Design of a Knife Mill with a Drying Adaptation for Lignocellulose Biomass Milling: Peapods and Coffee Cherry

Abstract

Effective grinding of residual agricultural materials helps to improve yield in the production of chemical compounds through hydrothermal technology. Milling pretreatment has different types of pre-treatment where ball mills, roller mills, and finally, the knife mill stand out. The knife mill being a mill with continuous processing, its multiple benefits and contributions highlight the knife milling process; however, it is a process that is generally carried out with dry biomass that generates extra processing of the biomass before grinding, implying longer times and wear than other equipment. This work presents the design of a knife mill with an adaptation of free convection drying as a joint process of knife milling and drying. The design is based on lignocellulosic biomass, and the knife milling results are presented for two biomasses: peapods and coffee cherries. The knife mill is designed with a motor, a housing with an integrated drive system, followed by a knife system and a feeding system with a housing and finally the free convection drying system achieving particle sizes in these biomasses smaller than 30 mm, depending on the time processed. The data demonstrate the significant impact of particle size on the yields of various platform chemicals obtained from coffee cherry and peapod waste biomass. For coffee cherry biomass, smaller particle sizes, especially 0.5 mm, result in higher total yields compared to larger sizes while for peapod biomass at the smallest particle size of 0.5 mm, the total yield is the highest, at 45.13%, with notable contributions from sugar (15.63%) and formic acid (19.14%).

1. Introduction

The construction of mills for lignocellulose milling is important. Grinding is a key process for the production of biofuels and other products derived from biomass. Grinding results in particle size reduction and thus improves the valorization potential of the process. Pretreatment is used in lignocellulosic biomass because it reduces particle size and crystallinity, benefiting downstream processes such as enzymatic hydrolysis and anaerobic digestion, among others [1,2]. In addition, it improves process efficiency by facilitating structural complexity and increasing surface area. By decreasing the crystallinity of the cellulose, it makes the biomass more suitable for hydrolysis. It is important to note that the milling process can be implemented when the biomass is wet or dry; however, this depends on the type of mill and the type of biomass. Several studies have shown that wet milling is more efficient than dry milling when the objective is to increase the pore volume for enzymatic hydrolysis [3].

Milling is an energy-consuming process; however, by selecting the milling pattern, a balance can be achieved between the energy required and the desired particle size. It should be noted that grinding efficiency and energy consumption are attributed to the type of mill, with hammer and knife mills being the most energy efficient for fibrous and hard lignocellulosic biomasses [2,3]. In addition, the cost is high, but process optimization can be achieved to obtain a higher process yield [4]. The milling process can be combined with other chemical or physicochemical processes to improve efficiency. For example, hydrothermal milling has been shown to reduce energy consumption and sugar yield. In biorefinery processes, which are of interest to the research group, milling increases surface area, allowing enzymes and other reagents greater accessibility. As mentioned, milling reduces the crystallinity of the cellulose and the degree of polymerization [1], which makes the biomass more suitable for downstream chemical processes. It also improves the reaction kinetics of downstream processes and leads to higher yields of sugars and biofuels [5]. Furthermore, milling facilitates the handling of biomass and increases its bulk density. Its process has no chemical residues, which means that no toxic materials are produced.

A fundamental parameter for the grinding process is the grinding time since it significantly influences the efficiency of particle size reduction, where the longer the time, the finer the particles, although the longer the time, the greater the energy expenditure, which is why we seek to optimize the process by providing a balance between energy consumption and particle size. It is necessary to optimize the grinding time and not to enter a point of decreasing performance since additional time increments do not reduce the particle size due to agglomeration but rather reduce the efficiency of the process [5].

The milling process is the most widely used physical pretreatment, with 68%, followed by grinding, 16%, refining 9%, ultrasound, 4%, and extrusion, 3%. There are different types of mills and each has a reduction mechanism such as shearing, compression, tearing or breaking. Figure 1 shows the types of mills and their classification according to the operating regime (batch and continuous). The most used types of mills are the ball mill with 65%, the knife mill with 13% and finally the other mills with 22% [1].

