ww.sciencedirect.com

wat e r r e s e a r c h 4 6 ( 2 0 1 2 ) 4 0 6 3e4 0 7 0

Available online at w

journal homepage: www.elsevier .com/locate/watres

Laboratory study on the adsorption of Mn2D on suspendedand deposited amorphous Al(OH)3 in drinking waterdistribution systems

Wendong Wang a,*, Xiaoni Zhang a, Hongping Wang b, Xiaochang Wang a, Lichuan Zhou a,Rui Liu c, Yuting Liang d

a School of Environmental and Municipal Engineering, Xi’an University of Architecture and Technology, Xi’an, Shaanxi 710055, ChinabQujiang Drinking Water Purification Plant, Water Industry Operations Company of Xi’an, Xi’an 710061, ChinacDepartment of Environmental Technology and Ecology, Yangtze Delta Region Institute of Tsinghua University, Zhejiang, Jiaxing 314006,

ChinadSchool of Environmental and Safety Engineering, Changzhou University, Changzhou 213164, China

a r t i c l e i n f o

Article history:

Received 18 November 2011

Received in revised form

6 May 2012

Accepted 10 May 2012

Available online 18 May 2012

Keywords:

Aluminum hydroxide

Drinking water

Pipe scale

Water quality

* Corresponding author. Department of EnviYanta Road, Xi’an, Shaanxi 710055, China. T

E-mail address: wwd@xauat.edu.cn (W. W0043-1354/$ e see front matter ª 2012 Elsevdoi:10.1016/j.watres.2012.05.017

a b s t r a c t

Manganese (II) is commonly present in drinking water. This paper mainly focuses on the

adsorption of manganese on suspended and deposited amorphous Al(OH)3 solids. The

effects of water flow rate and water quality parameters, including solution pH and the

concentrations of Mn2þ, humic acid, and co-existing cations on adsorption were investi-

gated. It was found that chemical adsorption mainly took place in drinking water with pHs

above 7.5; suspended Al(OH)3 showed strong adsorption capacity for Mn2þ. When the total

Mn2þ input was 3 mg/L, 1.0 g solid could accumulate approximately 24.0 mg of Mn2þ at

15 �C. In drinking water with pHs below 7.5, because of Hþ inhibition, active reaction sites

on amorphous Al(OH)3 surface were much less. The adsorption of Mn2þ on Al(OH)3 changed

gradually from chemical coordination to physical adsorption. In drinking water with high

concentrations of Ca2þ, Mg2þ, Fe3þ, and HA, the removal of Mn2þ was enhanced due to the

effects of co-precipitation and adsorption. In solution with 1.0 mg/L HA, the residual

concentration of Mn2þ was below 0.005 mg/L, much lower than the limit value required by

the Chinese Standard for Drinking Water Quality. Unlike suspended Al(OH)3, deposited

Al(OH)3 had a much lower adsorption capacity of 0.85 mg/g, and the variation in flow rate

andmajor water quality parameters had little effect on it. Improved managements of water

age, pipe flushing and mechanical cleaning were suggested to control residual Mn2þ.

ª 2012 Elsevier Ltd. All rights reserved.

1. Introduction Environment and Health/Institute of Occupational Medicine,

Manganese (Mn) is an element that is ubiquitous in the envi-

ronment. It is commonly found in drinking water and is

essential for humanhealth at low concentrations (Institute for

ronmental and Civil Engiel.: þ86 135 7254 7081; faang).

ier Ltd. All rights reserved

2004). However, excessive concentrations of Mn could result

inmetallic tasting water andmany health problems. Jose et al.

(2009) found that excessive intake of manganese could cause

nervous system damage, leading to Parkinson’s disease, and

neering, Xi’an University of Architecture and Technology, No. 13,x: þ86 29 8220 2729.

.

wat e r r e s e a r c h 4 6 ( 2 0 1 2 ) 4 0 6 3e4 0 7 04064

injure arterial breastwork and the myocardium. Wasserman

et al. (2006) identified a significant negative relationship

betweenMn2þ and the intellectual function of 142 10-years-old

children drinking well water. In addition, the accumulation of

manganese-containing sediments can lead to reduced water

pressure and flow in pipes (Sly et al., 1990; USEPA, 2008). The

World Health Organization has set a guideline of 0.4 mg/L for

manganese intake control from drinking water (World Health

Organization, 2008); and the limit value in the Chinese Stan-

dard for DrinkingWater Quality (GB5749-2006) is 0.1 mg/L.

