**3. Results and Discussion**

#### *3.1. Proposed Methodological Sequence to Evaluate Effectiveness of Construction and Demolition Waste for CCS*

A first outcome measure to take into consideration to provide the optimal construction residue is to have an important content of calcium oxide and/or magnesium as reactive compounds, since these elements are necessary for precipitation in the form of the carbonate, although the carbonation of other alkali metals is possible, as shown in the following equations [25]:

$$\rm{MOO} + \rm{CO}\_2 \rightarrow \rm{MCO}\_3.\tag{1}$$

This reaction is usually exothermic in nature, as per Lackner et al. (1995) [26]:

$$\text{CaO} + \text{CO}\_2 \rightarrow \text{CaCO}\_3 + 179 \text{ kJ/mole} \tag{2}$$

$$\text{MgO} + \text{CO}\_2 \rightarrow \text{MgCO}\_3 + 118 \text{ kJ/mole.} \tag{3}$$

Consequently, sample characterization requires a chemical analysis that is commonly used (WDXRF).

It has been widely described in the literature that the main possible carbonation minerals are oxides, silicates, and anhydrite; therefore, it was necessary to determine the minerals present in the sample in order to evaluate the candidate. Natural wollastonite is a widely studied mineral in several works, for example, Huijgen and coworkers [27], as a candidate for mineral carbonation. Different studies used other types of sample, such as industrial waste, that were mainly composed of wollastonite [28–30].

The proposed methodology includes a sample preselection using the characterization of chemical and mineralogical conditions, and in situ carbonation. The second part of the methodology is carbonation tests in the selected samples. The flow diagram of the proposed methodology is shown in Figure 1.

**Figure 1.** Flow diagram of proposed methodology.

For that reason, mineral characterization by powder X-ray diffraction is commonly used. In the same way, we propose an in situ carbonation test and process the evaluation by using a diffractometer equipped with a reaction chamber in a CO2-rich environment. We mixed 2 g of the powdered sample and a couple of pipetted drops, and took a scan every six hours over the course of 24 h. These tests made it possible to obtain the first analysis of carbonation evolution, selecting a feasible candidate for the carbonation tests.

Once the carbonation tests were completed, the treated samples had to be analyzed by XRD, showing the presence of carbonate phases as carbonation by-products. Rietveld refinement of both the treated and the original sample allowed the quantification of the present phases, determining the percentage of new stable carbonate phases and the destruction (or partial destruction) of phases that provided the necessary calcium and/or magnesium.

Likewise, the presence of new precipitated minerals could be observed by SEM. Combined with microchemical composition from energy-dispersive spectrometers (EDS), it confirmed the possible by-product phase. If the precipitate was large enough, then the optical microscopy supply provided information on how the precipitate grew on the surface of the sample and texture, as well as estimating sizes.

As an alternative to the quantification processes, and especially when these new phases did not represent a sufficient percentage for correct quantification, to quantify the amount of CO2 capture, two techniques were employed: elemental carbon measure and weight loss by differential thermal and thermogravimetric analysis (DTA-TG) in the temperature range of carbonate mineral decomposition. In the first case, CO2 capture can be calculated by the direct conversion of carbon to dioxide (CO2 (wt %) = 3.6641 × C (wt %)) for the difference of the carbon content in the original and the treated sample. In the second one, weight-loss measuring in the right temperature range corresponding to the thermal decomposition of carbonates for the difference between treated and the original sample corresponded to the percentage by weight of captured CO2. Since it was different, if any of the minerals constituting the samples had total or partial decomposition in the same temperature range, they would not be quantified, since they would be present in both samples pre- and post-treatment, for example, carbonates that already originally existed. The evolution of soluble ion content helps what follows mineral destruction.

Another aspect to take into consideration was the evaluation of the specific surface area by Brunauer–Emmett–Teller (BET) analysis. This technique can help us understand how the specific surface of a sample evolves at the microscopic level (the texture). Complete porosity analysis from the nano- to the macroscale also allows a study of the porous system after carbonation by isotherm adsorption of CO2 and N2. In this sense, it was possible to analyze how the precipitate completed the sample-surface pores, reducing their size to a smaller scale from the macro- or mesoscale to the microor nanoscale.

