llllllllllllll|||lllllIllllllllllllllllllllllll|l|||Illllllllllllllllllllll .

. USO05233306A
Unlted States Patent [19] [11] Patent Number: 5,233,306
Misra [45] ' Date of Patent: Aug. 3, 1993

[54] METHOD AND APPARATUS FOR Line: Test of an Improved Calibration Technique,
MEASURING THE PERMI'ITIVITY OF Misra, et al., Jan., 1990, pp. 8-14.
MATERIALS A Study on Coaxial Line Excited Monopole Probes for
[75] Inventor: Devendra K. Misra, White?sh Bay, In Situ Permittivity Measurements, Misra, Dec, 1987, 5
Wis. pages.
[73] Assignee: The Board of Regents of the Primary Examiner-Kenneth A. Wieder
University of Wisconsin System, Assistant Examiner-Jose M. Solis
Milwaukee, Wis. Attorney, Agent, or Firm--Foley & Lardner
[2 l] Appl. No: 655,755 [57] ABSTRACT
[22] Filed: Feb. 13, 1991 A method and apparatus for determining the permittiv
ity of a sample is disclosed. The method includes apply
[51] Int. Cl.5 ........................................... .. GOIR 27/04 ing an AC electrical signal in the microwave frequency
[52] US. Cl. .................................. .. 324/601; 324/646; range via a coaxial probe having an end position near
324/642; 324/643 the sample and measuring the re?ection coefficient of
[53] Field of Search ............. .. 324/630, 601, 605, 638, the sample. The complex permittivity of the sample is
324/642, 643, 646; 128/804 determined from an admittance parameter of the sam
[56] References Cited ple/probe combination and a system constant. The sys
tem constant is determined by measuring the re?ection
U.S. PATENT DOCUMENTS coefficients of four standards having known complex
4,507,602 3/1985 Aguirre ............................. .. 324/638 permittivity. The admittance parameter of the sample/
5,047,725 9/1991 Strid etal. ........................ .. 324/601 probe combination is determined from the-admittance
parameters of two standards/probe combinations and
OTHER PUBLICATIONS the measured re?ection coefficients of the sample and
A Quasi-Static Analysis of Open-Ended Coaxial Lines, the four standards. The admittance parameters of the
Misra, IEET transaction Oct., 1987, pp. 925-928. two standard/probe combinations are determined from
Re?ection of an Open-Ended Coaxial Line and Appli the known complex permittivities of the two standards
cation to Nondestructive Measurement of Materials, and the system constant. The apparatus includes a coax
Mosig, et al., Mar, 1981, pp. 46-51. ial probe having an end positioned near the sample, a
Measurement of Radio Frequency Permittivity of Bio microwave frequency generator and a device for mea
logical Tissues with an Open-Ended Coaxial Line: Part suring a re?ection coefficient are connected to the other
II-Experimental Results, Stuchly, et al., Jan., 1982, pp. end of the coaxial probe. A microprocessor determines
87-92. the complex permittivity of the sample from the mea
Dielectric Measurements with an Open-Ended Coaxial sured re?ection coefficients of the sample and four
Probe, Marsland and Evans, Aug. 1987, pp. 341-349. standards having known complex permittivities.
Noninvasive Electrical Characterization of Materials at
Microwave Frequencies Using an Open-Ended Coaxial 13 Claims, 3 Drawing Sheets

\ ' - - ' ' - -

/2oo
' - ' ' ' - ' n - _ ' ' _ _ _ _ "-1

l. msruv MICROPROCESSOR
3201.
l ' I

i \ I
; m 214 216 i
: w > w > :
. comma comma ;
I

i
l
l
l
I l

I 226 l
I l m 108
. __|SOLATOR_
: w 124
g 102 204
I--- _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ . _ _ . . _ _ . -_.| _/

\ m SAMPLE
242
US. Patent Aug. 3, 1993 ~ Sheet 1, of 3 5,233,306

, 100
136
l, .... A .... -r. .................... l

I I

; 132 129 I
, 015111111 mcaoraoczsson 2 vmon 1101111121511 \ ,
i
1
k, ' |
I06 I1
l - l

; 112 34 \"0 3 126 128 i

l I
. 11a 1211 .
125
i 116 I 1118
I l
1 [\J : TEST SH "T' 124
; v 122 ; 1
' 102 \ '
I 104 1 /
1 __________________________________ .1114 SAMPLE

/ 200 FIG. I
I ' - ' ' ' - ' ' - - _ ' - ' ' ' - ' _ - - ' ' ' - ' ' ' _ ' - " '1

msmv = MICROPROCESSOR - .
2211:

I.-_
1 M
\
112 214 216
1/11 > M) )
(01112111211 couvmsa

125 08 >

L_
as

If;i
s
1
124
I
l

I I
I
'

'0 is I I I I In12 I5 is I I I

/til
US. Patent Aug. 3, 1993 Sheet 2 of 3 5,233,306
US. Patent Aug. 3, 1993 Sheet 3 0f3 5,233,306 "

