268 IEEE TRANSACTIONS ON ELECTROMAGNETIC COMPATIBILITY, VOL. 46, NO. 2, MAY 2004
Effect of Dielectric Properties on the Peak 1- and
10-g SAR for 802.11 a/b/g Frequencies
2.45 and 5.15 to 5.85 GHz
Gang Kang, Senior Member, IEEE, and Om P. Gandhi, Life Fellow, IEEE
Abstract—Compliance with 1- or 10-g specific absorption rate
(SAR) safety guidelines is required in various countries for all
newly-introduced personal wireless devices such as Wi-Fi PCs.
Even though the dielectric properties of the human tissues are
known to be nonuniform and highly variable, relatively rigid ad-
herence to prescribed dielectric properties is required for
compliance testing of such devices. Using some typical near-field
radiators, we have examined the effect of dielectric properties for
SAR measurement fluids with conductivities varying by 2 : 1 to
show that both 1- and 10-g SARs vary by less than 2 –4 for
the 802.11a band 5.15 to 5.825 GHz and only slightly more at the
lower 802.11 b/g frequency of 2.45 GHz. This is due to higher sur-
face SAR but shallower depth of penetration of EM fields for the
higher conductivity media resulting in nearly identical SARs for
cubical volumes associated with 1- or 10-g of tissue, respectively.
Also studied, is the effect of lower fluids recommended in some
standards which results in slightly higher and, thus, a conservative
assessment of SAR.
Index Terms—2 : 1 variation of conductivity, compliance testing,
near constancy of specific absorption rate (SAR), wireless PCs.
I. INTRODUCTION
C
OMPLIANCE with the safety guidelines such as those
proposed by IEEE [1] and ICNIRP [2], etc., is required by
regulatory agencies in the United States and elsewhere for all
newly-introduced personal wireless devices such as Wi-Fi PCs
and cellular telephones, etc. These safety guidelines are set in
terms of maximum mass-normalized rates of electromagnetic
energy deposition (specific absorption rates or SARs) for any
1- or 10-g of tissue. The two most commonly-used SAR limits
today are those of IEEE [1], 1.6 W/kg for any 1 g of tissue, and
ICNIRP, [2] 2 W/kg for any 10 g of tissue, excluding extremi-
ties such as hands, wrists, feet, and ankles where higher SARs
up to 4 W/kg for any 10 g of tissue are permitted in both of
these standards. Experimental and numerical techniques using
planar or head-shaped phantoms have been proposed for de-
termining compliance with the SAR limits [3]–[5]. Dielectric
properties (dielectric constant and conductivity ) for the
tissue-simulant fluids have been prescribed in these standards
based on the properties measured for the various tissues for hu-
mans and other mammals [6] and equivalency with properties
needed for a homogeneous planar model to properly represent
Manuscript received June 15, 2003; revised October 7, 2003.
The authors are with the Electrical and Computer Engineering Department,
University of Utah, Salt Lake City, UT 84112 USA (e-mail: gkang@ieee.org).
Digital Object Identifier 10.1109/TEMC.2004.826875
absorption of incident planewaves to that for a skin-fat-muscle-
skull-sclera-CSF-brain layered planar model of a human [7].
In reality, the dielectric properties of the human tissues are
highly nonuniform and likely variable with age as well [8].
Thus, dielectric constants and conductivities reported for the
various tissues are highly variable and may vary by factors of
2 : 1 or more for some of the tissues [9]. This paper focuses on
the effect of the dielectric properties of the SAR measurement
fluids on the peak 1- and 10-g SARs for a planar phantom
used for compliance testing of wireless PCs and other body- or
torso-mounted devices that typically operate at frequencies of
2.4–2.484 GHz (802.11 b/g systems) and in the 5-GHz band for
frequencies of 5.15 to 5.35 and 5.745 to 5.825 GHz (802.11a
systems). For a 2 : 1 or 100% variability in conductivity of the
tissue-simulant fluid, the variation in peak 1- and 10-g SAR
is less than 2–4 for the higher frequency band 5.15 to
5.825 GHz, and only slightly higher for the lower 802.11 b/g
wireless systems band of 2.45 GHz. The reason for this is
the higher surface SAR but shallower depth of penetration
of electromagnetic fields for the higher conductivity media
which has the net effect of providing nearly identical SAR for
volumes in the shape of a cube, of dimensions 1 or 2.154 cm
associated with 1- or 10-g of tissue, respectively. Though
relatively negligible at 2.45 GHz, it is shown that the effect of
the changing dielectric constant (instead of conductivity) is
somewhat larger on both 1- and 10-g SARs for the 5–6 GHz
band with the lower phantom materials resulting in 10%
to 12% higher SARs. Thus the lower media recommended
by IEC PT62209 [5] for the 5–6 GHz band may be used if a
conservative determination of SAR is of interest.
II. ASSUMED EM SOURCES
The sources of microwave radiation typically used in wire-
less PCs are one or two dual band microstrip antennas either
fabricated on an insertable wireless card or built into the base
or at times behind the display screen of the PC. A wide va-
riety of antenna dimensions and locations are typically used.
For the purpose of calculations of peak 1- or 10-g SARs, we
have assumed a typical microstrip antenna (see Fig. 1) and a
couple of additional sources of EM radiation that are recom-
mended in the compliance standards [3]–[5] for SAR measure-
ment system validation. The three EM sources thus selected are:
a square microstrip antenna of dimensions; 30 30 mm placed
with a spacing of 4-mm above a ground plane of dimensions
0018-9375/04$20.00 © 2004 IEEE