w a t e r r e s e a r c h 4 3 ( 2 0 0 9 ) 5 5 3 – 5 6 1

Avai lab le a t www.sc iencedi rec t .com

journa l homepage : www.e lsev ie r . com/ loca te /wat res

Simultaneous colour and DON removal from sewagetreatment plant effluent: Alum coagulation of melanoidin

Jason Dwyera, Peter Griffithsb, Paul Lanta,*aThe Advanced Water Management Centre, The University of Queensland, St Lucia 4072, AustraliabCH2M Hill Australia Pty Ltd, Level 1, 33 Park Road, Milton, QLD 4064 Brisbane, Australia

a r t i c l e i n f o

Article history:

Received 9 January 2008

Received in revised form

28 June 2008

Accepted 21 October 2008

Published online 14 November 2008

Keywords:

Coagulation

Melanoidin

DON

DOC

Fractionation

Fluorescence

* Corresponding author. Tel.: þ61 7 3365 472E-mail addresses: jason.dwyer@uq.edu.a

0043-1354/$ – see front matter ª 2008 Elsevidoi:10.1016/j.watres.2008.10.053

a b s t r a c t

The aim of this study was to detect and characterise melanoidin in sewage treatment plant

(STP) effluent, and to study the ability of alum coagulation to remove the colour and dis-

solved organic nitrogen (DON) associated with melanoidin. The melanoidin is non-biode-

gradable due to the complex cyclic based structure and thus it directly contributes to

effluent nitrogen concentrations from the sewage treatment plant (STP). Lowering of

effluent total nitrogen limits and the link between colour and chlorinated disinfection by-

products have therefore driven a need to understand the structure, properties and treat-

ability of DON species found in STP effluent.

The focus of this paper is the refractory coloured, organic nitrogen compound melanoidin.

Wetalla STP effluent has relatively high colour (170 mg-PtCo L�1) and DON (2.5 mg L�1) for

a biological nutrient removal STP, owing to an industrial supply of melanoidin containing

molasses fermentation wastewater. Alum coagulation jar tests were performed on

synthetic melanoidin solution, STP effluent containing melanoidin (Wetalla, Toowoomba,

Australia) and STP effluent free of melanoidin (Merrimac, Gold Coast, Australia) to examine

the treatability of melanoidin and its associated colour and DON content when present in

STP effluent.

The removal of melanoidin from STP effluent resulted in significant colour and DON

reduction. An alum dose of 30 mg L�1 as aluminium was sufficient to reach maximum

removal of colour (75%), DON (42%) and dissolved organic carbon (DOC) (30%) present in

melanoidin containing STP effluent. Alum was shown to preferentially remove DON with

a molecular weight >10 kDa over small molecular weight DON. Fluorescence excitation-

emission matrix examination of the humic compounds present in the STP effluent indi-

cated that melanoidin type humic compounds were more readily removed by alum

coagulation than other humic compounds.

ª 2008 Elsevier Ltd. All rights reserved.

1. Introduction associated with melanoidin has been investigated in treating

This is the first study to investigate the use of alum coagula-

tion to remove melanoidin present in the effluent of a sewage

treatment plant (STP). Although the coagulation of the colour

6u (J. Dwyer), peter.griffither Ltd. All rights reserved

molasses wastewater (Migo et al., 1993, 1997; Levic and Gyura,

2000; Levic et al., 2005), no studies have focused on the

removal of residual colour, organic nitrogen and carbon

associated with the coagulation of melanoidin present in the

s@ch2m.com.au (P. Griffiths), paul.lant@uq.edu.au (P. Lant)..

w a t e r r e s e a r c h 4 3 ( 2 0 0 9 ) 5 5 3 – 5 6 1554

effluent of STPs. Total organic carbon removal during the

coagulation of molasses wastewater has been reported (Migo

et al., 1993), but as melanoidin is only one of many organic

compounds that make up the dissolved organic carbon (DOC)

in molasses wastewater, the DOC removal specific to mela-

noidin has not been determined. The removal of dissolved

organic nitrogen (DON) associated with melanoidin has not

previously been examined using any advanced treatment

methods.

Melanoidins are brown coloured, nitrogenous organic

compounds (Dignac et al., 2000), most commonly present in

wastewater fed to STPs due to the presence of molasses by-

product streams. They are refractory, thus result in elevated

effluent colour, organic nitrogen and carbon when fed to STPs.

Wetalla STP (Toowoomba, Australia) is a biological nutrient

removal plant that treats 19.5 mL day�1 of mainly municipal

wastewater. A significant industrial load of spent molasses

wastewater is present in the feed to the plant. Wetalla has

a distinctively brown coloured effluent (170 mg-PtCo L�1) and

above average dissolved organic nitrogen (2.5 mg L�1).

Melanoidins are a product of the Maillard reaction,

a complex reaction which occurs during the heating of a sugar

and an amine. Many attempts have been made to determine

the fundamental structure of melanoidin (Motai, 1974; Kato

and Tsuchida, 1981; Hayase et al., 1986; Cammerer and Kroh,

1995), but due to the complexity of the Maillard reaction, with

a multitude of reaction pathways and intermediates formed

(Martins et al., 2000), a precise structure has not been

determined.