0. , -60
500 700 900 1100 1500 1700 1330 2100 2300
MHZ
Fig. 4

I l l l l l
1 I I I I I

500
M 71X) 91X) 11oo1aoo1soo1eoo1aoo 2M0
Hz
Fig. 5
u-calculaied?eal o-Cabulatedlmagnary
l-Measured?ed 0 Meamedlmaginary
 5,233,306
1 2
sample because using such a method should not require
METHOD AND APPARATUS FOR MEASURING destruction of the sample and would be useful over a
THE PERMITTIVITY OF MATERIALS broad frequency band. Utilizing a coaxial line or probe
to electrically characterize a material entails measuring
FIELD OF THE INVENTION the re?ection coefficient or input impedance of a sample
This invention relates generally to measuring the material and, from the measured data, determining the
complex permittivity of materials and, in particular, to dielectric properties of the material. Prior art methods
using an open-ended coaxial probe to measure the com have not been successful in providing an effective way
plex permittivity of materials at microwave frequencies. to relate the measured data to the dielectric properties
of the material. 4
BACKGROUND OF THE INVENTION In one approach, nomograms are generated at three
Nondestructive methods for measuring the dielectric frequencies to determine the complex permittivity of a
properties of materials are needed in biomedical and material from the measured re?ection coefficient for an
industrial applications. Because the interaction of an SR7 type coaxial line at the three frequencies. See
electromagnetic ?eld with a material is highly depen 15 Mosig et. al., Re?ection of an Open-ended Coaxial
dent upon the electrical properties of that material, Line and Application to Nondestructive Measurement
many electrical properties of a material may be deter of Materials, IEEE Transactions Instrum. Meas., Vol.
mined by measuring how an electromagnetic wave IM. 30, No. 1, pp. 46-51, March, 1981. However, since
interacts with that material. However, the precision many more nomograms are necessary to cover the en
with which one can determine how an electromagnetic 20 tire frequency range, and each set of nomograms is
wave interacts with a given material depends upon how useful with only one probe, this method is not very
accurately the complex permittivity of the material is practical. Moreover, generating the nomograms re
known. A method for determining the complex permit quires a large number of time consuming numerical
tivity is thus needed. computations.
An accurate determination of the complex permittiv 25
Another prior art approach uses equivalent circuit
ity of a sample can be bene?cial in carrying out certain parameters determined by the numerical computations
biomedical procedures, such as electromagnetic thaw of the nomogram method and empirical relations to
ing of cryo-preserved organs and tissues, electromag provide an improved model having an acceptable accu
netically-induced hyperthermia in cancer treatment, racy up to only about 2 GHz. See Stuckly et. al., Mea
detection of pathological conditions in tissues and diag surement of Radiofrequency Permittivity of Biological
nostic monitoring applications such as lung water con
Tissues with an Open-ended Coaxial Line, IEEE Trans
tent. Since these processes are often carried out at mi
crowave frequencies, the complex permittivity, which actions Microwave Theory Tech., Vol. MTT-30, pp.
is frequency dependent, is of particular interest at mi 87-92, January, 1982.
crowave frequencies. 35 More recently, the permittivity has been determined
Industrial uses of complex permittivity measurements from measured data utilizing a bilinear transformation
include monitoring the electrical characteristics of a to account for imperfections in the measuring system in
material produced by or used in a manufacturing pro conjunction with an equivalent circuit model for a coax
cess. For example, water content of a material may be ial opening. See, Marzland and Evans,
monitored by measuring the electrical characteristics of 40 Dielectric Measurements With an Open-ended Co
the material. Industrial processes may require that the axial Probe, Proc. Inst. Elec. Eng, Vol. 134, pp.
complex permittivity be measured continuously or at ' 341-349, August, 1987. Because this technique is re
frequent intervals as a material passes the measurement stricted at high frequencies by the inadequate circuit
point on a production line. Such uses require that the model for the probe, a quasi-static analysis of a coaxial
permittivity be determined in real time (nearly instan 45 sensor has been proposed to formulate a more accurate
taneously), so that appropriate action may be taken if a equivalent model, Misra, A Quasi-Static Analysis of
?aw is detected. Open-ended Coaxial Lines, IEEE Trans. Microwave
Conventional methods of determining complex per Theory Tech. Vol. MTT-35, pp. 925-928, October,
mittivity at microwave frequencies are unsuitable for 1987. In this method, a quasi-static approximation to the
use in biomedical or industrial applications. For exam formula for the normalized aperture admittance (Y]_) of
ple, one method requires cutting, polishing and then an open-ended coaxial line terminated by a semi-in?nite
placing the sample in a suitable waveguide or cavity. medium on a ground plane is given as:
Another method, utilizing free-space techniques, is
based on the re?ection and transmission of electromag
netic waves radiated by a narrow beam antenna. This 55
method requires that the sample have a relatively large
plane surface. Samples used in biomedical or industrial
applications often do not provide such a large plane where k=mV pope, r= |p"+p"2pp'cos I4, and a
surface. and b are the inner and outer radii of the coaxial aper
Coaxial line excited monopole probes have been sug ture, k=w V poeoec, pgis the permeability of free space,
gested for use in the electrical characterization of mate 6' is the complex permittivity of the semi-in?nite me
rials. However, an electric monopole probe is fre dium, ec is the relative permittivity of the coaxial line
quency sensitive and useful only when it is inserted into and w is the angular frequency of the electromagnetic
the material medium. Further, an electric monopole ?elds. The calculations required to solve Equation 1 are
probe requires a relatively large sample volume which 65 time consuming and do not provide an easy method for
is not practical in many cases. determining the admittance of the sample.
