Results and Discussion

Organic thin film transistors are used for flexible electronics which could not sustain higher temperature processes. To understand the impact of substrate temperature for film deposition quality depending on surface morphology, thermodynamics and deposition rate, we choose different substrate temperatures to evaluate the insulator quality and gate leakage current of device.

Figure 2-1-3 shows leakage current density of silicon oxide deposited by APPJ.

The leakage current density of silicon oxide deposited by APPJ (APPJ-SiO2) is about 5E-7 A/cm² at E = 0.5 MV/cm. However, the deposition rate at room temperature for APPJ-SiO2 insulator on aluminum electrode is almost zero, possibly ascribed to poor chemical reaction rate on aluminum substrate. As we known, the surface active energy is highly related to substrate temperature based on chemical vapor deposition dynamics. The equation of chemical reaction rate was shown in Eq. (2-1).

) / exp(

.

.R A E kT

C   _a (2-1)

where A is a constant, Ea is active energy, T is substrate temperature, and k is

24

Boltzmann constant. According to chemical reaction mechanism, the surface reaction can occur easily for the case with a low active energy. Therefore, we can choose different material with a low active energy or lower the reaction barrier by increasing substrate temperature to improve the deposition quality. Here, the different metals including Ni, Ti, Ta, Pt, and Pd have been applied to be the bottom electrode of the MIM capacitor, but the deposition rate of SiO2 (APPJ) at room temperature is still poor, close to zero. We also used different metal oxide (HfO₂, AlO₂O₃) as the buffered materials between SiO2/electrode to decrease the active energy. However, the deposition quality is still poor.

To deposit silicon oxide on the metal by APPJ, we tried to increase the substrate temperature up to 100 ^oC, 150 ^oC, and 200 ^oC. We used aluminum as the bottom electrode for MIM structure. Figure 2-1-4 shows the thickness of APPJ-SiO2 at different temperatures, indicating deposition rate increased with substrate temperature, like surface reaction control. To clarify the effect of temperature-dependence deposition, a fixed flow rate of 200 sccm was applied to concise our experiment design. The detail discussion on the influence of flow rate would be investigated in next section. Figure 2-1-5 shows the leakage current density of silicon oxide deposited by E-gun and APPJ, respectively. The optimized APPJ-SiO2 at 150C presents a leakage current density of 1E-7 A/cm² at 0.5 MV/cm, much better than the silicon oxide deposited by E-gun with a leakage current of 4E-6 A/cm².

However, the SiO₂ deposition at room temperature is the future demand. The improvement of surface reaction between Si and bottom electrode is crucial important.

Therefore, we deposited silicon oxide on Al electrode with an amorphous Si buffered layer at room temperature. Figure 2-1-6(a) and (b) show C-V and I-V characteristics of Al/SiO2/amorphous Si/Al MIM structure, respectively. The good capacitance and leakage current characteristics for different scaling-times cases during demonstrate the

25

stable APPJ deposition process and reliable film quality. However, the experiment confirmed that the APPJ-SiO2 film could be deposited on a-Si and SiO2 substrate at room temperature since the active energy of a-Si and SiO₂ were lower than other materials. The experimental results support our previous assumption, which a lower surface active energy facilitate APPJ-SiO₂ deposition on Al electrode. As we mentioned before, the higher thermal budget and Si buffered layer can deposit SiO2

successfully on Al electrode at room temperature, but the approach aided by buffered layer could increase the thickness variability and process complexity except for extra cost. In addition, we emphasized the importance of process temperature, since most plastic substrates could not sustain the temperature higher than 200 ^oC. The APPJ processed SiO₂ at 150 ^oC not only maintains a good deposition rate for throughput but also presents a good film quality which show the potential for low-temperature flexible electronic device fabrication.

After the above discussion, we have proposed approaches by using lower temperature control (<150C) and Si buffered layer to solve the issue of film deposition. Sequentially, the transistor characteristics will be discussed in the following part.

The drain current (ID) versus drain-source voltage (VDS) at varied gate voltages (V_GS) shown in Figure 2-1-7. The output characteristic of I_D versus V_GS was shown in Figure 2-1-8. The carrier mobility at the saturation region was calculated by the equation:

^I_D 



^{W C i/2L}



^(V_G ^V_T⁾², [2]

Ci is the capacitance per unit area of gate insulator in Eq (2). It correspond to a device with channel width W = 200 μm and length L = 100 μm.

