interaction of mobile phones with superﬁcial … › assets › downloads › publications ›...

INSTITUTE OF PHYSICS PUBLISHING PHYSICS IN MEDICINE AND BIOLOGY

Phys. Med. Biol. 50 (2005) 2689–2700 doi:10.1088/0031-9155/50/11/017

Interaction of mobile phones with superficial passivemetallic implants

H Virtanen1, J Huttunen1, A Toropainen2 and R Lappalainen1

1 Department of Applied Physics, University of Kuopio, PO Box 1627, 70211 Kuopio, Finland2 Nokia Research Center, San Diego, CA, USA

E-mail: [email protected]

Received 4 February 2005, in final form 5 April 2005Published 18 May 2005Online at stacks.iop.org/PMB/50/2689

AbstractThe dosimetry of exposure to radiofrequency (RF) electromagnetic (EM) fieldsof mobile phones is generally based on the specific absorption rate (SAR,W kg−1), which is the electromagnetic energy absorbed in the tissues per unitmass and time. In this study, numerical methods and modelling were usedto estimate the effect of a passive, metallic (conducting) superficial implanton a mobile phone EM field and especially its absorption in tissues in thenear field. Two basic implant models were studied: metallic pins and ringsin the surface layers of the human body near the mobile phone. The aimwas to find out ‘the worst case scenario’ with respect to energy absorptionby varying different parameters such as implant location, orientation, size andadjacent tissues. Modelling and electromagnetic field calculations were carriedout using commercial SEMCAD software based on the FDTD (finite differencetime domain) method. The mobile phone was a 900 MHz or 1800 MHz genericphone with a quarter wave monopole antenna. A cylindrical tissue phantommodels different curved sections of the human body such as limbs or a head.All the parameters studied (implant size, orientation, location, adjacent tissuesand signal frequency) had a major effect on the SAR distribution and in certaincases high local EM fields arose near the implant. The SAR values increasedmost when the implant was on the skin and had a resonance length or diameter,i.e. about a third of the wavelength in tissues. The local peak SAR valuesincreased even by a factor of 400–700 due to a pin or a ring. These highestvalues were reached in a limited volume close to the implant surface in almostall the studied cases. In contrast, without the implant the highest SAR valueswere generally reached on the skin surface. Mass averaged SAR1 g and SAR10 g

values increased due to the implant even by a factor of 3 and 2, respectively.However, at typical power levels of mobile phones the enhancement is unlikelyto be problematic.

(Some figures in this article are in colour only in the electronic version)

0031-9155/05/112689+12$30.00 © 2005 IOP Publishing Ltd Printed in the UK 2689

http://dx.doi.org/10.1088/0031-9155/50/11/017

http://stacks.iop.org/pb/50/2689

2690 H Virtanen et al

Figure 1. Typical small metallic implants used in orthopaedics and dentistry.

1. Introduction

As radiofrequency (RF) mobile phones have become a part of daily life, possible risks related toexposure to RF electromagnetic (EM) fields of mobile phones have been studied from severalaspects, from cell level to macroscopic effects such as local heating (Fleming and Joyner1992, Riu and Foster 1999, Repacholi 2001, Vanderstraeten and van der Vorst 2004). Mostof the experimental research and numerical calculations have focused on estimating specificabsorption rate (SAR, W kg−1) and local heating. However, only a few studies focusing onthe possible interaction of mobile phones with passive implants exist (Hocking et al 1991,Fleming et al 1992, 1999, Cooper and Hombach 1996), although metallic implants such asthose shown in figure 1 are rather common. Pins, rods, wires, plates, screws, clips and implantsof special designs are widely used, for example, in orthopaedics and dentistry. Nowadays, thedosimetry and the regulation of exposure are mainly based on the SAR values. The maximumvalues for average SAR, averaged over tissue masses of 1 or 10 g, are regulated by IEEEand ICNIRP standards (IEEE 1999, ICNIRP 1998a and 1998b). According to ICNIRP SARshould be averaged over 10 g and the maximum allowable value for the general public is2 W kg−1 for the exposure of head and body excluding limbs. IEEE recommendation is to use1 g average SAR for estimating the local exposure and the corresponding maximum allowablevalue for the general public is 1.6 W kg−1.

