Progress In Electromagnetics Research, PIER 42, 131–142, 2003

MEASUREMENT OF COMPLEX PERMITTIVITY OF

LIQUIDS USING WAVEGUIDE TECHNIQUES

Y. Wang and M. N. Afsar

Department of Electrical and Computer Engineering
Tufts University
161 College Avenue, Medford, MA 02155, USA

Abstract—Complex permittivity of a number of liquids and binary

mixtures has been studied by measurement using the waveguide
techniques at the X and Ku band. Particular pieces of WR90
and WR62 waveguides were designed for the measurement of liquid
materials. The custom designed TRL calibration kits are applied
for calibration of the waveguide system. The measured results of
complex permittivity of liquid dielectrics, such as methanol, propyl
alcohol, ethyl alcohol, chlorobenzene, dioxane, cyclohexane and binary
mixtures, are presented. Particular pieces of open-ended waveguides
for the X and Ku bands were also designed for holding liquids and
the measured data using the open-ended waveguide technique were
compared with those measured using the waveguide technique. Some of
the measured results are also compared with calculated data using the
Debye equation and published data measured by the Fourier transform
spectroscopy.

1 Introduction
2 Analysis
2.1 Waveguide Technique
2.2 Open-Ended Waveguide Technique
3 Measurement
4 Measured Results
5 Conclusion
Acknowledgment
References
132 Wang and Afsar

1. INTRODUCTION

Accurate determination of complex permittivity data of liquid mat-

erials at microwave frequencies are necessary for applications such
as the evaluation of biological effects in biological molecules and in
solvents. Alcohols or other liquid dielectrics can be considered as good
solvent materials for a number of biological molecules, body tissues,
blood and bone marrow. Therefore, a thorough permittivity study of
various common solvents is essential for biomedical research. A number
of techniques has been developed which have the potential to be applied
for the measurement of complex permittivity of liquids such as the
waveguide technique [1–4], the open-ended waveguide technique [5, 6],
open-ended coaxial line technique [7, 8], dispersive Fourier transform
spectroscopy (DFTS) [9, 10] and microwave resonant method [11].
The waveguide technique and open-ended waveguide technique
are accurate for the measurement of complex permittivity. These
nondestructive techniques are based on the reflection coefficient and
transmission coefficient measurement from which the dielectric or
magnetic properties of materials can be determined. In addition,
these technique are well suited for the measurement of liquid solvent
materials. However, for the measurement of liquid materials, the
key problem is how to hold the liquid inside the waveguide for the
waveguide technique and outside the waveguide for the open-ended
waveguide technique.
In this paper, we employ the waveguide techniques for measuring
liquid dielectrics due to their high accuracy. Particular pieces of
liquid holder for WR90 and WR62 waveguides are designed for the
measurement of liquid materials. The complex permittivity of a
number of liquid solvents has been investigated experimentally using
the designed liquid holders at the X and Ku band. Particular pieces of
open-ended waveguides for the X and Ku bands are also designed for
holding liquids. The measured data using the waveguide technique
is compared with those measured using the open-ended waveguide
technique and with those from the literature.

2. ANALYSIS

2.1. Waveguide Technique

Fig. 1 shows the designed liquid holder and the measurement setup for
the waveguide technique. The liquid holder consists of WR62 or WR90
waveguide with a shallow pocket at both openings. This allows for a
thin sheet of Kapton polyamide film to be placed in the shallow pocket
and sealed with an epoxy adhesive. Threaded access holes are placed
Measurement of complex permittivity of liquids 133

Figure 1. Designed liquid holder and measurement setup for the

waveguide technique.

at the top and bottom of the holder to be used to inject and remove
the test liquid. These holes are then secured from liquid leakage using
sealing screws.
By the transmission line theory, the transmission coefficient S21
can be expressed by S21 = (Γ +Γ )2 eΓ1 L4Γ−(Γ 1 Γ0
2 −Γ1 L , where Γ0 =
1 0 1 −Γ0 ) e
jβ0 and Γ1 = α1 + jβ1 . Here the propagation coefficient of air
 1/2
filled waveguide β0 = 2π λ0 1 − ( λ0 2
λc ) , the propagation coefficient
 1/2
of loaded waveguide α1 + jβ1 = 2π λ0 ( λ0 2
λc ) − ε r , λ0 and λc are the
wavelength of free space and the cutoff wavelength of the waveguide
for TE10 mode, and εr = ε − jε is the relative complex dielectric
constant. The S21 is a function of complex permittivity of the
specimen, therefore, once the measured transmission coefficient S21 are
obtained, the complex permittivity of the specimen can be extracted
by an optimization procedure such as the Newton-Raphson method.

