Bioelectromagnetics 29:65^70 (2008)

Millimeter Wave Dosimetry of Human Skin

S.I. Alekseev, A.A. Radzievsky, M.K. Logani, and M.C. Ziskin*
Center for Biomedical Physics,Temple University Medical School,
Philadelphia, Pennsylvania

To identify the mechanisms of biological effects of mm waves it is important to develop accurate

methods for evaluating absorption and penetration depth of mm waves in the epidermis and dermis.
The main characteristics of mm wave skin dosimetry were calculated using a homogeneous unilayer
model and two multilayer models of skin. These characteristics included reflection, power density
(PD), penetration depth (d), and specific absorption rate (SAR). The parameters of the models were
found from fitting the models to the experimental data obtained from measurements of mm wave
reflection from human skin. The forearm and palm data were used to model the skin with thin and thick
stratum corneum (SC), respectively. The thin SC produced little influence on the interaction of mm
waves with skin. On the contrary, the thick SC in the palm played the role of a matching layer and
significantly reduced reflection. In addition, the palmar skin manifested a broad peak in reflection
within the 83–277 GHz range. The viable epidermis plus dermis, containing a large amount of free
water, greatly attenuated mm wave energy. Therefore, the deeper fat layer had little effect on the PD
and SAR profiles. We observed the appearance of a moderate SAR peak in the therapeutic frequency
range (42–62 GHz) within the skin at a depth of 0.3–0.4 mm. Millimeter waves penetrate into the
human skin deep enough (d ¼ 0.65 mm at 42 GHz) to affect most skin structures located in the
epidermis and dermis. Bioelectromagnetics 29:65–70, 2008.
2007 Wiley-Liss, Inc.

Key words: millimeter wave reflection; power density profile; penetration depth

INTRODUCTION non-uniform distribution of the local electromagnetic

field within the skin. It was suggested that the multilayer
Therapeutic application of millimeter waves in
organization of skin might result in a resonance type
medicine [Rojavin and Ziskin, 1998] has been used widely
absorption of mm waves [Balantsev et al., 1991].
in the Former Soviet Union. Strikingly high success rates
However, most data of mm wave reflection and
have been reported in the treatment of over fifty medical
absorption in the skin were obtained from theoretical
conditions including cardiovascular diseases, diabetes,
calculations based on a homogeneous skin model
dermatitis, gastrointestinal disorders, wound healing, pain
[Gandhi and Riazi, 1986].
relief, and cancer. Although the beneficial actions arise
Recently reflection of terahertz and mm wave
from induced systematic effects, the mm waves are totally
radiation was used to study electric properties and water
absorbed in the skin. The skin structures located within the
content of human skin [Pickwell et al., 2004; Alekseev
penetration depth of mm waves are potential targets
and Ziskin, 2007]. To fit the experimental reflection
[Betsky and Yaremenko, 1998]. This has stimulated great
data, Alekseev and Ziskin [2007] applied several
interest in better understanding the mechanisms of mm
multilayer skin models. The forearm and palm data
wave interaction with human skin. Accurate dosimetry is
were used to model the skin with thin and thick stratum
an important first step.
Dosimetry calculations require knowledge of ————— —
permittivity and geometry of skin layers. Permittivity Grant sponsor: NIH NCCAM; Grant number: P01-AT002025.
of biological tissue including skin in the gigahertz
*Correspondence to: M.C. Ziskin, Center for Biomedical Physics,
frequency range results mostly from the polarization of Temple University Medical School, 3400 North Broad Street,
free water molecules [Grant, 1982; Foster and Schwan, Philadelphia, PA 19140. E-mail: [email protected]
1986; Gabriel et al., 1996a]. Different skin layers
contain different amounts of water [Mashimo et al., Received for review 9 April 2007; Final revision received 16 July
1987; Warner et al., 1988; Imokawa et al., 1991; 2007
Richard et al., 1991, 1993; Caspers et al., 2001, 2003; DOI 10.1002/bem.20363
Mirrashed and Sharp, 2004a,b]. Hence, the skin has Published online 10 October 2007 in Wiley InterScience
heterogeneous electrical properties. This can cause a (www.interscience.wiley.com).

