Utilization of silica-enriched filter cake industry by-products as partial ordinary portland cement replacement

In recent years, industrial byproducts have been converted into useful and valuable commercial items. Reusing these byproducts plays a crucial role to ensure the circular economy and thereby safeguard the environmental impacts. In Ethiopia, the Aluminate Sulphate chemical factory disposes of filter-cake waste materials in landfills that have high silica content. The factory is using pure kaolin and other raw materials for the production of Aluminum Sulphate and Sulphuric Acid by burning at high temperatures. By-products materials were collected from the factory and then calcined (post-treated) at 600 °C for 2h in a muffle furnace. From Atomic Absorption Spectrometry measurement result, it is confirmed that the post-treated (at 600 °C/2h) silica-enriched filter-cake waste materials have a similar composition to Metakaolin (MK). Post-treated filter cake (named MK) became more amorphous having high reactive silica with very low impurities as it was calcined and quenched rapidly. In this study, the properties of blended Ordinary Portland Cement (OPC)-mortar samples were investigated with the addition of heat-treated filter cake waste materials (0%–20%) as a partial OPC replacement. X-ray diffraction, Fourier Transform-Infrared Spectroscopy, Differential Thermal Analysis, Scanning Electron Microscope, and Atomic Absorption Spectrometry were used to investigate the properties of mortar samples that contain post-treated filter-cake (MK) materials and OPC-cement. The flexural and compressive strengths of 10% MK + 90% OPC-mortar samples were enhanced at early curing ages, 7 & 28 days. Moreover, the flexural and compressive strengths of OPC mortars with 15% MK have been improved at 28 days of curing age. However, 20% MK + 80% OPC blended mortars have not shown any improvement in mechanical properties. Setting time, soundness, water absorption, and apparent porosity of cement pastes with the addition of post-treated filter cake (MK) are also analyzed.


Introduction
Currently, industrial byproducts materials are becoming a serious problem for human beings as well as ecology. Degradation of the environment and contamination of soil and water resources are brought on by incorrect disposal. The recycling of industrial wastes in the productive chain has emerged as an alternative material for minimizing damage. Awash Melkasa Aluminum Sulphate company in Ethiopia is producing 1,700 tons of useful chemicals annually and meantime filter-cake waste materials are disposed of in landfills [1]. Several worldwide researchers examined agricultural waste products (like bagasse ash, and rice ash), and industries' byproducts (like fly ash, and silica fume) to use as alternative cementitious materials. The main justification for this is that Ordinary Portland Cement (OPC) manufacture pollutes the environment by emitting almost a ton of CO 2 gas into the atmosphere for every ton produced. Cement factories also consume a significant quantity of natural Any further distribution of this work must maintain attribution to the author(s) and the title of the work, journal citation and DOI.
resources [2]. By 2030, it is anticipated that cement production would increase to four billion tonnes [3]. Cement production accounts for more than 5% of all man-made greenhouse gas emissions [4]. Such a negative impact on the environment opens the door for alternative materials that cut CO 2 gas emissions, reduce uncontrolled natural resource consumption, and recycle industrial and agricultural waste. Carbon dioxide emission from the cement industry is mostly due to: (a) the composition of fossil fuels in the industrial process since rotating kiln materials are required to burn up to 1450°C [5,6]. Limestone, CaCO 3 , is the basic raw material (∼70%) for the production of cement and decomposes into calcium oxide (CaO) and carbon dioxide (CO 2 ) at a high temperature [7]. Several types of research are undergoing in the world to overcome the above problems through substituting alternative materials. Metakaolin, fly ash (FA), and silica fume (SF) were used as partial replacements for cement, and these materials assist to reduce the setting time for mass concrete constructions while maintaining concrete long-term strength [8][9][10][11]. Ground granulated blast-furnace slag (GGBFS) and municipal solid waste materials were used as partial cement substitution, resulting considerable reduction of CO 2 gas emissions [12,13] and heat of hydration [14]. The amorphous nature and amount of pozzolanic oxides in industrial byproducts materials are crucial to use as a replacement for cement. Aluminate Sulphate chemical industry byproduct (filter cake waste materials) contains the required amount of pozzolanic oxides to consider as a cementitious material. With a simple thermal treatment process, the filter cake crystallinity state is changed to an active amorphous state. Currently, these filter cake waste materials are creating a problem for the nearby community of this industry here in Ethiopia. Several researchers conducted research on this waste filter-cake material and they reported that the AluminateSulphate filter-cake contains high silica materials [15,16]. Prior studies also stated that kaolin's crystalline structure can be broken down by calcination temperature which is necessary to generate the liquid phase and produce glassy liquid [17][18][19][20].