Mill selection depends largely on the type of biomass to be processed based on biomass properties such as ductility and biomass content, among others. For example, biomasses with more ductile properties, such as agricultural residues, are more suitable for shearing processes. On the other hand, seeds have anisotropic growth properties, which is why they are suitable for milling processes under compression and impact. It should be noted that depending on the type of biomass and the desired particle size, the amount of energy and the balance on them will be defined [2,6].

Taking into account that ball mills and knife mills are the most commonly used, the differences between them and the selection of the most suitable mill for the biomasses are presented below. The differences between ball mills and knife mills are significant in the design. Ball mills have a horizontal cylindrical structure that uses steel balls to grind the materials. It works by means of gears and transmission devices that make the cylinder rotate and the balls go up and down, producing impact and grinding the material by this mechanism. On the other hand, the knife mill has a rotor-driven cutting blade design that, when rotating, cuts and chops the material. Likewise, the mechanisms they use are different; the ball mill applies the compression and attrition process to reduce the particle size, while the knife mill applies the cutting and shearing process. The ball mill is used to grind chemicals, ceramics, ores, and minerals into fine powders both dry and wet. On the other hand, the blade mill is used for white materials, such as fresh fruits, vegetables, or fibrous or brittle materials, and its grinding is generally dry. Similarly, ball mills achieve ultrafine particle sizes down to 0.01 mm, while knife mills produce coarser particle sizes, generally used to grind materials before passing them through a mill where ultrafine powders are sought. Finally, ball mills are more commonly used in industry and knife mills in laboratories [7,8].

These blade mills have flowability problems due to moisture which causes poor feeding. Also, wet biomass generates excessive heat generation, which can damage the equipment and even the material properties when they are changed by the heat. Moisture content goes hand in hand with blade speed and size [9]; however, biomass with a very high moisture content tends to create abrasive settings on the cutting edges of the blades and makes maintenance more frequent. In addition, high moisture content increases energy, and this represents a cost increase [10,11]. In order to use the knife mill, it is necessary that the biomass has a low moisture content, and for this reason, a knife mill design with a drying adaptation is proposed in response to the problem posed.

2. Materials and Methods

2.1. Biomass: Peapods and Coffee Cherry

Table 1. Characterization of biomass used for testing the designed mill.

Assay	Coffee Cherry (%)	Peapods (%)
Moisture (initial biomass)	80.79	80.77
Moisture (BHP)	10.94	7.77
Ashes	7.79	4.22
Volatile matter	79.91	74.18
Fixed carbon	1.36	13.0
Cellulose	27.6	20.2
Hemicellulose	12.5	17.4
Lignin	13.7	5.0

shows the characteristics of the biomass used in order to test the efficiency of the milling process in the hydrothermal valorization of biomass. All the biomass data and meshing influence results were found by the researchers in order to test the efficiency of the knife mill. The moisture content of the pretreated biomass for hydrothermal treatment is defined as (BHP).

The results (carried out via acid detergent fiber and neutral detergent fiber) indicate the compositional differences between coffee cherry and peapods after biomass processing. Both initial biomasses had high moisture content, approximately 80.79% for coffee cherry and 80.77% for peapods, which significantly decreased after biomass hydrothermal pretreatment (BHP) to 10.94% and 7.77%, respectively. The ash content was higher in coffee cherry (7.79%) compared to peapods (4.22%). In terms of volatile matter, coffee cherry exhibited a higher percentage (79.91%) than peapods (74.18%), while fixed carbon was notably low in coffee cherry (1.36%) but higher in peapods (13.0%). The cellulose content was greater in coffee cherry (27.6%) compared to peapods (20.2%), whereas peapods had a higher hemicellulose content (17.4%) than coffee cherry (12.5%). Finally, lignin content was significantly higher in coffee cherry (13.7%) compared to peapods (5.0%), indicating distinct structural characteristics between the two biomass types [12,13].

