Selectively production of levoglucosenone from 1 cellulose via catalytic fast pyrolysis 2

Levoglucosenone (LGO) has a wide range of utilization in the field of organic synthesis. Magnetic 14 solid acid (Fe 3 O 4 /C-SO 3 H 600 ) was used in fast pyrolysis of cellulose to produce LGO. It was 15 demonstrated that the catalyst could promote the pyrolysis of cellulose to produce LGO, but the 16 yield was affected by the pyrolysis temperature and the relative amount of catalyst. The yield of 17 LGO reached 20.0 wt% from catalytic fast pyrolysis of cellulose at 300 °C, which was 18 significantly higher than that from cellulose (0.3 wt%). Furthermore, the kinetic analysis and 19 recycling results showed that the catalyst could not only reduce the required temperature of 20 cellulose in fast pyrolysis, but also still efficiently promote the production of LGO after recovery 21 and activation.


25
A great interest burgeons in renewable biomass as a source for fuel and 26 chemicals production for the demands to replace alternatives to fossil fuels (Jiang et  components in bio-oil obtained from the traditional pyrolysis process is relatively low, 35 and it is currently uneconomical to separate special compounds from bio-oil. There is 36 no doubt that bio-oil production with a high content of target products is essential LGO significantly, high expenses make them unavailable in large-scale applications 54 (Kudo et al., 2017). Cellulose impregnated with liquid inorganic acids (e.g. 55 phosphoric acid, sulfuric acid) is also performed to promote LGO production (Dobele 56 et al., 1999). However, this process shows several shortcomings: the impregnation of 57 cellulose and acid requires a complicated pretreatment process; a large amount of 58 waste acid is produced after the impregnation, which cannot be recycled and reused.

59
In contrast, solid acid is thermally stable and can be readily recovered after pyrolysis.

60
There is no need to go through a complicated impregnation process and only simple 61 mechanical mixing with the cellulose is required. Therefore, solid acid is a better   Regarding this procedure, it has been reported that both Fe3O4 and acid catalyst could 72 enhance the dehydration behavior and promote the yield of LGO in catalytic fast 73 pyrolysis cellulose (Lu et al., 2014a;Halpern et al., 1973). In this study, magnetic 74 cellulose to produce LGO. Different from previous reported investigations, this 76 research could not only efficiently promote the conversion of cellulose to LGO, but 77 also analyzed its kinetics and reaction mechanism in detail, and investigated the 78 recovery and reuse performance of the catalyst.

Catalyst preparation 87
The preparation of the catalyst required two steps: carbonization and sulfonation.    (1) 116 Due to the existence of the kinetic compensation effect, the activation energy, the 117 pre-reference factor, and the mechanism function were related to each other.

118
Assuming that the pre-reference factor was a constant, the activation energy was only 119 related to the mechanism function, as shown in formula (2).
where α is the conversion rate, %; β is the heating rate, °C/min; E is the 122 activation energy, kJ/mol; A is the pre-exponential factor, s -1 ; Ψ (E, T) is the 123 temperature integral:

Fast pyrolysis 126
The sample was subjected to fast pyrolysis in a semi-batch CDS reactor (CDS 5200, The SEM images of the catalysts are shown in Fig. 1. It could be seen that the 149 surface of Fe3O4/C-SO3H500 and Fe3O4/C-SO3H600 contained countless pores. This 150 facilitated the contact of the catalysts with cellulose. Although there were many pores 151 on the surface of Fe3O4/C-SO3H700, it was very messy, probably due to the excessive 152 temperature destroying the pore structure on the surface of Fe3O4/C-SO3H700. Among 153 the three catalysts, the surface area of Fe3O4/C-SO3H500 (1.8 m 2 /g) was larger than that 154 of Fe3O4/C-SO3H600 (0.7 m 2 /g) and Fe3O4/C-SO3H700 (0.6 m 2 /g) ( Table 1). This might 155 be caused by the higher calcination temperature causing Fe3O4 to be embedded in the 156 pores.

157
The XRD peaks of the three catalysts were all identified as the crystalline phase 158 of Fe3O4 (Fig. 2), which indicated that Fe3O4 remained stable after calcination and 159 sulfonation during the preparation of catalysts. In FTIR analysis, the absorption peak Fe3O4/C-SO3H600 and Fe3O4/C-SO3H700 was that catalyst Fe3O4/C-SO3H600 appeared-167 SO3stretching and O=S=O stretching in -SO3H at bands 1061 cm -1 and 1223 cm -1 , 168 respectively. This showed the presence of -SO3H functional group in Fe3O4/C-169 SO3H600. NH3-TPD characterization also indicated that the acid sites of Fe3O4/C-170 SO3H600 (0.7 mmol/g) were higher than that of Fe3O4/C-SO3H500 (0.1 mmol/g) and

