United States Environmental Protection Agency Water Engineering Research Laboratory Cincinnati OH 45268 Research and Development EPA/600/S2-86/067 Sept. 1986 <>EPA Project Summary Inactivation of Microbial Agents by Chemical Disinfectants RECEIVED John C. Hoff NOV181986 ENVIRONMENTAL PROTECTION AGENCY LIBRARY, REGION V Drinking water disinfection kinetics are used to evaluate Escherichia coli, poliovirus, and Giardia lamblia cysts with regard to their relative resistance to inactivation under a variety of physi- cal and chemical conditions. The report explains the concept of C-t product (the product of residual disinfectant, C, in mg/L and contact time, t, in minutes) and reviews the effects of temperature and pH on C-t values. The limitations and dangers of extrapolating C-t values beyond the range of experimental data are also discussed. This Project Summary was devel- oped by EPA's Water Engineering Re- search Laboratory, Cincinnati, OH, to announce key findings of the research project that is fully documented in a separate report of the same title (see Project Report ordering information at back). Introduction The primary purpose of drinking water disinfection is to control water- borne diseases by inactivating the pathogenic microorganisms in the water. Disinfection is the final (and sometimes only) engineering process barrier to the entry of viable pathogens into the water distribution system. After chlorine began to be used as a drinking water disinfectant around the turn of the century, interest in its bioci- dal effectiveness brought about disin- fection research. Information on the kinetics of disinfection was soon devel- oped. Since the early 1970's, concern about chemical by-products of chlorina- tion has resulted in a higher level of re- search activity involving alternative dis- infectants, including chloramine, ozone, and chlorine dioxide. The early disinfec- tion research was focused on inactiva- tion of bacteria. Viruses were studied later. Mostrecently, inactivation of Giar- dia cysts has been the topic of much research work. This report presents a comprehensive review of disinfection research. The concepts of disinfection kinetics that were developed by early researchers and later modified are used in this re- port to evaluate Escherichia coli, po- liovirus, and Giardia cysts with regard to their relative resistance to inactiva- tion under a variety of physical and chemical conditions. The document ex- plains the concept of C-t product (the product of residual disinfectant, C, in mg/L and contact time, t, in minutes) and reviews the effects of temperature and pH on C-t values. The limitations and dangers of extrapolating C-t values beyond the range of experimental data are also discussed. Disinfection Kinetics Inactivation of microorganisms can be considered to have the characteris- tics of a first-order chemical reaction, with the microorganism and the disin- fectants constituting the reactants. This concept was expressed as Chick's Law and is written as logN/N0=-K-t (1) where N0 = the original number of or- ganisms N = the number of organisms remaining at time t t = the contact time K = a proportionality constant ------- Ideally, plots of log N/N0 versus t for various contact times should provide a straight line (first-order kinetics). In ac- tual experiments, first-order kinetics are often not observed throughout the en- tire range of experimental conditions, but rather during only a portion of the experiment. Thus survival curves may depart from the ideal (Figure 1 a) and show (1) an initial lag period before first-order kinetics are observed (Figure 1 b), (2) a rapid initial decline in popula- tion (Figure 1 c), or (3) multiple kinetics sometimes referred to as "tailing off" (Figure 1 d). Experimental disinfection data commonly fail to follow first-order kinetics strictly (Figure 1 a). Other disin- fection kinetic models have been pro- posed, but they are not reviewed in this report. When the biocidal efficacy of disinfec- tants are compared, the major consider- ations are disinfectant concentration and time needed to inactivate a certain proportion of the population of exposed organisms. The C-t concept can be ex- pressed as k = Cn • t (2) where C = disinfectant concentration, mg/L n = a constant, also called the coefficient of dilution t = the contact time (minutes) required to inactivate a specified percentage of mi- croorganisms k = a constant for a specific mi- croorganism exposed under specific conditions To apply Equation 2 to disinfection data, the results are used from a number of individual experiments performed with different disinfectant concentrations under identical experimental