ownes and Blunt1
conducted a series of logical, ordered experiments to examine the behavior of bacteria and fungi when exposed to sunlight. Their experiments, which were conducted in the absence of many of the analytical tools that are common among academic and industrial laboratories today, demonstrated several interesting conclusions:
(Sun)light is detrimental to the development of bacterial and fungal communities.
The extent of exposure to sunlight appears to be related to the extent to which growth of these microbial communities is suppressed.
The actinic (ie, ultraviolet [UV]) portion of the solar spectrum is primarily responsible for this behavior.
This classic work may be viewed as the first demonstration of the effects of UV radiation on microbial communities. In the time since these basic discoveries were reported, the effects of UV radiation on microorganisms and microbial communities have been elaborated in detail. Today, UV-based disinfection systems are commonly applied in many settings. Numerous important contributions to our current understanding of the mechanisms of action of UV radiation on microorganisms and critical biomolecules were presented in the 1950s and 1960s. Subsequent expressions of concern with common forms of disinfection, such as chlorination, promoted research to incorporate these findings with developments of fundamental photochemical reactor theory, which in turn facilitated rapid increases in the usage of UV radiation for disinfection of water after roughly 1980. Today, UV disinfection is among the most commonly applied disinfectants for water; it is also being used for disinfection and preservation of solids, surfaces, and air. The goal of this chapter is to describe the fundamental concepts that are needed to understand the design, behavior, and validation of UV-based microbial inactivation systems used in water treatment. Similar principles apply with applications of UV disinfection to other media. Because the process relies on photochemistry, the basic principles of photochemistry are presented first, followed by a development of some of the practical issues related to UV-based microbial inactivation processes.
BASIC PRINCIPLES OF PHOTOCHEMISTRY
Photochemical processes are initiated by absorption of electromagnetic radiation. Two basic laws of photochemistry have been presented. The first law of photochemistry, also known as the Grotthus-Draper law, states that light (electromagnetic radiation) must be absorbed for a photochemical reaction to take place. The second law of photochemistry, also known as the Stark-Einstein law, states that for each photon of radiation absorbed, only one molecule will be excited. The second law also forms the basis for the concept of a quantum yield (Φ), which defines the ratio of molecules that react to photons absorbed. In turn, this allows for assignment of 1:1 equivalence between photons and molecules. The laws of photochemistry also indicate that two conditions must be met for a photochemical reaction to take place. First, a photon must be absorbed by a target molecule, or in some cases by a neighboring molecule, which may act as a photosensitizer. Second, the absorbed photon must have sufficient energy to break an existing bond or form a new one. If these two conditions are met, it is possible, although not certain, that a photochemical reaction may take place. If either condition is not met, a photochemical reaction will not take place.
The German physicist Max Planck is credited with the development of the basic mathematical expression used to define photon energy:
E = photon energy
h = Planck constant
c = speed of light
λ = wavelength
Because reacting molecules and photons normally react on a 1:1 basis (as described by the Stark-Einstein law), it is often useful to express Planck law in an alternative form (note that 1 einstein is defined as being equal to Avogadro’s number of photons):
Where: QE = energy per einstein
A = Avogadro’s number
Much in the same way that molecules (or atoms) may be viewed as the smallest discrete units of a chemical compound, the photon may be viewed as the smallest discrete unit of radiation energy. Across the electromagnetic spectrum, there exists a wide range of wavelengths and corresponding photon energies (Figure 9.1
). The focus of this chapter is on photochemical reactions that take place within biomolecules. It is impractical to develop a comprehensive summary of bond energies associated with all moieties that are relevant in these molecules because of the dependence on molecular substituent groups and other factors. It is possible, however, to summarize representative values of the bond energies that characterize some of the common bonds in biological systems (see Figure 9.1
). From this summary, it is evident that the bond energies of interest in microbial molecules are generally coincident with photon energies in the UV portion of the spectrum. Radiation with wavelengths less than approximately 320 nm is often sufficiently energetic to promote reactions in biomolecules. Radiation in the vacuum UV (VUV) portion of the spectrum is highly energetic but also tends to be absorbed strongly. Because of this, VUV radiation does not penetrate far into liquid or solid media; it is strongly absorbed by molecular oxygen (O2
), which limits its ability to be transmitted through gases. As such, most UV photoreactors are based on radiation in the UVB and UVC portions of the electromagnetic spectrum. UV radiation is defined as the portion of the electromagnetic spectrum that lies between visible light (ie, electromagnetic radiation that can be detected by a healthy human eye) and X-radiation. The UV spectrum is further subdivided into several subregions. Table 9.1
provides a summary of wavelength cutoffs for these regions (see Figure 9.1
) as well as distinguishing features of each wavelength range.
