4-1 Carbonation of Steelmaking slags through the RPB
Carbonation of steelmaking slags via High-Gravity Carbonation Process would be discussed in this section. It looks forward to take advantage of the steelmaking slags recycling and CO2 reduction. In this study, examination of the change in physicochemical properties of slags and the best operating conditions include the rotating speed, liquid to solid ratio and particle size would be achieved.
4-1-1 Effect of Carbonation on Characteristics of Feedstock
The chemical composition of four types of slags with and without carbonation are shown in Table 4-1. According to the X-ray fluorescence analyses, the main elements in slags are calcium, silica, iron and aluminum. The proportion of calcium and magnesium in slags is related to the potential in capturing CO2. Among all of the slags, electric arc furnace reducing slag (EAFRS) contains up to 54.36% of calcium, which is 1.5 to 2 times as much as that in other slags. Thus, it is expected to be the best materials for carbon capture in this study. As the calculating result, the amount of theoretical carbon captured by 1 ton of basic oxygen furnace slag (BOFS), refining slag (RFS), electric arc furnace reducing slag (EAFRS) and oxidizing slag (EAFOS) are 0.288, 0.277, 0.427, 0.172 tons respectively. Except calcium, the content of aluminum and iron also plays an important role in slags, which would affect their hardness and increase the
energy consumption during crushing and milling. Among these four materials, only EAFRS could be crushed by pestle in lab easily.
In comparison with differences of the slags before and after the carbonation, the free-CaO content in these slag reduces greatly after carbonation. It shows that the High-Gravity Carbonation Process did transform calcium oxide into calcium carbonated effectively and get to the effect of being stabilized. Although the proportion of calcium oxide under 1.5% is acceptable according to ASTM regulations, if the free-CaO turns into calcium hydroxide in the hydration process, it will bring about 97% of expansion, generating micro creaks and causing fatal defect (Neville et al., 2002). To improve the strength development and durability, less content of free-CaO should be residue in the slags.
Besides, there are higher proportion of alkali metal such as sodium and potassium in the EAFOS. Despite the amount of the sodium and potassium residue in EAFOS slightly reduces during the carbonation process due to its high solubility, the proportion of alkali metal in carbonated EAFOS still reaches 5.23%, which is far exceed the requirement of ASTM regulation (Na2O+0.66K2O < 0.6%).
Table 4-1 Chemical Composition of Fresh and Carbonated Steelmaking slags from XRF Analysis (F=fresh, C=carbonated).
[1] Not detected under 0.01%.
Table 4-2 is the crystal composition of steelmaking slags with and without carbonation. Comparing the analyzing result of XRD with Inorganic Crystal Structure Database (ICSD), the main compositions of slags are shown as below: FeO (no:01-075-1550) and CaCO3 (no: 01-076-2712) in BOFS, 2CaO•Al2O3•SiO2 (no: 01-076-7523) and CaO•Fe2O3 (no: 00-032-0168) in RFS, 2CaO·SiO2 (no: 00-049-1672) in EAFRS, SiO2 (no:01-087-2096) and CaCO3 (no: 01-083-1762) in EAFOS. According to figure 4-1 to 4-4, the calcium carbonated diffraction peaks have higher intensity than those before carbonation. Implying that the amount of calcium carbonate in carbonated slags increase significantly during the High Gravity Carbonation Process. Furthermore, some of the mineral compositions in the fresh slags such as CaO•Fe2O3 would decomposed due to calcium ions leach out and interact with CO2 in the solution. In the meanwhile,
Compositions
free-CaO in slags react with water forming calcium hydroxide, which can also combine with CO2 to form CaCO3.
