Valorization of White Lupin Straw Through Mild Dilute Acid Hydrothermal Treatment: A Sustainable Route for Monosaccharide and 5-Hydroxymethylfurfural Production

Abstract

This study investigates the potential use of white lupine straw (WLS), an underutilized agricultural by-product, as a raw material to produce valuable biochemicals such as monosaccharides and 5-hydroxymethylfurfural (5-HMF) through hydrothermal pretreatment. The aim was to optimize mild reaction conditions to maximize the recovery of these products while minimizing degradation. The hydrothermal treatment of WLS in subcritical water with trace amounts of sulfuric acid was performed, followed by a two-step approach to evaluate the yields of hemicellulose and 5-HMF. The highest monosaccharide yield (163 g/kg) was achieved at temperatures between 174 and 181 °C and a holding time of 7–14 min, while the 5-HMF production was 139.9 g/kg at 199–203 °C and after 0.5–4.5 min. These results suggest that optimal 5-HMF production also increases the remaining solid residue. This study highlights the feasibility of WLS as a sustainable, low-cost biomass resource. It highlights the balance between temperature and time to maximize valuable product yields. The results contribute to advancing biorefinery processes by demonstrating that WLS can be effectively converted into bioethanol precursors and industrial chemicals, supporting circular bioeconomy principles and providing an environmentally friendly alternative to burning crop residues.

1. Introduction

Lupin (Lupinus L., family Fabacae) has low soil and climatic requirements [1]. Still, grain legumes are not very common in the world. However, lupine seeds are rich in components, such as protein (33–55% expressed on a dry matter basis) and fat (5.5–12.9% expressed on a dry matter basis) [2,3], which may constitute the desired ingredients of food (also functional) and animal feed products [3,4].

Global lupin production is low [3,5,6], but over the last 10 years, an increase in lupine production has been observed in European Union (EU) countries, not without periodic declines [5], with Poland being one of the major producers of lupin in the EU (77.5% of supply in 2022) [5].

Among the many cultivars of lupine that have different chemical compositions [2,6,7], white lupins (Lupinus albus L.-WL) [8] are characterized by the highest yield-generating capabilities, containing (based on dry weight) 340–490 g·kg⁻¹ protein, 115–129 g·kg⁻¹ fat, 144–369 g·kg⁻¹ fibers, up to 45 g·kg⁻¹ carbohydrate, approximately 0.4 g·kg⁻¹ alkaloids and up to 45 g·kg⁻¹ ash [3,8]; this makes native WL a valuable product, increasing sustainability and biodiversity [4]. The growing demand for plant-based fats, protein-rich foodstuffs, and animal feed in the EU market suggests a promising opportunity for expanding the use of WL seeds [2,5,6,8,9]; at the same time, the growing amount of crop residues, like white lupine straw (WLS), and the accompanying need to develop this also suggests the need to expand the use of WL seeds, in addition to the already existing [2,8,10] effective solutions for its management. This is in accordance with the concept of biorefinery processes and in line with the concept of biorefinery processes [11].

In line with the concepts of biorefinery, circular bioeconomy, and biodiversity [12,13], chemical processing can be performed to yield valuable small-molecule products. Through the deconstruction of lignocellulosic materials via hydrolysis, primary products like C5 and C6 monosaccharides can be produced and, through biological or chemical valorization processes, converted into products of practical interest [14,15]. The monosaccharide fraction derived from white lupin straw can also be processed through acid-catalyzed dehydration to produce key intermediate platform chemicals like furfurals [16,17,18,19]. Among these, 5-hydroxymethylfurfural (5-HMF) stands out as a highly valuable biobased chemical. It is widely used to produce various bulk and fine chemicals, such as monomers, pharmaceuticals, agrochemicals, flavors, fragrances, and precursors, for liquid fuels and blending components [16,17,19,20,21,22,23]. These biobased products are increasingly replacing or supplementing those derived from petrochemical sources, thus playing a significant role in the chemical, food, and pharmaceutical industries [17,21,23].

