**2.2.2 Slot coupled 4x4 superstrate antenna array**

106 Wireless Communications and Networks – Recent Advances

Fig. 15. Measured and simulated E-plane radiation patterns of 2 x 2 superstrate antenna array.

The measured and simulated H-plane radiation patterns are shown in Fig. 16 for the optimised superstrate dimension. The radiation patterns are also plotted by taking into

Fig. 16. Measured and simulated H-plane radiation patterns of 2 x 2 superstrate antenna array.

58 GHz - measured 59 GHz - simulated

58 GHz - measured 59 GHz - simulated

59 GHz - measured 60 GHz - simulated .

59 GHz - measured 60 GHz - simulated .

Fig. 17 shows the photograph of a 4 x 4 array antenna array with superstrate. The antenna parameters and the distance between the elements are the same as explained for 2 x 2 superstrate antenna array in Section 2.2. The same substrate for the superstrate is used. Here also the superstrate should be optimised and it is found to be 6 λ0 x 6 λ0 which is the total size of the antenna. For manufacturing this prototype, because of the 4x4 array, two metal wedges of 3 mm width are used to position the superstrate at 2.3 mm (~ λ0/2 - 0.127 mm) above the patch array. This mechanical solution is found to be better in this case than foam due to the relation between superstrate position sensitivity and gain increase.

Fig. 17. Photograph of 4 x 4 superstrate antenna array prototype (a) Top view of superstrate antenna (b) view of antenna feed line network from bottom side. Ground plane taken is 6 λ<sup>0</sup> x 10 λ0, for connecting V band connector and for ease of measurement purpose.

Fig. 18 shows the measured and simulated S11, and measured gain with simulated directivity. It is found that the maximum gain measured is 19.7 dBi with an efficiency of 51% (simulated directivity = 22.6 dBi) which is far better than a classical 6 x 6 array antenna of gain 17.5 dBi with an efficiency of 40% at 59 GHz (Lafond et al. 2001), and an 8 x 8 array antenna (size = 6.5 λ0 x 6.5 λ0) of gain 19.7 dBi with an efficiency of ~ 40% as explained in (Nesic et al., 2001). The S11 bandwidth measured for the 4 x 4 superstrate antenna array is 57.9 GHz to 61.3 GHz (5.7%).

Fig. 18. Variation of S11 and gain with a square superstrate dimension of 6 λ0. Simulation measured .

It is clear from Fig. 18 that the gain measured and simulated are maximum from 58 GHz to 60 GHz. The simulated and measured H-plane radiation patterns are shown in Fig. 19 (a). It is noted that the simulated patterns are in good agreement with the measured results. The measured HPBW is 8° at 60 GHz.

The measured and simulated E-plane radiation patterns are shown in Fig. 19(b) for the optimised superstrate dimension. The radiation patterns are broad and the agreement between measurement and simulation are quite acceptable. The measured HPBW is found to be 10° at 60 GHz. In this case, the cross polarization level is lower than -25 dB on both the E and H-plane, and is lower than -16 dB at 45° cut plane in 3D pattern.

For both the presented antennas, a distance between the elements of 1.3 λ0 is used for the source array, which induces high ambiguity side lobes for both cases when there is no superstrate: -2 dB for a 2 x 2 array and -1.9 dB for 4 x 4 array as shown in Fig. 20. Adding a superstrate will strengthen the main lobe while suppressing the ambiguity side lobes to less than -10 dB for both the arrays without affecting the back radiation as shown in Fig. 20. It also strengthens the front to back ratio as shown in the figure. It is to be underlined that the size of the superstrate is a key point of the design of such structures. In fact the nature of the pattern is conditioned by the choice of this parameter: a broad pattern is obtained for a size limited to 3.2 λ0 x 3.2 λ0 for 2 x 2 array and 6 λ0 x 6 λ0 for 4 x 4 arrays.

Fig. 18. Variation of S11 and gain with a square superstrate dimension of 6 λ0.

E and H-plane, and is lower than -16 dB at 45° cut plane in 3D pattern.

limited to 3.2 λ0 x 3.2 λ0 for 2 x 2 array and 6 λ0 x 6 λ0 for 4 x 4 arrays.

