Monitoring of effluent DOM biodegradation using fluorescence, UV and DOC measurements.

t D
an
Ibrahim Saadi a,b, Mikhail Borisover c, Robert Armon b, Yael Laor a,*
was observed thus proposing (i) the formation of new fluorescing material associated with DOM biodegradation and/or
(ii) degradation of certain organic components capable of quenching DOM fluorescence. Based on the ratio between
tries in which the available resources of fresh water are higher concentrations of organic and inorganic constitu-
ents, the concentrations of which depend on the quality
and type of sewage and extent of treatment.
Dissolved organic matter (DOM) commonly ex-
pressed by its C content (DOC) is an important compo-
0045-6535/$ - see front matter � 2005 Elsevier Ltd. All rights reserved.
* Corresponding author. Tel.: +972 4 9539539; fax: +972 4
9836936.
E-mail address: laor@volcani.agri.gov.il (Y. Laor).
Chemosphere 63 (2006)fluorescence intensity and UV254, different biodegradation dynamics for fluorescent DOM constituents as compared
with other UV-absorbing molecules was delineated. Overall, about 50% of the total DOM was found to be readily
degradable such that residual resistant DOC levels were between 8 and 10 mg l�1. Enhanced levels of residual DOM
in effluent-irrigated soils may contribute to the DOM pool capable of carrying pollutants to groundwater.
� 2005 Elsevier Ltd. All rights reserved.
Keywords: Effluents; Dissolved organic matter (DOM); Excitation emission matrix (EEM); Biodegradation; Qishon (Kishon)
1. Introduction
The use of treated wastewater effluents to irrigate
agricultural land is a common practice in many coun-
limited. In Israel, a wide national program of waste-
water reclamation is being established (Rebhun et al.,
1987; Amiel et al., 1990; Fine et al., 2002). As compared
with freshwater, effluents used for irrigation containa Agricultural Research Organization, Institute of Soil, Water and Environmental Sciences, Newe Ya’ar Research Center,
P.O. Box 1021, Ramat Yishay 30095, Israel
b Technion-Israel Institute of Technology, Faculty of Civil and Environmental Engineering, Haifa 32000, Israel
c Agricultural Research Organization, Institute of Soil, Water and Environmental Sciences, Volcani Center,
Bet Dagan 50250, Israel
Received 22 April 2005; received in revised form 20 July 2005; accepted 31 July 2005
Available online 10 October 2005
Abstract
The potential of effluent DOM to undergo microbial degradation was assessed in batch experiments. Effluent sam-
ples from Haifa wastewater treatment plant and Qishon reservoir (Greater Haifa wastewater reclamation complex,
Israel) were incubated either with effluent or soil microorganisms for a period of 2–4 months and were characterized
by dissolved organic carbon contents (DOC), UV254 absorbance and by fluorescence excitation–emission matrices.
Three main fluorescence peaks were identified that can be attributed to humic/fulvic components and ‘‘protein-like’’
structures. During biodegradation, specific fluorescences (F/DOC) of the three peaks were increased at various extents,
suggesting selective degradation of non-fluorescing constituents. In some cases increase in the effluent fluorescence (F)Monitoring of effluen
using fluorescence, UVdoi:10.1016/j.chemosphere.2005.07.075OM biodegradation
d DOC measurements
530–539
www.elsevier.com/locate/chemosphere

I. Saadi et al. / Chemosphere 63 (2006) 530–539 531nent in secondary water resources. In addition to natural
soil DOM that is originated from plant litter, soil
humus, microbial biomass and root exudates (Kalbitz
et al., 2000), effluents may become a considerable source
of DOM in effluent-irrigated soils. Effluent DOM origi-
nated from biological wastewater treatment schemes
may include organic compounds of different groups,
from carbohydrates and proteins to more biologically
resistant components known as fulvic and humic materi-
als (Rebhun and Manka, 1971; Manka et al., 1974;
Ma et al., 2001; Imai et al., 2002; Ilani et al., 2005).
DOM in soil is affected by sorption as well as microbial
degradation whereas the magnitude of such processes is
strongly related to soil properties and the chemical com-
position of the DOM. Biodegradation of natural soil
DOM has been investigated in detail (Kalbitz et al.,
2000, 2003; Marschner and Kalbitz, 2003). Much less
work has been devoted to the biodegradation potential
of effluent DOM in effluent-irrigated soils. Proper moni-
toring and characterization of residual DOM in treated
effluents is essential for estimating the potential of
DOM to enhance contaminant transport. Of special con-
cern are dissolved aromatic humic and fulvic-type mole-
cules which are considered relatively stable against
microbial degradation, on one hand, andmay be of great-
er potential to bind organic compounds on the other
hand. However, due to the complexity of effluent
DOM, it is highly complicated to delineate the behavior
of individual constituents or groups of components in
the overall DOM dynamics.
