RF & ELF Mobile Phone Exposure Science Brunch, Zurich, May 2012
The Beginning
RF & ELF Mobile Phone Exposure - • early 80s - Motorola: developed first in-house SAR scanner Review and Newest Findings • late 80s - IT’IS/ETH & Motorola: published the absorption mechanism in the near-field of Marie-Christine Gosselin transmitter Sven Kühn Mark Douglas Andreas Christ Niels Kuster
1 2
RF & ELF Mobile Phone Exposure Science Brunch, Zurich, May 2012 RF & ELF Mobile Phone Exposure Science Brunch, Zurich, May 2012
The Beginning Today
• early 80s • each phone is certified to be compliant with the RF safety guidelines - Motorola: developed first in-house SAR scanner • the maximum exposure (spSAR) is provided in the user manual • late 80s • - IT’IS/ETH & Motorola: published the absorption mechanism for the near-field of lower values lead to a lower maximum exposure in the real world transmitters • technology to assess the average real-world exposure of CNS and • early 90s other tissues is ready - lawsuit: no knowledge about the phone exposure (except Motorola) • each phone is intrinsically compliant with the ELF restrictions - German Agency for Radiation Protection: requested phone certification - IT’IS/ETH: received a contract to develop a prototype of certification system & • main unresolved details: procedures by German Ministry of Telecom, D-Telecom, Mannesmann, Swiss-PTT - technical issues regarding measurement of latest technologies - hand effects on SAR - measurement distance for on-body testing
3 4 RF & ELF Mobile Phone Exposure Science Brunch, Zurich, May 2012 RF & ELF Mobile Phone Exposure Science Brunch, Zurich, May 2012
Basic Standard (>10 MHz)
ICNIRP FCC IEEE 2006 India Radio-Frequency (RF) Fields from Mobile Phones W/kg W/kg W/kg W/kg whole-body 0.08 0.08 0.08 0.08 SAR limit
peak spatial 2 1.6 2 1 SAR limit 10g contiguous 1g cube 10g cube 10g cube
• whole-body limit: only relevant for vascular diseases • spatial peak limit: relevant for local heating and potentially “athermal” effects
5 6
RF & ELF Mobile Phone Exposure Science Brunch, Zurich, May 2012 RF & ELF Mobile Phone Exposure Science Brunch, Zurich, May 2012
Dependence of Averaging Volume Problem
Source: Environment: Exposed body:
10 g cube P(t, ) G(qf ) E(t,x,y,z )inc SAR(t,x,y,z )ind 1g cube Z()t H(t,x,y,z ) T(t,x,y,z ) 1g contiguous inc ind
SAR in dB
Variables: Source power Distance Anatomy Frequency Coupling Mech. Posture 2450 MHz 5500 MHz Modulation +Scatterers +Tissue Parameters Antenna Char. Re-radiators
7 8 RF & ELF Mobile Phone Exposure Science Brunch, Zurich, May 2012 RF & ELF Mobile Phone Exposure Science Brunch, Zurich, May 2012
Goal for Exposure Assessment (Compliance) Research Effort of IT’IS/ETH
- establishment of absorption mechanism (Kuster et al., 92) • the measured quantity must be conservative - development of novel probes (various publications) ‣ >90th percentile of exposed population, i.e., all age groups - development of 1st dosimetric scanner DASY1/2 (Schmid et al., 96) • simplified but not greatly overestimated exposure - dependence on inner anatomy (Hombach 96, Meier 96, Drossos 01, Christ 08) ‣ i.e., w/o inhibiting technological progress - dependence on outer anatomy (Meier 96, Schoenborn 98, Christ 05,Kuehn 09) - dependence on the modeling of ear on the psSAR (Burkhardt 00, Christ 09) • low exposure in real life = low exposure in the test and vice versa - enhancements due to metallic implants (Thesis Meier 96, Kyriakou, 11) ‣ favors low exposure devices - development of phantom and tissue materials (MCL, SPEAG for IEEE1528) • field distribution is greatly non-homogeneous - dependence on the hand on head exposure (Meier 95, Li, 11, Li 12) - design rules for optimal OTA & minimal SAR (Tay et al., 98) ‣ large gradients in 3D, multiple maximum - calibration procedures (thesis Pokovic, 99, Kühn 09) • mobile specific - uncertainty assessment procedures & budget (thesis Pokovic, 99) - highest absorption close to the surface - procedure for body-worn devices (Christ et al. 06, Kühn et al. 09) ‣ measurements closest to the surface - validation of SAM head (Beard et al. 05, Christ et al. 06) - complex modulations - modulation dependent calibrations (Kühn et al. 11) ‣ very large peak-to-average ratios ‣ majority of relevant references in the standard generated by IT’IS/ETH
9 10
RF & ELF Mobile Phone Exposure Science Brunch, Zurich, May 2012 RF & ELF Mobile Phone Exposure Science Brunch, Zurich, May 2012
1. Interaction Mechanism 1. Consequences of Interaction Mechanism
H2 ~ j2/d2
