Electrochemical Atomic Layer Deposition of CdS on Ag Single

Article pubs.acs.org/JPCC

Electrochemical Atomic Layer Deposition of CdS on Ag Single Crystals: Effects of Substrate Orientation on Film Structure Francesco Carlà,*,† Francesca Loglio,‡ Andrea Resta,† Roberto Felici,† Elisa Lastraioli,‡ Massimo Innocenti,‡ and Maria Luisa Foresti‡ †

Experiment Division, ESRF, 6 rue J. Horowitz, F-38043 Grenoble, France Dipartimento di Chimica, Università degli studi di Firenze, Via della Lastruccia 3, 50019 Sesto Fiorentino, Italy

‡

ABSTRACT: Electrochemical atomic layer epitaxy (ECALE), also known as electrochemical atomic layer deposition, is a method based on under potential deposition, which allows the layer by layer growth of compounds on single crystal substrates. This paper reports on the electrochemical and structural investigation carried out on CdS thin films grown by ECALE on Ag(110) and Ag(100). Electrochemical experiments demonstrate that the CdS thin films, grown by ECALE, are highly ordered and epitaxial. The stoichiometric composition of the thin films was calculated using the charges associated with the anodic and cathodic stripping of CdS. The results indicate a 1:1 ratio between the elements. Characterization of the samples, performed by surface X-ray diffraction and X-ray reflectivity, suggests the presence of a crystalline CdS thin films epitaxial with the substrates. Moreover, a strong influence of the substrate orientation on the film structure is observed: CdS in wurtzite structure was found on Ag(100), while on Ag(110) the two CdS crystalline forms (wurtzite and zincblende) were present.

■

semiconductor film can be formed on the electrode surface by an alternate underpotential deposition (UPD) of the elements that form the compound. UPD is a surface-limited phenomenon, so the electrodeposition under UPD conditions is generally limited to one atomic layer.6 Compared with the codeposition process the ECALE method allows for a more strict control over the film thickness and stoichiometry, where the thickness is defined by the number of deposition cycles and the stoichiometry is determined by the amount of material involved in the UPD process. Another important aspect is that the layer-by-layer deposition, with respect to a 3-D nucleation process, is expected to promote an epitaxial growth and thus structural order. The possibility of controlling the composition and crystalline quality of the film makes ECALE a suitable method for application in several fields of technology such as photovoltaics and optoelectronics. For instance, the use of electrochemical methods in the electronic industry is much more than a remote possibility. The introduction of the damascene process for the Cu deposition in the microprocessors production7 proved that electrodeposition can satisfy the industry requirements in terms of precision, contamination, and large scale production. Nevertheless, to be competitive with today’s fabrication technology, the electrodeposition processes must be as reliable as any other vacuum deposition technique and should result in the production of high-quality films with well-controlled structural and electronic properties. This paper concludes our systematic work carried out to define the experimental parameters for growing CdS thin

INTRODUCTION Electrochemical deposition is an attractive method for the production of thin films of semiconductor compounds with high grade of crystallinity. Electrodeposition can be very competitive with respect to vapor phase deposition or vacuum methods for several reasons: low cost, room-temperature operation, the possibility of depositing on nonplanar substrates, and the direct control of film composition and thickness. Two methods for the electrodeposition of nonoxide semiconductor compounds have been reported so far in literature: codeposition and electrochemical atomic layer epitaxy (ECALE).1−4 Codeposition consists of the simultaneous electrochemical deposition of both elements forming the compound from a solution containing their precursors; despite the complexity of the chemistry involved, the codeposition process is rather fast and allows us to prepare good-quality films. Nevertheless, great care must be taken while choosing the deposition conditions to obtain films with well-defined properties. In fact, concentration, pH, and applied potential have an important role in determining stoichiometry, thickness, morphology, structure, and crystallinity of the film. It is also important to stress that problems related to the stability of the electrodeposition solution have sometimes been reported, as in the case of Cd-thiosulfite bath for CdS electrodeposition.1 Besides codeposition, well-ordered thin films of compounds can be obtained in a step-by-step process using a sequential electrochemical layer-by-layer growth. This method, first described by Stickney and coworkers5 and indicated as either ECALE or electrochemical atomic layer deposition (E-ALD), is effective for the growth of binary and ternary semiconductor compounds on metallic substrates with high crystalline properties. In the ECALE method, a binary or ternary © 2014 American Chemical Society

