Jakub Tolasz*ab,
Jiří Henychab,
Martin Šťastnýa,
Zuzana Němečkováa,
Michaela Šrámová Slušnáa,
Tomáš Opletalc and
Pavel Janošb
aInstitute of Inorganic Chemistry of the Czech Academy of Sciences, 25068 Husinec-Řež, Czech Republic. E-mail: tolasz@iic.cas.cz
bFaculty of the Environment, University of Jan Evangelista Purkyně, Králova Výšina 7, 400 96 Ústí nad Labem, Czech Republic
cRegional Centre of Advanced Technologies and Materials, Department of Analytical Chemistry, Faculty of Science, Palacký University, 17. Listopadu 12, 771 46 Olomouc, Czech Republic
First published on 8th April 2020
A simple low-temperature water-based and one-pot synthesis was developed for the preparation of nanocrystalline CeO2 that was used for degradation of the toxic organophosphate pesticide parathion methyl. By changing the reaction temperature in the range from 5 °C to 95 °C, several properties (i.e., crystallinity, grain size and surface area) of nanoceria can be easily controlled. The catalytic decomposition of parathion methyl to its degradation product 4-nitrophenol was highly dependent on the CeO2 preparation temperature. It was demonstrated that at low temperature (i.e. 5 °C), CeO2 with very small crystallites (<2 nm) and high surface area can be obtained. For practical use, it was demonstrated that highly crystalline CeO2 can be prepared at room-temperature (30 °C) in at least 100 g batches. It was shown that precipitated nanoceria had high thermal stability and its post-synthesis annealing up to 400 °C did not significantly alter the material properties and hence the catalytic activity. Furthermore, as shown by the reusability tests, the sorbent can be reactivated by simply washing with water which demonstrated its durability.
CeO2 nanoparticles can be prepared by various procedures, such as calcination of the suitable precursor, reflux or hydrothermal synthesis in an autoclave at elevated pressure, by low-temperature synthesis at 90 °C or using photochemical synthesis. These procedures were compared for preparation of CeO2 (ref. 14 and 15) or CeO2 composites with samarium.16 Many studies are focused on the shape engineering of ceria to obtain cubic, rod-like17 and plate-like18 particles. Nowadays very popular sol–gel methods are suitable for preparation of homogeneously doped materials, for example Ce-doped TiO2 (ref. 19) or Ce-doped YAG,20 and preparation of ultrafine pure metal oxides nanoparticles of Zn, Sn, Ti, Zr, Fe, Ni, Ga, Mn, In and Ce.21 However, the calcination at high temperatures or high-pressure treatment or multi-step approach is usually necessary. Modern green chemistry favours low-temperature one-pot syntheses, especially for potential industrial production, where the energy consumption and the complexity of the synthesis and hence its price must be reduced. For example, the low-temperature synthesis was used to prepare a photocatalytic composite with titanium nanorods,22 nevertheless the effect of reaction temperature and the effect of cerium salt concentration was not discussed in detail.
Another interesting application of nanoceria is decomposition of phosphoesters,23 (e.g. organophosphate pesticides, warfare agents and others). However, its catalytic activity can be inhibited by the degradation product24 that leads to inactivation of the catalyst for further use. It is therefore highly desirable to find a way how to prevent or eliminate catalyst inactivation and find an easy process how to regenerate its function. It has been previously described25 that both CeO2 bivalent character26 (also found in Mn3O4, MnFe2O4 and other metal oxides) and surface –OH groups play an important role in its catalytic properties. It was suggested that –OH surface groups are consumed during the degradation reactions but can be easily replenished and thus the catalyst can be regenerated.
In this work, we prepared nanocrystalline cerium oxides by simple low-temperature synthesis. We demonstrated that by simple changing of the reaction temperature, the properties such as particle size, specific surface area, porosity or surface chemical composition can be modified. The catalytic activity of prepared materials was tested by decontamination of parathion methyl in non-aqueous solution. Reusing of catalyst led to the decrease of its catalytic activity although this was not observed during testing in water in our previous study.27 We combined these two procedures and demonstrated that the catalyst used in a non-aqueous solution can be reactivated by water washing for further use.
