This article has been accepted for publication in a future issue of this journal, but has not been

fully edited. Content may change prior to final publication. Citation information: DOI
10.1109/ACCESS.2020.2970271, IEEE Access

Date of publication xxxx 00, 0000, date of current version xxxx 00, 0000.
Digital Object Identifier xx.xxxx/ACCESS.xxxx.DOI

Inductive Frequency Selective Surface:

An application for dichroic sub-reflectors
JUAN A. VÁSQUEZ-PERALVO1 , JOSÉ M. FERNÁNDEZ-GONZÁLEZ1 , (Senior Member, IEEE),
PAVEL VALTR2 , JONATHAN M. RIGELSFORD 3 , (Senior Member, IEEE)
1
Radiation Group, Department of Signals, Systems and Radio communications, Universidad Politécnica de Madrid, 28040 Madrid, Spain.
2
Department of Electromagnetic Field, Faculty of Electrical Engineering, Czech Technical University in Prague, 166 27 Prague, Czech Republic.
3
Sensata Technologies, Interface House, Bincknoll Lane, Royal Wootton Bassett, Swindon, SN4 8SY UK.
Corresponding author: Juan A. Vásquez-Peralvo (e-mail: [email protected]).
Authors would like to acknowledge the Spanish Government, Ministry of Economy, National Program of Research, Development and
Innovation for the support of this publication in the projects FUTURE-RADIO “Radio systems and technologies for high capacity
terrestrial and satellite communications in an hyperconnected world “ (project number TEC2017-85529-C3-1-R).”

ABSTRACT
An Inductive Frequency Selective Surface (IFSS) self-standing dichroic sub-reflector is presented, which
allows the transmission of two frequency bands and reflection of a third band. The operating frequency
of the sub-reflector is designed to work at Ku, K and Ka frequency bands, intended for earth to space
and space to earth satellite communications. The proposed unit cell is a Jerusalem cross intertwined with
an Brigid’s cross. The IFSS is investigated using transmission line theory, along with equivalent-circuit
model technique. It has been designed, and simulated using CST Microwave Studio and Advance Design
Simulator (ADS). The cascade configuration of the IFSS is also investigated to improve frequency roll-
off and bandwidth of the reflection and transmission coefficients. The IFSS has been manufactured using
two-sided Photo Chemical Machining (PCM) technique, and has been experimentally characterized using
an optical configuration, comprising two double ridged horn antennas connected to a VNA and a rotation
system. Measured results are in good agreement with theoretical and simulation data, which validates the
reliability of the design and manufacturing process.

INDEX TERMS Reflector antennas, equivalent circuits, frequency selective surfaces.

I. INTRODUCTION the E- and H-Fields at each point of the surface. This method
NDUCTIVE Frequency Selective Surfaces (IFSSs) are takes less computational resources than others, and gives a
I electromagnetic filters capable of working as high-pass
filters or band-pass filters depending on the resonant element
good approximation to real results, even though the FSS
is embedded in a stratified medium [4]. The drawbacks of
etched in a conductive surface [1]. The main advantage of this method are its accuracy, it cannot describe the operation
this type of FSS is the self-standing capability, that is, it can of the FSS at the grating lobe region, and it cannot be
be implemented without a dielectric substrate. Therefore, it used in certain structures like meandered dipole, quadrifilar
will not have frequency shifts due to change of mediums. spiral, and genetically optimized shapes [7]. The second are
Additionally, if the conducting surface is thick enough, and computational methods, which are the most accurate, and
manufactured with high hardness metal sheet, it will have complex methods of both. These offer good approximation
mechanical stability [2]. The drawback of dispensing with the at different scenarios, for example, in the case of an FSS
substrate are: reduced angular stability, reduced mechanical with multiple resonant shapes embedded within dielectric
stability, and increased manufacturing complexity compared loading, the case of an FSS illuminated at different incident
with a conventional FSS. angles, and the case of a multi-layer FSS configuration, just
Two methods have been proposed to analyze FSS reflection to mention a few. Some of those computational methods are:
and transmission coefficients. The first is equivalent circuit periodic method of moments (PMM) and finite difference
method, which was proposed in [3]–[7]. It uses a quasi-static time domain method (FDTD), which are the most used for
approximation, that is, it considers a uniform distribution of these matters [8].

VOLUME 0, 0000 1

This work is licensed under a Creative Commons Attribution 4.0 License. For more information, see https://creativecommons.org/licenses/by/4.0/.
 This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI
10.1109/ACCESS.2020.2970271, IEEE Access

Juan A. Vásquez-Peralvo et al.: Inductive Frequency Selective Surface: An applications for dichroic sub-reflectors

Reducing the weight on a spacecraft, and reducing the area The main difference between our work, and previous
in a satellite is critical, and involves many engineer fields works on dichroic surfaces for satellite antennas is the use
[2]. The satellite dish antenna design challenges comprise an a completely inductive frequency selective surface, which
effective reflector design, an optimum multifrequency feed means that this structure does not need a substrate to have
system, and frequency feed segregation. Each of those will mechanical stability.
be addressed in the following paragraph. First, an effective This paper is organized as follows: Section II comprises
reflector design can be done using a Cassegrain reflector sys- the unit cell design steps. It involves frequency requirements,
tem with an FSS as dichroic sub-reflector. Second, the FSS unit cell design of both Brigid’s and Jerusalem cross, and the
allows, for example, using X band and Ku or K frequency combination of both shapes. In addition, reflection and trans-
bands with the same antenna structure giving multifrequency mission coefficients of TE mode are presented for single and
operation. Finally, the feeds will be isolated due to the double layer IFSS. Section III comprises equivalent circuit
dichroic capability. Fig. 1 shows a ray analysis highlighting analysis of the designed FSS, as well as its mathematical
the antenna design challenges in a satellite. background. In addition, the reflection coefficient obtained
For satellite communications, several works have been done will be compared with the full wave simulation. Section IV
on dichroic sub-reflectors [9]–[19]. Most of these works comprises the parametric study of the IFSS. Section V com-
offers two, three, and even four operation bands, where the prises the experimental verification of our proposed IFSS. It
dichroic surface acts as a band pass filter or a band stop involves the effects of the PCM manufacturing process, the
filter. This allows working in the satellite with four frequency measurement setup, and a comparison between simulated and
bands using one parabolic dish, thus reducing the number of measurement results.
parabolic dishes in the satellite.
II. UNIT CELL DESIGN
First, to design a unit cell we need to know the frequency
bands of operation and the required bandwidth. This in-
formation can be found in Table 1. The frequency bands
presented are used in earth-to-satellite and satellite-to-earth
communications in Europe [20], and were chosen as a proof
of concept. Second, to fulfill the frequency operation and
bandwidth requirements, resonant shapes were chosen. Some
resonant shapes, and its frequency characteristics can be
found in [1]. Once the resonant shapes had been chosen,
the following characteristics must be obtained: dimensions,
periodicity, metal thickness, lattice type and material. Those
characteristics will dictate the resonant frequencies, band-
width, onset of grating lobes, angular stability and polariza-
tion independence.
The resonant shapes chosen, which best suit our require-
FIGURE 1. Ray analysis of a prime focus Cassegrain reflector using dichroic
sub-reflector. The dichroic surface will act as ground plane for one frequency
ments, are: the intertwined Brigid’s and Jerusalem cross
band, and will be transparent for two. apertures.

