Wideband 5G Antenna Gain Enhancement Using a Compact Single-Layer Millimeter Wave Metamaterial Lens

This paper presents a very compact, wideband, and enhanced-gain antenna for 5G applications. A simple single-layer millimeter wave (mm-wave) metamaterial lens (meta-lens) is used to improve the gain, aperture efficiency, and gain bandwidth of a slotted-patch antenna over a wide range of frequencies from 25 GHz to 31 GHz. The lens exhibits a metamaterial negative refractive index behavior, which is attributed to a substantial gain enhancement of around 4–5 dBi over the whole band compared to the gain values of the slotted patch antenna alone. The lens’s unit cell comprises a simple single-layer split ring resonator (SRR) whose dimensions are carefully chosen to improve transmitted power and suppress absorbed and reflected power. The meta-lens consists of <inline-formula> <tex-math notation="LaTeX">$8\times 8$ </tex-math></inline-formula> subwavelength SRR unit cells. Each cell has an area of <inline-formula> <tex-math notation="LaTeX">$1.6\times 1.6$ </tex-math></inline-formula> mm2, it is located in the near-field region closely above a slotted patch antenna to produce a total antenna size of <inline-formula> <tex-math notation="LaTeX">$12.8\times 12.8\times 7.27$ </tex-math></inline-formula>mm3 (<inline-formula> <tex-math notation="LaTeX">$1.2~\lambda \times 1.2~\lambda \times 0.68~\lambda $ </tex-math></inline-formula>, where <inline-formula> <tex-math notation="LaTeX">$\lambda $ </tex-math></inline-formula> is the free space wavelength at 28 GHz). The maximum gain of the proposed antenna is 12.7 dBi, the 1 dB gain bandwidth is 18%, the maximum aperture efficiency is 92%, and the −10 dB impedance bandwidth (10 dB B.W.) is 17%. This excellent combination of essential metrics is hard to realize at mm-wave using narrowband antenna structures (microstrip patch antennas), and the aperture efficiency is the highest thus far for such a class of antennas.


I. INTRODUCTION
Despite the mm-wave 5G bands, like the 28 GHz band, have been licensed by the Federal Communication Commission (FCC) and many other regulatory bodies across the world for years thus far [1], the existing technology (4G/WiFi/WLAN) is still using microwave sub-6 GHz bands. This delay in realizing a true high data rate system at mm-wave bands is caused by some challenges associated with the design and implementation of mm-wave systems and antennas [2]. For mm-wave 5G applications, antennas need to generally possess a high The associate editor coordinating the review of this manuscript and approving it for publication was Debabrata K. Karmokar . gain, wide bandwidth, diversity of polarization, and possibly a reconfigurable (agile) radiation pattern; while remaining very compact to fit into small packages [2], [3]. These characteristics are often inversely related; therefore, the structured artificial bi-and three-dimensional materials such as metamaterials and Electromagnetic/Photonic Bandgap Materials (E/PBG) can offer extra-ordinary solutions for these purposes [4], [5].
Among the metamaterial applications in the antenna field, antenna gain and bandwidth enhancements have been reported in [6], [7], [8]. In this regard, several works have introduced a superstrate of metamaterial/metasurface characteristics on top of a single radiating element to boost the overall antenna gain; while maintaining a small antenna footprint. This was mainly done to avoid the conventional way of achieving higher gain values by adding more radiating antenna elements hence, obtaining large aperture sizes. In [9], a single layer metamaterial superstrate comprises of 4×4 ring elements is placed on top of a diagonal radiating slot to improve its gain and operational bandwidth. The 10 dB B.W. achieved in [9] covers the entire licensed mm-wave range 24. 2-29.5 GHz with an actual small footprint of 1.1 λ×1.1 λ (at 27.5 GHz5GHz); however; the maximum achieved gain was only 11 dBi, which corresponds to an aperture efficiency of 82% for the antenna. Another exciting design reported in [10] has also utilized a single layer mm-wave FSS superstrate placed above a radiating dielectric resonator antenna (DRA) with an air gap of 0.5 λ in between. The 10 dB B.W. covers the frequency band from 26 GHz to 30.3 GHz with a maximum measured gain of 15.7 dBi. The large footprint of 2.9 λ×9λ× 2.9 λ of the superstrate in [10] has contributed to the gain improvement; however, this resulted in an aperture efficiency below 50% and the predicted 1 dB gain B.W. is only 4% (estimated from the gain versus frequency figure). A three-layer dual split ring mm-wave metamaterial lens is used in [4] to improve the gain of a magneto-electric dipole. The antenna in [4] supports circular polarization and has a wide 10 dB B.W. that spans from 29.5 GHz to 37 GHz; however, the maximum gain value achieved is 10.5 dBi. The design also utilizes multilayers and hence it requires advanced manufacturing facilities. A Substrate Integrated Waveguide (SIW) is used in [11] to excite a low-profile metasurface comprising of Maltese cross-shaped edge patches. Two frequency bands were covered efficiently namely, 23.7-29.2 GHz and 36.7-41.1 GHz however; the maximum gain values achieved are 7.4 dBi and 10.9 dBi in the lower and upper bands, respectively for an antenna footprint area of 2.8 λ× 1.5 λ m 2 .
