Highly Efficient Generation of Bessel Beams

with Polarization Insensitive Metasurfaces

M UHAMMAD R IZWAN A KRAM , 1 M UHAMMAD Q ASIM M EHMOOD , 2
TAUSEEF TAUQEER , 2 A HSAN S ARWAR R ANA , 2 R ONGHONG J IN , 1
I VAN D. R UKHLENKO, 3 AND W EIREN Z HU 1,*
1 Department of Electronic Engineering, Shanghai Jiao Tong University, Shanghai 200240, P. R. China
2 Department of Electrical Engineering, Information Technology University, Arfa Software Technology
Park, Lahore, Pakistan
3 Modeling and Design of Nanostructures Laboratory, ITMO University, Saint Petersburg 197101, Russia
* weiren.zhu@sjtu.edu.cn

Abstract: We present a generic approach for the generation of pseudo non-diffracting Bessel
beams using polarization insensitive metasurfaces with high efficiency. Cascaded unit cells,
which are fully symmetric, are designed for the complete 2π phase control in the transmission
mode. Based on the topological arrangements of such unit cells, two metasurfaces for the
generation of zero-order (i.e., single phase profile) and first-order (i.e., merger of two distinct
phase profiles) Bessel beams are designed and characterized. Both numerical simulations and
experimental measurements are in agreement with each other, confirming the electromagnetic
characteristics of the reported Bessel beams. Owing to the isotropy of the unit cells and the
rotational symmetry of the arrangements, the proposed metasurfaces are polarization insensitive,
providing a promising avenue for achieving such wave manipulations with any linear or circular
polarization.

© 2019 Optical Society of America under the terms of the OSA Open Access Publishing Agreement

