Fe₃O₄-based composite magnetic nanoparticles for volatile compound sorption in the gas phase: determination of selenium(IV)

Abstract

In this study, Fe₃O₄-based composite magnetic nanoparticles were found to separate volatile compounds directly in the gas phase for the first time. The phenomenon of H₂Se sorption on the magnetic nanoparticles was studied in detail and applied for separation and preconcentration. The developed approach was applied for the determination of selenium in dietary supplement samples after microwave digestion by ETA-AAS as a proof-of-concept example.

---

Recently, magnetic nanoparticles (MNPs) have been widely used for separation and preconcentration of diverse compounds from various matrixes (biological samples,¹ foods,² environmental samples³etc.). As is known, magnetic materials as sorbents have several advantages over conventional sorbents for sorption in the aqueous phase. The separation and preconcentration processes can be performed directly in aqueous solution containing MNPs, and after sorption MNPs can easily be collected and separated from the liquid phase using an external magnetic field.⁴ Due to the large surface-area-to-volume ratios of nanometer (nm)-sized sorbents, MNPs are particularly useful for extracting and enriching a large volume of diverse compounds without tedious centrifugation. The compounds absorbed on MNPs can easily be eluted by a suitable eluent for subsequent determination.

It should be pointed out that sorption procedures based on MNPs were only implemented for the separation of compounds from liquid phase sample matrixes.^5–7

In this work, a novel phenomenon – gas sorption on magnetic Fe₃O₄-based nanoparticles held in a gas phase by an external magnetic field was established and coupled with hydride generation. To the best of our knowledge, this discovered phenomenon has not been presented in the literature.

To demonstrate the efficiency of the suggested approach, the proposed strategy was applied to determine Se(IV) as a proof-of-concept analyte in dietary supplement samples after microwave digestion using electrothermal atomization-atomic absorption spectroscopy (ETA-AAS). Selenium is an essential micronutrient for humans and animals,⁸ and it is widely used as a supplement ingredient.⁹ It acts as an active center of selenoenzymes and selenoproteins and plays an important role in energy metabolism and gene expression in organisms.¹⁰ Currently, dietary supplements containing selenium are widely consumed over the world for health-related reasons. Even though, most of these supplements are beneficial for human health, they can also have side effects due to an excess of one of the supplement ingredients including selenium. It is then important to ensure their efficacy and safety.¹¹

The most frequently used magnetic MNPs are the materials based on Fe₃O₄.¹² Modification of the Fe₃O₄ MNP surface can improve the selectivity of the sorption procedure. Fe₃O₄@AlOOH was chosen because it is known as an effective sorbent for ion separation.¹³ In this research, Fe₃O₄ and Fe₃O₄@AlOOH MNPs with different shell parameters (thickness and crystallinity) were synthesized and investigated as sorbents for H₂Se separation in the gas phase obtained by analyte hydride generation based on Se(IV) reduction in the presence of acid and NaBH₄. H₂Se formation is widely used for analyte separation and preconcentration for its subsequent sensitive determination by various instrumental methods.^14–16

Synthesis of Fe₃O₄-based MNPs was performed as follows.¹⁷ Firstly, 10.5 g of FeSO₄·7H₂O, 15 mL of 3.7 mol L⁻¹ FeCl₃ (Sigma-Aldrich, USA) in the molar ratio 1 : 2 and 360 mL of H₂O were mixed in a round flask using an overhead stirrer under an argon flow. Then, 198 mL of NH₄OH solution (3 mmol L⁻¹) was dropwise added, and the reaction mixture was kept for 15 min. The resulting product was separated by centrifugation (Sigma 2-16P, 9000g, 5 min) and washed with distilled water six times followed by freeze-drying for 3 h.

To obtain composite Fe₃O₄@AlOOH nanoparticles 0.1 g of Fe₃O₄ MNPs was placed into Al(NO₃)₃ solution (75 mL, 0.1–4.3 g L⁻¹). Then, the pH value was adjusted to 4 by adding NaOH solution (6 mol L⁻¹). After that, the white-black suspension obtained was transferred into a 200 mL Teflon lined autoclave (SPBU, Russia), sealed and heated at 180 °C for 2 h. After cooling, the resulting product was separated by centrifugation (Sigma 2-16P, 8000g, 3 min) and washed with distilled water and ethanol four times followed by freeze-drying for 3 h.

All studied MNPs were stable during the storage under ambient conditions.