Manganese primarily exists in the form of divalent

manganese cation (Mn2þ) in drinking water. In neutral and

alkaline environments, Mn2þ is difficult to be oxidized

(Sommerfeld, 1999). Although chlorine or KMnO4 oxidation

combined with sand filtration is usually applied for Mn2þ

removal, its concentration in the treated water is still difficult

to meet the local requirements (Seppanen, 1993; Mo, 2010). It

was presumed that the concentration ofMn2þwas constant as

the water passes through pipelines (Schock, 2005). Conse-

quently, regulatory monitoring of Mn2þ is only required at

entry points of the drinkingwater distribution system (DWDS),

presuming that its concentration could not increase and that

suchmonitoring is sufficiently toprotectpublichealth (Schock,

2005; Wang et al., 2011). In fact, as the effects of pollutants

accumulationand sedimentation, the concentrationofMn2þ is

variable (USEPA, 2006), and the variation of water quality such

as solution pHmay cause a sudden release of Mn2þ.Aluminum-containing scale commonly exists in the DWDS

(Snoeyink et al., 2003; Wang et al., 2010). Many studies have

found that amorphous Al(OH)3 exists a good adsorption

capacity to many pollutants (Schock and Holm, 2003). Julien

et al. (1994) studied the adsorption of organic acids to pre-

formed Al(OH)3 and found a correlation between removal

efficiency and the number and ionization of functional

groups. Generally, the removal efficiency increased with

a greater number of functional groups on the organic acids

and thus with greatermolecular weight. Sudden reductions in

lead levelswere observed during pipe rig testing in a Rochester

case study that was coincided with the opening of a new

filtration plant and appeared to be related to aluminum

deposition on lead piping materials (Kirmeyer et al., 1999).

However, the adsorption of Mn2þ on amorphous Al(OH)3 has

not yet been considered in the literature.

The objective of this paper was to (1) study the adsorption

of Mn2þ in DWDS with suspended and deposited amorphous

Table 1 e Effecting factors and their levels selected in the expe

Experiments Para. Units

Adsorption of Mn2þ on Al(OH)3 scale pH e

Mn2þ mg/L

v mL/s

Adsorption of Mn2þ on Al(OH)3 suspension Ca2þ mg/L

Mg2þ mg/L

Fe3þ mg/L

HA mg/L

Para., parameters; Var., variable.

Al(OH)3 formation; (2) investigate the effects of major water

quality parameters including solution pH, dissolved organic

matter, and co-existing cations on the adsorption process; and

(3) reveal the reaction mechanism between amorphous

Al(OH)3 and Mn2þ. On the basis of the conclusions obtained in

this study, some DWDS operation management suggestions

were also given.

2. Experimental materials and methods

2.1. Adsorption of Mn2þ on suspended Al(OH)3

0.20 g of air-dried Al(OH)3 made from the hydrolysis of

Al2(SO4)3 at pH 6.5 was put into 500 mL test tubes with 240 mL

of deionized water. Then, 10.0 mL of 25.0 mg/L MnCl2 was

added. The tubes were stirred using a magnetic stirrer for

5 min. Then, the mixtures were sampled and centrifuged at

5000 rpm for 20 min. The amount of residual Mn2þ was

determined by analyzing the supernatant. The sorption coef-

ficient (Kd) was calculated as the ratio of the amount of Mn2þ

being adsorbed to the solid and the amount of residual Mn2þ

in the water at equilibrium, as shown in Eq. (1) (Dzombak and

Morel, 1990):

¼ AlOH0:5� þMn2þ/ ¼ AlOMn0:5þ þHþ (1)

The effects of the primary co-existing ions, including

calcium, magnesium, iron, and humic acids (HA), on the

adsorption of Mn2þ were investigated. HA was filtrated with

a 0.45 mm polycarbonate membrane to remove the insoluble

matter, and it was subsequently filtered with an ultrafiltration

membrane with a 500 Da molecular weight cutoff. In this

experiment, all thewater quality parametersweremaintained

at the values in the ranges of actual drinking water (Table 1).

The co-existing cations were added three times, correspond-

ing to their low, middle, and high levels in actual drinking

water respectively. Solution pH was adjusted with 0.50 mol/L

NaOH and 0.50 mg/L HNO3.

2.2. Manganese (II) adsorption on deposited amorphousAl(OH)3 scale

A laboratory scale DWDS was set up as shown in Fig. 1.

Drinking water was drawn from a tank and pumped to the

riments (15 �C).

Mn2þ pH Selected level

Low Middle High

1.0 Var. 6.5 7.5 8.5

Var. 7.5 0.0 1.0 2.0

1.0 7.5 0.2 0.5 1.0

1.0 7.5 20.0 60.0 100.0

1.0 7.5 5.0 25.0 50.0

1.0 7.5 0.1 0.3 0.5

1.0 7.5 1.0 5.0 10.0

water tankInfluent

Pump

Pressure gauge

Pipe system Discharge

Fig. 1 e Drinking water distribution system used in the

experiment.

wat e r r e s e a r c h 4 6 ( 2 0 1 2 ) 4 0 6 3e4 0 7 0 4065

pipe system. To exclude the effects of othermetals introduced

by pipe corrosion and metal element stripping, polyvinyl

chloride (PVC) pipes were used in the system. Deionized water

was used for modeling drinking water preparation. The water

quality parameters were adjusted by placing specified

volumes of 0.50 mol/L NaOH, 0.50 mg/L HNO3, 100 mg/L

Al2(SO4)3, and 50.0 mg/L MnCl2 into the tank. Reagent grade

chemicals were used except where noted.

The tank was placed in an incubator to maintain the

water temperature at 15 �C. Water samples were taken from

the effluent, and deposited Al(OH)3 were scraped from the

pipeline, respectively. To accelerate the precipitation of

Al(OH)3, the initial pH and aluminum concentration of the

solution were maintained at 6.5 and 50.0 mg/L, respectively.

