Zwitterionic Copolymer-Modified Mesoporous Silica Material: A Promising Adsorbent for Both Anionic and Cationic Dye Removal and an Efficient Stationary Phase for Liquid Chromatography

Abstract

The creation of innovative single materials with a variety of applications is a major need for materials in the future. The decolorization of both anionic and cationic dye-contaminated water with high adsorption capacities is one of the most challenging problems for the existing adsorbents. At the same time, separating highly polar compounds with high selectivity is also challenging with the existing stationary phases in liquid chromatography. The new material was designed with two important monomers such as methyl acrylate (MA) containing a carbonyl group and sulfobetaine methacrylate (SBMA) containing zwitterions (anionic and cationic sites) to interact with both anions and cations in dye-contaminated wastewater as well as for electrostatic interactions with polar analytes for their separation. 3-Mercaptopropyltrimethoxysilane (MPS) end-functionalized copolymer (telomer) of MA and SBMA was synthesized, characterized (by ¹H NMR), and then immobilized onto silica by the “grafting to” method to obtain Sil-(MA₁₀-co-SBMA₁₀). After immobilization onto silica, the novel material was characterized by elemental analysis, attenuated total reflectance Fourier transform infrared spectroscopy (ATR-FTIR), scanning electron microscopy (SEM), and thermogravimetric analysis (TGA). Cationic methylene blue (MB) was effectively adsorbed onto Sil-(MA₁₀-co-SBMA₁₀), with the highest adsorption capacity of 2000.00 mg/g at varying pH. Similarly, anionic Congo Red (CR) showed an adsorption capacity of 1666.67 mg/g under different pH conditions. The Sil-(MA₁₀-co-SBMA₁₀) material showed the highest adsorption capacities among the reported silica-based adsorbents. The material’s zeta potential was evaluated to identify the adsorption technique’s interaction mechanism. For hydrophilic interaction chromatography (HILIC), the same material was simultaneously packed onto a 150 × 2.1 mm HPLC column and assessed with Tanaka test probes. Remarkably, when contrasting it with the most popular commercial HILIC columns, the best segregation was achieved using the Sil-(MA₁₀-co-SBMA₁₀). Therefore, it will create a new way of thinking to design materials for more than one application.

Introduction

Zwitterionic polymers feature repeating units with both positive and negative charges, residing in the main chain or side chains. By adding additional charged or neutral monomers, referred to as zwitterionic copolymers, the characteristics of these polymers may be adjusted. Zwitterionic polymers and copolymers have fascinating properties such as very high hydrophilicity, pH- and salt-responsiveness, thermoresponsive behavior, and self-coacervation behavior. Zwitterionic polymers form a surface hydration layer due to the solvation of both anionic and cationic groups and therefore can bind huge amounts of water/hydration around the polymer through electrostatic interactions. Moreover, even after the grafting of zwitterionic polymers onto silica, gold, and glass, they can produce a highly hydrophilic surface. Due to the above exceptional properties, zwitterionic polymers/copolymers have applications in antifouling coatings, polyampholytes as dispersants and stabilizers, hydrogel materials, drug delivery, antifreeze materials, stationary phases for liquid chromatography, − and gene therapy.

Bangladesh has become one of the major textile-producing countries in the world. At the same time, the textile industry is growing in Bangladesh promptly. A growing and more extensive history of subpar working conditions and significant pollution has resulted from this expansion and the quickening pace of fashion manufacturing. This quicker economic development is driven by a rise in inbound remittances and an increase in the number of garment exports. For every kilogram of textile produced, typically 250 to 300 L of water are used for washing, dyeing, and finishing; that is the daily water requirement of two people. This water becomes seriously polluted after these steps, and if this kind of water is discharged without treatment, serious environmental pollution occurs. These disturb the ecological balance as well as being detrimental to human health. , Conversely, the amount of drinkable water is decreasing day by day, and it will be a big challenge for the 21st century. , Hence, wastewater treatment has become very important for the sustainable development of our society. In general, the activated sludge biodegradation or flocculation increases the effectiveness of pollutant removal from wastewater. , Nevertheless, the color is yet to remain pronounced even after the reduction of chemical oxygen demands (CODs) and efficient removal of pollutants. Dyes are difficult to remove, and a trace amount of dye can show a visible color. Adsorption is a comparatively suitable process for the elimination of dyes, as it does not produce secondary pollutants or other reaction products, and the dyes can be recovered. However, finding an adsorbent capable of adsorbing both anionic and cationic dyes with high adsorption capacities is challenging.

In high-performance liquid chromatography (HPLC), the packing material or stagnant phase is crucial for efficient separation. Usually, mesoporous silica surfaces modified with alkyl chains or embedded polar groups containing alkyl chains are used as packing materials for reversed-phase HPLC (RP-HPLC). However, these materials have limitations such as peak tailing for highly polar analytes. − Therefore, an alternative method and packing material were proposed for the study of highly polar substances in 1990 by Alpert and named hydrophilic interaction liquid chromatography (HILIC). There are many HILIC packing materials or columns that have been developed and are commercially available, such as zwitterionic, amide, diol, and amine. , Moreover, materials for the combination of RP-HPLC, HILIC, and sometimes others, called mixed-mode, have also been developed for the analysis of diverse analytes. − Nonetheless, it is still very desirable to produce materials with improved performance and a variety of uses.

In this work, a new material was designed with two important monomers, such as methyl acrylate (MA) containing a carbonyl group and sulfobetaine methacrylate (SBMA) containing zwitterions (anionic and cationic sites). The MA is suitable for carbonyl−π interactions and SBMA has the exact properties of a zwitterionic monomer as mentioned earlier. The designed materials are supposed to work successfully as an adsorbent and HILIC phase due to the multipole interaction sites required for dual functionalities. Initially, 3-mercaptopropyltrimethoxysilane (MPS) end-functionalized copolymer (telomer) of MA and SBMA was developed and embedded onto mesoporous silica by “grafting to” method to get Sil-(MA₁₀-co-SBMA₁₀) as shown in Figure . The choice of “grafting to” method rather than “grafting from” method was due to the ease of controlling the degree of polymerization and the analysis of the telomer without removing the copolymer from the silica. The material’s ability to adsorb cationic (MB) and anionic (CR) dyes at various pH values was assessed. Ultrahigh adsorption capacities were observed with the new material. In parallel, the substance was packed into a stainless steel narrow-bore column for use as an HILIC phase in HPLC. Evaluation of the HILIC column was carried out with the Tanaka test protocol for HILIC. Comparatively, higher separation ability was observed with the Sil-(MA₁₀-co-SBMA₁₀) column than with the most commonly used zwitterionic ZIC-HILIC column. ,

1.