Fluorescence spectroscopy has been utilized to probe
the chemical composition of DOM because of its exper-
imental simplicity and ability to distinguish between cer-
tain classes of organic matter. It is well known that
fulvic acids, humic acids and proteins (due to the pres-
ence of phenylalanine, tyrosine and triptophane) are
fluorophores (Senesi et al., 1991; Yamashita and
Tanoue, 2003). Recent advances in fluorescence spec-
troscopy ease the generation of high resolution three-
dimensional (3-D) excitation–emission matrices (EEM;
e.g. Marhaba et al., 2000; Baker, 2001; Baker and
Lamont-Black, 2001; Chen et al., 2003). There are multi-
ple assignments in the literature tempting to relate peak
locations in EEM to the nature of DOM components.
For example, the classification reported by Leenheer
and Croue´ (2003) suggests that fluorescence at 420–
480/330–350 and 380–480/250–260 nm (kem/kex) is
associated with humic-like material, and fluorescence
at 300–350/270–280 is related to protein-like structures.
Importantly, DOM fluorescence may be affected by
solution properties such as pH and ionic strength. It
can be affected also by solution interactions between
DOM molecules, or between DOM and organic or inor-
ganic co-species, both in the ground and excited states.
Such possible effects should be carefully considered
during fluorescence data interpretation. Nonetheless,the experimental simplicity of fluorescence spectroscopy
makes it a useful tool for DOM fingerprinting and for
on-line monitoring of natural and engineered water
systems. For these reasons the fluorescence mapping ap-
proach has been widely used to characterize and moni-
tor DOM in various ecosystems (e.g. marine, Coble,
1996; groundwater, Baker and Lamont-Black, 2001;
sewage-impacted rivers, Baker, 2001; wastewater and
effluents, Esparza-Soto and Westerhoff, 2001; Wester-
hoff et al., 2001; Her et al., 2003). Fluorescence spectros-
copy was also used to evaluate the ‘‘humification index’’
of soil DOM by quantifying the red-shifting of fluores-
cence emission that occurs with humification (Zsolnay
et al., 1999; Ohno, 2002a,b; Zsolnay, 2002). This index
was found to be negatively correlated with biodegrad-
ability of soil DOM (Kalbitz et al., 2003).
This research aims at assessing the biodegradability
potential and dynamics of effluent DOM in effluent-irri-
gated soils. Biodegradation experiments of effluent
DOM were performed using effluents from the munici-
pal wastewater reclamation complex of Greater Haifa
in northern Israel. DOM biodegradation was monitored
by DOC concentrations, UV absorbance and fluores-
cence EEM measurements.
2. Materials and methods
2.1. Effluent sampling
Effluent samples were obtained from two locations in
the Greater Haifa municipal wastewater reclamation
complex which is one of the two largest wastewater rec-
lamation systems in Israel. This complex renovates about
20–25 millions m3 of effluents per year (Halperin and
Ofir, 2003). The wastewater treatment scheme includes
a primary settling, activated sludge and secondary clari-
fiers. The treated effluents are chlorinated and then con-
veyed to the Yizre�el Valley, about 30 km to the east,
where they are impounded during the rainy winter
months in the Qishon (Kishon) reservoirs, with the aim
to use them for agricultural irrigation during the dry
summer months. ‘‘Haifa effluent’’ was sampled at the
treatment plant after the secondary clarification unit (be-
fore chlorination) on August 2003. The ‘‘Qishon efflu-
ent’’ was sampled from the surface layer of the
northern Qishon reservoir on October 2003. These two
effluent samples are considered to be representative to
the Greater Haifa complex based on the results of
DOC, UV254, EEM and pH observed in two more sam-
pling events on this year (Borisover et al., 2004).