Received: June 7, 2013 Revised: February 25, 2014 Published: February 26, 2014 6132

dx.doi.org/10.1021/jp405637g | J. Phys. Chem. C 2014, 118, 6132−6139

The Journal of Physical Chemistry C

Article

films on silver single crystal substrates. In the past years, the ECALE method was successfully used for the deposition of well-ordered films of CdS on Ag(111) with high grade of crystallinity and good semiconductor properties.8−12 Up to now, a complete structural characterization of an ECALE electrodeposited CdS thin film has been performed only on films grown on Ag(111) single crystal electrodes.12 The CdS deposition on Ag low-index faces has been investigated to bridge the gap between deposition on single crystals and polycrystalline materials and open the possibility of applying the ECALE method on substrates without any preferential orientation. It is well known that UPD processes on single crystals are strongly dependent on the substrate orientation, and thus different structures can be expected for UPD deposition on different facets. For sulfur deposition on Ag substrates, different voltammetric features have been observed and described.13,14 Structural characterization by in situ STM of the sulfur UPD structure on Ag(111) was reported in the past,15 and in more recent times the investigation was also completed for Ag(100) and Ag(110).16 The aim of this work is to investigate the relation between the substrate orientation and the film structure and to elucidate how the substrate orientation can affect the structural order of the film. Moreover, it is not known if after several deposition cycles the film will maintain an epitaxial relation with the substrate or if it will grow in a more disordered structure. We describe the ECALE growth of CdS on Ag(110) and on Ag(100) with particular emphasis on the Cd deposition process on the S UPD layer and report the results concerning the structural characterization of the films by surface X-ray diffraction (SXRD).

electrode was an Ag/AgCl (KCl saturated) placed in the outlet tubing. Further information about the electrochemical setup has been published in previous papers.8,12 All potentials reported in the paper are quoted with respect to the Ag/ AgCl (KCl saturated) reference electrode. Surface X-ray Diffraction and X-ray Reflectivity. The SXRD experiments were carried out at the ID03 beamline of the European Synchrotron Radiation Facility (ESRF) in Grenoble.19 SXRD and X-ray reflectivity (XRR) experiments were performed on the six circle vertical axis diffractometer installed in the first experimental hutch. Samples were mounted on the diffractometer stage enclosed in a mylar bag filled with argon to prevent surface damage by oxygen and ozone. X-ray energy was chosen to be 12.4 keV (1.00 Å in wavelength). Intensities of the CdS (00l) specular rod were corrected considering the area of the illuminated surface, the polarization of the incident beam, and the Lorentz factor. Modeling and simulation of the specular reflectivity are described in the next section. The reflectivity curves were fitted using the software PARRAT32,20 a general purpose program developed at the HMI Institute of Berlin. The code is based on Parrat formalism21 and includes a roughness factor in the fitting parameter calculated according to Névot and Croce.22

■

RESULTS AND DISCUSSION S and Cd UPD Deposition. The ECALE deposition process of CdS layers was always started with the UPD deposition of S atoms; the first S UPD monolayer acts as a barrier against the possible Cd diffusion in the crystal bulk. The formation of the first layer of S on Ag(111), Ag(100), and Ag(110) has been already extensively investigated using cyclic voltammetry13,14 and STM.15,16 Electrochemical measurements show that sulfur UPD deposition processes differ significantly on the three silver facets, and, accordingly, in situ STM experiments have evidenced the presence of differently ordered sulfur structures depending on substrate orientation. Moreover, on Ag(110) and Ag(100), the UPD processes occur at more negative potentials than on Ag(111). In fact, on both faces, the sulfur UPD layer can be obtained (in pH 13 solutions) by applying a potential of E = −0.8 V, whereas on Ag(111), it was obtained at E = −0.68 V. Here we present cyclic voltammetry used to investigate the Cd deposition on the bare Ag substrate and on silver covered by a S UPD layer (0). In the latter case, after sulfur UPD deposition, the cell was washed with ammonia buffer to eliminate HS− while keeping the potential at the value of potential of sulfur deposition. Then, the potential was shifted to −0.1 V and the ammonia buffer was replaced by a 5 mM CdSO4 in a pH 9.6 buffer solution. Figure 1 shows the UPD process of Cd on the three lowindex faces of silver and on the same faces covered by a sulfur UPD layer. It must be remarked that the Figure reports only a single reduction/oxidation peak for each system examined, whereas on S-covered silver the Cd under potential deposition yields two consecutive UPD peaks. However, on Ag(100) the second UPD peak can be observed well only during the anodic scan because the second reduction peak partially merges with the first one and is partially masked by Cd bulk deposition. For this reason, we limited the potential range of Figure 1 to the one corresponding to the first UPD peak. The comparison between the different Ag faces implies that the potential of the reduction/oxidation peaks of the Cd UPD process on the bare silver depends on the atomic density of the different faces. More precisely, the UPD process occurs at more positive

