Comprehending adsorption of methylethylketone and toluene and microwave regeneration effectiveness for beaded activated carbon derived from recycled waste bamboo tar

ABSTRACT

Beaded activated carbons (BACs) were derived from waste bamboo tar through carbonization (500°C for 2 hr) followed by physical activation using carbon dioxide (800–900°C for 2–4 hr). The adsorbent was examined for their physical and chemical properties, adsorption capacities toward methylethylketone (MEK) and toluene, and regenerabilities under microwave heating. It was found that the maximum total surface area reached for bamboo-tar-derived BAC after physical activation was 1364 m² g⁻¹, and more than 95% of the area was attributed to the microporous structures. Langmuir, Freundlich, and Dubinin-Radushkevich (D-R) isotherm models were applied to the adsorption isotherm fitting, and the minimum R² for each model was 0.986, 0.915, and 0.943, respectively. The isosteric heats of adsorption calculated based on D-R parameters for methylethylketone and toluene were 44.04 to 51.50 and 45.88 to 73.27 KJ mol⁻¹, respectively. They were slightly over the range of physisorption and increased with adsorbate loading, which might be related to the micropore filling mechanism. Microwave regeneration under 600 W of power output removed most of the adsorbate (>93.03%) within 8 min. The results of this study are intended to benefit future study on waste-derived adsorbent in environmental applications.

Implications

Recycling waste bamboo tar for the novel adsorbent preparation is shown feasible in this study. Beaded activated carbon (BAC) synthesized from this waste bamboo tar possessed a high specific surface area, which aided in the capturing of volatile organic compounds (VOCs). Three adsorption isotherms, Langmuir, Freundlich, Dubinin-Radushkevich (D-R) models can be applied in interpreting the experimental adsorption data, providing information on adsorption heat and possible adsorption mechanism. A potential microwave regeneration method for BAC is tested, showing high desorption efficiencies with minimum heel formation. These findings can provide a new pathway for waste bamboo tar management and VOC abatement using adsorbents.

Introduction

Air pollution kills millions of people yearly around the world (Lelieveld et al. 2015). Among the anthropogenic air pollutants, volatile organic compounds (VOCs) are emitted in tremendous amounts and found ubiquitously in modern society, causing adverse effects to both human beings and our environment (Zhang et al. 2017). For stationary VOCs emission sources, plenty of control strategies can be utilized, including absorption (Ozturk and Yilmaz 2006), oxidization, and catalytic conversion (Kamal, Razzak, and Hossain 2016), biotreatment (Mudliar et al. 2010), and adsorption using porous adsorbents such as activated carbon (AC) (Ghoshal and Manjare 2002; Zhang et al. 2017). Specifically, the adsorption removal of VOCs has the merit that economically recovering high-priced solvent can be fulfilled by regenerating the adsorbent, namely, reversing the adsorption process by heating the adsorbent followed by condensation (Johnsen et al. 2016; Wang et al. 2012).

The adsorption performance of AC has been studied for years. Factors influencing the adsorption capacity include physical properties of the adsorbent such as the pore structure, pore shape, and pore size distribution (Chiang, Chiang, and Huang 2001), and that of the adsorbate such as the molecular weight, concentration, dynamic diameter, and polarity (Li et al. 2012). The surface chemical properties of the adsorbent also play important roles in the adsorption processes, as the modification by oxidation or heat treatment of the surface of AC is typically made to increase the number of functional groups, enhancing the adsorption capacities toward certain compounds (Romero-Anaya, Lillo-Ródenas, and Linares-Solano 2015).

AC can be tailored into different morphologies (Luo et al. 2006), among which is the beaded activated carbon (BAC) that has several advantages including high fluidity, good packing ability, more controllable pore structures, high mechanical strength, and great adsorption performance (Qi et al. 2017; Romero-Anaya, Lillo-Ródenas, and Linares-Solano 2010). BAC can be derived from numerous polymeric precursors, including divinylbenzene-based polymer (Zhu et al. 2008), copolymer of vinylidene chloride and styrene (Qian et al. 2015), polystyrene-based resin (Wang et al. 2009), and phenol-formaldehyde (PF) resin (Qi et al. 2017; Zhang et al. 2018). However, the costs of these polymeric materials vary with fluctuating oil price, therefore, encouraging scientists to seek out possible replacements for BAC precursors.