It is clear from Fig. 18 that the gain measured and simulated are maximum from 58 GHz to 60 GHz. The simulated and measured H-plane radiation patterns are shown in Fig. 19 (a). It is noted that the simulated patterns are in good agreement with the measured results. The

The measured and simulated E-plane radiation patterns are shown in Fig. 19(b) for the optimised superstrate dimension. The radiation patterns are broad and the agreement between measurement and simulation are quite acceptable. The measured HPBW is found to be 10° at 60 GHz. In this case, the cross polarization level is lower than -25 dB on both the

For both the presented antennas, a distance between the elements of 1.3 λ0 is used for the source array, which induces high ambiguity side lobes for both cases when there is no superstrate: -2 dB for a 2 x 2 array and -1.9 dB for 4 x 4 array as shown in Fig. 20. Adding a superstrate will strengthen the main lobe while suppressing the ambiguity side lobes to less than -10 dB for both the arrays without affecting the back radiation as shown in Fig. 20. It also strengthens the front to back ratio as shown in the figure. It is to be underlined that the size of the superstrate is a key point of the design of such structures. In fact the nature of the pattern is conditioned by the choice of this parameter: a broad pattern is obtained for a size

Simulation measured .

measured HPBW is 8° at 60 GHz.

Fig. 19. Measured and simulated (a) H plane & (b) E plane radiation patterns of 4 x 4 superstrate antenna array.

Fig. 20. Comparison of simulated results of 2 x 2 and 4 x 4 arrays without and with superstrate in terms of main beam, side lobe and back radiation.

#### **2.3 Superstrate aperture antenna**

In this section we are using another source, as aperture, for exciting the antenna. The side view of an aperture antenna with superstrate is shown in Fig. 21(a). The aperture is optimised to 4.4 mm x 1 mm for maximum coupling with a stub length of 0.4 mm (Fig. 21(b)). To improve the rigidity of antenna, a ground plane of thickness t = 0.2 mm is used. For maintaining the exact air thickness in practical prototype, the superstrate is inserted within an air pocket realized in a Rohacell foam block of permittivity 1.05, as shown in Fig. 21(c). All the substrate and superstrate material used are the same as explained in section 2.2.

Fig. 20. Comparison of simulated results of 2 x 2 and 4 x 4 arrays without and with

In this section we are using another source, as aperture, for exciting the antenna. The side view of an aperture antenna with superstrate is shown in Fig. 21(a). The aperture is optimised to 4.4 mm x 1 mm for maximum coupling with a stub length of 0.4 mm (Fig. 21(b)). To improve the rigidity of antenna, a ground plane of thickness t = 0.2 mm is used. For maintaining the exact air thickness in practical prototype, the superstrate is inserted within an air pocket realized in a Rohacell foam block of permittivity 1.05, as shown in Fig. 21(c). All the substrate and superstrate material used are the same as explained in

superstrate in terms of main beam, side lobe and back radiation.

**2.3 Superstrate aperture antenna** 

section 2.2.

Fig. 21. (a) Cutting plane of aperture antenna with superstrate, ground plane size = 6 λ0 x 6 λ0. (b) Aperture and stub in details. (c) Overview of the Prototype: Details of the superstrate and air gap within foam.

In this case also we want to study the effect of superstrate size on antenna performance. To study the effect of superstrate size ' S ' and hence to optimize it, a parametric study has been performed, using commercial electromagnetic software CST Microwave studio. To highlight the effects of this parameter, results obtained for four sizes (1 λ0 x 2 λ0 , 2 λ0 x 2 λ0, 1.2 λ0 x 2.7 λ0 and 3 λ0 x 3 λ0 ) are reported in Fig. 22. It is observed that both the S11 and the directivity vary according to the size of the superstrate*.* A maximum directivity of 14.5 dBi is obtained for a superstrate size of 1.2 λ0 x 2.7 λ0. The corresponding 2:1 VSWR bandwidth is noted to be equal to 57.5 - 71 GHz i.e. 22.5%. It is also noticed that the directivity decreases when the size of the superstrate is above or below this optimized value**.** We plotted directivity only up to 65 GHz because of the decline in the values after that. When the superstrate size is higher than the optimized value, then there is a plunge in directivity as shown in Fig. 22. Hence the broad nature of the pattern moved out at 60 GHz as explained in (Vettikalladi et al., 2009a), i.e the radiation patterns change from broad side to sectorial / null at 60 GHz, which is also useful for some other applications.

