IEEE TRANSACTIONS ON ELECTROMAGNETIC COMPATIBILITY, VOL. 49, NO. 4, NOVEMBER 2007 825 Prediction of Temperature Increase in Human Eyes Due to RF Sources Concettina Buccella, Senior Member, IEEE, Valerio De Santis, Member, IEEE, and Mauro Feliziani, Senior Member, IEEE Abstract—A numerical study is proposed to investigate the ef- fects of different RF sources on the specific absorption rate (SAR) and maximum temperature increase in the human eye at differ- ent frequencies. In particular, a new model of the human head is presented and compared with an anatomical model of the visi- ble human. The high resolution (0.5 mm) of the proposed model allows to consider more eye tissues than previous studies distin- guishing the sclera from the retina and choroid. New values of blood perfusion and metabolic rate of these tissues are derived. A plane-wave field is considered as far-field exposure, while realistic models of mobile phone and dipole antennas are used as primary sources for near-field exposure. The obtained results show that the distributions of the SAR and temperature increase depend on the frequency, position, and kind of sources. Finally, attention is paid to the maximum temperature increase in the lens for the SAR values prescribed by the Commission on Non-Ionizing Radiation Protec- tion. To this aim, a scaling approach is proposed, and significant values of temperature increase are found (about 0.3 ◦ C for general public exposure and about 1.5 ◦ C for occupational exposure) for the most critical cases of near-field exposures. Index Terms—Cellular phones, finite difference method, human exposure to electromagnetic fields (EMF), human eye modeling, numerical dosimetry, thermal simulation. I. I NTRODUCTION I N RECENT years, the enormous developments of wireless systems and personal communication devices have signif- icantly increased the exposure to radio-frequency (RF) elec- tromagnetic (EM) waves. As a consequence, it is important to consider the possible health hazards due to these kinds of de- vices (e.g., mobile phones) that are used close to the human head. It is well known that short-term acute effects of intense RF exposure can occur in sensitive tissues that exhibit significant thermal damage. The RF fields have been reported to cause a variety of ocular effects, primarily cataracts in the lens, and also effects on the retina, cornea, and other ocular systems. The results described in the literature lead to the conclusion that EM fields induce cataracts in rabbits for a temperature rise of 3–5 ◦ C [1]–[5]. On the other hand, no cataract formation was observed in monkey eyes for the same kind of exposure. This inconsistency was attributed to the difference in RF energy absorption due to the different anatomical shape of the head of rabbits and monkeys Manuscript received August 3, 2006; revised March 2, 2007 and May 1, 2007. C. Buccella and M. Feliziani are with the Department of Electrical Engineer- ing, University of L’Aquila, 67040 L’Aquila, Italy (e-mail: c.buccella@ieee.org; felizian@ing.univaq.it). V. De Santis is with the Department of Electrical and Computer Engineering, University of L’Aquila, 67040 L’Aquila, Italy (e-mail: v.desantis@ieee.org). Color versions of one or more of the figures in this paper are available online at http://ieeexplore.ieee.org. Digital Object Identifier 10.1109/TEMC.2007.909024 [4]. Thus, it is interesting to investigate the temperature increase in the human eye due to both far-field [6]–[9] and near-field exposures [10], [11]. To predict the temperature increase in human eyes, numeri- cal methods are adopted here. Anatomically based models de- rived from the Visible Human Project (VHP) [12] or magnetic resonance imaging (MRI) of human volunteers are usually em- ployed, but their small resolution of 1–2 mm is not enough for an accurate model of the eye. This problem was overcome in [13] and [14] where a 0.25-mm 2-D model of the eye and head has been developed for an implanted retinal stimulator. In the present paper, a 3-D computer-aided design (CAD) model of the head with seven eye tissues and high resolution is proposed and compared with the VHP anatomical model with a resolution of 2 mm. The advantage of the proposed model is the flexibility to enhance anatomical structure, especially near the eye zone, and increase accuracy in future. Another problem of anatomical models is that the retina and choroid layers, which are about 0.25 mm thick, are not usu- ally taken into account, but they are fundamental from a ther- mal point of view due to their huge blood flow and metabolic rate [15]. Published works do not give explicit values of blood perfusion and metabolic rate of eye tissues. In this study, we have derived realistic values for these parameters using the data from recent ophthalmic studies on patients with vascular dis- eases, such as hypertension and diabetes [16]–[18]. Finally, since near-field exposure is actually considered more important than plane-wave exposure, an accurate CAD model of a mobile phone with a triband planar inverted F antenna (PIFA) and an isolated dipole antenna