Sample CeS_30 was prepared analogously at room temperature (30 °C) in a larger amount. Concretely, 12 L of 60 mM cerium salt solution was used in a 15 L reactor to yield 100 g of the powder sample that was used for extensive degradation and reusability tests and for preparing samples annealed at 400, 600, 800 and 1000 °C. The samples were denoted as CeA_yyyy, where yyyy is the annealing temperature.
To monitor parathion-methyl (PM) and its degradation product 4-nitrophenol (4-NP) high-performance liquid chromatography (HPLC) system with a diode array detector (DAD) DIONEX UltiMate 3000 (Thermo Scientific™, Palo Alto, USA) was used. Chromatographic analysis was carried out in a reverse phase system (RPLC-C18) on AccucoreTM column, 2.6 μm, PFP, 150 × 4.6 mm. Acetonitrile (MeCN), methanol (MeOH) and water (H2O) acidified with formic acid (HCOOH, 0.1%) were used as mobile phases for gradient elution. Further details are given in the electronic ESI in Table 1S.†
Sample | Grain size (XRD), (nm) | Grain size (TEM), (nm) | SSA, (m2 g−1) | MPD, (nm) | Vpore, (cm3 g−1) | Δm1000, (%) | Ce(III), (%) | Ce(IV), (%) | O-vac 600 cm−1, (Raman int.) |
---|---|---|---|---|---|---|---|---|---|
CeT_05 | 2 | 3 | 112 | 3.5 | 0.12 | 11.5 | 29.4 | 70.6 | 3.5 |
CeT_20 | 3 | 5 | 128 | 4.7 | 0.17 | 8.3 | 31.1 | 68.9 | 3.5 |
CeT_40 | 5 | 6 | 106 | 5.0 | 0.15 | 5.8 | 25.7 | 74.3 | 1.5 |
CeT_60 | 8 | 7 | 115 | 5.6 | 0.18 | 5.0 | 19.7 | 80.3 | 4.5 |
CeT_80 | 10 | 9 | 89 | 7.6 | 0.18 | 3.9 | 16.5 | 83.5 | 2.0 |
CeT_95 | 14 | 13 | 83 | 14.0 | 0.18 | 2.1 | 22.2 | 77.8 | 1.0 |
CeS_30 | 11 | 12 | 68 | 8.9 | 0.15 | 3.0 | 21.0 | 79.0 | 2.4 |
Fig. 2 The growth of crystalline nanoparticles and the loss of the amorphous fraction is visible in a series of TEM images of samples prepared at (a) 5, (b) 20, (c) 40, (d) 60, (e) 80 and (f) 95 °C. |
Fig. 3 High-resolution transmission electron images of the lattice structure of oriented crystals along the axis (111) for samples prepared at (a) 5, (b) 20, (c) 40, (d) 60, (e) 80 and (f) 95 °C. |
Fig. 4 Raman spectra after background correction of the three different regions from left: (350 to 550) cm−1; (550 to 700) cm−1; (700 to 1800) cm−1. |
Furthermore, D-band centered at about 600 cm−1 suggesting a formation of the oxygen vacancies30 was found in all prepared samples, but its intensity was the highest for samples CeT_05, CeT_20 and CeT_60. The bands in the region (700 to 1800) cm−1 result from dissociative adsorption of CO2 on the CeO2 surfaces31 that is typical for samples stored at ambient air. Interestingly, the intensity of these bands increases for the samples prepared at higher temperature suggesting different reactivity and adsorption properties of the prepared samples.