Metasurfaces have also been utilized at none mm-wave frequency bands for gain and bandwidth enhancement of a single radiating antenna element. Interesting examples of such works are reported in [12], [13], [14], and [15]. In [12], a metamaterial surface was used as a superstrate above a patch antenna designed at 10 GHz. The maximum gain achieved is 12.5 dBi; while the -10 dB fractional B.W. is below 10%. The footprint is around 2λ×2λ at 10 GHz. In [13] a dual-layer meta-lens is used to increase the gain and 10 dB impedance B.W. of a Reactive Impedance Surface (RIS) backed patch antenna. The design band in [13] covers 6.5-8.5 GHz with a maximum reported gain of 13.9 dBi and a -10 dB fractional B.W. of 13% (7.2 -8.2 GHz). The calculated 1 dB gain bandwidth from the available data in [13] is only 3%. Another dual-layer metasurface, reported in [14], is placed on top of an orthogonal shape aperture to improve its performance. It was designed to operate in the 4.5 -7.5 GHz band with a maximum gain of 7 dBi and a fractional B.W. of 36% (5.04 -7.21 GHz). A single-layer metamaterial lens is designed in [15] for LTE 46/WLAN applications in the 5-7 GHz band. The lens was placed on top of a truncated patch with a truncated ground plane. The maximum gain was 5.54 dBi with a 10 dB B.W. close to 7% (5. . Other techniques that were used to improve mm-wave single-element antenna gain; while maintaining a small footprint include Electromagnetic Band Gaps (EBGs) and Defected Groundplane (D.G.) [5], [16], [17]. The EBGs can be mainly used for gain improvement due to the known narrowband nature of resonant cavities [18]. In [16], a 2 × 2 printed Yagi structure was reinforced by an EBG structure with a suspended strip line to improve gain and fractional B.W. However, the structure is comprised of four layers designed to operate in the 25 -33 GHz band. Also, its footprint is quite large (13 λ×13 λ), produced a maximum gain of 11.9 dBi, and has a fractional bandwidth of 17%. A patch with a defected ground plane was proposed in [17] for mmwave applications (25)(26)(27)(28)(29)(30)(31)(32)(33)(34). By using a D.G., a fractional bandwidth of 22% was achieved (26.5-32.9 GHz), however, the maximum reported gain value was as low as 5.6 dBi. Element shaping was also used to improve antenna gain and bandwidth characteristics at different frequency bands with less gain improvement than metamaterial lenses and EBGs. Examples of such works are reported in [19], [20], and [21].
For mm-wave applications, due to the need for high gain and small antenna footprint, far-field parameters like the aperture efficiency of the antenna, and the 1 dB gain bandwidth must be seen as important as the impedance matching (the −10 dB B.W.). The aperture efficiency can be seen as the appropriate figure of merit in evaluating a mm-wave antenna performance rather than the maximum gain. The aperture efficiency relates the maximum achieved gain to the area of the antenna in wavelengths (λ) [12] hence; high aperture efficiency indicates excellent utilization of an antenna footprint. This indeed fits well with the need for antennas of high gain; while occupying small areas for mm-wave applications. Another equally important parameter is the 1 dB gain bandwidth. This parameter unfortunately is not given enough consideration in many of the reported works in the open literature. It indicates the antenna's far-field performance's stability against frequency change, which is crucial at mmwave frequencies. A narrow 1 dB gain bandwidth means a rapid change from the maximum gain achieved at close offcenter frequencies. This also indicates considerable deformation occurs on the antenna far-field radiation pattern at off-center frequencies. Therefore, it is quite crucial at mmwave frequencies to realize an excellent combination of the aforementioned figures of merit using a simple and compact antenna.
