Adsorption of berberine hydrochloride onto mesoporous carbons with tunable pore size

Abstract

Eleven mesoporous carbon adsorbents were synthesized and characterized, and the adsorption properties of berberine hydrochloride on these adsorbents were investigated. The mesoporous carbon adsorbents have high BET specific surface areas (1051.2 to 1554.9 m² g⁻¹), large pore volumes (1.42 to 2.31 m³ g⁻¹), and their average pore radii range from 1.53 nm to 3.28 nm. The adsorption capacities of berberine hydrochloride on these adsorbents strongly depend on the BET specific surface areas and pore volumes of the mesoporous carbons. The mesoporous carbon sample with the highest BET specific surface area was selected as the most promising adsorbent for adsorption of berberine hydrochloride because of its high adsorption capacities (415 mg g⁻¹ at 298 K, 0.10 mg mL⁻¹). The adsorption thermodynamic parameters of berberine hydrochloride on the selected adsorbent were calculated. The adsorption equilibrium of berberine hydrochloride on the selected adsorbent could be established within 240 minutes at 298 K. The dynamic adsorption capacity of berberine hydrochloride on the selected adsorbent is close to the theoretical maximum adsorption capacity; and 72.4% of the adsorbed berberine hydrochloride could be desorbed by a 70% alcohol aqueous solution. The excellent adsorption/desorption properties make the mesoporous carbons promising adsorbents for adsorption of berberine hydrochloride from aqueous solutions.

---

1. Introduction

Berberine hydrochloride is a quaternary ammonium chloride widely existing in herbal plants. Berberine hydrochloride has broad-spectrum antibiotic activity, and other pharmacological effects such as anti-tumor, anti-oxidant, anti-inflammatory properties; it is extensively used in clinic for treatment of gastroenteric discomfort and adjuvant therapy of type 2 diabetes mellitus, hyperlipidemia, and hypertension.^1–5 Berberine hydrochloride is generally obtained from extracts of herbal plants, and a series of separation steps such as dissolving, filtration, re-crystallization and column separation are needed for further purification. The conventional approach needs a plenty of heavy work and is difficult to achieve a high purity of berberine hydrochloride due to the large amount of impurities coexisted in herbal plant extracts.⁶ On the other hand, berberine hydrochloride is poisonous for microorganisms, difficult to be degraded,⁷ and could become one of the serious pollutants in water. Therefore, effective purification and separation of berberine hydrochloride is essential for efficient use of herbal extracts and treatment of berberine hydrochloride rich wastewater.

Adsorption is believed to be one of the most efficient separation processes and has been successfully applied for separation and purification of natural products like berberine hydrochloride from herbal material extracts. Silica gel (SiO₂) and alumina (Al₂O₃) are the most widely used adsorbents for natural product purification, and using other commercial adsorbents such as activated carbon, polymers and minerals^6,8–12 for purifying berberine hydrochloride has been studied, however, high adsorption capacity and selectivity are difficult to achieve on these commonly used commercial adsorbents. Mesoporous (2–50 nm) carbons have the properties of tunable pore structures, high specific surface areas, large pore volumes, high thermal stability and good chemical stability,^13,14 which could favor the adsorption of natural products like berberine hydrochloride, and these materials have been applied as adsorbents for removal of pollutants such as aromatic compounds,^14–16 organic dyes,^17,18 heavy metal ions^19–21 and radioactive substances²² from wastewater and purification of biomolecules from aqueous solutions.^23,24 Organic–organic self-assembly route has been widely applied to synthesize organic nanostructures and carbon materials including mesoporous carbons.^25–28 In the organic–organic self-assembly method for mesoporous carbon synthesis, a thin film containing amphiphilic block copolymers as soft template and suitable carbon precursors is synthesized from the solvent evaporation induced self-assembly of organic molecules, mesoporous carbon is produced after polymerization and carbonization. Pluronic F127 and P123 are two triblock copolymers commonly used as soft templates.^25,29–31 From our previous study, mesoporous carbons have been proven to be effective adsorbents for adsorption of berberine hydrochloride from aqueous solutions, and their adsorption properties for berberine hydrochloride heavily depend on their pore textural properties.^32,33 However, further experimental data about the relationship between synthesis conditions of mesoporous carbons and their adsorption/desorption properties for berberine hydrochloride is necessary for identifying suitable mesoporous carbon adsorbents.

The main objective of this work was to explore the adsorption equilibrium, kinetics and column breakthrough of berberine hydrochloride on mesoporous carbons with various pore textures to identify suitable mesoporous carbon adsorbents and to develop an efficient process for separation and purification of berberine hydrochloride using mesoporous carbons. Eleven mesoporous carbon adsorbents with different precursor compositions and carbonization temperatures were synthesized, characterized, and tested for adsorption of berberine hydrochloride from aqueous solutions. The relationship between synthesis conditions and pore textural properties of the mesoporous carbons, and the impacts of pore textural properties on adsorption properties for berberine hydrochloride were discussed, and suitable mesoporous carbon adsorbent was identified. These results obtained are essential for selecting and applying the mesoporous carbons for adsorptive purification of natural products like berberine hydrochloride from herbal materials and effective removal of pollutants like berberine hydrochloride from wastewater.

2. Experimental

2.1 Materials

The triblock copolymer Pluronic P123 and tetraethyl orthosilicate (TEOS, >99% purity) were provided by Sigma-Aldrich, berberine hydrochloride (C₂₀H₁₈ClNO₄, >98% purity) was obtained by Shanghai Darui Finechem Ltd. (Shanghai, China), AR grade phenol, HF, HCl, NaOH, ethanol and formaldehyde solution (36.5–38 wt%) were purchased from Sinopharm Chemical Reagent Co., Ltd. All these materials were used as received.

