Updated 03/08/2022
Abstract: This note is a compilation of various research studies performed in the late 1980s. This research was focused on the radio frequency radiation levels and conditions that could cause electric shock or tissue burns. Several instruments were used and various frequency levels were tested. The results were differentiated between male and female responses. The results of these and later research studies were used by regulatory agencies to provide current safe guidelines to
radio frequency contact. As with other research of this kind, the researchers suggested that further studies would be beneficial as more information was gleaned regarding shock and burn in humans. Indeed, subsequent research to 2018 yielded more information, but didn’t fundamentally change the information presented herein.
In the past 30 years research has been undertaken in an effort to provide safe guidelines for human exposure to radio frequency radiation. These scientific studies were also designed to form a basis on which future studies could be built. Furthermore, regulatory entities such as ANSI/IEEE have used the findings of these studies in developing safety guidelines for human exposure to radio frequency radiation. Much more is known about radio frequency radiation today but there remain questions to be answered. The following information is an overview of the scientific research to date on an often unappreciated aspect – the human shock and burn response.
Although it was known that radio frequency radiation could cause electric shock in the body or burns in tissue under certain circumstances, specific exposure limits were not included in previous standards for human exposure to radio frequency radiation. Guy (1985), however, noted that such effects were considered in choosing the 300-kHz lower frequency limit of the present ANSI (1982) standard.

Radio Frequency arc in crane cable hook due to induced current from AM station
Several studies were conducted recently to determine the radio frequency radiation levels and conditions that could cause electric shock or tissue burns, and it is now likely that the findings of such studies will serve as the basis for specifying, in future radio frequency radiation exposure standards, the appropriate electrical parameters and their maximum permissible levels under stated exposure conditions to avoid such effects.
Rogers (1981) stated the problem as follows: "When a person touches an electrically energized [sic] object, he may experience an adverse effect. If the object is energized by a radio frequency (RF) source, the predominant contact hazard is burning of tissue at the point of contact and arises when the current drawn from the object exceeds a certain value. This RF burn hazard exists on various transmitting aerials and simple precautions can be taken to avoid it. However, such a hazard can also arise on metallic objects excited by radiation from transmitting aerials in their vicinity and this paper is devoted to this aspect."
In the paper, a simple apparatus ("RF Burn Hazard Meter") was described to measure the RF currents passing through a human in shoes standing on a ground plane. Part of the apparatus consisted of a brass tube excited by an RF source, with the source connected to the ground plane. When the subject touched the tube with a finger, a loop consisting of the source, tube, body of the subject, and ground plane was closed to form the primary of a current transformer (single turn). The rest of the apparatus was the transformer secondary winding and its connections via diode detectors, resistors, and capacitors to a dc current meter. Using this apparatus, the current levels that yielded a barely perceptible sensation ("perception" current) and that caused discomfort ("let-go" or "hazard" current) were measured for frequencies in the MF (0.3-3 MHz) and HF (3-30 MHz) bands.
The author indicated that the perception current and let-go current for contact with the tip of the forefinger were both about twice those for contact with the back of the forefinger, and were even higher for large-area contact with the palm. The results for 50 persons tested (with the back of the forefinger) showed a mean hazard threshold current of about 200 mA for the band 2-20 MHz.
The paper was devoted primarily to possible shipboard hazards to humans from metallic structures in the vicinity of onboard radiating antennas. Included were measured and calculated data for various structures and distances. The author concluded: "The measurements described indicate an RF burn hazard threshold of about 200 mA for the HF band and show that many shipboard structures can be excited sufficiently by their own ship transmissions in this band to present rf burn hazards to personnel. They show also that cranes can be potent sources of rf burn hazards…It is to be noted that rf burn hazards are present on structures when irradiated at field strengths much lower than the maximum permissible for human exposure. For example the measurements on the crane reported above show that the electric field for rf burn hazard threshold is about 10 V/m compared, for example, to the American National Standards Institute for the band 3-30 MHz." (The author was referring to the ANSI (1974) standard, which specified a maximum electric field of 200 V/m for the frequency range 10 MHz to 100 GHz).
