DIY Disinfection Chamber
Design Calculations and Adenovirus Inactivation
Method
Very short wavelength of UV light, labelled UVC, has long been known to deactivate bacteria and viruses when a sufficient amount light energy is deposited on the micro-organism. From a brochure found at http://www.clordisys.com/ :
What is sufficient dose?
In the same brochure a large amount of data assembled from the literature is presented in tables for both bacteria and viruses. From this table a worst case fluence is taken for a particular adenovirus. In order to achieve a 3 log reduction (99.9%) of this virus on a surface required a fluence(dose) of 155 mJ/cm2 of 254 nm light from a low pressure (LP) mercury discharge tube. Many other viruses in the same table seem to require around 20-30 mJ/cm2 but certain other ones range up to this worst case number.
In a scientific paper[1] a G10T5/4P UVC LP tube was used (16W, 14 inch long, 5/8” diameter). At 20 centimeter distance the radiant flux was 80 uW/cm2. In order to deposit 155 mJ/cm2 of energy at this distance (about 8 inches) it requires about 32 minutes of radiation.
A chinese standard[2] specifies a minimum radiant flux of 70 uW/cm2 when measured one meter from a 30W UV tube of 253.7 nm radiation (measured presumable along the centerline perpendicular to the axis of the tube). Therefore, one can conclude typical LP tubes will be able to provide enough dose to inactivate worst case viruses in typically 30 minutes.
Low cost disinfection chamber.
A minimum requirement for a disinfection chamber is that during operation the chamber should be enclosed to substantially limit escape of UV from shining on any living subjects nearby, and it should maximally make use of the radiant energy in order to flood all surfaces of objects inside the chamber. This last requirement is met by covering all walls of the chamber with aluminum foil since aluminum has a very high reflectivity for UV[3]. The referenced paper demonstrates that the best surface to use is the bright side of the foil (such foil has a matte side and bright side). By covering all sides of the chamber the UV light from the tube bounces around the inside thus illuminating even the sides of a package not facing the tube. This is the principle of the so-called “integrating sphere” in optics.
In figure 1 below are seen the chamber and its components.
- A plastic storage bin.
- Aluminum foil taped to all surfaces (including the cover), bright side up.
- A 6W “germicidal LP tube assembly.
- Upper right hand corner is shown a corncob style UVC LED lamp that is “to-be-evaluated”.
Figure 1.
Operation
In figure 2 the chamber is shown in operation. The cover is propped open a bit to show lamp operation. The blue visible light that is seen results from a principal spectral line that is present at 546 nm. For a UVC LED light source no visible light will be present since the inherent design of the semiconductor will provide for a narrow band emission at the design wavelength, and is thus monochromatic.
Also shown in figure 2 is a web server operated relay to turn the lamp on/off. The lamp can be turned on/off either by a browser connection to the web server or by an on/off switch on the relay device.
Figure 2.
In figure 3 is shown an example of an uncrowded arrangement of packages being irradiated. The positioning is done so that substantially all sides receive radiation flux. Bottoms cannot receive flux but would typically have a much less probability of containing viruses than the sides. To disinfect bottoms the packages would need to be turned over and re-radiated. The major point is to place items in a manner to minimize shadowing. The highly reflecting aluminum surfaces greatly aid in achieving this result.
Figure 3.
Safety
UVC radiation is not only damaging to micro-organisms, but will cause burns to human skin and can cause serious burns to the cornea of the eyes. One needs to avoid, if possible, even brief exposure looking at the lamp. Also, the chamber should be operated in an environment where a person or pet cannot inadvertently open the chamber during operation.
UVC LED lamps
The cost has come down and the availability of UVC LED chips has increased to where lamp assemblies are now being offered commercially. More on this later.
The physics of semiconductors lead to an LED chip having a single peak frequency of light emission. There are no other lines of spectral emission such as in LP discharge tubes. Surveying various publications and data sheets shows the emission band is quite sharp, bell shaped with an approximate width at 10% of about 20 nm. Various peaks are specified, 255 nm, 265 nm, 275 nm, 280 nm, etc. Why the variety? This is best explained in a brochure that involves the company that produces these chips[4]:
“With the popularity of low-pressure mercury lamps, 254 nm has been thought of as the ideal wavelength, even though the peak germicidal effectiveness falls between 260 and 270 nm, depending on the specific pathogen. UV-C LEDs are quasi-monochromatic
(majority of output falls within a 10-nm gap) but can be engineered in a variety of
wavelengths within the germicidal range to target specific pathogens or the general
peak pathogen UV sensitivity. (e.g., 255, 265, 275 nm, etc.). The choice of wavelength
is usually dependent upon multiple factors………..”