In this paper, we propose, analyze and validate a simple single-layer metamaterial lens with negative refractive index behaviour. This behaviour is shown to improve the focusing characteristics of the lens as compared to other lenses available in the open literature leading to owning the highest aperture efficiency; while maintaining a very competitive 1 dB gain B.W. as well as a 10 dB broad 10B impedance bandwidth. Due to the low-loss nature of the proposed metalens, the radiation efficiency values sit above 90% at many frequencies with a maximum value of 95% at 28 GHz. This paper is organized as follows: Section II presents the design considerations and analysis of the unit cell used to design the metamaterial lens. An analysis of the unit-cell negative refractive index behaviour and its equivalent circuit model is also discussed in Section II. Section III addresses the design of the entire 8 × 8 element meta-lens and illustrates how it is used to improve the performance of a singleslotted patch antenna. Section IV presents a discussion of simulated and measured results, antenna performance, and other antenna radiation characteristics. Section V summarizes the current state-of-the-art with respect to the proposed work. Finally, Section VI concludes the paper.

II. DESIGN AND ANALYSIS OF THE LENS UNIT-CELL
This section presents the geometry of the proposed metalens unit cell with a ring element. The transmission characteristics of an infinite periodic structure comprised of the proposed unit cell is analyzed, and accordingly, the selection of design parameters is justified. The negative refractive index behaviour of the meta-lens unit-cell is then discussed to reason its novel focusing ability that makes it optimum for gain enhancement at mm-wave frequencies.

A. THE PROPOSED LENS UNIT-CELL DESIGN CONSIDERATIONS AND TRANSMISSION CHARACTERISTICS
Despite the use of split ring resonators in antennas design is not new [4], [6], the use of a simple split ring resonator printed on a highly capacitive material arranged in a subwavelength lattice for millimeter wave applications, can be considered new, in particular, for mm-wave 5G applications. Fig. 1(a) shows the unit cell model of the proposed element as part of a periodic arrangement created using the ANSYS High-Frequency Structure Simulator (HFSS) [22]. The unit-cell is placed inside a waveguide with perfect magnetic conductors (PMC) assigned as a boundary condition to the y-z walls of the model. In contrast, perfect electric conductor boundaries (PEC) are applied to the walls in the x-z plane orthogonal to the incident electric fields. A linearly y-polarized TEM wave is then incident from wave port #1 on the top side, allowing for transmission and reflection coefficients prediction at both port#1 and port#2.
The reference planes of both ports are defined at the top surface of the ring element. The mm-wave lens unit cell is composed, as shown in Fig. 1(b), of a split ring of length L ring , track width of W ring , metallization thickness t, and gap width g. The substrate used is Rogers TMM 10 (tm), which has a relative permittivity of 9.2, tangential losses of 0.0022, and a commercially available thickness of d = 1.27 mm. Unlike the design of mm-wave antennas of quite large aperture antennas, like transmitarrays and reflectarrays [23], [24], the use of such high capacitive material is economically feasible for small antennas; while the high relative permittivity  is utilized to accommodate more radiating elements within a small footprint area. This can be seen as a way to achieve a sort of miniaturization as it reduces the unit-cell size a and hence the length of the ring length L ring . This also fits well with the lens's targeted negative refractive index behavior as the use of a subwavelength (a small fraction of a free space wavelength) unit-cell can support the restoration of the amplitude of evanescent waves and therefore enable subwavelength focusing [25]. The proposed equivalent circuit, shown in Fig. 2(a), helped much in understanding the proposed element operation and the selection of all design element parameters namely, W ring , a, S and g except for the SRR length (L ring ). L ring was selected to be a quarter effective wavelength (λ eff /4) at 28GHz to support a strong resonance behavior, which is manifested by the current distribution shown in Fig. 2(b). The simple expressions that are used in calculating the effective wavelength, and consequently L ring [26] are λ eff = (λ o / ε eff ), L ring = (λ eff /4), respectively.
At a center frequency of 28 GHz, the calculated ε eff and λ eff , are 5.1 mm and 4.45 mm, respectively. The selected value of L ring in this work is 1.2 mm. In [27] an equivalent circuit model was proposed and verified for an array of identical square ring elements (loops) printed on a grounded dielectric substrate based on models reported in [28]. Rationally, in [27] and [28], inductors were used to represent the metallic arms of the square ring element along the polarization direction, and gaps between adjacent elements were replaced by coupling capacitors.
