You are on page 1of 7

View Article Online / Journal Homepage / Table of Contents for this issue

PAPER www.rsc.org/materials | Journal of Materials Chemistry

Solution growth of anatase TiO2 nanowires from transparent conducting


glass substrates†
Pavel A. Sedach,a Terry J. Gordon,a Sayed Y. Sayed,b Tobias F€urstenhaupt,c Ruohong Sui,a
Thomas Baumgartnera and Curtis P. Berlinguette*a
Received 3rd February 2010, Accepted 24th April 2010
Published on 20 May 2010. Downloaded by University of Michigan Library on 28/10/2014 04:42:11.

First published as an Advance Article on the web 20th May 2010


DOI: 10.1039/c0jm00266f

Sol–gel reaction conditions that enable the growth of one-dimensional (1D) anatase titanium dioxide
(TiO2) nanostructures from fluorine-doped indium tin oxide (FTO) substrates are described. The
generation of these linear nanostructures is achieved using acetic acid (HOAc) and titanium
isopropoxide (Ti(OiPr)4) in anhydrous heptane in the absence of an external bias or template. The
procedure requires the functionalization of base-treated substrates with Ti-oxide nucleation sites,
which serve as a foundation for the growth of linear Ti-oxide macromolecules. Calcination of these
macromolecules at 450  C under an ambient atmosphere produce uniform films of randomly oriented
anatase TiO2 nanowires. The nucleation and growth processes are both acutely sensitive to the relative
molar ratio (R) of HOAc to Ti(OiPr)4. Optimal surface coverage of the nucleation sites is observed
when the R value utilized for the nucleation phase (denoted Ri) is equal to 1.3. The highest quality
nanowire films were obtained when the R value employed during the gelation phase (denoted Rf) was
held between 8.5 and 14. Characterization of the films by electron microscopy revealed a uniform film
of disordered anatase TiO2 nanowires with high aspect ratios. The dimensions of the nanostructures
correspond to lengths of ca. 1–10 mm and widths of 54  10 nm. High-resolution transmission electron
microscopy (HRTEM) and X-ray diffraction (XRD) techniques demonstrate that the anatase
nanowires are a linear arrangement of crystallites ranging in size from 13 to 19 nm. A systematic
evaluation of how reaction conditions (e.g., solvent volume, stoichiometry of reagents, substrate base
treatment) affect the generation of these TiO2 films is presented.

Introduction the DSSC. Devices containing a TiO2 anode have been shown to
exhibit relatively high power conversion efficiencies;15,16
One-dimensional semiconducting metal-oxide nanostructures however, the growth of TiO2 nanowires from a substrate is
have received significant attention in a wide range of applications inherently more difficult due to the crystal structure and
primarily due to their ability to constrain the movement of symmetry of the respective materials.17
electrons and photons in one direction.1–3 In the context of While there is a broad set of electrochemical,11,18 sol-
electronic and photoelectrochemical applications (e.g., dye- vothermal19–24 and hydrothermal25,26 techniques capable of
sensitized solar cells (DSSCs)4–6), the electrical and structural generating TiO2 nanoscale wires, rods, and/or tubes in the bulk
anisotropy within this class of materials has proven to be phase, the attachment of these structures to a support is often
a particularly effective means of increasing the rate of electron more challenging.27,28 This shortcoming has important implica-
transport and charge collection at the supporting anode tions in the construction of device structures because the covalent
(Fig. 1).7–9 The vast majority of linear nanostructures that have attachment of the nanowire array to a substrate is often
been successfully incorporated into photovoltaic devices, such as a necessary prerequisite for optimal charge collection. Notwith-
the DSSC, have been based predominantly on ZnO7–9 and standing, a host of techniques leading to nanowire arrays have
TiO2.10,11 Although the crystalline nature of ZnO is more recently been documented; e.g., hydrothermal treatment;29
accommodating to axial growth,12,13 it is susceptible to degra- alumina or ZnO nanorod templating;18,30,31 thermal deposition;32
dation in acidic media,14 which compromises the power output of and electrochemical anodization.11,33 Despite these synthetic
advances, the scale of these methods is generally limited by the
a
Department of Chemistry and The Institute for Sustainable Energy,
Environment & Economy, University of Calgary, 2500 University Drive,
NW, Calgary, Alberta, Canada T2N-1N4. E-mail: cberling@ucalgary.
ca; Fax: +403 289 9488; Tel: +403 210-6263
b
National Institute for Nanotechnology, Department of Chemistry,
University of Alberta, Edmonton, Canada T6G-2G2
c
Microscopy and Imaging Facility, Health Sciences Centre, 3330 Hospital
Drive NW, Calgary, Alberta, Canada T2N-4N1
† Electronic supplementary information (ESI) available: A summary of
reaction conditions for all experiments and auxiliary AFM, SEM, and
XRD data. See DOI: 10.1039/c0jm00266f Fig. 1 Nanostructured film architectures.