26

The operational voltage was below -4 V for a EOT of 12.4 nm. The saturation mobility and the threshold voltage of the OTFT were about 0.066 cm² V^-1s^-1 and – 1.8 V, respectively. There are three possible reasons to explain why the mobility lower than pervious researches [14]-[17]. First, the grain size of the active layer would be affected by the roughness of the electrodes of source and drain due to bottom contact TFT structure. Secondly, poor oxide surface roughness could lead to the mobility degradation. In Figure 2-1-9, the roughness (RMS) of Al gate electrode was about 8.2 nm and a large RMS value of ~10.8 nm can be obtain after APPJ-SiO2 on the aluminum electrode. The uneven Al electrode could affect the roughness of following APPJ-SiO2. The contact angle of about 25 degree for APPJ-SiO2, showing hydrophilic characteristic on the oxide surface is another issue since the surface roughness and hydrophilic characteristics should be the main factors to influence the deposition of pentacene and then decrease the mobility of OTFT devices. To improve these issues, originated from bottom electrode with larger leakage current, we replaced Al (4.2eV) electrode by a higher work function Ni (5.1eV). Figure 2-1-10 shows the RMS of silicon oxide deposited at different temperatures on Al and Ni, respectively. The oxide on an evaporated Ni by E-gun has a RMS value of 2.2 nm, lower than that on Al. To reduce surface roughness of gate insulator and improve the gate leakage current, we used Ni as the bottom electrode of OTFT in the following experiment.

27 Silicon

substrate

Silicon

dioxide electrode

SAPPT silicon oxide

Metal Insulator

Metal

Figure 2-1-1 MIM structure for testing gate insulator’s quality.

50 nm Ni 50 nm Ni

Silicon substrate 500 nm Thermal oxide

100 nm Al 12.4 nm silicon oxide

50 nm pentacene

Figure 2-1-2 Scheme of the top contact structure of organic thin transistor

28

Figure 2-1-3 Leakage current density of silicon oxide deposited by APPJ and MIS structure.

Figure 2-1-4 Thickness of silicon oxide with 60 times at different temperatures.

A B C D

0 2 4 6 8 10 12 14 16 18 20

Thickness (nm)

SiO2 (APPJ) A @ RT B @ 100 OC C @ 150 OC D @ 200 OC

29

Figure 2-1-5 J-E of SiO2 deposited by APPJ in MIM structure. Leakage current density of SiO2 (APPJ), deposited at 150 ^oC, was about 1E-7 A/cm² at 0.5 MV/cm.

(a)

60 80 100 120

16 18 20 22 24 26 28 30 32

Scanning Times T h c ik n e s s ( n m )

Al/SiO2/a-Si/Al MIM Structure

SiO2 deposited by APPJ @ RT 60 times

80 times 100 times 120 times

30

(b)

Figure 2-1-6 (a) Thickness of silicon oxide with a a-Si buffer layer at different scanning times. (b) I-V characteristic of SiO2 using APPJ at different scanning times.

Figure 2-1-7 Scheme of the output characteristics of pentacene-based organic thin film transistor with a lower operation voltage about -3 V.

0.0 0.2 0.4 0.6 0.8 1.0

31

-4 -3 -2 -1 0

0.00E+000 5.00E-009 1.00E-008 1.50E-008 2.00E-008

(-IDS)0.5 (A)0.5 (-IDS) (A)

VG (Voltage)

-4 -3 -2 -1 0

0.00005 0.00010 0.00015

Figure 2-1-8 Scheme of transfer characteristics of penetacene-based organic thin film transistor with mobility about 0.066 cm²/V-S.

Figure 2-1-9 (a) AFM image of aluminum gate electrode with roughness about 8.7 nm.(b) AFM image silicon oxide (APPJ) deposited at 150 ℃ on the aluminum with the roughness around 10.8 nm

32

Figure 2-1-10 Scheme of the RMS of SiO2 deposited at different temperatures separately on Al and Ni.

33

References

[1] S. E. Babayan, J. Y. Jeong, V. J. Tu, J. Park, G. S. Selwyn, and R. F. Hicks,

“Deposition of silicon dioxide films with an atmospheric-pressure plasma jet,” Plasma Sources Sci. Technol., vol. 7, pp. 286-288, 1998.

