The Conversion of Wollastonite to CaCO3 Considering Its Use for CCS Application as Cementitious Material

Featured Application

Building materials and CCS.

Abstract

The reaction of wollastonite (CaSiO₃) with CO₂ in the presence of aqueous solutions (H₂O) and varied temperature conditions (296 K, 323 K, and 333 K) was investigated. The educts (CaSiO₃) and the products (CaCO₃ and SiO₂) were analyzed by scanning electron microscopy (SEM), powder X-ray diffraction (PXRD), and differential scanning calorimetry with thermogravimetry coupled with a mass spectrometer and infrared spectrometer (DSC-TG/MS/IR). The reaction rate increased significantly at higher temperatures and seemed less dependent on applied pressure. It could be shown that under the defined conditions wollastonite can be applied as a cementitious material for sealing wells considering CCS applications, because after 24 h the degree of conversion from CaSiO₃ to CaCO₃ at 333 K was very high (>90%). As anticipated, the most likely application of wollastonite as a cementitious material in CCS would be for sealing the well after injection of CO₂ in the reservoir.

1. Introduction

To reduce the amount of CO₂ in the atmosphere (currently 403 ppm [1]) several possibilities need to be considered. One possible technology being investigated and discussed is carbon capture and storage (CCS). The goal of CCS is reducing CO₂ emissions from burning of fossil fuel by storing considerable amounts in suitable deep geologic reservoirs. The reservoirs considered for injection could be on- or offshore formations which could be in part already former exploited gas fields [2]. After injection, the sealing of wells in deep geologic formations being considered for CCS must be sustainable. Therefore, the choice of a tight and durable cement, used for sealing, is critical. The material has to resist the extreme conditions at the point of injection considering pressure, temperature, pH, and the composition of aggressive fluids (chloride rich brines, supercritical (sc) CO₂ accompanied with NO₂ and SO₂). The high amount of CO₂ and the fluids will have a considerable impact on cement properties. Hence, the term carbonation is occupied in the literature as a post hydration reaction of cementitious materials [3,4]. Ordinary Portland cements (OPC) cannot be used under these conditions of well boring due to their instability. The current study describes carbonation as a curing method to obtain carbonates as reaction products. In the present study, a literature review describing the influence of CO₂ on conventional hydrated cements as an alteration reaction is presented. Studies on the reaction of natural wollastonite with CO₂ as curing method under different pressures and temperatures were investigated.

Although natural deposits are limited, wollastonite can be fabricated using a conventional cement rotary kiln production process. Wollastonite has a Ca:Si ratio of 1:1 and therefore during production a better CO₂ balance than conventional cements with Ca:Si ratio 2:1 and 3:1 (

Table 1. Raw meal CO₂-output of selected cement clinker phases. [5].

Phase	g of CO2 per g Phase	Raw Materials
C3S Alite	0.58	3 CaCO3/SiO2–OPC
C2S Belite	0.51	2 CaCO3/SiO2
CS Wollastonite	0.38	CaCO3/SiO2
C3A	0.49	CaCO3/Al2O3
CA	0.28	CaCO3/Al2O3
CA2	0.17	CaCO3/Al2O3
C2F	0.32	CaCO3/Fe2O3
C2A	0.41	CaCO3/Al2O3
C4AF	0.36	CaCO3/Al2O3/Fe2O3
C3A3Cs Sulfoaluminate	0.22	CaCO3/Al2O3/CaSO4

, [5]).

Additionally, the CO₂ will be consumed during the carbonation reaction thereby improving the CO₂ balance by forming carbonates.

The interaction of wollastonite and CO₂ [6] was investigated within the joint BMWi (German Federal Ministry for Economic Affairs and Energy) research project CLUSTER [7]. The project CLUSTER focusses on two main topics:

(I)

Potential impacts of temporal variations in CO₂ stream composition and mass flow that occur in larger pipeline networks on transport, injection, and geological storage of CO₂.

(I)

(II) Flexibility of individual steps and the whole CCS process chain in response to this temporal variability [8].

