Determining Optimal Characteristics Parameters of CO2 Adsorbent to Enhance its Fluidization
do not necessarily reflect the views of UKDiss.com.
Determining Optimal Characteristics Parameters of CO2 Adsorbent to Enhance its Fluidization by Using the Taguchi Method
Abstract
In this study, hydrophilic silica nanoparticles are used to modify Ca(OH)2 particles as CO2 adsorbent. Taguchi experimental design is used to identify the optimal characteristic parameters for improving the fluidization of modified adsorbents. Four parameters including SiO2 wt%, type of alcohols, sieved size of SiO2 and Ca(OH)2 particles are selected for conducting experiments in a fluidized bed. Through the ANOVA, sieved size and wt% of SiO2 nanoparticles (accounting for 72.87% and 16.74% of the total contribution of the four selected parameters, respectively) are found as the most significant parameters in fluidization quality of modified adsorbents. Fluidization experiments also confirm the effectiveness of these two parameters. CO2 adsorption tests; by measuring pH variation of pure water during adsorption, reveal that the improvement of Ca(OH)2 fluidizability serves to enhance the carbonation reaction. Modifying of Ca(OH)2 adsorbent by hydrophilic silica nanoparticles instead of hydrophobic ones is suggested based on the better adsorption results.
Keywords: Fluidization, Ca(OH)2, Silica nanoparticles, CO2 adsorption, Taguchi method
- Introduction
Fluidization is a very potential technique which provides a high specific surface area of contact between the solid and the gas phases. This advantage makes beds have a wide application in a variety of multiphase chemical reactions especially in CO2 capture technologies 1,2. A viable technological process for the capture of CO2 from post-combustion flue gas is based on the carbonation-calcination cycle, which makes use of CO2 sorbents based on lime (CaO or Ca(OH)2) powders in a potentially big demand 3-5 Lin et al. 6, 7 identified that the performance of Ca(OH)2 for CO2 capture is better than that of CaO. Ca(OH)2 is a highly cohesive powder for which fluidization behavior belongs to the Geldart C type 8. On the other words, Ca(OH)2 particles form large and compact agglomerates that cannot be disrupted by a gas flow leading to a highly heterogeneous fluidization behavior with slugs and channels 2,8. This difficulty in fluidization arises because the inter-particle forces are greater than those which the fluid can exert on the particle 9. Stable gas channels and slugs developed during fluidization of Ca(OH)2 particles, result in gas bypassing through the bed, hindering the efficiency of the gas-solids contact surface available for CO2 sorption and consequently very low efficiency of CO2 capture 2,3. Because to come in contact with unreacted sites, CO2 must penetrate into the aggregates by diffusion, which is a slow process 4. Therefore, destabilization of gas channels which means improvement of Ca(OH)2 fluidization, leads to remarkable enhancement of the CO2 capture performance in a lab-scale fluidized bed 8.
A number of techniques have been developed to improve the fluidizability of fine cohesive powders, all of which can be classified into two groups: external and intrinsic methods. The external methods employ external forces such as acoustic field 10-12, magnetic field 13,14 and vibrating field 15-17 to overcome the adhesion force between particles. The intrinsic method is altering the intrinsic properties of particles by reducing the attraction forces between them 18. One possibility to minimize the attractive forces between cohesive particles is to increase the surface roughness by coating the surface of the particles with additives 19. A lot of experiments have demonstrated that the fluidization quality of fine powders can be improved significantly as another kind of particle is added into the bed 8,18 . If the mixing method is effective enough to achieve a uniform coating of nanoparticles (NPs) onto the surface of the particles of the cohesive powder, interparticle attraction is decreased, thus reducing the formation of large and compact aggregates 8. Recently, Valverde et al. 2,8 showed that by mixing Ca(OH)2 with hydrophobic silica NPs, the easily fluidizable silica NP agglomerates cause to dispersion of the fine Ca(OH)2 particles and then serves to catalyze CO2 adsorption. In addition to increase the capture efficiency, using hydrophobic silica powders with a higher mechanical strength, reduces the attrition of the sorbent material when subjected to high mechanical stresses 20. It should be mentioned that, hydrophobic particles are produced by substituting the hydroxyl groups of the hydrophilic particles by organic groups in a process called hydrophobization 21. This surface treatment of hydrophilic NPs with; for example, dimethyl dichlorosilane to make them hydrophobic needs a very complex and expensive process by the supplier. Therefore, it seems that using hydrophilic silica NPs in improving the fluidization behavior of Ca(OH)2 powders might be economically more advantageous. Hydrophilic silica particles are characterized by a surface containing hydroxyl groups and hydrophobic silica have a surface that contains alkyl groups 21. van der Vegate and Hadziioannou 22 have shown that the mean adhesion force (measured by AFM) between hydroxyl groups is 0.9 nN, but is only 0.3 nN between alkyl groups. This implies that the adhesion force for the hydrophobic silica coated cohesive particles is smaller than the adhesion force for the hydrophilic silica coated cohesive particles. Therefore, at a first glance, silica hydrophilic NPs don’t seem a good choice of mixing additive for improving the fluidization. However, Tahmasebpoor et al. 23 reported that hydrophilic silica NPs in the presence of isopropanol (ISP) alcohol behave almost identically to their hydrophobic counterparts in the absence of ISP.
The influence of using hydrophilic NPs as a coating material to improve the fluidization of cohesive particles has not yet been studied. In this regard, the main objective of the present research is studying the fluidization behavior of Ca(OH)2 cohesive powders by adding hydrophilic silica NPs in order to increase the CO2 adsorption rate. To clarify the optimum CO2 capture efficiency, the effect of different parameters including; weight percentage of SiO2 additive, type of alcohol vapors including methanol, 2-propanol and ethanol, sieved size of SiO2 NPs and also sieved size of Ca(OH)2 powders are tested. Taguchi method is applied to design the experiments and to analyze the experimental data. Finally, in order to compare the results, the hydrophilic silica coated sample with the best fluidization behavior along with its corresponding hydrophobic sample are tested in CO2 adsorption experiment.
