Journal of Hazardous Materials B112 (2004) 87–94

Treatability and kinetics studies of mesophilic aerobic biodegradation of

high oil and grease pet food wastewater
Victor Lei Liu, G. Nakhla∗ , A. Bassi
Department of Chemical and Biochemical Engineering, The University of Western Ontario, London, Ont., Canada N6A 5B9

Received 25 June 2003; received in revised form 1 April 2004; accepted 5 April 2004

Available online 7 June 2004

Abstract

In this work, batch activated sludge studies were investigated for the treatment of raw pet food wastewater characterized by oil and grease
concentrations of 50,000–66,000 mg/L, COD and BOD concentrations of 100,000 and 80,000 mg/L, respectively, as well as effluent from an
existing anaerobic digester treating the aforementioned wastewater. A pre-treatment process, dissolved air flotation (DAF) achieved 97–99%
reduction in O&G to about 400–800 mg/L, which is still atypically high for AS. The batch studies were conducted using a 4-L bioreactor at room
temperature (21 ◦ C) under different conditions. The experimental results showed for the DAF pretreated effluent, 92% COD removal efficiency
can be achieved by using conventional activated sludge system at a 5 days contact time and applied initial soluble COD to biomass ratio of
1.17 mg COD/mg VSS. Similarly for the digester effluent at average oil and grease concentrations of 13,500 mg/L, activated sludge affected
63.7–76.2% soluble COD removal at 5 days. The results also showed that all kinetic data best conformed to the zero order biodegradation
model with a low biomass specific maximum substrate utilization rate of 0.168 mg COD/mg VSS day reflecting the slow biodegradability of
the wastewater even after 99% removal of oil and grease.
© 2004 Elsevier B.V. All rights reserved.

Keywords: Mesophilic aerobic; Activated sludge; Food industry wastewater; Oil and grease; Biokinetics constants; Batch reactors

1. Introduction settling, etc. have been explored for pre-treatment of high

oil and grease wastewater. Dissolved air flotation (DAF) is
The presence of high-strength oil and grease in indus- still considered to be the best due to its high oil and grease
trial wastewaters poses serious challenges for both aerobic reduction efficiency, which can be up to 99% [3].
and anaerobic biological treatment systems [1]. The devel- Anaerobic treatment of high-strength wastewater has
opments in rendering industries high-strength wastewater been achieved by using low rate covered anaerobic ponds
treatment during the last 10 years were summarized by and high rate anaerobic systems, as well as the traditional
Johns [2]. The latest applications and developments for dif- uncovered anaerobic ponds and low rate anaerobic systems
ferent treatment processes were evaluated below. While a [2]. Anaerobic treatment processes can favorably compete
series of primary treatment procedures, such as screening, with aerobic processes for the treatment of high O&G food
industry wastewater provided that the wastewater is high
in strength and is at high temperatures [4,5], particularly at
Abbreviations: AS, activated sludge; AnRBC, anaerobic rotating bio- thermophilic ranges where the solubility of oils is high. The
logical contactor; BOD, biological oxygen demand; COD, chemical oxy- use of surfactants to aid in the emulsification and removal of
gen demand; DAF, dissolve air flotation; DO, dissolve oxygen; HRT,
oils from wastewaters have been explored for the treatment
hydraulic retention time; O&G, oil and grease; RAS, returned activated
sludge; RBC, rotating biological contactor; SBOD, soluble BOD; SCOD, of wool-scouring wastewater, which is characterized by very
soluble COD; SP, soluble phosphorus; SRT, solid retention time; TBOD, high COD and high O&G, very similar to the high-strength
total BOD; TCOD, total COD; TP, total phosphorus; TSS, total suspended pet food wastewater described in this study. Application
solid; UASB, upflow anaerobic sludge blanket; VSS, volatile suspended of a large covered anaerobic pond to treat pork-processing
solid
∗ Corresponding author. Tel.: +1-519-661-2131; wastewater achieved BOD5 removal efficiency of 85–90%
fax: +1-519-661-3498. [6]. The anaerobic contact (AC) reactor was also considered
E-mail address: [email protected] (G. Nakhla). as a high rate anaerobic technology. AC reactor systems have

0304-3894/$ – see front matter © 2004 Elsevier B.V. All rights reserved.
doi:10.1016/j.jhazmat.2004.04.007
88 V.L. Liu et al. / Journal of Hazardous Materials B112 (2004) 87–94

full-scale completely mixed mesophilic anaerobic digester.