A recent study has demonstrated that BAC could potentially be derived from waste bioresources such as waste bamboo tar (Huang, Hsi, and Liu 2013). It was found that the vast amount of waste bamboo tar generated during the bamboo charcoal production was rich in phenolic content, suitable to be recovered, applied in the synthesis of beaded phenolic resin (Lee et al. 2011), and subsequently activated to produce BAC (Singh and Lal 2008, 2010). The method was further improved in the study by Chen, Huang, and Hsi (2019), yet the qualities of BAC still demand improvements in total surface area, microporous structures, and most importantly, the adsorption performance of some widely used organic solvents, in order to be qualified as a potential replacement for the commercial BAC. As bamboo remains one of the predominant bioresources in several tropical and subtropical regions of Asia, Africa, and Latin America, the problem of a proper management for waste bamboo tar is still challenging. Moreover, the related research on developing valuable materials such as BAC, carbon fiber, and carbon quantum dots from waste bamboo tar is still in its infant stage, requiring more thorough investigations (Liang et al. 2017; Qiao et al. 2005).

The objective of the current study aims at valorizing waste bamboo tar and improving the qualities of BAC derived from it. For this purpose, suspension polymerization and physical activation using carbon dioxide (CO₂) were applied to derived BAC from waste bamboo tar. The properties of BAC were examined using an elemental analyzer (EA), scanning electron microscope (SEM), and nitrogen adsorption at 77 K. The BAC was then tested in a bench-scale adsorption system to evaluate its capability of recovering organic solvent, where methylethylketone (MEK) and toluene were selected as targeted compounds. To better understand the adsorption performance of BAC, the results of VOCs adsorption were fitted with the Langmuir, Freundlich, and Dubinin-Radushkevich (D-R) isotherm model and the isosteric heat of adsorption was obtained analytically. The last part of this study devotes to the regenerabilities of the derived adsorbent since this characteristic is equally important for an ideal adsorbent. Comparing to the conductive heating method, microwave heating has been extensively studied in recent years due to its rapid temperature response, homogeneous heating, high energy efficiency, and less direct contact between heating source and targeted materials (Foo and Hameed, 2012; Foo, Lee, and Hameed 2013). The technology has been adopted in the heating process of material preparation (Cheng et al. 2018, 2017; Mao et al. 2015b), and regeneration of dielectric materials such as AC (Hashisho, Rood, and Botich 2005; Mao et al. 2015a). Moreover, the inherent capability of BAC to be operated in a fluidized configuration favors the utilization of microwave heating (Cherbanski and Molga 2009; Price and Schmidt 1998) and lowers the heel buildup (irreversible adsorption) after cycles of regeneration (Kamravaei et al. 2017). The saturated BAC was therefore regenerated at the end of the experiment using microwave heating.

Material and methods

Sample preparation and activation

The rectification was carried out to separate phenolic compounds from other substances. The rectified bamboo tar contained 73.7 wt% phenolic compounds, suggested by the examination of a high-resolution gas chromatography/mass spectrometer (JEOL JMS-700 and SHIMADZU QP2010). This rectified bamboo tar served as the main component participating in the suspension polymerization.

