Initial microstructural analyses (Figure 1) were conducted in order to observe the morphology of the primary material (MHC) used throughout the conversion studies.

Powder XRD analyses confirmed the sample composition to be phase-pure. The SEM analyses revealed that MHC crystallites were rice-like, up to 2 - 3 μm

long and about 0.5 μm wide. Helical ridges were observed along the length of the grain surfaces, as shown in the inset of Figure 1. To test the range of temperatures which may cause MHC conversion, the same sample was analysed by DTA/TG (Figure 2).

DTA confirms the loss of water in two stages: the evaporation of free water is indicated at temperatures 50˚C to 150˚C, whereas, the higher temperature endotherm 180˚C - 280˚C is likely to be structural water loss. XRD analysis confirmed that the exothermic event at around 330˚C is related to MHC to calcite

re-crystallisation. TG analysis of the MHC sample showed the release of water happening in stages, gradually between 70˚C and 270˚C. The first step took place between 70˚C and 210˚C, where a gentle continuous slope of mass loss was observed. The second step, attributed to the crystal-bound water loss, is marked by a significantly steeper slope, between 210˚C and 270˚C. The sample was thermally stable between 270˚C and 650˚C, after which it underwent carbonate decomposition.

MHC prepared at 5 and 20˚C (2 h aging duration) were used to assess the influence of MHC synthesis temperature on the resulting aragonite, also, 24 and 48 h reactions (at 20˚C) were carried out in order to test whether prolonging the MHC crystallisation process affects the conversion. Although differently synthesised, these samples were left to transform to aragonite over 4 weeks in identical conditions to those described in the experimental section. XRD patterns in Figure 3 displays partial MHC-aragonite conversion which depicted the said transformation.

Figure 4 shows the morphologies of partial MHC, aragonite conversion products, formed from naturally aged MHC samples which were synthesised at 5˚C and 20˚C. Ageing took place in a moderately humid environment (43 RH%), at 20˚C over 4 weeks. It was observed that the 5˚C sample converted the fastest and the opposite was true for the 24 h, 20˚C sample.

XRD phase identification was followed by the morphological assessment. The MHC loose needles seen in the reference image (Figure 1) were replaced by flakes and spherical clusters. The increase in the latter coincided with the observed XRD changes. Spherulites differed from those typically described in literature: reporting authors often observe aragonite products exhibiting a narrow-needle (acicular) shape, however flower, cauliflower and flake-like crystals are also common [62] .

The resultant aragonite products were found to be between 20 - 30 μm, in the cases where the precursor was synthesised at 20˚C. In the experiments where MHC was synthesized at 5˚C, the formed spheres were visibly smaller in size, with diameters ranging between 10 - 15 μm. These findings indicated that colder MHC precipitation temperatures favoured smaller aragonite spheres.

Having noted that MHC formation temperature had a significant effect on crystallite morphology and conversion time, further studies were carried out in order to test the effects of the precursor reaction duration on the resulting aragonite morphology. Figure 5 overviews the transition of an MHC sample, which was aged at 24 and 48 h in mother solution prior to the curing.

Although both samples were cured for the same duration of time and at the same conditions (over 4 weeks in a moderately humid (43 RH%) environment), the difference in size between the spherical aragonite products was clearly visible (Figure 5). The aragonite spheres that formed from a sample aged for 24 h, were between 50 - 70 μm in diameter, more than twice as large as those converted from 2 h MHC (Figure 4). In the case of the 48 h aged samples, they were between 80 - 100 μm across, further signifying the influence of MHC ageing

duration prior to its natural conversion. The resulting large aragonite spheres also displayed a clear cohesion between the particles across the grain boundary. The grains contained both surface cracks, and deep fractures. The former were shallow microcracks on an otherwise smooth surface, and the latter, a uniaxial diametric crack up to 2 µm wide.

Mg²⁺ used in the kinetic stabilisation of MHC were replaced by Co²⁺ and Cu²⁺. To test whether this would lead to a change in morphology, a selected number of MHC samples were left to age in “moderate humidity” environments and analysed by XRD and SEM. Presented in Figure 6 are their conversion products.

