ARTICLE IN PRESS

Available at www.sciencedirect.com

WA T E R R E S E A R C H 4 0 ( 2 0 0 6 ) 3 6 4 – 3 7 2

0043-1354/$ - see frodoi:10.1016/j.watres

�Corresponding aE-mail address: sh

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

Arsenic removal from high-arsenic water by enhancedcoagulation with ferric ions and coarse calcite

S. Song�, A. Lopez-Valdivieso, D.J. Hernandez-Campos, C. Peng, M.G. Monroy-Fernandez,I. Razo-Soto

Instituto de Metalurgia, Universidad Autonoma de San Luis Potosı, Avenida Sierra Leona 550, San Luis Potosı, C. P. 78210, Mexico

a r t i c l e i n f o

Article history:

Received 2 March 2004

Received in revised form

3 April 2005

Accepted 26 September 2005

Available online 15 December 2005

Keywords:

Coagulation

Arsenic removal

Coarse calcite

Ferric sulfate

Colloidal stability

nt matter & 2005 Publish.2005.09.046

uthor. Tel./fax: +52 444 82aoxian@uaslp.mx (S. Son

A B S T R A C T

Arsenic removal from high-arsenic water in a mine drainage system has been studied

through an enhanced coagulation process with ferric ions and coarse calcite (38–74 mm) in

this work. The experimental results have shown that arsenic-borne coagulates produced by

coagulation with ferric ions alone were very fine, so micro-filtration (membrane as filter

medium) was needed to remove the coagulates from water. In the presence of coarse

calcite, small arsenic-borne coagulates coated on coarse calcite surfaces, leading the

settling rate of the coagulates to considerably increase. The enhanced coagulation followed

by conventional filtration (filter paper as filter medium) achieved a very high arsenic

removal (over 99%) from high-arsenic water (5 mg/l arsenic concentration), producing a

cleaned water with the residual arsenic concentration of 13mg/l. It has been found that the

mechanism by which coarse calcite enhanced the coagulation of high-arsenic water might

be due to attractive electrical double layer interaction between small arsenic-borne

coagulates and calcite particles, which leads to non-existence of a potential energy barrier

between the heterogeneous particles.

& 2005 Published by Elsevier Ltd.

1. Introduction

In Mexico, hundreds of years of mining activities for gold,

silver and metallic sulfide minerals left a considerable

quantity of tailings, some of which have a high arsenic

content. Owing to rain, wind and other natural effects, the

arsenic dispersed into the nearest water courses or agricul-

tural fields in the forms of arsenate ion (As (V)) and arsenite

ion (As (III)), leading to a high arsenic concentration in the

water courses around the mines. For example, in Matehuala

city, state of San Luis Potosı, water from mine drainage

systems contained up to 7 mg/l As (Razo-Soto et al., 2005). As

it has been reported, serious health problems, such as

cancers, skin alteration, etc., have been linked to arsenic

ingestion even at a low concentration (20–50mg/l) (O’Connor,

2002). The Secretary of Health (SAA) of Mexico recently has

ed by Elsevier Ltd.

54326.g).

implemented a new maximum contaminant level (MCL) for

arsenic in drinking water of 20mg/l. Accordingly, the high

arsenic water from mine drainage systems would be a great

environmental concern through the contamination of shal-

low groundwater and surface water around the areas. The

water has to be treated to remove arsenic effectively, in order

to minimize the environmental effects.

Arsenic removal from water is an important subject

worldwide, which has recently attracted great attentions. A

variety of treatment processes has been developed for arsenic

elimination from water, including coagulation (precipitation)

(Wickramasinghe et al., 2004; Hering et al., 1997), adsorption

(Wang et al., 2002; Zhang et al., 2003; Katsoyiannis and

Zouboulis, 2002), ion exchange (Korngold et al., 2001),

membrane filtration (Sato et al., 2002), electrocoagulation

(Kumar et al., 2004; Arienzo et al., 2002), biological process

ARTICLE IN PRESS

WAT ER R ES E A R C H 40 (2006) 364– 372 365

(Katsoyiannis and Zouboulis, 2004), iron oxide-coated sand

(Thirunavukkarasu et al., 2003), high gradient magnetic

separation (Chiba et al., 2002) and natural iron ores (Zhang

et al., 2004), manganese greensand (Thirunavukkarasu et al.,

2005), etc. There are numerous review papers for the arsenic

removal technologies, some of which were recently made by

Jiang (2001), Bissen and Frimmel (2003), Dambies (2004) and

USEPA (2000). Coagulation and adsorption processes are most

promising for arsenic removal from high-arsenic water

because of the low cost and high efficiency, and are widely

used in the developing world. But, they have not been shown

to deeply eliminate arsenic from water and to produce

cleaned water with a very low arsenic concentration, say

10mg/l. However, membrane filtration process could lower

arsenic concentration in water from 48 to 1–2mg/l (Bissen and

Frimmel, 2003), and ion exchange process could remove

arsenic to levels lower than 5 mg/L for water with initial

arsenic concentration of 87mg/L (Wang et al., 2002), which are

currently used in the developed world. The other processes

are still at a laboratory or pilot scale.