There are several established methods for processing waste biomass [19,21,24,25]. One up-and-coming yet underutilized technique for converting lignocellulosic biomass is processing with subcritical water [26]. Water in its subcritical state exhibits significantly different properties compared to its behavior under normal conditions [27,28,29]. Under these conditions, subcritical water acts as both an acid and base catalyst [25,29], making it an ideal reagent, solvent, and catalyst for processes aligned with the principles of biorefinery, circular bioeconomy, and biodiversity [24,26,30,31]. To the best of our knowledge, no specialized investigations have been conducted on the hydrothermal treatment of white lupin straw.

In this study, we propose a biorefinery concept centered on the utilization of WLS, focusing on its conversion through hydrothermolysis into a mixture of monosaccharides and a separate fraction of 5-HMF. Integrating white lupin straw into biorefinery processes represents a significant opportunity to not only advance environmental sustainability but also promote circular economy practices [24].

The hydrothermal treatment of the polysaccharide components in WLS was conducted under mild subcritical conditions at 200 °C, utilizing trace amounts of an inexpensive and strong acid catalyst, H₂SO₄. This process was carried out in two stages: (1) the hydrolysis of WLS, which is rich in hemicellulose, to produce monosaccharides, and (2) the conversion of the remaining solid post-reaction residue (PSR), which is rich in cellulose, into a separated fraction of 5-HMF. The impact of key reaction parameters, including the reaction temperature (T) (°C) and holding time (i.e., the duration for which the reaction mixture was maintained at the prescribed temperature, denoted as t_h (min)) on the yields of monosaccharides, 5-HMF, and other primary and secondary low-molecular-weight products was thoroughly investigated. To achieve the optimal yields of monosaccharides and 5-HMF, we employed an Optimal Experimental Design (OED) methodology. This approach allowed us to establish a statistically significant reaction model and optimize the reaction conditions to maximize the production of both monosaccharides and 5-HMF.

2. Materials and Methods

2.1. Materials and Chemicals

For all experiments conducted in this study, the white lupin straw (WLS) was harvested from a farm in Drożków, a town in the Żary district of Lubuskie province, Poland. The farm is located at the geographical coordinates 51°41′01″ North, 15°04′17″ East. The WLS was planted in the third week of April 2023 and harvested between late August and early September 2022. Throughout the growing season, the agrometeorological conditions adhered to the long-term averages typical of a temperate oceanic climate.

Immediately following harvest, the WLS was transported to the laboratory, where it underwent preliminary processing. It was ground to a grain size of less than 1 mm, dried at 50 °C until the weight stabilized, and stored in a jute bag under room-temperature and dry conditions until further use. Prior to each experiment, the WLS was subjected to additional drying in a vacuum dryer at 50 °C to achieve a constant weight, ensuring consistent material properties for the hydrothermal treatment.

A detailed description of all chemicals used in this research is provided in the Supplementary Materials.

2.2. Reactor and Experimental Procedure

WLS and PSR hydrothermal treatment was performed using a 4576A-type batch reactor (Parr Instrument Company, Moline, IL, USA). A comprehensive description of the reactor equipment and the experimental procedure can be found in the Supplementary Materials.

Preliminary experimental results from tests with WLS and PSR helped define the initial reaction parameters, including T and t_h, which provided significant yields of monosaccharides and 5-HMF. To optimize production, we used an Optimal Experimental Design (OED), in two experimental series: the first series focused on maximizing the monosaccharide yield, while the second targeted the production of 5-HMF.

In the first experimental series, the hydrothermolysis of WLS was performed at a T ranging from 170 °C to 190 °C, with t_h from 0 to 20 min. The second series, focusing on 5-HMF production, was carried out at a T from 190 °C to 210 °C, with t_h from 0 to 10 min. In all experiments, the pressure inside the reactor corresponded to or slightly exceeded the vapor pressure curve at the selected temperature.

