Академический Документы
Профессиональный Документы
Культура Документы
IEEE TRANSACTIONS
ON
MICROWAVE
THEORY
AND
TECHNIQUES,
VOL.
MTT-14,
NO.
12,
DECEMBER,
1966
Quantitative
Comparison
of Solid$tate
Microwave
A. M. COWLEY,
MEMBER,
Detectors
AND
IEEE,
H. O. SORENSEN,
MEMBER,
IEEE
AbstractA method for quantitative comparison of solid-state microwave square-law detectors is presented. The threshold response of the square-law detectors are compared for unit video bandwidth using the concept of Noise Equivalent Power (NEP). NEP is the microwave input power required for unity signal-to-noise ratio at the output of the detector. Contours of conin a 1 Hz bandwidth stant NEP in the microwave (RF) and video frequency plane clearly describe the dependence of threshold sensitivity on both video and radio frequencies, and thereby provide comparison of the threshold sensitivities of devices over the entire video and RF frequency spectrum. A criterion for the upper RF power lid of square-law operation for detectors is also presented. Dynamic range for a device can be found using this criterion and the threshold sensitivity of the device. Six solid-state detection devices are described briefly, then compared on the basis of the foregoing concepts. Four of these devices are familiar: the point-contact and planar Schottky-barrier ((hot carrier) diodes, and the tunnel and back diodes. Two relatively new devices are also discussed: the so-called hot carrier thermoelectric detector, and the space-charge-limited (SCL) dielectric diode.
through current in
Saturation
of a hot video
J=
Current
k = Boltzmanns
KN = Noise
kl
corner
. . . kl = Coefficients
f~.<<f.
and
Effective band of
conduction
m = Electron
by
a device,
in-
SYMBOLS
voltage
amplitude,
device
area
PB = Microwave
PB(USL) = Upper
dissipated
susceptance
g = Electronic
for barrier-type detunnel, decurwhen mihot carrier, point-contact, for or barrier-type
= Tangential
low-frequency circuit to
T= Temperature
transit t = Noise .L = White time temperature noise) applied bias voltage
in video is applied
crowave device
nonlinear dev
function of
of nonlinear respect to
voltage voltage
f(v)
with
fRF = Lficrowave
(RF)
freqUenCY (RF)
Thickness linear
(SCLD) responsivityamperes/watt, mobility coefficient-cmZ/voltZ for for for a device small with signal with microseries ~ responsivity series resistance very device responsivity
nOn-
frequency frequency
Noise
Direct plied
corner
conductance current in power in video is applied circuit to when a detector the absence of apbias
~=
Current parasitic
@O= Current wave ~0 = Current resistance Y= voltage 70= Al= Low-level Square-law
current
responsivity,
Manuscript received May 31, 1966; revised September 8, 1966. The authors are with -hp- Associates, Palo Alto, Calif.
COWLEY AND
SORENSEN: COMPARISON
OF SOLID-STATE
DETECTORS
589
The
tangential
bears
simple and
fixed either
bandwidth, threshold
completely
specifies
performance
time
electron mobility
Electron
NONLINEAR
DEVICE
ANALYSIS
devices
considered on a
here nonlinear
depend
for
their
relaxation relaxation
properties It
is therefore device
useful
Momentum
relaxation frequency a method Six for basic hot diodes, backward dielectric threshold
a nonlinear
operated function
as a low-level of a nonlinear
The
current-voltage by i = j-(v).
be denoted
comdeare bardeand bases freExpanding the imthe stituting as follows: (4), (3) we ILlicrowave bias; v as therefore, the detectors for (3) operated write voltage with the
?70
microwave carrier
are of
often bias
a dc voltage and a
generality,
we will a dc
point-contact diodes,
microwave
A cos d Colt.
series about lFO, and and dc
(SCL) are
sensitivity, of or
v==-Vo+Acos
in a power the
(4)
subterms
and be
dynamic of
square-law of what
paper
obtain
fundamental
potential basis,
ce as detectors. to compare will can equivalent microwave I. devices that, the with to the on an equal appropriate RF power that by the all be made to tuning, source. of the
device is
be matched
device. where
f(l), . . . , f()
at
VO,
are
of ~(v)
with
rmpect
to V, evaluated M[e in will of express the as the RF of each device which an of The vice This average and terms Noise Equivalent of unity, (NEP) a bandwidth a measure the device video which eted current rent ratio can due of A; is evident term A;, to on the the can
and
P absorbed
(5) by for the one microwave
by input
the
de-
be found to the
multiplying
voltperiod.
to produce
signal-to-noise By using
expression
P:
hertz. threshold
definition,
we obtain of
(6)
almost by inspection. side of the of (5) The first brackdetected in curThe /3, and
itself,
completely
independent
associated
present-day this
availabilseems to to the time right-hand average application form is the increase power. responsivity
noise
video since
be a reasonable measure noise concept quencies among to may bandwidth for in detector is not the
practical amplifier
in circuits ri~ther A
where assumption
a very at
to P is called in the Ai
operating
be written
common detector
=p=(jo P
where
[11+1
l+
AI (7)
specify often
performance than
a large This
video
width,
practice
impossible
operating
~ j(2)
frequencies. a square-law Sensitivity detector, (TSS) TSS = the commonly tcl NEP by (1) used Tanand is related 2.5(NEP)@3 Equation
(8)
A!2 Al=FT
f(4)
(9)
A2 f(3)
where in dB
B is the video
becomes TSSdB =
bandwidth.
