Received: 5 October 2016
|
Revised: 19 January 2017
|
Accepted: 20 January 2017
DOI 10.1002/jemt.22848
RESEARCH ARTICLE
Nanoscale electrical properties of ZnO nanorods grown by chemical bath deposition Sy-Hann Chen
|
Chang-Feng Yu | Chia-Shan Chien
Department of Electrophysics, National Chiayi University, Chiayi, 600, Taiwan Correspondence Prof. Sy-Hann Chen, Department of Electrophysics, National Chiayi University, Chiayi 600, Taiwan. Email:
[email protected]
Abstract Well-aligned zinc oxide nanorod arrays (ZNAs) synthesized using chemical bath deposition were fabricated on a gallium-doped zinc oxide substrate, and the effects of varying the precursor concentrations on the growth and nanoscale electrical properties of the ZNAs were investigated. The as-synthesized ZNAs were characterized using field-emission scanning electron microscopy (FESEM), atomic force microscopy (AFM), conducting atomic force microscopy (CAFM), and scanning surface potential microscopy (SSPM). The FESEM and AFM images show that the growth rate in terms of length and diameter is highly sensitive to the precursor concentration. CAFM and SSPM analyses indicate that when concentrations of both the zinc acetate and hexamethylenetetramine solutions were 30 mM, the coverage percentages of the recordable and conducting regions on the ZNA surface were 48.3% and 0.9%, which is suitable for application in resistive random access memory devices.
KEYWORDS
zinc oxide nanorod arrays, chemical bath deposition, resistive random access memory
1 | INTRODUCTION
The development of RRAM was studied as early as the 1960s (Gibbons & Beadle, 1964; Hiatt & Hickmott, 1965); however, there
Nonvolatile memory (NVM) is indispensable in hand-held devices, such
have been no affirmative conclusions regarding the resistance-
as mobile phones and cameras, as well as in solid-state drives, which
switching mechanism or material selection until now. Many research
require an NVM module for data storage. At present, the conventional
teams aim to discuss the stability and feasibility of various materials for
type of NVM is charge trap flash (CTF) (King, King, & Hu, 2001; You &
RRAM and to explain and study physical models of the resistance
Cho, 2010). However, with the development of microminiaturization in
switching mechanism using different methods. According to several
fabrication, CTF is expected to fall out of use because of its physical
related references (Chung, Lai, Chen, & Chen, 2011; Fang et al., 2011;
limitations. Therefore, developing new types of NVM is an important
Kozicki, Gopalan, Balakrishnan, & Mitkova, 2006; Zhao et al., 2011),
topic in current industrial and academic circles. Currently, the new
the choice of material for the recording layer of RRAM is approximately
types of NVM with research value are approximately classified into
divided between four major materials: colossal magnetoresistance
magnetoresistive random access memory (MRAM) (Hu, Li, Chen, &
oxides, polymers, doped metal oxides, and binary metal oxides.
Nan, 2011; Tehrani et al., 2003), phase change random access memory
Because binary metal oxides (Panda & Tseng, 2013) are easily pro-
(PCRAM) (Loke et al., 2012; Raoux et al., 2008), and resistive random
duced, are compatible with the current semiconductor process, and
access memory (RRAM) (Ielmini, Nardi, & Cagli, 2011; Philip Wong
have good electrical characteristics, they are presently the key research
et al., 2012). RRAM offers several advantages such as good microminia-
material in the RRAM field. In terms of resistance switching mecha-
turization potential for simple structures, high response speed, low
nisms, binary metal oxide behavior can be classified into bipolar and
power consumption, and good reliability, and thus, it may become the
unipolar transfers. Similar to PCRAM (Makarov, Sverdlov, & Selberherr,
mainstream module for next-generation digital information storage
2012), RRAM in unipolar operation has a relatively high memory den-
devices.
sity and a more than thousand-fold change in resistance value, making it suitable for multilevel cells. Among the numerous binary metal
Review Editor: Dr. Chuanbin Mao
Microsc Res Tech. 2017;80:671–679
oxides, zinc oxide (ZnO) (Chang et al., 2008; Lee, Kim, Park, & Yong,
wileyonlinelibrary.com/journal/jemt
C 2017 Wiley Periodicals, Inc. V
|
671
672
|
T A B LE 1
CHEN
ET AL.
