Hot Carrier Stress

First reliability test method which we use is hot carrier stress, with stress condition by apply gate voltage 5 V over than threshold voltage, fixed drain voltage to 5 V, and connect source electrode to ground for 1000 s. Figure 4-4 show the transfer characteristic of control sample and fluorine implanted sample. It could be easily noticed the off state current segment in both the control and fluorine implanted sample been disappear after hot carrier stress. We suggest the damage does not happen in the channel, but in the gate dielectric. In order to examine our ideal, forward and reverse drain current were measured just like the method use in the two-bit operation memory. Figure 4-5 show the forward measurement setup, which drain voltage fixed to 0.1 V and source connect to ground. For the reverse measurement, source voltage fixed to 0.1 V and drain connect to ground. Figure 4-6 show the result that reverse measurement perform the off state current have the same order as the initial reverse measurement. It means that damage happened in the drain side gate dielectric, so that offer the frank-pool tunneling path for the current comes from gate, which the path marked by three start sign as show in Figure 4-5. As the gate voltage larger than drain voltage in Figure 4-4, the off state electron current will conduct upward, tile the inversion layer electron flow into drain larger than the gate current, and then curve again appear in the transfer characteristic. Figure 4-7 shows the transfer characteristic again but different with the current been take for absolute calculated, hence, the gate leakage current appeared as show in hallow sign. The fluorine implanted sample has the lower gate leakage current suggest that the damage

We speculate that the fluorine will diffuse into HfO₂ under the gate dielectric densify process and passivate the defect in the HfO2. Although the hot electron inject into gate dielectric, the fluorine bond inside the HfO₂ will not easy be broken. Thus the defect created damages can be less than the control one, so that have lower gate leakage current. And the phenomenon of fluorine passivate the defects in HfO₂ also can be see in the threshold voltage shift which is fewer than control one. Figure 4-8 show the gate dielectric capacitance of control and fluorine implanted sample, and the fluorine implanted one have lower gate capacitance than control due to the fluorine indeed have diffuse into gate dielectric as report in [68]. And this could also be show at the initial Gm maximum, which has lower gate capacitance by fluorine as show in Figure 4-9. And after hot carrier stress, the fluorine implanted sample performs the higher Gm maximum than the control one, thanks to the strong Si-F has work in passivate the grain boundary tail states in the drain side as discussed in the Figure 4-3.

Now let us take a look at Figure 4-10 and 4-11 which show the time evolutions of sub-threshold swing degradation and threshold voltage shift, respectively. The time evolution sub-threshold swing degradation of fluorine implanted sample exhibit better than control, which contribute by strong Si-F bond passivate the interface deep states in the drain side. And Figure 4-11 also show the fluorine implanted sample has lowered threshold voltage shift than control, which due to the fluorine passivate the defect states in HfO2, as discuss yet. Since sub-threshold swing region happened earlier than the threshold voltage, the threshold voltage shift should have the same trend with sub-threshold swing degradation, but it is not. It may be due to different mechanisms between these two parameters. In the past studies of prediction of threshold voltage shift, if it exhibit logarithm time dependent means that most of shift comes from bulk oxide charges, but for the power time dependent means that states

control and fluorine implanted sample both have the same logarithm time dependent, which suggest that our HfO2 perform the large amount of electron trapping even though fluorine have been passivate some of these defects. Table 4-1 and 4-2 list the characteristics before and after of control and fluorine implanted sample before and after hot carrier stress. We could easily see the fluorine implanted sample perform the better resist to hot carrier damage.

-1 0 1 2 3 4 5 6 7

Figure 4-4 Transfer characteristic with and without fluorine implantation, before and after 1000 s hot carrier stress at 25 ^oC with Vg-Vth=5 V, Vd=5 V, Vs=0 V

-1 0 1 2 3 4 5 6 7

Figure 4-6 Forward and reverse transfer characteristic with fluorine implantation, before and after 1000 s hot carrier stress at 25 ^oC with Vg-Vth=5 V, Vd=5 V, Vs=0 V,

Figure 4-7 Transfer characteristic with and without the fluorine implantation, before and after 1000 s hot carrier stress at 25 ^oC with Vg-Vth=5 V, Vd=5 V, Vs=0 V, where

solid line indicate the positive current, hallow line indicate the negative current

15 16 17 18

Control

Capacitance (pF)

Fluorine

W/L=100/100 μm

Figure 4-8 Gate dielectric capacitance with and without fluorine implantation

-1 0 1 2 3 4 5 6 7

0 5 10 15 20 25

Control Fluorine

After Stress Before Stress

G m(μs)

V_g (V) Hot Carrier Stress W/L= 100/10 μm, V

d=0.1 V

@ V_g-V_th=5 V, V_d=5 V, V_s=0 V, 1000 s, 25 ^oC

Figure 4-9 Transconductance with and without the fluorine implantation, before and after 1000 s hot carrier stress at 25 ^oC with Vg-Vth=5 V, Vd=5 V, Vs=0 V

10 100 1000 150

200

250 Subthreshold Swing under HCS

@V_g-V_th=5 V, V_d=5 V, V_s=0 V, 1000 s, 25 ^oC

S.S (mV/dec)

T_stress (s)

Control Fluorine

Figure 4-10 Time evolution of sub-threshold swing degradation under hot carrier stress with and without fluorine implantation

10 100 1000

0.0 0.5 1.0 1.5 2.0

Control Fluorine

ΔV th (V)

T_stress (s) Threshold Voltage Shift under HCS

@Vg-V

th=5 V, V

d=5 V, V

s=0 V, 1000 s, 25 ^oC

Figure 4-11 Time evolution of threshold voltage shift under hot carrier stress with and without fluorine implantation

Control Before stress

Control After stress

Fluorine Before stress

Fluorine After stress I_on(10^-5A)

8.51 6.37 7.80 6.61

I_off(10^-10A)

1.09 0.98 0.27 0.22

I_on/I_off(10⁵)

9 6 28 29

S.S.(mV/dec)

132 225 154 202

V_th(V)

0.91 2.34 1.42 2.18

Gm(μs)

16.4 10.3 15.1 12.3

Table 4-1 Characteristics of LTPS High-k TFTs with and without fluorine implantation before and after 1000 s hot carrier stress at 25 ^oC with V_g-V_th=5 V, V_d=5

V, V_s=0 V

Control Fluorine

ΔIon(%)

-27 -15

ΔIoff(%)

-10 -18

ΔS.S.(%)

70 29

ΔV_th(V)

1.43 0.76

ΔGm(%)

-36 -18

Table 4-2 Characteristics degraded percentage of LTPS High-k TFTs with and without fluorine implantation under 1000 s hot carrier stress at 25 ^oC with Vg-Vth=5 V, Vd=5 V,

Vs=0 V