sorption

important heavy metal

cement. A

and C-S-H (I) stands for
main mineral component of hardened
combination of techniques has been used

determine the relevant

nisms in this
laboratory model. Macroscopic wet-chemical techniques

experiments

carried

in order to

the interaction of Zn and C-S-H(I) in

aqueous environment.

Microscopic
spectroscopic techniques,

X-ray powder

fraction (XRPD), electron
probe microanalysis (EPMA),

X-ray absorption

fine structure

of Zn in the

spectroscopy (XAFS),

performed

investigate

the distribution and

binding

phase.

sorption experiments

The results of the Zn

distinct kinetics, which indicates

2-step

mechanism. After
very fast sorption in the first 12 hours, the

continues

least 3 months. This slow
kinetics correlates with the

uptake

by the C-S-H(I) particles,

After the initial

has been observed

by EPMA

different Zn concentration levels.

sorption step,

predominantly be
the surface of the C-S-H(I)

particles.

The Zn-rich rims then become thicker with time, Zn diffuses into the

particles

finally
uniform Zn distribution in the C-S-H(I)

is observed. This

is most
probably caused by incorporation

methods,

incorporation

solid-solution

of Zn into the C-S-H (I)
interlayer. However, this

spectroscopic

kind of

is not

the classical

sense.

Even with the

the exact mechanism

molecular level could used in this
conclusively identified. potential

reactions

With the combination oi

excluded

techniques

study,

mechanisms of Zn

C-S-H(I):

exchange

and the

precipitation of Zn minerals do

control the Zn

speciation
aqueous Zn concentrations

(corresponding

molar fraction of Zn sorbed to

C-S-H(I) of

Specific adsorption

the initial

dominating sorption mechanism, though

it could

play a

role in

sorption stage.

The formation of

crystalline

Zn minerals such

2-Zn(OH)?

and calcium zincate

is relevant
initial Zn concentrations,

highly supersatu

rated with respect to these minerals.

standing

prediction

behavior of heavy metals in cement-based
materials. Predictions, however, will

always

uncertainties.

Scope of

This Work

assessment

geochemical concept

for the work

from cement-based

materials formed the

background

following chapters.

scope of this work

extend the

knowledge

the relative

different

binding mechanisms

of heavy metal cations

to cement

minerals. The

following

specific questions

be examined:

Which methods
provide useful information

about the relevant

sorption mechanism in

minerals in the

specific laboratory model?
sorption of the heavy metals
Which mechanisms dominate the

laboratory system

metals?
for the measured aqueous concentrations of heavy
What is the relevance and

applicability of the

be drawn

results from the

in the labora

tory model for

complex,

real systems such

landfills with cement-based
materials? Which conclusions

regarding the

systems?
long-term leachability of heavy

metals in such

following way:

This includes

review
2 presents the evaluation of
appropriate investigation concept.

potential binding

mechanisms and the

investigation

heavy metal binding
mechanisms in cement-based waste materials, the choice of

appropriate

model substances for
minerals and the selection of methods.

macroscopic, microscopic

spectroscopic investigation

3 the chosen

substances and methods

in detail.

firstly illustrates

the results from

macroscopic,

wet-chemical

the interaction of the

metal and the cement

mineral in aqueous environment.

Secondly, it presents

structural information about the

metal in

the solid

obtained from

methods.

Chapter 5,

the results from the individual

methods

combined and metal in the labo

pellets.

glassware
and HDPE flasks used for the

analyses

leached with acid
M diluted from concentrated

HNO,) for using

least 24 hours. Filtration

polycarbonate

Whatmann)
filtration unit (Sartorius)

membrane filter discs

vacuum.

(0.45-pm nylon,

nitrogen atmosphere

measured

proton concentration using

combined

electrode (Metrohm

6.0202.100) connected

digital voltmeter (Metrohm 713).

The electrode

calibrated

dosimat

titrating

25 ml of

0.01 M HCl solution with up to J 5 ml 0.05 M NaOH

(Metrohm 665). Both

solutions
prepared by diluting commercially

available 1.0-M

standard solutions in the

glovebox with

of 0.1 M

(Na)Cl

electro

Standard

potentials

slopes

determined

linear

regression

of the measured

electro-motoric force (EMF) and the calculated scale. The calibrations and all

measurements

the acidic and alkaline part of the

under argon.