NEP~B
4 +
5 Iog~O B.
(2)
f(1)
(10)
590
IEEE TRANSACTIONS
ON
MICROWAVE
THEORY AND
TECHNIQUES
DECEMBER
Using
(6),
and
making
suitable
approximations,
(7)
can
A.
Hot
Hot
Carrier
carrier diodes by two
and Point-Contact
(planar are the Schottky
Diodes
barrier) and devices of rectification chief metal difference point Hot film Some closer into by caron obvicontrol shock. with mathe followed point-contact sometimes a metal pointwhich at
contact are
described the
[2 ]. The a fine
construction; pressing
(12)
diode
the surface The the quantity device. the The /?0 is the quantity low-level A shows current how, responsivity for from higher true of power squarean electrical rier the ous diodes prepared
semiconductor, forming by
operation.
levels, law
detector The
deviates
operation.
advantages
microwave which
better
resistance to make
prescribed
deviation
capacitance
as that point-contact in
deviation
conduction, of ordinary
operation are no
diodes. these
(1 + to
\ A I ) =
0.3 dB.
(13)
minority
carrier capable
effects,
are potentially approaching time rd of the this frequency other and in upper carrier by the v ; the hot the
reciprocal
0.56f(1j
P(USL) =
f(4) f(2)
(14) # j(l) range of crystal [1] deviation. the (pointchoose The comwhen the rms
tion
semiconductor will
crystal. order
be of the
of 1000
In a paper contact) 0.1 dB reason monly operated TSS. detector ammeter When for
dealing detectors,
with
dynamic and dB
Sorger of 0.3
point-contact at lower
placing
rather our
as a maximum
point-contact equivalent
diodes circuit
described in Fig. 1,
accepted as a the
a device to TSS, by an
shown
sensitivity
relationship
level for by
output,
as measured, is given
or voltmeter,
= s[exkii--v)-+
(17) where n is a number diodes, diodes. Referring respectively, associated to the with Fig. 1, the elements resistance or series barrier, resistance and
RB
greater nearly
unity hot
for car-
But
since
input in dB,
power is given
to
TSS, in
we
know
that power,
point-contact rier
(A;),i~..l expressed
= 2.5(Ai).0i~e
error
measuring
and
CB are,
of the -
and
capacitance region
100d(ttlY2= 101010(1+(+))1
device, while = 0.3 dB. A comparison for upper in equal of and error. OF DEVICES; of the electrical NEP CONTOURS of in this the pointthe (13) and (16) limits shows that our (16) definiwith the bulk tions result III. A brief the lower of square-law operation DESCRIPTION description detection in order. since method, the
R. is a parasitic
semiconductor
substrate
characteristics
&
Fig. 1. RF equivalent
RR
characteristics of operation.
of hot A useful
diodes
physical
circuit of barrier-type devices: hot
technique,
carrier,
section.
point-contact,
tunnel,
1966
COWLEY
AND
SORENSEN: COMPARISON
OF SOLID-STATE
DETECTORS
591
whisker, of cur-
The (22),
result the
will
be denoted current
by
fl.
Referring for
to
(20)
and
characteristic
low-level
responsivity
the
device
including and
resi:~tance and
R,. The
and semiconductor
becomes O; .
1, is called
to form with
saturation
depends
metal
_&
1 + +
1
[1 + (f/f.)
.
1
denominator. upper
(23)
dynamic from
resistance (17) as
asso-
[1
(18) Note entirely Using with be uniexlaw limit that the in the (13) frequency last and dependence term obtains for DO is contained in the the
RB =
where For form pressed
nk T q(lo + 18)
Io is the direct
a diode impurity as fabricated density,
bias the
square-
from
of P~ as
(24) cB=
~ ~. ~ ()
co
l/i-
(19,)
Using wave
(20),
the input
upper to
limit
for
total
micro-
VB
power
where
CO is the
junction dimensions
and
VB
and hot acabthe
P~~(USL)
= 0.56
is a parameter
as the upon carrier count sorbed of the of the tance Fig. power diffusion the or the applying fact
known
[%Y+J[
a
(25)
VB depends
impurity theory diode, all of the we
density. of Section must microwave In order to of l/f noise to in Fig. calculate or the device the threshold hot must sensitivity diode, Both of the either noise of devices point-contact noise, carrier
Before
be known. to a uniform
is absorbed nonlinear in of
portion
in addition
RB. A portion
series circuit ratio resisof
noise its
of either circuit
(Xln be
noise
temperature
mtio
diode its video resisbarrier
~.
