The effect of static and rotating magnetic fields on low-temperature

Subscriber access provided by UNIV OF SCIENCES PHILADELPHIA

Functional Inorganic Materials and Devices

The effect of static and rotating magnetic fields on lowtemperature fabrication of InGaZnO thin-film transistors Jeong Woo Park, Young Jun Tak, Jae Won Na, Heesoo Lee, Won-Gi Kim, and Hyun Jae Kim ACS Appl. Mater. Interfaces, Just Accepted Manuscript • DOI: 10.1021/acsami.8b02433 • Publication Date (Web): 23 Apr 2018 Downloaded from http://pubs.acs.org on April 23, 2018

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 30 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Applied Materials & Interfaces

The effect of static and rotating magnetic fields on low-temperature fabrication of InGaZnO thin-film transistors Jeong Woo Park, Young Jun Tak, Jae Won Na, Heesoo Lee, Won-Gi Kim, and Hyun Jae Kim* School of Electrical and Electronic Engineering, Yonsei University, 50 Yonsei-ro, Seodaemun-gu, Seoul, 03722, Republic of Korea.

KEYWORDS oxygen vacancy, ferromagnetism, InGaZnO, flexible thin-film transistor, static and rotating magnetic field, magnetic moment

ACS Paragon Plus Environment

1

ACS Applied Materials & Interfaces 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 2 of 30

ABSTRACT

We suggest thermal treatment with static magnetic fields (SMFs) or rotating magnetic fields (RMFs) as a new technique for the activation of indium-gallium-zinc oxide thin film transistors (IGZO TFTs). Magnetic interactions between metal atoms in IGZO films and oxygen atoms in air by SMFs or RMFs can be expected to enhance metal-oxide (M-O) bonds, even at low temperature (150oC), through attraction of metal and oxygen atoms having their magnetic moments aligned in the same direction. Compared with IGZO TFTs with only thermal treatment at 300°C, IGZO TFTs under an RMF (1150 rpm) at 150°C show superior or comparable characteristics: field effect mobility of 12.68 cm2 V-1 s-1, subthreshold swing of 0.37 V dec-1, and on/off ratio of 1.86 × 108. Although IGZO TFTs under an SMF (0 rpm) can be activated at 150°C, the electrical performance is further improved in IGZO TFTs under an RMF (1150 rpm). These improvements of IGZO TFTs under an RMF (1150 rpm) are induced by increases in the number of M-O bonds, due to enhancement of the magnetic interaction per unit time as the rpm value increases. We suggest that this new process of activating IGZO TFTs at low temperature widens the choice of substrates in flexible or transparent devices.


2

Page 3 of 30 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


1. Introduction Recently, amorphous oxide semiconductor thin film transistors (AOS TFTs) have been extensively investigated as next-generation semiconductor devices due to their superior performance compared with conventional a-Si TFTs, in turn due to their high mobility, low offcurrent and high optical transparency in the visible region.1-2 Even though amorphous indium-gallium-zinc oxide thin film transistors (IGZO TFTs), which are widely used in AOS TFTs, are usually fabricated by a sputtering process at room temperature, IGZO thin films deposited by sputtering require an additional post-annealing process over 300°C to achieve activation and satisfactory transfer characteristics. However, since most flexible substrates (except polyimide; PI) cannot sustain high temperatures over 200°C, the conventional activation process cannot be used for manufacturing flexible IGZO TFTs. Therefore, it is important to decrease the activation temperature when using various flexible substrates, which have a low glass transition temperature.3-5 In this study, we investigated activation of IGZO TFTs at low temperature using static magnetic fields (SMFs) and rotating magnetic fields (RMFs), to apply various flexible substrates to flexible devices. As shown in Figure 1 (a), IGZO films could be annealed on a magnetic stirring hot plate with thermal energy and magnetic field simultaneously. However, most investigations of magnetic oxide semiconductors have featured only magnetic property analysis, such as ferromagnetism behavior in oxide semiconductors, and did not lead to an application or device research according to the magnetic properties of oxide semiconductors.6 Especially, studies analyzing the magnetic properties of IGZO thin films, and the reaction between oxide semiconductors and oxygen via an external magnetic field, are rare. In this research, unlike previous studies of the magnetic properties of oxide semiconductors, we analyzed the activation of IGZO TFTs at low temperature for flexible


3


Page 4 of 30

TFTs, as well as the physical properties and mechanisms of IGZO thin films in the presence of external magnetic fields. We also observed changes in the characteristics of the IGZO TFTs according to various external magnetic field conditions, such as when the magnetic field is static or rotating (at a slow or fast speed).

2. Experimental Section 2.1 Device Fabrication Amorphous-IGZO films (40 nm-thick) were deposited on a heavily doped p+ Si wafer with a thermally grown SiO2 film (120 nm-thick), as shown in Figure 1 (a). Amorphous-IGZO films were fabricated with radio frequency (RF) sputtering for 5 min at 150 W. The working pressure and partial pressure of oxygen were set at 5 ×10-3, and 0 Torr, respectively. After IGZO films were deposited, an annealing process was performed on a magnetic stirring hot plate for activation. The magnetic stirring hot plate applied an SMF or RMF, as well heat, to the sample. Exactly, magnetic field and thermal annealing could be used simultaneously in the activation process of IGZO films. The intensity of the magnetic field was around 5 mT, as measured with a magnetometer. The temperature of the stirring hot plate ranged from 150 to 300°C. The number of rotations of the magnet around a fixed axis was controlled by the number of revolutions per minute (rpm), and ranged from 0 to 1150 rpm. The activation processes were performed at 150 to 300oC in air for 1 hour. After the activation process, Al electrodes were deposited by RF sputtering for source and drain electrodes. The thickness of the source and drain electrodes was set as 200 nm. The length and width of the channel were set at 150 and 1,000 μm, respectively, using a shadow mask.