For each hydrothermal treatment experiment, pure HPLC-grade water was degassed using an ultrasonic bath and purged with nitrogen to ensure optimal conditions. The WLS (10.0 g) used in series 1 or the PSR (10.0 g) used in series 2 were suspended in 90.0 g of water or a dilute sulfuric acid solution. The catalyst concentration was set at 0.049 mol L⁻¹, equivalent to a 0.479 wt% sulfuric acid solution [32,33,34]. Given the minimal ash content in the raw material, the effect of the cations present in WLS, such as Ca²⁺ and Mg²⁺, on the sulfuric acid neutralization capacity was not considered.

The combined influence of the reaction parameters—T, t_h, and the overall reaction time t (min) (assuming the instantaneous heating, holding, and cooling of the reaction mixture)—can be quantified using the severity factor (R_o). The severity function is defined as follows [35,36,37]:

$$R_o=t_h·\exp \left[{\displaystyle \frac{\left(T-100\right)}{14.75}}\right]={\displaystyle \underset{0}{\overset{t_h}{\int }}}\mathrm{e}\mathrm{x}\mathrm{p}[{\displaystyle \frac{\left(T\left(t_h\right)-100\right)}{14.75}}]· d{t}_h$$

(1)

In Equation (1), the constant 14.75 represents an empirical parameter related to the activation energy required for the cleavage of glycosidic bonds in polysaccharides and the influence of temperature. The results from the hydrothermal treatment are expressed as a function of the logarithmic severity factor (R_o), which quantifies the overall severity of the process [37].

The concept of severity can also be extended to acid-catalyzed hydrothermal treatments of WLS and PSR by transforming the standard severity factor equation into the combined severity factor (CSF). This transformation involves incorporating the effect of the acid concentration, specifically the pH, into Equation (2) [33,37,38]:

$$CSF=\log \left(R_o\right)-pH$$

(2)

2.3. Separation of Hydrothermal Treatment Products from White Lupin Straw

The hydrothermal treatment of both WLS and PSR resulted in two primary products: a water-soluble fraction (WS) and a solid post-reaction residue (WI fraction). The WS fraction was separated from the solid residue by filtration under reduced pressure. The dry matter content in the WS fractions was determined after evaporating the water in a vacuum dryer at 50 °C. Similarly, the WI fractions were dried under the same conditions until a constant solid mass was achieved.

The yields (Y_i) (g kg⁻¹) of all reaction products, namely the WS and WI fractions, along with individual substances contained in the WS fractions, were quantified. The yields were expressed in terms of the product mass (g) per kilogram of the initial raw material loaded into the reactor vessel. A more detailed explanation of the methodology and calculations can be found in the Supplementary Materials.

Isolation of 5-HMF

In experimental series 2, 5-HMF was isolated from the WS fractions using a liquid–liquid extraction method, as described in detail by Liu et al. [39]. A precise description of the 5-HMF isolation procedure from the reaction mixture is provided in the Supplementary Materials.

2.4. Analysis and Analytical Methods

The chemical composition of WLS and PSR, including the dry matter content, ash, total nitrogen, and protein (where a factor of 6.25 was used to convert total nitrogen into protein content), was determined using standard methods for plant biomass analysis [40,41,42,43,44]. The fat content was quantified via Soxhlet extraction, and the analysis also included measurements of acid detergent fiber, neutral detergent fiber, structural carbohydrates (hemicellulose and cellulose), sulfuric acid-insoluble lignin, and sulfuric acid-insoluble lignin.

The elemental composition of WLS and PSR (C, H, N, and O) was determined using a Vario EL III analyzer (Elementar Analysensysteme, GmbH, Langenselbold, Germany). Additionally, the WS fractions were analyzed for selected monosaccharides, carboxylic acids, 2-furfural (2-FA), and 5-HMF using high-performance liquid chromatography (HPLC). The operating conditions for all chromatographic analyses are described in the Supplementary Materials. All analytical determinations were conducted in triplicate, and the mean values were calculated for accuracy.

2.5. Modeling and Optimization Method

The experimental design methodology (OED) was employed to establish the optimal experimental conditions resulting from the hydrothermal conversion of WLS and PSR; specifically, the independent variables of temperature (T) and holding time (t_h) were used to maximize the yields of selected monosaccharides (Y_ms) and 5-HMF (Y_5-HMF) in the WS fractions.

