Liquid Precursors in Non-Classical Crystallization

Abstract

Many important phenomena depend on calcium carbonate nucleation and crystallisation, e.g. the incrustation of pipelines or the formation of coral reefs. The investigation of the early stages of phase separation of calcium carbonate was the topic of this work, addressing open questions within the frameworks of "non-classical" nucleation and crystallization. As established in previous work, the first species, which appear in the nucleation process, are solute pre-nucleation clusters (PNCs), which exhibit a highly dynamic, chain-like structural form called DOLLOP (dynamically ordered liquid-like oxyanion polymers).[25] In contrast to classical nucleation theory (CNT) these clusters are thermodynamically stable and occur already in undersaturated solutions. It is hypothesized that the dynamics of PNCs can significantly decrease upon reaching a certain ion activity product (IAP), and thereby, the PNCs transform into phase-separated nanodroplets or liquid mineral precursors. These nanodroplets then agglomerate and form amorphous calcium carbonate (ACC) via solidification, or a second nucleation event. Previous work showed that ACC that is formed via equilibrated PNC precursors exhibits short-range orders that relate to distinct anhydrous crystalline polymorphs of calcium carbonate, which is controlled by pH and temperature.[53, 56] The first part of this thesis examined the transition from PNCs to liquid mineral precursors, thereby proving and quantifying the aforementioned hypothesis. The solvent, in this case water, plays an important role for nucleation mechanisms. A titration assay in combination with THz spectroscopy revealed nonlinear changes in THz absorption during the early stages of CaCO3 nucleation. This provided evidence for altered coupled motions of hydrated calcium and carbonate ions. The direct link between these changes and the continuous development of the IAP during the early stages of precipitation revealed the locus of a liquid-liquid binodal limit in aqueous CaCO3 solutions. The liquid-liquid binodal was located at the ion activity product of the proto-structured ACCs for pH 9.0 and pH 10.0 (IAPpc = 3.1∙10-8 M² and IAPpv = 3.8∙10-8 M², respectively). The THz spectroscopy experiments in combination with the titration experiments strongly suggested that, after liquid-liquid phase separation, solid ACC was formed via solidification of the liquid precursors, and not a second nucleation event. Furthermore, advanced titration experiments employing different mixing conditions illustrated that calcium into carbonate titrations (Ca2+CO32-) follow a different pathway than carbonate into calcium experiments (CO32-Ca2+), as the liquid precursors are more stable in Ca2+CO32- titrations probably due to electrostatic stabilization. In addition, CO32-Ca2+ titrations in combination with cryo-TEM supported the notion that liquid precursors always occurred during CaCO3 nucleation. Here cryo-TEM clearly showed co-existing amorphous and crystalline structures, while the Ca2+-ISE detected a solubility product related to vaterite but not ACC, which is not expected as the ISE always detects the most soluble phase. These findings imply that the observed ACC was liquid or was 'hidden' in a dense liquid phase. The second part was advanced on this insight by establishing experiments towards the construction of a phase diagram for the metastable liquid-liquid coexistence in the aqueous CaCO3 system, in which three proto-structured ACCs were considered. This was achieved by means of temperature- and rate-dependent titration experiments. These experiments showed that various different, progressively more (thermodynamically) metastable ACCs were accessible by increasing the titration rates. The precipitated ACCs most likely differed in the amounts of water incorporated into their structure as indicated by FT-IR spectroscopy, leading to different solubility products – which directly reflected their thermodynamic stability. This observation was interpreted based on the notion that the ACCs had formed from liquid precursors via solidification. Faster addition rates allowed accessing a thermodynamically more metastable regime of the liquid-liquid miscibility gap, i.e., the second liquid phase was more metastable when faster addition rates were applied. This means that solidification of a more metastable second liquid phase also led to less stable ACC. All ACCs formed in the binodal regime of the liquid-liquid miscibility gap were transient towards more stable phases, like proto-structured ACCs or