Journal of Asian Earth Sciences xxx (2013) xxx–xxx

Contents lists available at SciVerse ScienceDirect

Journal of Asian Earth Sciences

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

Dissolution kinetics and biodurability of tremolite particles in mimickedlung fluids. Effect of citrate and oxalate

1367-9120/$ - see front matter � 2013 Elsevier Ltd. All rights reserved.http://dx.doi.org/10.1016/j.jseaes.2013.04.008

⇑ Corresponding author. Tel.: +34 958230000x190120; fax: +34 958552620.E-mail addresses: marisarozalen@ugr.es, mlrozalen@gmail.com (M. Rozalen).

Please cite this article in press as: Rozalen, M., et al. Dissolution kinetics and biodurability of tremolite particles in mimicked lung fluids. Effect ofand oxalate. Journal of Asian Earth Sciences (2013), http://dx.doi.org/10.1016/j.jseaes.2013.04.008

Marisa Rozalen a,⇑, M. Elena Ramos a, F. Javier Huertas a, Saverio Fiore b, Fernando Gervilla a,c

a Instituto Andaluz de Ciencias de la Tierra – IACT, CSIC-University of Granada, Avda. de las Palmeras 4, 18100 Armilla, Granada, Spainb Istituto di Metodologie per l’Analisi Ambientale – IMAA (CNR), Tito Scalo, PZ, Italyc Departamento de Mineralogía y Petrología, University of Granada, Avda. Fuentenueva 2, 18002 Granada, Spain

a r t i c l e i n f o a b s t r a c t

Article history:Available online xxxx

Keywords:TremoliteDissolution rateOxalateCitrateBiodurability

The effect of citrate and oxalate on tremolite dissolution rate was measured at 37 �C in non-stirred flow-through reactors, using modified Gamble’s solutions at pH 4 (macrophages), 7.4 (interstitial fluids) and5.5 (intermediate check point) containing 0, 0.15, 1.5 and 15 mmol L�1 of citrate or oxalate. The dissolu-tion rates calculated from Si concentration in the output solutions without organic ligands depend on pH,decreasing when the pH increases from �13.00 (pH 4) to �13.35 (pH 7.4) mol g�1 s�1 and following aproton-promoted mechanism. The presence of both ligands enhances dissolution rates at every pH,increasing this effect when the ligand concentration increases. Citrate produces a stronger effect as a cat-alyst than oxalate, mainly at more acidic pHs and enhances dissolution rates until 20 times for solutionswith 15 mmol L�1 citrate. However, at pH 7.4 the effect is lighter and oxalate solutions (15 mmol L�1)only enhances dissolution rates eight times respect to free organic ligand solutions. Dissolution is pro-moted by the attack to protons and organic ligands to the tremolite surface. Magnesium speciation inoxalate and citrate solutions shows that Mg citrate complexes are more effective than oxalate ones dur-ing the alteration of tremolite in magrophages, but this tendency is the opposite for interstitial fluids,being oxalate magnesium complexes stronger. The biodurability estimations show that the destructionof the fibers is faster in acidic conditions (macrophages) than in the neutral solutions (interstitial fluid).At pH 4, both ligands oxalate and citrate reduce the residence time of the fibers with respect to that cal-culated in absence of ligands. Nevertheless, at pH 7.4 the presence of ligands does not reduce significantlythe lifetime of the fibers.

� 2013 Elsevier Ltd. All rights reserved.

1. Introduction

The world has 200 million tons of identified resources of asbes-tos. Although the commercial use of these minerals is banned inmany countries, nowadays there are still many active mines, beingthe biggest producers some Eurasian countries like Russia(1000 million tons), and Asian countries like China (400 milliontons) and Kazakhstan (210 million tons) (USGS Mineral Commod-ity Summaries, January 2012). The leading consuming countriesin 2007 were, in decreasing order tonnage, China (30%), India(15%), Russia (13%), Kazakhstan and Brazil (5% each), and Thailand,Uzbekistan and Ukraine (4% each). These eight countries accountedfor about 80% of world asbestos consumption in 2007 (USGS Min-eral Industry Survey, 2003–2007).

Although the vast majority of Asian mines are ultramafic chrys-otile deposits, the presence of tremolite is common, as a naturalimpurity. Despite low concentrations, inhaled amphibole fibers

tend to accumulate in the lung tissue and may contribute signifi-cantly to the mesothelioma risk of chrysotile miners and millers(Case, 1991; McDonald and McDonald, 1997). National Instituteof Occupational Safety and Health (NIOSH) has developed a prior-ity list of 10 leading work-related illnesses and injuries whereoccupational lung disease is first on the list (silicosis, asbestosisand byssinosis).

Toxicological studies (in vivo) show that interactions betweenfibrous material and biological environment are strongly depen-dent on both geometry and crystal chemistry of mineral fibers(Donaldson and Tran, 2004; Maxim et al., 2006; Bulsari et al.,2007; Stettler et al., 2008; Donaldson, 2009; Osmond-McLeodet al., 2011). Another factor that control the tendency of a fiberto cause a disease is the residence time of the particles in the lungs.The residence time is related to the particle biodurability, which isintrinsic for each mineral. The biodurability is the resistance of aparticle to chemical dissolution in the body. Therefore to describethe dissolution in the lungs the solvent solution must be similar incomposition to those found in the lungs. Thus the use of geochem-

citrate

Fig. 1. SEM image of the micromorphology of the tremolite used as startingmaterial.

2 M. Rozalen et al. / Journal of Asian Earth Sciences xxx (2013) xxx–xxx

ical methods, as the traditional dissolution experiments, can bealso useful to understand the biological breakdown of the fibers.