After 5 days of operation, a white gelatinous pipe coating,

mainly composed of Al(OH)3 was generated. Then, stopped

adding Al2(SO4)3 and introduced 1.0 mg/L Mn2þ to the system

in stage I. To insure that the added metal elements were fully

mixed with the Al(OH)3 scale, the flow velocity was main-

tained at 0.5 mL/s, which was much lower than the normal

value.

When the concentration of Mn2þ residual in the effluent

was stable, the effects of flow velocity, solution pH, and

manganese concentration on Mn2þ adsorption were investi-

gated in stage II (Table 1). Water samples were taken at

different time intervals and filtered through a 0.45 mm poly-

carbonate membrane for water quality determination.

5 10 15 20

0.00

0.02

0.04

0.06

0.08 pH 8.5 pH 7.5 pH 6.5

Reaction time (min)

Res

idua

l Mn2+

(m

g/L

)

C = 1.0 mg/LC = 0.5 mg/LC = 0.1 mg/L

Fig. 2 e Manganese (II) adsorption in solutions with an

amorphous Al(OH)3 suspension.

2.3. Manganese (II) adsorption prediction method

The formation of surface complexes is conceptualized simi-

larly to that happened in solutions. The application of the

surface acid/base reactions could explain the protolytic

behavior of many different oxides (Schindler, 1981). The mass

action laws for surface complexation reactions are also

treated similarly to those involving soluteesolute

interactions.

The amount of Mn2þ adsorbed by suspended amorphous

Al(OH)3 was predicted applying the VISUAL MINTEQ software,

by defining the concentrations of all components, solution pH,

water temperature, ion strength, and adsorption parameters

including the number of the surfaces (1 surface), adsorption

model (Diffused Layer Model), solid concentration (0.05 mg/L),

site concentration (1.0 mmol/mmol), and the sorption coeffi-

cient (Kd) (Pommerenk and Schafran, 2005). Moreover,

manganese speciation including the solids that might

precipitate from solution was also calculated.

2.4. Water quality and solid analysis methods

The concentrations of aluminum, calcium, magnesium, iron,

and manganese were determined using an ICP-AES (IRIS

intrepid-II, Thermo Scientific, USA). Sample aliquots were

digested with trace metal grade nitric acid to pH < 2 for 12 h

prior to analysis. The concentration of HA was determined by

a TOC analyzer (TOC-Vwp, Shimazu Company, Japan); pHwas

determined using a pHmeter (Model 828, ThermoOrion, USA),

temperature was determined with a thermometer (Model

TTM1-JM-6200IM, Yuan-Da Technology Corporation, China).

All parameters were analyzed according to the standard

methods described in GB/T 5750.4-2006 of China.

To observe the surface characteristics of amorphous

Al(OH)3, scanning electron microscopy combined with an

energy dispersive spectrometer (SEM/EDS) (JSM-6490LV, JEOL

Ltd., Japan) and a X-ray diffractometry (Ultima IV, Rigaku

Corporation, Japan) were applied. The SEM/EDS samples were

sputter-coated with gold before conducting surface scanning

and elemental component analysis. To verify the valence

information of manganese, the X-ray photoelectron spec-

troscopy (K-Alpha, Thermo Scientific, USA) of the amorphous

Al(OH)3 solid after adsorption was also observed. Specific area

was measured using a Micromeritic Gemini model VII 2390

and an adsorption volumetric system with a pressure trans-

ductor (MKS Baratron 170 M). 100 mg of the sample was kept

under N2 flow at 60 �C for 4 h. The assays were performedwith

N2 as the adsorbate at 77 K.

3. Results and discussion

3.1. Manganese (II) adsorption in the suspended Al(OH)3solution

Static adsorption experiments were conducted in shaking

flasks. The residual concentrations of Mn2þ were stable with

reaction time at pH 6.5, while the addition of Mn2þ led to

a notable increase in residual Mn2þ (Fig. 2). By increasingMn2þ

input from 0.1 to 0.5 mg/L after 6 min of adsorption, the

0.0 0.5 1.0 1.5 2.0 2.5 3.0

0

5

10

15

20

25

q (

mg/

g)

Mn2+

input ( mg/L)

Model fitting results

Fig. 3 e Adsorption isotherm of Mn2D on amorphous

Al(OH)3 suspension at pH 7.5.

wat e r r e s e a r c h 4 6 ( 2 0 1 2 ) 4 0 6 3e4 0 7 04066

residual concentration of Mn2þ increased notably from

0.01 mg/L to approximately 0.04 mg/L. The increment of

residual Mn2þ was about 0.03 mg/L, much lower than the

increment of Mn2þ input. Increasing Mn2þ input further to

1.0 mg/L after 14 min of adsorption, similar experimental

results were obtained, indicating that the adsorption of Mn2þ

on Al(OH)3 was a fast process, andmost of the Mn2þ input was

removed. Compared with the solution at pH 6.5, residual

concentrations of Mn2þ were much lower (below 0.005 mg/L)

at pH 7.5 and 8.5 and less affected by the amount of Mn2þ

input (Fig. 2). Residual Mn2þ was not detected, even in the

solution with 1.0 mg/L Mn2þ input.