- Experiments
2.1 Materials
A commercial highly cohesive Ca(OH)2 powder (Merck) with the particle size in the range of microns, was used as CO2 sorbent. To study the effect of sieving factor on fluidization behavior, Ca(OH)2 powder was sieved using 75 and 106 μm sieves placed on a shaker. As an additive, we used Aerosil 130 (Evonik), which is an amorphous hydrophilic silica with a primary particle size of 16 nm. To compare the results with the previous works, Aerosil R974 (Evonik) which is a hydrophobic silica powder with a primary particle size of 16 nm, was also used as an auxiliary material. Primary silica NPs are sintered at contact points forming small aggregates of size of the order of microns, which in turn attract to each other due to van der Waals forces to form highly porous complex agglomerates of size up to hundreds of microns 20. To show the effect of removing large agglomerates on enhancement of fluidization of silica NPs, they were also sieved by using three different sieves including 300, 600 and 850 μm. Hydrophilic silica NP agglomerates can be fluidized uniformly in a bubble-less, highly expanded state and extremely large gas-solid contact efficiency in the presence of alcohol vapors 23. In this regard, three different alcohols including; methanol, ethanol and 2-propanol were used in this work. The list of all used materials are listed in Table 1.
Modified samples were prepared simply by hand dry-mixing of Ca(OH)2 and SiO2 powders. By mixing, the easily fluidizable silica NP agglomerates serve as dispersants and carriers of the fine Ca(OH)2 particles and then the gas-solids contact efficiency between Ca(OH)2 and the fluidizing gas is improved. The minimum weight percentage of nanosilica needed to ideally cover a certain percentage of the surface of Ca(OH)2, which is often referred as surface area coverage (SAC), is calculated as 15% 20. However, we used silica nanopowders with different weight percentages of 10, 15, 20, 25, 30 and 35 in our experiments to find the appropriate conditions of coating.
2.2 Set-up
The experimental set-up used to adsorb CO2 that contains solid fluidized bed system to fluidize the sorbent and liquid solution system to view the adsorption rate using pH changes is shown schematically in Fig.1. The fluidization and adsorption process were carried out in a gas-solid fluidized bed made of normal glass with 26 mm inner diameter and 800 mm height. High-purity nitrogen (99.98 %) at room temperature was entered into the column through a sintered porous distributor and alcohol vapor was added to nitrogen before it entered the bed using a bottle filled with alcohol. The gas flow leaving the system was cleaned using a HEPA filter to prevent the elutriation of NPs to the pH measuring solution system.
The bed expansion (H/H0) at different superficial gas velocities was studied to interpret the fluidization behavior of the particles. H is the height of the bed at a certain superficial gas velocity and H0is the height of the bed at zero gas velocity. After each change in gas velocity, 3 minutes waiting time was taken for the bed to stabilize before the bed expansion was measured. All the tests were carried out at room temperature and ambient pressure conditions.
The original sample; 1.69 g of Ca(OH)2, and also modified sample with the best fluidization behavior; 1.69 g of Ca(OH)2 with 0.56 g of Aerosil 130 (i.e., sample SiO2 25 wt%), were tested in CO2 adsorption experiment. We used a small inlet concentration of purified CO2 in the gas flow mixture (volume concentrations 2.5% CO2+98% N2). The N2 gas leaving the bubbler and CO2 gas were sufficiently mixed in a gas mixer at a constant temperature. Before adsorption was conducted, most parts of the experimental apparatus, including the gas line, empty space of the adsorber, and even the adsorbents themselves, were washed by N2 purging. To evaluate the adsorption, exhaust gas from the fluidized bed continuously entered into the tank contains 120 ml water through a sparser. pH variations of the solution during the adsorption of CO2 was measured every 5 s using a pH meter (ZAG CHEMIE, model PTR79). A stirrer bar was constantly rotated at the speed of 180 rpm in the water to ensure uniformly distributed absorption of exhaust CO2 in the water. The adsorption completion point was determined on the basis of converging pH to a constant value which was regarded as the standard solubility of CO2 in water at 25 oC. The data are average values obtained in three runs at each experiment.
2.3 Taguchi Experimental Design
The Taguchi method was applied as an experimental design to achieve high quality of fluidization because it provides a simple, efficient and systematic approach to optimize operating conditions under designated ranges of all selected parameters 24, 25. The working steps for the Taguchi experimental design include (1) identification of the quality characteristics and selection of design parameters; (2) determination of the number of levels for the design parameters and possible interactions between the design parameters; (3) selection of the appropriate orthogonal array and arrangement of operation parameters to the orthogonal array; (4) conducting experiments based on the arrangement of the orthogonal array; (5) analysis of the experimental results using the signal-to-noise (S/N) ratio and variance analyses; (6) recognizing the optimal experimental conditions (levels) among those listed in the orthogonal table; and (7) verification of the optimal design parameters through the confirmation experiment 26, 27. As shown in Table 2, four parameters (with their testing ranges), including SiO2 weight percentage (10, 15, 20, 25, 30 and 35%), silica NPs sieved size (300, 600 and 850 micron), Ca(OH)2 powder sieved size (75, 106 and without sieving, i.e., as received) and type of alcohol (methanol, ethanol and 2-propanol) were selected in this study.
There are many different possible S/N ratios, three of them are considered standard and are generally applicable in the most situations: larger the better (LTB), small the better (STB) and nominal the better (NTB) 28. In this study, the larger the better response category of Taguchi experimental design method was used to determine optimum fluidization conditions for maximizing the ratio of bed height to initial bed height, namely bed expansion (H/H0). The larger the better performance statistics is given by 29,30:
S/N = -10*log (mean square deviation) =
-10 log10(1n∑i=1n1yi2) (1)
where S/N ratio is the performance characteristics as “the larger the better”, y is the comparison variable in experiment i for a certain combination of control factor levels where n is the number of experiments performed for that combination. The Taguchi L18(61 ×33), orthogonal mixed fractional factorial array (six levels for parameter A and three levels for the other B, C and D parameters) was selected in order to find out the optimum levels of parameters.
