Easy Room Temperature Synthesis of High Surface Area Anatase Nanowires with Different Morphologies.

Anatase nanowires were synthesized in solution by using a simple mixing of titanium diisopropoxide bis(acetylacetonate), lactic acid and sodium hydroxide at room temperature. We discuss effects of reaction parameters and post treatment (annealing) on the nanowire morphology, surface area, and crystallinity, as well as the competing morphology directing effects of lactic acid and sodium hydroxide. Then the room temperature nanowires were directly grown onto fluoride doped tin oxide (FTO) glass to form photoanodes. Photoelectrochemical measurements of the different nanowires were performed and compared to conventional nanowires produced by high temperature synthesis. Clearly the nanowires introduced in this work show a significant increase in the maximum photocurrent, compared to classic hydrothermal nanowire layers.

Introduction

Due to its stability, photoelectrochemical properties, and low prize, titanium dioxide (TiO2) has attracted much attention over the last decades. Applications of titania electrodes involve photocatalytic pollutant degradation under UV‐light irradiation, energy storage, gas sensors, photoinduced H2 generation, or dye sensitized solar cells.1, 2 Most of these applications rely on efficient electron‐hole separation and a fast transfer of charge carriers to the environment. If used as a photoelectrode, additionally fast transport to the back contact is desired. This makes structures advantageous that provide directed electron pathways and high surface area.3 Nanostructures, especially 1d nanostructures such as nanotubes, nanorods and nanowires offer a variety of these features and are therefore highly investigated structures.4

A broad range of techniques is used for 1d TiO2 nanostructure synthesis, such as electrochemical methods, hydrothermal and sol‐gel methods, template assisted fabrication, and PVD deposition.5, 6 There are several inherent drawbacks to some of the methods, for example, high energy consumption for the production, high prize, expensive equipment, or substrate inflexibility. Some of the most promising methods for the creation of nanostructured TiO2, namely hydrothermal methods, where the nanostructure can be tailored by changing synthesis parameters, are often based on chemical decomposition (hydrolysis) of titanium precursors into TiO2, allowing the growth onto a variety of substrates under high temperature and high pressure conditions.6, 7, 8 However, these conditions are also a key reason for the limited application of the processes – that is, due to harsh reaction conditions, high temperature equipment is required that not only increases production costs and but also induces processing complexity.

A typical synthesis of TiO2 nanostructures involves liquid titanium precursors such as TiCl4, titanium isopropoxide and titanium diisopropoxide bis(acetylacetonate) that undergo rapid hydrolysis when mixed with water. To stabilize these precursors, acids such as HCl and organic acids such as lactic acid can be used to hinder the decomposition in water based solution.8 Additionally, these acids can have morphology directing properties which then can be used under hydrothermal conditions to produce nanostructures such as nanorods and nanowires.8, 9, 10 More recently, it has been reported that lactic acid can show an especially good stabilizing effect on titanium precursors which was demonstrated by synthesizing water soluble diammonium salts like diammonium tris(2‐hydroxypropionato)titanate(IV).11 Solutions of bis(ammonium lactato)titanium dihydroxide are known to be stable in water, and they have already been used successfully for the synthesis of anatase nanocrystals.12 Another hydrothermal approach to produce TiO2 nanostructures is the use of sodium hydroxide to dissolve and recrystallize solid titanium oxide powder into titanate and anatase nanotubes under hydrothermal conditions, whereby anatase is obtained after suitable annealing in air. This illustrates the morphology directing properties of sodium hydroxide.7, 13, 14 Since this reaction involves the dissolution of TiO2, the concentration of sodium hydroxide and high temperatures are crucial for the reaction. If instead typical titanium precursors are used in even mildly basic media, instant hydrolysis sets in which prevents the use of milder synthesis conditions.

Herein we present an easy to perform synthesis of titanium oxide nanowires from liquid precursors using the stabilizing effect of lactic acid on titanium diisopropoxide bis(acetylacetonate) and the promoting effect of sodium hydroxide on the formation of TiO2 nanowires. This method effectively replaces the need for high temperatures which are normally crucial to convert the titanium precursor compounds into TiO2, and utilizes both morphology directing capabilities of lactic acid and sodium hydroxide to produce high surface area titanium oxide nanowires of various morphologies. Moreover, we show that these nanostructures can also be directly grown onto conducting FTO substrates under the same mild conditions.

