Device characteristics of Schottky barrier diodes using In-Ga-Zn-O semiconductor thin films with different atomic ratios

Device characteristics of Schottky barrier diodes using In-Ga-Zn-O semiconductor thin films with different atomic ratios

Accepted Manuscript Device characteristics of Schottky barrier diodes using In-Ga-Zn-O semiconductor thin films with different atomic ratios Jae-Won K...

969KB Sizes 0 Downloads 13 Views

Accepted Manuscript Device characteristics of Schottky barrier diodes using In-Ga-Zn-O semiconductor thin films with different atomic ratios Jae-Won Kim, Tae-Jun Jung, Sung-Min Yoon PII:

S0925-8388(18)33184-0

DOI:

10.1016/j.jallcom.2018.08.289

Reference:

JALCOM 47377

To appear in:

Journal of Alloys and Compounds

Received Date: 24 April 2018 Revised Date:

27 August 2018

Accepted Date: 28 August 2018

Please cite this article as: J.-W. Kim, T.-J. Jung, S.-M. Yoon, Device characteristics of Schottky barrier diodes using In-Ga-Zn-O semiconductor thin films with different atomic ratios, Journal of Alloys and Compounds (2018), doi: 10.1016/j.jallcom.2018.08.289. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

ACCEPTED MANUSCRIPT

Device characteristics of Schottky barrier diodes using In-

ratios

RI PT

Ga-Zn-O semiconductor thin films with different atomic

SC

Jae-Won Kim, Tae-Jun Jung, and Sung-Min Yoon*

Department of Advanced Materials Engineering for Information and Electronics,

M AN U

Kyung Hee University, Yongin-shi, Gyeonggi-do 17104, Korea Electronic mail: (*)[email protected]

ABSTRACT

Oxide semiconductor Schottky barrier diodes (SBDs) were fabricated by using

TE D

amorphous In-Ga-Zn-O (IGZO) semiconducting thin films with different atomic ratios. Higher rectification ratios and Schottky barrier heights (SBHs) were obtained when the oxygen partial pressure was high during the sputtering deposition of the IGZO films.

EP

The increase in Ga composition effectively enhanced the device characteristics,

AC C

including the rectification ratio and the SBH of the SBDs. These properties were closely related to the control of oxygen vacancy concentration within the IGZO and the resulting conduction behaviors owing to Fermi-level pinning and tunneling current through the Schottky barrier. The fabricated SBD using IGZO with a higher Ga composition (In:Ga:Zn=1.0:0.8:0.3) exhibited a rectification ratio of 8.3×106 and an SBH of 0.79 eV.

Keywords: Oxide semiconductor, Schottky junction, Schottky barrier diode, In-Ga-Zn1

ACCEPTED MANUSCRIPT

O

1. Introduction

RI PT

Thin film transistors (TFTs) using the amorphous oxide semiconductor In-Ga-Zn-O (aIGZO) have been actively researched and developed for display backplane devices [1-3]. Furthermore, a-IGZO TFTs have attracted significant interest because they are

SC

promising candidates for various consumer electronics applications including the Internet of Things (IoT) technologies [4-6]. There have been several key challenges in

M AN U

these technical fields, such as ultra-low power consumption, easy manufacturing, higher device performance and reliability, associated with switching and driving device elements of the specified circuits and integrated systems. a-IGZO TFTs with conventional gate-stack structures of gate insulators (GIs) and active channel layers are

TE D

mainly employed for large-area electronic applications [7-8]. Even so, a metalsemiconductor field-effect transistor (MESFET) using a Schottky-gated junction between a metal and semiconductor can also be a good solution to overcome the

EP

limitations of the conventional oxide TFTs, including higher operating voltage and

AC C

undesirable charge trapping at the GI/active interfaces. From these backgrounds, the Schottky barrier diodes (SBDs) using the Schottky contacts (SCs) between the metal and oxide semiconductor have also been presented for various applications [9-12]. Alternatively, the MESFETs can be simply fabricated without GIs and be can operated at lower voltages with higher switching frequency thanks to their GI-free device structures [13-15].

However, to fully exploit the benefits of MESFETs, it is very important to obtain stable 2

ACCEPTED MANUSCRIPT

Schottky junctions. This is because it is difficult to form stable SCs at interfaces between the metal and a-IGZO due to the intrinsic defects, especially oxygen vacancies, located within the a-IGZO layers. Several approaches have been taken to solve these

RI PT

problems, such as oxygen plasma-treated metal electrodes [16-17], insertion of a thin insulator into the SC [18], thermal annealing of the anode metal [19], and additional heat treatments of semiconductors in an oxygen ambient prior to metallization [20].