■

EXPERIMENTAL SECTION Reagents and Electrochemical Preparation. The electrochemical growth was carried out using, without any further purification, Merck analytical reagent grade 3CdSO4· 8H2O, Fluka analytical reagent grade Na2S and Merck analytical reagent grade KClO4. Moreover, double-distilled water and Merck analytical reagent grade HClO4 (65%) and NH4OH (33%) were used to prepare solutions of 3CdSO4·8H2O 5 mM and Na2S 0.5 mM in pH 9.6 ammonia buffer. The solutions were freshly prepared just before the beginning of each series of measurements. Silver single crystals were made in our laboratory. Silver single crystal spheres were grown in a graphite crucible according to the Bridgman technique. They were aligned and cut in a crystallographic direction to obtain an Ag(110) and Ag(100) oriented electrode.17 Surfaces were then polished with emery paper (BuehlerMet SiC P1000, p2500, P4000) and successively finer grades of alumina powder down to 0.05 μm (Buehler Micropolish II).18 Before each measurement, the electrode was cleaned with water in an ultrasonic bath for 15 min and chemically polished using a patented procedure based on CrO3.17,18 An automated deposition apparatus consisting of Pyrex solution reservoirs, solenoid valves, a distribution valve, and a flow-cell was used under computer control. Both the distribution valves and the cell were designed and made in our workshop.8−11 The electrochemical cell was a Teflon cylinder with an internal diameter of 6.7 mm and a height of 40 mm. The electrochemical cell volume (1.5 mL) was delimited by the working electrode on one side and the counter electrode on the other side. The inlet and the outlet for the solutions were placed on the side walls of the cylinder. The counter electrode was a gold foil, and the reference 6133



Article

ECALE Deposition and Electrochemical Characterization. After the first UPD deposition of sulfur at −0.68 V, the deposition sequence of the ECALE cycles used to obtain CdS on Ag(110) and Ag(100) was: • deposition of cadmium at −0.6 V • washing of the cell with supporting electrolyte • deposition of sulfur at −0.68 V • washing of the cell with supporting electrolyte This basic sequence corresponds to the deposition of one CdS unit, and it is repeated as many times as necessary to obtain deposits with the desired thickness. To determine the amount of Cd and S deposited in a given number of cycles, we measured the transferred charge in the anodic stripping of Cd, followed by the cathodic stripping of S. As already observed in the case of the CdS deposition on Ag(111), the stripping potential of cadmium from CdS shifts toward more negative values as the number of deposition cycles increases, thus indicating that the compound is becoming more stable as the thickness of the deposit increases. The S layers remain after the Cd stripping. Except for the first layer, chemically bonded to the Ag substrate, the sulfur is reduced at the same potential as bulk sulfur. That is, during the subsequent reductive stripping, all sulfur is reduced at a more positive potential then the layer bound to the Ag surface. The charge deposited in each cycle is calculated by plotting the charges obtained in the stripping of Cd and S as a function of the number of ECALE cycles. The average values of the charge deposited in each ECALE cycle are 85 μCcm−2 for Ag(110) and 103 μCcm−2 for Ag(100). The charge associated with the deposition of S in the first ECALE cycle is bigger than the amount of charge calculated from the stripping of a thicker film. The amount of S deposited in the first UPD monolayer was in fact estimated to be 163 μCcm−2 for Ag(100) and 137 μCcm−2 for Ag(110), as reported in a previous paper.16 This behavior is not surprising because it has also been reported for CdS and ZnS on Ag(111).8,15 It is worth mentioning that the decrement of the deposited charge shows the same trend for Ag(100) and Ag(110). In fact, the ratio between the average S charges deposited in each cycle and the charge deposited in the first cycle is the same for Ag(100) and Ag(110). A summary of the S deposited charge on Ag calculated from S stripping is presented in Table 1.