The chemical properties of melanoidin resemble humic

substances, being acidic, polymeric and highly dispersed

colloids which are negatively charged due to the dissociation

of carboxylic and phenolic groups (Migo et al., 1993). The

molar mass of melanoidin increases the longer the sugar and

amino acid are heated, the higher the temperature and the

more alkaline the solution. They are colloidal compounds of

molecular weight between 40 and 70 kDa. Increasing solution

pH and alkalinity is reported to increase the degree of poly-

merisation, and thus the size of the colloids (Pena et al., 2003;

Coca et al., 2005a,b). Carbon to nitrogen ratios of synthetic

melanoidin solutions have been shown to range from 6.2 to

13.7, increasing with temperature and reaction time (Motai

and Inoue, 1974; Kato and Tsuchida, 1981; Benzingpurdie

et al., 1983; Cammerer and Kroh, 1995). The ratio also changes

depending on the types of amino acid and sugar precursors

and other reaction conditions.

Melanoidin, when present, contributes to the DON content

of STP effluent. DON has become a more significant

compound of effluent total nitrogen as effluent nitrogen limits

have become more stringent. Also, DON has been linked to

unwanted disinfection by-products during chlorination

(Richardson, 2003; Schrank et al., 2004; Sirivedhin and Gray,

2005).

DON is a small, but significant portion of the natural

organic matter (NOM) in STP effluent. The characterisation

and treatment of NOM have been extensively investigated

over recent years (Saadi et al., 2006). However, the DON

associated with NOM has received much less attention, and is

generally related to water treatment rather than wastewater

treatment (Egeberg et al., 1999; Westerhoff and Mash, 2002;

Pehlivanoglu-Mantas and Sedlak, 2006). The use of innovative

NOM characterisation methods with additional DON analysis

has the potential to provide quantitative and qualitative

information about the composition and properties of DON

containing compounds, and differentiate melanoidin related

DON from other NOM related DON.

Three characterisation techniques, important in the char-

acterisation of NOM, namely molecular weight character-

isation, specific ultraviolet absorbance (SUVA) and

fluorescence excitation-emission matrix (EEM) have recently

been described (Westerhoff et al., 2001; Westerhoff and Mash,

2002; Amy, 2007). Molecular weight fractionation can be per-

formed using size exclusion chromatography (SEC), in which

a mass distribution is given, or by ultrafiltration molecular

size fractionation which splits NOM into a variety of size

fractions. SEC gives a precise distribution of NOM compounds,

while ultrafiltration groups NOM compounds into a variety of

size ranges. The ability to make surrogate measurements

(DON, colour, SUVA, fluorescence, etc.) of the various size

fractions makes ultrafiltration advantageous over SEC.

The aromatic colloidal nature of melanoidin (Migo et al.,

1993) makes it possible to differentiate melanoidin from other

NOM. Colloidal compounds have relatively high SUVA values

for low carbon to nitrogen (C/N) ratio compared to other NOM

(Lee et al., 2006). Apart from colloidal compounds in natural

waters and wastewater, large molecular weight compounds

generally have low SUVA. Typically C/N is proportional to

SUVA values (Lee et al., 2006).

EEM is an increasingly popular tool used to specify the

types of compounds that contribute to NOM. EEM has been

used to differentiate between aromatic protein, fulvic-acid-

like compounds and humic-acid-like compounds (Chen et al.,

2003). Studies have shown that EEM can differentiate between

different types of humic substances, with melanoidin

compounds shown to fluoresce differently to terrestrial

humics (Coble, 1996).

This investigation used several NOM identification tech-

niques to identify melanoidin and study the ability of alum

coagulation to remove the melanoidin and the associated

colour and organic nitrogen from STP effluent. Specifically,

molecular weight ultrafractionation with surrogate measure-

ments of colour, DON, DOC and SUVA were used to identify

the presence of melanoidin and assess the effectiveness of

alum to remove melanoidins from STP effluent. EEM was used

to identify if melanoidin humics were preferentially removed

over other humics present in STP effluent.

2. Material and methods

2.1. Wastewater samples

Three different wastewaters were compared in this work. Two

municipal wastewater sources, one containing melanoidin

(Wetalla STP effluent) and the other with no melanoidin

(Merrimac STP effluent). The third water was a synthetic

melanoidin solution. Table 1 presents the values of the key

characteristics of each wastewater sample.

The synthetic melanoidin solution was made fresh daily by

diluting a concentrated solution of melanoidin. The

Table 1 – Characteristics of the three water samples.

Parameter Units Wetalla WWTP effluent Merrimac WWTP effluent Synthetic melanoidin

Flow mL day�1 19.5 35.0 –

pH 6.8–7.2 6.9–7.2 7

SS mg L�1 7.0 5.0 0

Colour mg-PtCo L�1 170 30 242

DOC mg L�1 20.7 13.3 19.4

DON mg L�1 2.52 0.96 3.09

NH4 mg L�1 0.03 0.64 0.16

NOx mg L�1 2.50 1.68 0.04

Phosphorous mg L�1 0.38 <0.5 0

SUVA L mg�1 cm�1 2.96 1.61 2.51

DOC/DON 8.2 13.9 6.3

Non-italicised number: Mean of weekly analysis 2006–2007.