This journal is ª The Royal Society of Chemistry 2010 J. Mater. Chem., 2010, 20, 5063–5069 | 5063
View Article Online

need for exotic reagents or templates, multiple reaction steps, form of TiO2. While we note that this acid dependence does not
and/or energy-intensive procedures. always hold,25 we demonstrate in this work that HOAc does
On this basis, recent independent reports by Grimes et al.27 and facilitate the preferential formation of linear TiO2 macromole-
Aydil et al.34 detailing the sol–gel growth of oriented rutile TiO2 cules to ultimately afford anatase nanowires on TCO substrates
nanowire arrays from transparent conducting oxide (TCO) that have been previously coated with Ti-oxide nucleation sites.
substrates in the absence of a template and/or external bias, We provide a systematic evaluation of how reaction conditions
represent important advancements in the field. The imple- affect the nucleation and growth of the Ti-oxide macromolecules,
mentation of these nanowire films in the DSSC produced and how these parameters affect the surface coverage and
reasonable power conversion efficiencies (h) of ca. 3–5%; morphologies of the anatase TiO2 nanostructured films.
Published on 20 May 2010. Downloaded by University of Michigan Library on 28/10/2014 04:42:11.

however, this value is still significantly less than the h > 10% that
can be produced with films based on anatase TiO2 nano-
particles.35 The lower performance of the films with the linear
Results and discussion
nanostructures can be ascribed to a lower available surface area
necessary for dye loading.9,34 Moreover, the rutile phase of TiO2 This study details how an appropriate ratio of HOAc and
typically exhibits higher charge recombination rates and lower Ti(OiPr)4 in anhydrous heptane fosters the axial growth of linear
transparencies relative to anatase TiO2 nanoparticulate films of Ti-oxide macromolecules from TCO substrates that render
comparative thickness.36 anatase TiO2 nanowires upon calcination. We have observed
We have therefore focused our efforts on the growth of anatase that the following steps are critical to achieving robust, uniform
TiO2 nanowires from TCO substrates using facile synthetic sol– films of anatase TiO2 nanowires attached to a TCO substrate: (i)
gel techniques.37,38 Achieving this goal is challenging because base treatment of the substrate; (ii) formation of Ti-oxide
anatase TiO2 adopts a distorted octahedron edge-sharing nucleation sites on the surface; (iii) growth of linear Ti-oxide
packing arrangement, whereas the octahedron corner-sharing macromolecules from the nucleation sites; (iv) gelation; and (v)
motif of the thermodynamically more stable rutile phase is more calcination (Fig. 3). While the morphology of the film is sensitive
accommodating to axial growth (Fig. 2).17,39 Moreover, the to a number of factors, we have determined that an optimal
lattice matching factor between FTO and rutile TiO2 is < 2%,34 molar ratio (R) of HOAc to Ti(OiPr)4 during two distinct phases
a feature that facilitates the linear growth of the rutile phase of the reaction is fundamental to nanowire growth. Specifically,
directly from the substrate. This same structural attribute cannot the first phase, or nucleation phase, requires an initial R value,
be exploited in the case of anatase TiO2 because of a large lattice denoted Ri (where Ri ¼ moli HOAc/moli Ti(OiPr)4), near unity,
mismatch of ca. 19%.40 Despite said obstacles, we outline herein while the subsequent growth phase requires a final R value,
a molecular self-assembly approach that overcomes the challenge denoted Rf (where Rf ¼ (moli HOAc + molf HOAc)/moli
of generating anatase TiO2 arrays on TCO substrates using sol– Ti(OiPr)4), greater than 5. Calcination conditions were held
gel chemistry without the aid of an external bias or template. constant for all trials: 450  C for 2 h at a ramp rate of 5  C min1
The synthetic conditions we use to grow films of anatase TiO2 (temperatures > 500  C led to the onset of the rutile phase, and
nanowires follow design principles that have been shown to are not discussed in this study). The specific roles of these reac-
produce high aspect ratio TiO2 nanostructures in the bulk tion parameters, as well as the effects of stirring, substrate
phase.22 For example, we have shown previously that the treat- pretreatment and solvent volume, on the growth process of the
ment of Ti(OiPr)4 with HOAc in non-polar media suppresses the nanostructures are addressed in sequence below (experimental
rapid precipitation of particles to favor the formation of linear conditions are summarized in Table S1, ESI†).
macromolecules that form anatase TiO2 nanowires upon calci-
nation.38 It has also been documented in other solvothermal
processes that the identity of the acid plays an integral role in
determining the crystalline phase of the TiO2 product; e.g., there
is a progressively greater preference for the anatase phase over
rutile when the hydrolyzing acid is HOAc > H2SO4 > HNO3 >
HCl, respectively.41,42 Indeed, the aforementioned studies by
Grimes et al.27 and Aydil et al.34 using HCl produced the rutile

Fig. 3 Reaction steps outlining the formation of anatase TiO2 nano-


Fig. 2 Distorted octahedron (a) edge-sharing and octahedron (b) wires from a substrate: (a) formation of Ti-oxide nucleation sites
corner-sharing atomic arrangements in the solid-state for the anatase and (affected by Ri); (b) growth of linear Ti-oxide macromolecules (affected
rutile forms of TiO2, respectively.39 by Rf); (c) gelation.

