**2.1 Microstrip fed parasitic patch antenna with superstrate**

Here the antenna configuration consists of a microstrip feed, patch and a parasitic patch, as the source, which are loaded by a superstrate. Fig. 2 shows the 3D view (a) and side view (b)

promising candidate for fulfilling the future needs for very high bandwidth wireless connections. It enables up to gigabit-scale connection speeds to be used in indoor WLAN

These new systems will need compact and high efficient millimeter wave front-ends including antennas. For antennas, printed solutions are often demanding for the researchers because of its low profile, lightweight and ease of integration with active components (Zhang et al., 2006). High gain and high efficient antennas are needed for 60 GHz communication due to high path losses at this range of frequencies. Conventional antenna arrays are used for high gain applications. But in these cases for achieving high gain, a large number of elements are needed, which not only increases the size of the antenna but also decreases its efficiency (Lafond et al., 2001), (Kärnfelt et al., 2006) & ( Soon-soo oh et al., 2004). It has been reported that for high gain, a superstrate layer can be added at a particular height of 0.5 λ0 above the

In this chapter the authors are explaining how to develop a wideband, high gain and high efficient antenna sufficient for 60 GHz communications using superstrate technology. Also explains the importance of different sources on antenna performance in terms of bandwidth,

Here the antenna configuration consists of a microstrip feed, patch and a parasitic patch, as the source, which are loaded by a superstrate. Fig. 2 shows the 3D view (a) and side view (b)

ground plane (Choi et al., 2003) , (Menudier et al., 2007) & (Meriah et al., 2008).

**2.1 Microstrip fed parasitic patch antenna with superstrate** 

networks or fixed wireless connections in metropolitan areas.

Fig. 1. Short range communication.

**2. Superstrate antenna technology** 

gain and efficiency.

of the microstrip fed stacked patch antenna with superstrate. It consists of a lower patch with an optimised dimension of 1.63 mm x 1.6 mm on a substrate RT Duroid 5880 (εr = 2.2, t1 = 0.127 mm). The upper patch with an optimised dimension of 1.63 mm x 1.63 mm is printed on the lower side of a parasitic substrate RT Duroid 5880 (εr = 2.2, t2 = 0.254 mm).

Fig. 2. Cutting plane of stacked patch antenna with superstrate.

The distance between the lower patch and the upper patch is optimized, by simulation using CST Microwave Studio**®,** as h1 = 0.35 mm for a resonance at 60 GHz for a larger bandwidth and gain. This antenna is then loaded with superstrate. The material used for the superstrate is Roger substrate RT6006 (εr = 7.5 at 60 GHz, t = 0.635 mm). The dimension of the superstrate and the height from the ground plane are optimized as explained below. The variation of gain and VSWR bandwidth with the variation of superstrate dimension (0.73λ0, 1.1λ0, 2λ0) for different heights (0.5λ0, 0.6 λ0, 0.7λ0) is shown in Figs. 3 (a-c). It is noted that the maximum gain with good bandwidth is achieved for a superstrate dimension of 1.1λ0 with a height = 0.6λ0, and is equal to 13.6 dB with almost flat over a frequency range of 59 GHz to 64 GHz (Vettikalladi et al., 2009). In all other cases either the gain is less than the above value or the VSWR bandwidth is poor. Also noted that when the superstrate dimension is higher than 1.1λ0, the gain goes down when the height h2 varies from 0.5λ0 to 0.7 λ0.

Fig. 3. variation of s11 and gain with superstrate dimension and height from ground plane. a) h2=0.5 λ<sup>0</sup> b) h2= 0.6 λ0 c) h2=0.7 λ0 for size =.73 λ0 size =1.1 λ<sup>0</sup> size =2 λ0 .

From the literature (Gupta et al., 2005), the theoretical height between the superstrate and ground plane is 0.5λ0, but in this work it is found to be 0.6 λ0 for maximum gain, it may be due to the stacked patch. Fig. 4 shows the return loss, simulated directivity with theoretical values, and simulated and measured gain of the prototype with superstrate. It is found that there is a gain reduction in the measurement, which is due to the variation of exact heights from the theoretical values as shown in Table I.

