Research on Chemical Intermediates

, Volume 39, Issue 4, pp 1623–1631

Preparation of large-area dye-sensitized solar cells based on hydrothermally synthesized nitrogen-doped TiO2 powders


DOI: 10.1007/s11164-012-0896-z

Cite this article as:
Liu, W., Feng, Z. & Cao, W. Res Chem Intermed (2013) 39: 1623. doi:10.1007/s11164-012-0896-z


Low-cost, yellowish, nanocrystalline nitrogen-doped titanium dioxide (N-doped TiO2) powder was synthesized by a hydrothermal method. The as-prepared N-doped TiO2 powder was characterized by X-ray diffraction, transmission electron microscopy (TEM), UV–Vis absorption spectra, X-ray photoelectron spectroscopy (XPS), and Brunauer–Emmett–Teller analysis techniques. The grain size of the prepared powder was around 13 nm as estimated by both Scherrer’s method and TEM images. The effect of the ratio of N-doped TiO2 particles to Degussa P25 on the photovoltaic performance of large-area dye-sensitized solar cells (DSSCs) was also investigated. The N-doped TiO2 electrode showed higher photovoltaic performance compared with that of pure P25 at constant irradiation of 100 mW cm−2, which is attributed to the large pore size and high surface area of N-doped TiO2 resulting in the introduction of extra charge carrier pathways that could be beneficial for overall charge transportation. Energy conversion efficiency of 5.12 % was achieved in a DSSC device with active area of 51.19 cm2.


Nitrogen-doped TiO2 Dye-sensitized solar cell Hydrothermal 


Dye-sensitized solar cells (DSSCs) are one of the low-cost alternatives to conventional silicon solar cells due to their lower-cost and simpler fabrication process [1]. Recently, a new efficiency record of 12.3 % (aperture area: 0.2 cm2) was reported by Ecole Polytechnique Fédérale de Lausanne (EPFL) under simulated sunlight [2]. The efficiency of large-size DSSCs (aperture area: 17.11 cm2) was reported as 9.9 % by Sony Corporation in 2010 [3]. However, further increase of the energy conversion efficiency remains critical for successful commercialization [4].

Recently, TiO2 photoelectrodes have been doped by metal and nonmetal atoms to enhance the light harvesting efficiency and overall conversion efficiency of DSSCs. Great progress has been made using TiO2 doped with metal or nonmetallic elements such as La, Zn, W, and N [5, 6, 7]. Among these materials, N-doped TiO2 materials proved to be excellent candidates for use in DSSCs [6, 7].

To improve both light harvesting and electron transport, in this work, nanostructured TiO2 films were prepared from mixed Degussa P25 and self-made nanocrystalline anatase N-doped TiO2 powder with smaller particle size. Anatase N-doped TiO2 powder was prepared by a quick, low-temperature hydrothermal method based upon our previous work [8, 9]. The effect of the ratio of self-made N-doped TiO2 powder to Degussa P25 on the photovoltaic performance of large-area DSSCs was investigated systematically.


Preparation and characterization of TiO2 powder and membranes

Desired amounts of industrial-grade TiOSO4 and CO(NH2)2 were dissolved in deionized water and then charged into a Teflon-lined autoclave. The chamber was flushed with nitrogen gas at pressure of 3.5 MPa. Then, the reaction chamber was heated to 150 °C, holding for 2 h. The synthesized products were cooled naturally and centrifugally separated. The separated products were washed with distilled water for several times, and dried at 150 °C for 12 h in a vacuum oven [8, 9]. The phase composition of the synthesized N-doped TiO2 powder was identified by X-ray diffraction (XRD) performed on a DMAX-RB diffractometer (40 kV, 30 mA, Cu Kα radiation; Rigaku, Japan). Grain size was estimated using Scherrer’s method. The Brunauer–Emmett–Teller (BET) specific surface area was measured by nitrogen adsorption at 77 K on a Quantasorb-18 (Quantachrome Co.). The morphologies of the synthesized nanopowder were observed by transmission electron microscopy (TEM, JEM-200CX; JEOL).

Powder mixtures with different ratios of N-doped TiO2 to Degussa P25 were introduced into 0.02 mol L−1 nitric acid. Then, the mixtures were treated for 2 min with a 600 W ultrasonic titanium probe at frequency of 15 pulses per second. After that, the obtained slurry was concentrated in a rotary evaporator at 35 °C for 3 h. Finally, the wet TiO2 was dispersed in ethanol and mixed with terpineol and a solution of ethyl cellulose in ethanol. After removing ethanol with a rotary evaporator, a paste consisting of 18 % TiO2 and 7 % ethyl cellulose in terpineol was prepared [10, 11]. Several pastes with different ratios of N-doped TiO2 powder to Degussa P25 were prepared using the same procedure, designated as NT-1 (N-doped TiO2:Degussa P25 = 1:1), NT-2 (N-doped TiO2:Degussa P25 = 1:2), NT-3 (N-doped TiO2:Degussa P25 = 1:3), and AT (pure Degussa P25).