Gandhi and Chatterjee (1982) used the quasi-static approximation (Deno, 1974; Bracken, 1976) to calculate the short-circuit currents induced in metallic objects (a 2.44-m x 1.22-m metal roof, a 50-ft metal fence, a compact car, and a fork-lift truck) and in a human (height 1.75 m, mass 68 kg), when each object is in a vertically polarized electric field at frequencies in the range 10 kHz to 10 MHz, and with each object assumed to be isolated from ground by 5 cm of insulation. They then calculated the incident electric fields necessary to produce threshold-perception and let-go currents for a human in conductive finger contact with each object. The threshold-perception current was defined as the smallest current that produces a tingling or pricking sensation due to nerve stimulation. The authors noted that the sensation changes from tingling to internal heat at frequencies above approximately 100-200 kHz. Let-go current was defined as the maximum value at which a human can still release an energized conductor with muscles directly stimulated by that current.
The authors used the experimental data on human perception-threshold and let-go currents of Dalzeil and Mansfield (1950), Dalzeil and Lee (1969), and Rogers (1981) for their calculations. These data were reproduced as log-log plots of perception-threshold and let-go currents vs frequency. The perception-threshold current showed a linear rise from about 0.4 mA at 10 KHz to about 14 mA at 150 kHz with a slower rise to about 100 mA at 20 MHz. The let-go current also showed a linear rise, from 6.4 mA at 10 kHz to about 85 mA at 150 kHz, and a slower linear rise to about 200 mA at 20 MHz.
By using the value of capacitance-to-ground for each object and ratio of its electrostatically coupled short-circuit current to the unperturbed vertical field measured at 60 Hz (Deno, 1974; Bracken, 1976), Gandhi and Chatterjee (1982) obtained the effective area (S) and height (h) of the object. They assumed that these values of S and h are also reasonably valid for frequencies in the range 10 kHz to 3 MHz because the fields are quasi-static for objects of largest dimension much smaller than the free-space wavelength.
Log-log plots of the calculated values of unperturbed electric field (E) necessary to create threshold-perception and let-go currents in a human in finger contact with each object were presented. For each object, E for threshold-perception was constant in the range approximately 10-100 kHz: about 250, 160, 80, and 20 V/m for the roof, fence, car, and truck, respectively; in the range 10 kHz to 10 MHz, E for the car and truck did not change substantially but those for the roof and fence decreased to about 35 and 20 V/m at 10 MHz, respectively. The plots of E for let-go current vs frequency were similar: the plateaus for the roof, fence, car, and truck in the range 10-100 kHz were about 1040, 850, 440, and 110 V/m, respectively, with diminution for the roof and fence to less than 100 V/m at 10 MHz and smaller decreases for the car and truck.
The authors stated: "A simple analysis based on the equivalent circuit representation of a human in conductive contact with an ungrounded, metallic object in a quasi-static HF field points out that there may be situations where the thresholds of perception and let-go can be exceeded for fields considerably lower than the ANSI recommended guideline of 615 [sic] V/m, the far-field equivalent E-field associated with a power density of 100 mW/sq cm in the frequency band 0.3 to 3.0 MHz [ANSI, 1982]…The above effects will not occur if the conducting objects are grounded or insulated at the points of possible contact."
Chatterjee et al. (1986) measured the complex body impedance (magnitude and phase) and the threshold currents for perception and pain for 197 men and 170 women between the ages of 18 and 70 years for the frequency range 10 kHz to 3 MHz. They defined the threshold-perception current for pain as the smallest current for which the subject reported "very uncomfortable sensations (similar to but more intense than that for perception) for which he/she will definitely not continue to touch the electrode any longer."
Mean body-impedance data (and SDs) vs frequency for barefoot subjects standing on a ground plane and grasping a brass-rod electrode that was insulated from the ground plane were shown separately for men and women. Both the magnitude and the phase decreased monotonically with frequency, but at corresponding frequencies, the impedance magnitude for women was significantly higher than for men. At 10 kHz, for example, the mean values for women and men were respectively about 630 and 520 ohms. The difference in mean phase at each frequency was not significant. The results for the subjects when they used an index finger moistened with 0.9% saline to touch a metal-plate electrode insulated from the ground plane were qualitatively similar, but of much higher magnitudes, about 900 and 1700 ohms at 10 kHz, respectively.