The different precursor concentrations used and the corresponding diameter and length of the produced ZNAs
Sample name
Zinc acetate (mM)
HMTA (mM)
Nanorod Diameter (nm)
Nanorod Length (nm)
Aspect ratio
S1
10
10
72 6 3
181 6 8
2.48 6 0.22
S2
30
30
140 6 3
727 6 13
5.19 6 0.20
S3
50
50
250 6 5
1,690 6 25
6.76 6 0.24
S4
70
70
399 6 7
1,150 6 20
2.88 6 0.10
*Errors are calculated by the experimental data from five-fabricated times.
2010; Panda & Tseng, 2013) is a potential recording material for the
under a constant ambient O2 pressure of 20 mTorr and at a substrate
unipolar resistance switching mechanism. ZnO can grow at a relatively
temperature of 2008C. A glass substrate was mounted parallel to the
low temperature, its raw materials are easily procured, and it is inex-
target surface at a distance of 5 cm. A Q-switched Nd:YAG laser (LS-
pensive, innocuous, and nonpolluting. In addition, diverse nano-
2137U, LOTIS) with a second harmonic operating at 355 nm and a
structures and preparation methods have been reported in the litera-
duration of 7 ns was used as the light source for target ablation. The
ture for ZnO, thus making it a very promising candidate for the record-
laser beam was focused using a 30 cm positive lens. The average
ing layer of RRAM modules in the future. (Ielmini et al., 2011; Philip
energy was approximately 160 mJ per shot, and the pulse rate was 10
Wong et al., 2012)
Hz. Each deposition process lasted 10 min, and the produced film
In this study, ZnO nanorod arrays (ZNAs) were grown by chemical
thicknesses were around 150 nm.
bath deposition (CBD) (Khranovskyy et al., 2012; Son, Noh, & Park,
Zinc acetate [C4H6O4Zn2H2O] and hexamethylenetetramine
2016; Urgessa, Oluwafemi, & Botha, 2012). Gallium-doped zinc oxide
(HMTA) [C6H12N4] were purchased from Sigma-Aldrich and used as
(GZO) was used as the seed layer, which was previously used as an
reagents without further treatment. Aqueous solutions of zinc acetate
anode for polymer light emitting diodes (Chen & Chen, 2012; Chen,
and HMTA were prepared in de-ionized (DI) water at various concen-
Chen, Yu, Lin, & Kao, 2013) and has been actively developed by
trations, and equal amounts of the zinc acetate solution and HMTA
researchers in recent years. The diameter and length of ZNAs are
solution were mixed in a beaker. The different concentrations used for
modulated by the precursor concentrations of the chemical agents in
the zinc acetate and HMTA solutions ranged between 10 and 70 mM,
the process. The effect of the precursor concentration in CBD on the
as shown in Table 1. A substrate with a pre-deposited GZO seed layer
morphology, dimensions, and electrical properties (Huang et al., 2014;
was vertically immersed into the mixed solution. Then, the beaker was
Yao et al., 2012) is important for the application of ZnO as NVM. The
sealed and placed in a water bath illuminated with UV light at 908C for
surface morphology and section depth of ZNAs were measured and
an hour. Finally, the samples were rinsed with DI water several times
analyzed using atomic force microscopy (AFM) and field-emission scan-
and dried at 278C for several hours. The ZNA samples synthesized
ning electron microscopy (FESEM). Conducting atomic force micros-
from precursor concentrations of 10, 30, 50, and 70 mM were labelled
copy (CAFM) (Chen et al., 2008; Lin, Chen, Perng, & Chen 2001; Yu
S1, S2, S3, and S4, respectively.