A microcolumn (5 mg of MNPs in a pipette tip (100 μL volume)) equipped with an external magnet was hermetically connected with a mixing chamber through a PTFE tube (Fig. 1). At the beginning 5 mL of 0.5 mg L⁻¹ Se(IV) solution (in 2 mol L⁻¹ HNO₃) was placed into the mixing chamber. Then, 3 mL of 3% (w/v) NaBH₄ (in 0.1 mol L⁻¹ NaOH) was delivered into the mixing chamber using a peristaltic pump over 6 min. Formation of H₂Se and H₂ was observed. H₂Se was collected on MNPs when the gas mixture was delivered through the microcolumn with MNPs held by an external magnetic field. Then, the microcolumn was disconnected and MNPs were transferred to a PTFE vial (5 mL). For analyte elution 1 mL of H₂O₂ (30%) was added to MNPs, and a sonication using an ultrasonic homogenizer (Bandelin Sonoplus HD2070, MS 72, Germany, power – 30%) for 10 min at 30% amplitude was carried out. After that, the vial was placed on the magnet (permanent neodymium iron boron magnet) to precipitate magnetic particles and the clear supernatant was analyzed by ETA-AAS (a Shimadzu Model AA-7000 atomic absorption spectrometer). MNPs were used once, since the particle structure was irreparably changed after sonication used for analyte elution.

Fig. 1 Schematic representation of H₂Se generation and sorption on the MNPs followed by selenium determination.

MNPs synthesized were characterized by XRD (ESI Fig. 1†), TEM (ESI Fig. 2 and 3†), SSA estimation, FTIR (ESI Fig. 4†) and VSM (ESI Fig. 5†). The data obtained were in good agreement with each other and indicated the presence of AlOOH shells with different parameters on the magnetic core. The observed decrease of a shell thickness (ESI Table 1†) with the increase of AlOOH amount up to 33 mol% (Fe₃O₄@33AlOOH) may indicate a gradual process of its crystallization on the Fe₃O₄ surface because a more crystalline structure has a higher packing density than the amorphous one. A further increase of AlOOH amount led to the appearance of more intensive peaks from OH groups in the 3400–3500 cm⁻¹ region (ESI Fig. 4†), which indicated the presence of a hydroxide amorphous phase.

To be effectively held in the microcolumn MNPs should have good magnetic properties. The magnetic characteristics of Fe₃O₄-based composite magnetic nanoparticles can be decreased because of the presence of a non-magnetic shell. According to the data obtained (ESI Table 1†) the decrease in the value of the saturation magnetization (M_S) for the Fe₃O₄@AlOOH MNPs was not critical (not more than 15%) in comparison with Fe₃O₄ MNPs.

The evaluation of magnetic interaction between the nanoparticles of bare and Fe₃O₄@AlOOH types in the gas and aqueous phases was also performed. We supposed that since the gas sorption and elution processes were carried out in the external magnetic field, they probably could be determined to a considerable degree by the magnetic interactions between the MNPs. For the gas phase the energy values (E) of interaction between MNPs under the external magnet field were calculated according to a handbook.¹⁸ MNPs with the lowest energy should be more effectively held in the external magnetic field. For the aqueous phase the velocity values (V) of nanoparticle movement under the external magnet field were also calculated according to a handbook.¹⁹ The V values determine the efficiency of MNP collection during the elution process. As expected both the parameters (ESI Table 1†) depended on the shell thickness and will be discussed below.

Seven types of MNPs (Fe₃O₄, Fe₃O₄@5AlOOH, Fe₃O₄@10AlOOH, Fe₃O₄@20AlOOH, Fe₃O₄@33AlOOH, Fe₃O₄@50AlOOH and Fe₃O₄@60AlOOH) were investigated for H₂Se sorption. To choose the optimal sorbent the efficiency of sorption and elution was studied. The sorption efficiency was calculated using the following equation:

where Q_st is the quantity of the analyte in Se(IV) standard solution and Q_ads is the quantity of the analyte determined in absorption solution.

The elution efficiency was calculated using the following equation:

where Q_el is the quantity of the analyte determined in the eluate.

After H₂Se sorption and its elution the concentration of Se was determined in the eluate and absorption solution by ETA-AAS.

The obtained experimental and calculated results (Fig. 2) indicated that the maximum sorption efficiency and the minimum interaction energy correspondingly were observed for bare Fe₃O₄. Both these values showed the same tendencies depending on the shell thickness for the Fe₃O₄-based nanoparticles under study. Note that the maximum difference between the experimental and calculated results and comparable to bare Fe₃O₄ sorption efficiency were observed for Fe₃O₄@10AlOOH MNPs with high SSA and shell thickness values. Thus, for composite MNPs not only magnetic interaction but also the surface composition influences the sorption properties.

