Editor’s Note—The paper on which this article is based was originally presented at the 2019 IEEE Product Safety Engineering Society Symposium. It is reprinted here, with permission, from the proceedings of the 2019 IEEE Product Safety Engineering Society International Symposium on Product Compliance Engineering. Copyright 2019 IEEE.

ESi (Engineering Systems Inc.) was asked to investigate on behalf of several injured parties after a radiation exposure event occurred that caused tissue damage in multiple parties at a hospital. In an effort to determine the effect of the improper usage of the lamps sans filters, measurements were taken by both ESi, for the plaintiffs, and another consultant, performing an occupational health survey and later referenced by the hospital. As these measurements were presented along with the calculations performed to determine the exposure severity, it came to light that the other involved consultant performed a miscalculation that significantly lowered the expected radiation dosage available. This paper outlines these errors and discusses the outcomes of our investigation.

Background on Radiation

The electromagnetic spectrum is continuous from low- frequency radio waves (tens of meters of wavelength) to microwaves (millimeters of wavelength); followed by infrared, visible, and ultraviolet waves (nanometers of wavelength), and finally to ionizing radiation (X-rays and gamma-rays—the highest frequency waves). Different living organisms sense different ranges of the electromagnetic spectrum and exposed tissues are also sensitive in different ways.

The ultraviolet (UV) portion of the spectrum has been defined to be wavelengths from 100 nm to 400 nm and is divided into three ranges: UV-A, UV-B, and UV-C. The UV-C spectrum is referred to as germicidal, due to its ability to destroy biological organic molecules [1]. UV-B is defined to be the range of 280 nm to 315 nm [2]; it is very damaging to skin tissue and is associated with erythema or “sun burn.” The erythemal dose rate for UV-B, the point at which abnormal redness of the skin or mucous membranes occurs due to capillary congestion, is about 250 milliwatts per square meter [3].

Humans experience no immediate sensory feedback that indicates that tissue is being damaged as it is being exposed to short wavelength radiation, since only radiation in the wavelengths from about 400 nm to 790 nm can be seen or felt by people. While radiation such as UV and X-rays are not immediately perceptible, they can cause severe tissue damage that manifests itself after a latent period post exposure. The time delay between exposure and symptoms confounds ready recognition that the exposure was causative.

Mammals can see and respond to the visible light wavelengths and experience the infrared (IR) portion of the electromagnetic spectrum as the sensation of heat. Some snakes have sensory organs that are sensitive to infrared signatures of their prey. Some insects can see in the lower energy portions of the ultraviolet (longer wavelength, UV-A) spectrum, however, humans do not have these levels of perception, and so have no immediate indication that they have been exposed to any UV band.

In this hospital operatory incident, the infrared (heat) portion of the halogen lamp produced light spectrum was the only indication of discomfort, however, even though the IR irradiance was 100 times more intense (as measured in watts per square centimeter) than the calculated UV radiation emitted by the lamps, the American Conference of Governmental Industrial Hygienists (ACGIH) Total Limit Values (TLV), indicated that it was the totalized UV-A and UV-B, not the IR wavelengths, that caused the tissue damage in the hospital operatory.

Theory of Operation

In order to provide background for the light measurements, some fundamentals are provided here. The degree of injury is proportional to the energy and exposure duration, and the energy is inversely related to the wavelength. The wavelength is normally represented by the Greek letter lambda (λ) and has the units of meters. Frequency is normally represented by the Greek letter nu (ν) and has the units of hertz or more intuitively as cycles per second.

The energy of electromagnetic radiation is proportional to the frequency:

E(energy) = h / λ (1)

Where energy is represented by the letter E and has the units of joules, h is Plank’s constant (value), and lambda (λ) is the wavelength. Since the damage is cumulative, that is, it occurs over time, it is natural to express the power of the incipient light as joules per second, or watts. Since the incident light is spread over an exposed area, the intensity of the incipient radiation is measured in watts per square area, in the case of the spectroradiometer used in this investigation, the units selected were watts per square centimeter. This is referred to as irradiance. Sunlight at noon near the equator has a power of about 1000 watts per square meter, or equivalently 0.1 watts per square centimeter.

