THE EFFECT OF PHYSICOCHEMICAL PROPERTIES OF SECONDARY TREATED WASTEWATER … · 2014. 3. 5. · ii The Effect of Physicochemical Properties of Secondary Treated Wastewater Flocs on
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
THE EFFECT OF PHYSICOCHEMICAL
PROPERTIES OF SECONDARY TREATED
WASTEWATER FLOCS ON UV DISINFECTION
By
Yaldah Azimi
A thesis submitted in conformity with the requirements
for the degree of Doctor of Philosophy,
Department of Chemical Engineering and Applied Chemistry
The Effect of Physicochemical Properties of Secondary Treated Wastewater Flocs on UV
Disinfection
Doctor of Philosophy 2013
Yaldah Azimi
Department of Chemical Engineering and Applied Chemistry
University of Toronto
ABSTRACT
The microbial aggregates (flocs) formed during secondary biological treatment of
wastewater shield microbes from exposure to ultraviolet (UV) light, and decrease the
efficiency of disinfection, causing the tailing phenomena. This thesis investigates whether
the formation of compact cores within flocs induces higher levels of UV resistance.
Moreover, it investigates the effect of secondary treatment conditions on the
physicochemical properties of flocs’, effluent quality, and UV disinfection performance.
Compact cores were isolated from the flocs using hydrodynamic shearing. The UV dose
response curves (DRC) were constructed for flocs and cores, and the 53-63 µm cores
showed 0.5 log less disinfectability, compared to flocs of similar size. Based on a
structural model developed for the UV disinfection of flocs, floc disinfection kinetics was
sensitive to the core’s relative volume, their density, and viability.
The UV disinfection and floc properties of a conventional activated sludge (CAS) system,
and a biological nutrient removal (BNR-UCT) system, including both biological nitrogen
and phosphorus removal, was compared. The 32-53 µm flocs and the final effluent from
the BNR-UCT reactor showed 0.5 log and 1 log improvement in UV disinfectability,
iii
respectively, compared to those from the CAS reactor. The BNR-UCT flocs were more
irregular in structure, and accumulated polyphosphates through enhanced biological
phosphorus removal. Polyphosphates were found to be capable of producing hydroxyl
radicals under UV irradiation, causing the photoreactive disinfection of microorganisms
embedded within the BNR-UCT flocs, accelerating their UV disinfection.
Comparing the UV disinfection performance and floc properties at various operating
conditions showed that increasing the operating temperature from 12 ºC to 22 ºC, improved
the UV disinfection of effluent by 0.5 log. P-Starved condition, i.e. COD:N:P of
100:10:0.03, decreased the average floc size and sphericity, both by 50%. Despite the
higher effluent turbidity of the P-Starved reactor, the final effluent’s UV disinfection
improved by at least 1 log compared to the P-Normal and P-Limited conditions. The
improvement in the floc and effluent disinfectability were accompanied by a decrease in
floc sphericity and a decrease in the number of larger flocs in the effluent, respectively.
iv
FOREWORD
I would like to express my sincere gratitude to the following people and institutions for all the people and institutions that supported me over the course of this research. I especially would like to thank: Prof. Ramin Farnood and Prof. Grant Allen, my supervisors, for their invaluable support and guidance over the past few years. I am thankful for their confidence in my work and their commitment to my education and professional development. Prof. Gideon Wolfaardt and Prof. Yuri Lawryshyn, members of my reading committee, for providing me with their feedback and helpful suggestions. The Natural Sciences and Engineering Research Council of Canada, and Trojan Technologies for financially supporting this project to completion. I would like to specifically thank Dr. William Cairns and Dr. Ted Mao for their constructive feedback The National Water Research Institute (NWRI), Environment Canada, in Burlington. They supported me through providing the facilities required for operating pilot scale experiments. Specifically, I thank Dr. Peter Seto and Scott Dunlop at the Water Science & Technology Directorate. The administrative and technical staff in the Department of Chemical Engineering & Applied Chemistry at the University of Toronto for their friendly support: Gorette Silva, Pauline Martini, Joan Chen, Daniel Tomchyshyn, Julie Mendonca, Anna Ho, Cindy Tam, Isabella Medina, Mary Butera, thank you all. The summer students and 4th year thesis students that helped me throughout my experiments: Lucile Paez, Benoit Barroso, Sara kirchner, Yao (Doria) Wang, Daphne Wilson, many thanks for all your hard work. And finally I would like to thank my friends and fellow lab mates for their moral support throughout my graduate studies.
et al. 2000, Das 2001). Liu et al. showed that floc particles formed by coagulation and
flocculation lead to lower (more than 1-log) inactivation of E.coli at UV doses ranging
from 10-40 mJ/cm2 compared to non particle associated E.coli at similar turbidity (Liu et al.
2007). Tan found that similar to the UV DRC of whole effluents, a suspension of
wastewater flocs with a narrow size distribution shows the above mentioned distinct
regions (Tan 2007). Since the effect of free-swimming microbes was eliminated in Tan’s
experiments, he attributed the initial steep slope and tailing regions in the UV DRC to the
inactivation of UV-susceptible flocs and the UV-resistant flocs, respectively. It has been
repeatedly reported that larger flocs have a more pronounced worsening effect on the
18
efficiency of UV disinfection (Batch et al. 2004, Blume & Neis 2004, Madge & Jensen
2006, Templeton 2005, Tan 2007, Kollu & Örmeci 2012). Therefore, prefiltration,
sonication, and hydrodynamic shearing have been among suggested methods of decreasing
floc size distribution and improving the efficiency of UV disinfection (Qualls et al. 1985,
Liltved &Cripps 1999, Gibson et al. 2008, Yong et al. 2009, Best 2012).
Several researchers have suggested that besides floc size, other factors such as floc
composition and structure, which affect the inter-floc absorbance of UV light, affect UV
disinfection. For instance, flocs containing humic substances were found to provide more
effective shielding of microorganisms compared to clay minerals and activated sludge flocs;
therefore, showing a stronger decrease in the efficiency of UV disinfection (Bitton et al.
1972, Templeton et al. 2005). Iron is a strong absorber of UV light (Darby et al. 1993,
Kozak et al. 2010); therefore, when embedded in the flocs or free in the solution, it is a
factor that could significantly reduce the efficiency of UV disinfection (Nessim & Gehr
2006). Loge et al. (1999) suggested that target organisms with a direct light pathway to the
bulk solution are easier to disinfect with UV light, which emphasizes the effect of floc
structure on UV disinfection.
2.3.3. Modeling the UV Dose Response
Microorganism populations are exposed to a distribution of UV radiation doses
depending both on their positioning in the UV reactor, and within the floc. The dose
distribution in the reactor affects the received UV dose at the surface of the flocs. However,
the effective dose received by a microorganism that is embedded within the floc is affected
by the flocs’ physiochemical properties.
19
Various kinetics models have been utilized to estimate the impact of wastewater
quality on the reactor performance and for effective reactor design. The one hit model
assumes that a single hit is sufficient to inactivate a microorganism (Zimmer 1961). The
probability of microorganism survival corresponds to the probability that the effective
cross-section of the microorganism escapes the photons. In this case the probability of
survival will correspond to first order kinetics with respect to UV dose:
Dke
N
N0
0
−=
where N is the number concentration of survived bacteria at dose D, and N0 is the number
concentration of bacteria at dose zero, and k0 is the inactivation constant. It should be
noted here that this model is applied in the absence of particulate matter. For particle
associated microorganisms the received dose within the flocs is less than D. Dose is a
product of UV intensity and exposure time, and as there is an exponential decay in UV
light intensity within the flocs (according to the Beer-Lambert’s law), therefore, the
effective dose decreases within the floc.
The double exponential model was later developed by Jagger (1977) that represents
the disinfection of a heterogeneous population with respect to radiation sensitivity. The
population consists of two groups, both inactivated in a single-hit fashion, but one group is
much more sensitive than the other. The model is as follows:
DkDkee
N
N21)1(
0
−− +−= ββ
where D is the delivered ultraviolet dose, N is the number of surviving microorganisms in
the irradiated sample, and N0 is the number of microorganisms in the unexposed sample.
Parameters k1 and k2 are the UV inactivation rate constants; k1 is the inactivation constant
20
for the group of microorganisms much more sensitive to irradiation, and k2 is the
inactivation constant of the less sensitive group of microorganisms. The parameter β
represents the ratio of microorganisms with lower sensitivity to UV disinfection to the total
number of microorganisms.
2.4. The Effect of Secondary Treatment Conditions on Floc
Properties
Altering the secondary treatment process operational conditions have shown to
affect both the flocs physicochemical characteristics, and the particle association of
bacteria (Loge et al. 2002, Scott et al. 2005, Liao et al. 2006). The most well studied
operational parameters in wastewater treatment, in terms of their effects on sludge
characteristics include: oxygen concentration, sludge retention time, operating temperature,
influent phosphorus levels, carbon source, and coagulant type. The effects of each of these
parameters are discussed in the following subsections.
2.4.1. The Effect of Oxygen Levels on Sludge and Floc Properties
Low concentrations of dissolved oxygen (DO) lead to the formation of porous flocs
with poor settling properties (Palmgren et al. 1998, Wilén & Balmér 1999, Martins et al.
2003). The effect is often quantified by measuring sludge volume index (SVI) which is the
volume (mL) that 1g of sludge occupies in suspension after settling by gravity for 30
minutes. Palmgren et al. (1998) showed that dissolved oxygen limitation (DO < 0.1 ppm)
leads to a decrease in cell surface hydrophobicity and the worsening of the solid-liquid
separation, in activated sludge systems. Sponza (2002) reported an increase in municipal
21
flocs’ hydrophobicity under oxygen limited conditions. Martins et al. (2003) attributed the
poor settleability (SVI > 250 mL/g) of sludge formed in low DO (< 1.1 mg/L) to the
overgrowth of filamentous bacteria. They showed that the effect was worsened at high
chemical oxygen demand loading rate, and suggested a DO level of 2.5 mg/L to maintain
good sludge settleability (SVI < 100 mg/L).
Bulking has been identified as a common problem in systems where the operation
shifts in time and/or space between aerobic, anoxic and anaerobic conditions (Eikelboom
1998, Yun et al. 2000, Martins et al. 2004, Vaiopoulou et al. 2007). Wilén & Balmér (1999)
showed that altering aerobic and anaerobic conditions (1-4 h) did not affect the settling
properties to a large extent. However, they reported a significant increase in turbidity and a
gradual increase in the fraction of smaller flocs (2-20 µm) during the anaerobic period.
Moon et al. (2004) used sequencing batch reactors to study the effect of anaerobic
conditions on floc size and concluded that anaerobic flocs are smaller than aerobic flocs.
Martins et al. (2004) concluded that the presence of microaerophilic zones in the anoxic
stages of BNR systems leads to worsening sludge settling characteristics. Tampus et al.
studied the effect of feeding pattern on sludge settleability in denitrifying systems under
anoxic conditions, and concluded that despite the diffused and irregularly shaped structure,
the sludge settled well (SVI < 60 ml/g) (Tampus et al. 2004). Wilén et al. showed that
activated sludge deflocculates under anaerobic conditions and that the deflocculated
fragments are mainly composed of bacteria and EPS (Wilén et al. 2000). They observed
reduced deflocculation when nitrate acted as an electron acceptor (anoxic conditions). It
has also been reported that the presence of PAOs causes the formation of more compact
22
structures and improved sludge settleability (Eikelboom et al. 1998, Crocetti et al. 2000,
Tampus et al. 2004).
Variations in the DO level during wastewater treatment could affect floc structure
by changing the oxidation state of the naturally occurring iron, which impacts floc
aggregation. In the aerated activated sludge systems, where higher levels of DO are
present, iron is mostly in the form of Fe3+ (Rockne 2007). However, in anaerobic and
anoxic environments iron commonly takes the form of Fe2+. Fe2+ is known to be a more
effective flocculent compared to Fe3+ due to its ability in forming stronger ionic bonds
under neutral pH (Oikonomidis et al. 2010). Oikonomidis et al. (2010) reported that Fe3+
forms flocs that are diffuse and irregular, whereas Fe2+ causes the formation of compact
and less filamentous flocs with well-defined boundaries. It has also been shown that
compared to Fe2+, Fe3+ has a high affinity for protein (Murthy et al. 2000), and during
anaerobic conditions as ferric iron converts to ferrous there is a loss in selective binding
between iron and protein, that causes a release of protein (Novak et al. 2003). This can
affect the composition of microbial flocs in aerobic and anaerobic treatments, and their
structural stability.
2.4.2. The Effect of Sludge Retention Time (SRT) on Sludge and Floc
Properties
SRT is the one of the most important operational parameters in biological
wastewater treatment. It is defined as the average time that sludge is retained in the
bioreactor. Liao et al. performed a comprehensive study on the effect of sludge retention
time on sludge flocs composition, morphology, and stability (Liao et al. 2002, 2006). They
23
simulated the wastewater biological treatment process, by operating fully aerated
sequencing batch reactors at 28 ºC with synthetic feed and at SRTs in a range of 4-20 days.
Their results showed that flocs formed at higher SRTs have a more spherical and compact
structure, were less hydrated, and showed higher chemical stability. Investigating the
effect of SRT on the microbial flocs inter-particle interactions, Liao et al. (2002) concluded
that ionic interactions and hydrogen bonds are the two dominant forces that maintain
stability on sludge flocs at lower SRTs, and physical enmeshment and van der Waals
and/or hydrophobic interactions are more important in controlling the stability of sludge
flocs at higher SRTs. A conceptual model was proposed for the floc structure based on
inter-particle interactions, in which flocs generated at higher SRTs form dense cores that
have a lower growth rate than that of the outer layer, due to the limitations of diffusion of
oxygen and nutrients from the bulk solution (Figure 2. 3).
Wang et al. also operated sequencing batch reactors on synthetic feed at 20 ºC and
at SRTs of 5, 10, and 20 days (Wang et al. 2013). They found that the loosely bound EPS
content decreased at higher SRTs. Similar observations were made by Xie and Yang, as
they examined the effect of SRT on flocculation and bacterial communities in wastewater
sludge. They concluded that the loosely bound EPS correlates positively with the sludge
volume index and effluent suspended solids (Xie & Yang 2009). Scott et al. operated
sequencing batch reactors at 27 ºC with synthetic feed at sludge retention times ranging
from 4-20 days to study the effect of SRT and feed source on the physicochemical
properties of sludge flocs (Scott et al. 2005). They found that the composition of EPS
(protein, carbohydrate, and DNA content) did not significantly change by varying SRT.
However, they reported that flocs at higher SRTs are more hydrophobic and a larger
24
fraction of E.coli adheres to them. Therefore, the number of free swimming E.coli
decreases in the final effluent by increasing SRT.
Liss et al. (2002) studied the effect of sludge retention time on microbial floc
structure using various microscopy techniques (optical microscopy, environmental
scanning electron microscopy, transmission electron microscopy). They reported that flocs
at lower SRT are more irregular in shape and have higher carbohydrate content.
Transmission electron microscopy revealed a denser EPS layer formed in flocs at higher
SRT, which protect the interior cells from disruption. In general, flocs generated at higher
SRTs are more stable, hydrophobic and less negatively charged. They suggested that the
hydrophobicity of sludge flocs is affected by the protein to carbohydrate ratio in particular,
and that it could also be related to the differences in the amounts of lipids in the slime and
capsular layers associated with EPS.
Loge et al. took mixed liquor and secondary effluents from eight wastewater
treatment plants, and used in-situ hybridization of a 16S rRNA oligonucleotide probe
specific to the family Enterobacteriaceae (Loge et al. 2002). They concluded that the
fraction of particle associated coliform decreases exponentially with increasing the mean
cell residence time. Emerick et al. (1999) similarly reported that the percentage of floc
association of coliform bacteria was 20, 12, and 4% for sludge retention times of 2.2, 11,
and 16 days respectively.
2.4.3. The Effect of Operational Temperature on Sludge and Floc Properties
Wilén et al. (2010) performed a comprehensive study on the settling and
flocculation of a full-scale activated sludge treatment plant in different seasons. They
25
found that the settling and compaction properties of sludge generally improved during the
summer months (variation of temperature from a minimum of 7 ˚C in the winter to a
maximum of about 20 ˚C in the summer). They also concluded that flocs formed during
the winter are more susceptible to breaking under shear stress, as the bacteria grow in small
colonies in chain-like structures. Tian et al. (1994) found that the hydrolysis rate of
influent particles decreases when the operating temperature decreased from 20 to 8 ºC, and
on average caused a 16% increase in volatile solid production in the colder (8 ºC) reactor.
They suggested that under cold weather the viable organisms will comprise a smaller
percentage of the total activated sludge.
Liao et al. (2011) studied the settling properties of sludge generated in sequencing
batch reactors running at different temperatures (35 ˚C and 55 ˚C) on synthetic feed. The
results showed that the settleability of thermophilic (55 ˚C) sludge was poorer (higher SVI)
than that of the mesophilic (35 ˚C) sludge, due to a higher level of loosely bound EPS in
the thermophilic sludge (32 vs. 17.7 mg/g MLSS for thermophilic vs. mesophilic sludge).
Morgan-Sagastume & Allen (2003) studied the effect of temperature shifts (from 35 ˚C to
45 ˚C) on aerobic sludge and found poor sludge settling at higher temperatures under
steady-state conditions. Krishna and Van Loosdrecht (1999) examined the effect of
operating temperature on sludge settleability using aerobic SBRs treating an acetate
medium. They found a continuous decrease in sludge settleability (SVI increase) when
increasing temperature (15 ˚C, 20 ˚C, 25 ˚C, 30 ˚C, 35 ˚C). They argued that at lower
temperatures of 15 and 20 ˚C larger flocs and many protozoa are found in sludge, whereas
at 30 and 35 ˚C, the flocs become smaller with more zoogloeal overgrowth which led to
higher SVI.
26
2.4.4. The Effect of Influent Phosphorus Content on Sludge and Floc
Properties
The effect of phosphorus limitation on floc structure has been studied by Liu and
Liss (2007). Larger, rounder and more compact flocs, with improved settling
characteristics were formed under the phosphorus limited conditions (100:5:0.3). The
darker regions seen in phase contrast images were attributed to the greater amount of EPS
in the sludge flocs. Dumitrache (2008) also studied the effect of phosphorus limitation on
floc structure. He found that flocs generated at phosphorus limited conditions (100:5:0.33)
were larger and observed darker areas with the optical microscope due to a more densely
packed biomass. Similar observations were made by Brei (2011).
Ericsson & Eriksson (1988) suggested that extreme pre-precipitation of phosphorus
prior to biological treatment, favors the growth of filamentous bacteria, and related that to
the depletion of easily assimilated phosphate. A decrease in the EPS protein content and
surface charge of activated sludge flocs under phosphorus deficiency has been reported by
Sponza (2002). Liu et al. (2006) showed a significant increase in the total polysaccharide
content of sludge and in the acid polysaccharide content of the EPS under the limited
phosphorus conditions.
Other factors in the biological wastewater treatment process, which change floc
structure, composition, and settling characteristics, include adding coagulants in the
primary treatment, varying carbon source, varying organic loading rate, etc. Gehr and
Nicell (1996) reported that replacing FeCl3 with alum significantly improved the efficiency
of UV disinfection and reduced the UV dose requirements. Xie & Yang 2009 showed that
by changing the carbon source from glucose to acetate, at SRTs of 5, 10 and 20 days, the
27
effluent suspended solids decreased and the sludge settled better. The effect of carbon
source on floc composition has been investigated by Scott et al. (2005). They showed that
by changing the main carbon source from glucose to equal portions of skim milk and
glucose, the protein content of the EPS increased.
2.5. Summary
This literature review has indicated that the presence of microbial flocs in final
effluents leads to poor UV disinfection performance, and that larger particles worsen this
effect. Some of the impacts of varying secondary biological treatment conditions on floc
morphological properties and chemical composition have been identified. For example low
oxygen concentrations lead to deflocculation, increase in turbidity, and the formation of
smaller flocs, or that more spherical and compact flocs are generated at higher SRT.
Nevertheless, the mechanism of UV resistance and the source of tailing are not well
understood. Moreover, the effect of biological treatment conditions on UV disinfection
performance and its relation to flocs structure and composition remain unclear. Therefore,
this research aims at identifying the source of UV resistance and relating the flocs
structural properties and chemical composition to UV disinfection kinetics.
2.6. References
Asano, T. (2007). Water reuse: Issues, technologies, and applications. New York: McGraw-Hill. Ayarza, J. M., Guerrero, L. D., & Erijman, L. (2010). Nonrandom assembly of bacterial populations in activated sludge flocs. Microbial Ecology, 59(3), 436-444. Batch, L. F., Schulz, C. R., & Linden, K. G. (2004). Evaluating water quality effects on UV disinfection of MS2 coliphage. Journal / American Water Works Association, 96(7), 75-87.