et al., 2008), and scanning surface potential microscopy (SSPM) (Chen,
The shapes and dimensions of the vertically grown ZNAs were
2005) were used to measure the electrical properties of ZNAs. A con-
characterized using FESEM (JEOL JSM-7500F) and AFM (Dimension
ductive probe was used as the top electrode, and the local contact cur-
3100, Bruker, USA). To analyze the lattice structure and crystal quality
rent and work function (WF) of ZNAs were integrally analyzed and
of the ZNA samples, the 18 grazing incidence X-ray diffraction (XRD)
discussed. These nanoscale experimental data and analysis are useful
patterns were recorded at room temperature in the 2u scan range from
for further understanding the filament effect of the resistance switch-
208 to 808, with a 0.0028 step using a high resolution diffractometer
ing mechanism in the RRAM module and may generate a more com-
(Cu Ka radiation; D8 DISCOVER, Bruker, USA). The local electrical
plete theoretical model, which is helpful for improving the RRAM
properties of the ZNAs were analyzed under ambient conditions using
module structure and the sophistication of the process.
CAFM and SSPM. Rectangular Si tips (PPP-EFM, Nanosensors, Switzerland) with a 2.50 N/m spring constant and a 75 kHz resonant frequency
2 | MATERIALS AND METHODS
were used for the CAFM and SSPM measurements. The contact force, which was determined from the force–distance plot, setpoint, spring
The fabrication process of vertically aligned ZNAs includes two steps:
constant, and deflection sensitivity of the cantilever, was maintained at
the preparation of textured GZO thin films using pulsed laser deposi-
approximately 60 nN by applying a constant setpoint. The tips were
tion (PLD) and ZnO nanorod growth using CBD.
precoated with a Cr layer and subsequently coated with a 20 nm PtIr
For the PLD step, a ceramic target of ZnO (99.999% purity) mixed
film by ion sputtering. The CAFM and SSPM procedures employed in
with 3.35 wt% Ga (99.999% purity) was used. The ambient gas was
this study have been previously described (Chen, 2005; Chen et al.,
high-purity O2 (99.999%). The GZO seed layer was prepared via PLD
2008; Lin et al., 2001; Yu et al., 2008). The current–voltage (I–V) curves
CHEN
|
ET AL.
FIGURE 1
673
The cross-sectional (left) and top-view (right) FESEM images of (a) S1, (b) S2, (c) S3, and (d) S4 samples
and current maps were measured by applying a positive voltage to the
ions will be attracted to the surface of the sample, modifying it and pre-
tip while the substrate was grounded. The applied voltage induces a
venting any collection of current. To avoid this local anodic oxidation of
high electrical field between the tip and the substrate that ionizes the
the sample, CAFM measurements (in air) are usually obtained by apply-
water droplet, and the OH2 ions produced provide the oxidant for the
ing a positive voltage to the tip while the substrate is virtually grounded
2
chemical reaction. If a negative voltage is applied to the tip, the OH
(namely, by electron injection from the substrate).
674
|
CHEN
ET AL.
the average diameter increased from 250 to 399 nm but the length decreased dramatically from 1,690 to 1,150 nm. This implies that the growth rate in terms of length is more sensitive to the precursor concentration. According to the FESEM images on the right side of the Figure 1d, the ZNAs of the S4 sample extrude each other, as the diameter is too large. In the same area, when large ZNAs are grown at high precursor concentrations, the nanorods are more likely to intersect each other because of spatial localization, so the ZNAs are likely to be tilted or damaged, thus making them inapplicable for RRAM modules. Figure 2 shows the XRD patterns of ZNAs grown on the GZO seed layers using different precursor concentrations and the detail data are exhibited in Table 2. For all samples, strong peaks appear at 2h 34.58, which are indexed to the (002) plane of hexagonal ZnO crystal and indicate that ZNAs have a strong c-axis orientation (Son et al., 2016). The ZnO (103) plane showed up in S2,S3, and S4 has a strong F I G U R E 2 XRD patterns of ZNAs grown with four different concentrations of zinc acetate and HMTA. [Color figure can be viewed at wileyonlinelibrary.com]
Contact potential difference (CPD, DU) mapping was performed in “lift mode” at a lift scan height (tip-sample separation) of 80 nm with a 5 V AC voltage applied to the tip, and DU is equal to Utip2Usample, where
Utip and Usample are the mean WFs of the tip and ZNA samples, respectively. The measurement was obtained simultaneously with the topography scan in tapping mode using the electrically conductive Cr/PtIrcoated tip. The tip was calibrated using a highly oriented pyrolytic graphite (HOPG) substrate.1 The chemical states of ZNA samples were identified by X-ray photoelectron spectroscopy (XPS) using a VG Scientific ESCALAB 250 spherical sector analyzer with a monochromatic AlKa radiation source (1,486.7 eV). The binding energy of the C1s level of adventitious carbon is widely taken as energy calibration. The elemental composition was determined from the relative peak areas at specific binding energies.