Post-synthesis, the samples were held in a moderately humid (43 RH%) environment; over 4 weeks. The micrographs showed that the samples had a varying degree of conversion, but nevertheless, all exhibited a resemblance to the spherical morphology observed in the Mg-stabilised MHC conversion products. Both of the 2 h MHC samples displayed a significant amount of flaky, poorly defined flocculants, whereas, the cases with increased ageing duration led to well-formed spheres. Cu-MHC formed aragonite aggregates of crystals up to 30 - 40 μm, while Co-MHC converted to slightly sintered, large and fractured spheres. Comparable to the Mg²⁺ doped samples, only the largest grains showed evidence of cracking. Co-MHC conversion grains were up to 70 μm across, larger than any previously described in this study. Thus, it could be concluded that, Mg²⁺ was not responsible for the shape of aragonite products, as Cu²⁺ and Co²⁺ products exhibited similar habits.

When it came to natural MHC to aragonite conversion, none of the tested MHC synthesis variables (reaction temperature, time and stabilising ion) affected

the shape of the product (i.e. the spherical habit), however, they did affect the size of individual grains (Figure 7). It must be noted that none of the experiments described thus far led to calcite formation, with aragonite being the only observed conversion phase with exception to minor unidentified peaks in Mg-MHC (most likely hydromagnesite) and malachite in Cu-MHC.

Whilst still concentrating on the natural conversion aspect, it was necessary to investigate the influence of the environmental humidity on the persistence of the MHC powder precursor. Assuming that, in a commercial setting, MHC would be stored in a warehouse or laboratory storage with minor temperature variations, the humidity aspect may be less predictable and more difficult to manage. It is very unlikely that industrially produced MHC would be stored at strictly humidity-controlled facilities, therefore knowing the implications it may have on the product is very important.

Low humidity experiments displayed a successful case of well-preserved, loose MHC powder, even after 2 months of curing. There were no significant differences between the resultant MHC grains and the benchmark reference (Figure 1). Aragonite was absent in both XRD and SEM analyses, suggesting that low humidity environments significantly increased MHC persistence. The X-ray

diffractograms confirmed the purity of the specimen, the only observed difference between the reference sample and the aged specimen were noted in the peak intensity (Figure 1). A partial conversion to aragonite was observed in the MHC sample held in a low humid environment (0 RH%) for 6 months. The micrographs in Figure 8 display the changes observed at the onset of aragonite conversion.

The MHC rice-like crystals appeared in agglomerated clusters, with interconnecting web-like bridges in between them. Some pseudo-spherical shapes were also identified amongst the MHC agglomerates. MHC was additionally kept in a system with a 100 RH% where a more advanced MHC conversion was observed to have taken place (Figure 9).

The MHC sample which was kept in a very humid environment (100 RH%) for 2 months had fully converted and MHC could no longer be detected via XRD. The initial conversion signs appeared after 2 weeks, when grains began to form webbed clusters, similar to those seen in Figure 8. After two months, the recovered sample was a paste and no longer a powder: the solid was suspended in water, adsorbed from the environment, as 100 RH% was used. The aragonite spheres were between 15 - 20 µm across and fully connected by a sheet-like layer. One may note that the spheres exhibited a uniform size throughout, which is an advantage for calcium carbonate applications which often rely on a specific grain size.

The humidity in the air encouraged the MHC conversion to aragonite, confirming the hypothesis regarding the importance of water in MHC conversion to aragonite. It must be noted that the very humid environment (100 RH%) accelerated the conversion with a very high efficiency comparing to the moderate humidity environment (43 RH%) described in the previous subsections. XRD

identified aragonite and small quantities of hydromagnesite; considering that aragonite does not incorporate Mg into its lattice, it’s not surprising that a Mg phase is present. An unidentified peak may belong to MHC (Figure 10).

Assuming that aragonite does not incorporate Mg²⁺ into its crystal lattice as per previously published studies [33] , and that Mg²⁺ is indeed responsible for the

stabilisation of MHC [63] , then the loss of Mg²⁺ is a key factor in the conversion mechanism. Equation (3) shows MHC dissolution and Equation (4) aragonite recrystallisation.

Ca 1 − x Mg x CO 3 ⋅ H 2 O ⇌ Ca 2 + + Mg 2 + + CO 3 2 − + H 2 O (3)

Ca 2 + + Mg 2 + CO 3 2 − + H 2 O ⇌ CaCO 3 + Mg 2 + + H 2 O (4)

As seen so far, humidity favours the conversion, turning the dry and loose MHC powder into aragonite paste, it can be concluded that moisture plays a crucial part in both, the loss of Mg²⁺ during the dissolution of MHC and the consequent reprecipitation of aragonite. Following that thought, it was hypothesised that in order to design an accelerated conversion method and ensure the availability of water molecules even at elevated temperatures, the powder could be submerged in water baths. A series of experiments testing the hypothesis are discussed below.