R.. A simple
1 yields absorbed the
equivalent
relation the
total
of the
2, the
device
a noise
current
video and
PB . pm
1
1 + : [1 [1 + (f/f.)]
RV is simply
RB +R,,
of value associated
Rv would
with by
excess
characterized
corresponding is 4kTBtfRv. As
(squared) lier,
the video
bandwidth
is taken
[1 2rCB(RJ?LI) 112
II that with the
to be unity.
(21)
of Section in underrefers carrier
referring corresponding
to
detected P.RF is
power
We
can
theory
P, as used
the is used when
section, or hot
P~
for (20)
i, = Ai = /30PRF.
(26)
a detector (8),
to total
of A; to PB, as obtained
(17)
l~
!l
E
The ing ratio
2nkT [ Ai/PB
engineer by multiplying will
nkT(IO
PB . + 1,) 1
(22)
[Hz
section.
However,
is= Ai
equivalent
=&
circuit
Prf
for barrier-type devices.
microwave obtained
is the
AijPRE,
IEEE TRANSACTIONS
ON MICROWAVE
THEORY AND
TECHNIQUES
DECEMBER
device
is
the
expression
can
be rewritten
4kTt iN= [1 RV
for unit The power power, the video noise which bandwidth. equivalent results Hence, in using power equal equal (26)
1/2 (27)
NEP = (NEP), 1 + 1
[
Note (NEP) output and (27) is signal noise that and current input noise at device in by Figs, (31) devices addition, that this 1 and whose are the expression equivalent 2, and threshold backward to the which
(fc)ll+fl]o
is general, circuits has are and applies forms Examples
31)
to any given of in
whose
or equivalently,
signal and
output.
tunnel
of course,
NEP=~.
(28)
of it
studies been
at the
authors corner
with
I/f
noise
has a noise
temperature
ratio
laboratory
fN
has and
expression
of hot of
point-contact
diodes
an equa-
tion
the
(29)
The where quency,
fN is
(32) depending for carrier hot indicating on the diode diodes l/f 3 this Figure during
constant It for
k-v
the less
familiar
~v the temperature
video
freratio;
is typically point-contact
and
tw the
than (27),
(28),
we obtain
after
some
shows study.
data
obtained
(30)
and
(31) carrier
are
convenient diodes, device, current TSS 4 dB (25). TSS and 10, for to The and
and
%[%112[+%12
KN,
(31)
all other
can be used
[1+(3T+:112
The video last two bracketed terms in (30) The depend remainder If by the on RF of
30)
and the
and
be calculated
(30)
bandwidth and
expression independent
independent. is denoted
frequency then
dynamic PRF(USL),
range
is
difference
between
(NEP)O,
expressed
50
40
30
10
100
1000
10000
100000
I mH.
I
10mHz lo(lmHa
FREQUENCY
Fig. 3. Noise characteristics of hot carrier and point-contact diodes,
1966
COWLEY
AND
SORENSEN: COMPARISON
OF SOLID-STATE
DETECTORS
593
B.
iVEP
The
Contours
responsivity of (23) the for PO for microwave the but hot a detector frequency carrier it and can may is, in genfRF, point-connot be be Simalways as
either
diode as unity
temperature
tw is typically be taken
eral, tact
calculations. range bias (30), diode. a is for the Hot veniently section. in order entire relatively corner can necessity general, and A (19) simple it of both conditions, (assuming Table other carrier One RF-video
these (NEP)
f<<f.
the For
This analytically,
functional
dependence in practice
devices
experiment with
as a graph excess
93.5
noise,
current
point-contact
diode
and
for and
iV in the
fv; video therefore
of video noise
calculated
iN is defined bandwidth
in this of
current (28),
hertz.
fRF
diodes method of the These calculate; found, current, each carrier with for the
are
a function NEP
that
presented devices if
NEp(fRF,
performance measure
frequency
(33),
for
a fixed
value
frequency
KN
for
is once
any bias
contou~.
of NEP com-
be calculated
picking
NEP, and
fN
these
contours
is observed obeyed
hot
specify the
of a detector. as NEP measured contours using plotted of on contours frequency of superior devices. to construct a hot carrier for of the in To (28). the of
are
experimentally can If same equal plane. threshold illustrate NEP diode, (NEP)o, two noise Fig. diode contact contact preference parison forward that, in the were such data 4. In has be constructed, NEP graph, NEP This
diode, (19).