In the nanocrystalline CeO2, the surface Ce(III)/Ce(IV) ratio is one of the important parameters that can be obtained by XPS measurements. Nevertheless, it is important to mention that measurements under strong vacuum may induce the formation of Ce(III) states. The results (Table 1 and Fig. 1S†) show that higher Ce(III)/Ce(IV) ratio was in samples prepared at lower reaction temperature (up to 60 °C). The amount of Ce(III) ions is well correlated with the formation of oxygen vacancies as show Raman data. This suggests that at lower temperatures, more defectious particles are formed that can be related to the soft synthesis procedure employed. The higher amount of amorphous fraction (detected by HRTEM) was present at lower temperatures and can be due to the higher amount of precipitated Ce(OH)3 gel that has not been fully oxidized (as suggest higher amount of Ce3+ detected by XPS). Different properties among the samples were also revealed by thermal analysis with MS detection (Fig. 5) (for comparison with CeS_30 sample see Fig. 4S in an ESI†). The sample weight loss linearly decreases with increasing synthesis reaction temperature from 12% (for CeT_05) to 2% (for CeT_95). Interestingly, a significant difference was observed in the releasing of residual (NO3)− ions from samples originating probably from Ce precursor. While samples synthesized at low temperatures (5, 20 and 40 °C) release weakly bound nitrates mainly as NO2 at relatively low temperature (up to 130 °C), CeT_60 sample released both NO2 and NO but the latter at significantly higher temperature (200 to 250) °C. In the sample CeT_80, mainly NO releasing at even higher temperatures (200 to 350) °C was detected while in the sample CeT_95, only negligible amount of NO without any NO2 was registered. This suggests the different binding of (NO3)− in the crystal lattice of the samples prepared at various temperature and it also may be strongly related to the surface structure, Ce(III)/Ce(IV) ratio and oxygen vacancies in the samples. Most of the mass from samples, especially water and carbon dioxide, is released between temperatures 112 °C (for CeT_95) and 130 °C (for the CeT_05). It is also evident that CO2, resulting from surface adsorbed carboxylates and carbonates,31 is released continuously up to temperatures (650–800) °C (Table 1).
Increasing reaction temperature also leads to a decrease of the specific surface area from 112 m2 g−1 at 5 °C and 128 m2 g−1 at 20 °C, to 83 m2 g−1 at 95 °C (Table 1). All samples showed type IV isotherm with hysteresis which is characteristic for finely non-aggregated (powdered or nanoparticulate) non-porous samples with capillary condensation in the interparticle space. Interestingly, as show the changing shape of the hysteresis (Fig. 6a) and its shift to higher p/p0 values, the shape of the pores and their size distribution (Fig. 6b) is strongly dependent on the reaction temperature. While the hysteresis of the CeT_05 sample is close to the H4 type with crevice pores in the microporous region, as the temperature rises, hysteresis transforms to H2 type suggesting formation of the complex porous structure consisting mainly of interconnected porous networks with pores with different widths and shapes. NLDFT calculation (Fig. 6b) shows that all materials are mesoporous with pores between (2 to 30) nm with a significant amount of micropores that are most evident in the CeT_05 sample. Similarly, to the calculated grain size (XRD data, Table 1), the mean pore diameter of the samples linearly increases with increasing reaction temperature. This demonstrates, in accordance with XRD and TEM observations, that the particle size and porosity can be easily controlled by simple changing of the reaction temperature. The total pore volume of all samples is between (0.1 to 0.2) cm3 g−1 (Table 1).
