Application of semiconductors in functional optoelectronic devices requires precise control over their doping and formation of junction between p- and n-doped semiconductors. While doped thin films have led to several semiconductor devices, need for high-surface area nanostructured devices for photovoltaic, photoelectrochemical, and photocatalytic applications has been hindered by lack of desired doping in nanostructures. Here, we show titanium-dioxide (TiO2) nanotubes doped with nitrogen (N) and niobium (Nb) as acceptors and donors, respectively, and formation of TiO2 nanotubes p-n homojunction. This TiO2:N/TiO2:Nb homojunction showed distinct diode-like behaviour with rectification ratio of 1115 at ±5 V and exhibited good photoresponse for ultraviolet light (λ = 365 nm) with sensitivity of 0.19 A/W at reverse bias of −5 V. These results can have important implications for development of nanostructured metal-oxide solar-cells, photodiodes, LED's, photocatalysts, and photoelectrochemical devices.

Titanium-dioxide (TiO2) is a wide-bandgap metal-oxide semiconductor with wide-ranging applications in solar cells, LED's, photocatalytic reactors, photoelectrochemical cells, etc.1–6 While desirable energy-level alignment of TiO2 electronic states and its chemical stability has led to several investigations in making nanostructured TiO2 devices, as-grown undoped TiO2 films show n-type doping due to oxygen vacancies which act as shallow donors.3,5–7 Several new investigations can be enabled by availability of desired p- and n-type nanostructured TiO2. But the difficulty in achieving desired p-type doping in these nanostructured films, due to self-compensation effects of shallow donor oxygen vacancies, has hindered progress. This has limited the investigations to heterojunction TiO2 semiconductor devices, with as-grown TiO2 nanostructures as nominal n-type material (charge carrier concentration is limited by synthesis method), along with another p-type semiconductor nanostructures. This approach limits the device architectures, choice of materials, device stability, and lacks precise control over the electronic states and the formation of semiconductor-heterojunction. Recently, we developed a new doping method using a modified electrochemical cell, which enabled growth of desired p-doped, n-doped, and co-doped TiO2 nanotubes, with wide range of optical, electronic, and magnetic dopants.8 Here, we show TiO2 nanotubes doped with Nb and N elements as donor and acceptor dopants.8–13 Addition of N as p-type dopants (from 0 to 1.2 wt. %) in nanotubes first decreases the n-type conductivity of as-grown TiO2 nanotubes (electrons are majority charge carriers), followed by conversion to p-type semiconductor (holes become the majority charge carriers) and a resulting increase in hole conduction. These electronic properties are supported by the measured shift of the Fermi-level in our scanning tunneling spectroscopy (STS) measurements with increasing N concentration. Using these desired p- and n-type TiO2 nanotubes, we show p-n homojunction TiO2 nanotubes. These nanostructured TiO2 homojunction devices demonstrate good rectification behavior (1115 at ±5 V), good photoresponse behavior, and can provide an important building block in future TiO2 nanoelectronic and catalytic devices.

TiO2 nanotubes were grown using electrochemical oxidation (anodization).8,12,13 In this method, platinum (Pt) counter-electrode and titanium (Ti) sheet (99.99% purity) were immersed in a modified electrochemical cell, and an alternating bias (rectangular-shaped voltage pulse) was applied.8,13 Electrolyte used in electrochemical growth solution consisted of solvent ethylene glycol with 1 wt. % ammonium fluoride (NH4F) and 2 wt. % water. Nb and N precursors, niobium (V) chloride and hexamethylenetetramine, were added during the growth in electrolyte. The N and Nb doping level was controlled by changing the amount of precursor (e.g., hexamethylenetetramine) in the electrolyte.8,13 All as-grown samples were amorphous and were annealed at 500 °C for 1 h in air to convert to anatase phase. The morphology of the grown TiO2 nanotubes was examined in field-emission scanning electron microscope (FE-SEM) and the well-defined tubular structure of samples was confirmed (representative images of nanotubes shown in Figure 1(a)). The crystal structure of undoped, Nb-, and N-doped TiO2 nanotube samples was analyzed by X-ray diffraction (XRD, Figure 1(b)), which showed anatase structure for samples, with diffraction peaks (101), (004), and (200) at 2θ = 25.3°, 37.85°, and 47.95°, respectively. Peak marked “Ti” in the XRD corresponds to the underlying Ti substrate from which TiO2 nanotubes were grown. No additional peaks were seen in the doped nanotubes XRD spectra, compared to undoped nanotubes, indicating no unwanted phase segregation and incorporation of dopants in the anatase TiO2 crystal lattice.