are considered, and the effect of these sources on the increase in eye temperature is evaluated for several configurations. The thermal model has been also applied to the human eye when considering the International Commission on Non-Ionizing Radiation Protection (ICNIRP) limits on the averaged specific absorption rate (SAR) by scaling up the SAR distributions obtained by the previous calculations for near-field exposures. II. M ODELS AND M ETHODS A. Human Head Models Two models of human heads are considered in this paper. The first model is based on the VHP and consists of a regular mesh with 293 (width) × 170 (depth) × 116 (height) voxel (air enclose), with a resolution of 2 mm as shown in Fig. 1(a). The VHP model is composed of 20 tissues (blood, bone cancel- lous, cortical and marrow, cartilage, cerebrospinal fluid (CSF), 0018-9375/$25.00 © 2007 IEEE 826 IEEE TRANSACTIONS ON ELECTROMAGNETIC COMPATIBILITY, VOL. 49, NO. 4, NOVEMBER 2007 Fig. 1. Human head models. (a) Anatomical model of the visible human. (b) CAD model. cornea, lens, sclera, vitreous and aqueous humor, fat, mucous membrane, muscle, nerve, gray matter, white matter, cerebel- lum, skin, and tooth). The second model of the head is obtained by a CAD software tool considering the human anatomy [19]. This CAD model comprises almost the same tissues as in the VHP model with the significant addition of the retina, choroid, optic nerve, orbit fat, and eye muscle. Particular attention has been devoted to the eye zone, nose, and the front of the head, as shown in Fig. 1(b), while for obvious reasons, the brain and back zone were given less attention. The advantage of the CAD model is that a variable mesh size of the computational region could be applied, saving computational time and increasing the accuracy of the obtained results. B. EM Modeling The EM problem is analyzed using Computer Simulation Technology (CST) Microwave Studio (MWS) code [20]. This solver is based on the finite integration technique (FIT) that in time-domain formulation becomes analogous to Yee’s finite- difference time-domain (FDTD) scheme. As a far-field source, we have considered a plane-wave excitation with a power den- sity of 5.0 mW/cm 2 for the sake of comparisons, even if the reference levels for general public exposure [21] at 1.0, 1.9, and 2.45 GHz are 0.5, 0.95, and 1.0 mW/cm 2 , respectively. An iso- lated half-wave dipole and an accurate CAD model of a mobile phone with a triband PIFA antenna are also considered as near- field sources for several human head–antenna configurations. Both anatomical VHP and CAD electrogeometrical models have been imported by CST MWS. In the VHP model, a uniform structured mesh is used, while in the CAD model, a nonuniform rectilinear mesh with a minimum cell size of 0.2 mm and a maximum cell size of 2.0 mm has been adopted. The more finely discretized zones are inside the eyeball to describe accurately the geometries of the different physical regions of the eye, as shown in Fig. 2. The Courant stability condition is satisfied in the whole computational domain. For the truncation of the computational region in the CST calculations, we have used Berenger’s perfectly matched layer (PML) with six layers and reflection coefficient equal to 1 × 10 −5 . Fig. 2. Variable mesh size in the eye zone for the CAD model. The dielectric properties of the tissues were determined by the 4-Cole–Cole extrapolation [22]. Note that averaged values for lens cortex and nucleus were used as material constants of the lens, and dielectric properties of vitreous humor were used as those of aqueous humor due to the lack of actual data. For the CAD model, we have used the same material constants of gray matter and blood for the retina and choroid, respectively, while for optic nerve, orbit fat, and eye muscle, we have used the same constants of nerve, fat, and muscle, respectively, as suggested in [13]. C. Cellular Phone Model A CAD model of a mobile phone with realistic compo- nents and a triband PIFA antenna is considered, as shown in Fig. 3(a). The PIFA’s metallic ground plane acts as a shield to prevent the EM waves from radiating to the user, thus minimiz- ing energy dissipation in the biological tissues. A structure of 14 mm × 55 mm × 105 mm is embedded in a FR4 substrate (ε r =4.6) and enclosed in a plastic case of acetal (ε r =2.8 and σ =0.002 S/m), which is commonly used for commercial phones, while a lithium battery (ε r =3and σ =0.8 S/m) is set on a typical position. Fig. 3(b) shows the radiating element proposed in [23], which is an E-shaped patch antenna with two shorting strips to improve the impedance matching at the triple bands of global system for mobile communications (GSM; 900 MHz), digital cellular system (DCS; 1800 MHz), and uni- versal mobile telecommunications system (UMTS; 2100 MHz), as shown in Fig. 3(c). D. SAR Calculation CST MSW provides whole-body averaged and local SAR values averaged over a cubic volume of specified tissue masses (1 or 10 g). However, to obtain the SAR distribution needed for the thermal model, we have used the well-known equation for BUCCELLA et al.: PREDICTION OF TEMPERATURE INCREASE IN HUMAN EYES DUE TO RF SOURCES 827 Fig. 3. Phone model. (a) Sketch of mobile phone. (b) PIFA antenna. (c) Return loss of the proposed antenna. time-harmonic EM fields SAR = σ 2ρ | ˆ E| 2 = σ 2ρ  | ˆ E x | 2 + | ˆ E y | 2 + | ˆ E z | 2  (1) where ˆ E x , ˆ E y , and ˆ E z are the peak values of the electric field components; and σ and ρ are, respectively, the conductivity and mass density of the tissue. The electric field, calculated by using a variable mesh size, was exported from the CST MWS by a linear interpolation procedure as a structured mesh of 0.5 mm cell size, and so this is the adopted mesh resolution for the calculation of the temperature increase distribution. E. Thermal Model For calculating the temperature increase inside the EM ex- posed tissues, the bioheat equation is used [24] Cρ ∂T ∂t = ∇(k∇T )+ρ(SAR) + A − B(T − T b ) (2) where T and T b denote the temperature of tissue and blood (Celsius degrees), respectively, C the specific heat of the tissue [J/(kg · ◦ C)], K the thermal conductivity of the tissue [J/(s · m · ◦ C)], A the basal metabolic rate (W/m 3 ), and B is the term associated to the blood perfusion [W/( ◦ C · m 3 )] that is usually given by B = C b W b = C b ρ b ρF (3) where C b = 3900 J/(kg · ◦ C) is the specific heat of blood, W b is the blood perfusion [kg/(m 3 · s)], ρ b = 1060 kg/m 3 is the mass density of blood, and F is the blood flow rate for mass unit [m 3 /(kg · s)]. Note that is valid when the temperature in- crease in tissues is sufficiently small where the thermoregula- tory mechanism is negligible. Moreover, for microwave expo- sure, the thermal elevation reaches the steady state after about 30 min, and so, this is the time of exposure adopted in this paper. It should be noted that the solution of (2) requires knowledge of the initial (or normo-thermal) temperature distribution. This is the temperature inside the human tissues without any RF field exposure, and can be obtained by the steady-state equation [24] ∇(K∇T )+A − B(T − T b )=0. (4) Equations (2) and (4) are here solved by a procedure based on the finite-difference method. To apply this method, it is relevant to individuate the closed computational domain with adequate boundary conditions. For the whole-head volume, the convective boundary conditions are usually applied on the skin–air and cornea–air interfaces [4], [24] −K  ∂T ∂n  S = H(T − T e ) (5) where H [W/(m −2 · ◦ C)] is the convection coefficient, T is the unknown surface temperature, and T e is the fluid temperature (corresponding to the air temperature). The convection coeffi- cient between the skin and the air is assumed to be H = H S = 10.5 W/(m 2 · ◦ C) [25], while H = H C =20W/(m 2 · ◦ C) is used between the cornea surface and the air [26]. It should be noted that the values of H include the following effects: 1) evaporation of the tear film on the cornea and insensible per- spiration (sweating) on the skin; 2) convective exchange with the air; and 3) radiative exchange with the surrounding objects. Furthermore, these values are obtained in the condition when the air temperature is T e =23 ◦ C. In order to reduce the computational cost, several previous works have considered for the thermal analysis of an undersized domain consisting of the only human eye region applying equiv- alent convective boundary conditions on the sclera [26], [27], but this simplified model was not able to take into account the EM energy absorbed in adjacent tissues and conducted into the eye [28]. However, due to the high resolution adopted here, i.e., 0.5-mm cell size, it is quite impracticable to apply the thermal 828 IEEE TRANSACTIONS ON ELECTROMAGNETIC COMPATIBILITY, VOL. 49, NO. 4, NOVEMBER 2007 Fig. 4. Reduced volume for the thermal model. (a) Discretized geometry with boundary conditions. (b) Normo-thermal temperature distribution. model to the whole head by using a so fine discretization, and therefore, a reduced volume containing the eyeball has been con- sidered for the solution of the transient thermal equation (2). In order to consider a reduced-volume region instead of the whole head, we have to solve the following problems: 1) how to define the dimensions and the position of the reduced-volume region; and 2) which are the boundary conditions of the reduced region and how to apply these conditions? To this aim, the thermal distribution calculated in the whole-head region by a coarse dis- cretization of 1.0-mm cell size has been used for the validation of the proposed reduced thermal model [29]. By these calculations, the reduced computational region has been therefore defined as the box-shaped domain whose boundary surfaces are placed at a distance of about 2.5 cm from the eyeball, as shown in Fig. 4(a). It should be noted that this region is large enough to take into ac- count the most part of the EM energy deposited near the eyeball, and therefore, small approximations in the boundary conditions inside the human head do not lead to significant inaccuracy. The energy concentration in the eyeball is due to the low water content in the tissues surrounding