Sample | k1 (PM), (min−1) | k2 (PM), (min−1) | ν1, (mmol L−1 min−1 g−1) | α60, (%) | k1 (4-NP), (min−1) | k2 (4-NP), (min−1) |
---|---|---|---|---|---|---|
CeT_05 | 0.674 | 0.038 | 10.3 | 93.7 | 0.864 | 0.050 |
CeT_20 | 0.378 | 0.026 | 7.4 | 84.9 | 0.544 | 0.016 |
CeT_40 | 0.431 | 0.023 | 8.3 | 80.0 | 0.476 | 0.002 |
CeT_60 | 0.740 | 0.035 | 10.4 | 90.6 | 0.772 | 0.028 |
CeT_80 | 0.395 | 0.009 | 8.0 | 68.0 | 0.603 | 0.009 |
CeT_95 | 0.280 | 0.001 | 3.4 | 28.4 | 0.145 | 0.006 |
CeS_30 | 0.496 | 0.022 | 9.6 | 68.1 | 0.378 | 0.009 |
The initial rate ν1 was calculated by formula (1). The degree of conversion α60 was calculated by formula (2):
(1) |
(2) |
The rate constants were calculated according to the model used in our previous work.33 Formula (3) is for PM degradation and formula (4) for 4-NP formation, respectively.
qPMt = qPM1e(−k1t) + qPM2e(−k2t) + qPM∞ | (3) |
q4NPt = 1 − (q4NP1e(−k1t) + q4NP2e(−k2t) + q4NP∞) | (4) |
As show the kinetics curves of both PM degradation and formation of 4-NP (Fig. 7), the degradation efficiency of prepared samples differs significantly. The highest initial rate of PM decomposition was observed on the samples Ce_T05 and Ce_T60, the latter also had the highest degree of conversion at 60 min. Note that the degradation product 4-NP is after its formation partially released to the solution but also adsorbed on the surface of the catalyst and therefore, it is not possible to calculate the reaction balance in this experimental arrangement. The highest amount of the degradation product 4-NP at 120 min was in the reaction mixture catalysed by CeT_80, even though the amount of PM decomposition on this sample was not the highest. It can be argued that the amount of 4-NP released into the solution is inversely proportional to the specific surface area of the catalyst. Therefore, we investigated also the reaction balance on selected sample CeS_30 (discussed further below).
Fig. 7 Kinetic curves for (a) degradation of parathion methyl (PM) and (b) formation of 4-nitrophenol (4-NP) on prepared samples. The solid lines are least-square best fits of the data to eqn (1). |
From the plots of the normalized materials characteristics (specific surface area, amount of oxygen vacancies and Ce(III) surface content) and normalized kinetic parameters for catalytic decomposition of PM in Fig. 8, it is clear that materials properties, mainly the amount of oxygen vacancies, significantly influence the catalytic activity. Note that the other parameters, such as content of the amorphous fraction especially in the samples prepared at low temperature, can also influence these relationships. CeT_60 sample that has the highest content of O-vac and still relatively high surface area showed the highest degree of conversion, high rate constants and the largest amount of 4NP released into solution. Therefore 60 °C seems to be the golden mean and a good compromise between the lowest and highest temperatures used in this low-temperature synthesis.
Sample | Grain size (XRD), (nm) | SSA, (m2 g−1) | MPD, (nm) | Vpore, (cm3 g−1) | ν1, (mmol L−1 min−1 g−1) | α60, (%) | k1 (PM), (min−1) | k2 (PM), (min−1) |
---|---|---|---|---|---|---|---|---|
CeS_30 | 11 | 68 | 8.9 | 0.15 | 9.6 | 68.1 | 0.496 | 0.022 |
CeA_400 | 11 | 67 | 9.3 | 0.16 | 7.7 | 70.7 | 0.383 | 0.023 |
CeA_600 | 21 | 53 | 10.6 | 0.14 | 6.3 | 60.0 | 0.409 | 0.028 |
CeA_800 | 100 | 14 | 14.1 | 0.05 | 0.8 | 14.5 | 0.090 | 0.090 |
CeA_1000 | 186 | <1 | 11.2 | <0.001 | — | — | — | — |
Fig. 9 The temperature-induced growth of nanoparticles is visible in a series of transmission electron microscope images of sample CeS_30 annealed to (a) 400, (b) 600, (c) 800 and (d) 1000 °C. |
All samples showed type IV isotherms with hysteresis (Fig. 10a), except the sample annealed at 1000 °C. The samples annealed at 400 °C and 600 °C are mesoporous and have relatively high surface area (67 and 53) m2 g−1. At 800 °C, the sample surface area decreased to 15 m2 g−1 with significant loss of pore volume. Annealing to 1000 °C leads to complete loss of mesoporosity as indicates the disappearing of the hysteresis loop and negligible surface area and pore volume of the sample. Pore size distribution calculated by the NLDFT method was practically unchanged for samples annealed up to 600 °C (Fig. 10b).
