www.elsevier.com/locate/wasman

Waste Management 27 (2007) 977–985

Effects of carbonation and leaching on porosity in cement-bound waste

T. Van Gerven *, G. Cornelis, E. Vandoren, C. Vandecasteele

Department of Chemical Engineering, K.U. Leuven De Croylaan 46, 3001 Leuven, Belgium

Accepted 2 May 2006Available online 14 July 2006

Abstract

Porosity is possibly an important parameter with respect to leaching of constituents from cement monoliths. During its lifetime, thepore structure of cementitious matrices changes due to carbonation and leaching. This paper discusses the effects of both acceleratedcarbonation and continuous leaching on the porosity, and, conversely, how porosity affects leaching properties. Two sample typesare investigated: a mortar with MSWI-bottom ash substituting the sand fraction and a cement paste with 30 wt% of the cement substi-tuted by a flue gas cleaning residue. The samples have been intensively carbonated in a 20% CO2 atmosphere for up to 60 days and weresubsequently leached. The porosity was investigated by mercury intrusion porosimetry.

Accelerated carbonation decreases total porosity by 12% in the case of 60 days of treatment of bottom ash mortars, whereas contin-uous leaching during 225 days increases it by 16%. Both carbonation and leaching decrease the amount of smaller capillary pores.

Carbonation decreases both porosity and pH. Decreasing porosity diminishes leaching of sodium and potassium, while the decrease inpH increases leaching. However, the former process dominates the latter, resulting in a net decreasing effect of carbonation on the releaseof sodium and potassium from these cement matrices.� 2006 Elsevier Ltd. All rights reserved.

1. Introduction

Leaching of heavy metals from wastes is a major con-cern for waste management. Depending on the heavy metalconcentration and the physico-chemical characteristics ofthe waste, landfill disposal or reuse as construction mate-rial may be viable management options (IAWG, 1997; Sab-bas et al., 2003). In both treatment options, the addition ofcement and water in order to form a monolithic product isa well-known technique (Conner, 1990; IAWG, 1997). Forlandfill disposal the goal is to decrease heavy metal solubil-ity and decrease permeability and geometric surface area tolimit transport of contaminants to the environment.Increasing material strength is an additional requirementwhen recycling materials.

Total porosity and pore size distribution affect both thetransport of metals to the environment and the strength of

0956-053X/$ - see front matter � 2006 Elsevier Ltd. All rights reserved.

doi:10.1016/j.wasman.2006.05.008

* Corresponding author. Tel.: +32 16 32 23 42; fax: +32 16 32 29 91.E-mail address: Thomas.VanGerven@cit.kuleuven.be (T. Van

Gerven).

the monolith and are in this respect important parametersto monitor (Bishop et al., 1992; Kumar and Bhattacharjee,2003). The pore system in well compacted cement-basedmaterials consists of two types of pores (Bishop et al.,1992; Garboczi and Bentz, 1996; Hewlett, 1998; Kumarand Bhattacharjee, 2003): (a) gel pores, which are microp-ores with pore radii smaller than 0.01 lm and are includedin the volume occupied by calcium silicate hydrate (CSH)and (b) capillary pores, which are mesopores with averageradii ranging from 0.01 to 10 lm. Capillary pores are theremnants of water-filled space during hydration and areresponsible for the reduction in strength and elasticity ofthe cement monolith (Kumar and Bhattacharjee, 2003).The exact threshold between gel and capillary pores israther arbitrary (Hewlett, 1998). In mortar or concrete athird type of pores, macropores, exists because aggregateparticles are enclosed by cement particles with an interfa-cial transition zone occurring between cement and aggre-gate. The thickness of this interface is considered to be30–60 lm and is thus included in the macropores (Hewlett,1998). Macropores can also occur in cement paste due to

978 T. Van Gerven et al. / Waste Management 27 (2007) 977–985

inadequate compaction and may also affect strength(Kumar and Bhattacharjee, 2003).

Once the monolith has been produced, leaching and age-ing, and more specifically carbonation, start changing theoriginal characteristics of the material. In nature, carbon-ation and leaching occur intermittently and alternately dur-ing the lifecycle of a monolith.

Carbonation involves the chemical reaction of cementhydration products (e.g., portlandite, Ca(OH)2, andCSH) with carbon dioxide from the atmosphere in the pres-ence of sufficient moisture leading to formation of calcite,CaCO3 (Bin Shafique et al., 1998; Garrabrants et al.,2004; Van Gerven et al., 2004). This reaction lowers thepH in the pore water of the monolith from pH 13 to ulti-mately about pH 8. Heavy metals, originally present ashydroxides in the matrix, are progressively converted tocarbonates, thus changing their solubility. In addition,matrix porosity can change. The molar volume of the cal-cite mineral is 11% larger than that of the portlandite min-eral (Mollah et al., 1993; Gervais et al., 2004). An increasein mineral volume within the physical boundaries of themonolith decreases the pore volume, provided the strainsaccompanying this reaction do not cause microcrackingof the monolith structure. Progressive carbonation eventu-ally leads to depletion of portlandite and subsequentlydecalcification of CSH to provide new portlandite for thecarbonation reaction (Gervais et al., 2004). This decompo-sition of CSH is reported to increase (Gervais et al., 2004)or decrease (Ngala and Page, 1997) porosity. Pore size dis-tribution has also been shown to change by carbonation,possibly due to decalcification of CSH (Ngala and Page,1997), with a decreasing volume of small pores(<0.03 lm) and consequently increasing proportion of cap-illary pores (>0.03 lm).