Synthesis of telomer or MPS-end functionalized copolymer with the monomers MA and SBMA (MA₁₀-co-SBMA₁₀). Grafting of the telomer onto silica to obtain the new material Sil-(MA₁₀-co-SBMA₁₀).

Experimental Section

Materials and Reagents

Methyl acrylate (MA), sulfobetaine methacrylate (SBMA), 3-mercaptopropyltrimethoxysilane (MPS), and benzoyl peroxide (BPO) were received from Sigma-Aldrich (Darmstadt, Germany) and Merck (Darmstadt, Germany). Different solvents such as tetrahydrofuran, methanol, diethyl ether, chloroform, and ethanol were acquired from VWR (Darmstadt, Germany).

The silica particles with a pore size of 10 nm, a diameter of 3 μm, and a surface area of 331 m²/g of were acquired from Dr. Maisch GmbH, Germany. Hydrochloric acid, sodium hydroxide, Congo Red (CR), and methylene blue (MB) were obtained from Merck (Darmstadt, Germany). The dye concentrations were determined during the adsorption experiments by using the UV-1800 spectrophotometer (Shimadzu Corporation, Japan). HILIC samples like Tanaka test probes (uridine, 2′-deoxyuridine, 5-methyluridine, theobromine, sodium p-toluenesulfonate, vidarabine, theophylline, adenosine, N,N,N-trimethylphenylammonium chloride, adenine, 2′-deoxyguanosine, cytosine, 3′-deoxyguanosine, cytidine, 4-nitrophenyl μ-d-glucopyranoside, and 4-nitrophenyl β-d-glucopyranoside) were obtained from Merck (Darmstadt, Germany) and Sigma-Aldrich (St. Louis, MO, USA). All test solutions were made by dissolving the samples in methanol, acetonitrile, or their mixture with water to a concentration of 10 mg/mL based on their solubility, then diluted to 0.1 mg/mL, and finally filtered through 0.45 μm syringe filters before injection.

After characterization, in order to remove dyes from aqueous solutions contaminated with dyes, half of the produced material, Sil-(MA10-co-SBMA10), was employed as an adsorbent. The remaining one was packed into a narrow-bore stainless steel column (150 mm × 2.1 mm i.d.). Reference HILIC columns, sulfobetaine (ZIC-HILIC from Merck, Germany), Luna-NH₂ (Phenomenex, USA), Luna-Diol (Phenomenex, USA), and iHILIC-Fusion (dichrom GmbH, Germany), were used to compare the chromatographic findings with the Sil-(MA₁₀-co-SBMA₁₀) column. To compare the chromatographic results, the physical characteristics of column materials are discussed in the Supporting Information (Table S1). In all chromatographic analyses, LC–MS quality solvents were utilized. Acetonitrile (ACN) was obtained from VWR (Darmstadt, Germany). Purification of tap water was completed with the Ultrapure water (resistivity 18.2 MΩ cm^–1) purchased from the Sartorius Ultrapure Water System (Goettingen, Germany).

Synthesis of Telomer (MA-co-SBMA)

50 mL of methanol were poured into a round-bottom flask containing MA (3.44 g, 0.04 mol), SBMA (11.1 g, 0.04 mol), and BPO (100 mg). To eliminate oxygen from the system, the mixture was bubbled with nitrogen gas for an hour. As seen in Figure , the mixture was subjected to heat at 70 °C for 4 h after the addition of the silane coupling agent, MPS (0.99 mL, 0.005 mol). The precipitated telomer was filtered and repeatedly cleaned with methanol to get rid of the unreacted monomers. After drying, the telomer was characterized by ¹H NMR to ascertain the degree of polymerization (Supporting Information, Figure S1).

Immobilization of the Telomer onto Silica

The telomer (5.0 g) and silica (5.0 g) were placed in a round-bottom flask with 50 mL of methanol. For 72 h, the mixture underwent reflux. Finally, the mixture was centrifuged to obtain telomer-grafted silica (Sil-(MA₁₀-co-SBMA₁₀)) (Figure ) and to remove the solvent and unreacted telomer. The detailed characterization techniques are provided in the Supporting Information.

Proton NMR

¹H NMR were captured with a Bruker DRX500 (Germany) instrument at 400 MHz in CDCl₃ solutions. Chemical shifts (δ) of ¹H were measured in parts per million (ppm).

Scanning Electron Microscopy (SEM) and Energy-Dispersive Spectroscopy (EDX)

Images from a scanning electron microscope (SEM) were captured with an Apreo S LoVac, Thermo Fisher Scientific (USA). Prior to analysis, each sample was covered with a 6 nm thick layer of gold–platinum coating. The surface and energy-dispersive spectroscopy (EDX) analysis of the adsorbents were performed using the Zeiss Sigma 300 VP FE-SEM machine.

FT-IR and Elemental Analysis

FT-IR experiments were performed with a Bruker (Germany) model ALPHA-Platinum FT-IR spectrometer with an ATR sample module. Elemental analyses were conducted using a EURO EA elemental analyzer (EA3100) and a EURO VECTOR (Italy).

Thermogravimetric Analysis (TGA)

The TGA experiments were accomplished with the HITACHI (Japan) Model STA7200, Thermal Analysis System. It was conducted in a steady N₂ gas flow (200 mL/min) at a temperature from 30 to 800 °C and a heating rate of 10 °C/min.

Zeta Analyzer

Zeta potential experiments were conducted by using a BM10 measurement instrument and a Litesizer 500 (Anton Paar GmbH, Austria). The Smoluchowski approximation was employed to calculate electrophoretic mobility and zeta potential through laser Doppler microelectrophoresis. A 658 nm laser source was employed to detect particle velocity under an electric field using a folded capillary cell (Omega cuvette, Mat. No. 225288). The dispersing medium was water (refractive index: 1.330, viscosity: 0.890 mPa·s), and the studies were carried out at 25 ± 0.1 °C with automated voltage adjustment (maximum 200 V). Before analysis, each sample was equilibrated for 2 min, and the average of 100 processed cycles was calculated.