4

Page 5 of 30 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


2.2 Analysis method of thin film characteristics Magnetization as a function of the magnetic field of IGZO films was measured with a superconducting quantum interference device (SQUID) magnetometer (MPMS3; Quantum Design). The magnetic field was varied from -10 to 10 kOe, and temperature was set to room temperature (300 K). The electrical characteristics of the IGZO TFTs were measured with a parameter analyzer (4156C; Hewlett-Packard). Amorphous-IGZO TFTs were tested with gate voltage sweep from 30 to 30 V, with 10.1 V of drain voltage. To analyze the stability of positive bias, a positive bias stress (PBS) test was conducted under VGS = 20 V and VDS = 10.1 V for 3,600 s. A negative bias stress (NBS) test was conducted under VGS = -20 V and VDS = 10.1 V for 3,600 s, and a negative bias illumination stress (NBIS) test was conducted under VGS = -20 V and VDS = 10.1 V for 3,600 s under 5,700 lux of white LED. O K-edge X-ray absorption spectroscopy (XAS) spectra were obtained in total electron yield mode with energy ranging from 520 to 555 eV. Spectroscopy ellipsometry (SE) was used to measure the optical band gap, with energy ranging from 1.24 to 5 eV (step: 0.02 eV) and an incident angle of 65o. X-ray photoelectron spectroscopy (XPS) spectra were obtained in constant analyzer energy mode (CAE) at 50 eV with an Al K alpha source.

3. Results and Discussion Figure 1 (b) shows the results of the magnetic field distribution simulation for the magnets used in this study. The simulation was performed with ViziMag software. The magnet size was set at 5 cm ×5 cm, the magnet strength was 3 m, and the permeability of the magnet was assumed to be 1.


5


Page 6 of 30

The simulation results showed that the magnetic field intensity applied to the sample ranged from 2.5 to 3 mT (Figure 1 and S1). However, as shown in Figure S1, the magnetic field intensity in the A-B section was not uniform especially within the sample size (2 cm × 2 cm). Figure 1 (c) shows a photograph of the magnetic field intensity of the magnet. The dark region indicates strong magnetic field intensity, while the bright region existed mainly at the boundary between the N pole and the S pole and corresponds to a weak magnetic field intensity. As the rotational speed of the magnet increased from 0 to 1150 rpm, the boundary between the N and S poles, which was clearly present at 0 rpm, gradually disappeared, and the magnetic field intensity became uniform. This suggests that the problem of non-uniformity of the magnetic field for the sample size described in Figure S1 can be resolved by rapid rotation of the magnet at around 1150 rpm. As shown in Figure 1 (d), room temperature magnetization (M) as a function of the magnetic field (H) of the IGZO films was measured to confirm room temperature ferromagnetism of the sputter-processed IGZO films. ‘No-treatment’ indicates that as-deposited IGZO films were subject to no thermal treatment or activation process. ‘Only-thermal’ pertains to post-thermal treatment of IGZO films at 300oC. Recently, it was reported that oxide semiconductor thin films exhibit ferromagnetism behavior due to point defects (oxygen vacancies), even when magnetic material is un-doped at room temperature.7-9 Especially, IGZO films have ferromagnetism property and IGZO’s ferromagnetism can be stronger by increasing amount of oxygen vacancies was reported by some researchers.10-11 Indeed, the no-treatment IGZO films in our study exhibited ferromagnetic properties, as shown in Figure 1 (d), since they have many oxygen vacancies. On the other hand, only-thermal IGZO films showed paramagnetism behavior, since oxygen vacancies inside IGZO films are filled with oxygen during the thermal treatment process. RMF IGZO films


6

Page 7 of 30 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


have medium amount of oxygen vacancies compared with no-treatment and only-thermal treated IGZO films. Therefore, RMF IGZO films exhibit weak ferromagnetism. These results mean that IGZO film’s ferromagnetism property can be dominantly affected by amount of oxygen vacancies. Consequently, the IGZO films used in this study exhibited ferromagnetism behavior under the notreatment condition, demonstrating that the films can have sufficient magnetic interaction with the external magnetic field.

Figure 1. (a) Fabrication process of IGZO films using magnetic field activation. (b) Simulation result of magnetic field distribution. (c) A photograph of the magnetic field intensity of magnet by the magnetic viewing film. (bright region: weak magnetic field intensity, dark region: strong


7


Page 8 of 30

magnetic field intensity) (d) Room-temperature magnetization (M) as a function of magnetic field (H) of the IGZO films. The magnetic fields were applied parallel to the plane of the films