Leaching is a process, in which constituents in the solidphase are transferred to the liquid phase in contact with it,followed by transport of these constituents from the poreliquid out of the monolith into the surrounding leachant,mainly by a diffusion process. The rate of leaching of met-als from porous matrices is dependent on, among others,its porosity (Bishop et al., 1992). The higher the porosityor the more pores are connected, the higher the effectivediffusivity of the constituents (Houst and Wittmann,1994; Garboczi and Bentz, 1996). Closed pores, i.e., poresnot connected with the outer atmosphere, do not come intocontact with the leachant (Bishop et al., 1992) and are thusnot relevant with respect to leaching. It has also beenreported that gel pores exhibit a conductance that is 400times lower than that of capillary pores (Bentz and Gar-boczi, 1992), indicating they present a negligible contribu-tion to leaching. Other factors affecting the rate andamount of leaching include the type and amount of theconstituent investigated, the alkalinity of the matrix, andthe chemical properties of the pore water and leachant(pH, Eh, ionic strength, etc.). Some of these parametersare further explored in a related paper (Van Gervenet al., 2006).

Although porosity influences the rate of leaching, leach-ing itself may also affect porosity. When leaching occurs,portlandite dissolves and new or larger pores develop (Gar-boczi and Bentz, 1996). Total pore volume, as well as poresize, increases, thus increasing the relative amount of cap-illary pores. Acid leaching for 90 days was reported(Bishop et al., 1992) to increase the percentage of large cap-illary pores (>6 lm) relative to total pore volume frombelow 5% to above 23%, while the percentage of smallpores (0.007–0.036 lm) decreased by 25% or more. Leach-ing with deionized water affected porosity much less.

Previous papers have investigated the effect of carbon-ation or leaching on porosity. Both effects, however, havenever been analyzed on the same specimen. This paper will,therefore, focus on the effect of both carbonation andleaching on the porosity of a monolith. The aim is to com-pare the importance of both effects. Moreover, the effect ofcarbonation on leaching is split into a porosity componentand a pH component by bringing together results from thispaper with data presented in a related paper (Van Gervenet al., 2006).

Two cement-based materials were produced using typ-ical residues from the municipal solid waste incinerationindustry: (i) a bottom ash replacing the sand fraction inmortar and (ii) a flue gas cleaning residue substitutingfor a portion of the cement in cement paste. These wasteresidues were chosen because they represent two wastestreams with distinctly different properties, but both con-taining heavy metals (Sabbas et al., 2003). Bottom ash is acoarse material consisting of mainly silica, alumina, ironand calcium oxides, as well as unburned organic matter.Bottom ash contains only a small amount of readilyleachable salts, the amount of which is even lower inthe investigated material than in untreated bottom ashdue to the wet-sieving treatment. Flue gas cleaning residueis finer material including the reaction products andexcess reagent of the air pollution control process, thusoften containing a high proportion of calcium (hydr)oxideand chloride salts. Samples from the two sample sets werecarbonated for 14, 30 or 60 days and were compared tosamples that were not carbonated. Different types of sam-ples were also leached during 225 days and compared tounleached samples. Porosity and pore size distribution isinvestigated by means of mercury intrusion porosimetry(MIP).

2. Materials and methods

2.1. Production of monoliths

Residues from a municipal solid waste incinerator(MSWI) were incorporated in cement matrices. One resi-due was a 0–2 mm fraction of a bottom ash that underwenta wet-sieving treatment as described in a previous paper(Van Gerven et al., 2005), while the other was a flue gascleaning residue originating from a scrubber with semi-dry injection of lime (Geysen et al., 2004).

T. Van Gerven et al. / Waste Management 27 (2007) 977–985 979

Mortars were produced by mixing 548 kg/m3 (28 wt%) ofordinary Portland cement (CEM I 42.5 R HES, Holcim),1096 kg/m3 (57 wt%) dried bottom ash and 281 kg/m3

(15 wt%) distilled water, giving a w/c = 0.5 (B-samples).Cement pastes were produced by mixing 900 kg/m3

(48 wt%) of ordinary Portland cement, 385 kg/m3 (21 wt%)flue gas cleaning residue and 578 kg/m3 (31 wt%) distilledwater (F-samples). The flue gas cleaning residue has pozzo-lanic properties and acts as a 30 wt% substitution of thecement. The resulting w/c = 0.45 (including the waste ascement).

The mixtures were poured in moulds of 150 · 150 ·150 mm and vibrated. After 24 h of setting time, the sam-ples were demoulded and cured for 28 days in a humidroom (20 �C, >95% RH, 0.035% CO2). At the end of thecuring period, approximately 1.5 cm of material was cutfrom the edges of the monoliths to obtain a fresh uncar-bonated surface and the remaining cubes were cut intosamples of 40 · 40 · 40 mm using a dry cutting technique.