Three sets of tests, in which the MHC product was immersed in water, were conducted in ovens set at different temperatures to observe their respective conversions to aragonite. The first experiment was carried out at 60˚C (Figure 11), followed by an identical protocol for temperatures 80˚C and 100˚C. An additional test was carried out at room temperature (20˚C), however, no conversion was observed over 9 hours.

The aragonite conversion tests conducted at 60˚C did not show any change in the phase composition within the first 2 h, a period which may be classed as the dissolution-dominated stage. After 3 h approximately 10% of the MHC content was observed to have converted to aragonite. After 6 h of heat treatment the

aragonite content overtook MHC as the major phase; both semi-quant XRD and peak intensity data displayed this change.

After 9 h in the heated water bath MHC had fully converted to aragonite. No additional phase or impurity was observed in the XRD pattern. The findings of this process proved to be an encouraging sign, as it decreased the conversion time from weeks to just hours.

AES, which was carried out on the solids recovered from the accelerated conversion experiments conducted at 60˚C, showed that Mg content within the powder decreased with increasing aragonite precipitation. Even within the first 2 hours a 2.7% decrease in Mg was observed, this was attributed to partial MHC dissolution. The final aragonite product contained around 20% less Mg than the MHC precursor. Although AES helps to identify patterns of decreasing Mg content, the data should not be interpreted as indication of the absolute aragonite elemental composition-aragonites are largely Mg-free, seldom containing as much as 1% MgCO₃ [33] , hence Mg detected by AES could have also been present elsewhere, e.g. an amorphous phase not detected by XRD.

Following the successful accelerated aragonite conversion, further experiments were conducted at 80˚C (Figure 12).

The aragonite conversion tests conducted at 80˚C showed a 10% drop in MHC content within the 1^st hour; further reduction in MHC over time followed a moreorless linear trend. Aragonite as the prevailing phase appeared in the experiments heated for over 3 hours. The final 30% of MHC converted at the highest rate. From the 5^th hour of reaction and onwards, there was no MHC detected by XRD. A 20˚C increase in the reaction temperature (from 60˚C to 80˚C), as shown, allowed for a significant reduction in the conversion time (up to 50%).

AES from the accelerated conversion experiments conducted at 80˚C showed

that Mg content within the powder decreased with increasing aragonite precipitation. The same pattern was observed for the conversion products at lower temperatures. During the 1st hour the powder had lost around 2.6% of the Mg. After 5 hours, the resulting aragonite product contained only 10% of the initial Mg.

To complement the above data, further experiments were conducted at an additional 20˚C increase (reaching 100˚C) (Figure 13).

A sudden 70% drop in MHC is observed within the first hour of the experiment, followed by a steady, near-linear decrease over the next hours. This strongly contrasts the linear trend, followed by a drop observed 80˚C experiment (Figure 6).

Although, in both 80˚C and 100˚C experiments, it took more than 4 h to completely dehydrate MHC and convert it to aragonite, the collected data suggested that the conversion kinetics were different in both cases. At 100˚C, the dissolution dominating period appeared to be significantly shorter than those seen in the 60˚C and 80˚C conversions. Due to carbonates’ unique characteristic of decreasing solubility with increasing temperature, it can be concluded that the aragonite reprecipitation stage was accelerated in higher temperatures and not the MHC dissolution.

The AES analysis suggested that after 1 h, as much as 3.1% percent of Mg was lost from the sample, more than observed at lower temperatures for the given time. Furthermore, approx. 1% of Mg was lost to solution each hour with increasing aragonite precipitation, the final product lost around 7% of the Mg from the precursor.

Aragonite formed using the accelerated conversion method had a very different morphology to the naturally converted spherical aragonites (Figure 14). The accelerated conversion grains took the shape of elliptical needles (between 15 -

20 µm long and up to 5 µm wide) and clusters thereof. The conversion temperature did not appear to have any significant effect on the morphology: the selected images (Figure 14) display partially converted MHC and aragonite mixtures resulting from the 80˚C water bath over 3 hours, however, all of the aragonite products which were converted in heated water baths at temperatures of 60˚C - 100˚C resulted in similar grain sizes and morphologies (Figure 15).

