Underestimation of EMF/NIR
Exposure for Children for Mobile
Telephones and for Electronic
Article Survellance(EAS) Systems
Om P. Gandhi
Department of Electrical & Computer Engineering
University of Utah
Salt Lake City, UT 84112 U.S.A.
Invited paper presented at the International NIR and Health
Workshop- Brazil, May 18,19,2009.
Some Pertinent Publications
1. O. P. Gandhi, G. Lazzi and C. M. Furse,
“Electromagnetic Absorption in the Human Head
and Neck for Mobile Phones at 835 & 1900 MHz,”
IEEE Trans on Microwave Theory & Techniques,
Vol. 44(10), pp. 1884-1897, 1996.
2. O. P. Gandhi and G. Kang, “Some Present
Problems and a Proposed Experimental Phantom
for SAR Compliance Testing of Cellular
Telephones at 835 and 1900 MHz,” Physics in
Medicine & Biology, Vol. 47, pp. 1501-1518, 2002.
3. O. P. Gandhi and G. Kang, “Inaccuracies of a
Plastic “Pinna” SAM for SAR Testing of Cellular
Telephones Against IEEE and ICNIRP Safety
Guidelines,” IEEE Trans. On Microwave Theory &
Techniques, Vol. 52(8), pp. 2004-2012, 2004.
Pertinent Publications - Continued
4. O. P. Gandhi, “Electromagnetic Fields: Human
Safety Issues,” Annual Review of Biomedical
Engineering, Vol. 4, pp. 211-234, 2002.
5. O. P. Gandhi and G. Kang,” “Calculation of
Induced Current Densities for Humans by
Magnetic Fields from Electronic Article
Surveillance Devices,” Physics in Medicine &
Biology, Vol. 46, pp. 2759-2771, 2001.
6. G. Kang and O. P. Gandhi, “SARs for PocketMounted Mobile Telephones at 835 and 1900
MHz,” Physics in Medicine & Biology, Vol. 47,
pp. 4301-4313, 2002.
In order to study the effect of head size and
pinna thickness on the absorption of electromagnetic energy radiated by cell phones, we
have used:
• Two different anatomic models of the head
(The Utah Model and the Visible Man Model),
and
• +11.1% larger and -9.1% smaller versions of
the above two models i.e. a total of six
anatomical models.
• The smaller versions of the head models
were used to correspond to the smaller
heads of the adults as well as the children.
Continued ….
• Different thicknesses of pinnas of the head
models (20, 14, 10, and 6 mm) to
correspond to the various individuals as
well as thinner pinnas of children.
• Different sizes/tilt angles of cell phone
antennas and handsets (2 types of
antennas, three sizes of handsets).
• Two frequencies: 1900 and 835 MHz.
Gandhi & Kang, Phys. Med. Biol., 47, 1501-18, 2002.
Gandhi & Kang, IEEE Trans. MTT, 52(8), 2004-12, 2004.
(a) The Utah Model
(b) The "Visible Man" model
Fig. 1. A visualization of the two anatomically-based
30º-tilted head models used for SAR calculations.
Gandhi & Kang, Phys. Med. Biol., 47, 1501-18, 2002.
Table 1. The calculated peak 1-g SARs for two models of the
human head for an irradiated power of 125 mW at 1900 MHz.
Gandhi & Kang, Physics in Med & Biol., 47, 1501-18, 2002.
Calculated
peak 1-g SARs
for two
the two
head models
at 835 MHz
Table 2.
The calculated
peak
1-gmodels
SARsoffor
of the
human head for an irradiated power of 600 mW at 835 MHz.
Gandhi & Kang, Physics in Med & Biol., 47, 1501-18, 2002.
Note that the peak 1-g SARs for both the
body tissue and the brain increase
monotonically with the reducing head
size (and pinna thickness) for both of the
head models (Utah and “Visible Man”),
all handset dimensions and the antennas
i.e. monopoles as well as helices.
Fig. 2. SAR distribution of Utah head model at 1900 MHz.
(a) 11.1% larger, (b) average, and (c) smaller.
Gandhi & Kang, Phys. Med. Biol., 47, 1501-18, 2002.
Fig. 3. SAR distribution of Utah head model at 835 MHz.
(a) 11.1% larger, (b) average, and (c) 9.09 % smaller.
Gandhi & Kang, Phys. Med. Biol., 47, 1501-18, 2002.
These figures calculated for the human
heads of various sizes are consistent with
the results reported for head models of
adults and children (Gandhi et al., IEEE
MTT, 44, 1884-97, 1996) in that there is a
deeper penetration of absorbed energy for
the smaller heads compared to that for the
larger heads, both for 1900 and 835 MHz
of radiated fields.