Zinc measured
The dissolved Zn concentrations above 0.1 ppm

absorption spectro

metry (AAS, Perkin-Elmer 5000)

at 213.9 nm

acetylene-air

flame. The calibration

carried out with 5 standard solutions between 0.050 and 1.0 ppm. All measurements

triplicate

checked for matrix effects

by analyzing

samples by
standard addition. Below 0.1 ppm, dissolved Zn

differential

stripping voitam

sample

metry (DP-ASV, Metrohm VA-Stand 694, VA-Processor 693). To achieve this, 10 ml
solution and I ml of in 100
ammonia-acetic acid buffer

C2H5OH

s.p. and 13.62

added into

Polarographie

vessel. After

degassing the
solution for 5 minutes recorded between -1150 determined

studies of the initial Zn

Chapter 3.9).

The observed Zn removal of
then slowed down considerably, and after 4

equilibration,

than 90%

sorbed (see

In order not to
strongly influence the sorption

by kinetic effects,

reactions in

minimal

equilibration

time of 4

chosen for the

sorption experiments.

continued up to 3 months to determine

larger

time scales.

Experiments

3 orders of

out at

three different
pFI values (11.7,12.48,12.78)
the influence of OH~ concentration. Initial total Zn concentrations

varied

nearly

magnitude.

The lowest concentration (4.8

given by the

determination The
limit (15 nM) for the aqueous Zn measurements

stripping (DP-ASV).

concentrations (980 and 4800

chosen

with respect

control and

and not

produce precipitates

preparation
of the Zn stock solutions.
obtain further information about the

experimental system,

experi

ments were

conducted without addition of the

C-S-H(I) suspension

only in

electrolyte

solution at the
value (without Ca and Si).

Table 3-1:

conditions of the
sorption study experiments. Equilibration times

5,15,30,60 rain

Initial total Zn concentrations

19 11.8,9

2,4,8,24 h 2,4,28,56,87 d

6,48,96,190,480,960

4,28,56,87 4

4.8,19,96

4,28,56,87 d

4,28,56,87 d 4,87 d

6,48,190,480,960,1900,4800

4 8,19,480

9 6,48,96,190,960,1900,4800

needed

Because of the
high solubility of C-S-H(I), the experiments

be carried out in

pensions that had been pre-equilibrated with respect

To avoid

changes

stoichiometry

of C-S-LI(I), Ca and Si from stock solutions

0.1 M NaCl

pensions. The composition of a

be carried

be determined

concentrations of Ca

out are

and Si at the three

values

were to

in Table 3-2. The values agree wcl with

ratio of 1

published

data for the C-S-H

solubility with

(Fujii

and Kondo, 1983; Gartner and

Jennings, 1997; Berner,

1988; Kindness et ab,

1992).

Table 3-2:

Composition

withCSH(I).

prcsaturated

OH"

12.48 12.78

0.060 0.050
produced by mixing appropriate volumes

1.0 mM

following stock solutions:

the solutions

SiCl, and 1.0 M NaOH. The ionic strength of

the necessary

amounts

always kept

at 0.1

2.0 M NaCl

solution. A stock

C-S-H(I) (S2)

prepared by adding

1 g of

to 0.5 L

presaturated

solution and

equilibrated

the rotary shaker

rpm. To

obtain the final

C-S-H(I) suspension for

experiments,

the stock

diluted

adding

1 ml of S2 to 50 ml SI in

50-ml HDPE bottle and

equilibrating again

stock solutions at

neutral pFI

prepared directly before

sorption experiments with

(except

hydroxide concentration).
of the Zn stock solutions

For the

sorption experiments,

aliquots

added to the
C-S-H(I) suspensions. Hie suspensions

150 rpm. After

value of the solution

measured in

aliquot.

The solid

remaining sample
filtered and acidified with 0.3 ml of concentrated

phase samples

freeze-dried and afterwards

analyzed by XRPD

determine

embedded in epoxy resin (Araldit, Ciba

Geigy)

and hardened under pressure

(6 bar). The

embedded

polished

subsequently
coated with about 200 A of carbon

(Balzers

Union Carbon

Evaporation

Device CED

surfaces, the powder

adhesive

graphite tape

subsequently coated

about 200 of carbon.

