Microwave-assisted conversion of novel biomass materials into levulinic acid

Levulinic acid is considered one of the most important platform chemicals. It is currently produced mainly from lignocellulosic biomasses. However, there are also other abundant biomass materials, which could be used as raw materials for levulinic acid production. In this work, levulinic acid was produced from two novel biomasses in the presence of Brønsted (H2SO4) and Lewis acid (CrCl3·6H2O or AlCl3·6H2O) catalysts. The studied materials were carbohydrate-rich potato peel waste and sporocarps of the fungus Cortinarius armillatus. Reaction conditions, i.e., time, temperature, H2SO4, and Lewis acid concentrations, were studied by utilizing full 24-factorial experimental designs. Microwave irradiation was used as the heating method. Based on the results, the reaction temperature and the H2SO4 concentration had the greatest impact on the yield of levulinic acid. The highest yield obtained in this study from potato peel waste was 49% with 180 °C for reaction temperature, 15 min for reaction time, and 0.5 and 0.0075 M for the concentrations of H2SO4 and CrCl3, respectively. When Cortinarius armillatus was used as the raw material, the highest yield was 62% with 180 °C for reaction temperature, 40 min for reaction time, and 0.5 and 0.0075 M for the concentrations of H2SO4 and CrCl3, respectively.


Introduction
Most of the industrial chemicals are currently being prepared from fossil resources. However, the increase in the fossil fuel price as well as the depletion of the resources is driving forward the search for new and alternative renewable feedstocks in the production of renewable platform chemicals, which could replace the petroleum-based ones [1]. One of the most important platform chemicals is levulinic acid (LA) [2,3]. Due to its chemical structure with ketone carbonyl and carboxylic functional groups, LA can be converted into various other important chemicals, and hence, can be used as a raw material for, e.g., resins, plasticizers, textiles, animal feed, coatings, antifreeze, fuel additives, polymer precursors, herbicides, pharmaceuticals, and flavor substances. Since new applications of LA are being constantly explored, the demand for it is expected to grow [4].
Presently, the research is mainly concentrated on the production of LA from lignocellulosic biomass [5][6][7][8]. However, there are other abundant, novel carbohydraterich materials, which could be utilized as well. Such materials include potato peel waste (PW) and sporocarps (S) of the fungus Cortinarius armillatus. PW is generated annually with considerable amounts by food processing industry. One peeling factory can process up to 1000 tons of potato per year and, depending on the peeling process, the amount of waste is 15-40% of the amount of processed potatoes. Regarding the sporocarps of edible macrofungi, it has been estimated that up to 15,000 DW tons of them are produced annually in the forests of Northern Finland [9] and the production of non-edible macrofungi even exceeds that. C. armillatus is one of the most productive non-food macrofungi in Northern Finland with a long term average sporocarp yield of 0.3 kg DW/ha in all forest types, while maximum yield is reported to be 4.7 kg DW/ha in all forest types [10].
In this work, we have studied the conversion of PW and sporocarps of Cortinarius armillatus into LA. Both studied materials are cheap and have few other usages. In order to accelerate the conversion reactions, microwave irradiation was used as the heating method. It has been found in previous studies with cellulose that besides accelerating the conversion reactions, microwave heating also enhances the product selectivity [11,12]. There is also a significant, up to 85-fold, energy saving involved in the microwave-assisted processes [13]. The effects of reaction conditions on the yield of LA were studied in detail by utilizing experimental design. To our knowledge, neither PW nor sporocarps from Cortinarius armillatus has been used as the raw material in LA production.

Reagents
PW, produced by abrasion peeling process, was provided by Tervakankaan Peruna Oy, Finland. PW was dried (105°C until constant weight) and ground into fine powder before use. The original water content of PW was 80%. All reactions were done using the same batch of PW. The Cortinarius armillatus sporocarps were collected in 2007 in a subarctic forest with mountain birch and Scots pine at Lapland Research Station, Kevo, Finland, dried and ground into fine powder with a homogenizer. Other reagents, i.e., sulfuric acid (95-97%, Merck), AlCl 3 · 6H 2 O (99%, Alfa Aesar), or CrCl 3 ·6H 2 O (98%, Alfa Aesar) were used as received from the suppliers.