For an array of very closely spaced split rings, we conjecture that the scenario is similar but with the gap introduced to tune the strong coupling between adjacent elements and to support the negative refractive index behaviour as explained in part B of this Section. A cursory look at the vector surface current distribution shown in Fig.2 (b), which is shown for one element as part of an infinitely periodic environment of identical elements, reveals explicitly that the vertical arms of the split ring resonator accommodate strong currents oriented along the direction of polarization. Not surprisingly, these time-varying currents will generate magnetic fields transverse to the current path and consequently to the incident electric field thus L 1 and L 2 are used to represent those arms in Fig. 2(a). On the other hand, the current distribution on the top and bottom arms of the ring, curls oppositely around the axis of symmetry (y-axis), thus the x-components of these currents cancel each other out; while the y-components survive and generate y-polarized electric near fields that are responsible for coupling between adjacent elements along the polarization direction. This explanation ties well with the polarity of charges indicated by current distribution; heads on the top arm and tails on the bottom arm. One can then introduce capacitors C 1 and C 2 between elements along the polarization direction to the left and right sides of the slit (the gap of width g). The slit will be shown useful in tuning the said coupling. The equivalent circuit model produced a periodic impedance surface comprised of an inductance L = L 1 + L 2 with total coupling capacitance C = C 1 + C 2 . Luckily, while the length of the element L ring is selected to yield resonance behavior, as pointed out earlier, expressions given in [27] and [28], for L and C reveal that W ring , a, and S can be used to control the values of L and C, respectively, and hence control the impedance on the lens surface.
Since a and S control the spacing between elements, as revealed by Fig. 2(a), while L ring is constant, the change in the unit cell size (a) would control C values. This occurs as  the change in S is directly covered by the unit cell size change. On the other hand, adding a non-radiating slit with width g can add another degree of freedom to tune the capacitance value as it controls the width of capacitive edges. For constant L ring , and selected a and g values, W ring would be the main parameter to tune the element inductance. It is illustrated next how parameters a, and hence S, g, and W ring are selected based on their effects on the transmission characteristics of the lens unit cell.
First, all designed parameters are fixed, except the unitcell size a, as given in Table 1, and the effect of changing the unit-cell size is studied. Fig. 3(a) shows the predicted transmission coefficient magnitude at different unit-cell sizes. It is noticeable that the transmission magnitude improves with the increase in the unit-cell size. It exceeds −1 dB over a VOLUME 11, 2023 wideband for a = 1.6 mm and a = 1.7 mm, respectively, and it even approaches −0.5 dB and beyond at some other frequencies. This means the transmission magnitude exceeds 0.9 to 0.95, this will lead to excellent power transmission through the meta-lens at a wide range of frequencies. A unitcell size of a = 1.6 mm is selected for the lens design in this work as it exhibits better performance below 28 GHz while having a smaller subwavelength period. The transmission magnitude improvement against the increase in the unitcell size stems from the fact that it reduces the coupling capacitance and produces more balance with the inductance generated along the loop.
A lens design also requires a slow change in the transmission coefficient phase versus frequency to avoid significant phase errors on a lens's surface. Phase errors lead to farfield deformations and hence poor gain values, associated with quite a narrow gain B.W., at off-center frequencies. The phase behavior versus frequency at several unit-cell sizes exhibited in Fig. 3(b), shows linear and stable variation with a phase change of around 8 degrees/GHz. This quite desirable behaviour will lead to stable far-field radiation hence, broadband gain operation.
Next, the effect of the slit width (g) and the ring width (W ring ) on the lens unit-cell transmittance is investigated. Fig. 4 shows progressive improvement in S 12 magnitude as W ring increases. For W ring = 0.4 mm, the S 12 magnitude is very promising, with values exceeding −1 dB from 27 to 30 GHz and sitting above −1.5 dB from 26 to 31 GHz. Smaller W ring values, namely, W ring = 0.3 mm and 0.2 mm are associated with higher inductance values that degrade S 12 magnitude, especially beyond 27 GHz. On the other hand, the change in slit width (g) exhibits less effect on S 12 magnitude as compared to W ring . For g = 0.2 mm, the S 12 magnitude is improved compared to g = 0.1 mm. This fine-tuning in S 12 magnitude, which is concluded from the difference between solid and dashed lines in Fig. 4, looks more significant at smaller W ring values as the slit width balances the larger inductance effect (at smaller W ring ). Therefore, W ring = 0.4 mm and g = 0.2 mm were selected for the meta-lens unitcell design as listed in Table 1.