Coagulation process is traditionally realized by adding ferric

or aluminum ions (Hering et al., 1996). In this process, fine

particles in water first aggregate into coagulates because

added ferric or aluminum ions strongly reduce the absolute

values of zeta potentials of the particles. Then, arsenic ions

(arsenate or arsenite) precipitate with the ferric or aluminum

ions on the coagulates, and thus concentrate in the coagu-

lates. After that, the coagulates are separated from water

through filtration, eliminating arsenic from the water. The

coagulates are termed arsenic-borne coagulates. Coagulation

with ferric ions for arsenic removal can be traced back to the

late 1960s in Taiwan to treat deep-well water with naturally

elevated arsenic concentrations (Shen, 1973). Gulledge and

O’Connor (1973) also reported that arsenic could be readily

removed from water to a high degree by conventional water

treatment using ferric or aluminum ions as coagulants. Since

then, there have been a lot of reports on coagulation process

for arsenic removal. It has been found that the coagulation is

much more effective for the removal of As (V) than As (III). In

the case when only As (III) is present, oxidation to convert As

(III) to As (V) is needed prior to coagulation. The effective pH

for arsenic removal was reported to be 5–7 with aluminum

ions, and 5–8 for ferric ions (Sorg and Logsdon, 1978). Besides

iron and aluminum compounds, manganese, calcium and

magnesium compounds are also of effective coagulants for

eliminating arsenic from water in neutral medium (Raje and

Swain, 2002; Jiang, 2001). Arsenic removal from water

achieved by coagulation process depends on initial arsenic

concentration in water (Thirunavukkarasu et al., 2005; Jiang,

2001). The arsenic removal could reach 99% (Jiang, 2001).

Recently, it was reported that modified coagulation/filtration

could give a residual arsenic concentration of 2mg/l or less for

treated well water (Han et al., 2003).

The arsenic removal is also dependent on the pore size of

the membrane filter disks used for coagulation process (Han

et al., 2003), since coagulates smaller than the pore can pass

through the filter and remain in water. As it is known, in

filtration, larger the filter pores, lower the capital and

operation costs, and higher the separation efficiency. There-

fore, there is a great significance in enlarging arsenic-borne

coagulates in order to improve coagulation/filtration process

for arsenic removal. However, there is little information in

this regard, although there are numerous reports on coagula-

tion process for arsenic removal from water.

Usually, coagulations are enhanced by adjusting pH and

electrolyte concentration to reduce the absolute values of zeta

potentials of particles, and by optimizing coagulation ki-

netics. Besides, it can also be realized by adding appropriate

coarse particles, in which fine particles in suspensions coat

coarse particles in the form of multilayers. Such coagulates

consist of fine particles and coarse particles. This phenom-

enon was observed early in mineral flotation systems, and

was termed ‘‘slime coating’’ (Fuerstenau et al., 1979; Soma-

sundaran, 1980). The coarse particle enhancement of coagu-

lation of fine particles not only strongly reduces particle

numbers in suspensions, but also greatly increases coagulate

size. The mechanism of the coarse particle effect might be

attributed to the increase of aggregation rate, because

collision rate between coarse particle and fine particle is

much higher than that between fine particles themselves

(Song and Lopez-Valdivieso, 2002). The commercial uses of

this effect appeared in a flotation circuit for the purification of

kaolin early in 1960s, in which anatase fines aggregated with

coarse calcite, followed by the flotation of the coated calcite to

remove the anatase from kaolin (Green and Duke, 1962).

In the present study, we attempted to make use of the

coarse particle effect to improve the coagulation process for

arsenic removal from water. The objective is to develop a new

cost-effective technique for eliminating arsenic from high-

arsenic water in the mine drainage located in Matehuala city,

Mexico. The possibility and effectiveness of this new en-

hanced coagulation process were determined while ferric

sulfate was used as coagulant and calcite was used as the

coarse particle. In order to understand the coarse particle

effect in the enhanced coagulation, the zeta potentials of

arsenic-borne coagulates and calcite particles were deter-

mined, and the settling rates of arsenic-borne coagulates

before and after coating coarse calcite were measured. In

addition, arsenic-borne coagulates were observed by using a

scanning electronic microscope (SEM) and an atomic force

microscope (AFM).

2. Experimental

2.1. Materials

The high-arsenic water sample used in this work was

originally collected from the Cerrito Blanco channel, located

in Matehuala city, state of San Luis Potosi, Mexico. The water

was first sieved with a 400 mesh screen, and then was

filtrated with an Ahlstrom Grade 610 filter paper (2.5 mm

aperture) to remove solid contaminants. The water past the

filter paper was used for the tests, which gave the initial

arsenic concentration of 5071mg/l. The chemical composition

and some properties of the water sample were listed in Tables

1 and 2, respectively.

Ferric sulfate (Fe2(SO4)3 � 5H2O) (J. T. Barker, analytic purity)

was used as coagulant; hydrochloric acid (HCl) (Fremont,

ARTICLE IN PRESS

Table 1 – Chemical composition of the high-arsenic water sample

Constituent Ca Mg Na K Fe Mn As Cl� NO3� SO4

2� HCO3� PO4

3�

Concentration (mg/l) 566 77 37 8.9 0.11 0.01 5.07 142 0.25 1340 185 40.8

Table 2 – Some properties of the high-arsenic water sample

pH Conductivity(mS/cm)

Eh (mV) Dissolved organiccarbon (mg/l)

Dissolved organicmaterials (mg/l)

7.62 1960 414 60.5 121

WA T E R R E S E A R C H 4 0 ( 2 0 0 6 ) 3 6 4 – 3 7 2366

analytic purity) and sodium hydroxide (NaOH) (J. T. Barker,

analytic purity) was used to adjust pH.