28
Blatchley III, E. R., Hunt, B. A., Duggirala, R., Thompson, J. E., Zhao, J., Halaby, T., Alleman, J. E. (1997). Effects of disinfectants on wastewater effluent toxicity. Water Research, 31(7), 1581-1588. Blatchley III, E. R., Dumoutier, N., Halaby, T. N., Levi, Y., & Laîné, J. M. (2001). Bacterial responses to ultraviolet irradiation. Water Science and Technology,43(10), 179-186. Bellouti, M., Alves, M. M., Novals, J. M., & Mota, M. (1997). Flocs vs granules: Differentiation by fractal dimension. Water Research, 31(5), 1227-1231. Best, R. (2012). The Effect of Physicochemical Properties of Wastewater Flocs on UV Disinfection Following Hydrodynamic Particle Breakage. Thesis (M.A.Sc.)- University of Guelph. Bitton, G., Henis, Y., & Lahav, N. (1972). Effect of several clay minerals and humic acid on the survival of klebsiella aerogenes exposed to ultraviolet irradiation. Applied Microbiology, 23(5), 870-874. Bitton, G. (1999). Wastewater microbiology (2nd ed. ed.). New York: Wiley-Liss Blume, T., & Neis, U. (2004). Improved wastewater disinfection by ultrasonic pre-treatment. Ultrasonics
Sonochemistry, 11(5), 333-336. Brei, E. (2011). Bacterial Adhesin Proteins Associated with Microbial Flocs and EPS in Activated Sludge. Thesis (Ph.D)- University of Toronto Bura, R., Cheung, M., Liao, B., Finlayson, J., Lee, B. C., Droppo, I. G., Liss, S. N. (1998). Composition of extracellular polymeric substances in the activated sludge floc matrix. Water Science and Technology, 37(4-5), 325-333. Cairns, W.L. (1993). Proc. Planning, Design and Operations of Effluent Disinfection Systems. Proceedings of the Water Environment Federation, Specialty Conference: Planning, Design and Operation of Effluent Disinfection Systems, Whippany, NJ, 1993 pp. 555-566. Cerf, O. (1977). Tailing of survival curves of bacterial spores. Journal of Applied Bacteriology, 42(1), 1-19. Chen, F., Yang, X., Xu, F., Wu, Q., & Zhang, Y. (2009). Correlation of photocatalytic bactericidal effect and organic matter degradation of TiO2 part I: Observation of phenomena. Environmental Science and
Technology, 43(4), 1180-1184. Cho, M., Chung, H., Choi, W., & Yoon, J. (2004). Linear correlation between inactivation of E. coli and OH radical concentration in TiO2 photocatalytic disinfection. Water Research, 38(4), 1069-1077. Crocetti, G. R., Hugenholtz, P., Bond, P. L., Schuler, A., Keller, J., Jenkins, D., & Blackall, L. L. (2000). Identification of polyphosphate-accumulating organisms and design of 16S rRNA-directed probes for their detection and quantitation. Applied and Environmental Microbiology, 66(3), 1175-1182. Darby, J. L., Snider, K. E., & Tchobanoglous, G. (1993). Ultraviolet disinfection for wastewater reclamation and reuse subject to restrictive standards. Water Environment Research, 65(2), 169-180. Das, T. K. (2001). Ultraviolet disinfection application to a wastewater treatment plant. Clean Products and
Processes, 3(2), pp. 69-80), August. Dietrich, J. P., Başaǧaoǧlu, H., Loge, F. J., & Ginn, T. R. (2003). Preliminary assessment of transport processes influencing the penetration of chlorine into wastewater particles and the subsequent inactivation of particle-associated organisms. Water Research, 37(1), 139-149. Dumitrache, R. G. (2008). Characterization of a hybrid sequencing batch reactor system for treatment of wastewater using suspended microbial aggregates and biofilms. Thesis (M.A. Sc.)- Ryerson University
29
Eikelboom, D. H., Andreadakis, A., & Andreasen, K. (1998). Survey of filamentous populations in nutrient removal plants in four european countries. Water Science and Technology, 37(4-5), 281-289. Emerick, R. W., Loge, F. J., Ginn, T., & Darby, J. L. (2000). Modeling the inactivation of particle-associated coliform bacteria. Water Environment Research, 72(4), 432-438. Emerick, R. W., Loge, F. J., Thompson, D., & Darby, J. L. (1999). Factors influencing ultraviolet disinfection performance part II: Association of coliform bacteria with wastewater particles. Water Environment
Research, 71(6), 1178-1187. Ericsson, B., & Eriksson, L. (1988). Activated sludge characteristics in a phosphorus depleted environment. Water Research, 22(2), 151-162. Eriksson, L., Steen, I., & Tendaj, M. (1992). Evaluation of sludge properties at an activated sludge plant. Water Science and Technology, 25(6), 251-265. Farnood, R., Flocs and ultraviolet disinfection. Chapter 18 in Flocculation in Natural and Engineering
Systems. I.G. Droppo, G.G. Leppard, S.N. Liss, and T.G. Milligan, Eds., CRC Press, Boca Raton, FL, 2005, pp. 385-395. Flemming, H., & Wingender, J. (2001). Relevance of microbial extracellular polymeric substances (EPSs) - part I: Structural and ecological aspects. Water Science and Technology, 43(6), 1-8. Forster, C. F. (1985). Factors involved in the settlement of activated sludge-II; the binding of polyvalent metals. Water Research, 19(10), 1265-127. Frølund, B., Palmgren, R., Keiding, K., Nielsen, P. H. (1996). Extraction of extracellular polymers from activated sludge using a cation exchange resin. Water Research, 30(8), 1749-1758. Gehr, R., & Henry, J. G. (1983). Removal of extracellular material. techniques and pitfalls. Water Research,
17(12), 1743-1748. Gehr, R., & Nicell, J. (1996). Pilot studies and assessment of downstream effects of UV and ozone disinfection of a physicochemical wastewater. Water Quality Research Journal of Canada, 31(2), 263-281. Gehr, R., Wagner, M., Veerasubramanian, P., & Payment, P. (2003). Disinfection efficiency of peracetic acid, UV and ozone after enhanced primary treatment of municipal wastewater. Water Research, 37(19), 4573-4586. Gibson, J. H., Yong, D. H. N., Farnood, R. R., & Seto, P. (2008). A literature review of ultrasound technology and its application in wastewater disinfection. Water Quality Research Journal of Canada, 43(1), 23-35. Gu, A. Z., Saunders, A., Neethling, J. B., Stensel, H. D., & Blackall, L. L. (2008). Functionally relevant microorganisms to enhanced biological phosphorus removal performance at full-scale wastewater treatment plants in the united states. Water Environment Research, 80(8), 688-698. Guan, J., Waite, T. D., Amal, R., Bustamante, H., & Wukasch, R. (1998). Rapid determination of fractal structure of bacterial assemblages in wastewater treatment: Implications to process optimisation. Water
Science and Technology, 38(2 pt 2), 9-15. Huang, Z., Maness, P., Blake, D. M., Wolfrum, E. J., Smolinski, S. L., & Jacoby, W. A. (2000). Bactericidal mode of titanium dioxide photocatalysis. Journal of Photochemistry and Photobiology A: Chemistry, 130(2-3), 163-170.
30
Ireland, J. C., Klostermann, P., Rice, E. W., & Clark, R. M. (1993). Inactivation of escherichia coli by titanium dioxide photocatalytic oxidation. Applied and Environmental Microbiology, 59(5), 1668-1670. Jagger, J., (1967) Introduction to Research in Ultraviolet Photobiology. Prentice-Hall, Inc. Englewood Cliffs, NJ Jenkins, D. (1993). In Richard M. G., Daigger G. T. (Eds.), Manual on the causes and control of activated
sludge bulking and foaming (2nd Ed.). Boca Raton: Lewis. Jin, B., Wilén, B., & Lant, P. (2003). A comprehensive insight into floc characteristics and their impact on compressibility and settleability of activated sludge. Chemical Engineering Journal, 95(1), 221-234. Jorand, F., Zartarian, F., Thomas, F., Block, J. C., Bottero, J. Y., Villemin, G., Manem, J. (1995). Chemical and structural (2D) linkage between bacteria within activated sludge flocs. Water Research, 29(7), 1639-1647. Jorand, F., Boué-Bigne, F., Block, J. C., & Urbain, V. (1998). Hydrophobic/hydrophilic properties of activated sludge exopolymeric substances. Water Science and Technology, 37(4-5), 307-315. Kitis, M. (2004). Disinfection of wastewater with peracetic acid: A review. Environment International, 30(1), 47-55. Kollu, K., & Örmeci, B. (2012). Effect of particles and bioflocculation on ultraviolet disinfection of escherichia coli. Water Research, 46(3), 750-760. Kozak, J. A., Lordi, D. T., Abedin, Z., O’Connor, C., Granato, T., & Kollias, L. (2010). The effect of ferric chloride addition for phosphorus removal on ultraviolet radiation disinfection of wastewater. Environmental
Practice, 12(4), 275-284. Krishna, C., & Van Loosdrecht, M. C. M. (1999). Effect of temperature on storage polymers and settleability of activated sludge. Water Research, 33(10), 2374-2382. Li, D., & Ganczarczyk, J. (1989). Fractal geometry of particle aggregates generated in water and wastewater treatment processes. Environmental Science and Technology, 23(11), 1385-1389 Li, D., & Ganczarczyk, J. J. (1987). Stroboscopic determination of settling velocity, size and porosity of activated sludge flocs. Water Research, 21(3), 257-262. Liao, B. (2000). Physicochemical studies of microbial flocs. Thesis (Ph.D)-University of Toronto Liao, B. Q., Allen, D. G., Droppo, I. G., Leppard, G. G., & Liss, S. N. (2001). Surface properties of sludge and their role in bioflocculation and settleability. Water Research, 35(2), 339-350 Liao, B. Q., Allen, D. G., Leppard, G. G., Droppo, I. G., & Liss, S. N. (2002). Interparticle interactions affecting the stability of sludge flocs. Journal of Colloid and Interface Science, 249(2), 372-380. Liao, B. Q., Droppo, I. G., Leppard, G. G., & Liss, S. N. (2006). Effect of solids retention time on structure and characteristics of sludge flocs in sequencing batch reactors. Water Research, 40(13), 2583-2591. Liao, B. Q., Lin, H. J., Langevin, S. P., Gao, W. J., & Leppard, G. G. (2011). Effects of temperature and dissolved oxygen on sludge properties and their role in bioflocculation and settling. Water Research, 45(2), 509-520. Liltved, H., & Cripps, S. J. (1999). Removal of particle-associated bacteria by prefiltration and ultraviolet irradiation. Aquaculture Research, 30(6), 445-450.
31
Linden, K. G., Shin, G., Faubert, G., Cairns, W., & Sobsey, M. D. (2002). UV disinfection of giardia lamblia cysts in water. Environmental Science and Technology, 36(11), 2519-2522. Liss, S. N., Methods for Analyzing Floc Properties. Chapter 1 in Flocculation in Natural and Engineering
Systems. I.G. Droppo, G.G. Leppard, S.N. Liss, and T.G. Milligan, Eds., CRC Press, Boca Raton, FL, 2005, pp. 385-395. Liss, S. N., Liao, B. Q., Droppo, I. G., Allen, D. G., & Leppard, G. G. (2002). Effect of solids retention time on floc structure. Water Science and Technology, 50(3), 431-438. Liss, S. N., Droppo, I. G., Flannigan, D. T., & Leppard, G. G. (1996). Floc architecture in wastewater and natural riverine systems. Environmental Science and Technology, 30(2), 680-686. Liu, J. R., Liu, C. T., Edwards, E. A., & Liss, S. N. (2006). Effect of phosphorus limitation on microbial floc structure and gene expression in activated sludge. Water Science and Technology, 54(1), 247-255. Liu, J. R., & Liss, S. N. (2007). The impact of reduced phosphorus levels on microbial floc properties during biological treatment of pulp and paper wastewaters. Water Science and Technology, 55(6), 73-79. Liu, R., Yang, Y., Li, G., He, W., & Han, H. (2007). Adsorptive behaviors of humic acid onto freshly prepared hydrous manganese dioxides. Frontiers of Environmental Science and Engineering in China, 1(2), 240-245. Loge, F. J., Emerick, R. W., Thompson, D. E., Nelson, D. C., & Darby, J. L. (1999). Development of a fluorescent 16S rRNA oligonucleotide probe specific to the family enterobacteriaceae. Water Environment
Research, 71(1), 75-83. Loge, F. J., Emerick, R. W., Ginn T. R., Darby, J. L. (2002). Association of coliform bacteria with wastewater particles: Impact of operational parameters of the activated sludge process. Water Research, 36(1), 41-48. Loge, F. J., Emerick, R. W., Heath, M., Jacangelo, J., Tchobanoglous, G., & Darby, J. L. (1996). Ultraviolet disinfection of secondary wastewater effluents: Prediction of performance and design. Water Environment
Research, 68(5), 900-916. Madge, B. A., & Jensen, J. N. (2006). Ultraviolet disinfection of fecal coliform in municipal wastewater: Effects of particle size. Water Environment Research, 78(3), 294-304. Maness, P., Smolinski, S., Blake, D. M., Huang, Z., Wolfrum, E. J., & Jacoby, W. A. (1999). Bactericidal activity of photocatalytic TiO2 reaction: Toward an understanding of its killing mechanism. Applied and
Environmental Microbiology, 65(9), 4094-4098. Martins, A. M. P., Heijnen, J. J., & Van Loosdrecht, M. C. M. (2003). Effect of feeding pattern and storage on the sludge settleability under aerobic conditions. Water Research, 37(11), 2555-2570. Martins, A. M. P., Heijnen, J. J., Van Loosdrecht, M. C. M. (2004). Bulking sludge in biological nutrient removal systems. Biotechnology and Bioengineering, 86(2), 125-135. Matasci, R. Weston, P. Lau, J. Cruver, S. Marek, D. Tomowich (1999), Wastewater Technology Fact Sheet: Ultraviolet Disinfection, United States Environmental Protection Agency, Office of Water, Washington, DC, EPA 832-F-99-064, R. Matsunaga, T., Tomoda, R., Nakajima, T., & Wake, H. (1985). Photoelectrochemical sterilization of microbial cells by semiconductor powders. FEMS Microbiology Letters, 29(1-2), 211-214.
32
Moon, B., Seo, G., Lee, T., Kim, S., Yoon, C. (2004). Effects of fill time on the fractal dimension and size of floc for SBR treating low C/N ratio wastewater. Journal of Environmental Science and Health - Part A
Toxic/Hazardous Substances and Environmental Engineering, 39(7), 1757-1765. Morgan-Sagastume, F., & Allen, D. G. (2003). Effects of temperature transient conditions on aerobic biological treatment of wastewater. Water Research, 37(15), 3590-3601. Murthy, S. N., Novak, J. T., & Holbrook, R. D. (2000). Optimizing dewatering of biosolids from autothermal thermophilic aerobic digesters (ATAD) using inorganic conditioners. Water Environment Research, 72(6-7), 714-721. Nelson, K. L. (2000). Ultraviolet light disinfection of wastewater stabilization pond effluents. Water Science
and Technology, 42(10-11), 165-170. Nessim, Y., & Gehr, R. (2006). Fouling mechanisms in a laboratory-scale UV disinfection system. Water
Environment Research, 78(12), 2311-2323. Novak, J. T., Sadler, M. E., & Murthy, S. N. (2003). Mechanisms of floc destruction during anaerobic and aerobic digestion and the effect on conditioning and dewatering of biosolids. Water Research, 37(13), 3136-3144. Oguma, K., Katayama, H., & Ohgaki, S. (2004). Photoreactivation of legionella pneumophila after inactivation by low- or medium-pressure ultraviolet lamp. Water Research, 38(11), 2757-2763. Oguma, K., Katayama, H., & Ohgaki, S. (2002). Photoreactivation of escherichia coli after low- or medium-pressure UV disinfection determined by an endonuclease sensitive site assay. Applied and Environmental
Microbiology, 68(12), 6029-6035. Oikonomidis, I., Burrows, L. J., & Carliell-Marquet, C. M. (2010). Mode of action of ferric and ferrous iron salts in activated sludge. Journal of Chemical Technology and Biotechnology, 85(8), 1067-1076. Örmeci, B., & Linden, K. G. (2005). Comparison of physical and chemical methods for extraction of coliform from wastewater particles and flocs. Environmental Engineering Science, 22(4), 459-471. Palmgren, R., Jorand, F., Nielsen, P. H., & Block, J. C. (1998). Influence of oxygen limitation on the cell surface properties of bacteria from activated sludge. Water Science and Technology, 37(4-5), 349-352. Paraskeva, P., & Graham, N. J. D. (2005). Treatment of a secondary municipal effluent by ozone, UV and microfiltration: Microbial reduction and effect on effluent quality. Desalination, 186(1-3), 47-56. Parker, J. A., & Darby, J. L. (1995). Particle-associated coliform in secondary effluents: Shielding from ultraviolet light disinfection. Water Environment Research, 67(7), 1065-1075. Pavoni, J. L. , Tenny M. W., & Echelberger Jr, W. F. (1972). Bacterial exocellular polymers and biological flocculation. Journal of the Water Pollution Control Federation, 44(3 pt 1), 414-431. Peeters, B., Dewil, R., Lechat, D., & Smets, I. Y. (2011). Quantification of the exchangeable calcium in activated sludge flocs and its implication to sludge settleability. Separation and Purification Technology,
83(1), 1-8. Perez, Y. G., Leite, S. G. F., & Coelho, M. A. Z. (2006). Activated sludge morphology characterization through an image analysis procedure. Brazilian Journal of Chemical Engineering, 23(3), 319-330. Pfeifer, G. P., You, Y., & Besaratinia, A. (2005). Mutations induced by ultraviolet light. Mutation Research - Fundamental and Molecular Mechanisms of Mutagenesis, 571(1-2 SPEC. ISS.), 19-31.
33
Piluso, L. G., & Moffatt-Smith, C. (2006). Disinfection using ultraviolet radiation as an antimicrobial agent: A review and synthesis of mechanisms and concerns. PDA Journal of Pharmaceutical Science and
Technology, 60(1), 1-16. Qualls, R. G., Ossoff, S. F., & Chang, J. C. H. (1985). Factors controlling sensitivity in ultraviolet disinfection of secondary effluents. Journal of the Water Pollution Control Federation, 57(10), 1006-1011. Rockne, K. J. (2007). Kinetic hindrance of fe(II) oxidation at alkaline pH and in the presence of nitrate and oxygen in a facultative wastewater stabilization pond. Journal of Environmental Science and Health - Part A Toxic/Hazardous Substances and Environmental Engineering, 42(3), 265-275. Schmid, M., Thill, A., Purkhold, U., Walcher, M., Bottero, J. Y., Ginestet, P., Wagner, M. (2003). Characterization of activated sludge flocs by confocal laser scanning microscopy and image analysis. Water
Research, 37(9), 2043-2052. Scott, H. E., Liss, S. N., Farnood, R. R., Allen, D. G. (2005). Ultraviolet disinfection of sequencing batch reactor effluent: A study of physicochemical properties of microbial floc and disinfection performance. Journal of Environmental Engineering and Science, 4(SUPPL. 1), S65-S74. Sears, K., Alleman, J. E., Barnard, J. L., & Oleszkiewicz, J. A. (2006). Density and activity characterization of activated sludge flocs. Journal of Environmental Engineering, 132(10), 1235-1242. Setlow, P. (1995). Mechanisms for the prevention of damage to DNA in spores of bacillus species. Annual
Review of Microbiology, 49, 29-54. Severin, B. F. (1980). Disinfection of municipal wastewater effluents with ultraviolet light. Journal of the
Water Pollution Control Federation, 52(7), 2007-2018. Shang, C., Cheung, L. M., Ho, C., & Zeng, M. (2009). Repression of photoreactivation and dark repair of coliform bacteria by TiO2-modified UV-C disinfection. Applied Catalysis B: Environmental, 89(3-4), 536-542. Sheng, G., Yu, H., & Li, X. (2006). Stability of sludge flocs under shear conditions: Roles of extracellular polymeric substances (EPS). Biotechnology and Bioengineering, 93(6), 1095-1102. Sponza, D. T. (2002). Extracellular polymer substances and physicochemical properties of flocs in steady- and unsteady-state activated sludge systems. Process Biochemistry, 37(9), 983-998. Tampus, M. V., Martins, A. M. P., & van Loosdrecht, M. C. M. (2004). The effect of anoxic selectors on sludge bulking. Water Science and Technology, 50(6), 261-268. Tan, T. C. (2007). Understanding the effect of particle-size on UV disinfection: Kinetics, mechanism and modeling. Thesis (M.A. Sc.)-University of Toronto. Tchobanoglous, G., Burton, F. L., Stensel, H. D., & Metcalf & Eddy. (2003). Wastewater engineering:
Treatment and reuse. Boston: McGraw-Hill. Templeton, M. R., Andrews, R. C., & Hofmann, R. (2005). Inactivation of particle-associated viral surrogates by ultraviolet light. Water Research, 39(15), 3487-3500. Templeton, M. R., Andrews, R. C., & Hofmann, R. (2008). Particle-associated viruses in water: Impacts on disinfection processes. Critical Reviews in Environmental Science and Technology, 38(3), 137-164. Tian, S., Lishman, L., & Murphy, K. L. (1994). Investigations into excess activated sludge accumulation at low temperatures. Water Research, 28(3), 501-509.
