The Step by Step Development of NRW Method

Alexandre Natã Vicente Gustavo Maciulis Dip Cynthia Junqueira

Division of Electronic, AEL Division of Electronic, AEL Division of Electronic, AEL
Institute of Aeronautics and Space, IAE Institute of Aeronautics and Space, IAE Institute of Aeronautics and Space, IAE
São José dos Campos, Brazil São José dos Campos, Brazil São José dos Campos, Brazil
[email protected] [email protected] [email protected]

Abstract— This paper explains in details the NRW method iterative, simple, and able to characterize dielectric and
highlighting its mathematical procedure and analyzing the most magnetic materials. This method can be applied to TEM
important points and limitations to characterize the complex waves, normally related with coaxial cables or TE and TM
permittivity and permeability of dielectrics. The step by step waves, applied with waveguides. Here, we will present the
procedure help the researchers to well understand the physical development of the expressions for TE waves, but all equations
phenomenona involved and also is a tool to facilitate its can be used for TEM waves if the cutoff wavelength
implementation. Explanations on the limitations of the method in approaching infinity ሺߣ௖ ՜ λሻ is considered.
situations where the return loss tends close to zero and the
disambiguation phase using two samples are detailed. The non- This paper comprises the Nicolson-Ross-Weir method
iterative method was evaluated as a candidate to characterize (NRW) that will be detailed discussed in section II. The
materials for radome applications. equation development due reflection and transmission
phenomena are presented in subsections A and B; the phase
Keywords— electric permittivity, magnetic permeability, radome ambiguity, in subsection C, and the divergence of the NRW
material, Nicolson-Ross-Weir measurement technique. method when the sample length is a integer length multiple of
half guided wavelength are described in subsection D. Finally
I. INTRODUCTION the concluding remarks are in section III.
In the early 1980’s, several USA companies were
competing for business sales on military and civil programs. II. THE METHOD NICOLSON-ROSS-WEIR
This competition fostered innovations in technology such that The Nicolson-Ross-Weir method (NRW) [2][3] is based on
internal costs could be lowered on each generation of the phenomenon of reflection and transmission of a certain
development. One technology area that the companies targeted material whose characteristics we want to know. These
for cost reduction was the ceramic radome. Different materials phenomena occur because of the difference between the
and processes were developed and many materials used in complex permittivity () and the complex permeability () of
electromagnetic applications, as antennas and radomes are the two environments involved (inside and outside the
lossy and need to be well characterized. The accurate material). As we get to know the effects of this phenomenon
measurement of lossy dielectric materials is challenging since and of the incident, reflected and transmitted waves it becomes
many resonant techniques lose sensitivity when applied to such possible to obtain the originating parameters of these effects (
materials and transmission-line methods are strongly e ). Most of the times, what we obtain aren’t the incident,
inuenced by metal losses. The goal was to assemble and reflected and transmitted waves but rather the scattering
know the relevant information needed to perform and matrix, which then gives us the values of those waves.
understand dielectric measurements on lossy materials. For
this, topics such as electromagnetics, circuit theory, related The literature generally presents the NRW algorithm as a
physics, measurement algorithms, and uncertainty analysis set of equations, without the important intermediate steps,
were extensively studied and reported in the literature [1], [4] usually separated from the explanation of the basis of
and [5]. transmission/reflection method bases resulting in a further
obstacle to the understanding of the method. In order to
There are several methods to obtain the electrical overcome this obstacle, the step by step procedure will be done
permittivity () and permeability () in non conductor materials in the following sequence: explanation of the equations related
using scattering parameters ൫ܵ௜௝ ൯ , and can be cited by the with the reflection and transmission phenomena, highlighting
Nicolson-Ross-Weir (NRW) algorithm [2][3], effective the equations system resolution utilized by Nicolson-Ross-
parameter method [1], nonlinear least-squares solution [1], Weir, presenting one possible ambiguity solver method using
free-space method [1][5] and Courtney method also called two samples and analyzing the half-wavelength divergence.
Hakki and Coleman method [1], amongst others.
A. The equations development due reflection and
Here, explanations will be given about the NRW method in transmission phenomena
a gradual way in order to facilitate the understanding and The main idea is to propose a scenario where we have a
explore its advantages as to be an analytical method, non- material under test (MUT) inside a wave guide or a coaxial air