2.2. Miling Pretreatment and Mathematic Modeling for Design

Based on the previous research and considering the properties of the biomasses to be processed (peapods and coffee cherry), it was decided to work with a knife mill with a drying adaptation to satisfy this need in the process without the need for sub-processes before milling. For the design process of the blade mill, the main component, the blades, must be considered. Several blades are used that can be changed and adjusted according to the material to be processed. They can have between three and eleven teeth and the degree of milling will depend on this. Blades are usually designed to achieve high efficiencies of large cuts per small units of time, e.g., curved blades. Its rotation speed can vary from 2000 to 25,000 rpm or more, depending on the model because some knife mills can grind hard materials and other knife mills can grind slightly softer materials. The mill feed is usually constant, between the rotor and the blades where the material is ground, there is a casing that is responsible for keeping the material inside the mill until the desired grain size is met. They are generally designed in steel and seek to relieve the rotor stresses in order to achieve good performance. In addition, a cooling system is installed to prevent the mill from overheating when it rotates at high revolutions. It can also have mixing adaptations for homogenization [14,15].

Also, the parts that a blade mill must have are the rotor blades, fixed knives, grinding chamber, screens, motor and power transmission, control systems, cooling system, feeding and discharging mechanisms, and finally, grinding vessels and accessories [8,16].

Theoretically, the energy required to reduce the particle size of solid foods can be calculated as shown in Equation (1). Where K and n are constants based on the milling technique and ground material, and dE is the energy needed to break a unit mass of diameter x to approximately size dx [17].

$$dE=-K{\displaystyle \frac{dx}{X^n}}$$

(1)

For milling processes, there are different theories which seek to explain energy consumption and grain size. Rittinger’s theory (1867) states that the energy required to reduce grain size is directly proportional to the new surface area created and is generally applied for particles less than 1 micrometer. Kick theory (1885) states that the energy required to reduce the particle size is directly proportional to the particle diameter and is used in the crushing of large particles. Also, Bond’s theory (1952) states that the energy used for crack propagation is proportional to the length of the new crack produced and is useful for the sizing of grinding mills. Finally, Walker’s theory, which proposed a model combining the three previous theories where he generalizes the relationship between energy and size in a differential way [18]. Based on the theories presented, the mill is sought to obtain fine particle sizes; for this reason, the Rittinger theory will be used. This theory has been adapted for different materials such as pellets and wood where energy consumption has been determined based on moisture content and particle size distribution [19,20]. Equation (2) shows mathematically the approach of the theory.

$$\gamma ={\displaystyle \frac{1}{particle diameter}}$$

(2)

Equation (3) shows the energy required for size reduction where $C_{VR}$ is a constant characteristic of the material and the x relates to the feed and the milled product. Index 1 being the initial particle diameter and index 2 the final particle diameter after milling.

$$E_{1-2}=C_{VR} \left({\displaystyle \frac{1}{X_2}}-{\displaystyle \frac{1}{X_1}}\right)$$

(3)

In addition, the diameters can be calculated with Equation (4), where G₂ nd g₁ represent the mass fractions of particles X₂ and X₁ [17].

$$X_2={\displaystyle \frac{1}{\sum _2\frac{G_2}{X_2}}}, X_1={\displaystyle \frac{1}{\sum _1\frac{g_1}{X_1}}}$$

(4)

To calculate the energy consumed in a size reduction test, we have Equation (5).

$$P={\displaystyle \frac{\sqrt{3}\ast I_{AVE}\ast V\ast t}{3600}}$$

(5)

where I_AVE corresponds to the average current, V to the voltage and t to the time. Finally, when this power is divided by the weight of the biomass particles, the energy consumed per unit mass is obtained [21].