34
Trivedi, H. K. (2009). Simultaneous nitrification and denitrification. In Advanced biological Treatment
Processes; Wang, L. K., Shammas, N. K., Hung, Y. T., Eds.; Humana Press: Totowa, NJ, pp. 186-189. Urbain, V., Block, J. C., & Manem, J. (1992). Bioflocculation in activated sludge, an analytic approach. Water Science and Technology, 25(4-5), 441-443. Urbain, V., Block, J. C., & Manem, J. (1993). Bioflocculation in activated sludge: An analytic approach. Water Research, 27(5), 829-838. Vaiopoulou, E., Melidis, P., & Aivasidis, A. (2007). Growth of filamentous bacteria in an enhanced biological phosphorus removal system. Desalination, 213(1-3), 288-296. Wang, H., Deng, H., Ma, L., & Ge, L. (2013). Influence of operating conditions on extracellular polymeric substances and surface properties of sludge flocs. Carbohydrate Polymers, 92(1), 510-515. Wang, L., Hu, H., & Wang, C. (2007). Effect of ammonia nitrogen and dissolved organic matter fractions on the genotoxicity of wastewater effluent during chlorine disinfection. Environmental Science and Technology,
41(1), 160-165. Water Environment Federation., & Environmental and Water Resources Institute (U.S.). (2006). Biological
nutrient removal (BNR) operation in wastewater treatment plants. New York: McGraw-Hill. Watts, R. J., Kong, S., Orr, M. P., Miller, G. C., & Henry, B. E. (1995). Photocatalytic inactivation of coliform bacteria and viruses in secondary wastewater effluent. Water Research, 29(1), 95-100. Wilén, B., Balmér, P. (1999). The effect of dissolved oxygen concentration on the structure, size and size distribution of activated sludge flocs. Water Research, 33(2), 391-400. Wilén, B. M., Lumley, D., Mattsson, A., & Mino, T. (2010). Dynamics in flocculation and settling properties studied at a full-scale activated sludge plant. Water Environment Research, 82(2), 155-168. Wilén, B. M., Keiding, K., & Nielsen, P. H. (2000). Anaerobic deflocculation and aerobic reflocculation of activated sludge. Water Research, 34(16), 3933-3942. Wilén, B. M., Lumley, D., Mattsson, A., & Mino, T. (2008). Relationship between floc composition and flocculation and settling properties studied at a full scale activated sludge plant. Water Research, 42(16), 4404-4418. Wojtenko, I. (2001). Performance of ozone as a disinfectant for combined sewer overflow. Critical Reviews
in Environmental Science and Technology, 31(4), 295-309. Wu, Y., Clevenger, T., & Deng, B. (2005). Impacts of goethite particles on UV disinfection of drinking water. Applied and Environmental Microbiology, 71(7), 4140-4143. Xie, B., Yang, S. (2009). Analyses of bioflocculation and bacterial communities in sequencing batch reactors. Environmental Engineering Science, 26(3), 481-487. Yong, H. N., Farnood, R. R., Cairns, W., & Mao, T. (2009). Effect of sonication on UV disinfectability of primary effluents. Water Environment Research, 81(7), 695-701. Yuan, Y. (2007). Investigating the relationship between the physical structure and shear strength of bioflocs. Ph.D. Dissertation #177, University of Toronto. Yuan, Y., & Farnood, R. R. (2010). Strength and breakage of activated sludge flocs. Powder Technology, 199(2), 111-119.
35
Yun, Z., Jo, W., Yi, Y., Choi, S., Choi, E., & Min, K. (2000). Effects of sludge settling characteristics in the BNR system. Water Science and Technology, 42(3-4), 283-288. Zimmer, K. G. (1961). Studies on quantitative radiation biology. Edinburgh: Oliver and Boyd.
36
3. The Effect of Wastewater Floc Structure on UV
Disinfection Kinetics1
3.1. Abstract
Ultraviolet disinfection is a physical method of disinfecting secondary treated
wastewaters. Flocs formed during secondary treatment harbor and protect microbes from
exposure to ultraviolet (UV) light, and significantly decrease the efficiency of disinfection
at high UV doses causing the tailing phenomena. However, the exact mechanism of tailing
and the role of floc properties and treatment conditions are not widely understood. The
objective of this work is to test whether a possible mechanism for tailing is that sludge
flocs are composed of an easily disinfectable loose outer shell, and a physically stronger
compact core that is more resistant to UV. In this chapter, hydrodynamic shear stress was
applied to the flocs to peel off the looser outer shell to isolate the cores. Floc and core
samples were fractionated into narrow size fractions by sieving and their UV disinfection
kinetics were determined and compared. The results showed that for flocs, the tailing level
elevated as the floc size increased, showing greater resistance to disinfection. However, for
the cores larger than 45µm, it was found that the UV inactivation curves overlapped, and
exhibited nearly identical inactivation kinetics. When comparing flocs and cores of similar
size fraction, the results consistently showed that cores were harder to disinfect with UV
1 .This Chapter is based on the following article: Azimi, Y., Allen, D. G., Farnood, R. R. (2012). Kinetics of UV Inactivation of Wastewater Bioflocs. Water
Research, 46 (12), 3827-3836.
37
light, and showed a higher tailing level. This study suggests that physical structure of flocs
plays a significant role in the UV inactivation kinetics.
3.2. Nomenclature
N (CFU): Final concentration (i.e. after UV irradiation) of viable flocs
N0 (CFU): Initial concentration of viable flocs
D (mJ/cm²): UV dose delivered to the wastewater. The product of UV intensity (mW/cm²)
and time (s)
δs (µm): The thickness of the loose shell
Deff (mJ/cm²): The effective dose delivered to a certain point inside a floc, which is a
function of δ, A, and floc porosity
δ (µm): The distance of the point of interest in the floc from the flocs surface that varies
between 0 and R
A (1/cm): The UV absorption coefficient of the floc material
Ac (1/cm): The UV absorption coefficient of the core material
AS (1/cm): The UV Absorption coefficient of the loose shell material
R (µm): The spherical radius of a floc
k (cm²/mJ): The inactivation rate constant of embedded coliform bacteria
k1 (cm²/mJ): Inactivation rate constant for the UV-susceptible or type 1 flocs
k2 (cm²/mJ): Inactivation rate constant for the UV-resistant of type 2 flocs
keff (cm²/mJ): The effective inactivation rate constant
f(δ,R): The probability density function of δ
38
β: The fraction of UV resistant flocs, or the ratio of the number of viable UV-resistant flocs
to the total number of viable flocs
δ1 (µm): The distance from the surface of the floc in which the UV inactivation occurs
with a rate constant of k1
δc (µm): The thickness of the viable layer
κ: Ratio of the density of the core to that of the loose shell
DRC: Dose Response Curve
3.3. Introduction
With increased restrictions on the quality of the treated effluents, there has been a
growing interest in applying ultraviolet (UV) disinfection for disinfecting secondary
wastewater effluents. UV disinfection is a physical method, which eliminates much of the
hazards associated with chemical disinfectants (Matasci et al. 1999). UV light, specifically
around the wavelength of 254 nm, can penetrate through the cell wall and alter the
microorganisms DNA and halt their ability to reproduce (Piluso & Moffatt-Smith 2006,
Pfeifer et al. 2005).
Suspended solids in secondary treated wastewater are known to decrease the
efficiency of UV disinfection by absorbing or scattering UV light, and hence protecting the
embedded microorganisms from exposure to UV light (Jorand et al. 1998, Das 2001, Loge
et al. 1999). The suspended solids in the secondary treated effluents are mostly generated
during the biological treatment process, which results in the formation and growth of
microbial flocs. These flocs consist of microorganisms, extracellular polymeric substances
(EPS), organic and inorganic colloidal particles (Sanin & Vesilind 1994, Urbain et al. 1993)
39
all of which are held together and stabilized by various mechanisms, including, EPS
enmeshment, salt bridging, polymer bridging, and hydrophobic interactions (Forster 1985,
Pavoni et al. 1972, Liss et al. 1996). Activated sludge flocs consist of 90-98% water, and
on a dry basis they contain about 60-90% cellular organic material (Jenkins 1993). They
are irregular in shape, vary in size from several microns up to about 1000 µm (Yuan 2007),
and are highly porous (45-90%) (Li & Ganczaraczyk 1987).
Figure 3. 1 is the schematic of a typical UV dose-response curve (DRC) for a
treated effluent. This curve consists of an initial steep slope followed by a tailing region
where a decrease in the rate of inactivation is observed for the applied UV dose. The
tailing phenomenon increases the usage of energy to reach acceptable degrees of
disinfection.
More Easily Disinfected
Lo
g (
N/N
0)
0
Tailing
UV Dose (D)
More Easily Disinfected
Lo
g (
N/N
0)
0
Tailing
UV Dose (D)
Figure 3. 1. Illustration of a typical UV dose-response curve of wastewater
In 1977, Cerf suggested that tailing of UV DRC occurs because of the existence of
a mixed population with different degrees of resistance to UV inactivation (Cerf 1997).
Later, many researchers suggested that the floc-associated microbes were responsible for
tailing (Tan 2007, Loge et al. 1996, Qualls et al. 1985). Recently, Tan studied the effect of
floc size on the degree of disinfectability and on the tailing level of activated sludge flocs
40
ranging in size from 20-120 µm and reported that the required UV dose for disinfection of
wastewater flocs increased with the floc size (Tan 2007). Tan also found that similar to the
UV DRC of whole effluents, the DRC of a suspension of flocs with a narrow size
distribution exhibited two regions, namely, an initial steep slope followed by a tailing
region, and suggested that these two regions represented the inactivation of UV-susceptible
flocs and UV-resistant flocs, respectively.
The kinetics of UV disinfection of wastewater flocs is expected to be a function of
the floc structure. It has been proposed that wastewater flocs have a two-layer structure
consisting of a compact core surrounded by a loose shell (Eriksson et al. 1992, Liao et al.
2002, Yuan 2007, Sheng et al. 2006). Liao et al. studied the stability of flocs generated at
various sludge retention times (SRT), and suggested that at higher SRTs the floc core is
more compact and more hydrophobic (Liao et al. 2002). Recently, by exposing wastewater
flocs to hydrodynamic shear in a Couette flow cell, Yuan and Farnood reported that the
shear strength of cores could be several times greater than that of loose shell (Yuan and
Farnood 2010). They suggested that the higher mechanical strength of cores may be due to
the densely packed structure of the core where microorganisms are held together by EPS
enmeshment and other intra-floc forces. Sheng et al. also studied the stability of flocs
under shear and proposed a multilayer structure for wastewater flocs with two distinct
regions. The outer part was suggested to be a loosely entangled readily-extractable EPS
fraction (Sheng et al. 2006) that is completely dispersed under shearing, while the inner
part is more stable and could not be dispersed except under extreme conditions.
Despite recent advances in the overall understanding of the floc structure, the
relationship between structure and UV disinfection kinetics of wastewater flocs is not well
41
understood. The objective of this study is to investigate the causes of the tailing
phenomenon and its relationship to the structure of wastewater flocs. In particular, it is
hypothesized that the compact cores are more UV-resistant and hence cause the tailing of
UV DRC of wastewater flocs.
3.4. A Structural Model for Floc UV Disinfection
A simplified model is presented here to justify the observed kinetics of UV
disinfection of flocs. In this model, a fraction of the flocs contain a compact ‘core’
surrounded by a loose outer shell; while others have a loose structure and do not contain
any core. Moreover, it is assumed that:
• The UV disinfection of a floc is governed by the inactivation of the most
shielded microorganism embedded within that floc
• The loose and porous outer shell is represented by a continuum spherical shell
layer with an average density and an average UV absorbance
• The compact core can be represented by a continuum spherical material with a
higher density and UV absorbance compared to those of the loose shell
• The probability of finding a microorganism in a given volume is proportional to
the mass of material within that volume
• The UV light intensity in various points of the core and shell may be expressed
by exponential decay (i.e. Beer-Lambert’s law)
• The floc surface is uniformly exposed to UV light
The above model is consistent with the floc structure proposed in the literature
based on floc stability studies; i.e. a double layer structure where a more stable inner core is
42
surrounded by a looser, more easily dispersible outer layer (Yuan et al. 2010, Liao et al.
2002, Sheng et al. 2006).
In the most general case, flocs are represented by a spherical particle of radius R
consisting of a compact core with the radius Rc and a loose shell with thickness δS (Figure
3. 2). Given that the UV disinfection of these model flocs is governed by the location of
the most shielded indicator organism; i.e. an embedded indicator organism that receives the
least amount of UV dose, model flocs can be classified into two types. In Type 1 (or UV-
susceptible) flocs, the most shielded indicator organism is located within the loose material
and near the floc surface. Hence, a Type 1 floc either has or does not have core (Figure 3.
2-a), and the indicator organism is embedded within its loose shell (Figure 3. 2-b). In
contrast, Type 2 (or UV-resistant) flocs are those in which the most shielded indicator
organism is embedded within the compact core and is more protected from UV light.
Compared to Type 1 flocs, Type 2 flocs are harder to disinfect and as will be seen later, are
responsible for the tailing of UV DRC.
Due to the UV absorption by the floc material, the UV dose received by embedded
organisms is less than the estimated average dose delivered to a free floating microbe
within the sample. Assuming that the most shielded indicator organism i.e. fecal coliform
in this study, is located at a depth δ from the floc surface, the effective dose, Deff (δ),
received by this organism can be estimated from the Beer-Lambert law:
Deff (δ) = D e –A(δ) × δ [3-1]
where D is the average UV dose delivered to the sample in mJ/cm2, and A(δ) is the UV
absorption coefficient of the floc material.
43
Type 1 Type 2
Loose shell Compact core
(a) (b) (c) (d)
Core
δSR R
Figure 3. 2. Schematic diagrams showing model spherical flocs: (a, b) Type 1 or UV-susceptible floc, (c)
Type 2 or UV-resistant floc, and (d) isolated compact core. The small white circle indicates the location
of the most shielded viable indicator organism.
It is worth noting that in the standard membrane filtration test, used to assess
wastewater UV disinfection performance (APHA 2001, # 9222 A), every floc that
“survives” UV irradiation leads to a single CFU count, regardless of the number of
surviving coliform bacteria embedded within that floc. This is an inherent limitation of the
membrane filtration technique. However, this technique is currently the industry/regulatory
standard for design and evaluation of disinfection technologies. If the inactivation rate
constant of embedded coliform bacteria, k, is equal to that of free-floating ones, the
inactivation kinetics of a spherical floc with radius R can be expressed by:
N(R, δ) = No(R, δ) e – k Deff
(δ) [3-2]
here, No(R, δ) and N(R, δ) are the initial concentration and the final concentration (i.e. after
UV irradiation) of viable flocs, a ‘viable floc’ being a floc with at least one culturable
embedded coliform bacteria, expressed as the number of colony forming units (CFU) per
44
100 ml. Since in practice the average UV dose is of significance, Equations [3-1] and [3-2]
can be rearranged to obtain:
N(δ, R) = No(δ, R) e – keff
(δ) D [3-3]
where keff (δ) is the effective inactivation rate constant defined by:
keff (δ) = k e −A(δ) × δ [3-4]
Given that δ varies from 0 to R; the UV DRC of ideal spherical flocs with radius R
would be:
N(R) /No(R) = ∫oδ
f (R, δ ) e – keff
(δ) D dδ [3-5]
where, N(R) and No(R) are the initial and final counts of viable flocs with radius R (in
CFU/100 ml), and f (δ,R) is the probability density function of finding a culturable
microorganism at the depth of δ.
Considering the structure of model flocs in Figure 3. 2, and replacing the
inactivation rate constants of Type 1 and Type 2 flocs with their respective average values,
i.e. k1 and k2, Equation [3-5] can be expanded as:
N(R) /No (R) = (∫oδS f (R, δ) dδ) × e – k
1 D + (∫δS
δ
f (R, δ) dδ) × e – k2
D [3-6]
Or:
N(R) /No (R) = (1 – β) e – k1 D + β e – k
2 D [3-7]
where β characterizes the fraction of UV-resistant flocs defined as ratio of the number of
viable UV-resistant flocs to the total number of viable flocs of radius R:
β = ∫δsδ
f (R, δ) dδ [3-8]
45
The first term on the right-hand-side of Equation [3-7] reflects the inactivation of
UV-susceptible or Type 1 flocs while the second term represents UV-resistant or Type 2
flocs.
Equation [3-7], known as the double exponential model, is used in this study to
analyze the inactivation kinetics of wastewater flocs. It is worthwhile to note that the
double exponential model has been used previously to study the UV disinfection of whole
effluents (Farnood 2005). However, in that case k1 is associated with the inactivation of
free-floating organisms and β is the fraction of particle-associated CFU.
3.5. Materials and Methods
3.5.1. Wastewater Samples
Mixed liquor samples were collected from Ashbridges Bay municipal wastewater
treatment facility, located at the eastern end of the city of Toronto. The plant has a capacity
of 818,000 m3/day, and has a suspended solid activated sludge biological system for
secondary treatment. The samples were taken on five different days during a period of
three months (July-October 2009). All experiments were conducted within 48 hours of
sample collection.
To test the effect of floc size on UV disinfection, the mixed liquor sample was
fractionated into narrow size fractions using appropriate sieve trays (U.S.A. Standard
Testing Sieve). The nominal sieve opening sizes were 32, 45, 53, 63, 75, and 90 µm;
therefore, the size fractions obtained were 32-45, 45-63, 53-63, 63-75, and 75-90 µm.
Collected floc fractions were gently washed with milliQ water for 15 minutes and
46
backwashed into a flask. The fractionated samples prepared in this way are referred to as
the “un-sheared floc” or just “floc” samples.
3.5.2. Shearing
In order to isolate the compact cores of the flocs, mixed liquor samples were subject
to hydrodynamic shear in a custom-made Couette flow cell developed by Yuan (Yuan &
Farnood 2010). During this process, the weak loose outer layer of the floc is removed,
leaving behind a mechanically strong compact core that is resistant to breaking under the
applied shear stress.
A schematic drawing of the Couette flow cell device is shown in Figure 3. 3. Ten
milliliters of wastewater floc sample was placed in the gap between the spindle and the cup
and speed of the spindle was set at 3500 rpm for 15 minutes. These shearing conditions
had no significant effect on the culturability of fecal coliforms (data not shown here). The
sheared samples were sieved to the following size fractions: 32-45, 45-63, 53-63, 63-75,
and 75-90 µm. The sieved fractions were gently washed using milliQ water to remove
small fragments and to obtain a narrow size distribution. Collected fractions of compact
cores were re-suspended in milliQ water and stored in separate flasks for further testing. In
this manuscript, these samples are referred to as “sheared” or “core” samples.
To determine the volume fraction and number concentration of compact cores, and
their relation to the floc size, activated sludge samples were fractionated according to the
method described earlier. The collected flocs were analyzed using a Multisizer 3 particle
size analyzer (Beckman Coulter Canada, Mississauga, Ontario, Canada), and diluted using
deionized water to obtain 0.06% suspensions (volume basis). Thirty milliliters of each
suspension was subsequently sheared using the Couette flow device at 3500 rpm for 15
47
minutes. The sheared suspension was filtered through a 45µm mesh sieve to ensure that
small fragments of loose shell (generated during the shearing process) are adequately
removed from the isolated core samples. The collected cores were subsequently analyzed
using a Multisizer 3 particle size analyzer to obtain the number and volume of cores for
each floc size fraction.
Cup
Spindle
Motor
4.4 cm
4 c
m0.2 cm
Cup
Spindle
Motor
4.4 cm
4 c
m0.2 cm
Cup
Spindle
Motor
4.4 cm
4 c
m0.2 cm
Figure 3. 3. Schematic of the Couette flow apparatus
3.5.3. UV Disinfection Assay
A collimated beam apparatus equipped with a low pressure mercury UV lamp was
used to irradiate the samples (VQ 650143 Trojan Technologies, London, Ontario, Canada).
The UV transmittance of the samples was measured at 254 nm using a Lambda 35 UV-Vis
Spectrometer (Perkin-Elmer, Woodbridge, Ontario, Canada). This spectrometer features a
double-beam all-reflecting system, installed with an integrating sphere device: Labsphere
RSA-PE-20 (Perkin-Elmer, Woodbridge, Ontario, Canada). Light intensity was measured
using an IL-1700 radiometer (International Light, Peabody, MA, USA). UV transmittance
and the measured UV intensity were used to estimate the irradiation time required to
48
achieve the desired UV dose according to method described by Bolton and Linden (Bolton
& Linden 2003).
To obtain the UV DRC, 20 mL of resuspended samples was placed in a glass Petri
dish and subjected to a range of UV doses between 10 and 60 mJ/cm2. The irradiated
samples were cultured using the membrane filtration method following the Standard
Methods for the Examination of Water and Wastewater (APHA 2001, # 9222 A). m-FC
agar (VWR, Mississauga, Ontario) was used as the culturing media and the buffer solution
used for rinsing the flocs on the filter (Fisher brand Water Testing membrane Filters, 47mm,
0.45µm, Mississauga, Ontario) was 13.6 g/L KH2PO4 at pH 7.2. The samples were then
incubated at a temperature of 44 ºC for 24 ± 2 hours, and then the colony forming units
were enumerated. The DRC data was analyzed using MathematicaTM v.7 (Wolfram
Research, Champaign, USA) to determine parameters of the double exponential model. UV
bioassay was performed on both “core” and “floc” samples of various size fractions.
Since floc and core samples were prepared by sieving and resuspension of mixed
liquor flocs in milliQ water, it is more meaningful to report the UV DRC data for these
samples as the normalized survival ratio; N/N0. The value of N0 in these samples was
typically between 104-105 CFU/ 100 mL.
