Introduction

Transparent electrodes are important materials that have been intensively investigated over the last decade (Hecht et al. 2011). Tin-doped indium oxide (ITO) (Tahar et al. 1998) is the most successful transparent electrode material for a wide variety of applications including displays (Bechinger et al. 1996), functional glasses (Bechinger et al. 1996; Chen et al. 2005), solar cells (Krebs 2009; Zhu et al. 2011), and light-emitting diodes (Chen et al. 2005; Ho et al. 2000; Ho and Granström 1998). The increased demand for ITO due to the rapid expansion of these applications makes the cost of ITO films a serious concern for industry. Thus, finding alternative transparent electrode materials or optimizing the processes for producing ITO films are emerging priorities in materials science and engineering (Minami 2005; Argun et al. 2003; Tvingstedt and Inganäs 2007; Otabe et al. 1998; Kim et al. 2009). Developing improved processes is the best solution in the short term because it is difficult to replace ITO because of the requirements for commercial applications that include not only “conductivity and transparency” but also “total cost, productivity, and reliability” (Spaid 2012). For ITO films, the total cost is influenced not only by the price of indium itself but also the inefficiencies of the deposition methods (e.g., sputtering methods). In fact, only 15 % of the indium source is typically utilized in the production of transparent electrodes in devices (Hecht et al. 2011; Li et al. 2011). In addition, sputtering deposition chambers have become larger and more expensive to fabricate large devices with high productivity. Recycling of the ITO waste introduces extra cost. Thus, there is the opportunity to replace the current deposition methods for ITO with more cost-effective deposition processes.

One of the promising routes for cost-effective production is the solution deposition of ITO films. Because a precursor “ink” can be used to deposit ITO films, existing roll-to-roll or ink-jet systems can be utilized for low-cost, reduced-waste, highly scalable, and high speed manufacturing (Chen et al. 2005; Bühler et al. 2007; Jeong et al. 2010). Compared to metal organic depositions of ITO precursors that require crystallization sintering at over 500 °C where thermal damage can occur to other components of the device, films formed from ITO nanoparticles do not require high temperature sintering because crystalline nanoparticles are already formed during chemical production. In addition, nanoparticle inks capped by organic ligands can be stable with reliable quality (Ba et al. 2006; Gilstrap et al. 2008; Sun et al. 2010). Given these advantages, interest in ITO nanoparticle inks has steadily increased (Bühler et al. 2007; Jeong et al. 2010; Goebbert et al. 1999; Ba et al. 2006; Gilstrap et al. 2008). Although solution deposition of ITO nanoparticle ink is a simple pathway to fabricate ITO thin films, a key step is to produce ITO nanoparticles with high crystallinity and homogeneous composition.

There are several reports of syntheses of monodispersed ITO nanoparticles (Bühler et al. 2007; Ba et al. 2006; Gilstrap et al. 2008; Sun et al. 2010). In a typical synthesis, pyrolysis of metal organic precursors is conducted in a high boiling point organic solvent. The decomposition of the metal salt generates nuclei that subsequently grow into well-defined nanoparticles. However, tight control of the reaction conditions is required to avoid undesirable side reactions. The need for such control of the reaction conditions is a practical disadvantage of these syntheses for producing nanoparticle inks.

Recently, Caruntu et al. has reported synthesis of In2O3 nanoparticles using oleyl alcohol as a solvent in air (Caruntu et al. 2010), but did not report tin-doping of these nanoparticles using this approach. In general, impurity doping by pyrolysis is difficult because each metal precursor has a different decomposition temperature. In this article, we have synthesized monodispersed ITO nanoparticles from indium acetylacetonate in oleyl alcohol in air by a unique temperature profile. Since oleyl alcohol is a non-toxic alcohol, this route offers the added benefit that it reduces environmental risks. In addition, we demonstrate characteristics of ITO nanoparticle films as transparent conductive films. The ITO nanoparticle films are able to deposit uniformly onto high-aspect-ratio structures.

Experimental

Synthesis of ITO nanoparticles

All chemicals were purchased from Sigma Aldrich. Indium (III) acetylacetonate (99.9 %, 0.45 mmol) and tin(II) acetylacetonate (99.9 %, 0.05 mmol) were mixed with 12.5 mL of oleyl alcohol at room temperature in a 50 mL three-neck round-bottom flask under a flowing stream of air (Caruntu et al. 2010). The reaction mixture was heated to 200 °C and kept for 30 min to dissolve all of precursors, then heated to 260 °C at a rate of 2 °C/min and maintained at 260 °C for 30 min under flowing of nitrogen to nucleate ITO clusters. Since the mixture of indium and tin precursors easily forms aggregated precipitation, 30 min of aging at 260 °C is important to form ITO clusters. During the aging at 260 °C, the color of the reaction mixture changed slowly from a light yellow to a dark blown. The solution temperature was then increased to 320 °C and maintained for 1.5 h. During the aging at 320 °C, the color of the reaction mixture changed from a dark brown to a dark blue solution. The resulting solution was cooled to room temperature, 25 mL of ethanol and 5 mL of methanol were added to precipitate the nanoparticles and the particles were isolated by centrifugation for 30 min at 7,500 rpm. The solid product was washed six times with ethanol and then dispersed in toluene. The resulting toluene dispersion was centrifuged for five min at 3,500 rpm to remove any insoluble solids.