Calcite used in this work was obtained from the mine of

Minera de Las Curvas, Mexico. The chunk mineral was first

crushed with a hammer, and then purified by hand sorting.

After that, the mineral was ground with a porcelain mortar,

followed by a size classification with screens to obtain two

size fractions of 52–74mm and 38–44mm, which were deter-

mined to contain 99.2% and 99.0% CaCO3, respectively.

2.2. Arsenic removal by coagulation with ferric ions

High-arsenic water (200 ml) was first mixed with a given

amount of Fe2(SO4)3 in a flask, and then was adjusted for pH

with HCl or NaOH by using a potential meter (Orion 720-A).

After that, it was stirred on a magnetic agitator (Digital hot

plate/stirrer 04644) at 240 rev/min for 30 min, while tempera-

ture was kept at 3070.2 1C. Arsenic-borne coagulates were

formed during the agitation. After that, the suspension was

filtrated through a Millipore white GSWP membrane (0.22mm

aperture). The filtrate was sent for arsenic analysis. Each test

was duplicated. The arithmetic average result of the two tests

was reported in this paper.

2.3. Arsenic removal by enhanced coagulation with ferricions and coarse calcite

High-arsenic water (200 ml) was first mixed with a given

amount of Fe2(SO4)3 and coarse calcite in a flask, and then pH-

adjusted and stirred in the same way as mentioned above.

After that, the suspension was filtrated through an Ahlstrom

Grade 610 filter paper (2.5mm aperture). The filtrate was sent

for arsenic analysis. Each test was duplicated. The arithmetic

average result of the two tests was reported in this paper.

2.4. Arsenic analysis

Arsenic concentration in water or filtrates was determined by

using a Perkin-Elmer Aanalyst 100 atomic absorption spectro-

meter coupled to a Perkin-Elmer FIAS 100 flow injection

system. Water or filtrate (20 ml) was first mixed with 2 ml

mixture of KI (10%) and ascorbic acid (5%) and 1 ml

concentrated hydrochloric acid at 75 1C for 5 min. This

treatment was to reduce As (V) to As (III) and to dissolve

solid particles in the water or filtrates. Then, the solution was

cooled to room temperature, from which a 10 ml solution was

taken for the analysis of As (III) concentration with the

instruments. Each analysis for arsenic concentration was

duplicated. The arithmetic average of the two analysis results

was reported in this paper. As soon as arsenic concentrations

were obtained, arsenic removal (E) was calculated by the

following expression:

E ¼C0 � C

C0, (1)

where C0 and C are the arsenic concentration of the high-

arsenic water sample and a filtrate from arsenic removal

tests, respectively.

2.5. Turbidity measurement

Turbidity of the high-arsenic water treated by coagulation

was determined by using a Bausch & Lomb Spectronic 20

spectrophotometer. First, coagulation was applied to water.

Then, 20 ml of the suspension was transferred to a tube for

turbidity measurement. Light transmissivity as a function of

settling time was obtained from this measurement, which

could provide information on settling rate of coagulates. At

the same settling time, a larger light transmissivity indicates

a higher settling rate, or stronger coagulation. Each measure-

ment was tripled. The arithmetic average of the three

measurement results was reported in this paper.

2.6. Microscope study

Coagulates from coagulations were observed by using a

Quesant Qscope 250 AFM and a Philips Microspec WDX SEM

in this work. AFM images and SEM micrographs of coagulates

were photographed. The samples for the observations were

prepared by putting a drop of coagulated suspension on a

metal dish for SEM study or on a mica for AFM study, and then

drying in room temperature. Before the SEM study, the

sample on the dish was coated by gold. In addition, an EDAX

energy dispersive spectrometer (EDS) attached to the SEM was

used to determine the existence of arsenic in coagulates.

2.7. Zeta potential determination

A micro-electrophoresis meter (Riddick) was used to measure

the zeta potentials of arsenic-borne coagulates from the

ARTICLE IN PRESS

0

20

0 4 8 12 16 20

40

60

80

100

CU

MU

LAT

IVE

UN

DE

RS

IZE

, %

PARTICLE SIZE, µm

Arsenic-borne coagulates

100 mg/l Fe2(SO

4)3 at pH 5

Fig. 2 – Size distribution of arsenic-borne coagulates from

the coagulation of high-arsenic water with 100 mg/

WAT ER R ES E A R C H 40 (2006) 364– 372 367

coagulation mentioned above and calcite fines in aqueous

solutions. The electrophoresis mobility of over 20 particles

was determined, and the average value was transformed to

zeta potential using the Smoluchowski equation.

2.8. Particle size analysis

A Shimadzu SALD-1100 laser diffraction particle size analyzer

with semiconductor laser as the light source was used to

measure the size distributions of arsenic-borne coagulates

from the coagulation mentioned above in this study. This

instrument reports optical diameter instead of the equivalent

Stokes diameter. In order to protect coagulates from breaking,

ultrasonic was not applied to the suspensions during the

measurements. The measurement was tripled. The arith-

metic average of the three measurement results was reported

in this paper.

l ferric sulfate at pH 5.