condi- tions. Disinfectant concentrations (C) and times (t) needed to attain the speci- fied degree of inactivation (e.g., 99%) are plotted on double logarithmic pa- per. Such plots should produce a straight line with a slope of n. When n = 1, the C-t value remains constant re- gardless of disinfectant concentration, and disinfectant concentration and ex- posure time are of equal importance. If n exceeds 1, disinfectant concentration is more important than contact time, and C-t values required for a specified kill decline as C increases. On the other hand, when n is less than 1, contact time is more important than disinfectant con- centration, and C-t values for a specified kill increase as C increases and t de- creases. The value of n is an important factor in determining the degree to which ex- trapolation may be valid beyond the range of experimental observations. In addition, evaluating n is valid only if the experimental data follow Chick's Law (Equation 1), which is often not the case. Values of n have been evaluated, and results generally fall in the range of 0.5 I I Time Exponential Kinetics Time Concave Upward Kinetics (Initial Shoulder Curve) I 1 Time Concave Downward Kinetics (Initial Rapid Rate Curve) Time Multiple Kinetics (Tailing Off Curve) Figure 1. Typical survival curves for disinfection experiments. Adapted from: Prokop. A., and A. E. Humphrey, 1970. Kinetics of disinfection. In: Disinfection, M. A. Bernarde, ed.. MarcelDekker. Inc.. N.Y. pp. 61-83. ------- to 2. Because of uncertainties about n, extrapolation of data from specific C and t conditions to other values under the assumption that n = 1 would be of questionable validity. Tables of n values for free chlorine, chloramine, chlorine dioxide, and ozone are given in the full report. Temperature Effects Effects of temperature change on dis- infection efficacy have been evaluated by a number of investigators. Disinfec- tion rates are generally increased by a factor of 2 to 3 as temperature increases by 10°C. This coefficient is referred to as the QIQ value. Reported Q10 values for viruses are usually in the range of 2 to 3. A slightly wider range of QIO values is found for disinfection studies in which ozone and chlorine dioxide have been used. Some concerns have been ex- pressed that as temperature ap- proaches 0°C, disinfection rates might decrease by a much greater factor than would be indicated by Q10 values. How- ever, no aspect of physical laws that govern chemical diffusion and reaction rates in aqueous media would support such a concept. Thus the common rule of a 2- to 3-fold increase in inactivation rates per 10°C increase in temperature seems fairly well substantiated. Characteristics of Disinfectants and Microorganisms Disinfectants The disinfectants reviewed in this re- port (free chlorine, chloramines, chlo- rine dioxide, and ozone) have individual characteristics that influence both the results of laboratory tests and their per- formance in the field. These characteris- tics are reviewed here briefly. Free chlorine exists in aqueous solu- tion as HOCI and OCI~. HOCI is a much more effective biocide than OCI~, so the efficacy of free chlorine is pH- dependent. Chloramines are formed when chlo- rine and ammonia react. Chloramines are generally much less effective than free chlorine, with equivalent inactiva- tion times about 25- to 100-fold higher for monochloramine than for equivalent concentrations of free chlorine. Chlo- ramine efficacy is also pH-dependent. Application of chloramine laboratory re- sults to field conditions is fraught with uncertainty. Most laboratory studies have been done with preformed chlo- ramine, whereas treatment plant prac- tice would result in at least some con- tact with free chlorine before chloramines are formed, even when the order of addition is ammonia first and chlorine second. Chloramine disinfec- tion as practiced in the field may be more effective than laboratory results would suggest, but the extent of this im- provement would be site-specific and would need to be evaluated on a plant scale at each site. Chlorine dioxide efficacy is less sub- ject to the influence of pH than either free chlorine or chloramine. Chlorine dioxide is a more effective disinfectant at pH 9 than at pH 6. This is a reversal of the behavior shown by free chlorine and chloramine. Because it is present in water as an undissociated dissolved gas, chlorine dioxide is more easily lost through volatilization than free chlorine or chloramine. This behavior could af- fect the kinetics of disinfection experi- ments with long exposure times, espe- cially at higher water temperatures. Ozone, like chlorine dioxide, is present in water as a