3.5.4. Optical Microscopy and Size Distribution
Optical microscopy images of core and floc samples were obtained using a Leica
Optical DM LA microscope (Leica Microsystems, Bannockburn, IL, US) equipped with a
10X objective lens. The field of view at this magnification was 0.870mm × 0.690mm, with
an image resolution of 0.67 µm/pixel.
49
Particle size analyses of core and floc samples were carried out using the Multisizer
3.0 particle analyzer. This device also allows for the measurement of particle concentration,
i.e. # of particles per unit volume of the sample. It should be noted that the Multisizer
operates based on Coulter principal, and hence underestimates the physical size of porous
particles, such as flocs (Kachel 1986).
3.5.5. Statistical Analysis
The statistical significance of the results was tested by performing a student t-test
on the double exponential parameters extracted from each data set. Since each UV DRC
was repeated in triplicates, three values were obtained for k1, k2 and β for each sample.
The student t-test was done in pairs for all size fractions for k1, k2, and β with α = 0.05.
Therefore, if the calculated p-value was smaller than 0.05 the difference between the UV
DRC for any two samples being compared was statistically significant. On the other hand,
if the p-value was larger than 0.05 the two samples were not statistically significantly
different in terms of UV inactivation kinetics.
3.6. Results and Discussion
UV DRC experiments conducted in this study were divided in two sets: 1- UV DRC
of un-sheared flocs; and 2- UV DRC of compact cores. In the first set of experiments,
mechanical sieving was used to isolate flocs of various size fractions. Subsequently, the
UV disinfection assay was performed to obtain the DRC for each floc size fraction. In the
second set of experiments, mixed liquor samples were first sheared and subsequently
sieved to isolate compact cores of various size fractions. The DRC’s of the isolated core
50
samples were then determined using UV disinfection assay. Results of these experiments
are provided and discussed in the following sections.
3.6.1. Flocs UV DRC
In this section, effect of floc size on the kinetics of UV disinfection is examined and
experimental results are compared to those obtained using the double exponential model.
The normalized UV DRC of the un-sheared floc samples are given in Figure 3. 4.
The DRC of each of these samples has a steep initial slope followed by near-plateau tailing
region. As discussed earlier, the initial slope of the UV DRC of flocs is due to the presence
of UV-susceptible (Type 1) flocs in the sample while the tailing region represents the
disinfection of the UV-resistant (Type 2) flocs. It should be pointed out that in this study,
since the floc samples were washed with MilliQ water, the number of free organisms in the
samples was negligible.
0.0001
0.001
0.01
0.1
1
0 20 40 60 80
UV Dose (mJ/cm²)
N /
N 0
75-90µm
53-63µm
32-45µm
Figure 3. 4. UV dose-response curves (DRCs) for floc samples of different size fractions. Solid curves in
figure represent the fitted double-exponential model. The data presented includes replicates of the
same experiment over three months with at least four samplings.
51
Table 3. 1. Inactivation rate constants (k1 and k2 ) and UV-resistant fraction (β) of un-sheared floc
samples. The ± values represent 95% confidence intervals.
Size k1 (cm2/mJ) k2 (cm
2/mJ) β
32-45 µm 0.23 ± 0.01 0.011 ± 0.004 0.04 ± 0.004
45-63 µm 0.22 ± 0.02 0.010 ± 0.006 0.07 ± 0.01
53-63 µm 0.19 ± 0.01 0.014 ± 0.006 0.09 ± 0.01
63-75 µm 0.18 ± 0.02 0.013 ± 0.004 0.12 ± 0.01
75-90 µm 0.17 ± 0.02 0.012 ± 0.003 0.14 ± 0.01
The double exponential model parameters for un-sheared floc samples; i.e. k1, k2,
and β, are given in Table 3. 1. This table shows that the initial inactivation rate constant for
un-sheared flocs, k1, ranged from 0.17 to 0.23 cm2/mJ that was significantly lower than the
Assuming that core density is twice that of the loose shell, and that the UV
absorbance of cores and loose shells are proportional to their densities, then for Ac = 1000
cm-1 the UV absorbance of loose shell will be AS= 500 cm-1. Using Equations [3-5] and [3-
10] together, the UV DRC’s of model spherical flocs with a diameter of 60 µm were
estimated and plotted in Figure 3. 11. According to this figure, in the absence of any
compact core (identified as ‘No core’ in Figure 3. 11), at high UV doses (> 20 mJ/cm2)
although the inactivation rate slowed down, the kind of tailing phenomenon that was
experimentally observed earlier (see Figure 3. 4) was not evident. With increasing core
diameter, or equivalently core volume, the tailing level increased and the initial slope of the
UV DRC of flocs became less steep. For comparison, the slope of the tailing region of UV
DRC for the flocs containing 40 vol% cores was about 0.027 cm2/mJ which is comparable
to the calculated value for the slope of tailing region of 60µm cores (0.02 cm2/mJ) seen in
Figure 3. 10-a. It is also important to notice the experimental values of 0.01-0.014
(cm2/mJ) found for the slope of tailing region of flocs (Table 3. 1) and isolated compact
cores (Table 3. 3) are of the same order of magnitude of the calculated values. These results
suggest that the loose shell-compact core model of floc structure presented in here
successfully captures qualitative and quantitative features of the UV dose response curve of
flocs. Furthermore, the similarities between the inactivation rate constants of compact
62
cores and loose shells support that the tailing of UV DRC of flocs is controlled by the
presence of compact cores.
0.001
0.01
0.1
1
0 10 20 30 40 50 60 70
N /
No
UV Dose (mJ/cm2)
10% (28 μm)
40% (44 μm)
70% (53 μm)
No core
Figure 3. 11. UV DRC spherical flocs for various 0%, 10%, 40%, and 70% core volume fractions. Core
diameter is given in parenthesis. Floc diameter is 60µm, and the UV absorbance of floc and core are
assumed to be 500 cm-1
and 1000cm-1
, respectively.
3.7. Conclusions
The conclusions of this study are:
• Floc structure has a significant effect on disinfectability with UV light. The
presence of compact cores in flocs reduces the UV penetration and hence
increases their resistance to UV disinfection.
• Cores have a higher tailing level compared to flocs of similar size.
• Larger flocs contain a larger number and a larger volume of cores. Therefore,
they are more likely to be resistant to UV disinfection and hence have a higher
UV DRC tailing level.
63
• In the tailing region, the inactivation rate constant of un-sheared flocs and
compact cores were essentially the same. This finding supports the idea that
compact cores, are responsible for the tailing phenomenon of UV DRC of
wastewater flocs.
• For the cores larger than 45 µm, UV inactivation kinetics are not affected by the
core size. In contrast for flocs, an increase in β was observed with increase in
floc size. This phenomenon was explained by the suggested model of having a
viable layer of a certain thickness for cores of all sizes.
3.8. References
APHA (2001), Standard methods for the examination of water and wastewater. 22nd Ed. Washington DC, USA Bolton, J. R., Linden, K. G. (2003). Standardization of methods for fluence (UV Dose) determination in bench-scale UV experiments. Journal of Environmental Engineering, 129 (3), pp. 209 – 215 Cerf, O. (1977). Tailing of survival curves of bacterial spores. Journal of Applied Bacteriology, 42(1), 1-19. Das, T. K. (2001). Ultraviolet disinfection application to a wastewater treatment plant. Clean Products and
Processes, 3(2, pp. 69-80), August. Eriksson, L., Steen, I., & Tendaj, M. (1992). Evaluation of sludge properties at an activated sludge plant. Water Science and Technology, 25(6), 251-265. Farnood, R., Flocs and ultraviolet disinfection. Chapter 18 in Flocculation in Natural and Engineering
Systems. I.G. Droppo, G.G. Leppard, S.N. Liss, and T.G. Milligan, Eds., CRC Press, Boca Raton, FL, 2005, pp. 385-395. Forster, C. F. (1985). Factors involved in the settlement of activated sludge-II; the binding of polyvalent metals. Water Research, 19(10), 1265-1271. Jagger, J., (1967) Introduction to Research in Ultraviolet Photobiology. Prentice-Hall, Inc. Englewood Cliffs, NJ Jenkins, D. (1993). In Richard M. G., Daigger G. T. (Eds.), Manual on the causes and control of activated
sludge bulking and foaming (2nd Ed.). Boca Raton: Lewis. Jorand, F., Boué-Bigne, F., Block, J. C., & Urbain, V. (1998). Hydrophobic/hydrophilic properties of activated sludge exopolymeric substances. Water Science and Technology, 37(4-5), 307-315. Kachel, V. (1986). Investigations into Coulter Sizing of Biological Particles; Theoretical Background and Instrumental Improvements. Particle Characterization, 3(2), 45-55.
64
Liao, B. Q., Allen, D. G., Leppard, G. G., Droppo, I. G., & Liss, S. N. (2002). Interparticle interactions affecting the stability of sludge flocs. Journal of Colloid and Interface Science, 249(2), 372-380. Li, D., & Ganczarczyk, J. J. (1987). Stroboscopic determination of settling velocity, size and porosity of activated sludge flocs. Water Research, 21(3), 257-262. Liss, S. N., Droppo, I. G., Flannigan, D. T., & Leppard, G. G. (1996). Floc architecture in wastewater and natural riverine systems. Environmental Science and Technology, 30(2), 680-686. Loge, F. J., Emerick, R. W., Heath, M., Jacangelo, J., Tchobanoglous, G., & Darby, J. L. (1996). Ultraviolet disinfection of secondary wastewater effluents: Prediction of performance and design. Water Environment
Research, 68(5), 900-916. Loge, F. J., Emerick, R. W., Thompson, D. E., Nelson, D. C., & Darby, J. L. (1999). Factors influencing ultraviolet disinfection performance part I: Light penetration to wastewater particles. Water Environment
Research, 71(3), 377-381 Matasci, R. Weston, P. Lau, J. Cruver, S. Marek, D. Tomowich (1999), Wastewater Technology Fact Sheet: Ultraviolet Disinfection, United States Environmental Protection Agency, Office of Water, Washington, DC, EPA 832-F-99-064, R. Pavoni JL, Tenney MW, & Echelbreger JR WF. (1972). Bacterial exocellular polymers and biological flocculation. Journal of the Water Pollution Control Federation, 44(3 pt 1), 414-431 Pfeifer, G. P., You, Y., & Besaratinia, A. (2005). Mutations induced by ultraviolet light. Mutation Research -
Fundamental and Molecular Mechanisms of Mutagenesis, 571(1-2 SPEC. ISS.), 19-31. Piluso, L. G., & Moffatt-Smith, C. (2006). Disinfection using ultraviolet radiation as an antimicrobial agent: A review and synthesis of mechanisms and concerns. PDA Journal of Pharmaceutical Science and
Technology, 60(1), 1-16. Qualls, R. G., Ossoff, S. F., & Chang, J. C. H. (1985). Factors controlling sensitivity in ultraviolet disinfection of secondary effluents. Journal of the Water Pollution Control Federation, 57(10), 1006-1011. Sanin, F. D., & Vesilind, P. A. (1994). Effect of centrifugation on the removal of extracellular polymers and physical properties of activated sludge. Water Science and Technology, 30(8 pt 8), 117-127. Sheng, G., Yu, H., Li, X. (2006). Stability of sludge flocs under shear conditions: Roles of extracellular polymeric substances (EPS). Biotechnology and Bioengineering, 93(6), 1095-1102 Tan, T. (2007). Understanding the effect of particle-size on UV disinfection: Kinetics, mechanism and modeling. M.A.Sc. Dissertation # 134, University of Toronto. Urbain, V., Block, J. C., & Manem, J. (1993). Bioflocculation in activated sludge: An analytic approach. Water Research, 27(5), 829-838. Wright, H.B., and Carins W.L. Ultraviolet Water Disinfection, USEPA Workshop on UV Disinfection of
Drinking Water, Arlington, VA, April 1999. Yuan, Y. (2007). Investigating the relationship between the physical structure and shear strength of bioflocs. Ph.D. Dissertation #177, University of Toronto. Yuan, Y., & Farnood, R. R. (2010). Strength and breakage of activated sludge flocs. Powder Technology,
199(2), 111-119.
65
4. The Effect of Biological Nutrient Removal on UV
Disinfection Kinetics2
4.1. Abstract
The microbial aggregates (flocs) generated in the secondary biological treatment
process protect microorganisms from exposure to UV light, decreasing the disinfection
efficiency. In this study the UV disinfection performance of two different secondary
treatment processes were compared. The two processes were: a conventional activated
sludge system (CAS), and a biological nutrient removal (University of Cape Town) (BNR-
UCT). In addition to UV disinfection performance, the physicochemical characteristics of
flocs were also compared. The BNR-UCT flocs were more open and less spherical
compared to the CAS flocs. Both the mixed liquor flocs and the secondary effluents
collected from the BNR-UCT process were easier to disinfect with UV light by 0.5 log, and
1 log, respectively. This study demonstrates the effect of the secondary process conditions
on floc properties and the UV disinfection rates, thereby providing valuable information for
modifying/designing upstream processes to increase disinfection efficiency.
4.2. Introduction
The effectiveness of ultraviolet (UV) disinfection decreases in the presence of
suspended microbial flocs formed in the activated sludge process (Qualls et al. 1983, Das
2 . This Chapter is based on the following article: Azimi. Y., Pileggi. V, Chen. X., Allen. D. G., Farnood. R., Droppo. I., Seto. P. (2013). The Effect of Secondary Biological Treatment Process Conditions on UV Disinfection of Wastewater. Water Science and
Technology, 64 (12), 2719-2723.
66
2001, Tan 2007). Tailing is representative of the disinfection efficiency reduction at higher
UV doses.
It is known that transmittance affects the UV disinfection efficiency. Loge et al.
(1999) suggested the light pathways concept inside a floc, and how target organisms that
have a relatively unobstructed pathway to the bulk medium are disinfected more easily with
UV light. Flocs are mainly composed of extracellular polymeric substances (EPS) which
are high absorbers of UV light (Farnood 2005). As floc density increases there is a higher
volumetric concentration of extracellular polymeric substances. Therefore, according the
Beer-Lamberts law there is a lower intensity of UV light inside the flocs that reduces the
rate of UV disinfection. In the previous Chapter, it was found that the compact core’s UV
DRC has a higher tailing level compared to flocs, implying that they are harder to disinfect
and likely responsible for the tailing of UV DRC (Figure 4. 1).
-3.5
-3
-2.5
-2
-1.5
-1
-0.5
0
0 20 40 60 80UV Dose (mJ/cm²)
Ln
(N
/N0
)
Cores 53-63µm
Un-sheared flocs 53-63µm
Figure 4. 1. UV DRCs of un-sheared flocs and cores of similar size
It is known that floc structure can be changed by varying the activated sludge
process conditions. Flocs generated at higher SRTs have more regular shapes, settle better,
and yield a cleaner secondary treated effluent (Liao et al. 2002, Xie &Yang 2009). Lower
67
DO concentrations in the secondary biological treatment process causes the formation of
porous flocs with poor settling properties (Wilén & Balmér 1999).
Nowadays many wastewater treatment plants have been constructed for the
purpose of biological nutrient removal (BNR). In a BNR process, anaerobic and anoxic
zones are incorporated with aerobic zones for simultaneous removal of nitrogen and/or
phosphorus in addition to carbonaceous removal (i.e., BOD, COD). Bulking is a common
problem in systems where the operation shifts in time and/or space between aerobic, anoxic
and anaerobic conditions (Eikelboom 1998, Yun et al. 2000, Martins et al. 2004,
Vaiopoulou et al. 2007). Moon et al. (2004) used sequencing batch reactors to study the
effect of anaerobic conditions on floc size and concluded that anaerobic flocs are smaller
than aerobic flocs. Martins et al. (2004) concluded that the presence of microaerophilic
zones in the anoxic stages of BNR systems leads to worsening sludge–settling
characteristics. Wilén and Balmér (1999) studied the effect of dissolved oxygen (DO)
concentrations on floc structure and settling properties. They concluded that low DO
concentrations produce sludge with poor settling characteristics mainly due to filamentous
bacterial growth and the formation of porous flocs. Moreover, they reported that by
alternating aerobic/anaerobic conditions, the number of small particles increased during the
anaerobic period and higher turbidities were observed.
Although BNR systems are widely used, there has been hardly any research done
on comparing the UV disinfection efficiency of these systems with that of a conventional
activated sludge (CAS) treatment process. Therefore, the main objective of this study was
to compare the structure and UV disinfectability of flocs generated under CAS conditions
68
and BNR conditions. The BNR was a modified UCT (University of Cape Town) process,
BNR-UCT.
4.3. Materials and Methods
4.3.1. Sample Collection from the CAS and BNR-UCT Reactors
The samples used for this study were collected from two pilot reactors in the
National Water Research Institute (NWRI), Environment Canada, Burlington, Canada.
Both reactors were equipped with identical upflow blanket clarifiers. One of the reactors
was a conventional fully aerated activated sludge system, and the other was a BNR-UCT,
which contained anaerobic, anoxic and aerobic zones. Each reactor had an operating
volume of 350 L with six identical cells in series. Baffles separated the reactor into the six
cells in order to simulate plug flow; mixed liquor flowed from one cell to the next via 25
mm diameter openings. In addition, a heating/cooling jacket was used to keep the reactors
at a temperature of 12 ± 1 ˚C, a common operating temperature at municipal wastewater
treatment plants in Ontario. The mixed liquor samples were collected from cell # 6 (the
last cell in the aerated zone) from both reactors.
The total sludge retention time for the BNR-UCT was 40 days (20 days anaerobic/
anoxic, and 20 days aerobic). From the 6 cells of the BNR-UCT reactor, cell #1 was
anaerobic, cells #2 and #3 were anoxic, and cells #4, #5, and #6 were aerobic, and hence
there was a ratio of 1:2:3 for the anaerobic, anoxic, and aerobic regions respectively. The
total sludge retention time for the CAS system was 20 days. Several operational
parameters for both reactors are listed in Table 4. 1. Schematics of the BNR-UCT and the
CAS reactors are displayed in Figure 4. 2.
69
835
626.25
1670
835
835
626.25
1670
835
(a)
835
626.25
835
626.25
835
626.25
835
626.25
(b)
Figure 4. 2. Schematic of the BNR-UCT (a), and the CAS (b) process
The influent to the pilot reactors was a primary treated sewage from the Skyway
Wastewater Treatment Plant (WWTP) located in Burlington, Ontario, Canada to NWRI.
The average influent characteristics are listed in Table 4. 2. For the BNR- UCT reactor, in
order to provide an easily accessible carbon source in the anaerobic zone for the
The collected mixed liquor samples were fractionated using sieve trays (U.S.A.
Standard Testing Sieve) with the nominal opening sizes of 32 and 53 microns. Collected
floc fractions were gently washed with UV disinfected RO water and back washed into a
flask. The fractionated samples prepared are referred to as “32-53 CAS” and “32-53,
BNR-UCT”, respectively from CAS and BNR-UCT reactors.
4.3.3. BNR-UCT and CAS Floc Characteristics
The particle size distribution and sphericity of mixed liquor flocs were measured
using Rapid VUE (Beckman Coulter, Miami, USA). This equipment uses image analysis
to calculate particle characteristics in the range of 10-1000µm equivalent area diameter.
Optical microscopy images were taken using an Olympus BX51 Microscope
(LECO Instruments, Mississauga, Ontario, Canada) and the Olympus Cellsens 1.4 software
using a 40X oil immersion objective lens, at a resolution of 0.16 µm.
71
4.3.4. UV Disinfection Assay
A collimated beam apparatus equipped with a low pressure mercury UV lamp was
used to irradiate the samples (Trojan Technologies, London, Ontario, Canada). Each
samples UV transmittance was measured at 254 nm using a UV-1700 Pharmaspec UV-Vis
Shimadzu Spectrometer (Shimadzu Scientific Instruments, Maryland, USA). An IL-1400A
radiometer (International Light, Peabody, MA) was used to measure the UV Light intensity.
The samples’ UV transmittance and the measured UV intensity of the lamp were used to
estimate the irradiation time required to achieve the desired UV dose according to the
method described by Bolton and Linden (2003).
To obtain the UV dose response curve (DRC), 100 mL of each sample was placed
in plastic Petri dishes and exposed to a range of UV doses between 10 and 60 mJ/cm2. The
irradiated samples were then cultured using the membrane filtration method following the
Standard Methods for the Examination of Water and Wastewater (# 9222 A, APHA 2001).
m-FC agar (VWR, Mississauga, Ontario) was used as the culturing media to culture the
surviving fecal coliform. The samples were incubated at 44 ºC for 24 ± 2 hours, and then
the colony forming units (CFU) were enumerated. The UV assay was performed on the
mixed liquor, as well as the secondary effluents of both reactors.
All the UV DRC data were fit into the double exponential model, which has been
previously used to characterize the kinetics of UV disinfection for wastewater flocs and
secondary effluents (Farnood 2005):
DkDkee
N
N21)1(
0
−− +−= ββ [4-1]
In this equation, D is the delivered ultraviolet dose, in mJ/cm2, N is the number of
colonies formed from the survived coliform per 100 mL of irradiated sample, and N0 is the
72
number of colonies formed per 100 mL of unexposed sample. Parameters k1 and k2 are the
UV inactivation rate constants; k1 is the inactivation constant for UV susceptible particles
(i.e. initial slope of the DRC). The parameter k2 is the inactivation constant for the UV
resistant particles (i.e. the slope of tailing region). The parameter β is the ratio of UV
resistant particles to the total number of viable particles.
