- Email: [email protected]
Graphene oxide reinforced ionic liquid-functionalized adsorbent for solid-phase extraction of phenolic acids
Journal of Chromatography B 1072 (2018) 123–129
Contents lists available at ScienceDirect
Journal of Chromatography B journal homepage: www.elsevier.com/locate/jchromb
Graphene oxide reinforced ionic liquid-functionalized adsorbent for solid-phase extraction of phenolic acids ⁎
Xiudan Houa,c, Xiaofeng Lua, Sheng Tangb, Licheng Wanga, , Yong Guoa,
MARK
⁎
a CAS Key Laboratory of Chemistry of Northwestern Plant Resources and Key Laboratory for Natural Medicine of Gansu Province, Lanzhou Institute of Chemical Physics, Chinese Academy of Sciences, Lanzhou 730000, China b School of Chemistry and Environmental Engineering, Wuhan Institute of Technology, Wuhan 430205, China c University of Chinese Academy of Sciences, Chinese Academy of Sciences, Beijing 100049, China
A R T I C L E I N F O
A B S T R A C T
Keywords: Adsorption energy Graphene oxide Ionic liquid Phenolic acid Solid-phase extraction
An environmental friendly sorbent of polymeric ionic liquids modified graphene oxide-grafted silica (PILs@GO@ Sil) was synthesized for solid-phase extraction (SPE) of phenolic acids. The sorbent was prepared via a chemical layer-to-layer fabrication including amidation reaction, surface radical chain-transfer polymerization and in situ anion exchange. After modification with PILs, the silica surface had higher positive potential so that it would exhibit stronger electrostatic interaction for acidic compounds compared with GO@Sil. The adsorption performance of phenolic acids was investigated through the theoretical calculation and static, kinetic state adsorption experiments. Under the optimized conditions, wide linear ranges were obtained with correlation coefficients ranging from 0.9912 to 0.9998, and limits of detection were in the range of 0.20–0.50 μg L−1. Compared with other reported methods, the proposed PILs@GO@Sil-SPE-HPLC showed higher extraction efficiency. Finally, the black wolfberry yogurt and urine were analyzed as real samples and good recoveries spiked with standard solution were obtained.
1. Introduction Phenolic acids are a class of secondary metabolites, which have drawn increasing attention due to their antioxidant properties and marked effects in the prevention of various oxidative stress associated diseases such as cardiovascular disease and cancer [1,2]. They widely spread throughout the plant kingdom including vegetables and fruits, and further exist in beverages as the food ingredient. Once ingested, most phenolic acids are extensively metabolized by the enzymes and partly adsorbed or excreted [3]. It is essential to enrich and purify before the instrumental analysis due to the low concentration of analytes and the interference of other complex substance in real samples. Therefore, sample pretreatment is one of the most important procedures in the whole analysis, especially for the analysis of biological and environmental samples with complex matrix [4], such as solid-phase extraction (SPE), solid-phase microextraction, liquid-phase microextraction, electromembrane extraction, microextraction by packed sorbents and so on. This stage would realize the pre-concentration of target compounds and elimination of most matrix interferences before introduction into the analytical instrument [5]. SPE is the most widely used separation and pre-concentration technique owing to its high chromatographic utility [6,7]. New various materials as adsorbents in ⁎
SPE have been exploited to extract organic and inorganic compounds from the complex media with higher adsorption capacity [8–10]. The studies indicated that nanosized SPE sorbents have high extraction capacities with rapid extraction kinetic performance [11]. Nanostructured materials, such as carbon-based nanomaterials, metal and metal oxide nanoparticles, electrospun nanofibers, and metal-organic frameworks, have gained great attention because of their physical and chemical properties including large surface area, desirable chemical and thermal stability, and favorable adsorption performance [4,12,13]. Graphene and its precursor graphene oxide (GO), as 2-dimensional nanoscale materials, have sparked intense research interest in sample preparation because of the outstanding properties. GO contains miscellaneous chemical functional groups on the basal planes and at the edges of GO sheets, such as hydroxyl, epoxy and carboxyl, which are expected to promote interfacial interactions between GO and adsorbate [14,15]. In the previous work, based on the excellent adsorption performance, GO-grafted silica was fabricated and used to determine trace phenolic acids in urine [16]. Besides, GO could be modified conveniently through the chemical fabrication procedure to improve the extraction yield and selectivity for some specific analytes [17,18]. Ionic liquids (ILs) are a class of organic salts composed entirely of organic cations (e.g. pyrrolidinium, pyridinium, imidazolium,
Corresponding authors. E-mail addresses: [email protected] (L. Wang), [email protected] (Y. Guo).
https://doi.org/10.1016/j.jchromb.2017.11.013 Received 29 August 2017; Received in revised form 8 November 2017; Accepted 10 November 2017 Available online 12 November 2017 1570-0232/ © 2017 Published by Elsevier B.V.
Journal of Chromatography B 1072 (2018) 123–129
X. Hou et al.
Fig. 1. Scheme of the synthetic route for poly(VHIm+PF6−)@GO@Sil (a) and the interactions between the extraction material and analytes (b).
water (18.2 MΩ), which was collected from a Molecular water purification system.
tetraalkyl ammonium or tetraalkyl phosphonium) and inorganic or organic anions (e.g. Br−, BF4−, CF3SO3−, NTf2, PF6−). Due to the unique physicochemical properties, such as low vapor pressure, good chemical and thermal stability, wide viscosities range and miscibility with water and organic solvents, ILs have been used as promising solvents and materials in many fields [19–21]. Imidazolium IL, a member of ILs family, is widely used for sample preparation on the basis of the pentacyclic structure and the easily tuned property. In the previous reports, polymeric ionic liquids (PILs) modified GO-grafted silica exhibited high extraction efficiency for flavonoids because of its capacity of most types of interactions with analytes (e.g., π-π, n-π, hydrogen bonding, dispersive, dipolar, ionic/charge-charge) [22,23]. The satisfactory results indicated that the presence of GO would increase the overall extraction recoveries of analytes and bring an enhancement in analyte transport. In this paper, the properties of an imidazolium-based PIL were tuned based on the reconstruction of its counter anion via in situ anion exchange. Firstly, poly(1-vinyl-3-hexylimidazolium hexafluorophosphate)-GO-grafted silica (poly(VHIm+PF6−)@GO@Sil) was synthesized and characterized. Next, the adsorption performance of phenolic acids on the proposed adsorbent was investigated through the theory calculation and a series of adsorption experiments. Comparisons of extraction efficiency with other prepared materials were also performed. Under the optimum conditions, the proposed poly (VHIm+PF6−)@GO@Sil-SPE-HPLC method was used to determine the selected phenolic acids in yogurt and urine.
2.2. Instrumentation and chromatographic conditions The SPE procedure was performed on a HGC-8 numerical control solid phase extraction system (Hegong Scientific Instrument Co., Shanghai, China). HPLC analysis were performed on an Agilent 1100 Series modular HPLC system (Agilent Technologies, USA) with a highpressure quaternary pump, a 20 μL sample loop and a UV–vis detector. Separation of the analytes was performed on a C18 column (Hypersil ODS2, 250 mm length × 4.6 mm i.d., 5 μm). The mobile phase was methanol and water with 0.25% acetic acid. The gradient elution condition was 0–23 min, 15–40% methanol; 23–30 min, 40–70% methanol. Flow rate was set at 0.8 mL min−1 while the detection wavelength was 298 nm. BET surface area was detected by a Micromeritics ASAP 2020 device. Surface properties of the prepared silica were characterized by scanning electron microscope (SEM, JSM-6701F, Japan). Energy dispersive spectrometry (EDS) was obtained on a low vacuum scanning electron microscope-X-ray energy dispersive spectrometer (JSM5600LV, Japan). The zeta potential was determined with a Zetasizer Nano series ZS instrument (Malvern Instruments, United Kingdom).
2.3. Preparation of PILs/GO-modified silica 2. Experimental
Hummers method was used to synthesize GO from the natural graphite [24]. Fig. 1 shows the preparation procedure of the PILs/GO-modified silica. Firstly, 8 g of silica was immersed in hydrochloric acid to be activated for 24 h and then rinsed with distilled water. The neutral silica and 8 mL of APTES were added in 100 mL of dry toluene to react and reflux at 120 °C. After 24 h, the obtained aminopropyl-modified silica (Sil-NH2), 10 mL of GO solution (0.1%, w/v), 100 mL of phosphate buffer solution (pH, 7.4), 0.05 g of EDC and 0.05 g of NHS were added into a 250 mL reaction flask. GO-grafted silica (GO@Sil) was obtained through the above amidation reaction. Mercaptopropyl-modified GO-grafted silica (GO@Sil-SH) was prepared being similar to that of Sil-NH2, in which APTES was replaced by MPTES. The fabrication of poly(VHIm+PF6−)@GO@Sil involved the following two processes (surface radical chain-transfer polymerization and in situ anion-exchange): (a) 8 g of VHIm+Br− and 0.4 g of AIBN were dissolved in 100 mL of DMSO and added into the conical flask with the prepared GO@Sil-SH, which was refluxed for 24 h at 60 °C under an N2 atmosphere to obtain poly(VHIm+Br−)@GO@Sil particles; (b) after cooling, rinsed and dried, the poly(VHIm+Br−)@GO@Sil particles and 8 g of NH4PF6 were added into a 250 mL three-neck round-bottomed flask with 150 mL of aqueous solution for reacting at 60 °C for 12 h. Finally, the obtained poly(VHIm+PF6−)@GO@Sil was dried, stored, characterized and further used as the extraction material.
