Chapter 8 Bipolar Junction Transistors: Question: What Is The Meaning of "Bipolar" ?

Chapter 8 Bipolar Junction Transistors
• Since 1970, the high density and low-power advantage of

the MOS technology steadily eroded the BJT’s early dominance.
• BJTs are still preferred in some high-frequency and analog

applications because of their high speed and high power output.
Question: What is the meaning of “bipolar” ?
Modern Semiconductor Devices for Integrated Circuits (C. Hu) Slide 8-1
-
Efn Ec
Efp -
VB E
8.1 Introduction
(b) to the EBJT
v VCB
NPN BJT:
B
E C
N+ P N IC VB E
Emitter Base Collector
(c)
VBE VCB
VCB
0
IC is an exponential
function of forward
VBE and independent
of reverse VCB.
Common-Emitter Configuration
Question: Why is IB often preferred as a parameter over VBE?

8.2 Collector Current depletion layers
N+ P N
emitter base collector
d 2 n n x
2
 2 0 WB
dx LB
B : base recombination lifetime
LB   B DB DB : base minority carrier (electron)
diffusion constant
Boundary conditions :
n(0)  nB 0 (e qVBE / kT  1)

n (WB )  nB 0 (e qVBC / kT
 1)  nB 0  0
 WB  x  8.2 Collector Current
sinh  
n( x )  nB 0 (e qVBE / kT  1)  L B  dn
sinh WB / LB  I C  AE qDB
n
------------- dx
n( x )n / n0 (0) 2
ni qV BE  kTn 2

Nn 
-------e
( x )  – 1 qVBE / kT
iB
(e  1) DB niB2 qVBE / kT
1 B
NB  AE q (e  1)
WB N B
I C  I S (e qVBE / kT  1)
It can be shown
qni2 qVBE / kT
I C  AE (e  1)
0 x/ B
x/W 1 GB
WB
n( x )  n(0)(1  x / WB ) ni2 p
GB   2 dx
niB2 qVBE / kT n DB
0 iB
 (e  1)(1  x / WB )
NB GB (s·cm4) is the base Gummel number
8.2.1 High Level Injection Effect
•At low-level injection, 10-2
inverse slope is 60 mV/decade IkF
10-4
•High-level injection effect :
IC (A)
10-6
At large VBE, n  p  N B 60 mV/decade
10-8
n  p   n  p
10-10
2 q ( EFn  EFp ) / kT
np  ni e  ni eqVBE / kT
2
10-12
0 0.2 0.4 0.6 0.8 1.0
 n  p  ni e qVBE / 2 kT VBE
GB  p  ni e qVBE / 2 kT
I C  ni e qVBE / 2 kT
When p > NB , inverse slope is 120mV/decade.

8.3 Base Current
Some holes are injected from the P-type base into the N+ emitter.
The holes are provided by the base current, IB .
(a) contact emitter base collector contact
electron flow
I
–
E IC
hole flow
+
I
B pE' nB'
(b)
WE WB
8.3 Base Current
(a) contact emitter base collector contact
electron flow
–
I IC
E
hole flow
+
I
B
qni2 qVBE / kT
I B  AE (e  1) For a uniform emitter,
GE
DE niE2 qVBE / kT
I B  AE q  1)
WE 2
n n (e
GE   dx i
2 WE N E
0
n DE iE
Is a large IB desirable? Why?
8.4 Current Gain
Common-emitter current gain, F : IC
F 
Common-base current gain: IB
IC   F I E
IC IC IC / I B F
F    
I E I B  IC 1  IC / I B 1   F
F
It can be shown that  F 
1F
GE DBWE N E niB2
F  
GB DEWB N B niE2
How can F be maximized?
EXAMPLE: Current Gain
A BJT has IC = 1 mA and IB = 10 mA. What are IE, F and F?
Solution:
I E  I C  I B  1 mA  10 μA  1.01 mA
 F  I C / I B  1 mA / 10 μA  100
 F  I C / I E  1 mA / 1.01 mA  0.9901
We can confirm
F F
F  and  F 
1 F 1F
8.4.1 Emitter Bandgap Narrowing
N E niB2 To raise F, NE is typically very large.

