Jaumann-Like Rasorber Offering Angular Stability With Wideband Absorption and Low Insertion Loss

In this letter, a Jaumann-like rasorber is examined, which offers wideband absorption, angular stability, and polarization insensitivity. The structure is defined by three layers with two resistive layers placed on top of a lossless layer. Also, all three layers use capacitors to ensure compactness for the metasurface rasorber. These features provide broad angular stability up to 55<inline-formula><tex-math notation="LaTeX">$^{\circ }$</tex-math></inline-formula> while also maintaining low insertion losses of about 1 dB and only 0.6 dB for measured broadside incidence. The rasorber is also shown to work as a radome in the near field of a planar antenna array.


Jaumann-Like Rasorber Offering Angular Stability With Wideband Absorption and Low Insertion Loss
Callum J. Hodgkinson, Dimitris E. Anagnostou , Senior Member, IEEE, and Symon K. Podilchak , Member, IEEE Abstract-In this letter, a Jaumann-like rasorber is examined, which offers wideband absorption, angular stability, and polarization insensitivity. The structure is defined by three layers with two resistive layers placed on top of a lossless layer. Also, all three layers use capacitors to ensure compactness for the metasurface rasorber. These features provide broad angular stability up to 55 • while also maintaining low insertion losses of about 1 dB and only 0.6 dB for measured broadside incidence. The rasorber is also shown to work as a radome in the near field of a planar antenna array.

I. INTRODUCTION
T HERE has recently been significant research devoted to antenna radomes realized using metasurfaces. Some applications include grating lobe mitigation [1], gain enhancement [2], and radar scattering reduction. One such low scattering structure is a rasorber (or radome absorber). This is a frequency-selective absorbing metasurface with a passband to offer practical antenna integration for radome scenarios.
The first rasorber design was reported in 2006 [3] but the name first appeared in 2014 [4]. Such structures are generally based on a classic Salisbury screen concept, in which a ground layer and a resistive sheet are spaced a quarter-wavelength apart. To create a passband, however, the layers (or sheets) of the rasorber are usually defined using frequency selective surfaces (FSSs). This signal transmission passband (for a back-end antenna system) can be between two absorption bands [4], below/above an absorption band [6], or be defined by multiple transmission bands [7]. The benefit of a rasorber over more conventional FSS radomes is the out-of-band absorbing functionality that the structure provides to reduce the back-end antenna scattering effects. Therefore, to enhance the radar scattering reduction over frequency, as examined in this letter, the absorption bands should be as wide as possible.  The proposed three-layered rasorber structure, which maximizes its absorption bandwidths (BWs) (see Figs. 1 and 2 inset), is achieved by stacking two resistive layers above a bottom lossless layer. This configuration means the proposed rasorber is more akin to that of a Jaumann absorber, where multiple resistive layers are placed above a metallic ground. A similar configuration to this classic architecture, yet also different, was chosen due to the absorption and angular stability benefits associated with Jaumann absorbers (when compared with a Salisbury screen). Another important consideration, when designing rasorbers, is to ensure that any design has minimal impact on the back-end antenna system or transceiver array. Therefore, any successful rasorber design must be as transparent as possible, i.e., have very small insertion losses, within its passband, to minimize any signal degradation or antenna interference.
Our proposed rasorber offers insertion loss values of about 0.6 dB (or less) for angles of incidence up to and including 40 • , when measured within a calibrated anechoic chamber. In addition, due to the symmetry and layering of the developed 1536-1225 © 2022 Crown Copyright unit cell, the design is polarization insensitive and has wideband absorption. The downside of such passive designs is that they can only be used in conjunction with narrowband antennas. Therefore, for the use of wideband antennas, frequency reconfigurable structures, such as varactor tunable FSSs [8] and rasorbers [9], [10] were designed, but these generally struggle to obtain the desirable low loss and polarization insensitive features due to the required bias lines. See, for example, [10], where insertion loss values were reported to be 2.1 dB.
This letter reports the design and test of the proposed Jaumann-like rasorber that offers wideband absorption, low insertion losses, and polarization insensitivity to incident TM and TE fields. Moreover, and as further outlined in Table I,  our proposed rasorber provides a −10 dB reflection coefficient relative BW of more than 130% as well as lower and upper 80% absorption BWs of about 104% and 33%, respectively, while also offering a narrow half-power bandwidth (HPBW) of 8% in the passband. As will also be shown in this letter, the fabricated rasorber was placed in the near-field (NF) of a six-element ultrawideband (UWB) array, and measurements show negligible degradation for the far-field (FF) beam pattern as well as the realized gain at the rasorber passband. To the best of our knowledge, no similar passive rasorber has been reported in the open literature that offers similar broadband absorption, polarization insensitivity, low loss, and wide-angle performance metrics.