Italicised numbers: Mean values obtained for samples used throughout the experiments.

w a t e r r e s e a r c h 4 3 ( 2 0 0 9 ) 5 5 3 – 5 6 1 555

concentrate was made with 1:1 mole of glucose and glycine

with a buffer of 0.5 M NaCO3. The solution was heated for 3 h

at 121 �C (Bernardo et al., 1997). The concentrate was stored

below 4 �C. The concentration of the synthetic melanoidin

solution was chosen to have the same DOC content to that of

the Wetalla STP effluent (20 mg L�1).

The two STP effluentswere sampled before disinfection. The

samples were stored in polyethylene containers and refriger-

ated. Both Wetalla and Merrimac STPs are biological nutrient

removal facilities with total phosphorous (TP) and nitrogen (TN)

limits of 1 and 5 mg L�1. At the time of sampling, both plants

were able to reach the TP limit. However, Wetalla had difficulty

in achieving its TN limit (Table 1), due to its high organic

nitrogen content. It is also clear that Wetalla had excessive

effluent colour in comparison to Merrimac, deemed to be

a result of an industrial feed of spent molasses wastewater.

2.2. Analytical

All samples were filtered using 0.45 mm syringe filters before

analysis to remove all non-dissolved solids, which affect the

various analytical techniques used.

Characteristic colour intensity was recorded in platinum–

cobalt (PtCo) units. A ThermoSpectronic (Helios Beta) spectro-

photometer at a wavelength of 475 nm was used to determine

the absorbance in a 1 cm path length cell. The absorbance at

this wavelength was characteristic of brown colour. The

absorbance value was compared to that of a platinum–cobalt

set of standards and PtCo units were inferred.

DON was calculated to be the difference between filtered

(0.45 mm Millipore express filters) total kjeldahl nitrogen (TKN)

and NH4–N nitrogen. TKN was measured using a Lachat

QuickChem method 10-107-06-2-D. Ammonium nitrogen was

measured on a Lachat flow injection analyser as per the

Lachat QuickChem method 31-107-06-1-A.

DOC was calculated as the difference between the dis-

solved total organic carbon and dissolved inorganic carbon

measurements of the filtered samples, by the high tempera-

ture combustion method (Standard Method 5310 B) (APHA,

1998). A Dohrmann DC-190 was used for analysis.

UV absorbance was measured on a ThermoSpectonic

(Helios Beta) spectrophotometer at 254 nm in a quartz 1 cm

path length cell.

Fluorescence intensity was measured on a Hitachi F-2000

fluorescence spectrophotometer. Twenty-four individual

emission spectra were collected at excitation wavelengths

from 220 to 450 nm, 10 nm apart.

2.3. Jar test procedure

Hydrated aluminium sulphate (Merck) was used as the alum

source. Rapid mixing in a 2 L reaction vessel at 200 rpm was

performed for 2 min during which acid (1 M H2SO4) or base

(1 M NaOH) was added to maintain the required pH.

Aluminium sulphate was dosed from 0 to 100 mg L�1 as

aluminium. The solution was allowed to further coagulate and

flocculate during a 20 min period of stirring at 30 rpm. A

60 min settling period followed. End point samples of 50 mL

(250 mL when molecular weight fraction was required) were

filtered using 0.45 mm Millipore express filters and stored

below 4 �C if not analysed immediately.

2.4. Molecular weight fractionation

The dissolved organic compounds were fractionated using

molecular sieves. Each fraction was analysed for DON, colour,

DOC, Abs254, Abs475, and fluorescence. Fractionation of samples

was performed using a 200 mL stirred cell (Amicon 8200; Milli-

ore Corp. Ma). Three membranes made of regenerated cellulose

with different molecular weight cut-offs were used: (1) PL 10 000

nominal molecular weight limit (NMWL), (2) 5000 NWML and (3)

1000 NMWL. The initial volume was 200 mL. The samples were

filtered through the membranes in series from (1) to (3); each

time a measured volume of 50 mL was retained for analysis,

and the permeate was passed through the next membrane.

DOC and DON were removed from the membranes by pre-

rinsing in accordance with the manufacturer’s instructions.

Volumetric losses were negligible due to the use of wet

membranes and insignificant hold-up volumes. The concen-

tration of colour, DON, DOC, UVA254 and fluorescence intensity

in each size range was calculated as follows:

½< 1 kDa� ¼ ½1 kDaPermeate� (1)

½1–5 kDa� ¼ ½1 kDaConcentrate� � ½< 1 kDa�2

(2)

Fig. 1 – Colour and DON removal at various aluminium

doses from the synthetic melanoidin solution.

w a t e r r e s e a r c h 4 3 ( 2 0 0 9 ) 5 5 3 – 5 6 1556

½5–10 kDa� ¼ ½5 kDaConcentrate� � ½1–5 kDa� � ½< 1kDa�3

(3)

½>10kDa�¼½10kDaConcentrate��½5–10kDa��½1–5kDa��½<1kDa�4

(4)

The percentage difference between the measured total mass

and the sum of the masses for each mass fraction was

calculated.