The distribution of manganese in the system was calcu-

lated using Visual MINTEQ software, while maintaining the

solution pH at 7.5e8.5. Because of the presence of amorphous

Al(OH)3, little manganese hydroxide or other manganese-

containing solids formed in the reaction. Nearly all of the

Mn2þ input was adsorbed by Al(OH)3 solid or remained in the

solution. From the X-ray diffraction spectrum of the solids

formed after adsorption, it was also difficult to observe the

peaks contributed by Mn(OH)2, MnCO3, and other manganese-

containing crystals. Both Mn(III) and Mn(IV) were not found in

amorphous Al(OH)3 after adsorption. Accordingly, the high

removal ratio of Mn2þ at pHs above 7.5 was not attributed by

chemical precipitation but mainly by the adsorption of Mn2þ

on amorphous Al(OH)3.

Table 2 e Adsorption equilibrium between amorphous Al(OH)

Total Mn (mmol) ¼AlOH0.5� (mmol) Mn2þ (mmol)

18.2 624 1.0

36.4 609 3.8

54.5 596 9.6

72.7 581 12.6

90.9 565 14.5

To observe the adsorption capacity of amorphous Al(OH)3,

the adsorption isotherm of Mn2þ was drawn in Fig. 3. The

adsorption of Mn2þ onto amorphous Al(OH)3 followed the rules

of type I adsorption isotherm at pH 7.5. When the total Mn2þ

input was 3 mg/L, 1.0 g amorphous Al(OH)3 could adsorb

approximately 24.0 mg of Mn2þ. On the basis of the concentra-

tions of total and residual Mn2þ, the adsorption constant (Kd) of

Mn2þ was calculated (Table 2). The specific surface area and

surface site density of amorphous Al(OH)3 were assumed to be

constant at 900 m2/g and 1.0 mmol/mmol, respectively

(Pommerenk and Schafran, 2005). Applying the diffuse layer

adsorption model, residual Mn2þ concentrations were pre-

dicted. In solutions with total Mn2þ input below 0.5 mg/L, the

experimentaldatafittedwellwith themodelfitting result;while

when the concentration of totalMn2þ inputwas above 1.0mg/L,

it was lower than the model fitting results, which might be

connectedwith the availability of the reaction sites in solutions

with high Mn2þ contents. Considering the concentration of

Mn2þ in drinking water usually lower than 0.5 mg/L (Mo, 2010),

the method was still valid for Mn2þ adsorption prediction.

At pHs above 7.5, the amount ofMn2þ removed fromdrinking

water was close to the amount adsorbed by amorphous Al(OH)3(Fig. 4), indicating that the reaction constant and adsorption

model selected were suitable and that amorphous Al(OH)3showed good adsorption capacity in a weak alkaline solution.

These observations agreed with the conclusions made by

Violante and Huang (1985), validating the data obtained in Fig. 4.

Besides, from Eq. (1), it could be concluded that the adsorption of

Mn2þ onto amorphous Al(OH)3 was highly pH dependent. The

chemical adsorption process would be inhibited in acidic condi-

tions (Gustafsson and Bhattacharya, 2007). However, the deter-

mined values of residual Mn2þ were lower than the predicted

values at pHs below 7.5.

Compared with chemical adsorption as the formation of

inner-sphere or strong outer-sphere bonds betweenMn2þ and

the surface structural cation, the contribution of physical

adsorption to Mn2þ removal was higher under acidic condi-

tions (Venema et al., 1996). As the existence of van der waals

force, part of the Mn2þ was removed. In Fig. 5, it was shown

that the original surface of amorphous Al(OH)3 solid was

irregular; while after Mn2þ adsorption it became much

smoother. However, as the combination strength caused by

physical adsorption was usually weaker than that caused by

chemical adsorption, flow velocity and drinking water quality

might lead to a sudden desorption of Mn2þ. In order to insure

drinking water supply safety, decreasing Mn2þ concentration

in the treated water and improving drinking water supply

network management are suggested.

3 and Mn2D at 15 �C.

¼AlOMn0.5þ (mmol) pH LogKd LogKd

17.2 6.5 �2.06 �2.43

32.5 6.5 �2.36

44.9 6.5 �2.61

60.1 6.5 �2.58

76.5 6.5 �2.53

6.5 7.0 7.5 8.0 8.5

0.0

0.2

0.4

0.6

0.8

1.0

Residual Mn2+

(calculated value)

Residual Mn2+

(determined value)

Adsorbed Mn2+

(calculated value)

Mn2+

(m

g/L

)

pH

Fig. 4 e Theoretical concentrations of Mn2D adsorbed by

amorphous Al(OH)3 at different solution pH values.

wat e r r e s e a r c h 4 6 ( 2 0 1 2 ) 4 0 6 3e4 0 7 0 4067

Elemental analysis showed that aluminum and oxygen

were the major elements of the aluminum-containing solids

that formed in both static and dynamic adsorption experi-

ments (Table 3). The Al/O molar ratio was approximately 1:3,

which corresponded well with the elemental composition

characteristics of amorphous Al(OH)3. Manganese was not

detected until after adsorption. The molecular weight percent

of Mn2þ was approximately 1.91%. Every gram of amorphous

Al(OH)3 could adsorb 18.5 mg of Mn2þ, close to the experi-

mental value. These results further indicated that amorphous

Al(OH)3 had a large adsorption capacity for Mn2þ, especially in

basic conditions.