Total Energy Calculation

The International Light Technologies model ILT950 spectroradiometer measures the irradiance in microwatts per square centimeter at each wavelength (nm) in one nm increments. ACGIH [4, Table 1] specifies the damage sensitivity of skin tissue to each wavelength, referred to as “relative spectral effectiveness.” The measured irradiance multiplied by the spectral effectiveness gives the effective irradiance. The irradiance over a range of wavelengths is the sum of the effective irradiances measured at each wavelength. For example, the unfiltered irradiance at 251 nm was measured as 3.45 µW/cm2, the spectral effectiveness was 0.43, resulting in an effective irradiance of 1.48 µW/cm2. The spectral effectiveness at 270 nm is 1.0. The total unfiltered energy contained in the range of wavelengths from 250 nm to 307 nm add up to 363.88 µW/cm2, where each wavelength has been scaled by its spectral effectiveness. In contrast, the total filtered energy for the same spectral range is 1.15 µW/cm2. Note that ACGIH does not specify sensitivity factors in 1 nm increments for the entire range, so ESi interpolated the spectral effectiveness for unspecified wavelengths.

Note that the International Light Technologies model ILT1400A does the above calculation internally over the detector range of 190 nm to 307 nm. The filter that was used limited the spectral range to between 235 nm and 307 nm. The ILT950D detector minimum wavelength was 250 nm.

ACGIH [4, Table 2] gives the allowed exposure time for a specific number of milliwatts of irradiance, tissue injury is then a function of energy. Using the measured values above, the measured irradiance of 363.88 µW/cm2 falls between the 300 µJ/cm2 and 3000 µJ/cm2 in the table, so exposure time falls somewhere between 10 seconds and 1 second (Table 1). The actual exposure time can be calculated by dividing the total allowed exposure energy of 3000 µJ/cm2 by 363.88 µW/cm2, giving an exposure time of approximately 8.2 seconds for unfiltered UV.

ACGIH Threshold Limit Value (TLV)
Exposure Per 8 Hour Effective Irradiance Dosage
8 hours 0.1 µW/cm2 2880 µJ/cm2
4 hours 0.2 µW/cm2 2880 µJ/cm2
2 hours 0.4 µW/cm2 2880 µJ/cm2
1 hour 0.8 µW/cm2 2880 µJ/cm2
30 minutes 1.7 µW/cm2 3060 µJ/cm2
15 minutes 3.3 µW/cm2 2970 µJ/cm2
10 minutes 5.0 µW/cm2 3000 µJ/cm2
5 minutes 10 µW/cm2 3000 µJ/cm2
1 minute 50 µW/cm2 3000 µJ/cm2
30 seconds 100 µW/cm2 3000 µJ/cm2
10 seconds 300 µW/cm2 3000 µJ/cm2
1 second 3000 µW/cm2 3000 µJ/cm2
0.5 second 6000 µW/cm2 3000 µJ/cm2
0.1 second 30 000 µW/cm2 3000 µJ/cm2

Table 1: ACGIH [4, Table 2] with added dosage calculations. The dosage column was added to the table based on ACGIH [4, eq. (7)].

The calculation method then becomes (1) measure the irradiance at a particular wavelength, (2) multiply each irradiance by the ACGIH spectral effectiveness, (3) add all the adjusted irradiances together, and (4) divide the number of joules of allowed energy [4, Table 1] by the effective irradiance; exposure time will be in seconds.

Performing this calculation for the filtered luminaire light from 250 nm to 307 nm gives a total effective irradiance of 1.155 µW/cm2. ACGIH gives a slightly higher exposure energy of 3060 µJ/cm2: [(3060 µJ/cm2)/(1.155 µW/cm2)] yields an exposure time of approximately 44 minutes [4, Table 1].

ACGIH Exposure Threshold Limit Values

Both UV and IR can cause injury to bodily tissues. The cornea of the eye is particularly susceptible to damage from UV and IR and is taken as the basis for the worst-case values set forth in the ACGIH TLV. It is worth noting that the skin offers a degree of protection from various wavelengths of light, as such, the ACGIH TLVs are for intact tissue, not unprotected, exposed subcutaneous tissue [5].