2.5. Fluorescence

Fluorescence contouring excitation-emission matrices (EEM)

are a compilation of emission scans at numerous excitation

wavelengths. EEM has been used widely in recent years for

a number of applications including:

� Qualification and quantification of the make-up of different

types of NOM in different water related environments

(Baker, 2001; Chen et al., 2003; Her et al., 2003; Wu et al.,

2003; Maie et al., 2007)

� A means of classifying the removal of different NOM origi-

nating compounds during various water treatments

methods (Sharpless and McGown, 1999; Saadi et al., 2006)

� Differentiating between humics of different origins

including synthetic, oceanic and terrestrial humics (Coble,

1996; Westerhoff et al., 2001; Seredynska-Sobecka et al.,

2007)

The peak fluorescence intensity obtained lExmax=lEmmax is

a characteristic property for different types of organic

compounds. The location of lExmax=lEmmax is different for

aromatic protein, fulvic-acid-like, soluble microbial product

and humic-acid-like compounds (Chen et al., 2003). Peak

locations of humic compounds differed depending on their

origin (Coble, 1996), with the identification of shallow marine

humic compounds with peak locations at much lower

lExmax=lEmmax than terrestrial and deep marine humics. That

study also found that the melanoidin type humic peak loca-

tions were at higher excitation and emission than the

synthetic humics and all other humic compound containing

source waters.

2.6. Statistical analysis

Eight replicates were analysed with dilution to within the

measurement range for the various analytical measurements.

Colour, DOC, TKN, NH4 and UVA measured values, were

diluted to have concentrations within the range 1–500 mg-

PtCo L�1, 1–100 mg L�1, 0.01–20 mg L�1 and 0.01–15 mg L�1

0.01–1.0 a.u. respectively. The results showed for single

samples the 95% confidence intervals were� 4.5%, � 3.1%, �2.0%, � 2.5% and� 3.8% respectively.

The fluorescence excitation-emission matrix measures

fluorescence intensity at 10.0 nm intervals, thus the location

of the peak is only accurate to� 5 nm. The error bars shown in

Fig. 3 are representative of this.

Analysis of DON was compared to the summation of the

dissolved organic nitrogen fractions. The average discrepancy

of the summation was 8.7% and did not exceed 18% of the total

nitrogen measures. Likewise, analysis of the total dissolved

organic carbon was compared to the summation of the dis-

solved organic carbon fractions. The average discrepancy of

the summations was only 6.2% and did not exceed 23% of the

total dissolved organic carbon measure.

3. Results and discussion

3.1. Removal of DOC, nitrogen, colour and UVA254

The optimum pH for colour removal from the synthetic mel-

anoidin solution was statistically similar for pH 5 and 6 for

applied alum doses of 10, 20 and 50 mg L�1 (Fig. 1). The

optimum DOC removal was at pH 5. The difference between

colour and DOC removal is probably a result of the presence of

other non-coloured organic Maillard reaction products in the

melanoidin solution. It is likely that these required a slightly

lower pH optimum for removal compared to the melanoidin

organics. The pH used for subsequent optimisation and

characterisation studies was pH 6.

Fig. 2 shows colour, DON, DOC and UVA254 remaining in

solution after the coagulation for the three wastewaters at pH

6. UVA254 is represented in these results instead of SUVA as

a means of showing the change in concentration of double

bond independently of carbon content. The aluminium doses

were standardised by dividing the concentration of

aluminium ions by the DOC content of each jar test.

The optimum alum dose for colour, DON, DOC and UVA254

removal was considered to be the minimum aluminium dose

required to reach the maximum removal of the various

constituents. For the fractionation experiments an aluminium

dose of 30 mg L�1 was used for all experiments. This dose was

at or in excess of the optimum dose required for the removal

of all colour, DOC, DON and UVA254 from all the wastewater. It

corresponded to standardised alum dose (aluminium dose/

DOC) of 2.2, 1.4 and 1.6 for Merrimac STP effluent, Wetalla STP

effluent and the synthetic melanoidin solution respectively

(Fig. 2). At this dose, the percentage colour removal was high

for all wastewaters, with 73, 75 and 89% colour removal from

Merrimac, Wetalla and the synthetic melanoidin solution

Fig. 2 – Removal of colour, DON, DOC and UVA254 after coagulation at various aluminium doses and pH 6.

w a t e r r e s e a r c h 4 3 ( 2 0 0 9 ) 5 5 3 – 5 6 1 557

respectively. The amount of colour removed from both the

synthetic melanoidin solution (215 mg-PtCo L�1) and Wetalla

effluent (127 mg-PtCo L�1) was high for the same dose of alum

(30 mg L�1) compared to the removal of colour from Merrimac

effluent (22 mg-PtCo L�1), indicating that alum was particu-

larly efficient at removing melanoidin associated colour,

when present.

The mass of organic removed relative to the mass of colour

removal was lower when melanoidin was present. At the

selected aluminium dose (30 mg L�1), DON removal was 14, 8.7

and 5.5 mg-DON g-PtCo�1 for Merrimac, Wetalla and the

synthetic melanoidin solution respectively. Likewise, the DOC

removal was 196, 49 and 29 mg-DON g-PtCo�1. This indicated

that when melanoidin was present, a much higher concen-

tration of colour was removed using alum coagulation with

respect to the total organic composition.

High SUVA values indicate a high content of organic

compounds with double bonds and the presence of hydro-

phobic, humic-like compounds (Lee et al., 2006). SUVA values

in natural environments range from below 1 to higher than

4 L mg�1 cm�1. The SUVA results are given in Table 2. Values

for the melanoidin containing wastewater were high (2.96 and

2.51 L mg�1 cm�1 for the synthetic melanoidin solution and

Wetalla SPT effluent respectively), compared to Merrimac STP

effluent (1.61 L mg�1 cm�1). This is because of the highly

aromatic nature of melanoidins.