Adsorption Studies

The effects of pH, dosage of adsorbent, initial concentration of dyes, and contact time were investigated since these factors can affect the adsorption of dye onto the copolymer-immobilized silica (Sil-(MA₁₀-co-SBMA₁₀)). To ascertain the equilibrium adsorption capacities of Sil-(MA₁₀-co-SBMA₁₀) at varying pH levels (ranging from 3 to 13), initial concentrations of dyes (50, 100, 150, 200, and 250 mg/L), contact times (ranging from 0 to 50 min), and amounts of Sil-(MA₁₀-co-SBMA₁₀) applied as adsorbent (0.002, 0.003, 0.004, 0.005, and 0.006 g) at room temperature, a specific quantity was added to 30 mL of MB and CR dye solution and shaken reciprocally at a rate of 180 strokes per minute for 50 min. Using a UV spectrophotometer, the initial and final dye concentrations of CR and MB were calculated. The Congo Red calibration curve at 498 nm and the methylene blue calibration curve at 660 nm were used in converting the adsorption values of CR and MB, respectively, to their equivalent concentrations. The calibration curves of MB and CR are included in the Supporting Information (Figure S2 and Figure S3, respectively). The equilibrium adsorption capacity Q_e and the percentage of dye removal by the Sil-(MA₁₀-co-SBMA₁₀) were calculated using eqs and .

$$\mathrm{A}\mathrm{d}\mathrm{s}\mathrm{o}\mathrm{r}\mathrm{p}\mathrm{t}\mathrm{i}\mathrm{o}\mathrm{n}\mathrm{c}\mathrm{a}\mathrm{p}\mathrm{a}\mathrm{c}\mathrm{i}\mathrm{t}\mathrm{y},{\mathrm{Q}}_{\mathrm{e}}=({\mathrm{C}}_{\mathrm{o}}-{\mathrm{C}}_{\mathrm{e}})\times (\mathrm{V}/\mathrm{m})$$ 1

$$\mathrm{D}\mathrm{y}\mathrm{e}\mathrm{r}\mathrm{e}\mathrm{m}\mathrm{o}\mathrm{v}\mathrm{a}\mathrm{l},\%=100\times ({\mathrm{C}}_0-{\mathrm{C}}_{\mathrm{e}})/{\mathrm{C}}_0$$ 2

where C₀ is the initial concentration (milligrams per liter) of dye for both MB and CR, C_e is the concentration at equilibrium of free MB or free CR dye molecules in the solution (milligrams per liter), Q_e (adsorption capacity at equilibrium) is the quantity of CR or MB dye molecules adsorbed by 1 g of Sil-(MA₁₀-co-SBMA₁₀) adsorbent (milligrams per gram), V is the volume of CR or MB dye solution (liters), and m is the mass of Sil-(MA₁₀-co-SBMA₁₀) adsorbent dosage (grams). With 0.005 g (dry weight basis) of the Sil-(MA₁₀-co-SBMA₁₀) in various initial concentrations of dye (50, 100, 150, 200, and 250 mg/L) and reciprocating shaking at 180 rpm for 50 min at pH 7 at room temperature, the adsorption isotherms and kinetic tests of the Sil-(MA₁₀-co-SBMA₁₀) were conducted. Plotting the equilibrium dye uptake versus the equilibrium dye concentration and fitting the data using a range of mathematical models allowed for the calculation of the adsorption isotherms. Using the dye uptake rate discovered during the tests, we also examined several kinetic models.

Liquid Chromatography

An Agilent 1290 Infinity II (Agilent, Walbronn, Germany), with a 1290 Infinity II autosampler (Vialsampler) module (G7129C), a 1260 Infinity II flexible pump (G7104C), a 1290 Infinity II MCT column compartment (G7116A), and a 1260 Infinity II DAD FS detector module (G7117A), was used for chromatographic analysis. The operation and control of the arrangement were performed by applying the program OpenLAB ChemStation Edition (Version C.01.07 SR3, Agilent, Santa Clara, USA). An ACN–ammonium acetate buffer (pH = 4.7, 20 mM in the aqueous portion) (90:10), a flow rate of 0.2 mL per minute, and a column temperature of 30 °C were applied for the isolation of Tanaka test probes. The disturbance caused by toluene injection was utilized as t ₀ or a dead time marker. The retention factors were measured as k = (t _R – t ₀)/t ₀, where t _R is the retention time of the interested analyte. The separation factors of two analytes were calculated as α = k of the most retained analyte of interest/k of the least retained analyte of interest. Tanaka test probes were analyzed in accordance with Table S2 in the Supporting Information. Each sample injection was performed at least in triplicate.

Results and Discussion

Characterization of the Material

Materials made of mesoporous silica have been functionalized with different ligands so that they may be utilized independently as adsorbents to treat wastewater , and as the HILIC stagnant phase for the isolation of polar analytes. End-functionalized copolymer/telomer with MA and SBMA was synthesized, and the degree of polymerization was identified to be 10 for both MA and SBMA units by ¹H NMR and represented as (MA₁₀-co-SBMA₁₀) (Figure ). After immobilization onto silica, the material Sil-(MA₁₀-co-SBMA₁₀) was characterized by elemental analysis, scanning electron microscopy (SEM), attenuated total reflectance Fourier transform infrared spectroscopy (ATR-FTIR), and thermogravimetric analysis (TGA). After characterization, half of the material was employed as an adsorbent for removing the anionic and cationic dyes, and half of the material was packed into an HPLC column for HILIC mode isolation of polar analytes. The elemental analysis data indicate the grafting of the zwitterionic copolymer onto silica (Table ). The degree of polymerization was also confirmed from the C/N ratio of the found and calculated data, as described in Table .

1. Results of Elemental Analysis of Copolymer-Immobilized Silica (Sil-(MA₁₀-co-SBMA₁₀).

	C (%)	H (%)	N (%)	C/N (found)	C/N (calculated)
Sil-(MA10-co-SBMA10)	13.50	2.60	0.81	16.56	16.66

The silica’s particle size of 3 μm was verified using the SEM. Moreover, it was observed that there was no broken silica due to stirring since it is crucial for usage as a column packing material (Supporting Information, Figure S4). Figure S5 (Supporting Information) illustrates the ATR-FTIR spectra of bare silica and Sil-(MA₁₀-co-SBMA₁₀). The characteristic band at 1724 cm^–1 is attributed to CO from both polymers MA and SBMA. A shoulder at 1649 cm^–1 indicates the C–N⁺ from the polymer SBMA. Moreover, the band at 1474 cm^–1 shows CH₃ stretching from both monomers. Therefore, the telomer was grafted perfectly onto the silica surfaces. Agreeing with the ATR-FTIR results, TGA also showed the perfect grafting of the telomer onto silica (Supporting Information, Figure S6), and the grafting amount (24.8%) is sufficient for any application of the material.