2 2 ‣ SAR = RFlosses ~ j /d ‣ exposure is not directly related to the radiated power! ‣ strongly design dependent
11 12 RF & ELF Mobile Phone Exposure Science Brunch, Zurich, May 2012 RF & ELF Mobile Phone Exposure Science Brunch, Zurich, May 2012
Research Effort of IT’IS/ETH Historical Note
- establishment of absorption mechanism (Kuster et al., 92) - development of novel probes (various publications) - development of 1st dosimetric scanner DASY1/2 (Schmid et al., 96) - dependence on inner anatomy (Hombach 96, Meier 96, Drossos 01, Christ 08) - dependence on outer anatomy (Meier 96, Schoenborn 98, Christ 05,Kuehn 09) - dependence on the modeling of ear on the psSAR (Burkhardt 00, Christ 09) - enhancements due to metallic implants (Thesis Meier 96, Kyriakou, 11) - development of phantom and tissue materials (MCL, SPEAG for IEEE1528) - dependence on the hand on head exposure (Meier 95, Li, 11, Li 12) - design rules for optimal OTA & minimal SAR (Tay et al., 98) - calibration procedures (thesis Pokovic, 99, Kühn 09) - uncertainty assessment procedures & budget (thesis Pokovic, 99) - procedure for body-worn devices (Christ et al. 06, Kühn et al. 09) - validation of SAM head (Beard et al. 05, Christ et al. 06) - modulation dependent calibrations (Kühn et al. 11) ‣ majority of relevant references in the standard generated by IT’IS/ETH
13 14
RF & ELF Mobile Phone Exposure Science Brunch, Zurich, May 2012 RF & ELF Mobile Phone Exposure Science Brunch, Zurich, May 2012
Smallest Isotropic Probes Research Effort of IT’IS/ETH
- establishment of absorption mechanism (Kuster et al., 92) - development of novel probes (various publications) - development of 1st dosimetric scanner DASY1/2 (Schmid et al., 96) - dependence on outer anatomy (Meier 96, Schoenborn 98, Christ 05,Kuehn 09) - dependence on inner anatomy (Hombach 96, Meier 96, Drossos 01, Christ 08) - dependence on the modeling of ear on the psSAR (Burkhardt 00, Christ 09) - enhancements due to metallic implants (Thesis Meier 96, Kyriakou, 11) - development of phantom and tissue materials (MCL, SPEAG for IEEE1528) - dependence on the hand on head exposure (Meier 95, Li, 11, Li 12) - design rules for optimal OTA & minimal SAR (Tay et al., 98) - calibration procedures (thesis Pokovic, 99, Kühn 09) - uncertainty assessment procedures & budget (thesis Pokovic, 99) - procedure for body-worn devices (Christ et al. 06, Kühn et al. 09) - validation of SAM head (Beard et al. 05, Christ et al. 06) - modulation dependent calibrations (Kühn et al. 11) ‣ majority of relevant references in the standard generated by IT’IS/ETH
15 16 RF & ELF Mobile Phone Exposure Science Brunch, Zurich, May 2012 RF & ELF Mobile Phone Exposure Science Brunch, Zurich, May 2012
2. Phantom Shape + 3. Liquid + 4. Hand Hand Issue (Open)
• phone parts - close as possible ‣ large head (90th percentile US army) ‣ touch and tilt
%/100) • liquid representing worst-case tissue composition ‣ liquid parameters derived from layered tissue model considering all tissue %/100 compositions • hand ‣ testing w/o hands since studies in the late 90s showed only reduction ‣ currently under reconsideration
17 18
RF & ELF Mobile Phone Exposure Science Brunch, Zurich, May 2012 RF & ELF Mobile Phone Exposure Science Brunch, Zurich, May 2012
Virtual Population Poser
19 20 RF & ELF Mobile Phone Exposure Science Brunch, Zurich, May 2012 RF & ELF Mobile Phone Exposure Science Brunch, Zurich, May 2012
Anatomical Characteristics Physics’ Based Morphing
Female"BMI" 0.25"
0.2" Virtual"Popula=on"Models:" B""Ella"""(26"yrs)"""=""22.0""kg/m2" B""Billie"(11"yrs)"""="16.5"kg/m2" " 0.15"
India"(n"="204877)"
Probabilty" 0.1" US"(n"="4313)" World"(n"=174611)"
0.05"
0" 0" 10" 20" 30" 40" 50" 60" 70" 80" 90" 2 BMI"(kg/m )"
21 22
RF & ELF Mobile Phone Exposure Science Brunch, Zurich, May 2012 RF & ELF Mobile Phone Exposure Science Brunch, Zurich, May 2012
Research Effort of IT’IS/ETH Implants - A Regulatory Gap
- establishment of absorption mechanism (Kuster et al., 92) - development of novel probes (various publications) - development of 1st dosimetric scanner DASY1/2 (Schmid et al., 96) - dependence on outer anatomy (Meier 96, Schoenborn 98, Christ 05,Kuehn 09) - dependence on inner anatomy (Hombach 96, Meier 96, Drossos 01, Christ 08) - dependence on the modeling of ear on the psSAR (Burkhardt 00, Christ 09) - enhancements due to metallic implants (Thesis Meier 96, Kyriakou, 11) - development of phantom and tissue materials (MCL, SPEAG for IEEE1528) - dependence on the hand on head exposure (Meier 95, Li, 11, Li 12) - design rules for optimal OTA & minimal SAR (Tay et al., 98) - calibration procedures (thesis Pokovic, 99, Kühn 09) - uncertainty assessment procedures & budget (thesis Pokovic, 99) - procedure for body-worn devices (Christ et al. 06, Kühn et al. 09) - validation of SAM head (Beard et al. 05, Christ et al. 06) - modulation dependent calibrations (Kühn et al. 11) ‣ majority of relevant references in the standard generated by IT’IS/ETH