2.2. Carbonation and leaching

One set of samples was dried for 1 h in a vacuum oven at40 �C to remove the excess of water on the surface of themonolith and subsequently stored in a bag filled with nitro-gen gas, awaiting further treatment (uncarbonated sam-ples: B00 and F00). Another set was placed in a closedchamber with the atmosphere at a temperature of 37 �C,RH over 90% and containing 20% of CO2. Some B-sampleswere carbonated for 14 days (B140), others for 30 days(B300) and yet another set for 60 days (B600). F-sampleswere carbonated for 30 days (F300) and 60 days (F600).After carbonation the samples were dried in a vacuum ovenfor 1 h at 40 �C and stored in a nitrogen atmosphere. Theduration of carbonation was based on the progressive car-bonation front in the B-samples. This front was qualita-tively assessed by regularly cutting a sample in two andspraying the inner surfaces with a solution of 1 g phenol-phthalein in 50 ml ethanol 95% and 50 ml distilled water.Phenolphtalein is a pH indicator with a red color atpH > 10 (non-carbonated cement matrix) and colorless atpH < 9 (carbonated cement matrix). After 60 days of car-bonation, phenolphthalein gave no color to the inner sur-faces of a B-sample, indicating that carbonation wascomplete.

Three samples of each mixture and each period of car-bonation were leached in a semi-dynamic monolith leach-ing procedure modified from the Dutch diffusion testNEN 7345 (NNI, 1995) or the similar Mass Transfer Testfor Monolithic Materials MT001.1 (Kosson et al., 2002).The test procedure was modified by prolonging exposureto a cumulative 225 leaching days and, in comparisonto MT001.1, the use of acidified leachant. Each cubewas immersed in 320 ml of distilled water (initiallyadjusted to pH 4 with concentrated HNO3) that was col-lected and renewed with fresh leachant 17 times aftercumulative leaching times of 6, 24, 54 and 96 h and 9,

16, 25, 36, 49, 64, 81, 100, 121, 144, 169, 196 and 225days. The volume of the leachant equals five times thevolume of the cube. The B-samples leached during themaximum period are labelled B0225, B14225, B30225and B60225 (depending on the degree of carbonation);F-samples are labelled F0225, F30225 and F60225. Afterleaching, the monoliths were immediately dried in a vac-uum oven at 40 �C until the weight decrease by dryingwas less than 1% in 24 h, and stored in a nitrogenatmosphere.

2.3. Analytical techniques

Bars of 1 · 1 · 2 cm were cut from the stored carbon-ated and/or leached monoliths in such a way that one side(1 · 1 cm) of the bar is the original outer surface of themonolith. Porosity of these bars was determined by mer-cury intrusion with a Micromeritics Pore Sizer 9310.

Disadvantages of MIP include: (1) the inability tointrude, and thus to determine, gel pores (<0.01 lm) aswell as closed pores, (2) the low accessibility of pores inthe bulk of the sample due to the presence of narrowinterconnections or choke points between pores (Dia-mond, 2000) and (3) the fact that it measures entry sizesrather than true pore sizes due to the ink bottle effect(Bentz and Garboczi, 1992; Kumar and Bhattacharjee,2003). With respect to the first disadvantage, closed poresand gel pores are not relevant to understand leachingproperties. The ink bottle and choke point effects, how-ever, lead to an overestimation of the fine pore volumeand an underestimation of the wide pore volume (Roels,2000; Wild, 2001). In addition to the limitations of theMIP technique itself, the technique of sample drying priorto mercury intrusion is also under debate (Diamond,2001; Galle, 2001; Wild, 2001; Diamond, 2003; Galle,2003). Incomplete drying reduces total intruded volumeand overestimates the amount of smaller pores (Diamond,2003). Therefore, severe drying (e.g., in an oven at 105 �C)would be preferred over a more gentle method (e.g., dry-ing to constant weight at 40 �C over silica gel in a closedenvironment, as suggested by Wild (2001)). On the otherhand, severe drying could lead to larger shrinkage andmicrocracking (Hewlett, 1998; Wild, 2001). The debateon the appropriateness of MIP and the drying treatmentprior to it is still on-going, but protagonists in the discus-sion agree that MIP is still an acceptable and valuabletool to comparatively study the pore structure evolutionof cement-based materials in various environments, aslong as the limitations are taken into account and as longas the same drying technique is consistently applied acrossthe series compared (Wild, 2001; Diamond, 2001; Galle,2003).

For the sake of simplicity, the MIP results are labeled as‘total porosity’ in this paper, although the technique actu-ally gives the volume of readily accessible capillary poresand macropores. Results are presented as the percentageof pore volume to the total specimen volume.

Table 1Total porosity of unleached B- and F-samples and decrease of porosityrelative to the uncarbonated sample

Duration ofcarbonation(days)

B-samples F-samples

Porosity (%) Decrease (%) Porosity (%) Decrease (%)

0 29.0 – 40.4 –14 28.1 3.0 – –30 27.2 6.2 40.6 �0.560 25.2 12.0 40.2 0.5

980 T. Van Gerven et al. / Waste Management 27 (2007) 977–985

3. Results and discussion

3.1. Degree of carbonation

Inner surfaces of specimens that were sprayed with phe-nolphthalein are shown in Fig. 1. The results show thatB600 is nearly completely carbonated, whereas B14 andB30 are partially carbonated. F300 and F600 are both onlyslightly carbonated: the carbonated shell is only 1 mm thickin F600 and even less in F300. B-samples are more carbon-ated than F-samples because the latter contain mainlysmall pores, as will be shown in the following paragraph.These small pores are likely to be filled with water becauseof the high RH (>90%) during the carbonation procedure,whereas the larger pores in the B-samples will be partly air-filled. As a consequence, CO2 diffusion is much slowerthrough the pores of the F-samples than through the poresof the B-samples.

