E
ffect of oxygen plasma on the surface states of ZnO
films used to produce thin-film transistors on soft plastic
sheets
Jagan Singh Meena,abMin-Ching Chu,aYu-Cheng Chang,aHsin-Chiang You,c Ranjodh Singh,aPo-Tsun Liu,bHan-Ping D. Shieh,bFeng-Chih Changa
and Fu-Hsiang Ko*a
Electronic displays andflexible electronics are poised to significantly impact emerging industries, including
displays, energy products, sensors and medical devices, building a market that will significantly grow in the
future. The implementation of transparent electronic devices requires the use of material components that
could be formed using controlled deposition in the appropriate orientation onto a transparentflexible
substrate. Here, we report a simple and efficient means of depositing onto a flexible polyimide (PI)
substrate a highly ordered and highly aligned zinc oxide (ZnO)film for use as a carrier transporting and
semiconducting layer with controlled surface charge density for thin-film transistor (TFT) applications.
The deposition approach is based on the solution-coating of a zinc-acetate suspension under controlled
conditions of the spreadflow rate, droplet size of the drops, speed limit, and the oxygen (ca. O2) plasma
treatment of the coatedfilm surface on the PI substrate. The plasma surface interactions on the surface
states of the ZnO films for various times (ca. 1–5 min) were studied using X-ray photoelectron
spectroscopy and Fourier transform infrared spectroscopy. Moreover, the effects of O2plasma and the
subsequent thermal annealing in an O2atmosphere at 250 C on the properties of ZnO films were
studied for its efficacy in TFT applications in terms of the charge carrier density and the change in the
mobility. ZnO thin-film-based TFTs on PI exhibited a very high electron mobility of 22.8 cm2V1s1at a
drain bias of 5 V after treatment with O2plasma for 2 min. Furthermore, the plasma treatment for long
durations of time caused a reduction in the charge carrier density from 1.58 1019cm3for the 2 min
treatment to 1.13 1017cm3for the 5 min treatment, and the corresponding electron mobility was
changed from 22.8 and 3.1 cm2V1s1for the treatment times of 2 min and 5 min, respectively. The
spin-coating technique used to deposit very thin ZnO films is currently used in microelectronics
technology, which helps to ensure that the described ZnO thin-film deposition approach can be
implemented in production lines with minimal changes in the fabrication design and in the auxiliary
tools used inflexible electronics production.
1.
Introduction
Zinc oxide (ZnO) shows great promise as an active layer in thin lm transistors (TFTs) due to its exceptional electronic and optoelectronic properties.1ZnO is a wide band gap (3.4 eV) II–VI compound semiconductor, which has a stable wurtzite struc-ture with lattice parameters of a¼ 0.325 nm and c ¼ 0.521 nm.2 ZnO is a promising material for short-wavelength light emitting
diodes and surface-acoustic wave devices. The high mobility and transparency of ZnO combined with a low temperature deposition process make ZnO a potential candidate for use in exible and transparent display applications. Notable efforts have been made in recent years to utilize the unique properties of ZnO in a wide variety of applications, including transparent electronics, ultraviolet (UV) light emitters, piezoelectric devices, chemical sensors and spin electronics.3High-quality ZnOlms can be grown at relatively low temperatures (<700C). The large exciton binding energy of approximately 60 meV enables an intense near-band-edge excitonic emission at room tempera-ture and above because the exciton binding energy is 2.4 times that of the room temperature thermal energy (kBT¼ 25 meV).4
Being inexpensive and nontoxic, ZnO is a good substrate material for devices based on gallium nitride (GaN) due to the good match of the lattice constants.5 Moreover, ZnO has
aDepartment of Materials Science and Engineering, National Chiao Tung University,
Hsinchu 30010, Taiwan, Republic of China. E-mail:[email protected]; Fax: +886-35744689; Tel: +886-35712121 ext. 55803
bDepartment of Photonics and Display Institute, National Chiao Tung University,
Hsinchu 30010, Taiwan, Republic of China
cDepartment of Electronic Engineering, National Chin-Yi University of Technology,
Taichung 41170, Taiwan, Republic of China Cite this:J. Mater. Chem. C, 2013, 1, 6613 Received 10th July 2013 Accepted 12th August 2013 DOI: 10.1039/c3tc31320d www.rsc.org/MaterialsC
Materials Chemistry C
PAPER
Published on 12 August 2013. Downloaded by National Chiao Tung University on 28/04/2014 01:51:38.
View Article Online
attracted intensive research efforts that are currently directed toward enhanced intrinsic transparent electric conductivity for potential use in theexible electronics industry. Many of the recent reports featured the integration of ZnO into TFTs on silicon substrates, exible plastic substrates, and even on stretchable rubber substrates.6To achieve many applications in exible and transparent electronics devices, the development of optimized semiconductor devices using ZnO as a channel layer remains critically dependent on low-temperature processing.
ZnO lms have been grown using vacuum-based deposition
techniques, such as radio-frequency magnetron sputtering,7 pulsed laser deposition,8 chemical vapor deposition,9 atomic layer deposition,10transfer printing11and sol–gel spin coating.12 On the basis of these methods, the preparation of ZnOlms and a wide variety of oxide semiconductors exhibiting high carrier mobility and low carrier concentrations has been demonstrated for applications in electronic devices.
