Preparation of a Highly Porous Clay-Based Absorbent for Hazard Spillage Mitigation via Two-Step Expansion of Vermiculite

: Expanded vermiculite (eVMT) has been studied as a risk-free, general-purpose absorbent for liquid hazards due to its excellent thermal and chemical stability. Here, vermiculite was expanded by two steps: exfoliation by 30 wt% H 2 O 2 treatment at 60 ◦ C and subsequent expansion by microwave heating. This two-step expansion produced more homogenously separated concertina-like eVMTs with a higher total pore volume of 7.75 cm 3 g − 1 than the conventional thermal method. The two-step eVMT was found to be greatly superior to the thermal and commercial silver counterparts in hazardous liquid-uptake performance. The uptake was simply interpreted as a physical inﬁlling process of a liquid into the eVMT pores, and the spontaneous hazard removal with a great capacity was discussed with the large pore volume of two-step eVMT and its suitable pore dimensions for capillary action. As a practical device, a prototype absorbent assembly made of these eVMTs demonstrated the successful mitigation of liquid hazards on an impermeable surface.


Introduction
Large-scale chemical spill accidents frequently occur worldwide on land or in oceans during the production, storage, transportation and usage of hazardous chemicals [1]. Sudden hazard spillage not only poses a fatal threat to human life, but also causes serious environmental contamination. Therefore, precautions should be taken in advance for mitigating the risks of spill accidents. Absorbent-based mitigation can efficiently clean up liquid hazards without secondary chemical pollutants that are usually produced in different mitigation methods, such as dispersion, in-situ burning and solidification [2][3][4]. A great number of absorbents have been developed for the removal of harmful chemical spillages using natural minerals, natural organics, synthetic polymers and carbon nanomaterials [5][6][7][8][9][10][11][12][13][14][15]. Natural organics and synthetic polymers can be degraded by toxic compounds and easily burned by fire, often accompanied by chemical accidents [16]. On the other hand, inflammability and high production cost of carbon nanomaterials hamper their broad utilization as absorbents under harsh conditions [9,10]. Therefore, inorganic minerals, such as zeolites, clays and silicon oxides have been considered as risk-free and general-purpose absorbents for the remediation of liquid hazard spillage due to their ideal properties, including non-flammability, chemical inertness, low environmental impact and a natural abundance. However, they have a crucial drawback. Their absorption capacity is inferior to advanced carbon or polymer absorbents. For realizing a high-performance absorbent, the absorption capacity of the mineral absorbents must be improved.
Vermiculite is a type of natural phyllosilicate clay with a 2:1 layer structure, possessing an octahedral alumina sheet sandwiched by tetrahedral silicate sheets. The trilayer units Figure 1a shows a schematic diagram for the preparation of eVMT, based on a twostep process consisting of wet H 2 O 2 treatment and microwave drying. Very thin or tiny vermiculite particles were discarded by means of a weight-based separation for better homogeneity in the sample dimension. The refined non-expanded vermiculite was first treated with an aqueous solution of 30 wt% H 2 O 2 for its chemical exfoliation. Typically, Minerals 2021, 11, 1371 3 of 14 10 g of the sample was spread in a glass Petri dish (diameter = 150 mm) containing a 60 mL H 2 O 2 solution, and then placed in a convection oven at 60 • C for 1 h covered by a lid. Next, the dish was transferred to a 1 kW microwave oven (KR-S341T, Daewoo, South Korea) in order to quickly evaporate the liquid solution. The wet-treated sample was heated by microwave radiation for 60 s. Afterward, the heated sample (~100 • C) was taken out in ambient atmosphere for cooling and further evaporation. This heating-cooling cycle was sequentially performed around five times until liquid was almost completely evaporated. The H 2 O 2 -expanded sample was finally dried at 150 • C in a vacuum oven for 1 h, and then stored in a desiccator before usage. For a comparison study, thermally expanded vermiculite at 1000 • C and a commercial silver sample were also utilized. The detailed thermal expansion conditions were previously reported for the same crude vermiculite [7]. vermiculite particles were discarded by means of a weight-based separation for better mogeneity in the sample dimension. The refined non-expanded vermiculite was f treated with an aqueous solution of 30 wt% H2O2 for its chemical exfoliation. Typically g of the sample was spread in a glass Petri dish (diameter = 150 mm) containing a 60 H2O2 solution, and then placed in a convection oven at 60 °C for 1 h covered by a lid. N the dish was transferred to a 1 kW microwave oven (KR-S341T, Daewoo, South Korea order to quickly evaporate the liquid solution. The wet-treated sample was heated by crowave radiation for 60 s. Afterward, the heated sample (~100 °C) was taken out in a bient atmosphere for cooling and further evaporation. This heating-cooling cycle was quentially performed around five times until liquid was almost completely evaporat The H2O2-expanded sample was finally dried at 150 °C in a vacuum oven for 1 h, and th stored in a desiccator before usage. For a comparison study, thermally expanded verm ulite at 1000 °C and a commercial silver sample were also utilized. The detailed therm expansion conditions were previously reported for the same crude vermiculite [7].

