Carbon - silicone hand phantoms for phone testing: material conductivity and permittivity compared to hand tissues

Abstract

We suggest that to best reproduce the interaction between a handheld wireless transceiver and the hand holding it, the dielectric properties of the hand phantom should be similar to those of the palm of the hand. The proposed target permittivity and conductivity values are presented and tabulated at a number of frequencies of interest. We developed suitable phantom hand materials based on carbon-loaded silicones and demonstrated that it was possible to match the proposed dielectric properties in the frequency range 600 to 6000 MHz.

1. Introduction

Phantom hands should be integral parts in any setup for testing wireless handheld transceivers. However, it was decided not to include a hand phantom in the assessment of the exposure of the head to electromagnetic fields from cellular telephones. In the testing of these telephones, the device under test is placed next to a head-shaped phantom shell filled with tissue equivalent liquid; a volumetric scan of the electric field induced in the head by the device is obtained and expressed as specific absorption rate (SAR). The head phantom, known as standard anthropomorphic model (SAM), has well defined internal and external contours; indeed the whole process is strictly described in standard procedures (IEC62209-1, EN62209:2006, and IEEE1528-2003). The reasons given for not using the hand were: (i) practical difficulties in specifying a unique hand holding position that is applicable to all devices, and (ii) with respect to SAR in the head, numerical studies suggest that not modelling the hand provides a conservative estimate (Kuster et al 1997). The latter is the compelling reason for not using a hand in SAR testing; the former is more of an observation because the practical problems cited are not insurmountable. Technical performance tests are made under realistic conditions, including having hand phantoms to hold the device under test as and where appropriate. SAM shells, filled with tissue equivalent liquid, are used to simulate the head (CTIA 2003), there is, however, no specified standard hand for use in these tests.

Phantom hands have been in use since the early 1990s. Hands available atthat time were more elaborate than practical, made from three types of tissue: bone, muscle and skin each positioned according to the basic hand anatomy, the dielectric properties (relative permittivity ε_r and conductivity σ of the materials matched those of the corresponding tissue in the range 900-1800 MHz (Gabriel 1997a). They served a purpose, demonstrated the effect of the proximity of the hand on the radiation from wireless transceivers but they were rigid and had only a single open grip to accommodate the size of the large cellular telephones of the time. In the late 1990s they were replaced with hands made from homogeneous skin equivalent material as it was argued that, for all practical purposes, the absence of internal anatomy would not affect the coupling mechanism, the interaction is localized and focussing effects are negligible (Kuster et al 1997). The shape and size of the hand were also changed in response to feedback from users and to accommodate smaller devices but the changes were not carried out in a systematic manner or in accordance with pre-determined criteria.

The anthropometric and dielectric properties of the hand phantom are the two most important aspects to be defined; this note covers the dielectric properties. Firstly, we will propose target dielectric properties; we will then show that it is possible to develop materials that are physically suitable for making hand phantoms and meet the target dielectric properties.

2. Rationale for the choice of hand material properties

In an ideal-case scenario, a phantom hand should couple to the electromagnetic field of the device in the same way and to the same extent as a human hand. In the case of a hand holding a transceiver, the interaction is in the near field of the device and the coupling is mostly inductive rather than capacitive; that is, the magnetic field induces eddy currents in the hand. The capacitive coupling is less effective because of the high relative permittivity (Kuster and Balzano 1992). For a given handheld device, grip is an important factor in the interaction of the near field with the hand. Grip determines, the proximity and position of the source with respect to the hand and hence the magnitude and distribution of current at the surface of the hand.

The dielectric properties of the palm and inner forearm have beenreported in the literature (Gabriel 1997b and Gabriel 2000), other tissues of the hand: muscle, bone, tendon, blood and fat have also been well characterized (Gabriel et al 1996).

Skin differs from other tissues in two ways that affect the measurement of its dielectric properties: (i) the water content of its outermost layer is uneven, ranging from about 20 percent at the surface, increasing gradually with depth to a plateau of 60 - 70 percent within 0.1 - 0.2 mm and (ii) skin has a microscopically rough surface which impedes good contact with topical contact dielectric measuring sensors.