SC

Sometimes, the oxygen partial pressure (PO2) was preferentially controlled during the deposition of a-IGZO to obtain both Schottky and ohmic contacts [21, 22]. It is also

M AN U

important to control the sputtering conditions including sputtering power, working pressure. These techniques are closely related to the charge distribution and defective layer formation on the a-IGZO surfaces. Alternatively, the defect structures and corresponding electronic nature of the a-IGZO thin films are sensitively influenced by

TE D

the film compositions. Consequently, controlling the composition of the a-IGZO can be an easy and powerful way to enhance the electrical characteristics of the SCs at metal/a-

EP

IGZO interfaces without performing any additional processes.

AC C

In this work, the SBDs were fabricated using a-IGZO thin films with two different compositions and were characterized prior to the fabrication of MESFETs. There have been few reports on the effects of composition variations in a-IGZO thin films with respect to the device characteristics of the SBDs. Variations in the chemical and electrical properties of the composition-modified a-IGZO had significant impacts on the device performance of the fabricated SBDs.

2. Experimental 3

ACCEPTED MANUSCRIPT

IGZO SBDs were fabricated on thermally oxidized Si substrates. First, a 200-nm-thick Pt thin film was deposited as a bottom electrode to form the SCs after the formation of a 10-nm-thick Ti adhesion layer. Firstly, PO2 was varied from 1 to 20% during the radio-

RI PT

frequency (RF) sputtering process to optimize the PO2 condition for the deposition of the IGZO layer. Investigations of PO2 variations during IGZO sputtering can provide meaningful information on the compositional effects that influence on the device

SC

characteristics of the IGZO SBDs. Then, two ceramic targets with different compositions were used for the IGZO thin film depositions. The atomic ratios (In:Ga:Zn)

M AN U

of sputtered films were analyzed by X-ray photoelectron spectroscopy (XPS) using the Al Kα line, and peak fitting and comparative analysis for the In 3d, Ga 2p, and Zn 2p spectra were performed. The atomic ratios were estimated to be 1.0:0.5:0.8 and 1.0:0.8:0.3, and these samples were referred to IGZO-A and IGZO-B, respectively. The

TE D

film thickness was adjusted to be 80 nm. Thermal treatment was carried out right after the IGZO deposition in a rapid thermal annealing (RTA) system at 200 oC in an oxygen ambient for 1 h. This mid-annealing process can help adjust the oxygen stoichiometry at

EP

Schottky junctions [23]. There were no significant changes in film composition before

AC C

and after the post-annealing process performed at 200 oC in an oxygen ambient. Finally, indium-tin-oxide (ITO) thin films were deposited and patterned via a lift-off process as top ohmic contacts providing a device area of 230×230 µm2. The device schematic of the fabricated SBDs is shown in Fig. 1(a). The detailed fabrication procedures and process conditions are summarized in Fig. 1(b).

The current-voltage (I-V) and capacitance-voltage (C-V) characteristics were measured using a semiconductor parameter analyzer (Keithley 4200SCS) in a dark box at room 4

ACCEPTED MANUSCRIPT

temperature. The composition and bonding characteristics of the IGZO films with different compositions were examined by XPS. The temperature-dependent electrical

RI PT

conductivity was estimated by the in-situ four-point probe method.

3. Results and discussion

SC

SBDs were first prepared using the IGZO-A with variations in PO2 from 1 to 20% to determine the optimum PO2 condition for the deposition of IGZO layers with given

M AN U

compositions. Fig. 2(a) shows the I-V curves of the fabricated IGZO SBDs. The device characteristics of the SBD were sensitively influenced by the PO2 conditions for the IGZO-A formation. The rectification performance and Schottky barrier height (SBH) of the SBDs were improved with increasing PO2 during the IGZO sputtering process. These

TE D

results originated from the fact that the conduction carrier concentration and oxygen vacancies within the IGZO film decreased with an increase in PO2. Defect-induced subgap states within the IGZO thin films commonly originate from oxygen vacancies [22].