A number of studies have reported the dissolution rates of var-ious silicate minerals as siliceous fibers (Scholze and Conradt,1987), chrysotile (Hume and Rimstidt, 1992; Gunter and Wood,2000; Oze and Solt, 2010), crocidolite (Werner et al., 1995), talc(Jurinski and Rimstidt, 2001) and tremolite (Mast and Drever,1998; Oze and Solt, 2010) in simulated lung fluids. Moreover, oth-ers studies have confirmed the catalytic effect of organic ligandspresents in lung fluids, as oxalate (Mast and Drever, 1998) or cit-rate (Ramos et al., 2011). These ligands are able to form strongcomplexes with cations as Al and Mg, promoting the release ofthese structural cations to solution and consequently enhancingdissolution rates. Therefore, the aim of this study is to evaluatetremolite dissolution rates in mimicked lung-fluids, including cit-rate and oxalate as a proxy for organic acids in alveolar fluidsand to estimate the biodurability of the tremolite particles.Although in vitro experiments do not reproduce the complexityof processes that can occur in human body, they provide a bench-mark to understand the degradation of these particles.

2. Materials and methods

2.1. Characterization of the mineral sample

All the experiments were carried out with tremolite collected inan old serpentine quarry next to K29 of A-395 road, in Sierra Neva-da (Granada, SE Spain). X-ray diffraction (XRD) analysis shows apure tremolite phase without accessory/companying minerals.We have not grinded the sample to avoid crushing the initial sizeof the fibers. The natural tremolite was pretreated to enrich thesample in fine particles (<4 lm equivalent spherical diameter) byrepeated sedimentation–suspension in water. The supernatantwas dried in an oven at 40 �C and the particles were recovered withacetone. Finally, it was stored in a polyethylene bottle as the start-ing material. Chemical analysis of major elements was performedby X-ray fluorescence (XRF). The calculated structural formula cor-responds to a tremolite:

Ca1:84Na0:17ðMg5:08Fe2þ0:17ÞðSi7:88Al0:12ÞO22ðOHÞ2:23

The corresponding atomic ratios Mg/Si, Ca/Si and Na/Si arerespectively 0.645, 0.233 and 0.0216. Scanning electron micros-copy (SEM) images show mainly prismatic and acicular structuresthat forms fiber bundles, with a wide variety of ratios length:diam-eter. The material is very friable and a very large number of extre-mely fine fibers are observed (Fig. 1). The specific surface area wasmeasured by BET using a 5-point N2 adsorption isotherm obtaininga value of 0.49 m2 g�1 with an associated uncertainty of at least10%.

2.2. Flow through dissolution experiments

Dissolution experiments were performed in single-pass, non-stirred, flow-through cells, which facilitated the measurement ofthe dissolution rate under fixed saturation state conditions bymodifying flow rate, initial sample mass and input solution con-centration. The reactors were fully immersed in a thermostaticwater-bath held at a constant temperature of 37 ± 1 �C. The flowrate was controlled with a peristaltic pump that injects the inputsolution into the bottom chamber of the cell (0.02 mL min�1).The tremolite sample is confined within the upper chamber (reac-tion zone) by using two membrane filters: a 5 lm nylon mesh plusa 1.2 lm Durapore membrane at the bottom and a 0.45 lm Dura-pore membrane at the top. The total volume of the cell was 46 mL

Please cite this article in press as: Rozalen, M., et al. Dissolution kinetics and band oxalate. Journal of Asian Earth Sciences (2013), http://dx.doi.org/10.1016/

and the solid mass added to each cell was approximately between0.1 and 0.5 g to yield a solid:solution ratio between 2 and 10 g L�1.

In each run, the flow rate and the input pH were held constantuntil steady-state conditions were achieved. The steady state wasassumed to prevail when the Si output concentration remainedfairly constant, differing by less than 6% between consecutive sam-ples (Rozalen et al., 2008). Reaction times were from 1000 to1800 h depending on the pH and citrate/oxalate concentration. Atsteady state, dissolution is expected to proceed under far-from-equilibrium conditions. All the experiments consisted of a singlestage; the cell was dismantled after the steady state was achieved.

After sampling every 24 h, the pH of the output solutions wereimmediately measured at room temperature by using a Crisoncombination electrode standardized with pH 4.01 and 7.00 buffersolutions. The reported accuracy was ±0.02 pH units. The differ-ence in the pH value between both room and experimental tem-perature was less than the accuracy of the measurement, thus notemperature correction was applied (Ramos et al., 2011). An ali-quot was separated for oxalate or citrate analysis. Then the outputsolutions were acidified to pH 3 with HNO3 to prevent the precip-itation of Mg-bearing phases during storage for further analyses.

The Si concentration in every output solution was determinedby colorimetry, using the molybdate blue method (Grasshoffet al., 1983) with a detection limit of 5 ppb and a 5% associated er-ror. The concentration of oxalate was measured by ion chromatog-raphy (IC) using a Metrohm 761 Compact Ion Chromatograph witha Metrosep A Supp 4-250 column with chemical suppression. Theeluent was prepared with 1.7 mmol L�1 NaHCO3 and 1.8 mmol L�1

Na2CO3. The detection limit is 0.007 mmol L�1 and the associatederror 5%. Citrate was also measured by IC with a Metrosep OrganicAcids column. The eluent was prepared with 0.5 mmol L�1 H2SO4/15% acetone. The detection limit is 0.03 mmol L�1 and the associ-ated error 3%.