II. DESIGN CONSIDERATIONS AND SIMULATIONS
The design of the proposed Jaumann-like rasorber consists of a lossless layer topped by two resistive layers, where each layer is separated by a gap of 12.2 mm filled with a low-loss and lowdielectric foam (Eccostock PP-4, r = 1.06). The resistive layers are defined by a Jerusalem cross surrounded by a square-ring structure and tapered dipole arms in the horizontal and vertical planes (see Fig. 1). This symmetry supports the noted polarization insensitivity. The unit cell of the bottom metamaterial layer (also referred to as the lossless layer) is composed of a small square-ring slot bridged with the same capacitors as the upper resistive layers. A larger gap of 0.95 mm is necessary between the inner and outer conductors of the bottom layer to ensure that the layers have the same resonant frequency. All of the printed circuit board (PCB) layers use a substrate of Rogers RO5008 ( r = 2.2) with a thickness of 0.508 mm.
For the unit cell design of the resistive layers, it was decided to use a printed Jerusalem cross. This is because such centre-cross elements generally provide greater angular stability when compared with other solid-interior implementations [5], as well as Jerulsaem crosses, in particular, having a low resonant frequency enabling compactness. This cross was then placed inside a square ring and lumped element capacitors were positioned between these shapes (see Fig. 1) to create a compact, subwavelength unit cell. Basically, with the use of the capacitors, our rasorber offers a small unit cell size and, therefore, a small periodicity of 15 (0.273 λ at 5.46 GHz) and 7.5 mm for the lossless metamaterial layer (0.136 λ). This small periodicity results in a high grating frequency [5] and should consequently create an angularly stable structure. Another method to improve angular stability is to include multiple resistive layers, but this can, however, also result in greater insertion loss [5]. It was, therefore, important to try and minimize these losses, and given the motivations to attain low losses, our rasorber employs tapered dipole arms (which offer more wideband matching). Details are outlined in Fig. 1 and Table II. It should also be mentioned that Hodgkinson et al. [9] reported initial findings of a reconfigurable rasorber. The newly optimized unit cell reported in this letter, however, offers wider absorption BWs, lower insertion losses, and greater angular stability when compared with [9]. Also, the rasorber proposed in this letter has been manufactured and experimentally verified, demonstrating these performance metrics, while also its suitability for antenna array integration.
The rasorber was initially designed and simulated in CST Microwave Studio as a unit cell with infinite periodic boundary conditions. Fig. 2 shows the simulated transmission (S 21 ), and results indicate that the design supports transmission at 5.46 GHz with a small insertion loss of about 1 dB and for angular variations up to 55 • . It can also be observed that the transmission coefficient offers a steep roll off away from this passband, with the transmission decreasing to less than −10 dB in a very short range, both below and above the 5.46 GHz passband frequency. These results also imply that out-of-band absorption is achieved since the reflection coefficient in Fig. 3 is still reasonably matched both inside and outside the passband. Also, the reflections remains below −10 dB over a UWB frequency range and for various incident angles.
The equation A = 1 − T − R was also used to calculate the absorption A of the structure, where T is the transmission coefficient (referring to |S 21 | 2 ) and R is the reflection coefficient   (|S 11 | 2 ) [11], [12], [13]. Following this definition and by inspection of Fig. 4, it can be observed that the out-of-band absorption is very high and is greater than 80% between 2.0-4.4 GHz and 7.0-9.5 GHz for all angles.
Although the unit cell design is square and is clearly symmetric in both the xand y-directions, simulation results were still needed to verify the polarization insensitivity. The design was, therefore, examined for both TE-and TM-oriented incident waves defining the vertical and horizontal linearly polarized cases, respectively. These results are shown for off-broadside incidence in Fig. 5, and it can be observed that the transmission and reflection remains stable when the polarization is varied, with no noticeable difference between the curves. Table III shows only a slight increase in insertion losses with increasing incident angle and no change in the resonant frequency. From these findings, the rasorber can be thought of as angularly stable and polarization insensitive.

