Pure α-Fe2O3 nanobelts are
configured as field effect transistors and
electrical transport studies demonstrate their
n-type behavior. In order to control the
electrical properties of the fabricated
transistor, the nanobelt channels are doped with
zinc element. Depending on the doping condition,
α-Fe2O3 nanobelts can be modified to either p-type
or n-type with enhanced conductivity and electron
mobility. This behavior change is exhibited in the
variation of I-V and I-Vg characteristics.
Quasi-one-dimensional (Q1D)
materials, such as nanotubes and nanowires, are
considered as highly promising nanoscale building
blocks for integrated electronic and photonic
circuits. 1,2 In this regard, the control of
electron (n-type) and hole (p-type) doping in
these nanostructures is of paramount importance.3
The doping approach is usually implemented by
incorporating impurity elements during the
synthesis procedures.4 In our work, α-Fe2O3
nanobelts are configured as field effect
transistors (FETs), followed by a controlled
in-situ doping method using zinc (Zn) as the
impurity element to achieve p- or enhanced n- type
semiconducting property. Carrier concentrations
and mobilities are obtained from electrical
transport studies. Furthermore, the mechanism of
p- and n-type doping using only one impurity
element is discussed.
Large-area vertically aligned
α-Fe2O3 nanobelts and nanowires have been
successfully synthesized via thermal oxidation of
iron substrate under appropriate flow of oxygen.5
It was observed that there was a morphology
transition from nanobelt to nanowire when
synthesis temperature was increased from 400 °C to
800 °C. In this work, the α-Fe2O3 nanobelts grown
at 700 °C were configured as field effect
transistors in order to characterize their
electrical properties and also explore their
potential application for integrated
nanoelectronics. The device fabrication process
can be briefly described as following. The
as-synthesized nanobelts suspended in isopropyl
alcohol were dropped onto a degenerately doped
p-type silicon substrate capped with 200 nm oxide
layer. Photoresist was then spin-coated onto the
substrate and photolithography was performed to
define an array of 100 mm2 pads. Finally, 10 nm
thick Ni and 100 nm thick Au are evaporated in
sequence onto this substrate, forming the
electrical contacts to the nanobelt. Individual
nanobelts with good contacts on both ends were
located with a high magnification optical
microscope or scanning electron microscope (SEM).
A SEM micrograph is shown in Fig. 1b.
Consequently, Fe2O3 nanobelt FETs were obtained
with metal contacts functioning as the source and
drain electrodes, and Si substrate acting as the
back gate, as illustrated in Fig. 1(c).
An array of vertically aligned
α-Fe2O3 nanobelts. (b) An individual α-Fe2O3
nanobelt (inset) is contacted by a pair of the
electrode pads. Scale bar: 3 μm. (c) A schematic
of nanobelt FET with metal contacts functioning as
the source and drain, and Si as the back gate.
Electrical transport
measurements on nanobelt FETs were conducted under
room temperature and ambient condition. Fig.2
demonstrates the typical results for a FET with
nanobelt width a = 80 nm, height b = 38 nm and
length L = 6.4 mm. From the I-V characteristics
[Fig. 2(a)] obtained under gate voltages (Vg) of
10, 0 and -10 V, it can be clearly seen that the
conductance of the nanobelt increases
monotonically as gate potential increases. This
indicates that the as-grown α-Fe2O3 nanobelts are
n-type semiconductor. In addition, the linearity
of the I-V curves around Vds = 0 V suggests that
ohmic contacts are formed between Ni and Fe2O3
nanobelt, which can be attributed to the higher
work function of Fe2O3 (5.4 eV) 6 than Ni (5.2
eV), as depicted in the inset of Fig. 2(a). The
pico-ampere current level reveals low conductivity
of the nanobelts. Considering the geometry of this
specific sample, the electrical conductivity σ =
2.2×10-3 (Ωcm)-1 can be estimated from the I-V
characteristic at Vg = 0 V. Fig. 2(b) shows the
I-Vg curve of the nanobelt FET at Vds = 2.0 V. As
known, charge carrier concentration and field
effect mobility in a typical cylindrical nanowire
system with radius r can be expressed as:
7
where Vgt is the threshold gate
voltage, e is the electron charge, εr and h are
the relative dielectric constant and thickness of
gate oxide layer (εr = 3.9 for SiO2 8), L is the
channel length, respectively. Although nanobelt
does not have a cylindrical geometry, it is
reasonable to estimate the capacitance using a =
2r as a first order approximation. Vgt and
transconductance dI/dVg can be extrapolated from
the linear region (-6 V~+6 V) of I-Vg curve as Vgt
= -27.1 V and dI/dVg = 8.2 ×10-12 A/V. Using Eqs.