2.3. Design of Knife Mill

The mathematical model described above has been followed in the design process, starting from the type of biomass and its properties. The breaking strength of lignocellulosic biomass was established as close to 95 MPa, with a transmission system from the motor to the blade to grind it and obtain particle sizes of less than 30 mm, where unification is achieved thanks to the meshes between the casing and the blades. To achieve the design of the blade mill, the breaking strength of the biomass was established as mentioned above, taking into account the length of the biomass. The thickness of the blade was set at 18 mm, as well as the thickness that is related to the particle size to be achieved, 30 mm. From the established data, the cutting force is obtained, and the cutting radius of 0.1 is established to obtain the required torque. Additionally, the motor torque is obtained from the torque required by the number of blades set. With this value of 51.3 Nm, the motor has been defined in the catalogs.

The milling design is supported by MOTOX Geared Motors’ D87.1.2011 catalog for motor selection based on a torque of 59.4 Nm. Likewise, a four-point blade was selected which is characteristic for cutting dense materials and irregular shapes. The design has a manual and vertical feeding system where the biomass enters directly in contact with the blades that are rotating at 572 revolutions per minute, in turn making a heat transfer by free conduction of ambient air, and will depend on the time left for the particle size to be obtained as well as the selection of the mesh that serves to separate the particle sizes and finally, remove the biomass manually (Figure 2).

Table 2. Characteristics of the knife mill.

Characteristics	Value	Unit
Knife thickness	18	mm
Shear force	51.3	N
Number of knives	10	-
Motor torque	116	Nm
Engine type	Worm-geared motor s	-
Power rating	1.1	kW
Transmission ratio	20	-

shows the characteristics of the design that was carried out. Figure 3 shows the parts of the design.

3. Results and Discussion

3.1. Simulation of the Main Part of the Mill

Finally, for the simulation of the blade, a fixed support was defined in the center corresponding to the axis on which it is anchored. We also defined the shear force on each side of the blade with the value of 51.3 N. We also defined a mesh predetermined by the program. Figure 4 shows a finite element analysis of the most important part of the design, which corresponds to the blade. For the design of the blades, it is necessary to have a material with good mechanical strength and good machinability for this cutting tool. That is why SAE 1045 steel was selected for the design of the blades because it is a carbon steel that has excellent mechanical properties and versatility; it also has a high mechanical strength that varies and has a high impact resistance [22], not to mention that it is a material that can be easily machined [23,24]. To select the steel, the properties mentioned above are taken into account, and the values based on this company are taken because they provide us with a commercial value and also general values because this steel can improve its properties with hot or cold thermal treatments. An SAE 1045 steel with a density of 10.85 g/cm³ and Young’s modulus of 206 GPa [25,26] was selected for the design. In the analysis of each tip or knife, the shear force mentioned above is possible, and thus it was obtained that the maximum effort is presented in the four tips and is 3.3 × 10⁵ Pa and a minimum effort of 1747.7 Pa, which is a minimum percentage. For this reason, it can be concluded that the material selected for the knife is adequate because it withstands the stress to which it is subjected. It can also be observed that the maximum deformation that is presented is 1.25 × 10⁻⁷ m in the corners of the knife and the minimum deformation is zero in the center that is supported on the shaft.

The prospects of this work would be to reach a detailed design for the model to achieve the construction of a prototype and to find an experimental efficiency by checking if the processing time is reduced. In addition, a grinding comparison between a conventional knife mill with a drying pre-processing and the knife mill with free convection drying.

The type of nucleated material can affect the grinding process and the material of the blades, due to the humidity that is trapped in the material. For this reason, it is convenient to establish a maximum of 10% moisture as a control parameter to enter the grinding process and avoid further wear of the material. Another factor to take into account is the rotation speed. Its rotation speed can vary from 2000 to 25,000 rpm or more, depending on the model because some knife mills can grind hard materials and other knife mills can grind slightly softer materials.

