TiO2 paste for DSSC photoanode: preparation and optimization of application method 140 D O I: 1 0. 15 82 6/ ch im te ch .2 02 0. 7. 4. 01 Selyanin I. O., Steparuk A. S., Irgashev R. A., Mekhaev A. V., Rusinov G. L., Vorokh A. S. Chimica Techno Acta. 2020. Vol. 7, no. 4. P. 140–149. ISSN 2409–5613 TiO2 paste for DSSC photoanode: preparation and optimization of application method I. O. Selyanin ab*, A. S. Steparukc, R. A. Irgashevbс, A. V. Mekhaevc, G. L. Rusinovbс, A. S. Vorokhab a Institute of Solid State Chemistry, Ural Branch of the Russian Academy of Sciences, 91 Pervomaiskaya St., Ekaterinburg 620990, Russia b Ural Federal University, 19 Mira St., 620002, Ekaterinburg, Russia с Postovsky Institute of Organic Synthesis, Ural Branch of the Russian Academy of Sciences, S. Kovalevskoi/Akademicheskaya 22/20 St., 620137, Ekaterinburg, Russia *email: theselyanin@yandex.ru Abstract. We propose a simple method of TiO2 paste preparation from titania pow- der (Degussa) and organic binders (terpineol, ethyl cellulose) for making a continuous photoactive layer of a dye-sensitized solar cell (DSSC). The prepared paste was charac- terized by using thermogravimetric and X-ray diffraction methods for comparison with commercial paste (Solaronix). The TiO2 layer parameters for applying and annealing were optimized by varying the layer thickness and using different masks. The surface morphology of annealed layers was controlled by optical microscopy. Before TiO2 paste applying and after annealing, the conductive glass (fluorine-tin oxide — FTO) was treated by TiCl4 hydrochloric acid solution. The structure of DSSCs were composed FTO- glass / TiO2 layer sensitized Ruthenium complex (N719 dye)/ iodide-based electrolyte / Pt counter electrode/ FTO glass. The DSSC photovoltaic characteristics were measured under AM 1.5G irradiation and demonstrated to be close to those of photoanodes based on the prepared and commercial pastes. Keywords: DSSC; Gratzel cell; TiO2 paste; photoanode; open-circuit photovoltage; short-circuit photocurrent density; I–V characteristics Received: 27.03.2020. Accepted: 20.10.2020. Published:30.12.2020. © Selyanin I. O., Steparuk A. S., Irgashev R. A., Mekhaev A. V., Rusinov G. L., Vorokh A. S., 2020 Introduction Over the  past two decades, DSSCs (dye-sensitized solar cells or Grat- zel cells) have become widely known due to  their environmental friendliness and low cost of  manufacture [1]. Improv- ing the efficiency of the solar irradiation conversion into electrical energy can be reached by modifying every individual cell element. A photoactive anode layer based on wide-band semiconductor oxides (usu- ally, nanocrystalline TiO2) plays the most important role [2, 3]. Titania matrix can be sensitized by organic molecules or quan- tum dots to  allow the  efficient injection of  light-excited electrons from dyes [4]. The combination of TiO2 phases (anatase, 141 rutile, brookite, and amorphous) improves the  conducting and catalytic properties of  anode material. For commercial and laboratory investigation of  photoanode, a crystalline phase of anatase or a mixture of anatase and rutile phases (for example, Degussa P25 with the ratio of anatase:rutile = 3:1) are used as sources of TiO2 for paste making [5, 6]. Together with the TiO2 phase compo- sition and electronic structure, the layer’s geometrical and morphological character- istics play an important role in electricity generation. In particular, the photoanode TiO2 layer thickness affects the photovol- taic cell characteristics directly: the thick oxide layer is  prone to  cracking after shrinkage [6]. The cracks impede the elec- tron transfer, provoke the short circuit, and reduce the area of the photoactive surface. A  balance between oxide layer continu- ity and thickness should be found to op- timize the  DSSC operation. When TiO2 nanocrystalline layer thickness increases from 0.5 to 2 μm, the solar cell character- istics behave as follows: the short-circuit current density increases by a third, and the open-circuit voltage increases by 5% [7]. The