The suspension polymerization was conducted in a 1-L four-neck round-bottom glass reactor equipped with a thermometer, a reflux condenser, and a Teflon stirrer. The rectified bamboo tar was blended with reagent-grade phenol (95% phenol) to adjust the formaldehyde to free phenol ratio (F/P). For a typical synthesis process, 1 wt% aqueous solution of polyvinyl alcohol (hydrolysis 98.5–99.2 mol%; degree of polymerization: 1700) as the stabilizer was added into the mixture of 1.6 F/P ratio. The pH value of the mixture was adjusted by adding trimethylamine (basic catalyst) to 7–8. After 90 min of homogenizing, hexamine was added to alter the F/P ratio to 2.9, and the crosslinking reaction occurred. The reaction ceased after around 4 hr and the mixture was cooled. The PF beads were collected on the filter and washed with deionized water and acetone. Cleaned PF beads were dried in an oven at 105°C and beads with the size of 20 × 30 mesh were sieved for the following experiment.

The BAC synthesis procedures involved carbonization under inert conditions and physical activation of PF beads using CO₂. All experiments were conducted in a cylindrical quartz tube with an inner diameter of 30 mm, heated by a conventional temperature-programmed furnace (DENGYNG, model D-35). The furnace utilized conductive heating method, and the heat was directed from outside to inside of the quartz tube. The PF beads were placed in the center of the quartz tube, carbonized under flowing of nitrogen (purity = 99.999%, flow rate = 200 mL min⁻¹) in a tubular furnace using a heating rate of 10°C min⁻¹ up to 500°C, dwell for 2 hr. The subsequent physical activation using high purity CO₂ (99.999%) as a mild oxidative agent removed the impurities blocking the pores generated after the carbonization step. The activation step was carried out under flowing of nitrogen (purity = 99.999%, flow rate = 200 mL min⁻¹) in the same tubular furnace using a heating rate of 10°C min⁻¹ up to 800°C, 850°C, and 900°C. Upon reaching the activation temperature, the CO₂ gas stream was directed into the furnace for 2, 3, and 4 hr of activation and then was switched back to nitrogen during the cooling step.

Physical and chemical characterization of materials

Morphology and pore structure

The morphologies and pore structures of the PF beads before and after the activation along with the commercial BAC were revealed using a scanning electron microscope (Hitachi, model TM3000) and nitrogen adsorption at 77 K (Micromeritics, model ASAP2020), respectively. Before the analysis, all samples were degassed at 10–20 Torr vacuum under 150°C for 24 hr. The total surface area (S_total) was calculated by the Brunauer-Emmett-Teller equation according to ASTM D4820-96a. The micropore surface area (S_micro) and micropore volume (V_micro) were obtained from a t-plot analysis using Jura-Harkins equation: t = [13.99/(0.0340 – log(P/P₀))]^0.5. The range of relative pressures used to determine S_micro and V_micro was based on thickness t values between 0.45 and 0.8 nm. Pore size distributions (PSDs) were determined using nonlocal density function theory (NLDFT) and Barrett-Joyner-Halenda (BJH) methods.

Elemental analysis

The chemical composition of the adsorbent was measured by an elemental analyzer (Elementar, model Vario EL Cube). The mass fractions of C, H, N, O, and S for the beads were obtained by complete oxidization of the samples followed by the detection of gaseous compounds with a thermal conductivity detector (TCD).

VOC adsorption test and isotherm construction

The adsorption testing system consisting of a VOCs vapor generating system, a temperature-controlled adsorption bed, a total hydrocarbon (THC) analyzer, and a data acquisition system (DAQ) is illustrated in Figure 1. A syringe pump (New Era Pump Systems Inc., model NE-1000) loaded with a gastight syringe (Hamilton Company) constantly injected MEK and toluene solvents into a small furnace. The temperature of the furnace was monitored by the thermocouple, K-type, and set well above the boiling point of both solvents (79.6°C and 110.6°C for MEK and toluene, respectively) where they immediately vaporized and mixed with the inert carrier gas (N₂, purity = 99.999%). The injection rates were adjusted according to the ideal gas law to generate a VOC gas stream with a desired concentration.

Figure 1. Schematic diagram of the adsorption testing system.

The on-line THC detecting system was a flame ionization detector (FID) (Ratfisch Analysensysteme GmbH, model RS53-T) continuously measuring the outlet concentrations of the adsorption bed. The concentrations of VOCs were recorded every second via a DAQ system consisted of a LabView software (National Instruments) and a data logger (National Instruments, model USB-6000).