Fig. 2 Sorption efficiency (left) and interaction energy values (right) in the gas phase under an external magnetic field for the MNPs synthesized (5 mg of MNPs).

Moreover, we can assume that the efficiency of sorbent and analyte binding in the gas phase was determined by two processes: magnetic interactions between MNPs and the direct adsorption process. In the case of Fe₃O₄-based MNPs, they were evenly distributed under the magnetic field on the inner surface of the microcolumn and formed a layer that interacted with the gas phase. The higher interaction energy of Fe₃O₄ MNPs led to the formation of more elongated needles, which significantly increased the surface area available for gas molecule interaction.

To understand the sorption mechanisms of H₂Se on the Fe₃O₄ MNP surface, an adsorption isotherm was plotted. Two widely used models based on Langmuir and Freundlich isotherm equations were applied. The linearization results indicated that in our case the adsorption process was polymolecular, the Freundlich model represented good agreement with the data obtained (adjusted R-square value for linear least squares fitting was equal to 0.998 (ESI Fig. 7†)). Thus, it can be concluded that H₂Se adsorption on Fe₃O₄ MNPs was characterized by the sorbent surface heterogeneity because the Freundlich isotherm model represents heterogeneity of solid surfaces.²⁰ The n value (0.947) obtained from the Freundlich model was smaller than 1. Thus, a nonlinear sorption process takes place.

The flow rate of NaBH₄ solution injected in a sample solution influenced the flow rate of the gas phase generated. The flow rate of NaBH₄ solution (3%) in the range of 0.2–1 mL min⁻¹ was varied, and the gas flow rate was measured using a gas flow meter. It was found that the NaBH₄ solution flow rate of 0.5 mL min⁻¹ provided optimal conditions for gas phase generation (gas flow rate 10 mL min⁻¹) and H₂Se sorption (sorption efficiency 95 ± 5%). At a higher reagent flow rate, excessive hydrogen generation was observed, which caused solution spraying and transferring of droplets of the aqueous phase into the microcolumn. In this case, the MNPs were wetted and H₂Se sorption was not observed.

As can be seen from Fig. 3 the optimal elution efficiency and the maximum velocity as expected were observed for bare Fe₃O₄. Similar to that discussed above the experimental and calculated data demonstrated the same behavior. But in the case of a high AlOOH amount the elution efficiency values became lower.

Fig. 3 Elution efficiency (left) and average velocity in the aqueous phase under an external magnetic field (right) for the MNPs synthesized (5 mg of MNPs).

Thus, Fe₃O₄ MNPs provided maximum sorption of H₂Se and its satisfactory elution, therefore Fe₃O₄ MNPs were chosen for future experiments. The amount of MNPs is an important factor affecting the sorption efficiency and reproducibility. The effect of MNP (bare Fe₃O₄) amount was investigated in the range from 2.5 to 10 mg. The optimal sorbent amount was determined as 5 mg (ESI Fig. 6†). At a lower sorbent amount, it was probably not enough to completely adsorb H₂Se from the gas phase. At a higher sorbent amount its surface area was less because the MNPs were tightly packed into the microcolumn resulting in a lower sorption efficiency. In addition, the low elution efficiency values at higher sorbent masses could be due to the shift of Se adsorption/desorption equilibrium in the aqueous phase during the sorption process.

In order to find the optimal conditions for H₂Se formation, the concentrations of NaBH₄ and HNO₃ were varied from 1 to 5% and from 1 to 5 mol L⁻¹, respectively. It was established, that the highest analytical signal was achieved using 3% NaBH₄ and 2 mol L⁻¹ HNO₃.

Taking into account the need to convert H₂Se to a water-soluble form for its elution from MNPs as well as the features of pyrolytic graphite tubes and the properties of MNPs, oxidants such as KMnO₄ and H₂O₂ were studied as components of elution solution. Initially, a solution of 10⁻⁵ mol L⁻¹ KMnO₄ was investigated. This concentration of KMnO₄ was chosen to eliminate background interferences taking place during the ETA-AAS analysis and to exclude its negative effect on the pyrolytic graphite tube. The results obtained showed a low elution efficiency (30%). Thus, another oxidant – H₂O₂ – was studied in the concentration range of 5–30%. It was established that the maximum absorbance peak area was observed by using 30% H₂O₂, thus, this eluent was chosen for further experiments. The elution efficiency was 2.5 times higher compared to the one with KMnO₄ solution.