3.2. Manganese (II) adsorption on deposited amorphousAl(OH)3

Experiments were conducted applying synthetic water with

pH, temperature, and initial Mn2þ concentration maintained

at 7.5, 15 �C, and 1.0 mg/L, respectively, after the formation of

amorphous Al(OH)3 pipe scale. The variation of residual Mn2þ

with flow volume was studied. Initially, the concentration of

residual Mn2þ in the bulk water was the lowest, at approxi-

mately 0.85 mg/L (Fig. 6). However, as the total flow increased,

the amount of Mn2þ that remained in the drinking water also

increased. When the total flow volume was 2.0 L, the

concentration of Mn2þ in the effluent increased to 1.0 mg/L

which was closed to the total Mn2þ input.

Manganese speciation was calculated applying Visual

MINTEQ under the same water quality condition. In the

absence of amorphous Al(OH)3 formation, more than 99% of

the Mn2þ would exist in the water. When amorphous Al(OH)3scale formed in the pipeline, part of the Mn2þ was adsorbed,

which had been shown in Section 3.1. At the end of stage I,

Al(OH)3 reached its maximum adsorption capacity. The

amount of Mn2þ adsorbed by deposited amorphous Al(OH)3was approximately 0.85 mg/g, much less than the value

removed from suspended amorphous Al(OH)3, whichmight be

related to its surface characteristics. Although the XRD

spectra of deposited and suspended Al(OH)3 solids were

similar to each other, both in an amorphous state, the specific

surface area of the deposited Al(OH)3 was about 160 m2/g,

much lower than that of suspended Al(OH)3, 920 m2/g. To

investigate the effects of major parameters on Mn2þ adsorp-

tion, the systems were operated under different conditions in

stage II, as shown in Fig. 6.

When the amount ofMn2þ addition decreased from 1.0mg/

L to 0 mg/L, the concentration of residual Mn2þ decreased

suddenly to 0.001mg/L, indicating that the adsorption of Mn2þ

on deposited amorphous Al(OH)3 was stable. The amount of

Mn2þ released from pipe scale was much lower than the

acceptable value 0.1 mg/L, which was good for drinking water

supply safety. However, when the added concentration of

Mn2þ was increased from 1.0 mg/L to 2.1 mg/L, the concen-

tration of residual Mn2þ reached 2.1 mg/L quickly, further

indicating that Al(OH)3 scale had reached its maximum

adsorption capacity (Fig. 6). When the flow rate increased

from 0.5 to 1.0 mL/s, the concentration of residual Mn2þ

increased from 0.94 mg/L to 1.12mg/L gradually, this might be

linked to the flushing effects of the pipe water at higher flow

velocity (Friedman et al., 2003). Maintaining low flow velocity

was suggested for residual Mn2þ control in the tap water.

Similar to the effects of flow velocity, the variation of

solution pH also had little effect on the adsorption of Mn2þ on

deposited amorphous Al(OH)3 in basic conditions (Fig. 6). At

pH 7.5e8.5, the amount ofMn2þ residual in drinkingwaterwas

stable at about 1.0 mg/L. Unlike the case at pH 8.5, the

concentration of residual Mn2þ fluctuated in the range of

0.85e1.05 mg/L at pH 6.5, which might be connected with the

weak combination between amorphous Al(OH)3 and Mn2þ

under acidic conditions (Fonseca et al., 2006). Elemental

analysis revealed that manganese was not detected in the

freshly formed scale; however, after adsorption, its molecular

weight ratio increased to approximately 0.15%. Every gram of

amorphous scale could adsorb 1.1 mg of Mn2þ, close to the

experimental value.

In this study, Al(OH)3 scale was prepared prior to Mn2þ

adsorption. Most of the Mn2þ was mainly adsorbed by the

reaction sites on its surface only, leading to much less Mn2þ

removed from water (Hiemstra and van Riemsdijk, 1996). In

actual DWDS, however, amorphous Al(OH)3 usually formed

simultaneously with the removal of pollutants. Metal

elements not only existed on the surface but also inside of

Al(OH)3 scale (Lauer and Lohman, 1994). It could be concluded

that in the formation process of amorphous Al(OH)3 scale, its

adsorption capacity for Mn2þ was strong; however, when

Al(OH)3 formation was halted or its structure was destroyed

due to the variation of water quality, its adsorption capacity

would drastically decrease or even leading to a suddenly

release ofMn2þ. Accordingly, removing contaminant reservoir

should be conducted regularly in DWDS management.