EP

Fermi-level pinning happens for the IGZO because there are a significant number of surface sub-gap state densities [16]. On the other hand, the defect densities can be

AC C

reduced with increasing the PO2. Hence, the Fermi-level pinning in the IGZO can also be avoided. Moreover, undesirable tunneling phenomenon through the Schottky barrier may be enhanced if a large number of defects are located near the space-charge region [24]. Degradation in the rectification performance of the SBDs is closely related to an excessive amount of defect states [25]. Consequently, a high SBH and good rectification characteristics could be obtained when the PO2 was controlled to be high. It is interesting to note that the ion bombardment effects of negatively-charged oxygen ions during the IGZO sputtering process also have an important effect on the device characteristics [26]. 5

ACCEPTED MANUSCRIPT

In other words, the conductive layer containing OH- can be eliminated by plasma treatment, and the oxygen vacancies located near the surface can also be filled up. Thus, the rectifying performance of the IGZO SBDs can be markedly improved [27]. The

RI PT

introduction of higher PO2 during the IGZO deposition can result in surface oxidation of the Pt bottom electrode. The rectifying characteristics of the IGZO SBDs can also be improved due to the higher work function of PtO formed on the surface of the Pt

SC

electrode. The obtained results suggest that higher PO2 conditions during the IGZO sputtering deposition are desirable for the preparation of good Schottky junctions of the

M AN U

devices.

Then, SBDs using two compositions of IGZO-A and IGZO-B were prepared to investigate the IGZO compositional effects, in which the PO2 condition was fixed at

TE D

20 %. Fig. 2(b) shows the I-V characteristics of the fabricated SBDs. The current transport via a Schottky junction can be generally described by the thermionic emission of majority carriers [28] over the junction barrier using Eq. (1). , 

  

  

EP

 = ∗   



 − 1!

(1)

AC C

A*, A, ΦB,IV, Rs, and n correspond to the effective Richardson constant, diode contact area, SBH obtained from I-V curves, series resistance and ideality factor, respectively. These parameters for both SBDs could be obtained by fitting the I-V curves in the forward bias regions with Eq. (1) and were summarized in Table I. While there was no marked difference in on-current (ION) between the two compositions, the off-current (IOFF) of the IGZO-B SBD was lower than that of the IGZO-A SBD. The Richardson constant was varied depending on the crystal planes and the applied strain conditions. Thus, although this constant is a variable that changes with the composition of the 6

ACCEPTED MANUSCRIPT

employed IGZO, the same value of the effective Richardson constant of 41 A·cm-2·K-2 was used to characterize the device parameters of the SBDs using two different compositions because there were not any marked variations in constituent elements

RI PT

between two IGZO compositions prepared in this work. The rectification ratios (ION/OFF) of the SBDs using IGZO-A and IGZO-B at ± 1 V were estimated to be approximately 2.1×106 and 8.3×106, respectively. It was suggested from the estimated device

SC

parameters that the IGZO-B SBD exhibited superior diode characteristics including

M AN U

higher SBH, lower n, and higher ION/OFF.

C-V characteristics were also evaluated to examine the device performances and were quantitatively compared for two SBDs, as shown in Fig. 3(a). Actually, the surface conditions and interface qualities between the metals and oxide semiconductors are

TE D

supposed to critically influence the device characteristics of the SBDs. Thus, the measurement frequency was set to 1 MHz to minimize the redundant capacitance induced by the interface trap densities [29] because of difficulties in extracting accurate

EP

values of interface trap densities. The measured capacitance at the Schottky junction can be plotted using Eq. (2) [4]. = %&



. /01 −

AC C

"#

$#

' &( )*+,-



− /

(2)

ԑs and ԑo correspond to the relative dielectric constant of the semiconductor and the vacuum dielectric constant, respectively. Ndepl and Vbi correspond to the charge concentration within the depletion region and the junction built-in potential, which can be obtained from the x-axis intercept and slope of the A2/C2-V curves, respectively. The obtained C-V characteristics were also found from lots of preliminary measurements to be sufficiently reproducible for deriving the device parameters of the fabricated IGZO 7

ACCEPTED MANUSCRIPT

SBDs. The SBH (ΦB,CV) obtained from the C-V curve can be represented as Φ3,45 = q/01 +

9: ;

>

ln ?, where the values of free charge concentration (Ne) can be calculated >@

RI PT

by assuming an electron mobility of 10 cm2 V-1 s-1, and a conduction band density of states (NC) of 5.2×1018 cm-3 can be assumed for the parameter calculations [30]. The estimated device parameters are summarized in Table I. There was a discrepancy (∆ΦB)

SC

between the values of ΦB,IV and ΦB,CV for both devices, which originated from the local fluctuations in SBH [14]. The SBD currents are primarily determined by the lowest

M AN U

barrier at SCs between the a-IGZO and anode, while the capacitance is measured by averaging the physical properties of SCs [31]. Thus, the small values of ∆ΦB (0.15 and 0.01 eV) for both devices reveal that high-quality SCs were formed over the uniform SBH. These values agree well with that of both devices, which exhibited values of n as

