Theory and Application of Field Effect Transistors
Theory and Application of Field Effect Transistors
TPJ\NSlf""70"\
by
/&?
College, 1952
A MASTER'S REPORT
FAST"?. 0? SCIENCE
'"
Approved by:
]
Major Professor
LU
Hi
CONTENTS
Page
"
INTRODUCTION
1 4 7
10
11
13
17
21
22
26
28
30
36
38
41
45
49 49
50
51 53
59
CONCLUSIONS
61
Ill
Page
REFERENCES
64
ACKNOWLEDGEMENT
LIST OF PRINCIPAL SYMBOLS
66
67
INTRODUCTION
device based on the modulation of a majority-carrier current. The principle of operation of this device, which he termed a
rather
width of a conducting channel through the narrowing or widening of a p-n junction depletion layer.
transistor was the high input impedance of the control electrode, which behaves essentially as a reverse-biased p-n
junction.
(7)
papor was published, the device in its pracIt was not until
reproducibility and with characteristics which offered considerable advantages over the junction transistor for certain
applications.
In fact,
it is only in the last few years that
channels
(7)
modi-
graded types.
An
since it and the epitaxial process enable the best control over
A "photoresist" pro-
Dif-
Another technique
(15)
Dram
Channel
Gate
*2
VX Y2
Fig.
Construction of a double-diffused 1. transistor. The p channel is formed by diffusing acceptors into the n substrate, which forms the second gate. A second diffusion through the masked surface forms gate 1.
Gate 1
Gate
log|N -N
D
log|N -N
D
Fig. 2. Approximate form of the doping profile for the double-diffused structure shown in Fig. 1.
b Along y_
(9)
Illustrated in Fig.
field-effect transistor.
terminal.
polar transistor".
nal is applied between source and gate terminals, then the effect
will be to widen and narrow the channel which carries current between source and drain terminals. This is closely analogous to
Fig.
3.
p(y)dy
(1)
where
Referring to equation
10
Published
presented here are general in that both the free carrier density
and space-charge density may vary arbitrarily with distance from
the gate junction
(1)
.
Both mutual
and
particular value V
The value of V
current to zero.
11
These
The Conductive Channel The physical structure for this analysis is illustrated in
Fig.
4.
Note that only the lower half p-n junction of the field3
is shown.
thickness
T,
he
T,
he
In Fig.
4,
drain contacts results when a voltage V is applied between the source and drain terminals. This current is restricted to the A field-effect
From a circuit point of view, the structures are the same except
12
terminal.
The values of T
Fig. 4.
Although a one-to-one
correspondence usually exists between free and space-charge densities, certain exceptions occur (2).
(y)
and p-(y),
13
that the reverse bias at both the gate and drain terminals are
the same so that the channel has substantially uniform thickness
(18).
Therefore,
Q(y)
= A f p(y)dy
(2)
where A is the effective area of the gate junction and has unit
directions.
<_
y < t is assumed
4)
where
(farads/cm)
The use of
14
Also
-
(y)
qp
p = impurity density
(3)
(N
Integration of equation
yields
(y)dy + C
(4)
is zero.
=0.
(y)dy
(5)
therefore
(y)dy - -
(y)dy
(6)
When equation
(2)
(6)
then
=
f-
[Q(y)
- Q(t)]
(7)
between
<
<
t.
15
V(t)
= -
[Q(y)
Q(t)] dy
(8)
Hence
V(t) =
[yQ(t)
(9)
are an integral of a
V(t)
= i
yp s (y)
dy
(10)
Equation
off voltage V
go to
T.
.
(10)
effect transistor has been taken, this limit then yields the reverse bias voltage that removes all the free charge from the con-
ducting channel; thus the current path has been pinched-off i.e.,
the channel no longer has the ability to conduct current.
The
pinch-off voltage V
put gate voltage, V
zero.
G
,
Hence,
16
.c
',-*
r y
i
s (y)
dy
(11)
now oe evaluated.
The
differential of equation
(10)
is
dV(y)
(yc
(y)dy)
<
<
(12)
When equation
(12)
(2)
(12)
equation
becomes
dV(y)
2ydQ(y)
Ac
<
<
(13)
where
dQ(y)
= Ap
(y)dy
(14)
therefore
_ eA
j
dV
2t
(y =
t)
(15)
Equation
(15)
that the space-charge layer thus acts as a parallel plate cap acitor with plate separation t.
17
Channel Current
An expression relating the drain current, I, to the physical
ponent E
must be present.
tween channel and the n-regions varies with distance x and therefore the channel thickness varies (18).