Figure 2 (a) shows a schematic of the conventional activation process of IGZO films through thermal treatment. In general, as-deposited IGZO films fabricated by a sputtering process contain many defect sites, including oxygen vacancies due to the bombardment effect, contrary to IGZO films deposited at high RF power12-13. These defect sites induce high electrical conductivity of IGZO films. The electrical conductivity of the sputtered IGZO films as a function of O2 partial pressure during sputtering can be reduced at the logarithmic scale by increasing O2 partial pressure.14 Also, in this study, the carrier concentration of sputter-processed IGZO films without O2 partial pressure was high (~1022 cm-3) due to the large number of oxygen vacancies. To reduce the number of defect sites in the IGZO films, and to increase the chemical bonds between the metal and the oxide, thermal treatment at temperatures above 300°C is required, as shown in Figure 2 (b). Figure 2 (c) provides a schematic of IGZO TFT activation under an SMF, in which thermal treatment and the SMF are simultaneously applied during the activation process. The transfer characteristics of IGZO TFTs with SMF activation (0 rpm) shown in Figure 2 (d) reveal good characteristics of the transistor, even at a relatively low temperature (150°C); overall, the offcurrent is slightly reduced at all temperatures compared with that shown in Figure 2 (b). To identify the effect on activation of IGZO TFTs exerted by the change in intensity of the SMF, the intensity of the magnetic field was controlled during the experiment by changing the distance between the magnet and the samples, while maintaining the temperature of the samples (Figure S2). Regarding the transfer characteristics of the IGZO TFTs shown in Figure S2, the off-current


8

Page 9 of 30 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


and Subthreshold swing (S.S) were decreased at 150°C as the distance between the magnet and the samples was decreased. This indicated that the SMF could influence the activation of IGZO TFTs even at a relatively low temperature (150°C), and that the activation effect of the IGZO TFTs increases with increasing intensity of the SMF. Figure 2 (e) shows a schematic of IGZO TFT activation under an RMF. Unlike the SMF activation described above (Figure 2 (d)), the RMF was induced by simultaneously rotating the magnet around the z-axis and applying the thermal treatment. The transfer characteristics of IGZO TFTs under an RMF (1150 rpm; Figure 2 (f)) show that the off-current is dramatically reduced at 100°C and 150°C. Unlike with an SMF (0 rpm), the activation of IGZO TFTs with an RMF (1150 rpm) was superior, especially at low temperatures of 100 or 150°C.


9


Page 10 of 30

Figure 2. Schematic of activation process of IGZO films: (a) conventional only-thermal treatment, (c) thermal treatment with static magnetic field (SMF) (e) thermal treatment with rotating magnetic field (RMF). Transfer characteristics of the IGZO TFTs for (b) only-thermal (d) SMF (0 rpm) (f) ‘RMF (0 rpm)’ samples.


10

Page 11 of 30 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


As shown in Figure S3, sequential treatments were conducted to investigate each role of magnetic and thermal treatments in depth; ‘RMF after thermal’, ‘thermal after RMF’, and ‘simultaneous RMF and thermal’. Thermal annealing was performed at 150oC, rpm of RMF was set at 1150 rpm, and treatment time was 1 hr. Figure S3 (a), (b), and (c) exhibit transfer characteristics of ‘RMF after thermal’, ‘thermal after RMF’, and ‘simultaneous RMF and thermal’ treated IGZO TFTs, respectively. From Figure S3, we confirmed that IGZO TFTs were fully activated only when RMF and thermal treatment were applied at the same time. Therefore, these results imply that the magnetic field can assists insufficient thermal energy at low temperatures (150°C) by aligning metal atoms and oxygen via magnetic interaction. Figure S4 (a) and (b) show transfer characteristics of only-thermal and RMF (1150 rpm) IGZO TFTs, respectively, with variation in treatment time. As a result, even though both only-thermal and RMF (1150 rpm) IGZO TFTs show unsatisfactory transfer characteristics for 30 minutes, they seem to be fully activated after 60 minutes of activation. Further, performances of both onlythermal and RMF (1150 rpm) IGZO TFTs were not changed significantly after treatment time over 60 minutes. From these results, we could confirm that a sufficient time (60 min) is required to form metal-oxide bonds within IGZO films for suitable semiconducting property, but over this time (60 min), electrical characteristics seem to have saturated because most of the metal atoms within IGZO films have fully reacted with external oxygen atoms. Figure 3 (a) shows the transfer characteristics of IGZO TFTs under an RMF at 150°C and increasing magnet rotation speed. To verify the effect of an RMF on the activation of IGZO TFTs, the rotation speed of the magnet in the stirring hot plate was set to 300, 800, and 1150 rpm, and the treatment time was set to 1 hour. Even though IGZO TFTs were not activated sufficiently under an SMF (0 rpm), the off-current value decreased significantly as the rpm value increased. It was


11


Page 12 of 30

confirmed that the off-current of IGZO TFTs activated by an RMF (1150 rpm) was reduced by about three orders of magnitude compared with that of IGZO TFTs activated by an SMF (0 rpm). Consequently, the IGZO TFTs activated by an RMF (1150 rpm) at 150°C showed relatively high, or similar, electrical performance compared with those activated at 300°C without any magnetic field (see Figure 2 (b)); the field-effect mobility (μFE) also increased, from 8.66 to 12.68 cm2 V-1 s-1. Figure S5 shows statistical data of field-effect mobility, S.S, and on/off ratio are extracted from 10 TFTs of only-thermal (300oC) and RMF, respectively. As can be seen in the data, RMF (150oC, 1150 rpm) IGZO TFTs indeed have better electrical characteristics and higher uniformity than only-thermal (300oC) IGZO TFTs. Figure 3 (b) and (c) show transfer characteristics with different treatment time for IGZO TFTs treated with RMF at 300 rpm and at 1150 rpm, respectively. From Figure 3 (b), we could see that the IGZO TFT did not show sufficient performance as a transistor after 60 minutes of treatment, but they showed better performance as the treatment time increased. In contrast, from Figure 3 (c), we could see that the IGZO TFT showed sufficient performance as a transistor after 60 minutes of treatment. Therefore, we could see that both time and rpm are important factors for IGZO activation.