Fig. 22. Simulated results of S11 and directivity with various superstrate dimensions. 1 λ0 x 2 λ0 ; 2 λ0 x 2 λ0 ; 1.2 λ0 x 2.7 λ0 ; 3 λ0 x 3 λ0 .

It concludes that in this case the dimension of the superstrate is critical for the optimum performance of the antenna. Also we can control the shape of the pattern by changing the dimension of the superstrate from broadside to sectorial / null. The comparison of measured and simulated S11, and measured gain with simulated directivity for the optimized superstrate size is shown in Fig. 23 (a). Table IV gives the summary of these results (Vettikalladi et al., 2010b). It is noted that the measured 2:1 VSWR bandwidth is 15% which is larger compared to the superstrate slot coupled antenna (Vettikalladi et al., 2009a), where the bandwidth was only 6.8%. The gain is measured using comparison technique with a standard gain horn. It is found to be 13.1 dBi. Moreover, it is almost flat (ripple ~ 0.5 dB) over a bandwidth of 5 GHz. To determine the efficiency, we compared the measured gain with the simulated directivity. The estimated efficiency is 79%. In order to highlight the effect of the superstrate for this configuration, the simulated comparison of E-plane radiation pattern of aperture antenna with superstrate and without supertstrate is shown in Fig. 23 (b). The ripples in the pattern without superstrate are due the diffraction from the edges of the limited ground plane. Also it is clear that aperture antenna is a bidirectional antenna, superstrate technology make this antenna to unidirectional without adding any reflector, which is a highlight of the superstrate with this kind of source. I.e. Superstrate makes the antenna pattern directive and there is a gain enhancement of 8 dB compared to its basic aperture antenna.

in (Vettikalladi et al., 2009a), i.e the radiation patterns change from broad side to sectorial /

Fig. 22. Simulated results of S11 and directivity with various superstrate dimensions. 1 λ0 x 2 λ0 ; 2 λ0 x 2 λ0 ; 1.2 λ0 x 2.7 λ0 ; 3 λ0 x 3 λ0 .

gain enhancement of 8 dB compared to its basic aperture antenna.

It concludes that in this case the dimension of the superstrate is critical for the optimum performance of the antenna. Also we can control the shape of the pattern by changing the dimension of the superstrate from broadside to sectorial / null. The comparison of measured and simulated S11, and measured gain with simulated directivity for the optimized superstrate size is shown in Fig. 23 (a). Table IV gives the summary of these results (Vettikalladi et al., 2010b). It is noted that the measured 2:1 VSWR bandwidth is 15% which is larger compared to the superstrate slot coupled antenna (Vettikalladi et al., 2009a), where the bandwidth was only 6.8%. The gain is measured using comparison technique with a standard gain horn. It is found to be 13.1 dBi. Moreover, it is almost flat (ripple ~ 0.5 dB) over a bandwidth of 5 GHz. To determine the efficiency, we compared the measured gain with the simulated directivity. The estimated efficiency is 79%. In order to highlight the effect of the superstrate for this configuration, the simulated comparison of E-plane radiation pattern of aperture antenna with superstrate and without supertstrate is shown in Fig. 23 (b). The ripples in the pattern without superstrate are due the diffraction from the edges of the limited ground plane. Also it is clear that aperture antenna is a bidirectional antenna, superstrate technology make this antenna to unidirectional without adding any reflector, which is a highlight of the superstrate with this kind of source. I.e. Superstrate makes the antenna pattern directive and there is a

null at 60 GHz, which is also useful for some other applications.