TE D

low as 0.9. Above all, the noteworthy result from the C-V measurement is that the value of Ndepl for the IGZO-B SBD was 10-times smaller than that for the IGZO-A SBD. It is known that the high Ndepl causes a tunneling current through the Schottky barrier and

EP

deteriorates the rectifying characteristics [25]. Thus, the larger IOFF for the IGZO-A SBD was because the Ndepl was higher than 1018 cm-3 owing to the generation of

AC C

tunneling current. Furthermore, it is also interesting to understand the effects of Ndepl on the device characteristics. In covalently-bonded semiconductors, the Ndepl of the SBD represents the ionized dopant density, which contributes to the free-charge density. On the other hand, depletion region in amorphous semiconductors like a-IGZO also contains ionized atoms, which cannot contribute to the free-charge density owing to sub-gap trap states [16]. Considering that the defect-induced sub-gap states are primarily composed of oxygen vacancies in the a-IGZO [22], IGZO-B may have a smaller number of defect-induced sub-gap states (including the oxygen vacancies) than 8

ACCEPTED MANUSCRIPT

the IGZO-A. The band-offset energy diagrams of the Schottky junctions fabricated using IGZO-A and IGZO-B were constructed by the calculated numerical values, as

RI PT

shown in Fig. 3(b), respectively. Fig. 4(a) and 4(b) show the I-V characteristics of the fabricated SBDs using IGZO-A and IGZO B, respectively, when the measurement temperature was varied from room

SC

temperature to 120 oC. For both devices, on- and off-currents increased with increasing measurement temperature, and the increases in the off-currents in the rectifying region

M AN U

were much larger than increases in on-currents. As a result, the ION/IOFF decreased with increasing measurement temperature. This behavior is because SBH is lower at higher measurement temperatures. It was found that there were no marked variations in the electrical conductivities of the a-IGZO because of the relatively low temperate below 120 oC, as shown in Fig. 6. IGZO composition effects also had an effect. The ION/IOFF of

TE D

the IGZO-A device increased to 3.6×103 at 120 oC. The increase in ION/IOFF was suppressed below 2.0×104 for the IGZO-B device. This result suggests that the SBD

EP

using IGZO-B showed a stronger immunity against the increase in operation temperature than that using IGZO-A owing to the higher SBH for the IGZO-B, as

AC C

illustrated in Fig. 3.

XPS analysis were performed for O 1s spectra to examine the bonding characteristics of oxygen-related species in the a-IGZO films and investigate their physical effects on the device characteristics. It is important to analyze these behaviors because oxygen vacancies can be a major contributor to Fermi-level pinning [23], which lowers the SBH, at contacts between the metal and IGZO. Fig. 5(a) and 5(b) show the O 1s spectra and peak deconvolution fitting for IGZO-A and IGZO-B, respectively. Three deconvoluted 9

ACCEPTED MANUSCRIPT

peaks correspond to metal oxygen bonds (peak A, 529.9 eV), oxygen vacancies (peak B, 530.8 eV), and absorbed surface contamination (peak C, 532.1 eV) including -OH, -CO3, and H2O species [32]. The relative areal ratios of peak B to the total area of peaks A, B

RI PT

and C in the O 1s spectrum were estimated to be 45% and 34% for the IGZO-A and IGZO-B, respectively. For the IGZO-A composition, the atomic ratio of Ga:In (0.5) was smaller than that (0.8) for the IGZO-B. Since Ga ions form stronger chemical bonds

SC

with oxygen than In and Zn ion [1], the amounts of oxygen vacancies within the film can be intrinsically reduced. Thus, the SBD employing an IGZO-B composition can be

M AN U

less affected by the Fermi-level pinning in SC regions and tunneling current generated through the Schottky barrier, which are the main reasons for the decreased rectification ratio compared to the IGZO-A composition. These results agree well with the

TE D

comparisons in ION/OFF and Ndepl values for both devices.

To further analyze the differences in the electronic natures of the a-IGZO thin films with

EP

different compositions, variations in electrical conductivity (σc) were measured as a function of temperature from 50 to 300 oC using the in-situ four-point probe method, as

AC C

shown in Fig. 6. The ramping rate was fixed at 5 oC/min. In the low temperature region, both a-IGZO films were almost insulating with a σc of approximately 10-12 S cm-1, and carriers did not transport to provide electronic conduction due to insufficient thermal energy. At a glance, both a-IGZO compositions exhibited very similar temperaturedependent conduction behaviors. However, there are two marked differences. Firstly, it is noteworthy that σc values started to drastically increase at 173 and 193 oC in IGZO-A and IGZO-B, respectively. A lower number of intrinsic conduction carriers in the IGZOB needed a higher temperature to be activated for the electronic conduction [33]. This is 10