4.3.5. Mechanical Strength of the BNR-UCT and CAS Flocs
In order to compare the mechanical strength of the CAS and BNR-UCT flocs, the
breakage percentage of the flocs was compared under identical hydrodynamic shear stress.
Solutions containing 32-63 µm flocs with similar solids content (0.06 %V) were prepared.
Ten milliliters of each floc sample was placed in the gap between the spindle and the cup in
a Couette flow cell (schematic provided in Figure 3. 3). The spindle speed was adjusted to
achieve the target shear rate and the sample was sheared for 10 minutes. During the
shearing process, flocs that have a higher mechanical strength than the applied turbulent
shear stress will not break.
After shearing at various degrees (controlled with the speed of the spindle ranging
from 2,000-5,000 rpm) the particle size distribution was measured using Multisizer 3
(Beckman Coulter, Miami, USA). Breakage was defined based on the number of flocs that
survive the shearing process (Yuan & Farnood 2010, Yong et al. 2009, Gibson et al. 2009):
% Breakage = (1 – 0P
Ps
N
N) x 100 [4-2]
where NP0 and NPs represent the initial (before shearing) and the final (after shearing)
number of flocs larger than the mode floc size before shearing. Based on this equation, the
73
percentage of breakage was calculated and is reported as a function of the turbulent shear
stress induced by the Couette flow.
4.3.6. Composition Analysis
To extract the EPS from sludge and measure the protein and carbohydrate content,
the cation exchange resin method was applied according to the method developed by
Frølund et al. (1996). Since at the time of this particular test the reactors were no longer in
operation, mixed liquor samples were prepared by resuspending previously freeze dried
sludge samples in milliQ water. To assess the effect of freeze drying the samples on the
measured composition, an extraction followed by composition analysis was done on an
identical sample in both the fresh and frozen dried states and no significant difference was
observed.
The carbohydrate content of the extracts was measured using the phenol-sulfuric
acid method with D-glucose as standard (Masuko et al. 2005), and the protein content was
measured using the BCATM Protein Assay kit and a modified Lowry method (Haff 1978)
using bovine serum albumin (BSA) as standard. Spectrometric readings were measured
using the multiwell-plate reader ThermoU Spectra III A-5082 from SLT-Labinstruments
(TECAN, Durham, NC, United States).
4.4. Results and Discussion
4.4.1. Microscopy and Morphological Characteristics
The optical microscopy images from the BNR-UCT and CAS flocs, taken under
identical illumination are presented in Figure 4. 3. The images show that the CAS flocs
had a higher optical density in comparison to the BNR-UCT flocs suggesting that the CAS
74
process produces denser flocs. Moreover, the BNR-UCT flocs show a more homogeneous
structure compared to the CAS flocs where there are denser regions within the flocs.
Figure 4. 3. Typical optical microscopy images of the mixed liquor flocs collected from the CAS and
BNR-UCT reactors
The BNR-UCT process generates smaller and less spherical flocs (Figure 4. 4).
Possible explanations to having smaller less spherical flocs in the BNR-UCT process
include:
1- The aerobic outer layer built up in the aerated region of the BNR-UCT process,
sloughs off when the flocs enter the anaerobic and anoxic regions. Therefore, the repeated
cycling between aerated and unaerated regions limits floc size. If methane is generated
during the anaerobic reduction of organic matter, it may diffuse out and, as a result of low
solubility in water, an accumulation of methane may occur in the internal zone of the flocs.
This may lead to the formation of bubbles which may ultimately cause sloughing off layers
75
from the flocs (Henze 2002). The formation of bubbles can also contribute to the lower
sphericity of the BNR-UCT flocs.
2- The formation of microaerophilic zones within the flocs in the anoxic region as a
result of nitrogen release during denitrification causes floc shape deformation (Martins et al.
2004). The release of nitrogen bubbles within the BNR-UCT flocs causes them to be less
spherical and more homogeneous than the CAS flocs.
0
2
4
6
8
10
15 24 38 60 96 151
Floc Size (µm)
Pa
rtic
le N
um
be
r %
BNR-UCT
CAS
(a)
0
5
10
15
20
25
0.2 0.4 0.5 0.7 0.8 1.0Sphericity
Pa
rtic
le N
um
be
r %
BNR-UCT
CAS
(b)
Figure 4. 4. Sludge floc size distribution (a) and sphericity (b) of the CAS and BNR-UCT flocs
76
4.4.2. UV Disinfection Kinetics of the BNR-UCT and CAS Flocs
UV Disinfection of BNR-UCT and CAS Mixed Liquor Flocs
In order to eliminate the effect of floc size on UV disinfection kinetics, flocs from
both processes were isolated into similar narrow size fractions (Tan 2007, Madge & Jensen,
2006). Figure 4. 5 demonstrates the normalized UV dose response curves for 32-53 µm
flocs of from both reactors as well as the fitted double exponential equation. The BNR-
UCT flocs’ DRC show a lower tailing level in comparison to the CAS flocs. The steeper
initial slope, and lower tailing level of the BNR-UCT’s UV DRC suggest that the BNR-
UCT flocs are easier to disinfect with UV light (e.g. for the same dose, there is a larger
reduction in CFU). The disinfection kinetics parameters for the data points in this graph
were extracted by fitting the DRC data into the double exponential equation using
MathematicaTM v.7 (Wolfram Research, Champaign, USA) and are listed in Table 4. 3.
The errors represent the standard deviation for each parameter from five different DRCs
constructed on different days.
The data in Table 4. 3 suggests that there is a significant difference in the UV
resistance (β) between CAS and BNR-UCT flocs (student’s t-test, p= 0.01). The double
exponential parameters values are comparable to those reported in Chapter 3.
77
0.00001
0.0001
0.001
0.01
0.1
1
0 20 40 60
UV Dose (mJ/cm²)
N /
N 0
BNR - UCTCAS
32-53 µm CAS
32-53 µm BNR-UCT
Figure 4. 5. Normalized UV dose response curves (data points and fitted double exponential model) for
the 32-53 µm flocs from the CAS and BNR-UCT flocs
Table 4. 3. UV disinfection kinetics parameters for the 32-53 µm flocs from the CAS and BNR-UCT
reactors
Reactor k1 k2 β
CAS 0.17 ± 0.03 0.019 ± 0.003 0.10 ± 0.01
BNR-UCT 0.25 ± 0.02 0.020 ± 0.003 0.06 ± 0.02
UV Disinfection of BNR-UCT and CAS Final Effluents
The normalized UV DRCs for the CAS and BNR-UCT secondary effluents are
shown in Figure 4. 6. The BNR-UCT effluent shows a higher log reduction at every UV
dose compared to the CAS effluent. As demonstrated in Figure 4. 6, at the UV dose of 10
mJ/cm2 the BNR-UCT effluent showed 3.5 log reduction whereas the CAS effluent shows
about 1.5. In terms of actual effluent fecal coliform counts the BNR-UCT had 5 times
more fecal coliform than the CAS (4×105 CFU / 100 mL vs. 8×104 CFU / 100 mL). At the
UV-dose 20 mJ/cm2 the corresponding fecal coliforms CFU / 100 mL were 40 and 1000 for
78
the BNR-UCT and CAS effluents, respectively. At the UV-dose 40 mJ/cm2 the numbers
dropped to 30 and 150 CFU/100ml.
0.000001
0.00001
0.0001
0.001
0.01
0.1
1
0 10 20 30 40 50 60
UV Dose (mJ/cm²)
N/N
0BNR - UCT
CAS
CAS Effluent
BNR - UCT Effluent
(a)
1.E+00
1.E+01
1.E+02
1.E+03
1.E+04
1.E+05
1.E+06
0 10 20 30 40 50 60
UV dose (mJ/cm²)
N
( C
FU
/ 1
00 m
L)
CAS
BNR - UCT
(b)
Figure 4. 6. Normalized (a) and actual (b) UV DRCs for the secondary effluents collected from both
the CAS and BNR-UCT reactors.
79
It should be noted that the UV DRC data was collected at least five times over the
period of December 2011 to March 2012, and the variations among the DRC of the
replicates from different days were within a third of a log (i.e. well within the experimental
error).
4.4.3. Mechanical Strength and Composition of the BNR-UCT and CAS Flocs
Although the BNR-UCT flocs have a more diffused structure and disinfect better,
they are more resistant to breaking under hydrodynamic shear stress (Figure 4. 7). The
median shear stress (τ50
), i.e. the shear stress at which the number of flocs reduced by 50
%, is listed for both the BNR-UCT flocs and the CAS flocs of a 32-63 µm fraction in Table
4. 4. Stronger flocs would have a higher τ50
, as a higher shear stress would be required to
break the flocs. The BNR-UCT flocs had a higher τ50
, implying that they were more
resistant to breakage. Similar results (higher mechanical strength of the BNR-UCT flocs)
were obtained when sonication was applied to break the flocs (not presented here). This is
despite the fact that the BNR-UCT flocs are irregular in shape and have a more open
structure.
Table 4. 4. Median 10%ile and 90%ile shear stress values for floc breakage for the BNR-UCT and
CAS flocs (32-63µm fraction)
Sample τ 50 (Pa) τ 10 (Pa) τ 90 (pa)
CAS 9.4 1.8 16.7
BNR-UCT 20 4 36
80
0
20
40
60
80
100
0 5 10 15 20 25Turbulent Shear Stress (Pa)
% B
rea
ka
ge
BNR-UCT
CAS
Figure 4. 7. Breakage percentage of the CAS and BNR-UCT flocs (32-63 µm) under various degrees of
hydrodynamic shearing with the Couette Flow
In Chapter 3 it was concluded that the compact cores in flocs are mechanically
stronger and harder to disinfect. In this study the BNR-UCT flocs are mechanically
stronger, less integrated and easier to disinfect with UV light. The new findings suggest
that the flocs’ mechanical strength and UV disinfection efficiency are not necessarily
correlated.
The composition analysis of the CAS and BNR-UCT flocs showed that the BNR-
UCT flocs have a significantly lower protein: carbohydrate ratio (2.6 ± 0.3 for the CAS
flocs vs. 1.2 ± 0.6 for the BNR-UCT flocs). This is consistent with the work of Sheng et al.
(2008) who reported that flocs with higher stability have EPS with a lower protein to
carbohydrate ratio. The finding is also in accordance with Wilén et al. (2008), as they
reported that mechanically weaker flocs were formed when the sludge protein content
increased. The lower protein to carbohydrate ratio and higher mechanical stability of the
81
BNR-UCT flocs compared to that of the CAS flocs may be explained with the oxidation
state of iron in both processes. In the BNR-UCT the sludge is cycling between aerobic and
anaerobic phases and iron is present as both ferric and ferrous, whereas in the fully aerated
CAS process ferric is the dominant iron form (Rockne 2007). As ferric iron has a higher
affinity to proteins (Murphy et al. 2000) it could cause a higher accumulation of proteins in
the CAS sludge compared to the BNR-UCT sludge. Moreover, ferrous builds stronger
bonds with EPS due to a higher ionic strength in neutral pH compared to ferric
(Oikonomidis et al. 2010). This may cause the overall sludge floc to have a higher
mechanical strength in the case of BNR-UCT process compared to CAS.
4.5. Conclusions
In this study, the effect of biological nutrient removal (specifically the BNR-UCT
process) on the UV inactivation of sludge flocs and secondary effluents was investigated.
Moreover, the mechanical strength and the composition of the flocs generated under CAS
and BNR-UCT were compared. The results showed that flocs of similar size generated in
the BNR-UCT process are more susceptible to UV disinfection than those generated under
CAS conditions. The BNR-UCT flocs show a more diffused, and less integrated structure
in comparison to the intact structure of the CAS flocs. The presence of the anaerobic and
anoxic zones and cycling between the aerated and unaerated zones in the BNR-UCT
process may be the cause of these differences observed. Although the BNR-UCT flocs
show a more open structure in the microscopy images and are less spherical in shape, they
show a higher mechanical strength in comparison to the CAS flocs. These findings show
82
that altering bioprocess conditions, such as adding extended unaerated zones, may improve
UV disinfectability.
4.6. References
American Public Health Association (APHA) (2001), Standard methods for the examination of water and wastewater. 22nd Ed. Washington DC, USA Bolton, J. R., Linden, K. G. (2003). Standardization of methods for fluence (UV Dose) determination in bench-scale UV experiments. Journal of Environmental Engineering, 129 (3), pp. 209 – 215. Das, T. K. (2001). Ultraviolet disinfection application to a wastewater treatment plant. Clean Products and
Processes, 3(2, pp. 69-80), August. Eikelboom, D. H., Andreadakis, A., Andreasen, K. (1998). Survey of filamentous populations in nutrient removal plants in four European countries. Water Science and Technology, 37(4-5), 281-289. Farnood, R., Flocs and ultraviolet disinfection. Chapter 18 in Flocculation in Natural and Engineering
Systems. I.G. Droppo, G.G. Leppard, S.N. Liss, and T.G. Milligan, Eds., CRC Press, Boca Raton, FL, 2005, pp. 385-395. Frølund, B., Palmgren, R., Keiding, K., Nielsen, P. H. (1996). Extraction of extracellular polymers from activated sludge using a cation exchange resin. Water Research, 30(8), 1749-1758. Gibson, J. H., Hon, H., Farnood, R., Droppo, I. G., Seto, P. (2009). Effects of ultrasound on suspended particles in municipal wastewater. Water Research, 43(8), 2251-2259. Haff, A. C. (1978). Mechanized micro-scale determination of protein in platelet pellet sonicates. Clinical
Chemistry, 24(11), 2031-2032. Henze, M. Ed. (2002). Wastewater treatment : Biological and chemical processes: New York: Springer. Liao, B. Q., Allen, D. G., Leppard, G. G., Droppo, I. G., Liss, S. N. (2002). Interparticle interactions affecting the stability of sludge flocs. Journal of Colloid and Interface Science, 249(2), 372-380. Loge, F. J., Emerick, R. W., Thompson, D. E., Nelson, D. C., & Darby, J. L. (1999). Factors influencing ultraviolet disinfection performance part I: Light penetration to wastewater particles. Water Environment
Research, 71(3), 377-381. Madge, B. A., Jensen, J. N. (2006). Ultraviolet disinfection of fecal coliform in municipal wastewater: Effects of particle size. Water Environment Research, 78(3), 294-304. Martins, A. M. P., Heijnen, J. J., Van Loosdrecht, M. C. M. (2004). Bulking sludge in biological nutrient removal systems. Biotechnology and Bioengineering, 86(2), 125-135. Masuko, T., Minami, A., Iwasaki, N., Majima, T., Nishimura, S., Lee, Y. C. (2005). Carbohydrate analysis by a phenol-sulfuric acid method in microplate format. Analytical Biochemistry, 339(1), 69-72. Murthy, S. N., Novak, J. T., & Holbrook, R. D. (2000). Optimizing dewatering of biosolids from autothermal thermophilic aerobic digesters (ATAD) using inorganic conditioners. Water Environment Research, 72(6-7), 714-721.
83
Oikonomidis, I., Burrows, L. J., & Carliell-Marquet, C. M. (2010). Mode of action of ferric and ferrous iron salts in activated sludge. Journal of Chemical Technology and Biotechnology, 85(8), 1067-1076. Qualls, R. G., Flynn, M. P., Johnson, J. D. (1983). The role of suspended particles in ultraviolet disinfection. Journal of the Water Pollution Control Federation, 55(10), 1280-1285. Rockne, K. J. (2007). Kinetic hindrance of fe(II) oxidation at alkaline pH and in the presence of nitrate and oxygen in a facultative wastewater stabilization pond. Journal of Environmental Science and Health - Part A Toxic/Hazardous Substances and Environmental Engineering, 42(3), 265-275. Sheng, G., Yu, H., Li, X. (2008). Stability of sludge flocs under shear conditions. Biochemical Engineering
Journal, 38(3), 302-308. Tan, T. (2007). Understanding the effect of particle-size on UV disinfection: Kinetics, mechanism and modeling. M.A.Sc. Dissertation # 134, University of Toronto. Vaiopoulou, E., Melidis, P., Aivasidis, A. (2007). Growth of filamentous bacteria in an enhanced biological phosphorus removal system. Desalination, 213(1-3), 288-296. Wilén, B., Balmér, P. (1999). The effect of dissolved oxygen concentration on the structure, size and size distribution of activated sludge flocs. Water Research, 33(2), 391-400. Wilén, B. M., Lumley, D., Mattsson, A., & Mino, T. (2008). Relationship between floc composition and flocculation and settling properties studied at a full scale activated sludge plant. Water Research, 42(16), 4404-4418. Xie, B., Yang, S. (2009). Analyses of bioflocculation and bacterial communities in sequencing batch reactors. Environmental Engineering Science, 26(3), 481-487. Yong, H. N., Farnood, R. R., Cairns, W., Mao, T. (2009). Effect of sonication on UV disinfectability of primary effluents. Water Environment Research, 81(7), 695-701. Yuan, Y., & Farnood, R. R. (2010). Strength and breakage of activated sludge flocs. Powder Technology,
199(2), 111-119.
84
5. Evidence of Disinfection by Advanced Oxidation
within Wastewater Flocs under UV Irradiation3
5.1. Abstract
In this paper, the role of naturally occurring polyphosphate in enhancing the ultraviolet
disinfection of wastewater flocs is examined. It was found that polyphosphate, which
accumulates within the wastewater flocs in enhanced biological phosphorus removal
(EBPR) processes, is capable of producing hydroxyl radicals under UV irradiation and
hence causing the photoreactive disinfection of microorganisms embedded within flocs.
This phenomenon is likely responsible for the improved UV disinfection of the biological
nutrient removal (BNR) effluent compared to that of conventional activated sludge
effluent. A mathematical model is developed that combines the chemical disinfection by
hydroxyl radical formation within flocs, together with the direct inactivation of
microorganisms by UV irradiation. The proposed model is able to explain the observed
improvement in the UV disinfection of the BNR effluents. This study shows that the
chemical composition of wastewater flocs could have a significant positive impact on their
UV disinfection by catalyzing the production of oxidative species.
3 . This Chapter is based on the following article: Azimi. Y., Allen. D. G., Droppo. I., Seto. P. , Farnood. R. (2013). Evidence of Disinfection by Advanced Oxidation within Wastewater Flocs under UV Irradiation. To be submitted
85
5.2. Nomenclature
DRC: Dose Response Curve
BNR: Biological nutrient removal
BNR-UCT: Biological nutrient removal - modified University of Cape Town
CAS: Conventional activated sludge
EPS: Extracellular polymeric substances
PAO: Phosphorus accumulating organisms
EBPR: Enhanced biological phosphorus removal
MLSS (g/L): Mixed liquor suspended solids
VSS (g/L): Volatile suspended solids
COD (mg/L): Chemical oxygen demand
SVI (mL/g): Sludge volume index
TEM: Transmission electron microscopy
RO: Reverse Osmosis
CFU: Colony forming units
TEM: Transmission electron microscopy
EDX: Energy dispersive X-ray
k” (mg.cm2/L.mW): production rate constant of •OH in the presence of polyphosphate and
UV light
k’ (L/mg.min): Inactivation rate constant of free E.coli with •OH
MB: Methylene blue
I (mW/cm2): Average UV intensity at the surface of a floc
kMB, •OH (L/mol. s): Reaction rate constant of Methylene blue and the hydroxyl radical
86
R (µm): Floc radius
δ (µm): The distance of the point of interest from the surface of the floc
Deff (δ) (mJ/cm2): The effective UV dose delivered in the distance of δ from the floc’s
surface
kUV (cm2/mJ): Inactivation rate constant of free E.coli with UV light
k•OH (L/mg.min): Inactivation rate constant of free E.coli with •OH, combined with the
production of •OH
N0(R, δ) (CFU/100mL): Initial concentration of viable flocs
N(R, δ) (CFU/100mL): Final concentration of viable flocs
keffUV(δ) (cm2/mJ): Effective UV inactivation constant at the distance δ from the surface of
the floc inwards
keff•OH(δ) (L/mg.min): Effective •OH inactivation constant at the distance δ from the surface
of the floc inwards
A (cm-1): UV absorbance of EPS
f (R, δ): Probability density function of R and δ
V (µm3): Total floc volume
δs (µm): Shell thickness in the shell-core model used for the CAS flocs
5.3. Introduction
Ultraviolet disinfection is a well established technology for disinfecting secondary
wastewater effluents. UV light, specifically around the germicidal wavelengths (200-300
nm), inactivates microorganisms by penetrating through the cell wall and altering the
microorganism’s DNA (Jagger 1967, Pfeifer et al. 2005, Piluso & Moffatt-Smith 2006).
87
This is widely believed to be the main mechanism for the UV disinfection of wastewater
effluents.
It has been reported that UV light can also cause disinfection through the
production of highly oxidative species such as hydroxyl radicals with the addition of
photocatalysts. Such species could result in the oxidation of cell wall and cause cell death
(Matsunaga et al. 1985, Maness et al. 1999). In particular Cho et al. (2004) measured the
release of •OH in the UV/TiO2 process, and found a linear correlation between the
concentration of hydroxyl radicals and the rate of free E.coli inactivation. Their study
showed that •OH is approximately 3 to 4 orders of magnitude more effective for E.coli
inactivation than common disinfectants such as chlorine and ozone.