1. Introduction
Bessel beams are paraxial solutions to the free-space Maxwell equations discovered by Durnin
in 1987 [1]. They take their name from their transverse intensity profile, which is described
by Bessel functions of the first kind. Owing to their non-diffracting nature, Bessel beams
have proven themselves useful over a broad spectral domain spanning from microwave to
optical frequencies and for various kinds of applications, including particle trapping [2], tractor
beams [3], microscopy [4], as well as lithography and laser machining [5, 6]. Bessel beams are
also beneficial for high-power applications such as microwave drilling [7,8], wireless charging [9],
high-resolution radar imaging [10], and the creation of through-the-wall radar [11]. Therefore, it
is beneficial to develop simple and practical approaches to generating such beams.
Many techniques have been utilized in the past to realize Bessel beams. These techniques involve
axicons [12], computer generated holograms [13], leaky waveguide modes [14], holographic
screens [15],near-field plates [16], transverse modes excited in a circular wave guide [8] to name
a few. However, these techniques suffer from issues such as low efficiencies, high manufacturing
complexities, and bulky sizes. The design proposed in [8] is based on transverse modes excited
in the circular waveguide, the electrical size of which is 5.8λ × 5.8λ with a thickness larger than
λ. Such a design suffers challenges in fabrication and compatibility issues with the integrated
circuits. Albeit numerous investigations have been focused on the generation of zero-order
Bessel beams, very few efforts have been made to realize higher-order Bessel beams which
have non-diffracting characteristics as well as twisting phase profile. In particular, Comite et.
al [17] have proposed the feasibility of Bessel beams carrying twisting waves, i.e., higher order
Bessel beam, using the radial line slot array (RLSA) antenna [18]. Circular antenna arrays
combined with RLSA [10] were also considered for the generation of Bessel beams carrying an
orbital angular momentum (OAM). Despite these demonstrations, realization of efficient and
polarization insensitive higher order non-diffracting twisted beams is on open problem.
Metasurfaces provide the freedom of introducing abrupt phase-change along the interface
between two media. This unique capability of metasurfaces received significant attention recently
due to their low profiles and great flexibilities in tailoring wavefronts [19–21]. The efficient
realization of wavefront manipulation with metasurfaces require complete phase control from 0
to 2π along with uniform high amplitude; ideally equal to unity. Such complete phase control
can be achieved by tailoring the geometries of artificially engineered subwavelength resonators.
Initial demonstration in this regard included subwavelength gold V-shaped antennas arranged on
a 2D surface [22]. However, the proposed metasurface exhibited meagre performance due to
lower polarization conversation efficiency and substantial ohmic losses associated with the gold.
Later on, it was concluded that the polarization conversion efficiency of such metasurfaces are
fundamentally limited to 25% [23], [24], which is inadequate for practical applications. Reflective
type metasurfaces [25–27] are also studied to achieve high generation efficiency. However, since
the source antenna is placed in front of the metasurface, it may partially block the generated
beam.
Metasurfaces based on Huygens principle were also proposed to control the 2π phase range
of the co-polarized components which have the potential for achieving a 100% efficiency in
theory [28]. Such a metasurface introduces both electric and magnetic discontinuities to control
the electromagnetic field. However, only the TM mode has been considered with Huygens
metasurface so far. It was proposed that the metasurface efficiency can be improved by increasing
the thickness of the metasurface [23]. Recently, a few works have been reported on achieving
higher transmission efficiency using metasurfaces made of multilayer structures [29–31]. These
designs employ the conversion of spin to OAM based on the photon spin Hall effect, where the
cross-polarized component of electromagnetic wave could be utilized efficiently. Later on many
cascaded metasurfaces are proposed to control the 2π phase range with good efficiency [32, 33].
However, these designed metasurfaces are typically limited to a single polarization because
of their asymmetric geometry. The direct comparison between the recently reported planar
transmissive metasurfaces with context to the realized efficiency are listed in Table I.
Metasurface lenses with both polarization insensitiveness and high conversion efficiency are
highly demanded. To the best of our knowledge, the only reported metasurface lens of such
type was proposed by Khorasaninejad et al. [36] and was made of standing dielectric resonators.
However, since such a lens is not a planar structure, its fabrication becomes very difficult and
costly. In this paper, we present highly symmetric planar metasurfaces with high efficiency for
generating Bessel beams, which can be easily manufactured using the standard printed circuit
board technique. Cascaded unit cells are employed in the design as they help to achieve high
efficiency. The unit cells have equal dimensions of (λ0 /3.75) × (λ0 /3.75) × (λ0 /11.97), which are
designed to achieve the full phase control from 0 to 2π with a weighted average efficiency of 71%.
As examples, we present the designs of two metasurfaces for the generation of the zero-order and
first-order Bessel beams. The measured results are in good agreement with the simulated ones,
confirming clearly the electromagnetic characteristics of the Bessel beams of different orders.
The rest of this paper is organized into the following sections. Section II introduces the design
of the cascaded unit cells for the full 0–2π phase control and presents the discussion of the high
transmission amplitudes and the physics behind their resonance modes. Section III describes the
strategy for designing metasurfaces that are capable of generating Bessel beams and presents
simulation results on the generation of the fundamental and first-order Bessel beams. The results
of experimental measurements are given in Section IV finally, the conclusions of our work are
given in Section V.
 Table 1. Efficiencies comparison for transmissive metasurfaces realized in literature
Design feature Isotropic Polarization Transmittance of Application Efficiency
periodic unit cell demonstrated
Single layer PB No CP Theoretical limit
phase [23], [24] = 25% [23], [24]
PB phase with cas- No CP 90% [31], 90% OAM [31], 48% [21]
caded layers [31], [21] [21]
[21]
PB phase, bi-layer No CP 81.6% OAM -
[29]
HuygenâĂŹs No LP 86% [28], 80.87% - -
Principle [28], [34]
[34]
Cascading multi- No LP min. eff. > Lens 62%
ple layers [35] 60%
Cascading multi- Yes LP&CP 42%-95% Bessel Beam J0 =1, 76%
ple layers (this weighted avg J1 =2, 83%
work) = 71%

2. Unit Cell Theory and Design

The unit cell under consideration is shown in Fig. 1. It consists of four circular copper patches,
of which two middle patches have holes of radius r = 1 mm. The radii R of the four patches
are all alike. The three dielectric spacers are made of Rogers RO4003C, which has dielectric
permittivity of 3.48 and thickness t = 0.813 mm. The period of the unit cell is p = 8 mm, which
equals 1/3.75 of the operation wavelength at 10 GHz. The amplitude and phase of the wave
transmitted by the metasurface consisting of such unit cells can be flexibly adjusted by changing
R. It is assumed that the incident plane wave propagates along the z direction, with the electric
field vector along the y axis and the magnetic field vector along the x axis.
The reflection coefficients of the metasurfaces made of periodically patterned unit cells are
plotted in Fig. 2. It is seen that there are three resonance dips in the reflection curves over
the frequency band of interest. It should be noted that in our case it is easier to show the
resonance behaviors in the reflection spectra, because the transmission is quite high within the
entire frequency band of interest. The reflection dips correspond to the resonance peaks of the
transmission spectrum. All the dips are seen to shift to higher frequencies, with the second and
third resonances becoming much closer to each other, as the radius of the patches decreases. The
two high-frequency resonances merge together for R = 3.53 mm. We will further analyze how
these resonances enable one to achieve the full phase control with high efficiency.
As electromagnetic fields are continuous naturally, the discontinuity in phase is achievable via
the introduction of a surface with electric and magnetic currents. These circulating electric and
magnetic currents result from resonances which can ultimately be controlled in metasurfaces. At
our operation (design) frequency of 10 GHz, the first resonance appears when the patch radius is
about R = 3.53 mm. As the patch radius increases, the second and third resonances shift close to
the design frequency while the first resonance shifts towards lower frequencies. The maximum of
three resonances corresponding to the number of the dielectric spacers are produced around the
designed frequency. By sweeping these resonances through the designed frequency, the whole
 Fig. 1. Unit cell on top and two designed metasurfaces at bottom. The geometrical parameters
are r = 1 mm, p = 8 mm, and t = 0.813 mm of the substrate.

phase range from 0 to 2π can be obtained.