The adsorption bed in this study was a tube made of Pyrex glass with an inner diameter of 1 cm. A K-type thermocouple (Omega) was inserted from one of the openings and extended into the center of the bed while the heating tape (Omega) wrapping around the tube. A proportional-integral-derivative (PID) controller connected to the thermocouple and the heating tape measured and controlled the temperature of the adsorption bed to the accuracy of ± 0.1°C. The experiments were conducted in a fixed-bed configuration. VOCs vapor came from the top to avoid expansions of the bed. Typically, BAC was filled in the tube with both sides covered by glass wool and fine glass meshes. The adsorption of VOCs onto the coverings was assumed negligible during the test.

The concentrations of VOCs ranged from 0.001 to 0.2 P/P₀ for MEK and from 0.004 to 0.28 P/P₀ for toluene, respectively. The operating temperature was set at 30°C, 40°C, or 50°C. The total flow rate was set at 1.6 SLPM. For each run, 1 g of BAC was loaded in the adsorption tube. The saturated adsorption capacity was calculated using weighted method and the following equation:

(1)

where W_f is the weight of BAC after adsorption and W_i is the weight of BAC before adsorption.

Microwave regeneration

The saturated BAC was regenerated in a microwave muffle furnace (Milestone, model Pyro 260). Prior to the microwave regeneration, 1 g of virgin BAC was saturated with MEK under the inlet concentration of 500 ppmv and operating temperature of 40°C. Once the BAC was saturated, it was transferred to an air-tight quartz reactor with a total volume of 1.5 L placed inside the microwave furnace. During the desorption step, the microwave furnace operated at 2.45 GHz with a constant power output of 600 W. A K-type thermocouple (Omega) measured the temperature change of the BAC, while 2.0 SLPM of inert purging gas (N₂, 99.999%) was directed into the reactor to purge the desorbed adsorbate from the reactor immediately. Desorption efficiencies were obtained by weighing the mass of BAC before and after the microwave regeneration.

Results and discussion

Adsorbent characterization

The surface morphologies and shapes of commercial BAC, termed “KBAC” (K stands for Kureha CORP) and self-prepared BAC with various synthesize conditions, termed “SBAC” were compared based on the SEM analysis. The beads which possess smooth surfaces and round shapes are considered suitable for mechanical interactions and fluidization. As can be seen from the SEM images (Figure 2), all the prepared adsorbents were in beaded shape. Macroscopically, the suspension polymerization reaction can successfully derive the beaded adsorbent so that both its packing in the fixed-bed and its fluidity in a fluidized system are superior to those in irregular shapes. However, except for SBAC-T800-t3 and KBAC, PF beads after the activation process appeared to have apparent fractures on the surface, microscopically. The extents of the fractures developed generally followed the order: SBAC-T900-t4 > SBAC-T900-t3 > SBAC-T900-t2 > SBAC-T850-t3 > SBAC-T800-t3 (T and t represent activation temperature in °C and time in hour, respectively.). This is in consistence with the results found in literatures, as the thermal degradation starts to enhance from 800°C. The vigorous degradation of the PF beads is attributed to the effective gasification agent, CO₂, and the suitable reaction temperature, especially above 800°C (Kim et al. 2004, 2002).

Figure 2. Scanning electron micrograph under 100× (upper) and 1500× (lower) magnification for (a) SBAC-T800-t3, (b) SBAC-T850-t3, (c) SBAC-T900-t2, (d) SBAC-T900-t3, (e) SBAC-T900-t4, and (f) KBAC.

The result of the elemental analysis indicates the degree of devolatilization and transformation from a raw material to a porous AC (Table 1). The carbonization process decomposed the raw material and increased its carbon content to 81.01 wt%. The subsequent activation further raised the carbon content up to 85.36– 89.47 wt%, consistent with the elemental composition of a typical AC.

Table 1. Physical and chemical properties of the adsorbents examined in this study.