the eye (fat and bone) and to the limited field penetration on human tissues at the considered frequencies. The boundary conditions in the reduced domain [see Fig. 4(a)] with fine discretization (0.5-mm cell size) are derived by a previous calculation in the whole-head domain with coarse discretization (1.0-mm cell size) [29]. In the reduced do- main the convective boundary conditions (5) applied for the whole-head thermal model have been imposed on the skin– air and cornea–air interfaces, while on the other bound- aries, with the exception of the most internal surface, the thermal isolation (or homogeneous Neumann) conditions are adopted −K  ∂T ∂n  S =0. (6) In the most internal surface, the convective boundary condi- tion (5) has been used assuming H = H i =40W/(m 2 · ◦ C) for the heat transfer coefficient and T e =37 ◦ C for the body-core fluid temperature [29]. Solving the normo-thermal temperature distributions (4) in the reduced domain, a skin and corneal average temperatures of 34.1 and 32.6 ◦ C are obtained, respectively, as shown in Fig. 4(b). These values are in good agreement with the measure- ments found in the literature: 34.0 ◦ C for the head surface [30] and 32.7 ◦ C for the cornea [31]. F. Thermal Parameters The thermal parameters of the tissues in the reduced volume considered are given in Table I. Note that the values shown in [11] are used, with the exception of the choroid/ retina. This is due to the better resolution of the proposed model that distin- guishes the sclera from the choroid and retina. Moreover, new values of blood perfusion and metabolic rate are adopted here. In fact, it is well known that the retina has a high metabolic rate due to the continual replacement of photoreceptor re- quired for vision [15]; consequently, it uses a lot of oxygen and nutrients, supplied by the highly vascularized choroid. Re- cently, several methods to estimate the retinal and choroidal blood flow rate have been developed for application in subjects with systematic vascular diseases, such as hypertension and diabetes. From these data, a retinal blood flow of about 65 µL/min [16] is obtained, while a choroidal pulsatile ocular blood flow (POBF) variable in the range of 600–1740 µL/min is found, ac- cording to the user’s manual [17]. The reason for this large vari- ability was attributed to a lot of factors, such as the age, gender, heart rate, and so on [18]. This suggests to perform a sensitivity analysis on the B value of the choroid/retina layer by reducing or increasing the choroidal blood flow to the values found on BUCCELLA et al.: PREDICTION OF TEMPERATURE INCREASE IN HUMAN EYES DUE TO RF SOURCES 829 TABLE I T HERMAL P ROPERTIES AND D ENSITY OF T ISSUES IN THE C ONSIDERED V OLUME the pathological cases. In fact, by applying a weighted proce- dure (choroid 0.78 g and retina 0.52 g), it is possible to estimate a choroid/retina blood flow variable in the range of 386–1070 µL/min, and then, knowing the weight of these tissues (about 1.3 g), we can calculate the blood flow rate for unit mass F used in (3). This leads to a choroid/retina B value variable in the range of 21 685–60 113 W/( ◦ C · m 3 ), and so, a mean value of 40 000 W/( ◦ C · m 3 ) is adopted here as a reference level (see Table I). It should be noted that in the whole-head calculations with a coarse resolution of 1.0 mm, the sclera (not modeled with blood flow) is also included in this weighted procedure; therefore, a reference value of about 20 000 W/( ◦ C · m 3 ) is ob- tained for the sclera/choroid/retina layer. This latter value is very similar to the 13 500 W/( ◦ C · m 3 ) used in [11] because in that paper, a different resolution (2.0-mm cell size) has been adopted. However in [11], the B value was derived from a theoretical pro- cedure by imposing the same temperature increase distributions in the eyes of two different heat transportation models [28]. The metabolic rates of the retina and choroid layers are not available explicitly, so we have assumed the metabolic rate of the choroid/retina to be proportional to the blood perfusion, as suggested in [32]. III. N UMERICAL R ESULTS A. Far-Field Exposure A plane wave field with vertical polarization and power den- sity of 5.0 mW/cm 2 is considered at different frequencies as a far-field source. In order to validate the proposed CAD model, we have compared the local SAR provided by CST MSW and averaged over 10 g for both human models (i.e., VHP and CAD models) exposed to a plane-wave field of 5.0 mW/cm 2 at 2.45 GHz. The results are reported in Fig. 5(a) and (b) and show that the SAR of both models is located in the upper right zone of the eye due to the diffraction of the nose, in agreement with ear- lier works [7], [8], [11]. It should be noted that some differences between the two models occur. These can be only justified by the different electrogeometrical configurations (tissue compo- sitions and anatomical structure), such as the less pronounced nose and orbit cavity of the visible human (see Fig. 1). Another possible reason of the discrepancies in the results regards the Fig. 5. Local SAR averaged