dependence of photoelectrochemi- cal characteristics on TiO2 layer thickness (6, 10, 14 μm) demonstrates that the opti- mal layer thickness is 10 μm for good light absorption, minimal charge recombina- tion, and the low resistance to charge trans- fer [8]. The short-circuit current and con- version efficiency decrease when the TiO2 layer thickness exceeds 15  μm [9–12]. The  good anode layer continuity and its certain thickness are necessary to obtain the high efficiency of DSSC. The purpose of this work is to develop methods for producing a paste based on TiO2 powder and forming a  continuous layer of optimal thickness on a conductive substrate. The phase and chemical compo- sition of commercial and prepared pastes were studied by X-ray diffraction and ther- mogravimetric analysis. The  search for optimal parameters of heat treatment and formation of the layer of a given thickness was carried out by studying the behavior (preservation of  continuity or cracking) of commercial paste on laboratory glass. The obtained parameters were used to cre- ate DSSC anodes, the photoactive layers of which are formed from both commercial and prepared pastes. A  dye N719 based on a ruthenium complex was adsorbed on the TiO2 layer. The photovoltaic properties of the collected DSSC were studied under the conditions of illumination of AM 1.5G by a xenon light source. Experimental Paste preparation and characterization. For the  paste preparation, 1.0  g of  TiO2 (Degussa P25, Sigma-Aldrich), 3.8  g of  α-terpineol (70% PS, Panreac), 0.5  g of ethyl cellulose (48.0–49.5% (w/w) eth- oxyl basis, Sigma-Aldrich), 0.15 g of nitric acid (65%, Vecton) in 20 ml of ethyl alco- hol were mixed. The mixture was stirred for 30  minutes at  room temperature. The titania-containing paste was obtained by  distilling excess volatile components on a Heidolph Hey-VAP Precision rotary evaporator at  a  water bath temperature of 60 °C and reduced pressure (175 mbar) for 30  minutes. Commercial paste Ti- Nanoxide D/SP (Solaronix SA) was used as  a  reference for comparison of  viscos- ity and homogeneity. The standard doctor blade method was used for applying pastes in all cases. The area of the applied layer was 2.0 × 2.0 cm. 142 The morphology of samples with pastes was studied using a  Carl Zeiss Neophot 30 optical microscope equipped with a  digital camera EverFocus EQ500A/P- IR. X-ray diffraction (XRD) studies were performed on the  Shimadzu MAXima- X XRD-7000 X-ray diffractometer (with Cu Kα radiation, λ=1.5406 Å) in the ge- ometry of  Bragg-Brentano with a  step of 0.03° at the angle of 2θ and an exposure time of 3 seconds. Particle size was deter- mined by  the  Igor Pro Multipeak2 pro- gram using the Scherrer formula D = Kλ/ (FWHM·cosθ), where K is the form factor, λ is the X-ray wavelength, FWHM is full width at half maximum and θ is the Bragg angle corresponding to  reflection. Ther- mogravimetric studies of commercial and prepared pastes based on titanium dioxide were performed on a Mettler Toledo TGA / DSC 1 device in the air (30 ml/min) with a heating rate of 10 °C/min in the range of 30–500 °C. Optimization of  annealing and paste application. Glass substrates were kept in  a  sulfochromic mixture for 24  hours, after ward rinsed with distilled wa- ter and isopropyl alcohol and dried at  100  °C.  To  regulate the  thickness of the paste layer, masks made of various materials were used, namely: adhesive tape (hereinafter  — a.t., 35  μm), aluminum foil (9 and 14 μm), screen printing mesh Sefar E-MESH 61/156-64W PW (thread thickness 64  μm, cell size 100×100  μm). The paste application onto the cover glass was made using a doctor blade technique. Samples with the paste were dried at 120 °C for 5  minutes, then annealed in  a  muf- fle furnace (SNOL with OVEN TRM-10 regulator) in the air at temperatures from 350 to 450 °C for 30 minutes (the heating rate was 5  °C/min). The  study of  the  ef- fect of 0.30 g nitric and perchloric acids on the continuity of the formed layer was carried out by adding HNO3 (65%, Vecton, >90% Sigma-Aldrich) or HClO4 (65%, Vec- ton) to the commercial paste. The  formation of  photoanodes. FTO (fluorine-doped tin oxide) conductive glasses (size 2.5 × 2.5 cm, surface resistiv- ity ~ 8 Ω/cm2, Sigma-Aldrich) were treated with isopropyl alcohol for 30 minutes and then dried at  50  °C.  