3.2. Products Obtained from Milling of Biomass

In the quest to optimize yields from the milling process, a series of 24 meticulously designed experiments were conducted, varying both temperature, biomass and particle size. The temperatures selected for this study are 180 °C, 220 °C, and 260 °C, while the particle sizes tested include 0.5 mm, 1 mm, 2 mm, and 5 mm. By systematically altering these parameters, we aim to uncover the intricate relationships between temperature, particle size, and milling efficiency. This exploration not only seeks to enhance yield outcomes but also to contribute valuable insights into the milling process, ultimately paving the way for improved industrial applications and resource management. The platform chemicals were obtained experimentally through hydrothermal reactions in batch reactors equipped with electronic temperature control systems. Furthermore, the measurements of sugars and organic platform molecules were obtained through instrumental chemical analysis with high-resolution liquid chromatography (HPLC) with samples weighed on four-digit analytical balances, generating absolute errors of 0.0001 ten-thousandths of a gram.

The temperature values studied were selected due to the operating conditions in hydrothermal processes that require subcritical or liquid water to attack the complex chemical structures in cellulose and hemicellulose polymers present in biomass, which from 180 °C begin to hydrolyze to obtain simpler chemical molecules such as sugars in the first stage and subsequently, with the increase in temperature, the structural changes begin towards becoming platform molecules such as levulinic acid, furfural and hydroxymethylfurfural with very defined reaction mechanisms that have been presented in our scientific articles related to this work.

3.2.1. Coffee Cherry Waste

Table 3. Results of changing the particle size at different temperatures in order to obtain platform chemicals from coffee cherry waste.

T (°C)	Particle Size (mm)	Y Sugar (%)	Y Formic Acid (%)	Y Levulinic Acid (%)	Y HMF (a) (%)	Y Furfural (%)	Y Total (%)
180	0.5	25.58	6.61	1.64	1.3	1.15	36.28
180	1	22.13	6.61	1.42	1.15	1.3	32.60
180	2	21.07	7.07	1.74	-	-	29.88
180	5	20.98	6.61	1.4	1.1	0.49	30.58
220	0.5	10.16	5.68	1.33	-	-	17.16
220	1	13.99	7.34	1.86	0.06	-	23.25
220	2	4.55	9.34	3.2	-	-	17.09
220	5	12.86	7.23	1.66	0.13	-	21.87
260	0.5	10.19	9.32	1.43	0.05	0.03	21.02
260	1	9.5	9.46	1.38	-	-	20.35
260	2	0.53	9.95	2.28	-	-	12.76
260	5	9.65	8.09	1.23	0.07	-	19.04

(a) Hydroxymethylfurfural.

shows the optimization results for the different particle sizes obtained from the designed mill. Particle sizes were established using sieve systems for granulometry.

The data show that particle size has a significant effect on the yields of various platform chemicals obtained from coffee cherry waste at 180 °C. Smaller particle sizes, especially 0.5 mm, result in higher total yields compared to larger sizes. This is primarily driven by the trend in sugar yield, which decreases from 25.58% at 0.5 mm down to 20.98% at 5 mm particle size. The yield of formic acid remains relatively constant around 6.6–7.0% regardless of particle size. Levulinic acid yield peaks at 1.74% for 2 mm particles, suggesting an optimal size in that range, but more data are needed to confirm this. The data for HMF and furfural yields are incomplete, with values missing for the 2 mm particle size. This makes it difficult to draw conclusions about the effect of particle size on these compounds. Overall, the total yields of all platform chemicals follow the same trend as sugar, with the highest total at 36.28% for 0.5 mm particles and decreasing to 29.88% and 30.58% for 2 mm and 5 mm, respectively. In summary, these data indicate that optimizing the particle size, especially reducing it below 1 mm, can significantly improve the overall yield of platform chemicals from coffee cherry waste at 180 °C, with formic acid and levulinic acid yields also impacted to some degree.