TABLE 1. Proposed frequency bands for the IFSS operation as an application

Key contributions of this paper are as follows: for satellite communications.
1) Design and optimize an IFSS, which works at three
frequency bands, using intertwined structures. This war-
rants improved angular stability and is self-standing. Band Central Required Application
2) A mathematical interpretation, based on quasi-static Frequency Bandwidth
analysis of the structure. 1 18.5 GHz 17.3-19.3 GHz Fixed satellite service
3) A physical interpretation, based on equivalent circuit 2 24.6 GHz 24.5-25.5 GHz ISM and Radio determination
3 30.1 GHz 29.1-31 GHz Fixed satellite service
model technique.
4) An understanding of the E- and H-Fields, which will
help us to identify inductances and capacitances in the
IFSS generated by an impinging plane wave. A. INTERTWINED BRIGID’S CROSS SHAPE
5) A parametric study of the unit cell, which let us tune The intertwined Brigid’s cross shape has been chosen due
different frequency bands. to its symmetry, and to its intertwined characteristics, which
6) A description of PhotoChemical Machining process, connects multiple adjacent unit cells. Both characteristics
which is used to manufacture the IFSS, and analyze the leads to a more packed unit cell, improving angular stability
effects of rounded corners and beveled edges. and bandwidth. The fundamental mode resonates at arm
2 VOLUME 0, 0000

This work is licensed under a Creative Commons Attribution 4.0 License. For more information, see https://creativecommons.org/licenses/by/4.0/.
 This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI
10.1109/ACCESS.2020.2970271, IEEE Access

Juan A. Vásquez-Peralvo et al.: Inductive Frequency Selective Surface: An applications for dichroic sub-reflectors

lengths multiple of λ0 /5. This shape will generate the Band

1 requirements.

B. JERUSALEM CROSS SHAPE

The Jerusalem cross shape has good angular stability, iso-
lated resonances, and can be highly packed [1]. Normally,
this shape has two highly separated fundamental modes,
the last is generally located at the grating lobe region and
sometimes is mistakenly confused with a second harmonic.
The first fundamental mode depends on the vertical aperture
and horizontal aperture, in the case of a symmetric shape
[1]. This etched shape will generate the frequency Band 3
requirements. Both, intertwined Brigid’s and Jerusalem cross
aperture shapes can be seen in Fig. 2.

FIGURE 3. Unit cell dimensions, and full structure of an intertwined

Jerusalem cross IFSS. (Shaded region represents metal)).

mode was analyzed in reflection and transmission, as can

be seen in Fig. 4. TM mode will not be analyzed due to
symmetry of the IFSS. The IFSS has been analyzed using

FIGURE 2. a) Intertwined Brigid’s cross unit cell aperture with its approximate
resonant frequency. b) Jerusalem cross unit cell aperture with its approximate
resonant frequency. (Shaded region represents metal).

C. SINGLE LAYER INTERTWINED JERUSALEM CROSS

IFSS
To acquire the final unit cell, and to generate the frequency
Band 3, we intersect both unit cells etching the intertwined
Brigid’s cross in the Jerusalem cross aperture unit cell. The
combination will generate a highly packed, multi-resonant
and polarization independent unit cell. The resonant frequen-
cies and bandwidth of this combined unit cell have a slight
variations compared to separated shapes. This phenomenon
can be explained due to coupling of the two etched shapes,
generated by surface currents that are mainly placed in the
apertures. These dimensions are a starting point to opti- FIGURE 4. Simulated reflection and transmission coefficients of the IFSS
mize the structure to resonate at the required frequencies presented in Fig. 3 at normal incidence of a vertically polarized wave.

and bandwidths. The structure was optimized using particle

swarm optimization, embedded in CST Microwave Studio. admittance Smith chart shown in Fig. 5, which helps us to
The goals were: resonant frequencies, bandwidth, and polar- characterize the IFSS operation at different frequency bands.
ization independency. The unit cell is presented in Fig. 3 and At frequency Band 1 and 2, the IFSS is well matched to the
1
its dimensions are: A = 1 mm, B = 0.3 mm, C = 0.15 free space admittance Yo ≈ 377 . Therefore, the structure will
mm, D = 0.425 mm, E = 0.2 mm, F = 0.3 mm, G = 0.2 let thorough almost all the incoming electromagnetic waves.
mm, and H = 0.3 mm. This structure has a thickness T = Conversely, at frequency Band 2, the admittance is poorly
0.2 mm and a lattice periodicity P = 4.3 mm. matched, acting as a ground plane reflecting all the incoming
Transmission and reflection coeffcients were obtained us- electromagnetic waves in the backscatter direction. It is worth
ing two Floquet ports separated λ40 from the IFSS. The solver to mention that the phase difference between reflected and
used was frequency domain and configured with a suitable incident electromagnetic wave is positive due to the inductive
mesh at the frequency range 0 − 90 GHz. This frequency operation of the surface.
range was selected to show the grating lobe region. TE
VOLUME 0, 0000 3

This work is licensed under a Creative Commons Attribution 4.0 License. For more information, see https://creativecommons.org/licenses/by/4.0/.
 This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI
10.1109/ACCESS.2020.2970271, IEEE Access

