Resolution - Microfabrication Technology - Solved Exam, Exams of Materials science

Main points of this past exam are: Resolution, Provide Units, Expect, Silicon Dioxide Layer, Metal Layer, Better or Worse, Faster or Slower

Typology: Exams

2012/2013

Uploaded on 03/22/2013

farhan
farhan 🇮🇳

3

(1)

55 documents

1 / 2

Toggle sidebar

This page cannot be seen from the preview

Don't miss anything!

bg1
1
EE143 Midterm Exam #1 Solutions Sp2001
Problem1
(i) The pad oxide is very thin (~100 Å). It can easily be prepared by thermal oxidation.
The silicon nitride is prepared by chemical vapor deposition (CVD)
(ii ) Peak concentration of implant profile = 0.4 φ / Rp. For the shallow channel stop implant, Rp ~
0.1 µm , peak concentration ~ 4E17/cm2
(iii) The gate oxide oxidation step is performed before poly-Si deposition.
(iv) We prefer As for shallow junctions because As has a slower diffusion constant than Phosphorus
[Other secondary reasons: As has a higher solid solubility than P.Therefore, lower sheet resistance and
contact resistance. Also, for the same implant energy , the projected range of As is smaller]
(v) We cannot use thermal oxidation step to replace CVD oxide. The oxide layer is to electrically isolate
the Al metallization from the poly-Si gate and substrate and require sufficient thickness. If too much poly-
Si is consumed by using thermal oxidation, the gate dimensions will be significantly reduced.
(vi) We cannot replace the poly-Si by Al as gate material in this process flow. There is a high temperature
annealing step (>900 oC ) after source/drain implantation. This step will melt the Al.
Problem2
( i) xox2 + Axox = B(t + τ)
For 900°C, O2 : B = 5600 (Å )2/minute, B
A = 2 Å /min A = 2800 Å
For xox = 1000Å, τ = 0 t = 680 min.
( ii) Let xi be the oxide grown during the steam oxidation step such that the final oxide thickness
xox = 5000Å after dry oxidation
(5000)2 + 2800 (5000) = 5600 (680) + xi2 + 2800 xi xi = 4700Å.
( iii) Using a high pressure for oxidation increases the oxidant concentration on the oxide surface ( Henry’s
Law). The oxidation rate will increase according to the Grove Model. Since it takes less time to grow the
0.1 µm gate oxide, the additional oxide grown on the field oxide is less. We need a larger xi initially to
make the final field oxide thickness = 0.5 µm .
( iv) (111) Si surface has more reaction sites than (100) Si. This will increase the surface reaction term (ks).
The oxidation rate will increase according to the Grove Model. Since it takes less time to grow the 0.1 µm
gate oxide, the additional oxide grown on the field oxide is less. We need a larger xi initially to make the
final field oxide thickness = 0.5 µm .
Problem 3
(a) (0.5 - x) 2.17 = 0.5 x = 0.27 µm
(b) (i) Rp = 0.5 µm Rp = 0.09 µm
The boron will need a kinetic energy of 186
keV
(A) For B +ions, we need 186kV accelerating
voltage
(B) For BF2+ions, we need 186 × 11+19+19
11 =
828 kV accelerating voltage
(ii) Np = 1013
2πRp = 4.4×1017/cm3
4.4×1017e- (xj - Rp)2/2Rp = 10 15
x=0
C(x)
x
0.5
µµ
m
x
j1
1E15/cm3
pf2

Partial preview of the text

Download Resolution - Microfabrication Technology - Solved Exam and more Exams Materials science in PDF only on Docsity!

EE143 Midterm Exam #1 Solutions Sp Problem (i) The pad oxide is very thin (~100 Å). It can easily be prepared by thermal oxidation. The silicon nitride is prepared by chemical vapor deposition (CVD) (ii ) Peak concentration of implant profile = 0.4 φ / ∆Rp. For the shallow channel stop implant, ∆Rp ~ 0.1 μm , peak concentration ~ 4E17/cm^2 (iii) The gate oxide oxidation step is performed before poly-Si deposition. (iv) We prefer As for shallow junctions because As has a slower diffusion constant than Phosphorus [Other secondary reasons: As has a higher solid solubility than P.Therefore, lower sheet resistance and contact resistance. Also, for the same implant energy , the projected range of As is smaller] (v) We cannot use thermal oxidation step to replace CVD oxide. The oxide layer is to electrically isolate the Al metallization from the poly-Si gate and substrate and require sufficient thickness. If too much poly- Si is consumed by using thermal oxidation, the gate dimensions will be significantly reduced. (vi) We cannot replace the poly-Si by Al as gate material in this process flow. There is a high temperature annealing step (>900 oC ) after source/drain implantation. This step will melt the Al.