3.2. Change in molecular weights

Molecular weight fractionation was performed before and

after alum coagulation using the selected aluminium dose

(30 mg L�1) for the synthetic melanoidin solution, Wetalla STP

effluent (contained melanoidin) and Merrimac STP effluent

(no melanoidin). The results are summarised in Table 2. The

objective was to identify the molecular size fraction that

contained melanoidin and compare alum induced aggrega-

tion of melanoidin to the aggregation of other organic matter.

The synthetic melanoidin solution was used as a control to

analyse the properties of melanoidin during coagulation.

Melanoidin was shown to reside in the>10 kDa fraction of the

synthetic melanoidin solution, which was expected as mela-

noidins have molecular weights of around 50 kDa (Pena et al.,

2003; Coca et al., 2005a). DON and DOC were also most preva-

lent in the >10 kDa fraction. The DON and DOC present in the

lower molecular weight fractions were other Maillard reaction

products, residual glucose and residual glycine, but not mela-

noidins due to the absence of colour. Alum coagulation caused

extensive colour removal from the >10 kDa molecular weight

fraction of the synthetic melanoidin solution with a decrease

from 203 to 11 mg-PtCo L�1. DOC and DON were also reduced

from 10.4 to 3.1 mg L�1 and 1.5 to 0.4 mg L�1 respectively in the

>10 kDa fraction, but no significant reduction occurred to the

<10 kDa fractions, indicating that alum was effective in

removing melanoidin from organic matter mixtures.

Wetalla and Merrimac STP effluents were hypothesised to

contain similar levels of ‘background’ organic compounds,

apart from the presence of melanoidin in the Wetalla STP

effluent. A comparison of the >10 kDa molecular weight

fractions of the STP effluents highlighted the presence of

melanoidin in Wetalla STP effluent, with elevated colour

content (95 compared to 6 mg-PtCo L�1), DON (0.8 compared to

0.3 mg L�1) and DOC (7.1 compared to 1.6 mg L�1) in Wetalla

compared to Merrimac STP effluent. The colour, DON and DOC

content of <10 kDa molecular weight fractions were similar

for both STP effluents.

Alum coagulation showed marked removal of the colour,

and thus melanoidin, associated with the >10 kDa molecular

weight fractions from STP effluent when present. Reduction in

colour (95 to 11 mg-PtCo L�1), DOC (7.1 to 2.9 mg L�1) and DON

(0.8 to 0.2 mg L�1) were also significant in that fraction. These

reductions were smaller compared to the reduction in the

>10 kDa fraction of the synthetic melanoidin, but were

comparable considering the initial concentration of colour

was far less for Wetalla STP effluent. In comparison, colour,

Ta

ble

2–

Mo

lecu

lar

weig

ht

fra

ctio

na

tio

no

fco

lou

r,D

ON

,D

OC

an

dS

UV

Ab

efo

rea

nd

aft

er

alu

mco

ag

ula

tio

n.

Co

lou

rm

g-P

tCo

L�

1D

ON

mg

L�

1N

H4�

Nm

gL�

1D

OC

mg

L�

1S

UV

AL

mg�

1cm

�1

DO

C/D

ON

Alu

min

ium

Do

se–

30

mg

L�

1–

30

mg

L�

1–

30

mg

L�

1–

30

mg

L�

1–

30

mg

L�

1–

30

mg

L�

1

Sy

nth

eti

cM

ela

no

idin

>10

kD

a203

11

1.5

30.4

20.1

4<

0.1

10.3

53.1

03.5

91.6

96.8

7.4

5–1

0k

Da

23

70.2

30.2

1<

0.1

<0.1

1.9

71.9

93.5

22.4

48.5

9.6

1–5

kD

a13

60.2

50.2

9<

0.1

<0.1

2.0

12.7

71.2

21.8

38.1

9.7

<1

kD

a4

41.0

70.9

90.1

0<

0.1

5.0

65.2

20.4

30.6

94.7

5.3

Tot

al

242

27

3.0

81.9

00.3

00.1

719.3

913.0

82.5

11.4

46.3

6.9

Weta

lla

WW

TP

Effl

uen

t>

10

kD

a95

11

0.7

50.2

4<

0.1

<0.1

7.1

02.8

82.9

51.3

79.4

12.1

5–1

0k

Da

42

10

0.3

70.3

7<

0.1

0.1

34.0

92.7

63.4

22.2

311.0

7.4

1–5

kD

a24

13

0.4

50.1

6<

0.1

<0.1

4.4

03.5

83.0

52.3

89.9

22.3

<1

kD

a9

91.0

60.7

5<

0.1

0.5

05.1

25.3

32.5

31.9

74.9

7.1

Tot

al

170

43

2.6

31.5

20.3

00.6

520.7

114.5

52.9

62.0

07.9

9.6

Merr

ima

cW

WT

PE

fflu

en

t>

10

kD

a6

30.2

70.1

7<

0.1

0.1

31.6

51.6

61.6

51.0

06.1

9.8

5–1

0k

Da

17

40.3

6<

0.1

0.1

2<

0.1

3.8

01.7

32.2

51.3

010.6

–

1–5

kD

a7

2<

0.1

0.2

2<

0.1

<0.1

3.4

02.3

41.5

32.4

5–

10.6

<1

kD

a0

00.3

00.2

20.4

80.4

84.4

93.2

81.1

01.7

615.0

15.2

Tot

al

30

90.9

20.6

10.6

40.6

113.3

39.0

11.6

11.7

114.4

14.9

No

va

lue

sho

wn

ifD

ON

or

DO

Cb

elo

w0.1

.

