HIGH PERFORMANCE ZNO BASED BACK REFLECTORS FOR THIN FILM SOLAR CELL APPLICATIONS
Project at the FCT webpage >
Main Research Area
Material Sciences and Engineering - Ceramics, Glass and Energy Materials Systems
· Photovoltaic cells
· Back reflector
· Zinc Oxide
Project cofunded by COMPETE
· Main Contractor
Universidade de Aveiro (UA)
· Main Research Unit
Centro de Tecnologia Mecânica e Automação (TEMA/UA)
· Principal Investigator
Victor Fernando Santos Neto
António Ferreira da Cunha
Joana Catarina Martins Mendes
Bárbara Filipa Casqueira Coelho Gabriel
Cátia Sofia Castanheira dos Santos
Li Qin Zhou
Raul Nunes Simões
Amorphous Si (a-Si) based thin film solar cells are one type of photovoltaic devices that convert sunlight directly into electricity. They have become very attractive in recent year due to the potential of large area production at low cost . The cells consist of one or more n-i-p junctions, each of which includes three silicon layers: n-type Si, intrinsic Si and p-type Si. The typical cell structure is SS/Ag/ZnO/n-i-p/TCO or Glass/TCO/p-i-n/ZnO/Ag if stainless steel (SS) or glass is used as substrate, respectively, where TCO is transparent conductive oxide coatings acting as front electrode.
In these cells, the Ag/ZnO (sometimes Al/ZnO) is called back reflector (BR) for light trapping to promote photocurrent. It has been shown that BRs with ideal lambertian surface can significantly increase the light pass, leading to ≈30% higher current density in practical solar cells . Therefore, without BR high efficiency solar cells cannot be achieved.
The present project focuses on high rate sputtering deposition of the ZnO based BR by developing and utilizing a novel ZnO:Zn target composition. Currently, the most common and successful method to prepare the Ag/ZnO BR is sputtering. While sputtering deposition of Ag film is fast and simple, ZnO deposition appears the bottleneck in the BR process.
Two sputtering techniques are widely used to deposit the ZnO films. One is RF sputtering from pure ZnO ceramic target. Another is DC reactive sputtering from a metal Zn target. Although RF sputtering results in high quality films it has the limitation of relatively low deposition rates. The ZnO film in a typical BR is around 600nm thick. Using RF sputtering to deposit such a thick film is not feasible in practical applications.
DC reactive sputtering meets the requirement for deposition rate. But there is a fundamental process stability problem due to oxide poisoning of the metal surface in reactive sputtering. To address these issues, several methods have been developed.
The most successful one is doping the ZnO with ≈3% Al2O3 to make the ceramic target conductive . This avoids the target poisoning problem while allowing high rate DC sputtering. However, Al ion migration into the Si layers becomes a big concern. In addition, the resulted films appear less transparent, especially in the red light region in which a-Si absorbs weakly.
Alternatively, AC reactive sputtering using a pair of magnetron cathodes with Zn metal target has been conducted . This technique eliminates arcing problems caused by target poisoning. However, besides the complicated hardware configuration, the AC discharge voltage is very high. This leads to large clusters being sputtered off from the target and deposited in the film. As a result, the solar cells often get shunted by those sharp points in the film surface.
The ideal ZnO sputtering process would combine the high rates of DC reactive sputtering with the film quality and process stability of RF sputtered ZnO films from a pure ceramic target. The ZnO:Zn target composition proposed in this project is for achieving these goals without introducing harmful impurities to the a-Si solar cells.
This idea is based on two essential findings. First, according to the percolation theory, it is feasible to obtain conductive ZnO:Zn composition once the conductive phase (metal) reaches a certain ratio . Second, we have recently proved that a SiO2:Si composition could become conductive once the fraction of Si (p-type doped) reached about 20% .
The conductive SiO2:Si composition had a ceramic nature and did not show target poisoning during reactive sputtering, while the deposition rate was high. Using the conductive ZnO:Zn composition as sputtering target, Ag/ZnO BRs will be systematically investigated to achieve the most suitable surface texture.