Juan A. Vásquez-Peralvo et al.: Inductive Frequency Selective Surface: An applications for dichroic sub-reflectors

grid. A more detailed equivalent circuit that englobe both

Jerusalem cross resonant frequencies was proposed in [21].
The equivalent circuit consists of two RLC series circuit con-
nected in parallel. This leads us to think that the equivalent
circuit for a Jerusalem cross aperture will be a two RLC
parallel circuit connected in series. To verify this hypothesis,
a full wave simulation was made, obtaining E- and H-Fields
at the structure. After examining those fields, capacitors are
represented by electric charge distribution in the apertures.
Whereas, the metallic strips, where the magnetic field de-
scribes a curl, are represented by inductors [22].
At the first resonant frequency, which is 29 GHz, the E-
Field is strong in the horizontal middle aperture. This can
be seen in Fig. 7a and can be represented using a capacitor.
On the other hand, the strip that separates both Jerusalem
cross shapes is surrounded by a curled H-Field. This can
FIGURE 5. Smith chart representation of the IFSS simulated admittance be seen in Fig. 7b and can be represented by an inductor
presented in Fig. 3. The frequency bands presented are described in Table 1. with its embedded resistance. The connection between the
capacitor and the inductor is parallel due to the inductive
D. DOUBLE LAYER INTERTWINED JERUSALEM CROSS
nature of the FSS. Neglecting the losses in the IFSS, it can
IFSS
The roll-off, flatter top, and bandwidth can be improved
by cascading two IFSS separated at a distance of less than
λ0 /4 [1]. The distance was acquired through optimization
at the frequency bands of interest. After the optimization
process, the most suitable distance is 2 mm. Fig. 6 shows
the simulation of a double layer IFSS.

FIGURE 6. Simulated reflection and transmission coefficients of a double layer

IFSS, presented in Fig. 3 at normal incidence of a vertically polarized wave.

FIGURE 7. a) E-Field 2D contour of a vertically polarized wave, working at 29

III. EQUIVALENT CIRCUIT ANALYSIS OF THE IFSS GHz, b) H-Field lines of a vertically polarized wave, working at 29 GHz.
For simplicity, equivalent circuit for both shapes will be
acquired separately, and finally, combined to obtain an equiv- be inferred through analysis that the equivalent circuit is
alent circuit, which describes its multi-band operation. Lp1 ||Cp1 , because of the bandpass operation of the IFSS.
It is worth to say that the second resonant frequency of the
1) Jerusalem cross aperture IFSS is placed at the grating lobe region. Therefore, it will
The work described in [3] is the first paper to our knowledge not be considered. The equivalent circuit of the Jerusalem
that proposes an equivalent circuit for a Jerusalem cross cross aperture can be seen in Fig. 11.
4 VOLUME 0, 0000

This work is licensed under a Creative Commons Attribution 4.0 License. For more information, see https://creativecommons.org/licenses/by/4.0/.
 This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI
10.1109/ACCESS.2020.2970271, IEEE Access

Juan A. Vásquez-Peralvo et al.: Inductive Frequency Selective Surface: An applications for dichroic sub-reflectors

The impedance generated by each of the inductors and The values computed are: Cp1 = 167.14 f F and Lp1 =
capacitor are: 180 pH. Full wave simulations and equivalent circuit model
−j simulations are shown in Fig. 8.
ZCp1 = (1)
ωCp1

ZLp1 = jωLp1 (2)

respectively; therefore, the impedance of the circuit reads:
Lp1
ZT = jωz1 2 L C ) (3)
(1 − ωp1 p2 p2

By Babinet’s principle, the grid array is complementary

to the aperture array [1]. Hence, the admittance expression
for a Jerusalem cross grid array calculated in [7] is equal
to the impedance calculated in our work. From (3), we can
derive the values of the inductors and capacitors as follows.
We can see that the impedance in (3) will be minimum at
ωz1 = 0, and maximum at a certain value of ωp1 . Clearly the
value of ωp1 is our second resonance frequency. The value
of the inductors and capacitors can be obtained through the
following relation.
FIGURE 8. Full wave and circuit approach simulation of the Jerusalem cross
1 aperture presented in Fig. 7 at normal incidence of a vertically polarized wave.
Lp1 = 2 (4)
ωp1 Cp1
The previous equation can be solved with the numer-
2) Intertwined Brigid’s cross aperture
ical value of Cp1 . The work described in [21] proposes
an equivalent-circuit-modeling technique, which allows to The intertwined Brigid’s cross aperture can be analyzed using
calculate an approximate value of reactance or susceptance the same methodology as the Jerusalem cross. The E-Field
of an FSS segment. By using this technique and (5)-(8), and H-Field obtained through full wave simulation can be
we will obtain an approximate value for Cp1 susceptance, seen in Fig. 9a and Fig. 9b respectively. It can be inferred that
named Bp1 . Then, to obtain the reflection and transmission the intertwined Brigid’s cross aperture can be represented
coefficients we use (9) and (10) respectively. For a more using a parallel LC circuit, as shown in Fig. 11. The structure
accurate value, we used an iterative procedure that compares equivalent impedance is presented in (11).
the IFSS reflection coefficient obtained through electromag- Lp3
netic simulation with equivalent circuit simulation. ZT = jωz3 2 L C (11)
1 − ωp3 p3 p3

Using (11) we can see that the IFSS will have a minimum
4d cos(φ) πg
Bp1 = ln csc( ) + G(p, g, λ, φ) (5) at ωz3 = 0 and a maximum at ωp3 , which in this case is
λ 2p our second resonance frequency. After applying the previous
where, formulas, we obtain, Lp3 = 1.27 nH and Cp3 = 55 f F . Full
wave simulation and equivalent circuit model simulation can
G(p, g, λ, φ) = be seen in Fig. 10.
1 2 2 β 2
2
The final equivalent circuit, which is shown in Fig. 11,
2 (1 − β ) [((1 − 4 )(C+ + C− ) + 4β C+ C− )]
(6) represents the intertwined Jerusalem cross IFSS. This circuit
2 2 4
1 − β4 + β 2 (1 + β2 − β8 )(C+ + C− ) + 2β 6 C+ C− consists on a series connection of branches Lp1 ||Cp1 , which
πw corresponds to the first resonance frequency of the Jerusalem
β = sin( ) (7) cross, with Lp3 ||Cp3 , which corresponds to the Brigid’s cross
2p
resonance frequency. This connection is possible because of
1 the band-pass nature of this structure. Equivalent circuit is
C± = q −1 (8) represented in the physical structure by drilling the Jerusalem
( p sin
λ
φ
± 1)2 − ( λp )2 cross structure into the Brigid’s cross, as shown in Fig. 3.
The previously calculated capacitances and inductances are
Z0
|S11 | = (9) affected by this combination, and new values were obtained
2Z + Z0
using the same curve fitting method. Basically, the previously
2Z calculated inductance Lp1 is lower than the new inductance
|S12 | = (10) LJC due to the magnetic field circulating in the plate between
2Z + Z0
VOLUME 0, 0000 5

This work is licensed under a Creative Commons Attribution 4.0 License. For more information, see https://creativecommons.org/licenses/by/4.0/.
 This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI
10.1109/ACCESS.2020.2970271, IEEE Access

Juan A. Vásquez-Peralvo et al.: Inductive Frequency Selective Surface: An applications for dichroic sub-reflectors

final unit cell was simulated in CST and ADS respectively,

and shown in Fig. 12.