515.9 nm

0

5.0 µm

10.0 µm

15.0 µm 5.0 µm10.0 µm

15.0 µm

0

500

400

300

200

100

0nm

Fig. 3 – AFM image of arsenic-borne coagulates from the

coagulation of high-arsenic water with 100 mg/l

ferric sulfate at pH 5.

3. Results and discussions

3.1. Arsenic removal by coagulation with ferric ions

Fig. 1 illustrates arsenic removal from the high-arsenic water

as a function of pH by coagulation with 100 mg/l ferric sulfate

and micro-filtration with the membrane. Error analysis for

the results of arsenic elimination presented in the graph was

made on the basis of the standard error of the coagulation/

filtration tests and arsenic analysis. The error was also

presented in the graph in the form of an error bar. As it can

be seen, in the acidic range, very high arsenic removal, about

99%, was achieved, while in the alkaline range, the arsenic

removal declined sharply with the increase of pH. This result

is in a good agreement with those obtained by coagulation

with ferric ions for low-arsenic concentration water

(20–100mg/l), as reported elsewhere (Hering et al., 1996;

Gulledge and O’Connor, 1973). Obviously, coagulation/micro-

filtration process is very effective in eliminating arsenic from

high-arsenic water at acidic to neutral pH.

754 5 6 7 8 9 10 11 12

80

85

90

95

100

100 mg/l Fe2(SO

4)3

Stirring: 240 rev/min, 30 min

AR

SE

NIC

RE

MO

VA

L, %

pH

Fig. 1 – Arsenic removal from the high-arsenic water as a

function of pH by coagulation with ferric ions and

filtration with the membrane.

Fig. 2 shows the size distribution of arsenic-borne coagu-

lates from the coagulation of the high-arsenic water with

100 mg/l ferric sulfate at pH 5. Error analysis for the weight

percentage of each size fraction showed that the maximum

error was 70.6%. Since the error bars were too small to draw,

only the arithmetic average values of cumulative undersize

were presented in the graph. As it is noted, the coagulates

were very fine, ranging between 0.5 and 20mm. There were

about 17% (weight percentage) coagulates smaller than

2.5mm, and about 83% coagulates smaller than 10mm.

Obviously, filtration with the membrane (0.22mm aperture)

is qualified for a perfect separation of the coagulates from

water. However, filtration with the filter paper (2.5 mm

aperture) would lead to low arsenic removal, because 17%

coagulates would pass through the paper into filtrates. In

addition, this result suggests that the separation of the

coagulates from water cannot be realized by gravitational

sedimentation, as the coagulates are in colloidal size range.

The AFM image of coagulates from the coagulation of the

high-arsenic water with 100 mg/l ferric sulfate at pH 5 is

given in Fig. 3. It clearly shows the three dimensions of

ARTICLE IN PRESS

100

WA T E R R E S E A R C H 4 0 ( 2 0 0 6 ) 3 6 4 – 3 7 2368

arsenic-borne coagulates. The coagulates have irregular

shape, and a size of 0.5–4 mm.

605 6 7 8 9 10 11 12

70

80

90

AR

SE

NIC

RE

MO

VA

L, %

pH

38 - 44 µm calcite 52 - 74 µm calcite

Coagulation: 100 mg/l Fe2(SO

4)3

5 g/l calcite

Fig. 5 – Arsenic removals from high-arsenic water as a

function of pH by the enhanced coagulation and

filtration with filter papers.

3.2. Arsenic removal by enhanced coagulation with ferricions and coarse calcite

Enhanced coagulation with ferric ions and coarse calcite

followed by filtration with filter papers (2.5mm aperture) was

tested to eliminate arsenic from the high-arsenic water. Fig. 4

illustrates arsenic removal from the water as a function of

calcite addition at pH 6 and 100 mg/l ferric sulfate addition.

The error bars obtained from the standard error analysis of

coagulation/filtration tests and arsenic analysis were also

given in the graph. Without coarse calcite, the arsenic

removal was only 85%, which was in correspondence with

the cumulative oversize of arsenic-borne coagulates as shown

in Fig. 2. Such a low arsenic removal should be attributed to

small arsenic-borne coagulates passing through the filter

papers. As the increase of coarse calcite addition, the arsenic

removal increased until it reached about 99%. This increase

was very sharp in the small calcite addition range, but was

mild after 2.5 g/l calcite addition. Indeed, the coagulation of

high-arsenic water could be greatly enhanced by adding

coarse calcite particles and, thus, arsenic can be effectively

removed from the water by filtration in a large pore filter

medium. Also, it can be observed from this graph that

38�44mm calcite enhanced arsenic removal more strongly

than 52�74mm calcite at the same amount of addition. The

mechanism might be due to the larger surface area on smaller

particles, leading to more coatings of arsenic-borne coagu-

lates on calcite surfaces.