Sanchez (1996) described a technique to derive theamount of free portlandite in cementitious samples fromthe results of an equilibrium-based leaching test on parti-cle-size reduced samples, assuming that all acid requiredto reach pH 11.9 is used to neutralize free portlandite.This method was applied here. The results of the equilib-rium-based leaching test are discussed in a related paper(Van Gerven et al., 2006). For the B-samples, thisapproach provides estimates of portlandite content of61, 11 and 1 mg/g of dry sample for samples B00, B140and B300, respectively. No free portlandite is present inthe carbonated B600 sample. This confirms that B600 iscompletely carbonated. In the F-samples, portlandite con-tents of 85, 72 and 71 mg/g were estimated for F00, F300and F600 samples, respectively. These results representthe average portlandite content for the whole sample. Itcan be assumed that in the partially carbonated monolithsmore portlandite than the average is present in the coreand less portlandite than the average is present in the car-bonated shell.

Fig. 1. Photographs of the inner surface of B- and F-samples that were sprayed

Assuming that by carbonation free portlandite is trans-formed into calcite, the calcite content can be deduced fromthe difference between the free portlandite in uncarbonatedand carbonated material. This estimate yields 68 mg and 81calcite/g of dry matter for B140 and B300, respectively.Because it was estimated that B600 contained no free port-landite, at least 82 mg calcite/g of dry sample is produceddirectly from portlandite. However, more calcite can beproduced from decalcification of calcium silicate hydrategels. In the F-samples, calculations result in a calcite con-tent of 18 mg/g for F300 and 19 mg/g for F600. This rela-tively high content of calcite is situated in the thincarbonated shell, resulting in a very dense zone with lowporosity, whereas carbonation in the B-samples induces amuch thicker carbonated zone that is less dense.

3.2. Influence of carbonation on porosity

Although the w/c is slightly higher, B-samples have alower overall porosity than F-samples (Table 1), due tothe fact that a considerable proportion of the monolith vol-ume consists of an impermeable aggregate. In the B-sam-ples, carbonation decreases total porosity, in agreementwith literature (Mollah et al., 1993; Gervais et al., 2004;Van Gerven et al., 2004). Carbonation of the F-sampleswas negligible and these samples thus exhibit no clearchange in porosity.

with phenolphthalein. Dark spots correspond to the uncarbonated phases.

T. Van Gerven et al. / Waste Management 27 (2007) 977–985 981

Fig. 2 shows the porosity as a function of calcite contentin B-samples. As calcite deposition increases, porositydecreases and the decrease becomes more important. Thisis due to the fact that with the used porosimetry technique,only the open porosity is determined. During carbonationcalcite is deposited on the inside of pores but also at theentrance, thus clogging some of these pores and makingthem unavailable for mercury intrusion.

Apart from the total porosity, mercury intrusion exper-iments also give results on the pore size distribution(Fig. 3). The pore structure in the B-samples is differentthan that in the F-samples. B-samples contain macropores(>10 lm), due to the interfacial transition zone betweenaggregate and hydrated cement particles, and capillarypores (0.01–10 lm). F-samples do not contain macroporesand the relative importance of small capillary pores(<0.1 lm) is higher. Gel pores are assumed to be presentin B- and F-samples but cannot be determined by MIP.These gel pores are, however, reported to be unimportantwith respect to leaching (Bentz and Garboczi, 1992).

Fig. 3 shows the influence of carbonation on pore sizedistribution. Both B- and F-samples show the same effects,

25.0

25.5

26.0

26.5

27.0

27.5

28.0

28.5

29.0

29.5

0 20 40 60 80 100

calcite content (mg/g DM)

po

rosi

ty (

%)

Fig. 2. Relationship between the calcite content and the porosity ofunleached B-samples.

0.0

0.5

1.0

1.5

2.0

2.5

3.0

3.5

4.0

4.5

0.001 0.010 0.100 1.000 10.000 100.000 1000.000

pore size (µm)

shar

e o

f to

tal

sam

ple

vol

ume

(%) B00

B600

Fig. 3. Pore size distribution of u

but to a lesser extent in the F-samples due to the limitedcarbonation. The relative importance of smaller capillarypores (<0.1 lm) diminishes while the proportion of largercapillary pores (0.1–10 lm) remains stable (in the case ofB-samples) or increases (in the case of F-samples). Thepore structure of B-samples includes also macropores(>10 lm), the amount of which increases with progressivecarbonation. Macropores are absent in the F-samples.These findings are in agreement with Ngala and Page(1997) and Anstice et al. (2005) where an increasing pro-portion of larger capillary pores was noted.

3.3. Influence of leaching on porosity

Total porosity increases after 225 days of leaching bothfor B- and F-samples by 16% on average (Table 2).