For men, the mean threshold-perception currents for finger contact rose linearly with frequency from about 4 mA at 10 kHz to about 40 mA at 100 kHz and remained at the latter value from 100 kHz to 3 MHz. The curve for women was parallel to that for men, but about 25% lower; by analysis of variance, the difference was highly significant. The results for grasping contact were similar. The curves of finger-contact threshold currents for pain vs frequency also rose roughly linearly to maxima at about 100 kHz, but diminished slightly with frequency in the range 100 kHz to 3 MHz. At 10 kHz, the mean pain-threshold currents for men and women were respectively about 10 and 6.5 mA, but their maxima at 100 kHz were nearly the same, about 14.5 mA. The authors noted that the values of threshold current for 10-year-old children could be obtained from those for male adults by using a scaling factor of about 60%.
The sensation reported by the 367 men and women for frequencies below 100 kHz was tingling or pricking, localized in the area adjacent to the region of contact on the finger or hand; for frequencies above 100 kHz and finger contact with the plate electrode, the sensation was warmth or heat in the area below and around the plate electrode; with grasping contact, warmth or heat was felt in the hand and wrist. To determine more accurately the frequency for transition from tingling to warmth, data were obtained for some subjects at 50 and 70 kHz. At 50 kHz, the sensation reported was always tingling, but at 70 kHz, some subjects reported tingling and others warmth. Moreover, when the current was raised slightly for those who reported tingling, the sensation changed to warmth. In addition, for frequencies above 100 kHz at which warmth was felt when the current was adjusted to be equal to the perception-threshold, pain was reported typically within 10-20 seconds, an effect that was not observed for frequencies below 100 kHz.
Also determined in this study were the short-circuit currents, at local AM broadcast stations (operating at 630, 700, and 1500 kHz) and at Coast Guard and Navy communication antenna sites in Hawaii (operating at 13.6, 23.4, 146, and 3105 kHz), induced in humans while barefoot or wearing safety, leather-soled, or rubber-soled street shoes; the short-circuit currents induced in various vehicles; and the currents induced in humans in contact with these vehicles. Among the findings were that electrical safety shoes and gloves respectively provide protection that is adequate only at frequencies less than about 1 and 3 MHz. The measurements of induced short-circuit current showed reductions to 55% of the barefoot values with rubber-soled shoes, 63% with safety shoes, and 85% with leather-soled shoes.
The mean body impedances and threshold-perception currents were used to calculate the E-fields for threshold-perception by grounded humans in finger contact with a compact car, van, and school bus, and the results for adult males and 10-year-old children were plotted vs frequency in one set of graphs, and for adult females in another set. Similar sets of curves were obtained for threshold-perception E-fields with grasping contact. Also presented were similar sets of curves of the threshold E-fields for pain vs frequency for finger contact with such vehicles.
The threshold-perception curves for finger contact by men, women, and children were all entirely below 632 V/m, the value recommended in ANSI (1982) for the range 0.3-3.0 MHz, with peaks at approximately100 kHz in ascending order respectively for the school bus, van, and compact car. (The curves for the bus and van crossed at frequencies above 100 kHz.) Thus, currents from all three vehicles could be sensed at fields smaller than those recommended by ANSI for the range 0.3-3.0 MHz. For grasping contact, the school-bus threshold-perception curves for all three groups were also below the ANSI value over the entire frequency range; those for the other vehicles were below that level except within frequency ranges of various sizes encompassing their respective peaks at 100 kHz.
The curves of pain-threshold E-field for finger contact with each type of vehicle were entirely below 632 V/m except for the curve for men in contact with the compact car. That curve exceeded 632 V/m in the range approximately 10-80 kHz, with a maximum of about 850 V/m at 30 kHz. All the other curves also had broader maxima at frequencies less than 100 kHz than those at 100 kHz for threshold perception.
The authors had measured the capacitance-to-ground of a GMC van that was well insulated from ground at a local AM broadcasting station operating at 700 kHz. The result was 1045 pf, which they used in a calculation of the current through the hand of a grounded human in conductive contact with the handle of such a van within a 3-MHz, 632-V/m field and obtained 879 mA. Based on this result and on an effective cross-sectional area of 11.1 sq cm for the wrist, they estimated that the corresponding local SAR in the wrist would be about 1045 W/kg.