over 10 g in the right eye for a plane wave exposure at 2.45 GHz. (a) SAR distribution in the eye surface of the anatomical VHP model. (b) SAR distribution in the eye surface of the CAD model. Fig. 6. SAR and temperature increase distributions calculated by the CAD model for plane wave exposures in the vertical plane across the center of the right eye. (a) SAR at 1.0 GHz. (b) Temperature increase at 1.0 GHz. (c) SAR at 1.9 GHz. (d) Temperature increase at 1.9 GHz. different position of the eyelid: closed eyelid in the VHP model, open eye in the CAD model. Fig. 6 shows the distributions of SAR and temperature in- crease calculated by the CAD model at the frequencies of 1.0 and 1.9 GHz. By these distributions, the small EM absorption of the lens is evident at this frequency [see Fig. 6(a) and (c)], due to low water content compared with surrounding tissues. The results also show that the peak SAR values increase as the frequency increases, but at the same time, the EM absorption decreases as the frequency increases due to the skin effect. The calculated maximum value of the temperature increase in the lens is about 0.2 ◦ C [see Fig. 6(d)]. However, this value is lower than the threshold temperature rise of 3–5 ◦ C needed for lens opacification. B. Near-Field Exposure The thermal elevation induced inside the human eye by near- field sources is studied. A triband mobile phone with a power 830 IEEE TRANSACTIONS ON ELECTROMAGNETIC COMPATIBILITY, VOL. 49, NO. 4, NOVEMBER 2007 Fig. 7. Mobile phone configurations. (a) Phone in tilted position. (b) Phone with display in front of the eye. (c) Phone with display opposite the eye. (d) Phone between two heads. TABLE I I SAR AND T EMPERATURE I NCREASE V ALUES I NSIDE THE R IGHT E YE FOR S EVERAL M OBILE P HONE C ONFIGURATIONS output of 0.25 W and a half-wave dipole antenna with a power output of 1.0 W have been considered. Fig. 7 shows four different human head–phone configurations [i.e., test cases (a)–(d)] considered of major interest for our applications, even if the test case configurations (c) and (d) are not very usual. The results of our investigation are summarized in Table II showing the SAR and temperature increases inside the exposed eye for the different configurations. The most critical situations can be found in the 900-MHz band and in the (c) and (d) test case configurations of mobile phone exposure. This is due to the considered PIFA antenna. In fact, the ground plane of the PIFA [see Fig. 3(a) and (b)] tends to shield the EM field in the display direction, while the geometry of the patch makes a good impedance matching especially in the 900-MHz band [see Fig. 3(c)]. It should be noted that only the 900-MHz band is considered for the mobile phone between two heads, i.e., test case (d), because the increase of the computational region dimension makes more stringent limits on the computer memory used. Fig. 8. Geometry configuration for the dipole antenna exposure. Fig. 9. SAR and temperature increase distributions calculated by the CAD model on the vertical plane across the center of the right eye for the dipole antenna at 1.5 GHz. (a) SAR at d = 1.2 cm. (b) Temperature increase at d = 1.2 cm. (c) SAR at d = 3.2 cm. (c) Temperature increase at d = 3.2 cm. For the dipole antenna exposure, the vertical polarization is adopted and the antenna center is assumed to be located in front of the eye at a separation distance d, as shown in Fig. 8. The eye–dipole distance d was chosen equal to 1.2 and 3.2 cm, while the radius and the length of the dipole were assumed to be 0.5 and 90 mm, respectively, according to the frequency range of 1.5 GHz. Fig. 9 reports the SAR and temperature rise distributions for this kind of exposure. In order to confirm the validity of our results, we compared the eye-averaged SAR and maximum temperature increase in the lens of our investigations with those reported in [11]. Table III shows how our results are clearly lower than those reported in the literature again due to the different anatomical shape and tissue layer thickness of the human heads [33]. In fact, the head model used in [11] was obtained started by an MRI of an Asiatic human volun- BUCCELLA et al.: PREDICTION OF TEMPERATURE INCREASE IN HUMAN EYES DUE TO RF SOURCES 831 TABLE III E YE —A VERAGED SAR AND M AXIMUM T EMPERATURE I NCREASE IN THE L ENS :C OMPARISON B ETWEEN O UR S TUDY AND [11] TABLE I V E FFECT OF CHOROID /R ETINA B LOOD F LOW ON THE M AXIMUM T EMPERATURE I NCREASE IN THE L ENS AND ON THE C ORRELATING S LOPE R teer, and so with nonpronounced lineaments of face. However, if we define R as the ratio between the eye-averaged SAR and maximum temperature increase in the lens, a value of about R =0.16 ◦ C · kg/W is found for both works, and this is due to the fact that in [11], the blood perfusion of the choroid is equivalently taken into account as stated previously. It should be noted that for plane-wave exposure, the correlating slope R between the eye-averaged SAR and maximum temperature increase in the lens was found to be about 0.18 ◦ C · kg/W. This slight difference is due to the different EM