Magnetic silica particles functionalized with guanidine derivatives for microwave-assisted transesterification of waste oil

This study aimed to develop a facile synthesis procedure for heterogeneous catalysts based on organic guanidine derivatives superbases chemically grafted on silica-coated Fe3O4 magnetic nanoparticles. Thus, the three organosilanes that were obtained by reacting the selected carbodiimides (N,N′-dicyclohexylcarbodiimide (DCC), N,N′-diisopropylcarbodiimide (DIC), respectively 1-ethyl-3-(3-dimethylaminopropyl) carbodiimide (EDC) with 3-aminopropyltriethoxysilane (APTES) were used in a one-pot synthesis stage for the generation of a catalytic active protective shell through the simultaneous hydrolysis/condensation reaction with tetraethyl orthosilicate (TEOS). The catalysts were characterized by FTIR, TGA, SEM, BET and XRD analysis confirming the successful covalent attachment of the organic derivatives in the silica shell. The second aim was to highlight the capacity of microwaves (MW) to intensify the transesterification process and to evaluate the activity, stability, and reusability characteristics of the catalysts. Thus, in MW-assisted transesterification reactions, all catalysts displayed FAME yields of over 80% even after 5 reactions/activation cycles. Additionally, the influence of FFA content on the catalytic activity was investigated. As a result, in the case of Fe3O4@SiO2-EDG, a higher tolerance towards the FFA content can be noticed with a FAME yield of over 90% (for a 5% (weight) vs oil catalyst content) and 5% weight FFA content.

In a 50 mL round bottom flask, 10 mmoles of the starting material carbodiimide (2.1 g N,N′ diciclohexyl carbodimides (DCC) or 1.3 g N,N′-diisopropylcarbodiimide (DIC), or 1.55 g 1-ethyl-3-(3-dimethylaminopropyl) carbodiimide (EDC), respectively) were dissolved in 20 mL of dry toluene under a nitrogen blanket. To this solution 10 mmoles (2.21 g) of 3-aminopropyltriethoxysilane (APTES) were added and the mixture was kept www.nature.com/scientificreports/ under continuous stirring at 100 °C for 24 h. The toluene was removed using a rotary evaporator at reduced pressure and yellowish oils were obtained and used directly for the preparation of the catalysts without any further purifications.
Synthesis procedure of the magnetic nanoparticles Fe 3 O 4 . The magnetic Fe 3 O 4 nanoparticles were prepared by a chemical co-precipitation method 31,39 . 18 mmoles FeCl 3 ·6H 2 O (4.864 g) and 8.9 mmoles FeCl 2 (1.134 g) were dissolved in 100 mL of bidistilled water. The solution was degassed by nitrogen bubbling for 15 min at room temperature under mechanical stirring (350 rpm) while being immersed in an ultrasonic water bath (40 kHz and 400 W). The US power absorbed by the solution was determined calorimetrically as 15.2 W with a corresponding power density of 0.152 W/mL). Then, 10 mL of 25% NH 4 OH were added to this solution and the reaction mixture remained under mechanical stirring and nitrogen atmosphere at room temperature for 1 h. The black precipitate was separated from the solution using magnetic decantation and was subsequently washed thoroughly (10 times) with deionized water and acetone. The resulting black solid was dried under vacuum at room temperature for 5 h, yielding 2.10 g of magnetic nanoparticles.