Certain metallic implants may be located on the skin or in the surface layers of the bodyclose to a mobile phone. They may act as an antenna and cause strong local enhancementof the EM field near the implant, leading also to stronger energy absorption, i.e. increaseof specific absorption rate (SAR, W kg−1). Especially, implants with resonance dimensionsand in appropriate orientation can be expected to cause the highest enhancement of thefield.

This study examines numerically two basic implant models: pins and rings in surfacelayers of the human body in the near field of a 900 or 1800 MHz mobile phone. The aim was tofind out ‘the worst case scenario’ with respect to energy absorption (SAR) by varying differentparameters such as implant location, orientation, size and adjacent tissues systematically inaccordance with the IEEE recommendations (IEEE 2002). This study provides the basis for

Mobile phones and implants 2691

Figure 2. Experimental arrangement for calculations. The λ/4 monopole antenna of a genericmobile phone is located 10 mm above the surface of a tissue phantom. The tissue phantom is acylinder with a diameter of 150 mm and has a 4 mm thick skin surface layer and a base normallyconsisting of muscle or bone. The vertical axis (z) is along the antenna and the cylinder axis, thex-axis is in the depth direction and the y-axis is perpendicular to x and z. Implants are located indifferent orientations on, inside or under skin.

similar calculations with more authentic implants in a realistic numerical human phantom orfor experimental measurements.

2. Materials and methods

The modelling and the calculations of electromagnetic fields were carried out using commercialSEMCAD software (Schmid & Partner Engineering AG, Switzerland) based on the FDTD(finite difference time domain) method. The software was run on a PC computer with aWindows NT 4.0 operating system. The simulations were run for 10 periods and eight-fold perfectly matched layers (PML) were used to truncate the calculation domain. In thegrid, the voxels in the region of interest (i.e., near to the implant) were substantially smallerthan elsewhere. The smallest voxel was 0.2 × 0.2 × 0.2 mm3 for a pin and 0.2 × 0.2 ×0.5 mm3 for a ring implant while the maximum step in the grid was 7 mm at 900 MHz and ahalf of this at 1800 MHz. Matlab 6.5 software (The Mathworks Inc., Natick, MA, USA) wasutilized in the analysis of results and in plotting.

Figure 2 illustrates the set-up for calculations. The mobile phone was a 900 MHz or1800 MHz generic phone with a conductive case (140 mm × 40 mm × 16 mm) and aquarterwave monopole antenna. The wavelength of 900 MHz (1800 MHz) EM field in theair, λ, is about 33 cm (16 cm) and in the muscle, λt, is about 4.4 cm (2.3 cm). The antennawas oriented along the axis of a cylindrical tissue phantom. The feed point of the antenna wasat 10 mm from the skin surface and it was used as an origin. The cylindrical tissue phantom(diameter 150 mm) modelled different curved sections of a human body such as limbs or ahead. It had a 4 mm thick skin layer on top of homogeneous tissue, which was generallymuscle but in some calculations bone or fat. In this study, skin was modelled as a single layer,although the skin thickness in the human body varies in the range of 1–4 mm and a layer of fat


Table 1. Electrical parameters of different tissues (relative permittivity εr and conductivity σ ) atfrequencies of 900 MHz and 1800 MHz (Gabriel 1996). Penetration depth δ is calculated usingthe formula 1/δ = 2πf (µεrε0/2)1/2((1 + (σ/(2πf εrε0))

2)1/2 − 1)1/2, where µ is roughly 4π ×10−7 H m−1 and f is the frequency of the mobile phone.

εr σ (S m−1) δ (cm)

Frequency (MHz) 900 1800 900 1800 900 1800Muscle 55.0 53.5 0.94 1.34 4.2 2.9Fat 11.3 11.0 0.11 0.19 16.4 9.3Bone 12.5 11.8 0.14 0.28 13.1 6.7Skin 46.1 43.9 0.84 1.23 4.3 2.9

and loose connective tissue (subcutis) of varying thickness is under skin. The parameters usedfor different tissues are given in table 1. At penetration depth, the field strength is attenuatedby a factor of 1/e, i.e. to 37% of the original value and SAR is decreased by a factor of 1/e2,i.e. to 14%. The implants were metallic pins and rings with thickness in the range of 0.5–8 mm and length or diameter in the range of 7–50 mm. They were modelled as perfect electricconductors (PEC), i.e. the electric field inside the implants was zero.

