Textural data obtained by nitrogen (77 K) and CO2 (273 K) isotherms, using BET and DFT methods.
The chemical modification of CWZ-22—commercial activated carbon (AC) with KOH‐ to enhance CO2 adsorption was examined. The effect of different impregnation ratios KOH:CWZ-22 from 1 to 4 was studied. The ACs were characterized by CO2 and N2 sorption, Fourier transform infrared (FTIR), SEM, and XRD methods.
- carbon dioxide
- activated carbon
- CO2 adsorption
The emission of CO2 originating mainly from industry has become a worldwide problem responsible for the global warming. The combustion of fossil fuel in power plants remains the main point source for CO2 emission to the atmosphere. Reduction in the CO2 concentration in the atmosphere is currently a hot topic [1, 2]. Among technologies proposed for reduction in CO2 emissions, adsorption is considered as a very promising process for CO2 capture. Alternatively, solid-based adsorbents have drawn substantial attention for CO2 capture.
Nowadays, many types of porous materials have been used in CO2 adsorption, such as zeolites , metal–organic frameworks (MOFs) , porous silica , and activated carbons (ACs) [2, 6–9]. Among these adsorbents, activated carbon (AC) has drawn great attention recently because of its high adsorption capacity, low cost, availability, large surface area, an easy-to-design pore structure, hydrophobicity (insensitiveness to moisture), and low energy requirements for regeneration [10, 11]. The adsorption performance of activated carbons depends on the selection of carbon sources and activation conditions. It has been reported that the activated carbons prepared with commercial carpet , eucalyptus sawdust , yeast , palm shells , peanut shell , pitch , and molasses  had high adsorption capacity for CO2.
Activated carbon can be mainly prepared by physical and chemical activation methods or by combination of both types of methods. Usually, physical activation is carried out using carbon dioxide, steam, air, or their mixture. Chemical activation involves agents such as zinc chloride , acids , and bases [10, 17]. KOH is one of the most widely used chemicals for activating the carbonaceous materials in preparation of activated carbon [1, 2, 8, 10]. Activated carbons obtained by chemical activation often possess a high specific surface area and well-developed micropores, which make them attractive materials for CO2 adsorption. Particularly, KOH activation has been applied in the preparation of activated carbons because it can produce lots of micropores favorable for CO2 adsorption. Therefore, the textural properties of the activated carbons depend on a type of carbon sources and the required amount of KOH for the preparation of efficient activated carbon to adsorb CO2.
Our motivation was to increase the porosity and CO2 adsorption capacity of commercial activated carbon CWZ-22. The increase of the CO2 adsorption on commercial activated carbon modified using KOH has not been yet described.
2. Experimental method
2.1. Materials and sample preparation
A commercial activated carbon CWZ-22 (Gryfskand Sp. z o.o. Hajnówka, Poland) was used as the raw material in this work. The CWZ-22 samples were mixed with the saturated KOH solution. The mass ratio of KOH:CWZ-22 was varied from 1 to 4. The soaking time was 3 h. The mixtures were dried at 200°C. The impregnated sample was activated at temperature of 800°C for 1 h under nitrogen flow. Then, the samples were washed repeatedly with a 5 M solution of HCl and with distilled water until they were free of chlorine ions. Finally, these samples were dried at 200°C for 12 h. The materials were named as KOH:CWZ-22 = X, where X is the mass ratio of KOH:CWZ-22.
2.2. Characterization and adsorption analyses
The textural properties of the ACs were determined by physical adsorption of N2 at 77 K and CO2 at 273 K using a Quadrasorb apparatus (Quantachrome Instruments). Before the experiments, the samples were outgassed under vacuum at 250°C overnight. The specific surface area was measured by the multipoint BET (Brunauer–Emmet–Teller) method. The total pore volume, Vp, which includes both the micropores and the mesopores, was estimated from the amount of nitrogen adsorbed at the highest relative pressure. The micropore volume with diameter in the range of 0.31–1.47 nm was calculated by applying the NLDFT model for CO2 adsorption at 273 K. Based on N2 adsorption isotherm, the volume of micropores larger than 1.5 nm was obtained by applying the QSDFT method for the slit/cylinder pore model using the software provided by Quantachrome.
