Effects of Non-thermal Plasma Treatment on the Geopolymerization of Kaolin Clay

Kaolin samples obtained from Cameroon were used to produce geopolymer binders. Prior to its application, the raw kaolin samples were activated through the gliding arc plasma treatment using both spatial post-discharge and direct mode. A mixture of sodium hydroxide and silicate was used as the alkaline solution. In order to study the influence of the modifications generated by the gliding arc plasma treatment on the geopolymerization process, X-ray diffraction, thermogravimetric analysis, differential scanning calorimeter and Fourier transform infrared spectroscopy were carried out. In addition, scanning electron microscopy, nitrogen physisorption and compression tests analysis were also carried out on the resulting geopolymer samples to access their mechanical performance. The results showed that the geopolymerization process was not completed at the curing temperature of 90 °C. Plasma spatial post-discharge mode treated kaolin led to 20.48% increase in compressive strength when compared with the geopolymer prepared from raw kaolin.


Introduction
In many activities field, especially in civil engineering, cement is of great interest because it forms a plastic paste capable of binding various materials. The production of hydraulic cement, also called ordinary Portland cement (OPC) requires high energy consumption [1]. The production of cement contributes to about 5-8% of CO 2 global emissions which is a major drawback from an environmental point of view [2].
Geopolymer cements can be used as an alternative to ordinary Portland cements in order to reduce CO 2 emissions [3,4]. For instance, the optimal production of 1 metric ton of geopolymeric cement generates 0.180 metric ton of CO 2 from combustion of carbon-fuel, compared with 1 metric ton of CO 2 for Portland cement, that is, six times less [5]. Geopolymers are a class of semi-crystalline aluminosilicate materials, typically synthesized by a chemical reaction between an amorphous aluminosilicate powder and a highly concentrated alkaline solution [6][7][8]. Geopolymerization is carried out in three phases. The first step consists of the dissolution of aluminosilicate material to form reactive precursor ions, [SiO(OH) 3 ] − and [Al(OH) 4 ] − [9,10].
This dissolution process depends on some characteristics of aluminosilicate such as its hydrophilicity, chemical composition, mineralogy, surface area and especially crystallinity. Kaolinitic clays are calcined to be amorphized and transformed into metakaolin before being used as an aluminosilicate material [11,12]. However, the use of kaolin instead of metakaolin could be beneficial as the high temperatures (500°-800°C) calcination step can be avoided. Additionally, metakaolin-based geopolymers require a lot water due to an increase in their porosity [13]. Therefore, developing a low cost and sustainable strategies to avoid the use of high temperatures is of great interest.
Recently, non-thermal plasma technology has received increasing interest for several applications such as surface treatment, textural properties modification, organic matter oxidation etc. Previous works demonstrated that the plasma treatment of kaolin improves its solubility and also lead to slight modification of its crystallinity [14][15][16][17]. The physicochemical modifications of activated kaolin particles could improve their geopolymeric reactivity. To the best of our knowledge, no literature work has investigated the gliding arc plasma processing kaolin powder in the geopolymerization process. For all these reasons, this paper investigates the use of gliding arc plasma instead of the heat treatment. Gliding arc plasma pre-treatment of kaolin might contribute to the greener, less energy-consuming production of green concrete.
The objective of this article was to verify whether the modifications that occur during the gliding arc plasma treatment of kaolin are likely to significantly and positively influence the geopolymerization process and compressive strength of the materials obtained.

Materials
The kaolin used as aluminosilicate raw material was obtained from natural deposits in Cameroon. It was enriched in argillaceous mineralogical phase by wet sieving then; the dried clay was crushed and sieved with an 80 μm mesh sieve.