In nature, wollastonite is formed during metamorphism. The present study considers the back reaction by carbonation of wollastonite forming CaCO₃ and SiO₂ as a base [6]. The carbonation reaction of wollastonite is described as extremely slow in absence of water at low temperatures [9]. Yet at 333 K and ambient pressure the $\overline{C}CS$ reaction is fast (several hours up to one day) in presence of water [9]. Microstructure aspects of the carbonation were investigated by Villani et al. [10] and Sahu and DeCristofaro [11]. Both studies showed a carbonation from the surface to the center of the wollastonite grains, similar to the shrinking core model [12]. Wollastonite reacts in the presence of CO₂ in aqueous conditions to calcite and amorphous silicon oxide:

$${\mathrm{CaSiO}}_3+{\mathrm{CO}}_2 \stackrel{{\mathrm{H}}_2\mathrm{O}}{\to }{\mathrm{CaCO}}_3+{\mathrm{SiO}}_2(\mathrm{amorphous})$$

(1)

The chosen temperature (333 K) defined by the project CLUSTER [8] equates to the conditions in deep geological formations envisaged for CCS. Gartner and Hirao [9] and Longo et al. [13] found that the carbonation reaction strongly depends on the presence of water. Our experiments were performed under aqueous conditions, because aqueous solutions are expected to be present at the point of injection in wellbores.

1.1. Interaction of CO₂ and Conventional Hydrated Cements

Carbonation (i.e., formation of CaCO₃) in hydrated cements causes a decrease of the pH-value from approximately 13 to approximately 8. The calcium carbonate ($C\overline{C}$) mostly appears in form of calcite but also as vaterite and aragonite [3,4]. Carbonation is responsible for volume changes and formation of carbonates, hemicarbonates and monocarbonates, which leads to changes in porosity and permeability, and the formation of a fracture network. Dissolution of phases also may occur [3,4]. Considering the chemical and mechanical stability for construction due to changes of the porosity, this process is very critical [3,4].

