Journal Search Engine
Search Advanced Search Adode Reader(link)
Download PDF Export Citaion korean bibliography PMC previewer
ISSN : 1225-0562(Print)
ISSN : 2287-7258(Online)
Korean Journal of Materials Research Vol.27 No.2 pp.94-99

Properties of Dinickel-Silicides Counter Electrodes with Rapid Thermal Annealing

Kwangbae Kim, Yunyoung Noh, Ohsung Song†
Department of Materials Science and Engineering, University of Seoul 163, Seoulsiripdae-ro, Dongdaemum-gu, Seoul 02504, Republic of Korea
Corresponding author : (S. Oh, Univ. of Seoul)
November 9, 2016 December 22, 2016 December 22, 2016


Dinickel-silicide (Ni2Si)/glass was employed as a counter electrode for a dye-sensitized solar cell (DSSC) device. Ni2Si was formed by rapid thermal annealing (RTA) at 700 °C for 15 seconds of a 50 nm-Ni/50 nm-Si/glass structure. For comparison, Ni2Si on quartz was also prepared through conventional electric furnace annealing (CEA) at 800 °C for 30 minutes. XRD, XPS, and EDS line scanning of TEM were used to confirm the formation of Ni2Si. TEM and CV were employed to confirm the microstructure and catalytic activity. Photovoltaic properties were examined using a solar simulator and potentiostat. XRD, XPS, and EDS line scanning results showed that both CEA and RTA successfully led to tne formation of nano thick-Ni2Si phase. The catalytic activity of CEA-Ni2Si and RTA-Ni2Si with respect to Pt were 68 % and 56 %. Energy conversion efficiencies (ECEs) of DSSCs with CEA-Ni2Si and RTA-Ni2Si catalysts were 3.66 % and 3.16 %, respectively. Our results imply that nano-thick Ni2Si may be used to replace Pt as a reduction catalytic layer for a DSSCs. Moreover, we show that nanothick Ni2Si can be made available on a low-cost glass substrate via the RTA process.


    © Materials Research Society of Korea. All rights reserved.

    This is an Open-Access article distributed under the terms of the Creative Commons Attribution Non-Commercial License ( which permits unrestricted non-commercial use, distribution, and reproduction in any medium, provided the original work is properly cited.


    Typical dye-sensitized solar cells (DSSCs) are composed of working electrode (WE), dye, electrolyte, and counter electrode (CE).1,2) For the catalytic layer of CE, Pt has excellent electrical conductivity due to low resistivity (0.1 × 10-6 Ω·cm), high reflectance (visible light of 380~ 780 nm > 99.9 %), and high work function (5.65 eV). However, Pt costs about 320 times more than a nickel, which is one of common metals, requires high heat treatment when processed by spin coating, and yields a lower energy conversion efficiency (ECE) because the dye adsorbed on WE degrades Pt and reduces the catalytic activity.3-5) In order to solve such problems, new materials including platinum-group, carbon-group, and intermetallic compound materials to replace conventional Pt are widely studied.

    First of all, research that using platinum-group metals, such as Ru6) and Ir,7) found that those metals are so expensive and provide 10~30 % less catalytic activity than Pt of the same thickness.

    For a research employing carbon-group, D. W. Zhang et al.8) employed graphene nanosheet and reported a relatively low ECE of 6.81 %, while the ECE of the device employing conventional Pt was observed at 7.59 %. Another research carried out by S. Widodo et al.9) reported the potential of the CNT as a catalyst material by reporting the ECE of 1.98 % while that of employing conventional Pt was 4.12 %.

    For researches using intermetallic compounds, J. Jia et al.10) used selenide-based Ni0.85Se as CE and reported the ECE of 8.88 %, which was 1.09 times improvement compared to the one employing conventional Pt catalyst, and W. Zhou et al.11) used sulfide-based NiS as CE and reported the ECE of 7.39 %, which was 1.04 times improvement compared to the one employing conventional Pt catalyst. However, the drawback of such S or Se intermetallic compounds is that they are less durable than Pt.

    Recently K. Kim et al.12) reported the possibility of employing nickel-silicide and cobalt-silicide, which are widely used in complementary metal-oxide-semiconductors, as DSSC device catalysts. However, silicide phase had the problem of having to use expensive quartz substrate instead of cheap glass substrate because of relatively high processing temperature.

