Provided for non-commercial research and educational use only. Not for reproduction or distribution or commercial use.

This article was originally published by IWA Publishing. IWA Publishing recognizes the retention of the right by the author(s) to photocopy or make single electronic

copies of the paper for their own personal use, including for their own classroom use, or the personal use of colleagues, provided the copies are not offered for sale and

are not distributed in a systematic way outside of their employing institution.

Please note that you are not permitted to post the IWA Publishing PDF version of your paper on your own website or your institution’s website or repository.

Please direct any queries regarding use or permissions to wst@iwap.co.uk

Integrated anaerobic treatment of dairy industrial

wastewater and sludge

Mauricio Passeggi, Ivan Lopez and Liliana Borzacconi

ABSTRACT

Mauricio Passeggi

Ivan Lopez

Liliana Borzacconi

Engineering Faculty,

University of the Republic,

J. Herrera y Reissig 565,

Montevideo,

Uruguay

E-mail: passeggi@fing.edu.uy

Performance parameters were studied in an alternative full-scale dairy effluent treatment system

comprising two anaerobic sludge-blanket reactors in parallel arrangement with upward flow,

internal fat-separation by flotation, external lamella settler and floated material digester. Reactors

were initially inoculated with flocculent sludge and granulated in a high-load stage. Using loading

rates up to a maximum 5.5 kgCOD/m3.d–hydraulic residence time of 17 hours- reactor efficiency

was found to remain stable around 90% of COD. Average sludge digester efficiency using a

loading rate of 3.5 kgVS/m3.d with a lipid content of 47% of COD amounted to 78% of VS (87% of

lipid removal). LCFA inhibition as assayed using palmitate was found to depend not only on the

palmitate concentration but also on the palmitate-to-biomass concentration ratio.

Key words | dairy wastewater, fat digester, full-scale, LCFA inhibition

INTRODUCTION

Depending on the type of product, equipment and unit

processes entailed in processing of dairy-based products,

effluent characteristics may vary widely according to indus-

trial plant. Thus, every individual situation must be con-

sidered separately with a view to ensuring appropriate

treatment design (Demirel et al. 2005). Nonetheless, on

account of process leakage of milk or milk-based products,

a number of constituents are systematically found in dairy

industrial wastewater: lactose, lipids, casein and other

proteins. Lipids, composed of triglycerides, are found mainly

in emulsions resulting from initial process stage of homo-

genization. In addition, effluents will be acidic or basic

according to chemical cleaning as used at any time of process.

The use of UASB reactors in dairy wastewater treatment

-as well as in the treatment of other complex effluents- has

found limited success in view of the fact that a considerable

amount of organic material hydrolyzes or degrades at an

excessively low rate, while normally building up within the

sludge blanket by entrainment or adsorption. The result is

the dilution of biomass, the affectation of mass transfer

properties and the impairment of sludge settling capacity.

Sludge activity is therefore reduced and sludge washed out

from the reactor in the outlet stream could occur (Sayed

1984; Sayed 1987; Rinzema 1993; Hwu et al. 1998).

Rapid adsorption and significantly slow degradation of

dairy fat emulsions were reported for batch studies using

biomass that was not adapted to degrading fats and a

substrate load as high as 12.6 grams of fat per gram of VSS

(Petruy & Lettinga 1997). Vidal et al. (2000) reported that

fat-rich dairy wastewater biodegradability rate is limited by

fats hydrolysis. They performed batch assays with non-

adapted biomass. However, using moderate loads and

substrate-adapted microorganisms, hydrolysis may not be

the rate limiting stage of total treatment process. Not limited

by hydrolysis, total process rate may be limited by a reduced

mass transfer rate due to LCFA adsorption onto sludge

(Pereira et al. 2003; Pereira et al. 2004), or by the inhibitory

effect of LCFA on acetotrophic methanogenic populations

(Perle et al. 1995; Hwu et al. 1996). According to research

reported by Hwu et al. (1998), sludge flotation in UASB

reactors due to LCFA adsorption may occur at lower LCFA

concentration than those that produce inhibitory effect.

doi: 10.2166/wst.2009.010

501 Q IWA Publishing 2009 Water Science & Technology—WST | 59.3 | 2009

With a view to studying the influence of the stage of

adsorption preceding lipid -and LCFA- degradation on the

performance of UASB reactors, several studies were made

using an intermittent effluent feed, leading to encouraging

results (Sayed 1984; Nadais et al. 2006). Cavaleiro et al.