5064 | J. Mater. Chem., 2010, 20, 5063–5069 This journal is ª The Royal Society of Chemistry 2010
View Article Online

Substrate and pretreatment

The surface of the substrate plays an integral role in the binding


of TiO2 nanostructures; thus, the initial phase of our studies
evaluated the growth processes from ITO and FTO substrates.
Atomic force microscopy (AFM) imaging shows that the surface
roughness of ITO (Rrms ¼ 2.2 nm) is less than that of the FTO
surface (Rrms ¼ 31.2 nm; Fig. S1, ESI†). Moreover, the AFM
images show that the ITO surface consists of small grain sizes
Published on 20 May 2010. Downloaded by University of Michigan Library on 28/10/2014 04:42:11.

that are on the order of tens of nanometres, while the FTO


substrate exhibits granular features with diameters in the 50–
150 nm range. The grain morphology and size of the FTO is
known to affect the growth of oriented nanostructures. Aydil
et al.34 noted that the convergent crystal lattice parameters of
FTO (a ¼ b ¼ 4.687 A)  43 and the rutile phase of TiO2 (JCPDS
No. 88-1175, a ¼ b ¼ 4.594 A)  enabled the growth of rutile
nanowires directly from the substrate. The lattice parameters of
anatase TiO2 (JCPDS No. 21-1272, a ¼ b ¼ 3.782 A),  however,
are markedly different than those of FTO; thus, this same feature
cannot be easily exploited to grow anatase structures in the same
manner.
These observations prompted us to implement a synthetic
protocol that involves the initial functionalization of the surface
with Ti-oxide nucleation sites. Initial base treatment of the
substrate was found to enhance the affinity of the nucleation sites Fig. 4 SEM images of anatase TiO2 nanowire films obtained at Ri value
1.3, Rf ¼ 14 (scotch taped surface).
for the surface. FTO substrates that are not subject to base
treatment lead to a poor coverage of TiO2 nanowires when
holding the other reaction parameters at parity. An overnight
a random collection of residual particulates that collect on the
treatment of FTO with 0.5 M NaOH, however, produced
surface of the wires during film growth that can be easily
significantly better surface coverage (Fig. S2, ESI†). The
removed by transparent tape.
hydrophilic OH sites present on the surface of the FTO facilitate
Cross-sectional field-emission scanning electron microscopy
the rapid hydrolysis of the Ti4+ precursor at the heptane/FTO
(FE-SEM) images of an FTO substrate after the nucleation
interface. The base treatment of the ITO substrates also
phase (prior to the growth phase) provide a snapshot of the
improved the surface coverage of the TiO2 nanostructures, but
nucleation sites on the surface (Fig. 5a). The plan view SEM
the resultant substrate properties were poor relative to the FTO
image highlights the uniformity of the nucleation sites over the
substrates (e.g., the electrical conductivity of the ITO was
entire FTO surface, while the cross-sectional image (Fig. 5a,
compromised by thermal sintering). Notably, a minor decrease
inset) reveals the aggregation of clusters approximately 30–50 nm
of 1 U cm2 in sheet resistance for the FTO substrates was
in diameter. It is evident at this early stage of the reaction that
observed after chemical and thermal treatment. Based on these
there is a preference for axial growth from the irregular surface of
collective observations, FTO films treated with 0.5 M NaOH for
the FTO. The cross-sectional FE-SEM image of the TiO2 film
24 h were used for our examination of the other reaction
following wire growth provided in Fig. 5b shows clear evidence
parameters.
that the wires originate from nucleation sites on the substrate
surface. The orientation of wire growth appears to be orthogonal
to the irregular faces of the FTO grains (Fig. 5b, inset), thereby
Substrate functionalization: role of Ri
driving wire growth in different directions.
The nucleation procedure involves the addition of base-treated A quantitative assessment of the surface coverage is a difficult
substrates to a reaction flask containing an initial molar ratio of task because many of our conclusions are made after gelation has
HOAc to Ti(OiPr)4 (Ri) in heptane at 60  C. In order to delineate occurred, thereby providing an indirect evaluation of the nucle-
the nucleation and subsequent growth processes, a series of ation mechanism. Notwithstanding, our results indicate that Ri
experiments were carried out where Rf was held constant at 14 directly affects the uniformity of the surface coverage of the wires
while adjusting Ri over the 0.5–14 range (Table S1, ESI†). While (Fig. 4 and 5), with a lesser effect on the subsequent macromo-
various TiO2 nanostructures consisting of variable aspect ratios lecular growth process and morphology of the calcined film. The
were obtained for all experiments, the Ri value appears to fact that the best coverage of non-aggregated nanowires was
primarily govern surface coverage (Fig. 4). When Ri ¼ 0.5, for obtained when Ri ¼ 1.3 is striking in that it resonates with an
example, there is non-uniform coverage of TiO2 nanowires on acid : Ti ratio that has been shown to promote the
the substrate. At Ri ¼ 1.3, optimal surface coverage is observed formation of hexanuclear molecular intermediates; e.g.,
(Fig. 4); reasonable surface coverage at Ri $ 2 is also observed, Ti6O6(OAc)6(OiPr)6.22,38,44,45 It has been shown that these
but the films are relatively poor in comparison with those molecular clusters can be generated in solution in < 1 min at R ¼
obtained at Ri ¼ 1.3. Note that the films at Ri ¼ 1.3 contain 1.3, and suggested that the ligand arrangement about the Ti-