Nomenclature The objective of this study is to investigate the aerobic batch
K maximum substrate consumption specific rate activated sludge treatment of high oil and grease wastewater
Ks Monod half velocity concentration with O&G concentration as high as 800 mg/L. The research
S0 initial substrate concentration will also delineate the kinetics of organic removal under such
Umax maximum specific growth rate conditions. During this AS study, different contact time and
X biomass concentration in reactor different initial applied substrate to biomass ratios (S0 /X0 )
X0 initial biomass concentration were investigated in order to determine the optimal organic
YX/S biomass growth yield coefficient loadings.

achieved better than 84% COD, 93% BOD5 and 75% TSS 2. Methodology
removal in slaughterhouse wastewater treatment [2]. While
anaerobic processes have been successful for the treat- 2.1. Existing system description
ment of soluble organic waste [7], their success in dealing
with high oil and grease wastewater has been limited. For The treatment system at this rendering facility in south-
example, in treating wool-scouring wastewater with high ern Ontario consists of an anaerobic digester that provides a
O&G strength, the thermophilic anaerobic process allowed hydraulic retention time of 30 days at the average flow rate
a COD degradation of about 40% at an HRT of 10 days of 60 m3 /day. The wastewater characteristics are shown in
[8]. Table 1 together with the anaerobic digester effluent. The
For aerobic biological wastewater treatments at high O&G very high COD and suspended solids concentrations in both
concentrations, an activated sludge system treating oil and the raw wastewater and digester effluent are noteworthy, de-
grease at concentrations of 400 and 600 mg/L, achieved spite the 60–70% reduction in SS achieved in the digester.
COD removal efficiencies of 86 and 75%, respectively, while It should be noted that oil and grease concentrations in the
at 800 mg/L oil and grease concentration in feed, the COD scum layer, which constituted well over 50% of the digester
removal efficiency of the bioreactor dropped markedly [9]. volume exceeded 100,000 mg/L. Gas production from the
Thus it is apparent that the treatment of high-strength oil digester was minimal while COD reduction was at approx-
and grease wastewaters is very challenging and novel tech- imately 30% well below normal for a wastewater with a
nologies utilizing either aerobic or anaerobic processes have BOD-to-COD ratio of >0.7.
not been very successful. The conventional AS treatabil-
ity study of high oil and grease and COD wastewater un- 2.2. Activated sludge experiment setup
der mesophilic condition or even at room temperature has
seldom been investigated before [10]. While aerobic ther- 2.2.1. Batch scale system setup
mophilic processes have been used for the treatment of high The activated sludge reactor using in this study is a batch
oil and grease wastewater including liquid pig manure [11], scale complete mixed reactor model as shown in Fig. 1. The
potato-processing wastewater [12,13], these systems are of- bioreactor system was made from glass with a working vol-
ten plagued by foaming problems, solids separability con- ume of 4 L. It was aerated through an air diffuser and me-
cerns, and volatilization of ammonia necessitating air phase chanically mixed by a magnetic stirrer (Corning Plate Stirrer
treatment. PC-351). Under these conditions, the dissolved oxygen con-
This project addresses the treatment of raw wastewater centration in the reactor was kept between 2 and 3 mg/L. The
from rendering operations prior to and post treatment by a batch aerobic treatability studies were conducted at room

Table 1
Raw and pre-treated wastewater characteristics
Parameters (mg/L) Raw waste (before digester) Anaerobic digester effluent DAF pretreated effluent
Range Average Range Average Range Average