Prior art methods have attempted using an open Accordingly, a need persists for a method and appa
ended coaxial line for electrical characterization of a ratus for easily and accurately determining the complex
 5,233,306
3 4
permittivity of materials at microwave frequencies mixers are further coupled to a ?rst and second ?rst and
using an open-ended coaxial line. Such a method and second analog to digital converters, respectively. A
apparatus should accurately relate, in real time, the microprocessor connected to the mixers has suitable
measured parameters of a material such as the re?ection means for receiving re?ection coef?cient data indica
coef?cient or input impedance of the sample to dielec tive of the sample and the standards, and determining
tric properties of the sample at frequencies from about the permittivity of the sample using the re?ection coef
100 MHz up to and above 20 GI-Iz. Moreover, such a ?cient data.
method and apparatus should be capable of measuring BRIEF DESCRIPTION OF THE DRAWING
complex permittivity continuously or at frequent inter
vals. 10 FIG. 1 is a schematic diagram of an apparatus accord
ing to the invention;
SUMMARY OF THE INVENTION FIG. 2 is a schematic diagram of an alternative appa
A method for determining the permittivity of a sam ratus of the invention;
ple according to the present invention, using a coaxial FIG. 3 is a ?ow chart showing the steps of a method
probe having a ?rst end positioned near the sample and of the invention;
a second end, comprises the initial steps of applying an FIG. 4 is a graph plotting the relative real component
AC electrical signal in the microwave frequency range of the complex permittivity on the left vertical axis and
to the second end of the probe and measuring the re?ec the relative imaginary component of the complex per
tion coefficient of a ?rst, second, third and fourth stan mittivity on the right vertical axis of a 0.1 N saline
dard, whose complex permittivities are known. From 20 solution versus frequency, shown on the horizontal axis,
the measured re?ection coefficients of four standards determined according the invention; and
and their known complex permittivities, a system con FIG. 5 is a graph plotting the relative real component
stant is determined. Admittance parameters of the com of the complex permittivity on the left vertical axis and
bination of the probe and three standards, preferably the relative imaginary component of the complex per
three of the four standards used previously, are then 25 mittivity on the right vertical axis of glycerol at 23 C.
determined from their known complex permittivities versus frequency, shown on the horizontal axis, deter
and the system constant. In a preferred embodiment, mined according to the invention.
wherein one of three standards is a short circuit having
in?nite admittance, only the ?rst and second of the four DETAILED DESCRIPTION OF THE
standards are needed to determine these admittance
PREFERRED EMBODIMENT
parameters. The steps of measuring the re?ection coef According to a preferred embodiment of the present
?cient of the sample and determining the admittance invention, the complex permittivity of materials is de
parameter of the sample and probe combination from termined from measured input re?ection coefficient
the determined admittance parameters of the two stan using a quasi-static analysis of an open-ended coaxial
dards and the measured re?ection coef?cients of the 35 line. In general, collected data are used to correct for
sample and the four standards are performed. Finally, approximations in calculations used to determine the
the complex permittivity of the sample is determined complex permittivity. More particularly, the complex
from the admittance parameter of the sample/probe permittivity of a sample is determined according to the
combination and the system constant. invention by first measuring the re?ection coef?cient of
According to a further aspect of the invention, an 40 four standard materials whose complex permittivities
apparatus for determining the permittivity of a sample are known, and from the re?ection coefficients deter
comprises a microwave source and a coaxial probe mining a system constant which is characteristic of the
having a ?rst end and a second end, wherein the ?rst probe. The admittance parameter of the probe/sample
end is positioned near the sample. A re?ection coef?ci combination is next determined from the system con
ent measuring device is coupled to the second end of the 45 stant and the re?ection coef?cient of the sample. Fi
probe and to the microwave source for measuring the nally, the complex permittivity is determined from the
re?ection coef?cients of four standard materials. A admittance parameter of the probe/sample combination
microprocessor is also coupled to the re?ection coef?ci and the system constant. By utilizing four standards and
ent measuring device and has suitable means for receiv determining the system constant, the complex permit
ing data indicative of the re?ection coefficients of the tivity may be quickly and accurately determined.
sample and the four standards and determining the per Referring now to FIG. 1, a permittivity measuring
mittivity of the sample using the re?ection coef?cient device designated generally as 100 comprises a micro
data in accordance with the method of the invention. wave source 102, a re?ection/transmission test set 104,
According to another embodiment of the invention, a vector voltmeter 106, a probe 108, a microprocessor
an apparatus for determining the permittivity of a sam 55 110, and a display or other output device 112 which
ple comprises a microwave source and ?rst and second gives the complex permittivity of a sample 114.
directional couplers. Each coupler has an input, a main Microwave source 102 should be capable of provid
output and an auxiliary output. The input of the ?rst ing AC signals having frequencies from about 100 MHz
directional coupler is coupled to the microwave source, to about 20 61-12, so that the complex permittivity may
and the input of the second directional coupler is cou 60 be obtained over a wide range of frequencies. Accord
pled to the main output of the ?rst directional coupler. ingly, both test set 104 and vector voltmeter 106 should
A coaxial probe has a ?rst end positioned near the sam be capable of operating at these frequencies. A network
