of the films surfaces obtained from AFM scans for the films grown at room temperature (RT) and at 400 °C.

View Table

 

Fig. 6 Scanning electron images of films grown at room temperature (RT) and 400°C for a) 12 at.% and b) 60 at.% of Cu₂O in the target. The size bar is 100 nm for all cases.

Download Full Size | PDF

The reflectance measurements of the films follow the same trend of the above results in that basically two types of responses are obtained: one for the amorphous films and other for the polycrystalline in which well-defined zinc oxide and copper selenide phases are present. Since the latter have rougher surfaces, lower reflectance is expected. Figure 7(a) shows a representative case ([Cu₂O] = 60 at.%) where the mean reflectance is about 17% for wavelengths above 500 nm in the case of the smooth-surface amorphous films. The reflectance of the polycrystalline Cu₂Se/ZnO films presents interference fringes in the same spectral window where the film transmits (Fig. 3c), i.e. light can reach the film/substrate interface and interfere with the reflected light from the air/film interface. For wavelengths below and above this spectral window (zone A in Fig. 7.a) the reflectance has a monotonic behavior, increasing to higher values as light gets deeper into the infrared. It is worth mentioning that in Fig. 7(a) the reflectance of such polycrystalline films (T_s = 300, 400°C) is around 13% at 2400 nm and below the average value of the amorphous films (zone B). A different electronic response is obtained once more in the films grown from the target with [Cu₂O] = 36 at.%. In Fig. 7(b) it may be observed that the reflectance in the infrared region increases significantly and gets even higher than that of the smooth-surface amorphous films above 1800 nm. For example, at 2400 nm R≈31% (Fig. 7(b)) for the film grown at 200 °C, nearly three times more reflective than the polycrystalline films in Fig. 7(a). In this case, the observed reflectivity of the films is strongly influenced by two competing factors: surface roughness and free carrier density. The former tends to reduce R, while the latter to increase it. Since the films with 36 at.% in the target are the least resistive films (see below), it may be considered that the higher reflectivity in the infrared of the films grown at T_s = 200 and 300°C with respect to the smoother film grown at 100°C, Fig. 7(b), is due to higher free carrier densities present in these two films. Indeed, in the following section is shown that the resistivity of the films grown at T_s = 200 and 300°C is four orders of magnitude lower than that of the film grown at 100°C. Assuming that this difference cannot be explained in terms of differences in mobility (a fairly reasonable assumption for four orders of magnitude difference in resistivity), its origin should be accounted for by the different values of free carrier density. That is, for these films the free carrier density effect dominated over that of the surface roughness.

 

Fig. 7 Reflectance spectra of the films grown at different substrate temperatures with [Cu₂O] in the target of a) 60 at.% and b) 36 at. %.

Download Full Size | PDF

3.3 Electrical properties

The films resistivity was measured by the four-point-probe method at room temperature. The films grown with [Cu₂O] = 12 at.% at T_s<200 °C had resistivities larger than the system upper limit (∼10⁴ Ω-cm). Figure 8 is a plot of the films resistivity as a function of the substrate temperature for the three Cu₂O concentrations reported here. The polycrystalline films have lower resistivities than the amorphous cases, and the more conductive films were those prepared from the target with [Cu₂O] = 36 at.%. The resistivity range covered by the films was from 10⁻³ (T_s = 300 °C, [Cu₂O] = 36 at.%) to 10³ Ω-cm (T_s = RT, [Cu₂O] = 60 at.%). The error bars correspond to +/− one standard deviation from the corresponding measurement sets. It is clear again in Fig. 8 that the values of the films grown from the target with [Cu₂O] = 36 at.% do not lie between the values of 12 and 60 at.%; another manifestation of the different electronic characteristics of these films. The causes that originate such behavior are beyond the scope of the present work.

 

Fig. 8 Resistivity of the films as a function of the substrate temperature. Inset: hot-point probe data of the film grown at 400°C from a target with [Cu₂O] = 60 at. %.

Download Full Size | PDF

Hot-point probe measurements were carried out to find out the conductivity type of the films. This was of particular interest since the films are a composite of materials that have usually different type of conductivity: ZnO is typically an n-type semiconductor, while Cu_2-xSe is a degenerate p-type semiconductor [20]. The hot-point probe results indicate that the dominant free-carriers in the composite system are electrons. The inset in Fig. 8 shows, as an example, the step-like graph obtained for the film grown at 400 °C and [Cu₂O] = 60 at.%. Similar results were obtained for the other samples. When holes are the free (positive) carriers, the voltage measured between the hot and cold leads falls towards negative values.

4. Conclusions

Composite ZnO-Cu_2-xSe films were grown by rf sputtering using a single target prepared by mixing ZnSe and Cu₂O powders in atomic ratios [Cu₂O]/[ZnSe] of 12, 36 and 60 at.%. Films grown with the substrate temperature lower than 200 °C were amorphous, except those with [Cu₂O] = 36 at.% that were polycrystalline even when Ts = 200°C. According to the X-ray diffraction studies, in all depositions polycrystalline copper selenide was formed at lower substrate temperatures than ZnO. The identified copper selenide phases in the polycrystalline films were berzelianite (Cu_2-xSe) and bellidoite (Cu₂Se). The former was present when [Cu₂O] = 12 and 36 at.%, while the latter when [Cu₂O] = 60 at.%. In addition, it was determined that the films were homogeneous in depth since the diffraction patterns were essentially unchanged independently of the X-ray beam angle of incidence. Given that the amorphous films had i) low surface roughness and ii) for each [Cu₂O] concentration the reflectance spectra were similar to each other, the transmittance per thickness unit length (TTUL: τ) was a useful quantity to show the different electronic response of the amorphous samples which tend to group as a function of the concentration of Cu₂O in the target. The lower transmittance in the infrared was observed for the amorphous films grown with [Cu₂O] = 36 at.%. Since the most conducting films were also those with [Cu₂O] = 36 at.%, it can be concluded that the lower IR-transmittance (and correspondingly high IR-reflectivity) of these films is due to their higher free carrier density with respect to those with 12 or 60 at.%. In the case of the polycrystalline films, the transmittance and TTUL spectra were dominated by the optical properties of the particular copper telluride phase present in each sample. Variations in the optical response (transmittance window) of the berzelianite Cu_2-xSe phases indicate a different copper vacancy density/distribution of these phases. On the other hand, the films where the stoichiometric copper selenide form Cu₂Se-bellidoite was present, the corresponding transmittance spectra were alike. Finally, the electrical characterization showed that the ZnO-Cu_2-xSe composite films are n-type with resistivity values between 3x10⁻³ and 3x10³ Ω⋅cm. The optical properties of the polycrystalline films were dependent basically on the copper selenide phases present in the films, which present some transparency in the visible region and very low transmittance in the infrared. This allows foreseeing potential applications such as coatings for windowpanes with strong infrared rejection.