Introduction
The use of hydrogen as an energy source is expected increase because of its environmentally friendly nature. Of the methods for generating hydrogen, the photocatalytic splitting of water using TiO2 1,2 and MTiO3 3,4 semiconductors has attracted considerable attention. The principle of photocatalytic water splitting is based on the conversion of light energy to electricity upon the exposure of a semiconductor to light. Upon exposure to incident light, the electrons of n-type semiconducting materials are emitted from the conduction band to the valence bands, leading to the holes in valence bands. These electrons and holes split water molecules into oxygen and hydrogen.5 The band gap required for hydrogen production by water splitting need to be at least 1.2 eV. Pure TiO2 and various MTiO3 photocatalysts (where M = a metal) have band gaps of 3.0-5.0 eV, making them ineffective in this reaction. In addition, hydrogen production is limited by the rapid recombination of holes and electrons.6 To overcome this rapid recombination, considerable efforts have been made to increase hydrogen evolution using methanol, ethanol, or a mixture of light alcohol and water, rather than water because lower energies of approximately 0.6-0.8 eV are needed.7,8
One issue in a liquid photocatalytic reaction is how to recover the catalyst from a liquid solution after the reaction. This paper introduces a core@shell multi-component catalyst, which has attracted considerable attention recently because of its potential applications in electronics, magnetism, optics, and catalysis.9-13 The magnetic core used can be collected conveniently and separated using a magnet. Therefore, magnetic-based TiO2 photocatalyst synthesis will be very useful. On the other hand, it is difficult to produce TiO2-coated particles with UV or visible-light photoactivity without sacrificing the magnetic properties. Some studies evaluated the use of magnetite (Fe3O4) cores.14-17 Recently, more stable transition-metal incorporated MFe2O4 (M = Ni, Co, Fe, Sr) super-paramagnetic materials have been used as core materials and have produced good results for the photocatalytic destruction of organic compounds.18-21 On the other hand, oxygenated cores are difficult to activate under visible-light, so materials containing sulfur or nitrogen are used. Therefore, the perfect core@shell structured MFeS2@TiO2 magnetic photocatalysts require further development. Furthermore, higher photocatalytic activity can be expected when a metal with higher reduction potentials, Cu or Ag, is inserted into MFeS2 core, but comparatively little developmental work has been conducted on redox applications.
The major objective of this study was to develop core@shell-structured CuFeS2@TiO2 photocatalysts with improved catalytic properties for the production of hydrogen from methanol/water aqueous systems, and with magnetic properties commensurate with the separations. The nature of the CuFeS2@TiO2 photocatalyst produced were examined by X-ray diffraction (XRD), UV-visible spectroscopy, photoluminescence (PL) spectroscopy, Brunauer-Emmett-Teller (BET) surface area measurements, and magnetic momentum using a vibrating sample magnetometer.
Experimental
Core@shell structured CuFeS2@TiO2 was prepared using sequential solvothermal and impregnation hybrid methods, as shown in Figure 1. In step 1 of a), to prepare the CuFeS2 sol mixture, iron chloride (FeCl3·6H2O, 99.95%, Junsei Chemical, Japan) and cupper chloride (CuCl·5H2O, 99.95%, Junsei Chemical, Japan) were used as the Fe and Cu precursors, respectively, and ethanol was used as the solvent. Briefly, 0.04 mol of FeCl3 and 0.04 mol of CuCl2 were added to 400 mL of ethanol, and stirred for 1 h.
Figure 1.Synthetic sequence for the TiO2, CuFeS2, and CuFeS2@ TiO2 photocatalyst.