A search for commercially available UVC LED lamps for germicidal application has resulted only in Chinese sourced lamps configured in corncob style with standard E26/E27 screw-in sockets. Although multiple Chinese distributors (“stores”) offer this product, it appears to be manufactured by a single company. The 60W lamp comes with a remote controller and timer. See figure 4.
There are considerable problems with the specifications presented with this lamp by the various vendors. Here is a list of some:
- The wavelength is listed as 253.7 nm. This is suspicious because this is the precise wavelength of a principal spectral line of a LP mercury discharge tube. Certainly one would not specify a UVC LED chip this way (typically would specify 255 nm, for example).
- Some stores specify the luminous flux to be 7200 lumens. This is ridiculous because luminous flux is a photometric parameter describing the spectral radiance over the visible spectrum adjusted by the response of the human eye.
- One store specifies the color temperature to be 7200 Kelvin. Another ridiculous number.
- All presentations show the lamp emitting bluish light as seen in figure 4. Clearly the emission from UVC LED chips is singularly monochromatic and invisible to the human eye.
This lamp may be legitimate but grossly miss-specified as a result of ignorance. There is a Chinese manufacturer of LED lighting products named Shenzhen Suntech Company Limited and at their web site one can find Germicidal UVC LED Strip. Figure 5 is a screenshot from the web site and the specification data presented appears valid.
Note added April 11, 2020: Reviewers on Amazon.com report this product as a scam. One reviewer reported his photochromic glasses were unresponsive to this light. Another reported he used an expensive UVC meter and found no UVC. My observations when I received this unit are that the unit simply emits bright blue light. The only way this could occur if the source is UVC is the LED lens is coated with a phosphor, which would be nonsense if the lamp is intended to emit UVC.
Therefore, if the lamp assembly in question is constructed with this strip then the product is valid, but miss-specified. When a sample of this lamp product is received it should emit no visible light. This seems to be only way to evaluate the lamp without a UV radiometer.
Figure 4.
Figure 5.
A last point but very important point regarding UVC LED lamps is the Wall Plug Efficiency. This value is the power in the emitting light radiation divided by the electrical power drawn out of the wall plug in terms of a percentage. For mercury discharge lamps this varies between 15 and 35 % while for UVC LED chips it is at best 4 %[5]. Therefore, it cannot be assumed that the 60W corncob UVC LED lamp will deliver higher dose (fluence) to a surface than a lesser rated LP tube.
Misleading CDC information
The CDC recently published guidance for Decontamination and Reuse of Filtering Facepiece Respirators. Not surprisingly, disinfection using UV germicidal irradiation is one of the methods recommended. Unfortunately the dose range recommended is quite wrong. See Figure 6.
The dose range is impractically much too high. In fact, to accumulate a dose of 950 J/cm2 in most chambers would take about 1000 hours. The numbers on the far right in the figure are references to the papers cited by the document. However, the dose range apparently came from only one of the references[6]. In that study the authors purposely dosed N95 respirators with excessively high doses in order to study the degradative effects repeated irradiation has on the respirators. Apparently the authors of the CDC recommendation failed to understand this and have now put out a ridiculous dose range. In fact, it can be shown that a dose below 0.5 J/cm2 is more than enough to inactivate even worse-case viruses.
Design Calculations and Adenovirus Inactivation
In order to gain the understanding of what is involved in applying UV lamps to the problem of disinfection the literature was surveyed. There appears two major uses that form the basis for many of the papers: UV disinfection of water systems and UV disinfection in air conditioning systems. Taken together this literature leads one to breakdown the titled topic into two parts: understanding UV sources and understanding how micro-organisms are inactivated.