Results and Discussion

Experiments were performed by mixing titanium diisopropoxide bis(acetylacetonate) with lactic acid in a glass beaker, aging the solution for 5 hours and adding this solution to a cooled solution of sodium hydroxide. After 16 hours the reaction was competed and the produced powder was filtered and washed. To evaluate the influence of the different constituents in solution on the nanowire formation process, namely lactic acid and OH− concentration, a series of screening experiments were performed. Figure 1 shows the morphological evolution of the resulting particles in dependence of the lactate ion concentration in solution varied from 1 % to 8 %. In the case of low lactate concentrations (Figure 1 a), agglomerations of nanoparticles were formed and no wire morphology could be obtained. When increasing the lactate concentration, agglomeration is reduced, as seen in Figure 1 b) and nanowires start to form, as shown in Figure 1 c). The nanowires get more pronounced and are further separated by increasing the lactic acid concentration, leading to well separated nanowires of roughly 10 nm to 20 nm diameter shown in Figure 1 d).

Figure 1

SEM pictures of TiO2 nanostructures formed by varying the amount of d‐lactic acid. a) 1 % b) 2 % c) 4 % d) 8 %.

Since the hydroxide concentration is the main driving force for the TiO2 evolution and has morphology directing capabilities on titanium based nanostructures, this key factor was also investigated by varying the OH− concentration in the growth solution from 0.001 mol/l to 14 mol/l. Figure 2 shows the morphological features of the as‐formed TiO2 structures for different hydroxide concentrations. If the OH− concentration is as low as 0.001 mol/l, no nanowires were formed (Figure 2 a) and the resulting nanoparticles agglomerate in a similar fashion as for low lactic acid concentrations. If the OH− concentration is raised to 0.2 mol/l, TiO2 nanowires start growing radially from nuclei and a hedgehog‐like morphology (as shown in Figure 2 b) is obtained. This structure resembles nanowires produced by hydrothermal methods with organic acids.9 By further increasing the OH− concentration to the range of 1 mol/l to 5 mol/l, the morphology directing properties of sodium hydroxide on nanowire formation are increased and the hedgehog spikes get elongated, the density of nanowires per particle is reduced and the separate particles start to get interlinked (Figure 2 c). At a hydroxide concentration of 6.4 mol/l, the morphology is changed into a net‐like structure (Figure 2 d), resembling nanowires grown under hydrothermal conditions with sodium hydroxide as the sole morphology directing agent.14 At a hydroxide concentration of 10 mol/l, the morphological evolution of the nanowires is hindered and particles composed of very short nanowires evolve as shown in Figure 2 e). Finally, when reaching an OH− concentration of 14 mol/l, no nanowire formation can be observed and assemblies of agglomerated nanoparticles shown in Figure 2 f) are produced.

Figure 2

Morphological evolution of TiO2 nanoparticles depending on hydroxide concentration a) 0.001 mol/l, b) 0.2 mol/l, c) 3 mol/l, d) 6.4 mol/l, e) 10 mol/l, f) 14 mol/l.

To investigate the properties of the different TiO2 morphologies, XRD, TEM, nitrogen adsorption, XPS and photocurrent spectra (incident photon to current efficiency measurements (IPCE)) were carried out for samples after different annealing conditions. Figure 3 a) shows XRD spectra for the as formed TiO2 nanostructures and Figure 3 b) shows the XRD spectra after thermal annealing of the hedge‐hog morphology in air for different times and temperatures. The as‐prepared samples show anatase peaks that can be identified for the net‐like and hedgehog‐like morphologies. The XRD peaks are similar in both cases. The main peaks of anatase are broadened and minor peaks cannot be detected. This is likely due to a small crystal size or due to a generally low crystallinity of the as‐prepared samples.15 In the case of the nanoparticle samples, no major peaks can be observed, which can be explained by a very low crystallinity and high amount of amorphous phase in these samples. The crystallinity can be increased by annealing – as a most promising morphology the hedgehog‐like nanowire structures were annealed in air for different temperatures ranging from 350 °C to 650 °C. As can be seen, most prominently in the three peaks between 35 ° and 40 ° (a zoomed in part of this region can be found in the supporting information Figure SI 1 a), crystallinity increases with increasing temperature. Since the peaks sharpen and get more pronounced with higher annealing temperatures, crystal size increases and the amount of amorphous phase is diminished. All structures remain anatase and no rutile phase was observed even when increasing the temperature to 650 °C. When comparing the annealing time, the sample annealed at 450 °C for 1 hour and 3 hours showed an increase in crystallinity with longer annealing times. By looking at TEM images of the as‐prepared samples, highly crystalline and well‐ordered areas, as shown in Figure 3c, can be observed; areas where the nanowires are composed of a polycrystalline structure or are of amorphous nature were also observed as shown in Figure SI 1 b). When looking at the SEM images of as‐formed and annealed samples, no major change in the morphology could be observed as shown in Figure SI 1 c) and d).

Figure 3

XRD measurements of a) different as‐formed morphologies, b) hedgehog‐like nanowires annealed at different temperatures and times, c) TEM image of as‐formed hedgehog‐like nanowire nanostructures, SEM pictures of d) hedgehog and e) net‐like nanowires directly grown on FTO and ICPE measurements of f) hedgehog and g) net‐like nanowires respectively.