Thioacetamide (C2H5NS) was added to the upper solution as a sulfur source. Thioacetamide is used widely in classical qualitative inorganic analysis as an in situ source of sulfide ions. Therefore, the treatment of aqueous solutions of many metal cations to a solution of thioacetamide affords the corresponding metal sulfide:M++M3+ + 2CH3C(S)NH2 + H2O→ MMS2 + 2CH3C(O)NH2 + 4 H+
In the next step, ethylenediamine was added to reduce Fe3+ to Fe2.5+. The final solution was stirred until it became homogeneous, and was then treated thermally in an autoclave at 200 ℃ for 2 h. After the thermal treatment, the obtained powder was cooled, washed with ethanol, and dried at 80 ℃ for 24 h. Here, pure TiO2 was also synthesized using a solvothermal method, which is the same process (b) to CuFeS2.22 Titanium tetraisopropoxide (TTIP, Junsei Chemical, Japan) was used as the Ti precursor. The solution in the synthesis step was fixed to pH = 3, and the thermal treatment was performed in an autoclave at 200 ℃ for 8 h. In step 3 of c), the magnetic CuFeS2 particles were coated with TiO2. The CuFeS2 and TiO2 with a 1.0 molar ratio were added to 100 mL of ethanol and its pH was fixed to 5 by HCl. The resulting colloid was stirred for 3 h. After stirring the solution until it became homogeneous, the colloid was evaporated at 70 ℃ for 3 h. The resulting precipitate was heated to 200 ℃ for 2 h under argon to remove the solvent and generate an anatase shell.
The TiO2, CuFeS2, and CuFeS2@TiO2 (analyzed as approximately CuFeS2:TiO2= 1:1) powders were examined by XRD (X'Pert Pro MPD PANalytical 2-circle diffractometer) using nickel-filtered CuKα radiation (30 kV, 30 mA) at 2θ angles from 5 to 70°, a scan speed of 10° min–1, and a time constant of 1 s. The sizes and shapes of the TiO2, CuFeS2, and 1CuFeS2@1TiO2 particles were determined by transmission electron microscopy (TEM, JEOL 2000EX) at 200 kV. The specific surface area and pore size distribution were calculated according to the BET theory, which gives the isotherm equation for multilayer adsorption by the generalization of Langmuir’s treatment of a multi-molecular layer. The adsorptiondesorption isotherm analysis to identify the BET surface area and pore size distribution of particles was performed using a Belsorp II instrument. All the particles were degassed under vacuum at 150 ℃ for 2 h before the measurements, and measured by nitrogen gas adsorption using a continuous flow method with a mixture of nitrogen and helium as the carrier gas. The UV-visible spectra of the TiO2, CuFeS2, and 1CuFeS2@1TiO2 powders were obtained using a Shimadzu MPS-2000 spectrometer (Kyoto, Japan) equipped with a reflectance sphere over the range, 200 to 800 nm. The magnetic properties were measured using a vibrating sample magnetometer (VSM) on a physical property measurement system (Quantum Design PPMS-9). Photoluminescence (PL) spectroscopy was carried out on the three powders to examine photo-excited electron hole pairs using 1.0 mm thick pellets of TiO2, CuFeS2, and 1CuFeS2@1TiO2 at room temperature using a He-Cd laser source at a wavelength of 325 nm.
The photosplitting of methanol/water was performed using a liquid photo reactor designed in this laboratory.23 To photosplit methanol/water (1:1 vol./vol., total volume = 1.0 L) solution, 0.5 g of powdered TiO2, CuFeS2, and 1CuFeS2@1TiO2 photocatalysts were added to 1.0 L of a methanol/water solution in a 2.0-L Pyrex reactor. UV-lamps (6 × 3 Wcm−2 = 18 Wcm−2, 30 cm length × 2.0 cm diameter; Shinan, Korea) emitting at 365 nm were used. The photosplitting of methanol/water was carried out over 1-8 h with stirring, and hydrogen evolution was measured after 1 h of operation. The hydrogen gas (H2) produced during methanol/water photosplitting was analyzed by TCD-type gas chromatography (GC; model DS 6200; Donam Instruments Inc., Korea). To identify the products and intermediates, GC was connected directly to a photo-reactor. The following GC conditions were used: TCD detector; Carbosphere column (Alltech, Deerfield, IL, USA); injection temperature 140 ℃; initial temperature 120 ℃; final temperature 120 ℃; and detector temperature 150 ℃.