Figure 1. Cyclic voltammograms of Cd on S-covered and bare Ag(111), Ag(100), and Ag(110), obtained from 5 mM CdSO4 in ammonia buffer pH 9.6 solutions. The scan rate was 10 mV/s.

potentials on the most compact (111) plane and then follows the order (111) > (100) > (110). This is not surprising because this is also the trend of comparable chemisorption processes on the same faces. On S-covered silver substrates, all underpotential deposition/reoxidation processes are expected to follow the same trend and to occur at even more positive potentials due to the stronger interaction of Cd with S than with Ag. Indeed, this is generally verified with the exception of Cd deposition on S-covered Ag(100) that practically coincides with the deposition of Cd on bare Ag(100). This occurrence is an intriguing exception that deserves deeper investigations. It is worth noting that also the S UPD deposition on Ag(100) presents some significant differences with respect to the S UPD on Ag(111) and Ag(110). Cyclic voltammograms of the S UPD on Ag(110) and Ag(111) show oxidation and reduction peaks with very narrow and sharp shape in correspondence of the formation of the high density S UPD monolayer, while on Ag(100) only a broad peak is observed. Moreover, the structural STM investigation on Ag(111) and on Ag(110) revealed the presence of only one structure of S in correspondence of each voltammetric feature, while patches of two different structures could be observed on Ag(100).15,16 We can tentatively assume that the negative shift in potential for the Cd deposition process on Ag(100) might be related to the properties of the S UPD layer underneath.

Table 1. S Deposited Charge on Aga

Ag(111)8,15 Ag(110)16 Ag(100)16

S UPD (calculated) (μCcm−2)

S UPD (experimental) (μCcm−2)

S (charge per cycle) (30 cycles CdS film) (μCcm−2)

189 136 192

137 163

70 85 103

a

Total charge of S UPD on Ag(111) can not be measured because the S stripping takes place in the hydrogen evolution region. Theoretical charges of S UPD layer on Ag are calculated on the basis of STM data reported in literature.15,16

Structural Characterization. X-ray diffraction experiments were carried out on samples prepared with 30 ECALE deposition cycles on Ag(100) and Ag(110) substrates, as previously described. CdS has two stable crystalline forms: wurtzite and zincblende. In our previous investigations on Ag(111), only CdS in wurtzite-like structure was found.9,12 Nevertheless, because the structures of the S and Cd UPD layers on Ag strongly depend on the substrate orientation, the 6134



Article

crystallographic structures of CdS thick films on Ag(100) and Ag(110) are expected to be different compared with the one observed on Ag(111). Diffraction patterns of wurtzite and zincblende are quite similar and have many diffraction peaks overlapping in the same 2θ position. However, the two structures can be identified using the (200) reflection for zincblende and the (101) and (103) reflections for wurtzite, which are well separated and characteristic for the two structures. To identify the two structures, we have recorded azimuthal scans at low values of the vertical exchanged momentum for different 2θ positions. (Further references to experimental geometry can be found elsewhere.23) The characteristic diffraction peaks of the CdS wurtzite structure were found for films grown on both substrates, while the (200) reflection of a zincblende was observed only in samples grown on the Ag(110). Figures 2−4 show the dependence of the Figure 4. Azimuthal scan showing the in-plane orientation of the zincblende CdS structures on Ag(110). The position of the (200) reflections indicates a 45° rotation of the structure with respect to the silver main axis.