Plasma Treatment of Kaolin
The resulting raw kaolin labelled K was treated with gliding arc plasma for 30 min using moist air as feeding gas at the flow rate of 800 L/h. Two methods of treatment were carried out: the direct treatment ( Fig.1), where the target was exposed in contact to the highly primary reactive species generated in the plasma discharge and the spatial post-discharge treatment or indirect treatment mode (Fig. 2), in which the target (kaolin) was placed in another reactor (secondary reactor) where it interacts with long-live plasma-generated species flowing from the primary reactor [16]. The plasma-treated kaolin sample in the direct mode was labelled K 1 and the plasma spatial post-discharge processed kaolin sample was labelled K 2 .

Geopolymers Synthesis
The alkaline activator solution was prepared by mixing 10 M sodium hydroxide solution with sodium silicate solution. The proportion of sodium silicate was 7% of the mass of the sodium hydroxide solution. The resulting alkaline solution was sealed to prevent reaction with CO 2 from the atmosphere and then stored for at least 24 h at ambient temperature before use to allow full silica dissolution, cooling and equilibration. The sodium water glass solution had a composition by weight of: 28.7% SiO 2 , 8.9% Na 2 O and 62.4% H 2 O and the bulk density was 1400 kg/m 3 .
The pastes were prepared as followed: various kaolin samples (K, K 1 and K 2 ) were mixed with alkaline solution for 10 min (4 mins at slow speed and 6 mins at high speed) in an M & O brand mixer, model N50-G. The powder/liquid activator mass ratio of 1.34 was used. The paste obtained was used for shaping the test specimens using cylindrical PVC molds (diameter of 20 mm and height of 40 mm). The specimens were labelled as follow: GP for the material produced with K, GP 1 for the material produced with K 1 and GP 2 for K 2 . After casting, the samples were vibrated for 3 mins on an M&O brand electric vibrating table, type 202, No. 106 to expel air trapped by particulate matter during mixing.
Thereafter, the paste specimens were cured for 24 h at ambient temperature in the laboratory (30 ± 3°C) then at 90°C in an oven (Heraeus VT 50.42 Ek) for 24 h to accelerate the hardening. Demoulding was carried out after the previous operations and the samples were then stored at ambient temperature in a dry place for 28 days. To ensure reproducibility, six duplicate specimens were prepared for each geopolymer paste.

Analytical Techniques
The chemical composition of the kaolin was determined using ICP-AES (Inductive Coupled Plasma-Atomic Emission Spectrometry).
The powder diffractograms of the kaolin samples and of the materials produced were obtained on a Bragg-Brentano type apparatus (Bruker D8 advanced). This device is equipped with a Ge (111) monochromator operating by reflection of Kα 1 radiation from copper (λ = 1.54056Ǻ); the scanned angular domain (2θ) is between 5 and 80°. Bruker DiffracPlus EVA software was used for the identification of crystalline phases.
The FTIR analyses were carried out using a BRUKER alpha-p spectrometer with pure ethanol as solvent. The FTIR spectra were collected over a wave number range of 400 to 4000 cm −1 . The correction of the baselines of the spectra obtained was performed using the ATR algorithm incorporated in an FTIR software package named OPUS. A linear model was chosen. All FTIR analyses were carried out at room temperature.
Morphological characteristics at the micrometric scale were carried out by scanning electron microscopy, using a Philips FEI XL γ0 FEG ("Field Emission Gun") device.
To obtain information on the pores in the materials, BET analyses were performed with a micrometric analyser; Tristar 3000 model by nitrogen adsorption/desorption at 77 K. Each sample was first degassed at 423 K under vacuum for 1 h. The detection of open porosity was then done by injecting a nitrogen/helium stream in a 30/70 ratio. The BET equations were used to explain the physisorption of gas molecules on the specimens.
Thermal analyses were carried out, using a LINSEIS model STA PT-1000 analyser. To carry out the analyses, 20 mg of each sample was introduced into the Fig. 1 Simplified schematic of gliding arc plasma reactor for direct processing [16] Fig. 2 Simplified schematic of gliding arc plasma reactor for indirect processing (spatial postdischarge) [16] alumina crucible of a sample holder equipped with a precision balance. The heating rate was set at 10°C/min under an air atmosphere (Self-Generated Atmosphere of Air).
The compressive strength was measured after 28 days using an M&O electro-hydraulic press, type 11.50, No. 21. These tests were carried out according to EN 196-1 [18].