Results and Discussion
Figure 2(a) and (b) shows XRD patterns and TEM images of the TiO2, CuFeS2, and 1CuFeS2@1TiO2 powders. First in A), TiO2 particles had a pure anatase structure with peaks at 25.3, 38.0, 48.2, 54, 63, and 68° 2θ, which were assigned to the (101), (004), (200), (105), (211), and (204) planes, respectively.24 CuFeS2 showed several peaks at 2θ values of 29.45 (d112), 34.19° (d200), 48.93° (d220), 57.95° (d321), 71.40° (d400), 79.07° (d332), and 90.99° (d415), indicating a tetragonal crystal system (I-42d).25 1CuFeS2@1TiO2 showed a mixed peaks of CuFeS2 and TiO2, indicating the exposed CuFeS2 or partially-covered TiO2 on the surfaces of CuFeS2 particles. On the other hand, peak broadening indicates a reduction in the crystallite size.26 The Debye-Scherrer’s equation, t = 0.9λ/βcosθ (where λ is the wavelength of incident X-rays, β is the full width at half maximum height in radians, and the θ is the diffraction angle) was used to determine the crystallite size.27 The calculated values at the representative 101 and 112 planes of TiO2 and CuFeS2 were 7.3 nm and 38.38 nm, respectively. The TEM images in Figure 2(b) showed slightly distorted tetragonal CuFeS2 particles, 20 nm in size. The 1CuFeS2@1TiO2 particles were partially covered by 5- 10 nm TiO2 particles, and the particle size increased to 100 nm after core@shell formation compared to the size of pure CuFeS2.
Figure 2.XRD patterns (a) and TEM photographs (b) of synthesized TiO2, CuFeS2, and CuFeS2@TiO2 powders.
Figure 3 shows the adsorption-desorption isotherm curves of N2 at 77 K for the TiO2, CuFeS2, and 1CuFeS2@1TiO2 powders. The isotherms of TiO2 and 1CuFeS2@1TiO2 belonged to IV type in the IUPAC classification28; this hysteresis slope has been observed in the presence of mesopores. The mesopores are considered to be bulk pores formed between the TiO2 particles. Otherwise, the isotherms of CuFeS2 mean non-pores with an III type. This suggests that a high surface area facilitates adsorption, which can generate adsorption activity. Therefore, some molecules are adsorbed more easily on the surfaces of the TiO2 and core@shell structured 1CuS@1TiO2 than CuFeS2. On the other hand, the surface area, pore volume, and pore size are listed in the table below. The specific surface areas of TiO2 and 1CuFeS2@1TiO2 were large, 154.21 and 73.70 m2g−1, respectively. On the other hand, it decreased in CuFeS2 to 24.50 m2g−1. Here, the specific surface area in this study might depend on the bulk pores formed by aggregation between the TiO2 particles in the shell, and the pore volumes showed the same tendency to the surface areas.
Figure 3.Adsorption-desorption isotherm curves of N2 at 77 K for the TiO2, CuFeS2, and 1CuFeS2@1TiO2 powders.
The UV-visible diffuse reflectance (a) and PL spectra (b) of TiO2 and CuFeS2 powders were also obtained (Figure 4). The absorption band corresponding to the octahedral symmetry of Ti4+ was observed at ~350 to 380 nm, indicating a bandgap of 2.90 eV.29 Generally, the band gaps of semiconductor materials are closely related to the absorption wavelength, where a higher wavelength indicates a smaller band gap. CuFeS2 exhibited a continuous absorption band in the range 200-700 nm, which concurs with its black color. Using Tauc’s equation,30 its band-gap was estimated to be approximately 0.75 eV. Figure 4(b) presents the photoluminescence (PL) spectra of TiO2, CuFeS2, and 1CuFeS2@1TiO2 pellets. The PL emission spectra are useful for examining the efficiency of the charge transfer behavior of photo-generated electrons and holes. The PL curves show that the electrons in the valence band are transferred to the conduction band, and then return to the valence band by photoemission. In general, the PL intensity increases with increasing number of photons emitted as a result of the recombination of electrons and holes, resulting in a decrease in photoactivity.31 Therefore, there is a strong relationship between the PL intensity and photoactivity. The PL spectrum of 1CuFeS2@1TiO2 showed that the PL intensity of TiO2 was quenched substantially by the CuFeS2 magnetic core. The CuFeS2 in 1CuFeS2@1TiO2 captures photo-generated electrons from the TiO2 conduction band, in particular Cu ions, which separates the photogenerated electron-hole pairs. The presence of Cu in the CuFeS2 magnetic core reduces the recombination rate, and reduces the PL spectrum intensity, indicating that the PL intensity depends on electron capture by Cu ions.
Figure 4.UV-visible diffuse reflectance spectra (a) and photoluminescence (PL) (b) spectra of TiO2, CuFeS2, and CuFeS2@TiO2 powders.