3.3. Effects of water quality on Mn2þ adsorption

In addition to solution pH and manganese concentration, the

influence of co-existing cations on the adsorption ofMn2þwas

also examined. Considering Ca2þ, Mg2þ, and Fe3þ commonly

existing in drinking water, their concentrations are usually

higher than other elements (Schock, 2005). Dissolved organic

Fig. 5 e Morphological characteristics of amorphous Al(OH)3 before (a) and after (b) Mn2D adsorption.

wat e r r e s e a r c h 4 6 ( 2 0 1 2 ) 4 0 6 3e4 0 7 04068

matter is mainly produced by living organisms or formed by

secondary synthesis reactions (Desroches and Dayde, 2000;

Liu et al., 2007). As a commonly existing organic matter in

drinking water, HA could react both with aluminum-

containing solids and Mn2þ by the formation of stable

complexes (Pommerenk and Schafran, 2005). Their effects on

Mn2þ adsorption on amorphous Al(OH)3 were also investi-

gated at pH 6.5 with 1.0 mg/L Mn2þ addition.

The presence of co-existing materials could promote the

removal of Mn2þ from drinking water. When the concentra-

tion of Ca2þ was 20 mg/L, residual Mn2þ was approximately

0.08 mg/L after 10 min adsorption (Fig. 7). Increasing the

concentration of Ca2þ to 60 mg/L, the residual Mn2þ increased

to about 0.035 mg/L after 22.0 min adsorption. When the

concentration of Ca2þ was at its highest value (100 mg/L), the

residual Mn2þ concentration was approximately 0.03 mg/L.

Similar experimental results were obtained when Mn2þ co-

existed with Mg2þ, Fe3þ, and HA in solution. However, the

contribution of these co-existing materials differed greatly

from each other. Their promoting capacity on Mn2þ removal

could be listed from strong to weak as follows:

HA > Fe3þ > Mg2þ > Ca2þ.Co-precipitation and enhanced adsorption induced by the

co-existing cations are the major reasons leading to further

removal of Mn2þ from drinking water (Knocke et al., 1991).

With a solution pH of 6.5, the solubility of Fe(OH)3 (III) was the

lowest, and the solubility of CaCO3 was the highest among the

cations selected in the experimental water quality condition.

Table 3 e Elements composition of amorphous Al(OH)3 before

Elements Before adsorption After adso

Weight (%) Mole (%) Weight (%)

O 62.84 74.33 62.7

Al 33.57 23.54 32.64

S 3.60 2.12 2.75

Mn 0.00 0.00 1.91

The amount of Fe(OH)3 solid formed was also more than

Mg(OH)2 and CaCO3. Besides the accessibility of adsorption

sites, the adsorption capacity of Fe(OH)3 for Mn2þ was also

stronger than that ofMg(OH)2 and CaCO3 (Buamah et al., 2008),

leading tomoreMn2þ removed from the systemwith Fe(III) co-

existing.

Unlike the co-existing cations, the HA used in this experi-

ment, with an average molecular weight below 500 Da, was

soluble. Its reactive characters can be explained by a series of

oxygen-containing functional groups, which includes

carboxyl, phenolic-OH, enolic, and alcoholic-OH, and carbonyl

(Wang et al., 2010). One HA molecular usually contains many

different active reaction sites (Tipping, 1994), which provide

more combining sites for Mn2þ and Al(OH)3 solid. With the

effects of bridging adsorption, nearly all of the Mn2þ was

removed.

Accordingly, in solutionswith CaCO3, Mg(OH)2, and Fe(OH)3solid formation and HA, Mn2þ removal ratio was higher due to

the effects of co-precipitation and adsorption. Even in solution

with only 1.0 mg/L HA, the concentration of residual Mn2þ in

the water body was determined to be below 0.005 mg/L.

However, it should be noted that this study has only investi-

gated the adsorption and precipitation processes of Mn2þ at

the lab scale, and its oxidation process was not considered.

Not withstanding its limitation, this study does suggest the

adsorption rules hold true for the adsorption of Mn2þ on

amorphous Al(OH)3 formed in DWDS when there is no

oxidation reaction with the residual Mn2þ.

(a) and after (b) adsorption.

rption (suspension) After adsorption (scale)

Mole (%) Weight (%) Mole (%)

74.51 63.33 74.05

23 33.29 23.12

1.84 3.23 2.76

0.65 0.15 0.07

0 5 10 15 20 25 30

0.00

0.05

0.10

0.15

Stage 3: coexisting cations at high-level

Stage 2: coexisting cations at mid-level

Stage 1: coexistingcations at low-level

Mn2+

con

cent

rati

on (

mg/

L )

Adsorption time ( min )

Ca Fe Mg HA

Fig. 7 e Effects of Ca2D, Mg2D, Fe3D, and HA on the

adsorption of Mn2D on amorphous Al(OH)3 suspension.