FIGURE 9.1 Photon energy (QE) as a function of wavelength. Superimposed on this figure are the wavelength ranges that define vacuum ultraviolet (UV), UVC, UVB, UVA, and visible radiation. Also included are typical bond energies (ΔH) for a number of bonds that are relevant to UV-based applications in water treatment and disinfection, as well as their corresponding wavelengths.
Although the preceding arguments suggest that essentially all photons within the UV range should be capable of promoting photobiochemical reactions, photons at wavelengths of less than 320 nm tend to be the most effective for this purpose. Solar UV disinfection processes are strongly influenced by the effects of UVA radiation, largely by mechanisms that involve heat and the formation of O2-based reactive intermediates. For some solar UV disinfection systems, incorporation of materials that allow for inclusion of UVB radiation can increase the rate of microbial inactivation; however, the vast majority of contemporary (nonsolar) UV disinfection systems employ “artificial” sources of UVC radiation to accomplish microbial inactivation.
The rate of a photochemical reaction is directly related to the rate at which the target molecule absorbs incident radiation. For the case of monochromatic radiation, the rate of a photochemical process can be described as4
r(λ) = rate of photochemical reaction
[B] = molar concentration (activity) of absorbing (reacting) compound
t = time
Ia(λ) = volumetric rate of photon (λ) absorption by the target molecule
Φ(λ) = quantum yield
TABLE 9.1 Summary of wavelength regions within the ultraviolet spectrum and some corresponding characteristics
Type of Radiation
Nominal Wavelength Range (nm)
Comprises >99% of ambient solar UV radiation at sea level near mid-latitudes; promotes microbial inactivation through heat and formation of O2-based reactive intermediates
Comprises roughly 1% of ambient solar radiation at sea level near mid-latitudes; promotes microbial inactivation through photochemical damage to nucleic acids; promotes skin cancer and damage to other external tissues
Absent from ambient solar spectrum at sea level near mid-latitudes because of absorption by stratospheric ozone; must be generated locally to the application; causes damage to nucleic acids, proteins, and other biomolecules
Strongly absorbed by gas-phase O2 and to a lesser degree N2; highly energetic radiation that can be used to photolyze H2O and many other molecules that are normally stable
Abbreviations: H2O, water; N2, dinitrogen; O2, molecular oxygen; UV, ultraviolet.
Parenthetical terms (λ) are included in equation (3)
to emphasize that these attributes of photochemical reactions may display wavelength dependence. From this point forward, the parenthetical terms will be dropped in the interest of simplifying the resulting mathematical expressions.
The volumetric rate of photon absorbance depends on the geometry and optical characteristics of the reaction vessel in which the reaction takes place. For highly transparent conditions, or for the situation where we examine the rate of a photochemical reaction at a point in a system, equation (3)
reduces to the following:
E0 = local fluence rate
ε = molar absorption coefficient
QE = energy per einstein
also indicates that the local rate of a photochemical reaction
is proportional to the local fluence rate
. As described in the following text, the fluence rate within a photochemical reactor often displays strong spatial gradients; as such, we observe equally strong spatial gradients in local photochemical reaction rates in these systems. Because photochemical reactions tend to be rapid, mixing behavior within photochemical reactors tends to have a strong influence on reactor performance.
In the case of UV disinfection, wherein the photochemical target is one or more biomolecules within a microbe, the physical interpretations of ε
] become nebulous. Moreover, data to describe these parameters in a disinfection context are rarely available. Because of this, the common practice is to resort to a mathematical expression to describe microbial inactivation kinetics (or UV dose-response behavior) based on the same logic that is used to describe the kinetics of purely photochemical reactions. Equation (5)
illustrates the most basic form of a UV disinfection kinetic model:
N = concentration of viable microbial target organisms
k = inactivation rate constant
In effect, the inactivation constant (k) is a combined parameter that accounts for efficiency of photon absorbance (analog of ε), the efficiency of absorbed photon use to inactivate the microbial target (analog of Φ), and the energy of an einstein of photons (analog of QE). This modeling approach implicitly assumes that a 1:1 correspondence exists between a photochemical event and inactivation of an individual microbial target. In other words, this model suggests that a single photochemical event leads to inactivation of the microbial target. This approach is valid for some simple organisms, over a limited range of conditions. As will be demonstrated later, this simple model fails to describe the intrinsic kinetics of all UV disinfection processes; however, it does provide a starting point for a discussion of this topic.