However, these vacuum-based deposition methods are not
likely to be compatible with exible TFT manufacturing
processes due to the need for expensive equipment, the high energy consumption and the low throughput of vacuum-based deposition techniques. High-vacuum and high-temperature based deposition processes are also not well suited for plastic substrates due to their intrinsically low melting temperatures. Moreover, a low manufacturing cost is the primary requirement for modern and mass produced large area electronic devices. Solution-processed thin-lm deposition can provide many advantages that facilitate the fabrication of high-performance and low-cost electronics, such as simplicity, low cost, and high throughput. Although the fabrication of a ZnO transistor on a plastic substrate at low temperature has been achieved, signi-cant challenges remain in the fabrication of electronics devices that meet the mechanical and optical specications for exible and transparent electronics. For the reasons stated above, we turned our attention to developing a solution-based deposition method to fabricate ZnO thinlms. High-quality oxide lms can be achieved by solution deposition followed by an O2 plasma
treatment step.13 Thus, solution deposition followed by a low power (i.e., 20–30 W) O2plasma treatment could be a desirable,
relatively low-temperature technique for producing high-quality, high-performance ZnO thinlms for use in TFT device applica-tions on plastic substrates. In addition to oxygen vacancies, doping can also change the electrical conductivity of ZnOlms. Oxygen vacancies, which provide the necessary free carriers for electrical conduction, are easily generated in oxidelms.14Thus, low power O2plasma treatment is expected to be a useful
tech-nique to inuence the electrical properties of a ZnO-based TFT device fabricated at a relatively low temperature. The carrier concentration could be controlled with the least amount of effort by using O2plasma power supply, with the time to create the
oxygen vacancies being determined by the following equation:15 Oo/1
2O2ðgÞ þ Voþ 2e (1)
where edenotes an electron, Vodenotes the threshold voltage
shi, and O2and Oorepresent the neutral and charged oxygen
molecules, respectively, in the ZnOlm associated with the TFT
device. Moreover, the electrical resistivity (r) of ZnO lms is determined by the carrier concentration (n) and the carrier mobility (m) by r ¼ 1/(nem), where e is the electron charge.16 Because ‘e’ is a constant, to obtain low resistivity, the carrier concentration and carrier mobility should be simultaneously maximized; the methods for achieving the maximum carrier concentration are through the creation of oxygen vacancies and through doping. Oxygen vacancies can be created by controlling the substrate temperature or the ambient oxygen pressure. Note that if an oxygen vacancy is created in a perfect crystal, two electrons are created in the crystal through the contribution from ionized donors. However, if excess oxygen vacancies are created in ZnO thinlms, sub-oxides will form, causing the resistivity to increase. The current research studies are mainly focused on achieving low resistivity in ZnO thinlms by increasing the free-carrier concentration through use of dopants and oxygen vacancies. Johnson and Horovitz17 stated that increasing the carrier density via doping or oxygen vacancies is self-limiting because increasing the number of free carriers decreases the mobility of the carriers due to carrier–carrier scattering. As a result, our focus is also on the inuence of O2plasma treatment
on the carrier concentrations and on the change in the carrier mobility of ZnO TFTs fabricated on plastic substrates.
In the present study, the inuence of O2plasma treatment on
ZnOlms prepared from a solution using a precursor of zinc acetate is studied, as well as its effect on a channel layer in ex-ible TFTs produced using the treated ZnO lms. The various solution-deposition and O2 plasma treatment conditions were
carefully controlled and systematically studied, from the prepa-ration of the precursor solution to the plasma surface interaction, to investigate the effects of each of the conditions. This study is focused on increasing the free-carrier concentration in thinlms through the use of dopants and oxygen vacancies created by plasma surface treatment. The interaction of the plasma ions with the surface of thelms on the molecular level was investi-gated using FTIR spectroscopy to determine the chemical composition and using XPS spectra for the O1s and C1s regions to determine the surface states. The O2plasma exhibited a strong
effect on the charge carrier density and caused a change in the mobility of the ZnO TFTs. Because the O2 plasma treatment
primarily affects the surface of the lms, the effect of this treat-ment is expected to be enhanced in systems with a high surface-to-volume ratio. The O2plasma treatment plausibly reduced the
surface dangling bonds and the carbon contaminations, thereby forming the oxygen vacancy-rich surface. The characteristics of the ZnO TFTs, including the threshold voltage shi and the carrier mobility, were changed by the O2plasma treatment. The
purpose of this work is to demonstrate the possibility of producing a high-mobility exible and transparent TFT with improved performance through a sol–gel process that is highly compatible with the other standard fabrication techniques.
2.
Experimental section
2.1 Preparation of ZnO thin-lms
The precursor solution for producing the ZnO thinlm to be used as a channel layer was prepared by dissolving zinc acetate
dehydrate [Zn(CH3COO)2$2H2O, Aldrich, St. Louis, MO] into
ethanol, abbreviated as EtOH [CH3CH2OH; Fluka; water content
<0.1%; Aldrich]. The concentration of zinc acetate was 0.05 M, and the volume of EtOH was 10 mL. The solution was rigorously
stirred for 1 h at 60 C and ltered through a 0.22 mm
membrane lter [polytetrauoroethylene (PTFE)] to obtain a
transparent and homogeneous solution. Prior to thin lm
deposition, a DuPont Kapton PI lm of 38 mm thickness was chosen from the PV9100 series for use as theexible substrate to fabricate the ZnO TFT and for the ZnO used for other char-acterizations. The PI substrate was ultrasonically cleaned using EtOH for 15 min and then rinsed with DI water for 10 min. Next, a high-pressure stream of N2gas was used to remove the water
and any remaining particles from the PI surface. The cleaned PI substrate was then annealed at 200C for 1 h under vacuum to achieve relative thermal stability and to enhance the adhesion strength. The precursor solution of zinc acetate was spin coated at a speed of 500 rpm for 30 s and at 1000 rpm for 20 s onto the exible PI substrate, with a piece of glass substrate being used
to support the PI. The O2 plasma treatment was applied
immediately aer the spin coating to obtain a pure ZnO thin lm and to remove any unnecessary organic impurities. The as-depositedlms were treated with O2plasma for 1 to 5 min in an
O2 plasma reactor (Harrick Scientic Corp., Japan), which
supplied a plasma power of 18 W. Scheme 1(a) illustrates the expected three-step reaction of (i) hydrolysis, (ii) condensation, and (iii) O2 plasma surface ion interaction for producing an
appropriate ZnO lm on a exible PI substrate. Scheme 1(b) shows a rough schematic of the preparation of the zinc acetate sol–gel solution, the spin coating process and the surface O2
plasma treatment. This scheme includes the successive removal of the hydroxide layer and organic impurities with the O2
plasma treatment. These O2 plasma-treated lms were also
annealed at a temperature of 250 C for 1 h under normal environmental conditions for surface passivation.