2.1. Chemicals and materials Protocatechuic acid (ProA), vanillic acid (VanA), N-hydroxy succinimide (NHS) and N-(3-dimethylaminopropyl)-N-ethyl-carbodiimide (EDC) were obtained from Aladdin Chemical Reagent Co. (Shanghai, China). Syringic acid (SyrA) and salicylic acid were purchased from Energy Chemical (Shanghai, China). The chemical structures of the selected phenolic acids are shown in Fig. S1. VHIm+Br− was purchased from Shanghai Chengjie Chemical Co. (Shanghai, China). Ammonium hexafluorophosphate (NH4PF6) was purchased from Shanghai Tiancheng Chemical Co. (Shanghai, China). Azobisisobutyronitrile (AIBN) was obtained from Shanpu Chemical Co. (Shanghai, China). Dimethyl sulfoxide (DMSO) was obtained from J&K scientific LTD. (Beijing, China). 3-Aminopropyltriethoxysilane (APTES) and 3-mercaptopropyltriethoxysilane (MPTES) were purchased from Chemical Industrial Corporation of Gaizhou (Guangzhou, China). SPE empty column (3 mL), polyethylene frits (5 mm pore size) and the commercial C18 SPE column were purchased from Shenzhen Biocomma Biotech Co. (Shenzhen, China). Stock solutions of phenolic acids were prepared in methanol with the concentration of 0.2 mg mL−1. Working solutions for extraction were freshly prepared by diluting the stock solutions with ultrapure 124
Journal of Chromatography B 1072 (2018) 123–129
X. Hou et al.
Fig. 2. SEM images of poly(VHIm+PF6−)@GO@Sil.
same conditions were performed and average values were provided as the final data. Recovery (%) of the analytes was calculated and evaluated by the following equation:
2.4. Preparation of sample solutions The black wolfberry yogurt was bought from a local supermarket (Lanzhou, China) and then centrifuged, filtered. The urine sample was collected from a healthy volunteer, which was first coarsely filtered through a Buchner funnel, and then through a 5 μm sieve plate. The pretreated samples were stored at 4 °C for further use.
Recovery =
where Csol and Csam are the enrichment concentration of analytes in eluent and the initial concentration of analytes in the working solution, respectively; Vsol and Vsam are the volumes of eluent and working solution. Csol is obtained from the detected peak area and the calibration curve, which is plotted by a series of concentrations of the standard solution and peak area determined by HPLC.
2.5. Adsorption experiments 2.5.1. Theoretical adsorption energy The calculation for adsorption energies of analytes on poly (VHIm+Br−) and poly(VHIm+PF6−) were performed with Gaussian 09 software, which was based on periodic density functional theory with M06-2X function [25].
3. Results and discussions 3.1. Characterization of the prepared materials
2.5.2. Investigation of the adsorption mechanism In static-state binding experiments, 20 mg of the poly(VHIm+PF6−) @GO@Sil particles was placed in a 10 mL glass flask with 1 mL of standard solutions at the concentrations of 5–130 μg mL−1. After shaking for 1 h, the mixture was centrifuged to separate the sorbent and solution; the upper solution was injected into HPLC to analyze the retained phenolic acids. In dynamic-state binding experiments, the poly (VHIm+PF6−)@GO@Sil particles (20 mg) was mixed with 1 mL of the phenolic acids solution (100 μg mL−1), and shaken at a regular time intervals (5, 10, 20, 30, 40, 50, 60 min). After centrifugation, the supernatants were determined to study the kinetic-adsorption process of poly(VHIm+PF6−)@GO@Sil. The adsorption capacities (q, μg g−1) were calculated by the following equation:
q=
CsolVsol × 100% CsamVsam
Surface morphology characteristic of the prepared poly (VHIm+PF6−)@GO@Sil was characterized by SEM (shown in Fig. 2). The uniform, rough surface with dense protuberances was observed, which greatly increased the surface area to volume ratio. The successful synthesis of poly(VHIm+PF6−)@GO@Sil was verified by the EDS analysis. As shown in Table S1, bromine element appeared on the surface of poly(VHIm+Br−)@GO@Sil, and fluorine element was occurred on the surface of poly(VHIm+PF6−)@GO@Sil, which sufficiently proved the successful preparation of two sorbents. To investigate the surface charge of the as-prepared extraction materials, the zeta potential was separately measured in different pH solutions (Fig. 3). After modified with PILs, it was obviously found that the sorbent had a bigger positive potential on account of the alteration of the surface functional groups. Therefore, poly(VHIm+PF6−)@GO@Sil was more suitable for the extraction of acidic compounds. The BET
(Csam-Cf )×Vsam W
where Csam (μg mL−1) and Cf (μg mL−1) are the initial and final concentrations of the sample solution, respectively; Vsam (mL) is the sample volume; W (g) is the weight of poly(VHIm+PF6−)@GO@Sil particles. 2.6. Solid-phase extraction procedure and recovery calculation For all SPE experiments, 30 mg of the prepared silica particles were packed into a 3 mL empty propylene cartridge with two sieve plates positioned at each end of the column. After that, the SPE packing column was activated by loading with 2 mL of eluent (methanol/water, 2:1, pH = 2) and 10 mL of methanol/water (v/v, 1:1). 40 mL of the working solution or sample solution was passed through the cartridge at a constant flow rate of 1.5 mL min−1 during the whole SPE procedure adjusted with the vacuum pump. After loading, the sorbent was washed with 1 mL of ultrapure water and dried in air for 5 min. The retained analytes on the cartridge were eluted with 1 mL of methanol/water (2:1, pH = 2) at a flow rate of 0.5 mL min−1. Then the desorption solution was used for HPLC-UV analysis. Triplicate analyses under the
Fig. 3. Zeta potential of poly(VHIm+PF6−)@GO@Sil and GO@Sil.
125
Journal of Chromatography B 1072 (2018) 123–129
X. Hou et al.
hydrogen bond and the stereochemcial structure resulted in the low adsorption amount of SalA on the sorbent.
Table 1 Molecular weight, octanol-water partition coefficient (log Kow) of phenolic acids and adsorption energy, adsorption amounts of analytes on the sorbent. Molecular Weight
ProA SyrA VanA SalA a
154.12 198.18 168.15 138.12
log Kowa
0.91 1.04 1.22 2.26
Adsorption energy (kJ mol−1) Br−-PIL
PF6−-PIL
−91.92 −97.02 −97.58 −88.95
−127.91 −111.49 −108.75 −111.14
3.4. Optimization of experimental conditions
Adsorption amount (μg g−1)
The prepared poly(VHIm+PF6−)@GO@Sil-SPE extraction column was applied to the extraction of phenolic acids as target analytes in the yogurt and urine. In order to evaluate the extraction and desorption performance, the experiment parameters including extraction conditions (volume of sample loading, sample loading rate, pH of sample) and elution conditions (eluent, volume of eluent, elution rate) were investigated. All the experiments were carried out in triplicate and the average was taken into the discussion.
3487 1866 1640 681
Kow: Octanol-water partitioning coefficients.
surface area, total pore volume and average pore width of commerical C18, poly(VHIm+Br−)@GO@Sil and poly(VHIm+PF6−)@GO@Sil were listed in Table S2. The discrepancy of two PILs-based silica was derived from the anion exchange reaction.
3.4.1. Volume of sample loading When the concentration of sample solution was constant, an increase of volume of sample loading would improve the enrichment factor [26]. Adsorption capacity of the extraction material had a main effect on the volume of sample loading. With the increase of sample volume, part of analytes might be loss due to the saturation of the sorbent for a large sample volume. In this work, on the basis of 100 μg L−1 of working solution, the sample volume in the range of 20–70 mL was investigated. Fig. 5a showed that when the volume of sample loading was above 40 mL, the extraction recovery decreased distinctly. Therefore, 40 mL was selected as the optimized volume of sample loading.
3.2. Adsorption energies The theoretical calculation results of adsorption energy of the selected phenolic acids on poly(VHIm+Br−) and poly(VHIm+PF6−) were presented in Table 1. The adsorption energies of ProA, SyrA and SalA on poly(VHIm+PF6−) were higher than those on poly(VHIm+PF6−). Therefore, the poly(VHIm+PF6−)@GO@Sil was selected as the suitable materials in the following experiments. The different adsorption energies of analytes on one sorbent were resulted from the molecular stereochemical structures and various interaction forces, including hydrogen bonding, π-π and electrostatic interactions.