Synthesis procedure for the magnetic catalysts. In a 250 mL two-necked round bottom flask equipped with mechanical stirring, 0.55 g cetyltrimethylammonium bromide (CTAB) were dissolved in 200 mL of water. In the solution obtained 1 g of magnetic nanoparticles were dispersed to serve as core for the catalyst generation. After the preparation of the dispersion, the solution was heated at 90 °C and 1. Transesterification procedure. The MW-assisted transesterification reactions were performed using a Biotage Initiator microwave equipment with a single-mode applicator (https:// selekt. biota ge. com/ hubfs/ ORGAN IC/ Premi um% 20Con tent_ Docum ents/ Initi ator/ PPS299. v4% 20-% 20Mic rowave% 20syn thesis% 20bro chure. pdf? hsLang= en). The reactions were carried out in 20 mL glass reactors capable of resisting at 10 bars. The reactor was charged with methanol (2.26 g), catalyst (0.35 g), and oil (10 g) (commercial sunflower oil) to ensure an oil: methanol molar ratio of 1:6 and a catalyst content of 3.5% (wt% versus oil). The microwave power was 400 W for a few seconds and then 40-60 W to maintain the reaction temperature at 80 °C for 180 min under magnetic stirring (780 rpm). The transesterification reactions for the conventional heating were performed in the same glass reactor using a silicon heating bath, the same temperature, stirring rate and reaction time. For the free fatty acid (FFA) content influence study, lauric acid was used together with the commercial sunflower oil to obtain the desired FFA content. After the reaction, the catalyst was removed by magnetic separation (using a neodymium magnet) and decanting. The glycerol was separated by settling or centrifugation. The upper layer formed was washed with acidified water and dried over calcium chloride (this wet method is used only at laboratory scale, for industrial process, dry washing may be considered 11 ). Then, the organic phase was processed with a rotary evaporator to remove unreacted methanol.
Fatty acid methyl ester content was determined by GC analysis using a HP 6890 gas chromatograph with HP-INNOWAX 19091N-133 column 30 m × 250 μm, according to European Standards 40 . The concentration of FAME was calculated from integration of the chromatograms using the area of the peaks. The esters content was calculated using the Eq. (1).
Unde: ∑Ai-total peak area from the FAME C 14:0 to C 24:1 ; As-peak area of internal standard-methyl heptadecanoate; Cs-concentration, in mg/ mL, of the internal standard methyl heptadecanoate; Vs-volume, in mL, of the internal standard solution, methyl heptadecanoate; m-mass, in mg, of the sample.
The transesterification reaction yield was calculated using the Eq. (2).
The activation of the catalyst consisted in the dispersion of the magnetic particles before each reaction in an alkaline solution (methanolic solution of NaOH 1 M, for 5 min ultrasound assisted-GT SONIC Professional Ultrasonic Cleaner, VGT-1620 T model, US Power 50 W) followed by magnetic separation and then washing with methanol up to a pH of 8 for the residual methanolic solutions. Before the transesterification reaction or between reactions in the case of reutilization of the catalyst, the catalyst was dried.
Characterization. FTIR analysis was performed using a Bruker VERTEX 70 spectrometer using 32 scans with a resolution of 4 cm −1 in 4000-600 cm −1 region. The catalysts and intermediates were analyzed using the KBr pellet technique for the fresh catalyst and attenuated total reflection (ATR) technique for the used catalysts.
The morphology of the catalysts was also evaluated by FEI Nova NanoSEM 630 Scanning Electron Microscope (FEG-SEM) using an ultra-high-resolution detector (TLD detector) at an acceleration voltage of 10 kV .
(2) yield(%) = mass of pure FAME obtained (g) theoretical mass of FAME obtained from the reaction (g) × 100 Scientific Reports | www.nature.com/scientificreports/ and a working distance of 5 mm. The elemental identification and quantification within the catalyst particles were performed with the Element energy dispersive spectroscopy (EDS) system (Smart Insight AMETEK). The Energy-dispersive X-ray spectroscopy spectra were acquired at an acceleration voltage of 10 kV. The thermogravimetric analyses (TGA) of the catalysts were performed using a Netzsch TG 209 F3 Tarsus equipment considering the next parameters: nitrogen atmosphere flow rate 20 mL min −1 ; samples mass: ∼ 3 mg; temperature range: room temperature − 700 °C; heating rate: 10 °C min −1 in an alumina (Al 2 O 3 ) crucible.
The digestion of the samples was performed using the speedwave XPERT microwave digestion system (Berghof Products-Instruments GmbH) involving a digestion protocol based on HCl/HF mixture (see Table 1). The HF excess have been neutralized with H 3 BO 3 . To determine the total nitrogen of the digested mixtures, the organic nitrogen to nitrate digestion by persulfate has been performed. The contents of the total nitrogen (TN) in digested samples were determined by using a Multi N/C 3100 TOC Analyzer (Analytik Jena, Germany) equipped with a CLD chemiluminescence detector for TN analysis. The nitrogen content was calculated as mg of nitrogen per gram of solid sample.