Among the various experimental designs based on a quadratic model, the Doehlert array was chosen as the experimental matrix for its efficiency and ability to explore a wide range of conditions. A detailed description of this methodology is provided in the Supplementary Material and has been previously published elsewhere [45,46].

3. Results and Discussion

3.1. Composition of WLS

The elemental composition of WLS was determined to be carbon (46.25%), hydrogen (6.45%), oxygen (46.03%), and nitrogen (1.27%), corresponding to the molecular formula [C_3.85H_6.45O_2.88N_0.09]. The dry matter content of the WLS was found to be 921.6 g·kg⁻¹ (±1.5 g·kg⁻¹), with an ash content of 50.2 g·kg⁻¹ (±2.4 g·kg⁻¹) and a total protein content of 48.7 g·kg⁻¹ (±0.4 g·kg⁻¹). The diethyl ether extractable substances were present in small amounts, at 0.8 g·kg⁻¹ (±0.1 g·kg⁻¹).

The composition further revealed a significant hemicellulose content, measuring 134.5 g·kg⁻¹ (±2.2 g·kg⁻¹), along with acetyl groups at 21.3 g·kg⁻¹ (±1.4 g·kg⁻¹), cellulose at 415.4 g·kg⁻¹ (±3.8 g·kg⁻¹), and sulfuric acid-insoluble lignin at 198.2 g·kg⁻¹ (±2.6 g·kg⁻¹). The density of the WLS was 0.129 g·mL⁻¹, while the pH of a 100 g·L⁻¹ water slurry measured at 5.88.

3.2. Modeling of Hydrothermolysis of WLS, Using Optimal Experimental Design

Both the preliminary experiments and optimization studies were carried out under mild hydrothermal conditions at 200 °C, as sulfuric acid is primarily dissociated at this temperature [45]. At the same time, its dissociation decreases at higher temperatures [47,48]. Additionally, monosaccharides, especially pentoses, exhibit instability in high-temperature and acidic environments [26].

The preliminary experiments with the hydrothermolysis of WLS helped establish the initial reaction parameters, precisely the T and t_h. These results were used to optimize monosaccharide productivity in the water-soluble (WS) fractions (series 1). The data obtained were analyzed using the optimal experimental design (OED) methodology, with T and t_h as the primary hydrothermolysis parameters.

The experimental points from the Doehlert matrix, presented in terms of both real and normalized variables, and the results obtained within the defined experimental area are displayed in

Table 1. Doehlert matrices for experimental series 1: normalized (x_i) and effective (u_i) variables, and values of corresponding experimental and calculated responses, i.e., sum of monosaccharides yield (Y_ms) for hydrothermolysis of WLS.

Experiment	x1	x2	u1 T (°C)	u2		Yms-exp. (2) (g·kg−1)	Yms-cal. (3) (g·kg−1)
				th (min)	t (min)
1	+1	0	190	10	12.4	116.7	118.8
2	−1	0	170	10	11.8	143.7	141.6
3	+0.5	+0.866	185	20	23.1	99.4	97.3
4	−0.5	−0.866	175	0	2.1	113.5	115.6
5	+0.5	−0.866	185	0	2.8	141.1	139.0
6	−0.5	+0.866	175	20	22.1	141.2	143.3
7	0	0	180	10	11.5	173.8	161.4
7′ (1)	0	0	180	10	12.4	155.9	161.4
7″ (1)	0	0	180	10	11.8	154.4	161.4

⁽¹⁾ Experiments repeated at the center of the experimental region to calculate the standard deviation on the response: s_Y = 9.3 (Y_ms). ⁽²⁾ experimental values, ⁽³⁾ calculated values.

.