Characterization of Raw and Plasma Modified Kaolin Samples
The chemical and mineralogical compositions of raw kaolin were reported in our previous work. From XRF analysis, we saw that raw kaolin mainly consists of silicon oxide (SiO 2 ), aluminium oxide (Al 2 O 3 ) and other trace compounds such as Iron III oxide (Fe 2 O 3 ), Titanium oxide (TiO 2 ) while the loss on ignition was 14.75%. However, the main mineral was kaolinite. There were other minor phases such as Quartz (Q), Gibbsite (G), Anatase (A), and Muscovite (M) in smaller proportions [16].
Some characteristics of plasma-treated kaolin were also reported in our previous study. The gliding arc plasma is able to functionalize kaolinite contained in kaolin powder. Exposure of kaolin clay material to gliding arc plasma resulted in an overall increase in FTIR absorbance peaks of the clay material, such as lower bands around the values 1622 and 1391 cm −1 due to the breakdown of certain Si-O-Si and Si-O-Al chemical bonds and the appearance of new peaks in the hydroxyl region (around the wave number values 3670 and 3520 cm −1 ). This corresponds to the formation of new hydroxyl groups on the surface of the material. Moreover, an increase in total dissolved solids for both direct and indirect treatment modes reflected the increase of the kaolin hydrophilicity [16]. Figure 3 gives the thermograms of the various kaolin samples. In general, DSC curves showed three endothermic phenomena and one exothermic phenomenon: It can be seen that the mass loss corresponding to the dehydroxylation of kaolinite was greater for sample K 1 (−8.33%) and sample K 2 (−9.68%) when compared with sample K (−6.62%). This difference confirms the formation of new aluminol groups on the surface of kaolinite after the plasma treatment of kaolin powder. In addition, the previous observation suggests that the spatial postdischarge treatment mode is more appropriate for the functionalization of the kaolinite in comparison to the direct plasma treatment mode for this purpose. As shown in Fig. 4b, the bands at 3619 and 3690 cm −1 corresponds to the elongation vibrations of the hydroxyl   (2) groups of kaolinite [20]. However, there was a sharp decrease in the intensity of these peaks (~70%). These observations indicate that the kaolinite particles were significantly dissolved by the alkaline solution but still remained in the system. The quantitative study of geopolymer's infrared spectra also revealed that, GP contains more kaolinite than GP 1 and GP 2 . This suggests that the treatment improves the dissolution of kaolinite during the first step of the geopolymerization process; with a more significant effect on K 2 sample. The bands at 3430 and 3520 cm −1 , are clearly distinguishable on the spectrum of the K sample, which correspond to the elongation vibrations of the O-H bond. Interestingly, theses bands were not very visible in the spectrum of GP, GP 1 and GP 2 . The appearance of the bands around 1400 cm −1 (Fig. 4c) is related to the vibration of the extended O-C-O bond, reflecting the presence of sodium carbonate in the geopolymer material [11].

