Chapter 2 Device Fabrication and Experimental Setup
3.4 Summary
In this chapter, the passivation effect of NH3 plasma treatment for bottom gate poly-Si TFTs is thoroughly studied. In section 3.2, the influences of different NH3
plasma treatment time are discussed. Compared with the control sample without any plasma treatment, pre-SPC-treatment degrades device performance; however, post-SPC-treatment and after-AA-treatment improve device performance. The poly-Si TFTs with NH3 plasma pre-SPC-treatment have even worse performance than control sample owing to the smaller grain size and the out-diffusion of radicals during high
temperature SPC annealing. On the other hand, the poly-Si TFTs with after-AA-treatment ones exhibit the best performance than the counterparts.
After-AA-treatment is the most efficient technique to passivate traps since after-AA-treatment additionally passivates traps by radicals diffusing laterally through the gate oxide. Unfortunately, the poly-Si TFTs with after-AA-treatment suffer from the heaviest ion bombardment damage than those with post-SPC-treatment ones. Thus, for plasma treatment time > 30 min, the samples with after-AA-treatment have lower Gm_max compared with the samples with post-SPC-treatment. In addition, longer plasma treatment time enhance performance and alleviate kink effect more effectively.
And the bulk oxide quality is insensitive to the plasma with power of 50 W, because all samples with different plasma treatment time exhibit comparable gate-leakage current.
In section 3.3, the influences of different NH3 plasma power are systematically investigated. The higher the plasma power, the better the performance of these poly-Si TFTs. Higher plasma power enhance performance and alleviate kink effect more effectively. Because nitrogen and hydrogen radicals have higher kinetic energy with plasma power of 200 W than those with 50 W ones, they reach poly-Si/SiO2 interface more easily. However, plasma treatment with power of 200 W results in heavier ion bombardment damage than 50 W one. Thus, for plasma power of 200 W, the sample with after-AA-treatment has lower Gm_max than the sample with post-SPC-treatment.
Moreover, the bulk oxide quality for the sample with plasma power of 200 W is better than the sample with 50 W based on the gate-leakage comparison.
-10 -5 0 5 10 15
Plasma Power = 50W control the samples with NH3 plasma pre-SPC-treatment (plasma power = 50 W, plasma treatment time = 15 min to 120 min).
-10 -5 0 5 10 15
Plasma Power = 50W control the samples with NH3 plasma post-SPC-treatment (plasma power = 50 W, plasma treatment time = 15 min to 120 min).
-5 0 5 10 15 10
-1110
-910
-710
-5W/L = 10m/10m VDS = 0.1V
Plasma Power = 50W control
15min 30min 60min 120min
V
GS(V) I
DS(A)
0 40 80 120 After-AA 160
G
m(nS)
Fig. 3-3 IDS–VGS and Gm–VGS characteristics of poly-Si TFTs for control sample and the samples with NH3 plasma after-AA-treatment (plasma power = 50 W, plasma treatment time = 15 min to 120 min).
(a)
(b)
Fig. 3-4 Top-view of bottom gate poly-Si TFTs with plasma-induced hole traps in the gate oxide at the corner edge portions with (a) wide channel width and (b) narrow channel width.
Fig. 3-5 The transistor can be equivalent to flat plate transistor and corner edge transistor in parallel.
-10 -5 0 5 10 15
Plasma Power = 50W
I
DS(A )
Fig. 3-6 IDS–VGS characteristics of poly-Si TFTs for the samples with 2 μm channel width with NH3 plasma post-SPC-treatment (plasma power = 50 W, plasma treatment time = 15 min to 120 min).
Plasma Power = 50W
I
DS(A )
Plasma Treatment Time
Fig. 3-7 IDS–VGS characteristics of poly-Si TFTs for the samples with 2 μm channel width with NH3 plasma after-AA-treatment (plasma power = 50 W, plasma treatment time = 15 min to 120 min).