FIG. 1.

(a) Representative SEM and (b) X-ray diffraction pattern of doped and undoped TiO2 nanotube arrays grown using our modified electrochemical anodization method; (c)–(e) Current-voltage (I-V) characteristics and corresponding ln(I) vs V plots of (c) N-doped, (d) Nb-doped, and (e) undoped TiO2 nanotube samples. (f) Conductance (I/V) as a function of temperature (1/kT), for Nb-doped (curve 1) and N-doped (curve 2) TiO2 nanotubes. The slope at higher temperatures (>160 K) revealed activation energies 0.018 eV and 0.135 eV for Nb-doped and N-doped samples, respectively.

FIG. 1.

(a) Representative SEM and (b) X-ray diffraction pattern of doped and undoped TiO2 nanotube arrays grown using our modified electrochemical anodization method; (c)–(e) Current-voltage (I-V) characteristics and corresponding ln(I) vs V plots of (c) N-doped, (d) Nb-doped, and (e) undoped TiO2 nanotube samples. (f) Conductance (I/V) as a function of temperature (1/kT), for Nb-doped (curve 1) and N-doped (curve 2) TiO2 nanotubes. The slope at higher temperatures (>160 K) revealed activation energies 0.018 eV and 0.135 eV for Nb-doped and N-doped samples, respectively.

Close modal

Electronic properties of the doped TiO2 nanotubes were studied using current-voltage (I-V) characteristics. Using room temperature I-V characteristics for undoped, N-doped, and Nb-doped TiO2 nanotubes (Figures 1(c)–1(e)), we extracted the charge carrier concentration (n) and mobility (μ) of these nanotubes. Charge carrier concentration was estimated using the intermediate bias regime, where the reverse-biased Schottky-barrier dominates the total current:8,14,15ln(I)=ln(S)+e(1kT1E0)V+lnJs: JS is the current density through the Schottky barrier, S is the contact area associated with this barrier, E0 is a parameter depending on carrier concentration n, e is electron charge, and k is the Boltzmann constant. The slope of ln(I) vs V is equal to e(1kT1E0), where E0=E00coth(E00/kT), and E00=(e/2)(n/m*ε)1/2, h is the Plank's constant, ε0 is vacuum permittivity, ε=31ε0 is the dielectric constant for TiO2, and m*=m0 is the electron effective mass. More details of these calculations can be found elsewhere.8,14,15 The n and μ values for undoped, N-doped, and Nb-doped TiO2 nanotubes were found to be 9 × 1017 cm−3 and 4.3 cm2/V·s; 3.6 × 1017 cm−3 and 3.8 cm2/V·s; and 5 × 1019 cm−3 and 0.097 cm2/V·s, respectively. Comparing the electronic properties of undoped and N-, Nb-doped nanotubes, the charge carrier concentration for Nb-doped TiO2 nanotubes was about two orders of magnitude higher than the undoped (9 × 1017 cm−3) and N-doped (3.6 × 1017 cm−3) nanotubes, and the mobility is reduced compared to these nanotubes. Similar I-V behavior of undoped and N-doped samples could be explained by similar charge carrier concentrations (electrons in undoped and holes in N-doped TiO2), since slope of ln(I) vs V determines the charge carrier concentration of the samples.8,14,15 Acceptor and donor natures of N- and Nb-dopants in TiO2 nanotubes were also confirmed from STS and Current Sensing Atomic Force Microscopy (CS-AFM) studies (Figure S1).8,16

Temperature dependent current-voltage characterization (I-V-T) was performed in temperature range of 20 K–400 K, to estimate the donor and acceptor activation energies. The conductivity σ = ln(I/V) was plotted as a function of 1/kT (Figure 1(f)) to measure the activation barrier for charge transport (or energy required to ionize charge carriers in our doped semiconductor). Activation energies (Ea) for Nb and N were found to be 0.027 eV and 0.135 eV, respectively. The activation energy for Nb-doped TiO2 nanotubes was close to the reported ionization energy values for Nb donors in anatase TiO2 thin films and bulk crystals, which was in the range of 10–40 meV.17,18 Nitrogen is known as a good acceptor in TiO2,9,11 with reported energy position 0.14 eV above the top of valence band.19 These measured activation energies were used to estimate the fraction of dopant ions which are thermalized at room temperature (to generate charge carriers for conduction), and hence balance the electron and hole charge carrier concentrations for forming a p-n homojunction TiO2 diode.