Material Characterizations
The extent of the expansion of vermiculite was evaluated by measuring the change volume [33], as shown in Figure 1b. To obtain an initial sample volume (V0), 1 g of the cru vermiculite was put into a graduated cylinder. The expanded volume after the H2O2 and crowave treatments (V) was also quantified by a graduated cylinder. The volumes were me ured after tipping samples gently for packing, and their average values were obtained fr five measurements. The expansion ratio (k) was calculated by the relationship of k = V/V0.
The morphology of eVMT was observed by scanning electron microscopy (SEM; S-48 Hitachi, Tokyo, Japan) under an electron beam with an accelerating voltage of 15 kV a deposition of a platinum nano-film. The chemical compositions of vermiculite were analy by X-ray fluorescence (XRF; ZSX Primus II, Rigaku, Tokyo, Japan). Crystal structures of samples were probed by powder X-ray diffractometer (PXRD; Uniflex 600, Rigaku) with C Kα1 (λ = 1.54056 Å) radiation. Nitrogen sorption (Belsorp-mini II, MicrotracBEL, Tokyo, Jap and mercury porosimetry (AutoPore IV 9500, Micromeritics, Norcross, GA, USA) were lized to characterize pore size distribution and a specific surface area.

Material Characterizations
The extent of the expansion of vermiculite was evaluated by measuring the change in volume [33], as shown in Figure 1b. To obtain an initial sample volume (V 0 ), 1 g of the crude vermiculite was put into a graduated cylinder. The expanded volume after the H 2 O 2 and microwave treatments (V) was also quantified by a graduated cylinder. The volumes were measured after tipping samples gently for packing, and their average values were obtained from five measurements. The expansion ratio (k) was calculated by the relationship of k = V/V 0 .
The morphology of eVMT was observed by scanning electron microscopy (SEM; S-4800, Hitachi, Tokyo, Japan) under an electron beam with an accelerating voltage of 15 kV after deposition of a platinum nano-film. The chemical compositions of vermiculite were analyzed by X-ray fluorescence (XRF; ZSX Primus II, Rigaku, Tokyo, Japan). Crystal structures of the samples were probed by powder X-ray diffractometer (PXRD; Uniflex 600, Rigaku) with Cu-Kα 1 (λ = 1.54056 Å) radiation. Nitrogen sorption (Belsorp-mini II, Mi-crotracBEL, Tokyo, Japan) and mercury porosimetry (AutoPore IV 9500, Micromeritics, Norcross, GA, USA) were utilized to characterize pore size distribution and a specific surface area.

Absorptive Removal Performance against Liquid Hazards
The absorption capacities and removal efficiencies of the three eVMTs were evaluated for 20 types of harmful liquid chemicals. The detailed experimental procedures were reported in [7]. Briefly, a stainless-steel mesh cell was charged by a 1.0 g eVMT sample, then fully immersed into a testing liquid for 30 min. It was lifted up to allow the excess liquid drained off. Shortly afterward, the weight of absorbed samples (M abs ) was measured by a balance. The dry mass (M dry ) was determined after drying in a vacuum oven at 150 • C for 1 h. The absorption capacity, defined by the relative weight of the absorbed liquid to absorbent, was determined by the following relationship.
These weights were determined by averaging values acquired by three measurements. For determining the removal efficiency, approximately 20 g of hazardous liquid was prepared in a reservoir. Its initial weight (M i ) was measured by a balance. A meshed container was filled with 10 g eVMT and then immersed into the hazard-containing vessel for 10 min. The eVMT container was lifted up and waits for dripping the excess liquid. The weight of the remaining solvent (M r ) was measured. The removal efficiency was defined by the percentage ratio of the extracted liquid mass with respect to its initial quantity, which was calculated by the following equation.

Fabrication and Demonstration of a Prototype Absorbent Assembly
A prototype absorbent assembly was fabricated to reveal the real application of eVMT as an absorbent for liquid hazard mitigation. Figure 2a shows the unlocked assembly consisting of an absorbent paper pouch and a stainless-steel mesh container. The absorbent pouch with a dimension of width × depth × height = 50 × 35 × 65 mm 3 was filled up by 7 g eVMT and was then sealed using a hand-press sealing machine. The pouch was placed inside of a stainless-steel mesh container (φ = 65 mm and h = 65 mm) and confined by a locked structure. The strong, heavy stainless-steel holder provides advantages to increase the specific gravity of the assembly for quick submersion and to handle it reliably. A hazard mitigation test of the prototype device was performed under the following protocol: (1) immersing the assembly into a beaker containing 300 mL of a liquid organic chemical for 1 min, (2) withdrawing it outside the beaker and measuring a weight of remaining liquid and (3) repeating the mitigation cycle (steps 1 and 2) using a new assembly until the beaker became nearly empty. The mitigation procedure can see in Video S1 in the Supplementary Materials. Figure 2b shows the experimental setup used for the mitigation test after the first mitigation of a blue-dyed ethanol solution. The absorption capacity of each mitigation trial and the cumulative removal efficiency were calculated from the measured weights.