ple and a second end coupled to the main output of the analyzer, which usually is capable of operating at
second directional coupler. A ?rst mixer is coupled to higher frequencies than a vector volt meter, may be
the auxiliary output of the first directional coupler and 65 used in place'of vector volt meter 106. Microwave
a local oscillator, and a second mixer is provided and source 102 is coupled to test set 104 and provides
coupled to the auxiliary output of the second directional thereto the AC microwave signal it generates (the inci
coupler and the local oscillator. The ?rst and second dent signal). Test set 104, in turn, is coupled to vector
 5,233,306
5 6
voltmeter 106 and probe 108, which is positioned near software incorporated in microprocessor 110 deter
sample 114. The incident signal is provided by test set mines the complex permittivity in a manner which will
104 to vector voltmeter 106 and to sample 114 through be described below.
probe 108. The incident signal is reflected by sample Display 112 may be a CRT, LCD, printer, or other
114 and provided through probe 108 back to test set apparatus having an input 134, which receives the out
104. The re?ected signal is then provided to vector put data from microprocessor 110, and is capable of
voltmeter 106 by test set 104. Vector voltmeter 106 displaying the complex permittivity in an understand
determines the re?ection coefficient of sample 114, able form. The entire apparatus, except probe 108, may
de?ned as the ratio of the magnitude and phase of the reside in a common housing 136 for ease of use.
re?ected signal to the incident signal, from the signals In operation, before testing sample 114, re?ection
provided to it by test set 104. Vector voltmeter 106 is coef?cient data are collected at a selected frequency for
coupled to microprocessor 110, and provides micro four standards, preferably an open-circuit, water, a
processor 110 with the re?ection coef?cient of sample short circuit, and methanol. The standards used for
114. Microprocessor 110 then determines the complex calibration should be chosen so their dielectric con
permittivity of sample 114 and provides it to display 112 stants (the real portion of the complex permittivity)
or any other similar output device. span a wide range of values. In any event, it is desirable
More particularly, test set 104 has an input 116 cou to use standards having dielectric constants that de?ne
pled to microwave source 102, an incident wave output a range in which the dielectric constant of the sample
118 coupled to vector voltmeter 106, a re?ected wave being tested falls. This type of selection will generally
output 120 also coupled to vector voltmeter 106, and a 20 provide more accurate results, especially at high fre
load port 122 coupled to probe 108 Input 116 receives quencies. The applicant found that using standards hav
the incident signal from microwave source 102 and ing dielectric constants from two to eighty was ade
provides it to incident wave output 118 and load port quate to determine the permittivity of glycerol and a 0.1
122. The incident signal is transmitted by port 122 to N saline solution. Also, as will be explained below, a
sample 114 through probe 108, and a signal is re?ected 25 short-circuit is chosen as one of the standards (the third
by sample 108 back through probe 108 to port 122. The standard in the equations below) to simplify the calcula
re?ected signal is provided by port 122 to re?ected tions performed by microprocessor 110. The re?ection
wave output 120. Output 118 thus provides the incident data for the standards is provided to microprocessor 110
signal to vector voltmeter 106, and output 120 provides for storage and lateruse in determining the permittivity
the re?ected signal to vector voltmeter 106 30 of sample 114. The frequency selected is the frequency
Probe 108 is a coaxial line with an open first end 124 at which the complex permittivity of sample 114 will be
placed near sample 114 and a second end 125 coupled to measured. Data from the standards may also be col
port 122 of test set 104. The incident signal is received lected at a number of frequencies, or during a predeter
from port 122 on end 125 and transmitted to sample 114 mined frequency scan, in order to'later determine the
on end 124. The incident signal is re?ected by sample complex permittivity of sample 114 at the frequencies
114 and received by end 124 and transmitted to end 125 scanned. Time may be saved by ?rst performing mea
The re?ected signal is then transmitted by end 125 to surements at all frequencies of interest using the ?rst
port 122 of test set 104. Since the analysis used by the standard, then at all frequencies of interest using the
present invention is based on the presence of an in?nite second standard, and so on.
sample, the size of probe 108 should be selected to en 40 Re?ection coef?cient data are collected for each
sure that fringing and radiation ?elds are contained standard by successively placing open end 124 of probe
within the material. 108 in contact with each standard. In the case of an
Vector voltmeter 106 has an incident wave input 126 open circuit, end 124 should be positioned so that it is
coupled to output 118 of test set 104, a re?ected wave not in contact with any material. When collecting data
input 128 coupled to port 122 of test set 104, and an 45 for water or methanol, end 124 should be inserted into
output 129. Vector voltmeter 106 samples the signals on the liquid, and when collecting data for the short cir
inputs 126 and 128 and determines the re?ection coef? cuit, end 124 should be in contact with a conducting
cient of the sample/probe combination. A signal indica plane such as a piece of copper. After end 124 of probe
tive of the re?ection coef?cient is provided on output 108 is properly positioned, microwave source 102 gen
129. Other devices, such as a network analyzer capable erates a signal at the selected frequency. Vector voltme
of providing an output indicative of the re?ection coef ter 106 samples the transmitted and re?ected signals and
?cient of sample 114, may be used in place of vector determines the re?ection coef?cient, which is then pro
voltmeter 106. The device used to determine the re?ec vided to microprocessor 110.
tion coefficients should be capable of operating at the Following the collection of the data for the stan
frequencies at which'measurements are being made. dards, end 124 of probe 108 is placed in contact with