The effect of pH on arsenic removal from the high-arsenic

water by coagulation with 100 mg/l ferric sulfate and 5 g/l

coarse calcite and filtration with the filter papers is shown in

Fig. 5. The error bars of arsenic removal for the points

appeared in the graph were 70.2% to 70.7%. They were not

presented in the graph because most of the bars overlapped

with the corresponding points. As it is noted, the arsenic

removal vs. pH curve in the presence of coarse calcite is very

850 5 10 15

90

95

100

100 mg/L Fe2(SO

4)3

pH = 6.0

AR

SE

NIC

RE

MO

VA

L, %

CALCITE ADDITION, g/L

38 - 44 µm calcite52 - 74 µm calcite

Fig. 4 – Arsenic removal from high-arsenic water as a

function of coarse calcite addition by the enhanced

coagulation at pH 6 and filtration with filter papers.

similar to that without calcite as shown in Fig. 1. High arsenic

removal was achieved in the range of pH 5–7. At the alkaline

pHs, the arsenic removal declined sharply with increasing pH.

Again, this graph also showed that 38–44mm calcite improved

arsenic removal from the high-arsenic water by enhanced

coagulation more strongly than 52�74mm calcite.

At pH ¼ 5, the enhanced coagulation and filtration with

filter papers produced treated water with the residual arsenic

concentration of 13.2mg/l, which is less than the Mexican MCL

for arsenic in drinking water. Compared with the results as

shown in Fig. 1, it is noted that the maximum arsenic removal

achieved by the enhanced coagulation was almost the same

as that produced by coagulation with the same amount of

ferric sulfate and filtration with membrane. It indicates that

the role played by coarse calcite in the enhanced coagulation

might be solely to increase the size of arsenic-borne

coagulates.

For high-arsenic water treatment, another effective process

is adsorption (Jiang, 2001; Bissen and Frimmel, 2003; Dambies,

2004; USEPA, 2000). Many kinds of materials, such as activated

alumina, activated carbon, granulated ferric hydroxide, zero-

valent iron, natural iron minerals and metal-loaded polymers,

etc., are used as the adsorbent. Ahn et al. (2003) studied

arsenic removal from high-arsenic water in a mine tailing

containment system by using waste steels and zero-valent

iron, and produced treated waters with the residual arsenic

concentrations of 44 and 3 mg/l arsenic from waters initially

containing 4137 and 1400mg/l arsenic, respectively. Jiang

(2001) reported that by activated carbon adsorption 96% of

arsenic can be removed from water with initial arsenic

concentration of 300mg/l, and by coagulation with ferric ions

99% of arsenic can be removed from water with initial arsenic

concentration of 1100mg/l. Clearly, the enhanced coagulation

process with ferric ions and coarse calcite showed a very high

efficiency for arsenic removal from high-arsenic water.

The results presented in this paper were obtained from

batch tests in laboratory. If the enhanced coagulation process

is applied to treat the high-arsenic water in pilot or plant

scale, the arsenic removal might be less than 99.5% that

was achieved in batch scale, and the residual arsenic

ARTICLE IN PRESS

00 10 20 30 40 50 60 70

20

40

60

80

100

LIG

HT

TR

AN

SM

ISS

IVIT

Y, %

SETTLING TIME, min.

Without calciteWith 5 g/l calcite (38 - 44 µm)

Coagulation with 100 mg/l Fe2(SO

4)3

pH 7

Fig. 7 – Sedimentation of arsenic-borne coagulates from

coagulation of high-arsenic water with ferric ions

at pH 7 in the absence or presence of coarse calcite

particles.

WAT ER R ES E A R C H 40 (2006) 364– 372 369

concentration of cleaned water might be greater than the

Mexican MCL for arsenic in drinking water (20mg/l As) and

much greater than the World Health Organization (WHO)

guideline value (10mg/l As). Therefore, the enhanced coagula-

tion process may be used as primary treatment for the high-

arsenic water from the mine drainage system. Another

process (which could be conventional coagulation, adsorp-

tion, ion exchange, etc.) has to be followed to deeply remove

arsenic from water in order to meet the Mexican MCL or WHO

guideline value.

3.3. Mechanisms of the enhanced coagulation

Coagulates from the enhanced coagulation were photo-

graphed by using a SEM. Fig. 6 gives two SEM micrographs.

In the images, the large particles which were similar to cubes

were of calcite, and the small irregular particles were of

arsenic-borne coagulates. As it can be observed, the calcite

surfaces (right sides and lateral sides) were coated by a lot of

small coagulates, some of which were very fine (less than

1 mm). Because of the coating, the small arsenic-borne

coagulates behave similar to the coarse calcite in gravita-

tional sedimentation and filtration, and thus are easily

removed from water. In addition, EDS attached to the SEM

was used to chemically analyze the small coagulates on the

calcite surfaces in the images. Numerous points on the

coagulates were chosen. Arsenic was found in the coagulates,

ranging from 0.8% to 5.3% As. Indeed, they were of arsenic-

borne coagulates.

Fig. 6 – SEM micrographs of arsenic-borne coagulates

coating coarse calcite particles.

The enhanced coagulation was also studied through

sedimentation tests. Fig. 7 illustrates the light transmissivity

of the high-arsenic water treated by coagulation with 100 mg/l

ferric sulfate at pH 7 as a function of settling time, in the

absence and presence of coarse calcite (38–44mm). Error

analysis for three tripled measurements showed that the

standard error for light transmissivity ranged from 0 to

70.8%. Such small error bars were difficult to be clearly

presented, so that only arithmetic average values of light

transmissivity appeared in the graph. As it is noted, without

calcite, the light transmissivity increased very slowly with

increasing settling time, indicating a slow coagulate sedi-

mentation. Clean water in the sedimentation tube (100% light

transmissivity) was obtained after 22 h settling. However, 5 g/l

coarse calcite coagulates settled down much more rapidly,

leaving clean water in the tube after 14 min. These results

suggest that coarse calcite considerably enhanced the coagu-

lation of the high-arsenic water, and greatly increased the

sedimentation of arsenic-borne coagulates.