FIGURE 11. Equivalent circuit of an intertwined Jerusalem cross aperture,

where LOC ||COC represents the first band pass generated by the Brigid’s
cross aperture, LJC ||CJC represents the second resonant frequency
generated by the Jerusalem cross aperture, and Zo represents the free space
impedance. LOC = 0.962nH, COC = 76f F , LJC = 0.4313nH,
CJC = 64.3f F .

FIGURE 9. a) E-Field 2D contour of a vertically polarized wave, working at

18.5 GHz, b) H-Field lines of a vertically polarized wave, working at 18.5
GHz.

FIGURE 12. Full wave and circuit approach simulation of the Jerusalem cross
aperture and intertwined Brigid’s cross aperture, at normal incidence of a
vertically polarized wave.

IV. PARAMETRIC STUDY OF THE PROPOSED IFSS

A parametric study was performed to understand the effects
of variation in frequency and bandwidth due to the change in
FIGURE 10. Full wave and circuit approach simulation of the intertwined length and width of the unit cell apertures. To keep the period
Brigid’s cross aperture, presented in Fig. 11 at normal incidence of a vertically constant, variations of the apertures will be compensated with
polarized wave.
variations in metal length. As the unit cell is symmetric, only
Jerusalem cross apertures. This increases as the effective TE mode reflection coefficient will be shown.
metallic area is reduced. The previously calculated Cp1 value 1) Parameter G (Metal strip between Brigid’s crosses)
is greater than the new CJC capacitance due to the mutual Since G variation only affects the size of the displaced
coupling of both structures, which reduces the electric field cross arms and not to the period of the unit cell; there-
in the center of the cross. The values of Lp3 and Cp3 , which fore, no other parameter will be modified to compensate
now are called LOC and COC , suffer variations but to a lesser this variation. Fig. 13 shows by increasing G, from 0.1
extent, since both values depend exclusively on the inter- mm to 0.4 mm, increases the first resonance frequency.
weaving generated in both arms of the Brigid’s cross. The Increasing this parameter, the electrical length of the
6 VOLUME 0, 0000

This work is licensed under a Creative Commons Attribution 4.0 License. For more information, see https://creativecommons.org/licenses/by/4.0/.
 This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI
10.1109/ACCESS.2020.2970271, IEEE Access

Juan A. Vásquez-Peralvo et al.: Inductive Frequency Selective Surface: An applications for dichroic sub-reflectors

Brigid’s cross aperture is reduced; therefore, the first the two crosses increases; therefore, and using equa-
resonance frequency is increased. It should be noticed tion (4), the second resonance frequency is reduced, as
that tuning of this first resonance frequency is totally shown in Fig. 15. It should be mentioned that these
independent of the second resonant frequency, but af- variations do not affect the first resonance frequency, but
fects the total reflection band, which is reduced as G affect the total reflection band, increasing it.
increases.

0 0

-5 -5

-10 -10

[dB]
|R|TE [dB]

-15 -15

TE
-20

|R|
-20

-25 -25
G=0.1 mm D=0.35 mm
G=0.2 mm -30 D=0.3875 mm
-30
G=0.3 mm D=0.425 mm
G=0.4 mm D=0.4625 mm
-35 -35
15 20 25 30 35 15 20 25 30 35
Frequency [GHz] Frequency [GHz]

FIGURE 13. Reflection coefficient of the proposed unit cell when G varies. FIGURE 15. Reflection coefficient of the proposed unit cell when the variated
The original value of G is 0.2 mm. parameter is H. The original value of H is 0.85 mm.

2) Parameter H (Brigid’s cross arms width) 4) Parameter B (Jerusalem cross loading arm width)
When increasing H the value of E is reduced to keep By varying the loading aperture B of the Jerusalem
the same period. Increasing H and reducing E, the cross we can control both the second resonance fre-
associated capacitance is reduced while the associated quency and the reflection band. When increasing B,
inductance increases; therefore, the quality factor Q F is reduced to keep the same period. This variation
is reduced, thus increasing the bandwidth of the first entails that the metallic strip placed above the loading
resonance frequency. This effect can be seen in Fig. 14. arm gets thinner, increasing the inductance associated;
It is worth mentioning that these variations do not affect therefore, the Q factor is reduced; thus, increasing band-
the second resonance frequency, but they do affect the width. On the other hand, as the aperture gets thicker
total reflection band, increasing it as H increases. the capacitance associated gets bigger, decreasing the
resonant frequency. Fig. 16 show that as B gets bigger,
the second resonance frequency is reduced, and the
0 bandwidth increases. Finally, as we mention earlier;
as B increases the reflective frequency band, and the
-5 bandwidth decreases.
-10
V. EXPERIMENTAL VERIFICATION
The proposed concept of an IFSS using intertwined shapes
|R|TE [dB]

-15
has been validated through experimentation. The material
-20 used for the manufacturing was stainless steel 304 with thick-
ness of 0.2 mm and σ =720 nΩ. Even though this material
-25
E=0.2 mm, H=0.4 mm
has less conductivity than copper, it has high strength, and
-30 E=0.25 mm, H=0.3 mm corrosion resistance. If we want to improve the conductivity,
E=0.3 mm, H=0.2 mm we could consider to metallize the structure with cooper. Two
E=0.35 mm, H=0.1 mm
-35 prototypes were designed using 36-unit cells with an overall
15 20 25 30 35
Frequency [GHz] dimension of 200x200 mm. Near the sided edges of the
structure 8 holes were placed to hold a frame for measuring
FIGURE 14. Reflection coefficient of the proposed unit cell when the variated
parameter are D and C. Original value of H is 0.3 mm and E is 0.2 mm.
purposes.