relationship with nanorod length. It can be found that the higher nanorod length has the larger distribution percentages of ZnO (103) plane by observing Tables 1 and 2. The result can reasonably conjecture that the length of growing nanorods would be as the main influencing factor to the c-axis orientation of the ZNAs. Moreover, according to fivefabricated times for XRD measurements, the lattice structure of S4 sample is in an instable status. The reason is that the S4 sample which has the larger nanorod diameter can lead to structural extrusion for each other. Topography (left) and current (right) images of the ZNA samples grown from the different precursor concentrations are presented in Figure 3. These images were obtained with a 180 mV bias applied to the tip during scanning, and they show that conducting, recordable, and insulating regions were present on the surfaces. The magnitude of the contact current was determined from the image brightness. According to the local current–voltage (I–V) measurements, when the tip was located at a point where the current exceeded 200 nA, the measured I–V data followed the relationships expected for Ohmic contacts, and the turn-on voltage was lower than 0.5 V. A typical I–V curve
3 | RESULTS AND DISCUSSION
is shown in Figure 4a and these regions are called conducting regions. When the probe was in a region in which the contact current was
Figure 1 shows the cross-sectional (left) and top-view (right) FESEM
between 5 and 200 nA, the typical I–V curve shown in Figure 4b shows
images of ZNAs grown using different precursor concentrations.
set and reset at the same polarity. Such a relationship is expected for
The FESEM images reveal that the nanorods have a well-formed hex-
unipolar memory behavior, and these regions are called recordable
agonal shape, suggesting the nanorods grow along the (002) direction
regions. However, in the regions where the current was less than 5 nA,
for the various precursor concentrations. The FESEM images also indi-
the Fowler-Nordheim electron tunneling mechanism (Chen et al., 2008;
cate that the precursor concentration has a strong impact on the diam-
Lin et al., 2001) shown in Figure 4c was responsible for the detected
eter, length, and aspect ratio of the ZNAs. As shown in Table 1, the
current, and the turn-on voltage was higher than 4 V; these regions are
average diameter increases significantly as the precursor concentration
called insulating regions. The coverage percentages of these three
increases. When concentrations of both the zinc acetate and HMTA
regions, which are presented in Figure 5, were determined using the
solutions were 50 mM, the ZNAs have a maximum length of 1,690 nm
current distribution histograms of the current images in Figure 3. As
and an aspect ratio of 6.76. However, when both concentrations of the
shown in Figure 4b, when a voltage was applied on the memory cell
zinc acetate and HMTA solutions were increased from 50 to 70 mM,
(swept from 0 to 5 V), it showed a very low current value (high resist-
1
HOPG has a stable work function of approximately 4.6 eV in ambient conditions, so it can be used as a standard sample. It is best to use a freshly cleaved HOPG substrate and to perform a measurement on it before measuring the ZNA surfaces.
ance state, HRS). With a higher bias, the current flow started to increase, and the cell switched to a high current state (low resistance state, LRS) at 1.69 V (set voltage). The above process is called the forming process, and a nonvolatile ON state is achieved. During a reverse
CHEN
|
ET AL.