While the two resonances are sufficient for achieving the full phase control from 0 to 2π [37–39],
maintaining the high efficiency would be difficult. By fully sweeping the radius of patches from
3.59 to 3.99 mm to adjust the second and third resonances, it is possible to obtain a full 2π range
of phase control. However, the magnitude decreases significantly when R is sufficiently large.
Particularly, the efficiency for R = 3.99 mm is only about 16%. For this reason, we have included
the contribution from another resonance at R = 3.53 mm (i.e., the first resonance) to achieve
the full 0–2π phase control with high efficiency. The magnitude and phase responses of the
designed unit cells are shown in Fig. 3. The proposed design is based on a frequency selective
surface acting as a low pass filter. The response of patch radii below the first resonance (i.e.,
R < 3.53 mm) acts as a non-resonant transparent surface where the transmission magnitude
is high but the transmission phase variation is very slow. The variation starts to increase at
resonances. This allows us to use smaller radii for better transmittance at these particular phase
values. The weighted average efficiency throughout the whole 0 − 2π range is around 71%, with
the highest and lowest magnitudes of the transmission coefficients around 95% and 65%.
The design frequency of the proposed unit cell depends on the periodicity p, the substrate
thickness t, and the hole radius r. By tuning these parameters, the operating frequency can be
shifted as desired. Apart for shifting the design frequency, the thickness t and holes in middle
patches have additional effects on the transmission magnitude and phase. By increasing the
thickness of the substrate, the transmission magnitude can be improved. However, a tradeoff is
necessary to be made between higher transmittance and lower profile. The periodic unit cells
can achieve phase control in the range of 0 − 2π. However, the cases with large phase values
 Fig. 2. Reflection spectra of three metasurfaces with uniform unit cells of different patch
radii.

usually suffer due to low transmittance as a result of sharp cut-off of the stop band. To maintain
relatively high transmittance at these high phase values, minor tuning is performed by decreasing
the capacitance via drilling holes of r = 1 mm in the two middle patches.
In Fig. 4, we show the field distributions of the unit cell at the resonance frequencies for better
understanding the physics behind these resonances. Magnetic fields around and inside the unit
cell are shown in Figs. 4(a)–4(c) and their equivalent models based on boundary conditions [40]
are shown in Figs. 4(d)–4(f). These fields and models are shown for three different resonances
at 10 GHz for three radii R = 3.53 mm, 3.94 mm, and 3.98 mm. Since the thickness of the
entire unit cell is subwavelength (λ0 /11.97), it is reasonable to treat the whole surface as a single
interface. For the patch radius of 3.53 mm, the magnetic field inside all three dielectric spacers
are directed in the +x direction and the currents induced on the middle sheets cancel each other.
The circulating currents outside the unit cell form the +x-directed magnetic fields and, hence,
the unit cell at the first resonance acts as a magnetic dipole. The second and third resonances
have contributions from both electric and magnetic resonances, because these resonances are
close to each other. For the second resonance, the electric response is dominant due to strong
outer circulating magnetic fields and the weak inner magnetic field. This electric response is not
a pure one because of the weaker inner magnetic field that gives contribution to the magnetic
response. For the third resonance, two oppositely directed magnetic fields generate an electric
dipole, while the third magnetic field contributes to a magnetic resonance. Both electric and
magnetic dipoles have nearly equal contributions to the third resonance.
 Fig. 3. Phase/magnitude of co-polarized transmission co-efficient (S21 ) versus radius R of
the patches at 10 GHz.

3. Generation of Bessel Beam with Metasurfaces: Numerical Simulations

In the previous section, unit cells were designed for the full 2π phase control with high efficiency
against geometric parameters. The next step is to design metasurfaces for realizing Bessel beams
of different orders with the predesigned unit cells. As examples, two metasurfaces are designed
for both the zero-order and first-order Bessel beams, which have dimensions of 304 × 304 mm2
along the x and y axes. The electric far field of the Bessel beam generated by the metasurface at
an arbitrary direction b̂ in space can be expressed as
M Õ
N
−j 2π
Õ
λ (®
rmn ·b̂0 )+jφ
E p (b̂) = rmn · b̂0 )A(b̂ · b̂0 )e
Fp A(® 0 , (1)
m=1 n=1

where p is the polarization of the incident electric field, m and n are the center coordinates of the
unit cell on the 2D metasurface, r®mn is the radius vector of the unit cell, Fp is incident plane
wave, A is scattering pattern of the unit cell, b̂0 is the desired beam direction, and φ is the desired
phase profile of the generated Bessel beam. Due to the isotropic nature of the unit cells, such a
metasurface is independent of the incident beam polarization.