Synthesis of ZnO nanopillars

ZnO nanopillar substrates were fabricated on Si substrates by a gold-nanoparticles-catalyzed vapor–liquid–solid method as described previously (Ito et al. 2008). In brief, ZnO seed films were prepared by spin casting a sol–gel Zn precursor onto Si wafers with a 3-μm thermal SiO2 layer, followed by an annealing step. A mixture of ZnO and carbon powder (roughly 1:1 by mass) was placed in a small quartz tube as a ZnO source, and the ZnO substrate was placed downstream from the source. The substrate temperature was held constant at 600 °C. The source temperature was raised to 900 °C and held for 20 min in N2 gas flow (2.5 SCFH). Then, the furnace was shut down and cooled to room temperature while maintaining the nitrogen flow.

Deposition of ITO nanoparticles

An ITO nanoparticle ink was prepared by dissolving the ITO nanoparticles (10 % of Sn) in a mixture of toluene and 1-octanol (1:20 by volume). Films were deposited from this ink on glass substrates or ZnO nanopillar substrates using a spin coating (3,000 rpm).

Measurements

Transmission electron microscope (TEM) observation was carried out with a FEI Titan operated at 300 kV. Measurements of size distribution of nanocrystals are expressed as a function of the standard deviations. TEM and X-ray photoelectron spectra (XPS) samples were prepared by drop-casting onto a holey carbon grid and Si substrates, respectively. XPS was carried out with a Thermo Scientific ESCALAB 250 X-ray Photoelectron Spectrometer using an aluminum X-ray source. The binding energy scale of all measurements was calibrated using the C1s peak at 284.8 eV. SEM was carried out using a ZEISS Ultra-55.

Results and discussion

To synthesize nearly monodispersed ITO nanoparticles in oleyl alcohol, careful control of reaction temperature is necessary. Figure 1 shows the temperature profile used to synthesize monodispersed ITO nanoparticles from the acetylacetonate complexes in oleyl alcohol. If the heating rate is higher than 3 °C/min, a precipitation of black powder (a mixture of indium metal, In2O3, and SnO2) occurs. By way of contrast, in the case of In2O3 nanoparticles, any heating rate lower than 5 °C/min is acceptable. In that case, the temperature of the precipitation is about 260 °C which corresponds with the nucleation temperature of In2O3 clusters (Caruntu et al. 2010). The nucleation step of ITO clusters is critical to control the generation of dispersed ITO nanoparticles. After the initial nucleation of ITO clusters is completed during an aging at 260 °C for 30 min, no precipitation occurs. Introduction of a nitrogen atmosphere during aging at 320 °C is also an important step for ITO nanoparticles because heating in this inert environment introduces oxygen vacancies in the ITO nanoparticles that enhance electric conductivity. Aging under air makes color of ITO nanoparticle solution brown.

Fig. 1
figure 1

The temperature profile of the reaction mixture for production of monodispersed In2O3 and ITO nanoparticles from acetylacetonate precursors in oleyl alcohol. The blanketing gas is changed from air to nitrogen at 320 °C

Full size image

Figure 2a and b show TEM images of In2O3 and 10 % tin-doped ITO nanoparticles, respectively. The images reveal that both In2O3 and ITO nanoparticles are nearly monodisperse and exhibit no aggregation. Measurements of 150 random particles indicate average diameters of 7.0 nm (±0.7 nm) for In2O3 and 8.6 nm (±0.9 nm) for ITO, both with narrow size distributions (Fig. 6c, d). The colors of In2O3 and ITO nanoparticle solutions are yellow and blue, respectively (insets of Fig. 7c, d). The color of ITO solution is the result of the effects of doping by tin impurities.

Fig. 2
figure 2

TEM images of a In2O3 and b 10 % tin-doped ITO nanoparticles synthesized at 320 °C in oleyl alcohol. Size distribution analyses of c In2O3 and d ITO; Those insets show optical photographs of In2O3 and ITO nanoparticles solution

Full size image

Figure 3a, b show selected area electron diffraction (SAED) patterns of In2O3 and ITO nanoparticles, respectively. The diffraction patterns indicate that the matrix phase of both types of nanoparticles is cubic In2O3 without any different phases. There is no SnO2 phase in ITO. Figure 3c, d show high-resolution TEM images of In2O3 and ITO nanoparticles, respectively. These images suggest that both types of nanoparticle are single crystal.