23 24 RF & ELF Mobile Phone Exposure Science Brunch, Zurich, May 2012 RF & ELF Mobile Phone Exposure Science Brunch, Zurich, May 2012
Evaluation at Basic Restriction Limit Research Effort of IT’IS/ETH
- establishment of absorption mechanism (Kuster et al., 92) - development of novel probes (various publications) - development of 1st dosimetric scanner DASY1/2 (Schmid et al., 96) - dependence on outer anatomy (Meier 96, Schoenborn 98, Christ 05,Kuehn 09) - dependence on inner anatomy (Hombach 96, Meier 96, Drossos 01, Christ 08) - dependence on the modeling of ear on the psSAR (Burkhardt 00, Christ 09) - enhancements due to metallic implants (Thesis Meier 96, Kyriakou, 11) - development of phantom and tissue materials (MCL, SPEAG for IEEE1528) - dependence on the hand on head exposure (Meier 95, Li, 11, Li 12) - design rules for optimal OTA & minimal SAR (Tay et al., 98) - calibration procedures (thesis Pokovic, 99, Kühn 09) - uncertainty assessment procedures & budget (thesis Pokovic, 99) - procedure for body-worn devices (Christ et al. 06, Kühn et al. 09) - validation of SAM head (Beard et al. 05, Christ et al. 06) - modulation dependent calibrations (Kühn et al. 11) ‣ majority of relevant references in the standard generated by IT’IS/ETH
25 26
RF & ELF Mobile Phone Exposure Science Brunch, Zurich, May 2012 RF & ELF Mobile Phone Exposure Science Brunch, Zurich, May 2012
Latest Scanning System (DASY5 NEO) Fast SAR Scanners
27 28 RF & ELF Mobile Phone Exposure Science Brunch, Zurich, May 2012 RF & ELF Mobile Phone Exposure Science Brunch, Zurich, May 2012
Spatial Peak SAR Values (System) Spatial Peak SAR (t)
29 30
RF & ELF Mobile Phone Exposure Science Brunch, Zurich, May 2012 RF & ELF Mobile Phone Exposure Science Brunch, Zurich, May 2012
Brain Exposure as a Function of Anatomy Brain Exposure (Phone Design)
31 32 Final Report SNF 405740-113591 25 ABCD%H7,IE3//%FG7.3% 45%/+-67.% >?@4%;-/3%/+-67.% )*%+,-./0123,% 89*"'$$%07;1E3%FG7.3% <*49%07;1E3%FG7.3% >?@4%H7,IE3//%FG7.3% 89*:$$%;-/3%/+-67.% 89*"'$$%;-/3%/+-67.% <*49%":=$%;-/3%/+-67.% ABCD%;-/3%/+-67.% ABCD%;-/3%/+-67.% 89*:$$%07;1E3%FG7.3% $%
!&$% 430F7,-E%E7;3% @7,+3J% !#$% 4G-E-0K/%
!($%
!"#$%&' !'$% RF & ELF Mobile Phone Exposure Science Brunch, Zurich, May 2012 RF & ELF Mobile Phone Exposure Science Brunch, Zurich, May 2012 !"$$% Estimation of Avg Brain Exposure of Phone X Various!"&$% Brain Regions Exposure by Sources !"#$%
18.1: heavy mobile/cordless phone user • Normal mobile/cordless phone user ABCD%H7,IE3//%FG7.3% 45%/+-67.% )*%+,-./0123,% >?@4%;-/3%/+-67.% 89*:$$%;-/3%/+-67.% 89*"'$$%;-/3%/+-67.% <*49%":=$%;-/3%/+-67.% ABCD%;-/3%/+-67.% ABCD%;-/3%/+-67.% 89*:$$%07;1E3%FG7.3% 89*"'$$%07;1E3%FG7.3% <*49%07;1E3%FG7.3% >?@4%H7,IE3//%FG7.3% $%
!&$% 430F7,-E%E7;3% @7,+3J% !#$% 4G-E-0K/%
!($%
!"#$%&' !'$%
!"$$%
+ !"&$% OTA Performance !"#$% 18.2: normal mobile/cordless phone user
33 34 ABCD%H7,IE3//%FG7.3% 45%/+-67.% )*%+,-./0123,% >?@4%;-/3%/+-67.% 89*:$$%07;1E3%FG7.3% RF & ELF Mobile Phone Exposure Science Brunch, Zurich, May 2012 RF & ELF Mobile Phone Exposure89*:$$%;-/3%/+-67.% 89*"'$$%;-/3%/+-67.% <*49%":=$%;-/3%/+-67.% ABCD%;-/3%/+-67.% ABCD%;-/3%/+-67.% 89*"'$$%07;1E3%FG7.3% Science<*49%07;1E3%FG7.3% >?@4%H7,IE3//%FG7.3% Brunch, Zurich, May 2012 $%
!&$% 430F7,-E%E7;3% @7,+3J% Mobile Phone Systems GSM:!#$% Time-Domain 4G-E-0K/%
!($%
Bands System max Prms avg Prms maxpsSAR avg psSAR remarks !"#$%&' !'$% (MHz) (mW) (mW) (W/kg) (W/kg) !"$$% NATEL/ 450/900 FM/FFSK 1000 ~500 1 - 4 0.25 - 2 NMT (1G) !"&$%
GSM/ 900/1800 TDMA, 250/125 ~50-100 0.1 - 2 0.03 - 0.8 !"#$% EDGE (2G) FDMA / GMSK 18.3: non mobile/cordless phone user UMTS/ 1950 CDMA / 125 ~1-5 0.1 - 2 0.001 - 0.04 HSPA (3G) QPSK Figure 18 – Distribution of the SAR in various brain regions for exposure to di erent sources of RF EMF with weighting representing di erent types of wireless users. LTE (3.9G) 2600? SC-FMDA / <200 n.a. 0.1 - 2 n.a avg. P in QPSK, network still 16QAM unknown WiFi 2450/ CSMA-CA / 100, 1000 usage dep: 0.05 - 1 0.0005 - 5200-5800 DxPSK, (DFS, TPC) <1% 0.001 xQAM Bluetooth 2450 FHSS / 100/2.5/1 usage dep: 0.001 - 0.5 0.00001 - GFSK, <0.1% 0.005 xDPSK