It is well known that sonication is widely used to activate the MNP surface.^21,22 Ultrasonic irradiation can improve the mass transfer between two immiscible phases and reduce the equilibrium time, and thus can be helpful for the elution process. It was established that the use of ultrasonic irradiation enhanced analytical signals at two different times. To choose the optimal sonication time, the elution process was implemented at 1, 5, 10, and 15 min of sonication (325 W, 35 kHz). The maximal absorbance peak area was observed at 10 min of sonication and no improvement was found at 15 min. Thus, 10 min of sonication was chosen for analyte elution from MNPs.

The effect of potentially interfering volatile compounds (AsH₃ and H₂S) on H₂Se sorption and selenium determination was also investigated. It was done by addition of known concentrations of KAsO₂ and Na₂S to 50 μg L⁻¹ selenium(IV) solution. The tolerable concentration of each taken foreign compound is considered to be less than 5% of relative error in the signal. It was shown that AsH₃ and H₂S generated did not interfere with H₂Se sorption and selenium determination up to the highest studied concentration of As(III) and S²⁻ (1 mg L⁻¹).

The developed procedure was applied for the determination of Se(IV) in three types of dietary supplements. The dietary supplements were complex matrices, which included selenoxanthen (sample no. 1), sodium selenite (samples no. 2, 3) as biologically active substances, and different other components (cellulose, dextrose, starch), anti-caking agents (magnesium stearate, silicon dioxide, talc), some vitamins ((2R)-2,7,8-trimethyl-2-[(4R,8R)-4,8,12-trimethyltridecyl]-6-chromanol (vitamin E) and L-ascorbic acid (vitamin C)), and salts/oxides of zinc, iron, and calcium. To destroy complex sample matrices and obtain selenium in the form of Se(IV) all samples were digested in a single run using a laboratory microwave digestor MDS-10 Master (Sineo, China).²³ 0.15 ± 0.5 g of powdered samples were placed in individual microwave reactors. The aliquots were treated with 6 mL concentrated HNO₃ and 1 mL H₂O₂. The temperature program is given in ESI Table 2.† The employed microwave power was up to 600 W.

Generally, the obtained results corresponded to the contents indicated on each package (Table 1). Accuracy and reliability of the resulting information were further studied by the add/found method. The procedure developed provides selenium separation from digested samples with recovery from 87 to 102%.

Table 1 The results of determination of selenium in dietary supplements (n = 3, P = 0.95)

Sample no.	Labeled value, μg	Added, μg	Found, μg	Recovery, %
1 (selenoxanthen)	50	0	45 ± 8	—
		50	90 ± 12	90
		100	147 ± 21	102
2 (sodium selenite)	50	0	54 ± 7	—
		50	101 ± 10	94
		100	146 ± 17	92
3 (sodium selenite)	40–60	0	52 ± 8	—
		50	99 ± 11	94
		100	139 ± 19	87

---

Moreover, the certified reference material SRM 7340-96 (Se(IV) solution, Ecohim, Russia) was used to validate the developed method. In this case the t-test was applied in order to establish the statistical significance of the results. The results showed no differences, which were found at a confidence level of 95%, between the certified value of 500 ± 5 μg Se(IV) L⁻¹ and the obtained value of 500 ± 5 μg Se(IV) L⁻¹ (n = 3). Thus, this result was in agreement with the certified value (t_exp < t_cr), where t critical and expected values were equal to 2.78 and 1.34, respectively.

The calibration plot was constructed using Se(IV) solutions under developed procedure conditions. The linear calibration range of 1–20 μg L⁻¹ selenium was obtained with a correlation coefficient of 0.998. The LOD, calculated from a blank test, based on 3σ, was found to be 0.3 μg L⁻¹. The LOQ, equal to 1 μg L⁻¹ was calculated that produced a peak with 10 times the signal-to-noise ratio. The obtained relative standard deviation values for 1 and 20 μg L⁻¹ of Se(IV) solutions were in the range from 4 to 8% (within-day precision) and from 6 to 9% (day-to-day precision), respectively.

In conclusion, Fe₃O₄-based MNPs were investigated for volatile H₂Se sorption directly in the gas phase for the first time. In this study sorption of H₂Se on Fe₃O₄-based MNPs was coupled with hydride generation and applied for selenium(IV) determination in dietary supplement samples by ETA-AAS. The proposed approach has the following advantages. First, Fe₃O₄-based MNPs can be easily held by an external magnetic field in the gas phase resulting in analyte sorption in the gas phase flow. Second, compared with conventional sorbents Fe₃O₄-based MNPs allowed the implementation of analyte elution without centrifugation and filtration. Thirdly, the Fe₃O₄-based magnetic nanoparticles can be easily prepared, and MNPs are inexpensive and readily available. The application of this method for other volatile compounds is under investigation.