Figure 3. (a) Transfer characteristics of IGZO TFTs for RMF samples at 150oC with increasing speed of magnet. Transfer characteristics with different treatment time for (b) RMF (300 rpm) and (c) RMF (1150 rpm) IGZO TFTs


12

Page 13 of 30 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


In order to confirm the effect of existence of O2 in RMF (1150 rpm) activation process, IGZO films were activated and their transfer characteristics were compared in air (including O2) and N2 atmosphere (not including O2). As shown in Figure S6, we can figure out that the role of O2 in the atmosphere is important in the activation process through the fact that activation is rarely performed in the N2 atmosphere compared with the air atmosphere. Figure S7 compares the carrier concentration of IGZO films under various activation conditions. SMF (0 rpm), and RMF (1150 rpm) indicate thermal treatment of IGZO films at 150°C and with simultaneous application of a magnetic field (SMF and RMF, respectively). The carrier concentration can be used to determine whether the IGZO films have appropriate semiconductor properties; in general, IGZO films exhibit semiconductor-like properties, such as a high Hall mobility, at a carrier concentration range of about 1016 to 1020 cm-3.1 The carrier concentration of no-treatment IGZO films was 9.1 × 1022 cm-3 (metal-like properties), and that of only-thermal IGZO films at 300°C was about 1.2 × 1016 cm-3 (semiconductor-like properties). The high carrier concentration of no-treatment IGZO films was attributed to the high number of oxygen vacancies and excess free electrons. On the other hand, the low carrier concentration of IGZO films at 300°C was attributed to suppression of free carrier generation by stronger metal-oxide (M-O) bonds.15 Also, IGZO films annealed with an SMF (0 rpm) and RMF (1150 rpm) at 150°C showed relatively low carrier concentrations (about 4.8 × 1017 and 9.9 × 1016 cm-3, respectively), indicating suitability as semiconductors. The PBS, NBS, NBIS test results of the stability of IGZO TFTs annealed at 300°C, and annealed with an RMF (1150 rpm) at 150°C, are shown in Figure S8 (a) - (f). In PBS test results, the positive threshold voltage (Vth) shift values of the activated IGZO TFTs annealed with an RMF (1150 rpm) at 150°C, and under the only-thermal condition at 300°C, were 2.8, and 5.4 V, respectively. In


13


Page 14 of 30

NBS test results, the negative Vth shift values of the only-thermal and RMF (1150 rpm) IGZO TFTs were similar. This is because the amount of positive charge which can be trapped at interface between gate insulator and channel layer is negligible in the IGZO films owing to their intrinsic n-type property. On the other hand, from NBIS test, RMF (1150 rpm) IGZO TFTs showed superior stability than only-thermal IGZO TFTs (Negative Vth shift of 8.4 V (RMF (1150 rpm) and 17.6 V (only-thermal). It means that RMF (1150 rpm) activation process can effectively decrease uncoordinated oxygen species in IGZO films compared to only-thermal activation process. XPS analysis of the In 3d5/2 metal peaks was performed for no-treatment, only-thermal, SMF (0 rpm), and RMF (1150 rpm) IGZO films, as shown in Figure 4 (a). A shift toward higher binding energy of In 3d5/2 peaks was observed in the order of no-treatment, only-thermal, SMF (0 rpm), and RMF (1150 rpm) IGZO thin films. The oxidized states of In–O, Ga-O, and Zn-O compounds have higher binding energy than In, Ga, and Zn, respectively.16-17 Therefore, with higher binding energy, more oxidation and better activation of IGZO films occur. Notably, the oxidation of In was more pronounced in RMF (1150 rpm) IGZO films than in SMF (0 rpm) IGZO films. This is also consistent with the transfer characteristics according to rpm speed (as described in Figure 3). In addition, as shown in Figure S9, not only the In 3d5/2 peak, but also the Ga 2p3/2 and Zn 2p3/2 peaks, presented with an increase of Ga-O and Zn-O compared with both the no-treatment and SMF (0 rpm) IGZO films. The binding energy peak shift under the no-treatment and RMF (1150 rpm) conditions showed different values depending on In 3d5/2, Ga 2p3/2 and Zn 2p3/2, which represents the different amount of oxidation states (0.42, 0.28, and 0.13 eV, respectively). According to Figure 4 (a) and (b), the oxidation states increase in the order of In-O, Ga-O and Zn-O. These results are closely associated with the magnetic moments (magnetic dipole moments) of In (6.40), Ga (2.00), and Zn (0.87), so