2.2 ZnO thin-lm characterizations and TFT device fabrication
The surface morphology of the ZnO lms on exible PI
substrates was evaluated using atomic force microscopy (AFM) with a Digital Instruments Nanoscope D-5000 at a scan size of 2mm and a scan rate of 1 Hz. The surface roughness charac-teristics of these O2 plasma treated lms were calculated in
terms of the average root-mean square (rms) using AFM data processing soware. The thicknesses of the ZnO lms on the PI substrate were measured using a step-prolometry technique. The optical absorption of the ZnOlms on exible PI substrates was measured at wavelengths between 300 nm and 900 nm using a UV-Vis absorption spectrometer. X-ray photo-spectros-copy (XPS) and Fourier transform infrared spectrosphoto-spectros-copy (FTIR, Bomem DA-8.3) analyses were also used to conrm the effects of the plasma treatment time (from 1 min to 5 min) on the surface properties of the ZnOlms and its interaction with the surface ions. To fabricate the ZnO TFTs onexible PI substrates, these O2plasma-treated lms were used as the active layer for the
channel on a SiO2/Au/Cr/PI substrate. To fabricate the transistor
with bottom-contact electrodes, Cr and Au with thicknesses of 20 and 80 nm, respectively, were sequentially deposited through a shadow mask using a thermal coater to function as the gate electrodes. Then, the 100 nm thick SiO2 lm was deposited
using the plasma-enhanced chemical vapor deposition (PECVD) technique as a gate insulator layer. Finally, the source and drain electrodes of Au with a thickness of 100 nm were deposited onto the ZnO/SiO2/Au/Cr/PI through a shadow mask, which yielded
the top-contact for the ZnO TFTs. The channel length (L) and width (W) were 70mm and 2000 mm, respectively. The electrical measurements for the ZnO TFT devices were performed using an Agilent-4156 probe station.
3.
Results and discussion
3.1 ZnOlm analysis: surface roughness, thickness and transmittance properties
The AFM images, displayed in Fig. 1(a)–(e), show that the O2
plasma-treated ZnO lms are well dispersed and micro-crys-talline, when deposited on exible PI substrates. The AFM images of the surface treated with O2plasma power for 1–5 min
are essentially smooth and uniform surfaces. The average rms surface roughness values of the ZnO thin-lms treated with O2
plasma for 1–5 min were 1.12 nm (1 min), 1.28 nm (2 min), 1.53 nm (3 min), 1.92 nm (4 min) and 2.17 nm (5 min). This result indicated that the ZnOlms treated with O2plasma for 1–
5 min exhibited good large-area quality, smoothness, a crack-free morphology and a uniform surface. The variation in the surface properties of the samples treated with O2plasma for 1–5
min is clearly observed in the AFM images. For 1–3 min of plasma treatment, thelm surface exhibited a dense and deep microcrystalline nature. However, as the plasma treatment time increased to 4 to 5 min or more, the quality of thelm surface Scheme 1 (a) Schematic representation of the reaction of zinc acetate mixed
into ethanol and the O2plasma growth mechanism for the fabrication of ZnO thin-films; (b) process of making the sol–gel solution, spin-coating the solution to fabricate the thinfilm, and finally using O2plasma processing to modify the conductivity.
degraded. The extended plasma treatment etched the native surface particles, including the organic residuals. The extended plasma treatment is also expected to degrade the charge carrier concentration and the mobility of the transistor device.
More-over, the surface morphology of our plasma-treated lms
deposited ontoexible PI substrates exhibits similar effects to those of O2plasma pre-treatment on ZnO thinlms grown on a
polyethersulfone substrate at various deposition tempera-tures.18 The AFM results demonstrated the unique ability of theselms with a microcrystalline surface to be used in device applications on aexible substrate to control the functionality of ZnO TFTs. Moreover, the O2 plasma treatment plausibly
reduces the surface dangling bonds and the carbon contami-nation, thereby resulting in an oxygen vacancy-rich surface. The roughness and–O–Zn– bonds to the –O–OH– could be slightly increased, but the post-annealing ambient has no further signicant effect on the morphology, except for the passivation of the surface. Aer the O2 plasma treatment, the relative
density of the –O–Zn– bonds was enhanced and that of the –O–OH– bonds decreased as the treatment time increased. Furthermore, for the use of O2plasma for a treatment time of 5
min or more, the surface residuals and the–O–Zn– bonds were seriously affected, which is expected to degrade the device properties.
The thicknesses of the ZnOlms on exible PI substrates were measured using a step-prolometry method: 12.1 nm for 1 min O2plasma, 11.6 nm for 2 min O2plasma, 11.2 nm for 3 min
O2plasma, 10.4 nm for 4 min O2plasma and 10.2 nm for 5 min
O2plasma. A minor reduction in the thickness of the ZnOlms
with increasing plasma treatment time was observed. Appar-ently, as the plasma strikes the surface of the ZnO lms, it engraves the surface particles. This engraving could be due to etching of organic residuals, which results in the reduction of thelm thicknesses. To conrm the transmittances of ZnO thin lms on exible PI substrates as well as on glass substrates as a function of O2plasma treatment time, we deposited an
ultra-thin ZnO lm onto a exible PI substrate that covers the university NCTU logo. The average optical transmission of the PI substrate in the visible part of the spectrum is approximately 88%, as shown in Fig. 2, which indicates a 12% transmission
loss from the ZnO-coated exible PI substrate. High optical transmittance over a large wavelength range from 350 nm to 750 nm is an important property for the transparent or semi-transparent ZnO layer in TFTs because one must minimize any optical loss due to the transparent semiconductor layer. The ZnO lm on a PI substrate exhibits a slightly lower trans-mittance compared with the ZnOlm on glass for different O2
plasma process treatments, but this lower transmission is compensated for by its slightly lower sheet resistance (27.2 U sq1), indicating that the transmittance to sheet resistance
ratio is similar for ZnO on either substrate. In addition to the high optical transparency on par with ZnO thin lms over exible substrates, solution-deposited ZnO offers excellent mechanical exibility while maintaining high conductivity, a signicant advantage over a traditional semiconductor layer that will crack under a large degree of bending. The mechanical exibility and recoverable conductivity of this ZnO semiconductor layer not only makes it compatible with low-cost, roll-to-roll manufacturing but also helps itnd promising applications in emerging technologies (such as foldable displays,exible solar cells, and TFTs), in which the semiconductor layer must with-stand mechanical deformation without a loss in conductivity. Transparent ZnO thinlms on glass and on exible substrates have been widely studied, and our experimental results have determined the transmittance functions of a ZnO thinlm.19
3.2 Effect of O2plasma treatment on the surface properties
of ZnOlms: XPS and FTIR analyses
To determine the effect of plasma-induced active oxygen ions on the electrical properties of ZnO TFTs, we used XPS to analyze the surface of the ZnO aer plasma treatment. The XPS data indi-cate that the O1s peak at the ZnO surface can be consistently tted by three Gaussians, centered at 532.4 (OI), 531.2 (OII) and
530.4 eV (OIII). Fig. 3 shows that the relative intensity of oxygen
continuously increases towards higher binding energies as the plasma-treatment time increased from 0 to 3 min, indicating a Fig. 1 (a–e) AFM images (scale: 2 mm 2 mm) of the ZnO films for the samples
after O2plasma treatments ranging from 1 to 5 min in duration.