(2008) reported that use of pulse-feeds results in increased

tolerance of acetotrophic methanogens to LCFA, and

suggest that satisfactory results for continuous operation

may be obtained following a stage of biomass acclimati-

zation by means of pulse-feed. Several studies demonstrated

that the use of flocculent sludge –rather than granular-

results in a higher efficiency in removing LCFA (Nadais

et al. 2003) and treating complex effluents (Sayed 1987). In

contrast, flocculent sludge appears to be less resistant to

LCFA inhibition (Hwu et al. 1996).

As a result of difficulties in treating complex effluents in

sludge-blanket reactors, pre-treatment methods are nor-

mally used, such as fat separation by means of flotation by

dissolved air. In order to obtain a compact system without

the need of previous fats separation, two UASB reactors

were modified to introduce a novel concept: Upflow

Anaerobic Sludge Blanket with internal fat separator and

external sludge recovery by settling. The system is com-

pleted by a floated material digester with the objective of

avoid sludge disposal. An analysis of this alternative

wastewater treatment system constructed at real scale in a

dairy industry in Melo City, Uruguay (Cooperativa de

Lecherıa de Melo) is presented in this paper. This dairy

industrial wastewater has lipid content higher than 40% of

total COD, and the mean flow is about 100 m3/d.

MATERIALS AND METHODS

System design and operation

In view of significant load and flow-rate fluctuations in

the plant’s effluent stream, a buffer tank (average hydraulic

residence time of 12 hours) was installed upstream of the

anaerobic reactors. Table 1 depicts the effluent characte-

ristics outlet the buffer tank, and Figure 1 shows the complete

treatment system lay-out, comprising the buffer tank, two

40 m3 anaerobic reactors (R1 and R2), a lamella settler and a

5 m3 digester of floated material.

Outlet effluents of both anaerobic reactors discharge to

a lamella settler with plates tilted 608 and spaced at 5 cm,

retaining part of the sludge carry-over from the reactors,

returning it to the feed box by means of pumping

equipment. The lamella settler operates at a residence

time of 15 minutes at full-load. The settling tank outlet is

discharged to municipal sewer.

A funnel-shaped sink for extraction of floating fat was

installed in both reactors, inside of the off-gas header and

two centimetres below liquid level. At each reactor, fat

extraction was performed by pumping from an external

chamber –acting as fat trap- receiving the discharge of the

sink, whereby float material is retained, returning the outlet

stream of this chamber to the reactor.

The reactors were inoculated with flocculent sludge

from an anaerobic treatment lagoon of an abattoir with

Specific Methanogenic Activity of 0.12 gCH4-COD/

gVSS.d. Intermittent and alternate feed rates were used at

start-up, with rate increasing gradually. Continuous feed

rate was used after 120 day operation, reaching the total rate

of wastewater generated by the plant 160 days after start-up.

Granulation process was analysed following the Jeison and

Chamy (1998) technique, immobilizing granules in agar and

then using the UTHSCA Image Tool software.

The float sludge digester was put in operation on day

290 in order to stabilize floating sludge extracted from the

reactors. Solid digester was inoculated with sludge from

the same lagoon as used for the reactors, and operated

under full mixing conditions with sludge recirculation by

means of pumping. The digester hold-up was recirculated

8 times per day. The digester outlet stream was discharged

Table 1 | Main characteristics of effluent outlet the buffer tank

Parameter Temp. (8C) pH BOD (mg/L) COD (mg/L) Fats (mg/L) TKN (mg/L) TSS (mg/L) VSS (mg/L)

Average 25.8 7.4 1710 2520 495 36 1020 710

Range 19–33 4.7–11 650–6240 430–15200 160–1760 14–90 250–2750 210–1890

502 M. Passeggi et al. | Integrated anaerobic treatment of dairy wastewater and sludge Water Science & Technology—WST | 59.3 | 2009

into R1 and R2 in order to retain the system’s biomass

hold-up.