The most damaging UV wavelength is accepted as 270 nm. ACGIH gives calculated exposure times for a 3000 µJ/cm2 dosage [4, Table 2]. The following explanation is given by the Lawrence Berkeley National Laboratory Environmental Safety & Health (ES&H) [6]:

The TLVs for UV radiation apply to electromagnetic radiation with a wavelength between 180 nm and 400 nm. Such radiation may present an eye and skin hazard. These TLVs apply to UV radiation from plasma discharges (e.g., welding, plasma etchers, etc.), solar simulators, unfiltered fluorescent and incandescent lights, and other incoherent UV sources. These TLVs do not apply to UV lasers.

The TLVs for UV radiation are wavelength and time dependent. The TLVs listed in the table below are the most restrictive. Contact the non-ionizing radiation subject matter expert for an evaluation of specific UV hazards and determination of specific application TLVs if these limits cannot be met.

Discovery of Injuries

Patient injuries were discovered upon regular post‑surgery follow-up. In several patients, the surgery sites did not heal as expected and long-term use of sterile compression bandages was required. The cause for the poor healing was not understood until the observation was made that the surgical sites resembled burn trauma.

Because the tissue damage observed by the attending physicians resembled burns, the plaintiff’s legal counsel had the mental image of a cooking broiler causing burns (thermal burns) to the exposed tissue. Such an imagining is natural given the earlier discussion of the ability of different organisms to sense different wavelengths of light—infrared can be felt as heat by humans, and this is what the physicians were reporting— however, the UV portion of the spectrum cannot be immediately seen or felt. The damage to the exposed patient tissue did manifest with the appearance and symptoms of a burn, but all damaging radiation manifests in this way [7].

Realizing that plaintiff’s counsel needed testing and analysis performed to determine what did and did not cause the tissue damage, and also realizing that defense counsel was going to argue that the “tissue burns” could not have been caused by the operatory lights, ESi personnel measured the intensity of each wavelength of light from 240 nm to 1050 nm, covering the UV- A and visible spectrum and a portion of the UV-B and IR spectrums. The results were imported into a spreadsheet to apply the ACGIH damage factors were. The results were used to account for differences in sensitivity of human tissue.

Halogen Light Bulbs

Everybody today is at least partially aware of what an incandescent bulb is. The principle of incandescent illumination relies on the fact that as an object is heated up it emits light according to the black body radiation curve [8]. The hotter it is, the more light is emitted in the visible portion of the spectrum. If the object is hot enough, some portion of the emitted spectrum will be in the UV region.

Halogen incandescent bulbs emit more total light in a smaller bulb size because the gases, bromine or iodine, enclosed inside the bulb allow the tungsten filament to survive higher temperatures. The higher filament temperatures cause it to emit light with a higher color temperature including potentially hazardous wavelengths including UV-A, UV-B, UV-C, and IR.

The light source that was the subject of this investigation was designed with IR and UV filters that would pass only exactly those wavelengths visible to the human eye. The luminaire instruction manual correctly specified the methods for cleaning the lenses and filters, as well as those methods which were unacceptable. In addition, the filters were replaceable components of the luminaire. For unknown reasons the hospital maintenance staff erroneously opted to remove the filters entirely and not replace them, thereby allowing the physicians and patients to be exposed to the full halogen lamp electromagnetic radiation spectrum.

The operatory luminaires are shown in Figures 1–3.

Figure 1: Operatory luminaire as presented.


Figure 2: Operatory luminaire showing one of the diffusers open for servicing.


Figure 3: Operatory luminaire diffuser and filter elements.

OSHA Exposure Guidelines

Occupational Safety and Health Administration (OSHA)
1091.97 is the only regulation regarding non-ionizing radiation and does not establish a radiation protection guide for the wavelengths that apply to this instance.


The degree to which a particular wavelength of light is damaging to living tissue is a function of the energy of the particular wavelengths of light and the nature of the tissue being exposed. Human skin has evolved to have a certain resistance to UV wavelengths, however, exposed subcutaneous tissue has little to no resistance. Further, the damage is cumulative and compounding, that is, the longer the tissue is exposed, the more tissue is damaged. Anybody who has experienced a sunburn has an intuitive understanding of the potential intensity of radiation and the amount of time that is required when being exposed to summer sunlight.