The EPMA

performed by Reto

automated electron micro-
microscope (SEM) (semi-quantitative

collected in the

element

distribution maps: IEOL

Superprobe JXA-8800L; quantitative analysis:

CAMECA SX-50; SEM

images:
IEOL SEM). The distribution maps
wavelength-dispersive mode
(WDS). llie characteristic X-rays of Zn (Zn-Ka),

(Ca-K ) and Si

of 100 nA)

(Si-Ka)

generated voltage

focused electron beam

(probe

current

acceleration their
15 keV. The intensities of the

generated X-rays

according to

wavelengths

following crystal
spectrometers: LiF for Zn, PET for Ca and TAP for Si.

distribution maps (stage

Data for the
semi-quantitative respective

and lines

collected for 140

each the

characteristic

of each element. The

quantitative analysis

selected

points

live time of 30

for each element.

Background

system was

by measuring

for 15

suitable distances

the front and back of

peaks.

zincite (99.99 wt % ZnO) and

pyroxene

Na O, 16.38

CaO, 56.88

standard

the measurements.

X-ray Absorption Fine Structure (XAFS) Spectroscopy

For the Zn-XAFS

analysis, samples

the C-S-H (I)

studies have to be

produced days

quantities, lliereto,

to 4000

added and

156.9 mg of

ml of the
solution SI in 5000-ml HDPE bottles. After 7

equilibra

tion, 200 ml of the appropriate Zn stock solution

equilibrated. The

Zn-treated

Table 3-4.

filtered and freeze-dried. The

Table 3-4:

conditions of the Zn-treated

measured bv Zn-XAFS.

Sample

Initial total

Zn concentration

Equ:ilibration time

11.7 11.7 11.7

30 min

Two additional

(except the sorption sample A^1).
with the lowest Zn concentrations,

limited up to f 0.0

Structural parameters

extracted

across

using non-linear

least square

fitting.

The fits

performed in

real space

the range of the first

shells

to 4.0

Theoretical the

scattering paths for

the fit
calculated with FEFF 7.02
(Zabinsky et ah, 1995), using
hemimorphite. As fitting parameters
served the interatomic distances (R), coordi

nation numbers

accounts
Debye-Waller factors (Ao~2)
for the individual shells. The latter
for static disorder and thermal vibration of the

backscattering

causes

amplitude damping

parameter for all
of the XAFS spectrum. The difference in threshold energy

was as a

theoretical reference functions and the unknown spectrum

single adjustable

of backscattering atoms for each

Fitting was

done until

best fit

obtained between the

predicted

curves.

Derived XAFS parameters

between the fitted and the

tested

by fitting the

reference

sample spectra.

The deviation

spectra is given
relative residuals in percent

(%Res), defined by

XbexpW-Y thecal
with N, the number of data

points,

yeri and ytheo, the

and theoretical data

points, respectively.
The accuracies of the derived structural parameters minerals
be estimated from the reference

by a comparison

data with literature values

(Fable

4-2 and Table
Thus the estimated accuracies for the interatomic distance

0.01-0.02

for the first shell and AR

tion number

0.03-0.05
A for the second shell. The accuracy for the coordina
Thermodynamic Equilibrium Calculations
Thermodynamic equilibrium

programs

with the aid of the computer
MQV40TIT (Furrer, 1995, based
and WestaU, 1997). The used
MICROQL by WestaU, 1986) and FITEQL

an are

(Herbelin

stability

shown in Table 3-5 and Table of 0.1 M in accordance
3-6. The values from literature

corrected for

by using activity

coefficients calculated

Davies

Equation:

coefficient of species

0.5 for water at 25 C,

of species i,

concentration of ionic

species I,

solution,

Table 3-5:

Stability

of formation used for the thermodynamic calculations:

phases (I

Equilibrium

reaction

2H" 2 IF

2ir -11.53

Baes and Mesmer
precipitated ZnO ignited ZnO amorphous Zn(0II)2

Zn:++II,O^ZnO

(1974)