(1) and (2), the electron concentration is
estimated to be n = 1.59×108 cm-1 which
corresponds to a bulk concentration of 5.3×1018
cm-3, and the mobility is obtained as μe = 2.8
×10-3
cm2/V.s.
FIG 2. (a) I-V characteristics of
an α-Fe2O3 nanobelt FET obtained at back gate
potentials of 10 V, 0 V and -10 V. Inset: energy
band diagram of Ni - Fe2O3 nanobelt contact, f is
work function. (b) I-Vg curve of the nanobelt FET
obtained at 2.0 V drain-source bias.
The native n-type behavior for
semiconducting metal oxide nanostructures, such as
ZnO and In2O3 nanobelts, has been well-documented
and attributed to the oxygen vacancies.7,9
Although in some context α-Fe2O3 is regarded as a
charge-transfer insulator,10 it tends to be an
n-type semiconductor in the existence of oxygen
vacancy.11 In addition, α-Fe2O3 has demonstrated
its peculiar behavior of the n- to p- type
transition under certain condition due to its
narrower band gap (Eg = 2.2 eV) compared with
In2O3 (Eg =3.6 eV) and ZnO (Eg = 3.4 eV). This
phenomenon has triggered many research interests
and is of particular importance for gas sensing
study.11-13 For pure α-Fe2O3, the n- to p- type
transition has been ascribed for the formation of
surface inversion layer due to oxygen
adsorption.11,12 In this work, Zn was introduced
into α-Fe2O3 nanobelts as dopant, it was observed
that both n- and p- type behavior could be
achieved stably by using different doping
conditions.
The doping experiment was carried out inside a
thermal furnace, as illustrated in Fig. 3(a). The
chip with the nanobelt FETs was placed at the
center of the furnace. An open-end quartz vial
containing small amount of pure Zn powder (~10 mg,
99.99%, Alfa Aeasar) was placed 12 cm away at
upstream. The system was first purged with pure Ar
three times to evacuate air, and then maintained
in 760 torr Ar. During doping, Zn vapor was
transported via 100 sccm Ar carrier gas. In order
to obtain p-type nanobelts, the furnace
temperature at the sample was quickly ramped up to
700 °C in 3 minutes and held for 5 minutes
followed by 20 minutes cooling time.14 During this
process, the highest temperature at Zn source was
around 600 °C due to the temperature gradient
inside the furnace. The doped nanobelts were then
subject to electrical transport measurement and
the typical results are plotted in Fig. 3(b). The
I-V curves and the I-Vg curve in Fig. 3(b) suggest
that α-Fe2O3 nanobelts have been converted to
p-type semiconductor, since the increase of gate
voltage results in the decrease of channel
conductance. In addition, the electrical
conductivity increased to 0.27 (Ω.cm)-1, which is
two orders of magnitude greater than that before
Zn doping. Combining Eqs. (1), (2) and the
geometry of this specific nanobelt (a = 66 nm, b =
27 nm, L = 3.9 mm), the hole concentration and
mobility were estimated to be p = 1.2×1020 cm-3
and μp = 1.3 ×10-2 cm2/V.s, respectively. On the
other hand, it was observed that n-type α-Fe2O3
nanobelts with enhanced carrier concentration and
mobility could be also obtained by using different
doping conditions. In this case, doping was
conducted at a temperature of 350 °C at the sample
substrate, while Zn source temperature was at 250
°C, and the duration for doping was extended to
one hour. Fig. 3(c) demonstrates the typical
electrical transport results of the nanobelts
subjected to such doping process. The I-V curves
and the I-Vg curve indicate that the nanobelt is
an n-type semiconductor with much higher
conductivity. An estimation of the electrical
conductivity gives σ = 17.2 (Ω.cm)-1, which is
almost four order of magnitude higher than that of
the nanobelts before doping. In addition, electron
concentration and mobility are estimated to be n =
8.9×1019 cm-3 and μe = 3.2 ×10-1 cm2/V.s for this
sample.
FIG 3. (a) A schematic of
experimental set up for doping α-Fe2O3 nanobelt
with Zn source. (b) I-V curves and I-Vg curve
(inset) of the p-type α-Fe2O3 nanobelt FETs doped
in 700 °C for 5 minutes. (c) I-V curves and I-Vg
curve (inset) of the n-type α-Fe2O3 nanobelt FETs
doped in 350 °C for one hour.