Instrumentation and Test Results

ESi used an ILT950D spectroradiometer to evaluate the radiation output of the subject lamp. The test setup is shown in Figure 4. This instrument can measure the intensity of light with wavelengths of about 250 nm to about 1050 nm, covering the ultraviolet UV-A and UV-B, visible light, and IR spectrums. A 10x attenuating flat spectrum filter was used to prevent overloading the detector; therefore, when interpreting the graphs in Figure 5 and Figure 6, the measured numbers on the vertical irradiance axis must be multiplied by 10 to correct for filter attenuation.

Figure 4: Operatory luminaire showing the ILT950D measurement set up by ESi.


Figure 5: Measured light output without the supplied luminaire filters. Note the values on the left must be multiplied by 10. Vertical axis units are microwatts per square centimeter. The horizontal axis is nanometers.


Figure 6: Measured light output with the luminaire optical filters. Note the values on the left must be multiplied by 10. Vertical axis units are microwatts per square centimeter. The horizontal axis is nanometers.

The following instruments were using in testing:

  • ILT950D w/integral detector, serial number 1403140U1
  • W2 diffuser, (flat spectrum) serial number 14373, included in the calibration
  • QNDS1, 10x attenuating flat spectrum filter, serial number 28938
  • QNDS2, 100x attenuating flat filter, serial number 29467

ESi measured the total unfiltered UV light between 280 nm and 315 nm UV-B band, to be approximately 25 µW/cm2, with the luminaire diffuser in-place and with no luminaire spectral filter. The 10x attenuating flat filter and the diffuser are shown in Figure 3.

When testing was performed without the filters in place the UV-B intensity peaked at about 100 µW/cm2
at 377 nm. With both the diffuser and optical filter in place, the UV irradiance between 250 nm and 377 nm was approximately 0.1 µW/cm2 (
Figure 6 on page 32). Using the ACGIH UV exposure effectiveness for each
measured wavelength the following values were obtained:

Spectrum: 250–307 nm (occupational health consultant. spectral range)

ACGIH UV Dosage: 3000 µJ/cm2

Luminaire Filtered UV: 43-minute exposure time

Luminaire Unfiltered UV: 8.2 seconds exposure time

The above UV range of 250 nm to 307 nm was chosen based on the measurement range of the equipment used for the occupational health personnel measurements.

The occupational health personnel used an ILT1400A to perform their evaluation of the operatory luminaire. They correctly identified the most damaging wavelengths and selected the SEL 240 detector sensitive to UV-B and UV-C, 190–320 nm, however, they also used a TC ACT5 filter limiting their measurement to UV-B, 235–307 nm. The use of the QNDS3 1000x filter reduced the light intensity to the lowest limit of what the SEL 240 detector could sense. The ILT1400A displayed measurement readings that totalized the energy per ACGIH [4, eq. (1) and (2)].

The following instruments were using in the testing by the occupational health personnel:

  • ILT1400A, serial number 7182
  • QNDS3, 1000x attenuating flat filter, serial number 28815
  • T2 ACT5, 2x attenuating actinic spectral filter, 235– 307 nm, serial number 25749
  • SEL 240, detector, 190–320 nm, (no serial number reported)

Summary of Analysis

Subcutaneous human tissue was exposed to unfiltered light from halogen light bulbs in a hospital operatory. The UV-A portion of the halogen light spectrum was determined to be the most damaging portion of the spectrum based on the ACGIH standard [4]. The state occupational health personnel, not understanding the principles of operation of their measurement equipment underestimated the UV light intensity by a factor of 2000. Correct measurement over the spectrum from just beyond UV-B (240 nm) to IR (1050 nm) showed that despite the differences in what physicians could feel, and what the lawyers’ intuition was telling them, that the UV was the most damaging radiation emitted from the lamp in the surgical facility with an exposure time of about 8 minutes for intact skin. Damage was likely more severe for the unprotected subcutaneous tissue that was exposed during the operations [9].