In the previous section,
3x
How-
x =
I*
<
C17)
in general
3x
p
However, if
y is 3x
very small
compared to
(18)
and t(x)
18
x is referred to by Shockley
(18)
The
electric field E
= E
<j(x)
(18)
Now
P^(y)
qp-
(y)
x=
"
PfW
<
19)
from which the total current through any cross section of the
1=2 x
J dS x
(20)
T,
he
p
= 2y
^W
(y)
The factor of
Equation
(y)
(21)
dy as given
19
by equation
(12).
Hence,
T
2Wuyp(y) dy
I
dx =
(y)cty
(22)
From equation
tions then,
(2)
direc-
/he
Q(T
hc
- Q(t)
=
J
(y)dy
(23)
(22)
it then
^f- tQ(T hc
Q(t)]yo s (y)dy
(24)
When equation
(24)
is integrated from x =
to x = L
the
to
* x
_ 2 "W eL
V
[Q(T
T
s
hc
Q(t)]yo s (y)dy
(25)
equals
20
From equation
a
(25)
and drain ends of the .channel which in turn are functions of the
= F(V
GS )
(26)
V GS
(27)
as the drain
current that flows when the external gate and source terminals
are shorted together.
By this definition,
but
has a value
(5)
jhc
[Q< t
4t^
Q<t)]yp
(y)dy
(28)
T,
the
Or
V
V_. DS
to go to zero by increasing
d
T.
he
21
max
then
= WJ
max
AT.
he
where
(2 8)
AT,
he
conducting channel.
mation.
Equation
greater than AT hc
(5)
field-effect transistor.
of the channel
(see Fig.
5)
is
vacuum pentode.
An examination of Fig.
5
voltage V
For this
22
(13).
Another region of
These two regions are separated by the dashed curve, which is the
= V locus of pinch-off points, i.e., the points where V nc DS
,
Fig. 5.
a
Mutual Transconductance
The most active and useful parameter of the field-effect
The amplifying
t.
the drain current is a function of the small increments of gateto- source voltage and the drain-to-source vo ltage,
i .e.
,
= f V
'
GS<
V 0S>
(29)
Equation
(29)
can be written as
3I 3I
AI
3V
d -
AV
d AV D 3V D
(30)
the re fore
AI
d =
V
D
^d
AV
(31)
whe re
3I
?m
3V
(32)
and
"
<?d
3V
(33)
so that
3I
3I
=
<?m
d
S
3
3T
9I
.
d D
3T
d G
3V
"
3T
3V
3T
3V
(34)
~
'.""
JI,
and
"s
j d
equation (25)
Thus
24
S^^l!
s
tQ(T
hc )
(T
(35)
and
St-^E a
t (T
hc'
-Q(T s )]Tdp s (T d
(36)
qyo(y)
V__)
W~
S
(38)
E
and
3V
3T
gs
d
qVs'V
(36),
(38)
nq
%- 3 k<V
"
Q(T s>
<
40 >
25
is extremely interesting;
it shows that the transconductance of the field-effect transistor is simply the conductance of the rectangular section of the de-
to y = T,.
Maximum g
G
is obtained when
D
he
and T
=0.
and V_ =
V. p
Hence,
*m,
(max)
=^ Q(T c>=9o
<
41 >
where g is the bulk conductance of the entire semiconductor body. o This value is independent of the space-charge density (1) . The
*d =
r <V
"
Q(T
d"
(42
>
Equation
(42)
According to equations
(28)
and (42)
when I, =
At V
^d
V DS
produces no in-
crease in I..
(1)
.
When T, approaches T
breakdown
(5)
26
Be-
cause pinch-off is not a sharply defined phenomena itself it becomes necessary to define somewhat arbitrarily how the pinch-off
are
D - *p *
(1
"
>"
P
(43)
Equation
(43)
The
'
The values of
and V
When equation
(43)
is differentiated
one obtains
nl
^m=^E P
<
-T
P
_
)
.
<
44)
E
27
DSS
Fig. 6.
(45)
GS=
Equation
(45)
for
VGS =
and V
<
(6)
v
p
-
gd
nl
(46)
28
= nl
(47)
are measured in
the pinch-off region and for all practical purposes are indepen-
(6)
have shown that all field-effect transistors have a fixed relationship between pinch-off voltage V bias forward transconductance
. ,
zero-bias current
and zero-
and V
second-order effects; for the most part they can be ignored (6).
Temperature Dependence of
Although
29
attests.
pendence of
Dif(T
)
A,
dl
rhc
dT
2W du eL dT
[Q(T
Q(t)]yp_(y)dy
(48)
mobility.