    Ni2Si, one of representative nickel silicides, is a white metal with an excellent durability and resistivity (~24 × 10-6 Ω·cm). It is widely used as an electrode material for semiconductors, and its processes are relatively wellknown. Also, the costs of Si and Ni are 1/322 and 1/214, respectively, in relative to that of Pt. Therefore, low production cost of Ni2Si offers an economical option.

    Hence, rapid thermal annealing (RTA) might be suitable to implement such nickel-silicides on glass substrates. In semiconductor fabrication processes, RTA is mostly adopted for salicide (self-aligned silicide) processes to form an ohmic contact intended to reduce contact resistance. This method shows little impact on the substrate because the heat treatment time needed is very short and its thermal energy is concentrated only at the top layer.13)

    The heat treatment by conventional electric furnace annealing (CEA) does not allow using glass substrate that cannot endure 800 °C-30min. If RTA process is applied to form a catalyst of a DSSC device, heat treatment can be finished within dozens of seconds, thus a glass substrate with a relatively low cost can be used instead of expensive quartz substrate intended for a high temperature process.

    In the study, we introduce a layered nano-thick Ni/Si/ glass structure as pre-form to form nickel-silicide on a glass substrate. Then RTA was applied to transform into nano-thick Ni2Si/glass and its suitability as DSSC catalyst material was confirmed.

    2.Experimental Procedure

    To employ soda-lime glass substrate and Ni2Si as a counter electrode, Ni/Si/glass was fabricated by using Egun evaporator (ZZZ550-2/D, Maestech), RTA (KVRTP- 020, Korea vacuum tech). To form Ni2Si, we used RTA for the heat treatment of 50 nm-Ni /50 nm-Si /glass structured specimens at 700 °C for 15 seconds. For comparison, we coated 50 nm-Ni/50 nm-Si on a quartz substrate through the same procedure as above and formed Ni2Si by using CEA for the annealing at 800 °C for 30 minutes. The process was followed by 20 minutes etching in 30 % H2SO4 to remove the nickel residue on the nickel silicide surface. Also for a comparison against Pt, the conventional DSSC catalyst, we prepared a sample with 100 nm-thick Pt film deposited on a flat glass substrate.

    We used the X-ray diffraction (XRD, X’pert-pro, PANalytical) to observe the properties of the produced nano-thick Ni2Si film.12) We used the X-ray spectroscopy (XPS, ThermoVG, U.K.) analysis to binding energy of the properties of the produced nano-thick Ni2Si film.14)

    To examine the microstructure of the as-prepared nanothick nickel silicide with a transmission electron microscope (TEM), a focused ion beam (FIB, NB 5000, Hitachi) was applied to prepare cross-sectional samples. Using TEM (HF-3300, Hitachi), we verified the crosssectional structure of Ni2Si in the dark field mode and investigated its crystallinity based on selected area diffraction pattern (SADP) analysis. In addition, using energy dispersive spectroscopy (EDS) line scanning attached to the TEM, we examined the elemental ratio of Ni and Si. The experimental conditions were set to a beam size of 5 nm, camera length of 30 cm, and wavelength of 0.00197 nm.

    We employed DY2113 potentiostat (Digi-Ivy, Austin, TX) for cyclic voltammetry (CV) measurement in order to analyze the catalytic activity.12) Later, we produced glass /FTO /300 nm-TiO2 blocking layer /8 μm-TiO2 /dye /electrolyte /Pt or Ni2Si /glass or quartz structured DSSC devices with the active area of 0.45 cm2. To measure impedance with the interface resistance and current-voltage (I-V) analysis, solar simulator (PEC-L11, Peccell) and potentiostat (Iviumstat, Ivium) were used.12) The impedance analysis confirmed the resistances of Rh, R1, R2, and R3. The I-V analysis determined short circuit current density (Jsc), open circuit voltage (Voc), fill factor (FF), and ECE.

    3.Results and Discussion

    Fig. 1 shows the XRD results of Ni2Si thin films produced by the RTA and CEA. In case of CEA, Ni2Si peak corresponding to (121), (110) plane direction at 45.1°, 47.3° confirmed the formation of Ni2Si, and SiO2 peak of quartz substrate that corresponds to (200), (201) phase appeared at 42.1°, 45.9°. In case of RTA, Ni2Si peak corresponding to (121), (002) plane direction at 45.1°, 48.5° confirmed the formation of Ni2Si, and NiSi peak that corresponds to (202) phase appeared at 47.1°. The existence of this NiSi peak in the RTA sample suggests a possibility that both Ni2Si and NiSi phases exist simultaneously in contrast to the CEA sample, which contains only Ni2Si.

