Mechanocatalytic partial depolymerization of lignocellulosic feedstock towards oligomeric glycans

: The depolymerization of lignocellulosic feedstock with a heterogeneous composition is a major challenge and usually leads to the formation of monosaccharides as main products. Our work aims to convert such feedstock into oligomeric glycans as more valuable products compared to sugars, by using mechanocatalysis in a planetary ball mill in a cost-efficient and resource-saving manner. Herein, we utilized raw materials such as wheat straw, beet pulp, cocoa shells and apple pomace as residual natural raw materials from food and feed production. Reaction parameters such as rotational speed, acid content and milling duration were investigated and optimized towards a maximum amount of soluble species and a minimum of monosaccharides. The optimization for cellulose as substrate resulted in a nearly full-soluble fraction containing oligomeric glycans. Based on these results the reaction parameters were transferred and further optimized for lignocellulosic feedstock. For wheat straw a solubility of over 90 % was achieved comprising a mixture of oligomeric glycans as well as partially depolymerized lignin.


Introduction
In the last decades mechanochemistry has advanced quickly and has been acknowledged as one of the top ten emerging technologies in chemistry. [1]One field of mechanochemistry is mechanocatalysis where a suitable catalyst for a reaction is added to the reactants.For these reactions many different reactor types such as vibration mills, mortar mills and rotary mills can be used. [2]In terms of green chemistry, this approach has the advantage that they are usually solventfree while there are also approaches with solvents, the so-called liquid assisted grinding (LAG). [3]chanochemistry is probably most known for processing inorganic materials such as alloys and intermetallic phases.Typical examples are the synthesis of oxides, nitrides and halides where long reaction times up to several days are required. [4]Besides the synthesis of inorganic materials there are also rising numbers of mechanocatalytic organic synthesis e.g.Knoevenagel-condensations and polymerizations of insoluble polymers whereby the synthesis of thermoplastics is not favored due to the released heat during the process. [2,5,6] I some cases, this typical solvent-free approach enables activation only due to mechanical forces, e.g.[13][14] Cellulose and lignocellulosic feedstocks are commonly insoluble and therefore difficult to process. [15]It is known that by mechanocatalytic processing of cellulosic biomass with a strong acid a soluble product mixture consisting of glycans (glucans and xylans), as well as partially depolymerized lignin is obtained. [16,17] his is an advantage compared to classic wood saccharification since the more valuable glycans instead of monomeric sugars can be formed.Furthermore, it was shown that glucose itself can oligomerize during mechanocatalysis, which means that complete depolymerization to monomeric sugars is not possible (Figure 1). [18]gure 1: Mechanocatalytic partial depolymerization of cellulose to oligomeric glycans without formation of monomeric sugars.
It should be emphasized, that the energy required for a reaction can be reduced by scale up which makes this method even more attractive. [19]For successful depolymerization of the biomass a strong acid with a pKa below -1.8 such as hydrochloric or sulfuric acid is needed. [18]The approach we investigated is based on sulfuric acid as catalyst and a planetary ball mill.Hereby we aimed to obtain oligomeric glycans with a defined molecular weight of 1-10 kDa, which is precipitated in a second step followed by functionalization.These defined functionalized oligomeric glycans can be used in the packaging industry as binder in coating resins as well as in printing resins. [20]

Composition analysis of lignocellulosic feedstock
For composition analysis we refer to the regulations of the National Renewable Energy Laboratory.Within these regulations total solids [21] , extractives [22] , ash [23] , structural carbohydrates and lignin [24] were analyzed (see SI).

Biomass impregnation
Before impregnation with sulfuric acid the lignocellulosic feedstock was shred to a fine powder in a coffee grinder.For impregnation the required amount of 96 % sulfuric acid was weighed in a 500 mL round bottom flask and diluted with about 50 mL diethyl ether respectively ethanol, the latter for the impregnation at Fraunhofer IWKS, per 5 g biomass.Here, the acid content is given in % based on the total mass including acid load itself.After adding the biomass the resulting suspension was stirred for 1 h at room temperature.The solvent was then removed under reduced pressure and the product was dried on the Schlenk line.