Expansion Behavior of Palabora Vermiculite
The expansion ratio was measured at different temperatures and times of H2O2 exfoliation under a subsequent thermal drying. Figure 3 shows variations of the expansion ratio as a function of the H2O2 reaction time obtained at three different temperatures of 25 °C, 60 °C and 80 °C. At the lowest temperature of 25 °C, the k magnitude steadily increased and reached 8.0 at a reaction time of 3 h. As the temperature increased, the expansion happened more quickly and was then saturated to a maximum value: k = 16.2 at a reaction time of 1 h and k = 15.4 at 0.5 h for expansion at 60 °C and 80 °C, respectively. A more rapid expansion at a higher temperature was previously reported and interpreted as the efficient evolution of atomic oxygen via decomposition of hydrogen peroxide penetrated into the interlayer spaces of vermiculite [29]. The atomic oxygen could result in the separation of the silicate layers from each other. On the other hand, the maximum k magnitude at 80 °C was slightly lower than that at 60 °C. This behavior seems to be attributed to a decrease in H2O2 concentration with the elapse of reaction time due to the considerable H2O2 decomposition outside of the interlayer. The expansion was known to greatly diminish below the concentration of 30 wt% H2O2 [28]. In addition, extremely exfoliated vermiculites were observed to break down to smaller fragments at 80 °C, which resulted in a decrease in k magnitude. As a result, the 60 °C H2O2 treatment for 1 h was chosen as an optimized exfoliation condition, and utilized to prepare highly expanded, mechanically durable vermiculite.
A volume shrinkage of the H2O2 treated sample was observed during the subsequent thermal drying, induced by the evaporation of water molecules with a high surface tension. For preventing the shrinkage and reducing the dry process time, H2O2-treated particles were dried by the heating-cooling cycle of the microwave irradiation for 1 min and standing in ambient air for 1 min. Complete drying was possible within approximately 10 min. Interestingly, the sample was found to be further expanded during the microwave drying process: the volume was increased by approximately 23% when compared with that after the H2O2 treatment (see Figure 1b). The k values of three different eVMTs used in this work are summarized in Table 1, together with their physical dimensions and apparent density. The two-step method results in a highly expanded sample up to k = 20 while preserving its porous structure. Compared with the k = 6.6 obtained from thermally expanded samples at 1000 °C, the two-step expansion provides a higher k magnitude of more than three times for the same Palabora vermiculite. The microwave expansion has been previously reported in several different types of vermiculites [24][25][26]. This process can be understood as a result of the build-up pressure formed by the prompt evaporation of crystal water remaining in the vermiculite interlayers after H2O2 treatment. Therefore, H2O2 treatment coupled with microwave drying is an efficient method to prepare a highly porous, robust absorbent for the mitigation of liquid hazards.

Expansion Behavior of Palabora Vermiculite
The expansion ratio was measured at different temperatures and times of H 2 O 2 exfoliation under a subsequent thermal drying. Figure 3 shows variations of the expansion ratio as a function of the H 2 O 2 reaction time obtained at three different temperatures of 25 • C, 60 • C and 80 • C. At the lowest temperature of 25 • C, the k magnitude steadily increased and reached 8.0 at a reaction time of 3 h. As the temperature increased, the expansion happened more quickly and was then saturated to a maximum value: k = 16.2 at a reaction time of 1 h and k = 15.4 at 0.5 h for expansion at 60 • C and 80 • C, respectively. A more rapid expansion at a higher temperature was previously reported and interpreted as the efficient evolution of atomic oxygen via decomposition of hydrogen peroxide penetrated into the interlayer spaces of vermiculite [29]. The atomic oxygen could result in the separation of the silicate layers from each other. On the other hand, the maximum k magnitude at 80 • C was slightly lower than that at 60 • C. This behavior seems to be attributed to a decrease in H 2 O 2 concentration with the elapse of reaction time due to the considerable H 2 O 2 decomposition outside of the interlayer. The expansion was known to greatly diminish below the concentration of 30 wt% H 2 O 2 [28]. In addition, extremely exfoliated vermiculites were observed to break down to smaller fragments at 80 • C, which resulted in a decrease in k magnitude. As a result, the 60 • C H 2 O 2 treatment for 1 h was chosen as an optimized exfoliation condition, and utilized to prepare highly expanded, mechanically durable vermiculite.
A volume shrinkage of the H 2 O 2 treated sample was observed during the subsequent thermal drying, induced by the evaporation of water molecules with a high surface tension. For preventing the shrinkage and reducing the dry process time, H 2 O 2 -treated particles were dried by the heating-cooling cycle of the microwave irradiation for 1 min and standing in ambient air for 1 min. Complete drying was possible within approximately 10 min. Interestingly, the sample was found to be further expanded during the microwave drying process: the volume was increased by approximately 23% when compared with that after the H 2 O 2 treatment (see Figure 1b). The k values of three different eVMTs used in this work are summarized in Table 1, together with their physical dimensions and apparent density. The two-step method results in a highly expanded sample up to k = 20 while preserving its porous structure. Compared with the k = 6.6 obtained from thermally expanded samples at 1000 • C, the two-step expansion provides a higher k magnitude of more than three times for the same Palabora vermiculite. The microwave expansion has been previously reported in several different types of vermiculites [24][25][26]. This process can be understood as a result of the build-up pressure formed by the prompt evaporation of crystal water remaining in the vermiculite interlayers after H 2 O 2 treatment. Therefore, H 2 O 2 treatment coupled with microwave drying is an efficient method to prepare a highly porous, robust absorbent for the mitigation of liquid hazards.