    Fig. 2 presents the XPS analysis results of Ni2Si manufactured using CEA and RTA. We verified that both CEA-Ni2Si and RTA-Ni2Si show distinct Ni 3p3/2, Ni 2p3/2, and Ni 2p1/2 peaks at 66.9 eV, 853.4 eV, and 870.6 eV, respectively. This is consistent with the XPS results of Ni2Si reported by Y. Cao et al.14) However, unlike in the case of the aforementioned XRD results, the existence of a NiSi phase in the RTA sample could not be verified by XPS analysis. It implies that NiSi may be formed much thinner than Ni2Si. Nevertheless, this analysis verified that Ni2Si was successfully formed using both RTA and CEA, which is in agreement with the previous XRD results.

    Fig. 3 displays the TEM cross-sectional images of the layered Ni2Si materials manufactured via CEA and RTA. The inset shows the corresponding SADP results. Fig. 3(a) exhibits the TEM cross-sectional image of Ni2Si/ quartz manufactured by the CEA process. From this image, it can be seen that the Ni2Si thickness is about 64 nm. Furthermore, the SADP displays regular peaks, confirming that a crystalline Ni2Si phase has been generated. This thickness agreed well with the theoretical thickness that appeared when the Ni2Si was formed from the resources of 50 nm-Ni and 50 nm-Si.

    Fig. 3(b) displays the TEM image of Ni2Si/glass manufactured using RTA. The thickness of the generated thin film was about 63 nm, which is similar to the film thickness of the CEA sample displayed in Fig. 3(a). As opposed to the CEA sample, a clearly observable contrast con-firms the existence of an additional phase on the surface of the RTA sample, which could be attributed to the NiSi phase formed on the surface, as indicated by the XRD results presented in Fig. 2. As this additional phase is extremely thin (about 5 nm, we confirmed that the most of the silicide layer with RTA is Ni2Si phase. Furthermore, all the SADP images display regular peaks verifying that crystalline Ni2Si had been formed.

    Fig. 4 displays the EDS line scanning data of the Ni2Si cross-section samples manufactured via CEA and RTA, respectively. Fig. 4(a) shows the EDS line scanning results of Ni2Si manufactured by CEA. It can be seen that the carbon coating employed as a protective layer exhibits a thickness of around 0.02 μm while the overall ratio of Ni to Si was found to be 2:1 confirming the formation of Ni2Si. Furthermore, the thickness of the formed Ni2Si was approximately 64 nm, which is consistent with the thickness displayed in the TEM image above (Fig. 3(a)).

    Fig. 4(b) exhibits the EDS line scanning results of Ni2Si manufactured by RTA. Similar to the observation in Fig. 4(a), a carbon layer can be identified with a thickness of about 0.02 μm. It is worth noting that this sample exhibits a Ni to Si ratio of 1:1 on the surface, which confirms the existence of NiSi in accordance with the TEM results of Fig. 3(b). Furthermore, the overall ratio of Ni to Si was also found to be 2:1, which confirms the formation Ni2Si. The thicknesses of the formed NiSi and Ni2Si layers were about 5 nm and 58 nm, respectively, which also consistent with the aforementioned TEM results (Fig. 3(b)).

    In other words, we verified the existence of a mixed layer mainly consisting of Ni2Si while NiSi was found in some parts of the surface of the RTA sample. In summary, XRD, XPS, and TEM results confirmed that Ni2Si was formed during heat-treatment of layered nano-thick Ni/Si/substrate structures at a temperature above 700 °C. These results are different from a mechanism that when the existing (100) silicon substrate to form Ni2Si, it was formed at a lower temperature of 400 °C.15) Thus, unlike traditional silicide processes that apply silicon substrates, the process introduced here employs individual silicon and nickel layers treated at a higher temperature to form Ni2Si. Therefore, we verified that was a phase that could not be acquired through CEA at a temperature lower than 500 °C, for which the glass substrate could be used.

