Materials

HNTs were obtained from Applied Minerals Inc. (New York, USA). Sodium hexametaphosphate (Junsei Co., Tokyo, Japan, 97% purity) and sulfuric acid (Daejung Co., Seoul, Korea, 98% purity) were used without further purification.

Modification of Natural HNTs

A 1.5 g portion of HNT was mixed with 50 mL of 1 M, 3 M, or 5 M sulfuric acid solution in a flask. The HNT was well dispersed in the acid solution by bath-type sonication for 20 min. An aspirator was used to remove atmospheric gases from the lumen of the tubes. As a result, the sulfuric acid solution from the flask filled the lumen of the halloysite nanotubes, and chemical reactions occurred to remove the alumina composition of the HNT. The flask was heated at 90°C for 24 h to react the sulfuric acid with the aluminum of the HNT (Abdullayev et al. Reference Abdullayev, Joshi, Wei, Zhao and Lvov2012). After reaction, the HNT modified by the 1 M, 3 M, and 5 M sulfuric acid were labeled “yHNT (x M)”, where y is the modification and x is 1, 3, or 5, depending on which sulfuric acid concentration was used. The sample was then washed with DI water 5 times, placed in a freezer for 6 h, and freeze dried for 3 days.

Purification and Separation of Natural HNTs

Purification of the unmodified HNTs followed the method introduced by Luo et al. (Reference Luo, Song, Feng, Run, Cui, Wu, Gao and Wang2013). 25 g of HNT was suspended in 100 mL of DI water by stirring for 2 h; then 2 g of sodium hexametaphosphate (SHMP) was added gradually to enhance dispersion. The system was then stirred for an additional 24 h and left to stand at room temperature for 6 h to allow solids to settle. The supernatant was decanted and discarded to remove readily soluble impurities. The remaining solid was then separated into various size fractions by centrifugation as follows. The sediment was resuspended in water then centrifuged at 2000 rpm (431×g) for 5 min and the supernatant collected and labeled HNT-1. The sediment was again resuspended in water and centrifuged at 4000 rpm (1722×g) for 10 min and the supernatant decanted and labeled HNT-2. These steps were repeated at 6000 rpm (3875×g), 8000 rpm (6888×g), and 10,000 rpm (10,763×g) to obtain samples labeled HNT-3, HNT-4, and HNT-5, respectively. This was repeated for each HNT specimen. Each such particle-size fractionation yielded tubes with different average lengths and outer and inner diameters (Table 4).

Acid Modification of HNT

A 0.5 g portion of each mHNT-x (m = modification; x = 1–5) sample was mixed with 16.7 mL of 5 M sulfuric acid in a flask and further dispersed by vortex mixing and sonication. The suspension was then aspirated to remove dissolved gases in the HNT suspension in order to facilitate the entrance of sulfuric acid solution into the lumen, thereby enabling reaction between the sulfuric acid and the Al in the tetrahedral sheet of the HNT inner wall. The reaction was further promoted by heating the suspensions at 90°C for 24 h. They were then cooled, washed five times with DI water, and freeze-dried. The process was repeated for all HNTs.

Modification of Natural HNTs

The porosities and surface areas of the HNTs were measured using N₂ adsorption. The volume of adsorbed gas at standard state (STP: T = 273.15 K, 101.3 kPa) per g of sample, represented by V _a, was measured by the Brunauer-Emmett-Teller (BET) method using a Belsorp-Mini II device (MicrotracBEL Corp., Osaka, Japan). Adsorption of volumetric N₂ proceeded on unmodified and modified HNTs. The N₂ gas-adsorption capacity of the unmodified HNT was 163.74 cm³ (STP)/g. The isotherm type was IUPAC type IV, revealing pore sizes in the mesoporous and macroporous ranges, but not in the microporous range. In the case of mHNTs (1 M), the value of V _a was 471.14 cm³ (STP)/g; for mHNTs (3 M), 746.75 cm³ (STP)/g. The value of V _a for mHNTs (5 M) was 837.31 cm³ (STP)/g, which is 5.11 times greater than for the natural HNTs (Table 1). All modified HNTs exhibited morphologies and plot shapes that follow the IUPAC type IV isotherms, similar to the natural HNTs.

Table 1 Adsorbed gas volume (V _a), specific surface area (S _BET), and total pore volume (V _total) of natural HNT, mHNT (1 M), mHNT (3 M), and mHNT (5 M).