dissolved gas, must be prepared onsite, and cannot be stored. Ozone is subject to losses by volatilization during disinfection experi- ments. The volatility and high reactivity of ozone make it very difficult to main- tain a stable concentration during ex- periments and in actual practice. For these reasons, C-t values for ozone tend to be less precise than C-t values for the other disinfectants. Microorganisms Waterborne pathogens of concern can be divided into three groups: bacte- ria, viruses, and protozoan cysts. They encompass a wide diversity of sizes, life cycles, and other biological characteris- tics, including resistance to chemical disinfectants. Note that even within dif- ferent isolates of the same species, re- sistance to disinfection can vary. Fur- thermore, differences in disinfection resistance have been observed between organisms that were cultured in the lab- oratory and those found naturally in the environment. Finally, differences exist in relative resistance to various chemi- cal disinfectants. Whereas organism A might be more resistant to chlorine than organism B, the opposite might be ob- served for chloramine or chlorine diox- ide. Application of the C-t Concept to Disinfection Practice In 1962, Watson's Law (k = Cn-t) was used as a basis for a procedure for mak- ing recommendations on disinfection practice. The C-t value recommenda- tions were based on constant C-t values, making an implicit assumption that n = 1. This use of the C-t concept may have been the first to relate disinfection laboratory data to recommended field practice. C-t values were used to compare bio- cidal efficiencies in 1980, but the back- ground of the concept was not ex- plained, and no attempt was made to extrapolate to other values for either C or t from those calculated from avail- able data. The use of C-t values to interpret dis- infection data has become more preva- lent in the 1980's. The 99% inactivation level has been used for calculating C-t values in most studies, probably be- cause it is the level at which exponential kinetics (N/N0 = K-t) are usually best ap- proximated. If exponential kinetics were followed, and if C-t values for 99% inac- tivation were known, C-t values for other levels of inactivation could easily be calculated. The ideal is not often ob- served, though. Problems associated with initial lags (Figure 1 b) and tailing off (Figure 1 d) make it difficult to calcu- late C-t values for conditions not directly observed in experiments. These diffi- culties should be noted when applying data from the following section of this report. Inactivation of Microorganisms Bacteria Though pathogenic bacteria are among the target organisms for disin- fection, little information is available on their inactivation. Most of the research related to bacteria has focused on indi- cator organisms. Studies in the 1940's did not reveal substantial differences in disinfection resistance between bacte- rial pathogens and members of the col- iform group. Thus data for E. coll should indicate the degree of disinfection needed for the pathogenic bacteria. Two factors that can influence disin- fection results are the relative resis- tance of laboratory-grown cultures ver- sus that of natural organisms and protection of bacteria by particulate matter. Cell cultures grown in the labo- ratory are more easily inactivated. Bac- teria that are within particles of feces or other organic matter or that are at- tached to activated carbon particles are not inactivated as readily as bacteria that are not associated with such partic- ulate matter. ------- In the full report, data show ranges of C-t values for 99% inactivation of E. coli by free chlorine, chloramine, chlorine dioxide, and ozone. With free chlorine, the range of experimental conditions for which data are available is some- what reduced for E. coli. The reason is that at low pH and high temperature (pH 6, 25°C), inactivation proceeds so fast that C-t measurements are difficult to attain with confidence. C-t values for free chlorine are given for pH 6 and 10, and for 5° and 15°C. The mean C-t for 99% inactivation at 5°C and pH 6 was 0.045 mg/L • minutes. For chloramine at 5°C and pH 7, the mean C-t was 22 mg/L • minutes. Chloramine data are given for pH 7 and 9, and for 5°, 15°, and 25°C. The mean C-t for chlorine dioxide at 5°C and pH 6.5 was 0.6 mg/L • minutes, a higher value than that observed for free chlo- rine. This level contrasts with the rela- tive efficacy of free chlorine and chlo- rine dioxide for poliovirus and Giardia cysts. In both of the latter cases, chlo- rine dioxide is the more powerful disin- fectant. Chlorine dioxide data span a temperature range of 5° to 