The calculated transmitted, reflected, and absorbed (lost) powers for the meta-lens unit-cell are shown in Fig. 5. Expressions given in [29], which count for the power loss, were used to calculate the powers from the extracted S-parameters matrix of the lens full-wave model. The transmitted power shown in Fig. 5 (a), confirms the excellent transmittance of the lens within the targeted frequency band. It is evident that more than 80 % of the power passed through the lens in the band from 27-30 GHz whereas the maximum transmitted power exceeded 85 % between 28-29 GHz.
Considering that the lens is designed with negative refractive index behavior, even the 70% transmitted power between 26-27 GHz and 30-31 GHz would yield good operation due to the meta-lens excellent power focusing. The behavior of the transmitted power and reflected power against W ring change, ties well with the S 12 magnitude shown in Fig. 5. It is quite  clear that tuning in W ring is crucial for successful lens design. Finally, the absorbed power (consumed power), shown in Fig.5 (b) confirms the very low loss nature of the proposed meta-lens despite the copper loss and tangent loss all included in the full-wave simulation model. The maximum power loss resides in the range of 1-1.25% over the entire band. This indicates that the meta-lens will not degrade the radiation efficiency of the exciting antenna element as detailed in Section III.

B. THE META-LENS UNIT-CELL NEGATIVE REFRACTIVE INDEX BEHAVIOR
A negative index material is a material that generates negative effective permittivity and negative effective permeability concurrently over some frequency band [30]. A negative refractive index allows a flat lens to bring E.M. waves into focus, whereas conventional materials always require curved surfaces to focus E.M. waves [25]. Although such a material does not exist, its characteristics can be artificially emulated using a certain type of radiating elements when printed on some dielectric slabs, which was experimentally demonstrated in [31]. The attractive focusing ability of negative index material, and hence the potential of using it as a focusing lens, was reported in [25]. Regarding Fig. 6, it was established in [25] that the refraction angle (α) is always negative with respect to the surface normal hence, the E.M. wave diverging from the source, which has an incident angle (β), will be converged effectively beyond the lens interface leading to more wave collimation at close point(s) from the lens surface. Such a phenomenon will be illustrated in Section III using the predicted total electric nearfield near the lens surface. Despite the work done in [25] deals with optical frequencies, optical principles were shown valid, despite being approximate, for antennas' design at mm-wave frequencies [32], [33]. The calculated refractive index and effective constitutive parameters for the proposed meta-lens unit cell shown in Fig. 1, using the design parameters given in Table 1, are depicted in Fig. 7. Equations (1), (2) and (3) [34] given below were used to calculate the effective permittivity (ε eff ), effective permeability (µ eff ) and the effective refractive index (n eff ) from the extracted S-parameters matrix at many frequencies from 25 GHz to 31 GHz.
where k and d are the propagation constant and the thickness of the metamaterial lens, respectively. Fig. 7(c) shows that the structure has a negative index region that extends over VOLUME 11, 2023 the whole frequency band from 25 to 31 GHz. Therefore, the meta-lens own negative refractive behavior, as confirmed by the results in Fig. 7. This would, as pointed out earlier, entail negative refractive angles that allow for more convergence and focusing of the incident waves; hence, the resulting refraction angle α must be negative with respect to the surface normal. Consequently, this will improve the incident wave focusing leading to real antenna gain and aperture efficiency enhancement. Fig. 7(a) and Fig. 7(b) exhibit drastic change in ε eff and µ eff near resonance (28 GHz), and more interestingly, the real ε eff and µ eff have close values from 27 to 29 GHz. This may indicate good matching to free space impedance and hence better transmission characteristics within the said frequency band. Since only one split ring element is used in the proposed lens, while the incident electric field is polarized along the slit, one can observe the meta-lens produces a full negative electric response as shown in Fig. 7(a).

III. WIDE-BAND HIGH-GAIN META-LENS ANTENNA CONFIGURATION AND CHARACTERIZATION A. ANTENNA CONFIGURATION
To use the proposed meta-lens as a superstrate for gain and hence aperture efficiency enhancement, the unit-cells are arranged in an array with a periodicity of 8 × 8 to form a single layer of negative refractive index material. The lens is placed in the near-field region closely above a slotted microstrip patch antenna at a distance of h as shown in the antenna perspective view in Fig. 8(b). This study uses a slotted patch reported in [36] to excite the meta-lens. As shown in Fig. 8(a), the slotted patch has two U-shaped slots engraved on a rectangular patch that is fed by a microstrip line and printed on a Rogers RT/Duroid 5880 substrate with a dielectric constant ε r = 2.2 and a thickness hs = 0.508 mm. All slotted patch parameters dimensions are given in the Fig. 8 caption.