Because equations (4)
describe photochemical reaction rates at any point in space, they are broadly applicable for description of the behavior of photochemical reactor systems. Their application requires detailed information regarding the spatial distribution of radiant energy (ie, the fluence rate field) within a photochemical reactor system; for many UV photoreactors, additional information is also required to develop quantitatively accurate predictions or reactor behavior.
SOURCES OF ULTRAVIOLET RADIATION
UV radiation can be generated from natural and manufactured sources. Although the source of radiation does not influence the effects of photons that it generates, it is useful to understand the basic physics of radiation generation by these sources as well as their corresponding output spectra.
The sun represents perhaps the most basic source of UV radiation.5
The sun emits radiation as a result of incandescence, a process by which any hot object emits radiation. Planck’s law can be used to simulate the spectrum of radiation emitted from a blackbody at any given temperature; blackbody emission represents the maximum possible emission spectrum for a surface at a given temperature. The spectrum of solar radiation received outside the earth’s atmosphere has been measured by satellite-based instruments and high-altitude aircraft. The spectrum of solar radiation received at earth’s surface is influenced by time of day, time of year, latitude, and altitude, as well as atmospheric composition. Numerical models have been developed to simulate these effects and to allow simulation of ambient solar spectra at essentially any time and location. One model that have been developed for this purpose is the Simple Model of the Atmospheric Radiative Transfer of Sunshine (SMARTS).6
provides a summary of several reference solar spectra. Included in this figure are two spectra that are used to describe solar radiation that is imposed on earth’s atmosphere. The term air mass zero
is used to describe these spectra. Air mass describes the path length of light through earth’s atmosphere relative to the path length that would be involved for radiation imposed at a zenith angle of zero (ie, perpendicular to earth’s surface).9
Therefore, air mass zero corresponds to solar radiation that has not yet encountered earth’s atmosphere (ie, the spectrum of radiation imposed on earth’s atmosphere). Another reference spectrum is included for air mass 1.5, which describes the spectrum of solar radiation imposed at an angle of 37 degrees. Also included in Figure 9.2
are spectra that were simulated using Planck’s law for blackbodies at 5777 K and 6000 K. These two temperatures span the nominal estimates of the surface temperature on the sun.
Reference solar spectra. Air mass zero spectra define radiation that is incident on the outside of earth’s atmosphere.7
The air mass 1.5 spectrum corresponds to solar radiation received at sea level imposed on earth’s atmosphere at a zenith angle of 37 degrees.8
The black-body spectra are presented for temperature estimates that have been used to describe the surface of the sun. For a color version of this art, please consult the eBook.
Solar radiation received at earth’s surface is substantially attenuated as compared to the air mass zero spectrum largely because of absorption by constituents of the atmosphere. Atmospheric ozone is a particularly strong absorber of UV radiation. The shortest wavelength of solar radiation received at near sea level elevations is roughly 290 nm. Earth’s atmosphere is essentially opaque to radiation at wavelengths shorter than 290 nm. This means that ambient solar radiation received at earth’s surface will include UVA and UVB radiation but essentially no UVC or VUV radiation. Although UVA and UVB radiation can be used to inactivate microbial pathogens, the most effective wavelengths for disinfection are found in the UVC range. Therefore, systems that employ UVC radiation for disinfection must rely on “artificial” sources of UV radiation. The most common artificial sources of UV radiation are mercury (Hg) lamps, of which there are two basic types: low pressure (LP) and medium pressure (MP).10
Hg lamps operate by striking an electrical arc across a space that contains Hg and an inert carrier gas, usually argon. Collisions between electrons and the atoms (or molecules) that occupy the space between the lamp electrodes result in electronic excitation of those atoms; electronically excited argon can transfer some of its energy to a nearby Hg atom. Relaxation of electronically excited Hg atoms to lower energy states results in photon emission. In addition to conventional LP lamps, there are also lamps that are based on an Hg amalgam rather than pure metallic Hg. The Hg amalgam lamps are characterized by essentially the same output spectrum as conventional LP Hg lamps (Figure 9.3
), but they can generate more power than their conventional LP Hg lamp analogs. For practical purposes, LP Hg lamps are often considered to be monochromatic (λ = 253.7 nm) sources of UV radiation; however, LP Hg lamps are also characterized by several other, smaller lines in their output spectra, including a line at 185 nm and a few lines in the visible range that are responsible for the pale blue color that is evident when these lamps are illuminated.