The effect of leaching on pore size distribution is consis-tent for both B- and F-samples: the relative importance ofsmall capillary pores (<0.1 lm in the case of B-samples,<0.04 lm in the case of F-samples) diminishes, while theproportion of larger capillary pores (0.1–10 lm in the caseof B-samples, 0.04–0.2 lm for the F-samples) increases. Inthe B-samples the relative importance of macroporesdecreases as well. These findings agree well with reportedresults (Bishop et al., 1992; Garboczi and Bentz, 1996;Kumar and Bhattacharjee, 2003).

shar

e o

f to

tal

sam

ple

vol

ume

(%)

0.0

0.5

1.0

1.5

2.0

2.5

3.0

3.5

4.0

4.5

0.001 0.010 0.100 1.000 10.000 100.000 1000.000

pore size (µm)

F00

F600

nleached B- and F-samples.

Table 2Total porosity of B- and F-samples after 225 days of leaching and increaseof porosity relative to that of the corresponding unleached sample, whichwas presented in Table 1

Duration ofcarbonation(days)

B-samples F-samples

Porosity (%) Increase (%) Porosity (%) Increase (%)

0 34.0 17.2 47.5 17.614 33.0 17.4 – –30 30.2 11.1 47.3 16.560 29.8 18.3 46.5 15.7

Table 3Cumulative release data (mg/m2) of sodium and potassium after 225 daysof leaching, measured in the diffusion test as reported elsewhere (VanGerven et al., 2006), and calculated following the approach presented inthe text

Measured Calculated

Na K Na K

B0225 24557 30929 21313 31489B14225 19522 15886 13370 11990B30225 20084 12927 10906 6858B60225 18816 9343 10481 3961

F0225 37075 36884 36803 38037F30225 37350 38021 37220 38741F60225 37038 37179 36654 37667

982 T. Van Gerven et al. / Waste Management 27 (2007) 977–985

3.4. The porosity and pH aspects of carbonation and their

relative effect on leaching

Carbonation has been shown to decrease total porosity,which in turn affects leaching properties. Carbonation,however, not only affects leaching because of its impacton porosity but also as it decreases pH in the pore waterand converts hydroxides to carbonates. To assess theimportance of the porosity and pH separately, the presentresults may be combined with data on leaching of sodiumand potassium, as reported in Van Gerven et al. (2006).It is assumed that the transformation of hydroxide intothe corresponding carbonate does not affect leaching inthe case of sodium and potassium, as both the hydroxidesand the carbonates are highly soluble. Results in Van Ger-ven et al. (2006) shows that this assumption is valid forsodium. The solubility of potassium, however, decreasessignificantly with carbonation, questioning the assumption.The following results based on potassium are, therefore,less reliable than those based on sodium (see Fig. 4).

The effect of porosity decrease on leaching is separatedfrom the effect of pH decrease as follows. Sodium andpotassium solubility increases with decreasing pH. A curveis fitted on each of the solubility sequences so that the sol-ubility can be deduced at any given pH. The highest pHobtained in the monolith leaching test is taken as a refer-ence pH: 12.2 for all B-samples and 12.5 for all F-samples.The solubility of sodium and potassium at the pH of eachleachate of the monolith leaching test is divided by the sol-ubility at the reference pH, yielding a conversion factor.Subsequently the conversion factors are weighed toaccount for the different time periods of the renewals andan average factor for the whole leaching period is calcu-lated. Finally, the measured cumulative concentration isdivided by the corresponding average conversion factorto give a calculated cumulative release at the referencepH (Table 3), which is plotted as a function of the porosityin Fig. 5a. This calculated value hence refers to the cumu-lative leaching from a monolith where pH is maintained atthe maximum value. Differences between the measured and

0.0

0.5

1.0

1.5

2.0

2.5

3.0

3.5

4.0

4.5

0.001 0.010 0.100 1.000 10.000 100.000 1000.000pore size (µm)

shar

e of

tot

al s

ampl

e vo

lum

e (%

) B00

B0225

Fig. 4. Pore size distribution of un

the calculated values in Table 3, therefore, refer to the effectof pH changes related to different degrees of carbonation inthe leachates during the test. The curves show that porositydecrease, due to carbonation, causes leaching of sodiumand potassium to decrease as well. The results for the F-samples are less clear due to limited carbonation.

From the comparison of the above calculated cumula-tive releases and the actually measured cumulative releasedata, which are also shown in Table 3, increase or decreaseof leaching due to carbonation, relative to leaching fromB0225, can be attributed to either the effect of porosity,or the effect of pH. The different contributions for the car-bonated B-samples are plotted in Fig. 5b. The ‘overalleffect’ refers to the change in leaching due to both theporosity decrease and the pH decrease. The ‘porosity effect’refers to the change in leaching due to the porositydecrease, whereas the ‘pH effect’ refers to the change inleaching due to the pH decrease. It can be seen that theeffect of porosity decrease exceeds the opposite effect ofpH decrease, resulting in a net decrease of leaching dueto carbonation.

The porosity component appears to decrease leaching ofpotassium more than that of sodium. Although this may becaused by the different ionic diameter, which is 0.099 nmfor Na+ and 0.137 nm for K+ (Lide, 1999), the main part

0.001 0.010 0.100 1.000 10.000 100.000 1000.000

pore size (µm)

shar

e of

tot

al s

ampl

e vo

lum

e (%

) F00

F0225

0.0

0.5

1.0

1.5

2.0

2.5

3.0

3.5

4.0

4.5

carbonated B- and F-samples.