crystalline forms. The THz experiments together with the advanced titration experiments demonstrated that the loci of liquid-liquid binodal limits were accessible by measuring the solubility of proto-structured ACCs. Hence, the phase diagram for the liquid-liquid miscibility gap of calcium carbonate was constructed via temperature- and pH-dependent measurements of the ACC solubilities. The phase diagram suggests the existence of a triple point (where three different liquid calcium carbonate phases co-exist, yielding proto-calcite, proto-vaterite and proto-aragonite upon solidification) at 35 °C, an IAP of 2.9∙10-8 M² and at a pH ~9.5. The spinodal limit in the system was determined by direct mixing experiments and completed the left branch of the phase diagram. In addition, the experiments allowed arriving at a model to calculate the critical temperature Tcrit for the miscibility gap by the use of the enthalpy (ΔHcluster) and entropy (ΔScluster) for the cluster formation (T_crit= 〖∆H〗_cluster⁄(〖∆S〗_cluster = ~98 K)). The influence of different additives on liquid precursors and ACCs was examined in the third chapter. THz experiments showed that polycarboxylates, which are known to stabilize liquid precursors of calcium carbonate, significantly enhanced the kinetic stability of the metastable liquid-liquid state, but they did not affect the locus of the corresponding binodal limit. A mixture of different additives, in this case polycarboxylates and magnesium ions, was also investigated by means of titration experiments. It was shown that both magnesium ions and poly(aspartic acid) (PAsp) inhibited nucleation of CaCO3. Interestingly, a combination of PAsp with magnesium ions led to synergistic effects that brought about a dramatic increase in the efficiency of nucleation and growth inhibition of nanoscopic CaCO3 precursors. Analyses of the precipitated phases showed that an amorphous phase containing calcium and magnesium ions in equimolar ratios occurred. Furthermore, PAsp triggered the crystallization process, because aragonite was observed only in the presence of PAsp. Therefore, it can be speculated that Mg2+ ions extended the existence region of liquid calcium carbonate precursors, while PAsp kinetically destabilized ACC and caused it to crystallize. These effects appear to be crucial for biomineralization processes, where polycarboxylate motifs occur in corresponding proteins, and magnesium ions are thought to play important roles. The last part of this thesis was concerned with the physical characterization of liquid calcium carbonate precursors, which were stabilized by polycarboxylates. These so-called "polymer-induced" liquid precursor (PILP) phases can play a role in non-classical calcium carbonate crystallization and growth processes, as thoroughly established in the literature.[6] In this thesis, the physical properties of PILPs were investigated by in-situ AFM studies. Here, for the first time, the observation of gel-like precursors was possible. An evaluation of the data yielded a Young’s modulus of ~2 MPa for the observed precursors, which was comparable to the stiffness of poly(acrylic acid) (PAA) gels. Crystal growth on a calcite surface was observed over time by means of an in-situ AFM. The time dependent measurements showed that crystal growth occurred via the addition of gel-like precursors, which were likely liquid before they attached to the surface. Overall, the liquid-liquid binodal limit in the aqueous calcium carbonate system was successfully quantified. In addition, it was shown that (proto-structured) ACCs form via solidification and not by a second nucleation event. Based on this knowledge, a phase diagram for the liquid-liquid miscibility gap was constructed, including three proto-structured ACCs. This is a strong indication that the aqueous CaCO3 system displays metastable liquid polyamorphism, where the three distinct liquid phases coexist at a triple point. Furthermore, the influence of additives – which are thought to be relevant for biomineralization – on the liquid-liquid miscibility gap was explored and the physical behaviour of stabilized dense liquids (i.e., PILPs) was investigated. Globally, this new data taken together helps to improve the understanding of "non-classical" nucleation and growth processes in general, and of biomineralization more specifically. Importantly, transferring the experimental techniques to other systems like calcium phosphates, or small organic molecules like amino acids, will allow testing the applicability of the mechanisms established for CaCO3 nucleation and growth in future work.