 Unfortunately, large NE makes niE2  ni2
N B niE2
(heavy doping effect).
 Eg / kT
ni2  NC NV e Since ni is related to Eg , this effect is
also known as band-gap
narrowing.
niE2  ni2e
EgE / kT
EgE is negligible for NE < 1018 cm-3,
is 50 meV at 1019cm-3, 95 meV at 1020cm-3,
and 140 meV at 1021 cm-3.
Emitter bandgap narrowing makes it difficult to raise F by
doping the emitter very heavily.
8.4.2 Narrow-Bandgap Base and Heterojuncion BJT
N E niB2 To further elevate F , we can raise niB by


N B niE2 using an epitaxial Si1-hGeh base.
With h = 0.2, EgB is reduced by 0.1eV and niE2 by 30x.
EXAMPLE: Emitter Bandgap Narrowing and SiGe Base
Assume DB = 3DE , WE = 3WB , NB = 1018 cm-3, and niB2 = ni2. What is
F for (a) NE = 1019 cm-3, (b) NE = 1020 cm-3, and (c) NE = 1020 cm-3
and a SiGe base with EgB = 60 meV ?
(a) At NE = 1019 cm-3, EgE  50 meV,

EgE / kT
niE2  ni2e  ni2e50 meV / 26 meV  ni2e1.92  6.8ni2
DBWE N E ni2 9 1019  ni2
F    18  13
DEWB N B niE 10  6.8ni
2 2
(b) At NE = 1020 cm-3, EgE  95 meV

niE2  38ni2  F  24
(c) niB2  ni2eEgB / kT  ni2e60 meV / 26 meV  10ni2  F  237
8.4.3 Poly-Silicon Emitter
A high-performance BJT typically has a layer of As-doped N+
poly-silicon film in the emitter.
F is larger due to the large WE , mostly made of the N+ poly-
silicon. (A deep diffused emitter junction tends to cause emitter-
collector shorts.)
N+-poly-Si
emitter
SiO2
P-base
N-collector
8.4.4 Gummel Plot and F Fall-off at High and Low Ic
F
SCR BE current
From top to bottom:
VBC = 2V, 1V, 0V
Why does one want to operate BJTs at low IC and high IC?
Why is F a function of VBC in the right figure?
Hint: See Sec. 8.5 and Sec. 8.9.
8.5 Base-Width Modulation by Collector Voltage
Output resistance :
1
 I C  V
r0     A
 VCE  IC
IC IB3
Large VA (large ro )
IB2
is desirable for a
VA : Early Voltage IB1 large voltage gain
0 VCE
VA
V BE
N+ P N
VCE
emitter base collector
WB3
WB2
WB1
} VCE 1 < VCE 2 <VCE 3
n'
x
How can we reduce the base-width modulation effect?
VBE
N+ P N
The base-width modulation emitter
VCE
base collector
effect is reduced if we
WB3
(A) Increase the base width,

(B) Increase the base doping n'
WB2
WB1 } VCE1< VCE2<VCE3
concentration, NB , or
(C) Decrease the collector doping
concentration, NC .
x
Which of the above is the most acceptable action?
8.6 Ebers-Moll Model
IC IB
saturation active region
region
VCE
0
The Ebers-Moll model describes both the active

and the saturation regions of BJT operation.
IC is driven by two two forces, VBE and VBC .
VB E VB C
When only VBE is present :
IB
I C  I S (e qVBE / kT  1)
IS
IB  (e qVBE / kT  1) E B C
F IC
Now reverse the roles of emitter and collector.

When only VBC is present :
I E  I S (e qVBC / kT  1) R : reverse current gain
IS F : forward current gain
IB  (e qVBC / kT  1)
R
1
I C   I E  I B   I S (1  )(e qVBC / kT  1)
R
In general, both VBE and VBC are present :
1
I C  I S (e qVBE / kT
 1)  I S (1  )(e qVBC / kT  1)
R
IS IS
IB  (e qVBE / kT  1)  (e qVBC / kT  1)
F F
In saturation, the BC junction becomes forward-biased, too.
VBC causes a lot of holes to be injected

into the collector. This uses up much
of IB. As a result, IC drops.
VCE (V)
8.7 Transit Time and Charge Storage
When the BE junction is forward-biased, excess holes are stored

in the emitter, the base, and even in the depletion layers.
QF is all the stored excess hole charge
QF
F 
IC
F is difficult to be predicted accurately but can be measured.

