Electrical Performance and Reliability Improvement of ... - MDPI

7 downloads 0 Views 1MB Size Report
May 17, 2018 - plasma treatment, both the electrical performance and reliability of the a-IGZO ... the a-IGZO TFT with a high-k gate dielectric layer is one of the ...
materials Article

Electrical Performance and Reliability Improvement of Amorphous-Indium-Gallium-Zinc-Oxide Thin-Film Transistors with HfO2 Gate Dielectrics by CF4 Plasma Treatment Ching-Lin Fan 1,2, *, Fan-Ping Tseng 2 and Chiao-Yuan Tseng 1 1 2

*

Department of Electronic Engineering, National Taiwan University of Science and Technology, 43 Section 4, Keelung Road, Taipei 106, Taiwan; [email protected] Graduate Institute of Electro-Optical Engineering, National Taiwan University of Science and Technology, 43 Section 4, Keelung Road, Taipei 106, Taiwan; [email protected] Correspondence: [email protected]; Tel.: +886-2-2737-6374  

Received: 21 April 2018; Accepted: 15 May 2018; Published: 17 May 2018

Abstract: In this work, amorphous indium-gallium-zinc oxide thin-film transistors (a-IGZO TFTs) with a HfO2 gate insulator and CF4 plasma treatment was demonstrated for the first time. Through the plasma treatment, both the electrical performance and reliability of the a-IGZO TFT with HfO2 gate dielectric were improved. The carrier mobility significantly increased by 80.8%, from 30.2 cm2 /V·s (without treatment) to 54.6 cm2 /V·s (with CF4 plasma treatment), which is due to the incorporated fluorine not only providing an extra electron to the IGZO, but also passivating the interface trap density. In addition, the reliability of the a-IGZO TFT with HfO2 gate dielectric has also been improved by the CF4 plasma treatment. By applying the CF4 plasma treatment to the a-IGZO TFT, the hysteresis effect of the device has been improved and the device’s immunity against moisture from the ambient atmosphere has been enhanced. It is believed that the CF4 plasma treatment not only significantly improves the electrical performance of a-IGZO TFT with HfO2 gate dielectric, but also enhances the device’s reliability. Keywords: a-IGZO TFT; HfO2 gate dielectric; plasma treatment; fluorine; reliability

1. Introduction Recently, amorphous indium-gallium-zinc oxide thin-film transistors (a-IGZO TFT) have attracted considerable attention in flat-panel displays (FPDs), because of its advantages, including high carrier mobility (>10 cm2 /V·s), good uniformity in large-area deposition and low temperature fabrication [1–5]. Compared with low-temperature poly-crystalline silicon (LTPS) TFTs, the a-IGZO TFTs are more suitable to utilize as driving-TFTs in large-area active-matrix organic light-emitting diode (AM-OLED) displays because of these aforementioned advantages. The a-IGZO TFT is a good candidate to replace silicon-based TFTs both in the active-matrix liquid crystal displays (AM-LCDs) and AM-OLED displays. In order to increase the carrier mobility of the a-IGZO TFT, many previous researchers used a high dielectric constant (high-k) material as the gate insulator to improve the electrical performance of a-IGZO TFT [6–11]. Among these gate insulator materials, HfO2 is one of the most promising high-k materials due to its advantages of a high dielectric constant (k > 20), a sufficient energy bandgap offset and a suitable interface for the IGZO semiconductor [12–17]. However, the electrical stability of the a-IGZO TFT with a high-k gate dielectric layer is one of the critical issues, especially under the positive-gate bias stress (PGBS) [10]. The threshold voltage (Vth ) is abnormally shifted in the negative Materials 2018, 11, 824; doi:10.3390/ma11050824

