Radiation Protection Dosimetry (2007), Vol. 124, No. 1, pp Advance Access publication 28 September 2007

Transcription

1 Radiation Protection Dosimetry (2007), Vol. 124, No. 1, pp Advance Access publication 28 September 2007 doi: /rpd/ncm370 EXPOSURE SET-UPS FOR IN VIVO EXPERIMENTS USING RADIAL WAVEGUIDES T. Reinhardt 1, *, A. Bitz 1, A. El Ouardi 1, J. Streckert 1, A. Sommer 2, A. Lerchl 2 and V. Hansen 1 1 Faculty of Electrical, Information and Media Engineering, Chair of Electromagnetic Theory, University of Wuppertal, Rainer-Gruenter-Str. 21, Wuppertal, Germany 2 School of Engineering and Science, Research II, Jacobs University Bremen ggmbh, Campus Ring 6, Bremen, Germany Radial waveguide set-ups are introduced as exposure devices for long-term experiments with large numbers of non-restrained animals exposed simultaneously. Methods are presented to ensure well-defined exposure conditions even for potentially overmoded waveguides and for the exposure of large groups of animals per cage. The proposed methods are applied for a four-generation study being performed on up to 2500 mice exposed to a generic UMTS test signal at prescribed averaged whole body specific absorption rates (SARs). The variation of the whole body SAR due to the movement of the mice inside the cage is calculated by using the finite-difference time-domain method and detailed animal models for selected configurations of the mice inside the cage for all stages of the study. INTRODUCTION In the framework of the German Mobile Telecommunication Research Programme, the Chair of Electromagnetic Theory of the University of Wuppertal is involved in three studies concerning the radio frequency (rf) exposure of freely moving animals kept in cages, which are performed in cooperation with the working group of Prof. Dr. A. Lerchl, Jacobs University Bremen, Germany. The first study Influence of 900 MHz electromagnetic fields on spontaneous leukaemia in AKR/J mice and the second study In vivo experiments on exposure to the high frequency electromagnetic fields of mobile telecommunication. B. Carcinogenesis were completed in 2004 and 2005 and the results were published in Sommer et al. (1,2) and Bitz et al., Lerchl et al. and Sommer et al. (3 5) respectively. In the third project Long-term study on the effects of UMTS signals on laboratory rodents, four generations [F0 F3] of mice were exposed to a generic UMTS test signal (6). The study will be completed in July One hundred and twenty-eight cages are equally distributed to four groups of averaged whole body specific absorption rate (SAR) of 1.3, 0.4, 0.08 and 0 W/kg. The flowchart of the study consists of consecutive stages with different numbers of animals per cage and with different exposure times: (1) Two pregnant mice and one male mouse per cage for 14 d. (2) One pregnant mouse and one male mouse kept in one cage for 7 d. (3) One female mouse together with one male mouse and six pups kept in one cage for 21 d. *Corresponding author: reinhardt@uni-wuppertal.de (4) Four mice kept in one cage until maturity and mating. This scheme is processed three times, i.e. four generations (F0 from the first day of pregnancy to F3 until the first litter) and in total 2688 mice are exposed during the study. The prescribed biological design of exposure of freely moving animals kept in cages allows for longterm exposure of up to 24 h d 21 and standard animal handling by the staff, but introduces variations of the exposure due to the following: (1) Different locations of the animals in the incident field, (2) Varying orientations ( postures) of the animals with respect to the direction of wave propagation and the polarisation of the field vector, (3) The superposition of the original incident field and the scattered field from neighbouring animals and (4) The age-dependent development of body mass. Whereas points (2) and (4) are inherent to the biological design and cannot be avoided, the effects of (1) and (3) can be reduced by a uniform, welldefined stable and thus reproducible exposure field inside the cage region. METHODS Exposure devices based on the concept of the radial waveguide are commonly used and well suited for the exposure of a large number of animals. Further, they comply well with all the above-mentioned requirements concerning the exposure field. Radial waveguides consist of parallel circular metal plates terminated at the radial boundary by absorbing # The Author Published by Oxford University Press. All rights reserved. For Permissions, please journals.permissions@oxfordjournals.org