T A B LE 2
675
XRD data for the ZNAs prepared at different precursor concentrations Sample name
XRD data S1
S2
S3
S4
Maximum intensities of ZnO (002) peak [counts]
327
11,810
7,239
9,140
Maximum intensities of ZnO (103) peak [counts]
N/A
1,534
2,842
1,658
Peak positions of ZnO (002) [degrees]
34.5
34.5
34.5
34.5
Peak positions of ZnO (103) [degrees]
N/A
62.8
62.8
62.8
Coverage ratio of ZnO (002) plane [%]
100
88.5
71.8
84.6
Coverage ratio of ZnO (103) plane [%]
N/A
11.5
28.2
15.4
*All experimental data are defined as average values analyzed from five-fabricated times.
Topography (left) and current (right) images of (a) S1, (b) S2, (c) S3, and (d) S4 samples. The tip was biased at 180 mV, and the samples were grounded. [Color figure can be viewed at wileyonlinelibrary.com]
FIGURE 3
Typical I–V curves from (a) the conducting regions, (b) the recordable regions, and (c) the insulating regions of all samples. [Color figure can be viewed at wileyonlinelibrary.com]
FIGURE 4
676
|
CHEN
ET AL.
shown in all current images in Figure 3, the conducting regions are distributed over the edge of the nanorod surface, because the charge accumulation forms a corona discharge-like region (Beinik et al., 2011; Xiao, Ong, Guo, Ho, & Zeng, 2015). The coverage percentage of the conducting region on sample S1 is as high as 46.6%, which severely limits the stability and lifetime of the RRAM module. The topography (left) and flattened CPD (right) images of the samples from a 2 3 2 mm2 scanning area are shown in Figure 6. The DC offset values for the right side images in Figure 6a–d, which were measured from a cross-section of the CPD images before the images were flattened, are 0.76, 0.78, 0.85, and 0.85 V, respectively. Because the WF of the Cr/PtIr-coated probe (Utip) calibrated by HOPG was 5.00 eV (Chen, 2005), the mean WF values of S1, S2, S3, and S4 samples (Usample) were calculated to be 4.24, 4.22, 4.15, and 4.15 eV, respectively. (see Table 3) The difference in the mean WF values on ZnO Histograms of the contact current distribution, which were obtained from the current images in Figure 3. Accurate data are shown in Table 3. [Color figure can be viewed at wileyonlinelibrary.com] FIGURE 5
samples has a strong relationship with the concentration of oxygen vacancies (Chen & Chen, 2012). Figure 7 shows the XPS spectrum of the O 1s region for all ZNA samples. The typical O 1s peak at the surface is consistently fitted by three nearly Gaussian curves centered at
bias sweep from 5 to 0 V, there was a drop in the leakage current at a
530.5, 531.7, and 533.2 eV. (Chen et al., 2000; Chen & Chen, 2012).
voltage of around 2.91 V (reset voltage). The device then switched to
The high-binding-energy component (OIII peak) located at 533.2 eV
the HRS, and a nonvolatile OFF state was achieved. The resetting
was reported as the presence of loosely bound oxygen related to envi-
(switching from ON to OFF state) and setting (switching from OFF to
ronmental contaminants on the surface of the ZnO associated with
ON state) occurred in the same polarity of a sweep, showing typical
adsorbed H2O or O22. The component on the low-binding-energy side
unipolar behavior. The current at 1 V is about 250 and 40 nA for LRS
of the O 1s spectrum at 530.5 eV (OI peak) is attributed to O22 ions in
and HRS, respectively, which will lead to an resistance ratio about 6-
the wurtzite structure associated with a hexagonal Zn21 ion array sur-
fold of magnitude.
rounded by Zn atoms with their full complement of nearest-neighbor
Table 3 lists the coverage percentage of insulating, recordable, and
O22 ions. The medium-binding-energy component (OII peak) centered
conducting regions on all samples, obtained from Figure 5. Sample S2
at 531.7 eV is associated with O22 ions in the oxygen-deficient regions
has the highest coverage percentage of the recordable regions of all
within the matrix of ZnO. Therefore, changes in the intensity of this
samples at 48.3%, while sample S3 has the lowest value at 17.3%. The
component may be connected, in part, to variations in the concentra-
resistance of ZNAs increases with length; in other words, for longer
tion of oxygen vacancies. The coverage ratio of oxygen vacancies can
ZNAs, transferring electrons from the nanorod surface to the GZO
be calculated by the OII/(OI 1 OII) area ratio. As shown in Figure 7, the
seed layer is more difficult. The length of the ZNAs on sample S3 is
difference between the coverage ratios of oxygen vacancies of all sam-
much higher than that of the other samples, and thus, the coverage
ples was only within 2% (see Table 3), so the reasons for the difference
percentage of the insulating regions on the sample greatly increased to
in the mean WF values on ZNA samples is not clear.