(18) and
and
a line
Back Diodes
and back diodes depend upon [9]. impurity produce can lower then tunnel direction at the for their interestof essenis narrow easily. ~made P-n Tun-
The ing
tunnel
current-voltage
has been
quantum-mechanical tial y p-n junctions high across current than operated in the in
contours and
fN
two
familiar
point-contact f. were using in lower Fig. calculated (18), left-hand threshold 3. The
purposely junction neling voltage as forward When ily what tunnel reports biased detector for ther signed diode. usually acteristic origin peak The this in Back biased lower the into
data
of applied increasing again. ordinarsomepeak diode a video data furdeare charof are
junction
the to region,
voltage
as detectors,
diode. diode
than
current
1P [10].
the negative to the
However,
author
be biased
50 pA
preliminary
is worthwhile diode
region
the
[11 ]; due mode this for They diodes are [12 ]. resistance current. RF equivalent the bulk paper. are
RF
carrier example, hot band The despite l/j hot It Figs. carrier-
at lower
tunnel current
carrier
a lower
tunnel
excess
is indeed chosen is
operated
however; the superior of than range and the the of in hot 1.0.5.
valu{es
high-frequency i.e., it lower compare for values the two are of n for at to lower video enables
diodes, for
noise carrier is
corner, better
same
as that diodes
interesting
diodes;
resistance conductci~w-rier
combination capacitance.
diodes,
respectively,
typically
is a majority
594
IEEE TRANSACTIONS
ON MICROWAVE
THEORY AND
TECHNIQUES
DECEMBER
, 1, 111 , I 1I I , 8 1 1 t , I , II I NEPCONTOURS NEPO -95dBm = HOT CARRIERIODEf,.21GH~, fN. I()()IIZ D [ { fe=22.5GHi, iN=30 POINT CONTACT DIODE~Epo =-93.5dB VIDEO BANDWIDTH = Itiz
,
I
1 1 , I [ 11
rAT v
1 [ 141 ,- 103
%ooA%lAS)
{A
---zs
++-
1,.1
d,m, l
1 1
Ill
11
, 1 Ill , II 10
# 1
IL(I 102
, t , 1 , , JIu 103
Ki2
Ii-
10
fv VIDEOFREQUENCY (KHz)
Fig. 4. NEP plots for hot carrier and point-contact diodes
, 1 , 1
, 1 1I
1 1 1 I
DEVICE PROPERTIES - (NEP)O NOISE EQUIVALENT POWER FOR >>f~ , frf < C fv
lo~ # 1 1 k1,
ii-2
PREFERENCE REGIONOR F HOTCARRIER IODE D WITH50PA BIAS f~ =2,1GHZ (NEP)O= -95df3m fN =IOOHZ
1(?
1 1 111i r1 I , , , , ,1 , 10 10
carrier and point-contact diodes whose
, 1,t
.
, 0 a ,
,1
102
NEP plots are shown in Fig. 4.
103
fv VIDEOFREQUENCY (KHz)
Fig.
5.
Preference
regions
for
hot
1966
COWLEY AND
SORENSEN: COMPARISON
OF SOLU3-STATE DETECTORS
.595
TABLE
R,(Q)
CB(PF)
~c (GHz)
~,v(kHz) voltl
(yBEJP)
(2) P&(u:,)
.
93.5 87 95 88.5 95.5 90 95.5 90.5 93 97.5 97.5 54 60.5 62..5 67.0 82.5 70.0 --30 18
--31..s --19.5 31.5 21.0 31.5 21.5
(Low Noise PC Diode PC Diode HC HC HC HC HC HC Back Tunnel Tunnel Diode Diode Diode Diode Diode Diode Diode Diode Diode Detector Detector Detector Detector
=50 5-600
20 20
20 30
= =
13 8 19 12 19.5 16 20 18 25
59.5 65
59.5 65 tiO 65 60 ti5
;: 10 10 5 5 10 5 5
u
5.5
o
1 mA ) -
90 45 45
11.2 11.2
m m m
50 1050 o 0 o 0
l.OpF
cc
0.57)
2.OV current responsitity
1.OPF
2350 H F-
I
t
,.O1 , ,,, ,,,,
IF
2
resistance
16
BIASCURRENT (mA) I
and capacitance carrier diode. vs. bias
10
Fig.
6.
Barrier
for
typical
hot
596 phenomenon; are therefore limitations fact, cisely diodes, for age vices. able the tunnel the the as hot and same tunnel to carrier back way and and
IEEE TRANSACTIONS
ON MICROWAVE
THEORY AND
TECHNIQUES
DECEMBER
back
diode the
for of
(=100 above, II
PA) for
bias
levels
for
subject
basically
tunnel
considerably
higher devices
values, biased
diodes hot
near from
current. in
presents laboratory
as the .&,
point-contact
of data
measured literature.
authors
tunnel
point-contact
Bias Condition
50100 PA 100 PA 1 mA
Source
Eng [15] Follmer [18] Authors messt Authors messt
fli
3050 <1 <1 50 kHz kHz kHz kHz
with design
useable microwave Recent facturers responsivity order reverse current, however, of dynamic noise. Dynamic very made forward near mission the high the back the tunnel narrow and low, of
devices
applications. publications specification of 30 bias. this By tunnel at current and the biasing voltsl [15], sheets, and zero the [16], back and tunnel as well diodes moderate diode can by instability as many that can near the be its manucurrent of the or peak loss I/f Let back current, the noise tively. us compare diode, using corner For are the unbiased and the back the performance diode of Chase cutoff 93,5 11 diode are The data, of a 100 Q unbiased biased and dBm, GHz, frequency, and near corner 97.5 near Chang the [1o]. (NEP) zero, ~p~~k, the dBm, is an with peak For O and respeccorand a tunnel data diode,
indicate
enormous range
sensitivities onset of
biased noise
increase
is accompanied
quantities
estimate data
on Table other
together
similar I.