Meanwhile, theoretical study on the features of ideal lambertian surface will be conducted to guide the development of the BRs. Finally, the BRs made from the ZnO:Zn composition will be integrated in a-Si based thin film solar cells to verify their effectiveness. Light soaking and acceleration aging tests will be conducted to check the stability of the resulted solar cells.
The research team has extensive experiences in ceramic science and engineering, thin film physics and sputtering technology, plasma enhanced chemical vapor deposition and a-Si thin film solar cells. Besides quality research publications in these areas, we have been collaborating with Martifer (an energy company in Portugal) on solar energy conversion and utilization for a few years. The combination of expertise in all these fields provides a unique advantage and solid base towards successful completion of the project.
Objectives / Milestones
1) Achieve conductive ZnO:Zn composition
2) Fully understand the sputtering of the ZnO:Zn
3) Achieve properly textured ZnO based back reflectors
4) Achieve ZnO/n-Si interface match
5) Achieve high efficiency a-Si based solar cells using ZnO
In general, the research work developed was in good agreement with what was initially purposed.
1) Develop key processes of synthesizing a conductive ZnO:Zn composite
One of the challenges in the synthesis of the ZnO:Zn composition was to effectively prevent the metal Zn in the composition from being oxidized and the ZnO from losing oxygen during the sintering. Therefore, sintering needed to be employed in a proper atmosphere. A tube furnace that could flow different gases was established for this research project, as shown in Figure 1. Two types of gases, nitrogen and argon, were tested. Although nitrogen gas could effectively prevent the oxidation of Zn, ZnO tended to get reduced during a long period of sintering. Argon seemed to be an effective gas in preventing both the oxidation of Zn and reduction of ZnO and therefore was chosen as sintering atmosphere. To eliminate the air in the furnace, Ar was flown for more than 2 hours prior to heating the samples. Ar gas was maintained during the sintering until the furnace was cooled to room temperature.
Figure 1. Tube furnace established for this research project.
Standard ceramic processing was used to prepare the ZnO:Zn composite compacts. The process included mixing ZnO powder with metal Zn powder, and high pressure pressing of mixed raw materials into tablet followed by sintering in Ar. The compositions studied were expressed as (1-x)ZnO+xZn, where x is a ratio in weight and x=0, 10%, 15%, 20%, 30%,40%, 50% and 60%. Small size pellets with 10 mm in diameter and around 5 mm in thickness were prepared at the initial materials investigation stage. The effects of the Zn/ZnO ratio and the sintering temperature on the density and conductivity of the resulted compacts were systematically investigated. The composite density increased with the sintering temperature up to ~800°C and then decreased with further increasing the sintering temperature, as shown in Figure 2, which represents for x=20% composition. This density variation was due to the vaporization of Zn.
Figure 2. Dependence of density on sintering temperature for (1-x)ZnO+xZn (x=20%) composition. The percentage of the theoretical density is the ratio of measured density to the theoretical density of the composition that is calculated by 80%´5.6 g/cm3+20%´7.14 g/cm3 where 5.6 and 7.14 g/cm3 are the theoretical density of ZnO and Zn respectively.
2) Structural and electrical properties of ZnO:Zn composites
The results of X-ray diffraction (XRD) analysis for the studied compositions (sintered at 800°C) are shown in Figure 3. It can be seen that all the ZnO:Zn composites contained two phases: ZnO and Zn. No additional phase was found. The amount of the Zn phase increased with the x value, as expected.
Figure 3. X-ray diffraction patterns of the compositions of (1-x)ZnO+xZn
The resistivity values at room temperature for the studied compositions are shown in Table 1. It can be seen that pure ZnO is a semiconductor. For the ZnO:Zn composites, the resistivity decreased very slightly when x<0.2 as the x value increased, dropped dramatically at x=0.2 and then decreased when x>0.2. This trend can be seen more clearly in Figure 4, where the resistivity values shown in Table 1 are plotted versus the x values. The abrupt decrease of the resistivity value at x=0.2 indicated that a finite conductive path was formed for this composition.