FTIR Analysis
In Fig. 4d, the band at 907 cm −1 for GP reflects the elongation vibration of the Al-OH bond, which shifted towards higher wave numbers in the spectra of the geopolymers, reflecting media alkalization [20].
Likewise, the bands at 1002 and 1026 cm −1 are attributed respectively to the vibrations of symmetrical and asymmetric Si-O-Si bond elongations in kaolinite [21]. These bands decrease in intensity and move towards large wavenumbers, which indicates structural rearrangement of Si-O-T (T = Si, Al) bonds within materials following the geopolymerization reaction and the formation of the geopolymer gel [22]. The intensity of the bands at 1002-1026 cm −1 gives an indication of the level of ordering within the geopolymer gel network [23,24]. The bands between 3600 and 3250 cm −1 and at 1650 cm −1 are attributed respectively to the vibrations elongation of H-O bond and distortion vibrations of (O-H) bond in the water molecules absorbed at the surface or present in the cavities of geopolymers [22].
The widening of the bands between 906 and 1025 cm −1 observed in Fig. 4d for the produced geopolymers confirms that these materials are different from the starting clay samples. Indeed, this broad band can be assigned to the asymmetric elongation vibration of the Si-O and Al-O bonds in the SiO 4 and AlO 4 tetrahedra of geopolymers [25,26].  Figures 5, 6, 7 and 8 show the XRD patterns of kaolin geopolymers. The diffractograms of geopolymers showed the formation of hydroxysodalite as a new crystalline phase. Usually, zeolitic phases are produced during the geopolymerization at elevated temperature [27,28]. It appears that the characteristic peaks of some minerals initially present in the kaolin samples still exist in the XRD patterns of the geopolymer products. These minerals are kaolinite, quartz and gibbsite. However, the decrease of peaks intensity after geopolymerization indicates they remained or partially reacted. Similar observations have been reported by Tchakoute [29]. Indeed, kaolinite has low reactivity in an alkaline medium due to its crystalline structure [30,31].

XRD Analysis
It should also be noted the appearance of two peaks on the XRD patterns of GP 1 and GP 2 respectively at approximately 2-theta angle of 76°and 74°. The corresponding minerals of these peaks have not been identified yet. Figure 9 shows the thermograms of geopolymers synthetized from raw kaolin and gliding arc plasmatreated kaolin in spatial post-discharge mode. These DSC curves highlighted four endothermic phenomena that occur around temperatures of 138, 289, 500 and 770°C. The phenomenon observed at 138°C corresponds to the removal of zeolitic water, while at 289°C, it could be the dehydroxylation of gibbsite for which an alkaline specie did not react with the alkaline activating solution. The most important endothermic phenomenon occurs at 502°C probably corresponding to the elimination of the constitution water of the residual kaolinite contained in the material [19] as indicated by the corresponding FTIR and XRD patterns.

Thermal Analysis
For these two materials of Fig. 9, three exothermic phenomena also occur at 40, 852 and 941°C. The exothermic phenomenon observed at 40°C corresponds to the reinitiation of the geopolymerization process of the remaining kaolinite. Considering the fact that the geopolymerization was partial and that the reagents remained in the materials obtained, the gradual increase in temperature during thermal analyses promotes a resumption of the geopolymerization reaction; knowing that the curing temperature improved the geopolymerization process [32]. This geopolymerization process ends around 852°C but, around 940°C, the structural reorganization of the little metakaolinite formed in spinel occurs. Figure 9 also shows the thermograms of differential scanning calorimeter (DSC) and thermogravimetric analysis (TGA) which are related to the geopolymer produced from kaolin treated with gliding arc plasma in direct mode (K 1 ). The DSC curve shows three endothermic phenomena which correspond to temperatures of 92, 142 and 283°C. The phenomenon observed at 92°C would correspond to a structural rearrangement of kaolinite and that at 142°C from zeolitic water while at 283°C, it would be the dehydroxylation of gibbsite. Three exothermic phenomena also occur, particularly at temperatures of 41, 209 and 650°C. The exothermic phenomenon observed at 41°C corresponds to the reinitiation of the geopolymerization process of the

BET Analysis
As shown in Table 1 the geopolymer particles had a larger specific surface area (geopolymers based on raw kaolin: 16.36m 2 /g; geopolymers based on kaolin treated in direct mode: 11.67m 2 /g; geopolymers based on kaolin treated in indirect mode: 23.67m 2 /g) than those of the kaolin samples used (untreated kaolin: 11.06m 2 /g; kaolin treated in direct mode: 8.68m 2 /g; treated kaolin in indirect mode: 7.14m 2 /g) for their production. This increase in the specific surface area is greatly due to the increase in the external surfaces area of the materials synthetized [33,34]. Figure 10 gives the BET isotherms analyses of the various geopolymers synthetized. From this figure and according to the IUPAC classification, the isotherms were of type IV indicating that the materials synthesized are mesoporous.
In addition, these isotherms contained an H3 type hysteresis loop. This hysteresis manifested itself at the relative pressure value of 0.5. It characterizes the pores between non-rigid aggregates of plate-shaped particles and indicates that the pores in materials are in uniform slit-shaped [35,36].