Most of the aragonite crystallites were of uniform size, between 20 - 30 µm

long and up to 15 µm wide. Notably, the aragonite crystallites synthesised at 60˚C were slightly larger (~2 μm), elliptical needle shape, as opposed to crosses and stars as seen in the other temperature experiments. Although the XRD of the 60˚C conversion product suggested a phase-pure aragonite, the small, flaky grains could suggest a secondary phase, i.e. hydromagnesite or small quantities of poorly crystalline MHC precursor. Figure 16 shows that accelerated conversion produces aragonite of higher yield relative to natural conversion.

Most of the aragonites exhibited a single spindle-resembling shape, however, there was an obvious progression to the conjoining pairs and clusters. These were either forming an “X” shape by joining in the middle or joining at the ends to form a “V” shape and mostly abundant in the higher temperature studies. Some of the crystallites had a spherical, web-like growth surrounding the crystal; the web was either very loose (Figure 17, early rose), with visible porosity, or very closely packed, resembling a flower (Figure 17, rose). Often there were spindles and their clusters attached to the spherical formations. Figure 17 displays some of the more interesting interaction between the two morphologies.

There was an intermittent abundance of hierarchical microspheres, exclusive to the fully converted samples at the higher temperature range (80˚C - 100˚C). The particle shape somewhat resembled the spheres described in the natural conversion studies, however, on closer inspection proved to be very different: much more porous, and significantly larger than aragonite spheres described in the natural conversion studies (Figure 18).

Despite somewhat similar appearance it was evident that the sphere formation had taken place via different mechanism and the availability of water played the

key role. In the natural conversion experiments, the powders could adsorb water due to the high relative atmospheric humidity. However, due to the absence of a suitable medium to disperse into, MHC dissolution may have caused locally saturated Mg environment around the reprecipitating crystallites. It could have been the reason for the formation a Mg-rich skin-like layer identified as hydromagnesite via XRD.

The XRD-amorphous spheres seen in the accelerated conversion studies (Figure 17) appeared as a visually distinct secondary phase. It exhibited morphological characteristics commonly associated with Mg(OH)₂ or Mg-carbonates e.g. hydromagnesite, dypingite [64] [65] [66] [67] . The locally saturated environments did not appear due to the significantly greater availability of water,

which acted as a medium for MHC dissolution and ion dispersion.

EDS elemental mapping showed that the largest Mg distribution coincided with the spheres, whilst the elliptical needle-clusters showed the highest Ca concentrations. These findings confirmed that Mg from MHC was lost during its conversion to aragonite, and eventually precipitated as Mg-minerals. Elemental mapping (Figure 19) validated the hypothesis that needles were a polymorph of very low-Mg CaCO₃ and spheres could potentially be identified as magnesite (MgCO₃).

Having evaluated the DTA data and the temperature range at which MHC lost its structural water, 10 samples, from the same MHC batch, weighing 1 gram each, were gently heated over a range of temperatures (150˚C - 250˚C) in an oven.

The structure of MHC broke down gradually during the second stage of water loss, as identified in the DTA (Figure 20). With increasing temperature MHC converted to calcite, the XRD patterns of the heating products are seen in Figure 21.

MHC, calcite transition patterns pictured in the figure above highlighted that the conversion is gradual, first decreasing MHC peaks and resulting in largely amorphous phase, anhydrous ACC (between temperatures 200˚C - 230˚C). The anhydrous ACC was rehydrated in order to check for the potential regenerative properties, which were successful according to XRD (Figure 22). The steps taken involved the immersion of the dehydrated MHC powder into a vessel filled with double-distilled water, before mixing and filtration of the suspension, followed by acetone drying and XRD analysis. Two examples of MHC regeneration are presented: 1) where MHC is partially decomposed with minor calcite conversion at 190˚C; 2) advanced MHC decomposition and conversion to calcite at 250˚C. In both cases it was observed that MHC regenerates once the dehydrated powder is immersed in water. The dehydration of MHC could be compared to the conversion of gypsum (CaSO₄·2H₂O) into calcium sulfate hemihydrate (CaSO₄·0.5H₂O) and anhydrites (γ-CaSO₄ and β-CaSO₄) [68] . A loss of 3/2 molecules of water transforms gypsum into hemihydrate; whereas, in the case of MHC, the resulting amorphous CaCO₃ is assumed to be anhydrous due to the DTA/TG indicated water loss described in the previous sections.