The generic phone is connected to the antenna through the feed point and works as avoltage source with an internal impedance of 50 �. In the simulations, the real impedancedepends on several factors such as surrounding materials. Therefore, all the results werenormalized to a peak input current of 100 mA. For an antenna with 50 � impedance thiswould theoretically correspond to an emitted power of 0.25 W.

3. Results

In order to study the effects of metallic implants, the SAR distributions with and withoutimplants were compared. Furthermore, the maximum local peak values SARpeak and the valuesaveraged over a mass of 1 (SAR1 g) or 10 g (SAR10 g) were determined. All the parametersvaried such as implant location, size, orientation, adjacent tissues and signal frequency had aconsiderable effect on the SAR distribution. In the following, the main results related to eachof these parameters will be demonstrated using figures and tables.

3.1. Orientation of the implant

The orientation of an implant has a major effect on its interaction with the EM field inthe tissues. A slice of SAR distribution for three perpendicular orientations of the implant areshown for a 14 mm long pin implant in figure 3(a) and for a 30 mm ring in figure 3(b). Thecorresponding SAR values are in tables 2(a) and (b). The slice was selected from the zx-planeclosest to y = 0, i.e. the slice dividing the implant into two equal halves. The surface ofthe 4 mm thick skin is at x = 1 cm and the depth in the phantom increases towards the right.The absorption is evidently highest when the longest dimension of the implant and the antennaare parallel.

3.2. Location of the implant

The longest dimension of the implant was selected to be parallel to the antenna and the positionof an implant was varied. The effect of a vertical pin position on the SAR distribution forthe 14 mm pin (φ = 2 mm) on the skin was just minor as long as the pin was within a fewcentimetres from the feed point of the antenna. SARmax

peak and SARmax1 g values for these cases


x (cm)

z (c

m)

xy-plane

2 4

-2

0

2

4

xz-plane

2 4

-2

0

2

4

yz-plane

2 4

-2

0

2

4

0

10

20

30 dB

(b)

x (cm)

z (c

m)

x axis

1 2 3-2

-1

0

1

2

3

4y axis

1 2 3-2

-1

0

1

2

3

4z axis

1 2 3-2

-1

0

1

2

3

4

0

5

10

15

20 dB

(a)

Figure 3. The effect of orientation of the implant on the SAR distribution in xz-plane for a 14 mmlong and 2 mm thick pin (a) and a 30 mm ring (b). The closest distance of the implant from theantenna is the same in all the cases.

Table 2. Relative enhancement of SARmaxpeak and SARmax

1 g values for the (a) pin in figure 3(a) and(b) ring in figure 3(b).

SARmax1 g (W kg−1)

SARmax1 g (W kg−1)ref

SARmaxpeak (W kg−1)

SARmaxpeak (W kg−1)ref

(a) Orientation of the pinx-axis 1.0 3.5y-axis 1.1 5.5z-axis 2.1 47

(b) Orientation of the ringxy-plane 1.0 3.0xz-plane 1.3 6.8yz-plane 1.9 490

are listed in table 3(a). Although the enhancement is highest in the middle position closest tothe antenna, the field is deformed towards the pin ends in all cases. Figure 4 shows the SARdistributions for the 30 mm ring (φ = 2 mm) placed in different depths in the tissue phantomand table 3(b) lists the corresponding values too. Note that although the highest local SAR isreached with a ring on the skin surface, the averaged SAR is highest when the ring is in themuscle. Similar results were obtained for the pin too. Furthermore, the energy absorption is


x (cm)

z (c

m)

on the skin

1 2 3-2

-1

0

1

2

3

4under the skin

1 2 3-2

-1

0

1

2

3

4in the muscle

1 2 3-2

-1

0

1

2

3

4

0

5

10

15

20

25

30 dB

Figure 4. The effect of position of implant on the SAR distribution in xz-plane for a ring atdifferent depths in the phantom.


1 g values for the (a) pin at different verticalpositions and (b) ring cases in figure 4.





(a) Position of the pin on the skinz = −1 cm 1.0 80z = 2 cm 1.2 190z = 4 cm 1.2 100

(b) Position of the ring in depth directionOn the skin 1.9 490Under the skin 1.9 5In the muscle 2.7 6

highest in tissues adjacent to the implant, not always on the skin surface as in the similar casewithout an implant.