H2O^ZnO-t

-11.36

-12.70

2H2O^Zn(OH), +

Schindler

(f 964)

rZn(0H)2 2-Zn(0II)2

Y-Zn(0H)2

Zn2+ 11H20 ^ Zn(OH),

-11.98

-12.02

Schindler Schindler

(1964)

(1964) (1964)

211,0 ^

Zn(OH)2

2II2O^Zn(OH)2-L2fI+

Zn(OH),

-11.96

-12.07

-Zn(0H)2

f-Zn(OII),

Zn-f + 2 HA) ^

ZifM- 211,0 # Zn(OH),

-11.IT -46.42 -13.59

-45.20 -21.02

Zn2" + Ca2'

H20 ^ Zn2Ca(0H)6

4fF 2 H'

Cockeetal.(1997)

2ZirA-II(Si04v^Zn,SiO14

Lindsay (1979)

Berner

H4SiO,

CaH2Si04

(1988)

2Il:0^Ca(0H)2

portlandite

Ca-|-+2H:O^Ca(OH)2

H,Si0,^Si02 +

-23.03 2,70 4.0

amorphous SiO,

quartz

Eikenberg(1990)

Grenthe(l992)

fb,Si0,^SiO2

Chapters

Table 3-6:

Stability constants

of formation used for the

thermodynamic

calculations:

dissolved species (1

Equilibrium reaction

H20 ^ H+

Zn2t Zn2'

-13.78

Smith and Martell (1976)
Bacs and Mesmer Bacs and Mesmer Bacs and Mesmer Baes and Mesmer

H2O^ZnOEf

-17 12

-28 10

(1974) (1974)

2H?O^Zn(OH)2 + 2ir

Zn2+-t Zif+ +

Zn(OH)f+

H20 #=

Zn(OH)42-+ 4 H>

-40.76

H2O^Zn2OH3++H+

Zn2+ +

H20 ^ Zn2(OH)62'

-57.58

( 1974)

Zn2h + Cl ^ZnCF
Smith and Martell Smith and Martell

(f 982) (1976)

Zn2++2CL^Zn(Cl),

3CL^Zn(CLY

Smith and Martell (1982)

Smith and Martell

Zn2+ + 4CF^Zn(Cl)(^

(1982)

Zn2,"+C032"^ZnCO30

1045 -13 07

der Sloot et ab,

"S*1

""**

{> 1 [IL

5"TU-t> o*

o 09 Q.

J3-'Lf.

> A

VX rf tews

| W* 1
Sorption process Specific adsorption (surface complexation)

either bv

={R,2~Ca2+}[NaT

XJNa 12

>aR.c.lR)
"{R-Naf [Cyi"xN;[Ca21
constancy (Sposito, 1984). LLsing typical values for the selectivity coefficient Q (1.7-3.3,

Kn=[Zn2+]fOLI12

and Reardon,

CaxZn{1_ TOH)2

be described in the
following way (according to Glynn

Ca2-(aq)

^Ci(0II)2

[Zirl I Of Lf

fZn2,l[OH

between the

activity (concentration) phase

1ISAlJ

^-iL-i iLl + KJA]

pH, there

logMeaq

(mol/dm3)

The two vertical lines represent the solubility concentrations of Me for the stable Me oxide and a metastablc precursor (e.g., a hydrated Mc oxide
called alite, belite, aluminate and

anhydrous

grains,

hydrated coating is

formed around
in cement pastes is often called "C-S-LI

of C-S-FI

ab, 1996), but their structural models partly differ. The probably
represent C-S-H domains, the open circles adsorbed on the.surfaces (from Classer, 1993b

Hie main

small number of Livers that

Lnci of 1 1

tobeimoute, the pictuies

Paits of the sttuauie ol

single pioicction along the

(Llamid,

) The tctiahcdia lepiesenl SiOt units, the black filled cucles the mam layer, the giav-shaded Ca m the mteilayei and the
open elides LI O molecules

Oil" in the

of Si-OLI groups in the main
increased content of Si-OLI groups (the

pLI declines

in cement (OFT,
probably ettringite (3CaO-ALO,-3CaSO/32LLO),

of AL' and diverse

(Zn,Ca(OH)(i-2H,0) during
L( shell of 4 oxygens and the results
calcium silicates. Llowever, the

imperfect

form of the

probably linked to

end of the silica chains in C-S-H (Moulin, 1999). fohnson and Kersten (1999) measured the

spcifie adsorption,

"C-S-H(I)
first step, the interaction of Zn wtith C-S-H(l) is of Zn to

O-Zn bonds and the spectra
the spectrum of hemimorphite

(Zn4Si207(OH)2-LI20).