The output is preferably digital so that it may be pro sample 114. If sample 114 is a solid, end 124 should be
vided directly to microprocessor 110 which receives it firmly in contact with sample 114. If sample 114 is a
on a digital input 130. Input 130 may receive a parallel liquid, end 124 should be inserted into sample 114. A
or series signal. If an analog output is provided by vec microwave signal at the frequency of interest is pro
tor voltmeter 106, it can be digitized by an analog-to 60 vided by microwave source 102, and the re?ection
digital (A/D) converter before it is provided to micro coefficient is determined by vector voltmeter 106. The
processor 110. re?ection coef?cient is then transmitted to micro
Microprocessor 110 is preferably able to receive in processor 110. When measuring the re?ection coef?ci
struction inputs directly from the user in addition to the ent of sample 114, microwave source 102 should oper
output of vector voltmeter 106. Microprocessor 110 65 ate at the same frequency at which the standards data
should also be capable of providing output data on was collected. Similarly, if a frequency scan was used to
output 132 indicative of the complex permittivity of acquire the standards data, an identical frequency scan
sample 114. Suitable logic in the form of hardware or should be used for testing sample 114.
 7 5,233,306 - 8
The complex permittivity of the sample is then deter 208 has an input 236 coupled to port 234 of directional
mined utilizing a set of instructions, e.g. a computer coupler 206, an output 238 coupled to mixer 212, and a
program, embodying the formulas described below, port 240 coupled to end 125 of probe 108. The incident
which correspond to an equivalent circuit model. The signal is received on input 236 and provided by port 240
calculated permittivity is provided as a signal by micro to sample 114 through probe 108. The incident signal is
processor 110 to display 112. System 100 thus deter re?ected by sample 114 back through probe 108 to port
mines the complex permittivity of sample 114. 240. The re?ected signal is provided by port 240 to
In industrial applications that require taking continu output 238. Thus, output 232 of directional coupler 206
ous or frequent permittivity measurements, the re?ec provides the incident signal to mixer 210 and output 238
tion data of the standards from one measurement at a of directional coupler 208 provides the re?ected signal
given frequency may be used for other measurements at to mixer 212.
the given frequency. This facilitates industrial applica Mixers 210 and 212 make the incident and re?ected
tions of the present invention because probe 108 may be signals more readily digitizable. Since it is difficult and
in contact with a moving web or membrane and contin expensive to digitize a signal in the 61-12 range, mixers
uously monitor the complex permittivity. Alternatively, 210 and 212 mix the signals received from directional
probe 108 could be in intermittent contact with samples couplers 206 and 208, respectively, with a signal from
passing by, or the probe could be moved in such a way local oscillator 218 The mixing lowers the frequency of
as to scan a sample. In either case, the standards data the data signals and, therefore, allows for easier digitiza
does not need to be regenerated for each measurement. tion. To accomplish this, local oscillator 218 should
However, periodically standards data should be reac 20 operate at a frequency f;, which is slightly different
quired to correct for instrument drift, temperature from f1, the frequency at which microwave source 102
change and other changing parameters. is operating. The outputs of mixers 210 and 212 each
An alternative embodiment of an apparatus of the have a frequency f3 equal to the difference between f1
invention is shown in FIG. 2, wherein a permittivity and f2. For example, if fl is selected to be 10.0 GHz and
measuring system 200 comprises microwave source 102, 25 an appropriate value for f; is 10.5 GHz, f3 would be 0.5
an isolator 204, a pair of directional couplers 206 and GHz. In any event, f2 should be chosen such that the
208, probe 108, a pair of mixers 210 and 212, a pair of output of mixers 210 and 212 may be digitized without
A/D converters 214 and 216, a local oscillator 218, a distorting the waveform. Preferably, f3 should be less
microprocessor 220, and display 112 which reside in a than one-half of the digitizing rate of A/D converters
unitary structure 242 and cooperate to determine the 214 and 216 to meet the Nyquist requirement for digitiz
permittivity of sample 114. Isolator 204 is coupled to ing a signal. When the user selects an operating fre
microwave source 102 and receives an AC microwave quency for microwave source 102 (the frequency at
signal from it (the incident signal). isolator 204 is also which permittivity is to be measured) the frequency of
coupled to directional coupler 206 and provides the local oscillator 218 is preferably automatically adjusted
incident signal to it. Directional coupler 206 is also 35 to an appropriate value. Such an automatic adjustment
coupled to directional coupler 208 and mixer 210, both is accomplished by combining the output of the fre
of which receive the incident signal from directional quency selection mechanism of microwave source 102
coupler 206. Directional coupler 208 is further coupled with an offset, and using the combination to select the
to mixer 212 and probe 108, which is positioned near frequency of local oscillator 218.
sample 114. The incident signal is provided by direc The outputs of mixers 210 and 212 are provided to
tional coupler 208 to sample 114 through probe 108, and A/D converters 214 and 216, respectively, where they
a signal is re?ected by sample 114 through probe 108 are digitized and the resulting signals sent to micro
back to directional coupler 208. The re?ected signal is processor 220. Microprocessor 220 is suitably pro
also provided by directional coupler 208 to mixer 212. grammed to determine the re?ection coef?cients from
Mixers 210 and 212 are further coupled to A/D con 45 the digital data and then perform the necessary steps to
verters 214 and 216, respectively, local oscillator 218, provide the complex permittivity of the sample.
and microprocessor 220. The incident and re?ected The operation of permittivity measuring system 200 is