978-1-4577-1664-5/11/$26.00 ©2011 IEEE 738

line, to verify the phenomena of reflection and transmission, as ૚ାડ ࢽ
ࣆ࢘ ൌ ቀ ቁ (9)
shown in the Fig. 1. Thus the problem to be solved is to obtain ૚ିડ ࢽ૙
 and ߤ complexes from the data that can be measured in the Following the next steps, (9) can be rewritten as shown in
samples, ie, measures of mechanical setup, such as the length (12), which is widely found in the literature. Based on (4), one
of the air gap between the calibration plan 1 until the sample ఊ
can obtain the ratio
ሺ‫ܮ‬ଵ ሻ, the length of the air gap between the calibration plan 2 ఊ೚
until the sample ሺ‫ܮ‬ଶ ሻ , the sample length ሺ‫ܮ‬ሻ , the physical
ഘమ ഋೝ ഄೝ మഏ మ
dimensions of the sample holder, the radii of the internal and ௝ඨ ିቀ ቁ
ఊ ೎మ ഊ೎
ఊ
external conductor, for coaxial air line, or the width and height ൌ ൌ భ భ
(10)
ఊ೚ ௝ଶగට ି
for rectangular waveguide, and the scattering parameters ഘ మ మഏ మ
௝ ඨቀ ቁ ିቀ ቁ ഊబ మ ഊ೎ మ
೎ ഊ೎
൫ܵ௜௝ ሺ݂ሻ൯, that is frequency dependent.
The first step in solving the problem is to explicit the Defining
relationship between the reflection coefficient () and the ଵ ఊ
ؔ (11)
complex propagation constant (ߛ) with the  e  of the MUT ஃ ௝ଶగ
[4], assuming that the material is not conductive (conductivity And replacing (11) in (10)
 = 0), and that the electromagnetic wave is transverse electric,
૚ାડ ૚
TE, or transverse electromagnetic, TEM. Analogous analysis ૄ‫ ܚ‬ൌ ቀ ቁ ૚ ૚
(12)
૚ିડ ઩ට ି
can be made for transverse magnetic waves, TM [5]. ઻૙ ૛ ઻‫ ܋‬૛

The complex propagation constant, ߛ, in the waveguide is a At this point, the relationship between ρ୰ and  and ߛ is
function of the cutoff wavenumber‫ܭ‬௖ and of the wavenumber explicit. The following will explicit the relationship between ߝ୰
‫ܭ‬. and  and ߛ. From (1), (2) and (3)
઻૛ ൌ ۹ ‫ ܋‬૛ െ ۹ ૛ (1) ߛ ଶ ൌ ሺʹߨሻଶ ቂ
ଵ
െ
ఓೝ ఌೝ
ቃ (13)
ఒ೎ మ ఒబ మ
ଶ஠
ୡ ൌ (2)
஛ౙ Solving for ߝ୰
 ൌ ɘξɂρ (3) 1
0 2 ቈ 2 -ቀ2ቁ ቉
 2
 c
Where ߣ௖ is the cutoff wavelength, ɘ is the angular r =
r
(14)
frequency. Considering the wave propagation in a given way,
(1) can be rewritten as Following the next steps, (14) can be rewritten as shown in
(16), which is widely found in the literature. Squaring both
னమ ρ౨ க౨ ଶ஠ ଶ sides of (11)
ɀ ൌ Œට െቀ ቁ (4)
ୡమ ஛ౙ
ଵ ఊమ
ൌ െ ሺଶగሻమ (15)
The wave impedance of TEM or TE waves is given by ஃమ