Characterization of Material Properties
SEM images of the eVMTs prepared by the two-step process and the rapid 1000 °C heating are shown in Figure 4a,b, respectively. The image obtained from the two-step expansion reveals a morphology with a more homogeneously separated concertina-like shape than that of the thermal expansion. The conical slit gaps have an opening size of less than 250 µm and the thickness of the unexfoliated flakes is in the range of 0.1-0.4 µm. The plate thickness is thinner than that of the thermally expanded sample, indicating more favorable interlayer delamination of the crude vermiculite. This observation is in good agreement with the greater expansion ratio of the two-step sample. These results indicate that H2O2 treatment coupled with microwave drying can produce more uniform and extensive exfoliation of vermiculite interlayers than simple thermal expansion. In addition, the chemical compositions analyzed by XRF are given in Table 2. The obtained compositions are in good agreement with previously reported results for the same vermiculite [22,28,34].
Porous characteristics of eVMTs were quantitatively evaluated by both Hg porosimetry and N2 sorption. Figure 5a shows pore-size distributions of the two-step and the thermally expanded samples quantified by Hg porosimetry. They display bimodal distributions: two peaks at 7.0 µm and 260 µm for the two-step sample, and two maxima at 1.9 µm and 284 µm for the thermal sample. These distributions exhibit a macropore component in the region larger than approximately 100 µm together with the small pores. The macroporosity might be considered as interparticle voids formed by packing eVMT granules. More importantly, the two-step sample presents higher intensity in intrusion volume, suggesting the existence of a larger pore volume within eVMT, than that of the thermal one. A total pore volume of 7.75 cm 3 g −1 was obtained from the two-step sample, more than two times greater when compared to the thermal sample. These total pore volumes