The molecular structure of berberine hydrochloride as well as its molecular size and optimized molecular structure model performed by the Gaussian 09 software package were listed in Table S1 in the ESI.†

2.2 Synthesis of the mesoporous carbon adsorbents

The mesoporous carbon adsorbents were synthesized via an organic–organic self-assembly method as described previously¹⁸ with a small modification. The synthesis of all the carbon samples was done in parallel and finished in one week.

A soluble phenolic resin (resol) was produced from phenol and formaldehyde solution under alkaline conditions using the method described in our previous work³² and the resultant product was dissolved in ethanol to obtain a 20 wt% resol ethanolic solution.

The synthesis of mesoporous carbon adsorbents was carried out using the preformed resol solution as a carbon precursor, and tetraethyl orthosilicate (TEOS) and triblock copolymer P123 as templates in an ethanol solution. The synthesis compositions were in the range of phenol/formaldehyde/P123/TEOS (molar ratio) = 1 : 0.75 : 0.021 : 1.0–1 : 0.75 : 0.044 : 1.4. In a typical synthesis process, 1.8580 mL of tetraethyl orthosilicate (TEOS) and 1.0005 mL of 0.2 mol L⁻¹ HCl aqueous solution were added into 5.0570 mL of ethanol, after stirring for 5 h at room temperature, 5.0 g of preformed resol solution, 1.0667 g of P123 and 10.1140 mL of ethanol were then added into it, and the mixture was stirred until P123 was completely dissolved. Afterwards, the solution was transferred into a plate to evaporate ethanol at room temperature, then held at 100 °C in air atmosphere for 24 h. The transparent film obtained was calcined under nitrogen atmosphere at 350 °C for 5 h first to remove the triblock copolymer template, then carbonized at 700–900 °C for 4 h, and the ramping rate was 5 °C min⁻¹. The product was soaked in a 10 wt% HF solution for 24 h at room temperature to remove silica component. The template-free carbon product was then filtered, washed with distilled water, dried, and crushed into powders, and the carbon powder obtained with particle size between 0.18 and 0.45 mm was ready for use.

2.3 Characterization of the mesoporous carbons

N₂ adsorption–desorption isotherms of the mesoporous carbon adsorbents were measured at 77 K on a sorptometer Quantachrome Autosorb iQ, the carbon samples were degassed under vacuum at 573 K for 6 h prior to the measurement. BET specific surface areas, pore diameter distributions and pore volumes were calculated by using the Autosorb-IQ built-in software ASIQwin v2.0. Transmission electron microscopy (TEM, Joel JEM-2100F) was used to image the mesoporous structures in the OMC sample. The FT-IR spectroscopy was obtained from a PerkinElmer Spectrum 400 FT-IR/FT-NIR spectrometer.

2.4 Adsorption isotherms

0.0200 g of the mesoporous carbon adsorbent was accurately weighed and mixed with 100 mL of berberine hydrochloride aqueous solution with a known initial concentration C₀ (mg mL⁻¹), and five different initial concentrations were set for each isotherm curve. The mixture was shaken at 298, 308 and 318 K at a constant speed of 200 rpm for 9 h. The final concentration of berberine hydrochloride C_e (mg mL⁻¹) was determined with a Metash UV-5500PC UV-visible spectrophotometer at 345 nm, and the adsorption capacity of berberine hydrochloride on the mesoporous carbon adsorbent Q_e (mg g⁻¹) was calculated as:

Q_e = (C₀ − C_e)V/W (1)

where V is the volume of the solution (mL) and W is the mass of the mesoporous carbon adsorbents (g).

2.5 Adsorption kinetics

0.0700 g of the selected mesoporous carbon adsorbent was weighed and introduced into 250 mL of 0.4 mg mL⁻¹ berberine hydrochloride aqueous solution, the mixture was then continuously shaken at 298, 308 and 318 K until the adsorption equilibrium was established. 0.5 mL of the solution was withdrawn at a preset interval during shaking process, the concentration of berberine hydrochloride at the contact time t C_t (mg mL⁻¹) was analyzed, and the corresponding adsorption capacity Q_t (mg g⁻¹) was calculated as:

Q_t = (C₀ − C_t)V/W (2)

2.6 Dynamic adsorption and desorption

About 3 mL of the selected mesoporous carbon adsorbent was densely packed into a glass column with an inner diameter of 10 mm, and continuously rinsed by de-ionized water before the dynamic experiment. A berberine hydrochloride aqueous solution with an initial concentration of 0.8 mg mL⁻¹ was passed through the mesoporous carbon column at a flow rate of 14 BV per h (0.7 mL min⁻¹, BV means bed volume, 1 BV = 3 mL). The effluent from the column was collected and the concentration of berberine hydrochloride, C (mg mL⁻¹), was analyzed until it was close to the initial concentration.

After that, the mesoporous carbon column was first rinsed with 6 mL of de-ionized water, then 70% of ethanol aqueous solution (v/v) was used to desorb berberine hydrochloride from the mesoporous carbon column, the flow rate was set at 6.4 BV per h, the effluent was collected and the concentration of berberine hydrochloride was analyzed until the concentration approached zero.