Photoanodes were prepared on fluorine-doped SnO2 (FTO) substrates (resistance = 15 Ω sq−1, transmittance = 90 %; Nippon Sheet Glass Co., Japan) using a screen-printing method. The electrode contained a nanocrystalline TiO2 particle layer at the bottom to enhance dye adsorption, with larger TiO2 particles on top to serve as a light scattering layer for better light harvesting. The working electrode was treated with 40 mM TiCl4 and heat-treated at 500 °C for 30 min to obtain membranes, where were immersed in ethanolic solution (0.5 mM) of cis-di(thiocyanato)-N-N’-bis(2,2’-bipyridyl-4-carboxylic acid-4’-tetrabutylammonium carboxylate) ruthenium (II) (N719; Solaronix SA, Switzerland) and kept overnight at room temperature to dye the TiO2 membranes.

Membrane morphology was studied by field-emission scanning electron microscopy (FESEM, JEM-2010; JEOL, Japan) at accelerating voltage of 20 kV. Membrane thickness was measured using a surface profiler (Dektak 150; Veeco, USA).

Preparation of dye-sensitized solar cells

Dye-sensitized TiO2 electrode and platinum-coated FTO were used as working electrode (aperture area of TiO2 membrane was 51.19 cm2) and counterelectrode, respectively. Both electrodes were printed with silver paste lines as electron collector before coating with TiO2 membrane. The electrodes were placed face-to-face with a Surlyn spacer on the top of the silver electron collector, then heated up to seal the cell. The Surlyn spacer also works to avoid short-circuiting and prevent corrosion of the silver electron collector by the electrolyte. The redox electrolyte was a 0.5 M lithium iodide, 0.05 M iodine, 0.4 M 1,2-dimethyl-3-propylimidazolium iodide (DMPII), 0.5 M 4-tert-butylpyridine (TBP), and 0.1 M guanidine thiocyanate (GuSCN) solution in dehydrated acetonitrile (all chemicals in electrolyte from Acros Organics). The electrolyte was introduced into the gap between the electrodes using two holes predrilled in the counterelectrode. Then, a patch of Surlyn foil was used to seal the hole, and the cell was prepared successfully (Fig. 1).
Fig. 1

Prepared 10 cm × 10 cm size DSSC module (active area: 51.19 cm2)

Photovoltaic measurements of DSSCs

The photovoltaic performance of the DSSC was tested by recording current density–voltage (IV) curves using a Keithley 2400 source meter (Oriel) under illumination by simulated AM1.5 solar light from a solar simulator (Oriel-91193 equipped with a 1,000 W Xe lamp and an AM1.5 filter). The light intensity was calibrated using a standard Si solar cell (Newport).

Results and discussion

Phase composition and morphologies of TiO2 powder and membranes

Figure 2 shows the XRD patterns of the as-synthesized yellowish N-doped TiO2 powder. All peaks in the XRD patterns in Fig. 2 correspond to anatase, indicating that the powder was mainly composed of anatase TiO2. The mean grain size of TiO2 was estimated as 13 nm by Scherrer’s method. The specific surface area (SSA) was measured as 200 m2 g−1 by the BET method. Larger specific surface area benefits the amount of dye adsorbed in the photoanode. As shown in Fig. 3, most of the particles were spherical, and the grain size was uniformly distributed at about 12 nm.
Fig. 2

XRD patterns of N-doped TiO2 powder

Fig. 3

TEM image of N-doped TiO2 powder

The atomic composition and surface state of the N-doped TiO2 particles were investigated by X-ray photoelectron spectroscopy (XPS), as shown in Fig. 4. The XPS survey spectra showed that the surface of the particles was composed of Ti, O, N, and carbon. The atomic percentage of N was about 0.3 %. Substitution of N atoms will introduce an N 2p isolated band above the O 2p valence band at lower N-doping level [12, 13], consistent with the UV–Vis results. Figure 4b shows the peak binding energy of N 1s at 396.5 and 399.8 eV, corresponding to Ti–N and N–O bonds, respectively. It is obvious from Fig. 5 that the absorption edge of N-doped TiO2 powder has been extended to the visible light region. The absorption edge is around 460 nm as calculated by the straight-line extrapolation method.
Fig. 4