Effect of the pinna thickness on the brain and body
tissue SAR is given in the following table:
Table 3. The calculated peak 1-g body tissue and brain SARs
for the Utah Model of the head. Assumed is a handset of
dimensions 22 x 42 x 122 mm held at an angle of 30º relative
to the head.
Frequency
(MHz)
(irradiated
power in mW)
1900 (125 mW)
835 (600 mW)
Pinna thickness (mm)
Antenna
length (mm)
Tissue
40
Body tissue
Brain
0.51
0.20
0.83
0.27
1.02
0.33
1.29
0.46
80
Body tissue
Brain
2.44
0.85
2.95
1.06
3.20
1.24
3.50
1.37
20
14
10
6
Peak 1-g SAR
Gandhi & Kang, Physics in Med & Biol., 47, 1501-18, 2002.
As expected, both the body tissue and
brain SARs increase monotonically with
the reducing pinna thickness (e.g. for the
children). This is due to the closer
placement of the radiating antenna to the
body tissues and to the brain.
We have also reported [Gandhi & Kang, IEEE Trans.
MTT, 52, 2004-12, 2004] that use of a plastic “pinna”
for the specific anthropomorphic mannequin (SAM)
head model used for SAR compliance testing of cell
phones underestimates both the peak 1-g SAR as
well as the 10-g SAR required for ICNIRP Guidelines
by a factor of 1.6-2.0 or more, even for adults.
We have also reported that use of the so-called
“Visible Man” Model based on the CT scans of a
fairly husky (105 kg) man’s cadaver tends to
underestimate both the peak 1- and 10-g SARs.
This problem is further compounded by the fact that
SARs are higher for children as compared to adults.
RE
LE
M
(a) Side view.
(b) Cut through reference
plane R passing through
mouth M.
(c) A cut 30 mm
below plane R.
Fig. 4. SAM head model with three cross-sectional cuts
defining the 5-10 mm thickness of the plastic shell.
(Source: IEEE Std., 1528, 2005)
Gandhi, IEEE Trans. on Microwave Theory & Techniques,
52(8), 2004.
Table 4. Comparison of peak 1- and 10-g SARs obtained for SAM and
Anatomic Models for the “cheek” and “15º-tilted” positions of the 22 x 42 x
122 mm handsets with different antennas. The SARs are normalized to a
radiated power of 600 mW at 835 MHz.
Cheek
Position
15-Tilted
position
Peak 1-g SAR (W/kg)
Antenna
Axial
Length mm
SAM 5-10 mm
plastic pinna
Utah Model,
6 mm-thick pinna
Utah Model,
14 mm-thick pinna
Visible Man,
6 mm-thick pinna
80 mm
monopole
3.30
10.82
9.58
6.43
20 mm helix
4.18
14.40
14.51
9.07
80 mm
monopole
2.49
10.90
8.80
7.12
20 mm helix
2.65
14.96
12.96
9.95
Peak 10-g SAR (W/kg)
Cheek
Position
15-Tilted
position
80 mm
monopole
2.36
3.67
4.55
2.34
20 mm helix
3.03
4.79
6.61
3.21
80 mm
monopole
1.69
3.83
4.46
3.17
20 mm helix
1.77
5.06
5.94
4.15
Gandhi, IEEE Trans. on Microwave Theory & Techniques, 52(8), 2004-12, 2004.
Table 5. Comparison of peak 1- and 10-g SARs obtained for SAM and
anatomic models for the “cheek” and “15º-tilted” positions of the 22 x 42 x
122 mm handsets with different antennas. The SARs are normalized to a
radiated power of 125 mW at 1900 MHz.
Cheek
Position
15-Tilted
position
Peak 1-g SAR (W/kg)
Antenna
Axial
Length mm
SAM 5-10 mm
plastic pinna
Utah Model,
6 mm-thick pinna
Utah Model,
14 mm-thick pinna
Visible Man,
6 mm-thick pinna
80 mm
monopole
0.73
2.05
2.31
1.97
20 mm helix
0.93
2.24
2.54
2.66
80 mm
monopole
1.13
3.06
3.00
3.06
20 mm helix
1.30
3.52
3.45
3.96
Peak 10-g SAR (W/kg)
Cheek
Position
15-Tilted
position
80 mm
monopole
0.49
0.85
1.04
0.90
20 mm helix
0.59
0.94
1.16
1.02
80 mm
monopole
0.74
1.16
1.33
1.36
20 mm helix
0.82
1.34
1.52
1.61
Gandhi, IEEE Trans. on Microwave Theory & Techniques, 52(8), 2004-12, 2004.
