We first performed a preliminary survey of university libraries in Japan to identify commonly used EAS gates. The survey suggested that most of the university libraries with EAS gates are equipped with EM-EAS gates. We selected five models with relatively high adoption rates. These five models cover about 90% of the EAS gates used in university libraries in Japan.

Table 1 summarizes the frequency and external specifications of the investigated five EM-EAS gates G1–G5. It should be noted that the operating frequency ranges from 220 Hz to 14 kHz and the frequencies of G4 and G5 are 220 Hz, which is not in the IF region (> 300 Hz). Those are included in this study considering the common nature of the exposure.

Those five models have built-in coils (current-carrying coils) inside the gateposts. The coils are excited by continuous sinusoidal currents to generate sinusoidal magnetic fields of continuous waves. Although their frequencies are different as shown in Table 1, the magnetic fields generated by the coils have similar characteristics regarding spatial distributions. Normally, two gateposts are erected vertically, facing each other at a predetermined interval. The monitoring zone is formed between the gateposts to detect the magnetic tags.

It was confirmed by the measurement that the waveforms of magnetic fields from EAS gates are continuous sinusoidal waves at frequencies nearly equal to the nominal frequencies shown in Table 1. Figure 3 shows the distribution of the resultant rms values of magnetic flux density for G1 (366 Hz) in zone A. Measured points were on the 10 cm grid and the measured data were interpolated to 2 mm resolution using bi-cubic interpolation to apply to dosimetry calculations with 2 mm resolution. The maximum was 430 μT, the average 100 μT, and the minimum 30 μT in zone A. The maximum, average, and minimum for G2 (14 kHz) in zone A were also obtained and were 439, 128, 19 μT, respectively.

Figure 4 illustrates ellipticity in zone A for G1. The ellipticity is larger near the center of the gate while it is small (or close to linearly polarized) near the gatepost (see also Figure 8). It should be noted that significant ellipticity is found only in the region where the magnetic flux density is rather small.

Measurements were carried out for G2 in zone B, or just outside of the gate aisle. The maximum, minimum, and average values were 371, 70, 8 μT, respectively. It was suggested that similar exposure occurs either passing through the aisle or passing by the gatepost.

The IEC averages for G1–G5 were 87, 111, 68, 87, and 105 μT, respectively. The ratios to ICNIRP reference levels range from 34 to 411 % for the general public and 8.2–111 % for occupational exposures as shown in Table 3. The results confirmed that EM-EAS gates are sources of significantly strong IF-EMF encountered in daily lives.

Figure 5 shows the distribution of rms magnetic flux density in zone D for G2 in contour lines. The magnetic flux density is approximately constant in the vertical direction and decays steeply with distance.

Dependency on the distance to the nearest gatepost is shown in Figure 6. The magnetic flux density is normalized by the IEC average B_IEC to allow comparison with different gates. The normalized magnetic flux density depends similarly on the distance to the nearest gatepost regardless of the direction. The dependency is approximated by Equation (3).

where r is the distance to the nearest gatepost in cm and B_R(r) is the resultant magnetic flux density at the distance r. The coefficient of determination R² is 0.976 for this approximation. The approximation is valid in the region r > 50 cm. The magnetic flux density at r = 50 cm is nearly equal to B_IEC. It decays to 13% of B_IEC at 1 m from the nearest gatepost. It becomes 1.7 % at 2 m, and 0.6 % at 3 m distance from the nearest gatepost.

The data measured in situ for G3 and G4 are superimposed on Figure 6. It is found that the magnetic flux density normalized by B_IEC is well approximated by Equation (3) for different EAS gate models.

Equation 3 suggests that magnetic field from EAS gates decays in proportion to the approximately third power of the distance. The dependency is similar to the magnetic field from an equivalent dipole source (27). The approximation holds in the region r > 50 cm while the source is larger than the distance where the approximation holds. It is not natural for the large current source to be equivalent to a dipole source having the third power decay, but the relationship holds for the data measured for three different EAS gates consisting of two gateposts. The approximation formula is empirically obtained from a limited number of samples but it may be useful for estimation of exposure in a room equipped with EAS gates. It should be noted that the dependency is well approximated by Equation 3 when the distance is measured to the closer post of the pair of gateposts.

Figure 7 shows the calculated result of induced electric field E_in in the human body passing through gate G1. The human model is located in zone A. The 99th and 99.9th percentile values (p99 and p99.9) in the whole body were 0.017 and 0.032 V/m, respectively, for G1. The p99 and p99.9 for G2 were 0.88 and 1.72 V/m, respectively. The values will be discussed in the next section.

The magnetic fields from the EM-EAS gates are not linearly but elliptically polarized. Hence, the induced electric field in the body should also be elliptically polarized. Figure 8 shows the distributions of the ellipticity of incident magnetic flux density (left) and the ellipticity of induced electric field (right) in the mid-coronal and mid-sagittal sections of the body. It is confirmed that an elliptically polarized electric field is induced in tissue.

Figure 9 shows the frequency of ellipticity in voxels in the human model as a function of ellipticity and induced resultant electric field in the decile scale. Small ellipticity is significantly frequent in the 10th decile of the induced electric field in the tissue. This means that ellipticity is rather small in the tissue where the induced electric field is large.