Over the past two decades, many wastewater treatment plants have been
constructed for the purpose of biological nutrient removal (BNR) (Trivedi 2009). In a
BNR process, anaerobic and anoxic zones are incorporated with aerobic zones for
biological removal of nitrogen and/or phosphorus in addition to carbonaceous nutrients.
Enhanced biological phosphorus removal (EBPR) is one of the processes involved in some
biological nutrient removal systems, where polyphosphate accumulating organisms (PAOs)
remove phosphorus from wastewater and store it in the form of polyphosphates. In the
aerobic zone of EBPR, the PAOs oxidize the polyhydroxyalkanoates (which were
synthesized by the PAOs in the anaerobic region) and the energy reserves from oxidation
are stored through phosphate uptake and polymerization (WEF & ASCE 2005, Hirota et al.
2010). The total P-content in the sludge can reach 10–15 % (Stowa 2002). Hirota et al.
(2010) heated EBPR sludge to extract the polyphosphates, and applied polyacrylamide gel
electrophoresis to determine the chain length of the released polyphosphate. They
88
concluded that the main compounds stored by the PAOs are polyP (with chain lengths of
100-200 Pi residues) and trimetaphosphate. Lee et al. (2010) studied the positioning of
PAOs in flocs generated in the biological phosphorus removal process and concluded that
they were distributed evenly throughout the floc. However, they reported that phosphate
concentrations increased as the depth increased to the centre of the flocs, which implies that
some of the polyphosphate likely exists within the EPS outside of the PAO cells. As seen in
chapter 4, flocs formed in the BNR-UCT process (where EBPR occurs) are more readily
disinfected with UV light compared to the conventional activated sludge flocs.
In this study it was hypothesized that the phosphorus compounds, accumulated in
EBPR flocs, can catalyze the production of oxidative species within the flocs under UV
irradiation and enhance UV disinfection. The main objectives of this study were to: 1)
explore the effect of polyphosphate storage on the UV disinfection kinetics of PAO-
containing sludge flocs, and 2) develop a mathematical model for the UV disinfection of
PAO-containing flocs that takes into account both the photoreactive disinfection due to the
production of oxidative species within these flocs and the direct UV inactivation of
microorganisms.
5.4. Materials and Methods
5.4.1. Sample Collection
The samples used for this study were collected from two continuous pilot scale
reactors (effective volume of 350 L), treating real municipal wastewater, and operating at
the National Water Research Institute (NWRI), Environment Canada, Burlington, Canada.
Both reactors were equipped with identical secondary clarifiers, and had recycling sludge
streams. One of the reactors was a conventional fully aerated activated sludge system, and
89
the other was a modified University of Cape Town biological nutrient removal system
(BNR-UCT), which contained anaerobic, anoxic and aerobic zones for simultaneous
nitrogen and phosphorus removal. A heating/cooling jacket was used to keep the reactors at
a temperature of 12 ± 1 ˚C, a common operating temperature at municipal wastewater
treatment plants in Canada. The performance and stable operating conditions of the
reactors were determined by monitoring the mixed liquor suspended solids (MLSS),
volatile suspended solids (VSS), chemical oxygen demand (COD), and sludge volume
index (SVI). Details of the influent COD, total nitrogen and phosphorus in addition to the
operating conditions and reactor performance have been described in Chapter 4.
5.4.2. UV Disinfection Assay
The UV assay was performed as section 4.3.4.
5.4.3. Transmission Electron Microscopy
For transmission electron microscopy (TEM) imaging, the sample preparation
technique was adopted from Liss et al. (1996). The sections made after polymerizing the
samples were observed in transmission mode (TEM) at an accelerating voltage of 80 kV
using a JEOL 1200 EXII TEMSCAN scanning transmission electron microscope.
5.4.4. Photoreactive Properties of Polyphosphates
Trisodium trimetaphosphate (Na3P3O9 > 95%, Sigma Aldrich), was used as the
model compound for this study. This compound is one of the most commonly occurring
derivatives of polyphosphates in natural environments and bacterial cells (Hupfer et al.
2008, Hirota et al. 2010). The photoactivity of this compound was examined by measuring
the discoloration of methylene blue under UV light according to the method described by
90
Zhang et al. (2012). The UV light source used for this experiment was a low pressure
mercury lamp with a main emission line at 254 nm (Trojan Technologies, London, Ontario,
Canada). Several one hundred milliliter oversaturated solutions of trisodium
trimetaphosphate (250 g/L) were prepared, and 10 mg/L methylene blue (0.1% w/v
aqueous solution, Fisher Scientific) was added to each. The solutions were then exposed to
UV light and samples were taken using a 4 mL syringe at various exposure times. These
samples were then passed through a 0.22 µm filter and their absorbance was measured at
665 nm (peak absorbance of methylene blue) using a Lambda 35 UV-Vis Spectrometer
(Perkin-Elmer, Woodbridge, Ontario, Canada). The discoloration percentage of methylene
blue was calculated at each exposure time. For comparison, the test was repeated by
compared to the shell-core floc with no advanced oxidation (representing CAS). It is worth
noting that the no “fitting” parameter was used in constructing the predicted UV DRCs of
BNR-UCT and CAS flocs in this figure. In spite of this, the predicted DRCs in Figure 5. 3
agree well with the experimental UV DRCs of the CAS and BNR-UCT flocs.
99
0.0001
0.001
0.01
0.1
1
0 20 40 60
UV Dose (mJ/cm²)
N /
N 0
Core-shell Model with UVDisinfection (CAS)
Homogeneous Model with UV andAdvanced Oxidation Disinfection(BNR-UCT)
Figure 5. 3. The estimated UV DRC for the shell-core model with only direct UV inactivation
(representing CAS flocs), and for the homogeneous model with direct UV inactivation and advanced
oxidation (representing BNR flocs). UV absorbance of EPS: 1000 cm−1
, floc size: 48 µm. The solid line
is estimated UV DRCs and the scattered data points are experimental data.
5.6.3. The Signifficance of Advanced Oxidation in Floc Disinfection
To better illustrate the relative signifficance of advanced oxidation in terms of floc
disifnection, a parametric study was conducted. Using the above mathematical model, the
UV DRC of model BNR-UCT flocs without advanced oxidation and with various
concentrations of hydroxyl radicals were calculated. Figure 5. 4 shows that in the presence
of advacned ocidation, the UV disinfectability of flocs improves significantly; and as
expected, increasing the concentration of •OH further improves the UV disinfection rate
and lowers the tailing level of UV DRC. A closer examination of this figure shows that by
increasing the •OH concentration within the BNR-UCT floc from 0 mg/L to 5×10-9 mg/L
the tailing level decreased by about 2 logs. Furthermore, the predicted value for the UV
dose required for 3 log reduction of CFU decreased by a factor of more than 5, from ~ 55
mJ/cm2 to merely ~10 mJ/cm2.
100
0.00001
0.0001
0.001
0.01
0.1
1
0 10 20 30 40 50 60
UV Dose (mJ/cm²)
N /
N 0
No Advanced Oxidation (AO)
With AO, •OH = 10^-9 mg/L
With AO, •OH = 2 x 10^-9 mg/L
With AO, •OH = 5 x 10^-9 mg/L
Figure 5. 4. The effect of •OH on the disinfectability of homogeneous spherical flocs with UV light, and
the estimated effect of increasing •OH concentration. (floc size: 48 µm)
It should be noted here that the photoactive role of polyphosphate and its effect on
the additional disinfection of BNR-UCT flocs is one explanation for the lower tailing level
observed in the DRC. As discussed earlier, the morphological differences induced by
biological nitrogen removal (lower sphericity) may also affect the flocs’ UV susceptibility.
Further investigation regarding the effects of each process is required to clarify their
relative contributions to the observed improvements in the UV disinfection of BNR flocs.
5.7. Conclusions
This study shows that advanced oxidation due to the naturally occurring
polyphosphate material may play a significant role in improving disinfection of EBPR
effluents under UV irradiation. Polyphosphate, naturally formed within the flocs during
the EBPR processes, is proposed to act as a photocatalyst to produce hydroxyl radicals
under UV irradiation and accelerate disinfection through cell lysis. This finding provides
101
new and significant opportunities in terms of optimizing upstream treatment processes to
improve the UV disinfection performance of effluents.
5.8. References
Alpert, S. M., Knappe, D. R. U., Ducoste, J. J. (2010). Modeling the UV/hydrogen peroxide advanced oxidation process using computational fluid dynamics. Water Research, 44(6), 1797-1808. Bolton, J. R., Linden, K. G. (2003). Standardization of methods for fluence (UV Dose) determination in bench-scale UV experiments. Journal of Environmental Engineering, 129 (3), pp. 209 – 215. Chen, Y., Yang, S., Wang, K., Lou, L. (2005). Role of primary active species and TiO2 surface characteristic in UV-illuminated photodegradation of acid orange 7. Journal of Photochemistry and Photobiology A:
Chemistry, 172(1), 47-54. Cho, M., Chung, H., Choi, W., Yoon, J. (2004). Linear correlation between inactivation of E. coli and OH radical concentration in TiO2 photocatalytic disinfection. Water Research, 38(4), 1069-1077. Hirota, R., Kuroda, A., Kato, J., Ohtake, H. (2010). Bacterial phosphate metabolism and its application to phosphorus recovery and industrial bioprocesses. Journal of Bioscience and Bioengineering, 109(5), 423-432. Hriberšek, M., Žajdela, B., Hribernik, A., & Zadravec, M. (2011). Experimental and numerical investigations of sedimentation of porous wastewater sludge flocs. Water Research, 45(4), 1729-1735. Hupfer, M., Glöss, S., Schmieder, P., Grossart, H. (2008). Methods for detection and quantification of polyphosphate and polyphosphate accumulating microorganisms in aquatic sediments. International Review
of Hydrobiology, 93(1), 1-30. Jagger, J., (1967) Introduction to Research in Ultraviolet Photobiology. Prentice-Hall, Inc. Englewood Cliffs, NJ Lee, W. H., Bishop, P. L. (2010). In situ microscale analyses of activated sludge flocs in the enhanced biological phosphate removal process by the use of microelectrodes and fluorescent in situ hybridization. Journal of Environmental Engineering, 136(6), 561-567. Li, D., & Ganczarczyk, J. J. (1987). Stroboscopic determination of settling velocity, size and porosity of activated sludge flocs. Water Research, 21(3), 257-262. Liss, S. N., Droppo, I. G., Flannigan, D. T., & Leppard, G. G. (1996). Floc architecture in wastewater and natural riverine systems. Environmental Science and Technology, 30(2), 680-686. Maness, P., Smolinski, S., Blake, D. M., Huang, Z., Wolfrum, E. J., & Jacoby, W. A. (1999). Bactericidal activity of photocatalytic TiO2 reaction: Toward an understanding of its killing mechanism. Applied and
Environmental Microbiology, 65(9), 4094-4098. Matsunaga, T., Tomoda, R., Nakajima, T., & Wake, H. (1985). Photoelectrochemical sterilization of microbial cells by semiconductor powders. FEMS Microbiology Letters, 29(1-2), 211-214. Pfeifer, G. P., You, Y., & Besaratinia, A. (2005). Mutations induced by ultraviolet light. Mutation Research - Fundamental and Molecular Mechanisms of Mutagenesis, 571(1-2 SPEC. ISS.), 19-31.
102
Piluso, L. G., & Moffatt-Smith, C. (2006). Disinfection using ultraviolet radiation as an antimicrobial agent: A review and synthesis of mechanisms and concerns. PDA Journal of Pharmaceutical Science and
Technology, 60(1), 1-16 Stowa, . (2002). Biological phosphorus removal. London: IWA Trivedi, H. K. (2009). Simultaneous nitrification and denitrification. In Advanced biological Treatment
Processes; Wang, L. K., Shammas, N. K., Hung, Y. T., Eds.; Humana Press: Totowa, NJ, pp. 186-189. Water Environment Federation., & Environmental and Water Resources Institute (U.S.). (2006). Biological
nutrient removal (BNR) operation in wastewater treatment plants. New York: McGraw-Hill. Wright, H.B., and Carins W.L. Ultraviolet Water Disinfection, USEPA Workshop on UV Disinfection of
Drinking Water, Arlington, VA, April 1999. Zhang, Q., Li, C., Li, T. (2012). Rapid Photocatalytic Degradation of Methylene Blue under High Photon Flux UV Irradiation: Characteristics and Comparison with Routine Low Photon Flux. International Journal
of Photoenergy, vol. 2012, Article ID 398787, 7 pages.
103
6. The Effect of Activated Sludge Process Conditions
on UV Disinfection: Temperature, Sludge Retention Time,
Influent Phosphorus Levels4
6.1. Abstract
The effect of sludge retention time (SRT), temperature and influent phosphorus levels on
the ultraviolet (UV) disinfection of effluent was investigated using a pilot-scale sequencing
batch reactor (SBR). Increasing the operating temperature from 12 ºC to 22 ºC improved
the tailing level of UV dose response curve (DRC) of the final effluent by 0.5 log. A
similar improvement was observed for the 75-90 µm flocs that were collected and
fractionated from the mixed liquor. The improvement in the floc disinfectability was
accompanied by an increase of 0.1-0.2 in the fractal dimension. Introducing a feed with
phosphorus limitation, i.e. COD:N:P of 100:10:0.3, caused the formation of rounder flocs
with improved settling characteristics, however these flocs were more resistant to UV
disinfection resulting in a higher tailing level of UV DRC (by 1 log). Starving phosphorus
conditions, i.e. COD:N:P of 100:10:0.03, decreased the average floc size and sphericity,
both by 50%, and improved the final effluent UV disinfection by at least 1 log, compared
to the normal and limited phosphorus conditions. The results also showed that the UV
disinfectability of the mixed liquor flocs and that of the final effluent did not change
4 This Chapter is based on the following article: Azimi. Y., Allen. D. G., Farnood. R., Seto. P.. The Effect of Activated Sludge Process Conditions on UV Disinfection: Temperature, Sludge Retention Time, Influent Phosphorous Levels. To be submitted
104
significantly when increasing the SRT from 7 to 20 days. Similarly, no clear trend was
observed between either of the flocs’ protein to carbohydrate ratio or the flocs’ mechanical
stability with UV disinfectability.
6.2. Introduction
Ultraviolet disinfection is a well-established, environmentally friendly technology
for effluent disinfection. However, the effectiveness of UV disinfection decreases in the
presence of suspended microbial flocs formed in the activated sludge process (Qualls et al.
1983, Das 2001). In a typical UV dose response curve (DRC) this decrease in the
efficiency is detected at higher UV doses, referred to as tailing.
The tailing phenomenon is generally attributed to the absorbance of UV light by
the floc material and hence shielding of the embedded microorganisms. The reduction of
UV transmittance within flocs is due to the presence of extracellular polymeric substances
(EPS), which are strong absorbers of UV light (Kalisvaart 2004, Farnood 2005). Hence,
UV disinfectability of flocs is expected to be a function of floc structure and composition.
Loge et al. (1999) suggested that the transmittance of UV light within a floc would
affect its UV disinfectability, and proposed the light pathways concept. They suggested that
the target organisms with a relatively unobstructed pathway to the bulk medium are
disinfected more easily. In Chapter 3 it was suggested that the formation of compact EPS
regions could create two types of flocs: 1) flocs with a homogeneous loose structure, and 2)
flocs with a double layer structure of a compact core surrounded by a loose shell.
Accordingly, they developed a structure-based model to describe the UV disinfection
kinetics of wastewater flocs.
105
It is expected that as floc compactness increases, the UV transmittance within the
floc decreases, and hence the effective dose delivered to embedded microorganisms within
the floc decreases. Floc compactness is expected to be related to its chemical and
mechanical stability (Liao et al. 2002, Yuan & Farnood 2010, Sheng et al. 2008). Flocs
generated at higher SRTs were found to have higher protein content, settle better, and have
a higher chemical stability (Liao et al. 2002, Xie and Yang 2009). Xie and Yang (2009)
reported that the loosely bound EPS content decreased as SRT increased and was the cause
of good settleability and low sludge volume index. However, it has been reported that
neither the EPS composition (protein, carbohydrate, and DNA content) nor the UV
disinfectability of the flocs was significantly affected by varying SRT from 4 to 20 days
(Scott et al. 2005). Xie and Yang (2009) as well as Li and Yang (2007) showed that
changing the carbon source from glucose to acetate had a significant effect on improving
sludge settleability. Li and Yang (2007) studied the effect of SRT ranging from 5 to 20
days on the loose, and tightly bound EPS content, and the settling properties of sludge flocs.
They found that sludge at lower SRTs had a higher loosely bound EPS content which
caused a higher sludge volume index. However, the tightly bound EPS content was
independent of the SRT in the range tested. Liss et al. (2002) investigated the effect of SRT
on microbial floc structure using various microscopy techniques and reported that flocs
formed at higher SRTs were more stable, more hydrophobic and less negatively charged.
Liao et al. (2002) examined the chemical stability of wastewater flocs and found that flocs
at higher SRTs are more stable, and have a dense core associated with their structure that
due to the limitations of diffusion of oxygen and nutrients has a growth rate lower than that
of the outer layer. In addition to the effect of SRT on floc structure, increase in SRT has
106
been reported to result in an exponential decrease in the fraction of flocs with associated
coliform (Loge et al. 2002).
Similarly, earlier studies have shown that the operating temperature of the
wastewater treatment process could have a significant impact on the floc composition and
structural characteristics. Liao et al. (2011) found that aerobic flocs generated under
mesophilic (35ºC) conditions settle better compared to those generated under thermophilic
(55ºC) conditions, due to a higher level of filaments. Morgan-Sagastume and Allen (2005)
reported that temperature shifts from 30˚ to 45 ˚C causes deflocculation and an increase in
the final effluent’s suspended solids. Tian et al. (1994) examined the effect of operating
temperature from 20 ˚C to 8 ˚C and suggested that under cold weather conditions the viable
organisms will comprise a smaller percentage of the total activated sludge. Wilén et al.
(2010) showed that flocs generated in the summer have more and larger microbial colonies
and correlated this observation to the metabolic activity differences which affects the
growth pattern of the microbial species.
The effect of phosphorus limitation on floc structure has been studied by Liu and
Liss (2007). The response of the activated sludge process to the phosphorus limited
conditions (COD:N:P, 100:5:1 switching to 100:5:0.3) was the production of flocs that
were larger, rounder, and more compact, which improved their settling characteristics. In a
more recent study, Dumitrache (2008) found that flocs generated at phosphorus limited
conditions (100:5:0.33) were larger with more densely packed biomass, observed as darker
clusters in optical microscopy images. Similar observations were made by Brei (2011).
Although the above literature suggest that wastewater treatment process conditions
could have a significant effect on the UV disinfectability of final effluent, very little
107
systematic study has been done to better quantify such effects. Therefore, the main
objective of this chapter was to compare the UV disinfectability of flocs generated under
various SRTs, temperatures, and nutrient (phosphorus) conditions in relation to floc
characteristics. Results from this study could provide a basis for a better and holistic
design and operation of wastewater treatment processes for improved disinfection
performance of final effluents.
6.3. Materials and Methods
6.3.1. Sample Collection
The samples used for this study were collected from two sets of pilot sequencing
batch reactors (SBR), operating at Environment Canada’s National Water Research
Institute (NWRI), located in Burlington, Canada. Schematics of the reactors are presented
in Figure 6. 1.
The first set of reactors consisted of four SBR systems, each having an operating
volume of 20L and using primary treated sewage from the Skyway Wastewater Treatment
Plant (WWTP) located in Burlington as the feed. In this chapter, these reactors are referred
to as the “large reactors”. The COD, BOD5, TKN and TP values for the influent to the
The morphological characteristics of the flocs generated in the SBR reactors are
listed in Table 6. 4. The results suggest that flocs formed at the lower operating
temperature of 12 ºC were smaller and more spherical compared to those generated at 22 ºC,
at both 7 days and 20 days SRT. Reactors operating at 22 ºC produced larger flocs with
higher fractal dimension and improved settleability characteristics; i.e. lower ESS and SVI
as seen in Table 6. 3. By increasing SRT from 7 days to 20 days at both 12 ºC and 22 º) the
average mixed liquor floc size increased.
116
In general, limiting or starving the SBR reactors on phosphorus caused the
formation of smaller flocs. Moreover, under the P-Starved conditions flocs were notably
less spherical and more irregular in shape (higher fractal dimension).
Table 6. 4. The effect of increasing SRT, T and changing feed phosphorus level on floc size, sphericity,
fractal dimension (D1 using Logan & Wilkinson 1991), protein to carbohydrate ratio, and mechanical
stability (τ 50)
Reactor
% Flocs with
Sphericity > 0.5
(35-120 µm)
% ML Flocs
> 50 µm
D1
(35-120 µm)
ML Protein:
Carbohydrate
τ 50*
SRT 20 T 12 20 28 1.25 2.5 ± 0.04 9.0 ± 0.4
SRT 7 T 12 15 17 1.3 1.8 ± 0.02 9.4 ± 0.3
SRT 20 T 22 8 22 1.4 1.9 ± 0.01 9.9 ± 0.5
SRT 7 T 22 11 23 1.5 1.5 ± 0.02 10.0 ± 0.6
P-Normal 12 27 1.36 3.2 ± 0.04 11.7 ± 0.5
P-Limited 14 20 1.3 1.0 ± 0.03 14.6 ± 0.3
P-Starved 5.5 12 1.4 3.7 ± 0.01 9.4 ± 0.4
*: Calculated for the 63-106 µm floc fraction in the large reactor and 45-75 µm floc fractions in the
small reactors.