2.0
V
7
1.5
9s
1.0
DS
= -]
0.5
-75
-25
50
100
150
Free-Air Temperature, C
Fig.
7.
30
with temperature can be held to very low levels by proper selection of the bias point. This zero bias point depends on the physi-
As temperature increases,
This
The rate of decrease for most FET's is between 0.6 and 0.8 per
cent/deg C
(3)
Any p-n
31
If the
dz
^mz
(49)
dz ~-
= 0.315 volts
"mz
where g and mz
&z
mz
must
be developed.
(see Fig.
3)
the drain
pinch-off voltage V
is.:
d *
V CS
"
-F>
P
(50)
where
GS
= 0.
Equation
(50)
is called
32
d = g 'mz 3V GS
3I
(51)
or
= g y mz
_ v
P
31
V
(1
-f.) p
(52)
JL a
'mz
_E
2
/I (1
VGS,
(53)
substituting equation
(53)
gives
V
0.315 =
-2-
V
(1
GSt,
'
or
V GSZ = P " - 63
(54)
V GS
,
FET with V
P
- 0.6 3 volts
the
or
*dl
(3)
and g
Thus
at zero
drift,
33
-I (2^, p
(55,
a y mz
= _JP- (- 63
) '
(56) '
(
Equation
at an I,
(55)
must operate
For example,
if V
and
(10)
= KV p
(57)'
Substituting equation
(57)
K 9mz = l<^>P
(58 '
K 9mz = 2<f> P
0,4
<")
V.
p
Therefore, low V
p
.
34
have on V
'
It has already been found that the change in depletion layer thick-
Assuming
(I,/g
_
change, in volts/det C, of V
(3).
I.,
D = 0.0022 -
0.007(
<?m
(60)
in volts/det C.
Dividing equation
(50)
by equation
(52)
reveals that
I.
V
(1
V
- -i)
-S = -^
(61)
Substituting equation
(61)
D = 0.0022
0.007 -(1
- -^.)
(62)
Rearranging equation
(50)
(1
- -.)
V GS
d (~) P
1/2
(63)
Substituting equation
(63)
35
V
D = 0.0022 - 0.007
1/2
-^-(jF
5
P
-)
(64)
(I 1
1/ 2
(65) 3;
Then
(66)
D = 2.2(1 -
/=J
/
DSS
(67 )
Source resistor R
is large enough
&
constant.
s
adjusting R
until V
at 100C is equal to
(58)
V Gb
at 25C (10).
To confirm equations
and (59)
field-effect transistors
1^
The measured
36
+ 30 v
Fig.
8.
8.
permit design of d-c amplifiers which exhibit little change in operating point even with generator impedances greater than one
megohm
(10)
have higher
P
37
4
3
0)
10
\
?
V
V
o e
0. 8
\
100
0.6
\h
\
400
600
100
I., microamperes
Pig.
9. Drain current at zero drift for various The cross reprevalues of pinch-off voltage. sent measurements made in the circuit of Fig. 8. The line represents values calculated from eg. (58)
4
3
01
01
c
1
x\
n.R
n 6
V
\
0.5
2000 3000
5000
micromhos
10. Transconductance at zero drift for various values of pinch-off voltage. The line represents calculated values from eq (59)
38
stage.
put impedance and the low input current drift of the small-geometry field-effect transistor and the ease of circuit design afforded by
the higher drain current of large field-effect transistors.
Circuit Design
As any given FET type may have as much as 5:1 spread in
I
perform satisfactorily
(16)
The range of
the ouiescent I, has now been determined; however, the peak sig-
nal current
Dm a v
must be added.
39
To maintain a
(max)
determined.
1.5 V
,
then
V DG " D
%
-
(max)
,\
<
68 >
and
V__
1.5V
(max
= -I
>
&
(69)
(max)
R_
To do this
reduce i, and/or V
I
or select
(17)
.
As
spreads of 3:1
(17)
40
'
(max
"out
1
Fig. 11(a).
Self-biased amplifier.
(max)
'
(min)
P(mirK
(max)
'
GS
Fig.
11(b).
Transfer characteristics.
41
^SG
Fig,
12.
42
process (9).
(9)
by
V
V'
=
l
9s
VSB
(70)
The value of g
by the equations
= y. + y v. v J is gs rs ds
(71)
+ y i, = y, v v. J d fs gs -'os ds
(72)
43
Fig. 13.
Two-port y-network.
Output shorted:
y.