3) Parameter D (Jerusalem cross loading arm length) A. IFSS MANUFACTURING PROCESS

By increasing the loading length of the Jerusalem cross, Before choosing a manufacturing method, we need to con-
the inductance generated by the metal plate separating sider some design parameters, such as: smallest aperture size,
VOLUME 0, 0000 7

This work is licensed under a Creative Commons Attribution 4.0 License. For more information, see https://creativecommons.org/licenses/by/4.0/.
 This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI
10.1109/ACCESS.2020.2970271, IEEE Access

Juan A. Vásquez-Peralvo et al.: Inductive Frequency Selective Surface: An applications for dichroic sub-reflectors

-5

-10
|R|TE [dB]

-15

-20

-25
B=0.2 mm, F=0.4 mm
-30 B=0.3 mm, F=0.3 mm FIGURE 17. a) Cross sectional view of a prototyped metal sheet using
B=0.4 mm, F=0.2 mm two-sided PCM. b) Cross sectional view of a prototyped metal sheet using
B=0.5 mm, F=0.1 mm one-sided PCM. c) Front view of the unit cell after etching using PCM.
-35
15 20 25 30 35 40
Frequency [GHz]
FIGURE 16. Reflection coefficient of the proposed unit cell when the variated
parameter are B and F . The original value of B is 0.3 mm and F 0.3 mm.

which is ≈ 200 µm, hardness, and thickness of the material.

After a research of the available manufacturing techniques,
the better that suits our purposes is Photochemical Machining
(PCM).

1) PhotoChemical Machining
This method follows the same principle as PCB manufac-
turing. It has an accuracy of 13 µm and the sheet thickness
varies from 0.013 to 2.032 mm. It has a major limitation
represented by non-uniform etching through apertures. That
FIGURE 18. IFSS prototyped using a 0.2 mm thick sheet of stainless steel
is, the etchant will not attack the metal perpendicularly to 304 and a zoomed view of a unit cell.
the resistive sheet; instead, it will attack sidewalls and under
the resistive film. Depending on the manufacturer, there are dations of the manufacturer for both etching methods are:
a variety of etching techniques, which offers different bevel inner radius r=t, outer radius R=0.5t, bevel for one-sided
finishing at the edges. The most common PCM techniques etching d=0.4t and for two-sided etching D=0.2t. Stainless
are: one and two-sided etching. On One-sided etching, the steel 304 was selected as material for the IFSS unit cell. S-
resistive stencil is placed over the upper face of the metal Parameters comparison between those two etching methods
sheet and a complete resistive sheet over the other face, as and the original unit cell design are shown in Fig. 19.
illustrated in Fig. 17b. The bevel in one-sided etching is
0
approximate d = 0.4 t, being t the thickness of the metal
sheet. In Two-sided etching, the resistive stencils are placed -5
over both faces of the metal sheet, as illustrated in Fig. 17a. -10
The bevel in two-sided etching is approximate D = 0.2 t
Magnitude [dB]

-15
to D = 0.1 t and the bevel is uniform in both faces. Both
etching techniques have an under etching in the resistive film -20
that is approximate to g = 0.02 t. -25
In addition, rounded corners are created in both etching
-30
techniques. Depending on the corner, the radius could be
external R or internal r, as illustrated in Fig. 17c. External -35 |R| TE Original |T| TE original

radius is approximate R = 0.4 t to R = 0.6 t and internal |R| TE One side etching |T| TE one side etching
-40
radius is r = 0.8d to r = d [23]. |R| TE Two side etching |T| TE two side etching
-45
15 20 25 30 35
Frequency [GHz]
B. EFFECTS IN ROUNDED CORNERS AND BEVELED
EDGES. FIGURE 19. Reflection and transmission coefficients simulation of an original
designed IFSS, one-sided etching and two-sided etching method. Both
The manufacturer offered both one-sided and two-sided etch- methods use an inner radii r = t and outer radii R = 0.5t.
ing. To understand the effects of either, and to choose one
that better suits our research purposes, the original unit cell In the first resonance frequency the S-Parameters for the
was modified to simulate both etching methods. Recommen- original and one-sided etching are nearly the same, with a
8 VOLUME 0, 0000

This work is licensed under a Creative Commons Attribution 4.0 License. For more information, see https://creativecommons.org/licenses/by/4.0/.
 This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI
10.1109/ACCESS.2020.2970271, IEEE Access

Juan A. Vásquez-Peralvo et al.: Inductive Frequency Selective Surface: An applications for dichroic sub-reflectors

light increase in bandwidth of the latter. On the other hand,

two-sided etching has a resonance frequency shift of 0.3
GHz. In the second resonance frequency, the S-Parameters
have a more marked difference. The resonance frequency is
shifted upward 0.4 GHz for two-sided etching and 1.3 GHz
for one-sided etching with respect to the original design.
This variation can be explained due to the reduction on the
Jerusalem cross arm length and width due to the etching
process. This shift is bigger in one-sided etching because
the bevel edge is bigger. The transmission coefficient shows
a bigger shift upward in one-sided etching than two-sided
due to the same reasons previously explained. As a result,
we choose two-sided etching to avoid unwanted frequency
displacement and to keep cardinal symmetry.