Result and discussion
The first step of this study consisted in the synthesis of magnetic SiO 2 particles with three different guanidine derivatives, organic superbases. The synthesis strategy adopted involved the generation of Fe 3 O 4 nanoparticles by precipitation technique starting from a mixture of FeCl 3 and FeCl 2 using ammonia under ultrasonic treatment. The magnetic nanoparticles were then coated with a silica shell using TEOS and a silane-superbase precursor, obtained by reacting different carbodiimides with APTES. Thus, magnetic particles presenting a SiO 2 -superbase shell were obtained through a one-pot synthesis approach (Schemes 1 and 2).
The successful synthesis of the magnetic catalyst particles was investigated by FTIR analysis (Fig. 1). The presence of the organic bases is confirmed by the characteristics signals for the organic derivatives such as a large band at 3400-3000 cm −1 specific for O-H and N-H stretching 31,43 , 2930 cm −1 and 2850 cm −1 specific for the asymmetric and symmetric C-H vibration 44 , 1630 cm −1 , and 1448 cm −1 attributed to C=N and C-N (from guanidine) stretching 31,35,38,45 . Further, the presence of the magnetite and silica shell leads to the presence of a www.nature.com/scientificreports/ specific stretching vibration band of the siloxane groups (Si-O) at 1051 cm −1 and an intense band due to Fe-O bond vibrations 46 split into two peaks at 635 and 588 cm −1 . This band splitting has been previously associated with the nanometric size of magnetic nanoparticles (Fe 3 O 4 ) and was observed for all of the prepared catalysts 31 . TGA analysis was employed to investigate the thermal stability of the catalysts and to assess the amount of organic superbase covalently bonded to the silica shell (Fig. 2). The initial weight loss observed in all cases is attributed to physiosorbed water and solvent molecules on the surface of the catalysts (less than 2% in all cases). The second weight-loss step is comprised of two consecutive weight-losses, and it can be attributed to the detachment of the organosilane and the guanidine derivatives from the silica shell. The DTG curves indicated that the second weight loss step begins at temperatures above 195 °C and the highest thermal stability was observed for Fe 3 O 4 @SiO 2 -EDG. The high temperature confirms a chemical attachment between the organosilane derivatives and the silica. The weight loss for this step is approximately 36%, 15%, and 18.8% for Fe 3 O 4 @SiO 2 -DCG, Fe 3 O 4 @SiO 2 -DIG, and Fe 3 O 4 @SiO 2 -EDG, respectively. This difference in weight loss is following the difference in the molecular weight of the superbases, indicating that a comparable molar concentration of superbase was successfully attached to the silica.
The morphology of the catalysts was investigated by SEM analysis to assess the influence of the organosilane derivative on the SiO 2 shell. Considering that the SiO 2 shell formation involves the simultaneous hydrolysis/ polycondensation of both TEOS and the organosilane derivative, leading to a functionalized silica shell. The SEM images revealed the development of spherical-shaped catalysts particles with particle sizes of 100-200 nm (Fig. 3). The catalyst particles dimensions vary only slightly between the samples with the size decreasing in the order Fe 3 O 4 @SiO 2 -DCG > Fe 3 O 4 @SiO 2 -DIG > Fe 3 O 4 @SiO 2 -EDG, which can be related to an increase of the basicity of the superbases and the presence of a tertiary amine group in the case of Fe 3 O 4 @SiO 2 -EDG. The relative agglomerated characteristics of the catalyst particles can be related to the high content of guanidine derivatives 43 .  www.nature.com/scientificreports/ The magnetic catalyst particles were also investigated by EDS analysis to ascertain their surface chemical composition (Fig. 4). The analysis confirmed the presence of Fe, O, Si, C and N atoms which would be in accordance with a core-shell structuring of the material and confirms the presence of organic derivatives on the surface of the Fe 3 O 4 @SiO 2 -DCG catalyst. The higher C content denoted by the EDS analysis compared to the TGA analysis results suggests that the organic derivatives are predominantly present at the surface of the particles. This could be explained by the hydrophilic characteristic of the organic functional groups that influence the formation and structuring of the micelles during the silane precursors hydrolysis step. Further, the characteristic signals for Fe may indicate a somewhat deficient covering of magnetite particles by the silica shell or a covering by a very thin layer.