UV Sources and calculation irradiance
A single good source for understanding UV lamps is found in CIE 155:2003[7]. In particular the properties of low pressure (LP) and medium pressure mercury discharge lamps are described. The majority of germicidal UV lamps are of the low pressure tubular type where the majority of the optical output is in the spectral line at 253.7 nm. These lamps are called UVC LP and convert 30 % of the electrical input energy to optical energy at this wavelength. This percentage is called the Wall Plug Efficiency. The manufacturer of UVC LP lamps specifies both the wall power and the UV output power, for example, a 6W G6T5 lamp outputs 1.8W of UVC. As taught in many of the references this particular wavelength is nearly optimum for breaking apart RNA bonds that have a peak absorption at 265 nm.
For purposes of both specification and design a straightforward model[8] has been developed for the irradiance E of UVC at a distance a from an LP lamp of length L. Shown in figure 7 is the expression for irradiance along the center line that bisects the lamp perpendicular to the lamp axis. The value of greek letter phi is the UVC output in watts and greek letter alpha is the angle shown in figure 8.
This expression can now be used to evaluate the 6W UVC G6T5 lamp used in the chamber. Data on this particular lamp is shown in figure 9.
The length and UVC output is used in the equation and the irradiance along the center line is shown in figure 10. Figure 10 gives the fluence on an object receiving power directly from the lamp source. In the chamber this is a minimum value of fluence since the walls are highly reflective and the object receives additional fluence. So actual practice is better than what is being calculated here.
Adenovirus inactivation
How the 6W UVC lamp performs is evaluated by first considering inactivation of adenoviruses. These viruses are the most UV-resistant group of viruses currently known. Therefore, finding the dosing times for this lamp that inactivate these viruses gives one the performance capability of this simple low wattage lamp. A valuable reference that permits one to find these dosing times is “Effect of adenovirus resistance on UV disinfection requirements: A report on the state of adenovirus science”[9]. The authors have assembled a great deal of data from studies on the inactivation of adenoviruses, and this information is summarized in figure 11.
Normally one doses a surface until there is at least a 3 log reduction (99.9%) in virus particles. For UVC LP lamps one sees this corresponds to about 200 mJ/cm2. Now one can carry the result in figure 10 further, resulting in figure 12.
One can conclude that the irradiation times and distances are quite practical for disinfecting the difficult adenoviruses. Many micro-organisms are much easier to obtain a 3 log reduction and many require only a dose 10 times smaller than adenoviruses. For examples one can consult the tables in http://www.clordisys.com/.
Excellent set of posts on UVC!
A June 2020 paper by Buonanno et al https://www.nature.com/articles/s41598-020-67211-2 suggests 1.2 – 1.7 mJ/cm^2 inactivates 99.9% of aerosolized coronaviruses, quite a bit lower than the numbers you cite from the Clordisys and the CDC. Do you use your disinfection chamber to deactivate COVID-19–if so, what dose do you target?
Keep up the good work!
When you saw this paper did you take note that it regards excimer lamps that emit in the FarUVC, specifically at 222 nm? There seems to be a current body of research where this particular wavelength is quite effective inactivating viruses but does not penetrate into skin and eyes whereas 254 nm does and is harmful. These lamps are not yet commercial and USHIO is actively in the process to commercialize them. I cannot find pricing but it is my understanding that excimer lamps are quite pricey. Until these lamps come into volume production and prices become competitive it is not likely one will see these for a while. The low pressure (LP) mercury discharge lamps have existed now for many decades and are produced by many companies at low cost and high volumes. The only thing unique about UVCG lamps compared to generic fluorescent lamps is the transmission properties of the tube, the lack of a phosphor coating, and the gas composition and pressure is optimized to put the power into the 254 nm line (~10% of power goes into the ozone-producing 185 nm line). In my post I simply looked at what dose it takes to inactivate viruses at 254 nm for which there is a large body of studies. The adenoviruses are known to be the most UV resistant and ~150 mJ/cm2 is required to achieve 3 log reduction. Unless someone discovers that covid-19 is some kind of ultra-UV resistant virus at 254 nm one can be safe to dose objects at this level. Therefore I use what I teach in the post to find exposure times to achieve doses in this range. I shoot for no longer than 15 minutes. I have absolutely no capacity to study viruses in this regard and I just use the chamber to do personal disinfections, mostly on objects with porous surfaces like my N95 mask. I find it much easier to wipe down most surfaces with any squirt bottle cleaner containing long chain quaternary ammonium compounds like Lysol, 409, Fantastik, etc. Thanks for pointing out this paper.