resistance often bias, peak of the while current Series diode. to junction tunnel Burrus for
of order
the
tunnel
devices
is usually can be
for
the
is summarized tunnel and without devices high operates poorer as As is best into the
diodes or often
Dynamic cult pression dynamic amplifier i.e., when Dynamic nearer to for
to have
50 Q total
resistance to obtain
moderate be biased
an analytical the
function. impedance
line. compared
to the detector,
current. us first
a crude that
estimate
dynamic written
observe
(22)
However, than nel 0.1 pF diodes Noise number [17] one related current tunnel current; the narrow
capacitances point-contact
special
experimental
used in tunnel
as detectors. and l/~ way occurs peak back in was to in and the the diodes [17], has been Yajima tunnel noise iexcess direction onset of p-n I/f studied and seems after by a Esaki diodes; to be a the current, for have L from the hot (34) carrier (18). as [15], noise [18]. that diode forward the used
B = Bo[l
and find We the
*@oRBP.]
(34)
diodes, where we term
point-contact square-law
of authors measured
deviation
germanium
observations some
,,02RB+(1+EJ
frequencies as less than .P~~(USL)
before the
junction is due to microwave write of the noise in genthe back bias Eng back higher his back lowwe of the using holds carrier clynamic and the for and criterion the range of tunnel for (36) in (12) the (13). P~~(USL)
(35)
~,. From
is thought of defect junction. ratio extremely typical data while that 10 to mixer measured
that levels
excess [18]
presence
forbidden
gap the
Follmer
temperature eral diode levels, [15] diodes, l/f noise have back noise agreement has with presents and than
t for back
Yajima low noise for his of
0.14
=
B02RB
with
(36)
(36) hot the
If
we diodes,
devices
data
indicates
the devices
of R~, R,,
diodes
15 dB
than for
that
the
i-v
a small
are the
devices diodes.
tunnel
diodes,
noise
corners
are for
point-contact
1966
COWLEY AND
SORENSEN: COMPARISON
OF SOLID-STATE
DETECTORS
597
n
in
an
of tunnel
the
dy-
diode
the
sensitive region
microwave
a consequent
resistance.
D.
Thermoelectric
The thermoelectric upon carriers by the
Detector
detector the in depends for effect r. The field, and its detection for carriers the hot are lattice to hemi[19] a I i
SEMICONDUCTOR CRYSTAL
1
r. ~.
properties majority heated temperature Fig. have detailed an analysis linear the an seen Its 7, the
thermoelectric a semiconductor RF
LARGEAREAOHMIC CONTACT
be readily
spherical
Harrison device.
Zucker present
performed
=IRV VP
o
RF E(lUIV CKT.
Fig.
into
7.
Thermoelectric
section. square-law
II
THERMOELECTRIC OETECTOR Rv=Ioofz I 1
device standard
without
the for a
is often
or planar,
r~, the
resistance
given by
R, = p/4ro (37)
~.uL.J
Ii
3
Fig. 8,
Iiz
Voltage ~e~p~n~ivity vs. applied
1(
POWER INPUT (mW)
10
detector
p is resistivity
semiconductor. has been chosen conformity for per for conand with
of thermoelectric RF power.
device,
expression
7, the
watt is,
for according
low
levels, This
but
voltage
responsivity
powers
=*[++ln%l[+08-7*:ro31
38)
-hpFig.
is shown
a typical
= TO(1 + @
where
(39)
operation at of
of the
the small
TED
(40) =z%[:+]n%
to move
contact of operation
=08[$z:roJ
~E = energy
70 = small ~ = carrier
(41)
of the time
relaxation
time
for h
(hot)>
carriers
detector Generally
speaking,
microwave time,
N,=
density
states coefficient
approaches
case
TP=10*3
Thus, when
the de-
in semiconductor power
used
as a microwave limit.
PRY = absorbed
q = electronic The open circuit
virtually
no upper
frequency
developed
for
unit
absorbed
Alto,
was kindly
furnished
by L. Wright
of -hp- Labs.,
Palo
598
IEEE TRANSACTIONS
ON MICROWAVE
THEORY AND
TECHNIQUES
DECEMBER
Noise ple: whose corner result The sistor ance may given also the
the
TED
are
also of
noise
properties
a resistor
Thus, is no l/~
as a simple
is equal
video only
impedance, changing
resistance
of the
contact.
semiconductor, higher as the contact resistance. contact radius both practical and the
a device
sensitivity for
resistivity. resistivity There desires sistance mission (37) n, the we is an for that line.
fixed,
increases upper
one retransby
p Fig. 9. The
be matched quantity
P
CB, ,. / ,., V B.