Generated biogas was used to increase reactor influent

stream temperature by 3–48C, in addition to keeping the

digester contents at a temperature between 328C and 388C.

A concentric tubes exchanger constructed in stainless steel

was used to increase the temperature of the influent and of

the digester content.

Data recording and system operation analysis

Reactor performance was evaluated by monitoring the

organic matter removal on a COD and a fat-and-oil basis.

In view of wastewater fluctuations, R1 and R2 influent and

effluent streams were analyzed weekly using refrigerated 24-

hour samples collected daily. COD was determined for all

combined samples. A settling stage was used prior to

analysis, so that COD determinations include soluble

COD and colloidal COD, avoiding distorted measurement

resulting from variable sludge carry-over. Fats and oils were

analyzed monthly using combined samples of influent and

effluent streams (Standard Methods, APHA, AWWA, WEF

1995). Temperature and pH measurements were recorded

daily for point samples of the influent and effluent streams;

and total alkalinity, bicarbonate alkalinity and volatile fatty

acids were determined by titration for point samples on a

three-times-per-week basis. Generated biogas was moni-

tored daily by meters connected to either reactor.

Lamella settler performance was evaluated by daily

measurement of settleable solids in after 30 minutes, for

both influent and effluent stream samples.

Digester performance was evaluated by monitoring the

rate of volatile solid removal and the rate of fat-and-oil

removal. Daily volume measurements were made of floated

sludge extracted from the reactors and fed to the digester;

and digester inlet and outlet samples were analysed on a

weekly basis for Total Solids, Volatile Solids and Fats and

Oils. Digester temperature was measured twice a day, at the

beginning and end of the heating period.

Assaying LCFA inhibition

To evaluate the effect of LCFA inhibition, batch tests were

carried out with different relationships of LCFA/VSS.

LCFA inhibition was assayed in batch runs using sodium

palmitate — the largest abundant of milk triglycerides and

primer of smaller LCFA present in the milk. Granular

adapted sludge from the UASB reactors was used to

perform the tests. Triplicate measurements were made in

150 mL vials with varying concentration of biomass

(expressed as VSS) and palmitate, while the acetate

concentration was kept constant as shown in Table 2.

Figure 1 | Treatement plant lay-out. (—) continuous flow; (– – –) intermittent flow.

503 M. Passeggi et al. | Integrated anaerobic treatment of dairy wastewater and sludge Water Science & Technology—WST | 59.3 | 2009

RESULTS AND DISCUSSION

Sludge blanket reactors

Figure 2 shows reactor feed loads and removal efficiency

in COD (weekly average). Removal efficiency remained at a

value around 90% of COD for load values increasing up to

5.5 kg/m3.d; the hydraulic residence time remaining at a

value around 17 hours during the maximum load stage.

Reactor temperature fluctuated within the range of 208C to

308C throughout the experiment and did not appear to have

a significant effect on the COD removal efficiency.

Operation of the reactors R1 and R2 was started with

440 and 400 kgVSS, respectively (mean concentration

values of 11 and 10 kgVSS/m3). Figure 3 shows the VSS

evolution. A reduction in VSS concentration to 5 g/L during

the initial stage was due to carry-over. About 290 kgVSS of

sludge were added to both reactors between days 162 and

250 to recover the initial concentration of sludge. Granula-

tion started around days 350 and 400 for R2 and R1

respectively; on day 400 small granules of an average size of

0.57 mm and 0.45 mm were measured in R2 and R1

respectively. Granulation resulted in a reduction in sludge

carry-over and led to an increase in VSS concentration to

above 20 g/L. The sludge volume index decreased from

37 mL/g in the inoculum to12 mL/g upon completion of

granulation. In despite flocculent sludge has been reported

to be more efficient in the treatment of effluents with a high

LCFA concentration (Pereira et al. 2002; Nadais, et al. 2003)

as well as complex effluents (Sayed 1987), the sludge has

granulated spontaneously in the higher-organic load stage,

resulting in improved sludge retention capacity.