Mechanical Stability
The breakage curves of flocs as a function of shear stress for the SBR reactors are
plotted in Figure 6. 2. The median shear stress (τ50
), i.e. the shear stress at which the
number of flocs reduced by 50 %, is reported for flocs of narrow size fractions in Table 6. 4.
Mechanically stronger flocs would have a higher τ50
, as a higher shear stress would be
required to break the flocs. Although there is significant variability in the data, increasing
the operating temperature from 12 to 22 oC caused a decrease in floc breakage, i.e.
produced somewhat stronger flocs. No clear relationship was found between SRT and the
flocs’ mechanical stability. Phosphorus limitation on the other hand, had a noticeable
effect on floc mechanical stability (Figure 6. 2-b). At similar shear stress levels, P-Limited
flocs showed a lower breakage tendency (i.e. higherτ50
), indicating that these flocs were
117
mechanically stronger. This is consistent with the previous studies that reported flocs
under phosphorus limitation condition tend to have a more compact and spherical structure,
with more densely populated biomass and better settleability characteristics (Liu & Liss
2007, Dumitrache 2008).
0
20
40
60
80
100
6.7 8.6 10.6 12.6
Turbulent Shear Stress (Pa)
%B
rea
ka
ge
C SRT20T12 (63-106 µm)
D SRT7T12 (63-106 µm)
B SRT20T22 (63-106 µm)
E SRT7T22 (63-106 µm)
(a)
0
20
40
60
80
6.7 8.6 10.6 12.6
Turbulent Shear Stress (Pa)
%B
reak
ag
e
P-Normal (45-75 µm)
P-Limted (45-75 µm)
P-Starved (45-75 µm)
(b)
Figure 6. 2. Percentage of breakage obtained at various degrees of shearing, for (a) the large (20 L)
reactors, and (b) for the small (2 L) reactors
118
Composition
The protein to carbohydrate ratio of the EPS samples extracted from the SBR
reactors are shown in Table 6. 4. The P-Limited flocs, which had significantly higher
mechanical strength, also showed a lower protein to carbohydrate ratio compared to the P-
Normal and P-Starved conditions. At both SRTs of 7 & 20 days, increasing the operating
temperature from 12 to 22 ºC decreased the protein to carbohydrate ratio and increased the
mechanical strength. The calculated protein to carbohydrate ratio is negatively correlated to
mechanical strength.
6.4.2. UV Disinfection of Mixed Liquor Flocs
As discussed in pervious chapters, the tailing of the UV DRC is controlled by the
presence of larger flocs in the effluent. Hence, it is expected that the physicochemical
characteristics of the flocs affect the level of tailing and the resistance of the effluent to UV
disinfection. For this reason, it is important to asses the UV disinfectability of isolated
flocs in relation to the wastewater treatment conditions.
The Effects of Temperature and Sludge Retention Time
The UV DRC of 32-53 µm and 75-90 µm flocs, collected from the large SBRs are
displayed in Figure 6. 3. The double exponential model parameters for these samples are
also provided in
Table 6. 5. For the 32-53 µm flocs (Figure 6. 3- a), regardless of variations in SRT
and temperature, the kinetics of UV disinfection were nearly the same. However, in the
case of 75-90 µm flocs (Figure 6. 3-b), increasing the operating temperature from 12 to 22
119
ºC at both SRTs resulted in a lower tailing level of the UV DRC by 0.5 log, while SRT had
no significant effect on the tailing level.
Higher operating temperatures are known to favor the growth of filamentous
bacteria over the floc-forming bacteria in flocs (Eckenfelder. 2000, Rossetti et al. 2002).
The presence of filamentous bacteria is known to give the flocs a more open structure (Liu.
2006, Tian & Su. 2012). Liu et al. (2005) showed that the abundance of filamentous
bacteria increased when increasing the operating temperature from 17 ºC to 25 ºC.
Furthermore, filamentous networks direct floc growth into shapes other than spherical
(Jenkins et al. 2004). The morphological data presented in Table 6. 4 also show that flocs
formed at higher temperatures were less spherical. Higher irregularity in floc surface (or
higher fractal dimension) is generally associated with higher porosity (Perez et al. 2006),
and would allow for better penetration of UV light. Elongated filamentous bacteria (that
could be hundreds of microns in size) are more likely to be embedded in larger flocs. This
could be the reason that temperature showed a noticeable effect on the UV disinfection of
the larger floc fraction (75-90 µm) but had no influence on the tailing level of the 32-53 µm
floc fraction.
120
0.0001
0.001
0.01
0.1
1
0 10 20 30 40 50 60
UV Dose (mJ/cm²)
N /
N 0
SRT 20 T 12 (32-53 µm)
SRT 7 T 12 (32-53 µm)
SRT 20 T 22 (32-53 µm)
SRT 7 T 22 (32-53 µm)
(a)
0.001
0.01
0.1
1
0 10 20 30 40 50 60
UV Dose (mJ/cm²)
N /
N 0
SRT 20 T 12 (75-90 µm)
SRT 7 T 12 (75-90 µm)
SRT 20 T 22 (75-90 µm)
SRT 7 T 22 (75-90 µm)
(b)
Figure 6. 3. The averaged UV DRC of (a) 32-53 µm and (b) 75-90 µm activated sludge flocs collected
from the SRT20T12, SRT7T12, SRT7T22, and SRT20T22 reactors. The lines represent the fitted
double exponential model.
The Effect of Influent Phosphorus Content
The averaged double exponential UV DRCs of 45-75 µm mixed liquor flocs,
collected from the small SBRs are displayed in Figure 6. 4. The flocs isolated from the P-
Limited reactor showed significantly higher resistance to UV disinfection. However, the P-
121
Starved flocs were easier to disinfect and showed a lower tailing level. The total
phosphorus in the mixed liquor (TP/MLSS) was estimated for the P-Normal, P-Limited,
and P-Starved reactors were 1.3, 0.34, and 0.02 mg/g, respectively. At lower feed
phosphorus content, there was less accumulation of phosphorus in the mixed liquor.
P-Starved flocs were more irregular in shape with poor settling characteristics
compared to the P-Limited and P-Normal flocs (Table 6. 4 &Table 6. 3). The UV DRC
also showed a lower tailing level by 1 log and 0.5 log compared to the P-Limited and P-
Normal flocs, respectively (Figure 6. 4). The lower tailing level of the P-Starved flocs
suggests that these flocs were less likely to contain viable coliform within their structure.
In fact it has been reported that coliform bacteria were less readily present under
phosphorus deficiency (An et al. 2002, Davies et al. 1995). Therefore, a lower percentage
of flocs grown under these conditions would harbor and shield coliform bacteria, and hence
they are less likely to be UV resistant (i.e. have lower tailing level).
0.00001
0.0001
0.001
0.01
0.1
1
0 10 20 30 40 50 60
UV Dose (mJ/cm²)
N /
N 0
P-Normal 45-75
P-Limited 45-75
P-Starved 45-75
Figure 6. 4. The averaged UV DRC of the 45-75 µm activated sludge flocs collected from the P-Normal,
P-Limited, and P-Starved reactors. The lines represent the fitted double exponential model
122
The UV DRC double exponential kinetic parameters are listed in Table 6. 5.
Higher values of β and lower values of k1 correspond to higher UV resistance. For a given
SBR system, the UV resistance (i.e. β) of flocs was positively correlated with the
percentage of flocs with a sphericity > 0.5. This correlation was particularly strong in the
case of larger reactors (r2= 0.99). Sphericity affects the path length of UV light within the
floc and hence could alter the average dose delivered to the embedded microorganisms
within the floc. More elongated (less spherical) flocs could provide a larger average dose
and hence would be less UV resistant. It appears that higher UV resistance occurred when
the percentage of particles with a sphericity > 0.5 increased for all of the SBR floc samples
(Figure 6. 5).
Table 6. 5. Double exponential UV disinfection kinetic parameters for the isolated mixed liquor flocs
from the large and small SBRs
Large Reactors (75-90 µm) k1 k2 β
SRT 20 T 12 0.14 ± 0.02 0.01 ± 0.005 0.21 ± 0.03
SRT 20 T 22 0.16 ± 0.04 0.01 ± 0.006 0.12 ± 0.03
SRT 7 T 12 0.15 ± 0.02 0.009 ± 0.005 0.18 ± 0.04
SRT 7 T 22 0.16 ± 0.03 0.012 ± 0.007 0.15 ± 0.02
Small Reactors (45-75 µm)
P-Normal 0.2 ± 0.05 0.02 ± 0.006 0.09 ± 0.02
P-Limited 0.12 ± 0.01 0.01 ± 0.006 0.06 ± 0.03
P-Starved 0.26 ± 0.01 0.02 ± 0.003 0.05 ± 0.07
123
0
0.05
0.1
0.15
0.2
0.25
0.3
8 11 15 20 5.5 12 14
% Flocs with Sphericity > 0.5
β o
r k
1
β
k1
Large Reactors Small Reactors
Figure 6. 5. UV DRC double exponential model parameters (β &k1) for mixed liquor flocs with
different percentages of flocs with sphericity larger than 0.5. Please note that the double exponential
parameters are reported for the 75-90 µm, and 45-75 µm flocs for the large and small reactors,
respectively. The percentage of flocs with sphericity larger than 0.5 is estimated for 35-120 µm flocs.
6.4.3. Final Effluent Quality and UV Disinfection
The Effects of Temperature and Sludge Retention Time
The final effluent quality parameters from all seven reactors are listed in Table 6. 6.
For the large reactors, operating at the higher temperature (22˚C) improved the overall
effluent quality (i.e. lower ESS, lower turbidity, and lower percentage of flocs larger than
50µm). There was no clear relationship between SRT and the effluent quality parameters.
For the small reactors, the P-Limited condition showed improved effluent characteristics
(i.e. lower turbidity, lower TSS and higher UVT) as well as improved settleability
compared to the P-Normal and P-Starved conditions. However, the quality of final effluent
worsened significantly in the P-Starved condition. It should be noted here that the
percentage of particles larger than 50 µm plays a significant role in the UV disinfectability
of the final effluent, as larger particles are harder to disinfect (Tan 2007, Yong et al. 2009,
124
Gibson et al. 2008). Effluents with a significantly lower percentage of particles larger than
50 µm included the P-Starved and the large reactors operating at 22 ºC. The P-Starved
effluent showed a higher turbidity and lower UVT compared to the P-Limited and P-
Normal conditions.
Table 6. 6. Effluent quality data for all seven pilot reactors
Reactor
Turbidity
(NTU)
ESS
(mg/L)
UVT% at
254nm
#Particle/m
L
% of
particles
> 30µm
% of
particle
s >
50µm
C (SRT20T12) 2.8 ± 1.3 9 ± 4 67 ± 3 108 ± 10 54 9
D (SRT7T12) 3.9 ± 0.2 5 ± 3 71 ± 3 102 ± 10 54 9
B (SRT20T22) 1.8 ± 0.2 4 ± 3 72 ±3 82 ± 10 50 4
E (SRT7T22) 1.4 ± 0.6 3 ± 2 75 ± 2 73 ± 10 57 5
P-normal 3.7 ± 0.3 24 ± 10 73 ± 4 110 64 17
P-limited 2.3 ± 0.3 10 ± 10 84 ± 3 87 58 13
P-starved 6.9 ± 0.6 17 ± 12 64 ± 4 100 53 6
The average UV DRCs of the secondary effluents collected from all reactors are
shown in Figure 6. 6 & Figure 6. 7. These DRCs are displayed both based on the
concentration of colony forming units (N) in CFU / 100 mL as well as the normalized
(N/N0) scale.
Based on Figure 6. 6, operating at higher temperature (22 ºC) improved the UV
disinfection performance of final effluent (i.e. lower CFU / 100mL at each UV dose).
Table 6. 6 shows that the secondary effluents collected from the 22 ºC reactors contained a
significantly lower percentage of particles smaller than 50 µm. According to Figure 6. 3-b,
larger flocs at higher operating temperatures were found to be more susceptible to UV
irradiation. Therefore, it is expected that the secondary effluents collected from the
reactors operating at 22 ºC would show an improved disinfectability. It should be
125
emphasized that the above effects are limited to temperatures examined in this study (i.e.
12 ºC and 22 ºC).
10
100
1000
10000
100000
1000000
0 20 40 60
UV Dose (mJ/cm²)
N
(CF
U /
10
0 m
L) B SRT 20 T 22
E SRT 7 T 22
10
100
1000
10000
100000
1000000
0 20 40 60
UV Dose (mJ/cm²)
N
(CF
U /
10
0 m
L) C SRT 20 T 12
D SRT 7 T 12
10
100
1000
10000
100000
1000000
0 20 40 60
UV Dose (mJ/cm²)
N (C
FU
/ 1
00
mL
) C SRT 20 T 12
B SRT 20 T 22
10
100
1000
10000
100000
1000000
0 20 40 60
UV Dose (mJ/cm²)
N
(CF
U /
10
0 m
L) D SRT 7 T 12
E SRT 7 T 22
a - the effect of SRT at high T b - the effect of SRT at low T
c - the effect of T at high SRT d - the effect of T at low SRT
10
100
1000
10000
100000
1000000
0 20 40 60
UV Dose (mJ/cm²)
N
(CF
U /
10
0 m
L) B SRT 20 T 22
E SRT 7 T 22
10
100
1000
10000
100000
1000000
0 20 40 60
UV Dose (mJ/cm²)
N
(CF
U /
10
0 m
L) C SRT 20 T 12
D SRT 7 T 12
10
100
1000
10000
100000
1000000
0 20 40 60
UV Dose (mJ/cm²)
N (C
FU
/ 1
00
mL
) C SRT 20 T 12
B SRT 20 T 22
10
100
1000
10000
100000
1000000
0 20 40 60
UV Dose (mJ/cm²)
N
(CF
U /
10
0 m
L) D SRT 7 T 12
E SRT 7 T 22
a - the effect of SRT at high T b - the effect of SRT at low T
c - the effect of T at high SRT d - the effect of T at low SRT
Figure 6. 6. The UV DRCs of treated effluents based on actual CFU / 100 mL.
The Effect of Influent Phosphorus Content
The UV DRCs of final effluents collected from P-Normal, P-Limited, and P-
Starved conditions are given in Figure 6. 7- a. This figure shows the P-Starved effluent
reached about 10 CFU/ 100 mL at a dose of 15 (mJ/cm2) whereas the P-limited and P-
normal effluents reached the same target at a UV dose of about 30 (mJ/cm2). At higher
126
doses, however, there was no significant difference among the UV DRC of the effluents
collected from the small reactors. In contrast, in their normalized form (Figure 6. 7-b), UV
DRC of the P-Starved effluent had a notably lower tailing level (about 1 log). The lower
tailing level of P-Starved effluent was likely a combination of two factors: 1) the P-starved
effluent contained a lower percentage of larger flocs (Table 6. 6), which are known to be
harder to disinfect, and 2) the P-Starved flocs were more susceptible to UV disinfection
compared to the P-Normal and P-Limited flocs as discusses in the previous section. In
contrast, the final effluent of the P-Limited reactor showed a higher tailing level implying
that although the effluent had a lower number of flocs, the flocs were harder to disinfect (in
accordance with the results in Figure 6. 3).
This information may be applied in the selection of upstream processes. For
example, it is known that the P-limited condition produces fewer flocs in the final effluent,
but with much higher UV resistance compared to the P-Starved and P-Normal conditions.
If filtration is the subject upstream process of interest, it would be more effective to apply
P-Limited conditions due to the presence of fewer particles in the effluent. Whereas, if
sonication or hydrodynamic shearing is the upstream process of interest, a process should
be selected that produces flocs with lower mechanical stability (P-Starved). However, in
the selection of upstream processes, the changes in UV dose demand should also be
accounted for.
127
0.1
10
1000
100000
10000000
0 20 40 60
Dose (mJ/cm²)
N (
CF
U /
10
0 m
L)
P-Normal
P-Limited
P-Starved
(a)
0.000001
0.0001
0.01
1
0 20 40 60
Dose (mJ/cm²)
N /
N 0
P-Normal
P-Limited
P-Starved
(b)
Figure 6. 7. UV DRCs (actual CFU/100ml-a, and normalized-b) of final effluents collected from the P-
Normal, P-Limited, and P-Starved conditions
6.4.4. UV Dose Demand of Final Effluent
In all of the UV DRCs presented so far in this study, the UV dose displayed on the
x axis was calculated based on the UV transmittance, and hence the effect of effluent
quality (turbidity, UVT) on UV dose was eliminated. However, for design purposes, it is
128
necessary to determine the total dose demand (i.e. energy demand) that includes the effect
of effluent quality parameters. To observe the effect of effluent quality in the UV DRC,
the fecal coliform survival ratio must be plotted against UV energy delivered to the sample.
The effluents’ UV dose demand for the large and small SBR reactors are displayed in
Figure 6. 8. UV dose demand in this section is defined as the time for achieving 3 log
reduction for the large reactors and 4 log reduction for the small reactors. The results
showed that for 3 log reduction of the large reactors’ effluents, SRT20T22 condition
needed almost half the UV dose than that of the reactors operating at lower temperature, or
lower SRT. The P-Starved condition effluent also showed a significantly lower (less than
half) UV dose demand to reach 4 log reduction compared to the P-limited and P-Normal
conditions.
In Figure 6. 8, the P-Starved effluent had lower effluent quality parameters
compared to the other small reactors, i.e. lower UVT and higher turbidity. However, the P-
Starved effluent had a much lower UV dose demand due to much higher susceptibility of
the flocs. This represents a case in which the conventionally measured effluent quality
parameters do not correlate with the UV resistance of the effluent. These results further
confirm that operating an activated sludge system at 22 ºC compared to 12 ºC, and P-
Starved compared to P-Limited and P-Normal produced effluents with significantly lower
UV dose demand.
129
10
20
30
40
50
SRT20
T12
SRT7
T12
SRT20
T22
SRT7
T22
P-
Normal
P-
Limited
P-
Starved
UV
Do
se D
em
an
d (
mJ /
cm
²)
3 log Reduction 4 log Reduction
Figure 6. 8. UV dose demand of the final effluents collected from all SBRs.
6.5. Conclusions
In this study the effects of SRT, temperature, and influent phosphorus levels on UV
disinfection kinetics were investigated. All reactors operated aerobically, and the results
showed that in general, more irregularly shaped flocs with lower sphericity and higher
fractal dimension were susceptible to UV disinfection. Increasing the operating temperature
from 12 ºC to 22 ºC, at both 7 & 20 days SRT, induced the formation of less spherical flocs
with lower resistance to UV disinfection (0.5 log lower tailing for the 75-90 µm flocs).
The 22 ºC final effluent showed 40% reduction in the number of particles larger than 50
µm compared to that of the 12 ºC reactors. Varying the SRT did not significantly change
the UV dose response of the sludge flocs and secondary effluents. Mechanical strength was
negatively correlated to UV resistance when varying temperature and SRT, but no
130
correlation was found when changing the influent phosphorus levels. P-Limited condition
(COD:N:P of 100:10:0.3) causes the formation of more spherical flocs with better
settleability and higher UV resistance compared to the P-Normal and P-Starved conditions.
However, P-Starved condition promotes the formation of small and irregularly shaped flocs
with poor settling characteristics. Despite having a lower effluent quality (high ESS and
high turbidity), the P-Starved condition showed improved UV disinfection performance, i.e.
less than half the UV dose required for 4 log removal of fecal coliform, compared to the P-
Limited and P-Normal effluents.
6.6. References
American Public Health Association (APHA) (2001), Standard methods for the examination of water and wastewater. 22nd Ed. Washington DC, USA. An, Youn-Joo, Donald H. Kampbell, and G. Peter Breidenbach. (2002). Escherichia coli and total coliforms in water and sediments at lake marinas. Environmental Pollution 120, 771-778. Bolton, J. R., Linden, K. G. (2003). Standardization of methods for fluence (UV Dose) determination in bench-scale UV experiments. Journal of Environmental Engineering, 129 (3), pp. 209 – 215. Bellouti, M., Alves, M. M., Novals, J. M., & Mota, M. (1997). Flocs vs granules: Differentiation by fractal dimension. Water Research, 31(5), 1227-1231. Brei, E. (2011). Bacterial Adhesin Proteins Associated with Microbial Flocs and EPS in Activated Sludge. Thesis (Ph.D)- University of Toronto Das, T. K. (2001). Ultraviolet disinfection application to a wastewater treatment plant. Clean Products and Processes, 3(2, pp. 69-80), August. Davies, Cheryl M., Julian A. H. Long, Margaret Donald, and Nicholas J. Ashbolt. 1995. Survival of fecal microorganisms in marine and freshwater sediments. Applied and Environmental Microbiology 61, 1888-1896. Dumitrache, R. G. (2008). Characterization of a hybrid sequencing batch reactor system for treatment of wastewater using suspended microbial aggregates and biofilms. Thesis (M.A. Sc.)- Ryerson University Eckenfelder WW. Industrial water pollution control. Singapore: McGraw-Hill; 2000. Farnood, R., Flocs and ultraviolet disinfection. Chapter 18 in Flocculation in Natural and Engineering
Systems. I.G. Droppo, G.G. Leppard, S.N. Liss, and T.G. Milligan, Eds., CRC Press, Boca Raton, FL, 2005, pp. 385-395.