www.mdpi.com/journal/materials

Materials 2018, 11, 824

2 of 8

direction, which is opposite to the a-IGZO TFT with a SiO2 gate insulator [18,19]. This is due to the electron density generated by the density of states (DOS) near the dielectric/channel layer interface or from absorbed moisture from the air. In addition, the higher dielectric constant of the high-k gate insulator would induce a higher electric-field through the active layer, which makes the a-IGZO TFTs more sensitive to the ambient molecules [20]. Therefore, the reliability of the a-IGZO TFT with high-k gate dielectric under PGBS and under ambient atmosphere should be enhanced. It is reported that fluorine incorporated into the a-IGZO TFT can effectively improve TFT reliability [20–26], because F has several advantages: (i) it has the highest electron affinity among chemical elements, which makes F simple to bond with misoriented metal atoms in the IGZO active layer and reduce the DOS, (ii) compared to hydrogen, F can provide stronger bonding with the metal ion in the active layer to improve the a-IGZO TFT stability [27], and (iii) extra free electrons can be generated by the fluorine ions by replacing the oxygen sites, due to the difference in electrovalence between the oxygen ion (O2− ) and the fluorine ion (F− ). The higher electron density induced in the active layer will improve the carrier mobility of the a-IGZO TFT [28]. Recently, some researchers reported using carbon fluoride mixed with oxygen gas (CF4 /O2 and CHF3 /O2 ) to enhance the stability of a-IGZO TFT [20–22]. Although the stability under the PGBS of the a-IGZO TFT was enhanced, the electrical performance of the TFT was not significantly improved. It is well-known that the oxygen molecule acts as a carrier suppressor in the IGZO thin film. The merit of F ions that can generate extra electrons in the IGZO film might be diluted by the incorporation of oxygen. In addition, there is no research that has focused on the effects of plasma treatment in the a-IGZO TFT with the HfO2 material as the gate insulator. The a-IGZO TFT with the HfO2 gate insulator and treatment by CF4 plasma is studied here for the first time. In this work, both the electrical performance and reliability of the a-IGZO TFTs were improved by the CF4 plasma treatment, and with HfO2 as the gate insulator. After the CF4 plasma treatment of the active layer, the carrier mobility significantly improved from 30.2 cm2 /V·s to 54.6 cm2 /V·s. The improvement is attributed to the extra electrons generated in the IGZO film and the DOS near the gate insulator/channel interface passivated by the incorporation of fluorine. The effect of the reduced DOS near the interface also can be distinctly observed in the hysteresis measurement. In addition, the moisture absorption effect was also investigated in this work. It was found that the CF4 plasma treated a-IGZO TFT has higher immunity against ambient moisture. 2. Device Fabrication Figure 1 shows the schematic cross-sectional diagrams of the bottom-gate, top-contact a-IGZO TFTs. First, ITO film was patterned using photolithography and wet etching to form the gate electrode. The 180-nm thick HfO2 was then deposited as the gate insulator layer by RF sputtering in the mixed gas O2 /Ar = 33% and annealed at 250 ◦ C for 60 min. After patterning the contact hole, a 30 nm a-IGZO thin-film was deposited by RF sputtering in an Ar environment at room temperature. The active layer was then patterned by wet etching. A 160 nm Ti layer was deposited by thermal evaporation and using lift-off process to form source and drain (S/D) electrodes. Finally, a CF4 plasma treatment was carried out at the RF power of 15 W for 20 s, followed by a post-anneal at 250 ◦ C for 60 min. Note that the low power and short treatment time of the plasma treatment ensured that the plasma treatment mainly acted on the IGZO active layer without serious damage. The electrical characteristics and the bias stress voltage were examined using an Agilent 4145B semiconductor parameter analyzer in the dark. The transfer curves were measured at a source-to-drain voltage (VDS ) of 5 V and the Vth was extracted from linear extrapolations of the square root plot of the drain current (IDS ). The channel width (W) and length (L) of the a-IGZO TFT were 50 µm and 5 µm, respectively. The a-IGZO TFT without the passivation layer was used in the positive-gate bias stress (PGBS) to examine the ambient effect. The PGBS was carried out at the gate electrode (VGS = 6 V) with the source and drain electrodes grounded.

Materials 2018, 11, x FOR PEER REVIEW    Materials 2018, 11, x FOR PEER REVIEW    Materials 2018, 11, 824

3 of 8  3 of 8  3 of 8

   

Figure 1. Schematic cross‐sectional diagrams of a‐IGZO TFT with CF4 plasma treatment.  Figure 1. Schematic cross‐sectional diagrams of a‐IGZO TFT with CF Figure 1. Schematic cross-sectional diagrams of a-IGZO TFT with CF44 plasma treatment.  plasma treatment.

3. Results and Discussion  3. Results and Discussion  3. Results and Discussion Figure  2a  shows  the  secondary  ion  mass  spectrometry  (SIMS)  depth  profile  in  the  stack  of  Figure  2a thin  shows  the  secondary secondary  mass  spectrometry  (SIMS)  depth  profile  in  stack  of of  Figure2/Si  2a shows the ion mass (SIMS)the  depth profile in the  the stack IGZO/HfO films.  Compared ion  with  the spectrometry untreated  sample,  higher  F  concentration  is  IGZO/HfO 2/Si  thin  films.  Compared  with  the  untreated  sample,  the  higher  F  concentration  is  IGZO/HfO films. Compared with thetreatment,  untreatedwhich  sample, the higher concentration is observed  in 2 /Si the  thin sample  with  the  CF4  plasma  indicates  that  F the  F  atoms  were  observed  in  the  sample  with  the  CF 4  plasma  treatment,  which  indicates  that  the  F  atoms  were  observed in the sample with the CF4 plasma treatment, which indicates that the F atoms were successfully introduced into the IGZO film by the CF 4 plasma treatment. In addition, the F atoms  successfully introduced into the IGZO film by the CF successfully introduced into the IGZO film by the CF44 plasma treatment. In addition, the F atoms  plasma treatment. In addition, the F atoms diffused towards the bulk of the IGZO film, and piled up at the IGZO/HfO 2 interface. Figure 2b shows  diffused towards the bulk of the IGZO film, and piled up at the IGZO/HfO 2 interface. Figure 2b shows  diffused towards the bulk of the IGZO film, and piled up at the IGZO/HfO Figure 2b the  fourier‐transform  infrared  spectroscopy  (FTIR)  measurement  of  the  IGZO  film.  The  wave  2 interface. the  fourier‐transform  infrared  spectroscopy  (FTIR)  measurement  of  the  IGZO  film.  The  wave  −1 −1 shows the fourier-transform infrared spectroscopy3 bond and CF (FTIR) measurement of the IGZO film. The wave numbers of 980 cm  and 1240 cm  refer to the CF 2 bond, respectively [29]. Evidently,  −1 −1 1 numbers of 980 cm 2 bond, respectively [29]. Evidently,  numbers of 980 cm−1 and 1240 cm and 1240 cm− refer to the CF refer to the CF3 bond and CF bond and CF bond, respectively [29]. Evidently, the IGZO film with 45 W treatment power has the higher CF 2  peak, which indicates the more serious  3 2 the IGZO film with 45 W treatment power has the higher CF 2 peak, which indicates the more serious  the IGZO film with 45 W treatment power has the higher CF peak, which indicates the more serious etching effect during the plasma treatment. CF4 is well‐known as one of the reactive etching gas used  2 etching effect during the plasma treatment. CF 4 is well‐known as one of the reactive etching gas used  etching effect during the plasma treatment. CF4 is well-known as one of the reactive etching gas used in dry etching process. As the RF power increased, CF 4 begins to dissociate into smaller components,  in dry etching process. As the RF power increased, CF in dry etching process. As the RF power increased, CF44 begins to dissociate into smaller components,  begins to dissociate into smaller components, such as CF 2, CF 3 and F radicals, which etch the underlying IGZO thin film. A small treatment power  such as CF such as CF22, CF , CF33 and F radicals, which etch the underlying IGZO thin film. A small treatment power  and F radicals, which etch the underlying IGZO thin film. A small treatment power can reduce plasma damage in the IGZO active layer.  can reduce plasma damage in the IGZO active layer.  can reduce plasma damage in the IGZO active layer.