- Results and discussion
3.1 Analysis of the S/N ratio
The collected data were analyzed by using Minitab ®17 Statistical Software for the evaluation of the effect of each parameter on the optimization criteria. Table 3 shows the bed expansion results for each run of the selected orthogonal arrayin the gas velocity of 0.025 m/s and also the corresponding S/N ratios obtained from Eq (1). Since the experimental design is orthogonal, it is then possible to separate out the effect of each parameter at different levels. For example, the mean S/N ratio for the SiO2 wt% parameter at levels 1, 2, 3, 4, 5 and 6 can be calculated by averaging the S/N ratios for the experiments 1–3, 4–6, 7–9, 10–12, 13–15 and 16–18, respectively. The mean S/N ratio for each level of the other parameters can be computed in the similar manner. The mean S/N ratio for each level of the parameters is summarized and called the mean S/N response table for bed expansion (Table 4). The trend of the resultant mean S/N ratios for selected parameters B, C, D at the three designated levels and selected parameter A at the six designated levels is shown in Fig. 2. Both parameters of SiO2 wt% and Ca(OH)2 sieved size share the same trend in their resultant S/N ratios (i.e., first increases and then decreases) which is different from that of SiO2 sieved size (i.e., always decreases) and type of alcohols (i.e., always increases). Based on this figure, the combination of A4 (= 25 wt%), B1 (= 300 μm), C2 (=106 μm), and D3 (= 2-propanol) were found with the highest S/N ratio for each of the four selected parameters, and hence is considered as the optimum condition for increasing the bed expansion.
3.2 Analysis of variance (ANOVA)
The knowledge of the contribution of individual parameters is critically important for the control of the final response. In this regard, ANOVA was performed to identify significant factors affecting bed expansion (H/H0) and the results are shown in Table 5. In the ANOVA, the Fischer ratio (or F-ratio) which is a statistically valid measure of how well the factors describe the variation in the data about its mean, can be calculated by dividing the mean of the squared deviations to the mean of the squared error. In addition, the minimum or critical values for the Fischer ratio (F-cr) can be found in most of the statistics and experimental design handbooks 31. A F-ratio is calculated from the experimental results and then compared to the critical value. If the calculated F-ratio is bigger than the F-cr value, it is an indication that the statistical test is significant at the confidence level selected. As it can be seen in Table 5, the F-cr value of parameter A is 4.39 at a confidence level of 95% and it is 5.14 for B, C and D parameters at a confidence level of 95%. Comparison between the calculated F-ratio and the F-cr values indicates that weight percentage of SiO2 NPs (A) and also their sieved size (B) are two significant parameters affecting bed expansion, whereas Ca(OH)2 sieved size (C) and type of alcohols (D) do not have significant influence on experiments. The percentage contribution (P (%)), which is defined as a ratio of the parameters’ corrected sum of square to the total sum of square, also indicates the contribution of parameters. For example, the value of P (%) for SiO2 NPs sieved size given in Table 5 is calculated as follows:
PB=SSB-DOFB(MSerror)SStotal×100=5.09-2×0.036.89×100=72.87
The results of P(%) values also confirm that parameters B and A have the most significant effect on bed expansion of modified adsorbents, respectively.
3.3 Experimental observations
In order to find the mainreason of difference between fluidization behavior of various samples, two experimental runs with trial numbers 13 and 15 based on the L18 orthogonal array (see Table 3) were selected to further study. These two cases show the best and worst fluidization behaviors, respectively, which is clearly appreciated in the photographsof their fluidized beds (Fig. 3a and b). SEM pictures of both cases are also presented in Fig. 3c and d to see the distribution and coating quality of particles of each sample. As can be seen, in the case of experiment RUN 13, silica NP aggregates have become uniformly coated with the Ca(OH)2 fine particles which is indicative of materials mixing homogeneity (Fig. 3c). While in the case of experiment RUN 15, Ca(OH)2 and silica NP aggregates are seen more as separate and uncoated which is indicative of poor coverage of the materials over each other (Fig. 3d). For modified samples, the bed expansion ratio in the case of experiment RUN 13 and 15 reaches up to about 3.25 and 1.25, respectively, in the gas velocity of 0.05 m/s. As it obvious, Fig. 3a shows a homogeneous fluidization behavior with the type of APF, whereas Fig. 3b shows a non-uniform fluidization with stable channels and cracks through which most of the gas flow bypass the bed. Therefore, it can be concluded that by a proper modification of Ca(OH)2 fine powder with silica NPs, its fluidization behavior will be homogenized, which, as will be shown, yields a substantial increase of the gas-solids contact efficiency and a higher CO2 adsorption rate. In this regard, the effect of all previously mentioned parameters on fluidization quality of Ca(OH)2 + Aerosil 130 adsorbent will be discussed experimentally in the following.