ACCEPTED MANUSCRIPT

closely related to strong Ga-O chemical bonding, which can suppress the excessive generation of oxygen vacancies. Alternatively, the activation energy (Ea) for carrier conduction was estimated using an Arrhenius plot. As can be seen in the figure, the

RI PT

temperature-dependent variations in the σc could be divided into lower and higher temperature regions. In the lower temperature region, the σc abruptly increased with the Ea value higher than 10 eV for both a-IGZO composition, and hence these behaviors

SC

cannot be simply explained by defect-induced electronic conduction. On the other hand,

M AN U

in the higher temperature region, the values of Ea were estimated to be approximately 2.30 and 2.58 eV for the IGZO-A and IGZO-B, respectively, as plotted in Fig. 6. This is the second noteworthy difference between the two compositions. This suggests that a larger quantity of energy would be required for the IGZO-B with a larger Ea to activate

4. Conclusions

TE D

the conduction electrons within the film.

SBDs using the oxide semiconductor IGZO thin films were fabricated and characterized.

EP

The effects of PO2 conditions during the IGZO sputtering were firstly examined to figure

AC C

out the role of oxygen in the formation of the SBH. A higher ION/OFF and SBH could be obtained when the PO2 was controlled to be high. This phenomenon is closely related to the amount of defect states and ion bombardment effects during the IGZO sputtering. Then, the effects of composition variations in the a-IGZO thin films on the device characteristics of the SBDs were investigated by modulating the Ga composition of aIGZO. Higher ION/OFF and SBH were obtained for the SBD using IGZO-B with higher Ga composition. XPS analysis reveals that the amounts of oxygen vacancies in the IGZO-B were estimated to be smaller than those for the IGZO-A. Excessive oxygen 11

ACCEPTED MANUSCRIPT

vacancies within the IGZO-A were suggested as the main origin for degradations in device performance owing to the Fermi-level pinning and tunneling current through the Schottky barrier. It was also found that defect-related conduction behaviors were not

RI PT

easily activated for the IGZO-B. These analyses of the temperature-dependent σc and bonding natures for the a-IGZO films with two different compositions suggest that the IGZO-B with fewer oxygen vacancies is preferable for SBDs with better device

Acknowledgements

M AN U

SC

performance.

This work was supported by the Kyung Hee University–Samsung Electronics Research and Development Program entitled Flexible Flash Memory Device Technologies for

AC C

EP

TE D

Next-Gen Consumer Electronics.

12

ACCEPTED MANUSCRIPT

References [1]

K. Nomura, H. Ohta, A. Takagi, T. Kamiya, M. Hirano, and H. Hosono, Roomtemperature fabrication of transparent flexible thin-film transistors using

[2]

RI PT

amorphous oxide semiconductors, Nature 432 (2004) 488-492.

N. On, Y. Kang, A. Song, B. D. Ahn, H. D. Kim, J. H. Lim, K. B. Chung, S. Han, and J. K. Jeong, Origin of electrical instabilities in self-aligned amorphous In–Ga–

[3]

SC

Zn–O thin-film transistors, IEEE Trans. Electron Dev. 64 (2017) 4965-4973.

S. M. Yoon, N. J. Seong, K. Choi, G. H. Seo, and W. C. Shin, Effects of

M AN U

deposition temperature on the device characteristics of oxide thin-film transistors using In−Ga−Zn−O active channels prepared by atomic-layer deposition, ACS Appl. Mater. Interfaces 9 (2017) 22676-22684. [4]

A. Chasin, V. Volskiy, M. Libois, K. Myny, M. Nag, M. Rockelé, G. A. E.

TE D

Vandenbosch, J. Genoe, G. Gielen, and P. Heremans, An Integrated a-IGZO UHF Energy Harvester for Passive RFID Tags, IEEE Trans. Electron Dev. 61 (2014) 3289-3295.

X. Du, Y. Li, J. R. Motley, W. F. Stickle, and G. S. Herman, Glucose Sensing

EP

[5]

AC C

Using Functionalized Amorphous In-Ga-Zn-O Field-Effect Transistors, ACS Appl. Mater. Interfaces 8 (2016) 7631-7637.

[6]

S. J. Kim, M. J. Park, D. J. Yun, W. H. Lee, G. H. Kim, and S. M. Yoon, High

performance and stable flexible memory thin-film transistors using In–Ga–Zn–O channel and ZnO charge-trap layers on poly(ethylene naphthalate) substrate, IEEE Trans. Electron. Dev. 63 (2016) 1557-1564.