TSS 17300–61700 36857 11400–17200 14467 1160–2250 1685

VSS 15180–59800 34383 10600–17100 13867 1060–1850 1500
TCOD 74925–154100 96660 52300–98525 77300 16940–20500 18810
SCOD 13125–18450 16757 13150–23375 18855 11060–16940 13700
TBOD 77800 77800 11800–13200 11900
SBOD 8820 8820 8800–10500 10000
NH3 -N 197.5–400 328 680–1485 1353 674–1348 1186
PO4 3− 500–830 665 240–355 286 210–360 249
O&G 38800 38800 5942–21500 13500 404–820 668
 V.L. Liu et al. / Journal of Hazardous Materials B112 (2004) 87–94 89

of the pilot system. All the vessels are made from stain-
less steel with a glass observation window. DAF pretreated
waste from a storage tank was pumped (Milton Roy LMI
metering pump, 24.4 GPD at 125 psi) into AS reactor. The
AS reactor was then aerated by pressurized air through tube
type air diffusers. The pressurized air for both DAF and AS
reactor provided by a 5 hp air compressor. The operational
conditions were similar to lab scale batch conditions with
batches running for 6 days and grab samples of the mixed
liquor were collected daily as well.
Fig. 1. Batch activated sludge experimental setup: (1) 4 L bioreactor, (2)
magic stirrer (Corning Plate Stirrer PC-351), (3) DO meter (YSI Dissolve 2.2.3. Analytical methods
Oxygen Meter Model 50), (4) DO probe, (5) air diffuser, (6) mixing bar. During the batch sample analysis, mixed liquor samples
were withdrawn from the reactor once a day and analyzed
temperature (21 ◦ C) under different initial food to microor- for total suspended solid (TSS), volatile suspended solid
ganism ratio (S0 /X0 ). The wastewater samples used in AS (VSS), dissolved oxygen (DO), total COD (TCOD), soluble
reactor were the raw waste, DAF pre-treated waste, and also COD (SCOD), ammonia nitrogen (NH3 -N), nitrate nitrogen
from the aeration tank of an on-site pilot scale activated (NO3 -N), total phosphorus (TP), soluble phosphorus (SP)
sludge system, which is described later in Section 2.2.2. and oil and grease (O&G).
Wastewater was mixed with a typical concentrated returned Portions of the samples were collected daily from the
activated sludge (RAS) collected from Adelaide municipal mixed liquor in the bioreactor, were filtered through a 45 ␮m
wastewater treatment plant in London, Ontario, which em- fiber glass filter paper (Whatman, 47 mm 1822 047) in order
ploys conventional activated sludge system for BOD re- to analyze the soluble wastewater characteristics as well as
moval and nitrification. The concentrated sludge was ob- TSS and VSS using Standard Method (No. 2540D for TSS
tained by settling the regular RAS to 1/3 of the total vol- and No. 2540E for VSS [14]). Dissolved oxygen was con-
ume. The concentrated sludge had the following character- stantly monitored by a DO meter (YSI Dissolved Oxygen
istics: TSS = 9900–12,400 mg/L, VSS = 7000–9130 mg/L, Meter Model 50). Oil and grease testing was sent out to a
TCOD = 12,800–16,200 mg/L. certified commercial laboratory, while all the other parame-
In the batch scale study, the system was initially operated ters were determined by HACH equipment (HACH Odyssey
at different reaction times to determine the optimum con- Analyzer and COD heating reactor) using standard HACH
tact time from an SCOD removal standpoint. Grab samples testing kits for different analyses.
of the mixed liquor were collected daily, filtered through
0.45 ␮m filter paper to determine TSS, VSS, and SCOD.
2.3. Kinetics modeling
The results showed that SCOD reduction dropped remark-
ably when the reaction time was longer than 5 days. There-
The kinetics modeling used in this study were based on
fore, the maximum reaction time of the selected batches
basic Monod model. Two limiting cases of the Monod model
was controlled at 5 days.
were considered.
2.2.2. On-site pilot scale reactors setup
Subsequent to the laboratory scale batch studies, an 2.3.1. Zero order model
on-site pilot batch scale system was set up. The system In the cases of constant biomass concentrations with low
consisted of a 20 L DAF tank, a 150 L storage tank and a biomass change, i.e., X  X0 , and high substrate concen-
28 L activated sludge reactor. Fig. 2 shows the schematic tration (S  Ks ), Monod equation can be reduced to a zero