Fig. 4. (a) Comparison of return loss and gain with superstrate.

96 Wireless Communications and Networks – Recent Advances

Fig. 3. variation of s11 and gain with superstrate dimension and height from ground plane.

a) h2=0.5 λ<sup>0</sup> b) h2= 0.6 λ0 c) h2=0.7 λ0 for size =.73 λ0 size =1.1 λ<sup>0</sup>

size =2 λ0 .


Table I. Variation of prototype parameters from exact values.

It is very difficult to maintain the exact thickness h1, hence we inserted a substrate cut in the form of a rectangular U shape (Fig. 4(b)), with width 2mm, thickness 0.38mm and permittivity 2.2. Also the thickness h2 is varied to 3.48mm instead of 3mm (0.6 λ0) and hence is the reason for the reduction of gain to nearly 10 dB. The E and H planes radiation patterns at 57 GHz and 58 GHz are shown in Figs. 5(a-b). The radiation patterns are found to be broad. There is a cross polar level of less than -20 dB on both the planes. The measured half-power beam widths are found to be 37° for E and H planes at 57 GHz, and 38° and 41° for E and H planes respectively at 58 GHz.

Fig. 5. Measured and simulated E-plane & H-plane radiation patterns of the parasitic patch Superstrate antenna (a) 57 GHz, (b) 58 GHz.

It is noted that for microstrip fed stacked patch antenna, the optimized superstrate size is 1.1 λ0 for getting maximum gain and broad pattern. This value is considered as the limitation of size in this case. It is also observed from Fig. 3, that when the superstrate size is higher than 1.1 λ0, and when the height varies from 0.5 λ0 to 0.7λ0 the broad nature of the gain decreases and starts coming down at 60 GHz .I.e. the pattern changes from broadside to sectorial and then to conical for different frequencies in the band as shown in Fig. 6 (for a superstrate size of 2 λ0 & h2=0.6 λ0), which may suitable for some other application (Vettikalladi et al., 2009b). Here, the small superstrate size is due to the presence of the parasitic patch that disturbs the field in the cavity (thickness = 0.6 λ0).

Fig. 6. Gain pattern for a microstrip fed parasitic patch superstrate with a superstrate size = 2 λ0 , for different frequencies in the band.

Since this kind of prototype is very difficult to manufacture and hence we are going to discuss with other kind of technology.

### **2.2 Slot coupled superstrate antenna**

98 Wireless Communications and Networks – Recent Advances

(a)

57GHz

58GHz

Fig. 5. Measured and simulated E-plane & H-plane radiation patterns of the parasitic patch

(b)

It is noted that for microstrip fed stacked patch antenna, the optimized superstrate size is 1.1 λ0 for getting maximum gain and broad pattern. This value is considered as the limitation of size in this case. It is also observed from Fig. 3, that when the superstrate size is higher than 1.1 λ0, and when the height varies from 0.5 λ0 to 0.7λ0 the broad nature of the gain decreases and starts coming down at 60 GHz .I.e. the pattern changes from broadside to sectorial and then to conical for different frequencies in the band as shown in Fig. 6 (for a superstrate size of 2 λ0 & h2=0.6 λ0), which may suitable for some other application (Vettikalladi et al., 2009b). Here, the small superstrate size is due to the presence of the parasitic patch that disturbs the

Superstrate antenna (a) 57 GHz, (b) 58 GHz.

field in the cavity (thickness = 0.6 λ0).

In this section, we are explaining a superstrate antenna with aperture coupled source as the excitation. We are also showing the importance of the size of the superstrate for getting maximum gain and also for getting consistent radiation pattern all over the frequency range of interest. Fig.7 shows the side view and the 3D view of a slot coupled patch antenna with superstrate. The slot is optimised to 0.2 mm x 1 mm for maximum coupling with a stub length of 0.75 mm. In order to consider the easiness of implementation; we used a thick ground plane of thickness t=0.2 mm. The antenna consists of a patch with optimised dimension 1.3mm x 1.3mm on a substrate RT Duroid 5880 of permittivity 2.2 and a loss tangent tanδ = 0.003 with a thickness t1 = 0.127 mm. Low thickness and low permittivity substrate are used for reducing surface waves. A dielectric superstrate is added above the slot coupled patch antenna (Vettikalladi et al., 2009a). Here we used only one layer to avoid the technological manufacturing problems when many layers are used at 60 GHz. The material used for the superstrate is Roger substrate RT6006 with a relative permittivity of 7.5 at 60 GHz. Theoretically the thickness of superstrate must be λg/4 (0.456 mm), but here we took the thickness (t2 =0.635 mm) close to the theoretical thickness available in market for good antenna performance. The distance between the superstrate and ground plane is 0.5 λ<sup>0</sup> as per the theory (Gupta & Kumar, 2005). A Rohacell foam layer of permittivity 1.05 is sandwiched between base antenna and superstrate to fix all the layers.