In this paper, it has been verified that post-treated filter-cake materials (MK) are used as supplementary cementitious material, and also analyzed the performance of post-treated filter cake containing OPC-mortar samples. The mortar samples were prepared using OPC cement, filter cake (0%-20%), sand, and water. Posttreated filter cake (MK) materials have higher pozzolanic reactivity and are used as partial cement substitution. Flexural and compressive strengths of MK-contained OPC mortar samples and pure OPC-mortar samples are compared. Moreover, the physical properties of OPC-cement paste samples with filter-cake materials are measured and compared.

2. Experimental procedures
2.1. Raw filter cake materials processing 2.1.1. Preparation and characterization of raw filter-cake material Figures 1 and 2 depict the byproduct residue material created during the manufacture of aluminum sulfate chemicals at Awash Melkassa Aluminum Sulphate PLC in Ethiopia. These solid waste residues were collected and pulverized and then sieved using mesh sizes of 45 μm. As illustrated in figure 2, raw filter-cake materials were quickly cooled from heat treatment at 600°C to achieve an amorphous state of materials (MK). Table 2 revealed the chemical compositions of unprocessed filter cake, post-treated filter-cake materials, and OPCcement (MK). It was investigated that filter-cake materials (before & after heat treatment) contained high silica content as prior reported [21,22]. Heat-treated filter cake material was analyzed by Differential Thermal Analysis/Thermogravimetry Analysis (DTA/TGA) (Shimadzu, DTG-60H). XRD was employed (Shimadzu, XRD 7000, λ CuK α = 1.5418Å), and Fourier Transform Infrared Spectroscopy (FT-IR) (FT/IR-6600typeA). Atomic Absorption Spectrometry (AAS-model, spectra AA-20 plus) was also used to analyze the major and minor oxide chemical compositions of post-treated filter cake materials. Moreover, the flexural and compressive strengths of post-treated filter cake-contained OPC-mortar samples were measured. Initial and final setting times of all the samples were measured using the Vicat needle under ASTM standard C191-82 [23], and Le-Chatelier expansion was measured according to ASTM C151 [24]. Finally, water absorption and apparent porosity tests for OPC mortars with and without filter cake were performed to check the durability of the specimens according to ASTM C20-00 [25].