According to Stokes equation of sedimentation, the term-

inal settling velocity of a particle in a liquid is proportional to

the square of the diameter of the particle. So, the coarse

calcite should have a much higher settling rate than small

arsenic-borne coagulates. If small coagulates coat on coarse

calcite particles, they would settle down in the same velocity

(or a little bit higher) as the calcite particles. This is the

mechanism by which the sedimentation of arsenic-borne

coagulates was considerably improved by adding coarse

calcite particles.

Fig. 8 shows light transmissivity of the high-arsenic water

treated by enhanced coagulation with 100 mg/l ferric sulfate

and 5 g/l coarse calcite as a function of settling time at various

pH. As it can be seen, the settling rate of arsenic-borne

coagulates decreased with the increase of pH. Light trans-

missivity of 100% appeared after 9 min at pH ¼ 6, compared

with 29 min at pH ¼ 9. These results might be attributed to

that coarse calcite had a stronger ability to carry small

arsenic-borne coagulates at pH ¼ 6 than at pH ¼ 9. In other

ARTICLE IN PRESS

00 5 10 15 20

20

40

60

80

100LI

GH

T T

RA

NS

MIS

SIV

ITY

, %

SETTLING TIME, min

pH 6 pH 7 pH 8 pH 9

Coagulation: 100 mg/l Fe2(SO

4)3

5 g/l calcite (38 - 44 µm)

Fig. 8 – Effect of pH on the sedimentation of arsenic-borne

coagulates from the enhanced coagulation of high-

arsenic water.

-302 4 6 8 10 12 14

-20

-10

0

10

20

30

ZE

TA

PO

TE

NT

IAL,

mV

pH

Arsenic-borne coagulatescalcite

0.01 mol/l KNO3

25°C

Fig. 9 – Zeta potentials of calcite particles and arsenic-borne

coagulates from the coagulation of high-arsenic

water with 100 mg/l ferric sulfate in 0.01 mol/l

KNO3 solution as a function of pH.

WA T E R R E S E A R C H 4 0 ( 2 0 0 6 ) 3 6 4 – 3 7 2370

words, coarse calcite enhanced the coagulation more strongly

at acidic pH than at alkaline pH.

The zeta potentials of calcite particles and arsenic-borne

coagulates from the coagulation with 100 mg/l ferric sulfate in

0.01 mol/l KNO3 solution as a function of pH are illustrated in

Fig. 9. It shows that the isoelectric points (IEP) of the calcite

particle and the arsenic-borne coagulate were pH 9.3 and pH

4.9, respectively. In the range of pH 4.9–9.3, calcite particles

were positively charged, while arsenic-borne coagulates were

negatively charged.

The aggregative stability of colloidal suspensions is due to

the existence of a potential energy barrier between particles,

preventing the proximity of the particles. Such a study of

stability is known as the Derjaguin–Landau–Verwey–Over-

beek (DLVO) theory (Derjaguin and Landau, 1941; Verwey and

Overbeek, 1948). According to this theory, the potential energy

barrier arises as a result of interaction energies of electrical

double layer and van der Waals. In a suspension of hetero-

geneous particles, the potential energy of electrical double

layer interaction between two heterogeneous spheres (VR) is

expressed by (Wang, 1991)

VR ¼ 32peoer2a1a2

a1 þ a2

kTeu

� �2

tanheuc1

4kT

� �tanh

euc2

4kT

� �expð�khÞ,

(2)

where a1 and a2 are the radii of the particles 1 and 2,

respectively; eo is the vacuum dielectric permittivity; er is the

relative dielectric permittivity of the medium; k is the

Boltzmann constant; T is the absolute temperature; e is the

elementary charge, u is the ionic valence of electrolyte, c1 and

c2 are the outer Helmholtz plane (OHP) potentials or zeta

potentials of particles 1 and 2, respectively, k is the Debye

reciprocal length and h is the shortest separation between the

two particles. The potential energy of van der Waals interac-

tion between two heterogeneous particles (VA) is expressed by

(Hiemenz and Rajagopalan, 1997)

VA ¼ �a1a2

a1 þ a2

A132

6 h, (3)

where A123 is the Hamaker constant of particles 1 and 2 in

medium 3, which may be obtained by (Israelachvili, 1992)

A132 ¼ffiffiffiffiffiffiffiffiA11

p�

ffiffiffiffiffiffiffiffiA33

pÞð

ffiffiffiffiffiffiffiffiA22

p�

ffiffiffiffiffiffiffiffiA33

p� �, (4)

where A11, A22 and A33 are the Hamaker constants of particles

1, 2 and medium 3 in vacuum.

In the case of this work, the Hamaker constants of the

arsenic-borne coagulates and calcite are larger than water, i.e.