-40

-20

0

20

40

60

80

100

B14225 B30225 B60225 B14225 B30225 B60225Na K

chan

ge

in l

each

ing

(%

)in

crea

se

dec

reas

e porosity effect pH effect overall effect

0

5000

10000

15000

20000

25000

30000

35000

40000

45000

20 25 30 35 40 45

total porosity (%)

calc

ula

ted

cu

mu

lati

ve r

elea

se

(mg

/m²) Na, F-samples

K, B-samples

Na, B-samples

K, F-samples

a bFig. 5. Cumulative release, normalized for pH effects, of sodium and potassium in relation to total porosity (a), and contributions of porosity and pH tothe change in leaching relative to leaching from B0225 (b).

T. Van Gerven et al. / Waste Management 27 (2007) 977–985 983

of this decrease can be attributed to an underestimation ofpotassium leaching from carbonated samples. This under-estimation is caused by the important decrease of potas-sium solubility at alkaline pH in carbonated samplescompared to uncarbonated samples, as reported in VanGerven et al. (2006). Sodium leaching is also underesti-mated, but to a lesser extent than potassium.

The smaller increase of sodium and potassium leachingdue to pH in B60225 compared to B30225 may also becaused by undeservedly neglecting carbonate-formationand subsequently the conversion of carbonate to bicarbon-ate. This conversion, which occurs mainly in B60225 ratherthan B30225 and B14225 because of the low pH in the dif-fusion test, increases the release of sodium and potassiumin the monolith leaching test, analogous to the release ofcalcium and other metals (Van Gerven et al., 2006). Whencomparing the higher release in the monolith leaching testwith the reference pH of 12.2 in the monolith leaching testof B0225, where the alkaline nature of the pore water stillfavours the presence of carbonate over bicarbonate, thecalculated decrease in leaching due to the porosity effectincreases, and the corresponding increase in leaching dueto the pH effect would increase as well.

The above-applied approach to distinguish between theporosity component and the pH component within the car-bonation influence on leaching, gives satisfying results forsodium, but less for potassium. Even for sodium, the contri-butions of the different effects should not be judged quanti-tatively, but rather qualitatively. It cannot be used for othermeasured constituents such as calcium, cobalt or leadbecause the influences of carbonate formation and subse-quently carbonate/bicarbonate speciation become majorparameters in the monolith leaching test results of theseconstituents and interfere too much with the other twoeffects (porosity and pH). Due to the fact that, for these con-stituents, a change in pH can alter leaching by three or fourorders of magnitude (Van Gerven et al., 2006), it may beassumed that the effect of pH and of carbonate formationwill be much more important to leaching than the porosity

effect of carbonation. For the assessment of heavy metalleaching, the porosity effect of carbonation is thus consid-ered to be of smaller importance.

3.5. Comparison of the effects of carbonation and leaching on

porosity

Carbonation and leaching often occur intermittently innature with carbonation during non-leaching periods alter-nating with leaching periods. Carbonation and leachingexhibit opposite effects on total porosity. While continuousleaching during 225 days increased porosity by 16% onaverage, accelerated carbonation in a 20% carbon dioxideatmosphere during 60 days decreased total porosity by12% in the B-samples and by 2% at most in the F-samples.Combining these two effects, there was a net porosityincrease by about 4% in most B-samples and by more than14% in the F-samples.

Accelerated carbonation differs from natural carbon-ation in two ways. First, the degree of carbonation is muchhigher. The carbonation procedure in this investigation isreported to intensify natural carbonation by a factor ofmore than 40 (CUR, 1975). Second, due to the high CO2

pressure, newly formed carbonate near the outer surfaceof the matrix will react to soluble bicarbonate, which willdiffuse deeper in the monolith and convert back to carbon-ate when the CO2 pressure is low enough (CUR, 1975;Hewlett, 1998). This phenomenon will induce carbonatedislands in the uncarbonated zone. Natural carbonation,on the contrary, produces a more homogeneous carbon-ated shell around the uncarbonated core (Houst and Witt-mann, 2002). Overall, natural carbonation will decreaseporosity much less in the affected zone, and the carbonatedshell itself will be less thick than in the case of acceleratedcarbonation.

Whether intermittent leaching will increase porositymore than natural carbonation will decrease it, dependson the cumulative time length of the leaching periods rela-tive to that of the non-leaching periods and on the effect the

0.0

0.5

1.0

1.5

2.0

2.5

3.0

3.5

4.0

4.5

0.001 0.010 0.100 1.000 10.000 100.000 1000.000

pore size (µm)

shar

e of

tota

l sam

ple

volu

me

(%) B00

B60225

0.0

0.5

1.0

1.5

2.0

2.5

3.0

3.5

4.0

4.5

0.001 0.010 0.100 1.000 10.000 100. 000 1000. 000

pore size (µm)

shar

e of

tota

l sam

ple

volu

me

(%)

F00

F60225

Fig. 6. Pore size distribution of unleached uncarbonated samples and maximum leached and carbonated samples.

984 T. Van Gerven et al. / Waste Management 27 (2007) 977–985

non-leaching period may have on leaching of constituentsduring the subsequent leaching period.