Fig. 2 Transmission spectra for transparent ZnO thinfilms on PI; inset: the ZnO film over the university NCTU logo.
clear increased concentration of the oxygen chemical state on the ZnO surface. However, the relative intensity of O2reverses
for treatment times longer than 3 min because the active oxygen ions reacted with residual hydrocarbons at the surface of the ZnO lm.20 The XPS data indicate that the peak intensity decreased. The high binding energy component located at 532.4 eV is usually attributed to the presence of loosely bound oxygen on the ZnO surface, such as–CO3, adsorbed H2O, and
adsorbed O2. The oxygen components were observed to increase
aer the ZnO lm was treated with O2plasma for 1 min. The
results implied that the O2plasma caused an accumulation of
absorbed O2on the ZnO surface with increasing treatment time.
The medium binding energy component, which is centered at 531.2 eV, is associated with oxygen decient regions within the matrix of ZnO.21The intensity changes of the medium oxygen component are related to variations in the concentration of oxygen vacancies. In other words, medium oxygen-related oxygen vacancies provide free electrons in the active channel layers of ZnO TFTs. The intensity of the medium oxygen related to oxygen vacancies was reduced by the O2plasma treatment of
3 min. The low binding energy component (i.e., third energy component) of the O1s spectrum at 530.4 eV is attributed to O2 ions within the micro-crystalline structure of the hexagonal Zn2+ion array. In other words, the intensity of the third energy component is a measure of the amount of oxygen atoms in a fully oxidized stoichiometric surrounding. Thus, we concluded that an O2plasma treatment of 5 min or more resulted in the
reaction of active oxygen ions with the residual hydrocarbon in
the ZnO lm and in the etching of the Zn2+ ion array. This reaction and etching affect the semiconducting properties of the ZnO surface and consequently degrade the ZnO TFT performance to lower the mobility. This degraded performance
indicates that the number of –OH–Zn–O–Zn–OR– bonds
increased and the O–OH and –Zn–OR– bonds lost from the surface were transformed into Zn–O– bonds, meaning that the
ZnO was formed at the Zn surface aer the O2 plasma
treatment.
In addition, XPS analysis was performed to identify the chemical reactions on the carbon surface of the zinc acetate precursor-basedlms on the exible PI substrate resulting from O2plasma treatment. The XPS analysis was conducted for the
C1s peaks for the untreated lm and lms treated using O2
plasma for durations ranging from 1 to 5 min. The XPS survey spectra of the untreated and plasma-treated carbon bers exhibited prominent peaks of carbon in the binding energy range of 270–295 eV. Fig. 4 shows the original and the decon-voluted C1s core level spectra of the untreated and plasma-treated carbonbers. For the untreated sample, the C–C peak was located at 284.7 eV, C–OH at 285.9 eV, C–O–C at 286.1 eV, C]O at 287.5 eV, and the COOR or COOH group at 288.8 eV. The C–O–C and COOR peak area ratios increased, accompanied by a decrease of the C–O peak area aer the plasma treatment. When the O2plasma was used for durations ranging from 1 to 5
min, shis in the binding energy were observed, with the peak shiing towards the chemical shi for native carbon peak. The change is clearly observed between the untreated sample and
Fig. 3 XPS spectra of the O1s region of ZnOfilms for O2plasma treatments with durations from 0 to 5 min.
Fig. 4 C1s spectra for untreated and O2plasma-treatedfilms on flexible PI substrates.
the samples that were treated with O2 plasma for durations
ranging from 1 to 5 min. This observation conrmed that the plasma treatment increased the oxidation of theber surfaces, transforming C–O groups into C]O and COOR groups. The XPS spectra of the C1s region yieldedve major peaks composed of ve components, of which the binding energies of C–C and C–H were normally assigned at 284.6–285.8 eV. Chemical shis in the binding energy are usually assigned for each functional group, such as C–O, C]O, and COOH. Thus, the plasma treatment makes the support surface rich in oxygen functional groups. These three C1s peaks (284.7, 285.9, and 288.8 eV) are attributed to adventitious surface contamination arising from (CHx)-like carbon and carbon oxides. For the O2plasma
treat-ment and the subsequently annealed lms (at 250 C),
complicated and unassigned peaks were observed with core level carbon peaks of C–C and C–H at 284.6 eV. These assign-ments from data on aexible PI substrate are in agreement with the observed C1s core level peaks.22The O
2plasma treatment
leads to the creation of active species and free radicals on the carbon support during the plasma treatment. Weak or unstable boundary layers on the surfaces of the carbon supports are removed by the O2plasma treatment. Thus, such changes easily
introduce oxygen functional groups onto the surface of the carbon supports. However, using a plasma power for a longer duration results in etching of the–Zn–O– surface, which affects the electrical performance of thelm. We have examined these effects on a ZnO lm used in a TFT application to demonstrate the tuning of the carrier mobility with facile O2plasma. From
these analyses and the measured electrical properties from the ZnO-TFT devices, we conclude that only short duration plasma treatments enhance the properties of the ZnO thinlm-based electrical devices.