Equilibrium Adsorption Studies

To determine its adsorption properties, the newly synthesized material, Sil-(MA₁₀-co-SBMA₁₀), was added to aqueous solutions of anionic CR and cationic MB dyes. As described in the Supporting Information (Figure S7), when the aqueous solutions of MB and CR were compared after 50 min, the Sil-(MA₁₀-co-SBMA₁₀) showed bright blue and red coloring, respectively, in contrast to the initially developed solutions.

Effect of pH on Dye Adsorption

Making use of a 0.005 g Sil-(MA₁₀-co-SBMA₁₀) adsorbent and an initial concentration of dye of 150 mg/L in a 30 mL solution, the impact of pH on the adsorption at equilibrium of MB and CR has been studied in the pH range of 3 to 13. However, the highest quantity of MB adsorption by Sil-(MA₁₀-co-SBMA₁₀) was observed at pH 7, as shown in Figure a. While MB adsorption by Sil-(MA₁₀-co-SBMA₁₀) was favorable in the basic medium and approached the maximum value, the adsorption values decreased in the acidic medium. This indicates that, possibly because of the presence of the adsorbent’s sulfite group sites in the SBMA monomer as it is negatively charged, it worked most effectively for MB adsorption in neutral pH and basic conditions. The findings demonstrate that the negatively charged areas on the adsorbent’s surface become extremely reactive with the cationic MB molecules at neutral and basic pH values. Moreover, there is a possibility of carbonyl−π interactions from both MA and SBMA monomers with the MB dye, and therefore, it is possible even at neutral pH. , However, as shown in Figure b, the maximum CR adsorption by Sil-(MA₁₀-co-SBMA₁₀) was recorded at pH 3 and thereafter declined as pH increased to 13, suggesting that the adsorbent is best suited for acidic environments, perhaps as a result of the positively charged amine group sites in the SBMA monomer. The results demonstrated that the positively charged sites on the adsorbent’s surface are highly reactive toward the CR molecules when the pH is acidic, because pH 3 has the maximum adsorption capability.

2.

Effect of pH on the adsorption of MB (a) and CR (b) onto Sil-(MA₁₀-co-SBMA₁₀).

Effect of Adsorbent Dosages

Using varying dosages of adsorbent of 2–6 mg per 30 mL and an initial concentration of dye of 150 ppm, the effect of the concentration of the adsorbent on Q_e was determined at pH 7 and room temperature. It is easy to see from Figure a that MB dye removal increased from 66.27% to 86.94% and CR dye elimination rose from 34.46% to 49.68% when the dosage was improved from 0.002 to 0.006 g, as shown in Figure b. A larger region of accessible surfaces of the adsorbent and more reactive sites of adsorption at higher dosages may be the cause of this phenomenon. Conversely, Q_e declined from 1491.14 to 652.06 mg/g for MB and from 775.42 to 372.61 mg/g for CR, respectively, as adsorbent dosages increased.

3.

Effect of adsorbent dosage on the adsorption of MB (a) and CR (b) onto Sil-(MA₁₀-co-SBMA₁₀); effect of initial concentration on the adsorption of MB (c) and CR (d) onto Sil-(MA₁₀-co-SBMA₁₀); effect of contact time on the adsorption of MB (e) and CR (f) onto Sil-(MA₁₀-co-SBMA₁₀).

Adsorption sites aggregating or overlapping may result in a decline in the overall area of adsorption per unit mass of the adsorbent, which accounts for the observed pattern.

Effect of Initial Concentration of Dye

Figure c and d illustrates the capability of change in MB and CR adsorption on Sil-(MA₁₀-co-SBMA₁₀), respectively, for initial concentration. Making use of a 0.005 g Sil-(MA₁₀-co-SBMA₁₀) adsorbent, the effects of various initial concentrations of dye (50, 100, 150, 200, and 250 mg/L) on the adsorption of MB and CR on Sil-(MA₁₀-co-SBMA₁₀) were examined in a 30 mL solution at pH 7 and room temperature. According to the findings, the adsorption capacity was raised from 251.02 to 1115.77 mg/g, while the initial concentration of MB dye (C_o) was improved from 50 to 250 mg/L, and from 166.61 to 666.1 mg/g when the first concentration of CR dye (C_o) was raised from 50 to 250 mg/g. The reason behind this is that the gradient of concentration created by the larger initial concentration of dye strengthened the driving force, thereby increasing the diffusion of dye moleculesboth MB and CR tend to diffuse toward the adsorbent. Although the adsorption capacity (Q_e) increased significantly with increasing initial dye concentration, the percentage removal did not change proportionally, particularly at lower concentrations. This is because, at low initial concentrations, the number of available adsorption sites on the adsorbent surface is relatively high compared to the number of dye molecules present. As a result, a large fraction of dye molecules is removed even at low concentrations, leading to high removal efficiency. However, further increases in initial concentration primarily enhance the amount of dye adsorbed per unit mass (Q_e) due to a higher concentration gradient and driving force, rather than significantly increasing the percentage removal. This behavior is commonly observed in adsorption systems where adsorption sites are not saturated at low concentrations but become progressively occupied as the initial dye concentration increases.

Effect of Contact Time

Figure e and f shows the change in MB and CR adsorption capacity on Sil-(MA₁₀-co-SBMA₁₀) with respect to contact time. The effects of contact time on the adsorption of MB and CR on Sil-(MA₁₀-co-SBMA₁₀) were examined at initial dye concentrations of 150 mg/L at pH 7 and room temperature in a 30 mL solution containing 0.005 g of Sil-(MA₁₀-co-SBMA₁₀) adsorbent. The effect of contact duration on the MB adsorption on Sil-(MA₁₀-co-SBMA₁₀) was examined. The dye was rapidly adsorbed during the first 10 min, then gradually decreased and reached equilibrium after 30 min, at which point very little dye remained adsorbed. When CR was adsorbed by Sil-(MA₁₀-co-SBMA₁₀), it was found that the dye was quickly adsorbed in the first 30 min and achieved equilibrium after 40 min. The dye’s initial rapid adsorption may have been caused by the increasing adsorption and the large number of dye molecules on the Sil-(MA₁₀-co-SBMA₁₀) surface.