XPS spectra of N-doped TiO2 powder

Fig. 5

UV–Vis diffuse reflectance spectra of N-doped TiO2 powder

The thickness of the membrane was controlled as 13 ± 0.3 μm by adjusting the screen-printing processing parameters. Figure 6 shows typical SEM images of the prepared TiO2 membranes coated on the FTO glass substrate. Lots of pores existed in the membrane, as indicated in Fig 6a, b. With increasing amount of N-doped TiO2 powder, the connection between the nanoparticles was improved in the membrane due to the graded TiO2 particles.
Fig. 6

FESEM images of different thin membranes coated on glass substrate: a, b pure AT; c, d NT-1; e, f NT-2; g, h NT-3

Figure 7a, b shows the micromorphology of the light scattering layer. The presence of the TiO2 particles in the light scattering layer could enhance the light trapping.
Fig. 7

FESEM images of the light scattering layer of the cell

Photocurrent–voltage measurement

The effect of the ratio between the N-doped TiO2 particles and Degussa P25 on the photovoltaic performance of large-area DSSCs was investigated. Photocurrent–voltage measurements were carried out under irradiation with white light from a xenon lamp. Figure 8 presents current density–voltage curves for a 10 cm × 10 cm DSSC cell under full sun illumination (air mass 1.5, 100 mW cm−2). Three parallel cells were prepared under the same conditions to enhance the reliability of the obtained data.
Fig. 8

IV curve of DSSC cells

From Fig. 8, the energy conversion efficiency of sample AT was 2.93 %, while that for NT-2 reached up to 5.12 %, which is the largest among the five samples. Compared with that of sample AT, the short-circuit current density Jsc increased from 11.55 to 12.55 mA cm−2 and the open-circuit voltage (Voc) increased from 0.71 to 0.73 V. These results are in accordance with theoretical studies that predict an increase of both Jsc and Voc when the recombination rate at the TiO2–electrolyte interface decreases. The energy conversion efficiency of samples NT-1, NT-2, and NT-3 with different N-doped TiO2 powder ratios was higher by 69.9, 74.7, and 5.1 %, respectively, compared with that of sample AT.

There are two main reasons for using N-doped TiO2 powder to improve both light harvesting and electron transport: firstly, nitrogen doping can improve the light absorption efficiency of the TiO2 photoelectrode; secondly, the smaller particles may favor greater film compactness (as shown in Fig. 6) and reduce electron loss during the transport process. However, greater film compactness can decrease the efficiency of the electrolyte with the oxidized dye molecules on the TiO2 film and thus result in lower conversion efficiency. Therefore, there must be an optimal value of the proportion of small N-doped TiO2 particles in the nanometer powder. Here, the maximum energy conversion efficiency of 5.12 % was achieved for the sample with ratio of N-doped TiO2:Degussa P25 of 1:2.


Nanocrystalline N-doped TiO2 powder was synthesized by a quick, low-temperature hydrothermal method using industrial-grade TiOSO4 and CO(NH2)2 as starting materials. The grain size of the as-prepared particles was estimated as 13 nm, and the specific surface area was 200 m2 g−1. Nanostructured TiO2 films were prepared by mixing Degussa P25 with N-doped TiO2 powder. The porosity of the films decreased slightly and their absorption edges were red-shifted due to the filling of the larger pores between the P25 particles by the smaller N-doped TiO2 nanoparticles. Energy conversion efficiency of 5.12 % was achieved for the sample with ratio of N-doped TiO2 to Degussa P25 of 1:2 in an assembled DSSC cell (active area of 51.19 cm2). Compared with a pure Degussa P25 membrane, the photoelectric conversion efficiency of sample NT-2 was increased by 74.7 %. This is attributed to the lower porosity of the composite film. Work is in progress to gain further insight into mechanistic aspects as well as to study the stability of such N-doped TiO2 DSSC cells.


This work has been financially supported by the National Natural Science Foundation of China under grant number 51072019, the National High-tech Research and Development Program under grant number 2012AA030302, and the Opening Project of the State Key Laboratory of High Performance Ceramics and Superfine Microstructure under grant SKL201112SIC.

Copyright information

© Springer Science+Business Media Dordrecht 2012

Authors and Affiliations

  1. 1.Department of Inorganic Nonmetallic MaterialsSchool of Materials Science and Engineering, University of Science and Technology BeijingBeijingPeople’s Republic of China
  2. 2.School of Mechanical Engineering, North China University of Water Resource and Electric PowerZhengzhouPeople’s Republic of China

Personalised recommendations