A114A33 and A224A33. So, A12340, and thus VAo0 according

to Eqs. (4) and (3), indicating an attractive interaction of van

der Waals between the two heterogeneous particles. In the

range of pH ¼ 4.9–9.3, c1o0 (arsenic-borne coagulates), and

c240 (calcite), leading to tanh euc14 kT

� �o0 and tanh euc2

4 kT

� �40, and

thus VRo0 according to Eq. (1). It means that the electrical

double layer interaction was also attractive between the two

heterogeneous particles. Therefore, in the pH range, the total

potential energy of interaction between the arsenic-borne

coagulate and the calcite particle was attractive at every

distance. In other words, there was no potential energy

barrier between the two particles. A strong coagulation

should employ to the arsenic-borne coagulates and calcite

particles. This is the mechanism by which small arsenic-

borne coagulates coated on coarse calcite, as shown in Figs. 6

and 8.

In addition, there are interests to understand the impacts of

other factors on arsenic removal from the high-arsenic water

by the enhanced coagulation process, e.g., ions, organic

carbon and ionic strength in initial water, chemiosorption of

arsenate ions on calcite surfaces, etc., which are our current

and future research interests. Another further work in this

regard is to build up a pilot system for the enhanced

coagulation process to eliminate arsenic from the high-

arsenic water in the mine drainage system. The state

government of San Luis Potosı has approved financial support

for the pilot-scale tests.

The experimental results from this study have shown that

the enhanced coagulation process is a very cost-effective

technique for arsenic removal from high-arsenic water.

ARTICLE IN PRESS

WAT ER R ES E A R C H 40 (2006) 364– 372 371

Because of the high settling rate, the arsenic-borne coagulates

from the enhanced coagulation can be first concentrated

from water by gravitational sedimentation, and then sepa-

rated from water by filtration with a large-pore filter medium,

producing a very low arsenic concentration of cleaned water.

Compared with conventional coagulation/filtration process

for arsenic removal, the enhanced coagulation process

features lower capital and operation costs, simpler in opera-

tion and lower task in maintenance. Since coagulation is a

commonly and widely used process for water treatment

worldwide, it will not be hard to realize the practical uses of

the enhanced coagulation process. Coarse calcite may be

added together with ferric sulfate and other chemical

reagents, which would not bring great difficulty. For the steps

of conditioning, sedimentation and filtration, the enhanced

coagulation could be the same as conventional coagulation/

filtration process. However, the disposal of the sludge from

the enhanced coagulation process has to be studied to meet

the Mexican regulation for toxic waste disposal, which will be

our further research subject together with the pilot-scale

studying.

4. Conclusions

(1)

Coagulation/filtration for arsenic removal from a high-

arsenic water with ferric ions as the coagulant could be

considerably enhanced by adding an appropriate amount

of coarse calcite (38–74mm). This enhancement may be

attributed to the coating of small arsenic-borne coagulates

on calcite surfaces, improving the gravitational sedimen-

tation and filtration of the coagulates greatly.

(2)

The coating of small arsenic-borne coagulates on coarse

calcite may be due to the attractive electrical double layer

interaction between the coagulates and coarse calcite

because of the reverse zeta potential of the two particles,

which leads to non-existence of a potential energy barrier

between the two particles.

(3)

The enhanced coagulation with ferric ions and coarse

calcite and conventional filtration (filter paper as filter

medium) achieved a very high arsenic removal (over 99%)

from a high-arsenic water in mine drainage system (5 mg/l

As), giving the residual arsenic concentration of 13mg/l.

Acknowledgements

The financial support from the Consejo Nacional de Ciencia y

Tecnologıa (CONACyT) of Mexico and the government of the

State of San Luis Potosı under the Grant FMSLP 2002-5537 is

gratefully acknowledged.

R E F E R E N C E S

Ahn, J.S., Chon, C.M., Moon, H.S., Kim, K.W., 2003. Arsenicremoval using steel manufacturing byproducts as permeablereactive materials in mine tailing containment systems. WaterRes. 37, 2478–2488.

Arienzo, M., Adamo, P., Chiaenzelli, J., Bianco, M.R., de Martino, A.,2002. Retention of arsenic on hydrous ferric oxides generatedby electrochemical peroxidation. Chemosphere 48, 1009–1018.

Bissen, M., Frimmel, F.H., 2003. Arsenic—a review. Part II:oxidation of arsenic and its removal in water treatment. ActaHydrochim. Hydrobiol. 31 (2), 97–107.

Chiba, A., Okada, H., Tada, T., Kudo, H., Nakazawa, H., Mitsuhashi,K., Ohara, T., Wada, H., 2002. Removal of arsenic fromgeothermal water by high gradient magnetic separation. IEEETrans. Appl. Superconductivity 12 (1), 952–954.

Dambies, L., 2004. Existing and prospective sorption technologiesfor the removal of arsenic in water. Separation Sci. Technol.39, 599–623.

Derjaguin, B.V., Landau, L., 1941. Theory of molecular interaction.Acta Physicochem. 14, 633–668.

Fuerstenau, D.W., Chander, S., Abouzeid, A.M., 1979. The recoveryof fine particles by physical separation methods. In: Soma-sundaran, P., Arbiter, N. (Eds.), Beneficiation of MineralFines—Problems and Research Needs. AIME, New York,pp. 3–59.

Green, E.W., Duke, J.B., 1962. Selective froth flotation of ultrafineminerals or slimes. Trans. AIME 223, 389–395.