Conversion of potato peel waste or sporocarps into levulinic acid in microwave reactor
In a typical experiment C. armillatus sporocarps or PW (0.5 g) was weighted into a microwave reactor vessel (size 2-5 ml) equipped with a magnetic stirring bar.  Tables 1 and 3). After the reaction, a sample (1 ml) was taken from the reaction mixture, filtered with a syringe filter and analyzed with HPLC-PDA.
The energy input for the conversion reactions was calculated with Eq. 1:

HPLC analysis
The LA concentration of the samples taken after each reaction was analyzed with high-performance liquid chromatography (HPLC; Waters 2695 Separation module) fitted with an Atlantis dC18 (5 μm, 4.6 × 150 mm) column and a photodiode array (PDA) detector (Waters 996). Water:methanol (90:10) mixture with 0.1% (v/v) of TFA was used as the mobile phase with a flow rate of 1 ml/min. The injection volume was 2 μl. The column temperature was kept constant at 30°C and the calibration was performed using LA analytical standard (Sigma-Aldrich). The UV detection was done at 267 nm for LA.
The yield of LA was calculated with Eq. 2: The theoretical maximum yield of LA was calculated from the total carbohydrate content of the starting material, which, according to the supplier, was 80% for the dry PW. No data on the carbohydrate content of C. armillatus sporocarps is available but the carbohydrate content of 48.6% DW has been reported for the sporocarps of mixed Cortinarius species [14]. However, a high yearly variation occurs in the general sporocarp production, and the composition of the main constituents of sporocarps varies. This can be partially due to varying methodology and conversion factors used, typically overestimating protein and underestimating carbohydrate contents [15]. In a review on 11 macrofungal species in Agaricales studied, maximum carbohydrate content of 75% DW for the sporocarps was measured [15]. In another European study on six species of Agaricales [16], maximum carbohydrate content of 71.2% DW was measured. In order not to overestimate the LA yield, the total carbohydrate content of 80% was used for the sporocarps in this study.

Experimental design
The full 2 4 -factorial design was chosen as the experimental design in order to study the effect of the reaction conditions on the conversion of sporocarps and PW into LA. The factors (time, temperature, H 2 SO 4 , and Lewis acid concentration) and their levels used in the experiments are given in Table 1. The factorial design consisted of four factors with two levels (high and low), including three center points. The levels for the factors were chosen based on the literature [17,18] and some preliminary experiments. They were also selected to be moderate but different enough from each other. Nineteen experiments were carried out including three replications determined at the center point of the design in order to obtain the estimate for the experiment uncertainty. All experiments were carried out in a random order and LAyield (%) was used as the response. Once the yields were gained, the data was fitted using the multiple linear regression method in MODDE 9.1 (Umetrics) computer software. During modeling, the data points, which were considered as outliers (set 1, exp. 3, Table 2), were excluded to improve the model. The decision was based on the normal probability plot of residuals. The statistical validation was determined using the ANOVA test at a 95% confidence level.
Based on the results from the factorial experiments, some additional reactions (eight reactions) were performed (Table 3). In these reactions, the reaction temperature was set for 180°C and the concentration of H 2 SO 4 for 0.5 M.   3 Results and discussion