The induced electric field is also calculated for a human passing beside the gate in zone B for G2 to compare the cases between passing inside and outside of the gate aisle. The p99 and p99.9 values were 0.95 and 1.57 V/m, respectively, for zone B while they were 0.88 and 1.72 V/m for zone A as mentioned above. The results suggested that exposure is almost of the same level in terms of the induced electric field between the cases during passing inside and outside of the gate aisle. The detailed comparison between exposures in zones A and zone B for G2 is shown in Table 4.

The induced electric field is proportional to both frequency and magnitude of the magnetic field (represented by B_IEC). Then, the induced electric field is represented by Equation (4).

where k is a constant, f is the operating frequency in Hz, and B_IEC is the IEC average in μT of the gate of interest. The proportional constant k corresponding to, for example, the p99 is expected to be close to each other assuming that the spatial distributions of incident magnetic flux density are similar for different gates and the electric constants of tissues are not significantly dependent on frequency. We have calculated the p99 of induced electric fields E_in = 0.017 V/m for G1 (f = 366 Hz, B_IEC=87 μT) and E_in= 0.88 V/m for G2 (f = 14 kHz, B_IEC=111 μT). Then the constant k for the p99 is obtained as 5.34 × 10⁻⁷ (V m⁻¹Hz⁻¹μT⁻¹) from the values for G1, and 5.66 × 10⁻⁷ from the values for G2. There is a 6% difference between the estimated k values but they are reasonably close. The result suggests that the maximum (e.g., p99) induced electric field should be roughly estimated for different EAS gates without detailed measurements and dosimetry calculations if B_IEC is obtained. It should be noted that the constant k may not be so close if the configuration of the coils is different from each other.

The exposure levels near the gates are compared with the ICNIRP guidelines (21). Table 3 summarizes the obtained data and results of the comparison. Table 3 also includes the spatial maximum/average/minimum of incident magnetic flux density B_inc, that is, magnetic flux density in the space where the human body is supposed to occupy. The human body is assumed here to be at the location described in 2.3.2 in zone A. The average of B_inc is the quantity to be compared with the reference level for non-uniform EMF exposures.

Regarding the reference levels, the maximum magnetic flux density in zone A exceeds the reference level for the general public both for G1 and G2. The spatial average of incident field B_inc for G1 does not exceed the reference level but that of G2 still exceeds the reference level for the general public.

B_IEC is a quantity to be used for compliance assessment instead of the spatial average of B_inc according to the standardized assessment procedure. The measured values of B_IEC for G1 and G2 are both close to but sufficiently larger (i.e., conservative) than the average of B_inc. This suggests that B_IEC is a useful quantity to evaluate exposure due to EAS gates. For G3– G5, the values of B_IEC are well below the reference level.

We defined exposure ratio to reference level (ER_RL) as B_IEC / B_RL,where B_RL is the reference level of ICNIRP guidelines (21). This index is useful for whole-body uniform exposure but not relevant for localized exposures such as exposure of humans passing the EAS gate. The exposure ratio of G2 to the reference level exceeded 100% of the reference levels for both general public and occupational exposures.

We defined the exposure ratio to basic restriction (ER_BR) as E_in / E_BR. This index is useful to evaluate exposures including localized exposures. The exposure ratios of G1 and G2 to basic restriction did not exceed 100% of the basic restrictions for both general public and occupational exposures.

It was shown that magnetic fields from EAS gates are elliptically rotating fields. It should be noted that the orientation of the magnetic field relative to the human body changes with time. This characteristic results in different coupling with the body from linearly polarized fields. In addition, some studies had suggested that rotating fields might specifically affect melatonin secretion of animals (28) while no change was found for linearly polarized fields (29). Though the evidence was evaluated weak due to the problem of experimental conditions (30), polarization may be taken into account for the health risk assessment of magnetic field exposures.

This study has been initiated to contribute to epidemiological studies on IF-EMF exposures of library workers. The results of this study provide useful information about the exposure assessment of library workers due to EAS gates.

Library workers have opportunities to be exposed to IF-EMF from EAS gates when passing through or by the gate. The exposure, in this case, is relatively high-level transient exposure repeated many times a day for a long term. It should be noted that exposures are similar between passing through and passing by the gate. Exposure ratio to basic restriction ER_BR was obtained and it will provide a useful index for this type of exposure, which allows comparing exposures from EAS gates with different operating frequencies and different magnitudes of magnetic fields.

There is another type of exposure for library workers at their working desks. It was found that IF-EMF from the EAS gate spreads a few meters from the gatepost. Library workers can be exposed the whole day long if the working desk is located within a few meters from the gate. This type of exposure can be evaluated based on the magnetic flux density estimated by Equation (3) described in 3.3.

In addition to these exposures, library workers are exposed to magnetic fields from the activator/deactivator of anti-theft magnetic strips for the EAS systems attached to the books (book-check unit, BCU). The device generates strong pulsed magnetic fields. The pulsed fields include IF components. Library workers operate BCU many times a day to cause repeated exposure to the pulsed magnetic field. This type of exposure should also be taken into account when an epidemiological study is planned. This paper does not cover assessment for this type of exposure but we will report it when we get another opportunity.

It is a challenge for exposure assessment of library workers to IF-EMF how to evaluate the combined exposure of those different types of exposures in epidemiological studies.