This journal is ª The Royal Society of Chemistry 2010 J. Mater. Chem., 2010, 20, 5063–5069 | 5065
Published on 20 May 2010. Downloaded by University of Michigan Library on 28/10/2014 04:42:11. View Article Online

Fig. 5 (a) Plan-view SEM and cross-sectional FE-SEM (inset) images of


TiO2 films prior to wire growth highlighting a uniform film of nucleation
sites. (b) Cross-sectional FE-SEM images of TiO2 nanowire films at Ri ¼
1.3 and Rf ¼ 8.5; the thickness of the film is 5 mm. Fig. 6 SEM images of TiO2 films obtained at Ri ¼ 1.3 and Rf ¼ 7. (a)
Solution perturbation after nucleation and substrate addition; (b) solu-
tion is not stirred. Rf ¼ 7 conditions were chosen to illustrate the densi-
oxide core of the cluster facilitates axial condensation at higher R fication of nanowire growth because at Rf > 8.5 a higher density of
values to form macromolecular chains.38 On this basis, we nanowires makes the changes in solution perturbation more difficult to
postulate that the Ti-oxide clusters form in solution and then distinguish.
dehydratively couple to the surface hydroxyl groups of the base-
treated substrate; axial condensation is not likely to occur until
an additional acid treatment (i.e., Rf). Inspection of the SEM images reveals that both the Rf value
In experiments where Ri < 1, a limited quantity of Ti(OiPr)4 is and stirring have a dramatic effect on the density and
hydrolyzed; thus, the production of precursors that can bind to morphology of the TiO2 nanostructures. Reactions carried out at
the substrate surface is restricted. As a consequence of a low Ri, Rf # 7 produce isolated islands of particulates on the FTO
the subsequent addition of HOAc to this reaction mixture surface, whereas stirring affects the size and density of the islands
facilitates the rapid reaction of unreacted Ti(OiPr)4 to produce (Fig. 6a). While stirring solutions over the Rf ¼ 7–10 range
aggregated particulates that stunt the growth of nanostructures improves surface coverage, a uniform bed of wires is not
at the substrate surface. Conversely, Ri > 2 favors the rapid observed at Rf < 14. This observation is presumably because the
hydrolysis of Ti(OiPr)4 resulting in large aggregated wire-like perturbation of the solution reduces the residence time of the
clusters that also lead to poorly covered FTO. In effect, values of macromolecular precursors near the FTO surface. Poor surface
Ri that deviate substantially from 1.3 lead to similar results; i.e., coverage was observed at Rf ¼ 20 (not shown), which indicates
incomplete surface coverage and non-uniformity of TiO2 nano- that a specific range of acid ratio is, indeed, required. The
structures. unperturbed solutions show a lower sensitivity to Rf value; i.e.,
a uniform coverage of nanowires is obtained within the Rf ¼ 8.5–
14 range. It was determined that the parameters Ri ¼ 1.3 and Rf
Growth of nanowires: role of Rf
¼ 8.5–14, with no stirring after introduction of substrate,
With the reaction conditions affording uniform surface coverage produced the best reaction conditions for nanowire nucleation
of the nucleation sites established, the effects of the subsequent and growth. A higher density of nanowires is observed at higher
acid addition (expressed by Rf) on the TiO2 film morphologies Rf values, presumably due to the static solution favoring the self-
were investigated. Recall that Rf represents the molar ratio of assembly of macromolecular precursors.
HOAc to Ti(OiPr)4 over the entire reaction sequence. Experi- The importance of the volume of solvent on the formation of
ments were carried out over the Rf ¼ 1–20 range using substrates the nanostructured films was also examined by conducting
treated with 0.5 M NaOH and an Ri value of 1.3; the excess a series of experiments where Ri ¼ 1.3 and Rf ¼ 14 using 0, 5, 10,
HOAc addition was delivered 10 min after the nucleation phase 20, 30 and 40 mL of anhydrous heptane, respectively. This
was initiated. SEM images of a representative film are presented procedure was performed with stirring in order to maintain
in Fig. 6. homogeneity of the reagents in solution. Representative images

