MM-5
DISTORTION
IN
MICROWAVE
REVERSE
Robert
BIASED
Gerald
of Electrical
system
reverse
compromise
This
and experimental
biased
PIN
are presented
published
University
bias
frequency
presents
an
of distortion
with
As a result
original
expressions
by
order
relationships
the authors’
01803
degradation
distortion
intercept
The
will
can be worse than
further
of an analysis
to the first
points
compare
PIN
contrary
diodes
Many microwave
causes undesired
which
may
particularly
distortion.
harmonic
result
in
worsens
Distortion
contributing
system
and reverse
introduced
has been analyzed
data.
It
increasing
frequency,
biased PIN diode.
of the distortion
introduced
by the reverse
time
biased diodes,
i(t) = dQ/dt = d[C vi(t)]
/ dt
(1)
each
where C is the PIN diode reverse bias capacitance
by a forward
and is now predictable
varying
limitations,
A PIN diode switch
a measure of distortion.
The distortion
with
of the
predicted
generated by reverse
biased PIN diode is based on the nonlinear
current passing through the diode [4,5]:
in switches
signals or spurious in-band signals
significant
both forward
The
ANALYSIS
in sensitive receiver circuits.
may contain
derivatives
to experimental
that distortion
to the case of the forward
The analysis
PIN diode switches have a requirement
permissible
tail capacitance,
second and third
characteristic.
favorably
INTRODUCTION
for maximum
relating
and second
(C-V)
also be demonstrated
biased
as the
increases.
of diffusion
have been developed
capacitance-voltage
previously
biased PIN diode distortion.
with
MA
because of the diffusion tail phenomena, will exhibit a definite
capacitance change and will generate real distortion.
can
generated
distortion
that reverse bias distortion
distortion,
switches
Division
Products
Inc.
and RF communication
Concise
diodes.
and compared
results indicate
diode
paper
study
work on forward
forward
PIN
microwave
performance.
analytical
by
Hiller
M/A-COM,
South Ave.
Burlington,
ABSTRACT
caused
BY
Semiconductor
and Computer
Engineering
Southeastern Massachusetts
N. Dartmouth,
MA 02747
significantly
RF SWITCHES
H. Caverly
Department
Distortion
AND
PIN DIODES
biased PIN diode
the microwave
[1-3].
capacitance,
Distortion
voltage across the diode.
then
i(t)
will
is
contain
and
intermodulation
account because it has been assumed to be much lower
components,
that will
produce
signal
distortion
in a
microwave
or RF switching
circuit consisting
of a series
been reported
biased diode and no analysis
to predict
its distortion
generating
This paper reports the results of an original
than
to date has
and
experimental
investigation
of the reverse biased PIN diode.
Concise
distortion
prediction
expressions
have been
developed
and conclude
generate more distortion
The primary
varying
that reverse biased PIN diodes may
than forward
capacitance
assumed
will
exhibit
and therefore
higher
than
no change
generate
the
dielectric
in capacitance
no distortion.
The
actual
reverse
voltage,
PIN
of the ideal
for all reverse
voltage.
PIN
diode
bias voltages
In actual PIN diodes,
between the intrinsic
is generally
beyond
the
however,
the
layer and the end regions are
not exactly abrupt, but rather exhibit a continuously
increasing
doping concentration
that approaches the end region doping
These
so-called
diffusion
tails cause the
concentration.
frequency
increasing
constant
boundties
induced by the
relaxation
with
capacitance
punchthrough
microwave
or RF signal. An ideal PIN diode operated at a
reverse bias voltage beyond the punchthrough
voltage and at
frequencies
AC voltage.
The
biased PIN diodes.
cause of reverse biased PIN diode distortion
is the small but finite time varying
to the fundamental
connected
PIN diode between generator and load.
The
distortion analysis focuses on the variation of the PIN diode’s
reverse bias capacitance
and its relationship
to the time
properties.
analytical
in addition
harmonic
generated by a reverse biased PIN diode is often not taken into
that of the forward
products,
and vi(t)
If C is a time-vmying
biased
for voltages
&lode,
PIN diode’s
beyond
capacitance
punchthrough,
to continue
independent
to decrease
of frequency,
1073
CH2725-O/89/0000-
1073$01.00
(!3 1989 IEEE
1989 IEEE MTT-S
Digest
since the depletion
with increasing
change
profile
layer will
advance
reverse bias voltage.
with reverse voltage
is dependent
shape, but may be directly
C-V data.
and 100
Figures
micron
measurement
illustrates
profiles
frequencies
of
Thicker
than thinner
device
capacitance
determined
1 MHz
PIN
exhibit
forward
portion
computed
from
1 and 2) that demonstrate
and third
circuit
contributor
distortion
measurements
the theoretical
have
reverse
analysis.
