The views expressed in articles published on FIRES do not necessarily reflect those of IES or represent endorsement by the IES.
By Michael L. Grather, CTO, LightLab International Allentown, LLC
We have all seen the wide range of new products claiming to be effective against pathogens of all sorts. Common claims range from the rather ambiguous “fights germs” to the more specific but questionable claim “can disinfect up to 40 square meters of the surface at a time.” Maybe it’s just the times we’re living in, but my first reaction is to take a very skeptical view of these claims. It also seems that I’m not the only one. My neighbor, who is not in the lighting field, asked me if I thought “these UV lights” are safe. Almost without taking another breath he followed it up by asking “… do they actually work?”
If we are truly to get to the bottom of these issues, we must first decide what it means for these devices to be “safe” and if they “actually work.” The answers to these questions are still being studied and discussed in the photobiology community, but it is certain that there will be a need for good measurements, both in the laboratory and in the field, to support research, select products, make predictions, and ultimately verify performance. During my career in photometric and radiometric testing laboratories, I have been a long-time advocate for the importance of understanding the intensity distribution of sources. This information elevates a conversation from discussing source efficacy to discussing the application itself. For example, instead of simply considering the total radiant power or whether the source meets an efficacy requirement, we can now discuss whether the source is actually accomplishing its purpose in the application. The field of germicidal ultraviolet (GUV) application takes this conversation to a new level because of the safety implications of either an ineffective application or an application that may be dangerous to its inhabitants.
Fortunately, there is an entire field of lighting application with well-defined practices that can be applied to the field of GUV application. Many of these practices can be translated easily to GUV applications, but some present new challenges or push common assumptions to a point where errors can become too large to ignore. For example, wall reflectance that allows visible light to bounce around a room may absorb much of the UV-C energy and leave “dark corners” where GUV radiation is not effective. I believe that with good dosing recommendations from the photobiology community, good measurements from the radiometric measurement community, and good practices for application and field measurement, GUV systems can be applied both safely and effectively.
The first kind of safety that most people think about is intended to keep people from being cut, burned, or electrocuted in the normal use and maintenance of the device. In the case of GUV products, however, there is also a need for safety related to the adverse effects that can be caused by the exposure of skin (erythema) and eyes (photokeratitis) to UV-C radiation. Of course, the easiest way to make sure that the GUV radiation doesn’t harm people and pets in the room is to make sure they aren’t in the room during its use. For this reason, many devices employ occupancy sensors and/or timers. Other devices are designed to only treat the air in unoccupied portions of the room, including the upper volume of the room, within enclosed chambers, or in the HVAC duct work. Upper-room devices are specifically designed with very restrictive baffling to ensure that the emitted radiant energy is confined to a very narrow angular range. It should also be noted that UV-C radiation can also damage plants that may be positioned in its path.
Devices that are intended to decontaminate surfaces often come in the form of wands, chambers, and designs similar to those of task lights (for example, under-cabinet and computer-keyboard disinfection applications). These devices are certainly more difficult to control in terms of the safety of the user, although some are also outfitted with interlocks, timers, and motion sensors. In the case of UV-C wands, many are provided with mechanisms that only allow their operation when they are not pointed up. Surface decontamination can also present a problem, depending on the UV-reflectivity of the surface that is being decontaminated. For example, with a powerful source, materials with even a relatively low reflectivity in the UV-C region could cause a reflection of dangerous levels of UV-C energy back to the user.
Another safety concern for the use of GUV products is based on the assumed effectiveness of the device. In this case, if the claims for the product (or the assumption of the user) are not consistent with its actual effectiveness at disinfection, there can be a large difference between the perceived safety of the environment and reality. For example, suppose a manufacturer makes a wild claim about its product, claiming that it will thoroughly disinfect a surface after a one-minute treatment. After reading this claim, the user will likely assume that the surface has been sterilized after a treatment. In this situation, the user may engage in much riskier activities than if it was presumed to be contaminated. This situation has been called “security theater” and can be of particular concern if the people in the environment have a false sense of security and let their guard down. See Ian Ashdown’s FIRES article “Designing a UV-C Germicidal System” for a more detailed discussion on this topic1.