F determines the high-frequency limit of BJT operation.
8.7.1 Base Charge Storage and Base Transit Time
Let’s analyze the excess hole charge and transit time in

the base only.
QFB  qAE n(0)WB / 2
QFB WB2
  FB 
IC 2 DB
np' = n'p
niB22iB qV  kT
n
 
( 0) 
n 0 = ------- e 
( e qVBEBE/ kT
 1–)1 
N BB
x
0 WB
EXAMPLE: Base Transit Time
What is FB if WB = 70 nm and DB = 10 cm2/s?
Answer:
WB2 (7 106 cm) 2

 FB    2.5  10 12
s  2.5 ps
2 DB 2 10 cm /s
2
2.5 ps is a very short time. Since light speed is

3108 m/s, light travels only 1.5 mm in 5 ps.
8.7.2 Drift Transistor–Built-in Base Field
The base transit time can be reduced by building into the base
a drift field that aids the flow of electrons. Two methods:
• Fixed EgB , NB decreases from emitter end to collector end.
E - B C
Ec Ef
Ev
• Fixed NB , EgB decreases from emitter end to collector end.
E - B C
1 dEc
Ec Ef E
q dx
Ev
8.7.3 Emitter-to-Collector Transit Time and Kirk Effect
• To reduce the total transit time, emitter and depletion layers must be thin, too.
• Kirk effect or base widening: At high IC the base widens into the collector. Wider
base means larger F .
Top to bottom :
VCE = 0.5V, 0.8V,
1.5V, 3V.
Base Widening at Large Ic
E
N N+
base collector collector
I C  AE qnvsat
  qN C  qn x
IC
 qN C  base
width
depletion
layer
AE vsat
E
N N+
dE base collector collector
  /es
dx
“base depletion
width” layer
8.8 Small-Signal Model
B C
I C  I S e qVBE / kT +
C vbe r gm vbe
Transconductance: 
E E
dI C d
gm   ( I S e qVBE / kT )
dVBE dVBE
q
 I S e qVBE / kT  I C /( kT / q)
kT
g m  I C /( kT / q) At 300 K, for example, gm=IC /26mV.
8.8 Small-Signal Model
B C
1 dI B 1 dI C g
   m +
r dVBE  F dVBE  F
C vbe r gm vbe

r   F / g m
E E
dQF d
C    F IC   F gm
dVBE dVBE
This is the charge-storage capacitance, better known as the

diffusion capacitance.
Add the depletion-layer capacitance, CdBE :
C   F g m  CdBE
EXAMPLE: Small-Signal Model Parameters
A BJT is biased at IC = 1 mA and VCE = 3 V. F=90, F=5 ps,

and T = 300 K. Find (a) gm , (b) r , (c) C .
Solution:
1 mA mA
(a) g m  I C /( kT / q)   39  39 mS (milli siemens)
26 mV V
90
(b) r   F / g m   2.3 kΩ
39 mS
(c) C   F g m  5 1012  0.039  1.9 1014 F  19 fF (femto farad)
Once the model parameters are determined, one can analyze
circuits with arbitrary source and load impedances.
The parameters are routinely

determined through comprehensive
measurement of the BJT AC
and DC characteristics.
8.9 Cutoff Frequency
B C
+
Signal C Load
source  vbe r gm vbe 1
  1 at fT 
2 ( F  CdBE kT / qI C )
-
E E
The load is a short circuit. The signal source is a current source,

ib , at frequency, f. At what frequency does the current gain
 ( ic / ib ) fall to unity?
ib ib
vbe   , C   F g m  CdBE
input admittance 1 / r  jC
ic  g m vbe
ic gm 1
 ( )   
ib 1 / r  jC 1 /  F  j F  jCdBE kT / qI C
8.9 Cutoff Frequency
fT = 1/2(F + CdBEkT/qIC)
fT is commonly used to compare the speed of transistors.
• Why does fT increase with increasing IC?
• Why does fT fall at high IC?
BJT Structure for Minimum Parasitics and High Speed
• Poly-Si emitter
• Thin base
• Self-aligned poly-Si base contact
• Narrow emitter opening
• Lightly-doped collector
• Heavily-doped epitaxial subcollector
• Shallow trench and deep trench for electrical isolation
8.10 Charge Control Model
QF
•For the DC condition, IC(t) = QF(t)/F I B  IC /  F 
 F F
•In order to sustain an excess hole charge in the transistor,
holes must be supplied through IB to susbtain recombination at
the above rate.
•What if IB is larger than QF /  F  F ?
dQF QF
 I B (t ) 
dt  F F
Step 1: Solve it for any given IB(t) to find QF(t).
Step 2: Can then find IC(t) through IC(t) = QF(t)/F .
Visualization of QF(t)
II B( t)(t )
B
dQF QF
 I B (t ) 
dt  F F
Q F (t)
QF
 F F Q F / F  F
EXAMPLE : Find IC(t) for a Step IB(t)
IB
IC (t) I B0
t
I B(t)
IC (t)
dQF QF
The solution of  I B (t )  is
dt  F F n
QF   F  F I B 0 (1  e t / F  F ) E B C
t
t /  F  F
I C (t )  QF (t ) /  F   F I B 0 (1  e )
What is I B () ? QF (0)? QF () ?
QF
8.11 Model for Large-Signal Circuit Simulation
• Compact (SPICE) model contains dozens of parameters,
mostly determined from measured BJT data.
• Circuits containing tens of thousands of transistors can C
be simulated.
rC CCS
• Compact model is a “contract” between
device/manufacturing engineers and
circuit designers. QR
rB CB C
B IC
QF
CB E
  VCB  I S qVBC / kT rE
I C  I S (e qVBE / kT
e qVBC / kT
)1    (e  1) E
 VA   F
8.11 Model for Large-Signal Circuit Simulation
A commonly used BJT compact model is the Gummel-Poon

