Reusability and regeneration of solid catalysts used in ultrasound assisted biodiesel production

Reusability of two heterogeneous catalysts in ultrasound (US) assisted biodiesel production was investigated in comparison to each other. An ultrasound (US) generator (200 W, 20 kHz) equipped with a horn type probe (19 mm) was used. Regeneration experiments were planned according to second order central composite design (CCD) method. After the eighth use of the catalysts, biodiesel yield decreased from 99.1% to 90.4% for calcined calcite (CaO) and from 98.8% to 89.8% for calcined dolomite (CaO.MgO). Furthermore, regeneration of spent catalysts by calcination was investigated; optimum temperature and time were found as 750 °C and 90 min, lower than fresh catalyst preparation conditions. The regenerated catalysts were reused in a second process cycle; biodiesel yield was calculated as 97.2% for CaO and 96.5% for CaO.MgO. Finally, the process showed that calcination is an energetically favorable regeneration process of spent catalysts.

catalyst, KF/CaO-Fe 3 O 4 , using an impregnation method for biodiesel production and reused the catalyst 16 times [13]. Yu et al. synthesized different compositions of CaO-CeO 2 mixed metal oxides for using solid base catalysts and found that the biodiesel yield decreased slowly until the fifth reuse when a significant loss in the catalytic activity was observed [14].
On the other hand, catalyst regeneration methods can be classified into two groups. In the first group, after catalyst filtration, regeneration is accomplished by washing and drying the catalyst at a temperature usually lower than 200 °C. In the second one, calcination alone or combined with other methods is applied [6]. Madhu et al. found that reused catalysts must be calcinated in order to activate the catalyst active sites [15]. Boey et al. reused the waste cockle shell derived CaO catalyst at least for three times, with a purity above 96.5%. Before it was reused, the spent catalyst was washed with methane and n-hexane to remove the adsorbed materials and calcined at 900 °C for 2 h. The washed uncalcined spent catalyst consisted of Ca(OH) 2 and traces of Ca(C 3 H 7 O 3 ) 2 (calcium diglyceride, CaDG). After calcination at 900 °C for 2 h, the structure of the calcined dolomite catalyst was changed to CaO [16].
The literature survey shows that calcination is a more appropriate method to regenerate the activity of CaO and Ca based catalysts; it is a simple process with no waste solvent or solid waste product formation, also the heat of combustion of some organic compounds formed during the esterification such as CaDG or adsorbed impurities on the spent catalyst may diminish the heat duty of the calcination process.
By considering these points, calcination was selected in this study for catalyst regeneration, which it is expected to be more energetically favorable process compared to fresh catalyst preparation.

Catalyst reuse experiments
The schematic representation and details of the experimental setup were given in a previous study [1]; an ultrasonic generator (Bandelin 2200 sonopuls, 200 W, 20 kHz, Sigma-Aldrich, city, country?) equipped with a horn type probe was used to deliver pulsed ultrasound (US) with controllable power in a 300 mL three-necked cylindrical glass reactor equipped with a reflux condenser and a magnetic stirrer.
The experiments were carried out at the optimum conditions determined in this study. At the end of the run, the reaction slurry was poured into specially designed bottom screwed centrifuge tubes, which allowed the catalyst recovery almost completely? Until the next run, the catalyst was kept in refrigerator in CH 3 OH to avoid contact with atmospheric CO 2 and H 2 O. The next run required CH 3 OH for the desired feed CH 3 OH/oil molar ratio, and oil was added to the sonoreactor with some fresh catalyst to compensate for the catalyst loss (approximately 5%).