3.4.2. Sample pH Sample pH played a vital role in the extraction process since it not only had an influence on the state of analytes in the solution (ionic or neutral molecule), but also affected the sorbent surface charge [27]. The effect of sample pH on the extraction performance ranging from 3.0 to 10.2 was investigated and the results were shown in Fig. 5b. For all phenolic acids, the extraction recoveries decreased from 8.8 to 3.0. This can be attributed that phenolic acids became the protonated molecules gradually with the decrease of sample pH so that the electrostatic repulsion between ILs and analytes would appear and the intramolecular hydrogen-bonding force would become stronger. However, when the pH value was over 8.8 (the neutral molecule), the extraction recovery was also decreased. Although they would be deprotonation in the alkaline condition, phenolic acids with the ionic state were not conducive to escape from the aqueous solution. The original pH value of analytes in the aqueous solution was the optimized condition of sample loading.
3.3. Adsorption characteristics It is crucial for the evaluation of adsorption capacity and characteristics of analytes on the sorbent to demonstrate the adsorption mechanism. Therefore, static-state and dynamic-state adsorption experiments were performed. Fig. 4a exhibited the static adsorption curve of each phenolic acid on the poly(VHIm+PF6−)@GO@Sil. Adsorption amounts in the static equilibrium-state firstly increased with an increase in the initial concentration of analytes, and the adsorption site reached saturation when the concentrations of phenolic acids were above 120 μg mL−1. The dynamic equilibrium adsorbance of poly (VHIm+PF6−)@GO@Sil for analytes gradually increased in the investigated time range (5–40 min) and it became constant after 30 min (Fig. 4b), demonstrating the quick adsorption process and the strong interactions between the sorbent and analytes. Table 1 summarizes the maximum adsorption capacity of each phenolic acid on the poly(VHIm+PF6−)@GO@Sil sorbent. Adsorption amounts of ProA, SyrA, VanA and SalA were 3487, 1866, 1640 and 681 μg g−1, respectively. In general, the adsorption amount was collectively affected by the adsorption interactions and the hydrophilichydrophobic properties of analytes. Besides, the intramolecular
3.4.3. Sample loading rate The effect of sample loading rate on the extraction capacity was also two-sided. Generally, operating time can be saved at a high rate of sample loading while the loss of target compounds would be happened due to the incomplete adsorption of analytes by the sorbent [28]. However, when the sample loading rate was slow, the extraction efficiency of analytes would also reduce due to the appearance of backFig. 4. Static (a) and dynamic (b) adsorption curves of phenolic acids on the poly(VHIm+PF6−)@GO@ Sil.
126
Journal of Chromatography B 1072 (2018) 123–129
X. Hou et al.
Fig. 5. Effect of extraction and desorption conditions on the extraction recoveries of analytes.
3.4.4. Eluent pH For acidic compounds, the pH value would affect the existing form, so the extraction recovery would change with the variation of eluent pH. Eluent pH from 1.5 to 3.5 was tested. Fig. 5d showed that when the eluent pH was higher than 2.0, the recoveries of analytes decreased rapidly. With the enhancement of acidity, the ion exchange interactions would be obvious to improve the elution performance. In the following experiments, pH 2.0 of eluent was adopted.
extraction from the long extraction period [29]. In order to obtain higher recovery within shorter loading time, different sample loading rate (0.5–2.5 mL min−1) was investigated to evaluate their effects on the recoveries. As shown in Fig. 5c, the extraction recoveries of phenolic acids decreased below to 1.5 mL min−1 owing to the back-extraction, while it also reduced above 1.5 mL min−1 due to the insufficient contact between analytes and the sorbent. According to the multiple influences, 1.5 mL min−1 was chosen as the optimized sample loading rate. 127
Journal of Chromatography B 1072 (2018) 123–129
X. Hou et al.
3.4.5. Volume of eluent The volume of eluent had a great effect on the elution efficiency, which was determined by the adsorption amount (the concentration and volume of sample). It was necessary to elute all the analytes from the sorbent with the smallest volume [30]. In this experiment, the volume of eluent (0.5, 0.8, 1.0, 1.5, 2.0 mL) was investigated. Fig. 5e showed that the extraction recoveries of phenolic acids increased with the increase of the eluent volume from 0.5 to 1.0 mL, and remained stable from 1.0 to 2.0 mL. The results indicated that 1.0 mL of methanol/water (2:1, pH 2.0) was enough to obtain the desirable extraction recovery as well as high enrichment factor. Therefore, 1.0 mL was selected as the optimum volume of eluent for the following studies.
Table 2 Linear regression, LODs and RSDs of the developed method for analytes.
ProA SyrA VanA SalA
Linear range (μg L−1)
r
LOD (μg L−1)
RSD (n = 5, %)
1–100 1–100 1–100 1–100
0.9998 0.9998 0.9992 0.9912
0.20 0.25 0.25 0.50
4.0 4.7 4.5 2.3
indicated that GO played an important role in the enhancement of extraction efficiency. The reasons were as follows: (1) GO with large surface area increased the covering amount of ILs and the adsorption site; (2) the delocalized π system on the cation of IL and the surface of GO could interact with the aromatic rings of phenolic acids through π-π interaction; (3) the cation of IL with positive charge would provide electrostatic interaction with the acidic compound; (4) GO and IL were all hydrophilic and poly(VHIm+PF6−)@GO@Sil had a good extraction ability to the polar compound.
3.4.6. Elution rate As for the elution rate, it was also necessary to achieve the best desorption ability in an appropriate value. The elution capacity is mainly determined by the solubility of analytes in the eluent and the flush force. The flush force of eluent is proportional to the elution velocity, while the solubility equilibrium of analytes between the eluent and sorbent is inversely proportional to the elution velocity. The opposite effects of these two factors would lead to the existence of the optimum elution rate. Different elution rates (0.2, 0.3, 0.5, 0.8, 1.1 mL min−1) were tested to evaluate their effects on the extraction recovery of each analyte. As shown in Fig. 5f, the recoveries of phenolic acids decreased due to the incomplete elution after 0.5 mL min−1. Considering the time-saving, 0.5 mL min−1 was selected as the optimal elution rate. Based on the above discussion, all the following experiments were carried out under the appropriate experimental conditions: the sample solution was loaded onto the extraction column at 1.5 mL min−1 of sample loading velocity, and desorbed by 1 mL of methanol-water (2:1, pH 2.0) at a flow rate of 0.3 mL min−1.
3.6. Analytical performance Under the optimized conditions, several analytical parameters, including linear range, correlation coefficients (r), limits of detection (LODs) and reproducibility, were tested to evaluate the performance of poly(VHIm+PF6−)@GO@Sil extraction column. The results were listed in Table 2. The linear range of each analyte was from 1 to 100 μg L−1 with correlation coefficients in the range of 0.9912-0.9998. LODs were 0.20 μg L−1 for ProA, 0.25 μg L−1 for SyrA and VanA, and 0.50 μg L−1 for SalA, which were calculated as three times the signal to noise ratio, and investigated by the extraction of a series of standard solutions spiked at different levels to achieve such a signal. The reproducibility was examined by extracting five identical solutions with five different extraction columns, and the relative standard deviations (RSDs) of the concentrations for different compounds were in the range of 2.3–4.7%. The proposed method was also compared with the other reported sample pretreatment method with the same analytes. As listed in Table 3, LODs of the proposed method for most analytes are much lower than those of the other methods [31–35], indicating the high extraction efficiency of the prepared poly(VHIm+PF6−)@GO@Sil. In order to exhibit the reusability of the sorbent, 30 adsorptiondesorption cycles were repeated using one poly(VHIm+PF6−)@GO@Sil packed SPE cartridge and the result was shown in Fig. S2. The peak areas fluctuated for about 23–25%, partly due to the change of measurement conditions. This indicated that the prepared extraction material was quite stable and durable.