Adsorption Isotherms

Different initial concentrations of dye (50–250 mg/L) were used for the MB and CR adsorption isotherm investigations onto Sil-(MA₁₀-co-SBMA₁₀). Eqs 3 and were utilized to relate the observed data of adsorption at equilibrium to the Freundlich and Langmuir models, respectively. The Langmuir equation is displayed here in its linear version:

$$\frac{{\mathrm{C}}_{\mathrm{e}}}{{\mathrm{Q}}_{\mathrm{e}}}=\frac{{\mathrm{C}}_{\mathrm{e}}}{{\mathrm{Q}}_{\max }}+\frac{1}{{\mathrm{Q}}_{\max }{\mathrm{K}}_{\mathrm{L}}}$$ 3

where K_L = Langmuir isotherm constant, the adsorption-to-desorption rate ratio, and Q_max = maximal specific adsorption of MB or CR at equilibrium connected to site saturation (coverage capacity at monolayer) (milligram per gram). The Langmuir model of MB and CR can be examined using the C_e/Q_e vs C_e plot displayed in Figure a and b, respectively. At room temperature, C_e/Q_e and C_e exhibit a linear connection in both figures, and the slope and intercept, respectively, were used to determine the values of Q_max and K_L. The dimensionless variable known as adsorption intensity, or R_L, determines the Langmuir isotherm’s favorable character. The following formula (eq ) depicts the R_L equation:

$${\mathrm{R}}_{\mathrm{L}}=\frac{1}{1+{\mathrm{K}}_{\mathrm{L}}{\mathrm{C}}_{\mathrm{o}}}$$ 4

here C_o represents the initial concentration of dye used in the adsorption experiments.

4.

Langmuir adsorption isotherm for the adsorption of MB (a) and CR (b) onto Sil-(MA₁₀-co-SBMA₁₀). Freundlich adsorption isotherm for the adsorption of MB (c) and CR (d) onto Sil-(MA₁₀-co-SBMA₁₀).

Eq uses the Freundlich isotherm model to calculate the adsorbent’s adsorption strength toward the dye molecules:

$$\log {\mathrm{Q}}_{\mathrm{e}}=\log {\mathrm{K}}_{\mathrm{F}}+\log {\mathrm{C}}_{\mathrm{e}}$$ 5

here the heterogeneity factor is described by 1/n which is linked with the adsorption’s intensity and capacity, C_e is the concentration at equilibrium (milligrams per liter), and K_F indicates the Freundlich constant. Charting log C_e vs log Q_e allows one to verify the Freundlich model, as seen in Figure c and d. K_F and 1/n were achieved from the slope and intercept, respectively. Table demonstrates both of the isotherm models’ adsorption parameter values. According to the Langmuir model, adsorption occurs at specific homogeneous locations on the adsorbent and is effective in monolayer adsorption methods. The highest monolayer adsorption (Q_max) onto Sil-(MA₁₀-co-SBMA₁₀) is predicted by the Langmuir model to be 2000.00 mg/g for MB and 1666.67 mg/g for CR. At an initial concentration of 250 mg/L, equilibrium dye adsorption experiments represented that the highest adsorption capacity for MB was 1116 mg/g and for CR it was 666 mg/g. R_L presents a qualitative assessment of the adsorption process’s favorable conditions in the Langmuir model. There are four types of adsorption: unfavorable (R_L > 1), linear (R_L = 1), favorable (0< R_L < 1), and irreversible (R_L = 0). The R_L values for the 50, 100, 150, 200, and 250 mg/L solutions of MB dye in our experiment are 0.5192, 0.3506, 0.2647, 0.2126, and 0.1776, respectively. Additionally, in our experiment, the R_L values for solutions of CR dye containing 50, 100, 150, 200, and 250 mg/L were, respectively, 0.8082, 0.6781, 0.5841, 0.513, and 0.4573. These suggest a very beneficial process of adsorption.

2. Adsorption Isotherm Parameters for the Adsorption of MB and CR on Sil-(MA₁₀ -co-SBMA₁₀).

	Langmuir Isotherm				Freundlich Isotherm
	Qmax (mg/g)	KL	R2	RL	n	KF (mg/g)	R2
Methylene Blue (MB)	2000.00	0.0185	0.9785	0.1776	1.3824	59.7723	0.9784
Congo Red (CR)	1666.67	0.0047	0.9305	0.4573	1.3389	16.4665	0.9984

The Freundlich model states that as n declines, adsorption gets harder. Depending on the order of the isotherm, 1/n can be labeled as unfavorable (1/n > 1), favorable (0 < 1/n < 1), or irreversible (1/n = 0). The value of 1/n at room temperature determined by our experiment was 1.3824 for MB and 1.3389 for CR. This suggests that the adsorption intensity parameter does not follow classical favorable adsorption behavior strictly and may reflect complex adsorption phenomena involving electrostatic interactions, surface heterogeneity, and possible cooperative adsorption effects at higher surface coverage. This nonideal behavior is attributed to the zwitterionic nature of the adsorbent, which provides both positively and negatively charged sites, enabling simultaneous interactions with different dye species.

The adsorption capacity may be roughly inferred from the anticipated K_F number; the greater the K_F value, the larger the adsorption capacity. The Langmuir model produced a higher regression coefficient than the Freundlich model and an adsorption capacity that was nearer to the measurement obtained from experiments. The Langmuir isotherm identifies the surface coverage of the adsorbent by using the monolayer generation of the adsorbate. In this study, we assume that the zwitterionic ion Sil-(MA₁₀-co-SBMA₁₀) supports the concept of the formation of a monolayer, the positive ion MB and the negative ion CR undergo an adsorption method driven by the attraction of electrostatic forces. Of course, other interactions also took part in this kind of ultrahigh adsorption capacity of Sil-(MA₁₀-co-SBMA₁₀) as mentioned earlier. , Conversely, the Freundlich model assumes that the adsorbate forms many layers and illustrates their link of attraction. The adsorbate’s molecules (Congo Red and Methylene Blue) were not attracted to one another in this case. Consequently, only a portion of the adsorbed molecules can be explained by the Freundlich model, which is inconsistent with the actual results (Table ).