Figure 5 shows the saturation magnetization (emu/g) versus coercivity (Hc) plots obtained at room temperature for CuFeS2 and CuFeS2@TiO2, and the inset shows an enlarged figure at an almost zero applied magnetic field. The two samples have hysteresis cycles that are characteristic of ferromagnetism. The saturation magnetizations of CuFeS2 and CuFeS2@TiO2 were similar but the value of CuFeS2 was approximately two times larger than that of CuFeS2@TiO2. This suggests that the TiO2 shell was responsible for the observed reduction in ferromagnetism. The TiO2 anatase structure is diamagnetic.32,33 Occasionally, some Ti4+ ions become Ti3+ ions if defects are present in the TiO2 structure, resulting in the detection of weak ferromagnetism.34 Therefore, the ferromagnetic effect of TiO2 depends on the method of synthesis, due to the presence of defects, particularly oxygen vacancies in TiO2 nanomaterials. In addition, an advantage of CuFeS2@TiO2 is that it can be recovered by a magnet after the reaction.
Figure 5.Saturation magnetization (M) versus coercivity (Hc) plots obtained at room temperature for CuFeS2 and CuFeS2@ TiO2.
The CV results for TiO2 and CuFeS2 (Figure 6) were strongly dependent on the analytical conditions used, and a semiconductor needs to be redox active within the experimental potential window. In this study, the potentials were measured in distilled water using a pelletized sample as the working electrode, Ag/AgCl as the reference electrode, and 0.1 M KCl as the supporting electrolyte. A reversible wave was observed, which gives the following information: reversible reactions display a hysteresis of the absolute potential between the reduction (Epc) and oxidation (Epa) peaks. Reversible reactions show the ratios of the peak voltages under reduction and oxidation conditions that are near unity (1 = Epa/Epc). When such reversible peaks are observed, the thermodynamic information in the form of half-cell potentials, E0 1/2 (Epc+ Epa/2) can be determined. In particular, when the waves are semi-reversible, such as when Epa/Epc is
Figure 6.CV curves of TiO2 and CuFeS2 powders.
The evolutions of H2 from the photo-splitting of methanol/ water over TiO2, CuFeS2, and CuFeS2@TiO2 powders at various molar ratios were measured in a batch-type liquid photo system and are presented in Figure 7. No H2 was collected from the photodecomposition of methanol/water over pure anatase CuFeS2 after 10 h, whereas 1.80 mmol of H2 was collected over 1CuFeS2@0.05TiO2. This was attributed to excited electron capture by the CuFeS2 core and a reduced recombination rate. Otherwise, evolved oxygen and carbon dioxide gases could not be observed because they were partially transformed into some by-products, such as formaldehyde, acetaldehyde, formic acid, and acetic acid, by oxidation reactions with methanol molecules. On the other hand, the expected mechanism of charge separation and the photocatalytic process of CuFeS2 and TiO2 are shown in the adjoining diagram based on the UV-visible and CV spectra results. The conduction band position in the CuFeS2 core is at a lower energy than that in the TiO2 shell. Therefore, the core could act as a sink for photo-generated electrons. Under UV-light irradiation, the excited electrons from the valence band of the TiO2 shell move into its conduction band and flowed to the surface of the shell, and they fell into the conduction band of the CuFeS2 magnetic core. XPS confirmed that the Cu ions in CuFeS2 were reduced after the photoreaction, meaning that the core attracts electrons during the photoreaction. Therefore, Cu+ ions (3d9) can accept electrons, and the electrons that move from the TiO2 shell into CuFeS2 core can be located in the 3d orbital of Cu ions.
Figure 7.Evolution of H2 from the photo-splitting of methanol/water over TiO2, CuFeS2 and CuFeS2@TiO2 powders along with the expected mechanism of charge separation and the photocatalytic process of CuFeS2@TiO2.
Conclusion
Core@shell-structured CuFeS2@TiO2 was synthesized for the production of H2 gas from the photodecomposition of methanol/water in a batch-type liquid photo system. The core@shell morphology of synthesized CuFeS2@TiO2 was examined by XRD and TEM. The PL intensity of CuFeS2@ TiO2, indicating electron/hole recombination, decreased. The ferromagnetic property of CuFeS2 was decreased slightly by the presence of a TiO2 shell. Hydrogen production from methanol/water was remarkably higher for 1CuFeS2@ 0.05TiO2 than for pure TiO2, and 1.80 mmol of H2 was collected after 10 h when 0.5 gL−1 of the powder was used. These results suggest that hydrogen production by methanol/ water splitting can be achieved more readily over the CuFeS2@TiO2 magnetic material. The significant enhancement in photoactivity for methanol/water mixture splitting was attributed to the synergism between CuFeS2 and TiO2, i.e., to effective charge transfer from TiO2 to CuFeS2 and the suppression of electron/hole pair recombination.