Figures
distortion
[1,2,6].
Good
agreement
between
the
measured and computed distortion intercept points is noted.
The second order intercept point for the 50 micron PIN diode
They are related to the
is approximately
10 dB higher than that of the 7 micron diode.
reverse bias voltage
Z03
by 60 dB. Also, at 10 MHz
at 1000 MHz using series connected
The third order intercept
(dc/dv)2
relative
increases
range the
5 and 6 show the results of these measurements
using
previously
described techniques for measuring PIN diode
as follows:
= l/32
order
reverse bias voltage.
point shows similar
that the reverse bias distortion
IP2
that for the 7 micron
third
is the significant
order
biased PIN diodes to verify
1 MHz
PIN diodes
of the C-V data, measured at the input frequency,
and are expressed
with increasing
been performed
this point.
reverse biased PIN diode.
derivatives
improves
bias distortion
Second
Equations for the two tone intermodulation
distortion
intercept points have been derived from Eqn. 1 for the series
connected
bias distortion
improves
of
This results in a lower rate of
C-V data for 7 and 100 micron
(- 10v)
at 1000 MHz it is the reverse bias dktortion
that
In all cases, the reverse bias distortion
predominates.
change with reverse voltage for thick PIN diodes.
and 1000 MHz
bias
whereas
C-V
diodes since the depth that the depletion
thickness.
These results imply
the reverse
the forward
that
flatter
diode
distortion
(dBc) degrades by 32 dB as frequency
from 10 to 1000 MHz.
In the same frequency
from measured
and 1000 MHz
diodes
Figures 3 and 4 show plots of dC/dV
(Figures
PIN
on this diffusion
extends into the end region is a smaller
the overall
charge data [1,2].
1 and 2 show the C-V characteristic for 7
I-region
width
PIN diodes
using
this.
boundary
into the end regions
The rate of capacitance
intercept
behavior.
Note
point increases as the
increases in all cases.
(2a)
(OX +C02)2
CONCLUSION
The
and
objective
understanding
of this
PIN diode switches.
IP3
where
= l/12
COland
0J2 are the
intercept
point
available
from
evident
that
increases)
@2c/dv2)zo2
intercept
as frequency
diodes exhibiting
perform
are
the source.
the
+on)
referenced
to
the
From
these expressions
point
decreases
C-V characteristic
with better distortion
properties
order distortion)
These
forward
power
distortion
it is
The fundamental
and frequency.
bias PIN
improves
diode
switch
a better
in reverse biased
analytical
conclusion
is inversely
(for second
This is in contrast
distortion
with increasing
to the
case where
frequency.
shows that thick PIN diodes tend to exhibit
(distortion
increases. Also, thicker I-region
a flatter
was to provide
mechanism
is that the magnitude of the reverse bias distortion
proportional
to the slope of the C-V characteristic
(2b)
tone input ft_qUetICkS.
two
equations
(2tol
paper
of the distortion
the
The analysis
less distortion
than
thin diodes. Experimental
distortion data taken at 1000 MHz
on diodes of different thicknesses confirm this phenomenon.
PIN
are expected to
than thinner diodes.
ACKNOWLEDGMENT
This
APPLICATIONS
Reverse
using
bias distortion
Eqn.
operating
2 for
around
a series
National
intercept
points
connected
frequencies
PIN
material
were derived
by the Naval
diode
NOOO 14-87-C-2449.
of 10 and 1000 MHz.
switch
These
Goodrich
is based
Science Foundation
Research
and Theresa
results are shown in Table I. The results are based on C-V
their help and expertise
measurements
devices.
and their
derivatives
1000 MHz using the HP4 192A
the HP4191A
Table
same PIN
RF Impedance
I also illustrates
diodes
taken
at 10 MHz
LF Impedance
Analyzer,
forward
based on 10 mA
Analyzer
and
and
respectively.
bias distortion
resistance
on the
and stored
1074
upon
work
Laboratories
The
supported
under Grant MIP-861
authors
Suscavage
under
wish
Contract
No.
to thank
Joel
of M/A-COM,
in the processing
by the
1239, and
Inc.,
for
of some of the test
REFERENCES
0,46 CAPACITANCE
1.
Caverly,
Circuits”,
“Distortion
IEEE Trans. Microwave
MTT-35
2.