5066 | J. Mater. Chem., 2010, 20, 5063–5069 This journal is ª The Royal Society of Chemistry 2010
View Article Online

of sintered films are collected in Fig. S3, ESI†. Ideal wire growth
was observed when the volume of heptane was set at 20–30 mL.
Solvent volumes of less than 20 mL resulted in a collection of
bundles of agglomerated wires and other particulates on the FTO
surface. Excess heptane (> 30 mL) improves the dispersion of the
wires, but also resulted in poor FTO surface coverage. Although
the scale of the reaction likely dictates the optimum volume of
heptane, the experiments in this study were all carried out in
20 mL of heptane for the sake of brevity.
Published on 20 May 2010. Downloaded by University of Michigan Library on 28/10/2014 04:42:11.

Physical characterization
Fig. 8 (a) Powder X-ray diffraction patterns of bulk anatase TiO2
The nanowires grown from the substrate, as well as those grown nanowires obtained at varying Rf values (Ri ¼ 1.3). The diffraction
concomitantly in the bulk phase of the reaction vessel, were pattern for 20 nm spherical anatase TiO2 nanoparticles is provided for
evaluated by extensive electron microscopy and powder XRD comparison. (b) Plot of the ratio of anatase (004) and (200) diffraction
experiments. TEM images of the bulk TiO2 nanowires at Ri ¼ 1.3 peaks indicating anisotropic growth along the c-axis of the crystallites.
and Rf > 8.5 reveal the amorphous nature of the linear macro-
molecules prior to annealing (Fig. 7a, inset), which is verified by peaks (located at 2q ¼ 37.8 and 48.0 , respectively) suggest
powder XRD measurements. Sintering the smooth linear struc- anisotropic growth of the crystallites along the a and c axes on
tures at 450  C for 2 h results in anatase TiO2 nanowires with the basis that a higher ratio corresponds to axial growth.46 As
relatively uniform morphologies; e.g., widths of 54  10 nm and a benchmark comparison, 20 nm spherical anatase TiO2 nano-
lengths in excess of 1 mm (Fig. 7). The TEM images indicate that particles (Alfa Aesar 10 nm APS Powder, S.A. 100–130 m2 g1)
the nanowires are made up of anatase crystallites that are 16  exhibit a (004) to (200) ratio of 0.52, while high aspect ratio
3 nm in diameter (Fig. 7a). The active surface area of the wires,
which was determined by spectrophotometrically monitoring the
adsorption of dye molecules to the TiO2 films, was measured to
be 3  109 moles per cm2 – a value that is resonant with other
low-density nanowire films.3,9
Powder XRD patterns were obtained for nanowires in the bulk
phase (in the absence of an FTO substrate) over Rf ¼ 5–14 to
gain insight into the growth process. These materials display
behavior that is consistent with the thin-film data (Fig. S4, ESI†):
the nanowires are amorphous prior to sintering and take on the
anatase form of TiO2 after calcination at 450  C (characteristic
diffraction peaks are observed at 2q ¼ 25.3 , 37.8 , and 48.0
corresponding to the (101), (004), and (200) crystal planes,
respectively; Fig. 7a). Analysis of the most intense diffraction
peak located at 2q z 25 (i.e., the anatase (101) diffraction peak)
revealed that the average crystallite sizes were on the order of 11–
12 nm (according to the Scherrer equation19), and sensitive to the
Rf value. A plot of the average crystallite size versus Rf reveals
a decreasing trend with a slope of 0.16 nm/Rf (not shown),
while the relative intensities of the (004) and (200) diffraction

Fig. 7 (a) TEM image of anatase TiO2 nanowires generated in the bulk Fig. 9 (a) SAED diffraction pattern and corresponding HRTEM image
phase before (inset) and after calcination. (b) Powder XRD pattern of (inset) of anatase TiO2 nanocrystals (Ri ¼ 1.3, Rf ¼ 14). (b) Higher
anatase TiO2 nanowires (bulk phase). Vertical bars represent the stan- magnification HRTEM image of a crystallite and its fast fourier trans-
dard diffraction pattern of anatase TiO2 (JCPDS No. 21-1272). form (inset) highlighting the anatase diffraction pattern.