2.2. Determination of pH, DOC and UV absorbance
in effluent samples
Effluents were filtered on the sampling day through a
1.6 lm filter (GF/1, glass fiber; Macherey-Nagel,

532 I. Saadi et al. / Chemosphere 63 (2006) 530–539Germany) followed by 0.7 lm filter (GF/F, glass fiber;
Whatman). The pH of the effluents was determined
immediately after filtration. Samples were then refriger-
ated (4 �C) for further analyses. The DOC was measured
using Shimadzu TOC-5000 Total Organic Carbon Ana-
lyser (catalyzed combustion). Before DOC analyses sam-
ples were acidified to pH < 2 using 2 N HCl and purged
with oxygen for 4 min to remove CO2. Absorbance at
254 nm (UV254) was obtained in 1.0 cm quartz cuvette
on a computer-controlled CARY 50 Bio UV–Visible
Spectrophotometer. Absorbance (cm�1) is expressed as
the ratio between optical density and optical length
(1.0 cm). SUVA values (specific UV absorbance; l mg�1
m�1) as a measure of the relative contents of aromatic
structures in the overall DOM (Weishaar et al., 2003)
were calculated as (UV254/DOC) · 100.
2.3. Fluorescence measurements
Fluorescence spectra of effluent DOM were measured
in a standard 1.0 cm quartz cell using a spectrofluorom-
eter (Shimadzu, RF 5301PC) equipped with 150 W
Xenon lamp (Ushio Inc., Japan). Lamp decay, as re-
vealed by measuring the fluorescence of pyrene standard
solutions (Fluca, puriss. for fluorescence), did not exceed
5% during the experimental period and had no effect on
comparing the fluorescence intensities of different DOM
solutions. The Raman peak of water at 348 nm was also
used to test instrument stability.
Three-dimensional excitation–emission matrices
(EEM) were generated at 23 ± 2 �C, at excitation and
emission slit widths of 5.0 nm band pass. Fluorescence
emission spectra were collected for excitation wave-
lengths between 220 and 550 nm at 10 nm intervals.
All emission data were corrected for inner filter effect
(IFE; Gauthier et al., 1986) and then, the 3-D maps were
built using Surfer 8.0 software. Coordinates of the three
main noticeable peaks were established in these maps.
Peak positions were not affected by five-fold DOM
dilution (as tested for selected samples of this study
and observed in many other effluent DOM systems
(unpublished data)), thus indicating that IFE correc-
tions did not affect peak location.
2.4. Biodegradation experiments
Biodegradation of Haifa effluents was tested for a
period of 60 days in 20 ml glass vials (10 ml aliquots),
scarifying triplicate vials at interval analyses. A longer
term (up to 4 months) biodegradation experiment was
conducted on the Qishon effluent as these effluents are
mostly used for irrigation and their long-term biode-
gradability is of more importance for environmental
concerns related to effluents irrigation. The Qishon efflu-
ents were tested in triplicate air-tight 2 1 glass jars (11
aliquots) to enable multiple sub sampling.The experiments were prepared by transferring ali-
quots of the filtered effluents into each incubation vessel
and inoculating it either with non-filtered effluents or
with soil. Effluent microorganisms were collected on
0.7 lm GF/F filter and suspended overnight in 0.85%
saline (0.85% NaCl). Soil inocula were prepared from
the top 10 cm layer of a vertisol type soil from Ginnegar,
located in Yizre�el Valley, in a field that is being irrigated
with the Qishon effluents for more than 10 years. This
soil was selected since it might exhibit acclaimed micro-
bial populations capable of degrading effluent DOM. A
prewashed soil inoculum was prepared from 1 g of fresh
soil suspended in 30 ml of 0.85% saline and shaken over-
night on a rotary shaker (150 rpm) at 25 �C. The suspen-
sion was centrifuged for 10 min at 6500 rpm, and the
supernatant was discarded. This procedure was repeated
three times to minimize contribution of DOM by the
inoculum. The precipitate was then re-suspended, using
vortex, in additional 10 ml of saline. Approximately
105�106 CFU ml�1 of both effluent and soil bacteria
suspensions were transferred to each incubation vessel.
Incubation was performed in the dark at 25 �C.
Vessels were mixed periodically. In all incubation exper-
iments, oxygen in the headspace was in excess relatively
to low carbon concentrations. To assess any possible
contribution of DOM from the soil inoculum, control
vials containing saline and soil inoculum were prepared
as well. The effluent inoculum made a negligible (non-
measurable) input of DOC to the incubation vessel.
Sterile control samples were additionally filtered
through a 0.2 lm filter (prewashed FP, Schleicher &
Scuell). At different time intervals, triplicate vials with
Haifa effluent or 10 ml aliquots from triplicate jars with
Qishon effluent were filtered through a prewashed
0.45 lm syringe filter (Schleicher & Scuell) and used
for DOC, UV, and fluorescence EEM analyses. Micro-
bial counts (R2A agar, incubation for 24 h at 30 �C)
were performed on the samples before filtration.