୨னρ Finally, the substitution of (15) into (14) yields

Ʉൌ (5)
ஓ
ૃ૙ ૛ ૚ ૚
The reflection coefficient for a propagation wave from ઽ‫ ܚ‬ൌ ቂ ൅ ቃ (16)
ૄ‫ܚ‬ ૃ‫ ܋‬૛ ઩૛
material 1 to material 2 can be obtained from the boundary
condition of continuality of tangential components of electric The analysis to this stage showed the expressions of ߝ e ߤ
and magnetic fields across the interface, and is given by as a function of Ȟ and ߛ. The next step aims at clarifying the
஗మ ି஗భ
relationship between  and ߛ, which carries the information of
Ȟൌ (6) the incident, reflected and transmitted waves, and the scattering
஗మ ା஗భ
matrix. For the conditions of problem under study (TE or TEM
The substitution of (5) into (6) yields waves and  = 0), one can write the equations of the electric
ഋ ഋ೚
ି
fields shown in Fig. 1 [1][5].
ം ം೚
Ȟൌഋ ഋ (7)
ା ೚
ം ം೚

Where ߛ௢ is the complex propagation constant in the air.

Then
ംబ
ఓೝ ିଵ
Ȟ ൌ ംംబ  (8)
ఓೝ ାଵ
ം

Solving for ߤ௥ Figure 1. – Electromagnetic waves transmitting through and reflect from a
sample in a transmission line [1]

978-1-4577-1664-5/11/$26.00 ©2011 IEEE 739

 ሬሬሬԦ
୍ ൌ ሺ‡ିஓబ୶ ൅ ଵ ‡ஓబ୶ ሻ›ො (17) The substitution of (25) into (24), followed the correction
yields
ሬሬሬሬԦ
 ୍୍ ൌ ሺଶ ‡
ିஓ୶
൅ ଷ ‡ஓ୶ ሻ›ො (18)
ડሺ૚ି‫ ܈‬૛ ሻ
ሬሬሬሬሬሬԦ ‫܁‬૚૚ ൌ ‫ ܀‬૚ ૛ (26)
୍୍୍ ൌ ሺସ ‡ିஓబ୶ ሻ›ො (19) ૚ିડ ૛ ࢆ૛

Assuming normal incidence in the material, using the ray This occurs becauseȞ is defined at interface as
tracing method or the method of impedance [6] it is possible to ாೝ೐೑೗೐೎೟೐೏
relate the coefficient of  and ɀ with the scattering parameters Ȟൌ (27)
ா೔೙೎೔೏೐೙೟
൫ܵ௜௝ ൯. Additional remarks suppose that the material is isotropic, When the wave propagates a distance ୧ in a material with
so, ଶଵ ൌ ଵଶ . Following, the ray tracing method is used. propagation constant , we correct its amplitude and phase by
݁ ఊబ ୐౟ ‫ܧ‬௜௡௖௜ௗ௘௡௧ . (28)
Each term of the reflected wave at a distance ୧ of the
interface is given by
‫ܧ‬௥௘௙௟௘௖௧௘ௗ ݁ ିఊబ ୐౟ (29)
Dividing the (29) by (28) and computing the sum of the
reflected electric field, we obtain
Figure 2. – Multiple reflection at three materials interfaces σ ாೝ೐೑೗೐೎೟೐೏ ୻ሺଵି୞మ ሻ
݁ ିଶఊబ୐భ  = ଵଵ ൌ ଵ ଶ (30)
ா೔೙೎೔೏೐೙೟ ଵି୻మ ௓ మ
Since the incident wave has a unit amplitude, material 1 is
equal to material 3 and material 2 is the MUT, from Fig. 2, we Similarly for ଶଶ , we obtain
find that the sum of reflected fields at the interface with the ‫܀‬૛ ૛ ડ൫૚ି‫ ܈‬૛ ൯
sample (Ȟ୧୬ ሻ is given by ‫܁‬૛૛ ൌ (31)
૚ିડ ૛ ࢆ૛
௡ ௡ ିଶఊ௅௡
Ȟ୧୬ ൌ  Ȟଵଶ ൅ ଵଶ ଶଵ Ȟଶଷ ‡ିଶஓ୐ σஶ
௡ୀ଴ Ȟଶଵ Ȟଶଷ ݁ (20) Alternatively, the equations of ଵଵ and ଶଶ can be obtained
by the impedances transfer equation, well explained in
Where Ȟ୧୨ and ୨୧ are, respectively, the reflection coefficient
reference [6].
and the transmission coefficient of the wave that propagates
from material “i” to material “j”. This sum is a well known For ଶଵ ǡ remembering that, in our case ଶଵ ൌ ଵଶ , we can
geometric progression. Then the equation can be rewritten as apply similar reasoning, than
௡ ௡ ିଶఊ௡௅
୘భమ ୘మభ ୻మయ ୣషమಋై ସ ൌ ଷଶ ଶଵ ݁ ିఊ௅ σஶ
௡ୀ଴ Ȟଶଵ Ȟଶଷ ݁ (32)
Ȟ୧୬ ൌ  Ȟଵଶ ൅ (21)
ଵି୻మభ ୻మయ ୣషమಋై
Again, we have a well known geometric progression, so the
At the traditional setup, material 1 and material 3 are air. equation can be rewritten as
Based on (6), we can conclude that
୘యమ ୘మభ ௘ షംಽ
Ȟଵଶ ൌ െȞଶଵ ൌ െȞଶଷ ସ ൌ  (33)
(22) ଵି୻మభ ୻మయ ୣషమಋై