Fig. 2. Pilot activated sludge experimental setup: (1) 20 L pressure tank, (2) 150 L storage tank, (3) 28 L AS bioreactor, (4) air diffuser, (5) motor mixer,
(6) booster pump (Milton Roy LMI).
90 V.L. Liu et al. / Journal of Hazardous Materials B112 (2004) 87–94

order reaction [15]: SCOD

dS 7000
= kX (1)
dt 6000

Therefore, the kinetics constants ‘kX’ can be measured by 5000

SCOD (mg/L)

zero order linear regression using substrate S versus time

4000
plot, with the slope being equal to the product of ‘k’ and
X. Thus ‘k’ is the slope of the zero order coefficient versus 3000
biomass (X) concentration. 2000

2.3.2. First order model 1000

On the other hand, based on the same constant biomass 0

concentration condition, with Ks  S, Monod equation can 0 2 4 6 8 10
be simplified to a first order reaction: Time (d)
dS kXS
= (2) Fig. 3. Soluble COD batch removal kinetics.
dt Ks
Therefore, the first order biodegradation kinetics coefficient this test clearly show that approximately 10% of the initial
‘kX/Ks ’ can be determined from ln(S/S0 ) versus time plot. soluble COD was non-biodegradable even after 10 days of
The slope of the first order biodegradation coefficient versus treatment.
biomass is thus k/Ks . Table 2 summarizes the initial O&G, substrate (as SCOD)
and biomass concentrations and SCOD removal efficiencies
for the various waste streams investigated. It is noteworthy
3. Results and discussion that COD removal efficiencies for the DAF pretreated efflu-
ents both in lab and pilot scale studies were based on solu-
3.1. Overall performance ble COD since it accounted for over 70% of the total COD,
while the soluble fraction of the BOD was 0.87. For the
The temporal variation of soluble COD in the batch scale particulate-laden digester effluent, COD removals have been
operated with DAF effluent wastewater at initial substrate to based on both soluble and total COD data. It is conceded
microorganism ratio of 1.17 is depicted in Fig. 3. As apparent however that the total COD removal efficiencies are affected
from Fig. 3, SCOD removal was accomplished within 4 days, by the distribution of particulate COD between biomass and
and no further reduction in SCOD was observed between substrate.
day 5 and day 9, with the steady state SCOD stabilizing The initial substrate to biomass (S0 /X0 ) has been calcu-
at 300–500 mg/L after day 5. Thus it was decided that all lated using the initial measured VSS, which includes par-
batches would be run for 5 days. Furthermore, the results of ticulate COD, and the initial VSS originating from the seed

Table 2
Initial conditions in different activated sludge batches
Batch name S0 (SCOD) (mg/L) X0 (VSS)a S0 /X0 X average X0 (X)b S0 /X0 (X)b O&Gc (mg/L) COD removal (%)
(mg/L) (VSS)a (VSS)a (mg/L)
(mg/L)
DAF 1 5930 5060 1.17 5208 4565 1.30 >600 92.0
DAF 2 2485 2110 1.18 2625 1750 1.42 >600 92.0
DAF 3 10100 4125 2.45 5861 1160 8.71 >600 64.4
DAF 4 7850 2833 2.77 3375 2455 3.20 >600 60.8
DAF 5 7565 5200 1.45 5562 4055 1.87 >600 70.8
DAF stored 1 7600 3700 2.05 4295 2455 3.10 660 47.6
DAF stored 2 7740 6430 1.20 5643 4055 1.91 660 70.8
Digester effluent 1 10500 (35600)d 13900 0.76 17314 4565 2.30 13500 63.7 (18.9)e
Digester effluent 2 12600 (27100)d 13600 0.93 16270 4565 2.76 13500 76.2 (44.2)e
Digester effluent 3 9450 (30600)d 19000 0.50 22130 3043 3.11 13500 69.3 (37.2)e
On-site 1 12750 5130 2.49 7045 5130 2.49 664 88.7
On-site 2 8530 4830 1.77 6388 3830 2.23 664 78.4
a
These values were calculated based on VSS values being identical to biomass concentration.
b
These values were calculated based on assuming only seed sludge VSS as biomass concentration.
c O&G parameter values were not measured in each batch, while the samples were collected and tested periodically. The above shown O&G values

were measured during selected batch experiments.