signals are mixed with a signal from local oscillator 218 similar to that of permittivity measuring system 100 in
by mixers 210 and 212, respectively, and provided to that, prior to testing sample 114, re?ection data are
A/D converters 214 and 216, respectively, for digitiza 50 collected for four standards. Again, the standards are
tion. The digital outputs of A/D converters 214 and 216 preferably an open circuit, a short circuit, water and
are provided to microprocessor 220, which determines methanol. The data collected for these standards is used
the re?ection coefficient of sample 114. From the re to determine the permittivity of sample 114. Re?ection
?ection coefficient, and a system constant, micro coefficient data are collected for each standard by ?rst
processor 220 determines the complex permittivity of 55 placing the end of probe 108 in contact with a sample of
sample 114 quickly and accurately. each standard.
Isolator 204 has an input 226 coupled to microwave The formulas derived below allow the admittance
source 102 and an output 228 coupled to directional parameter of sample 114 to be determined from mea
coupler 206. Isolator 204 provides the incident wave sured re?ection coefficients, and the permittivity to
received from microwave source 102 to directional 60 then be determined. The stationary formulation for the
coupler 206 and protects microwave source 102 by input admittance of a coaxial line terminated by a semi
preventing any re?ected signals from reaching it. Di in?nite medium on a ground plane given in Equation 1 -
rectional coupler 206 has an input 230 coupled to output above is simplified for practical evaluation of permittiv
228 of isolator 204, an output 232 coupled to mixer 210 ity. Using the assumption that the coaxial opening is
and a port 234 coupled to directional coupler 208. The 65 electrically very small, Equation 1 may be approxi
incident signal is received through input 230 and pro mated by taking the ?rst few terms of the Taylor series
vided to mixer 210 on output 232, and to directional expansion for the exponential term in the integral in
coupler 208 on port 234. Similarly, directional coupler Equation 1. The second term of this expansion goes to
 5,233,306
9 10
zero and the fourth term reduces to the well known
radiation term of coaxial probes, Following the expan
sion and approximations, Equation 1 may be rewritten
as:
l1 and 12, which require time consuming calculations to
solve, are not evaluated according to the present inven
tion, as explained below. 25
The ?rst term of Equation 2 represents a capacitance
of the sample, the second term represents a capacitance
of the sample that changes with angular frequency (w),
and the third term represents the radiation conductance
of the sample. In formulating Equation 2, an in?nite
conducting ?ange is assumed over the coaxial aperture.
However, an in?nite conducting ?ange is not used in
practice because it is inconvenient. Also, small disconti
nuities between the aperture and vector voltmeter 106
(due to connectors, etc.) cannot be avoided.
In order to account for these imperfections, the cir 35
cuit is modeled by an equivalent two-port network
connected between vector voltmeter 106 and open end
of probe 108. The actual admittance of the aperture
terminated by a sample is evaluated from the measured
re?ection coef?cient as follows:
5:153:
5:251:
where Y5 is the desired aperture admittance terminated
by the sample material, Y1, are aperture admittances
with standards one, two and three terminating the
probe, respectively, and 8,], which is called re?ection
difference data, is equal to l",-I"j, with 1",, representing
the measured re?ection coefficient for the nth material.
Equation 3 is shown with the ?rst, second and third
standards, but it could have been written using any
three standards.
The right-hand side of Equation 3 is determined from 55 where
the measured data, Y1 and Y; are calculated from Equa
tion 2 using the known values of e for the standards,
and a short circuit is used as the third standard so that
Y3: on. To avoid having to evaluate integrals I1 and I;
to calculate Y1 and Y;, a fourth standard is used, as will
be shown below. Equation 2 may be rewritten as:
7L=,'+e,'2+E1e,'2'5 (4)
where e, is the complex relative permittivity of the 65
material, 2 and 1 are constants dependent on the fre
quency and the dimension of the aperture, and Y L repre
sents a transformed parameter of aperture admittance
YL. Because 5 and 1 are not dependent on the sample
being tested, they may be determined using the data
from the standards.
Generally, the radiation from the coaxial aperture,
which is represented by the third term in Equation 4,
may be neglected at lower microwave frequencies.
Therefore, Equation 4 can be approximated as follows:
inner" (5)
Equations 3 and 5 can be used to determine the admit
tance parameter y, which, in turn, can be used to calcu
late the complex permittivity of the sample as shown
below. Solving Equation 3 for L and, using a short
circuit as the third standard so that 3= m, we obtain:
which we will call the re?ection difference ratio for the
sample, _standard 1, standard 2 and standard 3. Of
course, y, may be determined if one of the'three stan
dards used in Equation 3 is not a short circuit, but the
calculations will be more time-consuming. The values
of ylvz are determined as part of the probe calibration
and are calculated from Equation 5 for given complex
relative permittivities of standards 1 and 2 (an open
circuit and water in the preferred embodiment), respec
tively, at the operating frequency.
The unknown constant E in Equation 5, which is
called the system constant, is determined from the mea
sured re?ection coefficients and the known complex
permittivities for standards 1, 2, 3, and 4. Equation 3
may be rewritten, substituting the fourth standard for
the sample, to obtain:
(3A)
45 [ll Y4 - Y1 541632
542513
Substituting the expression given in Equation 5 for y,,,
which gives ,, in terms of 6'", for standards 1, 2 and 4,
and using 3= 00, Equation 3A may be solved for g to
obtain:
_ 641632
' 542513
and 6,; were de?ned earlier.
The values for E and i, are used in Equation 5 to solve
for e';, thus determining the complex permittivity of the
sample.
According to the method of the invention, the com
plex permittivity of a sample may be determined by
?nding a system constant for a given frequency using
