240.437.4615 | Contact Us

Action Spectra: A Review

James R. Bolton

Bolton Photosciences Inc., 628 Cheriton Cres. NW, Edmonton, AB, Canada T6R 2M5
Click Image to View in Gallery Mode


Table 1. Comparison of GF(λ) values for Bacillus subtilis spores using various methods (taken from data sets in Chen (2007).

Supplemental material

Click here for data (PDF) from the literature for which action spectra have been determined.

Introduction and definitions

Following the publication of an extensive review of the UV sensitivities of many microorganisms (Haji Malayeri, et al. 2016), it is appropriate to review the action spectra that have been determined for some of these microorganisms.

Action spectra are important in analyzing data for UV experiments involving polychromatic UV lamps (e.g., medium pressure UV lamps). They are also very useful for modeling medium pressure UV systems using computational fluid dynamics (CFD)-based UV dose models.

The germicidal factor [GF(λ)] or “action” of a microorganism is defined as the UV sensitivity of that microorganism at wavelength λ normalized to 1.00 at 253.7 nm (the primary emission wavelength of a low pressure UV lamp). In general terms, the action spectrum of a microorganism is a plot of GF(λ) vs. wavelength. However, there are several possible methods for determining GF(λ) from experimental data.

Methods for determining germicidal factors

The various methods for determining the GF(λ) values for a microorganism involve analysis of fluence (UV dose) – response curves (or photon fluence – response curves) carried out at several wavelengths. Ideally, the spectral width of the UV source should be very narrow (< 1 nm); however, in many cases, such as the use of interference filters or a monochromator with a medium pressure UV lamp, the spectral width is 10-12 nm. The best case is the use of a tunable laser, where the spectral width is <1 nm.

As pointed out by Bolton et al. (2015), the rate of photochemical and photobiological processes is driven by the rate of photon absorption. Thus ideally, the GF(λ) values should be derived from plots of log(N0/N) vs. the photon fluence (einstein cm-2), where N is the number of viable microorganisms per mL at photon fluence Fo,p(λ) and N0 is number of viable microorganisms per mL at Fo,p(λ) = 0.

However, in most cases the GF(λ) values have been derived on an energy basis from plots of log(N0/N) vs. the fluence (UV dose) (mJ cm-2). Bolton et al. (2015) indicated that GF(λ) values derived from energy-based data can be converted to GF(λ) values derived from photon-based data by multiplying the former by (253.7/?). Henceforth we designate GFε(λ) as germicidal factors derived from energy-based data and GFp(λ) as germicidal factors derived from photon-based data.

The various methods for the determination of GFε(λ) values are as follows:

1. Fluence-based rate constant method
Plots of log(N0/N) vs. the fluence often have a linear region, where one can derive a first order rate constant . GF(λ)then is defined as

2. Fixed fluence method
Plots of log(N0/N) vs. the fluence are often non-linear. The International Union of Pure and Applied chemistry (IUPAC) (Braslavsky 2005) thus proposed that “In cases where the fluence dependence of the photoresponse is not linear (as is often the case in biological photoresponses), a plot of the photoresponse vs. fluence should be made at several wavelengths and a standard response should be chosen (e.g., two-log reduction). A plot of the inverse of the “standard response” level vs. wavelength is then the action spectrum of the photoresponse.” Thus

where n is the fixed log reduction chosen (e.g., 2). Gates (1930) used this method.

3. Fluence-response mapping (curve mapping) method
Coohill (1991) and Sutherland (2002) have pointed out that if the fluence-response curves at various wavelengths have a very similar fluence-response profile, then it is possible to apply a single constant that will map the fluence-response curve at given wavelength to that at 253.7 nm. This constant is 1/GF(λ). The advantage of this method is that all the data in the fluence-response curve are used to determine GF(λ). Beck et al. (2015) used this method to determine their germicidal factors.

If the dataset are plots of log(N0/N) vs. photon fluence, the corresponding GF values are GF1p(λ), GF2p(λ) and GF3p(λ), where

where i = 1, 2 or 3.

The choice of GF values based on fluence-response curves or those based on photon-fluence response curves depends on how the GF values are to be used. For example, if one wishes to calculate the average germicidal fluence rate over a certain range, the following equation should be used (Bolton et al. 2015)

where Eλ is the spectral irradiance (W m-2 nm-1) at wavelength λ.

On the other hand, if one wishes to calculate the average germicidal photon irradiance, the following equation should be used

where Ep,λ is the spectral photon irradiance (einstein m-2 s-1 nm-1) at wavelength λ.

As an illustration of how much variation is caused by use of the various methods, the data sets from Chen (2007) were used to generate GF values using several methods. As can be seen, there is not much difference among the methods for the GFe(λ) values. The GFp(λ) values increase at wavelengths less than 254 and decrease at wavelengths greater than 254 because of Eq. 3.

Various methods have been used to select wavelengths from a broadband UV light sources for the determination of action spectra, such as:

1. Monochromators – the early work by Gates (1930) used a quartz monochromator; however, modern monochromators are usually are grating monochromators. A good monochromator can produce a narrow beam with a width of only a few nm.

2. Interference filters – these filters can be manufactured with specific center wavelengths, so a set of filters can be used to determine an actions spectrum. The problem is that the bandwidth is 10-12 nm (full width at half height), so the action values are somewhat “smoothed” by this bandwidth.