[2] A. Ladwig, S. Babayan, M. Smith, M. Hester, W. Highland, R. Koch, R.

Hicks, “Atmospheric plasma deposition of glass coatings on aluminum,” Surface &

Coatings Technology, vol. 201, pp. 6460-6464, 2007.

[3] G. R. Nowling, M. Yajima, S. E. Babayan, M. Moravej, X. Yang, W. Hoffman, and R. F. Hicks, “Chamberless plasma deposition of glass coatings on plastic,”

Plasma Sources Sci. Technol. vol. 14, pp.477-484, 2005.

[4] J. Y. Jeong, S. E. Babayan, A. Schu¨ tze, V. J. Tu, and R. F. Hicks, “Etching polyimide with a nonequilibrium atmospheric-pressure plasma jet,” J. Vac. Sci.

Technol. A, vol. 17, no. 5, 1999.

[5] K. Maruyama, I . Tsumagari, M. Kanezawa, Y. Gunji, M. Morita, M. Kogoma, and S. Okazaki, “Preparation of ZnO films from Zn2+ aqueous mist using atmospheric pressure glow plasma,” J. Mater. Sci. Lett. vol. 20, pp. 481-484, 2001.

34

2-2 The Effect Carrier Gas’s Flow Rate on The Surface Morphology and Film Quality

Carrier gas’s flow rate may be an influence factor for depositing SiO2, discussed in many studies. In APPJ, flow rate would alter the concentration of the precursor at nozzle, influencing the deposition rate and the quality of silicon oxide. In this part, bubble method was utilized to carry the precursor into the plasma region and Ar was used as the carrier gas. The experimental detail was shown in Table 2-2-1. We selected substrate temperature at 150 ℃ because every experimental parameter must be suitable for plastic substrate. From Figure 2-2-1, we could obviously found that the leakage current density was increased with increasing the carrier gas’s flow rate.

When the gas flow rate was less than 100 sccm, the leakage current density could be suppressed below 2 E-8 A/cm² at 0.5 MV/cm. However, a larger leakage current of two order of magnitude (1E-6 A/cm²) than optimized one was obtained as introducing an excess follow rate of > 200 sccm. From Figure 2-2-2, the thickness and roughness were increased with increasing the flow rate of carrier gas. The AFM plots of SiO2

were also shown in Figure 2-2-3. The roughness of silicon oxide increased rapidly when the flow rate over 200 sccm. The rapid deposition rate would cause silicon oxide too loose to suppress leakage current. In addition, we used SEM to get the top view of silicon oxide surface shown in Figure 2-2-4 and found that these pellets on the silicon oxide surface was also increased obviously at higher flow rate. These pellets must be decreased because the surface roughness would scatter transport carriers which would decrease the mobility of OTFT. In addition, organic molecules

35

ordering and grain size would be degraded with high surface roughness [1-2] which would decrease the mobility of organic semiconductor. In our experiment, the roughness of silicon oxide could be reduced by decreasing the flow rate and the pellets also could be decreased, which would increase the possibility of application for OTFT. Figure 2-2-5 shows the XPS of silicon oxide deposited with different flow rates and the percentage of carbon existed in silicon oxide, shown in Table 2-2-2, was increased with increasing the flow rate. Some precursor may not be decomposed completely and then deposited on the surface of sample to form the impurities which may consist of C-H, C-O, and C-O-O in our demonstrated silicon oxide. This phenomenon may be more serious when the flow rate increases. CVD reactions are homogeneous, heterogeneous, or a combination of both. Homogeneous reactions nucleate in the gas phase and lead to particle formation. The greatest single problem in CVD technology. Most CVD processes are chosen to be heterogeneous reactions [3-4]. That is, they take place at the substrate surface rather than in the gas phase and form the desirable film deposit. In general, increasing temperature leads to increase film deposition rate, greater density, and improved structural perfection and crystallinity of the deposits. The quality of silicon oxide deposited by APPJ in this thesis was also improved with increasing temperature in lower flow rate. It is very important to control temperature and flow rate for depositing silicon oxide at heterogeneous reactions. Although silicon oxide deposited by APPJ in air may combined some particles or impurities, low leakage current density of insulator was obtained by well controlling temperature, flow rate, and other parameters of APPJ.

36

Table 2-2-1 The details of experimental parameters of silicon oxide deposited with different flow rates by APPJ.