Finally, it is common in the literature to use the Steinour formula [21,24,31–33] to determine the efficiency of the reaction from the theoretical maximal CO2 sequestration value obtained from the following stoichiometric formula (Equation (4)):

$$\rm CO\_2(wt\ \%) = 0.785(\% \rm CaO - 0.7\% \rm SO\_3) + 1.09\% \rm MgO + 0.71\% \rm Na\_2O + 0.468\% \rm K\_2O.\tag{4}$$

#### *3.2. Validation of Proposed Methodology Using Ca-Silicate-Rich Brick*

For each of the techniques described in the previous section, an example of the obtained results for the selected sample for this study (MPC2) and the correlations between them is presented. However, this example sample was already studied in more detail in a previous work of some of the authors in mineral carbonation processes [5].

In the MPC2 sample, CaO was around 15 wt % and 2 wt % for MgO (Table 1 and Tables S1–S3 in Supplementary Materials), which were appropriate values to be considered candidates for the mineral-carbonation process. This sample was composed of quartz, k- and alkali feldspar, plagioclases (orthoclase, anorthite, and albite), and calcium-rich silicate as wollastonite (Figure 2 and Table 2).

**Table 1.** Chemical composition of original brick MPC2 (wt %) by X-ray fluorescence (XRF) (modified from [5]).

**Figure 2.** X-ray diffraction (XRD) pattern of MPC2 (and Rietveld refinement). Abbrv.: Qtz—quartz; Ab—Albite; Wo—wollastonite; Or—orthoclase; Di—diopside; Zi—zincite (Internal Standard).

**Table 2.** Mineralogical composition (wt %) of selected brick MPC2 (Qtz: quartz; Wo: wollastonite; Or: orthoclase; Ab: albite; Di: diopside; An: anorthite; Amor: amorphous phase). Rexp, Rwp, and GOF are numerical indicators of how well the Rietveld model was refined. Rwp, residual of least-squares refinement (weighted), which must be improved in refinement (with common sense); Rexp evaluates data quality; and GOF, goodness of fit parameter.


An in situ carbonation test (Figure 3) showed the evolution of newly grown calcite over time in a CO2-rich environment. Intensity for the main calcite's peak (at d = 3.04 Å) increased directly proportional to time.

The next step was to perform carbonation tests on a laboratory scale with the selected bricks according to previous analysis, the high content of CaO, mineralogical composition rich in Ca silicate, and the presence of calcite in the in situ carbonation test. Tests were performed at room temperature and 10 bar pressure for different reaction times, and three fractions of particle sizes were representative of the total. Additionally, a test was carried out for the original sample at low pressure (1 bar), room temperature, and a 4:1 solid–water ratio in a 5 L volume hermetic reactor for 720 h of reaction time.

In the X-ray patterns of treated samples, compared with the nontreated, the most obvious differences were the presence of calcite, and the partial destruction of wollastonite and some orthoclase. The attack on wollastonite with carbonic acid allowed for the release of calcium ions and calcite precipitation [4–6,27,28,34–36].

**Figure 3.** In situ XRD MPC2 pattern where d = 3.04 Å, reflection corresponding to precipitated newly formed calcite.

Therefore, the carbonation process occurred in the following two steps: a) silicate mineral dissolution and b) carbonate precipitation. Several studies based on the mineral carbonation of wollastonite [27,28] described the process in an aqueous carbonation route as:

1. Dissolution of CO2 in water for the production of a (bi-)carbonate:

$$\text{CO}\_2\text{ (g)} + \text{H}\_2\text{O (l)} \rightarrow \text{H}\_2\text{CO}\_3\text{ (aq)} \rightarrow \text{HCO}\_3^-\text{ (aq)} + \text{H}^+\text{ (aq)}.\tag{5}$$