2 material, which is mandatory for experiments with free-running animals in order to minimise the standing wave ratio. A rotational symmetrical cone antenna in the centre excites the radial propagating waves. The cages are arranged at a constant distance from the centre. Additional advantages of the radial waveguide are a uniform exposure due to the highly symmetrical configuration and a high ratio of power density to generated power. Since it is an electrically closed system, there is no need for additional shielding of laboratory, thus exposed and sham-exposed waveguide can be located close to each other. With regard to the numerical analysis, it is possible to describe the field distribution by only modelling one sector with proper boundary conditions containing field excitation, exposed animals in the cage and termination of the waveguide. For a height of the waveguide smaller than half a wavelength and a rotationally symmetrical excitation, the exposure field is given by the fundamental transverse electromagnetic (TEM) r -wave only described by the field components and E z ðr; w; zþ ¼ ja k2 0 H ð2þ 0 ðk 0 rþ ð1þ vm H w ðr; w; zþ ¼ A k 0 H ð2þ 0 0 ðk 0 rþ ð2þ m 0 The field strength is constant in the cross-section of the waveguide (r ¼ constant) and the dependency in radial direction is described by the Hankel functions H (2) (2) p 0 (k 0 r) and H 0 0 (k 0 r)(a is constant and k 0 ¼ ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi v 2 m represents the wave number in free space). In the actual study, the height of the radial waveguide is 17 cm and is chosen according to the height of the cages of approximately 15.5 cm, which is prescribed by the animal protection authority. Thus, the height of the waveguide exceeds the wavelength of about 15 cm at the UMTS-signal carrier frequency of 2 GHz. Therefore, the waveguide is able to guide higher-order modes with respect to the vertical direction in addition to the TEM wave, which can lead to an inhomogeneous field distribution in the exposure region and to instabilities due to interferences. Single-mode excitation and, by thus the uniformity of the exposure field, is assured by a conical transition, i.e. by increasing the height from smaller than half a wavelength in the centre of the radial waveguide, where the feed is located, to the required height in the exposure region (Figure 1). Another measure to provide stable exposure conditions is to decouple the field of adjacent cages by walls with surface impedance that tends to infinity. This measure is necessary because in some stages of T. REINHARDT ET AL. Figure 1. Design of a single mode excitation of the exposure region by conical transition. Figure 2. Sketch of sector with high impedance -walls; upper and lower metal plates are removed for better illustration. this study a large number of animals per cage are exposed. For such configurations, the scattered fields of the animals, which depend on the positions and postures of the animals, can in general affect the field distribution in adjacent cages. The high impedance walls consist of a multilayered structure of metallic webs and dielectric bars (1 r,m r ) covering a metal plate (Figure. 2). The careful optimisation of the thickness t of the high impedance walls with respect to the operating frequency is essential, because they operate properly in a very small band of around 80 MHz width. To optimise the following set of equations for the wall impedance, Z w and the optimum thickness t opt is used: qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi rffiffiffiffiffitan 2pt m opt =l 1 ðml=2wþ 2 Z W ¼jZ r 0 q ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi!j1 1 r 1 ðml=2wþ 2 qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi 1 witht opt ¼ 1 r m r ðml=2wþ 2 and p Z 0 ¼ ffiffiffiffiffiffiffiffiffiffiffiffi m 0 =1 0 ð3þ with length of the walls w and mode number m of the TEM 0 -wave inside the bars (TEM 0 -wave is transverse electromagnetic with respect to the thickness of the bars). The optimisation was performed by means of extensive computer simulations and supporting measurements in radial waveguides and equivalent rectangular waveguides. Figure 3 exemplarily, shows 22

3 EXPOSURE SET-UPS FOR IN VIVO EXPERIMENTS Figure 3. Computed field distribution for sector with high impedance walls. the computed electric field distribution inside a sector with high impedance walls. RESULTS Four stacks, each containing two radial waveguides are installed in the animal laboratory at IU Bremen (Figure 4). The waveguides have a diameter of 2 m and every device contains 16 cages. Lids with wire mesh above each cage ensure that light and gas can penetrate, whereas electromagnetic radiation is effectively shielded. A light shield on the top of each stack assures similar light conditions in the upper and lower device during the study. Figure 5 shows the result of field measurements in the eight radial waveguides. Each point represents the electric field value averaged over 76 measuring points inside a cage. The standard deviation of the measured field is 25% in each device. Figure 4. Four stacks with two radial waveguides each implemented at IU Bremen. 23

4 T. REINHARDT ET AL. Figure 5. Results of the field measurements in the eight radial waveguides. In order to analyse the variation of the whole body SAR, it must be taken into account that the mice can freely move inside their cage. Therefore, different configurations of mice per cage with different positions and postures towards the direction of the incident wave have to be considered. The decision which constellations of mice in a cage should be considered for a representative numerical dosimetry is not trivial, since the movement of the animals is a dynamic, non-deterministic process. Some constellations are more typical than others, even small variations of a configuration can change the SAR distribution, and the number of computer runs must be limited. Ideally, many calculations and a weighting of different configurations according to their occurrence might be reasonable. However, for a reliable statistic, the required data are not available. For a practicable assessment of the whole body SAR without weighting, typical configurations were identified in agreement with the IU Bremen (e.g. agglomerating mice, pubs close to the dam), and additionally, some configurations known as worst cases were included in order to avoid an underestimation of the SAR (e.g. mice erected on the hind legs, isolated animals). Since the material parameters are measured quantities, an uncertainty assessment was done and calculations with altered material parameters (+10%) were made for each stage of the study considering the typical configurations of mice. For the numerical calculations, in-house and commercial FDTD-codes (7) with an implemented subgridding algorithm are used. The mouse models were scaled from anatomical models for a pregnant rat and for a male 7 d old rat, which were developed by IT IS, Zürich. The electric material parameters were chosen according to the Brooks database (8). Exemplarily, a configuration of eight mice inside a cage is shown in Figure 6. Figure 7 shows the field distribution inside the mouse located in front of the cage and the corresponding SAR distribution. Standing waves inside the cage arise due to reflections at the mice and due to reflections at the absorber. The prescribed SAR values of 1.3, 0.4, 0.08 W/kg were adjusted according to the first stage of the study (three adult mice per cage) and the corresponding input power was kept constant during all stages because of the animals rapid growth. 24