References
- Lin, Y. C.; Liu, S. H.; Syu, H. R.; Ho, T. H. Spectrochim. Acta A Mol. Biomol. Spectrosc. 2012, 95, 300. https://doi.org/10.1016/j.saa.2012.03.080
- Liu, S. H.; Syu, H. R. Appl. Energ. 2012, 100, 148. https://doi.org/10.1016/j.apenergy.2012.03.063
- Khemthong, P.; Photai, P.; Grisdanurak, N. Int. J. Hydrogen Energy 2013, 38, 15992. https://doi.org/10.1016/j.ijhydene.2013.10.065
- Jin, Z.; Zhang, X.; Li, Y.; Li, S.; Lu, G. Catal. Commun. 2007, 8, 1267. https://doi.org/10.1016/j.catcom.2006.11.019
- Zoua, Z.; Ye, J.; Arakawa, H. Int. J. Hydrogen Energy 2003, 28, 663. https://doi.org/10.1016/S0360-3199(02)00159-3
- Mizoguchi, H.; Ueda, K.; Orita, M.; Moon, S. C.; Kajihara, K.; Hirano, M.; Hosono, H. Mater. Res. Bull. 2002, 37, 2401. https://doi.org/10.1016/S0025-5408(02)00974-1
- Kandiel, T. A.; Dillert, R.; Robben, L.; Bahnemann, D. W. Catal. Today 2011, 161, 196. https://doi.org/10.1016/j.cattod.2010.08.012
- Ruiza, S. O.; Zanella, R.; Lopezb, R.; Gordillo, A. H.; Gomez, R. J. Hazard. Mater. 2013, 263, 2. https://doi.org/10.1016/j.jhazmat.2013.03.057
- Fu, W.; Yang, H.; Li, M.; Chang, L.; Yu, Q.; Xu, J.; Zou, G. Mater. Lett. 2006, 60, 2723. https://doi.org/10.1016/j.matlet.2006.01.078
- Li, C. J.; Wang, J. N.; Wang, B.; Gong, J. R.; Lin, Z. Mater. Res. Bull. 2012, 47, 333. https://doi.org/10.1016/j.materresbull.2011.11.012
- Wilson, A.; Mishra, S. R.; Gupta, R.; Ghosh, K. J. Magn. Magn. Mater. 2012, 324, 2597. https://doi.org/10.1016/j.jmmm.2012.02.009
- Liang, H.; Niu, H.; Li, P.; Tao, Z.; Mao, C.; Song, J.; Zhang, S. Mater. Res. Bull. 2013, 48, 2415. https://doi.org/10.1016/j.materresbull.2013.02.066
- Amarjargal, A.; Tijing, L. D.; Im, I. T.; Kim, C. S. Chem. Eng. J. 2013, 226, 243. https://doi.org/10.1016/j.cej.2013.04.054
- Xin, T.; Ma, M.; Zhang, H.; Gu, J.; Wang, S.; Liu, M.; Zhang, Q. Appl. Surf. Sci. 2014, 288, 51. https://doi.org/10.1016/j.apsusc.2013.09.108
- Wu, S. H.; Wu, J. L.; Jia, S. Y.; Chang, Q. W.; Ren, H. T.; Liu, Y. Appl. Surf. Sci. 2013, 287, 389. https://doi.org/10.1016/j.apsusc.2013.09.164
- Chi, Y.; Yuan, Q.; Li, Y.; Zhao, L.; Li, N.; Li, X.; Yand, W. J. Hazard. Mater. 2013, 262, 404. https://doi.org/10.1016/j.jhazmat.2013.08.077
- Ma, P.; Jiang, W.; Wang, F.; Li, F.; Shen, P.; Chen, M.; Wang, Y.; Liu, J.; Li, P. J. Alloys Compd. 2013, 578, 501. https://doi.org/10.1016/j.jallcom.2013.07.026
- Mohapatra, S.; Rout, S. R.; Panda, A. B. Colloids Surf., A: Physicochem. Eng. Aspects 2011, 384, 453. https://doi.org/10.1016/j.colsurfa.2011.05.001
