Journal of Colloid and Interface Science 351 (2010) 50–56

Contents lists available at ScienceDirect

Journal of Colloid and Interface Science

www.elsevier .com/locate / jc is

Phase transformation in CaCO3 polymorphs: A spectroscopic, microscopicand diffraction study

Moritz Schmidt a,1, Thorsten Stumpf a,b,*, Clemens Walther a, Horst Geckeis a,b, Thomas Fanghänel c,d

a Institut für Nukleare Entsorgung, Forschungszentrum Karlsruhe, Karlsruhe Institute of Technology, P.O. Box 3640, 76021 Karlsruhe, Germanyb Institut für Anorganische Chemie, Universität Karlsruhe (TH), Engesserstr. 15, 76131 Karlsruhe, Germanyc Physikalisch-Chemisches Institut, Ruprecht-Karls-Universität Heidelberg, Im Neuenheimer Feld 253, 69120 Heidelberg, Germanyd European Comission, Joint Research Centre, Institute for Transuranium Elements, P.O. Box 2340, 76125 Karlsruhe, Germany

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

Article history:Received 8 April 2010Accepted 13 July 2010Available online 17 July 2010

Keywords:Solid solutionCalciteVateriteCm(III)ActinidesLanthanidesTRLFSXRDSEM

0021-9797/$ - see front matter � 2010 Elsevier Inc. Adoi:10.1016/j.jcis.2010.07.026

* Corresponding author at: Institut für Nukleare EntKarlsruhe, Karlsruhe Institute of Technology, P.O.Germany. Fax: +49 (0)7247 82 3927.

E-mail addresses: Thorsten.Stumpf@kit.edu,(T. Stumpf).

1 Current address: Chemical Sciences and EngineeriLaboratory, 9700 South Cass Avenue, Argonne, Illinois

This study presents results of the phase transformation from Cm(III) and Eu(III) doped vaterite to calcite.This transformation of one solid solution (An/Ln:vaterite) to another (An/Ln:calcite) was observed bypowder X-ray diffraction and scanning electron microscopy. These observations were combined withsite-selective time-resolved laser fluorescence spectroscopy (TRLFS), using Eu3+ and Cm3+ as atomicprobes, which give an internal view of the structure. The transition from vaterite to the thermodynam-ically stable CaCO3 polymorph calcite lasts several days. It could be shown that the transformation is tak-ing place in four steps: initial precipitation of low crystalline vaterite, followed by transformation into thecrystalline phase, upon suspending the vaterite in CaCO3 solution the phase transformation to calcitestarts. As third step a transition state with again partly hydrated Eu3+ can be observed before the trans-formation is completed after 72 h. No transition is observed in vaterite kept in vacuum, demonstratingthat the transition follows a dissolution/precipitation mechanism.

Comparison with Eu3+-doped calcite directly synthesized under near-equilibrium conditions showsthat identical solid solutions are formed, independent of the reaction path. Moreover the trivalent guestcations are fully transferred to the newly formed phase. This is strong evidence for a thermodynamicdriving force for the solid solution formation in these systems.

� 2010 Elsevier Inc. All rights reserved.

1. Introduction

Calcium carbonate, CaCO3, is one of the most frequently foundminerals in nature. It is important not only for geochemistry, butalso for large scale technological challenges like CO2 sequestration[1,2] or nuclear waste disposal as well as for every day applica-tions, e.g., pipe scaling [3]. Moreover, CaCO3 is frequently foundas biogenic material in forms of shell or skeleton material [4]. Car-bon dioxide sequestration as carbonate materials seems to be apromising way to bury billions of tons of CO2 without the risk ofrelease. In nuclear waste management calcite has proven to be avaluable mineral phase for the retention of highly radiotoxic actin-ides [5–9]. In this context the incorporation of the trivalent

ll rights reserved.

sorgung, ForschungszentrumBox 3640, 76021 Karlsruhe,

Thorsten.Stumpf@ine.fzk.de

ng Division, Argonne National60439, USA.

actinide curium into CaCO3 mineral phases while retaining thehost’s crystal structure – the formation of a solid solution – is par-ticularly interesting. There has been an ongoing interest in thepolymorphic CaCO3 system and the transitions therein from vari-ous scientific communities [4,10–14].