R. and G. Hiller,
(5); pp. 492-501,
May,
Theory
Microwaves
1
and Tech., vol.
1987.
Holler, G. and R. Caverly,
distortion”,
(PF)
in p-i-n Diode
“Predict
0.35
PIN diode switch
and RF, vol. 25(l),
p. 111, Jan.
1986.
0.26
3.
Caverly, R. and G. Phaneuf, “Nonlinem
and Transient
Microwave
and RF Modeling
of the PIN Diode”, to be
presented at the 1989 IEEE Intl. Symp. Circuits and
..
**
“.. ...* . . . . . . . . . . . . . . . . . ..? . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
0.16
10
0
Systems.
Penfield,
P. and R. P. Rafuse,
M. I. T. Press, Cambridge,
MA,
“Varactor
6.
Hingham,
Microwave
Technology,
VOLTAG:(VOLTS)
Applications”.
FREQUENCY
1962.
‘
5.
Mortenson, K., Variable
House, Dedham, MA, 1974.
Capacitance
E., “Measuring
50
40
REVERS:
4.
*1--1
..
*
Distortion
Diodes,
Artech
Figure 2.
1
*
MHZ
1000 MHZ
Reveree
blae PIN diode
capacitance
for
100 micron
device
at 1 and 1000 MHz.
a
in PIN Diodes”,
System News and Communication
p. 97, May,
_,* LOG[ ldC/dVl
1988.
IN F/V]
—,-12
I
-13
*
-14
1
-15
CAPACITANCE
*
1“
(PF)
-16
-
.*.
-13
- -14
.*
- -15
*
- -16
t
1
0.4 “
0.3
0.2
I
-171
o
. . . . .. . .. . . . . .. . . . . . . . . .. . . .
Figure
*
‘iOLAi
35
(voYTs)
3.
1 MHZ
1000
MHZ
Reveree
blaa dC/dV
at 1 and
a 7 micron
PIN diode,
*
1000
MHz
for
I
40
10
REVERS:
60
_12 LOG [ldC/dVl
IN F/V]
-12
VOLTAG:(VOLTS)
FREQUENCY
.
Figure
—-~_,7
FREQUENCY
%-... . . . . . . . . . . . . . . .
1
0’
o
,
30
R&RsE
Y.+ ....= .......
0.1
,
5
1 MHz
*
1000
-13
-.
-13
T
MHZ
1. Reverse
bias PIN diode capacitance
7 micron
diode at 1 and 1000 MHz.
for
a
-14
1
“
I
.
-14
-15
-15
1
.
*
-16-
.
*
~
-17
0
-17
5
30
RE!ERSE
‘;OLTA3E
(VOyTs)
FREQUENCY
“
Figure
1075
4.
“d
-16
*
1 MHZ
*
1000
Reverse
bias dC/dV
at 1 and
a 100 micron
PIN diode.
MHz
1000
MHz
for
36
lnterceDt
Point
(dBm)
TABLE I
I
90 I
Predicted
Diode
Intercept
pOhS
*
80 0
70 -
REVERSE BIAS
0
60 -y:+
.
Width
40 -*
,
6
30
o
,
10
Reverse
*
Voltage
I
,
16
20
+
w2-50
(MEASUREDl
IP2-7
(TbtEORy)
0
w2-60
(THEORY)
5.
1000 MHz second
ordsr
Intsrcept
point
measurements
on a T snd 50 micron
PIN
Point
(dBm)
*
o
60 -
*
40 -W*
*
30 *
,
6
20 I
o
,
10
Reverse
16
Voltage
,
20
(MEASUREDJ
+
lp3-60
(MEA5URED)
IFS-7
{THEORY)
m
w%-so
(THEORY)
6.
I
26
(volts)
tp9-7
Figure
1000 MHz Third Order Interoept
Point
on a 7 and 50 micron
PIN diode
IP2
[P-3
IPZ
IP3
IP2
1P3
-l@
-1OV
-20T
-20V
+lomA
+lomA
10MHz
94dBm
58dBm
102dBm
62dBm
39dBm
28dBm
llX)OMHz
6 ldBm
42dBm
72dBm
47dBm
79dBm
58dBm
10MHz
92dBm
60dBm
l12dBm
70dBm
77dBm
56dBm
lfXIOMHz
85dBm
57dBm
91dBm
6C!dBm
117dBm
86dBm
diode.
lCO
GoIntercept
BIAS
25
7
(MEA5uRED]
Frequency
microns
(Volts)
IP2-7
Figure
FORWARD
*
50 -.*
*
PIN
Measurements
1076