Characterization of Material Properties
SEM images of the eVMTs prepared by the two-step process and the rapid 1000 • C heating are shown in Figure 4a,b, respectively. The image obtained from the two-step expansion reveals a morphology with a more homogeneously separated concertina-like shape than that of the thermal expansion. The conical slit gaps have an opening size of less than 250 µm and the thickness of the unexfoliated flakes is in the range of 0.1-0.4 µm. The plate thickness is thinner than that of the thermally expanded sample, indicating more favorable interlayer delamination of the crude vermiculite. This observation is in good agreement with the greater expansion ratio of the two-step sample. These results indicate that H 2 O 2 treatment coupled with microwave drying can produce more uniform and extensive exfoliation of vermiculite interlayers than simple thermal expansion. In addition, the chemical compositions analyzed by XRF are given in Table 2. The obtained compositions are in good agreement with previously reported results for the same vermiculite [22,28,34].
Porous characteristics of eVMTs were quantitatively evaluated by both Hg porosimetry and N 2 sorption. Figure 5a shows pore-size distributions of the two-step and the thermally expanded samples quantified by Hg porosimetry. They display bimodal distributions: two peaks at 7.0 µm and 260 µm for the two-step sample, and two maxima at 1.9 µm and 284 µm for the thermal sample. These distributions exhibit a macropore component in the region larger than approximately 100 µm together with the small pores. The macroporosity might be considered as interparticle voids formed by packing eVMT granules. More importantly, the two-step sample presents higher intensity in intrusion volume, suggesting the existence of a larger pore volume within eVMT, than that of the thermal one. A total pore volume of 7.75 cm 3 g −1 was obtained from the two-step sample, more than two times greater when compared to the thermal sample. These total pore volumes and areas are  Table 3, together with complementary values analyzed by N 2 sorption. These distributions clearly indicate that the two-step eVMT has a highly porous structure with a pore dimension of 0.1-500 µm, as observed by SEM. N2 sorption. These distributions clearly indicate that the two-step eVMT has a highly porous structure with a pore dimension of 0.1-500 µm, as observed by SEM.
(a) (b)  3.57 11.8 Furthermore, for investigating the distribution of pores smaller than sub-micrometer, N2 sorption isotherms of the two samples were measured as shown in Figure 5b. In contrast to Hg porosimetry, the isotherm curve of the thermal sample exhibits higher intensity in adsorption volume and a more obvious hysteresis loop in the region of p/p0 = 0.4-0.95 than those of the two-step sample. The higher adsorption indicates that the smaller pore  N2 sorption. These distributions clearly indicate that the two-step eVMT has a highly porous structure with a pore dimension of 0.1-500 µm, as observed by SEM.
(a) (b)  3.57 11.8 Furthermore, for investigating the distribution of pores smaller than sub-micrometer, N2 sorption isotherms of the two samples were measured as shown in Figure 5b. In contrast to Hg porosimetry, the isotherm curve of the thermal sample exhibits higher intensity  Furthermore, for investigating the distribution of pores smaller than sub-micrometer, N 2 sorption isotherms of the two samples were measured as shown in Figure 5b. In contrast to Hg porosimetry, the isotherm curve of the thermal sample exhibits higher intensity in adsorption volume and a more obvious hysteresis loop in the region of p/p 0 = 0.4-0.95 than those of the two-step sample. The higher adsorption indicates that the smaller pore structure is more extensively developed during the thermal expansion than the counterpart. This interpretation is well supported by the values of the surface area and total pore volume, obtained from the BET adsorption model (see Table 3). The thermal sample has a BET surface area of 11.8 m 2 g −1 , which is greater than the 3.57 m 2 g −1 of the two-step. On the other hand, the hysteresis loop suggests the existence of a mesoporous structure, and the BJH analysis exhibited a distinct peak at around 30 nm in pore-size distribution (see Figure S1). These results indicate that the porosity characteristics are strongly dependent on the expansion method used: the H 2 O 2 -treatment coupled with the microwave drying method produces a highly expanded form with various conical slit gaps in micrometer dimensions, whereas the thermal expansion is favorable to provide a micro/mesoporous structure with a large surface area. The two-step eVMT with a large total pore volume seems to be a promising absorbent for liquid hazard removal. In addition, the thermally expanded sample may have many advantages in surface-involved adsorption applications. Figure 6 shows the PXRD spectra of the crude and two eVMTs in a low 2θ angle region for observing a change in the first-order basal spacing (d 002 ). The crude sample exhibits three intense reflections with d-spacings of 1.43 nm, 1.25 nm and 1.19 nm and one weak peak of d 002 = 1.01 nm. These different d-spacings have been interpreted with different water-layer hydration-states (WLHS) and interstratification: 2-WLHS for 1.43 nm, interstratified phase for 1.25 nm, 1-WLHS for 1.19 nm and 1/0-WLHS for 1.01 nm [22,25,35]. In the case of the two-step eVMT, the four peaks still survive with a slight broadening, indicating the occurrence of significant exfoliation of the basal layers without a serious change in crystalline structure. In addition, the decrease in the 2-WLHS peak intensity implies that the H 2 O 2 and subsequent microwave treatments induce a slight diminishment in the hydrated states of the interplanar cations. Similar intensity reduction was previously reported more clearly when a Mg-vermiculite sample was treated by 80 • C H 2 O 2 (aq) and microwave radiation for 40 min [31]. The high H 2 O 2 temperature and the long microwave time may be attributed to induce great exfoliation and extensive water evaporation from the fully hydrated 2-WLHS, respectively. In contrast, the PXRD pattern of the thermal sample displays only the dehydrated 1/0-WLHS peak with the smallest d-spacing. The hydrated and interstratified peaks completely disappeared. These observations are in good agreement with the previous works [7,22]. The prompt heating seems to cause almost complete dehydration and break-down of the interstratified structure. Consequently, the resultant crystal structures of the eVMTs could be related with the above-mentioned different characteristics in porosity according to the expansion methods. structure is more extensively developed during the thermal expansion than the counterpart. This interpretation is well supported by the values of the surface area and total pore volume, obtained from the BET adsorption model (see Table 3). The thermal sample has a BET surface area of 11.8 m 2 g −1 , which is greater than the 3.57 m 2 g −1 of the two-step. On the other hand, the hysteresis loop suggests the existence of a mesoporous structure, and the BJH analysis exhibited a distinct peak at around 30 nm in pore-size distribution (see Figure S1). These results indicate that the porosity characteristics are strongly dependent on the expansion method used: the H2O2-treatment coupled with the microwave drying method produces a highly expanded form with various conical slit gaps in micrometer dimensions, whereas the thermal expansion is favorable to provide a micro/mesoporous structure with a large surface area. The two-step eVMT with a large total pore volume seems to be a promising absorbent for liquid hazard removal. In addition, the thermally expanded sample may have many advantages in surface-involved adsorption applications. Figure 6 shows the PXRD spectra of the crude and two eVMTs in a low 2θ angle region for observing a change in the first-order basal spacing (d002). The crude sample exhibits three intense reflections with d-spacings of 1.43 nm, 1.25 nm and 1.19 nm and one weak peak of d002 = 1.01 nm. These different d-spacings have been interpreted with different water-layer hydration-states (WLHS) and interstratification: 2-WLHS for 1.43 nm, interstratified phase for 1.25 nm, 1-WLHS for 1.19 nm and 1/0-WLHS for 1.01 nm [22,25,35]. In the case of the two-step eVMT, the four peaks still survive with a slight broadening, indicating the occurrence of significant exfoliation of the basal layers without a serious change in crystalline structure. In addition, the decrease in the 2-WLHS peak intensity implies that the H2O2 and subsequent microwave treatments induce a slight diminishment in the hydrated states of the interplanar cations. Similar intensity reduction was previously reported more clearly when a Mg-vermiculite sample was treated by 80 °C H2O2(aq) and microwave radiation for 40 min [31]. The high H2O2 temperature and the long microwave time may be attributed to induce great exfoliation and extensive water evaporation from the fully hydrated 2-WLHS, respectively. In contrast, the PXRD pattern of the thermal sample displays only the dehydrated 1/0-WLHS peak with the smallest d-spacing. The hydrated and interstratified peaks completely disappeared. These observations are in good agreement with the previous works [7,22]. The prompt heating seems to cause almost complete dehydration and break-down of the interstratified structure. Consequently, the resultant crystal structures of the eVMTs could be related with the abovementioned different characteristics in porosity according to the expansion methods.