3.3.1 Effect of the alcohols type
It is believed that the use of a fluidizing gas that includes vapor of a polar solvent, like used alcohols here, will be effective in dissipating electrostatic charge within the fluidization chamber of hydrophilic NPs. This phenomenon happens by binding the polar section of alcohols to the surface of hydrophilic NPs (e.g., Aerosil 130). Through this bonding, alcohol molecules expose their organic groups result in decreasing the interaction between silica NPs and consequently friction between these molecules. That is the main reason of increased bed expansion of hydrophilic silica NPs in the presence of alcohols 32. In order to investigate the effect of different alcohols on the fluidization behavior of modified absorbents, number of experiments were carried out in the presence of methanol, ethanol, 2-propanol and 2-butanol. We chose Ca(OH)2 + Aerosil 130 (25 wt%) modified adsorbent to further study. The results of the relative bed expansion during fluidization of this adsorbent in the presence of used alcohols are shown in Fig. 4. As can be seen, the bed expansion of original adsorbent; Ca(OH)2 in the absence of silica additive, is very low and the bed expands up to about 1.4 times the initial bed height at gas velocity of 0.05 m/s. This low expansion of Ca(OH)2 particles is attributed to the strong cohesive forces between them.In a similar way,the bed expansion of modified adsorbent in the absence of alcohols is very low and the bed expands up to only about 1.9 times the initial bed height at gas velocity of 0.05 m/s. However, it is shown that the bed expansion, which is an indicative of the fluidization homogeneity, increases with adding alcohols vapor to the system. In the case of using methanol, ethanol, 2-propanol and 2-butanol, the bed expansion ratio reaches up to about 3.4, 3.2, 3.1 and 2.5 respectively, in the gas velocity of 0.05 m/s. Therefore, it seems that the influence of alcohols still remains important even when the particles of Ca(OH)2 are mixed with hydrophilic silica NPs. When powders are mixed, porous silica agglomerates become coated by a monolayer of Ca(OH)2 particles. In this way, the easily fluidizable hydrophilic silica agglomerates in the presence of alcohols become carriers of the Ca(OH)2 fine particles and result in better fluidization behavior. As can be seen in Fig. 4, methanol, ethanol and 2-propanol are the most effective alcohols in enhancing the fluidization quality of modified sample. It might be attributed to the better bonding of these alcohols to the surface of silica NPs 32.
We discussed how the alcohols affect the fluidization behavior of hydrophilic silica and alumina NPs by two parameters of surface potential (Ψ0) and dielectric coefficient (ε) in our last paper 32. These two parameters play principal role in controlling the agglomerates size and consequently their fluidization behavior by controlling the bonding intensity of alcohol molecules to the surface of hydrophilic NPs. It was shown that the more suitable alcohol in enhancing the fluidization behavior of mentioned NPs is the one that has the higher surface potential and also dielectric constant values. The highest Ψ0 values of the alcohols correspond to 2-propanol (~ -56), methanol (~ -44) and ethanol (~ -39), respectively 33. On the other hand, the highest ε values between the alcohols correspond to methanol (~ 28.42), ethanol (~ 20) and 2-propanol (~ 11.48), respectively 34. So, it is expected that the interaction between NPs by using these three alcohols will be lower than that by using 2-butanol (Ψ0 ~ -25, ε ~ 7.79). This is in good agreement with the experimental observations in Fig. 4. Finally, it should be noted that due to the poor flowability of modified adsorbent in the presence of 2-butanol, this alcohol was not considered in the Taguchi experimental design array.
We also used Richardson–Zaki (R–Z) equation to interpret the difference between fluidization behaviors of adsorbents in the presence of different alcohols. This equation relates the superficial gas velocity U with the bed voidage and the terminal velocity for an agglomerate Ut as:
U=Utεbn
(2)
Which can be rewritten as the following linear equation:
log U=log Ut+nlogεb
(3)
Where n is the index of R–Z equation and it is believed to be an indicator for the characteristic of particles fluidization 35,
εbis the bed voidage and can be given by mass balance of the particle in the fluidized bed as:
εb=1-H0H(1-εb0)
(4)
The initial bed voidage (εb0) of single SiO2 NPs is within the range from 0.2 to 0.25 36 and εb0 = 0.22 was chosen for the calculations. The value of εb0 for pure Ca(OH)2 particles for which bed exhibits higher bulk and particle densities than that of SiO2 NPs, is assumed to be 0.14 35, 36. For the Ca(OH)2 + Aerosil 130 (25 wt%) modified adsorbent bed, εb0 will be calculated as 0.16 by using the relation of
0.25×ε0 SiO2+0.75×ε0 CaOH2. By drawing a plot of lg U vs. lg
εbfor original and modified adsorbents in the presence of alcohols (see Fig. 5), the R–Z index n can be obtained. All the linear fitting trend lines of the experimental data shown in this figure has the correlation coefficients above 0.97, so the fluidized systems obey the R–Z equation very well. Nam et al. 37have shown that a R–Z exponent of n = 5 is valid for APF behavior of NP agglomerates. Therefore, it can be concluded that those modified adsorbent-alcohol systems which have index n ≥ 5 show APF behavior and those with n < 5 present ABF behavior. Type of fluidization of original and also Ca(OH)2 + Aerosil 130 (25 wt%) modified adsorbents in the absence and presence of different alcohols are presented in Table 6 based on index n quantities. According to this table, modified adsorbent in the presence of methanol, 2-propanol and ethanol has the index n more than 5 which means the fluidization behavior is as APF. Whilethis value for original adsorbent and also modified adsorbent in the presence of 2-butanol is less than 5 which shows the behavior of ABF. These results confirm the experimental observations (see Fig.4) and Taguchi predicted results very well.
3.3.2 Effect of Ca(OH)2 powder sieved size
In order to investigate the effect of Ca(OH)2 sieved size on the fluidization behavior of modified adsorbents, number of experiments were carried out with different sieved sizes of Ca(OH)2 powder (as received, 75 and 106 micron). The modified sorbents were prepared by adding 25 wt% of Aerosil 130 with the sieved size of 300 micron to Ca(OH)2 samples. All the experiments were carried out in the presence of methanol. The results of bed expansion for the mentioned adsorbents during fluidization at different superficial gas velocities are shown in Fig. 6.As it is depicted in this figure, there is no any sensible difference between the results and diagrams of all cases (specially two cases of 75 and 106 micron) are very similar. Despite the negligible difference between the results, the ratio of the bed height to the initial height for the modified sorbent contains Ca(OH)2 without any sieving is lower to some extent.This lower ratio could be due to the presence of Ca(OH)2 agglomerates which might not be very well distributed on the agglomerates of silica NPs. The obtained results are also confirmed by the R–Z model; as all drawn graphs (lg U vs lg εb) shows almost the same slope (figure not shown) and consequently the calculated index n values for all three samples are very close to each other (i.e., 5.14, 5.2 and 5.3 for adsorbents contain as received, 75 and 106- micron Ca(OH)2 powder, respectively). Finally, it should be noted that based on the results of variance analysis, Ca(OH)2 powder sieved size (parameter C) is introduced as an insignificant parameter on the fluidization quality of modified sorbents (see Table 5) which is in consistent with the experimental observations.