Preparation and characterization of Post-treated filter cake (MK) contained OPC-mortar samples
Raw filter-cake waste material was burnt at a calcination temperature of 600°C for 2h and then quenched rapidly. Then, the samples were crushed to a particle size of 45 μm, and then mixed with cement (OPC 42.5R), standard fine aggregate (sand), and H 2 O, as listed in table 1. These mortar mixtures were used to make mortar blocks of sizes 40 mm×40 mm×160 mm. A water-to-binders (i.e. 95%OPC + 5%MK powder and 90%OPC + 10%MK powder) ratio of 0.5 was used during mortar sample block preparation. However, the workability of MK-contained mortar samples decreased a lot when the percentage of MK substitution rose by 15% to 20%. Thus, additional water (15 ml) was added while MK 15% and 20% were added for blended OPC-mortar sample preparation. The materials were thoroughly combined until the consistencywas achieved using an automatic mortar mixer machine. Then the mixed mortar materials were poured into each mold in a single layer (to 1/2 the penetration). The assembled mold was fixed in position on a vibrating machine, and a suitable hopper was employed to make filling easier. A jolting machine was used to vibrate the mold for 120 s. After 24 h, the produced samples were de-molded and cured as usual in curing conditions (at a temperature of 20 ± 3°C and relative humidity of 95%) for the specified period. After 3, 7, and 28 days of curing, the flexural and compressive strengths of mortar samples were tested.  3. Results and discussion 3.1. Characterization of untreated filter-cake 3.1.1. Thermal analysis of untreated (raw) filter cake There is a phase transformation of raw filter cake material to mullite and cristobalite crystalline at a heating temperature above 900°C. The crystalline phase-contained materials are not preferred to replace Ordinary Portland Cement (OPC). During the heating of raw filter-cake waste materials (as shown in figure 3), an endothermic peak appeared at 82°C which is linked to the loss of absorbed water (2.3%). An endothermic peak at 274.2°C is due to the breakdown of organic volatile components of filter-cake materials. At 508.5°C, the maximum mass loss from raw filter cake material occurred, indicating a substantial mass loss (4.9%) due to the dehydroxylation of raw filter cake materials. After calcination at 450°C-600°C, the raw filter cake material changes to non-crystalline materials (Al 2 Si 2 O 7 ), releasing the chemically bonded water molecules. The posttreated filter cake (MK) at 600°C/2h has increased pozzolanic reactivity, and dehydroxylation was carried out up to 850°C [26]. The DTA curve revealed a broad exothermic of around 940°C. This is owing to the recrystallization (0.42%) and transformation of dehydrated substances into mullite, cristobalite, and quartz, which is typical of metakaolin dissociation and spinel production [27]. The total mass loss of raw filter cake samples up to 1000°C is 8.04% (figure 3). Figure 3 inllustrated illustrated that the higher weight loss occurred at temperatures 450°C-600°C. In this temperature range, it is also expected that aluminum silicate hydrate in materials would be completely transformed into dehydrated aluminum silicate. At higher calcination temperatures (> 600°C), the weight loss of the materials was much smaller.  Table 2 displays the mineralogical compositions of the pre-treated and post-treated filter cake waste materials, and OPC cement. An Atomic Absorption Spectrometry measurement was employed to examine the filter cake pozzolanic contents of oxides (SiO 2 + Al 2 O 3 + Fe 2 O 3 ), alkaline oxides, and loss on ignition. The percentage of pozzolanic oxides (SiO 2 + Al 2 O 3 + Fe 2 O 3 ) in the post-treated filter cake (Metakaolin) material is 85.62%, which meets the ASTM C-618 requirement of 70% and above [28]. Thus, the result testified to the pozzolanic nature of the post-treated filter cake materials (MK), thus it is possible to take it ascement substitution. Furthermore, the MK's loss on ignition (LOI) is 7.56%, which is less than 10% and close to the preceding studies [22]. Due to its pozzolanic activity, the effect of partial Metakaolin replacement on the strength and durability of cementconcretes was reported [29]. MK has a three-foldhigher silica content (i.e., 69.2%) than OPC cement (i.e., 19.74%). It also has a low alkali concentration (Na 2 O + K 2 O = 1.61%), predicting a lower alkali-silica reaction potential.

3.2.2.
Phase identification of post-treated filter cake (MK) X-ray diffraction of the Unprocessed filter-cakeand post-treated filter cake (600°C/2h) samples were evaluated to validate the conversion of raw filter cake to an amorphous state. Figure 3 shows the Unprocessed filter cake contained three separate phases: kaolinite, illite, and quartz. The presence of kaolinite in unprocessed filter cake  is confirmed by the characteristics reflections at 2θ angles 12.4 and 24.9 by reference code (ICDD 01-083-0971).