2. Leaching Ca from wollastonite by acidic attack:

$$\text{CaSiO}\_3\text{ (s)} + 2\text{H}^+\text{ (aq)} \rightarrow \text{Ca}^{2+}\text{ (aq)} + \text{H}\_2\text{O (l)} + \text{SiO}\_2\text{ (s)}.\tag{6}$$

3. Nucleation and growth of calcium carbonate:

$$\text{Ca}^{2+}\text{ (aq)} + \text{HCO}\_{3}^{-}\text{ (aq)} \rightarrow \text{CaCO}\_{3}\text{ (s)} + \text{H}^{+}\text{ (aq)}.\tag{7}$$

These new carbonate crystals were observed by stereomicroscope (optical microscopy) and scanning electron microscopy (Figures 4 and 5), with a composition close to theoretical calcite, as shown by the results of elemental analysis by EDS.

**Figure 4.** Layers of calcite on the MPC2 surface, observed by stereomicroscope after 720 h of CO2 treatment.

**Figure 5.** Scanning-electron-microscopy (SEM) image, and the energy-dispersive-spectrometer (EDS) spectrum and elemental quantification of MPC2.

The measurement of elemental carbon and weight loss by DTA-TG in the temperature range of the calcite decomposition (Table 3 or Figure 6a) was used to quantify the amount of captured CO2. Both results could be expressed in terms of calcite as carbonate ore. Obviously, it was an indirect method due to it being assumed, on the one hand, that the only phase that decomposes in that temperature range was calcite and, on the other hand, that it only precipitates calcite as a carbonated material. This weight loss also corresponded to the expulsion of CO2 and carbon content that was related to the theoretical content of CO2 in the chemical composition of calcite (Equations (8) and (9)) [37,38].

 $\% \text{ Calculate}$  $\text{TIA-TG} = \frac{\text{Ann}\_{450-900^{\circ}\text{C}}}{\text{CO}\_{2}\text{Theoretical}} \times 100 = \frac{\text{Ann}\_{450-900^{\circ}\text{C}}}{43.97} \times 100 = 2.274 \times \text{Ann}\_{450-900^{\circ}\text{C}}$  $\text{CalciteDIA-TG} = \frac{\text{Ann}\_{450-900^{\circ}\text{C}}}{\text{CO}\_{2}\text{Theoretical}} \times 100 = \frac{\text{Ann}\_{450-900^{\circ}\text{C}}}{43.97} \times 100 = 2.274 \times \text{Ann}\_{450-900^{\circ}\text{C}}$ 

$$\% \text{ Calculate}\_{\text{C-Element}} = \frac{\text{C-content}}{\text{C}\_{\text{Thavel}}} \times 100 = \frac{\text{C-content}}{12} \times 100 = \frac{25}{3} \text{C}\_{\text{content}} \tag{9}$$

**Figure 6.** (**a**) MPC2 thermogravimetry (solid line) and differential thermal analyses (dashed line); >4 mm and 720 h reaction time. (**b**) Calculated calcite content by DTA-TG (circle) and by C element (square) as function of reaction time and particle size fraction (modified from [5]).


**Table 3.** Carbon elemental content, mass loss by differential thermal and thermogravimetric analysis (DTA-TG), and calcite content calculated by both techniques.

The difference compared to carbon analysis using the elemental analyzer could be attributed to the sum of experiment errors and analysis sensitivity. Yet, it was also necessary to take account of the adsorbed CO2, which could be measured in CHNS instead of DTA. However, both followed the same tendency with respect to reaction times (Figure 6b). In both instances, calcite content was higher than the original and directly proportional to the reaction time in the studied size fractions.

The wollastonite attack with carbonic acid allowed for the release of calcium ions and calcite precipitation. Concerning the presence of soluble ions (Figure 7 and Table S4 in Supplementary Materials), the amount of Ca ions decreased over reaction time because of calcite precipitation, while Si increased as a consequence of partial silicate destruction. This partial destruction of silicates resulted in a new specific surface on the bulk (Figure 8). The increase of the specific surface had a direct relationship reaction time. The newly precipitated calcite on the sample surface also increased BET. However, both possibilities must have had a greater effect than physical CO2 absorption that would result in a reduction of the specific surface area.