Nitrogen adsorption measurements where conducted to determine the surface area of the as‐formed and annealed samples. Figure SI 2 a)–d) show the N2 adsorption isotherms for the as‐formed hedgehog‐like and net‐like, as well as the annealed hedgehog‐like and net‐like structures, respectively. High surface areas of 300 m2/g could be achieved by the net‐like nanowires, whereas the hedgehog morphology led to surface areas of roughly 240 m2/g. The isotherm of hedgehog‐like nanowires follows the characteristics of a type IV isotherm (UPAC) and shows type H2 ink bottle mesoporosity. In contrast, the net‐like nanowire isotherms are of type II with no pores present. When looking at the annealed samples, the surface area is decreased due to sintering effects but high surface areas of 190 m2/g and 120 m2/g, respectively, are still retained even after 3 hours of annealing which are still comparable or even higher than the surface area of hydrothermal titania nanotubes.7 Surface composition of the different nanowires was determined by XPS and show the presence of sodium and an increase in carbon content for the low temperature samples as compared to high temperature samples (Figure SI 3 a). Moreover an increase in OH− content with increased sodium hydroxide concentration in the reaction solution can be determined by the fitted O1 S peaks shown in Figure SI 3 b)–d). This may indicate a stronger OH− termination of the TiO2 nanowires when synthesized with higher NaOH concentration.

To evaluate the feasibility to easily fabricate photoelectrodes, we immersed a piece of FTO into the growth solution placed at an angle against the wall with the conducting side face down. For reference, a comparable solution was prepared without the addition of sodium hydroxide and placed in a Teflon lined autoclave and heated to 180 °C for 16 hours to conduct the transformation of the precursor into rutile TiO2, in line with the most conventional hydrothermal synthesis reaction for titania nanowires. The coated FTO substrates were then investigated by photocurrent spectroscopy. The photoelectrodes were all prepared in a one batch approach under typical conditions, to compare the result of a single deposition approach, since for all morphologies increasing the number of coating layers by repeating the growth step led to a diminished photoelectrochemical performance as shown in Figure SI 4 a) and b). Figure 3 d) and e) show the resulting layers of the net‐like and hedgehog nanowires on FTO, respectively. When considering the photoelectrochemical performance, the anatase nanowires show a clearly increased IPCE with elevated annealing temperatures and annealing times due to increased crystallinity as discussed before. Even though the net‐like morphology nanowires show higher surface areas per mass, in general the hedgehog‐like nanowires have higher IPCE, whereas the net‐like nanowires outperform the hedgehog nanowires at a wavelength of 300 nm which can be explained by the lower layer thickness of the net‐like structure and the therefore decreased surface area of the photoanode. As can be seen in Figure 3 d), the net‐like nanowires attach flatly on the FTO surface resulting in very thin layers of around 20 nm thickness. The hedgehog nanowires grow radially from the initial nucleation points on the surface, leading to an increased layer thickness, which was determined by SEM of the cross section shown in Figure SI 4 c) to be roughly 400 nm and therefore is comparable to the high temperature nanowire layer, which also reached this thickness after 16 hours as shown in Figure SI 4 d). These results are also supported by surface area measurements performed by dye adsorption tests in which a comparable surface area of 22.5 cm2 and 22.3 cm2 per cm2 of electrode area was achieved for the high temperature and hedgehog nanowires respectively, whereas the net‐like structure anode only reached 7.6 cm2 per cm2 of electrode area. The surface areas were calculated on the assumption that 1 nm2 is occupied by 1.95 N719 dye molecules.16 In general, nanowire absorption is shifted to lower wavelengths compared to hydrothermal nanowire layers. This can be explained by the different crystal structure of the room temperature nanowires compared to the high temperature ones. Nanowires grown by the hydrothermal method in assistance of lactic acid have been reported to be of rutile crystal structure, whereas our room temperature nanowires are anatase. This difference in crystal structure can also explain the higher IPCE values of the room temperature nanowires since anatase has been shown to have better electron mobility than rutile.17 To validate this assumption, UV‐vis was performed for the best samples and is shown in Figure SI 5 a). Adsorption edge is shifted to lower wavelength for the room temperature nanowires compared to the rutile structures which is in accordance to the ICPE results. Also when looking at the difference of net‐like and hedgehog nanowires adsorption, the above given explanation for the different IPCE performance is supported. Comparing the maximum IPCE of samples with identical growth time, the low temperature nanowire anodes show good photoelectrochemical performance and an increase of about 60 % and 30 % is achieved by the hedgehog and net‐like nanowire anodes, respectively, compared to classic hydrothermal synthesis. To determine the effect of decreased growth time and therefore reduced sheet thickness in a single coating circle on the efficiency of the nanowires, hedgehog nanowire anodes with lower reaction time were also prepared and are shown in Figure SI 5 b). Decreasing the reaction time to 8 hours had no significant effect on ICPE of the hedgehog structure anode whereas reaction times of 4 hours and 2 hours led to a strong decrease in maximum IPCE showing that a smaller layer thickness is not beneficial in case of the hedgehog nanowires.