Then FTO glasses were treated by holding in the TiCl4 hy- drochloric acid solution (≥ 98.0%, Fluka) for 30  minutes at  a  solution tempera- ture of  70  °C, followed by  washing with isopropyl alcohol and distilled water. After applying the  oxide paste, the  sam- ples were annealed in  a  muffle furnace for 30  minutes at  450  °C.  The  resulting layer was re-treated by  holding in  TiCl4 hydrochloric acid solution according to the method described above. Adsorp- tion of the dye N719 [di-tetrabutylammo- nium cis-bis(isothiocyanato)-bis(2,2-bi- pyridyl-4,4-dicarboxylate)ruthenium(II)] (Solaronix SA) onto the  TiO2 layer was carried out by holding the layer in a 5·10–4 M methanol solution of N719 for 18 hours. The cell elements and assembly. To create a counter electrode, a thin layer of platinum was applied to the FTO glass using the fol- lowing method: a  few drops of  H2PtCl6 solution in ethanol were distributed over the  surface and annealed at  450  °C for 30 minutes. A mixture of 0.5 M LiI (99%, Sigma-Aldrich) + 0.05 M I2 (sublimated for analysis, Sigma-Aldrich) in acetonitrile (99.9%, NPO Reaktivy OSC) or 3-meth- oxypropionitrile (≥98.0%, Sigma-Aldrich) was used as the electrolyte. The assembly of the elements was performed as follows: the Pt cathode was placed on top of the pho- toanode side soaked TiO2 layer, then a few drops of electrolyte were added in the gap 143 between the two glasses, and then the glass- es were stuck together using office clips. Measurement of  I–V characteristics. The assembled cell was illuminated with a  xenon source Zolix Gloria-X500A (the  illumination power at  a  distance of  25 cm is  100 mW/cm2, determined using a  silicon calibration element Zo- lix QE-B1). The  source radiation spec- trum was in the 250–1000 nm band with the  maximum intensity in  the  range of 350 to 550 nm. The illuminated surface of  the  photoanode was 4 cm2. Lighten- ing was performed from the photoanode side with the adsorbed dye. The I–V char- acteristic was recorded using a  Keithley 2450 source-meter with a 50 mV step with a measurement error of ±10 nA for current and ±1 μV for voltage. Results and discussion The prepared paste characterization and comparison with the  commercial analog. The paste was made of TiO2 De- gussa P25 with the addition of α-terpineol, ethyl cellulose, and nitric acid (hereinafter “Prepared”). Three stages are distinguished on thermogravimetric curves (Fig.  1): the evaporation of volatile components (1st stage), oxidation of α-terpineol (2nd stage, Tb.p. = 215 °C), and the combustion of cellu- lose derivatives (3rd stage, Tautoignit. = 370 °C). After annealing, with the organic compo- nents removed, the remaining weight (pure TiO2) was found to be ~15% of the initial one. The result obtained for the laborato- ry paste is comparable to the commercial paste Ti-Nanoxide D/SP (Solaronix SA). The samples annealed at temperatures of 350–400 °C have a brownish tinge asso- ciated with incomplete burning of organic components. This result is consistent with the TGA data. Based on the obtained data, a temperature of 450 °C was used for fur- ther research. It is evident from Table 1 that annealing leads to cracking of the result- ing layer, which can be avoided by varying the thickness of the applied layer. According to  XRD, the  composi- tion of  the  obtained layers at  450  °C is  nanocrystalline TiO2 powder (Fig.  2). The  combustion of  organic components defined from thermogravimetric data is  proved by  the  absence of  any reflec- tions or amorphous halo onto the diffrac- Fig. 1. Thermogravimetric curves of prepared and commercial pastes 144 tion pattern. The  powder obtained from the  annealed commercial paste consists of anatase nanoparticles with an average particle size of 20.7 nm. The