We fabricated p-n diodes using respective doped TiO2 nanotubes and measured their I-V characteristics. These measurements were performed using a Keithley source meter (Keithley 2612A, Tektronix, Inc.) and home-made probe station. The schematic of the experimental setup is shown in Figure 2. The TiO2 homojunctions were formed by contacting p-n, n-n, and p-p junction diodes (schematic shown in Figures 3(a)–3(c)). Square-shaped 2 × 2 mm2 size TiO2 nanotubes grown on titanium foils (p-type or n-type) were cut and mechanically pressed using the probes onto a bigger 12.7 × 25.4 mm2 sized nanotube substrates (n-type or p-type depending whether p-n, or p-p, or n-n structure to be fabricated). Since nanotubes grow on both surfaces of Ti sheet, one side was cleaned by mechanically removing the nanotubes (using a blade) and used for electrical contact. Figure 3(d) shows I-V characteristics for the p-n TiO2:N/TiO2:Nb structure, which clearly shows the characteristic rectifying diode behavior (asymmetric curves). The forward current Ifwd and reverse bias current Irev at ±5 V were 2.9 × 10−3 A and 2.6 × 10−6 A, respectively, resulting in a rectification ratio Ifwd/Irev = 1115. In contrast, I-V characteristics of n-n and p-p structures, shown in Figures 3(e) and 3(f), were nearly symmetric, without any rectifying behavior. This provides confirmation for the formation of the desired p-n homojunction in TiO2:N/TiO2:Nb nanotubes. Slightly non-linear behavior of I-V characteristics of N-doped sample in both positive and negative bias (Figure 3(f)) can be explained by space-charge-limited conduction observed in these doped nanotubes.20 

FIG. 2.

Schematic showing experimental setup used to measure current-voltage characteristics of p-TiO2/n-TiO2, p-TiO2/p-TiO2, and n-TiO2/n-TiO2 diode structures.

FIG. 2.

Schematic showing experimental setup used to measure current-voltage characteristics of p-TiO2/n-TiO2, p-TiO2/p-TiO2, and n-TiO2/n-TiO2 diode structures.

Close modal
FIG. 3.

Schematics of the p-n, n-n, and p-p diodes ((a)–(c)) structures and their corresponding I-V characteristics ((d)–(f)). Comparison shows clear rectifying diode-like behavior for p-n TiO2:N/TiO2:Nb homojunction diode.

FIG. 3.

Schematics of the p-n, n-n, and p-p diodes ((a)–(c)) structures and their corresponding I-V characteristics ((d)–(f)). Comparison shows clear rectifying diode-like behavior for p-n TiO2:N/TiO2:Nb homojunction diode.

Close modal

Photoresponse properties of the p-n TiO2:N/TiO2:Nb diodes were studied by irradiating the homojunction with 365 nm wavelength ultraviolet (UV) light (above TiO2 nanotube bandgap). A significant photoresponse was observed in the I-V characteristics. Figure 4 shows I-V curves before (dark current) and during (photocurrent) light irradiation. The reverse current increased by two orders of magnitude upon light irradiation (photodetector). Photosensitivity (η) was calculated using equation: η=IP, where I—photocurrent = Ilight-Idark and P—incident light intensity on the nanotube photodiode. Using the light intensity (5.1 mW/cm2) and contact area (estimated using the density, and inner and outer diameters of the hollow nanotubes measured using SEM), the photosensitivity for the p-n diode at −5 V was estimated as η = 0.19 A/W. Considering the light-scattering in TiO2 nanotubes, indirect bandgap of the semiconductor, and incomplete light absorption near the bandedge, good photodiode sensitivity and the high-rectification factor indicate promising optoelectronic applications for doped-TiO2 homojunction devices.

FIG. 4.

Photoresponse I-V characteristics of TiO2:N/TiO2:Nb p-n diode under illumination with 365 nm wavelength UV light. Two orders-of-magnitude increase in current at reverse bias was observed on light-illumination (photodetection).

FIG. 4.

Photoresponse I-V characteristics of TiO2:N/TiO2:Nb p-n diode under illumination with 365 nm wavelength UV light. Two orders-of-magnitude increase in current at reverse bias was observed on light-illumination (photodetection).