3. Results and discussion

3.1 Pore textures of the mesoporous carbons

The synthesis conditions and the main pore textural properties including BET surface areas, pore volumes and average pore diameters of the mesoporous carbon adsorbents are listed in Table 1. The N₂ adsorption and desorption isotherms and pore diameter distributions calculated from the desorption isotherm branches of the mesoporous carbons are shown in Fig. S1 and S2, respectively, in ESI.† Overall, the pore textural properties of the these mesoporous carbon samples are consistent with those of mesoporous carbon materials reported in literatures and our previous work.^18,33 The BET specific surface areas of the mesoporous carbon adsorbents were in the range of 1051.2 to 1554.9 m² g⁻¹, and their pore volumes were measured to be between 1.42 and 2.31 cm³ g⁻¹. The average pore radii of the mesoporous carbon samples were measured to be between 1.53 to 3.28 nm, and mesopores with the radius distributions between 1 and 10 nm are the most predominant pores in the samples. Of all the 11 samples, sample 8-1 with the molar compositions of phenol/formaldehyde/P123/TEOS (molar ratio) = 1 : 0.75 : 0.044 : 1.2 showed the highest BET specific surface area of 1554.9 m² g⁻¹, and sample 9 with the molar compositions of phenol/formaldehyde/P123/TEOS = 1 : 0.75 : 0.044 : 1.4 displayed the largest pore volume of 2.31 cm³ g⁻¹.

Table 1 The synthesis conditions and the main pore textural properties of the mesoporous carbons

Samples	Molar ratio of (phenol : formaldehyde : P123 : TEOS)	Carbonization temperature (°C)	BET specific surface area (m^2 g^−1)	Pore volume (cm^3 g^−1)	Average pore radius (nm)
Sample 1	1 : 0.75 : 0.021 : 1.0	900	1106.0	1.49	1.70
Sample 2	1 : 0.75 : 0.021 : 1.2	900	1051.2	1.42	1.91
Sample 3	1 : 0.75 : 0.021 : 1.4	900	1347.8	2.04	1.71
Sample 4	1 : 0.75 : 0.033 : 1.0	900	1224.3	1.58	1.71
Sample 5	1 : 0.75 : 0.033 : 1.2	900	1365.8	1.91	2.44
Sample 6	1 : 0.75 : 0.033 : 1.4	900	1537.0	1.92	1.70
Sample 7	1 : 0.75 : 0.044 : 1.0	900	1386.6	1.93	1.70
Sample 8–1	1 : 0.75 : 0.044 : 1.2	900	1554.9	2.03	1.53
Sample 8–2	1 : 0.75 : 0.044 : 1.2	800	1526.7	1.80	2.81
Sample 8–3	1 : 0.75 : 0.044 : 1.2	700	1292.9	1.74	3.28
Sample 9	1 : 0.75 : 0.044 : 1.4	900	1533.5	2.31	1.70

---

The BET specific surface areas, pore volumes and average pore radii of the 9 mesoporous carbon samples with various synthesis compositions were plotted in Fig. 1a–c. As shown in Fig. 1a and b, the BET surface areas and pore volumes of the mesoporous carbon samples showed significant differences and a similar tendency, the increase of the TEOS composition as well as the P123 amount has a positive effect on both the BET surface areas and pore volumes, and large BET surface areas and large pore volumes were obtained from the mesoporous carbon samples with a higher TEOS composition and larger P123 amount. However, sample 6, sample 8-1 and sample 9 with the highest P123 amounts and TEOS compositions showed similar BET surface areas, which indicated that a too high P123 and TEOS composition could not further improve the surface areas of the mesoporous carbons. On the other hand, the median pore radii of the mesoporous carbon samples don't present a linear correlation with neither their TEOS compositions nor P123 amounts. These results might attribute to the different effects from the hard template and soft template on the pore textural properties of mesoporous carbons.³³ TEOS was used as a hard template, large amount of TEOS has a positive effect on pore forming to obtain large surface areas and pore volumes for mesoporous carbons.³³ P123 was used as a soft template, large amount of P123 is favourable for expanding pores to obtain large surface areas and pore volumes.³³ In the range of the synthesis compositions for mesoporous carbons in this study, the compositions of both the two templates did not show significant effects on the pore radii of the mesoporous carbons (Fig. 1c).

Fig. 1 The pore textural properties of the mesoporous carbons with different synthesis compositions: (a) BET specific surface area; (b) pore volume; (c) average pore radius.

On the other hand, the mesoporous carbons studied in this work have lower BET specific surface areas (1051.2 to 1554.9 m² g⁻¹) and smaller pore volumes (1.42 to 2.31 cm³ g⁻¹) than those of the mesoporous carbons reported in our previous work (BET surface areas of 1590.3 to 2193.5 m² g⁻¹ and pore volumes of 1.72 to 2.56 m³ g⁻¹) with F127 as a soft template at similar synthesis conditions, while their pore sizes are similar.³³ F127 has a larger molecular weight (12 600) and a longer chain length than those of P123 (molecular weight of 5800), which might lead to a stronger pore expanding effect than P123 to obtain larger surface areas and pore volumes, on the other side, similar pore sizes of the mesoporous carbons with either P123 or F127 as soft templates might attribute to the combination of the pore forming effect from TEOS and the pore expanding effect from F127 or P123.

From the BET specific surface areas, pore volumes and average pore radii of the mesoporous carbon sample 8-1, 8-2 and 8-3 (Table 1) with same synthesis compositions, approximately linear correlations were observed between the pore textural properties and the carbonization temperatures, and it can be found out that with the increase of carbonization temperature, the BET specific surface area as well as the pore volume of the mesoporous carbon samples increased, smaller pore radius was obtained. These results indicated that in a certain range, high carbonization temperature is favourable for producing a more cross-link mesoporous carbon to obtain larger surface area, larger pore volume but narrower pores.