    Fig. 5 is CV data of Ni2Si formed by the RTA and CEA, and Pt as CE. The lower left corner represents the reduction reaction of 3I2+ 2e → 2I3 and I3+ 2e → 3I from iodine. The catalytic reaction at DSSC is related to reduction reaction and can be proportional by area at the lower left corner. Ni2Si formed by the CEA-Ni2Si and RTA-Ni2Si showed the catalytic activity that is about 68 % and 56 % of the one formed by employing Pt for the same area, respectively. Therefore, both Ni2Si produced by using RTA-Ni2Si and CEA-Ni2Si can similarly function as catalysts for DSSC.

    Fig. 6 shows the Nyquist plot consisted of real part (Z') and imaginary part (Z") obtained at an applied frequency for DSSC devices employing Ni2Si formed by the CEANi2Si and RTA-Ni2Si, and Pt as CE. Each resistance of Rh, R1, R2, R3 that appear on the Nyquist plot of typical DSSC devices are shown at the left-upper inset.16) For instance, Rh value represents the real part resistance, which is dependent on sheet resistance of TCO substrate used for WE and sheet resistance of CE, and the resistance of the TCO/electrolyte interface.

    The Rh values of the DSSC devices employing CEANi2Si and RTA-Ni2Si, respectively, were found to be higher than that of the device employing Pt. These results can be attributed to the resistivity of Ni2Si and Pt of 24 × 10-6 Ω·cm and 0.1 × 10-6 Ω·cm, respectively.

    The R1 value is the interface resistance in the 103-105 Hz region that is related to the TCO/TiO2 interface resistance and the electrolyte/CE interface charge transfer resistance. CEA-Ni2Si’s R1 value was found to be higher than that of RTA-Ni2Si. As examined in the previous EDS line scanning results (Fig. 4), we believe that the existence of NiSi affected the catalytic activity of the RTA-Ni2Si sample.

    The R2 value represents the electron movement resistance and electron recombination resistance in WE in the 1-103 Hz region while the R3 value denotes the Warburg impedance of the electrolyte over 106 Hz. Since identical WEs and electrolytes were used in these experiments, the R2 and R3 values are consequently identical.

    Fig. 7 is the I-V data of DSSC devices employing CEA-Ni2Si, RTA-Ni2Si, and Pt as CE. Jsc values were found to be high in the order of Pt, CEA-Ni2Si, and RTA-Ni2Si, which indicates that CEA-Ni2Si provides improved performance compared to RTA-Ni2Si. For Voc values, all CEA-Ni2Si, RTA-Ni2Si, and Pt were close to each other. In case of FF, CEA-Ni2Si and RTA-Ni2Si showed the value lower than Pt. But RTA-Ni2Si was smaller than CEA-Ni2Si. Therefore, we confirmed the possibility that Ni2Si can replace Pt as a DSSC catalyst. Furthermore, we provided evidence that RTA-Ni2Si manufactured on a glass substrate can provide similar levels of Jsc and FF compared to Ni2Si/quartz materials, demonstrating its suitability to be employed as a DSSC CE catalyst material.

    Table 1 shows the I-V result of Fig. 7 in detail. Jsc of DSSC device employing, CEA-Ni2Si, RTA-Ni2Si and Pt were recorded as 10.43 mA/cm2, 10.03 mA/cm2, and 11.45 mA/cm2 respectively. The device using CEA-Ni2Si showed improved Jsc values than the device employing RTANi2Si. The reason for this could be the fact that the existence of NiSi on the surface affected the catalytic activity of the RTA-Ni2Si sample.

    Voc showed similar values of 0.67V within the error range, and the similar values of Voc considered to be attributable to the use of the same electrolyte in all Fermi levels as an element related to the oxidation-reduction reaction of the electrolyte.

    FF, which is a factor influenced by the interface resistance of the device, was 0.53 and 0.47 for employing CEA-Ni2Si and RTA-Ni2Si, respectively. As with the previous impedance result, it confirmed to the change of R1 value. The final ECEs obtained by employing CEANi2Si, RTA-Ni2Si, and Pt were 3.66 %, 3.16 %, and 5.05 %, respectively, and this result was due to the change of Jsc and FF.

    Therefore, regardless of whether the CEA or RTA process is employed, the nano-thick of Ni2Si showed catalytic properties to the extent that it could be employed as a DSSC CE layer. Even when using RTA and glass substrates, our results verified that Ni2Si can be applied as a catalyst, even though it showed slightly reduced ECE because of the very thin NiSi layer formed on the surface.