3.3.3 Effect of SiO2 NPs sieved size
The results of variance analysis showed that the sieved size of SiO2 NPs has the most significant effect (P (%) ~ 72.87) on the fluidization quality of modified sorbents. To investigate this effect, fluidization experiments were carried out for samples modified by adding 25 wt% of Aerosil 130 with three different sieved sizes (300, 600 and 850 micron). The sieved size of Ca(OH)2 powder was kept as 106- micron and all the experiments were performed in the presence of methanol. The results of bed expansion for the prepared adsorbents during fluidization at different superficial gas velocities are shown in Fig. 7.As shown in this figure, the bed expansion ratio of adsorbents modified by300, 600 and 850-micron silica NPsreaches up to about 3.4, 2.1 and 1.7, respectivelyin the gas velocity of 0.05 m/s. This behavior is in agreement with original Rumpf model based on which the flowability improvement is inversely proportional to the guest particle size for a given host particle size 38. The better fluidization behavior of the adsorbents modified by 300 micron SiO2 NPs may be attributed to the formation of smaller silica agglomerates. Because smaller aggregates become coated more uniformly with the Ca(OH)2 fine particles which is indicative of materials mixing homogeneity (see Fig. 3c). In addition, it is clear that alcohol could play its role more effectively on smaller agglomerates of silica carrier particles to make them better fluidized. On the other hand, Ca(OH)2 and SiO2 NP aggregates are seen more as separate and uncoated in the case of adsorbents modified by bigger silica NPs (850 micron) which is indicative of poor coverage of the materials over each other (see Fig. 3d). Therefore, alcohol cannot be able to play an appropriate role in improving their fluidization behavior. The index n values calculated from the R-Z model are also in agreement with the obtained results as this index decreases by increasing the silica NPs sieved size; 5.30, 3.32 and 2.68 for samples prepared by 300, 600 and 850-micron silica NP additives, respectively.
3.3.4 Effect of weight percentage of SiO2 NPs
The last parameter which was studied is the weight percentage of added silica NPs to the modified adsorbents. Based on the variance analysis, this parameter has the contribution of 16.74% on improving the fluidization quality of Ca(OH)2 particles. The results of the relative bed expansion during fluidization of Ca(OH)2 adsorbents modified by Aerosil 130 NPs with different weight percentages are shown in Fig. 8. The sieved size of SiO2 NPs and Ca(OH)2 powder was kept as 300 and 106 -micron, respectively, and all the experiments were performed in the presence of methanol. As can be seen, the expansion of the fluidized bed increases gradually when the weight percentage of SiO2 NPs increases from 10 to 25%. By further increasing, the bed height declines somewhat and then remain relatively constant. Therefore, it can be concluded that an optimum concentration of added hydrophilic SiO2 NPs is required to reach the maximum bed expansion. The weight percentage of 25 was selected in this work as the optimum SiO2 concentration value for the improvement of Ca(OH)2 fluidizability which is in accordance with the Taguchi prediction. For this sample (Ca(OH)2 + Aerosil 130 (25 wt%)), the fluidized bed reaches a height larger than 3.5 times the initial height of the settled bed, in the gas velocity of 0.05 m/s, while gas channels or large bubbles are not seen during fluidization. The same result is obtained by the R–Z model. As shown in Fig.9 and Table 7, the R–Z index n for modified adsorbents increases with increasing the weight percentage of SiO2 NPs up to 25 wt% and then with a slight decline remains relatively constant. According to Table 7, the sample of (Ca(OH)2 + Aerosil 130 (25 wt%)) has the highest index n = 5.30and the sample of (Ca(OH)2 + Aerosil 130 (10 wt%)) has the lowest value n = 2.64 among all modified adsorbents. It should be noted that Valverde et al. 2 observed the similar behavior while studying the flowability behavior of Ca(OH)2 adsorbent modified by hydrophobic silica NPs (Aerosil R974) (see Fig. 10). As can be seen in Fig. 10, the expansion of the bed increases with the weight percentage of Aerosil R974. In addition, there is not a significant difference between the final bed height of samples contain Aerosil R974 (20 wt%) and Aerosil R974 (30 wt%). These researchers reported weight percentage of 20 as the optimum SiO2 concentration value to achieve the best fluidization quality of modified adsorbents. For the sample of (Ca(OH)2 + Aerosil R974 (20 wt%)), the bed expands up to about 2 times the initial bed height, in the gas velocity of 0.03 m/s, while the expansion ratio is about 3.2 for the sample of (Ca(OH)2 + Aerosil 130 (25wt %)), in the same gas velocity. This observation shows the considerable importance of using hydrophilic SiO2 NPs rather than their hydrophobic counterparts to modify Ca(OH)2 adsorbent in order to improve its fluidizablity.
Finally, the SEM picture of sample (Ca(OH)2 + Aerosil 130 (25 wt%)); presented as Fig. 11, reveals the distribution and coating quality of the materials over each other. As can be obviously seen, Aerosil 130 aggregates; which are easily fluidizable in the presence of methanol, have become uniformly coated with the Ca(OH)2 fine particles and this is indicative of materials mixing homogeneity and hence high fluidization quality.
3.3.5 Confirmation test
Confirmation test is a crucial step recommended by Taguchi to verify experimental conclusions. Once the optimal level of the design parameters has been selected, the final step is to predict and verify the improvement of the quality characteristic using the optimal level of the design parameters. Therefore, a final fluidization test is conducted at different gas velocities to verify the bed expansion at optimum level of A4 B1 C2 D3. The results of these experiments along with their corresponding predictive values by the Taguchi are presented in Table 8. As can be seen, the experimental values agree reasonably well with predictions ones.