3.1. Zero-Order Bessel Beam

The phase profile of Bessel beams is described by Bessel functions of the first kind. The phase
distribution for the zero-order (J0 ) beam is given by [41]
q
φ(x, y) = 2π − (2π/λ0 ) x 2 + y 2 NA, (2)
 Fig. 4. Magnetic field distribution and equivalent current distribution for three different
resonances at 10 GHz: (a) Hy field and (b) equivalent current distribution for R = 3.55 mm;
(c) Hy field and (d) equivalent current distribution for R = 3.94 mm; (e) Hy field and (f)
equivalent current distribution for R = 3.98 mm.

where x and y are the Cartesian coordinates on the metasurface, x 2 + y 2 = r 2 , and NA is the
numerical aperture of the metasurface. The designed metasurface for the zero-order Bessel
beam with NA = 0.7 is shown in Fig. 1 (bottom left side). One can see from Eq. (2), that the
phase profile for the zero-order Bessel beam varies in the radial direction. The metasurface is
distributed into annular strips and each strip has the same phase profile. Each annular strip has
its specific unit cells, whose phase profiles are calculated using Eq. (2). Such a distribution is not
ideally given in square lattices. The circular strips are arranged in such a way that the distances
between adjacent cells are nearly equal to the periodicity and the relative differences are below
4.46% which are even lower for outer circular strips. Such small differences would not affect
significantly on the performance of the metasurface lens.
The performance of the metasurface is assessed numerically using full-wave simulator CST
Microwave Studio. An incident plane wave polarized along the x axis is used to illuminate the
metasurface and the electric field at the other side of the metasurface is recorded. The metasurface
is assumed to be placed in plane z = 0. In Figs. 5(c) and 5(d), we show the magnitude and phase
profiles of the x component of the electric field in plane z = 100 mm. It can be seen that the
most of the power is contained in the main lobe whereas the side lobes are sufficiently weaker.
The phase profile matches well with the analytical phase profile in Fig. 5(b) except a minor
asymmetry around the radial distance of 90 mm. Such an asymmetry is mainly caused by the
 Fig. 5. theoretical zero-order Bessel profile; (a) magnitude and (b) phase; Simulation results
of the zero-order Bessel beam; (c) magnitude and (d) phase distributions of electric field in
plane z = 100 mm; (e) magnitude of electric field for y = 0 and z = 100 mm; (f) electric
field magnitude in plane x = 0.

diffraction of incident waves from the edges of the metasurface lens due to its limited size. It is
worth noting that an ideal Bessel beam should be generated from an infinitely large lens, which is
impossible to be implemented practically. However, using finite metasurface lens, it is possible
to approximate the Bessel profile to few rings as have been demonstrated here. The magnitude
curve for y = 0 in this cut-plane, plotted in Fig. 5(e), clearly shows a profile similar to the Bessel
function. The electric magnitude distribution of the Bessel beam in the yz plane is also shown in
Fig. 5(f). It can be observed that the Bessel beam contains a high-energy region in the form of a
cone. The focal depth of the Bessel beam is given by
r
D 1
fd = − 1. (3)
2 NA2
For D = 304 mm and NA = 0.7 the non-diffracting region of the Bessel beam is found from this
equation to be 153 mm, which nearly coincides with the simulated focal depth of 155 mm. The
full width at half maximum (FWHM) of the zero-order Bessel beam J0 is defined as the waist of
 Fig. 6. theoretical first-order Bessel profile; (a) magnitude and (b) phase;Simulation results
of the first-order Bessel beam; (c) magnitude and (d) phase distributions of electric field in
plane z = 100 mm; (e) magnitude of electric field for y = 0 and z = 100 mm; (f) electric
field magnitude in plane x = 0.

this beam where the intensity of its center bright spot is at half of its peak value,

2.25 0.385λ0
FWHMJ0 = = . (4)
kr NA
This equation gives the zero-order Bessel beam waist of 15 mm, which correlates well with the
simulated value of 17 mm.