Conclusion

In summary, we present an easy to perform method of synthesizing high surface area titanium oxide nanowires of different morphologies with anatase crystal structure under room temperature and normal atmospheric conditions. The effects of different morphology directing agents such as lactic acid and sodium hydroxide as well as different annealing temperatures were explored, and their effect on surface area and crystallinity were evaluated. High surface area of up to 300 m2/g and 240 m2/g for the as‐formed nanostructures were achieved, whereas elevated annealing helped to increase crystallinity of the wires. Photelectrochemical properties of the room temperature grown structures were compared to classic high temperature grown hydrothermal nanowire layers. For this, we achieved growth on FTO substrates. Photocurrent spectra show that the absorption is shifted to lower wavelengths, which is due to a larger band‐gap of anatase compared to rutile. Nevertheless, when comparing the maximum IPCE at 330 nm and 300 nm wavelength of our anatase room temperature synthesized anodes, an increase of 60 % and 30 % to the hydrothermal ones can be seen for the hedgehog‐like and net‐like nanowires synthesized for 16 hours, respectively.

Experimental Section

Anatase nanowires were grown by preparing titanium precursor solutions of titanium diisopropoxide bis(acetylacetonate) (75 wt% in isopropanol) with varying concentrations in d‐lactic acid (0.8 ml titanium precursor in 4 ml to 16 ml lactic acid) and aging the solution for 5 hours. Afterwards the precursor solution was added to a sodium hydroxide solution (the mass of sodium hydroxide in 284 ml water was varied from 7.75 g to 175.74 g) which was cooled down beforehand in a freezer at −15 °C for 30 minutes to compensate for heat evolution during the mixing step. The prepared solutions were mixed under stirring and kept at room temperature for 16 hours. The actual chemical composition for the hedgehog and net‐like nanowires solutions were 16 ml lactic acid (90 %), 284 ml di water, 0.8 ml titanium precursor as well as 7.98 g and 84.54 g sodium hydroxide respectively. The precipitated white powder was filtered and repeatedly washed with water and 0.1 molar hydrochloric acid until neutral conditions were reached. The as‐formed powder was dried in an oven at 70 °C for 12 hours and afterwards annealed at temperatures between 450 °C to 650 °C for 1 h to 3 h to increase crystallinity and to remove excess carbon. For photocurrent measurements anatase nanowires were directly grown on a 1.5 cm×2.5 cm piece of FTO (fluoride doped tin oxide coated glass), which was cleaned ultrasonically for 60 min in a 1 : 1 : 1 solution of water, isopropanol and ethanol and afterwards immersed in 8 ml of the reaction solution, produced as described above in a 10 ml glass bottle, leaned against the wall with the conducting side face down to prevent precipitation of nanoparticles due to gravity. To further reduce precipitation in solution, precursor concentration was reduced to halve of the corresponding powder synthesis. For reference, a TiO2 nanowire sample was produced without sodium hydroxide and conducting the synthesis at 180 °C in a Teflon lined autoclave for 16 h in an oven.

Nitrogen sorption experiments at liquid nitrogen temperature were performed using a volumetric gas sorption analyzer (3D Instruments). Degassing under vacuum conditions was performed for 6 h at 150 °C. Mass specific surface areas (Sm) were evaluated in a p/p0 range from 0.02 to 0.35 according to the Brunauer, Emmett, Teller theory.18 The “as‐formed” samples were annealed in 250 °C for 4 hours to reduce carbon contamination.

SEM pictures were taken by a Hitachi S4800.

For IPCE measurements, a standard 3‐electrode electrochemical cell with a 0.5 cm×0.5 cm Pt plate counter electrode, an Ag/AgCl reference electrode (3 M KCl) and a 0.1 M Na2SO4 electrolyte were used without nitrogen purging and a standard bias of +500 mV against Ag/AgCl reference electrode was applied.

Dye desorption measurements were conducted immersing dye loaded samples in a solution of 10 mM KOH (5 mL) for 30 min. The concentration of fully desorbed N719 dye was measured spectroscopically (by using a Lambda XLS UV/VIS spectrophotometer, PerkinElmer) at λ=520 nm.

Conflict of interest

The authors declare no conflict of interest.

As a service to our authors and readers, this journal provides supporting information supplied by the authors. Such materials are peer reviewed and may be re‐organized for online delivery, but are not copy‐edited or typeset. Technical support issues arising from supporting information (other than missing files) should be addressed to the authors.

Supplementary

Click here for additional data file.