14

Page 15 of 30 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


that In having a higher magnetic moment than those of Ga and Zn can be considered to be due to oxidation caused by the attraction of more oxygen atoms.18-20 Metal atoms with large magnetic moments form stronger attractions and bonds with oxygen when they have the same magnetic moment direction, as explained by ‘Atom models of Ferman’, which is associated with magnetic interaction in atoms. This is discussed in detail in the caption of Figure 7. To investigate magnetic interaction of metal and oxygen during the treatment, wet etching rate test was conducted for IGZO films as shown in Figure S10. Generally, films with a high density and a low void fraction have a relatively low amount of defect sites and more amount of M-O bond.21 Since these films exhibit a low wet etching rate, we can conjecture the amount of M-O bond in the films through the wet etching rate value. The great number of M-O bonds in IGZO films indicates that IGZO films are well activated by the magnetic interaction of metal and oxygen. We used etchant for BOE (H2O2:HF=1000:1) and prepared samples (IGZO(40nm)/p+-Si). As shown in Figure S10, the wet etching rate data indicate RMF (150oC, 1150 rpm) IGZO films have a high density and more amount of M-O bond than the only thermal (150oC) and no-treatment IGZO films. Through the analysis of RMF (150oC, 1150 rpm) IGZO films with a greater amount of M-O bond than only-thermal (150oC), we could support the activation effect by the only magnetic field and the magnetic interaction of metal and oxygen. Figure 4 (c) shows the O 1s peaks of XPS spectra for the IGZO films under the no-treatment, only-thermal and RMF (1150 rpm) conditions. XPS O 1s spectra can be used to index changes in oxygen composition. In addition, the relative metal-oxide (M-O) bond, oxygen vacancy (Vo), and hydroxyl group peak values (%) can be compared by deconvolution of the spectra based on a Gaussian distribution. The O 1s peaks were located at 530.1 ± 0.2, 530.8 ± 0.1, and 531.8 ± 0.1 eV. In the no-treatment IGZO films, the peak corresponding to the M-O bonding was lower and


15


Page 16 of 30

that corresponding to oxygen vacancy was higher (M-O: 46.3%, Vo: 43.3%). Indeed, the notreatment IGZO films contained a large number of oxygen vacancies in the as-deposited state, which means that many metal atoms were not yet bonded to oxygen. On the other hand, the onlythermal and RMF (1150 rpm) IGZO films had a large number of M-O bonds and a small Vo value compared with the no-treatment IGZO films; both the only-thermal and RMF (1150 rpm) IGZO films were considered to have been oxidized sufficiently for activation. In particular, the RMF (1150 rpm) IGZO films at 150°C showed a slight decrease in M-O bonding, and an increased value of Vo, compared with the only-thermal IGZO films (M-O decreased from 63.5 to 60.0%, and Vo increased from 28.1 to 32.2%). This suggests that IGZO films activated by an RMF, even at a low temperature of 150°C, were as effective as that at a high temperature of 300°C. The increase in MO bonding, which forms conduction pathways for charge carriers, was caused by the increase in probability per unit time of M-O bonding through application of a rapidly RMF. In addition, the M-O bonding increase and Vo decrease seen in the RMF (1150 rpm) IGZO films was also consistent with the increase in oxidation of the In-O, Ga-O, and Zn-O compounds, as shown in Figure 4 (a) and (b). Also, the order of Vo concentration (only-thermal, RMF (1150 rpm), and notreatment IGZO films) coincide with the order of ferromagnetic strength concentration. These results support that Vo plays an important role in the room temperature ferromagnetism (RTFM) observed in IGZO films (see Figure 1 (d)).


16

Page 17 of 30 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Figure 4. XPS spectrum analysis for the IGZO films. (a) In 3d5/2 metal peaks. (b) Ga 2p3/2, Zn 2p3/2 and metal peaks. (c) O 1s peaks.

Figure 5 shows the electronic structures of the IGZO films obtained by O K-edge XAS, including the conduction band features. The conduction band consists of unoccupied hybridized orbitals for In 5sp, Ga 4sp, Zn 4sp, and O 2p from 530 to 550 eV.22 The conduction band spectra of the IGZO films are increased in the order of the no-treatment, only-thermal and RMF (1150 rpm)conditions, and a distinct difference in conduction band spectra between RMF (1150 rpm) and no-treatment IGZO films is seen at both peak A and peak B. Peaks A and B of the spectra of the IGZO films can be mainly attributed to In 5sp state, and a composition of In 5sp, Ga 4sp, and Zn 4sp states, respectively.22 The RMF (1150 rpm) IGZO films have a larger conduction band than


17


Page 18 of 30

the IGZO films under the other two conditions (no-treatment and only-thermal), which is closely related to enhanced charge transport due to unoccupied states, in turn caused by oxygen vacancies in the RMF (1150 rpm) IGZO film (and where, unoccupied states in the conduction band facilitate the transfer of electrons).23 The superior charge transport properties of the RMF (1150 rpm) IGZO films, even in comparison to the SMF (0 rpm) condition, is illustrated in Figure S11.

Figure 5. XAS spectra of the IGZO films at the O K-edge. The denoted by ‘A’ and ‘B’ in the spectra of the IGZO films can be mainly attributed to In 5sp state and a composition of In 5sp / Ga 4sp / Zn 4sp states, respectively.

Figure 6 (a) shows the valence band offsets (EFV) between the Fermi level (EF) and the valence band maximum (EV), according to XPS spectra in the vicinity of the valence band energy. The imaginary part of the absorption coefficient derived from the SE spectra of IGZO films is shown in Figure 6 (b). Optical band gap energies of IGZO thin films can be derived from the SE spectra. The band alignments of the IGZO films can be determined according to the conjugation between the optical band gap and the EFV data. The band alignments are shown in Figure 6 (c) for three conditions: no-treatment, only-thermal, and RMF (1150 rpm) (the band alignment analysis for the


18

Page 19 of 30 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


SMF (0 rpm) condition is shown in Figure S12). The Fermi energy level is located above the conduction band, indicating that IGZO films have metallic properties due to the excessive carrier concentration caused by Vo. On the other hand, under the only-thermal and RMF (1150 rpm) conditions, the Fermi energy was located relatively far from the conduction band, resulting in proper semiconductor characteristics.