w a t e r r e s e a r c h 4 3 ( 2 0 0 9 ) 5 5 3 – 5 6 1558

DON and DOC reduction were minimal in the largest molec-

ular weight fraction of Merrimac STP effluent as it contained

no melanoidin. The DOC and DON reduction from the<10 kDa

fraction was minimal for both Merrimac and Wetalla STP

effluents.

Wetalla STP effluent >10 kDa molecular weight fraction

had a high SUVA value (2.95 L mg�1 cm�1) for low DOC/DON

(9.4). Alum coagulation caused the SUVA value to reduce to

1.37 L mg�1 cm�1 for DOC/DON value of 12.1. This change in

DOC/DON is not considered to be significant as DOC/DON

values have been shown to range from 5 to 32 mg L�1 in raw

waters (Lee et al., 2006). This change in the SUVA value, with

a stable DOC/DON value, represents NOM with less conjuga-

tion, as expected for melanoidin. The SUVA value of Wetalla

STP effluent post-coagulation was similar to that of Merrimac

STP effluent, which was much less affected by coagulation.

The synthetic melanoidin solution contained a <1 kDa

DOC/DON ratio that was lower than the other water sources.

The melanoidin solution was manufactured by heating sugar

and amino acid. A fraction of the amino acid will remain

unreacted and will be detected in the<1 kDa molecular weight

fraction. This results in the low DOC/DON value observed.

The coagulation of Wetalla STP effluent compared to the

synthetic melanoidin solution resulted in slightly higher DOC

per colour removal (50 compared to 38 mg g-PtCo�1) and DON

per colour removal (6.1 compared to 5.8 mg g-PtCo�1). This

was reasonable assuming Wetalla STP effluent contained

other NOM in the >10 kDa fraction. Fluorescence EEM was

used to determine if these were other humic substances.

3.3. Fluorescence change with melanoidin removal

The objective of using fluorescence EEM was to determine if

coagulation preferentially removed melanoidin rather than

other humic organic compounds. The colour results of the

>10 kDa fraction indicated that melanoidins were effectively

removed from Wetalla STP effluent. However, residual DON

and DOC content after alum coagulation indicated that

residual NOM contributing compounds remained. EEM was

used to establish if these compounds were residual melanoi-

dins or other humic compounds.

Coble (1996) performed an extensive EEM study on various

humic containing wastewaters. A vast variation in peak

location of humic substances occurred depending on their

origin (Fig. 2). Increased lExmax=lEMmax values indicated the

presence of compounds with increased aromaticity,

compounds with conjugated double bonds and carboxyl,

hydroxyl and amine containing compounds (Senesi, 1990),

which explains why synthetic melanoidin had higher

lExmax=lEMmax than the other synthetic humic isolates tested, in

fact the highest of all humic compounds tested (Coble, 1996).

Dwyer and Lant (2008) showed that during hydroxyl radical

oxidation of melanoidin, the concentration of colour, DOC and

DON decreased, while the location of the EEM peak moved

from 361/345 to 337/421, further highlighting how the removal

of melanoidin showed a shift to lower lExmax=lEMmax . The EEM

peak location of >10 kDa fraction of Wetalla STP effluent,

Merrimac STP effluent and the synthetic melanoidin solution

before and after alum coagulation are plotted in Fig. 3. The

Fig. 3 – Position of wavelength independent lEXmax and lEMmax for humic samples examined in (a) literature and (b) this

study.

w a t e r r e s e a r c h 4 3 ( 2 0 0 9 ) 5 5 3 – 5 6 1 559

lExmax=lEMmax peak locations are consistent with the observed

data obtained by Coble (1996).

As expected, the synthetic melanoidin solution had the

highest lExmax=lEMmax values. No significant lExmax=lEMmax shift

occurs to the synthetic melanoidin solution after alum-

induced coagulation, indicating a small residual concentra-

tion of melanoidin remains after coagulation.

Sewage treatment plant effluent contains many different

humic compounds, with different fluorescent properties,

which combine to give a prominent lExmax=lEMmax peak.

lExmax=lEMmax values of the >10 kDa fraction of Wetalla STP

effluent (361/439 nm) were significantly higher than Merrimac

STP effluent (348/434 nm). The high lExmax=lEMmax was indica-

tive of the presence of a significant content of melanoidin in

the >10kD fraction, with the Wetalla STP effluent peak loca-

tion close to that of the synthetic melanoidin solution (361/

445 nm). Merrimac peak location was not significantly

changed after alum coagulation. However, coagulation of

Wetalla STP effluent caused a significant change in peak

location to 346/428 nm. This showed that melanoidin type

humics were favoured for removal by alum coagulation when

present in STP effluent over other humic substances. More

significantly this indicated that non-melanoidin humic

compounds remained after alum coagulation. Also, the

residual humic peak location of Merrimac (344/431 nm) and

Wetalla (346/428 nm) STP effluent was not significantly

Table 3 – Change in fluorescence peak intensities of with resp

Fluorescence Intensity

Synthetic Melanoidin Wastewater 0 12,250

30 mg L�1 4200

Wetalla WWTP Effluent 0 2550

30 mg L�1 1000

Merrimac WWTP Effluent 0 550

30 mg L�1 400

different after alum coagulation implying that the residual

humics were similar for both STP effluents.