The output spectra of MP Hg lamps are considerably more complex than those of LP Hg lamps. The MP lamp output spectra are characterized by numerous discrete lines superimposed on a continuum. The MP lamps tend to be much less efficient than LP Hg lamps in terms of
conversion of electrical power (input) into UV photons (output). The MP lamps can be much more powerful than LP lamps, and as such, they can be used to develop systems that are characterized by a smaller footprint and less complexity than equivalent systems based on LP Hg lamp technology. A normalized MP lamp spectrum is included in Figure 9.3
. The spectra presented for LP (amalgam) and MP Hg lamps were normalized against the maximum values that were measured for each lamp type. Normalization was applied because the output spectrum measured for any lamp will depend on the location of the measuring device relative to the source as well as the composition of the media located between the source and the measurement location.
FIGURE 9.3 Normalized output spectra of Hg lamps. For both lamp types, spectra are normalized to the maximum value measured in the spectrum. Abbreviation: MP, medium pressure. Data courtesy of Trojan Technologies. For a color version of this art, please consult the eBook.
Historically, Hg lamps have been viewed as the default option for generation of germicidal UV radiation. However, the inclusion of Hg in these devices presents safety risks that motivate the development of alternative sources of UV radiation. From the regulatory perspective, some pressure to develop alternatives to Hg lamps can be found in the Minamata Convention on Mercury
an international treaty that was developed to protect human health and the environment from anthropogenic emissions of Hg. A similar piece of legislation in Europe has also resulted in restrictions on the use of Hg.12
As with the Minamata Convention
, Hg-based lamps for general lighting are allowed, but there are restrictions on the mass of Hg that can be included in these lamps.
Excilamps represent a family of UV and VUV sources that include excimers (exci
) and exiplexes (exci
). In both cases, a diatomic molecule or complex is electronically excited; relaxation back to a ground state results in photon emission at a characteristic wavelength. Excilamps are characterized by relatively narrow emission spectra, typically in the range of 2 to 30 nm at half maximum.10
,14 Figure 9.4
illustrates the normalized spectral output of two common exciplex lamps: krypton chloride and xenon bromide. The output spectrum of an exciplex lamp can be adjusted through selection of the constituent gases that are used to form the exciplex. The most common configuration for these lamps is a so-called dielectric barrier discharge lamp,15
which usually is manufactured with a cylindrical geometry. Many other geometric configurations are possible with these lamps, including flat-panel displays. This geometric flexibility may provide opportunities to improve the optical characteristics of UV photoreactors.
FIGURE 9.4 Normalized output spectra of krypton chloride (KrCl) and xenon bromide (XeBr) exciplex lamps. For a color version of this art, please consult the eBook.
UV light-emitting diodes (LEDs) have emerged as important sources of antimicrobial UV radiation in recent years.18
Diodes are semiconductor materials that have been doped with chemicals to alter their ability to carry electrical current. Doping materials are added to semiconductors, such as silicon, to allow formation of charge carriers. A so-called p-n (positive-negative) junction is established by bringing into contact materials that have been doped to promote formation of positive charge carriers (holes) and negative charge carriers (electrons). Holes and electrons naturally migrate toward each other because of electrostatic attraction, but their migration can be influenced by application of electrical current. When applied properly, this current promotes migration of holes and electrons toward each other, resulting in their recombination. In turn, this causes the electron to experience a drop in energy, yielding a photon that corresponds to the bandgap of the dopant materials. Visible and infrared LEDs are commonly applied in electronic devices, including cell phones, televisions, room lighting, etc. The production of UV LEDs is possible through the use of group III nitrides as doping materials, especially including the aluminum-gallium nitride doping system. The spectral
output of these devices depends on the composition of the dopants. Figure 9.5
provides an illustration of the range of outputs observed among commercially available UV LEDs. Because of this, it is possible to select UV LEDs to have an output spectrum that is tailored to a given application.
FIGURE 9.5 Normalized output spectra of ultraviolet light-emitting diodes as a function of aluminum-gallium composition (data from AquiSense Technologies). Parenthetical entries in the legend indicate peak width (nm) at half maximum. For a color version of this art, please consult the eBook.
UV LEDs have a number of other important advantages relative to conventional sources of UV radiation, such as Hg lamps. They contain no Hg, which is relevant in the context of the Minamata Convention and general public pressure. The UV LEDs are compact and durable and allow for instant on/off switching. Packaging of LEDs (including UV LEDs) allows for inclusion of optical elements that can provide for directional irradiation. As such, UV LEDs can function as nearly point sources of radiation, which offers considerable flexibility in reactor design. On the other hand, it is also important to recognize the drawbacks of UV LEDs. At present, the efficiency of these devices is lower than Hg lamps, and their costs are higher. Historic trends in LED development have indicated that as applications for these devices emerge, production processes improve to yield increasingly efficient and powerful LEDs.