The CaCO3 mineral system consists of three anhydrous poly-morphs (in order of stability: calcite, aragonite and vaterite) thatcan all be found in nature. The thermodynamically stable poly-morph under normal conditions is calcite. Aragonite is formed bio-genically, e.g., as shell material of molluscs, or precipitates underspecial conditions. Aragonite is metastable but its phase transfor-mation to calcite is sufficiently slow to form geological formations.The other metastable polymorph vaterite precipitates as a precur-sor in conditions of high supersaturation and transforms to the sta-ble phase within hours to days. It is, however, still unknown howthis phase transformation takes place. Both, an internal rearrange-ment and a dissolution/precipitation mechanism are discussed[15–19]. To distinguish between these two possibilities we useda combination of microscopy, diffraction and spectroscopy andthe results are presented in this paper.

X-ray powder diffraction (XRD) is a well established standardmethod to gain bulk structural information. It is combined with

M. Schmidt et al. / Journal of Colloid and Interface Science 351 (2010) 50–56 51

scanning electron microscopy (SEM) providing for the external,microscopic information. In addition we obtain an internal viewon the structure of the samples by the application of fluorescentatomic probes. Trivalent lanthanides and actinides have beenwell-known to interact with the two stable polymorphs of CaCO3

by the formation of solid solutions [20,21]. Previous work of ourgroup could show that the trivalent ions replace Ca2+ on its latticesite in the host structure. Hence, their fluorescence emission can beused to gain internal structural information on the investigatedmineral phase. Eu3+ fluorescence is particularly suitable for identi-fication of the local symmetrical environment. The splitting of the7F1, 2 levels together with the hypersensitivity of the 5D0 ?

7F2

transition yields valuable information of the point symmetry ofthe Eu(III) in the crystal lattice [22–24]. Cm3+ fluorescence spec-troscopy is most prominently characterized by its outstanding sen-sitivity. In solution speciation down to 10�9 mol l�1 and detectionlimits of 10�12 mol l�1 is possible [25,26]. In the solids used in thisstudy speciation is still possible at 0.1 ppm Cm3+. Due to spin–or-bit-coupling the Cm3+ 5f-5f-transitions exhibit strong shifts withchanges in the ligand field. Stronger coordination induces a stron-ger bathochromic shift of the fluorescence signal [26]. Due to thelarge energy gap between the emitting state and the ground statefor both ions (Eu and Cm), the only efficient quenching mechanismis O(N)–H-vibrational quenching. The fluorescence decay constantk (the inverse fluorescence life time, k = 1/s) correlates with thenumber of coordinated water molecules in the first coordinationsphere [27,28]. This allows one to distinguish between innersphere, outer sphere and bulk incorporated species. Comprehen-sive reviews of Eu3+ and Cm3+ fluorescence spectroscopy can befound in literature [25,29].

Our previous results on calcite samples grown in a mixed-flowreactor at near equilibrium ‘‘steady state” conditions showedstrong interactions with both, Eu3+ and Cm3+. Three different spe-cies – A, B, and C – could be identified in both systems. Species Awas identified as a surface incorporation species by its signal’sshort fluorescence lifetime indicating partial hydration. The fluo-rescence signal is red shifted weakly to 578.1 nm (Eu3+ 7F0 ?

5D0)and 604.5 nm (Cm3+ 8S7/2 ?

6D7/2) respectively. As expected for asurface species the emission spectra show splitting patterns indi-cating low symmetry coordination environments. Species B is anincorporation species, evident from its signal’s long fluorescencelifetime. The emission spectra, however, indicate that the coordi-nation environment has low symmetry, which is consequence ofa distortion of the calcite lattice. The signal positions for this spe-cies are more strongly red shifted compared to species A and arefound at 578.4 nm (Eu3+ 7F0 ?

5D0) and 616.9 nm (Cm3+ 8S7/

2 ? 6D7/2) respectively. Species C finally is Eu3+/Cm3+ incorporatedinto calcite on the nearly undistorted Ca2+ lattice site. This becomesevident by very long fluorescence lifetimes, strongly red shiftedsignals and emission spectra showing splitting patterns in agree-ment with the trigonal symmetry of the lattice site. This species’signals are found at 579.6 nm (Eu3+ 7F0 ?

5D0) and 624.4 nm(Cm3+ 8S7/2 ?

6D7/2) respectively.