25°C and a pH range of 6.5 to 7. A mean C-t value of 0.2 mg/L • minutes was obtained for 99% inactivation of E. coli by ozone at pH 7.2 and 1°C. Ozone data are also available at 12°C and pH 7.0. For the four disinfectants, the n values were generally near 1 when E. coli was the target organism. The ranges of C-t values were relatively narrow for exper- iments conducted at the same pH and temperature using different disinfectant concentrations. This result suggests that C-t values for E. coli are relatively reliable. Viruses The most extensive research on virus inactivation has been done with mem- bers of the enterovirus group because the viral agents responsible for water- borne disease (Hepatitis A virus, ro- ta virus, Norwalk virus, etc.) were identi- fied only recently. Methods for laboratory growth and enumeration of the pathogenic viruses are difficult or not yet available. Most of the disinfec- tion data presented in this report are for poliovirus. Factors involved in viral resistance to disinfection are their natural or innate resistance, aggregation into virus clumps, and association with particu- late materials. Research results suggest that viral aggregation or clumping can cause deviation from exponential inacti- vation kinetics, particularly the tailing off curves (Figure 1 d). The protective effects of paniculate matter are similar for viruses and bacteria. The best pro- tection is offered by virus-particle com- plexes associated with human fecal ma- terial. Viral clumps in fecal particles are most likely to be highly protected from inactivation. The viral C-t data base is largest for poliovirus 1. For 99% inactivation with free chlorine, C-t values averaged 1.1 and 2.0 mg/L • minutes for two different researchers. For 5°C and pH 10, C-t aver- aged 10.5 mg/L • minutes. Data are available for 5° and 15°C, and for pH 6 and 10. In contrast to these values for free chlorine, the 99% inactivation value for chloramine at 5°C and pH 9 averaged 1420 mg/L • minutes, indicating that chloramine is a very weak viral disinfec- tant. Chloramine data are available at pH9and5°, 15°, and 25°C. Chlorine dioxide was as effective as free chlorine, with a mean C-t of 3.6 mg/ L • minutes for 99% inactivation at 5°C and pH 7. Data are available for pH 7 and 9, and for 5° to 25°C. Ozone was the most effective agent. A 99% inactivation was attained at 5°C and pH 7.2 with a mean C-t of 0.2 mg/L • minutes. Ozone data are available for 5°, 10°, and 20°C, and for pH 7.0 or 7.2. A limited number of other data are also presented in the full report. Overall, the C-t values for poliovirus 1, rotavirus, and bacteriophage f2 are similar. Labo- ratory studies done with preformed chloramine indicate that all of these viruses are extremely resistant to chlo- ramine. The apparent biocidal efficiency of chloramine as it is used in water- works practice would be higher because of the free chlorine that is present for a short time. Protozoan Cysts The inactivation of Endomoeba his- tolytica cysts by chlorine and other dis- infectants was studied extensively dur- ing the 1940's and 50's, mainly because of concerns about waterborne transmis- sion of amoebic dysentery in military forces operating in areas where this dis- ease was prevalent. These studies es- tablished conclusively that the cysts of E. histolyticawere very resistant to inac- tivation. The appearance of giardiasis as an important waterborne disease in the United States stimulated disinfec- tion research on the inactivation of cysts of the etiologic agent Giardia lamblia. A method for determining cyst viability by in vitro excystation was developed, but problems developed in obtaining G. lamblia cysts, and deficiencies occurred in the excystation procedure. Thus most disinfection research is currently con- ducted using G. muris cysts (a species infective for mice) as a model for G. lamblia cysts. This approach seems to work well. A comparative study of ex- cystation and mouse infectivity for measurement of chlorine-exposed G. muris cysts indicated that the results were similar for both methods. Giardia lamblia has a complex life cycle. The conversion from the active trophozoite to the inactive, resistant cyst occurs in the lower portion of the intestinal tract. The cysts do not multiply and are rela- tively inert in the environment, excyst- ing to form the trophozoite stage only after ingestion by the host. Because the cysts are relatively large (ovoid bodies 8 to 12 by 7 to 10 urn in diameter), protection from inactivation by association with particulate matter may be less important for them than for smaller, more easily occluded patho- gens. Little information has been devel- oped on this