Table 1. Resistivity for the composite (1-x)ZnO+xZn (sintered at 800°C in Ar)
Figure 4. Resistivity versus x value for (1-x)ZnO+xZn compositions
According to percolation theory, when conductive particles are dispersed into an non-conductive matrix and the amount of the conductive phase increases from zero up to a critical volume fraction of percolation, one particle contact with neighbors and form finite conductive path where the entire composite is non-conductive. Near the critical volume, infinite conductive network is formed and the resistivity of the entire composite abruptly decreases. Our results appeared in good agreement with the percolation theory. The critical fraction of the conductive Zn phase is ~20% in weight in the present study.
SEM/EDX analysis was performed to confirm this assumption. The SEM image in Figure 5 depicts the surface microstructures of a ZnO:Zn composite with 40 wt% Zn. The composite is reasonably dense, which is in agreement with the density measurement. The morphology features globules (brighter color) distributed uniformly in a matrix (relatively dark color). The globules appeared to be the location where ZnO exists, as evidenced by EDX mapping shown inFigure 5, which indicates uniform distribution of Zn (contributed from both metal Zn and ZnO) and scattered O (contributed mainly from ZnO). EDX scanning focused at the globule region shows that this region consisted of 13.5 wt% O and 86.5 wt% Zn (Table 2), indicating that the ZnO is completely surrounded/covered with metal Zn. On the other hand, scanning of the region in between the globules indicates that it is Zn-rich (Table 2), which implies that the metal Zn forms a conductive matrix. Furthermore, the excess zinc could also contribute to improving the conductivity of the ZnO:Zn composite due to ionized Zn defects in ZnO lattice and Zn interstitial defects that act as donors. This was evidenced by the previous research reported in the literature.
Figure 5.(a) SEM image, (b) EDX mapping of O, and (c) EDX mapping of Zn of a ZnO:Zn composite with 40wt% Zn.
Table 2. EDX analysis results of O and Zn concentration in different regions in the ZnO:Zn composite surface.
3) Fabricate ZnO:Zn composite target
With the identification of the critical Zn fraction and the appropriate sintering temperature, we fabricated 2” diameter targets. For comparison, four compositions were attempted: pure ZnO and with Zn fraction of 20, 30 and 40 wt%, respectively. Pure ZnO was sintered at 1300 °C as there was no issue of Zn vaporization. Other compositions were sintered at 800 °C. Ar atmosphere was used in sintering for all samples. All targets showed fairly high density (>80% of the theoretical density). Initial sputtering tests were then subsequently conducted for those targets. Although the sputtering rate of ZnO:Zn composite targets was significantly higher than that of pure ZnO, the targets seemed to need further improvement for sputtering high quality films. As can be seen from Figure 6, the targets showed some defects after initial sputtering tests. Hot press technique is being employed in attempt to maximally increase the density of the target.
Figure 6. Photos of the 2” (1-x)ZnO+xZn sputtering targets after initial sputtering.
4) Sputtering characterization of conductive ZnO:Zn targets
ZnO:Zn sputtering targets with different Zn concentration were fabricated. The sputtering characteristics of the ZnO:Zn target in DC, pulse DC, and RF plasma discharge were studied in comparison with metal Zn, Al doped ZnO (AZO), and pure ZnO targets. The results obtained from the ZnO:Zn target with 40 wt% Zn are presented here, while other composition targets with x=20% and x=30% showed similar behavior as the x=40% target and were not repeated. Figure 7 shows the DC discharge characteristics in pure Ar gas. The ZnO:Zn target exhibits discharge voltage slightly higher than the AZO target, but considerably lower than the metal Zn target. The reason is well known: metals usually have lower secondary electron emission coefficient than their corresponding oxides. This is a clear indication that the ZnO:Zn composite target has both metal and oxide characteristics.
Figure 7. DC discharge characteristics of metal Zn, ZnO:Zn, and AZO targets.