2.2. Open-Ended Waveguide Technique

For comparison purpose, the open-ended waveguide technique was
also developed following the procedures presented in [5]. However,
the common used open-ended flange is flat, so it is difficult to hold
liquids. In this work, two rectangular flanges with stepped walls for
WR90 and WR62 waveguides, respectively, were designed for holding
liquid materials. Fig. 2 shows the stepped flange designed for open-
ended waveguide technique. The aperture size of the liquid holders
134 Wang and Afsar

Figure 2. The stepped flange designed for open-ended waveguide

technique.

is 150 mm×150 mm and the depth of each stepped wall is 5mm. The
stepped walls were designed for the liquid holders so that the liquid
samples can be measured at different thickness. The technique is
based on the reflection coefficient S11 measurement from which the
dielectric properties of materials can be determined. The reflection
coefficient from the open-ended waveguide can be calculated by various
methods. In this paper, the reflection coefficient is derived by solving
the boundary value problem using a Rayleigh-Ritz technique, following
the procedures presented in [5]. The reflection coefficient of a slab-like
sample backed by a conducting plate can be expressed by S11 = 1−y 1+y ,
jab        
where y = 2k1 D2 S  S f (x, y)f (x , y )G(x, y; x , y )dxdydx dy , D =
 
λ ) − ( a ) , a and b are the dimensions
πx
s f (x, y) sin a dxdy, k1 = ( 2π 2 π 2

of the rectangular waveguide, λ is the wavelength in free space,

G(x, y; x , y  ) is the Green function valid in the dielectric medium
outside of the waveguide which implicitly contains the permittivity
of the liquid, S or S  is the area of the waveguide aperture, and
f (x, y) is the trial function in the waveguide aperture plane which
can be expanded over the completed set of the orthonormal waveguide
modes. The complex permittivity of the sample can be determined
by minimizing the difference between the calculated and measured
reflection coefficients.

3. MEASUREMENT

The waveguide system is connected with Agilent 8510C Vector Network

Analyzer for measuring the S-parameters. To calibrate the waveguide
system, a few calibration techniques can be employed that include
three-reference standards, waveguide calibration kits or thru-reflect-
line (TRL) calibration kits. The procedure using three-reference
standards works as follows. First, following the one port S11 calibration
Measurement of complex permittivity of liquids 135

procedure of open-short-load, calibrate the coaxial line that connects

the network analyzer and the coax-to-waveguide adapter. Then,
calibrate the waveguide using three references such as a matched load
and two offset short circuits or a sliding short circuit [6]. A waveguide
calibration kit is normally composed of two offset short circuits with
known depths and a matched load. These calibration circuits are
connected to the waveguide following the procedures of the calibration,
respectively, and the calibrated results are sound. The waveguide
calibration kit can be custom designed using VNA Cal Kit Manager.
The TRL calibration technique is applied for the work described in this
paper because it is the most promising method. The TRL calibration
kit is custom designed using VNA Cal Kit Manager. The kit simply
consists of a short circuit and a waveguide line section with a known
length. The line section length should not be greater than half a
waveguide wavelength. It is chosen to be a quarter nominal waveguide
2(λ ×λgh )
wavelength λgn = (λglgl+λgh ) , where λgl and λgh are wavelengths of the
low frequency end and the high frequency end at the corresponding
frequency band, respectively. The accuracy of these standards is
excellent because it is not difficult to make an accurate short circuit and
a quality piece of waveguide in terms of its accuracy of characteristic
impedance. A thru, a short for each port and a delay line were applied
but the isolation was omitted in this measurement.
After the TRL calibration is performed on the waveguide ports,
for the thru case, it was observed that the amplitude variation of S21
is less than −0.01 dB and the phase is less than 200 m◦ for both the
X-band and Ku band. The effect of the designed liquid holder (air
filled) on the S21 was studied by comparing the S21 measured using a
same length waveguide (air filled). It was observed that the decay and
phase shift due to the particular design is less than −0.1 dB and less
than 3◦ , respectively. The decay and phase shift are attributed from
the thin dielectric sheet sealing both waveguide openings of the liquid
holder.