2. Experimental details

Eu3+-doped vaterite was precipitated by a method described byTurnbull [30]. The precipitation solution is a 2 M NH4OH solution(Merck, suprapur) of 1 M CaCl2 (Merck, p.a.) with either 20 ppmEu3+ or 0.2 ppm Cm3+ relative to Ca2+. The curium isotope used is248Cm, an alpha-emitter with a half life of 3.4 � 105 a. This solutionwas bubbled with CO2 for 30 min. Precipitation started after 5 to10 min, indicated by a clouding of the solution.

The solid was then filtered, washed with Milli-Q water and eth-anol (Merck, 99.8% pure) and brought into a desiccator, where it

was stored in vacuum. After 24 h the solid was suspended in satu-rated CaCO3 solution in a solid/liquid ratio of 10 g/L. The samplestaken in the course of the transformation were treated in the samemanner as the precipitated solid.

The obtained samples contained 100 ppm Eu3+ and 0.3 ppmCm3+, as confirmed by ICP-MS. There was no significant changeof the Eu3+/Cm3+ concentration during the transformation.

The laser setup consisted of a XeCl excimer pumped dye laserdirectly exciting Eu3+’s 5D0 level or Cm3+’s 6D7/2 level, respectively.To achieve the desired spectral resolution the powder sampleswere cooled to temperatures below 20 K in a helium refrigeratedcryostat. The dyes coumarin 153 and rhodamine B were used fordirect excitation of Eu3+ and Cm3+, respectively. The luminescenceemission signal was recorded time resolved by a fiber coupled opti-cal multi-channel system consisting of a polychromator with 300,600 and 1200 lines/mm gratings and a gated, intensified diodearray.

For measurement of the powder diffraction patterns a BrukerD5000 diffractometer (Cu radiation at 40 kV and 40 mA) was used.The measurements were carried out with a range of 5–70�, a stepwidth of 0.01� and one second counting time.

The samples (XRD, TRLFS and SEM) were filtered where re-quired, washed with ethanol and resuspended in Milli-Q water.These suspensions were dried on a sample holder and measured.The SEM micrographs were measured in a CamScan FE44microscope.

3. Results

3.1. SEM results

Fig. 1 shows the SEM micrographs taken 2 h after precipitation(a), after 24 h in vacuum (b), as well as after 24 h, 36 h, 48 h and72 h in saturated CaCO3 solution (c–f, respectively). The two occur-ring mineral phases, vaterite and calcite, can clearly be distin-guished by their characteristic morphologies. Vaterite can berecognized as framboidal spheres [31,32] while calcite formsrhombohedral crystallites [33]. Fig. 1a shows the CaCO3 whichwas taken directly from the precipitation solution, filtered anddried from ethanol suspension. In total, approximately 2 h afterstarting the CO2 bubbling (methods section). While spherical vate-rite particles are already recognizable a considerable ratio of thematerial is present in amorphous form without apparent micro-crystalline morphology. After 24 h in vacuum the only particlespresent in the sample are the framboidal spheres which are typicalfor vaterite (Fig. 1b).

Once this vaterite is suspended into saturated CaCO3 solution atransformation can be observed in the micrographs (c–f). Rhombo-hedral calcite crystallites appear on the vaterite spheres after 24 h.Calcite particles are only observed in direct contact with vaterite aslong as both phases are present. With prolonged contact time theamount of vaterite decreases and more calcite is formed on thevaterite surface. After 48 h suspension time only calcite particlescan be found. Many of them, however, still show spherical cavitiesindicating the prior location of a vaterite particle. After 72 h thetransformation is complete, as far as can be observed by SEM. Onlyrhombohedral calcite is present and the number of spherical defor-mations is strongly reduced.

3.2. X-ray powder diffraction

To monitor the bulk structural changes of the CaCO3 samples X-ray powder diffraction patterns were measured in the range from5� to 70�. Fig. 2 shows the region from 10� to 45�, containing themost intense peaks for both mineral phases.

Fig. 1. SEM micrographs of the various stages of the vaterite calcite transformation: (a) 2 h after precipitation, (b) after 24 h in vacuum, (c–f) after 24 h in vacuum and 24 h,36 h, 48 h, and 72 h in saturated CaCO3 solution, respectively (inserts b: 10 lm, f: 3 lm scale).