[7]

E. Fortunato, P. Barquinha, and R. Martins, Oxide Semiconductor Thin-Film Transistors: A Review of Recent Advances, Adv. Mater. 24 (2012) 2945-2986. 13

ACCEPTED MANUSCRIPT

[8]

J. Sheng, H. J. Jeong, T. Hong, and J. S. Park, Review of recent advances in flexible oxide semiconductor thin-film transistors, J. Inf. Display 18 (2017) 159172. J. Zhang, H. Wang, J. Wilson, X. Ma, J. Jin, and A. Song, Room temperature

RI PT

[9]

processed ultrahigh-frequency indium-gallium-zinc-oxide Schottky diode, IEEE Electron. Dev. Lett. 37 (2016) 389-392.

SC

[10] S. Muller, H. von Wenckstern, F. Schmidt, D. Splith, H. Frenzel, and M. Grundmann, Method of choice for the fabrication of high-quality β-gallium oxide-

M AN U

based Schottky diodes, Semicond. Sci. Technol. 32 (2017) 065013. [11] P. Schlupp, H. von Wenckstern, and M. Grundmann, Schottky barrier diodes based on room temperature fabricated amorphous zinc tin oxide thin films, Phys. Status Solidi a 214 (2017) 1700210.

TE D

[12] J. Kaczmarski, M. A. Borysiewicz, K. Piskorski, M. Wzorek, M. Kozubak, and E. Kaminska, Flexible IGZO Schottky diodes on paper, Semicond. Sci. Technol. 33 (2017) 015010.

EP

[13] S. Lee and A. Nathan, Subthreshold Schottky-barrier thin-film transistors with

AC C

ultralow power and high intrinsic gain, Science 354 (2016) 302-304. [14] Y. Magari, S. Hashimoto, K. Hamada, and M. Furuta, Low-temperature processed metal-semiconductor field-effect transistor with In-Ga-Zn-O/AgOx schottky gate,

ECS Trans. 75 (2016) 139-144.

[15] J. Kaczmarski, A. Taube, M. A. Borysiewicz, M. Mysliwiec, K. Piskorski, K. Stiller, and E. Kaminska, Transparent Ru–Si–O/In–Ga–Zn–O MESFETs on Flexible Polymer Substrates, IEEE Trans. Electron Dev. 65 (2018) 129-135. [16] A. Chasin, S. Steudel, K. Myny, M. Nag, T.-H. Ke, S. Schols, J. Genoe, G. Gielen, 14

ACCEPTED MANUSCRIPT

and P. Heremans, High-performance a-In-Ga-Zn-O Schottky diode with oxygentreated metal contacts, Appl. Phys. Lett. 101 (2012) 113505. [17] J. Jin, J. Zhang, A. Shaw, V. N. Kudina, I. Z. Mitrovic, J. S. Wrench, P. R. Chalker,

RI PT

C. Balocco, A. Song, S. Hall, A high speed PE-ALD ZnO Schottky diode rectifier with low interface-state density, J. Phys. D. 51 (2018) 065102.

[18] M. J. Choi, M. H. Kim, and D. K. Choi, A transparent diode with high rectifying

SC

ratio using amorphous indium-gallium-zinc oxide/SiNX coupled junction, Appl. Phys. Lett. 107 (2015) 053501.

M AN U

[19] L. Du, H. Li, L. Yan, J. Zhang, Q. Xin, Q. Wang, and A. Song, Effects of substrate and anode metal annealing on InGaZnO Schottky diodes, Appl. Phys. Lett. 110 (2017) 011602.

[20] H. Frensel, A. Lajn, H. von Wenckstern, M. Lorenz, F. Schein, Z. Zhang, and M.

TE D

Grundmann, Recent Progress on ZnO-Based Metal-Semiconductor Field-Effect Transistors and Their Application in Transparent Integrated Circuits, Adv. Mater. 22 (2010) 5332-5349.

EP

[21] L. Yan, Q. Xin, L. Du, J. Zhang, Y. Luo, Q. Wang, and A. Song, High performance

AC C

InGaZnO-based Schottky diodes fabricated at room temperature, Phys. Status Solidi C, 13 (2016) 618-622.