Defining the electric field incident at interface  , From (22), (23) and (33) we can obtain
propagating from material “i” to material “j”, the reflected field ൫ଵି୻మ ൯௘ షംಽ
Ȟ୧୨ , and the transmitted field ୧୨ , applying the boundary ‫ܥ‬Ͷ ൌ (34)
ଵି୻మ ௘ షమംಽ
condition of continuality of tangential component of electric
The scattering parameter ଵଶ measured at reference planes
fields across the interface then
showed at Fig. 1 is given by a correction in (34). The
୨୧ ൌ  ൅ Ȟ୧୨ substitution of (25) into (34), followed the correction yields
୨୧ ൌ ͳ ൅ Ȟ୧୨ (23) ‫܀‬૚ ‫܀‬૛ ൫૚ିડ ૛ ൯ࢆ
‫܁‬૛૚ ൌ (35)
૚ିડ ૛ ࢆ૛
From (21), (22) and (23) we can obtain
B. Core NRW method: solving the system of equations
୻భమ ሺଵିୣషమಋై ሻ
Ȟ୧୬ ൌ (24) It is evident that to determine the complex permittivity and
ଵି୻మభ మ ୣషమಋై
permeability of the material, one only needs to solve for Ȟ and
The scattering parameter ୧୧ measured at a distance ୧ from ߛ the equations that relate the scattering parameters with Ȟ and
the sample (Fig. 1) is given by a correction in (24). Defining ߛ (in (26), (31) and (35)). The NRW method consists in
solving this system of equations. After that, to obtain the
 ൌ ‡ିஓ୐ , Ȟ ൌ Ȟଵଶ ,  ୧ ൌ ݁ ିఊబ ௅೔ ǡ ݅ ൌ ͳǡʹ (25) complex permittivity and permeability the solution is applied
into (9) and (14).

978-1-4577-1664-5/11/$26.00 ©2011 IEEE 740

 Changing the reference plane from calibration plans for the initial frequency, then, after each ߶௓ jump, the value of n
ends of the MUT: Do this by removing the correction factors increases. But for this way, know that it may not always be
entered,  ୧ . possible to have such a low frequency (when you use
ୗభభ ି୻ሺଵି୞మ ሻ
waveguides). Finally you can use two samples of different
 େଵଵ ൌ ൌ (36) sizes. In this paper you will be presented with a solution for
ୖభ మ ଵି୻మ ୞మ
using two samples of different lengths.
ௌమభ ௓൫ଵି௰ మ ൯
ܵ ஼ ଶଵ ൌ ൌ (37) C. Disambiguation phase method: A method of two samples
ோభ ோమ ଵି௰ మ ௓ మ

Defining the auxiliary variables to solve the system As a general rule in an ambiguity problem, the solution is
based on more information to resolve an ambiguity. In this
ଵ ൌ  େ ଶଵ ൅  େଵଵ (38) method, additional information is coming from another sample
ଶ ൌ  େ
െ େ
(39) of different length.
ଶଵ ଵଵ