The spacing, from the radiator (slotted patch) to the bottom side of the metamaterial superstrate is h. This is optimized to obtain better antenna performance as detailed in part B of this section. The metamaterial superstrate layer is of a standard thickness d = 1.27 mm. The antenna ground plane and metamaterial superstrate have dimensions of 12.8 × 12.8 × 7.27mm 3 (1.2 λ×1.2 λ× 0.68 λ, where λ is the free space wavelength at 28 GHz). These dimensions are changed a bit upon fabrication hence, calculated aperture efficiency was made based on a footprint of 1.25 λ×1.3 λ.
The slotted patch is used in this work to improve the poor matching obtained from a simple patch designed at 28 GHz. The slots improve the impedance matching by altering the current distribution on the patch. The extracted surface currents on the simple patch and the slotted patch calculated at 28 GHz are shown in Fig. 9(a) and Fig. 9(b), respectively. It is evident that adding slots to the patch changed the current distribution and hence improved the resonant effect by increasing the current densities at the radiating edges of the slotted patch as compared to the simple patch.
Upon the optimization of the U slots' location and dimension, the S 11 was significantly improved over the entire band and the radiation from the two U slots improved the patch antenna gain as shown in Fig.10 (a) and Fig. 10(b), respectively.
The selection of the 8 × 8 array size for the proposed metalens was at first mainly governed by the antenna aperture size, which equals 13mm, and the optimized unit cell size, which equals 1.6mm selected based on the results shown in Fig. 3; these figures entails an array size of 8×8 (13/1.6 = 8). However, it is tentatively known in advance that a smaller array size than 8 × 8 may result in a smaller gain due to the less focusing of the radiation emanating from the slotted patch antenna while a larger array will be oversized with respect to the slotted patch antenna and hence may reduce the antenna aperture efficiency. To investigate the effect of the array size on the meta-lens performance, 6 × 6 and 10 × 10 array meta-lenses were designed and simulated. Fig. 11(a) and Fig. 11(b) show the 3D models of the 6 × 6 meta-lens  and the 10 × 10 meta-lens, respectively. Fig. 12 exhibits the calculated gain versus frequency when the meta-lens of the concerned sizes are placed on top of the slotted patch antenna. It is clear that the 8 × 8 array substantially outperforms the 6 × 6 array meta-lens while it provides overly close to better performance as compared to the 10×10 array meta-lens. This makes the 8 × 8 array size a viable choice for the proposed SRR meta-lens.

B. COMPLETE ANTENNA CHARACTERIZATION AND PERFORMANCE ANALYSIS
Next, the effect of changing the spacing h between the slotted patch and the meta-lens is investigated. Fig. 13(a) exhibits VOLUME 11, 2023 the change of the antenna gain and reflection coefficient (S 11 ) versus h. It is evident that h values slightly below 6 mm provide the best gain-S 11 combination; hence h = 5.9 mm was selected for antenna simulation and measurement. Apparently, within the close vicinity of h = 5.9 mm the transmission through the lens is excellent as the gain is high and the reflection is reasonably low. The strong effect of changing h on the impedance matching over the entire band is illustrated in Fig. 13(b). Fig. 14(a) shows the fabricated prototype and Fig. 14(b) shows the slotted patch mounted for measurement in a farfield anechoic chamber. The measurement was done using a standard far-field chamber in which the received power at the Antenna under Test (AUT) emanated from a Ka-band horn antenna is used to predict the principle pattern cuts. The AUT gain was measured using the gain substitution method by comparing the received power by the AUT to the received power by a reference antenna of known gain. Fig. 15 shows the antenna's measured versus calculated S 11 values over the entire band. Both curves in Fig. 15 compare  well and follow the same changing trend, except the shift toward lower frequencies resulted from measurement. Upon a microscopic check of the etching tolerance of the split rings, there was a systematic error that might have created such a frequency shift. The -10dB fractional B.W. predicted from the measured results equals 17%. This is quite reasonable as compared to the ones obtained in works for mm-wave applications.  To illustrate the excellent focusing characteristics of the meta-lens, the total near-fields were predicted on different planes close to the lens surface. In Fig. 16(a), the near-fields predicted on planes parallel to the lens surface, one below and the other above the lens, show explicitly how the wildly scattered electric fields emanated from the slotted patch were focused efficiently to converge over a much smaller zone on the transmission side of the lens, which looks like a focal region. The formation of the said focal region, and hence the strong focusing effect of the meta-lens, becomes very obvious from the fields depicted in Fig. 16(b) where the field is predicted on an orthogonal plane, which cuts through the plane in Fig.16(a), in the transmission side of the lens. These observations strongly support the argument that the negative refractive index leads to negative refractive angles as discussed in Section II-B and reported in [25]. A focal point was indeed obtained at mm-wave frequencies, which confirms the strong focusing effect caused by the negative refractive index behaviour of the meta-lens.
