Capacitive Loading Shrinks Mobile PIFAs

Shrinking cellular telephones call for smaller antennas. One of the more effective antenna designs for mobile applications is the planar inverted-F antenna (PIFA). Fortunately, this proven design can be shrunk without appreciable loss in electrical performance through the use of simple capacitive top loading.

Several methods have been applied for antenna size reduction, all at the expense of lower antenna gain and bandwidth.¹ This is due to the fact that the antenna is used to transform a bounded wave into a radiated wave.³ Of course. this transformation suffers in efficiency when the antenna is much smaller than the design wavelength.⁴ Although the loss in signal level can often be corrected by amplification, the same is not true of the loss in bandwidth. If the impedance match is better than required, broadbanding techniques can be used to increase the bandwidth.⁵

For a cellular configuration, the antenna should use the total volume available.^6,7 For a fixed working volume, the design of a small antenna is a trade-off between bandwidth and gain.^2,8

One method of reducing antenna size is simply by shortening the antenna. However, this approach affects the impedance at the antenna terminals such that the radiation resistance becomes reactive as well. This can be compensated for with capacitive top loading. In practice, the missing antenna height is replaced with an equivalent circuit,² which improves the impedance match and the efficiency.^9,10 This approach has commonly been used with low-frequency antennas which can be several hundred meters in length,² but has found little application in mobile telephony.¹¹

The idea of capacitive top loading has been discussed previously in connection to monopoles and dipoles, but here the technique is adapted to PIFAs.¹¹ For many practical applications a lumped capacitor as well as a distributed capacitor could be used for top loading the antenna. Here, it is more suitable to use a distributed plate capacitor where the open end of the PIFA forms one of the plates.

To demonstrate the effectiveness of a top-loaded distributed capacitor for shrinking PIFAs, an intermediate target was set to reduce the center frequency by 33 percent for a fixed antenna size. The antenna performance in terms of the radiation properties, scattering parameters, electrical near-field distribution, and current distribution were simulated and verified by measurements.^12,13 The results were also compared to an inductor loaded PIFA with the same size and a PIFA that is 60 mm long.

The experimental antenna configuration consists of a 40-mm-long, 1.5-mm-wide, 5-mm-high PIFA located on a 40 × 100-mm ground plane. In all the prototypes, Rohacell material (ε_r = 1.06) is used as the supporting structures of the antenna. The antenna is located at the edge and parallel to the 100-mm edge (Fig. 1). The feed point is located 5 mm from the edge where a 90-deg. bend forms the short to the ground plane. Copper tape (1.5 mm wide) of different lengths was added between the open end of the PIFA and the ground plane. The distance, d, between the two plates in the capacitor, i.e., the open end of the PIFA and the copper tape, is illustrated in Fig. 1. The lower plate of the capacitor is formed by the copper tape; the upper plate is the PIFA.

A 40-mm-long unloaded PIFA has a center frequency that is somewhat higher compared to that of loaded antennas. Therefore, a larger unloaded antenna that has the same center frequency (min. |S₁₁|) as the loaded antenna is presented as well (60 × 1.5 × 5 mm). This offers a more realistic comparison of antennas with and without capacitive top loading.

For the 60-mm-long PIFA, the simulated center frequency (min. |S₁₁|) is 1.2 GHz or 33 percent lower than the center frequency (min. |S₁₁|) for the 40-mm-long PIFA (1.8 GHz). For both antennas, the measured frequencies with the lowest reflection coefficient are approximately 10 percent lower than the simulated results. This difference could be caused by slight differences between the simulated model and the prototype. Also, the resolution used in the simulation can cause some discrepancy (where converged results were obtained using 20 cells per wavelength and edge cells).¹³

The impedance curves for the 40-mm antenna shown on the Smith chart of Fig. 2 indicate the location of the center frequency 11|> (1.8 GHz), where the return loss is 6 dB (1.78 GHz and 1.87 GHz), and also the center frequency of the other antenna (1.2 GHz). The same is shown for the 60-mm antenna. In the 40-mm PIFA case, it is the 1.2-GHz point that must be shifted toward the 1.8 GHz point by proper loading of the antenna. The reduction in center frequency (min. |S₁₁|) from 1.8 GHz for the 40-mm-long PIFA to 1.2 GHz is accomplished by the capacitor. This frequency reduction corresponds to a size reduction of 33 percent, from 60 to 40 mm.