3.2. First-Order Bessel Beam

The phase profile of higher-order Bessel beams is similar to that of the zero-order Bessel beam
except that an additional phase term lϕ is introduced,
q
φ(x, y) = 2π − (2π/λ0 ) x 2 + y 2 NA + lϕ, (5)

where ϕ = arctan(y/x) is the azimuthal phase and l is the Bessel beam order. In addition to the
non-diffracting nature, higher-order Bessel beams have helical phase profiles and can be regarded
as non-diffracting OAM beams. The twist of the phase profile is determined by the order l of the
Bessel beam, which also coincides with the OAM mode number. For the first-order (J1 ) beam,
i.e., l = 1, the phase profile undergoes rotation from 0 to 2π through the azimuthal direction.
To realize the first-order Bessel beam, we designed a 304 × 304 mm2 metasurface with a
numerical aperture of 0.7 as shown in Fig. 1 (bottom right side). Such a metasurface consists of
a 38 × 38 array of unit cells. Each unit cell is capable of adding an additional phase shift to the
incident wave given by Eq. (5).
Next, the performance of the designed metasurface is analyzed using numerical simulations
similar to how it was done earlier. The electric field distributions in the transmitted wave are
examined for a normally incident wave at 10 GHz. Figs. 6(c) and 6(d) show the magnitude and
phase of the x component of the electric field at the cut-plane z = 100 mm. Unlike the zero-order
Bessel beam, the first-order Bessel beam has a ring-like magnitude distribution and a helical
phase profile with a topological charge of l = +1. The field profile along the line y = 0 in the
same cut-plane is plotted in Fig. 6(e). The first-order beam is seen to have zero intensity at the
origin. The electric field at the cut-plane x = 0 is shown in Fig. 6(f). It should be noted that
obtained Bessel profiles for both the zero-order and the first-order modes show some variations
in demonstrating circular symmetry. The major difference between the two results is due to the
effect of numerical aperture which is not considered analytically, as pointed out in a recent paper
[41] Another fact is that the transmission magnitudes of different unit cells are not equal, which
contributes to the non-uniform of field distributions along the angle direction.
The first-order beam is non-diffracting until a certain distance. The non-diffracting region of
the first-order Bessel beam is around 150 mm according to the simulations and is nearly the same
as predicted by theory, which is 153 mm for NA = 0.7 and D = 304 mm. The FWHM for the
first-order Bessel beam J1 is twice the distance from the center dark spot to the closest point on
the ring at the half of the maximum intensity,
2.25 0.292λ0
FWHMJ1 = = . (6)
kr NA
The waist of the first-order Bessel beam calculated from Eq. (7) is 13 mm, which is in good
agreement with the numerically obtained value of 15 mm.
The metasurfaces designed for real applications as presented here are not exactly the same as
obtained from periodic unit cells, which require patches with different radii on the sides due to
different phase distribution. Such a discontinuity will be even larger when one phase cycle 0 − 2π
complete and starts from 0. However, the phase distribution calculated according to Eqs. (2) and
(5) are uniform and changes are gradual from one cell to another and relative for consecutive
cells. This leads to a consistent relative phase shift and does not affect much. Issue arises only
when a phase cycle completes and a large patch radius difference is observed for neighboring
cells. This, however, can be improved using higher patch radius for the starting phase value i.e.
R = 3 mm. In spite of these irregularities, the obtained efficiencies are still quite good. For both
Bessel beams, the obtained magnitude and phase profiles are quite matched to those of analytical
profiles.
It is worth noting that the FWHM and the focal depth are directly related to the numerical
aperture of the metasurface. With the increase in NA, it is possible to narrow the waist of the
Bessel beam at the expense of smaller focal depth. Such a trade-off between the length and the
width of the beam should be considered depending on the specific application of the metasurfaces.

4. Generation of Bessel Beam with Metasurfaces: Experimental Results

In order to experimentally assess the performance of our designs, two metasurfaces for the
zero-order and first-order Bessel beams are fabricated using the standard printed circuit board
technique as shown in Figs. 8(a0 ) and 8(a1 ). The experimental setup is shown in Fig. 7. A horn
antenna connected to a vector network analyzer is used for generating quasi-plane waves, and
 Fig. 7. Experimental setup.

