The major elemental composition expressed in oxides was performed by X-ray fluorescence (XRF) with an automated Panalytical Axios (Marverl Panalytical Ltd, Netherlands) model spectrometer. The samples were prepared for analysis as glass discs to reduce the “matrix effect”.

The mineralogical composition of the untreated and treated samples was determined by X-ray diffraction (XRD), using a Bruker D8 Advance diffractometer (Bruker, Germany) with standard monochromatic Cu Kα radiation and operating at 40kV and 30mA. Scanning was performed with a 0.015° 2θ step size, and at 0.1s per step from 3° to 70°. Semi-quantifications were performing using the software Profex (version 3.14.0) [23].

The carbonate content of treated samples was determined by two analytical methods: differential thermal and thermogravimetric analysis (DTA-TG) and an elemental analyzer. DTA-TG were performed on a TG Netzsch STA 409 PC (Netzsch, Germany). Samples (around 150mg) were heated in an aluminium oxide crucible under a nitrogen atmosphere at 10°Cmin−1 from room temperature to 1200°C. Weight loss was measured by thermogravimetric analysis in the temperature range of 600–1000°C relative to the total carbonated decomposition. Elemental carbon content was measured using an elemental analyzer, Leco Truspec CHNS Micro (Leco, Michigan, USA), which calculated the carbonated ratio assuming that the whole carbon content was calcite.

The specific surface area (SSA-BET), micro- and nano-porosity were measured with an ASAP 2420 (Micromeritics, Georgia, USA) instrument using the adsorption of N2 at liquid nitrogen temperature (−196°C) and CO2 at room temperature. Before measurement, all samples were degassed at 150°C for 180min and finally outgassed to 10−3Torr. SSA was calculated using the classical Brunauer-Emmett-Teller theory (BET), and H2 and CO2 isotherms are analyzed using the Barrett–Joyner–Halenda (BJH) [24] method to yield micro- and nano- pore size distribution, respectively.

Macro- and meso-porosity were studied using a mercury porosimeter Pore Master 60-GT (Quantachrome Instruments Anton Paar, Austria). Low pressure mercury porosimetry applied to macropores while high pressure measured mesopores. The relationship between the pore diameter and Hg-intrusion pressure was calculated by the Washburn equation [25]. The pressures used in mercury porosimetry were from 1bar to 4000bar.

Soluble Ca, K, Na, Si and S ions were measured by a simultaneous inductively coupled plasma - optical emission spectrometry (ICP-OES) analysis using an ULTIMA 2 (Horiba Jobin Yvon, New York, USA) instrument. The specimens were prepared by mixing the solid powder samples with water and stirring for 24h, then concentrating the liquid phase by centrifugation followed by filtering through a Nylon 0.22μm syringe filter.

Micro-observations of morphological changes were obtained by scanning electronic microscopy (SEM) using a FEI Teneo (Thermo Fisher Scientific, Waltham, MA USA) microscope equipped with energy dispersive spectrometers (EDAX).

Due to the limits of detection and quantification of the different techniques used, the combination of analytical techniques allowed the qualitative analysis of the capture mechanisms and the indirect quantification of the amount of carbon captured [8].

The analysis of the majority elements of the samples analyzed (Table 2) shows an inverse relationship between the silicon and calcium content related to the compressive strength, being higher to greater SiO2.

The calcium oxide content in the concretes ranges from 9–50wt% and the magnesium oxide content ranges from 0.5–5wt%. Such high Ca and/or Mg materials seem to be appropriate for mineral carbonation.

The concretes, as it was expected, are mainly composed of quartz, carbonates and feldspars (alkali- and plagioclase). Minor phyllosilicates, gypsum and portlandite (calcium hydroxide) were also detected in some concretes (Table 3).

In limestone concretes the presence of carbonates (calcite, dolomite and ankerite) was predominant, varying from 53–89wt%, being these mainly the minerals rich in Ca (and/or Mg). Hc1 showed a percentage not much higher than 5wt% of other Ca-minerals that could be considered for carbonation (plagioclase and portlandite).

Siliceous concretes were mainly composed of quartz (65–68wt%). Hs2 presented 10wt% of calcite as main Ca-rich mineral, against Hs1 with a higher calcium content assigned to portlandite, plagioclase and ettringite (hydrated calcium aluminium sulphate) as possible minerals for carbonation.

For this reason, the siliceous concrete sample Hs1 was selected for carbonation tests in different particle size fractions, reaction times and different pressures.

The selected samples were tested for carbonation at room temperature, 20% humidity and 20bar CO2 pressure for different fractions and reaction times.