/zz/-BlAS APPLIED
diode band
(SCLD); diagrams.
device
configura-
have (40)
is range the of 50 to device 100 kfl. The thermal noise voltage from
inversely a desired
density device, contact contact units formed technique by by conshock, from to fabrihere is in techand NEP dBm, lent to solving from the for (2). NEP. dBm. detector, statements (13), the device dB, for Table for for For the to NEP device (31), in question sensitivity notice that = (NEP)
ancf
voltage
is proportional making
be maximized possible. in
(42)
10-9 NEP
radius size using heating has tacts, and unit cate the nology diameter. Let practical the
Extremely some
been
obtained contact,
a video for
bandwidth
R.=
10000.
by
be obtained
disadvantages
associated
point
of uniformity
= 62.5
Tangential
Referring
contact. planar
is practical
thermoelectric
o; this
is equiva-
microns
f,v=O
for the upper
f.-+ co.Applying
limit has with a dynamic p=O.2 a dynamic in summary L?-cm range
criterion us make some estimates detector. of the sensitivity silicon, of a we can This thermoelectric ~m2 For we have
thermoelectric
square-law
therefore a 1 Hz video
of about
bandwidth. and
,8=5 X1
O-8-) Voltz NC ~
PO = 1350 cm2/volt-see,
data
3 X 1019 cms.
of sensitivity Let to of us the say to From 7.-7 range point and Al. previously 100 fix 1000 a carrier (40) This we is disvideo Q !J
On video
the
basis we resistance of
of
E.
The Space-Charge-Limited
The space-charge-limited equivalent diode and will the in SCLD, power For
Dielectric
dielectric of the
Diode
diode (SCLD) is
devices resistance A practical (37) find indeed cussed, Al. have diodes, ~.
the the
Space-charge is well [21] considered transport to the in with e.g., the is most
contact of
radius
is 0.1 mil
or 2.5 microns.
From
in insulators device
semiconductors
p = 1 a-cm,
corresponding
density
6 X 1015 electrons/cm3. responsivitity to the To =floRv detectors to hot in the using carrier high video detectors
gives
to a square-law
a semiconductor of carriers,
extremely
resistance,
concentration
1966
COWLEY AND
SORENSEN: COMPARISON
OF SOLID-STATE
DETECTORS
599
lo3rrr3
SPACECHARGE LIMITED OIOOE I I
03pI-T--7]
102
J JO
10: : -:
/ EXACT APPROX I
10.
!iiul i
2 i
Fig. 11. SCLD
10
IO
Ii -
generated from
by data
subtracting points,
the leaving
extrathe
current
in in
the the
pre-
third-
fourth-order
ch arac-
teristic
of the
we use
MO(I
relation
/3.E2)
mobility with
(45) electric
to exact analysis of be transit to current obtained time calculate in Fig. the the the flow in is complicated, parametric If voltage form, one function but can with naively of the [22], nevertheless electron proceeded the SCL one device current would
describe we
the obtain
J =
behavior
of
the
field,
T as a parameter.
current simply 9 by low-level
(46]
where
the
first
superimposing current
(ohmic)
equation
(47)
(44)
where the n is the carrier semiconductor e the carrier dielectric concentration, constant, In Fig. the V ~, the These factor tuting For are approximate than two, relations, and and obtained accurate by simply within su~bstia
mobility, and of
J the current
and in The compared good. function introduces Referring voltages at current higher higher
10 the
approximating @ is
E by
about thick, to from thermal of the we be [24],
V/W.
have space[25]. noise dewice. at this For
Shockley
coefficient
5 X 10-8
are
cm2/volt2
a device devices
is indeed current and the that prevails, analysis. for low while (SCL) with linear region
@/ Tv=o.o5
The charge noise Ziel should
is an adequate
engineering simplificaFig. or the 10, we linear spaceFor free of SCL experiin the the a car-
suppressed, der
Van device
an SCL
observe
begins resistivity
to dominate. onset
is available theories.
is smaller, voltage.
device This
a typical
SCLD circles
at ambient
temperature. Webb
laboratories:
points,
Wright
600 For device, low from frequencies, (8) and the (46), is
IEEE TRANSACTIONS
ON MICROWAVE
THEORY AND
TECHNIQUES
DECEMBER
current
responsivity
of
the
at
about
one
half 0.
its It
lowhas cur-
detected
(49)
practitherefore
to ON27r.
we a device
have with
terms
containing response,
ks and
the
kd.
cur-
device and
could
be useful
square-law be
UHF
frequencies. make it
responsivity
dynamic power
measuring response
where wide
/30(perf
sq. law)
--+
(50)
accurate range.
square-law
The
dynamic
video
resistance 1
is 1 (51) A several cost where 1.25X Using we 103 (49), have cmz, Q-cm (52), volt-, The and most quantitative detectors burnout, important
IV.
SUMMARY between if equal which threshold response. for comparing viewpoint, method over are frequency method bandwidth This other or not are two, or among e.g.,
F)
(k, + k,V)A
area, with to 0.5 and area from volt.
factors,
A is the
neglected us
cross-sectional
properties
ks and
microns, and (8)
kd.
fabricated 10000
consider 10
a convenient
devices
(49),
we obtain
6,@l.O
contour
method. spectrum.