In two different simulation tests, the plate capacitor loading principle was utilized. First, with a fixed plate area of 4.0 × 1.5 mm, the plate distance, d, is varied from 0.05 to 0.60 mm. Hereafter, the distance is fixed at d = 0.1 mm and the area changed from 0.8 to 9.0 mm. The areas, A of the two plates together with the plate distance, d gives the capacitances shown in Table 2 sorted in order of ascending capacitance. The capacitance can be calculated as:

C = Aε₀ε_r /d (1)

where:

ε₀ = the free space permittivity, 8.8542 × 10^-12 F/m .

Even though sticky tape is used as the spacer in the prototype, the relative permittivity ε_r, is set to 1. This means the measurement values are somewhat higher than the simulation values. In this way, overlapping capacitance values in the range from 0.13 to 2.13 pF are used. Prototypes were fabricated with plate area and separation covering the simulated range. Simulated results based on varying the plate capacitance (Table 1) are shown in Figs. 3 and 4 as solid lines, while measured results are shown with crosses. Linear trend lines are added for easier visualisation of the behavior of the results as a function of the capacitance.

The simulated center frequency (min. |S₁₁|) decreases almost linearly from 1.8 to 0.8 GHz for capacitances below 1.5 pF. The experimental results follow the same trend, at lower values. The simulated bandwidths are almost constant at 120 MHz (9 percent) when adding a capacitor that is smaller than 1.1 pF. Above 1.1 pF, the impedance match reduces to above –6 dB, hence no 6-dB bandwidths are observed. The 2.1-pF unit deviates from the remaining cases, since it yields a –10-dB impedance match and 7-percent bandwidth (80 MHz). The radiation efficiency for this value is also higher compared to the 1.6-pF case. Again, the measured results has lower values, due to the poor impedance match, i.e., between –8 dB and –2 dB for the unloaded and the 2.2 pF case.

The simulated and measured radiation efficiency is almost constant 80 percent for capacitances below 1.1 pF followed by a decrease for higher capacitances. The bandwidth in which the radiation efficiency are above 50 percent are shown, illustrating more or less the same trends, i.e., rather constant (a slight decrease) below 1.1 pF and rapidly decreasing above 1.1 pF.

It is possible to improve the PIFA performance by adding a plate capacitor at the open end of the PIFA. Generally speaking, the results could be divided into two groups, one for capacitances below 1.1 pF and one above 1.1 pF.

Below 1.1 pF, the results are continuous and the best case with respect to center frequency (min. |S₁₁|) reduction and unchanged bandwidth and efficiency are obtained for a capacitance of approximately 1.1 pF. Here the simulated center frequency (min. |S₁₁|) is decreased by 32 percent from 1.80 to 1.22 GHz. The reflection coefficient is –12 dB, the bandwidth is 9 percent, and the radiation peak efficiency is 91 percent. Measurements verified the trends, although at somewhat lower values, most likely due to loss in the plate capacitor. Above 1.1 pF, the simulated as well as the measured results show rather decreasing performance in terms of poor impedance match, hence lower bandwidth and lower radiation efficiency. Therefore, capacitances below 1.1 pF should be used. A prototype loaded with a capacitance of 1.11 pF will be investigated further here for this reason.

The simulated center frequency (min. |S₁₁|) of the capacitor-less PIFA is 1.80 GHz, but it decreases to 1.22 GHz when a 1.06-pF plate capacitor is added. The unloaded configuration yields peak radiation efficiency of 91 percent, which is 6.5 percent higher compared to the case with the 1.06-pF plate capacitor. The measured center frequency (min. |S₁₁|) for the prototype PIFA without any capacitor is 1.63 GHz, dropping to 0.99 GHz with a 4.0 × 1.5-mm plate capacitor separated by 0.1 mm (an added capacitance of 1.11 pF). Figure 5 also shows the measured radiation efficiency. The peak efficiency decreases from 83 to 62 percent when the capacitor is added.

The deviation between the simulated and measured center frequencies is mainly due to the cable used in the measurements, the simulated ideal assumptions, i.e., loss less and free space, as compared to the Rohacell material used for the prototype, and the deviation between the actual prototype and the model used in the simulation. The relative decrease in center frequency (min. |S₁₁|) is approximately 39 percent for the experimental results and 32 percent for the simulated results. The reason for this deviation, and the lowered radiation efficiency, comes from the slightly higher value of the capacitance, and the use of sticky tape as the spacer in the prototype.

The total electrical field components of the radiation patterns shown in Fig. 6 indicate almost omni-directional properties for the 40-mm-long top-loaded PIFA. Good resemblances between the simulation and measurement results are obtained. The measured maximum gain is 2.1 dBi, slightly lower than the simulated gain.