To properly assess the effectiveness and safety of a GUV system, it is necessary to know how much UV energy is being delivered to the targeted pathogens, either in the air or on the surface. It is also important to know how much UV energy may be delivered to people or pets that inhabit the space. This evaluation requires careful measurements of the source as well as consideration of the application environment. Fortunately, this is not an entirely new concept. The lighting industry has been using the measurements of lamps and luminaires to predict levels of visible-light illumination for over a century. Since visible light and UV-C radiation are both forms of electromagnetic energy, there are many analogous properties and calculation methods that can be borrowed for the evaluation of GUV products.
Balancing the effectiveness and safety of a GUV system involves ensuring that the radiant energy gets to the surface or volume of air that is intended to be treated. It also requires ensuring that the radiant power does not get to parts of the room where it might be dangerous. In order to make these evaluations, it is important to know more than just the total radiant power generated by the source. The radiant intensity distribution is equally important in these circumstances, since it can be used to predict the amount of radiant power within a given application environment. The radiant intensity distribution is measured in a laboratory using a specially designed instrument called a gonioradiometer. The gonioradiometer is used to measure the radiant intensity as a function of angle from all angles around the source. Figure 1 shows an example of a polar plot of radiant intensity measured on a gonioradiometer. In the case of an upper-room GUV product, these measurements are used to ensure that the baffles are working as designed by directing the majority of the radiant energy across the volume of air in the upper room while minimizing the amount of radiant energy that hits the ceiling or reaches the occupied portion of the room directly 2. For surface-disinfection applications, the goal is to irradiate the target surface uniformly while minimizing the radiant energy that might reach occupants or cause UV degradation of other surfaces within the environment3.
There are also multiple considerations related to the spectral content of the energy produced by a GUV system. To evaluate these considerations, it is necessary to understand the amount of radiometric power that the UV source produces as a function of wavelength across the UV portion of the electromagnetic spectrum. This is known as the spectral power distribution (SPD) of the source. Generally, DNA and RNA are most susceptible to energy at wavelengths around 265 nm. Because of this, a system designed to produce energy in this part of the spectrum can be very effective at inactivation of microbes and viruses. Simultaneously, keeping the power low in parts of the spectrum that may harm humans is necessary when considering the safety of people in the room.
The susceptibility of viruses and microbes to UV energy can be shown as a function of wavelength. This function of susceptibility is also referred to as an “action spectrum” and can be used to predict a particular UV source’s ability to inactivate pathogens. The action spectra for different types of pathogens can vary depending on the inactivation mechanism employed. Figure 2 shows a graph of the action spectrum for coliphage MS2, with a peak spectral sensitivity between 260 and 270 nm4. It should also be noted that there have been studies showing that UV energy near the far end of the UV-C portion of the spectrum (wavelengths shorter than around 240 nm) may be even more effective than the range of 265-nm to 270-nm range for inactivating many types of viruses 5. Currently, there is no conclusive evidence of exactly how far-UV energy inactivates viruses, but it is likely that studies will increase as excimer UV-C sources (which produce a strong peak in the far-UV range at 222 nm) become more prevalent.
The SPD of a source can be roughly evaluated for its effectiveness by comparing the location of the peaks in its SPD with the peak of the action spectrum of the pathogen that it is intended to inactivate. This is one of the reasons low-pressure mercury lamps (SPD with a strong peak at 253.7 nm) have been used widely in GUV applications.
Another issue related to the SPD of the source pertains to radiant efficiency. The radiant efficiency of a source is the amount of radiant power produced per unit of electrical power consumed. The radiant efficiency can be used to evaluate the amount of electrical power needed to provide an effective amount of UV-C power. Note that low-pressure mercury lamps typically have a radiant efficiency of around 40%, while current UV-LED and excimer sources can be as low as a fraction of a percent.