For methylene blue (MB), both Langmuir and Freundlich models exhibited comparable correlation coefficients (R² ≈ 0.978), indicating that the adsorption process may involve both monolayer adsorption and surface heterogeneity. However, the Langmuir model provided a slightly better agreement between experimental and calculated adsorption capacities, suggesting that monolayer adsorption on relatively homogeneous sites is the dominant mechanism. For Congo red (CR), the Freundlich model showed a significantly higher correlation coefficient (R² = 0.9984) compared to the Langmuir model (R² = 0.9305), indicating that adsorption occurs on a heterogeneous surface with possible multilayer formation.

Overall, the adsorption process is governed by a combination of electrostatic interactions, carbonyl−π interactions, and surface heterogeneity, rather than a single ideal adsorption model. In this study, linearized forms of the Langmuir and Freundlich models were used for parameter estimation and comparison with literature data. However, it is acknowledged that nonlinear fitting methods may provide improved accuracy by minimizing errors associated with linear transformation. Future work will focus on applying nonlinear regression models for more precise evaluation of adsorption parameters.

It should be noted that slight deviations in correlation coefficients (R²) may arise because of experimental variability and the inherent complexity of the adsorption process on a heterogeneous surface. The zwitterionic nature of the adsorbent introduces multiple interaction mechanisms that cannot be fully captured by a single ideal isotherm model. Despite these variations, all experiments were performed in replicate and consistent adsorption trends were observed, confirming the reproducibility of the data.

Adsorption Kinetics

Adsorption kinetics explain how quickly an adsorbate is taken up, defining the time required for adsorption to occur at the interface of a solid and liquid. The widely used pseudo-first-order model developed by Lagergren can be expressed in its linearized integral form (eq ).

$$\ln ({\mathrm{Q}}_{\mathrm{e}}-{\mathrm{Q}}_t)=\ln {\mathrm{Q}}_{\mathrm{e}}-{\mathrm{k}}_1\mathrm{t}$$ 6

The pseudo-second-order model, given in eqs and , can also be represented in a linearized form as outlined below.

$$\frac{\mathrm{t}}{{\mathrm{Q}}_{\mathrm{t}}}=\frac{1}{{\mathrm{k}}_2{\mathrm{Q}}_{\mathrm{e}}^2}+\frac{\mathrm{t}}{{\mathrm{Q}}_{\mathrm{e}}}$$ 7

$$\mathrm{h}={\mathrm{k}}_2{{\mathrm{q}}_{\mathrm{e}}}^2$$ 8

Here, Q_e and Q_t represent the adsorption capacities (milligrams per gram) at equilibrium and at a given time t (minutes), respectively. The initial adsorption rate is denoted as h (mg g^–1 min^–1), while the rate constants for pseudo-first-order and pseudo-second-order adsorption are expressed as k₁ (min^–1) and k₂ (g mg^–1 min^–1), respectively. For pseudo-first-order kinetics, a plot of ln­(Q_e – Q_t) versus t should yield a straight line, from which the slope and intercept provide the values of k₁ and Q_e. In the case of pseudo-second-order kinetics, linearity is verified by plotting t/q_t against t, with the slope and intercept used to measure Q_e and k₂. Beyond the determination coefficient (R²), the applicability of these models was further assessed using an error analysis function, specifically the Chi-square (χ²) test (eq ).

$${\mathrm{X}}_2=\mathrm{\sum }\frac{{({\mathrm{Q}}_{\mathrm{e}}\exp -{\mathrm{Q}}_{\mathrm{e}}\mathrm{model})}^2}{{\mathrm{Q}}_{\mathrm{e}}\mathrm{model}}$$ 9

Q_eexp (milligram per gram) and Q_emodel (milligram per gram) are the experimental and model values of the adsorption capacity, respectively. Figure a and b, respectively, displays the linearized forms of the first- and second-order pseudoadsorption processes of MB on Sil-(MA₁₀-co-SBMA₁₀) fitted to the experimental data.

5.

Pseudo-first-order (a) and pseudo-second-order (b) kinetic plots for the adsorption of MB onto Sil-(MA₁₀-co-SBMA₁₀). Pseudo-first-order (c) and pseudo-second-order (d) kinetic plots for the adsorption of CR onto Sil-(MA₁₀-co-SBMA₁₀).

The linearized form of CR’s adsorption to the Sil-(MA₁₀-co-SBMA₁₀) model fitted to the experimental data in pseudo-first and pseudo-second order, respectively, is displayed in Figure c and d. Table summarizes the correlation coefficients (R₁ ²; R₂ ²), chi-square values (χ₁ ²; χ₂ ²), and kinetic parameters obtained by using the two previously discussed kinetic models for MB and CR. According to this table, χ₂ ² (1.5031) was an extremely low number according to the pseudo-second-order model for MB. Consequently, the adsorption kinetics of MB on Sil-(MA₁₀-co-SBMA₁₀) are discussed by the pseudo-second-order model, confirming that adsorption is a chemical process. The pseudo-first-order and pseudo-second-order models for CR, however, indicated that χ₁ ² (10.9845) and χ₂ ² (12.8) were both carrying low values.

3. Equation Parameters of Pseudo 1st Order and Pseudo 2nd Order Kinetic Models for the Adsorption of MB and CR on Sil-(MA₁₀ -co-SBMA₁₀).

Dye	Model	Parameter	Initial Dye Concentration (150 ppm)
Methylene Blue (MB)	Pseudo 1st Order	K1	0.0811
		Qe1	177.7539
		R1 2	0.9443
		χ1 2	1748.3566
	Pseudo 2nd Order	K2	0.0009
		Qe2	769.2308
		R2 2	0.9999
		χ2 2	1.5031
	Intraparticle Diffusion Model	Kipd	14.9160
		C	633.8500
		R2	0.7571
Congo Red (CR)	Pseudo 1st Order	K1	0.1036
		Qe1	493.6368
		R1 2	0.9606
		χ1 2	10.9845
	Pseudo 2nd Order	K2	0.0002
		Qe2	500.0000
		R2 2	0.9922
		χ2 2	12.8000
	Intraparticle Diffusion Model	Kipd	51.0280
		C	86.4280
		R2	0.9543

The overall pace is determined by the slowest stage of the multistep adsorption procedure. After being carried from the liquid phase to the solid adsorbent’s surface, adsorbate molecules diffuse into the internal pores. The intraparticle diffusion model was built utilizing the experimental data on adsorption kinetics to identify the rate-limiting step. The intraparticle diffusion model that Weber and Morris suggested has a linear mathematical form, represented by eq :

$${\mathrm{Q}}_{\mathrm{t}}={\mathrm{K}}_{\mathrm{i}\mathrm{p}\mathrm{d}}{\mathrm{t}}^{1/2}+\mathrm{C}$$ 10

where Q_t (milligrams per gram) shows the adsorption capacity at time t, constant C is proportional to the boundary layer thickness, and K_ipd (mg g^–1 min^–0.5) determines the rate constant for intraparticle diffusion.