2007 Wiley-Liss, Inc.

66 Alekseev et al.

corneum (SC), respectively. The skin with thin SC was calculation of permittivity we used the Debye equation
well described even with the homogeneous skin model. with a single relaxation time t equal to that of pure water
Permittivity of the skin obtained using a homogeneous at a skin temperature of 32.6 8C [Schwan and Foster,
one-layer model was close to the corresponding skin 1977; Foster et al., 1980; Gandhi and Riazi, 1986]:
permittivity given by Gabriel et al. [1996b]. To fit
Dei si
reflection from the palmar skin with a thick SC, a skin e
i ¼ e1i þ þ ð1Þ
model containing at least two layers was required. 1 þ jot joeo
Multilayer models provided better fitting to both the
where o ¼ 2pf, f is the frequency, j ¼ (1)1/2,
forearm and palmar skin reflection data. The fitting
Dei ¼ esie1i is the magnitude of the dispersion of the
parameters obtained with different models included
free water fraction of the skin layer i, esi is the
permittivity and thickness of different skin layers.
permittivity at ot  1, e1i is the optical permittivity,
We did not find in the literature any detailed
eo ¼ 8.85  1012 F/m, si is the skin ionic conductivity.
analysis of the mm wave interaction with heteroge-
The magnitude of the dispersion Dei increases with the
neous skin, i.e., we could not find any calculations
concentration of free water in the skin layers and can
of mm wave intensity distributions within skin layers
be used for determination of the free water content.
based on their different electrical properties. Therefore,
The values of t, Dei, e1i, and si used in the models are
the aim of the present study was to calculate the power
presented in Table 1.
density and SAR profiles and the penetration depth
of mm waves in skin based on homogeneous and Calculation of mm Wave Power Deposition
multilayer skin models.
Millimeter waves were considered to be at normal
incidence to the skin surface. We denoted the refractive
METHODS indices and permittivity of skin layers as ni and ei,
respectively, where ni ¼ (ei)1/2, i ¼ 0, 1, 2, 3. For air,
Skin Models n0 ¼ 1; for skin layers, ni ¼ n0i  jn00i ¼ ðe0i  je00i Þ1=2 . To
In modeling mm wave interaction with skin, we calculate the power reflection coefficient (R) and power
used two types of skin: the forearm and palmar skin as density profiles PD(z) as a function of skin depth we
representative of skin with thin and thick SC, applied the Fresnel equations for multilayer systems
respectively. For both types of skin we applied a [Born and Wolf, 1975; Ryakovskaya and Shtemler,
homogeneous unilayer model and two multilayer skin 1983; Alekseev and Ziskin, 2000, 2001, 2007]. We used
models (Fig. 1). We made several assumptions. First, the same tissue mass density for all skin layers equal to
the skin layers used in models were considered to have 1000 kg/m3. In the homogeneous skin model d was
distinct plane boundaries. Second, the thicknesses and calculated as d ¼ c/on00 , where c is the velocity of light.
permittivity of skin layers were equal to those found The penetration depth for the heterogeneous skin was
from fitting skin models to reflection data [Alekseev and calculated as the distance into the skin where intensity
Ziskin, 2007]. It was shown that the permittivity of each decreased to e2 of its original value.
skin layer (i) could be described by the permittivity of
the fraction of free water contained in the layer. For
RESULTS
Skin Models With Thin SC
Using the permittivity parameters and thicknesses
of skin layers given in Table 1 for the forearm skin, we
calculated the frequency dependence of the power
reflection coefficient for the plane mm wave field
(Fig. 2). In our calculations of reflection we used the
entire mm wave frequency range. As shown recently,
water permittivity throughout the 0.2–410 GHz range
could be described by two relaxation times [Buchner
Fig. 1. Models used for calculation of mm wave reflection and et al., 1999]. However, there was only a small difference
powerdepositioninskin.Model1representedahomogeneousskin between the permittivity dispersion curves calculated
model. Model 2 included the stratum corneum (SC) (1) and under- using the Debye equation with either one or two
lying homogeneous skin tissue (2). Model 3 contained the SC (1),
the rest of epidermis plus dermis (2), and fat. The skin in the first relaxation times. Therefore, we assumed that the
model and the deepest layer in each of the other 2 models had parameters of the Debye equation found from modeling
semi-infinite thicknesses. in the relatively narrow frequency range (37–75 GHz)
Bioelectromagnetics
 Millimeter Wave Dosimetry of Human Skin 67
TABLE 1. Parameters of Skin Models Used in Calculation of the Power Reflection Coefficient, Power Deposition and Penetration
Depth