The data collected at 220 °C reveal interesting trends in the yields of platform chemicals derived from coffee cherry waste, particularly in relation to particle size, even though the yields decrease compared to those obtained at 180 °C, due to the beginning of the carbonization processes and decrease in the extraction of the lignocellulosic material. At the smallest particle size of 0.5 mm, the total yield is the lowest at 17.16%, with sugar yield at 10.16%, formic acid at 5.68%, and levulinic acid at 1.33%. As the particle size increases to 1 mm, the total yield rises to 23.25%, driven by a notable increase in sugar yield to 13.99% and formic acid yield to 7.34%. However, the total yield drops significantly at the 2 mm size to 17.09%, despite a higher levulinic acid yield of 3.20%, indicating that while larger particle sizes may enhance certain yields, they can detrimentally affect the overall yields of other compounds. At the larger particle sizes of 5 mm, the total yield is 21.87%, which is slightly lower than that of the 1 mm size but still higher than the 0.5 mm and 2 mm sizes. The formic acid yield remains relatively stable across the 1 mm and 5 mm sizes, while the levulinic acid yield is highest at 2 mm. Notably, HMF and furfural yields are absent for the smaller particle sizes, suggesting that these compounds may not be produced effectively under the conditions tested. Overall, the data indicate that a particle size of 1 mm appears to optimize the yield of platform chemicals at 220 °C, balancing the production of sugar and formic acid while maintaining reasonable levels of levulinic acid. Further investigation into the missing data for HMF and furfural would provide a more comprehensive understanding of the conversion process.

The data collected at 260 °C reveal a distinct shift in the yields of platform chemicals derived from coffee cherry waste compared to those at lower temperatures. At the smallest particle size of 0.5 mm, the total yield is 21.02%, with sugar yield at 10.19%, formic acid at 9.32%, levulinic acid at 1.43%, HMF at 0.05%, and furfural at 0.03%. As the particle size increases to 1 mm, the total yield decreases slightly to 20.35%, driven by a small decrease in sugar yield to 9.5% and levulinic acid to 1.38%. However, the formic acid yield increases to 9.46%, indicating that larger particle sizes may enhance the production of certain compounds at higher temperatures. At the 2 mm particle size, the total yield drops significantly to 12.76%, despite a higher levulinic acid yield of 2.28% and formic acid yield of 9.95%. This suggests that while larger particle sizes may enhance the yields of certain compounds, they can also detrimentally affect the overall yields of other compounds. At the larger particle size of 5 mm, the total yield is 19.04%, which is higher than the 2 mm size but lower than the 0.5 mm and 1 mm sizes. The formic acid yield remains relatively stable across the 0.5 mm, 1 mm, and 2 mm sizes, while the levulinic acid yield is highest at 2 mm. HMF yield is present for the 0.5 mm and 5 mm sizes, while furfural yield is only present for the 0.5 mm size. Overall, the data indicate that a particle size of 0.5 mm appears to optimize the yield of platform chemicals at 260 °C, balancing the production of sugar, formic acid, and levulinic acid while maintaining reasonable levels of HMF and furfural.

In conclusion, the data presented demonstrate the significant impact of particle size on the yields of various platform chemicals obtained from coffee cherry waste at different temperatures. At 180 °C, smaller particle sizes, particularly 0.5 mm, result in the highest total yields, driven primarily by increased sugar production. While formic acid yields remain relatively constant, levulinic acid peaks at 2 mm, suggesting an optimal size in that range. At 220 °C, a particle size of 1 mm appears to strike a balance between maximizing sugar and formic acid yields while maintaining reasonable levels of levulinic acid. However, the missing data for HMF and furfural at this temperature limit the conclusions that can be drawn. At the highest temperature of 260 °C, the trends shift, with 0.5 mm particles yielding the best overall results. While larger sizes enhance certain compound yields, such as formic acid, they can also detrimentally impact the total yield. The presence of HMF and furfural at 260 °C, albeit in small amounts, further highlights the differences in the conversion process at higher temperatures. Overall, these data emphasize the importance of optimizing particle size to maximize the production of platform chemicals from coffee cherry waste, with the optimal size varying depending on the target compounds and the operating temperature.

3.2.2. Peapod Waste

Table 4. Results of changing the particle size at different temperatures in order to obtain platform chemicals from peapod waste.