The intraparticle diffusion model plots for MB and CR onto Sil-(MA₁₀-co-SBMA₁₀) are shown in Figure S8a and S8b, respectively, and the formula parameters are presented in Table .

With excellent correlation coefficients (R² = 0.9999 for MB and 0.9922 for CR), the pseudo-second-order model provided the most precise explanation for the MB and CR adsorption kinetics, although this does not exclusively confirm chemisorption. However, the slightly lower pseudo-first-order R² values (0.9443 for MB, 0.9606 for CR) suggest that physical adsorption may also occur, particularly in the initial phase. To explore this, an intraparticle diffusion model was applied. The results (R² = 0.9543 for CR and 0.7571 for MB) indicate that intraparticle diffusion contributes more significantly to CR.

It was observed that the newly developed zwitterionic copolymer-grafted silica material Sil-(MA₁₀-co-SBMA₁₀) was excellent for the adsorption of both anionic (CR) and cationic (MB) dyes with large adsorption capacities, which is challenging for an adsorbent. Moreover, the MB dye may be adsorbed at a neutral pH. The highest rate of adsorption capacities for MB and CR were extremely high (2000.00 and 1666.67 mg/g, respectively) on Sil-(MA₁₀-co-SBMA₁₀) material compared to the other developed silica-based adsorbents (Table ), and most of the adsorbents could not adsorb both dyes including our recent work with isoniazid-modified silica which worked only for anionic dyes and had a highest adsorption capacity of 714.3 mg/g. Furthermore, similar zwitterionic polymer-based membranes, polymer-modified nanosilica, zwitterionic copolymer, and zwitterionic polymer hydrogel for both anionic and cationic dyes did not demonstrate greater adsorption capacities. −

4. Silica-Based adsorbents’ Adsorption Capacities for MB and CR Removal.

Dye	Adsorbents	Adsorption capacity, Qe (mg/g)	References
Methylene Blue (MB)	Poly(methacrylate)/silica	91.32
	Silica/iron oxide composite	96.5
	Cyclodextrin/silica	212.0
	Silica-acrylic acrylamide	375.9
	Graphene oxide/hollow mesoporous silica	476.19
	pMS cross-linked silica nanoparticles	588.23
	pAS cross-linked silica nanoparticles	666.65
	Calcium Alginate-Silica Dioxide	1158.33
	Activated carbon	381.88
	Metal–organic framework	39.50
	Polyelectrolyte-functionalized silica	15.13
	Sil-(MA10-co-SBMA10)	2000.00	This study
Congo Red (CR)	Aniline propyl silica xerogel	40.86
	Ni0.5Cu0.5Fe2O4/SiO2 Nanocomposites	63.49
	ZnO-modified SiO2	90.10
	Montmorillonite-Silica Nanocomposite	172.40
	Xanthan gum–graft-poly(acrylamide)/SiO2 Nanocomposite	202.43
	Graphene oxide-silica	333.33
	Polyethylenimine-functionalized lignin/SiO2	442.48
	Fe3O4@nSiO2@mSiO2 core–shell microspheres	1429.0
	Activated carbon	632.13
	Sil-(MA10-co-SBMA10)	1666.67	This study

Adsorption–Desorption Study

The reusability of the Sil-(MA₁₀-co-SBMA₁₀) adsorbent was evaluated for two dyes: MB and CR over six adsorption–desorption cycles. For MB, at pH 7, the initial concentration was 150 mg/L, with a 3-mg dosage of adsorbent in 30 mL, under 40 min of contact time and 180 rpm. After an hour of desorption with 0.1 M HCl, centrifugation and drying at 60 °C were performed. The conditions were the same for CR, except 0.1 M NaOH was used for desorption, and the pH was 3. The adsorbent’s performance, regeneration efficiency, and reusability were assessed for both dyes across all cycles. Figures S9 and S10 (Supporting Information) show that after six cycles, both MB and CR’s adsorption capacities declined; additionally the dye removal efficiency remained around 70% for MB and over 50% for CR. Despite this, the adsorption capacities for both dyes remained relatively high. The decline in efficiency is likely due to pore filling. Overall, the Sil-(MA₁₀-co-SBMA₁₀) adsorbent demonstrates excellent performance in dye adsorption, regeneration, and reusability.

As we mentioned earlier, to find out the diverse applications of the developed material Sil-(MA₁₀-co-SBMA₁₀), half of the material was packed into an HPLC column for the isolation of polar analytes in HILIC mode. Therefore, for instance, Tanaka test probes (Supporting Information, Table S2) were separated on the developed column and compared with the most widely utilized HILIC columns (ZIC-HILIC, iHILIC-Fusion, Luna-Diol, and Luna-NH₂) to clarify the effectiveness of Sil-(MA₁₀-co-SBMA₁₀). The Tanaka test protocol for HILIC column evaluation with the separation of different types of analytes (Supporting Information, Figure S11) is very famous for evaluating a recently developed column to determine its suitability for application.

Comparative separation of these Tanaka test probes (Figure a, b, and c) indicates the best separation abilities (hydrophobic, hydroxyl group, and configurational selectivity, respectively) of Sil-(MA₁₀-co-SBMA₁₀) compared to the best HILIC commercial column, ZIC-HILIC. The Supporting Information (Table S3) provides details for a number of parameters examined using the Tanaka test probes, demonstrating Sil-(MA₁₀-co-SBMA₁₀)’s comparable performance.

6.

Separation of Tanaka test probes (a, b, and c) on Sil-(MA₁₀-co-SBMA₁₀) and reference columns. Mobile phase: ACN-ammonium acetate buffer (20 mM in the aqueous portion, pH 4.7) (90:10, v/v); column temperature: 30 °C; flow rate: 0.2 mL/min; UV detection: 254 nm; injection volume: 5 μL.