re?ection data from four standards, having known com
 5,233,306
11 12 '
plex permittivities at the given microwave frequency, determined and compared with the corresponding data
measuring the re?ection coef?cient of the sample, and available in the literature. Since Equation 5 is quadratic
then determining the permittivity of the sample. The in 6,, mathematically two solutions of complex permit
standards are preferably an open circuit, a short circuit, tivity are found. However, the linear approximation of
water and methanol. The system constant does not need this equation (i.e. i=0) can be used at this point to
to be redetermined before each sample measurement, determine the correct permittivity from the two mathe
but should be re-determined at such intervals to prevent matical solutions.
instrument drift, changing temperatures or other chang FIG. 4 shows the complex relative permittivity of a
ing parameters from signi?cantly reducing accuracy. 0.1 N saline solution determined by the present inven
The system constant may also be determined for a tion at room temperature compared with values calcu
number of frequencies, or using a frequency scan. The lated by Equation 8, which was formulated in accor
order of the standard measurements does not matter, dance with the literature. See J. A. Saxton et. al., Elec
but the same frequency or frequencies should be used trical Properties of Sea Water-Reflection and Attenu
for each standard. Thus, it may be time et'?cient to ation Characteristics at V.H.F.," Wireless Engn, pp
collect reflection coef?cient data using standard 1 with 269-275, Oct. 1952.
a frequency scan, then standard 2 with an identical scan,
etc. In any event the sample should be tested at the same
72.72 (8)
frequencies that the calibrations were performed at. . _
e, - 4.9 + -
. (2l.O6)(lO9)
f
One bene?t of not having to redetermine the system (1 + j54.76) (10~'2)/
constant before each measurement is that in industrial 20
applications the probe may be scanned over a material The real component of the measured values, shown by
and determine the complex permittivity as-a function of , solid squares, is seen to be in close agreement with the
position on the material. Similarly, in a biomedical ap real component of the calculated values, shown by
plication the probe could be moved across a patients hollow squares. The left-hand vertical axis shows the
skin to determine the complex permittivity of the pa 25 relative magnitude of the real components and the hori
tients skin as a function of position. Such information is zontal axis is frequency in MHz. The imaginary compo
useful in locating tumors or other abnormal growths nent of the measured values, shown by solid circles, is
which may have complex permittivities different than also seen to be in close agreement with the imaginary
that of normal tissues. component of the calculated values, shown by hollow
The method may be performed using the apparatus 30 circles. The right-hand vertical axis shows the relative
disclosed above, or, alternatively, with conventional magnitude of the imaginary components. For example,
re?ection/transmission equipment. For example, the at the signal frequency of 700 MHz, the measured value
complex permittivity of a sample was successfully mea of 77-332 was found, which is in excellent agreement
sured using an HP-8620C sweeper main frame capable with the 78-j33 obtained from Equation 8.
of generating 16 MHz to 20 GHz signals as microwave 35 The complex permittivity of glycerol determined by
source 102. The HP-8620C sweeper main frame was the present method at 23 C. is shown in FIG. 5, using
connected to an HP-85044A transmission/re?ection the same symbols as FIG. 4. These results are also very
test set capable of operating from 300 kHz to 3 GI-Iz. An close to the values calculated by Equation 9 which was
I-IP-8508A vector voltmeter capable of operating from formulated in accordance with the literature. See Buck
300 kHz to 2,4 61-12 was connected to the HP85044A 40 ley et. al., Tables of Dielectric Dispersion Data for
transmission/re?ection set, as was a typical open ended Pure Liquids and Dilute Solutions, National Bureau of
coaxial probe having an outer diameter of 3.6 millime Standards Circular 598, Nov., 1958, Davidson et. al.,
ters. An HP-8510 automatic network analyzer was used Dielectric Relaxation in Glycen'ne, J. Chem. Phys.,
in place of the vector voltmeter for frequencies greater Vol. 18, pl4l7, 1950, and Morgan, Two Types of Di
than 2.4 GI-Iz. The equipment (available from Hewlett 45 electric Polarization, Trans. Amer. Electrochem. Soc.,
Packard Corporation) was used to gather the reflection Vol. 65, pp 109-118, 1934.
coefficient data. The data were manually read off of the
HP-8508A vector voltmeter or the HP-8510 automatic 38.32
network analyzer and the steps to provide the complex
permittivity were performed on a personal computer. 50
Referring now to FIG. 3, one method of the present where w is the angular frequency of the microwave
invention includes the steps of selecting a frequency of signal and e is the complex relative permittivity of
operation (step 300) and collecting reflection data for glyercol.
standards 1, 2, 3 and 4 (step 302). In step 304 E is calcu The examples given above are not intended to limit
lated according to Equation 7, and in step 306 A is cal the scope of the appended claims, but are for exemplary
culated. In steps 308 and 310, y] and y; are calculated, purposes only. As one skilled in the art will recognize,
respectively, according to Equation 5. In step 312 re various components and steps may be modi?ed or sub
?ection data for the sample is collected and A is calcu stituted and still remain within the scope of the inven
lated at step 314. The complex permittivity is deter tion.
mined at step 316, using the data collected in step 310 60 I claim:
and the values of E, A, and A, determined in steps 304, 1. A method for determining the permittivity of a
306, and 314, respectively, according to Equation 6. If sample using a coaxial probe having a ?rst end posi
the complex permittivity of other samples is to be mea tioned at the sample and a second end, comprising the
sured, the system returns to step 312. Each of steps steps of:
302-316 may be performed at a number of different 65 applying an AC electrical signal in the microwave
frequencies, if so desired. frequency range to the second end;