131
Frølund, B., Palmgren, R., Keiding, K., Nielsen, P. H. (1996). Extraction of extracellular polymers from activated sludge using a cation exchange resin. Water Research, 30(8), 1749-1758. Gibson, J. H., Hon, H., Farnood, R., Droppo, I. G., Seto, P. (2009). Effects of ultrasound on suspended particles in municipal wastewater. Water Research, 43(8), 2251-2259. Haff, A. C. (1978). Mechanized micro-scale determination of protein in platelet pellet sonicates. Clinical
Chemistry, 24(11), 2031-2032. Jenkins, D., Richard, M. G., Daigger, G. T., & Jenkins, D. (2004). Manual on the causes and control of activated sludge bulking, foaming, and other solids separation problems. Boca Raton, Fla: Lewis Publishers. Kalisvaart, B. F. (2004). Re-use of Wastewater: Preventing the Recovery of Pathogens by using Medium-pressure UV lamp technology. Water Science and Technology, 50(6), 337-344. Li, X. Y., Yang, S. F. (2007). Influence of loosely bound extracellular polymeric substances (EPS) on the flocculation, sedimentation and dewaterability of activated sludge. Water Research, 41(5), 1022-1030. Liss, S. N., Liao, B. Q., Droppo, I. G., Allen, D. G., & Leppard, G. G. (2002). Effect of solids retention time on floc structure. Water Science and Technology, 50(3), 431-438. Liao, B. Q., Allen, D. G., Leppard, G. G., Droppo, I. G., Liss, S. N. (2002). Interparticle interactions affecting the stability of sludge flocs. Journal of Colloid and Interface Science, 249(2), 372-380. Liao, B. Q., Lin, H. J., Langevin, S. P., Gao, W. J., & Leppard, G. G. (2011). Effects of temperature and dissolved oxygen on sludge properties and their role in bioflocculation and settling. Water Research, 45(2), 509-520. Liss, S. N., Liao, B. Q., Droppo, I. G., Allen, D. G., & Leppard, G. G. (2002). Effect of solids retention time on floc structure. Water Science and Technology, 50(3), 431-438. Liu, Y., Liu, Q. (2006). Causes and control of filamentous growth in aerobic granular sludge sequencing batch reactors. Biotechnology Advances, 24(1), 115-127. Liu, J. R., & Liss, S. N. (2007). The impact of reduced phosphorus levels on microbial floc properties during biological treatment of pulp and paper wastewaters. Water Science and Technology, 55(6), 73-79. Logan, B. E., & Kilps, J. R. (1995). Fractal dimensions of aggregates formed in different fluid mechanical environments. Water Research, 29(2), 443-453. Loge, F. J., Emerick, R. W., Thompson, D. E., Nelson, D. C., & Darby, J. L. (1999). Factors influencing ultraviolet disinfection performance part I: Light penetration to wastewater particles. Water Environment
Research, 71(3), 377-381. Loge, F. J., Emerick, R. W., Ginn T. R., Darby, J. L. (2002). Association of coliform bacteria with wastewater particles: Impact of operational parameters of the activated sludge process. Water Research, 36(1), 41-48. Masuko, T., Minami, A., Iwasaki, N., Majima, T., Nishimura, S., Lee, Y. C. (2005). Carbohydrate analysis by a phenol-sulfuric acid method in microplate format. Analytical Biochemistry, 339(1), 69-72. Morgan-Sagastume, F., Allen, D. G. (2005). Activated sludge deflocculation under temperature upshifts from 30 to 45°C. Water Research, 39(6), 1061-1074. Perez, Y. G., Leite, S. G. F., & Coelho, M. A. Z. (2006). Activated sludge morphology characterization through an image analysis procedure. Brazilian Journal of Chemical Engineering, 23(3), 319-330
132
Qualls, R. G., Flynn, M. P., Johnson, J. D. (1983). The role of suspended particles in ultraviolet disinfection. Journal of the Water Pollution Control Federation, 55(10), 1280-1285. Rossetti, S., Tomei, M. C., Levantesi, C., Ramadori, R., & Tandoi, V. (2002). "Microthrix parvicella": A new approach for kinetic and physiological characterization. Water Science and Technology, 46 (1-2), 65-72. Scott, H. E., Liss, S. N., Farnood, R. R., Allen, D. G. (2005). Ultraviolet disinfection of sequencing batch reactor effluent: A study of physicochemical properties of microbial floc and disinfection performance. Journal of Environmental Engineering and Science, 4(SUPPL. 1), S65-S74. Sheng, G., Yu, H., Li, X. (2008). Stability of sludge flocs under shear conditions. Biochemical Engineering
Journal, 38(3), 302-308. Tan, T. (2007). Understanding the effect of particle-size on UV disinfection: Kinetics, mechanism and modeling. M.A.Sc. Dissertation # 134, University of Toronto. Tian, S., Lishman, L., & Murphy, K. L. (1994). Investigations into excess activated sludge accumulation at low temperatures. Water Research, 28(3), 501-509. Tian, Y., Su, X. (2012). Relation between the stability of activated sludge flocs and membrane fouling in MBR: Under different SRTs. Bioresource Technology, 118, 477-482. Wang. Y. (2012). Investigation of Biofloc Composition and its Influence on UV Disinfection of Wastewater. B.A.Sc. Dissertation, University of Toronto Wilén, B. M., Lumley, D., Mattsson, A., & Mino, T. (2010). Dynamics in flocculation and settling properties studied at a full-scale activated sludge plant. Water Environment Research, 82(2), 155-168. Xie, B., Yang, S. (2009). Analyses of bioflocculation and bacterial communities in sequencing batch reactors. Environmental Engineering Science, 26(3), 481-487. Yong, H. N., Farnood, R. R., Cairns, W., Mao, T. (2009). Effect of sonication on UV disinfectability of primary effluents. Water Environment Research, 81(7), 695-701. Yuan, Y., & Farnood, R. R. (2010). Strength and breakage of activated sludge flocs. Powder Technology,
199(2), 111-119.
133
7. Overall Discussion & Conclusions
This research provides a comprehensive understanding on the effects of wastewater
flocs’ structure and composition on UV disinfection kinetics. The objectives of this work
were to investigate the causes of tailing, and to evaluate the effect of secondary biological
treatment conditions on flocs physicochemical characteristics and UV disinfection kinetics.
Figure 7. 1 shows a summary of the project plan.
The three significant contributions of this study are listed below:
(1) The formation of compact cores in wastewater flocs significantly increases their
resistance to UV disinfection. A conceptual model for UV disinfection of flocs was
proposed that takes into account the floc structure;
(2) The flocs’ chemical composition could significantly affect their UV disinfection
kinetics. Finding of this study suggest that the presence of naturally occurring
polyphosphates in flocs collected from biological removal processes was likely inducing
additional UV disinfection of this type of effluent. This effect was attributed to the
photoreactive release of oxidative species that could results in cell lysis and accelerate UV
disinfection. On this basis, an invention disclosure was filed with the University of
Toronto (Disclosure # 10002579, Confidential Intellectual Property Center at the
University of Toronto);
(3) The effect of activated sludge process operational temperature, sludge retention
time, and phosphorus levels on floc structure and UV disinfection was investigated.
These topics are discussed in more detail in the following sections.
134
Figure 7. 1. Summary of thesis project plan
7.1. Identification of Compact Cores as UV Resistant Constituents
This study showed that activated sludge flocs may contain mechanically strong and
compact cores that exhibit significantly higher resistance (i.e. higher tailing level in DRC)
to UV disinfection. The volume and number of cores and their viability increased as floc
size increased. Therefore larger flocs were more likely to be resistance to UV (larger β)
and showed higher tailing level in the UV DRCs, as indicated in Chapter 3. The worsening
effect that the compactness of microbial aggregates has on UV disinfection and increasing
the UV dose demand (measured by comparing DRCs of flocs and cores of similar size), is
comparable to the worsening effect of increasing the average floc size by 20 µm.
135
Based on a novel structural model for the UV disinfection of floc, it was found that
floc disinfection kinetics was sensitive to the relative volume, the density (UV absorbance),
and the viability of dense core. However, the exact mechanisms and conditions that control
the formation of such cores are not widely understood, and additional work is required.
Gaining such fundamental understanding would be an important step towards designing
more effective wastewater treatment processes.
7.2. The Potential of Polyphosphates as Photooxidative Agents in
Sludge
This study showed that flocs harvested from BNR-UCT process exhibited improved
UV disinfection kinetics compared to the conventional activated sludge. It was suggested
that the origin of such differences was due to the chemical composition and structure of
flocs under BNR conditions. This finding suggests that through the holistic design of
wastewater treatment facilities it is possible to manipulate the physicochemical properties
of activated sludge flocs and improve their UV disinfection kinetics. Manipulating the
physicochemical properties of sludge flocs by integrating enhanced biological nutrient
removal processes improves the UV disinfection performance. The main contribution was
reporting evidence of in-situ advanced oxidation when disinfecting biological phosphorus
removal sludge with UV light, due to the photoreactive properties of polyphosphates in
releasing oxidative species under UV exposure.
The causes of higher UV susceptibility of the BNR-UCT flocs were proposed to be:
(1) the release of oxidative hydroxyl radicals within the flocs due to the photoreactive
effect of polyphosphates (embedded in BNR flocs), which can cause cell lysis; (2) the
release of micro-bubbles within the BNR-UCT flocs in the anaerobic and anoxic regions
136
gives them a more open and homogeneous structure, and limits the formation of compact
EPS zones, which further protect fecal coliform from exposure to UV.
Mathematical modeling showed that the UV disinfectability of flocs, as expected,
was quite sensitive to the generation of •OH within flocs. Formation of hydroxyl radicals
could improve the disinfection rate and significantly lower the tailing level of UV DRC.
7.3. Reducing Wastewater Flocs’ UV Resistance by Changing the
Secondary Treatment Operational Conditions
This study showed that the UV dose demand of final effluent, which is related to
the effluent quality and sludge flocs’ UV susceptibility, can be significantly altered by
changing the operational conditions of the activated sludge system. A strong correlation
was found between sphericity and UV resistance. As sphericity decreases, flocs are more
susceptible to UV disinfection. This finding was consistent among all the samples tested
over the course of this study. Compact cores had higher resistance to UV and were more
spherical compared to flocs; the BNR-UCT flocs were less spherical and had lower UV
resistance compared to CAS flocs; the flocs formed at 22 ºC were less spherical and more
susceptible to UV compared to those formed at 12 ºC; and the P-Staved flocs were less
spherical and more UV susceptible compared to the P-Limited and P-Normal flocs.
Moreover, it was found that compared to UVT, the percentage of flocs larger than
50 µm in the effluent correlated better to the UV disinfection performance. Lower UV
dose demand was estimated for two effluents, one with a lower turbidity (22 ºC) and the
other with higher turbidity (P-Starved). However, both effluents with low UV dose
137
demand had a lower percentage of flocs larger than 50 µm compared to the other
conditions.
Mechanical strength did not show a consistent trend with UV disinfection. The
cores were mechanically stronger than the flocs and showed higher UV resistance
compared to flocs; the BNR-UCT flocs were mechanically stronger with higher UV
susceptibility; the flocs produced at 22 ºC appeared to be mechanically stronger and
showed a lower UV resistance; and the P-Starved flocs which were mechanically less
stable showed high susceptibility to UV disinfection. Therefore, no conclusions can be
drawn regarding the effect of mechanical strength on UV resistance. However, the protein
to carbohydrate ratio has a strong negative correlation with mechanical strength. In Chapter
3, it was found that the mechanical strength of the compact cores was higher than that of
the flocs, and it was found that they contained a lower protein to carbohydrate ratio (Wang
2012). Moreover, the BNR-UCT flocs that were mechanically more stable had a lower
protein to carbohydrate ratio (see Chapter 4) than the CAS flocs. In Chapter 6, the flocs
formed at 22 ºC or at the P-Starved condition were stronger with lower protein to
carbohydrate ratio.
The addition of upstream processes that cause an increase in floc porosity or break
larger flocs would increase UV disinfection efficiency. Breaking the flocs by sonication or
hydrodynamic shearing prior to UV disinfection are suggested methods to improve the
flocs disinfectability. Floc breakage causes a size distribution shift towards smaller flocs,
while increasing the number of flocs. This increases effluent turbidity and decreases UV
transmittance, which adversely affect UV disinfection performance. Therefore, it is
suggested that processes that reduce sphericity and increase the surface roughness of the
138
microbial aggregates without breaking them would be more effective in improving their
UV disinfectability.
7.4. Conclusions
Considering the conducted research in this thesis, the following conclusions can be
made:
1- The formation of compact cores within the flocs contributes significantly to
their UV resistance. The compact cores were isolated from the flocs by shearing and
sieving, which peels off the loose outer layer. When the UV DRCs of flocs and cores of
similar size were compared, the tailing level was 0.5 log higher for the cores.
2- Enhanced biological nitrogen/phosphorus removal increases the UV
susceptibility of the sludge flocs. Denitrification was suggested to give the flocs a more
porous structure due to the internal release of nitrogen bubbles, hence increasing UV
transmittance within the flocs. Moreover, the accumulation of polyphosphate as a result of
EBPR causes significant changes in the mechanism of UV disinfection. In addition to
direct DNA damage from UV light, the photoreactive effect of polyphosphate causes the
release of •OH, which ruptures cell membrane and causes death. The UV DRCs showed
that at a UV dose of 15 mJ/cm2 the BNR-UCT final effluent shows 4 log disinfection,
while the CAS final effluent showed 2 log disinfection. At the UV dose of 60 mJ/cm2 the
BNR-UCT final effluent shows a lower tailing level than the CAS (1 log).
3- Polyphosphates, which occur naturally in sludge during the enhanced
when exposed to UV light. For the first time it has been shown in this study that
139
photoactive disinfection occurs within the flocs, not by adding a photocatalyst, but merely
by changing the type of wastewater treatment process and the accumulation of
polyphosphates in the sludge.
4- Floc sphericity is directly correlated to UV resistance. It was consistently
reported in this study that where flocs are less spherical, the embedded fecal coliform had a
lower chance of surviving UV disinfection.
5- The UV dose demand is dependent on secondary process conditions. Among
the conditions studied, two of them significantly reduced the UV dose demand of
secondary effluents: 1) the BNR-UCT process compared to conventional activated sludge
systems; and 2) conventional activated sludge system operating at COD:N:P of 100:10:0.03
(i.e. P-Starved condition) compared to 100:10:1 (i.e. P-Normal condition).
6- P-Limited conditions induce the formation of rounder and more compact
flocs with higher UV resistance, and higher mechanical stability. The P-Limited
condition appears to also improve the effluent characteristics such as decreasing turbidity,
decreasing ESS, and increasing the UVT. However, the flocs were more compact and
showed higher resistance to UV disinfection.
7- P-Starved conditions cause the formation of irregular flocs with high UV
susceptibility. The 45-75 µm mixed liquor flocs generated under P-Starved conditions
show improved UV disinfectability compared to the P-Limited (2 log improvement) and
the P-Normal (1 log improvement) flocs of similar size. The UV Dose demand of the P-
Starved effluent was also significantly lower than that of the P-Limited and P-Normal
conditions.
140
8- Operating a conventional activated sludge system at 22 ºC, compared to 12
ºC, decreases floc sphericity and the UV resistance of the final effluent. Flocs formed
at 22 ºC are more irregular in shape, and show improved settling properties compared to
those formed at 12 ºC (at both SRTs of 7 and 20 days). The final effluent collected from
the 22 ºC reactor had lower ESS, lower turbidity, higher UVT, and a lower percentage of
particles larger than 50 µm compared to the one collected from the 12 ºC reactor.
Therefore, UV disinfection improves when operating conventional activated sludge
systems at higher temperature (comparing 12 & 22 ºC). It should be noted here that the
above effects are limited to temperatures examined in this study (i.e. 12 ºC and 22 ºC). It is
known that higher operating temperatures (e.g. above 35 ºC) result in deflocculation, which
decreases secondary effluent quality. Therefore, the UV disinfection kinetics of flocs
formed at temperatures higher than 22 ºC remains unknown.
9- For the final effluents, the percentage of particles larger than 50% seems to
be a more relevant parameter to UV resistance compared to UVT and turbidity. In all
of the conditions studied, it appears that the effluent’s percentage of particles larger than 50
µm is more consistently related parameter to UV resistance. From the two operating
conditions that had the lowest UV dose demand, one had a comparatively low turbidity and
high UVT and the other had a higher turbidity and lower UVT.
10- Changing the sludge retention time from 7 to 20 days does not significantly
affect the UV disinfectability of the mixed liquor flocs and final effluents. Increasing
the sludge retention time slightly increases the mixed liquor size distribution, but no impact
on UV disinfectability was observed.
141
7.5. Recommendations for Future Work
1- In this study, the enhanced biological nutrient removal process was found to
generate flocs with higher susceptibility to UV disinfection. It was suggested that a
combination of polyphosphate’s photoreactivity and the increase in floc porosity by the
release of nitrogen gas was the origin of higher UV susceptibility. Polyphosphates are
produced in the enhanced biological phosphorus removal process, and the release of
nitrogen gas occurs during the denitrification (part of the enhanced biological nitrogen
removal) process. However, the effect of each of the enhanced biological
nitrogen/phosphorus removal on floc properties and UV disinfection remains unknown.
Assessing the effects of each process on its own would clarify the contribution of porosity
increase, and the presence of polyphosphates on UV disinfectability.
2- Studying the effect of mixing a pure culture of polyphosphate accumulating
organisms such as Accumulibacter Phosphatis on the UV disinfection kinetics of non-
particle associated fecal coliform. In this study it was shown that polyphosphate
accumulating organisms have a significant effect on UV disinfection mechanism, due to the
photoreactive effect of polyphosphates and the release of oxidative species. Moreover, it
would be helpful to systematically study the effect of polyphosphate accumulation
(different amounts of accumulated polyphosphate) on the UV disinfectability of e.coli.
3- Assessing the photoreactivity of polyhydroxylalkenoates (PHAs). In this study
the photoreactivity of polyphosphate was assessed. However, the BNR-UCT sludge also
142
contains PHA in large quantities and they may also be contributing to the accelerated
disinfection in the BNR-UCT flocs. More specifically, the photoreactivity of
polyhydroxybutyrate (PHB) should be studied.
4- Assessing the effect of nitrogen limitation on the amount of extracellular
polymeric substances and proteins, effluent quality and UV disinfectability. In this study
the effect of phosphorus limitation on some floc physicochemical characteristics was found,
and the UV disinfectability of mixed liquor flocs and final effluents were compared for
various degrees of phosphorus limitation.
5- Understanding the effect of feeding pattern on the flocs structural properties,
viability and UV disinfection kinetics. Throughout this study, although not reported,
observations were made that feeding pattern can significantly impact the mechanical
strength and viability-culturability of the microbial flocs. It is suspected that the feeding
pattern would change the response of microbial aggregates to UV disinfection.
6- In this study, composition was a general protein to carbohydrate ratio
measurement. However, further characterization of the type of sugars and proteins in the
flocs, their UV absorbance and contribution to floc UV resistance is needed.
7- UV disinfection performance in the industry and in this research is assessed
using the membrane filtration method. In this method, if a floc is cultured and produces
one colony, this is independent of whether the floc has one or a thousand viable-culturable
143
microorganisms. It would be of value to develop methods to assess the total number of
fecal coliform in the flocs.
144
8. Appendices
8.1. Appendix 1: Diffusion-Reaction Equation of OH• in the
Polyphosphate Accumulating cells
The diffusion and reaction of hydroxyl radical produced at the surface of inter-
cellular polyphosphate in the polyphosphate accumulating organisms is considered the
problem of Fickian diffusion including a first order irreversible reaction. The system can
be described as the following in one-dimensional coordinates:
x
CCk
x
CD r
∂
∂=−
∂
∂2
2
The solution to this problem can be found in Basmadjian & Farnood (2010) as
following:
−+
−−= tk
tD
xerfcDkxtk
tD
xerfcDkx
C
Crrrr
2)/exp(
2
1
2)/exp(
2
1
0
where C0 is the initial concentration of OH•, and C is the concentration of OH• at distance
x from the polyphosphate granule, at time t. D is the diffusion coefficient of OH• in the
cytoplasm, and kr is and the reaction rate constant of OH• with cell material.
The value of D of OH• in water is estimated by Campo & Grigera (2005) as 7.1E-9
m2s-1. The value of kr is extracted from Liu et al. (2007), where they reported the
degradation rate of proteins and carbohydrates in the presence of hydroxyl radicals as
1.25E-4 s-1. Substituting the values of D and kr and t as 60 seconds (a smaller value than
145
the exposure time for the lowest UV dose), the distance for a 1 log reduction of OH• can be
estimated as 100 µm.
8.1.1. References
Basmadjian, D., Farnood, R., & Basmadjian, D. (2007). The art of modeling in science and engineering with Mathematica. Boca Raton: Chapman & Hall/CRC. Campo, M. G., & Grigera, J. R. (2005). Classical molecular-dynamics simulation of the hydroxyl radical in water. The Journal of Chemical Physics, 123(8), 084507. Liu, Y., Li, J., Qiu, X., Burda, C. (2007). Bactericidal activity of nitrogen-doped metal oxide nanocatalysts and the influence of bacterial extracellular polymeric substances (EPS). Journal of Photochemistry and