Experimental parameters

Main gas CDA

Speed (mm/sec) 30

Gap distance ( cm² ) 2.2

Scanning times 60

Ar flow rate (sccm) 60~300

Substrate temperature ( ℃ ) 150

37

Figure 2-2-1 Plot of leakage current density versus electric field (J-E) of our APPJ- SiO2 deposited with different carrier gas’s flow rates.

50 100 150 200 250 300

Figure 2-2-2 Deposition rate and RMS of silicon oxide deposited with different flow rates of atmospheric pressure plasma jet.

38

(a) (b)

(c ) (d)

Figure 2-2-3 AFM images of silicon oxide fabricated with different Ar flow rates of APPJ. (a) 60 sccm (b) 100 sccm (c) 200 sccm (d) 300 sccm

39

(a) (b)

(c ) (d)

Figure 2-2-4 SEM images of silicon oxide fabricated with different Ar flow rates of atmospheric pressure plasma jet.

(a) 60 sccm (b) 100 sccm (c) 200 sccm (d) 300 sccm

40

Figure 2-2-5 X-ray Photoelectron Spectroscopy spectra of SiO2 (APPJ) deposited with different Ar flow rates.

Table 2-2-2 The concentration of carbon in silicon oxide, deposited with different flow rates, was analyzed by X-ray Photoelectron Spectroscopy.

5.87

41

Reference

[1] J. H. Seo, J. H. Kwon, and S. I. Shin, “Organic thin film transistors with polyvinyl alcohol treated dielectric surface,” Semiconductor Science and Technology, vol. 22, no. 9, pp. 1039-1043, Sep. 2007.

[2] W. H. Lee and C. C. Wang, “Effect of nanocomposite gate dielectric roughness on pentacene field-effect transistor,” J. Vac. Sci. Technol. B, vol. 27, no. 3, pp. 1116-1121, May 2009.

[3] K. J. Huttinger, “CVD in Hot Wall Reactors - The Interaction Between Homogeneous Gas-Phase and Heterogeneous Surface Reactions,” Chemical Vapor Deposition, vol. 4, no. 4, pp. 151-158, pp. JUN 1998.

[4] G. G. Condorelli, A. Baeri, I. L. Fragala, “Homogeneous and heterogeneous reactions in the decomposition of precursors for the MOCVD of high-k and ferroelectric films,” Materials Science in Semiconductor Processing, vol. 5, no. 2-3, pp. 135-139, Apr. Jun. 2002

[5] V. Nehra, A. Kumar,and H. K. Dwivedi, “Atmospheric Non-Thermal Plasma Sources,” International Journal of Engineering, vol. 2, no. 1, pp. 53-68,

42

2-3 The Influence of Gap Distance on Deposition Rate and Film Leakage

In this section, we only focused the study on the gap distance from nozzle to sample surface, as shown in Figure 2-3-1. The detail of experimental parameters were shown in Table 2-3-1. CDA was used as the main gas to investigate the effect of gap distance on deposition process since the deposition rate for APPJ-SiO2 with a main gas of CDA was more stable than the others (O₂, N₂). In Figure 2-3-2, the deposition rate of silicon oxide increasing with gap distance decreasing could be observed. This may be attributed to the higher species concentration near Al surface participate the film deposition at a small gap distance, which would increase the surface reaction probability and deposition rate.

In previous section of 2-1 and 2-2, we demonstrated the deposition rate would be influenced by temperature, increasing with substrate temperature due to a decrease of active energy. In addition, modifying species concentration by an appropriate flow rat to avoid the occurrence of homogeneous reactions during deposition is also important.

As we known, the mean free path (MFP) of reactive species would be decreased in atmospheric pressure, compared to high vacuum environment. Also, the concentration of reactive species near the Al substrate would be decreased while the gap distance increases. In Figure 2-3-2, the deposition rate increase with gap distance deceasing but no film deposition process can be observed at an excess gap (>2.5cm), possibly caused by the lack of reactive species.