To evaluate the nitrogen amount in the bulk of the catalysts, total nitrogen (TN) content analysis was performed. The results are presented in Table 2. Compared with the EDS information the TN analysis indicates the organic derivatives are predominantly present at the surface of the catalysts which will improve their activity.
A qualitative base strength (H_) assessment of the samples was performed using Hammet indicator method. The value obtained were between 15.0 ≤ H ≤ 17 for all the catalysts 35,42 . However, due to the initial brown shade of the solid catalysts the change of color of the solid catalyst is very difficult to be properly evaluated. The titration The BET analysis was performed to assess the textural properties of the magnetic catalysts ( Table 3). The analysis revealed comparable specific surfaces with a slight increase in the case of Fe 3 O 4 @SiO 2 -DIG. Also, the pore volume is comparable between Fe 3 O 4 @SiO 2 -DIG and Fe 3 O 4 @SiO 2 -EDG, while an increased size was registered for Fe 3 O 4 @SiO 2 -DIG and Fe 3 O 4 @SiO 2 -DCG. Considering that the reaction takes place in liquid media, the values for the specific surface area are reasonable, while an increased pore size should facilitate the adsorption/ desorption and diffusion processes.
The XRD analysis of the magnetic catalysts (Fig. 5) indicates the presence of a spinel-structured oxide identified as either magnetite (Fe 3 O 4 ) or a mixture between magnetite (Fe 3 O 4 ) and maghemite (γ-Fe 2 O 3 ) (in the case of Fe 3 O 4 @SiO 2 -DCG). No characteristic peaks could be attributed to the silica shell, possibly due to its small thickness and amorphous characteristics.
Different preparation and use of guanidine derivatives anchored on silica gel as catalysts for biodiesel production were previously reported in the literature 35,37,39 . The yield of biodiesel was within 90-95% at 80 °C and 180 min of reaction time with the use of dicyclohexylguanidine but using a methanol/oil molar ratio in the range of 20-30. In the case of tetramethylguanidine, the yield of biodiesel was 86% at similar reaction conditions.
Our results of the investigation of the catalytic activity of the synthesized magnetic catalysts for the biodiesel synthesis and the possibility for process intensification using microwave (MW) irradiation are presented in Fig. 6. The reactions were performed in our case at a temperature of 80 °C, a reaction time of 180 min, molar ratio oil/ MeOH 1/6, a 3.5% weight ratio catalyst vs. oil, and a stirring rate of 780 rpm. The catalytic activity variation was in the order:     www.nature.com/scientificreports/ good microwave energy absorption properties 47 . Even if the size of the catalyst particles is very small, a selective heating effect of the catalyst particles is possible, which may justify the favorable effect of the microwaves 48 . Nevertheless, the catalyst activation stage proves to have a dramatic role in the catalyst activity. The activation of the catalyst aims to assure a deprotonated C=N bond from guanidine derivatives prior to the transesterification process, which promotes the deprotonation of methanol during the first stage of the base-catalyzed transesterification process 2,49 . Thus, by activation, all catalysts display FAME yields of over 70% at a temperature of 80 ºC, a reaction time of 180 min, molar ratio oil/MeOH 1/6, a 3.5% weight ratio catalyst vs. oil (Fig. 6). The process intensification induced by the MW-assisted transesterification was also observed for the activated catalysts. Thus, FAME yields of over 80% were obtained for all catalysts with the yield reaching 91% in the case of Fe 3 O 4 @ SiO 2 -DIG. The catalytic activity variation of both MW-assisted transesterification reaction using the activated catalysts followed the same trend Previous examples of guanidines derivatives use as catalysts for vegetable oil transesterification described high reaction yields in homogeneous 50 or heterogeneous systems 32,35 , but the reuse of heterogeneous catalysts usually provided a challenge due to the leaching 18,30,31,34,51 from the support and require covalent bonding to the support 37 . In our case, through the simultaneous hydrolysis/polycondensation of both TEOS and of the organosilane, the chemical attachment of the guanidine derivatives to the catalysts structures is assured, therefore the leaching of the superbase problems should be eliminated. To confirm this, catalyst reutilization experiments were