**Figure 7.** Soluble Ca, Mg, and Si ions measured untreated and treated after 240 and 720 h of reaction for MPC2 (modified from [5]).

After the carbonation tests, the brick samples revealed macro- and mesoporosity as determined by Hg porosity, which affects the proportion and size of the pores (Figure 9a,b), with a decrease in microporosity and increase in nanoporosity by N2 and CO2 absorption (Figure 9c,d). All of them were a result of two differentiated processes: (a) the action of carbonic acid destroying calcium silicates that produced an irregular surface and an increment of macro- and mesoporosity, and (b) the precipitation of carbonates that filled the micropores and probably reduced them to nanopores.

**Figure 8.** Specific surface evolution by specific surface area (BET) as a function of reaction time for MPC2 (modified from [5]).

**Figure 9.** (**a**,**b**) Histograms of Hg porosity, (**c**) N2 adsorption, and (**d**) CO2 adsorption isotherms for original brick and treated brick layer.

Calculating efficiency was as the quotient between the percentage of captured CO2 (experimental) and theoretical CO2 by Steinour's formulae (Equation (4); Table 4). The calculated capture efficiency was proportional to the reaction time (the longer the time was, the higher the carbonation amount), whereas carbonation did not seem to significantly depend on particle size in the studied conditions. Results obtained at low pressure (1 bar) after 720 h of reactions for the total sample were very similar to those obtained for the finer fractions (2–4, 1–2 mm) at 10 bars.


**Table 4.** Carbonation efficiency (CE) according to Steinour's equation.

## **4. Conclusions**

The present research showed a suitable methodology to evaluate the possibility of using ceramic construction waste (or other types of construction and demolition waste) for mineral carbonation under surface conditions in short periods of time. Even though the experiment was conducted in comparatively large scale, each test method yielded only a very small fraction. Therefore, to increase reliability each test could be conducted several times.

Specifically, Ca-rich bricks were successfully used as raw material for direct mineral carbonation by the destruction of Ca silicates. The amount of carbonation was proportional to the reaction time, whereas it did not seem to significantly depend on particle size or pressure in the studied conditions.

Acceptable carbonation efficiency was achieved under the favorable conditions of low pressure and temperature.

These results open the possibility for future studies using the proposed methodology for other types of construction and demolition waste rich in calcium silicates or other calcium compounds, which could be directly carbonated by in situ injections of CO2.

**Supplementary Materials:** The following are available online at http://www.mdpi.com/2075-163X/9/10/612/s1, Table S1: List of certificated standars used for XRF calibration.; Table S2: Detection limit (L.D.), quantification limit (L.C.) and relative error for the measuremento of the mayor elements by XRF in Panalytical Axios spectrometer of the SGI Laboratorio de Rayos X (University of Seville) [September 2014]; Table S3: Mayor elements comparative between monitor standards (calibration validation) in white box and the certificate value in grey box [September 2014]; Table S4: Soluble Ca, Mg, and Si ions measured untreated and treated after 240 and 720 h of reaction for MPC2 and their standard deviations.

**Author Contributions:** D.M. collected and analyzed the data and wrote the paper; V.F.-A. conceived and designed the ideas; and revised the paper; P.A. made the formal Investigation, discuss the data and revised/edited the manuscript.

**Funding:** This research was funded by the Junta de Andalucía (Consejería de Economía y Conocimiento) (P12-RNM-568 MO project) and the contract of Domingo Martín granted by the V Plan Propio de Investigación de la Universidad de Sevilla.

**Acknowledgments:** Authors are grateful to editor and reviewers for their comments which improved the manuscript. XRD, XRF, ICP-OES, C-elemental and SEM analysis were performed using the facilities of the General Research Center at the University of Seville (CITIUS).

**Conflicts of Interest:** The authors declare no conflict of interest.