The other consideration related to the spectral content of GUV sources is the concern about producing too much power in a portion of the spectrum that could cause a safety hazard. Although radiant energy in the UV-B region of the spectrum (280 to 315 nm) can be effective for GUV purposes, it also presents a higher risk of damage to skin and eyes, since it is capable of penetrating deeper into the skin than UV-C 6. Also, too much power in the far end of the UV-C region (less than 200 nm) can generate potentially dangerous levels of ozone, which can be highly toxic in excess concentrations7.
Once the SPD of the source has been measured to ensure that there is radiant energy present where it should be (and not where it shouldn’t be) and the radiant intensity has been measured in all directions around the source, this source data is ready to be applied to the application. Software packages are available that will calculate surface irradiance, eye level readings, and fluence rates for a modeled room. These software packages require source data in the form of an IES file (ANSI/IES LM-63-19 format) modified to contain UV-C radiant intensity (mW/sr) at each measured angle instead of the standard luminous intensity (cd). The accuracy of these predictive calculations is also limited by the accuracy of the model information used in the application. This is an extremely important concern in application calculations, since the reflectance of materials in the UV-C portion of the electromagnetic spectrum can be very different from their reflectance in the visible range. For example, a study on the UV-C reflectance of ceiling tiles has shown that their reflectance can range from 4% to 46%8. As you might imagine, this order-of-magnitude difference in reflectance (especially for ceiling tiles in an upper-room GUV application) can lead to significant errors in the calculations of dosing and occupant safety. As another example, since the reflectivity of typical room surface materials is often lower in the UV-C region than in the visible part of the spectrum, there will typically be less UV-C energy bouncing around the room. This means that the UV-C energy may not be as good at reaching the extreme portions of a room as might be expected if the designer expects it to act similarly to visible light. It is also important to note that the transmittance of materials is also highly affected by wavelength. For example, most glass and transparent plastic is opaque to UV energy.
The effectiveness of a GUV system is normally evaluated by its ability to kill or inactivate pathogens. This is often based on a prediction of the percentage of pathogens that can be expected to be killed or inactivated by a specified dose of energy in the GUV-effective portion of the spectrum. Claims from studies will often state the radiant power delivered to the target pathogen for a period of time (i.e., the radiant energy delivered to the target). This amount of radiant energy delivered (measured in units of millijoules per square centimeter) is called the radiant exposure or dose of UV-C, which is often associated with a predicted percentage of pathogens killed or inactivated. This is often expressed with the term “log kill.” For example: a 2-log kill means 99% inactivation, and a 3-log kill means 99.9% inactivation. Since the recommendations for dosing often come from controlled laboratory studies, the effectiveness of the application will depend on how much the laboratory studies represent real life.
For surface disinfection, the evaluation of the application’s effectiveness will be based on the radiant intensity distribution of the UV sources, their location within the room, the room’s geometry, and the UV-C reflectance of the surfaces. With this information, the software can predict the UV-C irradiance (UV-C power per unit area) for the surfaces in the room. The irradiance combined with the time of exposure will predict the dose (in millijoules per square centimeter) applied to the surfaces in the room. Once again, note that the prediction of UV-C irradiance is only as good as the quality of the information used for the calculations. Of particular interest are the UV-C reflectance characteristics of the surfaces in the room. It should also be noted that irradiance is typically calculated based on a cosine-receptor model of irradiance. This can introduce errors when the actual application may have rough surfaces and therefore be less sensitive than expected to energy from high angles of incidence.