3.3. Size of the implant

The effect of the size of a pin or a ring is illustrated in figures 5(a) and (b) and tables 4(a)and (b). In the case of the pins, the absorption is highest with the 14 mm long pin, the lengthof which is about a third of the wavelength λt, i.e. a resonance length (Fleming et al 1999,McIntosh et al 2005). In the case of the rings, a major change in the SAR distribution can beseen as the diameter of the ring increases. Although the local SAR is highest for the smallestring, SAR1 g is clearly highest for the ring of 30 mm in diameter. Furthermore, the thickness ofthe pin or the ring has an effect on the density of EM field lines near the implant and especiallyon the SARmax

peak values. This is shown for a 14 mm pin of three thicknesses in figure 6 andtable 5.

3.4. Surrounding tissues

The effect of tissue type was studied for a pin implant under the skin by varying the surroundingtissue beneath the skin (muscle to bone or fat). The results are given in figure 7 and table 6.The major difference in electrical parameters of the tissues is reflected in SAR distributionswhich can be seen also in the case without an implant in figure 7(b).


x (cm)

z (c

m)

15 mm

1 2 3-2

-1

0

1

2

3

430 mm

1 2 3-2

-1

0

1

2

3

450 mm

1 2 3-2

-1

0

1

2

3

4

0

5

10

15dB

(b)

x (cm)

z (c

m)

7 mm

1 2 3-2

-1

0

1

2

3

414 mm

1 2 3-2

-1

0

1

2

3

428 mm

1 2 3-2

-1

0

1

2

3

4

0

5

10

15

20dB

(a)

Figure 5. The effect of size of the implant on the SAR distribution in the xz-plane for a pin(a) and a ring (b). Note the maximum SAR in the case of an implant with size close to optimumfor resonance effects.


1 g values for the (a) pins in figure 5(a) and(b) rings in figure 5(b).





(a) Length of the pin7 1.2 13

14 2.1 4728 1.4 27

(b) Diameter of the ring15 1.6 1430 2.7 6.250 1.5 2.4

3.5. Frequency of the phone

Table 7 compares the SAR values of the distributions for the frequencies 900 and 1800 MHzwith a pin implant placed at different depths in the tissues. At both frequencies, the absorption


x (cm)

z (c

m)

0.5 mm

1 2 3-2

-1

0

1

2

3

42 mm

1 2 3-2

-1

0

1

2

3

48 mm

1 2 3-2

-1

0

1

2

3

4

0

5

10

15

20

25 dB

Figure 6. The effect of thickness of the implant on the SAR distribution in the xz-plane for a pinat 900 MHz. The pin is just under the skin.


1 g values for the pins in figure 6.

Thickness of SARmax1 g (W kg−1)



SARmaxpeak (W kg−1)refthe pin (mm)

0.5 1.8 1752 2.1 478 1.6 1.1

was highest with the implant on the skin. However, the field penetrates less at 1800 MHz thanat 900 MHz. This is also the case when an implant is not present since the energy absorptionin tissues is more efficient at high frequencies.

As resonance effects may occur in different cavities, e.g. in an implant with an open gap,rings with different sizes of gaps were simulated on the skin or in the skin. The results of thesimulations showed that in the worst cases of these open rings, the highest SAR values wereabout the same as in other worst cases.

The dosimetry and safety limits for RF exposure are based on the average SAR values for1 or 10 g of tissues (IEEE 1999, ICNIRP 1998a and 1998b). Figure 8 compares the calculatedSAR, SAR1 g and SAR10 g distributions for a pin implant. Evidently due to averaging the localhigh SAR values are smoothed. Furthermore, the averaging procedure affects the locationwhere the highest SAR values occur.

Of the studied cases, the highest increase in SAR values was obtained for the implants onthe skin. The local peak SAR values increased even by a factor of 400–700 for a pin and aring. These high values were reached in a limited volume close to an implant surface. On theother hand, the SAR1 g and SAR10 g values were increased due to an implant even by a factorof 3 and 2, respectively. The cases corresponding to the highest SAR values are summarizedand compared to the cases without an implant in table 8.