FTIR spectroscopy was used to examine the functional groups in the spectrum of the ZnO precursorlm prepared from zinc acetate dissolved in ethanol to gain more insight into the effect of the O2 plasma treatment. Fig. 5 shows the FTIR
absorption spectrum of the as-deposited (i.e., ZnO precursor lm) and the O2plasma-treatedlms for treatment durations of
1–5 min on PI substrates. For the untreated sample, the band at
3418 cm1 is due to O–H species and the absorbed water
molecules in thelm, and the band at 29 100 cm1is due to C– H stretching frequencies. The band observed at 1592 cm1is due to C]O arising from the bridging-type bond turning into acetate bonding (OCOO). The bands at 1430 and 1015 cm1 are due to the C–O stretching frequencies, and the band at 1330 cml is due to weakly bound acetate molecules (HOOC–R), which is consistent with the previous reports.23 The band
observed at 450–580 cm1 is due to the Zn–O stretching
frequencies.24For a sample treated with O
2plasma for 1 min,
the spectra were nearly identical with those of an untreated sample, except that the O–H band shied slightly towards lower frequencies. The spectra of ZnO precursorlms treated with O2
plasma for 2–5 min exhibited an absence of absorption bands corresponding to organics and hydroxyls, indicating the complete removal of the organics and hydroxyls. According to the XPS spectra shown in Fig. 4, most of the organic part of the CH3COO group of zinc acetate and other volatile parts (C–H,
O–H, C–O, etc.) were removed for O2plasma treatments with
durations from 2 to 5 min. These FTIR spectral results clearly conrmed that the bonding force of the acetate anion with zinc cations decreases with the phase transformation, the OH-groups gradually replaced the acetate OH-groups coordinated to the matrix zinc cations, and the acetate groups werenally released completely, thus enabling ZnO to be formed with the release of the acetate anion.
3.3 Effect of the O2plasma on the electrical properties and
the charge carrier density of the ZnO TFTs fabricated on exible PI substrates
To comprehensively investigate the application and the effects of the O2plasma-treatment on ZnOlms to be used as the active
semiconductor layer in TFTs, we used the TFT device congu-ration shown in Fig. 6(a), which was also used to determine the O2plasma treatment effect on mobility. A high-quality optical
image of the TFT devices on aexible PI substrate is shown in Fig. 6(b). Fig. 6(c) shows the transfer characteristics of the plasma-treated ZnO TFTs of samples that underwent treatment with O2plasma for durations ranging from 1 to 5 min. The Ion/
Ioffratio is 108with a mobility of 0.74 cm2V1s1for the sample
treated with O2plasma for only 1 min. As the plasma power was
used for a longer duration, the off-state leakage currents (Ioff) of
the ZnO TFTs of the plasma-treated samples, which were treated for durations ranging from 1 to 2 min, exhibited signicant improvement from 1.86 1011A to 5.90 1012A
and hence an increase in the Ion/Ioffratio from 107to 109, while
the mobility increased to 22.8. However, the degradation of Ioff indicates considerable changes in the carrier concentration and distribution aer O2plasma treatment for 2 min. The Iofflevel
increased by over two orders of magnitude for plasma treatment durations increasing from 3 to 5 min, which consequently degraded the ZnO-TFT performance by decreasing the mobility. To understand the performance of the ZnO TFTs with various Fig. 5 FTIR spectra for untreated and O2plasma-treatedfilms on flexible PI substrates.
O2 plasma treatment durations, the extracted TFT-device
parameters are listed in Table 1. The saturation mobility in the thin-lm transistors was generally estimated using the following equation:25
ID¼W
2LmsatCiðVG VthÞ2 (2)
wheremsatis the eld-effect charge carrier mobility, Ci is the
capacitance per unit area of the dielectric, Vthis the threshold
voltage, and W and L are the ZnO TFT's channel width and length, respectively. In Fig. 6(d), a curve is also shown for the variation in mobility at different treatment times, ranging from 1 to 5 min, for the ZnO lms used as a channel layer. As described in Table 1, the ZnO TFT sample treated with O2
plasma for 1 min exhibited a good saturation mobility, threshold voltage and Ion/Ioff ratio. However, the ZnO TFTs
treated by O2plasma for 2 and 3 min exhibited excellent
satu-ration mobilities, which were 10 times greater than the mobility of the sample treated for 1 min under plasma power. Similar reductions in Ioff, Vthand mobilities were observed for the ZnO
TFT devices with the ZnO lm surface treated with plasma power for 5 min or longer. This degradation can be interpreted as the adsorption of highly electronegative active O2ions that
affect the depletion layer in the ZnO lm. Remote O2plasma
treatments at 18 W for 5 min or longer led to a shi in the
turn-on voltage and a reductiturn-on in the off-current by more than two orders of magnitude in the ZnO thinlm transistors. The effect of O2plasma power on solution-processed ZnOlm used into
TFTs application by XPS analyses is supported by results from the changes in the mobilities with time duration of the O2
plasma process.
Moreover, the effect of the O2plasma can also be interpreted
as the adsorption of active oxygen ions with highly electroneg-ative effects on the depletion layer in the ZnO lm. The carrier mobilities of the ZnO TFTs can be used to calculate the charge carrier concentrations using the equation26
n ¼ IDL
qVDmsatWd (3)
where n is the carrier concentration, IDis the measured drain
current at VD, L is the channel length, W is the channel width, d
is the gate insulator thickness, q is the electron charge, andmsat
is the carrier mobility. The carrier concentrations for the ZnO thin-lms treated with O2plasma treatments for durations of
1–5 min were calculated as 1.52 1017 cm3 for the 1 min
sample, 1.58 1019cm3for the 2 min sample, 4.3 1018cm3 for the 3 min sample, 1.75 1018cm3for the 4 min sample and 1.58 1017 cm1 for the 5 min sample. The resulting charge carrier concentrationrst increased for the samples with 1 to 2 min treatments and then decreased for the samples with 3 to 5 min treatments. Further O2 plasma treatment of longer
than 5 min resulted in devices with carrier concentrations that decreased as the O2ratio increased. This observation could be
due to the oxygen vacancies compensated by oxygen, causing the channel layer to be less conductive by reducing the supply of free electron carriers to the conduction band. In other words, a higher O2 ratio induces a lower density of oxygen vacancies,
which usually act as shallow donors. The remaining electrons of the deposited ZnOlm should be lower for a lm with a high O2
ratio caused by O2 plasma treatment of longer duration (3–5
min) than for alm with a low O2 ratio (i.e., low O2 plasma
treatment time).27 Moreover, the interaction among the O 2
plasma surface ions and the oxygen vacancies inuences the carrier concentration in a positive manner.