The ZnO:Zn composite appeared not a simple mixture of ZnO and Zn, as evidenced by the dependence of the discharge voltage on the oxygen flow rate in DC sputtering at 30 W power, 10 sccm total gas flow (Ar + O2) and 4 mTorr gas pressure. As can be seen in Figure 8, increasing oxygen flow (simultaneously decreasing Ar flow to maintain a constant pressure) has only a little effect on the discharge voltage for the ZnO:Zn target (x=40%), while the metal Zn target exhibits the typical sudden drop in the discharge voltage as the O2 flow increases to~30%, which is a result of strong secondary electron emission from the oxide layer formed on the target surface. As expected, the AZO target does not show voltage drop, as it is already fully oxidized. It is worth to point out that the discharge voltage increases slightly with the oxygen ratio, which is due to the lower ionization coefficient of O2 as compared to Ar.
Figure 8. Oxygen effects on discharge voltage of ZnO:Zn, metal Zn, and AZO targets at 30 W DC sputtering power.
Discharge stability of the ZnO:Zn target was tested at 30 W DC power with 30% O2 ratio, which was more than enough to produce fully oxidized ZnO film. During 240 minutes sputtering, no arcing was recorded from the power supply. This stable discharge indicated that target poisoning did not occur in the ZnO:Zn target.
Figure 9 shows the deposition rates of the ZnO films DC and RF sputtered from the ZnO:Zn, metal Zn, AZO, and pure ZnO, respectively. Note that the pure ZnO target was not sputtered by DC because of its high resistance. Interestingly, the ZnO:Zn composite target yielded clearly higher deposition rates than metal Zn, AZO and ZnO targets in both DC and RF sputtering.
Figure 9. Comparison of deposition rates from the ZnO:Zn, Zn, AZO, and ZnO target in (a) DC and (b) RF sputtering.
The high deposition rate from the ZnO:Zn composite may be due to the difference in the kinetic energy transfer that occurs in this target compared to the metal Zn, pure ZnO, and AZO targets. It is likely that the energy of ions striking the target surface of Zn, ZnO, or AZO may be more readily dissipated in a material such as polycrystalline Zn or ZnO that have highly ordered structures. In the case of the mixed phase ZnO:Zn material the interface between the two different materials may act as a barrier for energy dissipation within the body of the target material. As a result, the kinetic energy of the sputtering ions is localized in a small region on the surface. This in turn promotes a higher sputtering yield. We are conducting further work to verify the mechanisms that accounts for the high sputtering yield of the ZnO:Zn composite target.
5) Evaluation of sputtered thin films
A stable sputtering process and high deposition rates of ZnO thin films are very useful benefits for coatings and optoelectronic device applications such as solar cells. Hence, a preliminary study on the electrical and optical properties of the resulting films was conducted. Figure 10 shows the transmittance spectra of the ZnO films prepared by RF sputtering from ZnO:Zn, Zn, AZO, and ZnO targets. It can be seen that all these films have very good transmittance in the whole visible light range. The high transmittance implies that these ZnO films were fully oxidized. Note that the RF sputtering power was 50 W and the process gas contained 30% O2 mixed with Ar. This amount of oxygen was needed for the reactive sputtering of the metal Zn target (see Fig. 8) to obtain ZnO films. However, such an oxygen ratio was way too high than needed for the sputtering of the AZO, ZnO, and ZnO:Zn targets. As a result, the ZnO films prepared from the Zn, ZnO, and ZnO:Zn targets exhibited resistivity in the order of 14 kW.cm, while the ZnO film prepared from the AZO target had a resistivity about 610´10-4W.cm. It is known that the resistivity of ZnO thin films strongly depends on the oxygen concentration introduced in the sputtering deposition. Further work is underway to systematically study the variation of resisitivity of the ZnO films prepared from the ZnO:Zn target under different sputtering conditions.
Figure 10. Optical transmittance of ZnO films RF sputtering deposited using different targets. The RF sputtering power was 50 W and the process gas was 30% O2 mixed with Ar.
6) ZnO thin films via sol-gel method
A UV-vis transmission spectra of ZnO annealed at different temperature is shown in Figure 11. Four different ZnO coated film substrates were annealed at 200°C, 300°C 400°C and 500°C for 15 minutes each. Film annealed at 200°C showed the highest transmittance in 450 nm-750 nm range with an average transmittance above 90%. With increase in annealing temperature from 200°C to 500°C, a decreasing trend in optical transmittance was observed. Nevertheless, the transmittance for all ZnO coated films were greater than 85%. This reduction in transparency is attributed to roughening of the film at higher annealing temperatures.