3.5. Comparison of extraction capacity The extraction capacity of poly(VHIm+PF6−)@GO@Sil for phenolic acids was compared with those of the commerical C18, GO@Sil, poly (VHIm+Br−)@GO@Sil, poly(VHIm+Br−)@Sil and poly(VHIm+PF6−) @Sil. As shown in Fig. 6, the recovery efficiencies of poly(VHIm+PF6−) @GO@Sil for ProA, SyrA and SalA were little higher than that of poly (VHIm+Br−)@GO@Sil. However, the poly(VHIm+PF6−)@GO@Sil was selected as the extraction material in this experiment mainly due to its stability and durability. Although there was little difference between the surface area of the prepared PILs-modified silica and the commercial C18 silica, the former presented much higher extraction recovery (about 4 times) and extraction capacity under the same experimental parameters. In addition, the extraction efficiencies of poly (VHIm+PF6−)@GO@Sil for the selected analytes were all superior to those of the GO@Sil, poly(VHIm+Br−)@Sil and poly(VHIm+PF6−)@ Sil, which reflected the synergistic effect of PILs and GO. This also
3.7. Application in real samples In order to test the applicability of the proposed poly(VHIm+PF6−) @GO@Sil-SPE-HPLC method, the black wolfberry yogurt and urine samples were analyzed. Fig. 7 showed the chromatograms of the real samples and samples spiked with the standard solution. The selected phenolic acids were not detected in two real samples. In order to demonstrate the reliability of this method and the matrix effect in the complex sample, recoveries of analytes were determined spiked at 15 μg L−1. As can be seen from Table S3, the recoveries are 71.6%–112.1% and 77.0%–109.4% for the black wolfberry yogurt and urine, respectively. 4. Conclusion In this paper, the poly(VHIm+PF6−)@GO@Sil particle was fabricated by a simple protocol based on the metathesis reaction of ILs anion and packed in a SPE cartridge. The prepared PILs@GO@Sil exhibited the enhanced hydrophilicity and stability. It was used for 15 replicate
Fig. 6. Comparison of the recoveries of phenolic acids extracted by different extraction materials.
128
Journal of Chromatography B 1072 (2018) 123–129
X. Hou et al.
Table 3 Comparison of the proposed method with other methods for the determination of phenolic acids. Sorbent/extractant
MSPE-HPLC-UV a
IP-DLLME-SFO -UHPLCPDA MAEb-HPLC-PDA μ-SPE-UHPLC Q-TOF MS SPE-HPLC-DAD SPE-HPLC-UV a b c
Sample
Sorbent (mg)/extractant (mL)
Volume of sample (mL)
LOD (μg L−1)
Ref.
ProA
SyrA
VanA
SalA
Molecularly imprinted polymers –
River water
10
40
10
–
–
20
[31]
Wine
0.5
5
75
–
50
–
[32]
Methanol OASIS HLB plate
Rice Human plasma and urine Pollen Typha angustifolia Yogurt and urine
25,000 –
2500 0.6
760 0.1
140 9.3
– 1.4
– –
[33] [34]
40
4
10
–
10
–
[35]
30
40
0.2
0.25
0.25
0.5
This work
Hollow porous PILsc +
−
poly(VBIm PF6 )@GO @Sil
IP-DLLME-SFO, ion pair dispersive liquid-liquid microextraction-floating organic droplet. MAE, microwave-assisted extraction. PILs, ionic liquids composite polymers. [4] F. Tang, Q.W. Yu, B.F. Yuan, Y.Q. Feng, Trends Anal. Chem. 86 (2017) 172–184. [5] B.H. Fumes, M.R. Silva, F.N. Andrade, C.E.D. Nazario, F.M. Lancas, Trends Anal. Chem. 71 (2015) 9–25. [6] J. Płotka-Wasylka, N. Szczepańska, M. Guardia, J. Namieśnik, Trends Anal. Chem. 77 (2016) 23–43. [7] S.A. Ansari, P.K. Mohapatra, J. Chromatogr. A 1499 (2017) 1–20. [8] A. Andrade-Eiroa, M. Canle, V. Leroy-Cancellieri, V. Cerda, Trends Anal. Chem. 80 (2016) 641–654. [9] A. Andrade-Eiroa, M. Canle, V. Leroy-Cancellieri, V. Cerda, Trends Anal. Chem. 80 (2016) 655–667. [10] M. Tuzen, S. Sahiner, B. Hazer, Food Chem. 210 (2016) 115–120. [11] E. Yavuz, Ş. Tokalıoğlu, H. Şahan, Ş. Patat, Talanta 128 (2014) 31–37. [12] M. He, L.J. Huang, B.S. Zhao, B.B. Chen, B. Hu, Anal. Chim. Acta 973 (2017) 1–24. [13] B. Hu, M. He, B. Chen, Anal. Bioanal. Chem. 407 (2015) 2685–2710. [14] Q. Liu, J. Shi, J. Sun, T. Wang, L. Zeng, G. Jiang, Angew. Chem. 123 (2011) 6035–6039. [15] J. Kim, L.J. Cote, F. Kim, W. Yuan, K.R. Shull, J. Huang, J. Am. Chem. Soc. 132 (2010) 8180–8186. [16] X.D. Hou, X.S. Wang, Y.X. Sun, L.C. Wang, Y. Guo, Anal. Bioanal. Chem. 409 (2017) 3541–3549. [17] S. Stankovich, D.A. Dikin, G.H.B. Dommett, K.M. Kohlhaas, E.J. Zimney, E.A. Stach, R.D. Piner, S.T. Nguyen, R.S. Ruoff, Nature 442 (2006) 282–286. [18] X.D. Hou, L.C. Wang, X.F. Tang, C.M. Xiong, Y. Guo, X. Liu, Analyst 140 (2015) 6727–6735. [19] R.D. Rogers, K.R. Seddon, Science 302 (2003) 792–793. [20] B. Liu, N. Jin, Curr. Org. Chem. 20 (2016) 2109–2116. [21] K. Goossens, K. Lava, C.W. Bielawski, K. Binnemans, Chem. Rev. 116 (2016) 4643–4807. [22] X.D. Hou, S.J. Liu, P.P. Zhou, J. Li, X. Liu, L.C. Wang, Y. Guo, J. Chromatogr. A 1456 (2016) 10–18. [23] J.L. Anderson, J. Ding, T. Welton, D.W. Armstrong, J. Am, Chem. Soc. 124 (2002) 14247–14254. [24] J.W.S. Hummers, R.E. Offeman, J. Am. Chem. Soc. 80 (1958) 1339. [25] Y.X. Sun, S. Han, J. Mol. Model. 19 (2013) 5407–5422. [26] Y. Suneetha, B.N. Kumar, Y. Harinath, D.H.K. Reddy, K. Seshaiah, Microchim. Acta. 176 (2012) 169–176. [27] T.T. Wang, Y.H. Chen, J.F. Ma, M.L. Chen, C.G. Nie, M.J. Hu, Y. Li, Z.J. Jia, J.H. Fang, H.Q. Gao, J. Chromatogr. A 2013 (2013) 63–72. [28] J.P. Ma, R.H. Xiao, J.H. Li, J.B. Yu, Y.P. Zhang, L.X. Chen, J. Chromatogr. A 1217 (2010) 5462–5469. [29] J.L. Benede, A. Chisvert, D.L. Giokas, A. Salvador, J. Chromatogr. A 1362 (2014) 25–33. [30] N. Wang, J.M. Wang, Y. Liao, S.J. Shao, Talanta 151 (2016) 1–7. [31] S.Y. Shi, J.F. Guo, Q.P. You, X.Q. Chen, Y.P. Zhang, Chem. Eng. J 243 (2014) 485–493. [32] J. Li, S.D. Jia, S.J. Yoon, S.J. Lee, S.W. Kwon, J. Lee, J. Food Compos Anal. 45 (2016) 73–79. [33] W. Setyaningsih, I.E. Saputro, M. Palma, C.G. Barroso, Food Chem. 169 (2015) 141–149. [34] R.P. Feliciano, E. Mecha, M.R. Bronze, A. Rodriguez-Mateos, J. Chromatogr. A 2016 (1464) 21–31. [35] X.P. Dai, D.S. Wang, H. Li, Y.Y. Chen, Z.C. Gong, H.Y. Xiang, S.Y. Shi, X.Q. Chen, J. Chromatogr. A 1484 (2017) 7–13.
Fig. 7. Chromatograms of the black wolfberry yogurt (a), urine (c) and samples spiked with 15 μg L−1 of the standard solution (b, d).
extractions without an obvious loss. The introduction of PIL not only affected the zeta potential of GO surface, but also offered electrostatic interactions to extract the acidic compound. Compared with other adsorbents (C18, GO@Sil, poly(VHIm+Br−)@Sil and poly(VHIm+PF6−) @Sil), the as-prepared poly(VHIm+PF6−)@GO@Sil showed higher extraction recovery for the polar phenolic acid. Under the optimized conditions, good linearity, low LODs and desirable spiked recoveries were obtained for the real sample. The results indicated that poly (VHIm+PF6−)@GO@Sil had a great potential as a polar adsorbent for acidic analytes. Acknowledgements This work was supported by the National Natural Science Foundation of China (21575149, 21505146 and 21575148); and Key Laboratory of Chemistry of Northwestern Plant Resources, Chinese Academy of Sciences, China (CNPR-2015kfkt-01). References [1] J. Dai, R.J. Mumper, Molecules 15 (2010) 7313–7352. [2] R.P. Feliciano, E. Mecha, M.R. Bronze, A. Rodriguez-Mateos, J. Chromatogr. A 2016 (1464) 21–31. [3] I.A. Ludwig, P. Mena, L. Calani, G. Borges, G. Pereira-Caro, L. Bresciani, D.D. Rio, E.J.L. Michael, C. Alan, Free Radic. Biol. Med. 89 (2015) 758–769.