prepared paste is based on TiO2-P25 and consists of a mix- ture of anatase (92 mass %) and rutile (8 mass %) phases. The size of anatase parti- cles in TiO2-P25 was 19.8 nm, and the size of rutile particles was 27.9 nm. We proposed that titania nanoparticles can be ordered along chains of organic mol- ecules (ethyl cellulose) and can form aniso- tropic agglomerates during heat treatment. This process can provoke the texturizing of the TiO2 layer. In support of this proposi- tion, annealing was performed directly on the glass substrate used in the XRD experi- ment. The peak intensities of both layer and powder are the same as the standard peak intensities defined by PDF cards (21–1272 for anatase, and 21–1276 for rutile). The dif- ference curve between the layer and powder diffraction patterns shows a lack of crys- tallographic orientations of the layer. We can conclude that the  procedure of  heat treatment allows decomposing organic compounds and obtaining the titania layer with a random orientation of nanostruc- tured grains. An amorphous halo occurs for both the prepared powder pattern and the difference curve because of the very thin oxide layer and the reflection from the glass substrate. Fig. 2. Diffraction patterns of commercial and prepared TiO2 layers and initial TiO2 powder Table 1 Microphotographs of the TiO2 layer surface, the insets show layer surfaces on the glass substrate 145 Effect of various masks on the thick- ness of  the  TiO2 layer. The  commercial paste was used as a reference for searching the optimal method of applying the paste to  form a  continuous TiO2 layer. This paste has a uniform small particle size and a verified ratio between the components of the paste: TiO2, thickener, and binder. To vary the thickness of the resulting layer on glass substrates, various combinations of  a.t., aluminum foil of  different thick- nesses, and Sefar screen printing mesh were used. A.t. and foil were used as masks that were applied to the glass. The squeegee was used to apply the paste evenly. The screen mesh was applied on top of  a  glass or a mask, a layer was applied on top of a mask, and then the mesh was removed. The sam- ples were exposed to temperature treatment corresponding to  the  “annealing” stage discussed in the previous section. Various “mask-mesh” combinations for applying TiO2 paste to glass substrates and corre- sponding microscopic images of samples are shown in Table 2 (sections I–III). A n a ly s i s of   m i c rophoto g r aphs of the obtained layers showed that the best results are obtained when using one layer of a.t. in the case of using the only a.t. It corresponds to the thickness of the applied paste of  ~ 35  μm. In  this case, the  layer remains continuous; almost no cracks emerge (Table  2, section I). A  compara- tive study by optical microscopy showed that the layer thickness after annealing was ~10 μm. The experiment with using more than one layer of the adhesive tape showed that the  increase in  the  layer’s thickness results in intense cracking of the TiO2 layer after annealing. It negatively affects the ef- ficiency of the cell. Foil without mesh does not allow getting a  layer of  appropriate quality, the combination of foil and mesh produces an effect similar to the combina- tion of mesh and a.t. (Table 2, section II). Besides, the  foil deforms easily when it overlays the glass. This suggests that the use of a foil mask is problematic. Applying two layers of paste with intermediate drying, in contrast, demonstrates good continuity (Table 2, section III). To  test the  hypothesis that cracking is  primarily caused by  uneven evapora- tion of the liquid phase and the formation of bubbles, an experiment with the addi- tion of an oxidizer to a commercial paste to accelerate the oxidation process of or- ganic components was performed. HNO3 and HClO4 acids were used as  oxidants. Based on microphotographs of photoanode surfaces, it can be concluded that the layers have the best appearance after adding 65% HNO3, and adding HClO4, on the contrary, leads to greater cracking of the photoanode layer (Table 2-IV). In general, adding 65% and 90% HNO3 oxidizer reduces the crack- ing of the oxide layer. Having established the optimal mode of application, the samples were obtained with the prepared paste with combinations of  masks presented in  Table  2-V.  