d The numbers in parentheses are the initial values of total COD for the batches.
e The numbers in parentheses are the total COD removal efficiencies.
 V.L. Liu et al. / Journal of Hazardous Materials B112 (2004) 87–94 91

Xo as Seeding VSS

Xo as measured VSS

100.0

80.0
Removal (%)

60.0

40.0

y = -23.62x + 121.84
20.0
R2 = 0.9327
0.0
0.00 0.50 1.00 1.50 2.00 2.50 3.00 3.50
So/Xo

Fig. 4. Soluble COD removal efficiencies in different batches.

sludge obtained from the municipal wastewater treatment grease adversely impacts aerobic biological systems is by
plant. Perhaps the most remarkable observation is that the coating the biological floc and hindering oxygen transfer.
activated sludge affected 63.7–76.2% soluble COD removal Based on the five laboratory batches conducted with DAF
and 18.9–44.2% total COD removal at O&G concentrations pretreated wastewater from the anaerobic digester, COD re-
of 13,500 mg/L. The reproducibility of the data is evident moval efficiencies varied widely from 60.8 to 92% at S0 /X0
since all three batches were at S0 /X0 values within 10–20% values of 1.3 to 8.71 with removal efficiency generally de-
of each other and also affected close removal efficiencies. creasing with the increase in S0 /X0 . The relationship be-
This is contradictory to the findings of other researchers [9] tween the COD removal efficiencies for the DAF pretreated
who reported that oil and grease concentrations above 600 digester effluent and the initial substrate to biomass ratio is
and 800 mg/L affected significant reductions in COD re- depicted in Fig. 4. It is evident that generally removal ef-
moval. While the reasons for these discrepancies are largely ficiency decreased with an increase in S0 /X0 with the peak
unknown, it is postulated that the nature of oil and grease SCOD removal efficiency of 92% occurring at S0 /X0 in the
in this rendering wastewater being predominantly of animal range of 1.3–1.42 mg COD/mg VSS. Since in this study all
origin rather than mineral, is not only non-inhibitory but also batches were run for the same time, the S0 /X0 is therefore
biodegradable. The principal mechanism by which oil and synonymous with food to microorganism ratio (F/M) and

Table 3
Comparative performance of various treatment processes
Wastewater type Treatment methods Influent COD (mg/L) Oil and grease (mg/L) Substrate removal (%) Ref.

Previous studies
Beer brewery Thermophilic aerobic reactor (50 ◦ C) – – 67 [23]
Slaughterhouse Anaerobic lagoons 2291 624 79 [17]
Slaughterhouse Anaerobic reactor 3600 1700 33–52 [17]
Slaughterhouse UASB 1610 – 57 [17]
Dairy AnRBC – – 60–81 [24]
High strength AnRBC 12000 – 71 [19]
Food processing UASB 4700 – 80–85 [25]
Wool scouring Thermophilic aerobic reactor (65 ◦ C) 77000 1100 40 [20]
Wool scouring Mesophilic anaerobic reactor 40000 – 40 [8]
Dairy Activated sludge 2521 400 86 [9]
600 75
800 0
Dairy Pre-hydrolysis + activated sludge 2521 400 93 [9]
600 92
800 82
Meat processing UASB 1544 144 56 [18]
Meat processing UASB + RBC 1544 144 91.5 [18]
Current study
Pet food Conventional activated Sludge 13700 404–820 70–92 –
92 V.L. Liu et al. / Journal of Hazardous Materials B112 (2004) 87–94