82.7%, which explains why the coverage percentage of the recordable
The variations in the brightness of the CPD images represent WF
regions on sample S3 is so low. Therefore, sample S3 would not be
fluctuations, with brighter regions representing regions with lower
applicable as the recording layer of an RRAM module. In addition, the
WFs. In the CPD images in Figure 6, the edges of the nanorod surfaces
conducting regions of RRAM often generate high electric fields during
are relatively bright, meaning these regions have a low WF. According
module operation, disintegrating the region (Long et al., 2013). As
to the analysis of the CPD image roughness, the differences between
T A B LE 3
CAFM, SSPM, and XPS data for the ZNAs based on the different precursor concentrations Sample name
CAFM, SSPM, and XPS data S1
S2
S3
S4
Coverage percentages of insulator regions [%]
16.2 6 0.5
50.7 6 1.5
82.7 6 2.4
57.4 6 1.7
Coverage percentages of recordable regions [%]
37.1 6 1.2
48.3 6 1.3
17.3 6 0.6
36.8 6 1.0
Coverage percentages of conducting regions [%]
46.6 6 2.1
0.9 6 0.1
N/A
5.7 6 0.2
Mean work functions [eV]
4.23 6 0.02
4.22 6 0.02
4.15 6 0.02
4.15 6 0.02
Coverage ratio of oxygen vacancies [%]
63.9
64.3
64.0
65.4
*Errors and means are calculated by the experimental data from five-fabricated times.
CHEN
|
ET AL.
Topography (left) and CPD images (right) of (a) S1, (b) S2, (c) S3, and (d) S4 samples. The CPD images have been flattened. [Color figure can be viewed at wileyonlinelibrary.com]
FIGURE 6
FIGURE 7
XPS of the O 1s region for (a) S1, (b) S2, (c) S3, and (d) S4 samples. [Color figure can be viewed at wileyonlinelibrary.com]
677
678
|
the maximum and minimum intensity of the WF on samples S1, S2, S3, and S4 are 614.95, 69.80, 67.70, and 65.45 meV, respectively. The ZNA interval density decreases as the precursor concentration increases, and thus, the higher precursor concentration leads to better WF uniformity.
4 | CONCLUSION We have successfully synthesized ZNAs using a simple CBD method on glass substrates pretreated with GZO films by PLD. A systematic study of various precursor concentrations showed that these concentrations have a strong impact on the diameter and length of the ZNAs. XRD analysis demonstrated that all the as-grown ZNAs are crystalline and are preferentially oriented along the c-axis. FESEM and AFM images show that when concentrations of both the zinc acetate and HMTA solutions were lower than 50 mM, the ZNAs are particularly straight. The CAFM and SSPM analysis indicated that the coverage percentage of the recordable regions on the ZNA surfaces had a maximum value of 48.3% when the concentrations of the zinc acetate and HMTA solutions were both 30 mM, while the coverage percentage of the conducting regions on the ZNA surface decreased to 0.9%. This simple and economical CBD method is promising for the large-scale production of high-performance ZNAs for application in RRAM devices.
ACKNOWLE DGMENTS The authors would like to thank the Ministry of Science and Technology of Taiwan for financially supporting this research under Contract No. 105-2112-M-415-003.
REFERENCES Beinik, I., Kratzer, M., Wachauer, A., Wang, L., Lechner, R. T., Teichert, C., . . . Djurisić, A. B. (2011). Electrical properties of ZnO nanorods studied by conductive atomic force microscopy. Journal of Applied Physics, 110(5), 052005. doi:10.1063/1.3623764.
CHEN
ET AL.