- Wang, L.; Li, J.; Wang, Y.; Zhao, L.; Jiang, Q. Chem. Eng. J. 2012, 181-182, 72. https://doi.org/10.1016/j.cej.2011.10.088
- Goyal, A.; Bansal, S.; Singha, S. Int. J. Hydrogen Energy 2014, 39, 4895. https://doi.org/10.1016/j.ijhydene.2014.01.050
- Xiong, P.; Fu, Y.; Wang, L.; Wang, X. Chem. Eng. J. 2012, 195-196, 149. https://doi.org/10.1016/j.cej.2012.05.007
- Lee, Y.; Chae, J.; Kang, M. J. Ind. Eng. Chem. 2010, 16, 609. https://doi.org/10.1016/j.jiec.2010.03.008
- Kim, H. S.; Kim, D.; Kwak, B. S.; Han, G. B.; Um, M. H.; Kang, M. Chem. Eng. J. 2014, 243, 272. https://doi.org/10.1016/j.cej.2013.12.046
- Lee, G.; Kang, M. Curr. Appl. Phys. 2013, 13, 1482. https://doi.org/10.1016/j.cap.2013.05.002
- Wang, Y. H. A.; Bao, N.; Gupta, A. Solid State Sci. 2010, 12, 387. https://doi.org/10.1016/j.solidstatesciences.2009.11.019
- Liu, H.; Wang, M.; Wang, Y.; Liang, Y.; Cao, W.; Su, Y. J. Photochem. Photobiol., A: Chem. 2011, 223, 157. https://doi.org/10.1016/j.jphotochem.2011.06.014
- Leong, K. H.; Monash, P.; Ibrahim, S.; Saravanan, P. Sol. Energy 2014, 101, 321. https://doi.org/10.1016/j.solener.2014.01.006
- Rashad, M. M.; Elsayed, E. M.; Al-Kotb, M. S.; Shalan, A. E. J. Alloys Compd. 2013, 581, 71. https://doi.org/10.1016/j.jallcom.2013.07.041
- Kim, J.; Kang, M. Int. J. Hydrogen Energy 2012, 37, 8249. https://doi.org/10.1016/j.ijhydene.2012.02.057
- Dua, J.; Chen, H.; Yang, H.; Sang, R.; Qian, Y.; Li, Y.; Zhu, G.; Mao, Y.; He, W.; Kang, D. J. Microporous Mesoporous Mater. 2013, 182, 87. https://doi.org/10.1016/j.micromeso.2013.08.023
- Zhang, W.; Zhao, J.; Liu, Z.; Liu, Z.; Fu, Z. Appl. Surf. Sci. 2010, 256, 4423. https://doi.org/10.1016/j.apsusc.2009.12.064
- Lyubutin, I. S.; Lin, C. R.; Starchikov, S. S.; Siao, Y. J.; Shaikh, M. O.; Funtov, K. O.; Wang, S. C. Acta Mater. 2013, 61, 3956. https://doi.org/10.1016/j.actamat.2013.03.009
- Fan, L.; Dongmei, J.; Yan, L.; Xueming, M. Phys. B: Condens. Matter. 2008, 403, 2193. https://doi.org/10.1016/j.physb.2007.11.014
- Wang, Q.; Wei, X.; Dai, J.; Jiang, J.; Huo, X. Mater. Sci. Semicond. Process. 2014, 21, 111. https://doi.org/10.1016/j.mssp.2014.01.004
- Kim, Y.; Jeong, J. H.; Kang, M. Inorg. Chim. Acta 2011, 365, 400. https://doi.org/10.1016/j.ica.2010.09.041
Cited by
- A degradation column for organic dyes based on a composite of CuFeS2 nanocrystals and sawdust vol.51, pp.11, 2016, https://doi.org/10.1007/s10853-016-9844-4
- Catalytic oxidation of lignin to dicarboxylic acid over the CuFeS2 nanoparticle catalyst vol.7, pp.4, 2018, https://doi.org/10.1515/gps-2017-0056
- nanoparticles pp.00094536, 2019, https://doi.org/10.1002/jccs.201800355