Mechanocatalytic partial depolymerization
For the experiments a planetary ball mill Retsch PM 100 was used.5 g of the respective sample was put into the grindig vessel followed by 163.48 g zirconia balls with a diameter of 3 mm.At Fraunhofer IWKS, mechanocatalytic partial depolymerization was analogous carried out in a Fritsch Pulverisette 6 with 144.15 g zirconia balls with a diameter of 10 mm.Due to heat development during the grinding process a program with a milling duration of 5 min at 500 rpm followed by a cooling cycle of 2 min was chosen.For clarity, only the effective milling duration without the cooling cycles is given.To show that mechanocatalytic partial depolymerization is a reproducible method for polysaccharide digestion towards oligomeric glycans, three identical experiments with apple pomace pre-extracted with hot water at the Fraunhofer IWKS were carried out and compared with each other

Product characterization after partial depolymerization
Directly after the milling process the temperature inside the grinding vessel is measured to confirm that the temperature is below 90 °C and no Maillard reaction as well as no following reactions like dehydration of sugars to furfural and hydroxymethylfurfural occurs. [25]In a first experiment without acid impregnation the influence of ball milling towards the crystallinity of cellulose was determined by milling for 86 min followed by a XRD-measurement to verify the amorphization while milling.Therefor a Bruker D2 Phaser 2nd Generation was used with a wavelength of Cu Kα =1,5406  .The measurements were taken in a 2-theta range from 8 ° to 60 ° with a measurement speed of 0,239 °•(1.1 s) -1 using a background-free silicon sample holder.After mechanocatalytic partial depolymerization the amount of product equal to a mass of 400 mg without acid load was weighed in and stirred with 40 g demineralized water at room temperature respectively 100 °C (Fraunhofer IWKS).Afterwards the insoluble fraction was filtered off, dried under vacuum at 40 °C overnight and weighed in again to determine the mass of the insoluble fraction.The filtrate was analyzed by HPLC (Agilent 1260 Infinity equipped with a HPX-87H-column at a flow rate of 0.5 mL•min -1 at 60 °C).Additionally precipitation and purification was performed as reported by Wolf et al. for the experiments at Fraunhofer IWKS. [20]C-measurements of the samples after milling were taken to determine the molecular weight distribution of the soluble fraction.For samples with cellulose as substrate a PSS Suprema column at a flow rate of 1 mL•min -1 0.05 % NaN3 at 25 °C was used.The lignocellulosic samples were measured with a MCX 1000A column with a mixture of 0.1 mol•L -1 NaOH and 0.1 mol•L -1 NaCl at the same flow rate.At Fraunhofer IWKS SEC-MALS was performed using an Agilent 1260 Infinity II HPLC system combined with an Optilab and Dawn 8 detector (both Wyatt Technology Europe GmbH) A PSS Suprema column with a flow rate of 0.8 mL min -1 100 mM NaNO3/0.05% NaN3 at 30°C was used.

Influence towards crystallinity
Initially, an experiment with pure cellulose was carried out to verify the expected amorphization of cellulose after milling in a planetary ball mill.The resulting XRD-pattern (figure S1) was obtained where the observed reflexes of untreated cellulose are in agreement with the ones from literature. [26]Based on the XRD-pattern after milling we could show that the amorphization was successful in the absence of depolymerization.