20 20 15 15

a-InGaZnO ( 30 nm ) a-InGaZnO ( 30 nm ) Fluorine distribution Fluorine W/Odistribution treatment W/O 15 W treatment 20 s 15 W 20 s

10 10 5 5 0 0

5 10 15 20 25 30 35 40 45 50 5 10 15 20 25 (nm) 30 35 40 45 50 Depth   Depth (nm)   (a)  (a) 

Plasma treatment parameter Plasma treatment 15 W 20 s parameter 15 W W 20 20 ss 45 45 W 20 s

HfO2 HfO2

Absorbance (%) Absorbance (%)

4

4 Intensity (10atom atom counts Intensity (10 counts ) )

25 25

CF2 CF2

CF3 CF3

800 800

1000 1000

1200 1200

1400 1400 -1

Wavenumber (cm -1) Wavenumber (cm )

1600 1600

   

(b)  (b) 

Figure 2. (a) SIMS depth profiles; (b) FTIR measurement for the untreated sample and the sample  Figure 2. (a) SIMS depth profiles; (b) FTIR measurement for the untreated sample and the sample with Figure 2. (a) SIMS depth profiles; (b) FTIR measurement for the untreated sample and the sample  4 plasma treatment after annealing at 250 °C for 60 min.  with the CF the CF4 plasma treatment after annealing at 250 ◦ C for 60 min. 4 plasma treatment after annealing at 250 °C for 60 min.  with the CF

Figure  3a  shows  the  transfer  curves  of  the  a‐IGZO  TFT  with  and  without  the  CF4  plasma  Figure 3a curves Figure  3a  shows shows  the the  transfer transfer  curves  of of  the the  a-IGZO a‐IGZO  TFT TFT  with with  and and  without without  the the  CF CF44  plasma plasma  treatment. The linear mobility (μ linear), maximum transconductance (gmmax), threshold voltage (Vth),  treatment. The linear mobility (µlinear ), maximum transconductance (gmmax ), threshold voltage (Vth),  ), treatment. The linear mobility (μ linear), maximum transconductance (gm max), threshold voltage (V subthreshold swing (S.S.) and the on/off current ratio (Ion/Ioff) are summarized in Table 1. The μlinear is  subthreshold swing (S.S.) and the on/off current ratio (I /I ) are summarized in Table 1. The µ is onon/Ioff subthreshold swing (S.S.) and the on/off current ratio (I off) are summarized in Table 1. The μ linear is  linear calculated from the gm max measured at the VDS = 0.1 V using the linear‐region drain current function,  calculated from the gmmax measured at the VDS = 0.1 V using the linear-region drain current function, calculated from the gm max measured at the V DS = 0.1 V using the linear‐region drain current function,  which  is  shown  in  the  inset  of  Figure  3a.  It  is  obvious  that  the  a‐IGZO  TFT  with  the  CF4  plasma  which of which  is is  shown shown  in in the the inset inset  of Figure Figure 3a. 3a.  It It  is is  obvious obvious  that that  the the a-IGZO a‐IGZO TFT TFT with with the the CF CF44  plasma plasma  treatment has a higher gm max compared to the TFT without treatment, and that the linear mobility of  treatment has a higher gm compared to the TFT without treatment, and that the linear mobility max treatment has a higher gm max  compared to the TFT without treatment, and that the linear mobility of  2/V·s (with CF4  the device is significantly improved from 30.2 cm2/V·s (without treatment) to 54.6 cm 2 /V·s (without treatment) to 54.6 2cm 2 /V·s (with 2/V·s (without treatment) to 54.6 cm of the device is significantly improved from 30.2 cm the device is significantly improved from 30.2 cm /V·s (with CF 4  plasma treatment). Moreover, the Vth and S.S. were slightly decreased from 1.50 V to 1.05 V, and 0.17  plasma treatment). Moreover, the Vth and S.S. were slightly decreased from 1.50 V to 1.05 V, and 0.17 