There have been multiple methods proposed for the evaluation of the effectiveness of upper-room GUV systems, and it is likely that improvements will continue to be made as more research is presented on the topic. Dosing recommendations are based on predicted percentages of airborne microbes inactivated when exposed to a given average fluence rate for a given amount of time9. The area in-between the measurement of the UV source and the pathogen’s dosing can have many assumptions and sources of error. Some of the challenges involve the way that the UV energy reaches an aerosolized pathogen and the way that this is modeled in the software (for example, whether it is treated as a direct or indirect spherical receptor). The evaluation of software modeling methods is beyond the scope of this article, but it can have an impact on the results of the calculations. Airflow in the upper-room environment affects the amount of UV exposure the pathogens may receive and therefore is an important factor in the efficacy of the system. Effectiveness also depends on the radiant-intensity distribution of the UV sources, their location within the room, the room’s geometry, and the UV-C reflectance of the surfaces. In addition, the viability of pathogens is often highly influenced by temperature and humidity, which can therefore influence the efficacy of the GUV system. All these complications to the estimation of GUV system effectiveness can be combined into recommendations and best practices for applications. Ultimately, with the careful evaluation of an application’s contributing factors, a plan for the modeling and validation of the application can be developed with appropriate allowance for unknown quantities. In a similar manner, the safety of a GUV system can be evaluated by predicting the amount of energy that is expected to reach the skin or eyes of occupants. In many cases, these predictions can be carried out by the same software used to calculate the system effectiveness. The calculations are based on the above application information as well as the predicted levels of exposure for assumed standard occupants. The maximum predicted levels of exposure for the occupants can then be verified against the recommended levels for exposure to ultraviolet irradiance established by the IES10.
In-Application (Field) Measurements
After a GUV system has been installed, field verification of the performance of the system is highly recommended. The field verification is performed before the system is put into service to ensure that the system is performing as expected and not causing safety hazards for the occupants of the space. It should be noted that there are often modifications made to the design of a space and the materials that were originally planned. As noted above in the software modeling section, these changes can have a large impact on the performance of a GUV system within its application environment. Furthermore, depending on the source type employed in the GUV system, sources will depreciate at different rates over their lifetimes. For these reasons, validation of the system performance should also be conducted according to a planned schedule to ensure the safety and effectiveness of the system. As a minimum, validation should occur when any significant changes to the environment are made (for example, replacement of ceiling tiles, re-painting of walls, or re-positioning of large furniture). If a field measurement of the system indicates that it is not performing to an adequate level, this could mean actions as simple as re-lamping fixtures or increasing dosing time are needed, but may also require adding additional UV sources, or even replacing the system. The process of replacing a system can be a costly endeavor, which is one of the main reasons for employing an accurate software model of the application, including an accurate study of the UV source and the UV-C reflectances of surfaces in the room.
There are many different types of meters designed to measure irradiance in the UV portion of the spectrum. These different designs offer tradeoffs in multiple areas, including cost, portability, ease of use, and of course, accuracy. Meters for the field measurement of UV irradiance generally fall into three categories: broadband irradiance meters, spectral irradiance meters, and dosimeters; see Figure 3.
Broadband irradiance meters are relatively inexpensive, relatively easy to use, and a good choice for quick measurements of irradiance within a space. The meter typically uses a silicon-carbide detector with wavelength filtering designed to only allow the measurement of energy within a certain wavelength range (the passband), and exclude energy outside of the passband (the stopband). For example, the perfect broadband radiometer would have a flat response to wavelengths within the passband and no response in the stopband. Additionally, some broadband irradiance meters are designed to simulate a specific action spectrum. For example, some broadband irradiance meters employ filtering that is intended to simulate the response of many common pathogens to wavelength. Other irradiance meters are filtered for and calibrated based on a specific source. For example, some meters are designed and calibrated for the 254-nm energy that is prevalent in low-pressure mercury lamps. If this is the case, the meter may accurately measure low-pressure mercury lamps, but could yield erroneous results when measuring a UV-LED or excimer product. In all cases, it is important to review the manufacturer’s specifications and make sure that the response of the meter is appropriate for the source to be measured and for the intended measurements.