However, these spectra do

significantly differ

from spectra of
low Zn concentrations, where

from bulk solution

therefore either very similar in their local Zn coordination

the observed

absolutely unstructured

and hence

contribution to the XAFS
spectra. Since both of the

coprecipitated samples

essentially

XAFS spectrum

sorption samples, the first possibility

unlikely

Hoyvever, with the present
ments, the nature of the Zn-rich

conclusively.

literature, the formation of 2:1 layer Zn silicates from alkaline solutions (pH 7.5-10)

20 C and

atmospheric pressure

is described

(Tiller

Pickering, crystalline

Leggett, 1978).

products

x-rays, but became

aging for

about I year,
stevensite type structure

(Ca03Mg,SisO10(OH)o,

Pickering (1974)

layer silicate)

and Zn:Si ratios

ranging from

1.8 to 0.8. Tiller and

concluded that Zn

formed in the presence of Si and under conditions that
des in the absence of Si.

orthorhombic zinc

hydroxi

This is the

for the present

since the orthorhombic

2-zinc

is formed at

ZnO, which is

hexagonal

formed in the present

experiments).

Llence, it

likely that

Zn silicate,

perhaps

layer silicate,

could be

from solution under the

conditions. But the such

apparently high

Zn content in

formation.

Perhaps,

Zn silicate and e.g.

High highest

Initial Zinc Concentrations initial Zn concentrations (1000 pM and above), the

12.48 and 12.78. The

be identified

by XRPD

1-Zn(OH),

11.7 and calcium zincate

(Zn2Ca(OH)6-2H?0)
12.48 and 12.78. The observed aqueous Zn concentrations agree
with literature data for the

^-zinc hydroxide,

of calcium zincate, of this

be confirmed

by the presented determination

mineral.

The kinetics of Zn indicates
lower initial Zn concentrations
2-step sorption mechanism with

fast and

of Winterthur, 1998). The

pH ofthe leachate

in the leachate

relatively loyv

with values between 10.0 and

10.5; and Zn

ments to

originate

accompanying

1998).

a case

study with

students from ETLI Zurich (VTB,
Zn concentrations in leachates from these tyvo landfill compartments

together with

data from the

studies in

figure

6-2. The

landfill leachates

beloyv the

solubility of

units. At these

2-Zn(OLI)2, although the leachates
pll values differing by more than

observed in the

laboratory experiments

important.

5<Zn2Ca(0H)6-2H20

^"*-^^

'_ 2-Zn(0H)2

^^-^"^"^

4.9%C**^4%

"/O

3.4% Q

-""

1.8% O

Comparison of dissolved

Zn concentrations in
concentrations measured in leachates from landfills yvith
cement-stabilized APC residues. The open circles (O) indicate the isotherm measurements at loyv initial Zn concentrations, the filled

symbols represent

zincate
field data from the landfills "Teuftal" () and "Riet" (A). The lines denote the solubilities of pure solid Zn minerals

laboratory system,

,-Zn(OH)2

(Zn2Ca(OH)fi-2H20,

at a total

Ca concentration of 2

mM). The

dashed line is

parallel

,-Zn(OH),.

and the other Zn

crosses

modifications is drawn
averages of field measurements. This dashed line
laboratory sorption studies
values which represent Zn concentrations in
the solid C-S-H(I) from 3.0 to 4.5 atom 'o

approximately (compare

A calcu
lation ofthe Zn amount present in the cement-stabilized APC residues with

In the first case, Zn dissolution is controlled

primarily by the

release
from clinker, but afterwards

by the sorption

compounds,
the present C-S-LI. The second

similar.