Fig. 7. Cutting plane and 3D view of aperture coupled antenna with superstrate, ground plane size = 30 x 30 mm².

Usually in all the known superstrate antennas large superstrates are used for improving the gain which not only increases the size of the antenna but also decreases the S11 bandwidth. But our objective is different, we want to use a small superstrate for obtaining high stable gain and consistant radiation pattern all over the frequency band of interest. To study the effect of superstrate size ' S ' and hence to optimize, we considered four square sizes (1 λ0, 2 λ0, 4 λ0 and 6 λ0). Simulations are done using CST Microwave studio**®**. Fig. 8 shows the CST results of S11 and gain variations of the slot coupled antenna without superstrate and with varying superstrate size. It is observed that the S11 and gain vary with various size of the superstrate. When there is no superstrate, the antenna radiates at 60 GHz with a bandwidth of 3.7% over a frequency range of 58.9 to 61.1 GHz with a maximum gain of 5.9 dBi. It is noted that with superstrate the gain is highest for a superstrate size of 2 λ0. The 2:1 VSWR bandwidth is noted to be BW = 58.7 - 62.7 GHz i.e. 6.7% with a maximum gain of 14.9 dBi. It is also noticed that the gain decreases when the size of the superstrate is above or below 2 λ0.

Fig. 8. Variation of S11 and gain without superstrate and with various superstrate dimensions. without superstrate 1 λ0 2 λ0 4 λ0 6 λ0 .

Superstrate

Patch

Ground plane

Feedline

Substrate

Fig. 7. Cutting plane and 3D view of aperture coupled antenna with superstrate, ground

foam

gain decreases when the size of the superstrate is above or below 2 λ0.

Fig. 8. Variation of S11 and gain without superstrate and with various superstrate

dimensions. without superstrate 1 λ0 2 λ0 4 λ0

Usually in all the known superstrate antennas large superstrates are used for improving the gain which not only increases the size of the antenna but also decreases the S11 bandwidth. But our objective is different, we want to use a small superstrate for obtaining high stable gain and consistant radiation pattern all over the frequency band of interest. To study the effect of superstrate size ' S ' and hence to optimize, we considered four square sizes (1 λ0, 2 λ0, 4 λ0 and 6 λ0). Simulations are done using CST Microwave studio**®**. Fig. 8 shows the CST results of S11 and gain variations of the slot coupled antenna without superstrate and with varying superstrate size. It is observed that the S11 and gain vary with various size of the superstrate. When there is no superstrate, the antenna radiates at 60 GHz with a bandwidth of 3.7% over a frequency range of 58.9 to 61.1 GHz with a maximum gain of 5.9 dBi. It is noted that with superstrate the gain is highest for a superstrate size of 2 λ0. The 2:1 VSWR bandwidth is noted to be BW = 58.7 - 62.7 GHz i.e. 6.7% with a maximum gain of 14.9 dBi. It is also noticed that the

plane size = 30 x 30 mm².

6 λ0 .

There is a gain enhancement of 9 dB with the superstrate. Fig. 9 shows the comparison of measured and simulated S11 and gain for the optimised superstrate size of 2 λ0. Table II gives the comparison of measured and simulated S11 and gain for the optimised superstrate size. It is noted in S11 that there is a frequency band shift of 2.8% (1.7 GHz), when a Vconnector is used and a frequency band shift of 1.5% when a V-band test fixture is used. These frequency shifts are maybe due to the combined effect of connectors and the inaccuracy of the distance between patch and superstrate for the experimental prototype.