The results of this study were presented in considerably more detail in the final report by Gandhi et al. (1985a). The very high SAR value above for the wrist was not mentioned. However, given in Appendix B of that report (although not directly related to possible shock and burn hazards from contact with metallic objects) were formulas for calculating local SARs in cross sections of the human leg (ankles, just below and above the knee, and two other thigh locations) for a human (*barefoot and with safety shoes) immersed in a vertical field at the levels specified in ANSI (1982) for the range 0.3-30 MHz. (These levels were 632 V/m for the frequencies f less than 0.3 MHz and 1897/f for f in the range 3-30 MHz.)
The formulas include the currents induced in the bodies, derived by the quasi-static approximation, and the effective cross-sectional areas of interest, derived from the geometric areas and specific conductivities of the tissues involved. Also presented were experimentally determined human values of normalized current (in mA per V/m that demonstrated the validity of the quasi-static approximation to frequencies up to about 40 MHz. With these formulas, calculations based on assuming that half the body current flows through each leg indicated that SARs as high as 182 W/kg could occur in the ankle region. These results were also published (Gandhi et al., 1985b).
Guy and Chou (1985) performed an extensive study on possible hazards to humans from exposure to fields in the VLF-MF (10 kHz to 3 MHz) range, directed principally toward quantitation of thresholds and establishment of safety standards against such hazards. They noted that though SAR is generally used to quantify internal energy absorption, other quantities may be more important in the VLF-MF band for this purpose, because the amounts of energy absorbed by humans exposed to fields in that frequency range are relatively low but can cause direct neuromuscular effects from electric shock, and local tissue damage may result from electric contact between subject and metallic objects in the field. In addition to SAR, the important quantities include total electric current, I, in the body resulting from exposure while in contact with objects or surfaces in free space and current density, J, through various cross sections of the body.
The authors also noted that maximum energy coupling occurs when the longest dimension of the body is parallel to the electric vector, and that for exposure in this orientation to plane waves and most VLF-MF antenna sources, absorption from the magnetic component is more than an order of magnitude smaller than absorption from the electric component, so any restrictions on electric-field exposure will ensure restrictions on the corresponding magnetic fields that are more conservative by at least a factor of ten.
In this report, the prior work on possible shock and burn hazards was reviewed, and vast quantities of experimental and calculated data were presented, including:
1. Impedances for humans and various vehicles, and for humans in contact with such vehicles under various ground conditions.
2. Measurements of open-circuit voltage and short-circuit current for humans and vehicles in high-strength electric fields, with the humans barefoot or wearing standard leather-soled shoes, standard rubber-soled shoes, sandals, or stockings.
3. Impedance distributions along the axis of the body and limbs vs frequency for 275 adult men and women of varying ages, shapes, and sizes.
4. Weight and heights of the populations above, and circumference and diameter of middle finger, legs, arms, torso, shoulders, neck, and head at 5-cm intervals along their body and limb axes.
5. Thresholds for electrical-stimulation perception vs frequency for the finger, arm, and ankle.
6. Threshold currents and current densities for perception of electric shock by the average-sensitive, 0.5-percentile least-sensitive, and 0.5-percentile most-sensitive subpopulations.
7. Induced body currents, current densities, SAR distributions, and average SARs in all parts of the bodies of some subjects in the low-, medium-, and high-weight ranges for exposure to fields under various conditions, including: free space, feet grounded, hand grounded, and hand in contact with an object drawing 1 mA of current.
8. Current flow in the bodies of subjects when in contact with various vehicles in VLF and MF fields at field strengths measured from ground level to a height of 7 ft.
9. Human exposures at the following sites to the frequencies indicated:
• 10.2 kHz Haiku, HI
• 23.4 kHz Lualualei, HI
• 24.8 kHz Jim Creek, WA
• 146 kHz Lualualei, HI
• 1 MHz KOMO Radio, Vashon Island, WA
An important finding was that highest local SARs occur in the ankles of the subjects, a result also found by Gandhi et al (1985a). In the rationale of the ANSI (1982) guidelines, exposures at maximum local spatial SARs as high as 8 W/kg (averaged over any gram of tissue and over any 0.1-hr period) would be permitted, provided that the whole-body-averaged SAR does not exceed 0.4 W/kg. In this context, Guy and Chou (1985) noted that exposure to fields in the VLF-MF range would have to be restricted to 97 V/m to avoid exceeding the 8-W/kg limit.