energy de- position outside the orbit eye that is larger for the plane-wave exposure. Furthermore, the effects of the pathological variations of choroid blood flow on the eye thermal elevation have been considered for the plane wave and dipole antenna exposures. Table IV shows how these variations are reflected in a chang- ing of the correlating slope R of about ±0.1 ◦ C · kg/W from the reference levels (R = 0.16 ◦ C · kg/W for the dipole antenna and R = 0.18 ◦ C · kg/W for the plane-wave exposure) obtained with the control value of B = 40 000 W/( ◦ C · m 3 ). C. ICNIRP Exposure For public health safeguards, it is important to investigate the effects on temperature rise in the lens for the SAR values pre- scribed by the safety guidelines. According to the ICNIRP [21], the upper limits for local SAR in the head are equal to 2 and 10 W/kg for general public and occupational exposures, respec- tively. These values must be averaged over a contiguous tissue mass of 10 g. Thus, we have scaled the SAR distribution previ- ously calculated in the reduced volume to obtain the averaged SAR in the eye (28 mm diameter for 9.9 g in our CAD model) equal to the basic restrictions indicated by the ICNIRP guide- line. However, this procedure cannot be applied for any kind of exposures because the eye-averaged SAR does not strictly TABLE V M AXIMUM T EMPERATURE I NCREASE IN THE L ENS FOR THE ICNIRP L IMITS ON G ENERAL P UBLIC AND O CCUPATIONAL E XPOSURES correspond to the maximum averaged SAR in the head. To this aim, only the dipole antenna and cell phone in the test cases (b), (c), and (d) exposures have been considered, as listed in Table V. For the ICNIRP limits on general public exposure (2 W/kg), the maximum temperature increases in the lens are found to be 0.291–0.320 ◦ C, very similar to the 0.309–0.348 ◦ C for an isolated dipole antenna, and the 0.303–0.342 ◦ C for a monopole antenna on a metallic box reported in [11]. The respective lens temperature rises obtained for the ICNIRP limits on occupa- tional exposure (10 W/kg) are instead in the range 1.49–1.60 ◦ C. Even if not negligible, these values are lower than the threshold limits needed to induce adverse thermal effects in the eye. Fur- thermore, it should be noted that our thermal model does not take into account some thermoregulatory mechanisms as eye- lid closure, evaporation of tear liquid, and blood flow increase. Therefore, leaving out these mechanisms leads to overestimation of the temperature increase in the eye for intense EM exposure, such as those of the ICNIRP limits for occupational exposure. IV. C ONCLUSION SAR and temperature distributions in the human eye have been numerically calculated for several configurations and RF sources, both for near-field and far-field exposures. The SAR has been derived by a commercial software tool (CST MWS) importing the electrogeometrical configurations under examina- tion. The thermal problem has been solved by a finite-difference procedure developed by the authors importing the SAR values previously calculated. A new CAD model of the human head has been proposed considering many different tissues in the hu- man eye due to the very high discretization of 0.5 mm. New values of blood perfusion and metabolic rate have been esti- mated for the retina/choroid layers and used for the first time in the thermal computation. Furthermore, a sensitivity analysis on the large variation of the choroid blood flow present in several pathological cases has been considered, highlighting the effect on the eye thermal elevation. A triband integrated PIFA cellular phone and half-wave dipole antennas have been adopted as primary sources of near-field exposure, while the far-field exposure has been modeled by a plane-wave field. The simulation results obtained by the solution of the bioheat equation for several mobile phone configurations have clearly shown that there are no significant temperature increases in the lens for this kind of exposure. The results of our investigations have shown that SAR and temperature increase in human eye are largely dependent on the frequency, position, and kind of the sources, according to recent 832 IEEE TRANSACTIONS ON ELECTROMAGNETIC COMPATIBILITY, VOL. 49, NO. 4, NOVEMBER 2007 studies [8], [11]. Moreover, for a fixed exposure, the comparison of the CAD model with the anatomical VHP and MRI models of the head has shown that there are some differences in the SAR distributions due to the different human head shapes. Neverthe- less, the CAD model, even if less accurate in the human design, can be much more accurate for its fine discretization, and can also be very suitable to simulate different anatomical shapes of individual people, which can produce quite different numerical results, especially for plane-wave exposure. The thermal model has also been applied to the human head when considering the ICNIRP limits of the averaged SAR. To this aim, the results of the previous calculations in the eye have been scaled up in order to obtain an averaged SAR equal to the safety ICNIRP guideline. By this procedure, the maximum temperature elevations