Therefore, extensive research on the influence of CO₂ on OPC and wellbore cement was performed. Shi et al. [14] investigated the carbonation of white portland cement (wPc, CEM I 52,5 N) from Aalborg Portland A/S, Denmark, metakaolin (MK) and limestone (LS) mortars. Their mortars were hydrated for 91 days and exposed to a controlled atmosphere with one percent CO₂ at 293 K and 57% relative humidity (RH) for 280 days. They found that the Portland cement had the highest CO₂ binding capacity followed by limestone and metakaolin. Abdoulghafour et al. [15] and Luquot et al. [16] performed flow-through experiments with CO₂ and brine on fractured Portland cement. Their experiments were carried out at 333 K and 10 MPa for more than 100 h. They found that the leakage potential is controlled by the aperture of the fracture. Ozyurtkan and Radonjic [17] investigated the effect of CO₂ rich brine (two percent salt solution) on artificially fractured well-cement (class H) in a flow-through experiment over 30 and 100 days. After 100 days, a secondary fracture network was developed which was not present after 30 days. The secondary fractures caused higher leaching rates. The obtained data by Ozyurtkan and Radonjic [17] indicates a negative effect on porosity of cement under dynamic conditions which can be dependent on time. Yalcinkaya et al. [18], performed an experimental study on the dynamic alteration process of class-H wellbore cement caused by acidic brine. Their experiments confirmed leaching of Ca²⁺ from the reacted cement. Carey et al. [19] investigated the wellbore integrity and CO₂-brine flow along the casing-cement. They performed flow-through experiments at 313 K and 14 MPa pore pressure for 394 h. The brine was a 50:50 mixture of scCO₂ and 30,000 ppm NaCl-rich brine. They were able to confirm that degradation of Portland cement is slow. Barlet-Gouédard et al. [20] compared the behavior of a Portland cement with a Schlumberger CO₂ resistant cement (SCRC) with expanding agent under CO₂ geological storage environment (363 K, 28 MPa). They used brine containing 220 g/L NaCl. They focused on the expansion of the cement because fractures caused by the expansion are one possible occurrence of leakages. The simulated injection scenario presented by Barlet-Gouédard et al. [20] showed, that conventional Portland cement does not fulfill the requirements for the well integrity. Wigand et al. [21] performed flow-through experiments which simulated the diffusion of brine (1.65 M) and scCO₂ from the interface between wellbore cement and caprock into a fracture-bearing Portland cement. The experiments were performed at 327 K with a pore pressure of 19.9 MPa and a confining pressure of 26.2 MPa for 174 h. They showed that the carbonation of the wellbore cement at first caused closing and blocking of pores and fractures but opened other fractures in the following process, which increased the permeability. The chemical interaction along the cement–sandstone interface with scCO₂ was investigated by Nakano et al. [22]. They prepared a well composite sample consisting of casing, Portland cement and sandstone. The experiments were performed under 10 MPa and 323 K with a 0.5 M NaCl-brine. Their results indicate that the formation of carbonation layers act as a barrier to further CO₂-attack. Zhang et al. [23,24] investigated the reaction of two different pozzolan-amended (Class F fly ash, with ~70% silica, alumina and Fe₂O₃, ~5% sulfate, and less than 20% CaO) wellbore cements (API RP 10B; 35 vol % pozzolan and 65 vol % pozzolan) being exposed to CO₂ and H₂S gas mixtures and brine (one percent NaCl) at 323 K and 15.1 MPa for 2.5, 9, 28 and 90 days. The samples of the pozzolan-amended wellbore cement with 35 vol % pozzolan showed higher resistance against H₂S. Mito et al. [25] and Wolterbeek et al. [26] investigated the influence of CO₂-induced reactions on the mechanical behavior of fractured wellbore cement (class-G Portland cement). They found that the reaction with CO₂ did not produce further geomechanical weakening. In fact, samples with six weeks reaction time recovered peak-strength (83%) and increased frictional strength (15–40%). Kutchko et al. [27] investigated the reaction of hydrated well cement (class-H) with H₂S-CO₂ (21 mol % H₂S) acid-gas for 28 days at 323 K and 15 MPa. The H₂S added to the experimental setup resulted in significant formation of ettringite (interior region of the cement) and pyrite (carbonated rim of the cement). Although secondarily formed ettringite is known to cause cracking and strength loss, a mechanical damage caused by formation of ettringite was not found. Even the long-term cement corrosion in chloride-rich solutions was investigated by Bube et al. [27]. They performed leaching experiments and thermodynamic simulations on cement (OPC; CEM I 32,5 R; w/c ratio of 0.4) with brine (4.3 mol Mg²⁺, 0.41 mol Na⁺, 0.54 mol K⁺, 8.9 mol Cl⁻, and 0.23 mol SO₄²⁻) for 200 days. Comparing the results of laboratory and full-scale experiments, they were able to establish a link between the results of short-term laboratory experiments to long-term full-scale experiments, using thermodynamic equilibrium calculations.

The above research shows that carbonation can negatively impact the sealing qualities of OPC wellbore cements. Therefore, alternative materials are required to ensure wellbore integrity during the subsurface storage of CO₂. One such alternative is a cement based on wollastonite (CaSiO₃), which will create a seal through a carbonation reaction. In this study, experimental work addressing wollastonite carbonation kinetics under in-situ conditions relevant for subsurface CO₂ storage in aquifers or depleted reservoirs is presented.

1.2. Interaction of Wollastonite and CO₂

CO₂ has an impact on the stability of conventional hydrated cements. Therefore, the use of a cement, based on calcium silicate (CS) curing in the presence of CO₂ ($\overline{C}$), seems very promising [28,29,30]. Yet knowledge considering this approach is rather scarce.