In order to check reproducibility of calculations, some of them were repeated usingslightly different choices for parameters in the SEMCAD program. The relevant values ofSAR were the same within 5%, which agrees well with an average agreement of around 5–6%for comparison of calculations and experimental measurements with a head phantom at 900and 1800 MHz provided by the manufacturer of the program. In practice, for example, thevariation of electrical parameters in tissues (Gabriel et al 1996) can be expected to causedifferences in SAR estimates typically within 10%.


x (cm)

z (c

m)

muscle tissue

0.5 1 1.5 2 2.5-1

0

1

2

3fatty tissue

0.5 1 1.5 2 2.5-1

0

1

2

3bony tissue

0.5 1 1.5 2 2.5-1

0

1

2

3

0

5

10

15

20 dB

x (cm)

z (c

m)

muscle tissue

0.5 1 1.5 2 2.5-1

0

1

2

3fatty tissue

0.5 1 1.5 2 2.5-1

0

1

2

3bony tissue

0.5 1 1.5 2 2.5-1

0

1

2

3

0

5

10

15

20 dB

(b)

(a)

Figure 7. SAR distributions in different tissues with (a) and without (b) a pin implant. In eachcase the surface layer is skin (4 mm) and the location of the implant is the same, i.e. just under theskin, but the tissue under the skin is different.


1 g values for the pin in figure 7. In eachcase the reference SAR values were calculated in the same tissue but without implant.

Surrounding tissue SARmax1 g (W kg−1)



SARmaxpeak (W kg−1)refunder the skin

Muscle 2.1 47Fat 1.6 53Bone 2.2 50

4. Discussion and conclusions

In the present study, all the relevant parameters (location, orientation, size, surrounding tissuesand frequency) were systematically varied for two model implants (pins and rings) to find outthe worst case, i.e. the highest SAR enhancement due to a passive metallic implant. Onenovelty of this study was to show systematically the effects of different parameters. Basedon the simulations of this study, the local absorption of EM field in a limited volume maybe significantly (even by a factor of 700) enhanced by a conductive implant in the surfacelayer of a human body. The mobile phone and the metallic implant are strongly coupled,especially when the implant is close to a mobile phone, its length is in resonance with thefield, and it is aligned with the antenna. This coupling can be either conductive, magneticor both (Troulis et al 2003). When this coupling is strong, i.e. the mutual inductance of the


x (cm)

z (c

m)

SARpeak

0.5 1 1.5 2 2.5-1

0

1

2

3

SAR1 g

0.5 1 1.5 2 2.5-1

0

1

2

3

SAR10 g

0.5 1 1.5 2 2.5-1

0

1

2

3

50

100

150

2

4

6

0.5

1

1.5

2W/Kg W/Kg

W/Kg

Figure 8. Comparison of SARpeak, SAR1 g and SAR10 g distributions in a linear scale for a pinimplant under the skin.


1 g values for the pin at different depths ofthe tissue phantom and for 900 and 1800 MHz.

Position of the implant SARmax1 g (W kg−1)



SARmaxpeak (W kg−1)refand frequency (MHz)

On the skin, 900 1.4 130Under the skin, 900 2.1 46In the muscle, 900 1.7 36On the skin, 1800 2.6 360Under the skin, 1800 1.2 11In the muscle, 1800 1.9 6

Table 8. The cases with the highest SARmax1 g and SARmax

10 g values, i.e. the worst cases. Note that theresults are for a peak input current of 100 mA. The values in the brackets are SAR values relativeto the reference case without an implant.

Implant, positionand frequency (MHz) SARmax

peak (W kg−1) SARmax1 g (W kg−1) SARmax

10 g (W kg−1)

No implant, reference, 900 4.6 (1.0) 2.9 (1.0) 1.8 (1.0)No implant, reference, 1800 6.3 (1.0) 3.7 (1.0) 2.1 (1.0)Standard pin in z-direction, 215 (46) 6.2 (2.1) 2.0 (1.1)(14, 0, 10), under the skin, 900Standard pin in z-direction, 2280 (360) 9.4 (2.5) 3.7 (1.8)(8, 0, 10), on the skin, 1800Standard ring in yz-plane, 28 (6.1) 7.8 (2.7) 3.7 (2.0)(18, 0, 10), in the muscle, 9004 mm thick ring in yz-plane, 20 (4.4) 9.6 (3.3) 4.2 (2.3)(18, 0, 10), in the muscle, 9002 mm thick ring in yz-plane, 4420 (700) 8.4 (2.3) 2.5 (1.2)(8, 0, 10), on the skin, 18002 mm open ring in yz-plane 3590 (570) 8.6 (2.3) 2.5 (1.2)(8, 0, 10), on the skin, 1800