Fig. 23. (a) Results of S11 and gain with a superstrate dimension of 1.2 λ0 x 2.7 λ0 . Simulation ; measurement . (b) Simulated comparison of E-plane antenna radiation pattern with and without superstrate.

The measured and simulated H- and E-plane radiation patterns at 57 GHz, 60 GHz and 62 GHz are shown in Fig. 24 for the optimized superstrate dimension. It is noted that the radiation patterns are found to be broad and in agreement with simulations. The cross polar level is less than -25 dB for H-plane and -20 dB for E-plane respectively, for all the frequencies in the band. The measured half-power beam widths (HPBW) at 60 GHz are 26° for H-plane and 30° for E-plane respectively. The measured cross polarization level is lower than -17 dB at 45° cut plane in 3D patterns of both planes.


Table IV. Comparison between simulated and measured results superstrate aperture antenna.

Fig. 24. Measured and simulated H-plane & E- plane radiation patterns of superstrate antenna (Co and Cross polarisation). 57 GHz -Simulated ; 57 GHz - Measured ; 60 GHz -Simulated ; 60 GHz -Measured ; 62 GHz - Simulated ; 62 GHz - Measured .

#### **2.3.1 2x2 superstrate aperture antenna array**

The side and 3D view of a 2 x 2 aperture antenna array with superstrate are shown in Figs. 25(a) and (c). All the parameters of the antenna are the same as explained in section 2.3. For maintaining the exact air thickness, the superstrate is inserted within an air pocket realized in Rohacell foam as shown in Fig. 25(c). The distance between the elements in the array is optimized as d = 1.3 λ0 for obtaining maximum gain and to minimize coupling. The 2 x 2 feeding network is exposed in Fig. 25(b). In a classical array (without superstrate), when the

57.5-71 GHz (22.5%) 55 - 64GHz (15%) 14.1 dBi 13.1 dBi 79%

Fig. 24. Measured and simulated H-plane & E- plane radiation patterns of superstrate

E-plane

H-plane

The side and 3D view of a 2 x 2 aperture antenna array with superstrate are shown in Figs. 25(a) and (c). All the parameters of the antenna are the same as explained in section 2.3. For maintaining the exact air thickness, the superstrate is inserted within an air pocket realized in Rohacell foam as shown in Fig. 25(c). The distance between the elements in the array is optimized as d = 1.3 λ0 for obtaining maximum gain and to minimize coupling. The 2 x 2 feeding network is exposed in Fig. 25(b). In a classical array (without superstrate), when the

57 GHz -Simulated ; 57 GHz - Measured ; 60 GHz -Simulated ; 60 GHz -Measured ; 62 GHz - Simulated ; 62 GHz - Measured .

antenna (Co and Cross polarisation).

**2.3.1 2x2 superstrate aperture antenna array** 

Table IV. Comparison between simulated and measured results superstrate aperture antenna.

Maximum Directivity of the prototype (simulated)

Maximum Gain (measured) Efficiency Estimated η

Return loss bandwidth (measured)

Return loss bandwidth (simulated)

distance between the patches is d = 1.3 0, high ambiguity side lobes appear with almost the same level as the main lobe. Adding a superstrate strengthens the main lobe while reducing the ambiguity side lobes to less than -10 dB (E-plane) as shown in Fig. 26. It also strengthens the front to back ratio as shown in the figure. It has to be underlined that the size of the superstrate is a key point for the design of such structures as explained in previous cases.

Fig. 25. (a) Cutting plane of a 2 x 2 aperture antenna array with superstrate, ground plane size = 6 λ0 x 10 λ0 for connecting V band connector and for ease of measurement purpose. (b) 2 x 2 feed network. (c) Overview of 2 x 2 array prototype: details of the two separate superstrate sheets and air gap within foam.

Fig. 26. Simulated comparison of 2x2 antenna array pattern (d = 1.3 0) with and without superstrate (E-plane).