It should also be noted that if the intent is to measure the total irradiance within a specified wavelength range, the accuracy of the measurements will depend on the source being measured, the response of the meter, and the spectral content of the source used for its calibration. Generally, the more the spectrum of the source being measured deviates from the calibration source, the more uncertainty will be introduced into the measurement. Irradiance measurements are usually performed in a space where the radiant energy can come from a variety of different directions within the room, including multiple sources and reflections. To properly measure irradiance from these many different angles, irradiance meters are usually equipped with a diffuser designed to accept light with a weighting function similar to a cosine function. These diffusers are called cosine-correcting diffusers, and they can highly influence the uncertainty in measurement of total irradiance. The materials that comprise a good cosine-correcting diffuser in the UV region of the spectrum can be relatively expensive, so a buyer should be cautious of very inexpensive meters. In all cases, it is best to confirm the cosine correction capability of an irradiance meter before making a purchase. The metric for evaluating the cosine correction of a detector is called the f2 response. In general, the lower the f2 number, the more its angular responsivity will match a cosine function. As an example, requirements for photometric (visible light) detectors often specify an f2 less than 2%. Because of the expected uncertainty in measurement of UV-C applications, it is recommended that UV-C detectors have cosine correction with an f2 less than 5%.
Finally, many irradiance meters require a “dark measurement” to ensure that the noise inherent in the meter is properly zeroed out. Check with the manufacturer’s literature for the recommended procedure for using the equipment.
Spectral irradiance meters allow the measurement of irradiance as a function of wavelength across the measurement range. These measurements can be very accurate but will often require more time for measurement. The meters can also be larger and more delicate, making them less practical for field measurements. They are also generally more expensive and can be more complicated to operate because of the much larger amount of information that is collected during a measurement. As with broadband irradiance meters, the quality of the calibration is an important component of the accuracy of the measurements made with spectral irradiance meters. Users should check with the manufacturer of the equipment and the calibration lab for the expected uncertainty of the equipment within in the wavelength range that is intended to be measured. The evaluation of spectral irradiance meters can vary greatly based on the design of the meter and the correction techniques employed. See ANSI/IES LM-58-20 for a more thorough discussion of this topic11. It is also to note that the cosine correction of the detector head for spectral irradiance meters is also very important in the measurement of spectral irradiance. As with broadband irradiance meters, it is best to confirm the cosine correction capability of a spectral irradiance meter with the manufacturer before making a purchase.
Paper UV-C dosimeters can be found in the form of cards and stickers of varying shapes and sizes. They are designed for one-time use and are the least expensive and easiest to use of the UV measurement devices. As the name implies, dosimeters are designed to indicate the dose of UV energy that the device is exposed to. A typical application would be to place the dosimeters within a room, and then verify that the treatment was effective after exposure for a given period of time. They are designed to indicate the amount of UV-C exposure incident on the target within a specific range of energy (dose) levels. It is also important to note that most paper UV-C dosimeters are calibrated specifically for 254-nm (low-pressure mercury) responsivity, so UV sources that emit power in different parts of the UV-C spectrum will not be accurately measured. The cosine response of paper UV-C dosimeters is likely to be reasonably good.
There are other technologies that may be useful in the field evaluation of UV-C applications. One of these technologies is called solar-blind imaging, which employs a charge-coupled device (CCD) element that is sensitive in the UV-C region while being filtered from most other radiant energy. Although the accuracy of these imaging devices may not be at an acceptable level for UV-C evaluation directly, they may be useful for detecting minima and maxima within an application. This information can then be used to locate the best areas to evaluate with an irradiance meter.
Field Measurement Technique
There are many potential complications that can cause differences between the predicted performance of a GUV installation and the field measurements when it is finally installed. Some of the differences are the result of physical changes in the installation (for example, the paint in the room may have a different reflectivity in the UV-C region than what was expected). Other differences may be the result of the physical environment (for example, the ambient temperature in the room may have a large effect on the output of the UV-C source). Additionally, some differences can be caused by the condition of the GUV devices. The number of hours that the lamps have been operating and build-up of dirt inside and outside of the UV-C source can be significant sources of differences in expected irradiance measurements.