Table 4-2 Major Crystal Structure of Steelmaking Slags from XRD Analysis
Figure 4-1 Diffraction Peak Comparison of BOFS
Items Major structure
BOFS Fresh FeO, CaCO3
Carbonated CaCO3, FeO
RFS Fresh 2CaO•Al2O3•SiO2, CaO•Fe2O3
Carbonated 2CaO•Al2O3•SiO2, SiO2
EAFRS Fresh 2CaO•SiO2, SiO2
Carbonated CaCO3
EAFOS Fresh SiO2, CaCO3
Carbonated SiO2, CaCO3
Figure 4-2 Diffraction Peak Comparison of RFS
Figure 4-3 Diffraction Peak Comparison of EAFRS
Figure 4-4 Diffraction Peak Comparison of EAFOS
Composition change during carbonation could also be presented by using the differential thermal gravity (DTG) curve, which is a technique to analyze the components with different derivative weight peak at a specific temperature. As shown in figure 4-5, the first peak of weight loss is illustrated at 50 to 110 ℃, corresponding to the removal of moisture. The second peak of weight loss is related to decomposition of amorphous C-S-H gel at 100 to 150 ℃. Some unobvious weight loss between 200-400 ℃ may related to dehydration of iron hydroxide and alumina hydroxide. The third peak of fresh BOFS at around 441℃ is associated to the calcium hydroxide dehydration.
Then the last peak between 600 and 850 ℃ is related to the decomposition of calcium carbonated and magnesium oxide. Different from figure 4-5, there is just one significant peak being preserved in the DTG curve, which is the peak of CaCO3 and MgCO3
decomposition of figure 4-6. By comparing Figure 4-5 and 4-6, it implies that calcium in structures such as calcium hydroxide and amorphous C-S-H gel may be leached out,
Figure 4-5 DTG Curves of Fresh Steelmaking Slags.
Figure 4-6 DTG Curves of Carbonated Steelmaking Slags
4-1-2 Effects of Operating Parameters for Carbonation
The first variance in the carbonation process is the liquid-solid ratio (L/S ratio). In this study, L/S is set between 10 to 50 ml/g. Higher liquid-solid ratio represents the proportion of slags in the whole system is relatively less, and lower liquid-solid ratio represents a relatively lower reaction kinetics of carbonation process. As shown in
affect the carbonation conversion yield. However, the conversion yield at higher L/S ratio (50ml/g) is slightly higher than that at lower L/S ratio. For example, EAFRS conversion yield at L/S=50 ml/g is 2.65% higher than that at L/S=10 ml/g. This phenomenon may be the combination effect of two opposite theories. At lower L/S ratio, higher calcium concentration condition can be produced, which is advantage to react with CO2 and nucleate. However, higher calcium concentration makes the slag particles hard to leach out according to the Le Châtelier's principle. In this study, leaching seems more important than nucleation, resulting better conversion yield at higher L/S ratio.
Moreover, owing to carbonation conversion yield is calculated by subtracting the calcium content tested by TGA and XRF. Conversion yield between different slags may associate to the difference of calcium-related crystal forms. The higher amount of free-CaO and calcium hydroxide in slags, the easier calcium ions can be leached out from crystal and reacts with CO2 in water. Unfortunately, some calcium mineral like dicalcium silicate which especially crystallized in the γ phase is hard to leached out owing to its stable regular-octahedron structure. This type of structure doesn’t provide enough cavities for water molecules to penetrate in and break down the Ca-O bond.
Thoroughly considering the efficiency and the convenience of operation, including processes of the reactor cleaning and the slags filtration. The slags which offered to be used later in the cement substitution experiments are produced at the proportion of L/S=20 ml/g.
Table 4-3 Effect of Carbonation Conversion Efficiency Under Different Rotating
Figure 4-7 Effect Carbonation Conversion Yield under Different L/S Ratio (Rotating Speed = 1100rpm, Particle Size = 32μm).
Figure 4-8 shows the conversion yield change against to the rotating speed from 700-1300 rpm at L/S=20ml/g. Among the range from 700 to 1000 rpm, conversion yield increases apparently as the rotating speed accelerates. That’s because higher rotating speed can break the slurry into smaller liquid drops on the stainless wire in rotating packed bed, which provides a larger surface area for CO2 to dissolve from gas, forming carbonic acid and interacting with calcium ions in the slurry drops. Thus, the mass transfer efficiency can be elevated.