129
Contents lists available at ScienceDirect
Journal of Chromatography B journal homepage: www.elsevier.com/locate/jchromb
Graphene oxide reinforced ionic liquid-functionalized adsorbent for solid-phase extraction of phenolic acids ⁎
Xiudan Houa,c, Xiaofeng Lua, Sheng Tangb, Licheng Wanga, , Yong Guoa,
MARK
⁎
a CAS Key Laboratory of Chemistry of Northwestern Plant Resources and Key Laboratory for Natural Medicine of Gansu Province, Lanzhou Institute of Chemical Physics, Chinese Academy of Sciences, Lanzhou 730000, China b School of Chemistry and Environmental Engineering, Wuhan Institute of Technology, Wuhan 430205, China c University of Chinese Academy of Sciences, Chinese Academy of Sciences, Beijing 100049, China
A R T I C L E I N F O
A B S T R A C T
Keywords: Adsorption energy Graphene oxide Ionic liquid Phenolic acid Solid-phase extraction
An environmental friendly sorbent of polymeric ionic liquids modified graphene oxide-grafted silica (PILs@GO@ Sil) was synthesized for solid-phase extraction (SPE) of phenolic acids. The sorbent was prepared via a chemical layer-to-layer fabrication including amidation reaction, surface radical chain-transfer polymerization and in situ anion exchange. After modification with PILs, the silica surface had higher positive potential so that it would exhibit stronger electrostatic interaction for acidic compounds compared with GO@Sil. The adsorption performance of phenolic acids was investigated through the theoretical calculation and static, kinetic state adsorption experiments. Under the optimized conditions, wide linear ranges were obtained with correlation coefficients ranging from 0.9912 to 0.9998, and limits of detection were in the range of 0.20–0.50 μg L−1. Compared with other reported methods, the proposed PILs@GO@Sil-SPE-HPLC showed higher extraction efficiency. Finally, the black wolfberry yogurt and urine were analyzed as real samples and good recoveries spiked with standard solution were obtained.
1. Introduction Phenolic acids are a class of secondary metabolites, which have drawn increasing attention due to their antioxidant properties and marked effects in the prevention of various oxidative stress associated diseases such as cardiovascular disease and cancer [1,2]. They widely spread throughout the plant kingdom including vegetables and fruits, and further exist in beverages as the food ingredient. Once ingested, most phenolic acids are extensively metabolized by the enzymes and partly adsorbed or excreted [3]. It is essential to enrich and purify before the instrumental analysis due to the low concentration of analytes and the interference of other complex substance in real samples. Therefore, sample pretreatment is one of the most important procedures in the whole analysis, especially for the analysis of biological and environmental samples with complex matrix [4], such as solid-phase extraction (SPE), solid-phase microextraction, liquid-phase microextraction, electromembrane extraction, microextraction by packed sorbents and so on. This stage would realize the pre-concentration of target compounds and elimination of most matrix interferences before introduction into the analytical instrument [5]. SPE is the most widely used separation and pre-concentration technique owing to its high chromatographic utility [6,7]. New various materials as adsorbents in ⁎
SPE have been exploited to extract organic and inorganic compounds from the complex media with higher adsorption capacity [8–10]. The studies indicated that nanosized SPE sorbents have high extraction capacities with rapid extraction kinetic performance [11]. Nanostructured materials, such as carbon-based nanomaterials, metal and metal oxide nanoparticles, electrospun nanofibers, and metal-organic frameworks, have gained great attention because of their physical and chemical properties including large surface area, desirable chemical and thermal stability, and favorable adsorption performance [4,12,13]. Graphene and its precursor graphene oxide (GO), as 2-dimensional nanoscale materials, have sparked intense research interest in sample preparation because of the outstanding properties. GO contains miscellaneous chemical functional groups on the basal planes and at the edges of GO sheets, such as hydroxyl, epoxy and carboxyl, which are expected to promote interfacial interactions between GO and adsorbate [14,15]. In the previous work, based on the excellent adsorption performance, GO-grafted silica was fabricated and used to determine trace phenolic acids in urine [16]. Besides, GO could be modified conveniently through the chemical fabrication procedure to improve the extraction yield and selectivity for some specific analytes [17,18]. Ionic liquids (ILs) are a class of organic salts composed entirely of organic cations (e.g. pyrrolidinium, pyridinium, imidazolium,
Corresponding authors. E-mail addresses: [email protected] (L. Wang), [email protected] (Y. Guo).
https://doi.org/10.1016/j.jchromb.2017.11.013 Received 29 August 2017; Received in revised form 8 November 2017; Accepted 10 November 2017 Available online 12 November 2017 1570-0232/ © 2017 Published by Elsevier B.V.
Journal of Chromatography B 1072 (2018) 123–129
X. Hou et al.
Fig. 1. Scheme of the synthetic route for poly(VHIm+PF6−)@GO@Sil (a) and the interactions between the extraction material and analytes (b).
water (18.2 MΩ), which was collected from a Molecular water purification system.
tetraalkyl ammonium or tetraalkyl phosphonium) and inorganic or organic anions (e.g. Br−, BF4−, CF3SO3−, NTf2, PF6−). Due to the unique physicochemical properties, such as low vapor pressure, good chemical and thermal stability, wide viscosities range and miscibility with water and organic solvents, ILs have been used as promising solvents and materials in many fields [19–21]. Imidazolium IL, a member of ILs family, is widely used for sample preparation on the basis of the pentacyclic structure and the easily tuned property. In the previous reports, polymeric ionic liquids (PILs) modified GO-grafted silica exhibited high extraction efficiency for flavonoids because of its capacity of most types of interactions with analytes (e.g., π-π, n-π, hydrogen bonding, dispersive, dipolar, ionic/charge-charge) [22,23]. The satisfactory results indicated that the presence of GO would increase the overall extraction recoveries of analytes and bring an enhancement in analyte transport. In this paper, the properties of an imidazolium-based PIL were tuned based on the reconstruction of its counter anion via in situ anion exchange. Firstly, poly(1-vinyl-3-hexylimidazolium hexafluorophosphate)-GO-grafted silica (poly(VHIm+PF6−)@GO@Sil) was synthesized and characterized. Next, the adsorption performance of phenolic acids on the proposed adsorbent was investigated through the theory calculation and a series of adsorption experiments. Comparisons of extraction efficiency with other prepared materials were also performed. Under the optimum conditions, the proposed poly (VHIm+PF6−)@GO@Sil-SPE-HPLC method was used to determine the selected phenolic acids in yogurt and urine.
2.2. Instrumentation and chromatographic conditions The SPE procedure was performed on a HGC-8 numerical control solid phase extraction system (Hegong Scientific Instrument Co., Shanghai, China). HPLC analysis were performed on an Agilent 1100 Series modular HPLC system (Agilent Technologies, USA) with a highpressure quaternary pump, a 20 μL sample loop and a UV–vis detector. Separation of the analytes was performed on a C18 column (Hypersil ODS2, 250 mm length × 4.6 mm i.d., 5 μm). The mobile phase was methanol and water with 0.25% acetic acid. The gradient elution condition was 0–23 min, 15–40% methanol; 23–30 min, 40–70% methanol. Flow rate was set at 0.8 mL min−1 while the detection wavelength was 298 nm. BET surface area was detected by a Micromeritics ASAP 2020 device. Surface properties of the prepared silica were characterized by scanning electron microscope (SEM, JSM-6701F, Japan). Energy dispersive spectrometry (EDS) was obtained on a low vacuum scanning electron microscope-X-ray energy dispersive spectrometer (JSM5600LV, Japan). The zeta potential was determined with a Zetasizer Nano series ZS instrument (Malvern Instruments, United Kingdom).
2.3. Preparation of PILs/GO-modified silica 2. Experimental
Hummers method was used to synthesize GO from the natural graphite [24]. Fig. 1 shows the preparation procedure of the PILs/GO-modified silica. Firstly, 8 g of silica was immersed in hydrochloric acid to be activated for 24 h and then rinsed with distilled water. The neutral silica and 8 mL of APTES were added in 100 mL of dry toluene to react and reflux at 120 °C. After 24 h, the obtained aminopropyl-modified silica (Sil-NH2), 10 mL of GO solution (0.1%, w/v), 100 mL of phosphate buffer solution (pH, 7.4), 0.05 g of EDC and 0.05 g of NHS were added into a 250 mL reaction flask. GO-grafted silica (GO@Sil) was obtained through the above amidation reaction. Mercaptopropyl-modified GO-grafted silica (GO@Sil-SH) was prepared being similar to that of Sil-NH2, in which APTES was replaced by MPTES. The fabrication of poly(VHIm+PF6−)@GO@Sil involved the following two processes (surface radical chain-transfer polymerization and in situ anion-exchange): (a) 8 g of VHIm+Br− and 0.4 g of AIBN were dissolved in 100 mL of DMSO and added into the conical flask with the prepared GO@Sil-SH, which was refluxed for 24 h at 60 °C under an N2 atmosphere to obtain poly(VHIm+Br−)@GO@Sil particles; (b) after cooling, rinsed and dried, the poly(VHIm+Br−)@GO@Sil particles and 8 g of NH4PF6 were added into a 250 mL three-neck round-bottomed flask with 150 mL of aqueous solution for reacting at 60 °C for 12 h. Finally, the obtained poly(VHIm+PF6−)@GO@Sil was dried, stored, characterized and further used as the extraction material.