Close modal

In conclusion, we demonstrated a p-n TiO2 nanotube homojunction using TiO2:N/TiO2:Nb doped nanotubes. The nanotubes electronic properties were easily tailored using a facile electrochemical doping method, and the homojunction p-n TiO2 nanotube diode showed good rectifying diode-like behavior, and excellent reversed-bias photoresponse (photosensitivity 0.19 A/W). Considering their chemical stability, optimal alignment of the TiO2 energy levels, and their wide-ranging applications for nanostructured solar-cells, photodiodes, LED's, photocatalysts, and photoelectrochemical cells, these results can have important implications for development of nanostructured metal-oxide optoelectronic and catalytic devices.

1.
B.
Oregan
and
M.
Gratzel
,
Nature
353
,
737
(
1991
).
2.
X. B.
Chen
,
L.
Liu
,
P. Y.
Yu
, and
S. S.
Mao
,
Science
331
,
746
(
2011
).
3.
A. L.
Linsebigler
,
G. Q.
Lu
, and
J. T.
Yates
,
Chem. Rev.
95
,
735
(
1995
).
4.
A.
Fujishima
and
M.
Honda
,
Nature
238
,
37
(
1972
).
5.
Y. C.
Nah
,
I.
Paramasivam
, and
P.
Schmuki
,
ChemPhysChem
11
,
2698
(
2010
).
6.
V.
Singh
,
I. J. C.
Beltran
,
J. C.
Ribot
, and
P.
Nagpal
,
Nano Lett.
14
,
597
(
2014
).
7.
Q. J.
Xiang
and
J. G.
Yu
,
J. Phys. Chem. Lett.
4
,
753
(
2013
).
8.
Y.
Alivov
,
V.
Singh
,
Y.
Ding
,
L. J.
Cerkovnik
, and
P.
Nagpal
,
Nanoscale
6
,
10839
(
2014
).
9.
A. K.
Chandiran
,
F.
Sauvage
,
M.
Casas-Cabanas
,
P.
Comte
,
S. M.
Zakeeruddin
, and
M.
Graetzel
,
J. Phys. Chem. C
114
,
15849
(
2010
).
10.
T.
Hitosugi
,
A.
Ueda
,
S.
Nakao
,
N.
Yamada
,
Y.
Furubayashi
,
Y.
Hirose
,
T.
Shimada
, and
T.
Hasegawa
,
Appl. Phys. Lett.
90
,
212106
(
2007
).
11.
W. G.
Zhu
,
X. F.
Qiu
,
V.
Iancu
,
X. Q.
Chen
,
H.
Pan
,
W.
Wang
,
N. M.
Dimitrijevic
,
T.
Rajh
,
H. M.
Meyer
,
M. P.
Paranthaman
,
G. M.
Stocks
,
H. H.
Weitering
,
B. H.
Gu
,
G.
Eres
, and
Z. Y.
Zhang
,
Phys. Rev. Lett.
103
,
226401
(
2009
).
12.
Y.
Alivov
,
M.
Pandikunta
,
S.
Nikishin
, and
Z. Y.
Fan
,
Nanotechnology
20
,
225602
(
2009
).
13.
Y.
Alivov
,
V.
Singh
,
Y.
Ding
, and
P.
Nagpal
,
Nanotechnology
25
,
385202
(
2014
).
14.
Z. Y.
Zhang
,
C. H.
Jin
,
X. L.
Liang
,
Q.
Chen
, and
L. M.
Peng
,
Appl. Phys. Lett.
88
,
073102
(
2006
).
15.
X. D.
Bai
,
D.
Golberg
,
Y.
Bando
,
C. Y.
Zhi
,
C. C.
Tang
,
M.
Mitome
, and
K.
Kurashima
,
Nano Lett.
7
,
632
(
2007
).
16.
See supplementary material at http://dx.doi.org/10.1063/1.4905218 for STS and CS-AFM data.
17.
J.
Osorio-Guillen
,
S.
Lany
, and
A.
Zunger
,
Phys. Rev. Lett.
100
,
36601
(
2008
).
18.
Y. Q.
Gai
,
J. B.
Li
,
S. S.
Li
,
J. B.
Xia
, and
S. H.
Wei
,
Phys. Rev. Lett.
102
,
036402
(
2009
).
19.
C.
Di Valentin
,
E.
Finazzi
,
G.
Pacchioni
,
A.
Selloni
,
S.
Livraghi
,
M. C.
Paganini
, and
E.
Giamello
,
Chem. Phys.
339
,
44
(
2007
).
20.
S. M.
Sze
,
Physics of Semiconductor Devices
, 2nd ed. (
Wiley-Interscience
,
2007
).

Supplementary Material