a waveguide probe of cross section area of 25 × 15 mm2 is used for detecting the transmitted
electric fields. The probe gives the field value averaged over the entire cross section. The incident
electric field is polarized along the x axis and the incident magnetic field is along the y axis. The
designed metasurfaces are placed at a sufficient distance away from the horn antenna to receive
approximately plane waves. The probe is set to be x-polarized and the transmitted electric field
is recorded at the plane which is 100 mm away from the designed metasurface with a step size of
3 mm along x and x directions.
The results of our experimental assessment of the fabricated metasurfaces presented in Fig. 8
are in good agreement with the results of numerical simulations. The phase profiles of the beams
indicate that there is no azimuthal phase rotation for the zero-order Bessel beam and that the
first-order Bessel beam has a twisting phase profile. The experimental results clearly validate
the simulated results except for a few discrepancies such as that the magnitude profile of the
first-order Bessel beam is not a uniform ring. This is not the case for the zero-order Bessel
beam, because the corresponding metasurface has only radial phase variations and is symmetrical
in nature, whereas the metasurface for the first-order Bessel beam has variations in both the
radial and azimuthal directions. The main reason behind the difference between the simulation
and measurement is due to the misalignment of the source horn antenna to the center. This
causes asymmetry for first-order Bessel beam and focus more energy towards one side of the
ring as shown in Fig. 9. Another source of error could be due to fabrication limitation. The
accuracy of the fabrication technology used is limited to a minimum dimension of 0.01 mm. The
maximum sensitivity of the transmission phase on R in our metasurface design is 23◦ /0.01 mm.
 Fig. 8. (a) Fabricated metasurfaces, (b) magnitude and (c) phase distributions of electric
field in plane z = 100 mm. Subscripts 0 and 1 mark panels representing the zero-order and
first-order Bessel beams.

Considering the fabrication accuracy, such sensitivity may introduce a phase shift no higher than
23◦ , which is lower than the minimum phase difference (60◦ ) between the adjacent cells. Such
an error will not show very significant effect on the metasurface’s performance.
The FWHM of the zero-order Bessel beam is found to be 15 mm, which is slightly smaller than
the value predicted by numerical simulations. On the other hand, the FWHM of the first-order
Bessel beam is 17 mm and is slightly larger than the theoretical prediction. There are two reasons
for this discrepancy. One is the limitation in the fabrication of the minimum dimension and
another is a comparatively large cross section area of the waveguide probe, which makes the
measured response being averaged over the entire waveguide cross section.
We have thus successfully demonstrated the feasibility of generating the zero-order Bessel
beam and the non-diffracting OAM beam by both simulations and experiments. Also noteworthy
is that the proposed metasurfaces are polarization insensitive and provide a promising solution
for achieving Bessel beams with linear polarization in any direction or even circular polarization.
As a concluding remarks, we have analyzed the efficiency of our designed Metasurfaces. The
transmission efficiency, η, of the designed metasurfaces can be estimated using the following
equation, ∬
E®t × H®t∗ dS®
η = ∬S × 100%, (7)
E®i × H® ∗ dS®
S i

where E®t and H®t are the electric and magnetic fields at the transmission side, E®i and H®i are on
the incident side of the metasurface and S is the integration plane of 200×200mm2 size, 100mm
 1.0 1.0
(a) (b) Simulation
Simulation
Ideal Ideal
0.8 Measured 0.8 Measured

0.6 0.6
|Ex|

|Ex|
0.4 0.4

0.2 0.2

0.0 0.0
-160 -80 0 80 160 -160 -80 0 80 160
y (mm) x (mm)
1.0 1.0
(c) Simulation (d) Simulation
Ideal Ideal
0.8 Measured 0.8 Measured

0.6 0.6

|Ex|
|Ex|

0.4 0.4

0.2 0.2

0.0 0.0
-160 -80 0 80 160 -160 -80 0 80 160
y (mm) x (mm)

Fig. 9. Comparison between theory, simulation and measured results: zero-order Bessel
beam(a) x = 0, (b) y = 0; first-order Bessel beam (c) x = 0, (d) y = 0.

away from the metasurface. The simulated efficiency for the J0 beam metasurface is obtained to
be 76% while for the J1 beam metasurface is 83%. The difference of the two obtained efficiencies
is mainly due to the different distributions of the unit cells required for the two different beams.
The detailed discussion on achieved efficiencies for the transmissive metasurfaces based on
periodic units cell have been included in section I. However, to the best of authors knowledge
the transmission efficiency for the whole metasurfaces generating Bessel beams is not discussed
before. Moreover, converging OAM Beam (First Order Bessel Beam) in this work achieved
higher efficiency of 83% compared to prior works [14, 29, 42] which is not more than 50%.

5. Conclusion
In summary, we have presented the design, fabrication, and characterization of planar high-
efficiency metasurface for generating Bessel beams. Cascaded unit cells were designed for the full
phase control in the transmission mode, where the underlying resonance modes were analyzed
based on the electric/magnetic dipole theory. As two examples, metasurfaces for the generation of
the zero-order and first-order Bessel beams were designed and manufactured. The functionalities
of the proposed metasurfaces were examined in both simulations and measurements, which
clearly confirmed the effectiveness in generating Bessel beams of different orders.