T (°C)	Particle Size (mm)	Y Sugar (%)	Y Formic Acid (%)	Y Levulinic Acid (%)	Y HMF (%)	Y Furfural (%)	Y Total (%)
180	0.5	15.63	19.14	3.99	3.69	2.67	45.12
180	1	14.46	16.70	3.28	4.57	6.50	45.51
180	2	11.24	14.53	3.69	3.04	4.06	36.56
180	5	11.27	15.54	3.67	2.73	0.85	34.06
220	0.5	1.51	19.09	3.72	1.00	0.09	25.41
220	1	1.53	21.63	3.41	0.87	1.01	28.45
220	2	1.19	17.83	3.1	-	-	22.12
220	5	1.41	19.14	3.77	1.07	0.51	25.90
260	0.5	0.54	20.89	1.65	-	-	23.08
260	1	0.54	28.78	2.17	-	-	31.49
260	2	0.78	19.20	1.72	-	-	21.70
260	5	0.85	19.86	1.77	-	-	22.48

shows the optimization results for the different particle sizes obtained from the designed mill with peapod waste. The data collected at 180 °C indicate that particle size significantly influences the yields of various platform chemicals derived from peapod waste. At the smallest particle size of 0.5 mm, the total yield is highest at 45.13%, with notable contributions from sugar (15.63%) and formic acid (19.14%). The yields of levulinic acid (3.99%), HMF (3.69%), and furfural (2.67%) also contribute to this total, suggesting that smaller particle sizes enhance the overall extraction of valuable compounds. As the particle size increases to 1 mm, the total yield slightly increases to 45.52%, driven by a higher yield of HMF (4.57%) and furfural (6.50%), indicating that while sugar yield decreases to 14.46%, the increase in these specific compounds compensates for the loss, maintaining a high overall yield. However, as the particle size continues to increase to 2 mm and 5 mm, the total yield declines significantly, dropping to 36.55% and 34.05%, respectively. The decrease in sugar yield to 11.24% at 2 mm and 11.27% at 5 mm, coupled with a reduction in formic acid yield, indicates that larger particle sizes may hinder the extraction efficiency of these compounds. While levulinic acid yields remain relatively stable across the different sizes, the notable drop in HMF and furfural yields at larger particle sizes suggests that the conversion of peapod waste to platform chemicals is less effective at these sizes. Overall, the data indicate that optimizing particle size, particularly at 0.5 mm or 1 mm, is crucial for maximizing the yield of platform chemicals from peapod waste at 180 °C, highlighting the importance of size reduction in enhancing extraction and conversion efficiency.

The data collected at 220 °C reveal notable trends in the yields of platform chemicals derived from peapod waste, particularly as particle size varies. At the smallest particle size of 0.5 mm, the total yield is 25.41%, with the formic acid yield being the most significant at 19.09%. Sugar yield is relatively low at 1.51%, while levulinic acid contributes 3.72%. As the particle size increases to 1 mm, the total yield rises to 28.44%, driven primarily by an increase in formic acid yield to 21.63% and a slight increase in furfural yield to 1.01%. This suggests that smaller particle sizes may enhance the extraction of certain compounds, particularly formic acid, while maintaining reasonable levels of levulinic acid. However, as the particle size continues to increase to 2 mm and 5 mm, the total yields decline to 22.12% and 25.90%, respectively. The yield of sugar remains low across all sizes, with the highest being 1.53% at 1 mm, while formic acid yields decrease at 2 mm to 17.83% and then slightly recover at 5 mm to 19.14%. Levulinic acid yields remain relatively stable, around 3.1 to 3.77%. The absence of HMF and furfural yields at 2 mm indicates a potential loss of these compounds with larger particle sizes. Overall, the data suggest that a particle size of 1 mm optimizes the yield of platform chemicals at 220 °C, particularly enhancing the production of formic acid, while larger sizes may hinder the extraction efficiency of other valuable compounds. This highlights the importance of optimizing particle size to maximize the overall yield from peapod waste.