In their conclusions, Guy and Chou (1985) stated: "At the present time this study is far from complete and will require additional work before any final conclusions can be made concerning safety guidelines for the VLF-MF frequency range." They also noted: "Based on the measurements carried out under this contract, it appears that not enough attention has been placed on spark discharge hazards. This type of insult can produce a very unpleasant stimulus and an involuntary startle reaction which can occur at exposure fields much less than that required to produce a perceived steady state current:
In a presentation, Guy (1985) summarized the work described in Guy and Chou (1985) as well as that of other investigators on this topic.
As suggested at the beginning of this section, exposure standards, specifically the 1987 ANSI standard, in force and those under development, will undoubtedly include maximum permissible levels in the VLF-MF range for avoidance of shock and burn hazards.
REFERENCES:
ANSI (American National Standards Institute), C95.1-1974. "Safety Level of Electromagnetic Radiation With Respect to Personnel."
Published by the Institute of Electrical and Electronics Engineers, New York (1974).
ANSI, C95.1-1982. "Safety Levels With Respect To Human Exposure to Radiofrequency Electromagnetic Fields, 300 KHZ TO 100 GHZ."
Published by the Institute of Electrical and Electronics Engineers, New York (1982).
Bracken, T.D. "Field Measurements and Calculations of Electrostatic Effects of Overhead Transmission Lines." IEEE Trans. Power App. Syst., Vol. 95, pp. 494-504 (1976).
Chatterjee, I., D. Wu, and O.P. Gandhi. "Human Body Impedance and Threshold Currents for Perception and Pain for Contact Hazard Analysis in the VLF-MF Band." IEEE Trans. Biomed. Eng., Vol.33, No. 5, pp. 486-494 (1986).
Dalziel, C.F. and T.H. Mansfield. "Effects of Frequency on Perception Currents."
Trans. AIEE, Vol. 69, Pt. II, pp. 1162-1168 (1950).
Dalziel, C.F. and W.R. Lee. "Lethal Electric Currents." IEEE Spectrum, Vol. 6, pp. 44-50 (1969).
Deno, D.W. "Calculating Electrostatic Effects of Overhead Transmission Lines."
IEEE Trans. Power App. Syst., Vol. 93, pp. 1458-1471 (1974).
Gandhi, O.P., I. Chatterjee. "Radio-Frequency Hazards in the VLF to MF Band." Proc. IEEE, Vol. 70, No. 12, pp. 1462-1464. (1982).
Gandhi, O.P. and I. Chatterjee, D. Wu, J.A. D'Andrea, and K. Sakamoto. "Very Low Frequency (VLF) Hazard Study." USAF School or Aerospace Medicine, Brooks AFB, Texas: Final Report on Contract F33615-83-R-0613, submitted by University of Utah, Salt Lake City, UT, (31` January 1985a).
Gandhi, O.P., I. Chatterjee, D. Wu, and Y.G.Gu. "Likelihood of High Rates of Energy Deposition in the Human Legs at the ANSI Recommended 3-30 MHz RF Safety Levels." Proc. IEEE, Vol. 73, No. 6, pp. 1145-1147. (1985b).
Guy, A.W. and C.K. Chou. "Very Low Frequency Hazard Study. USAF School of Aerospace Medicine, Brooks AFB, Texas; Final Report on Contract F33615-83-C-0625, Submitted by University of Washington, Seattle, WA. (May 1985).
Guy, A.W. "Hazards of VLF Electromagnetic Fields." Advisory Group for Aerospace Research and Development (AGARD) Lecture Series No. 138, The Impact of Proposed Radio Frequency Radiation Standards on Military Operations, pp. 9-1 to 9-20 (1985).
Rogers, S.J. "Radiofrequency Burn Hazards in the MF/HF Band." In J.C. Mitchell (ed.)., USAFSAM Aeromedical Review 3-81, Proceedings of a Workshop on the Protection of Personnel against RFEM, pp. 76-89. (1981).
Much of this research was developed as a result of a contract between the researchers and the USAF School of Aerospace Medicine, Brooks AFB, Texas. The above paper includes the results of the researchers final report.
Dr. Stock is retired Director, RF Health and Safety Programs at Lawrence Behr Associates, Inc.