phone and the implant is high, significant surface currents are induced on the implant surface,and they produce a secondary EM field which is absorbed by the tissues (Hocking et al 1991).Although in the present study the enhancement in local SAR values due to an implant wasquite high (<400–700), the possible increase in averaged absorption and thus heating wasmuch smaller, i.e. maximum enhancement was by a factor of 2–3 for SAR10 g and SAR1 g

values. Generally, the results agree well with the simulations of other studies (Fleming et al1992, Cooper and Hombach 1996, Hocking et al 1991, McIntosh et al 2005).


In this study, the resonance effect in absorption was strongest with the 14 mm long pinin parallel with the antenna. This is very close to one-third of the wavelength λt/3 in muscle(44 mm at 900 MHz). Similar results were observed for the rings, i.e. absorption was highestwhen the diameter of the ring was 1/3–1/2 of the wavelength. These results agree well withthe resonance lengths of λt/3–λt/2 obtained in other similar studies too (Cooper and Hombach1996, Joyner et al 1998, Fleming et al 1999).

High local SAR values are not always obtained for the same implants, surrounding tissuesand positions as high average SAR values. Compared to these small implants, the averagingvolume is quite large, e.g. 10 g of muscle tissues corresponds to a cube of about 2 cm × 2 cm× 2 cm. In our simulations, the voxel size is smallest (0.2 mm × 0.2 mm × 0.2 mm) close tothe implant and corresponds to a mass of about 0.01 g. Although the highest values for SARdistribution were obtained for an implant on the skin, the maximum SAR1 g and SAR10 g valueswere obtained when the implant was under the skin. This is due to the fact that on the skin,the averaging volume can not be obtained close to the implant, while deeper in the tissues theimplant is completely surrounded by tissues. Without an implant the highest absorption oftenoccurs on the skin and the absorption decreases as the distance from the antenna increases andthe field attenuates. When an implant is in appropriate position and alignment, the incomingEM field and reradiated field by the implant interfere constructively leading to high-energyabsorption deeper in tissues adjacent to the implant.

The absorption naturally depends on the type of tissue too. As the electrical parametersof skin and muscle are quite close, there is just a minor change in the distribution at theskin–muscle interface. At the skin–bone and skin–fat interface the electromagnetic wave ispartially reflected. In general, the highest SAR values were obtained either in skin or muscle,which is evidently due to the fact that they have significantly higher conductivity than bone orfat. Furthermore, skin is close to the source where the incoming fields are generally strongest.

In addition to the effective coupling, high local fields in the tissues can be caused by thegeometry of an implant. For example, in the case of implants which are thin or have sharpedges, the EM field lines are focused on an edge or an end, leading to a high local flux. Inlossy tissues, this leads to significant increase in the absorption in comparison to the same casewithout an implant. In this study, the local SAR values increased by a factor of about 4 as thethickness of the pin decreased from 2 mm to 0.5 mm, which agrees well with the increase offlux due to geometry. It can be expected that the local SAR increases even further as the pingets thinner, although very thin structures are difficult to simulate with these programs due tolimitations in the size of voxels. However, the average SAR values may even decrease as inthe case of the pins in this study. The thinner structures do not block the incident EM waveand allow easier penetration of the field deeper in the tissues than larger implants.

In this study, the highest SAR1 g values were observed when the implant was under theskin or in the muscle. One of the reasons for this is the fact (also shown in figure 1) that fora resonant implant the interference between incident and reradiated fields is effective close tothe implant and tissues are fully surrounding the implant. However, in real clinical cases theseresonance effects are rare, because the worst cases occur for the implants with an appropriatesize and orientation and close to a mobile phone. Furthermore, with the current mobilephones the power is typically much less than the maximum allowed and in some modelsthe user is more effectively protected by EM shielding of the phone than in the simple genericphone of this study. Thermal input, e.g. in the cases of table 8, would indicate a dramaticand physiologically significant temperature rise if it were not for the small volume, whenconduction and convection distribute the heat effectively in surrounding tissues. Furthermore,guidelines involve time averaging, e.g. over a period of 6 min, too. So the presence of passivemetallic implants should seldom be a risk for a mobile phone user. However, tissue heating


near metallic medical implants may be a potential risk with high power RF equipment such asthose used in pulsed radio frequency diathermy (Ruggera et al 2003).