Using the method and apparatus described above, the measuring the resulting reflection coefficient of the
complex relative permittivities of several samples were sample;
 5,233,306
13 14
determining an admittance parameter of the sample third standard is a short circuit, and said fourth standard
and probe combination using the admittance pa is ethanol.
rameters of a pair of standards determined from 8. The method of claim 4, wherein the standards have
their known complex permittivities and a system dielectric constants different from each other and span
constant, and measured re?ection coefficients of 5 ning a range of values, and the dielectric constant of the
the sample and four standards having known com sample being tested falls within the range of dielectric
plex permittivities, wherein the system constant has constants de?ned by the standards.
been determined from the measured re?ection co 9. An apparatus for determining the permittivity of a
ef?cients of the four standards and the known com sample comprising: .
plex permittivities of the four standards; and a coaxial probe having a ?rst end positioned at the
determining the complex permittivity of the sample sample and a second end;
from the admittance parameter of the sample and means for applying an AC electrical signal in the
the system constant. microwave frequency range to the second end;
2. The method of claim 1 wherein said AC electric means for measuring the resulting re?ection coef?ci
signal has a frequency of between 100 MHz and 20 ent of the sample;
GHz. 15
means for determining an admittance parameter of
3. In a method for determining the permittivity of a the sample and probe combination using the admit
sample using a coaxial probe having a ?rst end posi tance parameter of a pair of standards determined
tioned at the sample and a second end, including the from their known complex permittivities and a
steps of applying an AC electrical signal in the micro system constant, and measured re?ection coef?ci
wave frequency range to the second end, measuring the 20 ents of the sample and four standards having
resulting re?ection coefficient of the sample, determin known complex permittivities, wherein the system
ing an admittance parameter of the sample and probe constant has been determined from the measured
combination, and determining the complex permittivity re?ection coef?cients of the four standards and the
of the sample from an admittance parameter of the known complex permittivities of the four stan
sample, the improvement which comprises: 25 dards; and _
determining a system constant g to determine com means for determining the complex permittivity of
plex relative permittivity 6, according to the equa the sample from the admittance parameter of the
tion: sample and the system constant.
10. The apparatus of claim 9, wherein said means for
30 determining the complex permittivity further comprises
wherein L represents the admittance parameter of ap means for solving ?S=e,'+e,'2, where e, is relative
erture admittance y;_. complex permittivity of the sample, y, is an admittance
4. A method for determining the permittivity of a parameter of the sample and E is a system constant.
sample using a coaxial probe having a ?rst end posi 11. An apparatus for determining the permittivity of a
tioned at the sample and a second end, comprising the 35 sample comprising:
steps of: a microwave source;
applying an AC electrical signal in the microwave ?rst and second directional couplers, each of said ?rst
frequency range to the second end; and second couplers having an input, a main output
measuring the re?ection coefficient of a ?rst, second, and an auxiliary output, wherein said input of said
third and fourth standard having known complex ?rst directional coupler is coupled to said micro
permittivities; wave source and said input of said second direc
determining a system constant from the measured tional coupler is coupled to said main output of said
re?ection coef?cients of the ?rst, second, third and ?rst directional coupler;
fourth standards and the known complex permit a coaxial probe having a ?rst end and a second end,
tivities of the ?rst, second, third and fourth stan 45 said ?rst end positioned near the sample and said
dards; second end coupled to said main output of said ?rst
determining an admittance parameter of the combina directional coupler;
tion of the probe and the ?rst standard, and the a local oscillator coupled to a ?rst and second mixer,
combination of the probe and the second standard, wherein said ?rst mixer is further coupled to said
from their known complex permittivities and the auxiliary output of said ?rst directional coupler and
system constant; said second mixer is further coupled to said auxil
measuring the re?ection coef?cient of the sample; iary output of said second directional coupler;
determining the admittance parameter of the sample ?rst and second analog to digital converters, wherein
and probe combination from the determined admit said ?rst analog to digital converters is coupled to
tance parameter of the ?rst and second standards 55 said ?rst mixer and said second analog to digital
and the measured re?ection coef?cients of the converters is coupled to said second mixer;
sample and the ?rst, second, third and fourth stan microprocessor means coupled to said ?rst and sec
dards; and ond mixers for receiving re?ection coef?cient data
determining the complex permittivity of the sample indicative of the sample and for determining the
from the admittance parameter of the sample and permittivity of the sample using the re?ection coef
the system constant. ?cient data.
5. The method of claim 4, wherein the step of apply 12. The apparatus of claim 10, further comprising a
ing an AC electric signal in the microwave range in display coupled to said microprocessor means.
cludes the step of applying an AC electric signal having 13. The method of claim 1, wherein the standards
a frequency of between 100 MHz and 20 GHz. have dielectric constants different from each other and
6. The method of claim 4, wherein each step is per 65 spanning a range of values, and the dielectric constant
formed at a plurality of frequencies. of the sample being tested fails within the range of di
7. The method of claim 4, wherein said ?rst standard electric constants defined by the standards.
i i t i
is an open circuit, said second standard is water, said