Fig. 9. Variation of S11 and gain with a superstrate dimension of 2 λ0. Simulation ; measured with V coaxial mounting connector ; measured with V test fixture .


Table II. Comparison between simulated and measured results of aperture coupled superstrate antenna.

Also the gain measured and simulated are in good agreement but with a frequency shift as explained. The gain is measured using comparison technique with a standard horn of known gain. For calculating the efficiency, we compared the measured gain with the simulated directivity. The measured and simulated E plane radiation patterns are shown in Fig. 10a for the optimised superstrate dimension. It is clear from Fig. 9 that the measured S11 and gain are shifted; the measured gain is maximum between 57 to 59 GHz and simulated gain is from 59 to 61 GHz. Hence the radiation patterns are plotted by taking in account of this frequency shifting (e.g.; that is radiation pattern plotted is, 60 GHz simulation and 58 GHz measurement, and so on). It is noted that the radiation patterns are found to be broad and in good agreement with measurements, and there is a cross polar level of less than -28 dB at all frequencies. The radiation patterns are verified to be the same in all the frequencies in the band of interest. The measured half-power beam width is found to be 23° at 58 GHz. Also verified by simulation that the back radiation in this case is below - 22 dB as compared to the antenna without superstrate (-12 dB).

Fig. 10a. Measured and simulated E-plane radiation patterns of superstrate antenna.

The measured and simulated H plane radiation patterns are shown in Fig. 10b for the optimised superstrate dimension. The radiation patterns are also plotted by taking in account of shifting as explained in E plane radiation pattern. It is noted that the radiation patterns are found to be broad and in good agreement with measurements, and there is a cross polar level of less than -28 dB at all frequencies. The measured half-power beam width is found to be 22°at 58 GHz.

Fig. 10b. Measured and simulated H-plane radiation patterns of superstrate antenna.

When the superstrate size is higher than 2 λ0, the broad nature of the pattern disappeared at 60 GHz. Fig. 11 shows the simulated (60 GHz) and measured (58 GHz) H plane radiation patterns of the antenna with a superstrate dimension of 6 λ0**.** It is noted that the radiation patterns change from broad side to sectorial / null at 60 GHz, which is also useful for some other applications. It concludes that the dimension of the superstrate is critical for the optimum performance of the antenna. To conclude, the dimension of the superstrate is very important in order to get the consistent radiation pattern for the entire frequency band and it is found to be 2 λ<sup>0</sup> in this case. This is the main difference from the already developed superstrate antennas published in the literature.

Fig. 11. Measured and simulated H-plane radiation pattern for a superstrate dimension of 6 λ0. Co-simulated Co-measured Cross-simulated Cross-measured .

#### **2.2.1 Slot coupled 2x2 superstrate antenna array**

102 Wireless Communications and Networks – Recent Advances

Co-polar

59 GHz -sim 60 GHz -sim 61 GHz -sim


Cros-polar

Angle(Degree)

Fig. 10a. Measured and simulated E-plane radiation patterns of superstrate antenna.

57 GHz -meas 58 GHz -meas 59 GHz -meas

The measured and simulated H plane radiation patterns are shown in Fig. 10b for the optimised superstrate dimension. The radiation patterns are also plotted by taking in account of shifting as explained in E plane radiation pattern. It is noted that the radiation patterns are found to be broad and in good agreement with measurements, and there is a cross polar level of less than -28 dB at all frequencies. The measured half-power beam width

Co-polar

59 GHz -sim 60 GHz -sim 61 GHz -sim


Cros-polar

Angle(Degree)

Fig. 10b. Measured and simulated H-plane radiation patterns of superstrate antenna.

57 GHz -meas 58 GHz -meas 59 GHz -meas





Magnitude(dB)


0

is found to be 22°at 58 GHz.



Magintude(dB)


0

Fig.12 (a) & (b) show the side view of an aperture coupled 2 x 2 patch antenna array with superstrate and the feeding network. The distance between the elements in the array are optimized to be d = 1.3 λ0 for obtaining maximum gain and to minimize coupling. All the base antenna parameters and substrate and superstrate are same as explained in section 2.2.