accordingly the observed pattern of COD removal with in- processes offer the advantages of low energy consumption,
creasing S0 /X0 or F/M is typical [16]. This is to be excepted sludge production, and nutrient requirements, their vulner-
since any organic loading above the maximum microbial ability to high oil and grease concentrations is notable. The
uptake will be untreated. relative advantages of aerobic thermophilic processes are
A similar trend was observed for the stored DAF pre- offset by inconsistent performance, and operational prob-
treated and the raw wastewater with COD removal efficien- lems i.e., foaming and the need for off-gas to treatment for
cies decreasing from 70.8 to 47.6% with an increase in S0 /X0 ammonia. Based on the findings of this study and that of
from 1.91 to 3.1 mg COD/mg VSS in the case of stored DAF Jung et al. [9], the activated sludge process can effectively
pretreated and from 76.2 to 69.3% at S0 /X0 values of 2.76 treat wastewater at oil and grease concentrations of up to
and 3.11 mg COD/mg VSS for the raw wastewater. For the 800 mg/L. In fact it has been shown that membrane bioreac-
two on-site batches, the COD removal efficiency increased tor (MBR), which is a modification of the conventional acti-
mildly from 78.4 to 88.7% with the increase in S0 /X0 from vated sludge system can achieve 94–96% COD reduction at
2.23 to 2.49 mg SCOD/mg VSS. It is unlikely that the on-site influent oil concentration of 500–3000 mg/L at a hydraulic
batch would exhibit substantially different trends than the retention time of 13.3 h with sludge and oil concentration
laboratory ones and this slight discrepancy may be attributed in the MBR as high as 48 and 30 g/L, respectively [21].
to experimental errors. This study clearly demonstrates that biomass retention and
Table 3 compares the performance of various aerobic long sludge ages are critical to the performance of activated
and anaerobic treatment processes for food processing sludge systems.
wastes. As evident from the table, COD removal efficiencies
achieved in this study using the activated sludge processes 3.2. Biokinetic modeling
are at the high end of literature values. For example, at sim-
ilar O&G concentrations anaerobic lagoons achieved 79% As elaborated upon earlier, both limiting cases of the
[17] at much lower influent COD concentrations. Upflow Monod model i.e., zero order and first order kinetics were
anaerobic sludge blanket reactors [17,18] as well as anaer- investigated. A summary of the zero order and first order
obic rotating biological contactors [19] achieved 56–70% coefficients for the various batches is listed in Table 4 to-
COD removal. Thermophilic aerobic process, which can gether with the various correlation coefficients. Figs. 5 and
operate at very high oil and grease concentrations [20] due 6 illustrate graphically the fit of the data from both lab and
to solubilization of oil, achieved highly variable removal on-site batches to the zero order and first order kinetic mod-
efficiencies in the range of 40–90%. Thus while anaerobic els, respectively. It is apparent from the data that both model

Raw
DAF Fresh
14000
DAF stored
On-site
12000
Linear (DAF

10000
Raw: y = -1223.2x + 11771
R2 = 0.7735
SCOD (mg/L)

8000
DAF: y = -1202x + 7980
R2 = 0.9663

6000
DAF Stored: y = -1107.1x + 7044.5
R2 = 0.9139
4000
On-site: y = -951.67x + 7281.7
R2 = 0.8155
2000

0
0 2 4 6 8
Time (day)

Fig. 5. Zero order soluble COD removal kinetics.

 V.L. Liu et al. / Journal of Hazardous Materials B112 (2004) 87–94 93

Raw
DAF

0.5 DAF Stored

On-site
0
0 2 4 6 8 Linear (Raw)

-0.5
Raw: y = -0.1927x + 0.2764
R2 = 0.7008
-1
Ln(S/So) in SCOD

DAF: y = -0.2664x + 0.1464

-1.5 R2 = 0.9635

-2 DAF Stored: y = -0.2426x - 0.3497

R2 = 0.9477

-2.5
On-site: y = -0.4886x - 0.0322
R2 = 0.9099
-3

-3.5

-4
Time (d)

Fig. 6. First order soluble COD removal kinetics.