0 30 60 90 120 4
6 8 10
W/L = 10m/10m VDS = 0.1V
Plasma Power = 50W
V
TH(V)
Plasma Treatment Time (min)
Pre-SPC Post-SPC After-AA
Fig. 3-8 Threshold voltage (VTH) as a function of plasma treatment time.
0 30 60 90 120
60 90 120
W/L = 10m/10m VDS = 0.1V
Plasma Power = 50W
G
m_max(nS)
Plasma Treatment Time (min)
Pre-SPC Post-SPC After-AA
Fig. 3-9 Maximum transconductance (Gm_max) as a function of plasma treatment time.
Fig. 3-10 NH3 plasma after-AA-treatment passivates channel traps not only by radicals diffusing vertically through the poly-Si channel, but also by radicals diffusing laterally through the gate oxide.
Ion Bombardment Damage
Fig. 3-11 The extra damage on the sidewall of the channel for the poly-Si TFTs with NH3 plasma after-AA-treatment.
0 2 4 6
Plasma Power = 50WI
DS( A)
Fig. 3-12 IDS–VDS characteristics of poly-Si TFTs for control sample and the samples with NH3 plasma post-SPC-treatment (plasma power = 50 W, plasma treatment time = 15 min to 120 min). Plasma Power = 50W
I
DS( A)
Fig. 3-13 IDS–VDS characteristics of poly-Si TFTs for control sample and the samples with NH3 plasma after-AA-treatment (plasma power = 50 W, plasma treatment time = 15 min to 120 min).
0 30 60 90 120 0.2
0.4 0.6 0.8
Post-SPC
Sl op e ( A/V)
Plasma Treatment Time (min)
VGS-VTH=1V VGS-VTH=2V VGS-VTH=3V
Fig. 3-14 Slopes of IDS–VDS characteristics (post-SPC-treatment) as a function of plasma treatment time (plasma power = 50 W, plasma treatment time = 15 min to 120 min).
0 30 60 90 120
0.2 0.4 0.6 0.8
After-AA
Sl op e ( A/ V)
Plasma Treatment Time (min)
VGS-VTH=1V VGS-VTH=2V VGS-VTH=3V
Fig. 3-15 Slopes of IDS–VDS characteristics (after-AA-treatment) as a function of plasma treatment time (plasma power = 50 W, plasma treatment time = 15 min to 120 min).
-5 0 5 10 15 20 25
Plasma Power = 50W
Post-SPC
Fig. 3-16 Gate-leakage current characteristics of poly-Si TFTs for control sample and the samples with NH3 plasma post-SPC-treatment (plasma power = 50 W, plasma treatment time = 15 min to 120 min).
Plasma Power = 50W
After-AA
Fig. 3-17 Gate-leakage current characteristics of poly-Si TFTs for control sample and the samples with NH3 plasma after-AA-treatment (plasma power = 50 W, plasma treatment time = 15 min to 120 min).
-10 -5 0 5 10 15
Plasma Time = 15min
Pre-SPC
G
m(nS)
Fig. 3-18 IDS–VGS and Gm–VGS characteristics of poly-Si TFTs for control sample and the samples with NH3 plasma pre-SPC-treatment (plasma power = 50 W or 200 W, plasma treatment time = 15 min).
-10 -5 0 5 10 15
Plasma Time = 15min
Post-SPC
control 50W200W
Fig. 3-19 IDS–VGS and Gm–VGS characteristics of poly-Si TFTs for control sample and the samples with NH3 plasma post-SPC-treatment (plasma power = 50 W or 200 W, plasma treatment time = 15 min).
-5 0 5 10 15
Plasma Time = 15min
After-AA
control 50W200W
Fig. 3-20 IDS–VGS and Gm–VGS characteristics of poly-Si TFTs for control sample and the samples with NH3 plasma after-AA-treatment (plasma power = 50 W or 200 W, plasma treatment time = 15 min).