  1. The surgical luminaires were intended to operate only with the optical filters in place. However, even with the filters in place, based on ACGIH [4, Table 1] and measurement of the filtered operatory luminaire, calculations indicate that UV exposure limits would occur within 44 minutes of surgery.
  2. The totalized unfiltered measured light intensity between 250 nm and 307 nm (UV-A and UV-B), was approximately 363.88 µW/cm2. More than eight seconds of exposure would exceed the ACGIH Threshold Limit Value of 3000 µJ/cm2 for unbroken skin.
  3. The measured unfiltered IR irradiance of 1150 µW/cm22 at a single wavelength of 1050 nm, would cause tissue damage in approximately 1.4 hours, while the total energy in the spectrum from 780 nm to 1050 nm would have caused tissue damage in 1.6 minutes. No IR exposure limits have been established for exposed subcutaneous tissue.
  4. The UV actinic/germicidal wavelengths are considered cytotoxic and are routinely used for germicidal disinfection [1]. While not quantified in the literature, human skin or subcutaneous tissue, should never be exposed to such UV radiation.
  5. ACGIH Table 1 presents Threshold Limit Values intended for exposure of intact skin tissue. No exposure limits have been established for surgically exposed subcutaneous tissue.
  6. The AIHA Quick Reference Sheet states that acute overexposure may cause erythema (redness or burning) of the skin and photo keratitis (inflammation of the cornea). International Agency for Research on Cancer classifies UV as a Group 1 Human Carcinogen.
  7. Based on the exposure limits and the measured light intensities in the UV and IR portions of the light spectrum, exposure to exposed subcutaneous tissue is likely to cause permanent injury as well as latent cancer with the possibility of future medical complications.
  8. Use of the 2000x attenuation by the occupational health personnel ILT1400A reduced the light intensity of the detector to the limit of its sensitivity, and thereby likely created measurement errors.
  9. The occupational health personnel did not account for the 2000x light reduction of their measurement filters. Their reported values matched ESi measured values when multiplied by this factor; reported values ranged from 0.01 µW/cm2 to 0.03 µW/cm2. Corrected values would fall between 20 µW/cm2 and 60 µW/cm2, resulting in TLV exposure times ranging from 30 seconds to 2.5 minutes.
  10. Irrigation of the surgical site during surgery would have reduced the potential for heat (thermal) damage from the IR part of the spectrum but would not mitigate cellular damage form UV.
  11. The operatory luminaire could have been designed with interlocks that prevented its use with the filters removed.


Thank you to the lawyers representing the injured parties, they deserve good representation. Thank you to the scrupulous review of my colleagues.


  1. AIHA, “Ultraviolet Radiation Quick Reference Sheet,” AIHA Non- Ionizing Radiation Committee, Rev. 1, March 4, 2013.
  2. ISO 21348 Definitions of Solar Irradiance Spectral Categories. Courtesy of Space Environment Technologies, http://spacewx.com.
  3. J. R. Herman, N. Krotkov, E. Celarier, D. Larko, and G. Labow, “Distribution of UV Radiation at the Earth’s surface from TOMS- measured UV-backscattered radiances,” Journal of Geophysical Research, vol. 104, no. D10, pp. 12,059–12,076, May 27, 1999.
  4. ACGIH, 2013 TLVs and BEIs: Based on the Documentation of Threshold Limit Values for Chemical Substances and Physical Agents & Biological Exposure Indices. Cincinatti: ACGIH, 2013.
  5. M. Whelan, “The Halogen Lamp,” Edison Tech Center, 2013. http://edisontechcenter.org/halogen.html, accessed February 3, 2019.
  6. Lawrence Berkeley National Laboratory, “ES&H Manual (Pub 3000),” chap. 43, 2018. https://www2.lbl.gov/ehs/pub3000/CH43.html.
  7. R. P. Rapini, Practical Dermatopathology, 1st ed. Philadelphia: Elsevier Mosby, 2005.
  8. R. Loudon, The Quantum Theory of Light, 3rd ed. New York: Cambridge University Press, 2000.
  9. IQWiG, “How Does Skin Work?” https://www.ncbi.nlm.nih.gov/books/NBK279255, 2016. accessed February 3, 2019.

Leave a Reply

Your email address will not be published.