Fig. 2. X-ray powder diffraction patterns of CaCO3 samples at different stages of thetransformation [34,35].

52 M. Schmidt et al. / Journal of Colloid and Interface Science 351 (2010) 50–56

The diffraction pattern obtained after 24 h in vacuum shows avaterite pattern with no recognizable impurities [34]. This patternremains unchanged after 4 months in vacuum. In CaCO3 solutionthe calcite (1 0 4) peak at 29.41� [34] becomes observable after24 h, indicating the start of calcite formation. All following mea-surements yielded pure calcite patterns with no observable diffrac-tion from vaterite.

3.3. Results from time-resolved laser fluorescence spectroscopy

Selective excitation of the Cm3+ 8S7/2 ?6D7/2 transition at low

temperatures allows discrimination of different coordinative envi-ronments of Cm3+ in a sample. The Cm(III) containing sampleswere cooled to below 20 K and subsequently excited through thespectral range of interest from 595 to 625 nm. In the case of Eu(III)doped CaCO3 the 7F0 ?

5D0 transition was excited between 574and 583 nm, which is non-degenerate due to the J = 0 characteris-

tics of both energy levels involved in the transition. The excitationspectrum was obtained as the integral emission intensity as a func-tion of excitation wavelength. Using this method, the resolution isnot limited by the spectral resolution of the camera setup but onlyby the tuning resolution of the laser system. The tuning resolutionis 0.001 nm in the present case.

When the sample was frozen down directly after precipitation,the spectra shown in black in Fig. 3 were acquired. The Cm3+ spec-trum consists of two peaks at 612.1 nm and 601.8 nm, respectively,corresponding to two sublevels of the 6D7/2 level, usually labeled A1

and A2. Both transitions are rather broad and do not feature a rec-ognizable splitting pattern. They do not, however, coincide withany of the known transitions of Cm3+ species in calcite (A:604.5 nm, B: 616.9 nm and C: 624.4 nm, respectively [20]). TheEu3+ excitation spectrum corroborates these observations. Onebroad peak centered at 579.3 nm is observed, with a distinctiveshoulder to the blue side of the spectrum. Once again, it can be ru-led out, that any of the calcite species are observed, whose signalsare at 578.1 nm (A), 578.4 nm (B) and 579.6 nm (C).

After 24 h in vacuum the excitation spectra changed in accor-dance with the previously described observations. While the peakcenters stayed the same for both Cm3+ and Eu3+, the peaks becamesignificantly sharper. Furthermore, an additional peak appeared inthe Cm3+ spectrum at 619.1 nm indicating the presence of a secondminor species. Due to the lower sensitivity of Eu3+ fluorescencespectroscopy such a species is not observable in the Eu3+ excitationspectrum.

Although macroscopically minor calcite traces were already de-tected after 24 h in the CaCO3 solution no significant changes areobservable in the excitation spectra. Apparently no exchange ofEu(III) or Cm(III) to calcite takes place as long as vaterite is thedominant mineral phase.

The phase transformation became prominent after 36 h in solu-tion. The Eu3+ excitation spectrum now features a total of five dif-ferent species. All three well-known Eu-calcite species are presentnow. The surface incorporation species A is found at 578.1 nm, andthe asymmetrically incorporated species B at 578.4 nm. The sym-metrically incorporated species C, with Eu3+ replacing Ca2+ on itsnearly undisturbed lattice site is centered at 579.6 nm. Species C

Fig. 3. Excitation spectra of the Cm3+ samples (below) and Eu3+ samples with the same time steps described above.

M. Schmidt et al. / Journal of Colloid and Interface Science 351 (2010) 50–56 53

is dominant in the spectrum, however; a strong shoulder indicatesthe presence of the Eu-vaterite species at 579.3 nm. In addition anew species at 579.1 nm is observed, that matches neither aknown calcite nor vaterite species.

After another 12 h in solution only the known calcite specieswere found for both, Cm3+ and Eu3+. All three known species arereproduced though the relative intensities vary from those ob-tained by direct near-equilibrium precipitation of calcite. The

Cm3+ excitation spectrum in the sample from this work is domi-nated by the surface incorporation species at 604.5 nm which coin-cides with the A2 level of the incorporation species C, which is herefound at 623.8 nm. Its peak is extremely broad and does not showthe splitting pattern known from calcite samples obtained underdifferent conditions. The second incorporation species B at616.9 nm is strongly suppressed and therefore is hardly visible inthe broad shoulder of the signal from species C. This may be due

Table 1Summary of measured lifetimes for all species and the correlated amount ofcoordinating water molecules.