	Component (wt% on dry basis)
Sample	C	H	O	N	S	Ash	Smicro (m^2 g^−1)	Stotal (m^2 g^−1)	Vmicro (cm^3 g^−1)	Vtotal (cm^3 g^−1)	davg (nm)
KBAC	88.60	1.58	4.23	0.46	0.03	5.10	956	973	0.494	0.526	2.162
SBAC-T800-t3	85.71	1.49	6.35	0.92	0.05	5.48	666	684	0.343	0.376	2.193
SBAC-T850-t3	85.36	1.49	7.47	0.87	0.05	4.76	903	927	0.465	0.512	2.195
SBAC-T900-t2	89.47	1.02	4.08	1.08	0.07	4.28	966	998	0.497	0.544	2.178
SBAC-T900-t3	88.90	1.05	4.51	0.99	0.07	4.48	1222	1266	0.631	0.692	2.186
SBAC-T900-t4	85.55	1.30	6.30	0.99	0.07	5.79	1306	1364	0.674	0.751	2.201

The pore structures of BAC samples were evaluated by using nitrogen adsorption at 77 K. The nitrogen adsorption and desorption curves for KBAC and SBACs are presented in Figure 3(a). All the adsorbents followed the type I(b) adsorption isotherm classified by the latest version of the IUPAC definition, indicating a microporous material having a broader pore size distribution including wider micropores (of width >0.7 nm) (Thommes et al. 2015). The results of pore structure analysis are also shown in Table 1. High-level trends were found that the surface area and the pore volume increased with activation temperature and duration, with SBAC-T900-t4 having the optimal S_total of 1364 m² g⁻¹ and V_total of 0.751 cm³ g⁻¹. The PSDs of BAC samples were also examined based on NLDFT and BJH methods (Figure 3(b–c)). It was found that the activation temperature and duration increased the fractions of pores with larger pore widths (revealed by two sharp peaks at 1.5–2 nm and 2.1–3.0 nm in Figure 3(b) and curves in Figure 3(c)). This might be due to the overgasification by CO₂ that causes the widening of micropores into mesopores.

Figure 3. (a) Nitrogen adsorption isotherm; (b) micropore; and (c) mesopore size distribution for the adsorbents examined in this study.

Adsorption and desorption performance

Isotherm construction

MEK and toluene were selected to compare the influence of activation condition on the adsorption performance of the SBACs. These two solvents have distinct physical and chemical properties. A preliminary evaluation of SBACs adsorption was conducted at 40°C and 500 ppmv of MEK and toluene. The results are shown in the Supplemental Data. The adsorption capacities of toluene were up to 85% greater than that of MEK under the same conditions, suggesting the stronger interaction between the toluene molecules and the adsorbent. This strong interaction can be attributed to two reasons. First, both toluene and the AC’s surface are nonpolar, which strongly enhances their mutual interaction (Li et al. 2012). Toluene’s aromatic structure also facilitates π-π interactions with the aromatic structure of the AC surface, increasing the amount adsorbed. It was also found that both MEK and toluene adsorption capacities increased linearly with the surface area and pore volume, indicating that these two are the dominating factors determining the overall adsorption performance (Ouzzine et al. 2019; Yu et al. 2018).

A great way to investigate and predict adsorption characteristics under various conditions is using the adsorption isotherm constructed based on the experimental results. The data of MEK and toluene adsorption by two bamboo-tar-derived BACs, SBAC-T900-t3, and SBAC-T900-t4, and the commercial BAC were fitted with Langmuir, Freundlich, and D-R isotherm models. The two SBAC samples were selected not only because of their high S_total but also of their abundant microporous structures, which have been demonstrated to facilitate the capture of gaseous molecules especially under a lower pressure range (Ghafari and Atkinson 2018; Lillo-Ródenas, Cazorla-Amoros, and Linares-Solano 2005; Romero-Anaya, Lillo-Ródenas, and Linares-Solano 2014, 2010).