Incubation of the Qishon effluent samples was fol-
lowed 6 months until further degradation could not be
measured. Then, these samples were probed by adding
a readily available co-substrate. For this purpose, tripli-
cate aliquots of 80 ml were withdrawn from selected jars
(effluent-inoculated samples) and transferred into 250 ml
glass Erlenmeyer flasks. Samples were then supple-
mented with L-asparagine (99%; Sigma) or yeast extract
(Acumedia manufactures, Inc.; 44% carbon content as
determined by the TOC) at concentration of 10 mg car-
bon per liter such that the total DOC concentration was
brought to a comparable level as at the beginning of the
biodegradation experiment. Control effluent samples did
not receive any additional substrate. Yeast extract was
selected as a general complex organic substrate that
may provide limiting growth factors. L-asparagine was
selected on the basis of a preliminary substrate utiliza-
TMtion test using BIOLOG EcoPlate . The response of

L-asparagine (as indicated by the formation of a purple
color) was relatively strong as compared to the response
of the other substrates in the plate. After L-asparagine or
yeast extract had been added, the flasks were covered
with aluminum foil and incubated in the dark for 17
days (at 25 �C). Ten ml aliquots were withdrawn from
triplicate flasks at time intervals, filtered through a pre-
washed 0.45 lm syringe filter and analyzed for DOC and
UV254. Bacterial counts were performed on the samples
before filtration.
3. Results and discussion
3.1. General description of Haifa and Qishon effluents
Table 1 represents some general and fluorescence
properties of Haifa and Qishon effluent samples used in
the present study. Improved effluent quality during
stabilization in the Qishon reservoir is reflected by lower
DOC and microbial count values. The DOC level of Hai-
fa effluent in August (19.5 mg l�1) is similar to the DOC
levels recorded in March (15.2 mg l�1) and June
(17.7 mg l�1) of 2003 (Borisover et al., 2004). The DOC
of Qishon effluent in October (15.8 mg l�1) was about
50% higher than the level recorded in June 2003
(9.9 mg l�1; Borisover et al., 2004), which is apparently
related to late summer algal bloom, a well known phe-
nomenon in effluents reservoirs in Israel (Friedler et al.,
2003). At a specific sampling event, direct relationships
between the quality of Haifa and Qishon effluents may
not be sought as the residence time of Qishon effluent
may vary between 2 and 6 months depending on rainfall.
EEM maps generated for Haifa and Qishon effluents
(before biodegradation experiments) are presented in
Fig. 1. Based on these maps, three main fluorescence
peaks were identified (Table 1); two peaks were pro-
posed as humic/fulvic components, and additional less
Table 1
General and fluorescence properties of Haifa and Qishon effluents
Haifa effluent
(August 2003)
Qishon effluent
(October 2003)
General properties
Dissolved organic carbon (DOC; mg l�1) 19.5 15.8
UV absorbance at 254 nm (UV254; cm
�1) 0.23 0.17
Specific UV absorbance (SUVA; 1 mg�1 m�1) 1.18 1.07
pH 8.00 8.08
Microbial count (colony forming units; CFU ml�1) 9 · 105 2.5 · 104
Fluorescence properties
DOC-normalized peak intensities/
coordinates (emission/excitation wavelengths)
Peak 1 (humic/fulvic) 15.7/(430–434/330–339) 9.4/(430–437/330–339)
Peak 2 (humic/fulvic) 13.7/(437–440/272–281) 11.9/(434–443/249–259)
Peak 3 (protein-like) 12.5/(348–359/281) 10.9/(346–359/281)
hon e
bers
I. Saadi et al. / Chemosphere 63 (2006) 530–539 533Fig. 1. Excitation–emission matrices (EEM) for Haifa and Qis
(DOC)-normalized fluorescence intensity values (F/DOC). Num
(1,2) and protein-like (3) constituents.ffluents. Contour lines are based on dissolved organic carbon
identify the major fluorescence peaks assigned to fulvic/humic

Fig. 2(a)) that some DOM was still released from the soil
inoculum during incubation. Part of this soil-borne
DOM was probably degraded during the experimental
period, but about 2–3 mg l�1 of residual DOC was more
biologically resistant. Therefore, it could be that some of
the DOM remained in soil-inoculated effluent samples
originated from the soil inoculum such that reduction
in effluent-borne DOM was underestimated. Absor-
bance at 254 nm decreased by 5% in effluent-inoculated
samples, and by 25% in soil-inoculated samples. UV
absorbance of control samples of saline and soil inocu-
lum were slightly increased presumably due to DOM
release.