Speciesdescription

Exc. wavelength(nm)

Lifetime(ls)

Calcite Ref. [19](ls)

Vaterite incorporationEu3+ 579.3 4069 ± 244 –Cm3+ 612.1 1802 ± 216

Minor vateriteEu3+ – – –Cm3+ 619.1 2569 ± 308

TransitionEu3+ 579.1 1106 ± 133 –

Calcite sorption AEu3+ 578.1 412 ± 49 460Cm3+ 604.6 775 ± 92 820

Calcite incorporation BEu3+ 578.4 4368 ± 524 3610Cm3+ 616.9 1543 ± 185 1453

Calcite incorporation CEu3+ 579.6 4352 ± 261 3663Cm3+ 623.7 5067 ± 456 4608

54 M. Schmidt et al. / Journal of Colloid and Interface Science 351 (2010) 50–56

to the very low Cm3+ concentration in the sample. While the Cm3+

excitation spectrum remained widely unchanged after another24 h in the CaCO3 solution, in the Eu3+ spectrum another redistri-bution of the three species can be observed. The asymmetricallyincorporated species B became more pronounced, while the peaksgenerally appear sharper.

The spectrum obtained from the sample that remained in vac-uum for 4 months after precipitation, is identical with that fromthe pure vaterite sample acquired after 24 h. In the absence of asolution phase no transformation is observed.

For the determination of fluorescence lifetimes the camerashutter (gating of the intensifier) is opened for 10 ms after a vari-able delay following the laser pulse. The first spectrum in each de-cay profile is taken after 1 ls with steps between 20 and 250 ls,adjusted to acquire around 50 points in a decay range of two log-arithmic units. The fluorescence lifetime is determined by fittinga first or second order exponential decay function to the plot ofintegrated fluorescence intensity over delay time. Representativefluorescence decay profiles for all five species in the Cm3+ andEu3+ systems are plotted in Fig. 4. Exponential fitting of the plotsyield the fluorescence lifetimes that can be correlated to theamount of water present in the first coordination sphere.

The Eu3+ incorporation species in vaterite (kexc = 579.3 nm) hasa lifetime of sEu = 4069 ± 244 ls, while the Cm3+ incorporation spe-cies in vaterite (kexc = 612.1 nm) has a lifetime ofsCm = 1802 ± 216 ls. Both lifetimes correspond to a complete lossof hydration, according to Kimura’s and Horrocks’ equations,respectively, indicating that the fluorescent probes are incorpo-rated into the vaterite structure.

The minor species found in the Cm3+ vaterite samples at619.1 nm has a fluorescence lifetime of 2569 ± 308 ls. Because ofthe dominance of the major species at 612.1 nm, it was not possi-ble to obtain a mono-exponential decay pattern. A bi-exponentialfit with the major species’ lifetime as second component yieldedsound results.

The transition species found in the Eu3+ sample after 36 h inCaCO3 solution could not be excited separately, due to the verynarrow distribution of five distinct species. Therefore, a bi-expo-nential decay pattern was obtained. Reasonable fits could beachieved assuming an overlap with either the lifetime of the vate-rite incorporation species or the symmetrically incorporated spe-cies in calcite, as both are nearly identical. For both fits a lifetimeof 1106 ± 133 ls can be calculated. This lifetime clearly shows, thata unique species is obtained, as it matches neither a calcite norvaterite species. It correlates with 0.5 water molecules in the firstcoordination sphere of Eu3+.

Fig. 4. Representative fluorescence decay profiles of all Cm3+ (left) and

For the species which were present when calcite was the dom-inating mineral phase the expected lifetimes known from previousworks were finely reproduced. The sorption species A was not ex-cited individually, as can be recognized by the bi-exponential de-cay patterns. The plots can be fitted with the incorporationspecies lifetime as a second component. The resulting lifetimesare sEu = 412 ± 49 ls for Eu3+ and sCm = 775 ± 92 ls for Cm3+. Thisindicates partial hydration of this species. The lifetimes obtainedfor the first incorporation species B are sEu = 4368 ± 524 ls andsCm = 1543 ± 185 ls, respectively. In both cases no mono-exponen-tial decay patterns could be obtained, due to the very low intensityin the Cm3+ spectra and overlap in the case of Eu3+. The very longlifetimes indicate that the An/Ln species is not coordinated bywater anymore. The lifetimes found for the other calcite incorpora-tion species C (sEu = 4352 ± 261 ls and sCm = 5067 ± 456 ls) are ingood agreement with the known lifetimes from the previous cal-cite experiments. All lifetimes are summarized in Table 1.

