A sequence of developmental simulations commenced on a 40 X 40 mm ground plane, to investigate the options available for varying the block length to width ratio to optimize η and BW while positioning the TM₁₀₁, TM₁₀₂ mode frequencies to 1800 and 2450 MHz, respectively. The third 1900 MHz band was introduced by the addition of two metal strip rings around the block, realizing a tri-band antenna. The ground plane was then extended to 40 X 100 mm to comply with typical handset dimensions and by varying μ’ and ε’, monopole length, position, width and gap of the rings, the fourth UMTS band was achieved, resulting in the realization of a quad-band material loaded antenna. One final constraint was imposed on the simulation: the maintenance of the nulls in the radiation patterns in line with the monopole axis. It was achieved by metallizing the block face overlooking the ground plane, leaving a 0.1 mm gap at the block base to electrically isolate the metallization and lower ring from the ground plane itself. The finalized design is illustrated in Fig. 5 and the tri-band and quad-band coverage achieved at an S₁₁ level of -6 dB are shown in Fig. 6. With no head present, the pattern cuts for the first three quad-bands in Fig. 7(a) and (b) are predominantly dipole-like with well-positioned nulls for the head and hence low SAR operation.

Fig. 5. Layout of simulated handset showing quad-band rectangular material coated antenna with central dipole excitation mounted on 100 X 40 mm ground plane.

Fig. 6. Simulated S₁₁ characteristic of triband and quadband handset antenna of Fig. 5; tan δ_ε = tan δ_μ = 0.03 (triband); tan δ_ε~ 0, tanδ_μ= 0.06 (quadband).

Fig. 7. Simulated radiation patterns of quad-band antenna shown in Fig. 5. (a) yx plane without head, (b) yz plane without head, (c) yx plane with head, (d) yz plane with head.

The Bluetooth band pattern cuts resemble those of the UMTS band and give smooth overall coverage. The scattering from the head is far less than that experienced by other types of handset antennas due to the presence of the positioned pattern nulls and the pattern cuts Fig. 7(c) and (d) show less scattering interference in the yx cut than the yz cut.

The quad-band antenna was modelled for two different cases of losses: i) tan δ_ε = tan δ_μ = 0.03 and ii) tan δ_ε~0, tan δ_μ=0.06, to check the antenna performance against material variation. The results were satisfactory and in both cases the four bands were covered by slight modifications to the μ’ and ε’ values. The variation of the design parameters with losses as well as a summary of the overall antenna erformance is presented in Table II.

Other types of handset antennas such as the PIFA and meander- line antennas typically have radiation efficiencies of 60% that reduce to 20% with the head present, in which case 10 g (W/kg) SAR values between 1 and 2 are encountered have been measured for a head separation distance of 8 mm. The normalizing input power is not stated in. In contrast, the quad-band handset antenna offers a significant reduction in SAR together with comparable radiation efficiencies in the presence of the head. In addition, the ground plane currents were minimized in the design process by viewing them in the simulation stages and iterating the positions of the various metallized rings and strips. Other ways of minimizing the ground plane currents have been reported, with the common aim of creating a stand-alone plug-in antenna component that is less sensitive to its environment. The same head model was also simulated in the presence of a hand, 10 mm away from the antenna. Compared to other types of antennas whose efficiency drops by 10% when 15 mm away from the hand, it was found that the quad-band antenna is less sensitive to the presence of the hand. The radiation efficiency reduces by a maximum of 12% at 1800 MHz as seen in Table II.

TABLE II

Simulated design parameters and performance of quad-band rectangular material coated antenna of fig. 5 with central monopole excitation mounted on a 100 mm X 40 mm ground plane; W1 = 32 mm, W2 = 20 mm, H = 10 mm; monopole radius = 0.45 mm, ring widths = 2.9 mm for lower and 0.5 mm for upper, the former being 0.1 mm above the ground and the latter having a central 5 mm tuning gap