Ashraf et al. [31] investigated the effects of high temperature (723–783 K) on calcium silicate-based cement (CSC), OPCs and OPCs with 20% fly ash. During the experiments, the OPCs showed a reduction of stiffness (60%), an increase of porosity (30%) and a loss of mass (13–15%) higher than the CSC (stiffness: 6%; porosity: 20%; mass loss: 3%). Huijgen et al. [32] investigated the mechanisms of aqueous wollastonite carbonation, using three different grain sizes (<38 µm, <106 µm and <500 µm) at short time intervals (5–60 min), a temperature range from 298 to 498 K and a pressure variation up to 4 MPa. A maximum conversion of 70% in 15 min at 473 K, 2 MPa CO₂ partial pressure and with a particle size of <38 µm was achieved. They were able to show that the aqueous carbonation mechanisms of wollastonite, steel slag and olivine are generally similar, although they showed that wollastonite carbonation was rapid, relative to the carbonation of Mg-silicates. Longo et al. [13] showed that water has a huge impact on the carbonation reaction of wollastonite. Without water the wollastonite sample carbonated on the surface rather fast but after a carbonate monolayer was formed, the reaction comes to a standstill. The presence of water completely changes the reaction and caused ongoing carbonation. Daval et al. [33] investigated the carbonation reactions of wollastonite at conditions relevant to geologic CO₂ sequestration in subsurface environments. They performed experiments in batch reactors at 363 K and 25 MPa. After carbonation, they found two reaction products: Calcite and a silica-rich phase. No other carbonate polymorph was found. Min et al. [34] performed carbonation experiments on wollastonite at typical geological storage conditions (333 K and 10 MPa). They investigated the microstructure of the carbonated wollastonite and found that after 20 h the reaction was controlled by the diffusion of water-bearing scCO₂ across the product layer on wollastonite surfaces. A layer of amorphous SiO₂ covered the surface of the wollastonite and functioned as a barrier.

2. Materials and Methods

To determine carbonation rates, several series of experiments were performed and different analytic techniques were applied. Educts and products were analyzed with powder X-ray diffraction (PXRD) and differential scanning calorimetry with thermogravimetry (DSC-TG) coupled with mass spectrometer (MS) and infrared spectrometer (IR). The chemical composition of the educts was analyzed with wave length dispersive X-ray fluorescence (XRF).

2.1. Materials

For the experiments, steel reactors with teflon liners from the company Büchi AG were used. According to the investigations of Huijgens et al. [32] (best conversion rate of 70% after 15 min at 473 K and 2 MPa) a sample material with a grain size <38 µm was chosen and the experiments were performed with 2 MPa. In this work, experiments at lower temperatures were performed because this work is connected to a specific CCS process, in which temperatures do not exceed 333 K. The used raw material (natural CaSiO₃, calcium silicate, meta, reagent grade, <20 µm powder) was supplied by Alfa Aeser. The chemical composition of the used raw material was analyzed with an XRF spectrometer (Siemens SRS 3000, Table 1). As expected the main elements were Si and Ca, but in addition the raw material contained small amounts of Fe, Al, Mg, Mn, Ti, K, and Sr (

Table 2. Chemical composition of the used raw material determined by Siemens SRS 3000.

Element (Oxides)	Content (wt %; Normalized)
SiO2	50.55
CaO	45.95
Fe2O3	0.52
Al2O3	0.52
MgO	0.17
MnO	0.13
TiO2	0.05
K2O	0.03
SrO	0.01
LOI 1	2.09
Total	100.02

¹ LOI: loss of ignition at 1323 K.

), which is due to the natural origin of this material.

The pressure in the reactors was adjusted using an ISCO-pump (Teledyne ISCO, Model 500D Syring pump). CO₂ was supplied by Westfalen AG and had a purity of 4.5 (99.995%).

2.2. Analytical Methods

The mineralogical composition of the raw materials was measured by PXRD (Panalytical X’PERT³ Powder with Pixcel 1D detector). The PXRD measurements were carried out with 45 kV and 40 mA. As apertures 0.04 rad soller slit and fixed slits (0.125° and 0.25°) were used. Diffractograms were recorded from 5° to 100° 2Θ with 49.725 s per step and stepsize: 0.013° 2Θ. For qualitative investigation the Highscore Plus Suite [35] and for quantitative phase analysis the Rietveld method [36] with Profex-BGMN [37,38] was used. To determine the amorphous content, 10 wt % rutile (Kronos 2900–TiO₂; Rutile, supplied by KRONOS TITAN GmbH, Leverkusen, Germany) was added as internal standard. Chemical composition was analyzed with wave length dispersive X-ray fluorescence (SRS 3000; Siemens, Karlsruhe, Germany; Rh-tube; vacuum: 200 mbar; measurement, 30 s for each line). Therefore, melting tablets with 1 g sample material and 8 g Di-Lithium Tetraborate (FX-X100-2) supplied by FLUXANA (Bedburg-Hau, Germany) were prepared.