The 3D far-filed radiation patterns shown in Fig. 17 tie well with the near-fields in Fig. 16. It is quite clear how the use of the meta-lens narrowed the broad far-field pattern of the slotted patch alone to a much more directive beam. The gain behaviour versus frequency is shown in Fig. 18. The measured gain values compare well to the simulated ones for both the patch alone and the entire antenna comprised of the meta-lens on top of the patch. The maximum measured gain is 12.7 dBi, while the aperture efficiency at this gain is 92%. These excellent metrics slightly change at off-center frequencies, as the measured 1dB gain bandwidth is 18%. Finally, Fig. 19 exhibits the measured versus simulated normalized far-field principle planes pattern cuts at frequencies 26 GHz, 27 GHz, and 28 GHz, respectively. The measured patterns have good agreement with simulated ones with some discrepancies that can be attributed to some slight misalignment of the antenna during measurement and to the gain reduction at some measured frequencies as compared to the simulated ones. Overall, it is evident that the antenna performance is stable versus frequency. VOLUME 11, 2023

C. THE META-LENS PERFORMANCE VERSUS A CONVENTIONAL LENS
The performance of the meta-lens is then compared to an equivalent conventional lens of the same footprint. A spherical conventional lens was designed [37] using equation (4) [37]. The dielectric material used is polytetrafluoroethylene (PTFE) with ε 2 = 2.2, which is common for lens design.
In (4), ε 1 and ε 2 are the relative dielectric permittivity of the free space and the lens material, respectively, and h = 5.9mm. The slotted patch antenna, as shown in Fig. 20, is used to excite the conventional lens and the structure is fully simulated using HFSS [22]. The simulated results of the gain and S 11 are compared in Fig. 21 for the conventional lens versus the proposed SRR meta-lens, respectively. It is clear that the meta-lens and the conventional lens generate comparable gain values over the entire band with a slight gain improvement achieved by the meta-lens at some frequencies. Such higher gain values obtained by a flat lens as compared to an equivalent conventional lens are reported in [38]. Regarding the impedance matching, despite the conventional lens offering better matching from 27GHz to 30GHz (lower S 11 values), the 10-dB fractional bandwidth of the SRR meta-lens is wider. These results confirm the good performance of the proposed metalens as it offers comparable performance to a conventional lens while it has a much smaller profile.
To further confirm the proposed meta-len's excellent focusing ability, a different source antenna was used to excite the meta-lens. The element was proposed in [9] to operate at mm-wave frequencies. Fig. 22 shows the antenna element with the proposed meta-lens located at h = 5.9mm on top of it. Fig. 23 exhibits the calculated gain values over the entire frequency band for the antennas alone and the antennas topped with the proposed meta-lens. Fig. 23 shows clearly that the SRR meta-lens functions very well in improving the gain for both types of antennas used to excite it. This is evident from the 4-5 dBi gain increase achieved over the entire band. This indeed further confirms the excellent focusing performance of the proposed SRR meta-lens irrespective of the considered antenna type used to excite it.

D. PROPOSED META-LENS POSSIBLE APPLICATION
As pointed out in the introduction, despite the millimeter wave (mm-wave) bands in the range of 25-40 GHz being licensed for a decade, the existing technology (4G/WiFi/WLAN) is still using microwave sub-6GHz bands due to the challenges associated with mm-wave network implementation and the strong signal fading at mm-wave frequencies. Despite this, we strongly believe that mm-wave wireless communication systems will be implemented worldwide within the coming years due to the aggressive demand for higher data rates.
