A) Measured and simulated S-parameters

The measured S-parameters of the proposed antenna are shown in Fig. 2 along with the simulated ones. Ultra wide impedance bandwidth with a lower edge frequency of 3.4 GHz for port 1 excitation and 3.8 GHz for port 2 excitation is noted from the S ₁₁ and S ₂₂ curves. The difference in the lower edge frequencies is because the two circuits are not identical.

Fig. 2.

Measured and simulated S-parameters.

The presence of multiple resonances in the reflection coefficient characteristics is due to the increased number of reactive elements in the circuit. A single uniform slot is electrically a tank circuit consisting of inductance and capacitance which depend on the slot length and width. The presence of multiple sections of different lengths and widths leads to a number of such tank circuits in parallel. This increases the number of reactive elements causing multiple resonances to appear. The merging of these resonances gives the ultra wide bandwidth. A symbolic equivalent circuit of one of the arms of the proposed antenna is shown in Fig. 3. In the circuit each slot step is represented by an R–L–C circuit while the metallic short at the slot end is shown by the inductance Ls. Although the resistance “R” and inductance “L” are dependent on the length of the slot segment, the capacitance “C” will depend on the width. Similarly, the steps in the microstrip feed are represented by their R–L–G–C parameters while C _o represents the open circuit capacitance at the feed end. The third step in the feed line behaves as a patch antenna and hence its radiation resistance R _rad has been included in series with the R, L, G, C parameters. The coupling of electromagnetic energy from the microstrip feed to the slot is depicted by means of a transformer of turns ratio 1:n. The coupling takes place exactly at the center of the slot.

Fig. 3.

Symbolic equivalent circuit for one of the arms of the proposed antenna.

The measured isolation seen from Fig. 2 varies between 20 and 30 dB in the operating region. The isolation of the proposed antenna has been improved by (1) using L-shape feed line for exciting the horizontal slot arm and (2) by inserting a slant slot stub “I” at the junction of the arms. A slant stub as proposed in [Reference Lee, Chen and Hsu5] restricts the spread of the ground plane currents and thus reduces the coupling over certain frequencies. The simulated isolation S ₂₁ without the stub “I” is also shown in Fig. 2 (dotted green line). The maximum improvement with the stub is seen around 7.5 GHz where the stub length (13 mm) is close to λ _eff/2.

The effect of the PS on S ₁₁ is shown in Fig. 4. The S ₁₁ of the optimized design without the PS (red line, Fig. 4) shows a peak (>−10 dB) near 8.5 GHz. The profile is corrected by the introduction of PS whose length (11 mm) is tuned to create a resonance at the correct position (8.3 GHz). The impedance matching in the neighborhood of the resonance is also improved and the overall response becomes UWB. As the length of the PS is decreased, the introduced resonance shifts upwards. The variations in S ₁₁ with a change in the position of the PS for a fixed slot length (11 mm) are shown in Fig. 5. The resonance is seen to shift toward higher frequencies as the PS is moved away from the patch in the vertical direction or away from the slot in the horizontal direction. Further, there exists an optimal placement for obtaining the best performance. In the antenna, the feed section M3 also behaves as a microstrip fed patch placed on a defected ground. The role of the PS then is to create an additional resonance by itself (where its length l = λ/2) and to shift the resonance of the patch M3 by disturbing the ground plane currents underneath it. At a particular position of PS, when these two resonances favorably combine with the impedance behavior of the main slot, overall wideband characteristics are obtained.

Fig. 4.

Effect of PS length on S ₁₁ and S _21.

Fig. 5.

Variation in S ₁₁ with position of parasitic slot.

B) Aperture field distribution and radiation patterns

To verify the dual polarized behavior, the aperture field distribution for different port excitations is plotted at 5.7 GHz in Fig. 6. When port 1 is excited, horizontally directed electric field gets established in the vertical arm of the slot while little cross polarization is seen in the horizontal arm. The converse is true for port 2 excitation which results in vertical polarization.

Fig. 6.

Aperture electric field at 5.7 GHz.

The measured and simulated radiation patterns of the proposed antenna are shown in Fig. 7. The simulated cross polarization is seen to be around −15 dB along the boresight direction over most of the frequencies. The measured cross-polarization is however higher owing to the poor characteristics of the reference horn antenna used in the measurement system.

Fig. 7.

Measured and simulated radiation patterns.

A) Envelope correlation coefficient

An important quality indicator for a two port antenna is the envelope correlation coefficient ρ _e calculated using (1). A lower value of the correlation coefficient indicates less cross talk during simultaneous usage of the two ports.

(1)$${\rho_e} = \displaystyle{{\vert {S_{11}^\ast} {S_{12}} + S_{21}^\ast {S_{22}}{\vert ^2}} \over {(1 - (\vert {S_{11}}{\vert ^2} + \vert {S_{21}}{\vert ^2}))(1 - (\vert {S_{22}}{\vert ^2} + \vert {S_{12}}{\vert ^2}))}}.$$

The simulated and measured correlation coefficients for the proposed antenna (without notch) are shown in Fig. 13. In the same figure, the correlation coefficients for the band-notch design are also shown. It can be observed that both the simulated and measured correlation coefficients for both the antennas are almost equal and found to have low values (<0.002). The band-notch seen in the S-parameters and peak gain is seen to have little effect on the correlation coefficient as also observed in [Reference Chacko, Augustin and Denidni12] and [Reference Zhao, Zhang, Zhang and Wang15].

Fig. 13.

Measured and simulated envelope correlation coefficients.