Materials2018, 2018,11, 11,824 x FOR PEER REVIEW Materials

of88 44of

V/decade to 0.14 V/decade, respectively. The carrier mobility of TFT mainly depended on the carrier CF the Vth and were slightly decreased 1.50concentration V to 1.05 V, 4 plasma scattering ortreatment). trapping inMoreover, the channel layer. It is S.S. reasonable to suggest that thefrom carrier and 0.17 V/decade to 0.14 V/decade, respectively. The carrier mobility of TFT mainly and trap density will dominate the carrier mobility. It is important to consider that the roledepended of F atom on the IGZO carrierfilm scattering trapping in the layer. It is reasonable tobonded suggestoxygen that theatoms carrier in the can notor only passivate thechannel defects by replacing the weakly or concentration and trap density will dominate the carrier mobility. It is important to consider that the by directly filling the oxygen vacancy, but will also donate extra electrons due to the difference in role of F atom in the IGZO can not the defects the weakly bonded electrovalence between thefilm oxygen ion only (O2−)passivate and the fluorine ionby (F−replacing ) [28]. A lower interface trap oxygen the oxygen will also donate extra electrons to density atoms resultsor inby lessdirectly carrier filling scattering, and thevacancy, gate biasbut voltage can effectively induce moredue carrier 2− ) and the fluorine ion (F− ) [28]. A lower the difference in electrovalence between the oxygen ion (O concentration to transport in the IGZO channel due to the decreased trap density and the donated interface trap density carrier scattering, and the gate biasisvoltage canby effectively extra electrons. As a results result, in theless carrier mobility of the a-IGZO TFT improved the CF4 induce plasma more carrier concentration to transport in the IGZO channel due to the decreased trap density and treatment. Figure 3b shows the output curves of the two a-IGZO TFTs measured at VGS = 4 V, 6the V, donated electrons. As a result, carrier mobility thethe a-IGZO TFT is improved CF4 and 8 V. extra Both the TFTs exhibit n-typethe enhancement mode,ofand IDS saturated at the highby VDSthe region. plasma treatment. Figure 3b shows thethe output curves of the two TFTsTFT measured VGS = 4 V, Compared with the untreated device, higher IDS observed in a-IGZO the a-IGZO with theatCF 4 plasma 6treatment V, and 8 V. Both the TFTs exhibit n-type enhancement mode, and the I saturated at the high DS resulted from the increased carrier mobility. The on currentsDSmeasured at VDS = 6 VVand region. Compared with treatment the untreated higher TFT IDS observed in the with the VGS = 8 V of the plasma TFT device, and thethe untreated are 1180 µA anda-IGZO 623 µA,TFT respectively. CF resulted from theof increased carrier The on currents measured at VCF DS4 As4 aplasma result, treatment the electrical performance the a-IGZO TFTmobility. can be significantly improved by the =plasma 6 V and V = 8 V of the plasma treatment TFT and the untreated TFT are 1180 µA and 623 µA, GS treatment. respectively. As a result, the electrical performance of the a-IGZO TFT can be significantly improved by theTable CF4 plasma 1. Effectstreatment. of the CF4 plasma treatment on electrical performance parameters of the fabricated aIGZO TFTs with HfO2 gate dielectric. Table 1. Effects of the CF4 plasma treatment on electrical performance parameters of the fabricated Electrical Parameters Without Treatment With CF4 Plasma Treatment a-IGZO TFTs with HfO 2 gate dielectric.

µlinear (cm2/V∙s) gmParameters max (A/V) Electrical Vth2(V) µlinear (cm /V·s) S.S. (V/decade) gmmax (A/V) VthI(V) on/Ioff

30.2 3.02 10−6 Without×Treatment 1.50 30.2 0.17 3.02 × 10−6 3.5 ×1.50 106

54.6 5.46 × 10−6 Treatment With CF4 Plasma 1.0554.6 0.14× 10−6 5.46 7.44 × 1.05 107

S.S. (V/decade) Ion /Ioff

0.17 3.5 × 106

0.14 7.44 × 107

12

W/O treatment VGS

10

4V 6V 8V CF4 plasma treatment

-4

IDS (10 A)

8 6

VGS 4V 6V 8V

4 2 0

0

1

2

3

4

5

6

VDS (V) (a)

(b)

Figure 3. (a) Transfer curves (IDS -VGS ); (b) Output curves (IDS -VDS ) of the a-IGZO TFT without Figure 3. (a) Transfer curves (IDS-VGS); (b) Output curves (IDS-VDS) of the a-IGZO TFT without treatment and with a 15 W, 20 s CF4 plasma treatment. Inset of Figure 3a shows the transconductance treatment and with a 15 W, 20 s CF4 plasma treatment. Inset of Figure 3a shows the transconductance measured at VDS = 0.1 V. measured at VDS = 0.1 V.