subject. The largest data base available for Gi- ardia is from disinfection research with G. muris cysts. The available data for cysts of the human pathogen G. lamblia also are included in the report. Mean C-t values for 99% inactivation of G. lamblia cysts by free chlorine range from 65 to greater than 150 mg/L • minutes for 5°C and pH 6. Data are available for 5°, 15°, and 25°C, and for pH 6, 7, and 8. Data on G. muris cover the range of 3° to 25°C, and pH 5 to 9. At 5°C and pH 6, a C-t of greater than 150 mg/L - minutes has been reported. How- ever, researchers have also obtained a C-t of 68 mg/L • minutes for 3°C, pH 6.5, and 99% inactivation. Chloramine data are available for G. muris in a temperature range of 3° to 18°C and a pH of 6.5 to 8.5. A mean C-t of 463 mg/L • minutes was obtained for 99% inactivation at 3°C, pH 6.5, but the chloramine was not preformed. At 18°C and pH 7, a C-t of 184 mg/L • minutes was obtained when chloramine was not preformed. In contrast, at 15°C and pH 7, a mean C-t of 848 mg/L • minutes re- sulted from use of preformed chlo- ramine. Data for 99% inactivation of G. muris by chlorine dioxide are available for 5° and 25°C at pH 7 and 9. At 5°C and pH 7, the mean C-t is 11.2 mg/L • minutes. This value is about one order of magni- tude lower than those for free chlorine ------- and chloramine, and it suggests that chlorine dioxide is a powerful cysticidal agent. Ozone disinfection data are available for G. muris and G. lamblia at pH 7 from 5° to 25°C. At pH 7 and 5°C, the mean C-t value was 1.9 mg/L • minutes for 99% inactivation of G. muris and 0.6 mg/L • minutes for G. lamblia. Ozone appears to be somewhat more effective against cysts than chlorine dioxide. A summary of the comparative effi- ciency of free chlorine, chloramine, chlorine dioxide, and ozone for inactiva- tion of specific bacteria, viruses, and protozoan cysts appears in Table 1. Ozone shows the highest efficiency, in- activating 99% of all types of microor- ganisms at very low C-t values. Chlo- ramine shows the lowest efficiency. Chloramine C-t values for viral agents are particularly high. Free chlorine at pH 6 to 7 and chlorine dioxide at pH 7 are approximately equivalent for poliovirus 1 inactivation. Free chlorine appears to be considerably more effective than chlorine dioxide for inactivation of ro- tavirus, bacteriophage f2, and E. coli, whereas chlorine dioxide appears to be much more effective than free chlorine for G. muris cysts. The data in Table 1 also show the relative variability in re- sistance among and within the groups of microorganisms. The general pattern of greater resistance of cysts compared with viruses, and of viruses compared with bacteria is evident for free chlorine, chlorine dioxide, and ozone. Although cyst C-t values for preformed chlo- ramines at 5°C are not yet available, the available values at 15°C suggest that cysts may be more sensitive to pre- formed chloramine than the viruses. The bacteriophage f2 C-t values sug- gest that the use of this virus to indicate virus inactivation is questionable. On the other hand, poliovirus appears to be a relatively good indicator, since it is substantially more resistant to free chlorine and chlorine dioxide than ro- ta virus and bacteriophage f2. Rotavirus, however, appears to be somewhat more resistant to preformed chloramine than poliovirus 1. Finally, G. muris cysts appear to be somewhat more resistant than G. lam- blia cysts to free chlorine and ozone. Also, considerable uncertainty exists with regard to C-t values for 99% inacti- vation of cysts by free chlorine. Values derived from studies by different inves- tigators show substantial variation. Table 1. Summary of C-t Value Ranges for 99% Inactivation of Various Microorganisms by Disinfectants at 5°C Disinfectant Micro- organism E. coli Polio 1 Rotavirus Bacterio- phage f2 G. lamblia cysts G. muris cysts Free Chlorine, pH6to7 0.034-0.05 1.1-2.5 0.01-0.05 0.08-0.18 47- > 150 30-630 Preformed Chloramine, pH8to9 95-180 768-3740 3806-6476 — — — Chlorine Dioxide, pH6to7 0.4-0.75 0.2-6.7 0.2-2. 1 — — 7.2-78.5 Ozone, pH6to7 0.02 0. 