3. Tunable laser – this method is by far the best because the bandwidth for a tunable laser is much less than 1 nm. Most of the data in Beck et al. (2015) used this method.

Action spectra data and discussion

Action spectra data has been extracted from the various references either by extracting data from tables or by interpolating off charts using the Web Plot Digitizer (http://arohatgi.info/WebPlotDigitizer/app/?). The data from the literature for which action spectra have been determined are collected in an Excel file (supplementary material), which is available as a pdf in the e-version of IUVA News (posted at www.iuvanews.com).

The type of wavelength filter and the method for determining the action spectra are specified in each worksheet. The supplementary material also contains curve fits of these data using the cubic spline method. The figure numbers in the following discussion refer to worksheets in the supplementary online file.

For microorganisms where more than one action spectrum study has been carried out, there are significant differences (e.g., adenovirus 2 and Bacillus subtilis spores) (see Figs. S1a and S1c), and for others they agree well (e.g., Bacillus pumilis spores, Cryptosporidium parvum, and MS2 coliphage) (see Figs. S1b, S1d and S1e). Perhaps future work will help resolve differences.

It often has been implied that microorganism action spectra match well to the DNA absorbance spectrum. However, as seen in Figures 1a-1g, this is true only for a few microorganisms (e.g., T1 and T2 coliphages). In other cases (e.g., MS2 coliphage and Adenovirus 2), the action spectra differ markedly from the DNA absorbance with sharp increases below 240 nm and an increase around 280 nm. These deviations appear to be correlated with the absorption spectra of aromatic amino acids. This could indicate a mechanism whereby these aromatic amino acids absorb UV and transmit the excitation energy to DNA nucleotides (e.g., thymine) by excitation energy transfer.


Beck, S.E.; Wright, H.R.; Hargy, T.M.; Larason, T.C.; and Linden, K.G. 2015. Action spectra for validation of pathogen disinfection in medium-pressure ultraviolet (UV) systems, Water Res., 70: 27-37.

Bolton, J.R.; Mayor-Smith, I.; and Linden, K.G. 2015. Rethinking the concepts of fluence (UV dose) and fluence rate: The importance of photon-based units – A systemic review, Photochem. Photobiol., 91: 1252-1262.

Braslavsky, S.E. 2007. Glossary of terms used in photochemistry, 3rd ed., Pure Appl. Chem., 79(3): 293-465.

Cabaj, A.; Sommer, R.; Pribil, W.; and Haider, T. 2002. The spectral UV sensitivity of microorganisms used in biodosimetry, Water Sci. Technol. – Water Supply, 2(3): 175-181.

Chen, R-Z. 2007. Comparison of action spectra of microorganisms and DNA absorbance spectra for UV disinfection of Water, M.Sc. Thesis, University of Alberta, Edmonton, AB, Canada.

Chen, R-Z; Craik, S.A.; and Bolton, J.R. 2009. Comparison of the action spectra and relative DNA absorbance spectra of microorganisms: information important for the determination of germicidal fluence (UV dose) in an ultraviolet disinfection of water, Water Res., 43: 5087-5096.

Coohill, T.P. 1991, Action spectra again? Photochemistry and Photobiology, 54(5): 859-870.

Fluke, D.J.; Pollard, E.C. 1949. Ultraviolet action spectrum of T1 bacteriophage, Science, 110: 274-275.

Gates, F.L. 1930. A study of the bacterial action of ultra violet light. III The absorption of ultra violet light by bacteria, J. Gen. Physiol., 14: 31-42.

Haji Malayeri, A.; Mohseni, M.; Cairns, B. and Bolton, J. 2016. Fluence required to achieve incremental log inactivation, IUVA News, 18(3).

Linden, K.G.; Shin, G.; and Sobsey, M.D. 2001. Comparative effectiveness of UV wavelengths for the inactivation of Cryptosporidium parvum oocysts in water, Water Sci. Technol., 43(12): 171-174.

Linden, K.G.; Thurston, J.; Schaefer, R.; and Malley, J.P., Jr. 2007. Enhanced UV inactivation of adenoviruses under polychromatic UV lamps,” Appl. Environ. Microbiol, 73(23): 7571-7174.

Mamane-Gravetz, H.; Linden, K.G.; Cabaj, A.; and Sommer, R. 2005. Spectral sensitivity of Bacillus subtilis spores and MS2 Coliphage for validation testing of ultraviolet reactors for water disinfection, Environ. Sci. Technol., 39(20): 7845-7852.

Rauth, A.M. 1965. The physical state of viral nucleic acid and the sensitivity of viruses to ultraviolet light, Biophys. J., 5(3): 257-273.

Rochelle, P.A.; Blatchley III, E.R.; Chan, P.; Scheible, O.K.; and Shen, C.Y. 2010. Challenge organisms for inactivation of viruses by ultraviolet treatment. Final Report. Water Research Foundation and US EPA, Denver, CO.

Sutherland J.C. 2002. Biological effects of polychromatic light. Photochem. Photobiol., 76:164-170.

Zelle, M.R.; Hollaender, A. 1954. Monochromatic ultraviolet action spectra and quantum yields for inactivation of T1 and T2 Escherichia coli bacteriophages, J. Bacteriol., 68(2): 210-215.