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

The combined method was developed and supported by the Russian Scientific Foundation (project no. 16-13-10117). Irina Timofeeva gratefully acknowledges the Russian Foundation for Basic Research (project no. 16-33-60126 mol_a_dk) for the financial support in the on-line sample preparation investigations. Mikhail Osmolowsky acknowledges the Russian Scientific Foundation (project no. 18-03-01066) for the financial support in sorbent synthesis and characterization. Scientific research was performed using the equipment of the Research Park of St Petersburg State University (Centre for X-ray Diffraction Studies, Chemical Analysis and Materials Research Centre, Centre for Innovative Technologies of Composite Nanomaterials).

Notes and references

1. B. Hashemi, P. Zohrabi and M. Shamsipur, Talanta, 2018, 187, 337–347.
2. J. González-Sálamo, B. Socas-Rodríguez, J. Hernández-Borges and M. A. Rodríguez-Delgado, TrAC, Trends Anal. Chem., 2016, 85, 203–220.
3. J. Li, Y. Wang, K. Li, Y. Cao, S. Wu and L. Wu, TrAC, Trends Anal. Chem., 2015, 72, 141–152.
4. J. Meng, J. Bu, C. Deng and X. Zhang, J. Chromatogr. A, 2011, 1218, 1585–1591.
5. C. Vakh, M. Alaboud, S. Lebedinets, D. Korolev, V. Postnov, L. Moskvin, O. Osmolovskaya and A. Bulatov, Anal. Chim. Acta, 2018, 1001, 59–69.
6. F. Maya, C. P. Cabello, J. M. Estela, V. Cerdà and G. T. Palomino, Anal. Chem., 2015, 87, 7545–7549.
7. Q. Gao, H. B. Zheng, D. Luo, J. Ding and Y. Q. Feng, Anal. Chim. Acta, 2012, 720, 57–62.
8. T. Ferri and De L. L. Ticconi, Anal. Lett., 2001, 34, 975–988.
9. A. Rocco, E. Donati, E. Touloupakis and Z. Aturki, TrAC, Trends Anal. Chem., 2018, 103, 156–183.
10. F. Zhou, W. Yang, M. Wang, Y. Miao, Z. Cui, Z. Li and D. Liang, Food Chem., 2018, 265, 182–188.
11. A. Rocco, E. Donati, E. Touloupakis and Z. Aturki, TrAC, Trends Anal. Chem., 2018, 103, 156–183.
12. H. Hu, L. Yang, Z. Lin, X. Xiang, X. Jiang and L. Hou, Int. J. Biol. Macromol., 2018, 114, 256–262.
13. R. Massart, IEEE Trans. Magn., 1981, 17, 1247–1248.
14. A. Shishov, M. Wieczorek, P. Kościelniak, D. Dudek-Adamska, A. Telk, L. Moskvin and A. Bulatov, Talanta, 2018, 181, 359–365.
15. X. Tang, Z. Xu and J. Wang, Spectrochim. Acta, Part B, 2005, 60, 1580–1585.
16. N. Zhang, G. Sun and H. Ma, Miner. Eng., 2007, 20, 1397–1400.
17. A. A. Hernández-Hernández, G. A. Álvarez-Romero, E. Contreras Lopez, K. AguilarArteaga and A. Castañeda-Ovando, Food Anal. Methods, 2017, 10, 2974–2993.
18. A. Stone, The Theory of Intermolecular Forces, Oxford University Pres, 2nd edn, 2016.
19. L. D. Landau and E. M. Lifshitz, Fluid Mechanics: Volume 6 (Course of Theoretical Physics), Butterworth-Heinemann, 2nd edn, 1987.
20. L. Yin, S. Song, X. Wang, F. Niu, R. Ma, S. Yu, T. Wen, Y. Chen, T. Hayat, A. Alsaedi and X. Wang, Environ. Pollut., 2018, 238, 725–738.
21. N. Jaafarzadeh, A. Takdastan, S. Jorfi, F. Ghanbari, M. Ahmadi and G. J. Barzegar, Mol. Liq., 2018, 256, 462–470.
22. A. Keramat and R. Zare-Dorabei, Ultrason. Sonochem., 2017, 38, 421–429.
23. Q.-X. Zhao, Y.-W. Chen, N. Belzile and M. Wang, Anal. Chim. Acta, 2010, 665, 123–128.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c8an01894d

---