35 36 RF & ELF Mobile Phone Exposure Science Brunch, Zurich, May 2012 RF & ELF Mobile Phone Exposure Science Brunch, Zurich, May 2012
GSM Time Domain Signals UMTS Time Domain Signals
GSM Handset Time Domain Signal GSM Handset Time Domain Signal (DTX) ZH city, MotorolaV1050, orange(dual), run: 1, ID: 1 Length: 3408.2469m
0
−10 ZH city, MotorolaV1050, orange(gsm), run: 1, ID: 4 Length: 3835.7921m ! ! −20 0 power / dB
−30 −10
−40 −20 0 200 400 600 800 1000 1200
power / dB t/s −30 GSM: Power Control 4 −40 0 200 400 600 800 1000 1200 t/s 3
4 2 band 37 38 3 1
2 RF & ELFband Mobile Phone Exposure Science Brunch, Zurich, May 2012 RF & ELF0 Mobile Phone Exposure Science Brunch, Zurich, May 2012 0 200 400 600 800 1000 1200 t/s 1 LTE: Time-Domain WiFi: Time-Domain 0 0 200 400 600 800 1000 1200 t/s
39 40 RF & ELF Mobile Phone Exposure Science Brunch, Zurich, May 2012 RF & ELF Mobile Phone Exposure Science Brunch, Zurich, May 2012
Communication Systems: Frequency Domain
Guidelines for limiting exposure to time-varying electric, magnetic, and electromagnetic fields ● ICNIRP GUIDELINES 509 0 In the frequency range from a few Hz to 1 kHz, for In the low-frequency range, there are currently few levels of induced current density above 100 mA mϪ2, the data relating transient currents to health effects. The -10 thresholdsLow-Frequency for acute changes in central (LF) nervous Fields system fromICNIRP Mobile therefore recommendsPhones that the restrictions on excitability and other acute effects such as reversal of the current densities induced by transient or very short-term -20 visually evoked potential are exceeded. In view of the peak fields be regarded as instantaneous values which safety considerations• FSM Project above, it was decided that, for should not be time-averaged. frequencies in the range 4 Hz to 1 kHz, occupational The basic restrictions for current densities, whole- -30 exposure should be limited to fields that induce current body average SAR, and localized SAR for frequencies Ϫ2 P dB densities less than 10 mA m , i.e., to use a safety factor between 1 Hz and 10 GHz are presented in Table 4, and -40 of 10. For the general public an additional factor of 5 is those for power densities for frequencies of 10–300 GHz applied, giving a basic exposure restriction of 2 mA mϪ2. are presented in Table 5. Below 4 Hz and above 1 kHz, the basic restriction on -50 induced current density increases progressively, corre- REFERENCE LEVELS sponding to the increase in the threshold for nerve -60 stimulation for these frequency ranges. Where appropriate, the reference levels are obtained Established biological and health effects in the from the basic restrictions by mathematical modeling and frequency range from 10 MHz to a few GHz are by extrapolation from the results of laboratory investiga-
-15 -10 -5 frequency/MHz 5 10 15 consistent with responses to a body temperature rise of tions at specific frequencies. They are given for the condi- more than 1°C. This level of temperature increase results tion of maximum coupling of the field to the exposed from exposure of individuals under moderate environ- individual, thereby providing maximum protection. Tables mental conditions to a whole-body SAR of approxi- 6and7summarizethereferencelevelsforoccupational mately 4 W kgϪ1 for about 30 min. A whole-body exposure and exposure of the general public, respectively, 41 42 average SAR of 0.4 W kgϪ1 has therefore been chosen as and the reference levels are illustrated in Figs. 1 and 2. The the restriction that provides adequate protection for reference levels are intended to be spatially averaged values occupational exposure. An additional safety factor of 5 is over the entire body of the exposed individual, but with the introduced for exposure of the public, giving an average important proviso that the basic restrictions on localized RF & ELF Mobile Phone Exposure Science Brunch, Zurich, May 2012 whole-body RF & ELF SAR Mobile limit Phone of 0.08 Exposure W kg Ϫ 1 . exposure are not exceeded. Science Brunch, Zurich, May 2012 The lower basic restrictions for exposure of the For low-frequency fields, several computational and general public take into account the fact that their age and measurement methods have been developed for deriving Absorption Mechanism (ELF) health statusICNIRP may differ from Basic those of workers.Restrictionsfield-strength (1998) reference levels from the basic restrictions.