Regarding the possible applications of the proposed metalens antenna, the focus of the scientific literature thus far has been on designing mm-wave antennas for base station platforms [39] and mobile phones [40]. However, authors strongly believe that mm-wave implementations will not be restricted to cell phone networks, and will be extended to replace the current indoor WiFi 4G and Sub-6GHz systems. This is natural as there is no point in implementing a very expensive mm-wave 5G network without upgrading all terminal devices despite both 4G and 5G systems can stay functional simultaneously. This will certainly require access points and routers with antennas of much higher gain values than the existing ones used in the 4G and Sub 6G spectrums, which are typically in the 4-6 dBi range [41], to provide an acceptable power level at the terminals of the receiving device (like cell phones, tablets, and laptops). Therefore, the proposed antenna can be used for routers and access points at mm-wave frequencies as the achieved gain range (11-12.7 dBi) is quite reasonable to overcome fast fading in indoor environments like big halls and meeting rooms. To ensure full coverage and polarization diversity multiple of the proposed antennas can fit easily on the side, top, or bottom facets of an access point with different orientations [41]. An illustration of such an implementation can be grasped from Fig. 24 shown below. Radiation patterns can be titled mechanically as illustrated by shaping the access point chassis or by defining phase gradients on the meta-lens surface similar to the principle explained in [42].

IV. STATE OF-THE-ART COMPARISON
In order to highlight the merits of the proposed design compared to previous related works, Table 2 compares the performances of some recent millimeter-wave antennas that utilize a lens or a metasurface to improve the gain and/or the bandwidth of another antenna element. It is evident from Table 2 that the aperture efficiency of the proposed design (92%) is the highest among all works. This indeed came as a result of the excellent focusing characteristic generated by the selected meta-lens. To our knowledge, this is the highest aperture efficiency value available in the open literature for similar antennas. It is worth mentioning here that the aperture efficiency was calculated, for all antennas in Table 2, using the expression given for aperture efficiency in [12] and based on the antenna area and maximum measured gain values reported in the corresponding references.
One can also realize that the proposed antenna provides the best impedance-bandwidth/gain-bandwidth combination (17%/18%) despite other designs like the ones in [4] and [9] do provide higher 10dB B.W. alone. This good performance at the port level as well as in the farfield confirms, as reinforced by the far-field pattern stability illustrated in Section III, the superior performance of the proposed design. It is also obvious from the comparison table that the proposed design belongs to the group of antennas that have a quite small footprint [4], [9]. Moreover, the low loss nature of the negative refractive index material yielded a quite high radiation efficiency of a maximum value of 95%. This is the highest value among all recent designs reviewed in this paper. Finally, unlike many of the complex designs available in the  open literature, the proposed design utilizes a simple singlelayer meta-lens to improve the gain and bandwidth mounted on top of another simple slotted patch antenna. As listed in Table 2, most designs include some complexity either in the FIGURE 23. Gain improvement obtained using the proposed meta-lens for the slotted patch against the antenna proposed in [9]. lens geometry or in the antenna element itself, except the proposed method and the design in [9].

V. CONCLUSION
A simple single-layer mm-wave meta-lens is proposed for 5G antenna performance enhancement. The proposed meta-lens is composed of 8 × 8 split ring elements that all occupy a minimal footprint of 1.25λ×1.3λ. Hence, it can be considered one of the most compact lenses available in the open literature. Despite this small footprint, the proposed meta-lens were shown, via both simulation and measurement, to be very effective in improving the gain of a slotted patch antenna by 4-5 dBi over a wide frequency band from 25 GHz to 31 GHz. This was justified by the negative refractive index behavior that significantly improved the lens focusing characteristics. The achieved aperture efficiency of the proposed antenna, which is 92%, is the maximum reported in the literature for similar mm-wave antennas. This is a very desirable feature as it yields a high-gain antenna with quite a small footprint. The stability of the proposed antenna performance against frequency was established by producing a wide impedance bandwidth of 17% associated with a wide 1dB gain bandwidth of 18%. Such a combination is indeed novel, especially because a high radiation efficiency accompanied it at many frequencies with a maximum value of 95%. Due to the simplicity, compactness, and excellent performance metrics, the proposed meta-lens represents a promising candidate for lowprofile mm-wave 5G applications.

VI. ACKNOWLEDGMENT
The authors appreciate and acknowledge the support provided by the University of Sharjah, United Arab Emirates, by providing all their resources to conduct this research.