With regards to the latter, it has been reported (Garra-brants et al., 2002) that relaxation, i.e., redistribution inthe pore water, of matrix components during the non-leaching period has only a marginal effect on their leachingduring the subsequent leaching period, compared to theeffect of the length of the non-leaching period. The porosityincrease, being dependent on the amount of matrix ions(Ca2+, Na+, K+, OH� and Cl�) leached, will thus not beinfluenced by relaxation.

During the non-leaching period, not only relaxationoccurs, but also carbonation. This intermittent carbonationcould affect leaching during the subsequent leaching per-iod, and consequently the porosity increase due to thisleaching. In a related paper, sodium, potassium, calciumand hydroxide ions were reported to leach less after car-bonation (Van Gerven et al., 2006). Considering the factthat these results were obtained with accelerated carbon-ation and that slight carbonation has a negligible effecton leaching, as shown in the case of F-samples, the effectof intermittent natural carbonation on the porosityincrease by leaching seems minimal.

Overall, we speculate that the cumulative time length ofthe leaching period in an intermittent wetting sequenceappears to be the decisive factor as to whether leaching willincrease porosity more than natural carbonation willdecrease it.

Both carbonation and leaching decrease the amount ofsmaller capillary pores. This is illustrated in Fig. 6.

4. Conclusion

Carbonation decreases total porosity of mortar withMSWI-bottom ash as sand fraction, while leaching hasthe opposite effect. Specimens that were carbonated for 2months in an atmosphere containing 20% CO2 and subse-quently leached for 225 days, showed a net increase ofporosity, compared to samples neither carbonated nor lea-ched. In nature, monoliths will undergo intermittent leach-ing with natural carbonation during the non-leaching

period. Whether intermittent leaching will increase poros-ity more than natural carbonation decreases it, dependson the cumulative time length of the leaching periods inthe intermittent wetting sequence. With respect to the effecton pore size distribution, both leaching and carbonationdecrease the amount of smaller capillary pores.

Carbonation affects leaching at three levels: decrease ofporosity, decrease of pH and transformation of hydroxidesto carbonates. By combining the porosity results with datafrom a related paper, it was possible to qualitatively distin-guish between the porosity effect of carbonation and thepH effect of carbonation on leaching of sodium and, to alesser extent, potassium. Porosity decrease diminishesleaching of these constituents, while decrease in pHincreases leaching. The effect of porosity, however, exceedsthat of the pH, resulting in a net decrease of leaching due tocarbonation. The porosity aspect of carbonation is, how-ever, assumed to be of smaller importance for the assess-ment of heavy metal leaching, because pH can affectleaching of these constituents by a much larger order ofmagnitude.

Acknowledgements

The Laboratory of Building Materials and BuildingTechnology and the Laboratory of Building Physics ofthe K.U. Leuven are gratefully acknowledged for, respec-tively, the use of the appropriate tools for the productionof cementitious monoliths and the use of the mercuryporosimeter.

References

Anstice, D.J., Page, C.L., Page, M.M., 2005. The pore solution ofcarbonated cement pastes. Cement and Concrete Research 35, 377–383.

Bentz, D.P., Garboczi, E.J., 1992. Modeling the leaching of calciumhydroxide from cement paste: effects on pore space percolation anddiffusivity. Materials and Structures 25, 523–533.

Bin Shafique, M.S., Walton, J.C., Gutierrez, N., Smith, R.W., Tarquin,A.J., 1998. Influence of carbonation on leaching of cementitiouswasteforms. Journal of Environmental Engineering 124 (5), 463–467.

T. Van Gerven et al. / Waste Management 27 (2007) 977–985 985

Bishop, P.L., Gong, R., Keener, T.C., 1992. Effects of leaching on poresize distribution of solidified/stabilized wastes. Journal of HazardousWastes 31, 59–74.

Conner, J.R., 1990. Chemical Fixation and Solidification of HazardousWastes. Van Nostrand Reinhold, New York, USA.

CUR, 1975. Carbonatatie lichtbeton: literatuurstudie (Carbonation oflightweight concrete: a literature study). CUR Rapport 73. Betonve-reniging, Zoetermeer, The Netherlands.

Diamond, S., 2000. Mercury porosimetry – an inappropriate method forthe measurement of pore size distributions in cement-based materials.Cement and Concrete Research 30, 1517–1525.

Diamond, S., 2001. Reply to the discussion by S. Wild of the paper‘‘Mercury porosimetry – an inappropriate method for the measure-ment of pore size distributions in cement-based materials’’. Cementand Concrete Research 31, 1655–1656.

Diamond, S., 2003. A discussion of the paper ‘‘Effect of drying on cement-based materials pore structure as identified by mercury porosimetry – acomparative study between oven-, vacuum-, and freeze-drying’’ by C.Galle. Cement and Concrete Research 33, 169–170.

Galle, C., 2001. Effect of drying on cement-based materials pore structureas identified by mercury porosimetry – a comparative study betweenoven-, vacuum-, and freeze-drying. Cement and Concrete Research 31,1467–1477.

Galle, C., 2003. Reply to the discussion by S. Diamond of the paper‘‘Effect of drying on cement-based materials pore structure as identifiedby mercury porosimetry – a comparative study between oven-,vacuum-, and freeze-drying’’. Cement and Concrete Research 33,171–172.