1-0.2 0.006-0.06 — 0.5-0.6 7.8-2.0 All of the C-t values discussed above are based on 99% inactivation of the mi- croorganisms. As indicated, the nature of inactivation curves prevents extrapo- lation from the 99% inactivation C-t val- ues to obtain reliable C-t values for other levels of inactivation (e.g., 50%, 90%, 95%, 99.9%, etc.). The curves nearly always show either an initial shoulder, tailing off, or other more com- plex configurations (see Figure 1). Ex- trapolation from initial shoulder curves on the 99% inactivation level (assuming exponential inactivation rates*) will underestimate C-t values for less than 99% inactivation and overestimate C-t values for greater than 99% inactiva- tion. Extrapolation from initial rapid rate and tailing off curves will usually have the opposite effect, depending on the point at which the inactivation rate be- gins to decrease. Thus extreme caution must be used if any attempt is made to extrapolate C-t data to other inactiva- tion percentages. Conclusions 1. The C-t values compiled provide a basis for comparing the effective- ness of different disinfectants for in- activation of specific microorgan- isms and for comparing the relative resistance of different microorgan- isms to specific disinfectants. In some cases, the C-t values derived from exposure to different concen- trations of the same disinfectant under specific pH and temperature 3. *A straight line extended from 100% survival at time 0 through the 99% inactivation time point. conditions show little variation, and in other cases, a wide range of C-t values occurs. Discerning the rea- sons for widely differing values is al- most always difficult, whether con- sidering the results from only one in- vestigation or from several. These factors make it difficult to pinpoint disinfection requirements. C-t values must be used cautiously to evaluate disinfection practice or to establish disinfection criteria for use in the field, and appropriate safety factors must be incorporated into the C-t values. Some major problems in applying the results of C-t values to develop- ing disinfection requirements are as follows: a) the failure of disinfection data to follow the exponential rates described by the empirical C-t equa- tion, b) differences in disinfection re- sistance between different isolates of the same species and between differ- ent species within groups (bacteria, viruses, cysts), c) state-of-the- microorganism effects such as ag- gregation, prior growth conditions, and protective effects that cannot be factored into the values, d) influence of experimental conditions (mixing intensity, disinfectant concentration variations, etc.) on inactivation rates, e) problems relating to the relevance of laboratory data to field conditions. Because of the limited data available, uncertainties still remain regarding disinfection requirements for Giardia lamblia cysts. Most of the data avail- able for G. lamblia cysts indicate lower requirements than do the ------- more extensive data available for the model G. muris cysts. These uncer- tainties are very important because this pathogen is the most resistant of all waterborne pathogens of con- cern. Additional research is in prog- ress using G. muris cysts and chlo- ramines and free chlorine at low temperatures. Other research indi- cates that an alternative to the excys- tation method for determing G. lam- blia and G. muris viability may soon be available. Such an alternative would facilitate disinfection research on this species. The method would still involve microscopic observation and therefore would not result in im- proved detection of viability at low cyst concentrations (greater than 99% inactivation). 4. For some disinfectants (mainly chlo- ramines), utilities should perhaps be required to demonstrate the efficacy of their disinfection practices for con- trolling pathogens of concern. This alternative approach may well be warranted because of the extreme dependence of chloramine disinfec- tion efficiency on field conditions that cannot all be taken into account in developing overall C-t values. 5. For inactivation by free chlorine, pH is a very important factor because of the rapid decrease in the more effec- tive disinfectant chemical species (HOCI) that occurs over a pH range of 7 to 8. Many natural waters fall into this pH range. Monitoring of pH and subsequent pH modification may be advisable in some cases to enhance disinfection efficiency, particularly at low temperatures. ------- TheEPA author JohnC. Hoff (see below) is withthe Water Engineering Research Laboratory. Cincinnati, OH 45268. The complete report, entitled "Inactivation of Microbial Agents by Chemical Disinfectants, "(Order No. PB 86-232 568/AS; Cost $11.95, subject to change) will be available only from: National Technical Information Service 5285 Port Royal Road Springfield, VA 22161 Telephone: 703-487-4650 John C. Hoff can be contacted at: Water Engineering Research Laboratory U.S. Environmental Protection Agency Cincinnati, OH 45268 United States Environmental Protection Agency Center for Environmental Research Information Cincinnati OH 45268 BULK RATE POSTAGE & FEES PAI EPA PERMIT No G-35 Official Business Penalty for Private Use S300 EPA/600/S2-86/067 CM It 60604 ------- |