SEM Observations
Ground samples were used for both BET and SEM analysis. Figures 11, 12 and 13 show the surface morphology at different scales of the milled geopolymers ( Fig. 11 from GP, Fig. 12 from GP 1 and Fig. 13 from GP 2 ).
As seen in 10 μm scale images of all geopolymer materials, the geopolymerization process resulted in aggregation of particles. The images obtained at the scale of one micron confirm that the particles have aggregated and they are in the form of plaques. This agrees with the porous slit-shaped network suggested by the type of hysteresis loop (H3 type hysteresis) obtained after the nitrogen physisorption. Figure 14 gives the 28-day compressive strength results of the geopolymers made with different type of kaolin sample. It can be seen that the geopolymer material from the kaolin treated in spatial post-discharge mode gave the highest compressive strength (6 MPa) while the lowest compressive strength (1.245 MPa) was recorded with direct mode treated kaolin based geopolymer. Thus, compared to raw kaolin, the indirect treatment increased the compressive strength of the geopolymer material by 20.48%, while the direct treatment reduces the strength of the geopolymer material by almost 75%. From Table 1 in relation to the micropore volume, we noticed that the compressive strength of the geopolymers was inversely proportional to the pores size within the material. Therefore, the less porous sample had the greatest compressive strength. These observations are in agreement with those reported by Izquierdo et al. [37], demonstrating that the increase in porosity negatively impacted the mechanical properties of geopolymers.

Compressive Strength
Also, an increase in the compressive strength of the geopolymer material produced from indirectly treated kao l i n m a y b e a t t r i b u t e d t o a n i n c r e a s e i n t h e geopolymerization degree with regard to the greater functionalization (hydroxylation) and the greater hydrophilicity of the kaolin obtained following the treatment in spatial post-discharge mode [16]. The high solubility of this sample of kaolin in the alkaline solution may justify fast deconstruction of the aluminosilicate into a reactive precursor SiO(OH) 3 − and Al(OH) 4 − ) which could explain the greater geopolymeric reactivity [38]. Further, the compressive strength increases with the extent of geopolymerization [37]. This increase of the compressive strength is in agreement with the results obtained by Xu and Van-Deventer reporting that a greater solubility of aluminosilicates in the alkaline solution is likely to increase the compressive strength of geopolymers [10]. On the other hand, the low strength development of geopolymer based on kaolin treated in direct mode is due to the increase in the degree of crystallinity of kaolinite, which reduced its geopolymeric reactivity. Elimbi et al. [12] demonstrated that the decrease in crystallinity of kaolinite promotes the geopolymerization. As demonstrated in previous work, the high percentage of gibbsite present in raw kaolin as well as its non-reactivity during the geopolymerization process may also contributes to the overall reduction in the compressive strength of the materials produced [29].

Conclusion
The results obtained in this work show that kaolin partially reacts during the geopolymerization process at the curing temperature of 90°C. The geopolymer synthesized from gliding arc plasma treated kaolin in spatial post-discharge mode has the greatest compressive strength representing an improvement rate of 20.48% when compared to the geopolymer based on raw kaolin. On the other hand, a reduction of the compressive strength of 75% is recorded using kaolin treated with gliding arc plasma in direct mode. Thus, the gliding arc plasma treatment of kaolin clay significantly influences its geopolymerization process as well as the compressive strength of the resulting geopolymers. However, these geopolymers have low compressive strength compared to metakaolin-based geopolymers. For this reason, gliding arc plasma activation of kaolin cannot efficiently replace its heat treatment for geopolymerization process.