The anatase phase was also observed at 25.28°(ICDD 00-021-1272). The presence of anatase in filter cakekaolinite is widespread owing to the geological location where the clay is mined, and it is regarded as an impurity in the finished product. Some of the investigations have found impurities such as Fe, Ti, and Al minerals in these clays [30,31]. One of the main minerals of clay illite was also observed at 2θ angles 23.05, 29.8, and 47.4. Thus, the XRD measurement result revealed that the pre-treated byproduct (unprocessed filter cake) has several peaks which are an indication of crystal phases in the materials. However, post-treated materials (MK) have shown a broad hump around 20°(figure 4) which confirmed that post-heat-treatment at 600°C/2h with rapidly quenching processes converted the crystalline phase of materials to an amorphous phase as stated in a previous study [32]. The calcination stage was found to be effective, and the de-hydroxylation process destroyed the kaolinite structure. As a result, an amorphous material was identified by an increase in the diffractogram background, as previously described [33,34]. In this work, amorphous silica was produced by calcining filter cake at 600°C/2h followed by water quenching. A similar result was also reported as metakaolin materials had high free energy and pozzolanic properties as the most reactive clay [35]. Metakaolin clay material is extremely reactive and appropriate for applications such as the creation of binding materials since it contains less quartz and illite [36]. In the present study, XRD results of the unprocessed and post-treated filter cake materials have shown closely resembled with previous work [37].
3.2.3. FTIR spectra of unprocessed filter cake and post-treated filter cake (MK) Figure 5 illustrates FTIR spectra of unprocessed filter cake and post-treated filter cake. Unprocessed filter cake shows notable broad stretching at 3435 cm −1 ; this is mostly due to H-O-H stretching with a compound class of absorbed water [38]. Analyzing the x-ray diffraction results using X'Pert High Score Plus software, OPC cement paste samples with and without MK (after 28 curing days) revealed calcium silicate hydrates, calcium hydroxide, mono-sulfate hydrate, unreacted SiO 2 , and free CaO. The pure tri-calcium aluminates phase is not found in the cement paste due to the rapid rate of hydration processes. However, the tri-calcium aluminates phases with gypsum are caused for the formation of ettringite (C 3 A.     According to a prior study, MK was employed as a partial substitute for cement. It reacts with calcium hydroxide to form additional calcium silicate hydrate, which strengthens the mortar structure [44]. Apart from calcium silicate hydrates, calcium hydroxide (CH) and calcium aluminosilicate hydrates (CASH) are the main hydration products of OPC-cement pastes with and without 10% MK.
Post-treated filter cake (MK) also contains alumina which produced additional alumina-containing phases some of which are crystalline as explained in a previous study [44]. 3CaO.SiO2 & 2CaO.SiO2 phases in clinker are reacted with moisture (H 2 O), and produced Calcium Silicate Hydrates (CSH) and Ca(OH) 2 . Post-treated filter cake's composition of Al 2 O 3 .2SiO 2 phase is also reacted with Ca(OH) 2 to produce secondary Calcium Silicate hydration products and C 2 ASH 8 , C 4 AH 13 , and C 3 AH 6. From figure 7, at a peak of 29.5°, it is confirmed that 10%MK + 90% OPC cement paste has a higher amount of calcium aluminates silicate hydrates than 100% OPC cement paste. XRD pattern revealed that calcium aluminates silicate gel formed in both MK + OPC and OPC pastes by referring code 01-086-0402, i.e., CSH peaks appear at 2 q angles: 27.57, 29.44, 32.19, 32.54, 38.75, and 39.48. It is also confirmed that unreacted Quartz existed in both samples by referencing the code (01-085-0335), which has been identified at 2θ angles of 20.9, 26.74, 46.01, and 50.1 as indicated in a previously reported study. [45].

Thermal analysis of mortar samples
DTA/TGA analysis is employed to determine the thermal stability of hardened mortar samples with and without MK (figure 8). As the temperature increased, the weight of both samples decreased due to the loss of water, the breakdown of oxides, and the escape of volatile particles/elements from mortars. The endothermic peak at 83.4 is mainly due to the removal of free water molecules and the disintegration of the amorphous component of calcium silicate hydrates (CSH) ( figure 8). An endothermic peak at 447.5°C is also related to the loss of water combined with calcium hydroxide [46]. The breakdown of CaCO 3 to lime (CaO) and carbon dioxide (CO 2 ) causes the endothermic peak at 703.8°C as in a priorstudy [47]. The dehydration of calcium hydroxide is represented by the endotherm at 447°C []. Disintegration of CaCO 3 to CaO and CO 2 (g) 703.8°C allowed the volatile particles to escape and cause the weight loss of the samples above 600°C [49,50]. The total weight loss of the 10%MK + 90%OPC mortar sample is 14.7% and the 100% OPC mortar sample is 13.8%. This is due to the dehydration of the interlayer calcium silicate hydrates (CSH), calcium aluminate hydrates (CAH), and calcium sulphoaluminate hydrates. The addition of MK causes an increment inCSH production, as revealed by XRD data. This CSH is responsible for interlayer dehydration, which decrements in the mass percentage. The more CSH, calcium aluminate hydrates (CAH), and calcium sulphoaluminate hydrates, leads to the more interlayer dehydration.

