Chapter 4 Reliability in LTPS High-k TFTs
4.2 Fluorine Implantation
4.2.2 Positive Bias Temperature Instability
After discuss the hot carrier stress which happened locally in the drain side, we look for the PBTI which qualify all the material degrade issue. Figure 4-12 show the transfer characteristic of control and fluorine implanted sample, before and after PBTI stress at 25 oC with Vg-Vth=5 V, Vd=Vs=0 V. One could easily see that both the drain current of control and fluorine in the upper region of sub-threshold segment become not steeped, and we suggest that was due to the donor like states created in the upper of band gap[71, 72]. Figure 4-13 and 4-14 show the process of defect generated which have the same ideal with defect pool model. Since Fermi level had been operating in the on state for n-channel device, the band bending curve would be show like Figure 4-13. The defect states have been generated at the time device operate in channel turn on, but the states still under Fermi level which will not contribute transfer curve. As the time our device measure after stress, the Fermi level will move downward, and the electron in the states will flow out and the defect states will lost the energy which electron carry away, followed by the defect energy level move download but still locate in the half of upper band gap. At the time we sweep the gate voltage again, Fermi level move upward and the electron would be trapped into the states created by stress, hence the hump in the transfer characteristic would be found. But we should notice that this phenomenon could not be see in the hot carrier stress, due to the drain bias in hot carrier operation will reduce the electric field upward in the drain side.
Therefore, the donor like states would generate fewer in HCS than PBTI.
Figure 4-15 show the transconductance of control and fluorine implanted sample before and after PBTI stress. We could notice the larger Gm maximum degradation in control than fluorine implanted sample, which contribute to the strong Si-F bond exist
the degradation percentage could also be seen in compare PBTI with HCS, the improvement in fluorine in HCS is 9.7 % better than control, but for the PBTI is 46.76
%. Thus the effect of material degraded will be expand under PBTI operation, which can help as much more easily see the passivated effect by fluorine incorporation.
Figure 4-16 and 4-17 show the time evolution of threshold voltage shift and sub-threshold swing degradation, respectively. We could see the same phenomenon with HCS in PBTI, the less threshold voltage shift contribute the defect in HfO2 been passivate by the fluorine diffuse and resist the to the gate voltage stress damage. And here one could see the logarithm time dependent in the threshold voltage shift in both the control and fluorine implanted sample, which mean that electron trapping in HfO2
still play the major role. For the better sub-threshold swing in fluorine implanted sample, which suggest that the better deep state reduction in the interface have been passivate by strong Si-F bond.
Figure 4-18 and 4-19 show the transfer characteristic of control and fluorine implanted sample with different temperature operation, respectively. The fluorine implanted performs better off state current than control one even after elevate temperature, means that the drain side junction defect really reduce by fluorine passivated. Since device operated at the higher temperature, the off state current dominated by the thermionic emission, still perform lower in fluorine implanted sample then in control.
Moreover, the time evolution of threshold voltage shift of control and fluorine implanted sample with different temperature operation also be show in Figure 4-20 and 4-21. One could easily see that the curve of 75 oC of threshold voltage shift lower than curve 25 oC in both control and fluorine implanted samples, which we suggest the trapping sites in those samples would be de-trap in the higher temperature, which
de-trap could be found in the fluorine implanted sample, which we suggest that thinner interfacial oxide layer growth which due to fluorine like to grab the oxygen and prevent the oxygen reactive with Si[73]. The sketch map describe above as show in Figure 4-22.
Finally, Table 4-3 and 4-4 list the characteristics of control and fluorine implanted samples before and after PBTI stress. We could easily see that fluorine implanted sample still perform the better resist to PBTI.
-1 0 1 2 3 4 5 6 7 10-13
10-12 10-11 10-10 10-9 10-8 10-7 10-6 10-5 10-4 10-3 10-2
After Stress Before Stress I d (A)
Vg (V)
Control Fluorine
PBTI
W/L=100/10 μm, Vd=0.1 V
@Vg-Vth=5 V, Vd=Vs=0 V, 1000 s, 25 oC
Figure 4-12 Transfer characteristic with and without the fluorine implantation, before and after 1000 s PBTI at 25 oC with Vg-Vth=5 V, Vd=Vs=0 V
e
E
CE
VE
FPoly-Si
HfO
2Donor State
E C
E
VE
FPoly-Si
HfO 2
Figure 4-14 Defect pool model after states creation
-1 0 1 2 3 4 5 6 7
0 5 10 15 20 25
After Stress Before Stress
Control Fluorine
G m(μs)
Vg (V) PBTI
W/L=100/10 μm, V
d=0.1 V
@Vg-Vth=5 V, Vd=Vs=0 V, 1000 s, 25 oC
Figure 4-15 Transconductance with and without the fluorine implantation, before and after 1000 s PBTI at 25 oC with Vg-Vth=5 V, Vd=Vs=0 V
10 100 1000
Figure 4-16 Time evolution of threshold voltage shift under PBTI with and without fluorine implantation
Figure 4-17 Time evolution of sub-threshold swing degradation under PBTI with and
-1 0 1 2 3 4 5 6 7
Figure 4-18 Transfer characteristic without fluorine implantation, before and after 1000 s PBTI at 25 oC and 75 oC with Vg-Vth=5 V, Vd=Vs=0 V
Figure 4-19 Transfer characteristic with fluorine implantation, before and after 1000 s PBTI at 25 oC and 75 oC with Vg-Vth=5 V, Vd=Vs=0 V
10 100 1000
Figure 4-20 Time evolution of threshold voltage shift in control sample with temperature of 25 oC and 75 oC
Figure 4-21 Time evolution of threshold voltage shift in fluorine implanted sample
F&poly-Si HfO 2
poly-Si HfO 2
SiO 2 x x x x x x x x x x
Figure 4-22 Interfacial oxide growth with and without fluorine implanted poly-Si
Control
Before stress
Control After stress
Fluorine Before stress
Fluorine After stress Ion(10-5A)
9.13 3.37 7.55 3.27
Ioff(10-10A)
1.28 1.77 0.30 0.45
Ion/Ioff(105)
7 1 24 7
S.S.(mV/dec)
135 320 158 250
Vth(V)
0.91 2.99 1.38 3.14
Gm(μs)
17.0 6.95 15.8 10.2
Table 4-3 Characteristics of LTPS High-k TFTs with and without fluorine implantation before and after 1000 s PBTI at 25 oC with Vg-Vth=5 V, Vd=Vs=0 V
Control Fluorine
ΔIon(%)
-63 -55
ΔIoff(%)
37 49
ΔS.S.(%)
135 57
ΔVth(V)
2.08 1.76
ΔGm(%)
-59 -35
Table 4-4 Characteristics degraded percentage of LTPS High-k TFTs with and without fluorine implantation under 1000 s PBTI at 25 oC with Vg-Vth=5 V, Vd=Vs=0 V