This journal is ª The Royal Society of Chemistry 2010 J. Mater. Chem., 2010, 20, 5063–5069 | 5067
View Article Online

nanowires have been shown to exhibit a (004) to (200) ratio of received. FTO glass substrates (F:SnO2; Tec 8; 8 U cm2) were
>0.70 (Fig. 8b).46 Analysis of materials produced at Rf ¼ 5 purchased from the Hartford Glass Company. All manipulations
reveals a (004) to (200) ratio of 0.69 indicating only a slight were carried out under ambient atmosphere unless otherwise
preference for crystallite growth along the c-axis; however, the specified.
(004) to (200) ratio is found to increase to 1.06 as Rf approaches
14. This feature indicates a preference for axial growth as Rf Preparation of substrates
increases.
The crystalline phase of the TiO2 nanowires grown directly FTO glass substrates approximately 2 cm  2 cm in size were
from the substrate were detached and characterized by HRTEM subject to consecutive ultrasonic treatments in Alconox
Published on 20 May 2010. Downloaded by University of Michigan Library on 28/10/2014 04:42:11.

(Fig. 9). The selected-area electron-diffraction (SAED) ring detergent for 15 min, deionized H2O for 5 min, and ethanol for
patterns confirm that the nanowires are composed of poly- 15 min. The substrates were exposed to UV light and O3(g) for
crystalline anatase TiO2; i.e., the Debye–Scherrer concentric 15 min, and then submerged in 0.5 M NaOH(aq) overnight. The
rings of (101), (004), (200), (105), (204), (220), and (215) planes in substrates were removed from the solution, air dried for 15 min,
Fig. 9a match the atomic spacing of the anatase crystal lattice. and used immediately.
Powder XRD techniques also corroborated the anatase crystal-
line phase of the films on the surface of the FTO slides based on Preparation of TiO2 nanostructures on FTO Glass
the observation of 2q ¼ 25.3 and 48.0 diffraction peaks arising
Base functionalized FTO substrate was placed in a dry 50 mL
from the (101) and (200) crystal planes.
round bottom flask charged with 20 mL of anhydrous heptane
and heated to 60  C. To the flask was added stepwise 1.30–
Conclusions 1.68 mL of HOAc (23.0–29.2 mmol; Ri ¼ 1–1.3) and 6.75 mL of
This study establishes the conditions necessary to grow anatase Ti(OiPr)4 (22.0 mmol) (Table S1, ESI†). After the reaction flask
TiO2 nanowires from the surface of transparent conducting FTO was left to stand for 10 min, 4.90–17.0 ml of HOAc (92.0–
glass substrates. The procedure presented herein requires base 296.0 mmol; Rf ¼ 5–14) was added followed by gentle stirring
pretreatment of the substrate to facilitate nucleation and growth until the solution became translucent. After the reagents were
of nanostructures on the surface. While nanowire growth is stirred for 10 min, the magnetic stirrer was stopped (in some
sensitive to a number of factors, we tentatively assign a growth experiments; refer to Table S1 for details†) and the reaction
mechanism that requires, as a first step, the binding of Ti-oxide mixture was left to stand at 60  C to facilitate gelation (4 h) and
molecular clusters to the base-treated surface. A series of aging (2 d). The substrates were then dried under reduced pres-
subsequent hydrolysis steps facilitate the axial growth of linear sure for 12 h, blown free of debris using compressed air and
macromolecules normal to the plane of the FTO crystal faces. calcined at 450  C for 2 h (ramp rate of 5  C min1).
The linear nature of these structures is retained after calcination
resulting in a thin film of anatase TiO2 nanowires tethered to the Dye loading studies
FTO surface. Extensive imaging and XRD diffraction studies
The TiO2 films were immersed in a 1.0 M EtOH solution of the
reveal that the nanowires consist of a collection of anisotropic
Ru(2,2-bipyridyl-4,40 -dicarboxylate)2(NCS)2 (Solaronix) for
crystallites that exhibit a preference for axial growth.
12 h. The dye was then desorbed from the TiO2 film by soaking
The relative quantities of HOAc and Ti(OiPr)4 appear to
for 30 min in 1.0 M NaOH. The absorbance band at ca. 500 nm
govern the nucleation and growth processes. While we have
was used to determine the amount of desorbed dye against
found that there is a strong correlation of nucleation to Ri and
a calibration curve ranging from 6.0  105 M to 6.0  108 M in
growth to Rf, there is clearly interplay between the two param-
1.0 M NaOH.
eters. Of the reaction conditions explored in this investigation,
the best anatase TiO2 films were achieved at Ri ¼ 1.3 and Rf ¼
8.5–14. Films produced at Ri s 1.3 led to poor coverage of FTO Physical methods
by TiO2 nanostructures regardless of Rf value. While the wire UV-ozone treatment of the glass substrates was carried out using
morphology was similar over the Rf ¼ 8.5  14 range, the XRD a PSD-UVO3 (Novascan Technologies) system. The powder
data indicate that a higher Rf value enhances the anisotropy of XRD patterns of the films were recorded using a Rigaku Mini-
the crystallites. Lower quality films were obtained at Rf ¼ 20, Flex II Desktop diffractometer (scan speed ¼ 2.000 min1, Cu
thereby indicating the relationship between Rf value, crystallite  Morphological and lattice struc-
Ka radiation, l ¼ 1.5406 A).
anisotropy and film quality should be considered in the further tural information were examined by tungsten-filament SEM
development of this chemistry. Studies are currently underway to (FEI XL 30, accelerating voltage 20 kV) and tungsten-filament
examine how other molecular precursors affect the growth of TEM (Hitachi H7650, AMT 16000 digital camera, accelerating
these wires, as well as an investigation defining the electrical voltage 80 kV). SEM imaging of cross sections was performed on
properties of this class of materials. a Hitachi S4800 with an accelerating voltage of 30 kV. HRTEM
imaging and SAED patterns were collected on an FEI Tecnai
Experimental F20 FEG TEM using an accelerating voltage of 200 kV. Contact
mode AFM was performed on a Molecular Imaging PicoSPM
Materials
instrument with a scan rate of 0.25 Hz. Samples for cross-
Materials Ti(OiPr)4 (97%), AcOH (99.7%) and anhydrous sectional TEM imaging were embedded in an EPON resin
heptane (99%) were purchased from Sigma Aldrich and used as mixture and cured at 60  C, trimmed for sectioning (70–90 nm,