Absorption Capacity
The absorption capacity of the two-step eVMT was quantified for twenty harmful organic compounds with different chemical properties, including ethanol, methanol, methyl ethyl ketone, ethyl acetate, furfuryl alcohol, acrylic acid, methyl acrylate, phenol, allyl alcohol, methyl vinyl ketone, acrylonitrile, benzene, toluene, m-cresol, allyl chloride, 2-nitrotoluene, isophorone diisocyanate, benzyl chloride, toluene-2,4-diisocyanate and vinyl acetate. The same measurements were also performed using thermally expanded and commercial silver eVMTs for the purpose of comparison. Figure 7 shows horizontal bar graphs for the liquid hazards, displaying variations in the absorption capacity for the three eVMTs. The hazardous compounds are arbitrarily grouped together according to the hydrophilicity. Hydrophilic and hydrophobic chemicals are positioned in the upper and bottom regions of the graph, respectively. The two-step sample distinctly exhibits great absorption capacities, ranging from 2.4 g g −1 to 7.9 g g −1 . These capacities are approximately 2-3 times higher than those of the thermal and silver samples for the same chemical, except for the three hydrophobic chemicals of isophorone diisocyanate, benzyl chloride and toluene-2,4-diisocyanate with a lower capacity of less than 4.0 g g −1 . There is no clear correlation with the absorption capacity and chemical properties of an absorbate, such as hydrophilicity/hydrophobicity and the existence of a specific functional group, as in our previous results obtained from the thermally expanded sample [7]. Although it is not a major factor, the inferior absorption capability to the specific hydrophobic compounds suggests that the absorption can be affected by chemical interactions between the liquid hazard and eVMT. In addition, the obtained absorption capacities for various hazards are greater than 0.53 g g −1 of hydrophobic dolomite [12] for oil and 2.5 g g −1 of modified VMT [20] for water, and comparable with 3.2-7.5 g g −1 of expanded perlite [6] for oil-water mixture.