Cellulose as substrate
The experiments carried out were run with milling cycles of five minutes milling followed by two minutes cooling so that temperatures of 90 °C were not exceeded to prevent subsequent reactions.
Possible subsequent products starting from monomeric sugars which were formed in small amounts could be furfural and hydroxymethylfurfural. [27,28] irst insights towards the effect of sulfuric acid as catalyst were acquired through experiments with effective milling durations of 86 min and different acid contents.The acid content was varied between 0 % and 9.48 % as shown in Figure 2. The figure shows that the acid content has a large impact on the oligomer fraction.Without catalyst the amount of desired oligomers is below 10 % which increases to above 90 % by adding 2.37 % sulfuric acid.What we also noticed is that a higher acid content reduces the oligomer fraction.This trend can be attributed to the higher catalyst content which brings about an even higher reaction rate and therefore a higher degree of depolymerization towards the sugar components.The desired 1-10 kDa glycans defined during this work can be used as starting material for coating resins. [20] verify the desired molecular weight range the GPC measurement of the sample partially depolymerized with 2.37 % sulfuric acid is shown (Figure 2b).
The GPC measurement exhibits a broad molecular weight distribution with a maximum in the range of the desired molecular weight of 1-10 kDa.Based on this result we wanted to transfer the reaction parameters to the lignocellulosic feedstock.

Composition of lignocellulosic feedstock
Since every lignocellulosic feedstock has its own composition regarding the contents of water, extractives, ash, carbohydrates and lignin it is necessary to determine the composition for all used feedstocks which are shown in Figure 3. are predominant in the lignin fraction of cocoa shells. [29]he composition of the lignocellulosic feedstock is completely heterogeneous.The largest amount of the natural raw material, with exception of wheat straw, consists of extractives such as nonstructural low molecular weight saccharides and carboxylic acids.The valuable glycans will be obtained from cellulose and hemicellulose.Hereby wheat straw has the highest content of cellulose and hemicellulose with 19.2 % and 18.9 % respectively which makes up nearly 40 % of the raw material.This amount is lower in the other natural raw materials with 31 % for beet pulp, 23 % for apple pomace, and 12 % for cocoa shells.On the other hand, the content of tannin-like substances with small amounts of lignin is highest in the cocoa shells with 32 % [29] , while it ranges between 12 % and 19 % in the other natural raw materials used.The structure of the biomass has a decisive influence on the mechanocatalytic partial depolymerization.Strong binding of the polysaccharides into the lignin matrix prevents depolymerization.Since cocoa shells have the highest content of rigid aromatic molecules in combination with a lower content of hemicellulose and cellulose we expect that the mechanocatalytic partial depolymerization under the same conditions will be more hindered than for wheat straw since it has the highest content of hemicellulose as well as cellulose.

Lignocellulosic feedstock as substrate
This section discusses the results of wheat straw, cocoa shells, apple pomace and beet pulp used as lignocellulosic feedstock.Since the composition is more complex, as shown in the previous section, the following results are shown as soluble products and solid residue without further separation of the soluble products.

Wheat straw
In an experiment with 3.6 % acid content and a similar milling duration as for cellulose a lower soluble fraction was achieved compared to impregnated cellulose (Figure 4a, second bar).This experiment showed that the milling duration has to be increased for the lignocellulosic feedstock, which is probably due to the lignin content that hardens the structure.The results after increasing the milling duration as well as varying the acid content are shown in Figure 4a.Without acid a soluble fraction of around 25 % is obtained.Hereby we have to keep in mind that wheat straw has 14.3 % extractives which are included in the soluble products.Substracting the fraction of extractives from the soluble product fraction leads to an uptake around 10 % which is in the same range as for cellulose alone.For higher milling durations of 255 min and acid contents of 3.6 and 6 %, a soluble fraction above 90 % is achieved.Hereby, sufficient energy was supplied to the system to break the rigid structure and then depolymerize it.

Cocoa shells
The reactions with cocoa shells were carried out in the same way as those of wheat straw.For cocoa shells a 7-8 % increase of the soluble fraction could be obtained (Figure 4b).Compared to wheat straw this increase seems quite small.This small increase can be attributed to the high phenolic content with 32.2 % as well as the low carbohydrate content with 11.4 % where the molecular structure is more rigid.This is consistent with the expectation of a more hindered reaction for this natural raw material.Furthermore, condensation reactions of phenolic molecules (tannins) and between phenolic molecules and other macromolecules (proteins, polysaccharides) may counteract depolymerization.In sum depolymerization of cocoa shells without pretreatment has a low efficiency.