To investigate the degree of carbonation depending on pressure (p) and temperature (T), the contents of wollastonite accompanied by the increase of calcium carbonate and increase of the amorphous phase contents different approaches were applied to identify and quantify educts and products. For quantitative phase analysis, PXRD with the Rietveld method [36] and differential scanning calorimetry with thermogravimetry (DSC-TG; STA 449 F3 Jupiter, Netzsch, Selb, Germany) coupled with mass spectrometer (MS; QMS 403D Aeolos Netzsch; atom mass units 1 to 50 u were recorded) and infrared spectroscopy (IR; Tensor II with TGA-IR extension compartment from Bruker, Karlsruhe, Germany) were used. The thermoanalytical measurements were carried out from ambient temperature (296 K) to 1273 K (1000 °C) with a heating rate of 10 K per minute. For the DSC-TG measurements ~15 to ~25 mg sample material per measurement were used.

With DSC-TG/IR/MS the amount of X-ray amorphous content of CaCO₃ in the product could be analyzed as well.

The content of carbonate was calculated from the mass loss associated with CO₂:

$${\mathrm{CaCO}}_3[\mathrm{wt} \%]=\frac{\mathrm{mass} \mathrm{loss} {\mathrm{CO}}_2 [\mathrm{wt} \%]}{(\frac{\mathrm{mass} {\mathrm{CO}}_2 [\mathrm{g}/\mathrm{mol}]}{\mathrm{mass} {\mathrm{CaCO}}_3 [\mathrm{g}/\mathrm{mol}]})}$$

(2)

Microstructure of educts and products were characterized by SEM (JEOL JSM 6300 combined with X-Flash 5010 detector from Bruker; secondary electron pictures were taken with 15 keV; samples were sputtered with gold).

2.3. Experimental Setup

Carbonation of wollastonite at three different temperatures (296 K, 323 K, and 333 K) was investigated. As starting conditions 1 g sample material was mixed with 5 mL H₂O in the steel reactor. 20 mL CO₂ (5.5 MPa) were pumped by the ISCO-pump into the steel reactor, to obtain a pressure of 2 MPa inside the reactor. The temperature in steel reactors was adjusted by placing them in a drying chamber. For each temperature (296 K, 323 K, and 333 K) several experiments with different time intervals and constant pressure (2 MPa) were performed (

Table 3. Overview of performed series of experiments at constant pressure and varied time and temperature.

Temperature (K)	Pressure (MPa)	Reaction Time (h)
296	2	0, 24, 48, 72, 96, 120, 144
323	2	0, 6, 18, 24, 48
333	2	0, 3, 6, 12, 18, 24

).

In addition, a series of experiments with constant time and temperature but varied pressure were performed. The applied pressure ranged from 0.2 MPa to 5 MPa (

Table 4. Overview of performed series of experiments at varied pressure and constant time and temperature.

Temperature (K)	Pressure (MPa)	Reaction Time (h)
333	0.2, 0.5, 1.0, 1.5, 2.0, 2.6, 4.0, 5.0	6

).

To determine carbonation and reactions rates educts and products were analyzed by PXRD, DSC-TG/MS/IR and SEM.

3. Results and Discussion

The results of the quantification of the educts and products at different temperatures and pressures showed that temperature has a higher impact on the conversion of wollastonite to CaCO₃ than applied pressure.

3.1. Microstructure (SEM)

For the microscopic analysis of the morphology of educts and products SEM was used. The raw material exhibited wollastonite grains, showing elongated crystals of 5–6 µm × 10–12 µm (Figure 1, white arrows).