5068 | J. Mater. Chem., 2010, 20, 5063–5069 This journal is ª The Royal Society of Chemistry 2010
View Article Online

Reichert-Jung Ultracut E microtome with a diamond knife) and 16 I. C. V. Baek, M., J. A. Chang, J.-H. Yum, M. K. Nazeeruddin,
collected on single-hole grids with Formvar supporting film. M. Gr€atzel, Y.-C. Chung and S. I. Seok, Electrochem. Commun.,
2009, 11, 909–912.
Alternatively, TEM samples were prepared by detaching the 17 R. L. Penn and J. F. Banfield, Geochim. Cosmochim. Acta, 1999, 63,
TiO2 nanowires from the substrate and transferred to a holey 1549–1557.
carbon TEM grid. 18 Z. Miao, D. Xu, J. Ouyang, G. Guo, X. Zhao and Y. Tang, Nano
Lett., 2002, 2, 717–720.
19 M. Niederberger, H. J. Muhr, F. Krumeich, F. Bieri, D. Guenther and
R. Nesper, Chem. Mater., 2000, 12, 1995–2000.
Acknowledgements 20 C.-T. Dinh, T.-D. Nguyen, F. Kleitz and T.-O. Do, ACS Nano, 2009,
3, 3737–3743.
Published on 20 May 2010. Downloaded by University of Michigan Library on 28/10/2014 04:42:11.