As did in previous sections, we studied the effect of superstrate size ' S ' by simulating different sizes and the optimized solution is found to be two pieces of dimension 1.2 λ0 x 4 λ0, one sheet for two aperture antenna, with a spacing of 1mm as shown in Fig. 25(b). If we use a single piece with a size of 2.6 λ0 x 4 λ0 , then the gain is little lower than in the previous case. The comparison of measured and simulated S11, and measured gain with simulated directivity for the optimized superstrate size is shown in Fig. 27. The highest directivity of 17.9 dBi is obtained for the optimized superstrate size. The resulting simulated 2:1 VSWR bandwidth is 11.3% from 57.2 to 64 GHz**.** 

Fig. 27. Variation of S11 and gain for a 2 x 2 superstrate aperture antenna array. Simulation ; measurement .

Table V gives the comparison of measured and simulated results for the optimized superstrate size. It is found that the maximum measured gain is 16.6 dBi with S11 bandwidth of 13.3% (56 GHz - 64 GHz), and an estimated efficiency of 74%. This gain is comparable to a classical 4 x 4 array at 60 GHz but with better efficiency (Lafond 2000). Also the measured gain is almost stable (ripple ~ 0.8 dB) over 5 GHz (57 GHz - 62 GHz) in the band of interest (Vettikalladi et al., 2010c).


Table V. Comparison between simulated and measured results of a 2 x 2 superstrate aperture antenna array.

The measured and simulated H- and E-plane radiation patterns at 57 GHz, 60 GHz and 62 GHz are shown in Figs. 28 (a) & (b) respectively for the optimized superstrate dimension. It is noted that the radiation patterns are found to be broad and in good agreement with simulations. The measured cross polar levels are -26 dB for H-plane and -20 dB for Eplane respectively. The radiation patterns are verified to be the same in all the frequencies in the band of interest. The measured HPBWs are 17° for H-plane and 16° for E-plane respectively at 60 GHz.

As did in previous sections, we studied the effect of superstrate size ' S ' by simulating different sizes and the optimized solution is found to be two pieces of dimension 1.2 λ0 x 4 λ0, one sheet for two aperture antenna, with a spacing of 1mm as shown in Fig. 25(b). If we use a single piece with a size of 2.6 λ0 x 4 λ0 , then the gain is little lower than in the previous case. The comparison of measured and simulated S11, and measured gain with simulated directivity for the optimized superstrate size is shown in Fig. 27. The highest directivity of 17.9 dBi is obtained for the optimized superstrate size. The resulting simulated 2:1 VSWR

Fig. 27. Variation of S11 and gain for a 2 x 2 superstrate aperture antenna array.

Return loss bandwidth (measured)

57.2 - 64 GHz (11.3%) 56 - 64 GHz (13.3%) 17.9 dBi 16.6 dBi 74% Table V. Comparison between simulated and measured results of a 2 x 2 superstrate

The measured and simulated H- and E-plane radiation patterns at 57 GHz, 60 GHz and 62 GHz are shown in Figs. 28 (a) & (b) respectively for the optimized superstrate dimension. It is noted that the radiation patterns are found to be broad and in good agreement with simulations. The measured cross polar levels are -26 dB for H-plane and -20 dB for Eplane respectively. The radiation patterns are verified to be the same in all the frequencies in the band of interest. The measured HPBWs are 17° for H-plane and 16° for E-plane

Table V gives the comparison of measured and simulated results for the optimized superstrate size. It is found that the maximum measured gain is 16.6 dBi with S11 bandwidth of 13.3% (56 GHz - 64 GHz), and an estimated efficiency of 74%. This gain is comparable to a classical 4 x 4 array at 60 GHz but with better efficiency (Lafond 2000). Also the measured gain is almost stable (ripple ~ 0.8 dB) over 5 GHz (57 GHz - 62 GHz) in the

> Maximum Directivity (simulated)

Maximum Gain (measured) Efficiency Estimated η

bandwidth is 11.3% from 57.2 to 64 GHz**.** 

Simulation ; measurement .

band of interest (Vettikalladi et al., 2010c).

Return loss bandwidth (simulated)

aperture antenna array.

respectively at 60 GHz.

Fig. 28. Measured and simulated H-plane (a) & E-plane (b) radiation patterns of the 2 x 2 superstrate antenna array (Co and Cross polarisation).