in the lens have been found to be 0.291– 0.320 ◦ C for general public exposure (assuming an averaged SAR =2W/kg) and 1.49–1.60 ◦ C for occupational exposure (assuming an averaged SAR =10W/kg), respectively. These values are protective against the threshold temperature rise of 3–5 ◦ C needed for cataract formation; however, they are not negligible and higher than the safety margin of ten, typically used in terms of SAR. This suggests to relate the threshold limits, for the sensitive organs like the eye, in terms of temperature rises and not only in terms of average SAR, as currently used. A CKNOWLEDGMENT The authors would like to thank Prof. G. Macchiarelli (the Head of the Anatomy Department of University of L’Aquila) and Dr. J. Elder (Motorola, Inc., Plantation, FL) for their useful suggestions. R EFERENCES [1] A. Taflove and M. E. Brodwin, “Computation of the electromagnetic fields and induced temperatures within a model of the microwave irradiated human eye,” IEEE Trans. Microw. Theory Tech., vol. MTT-23, no. 11, pp. 888–896, Jun. 1975. [2] J. C. Lin, “Cataract and cell-phone radiation,” IEEE Antennas Propag. Mag., vol. 45, no. 1, pp. 171–174, Feb. 2003. [3] J. Elder, “Ocular effects of radiofrequency energy,” Bioelectromagnetics Suppl., vol. 24, no. S6, pp. S148–S161, Nov. 2003. [4] P. Kramer, C. Harris, A. F. Emery, and A. W. Guy, “Acute microwave irradiation and cataract formation in rabbits and monkeys,” J. Microw., vol. 13, pp. 239–249, 1978. [5] A. W. Guy, J. C. Lin, P. O. Kramar, and A. Emery, “Effect of 2450- MHz radiation on the rabbit eye,” IEEE Trans. Microw. Theory Tech., vol. MTT-23, no. 6, pp. 492–498, Jun. 1975. [6] A. Taflove and M. E. Brodwin, “Computation of the electromagnetic fields and induced temperatures within a model of the microwave-irradiated human eye,” IEEE Trans. Microw. Theory Tech., vol. MTT-23, no. 11, pp. 888–896, Nov. 1975. [7] P. Bernardi, M. Cavagnaro, S. Pisa, and E. Piuzzi, “SAR distribution and temperature increase in an anatomical model of the human eye exposed to the field radiated by the user antenna in a wireless LAN,” IEEE Trans. Microw. Theory Tech., vol. 46, no. 12, pp. 2074–2082, Dec. 1998. [8] A. Hirata, S. Matsuyama, and T. Shiozawa, “Temperature rises in the human eye exposed to EM waves in the frequency range 0.6–6 GHz,” IEEE Trans. Electromagn. Compat., vol. 42, no. 4, pp. 386–393, Nov. 2000. [9] A. Hirata, H. Watanabe, and T. Shiozawa, “SAR and temperature rise in the human eye induced by obliquely incident plane waves,” IEEE Trans. Electromagn. Compat., vol. 44, no. 4, pp. 594–596, Nov. 2002. [10] M. Kita, A. Hirata, and T. Shiozawa, “Temperature rises in the human eye exposed to EM waves emitted from a dipole antenna at various microwave frequencies,” (in Japanese), Trans. Inst. Elect. Inf. Commun. Eng. Jpn., vol. J85-B, no. 1, pp. 132–139, 2002. [11] A. Hirata, “Temperature increase in human eye due to near-field and far-field exposures at 900 MHz, 1.5 GHz, and 1.9 GHz,” IEEE Trans. Electromagn. Compat., vol. 47, no. 1, pp. 68–77, Feb. 2005. [12] The Visible Human Project (2000) [Online]. Available: http://www.nlm. nih.gov/research/vis-ible/visible_human.html [13] S. C. DeMarco, G. Lazzi, W. Liu, J. D. Weiland, and M. S. Humayun, “Computed SAR and thermal elevation in a 0.25-mm 2-D model of the human eye and head in response to a retinal implanted stimulator— Part I: Models and methods,” IEEE Trans. Antennas Propag., vol. 51, no. 9, pp. 2274–2285, Sep. 2003. [14] G. Lazzi, S. C. DeMarco, W. Liu, J. D. Weiland, and M. S. Humayun, “Computed SAR and thermal elevation in a 0.25-mm 2-D model of the human eye and head in response to a retinal implanted stimulator—Part II: Results,” IEEE Trans. Antennas Propag., vol. 51, no. 9, pp. 2286–2295, Sep. 2003. [15] M. Morales, J. Ramon, I. Rodrigo, and P. Bovolenta, “Eye development: A view from the retina pigment epithelium,” Bioessays, vol. 26, pp. 766– 777, 2004. [16] J. P. Garcia, Jr., P. T. Garcia, and R. B. Rosen, “Retinal blood flow in the normal human eye using the canon laser blood flowmeter,” Ophthalmic Res., vol. 34, no. 5, pp. 295–299, Sep.–Oct. 2002. [17] R. Banfiel, Dicon Diagnostic Blood Flow Analyzer User’s Manual.San Mateo, CA: Paradigm Medical Industries, Inc., 2001. [18] A. Aydin, G. Wollstein, L. L. Price, and J. S. Schuman, “Evaluat- ing pulsatile ocular blood flow analysis in normal and treated glau- comatous eyes,” Amer. J. Ophthmol., vol. 136, no. 3, pp. 448–453, 2003. [19] Wolf -Heidegger, Atlas of Human Anatomy. Milan, Italy: Edi-Ermes, 2005. [20] CST Microwave Studio (2006) [Online]. Available: http://www.cst.com [21] International Commission on Non-Ionizing Radiation Protection, “Guide- lines for limiting exposure to time-varying electric, magnetic, and elec- tromagnetic fields (up to 300 GHz),” Health Phys., vol. 74, pp. 494–522, 1998. [22] C. Gabriel, “Compilation of the dielectric properties of body tissues at RF and microwave frequencies,” RFR Division, Brooks Air Force Base, San Antonio, TX, Final Tech. Rep. AL/OE-TR-1996–0037, 1996. [23] W. Dou and M. Y. W. Chia, “E-Shaped planar inverted-F antenna with low