The authors are grateful to Drs W. M. Schoel, V. Thangadurai, 21 L. s. Li, J. Hu, W. Yang and A. P. Alivisatos, Nano Lett., 2001, 1,
S. S. Bella, W.-X. Dong, Ira Probodh, R. Vaidhyanathan, 349–351.
S. Park, M. Sajjadi and J. Grant (University of Calgary) and 22 R. Sui, A. S. Rizkalla and P. A. Charpentier, Langmuir, 2005, 21,
6150–6153.
Dr J. Buriak (National Institute for Nanotechnology) for
23 C. Xiong and K. J. Balkus, Chem. Mater., 2005, 17, 5136–5140.
consultation and experimental assistance. This work was finan- 24 T. J. Trentler, T. E. Denler, J. F. Bertone, A. Agrawal and
cially supported by the Canadian Natural Science and Engi- V. L. Colvin, J. Am. Chem. Soc., 1999, 121, 1613–1614.
neering Research Council (NSERC), Canada Research Chairs, 25 C.-C. Wang and J. Y. Ying, Chem. Mater., 1999, 11, 3113–3120.
26 T. Kasuga, M. Hiramatsu, A. Hoson, T. Sekino and K. Niihara,
Canadian Foundation for Innovation (CFI), and the Canada Langmuir, 1998, 14, 3160–3163.
School of Energy and Environment (CSEE). This work utilized 27 X. Feng, K. Shankar, O. K. Varghese, M. Paulose, T. J. Latempa and
the University of Calgary Microscopy Facilities, which receive C. A. Grimes, Nano Lett., 2008, 8, 3781–3786.
support from the Canadian Foundation for Innovation and the 28 V. G. Pol and A. Zaban, J. Phys. Chem. C, 2007, 111, 14574–14578.
29 W. Wang, H. Lin, J. Li and N. Wang, J. Am. Ceram. Soc., 2008, 91,
Alberta Science and Research Authority. 628–631.
30 A. S. Maria Chong, L. K. Tan, J. Deng and H. Gao, Adv. Funct.
Mater., 2007, 17, 1629–1635.
Notes and References 31 J.-H. Lee, I.-C. Leu, M.-C. Hsu, Y.-W. Chung and M.-H. Hon,
J. Phys. Chem. B, 2005, 109, 13056–13059.
1 A. I. Hochbaum and P. Yang, Chem. Rev., 2010, 110, 527–546. 32 B. Xiang, Y. Zhang, Z. Wang, X. H. Luo, Y. Zhu, H. Z. Zhang and
2 Y. Xia, P. Yang, Y. Sun, Y. Wu, B. Mayers, B. Gates, Y. Yin, F. Kim D. P. Yu, J. Phys. D: Appl. Phys., 2005, 38, 1152–1155.
and H. Yan, Adv. Mater., 2003, 15, 353–389. 33 D. Gong, C. A. Grimes, O. K. Varghese, W. Hu, R. S. Singh, Z. Chen
3 X. Chen and S. S. Mao, Chem. Rev., 2007, 107, 2891–2959. and E. C. Dickey, J. Mater. Res., 2001, 16, 3331–3334.
4 G. K. Mor, S. Kim, M. Paulose, O. K. Varghese, K. Shankar, 34 B. Liu and E. S. Aydil, J. Am. Chem. Soc., 2009, 131, 3985–3990.
J. Basham and C. A. Grimes, Nano Lett., 2009, 9, 4250–4257. 35 M. Gr€atzel, Chem. Lett., 2005, 34, 8–13.
5 B. O’Regan and M. Graetzel, Nature, 1991, 353, 737–740. 36 S. K. Deb, Sol. Energy Mater. Sol. Cells, 2005, 88, 1–10.
6 M. Gr€ atzel, Nature, 2001, 414, 338–344. 37 R. Sui, J. L. Young and C. P. Berlinguette, J. Mater. Chem., 2010, 20,
7 A. B. F. Martinson, J. W. Elam, J. Liu, M. J. Pellin, T. J. Marks and 498–503.
J. T. Hupp, Nano Lett., 2008, 8, 2862–2866. 38 R. Sui, V. Thangadurai and C. P. Berlinguette, Chem. Mater., 2008,
8 A. B. F. Martinson, M. R. S. Goles, F. Fabregat-Santiago, 20, 7022–7030.
J. Bisquert, M. J. Pellin and J. T. Hupp, J. Phys. Chem. A, 2009, 39 U. Diebold, Appl. Phys. A: Mater. Sci. Process., 2003, 76, 681–687.
113, 4015–4021. 40 C. J. Howard, T. M. Sabine and F. Dickson, Acta Crystallogr., Sect.
9 M. Law, L. E. Greene, J. C. Johnson, R. Saykally and P. Yang, Nat. B: Struct. Sci., 1991, 47, 462–468.
Mater., 2005, 4, 455–459. 41 M. Wu, G. Lin, D. Chen, G. Wang, D. He, S. Feng and R. Xu, Chem.
10 M. Adachi, Y. Murata, J. Takao, J. Jiu, M. Sakamoto and F. Wang, Mater., 2002, 14, 1974–1980.
J. Am. Chem. Soc., 2004, 126, 14943–14949. 42 Z. Wang, D. Xia, G. Chen, T. Yang and Y. Chen, Mater. Chem.
11 G. K. Mor, O. K. Varghese, M. Paulose, K. Shankar and Phys., 2008, 111, 313–316.
C. A. Grimes, Sol. Energy Mater. Sol. Cells, 2006, 90, 2011–2075. 43 M. Abd-Lefdil, R. Diaz, H. Bihri, M. A. Aouaj and F. Rueda, Eur.
12 W.-J. Li, E.-W. Shi, W.-Z. Zhong and Z.-W. Yin, J. Cryst. Growth, Phys. J.: Appl. Phys., 2007, 38, 217–219.
1999, 203, 186–196. 44 A. Rammal, F. Brisach and M. Henry, C. R. Chim., 2002, 5, 59–
13 T. Maruo, N. Ueno, S. Ichikawa, N. Nishiyama, Y. Egashira and 66.
K. Ueyama, Mater. Lett., 2009, 63, 2373–2376. 45 T. J. Boyle, R. P. Tyner, T. M. Alam, B. L. Scott, J. W. Ziller and
14 K. Keis, E. Magnusson, H. Lindstr€ om, S.-E. Lindquist and B. G. Potter, J. Am. Chem. Soc., 1999, 121, 12104–12112.
A. Hagfeldt, Sol. Energy Mater. Sol. Cells, 2002, 73, 51–58. 46 J. Jiu, S. Isoda, F. Wang and M. Adachi, J. Phys. Chem. B, 2006, 110,
15 M. Gratzel, J. Photochem. Photobiol., A, 2004, 164, 3–14. 2087–2092.

This journal is ª The Royal Society of Chemistry 2010 J. Mater. Chem., 2010, 20, 5063–5069 | 5069

You might also like