Figure 6. (a) Valence band offsets (EFV) between the Fermi level (EF) and the valence band (EV) maximum using XPS spectra. (b) The imaginary part of the absorption coefficient from SE spectra. (c) The band alignments of the IGZO films determined from conjugation between optical band gap and the valence band offset data.


19


Page 20 of 30

In general, as the conduction band offsets (ECF) decreases, the carrier concentration increases exponentially, and the Hall mobility is proportional to the carrier concentration (as explained by the percolation model of oxide semiconductors).24 In addition, the μFE is proportional to the Hall mobility in oxide semiconductors.25-26 Therefore, the ECF of RMF (1150 rpm), which is smaller than the ECF of only-thermal, indicates that RMF (1150 rpm) IGZO films have a higher carrier concentration and μFE (in the case of IGZO TFTs) versus films under the only-thermal condition, which is consistent with the electrical performance shown in Figure S4 and Table S1, respectively. Every electron has an intrinsic angular momentum, called spin, the magnitude of which is the same for all electrons; associated with this angular momentum is a magnetic moment. 27 The magnetic moment is related to the external magnetic field, and the change in the external magnetic field is expressed as torque, which can be described by the following equation: ⃗⃗ 𝜏⃗ = 𝜇⃗ × 𝐵

(1)

⃗⃗ are the torque, magnetic moment, and external magnetic field, respectively. Where 𝜏⃗, 𝜇⃗, and 𝐵 This equation shows the relationship between the magnetic moment and the external magnetic field. The existence of a magnetic moment in atoms means that they are subjected to a torque that makes them rotate within an external magnetic field. The degree of alignment with the direction of the external magnetic field depends on the magnetic moment value; as the magnetic moment value becomes larger, the magnetic moment of the atom tends to rotate and align well with the external magnetic field. Activation of the IGZO TFTs by an SMF (0 rpm) and RMF (300, 800, 1150 rpm), as described above, can be explained by the magnetic interaction between oxygen and the metal whose magnetic moments are aligned with the magnetic field. Figure 7 (a) shows a magnetic interaction


20

Page 21 of 30 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


between metal atoms and oxygen atoms based on ‘Atom models of Ferman’, which can be expected to enhance the bond between the metal and the oxygen.28 Oxygen atoms in air, which have a small magnetic moment (1.50 × 10-6) and exhibit a weak paramagnetism characteristic, show little change in torque according to the external magnetic field, so that magnetic moment of oxygen atom cannot be easily aligned in the direction of the external magnetic field.29 On the other hand, in the case of In, Ga, and Zn atoms in IGZO films having ⃗⃗, and 𝜇⃗ large moment values (6.40, 2.00, and 0.87, respectively), the 𝜏⃗ changes greatly with the 𝐵 of metal atoms so that it is easily aligned in the direction of the external magnetic field. Thus, as shown in Figure 7 (b), if the magnetic moment directions of an oxygen atom and a metal atom aligned by the external magnetic field are the same, the magnetic interaction between the two atoms occurs due to the ‘Atom models of Ferman’. If the magnetic moment directions of the oxygen atom and metal atom are in opposition, there will be less attraction between the two atoms, and the bond between metal and oxygen will weaken, as shown in Figure 7 (a). Figure 7 (c) and (d) show the activation mechanism of IGZO films by SMF (0 rpm). The magnetic moment directions of metal atoms in IGZO films are randomly arranged without a magnetic field. However, with an SMF applied, metal atoms align in the direction of the magnetic field, and only metal atoms having the same magnetic moment direction to that of oxygen atoms can combine with oxygen atoms to form M-O bonds. Then, the number of M-O bonds in IGZO films increases slightly and the number of oxygen vacancies in IGZO films decreases slightly; that is, IGZO films can be relatively well-activated by the SMF, even at relatively low temperatures. In order to activate IGZO films, thermal annealing should be applied with magnetic field at the same time. This is because magnetic field just induces an improvement of magnetic interaction between metal and oxygen by aligning their directions, but simultaneously, thermal annealing


21


Page 22 of 30

facilitates chemical reaction between metal and oxygen. As a result, a simultaneous magnetic field and thermal annealing must be applied to activate IGZO TFTs.

Figure 7. Schematic of magnetic interaction between oxygen and metal atoms; (a) less-attraction with different magnetic moment direction, (b) attraction with same magnetic moment direction. Illustration of SMF activation; (c) Magnetic moment directions of metal atoms are randomly arranged (w/o magnet). (d) Magnetic moment directions of metal atoms are aligned in the direction of the magnetic field (w/o magnet).

On the other hand, when an RMF (1150 rpm) is applied to IGZO films, the magnetic moments of the metal atoms are rotated at high speed along the magnetic field. In this case, the probability


22

Page 23 of 30 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


that the magnetic moment directions of the oxygen atoms, which are hardly affected by the magnetic field and the metal atoms will coincide with each other increases with an RMF (1150 rpm) compared with an SMF (0 rpm). Accordingly, the probability that a metal atom and an oxygen atom will combine also increases, and an M-O bond could be easily formed with an RMF (1150 rpm), as shown in Figure 8. In order to increase the number of bonds between M and O, the total number of times that M and O are aligned with each other must be increased, which is proportional to the total number of revolutions of the magnetic field. The total number of revolutions of the magnetic field can be expressed in rpm × total time. For these reasons, as the rpm value increases, the rotational speed of the magnetic field, and of the magnetic moment of the metal atoms, also increases. Therefore, the number of M-O bonds increases with the magnetic interaction per unit time, as the rpm value increases. If the time is the same, it can be concluded that the number of M-O bonds will increase, and activation of the IGZO films can proceed efficiently, with a high-rpm magnetic field. Since the total time is also an important factor in terms of forming M-O bond, and we already confirmed that activation was effective by increasing the treatment time at low rpm conditions (see Figure 3 (b) and (c)). As a result, activation by RMF depends not only on rpm but also on treatment time.