Forearm model Palm model

Parameter 1 2 3 2 3
SC e1 — 2.96 2.96 3.63 3.63
De — — — 9.7 9.5
d (mm) — 0.015 0.015 0.43 0.42
s (S/m) — 0 0 0 0
E þ D e1 4.0 4.0 4.0 4.52 4.52
De 32.0 32.6 32.4 27.2 26.4
d (mm) 1 1 1.45 1 1.85
s (S/m) 1.4 1.4 1.4 1.4 1.4
t  1012 (s) 6.9 6.9 6.9 6.9 6.9
The stratum corneum and the rest of epidermis plus dermis are denoted as SC and E þ D, respectively; d is the thickness and s is the ionic
conductivity of each skin layer. In homogeneous skin (model 1), E þ D stands for the whole skin. Other symbols used in the table are given
in Methods section.

[Alekseev and Ziskin, 2007] were also valid for the

entire 30–300 GHz range.
To test the influence of higher hydration of the SC on
reflection, we calculated the power reflection coefficients
at two free water concentrations in the SC, 0 and 0.24
weight fractions, using model 3. The higher hydration
levels of the SC may be achieved during application of
moisturizing lotions. As shown in Figure 2, the difference
in reflection was not significant especially in the thera-
peutic frequency range (42–62 GHz). The models 1 and 2
provided practically the same result.
The PD profiles of mm waves at two therapeutic
frequencies of 42 and 61 GHz in relation to the skin
layers are shown in Figure 3. For these calculations we
used model 3. The other models gave similar results. As
shown in Figure 3, the 61 GHz exposure provided a
greater PD in the epidermis while the 42 GHz exposure
was more intense in the dermis. The frequency depend-
ence of the SAR at different skin depths calculated for
the homogeneous skin model is illustrated in Figure 4. As
expected, the greatest SAR values occurred at the skin
surface. In general, the SAR increased with frequency.
However, at the highest frequencies the SAR in the
deeper regions of skin decreased because of the reduced
penetration depth at these frequencies. Thus, over the
therapeutic range of frequencies, the SAR value
gradually increased reaching a peak, and then decreased
as the frequency increased.
The penetration depth of the plane mm wave field
in the forearm skin at different frequencies is shown in Fig. 2. Frequency dependence of the power reflection coefficient
Figure 5. In order to take into account the thickness of for the forearm and palm calculated for the plane mm wave field
the SC, the calculations of d were made using both using model 3. Forearm: solid and dashed lines correspond to the
model 2 and model 3. Both gave similar results. Our weight fractions of free water in the SC of 0 and 0.24, respectively.
calculations yielded a slightly higher penetration depth Palm: Solid, dashed, and dash-dotted lines correspond to the
weight fractions of free water in the SC of 0, 0.12, and 0.24, respec-
of mm waves into the skin with a thin SC than those tively.The free water contents equal to 0 in the forearm SC and 0.12
obtained using the permittivity values for skin given by in the palm SC were obtained from the reflection measurements
Gandhi and Riazi [1986]. under normal conditions.

Bioelectromagnetics
68 Alekseev et al.