The thickness of the adsorption layer is defined as t = 0.354×V _a/V _m [nm]_. The value of 0.354 comes from the assumption that the thickness of a single layer of N₂ molecules is 0.354 nm (Lippens and De Boer Reference Lippens and De Boer1965; Kaneko Reference Kaneko1994). According to one hypothesis, adsorbed N₂ molecules form an hexagonally close-packed layer on the solid surface (Schull Reference Schull1948). The linear-like graph shapes of the samples in Fig. 1b, which increase at the same rate as standard isotherms, means that they have macroporous characteristics. The graphs of the modified HNTs have a steeper slope than those for unmodified HNTs. This is because the V _m values, which indicate the monolayer volume of the samples, are larger in the modified HNTs. The V _a value of mHNTs (5 M) at 1.2 nm of pore thickness was 270.17 cm³ (STP)/g, which is 8.5 times larger than the V _a value of the natural HNTs of 31.703 cm³ (STP)/g at 1.2 nm. The modified HNTs had a larger capacity for gas-adsorption than the natural HNTs.

Fig. 1 a N₂ adsorption isotherms and b thickness plots of natural HNT, mHNT (1 M), mHNT (3 M), and mHNT (5 M).

The internal structures of the natural HNTs and the mHNTs were revealed by transmission electron microscopy (TEM, JEOL JEM-2100F, Tokyo, Japan) (Fig. 2). A clear crystal morphology with a relatively small lumen diameter in the natural HNTs was verified. In contrast, the mHNTs lost their crystallinity after dealumination, and their lumen diameter was expanded.

Fig. 2 TEM images of: a natural HNT, b mHNT (1 M), c mHNT (3 M), d mHNT (5 M).

The crystallinities of both natural HNTs and modified HNTs were analyzed using X-ray diffraction (XRD). The diffraction patterns of natural HNTs have a monoclinic-domatic crystallographic character. The characteristic sharp peak at 12°2θ (Fig. 3) corresponding to the (001) reflection occurred due to the HNT multiwall reflection (Costanzo and Glese Reference Costanzo and Giese1985; Guimaraes et al. Reference Guimaraes, Enyashin, Seifert and Duarte2010). Modified HNTs lost the characteristic peaks in the corresponding 2θ region, however. Instead of the sharp peaks indicating alumina crystals, modified HNTs showed a broad band observed at ~20–25°2θ. This means that the aluminosilicate composition of the natural HNTs was changed to amorphous silica by acid modification.

Fig. 3 XRD traces of natural HNT and mHNT (5 M).

Energy-dispersive X-ray spectroscopy (EDS) mapping was used to track changes in elemental composition and to verify structural changes during acid modification. Comparison of unmodified and modified HNTs (Fig. 4) revealed that the unmodified HNT was composed of evenly combined and distributed Al, Si, and O, whereas EDS mapping of the mHNTs (5 M) sample revealed that most of the Al was lost but the morphology of the silica structure remained.

Fig. 4 EDS mapping images of a natural HNT and b mHNT (5 M).

Modification of Separated HNTs

The five particle-size fractions of the unmodified HNT (using the BET method) yielded similar shapes in their gas-adsorption isotherms (type IV), as would be expected because of their similar atomic compositions and morphologies. A notable difference, however, was in the value of V _a, which is a reflection of characteristics such as size, mass, and porosity. The samples showed a consistent tendency of increasing V _a, except for HNT-5, which was assumed to differ because of an imperfect structure in the tubes that hindered gas adsorption.

The largest value for V _a was 226.43 cm³ (STP)/g, for HNT-4. This was 1.74 times larger than the smallest V _a value, which was exhibited by HNT-1 (130.37 cm³ (STP)/g. Another trend was that the adsorption capacity increased consistently in the lower p/p ₀ range, from HNT-1 to HNT-5 (Table 2).

Table 2 Adsorbed-gas volume (V _a), specific surface area (S _BET), and total pore volume (V _total) of 5 groups of natural HNTs.

Compared to those of unmodified HNTs (Fig. 5a), the adsorption isotherms of the five particle-size fractions of the acid-modified HNTs revealed a consistent tendency for the V _a value to decrease as the particle size became smaller (Fig. 5b). The largest V _a was 1223.1 cm³ (STP)/g, for mHNT-5, which was 5.6 times larger than for HNT-5 before acid treatment (Tables 2 and 3). The results showed that the modification of halloysite nanotubes increased the gas-adsorption efficiency of the unmodified halloysite by a factor of 7.47 (1223.1/163.74) (Tables 1 and 3). Overall, acid modification of the HNT increased its porosity, gas-adsorption capacity, and macroporosity. The smallest value for V _a among the mHNT samples was 638.84 cm³ (STP)/g for mHNT-1, which is 4.9 times larger than for HNT-1. Both acid activation and smaller particle size improved the gas-adsorption capacity of HNT tubular materials, therefore.

Fig. 5 Adsorption isotherms of five groups of: a natural HNT, b mHNT (5 M).

Table 3 Adsorbed gas volume (V _a), specific surface area (S _BET), and total pore volume (V _total) of five groups of mHNTs (5 M).