From the presented images it is clear that the lab- oratory paste during the  application has good continuity. The  fabrication of  the  photoanodes and the assembly of Gratzel cells. Based on the above experiments, the application and annealing conditions were selected for the preparation of photoanodes. In all cas- es, the commercial and the prepared pastes were applied through a mesh to a conduc- tive FTO glass. To improve the continuity of the layer, preliminary (before applying the paste to the glass) and final (after ap- plying and annealing the paste) treatments with TiCl4 hydrochloric acid solution were performed. 146 Strong adhesion of oxide layers plays an  important role when they have been immersed in a dye solution for a long time. In the presence of microcracks and poor adhesion, even in the presence of distilled water, it is possible to observe the peeling of the oxide layer. For this reason, an ad- ditional experiment was performed to test the adhesion properties. The prepared pho- toanode was immersed in distilled water for 24 hours, after which there was no de- lamination, the layer showed good resist- ance even under the mechanical influence. According to optical microscopy data, the  morphological differences between the layers of commercial and the prepared Table 2 Micrographs of the surface of TiO2 layers under various conditions for applying commercial paste (I–III) and its modification (IV), as well as methods for applying laboratory paste (V). The annealing temperature is 450 °C, the scale is shown on the upper-left microphotograph I 1 layer a.t. 2 layer a.t. 3 layer a.t. 4 layer a.t. II 9 μm foil 14 μm foil 9 μm foil– mesh 14 μm foil — mesh III mesh 1 layer a.t. — mesh 2 layer a.t. — mesh 1+1 layer a.t. — mesh* IV without oxidizer HNO3 (~65%) HNO3 (~90%) HClO4 V 1 layer a.t. mesh 1 layer a.t. — mesh TiCl4 treatment * 1 layer of a.t. — mesh, then drying at ~120 °C for 5 minutes, adding the 2nd layer of a.t, applying the 2nd layer of paste through the grid, followed by drying 147 pastes are minor in comparison with other layers shown in Table 2. Both layers have good continuity, but in the case of prepared paste (Fig. 3a) there are cracks and slight heterogeneity caused by  the  aggregation of TiO2 nanoparticles in the initial powder, while for commercial paste, small bubbles can be observed (Fig. 3b). А Pt thin layer applied to the conductive glass was used as a counter electrode (Fig. 3c). The surface of the assembled cell is shown in Fig. 3d. Fig.  4 shows the  I–V characteristics of  the  DSSCs. The  value of  the  open- circuit photovoltage Voc for the  cell with prepared paste was 0.33  V, for the  com- mercial paste Voc was slightly higher  — 0.38 V. At the same time, the short-circuit current density Isc for the cell with prepared paste was 0.96 mA/cm2, which is noticeably higher than for the element based on com- mercial paste — 0.6 mA/cm2. To prevent the evaporation of acetoni- trile from the electrolyte, the solvent was replaced with 3-methoxypropionitrile with a  higher boiling point. The  change of the electrolyte led to an increase in Voc for both the cell with prepared paste — 0.5 V, and for the cell with commercial paste — 0.47 V. The Isc values for the cell with pre- pared paste decreased to  0.9 mA/cm2, while for the cell with commercial paste, on the contrary, increased to 0.68 mA/cm2. According to  the  data obtained, we conclude that I–V characteristics of DSSCs based on the  prepared paste are close to the characteristics of DSSCs based on the commercial paste under the same con- ditions, namely: counter electrode, elec- trolyte, and assembly method. Replacing Fig. 3. Microscopic images of surfaces: photoanodes based on (a) prepared and (b) commercial pastes, (c) a cathode coated with