At microwave frequencies, the dielectric properties of tissues correlate strongly with their water content; a contact measurement carried out through the dry surface is dominated by the response of the outermost, low water content layer and would not be representative of the properties throughout whole thickness of the skin. Moistening evens out the water content of the outer skin layer and reduces the roughness of its surface. Consequently a measurement through moistened skin is more reproducible and, arguably, more representative of the dielectric properties of the skin. For these reasons the dielectric properties of the palm and forearm were measured dry and after moistening with water.

Moistening was found to greatly reduce the variability in the dielectricproperties of skin between different subjects; for this reason, the values for moist palm are considered fairly representative of the population. The values for dry palm are lower for both permittivity and conductivity reflecting the lower water content of the uppermost layer and, to a lesser extent the difficulties of ensuring a good contact. They represent the lower bound of the dielectric properties of the palm.

We therefore propose adopting the average between the two values as target dielectric properties for hand phantoms; the values are given in Table 1 at frequencies of interest.

Looking at the question from a different perspective, we have calculated the simple average of the permittivity and conductivity of all tissues of the hand: blood, muscle, tendon, bone, fat and skin; the result is shown in Figure 1 together with the corresponding values for dry and moistened palm skin. It is interesting to note that the average of all tissues fall within the bounds of the two values for palm. This observation reinforces the data in Table 1.

The difference in the permittivity of dry and moistened palm is of the order of 35 percent of their average; this difference is fairly constant across the reported frequency range, the figure is a little more for conductivity. This indicates the range by which we can relax the target values while maintaining the parameters within the bounds of dry and moist palm data. On this basis, a deviation from the target of up to ± 15 percent can be tolerated.

3. Development of phantom hand material

The material for phantom hands should have the prescribed dielectric properties over a wide frequency range, be mouldable, and sufficiently stiff to retain its shape under all test conditions; sturdiness and durability are also desirable. Aqueous gels (Gabriel 1997b) were not considered on account of their narrow bandwidth and fragility. Carbon and aluminium impregnated resins (Gabriel 2000) were also not considered suitable for hands on account of their extreme hardness. Carbon loaded silicones, however, could be made to comply with most of these requirements. The dielectric properties of the materials developed were measured with a coaxial probe using a well-established procedure (Gabriel and Peyman 2007).

4. Results

Four mixtures containing increasing amounts of carbon (graphite) were made and characterized; the results are shown in Figure 2 together with the proposed target values. The increase in permittivity and conductivity is commensurate with the concentration of carbon in the mixture; the mixtures reported in Figure 2 have between 30 and 60 percent carbon by weight. These are binary mixtures; the addition of a third component, such as carbon fibre or aluminium powder, would enable refinement of the process if required.

The data in Figure 2 encompass the range of dielectric properties of body tissue from the low permittivity and conductivity of fat and bone to the higher values for muscle and brain; the mixture with the second highest concentration of carbon is the best match to the values in Table 1. Its properties fall within ± 15 percent of the target permittivity and conductivity values in the frequency 600 - 6000 MHz. A hand made from the hand-equivalent mixture is shown in Figure 3.

5. Other applications

The results obtained show that, by adjusting the carbon content, this solid phantom material can be made to have dielectric properties similar to those of body tissues; this makes it suitable for applications other than hands. For example, it can be used to make SAM heads having dielectric properties close to those of the tissue equivalent liquid specified in CTIA 2003 (Figure 3b). To make an exact equivalent to a SAM shell filled with tissue equivalent liquid, a solid inner SAM head is made and used in conjunction with the ear section of the shell (Figure 3c). There are significant practical advantages to using a solid head instead of a liquid-filled shell in all situations where a liquid tissue phantom is not functionally prescribed.

6. Conclusions

We have assumed that the nature of the interaction between a handheld transceiver and the hand holding it is mostly inductive in nature and greatly affected by the dielectric properties of the tissue in closest proximity to it. On this basis we have recommended target permittivity and conductivity values derived from the dielectric properties of the palm of the hand. We have tabulated the target values at a number of frequencies of interest and proposed that deviations from the target by up to 15 percent can be tolerated without violating the basic values underpinning the target.

We have developed suitable materials based on carbon-loaded silicones and have shown that it is possible for such mixtures to exhibit dielectric properties that meet the target for hand phantoms over a wide frequency range.

7. References