Figure Figure44shows showsthe thehysteresis hysteresismeasurement measurementofofthe thea-IGZO a-IGZOTFT TFTwith withand andwithout withoutthe theCF CF44plasma plasma treatment. It can be observed that the hysteresis voltage (∆V ) abnormally shifted towards the negative H (∆VH) abnormally shifted towards treatment. It can be observed that the hysteresis voltage the direction and the a-IGZO TFT with the plasma treatment has a smaller ∆V . The ∆V of the a-IGZO H H negative direction and the a-IGZO TFT with the plasma treatment has a smaller ∆VH. The ∆VH of the TFT with TFT and without CF4 plasma treatment are −1.8 V and −0.3 V, respectively. The negative Vth a-IGZO with andthe without the CF 4 plasma treatment are −1.8 V and −0.3 V, respectively. The shift is due the carrier creation in the IGZO film, mainly from the enhanced control and negative Vthtoshift is due to the carrier creation in thewhich IGZOisfilm, which is mainly from the enhanced

control and from using the high-k HfO2 as the gate insulator. When the gate bias is applied to the

Materials 2018, 11, x FOR PEER REVIEW    Materials 2018, 11, 824

5 of 8  5 of 8

control and from using the high‐k HfO2 as the gate insulator. When the gate bias is applied to the  from using the high-k HfO2 as the gate insulator. When the gate bias is applied to 2the TFT, the HfO2 TFT, the HfO 2 gate insulator will provide a strong electric field near the IGZO/HfO  interface to bend  gate insulator will provide a strong electric field near the IGZO/HfO interface to bend the Fermi 2 the Fermi level in the deep of the band gap. The energy level of the neutral oxygen vacancies (V O),  level in the deep of the band gap. The energy level of the neutral oxygen vacancies (VO ), which is which is higher than the Fermi level, will be ionized (V O++) and two electrons will be contributed to  ++ higher thanfilm  the because  Fermi level, be ionized (Venergy  two in  electrons willband  be contributed to reach  the IGZO O ) and the  IGZO  the will highest  electron  level  the  IGZO  gap  cannot  the  film because the highest electron energy level in the IGZO band gap cannot reach the energy level of energy level of VO [10,18]. Therefore, the extra free electrons generated during the forward sweep of  V generated during the forward sweep of the gate bias will O [10,18]. Therefore, the extra free electrons the gate bias will cause the negative V th shift in the reverse sweep of the gate bias.  cause the negative Vth shift in the reverse sweep of the gate bias. -2

-2

10 -3 10 -4 10 -5 10 -6 10 -7 10 -8 10 -9 10 -10 10 -11 10 -12 10

CF4 Plasma Treatment Forward Reverse

IDS (A)

IDS (A)

10 -3 W/O treatment 10 Forward -4 Reverse 10 -5 10 -6 10 -7 10 -8 10 -9 10 -10 10 -11 10 -12 10

VH = 1.8 V

-6 -5 -4 -3 -2 -1 0 1 2 3 4 5 6 VGS (V) (a) 

 

VH = 0.3 V

-6 -5 -4 -3 -2 -1 0 1 2 3 4 5 6 VGS (V)

 

(b) 

Figure  4.  Hysteresis  thin‐film  Figure 4. Hysteresis measurement  measurement of  of amorphous  amorphous indium‐gallium‐zinc  indium-gallium-zinc oxide  oxide (a‐IGZO)  (a-IGZO) thin-film 4 plasma treatment.  transistor (TFT) with and without the CF transistor (TFT) with and without the CF plasma treatment. 4

Figure  5a,b  shows  the  transfer  curves  of  the  a‐IGZO  TFT  with  and  without  the  CF4  plasma  Figure 5a,b shows the transfer curves of the a-IGZO TFT with and without the CF plasma treatment  under  the  PGBS  (VGS  =  6  V).  The  TFTs  were  fabricated  without  the  passivation 4 layer  to  treatment under the PGBS (VGS = 6 V). The TFTs were fabricated without the passivation layer to examine the ambient effect. For the a‐IGZO TFT without treatment, the device suffers from a more  examine the ambient effect. For the a-IGZO TFT without treatment, the device suffers from a more serious negative Vth shift and a S.S. degradation, as shown in Figure 6a,b, which are mainly due to  serious negative Vth shift and a S.S. degradation, as shown in Figure 6a,b, which are mainly due to the the moisture from the ambient atmosphere absorbed onto the IGZO surface. The absorbed moisture  moisture from the ambient atmosphere absorbed onto the IGZO surface. The absorbed moisture will will bond to the metal element in IGZO to form metal‐hydroxide (M‐OH) bonds, which act as donor‐ bond to the metal element in IGZO to form metal-hydroxide (M-OH) bonds, which act as donor-like like states to provide extra electrons, but also increase the density of state within the IGZO film [18].  states to provide extra electrons, but also increase the density of state within the IGZO film [18]. Thus, Thus, the absorbed moisture will cause a negative Vth shift and an S.S. degradation. For the a‐IGZO  the absorbed moisture will cause a negative Vth shift and an S.S. degradation. For the a-IGZO TFT TFT with the CF4 plasma treatment, the oxygen vacancies or the weak oxygen bonds on the IGZO  with the CF4 plasma treatment, the oxygen vacancies or the weak oxygen bonds on the IGZO surface surface are replaced by the F ions, as shown in Figure 2b. The M‐F bonds are stronger and more stable  are replaced by the F ions, as shown in Figure 2b. The M-F bonds are stronger and more stable than than the M‐O bond, and thus the formation of the M‐OH bonds is suppressed by the incorporation  the M-O bond, and thus the formation of the M-OH bonds is suppressed by the incorporation of of F. In summary, the a‐IGZO TFT with HfO2 as the gate insulator and the CF4 plasma treatment not  F. In summary, the a-IGZO TFT with HfO2 as the gate insulator and the CF4 plasma treatment not only acquires a higher immunity against humidity from ambient atmosphere, but also significantly  only acquires a higher immunity against humidity from ambient atmosphere, but also significantly improves the device performance.  improves the device performance.