Figure 2 the presence of two different CdS domains rotated by 30° with respect to each other and ±15° with respect to the main axes of the substrate can be identified. A similar situation was found on Ag(110), where two wurtzite domains rotated by 30° with respect to each other and, respectively, aligned along the main axis of the Ag substrates were observed (Figure 3). Comparing the diffracted intensities of the two domains, it is evident that the growth of CdS is preferentially oriented along the K axis (or Ag [010] direction in the used reference system) of the Ag surface unit cell (Figure 3). As already mentioned, the presence of CdS in the zincblende structure on Ag(110) substrates was also observed. In this case, the CdS has only one domain oriented with a rotation of 45° with respect to the silver main axes (Figure 4). Comparing the normalized values of the diffracted intensities of the zincblende (200) and wurtzite (100) reflections on Ag(110), it appears that the CdS growth is favored in zincblende structure. Diffracted intensity maps in the reciprocal space shown in Figure 5 confirm the azimuthal scan results. The reciprocal space maps report the intensity using a coordinate system referred to the surface unit cell of the substrate. According to this convention,24 the correspondence between the bulk unitcell parameters (a0,b0,c0,α0,β0,γ0) and surface unit-cell parameters (a,b,c,α,β,γ) is given by a = b = a0/√2, c = a0, and α = β = γ = 90 for Ag(100) and a = a0/√2, b = c = a0, and α = β = γ = 90 for Ag(110). The intensity maxima visible in both maps at ⟨110⟩ correspond in the bulk unit cell reference system to the (200) reflection for Ag(100) and to the (101) reflection for Ag(110). The maxima at fractional h and k values are instead originated by the CdS film. The wurtzite unit cells are indicated in the Figure by red and blue solid lines, while dashed red and blue lines form a grid to facilitate the visualization of the positions in the film lattice. By comparing the intensity of the two CdS wurtzite domains in the Ag(110) map, we find a confirmation of the preferential growth along one of the two orientations (as previously observed in the azimuthal scans). The peaks of the domain oriented in the less favored growth direction (indicated by the blue lattice in Figure 5) are not always clearly visible but still distinguishable in position equivalent to (0.98,1.21,0), which corresponds to (110) in the CdS wurtzite surface unit-cell coordinates. As a matter of

Figure 2. Azimuthal scan showing the (101) reflections of the wurtzite CdS film on Ag(100). Two domains rotated ±15° with respect to the main axis are clearly visible.

Figure 3. In-plane orientation of CdS wurtzite film on Ag(110). Azimuthal scan shows the (101) reflections of the two CdS domains aligned along the substrate main axis. Peak intensities suggest preferential direction of growth along the h lattice vector.

(101) reflection of the CdS wurtzite reflection as a function of the azimuth. The wurtzite structure on Ag(100) is epitaxial and has two favored directions of growth. From data reported in 6135



Article

Figure 5. Mesh scans in the hk reciprocal space plane for l = 0.05. Diffraction pattern shows the in-plane orientations of CdS domains on Ag(100) and Ag(110). In both cases, two wurtzite domain rotated by 30° can be observed (red and blue solid lines indicate the CdS unit cells). On Ag(110), CdS with a zincblende structure is also present, this domain is rotated by 45° with respect to the substrate unit cell (solid black lines indicates the zincblende unit cell).

fact, the experiments show that CdS in both zincblende and wurtzite structures grows epitaxial on either Ag(100) and Ag(110) with a high level of order because no evidence of the presence of disordered CdS was observed. In this latter case, Debye rings should be present, but they are completely undetectable. By the position of the in-plane reflections we calculate that in both cases the bulk unit cells of wurtzite and zincblende are aligned with the c axis perpendicular to the surface. The results of the structural analysis are summarized in Figure 6, where a schematic diagram is reported showing the relative orientation of substrate and film unit cells along with a drawing representing the arrangement of first and second layer of atoms (respectively S and Cd) in the basal plane of the wurtzite and zincblende unit cell. In Figure 7, we show the large-angle specular reflectivity curve for the Ag(100) sample including the wurtzite CdS(002) and Ag (002) reflections. The blue line is a simulation to the data obtained by summing the calculated structure factors of Ag and CdS. Diffracted intensity and fringes periodicity of the CdS films depend on the number of times the CdS unit cell is repeated along the l direction perpendicular to the surface, while their damping is due to a disorder in the coherent thickness of the film. The first information that can be extracted from the curve is from the position of the maximum of the CdS(002) peak, which corresponds to a wurtzite c-axis lattice parameter of 6.67 Å, slightly shorter than the bulk lattice parameter value. To simulate the data, we used a model that assumes the presence of several CdS domains with different thickness whose intensity is summed incohorently. The distribution of the thickness follows a standard distribution. With this assumption, we can evaluate that the average thickness of the crystalline layer is ∼32 Å with a sigma of 2.3 Å. By performing an accurate analysis of the diffracted intensity in the reciprocal space, we can exclude the presence of other phases of the CdS and of disordered CdS. As already mentioned in the discussion of the data presented in Figure 5, in the presence of a disordered CdS phase, Debye rings should be present, but they are completely undetectable.