2.3. Simulation of lung fluids

The composition of the input solutions mimics the fluids foundin the human lung, the so-called Gamble’s solution. In this studythe solutions were prepared by using the formulation of Jurinskiand Rimstidt (2001) with additional modifications (Table 1). Salinesolutions have the same molar composition, but Mg and Ca saltswere substituted by Na salts to avoid structural cations of tremo-lite. Phosphate salts were also avoided, because phosphate is animportant interference in the Si analysis.

When fibers are respired the particles can come into contactwith both macrophage cytoplasm and interstitials fluids (Collieret al., 1992). We have chosen three working pHs: 4 (correspondingwith macrophages), 5.5 (middle check point) and 7.4 (correspond-

iodurability of tremolite particles in mimicked lung fluids. Effect of citratej.jseaes.2013.04.008

Table 1Salt concentration in the experimental solution (mmol L�1).

Jurinski and Rimstidt (2001) This study

MgCl2�6H2O 1.043 NaCl 112.3NaCl 109.6 Na2SO4 0.556CaCl2�2H2O 1.734Na2SO4�10H2O 0.556

M. Rozalen et al. / Journal of Asian Earth Sciences xxx (2013) xxx–xxx 3

ing with interstitial fluids). Both, intracellular fluids of macro-phages and interstitial fluids have grossly similar composition(with possible variations arising depending upon the cell function).

Interstitial fluids are generally similar in composition to bloodplasma, but the concentrations in proteins are much lower. Manyorganic acids like lactic, citric, malic, oxalic and salicylic are pres-ent. We have chosen to study the effect of two different ones, cit-rate and oxalate even though lactate is the most abundant organicacid in the interstitial fluids (Plumlee and Ziegler, 2003). However,previous studies (see Ramos et al., 2011) have shown that lactatehas non/little effect on the dissolution rate. By the other side, cit-rate and oxalate have proved to have a strong catalyst effect on dis-solution rates of smectites (e.g. Cambier and Sposito, 1991; Ramoset al., 2011), kaolinites (Ganor and Lasaga, 1994) or chrysotile (Roz-alen et al., 2012). The concentration of each organic anion in theinterstitial fluid is not well documented but some authors pointout a content of 0.115 mmol L�1 of citrate and 0.011 mmol L�1 ofoxalate in blood plasma fluids (Plumlee and Ziegler, 2003). In orderto assess the ligand effect on the dissolution rate the input solu-tions were prepared by adding oxalate or citrate in three differentconcentrations (0.15, 1.5 and 15 mmol L�1 for each ligand) to thesaline solution. The pH was adjusted at 4, 5.5 and 7.4 by addingHNO3 or NaHCO3 solutions. Sodium azide (NaN3, 2 ppm) wasadded as bactericide. The compositions of the input solution inevery dissolution experiment are gathered in Table 2. Dissolutionexperiments in ligand-free saline solutions were carried out forcomparison purposes covering the same pH range.

2.4. Kinetic calculations

In a well-mixed, flow-through reactor the dissolution rate, Rate(mol g�1 s�1), can be calculated based on the mass balance of a gi-ven mineral component j. Under steady-state conditions this is gi-ven by the following equation (e.g. Cama et al., 2000; Rozalen et al.,2008):

Table 2Specific composition of the input solutions.

Run Initial pH NaOx (mmol L�1) NaCit (mmol L�1)

Tr_37_4 4.08 – –Tr_37_5 4.98 – –Tr_37_7 7.42 – –Tr_C0.15_4 4.20 – 0.15Tr_C1.5_4 3.93 – 1.5Tr_C15_4 4.32 – 15Tr_C0.15_5 5.17 – 0.15Tr_C1.5_5 5.21 – 1.5Tr_C15_5 5.28 – 15Tr_C0.15_7 7.14 – 0.15Tr_C1.5_7 7.42 – 1.5Tr_C15_7 7.38 – 15Tr_Ox0.15_4 4.17 0.15 –Tr_Ox1.5_4 4.23 1.5 –Tr_Ox15_4 3.83 15 –Tr_Ox0.15_5 5.16 0.15 –Tr_Ox1.5_5 5.19 1.5 –Tr_Ox15_5 5.50 15 –Tr_Ox0.15_7 7.61 0.15 –Tr_Ox1.5_7 7.27 1.5 –Tr_Ox15_7 7.67 15 –

Please cite this article in press as: Rozalen, M., et al. Dissolution kinetics and band oxalate. Journal of Asian Earth Sciences (2013), http://dx.doi.org/10.1016/

Rateðmol g�1 s�1Þ ¼ � 1v j

qMðCj;out � Cj;inÞ ð1Þ

where mj is the stoichiometric coefficient of component j in the dis-solution reaction, q stands for the volumetric fluid flow through thesystem, M the mass of smectite and Cj,out and Cj,in correspond to theconcentrations of component j in the output and input solutions,respectively. The rate is defined as negative for dissolution and po-sitive for precipitation. The dissolution rate in silicate studies isgenerally calculated from Si concentrations (RSi) in the output solu-tions as it is considered as the rate-limiting step in the dissolutionreaction (e.g. Liu et al., 2006; Olsen and Rimstidt, 2008; Oelkersand Schott, 2001). The use of other cations as magnesium or cal-cium can lead to error due to multiple reasons as: non-stoichiome-tric dissolution, precipitation of secondary phases (e.g. hydroxidesand carbonates) and the possibility of exchange reactions with pro-tons (Liu et al., 2006; Olsen and Rimstidt, 2008). All dissolutionrates were normalized to initial mass of tremolite. The error inthe calculated rate is estimated by using the Gaussian error propa-gation method and is less than 6% in all cases, which corresponds toapproximately 0.1 logarithmic units.