High-resolution TEM images (Fig. 2) of sample 8-1 revealed an amorphous structure of the mesoporous carbon and homogeneous pores throughout the whole sample, these are similar to the reports in the literature.³⁴ The FT-IR spectroscopy of sample 8-1 was displayed in Fig. 3, few functional groups were observed on the surface of the carbon sample 8-1, which might due to the high temperature calcination process under nitrogen atmosphere. The adsorption band observed at 3432 cm⁻¹ is assigned to O–H stretching vibrations, the peaks between 2850 cm⁻¹ and 2920 cm⁻¹ and the band between 1370 cm⁻¹ and 1470 cm⁻¹ could be attribute to C–H groups, the adsorption peaks between 2850 cm⁻¹ and 2920 cm⁻¹ represent C–H stretching vibrations, the peak between 1600 cm⁻¹ and 1700 cm⁻¹ is attribute to C=O functional groups and aromatic C=C, and the week band between 1000 cm⁻¹ and 1200 cm⁻¹ represents C–O bending vibrations, these results suggested that a small amount of O-containing functional groups are present in the mesoporous carbon sample.

Fig. 2 TEM images of the mesoporous carbon sample 8-1.

Fig. 3 FT-IR spectroscopy of the mesoporous carbon sample 8-1.

3.2 Adsorption isotherms of berberine hydrochloride

Adsorption isotherms of berberine hydrochloride on the mesoporous carbon adsorbents are displayed in Fig. 4. It can be found from Fig. 4 that the equilibrium adsorption capacities of berberine hydrochloride on the mesoporous carbon samples are in the range from about 195 to 415 mg g⁻¹, at an equilibrium concentration of 0.10 mg mL⁻¹ and adsorption temperature of 298 K. Generally, the adsorption of berberine hydrochloride on all the mesoporous carbons studied in this work is effective, and the adsorption capacities strongly depend on the pore textural properties of the carbon samples. BET specific surface area and pore volume are known to be two predominant factors influencing adsorption processes in aqueous solution. The mesoporous carbons studied in this work have high BET surface areas (1051.2 to 1554.9 m² g⁻¹), large pore volumes (1.42 to 2.31 cm³ g⁻¹), which could result in the effective adsorption for berberine hydrochloride. Moreover, mesopores with the radius distributions between 1 and 10 nm are the most predominant pores in the mesoporous carbons samples, and the average pore radii are between 1.53 and 3.28 nm which are big enough for the adsorption of berberine hydrochloride with the molecular size of 1.32 nm × 0.66 nm × 0.32 nm.³⁵ However, these adsorption capacity values are less than that of the mesoporous carbons with F127 as a soft template obtained in our previous work³³ at similar conditions, which might because of the even higher BET specific surface areas (1590.3 to 2193.5 m² g⁻¹) and larger pore volumes (1.72 to 2.56 m³ g⁻¹) of the mesoporous carbons with F127 as a soft template.

Fig. 4 Adsorption isotherms of berberine hydrochloride on the mesoporous carbons from aqueous solution: (a) sample 1–3 with the molar composition of P123/phenol at 0.021; (b) sample 4–6 with the molar composition of P123/phenol at 0.033; (c) sample 7, 8-1 and 9 with the molar composition of P123/phenol at 0.044; (d) sample 8-1, 8-2 and 8-3 with the carbonization temperature at 900, 800 and 700 °C and (e) sample 8-1 with the adsorption temperature at 298, 308 and 318 K.

As shown in Fig. 4a–d, sample 8-1, 9 and 6 displayed relatively higher adsorption capacities for berberine hydrochloride at 298 K, while sample 1 served the lowest adsorption capacity, and the adsorption capacities of all the 11 mesoporous carbon samples are generally consistent with the order of their BET surface areas and pore volumes. The mesoporous carbon sample 5, 6 and 7 have similar pore volumes, and their adsorption capacities for berberine hydrochloride were measured to be 261, 380 and 328 mg g⁻¹, respectively, at the equilibrium concentration of 0.10 mg mL⁻¹, which agree the order of their BET surface areas. Additionally, adsorption capacities of berberine hydrochloride on the mesoporous carbon sample 6, 8-2 and 9 with similar BET surface areas were determined to be 380, 347, and 400 mg g⁻¹ at the equilibrium concentration of 0.10 mg mL⁻¹, respectively, which exhibit the same order of their pore volumes. These results further confirmed the dominant roles and positive impacts of the BET surface area and pore volume on the adsorption for berberine hydrochloride from aqueous solutions.

From the isotherms of berberine hydrochloride on the mesoporous carbon sample 8-1 at 298, 308 and 318 K shown in Fig. 4e, it can be found out that the adsorption capacities of berberine hydrochloride increased with increasing temperature from 298 K to 308 K, however, the adsorption amounts decreased with further increase of feed temperature from 308 K to 318 K, and the mesoporous carbon sample displayed similar adsorption capacities for berberine hydrochloride at 298 K and 318 K, which might suggest a complex adsorption mechanism (the combination of physical adsorption and weak chemical adsorption) of berberine hydrochloride on the mesoporous carbon.