Skin Models With Thick SC

As was shown in our previous paper [Alekseev
and Ziskin, 2007], the homogeneous skin model was
inappropriate for modeling the reflection from the palm.
Therefore, for calculations of the reflection, PD and
SAR profiles in palmar skin we used only models 2 and
3. With both models we obtained similar results.
The thick SC in the palm played the role of a
matching layer and caused a reduction in reflection
(Fig. 2). In addition, the palmar skin manifested a broad
peak in reflection within the 83–277 GHz range. To
model reflection from the dry and hydrated palm, we
reduced the free water content in the SC down to 0 and
increased it twice, respectively. The results of this
Fig. 3. The power density profile of the 42 and 61 GHz plane mm
modeling are shown in Figure 2. At reduced hydration
wave field within the forearm skin calculated using model 3.Calcu- the position of the peak was shifted to the left and the
lations were made for the IPD of 10 mW/cm2. The first thin layer peak amplitude was increased. The minimum values of
showsthe stratum corneum.Theboundarybetweenthe epidermis reflection appeared at D ¼ l0/4 and 3l0/4, where
(E) anddermis (at a depthofabout 0.1 mm) ismarkedbyawavyline. D ¼ nscdsc is the optical thickness of the SC, l0 is
The arteriolar subpapillary plexus is located at the boundary
between the papillary (PD) and reticular (RD) dermis (0.3 mm).
the mm wavelength in air, nsc and dsc are the refractive
index and thickness of the SC, respectively. Similar
properties of reflection were described for a dielectric
film bounded by air on one side and another dielectric
on its other side [Born and Wolf, 1975]. At higher
hydration reflection from the palm increased and the
broad peak disappeared.
The PD and SAR profiles for 42 GHz exposure
with the IPD ¼ 10 mW/cm2 are presented in Figure 6. In
this case the PD and SAR calculations were made using
model 2. Calculations performed with model 3 gave
similar results. Due to the presence of free water, SAR
in the SC was not equal to 0 as it was in the forearm skin.
The d values for the palmar skin are presented in
Figure 5. Penetration depth into the palmar skin is
greater than into the forearm skin mostly due to the low
water content in the thick SC.

DISCUSSION
The power density and SAR profiles in skin with
the thin SC were virtually unchanged if calculated using
either the homogeneous or multilayer models. This can
be explained by two factors of mm wave interaction
with skin. The SC of the forearm epidermis is too thin
(15 mm) to play the role of a matching layer.
Therefore, reflection and absorption of mm waves were
only weakly dependent on SC hydration. However, the
forearm skin had a thick viable epidermis plus dermis
(1.45 mm). This layer containing a large amount of free
water produced great attenuation of mm wave energy.
Fig. 4. The frequencydependence ofthe SARat different depthsin Internal reflection of mm waves from the fat layer was
skin calculated for the plane mm wave field at an IPD of10 mW/cm2. too small to significantly change the PD and SAR
A:SAR at 0 and 0.1 mm depth.B:SAR at 0.3,0.35, and 0.4 mm depth. distributions in the dermis at the edge of the fat layer.
Bioelectromagnetics
 Millimeter Wave Dosimetry of Human Skin 69

the forearm skin, the first minimum appears far above

the mm wave frequency range.
We did not find any sharp resonance-type
frequency dependence for either reflection from the
skin or in the PD and SAR profiles within the skin