PRF. Considering
equilibrium circuit
(13), sensitivity is 72.5 500 ohms. I. dB, The with
thermal earlier,
dBm. Iaw
to of Fig.
we
quantitative frequency
comparison
to calculate
2, we find ~RF(USL) is 78.5 video
using
6 from this
the
The
effectiveness in bias in order RF values of the NEP the examined Fig. current
by show
contours results
is
equivalent
Tangential range
to be and
emphasized creasing diodes 50 ohm those contour ary shape The has NEP no
of in-
carrier
point-contact match contours of bias change to a with on the boundin ~.~ the cc,
source.
5, the
a device in Table RF
W~25 has
of the preference
are obvious.
preference contours
experiSCL de-
thermoelectric
equivalent
for
susceptance; In terms of
O= constant.
the
Rv,
can
the
frequency
conductance
example This of in
be written (52)
12 a 100 Q thermoelectric carrier has radio of its I range range we the are poor at hot or PCD and point-contact low sensitivity has a very It high
detector
to the
region region
frequencies. extremely shows for the typical in for see the expense and the that most
(53)
comparison devices this paper. where increased sensitive, point-contact range. the increase case
dy-
where electron
O =uT transit
is the
time signal.
T is the
of the con-
device
considered
calculated
fV>>fN.
microwave veniently
fRF<<f.
from
the
general, devices
is bar-
T ~
where For the Shao ductance is steadily stant value C is the device and geometrical considered Wright find somewhat, roughly
3RvC
capacitance above, that and we when that find of d >0, the
(54)
device. ns. the con-
of threshold
carrier
T~2
dynamic
Dynamic by biasing in
range to lower
decreases increasing,
dynamic
as though
capacitance.
Above
O = 27r, the
conductance
which
1966
COWLEY
AND
SORENSEN: COMPARISON
OF SOLID-STATE
DETECTORS
601
}Rv=jgOQ~-----
----:!~
-Miwd==lo l----------t----------#
., ,., ,,,
diode
in the
low
video to the
is due in
shown
vantages,
it is clear
that
contact
the field
%# %----i
---------
(GHI)
capacitance,
some
have carrier
subis a
improved. planar barrier the field units newcomer microwave presently obey extremely and analyzed, at will this ideal detection available: Schc,ttky low theorckical and time circuits rather that the and indicate this time, Chief barrier replace I/f
~-&__
100000
fy (kHz) + Fig. contours and preference regions for hot carrier and point-contact diode biased to 50 ohms. Also included is a thermoelectric detector with R,= 100 ohms.
mixing. have
12.
NEP
perfectly, of their they are can It barrier diode indeed, carrier frequencies for the
performance,
improve in to the dynamic diode; it range, is obviously due to series resistance to bias the effects device
devices y.
impossible
seems
a video
resistance
Rv less than
noise due region, to an
R,. Also,
a further increase
frein noise
in
the
1/~
microwave recent
is suffered
can, devices
even is
is larger This
for device
the
than
for
in the
SO GHz
range.
the
barrier-type predicted in
larger its
range usefulness
to planar
be
overcome. devices
suggests
power-moni-
instrumentation. V. DISCUSSION of solid-state paper. paper, frequency the AND CONCLUSIONS detectors and present have dynamic some gen-
low-inductance better diodes. biased sensitivity speaking, which results range the sharp near burnout
packaging
characteristics the peak any current, other limited an extremely is, genof noise Schottky the deby
Tunnel
in this
for comparison
than
conclude
each has,
classes. adarea
devices
The vantage
technical low
nonlinearity
peak to
current. be to those
charbar-
small-junction
better
point-contact
development extending of
diodes. identical are during they Since have with to video as for to usually tunnel diodes in their they In condo this
of microwave thirty
mixing, stancipoint
years.
designed
so that
is the well
of these
responsivity
comparable
of
point-contact
empirical barrier
resistance high-frequency It carrier and diode, stemming offers millimeter complete of I/f would
Point-contact devices; show to noise problem detectors certain of the device that mostly never good the
Schottky
capacitance.
agreement
due
competition
point-contact
ble disadvantage,
is its suscepti-
or Doppler
zero-bias video
thermoelectric advantages
control
manufacturing 5 clearly
Indeed,
inherent
almost
the superiority
a cutoff
frequency,
602
IEEE TRANSACTIONS
ON MICROWAVE
THEORY AND
TECHNIQUES
disadvantage is the fact that high sensitivity in these devices is always obtained at the expense of extremely typically in the range of high video and RF resistance,
100 kfl. in the This RF poor implies circuit, inherently and poor in the narrow-band pulse video by fidelity, frequency nonlinear by burnout operation or equivacircuit. mobility and of
Proc. IRE (Correspondence), vol. 49, pp. 12261227, July 1961. Noise in silicon microwave diodes, Proc. IEE [5] G. R. Nicoll, (London), vol. 101, pt. 3, pp. 317-324, September 1964. Jr., Characterization of crystal diodes for low-level [6] A. Uhlir, ~~~wave detection, Microwave J., vol. 6, pp. 59-67, July
.,
[7]
[8] [9] [10]
.,.
lently,
Harrison resistance
that
to that
point-contact least It
[11]
attention
theoretically. done, it
has
would
[14]
detector in this
mixer.