5 EXPOSURE SET-UPS FOR IN VIVO EXPERIMENTS Figure 6. Exemplarily for the third stage of the study: eight mice are placed inside a cage. Figure 7. Field distribution and the corresponding SAR distribution inside the mouse located in the front of the cage. Table 1. Summary of dosimetric data during all stages of the study. Stage of the study Mice per cage s (%) SAR max,local / SAR adjusted,wb 1 Three adults Two adults Two adults Six pups 54 4 Four youngs g of contiguous tissue that occurs in one animal under consideration of all calculated configurations and related to the adjusted SAR value. From this, it is guaranteed that the experiments are performed below the maximum ratio of local SAR towards whole body SAR which is permitted, according to the guidelines for human exposure. The variation of the material parameters shows that the changes of the whole body SAR values lie within the standard deviation. Table 1 summarises the standard deviation of the whole body SAR for the different configurations and gives the maximum local SAR value averaged over CONCLUSION Radial waveguide rf-exposure set-ups for long-term experiments on non-restrained animals kept in cages 25

6 offer well-defined exposure conditions, although the biological design introduces variations of the whole body SAR. These variations can be minimised by implementation of special measures, e.g. conical transition region for single-mode excitation of the cage region with the fundamental TEM mode and decoupling of adjacent sectors by high impedance walls (corrugated metal walls with dielectric inserts) which do not alter the propagation properties of the TEM wave. During the exposure stages of the four-generation study being performed, the maximum standard deviation for the whole body SAR is 54% in case of six pups kept inside a cage and the minimum standard deviation is 31% for three adult mice per cage. FUNDING This work was funded by the German Federal Office for Radiation Protection (BfS) and the German Federal Ministry for the Environment, Nature Conservation and Nuclear Safety (BMU) within the German Mobile Telecommunication Research Program. REFERENCES 1. Sommer, A. M., Streckert, J., Bitz, A. K., Hansen, V. W. and Lerchl, A. No effects of GSM-modulated 900 MHz electromagnetic fields on survival rate and spontaneous development of lymphoma in female AKR/J mice. T. REINHARDT ET AL. BMC Cancer. 4, 77, (2004). Available on biomedcentral.com/ /4/ Sommer, A. M., Bitz, A. K., Streckert, J., Hansen, V. W. and Lerchl, A. No effect from 900 MHz electromagnetic fields on the spontaneous development of lymphoma in female AKR/J mice. In: 26th BEMS Annual Meeting, Washington, DC, June 2004, pp Bitz, A. K., Streckert, J., Sommer, A. M., Lerchl, A. and Hansen, V. W. 2 GHz-exposure of non-restrained AKR/J mice in a slightly over-moded radial waveguide. In: 26th BEMS Annual Meeting, Washington, DC, June 2004, p Lerchl, A., Sommer, A. M., Bitz, A., Streckert, J. and Hansen, V. UMTS-modulated electromagnetic fields do not influence the development of lymphoma in female AKR/J mice. In: 27th BEMS Annual Meeting, Dublin, Ireland, June 2005, p Sommer, A. M., Lerchl, A., Bitz, A. K., Streckert, J. and Hansen,V.W.UMTS-modulated electromagnetic fields do not affect hematological or histological parameters in lymphoma-prone mice. In: 28th BEMS Annual Meeting, Cancun, Mexico, June 2006, pp Ndoumbè Mbonjo Mbonjo, H., Streckert, J., Bitz, A., Hansen, V., Glasmachers, A., Gencol, S. and Rozic, D. A generic UMTS test signal for RF bio-electromagnetic studies. Bioelectromagnetics. 21(6), , Mafia Version 3.20, CST GmbH, Darmstadt, Germany, Gabriel, C. Compilation of the dielectric properties of body tissues at RF and microwave frequencies. Brooks Air Force Technical Report, 1996, AL/OE-TR Available on DIELECTRIC/Report.html 26