layers. It seems unlikely that the appearance of a
moderate SAR peak in the therapeutic frequency range
within the skin could be responsible for the distinct
frequency dependent effects reported in literature
[Sevastyanova, 1979; Pakhomov et al., 1998].
As shown in Figures 3 and 5, mm waves penetrate
deep enough into the skin to affect skin structures
located in the epidermis and dermis. At an IPD of
Fig. 5. Penetration depth (d) of the plane mm wave field in human
skin. Calculations were performed using model 2. Curves 1 and 2 10 mW/cm2, the power density within the epidermis
were obtained for the forearm skin using permittivity values given drops from 5.7 to 4.1 mW/cm2 at 42 GHz and from 6.2
inTable1and by Gandhi and Riazi [1986], respectively. to 4.1 mW/cm2 at 61 GHz, respectively. Main targets for
the mm wave effect in the epidermis are the keratino-
Therefore, mm wave interaction with the forearm skin cytes, melanocytes, and Langerhans cells as well as the
was well described even with the one layer homoge- Merkel cells located at the dermal–epidermal junction
neous skin model. [Bloom and Fawcett, 1968; Holbrook and Wolff, 1993;
In the palmar skin, the thick SC played a greater Mehregan et al., 1995]. In the papillary dermis, the mm
role in reflection and in PD and SAR profiles than the wave power density is still high. It drops from 4.1 to
thin SC in the forearm. Due to the presence of free water 2.2 and 1.7 mW/cm2 at 42 and 61 GHz exposures,
[Alekseev and Ziskin, 2007] the palmar SC absorbed a respectively. The papillary dermis contains free nerve
part of the mm wave energy and attenuated the PD endings, the most widespread and important sensory
within this layer. In the rest of the epidermis plus receptors. Some free nerve endings extend into the
dermis, the PD and SAR profiles were similar to those epidermis. Another type of sensory receptor, the
found for the forearm. The SC containing little or no Meissner corpuscle, is also located in the papillary
free water, and having a low permittivity, can be dermis. Monocytes and macrophages, as well as mast
considered as a dielectric film in contact with a medium cells are commonly present in this region.
of higher permittivity. This gives rise to the appearance The microcirculatory beds in skin include arterio-
of a reflection peak and two minimums. The positions of les, precapillary sphincters, arterial and venous capil-
the minimums depend on the optical thickness of the laries, postcapillary venules, and collecting venules
SC. The greater the optical thickness, the lower is the [Rhodin, 1967, 1968]. At the junction between the
frequency of the first minimum. For the palmar skin, papillary and reticular dermis, terminal arterioles form
the minimums lie in the mm wave frequency range. For a horizontal subpapillary plexus giving rise to the
capillary loops extending into the papillary dermis.
Venous capillaries drain into venules that lie above and
below the arteriolar subpapillary plexus. Thus, a large
amount of blood vessels and ultimately blood cells are
exposed to mm waves at a relatively high power density.
In the reticular dermis the power density decreases
from 2.2 and 1.7 to 0.11 and 0.02 mW/cm2 at 42 and 61
GHz exposures, respectively. Millimeter wave expo-
sures at therapeutic intensities are still capable of
affecting skin structures located in this layer [Pakho-
mov et al., 1998]. Thus, most skin structures present in
the epidermis and dermis could be the primary targets
for the mm wave action in human skin.