10-g SAR, “Visible Man” Model, cheek position, frequency = 1900 MHz
Radiated power = 125 mW
Fig. 5. Variation of peak 10-g SAR as a function of separation from
the absorptive tissues. Handset of dimensions 22 x 42 x 122 mm.
10-g SAR, Utah Model, 15º-tilted position, frequency = 1900 MHz
Radiated power = 125 mW
Fig. 6. Variation of peak 10-g SAR as a function of separation from
the absorptive tissues. Handset of dimensions 22 x 42 x 122 mm.
1-g SAR, Utah Model, 15º-tilted position, frequency = 835 MHz
Radiated power = 600 mW
Fig. 7. Variation of peak 1-g SAR as a function of separation from
the absorptive tissues. Handset of dimensions 22 x 42 x 122 mm.
The underestimation of SAR by SAM is due to the
fact that the cell phone under test is physically
separated from the tissue-simulant head model of
SAM by several millimeters.
It has been repeatedly shown both computationally
and experimentally* that each additional millimeter of
physical separation of the radiating source from the
tissue-simulant model results in an underestimation
of SAR by 13-15%. Thus, a factor of 2 or more
underestimation of SAR by the SAM SAR compliance model is understandable because of the 610 mm thickness of the plastic “pinna” used for SAM.
*Kang and Gandhi, Phys. Med. Biol., 47, 4301-13, 2002.
Gandhi & Kang, IEEE-MTT, 52, 2004-12, 2004.
Electronic Article Surveillance
(EAS) Systems
• Being introduced into stores, libraries, and hospitals
to prevent theft of items.
• Use alternating magnetic fields (at frequencies of
several kHz to several MHz.
• May take the form of one- or two-sided panels of
current-carrying loops near the exit door, loops hidden
in the floor or under the checkout counters.
• We have used the impedance method to calculate
induced current densities for 1mm resolution
anatomical models of adult and scaled models of 10and 5-year old children.
Gandhi & Kang, Phys. Med. Biology, 46, 2759-71, 2001.
Fig. 8. A few
representative EAS
systems.
Table 6. Some typical external dimensions and derived voxel
sizes used for the anatomic models of the male adult and 10and 5-year old boys.
Adult Male
10-Year Old Boy
5-Year Old Boy
Head circumference 1(cm)
53.8
52.3
51.0
Body weight 1 (kg)
71.7
30.5
18.9
Height 1 (cm)
176
138
112
17,400
9610
7510
1305
1048
984
2
Body surface area1 (cm )
2
Head surface area 1 (cm )
Voxel size for head (mm)
1.974  1.974  2.930 1.974  1.974  2.352 1.871  1.871  2.330
Voxel size for arms,
torso, and legs (mm)
1.974  1.974  2.930 1.400  1.400  2.282 1.196  1.196  1.782
1
C. Lentner, “Geigy Scientific Tables”.
Gandhi & Kang, Phys. Med. Biol., 47, 2759-71, 2001.
(a) Adult male
(b) 10-year old boy
(c) 5-year old boy
Fig. 9. The three anatomic models used for calculations of
induced electric fields and current densities.
Gandhi & Kang, Phys. Med. Biol., 47, 2759-71, 2001.
From: Gandhi & Kang, Phys.
Med. Biol., 46, 2759-71, 2001.
Fig. 10. Top view of the schematic of an assumed magnetic
deactivator coil and the placement of the human model relative
to it. All dimensions are in cm.
From: Gandhi & Kang, Phys.
Med. Biol., 46, 2759-71, 2001.
Fig. 11. An assumed EAS system using a
pair of rectangular coils with an overlap of
10 cm. The lower rung of the bottom coil
is assumed to be 20 cm off the ground
plane. The marked dimensions are in cm.
From: Gandhi &
Kang, Phys. Med.
Biol., 46, 2759-71,
2001.
Along the y-axis from y = 40 cm (frontal plane of the body) to y = 70 cm.
Fig. 12. The calculated variations of the magnetic fields with distance
y from the center of the deactivator solenoid and for the vertical z-axis
passing through the front of the human model (y = 40 cm) at the edge
of the table.
From: Gandhi &
Kang, Phys. Med.
Biol., 46, 2759-71,
2001.
Along the vertical z-axis passing through the front of the model.