Gulledge, J.H., O’Connor, J.T., 1973. Removal of arsenic (V) fromwater by adsorption on aluminum and ferric hydroxides.J. AWWA 65 (8), 543.

Han, B., Zimbron, J., Runnells, T.R., Shen, Z., Wickramasinghe,S.R., 2003. New arsenic standard spurs search for cost-effective removal techniques. J. AWWA 95 (10), 109–118.

Hering, J.G., Chen, P.Y., Wilkie, J.A., Elimelech, M., Liang, S., 1996.Arsenic removal by ferric chloride. J. AWWA 88 (4), 155–167.

Hering, J.G., Chen, P.J., Wilkie, J.A., Elimelech, M., 1997. Arsenicremoval from drinking water during coagulation. J. Environ.Eng. 8, 800–807.

Hiemenz, P.C., Rajagopalan, R., 1997. Principles of Colloid andSurface Chemistry, third ed. Marcel Dekker, New York.

Israelachvili, J.N., 1992. Intermolecular and Surface forces, seconded. Academic Press, Inc., New York.

Jiang, J.Q., 2001. Removing arsenic from groundwater for thedeveloping world—a review. Water Sci. Technol. 44 (6), 89–98.

Katsoyiannis, I.A., Zouboulis, A.I., 2002. Removal of arsenic fromcontaminated water sources by sorption onto iron-oxide-coated polymeric materials. Water Res. 36, 5145–5155.

Katsoyiannis, I.A., Zouboulis, A.I., 2004. Application of biologicalprocesses for the removal of arsenic from groundwaters.Water Res. 38, 17–26.

Korngold, E., Belayev, N., Aronov, L., 2001. Removal of arsenicfrom drinking water by anion exchange. Desalination 141,81–84.

Kumar, P.R., Chaudhari, S., Khilar, K.C., Mahajan, S.P., 2004.Removal of arsenic from water by electrocoagulation. Che-mosphere 55, 1245–1252.

O’Connor, J.T., 2002. Arsenic in drinking water. Part 2: humanexposure and health effects. Water Eng. Manage. 149 (3), 35–37.

Raje, N., Swain, K.K., 2002. Purification of arsenic contaminatedground water using hydrated manganese dioxide. J. Radioanal.Nucl. Chem. 253, 77–80.

Razo-Soto, I., Carrizales, L., Castro, J., Diaz-Barriga, F. and Monroy,M., 2005. Arsenic and heavy metal pollution of soil, water andsediments in a semi-arid climate mining area in Mexico.Water, Air Soil Pollut., in press.

Sato, Y., Kang, M., Kamei, T., Magara, Y., 2002. Performance ofnanofiltration for arsenic removal. Water Res. 36, 3371–3377.

Shen, Y.S., 1973. Study of arsenic removal from drinking water.J. AWWA 65 (8), 543.

Somasundaran, P., 1980. Role of surface chemistry of finesulphides in their flotation. In: Jones, M.J. (Ed.), ComplexSulphide Ores. The Institute of Mining and Metallurgy,London, p. 126.

ARTICLE IN PRESS

WA T E R R E S E A R C H 4 0 ( 2 0 0 6 ) 3 6 4 – 3 7 2372

Song, S., Lopez-Valdivieso, A., 2002. Parametric aspect of hydro-phobic flocculation technology. Miner. Process. Extr. Metall.Rev. 23, 101–127.

Sorg, J.T., Logsdon, G.S., 1978. Treatment technology to meet theinterim primary drinking water regulations for inorganics.Part 2. J. AWWA 70 (7), 379–392.

Thirunavukkarasu, O.S., Viraraghavan, T., Subramanian, K.S.,2003. Arsenic removal from drinking water using iron oxide-coated sand. Water Air Soil Pollut. 142, 95–111.

Thirunavukkarasu, O.S., Viraraghavan, T., Subramanian, K.S.,Chaalal, O., Islam, M.R., 2005. Arsenic removal in drinkingwater—impacts and novel removal technologies. EnergySource. 27, 209–219.

USEPA., 2000. Technologies y costs for removal of arsenic fromdrinking water. EPA/815/R-00/028, Washington.

Verwey, E.J., Overbeek, J.G., 1948. Theory of the Stability ofLyophobic Colloid. Elsevier, Amsterdam.

Wang, Q., 1991. Theoretical analysis of Brownian heterocoagula-tion of fine particles at secondary minimum. J. Colloid Inter-face Sci. 145, 305–313.

Wang, L., Chen, A.S.C., Sorg, T.J., Fields, K.A., 2002. Fieldevaluation of as removal by IX and AA. J. AWWA 94, 161–173.

Wickramasinghe, S.R., Han, B., Zimbron, J., Shen, Z., Karim, M.N.,2004. Arsenic removal by coagulation and filtration: compar-ison of groundwaters from the United States and Bangladesh.Desalination 169, 224–231.

Zhang, Y., Yang, M., Huang, X., 2003. Arsenic(V) removal with aCe(IV)-doped iron oxide adsorbent. Chemosphere 51, 945–952.

Zhang, W., Singh, P., Paling, P., Delides, F., 2004. Arsenic removal fromcontaminated water by natural iron ores. Miner. Eng. 17, 517–524.