    To employ a glass substrate and Ni2Si as a CE for a DSSC device, we formed Ni2Si /glass using 50 nm-Ni/ 50 nm-Si/glass in the RTA process. For comparison, we used the CEA method and manufactured Ni2Si/quartz as a reference material. From the phase and microstructure analysis results, all samples manufactured via RTA or CEA successfully formed 60 nm Ni2Si thin films. However, when RTA was employed, a 5 nm NiSi layer was formed on the surface. The final ECE results based on the catalytic activity of CEA-Ni2Si and RTA-Ni2Si were found to be 3.66 % and 3.16 %, respectively. We believe that the reduced catalytic activity and ECE of RTA-Ni2Si could be attributed to the NiSi layer generated on the surface. In conclusion, this study confirmed that nanothick Ni2Si could be utilized as counter electrode material for a DSSC. Furthermore, that the RTA process allows the employment of glass substrates to economically manufacture Ni2Si thin films.



    XRD peaks of dinickel-silicides with CEA and RTA.


    XPS peaks of Ni2Si counter electrodes with CEA and RTA.


    TEM datas of (a) CEA and (b) RTA. Insets are SADP images.


    EDS line scanning data of (a) CEA and (b) RTA.


    CV curves of using the electrolytes of MPN solutions measured at a scan rate of 50 mV/s.


    Nyquist plots of CEA-Ni2Si, RTA-Ni2Si, and Pt counter electrodes employed DSSCs. Inset is Nyquist plot with specified parameters.


    Current-voltage (I-V) characteristic of CEA-Ni2Si, RTANi2Si, and Pt counter electrodes employed DSSCs.


    Photovoltaic performance and energy conversion efficiency of DSSC.


    1. O’ Regan B. , Gratzel M. (1991) Nature, Vol.353 ; pp.737
    2. Longo C. , Paoli M.A. (2003) J. Braz. Chem. Soc, Vol.14 ; pp.889
    3. Kay A. , Gratzel M. (1996) Sol. Energy Mater. Sol. Cells, Vol.44 ; pp.99
    4. Li Z. , Ye B. , Hu X. , Ma X. , Zhang X. , Deng Y. (2009) Electrochem. Commun, Vol.11 ; pp.1768
    5. Olsen E. , Hagen G. , Lindquist S.E. (2000) Sol. Energy Mater. Sol. Cells, Vol.63 ; pp.267
    6. Noh Y. , Yu B. , Yoo K. , Ko M. , Song O. (2012) Korean J. Met, Mater, Vol.50 ; pp.243
    7. Mokurala K. , Kamble A. , Bhargava P. , Mallick S. (2015) J. Electron. Mater, Vol.44 ; pp.4400
    8. Zhang D.W. , Li X.D. , Li H.B. , Chen S. , Sun Z. , Yin X.J. , Huang S.M. (2011) Carbon, Vol.49 ; pp.5382
    9. Widodo S. , Wiranto G. , Hidayat M.N. (2015) Energy Procedia, Vol.68 ; pp.37
    10. Jia J. , Wu J. , Tu Y. , Huo J. , Zheng M. , Lin J. (2015) J. Alloys Compd, Vol.640 ; pp.29
    11. Zhou W. , Jia X. , Chen L. , Yin Z. , Zhang Z. , Gao G. (2016) Mater. Lett, Vol.163 ; pp.1
    12. Kim K. , Noh Y. , Choi M. , Song O. (2016) Korean J. Met Mater, Vol.54 ; pp.615
    13. Kim H.S. , Lee J.H. , Yeom G.Y. , Korean J. (1995) Inst. Electr. Electron. Mater. Eng, Vol.8 ; pp.619
    14. Cao Y. , Nyborg L. , Jelvestam U. (2009) Surf. Interface Anal, Vol.41 ; pp.471
    15. Colgan E.G. , Maenpaa M. , Finetti M. , Nicolet M.A. (1983) J. Electron. Mater, Vol.12 ; pp.413
    16. Santigo F.F. , Bisquert J. , Palomares E. , Oterio L. , Kuang D. , Zakeeruddin M. , Gratzel M. (2007) J. Phys. Chem. C, Vol.111 ; pp.6550