Flexural and compressive strength of MK contained OPC-mortar samples
The flexural strength of MK-blended OPC mortar samples is lower than conventional 100% OPC mortar samples at 3 days of curing (early curing ages). However, the flexural strength of 10%MK + 90%OPC mortar samples is higher than 100% OPC mortar at 7 and 28 days of curing (later curing ages) ( figure 9). This is most likely related to the substantial production of calcium silicate hydrates in MK blended OPC-mortar later curing ages. For instance, the flexural strength of 100% OPC mortar is 8.9 MPa, and 10%MK + 90% OPC mortar is 9.6 MPa at 28 days of curing. However, the flexural strength of MK + OPC blended mortar samples decreased with 15% and 20% MK replacement which is most likely related to a decrease in Alite (3CaO.SiO 2 ) content in mortar. Post-treated filter cake (MK) particles might also aggregate due to its particle large surface area (12,430 cm 2 g) −1 compared to the specific surface area of OPC particles (3,394 cm 2 g) −1 [50].
At 3 days of curing, the compressive strength of MK-contained OPC mortar samples decreased with the addition of MK content ( figure 10). This is due to the addition of MK in MK + OPC powder reducing the amount of alite phase, 3CaO.SiO 2 (C3S), which is responsible for increasing the early-age strength of a mortar. By the 7th and 28th days of curing, 10%MK + 90%OPC mortar samples produced greater compressive strength compared to 100% OPC mortar samples. For instance, the compressive strength of a conventional 100% OPCmortar sample has 44 MPa but the 10%MK + 90%OPC-mortar sample has 51.8 MPa by 28 days of curing. Thus, the compressive strength of the mortar sample increased by 17.73% with 10% MK replacement of OPC cement in 28 days of curing. However, the compressive strength of 15%MK + 85%OPC & 20%MK + 80OPC-mortar samples has lower than 100% OPC-mortar samples for all curing ages due to decreased amount of alite phase-3CaO•SiO 2 and belite phase-2CaO.SiO 2 in mortar materials is responsible for the early-age strength of mortar. For 15%MK + 85%OPC and 20%MK + 80%OPC blended mortar sample preparation, 15 ml of extra water is added to improve the workability. High water content most likely caused pores to form inside the dried

Setting time and expansion of MK-contained OPC pastes
The setting time and soundness of the MK + OPC-pastes and pure OPC-pastes are measured. The amount of water required for normal consistency rose as the amount of MK increased. As MK content in OPC mortars increased, the mixture mortar requests more water for proper consistency. This is due to MK's hygroscopic nature and its particles' high specific surface area of 12,430cm 2 g −1 which is 3 times higher than cement. The ASTM standard C191-82 specifies that Hydraulic cement paste has an initial setting time of not more than 45min and an ultimate setting time of not more than 420min [23]. All prepared OPC-pastes samples containing MK satisfied ASTM standard C191-82 requirements ( figure 11). The setting time of the MK + OPC cement pastes is longer than that of conventional OPC cement paste due to the lower amount of tricalcium aluminate (3CaO.Al 2 O 3 ) in MK + OPC blended pastes. Tricalcium aluminate in cement pastes is responsible for the release of high-heat energy and causes the rapid setting of the paste. The setting time of cement paste is controlled by the cement phases, 3CaO.Al 2 O 3 and 3CaO.SiO 2 , which is decreased in MK + OPC cement blended pastes.