The splitting and relative intensity of the two most intense tran-sitions in the Eu3+ emission spectrum 5D0 ?

7F1 and 5D0 ?7F2

yield information about the Eu3+ local crystal field. Consequently,each of the five species identified in the Eu3+-system has a uniquefluorescence emission. Each species is excited at its specific excita-

Eu3+ (right) species present in the course of the phase transition.

Fig. 5. Fluorescence emission spectra of all species present in the Eu3+-vaterite/calcite system with calcite Ref. [19] where applicable.

M. Schmidt et al. / Journal of Colloid and Interface Science 351 (2010) 50–56 55

tion wavelength, as obtained from the excitation spectra (seeabove). Thus, single species emission spectra should be obtained.They are shown in Fig. 5.

Immediately after precipitation, the spectrum consists of athreefold split 7F1 band, and a 7F2 band that has a structure thatis not resolved due to the broad line widths.

After 24 h in the vacuum, the 7F1 emission remains almost iden-tical, while a fourfold splitting of the hypersensitive 7F2 transitioncan be obtained. In addition the 7F1/7F2 ratio has changed from 0.54to 0.99. The same emission is observed after an additional4 months in vacuum.

The transition species features a strongly structured spectrumwith a very small 7F1/7F2 ratio (0.37). The single lines in both bandsare rather broad, so that a clear distinction is not easily possible.

For the three species found in the samples where calcite is thedominating mineral phase, the spectra could be compared to thepreviously known ones [19]. In each case an excellent agreementwith the literature data was achieved. For both sorption speciesA and incorporation species B a full splitting of both bands is ob-served. The high symmetry incorporation species C shows a two-fold splitting of 7F1 and a threefold splitting of 7F2 is observed, asit was expected based on the calcite samples.

4. Discussion and conclusion

The paper of Tsuno et al. [14] seems to contradict part of thefindings. However, Tsuno et al. investigated the inhibiting effectof La3+ on the phase transformation vaterite – calcite by using veryhigh La/Ca ratios (700 ppm) and short reaction times (24 h). More-over, they keep the originally precipitated vaterite in contact witha La3+ solution at all times during the experiment which is not thecase in the presented study.

Hence, the data which are presented indicate a four step mech-anism for the An/Ln doped vaterite–calcite precipitation/phasetransition process: initial precipitation of low crystalline vaterite,

followed by transformation into the crystalline phase, upon sus-pending the vaterite in CaCO3 solution the phase transformationto calcite starts. As third step a transition state with again partlyhydrated Eu3+ can be observed before the transformation is com-pleted after 72 h.

In the first step a fast precipitation of a poorly crystalline vate-rite-like mineral phase can be observed. This becomes obvious be-cause of the absence of spherical particles as expected for purevaterite in the SEM micrographs. Another clear indication of thelow crystallinity is the line broadening in the Cm3+ and Eu3+ exci-tation spectra, as well as in the Eu3+ emission spectrum.

This vaterite-like precursor phase crystallises to vaterite within24 h in vacuum and the SEM shows the expected framboidalspheres. The XRD confirms this, as the obtained pattern containsall vaterite reflections in the expected relative intensities with nodetectable impurities. In the TRLFS spectra this results in sharperlines, and clearly defined splitting patterns. The splitting patternobserved in the Eu3+ emission spectrum, can be used to identifythe symmetry of the Eu3+ coordination environment. A threefoldsplitting in F1 with a fourfold splitting of F2 indicates tetragonalpoint symmetry, e.g., D4h, C4v, D2d or S4. The Ca2+ lattice site inthe vaterite unit cell has a very low symmetry, which would inducea fivefold splitting in F2. Most likely this fifth transition is veryweak and thus not observed.