ACs were also analyzed by Fourier transform infrared (FTIR) on a Nicollet 380 (Thermo Scientific) spectrometer in order to identify the functional groups on the surface of the ACs. The CWZ-22 carbons were mixed with KBr. FTIR spectra were recorded within a range of 400–4000 cm−1. The equipment was run prior to each actual measurement to record a background spectrum, which was then automatically subtracted from the spectrum of each analyzed sample.
The crystal structures of the samples were determined by XRD with the PANalytical Empyrean X-ray diffractometer using a Cu Kα radiation (α = 1.5418A°) at room temperature.
The surface morphology of the materials was performed by ultra-high resolution field emission scanning electron microscope (UHR FE-SEM Hitachi SU8020) equipped with the secondary electron (SE) detectors.
CO2 adsorption isotherms were measured up to 2 bar at 25, 40, 60, 80, and 100°C temperature using the volumetric Sieverts’ apparatus (IMI—Hiden Isochema Corporation). The CO2 adsorption data were fitted to standard isotherm models such as Langmuir, Freundlich, Sips, and Toth.
where q is the adsorbed quantity (mmol of CO2 per g of AC), p is the pressure of CO2 in the bulk gas phase, qm is the maximum adsorption capacity (mmol g−1), and bL is the Langmuir adsorption affinity constant (bar−1).
where kF is the Freundlich constant, and nF is the Freundlich heterogeneity factor. Value of nF > 1 represents favorable adsorption. The higher nF value implies higher sorption capacity, such as forming new adsorption sites.
where bs is the Sips adsorption affinity constant, and ns is the Sips heterogeneity factor.
The Toth isotherm is a model useful in describing heterogeneous adsorption systems, which satisfies both low- and high-end boundaries of the concentration. The Toth’s isotherm is expressed as Eq. (4) [20, 22, 23]:
where bT is Toth adsorption affinity constant, and nT is the Toth heterogeneity factor.
Isotherm parameters were obtained through a nonlinear fit of experimental data to the model equations using a MatCad software. In order to evaluate the fit of an isotherm to experimental equilibrium data, algorithms based on the Levenberg–Marquardt were used.
In this study, five nonlinear error functions were examined. In each case, a set of isotherm parameters were determined by minimizing the respective error function across the pressure range studied. The error functions employed were as follows: the correlation coefficient (R2) , the sum of the squares of errors (SSE) , the hybrid fractional error function (HYBRID) , Marquardt’s percent standard deviation (MPSD) , and the average relative error (ARE) .
Isosteric heats of adsorption (Qi) can be calculated using Clausius–Clapeyron Eq. (5):
where ln(p) is the natural logarithm of the pressure at specific surface loading (θ), Qi is the isosteric heat of adsorption, R is the universal gas constant (8.314⋅10−3 kJ/mol⋅K), and T is the temperature [K]
The value of isosteric heats of adsorption was determined from the slop of ln(p) versus 1/T.
3. Results and discussion
3.1. Micropore size distribution of activated carbons
Figure 1 shows adsorption–desorption isotherms of N2 at 77 K for the CWZ-22 and CWZ-22 modified by different amounts of KOH.
All the isotherms present a high adsorption at low relative pressure, characteristic for microporous materials. The samples from KOH:CWZ-22 = 2–4 ratio showed the higher nitrogen adsorption capacity, whereas for the KOH:CWZ-22 = 1 and CWZ-22, nitrogen adsorption capacity was considerably lower. In addition, all samples showed hysteresis loop at relative pressure >0.4 indicating the presence of certain mesoporosity in the samples. The hysteresis loop type H4 was recognized, which is characteristic of slit shaped pores, such as those present in typical activated carbons. Thus, the isotherms of all samples appear to become a combination of types I and IV according to IUPAC classification. For the isotherms of AC with the highest impregnation ratio, the shape moves toward a type IV isotherm. This can be attributed to a formation of a greater fraction of mesoporosity, which is proved by the mesopore volume (Table 1). Table 1 shows the textural properties of the activated carbons prepared with different KOH:CWZ-22 ratios. Both specific surface area and total pore volume of the samples increased with increasing KOH:CWZ-22 ratios. When the KOH:CWZ-22 ratios are increased from 1 to 2 the specific surface area, total pore volume increases from 1093 to 1292 m2 g−1 and from 0.59 to 0.75 cm3 g−1, respectively. However, further increase of KOH:CWZ-22 ratios only results in slow increase in specific surface area and pore volume. A similar situation can be observed in the case of micropores development. This may be due to the effect of the increase in small pores and the decomposition of porous structure.