23


Page 24 of 30

Figure 8. Illustration of RMF activation; As the magnet continues to rotate, the probability that the magnetic moment directions of metal atoms and oxygen atoms become equal to each other increases. The number of M-O bonds increases, and the number of oxygen vacancies gradually decreases.


24

Page 25 of 30 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


4. Conclusion In this study, we have demonstrated the effect of SMFs and RMFs on the activation of IGZO TFTs. IGZO TFTs may be activated at low temperature (150°C) by applying thermal energy and an RMF simultaneously. Electrical characteristics can be improved by increasing the rpm of the magnet and the intensity of the magnetic field at 150°C. The positive bias stability of IGZO TFTs is also improved by applying an RMF (1150 rpm) at 150°C relative to that of IGZO TFTs only annealed at 300°C. During magnetic field activation, the magnetic field promotes the interaction and attraction of metal and oxygen atoms, thus facilitating M-O bonds in IGZO films. All of this was possible because sputter-processed IGZO films used in this study exhibited ferromagnetism behavior originated from many oxygen vacancies. Consequently, various types of flexible substrates can be applied to IGZO TFTs by applying an SMF or RMF during the activation of IGZO TFTs at low temperature.


25


Page 26 of 30

ASSOCIATED CONTENT Supporting Information. Magnetic field intensity for the position of the magnetic field distribution; schematic diagram and transfer characteristics of SMF IGZO TFTs; transfer characteristics of the IGZO TFTs for sequential treatments; transfer characteristics of the IGZO TFTs as a function of activation time; statistical data of field-effect mobility, S.S, and on/off ratio from IGZO TFTs; transfer characteristics of the IGZO TFTs with different ambient condition; the comparison of carrier concentration between IGZO films; PBS, NBS, NBIS test for stability of IGZO TFT; XPS spectrum analysis (Ga 2p3/2 and Zn 2p3/2 peaks) for the IGZO films; thickness of amorphous IGZO films varied by the etching time and etching rate; XAS spectra of the ‘SMF (0 rpm)’ a-IGZO films at the O K-edge; SE & XPS analysis for ‘SMF (0 rpm)’ a-IGZO films; the comparison of electrical performance between a-IGZO TFTs.

AUTHOR INFORMATION Corresponding Author Hyun Jae Kim: [email protected]

ACKNOWLEDGMENT This work was supported by the National Research Foundation of Korea (NRF) grant funded by the Korea government (MSIT) (No. 2017R1A2B3008719).


26

Page 27 of 30 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


REFERENCES (1) Nomura, K.; Ohta, H.; Takagi, A.; Kamiya, T.; Hirano, M.; Hosono, H. Room-temperature Fabrication of Transparent Flexible Thin-film Transistors Using Amorphous Oxide Semiconductors. Nature 2004, 432, 488-492. (2) Kim, H.; Kwon, S.; Choi, S.; Choi, K. C. Solution-processed Bottom-emitting Polymer Lightemitting Diodes on a Textile Substrate Towards a Wearable Display. J. Inf. Disp 2015, 16, 179184. (3) Kim, W.-G.; Tak, Y. J.; Ahn, B. D; Jung, T. S.; Chung, K.-B.; Kim, H. J. High-pressure Gas Activation for Amorphous Indium-gallium-zinc-oxide Thin-film Transistors at 100°C. Sci. Rep. 2016, 6, 23039. (4) Tak, Y. J.; Ahn, B.D.; Park, S. P.; Kim, S. J.; Song, A. R.; Chung, K.-B.; Kim, H. J. Activation of Sputter-processed Indium–gallium–zinc Oxide Films by Simultaneous Ultraviolet and Thermal Treatments. Sci. Rep. 2016, 6, 21869. (5) Lee, H.; Chang, K. S.; Tak, Y. J.; Jung, T. S.; Park, J. W.; Kim, W.-G.; Chung, J.; Jeong, C. B.; Kim, H. J. Electric Field-aided Selective Activation for Indium-Gallium-Zinc-Oxide Thin Film Transistors. Sci. Rep. 2016, 6, 35044. (6) Hoa Hong, N. Magnetic Oxide Semiconductors. In Handbook of Spintronics; Yongbing, X.; Awschalom, D. D.; Nitta, J., Eds.; Springer: Dordrecht, Netherlands, 2016; pp 563– 583 (7) Liu, S.-J.; Su, S.-H.; Juang, J.-Y. Room-temperature Ferromagnetism in Amorphous In–Ga– Zn–O Films Fabricated by Using Pulsed-laser Deposition. Appl. Phys. A 2014, 116, 1473-1476. (8) Chang, G.; Forrest, J.; Kurmaev, E.; Morozovska, A.; Glinchuk, M.; McLeod, J.; Moewes, A.; Surkova, T.; Hong, N. H. Oxygen-vacancy-induced Ferromagnetism in Undoped SnO2 Thin Films. Phys. Rev. B 2012, 85, 165319. (9) Wang, H.; Yan, Y.; Li, K.; Du, X.; Lan, Z.; Jin, H. Role of Intrinsic Defects in Ferromagnetism of SnO2: First‐principles Calculations. Phys. Status Solidi B 2010, 247, 444-448. (10) Svensson, B. G.; Pearton, S.; Jagadish, C. Oxide semiconductors. Academic Press: 2013; Vol. 88.