Apple pomace
As the previous lignocellulosic feedstock, the reactions with apple pomace were carried out under the same reaction conditions (Figure 4c).It should be noted that the experiment series with apple pomace showed no trend of increase of the soluble fraction for the reaction conditions of the first four treatments in Figure 4c.A possible reason could be non-structural inhibitors in the substrate which are blocking the acid effect and partial encapsulation in rigid structures derived from kernel fragments where the energy input is not sufficient in the same way as for wheat straw.The hypothesis of non-structural substances which hinder the reaction is supported by comparison with wheat straw where the content of extractives is low and the percentage of the soluble fraction is high.The experiment which increased the soluble fraction of apple pomace by 5 % compared to the condition without catalyst was the one with the harshest reaction conditions of 255 min milling duration and 6 % acid content.

Beet pulp
The reactions with beet pulp were carried out under the same conditions as the previous substrates (Figure 4d).Beet pulp shows an increase of the soluble fraction from 45 % without catalyst up to around 65 % which neither can be increased with a higher milling duration nor a higher acid content.The relatively high soluble fraction without acid can be explaind by the high contents of extractives in water which are of course also soluble after the mechanical treatment.

Influence of pre-extraction
Cocoa shells and beet pulp also show an increase in soluble products while apple pomace seems to follow no clear trend (Figure 4).When comparing the composition of all lignocellulosic feedstocks we noticed that wheat straw has the lowest amount of extractives with 14.3 % while the other substrates contain 39-47 %.Since wheat straw showed the highest amount of soluble products after mechanocatalysis the influence of the extractives towards the reaction is investigated under assumption that extractives may hinder a successful reaction to soluble products due to blocking some acid amount and because soluble non-structural glycans are favored to depolymerize.Therefore, pre-extractions based on the NREL regulations were carried out with beet pulp and apple pomace.Afterwards the processes were continued with 170 min milling duration and an acid content of 6 %.The results are shown in Figure 5. Combining the extractives with the soluble products leads in total to a higher soluble fraction than for the non-extracted substrates.Considering that the extractives make up 39 % and 47 % for apple pomace and beet pulp respectively, a huge increase up to 40 % of the desired soluble glycancontaining product fraction can be seen.Also a lower content of solid residue for beet pulp compared to apple pomace was observed which probably can be explained by the lower lignin content.As mentioned previously, the obtained molecular weight distribution of the soluble product fraction is of interest and is therefore shown in Figure 6 for the soluble glycan-containing fraction of the non-extracted as well as the pre-extracted apple pomace.The non-extracted apple pomace shows a bimodal molecular weight distribution in two nonfavored regions below 1 kDa and above 10 kDa, that supports the hypothesis that soluble non-structural glycans are favored to depolymerize while the structural larger ones depolymerize afterwards.In contrast the pre-extracted apple pomace shows a broad molecular weight distribution in the favorable region.This circumstance can be attributed to the pre-extraction where low molecular weight compounds like sugars as well as already soluble glycans are extracted from the sample and therefore the partial depolymerization of the macromolecules is less hindered.
Since the mechanocatalytic partial depolymerization is statistical, a broad molecular weight distribution was expected.

Reproducibility
The reactions were carried out with a milling duration of 170 min and an acid content of 6 %.In the following Figure 7 three overlayed HPLC-chromatograms of the soluble products that show identical profiles are shown as well as the amount of soluble products.The results are in good agreement to each other and lead to a soluble product fraction of (26.5±1.9)%.These reproducibility experiments show that the mechanocatalytic partial depolymerization is a reliable and robust method.