171
Fe3O4/C-SO3H700 (0.5 mmol/g) (Table 1).    The TG analysis of pure cellulose and cellulose mixed with the solid acid 186 catalyst is shown in Fig. 4 and Table 2. The main weight loss area of cellulose was  Table 3. E referred to the 202 global activation energy, that was, the energy that needed to be absorbed when the   Ti is the temperature of pyrolysis at which the main weightless zone begins, 5%. 223 Tt is the temperature of pyrolysis at which the main weightless zone finishes, 95%. 224 Tmax is the temperature corresponding to the highest point of the DTG peak. 225 Dmax is the weight loss rate corresponding to the highest point of the DTG peak. 226   LGO, which should be further analyzed. The higher the yield and relative content of

242
LGO was, the more obvious the decrease in the yield and relative content of LG The distribution of pyrolysis products could vary with temperature as the 265 dominant reaction pathways shifted in rate and overall extent (Maduskar et al., 2018). the cellulose pyrolysis process at 500 °C, and make the product distribution 296 concentrate on the anhydrosugars LG, LGO, and DGP (Fig. 8d). The formation of 297 LGO was mainly through glycosidic bond cleavage reactions in which cellulose was 298 depolymerized and then exposed on dehydration reactions of the pyran ring. The 299 dehydration process was the main reactions that affected the formation of LGO, 300 mainly occurred at low temperatures, and the cleavage of glycosidic bonds mainly 301 occurred in the medium temperature. As the pyrolysis temperature increased, the rate 302 of cellulose glycosidic bond cleavage reaction was accelerated. When it was faster 303 than the dehydration rate, only a small part of the glycosidic bond cleavage product 304 of LGO production at 300-500 °C in catalytic pyrolysis. These results also 306 demonstrated that Fe3O4/C-SO3H600 could not only promote the formation of LGO, 307 but also reduced the pyrolysis temperature of cellulose.  The ratio of Fe3O4/C-SO3H600 to cellulose is also an important factor affecting 321 the distribution of pyrolysis products. With the increase in the amount of catalyst, the 322 yield and relative content of LGO showed a trend of initially increasing followed by a 323 decrease, while the yield and relative content of LG continued to decrease as the 324 amount of catalyst increased (Fig. 9). This showed that when the amount of catalyst 325 was increased, the conversion of LG to LGO was promoted. The best ratio of 326 Fe3O4/C-SO3H600 to cellulose was 1:1, in which the yield and relative content of LGO 327 were 20.0 wt% and 50.7%, respectively. When the ratio of Fe3O4/C-SO3H600 to 328 cellulose rose from 1:3 to 1:1, the acidic sites of the catalyst gradually increased, and 329 the cellulose was able to completely contact the catalyst, so the yield of LGO 330 continued to rise. When the ratio of Fe3O4/C-SO3H600 to cellulose increased from 1:1 331 to 3:1, the relative content of LGO changed a little, but the yield of LGO decreased to  In order to evaluate the stability of the catalyst, the reuse of Fe3O4/C-SO3H600 for 346 LGO production was repeatedly conducted under the optimum conditions of 300 °C 347 and the ratio of cellulose to catalyst of 1:1. As shown in Fig. 10, the yields of LGO in 348 the three cycle experiments were 14.1 wt%, 11.6 wt%, and 8.3 wt%, respectively.

356
Compared with catalysts such as liquid acids and ionic liquids, Fe3O4/C-SO3H600 357 could not only efficiently catalyze the production of LGO from cellulose, but also 358 could be recycled and reused, which was more economical and practical.  According to currently proposed views, there were two main ways to produce 363 LGO. First, cellulose was pyrolyzed and depolymerized to produce LG, and LG was 364 further dehydrated to produce LGO. Second, cellulose was first dehydrated during 365 pyrolysis, and then the glycoside bonds at both ends were broken to form LGO 366 (Zhang et al., 2017;Lu et al., 2011). The latter process did not produce LG. In this 367 experiment, since the increase in LGO yield was consistent with the decrease in LG 368 yield, it was speculated that Fe3O4/C-SO3H600 could promote further dehydration of 369 LG to produce LGO. To verify it, pure LG was mixed with Fe3O4/C-SO3H600 for fast 370 pyrolysis at 300 °C. The typical ion chromatogram was shown in Fig. 11. The results 371 showed that the presence of Fe3O4/C-SO3H600 could indeed catalyze the dehydration 372 of LG to produce LGO and DGP. It is noticeable that although LGO and DGP were 373 both products of LG's further dehydration, the increase in the yield of LGO was 374 significantly more than that of DGP. This showed that the catalytic effect of Fe3O4/C-375 SO3H600 was also selective. After observing the chemical structure of the three 376 anhydrosugars, it could be found that LG and LGO had similar bicyclic structures