The resulting combined effect of microbial growth,

sludge supply from the digester and carry-over in the

effluent allowed operation of the reactors without sludge

blowdown.

Lamella settler

Until the sludge granulated, the external settling tank

operated at efficiency around 60% as measured in settleable

solids, thus preventing reactor wash-out. Following granu-

lation, sludge carry-over was notably reduced, thus reducing

the relevance of settling in overall process. Settleable solids

at the settling tank outlet averaged 2.7 mL/L.

Floated sludge extraction and digestion

The volume of extracted float material did not vary

considerably with load. Mean values of daily extractions

from R1 and R2 (containing fats, sludge and insoluble

protein) amounted to 79 L and 73 L, standard deviation

being 54 L and 52 L respectively (using 458 input data for

either reactor). The fat digester, discharging digested sludge

into the reactors, started operating on day 290 with a HRT

of 34 days and temperature ranging between 328C and 408C.

Table 2 | Experimental conditions used in inhibition assays

Assay 1 2 3 4 5 6 7

Biomass (gVSS/L) 2 2 1 2 4 2 2

Palmitate (g/L) 0 0.25 0.5 0.5 0.5 0.75 1

Ratio (gPa/gVSS) 0 0.125 0.5 0.25 0.125 0.375 0.5

Acetate (g/L) 3 3 3 3 3 3 3

Figure 2 | Organic load applied to reactors and COD removal efficiency of the system.

Figure 3 | Volatile suspended solids content in the reactors. concentration is referred

to total reactor volume.

504 M. Passeggi et al. | Integrated anaerobic treatment of dairy wastewater and sludge Water Science & Technology—WST | 59.3 | 2009

The applied load averaged 3.5 kgVS/m3.d, with a fat and oil

content amounting to 47% of VS in the feed. Over the study

period, average efficiency amounted to 78% on a VS basis

and 87% on a fat content basis. Outlet concentrations

measured monthly (n ¼ 10) were 18 ^ 9 g/L and

4.8 ^ 3.9 g/L for VS and fat content respectively. Fat

removal efficiency was considerably high for a sludge

digester. While not emulsified, the fat load is within the

floated sludge in intimate contact with a large amount of

biomass adapted to lipids as substrate. Besides, the flow

model, nearly equivalent to a stirred continuous reactor,

enables operation at a relatively low concentration of fat

and LCFA, which may inhibit degradation.

Assaying LCFA inhibition

Specific methanogenic activity was found to decrease with

increasing palmitate concentration and with increasing

palmitate-to-VSS ratio (Figure 4). According to results, a

reliable measurement of the inhibitory effect of palmitate on

said activity would require, in addition to the determination

of IC50, an indication of the inhibitor-to-biomass ratio used

for such determination. Further, not only it is recommended

to ensure proper contact and mixing in order to prevent

LCFA build-up, but it is also important to ensure the system

operates at a high biomass concentration.

CONCLUSIONS

A compact system consisting in two modified UASB

reactors and a floated sludge digester was assayed at

industrial conditions to treat dairy wastewater. Results

here reported demonstrate the feasibility of treating dairy

industrial effluent without prior fat separation in anaerobic

reactors, even when the content of lipidic material may be

higher than 40% of COD. COD removal efficiencies around

90% are achieved in stable operating conditions.

Inhibitory effects of palmitate (representing the LCFA)

depend not only of palmitate concentration but palmitate

to biomass ratio.

Even using moderate loading rates, in flocculent sludge

fed by dairy wastewater granulation can occur. Hence,

retention of sludge into the reactor is favoured, and this

enables the use of higher feed loads.

The floated sludge digester achieves high efficiency,

allowing operates the system without sludge discharge for

final disposal.

ACKNOWLEDGEMENTS

To managers and workers of COLEME, for the good

attitude to include changes and to operate the reactors.