[22] H. Qian, C. Wu, H. Lu, W. Xu, D. Zhou, F. Ren, D. Chen, R. Zhang, and Y. Zheng, Bias stress instability involving subgap state transitions in a-IGZO Schottky barrier diodes, J. Phys. D. 49 (2016) 395104. [23] D. H. Lee, K. Nomura, T. Kamiya, and H. Hosono, Diffusion-Limited a-IGZO/Pt Schottky Junction Fabricated at 200 oC on a Flexible Substrate, IEEE Electron Dev. Lett, 32 (2011) 1695-1697. 15

ACCEPTED MANUSCRIPT

[24] H. L. Mosbacker, Y. M. Strzhemechny, B. D. White, and P. E. Smith, Role of nearsurface states in ohmic-Schottky conversion of Au contacts to ZnO, Appl. Phys. Lett. 87 (2005) 012102.

RI PT

[25] K. H. Kim, B. S. Kang, M.-J. Lee, S.-E. Ahn, C. B. Lee, G. Stefanovich, W. X. Xianyu, K.-K. Kim, J. S. Kim, I. K. Yoo, and Y. Park, Defect-induced degradation of rectification properties of aged Pt/n-InxZn1-xOy Schottky diodes, Appl. Phys.

SC

Lett. 92 (2008) 233507.

[26] Q. Xin, L. Yan, L. Du, J. Zhang, Y. Luo, Q. Wang, and A. Song, Influence of

M AN U

sputtering conditions on room-temperature fabricated InGaZnO-based Schottky diodes, Thin Solid Film 616 (2016) 569-572.

[27] H. G. Kim, H. R. Kim, and D. W. Kim, Effect of oxygen plasma treatment on the electrical properties in Ag/bulk ZnO Schottky diodes, Vacuum, 101 (2014) 92-97.

New York, 1981.

TE D

[28] S. M. Sze and K. K. Ng, Physics of Semiconductor Devices, John Wiley & Sons,

[29] H. Y. Kim, Capacitance-Voltage (C-V) Characteristics of Cu/n-type InP Schottky

EP

Diodes, Trans. Electr. Electron. Mater. 17 (2016) 293-296.

AC C

[30] T. Kamiya, K. Nomura, and H. Hosono, Origins of High Mobility and Low Operation Voltage of Amorphous Oxide TFTs: Electronic Structure, Electron Transport, Defects and Doping, J. Display Technol. 5 (2009) 273-288.

[31] Dieter K. Schroder, Semiconductor Material and Device Characterization 3rd ed., Wiley, New York, 2006. [32] K. A. Kim, M. J. Park, W. H. Lee, and S. M. Yoon, Characterization of negative bias-illumination-stress stability for transparent top-gate In-Ga-Zn-O thin-film transistors with variations in the incorporated oxygen content, J. Appl. Phys. 118 16

ACCEPTED MANUSCRIPT

(2015) 234504. [33] D. J. Yun, H. B. Kang, and S. M. Yoon, Process Optimization and Device Characterization of Nonvolatile Charge Trap Memory Transistors Using In-Ga-

RI PT

ZnO Thin Films as Both Charge Trap and Active Channel Layers, IEEE Trans.

AC C

EP

TE D

M AN U

SC

Electron Dev. 63 (2016) 3128-3134.

17

ACCEPTED MANUSCRIPT

Table Caption Table 1. Summary of estimated device parameters of the fabricated SBDs using IGZOA and IGZO-B thin films including the ideality factor (n), rectification ratio (Ion/off) at

RI PT

±1V, series resistance (RS), free carrier concentration (Ne), built-in potential (Vbi), Schottky barrier height from I-V (ΦB,IV), C-V (ΦB,CV) measurements and the difference

depletion region (Ndepl). IGZO-A SBD

IGZO-B SBD

0.88

0.94

M AN U

n

SC

between these two SBHs (∆Φ=ΦB,IV –ΦB,CV), and charge concentration within the

2.1 × 106

8.3 × 106

146

1970

6.47 × 1014

4.80 × 1013

0.36

0.46

0.76

0.80

ΦB,CV (eV)

0.61

0.79

∆Φ (eV)

0.15

0.01

Ndepl (cm-3)

1.84 × 1018

1.18 × 1017

Ion/off Rs (Ω) Ne (cm-3) Vbi (eV)

AC C

EP

TE D

ΦB,IV (eV)

18

ACCEPTED MANUSCRIPT

Fig. Captions

fabrication procedures and process conditions.

RI PT

Fig. 1. (a) Schematic diagram of the fabricated IGZO SBDs and (b) flowchart of the full

Fig. 2. (a) I-V characteristics of the SBDs using IGZO thin films with different oxygen

SC

gas mixing ratios and (b) I-V characteristics of the SBDs using IGZO thin films with

M AN U

two different compositions.

Fig. 3. (a) A2/C2-V plots of the SBDs using the IGZO-A and IGZO-B thin films. (b) Schematic energy band diagrams of the IGZO/Pt Schottky junctions with IGZO-A and

TE D

IGZO-B, respectively.