In the future, experimental validation of theoretical predictions based on these simulationsand on a finite element solution of the heat equation for the worst cases will be carried outby measuring the RF-induced temperature rise in a tissue phantom. Thermal measurementsusing fibre optic thermometry or a high-resolution IR camera with proper precautions (IEEE2002) can provide relevant data of temperature increase and SAR even in small volumes closeto metallic implants where the conventional SAR probes would be too large.

Acknowledgments

Financial support from the Nokia Research Center for MSc thesis work (HV) and the HermoResearch grant by National Technology Agency of Finland and companies are acknowledged.

References

Cooper J and Hombach V 1996 Increase in specific absorption rate in human heads arising from implantations Electr.Lett. 32 2217–9

Fleming A H J, Anderson V and Rowley J 1999 Computational estimate of the frequency response of metallic implantsin biological tissues exposed to RF fields Conference Proceedings of the 1999 ACES 15th Annual Review ofProgress in Applied Computational Electromagnetics vol 1 pp 638–43

Fleming A H and Joyner K H 1992 Estimates of absorption of radiofrequency radiation by the embryo and the fetusduring pregnancy Health Phys. 63 149–59

Fleming A H J, Lubinas V and Joyner K H 1992 Calculation of electric field in tissue near metallic implants Conf.Proc. of 1992 Asia-Pacific Microwave Conference (Adelaide: IEEE) pp 229–32

Gabriel S, Lau R W and Gabriel C 1996 The dielectric properties of biological tissues: II. Measurements in thefrequency range 10 Hz to 20 GHz Phys. Med. Biol. 41 2251–69

Hocking B, Joyner K H and Fleming A H J 1991 Implanted medical devices in workers exposed to radio-frequencyradiation Scand. J. Work Environ. Health 17 1–6

Institute of Electrical and Electronics Engineers (IEEE) 1999 IEEE Standard for Safety Levels with Respect to HumanExposure to Radiofrequency Electromagnetic Fields, 3 kHz to 300 GHz (New York: IEEE) C95.1-1999

Institute of Electrical and Electronics Engineers (IEEE) 2002 IEEE Recommended Practice for Measurements andComputations of Radio Frequency Electromagnetic Fields with Respect to Human Exposure to Such Fields,100 kHz–300 GHz (New York: IEEE) C95.3-2002

International Commission on Non-Ionizing Radiation Protection (ICNIRP) 1998a Guidelines for limiting exposureto time-varying electric, magnetic, and electromagnetic fields (up to 300 GHz) Health Phys. 74 494–522

International Commission on Non-Ionizing Radiation Protection (ICNIRP) 1998b Response to questions andcomments on ICNIRP Health Phys. 75 438

Joyner K H, Hocking B, Fleming A H J and Macfarlane I P 1998 Metallic implants and exposure to radiofrequencyradiation Conference Proceedings of the 7th International Congress of the International Radiation ProtectionAssociation (10–17 April) pp 477–80

McIntosh R L, Anderson V and McKenzie R J 2005 Assessment of RF workers with metallic implantsBioelectromagnetics submitted

Repacholi M H 2001 Health risks from the use of mobile phones Toxicol. Lett. 120 323–31Riu P J and Foster K R 1999 Heating of tissue by near-field exposure to a dipole IEEE Trans. Biomed. Eng. 46 911–7Ruggera P S, Witters D M, von Maltzahn G and Bassen H I 2003 In vitro assessment of tissue heating near metallic

medical implants by exposure to pulsed radio frequency diathermy Phys. Med. Biol. 48 2919–28Troulis S E, Scanlon W G and Evans N E 2003 Effect of a hands-free wire on specific absorption rate for a

waist-mounted 1.8 GHz cellular telephone handset Phys. Med. Biol. 48 1675–84Vanderstraeten J and van der Vorst A 2004 Theoretical evaluation of dielectric absorption of microwave energy at the

scale of nucleic acids Bioelectromagnetics 25 380–9

interaction of mobile phones with superﬁcial … › assets › downloads › publications ›...

Documents