The experiments conducted at short holding times and low temperatures (Experiments 2, 4, and 5), as well as those performed with long holding times and high temperatures (Experiments 1, 3, and 6), resulted in relatively low monosaccharide yields (Y_ms). The highest yield of monosaccharides, 161.4 g kg⁻¹, was achieved near the center of the experimental region, at a T of 180 °C and a t_h of 10 min. Based on the experimental responses for Y_ms in the WS fractions obtained from the hydrothermolysis of WLS in series 2.1 (Table 1), the following quadratic polynomial model (Equation (3)) was calculated:

$${\mathrm{Y}}_{\mathrm{m}\mathrm{s}}=161.4-11.4{\mathrm{x}}_1-4{\mathrm{x}}_2-31.2{\mathrm{x}}_1^2-39.7{\mathrm{x}}_2^2-40.1{\mathrm{x}}_1{\mathrm{x}}_2$$

(3)

In experimental series 1, the standard error for the coefficients b₀, b₁, and b₂ was 5.4, while for b₁₁ and b₂₂, it was 8.5, and for b₁₂, it was 10.8. Based on this model, two-dimensional (Figure 1a) and three-dimensional (Figure 1b) representations of the monosaccharide yield in the WS fractions, as a function of the hydrothermolysis reaction temperature and holding time of WLS, were generated using the N.E.M.R.O.D.W. software (version 2002, France, www.nemrodw.com).

The hydrothermolysis of the hemicellulose and the partial hydrothermolysis of the cellulose present in the WLS into monosaccharides was most efficient at temperatures ranging from 174.6 °C to 181.1 °C, with t_h between 7.3 min (t = 9.4 min) and 14.1 min (t = 16.0 min) (Figure 1a). The predicted optimal yield of monosaccharides (Y_ms) was 162.5 g·kg⁻¹ ± 5.2 g·kg⁻¹, which occurred at T of 177.4 °C and a t_h of 10.8 min (t = 11.6 min). This was in close agreement with the experimental results, with a coefficient of determination (R²) of 0.942, indicating a strong correlation between the model and the observed data.

These findings suggest that the temperature and holding time are crucial in optimizing monosaccharide yields from WLS, which is consistent with previous studies on the hydrothermal treatment of lignocellulosic biomass [20,30,49,50,51]. The model provides a clear understanding of the ideal conditions required for maximum monosaccharide production and contributes to the broader goal of utilizing agricultural waste for bio-based chemical production.

3.3. Composition of WS and WI Fractions—Products of Hydrothermolysis of WLS

The dilute sulfuric acid hydrothermal treatment of white lupin straw led to the recovery of various pretreated solids and hydrolyzed compounds, depending on the material composition and treatment conditions. Dilute sulfuric acid fractionation typically causes the solubilization of hemicellulose and a small portion of lignin while hydrolyzing the solubilized hemicellulose and decreasing cellulose crystallinity. However, most lignin remains as a solid residue, with only the partial cleavage of ether and ester linkages, generating low-molecular-weight lignin fragments enriched with hydroxyl groups [15,37].

The composition of selected components in the WS fractions obtained from WLS hydrothermolysis in experimental series 1 is presented in Figure 2.

Under the tested conditions (Exps. 1, 3, 6, 7), hemicellulose hydrolysis was nearly selective, producing minimal amounts of glucose in the WS fractions. The presence of an acid catalyst and hydronium ions, generated through the autoionization of water and the release of acetic acid from WLS, facilitated the hydrolysis of hemicellulose and limited the breakdown of weakly bound cellulose fractions [33,49,50,51].

The yields of predominant monosaccharides, such as xylose and glucose, were lower than those obtained through conventional acid hydrolysis [21,49,52]. However, these monosaccharides were produced in a much shorter time. In the case of xylose, the results were comparable to those achieved by the hydrothermal treatment of other biomass, such as rapeseed straw [35].

An increase in the reaction temperature and holding time likely contributed to the partial decomposition of pentoses (Exp. 3), leading to the formation of non-fermentable products such as aldehydes, carboxylic acids, and furfurals [21,49,52,53].

Acetic acid and secondary reaction products like 2-FA and 5-HMF, which are present in the hydrolysates, can inhibit microbial activity during monosaccharide fermentation. However, in this study, the maximum concentration of acetic acid was 21.4 g·kg⁻¹ (Exp. 7), and the highest total furfural yield was 32.4 g·kg⁻¹ (Exp. 3). These concentrations are lower than the levels considered to be inhibitory to fermentation processes [54].