Though highly soluble during the

hydroxides,

first stage of

hydration,

resulting high
aqueous Zn concentrations will

be reduced

C-S-EI. In the third

only possibility by

high sorption

which C-S-LI minerals

reduce the aqueous Zn concentrations. With the

the sorption of Zn

the C-S-H

should control the

dissolved Zn concentrations
present in the cementitious material

providing that residence

time is

long enough. hypotheses

should be

that the above-mentioned

correct, the aqueous Zn concentrations

predictable

situations. This is indeed the

studies with

complex cement
for actual field systems with
stabilized APC residues. However, other

minerals, such

also be involved in the control of dissolved Zn concentrations.

of these indivi

would be necessary to
clarify their importance. phases
of cementitious materials, C-S-H

be carbonated and

finally trans

the solubi

formed into calcite and silica
result, the pH ofthe leachate

values below 10.

solubility of pure
Zn carbonate (smithsonite) is in the

lity ofthe

formed under these conditions, and
formation with calcite would

additionally lower

the aqueous Zn concentrations.

point of view,

barrier material

against

the release of
Zn concentrations: In the presence of C-S-H

dissolved Zn is buffered

centrations. Hie carbonation ofthe C-S-H

of the cement should not

These conclusions for Zn

simply be extrapolated

other divalent cations. As known

from research

solid-solution series, cations yvith

ionic radius

Zn21 such

Cu2\ Mg2', Ni2G Mir'

Co2' could behave

similarly. Hoyvever,

hydroxides/oxides

favorable

of these cations is several orders of

smaller than the

Hence, the formation of these
much lower concentrations

Cocke, D.E., Yousuf, A.M., LIess, T.R., Lin, T.-C,

calcium-metal

63-68.

high pLI

environments. Mat. Res. Soc.
chemistry of Symp. Proc. 432,
Cong, X., Kirkpatrick, R.J., 1996:29Si MAS NMR study hydrate. Advn. Cem. Bas. Mat. 3; 144-156.
ofthe structure of calcium silicate
Conner, J.R., 1993: Chemistry of cementitious solidified/stabilized yvaste forms. In: Spence, R.D. (Ed.), Chemistry and microstructure of solidified yvaste forms. Lewis, Boca

Raton, 41-81.

Crank, J.,
1975: The mathematics of diffusion.
2nd ed. Clarendon Press, Oxford.
Curti, E., 1999: Coprecipitation of radionuclides yvith calcite: estimation of partition coeffi

cients based

revieyv of
laboratory investigations

Geochem. 14, 433-445.

Davis, I.A., Kent, D.B., 1990: Surface complexation modeling in aqueous geochemistry. In: Hochella, M.F., White, A. F. (Eds.), Mineral-water interface geochemistry. Reviews in mineralogy Vol. 23. Mineralogical Society of America, Washington, 177-260.
Effenberger, H., Mereiter,

156,223-243.

K., Zeman, J., 1981: Crystal structure refinements of magnesite, calcite, rhodochrosite, siderite, smithsonite, and dolomite, with discussion of some
aspects of the stereochemistry of calcite type carbonates. Zeitschr. Kristallographie

Eighmy, TT, Eusden,

I.D., JR., Kryzanoyvski, J.E., Domingo, D.S., Stmpfli, D., Martin, LR-, Erickson, P.M., 1995: Comprehensive approach toward understanding element
speciation and leaching behavior m municipal solid precipitator ash. Environ. Sei. dechnol. 29, 029-646.
incinerator electrostatic

Eikenbcrg, k,

1990: On the

of silica

high pH.

PSI report

74. Paul
Scherrer Institut (PSI), Wrenhngen, Syvitzerland.

Farley, K.J., Dzombak,

D.A., Morel, F.M.M., 1985: A surface precipitation model for the on metal oxides. J. Colloid Interface Sei. 106, 226-242.
data for calcium silicate
Fujii, K., Kondo, W, 1983: Estimation of thermochemical (C-S-LI). L Am. Ceram. Soc. 66, C220-C221.

hydrate

Furrer, C, 1995: MQV40ddT (version 4.0), a computer program by Gerhard Furrer. Institute of Terrestrial Ecology (ITO), Schlieren, Syvitzeiiand. Gartner, E.M., Jennings, HAL, 1987: Tlierniodynamics of calcium silicate hydrates and their solutions. L Am. Ceram. Soc. 70, 743-749.