Garboczi, E.J., Bentz, D.P., 1996. Multi-scale picture of concrete and itstransport properties: introduction for non-cement researchers.National Institute of Standards and Technology Internal Report 5900.

Garrabrants, A.C., Sanchez, F., Gervais, C., Moszkowicz, P., Kosson,D.S., 2002. The effect of storage in an inert atmosphere on the releaseof inorganic constituents during intermittent wetting of a cement-based material. Journal of Hazardous Materials 91, 159–185.

Garrabrants, A.C., Sanchez, F., Kosson, D.S., 2004. Changes in consit-uent leaching and pore water characteristics of a Portland cementmortar as a result of carbonation. Waste Management 24, 19–36.

Gervais, C., Garrabrants, A.C., Sanchez, F., Barna, R., Moszkowicz, P.,Kosson, D.S., 2004. The effects of carbonation and drying duringintermittent leaching on the release of inorganic constituents from acement-based matrix. Cement and Concrete Research 34, 119–131.

Geysen, D., Imbrechts, K., Vandecasteele, C., Jaspers, M., Wauters, G.,2004. Immobilization of lead and zinc in scrubber residues from MSWcombustion using soluble phosphates. Waste Management 24, 471–481.

Hewlett, P.C. (Ed.), 1998. Lea’s Chemistry of Cement and Concrete.Arnold, London, Great Britain.

Houst, Y.F., Wittmann, F.H., 1994. Influence of porosity and watercontent on the diffusivity of CO2 and O2 through hydrated cementpaste. Cement and Concrete Research 24 (6), 1165–1176.

Houst, Y.F., Wittmann, F.H., 2002. Depth profiles of carbonates formedduring natural carbonation. Cement and Concrete Research 32, 1923–1930.

IAWG, 1997. Municipal solid waste incinerator residuesStudies inEnvironmental Science, 67. Elsevier, Amsterdam, The Netherlands.

Kosson, D.S., van der Sloot, H.A., Sanchez, F., Garrabrants, A.C., 2002.An integrated framework for evaluating leaching in waste managementand utilization of secondary materials. Environmental EngineeringScience 19 (3), 159–204.

Kumar, R., Bhattacharjee, B., 2003. Porosity, pore size distribution andin situ strength of concrete. Cement and Concrete Research 33, 155–164.

Lide, D.R. (Ed.), 1999. CRC Handbook of Chemistry and Physics,seventy-ninth ed. The Chemical Rubber Co., Cleveland (OH), USA.

Mollah, M.Y.A., Hess, T.R., Tsai, Y.N., Cocke, D.L., 1993. An FTIRand XPS investigation of the effects of carbonation on the solidifica-tion/stabilization of cement based systems – Portland type V with zinc.Cement and Concrete Research 23, 773–784.

Ngala, V.T., Page, C.L., 1997. Effects of carbonation on pore structureand diffusional properties of hydrated cement pastes. Cement andConcrete Research 27, 995–1007.

NNI (Nederlands Normalisatie Instituut), 1995. NEN 7345, Uitloogka-rakteristieken van vaste grond- en steenachtige bouwmaterialen enafvalstoffen – Uitloogproeven – Bepaling van de uitloging vananorganische componenten uit vormgegeven en monolithische mate-rialen met de diffusieproef. (Leaching properties of solid soil- andstone-like construction and waste materials – Leaching tests –Determination of leaching of inorganic components from monolithicmaterials using the diffusion test.) Delft, The Netherlands.

Roels, S., 2000. Modelling unsaturated moisture transport in heteroge-neous limestone. Doctoral thesis, UDC 532.51:552.541. Leuven,Belgium.

Sabbas, T., Polletini, A., Pomi, R., Astrup, T., Hjelmar, O., Mostbauer,P., Cappai, G., Magel, G., Salhofer, S., Speiser, C., Heuss-Assbichler,S., Klein, R., Lechner, P., 2003. Management of municipal solid wasteincineration residues. Waste Management 23, 61–88.

Sanchez, F., 1996. Etude de la lixiviation de milieux poreux contenant desespeces solubles: application au cas des dechets solidifies par liantshydrauliques (Leaching study from porous materials containingsoluble components: wastes solidified with hydraulic binders). Doc-toral thesis, 96 ISAL 0118. Lyon, France.

Van Gerven, T., Van Baelen, D., Dutre, V., Vandecasteele, C., 2004.Influence of carbonation and carbonation methods on leaching frommortars. Cement and Concrete Research 34, 149–156.

Van Gerven, T., Van Keer, E., Arickx, S., Jaspers, M., Wauters, G.,Vandecasteele, C., 2005. Carbonation of MSWI-bottom ash todecrease heavy metal leaching, in view of recycling. Waste Manage-ment 25, 291–300.

Van Gerven, T., Cornelis, G., Vandoren, E., Garrabrants, A.C., Sanchez,F., Kosson, D.S., Vandecasteele, C., 2006. Effects of progressivecarbonation on heavy metal leaching from cement-bound waste.AIChE Journal 52 (2), 826–837.

Wild, S., 2001. A discussion of the paper ‘‘Mercury porosimetry – aninappropriate method for the measurement of pore size distributions incement-based materials’’ by S. Diamond. Cement and ConcreteResearch 31, 1653–1654.