REFERENCES

APHA, AWWA, WEF 1995 Standard Methods for the Examination

of Water and Wastewater, 19th edition.

Demirel, B., Yenigun, O. & Onay, T. T. 2005 Anaerobic treatment

of dairy wastewater: a review. Process Biochem. 40,

2583–2595.

Cavaleiro, A. J., Pereira, M. A. & Alvez, M. 2008 Enhancement of

methane production from long chain fatty acid based effluents.

Bioresour. Technol. 99(10), 4086–4095.

Hwu, C.-S., Donlon, B. & Lettinga, G. 1996 Comparative toxicity of

long-chain fatty acids to anaerobic sludge from various origins.

Wat. Sci. Technol. 34(5-6), 351–358.

Hwu, C.-S., Tseng, S.-K., Yuan, C.-Y., Kulik, Z. & Lettinga, G. 1998

Biosorption of long-chain fatty acids in UASB treatment

process. Water Res. 32(5), 1571–1579.

Figure 4 | Palmitate inhibition effect on specific methanogenic activity.

505 M. Passeggi et al. | Integrated anaerobic treatment of dairy wastewater and sludge Water Science & Technology—WST | 59.3 | 2009

Jeison, D. & Chamy, R. 1998 Novel technique for measuring the

size distribution of anaerobic granular sludge, V Taller y

Seminario latinoamericano de tratamiento anaerobio de agues

residuales, Vina del Mar, Chile.

Nadais, M. H., Capela, M. I., Aroja, L. M. & Duarte, A. C. 2003

Biosorption of milk substrates onto anaerobic flocculent and

granular sludge. Biotechnol. Prog. 19, 1053–1055.

Nadais, M. H., Capela, M. I. & Aroja, L. M. 2006 Intermittent vs

continuous operation of upflow anaerobic sludge bed reactors

for dairy wastewater and related microbial changes. Wat. Sci.

Technol. 54(2), 103–109.

Pereira, M. A., Pires, O. C., Mota, M. & Alves, M. M. 2002

Anaerobic degradation of oleic acid by suspended and

granular sludge: identification of palmitic acid as a key

intermediate. Water Sci. Technol. 45(10), 139–144.

Pereira, M., Cavaleiro, A., Mota, M. & Alves, M. 2003 Accumulation

of long chain fatty acids onto anaerobic sludge under steady

state and shock loading conditions: effect on acetogenic and

metanogenic activity. Water Sci. Technol. 48(6), 33–40.

Pereira, M., Mota, M., Alvez, M. 2004. The important role of mass

transfer limitations caused by Long Chain Fatty Acids

accumulation onto the anaerobic sludge. Proceedings of the

10th IWA Congress on Anaerobic Digestion. Montreal,

Canada.

Perle, M., Kimchie, S. & Shelef, G. 1995 Some Biochemical Aspects

of the Anaerobic Degradation of Dairy Wastewater. Water

Res. 29(6), 1549–1554.

Petruy, R. & Lettinga, G. 1997 Digestion of a milk-fat emulsion.

Bioresour. Technol. 61, 141–149.

Rinzema, A. 1993 Anaerobic Digestion of Long-Chain Fatty Acids in

UASB and Expanded Granular Sludge Bed Reactors. Process

Biochem. 28, 527–537.

Sayed, S. 1984 Anaerobic treatment of slaughterhouse wastewater

using a flocculent sludge UASB reactor. Agric. Wastes 11,

197–226.

Sayed, S. 1987 Anaerobic treatment of slaughterhouse

wastewater using a granular sludge UASB reactor. Biol.

Wastes 21, 11–28.

Vidal, G., Carvalho, A., Mendez, R. & Lema, J. M. 2000 Influence of

the content in fats and proteins on the anaerobic

biodegradability of dairy wastewaters. Bioresour. Technol. 74,

231–239.

506 M. Passeggi et al. | Integrated anaerobic treatment of dairy wastewater and sludge Water Science & Technology—WST | 59.3 | 2009