The data collected at 260 °C reveal a distinct shift in the yields of platform chemicals derived from peapod waste compared to lower temperatures. At the smallest particle size of 0.5 mm, the total yield is 20.89%, with formic acid being the most significant contributor at 20.89%. Sugar yield is relatively low at 0.54%, while levulinic acid contributes 1.65%. As the particle size increases to 1 mm, the total yield rises to 29.28%, driven primarily by a notable increase in formic acid yield to 28.78%. This suggests that larger particle sizes may enhance the production of certain compounds, particularly formic acid, at higher temperatures. However, as the particle size continues to increase to 2 mm and 5 mm, the total yields decline to 19.65% and 20.32%, respectively. The sugar yield remains low across all sizes, with the highest being 0.85% at 5 mm. Formic acid yields decrease at 2 mm to 19.20% and then increase slightly at 5 mm to 19.86%. Levulinic acid yields remain relatively stable, around 1.65 to 1.77%. The absence of HMF and furfural yields across all particle sizes at 260 °C indicates a potential loss or complete conversion of these compounds at higher temperatures. Overall, the data suggest that a particle size of 1 mm optimizes the yield of platform chemicals at 260 °C, particularly enhancing the production of formic acid, while larger sizes may hinder the extraction efficiency of other valuable compounds. This highlights the importance of optimizing particle size to maximize the overall yield from peapod waste at higher temperatures.

In conclusion, the analysis of the data collected at different temperatures highlights the critical role of particle size optimization in maximizing the yield of platform chemicals from peapod waste. At 180 °C, smaller particle sizes, particularly 0.5 mm and 1 mm, result in the highest total yields, with significant contributions from sugar, formic acid, levulinic acid, HMF, and furfural. As particle size increases, the yields of these compounds decrease, indicating that size reduction is essential for enhancing extraction and conversion efficiency at this temperature. At 220 °C, a particle size of 1 mm appears to be the most effective, optimizing the production of formic acid while maintaining reasonable levels of levulinic acid. Larger particle sizes lead to a decline in total yields, likely due to the loss of HMF and furfural. At the highest temperature of 260 °C, the trends shift, with 1 mm particles yielding the best results, primarily driven by an increase in formic acid production. However, the absence of HMF and furfural at this temperature suggests that these compounds may have been completely converted or lost under the high-temperature conditions. Overall, the data emphasize the importance of carefully selecting the appropriate particle size and temperature to maximize the yield of specific platform chemicals from peapod waste. Continued research and optimization of these parameters will be crucial for developing efficient and sustainable conversion processes for this agricultural waste stream.

4. Conclusions

Milling is an effective pretreatment for lignocellulosic biomass, with knife milling being particularly popular for its results. This paper presents a model of a knife mill designed to process lignocellulosic biomasses, specifically peas and coffee cherry, achieving particle sizes below 30 mm with an output power of 1.1 kW. This is an innovative knife mill with an adaptation of free convection drying to reduce adverse processes before grinding. The biomasses were ground in a knife mill, and a valuation of the coffee cherry waste obtained using this mill shows that particle size reduction significantly affects product yield. At 180 °C, decreasing the size from 5 mm to 0.5 mm increases the total yield from 30.58% to 36.28%, a 5.70% improvement driven by a 4.60% increase in sugar yield. At higher temperatures, the effect is even more pronounced; for example, at 260 °C, the same reduction increases the total yield from 12.76% to 21.02%. Similarly, the valorization of peach waste highlights the importance of particle size reduction. At 180 °C, size reduction from 5 mm to 0.5 mm increases the total yield from 34.05% to 45.13%, an improvement of 11.08%, mainly due to a 4.36% increase in sugar yield. At 260 °C, size reduction from 5 mm to 1 mm results in an 8.96% increase in total yield, underscoring the need for thorough milling prior to hydrothermal treatment for optimal yield and process efficiency for both types of biomass.

The use of agricultural waste to obtain platform chemical molecules is a project of great economic interest due to the possibilities of finding alternatives for commercial products with added value. Biorefineries become productive alternatives on an industrial scale to valorize agricultural waste.