Chen, S. H. (2005). Work-function changes of treated indium-tin-oxide films for organic light-emitting diodes investigated using scanning surface-potential microscopy. Journal of Applied Physics, 97, 073713. doi:10.1063/1.1884245. Chung, Y. L., Lai, P. Y., Chen, Y. C., & Chen, J. S. (2011). Schottky barrier mediated single-polarity resistive switching in Pt layer-included TiOx memory device. ACS Applied Materials and Interfaces, 3, 1918– 1924. Fang, Z., Yu, H. Y., Li, X., Singh, N., Lo, G. Q., & Kwong, D. L. (2011). HfOx/TiOx/HfOx/TiOx multilayer-based forming-free RRAM devices with excellent uniformity. IEEE Electron Device Letters, 32, 566–568. Gibbons, J. F., & Beadle, W. E. (1964). Switching properties of thin NiO films. Solid-State Electronics, 7, 785–790. Hiatt, W. R., & Hickmott, T. W. (1965). Bistable switching in niobium oxide diodes. Applied Physics Letters, 6, 106–108. Huang, C. Y., Ho, Y. T., Hung, C. J., & Tseng, T. Y. (2014). Compact Ga-doped ZnO nanorod thin film for making high-performance transparent resistive switching memory. IEEE Trans. Electron Device, 61, 3435–3441. Hu, J. M., Li, Z., Chen, L. Q., & Nan, C. W. (2011). High-density magnetoresistive random access memory operating at ultralow voltage at room temperature. Nature Communications, 2, 553. Ielmini, D., Nardi, F., & Cagli, C. (2011). Universal reset characteristics of unipolar and bipolar metal-oxide RRAM. IEEE Transactions on Electron Devices, 58, 3246–3253. Khranovskyy, V., Yakimova, R., Karlsson, F., Syed, A. S., Holtz, P. O., Urgessa, Z. N., . . . Botha, J. R. (2012). Comparative PL study of individual ZnO nanorods, grown by APMOCVD and CVD techniques. Physica B, 407, 1538–1542. King, Y. C., King, T. J., & Hu, C. (2001). Charge-trap memory device fabricated by oxidation of Si1-xGex. IEEE Transactions on Electron Devices, 48, 696–700 Kozicki, M. N., Gopalan, C., Balakrishnan, M., & Mitkova, M. (2006). A lowpower nonvolatile switching element based on copper-tungsten oxide solid electrolyte. IEEE Transactions on Nanotechnology, 5, 535–544. Lee, S., Kim, H., Park, J., & Yong, K. (2010). Coexistence of unipolar and bipolar resistive switching characteristics in ZnO thin films. Journal of Applied Physics, 108, 076101. doi:10.1063/1.3489882. Lin, H. N., Chen, S. H., Perng, G. Y., & Chen, S. A. (2001). Nanoscale surface electrical properties of indium–tin–oxide films for organic light emitting diodes investigated by conducting atomic force microscopy. Journal of Applied Physics, 89, 3976–3979.
Chang, W. Y., Lai, Y. C., Wu, T. B., Wang, S. F., Chen, F., & Tsai, M. J. (2008). Unipolar resistive switching characteristics of ZnO thin films for nonvolatile memory applications. Applied Physics Letters, 92, 022110. doi:10.1063/1.2834852.
Loke, D., Lee, T. H., Wang, W. J., Shi, L. P., Zhao, R., Yeo, Y. C., . . . Elliott, S. R. (2012). Breaking the speed limits of phase-change memory. Science, 336, 1566–1569.
Chen, M., Wang, X., Yu, Y. H., Pei, Z. L., Bai, X. D., Sun, C., . . . Wen, L. S. (2000). X-ray photoelectron spectroscopy and auger electron spectroscopy studies of Al-doped ZnO films. Applied Surface Science, 158, 134–140.
~e , J. Long, S., Lian, X., Cagli, C., Perniola, L., Miranda, E., Liu, M., & Sun (2013). A model for the set statistics of RRAM inspired in the percolation model of oxide breakdown. IEEE Electron Device Letters, 34, 999–1001.