Langmuir, Freundlich and Sips isotherm models were employed to analyze the adsorption isotherm data of berberine hydrochloride on the mesoporous carbons, they can be represented as:

where Q_m (mg g⁻¹) is the theoretical maximum adsorption capacity, K_L is the Langmuir constant, K_F and n are the Freundlich constants. K_F is an indicator of the equilibrium adsorption capacity while n is an empirical constant related to the heterogeneity of adsorbent surface and the adsorption intensity.^17,36 The Sips model, known as Langmuir–Freundlich isotherm, is a combination of the Langmuir and Freundlich isotherm models, it reduces to the Freundlich isotherm at low adsorbate concentrations and reduces to the Langmuir isotherm at high adsorbate concentrations.^37,38

The calculated parameters such as Q_m, K_L, K_F and 1/n along with the correlation coefficients R² of Langmuir, Freundlich and Sips equations are listed in Table 2. It is found that all the three models fit the isotherms of berberine hydrochloride on the mesoporous carbons very well, and Sips equation is slightly better than the Langmuir and Freundlich models for describing the adsorption behavior of berberine hydrochloride. However, Q_m values from the Sips model for sample 8-1 at 308 K, 318 K and sample 8-2 at 298 K were considerably higher than the experimental adsorption capacities, which might due to the adsorbate–adsorbate interactions in the adsorption process according to the assumption of the model.³⁹ These results indicate a complex adsorption mechanism of berberine hydrochloride on the mesoporous carbons rather than a simple monolayer adsorption on the active adsorption sites with homogeneous distribution.

Table 2 Fitted results for the adsorption isotherms of berberine hydrochloride according to Langmuir, Freundlich and Sips models

Mesoporous carbon samples	Langmuir model			Freundlich model			Sips model
	Qm/(mg g^−1)	KL /(mL mg^−1)	R^2	KF/[(mg g^−1) (mL mg^−1)^1/n]	1/n	R^2	Qm/(mg g^−1)	KL (mL mg^−1)^1/n	1/n	R^2
Sample 1 at 298 K	297.6	19.35	0.997	355.0	0.2722	0.984	266.6	101.9	1.509	0.999
Sample 2 at 298 K	311.8	78.89	0.999	351.1	0.1204	0.996	314.1	64.16	0.9457	0.999
Sample 3 at 298 K	313.9	63.64	0.997	358.8	0.1392	0.996	340.0	16.97	0.6672	0.998
Sample 4 at 298 K	301.4	108.1	0.999	331.9	0.09486	0.996	303.2	85.79	0.9398	0.999
Sample 5 at 298 K	290.0	109.3	0.998	319.9	0.09458	0.999	368.4	5.01	3.090	0.999
Sample 6 at 298 K	404.3	192.7	0.999	447.0	0.08101	0.995	406.9	137.3	0.3237	0.999
Sample 7 at 298 K	360.1	123.8	0.999	402.4	0.09875	0.997	370.2	50.61	0.7813	0.999
Sample 8-1 at 298 K	436.4	197.7	0.999	487.6	0.08661	0.994	432.8	349.4	1.129	0.999
Sample 8-1 at 308 K	469.3	194.6	0.998	499.8	0.05833	0.999	1007	0.9742	0.1032	0.999
Sample 8-1 at 318 K	429.4	345.9	0.998	446.5	0.03479	0.999	940.7	0.8995	0.06168	0.999
Sample 8-2 at 298 K	427.2	61.54	0.992	520.2	0.1765	0.998	907.8	349.4	0.2861	0.999
Sample 8-3 at 298 K	401.1	99.73	0.999	453.4	0.1134	0.999	386.8	880.1	1.561	0.999
Sample 9 at 298 K	445.6	155.2	0.997	514.0	0.1090	0.997	497.2	15.49	0.5192	0.999

---

Generally, 1/n ≤ 1, stands for a preferential adsorption procedure, which indicates that the adsorption is easy to occur.^17,40 All the mesoporous carbon samples have a 1/n value of less than 0.3 for the adsorption of berberine hydrochloride indicating a favorable adsorption of berberine hydrochloride on all the mesoporous carbons.

Three thermodynamic parameters, adsorption enthalpy (ΔH⁰), adsorption entropy (ΔS) and adsorption free energy (ΔG) of berberine hydrochloride on the mesoporous carbon sample 8-1 were calculated according to the adsorption isotherms of berberine hydrochloride at 298, 308 and 318 K using Langmuir equation.^40–42 The calculated values of ΔG at 298, 308 and 318 K are −27.76, −28.65 and −31.10 kJ mol⁻¹, respectively, and the values of ΔS and ΔH⁰ are 167.1 J mol⁻¹ K⁻¹ and 22.30 kJ mol⁻¹, respectively. It was observed that the ΔG values of berberine hydrochloride at three different temperatures are negative, suggesting that the adsorption of berberine hydrochloride on the mesoporous carbon was thermodynamically spontaneous. Moreover, the positive ΔS value further confirmed the thermodynamical feasibility of the adsorption of berberine hydrochloride on the mesoporous carbon sample. However, the value of ΔH⁰ is positive, which implied that the adsorption of berberine hydrochloride is endothermic, and similar results were also reported in our previous work and earlier literatures.^32,43 The positive ΔH⁰ value confirmed the complex adsorption mechanism of berberine hydrochloride on the mesoporous carbon, indicated that the adsorption of berberine hydrochloride on the mesoporous carbon was not a totally physical process while a weak chemical adsorption might also be involved. The chemical adsorption of berberine hydrochloride on the mesoporous carbon might attribute to the structure of the berberine hydrochloride molecule and the surface chemistry of the mesoporous carbon. Berberine hydrochloride is an alkaloid salt which exists in an ionic state in its aqueous solution.⁴⁴ The residual oxygenic functional groups on the surface of the mesoporous carbon could possible adsorb berberine hydrochloride molecules through weak chemical interactions.

3.3 Adsorption kinetics of berberine hydrochloride on the mesoporous carbon

The mesoporous carbon sample 8-1 has a high BET surface area and large pore volume, and exhibited the largest adsorption capacity toward berberine hydrochloride, therefore was selected for the following kinetic and dynamic adsorption breakthrough tests for berberine hydrochloride.