Unfortunately, the conversion yield slightly goes down as rotating speed accelerates beyond 1100 rpm. This phenomenon may also come from the mass transfer
phase too fast to react with calcium ions. Large amount of carbonic acid results in the decline of pH in the solution. Part of carbonic acid reacts with the precipitated calcium carbonated, creating calcium bicarbonate re-dissolves, which consequently affects the amount of calcium in the slags being filtered from carbonation.
Table 4-4 Effect of Carbonation Conversion Yield under Different Rotating Speed.
Type of
Figure 4-8 Effect of Carbonation Conversion Yield under Different Rotating Speed (L/S = 20 ml/g, particle size = 32μm).
Figure 4-9 shows that under the condition of 1100 rpm and 20 ml/g, different particle sizes of EAFRS affect carbonation conversion yield. Apparently, as the particle sizes become smaller, carbonation conversion yield becomes higher. What’s more, according to table 4-5, the process of grinding and milling slags has little impact on chemicals. It means that the increasing of conversion yield completely results from higher surface area which smaller particles of EAFRS have. Therefore, calcium ions can leach out more easily and react with CO2.
Figure 4-9 Effect of Particle Size on Carbonation Conversion Yield. (L/S=20 ml/g, rotating speed = 1100rpm)
Table 4-5 Chemical Composition of Fresh EAFRS with Different Particle Size
Particle size (μm) 32 64 96 128 160
Composition (%)
SiO2 26.32 17.23 16.92 16.61 16.54
Al2O3 2.64 2.64 2.92 3.29 3.78
Fe2O3 1.89 1.94 2.00 2.57 3.37
CaO 48.67 48.16 48.24 46.93 46.35
MgO 5.65 5.60 5.60 5.80 6.33
SO3 1.31 1.34 1.48 1.62 1.64
K2O 0.01 0.01 0.01 0.01 0.02
Na2O 0.16 0.18 0.20 0.19 0.20
f-CaO 1.22 1.21 1.40 1.55 1.47
4-2 Cement Replacement by Steelmaking Slags
Slags contain calcium, silica and aluminum, which elements are similar to the Portland cement. Moreover, some lattice defects in slags which produce during the crystallization period can generate activity in slags. After grinding into microscale, these slags can be used as the supplementary cementitious materials (SCMs) in cement industry. Since cement manufacturing process will consume a great amount of original materials such as limestone, clay and coal and also emits a great tons of CO2. Using SCMs in concrete to reduce the amount of cement usage can not only lessen environmental destruction but also cut down the CO2 emission. In this research, slags within and without carbonation will replace partial of Ordinary Portland Cement at the substitution rate of 5, 10, 15% and tests are given to workability, strength, and durability to evaluate the feasibility.
4-2-1 Effect of Substitution on Workability of Cement
According to the standard consistency of pastes and standard fluidity of mortars in figure 4-10 and 4-11, the water demand for maintain the workability is getting more as the slags substitution rate getting higher. Compared with the Ordinary Portland Cement, the standard consistency of the pastes with 15% replacement of slags are 5.9-9.6%
higher. This phenomenon is related to kinds of reasons include the surface area and hydrophobic. In 2007, Kourounis and people pointed out that cement mixed with the materials with more surface area will promote its effective wet surface area and total amount of adsorbed water. Because of calcium ions dissolving after the carbonation of
slags can promote the nucleation, the mean particle size of carbonated slags becomes smaller than those before carbonation. Therefore, the hydroscopicity of carbonated slags slightly increases. Besides, the water requirement increase as the substitution rate goes up is also related to the hydrophilicity of slags. In this study, the hydrophilicity sequence of these slags are: OPC>BOFS>EAFOS>RFS>EAFOS. With some organic compounds residue, EAFOS is oilier and hydrophobic which makes it harder to scattered in water. This cause an abnormal result that the slag after carbonation needs even less water than those before carbonation while making mortar and paste.