Fig. 13. The calculated variations of the magnetic fields with distance
y from the center of the deactivator solenoid and for the vertical z-axis
passing through the front of the human model (y = 40 cm) at the edge
of the table.
From: Gandhi &
Kang, Phys. Med.
Biol., 46, 2759-71,
2001.
Along line xmxm passing through the center of the coils of the EAS system.
Fig. 14. The calculated variations of the magnetic fields with distance
y along line x m xm passing through the center of the EAS coils and
for a vertical line 11 at a distance x = 20 cm (y = 0) from the plane of
the EAS panel.
From: Gandhi &
Kang, Phys. Med.
Biol., 46, 2759-71,
2001.
11 at a distance x = 20 cm (y = 0) from the plane
Fig. 15. The calculated variations of the magnetic fields with distance
y along line x m xm passing through the center of the EAS coils and
for a vertical line 11 at a distance x = 20 cm (y = 0) from the plane of
the EAS panel.
Along the vertical line
Assumed for the calculation of induced current
densities for the various parts of the body is:
Frequency (F) of 1 kHz, 50,000 A turns rms for
the 85 cm high table top deactivator, and
Frequency of 30 kHz, 100 A turns for a second
panel type EAS system.
ICNIRP Basic Restriction for max 1 cm2 areaaveraged current density J for CNS tissues (brain
and spinal cord)
J max  2mA / m 2
at F  1 kHz
 60 mA / m 2 at F  30 kHz
Table 7. The calculated organ-averaged and maximum 1
cm2 area-averaged current densities (J) for the CNS
tissues for the models of the adult and 10- and 5-year
old children for the 1 kHz magnetic deactivator.
Organ
Adult
10-year old
child
5-year old
child
J(mA m-2)
J(mA m-2)
J(mA m-2)
Brain
Organ-averaged
Maximum (1 cm-2)
0.15
0.48
0.66
2.02
1.77
4.46
Pineal
gland
Organ-averaged
Maximum (1 cm-2)
0.02
---
0.10
---
0.51
---
Gandhi & Kang, Phys. Med. Biol., 47, 2759-71, 2001.
Table 8. The calculated organ-averaged and maximum 1
cm2 area-averaged current density (J) for the CNS
tissues for the models of the adult and 10- and 5-year
old children for the 30 kHz EAS pass-by system.
Organ
Adult
10-year old
child
5-year old
child
J(mA m-2)
J(mA m-2)
J(mA m-2)
Brain
Organ-averaged
Maximum (1 cm-2)
4.75
17.63
23.20
64.64
40.70
98.93
Pineal
gland
Organ-averaged
Maximum (1 cm-2)
0.92
---
17.42
---
36.27
---
Gandhi & Kang, Phys. Med. Biol., 47, 2759-71, 2001.
The point to note is that higher current densities
are induced for the CNS tissues of the 10- and
5-year old children for both of the assumed EAS
systems.
This is due to the fact that the head of the taller
adult is considerably above the deactivator or the
pass-by EAS panel and is thus in the weaker
magnetic field region.
The heads of the shorter children, on the other
hand, are in higher magnetic fields. The
maximum induced J for the children may exceed
ICNIRP basic restrictions if sufficiently strong
magnetic fields are used.
Conclusions – for Mobile
Telephones
• Use of six different anatomical models (two
different head shapes and three different scaled
versions i.e. average, larger, and smaller
versions of each, shows that the peak 1-g SAR
for the brain for the smaller models
representative of children may be up to 220% at
1900 MHz and 144% at 835 MHz of the SARs of
the larger models.
• This is due to the thinner pinna and the skull for
the smaller models which results in closer
placement of the mobile telephones to the brain
of children.
Conclusions -- Continued
• Use of the SAM (“standard anthropomorphic
mannequin”) model with a 5-10 mm thick plastic
spacer in the shape of “pinna” chosen by
industry for SAR compliance testing results in an
artificial, more distant placement of the mobile
telephone from the tissue-simulating fluid of
SAM. This gives an SAR that is up to two or
more times smaller than for the anatomic models
of the adult head, and an even larger
underestimation of the SAR for the heads of
children.
Conclusions -- Continued
• In Europe, compliance of the maximum
magnetic fields induced current densities for the
CNS tissues (brain and the spinal cord) against
ICNIRP guidelines is required for all EAS
systems.
• Because of the larger height, the adult head is
generally in the weaker magnetic field region
resulting in lower induced current densities for
the brain for adults.
• The heads of children, on the other hand, are in
the stronger magnetic field regions resulting in
higher induced currents for the brain as
compared to adults.