4. MEASURED RESULTS

The measurements were carried out at the X and Ku-band at room

temperature 23◦ C. Fig. 3 shows the measured permittivity of ethyl
alcohol at the X-band using the waveguide technique. Two types of
holders were used, the designed liquid holder and a piece of waveguide
that is sealed with Scotch tape. It is seen that the measured results
using the designed liquid holder agree reasonably well with those
measured using a piece of waveguide containing the liquid sample
sealed by the Scotch tape. However, the designed liquid holder is
136 Wang and Afsar

Figure 3. Permittivity of ethyl alcohol measured using the designed

waveguide liquid holder and Scotch tape at the X-band.

Figure 4. Permittivity of methanol measured using the designed

waveguide liquid holder and calculated by the Debye equation at the
X-band.
Measurement of complex permittivity of liquids 137

Figure 5. Permittivity of propyl alcohol measured using the designed

waveguide liquid holder and Scotch tape at the X-band.

Figure 6. Permittivity of propyl alcohol measured using the designed

waveguide liquid holder and the designed stepped flange for open-ended
waveguide technique at the Ku-band.
138 Wang and Afsar

Figure 7. Permittivity of chlorobenzene measured using the designed

waveguide liquid holder at the X- and Ku-band.

more stable and convenient than the Scotch tape because the liquids
may resolve the Scotch tape and this makes the Scotch tape unstuck.
Fig. 4 shows the complex permittivity of methanol measured using
the designed waveguide liquid holder at the X-band. It is seen that
the measured result agrees quite well with the calculated data using
the Debye equation [7]. Fig. 5 shows the measured permittivity of
propyl alcohol at the X-band together with the data measured using
the Scotch tape.
Fig. 6 shows the measured permittivity of propyl alcohol at the
Ku-band by using the designed waveguide liquid holder, compared with
the measured data using the designed stepped flange for the open-
ended waveguide technique. It is seen that the measured results by the
waveguide technique agree well with those measured using the open-
ended waveguide technique. It is also observed that the permittivity of
propyl alcohol at the overlapping frequency 12.4 GHz measured at the
Ku-band agrees well with that measured at the X-band. It should be
mentioned by our experimental experience that the designed waveguide
sample holders for the waveguide technique are easier to handle and
more reliable than the designed stepped flanges for the open-ended
waveguide technique. It is sometimes difficult to extract the complex
permittivity for the open-ended waveguide technique because the open-
Measurement of complex permittivity of liquids 139

Figure 8. Complex permittivity of binary mixture of ethyl alcohol

and water measured using the designed waveguide liquid holder at the
Ku-band (a) real part (b) imaginary part.
140 Wang and Afsar

Figure 9. Permittivity of dioxane and cyclohexane measured using

the designed waveguide liquid holder at the X-band.

ended waveguide technique is more sensitive to the experimental

errors than the waveguide technique. It is noted that the complex
permittivity is extracted from the reflection coefficient S11 for the open-
ended waveguide technique while the complex permittivity is extracted
from the transmission coefficient S21 for the waveguide technique. The
S11 is generally more difficult to be accurately measured than the S21 .
Fig. 7 shows the measured permittivity of chlorobenzene at the X-
and Ku-band using the designed waveguide liquid holder. It is seen that
the continuity between the two bands is quite good. The result agrees
with that measured by the dispersive Fourier transform spectrometry
in a higher frequency range. For example, the dielectric constant
of chlorobenzene is approximately 3.61 at 15 GHz [9]. Fig. 8(a) and
Fig. 8(b) show the measured real part and imaginary part of complex
permittivity of binary mixture of ethyl alcohol and distilled water at
the Ku-band, respectively. The difference of complex permittivity due
to the various percentages of the binary mixture is clearly shown,
and this is expected to be used for identifying the contents of binary
mixtures. Finally, Fig. 9 shows the complex permittivity of two low
loss liquids, dioxane and cyclohexane, measured using the designed
waveguide liquid holder at the X-band. This result is consistent
with that measured at higher frequencies [10], demonstrating that the
waveguide technique is equally effective for low loss liquids.
Measurement of complex permittivity of liquids 141