|Vmic (N2) 1.2–2 nm|
|Vmic (CO2) 0.3–1.5 nm|
|KOH:CWZ-22 = 1||1093||0.59||0.34||0.31||0.25|
|KOH:CWZ-22 = 2||1292||0.75||0.38||0.32||0.37|
|KOH:CWZ-22 = 3||1256||0.73||0.38||0.32||0.35|
|KOH:CWZ-22 = 4||1299||0.77||0.40||0.37||0.37|
CO2 and N2 adsorption isotherms are commonly used to determine the pore size distribution of activated carbons. The pore size distribution is shown in Figure 2. All activated carbons have peaks in almost the same pore size ranges. One peak is located in the range of 0.3–0.4 nm, the second one in the range of 0.45–0.7 nm, the next in the range of 0.7–0.9 nm, and then in the range of 2.1–4.0 nm. The micropores are dominant. The volume of mesopores (pores above 2 nm) is low. The pores with a diameter of more than 5 nm were no found. The pore volume of KOH:CWZ-22 = 4 at the pore size of above 0.45–0.7 nm and in the range of 0.7–0.9 nm is larger than for other carbons, indicated on a better developed microporous structure.
3.2. Surface morphology of activated carbon
The SEM images of the pristine CWZ-22 and modified CWZ-22 are shown in Figure 3. The surface of CWZ-22 was relatively smooth without large cavities except for some occasional cracks or crevices. All modified CWZ-22 samples have external surfaces covered with irregular holes of different sizes and the irregular shapes. It can be seen from the SEM images that the external surface of the activated carbons was full of cavities. It seems that the cavities on the surface of the carbons resulted from the removal of active agents leaving empty space previously occupied by the active agent. Consequently, the reaction with KOH was aiding in the creation of the porous structure and was proved to be an effective activating agent for the production of high-surface area CWZ-22. Moreover, Figure 3b–d shows irregular surface of ACs, the grooves that are considered helpful for the accessibility of gases to the adsorbent surface.
3.3. X-ray diffraction studies
The XRD patterns of pristine CWZ-22 and of the samples with mass ratio from KOH:CWZ-22 = 1 to KOH:CWZ-22 = 4 are shown in Figure 4.
According to the XRD patterns, the position of peaks is similar for all carbon samples. There are only two peaks at about 23 and 43 2θ in the XRD pattern corresponding to the (002) and (101) planes of the hexagonal graphite (JCPDS, 75-1621). The peaks are very broad because of the random turbostratic stacking of a limited number of layers.
3.4. FTIR analysis
To characterize surface groups on activated carbon, FTIR (Fourier transform infrared) transmission spectra were obtained. The spectra of starting commercial CWZ-22 and modified CWZ-22 were presented in Figure 5.
The FTIR spectroscopy provides information on the chemical structure of materials. It can be observed in the Figure 6 that irrespectively of a KOH/AC ratio, the overall shape of the spectra is very similar. The band at around 3440 cm−1 is assigned to the O–H stretching of physisorbed water [27, 28] and that at 1650 cm−1—to a presence of physisorbed water as well as of chemisorbed CO2. All the spectra show a band in the region 2360–2344 cm−1 due to CO2 in air . The broader peaks about 1760 cm−1 are characteristic for C=O stretching vibration in carboxylic groups. The peaks around 1380 cm−1 can be assigned to the deformation vibrations of an H–C–OH group . The bands located around 2750 and 820 cm−1 could be assigned to C–H group.