Table 4-6 Standard Consistency of Paste with Different Replacing Ratio of Slags
Items Replacing
Fresh Carbonated Fresh Carbonated
OPC - 26.9
Figure 4-10 Standard Consistency of Paste with Different Replacing Ratio of Slags.
Figure 4-12 and 4-13 show the initial setting time and the final setting time of slags separately at the replacement ratio of 5%, 10%, and 15%. These figures can be divided into two types to discuss. The former concludes BOFS, RFS and EAFRS, which initial setting time and final setting time get 5-10% longer as increasing of the replacement ratio and carbonation doesn’t affect the setting property of setting. This phenomenon isn’t much related to the property of the material added. The delay of setting time results from the fact that standard consistency pastes with high replacement ratio need more amount of water, so the pastes take more time to lose their plasticity. Additionally, in 2013, Camiletti indicated that calcium carbonate can help nucleation in the aqueous phase during hydration and accelerate the process to shorten its setting time.
The latter is EAFOS, which is extremely different from the former type. Adding fresh EAFOS in paste will significantly lengthen the initial and final setting time. The final setting time of 15% fresh EAFOS replacement is up to 390 minutes, which is 62%
higher than the OPC paste and even exceed the standard of CNS regulation. Besides, the final setting time of slags after carbonation is usually quite longer than slags before carbonation. However, the setting time of carbonated EAFOS pastes are obviously shorter than those partial replaced by fresh EAFOS and just 6-18% higher than the OPC paste. These phenomena can be explained that some organic compounds are residue in EAFOS, which act like adding retarders in cement paste. During the hydration process, the long chain shape organic substance will adsorb onto the particle surface of tricalcium silicate and tricalcium aluminate, blocking water from penetrating into the
compounds in EAFOS can be removed through the High-Gravity reactor during carbonation. Organic substance will dissolve in the water and be washed out.
Furthermore, owing to calcium carbonate crystal can helps nucleation, the initial setting time and the final setting time go back to the normal scale right away. In conclusion, the reaction of materials processed with High-Gravity Carbonation Process has little influence on the workability of cement. What’s more, the problem that EAFOS affects normal setting can be improved through the High-Gravity Carbonation Process
Table 4-7 Setting Time of Paste with Different Replacing Ratio of Slags
Items Replacing
Fresh Carbonated Fresh Carbonated CNS
Figure 4-12 Initial Setting Time of Pastes with Different Replacing Ratio of Slags.
4-2-2 Effect of Substitution on Strength of Cement
According to figure 4-14 to 4-16, the compressive strength of mortar specimens always getting worse as the substitution rate of slags growing higher and some specimens even can’t fit the standard of CNS regulation. This is result from the lacking of binder. As cooling down at a slower rate, these slags have more time to crystalized in better arrangement, which have less defects in the crystal structure to dissolve during hydration. Thus, comparing to C3S in OPC, these slags can’t produce enough C-S-H gel to provide strength. Additionally, the specimens using carbonated slags as pozzolanic material usually perform better than those partial replace by fresh slags. This phenomenon can be explained by 2 reasons. The first is calcium carbonate can promote nucleation and accelerated the hydration reaction. The second is providing gel as shown in eq. 4-1 and 4-2. In this equation, calcium carbonate can react with calcium hydroxide and tricalcium aluminate to produce the ettringite type of C-A-C̅-H structure, which can fill the crack and cavities in the specimens. However, there are still some detail in the compressive strength development trends of these four slags and they will be discussed as below.
2𝐶3𝐴 + 1.5CC̅ + 0.5𝐶𝐻 + 22.5𝐻
→ 𝐶3𝐴 ∙ 𝐶𝐶̅ ∙ 11𝐻 + 𝐶3𝐴 ∙ 0.5𝐶𝐶̅ ∙ 0.5𝐶𝐻 ∙ 11.5𝐻 (4-1) 𝐶3𝐴 + 𝐶𝐶̅ + 11𝐻 → 𝐶3𝐴 ∙ 𝐶𝐶̅ ∙ 11𝐻 (4-2) Compressive strength of specimens with fresh BOFS at different curing ages are 11.5-32.4% lower than the strength of OPC. However, the strength of specimens with carbonated BOFS even performed better than OPC at the early ages such as 3 and 7
days. The cause of this result has been mentioned above. Furthermore, it is noteworthy that the carbonated BOFS mortar specimens doesn’t maintain their advantage to the later age (28 days). This behavior is due to the properties of C-A-C̅ -H structure.