Catalyst regeneration experiments
The purpose of these experiments was to find calcination conditions energetically (and economically) favorable than the fresh catalyst production by calcination from raw materials. Catalysts (obtained after the eighth use) were divided into samples of 1-g after washing with methanol. Catalyst regeneration experiments were carried out using 1-g catalyst samples.
The statistical second order central composite (CCD) design was used in planning of regeneration experiments using Design-Expert software (demo version, Stat-Ease Inc, MN, USA?). In addition, temperature and time were selected as process factors in the experimental plan; as seen in Table 1, positive levels of factors were determined from fresh catalyst preparation conditions, and the negative levels were determined from similar catalyst regeneration studies cited in the literature [18,19].
The second order central composite (CCD) experimental plan is given in Table 2. Eleven regeneration experiments were made for this purpose. The response of the experimental plan was the mass loss (%) of the spent catalyst at the end

Catalyst reuse experiments
The serial experiments were carried out until the yield decreased to approximately 90%. Experimental results are depicted in Figure 1. Different heterogeneous catalyst deactivation mechanisms are proposed in literature; Oueda, et al. pointed out especially to the leaching of active sites on the catalyst surface, the surface poisoning and/or pore filling, and the structural collapse of catalysts [6]. To detect phase transformation during the process, X-ray diffraction (XRD) patterns of fresh and spent catalysts with crystal phases are shown in Figure 2 (1) CaDG is less active transesterification catalyst than CaO catalyst. Furthermore, CaDG is more soluble in the glycerol phase than CaO. Also, water reacts with CaDG according to reaction in Equation (2).
Furthermore, Ca 2+ reacts with the free fatty acid (FFA) in the oil and forms calcium soaps Ca(FFA) 2 according to the reaction in Equation (3), which results in the reduction of the reaction yield and biodiesel purity.

Catalyst regeneration experiments
The XRD data was used to check the success of calcination procedures and synthesis of CaDG. As seen in Figure 2, sharp and highly intense XRD peaks define the well crystallized structure of the regenerated spent catalysts. Because, the crystallinity or crystallite size of the catalyst is an important feature determining the catalytic activity. Therefore, the crystallite size for all phases was obtained from the XRD patterns using the Scherrer equation where, L is the mean average crystallite size (Å), α is a constant equal to 0.94, β is the full width at the half maximum in radians (obtained for the high intensity peak), and λ (Å) is the wavelength of the X-rays (  Table 3. When CaDG is completely oxidized to CaO, a mass loss of 74.7% occurs according to the reaction in Equation (5).    In the second cycle of the biodiesel production with regenerated catalyst, yield was realized as 97.2% for CaO. For dolomite catalyst the same calcination conditions was used for regeneration step and 96.5% yield was obtained.
Finally, for an overall assessment of the process, biodiesel production per kg of catalyst (BP) was calculated by Equation (6), where m oil is amount of canola oil (kg), α is the stoichiometric correction factor, x is the biodiesel content of product and m catalyst is the amount of used catalyst (kg). The cumulative biodiesel production per kg of catalyst value (CBP) is obtained by accumulating the BP values. The CBP results are given in Figure 3. As seen, the CBP values of both catalysts are very close to each other. and CaO (silent) respectively. The cumulative productivity rate of US assisted biodiesel production is 8% higher than the silent biodiesel production, but for a more detailed energy analysis of the biodiesel production process, the comparison may be more accurate by taking into account ultrasound energy used during the reaction.

Conclusion
The reusability performances of calcined calcite (CaO) and calcined dolomite (CaO.MgO) in the ultrasound assisted transesterification reactions were tested in successive transesterification runs. The yields decreased from 99.1% and 98.8% for CaO and calcined dolomite to 90.4% and to 89.8% after eighth reuse, respectively. These results are consistent with the average crystallite size results and also with XRD patterns.
The regenerated catalysts were reused in a second process cycle with negligible activity losses compared to fresh catalysts. Furthermore, an overall assessment of the process by means of various criteria shows that the performance of ultrasound assisted biodiesel production using CaO catalyst is 8% higher than the silent biodiesel production.
While the use of ultrasound in biodiesel production decreases the process cost and process time, the eighth reuse of calcined calcite and calcined dolomite catalysts reduces the energy requirement in catalyst preparation.
In conclusion, the experimental results confirmed that spent catalyst calcination at 750 °C for 90 min. is more energetically favorable when compared to fresh catalyst preparation at 840 °C for 3 h. Finally, reusability of heterogeneous catalyst is an important issue affecting technical, environmental and economic aspects of the biodiesel production process.

Acknowledgment
We thank the Gebze Technical University research fund for partial support.