The FTIR spectra of ACs modified KOH do not exhibit any differences compared to carbon CWZ-22, and these ACs present similar characteristics in their study by the FTIR technique. This suggests that the adsorption capacity of activated carbons CWZ-22 is not depending on the chemical reactivity of functional groups at surface.
3.5. Effect of KOH:CWZ-22 mass ratio of activated carbon on CO2 adsorption
Preliminary research on CO2 adsorption was performed at the temperature of 25°C and pressure up to of 1 bar. Chemical activation by KOH increased CO2 adsorption compared to pristine sample. The difference was small for KOH:CWZ-22 mass ratio 1–3, but when KOH:CWZ-22 was equal to 4, the CO2 adsorption was considerably higher.
3.6. CO2 adsorption and modeling of adsorption data
Two ACs samples were chosen for the modeling of adsorption data—the pristine sample and AC with the highest CO2 adsorption at the temperature of 25°C, namely KOH:CWZ-22 = 4.
The adsorption isotherms were measured at temperatures 25, 40, 60, 80, and 100°C up to the pressure equal to 2 bar. Adsorption isotherms of CO2 adsorbed on CWZ-22 and KOH:CWZ-22 = 4 at different temperatures are presented in Figures 7 and 8, respectively.
The amount of CO2 adsorbed on the KOH:CWZ-22 = 4 was higher than on the pristine carbon at all temperatures. The adsorption of CO2 on KOH:CWZ-22 = 4 at 2 bar and 25°C was of 38% higher than on the CWZ-22.
The Langmuir, Freundlich, Toth, and Sips models were employed to describe the adsorption data of CO2 onto the nonmodified and modified CWZ-22. The fitted adsorption isotherms are presented in Figures 7 and 8. Lines show the Langmuir, Freundlich, Toth, and Sips isotherm used to describe adsorption values. The experimental data were marked as symbols. The parameters of model isotherms of Langmuir, Freundlich, Toth, and Sips obtained from a nonlinear fit to experimental data at multiple temperatures for each adsorbate are shown in Tables 2 and 3.
|KOH:CWZ-22 = 4||25||6.98||0.80||0.999||2.97||0.62||0.999|
A good agreement can be noticed between the experimental data and the model fitting. This indicates that these models can accurately explain the correlation of the adsorption equilibrium with the adsorbent. According to the R2 parameter, all the fitted data for the four models are good. In all of the cases, the obtained regression coefficient R2 is higher than 0.999. Thus, in order to analyze the impact of various error functions on the predicted isotherms, four different error functions namely SSE, HYBRID, MPSD, and ARE were used by nonlinear regression method to recognize the error distribution between the experimental equilibrium data and the isotherms studied. Their values are presented in Table 4. The best fit isotherm was selected based on the error function. The lowest values of the error functions are bolded.
|KOH:CWZ-22 = 4||25||49.60||0.13||0.19||0.999||13.52||0.44||0.76||0.999|
|KOH:CWZ-22 = 4||25||0.0730||0.8836||8.3028||15.9486|
|KOH:CWZ-22 = 4||25||0.0340||0.7104||7.4561||18.0023|
|KOH:CWZ-22 = 4||25||0.0014||0.0360||1.5490||4.2991|
|KOH:CWZ-22 = 4||25||0.0046||0.1036||2.7307||6.7459|
The Sips model results in a much better fitting for CO2 adsorption than Langmuir, Freundlich, and Toth models under the entire experimental temperature range.