absorption for triple-band operation in cellular phones,” Int. J. Electron., vol. 88, no. 5, pp. 575–585, May 2001. [24] H. H. Pennes, “Analysis of tissue and arterial blood temperature in resting forearm,” J. Appl. Phys., vol. 1, pp. 93–122, 1948. [25] H. N. Kritikos, K. R. Foster, and H. P. Schwan, “Temperature profiles in spheres due to electromagnetic heating,” J. Microw. Power, vol. 16, no. 3/4, pp. 327–344, 1981. [26] J. A. Scott, “A finite element model of heat transport in the human eye,” Phy. Med. Biol., vol. 33, pp. 227–241, 1988. [27] J. J. W. Lagendijk, “A mathematical model to calculate temperature dis- tributions in human and rabbit eyes during hypertermic treatment,” Phys. Med. Biol., vol. 27, no. 11, pp. 1301–1311, 1982. [28] A. Hirata and T. Shiozawa, “Heat transportation models of the hu- man eye for microwave exposures,” presented at the Int. Workshop Ocular Side-Effects Non-Ionizing Radiation. Ishikawa, Japan, Nov. 2003. [29] C. Buccella, V. De Santis, and M. Feliziani, “Numerical prediction of SAR and thermal elevation in a 0.25-mm 3-D model of the human eye exposed to handheld transmitters,” in Proc. IEEE Int. Symp. EMC, Jul. 8–13, 2007, pp. 1–6. [30] W. F. Ganong, Review Medical Physics, 20th ed. New York: McGraw- Hill, 2002. [31] F. G. Shellock and J. V. Crues, “Corneal temperature changes induced by high-field-strength MR imaging with a head coil,” Radiology, vol. 167, pp. 809–811, 1988. [32] R. G. Gordon, R. B. Roemer, and S. M. Horvath, “A mathematical model of the human temperature regulatory system–transient cold exposure re- sponse,” IEEE Trans. Biomed. Eng., vol. BME-23, no. 6, pp. 434–444, Nov. 1976. [33] A. Hirata and T. Shiozawa, “Correlation of maximum temperature increase and peak SAR in the human head due to handset antennas,” IEEE Trans. Microw. Theory Tech., vol. 51, no. 7, pp. 1834–1841, Jul. 2003. BUCCELLA et al.: PREDICTION OF TEMPERATURE INCREASE IN HUMAN EYES DUE TO RF SOURCES 833 Concettina Buccella (M’92–SM’03) was born in L’Aquila, Italy. She received the Doctor degree in electrical engineering from the University of L’Aquila, L’Aquila, Italy, in 1988, and the Ph.D. de- gree in electrical engineering from the University of Rome “La Sapienza,” Rome, Italy, in 1994. In 1990, she joined the Department of Electri- cal Engineering, University of L’Aquila, where she is currently an Associate Professor. Her current re- search interests include electromagnetic compatibil- ity, power line communication, numerical methods, modeling techniques, lightning, microelectromechanical system, and ultra- wideband signal interferences. Dr. Buccella is a member of the Electrostatic Processes Committee of the IEEE Industry Applications Society. Valerio De Santis (M’05) was born in L’Aquila, Italy, on August 23, 1982. He received the telecom- munication engineering degree from the University of L’Aquila, L’Aquila, Italy, in 2005, where he is currently working toward the Ph.D. degree in the De- partment of Electrical and Computer Engineering. In 2006, he was with the University of Hannover, Hannover, Germany, where he was involved in a COST 286 European project for a Short Term Sci- entific Mission. From June to September 2007, he was a Visiting Researcher at the Motorola Corporate EME Research Laboratories, Plantation, FL. His current research interests in- clude biological effects of electromagnetic fields, electromagnetic compatibility, numerical methods and techniques, power line communication, and leaky line antennas. Mr. De Santis received the Best Student Paper Award at the IEEE EMC Inter- national Symposium, Honolulu, HI, in 2007, and the Second Best Student Paper Award at the Bioelectromagnetics Society Annual Meeting, Cancun, Mexico, in 2006. Mauro Feliziani (M’91–SM’00) received the degree in electrical engineering from the University of Rome La Sapienza, Rome, Italy, in 1983. Since 1994, he has been a Full Professor of electrical engineering at the University of L’Aquila, L’Aquila, Italy. He is the author or coauthor of many papers published in the fields of electromag- netic compatibility and in electromagnetic field nu- merical computation. His current research interests include ultra-wideband, wireless communications, power line communication, leaky line antennas, and microelectromechanical systems/film bulk acoustic resonators. Prof. Feliziani was the recipient of the Best Paper Award of the IEEE T RANSACTIONS ON I NDUSTRY A PPLICATIONS in 1995, the Electrostatics Process Committee and the EMC Europe Symposium, in 2000. From 1995 to 2000, he was an Associate Editor of the IEEE T RANSACTIONS ON E LECTROMAGNETIC C OMPATIBILITY . In March 2003, he was the Guest Editor of a special issue of the IEEE T RANSACTIONS ON M AGNETICS . He was the General Chairman of the EMC Europe Symposium, Sorrento, Italy, in 2002, and of the EMC Europe Workshop, Rome, in 2005. He is the Secretariat of the International Steering Committee of the EMC Europe. He has been the Program Committee Member, Editorial Board Member, Tutorial Session Organizer, Invited Speaker, and the Session Chairman of several international conferences.