performed under MW irradiation for the process intensification (Fig. 7). The results show that during five reaction cycles the FAME yield decreases gradually from 83.96%, 90.95%, and 86.12% to 76.9, 83.6%, and 80.7% for Fe 3 O 4 @SiO 2 -EDG, Fe 3 O 4 @SiO 2 -DIG, and respectively Fe 3 O 4 @SiO 2 -DCG. This decrease in the FAME  Considering possible differences in the absorption characteristics of the catalyst our next objective consisted in the investigation of the catalytic activity in the case of a starting material with a higher content of free fatty acids (FFA). The capability of a catalyst to be employed for the transesterification of oils with a higher FFA content is highly desirable when considering as starting material waste cooking oils. The results presented in Fig. 9 indicate a dramatic drop of the FAME yield with the increase of FFA content in the case of Fe 3 O 4 @SiO 2 -DIG and Fe 3 O 4 @SiO 2 -DCG. However, in the case of Fe 3 O 4 @SiO 2 -EDG, a higher tolerance towards the FFA content can be noticed with a FAME yield of 67.38% (for a 3.5% (weight) vs oil catalyst content) at a 2% weight FFA content. This activity can be correlated with the presence of a tertiary amine in addition to the guanidine functionality in the case of Fe 3 O 4 @SiO 2 -EDG. Thus, the functional groups present in Fe 3 O 4 @SiO 2 -EDG make possible the existence of an intramolecular hydrogen bond and intrinsic proton affinity characteristic in the case of Fe 3 O 4 @SiO 2 -EDG 52 .  www.nature.com/scientificreports/ To further investigate the activity of Fe 3 O 4 @SiO 2 -EDG at high FFA content in the starting material, experiments with different concentrations of catalyst were performed (Fig. 10). The results confirm that increasing the catalyst concentration leads to a higher FAME yield. However, at a higher concentration of FFA, the catalyst activity also decreases affording FAME yield below 10%. Thus, although the tertiary amine helps at extending the tolerance of the catalyst the intramolecular hydrogen bond at high FFA content leads to a deactivation of the catalyst active sites. However, this deactivation is reversible and by washing with alkaline methanol solution the catalyst can be reactivated with comparable results.

Conclusions
In conclusion, this study presented the synthesis, characterization, and catalytic activity evaluation of three heterogeneous catalysts based on guanidine derivatives both in conventional and MW-assisted transesterification reaction for biodiesel production. The catalyst consisted of magnetic particles encapsulated in a silica shell containing the guanidine derivative for their facile recovery from the reaction media. The catalysts were characterized by FTIR, TGA, SEM, BET and XRD analysis confirming the successful covalent attachment of the organic derivatives on the silica shell.
The catalytic activity evaluation aimed to confirm the good activity of the catalysts with FAME yields of over 70% being obtained at an oil/methanol ratio of 1/6, at 80 °C and 180 min reaction time using conventional heating. The activity of the catalysts showed improvement and good stability in MW-assisted transesterification reactions, all catalysts displaying FAME yields of over 80% after 5 reaction/activation cycles. The catalytic activity variation was in the order: Fe 3 O 4 @SiO 2 -DIG > Fe 3 O 4 @SiO 2 -DCG > Fe 3 O 4 @SiO 2 -EDG. Thus, the catalysts displayed good activity, stability, and reusability at a relative low reaction temperature for heterogeneous catalysts and low oil/methanol molar ratio. Furthermore, the influence of FFA content on the catalytic activity was . FAME yield depending on the FFA (wt% vs. oil) content and catalyst type (activated catalyst; temperature 80 °C (MW irradiation); reaction time 180 min; molar ratio oil/MeOH 1/6; 3.5% weight ratio catalyst vs. oil; 780 rpm). www.nature.com/scientificreports/ investigated. Thus, in the case of Fe 3 O 4 @SiO 2 -EDG, a higher tolerance towards the FFA content can be noticed with a FAME yield of over 90% (for a 5% (weight) vs oil catalyst content) and 5% weight FFA content. Thus, the combination of a guanidine functional group and tertiary amine in EDG allows for a higher tolerance towards FFA content. Thus, the facile synthesis procedure for the catalysts developed as well as their activity, stability, recyclability, and capability to process less expensive oil feedstocks and permit the improvement of biodiesel techno-economic parameters.