3.4 CO2 adsorption tests by modified absorbents
As mentioned before, we used the method of measuring pH variation of pure water during adsorption in order to study the CO2 adsorption capacity of adsorbents. Reactions that occur during the absorption of CO2 in the pure water are listed as eqs 5-8:
CO2 (g)
↔CO2(aq) (5)
CO2 (aq) + H2O (l)
↔H2CO3 (aq (6)
H2CO3 (aq)
↔H+(aq) + HCO3–(aq) (7)
HCO3–(aq)
↔H+(aq) +CO32-(aq) (8)
These reactions lead to a decrease in pH until CO2 saturation. Fig.12 shows the experimental result of the CO2 absorption into pure water at 25 0C. This experiment was carried out by passing the mixing gas (2.5% CO2+98% N2) through the empty bed continuously. As can be seen, the initial pH of water is 6.45 and proceeding the eqs 5-8 makes the final pH in the CO2-saturated pure water as approximately 4.75. This final constant value is regarded as the standard solubility of CO2 in atmospheric pressure and experiment temperature. It should be mentioned that pH starts to decrease right after starting the experiment and the adsorption time, i.e., the point at which the pH becomes constant, is only 200 s.
To compare the adsorption capacity of modified adsorbents, three different samples are selected to be tested in CO2 adsorption experiment. The selected samples are named as; sample 1: Ca(OH)2 modified by Aerosil 130 prepared at optimum condition (level of A4 B1 C2 D3), sample 2: Ca(OH)2 modified by Aerosil R974 prepared at the same condition and sample 3: unmodified Ca(OH)2 with the sieved size of 106 -micron. Fig.13 Shows the adsorption behavior when the bed fluidized with these three samples. For better comparison, the result of CO2 absorption in pure water when the fluidized bed was empty is also added to the figure. As can be seen, when the unmodified Ca(OH)2 powder (sample 3) is used as an adsorbent, pH changes right after starting the test which shows detecting of CO2 in the effluent gas from the fluidized bed. In addition, pH becomes constant after about 200 s in this case. Therefore, it seems that the behavior of CO2 absorption in water for sample 3 is very similar to the case of there is no any adsorbent in the bed. These results have a straightforward explanation on the basis of powder fluidizability (see Fig. 14). The bypass taken by most of the gas through channels in the Ca(OH)2 fluidized bed reduces the effective gas–solids contact surface and increases the local interstitial gas velocity, which decreases the adsorbent–CO2 contact time. Thus, CO2 rapidly appears in the effluent gas and changes the pH of pure water. Accordingly, based on the higher bed expansion ratio of sample 1 compared to sample 2 seen in Fig. 14, it might be predicted that the sample containing hydrophilic silica NPs has better CO2 adsorption behavior compared with the sample containing hydrophobic silica NPs. The adsorption behavior of these samples confirm the above mentioned prediction. Because as it can be seen in Fig. 13, pH starts to decrease at the longer times and the CO2 adsorption time is also longer for the sample containing Aerosil 130. In the other words, due to the effective contact surface and contact time of the adsorbent sample 1 with the fluidizing gas, CO2 emission from fluidized bed occurs at a longer time and until then pH stays unchanged and consequently saturation of pure water occurs at longer times. Finally, it may be suggested to use hydrophilic silica modified Ca(OH)2 adsorbent as an alternative for the hydrophobic silica modified one due to its higher amount of CO2 adsorption and also lower economical preparation costs.
- Conclusions
This paper presents an application of the Taguchi experimental design method in the optimization of characteristics parameters of Ca(OH)2+SiO2 modified adsorbent for enhancing its CO2 adsorption efficiency. The following conclusions can be drawn based on the experimental results of this study:
- Adding light agglomerates of hydrophilic/hydrophobic silica NPs to a Geldart C fine powder (Ca(OH)2 adsorbent) serves to catalyze CO2 adsorption in a fluidized bed.
- Based on the ANOVA, four parameters including SiO2 wt%, sieved size of SiO2 NPs, sieved size of Ca(OH)2 powder and type of alcohol have the percentage contributions of 16.74, 72.87, 2.38 and 0.72%, respectively in improving the flowability of Ca(OH)2 adsorbents modified by hydrophilic silica NPs.
- The resultant combination of SiO2 wt%(= 25), sieved size of SiO2 (= 300 μm), sieved size of Ca(OH)2 (= 106 μm), and alcohol (= 2-propanol) is considered as the optimum condition for increasing the bed expansion of Ca(OH)2 adsorbent modified by hydrophilic SiO2.
- By increasing fluidization homogeneity, the effective contact surface and contact time of the adsorbent with the CO2 are increased, which greatly increases the CO2 adsorption in fluidized bed.
- It is suggested to use Ca(OH)2 adsorbent modified by hydrophilic silica NPs as an alternative for the adsorbent modified by hydrophobic silica due to its higher amount of CO2 adsorption and also lower preparation costs.
- Two physical parameters of surface potential and dielectric coefficient of alcohols play principal role on controlling the agglomerates size of hydrophilic silica NPs and consequently fluidization behavior of particles.
References
1. Millán JMV. Fluidization of fine powders: cohesive versus dynamical aggregation. Vol 18: Springer Science & Business Media; 2012.
2. Valverde J, Pontiga F, Soria-Hoyo C, et al. Improving the gas–solids contact efficiency in a fluidized bed of CO 2 adsorbent fine particles. Physical Chemistry Chemical Physics. 2011;13(33):14906-14909.
3. Valverde J, Perejón A, Perez-Maqueda LA. Enhancement of fast CO2 capture by a nano-SiO2/CaO composite at Ca-looping conditions. Environmental science & technology. 2012;46(11):6401-6408.
4. Quintanilla M, Valverde J. Use of silica nanopowder to accelerate CO 2 sorption by Ca (OH) 2. Particuology. 2013;11(4):448-453.
5. Yu J, Zeng X, Zhang G, Zhang J, Wang Y, Xu G. Kinetics and mechanism of direct reaction between CO2 and Ca (OH) 2 in micro fluidized bed. Environmental science & technology. 2013;47(13):7514-7520.