• coupling with LF magnetic fields Table 4. Basic restrictions for time varying electric and magnetic fields for frequencies up to 10 GHz.a Current density for Whole-body Localized SAR - induced electric fields by eddy currents Exposure head and trunk average SAR (head and trunk) Localized SAR 38 Final Report - Assessment of ELFcharacteristics Current Frequency range Distributions(mA mϪ2) (rms) (W from kgϪ1) UMTS(W kgϪ1) and(limbs) (W GSM kgϪ1) Mobile Phones • exposure limited to prevent: Occupational up to 1 Hz 40 — — — exposure 1–4 Hz 40/f —— — - nerve and muscle stimulation 4 Hz–1 kHz 10 — — — 1–100 kHz f/100 — — — - retinal phosphenes 100 kHz–10 MHz f/100 0.4 10 20 10 MHz–10 GHz — 0.4 10 20 General public up to 1 Hz 8 — — — exposure 1–4 Hz 8/f —— — basic restrictions in terms of current density, J4 Hz–1. kHz The average2 current— density— over— a cross-section A of 2 1–100 kHz f/500 — — — 1 cm perpendicular to the direction of the current100 kHz–10 MHzdensityf/500 is calculated0.08 at2 each voxel4 r0: 10 MHz–10 GHz — 0.08 2 4 a Note: 1. f is the frequency in hertz. 2. Because of electrical inhomogeneity1 of the body, current densities should be averaged over a cross-section of 1 cm2 perpendicular to the current direction. 2 3.J For( frequenciesr0) A up to= 100 kHz, peak currentJ density(r values) canˆn beda. obtained by multiplyingA= the1cm rms value by ߛ2(ϳ1.414). For pulses (4) ϭ h of duration tpithe equivalentA frequencyA to apply in the· basic restrictions should be calculated as f 1/(2tp). 4. For frequencies up to 100 kHz and for pulsed magnetic fields, the maximum current density associated with the pulses can be calculated from the rise/fall times andZ the maximum rate of change of magnetic flux density. The induced current density can then be compared with the appropriate basic restriction. 5. All SAR values are to be averaged over any 6-min period. 43 6. Localized SAR averaging mass is any 10 g of contiguous tissue; the maximum SAR so obtained should be the value used for the 44 ICNIRP 2010 More recently, the ICNIRPestimation of exposure. has published guidelines dedicated to fields at frequencies
7. For pulses of duration tp the equivalent frequency to apply in the basic restrictions should be calculated as f ϭ 1/(2tp). Additionally, for pulsed exposures in the frequency range 0.3 to 10 GHz and for localized exposure of the head, in order to limit or avoid auditory between 1 Hz and 100 kHz [2] where theeffects caused limits by thermoelastic are expansion, now an additional defined basic restriction in is recommended. terms This is of that the induced SA should not exceed electric field, E, averaged over a small contiguous tissue10 volumemJ kgϪ1 for workerVs and 2ofmJ kg 2Ϫ1 for the 2general public2 mm, averaged2ove:r 10gtissue. ⇥ ⇥ 1 E(r ) = E(r) dv. (5) h 0 iV V ZV Numerically, the averaging is performed at each voxel via a summation of the contributions of n sur- rounding voxels of volume Vn, leading to 1 E(r ) E(r ) V , (6) h 0 iV ⇡ V n n n n X
where 0 n 1 is a filling factor that accounts for the partial contribution/volume of the outermost voxels of the averaging cube.
IEEE C95.6 In its standard limiting the exposure to LF EMF [?], the Institute of Electrical and Electronics Engineers (IEEE) defines basic restrictions in terms of electric field averaged over a straight line segment of 5 mm length, L,i.e.
ˆl E(r ) = 0 E(r) ˆl dl, (7) h 0 iL L · 0 ZL
where ˆl0 is the direction of the electric field in one particular voxel. The line segment is centered at each voxel and the basic restrictions are dependent on the surrounding tissue as well as on the frequency.
6.2 Development of the Numerical Source The measurement results presented in Sections 104(b) and 104(b) were used as the target envelop for the numerical exposure. Various loop configurations have been investigated (see Appendix D), separately for the fields associated to the communication network (Figure 104(b)) and the audio signal (Figure 104(b)). The envelope as well as the decay in the direction perpendicular to the handset (fields parallel and perpendicular to the phone surface treated individually) have been used as criterium to choose the most representative numerical source. For the signal due to the configuration network, the loop was simulated using a frequency of 217 Hz, most important component of the GSM technology (which was shown to lead to field XXX orders of magnitude higher than UMTS). The chosen loop (Figure 36) is composed of 7 concentric rectangular loops fed with the same current. The 2 smallest loops, closer to each other that the rest, were added to reach a more homogeneous field distribution of the x- and y-components (parallel to the plane of the loops). The H-field at a distance of 15 mm from the plane of the loops is shown in Figure 37. The field distribution is relatively uniform over the surface of the loops. Additionally, the fields (x-, y- and z-components) were compared to the corresponding field components of the measurements. Figure 38 shows the decay of the perpendicular and parallel components of the field with the distance to the handset. The best approximation was found for an o↵set of the simulation of 15 mm and a scaling factor of XX. As was stated in Section 104(b), the distance between the tip of the probe and the center of the sensor is 3 mm. The measurement data are taken 1 mm from the phone surface. Consequently, the loop was considered to be 11 mm inside the phone for the numerical simulations with the anatomical heads. A similar procedure was used to develop the numerical loop representing the exposure to the fields generated at the audio speaker. XXXX... 38 Final Report - Assessment of ELF Current Distributions from UMTS and GSM Mobile Phones Limiting exposure to time-varying electric and magnetic fields ● ICNIRP 825 