These results were also reflected in distinguishing six nucleosides and nucleobases with Sil-(MA₁₀-co-SBMA₁₀) and reference columns (Figure ).

7.

Separation of six nucleosides and nucleobases on Sil-(MA₁₀-co-SBMA₁₀) and reference columns. Mobile phase: ACN-ammonium acetate buffer (20 mM in the aqueous portion, pH 4.7) (90:10, v/v); column temperature: 30 °C; flow rate: 0.2 mL/min; UV detection: 254 nm; injection volume: 5 μL.

The highest retention was observed on the newly developed zwitterionic copolymer-grafted silica material. Therefore, the material Sil-(MA₁₀-co-SBMA₁₀) is very suitable for application as an adsorbent as well as a packing material for liquid chromatographic separation with high performance. The material Sil-(MA₁₀-co-SBMA₁₀) as an adsorbent and packing material showed excellent performance. The reasons behind Sil-(MA10-co-SBMA10) promising separating capabilities in HILIC and its ultrahigh adsorption capacity must be explained in light of the novel material’s adsorption process. In the zwitterionic copolymer, both monomers have carbonyl groups. Again, the SBMA monomer has both cationic and anionic sites. As shown in Figure , both MB and CR dyes can interact with the adsorbent material. In addition to ionic interaction, carbonyl groups from the adsorbent and π-electrons from dyes probably play an effective role. It has been reported that interaction in a complex model of HCHO-benzene is much greater (1.87 kcal mol^–1) than a CH₄-benzene (0.53 kcal mol^–1) interaction and a plane-to-plane interaction linking two benzenes (0.49 kcal mol^–1). Due to the polymeric material with many interaction sites, multiple interactions are possible. The above-mentioned two interactions (electrostatic and carbonyl−π) were intended to serve as the factors that propel the novel material’s extraordinarily high adsorption capabilities (Figure a). Moreover, from the zeta-potential analysis (Figure b), electrostatic interaction was obvious at higher pH for MB and at lower pH for CR. Zeta potential tests were performed between pH 2 and pH 13 in order to assess the Sil-(MA₁₀-co-SBMA₁₀) surface charge behavior. As described in Figure b, the zeta potential became progressively more negative with increasing pH, shifting from −2.5 mV at pH 2 to −34.5 mV at pH 13. This trend reflects the deprotonation of surface functional groups. The adsorption capacity of MB (cationic dye) increased with pH because of a stronger electrostatic pull to the negatively charged surface. However, when the pH became low, CR (anionic dye) exhibited the highest adsorption, which drastically decreased as the surface grew more negatively charged due to increased electrostatic repulsion. These results confirm that electrostatic interactions have dominance over the adsorption mechanism, consistent with the zwitterionic nature of the adsorbent and the charge of the dyes (Figure a).

8.

(a) Possible interaction mechanism for the ultrahigh adsorption capacities of Sil-(MA₁₀-co-SBMA₁₀) for both MB and CR dyes; (b) variation of zeta potential and MB and CR’s pH-dependent adsorption capacities for the identical adsorbent.

To further confirm the adsorption of dye molecules onto the adsorbent, we recorded ATR-FTIR spectra of the materials before and after dye adsorption. The observed shifts in the characteristic absorption bands indicate the successful adsorption of the dye molecules onto the adsorbent surface (Figure ). In addition, energy-dispersive X-ray spectroscopy (EDX) was employed to further verify the adsorption of dyes onto the surface-modified silica adsorbent. Figure presents the EDX spectra of the Sil-(MA10-co-SBMA10) adsorbent after the adsorption of Congo Red (CR), providing additional evidence of the successful adsorption of dye molecules.

9.

ATR-IR spectra of Sil-(MA₁₀-co-SBMA₁₀) after adsorption of MB and CR dyes.

10.

EDX spectrum and corresponding elemental mapping images of the Sil-(MA10-co-SBMA10) adsorbent after adsorption of CR, showing the distribution of C, O, N, Si, S, and Na.

On the other hand, the interaction mechanism in HILIC mode separation is quite complex. However, the main mechanism was proposed by Alpert who established the HILIC method. The method involves the improvement of a stationary layer of water on the packing material’s outer layer due to the hydrophilicity of the packing material in HILIC. The polar solutes are then divided between the layer of water and the highly organic mobile phase. Moreover, there are many other reports on interactions between the packing material and the analytes, like ion exchange, adsorption, hydrogen bonding, carbonyl−π, π–π, and n−π interactions based on the structure of the stagnant phase and analytes. Sil-(MA₁₀-co-SBMA₁₀) demonstrated a greater selectivity for the copolymeric material with several interaction sites in comparison to the widely used reference HILIC columns. Therefore, the developed material Sil-(MA₁₀-co-SBMA₁₀) has very effective applications as a material for adsorption for removing both anionic and cationic dyes and for separating polar analytes in HILIC mode separation.

Conclusions

In this work, we describe a zwitterionic copolymer-modified mesoporous silica material. After characterization, both cationic (MB) and anionic (CR) dyes were removed by using the novel material as an adsorbent. It was discovered that the highest adsorption capabilities were 2000.00 mg/g and 1666.67 mg/g, respectively, which are far more than those of the silica-based adsorbents developed to date. Additionally, a stainless steel column containing a similar substance was utilized as a stationary phase for the polar analytical substances in the HILIC method isolation. When it came to the isolation of Tanaka test probes in HILIC, the Sil-(MA₁₀-co-SBMA₁₀) performed better than the commercial columns. Furthermore, in comparison to the reference columns, greater retention was noted for the separation of the six nucleosides and nucleobases. This work is a step forward for material design and improvement and its various uses, and additionally, it will lower the expense of designing and producing new materials for a range of uses. As a result, it will create opportunities for the use of a single high-performing material to be diversified.

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acsomega.6c08360.

• The characterization scheme for the Tanaka test; physical properties of the prepared and reference HILIC columns; ¹H NMR spectrum of the telomer (MA10-co-SBMA10); calibration curves for methylene blue and Congo red; SEM, ATR-FTIR, and TGA characterization of the synthesized material; adsorption kinetics and reusability studies; zeta potential analysis; chemical structures of the Tanaka test analytes; and comparative chromatographic parameters of the prepared stationary phase and commercial reference columns; additional experimental details, materials, and methods, including photographs of samples (PDF)

The authors declare no competing financial interest.