27


Page 28 of 30

(11) Hong, N. H.; Sakai, J.; Poirot, N.; Brizé, V. Room-temperature Ferromagnetism Observed in Undoped Semiconducting and Insulating Oxide Thin Films. Phys. Rev. B 2006, 73, 132404. (12) Kuo, S.-Y.; Liu, K.-C.; Lai, F.-I.; Yang, J.-F.; Chen, W.-C.; Hsieh, M.-Y.; Lin, H.-I.; Lin, W.T. Effects of RF Power on the Structural, Optical and Electrical Properties of Al-doped Zinc Oxide Films. Microelectron. Reliab. 2010, 50, 730-733. (13) Rim, Y. S.; Jeong, W.; Ahn, B. D.; Kim, H. J. Defect Reduction in Photon-accelerated Negative Bias Instability of InGaZnO Thin-film Transistors by High-pressure Water Vapor Annealing. Appl. Phys. Lett. 2013, 102, 143503. (14) Minami, T.; Nanto, H.; Takata, S. Highly Conductive and Transparent Aluminum Doped Zinc Oxide Thin Films Prepared by RF Magnetron Sputtering. Jpn. J. Appl. Phys. 1984, 23, L280- L282. (15) Nomura, K.; Takagi, A.; Kamiya, T.; Ohta, H.; Hirano, M.; Hosono, H. Amorphous Oxide Semiconductors for High-performance Flexible Thin-film Transistors. Jpn. J. Appl. Phys. 2006, 45, 4303-4308. (16) Hewitt, R.; Winograd, N. Oxidation of Polycrystalline Indium Studied by X‐ray Photoelectron Spectroscopy and Static Secondary Ion Mass Spectroscopy. J. Appl. Phys. 1980, 51, 2620-2624. (17) Wett, D.; Demund, A.; Schmidt, H.; Szargan, R. The Fe/ZnO (0001) Interface: Formation and Thermal Stability. Appl. Surf. Sci. 2008, 254, 2309-2318. (18) Millman, S.; Rabi, I.; Zacharias, J. On the Nuclear Moments of Indium. Phys. Rev. 1938, 53, 384-391. (19) Kusch, P.; Foley, H. Precision Measurement of the Ratio of the Atomic ‘g Values’ in the P3/2 and P1/2 States of Gallium. Phys. Rev. 1947, 72, 1256-1257. (20) Dharmatti, S.; Weaver Jr, H. The Magnetic Moment of Zn 67. Phys. Rev. 1952, 85, 927. (21) Smith, J. W.; Maeda, Y.; Iyer, R. S. Disilane-based Low Thermal Budget Silicon Dioxide Chemical Vapor Deposition Process in a Single-wafer Chamber. Electrochem. Solid State Lett. 2006, 9, G141-G143. (22) Ahn, B. D.; Park, J.-S.; Chung, K. Facile Fabrication of High-Performance InGaZnO Thin Film Transistor using Hydrogen Ion Irradiation at Room Temperature. Appl. Phys. Lett. 2014, 105, 163505.


28

Page 29 of 30 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


(23) Park, H.-w.; Park, J.-S.; Lee, J. H.; Chung, K.-B. Thermal Evolution of Band Edge States in ZnO Film as a Function of Annealing Ambient Atmosphere. Electrochem. Solid State Lett. 2012, 15, H133-H135. (24) Takagi, A.; Nomura, K.; Ohta, H.; Yanagi, H.; Kamiya, T.; Hirano, M.; Hosono, H. Carrier Transport and Electronic Structure in Amorphous Oxide Semiconductor, a-InGaZnO4. Thin solid films 2005, 486, 38-41. (25) Suzuki, T. I.; Ohtomo, A.; Tsukazaki, A.; Sato, F.; Nishii, J.; Ohno, H.; Kawasaki, M. Hall and Field‐Effect Mobilities of Electrons Accumulated at a Lattice‐Matched ZnO/ScAlMgO4 Heterointerface. Adv. Mater. 2004, 16, 1887-1890. (26) Kim, G. H.; Ahn, B. D.; Shin, H. S.; Jeong, W. H.; Kim, H. J.; Kim, H. J. Effect of Indium Composition Ratio on Solution-processed Nanocrystalline InGaZnO Thin Film Transistors. Appl. Phys. Lett. 2009, 94, 233501. (27) Uhlenbeck, G. E.; Goudsmit, S. Spinning Electrons and the Structure of Spectra Nature 1926, 117, 264-265. (28) Rodriguez, F. M. Atom and Stellar Models of Ferman: Planetary and Everything Theory. Int. j. res. eng. appl. sci. 2016, 4, 17-44. (29) Beltrán-López, V.; Blaisten, E.; Segovia, N.; Koo, E. L. Calculation of the Magnetic Moment of Atomic Oxygen. Phys. Rev. 1969, 177, 432-434.


29


Page 30 of 30

Table of contents


30

The effect of static and rotating magnetic fields on low-temperature

Recommend Documents