Fig. 11 Kinetic measurement of (a) degradation of parathion methyl (PM) and (b) formation of 4-nitrophenol (4-NP) on annealed samples. The solid lines are least-square best fits of the data to eqn (3) and (4). |
As can be seen from materials characterization, the sample annealed at 400 °C retained its structural properties and thus also its catalytic activity. It also demonstrates that post-synthesis annealing of the samples to remove any residuals from the synthesis is not necessary and that the nanoceria was sufficiently crystallized during low temperature used in our synthesis. Interestingly, the samples retained very good catalytic activity even after treatment at 600 °C. At higher annealing temperatures, the specific surface area and, therefore, also the degradation activity decreased significantly.
After 60 minutes of degradation in the first cycle, the sample was centrifuged and extracted twice with MeOH and then twice with MeCN; all aliquots were combined and analysed to determine the degree of conversion (α60). The catalyst was then air-dried and used in next cycle. This procedure is shown schematically in Fig. 12. The obtained c60/c0 molar ratio for PM and 4-NP are shown in Table 2S† together with degree of conversion at 60 minutes. It shows that by MeOH extraction a 1:1 molar ratio balance between PM degradation and 4-NP formation was achieved. The results of reusability experiments (9 cycles) are shown in Fig. 13 and Table 2S.† After the third cycle, the degree of conversion α60 decreased from the initial 50% to 19%. Therefore, to test a regeneration of the catalyst, the sorbent was immersed in water for 24 hours, dried and another three cycles were performed. The water regeneration of the sorbent caused improvement of the degree of conversion α60 back to 34%. After three more cycles, α60 decreased to only 9%.
Fig. 13 Nine PM degradation cycles with two water regeneration steps (first for 24 and second for 48 hours). |
Therefore, another catalyst regeneration using water washing was prolonged to 48 hours. Interestingly, the longer regeneration of the sorbent resulted in a substantial improvement of the degree of conversion α60 back to 41% after 6 cycles. After 9 cycles with two washing steps, the degree of conversion α60 decreased to 8%.
It is widely accepted that not only the available surface area but also its qualitative properties are important for achieving high catalytic activity in pollutant degradation. In the case of CeO2, a suitable Ce(III)/Ce(IV) ratio, as well as abundance of the surface –OH groups are the main factors. The formation of these active sites is feasible in defective nanocrystalline structures. According to our previous investigation,27 the catalytic activity of nanoceria can be preserved or recovered when the tests are performed in water, where a depletion of –OH groups can be compensated and the active sites can be regenerated. Contrary, in the non-aqueous environment, surface –OH groups that act as strong nucleophile in degradation reaction are consumed without regeneration and the catalytic activity decreases drastically after the first cycle. However, as demonstrated here, even the catalyst used in a non-aqueous solvent can be easily recovered by its simple immersion in water for a few hours. Nevertheless, we assume that this process is feasible only for nanoceria thanks to its unique properties (such as high oxygen mobility) and it cannot be simply applied to other metal oxide catalysts. Interestingly, during the degradation cycles, the colour of the catalyst gradually changes from yellow to orange and after water washing back to yellow. The orange colour was attributed to surface peroxo groups in many studies,4 but the mechanism of their formation in our catalytic test remains unclear.
Footnote |
† Electronic supplementary information (ESI) available. See DOI: 10.1039/d0ra00937g |
This journal is © The Royal Society of Chemistry 2020 |