Table 3 lists the optimal isotherm parameters obtained by simultaneously fitting of all the adsorption equilibrium data at multiple temperatures for carbons CWZ-22 and KOH:CWZ-22 = 4. The excellent agreement between the model fittings and the experimental data demonstrates that these isotherm models can be confidently employed to accurately correlate the adsorption equilibria of the adsorbates. As it is observed, decreasing the temperature increases the maximum of the adsorbed amount (qm) in all isotherms. In other words, increasing temperature will decrease the amount adsorbed at a given pressure. The parameter bs is the adsorption affinity, whereas ns qualitatively characterizes the heterogeneity of the adsorbate adsorbent system. Also, it can be seen that affinity parameter (bs) increases with a decrease in temperature. In fact, at lower temperatures, the surface is covered more by the molecules which lead to the stronger affinity of adsorbate molecules towards the surface. The bs parameters calculated for unmodified and modified carbons are very similar so affinity of CO2 molecules towards the surface of this two ACs is similar. The surfaces of both ACs were homogenous because ns was higher than 0.5. The values of ns for KOH:CWZ-22 = 4 were higher than for the CWZ-22. The modification makes the surface more homogenous. The fitting parameters (qm, bs, and ns) of Sips isotherm equations are temperature dependent and defined as Eqs. (6)–(8) .
The parameters qm0, bs0, ns0 are the parameters qm, bs, ns, respectively, at the same reference temperature (T0) The χand α are constant parameters. In Langmuir equation, Q is equal to isosteric heat of adsorption, constant with surface coverage. The parameter Q in the Sips equation is only the measure of the adsorption heat. The parameter Q defined in the affinity constant bs is the isosteric heat at the fractional loading of 0.5.
|KOH:CWZ-22 = 4||1.10||6.66||−4.91|
The temperature-dependent equations of the exponent ns and saturation capacity qm are empirical, and such forms in Eqs. (6) and (8) were chosen arbitrary because of their simplicity. Because Eqs. (6) and (8) have not been derived from physical laws, then, as a result, the parameters χ and α have no physical meaning.
The plots of observed versus predicted values of parameters qms, bs, and ns showed good agreement between these values (Figure 9).
3.7. Isosteric heat of adsorption
The heat of adsorption is described as an indicator of the interaction strength between the adsorbate molecules and the adsorbent. Isosteric heat of adsorption as a function of surface coverage (defined as number of adsorbed molecules on a surface divided by the number of molecules in a filled monolayer on that surface) at different temperatures was calculated by applying the Clausius–Clapeyron equation. The Sips isotherm model was used here for an evaluation of isosteric heat of adsorption. The Qi of CWZ-22 and KOH:CWZ-22 = 4 was determined from the slope of the straight line after plotting ln(p) against 1/T for different surface coverage (θ), as shown in Figure 10.
The isosteric heat of adsorption for CWZ-22 and KOH:CWZ-22 = 4 is shown as a function of surface coverage in Figure 11. The isosteric heats of adsorption decreased with the surface coverage increase. For the both ACs at the “zero loading,” isosteric heats of adsorption tend to 30 kJ/mol. This value indicates the physisorption. The modification of CWZ-22 reduced the isosteric heat of adsorption. By extrapolation, the curves to surface coverage equal to 0.5 were found that the isosteric heat of adsorption for CWZ and KOH:CWZ-22 = 4 was equal to 10 and 6 kJ/mol, respectively. These values are similar to values obtained on the basis of Eq. (7) and presented in Table 5.
The impregnation of CWZ-22 by KOH highly effectively increased the porosity, specific surface area, and pore volume of ACs. The specific surface area of KOH:CWZ-22 = 4 significantly increased to 1299 m2/g as compared with pristine CWZ-22 (856 m2/g). Similarly, the total pore volume raised from 0.51 to 0.77 cm3/g.
The chemical modification of CWZ-22 increased the CO2 adsorption capacity of 38%. The CO2 adsorption capacity of 4.24 mmol/g obtained for the KOH:CWZ-22 = 4 was considerably higher than that for pristine CWZ-22 (3.08 mmol/g) at 25°C and 2 bar. The Sips model described very good the CO2 adsorption on KOH:CWZ-22 = 4 and in the pristine sample. The surface of the both ACs was homogenous. The values of isosteric heats of adsorption indicated on physisorption.
The research leading to these results has received funding from the Polish-Norwegian Research Programme operated by the National Centre for Research and Development under the Norwegian Financial Mechanism 2009–2014 in the frame of Project Contract No Pol-Nor/237761/98/2014.