After these tests (in general) an increase in calcite intensity was observed by XRD (Fig. 2), as well as almost complete disappearance of the intensities corresponding to portlandite (Eq. (1)) and ettringite. The destruction of both minerals would provide calcium ions for the subsequent precipitation of the calcite, fixing CO2.

Fig. 2.

XRD patterns of the raw concrete (bottom) and carbonated concrete (top) after 30 days and fraction 0.5–1mm. Po: portlandite, Q: quartz; Ett: ettringite: Anot: Anortite; Alb: Albite; I: Illite; Ca Calcite.

(0.52MB).

Around 27.5° (asterisk in Fig. 2) a significant increase in intensity was observed. This increase could be due to a combination of two effects: (i) an increase in the calcium plagioclase content; and (ii) the appearance of a dehydrated phase of the ettringite due to drying of the samples after carbonation tests. This phase would have a structure equivalent to β-CaSO3[26]. Nonetheless, some ettringite should be destroyed due to an increase of silicon and sulfur ions compared with the original sample (Table 4). Also, an increase in free Ca ions were observed. Ca ions available for calcite precipitation due to the mineral destruction of portlandite and ettringite. Decrease in soluble K and Na may be related to increase in illite and albite respectively.

While Eq. (1) showed the carbonation reaction of portlandite, the following equation corresponds to ettringite [27]:

According to TG results (Table 5), as quantification of CO2 fixed, were proportional to the reaction time and inversely proportional to fraction size. In other words, 30 days and <0.5mm fraction was the higher CO2 capture amount to Hs1, around 6.5wt%. However, this value was higher calculated from C-elemental as CO2, around 7.8wt% (an increase of 2.1wt% of C respect to the original value). The difference was due to some proportion of CO2 was physically retained (physical adsorption).

The morphological time evolution was analyzed using the minor fraction size (<0.5mm). Macro- and meso-porous were determined by Hg-porosimeter, which affects the proportion and size of the pores (Fig. 3). The maximum volume of this porous diameter was in the range around 150 to 650μm (Fig. 3b) in front of the untreated sample, which maximum was around 200–400μm (Fig. 3a). So, an increase of porous diameter was observed due to the acidic attack on the surface in the carbonation process [20].

Fig. 3.

Histogram of Hg porosity for (a) original and (b) treated Hs1 concrete at 30 days and <0.5mm fraction size.

(0.93MB).

Whereas the smaller pore, micro- and nano-pore, determined by N2 and CO2 physisorption (Table 6) shown a decrease of the pore diameter to 5nm from around 15nm. This reduction was reached after 5 days of reaction, remaining constant at longer times. Combined with the decrease in the N2 and CO2 BET specific surface due to the smoothing of the irregular surface were related to the growth of calcite crystals that complete the pore by decreasing to smaller pore size as chemical CO2 retention.

Therefore, the mechanism would be as follows: (i) the mixture of CO2 and water, or CO2 in solution, begins to penetrate the empty pores; (ii) within the pore by the action of the acidity of the carbonic generated by the combination of gas and moisture, it dissolves calcium minerals (i.e. portlandite, ettringite); (iii) the released calcium ions combined with carbon ions precipitate as calcium carbonate filling the pores.

Nevertheless, some proportion of CO2 gas was physically adsorbed as shown in isothermal N2 and CO2 absorption curve (Fig. 4a and b). Both the treated samples presented a lower quantity of adsorbed gas than the original Hs1 concrete.

Fig. 4.

(a) N2 adsorption and (b) CO2 adsorption isotherms for original and treated Hs1 concrete at different reaction time and <0.5mm fraction size.

(0.28MB).

Micro observations confirmed the presence of precipitated calcium carbonate (Fig. 5). That neoformed calcite exhibits different habits. Mainly, geometric aggregates or with lamellar habit and dendritic growth were observed. Chemical analysis by EDS shows Ca, O and C (which was in the limit of proper quantification by the instrument) as principal elements. Also, a peak in the spectrum (not quantified) correspond to Pt, the metal used to the spattering.

Fig. 5.

Calcite crystal growth observed by SEM, EDS spectrum and elemental quantification for Hs1 (<0.5mm) at 30 days reaction time.

(0.68MB).

Some of these crystals were observed damaged, with cracks or fractures on the edges (Figs. 5 and 6). These alterations in the geometry of regular crystals may be due to the growth of new crystals and the destructive action of the acidic environment.

Fig. 6.

A geometrical habit of crystal growth and some alterations in borders.

(0.7MB).