The carbonated wollastonite still showed small and elongated (2.5 × 5 µm) wollastonite (Wo) crystals, covered with carbonate crystals (Figure 2a, white arrow). Yet aggregates of carbonate crystals (Cb, Figure 2b, white arrow) were observed in the micrograph shown in Figure 2 (white arrow).

3.2. PXRD Analysis

To quantify the carbonation reaction of natural wollastonite (

Table 5. Composition of the used raw material and carbonated wollastonite (24 h, 333 K), determined by powder X-ray diffraction (PXRD) and Rietveld refinement [38] (Rwp_{Raw material}: 7%, Rwp_{carbonated wollastonite}: 7%).

Phase	Raw Material	Carbonated Wollastonite
	Content (wt %)	Content (wt %)
Wollastonite	90 ± 1	4 ± 1
Calcite	5 ± 1	29 ± 1
Aragonite	1 ± 1	21 ± 1
Amorphous	4 ± 1	45 ± 1

), samples were measured with PXRD before and after the CO₂ treatment. The results of the PXRD-analysis showed three crystalline phases (wollastonite, calcite, and aragonite) and an amorphous content, determined by adding 10 wt % of rutile as internal standard (Table 5 and Figure 3 and Figure 4) in the starting material and the reaction products as well.

In Figure 4, the wollastonite with the fastest and most advanced conversion to carbonates (24 h, 333 K, 2 MPa) is shown. The results at the quantification of all carbonation reactions of wollastonite at 2 MPa is shown in part 3.4 of this study.

3.3. DSC-TG/MS/IR Measurements

DSC-TG/MS/IR measurement of the raw material and the carbonated wollastonite were carried out to determine the entire content of water and carbonate in the material because X-ray amorphous carbonate could not be determined with PXRD, when additional amorphous phases (SiO₂) are also present in the mixtures.

DSC-TG/MS/IR measurements as shown in Figure 5 were performed with the raw material. During the measurement, the sample lost 2.1% of its weight. 1.7 wt % were associated with CO₂, which results in original content of 3.9 wt % carbonate in the sample (Figure 5, Equation (2)). At the beginning of the experiment only H₂O were detected by mass spectrometry. The detected mass flow of H₂O (Figure 5, green line) decreased from 400 K to 800 K. From 800 K to 1050 K only CO₂ were measured by mass spectrometry. The mass flow of CO₂ reached its maximum at approximately 1000 K (Figure 5, magenta line). The measured weight loss (Figure 5, black line) decreased slowly from 400 K to 850 K. Starting with the increased mass flow of CO₂ the weight loss increased significantly (850 K to 1050 K). At temperatures higher than 1050 K, the mass of the sample did not change. The DSC-signal (Figure 5, blue line) showed amplitudes at 400 K and 1000 K, which were associated with the detected peaks of the mass flows of H₂O (400 K) and CO₂ (1000 K).

The results of the Rietveld analysis and the DSC-TG measurements of the raw material for the content of carbonate were in good agreement (Rietveld: 5%; DSC-TG: 4%).

DSC-TG/MS/IR measurements on the carbonated wollastonite, shown exemplarily in Figure 6, were performed. During the measurement, the sample lost 25.4% of its weight. 23.9% were associated with CO₂, which resulted in original content of 54.8 wt % carbonate in the sample (Figure 6, Equation (2)). At approximately 400 K the mass spectrometer detected a peak of the mass flow of H₂O (Figure 6, green line). The mass flow decreased rapidly until 500 K was reached. From 500 K to 750 K the mass flow of H₂O rose again and held a plateau until 950 K. The IR analysis did also confirm the investigation of the releasing behavior of H₂O from CaCO₃. The mass flow of H₂O decreased significantly at temperatures higher than 950 K. The detection of the mass flow of CO₂ (Figure 6, magenta line) started at 800 K and increased slightly until 950 K was reached, then it formed a sharp peak until 1100 K was reached. At temperatures higher than 1100 K, the mass flow of CO₂ was nearly zero. The measured weight loss (Figure 6, black line) decreased slowly until 900 K. Starting with the increase of the mass flow of CO₂ the weight loss increased significantly (900 K to 1100 K). At temperatures higher than 1100 K, the mass of the sample did not change. The DSC-signal (Figure 6, blue line) showed a negative (endothermic) amplitude at 1100 K which was associated with the detected peak of the mass flow of CO₂. At 1150 K the DSC-signal showed a positive amplitude, which could be ascribed to the formation of Ca₂SiO₄. PXRD-measurements were performed on the same sample which showed small contents of C₂S (larnite, i.e., Ca₂SiO₄).