2.1. Chemicals and materials Protocatechuic acid (ProA), vanillic acid (VanA), N-hydroxy succinimide (NHS) and N-(3-dimethylaminopropyl)-N-ethyl-carbodiimide (EDC) were obtained from Aladdin Chemical Reagent Co. (Shanghai, China). Syringic acid (SyrA) and salicylic acid were purchased from Energy Chemical (Shanghai, China). The chemical structures of the selected phenolic acids are shown in Fig. S1. VHIm+Br− was purchased from Shanghai Chengjie Chemical Co. (Shanghai, China). Ammonium hexafluorophosphate (NH4PF6) was purchased from Shanghai Tiancheng Chemical Co. (Shanghai, China). Azobisisobutyronitrile (AIBN) was obtained from Shanpu Chemical Co. (Shanghai, China). Dimethyl sulfoxide (DMSO) was obtained from J&K scientific LTD. (Beijing, China). 3-Aminopropyltriethoxysilane (APTES) and 3-mercaptopropyltriethoxysilane (MPTES) were purchased from Chemical Industrial Corporation of Gaizhou (Guangzhou, China). SPE empty column (3 mL), polyethylene frits (5 mm pore size) and the commercial C18 SPE column were purchased from Shenzhen Biocomma Biotech Co. (Shenzhen, China). Stock solutions of phenolic acids were prepared in methanol with the concentration of 0.2 mg mL−1. Working solutions for extraction were freshly prepared by diluting the stock solutions with ultrapure 124
Journal of Chromatography B 1072 (2018) 123–129
X. Hou et al.
Fig. 2. SEM images of poly(VHIm+PF6−)@GO@Sil.
same conditions were performed and average values were provided as the final data. Recovery (%) of the analytes was calculated and evaluated by the following equation:
2.4. Preparation of sample solutions The black wolfberry yogurt was bought from a local supermarket (Lanzhou, China) and then centrifuged, filtered. The urine sample was collected from a healthy volunteer, which was first coarsely filtered through a Buchner funnel, and then through a 5 μm sieve plate. The pretreated samples were stored at 4 °C for further use.
Recovery =
where Csol and Csam are the enrichment concentration of analytes in eluent and the initial concentration of analytes in the working solution, respectively; Vsol and Vsam are the volumes of eluent and working solution. Csol is obtained from the detected peak area and the calibration curve, which is plotted by a series of concentrations of the standard solution and peak area determined by HPLC.
2.5. Adsorption experiments 2.5.1. Theoretical adsorption energy The calculation for adsorption energies of analytes on poly (VHIm+Br−) and poly(VHIm+PF6−) were performed with Gaussian 09 software, which was based on periodic density functional theory with M06-2X function [25].
3. Results and discussions 3.1. Characterization of the prepared materials
2.5.2. Investigation of the adsorption mechanism In static-state binding experiments, 20 mg of the poly(VHIm+PF6−) @GO@Sil particles was placed in a 10 mL glass flask with 1 mL of standard solutions at the concentrations of 5–130 μg mL−1. After shaking for 1 h, the mixture was centrifuged to separate the sorbent and solution; the upper solution was injected into HPLC to analyze the retained phenolic acids. In dynamic-state binding experiments, the poly (VHIm+PF6−)@GO@Sil particles (20 mg) was mixed with 1 mL of the phenolic acids solution (100 μg mL−1), and shaken at a regular time intervals (5, 10, 20, 30, 40, 50, 60 min). After centrifugation, the supernatants were determined to study the kinetic-adsorption process of poly(VHIm+PF6−)@GO@Sil. The adsorption capacities (q, μg g−1) were calculated by the following equation:
q=
CsolVsol × 100% CsamVsam
Surface morphology characteristic of the prepared poly (VHIm+PF6−)@GO@Sil was characterized by SEM (shown in Fig. 2). The uniform, rough surface with dense protuberances was observed, which greatly increased the surface area to volume ratio. The successful synthesis of poly(VHIm+PF6−)@GO@Sil was verified by the EDS analysis. As shown in Table S1, bromine element appeared on the surface of poly(VHIm+Br−)@GO@Sil, and fluorine element was occurred on the surface of poly(VHIm+PF6−)@GO@Sil, which sufficiently proved the successful preparation of two sorbents. To investigate the surface charge of the as-prepared extraction materials, the zeta potential was separately measured in different pH solutions (Fig. 3). After modified with PILs, it was obviously found that the sorbent had a bigger positive potential on account of the alteration of the surface functional groups. Therefore, poly(VHIm+PF6−)@GO@Sil was more suitable for the extraction of acidic compounds. The BET
(Csam-Cf )×Vsam W
where Csam (μg mL−1) and Cf (μg mL−1) are the initial and final concentrations of the sample solution, respectively; Vsam (mL) is the sample volume; W (g) is the weight of poly(VHIm+PF6−)@GO@Sil particles. 2.6. Solid-phase extraction procedure and recovery calculation For all SPE experiments, 30 mg of the prepared silica particles were packed into a 3 mL empty propylene cartridge with two sieve plates positioned at each end of the column. After that, the SPE packing column was activated by loading with 2 mL of eluent (methanol/water, 2:1, pH = 2) and 10 mL of methanol/water (v/v, 1:1). 40 mL of the working solution or sample solution was passed through the cartridge at a constant flow rate of 1.5 mL min−1 during the whole SPE procedure adjusted with the vacuum pump. After loading, the sorbent was washed with 1 mL of ultrapure water and dried in air for 5 min. The retained analytes on the cartridge were eluted with 1 mL of methanol/water (2:1, pH = 2) at a flow rate of 0.5 mL min−1. Then the desorption solution was used for HPLC-UV analysis. Triplicate analyses under the
Fig. 3. Zeta potential of poly(VHIm+PF6−)@GO@Sil and GO@Sil.
125
Journal of Chromatography B 1072 (2018) 123–129
X. Hou et al.
hydrogen bond and the stereochemcial structure resulted in the low adsorption amount of SalA on the sorbent.
Table 1 Molecular weight, octanol-water partition coefficient (log Kow) of phenolic acids and adsorption energy, adsorption amounts of analytes on the sorbent. Molecular Weight
ProA SyrA VanA SalA a
154.12 198.18 168.15 138.12
log Kowa
0.91 1.04 1.22 2.26
Adsorption energy (kJ mol−1) Br−-PIL
PF6−-PIL
−91.92 −97.02 −97.58 −88.95
−127.91 −111.49 −108.75 −111.14
3.4. Optimization of experimental conditions
Adsorption amount (μg g−1)
The prepared poly(VHIm+PF6−)@GO@Sil-SPE extraction column was applied to the extraction of phenolic acids as target analytes in the yogurt and urine. In order to evaluate the extraction and desorption performance, the experiment parameters including extraction conditions (volume of sample loading, sample loading rate, pH of sample) and elution conditions (eluent, volume of eluent, elution rate) were investigated. All the experiments were carried out in triplicate and the average was taken into the discussion.
3487 1866 1640 681
Kow: Octanol-water partitioning coefficients.
surface area, total pore volume and average pore width of commerical C18, poly(VHIm+Br−)@GO@Sil and poly(VHIm+PF6−)@GO@Sil were listed in Table S2. The discrepancy of two PILs-based silica was derived from the anion exchange reaction.
3.4.1. Volume of sample loading When the concentration of sample solution was constant, an increase of volume of sample loading would improve the enrichment factor [26]. Adsorption capacity of the extraction material had a main effect on the volume of sample loading. With the increase of sample volume, part of analytes might be loss due to the saturation of the sorbent for a large sample volume. In this work, on the basis of 100 μg L−1 of working solution, the sample volume in the range of 20–70 mL was investigated. Fig. 5a showed that when the volume of sample loading was above 40 mL, the extraction recovery decreased distinctly. Therefore, 40 mL was selected as the optimized volume of sample loading.
3.2. Adsorption energies The theoretical calculation results of adsorption energy of the selected phenolic acids on poly(VHIm+Br−) and poly(VHIm+PF6−) were presented in Table 1. The adsorption energies of ProA, SyrA and SalA on poly(VHIm+PF6−) were higher than those on poly(VHIm+PF6−). Therefore, the poly(VHIm+PF6−)@GO@Sil was selected as the suitable materials in the following experiments. The different adsorption energies of analytes on one sorbent were resulted from the molecular stereochemical structures and various interaction forces, including hydrogen bonding, π-π and electrostatic interactions.