3. Results

The concentrations of Si, citrate/oxalate and pH were monitoredthroughout the duration of all the experiments. The variation withtime of the output solution composition for all the flow-throughexperiments is shown in Fig. 2. The experimental conditions ofall the series, the average pH and the concentration of Si at stea-dy-state are compiled in Table 3. Dissolution experiments followthe nomenclature pattern: Tr_37_4b, where Tr is tremolite in asimulated lung fluid (modified Gamble solution), 37 is tempera-ture, 4 corresponds with the initial pH. Finally, letter b correspondsto a replicate of the flow-through cell. In case of organic ligands theprefix C or Ox is used respectively for citrate and oxalate (e.g.Tr_C15_4) followed by the concentration of the ligand (0.15, 1.5or 15 mmol L�1).

In the experiments without organic acids and as it is typicallyobserved for flow-through experiments, high Si concentrationswere observed at the onset of most experiments, decreasing signif-icantly with elapsed time until a steady state is approached (Fig. 2).The solution pH remains constant with elapsed time in all theexperiments.

AcOH (mmol L�1) NaAcO (mmol L�1) KHCO3 (mmol L�1)

– – –0.64 1.86 –– – 0.034– – –– – –– – –0.64 1.86 –0.80 2.31 –– – –– – 0.044– – 0.044– – –– – –– – –– – –0.64 1.86 –0.64 1.86 –0.64 1.86 –– – 0.034– – 0.034– – –

iodurability of tremolite particles in mimicked lung fluids. Effect of citratej.jseaes.2013.04.008

(a) (d)

(b) (e)

(c) (f)

Fig. 2. Evolution of Si concentration in the output solutions of tremolite dissolution experiments conducted in flow-through cells at pH 4, 5 and 7.4, in presence of 0, 0.15, 1.5and 15 mM of citrate or oxalate. Output solutions used to calculate the average steady state are denoted by open symbols.

4 M. Rozalen et al. / Journal of Asian Earth Sciences xxx (2013) xxx–xxx

Parallel to experiments without ligands the tendency is the sim-ilar but at the onset of the experiments silica release increases untilreach a maximum and then starts to decrease asymptotical untilsteady state is approached. Silica released to the solution is higherwhen ligand concentration (oxalate or citrate) increases at everypH studied (4, 5 and 7.4). However, it is necessary to take in ac-count that the initial mass used in every experiment varies (see Ta-ble 3) due to experimental restrictions as detection limits.Consequently, the effect of ligands in the increase of Si releasecan be only observed in Fig. 2e (oxalate series at pH 5), since theinitial mass is the same.

Please cite this article in press as: Rozalen, M., et al. Dissolution kinetics and band oxalate. Journal of Asian Earth Sciences (2013), http://dx.doi.org/10.1016/

Citrate/oxalate concentration measured in the output solutionsfor all the experiments are constant and did not show any degrada-tion along time.

4. Discussion

4.1. Dissolution rates in ligand free solutions

The logarithm of dissolution rates calculated from Si concentra-tion in the output ligand free solutions depend on pH (Table 3),

iodurability of tremolite particles in mimicked lung fluids. Effect of citratej.jseaes.2013.04.008

Table 3Experimental conditions and results of flow-through dissolution experiments. Dissolution rates were normalized to mass.

Fig. 3. Experimental dissolution rates derived from Si in outlet solutions calculatedwith citrate (a) and oxalate (b). Dissolution rates in ligand-free solutions areincluded for comparison.

M. Rozalen et al. / Journal of Asian Earth Sciences xxx (2013) xxx–xxx 5

decreasing from �13.00 to �13.35 mol g�1 s�1 when the pH in-creases from pH 4 to 7.4. This dependence is not as strong as forother asbestos (e.g. chrysotile) but is in a good agreement with lit-erature results reported by Mast and Drever (1998) between pH 4and 9. In absence of organic ligands the species that attack themineral surface are protons and water molecules, which can formsurface complexes with cations at surface sites. An empirical ratelaw can be obtained considering dissolution rate proportional tothe proton activity powers to the proton reaction order:

Rateðmol=gsÞ ¼ 10�12:41 � a0:12Hþ ð2Þ

where the reaction order is nH = 0.12 and the rate constant iskH = 10�12.41. There are no data available about the dissolutionmechanism of tremolite in literature. However, the mechanism isknown for other magnesium-silicates as enstatite, antigorite andtalc. The mechanism, based on the concurrence of XPS adsorptionbands positions in fresh and acid leached enstatite, postulated thatH+ replace Mg2+ in the mineral structure during hydrolysis withoutrearranging the silica bonding network (Schott et al., 1981). Saldiet al. (2007) proposed that talc dissolution rates are controlled bythe detachment of partially liberated silica tetrahedral formed attalc edge surfaces from the exchange of Mg2+ for two protons, whatis consistent with a proton-promoted mechanism. When pH in-creases H–Mg exchange reaction is more extensive partially liberat-ing Si–O tetrahedra and dissolution rate increases. Theextrapolation of the results to the conditions found in the lungs, al-low us to affirm that dissolution is likely more effective in the mac-rophages (pH 4), since the calculated rate is more than doublecompared with interstitial fluid conditions (pH 7.4).