From that
response
reported
reciprocal
time,
is also high
pated
[20] [21]
resistance. ACKNOWLEDGMENT The authors would for like to express that their this appreciation be underduring Hall, and J. are
[25] [22] [23] [24]
suggesting
work
numerous
course. and
Discussions J. L. Mo1l
Lepoff, gratefully
helpful,
acknowledged.
[26]
REFERENCES [1] G. U. Sorger and B. O. Weinschel,. Comparison of deviations from square-law for RF crystal diodes and barrettes, IRE Tram. on lrsskw-nentatiors, vol I-8, pp. 1031 11, December 1959. vol. 15, [2] H. C. Torrey and C. A. Whitmer, Crystal Rect@rs.
M.I.T. Rad. Laboratory Ser. New York: McGraw-Hill, 1948. [3] K. Ishii and A. L. Brault, Noise output and noise figure of biased millimeter-wave detector diodes, IRE Tmns. on MicroY;KZ Theory and Techniques, vol. MTT-10, pp. 258-262, July [27]
[28]
[29]
noise in microwave
mixer
crystals,
A. Van der Ziel and A. G. T. Becking Theory of junction diode and junction transistor noise, Proc. IRE, vol. 46, pp. 589594, M arch 1958. H, 0. Sorensen, Using the hot carrier diode as a detector, ffezvlett-Packa~d J., vol. 17, PD. 2-5, December 1965. L. Esaki, New phenomenon in narrow germanium p-n junctions, Phys. Rev., vol. 109, pp. 603604, 1958. Tunnel diodes as millimeter P. E. Chase and K. K.. N. Chang, wave detectors and mixers, IRE Tram. on Microwave Theory and Techniques (Correspondence), vol. MTT-11, pp. 560561, November 1963. M. D. Montgomery, The tunnel diode as a highly sensitive microwave detector, Proc. IRE (Correspondence), vol. 49, pp. 826-827, April 1961. C. A. Burru~, Jr., Backward diodes for low-level millimeterwave detection, IEEE Trans. on Microwave Theory and Techrz@es, vol. MTT-11, pp. 357362, September 1963. T. P. Brody and P. H. Boyer, Evaluation of Esaki integrals and approximation of tunnel diode characteristics, Solid-State Electronics, vol. 2, pp. 209-215, May 1961. C. W. Bates, Tunnelinsz currents in Esaki diodes, . Phvs. Reo.. , . . vol. 121, p. 1071, 1961. S. T. Eng, Low-noise properties of microwave backward diodes, IRE Trans. On Microwave Theory and Techniques, vol. MTT-9, pp. 419-425, September 1961. R. B. Mouw and F. M. Schumacher, Tunnel diode detectors, Microwave J., vol. 9, pp. 2736, January 1966. T. Yajima and L. Esaki, Excess noise in narrow germanium p-n junctions, J. Phys. Sot. (Japan), vol. 13, pp. 1281-1287, November 1958. W. C. Follmer, Low-frequency noise in backward diodes, Proc. IRE (Correspondence), vol. 49, pp. 19391940, December 1961. R. I. Harrison and J. Zucker, Hot-carrier microwave detector, Proc. IEEE, vol. 54, pp. 588-595, April 1966. G. T. Wright, Mechanisms of space-charge-limited current in solids, Solid-State Electronics, vol. 2, pp. 1651 89, 1961. W. Shockley and R. C. Prim, Space-charge limited emission in semiconductors, P@Is. Rev., vol. 90, pp. 753758, June 1, 1953. M. A. Lampert, Injection currents in insulators, Proc. IRE, VO1. 50, pp. 17811796, August 1962. M. A. C. S. Brown, Deviations from Ohms law in germanium and silicon, J. Phys. Chem. Solids, vol. 19, pp. 218227, May 1961. V. Sergiescu, Space-charge reduction of shot noise for spacecharge-limited current in solids, Brd. J. A@@l. Phys., vol. 16, pp. 14351447, October 1965. A. Van der Ziel, Low frequency noise suppression in spacecharge-limited solid-state diodes, Solid-State Electronics, vol. 9, pp. 123-127, February 1966. P. W. Webb and T. T. Wright, Dielectric triode: A low-noise solid-state amplifier, J. Brd. IRE, vol. 23, pp. 11 11 13, February 1962. J. Shao and G. T. Wright, Characteristics of the space-charge limited diode at very high frequencies, Solid-State Electronics, vol. 3, pp. 291303, November/December 1961. G. T. Wright, Transit time effects in the space-charge-limited Electronics, vol. 9, pp. silicon microwave diode, Sohd-State 16, January 1966. D. T. Young and J. C. Irvin, Millimeter frequency conversion using Au-n-type GaAs Schottky barrier epitaxial diodes with a novel contacting technique, Proc. IEEE (Correspondence), vol. 53, pp. 2130-2131, December 1965.