The Langmuir model is one of the most widely applied theoretical isotherms. It assumes monolayer adsorption on the surface of the adsorbent and has the expression as follows:

(2)

where q (mg g⁻¹) is the adsorbed amount (or loading), P (Pa) the partial pressure of the adsorbate in the bulk phase, q_s (mg g⁻¹) the saturated adsorbed amount, and b (Pa⁻¹) the Langmuir constant representing the strength of adsorbate-adsorbent interactions (Do 1998).

The Freundlich model is an empirical one having the following form:

(3)

where K_F and n_F are empirical constants that are generally temperature dependent (Do 1998).

D-R isotherm has been shown to be able to reveal the extent of adsorbate-adsorbent interactions (Ghafari and Atkinson 2018; Mao et al. 2015b). It is derived based on Polanyi potential theory and expressed as:

(4)

and

(5)

where W is the cumulated volume of the adsorbed space, W₀ the limiting volume of the adsorbed space, E₀ the characteristic energy of adsorption for a reference vapor (usually benzene), β the similarity constant characterizing the polarizability of the adsorbate (β = 1 for the reference vapor), R the universal gas constant, T the temperature, and P₀ the saturated vapor pressure of the adsorbate. The fitting results are tabulated in Tables S1-S3. We found that the minimum R² for Langmuir, Freundlich, and D-R isotherm was 0.986, 0.915, and 0.943, respectively.

The plots of experimental data and results fitted by the Langmuir isotherm are presented in Figure 4. For the Langmuir isotherm model, q_s is greater in SBACs than in KBAC, both for MEK and toluene. This is mainly due to the drastic increase of adsorbed amounts for SBACs in higher inlet concentration ranges, which is attributed to the larger pore volumes of SBACs. Song et al. (2005) have pointed out that, in general, the larger pore volume leads to more adsorption of molecules due to the adsorbate condensation in pores.

Figure 4. Experimental and Langmuir-modeled isotherms; MEK adsorption on (a) SBAC-T900-t3, (b) SBAC-T900-t4, and (c) KBAC; toluene adsorption on (d) SBAC-T900-t3, (e) SBAC-T900-t4, and (f) KBAC.

Judging from the R² value, the Freundlich isotherm did not fit as good as the other two models. The heterogeneity of the carbon surface was revealed by n_F > 1 for all scenarios. The plots of the experimental and the Freundlich isotherm modeled data are presented in Figure 5. The Freundlich isotherm can hardly obtain a good fit for such a broad concentration range, as it does not conform to Henty’s law at lower concentration and fails to account for the more complicated adsorption mechanisms involved at higher concentration of adsorbate (Pei and Zhang 2012).

Figure 5. Experimental and Freundlich-modeled isotherms; MEK adsorption on (a) SBAC-T900-t3, (b) SBAC-T900-t4, and (c) KBAC; toluene adsorption on (d) SBAC-T900-t3, (e) SBAC-T900-t4, and (f) KBAC.

For both MEK and toluene adsorption, the micropore volumes (W₀) obtained in the D-R isotherm model are in the order of SBAC-T900-t4 > SBAC-T900-t3 > KBAC, which is identical to that obtained from nitrogen adsorption test. A slight difference between the values obtained might be due to different kinetic diameters among the probing molecules (N₂: 4.02 Å, MEK: 5.25 Å, and toluene: 5.85 Å) (Lashaki et al. 2012).

Generally, the adsorption capacity under lower vapor pressure (<0.1 P/P₀) is attributed to micropore filling mechanism (Lillo-Ródenas, Cazorla-Amoros, and Linares-Solano 2005), whereas the amount adsorbed under higher vapor pressure can be accounted for by surface layering or capillary condensation mechanism due to the presence of meso- and macropores (Mosher et al. 2013). The increase of adsorption capacities above P/P₀ = 0.05 should be considered as adsorption in meso- and macropores region, which was found to be much higher in SBACs than in KBAC (Table 1). This is in consistence with PSD analysis shown previously, where SBACs developed higher proportions of larger pores compared with KBAC. The characteristic energy, E (or E₀), is a parameter indicating the solid properties (Nguyen and Do 2001) and affinity between adsorbent and adsorbate (Ghafari and Atkinson 2018). The resulting E was consistently found decreasing with increasing adsorbent’s d_avg, which is the average pore size measured by the nitrogen adsorption. This implies that a narrower average pore size results in a higher adsorption potential and stronger tendencies for the molecules to adsorb. It can be utilized to estimate the slit-pore half-width, x₀, by adopting an empirical equation proposed by Dubinin (1989):

