Lecture 5
PN Junction and MOS
Electrostatics(II)
PN JUNCTION IN THERMAL EQUILIBRIUM
Outline
1.
2.
3.
4.
Introduction
Electrostatics of pn junction in thermal
equilibrium
The depletion approximation
Contact potentials
Reading Assignment:
Howe and Sodini, Chapter 3, Sections 3.3­3.6
6.012 Spring 2009
Lecture 5
1
1. Introduction
• pn junction
– p­region and n­region in intimate contact
Why is the p­n junction worth studying?
It is present in virtually every semiconductor device!
Example: CMOS cross­section
p-MOSFET
n+
p+
n-MOSFET
p+
n
n+
n+
p+
p
Figure by MIT OpenCourseWare.
Understanding the pn junction is essential to
understanding transistor operation
6.012 Spring 2009
Lecture 5
2
2. Electrostatics of p­n junction in equilibrium
Focus on intrinsic region:
Doping distribution of an abrupt p­n junction
6.012 Spring 2009
Lecture 5
3
What is the carrier concentration
distribution in thermal equilibrium?
First think of the two sides separately:
Now bring the two sides together.
What happens?
6.012 Spring 2009
Lecture 5
4
Resulting carrier concentration profile in
thermal equilibrium:
• Far away from the metallurgical junction: nothing
happens
– Two quasi­neutral regions
• Around the metallurgical junction: diffusion of
carriers must counter­balance drift
– Space­charge region
6.012 Spring 2009
Lecture 5
5
On a linear scale:
Thermal equilibrium: balance between drift and diffusion
J n (x) = J ndrift (x) + J ndiff (x) = 0
J p (x) = J pdrift (x) + J pdiff (x) = 0
We can divide semiconductor into three regions
• Two quasi­neutral n­ and p­regions (QNR’s)
• One space­charge region (SCR)
Now, we want to know no(x), po(x), ρ(x), E(x) and φ(x).
We need to solve Poisson’s equation using a simple
but powerful approximation
6.012 Spring 2009
Lecture 5
6
3. The Depletion Approximation
• Assume the QNR’s are perfectly charge neutral
• Assume the SCR is depleted of carriers
– depletion region
• Transition between SCR and QNR’s sharp at
– ­xpo and xno (must calculate where to place these)
x < −x po ;
ni2
po (x) = N a , no (x) =
Na
−x po < x < 0;
po (x), no (x) << N a
0 < x < xno ;
no (x),
x > xno ;
6.012 Spring 2009
po (x) << N d
no (x) = N d ,
ni2
po (x) =
Nd
Lecture 5
7
Space Charge Density
ρ(x) = 0;
= − qN a ;
−x po < x < 0
=
0 < x < xno
x > x no
= 0;
6.012 Spring 2009
x < −x po
qNd ;
Lecture 5
8
Electric Field
Integrate Poisson’s equation
E(x2 ) − E(x1 ) =
1
x2
ρ(x) dx
∫
εs x
1
x < − x po ;
− x po < x < 0;
E(x) = 0
E(x) − E( −x po ) =
x
1
εs − x∫
po
−qN a dx ′
 qN a  x
− qN a
= −
x
=
x + x po
ε
ε

 −x
s
s
po
(
qN d
0 < x < xno ;
E(x) =
x > xno ;
E(x) = 0
6.012 Spring 2009
εs
)
(x − xno )
Lecture 5
9
Electrostatic Potential
(with φ=0 @ no=po=ni)
φ=
kT
n
• ln o
q
ni
φ=−
kT
p
• ln o
q
ni
In QNRs, no and po are known ⇒ can determine φ
in p­QNR: po=Na ⇒
φp = −
N
kT
• ln a
q
ni
kT
Nd
φ
=
•
ln
in n­QNR: no=Nd ⇒ n q
ni
Built­in potential:
Nd Na
kT
φ B = φn − φ p =
• ln
q
ni2
This expression is always correct in TE!