Figure 6. Arrangement of first and second layer of atoms in the film unit cell (respectively S and Cd) along the basal plane of wurtzite (a) and zincblende (b). Schematic illustrations depicting the relative orientation of film and substrate lattices for Ag(100) (c) and Ag(110) (d). The drawings in panels c and d have been reported with the sole purpose of displaying the orientation of the two lattices. Any in-plane translation of the film unit cell on the Ag surface lattice is possible as the exact position of the first atomic layer of S on the Ag lattice cannot be determine with the experiments described in the paper.

Reflectivity. The low-angle specular reflectivity curve from a surface does not contain information on the crystalline properties of a film because it depends only on the electronic density profile as a function of the coordinate z perpendicular to the surface. Information, which can be gathered by such measurements, includes the total thickness of the film and its density together with the surface and interface roughnesses. The XRR curves, measured for the samples growth on the 6136



Article

These values differ by ∼14 Å, which corresponds to the thickness of two unit cells. The difference can be explained considering that the XRR measure depends only on the electron density profile along the surface normal, while only the ordered part of the film is contributing to the diffracted signal measured in the specular reflectivity.25 Thus, Cd or S layers laying in positions that do not match with the wurtzite lattice will not contribute to the diffracted signal in the specular reflectivity. For this reason, we suggest that the difference between the two data sets is due to the presence of distorted CdS cells either in contact with the silver surface lattice or at the surface. In fact, both of these regions would not contribute to the diffracted intensity. The CdS wurtzite structure oriented on Ag(100) as previously reported can be seen as a sequence of Cd and S layers stacked along the normal to the surface. The charge associated with the deposition of an S atomic layer calculated according to this structure is 216 μCcm−2, which is twice the value measured in electrochemical experiments (103 μCcm−2) for each ECALE cycle. It is clear from the experimental data that during the deposition of each atomic layer only about half of the charge needed to cover a full atomic plane with the wurtzite structure is deposited. This is in good agreement with the structural data. In fact, assuming that each ECALE cycle is depositing a charge equivalent to a CdS layer with the same structure reported in Figure 6, the total thickness of the sample after 30 ECALE deposition cycles should be 100.7 Å, while the thickness measured by XRR is about half of this value (46.5 ± 0.5 Å). We can conclude that the theoretical and experimental values of film thickness and deposited charge are in good agreement, proving the consistency of the experimental results. Electrochemical experiments show without a doubt that in UPD electrodeposition processes after the first ECALE cycle the deposited charge is constant but smaller than the charge associated with the first UPD layer and with the theoretical values calculated for a full monolayer. This phenomenon can not be easily explained, but it is not a surprising finding because the same behavior has already been observed in the past for similar systems.8,10−12 It is important to point out that an incomplete deposition of UPD monolayers during the film growth does not affect the crystalline quality of the film; however, the presence of a thick disordered layer at the surface can raise some doubts about the ECALE growth. It is assumed that in each step of the ECALE cycles the atoms are supposed to occupy a well-defined positions on the surface lattice and that they do not move anymore. However considering that after the first cycle we deposit a fraction of the charge associated with the deposition of the first UPD monolayer and that we have a structural disorder at the surface with a thickness of about one unit cell (or a CdS bilayer), we can suppose that during one cycle only half of the monolayer is deposited and to fill up the monolayer it is necessary to stabilize the top layer by depositing the relative counterion. In other words, it is possible that during the deposition process the top layer (or layers) may minimize the energy of the truncated lattice, rearranging the atoms in a structure different from the bulk. Following the deposition of another UPD layer, the surface will again modify its structure, rearranging the top layer to minimize the energy while the lower distorted layer underneath will order in the bulk structure. According to this interpretation, we can imagine the deposition of a monolayer as a dynamic process occurring in steps where rearrangements and reordering of the atoms can

Figure 7. Specular rod scan on CdS films on Ag(100).