Materials 2018, 11, x FOR PEER REVIEW

6 of 8

11, 824 x FOR PEER REVIEW Materials 2018, 11,

6 of 8 -1

10 CF plasma treatment -2 4 10 -1 condition VGS = 6 V 10-3 Stress 10 -2 CF4 plasma treatment 10-4 Under ambient environment 10 -3 Stress condition VGS = 6 V 10-5 10 -4 Under ambient environment 10-6 10 -5 10-7 10 -6 10-8 10 -7 10-9 10 -8 10-10 10 -9 10-11 10 -10 10-12 10 -11 10 -5 -4 -3 -2 -1 0 1 -12 10 V (V) -5 -4 -3 -2 -1GS 0 1 (b) V (V)

IDS (A) IDS (A)

0s 100 s 0s s 300 100 ss 500 300 ss 1000 500 s 2 1000 3 4s

2

3

IDS (A) IDS (A)

-1

10 W/O treatment -2 10 -1 Stress condition V = 6 V 10-3 W/O treatment GS 10 -2 Under ambient environment 10-4 10 -3 Stress condition VGS = 6 V 10-5 10 -4 Under ambient environment 10-6 10 -5 10-7 10 -6 10-8 10 -7 10-9 10 -8 10-10 10 -9 10-11 10 -10 10-12 10 -11 10 -5 -4 -3 -2 -1 0 1 -12 10 V (V) -5 -4 -3 -2 -1GS 0 1 (a) V (V)

4

2 2

0s 100 s 0s s 300 100ss 500 300 ss 1000 500 s 1000 3 4s

3

4

GS

GS

(a)

(b)

Figure 5. Transfer curves (IDS-VGS) of a-IGZO TFT with and without the CF4 plasma treatment under Figure 5.gate Transfer curves(PGBS) (IDS -V(V ) of= a-IGZO TFT with and without the CF4 plasma treatment under positive bias stress 6 V). GSGS Figure 5. Transfer curves (IDS-VGS) of a-IGZO TFT with and without the CF4 plasma treatment under positive gate bias stress (PGBS) (VGS = 6 V). positive gate bias stress (PGBS) (VGS = 6 V).

Threshold voltage Threshold voltage (V) (V)

1 0 0 -1 -1 -2 -2 -3 -3

0

200

400

600

800

1000

0

200Stress 400Time 600(sec) 800 (a)

1000

Stress Time (sec) (a)

1.4

W/O treatment CF4 Plasma Treatment W/O treatment CF4 Plasma Treatment

1.4 1.2 (V/decade) S.S. S.S. (V/decade)

W/O treatment CF4 plasma treatment W/O treatment CF4 plasma treatment

1

1.2 1.0 1.0 0.8 0.8 0.6 0.6 0.4 0.4 0.2 0.2

0

200

400

600

800 1000

0

200Stress 400Time 600(sec) 800 1000 (b)

Stress Time (sec) (b)

Figure 6. (a) Vth ; (b) S.S. as function of PGBS time of a-IGZO TFT. Figure 6. (a) Vth; (b) S.S. as function of PGBS time of a-IGZO TFT.

4. Conclusions Figure 6. (a) Vth; (b) S.S. as function of PGBS time of a-IGZO TFT. 4. Conclusions We have demonstrated that the CF plasma treatment can significantly improve the electrical 4. Conclusions We have demonstrated that the CF44 plasma treatment can significantly improve the electrical performance and reliability of the a-IGZO TFT with a HfO2 gate insulator. After the plasma treatment performance and reliability of the a-IGZO TFT with a HfO2 gate insulator. After the plasma treatment have that the CFthe 4 plasma can significantly improve the electrical at 15 We W for 20 sdemonstrated on the IGZO active layer, carrier treatment mobility significantly improved from 30.2 cm2 /V·s at 15 W for220 s on the IGZO active layer, the carrier mobility significantly improved from 30.2 cm2/V∙s performance and reliability of the a-IGZO TFT with a HfO 2 gate insulator. After the plasma treatment to 54.6 cm /V·s. This can be attributed to the incorporated F providing additional electrons in the to 54.6 cm2/V∙s. This can be attributed to the incorporated F providing additional electrons in2 the at 15 Wand for 20 s on the passivating IGZO activethe layer, the carrier significantly improved from 30.2 cmtraps /V∙s IGZO, in parallel interface trapsmobility at the IGZO/HfO interface. Low interface IGZO, and 2in parallel passivating the interface traps at the IGZO/HfO22 interface. Low interface traps to 54.6 cm /V∙s. This can be attributed to the incorporated F providing additional electrons in the result in less carrier scattering, which results in an increased carrier mobility. Moreover, it was found result in less carrier scattering, which results in an increased carrier mobility. Moreover, it was found IGZO, parallel passivating interfacesuppress traps at the hysteresis IGZO/HfOeffect 2 interface. Low interface traps that theand CF inplasma treatment can the effectively and enhance the device’s that the CF44 plasma treatment can effectively suppress the hysteresis effect and enhance the device’s result in less carrier scattering, which results in an increased mobility. Moreover, it was found immunity against humidity from the ambient atmosphere. It iscarrier believed that the CF4 plasma treatment immunity against humidity from the ambient atmosphere. It is believed that the CF4 plasma thatimprove the CF4 plasma treatment can effectively suppressofthe effect anda enhance theinsulator. device’s can the electrical performance and reliability thehysteresis a-IGZO TFT with HfO2 gate treatment can improve the electrical performance and reliability of the a-IGZO TFT with a HfO2 gate immunity against humidity from the ambient atmosphere. It is believed that the CF4 plasma insulator. Author Contributions: Conceptualization, C.-L.F. and F.-P.T.; Formal Analysis, F.-P.T. and C.-Y.T.; treatment can improve the electrical performance and reliability of the a-IGZO TFT withInvestigation, a HfO2 gate F.-P.T. and C.-Y.T.; Resources, C.-L.F.; Writing-Original Draft Preparation, F.-P.T. and C.-Y.T.; Writing-Review & insulator. Editing, C.-L.F.; Supervision, C.-L.F. Funding: This research was funded by the National Science Council of Taiwan under contract grant number NSC 106-2622-E-011-015-CC3.