We did not use depletion approximation.
6.012 Spring 2009
Lecture 5
10
To obtain φ(x) in between, integrate E(x)
x2
φ(x2 ) − φ(x1) = − ∫ E(x′)dx′
x1
x < −x po ;
−x po < x < 0;
φ(x) = φp
x
φ(x) − φ(−x po ) = −
∫
−
qNa
−x po
(
εs
(x ′ + x po )dx ′
)
2
qNa
x + x po
2ε s
2
qN
φ(x) = φ p + a x + x po
2ε s
=
(
0 < x < xno ;
φ(x) = φn −
)
qNd
2
x − xno )
(
2ε s
φ (x) = φn
Almost done ….
x > xno ;
6.012 Spring 2009
Lecture 5
11
Still do not know xno and xpo ⇒ need two more equations
1. Require overall charge neutrality:
qN a x po = qN d x no
2. Require φ(x) to be continuous at x=0;
φp +
qN a 2
qN d 2
x po = φ n −
x no
2εε s
2εε
s
Two equations with two unknowns — obtain solution:
2εε s φ B N a
x no =
q(N a + N d )N d
2εε s φ B N d
x po =
q(N a + N d )N a
Now problem is completely solved!
6.012 Spring 2009
Lecture 5
12
Solution Summary
Space Charge Density
Electrostatic Field
Electrostatic Potential
6.012 Spring 2009
Lecture 5
13
Other results:
Width of the space charge region:
x do = x po + x no =
2εε s φ B (N a + N d )
qN a N d
Field at the metallurgical junction:
2qφ
φB Na Nd
Eo =
ε s (N a + N d )
6.012 Spring 2009
Lecture 5
14
Three Special Cases
• Symmetric junction: Na = Nd
x po = x no
• Asymmetric junction: Na > Nd
x po < x
no
• Strongly asymmetric junction
– p+n junction: Na >> Nd
2 εs φB
x po << xno ≈ x do ≈
qNd
Eo ≈
2qφ B Nd
εs
The lightly­doped side controls the electrostatics of
the pn junction
6.012 Spring 2009
Lecture 5
15
4. Contact Potential
Potential distribution in thermal equilibrium so far:
Question 1: If I apply a voltmeter across the pn junction
diode, do I measure φB?
yes
no
it depends
Question 2: If I short terminals of pn junction diode, does
current flow on the outside circuit?
yes
6.012 Spring 2009
no
sometimes
Lecture 5
16
We are missing contact potential at the metal­
semiconductor contacts:
Metal­semiconductor contacts: junction of dissimilar
materials
⇒ built­in potentials at contacts φmn and φmp.
Potential difference across structure must be zero
⇒ Cannot measure φB.
φ B = φmn + φmp
6.012 Spring 2009
Lecture 5
17
5. PN Junction-Reverse Bias Assume: No Current Flows
-
-
t
+
ZDsOA
+
O
v, (<OW
ohmic contact to
%wn
p side
-w,~///////////A
-
-PI
-
++j
X
-
%I
qmn
+
-
(a>
t+'"'
Same Analysis applies:
(b)
Substitute
6.012Spring 2009
Lecture B
ohmic contact
to n side
What did we learn today?
Summary of Key Concepts
• Electrostatics of pn junction in equilibrium
– A space­charge region surrounded by two
quasi­neutral regions formed.
• To first order, carrier concentrations in space­charge
region are much smaller than the doping level
– ⇒ can use Depletion Approximation
• From contact to contact, there is no potential build­
up across the pn junction diode
– Contact potential(s).
6.012 Spring 2009
Lecture 5
19
MIT OpenCourseWare
http://ocw.mit.edu
6.012 Microelectronic Devices and Circuits
Spring 2009
For information about citing these materials or our Terms of Use, visit: http://ocw.mit.edu/terms.