Full 2 4 -factorial designs
In this study dried and ground PW and Cortinarius armillatus sporocarps were used in the production of LA. The full 2 4factorial experimental design was first used to study the effect of various factors on the yield of LA, which was used as the response in the design ( Table 2). According to the analyzed data, time, temperature, and the concentration of H 2 SO 4 had a statistically significant effect on the LA yield in all sets. This can also be seen from the data presented in Table 2 and Supplementary material (Tables S1, S3, and S5). The LA yields generally increase, when time, temperature, and H 2 SO 4 concentration increase (see also Figures S7-S9). The concentration of Lewis acid catalyst (CrCl 3 ) had a statistically significant negative effect on the LA yield, when sporocarp powder was used as the starting material. On the other hand, no effect was found when CrCl 3 and AlCl 3 catalysts were used for PW or sporocarp powder in sets 1 and 3, respectively. All sets contained an interaction term between temperature and H 2 SO 4 concentration. In addition, there was an interaction term between time and temperature as well as a square term of time in the model of set 1 and an interaction between H 2 SO 4 concentration and AlCl 3 catalyst in the model of set 3 (see Supplementary material, Tables S1-S6).
Based on the obtained models from the factorial designs, it was concluded that the reaction temperature and the concentration of H 2 SO 4 had the greatest impact on the yield of LA. PW seemed to react faster than sporocarps since for PW, the highest LA yield (49%, exp. 7; Table 2) was achieved in 15 min at 180°C with H 2 SO 4 concentration of 0.5 M and the concentration of the CrCl 3 catalyst 0.0075 M. For sporocarps, the highest LA yield, 46% or 53%, with CrCl 3 or AlCl 3 catalyst, respectively, was achieved at 180°C in 60 min with 0.5 M for H 2 SO 4 concentration and 0.0075 M for the concentration of the catalyst (exp. 8; Table 2).

The effect of Lewis acid catalyst and the reaction time on levulinic acid yield
It was found intriguing that according to the factorial experiments, the concentration of the Lewis acid catalyst did not have a positive effect on the yield of LA in any of the sets. In studies with glucose, the Lewis acid catalysts have been found to improve the conversion of biomass into LA, since they catalyze the glucose to fructose isomerization step, which is necessary for the conversion reaction to occur [1,19]. The exact route for the CrCl 3 catalyzed isomerization of glucose to fructose is not known but based on the recent literature, a plausible route is presented in Fig. 1 [20,21]. Therefore, the effect of the Lewis acid was studied in more detail with some additional reactions (Table 3). First, the abovementioned experiments, 7 and 8 for PW and sporocarps, respectively, were repeated without the Lewis acid catalyst. For PW, the LA yield was 38% and for sporocarps, 25% (exps. 20 and 21; Table 3). The decreased yields indicated that the additional catalyst had some effect on the LA yield. Next, the effect of the Lewis acid catalyst was studied with reduced amounts, 0.0012-0.0086 M (Table 3), of catalyst compared to the amounts used in the full factorial designs (0.0075-0.03 M; Table 2). The amount of the Lewis acid was reduced since in the full factorial designs, higher LA yields were achieved with small amounts of the Lewis acid. Also, the reaction time was varied since its effect on the LA yield was not as apparent in the full factorial designs as that of the reaction temperature or H 2 SO 4 concentration. The studied reaction times were 40 and 50 min. Seeing that with PW, the highest LA yield was reached already in 15 min in the factorial experiments, these additional reactions were only performed with sporocarps. Also, the reactions were performed only with CrCl 3 , since the results achieved for sporocarps with AlCl 3 and CrCl 3 in the full factorial designs were quite similar. The LA yields as well as the reaction conditions for each additional experiment are presented in Table 3. Based on the results, the LA yield increased with the increasing amount of CrCl 3 , when the reaction time was kept at 50 min (exps. 22-24; Table 3). This indicated that the Lewis acid concentration had some effect on the LA yield. In addition, the reaction time had some effect on the LA yield, i.e., with shorter reaction time, 40 min, the LA yield was higher (exps. 25 and 26; Table 3). This indicated that LA started to decompose with prolonged reaction times. The highest LA yield (62%, exp. 26; Table 3) in the study was achieved with the sporocarps at 180°C in 40 min with 0.5 M for the concentration of H 2 SO 4 and 0.0075 M for the CrCl 3 catalyst. Finally, the reaction conditions, which were found optimal for sporocarps, were also used for PW conversion into LA (exp. 27; Table 3). However, with PW, the LA yield was only 38% when reaction time was increased into 40 min. The result verified that for PW, the optimal reaction time is shorter than for sporocarps (15 min, exp. 7; Table 2).
Overall, based on the results, PW and Cortinarius armillatus sporocarps proved to be excellent starting materials for the LA production. Previously reported results for the LA yields include, for example, Weiqi and Shubin [19] with 54% LA yield from glucose at 170°C in 4 h, Mukherjee and Dumont [22] with 55% LA yield from corn starch at 180°C in 15 min, Jeong [18] with 25% LA yield from glucosamine at 188°C in 49 min, and Shen et al. [23] with 39% LA yield from cellulose in 2 h. It should be noted that in the abovementioned studies, the starting materials were pure carbohydrates; where as in this study, biomass was used as received without prior separation of carbohydrates. Thus, the LA yields achieved in this study are highly comparable with the yields reported in literature.