We concluded that the increase of deposition rate with gap distance may be attributed to two reasons. One is the concentration of reaction species and the other may be contributed from the “varied” fixed substrate temperature. Although the fixed substrate temperature of 150 ^oC has been applied in this experiment, the temperature near the substrate still possibly varied with changing the gap distance. We believed

43

that the additional thermal budget for the “fixed” substrate temperature is from the glowing plasma source. However, as the gap distance was decreased to 1.8 mm, the deposition rate decreased abruptly and the roughness increased. We suggested that the strong plasma flow would bombard the substrate surface with the decrease of gap distance, which may leave damages on the substrate surface to lower the deposition rate. In addition, we also think that some reactive species with high dynamic energies would be reflected and then formed nucleated particles in the air to fall on the substrate, which may lead to the large surface roughness. Fig. 2-3-3 shows the leakage current density versus electric field of silicon oxide deposited with atmospheric pressure plasma at different gap distances. We could found that the leakage current density of silicon oxide with a gap distance of 1.8 mm is the largest, suggesting that the plasma damage is major contribution in the leakage performance, as we described earlier.

The leakage current density and surface roughness for APPJ-SiO2 with different gap distances have been summarized in Table 2-3-3. While the gap distance was decreased from 2 mm to 1.8 nm, the leakage current density can be lowered about 2 times. However, the large surface roughness (3.618 nm) could influence the coverage of organic layer and device mobility. In this study, we selected the gap distance of 2.2 mm as process parameter to fabricate gate insulator of OTFT because the condition has good insulator quality and an appropriate deposition rate.

According to the previous discussion, we consider that the concentration of ionization species would influence the deposition rate of APPJ-SiO2 with various gap distances and oxide quality might be degraded at a small gap distance between nozzle and substrate (<2.0 mm).

44

Figure 2-3-1 Schematic diagram of gap distance of APPJ. Gap distance means the distance from nozzle to the surface of sample.

Table 2-3-1 Experimental parameters of APPJ-SiO2 deposited with different gap distances.

45

Figure 2-3-2 Deposition rate and RMS of APPJ-SiO2 deposited with different gap distances.

Table 2-3-2 Thickness and deposition rate of APPJ-SiO2 deposited with different gap distances.

46

Figure 2-3-3 Leakage current density versus electric field of APPJ-SiO2 with different gap distances.

Table 2-3-3 The leakage current density and surface roughness of silicon oxide deposited with different gap distances.

Gap distance (cm)

Leakage current density (A/cm²) at 0.5 MV/cm

Root Mean Square (nm)

1.8 2.29E-7 3.618

2 5.34E-8 2.79

2.2 2.12E-8 2.47

2.5 1.46E-8 2.26

2-4 The Effect of Main Gas on The Deposition Rate and Electrical

47

Characteristics

2-4-1 Introduction

Main gas, used to generate plasma, is a key factor for precursor dissociation and film quality. During process, main gas was introduced into the high electric field region between anode and cathode to generate plasma source, including electrons, ions, metastable atoms. Under deposition process, deposition rate, hardness, and refraction index deposited with different main gases were widely studied. Some researches used low oxygen-ratio main gas to increase the deposition rate of silicon oxide [1]-[3], since oxygen could increase the dissociation of organic precursor. In this section, we would investigate the influence of different main gases on deposition rate, surface roughness, and leakage current of device.

2-4-2 Experiment

Main gas flow into the high electric field region between electrodes of APPJ was shown in Figure 2-4-1. In this section, we utilized oxygen, nitrogen, and CDA as the main gas for APPJ to decompose tetraethoxysilane (TEOS). The Metal-Insulator-Metal (MIM) capacitor was used to study electrical characteristics.

First, a 500-nm-thick SiO₂ insulation layer was grown on the p-type silicon substrate by using a wet oxidation method at 1000 ^oC. Then, a 50-nm-thick Ni deposited by E-gun evaporator was used as bottom electrode and low temperature SiO₂ (APPJ) was deposited on it using different main gases including oxygen, nitrogen, and CDA for this experiment. After the formation of the low temperature SiO2, a 50-nm-thick Ni was deposited on the SiO2 (APPJ) by E-gun evaporator. The optimized substrate

48

temperature was about 150 ^oC. As for material analysis, the surface roughness and film compound were analyzed by AFM and XPS, respectively. The C-V and I-V characteristics were measured by HP4156 and HP4284, respectively.

2-4-3 Results and Discussion

Oxygen, nitrogen, and CDA were used to investigate the influence of different

Oxygen, nitrogen, and CDA were used to investigate the influence of different

在文檔中 以噴射式大氣電漿在低溫下開發高品質二氧化矽應用在低電壓操作之有機薄膜電晶體之研究 (頁 38-0)