References
1. J. Durnin, “Exact solutions for nondiffracting beams. I. The scalar theory,” J. Opt. Soc. Am. A 4, 651–654 (1987).
2. G. Rui, X. Wang, and Y. Cui, “Manipulation of metallic nanoparticle with evanescent vortex Bessel beam,” Opt.
Express 23, 25707–25716 (2015).
3. A. Novitsky, C.-W. Qiu, and H. Wang, “Single gradientless light beam drags particles as tractor beams,” Phys. Rev.
Lett. 107, 203601 (2011).
4. F. O. Fahrbach, P. Simon, and A. Rohrbach, “Microscopy with self-reconstructing beams,” Nat. Photon. 4, 780
(2010).
5. M. Duocastella and C. B. Arnold, “Bessel and annular beams for materials processing,” Laser Photon. Rev. 6,
607–621 (2012).
6. G. Rui, J. Chen, X. Wang, B. Gu, Y. Cui, and Q. Zhan, “Synthesis of focused beam with controllable arbitrary
homogeneous polarization using engineered vectorial optical fields,” Opt. Express 24, 23667–23676 (2016).
7. E. Jerby and V. Dikhtyar, “Drilling into hard non-conductive materials by localized microwave radiation,” in Adv.
Microwav. Rad. Freq. Process., (Springer, 2006), pp. 687–694.
8. S. Costanzo, G. D. Massa, A. Borgia, A. Raffo, T. W. Versloot, and L. Summerer, “Microwave Bessel beam launcher
for high penetration planetary drilling operations,” in Proc. 10th Eur. Conf. Antennas Propag., (2016), pp. 1–4.
9. G. A. Landis, “Charging of devices by microwave power beaming,” (2005). US Patent US6967462B1.
10. A. Mazzinghi and A. Freni, “Simultaneous generation of pseudo-Bessel vortex modes with a RLSA,” IEEE Antennas
Wirel. Propag. Lett. 16, 1747–1750 (2017).
11. S. Costanzo and G. Di Massa, “Near-field focusing technique for enhanced through-the-wall radar,” in Proc. 11th Eur.
Conf. Antennas Propag., (2017), pp. 1716–1717.
12. S. Monk, J. Arlt, D. Robertson, J. Courtial, and M. Padgett, “The generation of Bessel beams at millimetre-wave
frequencies by use of an axicon,” Opt. Commun. 170, 213–215 (1999).
13. J. Salo, J. Meltaus, E. Noponen, J. Westerholm, M. M. Salomaa, A. Lonnqvist, J. Saily, J. Hakli, J. Ala-Laurinaho,
and A. V. Raisanen, “Millimetre-wave Bessel beams using computer holograms,” Electron. Lett. 37, 834–835 (2001).
14. M. Ettorre and A. Grbic, “Generation of propagating Bessel beams using leaky-wave modes,” IEEE Trans. Antennas
Propag. 60, 3605–3613 (2012).
15. D. Blanco, J. L. Gómez-Tornero, E. Rajo-Iglesias, and N. Llombart, “Holographic surface leaky-wave lenses with
circularly-polarized focused near-fields — Part II: Experiments and description of frequency steering of focal length,”
IEEE Trans. Antennas Propag. 61, 3486–3494 (2013).
16. M. F. Imani and A. Grbic, “Generating evanescent Bessel beams using near-field plates,” IEEE Trans. Antennas
Propag. 60, 3155–3164 (2012).
17. D. Comite, W. Fuscaldo, S. C. Pavone, G. Valerio, M. Ettorre, M. Albani, and A. Galli, “Propagation of nondiffracting
pulses carrying orbital angular momentum at microwave frequencies,” Appl. Phys. Lett. 110, 114102 (2017).
18. A. Mazzinghi, M. Balma, D. Devona, G. Guarnieri, G. Mauriello, M. Albani, and A. Freni, “Large depth of field
pseudo-Bessel beam generation with a RLSA antenna,” IEEE Trans. Antennas Propag. 62, 3911–3919 (2014).
19. M. Kang, H.-T. Wang, and W. Zhu, “Wavefront manipulation with a dipolar metasurface under coherent control,” J.
Appl. Phys. 122, 013105 (2017).
20. W. Luo, S. Xiao, Q. He, S. Sun, and L. Zhou, “Photonic spin Hall effect with nearly 100% efficiency,” Adv. Opt.
Mater. 3, 1102–1108 (2015).
21. S. Jiang, C. Chen, H. Zhang, and W. Chen, “Achromatic electromagnetic metasurface for generating a vortex wave
with orbital angular momentum (OAM),” Opt. Express 26, 6466–6477 (2018).
22. N. Yu, P. Genevet, M. A. Kats, F. Aieta, J.-P. Tetienne, F. Capasso, and Z. Gaburro, “Light propagation with phase