The results of the Rietveld method and the DSC-TG measurements of the carbonated wollastonite for the content of carbonate were again in good agreement (Rietveld: 50%; DSC-TG: 55%).

3.4. Discussion of the PXRD and DSC-TG/MS Results

The results of the PXRD-measurements with the Rietveld method and the DSC-TG/MS/IR-measurements were collected and plotted in three diagrams (Figure 7, Figure 8 and Figure 9). The quantity of the phases wollastonite (black squares), amorphous (blue triangles), calcite (magenta circles) and aragonite (green circles) were determined by the Rietveld method. The red circles symbolized the sum of calcite and aragonite and therefore the content of carbonate in the sample detected by PXRD. The amount of CaCO₃ determined by TG measurements were shown by cyan circles.

In Figure 7, the carbonation of wollastonite at 296 K is shown with seven different reaction times. To show the good reproducibility several experiments were performed two times (Figure 7). At 296 K, the reaction was not completed after 144 h. The content of calcite (magenta circles), the amount of CaCO₃ determined by TG measurements (cyan circles) and of the amorphous content (blue triangles) in the sample increased exponentially while the content of wollastonite (black squares) decreased similarly. No further aragonite (green circles) was formed (Figure 7).

At 323 K, a high degree of conversion (>90%) was already achieved after 48 h (Figure 8). The content of calcite (magenta circles), aragonite (green circles), the amount of CaCO₃ determined by TG measurements (cyan circles) and the amorphous content (blue triangles) increased exponentially, while the content of wollastonite (black squares) decreased exponentially in an analog manner. At 24 h, the contents of calcite and aragonite almost reached the same amount (~20 wt %), while at the beginning of the experiment the content of calcite was higher (Figure 8, grey circles marked by arrows).

At 333 K, the conversion was nearly completed already after 24 h (Figure 9). The contents of calcite (magenta circles), aragonite (green circles) and the total amount of CaCO₃ determined by TG measurements (cyan circles) and the amorphous content (blue triangles) increased exponentially, while the content of wollastonite (black squares) decreased exponentially. At 12 h the contents of calcite and aragonite almost reached the same amount (~20 wt %) whereas at the beginning (6 h and 18 h) of the experiment the content of calcite was higher (Figure 9, grey circles marked by arrows).

At each investigated temperature (296 K, 323 K, 333 K), an exponentially decrease of wollastonite was observed. This was also true for the increase of calcite and the amorphous content vice versa. At higher temperatures (323 K and 333 K), aragonite exhibited also exponential increase (Figure 8 and Figure 9).

An additional experiment was performed, in which the temperature (333 K) and time (6 h) were held constant and only the pressure varied. The results showed that the reaction was mostly independent of pressure (Figure 10).

At low pressure (0.2 MPa) the decrease of wollastonite was slower and newly formed aragonite was not observed. At higher pressures (>1 MPa) the carbonation of natural wollastonite showed no further dependence of the applied pressure (Figure 10). From 1 MPa to 5 MPa no differences considering the rate of wollastonite transformation were observed. However the formation of aragonite was dependent on pressure. From 0.2 MPa to 2 MPa the formation of aragonite increased, from 2 MPa to 4 MPa it was constant and from 4 MPa to 5 MPa the formation of aragonite decreased. Figure 11 shows this observation as normalized amount of the wollastonite conversion.

Considering the conversion of wollastonite to CaCO₃ in the presence of water the reaction rate of the carbonation depended strongly on the temperature, at higher temperatures the reaction rate was significantly higher (Figure 7, Figure 8 and Figure 9).