3.4.2. Sample pH Sample pH played a vital role in the extraction process since it not only had an influence on the state of analytes in the solution (ionic or neutral molecule), but also affected the sorbent surface charge [27]. The effect of sample pH on the extraction performance ranging from 3.0 to 10.2 was investigated and the results were shown in Fig. 5b. For all phenolic acids, the extraction recoveries decreased from 8.8 to 3.0. This can be attributed that phenolic acids became the protonated molecules gradually with the decrease of sample pH so that the electrostatic repulsion between ILs and analytes would appear and the intramolecular hydrogen-bonding force would become stronger. However, when the pH value was over 8.8 (the neutral molecule), the extraction recovery was also decreased. Although they would be deprotonation in the alkaline condition, phenolic acids with the ionic state were not conducive to escape from the aqueous solution. The original pH value of analytes in the aqueous solution was the optimized condition of sample loading.
3.3. Adsorption characteristics It is crucial for the evaluation of adsorption capacity and characteristics of analytes on the sorbent to demonstrate the adsorption mechanism. Therefore, static-state and dynamic-state adsorption experiments were performed. Fig. 4a exhibited the static adsorption curve of each phenolic acid on the poly(VHIm+PF6−)@GO@Sil. Adsorption amounts in the static equilibrium-state firstly increased with an increase in the initial concentration of analytes, and the adsorption site reached saturation when the concentrations of phenolic acids were above 120 μg mL−1. The dynamic equilibrium adsorbance of poly (VHIm+PF6−)@GO@Sil for analytes gradually increased in the investigated time range (5–40 min) and it became constant after 30 min (Fig. 4b), demonstrating the quick adsorption process and the strong interactions between the sorbent and analytes. Table 1 summarizes the maximum adsorption capacity of each phenolic acid on the poly(VHIm+PF6−)@GO@Sil sorbent. Adsorption amounts of ProA, SyrA, VanA and SalA were 3487, 1866, 1640 and 681 μg g−1, respectively. In general, the adsorption amount was collectively affected by the adsorption interactions and the hydrophilichydrophobic properties of analytes. Besides, the intramolecular
3.4.3. Sample loading rate The effect of sample loading rate on the extraction capacity was also two-sided. Generally, operating time can be saved at a high rate of sample loading while the loss of target compounds would be happened due to the incomplete adsorption of analytes by the sorbent [28]. However, when the sample loading rate was slow, the extraction efficiency of analytes would also reduce due to the appearance of backFig. 4. Static (a) and dynamic (b) adsorption curves of phenolic acids on the poly(VHIm+PF6−)@GO@ Sil.
126
Journal of Chromatography B 1072 (2018) 123–129
X. Hou et al.
Fig. 5. Effect of extraction and desorption conditions on the extraction recoveries of analytes.
3.4.4. Eluent pH For acidic compounds, the pH value would affect the existing form, so the extraction recovery would change with the variation of eluent pH. Eluent pH from 1.5 to 3.5 was tested. Fig. 5d showed that when the eluent pH was higher than 2.0, the recoveries of analytes decreased rapidly. With the enhancement of acidity, the ion exchange interactions would be obvious to improve the elution performance. In the following experiments, pH 2.0 of eluent was adopted.
extraction from the long extraction period [29]. In order to obtain higher recovery within shorter loading time, different sample loading rate (0.5–2.5 mL min−1) was investigated to evaluate their effects on the recoveries. As shown in Fig. 5c, the extraction recoveries of phenolic acids decreased below to 1.5 mL min−1 owing to the back-extraction, while it also reduced above 1.5 mL min−1 due to the insufficient contact between analytes and the sorbent. According to the multiple influences, 1.5 mL min−1 was chosen as the optimized sample loading rate. 127
Journal of Chromatography B 1072 (2018) 123–129
X. Hou et al.
3.4.5. Volume of eluent The volume of eluent had a great effect on the elution efficiency, which was determined by the adsorption amount (the concentration and volume of sample). It was necessary to elute all the analytes from the sorbent with the smallest volume [30]. In this experiment, the volume of eluent (0.5, 0.8, 1.0, 1.5, 2.0 mL) was investigated. Fig. 5e showed that the extraction recoveries of phenolic acids increased with the increase of the eluent volume from 0.5 to 1.0 mL, and remained stable from 1.0 to 2.0 mL. The results indicated that 1.0 mL of methanol/water (2:1, pH 2.0) was enough to obtain the desirable extraction recovery as well as high enrichment factor. Therefore, 1.0 mL was selected as the optimum volume of eluent for the following studies.
Table 2 Linear regression, LODs and RSDs of the developed method for analytes.
ProA SyrA VanA SalA
Linear range (μg L−1)
r
LOD (μg L−1)
RSD (n = 5, %)
1–100 1–100 1–100 1–100
0.9998 0.9998 0.9992 0.9912
0.20 0.25 0.25 0.50
4.0 4.7 4.5 2.3
indicated that GO played an important role in the enhancement of extraction efficiency. The reasons were as follows: (1) GO with large surface area increased the covering amount of ILs and the adsorption site; (2) the delocalized π system on the cation of IL and the surface of GO could interact with the aromatic rings of phenolic acids through π-π interaction; (3) the cation of IL with positive charge would provide electrostatic interaction with the acidic compound; (4) GO and IL were all hydrophilic and poly(VHIm+PF6−)@GO@Sil had a good extraction ability to the polar compound.
3.4.6. Elution rate As for the elution rate, it was also necessary to achieve the best desorption ability in an appropriate value. The elution capacity is mainly determined by the solubility of analytes in the eluent and the flush force. The flush force of eluent is proportional to the elution velocity, while the solubility equilibrium of analytes between the eluent and sorbent is inversely proportional to the elution velocity. The opposite effects of these two factors would lead to the existence of the optimum elution rate. Different elution rates (0.2, 0.3, 0.5, 0.8, 1.1 mL min−1) were tested to evaluate their effects on the extraction recovery of each analyte. As shown in Fig. 5f, the recoveries of phenolic acids decreased due to the incomplete elution after 0.5 mL min−1. Considering the time-saving, 0.5 mL min−1 was selected as the optimal elution rate. Based on the above discussion, all the following experiments were carried out under the appropriate experimental conditions: the sample solution was loaded onto the extraction column at 1.5 mL min−1 of sample loading velocity, and desorbed by 1 mL of methanol-water (2:1, pH 2.0) at a flow rate of 0.3 mL min−1.
3.6. Analytical performance Under the optimized conditions, several analytical parameters, including linear range, correlation coefficients (r), limits of detection (LODs) and reproducibility, were tested to evaluate the performance of poly(VHIm+PF6−)@GO@Sil extraction column. The results were listed in Table 2. The linear range of each analyte was from 1 to 100 μg L−1 with correlation coefficients in the range of 0.9912-0.9998. LODs were 0.20 μg L−1 for ProA, 0.25 μg L−1 for SyrA and VanA, and 0.50 μg L−1 for SalA, which were calculated as three times the signal to noise ratio, and investigated by the extraction of a series of standard solutions spiked at different levels to achieve such a signal. The reproducibility was examined by extracting five identical solutions with five different extraction columns, and the relative standard deviations (RSDs) of the concentrations for different compounds were in the range of 2.3–4.7%. The proposed method was also compared with the other reported sample pretreatment method with the same analytes. As listed in Table 3, LODs of the proposed method for most analytes are much lower than those of the other methods [31–35], indicating the high extraction efficiency of the prepared poly(VHIm+PF6−)@GO@Sil. In order to exhibit the reusability of the sorbent, 30 adsorptiondesorption cycles were repeated using one poly(VHIm+PF6−)@GO@Sil packed SPE cartridge and the result was shown in Fig. S2. The peak areas fluctuated for about 23–25%, partly due to the change of measurement conditions. This indicated that the prepared extraction material was quite stable and durable.