fit the data well, with R2 values of zero order kinetics for the VSS measurement as shown in Fig. 7. The high
the DAF pretreated wastewater ranging from 0.66 to 0.97. R2 value of 0.634 reflects the reliability of the esti-
High correlation coefficients were also observed for all other mated value of k of 0.168 mg COD/mg VSS day. This
waste streams. By comparison, similar R2 values were ob- value is much lower than municipal wastewater values of
served for the first order kinetics. The goodness fit of the 4.3–28.7 mg COD/mg VSS day [22], which reflects the very
data to both first and zero order kinetics is rather intriguing slow biodegradability of the wastewater even after remov-
and need further investigations. ing oil and grease. It is noted from Fig. 7 that the plot of
The maximum specific substrate utilization rate, ‘k’
for the DAF pretreated digester effluent was obtained VSS avg. (DAF)
by linear regression of the zero order coefficient and k' Vs Biomass in 0-order Linear (VSS avg. (DAF))
the initial biomass concentration, X0 as reflected by 1500
k' (mg/L.d)

1000
vss avg.: y = 0.1683x + 153.28
Table 4 500 R2 = 0.6342
Kinetics of zero order and first order modeling
0
Batch Zero order kinetics First order kinetics 0 2000 4000 6000 8000

k (kX) R2 k (kX/Ks ) R2 Biomass (mg/L)

(mg/L day) (1/day)
First Order DAF K" Vs Biomass VSS
DAF 1 917.5 0.660 0.486 0.882 0.6 Linear (VSS)
DAF 2 439.6 0.825 0.526 0.924 0.5
DAF 3 1120 0.899 0.156 0.897 0.4
DAF 4 1019 0.916 0.215 0.906
K"

0.3 VSS: y = -5E-05x + 0.5083

DAF 5 1202 0.966 0.266 0.964 R2 = 0.1436
0.2
DAF stored 1 750 0.815 0.128 0.839
0.1
DAF stored 2 1107 0.914 0.243 0.938
Digester effluent 1 1223 0.774 0.193 0.701 0
Digester effluent 2 2217 0.787 0.347 0.898 1000 3000 5000
Digester effluent 3 1995 0.516 0.352 0.635 Biomass (mg/L)
On-site 1 1708 0.693 0.823 0.848
On-site 2 951.7 0.816 0.489 0.909 Fig. 7. Variation of zero order and first order biodegradation coefficients
with biomass.
94 V.L. Liu et al. / Journal of Hazardous Materials B112 (2004) 87–94