Then Samples of different sizes have phase variations with

different frequency, usually generating phase jumps in
ଵି୚భ ୚మ ୗి భభ ିୗి మభ ାଵ different frequencies and usually generating different
ൌ ൌ (40)
୚భ ି୚మ ଶୗి భభ ɂ୰ ሺǡ ˆሻƒ†ρ୰ ሺǡ ˆሻ, for each sample. For a given frequency,
Ȟ ଶ െ ʹȞ ൅ ͳ ൌ Ͳ (41) ˆ଴ , ther is only one correct value of ɂ୰ ƒ†ρ୰ , so, if we
compare ɂ୰ఽభ ሺ୅ଵ ǡ ˆ଴ ሻƒ†ρ୰ఽభ ሺ୅ଵ ǡ ˆ଴ ሻ , Ͳ ൑ ୅ଵ ൑ ୤ , ୤
Ȟൌേ ξ ଶ െͳ (42) arbitrary but large enough to ensure that the right “n” is smaller
The quadratic equation of complex constant coefficients than “ ୤ ”, obtained from one sample, ଵ ǡ with
has two solutions. To determine the correct solution for a ɂ୰ఽమ ሺ୅ଶ ǡ ˆ଴ ሻƒ†ρ୰ఽమ ሺ୅ଶ ǡ ˆ଴ ሻ , ሻ , Ͳ ൑ ୅ଶ ൑ ୤ , obtained
passive medium, we use physical restriction: Ȟ ൑ ͳǤ From from another sample, ଶ , there is only one ୅ଵ and one
(36) and (37) one can also show that ୅ଶ such that ɂ୰ఽభ ሺ୅ଵ ǡ ˆ଴ ሻ ؆ ɂ୰ఽమ ሺ୅ଶ ǡ ˆ଴ ሻƒ†ρ୰ఽభ ሺ୅ଵ ǡ ˆ଴ ሻ ؆
ௌమభ ಴ ାௌభభ ಴ ି௰
ρ୰ఽమ ሺ୅ଶ ǡ ˆ଴ ሻ, with these “n” the correct for the disambiguation
ܼൌ ൌ ݁ ିఊ௅ (43) phase. This comparison must be made by any metric. One
ଵି൫ௌమభ ಴ ାௌభభ ಴ ൯௰
possible metric is the following
Solving (43) for ߛ, the phase ambiguity problem arises. The ଶ
measurement of phase of ଶଵ and of ଵଵ are ambiguous, since ‫ܴܱܴܴܧ‬ሺ݂଴ ሻ ൌ ൣߝ ᇱ ௥ಲభ ሺ݊஺ଵ ǡ ݂଴ ሻ െ ߝ ᇱ ௥ಲమ ሺ݊஺ଶ ǡ ݂଴ ሻ൧ ൅
the network analyzer is only able to detect a phase between െߨ ଶ
ൣߝ ᇱᇱ ௥ಲభ ሺ݊஺ଵ ǡ ݂଴ ሻ െ ߝ ᇱᇱ ௥ಲమ ሺ݊஺ଶ ǡ ݂଴ ሻ൧ ൅ ቂߤ ᇱ ௥ ሺ݊஺ଵ ǡ ݂଴ ሻ െ
and ൅ߨ. Moreover, the operations of addition and subtraction ಲభ
ଶ
ᇱ ଶ ᇱᇱ ᇱᇱ
are only defined for a domain of length equal to one period, i.e. ߤ ሺ݊ ǡ ݂ ሻ ൅ ቂߤ
௥ಲమ ஺ଶ ଴ ൧ ௥ಲభ
ሺ݊஺ଵ ǡ ݂଴ ሻ െ ߤ ሺ݊ ǡ ݂ ሻቃ )
௥ಲమ ஺ଶ ଴
(45)
ʹߨ . So even if the phase measurements would not be
ambiguous, the ambiguity would still exist. Thus solving (43) The pair ሺ݊஺ଵ ǡ ݊஺ଶ ሻ that provides the smallest ERROR for a
given frequency is considered correct for the disambiguation
૚
઻ ൌ Ǥ ቀ‫ܖܔ‬ሺ۷‫܈‬۷ ି૚ ሻ െ ‫ܒ‬ሺ૖‫ ܈‬൅ ૛‫ૈܖ‬ሻቁǡ ‫ א ܖ‬Գ (44) phase.
‫ۺ‬