= -^2 gs
'fs
gs
Input shorted:
CIS
"us
3< C DG + C SG>
(73)
44
'rs =
"^ C DG
<
74 >
^ C DG
^C DG
(75 >
" *DS +
(76)
It has been
in the oinch-off
(4 4)
45
and (72)
Both
equations (71) and (72) have dimensions of current, and thus the
right hand side of each equation suggest the summing of currents
entering a junction.
and (72)
effect transistor.
to be negligible
is assumed
(71)
(let I
both
y.
and y
in equation
Hence,
= V gs +
ds
<
77 >
* "
Vgs
Vds
(78 >
redefining,
= g
'fs
and
o.
where g
device.
46
The equivalent small-signal circuit for the field-effect transistor is shown in Fig. 14.
The assumption that
1=0
Fig.
If a load JL
is applied to the
(79)
gs
47
gs
^d
ds
(80)
'ds
1
1*\
(81)
therefore
r+
As long as g.R.
<<
1
Vl
^
= g_Rr
(82)
then A
Tne value of A
varies
linearly with
(20).
|a
P_
When
:
R_
I 1
g g d
(20)
amplification factor
of
v
The value
characteristics.
48
Z.
in
GS -S I
g
- -
(83)
49
output terminals pairs, giving altogether six possible combinations, but the three connections shown are the only ones that
have any major applications. The field-effect transistor symbol shown in Fig. 15, indicates the interchange-ability of the drain and the source terminals.
Without some special marking, the drain is indistinHereafter, the drain will be desigThe symbol shown indicates
terminal is reversed.
output
input
output
inout
output
input
(a)
(b)
(c)
Fig. 15. Basic amplifier configurations: (a) common source, (b) common gate, (c) common drain.
50
effect transistors.
conventional transistors.
out previously.
The d-c
-9
x 10
amps
(8)
This cur-
bias is
V GS = G -
VG
(84)
51
<R
A.
D
-o
iC
cr
G R
<U
(a)
(b)
Pig.
(a)
16.
gS
T
Fig.
-> r ^
Qe
"o
<s.
17.
52
If,
value of R
(8),
(8)
upon temperature.
If V
is
held constant,
creased.
Since transconductance g
production spread of device parameters, some form of bias stabilization must be used
(8)
Capaci-
tor C
serves to bypass R
53
G *L
D
-e
l
N?
^F
as
'gs
e g ~m gs
>
r
Fig.
i
(a)
(b)
S.
18.
(a) (b)
9~
*
-<3*T
'gs
S
a e
\
L
m gs
o<
S
(a)
(b)
Fig.
19.
(a) (b)
54
rV
^
$
(a)
(b)
Fig.
20
(a) (b)
55
R^ g
v =
(86)
which is quite familiar to vacuum-tube engineers as the approximate gain of a pentode stage.
In general, the amplifier input admittance is complex and
is made up of the sum of 1/R_, plus the
gate-source admittance,
This
combination of
R,
of interest
(8)
(Fig.
18a)
+ e iN - gs
Vs
(88)
e + e iN = gs 57 o
(89)
56
is the ratio of R
amplification is given by
\
1 +
%RT V
u
<
90 >
substituting equation
(86)
1 + g
^m
<
91
>
(91)
becomes
A R L v = ig
<
92
>
a source-
emitter follower and cathode follower circuits and has the same
types of advantages.
_ o "
J gs g m
(1
+ g R ) d s
<>
,-,
57
e e
- i R
os
,f + (l
gm R s
p glRJ
.
(94)
g
e + e e e in = gs o gs
l
(i +
gd R s
<
95)
g (i
P.
v "
i in
Rjg+g s m 3a
(96)
)
= i
+\(g m +g a d s ^m
(9?)
>
C gs
where C
= C
gs
(1
(98)
= gate-to-source capacitance.
58
The
FET Q. by the h-
If the out-
gj, is assumed to
and 1/R
-2_ =
e.
R g m h!. fe 5
1
In
+ Q h' R -m fe s
In equation (99)
fe
fe
R+h. D ie
(100)
where h.
and h.
Equation
is similar to Equation
(96)
100-200 KHz.
(3)
59
Junction power field-effect transistors have some provocative characteristics for applications through VHF.
Currently
of
(core drivers?)
offset voltage.
switches?)
(17)
current dictates large junction area and high feedback capacitance, while high voltage means large base width and reduced
gain.
In the FET, junctions can be kept small because current
Also,
For
(17).
FETs in general, and the low noise resistance of these new high
current devices in particular, an evaluation unit has demonstrated over 140 DD of dynamic range operating as an RF front
end (17).