Process application with extracted apple pomace
The extraction process of apple pomace in pilot plant scale to yield about 5 % glycans (purity ≥ 90 %) has been established at Fraunhofer IWKS using subcrictical water. [20]After extraction approx.70 % of unsoluble apple pomace residue is recovered after removing the excess water by oven drying (80 °C).Apple pomace residue has been prepared as described and processed in a planetary ball mill.The objective of adding mechanocatalytic partial depolymerization to the process chain was to increase the yield of glycans.Therefore, the milled apple pomace was extracted under atmospheric pressure in either hot or cold water.

Conclusion
Mechanocatalytic partial depolymerization is proven to be a robust and reliable method for the depolymerization of cellulose and different lignocellulosic feedstocks towards glycans.Testing different acid contents of sulfuric acid as well as different milling durations lead to a feasible partial depolymerization with an acid content of 6 %.Milling durations were best at 85 min and 170 min depending on the composition of the feedstock.It could be shown that pretreatment by extraction increased the yield of the total soluble fraction up to 40 % and additionally lead to a molecular weight distribution of the desired range 1-10 kDa. [35]In comparison to glycans from hot water extraction process (pilot plant) the molecular weight of precipitated glycans was reduced to 1/3 and the polydispersity narrowed down from 2.9 to 1.4.With this, a molecular weight of 13.0 kDa respectively 12.1 kDa was achieved increasing the yield from 5 % to nearly 10 %.This makes the approach by mechnocatalytic partial depolymerization very promising for increasing the yield and reaching the molecular weight range in one step.To further assess and improve the energy demand/product-yield-ratio a scale-up of the process is planned.It could be shown that mechanocatalytic partial depolymerization is a robust and reproducible method to obtain lignocellulose-based glycans as sustainable biobased chemicals.

TOC Graphic
Mechanocatalytic partial depolymerization is used to obtain glycans from lignocellulosic feedstock such as wheat straw, cocoa shells, apple pomace and beet pulp.It can be shown that the yield of the glycans is increased by mechanocatalytic partial depolymerization.In addition, with a suitable pre-treatment the yield can be increased further.

Figure 2 :
Figure 2: a) Composition of the product mixture after mechanocatalytic partial depolymerization of cellulose substrate with different acid contents; Reaction conditions: 500 rpm, 86 min effective milling duration, 5 g substrate and 163.48 g zirconia balls.b) Molecular weight distribution from GPC-measurement of the experiment with 2.37 % sulfuric acid.

Figure 3 :
Figure 3: Composition of used wheat straw, cocoa shells, beet pulp and apple pomace analyzed according to the NREL regulations.The lignin fraction includes other phenolic substances which

Figure 4 :
Figure 4: Amount of soluble products for a) wheat straw, b) cocoa shells, c) apple pomace and d) beet pulp dependent on the acid content and the milling duration.Reaction conditions: 500 rpm, 5 g substrate and 166.163.48 g zirconia balls.

Figure 5 :
Figure 5: Amount of solid residue with and without pre-extraction of beet pulp and apple pomace

Figure 6 :
Figure 6: Molecular weight distribution from GPC-measurement of the experiment with a) nonextracted apple pomace and b) pre-extracted apple pomace.Reaction conditions: 6 % acid content, 170 min milling duration, 500 rpm, 5 g substrate and 166.48 g zirconia balls.

Figure 7 :
Figure 7: a) Amount of soluble products and b) HPLC-chromatograms for three experiments of pre-extracted apple pomace.Reaction conditions: 6 % acid content, 170 min milling duration, 500 rpm, 5 g substrate and 166.48 g zirconia balls.

Figure 8 :
Figure 8: Composition of different fractions obtained by cold and hot water extraction (CWE/HWE) of milled apple pomace.Reaction conditions:500 rpm, 170 min milling duration, 5 g substrate, 6 % acid content and 144.15 g zirconia balls.