Fig. 3
figure 3

Selected area electron diffraction patterns of a In2O3 and b ITO nanoparticles and high-resolution TEM images of c In2O3 and d ITO nanoparticles

Full size image

Figure 4 shows normalized XPS of the O1s core levels for ITO nanoparticles. The binding energy scale of all measurements was calibrated using the C1s peak at 284.8 eV. Peaks of O1s shift approximately 0.4 eV toward high binding energy by adding tin into In2O3 matrix. The peaks can be assigned based upon literature reports for ITO. Peak I of O1s at 529.6 eV ± 0.2 eV is due to bulk O2− ions whose neighboring indium atoms possessed a full complement of six nearest O2− ions, and peak II at 530.6 ± 0.2 eV corresponds to O2− ions in oxygen deficiency regions (Donley et al. 2002). Peaks III and IV at 531.6 ± 0.2 eV and 532.6 ± 0.2 eV are assigned to InOOH or In(OH)3 and absorbed H2O or contaminants (Donley et al. 2002), respectively. Tin-doping into In2O3 matrix causes a decrease of peak I and increases in peaks II and III, which indicates introduction of oxygen vacancies and surface defects. These results are consistent with the idea that the effect of tin-doping is not only substitution of indium sites but also introduction of oxygen vacancies (Fan and Goodenough 1977; Lin et al. 2005). The oxygen vacancies also work as n-type dopants to enhance conductivity. The results of XPS suggest that the oxygen vacancies are a result of the tin-dopant ions.

Fig. 4
figure 4

XPS spectra of In2O3 and ITO nanoparticles. O1s spectra overlaid for comparison (The counts of the largest peak in each of the two spectra were normalized)

Full size image

Figure 5a shows UV–Vis transmission spectra of In2O3 and ITO nanoparticles solutions. Both solutions are transparent throughout most of the visible range. The UV–Vis spectra of the ITO nanoparticles solution show absorptions at longer wavelength that are likely the result of free-electron vibrations (Gilstrap and Summers 2009) of highly conductive materials. The blue shift of the bandgap absorption in ITO nanoparticles relative to In2O3 nanoparticles can be explained by Burstein-Moss effect (increasing of free-electron density). Taken together, these results suggest that the ITO nanoparticles are optically semiconducting, but electrically metallic even though the size is below 10 nm.

Fig. 5
figure 5

UV–Vis transmission spectra of In2O3 and ITO nanoparticle solutions

Full size image

We have produced transparent electrodes by depositing ITO nanoparticles films from the ITO nanoparticle ink. A suitable ink was produced by dispersing the ITO nanoparticles in a mixture of toluene and 1-octanol (1:20 by volume). ITO nanoparticle films were deposited by spin coating onto Si or glass substrates and the samples were annealed at 300 °C to remove the surface ligands of the nanoparticles. The thickness of the film can be increased by performing additional depositions, each followed by annealing at 300 °C. Figure 6a, b shows an optical photograph and a cross-section SEM image of an ITO nanoparticle film, respectively. The ITO film (thickness: 110 nm) is transparent in visible range and the film consists of nearly close-packed ITO nanoparticles. Figure 6c, d show a UV–Vis transmission spectrum and the dependence of electric conductivity of the In2O3 and ITO nanoparticle films upon the post-deposition annealing temperature. The ITO nanoparticle film has 98 % of transparency in the visible range and has a similar spectrum to the ITO nanoparticle solution. Electrical resistivity values of In2O3 and 10 % tin-doped ITO nanoparticles films after annealing at 300 °C show 1 × 102 and 7 × 10−3 Ω cm, respectively. The difference of resistivity with and without tin-doping is larger than 10,000 times.

Fig. 6
figure 6

a An optical photograph, b a cross-section SEM image, and c a UV–Vis transmission spectrum of the ITO nanoparticle film. d Dependence of the electric conductivity of In2O3 and ITO nanoparticle films on the post-deposition annealing temperature

Full size image

We aimed to investigate whether these inks could also be deposited on non-planar substrates. Figure 7 shows SEM images of ITO nanoparticles deposited by spin coating on a substrate containing vertical ZnO nanopillars (Ito et al. 2008). Since 1-octanol is highly wettable on oxide surfaces, the ITO nanoparticle ink in 1-octanol is able to be deposited onto ZnO nanopillars. After repeating the spin coating (five times), ZnO nanopillars are completely coated by ITO nanoparticles film from the bottoms to tips. These results show that, in comparison with conventional sputtering methods, ITO inks can be deposited onto high-aspect-ratio structures or porous structures if ITO inks can penetrate into structures.

Fig. 7
figure 7

Cross-section SEM images of ITO nanoparticles coating onto a ZnO nanopillar substrate. a Low magnification. b High magnification of bottom of ZnO nanopillars. c High magnification of tip of ZnO nanopillars

Full size image

Conclusions

In conclusion, we have synthesized nearly monodispersed ITO nanoparticles using a benign solvent (oleyl alcohol) and produced ink that is suitable for forming thin films by spin coating. The deposited ITO nanoparticles film showed high enough conductivity for transparent electrodes and can be coated onto high-aspect-ratio substrates. The ITO nanoparticle ink offers a number of advantages as a deposition method of transparent electrode onto a variety of structured substrates for technologically important applications.