reference levels are derived from relevant basic restric- account for the uncertainties described above. Such tions using measurement and/or computational tech- restrictions rise above 3 kHz. niques but some address perception (electric field) and For the general public for CNS tissue of the head a adverse indirect effects of exposure to EMF. The derived reduction factor of 5 is applied, giving a basic restriction quantities are electric field strength (E), magnetic field of 10 mV mbasicϪ1 between restrictions 10 and 25 Hz. Above and in below terms of current density, J. The average current density over a cross-section A of strength (H), magnetic flux density (B) and currents these values, the basic2 restrictions rise. At 1,000 Hz it flowing through the limbs (IL). The quantity that ad- intersects with1 cm basic restrictionsperpendicular that protect against to the direction of the current density is calculated at each voxel r0: dresses indirect effects is the contact current (IC). In any peripheral and central myelinated nerve stimulation. particular exposure situation, measured or calculated Here, the reduction factor of 10 results in a basic values of any of these quantities can be compared with restriction of 400 mV mϪ1, which should be applied to 1 the appropriate reference level. Compliance with the the tissues of all parts of the body. J(r0) A = J(r) ˆn da. (4) reference level will ensure compliance with the relevant The basic restrictions are presented in Table 2 and h i A A · basic restriction. If the measured or calculated value Fig. 1. Z exceeds the reference level, it does not necessarily follow that the basic restriction will be exceeded. However, Time averaging 38 Final Report - Assessmentwhenever of a ELF reference Current level is exceeded Distributions it is necessary to fromICNIRP UMTSICNIRP recommends and that GSM the 2010 restrictions Mobile onMore internal Phones recently, the ICNIRP has published guidelines dedicated to fields at frequencies test compliance with the relevant basic restriction and to electric fields induced by electric or magnetic fields includ- determine whether additional protective measures are ing transientbetween or very short-term 1 peak Hz fields and be regarded 100 as kHz [2] where the limits are now defined in terms of induced electric field, E, necessary. instantaneousaveraged values which should over not be a time small averaged contiguous tissue volume V of 2 2 2 mm2: (see also section on non-sinusoidal exposure). ⇥ ⇥ BASIC RESTRICTIONS basic restrictions in terms of current density, J. The average currentSpatial density averaging of over induced a electric cross-section field A of The main objective of this publication is to establish 1 1 cm2 perpendicular to the direction of the current density is calculatedWhen at restricting each adverse voxel effectsr : of induced electric guidelines for limiting EMF exposure that will provide fields to nerve cells and networks,0 it is important to E(r0) V = E(r) dv. (5) protection against adverse health effects. As noted above, define the distance or volume over which the local h i V V the risks come from transient nervous1 system responses induced electric field must be averaged. As a practical Z including peripheralJ(r ) (PNS)= and central nerveJ(r stimulation) ˆn da. compromise, satisfying requirements for a sound biolog- (4) (CNS), the induction0 ofA retinal phosphenes and possible h i A A · ical basis andNumerically, computational constraints, the ICNIRP averaging recom- is performed at each voxel via a summation of the contributions of n sur- effects on some aspects of brain function.Z mends determining the induced electric field as a vector In view of the considerations above for frequencies average of therounding electric field in voxels a small contiguous of volume tissue Vn, leading to RF & ELF Mobile Phone Exposure Science Brunch, Zurich, May 2012 RF & ELF Mobile Phone Exposure Science Brunch, Zurich, May 2012 in the range 10 Hz to 25 Hz, occupational exposure volume of 2 ϫ 2 ϫ 2mm3. For a specific tissue, the 99th ICNIRP 2010 More recently,should the be ICNIRP limited to fields has that published induce electric guidelines field percentile dedicated value of the electric to fields field is the at relevant frequencies value strengths in CNS tissue of the head (i.e., the brain and to be compared with the basic restriction. 1 between 1 Hz and 100 kHz [2] whereretina) of lessthe than limits 50 mV m areϪ1 innow order to defined avoid the in terms of induced electric field, E, IEEE ICNIRP Basic Restrictions (2010) HUMAN EXPOSUREEIEEE(r0 )TO Basic ELECTROMAGNETICV Restrictions FIELDS,E( 0–3rn kHz) n V(C95.6,n, 2002) Std C95.6-2002 (6) averaged over a small contiguousinduction tissue of volume retinal phosphenes.V of 2 These2 restrictions2 mm2: h i ⇡ V should also prevent any possible transient effects on n ⇥ ⇥ Table 2. Basic restrictions for human exposure to time-varying Table 1—Basic restrictions applying to various regions of the bodya, b brain function. These effects are not considered to be electric and magnetic fields. X adverse health effects; however, ICNIRP recognizes that 1 Internal electric field they may be disturbing in some occupational circum- Ϫ1 Controlled E(r0) V = E(r) dv. Exposure characteristicwhere Frequency 0 range n (V1 m is) a filling(5) factor that accounts for the partialGeneral contribution/volume public of the outermost f environment stances andh should be avoidedi butV no additionalV reduc- Occupational exposure Exposed tissue e (Hz) tion factor is applied. Phosphene thresholdsZ rise rapidly CNS tissuevoxels of the head of 1− the10 Hz averaging0.5/f cube. at higher and lower frequencies, intersecting with the 10 Hz−25 Hz 0.05 E0 - rms (V/m) E0 - rms (V/m) 25 Hz−400 Hz 2 ϫ 10Ϫ3f Numerically, the averaging is