Comparing with C-S-H gel, the C-A-C̅-H structure isn’t strong enough that it can just provide the filling function.
Unlike other slags, compressive strength of fresh RFS mortar specimens at the early age are stronger than OPC. This special phenomenon may associate with the crystal structure. According to table 4-2, minerals such as 2CaO•Al2O3•SiO2 may have higher reacting activity owing to Al3+ displaces Si4+ in the structure, causing inhomogeneous of charge distribution. Besides, the good performance of compressive strength in the early age may be attributed to its small particle size. Small particle blending in mortars may also provide a denser stacking, which enhance the strength in a physical view. Unfortunately, the advantage of strength just last until 7 days. As shown in figure 4-16, RFS mortar specimens do not perform remarkably. It’s also due to the weaker strength of the hydration product.
Comparing with BOFS and RFS, EAFRS mortar specimens have weaker strength.
Specimens with fresh EAFRS can barely meet the CNS regulation at the early stage such as 3 days and 7 days. At 28 days, the compressive strength of fresh EAFRS mortar specimens fail to reach the regulation because of lack of binder and the micro-cracks resulting from free-CaO expansion. Besides, the poor performance of compressive strength of EAFRS may also due to the larger particle size. Because EAFRS is less hard
miller. Using pestle to crush, the EAFRS has larger mean diameter and smaller specific surface. Thus, the compressive strength of fresh EAFRS mortar specimens at 3 days are slightly worse than the those before carbonation. However, the calcium carbonated in EAFRS after carbonation reacts with the calcium hydroxide which follows eq. 4-1 to 4-2 and produces C-A-C̅ -H hydration product, which can fill some micro-cracks.
Therefore, the later compressive strength of carbonated EAFRS mortar specimens is higher than that before carbonation. Surprisingly, the 28-day compressive strength of carbonated EAFRS mortar even meets CNS regulation at 15% replacement ratio.
The most special one of these slags is EAFOS. EAFOS’s strength development is similar to EAFRS’s; however, there is another reason why the mortar specimens with slags EAFOS have a weak strength in later stage. According to table 4-1, EAFOS has excessive Na and K. According to CNS regulation, cement’s total alkalinity
（Na2O+0.66K2O）needs to be less than 0.6%. The total alkalinity in this EAFRS is 1081% over the regulation, which will cause serious Alkali-aggregate reaction, AAR).
After being poured into molds, it will absorb water and produce fatal swelling as well as a large amount of micro-cracks, which does harm to the long-term strength just as shown in figure 4-17. On the other hand, carbonated EAFOS produces alkali-silica gel and absorbs water during the carbonation process, causing a large amount of swelling.
After it is filled up in the molds, the second swelling won’t occur. Thus, the long-term compressive strength is higher than mortar specimens with fresh EAFOS mixing inside.
To sum up, mortar specimens with partial replacement of carbonated slags via High-Gravity carbonation process have higher compressive strength in the later stage
than those before carbonation. It’s a good replacement material for CO2 capture as well as cement. Although carbonated EAFOS’s property is improved, only the specimens with 5% replacement ratio of slags can meet CNS strength regulation at 28 days.
Table 4-8 Compressive Strength of Cement Specimens at Different Curing Ages
Items Replacement
Figure 4-14 The 3rd day Compressive Strength of Mortars with Different Replacing Ratio of Slags.
Figure 4-15 The 7th day Compressive Strength of Mortars with Different Replacing Ratio of Slags.
Figure 4-16 The 28th day Compressive Strength of Mortars with Different Replacing Ratio of Slags.
4-2-3 Effect of Substitution on Durability of Cement
4-2-3 Effect of Substitution on Durability of Cement