Chen, S. H., Chen, W. C., Yu, C. F., Lin, C. F., & Kao, P. C. (2013). Influence of gallium-doped zinc-oxide thickness on polymer light-emitting diode luminescence efficiency. Microscopy Research Technique, 76, 783–787.
Makarov, A., Sverdlov, V., & Selberherr, S. (2012). Emerging memory technologies: Trends, challenges, and modeling methods. Microelectronics Reliability, 52, 628–634.
Chen, S. H., & Chen, Y. C. (2012). High-performance polymer lightemitting diodes based on a gallium-doped zinc oxide/polyimide substrate. IEEE Transactions on Electron Devices, 59, 1709–1715. Chen, S. H., Yu, C. F., Lin, Y. S., Xie, W. J., Hsu, T. W., & Tsai, D. P. (2008). Nanoscale surface electrical properties of aluminum zinc oxide thin films investigated by scanning probe microscopy. Journal of Applied Physics, 104, 114314. doi:10.1063/1.3042237.
Panda, D., & Tseng, T. Y. (2013). One-dimensional ZnO nanostructures: Fabrication, optoelectronic properties, and device applications. Journal of Materials Science, 48, 6849–6877. Philip Wong, H. S., Lee, H. Y., Yu, S., Chen, Y. S., Wu, Y., Chen, P. S., . . . Tsai, M. J. (2012). Metal–oxide RRAM. Proceedings of the IEEE, 100, 1951–1970 Raoux, S., Burr, G. W., Breitwisch, M. J., Rettner, C. T., Chen, Y. C., Shelby, R. M., . . . Lam, C. H. (2008). Phase-change random access
CHEN
|
ET AL.
memory: A scalable technology. IBM Journal of Research and Development, 52, 465–479. Son, N. T., Noh, J. S., & Park, S. (2016). Role of ZnO thin film in the vertically aligned growth of ZnO nanorods by chemical bath deposition. Applied Surface Science, 379, 440–445. Tehrani, S., Slaughter, J. M., Deherrera, M., Engel, B. N., Rizzo, N. D., Salter, J., . . . Grynkewich, G. (2003). Magnetoresistive random access memory using magnetic tunnel junctions. Proceedings of the IEEE, 91, 703–714. Urgessa, Z. N., Oluwafemi, O. S., & Botha, J. R. (2012). Effect of precursor concentration on the growth of zinc oxide nanorod arrays on pre-treated substrates. Physica B, 407, 1543–1545. Xiao, J., Ong, W. L., Guo, Z., Ho, G. W., & Zeng, K. (2015). Resistive switching and polarization reversal of hydrothermal-method-grown undoped zinc oxide nanorods by using scanning probe microscopy techniques. ACS Applied Materials and Interfaces, 7, 11412–11422. doi:10.1021/acsami.5b01988. Yao, I. C., Lee, D. Y., Tseng, T. Y., & Lin, P. (2012). Fabrication and resistive switching characteristics of high compact Ga-doped ZnO nanorod thin film devices. Nanotechnology, 23, 145201.
679
You, H. W., & Cho, W. J. (2010). Charge trapping properties of the HfO2 layer with various thicknesses for charge trap flash memory applications. Applied Physics Letters, 96, 093506. doi:10.1063/ 1.3337103. Yu, C. F., Chen, S. H., Xie, W. J., Lin, Y. S., Shen, C. Y., Tsai, S. J., . . . Ay, C. (2008). Nanoscale surface electrical properties of zinc oxide films investigated by conducting atomic force microscopy. Microscopy Research Technique, 71, 1–4. Zhao, L., Zhang, J., He, Y., Guan, X., Qian, H., & Yu, Z. (2011). Dynamic modeling and atomistic simulations of SET and RESET operations in TiO2-based unipolar resistive memory. IEEE Electron Device Letters, 32, 677–679.
How to cite this article: Chen S-H, Yu C-F, Chien C-S. Nanoscale electrical properties of ZnO nanorods grown by chemical bath deposition. Microsc Res Tech. 2017;80:671–679. https:// doi.org/10.1002/jemt.22848