5. CONCLUSION

Complex permittivity of a number of liquid dielectrics such as metha-

nol, propyl alcohol, ethyl alcohol, chlorobenzene, dioxane, cyclohexane
and binary mixtures, have been measured using the waveguide tech-
niques at the X and Ku band. The custom designed TRL calibration
kits are applied for the calibration of the waveguide system. Particular
pieces of waveguides for both the waveguide technique and open-
ended waveguide technique were designed for holding the liquids. The
measurement of complex permittivity for the liquids manifested the
dielectric properties of the liquid materials. This work demonstrates
that the designed liquid holders are effective for the measurement
of complex permittivity of low loss, medium loss and lossy liquids.
The research work conducted in this paper is expected for microwave
biomedical applications.

ACKNOWLEDGMENT

The authors would like to thank Anthony N. Andreucci for design of

the waveguide sample holder and Rene Grignon for his assistance in
the measurement.

REFERENCES

1. Ball, J. A. R. and B. Horsfield, “Resolving ambiguity in broadband

waveguide permittivity measurements on moist materials,” IEEE
Trans. Instrum. Meas., Vol. 47, No. 2, 390–392, April 1998.
2. Hallikainen, M. T., F. T. Ulaby, M. C. Dobson, M. A. El-Rayes,
and L. K. Wu, “Microwave dielectric behavior of wet soil-part
I: empirical modes and experimental observation,” IEEE Trans.
Geosci. Remote Sensing, Vol. 23, No. 1, 25–34, Jan. 1985.
3. Weir, W. B., “Automatic measurement of complex dielectric
constant and permeability at microwave frequencies,” Proceedings
of the IEEE, Vol. 62, No. 1, 33–36, Jan. 1974.
4. Huygen, I., C. Steukers, and F. Duhamel, “A wideband line-line
dielectric method for liquids, soils and planar substrates,” IEEE
Trans. Instrum. Meas., Vol. 50, No. 5, 1343–1348, Oct. 2001.
5. Decreton, M. C. and M. S. Ramachandraiah, “Nondestructive
measurement of complex permittivity for dielectric slabs,” IEEE
Trans. Microwave Theory Tech., Vol. 23, 1077–1080, Dec. 1975.
6. Chang, C. W., K. M. Chen, and J. Qian, “Nondestructive de-
termination of electromagnetic parameters of dielectric materials
142 Wang and Afsar

at X-band frequencies using a waveguide probe system,” IEEE

Trans. Instrum. Meas., Vol. 46, No. 5, 1084–1092, Dec. 1997.
7. Misra, D., M. Chabbra, B. R. Epstein, M. Mirotznik and
K. R. Foster, “Noninvasive electrical characterization of materials
at microwave frequencies using an open-ended coaxial line: test
of an improved calibration technique,” IEEE Trans. Microwave
Theory Tech., Vol. 38, No. 1, 8–14, Jan. 1990.
8. Bao, J., M. L. Swicord, and C. C. Davis, “Microwave dielectric
characterization of binary mixtures of water, methanol and
ethanol,” Journal of Chemical Physics, Vol. 104, No. 12, 4441–
4450, 1996.
9. Afsar, M. N., J. Chamberlain, G. W. Chantry, R. Finny, and
R. Van Loon, “Assessment of random and systematic errors in
microwave and submillimeter dielectric measurements,” Proc. of
IEE, Vol. 124, No. 6, 575–577, 1977.
10. Afsar, M. N., J. Chamberlain, and G. W. Chantry, “High precision
dielectric measurements on liquids and solids at millimeter and
submillimeter wavelength,” IEEE Trans. Instrum. Meas., Vol. 25,
No. 4, 290–294, Dec. 1976.
11. Yu, K. B., S. G. Ogourtsov, V. G. Blenky, A. B. Maslenikov, and
A. S. Omar, “Accurate microwave resonant method for complex
permittivity measurements of liquids,” IEEE Trans. Microwave
Theory Tech., Vol. 48, No. 11, 2159–2164, Nov. 2000.