The adsorption kinetics of berberine hydrochloride on the mesoporous carbon sample 8-1 from aqueous solutions at 298 K, 308 K and 318 K were displayed in Fig. 5. The adsorption of berberine hydrochloride on the mesoporous carbon was shown to be fast, and shorter time was required for establishing adsorption equilibrium at higher temperature. At 298 K, the adsorption rate of berberine hydrochloride was quite fast in the first hour and about 77% of the equilibrium adsorption capacity had been reached, and the final equilibrium was arrived within 240 min. These results exhibited the good kinetic properties of the mesoporous carbon, and implied the efficiency of the mesoporous carbon for adsorptive removal or purification of berberine hydrochloride from aqueous solutions. However, as compared with the time required for reaching the adsorption equilibrium for berberine hydrochloride on sample 8-1 and the mesoporous carbon sample reported in our previous work with F127 as a soft template at 298 K, even shorter time (90 min) was needed for berberine hydrochloride on that mesoporous carbon sample reported in our previous work.³² The mesoporous carbon sample reported in our previous work with F127 as a soft template has an average pore diameter of 4.21 nm,³² which is larger than that of the mesoporous carbon sample 8-1 in this study, larger pore size is more favourable for the diffusion of berberine hydrochloride in the pores of the mesoporous carbon which might lead to a better kinetic property.

Fig. 5 Adsorption kinetics of berberine hydrochloride on the mesoporous carbon sample 8-1 from aqueous solution.

Pseudo-first-order rate equation (Lagergren's rate equation) and pseudo-second-order equation proposed by Ho and McKay,^45,46 were applied to fit the adsorption kinetics of berberine hydrochloride on the mesoporous carbon sample 8-1, the two equations are presented as:

where k₁ and k₂ are the pseudo-first-order rate constant (min⁻¹) and the pseudo-second-order rate constant (g mg⁻¹ min⁻¹), Q_t (mg g⁻¹) and Q_e (mg g⁻¹) are the adsorption capacities at uptake time t (min) and equilibrium, respectively, in the pseudo-second-order rate equation, v₀ is the initial adsorption rate, ν₀ = k₂Q_e².

The fitted parameters including k₁, k₂, Q_e, v₀ as well as the correlation coefficients R² from both kinetic equations were summarized in Table 3. It can be found that the pseudo-second-order rate equation is slightly better than the pseudo-first-order rate equation for describing the kinetic data of berberine hydrochloride. Meanwhile, the initial adsorption rate v₀ calculated in Table 3 further confirmed the high efficiency of the mesoporous carbon for adsorption of berberine hydrochloride from aqueous solutions at all the three selected temperatures, and the v₀ values clearly increased with increasing temperature, which suggested the positive effect of temperature on the adsorption efficiency.

Table 3 Parameters in the pseudo-first-order and pseudo-second-order rate equations for berberine hydrochloride on the mesoporous carbon sample 8-1

Temperature/(K)	Pseudo-first-order			Pseudo-second-order
	Qe (mg g^−1)	k1 (min^−1)	R^2	Qe (mg g^−1)	k2	v0 (mg g^−1 min^−1)	R^2
298	391.6	3.245 × 10^−2	0.966	428.8	1.100 × 10^−4	20.23	0.995
308	426.8	3.599 × 10^−2	0.980	461.1	1.200 × 10^−4	25.51	0.979
318	409.4	5.632 × 10^−2	0.979	435.0	2.300 × 10^−4	43.52	0.988

---

3.4 Dynamic adsorption and desorption of berberine hydrochloride

The dynamic adsorption breakthrough curve and desorption curve of berberine hydrochloride on the mesoporous carbon sample 8-1 at room temperature were displayed in Fig. 6(a and b). As shown in Fig. 6a, the breakthrough point (C/C₀ = 0.05) and total adsorption volume (C/C₀ = 0.95) of berberine hydrochloride from the mesoporous carbon column were measured to be 120 BV and 192 BV, respectively, and the dynamic adsorption capacity calculated from the numerical integration of the breakthrough curve was to be 425.3 mg g⁻¹, which is close to the theoretical maximum adsorption equilibrium capacity (436.4 mg g⁻¹) calculated from the isotherm data of berberine hydrochloride on this sample at 298 K according to Langmuir equation.

Fig. 6 Adsorption breakthrough curve of berberine hydrochloride on the mesoporous carbon sample 8-1 at room temperature (a); desorption curve of the saturated mesoporous carbon column at room temperature (b).

Various mathematical models have been developed to predict the dynamic adsorption behavior of the column. Adams–Bohart, Thomas, Yoon and Nelson, and Clark models were used to analyze the dynamic adsorption breakthrough data of berberine hydrochloride,^39,47–50 they are presented as:

where C (mg mL⁻¹) is the outlet concentration of berberine hydrochloride. In the Adams–Bohart model, k is the rate constant [mL min⁻¹ mg⁻¹], N₀ is the adsorption capacity (mg mL⁻¹), Z is the height of the adsorbent bed (cm) and U is the linear flow rate (cm min⁻¹). In the Thomas model, k_T is the Thomas rate constant [mL min⁻¹ mg⁻¹], q₀ is the maximum adsorption capacity (mg g⁻¹), m_c is the weight of the adsorbent (g) and Q is the flow rate (mL min⁻¹). In the Yoon and Nelson model, k_YN is the rate constant (min⁻¹) and t_0.5 is the time required for reaching C/C₀ = 0.50. In the Clark model, n is the Freundlich constant, A and r are the Clark constants related to the breakthrough point and the shape of the breakthrough curve.