performedthresholds for peripheral at each and central voxel myelinated via3 a nerve summation of the contributions400 Hz−3 kHz 0.8 of n sur- Brain 20 5.89 × 10–3 1.77 × 10–2 stimulation atV= 400 Hz.2x2x2mm At frequencies above 400 Hz, 3 kHz−10 MHz 2.7 ϫ 10Ϫ4f rounding voxels of volume Vn, leading to All tissues of head and 1 Hz−3 kHz 0.8 Heart 167 0.943 0.943 limits on peripheral nerve stimulation apply in all parts of Ϫ4 body IEEE3 C95.6kHz−10 MHz 2.7Inϫ 10 itsf standard limiting the exposure to LF EMF [?], the Institute of Electrical and the body. General public exposure Hands, wrists, feet and ankles 3350 2.10 2.10 Exposure in controlled environments, where work- CNS tissue of the head 1−10 Hz 0.1/f 1 Electronics10 Hz−25 Hz Engineers0.01 (IEEE) definesOther basic tissue restrictions in terms3350 of 0.701 electric field2.10 averaged over a straight ers are informedE(r about) the possible transientE(r effects) ofV , 25 Hz−1000 Hz 4 ϫ 10Ϫ4f (6) 0 V n n n a such exposure, should be limited to fields that induce line segment1000 Hz−3 kHz of 50.4 mm length, L,i.e. Interpretation of table is as follows: Ei = E0 for f ≤ fe; Ei = E0 (f / fe) for f ≥ fe. h i ⇡ V Ϫ4 b electric fields in the head and body ofn less than 800 mV 3 kHz−10 MHz 1.35 ϫ 10 f In addition to the listed restrictions, exposure of the head and torso to magnetic fields below 10 Hz shall All tissues of head and 1 Hz−3 kHz 0.4 Ϫ1 X be restricted to a peak value of 167 mT for the general public, and 500 mT in the controlled environment. m in order to avoid peripheral and central myelinated body 3 kHz−10 MHz 1.35 ϫ 10Ϫ4f nerve stimulation. A reduction factor of 5 has been Note: In the frequency range above 100 kHz, basic restrictions preventing ˆ where 0 n 1 is a filling factor that accounts for theϪ1 partial contribution/volume of the outermost l0 applied to a stimulation threshold of 4 V m in order to the adverse effects of RF hearting also need to be considered. 5.2 MaximumE(r permissible) = exposureE (MPE)(r) values:ˆl dl, Magnetic L=flux density 5mm (7) voxels of the averaging cube. h 0 iL L · 0 5.2.1 Exposure of the head and torsoZL to sinusoidal fields IEEE C95.6 In its standard limiting the exposure to LF EMF [?], the Institute of Electrical and Table 2 lists maximum permissible magnetic field limits (flux density, B, and magnetic field strength, H) for where ˆl0 is the direction of the electricexposure of field the head and in torso. one The particularaveraging time for an voxel.rms measure is The 0.2 seconds line for frequencies segment above is centered at each Electronics Engineers (IEEE) defines basic restrictions in terms of electric field averaged over a straight45 25 Hz. For lower frequencies, the averaging time is such that at least 5 cycles are included in the average, but 46 voxel and the basic restrictions arewith dependent a maximum of 10 seconds. on the surrounding tissue as well as on the frequency. line segment of 5 mm length, L,i.e. RF & ELF Mobile Phone Exposure Science Brunch, Zurich, May 2012 RF & ELF MobileTable Phone2—Magnetic Exposure maximum permissible exposure (MPE) levels: Science Brunch, Zurich, May 2012 ˆ exposure of head and torsoa, b l0 6.2 Development of the Numerical Source E(r0) L = E(r) ˆl0 dl, (7) LF Fields from Mobile Phones MeasurementGeneral of public Incident FieldsControlled environment h i L L · Frequency range Z (Hz) The measurement results presented in Sections 104(b)B - rms and H 104(b)- rms wereB - rms used asH - the rms target envelop for the where ˆl is the direction of the electric• fieldpower in of one digital particular circuits voxel. The line segment is centered at each • 10 mobile phones(mT) (A/m) (mT) (A/m) 0 numerical exposure. Various loop configurations have been investigated4 (see Appendix D), separately for • 6.2 Development of the Numerical Source perpendicular to the phone surface treated individually) have5 been used as criterium6 to choose the most • 759–3000audio OFF/ON (1kHz)687/f 5.47 × 10 /f 2060/f 1.64 × 10 /f a representative numerical source. f is frequency• DASY52 in Hz. NEO The measurement results presented in Sections 104(b) and 104(b) were used as the target envelop for the bMPEs refer to spatial maximum. numerical exposure. Various loop configurations have been investigated (see AppendixFor the signal D), separately due to for the configuration• T-coil uniaxial network, probe the loop was simulated using a frequency of 217 Hz, the fields associated to the communication network (Figure 104(b)) and themost audio important signal (Figure component 104(b)). of theCompliance GSM• time-domain with technology Table via 2 ensures python compliance (which with the was basic restrictions shown of Table to 1. lead However, to lack of field XXX orders of complianceimplementation with Table 2 does not necessarily imply lack of compliance with the basic restrictions, but rather The envelope as well as the decay in the direction perpendicular to themagnitude handset (fields higher parallel than and UMTS). Thethat it may chosen be necessary to loop evaluate whether (Figure the basic 36)restrictions is have composed been met. If the basic of restrictions 7 concentric in rectangular perpendicular to the phone surface treated individually) have been used as criterium to choose the most Table• 1 areprobe not exceeded, tip 1mm then from the phone MPE values in Table 2 can be exceeded. Consequently, it is sufficient to loops fed with the same current.demonstrate Thesurface 2 compliance smallest (=4mm with from either sensor loops, Table 1 or Table closer 2. to each other that the rest, were added representative numerical source. to reach a more homogeneous field distributioncenter) of the x- and y-components (parallel to the plane of the For the signal due to the configuration network, the loop was simulatedloops). using a The frequency H-field of at217 a Hz, distance of 15 mm from the plane of the loops is shown in Figure 37. The field most important component of the GSM technology (which was shown todistribution lead to field is XXX relatively orders uniformof over the surface of the loops. magnitude higher than UMTS). The chosen loop (Figure 36) is composed of 7 concentric rectangular47 Copyright © 2002 IEEE. All rights reserved. 