The porosity distributions of the HNTs were plotted using the Barrett-Joyner-Halenda (BJH) method. Adsorbents that have type IV adsorption isotherms can be characterized by BJH adsorption-desorption plots (Luan et al. Reference Luan, He, Zhou, Cheng and Klinowski1995; Baiju et al. Reference Baiju, Shukla, Biju, Reddy and Warrier2009). Hysteresis, the difference between adsorption and desorption curves, occurs due to capillary condensation of N₂ gas in the mesopores of the materials (Barrett et al. Reference Barrett, Joyner and Halenda1951). The desorption isotherm of the material and the Kelvin equation (Eq. 1) demonstrate the relationships between critical condensation pressure and pore size and were used to calculate the pore-size distribution in the BJH plots (Schull Reference Schull1948).

(1) $ln\frac{p}{p_0}=-\frac{2\gamma V_L}{RT{r}_m}$

where p is equilibrium pressure in pores with a radius r _m; p ₀, pressure on the planar surface; γ, the surface tension; V _L, the molar volume; R, the gas constant; and T, the absolute temperature. The cumulative pore-volume graph was obtained for various pore volumes. The observed tendency (Fig. 6a) was that the pore size decreased from HNT-1 to HNT-5. Samples HNT-1 and HNT-2 were macroporous, whereas HNT-3, HNT-4, and HNT-5 were mesoporous. HNT-1 exhibited the tallest peak at d _p = 70 nm, revealing obvious microporous behavior. On the other hand, sample HNT-5 was mesoporous because the porosity distribution was clearly <50 nm. Because the volume distributions of HNT-4 and HNT-5 were close to 0 with a porosity distribution >50 nm, these samples were non-macroporous tubes.

Fig. 6 BJH plots of 5 groups of: a natural HNT, b mHNT (5 M).

Plots of the BJH curves for the five groups of modified HNTs (Fig. 6b) revealed that the porosities of all natural HNTs were expanded by sulfuric acid modification. The pore sizes of samples mHNT-3, mHNT-4, and mHNT-5 (on the right side of Fig. 6b) compared to those of the natural HNTs groups were enlarged from mesopores to macropores.

Regarding pore-thickness curves of the separated HNTs7, the slopes of the curves increased with decreasing particle size (Fig. 7a). This indicates that samples separated later had larger V _m values and gas-adsorption efficiencies. The gas adsorption of mHNT-5 at 1.2 nm pore thickness was 83.816 cm³ (STP)/g, which is 2.98 times greater than the corresponding value for the mHNT-1, 28.313 cm³ (STP)/g.

Fig. 7 Thickness plots of five groups of: a natural HNT, b mHNT (5 M).

The thickness plot patterns of the modified HNTs were similar to those of the unmodified HNTs (Fig. 7b). The inclination increased according to the order of separation, because the samples separated later had larger V _m values and gas-adsorption efficiencies. The gas-adsorption of mHNT-5 at 1.2 nm pore thickness was 191.81 cm³ (STP)/g, which is 1.52 times greater than that of mHNT-1, which was 125.87 cm³ (STP)/g.

Transmission electron microscopy images of natural and modified HNTs (Fig. 8, Table 4) were used to determine the average lengths and outer and inner diameters of the HNTs. The TEM images were also utilized to compare the relative differences between the five difference sizes of natural HNTs. The size of the nanotubes tended to decrease with increasing centrifugal force. The nanotube groups were not clearly divided in terms of gas-adsorption capacity because the separation of HNTs was affected by various factors such as mass, hydrophilicity, and the size of each particle.

Fig. 8 TEM images of natural HNT groups: a HNT-1, b HNT-2, c HNT-3, d HNT-4, e HNT-5. TEM images of modified HNT groups: f mHNT-1, g mHNT-2, h mHNT-3, i mHNT-4, j mHNT-5.

Table 4. Average lengths and outer (O.D.) and inner (I.D.) diameters of the HNTs.

The largest contribution to separation was assumed to be the mass of the particles. Because the atomic composition of each halloysite is almost the same, as the HNT particle size decreased, the mass of the particle would be decreased proportionally. Therefore, in principle, the gas adsorption capacity of the tubes is inversely proportional to the size of the particles.

HNT-5 had nano-size debris (Fig. 8e). The hindrance of gas-adsorption for this sample was because of these light nanotube fragments that were mixed with the intact tubes.

Separated samples of modified HNTs are shown in Fig. 8f–j. They exhibited an increasing gas-adsorption tendency with separation process; the size of the nanotubes decreased according to the order of separation. Therefore, the assumption that the gas-adsorption capacity is inversely proportional to the mass or size of the particles is applied from the same perspective as with separated natural HNTs. The fragments of nanotubes are detectable in the mHNT-5 sample also (Fig. 8j). These experiments demonstrated that the porosity of halloysite nanotubes could be increased by selection of a small size and acid treatment.