The parameters calculated from the four models together with the correlative coefficients R² were summarized in Table 4. It can be seen that the correlation coefficients for all the four models exceeded 0.99, which indicated that all the four models are suitable for predicting the dynamic behaviour of berberine hydrochloride to be adsorbed from solution in a fixed mesoporous carbon column bed. The N₀ calculated from the Adams–Bohart model was 118.7 mg mL⁻¹ (430.7 mg g⁻¹), and the q₀ calculated from the Thomas model was 430.8 mg g⁻¹, both of them were close to the experimental dynamic adsorption capacity, additionally, the t_0.5 value from the Yoon and Nelson model was calculated to be 636.0 min, which was also close to the experimental value of 630.0 min. These results further confirmed the validity of these three models and the high adsorption efficacy of the mesoporous carbon adsorbent for berberine hydrochloride.

Table 4 The fitted results for the dynamic breakthrough curves of berberine hydrochloride on the mesoporous carbon sample 8-1 according to the Adams–Bohart, Thomas, Yoon and Nelson, and Clark models

Adams–Bohart model			Thomas model
k (mL min^−1 mg^−1)	N0 (mg mL^−1)	R^2	kT (mL min^−1 mg^−1)	q0 (mg g^−1)	R^2
0.02501	118.7	0.999	0.02501	430.8	0.999

---

Yoon and Nelson model			Clark model
kYN (min^−1)	t0.5 (min)	R^2	A	r	R^2
0.02001	636.0	0.999	558.2	0.01431	0.999

---

As can be seen from Fig. 6b, about 60 bed volume of 70% (v/v) ethanol aqueous solution could elute about 72.4% of berberine hydrochloride from the mesoporous carbon column, and 100 bed volume was needed to reach a desorption ratio of 82.9% for berberine hydrochloride, these results implied that the adsorbed mesoporous carbon adsorbent can be regenerated and reused easily.

4. Conclusions

Eleven mesoporous carbon adsorbents were synthesized, characterized and evaluated their adsorption/desorption properties of berberine hydrochloride from aqueous solutions. The mesoporous carbon adsorbents synthesized in this work exhibit high BET specific surface areas (1051.2 to 1554.9 m² g⁻¹), large pore volumes (1.42 to 2.31 m³ g⁻¹), and their average pore radii are between 1.53 nm and 3.28 nm. Pore textural properties of the mesoporous carbon adsorbents strongly effects the adsorption of berberine hydrochloride from aqueous solution, and the mesoporous carbon sample with the largest surface area showed the highest adsorption capacity (415 mg g⁻¹) at 298 K and 0.10 mg mL⁻¹ for berberine hydrochloride and was selected as the most promising adsorbent. Moreover, the adsorption rate of berberine hydrochloride on the selected mesoporous carbon is fast, less than 240 min is needed to reach equilibrium at 298 K. The dynamic adsorption tests indicated the effective dynamic adsorption and easy regeneration of the selected mesoporous carbon for berberine hydrochloride. The mesoporous carbons especially the selected carbon sample studied in this work are promising adsorbent materials for adsorptive removal and purification of berberine hydrochloride and similar alkaloids from aqueous solutions.

Acknowledgements

This work was partially supported by the Zhejiang Provincial Natural Science Foundation of China (No. LQ15B060003), the Foundation of Key Laboratory of Biomass Chemical Engineering of Ministry of Education, Zhejiang University, the Zhejiang Province Research Project of Public Welfare Technology Application (No. 2015C33006), and the Analysis & Test Funding Project of Zhejiang Province (No. 2015C37059).