6. Lin S-Y, Suzuki Y, Hatano H, Harada M. Hydrogen production from hydrocarbon by integration of water− carbon reaction and carbon dioxide removal (HyPr− RING method). Energy & Fuels. 2001;15(2):339-343.
7. Lin S, Harada M, Suzuki Y, Hatano H. Comparison of pyrolysis products between coal, coal/CaO, and coal/Ca (OH) 2 materials. Energy & Fuels. 2003;17(3):602-607.
8. J. M. Valverde AP, L. A. Perez-Maqueda2, F. J. Duran, F. Pontiga,C. Soria-Hoyo, M.A.S. Quintanilla, H. Moreno, M. J. Espin. Enhancement of CO2 capture capacity by a fine cohesive powder 2012.
9. Chen Y, Yang J, Dave RN, Pfeffer R. Fluidization of coated group C powders. AIChE journal. 2008;54(1):104-121.
10. Chirone R, Massimilla L, Russo S. Bubbling fluidization of a cohesive powder in an acoustic field. Fluidization VII: Engineering Foundation New York, USA; 1992:545.
11. Levy EK, Shnitzer I, Masaki T, Salmento J. Effect of an acoustic field on bubbling in a gas fluidized bed. Powder Technology. 1997;90(1):53-57.
12. Nowak W, Hasatani M, Derczynski M. Fluidization and heat transfer of fine particles in an acoustic field. Paper presented at: AIChE Symposium Series1993.
13. Jaraiz-M E, Levenspiel O, Fitzgerald T. The uses of magnetic fields in the processing of solids. Chemical Engineering Science. 1983;38(1):107-114.
14. Rosensweig RE. Fluidization: hydrodynamic stabilization with a magnetic field. Science. 1979;204(4388):57-60.
15. Dutta A, Dullea L. Effects of external vibration and the addition of fibers on the fluidization of a fine powder. Paper presented at: AIChE Symp. Ser1991.
16. Jaraiz E, Kimura S, Levenspiel O. Vibrating beds of fine particles: estimation of interparticle forces from expansion and pressure drop experiments. Powder technology. 1992;72(1):23-30.
17. Tang H, Zhao J. The agglomerating behavior of fine particles in a Vibro-fluidized bed. Chem. Ind. Eng. 1996;13(3):6-10.
18. Liu H, Li Y, Guo Q. Fluidization quality improvement for cohesive particles by fine powder coating. Industrial & engineering chemistry research. 2006;45(5):1805-1810.
19. Roßkopf C, Haas M, Faik A, Linder M, Wörner A. Improving powder bed properties for thermochemical storage by adding nanoparticles. Energy Conversion and Management. 2014;86:93-98.
20. Perez-Vaquero J, Valverde J, Quintanilla M. Flow properties of CO 2 sorbent powders modified with nanosilica. Powder technology. 2013;249:443-455.
21. Pontiga F, Valverde J, Moreno H, Duran-Olivencia F. Dry gas–solid carbonation in fluidized beds of Ca (OH) 2 and nanosilica/Ca (OH) 2 at ambient temperature and low CO 2 pressure. Chemical Engineering Journal. 2013;222:546-552.
22. van der Vegte EW, Hadziioannou G. Scanning force microscopy with chemical specificity: an extensive study of chemically specific tip− surface interactions and the chemical imaging of surface functional groups. Langmuir. 1997;13(16):4357-4368.
23. Tahmasebpoor M, de Martín L, Talebi M, Mostoufi N, van Ommen JR. The role of the hydrogen bond in dense nanoparticle–gas suspensions. Physical Chemistry Chemical Physics. 2013;15(16):5788-5793.
24. Fraley S, Oom M, Terrien B, Date J. Design of experiments via Taguchi methods: orthogonal arrays. The Michigan chemical process dynamic and controls open text book, USA. 2006;2(3):4.
25. Pourjavadi A, Amini-Fazl M, Ayyari M. Optimization of synthetic conditions CMC-g-poly (acrylic acid)/Celite composite superabsorbent by Taguchi method and determination of its absorbency under load. Express Polym. Lett. 2007;1(8):488-494.
26. Chen Y-C, Tsai P-J, Mou J-L. Determining optimal operation parameters for reducing PCDD/F emissions (I-TEQ values) from the iron ore sintering process by using the Taguchi experimental design. Environmental science & technology. 2008;42(14):5298-5303.
27. Yang Wp, Tarng Y. Design optimization of cutting parameters for turning operations based on the Taguchi method. Journal of materials processing technology. 1998;84(1):122-129.
28. Kaminari N, Schultz D, Ponte M, Ponte H, Marino C, Neto A. Heavy metals recovery from industrial wastewater using Taguchi method. Chemical Engineering Journal. 2007;126(2):139-146.
29. Engin AB, Özdemir Ö, Turan M, Turan AZ. Color removal from textile dyebath effluents in a zeolite fixed bed reactor: Determination of optimum process conditions using Taguchi method. Journal of hazardous materials. 2008;159(2):348-353.
30. Liao H-T, Shie J-R, Yang Y-K. Applications of Taguchi and design of experiments methods in optimization of chemical mechanical polishing process parameters. The International Journal of Advanced Manufacturing Technology. 2008;38(7-8):674-682.
31. Montgomery DC. Design and analysis of experiments: John Wiley & Sons; 2008.
32. Tahmasebpoor M, Ghasemi Seif Abadi R, Rahimvandi Noupoor Y, Badamchizadeh P. Model Based on Electrostatic Repulsion and Hydrogen Bond Forces To Estimate the Size of Nanoparticle Agglomerates in Fluidization. Industrial & Engineering Chemistry Research. 2016;55(50):12939-12948.
33. Barbagini F. A Fundamental Study of Particle-Substrate Interactions in Liquids of Low Polarity (Een fundamentele studie van deeltjes-substraat interacties in vloeistoffen met lage polariteit). 2009.