A first order reaction was assumed because the reaction was only dependent on wollastonite. The amount of CO₂ was one order of magnitude higher than the amount of wollastonite. This assumption could be supported by plotting ln (C) vs time (Figure 12) from which a linear dependency after a conversion of ~90% could be derived.

In Figure 13 the wollastonite hkl reflection (100) at 12 ° 2Θ decreased from 296 K to 333 K.

Considering the conversion rates from wollastonite to calcite the rate constants and activation energies could be estimated (

Table 6. Rate constants k and activation energies E_A of CaSiO₃/CaCO₃ conversion.

Activation Energies	Rate Constants	kJ/mol
EA (323–296 K)	−3.79 × 10−6 ± 2.36 × 10−7	63.88 ± 0.2
EA (333–296 K)	−3.32 × 10−5 ± 1.87 × 10−6	58.88 ± 0.6
EA (333–323 K)	−5.42 × 10−5 ± 1.95 × 10−6	43.70 ± 2.0

).

The carbonation reaction of wollastonite starts at the surface of a grain and moves towards the center [10,11,34]. The reaction rate will decrease because of the growing layer thickness of CaCO₃ between unreacted CaSiO₃ and CO₂. Consequently, low amounts (5% to 10%) of material will remain unreacted within experimental periods. At 323 K, the reaction rate is three times faster, whereas at 333 K the reaction rate increased by a factor of six (Figure 14). The values of the activation energy E_A for the carbonation reaction ranged from 44 to 64 kJ·mol⁻¹ (Table 6) depending on ∆T, with the latter being in a good agreement with the results reported in Brady [39] (~72 kJ·mol⁻¹).

The carbonation rate of wollastonite at different temperatures (when ~10% of wollastonite were still present) showed a non-linear behavior (exponential decrease) of reaction time (Figure 14).

The p-T conditions in the steel reactor were thermodynamically not suitable [40] for the formation of aragonite. Although metastable aragonite formation can be stabilized in presence of Sr [41], the content of Sr in the material was very low (0.01 wt %). Despite in preliminary performed experiments no aragonite was formed when the carbonate was previously removed from the raw material. Therefore, it was assumed that the small content of aragonite in the raw material functioned as a seed crystal for further metastable aragonite growth.

4. Conclusions

Experiments considering the conversion of CaSiO₃ to CaCO₃ in aqueous conditions were performed at different temperatures and pressures. The kinetics of this reaction was strongly dependent on temperature (Figure 13). Pressure seemed not to be critical for reaction kinetics (Figure 10 and Figure 11). The degree of conversion, i.e., the amount of reaction products (amorphous and crystalline), could be determined by XRD and coupled DSC-TG/MS/IR analysis. Calcite, aragonite and amorphous CaCO₃ and amorphous SiO₂ were observed. Depending on the temperature a nearly complete reaction was achieved within the experimental periods of 1–144 h. Activation energies E_A could be determined and were in better agreement with the values reported by Brady [39] compared to those given by Huijgen (22 kJ·mol⁻¹ at <38 µm, 20 kJ·mol⁻¹ at <106 µm, 16 kJ·mol⁻¹ at <500 µm) [32].

At the defined conditions (333 K, 2 MPa) relevant for the point of injection of CO₂, the carbonation reaction of wollastonite was already completed after 24 h (Figure 9). At lower temperatures, the carbonation reaction was rather slow (144 h at 296 K, Figure 7 and 48 h at 323 K, Figure 8). Hence, it could be proven that the use of wollastonite as a cementitious material suitable for sealing the borehole in CCS is applicable.

To characterize mineral association and their stability as well, the reaction at the interfaces between rocks and casing materials, more experiments at selected pressure, temperature and composition of formation waters will be performed. The long-term stability is still under investigation. As anticipated, the most likely application of wollastonite as cement in CCS would be for sealing the well after injection of CO₂ in the reservoir. In that case, the wollastonite will be mixed with water and pumped into the well like conventional cement, mortar or concrete. The CO₂ stored in the reservoir will cause the hardening and therefore the sealing of the borehole.