Materials 2018, 11, 824

7 of 8

Acknowledgments: The authors would like to acknowledge the financial support of the National Science Council of Taiwan under contract no. NSC 106-2622-E-011-015-CC3 and the Taiwan Building Technology Center (TBTC) of National Taiwan University of Science and Technology (NTUST). Conflicts of Interest: The authors declare no conflict of interest.

References 1.

2.

3. 4.

5. 6.

7. 8. 9. 10. 11.

12. 13.

14. 15. 16. 17. 18.

19.

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. [CrossRef] [PubMed] Yabuta, H.; Sano, M.; Abe, K.; Aiba, T.; Den, T.; Kumomi, H.; Nomura, K.; Kamiya, T.; Hosono, H. High-mobility thin-film transistor with amorphous InGaZnO4 channel fabricated by room temperature rf-magetron sputtering. Appl. Phys. Lett. 2006, 89, 112123. [CrossRef] Suresh, A.; Wellenius, P.; Dhawan, A.; Muth, J. Room temperature pulsed laser deposited indium gallium zinc oxide channel based transparent thin film transistors. Appl. Phys. Lett. 2007, 90, 123512. [CrossRef] Kamiya, T.; Nomura, K.; Hosono, H. Origins of High Mobility and Low Operation Voltage of Amorphous Oxide TFTs: Electronic Structure, Electron Transport, Defects and Doping. J. Disp. Technol. 2009, 5, 273–288. [CrossRef] Mativenga, M.; Choi, M.H.; Choi, J.W.; Jang, J. Transparent Flexible Circuits Based on Amorphous-Indiumgallium-Zinc-Oxide Thin-Film Transistors. IEEE Electron Device Lett. 2011, 32, 170–172. [CrossRef] Lee, I.; Lee, S.; Gu, J.; Kim, K.; Cho, W. Comparative Study of Device Performance and Reliability in Amorphous InGaZnO Thin-Film Transistors with Various High-k Gate Dielectrics. Jpn. J. Appl. Phys. 2013, 52, 06GE05. [CrossRef] Hsu, H.; Chang, C.; Cheng, C.; Chen, P.; Chiu, Y.; Chiou, P.; Cheng, C. High Mobility Field-Effect Thin Film Transistor Using Room-Temperature High-k Gate Dielectrics. J. Disp. Technol. 2014, 10, 875–881. [CrossRef] Pan, T.; Chen, C.; Liu, J.; Her, J.; Koyama, K. Electrical and Reliability Characteristics of High-k HoTiO3 a-InGaZnO Thin-Film Transistors. IEEE Electron Device Lett. 2014, 35, 66–68. [CrossRef] Shao, Y.; Xiao, X.; He, X.; Deng, W.; Zhang, S. Low-Voltage a-InGaZnO Thin-Film Transistors With Anodized Thin HfO2 Gate Dielectric. IEEE Electron Device Lett. 2015, 36, 573–575. [CrossRef] Huang, X.D.; Song, J.Q.; Lai, P.T. Positive Gate Bias and Temperature-Induced Instability of a-InGaZnO Thin-Film Transistor with ZrLaO Gate Dielectric. IEEE Trans. Electron Devices 2016, 63, 1899–1903. [CrossRef] Park, J.C.; Cho, I.; Cho, E.; Kim, D.H.; Jeong, C.; Kwon, H. Comparative Study of ZrO2 and HfO2 as a High-k Dielectric for Amorphous InGaZnO Thin Film Transistors. J. Nanoelectron. Optoelectron. 2014, 9, 67–70. [CrossRef] Kim, H.; Mclntyre, P.C.; Saraswat, K.C. Effects of crystallization on the electrical properties of ultrathin HfO2 dielectrics grown by atomic layer deposition. Appl. Phys. Lett. 2003, 82, 106–108. [CrossRef] Hu, H.; Zhu, C.; Lu, Y.F.; Wu, Y.H.; Liew, T.; Li, M.F.; Cho, B.J.; Choi, W.K.; Yakovlev, N. Physical and electrical characterization of HfO2 metal-insulator-metal capacitors for Si analog circuit applications. J. Appl. Phys. 2003, 94, 551–557. [CrossRef] Lin, Y.; Chou, J. Temperature Effects on a-IGZO Thin Film Transistors Using HfO2 Gate Dielectric Material. J. Nanomater. 2014, 2014, 72. [CrossRef] Son, D.; Kim, D.; Sung, S.; Jung, E.; Kang, J. High performance and the low voltage operating InGaZnO thin film transistor. Curr. Appl. Phys. 2010, 10, e157–e160. [CrossRef] Chun, Y.S.; Chang, S.; Lee, S.Y. Effects of gate insulators on the performance of a-IGZO TFT fabricated at room-temperature. Micoelectron. Eng. 2011, 88, 1590–1593. [CrossRef] Na, J.H.; Kitamura, M.; Arakawa, Y. High field-effect mobility amorphous InGaZnO transistors with aluminum electrodes. Appl. Phys. Lett. 2008, 93, 063501. [CrossRef] Choi, S.; Kim, H.; Jo, C.; Kim, H.; Choi, S.; Kim, D.M.; Park, J.; Kim, D.H. A Study on the Degradation of In-Ga-Zn-O Thin-Film Transistors under Current Stress by Local Variations in Density of States and Trapped Charge Distribution. IEEE Electron Device Lett. 2015, 36, 690–692. [CrossRef] Lee, J.; Cho, I.; Lee, J.; Kwon, H. Bias-stress-induced stretched-exponential time dependence of threshold voltage shift in InGaZnO thin film transistors. Appl. Phys. Lett. 2008, 93, 093504. [CrossRef]