11 48 Additionally, the fields (x-, y- and z-components) were compared to the corresponding field components loops fed with the same current. The 2 smallest loops, closer to each other that the rest, were added Authorized licensed use limited to: ETH BIBLIOTHEK ZURICH. Downloaded on August 30,2011 at 07:45:08 UTC from IEEE Xplore. Restrictions apply. to reach a more homogeneous field distribution of the x- and y-componentsof (parallel the measurements. to the plane of Figure the 38 shows the decay of the perpendicular and parallel components of the loops). The H-field at a distance of 15 mm from the plane of the loops isfield shown with in Figure the distance 37. The fieldto the handset. The best approximation was found for an o↵set of the simulation distribution is relatively uniform over the surface of the loops. of 15 mm and a scaling factor of XX. As was stated in Section 104(b), the distance between the tip Additionally, the fields (x-, y- and z-components) were compared to the correspondingof the probe field and components the center of the sensor is 3 mm. The measurement data are taken 1 mm from the of the measurements. Figure 38 shows the decay of the perpendicular andphone parallel surface. components Consequently, of the the loop was considered to be 11 mm inside the phone for the numerical field with the distance to the handset. The best approximation was foundsimulations for an o↵set of with the thesimulation anatomical heads. of 15 mm and a scaling factor of XX. As was stated in Section 104(b), theA distance similar between procedure the was tip used to develop the numerical loop representing the exposure to the fields of the probe and the center of the sensor is 3 mm. The measurement data are taken 1 mm from the phone surface. Consequently, the loop was considered to be 11 mm insidegenerated the phone at for the the audio numerical speaker. XXXX... simulations with the anatomical heads. A similar procedure was used to develop the numerical loop representing the exposure to the fields generated at the audio speaker. XXXX... Final Report - Assessment of ELF Current Distributions from UMTS and GSM Mobile Phones 23 5.2.7 Measurement Procedure For each DUT, a number of configurations were measured. The measurement planes were defined as parallel to the surface and at a distance of 1 mm from the probe tip at the highest phone location (since the phone surfaces are irregular). As the sensor is located 3 mm from the bottom part of the probe tip, the actual measurements data are located 4 mm away from the highest point of the phone surface. For the area scans, the resolution used was 5 7 mm2 for the entire area. This was done at 2 di↵erent PCLs such that we can distinguish the frequency⇥ components due to the communication system from the rest. A higher-resolution scan (3 3 mm2) was performed around the audio output of the phone when the audio signal was turned ON. Additionally,⇥ the B-field decay from the phone surface was assessed by performing z-scans (1D scan perpendicular to the phone surface) at one location on the front size of the phone. All the following configurations have been measured for each DUT (unless specified otherwise) in the GSM900, GSM1800, and UMTS bands: maximum PCL, audio ON: area scan, front, back, speaker; • PCL -6 dB, audio OFF: area scan, front1; • 3 PCLs, audio OFF: z-scan, front, one location; • 3 PCLs, audio ON: 4 DUTs, area scan, speaker; • 3 PCLs, audio ON: 4 DUTs, z-scan, speaker, one location. • For all measurements, the microphone was turned o↵ to prevent the noises in the lab from generating additional signal at the speaker. The action process during the measurement of each configuration is illustrated in Figure 20. 5.2.8 Devices Under Test Table RF 7& ELF shows Mobile the Phone list and Exposure main characteristics of the ten mobile phones, Science or devices Brunch, under Zurich, testMay 2012 (DUTs), RF & ELF Mobile Phone Exposure Science Brunch, Zurich, May 2012 measured in this study2. Pictures of the phones can be found in Appendix B. Measurement of Incident Fields - 10 DUTs Measurement of Incident Fields - 10 DUTs Table 7: List of DUTs and their main characteristics. ID Phone Model Type OS Release Date Nokia6120 Nokia 6120 bar April 2007 SonyEricssonW910 Sony Ericsson W910i slide Oct 2007 SonyEricssonW760i Sony Ericssion W760i slide May 2008 MotorolaV1050 Motorola V1050 flip January 2005 HTCdiam100 HTC Diam100 Touch Diamond smart Windows Phone May 2008 HTCtopa100 HTC Topa100 Touch Diamond2 smart Windows Phone April 2009 iPhone3g Apple iPhone 3g smart iOS July 2008 iPhone4 Apple iPhone 4 smart iOS June 2010 SamsungGT-I9001 Samsung Galaxy GT-I9001 smart Android June 2010 LG LG P920 Optimus 3D smart Android July 2011 1For the slide and flip phones, only the surface of the phone resulting in the highest fields was measured at this second power level, i.e. the bottom part, closer to the battery. 2As indicated in Section 5.2.7, every DUT has been measured in the GSM900, GSM1800, and UMTS band. Although the specifications of the MotorolaV1050 phone include UMTS, the connection in the band was very unstable and did not allow any complete measurement. Similarly for GSM900 and GSM1800 at lower PCLs. 49 50 RF & ELF Mobile Phone Exposure Science Brunch, Zurich, May 2012 RF & ELF Mobile Phone Exposure Science Brunch, Zurich, May 2012 Results - Frequency Spectrum Results - Spatial Distribution