#### **2.3.2 16 x 16 superstrate aperture antenna array**

Finally we developed a big array to obtain very high gain of nearly 30 dBi for 60 GHz outdoor communication, for example from one department to another department inside a university (< 1km). Fig. 29 (a) depicts the 3D side view of the 16 x 16 array prototype. The antenna parameters and the distance between the elements are all same as explained in Section 3.2. For maintaining the exact air thickness, the superstrate is inserted within an air pocket realized in Rohacell foam as shown in Fig. 29(a). The 16 x 16 feeding network is showing in Fig. 29(b).

As pointed out in previous section , we want to use the smallest superstrate for obtaining high stable gain and consistent radiation pattern in the frequency band. We studied the effect of superstrate size ' S ' by simulating different sizes and the optimized size is found to be 16 pieces of dimension 1.2 λ0 x 21.8 λ0, one sheet for 16 aperture antenna, with a spacing of 1mm as shown in Fig. 29(a). If we use a single piece with a size of 20.6 λ0 x 21.8 λ0, then the gain is little lower than in the previous case.

Fig. 29. (a) Side overview of 16 x 16 array prototype: details of the 16 separate superstrate sheets and air gap within foam, total size = 20.6 λ0 x 21.8 λ0. (b) 16 x 16 feed network.

The comparison of measured and simulated S11, and measured gain with simulated directivity for the optimised superstrate size is shown in Fig. 30. The highest simulated directivity of 33.3 dBi is obtained for the optimised superstrate size. The resulting simulated 2:1 VSWR bandwidth is 22 %. It is found that the maximum measured gain is 29.4 dBi ( at point 'P' in Fig. 29 (b)) with S11 bandwidth of 16.7 % (54 GHz - 64 GHz), and an estimated

pocket realized in Rohacell foam as shown in Fig. 29(a). The 16 x 16 feeding network is

As pointed out in previous section , we want to use the smallest superstrate for obtaining high stable gain and consistent radiation pattern in the frequency band. We studied the effect of superstrate size ' S ' by simulating different sizes and the optimized size is found to be 16 pieces of dimension 1.2 λ0 x 21.8 λ0, one sheet for 16 aperture antenna, with a spacing of 1mm as shown in Fig. 29(a). If we use a single piece with a size of 20.6 λ0 x 21.8 λ0, then the

(a)

Fig. 29. (a) Side overview of 16 x 16 array prototype: details of the 16 separate superstrate sheets and air gap within foam, total size = 20.6 λ0 x 21.8 λ0. (b) 16 x 16 feed network.

(b)

The comparison of measured and simulated S11, and measured gain with simulated directivity for the optimised superstrate size is shown in Fig. 30. The highest simulated directivity of 33.3 dBi is obtained for the optimised superstrate size. The resulting simulated 2:1 VSWR bandwidth is 22 %. It is found that the maximum measured gain is 29.4 dBi ( at point 'P' in Fig. 29 (b)) with S11 bandwidth of 16.7 % (54 GHz - 64 GHz), and an estimated

showing in Fig. 29(b).

gain is little lower than in the previous case.

efficiency of 41%. The measured and simulated E- and H-plane radiation patterns at 57 GHz, 60 GHz and 62 GHz are shown in Figs. 31(a) and (b) respectively for the optimised superstrate dimension. It is noted that the radiation patterns are found to be broad and in good agreement with simulations. The measured cross polar levels are -28 dB for H-plane and -26 dB for E-plane respectively. The radiation patterns are verified to be the same in all the frequencies in the band of interest. The measured HPBWs are 2.5° for H-plane and Eplane respectively at 60 GHz.

Fig. 30. Variation of S11 and gain for a 16 x 16 superstrate aperture antenna array. Simulation ; measurement .

Fig. 31. Measured and simulated E-plane (a) & H-plane (b) radiation patterns of the 2 x 2 superstrate antenna array (Co and Cross polarisation).

It is noted from the study that single/small array superstrate antenna technology is very good for high gain, wide bandwidth and high efficiency, but is not suggestive for big arrays because of the gain go up is not upto the point but is good in terms of efficiency.