Absorption Capacity
The absorption capacity of the two-step eVMT was quantified for twenty harmful organic compounds with different chemical properties, including ethanol, methanol, methyl ethyl ketone, ethyl acetate, furfuryl alcohol, acrylic acid, methyl acrylate, phenol, allyl alcohol, methyl vinyl ketone, acrylonitrile, benzene, toluene, m-cresol, allyl chloride, 2nitrotoluene, isophorone diisocyanate, benzyl chloride, toluene-2,4-diisocyanate and vinyl acetate. The same measurements were also performed using thermally expanded and commercial silver eVMTs for the purpose of comparison. Figure 7 shows horizontal bar graphs for the liquid hazards, displaying variations in the absorption capacity for the three eVMTs. The hazardous compounds are arbitrarily grouped together according to the hydrophilicity. Hydrophilic and hydrophobic chemicals are positioned in the upper and bottom regions of the graph, respectively. The two-step sample distinctly exhibits great absorption capacities, ranging from 2.4 g g −1 to 7.9 g g −1 . These capacities are approximately 2-3 times higher than those of the thermal and silver samples for the same chemical, except for the three hydrophobic chemicals of isophorone diisocyanate, benzyl chloride and toluene-2,4-diisocyanate with a lower capacity of less than 4.0 g g −1 . There is no clear correlation with the absorption capacity and chemical properties of an absorbate, such as hydrophilicity/hydrophobicity and the existence of a specific functional group, as in our previous results obtained from the thermally expanded sample [7]. Although it is not a major factor, the inferior absorption capability to the specific hydrophobic compounds suggests that the absorption can be affected by chemical interactions between the liquid hazard and eVMT. In addition, the obtained absorption capacities for various hazards are greater than 0.53 g g −1 of hydrophobic dolomite [12] for oil and 2.5 g g −1 of modified VMT [20] for water, and comparable with 3.2-7.5 g g −1 of expanded perlite [6] for oilwater mixture.  Assuming that the absorption happens by infilling the macro-porous total pore volume of eVMT, the absorption capacity defined by the relative weights strongly depends on the density of a liquid hazard. The total pore volume (V t ) of 7.75 cm 3 g −1 determined by Hg porosimetry corresponds to the maximum amount in absorption volume capacity, i.e., absorbed volume per unit of absorbent weight. The filling percentage of the pore volume can be expressed by the following equation: where V is the absorbed liquid volume per 1 g eVMT. This value can be calculated by the division of the absorption capacity (g g −1 ) by the liquid density (g cm −3 ). Figure 8 displays the variations of the filling percentages for the two-step and the thermally expanded samples. Overall, the two-step eVMT exhibits higher filling percentages than those of the thermal one, presenting better efficiency in absorptive liquid uptake. However, the filling percentage exhibits an opposite tendency in the case of furfuryl alcohol and the three hydrophobic compounds with a low absorption capacity, which is further evidence for the involvement of the chemical characteristics of a hazard in an absorption process. The two-step eVMT demonstrates a high filling percentage of 63-96%, except for the four exceptions. Assuming that the absorption happens by infilling the macro-porous total pore volume of eVMT, the absorption capacity defined by the relative weights strongly depends on the density of a liquid hazard. The total pore volume (Vt) of 7.75 cm 3 g −1 determined by Hg porosimetry corresponds to the maximum amount in absorption volume capacity, i.e., absorbed volume per unit of absorbent weight. The filling percentage of the pore volume can be expressed by the following equation: where V is the absorbed liquid volume per 1 g eVMT. This value can be calculated by the division of the absorption capacity (g g −1 ) by the liquid density (g cm −3 ). Figure 8 displays the variations of the filling percentages for the two-step and the thermally expanded samples. Overall, the two-step eVMT exhibits higher filling percentages than those of the thermal one, presenting better efficiency in absorptive liquid uptake. However, the filling percentage exhibits an opposite tendency in the case of furfuryl alcohol and the three hydrophobic compounds with a low absorption capacity, which is further evidence for the involvement of the chemical characteristics of a hazard in an absorption process. The twostep eVMT demonstrates a high filling percentage of 63-96%, except for the four exceptions.

Removal Efficiency
The removal efficiencies of the three eVMTs were also measured for the same harmful chemicals in order to evaluate the effectiveness in hazard removal via spontaneous wicking. The obtained quantities are presented in Figure 9. The removal efficiency of the two-step sample is greater than 95% for all the tested chemicals, which discloses a superb capability in hazard removal. In addition, the removal efficiency reveals a little variation in magnitude, irrespective of the tested hazard chemicals. This outcome outperforms 90% of the thermal eVMT and 73-95% of silver.

Removal Efficiency
The removal efficiencies of the three eVMTs were also measured for the same harmful chemicals in order to evaluate the effectiveness in hazard removal via spontaneous wicking. The obtained quantities are presented in Figure 9. The removal efficiency of the two-step sample is greater than 95% for all the tested chemicals, which discloses a superb capability in hazard removal. In addition, the removal efficiency reveals a little variation in magnitude, irrespective of the tested hazard chemicals. This outcome outperforms 90% of the thermal eVMT and 73-95% of silver.

Analysis of the Absorption Performance
The absorption capacity has a strong correlation with the total pore volume of eVMT and the density of hazardous liquids rather than any other physicochemical properties. Therefore, the adsorption can be regarded as the physical process of infilling a liquid hazard into the pores of eVMT, even though it is slightly affected by chemical interactions between the absorbent and absorbate. For preparing eVMT with a great absorption capacity, one of the most important requirements is a high expansion ratio, i.e., the great pore volume in eVMT. Another key factor is the size distribution of the pores. Since liquid wicking behavior strongly depends on the dimensions of a flow channel, the size of the empty gaps in eVMT can influence the absorption characteristics, such as the filling percentage and the removal rate. In fact, the spontaneous imbibition degrades along a channel that is too narrow or wide due to the considerable flow resistance or negligible capillarity, respectively. An ideal eVMT pore size for absorptive hazard removal seems to be in the micrometer range of roughly 1-100 µm.
In studying the hazard uptake performance, the two-step sample demonstrates the best performance among the three eVMTs in terms of the absorption capacity, the filling percentage and the removal efficiency. The excellent performances can be interpreted as an outcome of producing a highly and uniformly expanded structure with a large pore volume of 7.75 cm 3 g −1 with suitable conical slit gaps of 0.1-500 µm for capillary action. It is worth remembering that thermal eVMT with a larger surface area does not result in higher absorption capacity. The weak dependence of uptake behaviors on hydrophilicity or hydrophobicity can be explained with the unique wetting property of eVMT having both superhydrophilicity and oleophilicity, as observed in our previous work [32]. Hydrophobic organic molecules have been observed to spontaneously penetrate into a physically packed eVMT column, although the flow rate is slower than that of aqueous solution. Consequently, these uptake tests obviously indicate that the two-step eVMT can be utilized as an effective absorbent for the remediation of various types of liquid hazards.