0 2500 5000 25000 26000 27000 28000 29000

0.8

1.0

2.0

2.2

Stage ΙΙ

Mn

conc

entr

atio

n (

mg/

L)

V ( mL )

v=0.2 mL/s v=1.0 mL/s pH 6.5 pH 8.5 C = 2.1 mg/L

Stage Ι

Fig. 6 e Variation of Mn2D remained in drinking water with

the total flow volume before and after water quality and

flow velocity adjustment in the system with amorphous

Al(OH)3 scale formation.

wat e r r e s e a r c h 4 6 ( 2 0 1 2 ) 4 0 6 3e4 0 7 0 4069

4. Conclusions

The adsorption of Mn2þ on amorphous Al(OH)3 is a complex

process. In drinking water with solution pHs above 7.5,

chemical adsorption mainly took place; Al(OH)3 suspension

showed a strong adsorption capacity to Mn2þ. In drinking

water with solution pHs below 7.5, as the inhibition of Hþ

increased, active reaction sites on the solid surfaceweremuch

less. Thus, the adsorption process of Mn2þ on amorphous

Al(OH)3 changed gradually from chemical coordination to

physical adsorption with a decreasing solution pH.

In drinking water with high concentrations of Ca2þ, Mg2þ,and Fe3þ, the adsorption of Mn2þ on amorphous Al(OH)3 was

enhanced due to the effects of co-precipitation and adsorption

contributed by the newly formed CaCO3, Mg(OH)2, Fe(OH)3,

and other solids. Different from the effects of co-existing

cations, dissolved organic matter, especially HA, enhanced

the adsorption of Mn2þ due to the effect of bridge connecting.

Even in solution with only 1.0 mg/L HA (average molecular

weight of less than 500 Da), the concentration of residualMn2þ

in the water body could be controlled below 0.005 mg/L, much

lower than the standard required level.

In the formation of amorphous Al(OH)3 scale, the adsorp-

tion capacity for Mn2þ was the highest, at about 24.0 mg/g,

when the Mn2þ input was 3 mg/L at 15 �C. In solution with HA

and CaCO3, Mg(OH)2, and Fe(OH)3 solids, its removal ratio was

higher. When the Al(OH)3 scale stopped growing, its adsorp-

tion capacity decreased significantly to approximately

0.85 mg/g. Better management of water age and sediment

accumulation was suggested to control the concentration of

Mn2þ in the tap water. Moreover, flushing and mechanical

cleaning should also be conducted regularly.

Acknowledgments

This research was supported by Changjiang Scholars and

Innovative Research Team in University (PCSIRT) (IRT0853),

the National Natural Science Foundation of China (21007050),

and Changzhou University Research Fund (ZMF1002127). We

are also grateful for anonymous reviewers for their helpful

suggestions and advices.

r e f e r e n c e s

Buamah, R., Petrusevski, B., Schippers, J.C., 2008. Adsorptiveremoval of manganese(II) from the aqueous phase using ironoxide coated sand. Journal of Water Supply Research andTechnology e AQUA 57 (1), 1e11.

Desroches, S., Dayde, S., 2000. Aluminum speciation studies inbiological fluids: Part 6. Quantitative investigation ofaluminum(III)-tartrate complex equilibria and their potentialimplications for aluminum metabolism and toxicity. Journalof Inorganic Biochemistry 81 (4), 301e312.

Dzombak, D.A., Morel, F.M.M., 1990. Surface ComplexationModelling: Hydrous Ferric Oxide. Wiley-Interscience,New York.

Fonseca, M.G., Oliveira, M.M., Arakaki, L.N.H., 2006. Removal ofcadmium, zinc, manganese and chromium cations fromaqueous solution by a clay mineral. Journal of HazardousMaterials 137 (1), 288e292.

Friedman, M.J., Martel, K., Hill, A., Holt, D., Smith, S., Ta, T.,Sherwin, C., Hiltebrand, D., Pommerenk, P., Hinedi, Z.,Camper, A., 2003. Establishing Site-Specific FlushingVelocities. AwwaRF, Denver, Colorado.

Gustafsson, J.P., Bhattacharya, P., 2007. Geochemical modelling ofarsenic adsorption to oxide surfaces. In: Bhattacharya, P.,Mukherjee, A.B., Bundschuh, J., Zevenhoven, R., Loeppert, R.H.(Eds.), Arsenic in Soil and Groundwater Environment. TraceMetals and Other Contaminants in the Environment, vol. 9,pp. 159e206.

Hiemstra, T., van Riemsdijk, W.H., 1996. A surface structuralapproach to ion adsorption: the charge distribution (CD)model. Journal of Colloid and Interface Science 179 (2),488e508.

Institute for Environment and Health/Institute of OccupationalMedicine, 2004. Occupational Exposure Limits: Criteria

wat e r r e s e a r c h 4 6 ( 2 0 1 2 ) 4 0 6 3e4 0 7 04070

Document for Manganese and Inorganic ManganeseCompounds. Web Report W17. Medical Research Council,Institute for Environment and Health, Leicester, UnitedKingdom,. Available from: http://www.le.ac.uk/ieh.

Jose, A.M.F., Maryse, B., Paula de, N.S., Josino, C.M., 2009.Manganese exposure and the neuropsychological effect onchildren and adolescents: a review. American Journal of PublicHealth 26 (6), 541e548.

Julien, F., Gueroux, B., Mazet, M., 1994. Comparison of organiccompounds removal by coagulation focculation and byadsorption onto preformed hydroxide focs. Water Research 28(12), 2567e2574.

Kirmeyer, G.J., Kirby, L., Murphy, B.M., Noran, P.F., Martel, K.D.,Lund, T.W., Anderson, J.L., Medhurst, R., 1999. Maintainingand Operating Finished Water Storage Facilities. AWWA andAWWARF, Denver, Colo.