Problem

( i) xox^2 + Axox = B(t + τ)

For 900°C, O 2 : B = 5600 (Å )^2 /minute,

B

A = 2 Å /min^ ∴^ A = 2800 Å ∴ For xox = 1000Å, τ = 0 ⇒ t = 680 min.

( ii) Let xi be the oxide grown during the steam oxidation step such that the final oxide thickness

xox = 5000Å after dry oxidation

∴ (5000)^2 + 2800 • (5000) = 5600 (680) + xi^2 + 2800 xi ⇒ xi = 4700Å. ( iii) Using a high pressure for oxidation increases the oxidant concentration on the oxide surface ( Henry’s Law). The oxidation rate will increase according to the Grove Model. Since it takes less time to grow the 0.1 μm gate oxide, the additional oxide grown on the field oxide is less. We need a larger xi initially to make the final field oxide thickness = 0.5 μm.

( iv) (111) Si surface has more reaction sites than (100) Si. This will increase the surface reaction term (ks). The oxidation rate will increase according to the Grove Model. Since it takes less time to grow the 0.1 μm gate oxide, the additional oxide grown on the field oxide is less. We need a larger xi initially to make the final field oxide thickness = 0.5 μm.

Problem 3

(a) (0.5 - x) • 2.17 = 0.5 ⇒ x = 0.27 μm (b) (i) Rp = 0.5 μm ∆Rp = 0.09 μm

The boron will need a kinetic energy of 186 keV (A) For B +ions, we need 186kV accelerating voltage

(B) For BF 2 +ions, we need 186 ×

828 kV accelerating voltage

(ii) Np =

2 π∆Rp

= 4.4× 1017 /cm^3

∴ 4.4× 1017 e- (xj - Rp)

2 /2∆R

p = 10^15

x=

C(x)

x

0.5 μμ (^) m

x (^) j

1E15/cm

∴ (xj - Rp)^2 = 2(0.09)^2 • ln(

4.4× 1017

) = 0.099 μm^2

∴ xj = Rp ± 0.314 μm

For x j1; x j1 = Rp + 0.314 = 0.814 μm

(iii) No=4.4 × 1017 /cm^3 , NB=10^15 /cm^3 , half-gaussian profile of p into n-type substrate.

The Irvin’s curves gives RS xj ≈ 2000 ohm-μm.

Since xji is only 0.314um underneath the SiO2 interface , xj= 0.314 μμ m , therefore RS ≈ 6400 ohm/square

(d) For some particular angles of incidence along crystallographic axis or planes, a fraction of the incident ions can transverse the crystalline substrate with little nuclear scattering. This will give a deeper implantation profile than that of an amorphous substrate. Although a tilt-and-rotate geometry can minimize the primary channeling effect, we can still have a small fraction of the scattered ions bouncing into various axis or planes, giving a “channeling tail” of the doping profile.

Problem 4

(i) We just consider the diffusion of phosphorus under the poly region

1021 erfc[

xj 2 Dt

] = 10^15 ⇒

xj 2 Dt

= erfc-1^ (10-6) = 3.

2 Dt = 2 3 × 10 -13^ × 3600 = 6.6× 10 -5^ cm ∴ xj = 6.6× 10 -5cm × 3.45 = 2.26 μm

(ii)

Q

πDt

e-xj (^2) /4Dt = 10^15 /cm^3

Dt for As = 3× 10 -14^ × 3600 = 1.08× 10 -10^ cm^2 at 1100°C

Q

πDt

π • 1.08× 10 -^

= 5.4× 1019 /cm^3

∴ xj^2 = 4Dt ln(

5.4× 1019

) = 4.7× 10 -9^ cm^2

∴ xj = 0.69 μm

(iii) 2 Dt of As = 2.1× 10 -5^ cm

At a depth of 2.26 μm below surface y =

x 2 Dt

Since both the erfc(y) and exp(-y^2 ) functions decrease rapidly with y, the contribution due to As at depths below 2.26 μm is << NB = 10^15. Therefore, the As profile contribution is negligible.

( iv) Electric field enhancement and Charge Point Defects enhancement

(v) Ion implanation will create a large concentration of vacancies and interstitials. They should recombine to form perfect Si again after subsequent annealing. However , during the initial stage of annealing ( ~ seconds) , the dopant diffusion will experience a very high diffusion constant already due to point defect enhancement. The depth junction will increase rapidly. After the vacancies and interstitials recombine with sufficient annealing time, this extra enhancement effect will disappear. Unfortunately, the junction depth may be too deep already if we want to form ultra-shallow junctions.