discontinuities: Generalized laws of reflection and refraction,” Science 334, 333–337 (2011).
23. F. Monticone, N. M. Estakhri, and A. Alù, “Full control of nanoscale optical transmission with a composite
metascreen,” Phys. Rev. Lett. 110, 203903 (2013).
24. A. Arbabi and A. Faraon, “Fundamental limits of ultrathin metasurfaces,” Sci. Rep. 7, 43722 (2017).
25. M. L. Chen, L. J. Jiang, and W. E. I. Sha, “Artificial perfect electric conductor-perfect magnetic conductor anisotropic
metasurface for generating orbital angular momentum of microwave with nearly perfect conversion efficiency,” J.
Appl. Phys. 119, 064506 (2016).
26. S. Yu, L. Li, and N. Kou, “Generation, reception and separation of mixed-state orbital angular momentum vortex
beams using metasurfaces,” Opt. Mater. Express 7, 3312–3321 (2017).
27. J. Lončar, A. Grbic, and S. Hrabar, “A reflective polarization converting metasurface at X-band frequencies,” IEEE
Trans. Antennas Propag. (2018).
28. C. Pfeiffer and A. Grbic, “Metamaterial Huygens surfaces: Tailoring wave fronts with reflectionless sheets,” Phys.
Rev. Lett. 110, 197401 (2013).
29. M. L. Chen, L. J. Jiang, and W. E. I. Shao, “Ultrathin complementary metasurface for orbital angular momentum
generation at microwave frequencies,” IEEE Trans. Antennas Propag. 65, 396–400 (2017).
30. J. Yuan, Y. Zhou, R. Chen, and Y. Ma, “Photonic spin Hall effect with controlled transmission by metasurfaces,” Jap.
J. Appl. Phys. 56, 110311 (2017).
31. C. Liu, Y. Bai, Q. Zhao, Y. Yang, H. Chen, J. Zhou, and L. Qiao, “Fully controllable Pancharatnam-Berry metasurface
array with high conversion efficiency and broad bandwidth,” Sci. Rep. 6, 34819 (2016).
32. N. K. Grady, J. E. Heyes, D. R. Chowdhury, Y. Zeng, M. T. Reiten, A. K. Azad, A. J. Taylor, D. A. Dalvit, and
H.-T. Chen, “Terahertz metamaterials for linear polarization conversion and anomalous refraction,” Science 340,
1304–1307 (2013).
33. C. Pfeiffer and A. Grbic, “Controlling vector Bessel beams with metasurfaces,” Phys. Rev. Appl. 2, 044012 (2014).
34. J. P. Wong, M. Selvanayagam, and G. V. Eleftheriades, “Polarization considerations for scalar huygens metasurfaces
 and characterization for 2-d refraction,” IEEE Trans. Microw. Theory Techn. 63, 913–924 (2015).
35. C. Pfeiffer and A. Grbic, “Cascaded metasurfaces for complete phase and polarization control,” Appl. Phys. Lett.
102, 231116 (2013).
36. M. Khorasaninejad, A. Y. Zhu, C. Roques-Carmes, W. T. Chen, J. Oh, I. Mishra, R. C. Devlin, and F. Capasso,
“Polarization-insensitive metalenses at visible wavelengths,” Nano Lett. 16, 7229–7234 (2016).
37. I. Staude, A. E. Miroshnichenko, M. Decker, N. T. Fofang, S. Liu, E. Gonzales, J. Dominguez, T. S. Luk, D. N.
Neshev, I. Brener et al., “Tailoring directional scattering through magnetic and electric resonances in subwavelength
silicon nanodisks,” ACS Nano 7, 7824–7832 (2013).
38. J. Cheng, D. Ansari-Oghol-Beig, and H. Mosallaei, “Wave manipulation with designer dielectric metasurfaces,” Opt.
Lett. 39, 6285–6288 (2014).
39. M. Decker, I. Staude, M. Falkner, J. Dominguez, D. N. Neshev, I. Brener, T. Pertsch, and Y. S. Kivshar, “High-efficiency
dielectric Huygens surfaces,” Adv. Opt. Mater. 3, 813–820 (2015).
40. R. F. Harrington, Time-Harmonic Electromagnetic Fields (McGraw-Hill, 1961).
41. W. T. Chen, M. Khorasaninejad, A. Y. Zhu, J. Oh, R. C. Devlin, A. Zaidi, and F. Capasso, “Generation of
wavelength-independent subwavelength Bessel beams using metasurfaces,” Light. Sci. Appl. 6, e16259 (2017).
42. S. Jiang, C. Chen, H. Zhang, and W. Chen, “Achromatic electromagnetic metasurface for generating a vortex wave
with orbital angular momentum (oam),” Opt. Express 26, 6466–6477 (2018).