Materials 2018, 11, 824

20. 21.

22. 23. 24.

25.

26.

27.

28.

29.

8 of 8

Huang, X.D.; Song, J.Q.; Lai, P.T. Improved Stability of a-InGaZnO Thin-Film Transistor under Positive Gate Bias Stress by Using Fluorine Plasma Treatment. IEEE Electron Device Lett. 2017, 38, 576–579. [CrossRef] Liu, X.; Wang, L.L.; Hu, H.; Lu, X.; Wang, K.; Wang, G.; Zhang, S. Performance and Stability Improvements of Back-Channel-Etched Amorphous Indium-Gallium-Zinc Thin-Film Transistors by CF4 +O2 Plasma Treatment. IEEE Electron Device Lett. 2015, 36, 911–913. [CrossRef] Qian, L.Z.; Lai, P.T. Fluorinated InGaZnO Thin-Film Transistor with HfLaO Gate Dielectric. IEEE Electron Device Lett. 2014, 35, 363–365. [CrossRef] Jiang, J.; Toda, T.; Hing, M.P.; Wang, D.; Furuta, M. Highly stable fluorine-passivated In-Ga-Zn-O thin-film transistors under positive gate bias and temperature stress. Appl. Phys. Express 2014, 7, 114103. [CrossRef] Yamazaki, H.; Ishikawa, Y.; Fujii, M.; Ueoka, Y.; Fujiwara, M.; Takahashi, E.; Andoh, Y.; Maejima, N.; Matsui, H.; Matsui, F.; et al. The Influence of Fluorinated Silicon Nitride Gate Insulator on Positive Bias Stability toward Highly Reliability Amorphous InGaZnO Thin-Film Transistors. ECS J. Solid State Sci. Technol. 2014, 3, Q20–Q23. [CrossRef] Jiang, J.; Furuta, M.; Wang, D. Self-Aligned Bottom-Gate In-Ga-Zn-O Thin-Film Transistor with Source/Drain Regions Formed by Direct Deposition of Fluorinated Silicon Nitride. IEEE Electron Device Lett. 2014, 35, 933–935. [CrossRef] Furuta, M.; Jiang, J.; Hung, M.P.; Toda, T.; Wang, D.; Tatsuoka, G. Suppression of Negative Gate Bias and Illumination Stress Degradation by Fluorine-Passivated In-Ga-Zn-O Thin-Film Transistors. ECS J. Solid State Sci. Technol. 2016, 5, Q88–Q91. [CrossRef] Um, J.K.; Lee, S.; Jin, S.; Mativenga, M.; Oh, S.Y.; Lee, C.H.; Jang, J. High-Performance Homojunction a-IGZO TFTs with Selectively Defined Low-Resistive a-IGZO Source/Drain Electrodes. IEEE Trans. Electron Devices 2015, 62, 2212–2217. [CrossRef] Seo, J.; Jeon, J.; Hwang, Y.H.; Park, H.; Ryu, M.; Park, S.K.; Bae, B. Solution-Processed Flexible Fluorine-doped Indium Zinc Oxide Thin-Film Transistors Fabricated on Plastic Film at Low Temperature. Sci. Rep. 2013, 3, 2085. [CrossRef] [PubMed] Banuerfeldt, G.F.; Arbilla, G. Kinetic Analysis of the Chenical Processes in the Decomposition of Gaseous Dielectrics by a Non-Equilibrium Plasma-Part1:CF4 and CF4 /O2 . J. Braz. Chem. Soc. 2000, 11, 121–128. [CrossRef] © 2018 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).