Fig. 4. Sets of I-V characteristics of the SBDs using (a) IGZO-A and (b) IGZO-B when

EP

the measurement temperatures were varied from room temperature to 120 oC.

AC C

Fig. 5. Variations in oxygen 1s spectra for (a) IGZO-A and (b) IGZO-B thin films deposited at an oxygen gas mixing ratio of 20% and annealed at 200 oC in an oxygen ambient for 1 h.

Fig. 6. Variations in in-situ measured electrical conductivity for the 80-nm-thick IGZOA and IGZO-B films deposited at an oxygen gas mixing ratio of 20% and annealed at 200 oC in an oxygen ambient based on an Arrhenius plot. Samples were prepared on insulating SiO2/Si substrates. 19

Fig. 1 ACCEPTED MANUSCRIPT

(a)

(b)

1. Bottom Electrode Deposition

RI PT

·Pt/Ti (200/10 nm) deposition on SiO2/Si

2. Active Layer Deposition

·In-Ga-Zn-O (80 nm) sputtering (PO2=20%)

SC

3. Mid-Annealing Process

·200oC in oxygen ambient for 1 h in RTA

M AN U

4. Active Layer Formation ·IGZO patterning by wet etching

5. Top Electrode Deposition

AC C

EP

TE D

·ITO (150 nm) sputtering

Fig. 2

-1

10 (b) -2 10 -3 10 -4 10 -5 10 -6 10 -7 10 -8 10 -9 10 -10 10 -11 10

M Current [A] AN US C

IGZO A PO2=1% IGZO A PO2=5%

IGZO A PO2=10%

-1

0

TE D

IGZO A PO2=20%

EP

Voltage [V]

AC C

Current [A]

-1

10 (a) -2 10 -3 10 -4 10 -5 10 -6 10 -7 10 -8 10 -9 10 -10 10 -11 10

RI PT

ACCEPTED MANUSCRIPT

1

-1

IGZO A PO2=20% IGZO B PO2=20%

0

Voltage [V]

1

Fig. 3 ACCEPTED MANUSCRIPT

(a)

20

2

RI PT

10

f = 1MHz IGZO A IGZO B

5 0 -1

M AN U

0

SC

15

2

12

4

2

A /C [x10 cm /F ]

25

1

Voltage [V]

EVAC

χs (4.79eV)

TE D

(b)

ΦB,CV (0.61eV)

EP

Φm (5.4eV)

AC C

Pt

EVAC

Φm (5.4eV)

qVbi (0.36eV)

IGZO-A

χs (4.61eV) ΦB,CV (0.79eV)

Pt

qVbi (0.46eV)

IGZO-B

Fig. 4

SC

RI PT

ACCEPTED MANUSCRIPT

-1

10 -2 10 -3 10 -4 10 -5 10 -6 10 -7 10 -8 10 -9 10 -10 10 -11 10

RT 40 60 80 100 120

M Current [A] AN U TE D IGZO A

0

EP

Voltage [V]

AC C

Current [A]

-1

10 RT -2 40 10 60 -3 10 80 -4 100 10 120 -5 10 -6 10 -7 10 -8 10 -9 10 -10 10 (a) -11 10 -1

1

(b) -1

IGZO B 0

Voltage [V]

1

Fig. 5 ACCEPTED MANUSCRIPT

Sum Back ground O 1s A O 1s B O 1s C

536

534

M AN U

SC

Intensity

RI PT

(a)

532

530

528

526

524

Binding energy (eV) Sum Back ground O 1s A O 1s B O 1s C

TE D

AC C

EP

Intensity

(b)

536

534

532

530

528

526

Binding energy (eV)

524

Fig. 6

RI PT

16

IGZO A IGZO B

12

SC

8 4

Ea=2.30 eV (IGZO A)

0

-8 -12 -16 2.0

M AN U

Ea=2.58 eV (IGZO B)

-4

TE D

Electrical Conductivity (lnc , Scm-1)

ACCEPTED MANUSCRIPT

2.5

AC C

EP

1000/T [1/K]

3.0

ACCEPTED MANUSCRIPT We fabricated Schottky barrier diodes using In-Ga-Zn-O oxide semiconductors.

·

Effects of oxygen partial pressures (PO2) and film compositions were investigated.

·

PO2 conditions during the sputtering has marked impacts on device characteristics.

·

Increase in Ga composition is effective for enhancing the diode performance.

·

Less amount of oxygen vacancies within the IGZO is preferred for the diodes.

AC C

EP

TE D

M AN U

SC

RI PT

·