Notes and references

1. X. Luo, J. Li, L. Guo, X. Cheng, T. Zhang and Y. Deng, Asian J. Pharm. Sci., 2013, 8, 261–266.
2. Y. Wang, Q. Liu, Z. Liu, B. Li, Z. Sun, H. Zhou, X. Zhang, Y. Gong and C. Shao, Mutat. Res., 2012, 734, 20–29.
3. J. Lan, Y. Zhao, F. Dong, Z. Yan, W. Zheng, J. Fan and G. Sun, J. Ethnopharmacol., 2015, 161, 69–81.
4. E. Küpeli, M. Koşar, E. Yeşilada and K. H. C. Başer, Life Sci., 2002, 72, 645–657.
5. A. Shirwaikar, A. Shirwaikar, K. Rajendran and I. S. R. Punitha, Biol. Pharm. Bull., 2006, 29, 1906–1910.
6. H. Li, Y. Li, Z. Li, X. Peng, Y. Li, G. Li, X. Tan and G. Chen, Appl. Surf. Sci., 2012, 258, 4314–4321.
7. M. Ren, Y. Song, S. Xiao, P. Zeng and J. Peng, Chem. Eng. J., 2011, 169, 84–90.
8. Y. Li, J. Huang, J. Liu, S. Deng and X. Lu, J. Chem. Eng. Data, 2013, 58, 1271–1279.
9. J. Votruba and M. Flieger, Biotechnol. Lett., 2000, 22, 1281–1285.
10. D. X. Wang, A. Sakoda and M. Suzuki, Adsorption, 1999, 5, 97–108.
11. G. Rytwo, H. Varman, N. Bluvshtein, T. N. König, A. Mendelovits and A. Sandler, Appl. Clay Sci., 2011, 51, 43–50.
12. E. Cohen, T. Joseph, I. Lapides and S. Yariv, Clay Miner., 2005, 40, 223–232.
13. J. Górka and M. Jaroniec, J. Phys. Chem. C, 2010, 114, 6298–6303.
14. J. Cai, S. Bennici, J. Shen and A. Auroux, Water, Air, Soil Pollut., 2014, 225, 2088.
15. K. Wang, B. Huang, D. Liu and D. Ye, J. Appl. Polym. Sci., 2012, 125, 3368–3375.
16. M. Anbia and S. Salehi, Chem. Eng. Res. Des., 2012, 90, 975–983.
17. X. Peng, D. Huang, T. Odoom-Wubah, D. Fu, J. Huang and Q. Qin, J. Colloid Interface Sci., 2014, 430, 272–282.
18. X. Zhuang, Y. Wan, C. Feng, Y. Shen and D. Zhao, Chem. Mater., 2009, 21, 706–716.
19. G. Zeng, Y. Liu, L. Tang, G. Yang, Y. Pang, Y. Zhang, Y. Zhou, Z. Li, M. Li, M. Lai, X. He and Y. He, Chem. Eng. J., 2015, 259, 153–160.
20. B. Acevedo and C. Barriocanal, RSC Adv., 2015, 5, 95247–95255.
21. G. Yang, L. Tang, Y. Cai, G. Zeng, P. Guo, G. Chen, Y. Zhou, J. Tang, J. Chen and W. Xiong, RSC Adv., 2014, 4, 58362–58371.
22. Z.-b. Zhang, Y.-d. Zhou, Y.-H. Liu, X.-h. Cao, Z.-w. Zhou, B. Han, P. Liang and G.-x. Xiong, J. Radioanal. Nucl. Chem., 2014, 302, 9–16.
23. A. Vinu, K. Z. Hossian, P. Srinivasu, M. Miyahara, S. Anandan, N. Gokulakrishnan, T. Mori, K. Ariga and V. V. Balasubramanian, J. Mater. Chem., 2007, 17, 1819–1825.
24. Z. Wang, X. Liu, M. Lv and J. Meng, Carbon, 2010, 48, 3182–3189.
25. Y. Meng, D. Gu, F. Zhang, Y. Shi, L. Cheng, D. Feng, Z. Wu, Z. Chen, Y. Wan, A. Stein and D. Zhao, Chem. Mater., 2006, 18, 4447–4464.
26. Q. Wang, W. Zhang, Y. Mu, L. Zhong, Y. Meng and Y. Sun, Microporous Mesoporous Mater., 2014, 197, 109–115.
27. S. Tanaka, N. Nishiyama, Y. Egashira and K. Ueyama, Chem. Commun., 2005, 28, 2125–2127.
28. C. Liang, Z. Li and S. Dai, Angew. Chem., Int. Ed., 2008, 47, 3696–3717.
29. C. Liang and S. Dai, J. Am. Chem. Soc., 2006, 128, 5316–5317.
30. W. Libbrecht, F. Deruyck, H. Poelman, A. Verberckmoes, J. Thybaut, J. De Clercq and P. Van Der Voort, Chem. Eng. J., 2015, 259, 126–134.
31. J. Jiao, Y. Cai, P. Liu and P. Liu, J. Porous Mater., 2013, 20, 1247–1255.
32. Y. Li, B. Yuan, J. Fu, S. Deng and X. Lu, J. Colloid Interface Sci., 2013, 408, 181–190.
33. Y. Li, J. Fu, S. Deng and X. Lu, J. Colloid Interface Sci., 2014, 424, 104–112.
34. B. Yuan, X. Wu, Y. Chen, J. Huang, H. Luo and S. Deng, J. Colloid Interface Sci., 2013, 394, 445–450.
35. Y. Li, R. Cao, X. Wu, J. Huang, S. Deng and X. Lu, J. Colloid Interface Sci., 2013, 400, 78–87.
36. Y. Fu, Y. Zu, W. Liu, C. Hou, L. Chen, S. Li, X. Shi and M. Tong, J. Chromatogr. A, 2007, 1139, 206–213.
37. A. B. Pérez-Marín, V. M. Zapata, J. F. Ortuño, M. Aguilar, J. Sáez and M. Lloréns, J. Hazard. Mater., 2007, 139, 122–131.
38. J. He, R. Lin, H. Long, Y. Liang and Y. Chen, J. Colloid Interface Sci., 2015, 454, 144–151.
39. Z. Aksu, Process Biochem., 2005, 40, 997–1026.
40. R. A. Chi, F. Zhou, K. Huang and Y. F. Zhang, Chin. J. Nat. Med., 2011, 9, 120–125.
41. M. D. Saikia and N. N. Dutta, Colloids Surf., A, 2006, 280, 163–168.
42. J. Goscianska, M. Marciniak and R. Pietrzak, Chem. Eng. J., 2014, 247, 258–264.
43. S. Wang, Y. Boyjoo, A. Choueib and Z. H. Zhu, Water Res., 2005, 39, 129–138.
44. J. Dostál, J. Chem. Educ., 2000, 77, 993–998.
45. C. Tien and B. V. Ramarao, Sep. Purif. Technol., 2014, 136, 303–308.
46. Y. S. Ho and G. McKay, Process Biochem., 1999, 34, 451–465.
47. G. S. Bohart and E. Q. Adams, J. Am. Chem. Soc., 1920, 42, 523–544.
48. Y. H. Yoon and J. H. Nelson, Am. Ind. Hyg. Assoc. J., 1984, 45, 509–516.
49. H. C. Thomas, J. Am. Chem. Soc., 1944, 66, 1664–1666.
50. R. M. Clark, Environ. Sci. Technol., 1987, 21, 573–580.

---

Footnote

† Electronic supplementary information (ESI) available: Table S1, Fig. S1 and S2. See DOI: 10.1039/c6ra01257d

---