Figure 11. Optical transmittance spectra of ZnO films annealed at different temperature.
Figure 12 shows the XRD pattern of the ZnO films annealed at different temperatures. The XRD spectra of ZnO film annealed at temperatures lower than 300°C clearly reveals an absence of a peak at 34.42 (002 plane). The absence of a peak indicates that the films are in amorphous state when annealed at a temperature lower than 300°C. An increase in peak intensity is observed when the films are annealed at temperatures higher than 300°C (i.e. at 400°C and 500°C in the figure). The gain in peak intensity indicates an increase in crystallinity with temperature.
Figure 12. XRD pattern of ZnO films at different annealing temperature.
The crystal growth of ZnO film in a specific plane (002) after thermal treatment indicates the highest density of Zn atoms in the 002 planes as such growth is kinetically favorable. As the treatment temperature is increased, excess energy acquired by ZnO crystallites allows them to orient themselves along the 002 plane where the surface energy is minimum.
Figure 13 shows the effect of annealing temperature on the crystal size of the ZnO. It was observed that the average crystal size was increased with increase in annealing temperature. The crystallite size may be estimated from the full-width at half-maximum (FWHM) of (002) peak using Scherrer’s formula:
Where K is 0.9, λ is the X-ray wavelength of 1.54 Å, θB is Bragg diffraction angle, and B is the FWHM of ZnO diffraction peak. Crystal size increased from 5.86 nm at 300°C to 23.33 nm at 500°C. The increase in thermal annealing temperature causes coalescence/merging of small grains by grain boundary diffusion resulting in major grain growth/ crystal size.
Figure 13. Effect of annealing temperature on the crystal size of ZnO films.
7) Alternative or complementary materials
ZnS and carbon allotropes where also investigated for the possibe integration on photovoltaic solar cells.
Figure 14. Nanocrystaline films tested to be used was protection film of the ZnO layer
8) Application on a-Si based solar cells
Optimized ZnO:Zn has been used to grow conductive and transparent back reflectors on a-Si based solar cells presenting good results
9) Disseminations actives
have been disseminated within the scientific photovoltaic community and
industry. Dissemination to the general public was also performed. Several
research and industrial contacts are now being established in order to
incorporate the finding in industrial processes.
The conductive ZnO:Zn composite target can be sputtered by use of DC or RF power supplies todeposit ZnO thin films. The target-poisoning problemis avoided by use of the ZnO:Zn composite target.Both DC and RF sputtering from the ZnO:Zntarget yield deposition rates higher than thoseobtained for ZnO films sputtered from metallic Zn,AZO, and pure ZnO targets. The optical transmittanceof RF sputtered ZnO films from the ZnO:Zntarget is comparable with that of films obtainedfrom other RF sputtered targets.