0 50 100 150 200
Plasma Time = 15min
V
TH(V)
Fig. 3-21 Threshold voltage (VTH) as a function of plasma power.0 50 100 150 200
Plasma Time = 15min
G
m_max(n S)
Fig. 3-22 Maximum transconductance (Gm_max) as a function of plasma power.0 2 4 6
Plasma Time = 15min
I
DS( A)
Fig. 3-23 IDS–VDS characteristics of poly-Si TFTs for control sample and the samples with NH3 plasma post-SPC-treatment (plasma power = 50 W or 200 W, plasma treatment time = 15 min).
0 2 4 6
Plasma Time = 15min
I
DS( A)
Fig. 3-24 IDS–VDS characteristics of poly-Si TFTs for control sample and the samples with NH3 plasma after-AA-treatment (plasma power = 50 W or 200 W, plasma treatment time = 15 min).
Fig. 3-25 Slopes of IDS–VDS characteristics (post-SPC-treatment) as a function of plasma treatment time (plasma power = 50 W or 200 W, plasma treatment time = 15 min).
0 50 100 150 200
Fig. 3-26 Slopes of IDS–VDS characteristics (after-AA-treatment) as a function of plasma treatment time (plasma power = 50 W or 200 W, plasma treatment time = 15 min).
Plasma Time = 15min
Post-SPC
Fig. 3-27 Gate-leakage current characteristics of poly-Si TFTs for control sample and the samples with NH3 plasma post-SPC-treatment (plasma power = 50 W or 200 W, plasma treatment time = 15 min).
-5 0 5 10 15 20 25 10
-1310
-1110
-910
-710
-510
-3After-AA
W/L = 10m/10m VD = VS = 0V
Plasma Time = 15min
I
GS(V)
V
GS-V
TH(V)
control 50W 200W
Fig. 3-28 Gate-leakage current characteristics of poly-Si TFTs for control sample and the samples with NH3 plasma after-AA-treatment (plasma power = 50 W or 200 W, plasma treatment time = 15 min).
Chapter 4
Impacts of Post-Metal-Annealing on Plasma-Treated Bottom Gate Poly-Si TFTs
4.1 Introduction
Due to the passivation effect of hydrogen and nitrogen radicals, NH3 plasma treatment has been widely used to improve the performance and the hot-carrier reliability of poly-Si TFTs [4.1]-[4.2]. However, poly-Si TFTs with plasma treatment may lose their passivation effect after subjecting to high temperature annealing [4.2]-[4.3]. In this chapter, the impacts of post-metal-annealing on plasma-treated bottom gate poly-Si TFTs will be thoroughly studied.
4.2 Plasma Treatment Time Dependence
The comparisons of IDS–VGS and Gm–VGS characteristics between the poly-Si TFTs with and without post-metal-annealing (PMA) are shown in Fig. 4-1 – Fig. 4-3.
The samples with PMA were annealed at 400℃ for 30 minutes in N2 ambient. The subthreshold double-hump phenomenon has been completely eliminated by PMA. In chapter 3, we hypothesized the subthreshold double-hump phenomenon was caused by the hole traps in the gate oxide at the corner edge portions, which were generated by plasma dry etching [4.4]-[4.5]. As a result, the elimination of the subthreshold double-hump phenomenon comes from the elimination of the hole traps in the gate oxide. Moreover, as shown in Fig. 4-1 (a) (curve: W/ PMA), because the hole traps in the gate oxide at the corner edge portions have been eliminated, the flat plate transistor with high VTH dominates the IDS–VGS characteristics. PMA in N2 ambient may be a method to annihilate the hole traps generated by plasma dry etching. In
addition, the IDS–VGS characteristics for the samples with 2 μm channel width are shown in Fig. 4-4 and Fig. 4-5. In chapter 3, we hypothesized the hole traps at the corner edge portions may construct a leakage path and lead to a large IOFF. Nevertheless, after PMA, the IOFF of all samples decrease significantly because PMA can eliminate the leakage path constructed by the hole traps.