Mathematically, “n” could be any integer, but to have a D. Analysis of the divergence of the NRW method when the
physical sense, the imaginary part of ߛ should be negative, sample has a length multiple of half guided wavelength
since it is assumed that the wave propagates in the direction of This section will briefly show that for lossless materials,
+x, as Fig. 1, then, considering that Ԅ୞ ‫ א‬ሾെɎǡ Ɏሾ, “n” should ஛
when the length of MUT is a multiple of ଶౝ, where ɉ୥ is the
be natural, including zero. An important physical constraint guided wavelength, then ଵଵ  ൌ Ͳ. Although the materials of
can be checked at this point, it is: ‫݈ܽ݁ݎ‬ሺߛሻ ൒ Ͳ, for passive interest are not lossless, remains valid the fact that the closer
material. ஛
the thickness is a multiple of ଶౝ, ଵଵ  is closer to zero.
This phase ambiguity is the main problem of the NRW
method, since this is supposed to be an analytical resolution.
The solution isn’t entirely analytical because of this ambiguity,
meaning that the NRW method raises a nearly closed formula
to obtain the material’s complex permittivity and permeability.
There are several ways to solve this problem. The most
ఒ
common is to use a very thin sample, in this case ‫ ܮ‬൏ ଶ , so
݊ ൌ Ͳ. Another possible way would be to use the group delay,
comparing the group delay calculated and the measured [3].
You can also start with a frequency closer to zero and small Figure 3. – The results of permittivity as a function of frequency using the
ఒ
frequency steps, in this case you may consider ‫ ܮ‬൏ ଶ for the Nicolson-Ross algorithm (solid line) [5]