When the An/Ln doped vaterite came into contact with CaCO3

solution the transition began. After 24 h exposure there was evi-dence of small quantities of calcite in the sample by SEM micro-graphs and XRD patterns. SEM shows the appearance of the firstisolated rhombohedral crystallites on top of the dominating vate-rite particles. The corresponding powder X-ray diffraction patternincludes the calcite (1 0 4) peak with low intensity. The An/Ln solidsolution apparently does not change from vaterite to calcite in suf-ficient quantity to be observed by TRLFS, as long as vaterite is thedominant mineral phase.

After 36 h in carbonate solution, no vaterite was detected byXRD, while SEM showed that single vaterite-like particles were stillpresent. The TRLFS clarified, that this stage of the transformationcan be understood as a transition state. This became obvious bythe observation of an additional species in the system, featuringall Eu-calcite and the Eu-vaterite species. The transition speciesis characterized by a short lifetime and a strongly split emissionspectrum, with a very strong 7F0 transition. As the lifetime corre-sponds to 0.4 ± 0.5 H2O ligands we most likely observe Eu3+ inthe state of reincorporation while it is still coordinated by onewater molecule or hydroxide. This observation can only be under-stood as proof of a phase transition by dissolution andprecipitation.

After 2 days of having the sample in suspension, only calciteis detected by SEM and XRD. The SEM micrographs give anotherclear indication for the mechanism of the transition. The pres-ence of rhombohedral calcite particles with spherical cavitiescan only be explained by a dissolution/precipitation mechanism,where precipitation occurs immediately after dissolution so thatno long distance transport can occur. This agrees with the obser-vation that calcite is solely found in close proximity to vaterite.The TRLFS results corroborate these findings. Both An/Ln solidsolutions exhibit excitation spectra equivalent to those knownfrom Eu3+/Cm3+ doped calcites. The peaks are still rather broadindicating that the obtained calcite is not very crystalline andthat the transformation is not finished at this state. Anotherday in the CaCO3 solution did not change the SEM or XRD signif-icantly, also the Cm3+ excitation spectrum remained widely un-changed. Only in the Eu3+ excitation spectrum there isobserved evidence of the ongoing alteration of the material.The species show a significant redistribution, which goes alongwith a narrowing of the peaks.

56 M. Schmidt et al. / Journal of Colloid and Interface Science 351 (2010) 50–56

The results which were obtained by the investigation of the An/Ln doped vaterite in absence of carbonate solution confirm theconclusion that dissolution/precipitation is the dominant transfor-mation mechanism. When the vaterite sample is kept in vacuumeven after 4 months no alterations are observed. In vacuum intra-crystalline reorganization is the only available mechanism for thephase transformation. Consequently, one can conclude that atroom temperature the transformation from vaterite to calcite isstrongly preferred via the aqueous phase.

Furthermore, we could shed more light on the important An/Ln:calcite solid solution system. Eu3+ as well as Cm3+ were takenup quantitatively into vaterite and were completely transferredinto calcite during the phase transformation. All An/Ln:calcite solidsolution species which are known from the near-equilibriumexperiments when Eu(III) and Cm(III) were in direct contact withcalcite are reproduced during the phase transformation process.This path-independent behavior by the formation of the An/Ln:cal-cite solid solution is a clear indication that there is a thermody-namic driving force for the formation of solid solutions in theinvestigated systems.

Acknowledgments

Sebastian Büchner from INE (Karlsruhe, Germany) is gratefullyacknowledged for his for technical assistance. We want to thankKiel Holliday for the critical reading of the manuscript. This workwas co-financed by the Helmholtz Gemeinschaft Deutscher Fors-chungszentren (HGF) by supporting the Helmholtz-Hochschul-Nachwuchsgruppe ‘‘Aufklärung geochemischer Reaktionsmecha-nismen an der Wasser/Mineralphasen Grenzfläche”.

References

[1] K.S. Lackner, Science 300 (5626) (2003) 1677–1678.[2] W. Aeschbach-Hertig, Nature 458 (7238) (2009) 583–584.[3] E. Dalas, P.G. Koutsoukos, Geothermics 18 (1–2) (1989) 83–88.[4] S. Weiner, I. Sagi, L. Addadi, Science 309 (5737) (2005) 1027–1028.

[5] E. Curti, Appl. Geochem. 14 (1999) 433–445.[6] M.M. Fernandes, T. Stumpf, T. Rabung, D. Bosbach, T. Fanghänel, Geochim.