Maximum transconductance (Gm_max) as a function of plasma treatment time for the samples with and without PMA are shown in Fig. 4-6. Substantial improvements of Gm_max for the poly-Si TFTs (pre-SPC-treatment) with PMA are measured.
Surprisingly, control sample without any NH3 plasma treatment has a large improvement of Gm_max with PMA. These improvements of Gm_max for the poly-Si TFTs with PMA are possibly due to the elimination of the hole traps in the gate oxide.
Besides, the amounts of the improvements of Gm_max for each sample with different NH3 plasma treatment time are almost the same.
However, for the poly-Si TFTs with NH3 plasma post-SPC-treatment as shown in Fig. 4-7, the saturation phenomena of Gm_max for the samples with PMA have been measured. Without PMA, Gm_max become higher as plasma treatment time increases because more nitrogen and hydrogen radicals passivate the intra-grain defects. On the other hand, with PMA, the samples with and without NH3 plasma have the Gm_max
saturation value of about 113 nS. Even though longer NH3 plasma treatment time yields better passivation effect, some nitrogen and hydrogen radicals release from the defect sites during high temperature PMA processing [4.2]-[4.3]. Therefore, the saturation phenomena of Gm_max can be attributed to the combination of following mechanisms: 1) the elimination of the hole traps in the gate oxide by PMA, 2) the out-diffusion of nitrogen and hydrogen radicals during PMA processing, and 3) the remaining passivation effect by NH3 plasma.
Nevertheless, for the samples with NH plasma after-AA-treatment as shown in
Fig. 4-8, although Gm_max saturates to 113 nS for plasma treatment of 60 min, it decreases from 113 nS to 105 nS as plasma time increases from 60 min to 120 min owing to heavier ion bombardment as plasma treatment time is further increased.
Hence, the results of the Gm_max are attributed to the combination of following mechanisms: 1) the elimination of the hole traps in the gate oxide by PMA, 2) the out-diffusion of nitrogen and hydrogen radicals during PMA processing, 3) the remaining passivation effect by NH3 plasma, and 4) the plasma-induced damage by NH3 plasma.
4.3 Plasma Power Dependence
Fig. 4-9 – Fig. 4-11 illustrate the comparisons of IDS–VGS and Gm–VGS
characteristics between the poly-Si TFTs with and without PMA. Also, the subthreshold double-hump phenomenon has been completely suppressed because of the elimination of the hole traps in the gate oxide by PMA.
Maximum transconductance (Gm_max) as a function of plasma power for the samples with and without PMA are shown in Fig. 4-12 – Fig. 4-14. As shown in Fig.
4-12, for the poly-Si TFTs with NH3 plasma pre-SPC-treatment, substantial improvements of Gm_max for the samples with PMA process have been measured due to the elimination of the hole traps in the gate oxide. Moreover, after PMA, the extent of the improvements of Gm_max for each sample with different NH3 plasma power is almost the same.
The saturation phenomena of Gm_max for the samples (post-SPC-treatment) with PMA have been measured as shown in Fig. 4-13 because of 1) the elimination of the hole traps, 2) the out-diffusion of radicals, and 3) the remaining passivation effect.
Meanwhile, with plasma power of 200 W, the samples without PMA have higher G than those with PMA ones since the out-diffusion of radicals during PMA
leads to the lost passivation effect.
Furthermore, as shown in Fig. 4-14, the sample with after-AA-treatment for 120 min has lower Gm_max than the sample with plasma treatment time of 60 min due to heavier ion bombardment as plasma treatment time is further increased. Thus, the results of the Gm_max are attributed to 1) the elimination of the hole traps, 2) the out-diffusion of radicals, 3) the remaining passivation effect, and 4) the plasma-induced damage.