The trends of SUVA are opposite. SUVA increased
by 56% (effluent-inoculated samples) or by 63% (soil-
inoculated samples), and also increased in control sam-
ples of saline + soil inoculum. These trends indicate
selective degradation of DOM with lower UV absorp-
tion, while higher SUVA values have been attributed
to more resistant aromatic humic-like matter (Edzwald,
1993; Laor and Avnimelech, 2002). Similar trends have
also been observed in biodegradation of natural soil
DOM (Kalbitz et al., 2003; Marschner and Kalbitz,
2003). The slight increase in both UV and SUVA in con-
trol samples containing the soil inoculum can thus reflect
a net contribution of a relatively resistant DOM fraction
534 I. Saadi et al. / Chemosphere 63 (2006) 530–539distinct peak was assigned to protein-like structures
(Baker, 2001; Leenheer and Croue´, 2003). Haifa effluent
was different from Qishon effluent mostly by the loca-
tion of peak 2 (different excitation wavelength) and by
the DOC-normalized intensity of peak 1 (by factor of
1.7). Biodegradation, precipitation and sedimentation
processes during effluent conveyance and stabilization
in the Qishon reservoir may result in reduction of the
overall levels as well as modifying the structure and
composition of the DOM. Chlorination of the effluents
before entering the Qishon reservoir may also alter the
chemical structure of fluorescing constituents in the
DOM. Such differences in fluorescence properties need
to be interpreted with caution after evaluating possible
effects of solution DOM interactions. In general, we
believe that the observed fluorescence peaks were not
related to the formation of excited DOM complexes
(exciplexes), as no decrease in DOC-normalized fluores-
cence intensity was observed upon dilution (as tested in
selected samples of this study and in many other DOM
systems; unpublished results). If DOM exciplex forma-
tion would be responsible for fluorescence peaks,
then, lesser probability of exciplex formation would be
expected at lower DOM concentrations thus reducing
fluorescence intensity.
Also, the lower fluorescence intensity of peak 1 in the
Qishon effluent, as compared with Haifa effluent, can
hardly be related to fluorescence quenching of the
DOM by heavy metals. Overall concentration of 18
heavy metals in Haifa and Qishon effluents was found
similar and close to 1 mg l�1 (Halperin and Ofir,
2003). These concentrations were determined for non-fil-
tered samples thus included free metal ions, DOM-
bound metals, and metals adsorbed onto particulate
matter. Obviously, particulate-associated metals were
eliminated in this study during sample filtration such
that actual heavy metal concentrations were even lower.
We found that, for example, 1 mg l�1 of a strong DOM
fluorescence quencher, such as copper, would reduce
DOM fluorescence by approximately 5% at the em/ex
wavelengths of interest. Hence, it is not believed that
metal-induced quenching effects are responsible for the
40% differences observed between fluorescence intensi-
ties of peak 1 in the Qishon and Haifa effluents.
3.2. Biodegradation of Haifa effluent-DOM
The results of the biodegradation experiment with
Haifa DOM are presented in Fig. 2. During incubation
the DOC was reduced by 54% in samples inoculated
with soil microorganisms and by 39% in samples inocu-
lated with effluent microorganisms. Importantly, no
reduction in DOC was observed in control sterile efflu-
ent samples, thus indicating that biodegradation was
rather the process governing DOM reduction in inocu-
lated effluents but not aggregation or precipitation.Although the soil inocula were prewashed three times
to minimize contribution of soil DOM, it can be seen
from the control samples (saline + soil inoculum;
0 10 20 30 40 50 60
0
5
10
15
20
25
ba
sterile effluent effluent+effluent inoculum
effluent+soil inoculum saline+soil inoculum
D
O
C
(m
g l
-
1 )
Time (days)
0 10 20 30 40 50 60
0.00
0.05
0.10
0.15
0.20
0.25
Time (days)

UV
25
4
(cm
-
1 )
0 10 20 30 40 50 60
0.0
0.5
1.0
1.5
2.0
2.5
dc Time (days)

SU
VA
(l
mg
-
1
cm
-
1 )
0 10 20 30 40 50 60
100
101
102
103
104
105
106
107
108
Time (days)

CF
U
m
l-1
Fig. 2. Dynamics of Haifa effluent DOM during 60 days of
incubation: (a) DOC; (b) absorbance at 254 nm (UV254); (c)
Specific UV Absorbance (SUVA); (d) microbial counts (colony
forming units, CFU). Error bars represent ± standard devia-
tion of three replicates.released from the inoculum. No considerable changes in