a layer of Pt, (d) a cell assembled from the resulting electrodes 148 the solvent in the electrolyte increases both the current values and the voltage between the cell electrodes. Conclusions To produce a homogeneous paste of nan- odisperse TiO2, a combination of α-terpineol and ethyl cellulose as binding components was used. Applying the  prepared paste to the pretreated glasses using a screen print- ing mesh and annealing at a temperature of 450 °C allows forming a continuous layer of TiO2 on the glass substrate. The  DSSCs assembled from the  pre- pared and the commercial pastes demon- strated comparable photovoltaic charac- teristics of photoanodes. The open-circuit photovoltage value reaches 0.4–0.5  V; the short-circuit current density reaches 1 mA/cm2. The TiO2 layers produced by us- ing either prepared or commercial paste have similar properties, but the prepared paste is  less expensive to  manufacture. By changing other components (electro- lyte, a counter electrode, etc.) and improv- ing the assembly method, higher efficiency of DSSC can be achieved. Acknowledgements The team of authors is grateful to the Russian Science Foundation for financial support, grant No. 17-79-20165. A. S Steparuk would like to acknowledge the financial support for the analytical studies of synthesized compounds from the Ministry of Education and Science of the Russian Federation within the framework of the State Assignment for Research (project no. AAAA-A19-119012490006-1). References 1. Gratzel M. Mesoporous oxide junctions and nanostructured solar cells. Curr. Opin. Colloid Interface Sci. 1999;4(4):314–21. doi:10.1016/S1359-0294(99)90013–4 Fig. 4. I–V characteristics of assembled solar cells with various electrolytes: E1 — acetonitrile, E2–3-methoxypropionitrile 149 2. Joshi PH, Korfiatis DP, Potamianou SF, Thoma K-AT. Selected parameters leading to an optimized DSSC performance. Russ. J. Electrochem. 2013;49(7):628–32. doi:10.1134/S1023193513070045 3. Amit K, Sharma R, Pransu G, Boxman RL. Evaluation of the photo electrode degra- dation in dye sensitized solar cells. Russ. J. Electrochem. 2019;55(9):829–40. doi:10.1134/S1023193519090039 4. O’Regan B, Grätzel M. A low-cost, high-efficiency solar cell based on dye-sensitized colloidal TiO2 films. Nature. 1991;353(6346):737–40. doi:10.1038/353737a0 5. Nizard H, Meyer T, Oberpriller H, Meyer M, Benien H. Assessment of photocata- lytic activity of SolaronixTM nanostructured anatase, Journal of Photochemistry and Photobiology A: Chemistry. 2008;195(1):99–104. doi:10.1016/j.jphotochem.2007.09.012 6. Khalid Hossain M, Rahman MT, Basher MK. Influence of thickness variation of gam- ma-irradiated DSSC photoanodic TiO2 film on structural, morphological and optical properties. Optik. 2019;178:449–60. doi:10.1016/j.ijleo.2018.09.170 7. Kao MC, Chen HZ, Young SL, Kung CY, Lin CC. The  effects of  the  thickness of TiO2 films on the performance of dye-sensitized solar cells. Thin Solid Films. 2009;517(17):5096–9. doi:10.1016/j.tsf.2009.03.102 8. Baglio V, Girolamo M, Antonucci V. Influence of TiO2 Film Thickness on the Elec- trochemical Behaviour of  Dye-Sensitized Solar Cells. Int. J.  Electrochem. Sci. 2011;6:3375–84 9. Kumari JMKW, Sanjeevadharshini N, Dissanayake MAKL, Senadeera GKR, Thota- watthage CA. The effect of TiO2 photo anode film thickness on photovoltaic proper- ties of dye-sensitized solar cells. Ceylon J. Sci. 2016;45(1):33–41. doi:10.4038/cjs.v45i1.7362 10. Sedghi A, Miankushki HN. The Effect of Drying and Thickness of TiO2 Electrodes on the Photovoltaic Performance of Dye-Sensitized Solar Cells. Int. J. Electrochem. Sci. 2015;10:3354–62 11. Fitra M, Daut I, Irwanto M, Gomesh N, Irwan YM. Effect of TiO2 Thickness Dye Solar Cell on Charge Generation. Energy Procedia. 2013;36:278–86. doi:10.1016/j.egypro.2013.07.032 12. Mathew A, Rao GM, Munichandraiah N. Effect of TiO2 electrode thickness on pho- tovoltaic properties of dye sensitized solar cell based on randomly oriented Titania nanotubes. Mater. Chem. Phys. 2011;127(1-2):95–101. doi:10.1016/j.matchemphys.2011.01.032