4.2. Catalytic effect of ligands

The results obtained in this study show that both ligands en-hance dissolution rates. This effect also depends on pH and concen-tration of the ligand, as it is shown in Fig. 3. For citrate and oxalatethe catalytic effect increase when concentration increases from0.15 to 15 mmol L�1. For citrate (Fig. 3a), in the presence of thehighest concentration at pH 4 (15 mmol L�1) the dissolution rateincreases up to 30 times respect to the rate calculated without or-ganic ligands at the same pH. However, at pH 7.4 the rate obtainedis only three times higher than that without citrate. For oxalate(Fig. 3b), in the presence of the highest concentration at pH 4

Please cite this article in press as: Rozalen, M., et al. Dissolution kinetics and band oxalate. Journal of Asian Earth Sciences (2013), http://dx.doi.org/10.1016/

(15 mmol L�1) the dissolution rate increases up to 20 times respectto the rate calculated without organic ligads at the same pH. Paral-lel to citrate, the effect at pH 7 is smaller, and rates increase justeight times respect to those without citrate.

iodurability of tremolite particles in mimicked lung fluids. Effect of citratej.jseaes.2013.04.008

6 M. Rozalen et al. / Journal of Asian Earth Sciences xxx (2013) xxx–xxx

Tremolite dissolution in acidic and neutral solutions containingligands may occur through contributions of proton and ligand-pro-moted reactions. At the same time the contribution of the ligand tothe overall rate may occur through the formation of surface com-plexes or a reduction in the ion activity product by complexationof the released cations to the solution. The first mechanism hasbeen observed by Liu et al. (2006) or Olsen and Rimstidt (2008)for forsterite in presence of oxalate. They proposed that protons

Table 4Stability constants for citric and oxalic acid and aqueous Mg-oxalate and Mg-citrate complexes (NIST, 2001).

Reaction Constant

OxalateOxalic acid = HOx� + H+ pK1 = 1.27HOx� = Ox2� + H+ pK2 = 4.26Mg2+ + Ox2� = MgOxaq logK1 = 2.10

CitrateH3Cit = H2Cit� + H+ pKa1 = 3.10H2Cit� = HCit2� + H+ pKa2 = 4.80HCit2� = Cit3� + H+ pKa3 = 6.40Mg2+ + Cit3� = MgCit� logK = 3.33Mg2+ + Cit3� + H+ = MgH(Cit) logK = 7.48Mg2+ + Cit3� + 2H+ = MgH2(Cit)+ logK = 11.01

(a)

(b)

Fig. 4. Relative abundances of citrate and oxalate species in aqueous solution as a functioof citrate and oxalate as a function of the solution pH (b and d).

Please cite this article in press as: Rozalen, M., et al. Dissolution kinetics and band oxalate. Journal of Asian Earth Sciences (2013), http://dx.doi.org/10.1016/

and ligands must be present in the activated complex, which im-plies that the release of silica into solution requires simultaneoushydrogen ion attachment to the underbonded Mg–O–Si atom andoxalate attachment to the Mg site.

The second mechanism is related to the trend of both ligands toform soluble complexes. The aqueous speciation of citrate and oxa-late was derived from their acidity constants (Table 4) using VisualMINTEQ (Gustafsson, 2010) and are shown in Fig. 4. The capacity ofcitrate and oxalate to form aqueous complexes with magnesiumwas also calculated using the corresponding constant values listedin Table 4.

The Mg speciation diagram was different for each ligand (Fig. 4band d). For citrate, in the pH range 2–4.5 three species coexistMgH2Cit+, MgHCit and MgCit�. The amount of MgH2Cit+ increasesuntil reach a maximum of 20% at pH 3 decreasing progressively un-til disappears at pH 5. The amount of MgHCit specie increases untilreach a maximum of 29% around pH 4, decreasing progressivelyuntil disappear at pH 6. Finally, the amount of MgCit� increasesprogressively until reach 100% from pH 5.5 and above. Taking intoaccount the existence of a rough correlation between the stabilityconstants for aqueous and surface complexes (Kummert andStumm, 1980; Stumm et al., 1980; Sigg and Stumm, 1981) the cat-alytic behavior of citrate and oxalate can be explained by the for-mation of aqueous Mg complexes. In the case of macrophagefluids (pH 4) the formation of 9% of MgH2Cit+, 30% of MgHCit and

(c)

(d)

n of pH (a and c). Magnesium (1 mmol L�1) speciation in solutions with 1 mmol L�1

iodurability of tremolite particles in mimicked lung fluids. Effect of citratej.jseaes.2013.04.008

Fig. 5. Estimated time required to dissolve tremolite fibers versus particlediameter. Citrate (a) and oxalate (b) concentration was 0.15 mmol L�1. The curvefor the dissolution time calculated without any ligand was included for comparison.

M. Rozalen et al. / Journal of Asian Earth Sciences xxx (2013) xxx–xxx 7

36% of MgCit� enhances dissolution rates up to 30 times respect toligand free solutions. In the case of oxalate, the effect is also verysignificant due to the formation of the complex MgOx. In the caseof interstitial fluids (pH 7.4) the effect of citrate is due to the for-mation of almost 100% of MgCit� and MgOx complexes for citrateand oxalate respectively.