Analysis of the Absorption Performance
The absorption capacity has a strong correlation with the total pore volume of eVMT and the density of hazardous liquids rather than any other physicochemical properties. Therefore, the adsorption can be regarded as the physical process of infilling a liquid hazard into the pores of eVMT, even though it is slightly affected by chemical interactions between the absorbent and absorbate. For preparing eVMT with a great absorption capacity, one of the most important requirements is a high expansion ratio, i.e., the great pore volume in eVMT. Another key factor is the size distribution of the pores. Since liquid wicking behavior strongly depends on the dimensions of a flow channel, the size of the empty gaps in eVMT can influence the absorption characteristics, such as the filling percentage and the removal rate. In fact, the spontaneous imbibition degrades along a channel that is too narrow or wide due to the considerable flow resistance or negligible capillarity, respectively. An ideal eVMT pore size for absorptive hazard removal seems to be in the micrometer range of roughly 1-100 µm.
In studying the hazard uptake performance, the two-step sample demonstrates the best performance among the three eVMTs in terms of the absorption capacity, the filling percentage and the removal efficiency. The excellent performances can be interpreted as an outcome of producing a highly and uniformly expanded structure with a large pore volume of 7.75 cm 3 g −1 with suitable conical slit gaps of 0.1-500 µm for capillary action. It is worth remembering that thermal eVMT with a larger surface area does not result in higher absorption capacity. The weak dependence of uptake behaviors on hydrophilicity or hydrophobicity can be explained with the unique wetting property of eVMT having both superhydrophilicity and oleophilicity, as observed in our previous work [32]. Hydrophobic organic molecules have been observed to spontaneously penetrate into a physically packed eVMT column, although the flow rate is slower than that of aqueous solution. Consequently, these uptake tests obviously indicate that the two-step eVMT can be utilized as an effective absorbent for the remediation of various types of liquid hazards.

Hazard Mitigation Using the Prototype Absorbent
The prototype absorbent assembly was fabricated to determine its usefulness as a remedy for a liquid hazard existing on an impermeable surface. Video S1 demonstrates the almost complete removal of a 300 mL ethanol solution in a beaker via eight consecutive mitigations using a new eVMT assembly each time. Because the submersion and take-out times were controlled to be 1 min, the total process time was approximately 16 min. A metal chain linked to the mesh container was utilized to effectively move the assembly. Using an absorbent weight of about 7 g and the measured weight of the remaining liquid, the absorption capacity and the cumulative removal efficiency of each mitigation trial were calculated and are given in Table 4. This mitigation test was also performed for different chemicals of methanol, acrylonitrile, toluene and vinyl acetate. In spite of the short absorption time of 1 min, absorption capacities of the first mitigation were in the range of 5.9-7.2 g g −1 , slightly higher than the corresponding values for the same chemicals (see Figure 7). This range can be ascribed to the use of different amounts of absorbent and absorbate, and/or a dissimilar type in the absorbent container. The sorption capacity obviously decreased and the cumulative removal efficiency increased as the mitigation was repeated several times. The decrease in absorption capacity was due to the reduction in the liquid volume. It is worth mentioning that the assembly can absorb a considerable amount of chemicals for 1 min via spontaneous wicking action even when it is partially in contact with a small amount of liquid. All the tested chemicals exhibit cumulative removal efficiencies greater than 97% after six or seven consecutive mitigations. These tests successfully reveal that the eVMT-based prototype device can be utilized to remedy the spillage of various harmful organic compounds on an impermeable floor or soil. Because the absorbent assembly made of granular eVMT particles is reusable after thermal activation and can be flexibly fabricated with diverse sizes and shapes, our prototype device has the potential to become a cost-effective solution to safely remove various liquid hazards. a Calculated with an initial volume of 300 mL and corresponding to the cumulative removal efficiency.

Conclusions
The eVMT absorbent was prepared by the two-step method of H 2 O 2 treatment coupled with microwave drying as a remedy for hazard spillage. When compared with the conventional thermal expansion, the novel two-step process produced highly and uniformly exfoliated eVMTs with a greater pore volume and more suitable micro-sized pores for capillarity-driven imbibition. These porosity characteristics lead to excellent uptake performance (the absorption capacity = 4.0-7.9 g g −1 , the filling percentage = 63-96% and the removal efficiency >95%) for almost all types of tested harmful liquids. The prototype eVMTs assembly has presented fast, spontaneous removal of liquid organic compounds in a beaker. These results indicate that the eVMTs prepared by the two-step method is a promising general-purpose, risk-free absorbent for a large-scale remediation of hazard spillage due to its low cost, thermal stability, chemical inertness and mechanical durability. In addition, this two-step process could easily scale up using a larger reactor and a bigger microwave processor.
Supplementary Materials: The following are available online at https://www.mdpi.com/article/ 10.3390/min11121371/s1, Figure S1: pore size distributions of the two-step and thermal eVMTs obtained by BJH model, Video S1: the consecutive uptake mitigation of an ethanol solution in a beaker using 8 prototype absorbent assemblies.