Fig. 7(a)–(e) show the output characteristics for ZnO lm-based transistors treated using the O2plasma process for 1–5
min. A good ohmic contact property at the ZnO interfaces is observed for all samples due to the nearly identical switching on drain current (ID) at 1 V drain voltage (VD). Under various gate
voltages, all the drain currents sharply increase with a drain voltage in the range of 1–5 V, gradually increase with a drain Fig. 6 (a) Schematic configuration of the ZnO-based TFT structure on a PI
substrate, (b) an optical image showing the TFTs on aflexible PI substrate, (c) transfer (ID–VG) characteristics measured at a constantVD¼ 5 V for O2plasma treated ZnO precursorfilms for different treatment durations ranging from 0 to 5 min, and (d) a plot of the variation in mobility for different treatment durations ranging from 1 to 5 min.
Table 1 Various electrical parameters of ZnO TFTs onflexible PI substrates treated with different O2plasma surface treatment durations ranging from 1 to 5 min
O2Plasma-treatment
time (min) Ion(A) Ioff(A) Ion/Ioffratio Vth(V) msat(cm2V1s1)
1 2.46 104 1.86 1011 107 13.9 0.74
2 8.26 103 5.90 1012 109 12.9 22.8
3 6.48 103 6.39 1012 109 11.5 16.2
4 3.11 103 3.96 1011 108 13.2 8.7
5 1.25 103 6.84 1010 106 10.9 3.1
voltage in the range of 5–20 V, and saturate with a drain voltage in the range of 20–40 V. Upon repeated sweep tests, the ZnO TFTs exhibit consistent performance, with a clear linear curve and saturation behaviors. This observation reects the lack of degradation of the device performance from repeated ID–VD
tests only for the 1–2 min O2 plasma process. Furthermore,
upon increasing the duration of the plasma power to 3 to 5 min or longer, the maximal output drain current at a gate voltage of 28 V isrst increased for treatment durations ranging from 1 to 3 min and then decreased from 3 to 5 min. The low saturation currents and carrier mobilities of ZnO TFTs with the ZnOlm treated by O2plasma for longer durations of approximately 5
min or longer are caused by the large barrier heights for elec-tron injection at the Au/ZnO interface aer O2 plasma
treat-ment. This phenomenon was explained in the previous section as most likely being due to the number of electrons, oxygen vacancies and carrier scattering effect. Therefore, the annealing temperature strongly affects the output performance of the ZnO TFTs, and the selection of a suitable annealing temperature is a very critical parameter for zinc acetate solution-processed ultra-thin lm transistors in future industrial applications. We conclude that the electrical properties from our solution-deposited ZnO TFTs overexible substrates are best for ZnO lms treated by the low power plasma process for durations of up to 3 min; this result is similar to the previously published results of ZnO TFT's properties over other substrates.28 3.4 Model for O2plasma treatment of the ZnOlm surface
and the energy band diagram
Because of the uncertain knowledge of the conductive processes in eosin and ZnOlms, the incorporation of the experimental results into a model for sensitization must be speculative. O2
plasma is known to contain various oxygen ions and radicals,
such as O+, O, O2+and O*, especially with the atomic radical as the main species.29 The oxygen radical atoms, O*, are then expected toll oxygen vacancies during diffusion through the ZnO bulk, reducing the free carrier concentration and thus the dark current level. Based on these observations, we designed a possible model of O2 plasma for the threshold voltage shi
depending on the oxygen ratio, as shown in Fig. 8(a), and further studied how this effect is explained by an energy band diagram model, as illustrated in Fig. 8(b).
The properties of the ZnO TFTs were highly inuenced by oxygen ions during their O2plasma surface interaction, which
may be because of oxygen vacancies that provide the needed free carriers for electrical conduction because oxygen vacancies could be easily generated in the oxides.30Fig. 8(a) shows the schematic mechanisms for the proposed O2desorption model
under O2 plasma for a ZnO thin lm. The O2 concentration
effect on a ZnO thin lm is described for the control of the threshold voltage (Vth) parameter due to varying O2treatment.
This description involves understanding how the oxygen ratios inuence the electronic properties and how the electronic behavior can be controlled. The O2adsorbed on the ZnOlm,
which can capture electrons in the back channel, exists as O2 in the form of O2 + e / O2, leading to a decrease in the
carrier concentration of the ZnOlm and an increase in the Vth
shi. The properties of the ZnO TFTs were highly inuenced by the O2plasma surface interaction, which may be because of the
oxygen vacancies provide the necessary free carriers for elec-trical conduction in ZnOlm used as the conducting medium. However, under O2plasma treatment for longer durations (>5
min), the electrons generated from the plasma induce electron– hole pairs that react with the absorbed O2, thus reducing O2in
the back channel and resulting in a Vthshi because the energy
barrier is proportional to the square of the density of surface trap states at the grain boundaries and inversely proportional to the free charge carrier density. The observed decrease in charge carrier mobility is associated with either a decrease in the charge carrier density or an increase in the surface trap density.31Thus, we conclude that a reduction of charge carrier density occurs due to the oxygen adsorbed on ZnO surfaces. Furthermore, the adsorption of oxygen at the grain boundaries could also lead to an increase in the surface trap state density at the grain boundaries, which would also increase the barrier Fig. 7 (a–e) Output characteristics (ID–VD) for O2plasma-treated ZnO channel
layers prepared at different treatment times ranging from 1 to 5 min; the gate bias (positive voltage) ranges from 0 to 40 V.