Finally, the effect of the ligands in the overall rate of tremolitedissolution can be expressed and quantified as the contributionof proton- and ligand-promoted dissolution mechanism (e.g. Golu-bev and Pokrovsky, 2006; Golubev et al., 2006; Olsen and Rimstidt,2008):

RateTot ¼ RateH þ RateLigand ¼ kH � anHHþ þ kL � cnL

L ð3Þ

The first and second terms in this equation stand respectivelyfor the contribution of proton- and ligand-promoted dissolution.At each specific pH, the ligand-promoted dissolution can be esti-mated by subtracting the dissolution rate of ligand-free solutionsfrom the overall dissolution rate, RTot–RH and obtaining the corre-sponding empirical laws shown in Table 5. As it is observed in Ta-ble 5 ligand-promoted dissolution rates increase with increasingligand concentration. This effect is steeper at pH 4 than at 7.4 asreveals the increase in the reaction order, nL, with the solutionpH from 0.22 to 0.29 for citrate and from 0.17 to 0.48 for oxalate.This means that in interstitial fluids where the contribution of pro-ton to dissolution is very limited, the net effect of the ligands ishigher, when compared with macrophages environment.

Consequently, the results obtained in this study suggest that theligand-promoted dissolution mechanism needs to be considered inorder to quantify asbestos degradation in biological conditions.However, more studies are necessary to evaluate the species insolution, their interaction with the mineral surface and their rolein the dissolution process(es). In this sense electron microscopystudies of the solids obtained in these experiments can help usto monitor the changes in fiber morphology induced by the alter-ation reaction.

4.3. Biodurability estimations

The kinetic data obtained in this study can be used to estimatethe dissolution time of a fiber. Hume and Rimstidt (1992) proposedfor chrysotile a shrinking geometric model that considers a fiber asan infinitely long cylinder that dissolves over the lateral surface.The area in the ends of the fibers is negligible when it is comparedwith the lateral surface. Although the structure and morphology ofthe tremolite particles fit better to a prism, the Hume and Rims-tidt’s model can be used to get an estimation of the biodurabilityof tremolite fibers. The residence time or lifetime of a tremolite fi-ber can be estimated applying the following equation:

t ¼ 34

dVmR

ð4Þ

where t (s) is the residence time of the fiber, d (m) the diameter, Vm

the tremolite molar volume (2.73 � 10�4 m3 mol�1, Robie and Hem-ingway, 1995) and R is the dissolution rate normalized to specificsurface area (mol m2 s�1).

Table 5Empirical rate laws describing ligand-promoted dissolution rate RL (according to Eq.(3)) versus concentration of oxalate and citrate (CL) (0.15, 1.5 and 15 mM) at everyworking pH.

pH Citrate Oxalate

4 RateL ¼ 10�11:19 � C0:22L RateL ¼ 10�11:41 � C0:17

L

5.5 RateL ¼ 10�11:08 � C0:29L RateL ¼ 10�11:31 � C0:26

L

7.4 RateL ¼ 10�12:34 � C0:28L RateL ¼ 10�11:09 � C0:72

L

Please cite this article in press as: Rozalen, M., et al. Dissolution kinetics and band oxalate. Journal of Asian Earth Sciences (2013), http://dx.doi.org/10.1016/

Fig. 5 shows the calculated lifetime of tremolite fibers of severaldiameters at pH 4 (Fig. 5a) and pH 7.4 (Fig. 5b). To estimate thebiodurability, the rates used were those calculated in the mostsimilar conditions to biological fluids (Plumlee and Ziegler, 2003),with a concentration of 0.15 mmol L�1 for both ligands. The ratesin electrolyte solutions were also included for comparison pur-poses. The dissolution rates data used in the model are shadowedin Table 3.

Fig. 5 illustrates the lifetime of a fiber. The dissolution rate ishigher at pH 4 (macrophages) than at near-neutral pH (interstitialfluids) for all the compositions studied, what gives rise to a lifetimeof approximately one order of magnitude higher in the interstitialfluid compared to pH 4. At pH 4, both ligands oxalate and citratereduce approximately one order of magnitude the residence timeof the fibers with respect to that calculated in absence of ligands.Nevertheless, at pH 7.4 oxalate does not show any significant effecton the fiber lifetime. The destruction of the fibers is faster in themacrophages, where organic acids contribute creating an acidenvironment and supplying ligands. However, the complexingcapacity of the ligands in the interstitial fluids is not very effectiveto dissolve the tremolite particles, estimating lifetimes that exceedthe human life.

We may consider a particle of 0.1 lm in diameter as represen-tative of inhaled particles. Fig. 6 shows the lifetime for pH 4 and 7.4and ligands investigated. The presence of ligand in the fluids is away to improve the capacity of the tissues to dissolve inhaled par-ticles. At pH 7.4, the lifetime decreases from 92 to 52 years in cit-rate 0.15 mmol L�1. Increasing citrate concentration 100 timesonly reduce lifetime to 25 years. Lysosomal fluid in macrophagesis much more efficient than interstitial fluids. At pH 4, a fiber of0.1 lm in diameter would dissolve in 41 years in absence of organ-ic ligands. In a concentration of 0.15 mmol L�1 lifetime is reduceduntil 4.8 years in oxalate and 3.7 in citrate. Increasing citrate

iodurability of tremolite particles in mimicked lung fluids. Effect of citratej.jseaes.2013.04.008

Fig. 6. Lifetime dependence of a 0.1 mm diameter fiber with ligand concentrationat pH 4 and 7.4.

8 M. Rozalen et al. / Journal of Asian Earth Sciences xxx (2013) xxx–xxx

concentration does not have a stronger effect. In comparison, cit-rate is more efficient and would dissolve the fiber in 3.7 years.However, the increase of the ligand concentration 100 times doesnot reduce substantially the tremolite lifetime. The results fortremolite contrast with those for chrysotile, whose lifetime canbe estimated in about 1 month (Hume and Rimstidt, 1992).