Fluorescence intensity at the peak locations before and

after alum coagulation of the >10 kDa fractions is provided in

Table 3 as is the intensity/DOC and intensity/UVA measure-

ments. The impact of alum on the fluorescence intensity/DOC

was minimal for all the wastewater, while the fluorescence

intensity/UVA increased significantly for the synthetic mela-

noidin and Wetalla wastewater, but showed little change for

Merrimac wastewater. These results indicated that alum tar-

geted the fluorescing organics, which had carbon double

bonds. As melanoidin is known to be highly aromatic, it

follows that alum targets melanoidin removal over other

organics when melanoidin is present.

4. Conclusions

� Molecular weight fractionation highlighted that melanoidin

had a molecular weight >10 kDa with a majority of the

colour in that fraction. When melanoidin associated colour

was present in wastewater, coagulation reduced the DOC

and DON content by 38 and 5.8 mg g-PtCo�1 L�1 respectively.

� Alum coagulation of the Wetalla STP effluent (containing

melanoidin) showed a significant decrease in lExmax=lEMmax ,

while coagulation of Merrimac STP effluent (containing no

ect to DOC and UVA.

au Fluorescence/DOC au L mg�1 Fluorescence/UVA au au�1

1180 33,000

1350 80,300

350 12,100

350 25,200

300 20,800

250 23,700

w a t e r r e s e a r c h 4 3 ( 2 0 0 9 ) 5 5 3 – 5 6 1560

melanoidin) showed no significant change in peak position,

highlighting that melanoidin type humics were preferen-

tially removed, with residual, non-melanoidin humics

remaining after coagulation.

� An alum dose of 30 mg L�1 was required to obtain maximum

colour, DON and DOC removal from the STP effluents as the

imposed concentrations.

� No significant removal of colour, DOC and DON occurred in

the <10 kDa molecular weight fractions using 30 mg L�1 of

aluminium for coagulation from STP effluents, highlighting

that alum was more effective at removing compounds

which had MW >10 kDa.

� The peak locations of Wetalla and Merrimac STP effluents

after coagulation were similar, indicating that the residual

humic content was similar, thus alum coagulation at

30 mg L�1 was not able to remove all humic compounds.

Acknowledgements

This work is funded by Toowoomba City Council (Australia)

and the Australian Research Council (Project LP0453685). All

experiments were conducted in the Advanced Water

Management Centre Laboratories at the University of

Queensland. Aravinthan Vijayaragavan made a significant

contribution to the experimental work.

r e f e r e n c e s

Amy, G., 2007. Natural Organic Matter (NOM) Profiling as a Basisfor Treatment Process Selection and Monitoring. In:Proceedings of the Fourth Leading-Edge Conference,Singapore (International Water Association, 2007).

APHA, 1998. Standard Methods for the Examination of Water andWastewater. American Public Health Association,Washington, DC.

Baker, A., 2001. Fluorescence excitation-emission matrixcharacterization of some sewage-impacted rivers. Environ.Sci. Technol. 35, 948–953.

Benzingpurdie, L., Ripmeester, J.A., Preston, C.M., 1983. Elucidationof the nitrogen forms in melanoidins and humic-acid by N-15cross polarization magic angle spinning nuclear magnetic-resonance spectroscopy. J. Agric. Food Chem. 31, 913–915.

Bernardo, E.C., Egashira, R., Kawasaki, J., 1997. Decolorization ofmolasses’ wastewater using activated carbon prepared fromcane bagasse. Carbon 35, 1217–1221.

Cammerer, B., Kroh, L.W., 1995. Investigation of the influence ofreaction conditions on the elementary composition ofmelanoidins. Food Chem. 53, 55–59.

Chen, W., Westerhoff, P., Leenheer, J.A., Booksh, K., 2003.Fluorescence excitation-emission matrix regional integrationto quantify spectra for dissolved organic matter. Environ. Sci.Technol. 37, 5701–5710.

Coble, P.G., 1996. Characterization of marine and terrestrial domin seawater using excitation emission matrix spectroscopy.Mar. Chem. 51, 325–346.

Coca, M., Pena, M., Gonzalez, G., 2005a. Chemical oxidationprocesses for decolorization of brown-colored molasseswastewater. Ozone Sci. Eng. 27, 365–369.

Coca, M., Pena, M., Gonzalez, G., 2005b. Variables affectingefficiency of molasses fermentation wastewater ozonation.Chemosphere 60, 1408–1415.

Dignac, M.F., Ginestet, P., Rybacki, D., Bruchet, A., Urbain, V.,Scribe, P., 2000. Fate of wastewater organic pollution duringactivated sludge treatment: nature of residual organic matter.Water Res. 34, 4185–4194.

Dwyer, J., Lant, P., 2008. Biodegradability of DOC and DON forUV/H2O2 pre-treated melanoidin based wastewater. Biochem.Eng. J., doi:10.1016/j.bej.2008.05.016.