The reaction conditions applied in this study allowed for the prediction of non-sugar component yields in the water-soluble (WS) fractions based on the R_o and CSF experiments (

Table 2. The yield of WI fractions obtained in experimental series 1 and the content of unreacted hemicellulose, cellulose and sulfuric acid-insoluble lignin in solid post-reaction residues as a function of the severity factor and corrected severity factor.

Experiment	Experimental Series 1
	Log Ro	CSF	YWI (1) (g·kg−1)	Content in the WI Fractions (g·kg−1)
				H (2)	C (3)	L (4)
1	3.74	1.89	497.0	8.9	528.9	350.6
2	3.13	1.33	538.5	22.1	505.5	349.5
3	3.87	1.95	502.1	8.5	532.0	362.1
4	2.53	0.78	572.8	61.1	501.5	296.7
5	2.95	1.15	532.0	48.7	520.0	317.6
6	3.55	1.75	517.4	15.4	511.1	342.2
7 (5)	3.43	1.59	515.4	15.6	513.5	327.5

⁽¹⁾ Yield of the WI fractions. ⁽²⁾ hemicellulose. ⁽³⁾ cellulose. ⁽⁴⁾ sulfuric acid insoluble lignin. ⁽⁵⁾ mean values for three repeated experiments.

) [33,37]. The results demonstrated a clear interdependence between the yield of non-monosaccharide components and log Ro and CSF. For experiments where the log Ro and CSF values were relatively low, approximately 3 and 1, respectively (Exps. 2, 4, 5), the yields of secondary hydrothermolysis products were low. Conversely, in experiments where log R_o and CSF achieved higher values (Exps. 1, 3, and 6), there was a marked increase in the yield of secondary reaction products.

The detailed mechanism of the decomposition of hemicellulosic primary products and the partial hydrothermolysis of the cellulosic fractions of WLS remains complex. The degradation products are varied and can be formed through multiple reaction pathways, making it difficult to pinpoint the exact decomposition mechanism [53,55].

The yields of the remaining solid post-reaction residue (WI fractions) and the contents of unreacted hemicellulose, cellulose, and sulfuric acid-insoluble lignin are summarized as a function of the severity factor and combined severity factor in Table 2. The WI fraction yields ranged from 497.0 g·kg⁻¹ to 538.6 g·kg⁻¹, decreasing as hemicellulose and partial cellulose solubilization progressed. The hemicellulose content decreased with increasing temperature and holding time, with trace amounts of hemicellulose detected in WI fractions when the log Ro values were between 3.74 and 3.87 and the CSF values were between 1.89 and 1.95. The remaining unreacted lignin and cellulose contents in the WI fractions were still high, ranging from 501.5 g·kg⁻¹ to 528.9 g·kg⁻¹ for cellulose and from 296.7 g g·kg⁻¹ to 350.6 g·kg⁻¹ for sulfuric acid-insoluble lignin.

Under the applied reaction conditions, the cellulose and lignin in WLS were found to be more thermally stable than hemicellulose. They did not undergo significant hydrothermolysis, as indicated by the low numerical values of the severity and combined severity factors. The sulfuric acid-soluble lignin was absent from the WI fractions, highlighting the challenge of lignin dissolution. Studies have shown that during most hydrothermal treatment processes, only a small fraction of lignin is solubilized and depolymerized. In contrast, the remaining soluble lignin fragments may recondense or repolymerize [26,50].

This recondensation and repolymerization of lignin fragments might explain why the lignin content in the WI fractions was relatively high, or even comparable to, or higher than, the initial content prior to the reaction. This suggests that lignin did not dissolve effectively under the conditions used and supports the observation of its thermal stability during the hydrothermal process.