These estimations of lifetime point out the long residence time oftremolite particles in the lung and alert on the severe risk of lung dis-eases once inhaled. The estimated residence time for thin and shortparticles that can be engulfed by macrophages are in the order ofmonths or less. In addition to clearance mechanisms, the macro-phages are likely efficient to contribute to degradation of tremoliteby dissolution reactions. However, when thicker particles reachthe alveolar cavities, they are too big to be cleared and/or phagocy-ted. Thus they lay on the lining tissue in alveoli, where the interstitialfluid does not reach to degrade them efficiently, and may remain forlong time even longer than human lifespan.

5. Conclusions

The dissolution rate of tremolite in mimicked lung fluids con-taining citrate and oxalate was investigated at 37 �C at pH 4, 5.5and 7.4. The results obtained from these experiments support thefollowing conclusions:

1. Calculated dissolution rates (derived from Si release to solution)depend on pH, decreasing as pH increases from 4 to 7.4. Thispattern is consistent with a proton-promoted mechanism, dis-solving two times faster at pH 4 than at pH 7.4.

2. Both ligands used in this study, oxalate and citrate, have a sig-nificant effect enhancing dissolution rates of tremolite particles.This effect is stronger as concentration of ligand increases from0.15 to 1.5 mmol L�1.

3. Ligand effect also depends on pH due to the formation of aque-ous magnesium complexes. For pH 4 (macrophages), the forma-tion of magnesium citrate complexes (MgH2Cit+, MgHCit andMgCit�) is more effective as the formation of the correspondingoxalate complex (MgOx) increasing dissolution rates until 20times respect to ligand-free solutions. However, for pH 7.4(interstitial fluids) the formation of the oxalate complex (MgOx)increase dissolution rates eight times respect to the four timesthan citrate complex MgCit� is able to do it.

4. In presence of citrate and oxalate, two mechanisms contributeto the mineral alteration process: proton- and ligand-promoteddissolution mechanisms need to be considered in order toquantify the tremolite dissolution in lungs fluids conditions.

Please cite this article in press as: Rozalen, M., et al. Dissolution kinetics and band oxalate. Journal of Asian Earth Sciences (2013), http://dx.doi.org/10.1016/

5. The biodurability estimation developed from reaction kineticsshows that tremolite is less biodurable in macrophages (lowpH) than in interstitial fluids (near-neutral). For magrophages,both ligands help to reduce the residence time of the fibersbut citrate is more effective. However, for interstitial fluids,the presence of oxalate or citrate does not reduce significantlythe lifetime of the fibers.

Acknowledgements

Financial support was obtained from Junta de Andalucía byContract P07-RNM-02772 and Group RNM-264. MR benefited froma JAE-Doc contract from CSIC with contribution of FEDER funds.The authors thank Paloma Sanchez and Eduardo Flores for labora-tory work.

References

Bulsari, A., Bergman, N., Eusch, I., Fellman, J., Perander, M., Suvorov, D., 2007.Correlation between in vitro and in vivo dissolution behaviour of stonewools bynonlinear modelling techniques. Journal of the European Ceramic Society 27,1837–1841.

Cama, J., Ganor, J., Ayora, C., Lasaga, A.C., 2000. Smectite dissolution at 80 C and pH8.8. Geochimica Cosmochimica Acta 64, 2701–2717.

Case, B.W., 1991. Health effects of tremolite. Now and in the future. Annals of NewYork Academy of Sciences 643, 491–504.

Cambier, P., Sposito, G., 1991. Interactions of citric acid and synthetic hydroxy-aluminum montmorillonite. Clays & Clay Minerals 39, 158–166.

Collier, C.G., Pearce, M.J., Hodgson, A., Ball, A., 1992. Factors affecting the in vitrodissolution of cobalt oxide. Environmental Heath Perspectives 97, 109–113.

Donaldson, K., Tran, C.L., 2004. An introduction to the short-term toxicology ofrespirable industrial fibres. Mutation Research/Fundamental and MolecularMechanisms of Mutagenesis 553, 5–9.

Donaldson, K., 2009. The inhalation toxicology of p-aramid fibrils. Critical Reviewsin Toxicology 39, 487–500.

Ganor, J., Lasaga, A.C., 1994. The effects of oxalate acid on kaolinite dissolution rate.Mineralogical Magazine 58A, 315.

Golubev, S.V., Pokrovsky, O.S., 2006. Experimental study of the effect of organicligands on diopside dissolution kinetics. Chemical Geology 235, 377–389.

Golubev, S.V., Bauer, A., Pokrovsky, O.S., 2006. Effect of pH and organic ligands onthe kinetics of smectite dissolution at 25 �C. Geochimica Cosmochimica Acta 70,4436–4451.

Grasshoff, K., Ehrhardt, M., Kremling, K. (Eds.), 1983. Methods of Seawater Analysis,second ed.. Verlag Chemie, p. 29pp.

Gunter, M.E., Wood, S.A., 2000. Can chrysotile alter to tremolite in the human lung?AGU Spring Meeting 2000, M51A.

Gustafsson, J.P., 2010. Visual MINTEQ 3.0 For Windows XP/7.Hume, J.P., Rimstidt, J.D., 1992. The biodurability of chrysotile asbestos. American

Mineralogist 77, 1125–1128.Jurinski, J.B., Rimstidt, J.D., 2001. Biodurability of talc. American Mineralogist 86,

392–399.Kummert, R., Stumm, W., 1980. The surface complexation of organic acids and

hydrous-Al2O3. Journal of Colloid and Interface Science 75, 373–385.Liu, Y., Olsen, A.A., Rimstidt, J.D., 2006. Mechanism for the dissolution of olivine

series minerals in acidic solutions. American Mineralogist 91, 455–458.Mast, M.A., Drever, J.I., 1998. The effect of oxalate on the dissolution rates of

oligoclase and tremolite. Geochimica Cosmochimica Acta 51, 2559–2568.Maxim, L.D., Hadley, J.G., Potter, R.M., Niebo, R., 2006. The role of fiber durability/

biopersistence of silica-based synthetic vitreous fibers and their influence ontoxicology. Regulatory Toxicology and Pharmacology 46, 42–62.

