基于SILVACO ATLAS的a-IGZO薄膜晶体管二维器件仿真（01）

最近因为肺炎的缘故，宅在家里不能出门，就翻了下一些资料，刚好研究方向是这个，就简单研究了下。参考资料主要如下：

1.《半导体工艺和器件仿真软件Silvaco TCAD实用教程》 唐龙谷 2014

2.《长安大学 半导体工艺与器件仿真指导书》 张林 2015

　　

引用本科时社团一姐的一句话：学习PS的精髓在于毁图。个人浅见，学习仿真的话还是要根据实例拆解分析比较快。重复，我是新手，只是个人浅见。

官网实例：https://www.silvaco.com/examples/tcad/section41/index.html

tftex10.in : Amorphous IGZO TFT Simulation：

# (c) Silvaco Inc., 2019
# This example demonstrates simulation of amorphous IGZO (indium galium
# zinc oxide) TFT.  Here we reproduce the results from:
#
# Fung, T., Chuang, C., Chen, C., Katsumi, A., Cottle, R., Townsend, M.,
# Kumomi, H., and Kanicki, J., "Two-dimensional numerical simulation of
# radio frequency sputter amorphous In-Ga-Zn-O thin-film transistors",
# J. Appl. Phys., V. 106, pp. 084511-1 through 084511-10.
#
# Comparisons with experiment are included.
#
# This first part of the input deck simulates Id-Vg.
#
go atlas 

mesh width=180 outf=tftex10_1.str master.out

x.m l=0 s=0.25
x.m l=40 s=0.25
y.m l=0 s=0.0005
y.m l=0.02 s=0.0005
y.m l=0.12 s=0.005
#（划分网格）
# The device is composed of a 20 nm layer of IGZO deposited
# 100 nm oxide on a n++ substrate that acts as the gate.
#
region num=1 material=igzo y.min=0    y.max=0.02
region num=2 material=sio2 y.min=0.02 y.max=0.12
#（定义材料）
elec num=1 name=gate bottom
elec num=2 name=source y.max=0.0 x.min=0.0  x.max=5.0
elec num=3 name=drain  y.max=0.0 x.min=35.0 x.max=40.0
#（定义电极）
# We define the gate as N.POLY.  This pins the gate workfunction
# to the conduction band edge of silicon.
#
contact num=1 n.poly
#
# We also define a workfunction for the source and drain that
# is very close to the conduction edge.  In the reference the
# authors observed that without a workfunction the results for
# ohmic boundaries were not significantly different than the
# Schottky model.
#
contact num=2 workf=4.33
contact num=3 workf=4.33
#（定义接触条件）
models fermi
#
# Key to the characterization of amorphous materials is the
# definition of the states within the band gap.
#
defects nta=1.55e20 ntd=1.55e20 wta=0.013 wtd=0.12 \
    nga=0.0 ngd=6.5e16 egd=2.9 wgd=0.1 \
    sigtae=1e-15 sigtah=1e-15 sigtde=1e-15 sigtdh=1e-15 \
    siggae=1e-15 siggah=1e-15 siggde=1e-15 siggdh=1e-15 \
    dfile=tftex10_don.dat afile=tftex10_acc.dat numa=128 numd=64
#（定义缺陷分布）
# From here we simply extract the Id-Vg characteristic
#
solve init
solve prev
#？？？
solve vdrain=0.1
save outf=tftex10_0.str
log outf=tftex10_1a.log
solve vgate=0 vstep=-0.1 vfinal=-5 name=gate

log off
#？？？
load inf=tftex10_0.str master
solve prev
log outf=tftex10_1b.log
solve vstep=0.2 vfinal=20.0 name=gate
#
# And we compare the simulation with experimental data reported in
# the reference.
#
tonyplot -overlay tftex10_1a.log tftex10_1b.log tftex10_1.dat -set tftex10_1.set
#tonyplot 的 overlay与set命令 ？？？
tonyplot -overlay tftex10_don.dat -set tftex10_don.set
#绘制施主态密度分布曲线
tonyplot -overlay tftex10_acc.dat -set tftex10_acc.set
#绘制受主态密度分布曲线

go atlas
#
# In the next part of the input deck we extract the Id-Vd family of
# curves.  The structure definition is exactly the same.
#
mesh width=180

x.m l=0 s=0.25
x.m l=40 s=0.25
y.m l=0 s=0.0005
y.m l=0.02 s=0.0005
y.m l=0.12 s=0.005

region num=1 material=igzo y.min=0    y.max=0.02
region num=2 material=sio2 y.min=0.02 y.max=0.12

elec num=1 name=gate bottom
elec num=2 name=source y.max=0.0 x.min=0.0  x.max=5.0
elec num=3 name=drain  y.max=0.0 x.min=35.0 x.max=40.0

contact num=1 n.poly
contact num=2 workf=4.33
contact num=3 workf=4.33

models fermi

defects nta=1.55e20 ntd=1.55e20 wta=0.013 wtd=0.12 \
    nga=0.0 ngd=6.5e16 egd=2.9 wgd=0.1 \
    sigtae=1e-15 sigtah=1e-15 sigtde=1e-15 sigtdh=1e-15 \
    siggae=1e-15 siggah=1e-15 siggde=1e-15 siggdh=1e-15 \
    dfile=don afile=acc numa=128 numd=64
#
# Here we calculate the start structures for each of the IdVd
# family by performing an initial gate ramp.
#
#（下面就有点懵了...）

solve solve vstep=0.2 vfinal=4.0 name=gate save outf=tftex10_2.str 
solve vstep=0.2 vfinal=8.0 name=gate save outf=tftex10_3.str 
solve vstep=0.2 vfinal=12.0 name=gate save outf=tftex10_4.str 
solve vstep=0.2 vfinal=16.0 name=gate save outf=tftex10_5.str 
solve vstep=0.2 vfinal=20.0 name=gate save outf=tftex10_6.str 

load inf=tftex10_2.str master log outf=tftex10_2.log 
solve vdrain=0.0 vstep=0.5 vfinal=20.0 name=drain 

load inf=tftex10_3.str master log outf=tftex10_3.log 
solve vdrain=0.0 vstep=0.5 vfinal=20.0 name=drain 

load inf=tftex10_4.str master log outf=tftex10_4.log 
solve vdrain=0.0 vstep=0.5 vfinal=20.0 name=drain 

load inf=tftex10_5.str master log outf=tftex10_5.log 
solve vdrain=0.0 vstep=0.5 vfinal=20.0 name=drain 

load inf=tftex10_6.str master log outf=tftex10_6.log 
solve vdrain=0.0 vstep=0.5 vfinal=20.0 name=drain 

tonyplot -overlay tftex10_2.log tftex10_2.dat tftex10_3.log tftex10_3.dat tftex10_4.log tftex10_4.dat tftex10_5.log tftex10_5.dat tftex10_6.log tftex10_6.dat -set tftex10_2.set 

quit

官网给出的四张输出图像如下：

自己装的ATLAS还有点问题，这几天先折腾下在跑跑看。

不过毕竟是引用了别人的文献中的数据，与实验室的材料特性还有较大差距，估计要手调了。

总体来说，ATLAS的步骤主要为：

建立网格-定义材料-定义电极以及接触-定义缺陷分布或掺杂-引入模型以及求解方法-求解-Tonyplot输出

要解决的难点主要在于：

1.材料特性微调；

2.缺陷分布调整；

3.理解输出语句，并合理运用；

武汉加油！