Fortunately, there are many parallels that can be drawn between field measurements for visible light installations and for GUV installations. Many of the best practices are very applicable and should be followed to obtain meaningful and repeatable information from a field measurement12. Particular attention should be given to recording of environmental conditions, meter type and calibration, proper stabilization (warm-up) of the GUV system, condition of the sources, and location of measurements. Additionally, the best practices for the physical measurement of visible light also apply. For example, ensuring that the person making the measurement does not block the detector or cause additional reflections by standing near the measurement location is critical for accurate measurements. For these reasons, meters that employ remote detector heads or remote triggers may be useful.
It is important to note here that measurements of GUV installations will often involve being in the room while UV-C is present, and often at high levels. Proper care should be taken to avoid prolonged exposure to the skin and eyes while field measurements are being made. Examples of appropriate protection include wearing close-knit clothing with long-sleeves, long-pants, gloves, head and neck coverings, and UV-blocking goggles.
There are many factors to consider when evaluating the effectiveness and safety of a GUV device or installation. A proper evaluation requires careful laboratory measurements of the radiant intensity distribution of the source; knowledge of the spectral power distribution of the source; understanding of the physical properties of the room, including the reflectivity of its surfaces in the UV-C region of the spectrum; target doses of UV-C energy; and information about the planned use of the devices within the environment. Finally, it is important to measure and validate the effectiveness and safety of the installation on a planned schedule to ensure that changes in the performance of the devices or the surrounding environment have not caused any significant changes in the expected performance.
1 Ashdown I. Designing a UV-C germicidal system. Illuminating Engineering Society “FIRES” Forum; 2020.
2 IES Photobiology Committee. ANSI/IES RP-44-21, Recommended Practice: Ultraviolet Germicidal Irradiation (UVGI). Section 6 – Design and Use of UVGI Systems for Disinfection Effectiveness. New York: Illuminating Engineering Society; 2021.
3 IES Photobiology Committee. ANSI/IES RP-44-21, Recommended Practice: Ultraviolet Germicidal Irradiation (UVGI). Section 5.4 – Effects on Paints, Materials, Plants. New York: Illuminating Engineering Society; 2021.
4 Beck S E, Wright HB, Hargy TM, Larason TC, Linden KG. Action spectra for validation of pathogen disinfection in medium-pressure ultraviolet (UV) systems. Water Research. 2015;70:27-37.
5 Blatchley ER, Petri B, Sun W, Rieth LA. SARS-CoV-2 UV Dose-Response Behavior. IUVA; 2020.
6 IES Photobiology Committee. ANSI/IES RP-44-21, Recommended Practice: Ultraviolet Germicidal Irradiation (UVGI). Section 5.3 – Dangers to Humans. New York: Illuminating Engineering Society; 2021.
7 DiLaura D, Houser K, Mistrick R, Steffy S. (editors). The Lighting Handbook, 10th ed. Section 3.6.2. New York: Illuminating Engineering Society; 2011.
8 Wengraitis S, Reed NG. Ultraviolet spectral reflectance of ceiling tiles, and implications for the safe use of upper-room ultraviolet germicidal irradiation. Photochem Photobio. 2012;88:1480-8.
9 Miller SL. Upper room germicidal ultraviolet systems for air disinfection are ready for wide implementation (editorial). Am J Respir Crit Care Med. 2015;192(4):407-9.
10 IES Photobiology Committee. ANSI/IES RP-27-20, Recommended Practice: Photobiological Safety for Lighting Systems. New York: Illuminating Engineering Society; 2020.
11 IES Testing Procedures Committee. ANSI/IES LM-58-20, Approved Method: Spectroradiometric Measurement Methods for Light Sources. New York: Illuminating Engineering Society; 2020.
12 Richman E, PNNL, for the DOE. Standard Measurement and Verification Plan for Lighting Retrofit Projects for Buildings and Building Sites. Washington, DC: U.S. Department of Energy; 2012.