McDonald, J.C., McDonald, A.D., 1997. Chrysotile, tremolite and carcinogenicity. TheAnnals of Occupational Hygiene 41, 699–705.

NIST Standard Reference Database 46 Version 6.0. NIST critically selected stabilityconstants of metal complexes. In: Martell, A.E., Smith, R.M., 2001. NIST StandardReference Data, Gaithersberg, MD 20899, USA.

Oelkers, E.H., Schott, J., 2001. An experimental study of enstatite dissolution rates asa function of pH, temperature, and aqueous Mg and Si concentration, and themechanism of pyroxene/pyroxenoid dissolution. Geochimica et CosmochimicaActa 65, 1219–1231.

Olsen, A.A., Rimstidt, J.D., 2008. Oxalate-promoted forsterite dissolution at low pH.Geochimica et Cosmochimica Acta 72, 1758–1766.

Osmond-McLeod, M.J., Poland, C.A., Murphy, F., Waddington, L., Morris, H., Hawkins,S.C., Clark, S., Aitken, R., McCall, M.J., Donaldson, K., 2011. Durability andinflammogenic impact of carbon nanotubes compared with asbestos fibres.Particle and Fibre Toxicology 8, 15.

Oze, C., Solt, K., 2010. Biodurability of chrysotile and tremolite asbestos in simulatedlung and gastric fluids. American Mineralogist 95, 825–831.

iodurability of tremolite particles in mimicked lung fluids. Effect of citratej.jseaes.2013.04.008

M. Rozalen et al. / Journal of Asian Earth Sciences xxx (2013) xxx–xxx 9

Plumlee, G.S., Ziegler, T.L., 2003. The medical geochemistry of dust, soils, and otherearth materials. In: Holland, H.D., Turekian, K.K. (Eds.), EnvironmentalGeochemistry. Treatise on Geochemistry, vol. 9. Elsevier, pp. 263–310.

Ramos, M.E., Cappelli, C., Rozalen, M., Fiore, S., Huertas, F.J., 2011. Effect of lactate,glycine, and citrate on the kinetics of montmorillonite dissolution. AmericanMineralogist 96, 768–780.

Robie, R.A., Hemingway, B.S., 1995. Thermodynamic properties of minerals andrelated substances at 298.15 K and 1 bar (105 Pascals) pressure and at highertemperatures. U.S. Geological Survey Bulletin 2131, 1–416.

Rozalen, M.L., Huertas, F.J., Brady, P.V., Cama, J., Garcia-Palma, S., Linares, J., 2008.Experimental study of the effect of pH on the kinetics of montmorillonitedissolution at 25 �C. Geochimica Cosmochimica Acta 72, 4224–4253.

Rozalen, M., Ramos, M.E., Hernandez-Corbalan, M.D., Gervilla, F., Huertas, F.J., 2012.Oxalate-promoted dissolution of chrysotile. Macla 16, 156–157.

Saldi, G.D., Kohler, S.J., Marty, N., Oelkers, E.H., 2007. Dissolution rates of talc as afunction of solution composition, pH and temperature. GeochimicaCosmochimica Acta 71, 3446–3457.

Scholze, H., Conradt, R., 1987. An in vitro study of the chemical durability ofsiliceous fibres. The Annals of Occupational Hygiene 31 (4B), 683–692.

Please cite this article in press as: Rozalen, M., et al. Dissolution kinetics and band oxalate. Journal of Asian Earth Sciences (2013), http://dx.doi.org/10.1016/

Schott, J., Berner, R.A., Sjoberg, E.L., 1981. Mechanism of pyroxene and amphiboleweathering – I. Experimental studies of iron-free minerals. Geochimica etCosmochimica Acta 45, 2123–2135.

Sigg, L., Stumm, W., 1981. The interactions of anions and weak acids with thehydrous goethite surface. Colloids and Surfaces 2, 101–117.

Stettler, L.E., Sharpnack, D.D., Krieg, E.F., 2008. Chronic inhalation of short asbestos:Lung fiber burdens and histopathology for monkeys maintained for 11.5 yearsafter exposure. Inhalation Toxicology 20, 63–73.

Stumm, W., Kummert, R., Sigg, L., 1980. A ligand exchange model for the adsorptionof inorganic and organic ligands at hydrous oxide interfaces. Croatica ChemicaActa 53, 291–312.

U.S. Geological Survey, 2012. Mineral Commodity Summaries 2012. U.S. GeologicalSurvey, 198p.

USGS. Mineral Industry Surveys. World Asbestos Consumption 2003–2007. Reston,VA: U.S. Geological Survey.

Werner, A.J., Hochella, M.F., Guthrie Jr., G.D., Hardy, J.A., Aust, A.E., Rimstidt, J.D.,1995. Asbestiform riebeckite (crocidolite) dissolution in the presence of Fechelators: implications for mineral-induced disease. American Mineralogist 80,1093–1103.

iodurability of tremolite particles in mimicked lung fluids. Effect of citratej.jseaes.2013.04.008