Ag(100) and Ag(110) surfaces, are reported in Figure 8. Plots do not include data for values of the momentum transfer (Qz)

Figure 8. X-ray reflectivity measured from CdS films on Ag(100) and Ag(110). Fit of the data has been obtained using the same scattering density value of CdS canonical value, which confirms the exact 1:1 stoichiometry of the compound.

smaller than 0.20 Å−1. This is because at low incidence angles the illuminated surface is larger than the surface region covered by the CdS film (6.5 mm in diameter); therefore, the reflected intensity also includes the contribution of the Ag surface. The fit was performed assuming as fixed parameter the density of the Ag substrate. The fit provides for the density of the CdS films values, which match well the data reported for analogous CdS film on Ag(111).16 The determined scattering length density is (3.6 ± 0.2) × 10−5 Å−2, in agreement with the expected value of 3.63 × 10−5 Å−2. CdS film thickness is 52.5 ± 0.5 Å on Ag(110) and 46.5 ± 0.5 Å for Ag(100). In the case of the Ag(100) sample, we can compare the film thickness determined by low-angle and large-angle specular reflectivity. 6137



■

take place. Our findings could also be interpreted by the deposition-ordered islands covering only half of a monolayer at each cycle. However, in this way the film roughness would increase enormously with the film thickness, and this has never been observed. For this reason, we tend to believe that the ECALE deposition process for the CdS is probably a two-step phenomena in which two full cycles are necessary for the deposition of one CdS layer. It is important to note that during the growth the structure of the film remains epitaxial with the substrate, even at a distance of several unit cells from the substrate/film interface. Given the fact that zincblende and wurtzite structures are both based on a tetragonal bonding geometry, we could imagine a more disordered growth process, where film structure is changing during deposition, to be possible. Nevertheless deposition of ordered films with defined structure and epitaxial relation with the substrates was always observed.

REFERENCES

(1) Pandey, R.; Sahu, S.; Chandra, S. Handbook of Semiconductor Electrodeposition; CRC Press: New York, 1996. (2) Stickney, J. L. In Advances in Electrochemical Science and Engineering; Alkire, R., Kolb, D., Eds.; John Wiley & Sons: New York, 2001; Vol. 7, pp 1−105. (3) Stickney, J. L. In Electroanalytical Chemistry; Bard, A. J., Rubinstein, I., Eds.; CRC Press: New York, 1999; Vol. 21, pp 75−95. (4) Switzer, J. A. In Electrodeposition of Superlattices and Multilayers; Hodes, G., Ed.; Wiley- VCH: Weinheim, Germany, 2007; pp 67−101. (5) Gregory, B. W.; Stickney, J. L. Electrochemical Atomic Layer Epitaxy (ECALE). J. Electroanal. Chem. 1991, 300, 543−561. (6) Bockris, J.; Reddy, A.; Gambo, M. E. Modern Electrochemistry 2A: Fundamentals Of Electrodics; Kluwer Academic/Plenum Press: New York, 2000. (7) Andricacos, P. C.; Uzoh, C.; Dukovic, J. O.; Horkans, J.; Deligianni, H. Damascene Copper Electroplating for Chip Interconnections. IBM J. Res. Dev. 1998, 42, 567−574. (8) Innocenti, M.; Pezzatini, G.; Forni, F.; Foresti, M. L. CdS and ZnS Deposition on Ag(111) by Electrochemical Atomic Layer Epitaxy. J. Electrochem. Soc. 2001, 148, C357−C362. (9) Cecconi, T.; Atrei, A.; Bardi, U.; Forni, F.; Innocenti, M.; Loglio, F.; Foresti, M. L.; Rovida, G. X-ray Photoelectron Diffraction (XPD) Study of the Atomic Structure of the Ultrathin CdS Phase Deposited on Ag(111) by Electrochemical Atomic Layer Epitaxy (ECALE). J. Electron Spectrosc. Relat. Phenom. 2001, 114−116, 563−568. (10) Innocenti, M.; Cattarin, S.; Cavallini, M.; Loglio, F.; Foresti, M. Characterisation of Thin Films of CdS Deposited on Ag(111) by ECALE. A Morphological and Photoelectrochemical Investigation. J. Electroanal. Chem. 2002, 532, 219−225. (11) Foresti, M. L.; Capolupo, F.; Innocenti, M.; Loglio, F. Visual Detection of Crystallographic Orientations of Face-Centered Cubic Single Crystals. Cryst. Growth Des. 2002, 2, 73−77. (12) Foresti, M. L.; Pozzi, A.; Innocenti, M.; Pezzatini, G.; Loglio, F.; Salvietti, E.; Giusti, A.; D’Anca, F.; Felici, R.; Borgatti, F. In situ X-ray Analysis under Controlled Potential Conditions: An Innovative Setup and its Application to the Investigation of Ultrathin Films Electrodeposited on Ag(111). Electrochim. Acta 2006, 51, 5532−5539. (13) Hatchett, D. W.; White, H. S. Electrochemistry of Sulfur Adlayers on the Low-Index Faces of Silver. J. Phys. Chem. 1996, 100, 9854−9859. (14) Hatchett, D. W.; Gao, X.; Catron, S. W.; White, H. S. Electrochemistry of Sulfur Adlayers on Ag(111). Evidence for a Concentration- and Potential-Dependent Surface-Phase Transition. J. Phys. Chem. 1996, 100, 331−338. (15) Aloisi, G. D.; Cavallini, M.; Innocenti, M.; Foresti, M. L.; Pezzatini, G.; Guidelli, R. In Situ STM and Electrochemical Investigation of Sulfur Oxidative Underpotential Deposition on Ag(111). J. Phys. Chem. B 1997, 101, 4774−4780. (16) Lastraioli, E.; Loglio, F.; Cavallini, M.; Simeone, F.; Innocenti, M.; Carlà, F.; Foresti, M. L. In Situ Scanning Tunneling Microscopy Investigation of Sulfur Oxidative Underpotential Deposition on Ag(100) and Ag(110). Langmuir 2010, 26, 17679−17685. (17) Hamelin, A. In Modern Aspects of Electrochemistry; Plenum Press: New York, 1985; Vol. 16, pp 1−101. (18) Kurasawa, T. Patent Japan 35:5619, 1960. (19) Balmes, O.; van Rijn, R.; Wermeille, D.; Resta, A.; Petit, L.; Isern, H.; Dufrane, T.; Felici, R. The ID03 Surface Diffraction Beamline for In-Situ and Real-Time X-Ray Investigations of Catalytic Reactions at Surfaces. Catal. Today 2009, 145, 220−226. (20) Software Reflectivity Tool Parrat 32; HMI: Berlin, 1999. (21) Parratt, L. G. Surface Studies of Solids by Total Reflection of XRays. Phys. Rev. 1954, 95, 359−369. (22) Névot, L.; Croce, P. Caractérisation des Surfaces par Réflexion Rasante de Rayons X. Application à l′Étude du Polissage de Quelques Verres Silicates. Rev. Phys. Appl. 1980, 15, 761−780. (23) Vlieg, E. Integrated Intensities Using a Six-Circle Surface X-Ray Diffractometer. J. Appl. Crystallogr. 1997, 30, 532−543.