the first order biodegradation coefficient versus biomass • Although both zero order and first order models fit the
concentration did not yield a straight line relationship with biokinetics data for the batch studies, the zero order
a positive slope of k/Ks (L/mg VSS day) as expected. The was deemed more pertinent with an estimated maxi-
reasonably good fit of the data to both zero order and first mum biomass specific substrate utilization 0.17 mg COD/
order models approximations may be explained by a vary- mg VSS day.
ing biomass concentration or prevalence of wide various
in substrate concentrations within the vicinity of this Ks
value in any given batch. A comparison of the initial and Acknowledgements
average biomass concentration for the DAF pre-treatment
digester effluent indicates that generally the discrepancies The financial support for this study, provided by
between these two values were mostly in the 4–25% with CRESTech (under Grant No. WRPF022) is gratefully ap-
the exception of the on-site batches. While these values preciated.
clearly reflect some changes in the biomass concentration,
which undermine the zero order and first order models,
such changes are mostly within the expected variability for References
VSS measurement of 10% [14] and thus the delineation of
biokinetic constants is not compromised. [1] D. Barnes, C.F. Forster, S.E. Hrudey, Surveys in Industry Wastewater
By examining the initial and final SCOD presented in Treatment, vol. 1, Food and Allied Industries, 1984.
Table 2 for the DAF effluent and the on-site batches, which [2] M.R. Johns, Bioresour. Technol. 54 (1995) 203–216.
[3] D. Hopwood, Process. Biochem. 12 (8) (1977) 5–8, 4 and 46.
essentially treated DAF-pretreated effluent. It is evident that [4] T.I. Mercz, R. Cord-Ruwisch, Water Res. 31 (1997) 170–178.
the initial SCOD was mostly in the 7565–12,750 mg/L and [5] S. Di Berardino, S. Costa, A. Conberti, Bioresour. Technol. 71 (2000)
the final SCOD were in the 1500–4000 mg/L. The fit of the 261–266.
data to both models may suggest a value of Ks , i.e., the [6] R.R. Dague, R.F. Urell, E.R. Krieger, Proceedings of the 44th Purdue
Industrial Waste Conference, Lewis, Chelsea, MI, 1990, pp. 815–
batches are zero order initially and first order towards the
823.
end. The fact that in general the zero order fit was overall bet- [7] E.L. Stover, G.J. Samuel, Proceedings of the 52nd Industrial Waste
ter than the first order particularly for the on-site batches, and Conference, Purdue University, MI, USA, 1997.
also correlated better with biomass concentration indicated [8] P. Spies, Ph.D. Thesis, University of Hannover, Germany, 1986.
that the initial rates were indeed the maximum substrate uti- [9] F. Jung, M.C. Cammarota, D.M.G. Freire, Biotechnol. Lett. 24 (2002)
1797–1802.
lization rates, i.e., the value of Ks is much lower than the
[10] D.I. Masse, L. Masse, Can. Agric. Eng. 42 (3) (2000) 139–146.
7565–12,750 mg/L and is close to the 1500–4000 mg/L. The [11] R. Beaudet, C. Gagnon, J.G. Bisaillon, M. Ishaque, Appl. Environ.
same remarks can be made regarding the raw digester efflu- Microbiol. 56 (1990) 971–976.
ent where initial SCOD ranged from 9450 to 12,600 mg/L [12] C.F. Chiang, C.J. Cu, L.K. Surg, Y.S. Wu, Water Sci. Technol. 43
with final SCOD ranging from 2900 to 3700 mg/L. (2001) 251–258.
[13] A.F. Rozich, D. Strom, Proceedings of the WEFTEC Asia, Singapore,
1998, pp. 5435–5440.
[14] Standard Methods for the Examination of Water and Wastewater, 18th
4. Summary and conclusions ed., American Public Health Association, American Water Works
Association and Water Environmental Federation, Washington, DC,
Based on the batch activated sludge studies conducted the 1992.
[15] V.A. Vavilin, L.Ya. Lokshina, Bioresour. Technol. 57 (1996) 69–80.
high-strength pet food wastewater at a low temperature of
[16] R.S. Ramalho, Introduction to Wastewater Treatment Processes, 2nd
21 ◦ C, the following conclusions can be drawn: ed., Academic Press, 1983, p. 257.
• High oil and grease wastes can be treated aerobically [17] J. Martinez, L. Borzacconi, M. Galisteo, M. Vinas, Water Sci. Tech-
nol. 32 (12) (1995) 99–104.
with favorable kinetics in conventional activated sludge [18] R.A. Wahaab, M. Hamdy El-Awady, Environmentalist 19 (1999) 61–
systems. More than 90% COD removal efficiency can 65.
be achieved at oil and grease concentrations as high as [19] C. Lu, C. Yeh, Environ. Int. 21 (3) (1995) 313–323.
660 mg/L at contact time of 5 days. [20] P. Becker, D. Koster, M.N. Popov, S. Markossian, G. Antranian, H.
• At 13,500 mg/L of oil and grease and total COD concen- Markl, Water Res. 33 (3) (1999) 653–660.
[21] W. Scholz, W. Fuchs, Water Res. 34 (14) (2000) 3621–3629.
trations of 77,300 mg/L, activated sludge can still achieve [22] M. Henze, C.P.L. Grady, W. Gujer, GvR. Marais, T. Matsuo, Water
respectable soluble COD removals in the 63.7–76.2% Res. 21 (5) (1987) 505–515.
range at 5 days contact time. [23] R. Zvauya, W. Parawira, C. Mawadza, Bioresour. Technol. 48 (1994)
• Soluble COD removal efficiencies from the DAF pre- 273–274.
[24] K.V. Lo, P.H. Liao, Biomass 21 (1990) 207–218.
treated digester effluent peaked at an initial substrate to
[25] L.C.H.V. Oliva, M. Zaiat, E. Foresti, Water Sci. Technol. 32 (12)
biomass concentration of 1.17 mg COD/mg VSS and de- (1995) 157–163.
creased almost linearly beyond that.