Fig. 8 (a) Possible model of the effect of O2plasma for the threshold voltage shift dependence on the oxygen ratio. (b) Energy band diagram illustrating the photo-electric process details in ann-channel ZnO-TFT including the behavior of interface trap charges under O2plasma.
height and thus reduce the charge carrier mobility. Fig. 8(b)
shows the energy band of the ZnO TFTs under O2 plasma
treatment. The subgap electron excitation can occur from the deep-subgap to the conduction band under oxygen ions when the O2plasma provides sufficient energy (>1.7 eV), resulting in
increases in the free electron concentration in the ZnOlm. However, the electron concentration is seriously affected for O2
plasma treatments of longer duration. All the characterized ZnO TFTs with the long duration O2 plasma treatment exhibit a
reduced free charge carrier density and thus a lower carrier mobility.
4.
Conclusion
The recent advances in transparent exible TFTs have high-lighted the use of low-cost technology and materials to replace the most commonly used semiconductor materials. We have successfully used a simple and cost-effective technique for preparing ZnOlms, which includes a spin-coating technique using a zinc-acetate precursor solution and a low power O2
plasma process. A high-mobility ZnO TFT on a exible PI
substrate was demonstrated with an O2plasma process
dura-tion of 2 min. Moreover, the ZnO precursorlm treated with O2
plasma for 1–5 min conrmed the signicant changes in the electrical properties of the ZnO TFT caused by the active oxygen ions reacting with residual hydrocarbons and the production of excess oxygen vacancies near the Zn atoms. We demonstrated the unique ability of adjusting the charge carrier concentration of ZnO of the microcrystalline layers by controlling the humidity in which the thermal annealing at 250C step was performed aer the O2 plasma treatment. With all these benets shown
above, we believe that the present work demonstrates an effi-cient pathway for realizing the full potential of transparent exible TFTs in the near future.
Acknowledgements
The authors are grateful to the National Device Laboratories for their support in the device fabrication, the National Science Council of Taiwan for nancially supporting this research under the contract NSC 101-2113-M-009-007-MY3 and to the Ministry of Education of Taiwan fornancially supporting this research under the Aiming for the Top University Program.
Notes and references
1 A. Janotti and C. G. van De Walle, Rep. Prog. Phys., 2009, 72, 126501.
2 (a) Z. L. Wang, X. Y. Kong, Y. Ding, P. Gao, W. L. Hughes, R. Yang and Y. Zhang, Adv. Funct. Mater., 2004, 14, 943; (b) N. K. Hassan, M. R. Hashim, Y. Douri and K. Al-Heuseen, Int. J. Electrochem. Sci., 2012, 7, 4625.
3 (a) Y.-S. Choi, J.-W. Kang, D.-K. Hwang and S.-J. Park, IEEE Trans. Electron Devices, 2010, 57, 26; (b) K. Laurent, B. Q. Wang, D. P. Yu and Y. Leprince-Wang, Thin Solid Films, 2008, 517, 617; (c) O. Maksimov, Rev. Adv. Mater. Sci., 2010, 24, 26; (d) M. H. Huang, S. Mao, H. Feick,
H. Yan, Y. Wu, H. Kind, E. Weber, R. Russo and P. Yang, Science, 2001, 292, 1897; (e) B. K. Roberts, A. B. Pakhomov, V. S. Shutthanandan and K. M. Krishnan, J. Appl. Phys., 2005, 97, 10D310.
4 ¨U. ¨Ozg¨ur, Ya. I. Alivov, C. Liu, A. Teke, M. A. Reshchikov, S. Do˘gan, V. Avrutin, S.-J. Cho and H. Morkoç, J. Appl. Phys., 2005, 98, 041301.
5 P. K. Shishodia, H. J. Kim, A. Wakahara, A. Yoshida, G. Shishodia and R. M. Mehra, J. Non-Cryst. Solids, 2006, 352, 2343.
6 (a) K.-S. Jang, D. Wee, Y. H. Kim, J. Kim, T. Ahn, J.-W. Ka and M. H. Yi, Langmuir, 2013, 29, 7143; (b) S. Y. Cho, Y. H. Kang, J.-Y. Jang, S. Y. Nam, J. Lim, S. C. Yoon, D. H. Choi and C. Lee, Chem. Mater., 2012, 24, 3517; (c) Y. Jung, T. Jun, A. Kim, K. Song, T. H. Yeo and J. Moon, J. Mater. Chem., 2011, 21, 11879; (d) K. Kim, S. Park, J.-B. Seon, K.-H. Lim, K. Char, K. Shin and Y. S. Kim, Adv. Funct. Mater., 2011, 21, 3546; (e) S. Y. Park, B. J. Kim, K. Kim, M. S. Kang, K.-H. Lim, T. I. Lee, J. M. Myoung, H. K. Baik, J. H. Cho and Y. S. Kim, Adv. Mater., 2012, 24, 834.
7 (a) E. M. C. Fortunato, P. M. C. Barquinha,
A. C. M. B. G. Pimentel, A. M. F. Gonçalves,
A. J. S. Marques, L. M. N. Pereira and R. F. P. Martins, Adv. Mater., 2005, 17, 590; (b) C. C. Liu, M. L. Wu, K. C. Liu, S.-H. Hsiao, Y. S. Chen, G.-R. Lin and J. J. Huang, IEEE J. Display Technol., 2009, 5, 192.
8 J. B. Franklin, B. Zou, P. Petrov, D. W. McComb, M. P. Ryan and M. A. McLachlan, J. Mater. Chem., 2011, 21, 8178. 9 D. S. Bhachu, G. Sankar and I. P. Parkin, Chem. Mater., 2012,
24, 4704.
10 (a) E. Guziewicz, M. Godlewski, T. Krajewski,Ł. Wachnicki,
A. Szczepanik, K. Kopalko, A. W´ojcik-Głodowska,
E. Prze´zdziecka, W. Paszkowicz, E. Łusakowska,
P. Kruszewski, N. Huby, G. Tallarida and S. Ferrari, J. Appl. Phys., 2009, 105, 122413; (b) M. A. Thomas and J. B. Cui, ACS Appl. Mater. Interfaces, 2012, 4, 3122.