Microwave irradiation as the heating method
In this study, the maximum power output of the microwave reactor was set to 90 W in order to inhibit the overheating of the reaction solutions and also to maintain similar heating conditions between individual reactions. However, the full 90 W of power was required only to reach the reaction temperature, which took 4-5 min, depending on the said temperature. Once the temperature was reached, the power output of the reactor was quite constant, 40-50 W, corresponding to the reaction temperature of 140-180°C, respectively. The energy input for each reaction was calculated with Eq. 1 and was found to vary from ca. 36 to 180 kJ. For reactions, which produced the highest yields in this study from sporocarps or PW (exp. 27; Table 3 or exp. 7; Table 2, respectively), the energy input was ca. 120 or 45 kJ, respectively. Microwave irradiation is a non-contact heating due to which energy is not lost to the heating of the reaction vessel or, e.g., an oil bath. Instead, energy transfers directly to reacting compounds, which ensures rapid heating and low-power consumption [13]. To verify the superiority of microwave irradiation over conductive heating, the power consumption of an oil bath heating was also investigated by heating a laboratory-scale oil bath to temperature required by the conversion reaction. However, in this study, it proved to be impossible to heat the oil bath to 180°C. During 60 min of heating with the maximum power of the heat source (600 W), the oil temperature remained at 175-178°C.

Catalyst recovery
Catalyst recovery and reuse is an important point to consider, when using Lewis and Brønsted acid catalysts in biomass conversion reactions. In this study, some preliminary experiments regarding the recovery were performed with PW as the feedstock. Reaction temperature was 180°C, reaction time 30 min, H 2 SO 4 concentration 0.5 M, and CrCl 3 concentration 0.0075 M. The first reaction provided LA yield of 43%. LA was extracted with ethyl acetate and the filtered aqueous reaction liquid, containing the CrCl 3 and H 2 SO 4 catalysts, was used again in the conversion reaction of new batch of PW. Second cycle provided LA with the yield of 46%. LA was removed again by extraction and the reaction liquid was used for the third time. The third cycle gave LA yield of 42%. The results are in accordance with Havasi et al., who studied the recycling of H 2 SO 4 used in the conversion of household waste into LA [24]. However, more detailed catalyst recovery and reuse experiments are ongoing.

Conclusions
In this study, carbohydrate contents of PW and Cortinarius armillatus sporocarps were efficiently converted into LA with microwave irradiation as the heating method. Reaction conditions were studied by utilizing experimental design. Based on the results, the reaction temperature and the H 2 SO 4 concentration were found to have the greatest impact on the yield of LA. The Lewis acid concentration and reaction time had also some effect on the LA yield. The highest LA yield was 62% achieved with fungal sporocarps as the raw material in 40 min at 180°C, with H 2 SO 4 concentration of 0.5 M and CrCl 3 concentration of 0.0075 M. Experiments regarding the use of other novel biomasses such as other fungal species are ongoing.