60
61
CONCLUSIONS
The field-effect transistor, or FET, as discussed in this
reverse-biased diode
input characteristic providing 10 9 to 10 13 ohms d-c input resistance; square law transfer characteristic allowing only second
e.
In
= 20 nv/Hi"
erly biased.
electrometer
Each
junction.
be below 1 pA at 30 volts
this is equivalent to 10
ohms
input impedance.
62
Drain-gate voltage V,,,, should also be the DG lowest practical value in order to reduce gate-drain leakage to
The operating gate current
I
a minimum.
drain-gate V
voltages.
A single field-effect transistor connected in a commonsource or common-drain configuration may be biased to operate
This characteristic
V BE>
efficient.
temperature.
In the
63
Radiation reduces
transistor rather than the bipolar injection type now so prevalent, most circuits would be using field-effect transistors.
Injection types would play a special role only-a complete reversal of the actual situation today.
64
REFERENCES
(1)
Bockemuehl R. R. Analysis of field-effect transistors with arbitrary charge distributions, IEE Trans, on Electron Devices, Vol. ED-10, pp. 31-34, January 1963.
,
(2)
Bockemuehl, R. R. Field-effect modulation of photoconductance Vol. 31, pp. 2256-2259, December 1960.
J. Appl
Phys
(3)
Buckholz, W. Biasing a FET for low drift, Electronics, p. 92, May 30,
1966.
(4)
Bruncke, W. c. Noise measurement in field-effect transistors, Proc. IEEE, Correspondence, Vol. 51, p. 378, February 1963.
Cobbold, R. S. C. and Trof imenkof f F. N. Theory and application of the field-effect transistor, Proc. IEEE, Vol. Ill, No. 12, pp. 1981-1991.
,
(5)
(6)
Cowles, Laurence G. Measurement of FET pinch-off voltage, Proc. IEEE, p. 200, February 196 4.
Dacey, G. C. and Ross, I. M. The field-effect transistor, Bell System Tech. J., Vol. 34, 1149-1189, November 1955.
(7)
pp.
(8)
Evans, A. D. Analyzing high-input-impedance FET amplifiers, Electronic Equipment Eng., Vol. 11, p. 72, March 196 3.
Evans, A. D.
(9)
Characteristics of unipolar field-effect transistors, Electronic Industries, Vol. 22, p. 99, March 1963.
(10)
Evans, L. L.
Biasing FETs for zero D-C drift, Electro-Technology, p. 94, August 1964.
(11)
65
(12)
Phillips, Alvin B. Transistor engineering, McGraw Hill Book Company, 1962, Chapter 4
Radeka, V. Field-effect transistor-its characteristics and applications, IEEE Trans, on Nuclear Science, pp. 358-364, June, 1964.
(13)
(14)
Richer and Middlebrook. Power law nature of field-effect transistor, Proc. IEEE, p. 1145, August 1963.
Roosild, S. A., Dolan, R. P., and O'Neil, D. A unipolar structure applying lateral diffusion, Proc. Inst. Electrical Electronics Engrs p. 1824, 1963, 51.
. ,
(15)
(16)
Sevin, S. A simple expression for the transfer characteristics of a FET, Electronic Equipment Engineering (EEE) p. 59, August 1963.
,
(17)
(18)
Shockley, W. A unipolar field-effect transistor, Proc. IRE., Vol. 40, pp. 1365-1376, November 1952.
(19)
Uzunoglu, Vasil Semiconductor network analysis and design, McGraw-Hill Book Company , p. 68, 196 4. Van der Ziel, Albert Electronics, Allyn and Bacon, Inc., p. 103, 1966.
Wang, Shyh Solid State Electronics, KcOraw Hill Book Company, 1966, Chapter 3.
(20)
(21)
66
ACKNOWLEDGMENT
The writer is indebted to Dr. W. W. Koepsel, Chairman of
the Department of Electrical Engineering, for providing a faculty
67
dgs
DS
SG
D
E
x'
,
T
G
cutoff frequency
gate terminal
?m
gd
X i
dz g
V
i
channel current
small signal drain and source current
d'
*
D
A
p(y)
Q(y)
0(t)
q
ss
t
electronic charge
source terminal
68
T.
nC
half-thickness of channel
depletion layer thickness at the source
T.
V V V V
w
e
,
v,
drain-to-source voltage
gate-bias voltage (relative to source)
gate-to-source voltage
width of channel
dielectric constant
Pi
P
by
MASTER OF SCIENCE
The con-
Three characteristics of
(2)
Excellent stability.
The analysis
output conductance,