3.5. Comparison of extraction capacity The extraction capacity of poly(VHIm+PF6−)@GO@Sil for phenolic acids was compared with those of the commerical C18, GO@Sil, poly (VHIm+Br−)@GO@Sil, poly(VHIm+Br−)@Sil and poly(VHIm+PF6−) @Sil. As shown in Fig. 6, the recovery efficiencies of poly(VHIm+PF6−) @GO@Sil for ProA, SyrA and SalA were little higher than that of poly (VHIm+Br−)@GO@Sil. However, the poly(VHIm+PF6−)@GO@Sil was selected as the extraction material in this experiment mainly due to its stability and durability. Although there was little difference between the surface area of the prepared PILs-modified silica and the commercial C18 silica, the former presented much higher extraction recovery (about 4 times) and extraction capacity under the same experimental parameters. In addition, the extraction efficiencies of poly (VHIm+PF6−)@GO@Sil for the selected analytes were all superior to those of the GO@Sil, poly(VHIm+Br−)@Sil and poly(VHIm+PF6−)@ Sil, which reflected the synergistic effect of PILs and GO. This also
3.7. Application in real samples In order to test the applicability of the proposed poly(VHIm+PF6−) @GO@Sil-SPE-HPLC method, the black wolfberry yogurt and urine samples were analyzed. Fig. 7 showed the chromatograms of the real samples and samples spiked with the standard solution. The selected phenolic acids were not detected in two real samples. In order to demonstrate the reliability of this method and the matrix effect in the complex sample, recoveries of analytes were determined spiked at 15 μg L−1. As can be seen from Table S3, the recoveries are 71.6%–112.1% and 77.0%–109.4% for the black wolfberry yogurt and urine, respectively. 4. Conclusion In this paper, the poly(VHIm+PF6−)@GO@Sil particle was fabricated by a simple protocol based on the metathesis reaction of ILs anion and packed in a SPE cartridge. The prepared PILs@GO@Sil exhibited the enhanced hydrophilicity and stability. It was used for 15 replicate
Fig. 6. Comparison of the recoveries of phenolic acids extracted by different extraction materials.
128
Journal of Chromatography B 1072 (2018) 123–129
X. Hou et al.
Table 3 Comparison of the proposed method with other methods for the determination of phenolic acids. Sorbent/extractant
MSPE-HPLC-UV a
IP-DLLME-SFO -UHPLCPDA MAEb-HPLC-PDA μ-SPE-UHPLC Q-TOF MS SPE-HPLC-DAD SPE-HPLC-UV a b c
Sample
Sorbent (mg)/extractant (mL)
Volume of sample (mL)
LOD (μg L−1)
Ref.
ProA
SyrA
VanA
SalA
Molecularly imprinted polymers –
River water
10
40
10
–
–
20
[31]
Wine
0.5
5
75
–
50
–
[32]
Methanol OASIS HLB plate
Rice Human plasma and urine Pollen Typha angustifolia Yogurt and urine
25,000 –
2500 0.6
760 0.1
140 9.3
– 1.4
– –
[33] [34]
40
4
10
–
10
–
[35]
30
40
0.2
0.25
0.25
0.5
This work
Hollow porous PILsc +
−
poly(VBIm PF6 )@GO @Sil
IP-DLLME-SFO, ion pair dispersive liquid-liquid microextraction-floating organic droplet. MAE, microwave-assisted extraction. PILs, ionic liquids composite polymers. [4] F. Tang, Q.W. Yu, B.F. Yuan, Y.Q. Feng, Trends Anal. Chem. 86 (2017) 172–184. [5] B.H. Fumes, M.R. Silva, F.N. Andrade, C.E.D. Nazario, F.M. Lancas, Trends Anal. Chem. 71 (2015) 9–25. [6] J. Płotka-Wasylka, N. Szczepańska, M. Guardia, J. Namieśnik, Trends Anal. Chem. 77 (2016) 23–43. [7] S.A. Ansari, P.K. Mohapatra, J. Chromatogr. A 1499 (2017) 1–20. [8] A. Andrade-Eiroa, M. Canle, V. Leroy-Cancellieri, V. Cerda, Trends Anal. Chem. 80 (2016) 641–654. [9] A. Andrade-Eiroa, M. Canle, V. Leroy-Cancellieri, V. Cerda, Trends Anal. Chem. 80 (2016) 655–667. [10] M. Tuzen, S. Sahiner, B. Hazer, Food Chem. 210 (2016) 115–120. [11] E. Yavuz, Ş. Tokalıoğlu, H. Şahan, Ş. Patat, Talanta 128 (2014) 31–37. [12] M. He, L.J. Huang, B.S. Zhao, B.B. Chen, B. Hu, Anal. Chim. Acta 973 (2017) 1–24. [13] B. Hu, M. He, B. Chen, Anal. Bioanal. Chem. 407 (2015) 2685–2710. [14] Q. Liu, J. Shi, J. Sun, T. Wang, L. Zeng, G. Jiang, Angew. Chem. 123 (2011) 6035–6039. [15] J. Kim, L.J. Cote, F. Kim, W. Yuan, K.R. Shull, J. Huang, J. Am. Chem. Soc. 132 (2010) 8180–8186. [16] X.D. Hou, X.S. Wang, Y.X. Sun, L.C. Wang, Y. Guo, Anal. Bioanal. Chem. 409 (2017) 3541–3549. [17] S. Stankovich, D.A. Dikin, G.H.B. Dommett, K.M. Kohlhaas, E.J. Zimney, E.A. Stach, R.D. Piner, S.T. Nguyen, R.S. Ruoff, Nature 442 (2006) 282–286. [18] X.D. Hou, L.C. Wang, X.F. Tang, C.M. Xiong, Y. Guo, X. Liu, Analyst 140 (2015) 6727–6735. [19] R.D. Rogers, K.R. Seddon, Science 302 (2003) 792–793. [20] B. Liu, N. Jin, Curr. Org. Chem. 20 (2016) 2109–2116. [21] K. Goossens, K. Lava, C.W. Bielawski, K. Binnemans, Chem. Rev. 116 (2016) 4643–4807. [22] X.D. Hou, S.J. Liu, P.P. Zhou, J. Li, X. Liu, L.C. Wang, Y. Guo, J. Chromatogr. A 1456 (2016) 10–18. [23] J.L. Anderson, J. Ding, T. Welton, D.W. Armstrong, J. Am, Chem. Soc. 124 (2002) 14247–14254. [24] J.W.S. Hummers, R.E. Offeman, J. Am. Chem. Soc. 80 (1958) 1339. [25] Y.X. Sun, S. Han, J. Mol. Model. 19 (2013) 5407–5422. [26] Y. Suneetha, B.N. Kumar, Y. Harinath, D.H.K. Reddy, K. Seshaiah, Microchim. Acta. 176 (2012) 169–176. [27] T.T. Wang, Y.H. Chen, J.F. Ma, M.L. Chen, C.G. Nie, M.J. Hu, Y. Li, Z.J. Jia, J.H. Fang, H.Q. Gao, J. Chromatogr. A 2013 (2013) 63–72. [28] J.P. Ma, R.H. Xiao, J.H. Li, J.B. Yu, Y.P. Zhang, L.X. Chen, J. Chromatogr. A 1217 (2010) 5462–5469. [29] J.L. Benede, A. Chisvert, D.L. Giokas, A. Salvador, J. Chromatogr. A 1362 (2014) 25–33. [30] N. Wang, J.M. Wang, Y. Liao, S.J. Shao, Talanta 151 (2016) 1–7. [31] S.Y. Shi, J.F. Guo, Q.P. You, X.Q. Chen, Y.P. Zhang, Chem. Eng. J 243 (2014) 485–493. [32] J. Li, S.D. Jia, S.J. Yoon, S.J. Lee, S.W. Kwon, J. Lee, J. Food Compos Anal. 45 (2016) 73–79. [33] W. Setyaningsih, I.E. Saputro, M. Palma, C.G. Barroso, Food Chem. 169 (2015) 141–149. [34] R.P. Feliciano, E. Mecha, M.R. Bronze, A. Rodriguez-Mateos, J. Chromatogr. A 2016 (1464) 21–31. [35] X.P. Dai, D.S. Wang, H. Li, Y.Y. Chen, Z.C. Gong, H.Y. Xiang, S.Y. Shi, X.Q. Chen, J. Chromatogr. A 1484 (2017) 7–13.
Fig. 7. Chromatograms of the black wolfberry yogurt (a), urine (c) and samples spiked with 15 μg L−1 of the standard solution (b, d).
extractions without an obvious loss. The introduction of PIL not only affected the zeta potential of GO surface, but also offered electrostatic interactions to extract the acidic compound. Compared with other adsorbents (C18, GO@Sil, poly(VHIm+Br−)@Sil and poly(VHIm+PF6−) @Sil), the as-prepared poly(VHIm+PF6−)@GO@Sil showed higher extraction recovery for the polar phenolic acid. Under the optimized conditions, good linearity, low LODs and desirable spiked recoveries were obtained for the real sample. The results indicated that poly (VHIm+PF6−)@GO@Sil had a great potential as a polar adsorbent for acidic analytes. Acknowledgements This work was supported by the National Natural Science Foundation of China (21575149, 21505146 and 21575148); and Key Laboratory of Chemistry of Northwestern Plant Resources, Chinese Academy of Sciences, China (CNPR-2015kfkt-01). References [1] J. Dai, R.J. Mumper, Molecules 15 (2010) 7313–7352. [2] R.P. Feliciano, E. Mecha, M.R. Bronze, A. Rodriguez-Mateos, J. Chromatogr. A 2016 (1464) 21–31. [3] I.A. Ludwig, P. Mena, L. Calani, G. Borges, G. Pereira-Caro, L. Bresciani, D.D. Rio, E.J.L. Michael, C. Alan, Free Radic. Biol. Med. 89 (2015) 758–769.
129