■

CONCLUSIONS The results confirm that ECALE is an effective method to grow crystalline films with a high level of order. Structural analysis shows that film orientation and structure depend on the substrate lattice structure. Nevertheless, the film presents an high grade of crystallinity on all substrates. On both facets, Ag(100)and Ag(110), the growth of CdS in both zincblende and wurtzite structures has been proved to be epitaxial. Substrate lattice clearly influences the CdS structure; the presence of two wurtzite domains oriented along symmetrical direction with respect to the substrate main axes is consistent on Ag(100), Ag(110), and Ag(111). Besides, for Ag(111), a different relative orientation of the two wurtzite domains is reported. The experiments clearly show that disordered CdS is not present on the surface, proving the ECALE method to be a versatile and powerful tool for the electrodeposition of crystalline thin films of semiconductor compounds. Electrochemical measurements on Ag(100) and Ag(110) indicate that the charge associated with each CdS and S layer has an average value comparable to that found for films grown on Ag(111). XRR and diffraction data confirmed that film thickness can be controlled by the number of cycles. Conditions for CdS ECALE deposition (potential and deposition time) are similar to those previously used for deposition on Ag(111). This suggests that ECALE might be successfully used in the future for CdS deposition on substrates without a preferential orientation, and the crystallinity of the film should be probably unchanged. Because the method is very inexpensive, its application for the production of high-crystalline semiconductor thin films represents an interesting alternative to the other techniques used in the deposition of thin films of semiconductor compounds.

■

Article

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. Phone: +33 (0) 476 88 2680. Fax: +33 (0) 476 88 2325. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS We are grateful to Andrea Pozzi and Ferdinando Capolupo. 6138



Article

(24) Robinson, I. K.; Tweet, D. J. Surface X-Ray Diffraction. Rep. Prog. Phys. 1992, 55, 599−651. (25) Stierle, A.; Vlieg, E. Modern Diffraction Methods; Wiley-VCH: Weinheim, Germany, 2012.

6139


Electrochemical Atomic Layer Deposition of CdS on Ag Single

Recommend Documents