III. MEASUREMENTS AND DISCUSSIONS
For proof-of-concept of the proposed Jaumann-like rasorber, a finite 10×10 unit-cell structure measuring 150×150 mm was manufactured and tested (see Fig. 6). MuRata 0.2 pF capacitors (GRM1555C1HR20WA01) and Panasonic 300 Ω resistors (ERA2AEB301X) were selected. Fig. 7 shows the measurement setup where an absorbing wall was set in the FF between two reference horns (at a distance of 2.2 m from each antenna), with a hole in the wall the size of the rasorber. To obtain the insertion loss, the transmission (S 21 ) was measured with and without the rasorber. Then, the absorbing wall was rotated to study angular variation. The results in Fig. 8 and Table III suggest that simulations and measurements are in agreement. Also, the  low structural losses reported and similarity to the simulations suggest that the manufactured design is working as intended, with minimal insertion loss while offering good angular stability.
The higher than expected transmission, for frequencies above the passband in Fig. 9, could be related to some manufacturing tolerances or minor bending of the layers due to the thin substrate employed (which were selected to help support low insertion losses). This bending concern could be mitigated by performing similar experiments but with a larger version of the rasorber. Since, for a larger design, these small imperfections should have less of an effect on performance. Also, the finite size of the structure might be related to the variations that can be observed at frequencies below the passband when the size becomes less than 2.5 λ. Despite these minor deviations, there is a general agreement in the results, that is, the simulated structure (defined by an infinite number of unit cells) offers similar functionality when compared with the finite rasorber. Given these findings, we felt that further practical investigations and prototyping, whilst considering a larger PCB structure, was not required for proofof-concept realization.
To measure the reflection coefficient (S 11 ) of the rasorber, the reference antennas were both placed on the same side of the absorbing wall, being almost colocated. Transmission, S 21 , was then measured again for the following cases: 1) no rasorber, i.e., an open; 2) with microwave absorber; 3) with a metal plate; i.e., a short. From these results, the system could be calibrated for open and short representative reference standards, and from this, the calibrated reflection coefficient was then determined. These reflection and transmission measurements were then used to calculate the absorption, as shown in Fig. 10. The observed minor differences above and below the passband could be related to the practicalities mentioned earlier, and could also be linked Fig. 11. Normalized patterns for a 6×1 UWB planar array (see [14] for more details on the array) measured at 5.46 GHz with (red) and without (blue) the presence of the rasorber, which was placed in the NF of the array.
to spurious diffraction or possible edge scattering effects from the finite structure (and this would have also cascaded to the calibrations). However, generally consistent results are observed for the simulations and the measurements. Therefore, it can be surmised that the rasorber offers wideband absorption, low loss, polarization insensitivity, and 55 • angular stability as per design.
To compare the results of the rasorber with other similar designs, Table I was compiled. It can be observed that our Jaumann-like design has a far wider matching reflection BW when compared with other similar structures, and this also results in much wider absorption bands. The design achieves these results without sacrificing polarization or angular stability (here, angular stability or the angular stable range is defined where the rasorber still works as intended and with no change in resonance, and still with low reflections below −8 dB). The additional lossy layer does, however, cause a narrower HPBW in the transmission band. This is not considered to be of great detriment to the overall system, because only the peak of the transmission frequency band and its angular stability is generally of interest for covered arrays and other rasorber scenarios.
Since the main purpose of the rasorber is to act as a practical radome, this had to be tested as well. Therefore, the beam pattern of a compact 6×1 UWB antipodal Vivaldi array (see [14] for more details) was measured for uniform excitation in the rasorber passband at 5.46 GHz. In particular, the rasorber was placed in the radiative NF of the array (at a distance of 0.25 m), and the FF beam pattern was measured using a standard reference horn. This was then repeated without the rasorber. Patterns are shown in Fig. 11, and they demonstrate that the presence of the rasorber has little to no effect on the array, apart from a small drop in gain due to the minor transmission losses of the rasorber. For example, the measured realized gain without and with the rasorber was about 10 and 9.4 dBi, respectively. These findings highlight the important functionality of the design, that is, the proposed rasorber can act as an effective radome in the NF of an array without hampering its FF performance.

IV. CONCLUSION
This letter reports the design and measurement of a threelayered Jaumann-like rasorber. The results show that the additional resistive layer and miniaturized unit cell help provide 55 • of angular stability as well as wideband absorption and a matching BW of more than 135%. The structure also offers dual-polarization insensitivity and low insertion loss. Tests have also been completed to show that the rasorber can work in conjunction with an array to function as a practical radome.