(2)

As shown in Table S3, SBACs had a larger x₀ than KBAC. Results for SBACs showed good agreement using MEK or toluene as a probing molecule, while those for KBAC were deviated. This could be due to the fact that smaller and narrower pore sizes in KBAC might hinder diffusion of larger molecules, while wider PSDs found in SBACs compensate these effects (Hsi et al. 2011). It was also noticed that W₀ calculated for KBAC was less dependent on temperature than that calculated for SBACs. The “activated entry effect” defined as adsorption phenomena that the diffusion of molecules through narrow pores can be enhanced by increasing the surrounding temperature, which could explain this observation (Chiang, Chiang, and Huang 2001). The plots of experimental and D-R isotherm-modeled data are presented in Figure 6. A good agreement was found for the D-R modeling, and the results can be applied in predicting adsorption capacities under various conditions (Cal, Rood, and Larson 1997).

Figure 6. Experimental and Dubinin-Radushkevich-modeled isotherms; MEK adsorption on (a) SBAC-T900-t3, (b) SBAC-T900-t4, and (c) KBAC; toluene adsorption on (d) SBAC-T900-t3, (e) SBAC-T900-t4, and (f) KBAC.

Isosteric heat of adsorption

The isosteric heat of adsorption, ΔH_s, can be calculated analytically using the D-R parameters and Clausius-Clapeyron (C-C) equation (Ramirez, Qi, and Rood 2005). The derivation is shown in the Supplemental Data. ΔH_s for all the adsorbate-adsorbent systems is presented in Figure 7. It was found that ΔH_s for MEK adsorption onto all BACs was within 44.04 to 51.50 KJ mol⁻¹, which was similar to the results found in Long et al. (2012). On the other hand, ΔH_s for toluene adsorption ranged from 45.88 to 73.27 KJ mol⁻¹, which increased drastically at a higher loading. Such phenomena are not rare in other studies regarding the adsorption of organic compound in porous materials. Similar trends can be found in Long et al., where ΔH_s first decreased and then increased for both benzene and MEK adsorption onto the activated carbon. Ramirez, Qi, and Rood (2005) demonstrated that, as the loading increased, increasing lateral interactions due to the molecule aggregation within the micropores could possibly lead to a higher ΔH_s. Zheng et al. (2018) showed that due to the preferential adsorption of molecules with the adsorption sites, ΔH_s tended to decrease initially, and increase as the effect of intermolecular interaction became significant. This intermolecular interaction was negligible under lower partial pressure, but contributed greatly to the adsorption heat, as molecules needed to overcome the interactive forces with other molecules to reach certain adsorption sites under high pressure. Moreover, ΔH_s obtained in this study was all higher than typical bonding energy for physisorption (i.e., approximately 40 KJ mol⁻¹), suggesting that adsorption in the porous medium could involve complex interactions between molecules and higher energy released during the process.

Figure 7. Variation of isosteric heat of adsorption with surface loading; MEK adsorption on (a) SBAC-T900-t3, (b) SBAC-T900-t4, and (c) KBAC; toluene adsorption on (d) SBAC-T900-t3, (e) SBAC-T900-t4, and (f) KBAC.