34. Haynes WM. CRC handbook of chemistry and physics: CRC press; 2014.
35. Liang X, Duan H, Zhou T, Kong J. Fluidization behavior of binary mixtures of nanoparticles in vibro-fluidized bed. Advanced Powder Technology. 2014;25(1):236-243.
36. Yang J, Zhou T, Song L. Agglomerating vibro-fluidization behavior of nano-particles. Advanced Powder Technology. 2009;20(2):158-163.
37. Nam CH, Pfeffer R, Dave RN, Sundaresan S. Aerated vibrofluidization of silica nanoparticles. AIChE Journal. 2004;50(8):1776-1785.
38. Yang J, Sliva A, Banerjee A, Dave RN, Pfeffer R. Dry particle coating for improving the flowability of cohesive powders. Powder Technology. 2005;158(1):21-33.
Table Captions
Table 1: The materials used in this work
Table 2: Parameters, codes, and level values used for orthogonal array
Table 3: Taguchi L18 orthogonal array, bed expansion results and the computed S/N ratios in the gas velocity of 0.025 m/s
Table 4: Mean S/N ratio for each level of the parameters; response table for bed expansion
Table 5: Results of ANOVA for bed expansion
Table 6: The values of n obtained from the R−Z equation for original Ca(OH)2 and modified Ca(OH)2 + Aerosil 130 (25 wt%) adsorbents fluidized in dry nitrogen and also in the presence of different alcohols
Table 7: The values of n obtained from the R−Z equation forCa(OH)2 (106 micron) + Aerosil 130 (300 micron) modified adsorbents with different weight percentage of SiO2 NPs fluidized in the presence of methanol
Table 8: Results of the confirmation test for bed expansion at optimum level of A4 B1 C2 D3
Figure Captions
Fig. 1: Experimental set-up
Fig. 2: Mean S/N ratios of the four selected parameters at their corresponding designated levels
Fig. 3: Bed expansion and SEM pictures of modified samples with the best and worst fluidization behavior a, c) experiment RUN 13 b, d) experiment RUN 15
Fig. 4: Bed expansion curves for original Ca(OH)2 and modified Ca(OH)2 + Aerosil 130 (25 wt%) adsorbents fluidized in dry nitrogen and also in the presence of different alcohols
Fig. 5: lg U vs lg εb for original Ca(OH)2 and modified Ca(OH)2 + Aerosil 130 (25 wt%) adsorbents fluidized in dry nitrogen and also in the presence of different alcohols
Fig. 6: The effect of Ca(OH)2 powder sieved size on the bed expansion of modified adsorbents + Aerosil 130 (25 wt%) fluidized in the presence of methanol
Fig. 7: The effect of SiO2 NPs sieved size on the bed expansion of modified adsorbents + Aerosil 130 (25 wt%) fluidized in the presence of methanol
Fig. 8: The effect of weight percentage of SiO2 NPs on the bed expansion of Ca(OH)2 (106 micron) + Aerosil 130 (300 micron) modified adsorbents fluidized in the presence of methanol
Fig. 9: lg U vs lg εbforCa(OH)2 (106 micron) + Aerosil 130 (300 micron) modified adsorbents with different weight percentage of SiO2 NPs fluidized in the presence of methanol
Fig. 10: The comparison between bed expansion results of Ca(OH)2 + Aerosil 130 (this work) and Ca(OH)2 + Aerosil R974 2 modified adsorbents
Fig. 11: The SEM picture of Ca(OH)2 + Aerosil 130 (25 wt%) modified adsorbent
Fig.12: Variation of pH in pure water according to CO2 adsorption time in an empty bed
Fig.13: Variation of pH in pure water according to CO2 adsorption time when bed fluidized with/without different sample adsorbents
Fig.14: Bed expansion curves for different sample adsorbents
Fig. 2: Mean S/N ratios of the four selected parameters at their corresponding designated levels
Fig. 3: Bed expansion and SEM pictures of modified samples with the best and worst fluidization behavior a, c) experiment RUN 13b, d) experiment RUN 15
Fig. 4: Bed expansion curves for original Ca(OH)2 and modified Ca(OH)2 + Aerosil 130 (25 wt%) adsorbents fluidized in dry nitrogen and also in the presence of different alcohols
Fig. 5: lg U vs lg εb for original Ca(OH)2 and modified Ca(OH)2 + Aerosil 130 (25 wt%) adsorbents fluidized in dry nitrogen and also in the presence of different alcohols
Fig. 6: The effect of Ca(OH)2 powder sieved size on the bed expansion of modified adsorbents + Aerosil 130 (25 wt%) fluidized in the presence of methanol
Fig. 7: The effect of SiO2 NPs sieved size on the bed expansion of modified adsorbents + Aerosil 130 (25 wt%) fluidized in the presence of methanol
Fig. 8: The effect of weight percentage of SiO2 NPs on the bed expansion of Ca(OH)2 (106 micron) + Aerosil 130 (300 micron) modified adsorbents fluidized in the presence of methanol
Fig. 9: lg U vs lg εb forCa(OH)2 (106 micron) + Aerosil 130 (300 micron) modified adsorbents with different weight percentage of SiO2 NPs fluidized in the presence of methanol
Fig. 10: The comparison between bed expansion results of Ca(OH)2 + Aerosil 130 (this work) and Ca(OH)2 + Aerosil R974 2 modified adsorbents
Fig. 11: The SEM picture of Ca(OH)2 + Aerosil 130 (25 wt%) modified adsorbent
Fig.12: Variation of pH in pure water according to CO2 adsorption time in an empty bed
Fig.13: Variation of pH in pure water according to CO2 adsorption time when bed fluidized with/without different sample adsorbents
Fig.14: Bed expansion curves for different sample adsorbents
* Corresponding author, Maryam Tahmasebpoor; Tel: +98 41-33392936; fax: +98 41-33340191; E-mail: tahmasebpoor@tabrizu.ac.ir