C. MEASUREMENT SETUP
The manufactured IFSS has been characterized by measuring
transmission and reflection coefficients using an optical bed FIGURE 20. Schematic illustration of the measurement setup, which includes
configuration. The measurement setup shown in Fig. 20 was distance form antennas and from SUT.
configured as follows: Two double ridged horn antennas
model DRH40 were placed away form each other a distance
of 400 mm and 450 mm above a rail using a support sys-
tem. Although both antennas were placed at a distance that
warranties correct illumination, they are placed in the near-
field region to avoid using lents. At the middle, a motorized
rotation stage (MRS), model MOR-100-30 was placed over a
rail. This rotates the Surface Under Test (SUT) to a different
angles of incidence. The SUT was fastened using a 1 cm
thick frame made of a ROHACELL 31HF which has an r =
1.041. This frame was placed in the MRS fastening system. A
photo of the measurement setup can be seen in Fig. 21. Both
antennas outputs are connected through coaxial cable to a
Vector Network Analyzer (VNA) Rohde&Schwarz ZVA 67.
Surrounding the measurement setup absorbent material was
placed to reduce unwanted reflections. The VNA frequency
range was configured from 5 to 35 GHz with a step size
of 50 M Hz. The number of points was set to 1024 with a
frequency span of 13 GHz. To enhance measurements re- FIGURE 21. Photo of the measurement setup which includes both double
ridged horn antennas model DRH40, rail, motorized rotation stage, SUT, and a
sults, time gating was applied to the RAW data. This reduces VNA.
deceptive backscattering caused by multi-path propagation.
For reflection measurements, two stages were configured. In TE mode and for non-normal incidence angles, variations
Normal incident measurements were performed using one between measured and simulated results are seen in both
antenna. For angles above 0 degrees measurements were reflection and transmission coefficients. For the reflection
made using two antennas. The latter case, both antennas will coefficient, which is shown in Fig. 24, the first resonance
be separated by an angle centered in the middle of the IFSS. frequency position and bandwidth are in better accordance
with simulation, compared to the normal incidence results.
D. SINGLE LAYER IFSS Once more, this is due to the two antenna measurement
The prototype was characterized using the configuration pre- technique, which is more accurate. In addition, the measured
viously described. The S-Parameters measurements obtained reflection coefficients show a slight variation in bandwidth
to normal incidence illumination for both modes show accor- at the first resonance frequency, and a more pronounced
dance with simulations, with a slight variation in bandwidth variation at the second resonance frequency, compared to
and frequency displacement at the first resonance frequency, normal incidence case results because angular stability de-
as can be seen in Fig. 22 for TE mode, and Fig. 23 for TM creases for resonances above the first. On the other hand, for
mode. This variation in bandwidth is most likely due to: mea- transmission measurements, which are shown in Fig. 25, have
surement mismatches, one antenna measurement technique two nulls instead of one as for the case of normal incidence.
and manufacturing tolerances. This new null appears at 22 GHz and is present only for
VOLUME 0, 0000 9

This work is licensed under a Creative Commons Attribution 4.0 License. For more information, see https://creativecommons.org/licenses/by/4.0/.
 This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI
10.1109/ACCESS.2020.2970271, IEEE Access

Juan A. Vásquez-Peralvo et al.: Inductive Frequency Selective Surface: An applications for dichroic sub-reflectors

0 0

-5 -5

-10 -10
Magnitude [dB]

-15 -15

|R| TE [dB]
-20 -20

-25 -25

-30 -30
|R|TE =0° Sim. |T|TE =0° Sim. =5° Sim. =10° Meas.
-35 -35 =5° Meas. =15° Sim.
|R|TE =0° Meas. |T|TE =0° Meas.
=10° Sim. =15° Meas.
-40 -40
15 20 25 30 35 15 20 25 30 35
Frequency [GHz] Frequency [GHz]
FIGURE 22. TE mode measured and simulated reflection and transmission FIGURE 24. TE mode measured and simulated reflection coefficient of a
coefficient at normal incidence for a single layer IFSS. single layer IFSS at oblique incidence.

0 0

-5 -5

-10 -10
Magnitude [dB]

-15 -15
|T| TE [dB]

-20 -20

-25 -25

-30 -30
=5° Sim. =10° Meas.
|R| TM =0 ° Sim. |T| TM =0 ° Sim.
-35 -35 =5° Meas. =15° Sim.
|R| TM =0 ° Meas. |T| TM =0 ° Meas. =10° Sim. =15° Meas.
-40 -40
15 20 25 30 35 15 20 25 30 35
Frequency [GHz] Frequency [GHz]
FIGURE 23. TM mode measured and simulated reflection and transmission FIGURE 25. TE mode measured and simulated transmission coefficient of a
coefficient at normal angle of incidence for a single layer IFSS. single layer IFSS at oblique incidence.

TE mode at oblique angles of incidence, in this case angles shown in Fig. 28 and Fig. 29 respectively, and for TM mode
above 10 degrees. This null is generated by the so called in Fig. 30 and Fig. 31.
bend or crooked mode, which can be explained using a modal As is expected, for both modes the results show an increase
analysis, depicted in [1]. in bandwidth, flatter top, and faster roll-off compared with
For TM mode measured and simulation results, shown in single layer configuration. Some differences can be observed
Fig. 26 and Fig. 27, are almost the same as for TE mode with between simulated and measured results. First, for both TE
the difference that the crooked mode is not present anymore, and TM mode reflection coefficients, which are shown in
and the angular stability has been improved. This is because Fig. 28 and Fig. 30, the first resonance frequency measured
the surface rotation axis is not the same as the electric field bandwidth is slightly smaller than the simulated. This can be
orientation. explained by the mismatch alignment of both FSS. Second,
both reflection coefficient are shifted slightly to the left. This
E. DOUBLE LAYER IFSS is due to the addition of the foam that generates a frequency
As can be seen in the results for single layer IFSS, the shift. On the other hand, there are some slight variations
resonance frequencies and bandwidth are nearly at the limit between the measured results between different angles of
compared with the design requirements, shown in Table 1. To incidence. The bandwidth and reflection variation are still
improve the results we need to increase bandwidth, flatter the present due to the same reasons explained before in single
top, and faster roll-off. This can be done by placing two IFSS layer measurements. The crooked mode is still present at 23
in cascade, separated by a distance of 2 mm from each other. GHz, for TE mode. In TM mode the crooked mode is not
To have mechanical stability for measurement purposes, a present and the angular stability has been improved. Using
sheet of 2 mm ROHACELL 31HF was placed between both this cascade configuration we can see that the IFSS works
IFSS. Measured and simulated reflection and transmission fine over the frequency bands of interest.
coefficients at different angles of incidence for TE mode are
10 VOLUME 0, 0000

This work is licensed under a Creative Commons Attribution 4.0 License. For more information, see https://creativecommons.org/licenses/by/4.0/.
 This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI
10.1109/ACCESS.2020.2970271, IEEE Access

Juan A. Vásquez-Peralvo et al.: Inductive Frequency Selective Surface: An applications for dichroic sub-reflectors

0 0

-5
-10
-10
|R| TM [dB]

|T| TE [dB]
-15 -20

-20
-30
-25
=5° Sim. =10° Meas. -40 =5° Sim. =10° Meas.
-30
=5° Meas. =15° Sim. =5° Meas. =15° Sim.
=10° Sim. =15° Meas. =10° Sim. =15° Meas.
-35 -50
15 20 25 30 35 15 20 25 30 35
Frequency [GHz] Frequency [GHz]
FIGURE 26. TM mode measured and simulated reflection coefficient of a FIGURE 29. TE mode measured and simulated transmission coefficient of a
single layer IFSS at oblique incidence. double layer IFSS at oblique incidence.