model, consisting of
•Ebers-Moll model
•Current-dependent beta
•Early effect
•Transit times
•Kirk effect
• Voltage-dependent capacitances
• Parasitic resistances
•Other effects
8.12 Chapter Summary
• The base-emitter junction is usually forward-biased while
the base-collector is reverse-biased. VBE determines the
collector current, IC .
qni2 qVBE / kT
I C  AE (e  1)
GB
WB
ni2 p
GB   2 dx
n DB
0 iB
• GB is the base Gummel number, which represents all the

subtleties of BJT design that affect IC.
• The base (input) current, IB , is related to IC by the
common-emitter current gain, F . This can be related to
the common-base current gain, F .
I C GE IC F
F   F  
I B GB IE 1 F
• The Gummel plot shows that F falls off in the high IC
region due to high-level injection in the base. It also falls
off in the low IC region due to excess base current.
• Base-width modulation by VCB results in a significant slope
of the IC vs. VCE curve in the active region (known as the
Early effect).
• Due to the forward bias VBE , a BJT stores a certain amount
of excess carrier charge QF which is proportional to IC.
QF  I C F
F is the forward transit time. If no excess carriers are stored
outside the base, then
WB2
 F   FB  , the base transit time.
2 DB
• The charge-control model first calculates QF(t) from IB(t)
and then calculates IC(t).
dQF Q
 I B (t )  F
dt  F F
I C (t )  QF (t ) /  F
The small-signal models employ parameters such as
transconductance,
dI C kT
gm   IC /
dVBE q
input capacitance,
dQF
C    F gm
dVBE
and input resistance.

dVBE
r    F / gm
dI B

Chapter 8 Bipolar Junction Transistors: Question: What Is The Meaning of "Bipolar" ?

Caricato da

Informazioni sul documento

Titolo originale

Copyright

Formati disponibili

Condividi questo documento

Condividi o incorpora il documento

Opzioni di condivisione

Hai trovato utile questo documento?

Questo contenuto è inappropriato?

Copyright:

Formati disponibili

Chapter 8 Bipolar Junction Transistors: Question: What Is The Meaning of "Bipolar" ?

Caricato da

Copyright:

Formati disponibili

Chapter 8 Bipolar Junction Transistors

• Since 1970, the high density and low-power advantage of

• BJTs are still preferred in some high-frequency and analog

Question: What is the meaning of “bipolar” ?

Question: Why is IB often preferred as a parameter over VBE?

emitter base collector

When p > NB , inverse slope is 120mV/decade.

Is a large IB desirable? Why?

N E niB2 To raise F, NE is typically very large.

N E niB2 To further elevate F , we can raise niB by

With h = 0.2, EgB is reduced by 0.1eV and niE2 by 30x.

(a) At NE = 1019 cm-3, EgE  50 meV,

(b) At NE = 1020 cm-3, EgE  95 meV

(A) Increase the base width,

Which of the above is the most acceptable action?

The Ebers-Moll model describes both the active

Now reverse the roles of emitter and collector.

VBC causes a lot of holes to be injected

When the BE junction is forward-biased, excess holes are stored

F is difficult to be predicted accurately but can be measured.

Let’s analyze the excess hole charge and transit time in

What is FB if WB = 70 nm and DB = 10 cm2/s?

WB2 (7 106 cm) 2

2.5 ps is a very short time. Since light speed is

g m  I C /( kT / q) At 300 K, for example, gm=IC /26mV.

This is the charge-storage capacitance, better known as the

A BJT is biased at IC = 1 mA and VCE = 3 V. F=90, F=5 ps,

(c) C   F g m  5 1012  0.039  1.9 1014 F  19 fF (femto farad)

The parameters are routinely

The load is a short circuit. The signal source is a current source,

Step 1: Solve it for any given IB(t) to find QF(t).

Step 2: Can then find IC(t) through IC(t) = QF(t)/F .

A commonly used BJT compact model is the Gummel-Poon

• GB is the base Gummel number, which represents all the

and input resistance.

Potrebbero piacerti anche