As explained in section 2.2, usually, large superstrates are used for improving the gain. But our objective is different: we want to use the smallest superstrate for obtaining high stable gain and consistant radiation pattern in the frequency band. To study the effect of superstrate size ' S ' and hence to optimize it, here also we considered four square sizes (2.4 λ0, 3.2 λ0, 4 λ0 and 6 λ0). Simulations are done using CST Microwave studio. Fig. 13 shows the CST results of S11 and directivity variations of the 2 x 2 slot coupled antenna array with varying superstrate size. The S11 and directivity are affected by the size of the superstrate: the highest directivity of 18 dBi is obtained for a superstrate size of 3.2 λ0. The resulting 2:1 VSWR bandwidth is 5% from 58.6 to 61.6 GHz. It is also noticed that the directivity decreases when the size of the superstrate is above or below 3.2 λ0 and hence the optimised size of the 2 x 2 superstrate antenna array is 3.2 λ0 x 3.2 λ<sup>0</sup>**.** Fig. 14 shows the comparison of measured and simulated S11, and measured gain with simulated directivity for the optimised superstrate size of 3.2 λ0.

Fig. 12. a) Cutting plane of an aperture coupled 2 x 2 antenna array with superstrate, ground plane size = 6 λ0 x 10 λ0, for connecting V band connector and for ease of measurement purpose, b) 2x2 feeding network.

Fig. 13. Variation of return loss and directivity with various superstrate dimensions. 2.4 λ0 3.2 λ0 4 λ0 6 λ0 .

VSWR bandwidth is 5% from 58.6 to 61.6 GHz. It is also noticed that the directivity decreases when the size of the superstrate is above or below 3.2 λ0 and hence the optimised size of the 2 x 2 superstrate antenna array is 3.2 λ0 x 3.2 λ<sup>0</sup>**.** Fig. 14 shows the comparison of measured and simulated S11, and measured gain with simulated directivity for the

Fig. 12. a) Cutting plane of an aperture coupled 2 x 2 antenna array with superstrate, ground plane size = 6 λ0 x 10 λ0, for connecting V band connector and for ease of measurement

(a)

Fig. 13. Variation of return loss and directivity with various superstrate dimensions.

2.4 λ0 3.2 λ0 4 λ0 6 λ0 .

optimised superstrate size of 3.2 λ0.

purpose, b) 2x2 feeding network.

Fig. 14. Variation of S11 and gain with a superstrate dimension of 3.2 λ0. Simulation measured .

Table III gives the comparison of measured and simulated results for the optimised superstrate size (Vettikalladi et al., 2010a). It is found that the measured maximum gain is 16 dBi with S11 bandwidth of 6.7% and an estimated efficiency of 63%. With superstrate there is a gain enhancement of 4 dB compared to the classical 2 x 2 array (Liu et al., 2009, Book chapter 5, O. Lafond & M. Himdi). The measured gain is maximum at 58 GHz while the simulated directivity is maximum at 59 GHz, which corresponds to 1.7% frequency shift.


Table III. Comparison of simulated and measured 2 x 2 superstrate antenna array.

The simulated and measured E-plane radiation patterns are shown in Fig. 15 for the optimised superstrate dimensions. It is clear from Fig. 14 that the measured gain is maximum between 58 to 59 GHz and simulated is from 59 to 60 GHz. Hence the radiation patterns are plotted by taking in account of this 1.7% shift (e.g. ; that is radiation pattern plotted is, 60 GHz simulation and 59 GHz measurement, etc). It is noted that the radiation patterns are found to be broad and in good agreement with measurements. The measured half-power beam width (HPBW) is found to be 17° at 59 GHz.

Fig. 15. Measured and simulated E-plane radiation patterns of 2 x 2 superstrate antenna array. 58 GHz - measured 59 GHz - simulated 59 GHz - measured 60 GHz - simulated .

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 59 GHz - measured 60 GHz - simulated .

account the shift as explained for E-plane radiation patterns. It is noted that the radiation patterns are found to be broad and in good agreement with measurements. The measured HPBW is found to be 16° at 59 GHz. The cross polarisation level is lower than -26 dB on both the E and H-plane, and is lower than -19 dB at 45° cut plane in 3D pattern.