11 K. Park, D.-K. Lee, B.-S. Kim, H. Jeon, N.-E. Lee, D. Whang, H.-J. Lee, Y. J. Kim and J.-H. Ahn, Adv. Funct. Mater., 2010, 20, 3577.
12 (a) H. C. Cheng, C. F. Chen and C. Y. Tsay, Appl. Phys. Lett., 2007, 90, 012113; (b) C. S. Li, Y. N. Li, Y. L. Wu, B. S. Ong and R. O. Loutfy, J. Mater. Chem., 2009, 19, 1626.
13 (a) J.-S. Lee, S.-M. Song, Y.-H. Kim, J.-Y. Kwon and M.-K. Han, Phys. Status Solidi A, 2013, 1–5; (b) M.-C. Chu, J. S. Meena, P.-T. Liu, H.-P. D. Shieh, H.-C. You, Y.-W. Tu, F.-C. Chang and F.-H. Ko, Appl. Phys. Express, 2013, 6, 076501; (c)
H. Faber, J. Hirschmann, M. Klaum¨unzer,
B. Braunschweig, W. Peukert and M. Halik, ACS Appl. Mater. Interfaces, 2012, 4, 1693; (d) J. S. Meena, M.-C. Chu, S.-W. Kuo, F.-C. Chang and F.-H. Ko, Phys. Chem. Chem. Phys., 2010, 12, 2582.
14 (a) S.-Y. Han, G. S. Harman and C.-H. Chang, J. Am. Chem. Soc., 2011, 133, 5166; (b) R. Martins, E. Fortunato,
P. Nunes, I. Ferreira, A. Marques, M. Bender,
N. Katsarakis, V. Cimalla and G. Kiriakidis, J. Appl. Phys., 2004, 96, 1398.
15 (a) P.-T. Liu, Y.-T. Chou and L.-F. Teng, Appl. Phys. Lett., 2009, 95, 233504; (b) P. P. Edwards, A. Porch, M. O. Jones, D. V. Morgan and R. M. Perks, Dalton Trans., 2004, 2995; (c) T.-Y. Tseng and S. M. Sze, Nonvolatile Memories: Materials, Devices and Applications, American Scientic Publishers, Los Angeles, California, USA, 2012, vol. 1. 16 (a) R. Eason, Pulsed Laser Deposition of Thin Films:
Applications-LED Growth of Functional Materials, Wiley-Interscience, A John Wiley & Sons, INC., Publications, New York, 2007; (b) M. S. Hammer, D. Rauh, V. Lorrmann, C. Deibel and V. Dyakonov, Nanotechnology, 2008, 19, 485701. 17 V. A. Johnson and K. Lark-Horovitz, Phys. Rev., 1953, 92, 226. 18 J. Y. Lee, C. M. Shin, J. H. Heo, C. R. Kim, J. H. Park, T. M. Lee, H. Ryu, C. S. Son, B. C. Shin and W. J. Lee, Current Applied Physics, 2010, 10, S290.
19 (a) M. Mekhnache, A. Drici, L. S. Hamideche, H. Benzarouk, A. Amara, L. Cattin, J. C. Bernede and M. Guerioune, Superlattices Microstruct., 2011, 49, 510; (b) Z. L. Pei, X. B. Zhang, G. P. Zhang, J. Gong, C. Sun, R. F. Huang and L. S. Wen, Thin Solid Films, 2006, 497, 20.
20 P. K. Nayak, M. N. Hedhili, D. Cha and H. N. Alshareef, Appl. Phys. Lett., 2012, 100, 202106.
21 Z. Fan, D. Wang, P.-C. Chang, W.-Y. Tseng and J. G. Lu, Appl. Phys. Lett., 2004, 85, 5923.
22 (a) M.-G. Kim, H. S. Kim, Y.-G. Ha, J. He, M. G. Kanatzidis, A. Facchetti and T. J. Marks, J. Am. Chem. Soc., 2010, 132, 10352; (b) J. Xie, D. Xin, H. Cao, C. Wang, Y. Zhao, L. Yao,
F. Ji and Y. Qiu, Surf. Coat. Technol., 2011, 206, 191; (c) K. Y. Rhee, S. J. Park, D. Hui and Y. Qiu, Composites, Part B, 2012, 43, 2395.
23 (a) M. N. Kamalasanan and S. Chandra, Thin Solid Films, 1996, 288, 112; (b) A. Kelchtermans, K. Elen, K. Schellens,
B. Conings, H. Damm, H.-G. Boyen, J. D'Haen,
P. Adriaensens, A. Hardy and M. K. Van Bael, RSC Adv., 2013, 3, 15254.
24 S. K. Sharma, A. I. Inamdar, H. Im, B. G. Kim and P. S. Patil, J. Alloys Compd., 2011, 509, 2127.
25 (a) C. Busch, G. Schierning, R. Theissmann and
R. Schmechel, J. Appl. Phys., 2012, 112, 034506; (b) J. S. Meena, M.-C. Chu, C.-S. Wu, Y.-C. Chang, F.-C. Chang and F.-H. Ko, Org. Electron., 2011, 12, 1414; (c) L. Li, W. Hu, L. Chi and H. Fuchs, J. Phys. Chem. B, 2010, 114, 5315. 26 I. Abdel-Motaleb, N. Shetty, K. Leedy and R. Cortez, J. Appl.
Phys., 2011, 109, 014503.
27 C. G. Van de Walle, Phys. Rev. Lett., 2000, 85, 1012.
28 (a) B. S. Ong, C. Li, Y. Li, Y. Wu and R. Loutfy, J. Am. Chem. Soc., 2007, 129, 2750; (b) K. Song, J. Noh, T. Jun, Y. Jung, H.-Y. Kang and J. Moon, Adv. Mater., 2010, 22, 4308. 29 M. Liu and H. K. Kim, Appl. Phys. Lett., 2004, 84, 173. 30 C. W. Litton, D. C. Reynolds and T. C. Collins, Zinc Oxide
Materials for Electronic and Optoelectronic Device
Applications, Wiley, Chichester, UK, 2011.
31 S. Walther, S. Polster, M. P. M. Jank, H. Thiem, H. Ryssel and L. Frey, Adv. Powder Technol., 2011, 22, 253.