Egeberg, P.K., Eikenes, M., Gjessing, E.T., 1999. Organic nitrogendistribution in nom size classes. Environ. Int. 25, 225–236.

Hayase, F., Kim, S.B., Kato, H., 1986. Analyses of the chemicalstructures of melanoidins by c-13 nmr, c-13 and n-15 cp-masnmr spectrometry. Agric. Biol. Chem. 50, 1951–1957.

Her, N., Amy, G., McKnight, D., Sohn, J., Yoon, Y.M., 2003.Characterization of dom as a function of mw by fluorescenceeem and hplc-sec using uva, doc, and fluorescence detection.Water Res. 37, 4295–4303.

Kato, H., Tsuchida, H., 1981. Estimation of melanoidin structureby pyrolysis and oxidation. Prog. Food Nutr. Sci. 5, 147–156.

Lee, W., Westerhoff, P., Esparza-Soto, M., 2006. Occurrence andremoval of dissolved organic nitrogen in us water treatmentplants. J. AWWA 98, 102.

Levic, L., Gyura, J., 2000. Influence of aluminium sulphateconcentration on the change of electrokinetic potential ofmacromolecular compounds in molasses. Nahrung 44, 288–289.

Levic, L., Gyura, J., Djuric, M., Kuljanin, T., 2005. Optimizationof ph value and aluminium sulphate quantity in the chemicaltreatment of molasses. Eur. Food Res. Technol. 220, 70–73.

Maie, N., Scully, N.M., Pisani, O., Jaffe, R., 2007. Composition ofa protein-like fluorophore of dissolved organic matter in coastalwetland and estuarine ecosystems. Water Res. 41, 563–570.

Martins, S., Jongen, W.M.F., van Boekel, M., 2000. A review ofmaillard reaction in food and implications to kineticmodelling. Trends Food Sci. Technol. 11, 364–373.

Migo, V.P., DelRosario, E.J., Matsumura, M., 1997. Flocculation ofmelanoidins induced by inorganic ions. J. Ferment. Bioeng. 83,287–291.

Migo, V.P., Matsumura, M., Del Rosaria, E.J., Kataoka, H., 1993.Decolorization of molasses wastewater using an inorganicflocculant. J. Ferment. Bioeng. 75, 438–442.

Motai, H., 1974. Oxidative browning of melanoidin.2. Relationshipbetween molecular-weight and color intensity of colorcomponents of melanoidin from glycine–xylose system. Agric.Biol. Chem. 38, 2299–2304.

Motai, H., Inoue, S., 1974. Oxidative browning of melanoidin.1.Conversion of color components of melanoidin produced fromglycine–xylose system. Agric. Biol. Chem. 38, 233–239.

Pehlivanoglu-Mantas, E., Sedlak, D.L., 2006. Wastewater-deriveddissolved organic nitrogen: analytical methods,characterization, and effects – a review. Crit. Rev. Environ. Sci.Technol. 36, 261–285.

Pena, M., Coca, M., Gonzalez, G., Rioja, R., Garcia, M.T., 2003.Chemical oxidation of wastewater from molassesfermentation with ozone. Chemosphere 51, 893–900.

Richardson, S.D., 2003. Disinfection by-products and otheremerging contaminants in drinking water. Trac – TrendsAnalyt. Chem. 22, 666–684.

Saadi, I., Borisover, M., Armon, R., Laor, Y., 2006. Monitoring ofeffluent dom biodegradation using fluorescence, uv and docmeasurements. Chemosphere 63, 530–539.

Schrank, S.G., Jose, H.J., Moreira, R., Schroder, H.F., 2004.Comparison of different advanced oxidation process to reducetoxicity and mineralisation of tannery wastewater. Water Sci.Technol. 50, 329–334.

Senesi, N., 1990. Molecular and quantitative aspects of thechemistry of fulvic-acid and its interactions with metal-ions

w a t e r r e s e a r c h 4 3 ( 2 0 0 9 ) 5 5 3 – 5 6 1 561

and organic-chemicals.2. The fluorescence spectroscopyapproach. Anal. Chim. Acta 232, 77–106.

Seredynska-Sobecka, B., Baker, A., Lead, J., 2007. Characterisationof colloidal and particulate organic carbon in freshwaters bythermal fluorescence quenching. Water Res. 41, 3069–3076.

Sharpless, C.M., McGown, L.B., 1999. Effects of aluminum-induced aggregation on the fluorescence of humic substances.Environ. Sci. Technol. 33, 3264–3270.

Sirivedhin, T., Gray, K.A., 2005. 2. Comparison of the disinfectionby-product formation potentials between a wastewatereffluent and surface waters. Water Res. 39, 1025–1036.

Westerhoff, P., Chen, W., Esparza, M., 2001. Fluorescence analysisof a standard fulvic acid and tertiary treated wastewater.J. Environ. Qual. 30, 2037–2046.

Westerhoff, P., Mash, H., 2002. Dissolved organic nitrogen indrinking water supplies: a review. J. Water Sup. Res. Technol.Aqua 51, 415–448.

Wu, F.C., Evans, R.D., Dillon, P.J., 2003. Separation andcharacterization of nom by high-performance liquidchromatography and on-line three-dimensional excitationemission matrix fluorescence detection. Environ. Sci. Technol.37, 3687–3693.