LAUSR

Dear Editors: NEUFELD AND Kuster (2018) calculate “short temperature spikes in the skin” of people exposed to “bursts of a fewmilliseconds to seconds” of radiofrequency (RF) energy from wireless devices such as fifth generation (5G) communications and Internet of Things (IoT) devices, and they discuss what they consider to be inadequacies of RF exposure limits. But their use of a grossly inappropriate thermal model leads to extreme overestimates that negate the usefulness of the analysis for setting exposure limits for short pulses. For background, RF exposure limits above 100 kHz are expressed in terms of exposure averaged over defined intervals of time to account for the thermal inertia (resistance to change in temperature with time) of the body. While the averaging times vary with the limit, two major international limits, those of the International Commission on Non-ionizing Radiation Protection (ICNIRP) and IEEE (standard C95.1-2019) provide averaging times of 6 min for localized exposures. At millimeter-wave frequencies (30–300 GHz), local-body limits in IEEEC95.1-2019 range from 100–150Wm for “persons in restricted environments” (e.g., occupational limits); limits for “unrestricted environments” (e.g., for the general public) are a factor of 5 smaller (IEEE 2019). ICNIRP limits are similar, but as of this writing the most recent edition has not been finalized and is not discussed further. Neufeld and Kuster (2018) consider a worst-case exposure situation in which the “entire deliverable power over one time-averaging interval is deposited during a small fraction a of the averaging period Dt at an intensity I0.” In other words, they consider pulses of incident power density I0/a and fluence (total energy per unit area carried by the pulse) of I0Dt, where I0 is the exposure limit for continuous-wave (CW) exposure and Dt is the averaging time in the limit. For the limits in IEEE C95.1-2019 for “restricted environments” (e.g., occupational exposures), this would correspond to millimeter-wave pulses with fluence ranging from 54 to 36 kJ m between 30–300 GHz, depending on frequency. Limits for “unrestricted environments” (e.g., for the general public) are a factor of 5 smaller. Neufeld and Kuster base their analysis on the surface heating model introduced for millimeter waves (30–300 GHz) by Foster et al. (2016, 2017, 2018). The model assumes that energy is absorbed at the very surface of tissue, which is a reasonable approximation for many purposes given the very short energy-penetration depth of millimeter waves (0.4 to 0.1 mm between 30 and 300 GHz). However, the model has a singularity in its impulse response and is unphysical in some respects (Foster et al. 2018). Assuming adiabatic boundary conditions, the temperature increase from exposure I0u(t), where u(t) is the unit step function, in the surface heating model can bewritten (Foster et al. 2017):