Microwave regeneration

Desorption efficiencies were evaluated for different durations (2, 4, or 8 min) of microwave application under a constant flow rate of purging nitrogen gas (2.0 SLPM) and power output (600 W). The experimental results are shown in Figure 8. It was found that the desorption efficiencies were greater in SBACs than those in KBAC, mainly due to the higher temperatures achieved under microwave irradiation (the measured highest temperatures were 272°C, 300°C, and 308°C for KBAC, SBAC-T900-t3, and SBAC-T900-t4, respectively). This result indicates that AC with the same precursor could generate similar temperature raising pattern, probably resulted from the similar chemical composition and dielectric characteristics (Pan et al. 2016). Within 8 min of microwave irradiation, the desorption efficiencies reached 96.73 ± 2.32, 100 ± 0.00, and 93.03 ± 0.35% for SBAC-T900-t3, SBAC-T900-t4, and KBAC, respectively. Kim and Ahn (2010) regenerated AC loaded with MEK using conventional heating at 300°C with a desorption time of 1 hr, and the desorption efficiency was 99.1%. In the results of Shah, Pre, and Alappat (2014), 2 hr of conductive heating at 160°C was required for regenerating MEK-saturated AC to obtain a desorption efficiency of 98% while successive adsorption-desorption cycle resulted in the continuous degradation of MEK adsorption capacity. While cyclic regeneration is not shown in the current study, it is clear that comparing to the conventional thermal regeneration, the microwave regeneration is more time-saving and can easily achieve high desorption efficiencies. When using microwave heating, the electromagnetic energy is converted to thermal energy inside the adsorbent, thus creating a flow of heat from the hotter area (the core of the solid) to the cooler area (the surface of the solid). This phenomenon added a beneficial effect to further drive the desorbed adsorbate out of the adsorbent. In contrast, when regenerated in a traditional furnace, the temperature gradient around an adsorbent is opposite to that of the microwave-heated adsorbent, creating a barrier for the desorbed adsorbate to diffuse out, hampering the desorption process (Cherbanski et al. 2011).

Figure 8. Temperature variations and desorption efficiencies for microwave regeneration of BACs saturated with MEK (n = 3).

Conclusion

Bamboo-tar-derived BACs were successfully prepared through suspension polymerization and physical (CO₂) activation. Comparing to the commercial BAC, the bamboo-tar-derived BAC possessed higher total surface area when activated at 900°C for >3 hr. The specific surface area was up to 1364 m² g⁻¹, and are mostly contributed by micropores. Adsorption tests and isotherm fitting showed that VOCs adsorption on the derived adsorbent is mainly physisorption while increasing VOCs loading increases the amount of the heat released due to the micropore filling mechanism. Microwave heating technique was demonstrated to be effective in regenerating the spent BAC in this study. However, two limitations of this new adsorbent were found, including insufficient mechanical strength and lower adsorption capacity under low concentration ranges. Future studies to address these disadvantages will focus on optimizing the carbonization temperatures and durations to improve the mechanical strength, and developing pore structures with suitable size distributions by different activation methods.

Supplemental Material

Supplemental data for this article can be accessed on the publisher’s website.

Disclosure Statement

No potential conflict of interest was reported by the authors.

Additional information

Funding

This work was supported by the Ministry of Science and Technology, Taiwan [MOST 108-2218-E-002 −068 -MY3].

Notes on contributors

Yu-Ting Chen

Yu-Ting Chen is a graduated master degree student from Graduate Institute of Environmental Engineering, National Taiwan University, in Taipei, Taiwan.

Ying-Pin Huang

Ying-Pin Huang is a researcher at Industrial Technology Research Institute (ITRI) southern region campus, in Tainan, Taiwan.

Can Wang

Can Wang is a professor at School of Environmental Science and Engineering, Tianjin University, and Tianjin Key Lab of Indoor Air Environmental Quality Control, in Tianjin, China

Ji-Guang Deng

Ji-Guang Deng is a researcher at College of Environmental and Energy Engineering, Beijing University of Technology, in Beijing, China.

Hsing-Cheng Hsi

Hsing-Cheng Hsi is a professor at Graduate Institute of Environmental Engineering, National Taiwan University, in Taipei, Taiwan.