Expansion (soundness) of dried cement paste might occur due to the hydration of free oxides CaO and /or MgO inside the cement. The soundness of combined MK + OPC cement paste is measured using the Le-Chatelier apparatus conforming to IS: 5514-1969. Figure 12 clearly shows the expansion of dried pastes decreased with an increased amount of MK in OPC cement pates. Lower expansion of MK-blended OPC cement pastes might be related to the low amount of Mg(OH) 2 and Ca(OH) 2 formation in the blended cement pastes since lesser content of MgO & CaO in MK as observed in AAS data (table 2). According to the ASTM C151 [24], the upper limit of Le-Chatelier expansion is 10 mm. Based on the measurement result, the expansionof dried cement pastes decreased with increasing MK in blended OPC-cement paste. For instance, 20%MK + 80% OPC cement pastes (MK20) have lower expansion values (0.4 mm) than 100% OPC-cement paste (CO = 0.85 mm) ( figure 12).

Water absorption and apparent porosity of MK& OPC blended mortar samples
Since MK has a higher specific surface area (12,430 cm 2 g) −1 than cement (3,394 cm 2 g) −1 , water absorption of MK + OPC blended mortars at 3 days of curing rose from 17.88% to 30.49% as MK amount increased from 0 to 20% ( figure 13(a)). At 7 days of curing, water absorption of MK + OPC mortars increased from 13.9% to 26.55% with increasing MK amount from 0 to 20% MK respectively. The experiment results revealed that MKblended OPC mortars have higher water absorption at a curing age of 3 days compared to the 7 days for OPC mortars. This may be attributed to 7 days of curing being favorable for important compound formation compared to 3 days of curing. Pores spaces in the blocks decreased with increasing curing age. Apparent porosity of MK + OPC blended mortars decreased from 26.3% to 19.11% with increasing MK content to 10% ( figure 13(b)). This is due to the finer particle size of MK acting as filling voids/spaces in mortar. However, higher MK (above 10%MK) addition in MK + OPC mortar increased the apparent porosity probably because the finer particle size of MK started the agglomeration. Moreover, as explained before, extra water is added for 15% and 20% MK blended OPC mortars during sample preparation, which might result in more pore formation in dried mortars. The volume of the internal open pores in the specimen is indicated as a percentage of the specimen's outer volume.

4. Conclusion
Based on the present study, the following conclusions are drawn: The utilization of industrial waste products for producing eco-friendly construction materials is applicable for the partial replacement of OPC. The calcination of filter cake helps to break down the organic volatile components and mainly transformation of kaolinite to metakaolin. Amorphous reactive silica is successfully obtained by the calcination of filter-cake aluminum sulfate byproduct at 600°C/2h followed by rapid quenching. The summation calcined filter-cake contains above 70% of pozzolanic oxide materials. As a result, considerable calcium silicate hydrate generation is predicted in posttreated filter cake (MK) blended OPC-mortar samples at later curing ages, which assists to boost the later strength of mortar samples. During 3 days of the curing period, the addition of MK not has shown strength enhancement of MK + OPC blended mortar samples. However, by the 7th and 28th days of curing, the compressive strength of MK + OPC blended mortar samples increased with 10% MK replacement. For 10%MK + 90%OPC mortars, the flexural strength is 6.5MPa (@ 7 days cure) and 9.6 MPa (@ 28 days cure). The compressive strength of 10% MK + 90%OPC mortar samples is 31.1MPa (@ 7 days of curing) and 51.8MPa (@ 28 days of curing). The compressive strength of a conventional 100% OPC-mortar sample has 44MPa and OPCmortar sample with 10% MK has 51.8MPa at a curing age of 28 days. Thus, the compressive strength of the mortar sample increased by 17.73% with 10% MK replacement of OPC cement at 28 days of curing. Water absorption of MK + OPC blended mortar samples increased with increasing MK percent. At 3 days of curing, water absorption of 100% OPC mortar is 17.88% and 20% MK + 80%OPC blended mortar is 30.49%.