0 0

-5 -5

-10
-10
-15
|R| TM [dB]
|T| TM [dB]

-15
-20
-20
-25
-25
-30
=5° Sim. =10° Meas. =5° Sim. =10° Meas.
-30 -35
=5° Meas. =15° Sim. =5° Meas. =15° Sim.
=10° Sim. =15° Meas. =10° Sim. =15° Meas.
-35 -40
15 20 25 30 35 15 20 25 30 35
Frequency [GHz] Frequency [GHz]
FIGURE 27. TM mode measured and simulated transmission coefficient of a FIGURE 30. TM mode measured and simulated reflection coefficient of a
single layer IFSS at oblique incidence. double layer IFSS at oblique incidence.

0 0

-5
-10
-10

-15
|T| TM [dB]
|R| TE [dB]

-20
-20

-25 -30

-30
=5° Sim. =10° Meas. -40 =5° Sim. =10° Meas.
-35 =5° Meas. =15° Sim. =5° Meas. =15° Sim.
=10° Sim. =15° Meas. =10° Sim. =15° Meas.
-40 -50
15 20 25 30 35 15 20 25 30 35
Frequency [GHz] Frequency [GHz]
FIGURE 28. TE mode measured and simulated reflection coefficient of a FIGURE 31. TM mode measured and simulated transmission coefficient of a
double layer IFSS at oblique incidence. double layer IFSS at oblique incidence.

VOLUME 0, 0000 11

This work is licensed under a Creative Commons Attribution 4.0 License. For more information, see https://creativecommons.org/licenses/by/4.0/.
 This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI
10.1109/ACCESS.2020.2970271, IEEE Access

Juan A. Vásquez-Peralvo et al.: Inductive Frequency Selective Surface: An applications for dichroic sub-reflectors

VI. CONCLUSION [12] S. S. Roy, T. N. Sekhar, C. S. Padmavathy, K. Bhattachariya, M. N. Kumar

and C. Saha, “Design of double layers dichroic subreflector for S and X
An intertwined inductive FSS has been designed and man- band Cassegrain antenna,” 2016 IEEE Indian Antenna Week (IAW 2016),
ufactured working as a dichroic sub-reflector for Ku, K, Madurai, 2016, pp. 47-50.
and Ka bands. It has been successfully demonstrated that [13] P. Ingvarson, F. S. Johansson and L. E. Pettersson, “A dichroic subreflector
for a communication satellite,” Digest on Antennas and Propagation Soci-
an IFSS using intertwined shapes can be used to work as ety International Symposium, San Jose, CA, USA, 1989, pp. 1088-1091
a multi-band dichroic sub-reflector. E- and H-Fields on the vol.2.
structure were analyzed to understand their filter capabilities, [14] D. Bresciani, S. Contu, C. Bruno and G. Crone, “Design of a 1 m dichroic
subreflector for K/sub u//K/sub a/ frequency bands,” Digest on Antennas
and to visualize the generated inductances and capacitances and Propagation Society International Symposium, San Jose, CA, USA,
when a plane wave impinges on the surface. Additionally, 1989, pp. 1084-1087 vol.2.
an equivalent circuit model for the IFSS has been proposed, [15] C. Bruno, S. Contu, D. Marzi and G. Mascolo, “Design, manufacturing and
testing of a Ku/Ka dichroic subreflector for space communications,” 1993
allowing us to understand the effects of increasing or decreas- Eighth International Conference on Antennas and Propagation, Edinburgh,
ing the size of apertures or strips. A parametric study of the UK, 1993, pp. 178-181 vol.1.
unit cell was performed, showing the independence tuning [16] T. S. Mok, A. M. M. A. Allam, J. C. Vardaxoglou and E. A. Parker,
“Curved and plane frequency selective surfaces: a study of two offset sub-
for both resonant frequencies and bandwidth control. The reflectors,” in Journal of the Institution of Electronic and Radio Engineers,
effects of Photochemical Machining manufacturing process, vol. 58, no. 6, pp. 284-290, September-December 1988.
such as rounded corners and edge bevels has been analyzed [17] M. Han, M. He, H. J. Sun, G. Q. Zhao, and H. Cheng, “Analysis of
cassegrain antenna by using a dichroic sub-reflector,” ICMTCE 2013
and compared with the original IFSS determining its effects. - 2013 IEEE International Conference on Microwave Technology and
Two-sided etching method was selected due to the symmet- Computational Electromagnetics, Proceedings, vol. 1, pp. 58–60, 2013.
rical operation of the IFSS and less impact in frequency [18] T. K. Wu, “Double-square-loop FSS for multiplexing four (S/X/Ku/Ka)
bands,” Antennas and Propagation Society Symposium 1991 Digest, Lon-
and bandwidth shift. Finally, a dual layer configuration was don, Ontario, Canada, 1991, pp. 1885-1888 vol.3.
designed and manufactured improving flatter top and faster [19] R. Orr et al., “Circular Polarization Frequency Selective Surface Operating
roll-off. Simulations and measurements for both, single and in Ku and Ka Band,” in IEEE Transactions on Antennas and Propagation,
vol. 63, no. 11, pp. 5194-5197, Nov. 2015.
double layer IFSS show good agreement and good angular [20] E. C. C. (ECC), “The European Table Of Frequency Allocations And
stability for this application, specially in TM modes. Applications In The Frequency Range 8 . 3 Khz To 3000 Ghz ( ECA Table
),” European Conference of Postal and Telecommunication Administration
(CEPT), no. October, pp. 1–262, 2015.
[21] C. K. Lee and R. J. Langley, “Equivalent-circuit models for frequency-
REFERENCES selective surfaces at oblique angles of incidence,” in IEE Proceedings H
- Microwaves, Antennas and Propagation, vol. 132, no. 6, pp. 395-399,
[1] B. A. Munk, “Frequency Selective Surfaces: Theory and Design,” Hobo-
October 1985.
ken, NJ, USA: Wiley, pp. 1, 35-37, 2000.
[22] Li, W., Wang, C., Zhang, Y., and Li, Y. “A Miniaturized Frequency
[2] J. M. Rigelsford, S. M. Benito and A. Vallecchi, “A tri-band inductive Selective Surface Based on Square Loop Aperture Element,” International
frequency selective surface sub-reflector for satellite communications Journal of Antennas and Propagation, 2014, 1–6.
systems,” The 8th European Conference on Antennas and Propagation [23] D.M.Allen,“Photochemical Machining: from ‘manufacturing’s best kept
(EuCAP 2014), The Hague, 2014, pp. 898-900. secret’ to a $6 billion per annum, rapid manufacturing process”, CIRP
[3] I. Anderson, “On the theory of self-resonant grids,” in The Bell System Annals - Manufacturing Technology, vol. 53, no. 2, pp. 559–572, 2004.
Technical Journal, vol. 54, no. 10, pp. 1725-1731, Dec. 1975.
[4] M. Ohira, H.D eguchi, M. Tsuji, and H. Shigesawa,“Analysis of frequency
selective surface with arbitrarily shaped element by equivalent circuit
model,” Electronics and Communications in Japan (Part II: Electronics),
vol. 88, no. 6, pp. 9–17, jun 2005.
[5] A. L. P. S. Campos, A. M. Martins, and V. A. Almeida Filho, “Synthesis
of frequency selective surfaces using genetic algorithm combined with the
equivalent circuit method,” Microwave and Optical Technology Letters,
vol. 54, no. 8, pp. 1893–1897, aug 2012
[6] M. W. B. Silva and L. C. Kretly, “An efficient method based on equivalent-
circuit modeling for analysis of Frequency Selective Surfaces,” 2013
SBMO/IEEE MTT-S International Microwave and Optoelectronics Con-
ference (IMOC), Rio de Janeiro, 2013, pp. 1-4.
[7] F. Costa, A. Monorchio, and G.Manara, “An overview of equivalent circuit
modeling techniques of frequency selective surfaces and metasurfaces,”
Applied Computational Electromagnetics Society Journal, vol. 29, no. 12,
pp. 960–976, 2014.
[8] M. E. Bialkowski and F.-C. E. Tsai, “RF and Microwave Engineering,”,pp
1702, 2005.
[9] M. G. Floreani, R. E. Zich, and G. Aulisio, “Design and experimental
validations of a new shaped subreflector geometry for EMC oriented
Cassegrain system,” CEEM 2000 - 2nd Asia-Pacific Conference on En-
vironmental Electromagnetics, Proceedings, pp. 167–169, 2000.
[10] A. Vallecchi, J. M. Rigelsford and S. Martin Benito, “An inductive dichroic
surface for reflector antennas,” 2014 IEEE Antennas and Propagation
Society International Symposium (APSURSI), Memphis, TN, 2014, pp.
2070-2071.
[11] A. Vallecchi and A. G. Schuchinsky, “Analytical model of interwoven
spiral arrays,” IET Microwaves, Antennas and Propagation, vol. 8, no. 15,
pp. 1268–1276, 2014.