REFERENCES
Fig. 6. The power density (PD) and SAR profiles within the palmar
skin exposed to the 42 GHz plane mm wave field at an IPD of Alekseev SI, Ziskin MC. 2000. Reflection and absorption of
10 mW/cm2. The power density (solid line) and SAR (dotted line) millimeter waves by thin absorbing films. Bioelectromag-
profiles were calculated using model 2. netics 21:264–271.

Bioelectromagnetics
70 Alekseev et al.

Alekseev SI, Ziskin MC. 2001. Millimeter wave power density in Mashimo S, Kuwabara S, Yagihara S, Higasi K. 1987. Dielectric
aqueous biological samples. Bioelectromagnetics 22:288– relaxation time and structure of bound water in biological
291. materials. J Phys Chem 91:6337–6338.
Alekseev SI, Ziskin MC. 2007. Human skin permittivity determined Mehregan AH, Hashimoto K, Mehregan DA, Mehregan DR. 1995.
by millimeter wave reflection measurements. Bioelectro- Normal structure of skin. In: Pinkus’ guide to dermatology,
magnetics 28:331–339. 6th edition. Norwalk, Connecticut: Appleton and Lange.
Balantsev VN, Permyakov VA, Rakovskaya IL. 1991. The pp. 5–48.
millimeter wave range resonance absorption in multilayered Mirrashed F, Sharp JC. 2004a. In vivo quantitative analysis of the
models of the skin. In: Devyatkov ND, Betsky OV, editors. effect of hydration (immersion and Vaseline treatment) in
Digest of papers of International Symposium ‘‘Millimeter skin layers using high-resolution MRI and magnetisation
waves of non-thermal intensity in medicine’’. Moscow: transfer contrast. Skin Res Technol 10:14–22.
USSR Acad Sci. pp. 655–659. Mirrashed F, Sharp JC. 2004b. In vivo morphological character-
Betsky OV, Yaremenko YG. 1998. The skin and electromagnetic ization of skin by MRI micro-imaging methods. Skin Res
waves. Millimeter Waves in Biol Med N1(11):3–14. Technol 10:149–160.
Bloom W, Fawcett DW. 1968. A textbook of histology, 9th edition. Pakhomov AG, Akyel Y, Pakhomova ON, Stuck BE, Murphy MR.
Philadelphia: Saunders. pp. 479–509. 1998. Current state and implications of research on biological
Born M, Wolf E. 1975. Principles of optics. Electromagnetic theory effects of millimeter waves: A review of the literature.
of propagation, interference and diffraction of light. London: Bioelectromagnetics 19:393–413.
Pergamon Press. pp. 627–631. Pickwell E, Cole BE, Fitzgerald AJ, Pepper M, Wallace VP. 2004. In
Buchner R, Barthel J, Stauber J. 1999. The relaxation of water vivo study of human skin using pulsed terahertz radiation.
between 0 8C and 35 8C. Chem Phys Lett 306:57–63. Phys Med Biol 49:1595–1607.
Caspers PJ, Lucassen GW, Carter EA, Bruining HA, Puppels GJ. Rhodin JAG. 1967. The ultrastructure of mammalian arteriols and
2001. In vivo confocal Raman microspectroscopy of the skin: precapillary sphincters. J Ultrastruct Res 18:181–223.
Noninvasive determination of molecular concentration Rhodin JAG. 1968. The ultrastructure of mammalian venous
profiles. J Invest Dermatol 116:434–442. capillaries, venules, and small collecting veins. J Ultrastruct
Caspers PJ, Lucassen GW, Puppels GJ. 2003. Combined in vivo Res 25:452–500.
confocal Raman spectroscopy and confocal microscopy of Richard S, Querleux B, Bittoun J, Idy-Peretti I, Jolivet O,
human skin. Biophys J 85:572–580. Cermakova E. 1991. In vivo proton relaxation times analysis
Foster KR, Schwan HP. 1986. Dielectric properties of tissues. In: of the skin layers by magnetic resonance imaging. J Invest
Polk C, Postow E, editors. Handbook of biological effects of Dermatol 97:120–125.
electromagnetic fields. Boca Raton, Florida: CRC Press, Inc., Richard S, Querleux B, Bittoun J, Jolivet O, Idy-Peretti I, de
pp. 27–96. Lacharriere O, Leveque J-L. 1993. Characterization of the
Foster KR, Schepps JL, Schwan HP. 1980. Microwave dielectric skin in vivo by high resolution magnetic resonance imaging:
relaxation in muscle. A second look. Biophys J 29:271–282. Water behavior and age-related effects. J Invest Dermatol
Gabriel C, Gabriel S, Corthout E. 1996a. The dielectric properties of 100:705–709.
biological tissues. I. Literature survey. Phys Med Biol 41: Rojavin MA, Ziskin MC. 1998. Medical application of millimeter
2231–2249. waves. Q J Med 91:57–66.
Gabriel S, Lau RW, Gabriel C. 1996b. The dielectric properties of Ryakovskaya ML, Shtemler VM. 1983. Absorption of electro-
biological tissues. III. Parametric models for the dielectric magnetic waves of millimeter range in biological prepara-
spectrum of tissues. Phys Med Biol 41:2271–2293. tions with a plane layer structure. In: Devyatkov ND, editor.
Gandhi OP, Riazi A. 1986. Absorption of millimeter waves by Effect of nonthermal action of millimeter radiation on
human beings and its biological implications. IEEE Trans biological subjects. Moscow: USSR Academy of Sciences.
Microwave Theory Tech 34:228–235. pp. 172–181.
Grant EH. 1982. The dielectric method of investigating bound water Schwan HP, Foster KR. 1977. Microwave dielectric properties of
in biological material: An appraisal of the technique. tissue. Some comments on the rotational mobility of tissue
Bioelectromagnetics 3:17–24. water. Biophys J 17:193–197.
Holbrook KA, Wolff K. 1993. The structure and development of Sevastyanova LA. 1979. Peculiarities of biological effects of mm
skin. In: Fitzpatrick B, Eisen AZ, Wolff K, Freedberg IM, waves and their application to medicine. Vestnik AMN SSSR
Austen KF, editors. Dermatology in general medicine, 4th 2:65–68.
edition. New York: McGraw-Hill, Inc., pp. 97–145. Warner RR, Myers MC, Taylor DA. 1988. Electron probe analysis of
Imokawa G, Kuno H, Kawai M. 1991. Stratum corneum lipids serve human skin: Determination of the water concentration
as a bound-water modulator. J Invest Dermatol 96:845–851. profile. J Invest Dermatol 90:218–224.

Bioelectromagnetics