The simulated current distribution shown in Fig. 7 illustrates the capacitive top loading principle. For both antennas, i.e., the unloaded 60-mm PIFA and the capacitive-loaded 40-mm PIFA, the current density is highest near the feed point and zero at the open end. The effect of the plate capacitor located at the open end is clearly visible in Fig. 8. The ripple associated with the decrease of the current density as a function of the position on the antenna arm is due to the model definition in the IE3D electromagnetic (EM) simulation software from Zeland Software (Fremont, CA); in practice, it will have a continuously decreasing behavior. The two curves are parallel until 34 mm on the PIFA arm, which corresponds to the point where the plates are overlapping. Here, a discontinuity occurs and the current decreases to zero faster than for the unloaded 60-mm-long PIFA. Hence, in using the top-loading principle, it could be possible to remove fractions of the antenna arm. The capacitor replaces the last 20 mm on the 60-mm-long PIFA for a fixed center frequency.

The key results from the capacitor-loaded antenna can be compared in Table 2 with the 60-mm-long unloaded antenna as well as with a 40-mm-long inductor-loaded antenna.¹⁵ Comparing the 60-mm-long unloaded PIFA with the two 40-mm-long loaded antennas shows that the advantages of reduced size is a trade-off in decreased bandwidth. An 18-nH inductor yields the same frequency reduction as the 1.1-pF capacitor, although with higher gain, possibly due to the sticky-tape separator in the capacitor.

A small 0.3 cm³ PIFA located on a dielectric foam 5 mm above a 40 × 100-mm ground plane was also investigated. When comparing the 40-mm-long top loaded antenna with the 60-mm unloaded antenna, the major benefit is reduced size for a fixed center frequency (min. |S₁₁|), although at the expense of reduced efficiency and reduced bandwidth. A measured 39-percent reduction in center frequency (min. |S₁₁|) was accomplished by using a 1.1-pF distributed capacitor formed by a 4 × 1.5-mm plate located 0.1 mm below the open end of the PIFA, between the ground plane and the PIFA. The 1.1-pF distributed capacitor was also used with a 5-mm-high 0.3-cm³ PIFA located on a 40 × 100-mm ground plane. The one drawback is the increased complexity and hence the price of the antenna system.

ACKNOWLEDGMENT
This work has been supported by Nokia Denmark.

REFERENCES

1. R.F. Harrington, "Effect of antenna size on gain, bandwidth and efficiency," Journal of research of national bureau of standards, D-radio Propagation, Vol. 64D, pp. 1-12, Jan. 1960.
2. G. Hall, Ed., The ARRL Antenna Book, 15th ed,, third printing, Chap. 7, 1990.
3. "The IEEE standard definitions of terms for antennas," IEEE Transactions on Antennas and Propagation, Vol. 17, May 1969.
4. R.C. Hansen, "Fundamental limitations in antennas," Proceedings of the IEEE, Vol. 69, Feb. 1981.
5. H.A. Wheeler, "The wide-band matching area for a small antenna," IEEE Transactions on Antennas and Propagation, Vol. 31, May 1983, pp. 364–367.
6. L.J. Chu, "Physical limitations of omnidirectional antennas," Journal of Applied Physics, Vol. 19, December 1948, pp. 1163-1175.
7. J.S. McLean, "A re-examination of the fundamental limits on radiation Q of electrically small antennas," IEEE Transactions on Antennas and Propagation, Vol. 44, May 1996, pp. 672-675.
8. H.A. Wheeler, "Fundamental limitations of small antennas," Proceedings of the IRE, Vol. 35, December 1947, pp. 1163-1175.
9. C.W. Harrington, "Monopole with Inductive Loading," IEEE Transactions on Antennas and Propagation, July 1963, pp. 394–400.
10. G.S. Smith, "Efficiency of electrically small antennas combined with matching networks," IEEE Transactions on Antennas and Propagation, Vol. AP-25, May 1977, pp. 369–373.
11. R.E. Collin, Antennas and radiowave propagation, McGraw-Hill, New York 1985, pp. 97-104.
12. Zeland Software, Fremont, CA, www.satimo.com.
13. Zeland Software, Fremont, CA, www.zeland.com.
14. J. Appel-Hansen, "Antenna measurements," Chap. 8 in A.W. Rudge, Ed., The Handbook of Antenna Design, Peter Pelegrinus Ltd., 1982.
15. J. Thaysen and K.B. Jakobsen, "Size reduction technique for mobile phone PIFA antennas using lumped inductors," submitted 2005.