12 VOLUME 0, 0000

This work is licensed under a Creative Commons Attribution 4.0 License. For more information, see https://creativecommons.org/licenses/by/4.0/.
 This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI
10.1109/ACCESS.2020.2970271, IEEE Access

Juan A. Vásquez-Peralvo et al.: Inductive Frequency Selective Surface: An applications for dichroic sub-reflectors

JUAN A. VÁSQUEZ-PERALVO was born in JONATHAN M. RIGELSFORD (SM’13) re-

Quito, Ecuador. He received his B. Eng. in Elec- ceived the M.Eng. and Ph.D. degrees in electronic
tronics and Telecommunications from the Escuela engineering from the University of Hull, Hull,
Politécnica Nacional, Quito, Ecuador, in 2012; U.K., in 1997 and 2001, respectively. From 2000
M.Sc. degree in Wireless Communication Sys- to 2002, he was a Senior Design Engineer with
tems from The University of Sheffield, UK, in Jaybeam Limited. From late 2002, until 2014 he
2015. Currently he is pursuing his Ph.D degree was a Senior Experimental Officer for the Com-
at Universidad Politécnica de Madrid. His current munications Group within the Department of Elec-
interest are antenna design, Frequency Selective tronic and Electrical Engineering, University of
Surfaces, and Artificial Magnetic Conductors. Sheffield, Sheffield, U.K., where he is became
a Senior Research Fellow, researching in the areas of RF propagation,
biomedical electromagnetics, adaptive antennas, RFID and cyber security.
JOSÉ-MANUEL FERNÁNDEZ-GONZÁLEZ
In February 2019 he moved back into industry to become RF Engineering
(SM’19) was born in Lausanne, Switzerland. He
Lead at Sensata Technologies.
received the Diplôme dIngénieur en Électricité
degree from the École Polytechnique Fédérale de
Lausanne, Lausanne, in 2003, and the Ph.D. de-
gree from the Universidad Politécnica de Madrid,
Madrid, Spain, in 2009. In 2006, he joined the
Centre de Recherches Poly-Grames, l’École Poly-
technique de Montréal, Montreal, QC, Canada,
and in 2007, he joined the Chalmers University
of Technology, Göteborg, Sweden, as a Guest Ph.D. Student. From 2013
to 2019, he was an Assistant Professor with the Universidad Politécnica
de Madrid. In 2018, he was a Fulbright Visiting Researcher with Antenna
Research Group, University of Colorado at Boulder. Since 2019, he has
been an Associate Professor with the Universidad Politécnica de Madrid. He
has authored more than 90 publications in scientific journals, symposium
proceedings and seminars, and he holds 4 patents. He has participated in
more than 35 research projects and contracts. His current research interests
include phased array antennas, RF circuits, and metamaterial structures with
emphasis on planar antenna applications.

PAVEL VALTR received the Ing. (M.Sc.) and

Ph.D. degrees in radio electronics from the Czech
Technical University in Prague, Prague, Czech
Republic, in 2004 and 2007, respectively. From
2007 to 2009, he was a Research Fellow with
the University of Vigo, Vigo, Spain, working on
various topics in electromagnetic wave propaga-
tion including rough surface and vegetation scat-
tering and land mobile satellite channel modeling.
In 2009, he joined the European Space Agency
(ESA/ESTEC), Noordwijk, The Netherlands, as a Postdoctoral Research
Fellow. Since 2012, he has been with the Czech Technical University in
Prague as a Researcher. His research interests include radiowave propagation
for terrestrial and satellite applications, frequency selective surfaces and
computational methods in electromagnetics. Dr. Valtr was the recipient of
the Young Scientist Award of XXVIII General Assembly of the International
Union of Radio Science (URSI) in 2005.

VOLUME 0, 0000 13

This work is licensed under a Creative Commons Attribution 4.0 License. For more information, see https://creativecommons.org/licenses/by/4.0/.