978-1-4577-1664-5/11/$26.00 ©2011 IEEE 741

 When ଵଵ  ՜ Ͳǡ that is, when the signal power drops effects of uncertainties in the measurements. The NRW
drastically, the signal noise ratio severely deteriorates, causing method is also important because it is relatively simple, but is
inconsistent results that are displayed as peaks in the complex used as the basis for more complex methods; therefore, a deep
permittivity and permeability obtained, as showed in the Fig. 3. knowledge is vital to a perfect understanding of them.
To minimize uncertainties in low loss materials, one can use
One of the contributions of this work is to highlighted the
longer samples [5]. The ideal length of the sample depends on
two most important weaknesses of the NRW method: the
the setup. A rule to minimize the uncertainty would be to use
ambiguity of the phase and the divergence of the method when
sample length equal to ሺʹ݊ ൅ ͳሻȀͶǡ ݊ ‫ א‬Գ [8]. ఒ
the sample has a length multiple of ଶ೒ . The problem of
To demonstrated that ଵଵ  ൌ Ͳ, only if the length of MUT ambiguity phase arises due to limitations of measurement and
஛
is a multiple of ଶౝ, and MUT is lossless, equates (26) to zero, mathematics used. The most common method used to solve
then this problem is the method of group delay, but here was
୻ሺଵି୞మ ሻ
showed an alternative method that uses extra information
ଵଵ ൌ ଵ ଶ ൌͲ (46) coming from another sample that has different length from the
ଵି୻మ ௓ మ
first sample. Finally, the problem of divergence of the method
ȁ݁ ିଶఊబ ௅೔ ȁȁȞଵଶ ȁȁͳ െ݁ ିଶఊௗ ȁ
ൌͲ (47) ఒ
when the sample has a length multiple of ଶ೒ is due to the
ଶగ
Assuming thatȞଵଶ ് Ͳ and ߛ ൌ ߙ ൅ ݆ߚ, ߚ ൌ ǡ then severely deterioration of signal noise ratio when ଵଵ  ՜ Ͳǡ
ఒ௚
which leads to inconsistent results. An important source of
ȁͳ െ ݁ ିଶௗఈ ሾ…‘•ሺെʹ݀ߚሻ ൅ ݆‫݊݁ݏ‬ሺെʹ݀ߚሻሿȁ ൌ Ͳ (48) error not addressed in this article is the air gaps. A study of the
effects of uncertainties and air gaps in the results can be found
To solve (48), the real and imaginary parts must be zero.
in [5].
From the imaginary part
For the radome application, which uses low-loss materials,
݁ ିଶௗఈ ‫݊݁ݏ‬ሺʹ݀ߚሻ ൌ Ͳ ՜ ʹ݀ߚ ൌ ݇ߨǡ ݇ ‫ א‬Ժ (49)
the literature indicates that the NRW method is not enough
௞ఒ೒ accurate. For solid low-loss materials, literature indicates
݀ൌ ǡ݇ ‫ א‬Ժ (50)
ସ methods such as NIST iterative, new non-iterative and resonant
ͳ െ ݁ ିଶௗఈ …‘•ሺെʹ݀ߚሻ ൌ Ͳ (51) cavity techniques [9]. In spite of that, the NRW is basic
knowledge for several of these aforementioned methods and
From (48) and the real part equal to zero the step by step procedure is a powerful tool to better
ͳ െ ݁ ିଶௗఈ …‘•ሺെ݇ߨሻ ൌ Ͳ (52) understand and facilitate its implementation.
ିଶௗఈ ሺെͳሻ௞
ͳൌ݁ (53) ACKNOWLEDGMENT
This equation has solution only if This work has been supported by CNPq (project no.
559991/2010-0). The authors want to acknowledge the
௡ఒ೒
݇ ൌ ʹ݊ǡ ݊ ‫ ܼ א‬՜ ݀ ൌ ǡ݊ ‫ א‬Ժ (54) collaboration of researcher Dr. Antonio Carlos da Cunha
ଶ
Migliano from Instituto de Estudos Avançados (IEAv/DCTA).
Then, from (53) and (54)
REFERENCES
െʹ݀ߙ ൌ Ͳ (55)
[1] L. F. Chen,.C. K. Ong, C. P. Neo, V. V. Varadan, V. K. Varadan,
Since ݀ ് Ͳ, then ߙ ൌ Ͳ, in other words, MUT is a lossless Microwave electronics: measurement and materials characterization.
Chichester: Wiley, c2004, 537 p.
material. [2] A. M. Nicolson; G. F. Ross, Measurement of the intrinsic properties of
materials by time domain techniques. IEEE Transaction on Instrumentation and
III. CONCLUDING REMARKS Measurement, vol. IM-19: 377-382, 1970 November.
This paper described the NRW method in a way more [3] W. B. Weir, Automatic measurement of complex dielectric constant and
permeability at microwave frequencies. Proceedings of the IEEE, vol. 62: 33-
detailed and integrated than is usually found in the literature. It 36; 1974 January.
is the most commonly method used for materials [4] D. M Pozar,. Microwave engineering. 2. Ed. New York, Ny: Wiley, c1998.
characterization and has the advantage of being non-iterative, 716p.
applicable to TEM, TE and TM waves, although this paper has [5] J. Baker-Jarvis, et al. Measuring the Permittivity and Permeability of Lossy
Materials: Solids, Liquids, Metals, Building Materials, and Negative-Index
addressed only TEM and TE waves. It is also applicable to Materials, NIST Technical Note 1536, National Institute of Standards and
magnetic and non-magnetic materials, for the last ones it is Technology, U.S. Department of Commerce..
possible to improve accuracy and remove the phase ambiguity [6] C. A. Balanis, Advanced Engineering, chapter 5. New York, NY: Wiley,
by modifying the method, due to this additional information. c1989. 981 p.
[7] D. J. Kozakoff, Analysis of Radome – Enclosed Antennas., 2. Ed. Edition.
Another advantage of NRW method is that it is a closed [8] Materials measurement: Measuring the dielectric constant of solids with the
formula, unless the phase ambiguity, therefore, it is possible to HP 8510 network analyzer. Product Note 8510-3
study the sensitivity of the results according to each input [9] Measurement of dielectric material properties, Rohde&Schwarz, Aplication
parameter of the method, quantifying more precisely the Note. RAC0607-0019.

978-1-4577-1664-5/11/$26.00 ©2011 IEEE 742