Cosmochim. Acta 72 (2008) 464–474.[7] S.L.S. Stipp, J.T. Christensen, L.Z. Lakshtanov, J.A. Baker, T.E. Waight, Radiochim.

Acta 94 (9–11) (2006) 523–528.[8] T. Stumpf, M.M. Fernandes, C. Walther, K. Dardenne, T. Fanghänel, J. Colloid

Interface Sci. 302 (1) (2006) 240–245.[9] F. Heberling, M.A. Denecke, D. Bosbach, Environ. Sci. Technol. 42 (2) (2008)

471–476.[10] A.H. Cai, X.R. Xu, H.H. Pan, J.H. Tao, R. Liu, R.K. Tang, K. Cho, J. Phys. Chem. C 112

(30) (2008) 11324–11330.[11] Y. Politi, T. Arad, E. Klein, S. Weiner, L. Addadi, Science 306 (5699) (2004)

1161–1164.[12] D. Walsh, S. Mann, Nature 377 (6547) (1995) 320–323.[13] G. Wolf, E. Konigsberger, H.G. Schmidt, L.C. Konigsberger, H. Gamsjager, J.

Therm. Anal. Calorim. 60 (2) (2000) 463–472.[14] H. Tsuno, H. Kagi, T. Akagi, Bull. Chem. Soc. Jpn. 74 (3) (2001) 479–486.[15] X. Xu, J.T. Han, D.H. Kim, K. Cho, J. Phys. Chem. B 110 (6) (2006) 2764–2770.[16] Y. Nakahara, T. Tazawa, K. Miyata, Nippon Kagaku Kaishi 5 (1976) 732.[17] L. Brecevic, D. Kralj, Croat. Chem. Acta 80 (3–4) (2007) 467–484.[18] Y.S. Han, G. Hadiko, M. Fuji, M. Takahashi, J. Cryst. Growth 289 (1) (2006) 269–

274.[19] N. Spanos, P.G. Koutsoukos, J. Cryst. Growth 191 (4) (1998) 783–790.[20] M. Schmidt, T. Stumpf, M.M. Fernandes, C. Walther, T. Fanghaenel, Angew.

Chem., Int. Ed. 47 (31) (2008) 5846–5850.[21] M. Schmidt, T. Stumpf, C. Walther, H. Geckeis, T. Fanghänel, Dalton Trans.

(2009) 6645–6650.[22] J.-C.G. Bünzli, G.R. Choppin, Lanthanide Probes in Life, Chemical and Earth

Sciences, Elsevier, Amsterdam, 1989.[23] G.R. Choppin, D.R. Peterman, Coord. Chem. Rev. 174 (1) (1998) 283–299.[24] C. Görller-Walrand, K. Binnemans, Rationalization of crystal-field parameters,

in: K.A. Gschneider, L. Eyring (Eds.), Handbook on the Physics and Chemistryof Rare Earths, Elsevier, Amsterdam, 1996, pp. 121–244.

[25] N.M. Edelstein, R. Klenze, T. Fanghänel, S. Hubert, Coord. Chem. Rev. 250 (7–8)(2006) 948–973.

[26] J.I. Kim, R. Klenze, H. Wimmer, Eur. J. Solid State Inorg. Chem. 28 (1991) 347–356.

[27] T. Kimura, Y. Kato, H. Takeishi, G.R. Choppin, J. Alloys Compd. 271–273 (1998)719–722.

[28] W.D. Horrocks Jr., D.R. Sudnick, J. Am. Chem. Soc. 101 (1979) 334–340.[29] J.-C.G. Bünzli, C. Piquet, Chem. Soc. Rev. 34 (12) (2005) 1048–1077.[30] A.G. Turnbull, Geochim. Cosmochim. Acta 37 (1973) 1593–1601.[31] J.-P. Andreassen, M.J. Hounslow, AIChE J. 50 (11) (2004) 2772–2782.[32] G. Nehrke, P.V. Cappellen, J. Cryst. Growth 287 (2006) 528–530.[33] H. Nebel, M. Epple, Z. Anorg. Allg. Chem. 634 (2008) 1439–1443.[34] S.R. Kamhi, Acta Crystallogr. 16 (1963) 770–772.[35] D.L. Graf, Am. Mineral. 46 (1961) 1283–1316.