Published – Jyotshna Pokharel, Maheshwar Shrestha, Li Qin Zhou, Victor Neto* and Qi Hua Fan, “Oriented Zinc Oxide Nanocrystalline Thin Films Grown from Sol-Gel Solution”, Journal of Coating Science and Technology, 2 (2015) 46-50,https://dx.doi.org/10.6000/2369-3355.2015.02.02.2
Published – Li Qin Zhou, Qi Hua Fan, Raul Simões, Victor Neto, “High-rate sputtering deposition of high- and low-refractive index films from conductive composites”, MRS Communications, 5,2 (2015) 327-332. https://dx.doi.org/10.1557/mrc.2015.321
Presented – Li Qin Zhou, Qi Hua Fan, Raul Simões, Bárbara Gabriel, Victor Neto*, “High-rate Sputtering Deposition of High- and Low-refractive Index Films from Conductive Composites”, EU PVSEC 2015 – European PV Solar Energy Conference and Exhibition, 14-18 September 2015, Humburg, Germany
Presented – Barbara Gabriel,
Paula Marques, Victor Neto, “An analysis tool for decision making support on
nanomaterials applications - a preliminary case study”, TNT 2015 – Trends on
Nanotechnology, 7-12 September 2015, Toulouse, France (POSTER)
Presented – Li Qin Zhou, Raul Simões, Bárbara Gabriel, Qi Hua Fan, Victor Neto*, “Novel conductive ZnO:Zn composites for thin film solar cell back reflectors application”, TNT 2015 – Trends on Nanotechnology, 7-12 September 2015, Toulouse, France (POSTER)
Published – Raul Simões, V.F. Neto, “Nanodiamond Coated Glass as a Protective Layer in Solar Cells”, Special Issue of the 5th International Conference on Advanced Nanomaterials, Materials Today: Proceedings, 2 (2015) 230-235, https://dx.doi.org/10.1016/j.matpr.2015.04.027
Published – Li Qin Zhou, Mukul Dubey, Raul Simões, Qi Hua Fan, Victor Neto, Conductive ZnO:Zn Composites for High-Rate Sputtering Deposition of ZnO Thin Films, Journal of Electronic Materials, 44 (2015) 682-687. URL:http://link.springer.com/article/10.1007/s11664-014-3529-z
Presented – Raul Simões, Victor Neto, Nanodiamond Coated Glass as a Protective Layer in Solar Cells, ANM2014 – 5th International Conference on Advanced Nanomaterials, Aveiro, 2-4 July 2014. URL:http://www.researchgate.net/publication/263753708
Presented – R. K. Gupta, J. Hong, M. Dubey, X. Wang, N. Mandel, V. Neto, J. Gracio, Z. Gu, Q. Fan, Biochar Nanomaterials Activated by Oxygen Plasma for Energy Storage, ANM2014 – 5th Internantional Conference on Advanced Nanomaterials, Aveiro, 2-4 July 2014.
Presented – Raul Simões, Victor Neto, Application of nanocrystalline diamond films as protective coating in photovoltaic solar cells, CNME2014 – 9º Congresso Nacional de Mecânica Experimental, Aveiro, 15-17 de October de 2014. URL: http://www.researchgate.net/publication/265334765
Presented – Raul Simões, Victor Neto, Transparency of Different Diamond Films Prepared by Different Techniques, Diamond2014 International Conference on Diamond and Carbon Materials, Madrid, Spain, 7-11 September 2014. URL:http://www.researchgate.net/publication/265334757
Presented – Li Qin Zhou, Raul Simões, Qi Hua Fan , Victor Neto, Sputtered ZnO thin films from novel conductive ZnO:Zn composites, EVC13 – 13th European Vacuum Conference, Aveiro, 8-12 September 2014. URL:http://www.researchgate.net/publication/271210783
Published – Aiping Zeng, Victor F. Neto, Jose J. Grácio, Qi Hua Fan, “Diamond-like Carbon (DLC) Films as Electrochemical Electrodes”, Diamond and Related Materials 43 (2014) 12-22. URL:http://www.sciencedirect.com/science/article/pii/S0925963514000053
Published – Raul Simões, J.C. Mendes, L. Pereira, Victor Neto, "Diamond and other Carbon Related Materials Applications in Photovoltaic Solar Cells", IEEE International Electro-Information Technology Conference 2013, 9-11 May 2013, Rapid City, SD, EUA. URL: http://ieeexplore.ieee.org/xpl/articleDetails.jsp?arnumber=6632676
Presented – Raul Simões, J.C. Mendes, L. Pereira, V.F. Neto, "Diamond and other Carbon Related Materials Applications in Photovoltaic Solar Cells", IEEE EIT 2013 International Conference, May 9-11 2013, Rapid City, South Dakota, USA. URL: http://www.scopus.com/inward/record.url?eid=2-s2.0-84890108292&partnerID=MN8TOARS
Presented – Raul Simões, J.C. Mendes, L. Pereira, V.F. Neto, "Application of Diamond and other Carbon Related Materials Photovoltaic Solar Cells", International Conference on Diamond and Carbon Materials, 2-5 September 2013, Riva del Garda, Italia. URL:https://www.researchgate.net/publication/265334748