As shown in the experiments,
doping α-Fe2O3 nanobelts with Zn does not require
very high temperature and long time as compared
with doping bulk α-Fe2O3 with Magnesium (Mg).14
This can be reasoned by considering two factors:
(1). Zn has smaller atomic radius (Rzn = 0.142 nm)
than Mg (Rmg = 0.145 nm) and Fe (RFe = 0.156 nm),
which renders higher diffusivity of Zn in α-Fe2O3
lattice; (2). Nanobelt has large surface-to-volume
ratio, which makes the diffusion from the surface
of the sample a more significant effect. The
electrical behavior, n- or p-type, is attributed
to temperature dependence of Zn doping process in
Fe2O3. At low temperature, the introduction of Zn
induces more oxygen vacancies to be formed, which
serve as electron donors thus contributing to an
enhanced n-type conductivity; at high temperature,
Zn2+ substitutionally replaces Fe3+, consequently
increases hole concentration, thus giving rise to
p-type behavior. The doping effect on the initial
n-type behavior changing to p-type or enhanced
n-type behavior also manifests itself in the
modification of the contact property, as observed
in the increasingly non-linear I-V curves shown in
Fig. 3. As mentioned before, the linear I-V curves
(Fig. 2a) of undoped samples indicate ohmic
contacts are formed between the Ni/Au electrode
and the Fe2O3 nanobelt. As the sample is
compensation-doped to p-type, the Fermi level (EF)
approaches to the top of the valence band (EV),
thus the work function of the p-type Fe2O3
nanobelt is significantly increased, yielding
large Schottky barrier for hole transport, as
depicted in Fig. 4a, and pronounced non-linearity
in I-V characteristics (Fig. 3b). On the other
hand, as the sample is doped with more donors
showing enhanced n-type behavior, the Fermi level
in the Fe2O3 nanobelt shifts readily towards the
conduction band (EC), thus the work function of
the sample decreases to be smaller than that of
the electrode, which accounts for the Schottky
contact for electron transport (Fig. 4b) and
non-linearity of I-V characteristics (Fig. 3c).
FIG 4. Energy diagrams of Ni-
Fe2O3 nanobelt contact for (a) p-type Zn doping
and (b) n-type Zn
doping.
3
Conclusion
In summary, controlled p- and
n-type doping of α-Fe2O3 nanobelts with Zn as
dopant was successfully achieved. Electrical
transport investigations demonstrate enhanced
charge carrier concentration and mobility. With
both p- and n-channel field effect transistors
fabricated, Fe2O3 nanomaterials can be utilized as
potential building blocks for future nanoscale
electronic and magnetoelectronic
devices.
References
1. X. Liu, C. Lee, C. Zhou,
and J. Han, Appl. Phys. Lett. 79, 3329 (2001).
2. C.J. Barrelet, A.B.
Greytak and C.M. Lieber, Nano Lett. 4, 1981
(2004).
3. Y. Huang, X. Duan, Y.
Cui, L. J. Lauhon, K.-H. Kim, and C. M. Lieber,
Science 294, 1313 (2001).
4. Y. Cui, X. Duan, J. Hu,
and C. M. Lieber, J. Phys. Chem. B 104, 5213
(2000).
5. X. Wen, S. Wang, Y. Ding,
Z. L. Wang, and S. Yang, J. Phys. Chem. B 109, 215
(2005).
6. V. E. Hendrich, and P. A.
Cox, Surface Science of Metal Oxides (Cambridge
University Press, Cambridge, UK, 1994).
7. Z. Fan, D. Wang, P.
Chang, W. Tseng, and J. G. Lu, Appl. Phys. Lett.
85, 5923 (2004).
8. S. M. Sze, Modern
Semiconductor Device Physics (John Wiley &
Sons, New York, 1998).
9. B. Lei, C. Li, D.
Zhang, T. Tang, and C. Zhou, Appl. Phys. A 79, 439
(2004).
10. G. Dräger, and W.
Czolbe, Phys. Rev. B 45, 8283 (1992).
11. A. Gurlo, M. Sahm, A. Oprea,
N. Barsa and U. Weimar, Sens. Actuator B 102, 291
(2004).
12. A. Gurlo, N. Bârsan, A.
Oprea, M. Sahm, T. sahm, and U. Weimar, Appl.
Phys. Lett. 85, 2280 (2004).
13. M. Warnes, F. F. Aplan,
and G. Simkovich, Solid State Ion. 12, 271
(1984).
14. R. F. G. Gardner, R. L.
Moss, and D. W. Tanner, Brit. J. Appl. Phys. 17,
55
(1966).
Fe2O3 Nanowires