Knocke, W.R., Occiano, S.C., Hungate, R., 1991. Removal of solublemanganese by oxide-coated filter media: sorption rate andremoval mechanism issues. Journal of the American WaterWorks Association 83 (8), 64e69.

Lauer, W.C., Lohman, S.R., 1994. Corrosive Water Can ProtectPipes. American Water Works Association Water QualityTechnology Conference. American Water Works Association,San Francisco, California, pp. 1753e1760.

Liu, G.J., Zhang, X.R., Talley, J.W., 2007. Effect of copper(II) onnatural organic matter removal during drinking watercoagulation using aluminum-based coagulants. WaterEnvironment Research 79 (6), 593e599.

Mo, Z.X., 2010. Briefly explore the sources of manganese pollutionin drinking water. Journal of Disease Monitor and Control 4 (1),29e30.

Pommerenk, P., Schafran, G.C., 2005. Adsorption of inorganic andorganic ligands onto hydrous aluminum oxide: evaluation ofsurface charge and the impacts on particle and NOM removalduring water treatment. Environmental Science & Technology39, 6429e6434.

Schindler, P.W., 1981. Surface complexes at oxide-water interfaces.In: Anderson, M., Rubin, A. (Eds.), Adsorption of Inorganics atSolideLiquid Interfaces.AnnArborScIence,AnnArbor,Michigan.

Schock,M.R., 2005.Distributionsystemsasreservoirsandreactors forinorganic contaminants. Ch.6. In:MacPhee,M.J. (Ed.), DistributionSystemWater Quality Challenges in the 21st Century. AmericanWaterWorks Association, Denver, CO, pp. 105e140.

Schock, M., Holm, T., 2003. Are we monitoring in the right placesfor metals and radionuclides? Journal of New England WaterWorks Association 141 (2), 102e106.

Seppanen, H.T., 1993. Experience of biological iron andmanganese removal in Finland. Journal of the Institution ofWater and Environmental Management 6 (3), 333e341.

Sly, L.J., Hodgkinson, M.C., Arunpairojana, V., 1990. Deposition ofmanganese in a drinking water distribution system. Appliedand Environmental Microbiology 56 (3), 628e639.

Snoeyink, V.L., Schock, M.R., Sarin, P., Wang, L., Chen, A.S.C.,Harmon, S.M., 2003. Aluminum-containing scales in waterdistribution systems: prevalence and composition. Journal ofWater Supply Research and Technology e AQUA 52 (7),455e474.

Sommerfeld, E.O., 1999. Iron and Manganese Removal Handbook.American Water Works Association, Denver, CO, USA.

Tipping, E., 1994. WHAM e a chemical equilibrium model andcomputer code for waters, sediments and soils incorporatinga discrete site/electrostatic model of ion-binding by humicsubstances. Computers Geosciences 20 (6), 973e1023.

USEPA, 2006. Inorganic Contaminant Accumulation in PotableWaterDistribution Systems. Office of Groundwater and DrinkingWater. Available online at: http://www.google.com.hk/url?q¼http://www.epa.gov/ogwdw/disinfection/tcr/pdfs/issuepaper_tcr_inorganiccontaminantaccumulation.pdf&sa¼U&ei¼f3pyT6KYBIzBiQej8pjkDw&ved¼0CBcQFjAA&usg¼AFQjCNEFAfMVSlZ8ImwqVeT3e1HH3V4jiw.

USEPA, 2008. Integrated Risk Information System: Manganese(CASRN 7439-96-5). Available online at: http://www.epa.gov/ncea/iris/subst/0373.htm.

Venema, P., Hiemstra, T., Van Riemsdijk, W.H., 1996. Comparisonof different site binding models for cation sorption:description of pH dependency, salt dependency and cation-proton exchange. Journal of Colloid and Interface Science 181(1), 45e59.

Violante, A., Huang, P., 1985. Influence of inorganic and organicligands on the formation of aluminum hydroxides andoxyhydroxides. Clays and Clay Minerals 33 (3), 181e192.

Wasserman, G.A., Liu, X., Parvez, F., Ahsan, H., Levy, D.,Factor, L.P., 2006. Water manganese exposure and children’sintellectual function in Araihazar, Bangladesh. EnvironmentalHealth Perspectives 114 (1), 124e129.

Wang, W.D., Li, H., Wang, X.C., Liu, Y.J., 2010. Spatial variations ofaluminum species in drinking water supplies in Xi’an studiedapplying geographic information system. Journal ofEnvironmental Sciences 22 (4), 519e525.

Wang, W.D., Yang, H.W., Jiang, J., Ding, Z.Z., Zhu, W.P., 2011.Aluminum speciation in drinking water distribution system:a case study in northeastern China. Korean Journal ofChemical Engineering 28 (5), 1e6.

World Health Organization, 2008. Guidelines for Drinking-WaterQuality (third Ed. Incorporating the 1st and 2nd Addenda),Volume 1 Recommendations. World Health Organisation,Geneva. Available online at: http://www.who.int/water_sanitation_health/dwR/gdwq3rev/en/.