3.4. Modeling of Hydrothermal Treatment of PSR, Using Optimal Experimental Design

In the second experimental series, which focused on optimizing 5-HMF production, the solid reaction residue obtained from the first series helped to establish the optimal experimental conditions at 177.4 °C and a t_h of 10.8 min. The PSR (plant source residue) used in this study had a molecular formula of C_4.19H_6.45O_2.71, with an elementary composition of carbon (50.26%), hydrogen (6.4%), and oxygen (43.33%). The ash content of PSR was 5.0 g·kg⁻¹ ± 0.3 g·kg⁻¹, total protein was 2.8 g·kg⁻¹ ± 0.2 g·kg⁻¹, the diethyl ether extractable substance was 0.1 g·kg⁻¹ ± 0.01 g·kg⁻¹, hemicellulose was 14.1 g·kg⁻¹ ± 1.1 g·kg⁻¹, cellulose was 512.8 g·kg⁻¹ ± 4.0 g·kg⁻¹, and sulfuric acid-insoluble lignin was 329.1 g·kg⁻¹ ± 2.9 g·kg⁻¹. The density of PSR was 0.144 g·mL⁻¹, and the pH of the 100 g·L⁻¹ water slurry was 4.92.

The experimental conditions, presented in

Table 3. Doehlert matrices for experimental series 2: values of corresponding experimental and calculated responses, i.e., yield (Y_5-HMF) for hydrothermal treatment of PSR. Normalized (x_i) variables are presented in Table 1.

Experiment	u1 T (°C)	u2		Y5-HMF-exp. (2) (g·kg−1)	Y5-HMF-cal. (3) (g·kg−1)
		th (min)	t (min)
1	210	5	6.8	108.8	105.3
2	190	5	7.3	105.6	109.1
3	205	10	12.1	104.1	107.6
4	195	0	2.1	127.9	124.4
5	205	0	2.8	131.4	134.9
6	195	10	12.1	125.3	121.8
7	200	5	7.5	138.8	138.1
7′ (1)	200	5	7.4	138.0	138.1
7″ (1)	200	5	7.8	137.5	138.1

⁽¹⁾ Experiments repeated at the center of the experimental region to calculate the standard deviation on the response: s_5-HMF = 0.66 (Y_5-HMF). ⁽²⁾ experimental values. ⁽³⁾ calculated values.

, show that the production of 5-HMF was most favorable at around 200 °C and with a t_h of approximately 5 min (Exp. 7). On the other hand, lower temperatures combined with shorter holding times (Exp. 2) and higher temperatures with longer holding times (Exp. 1 and Exp. 3) did not result in high 5-HMF productivity. These conditions were insufficient for the effective hydrolysis of the cellulose fraction in PSR. Under such conditions, the cellulose hydrolyzed too slowly, leading to low yields of 5-HMF [56]. When the temperature and holding time were too high, the glucose formed through hydrothermolysis of the cellulose fraction further reacted into non-target compounds, reducing the yield of 5-HMF [53].

The highest 5-HMF yield was 138.1 g·kg⁻¹, achieved near the center of the experimental region (Exps. 7, 7′, and 7″).

The hydrothermal conversion of PSR to 5-HMF in the presence of dilute H₂SO₄ proceeds in two main stages: the hydrolysis of cellulose to monosaccharides, mainly glucose, and its dehydration to 5-HMF. During the hydrolysis reaction, which follows the Brønsted acid-catalyzed mechanism, H⁺ ions and water work together to cut off the C–O bonds in the β-1,4-glycosidic bond and in the pyranose ring of cellulose [57]. The course of both hydrolysis and dehydration is highly dependent on the ion concentration: from the self-dissociation of water, the H₂SO₄ catalyst and acetic acid, which is the product of the acetyl cleavage from acetylated saccharides monomers of residual amounts of the hemicellulose contained in the PSR [53]. Under subcritical water conditions, elevated temperatures increase the ionic product of water, resulting in a greater concentration of H⁺ and OH⁻ ions [58]. These ions can act as catalysts, facilitating the breakdown of glycosidic bonds in cellulose. H⁺ ions specifically aid in protonating oxygen atoms within glycosidic bonds, making them more susceptible to nucleophilic attack, which promotes cleavage and results in the release of glucose units [57]. The presence of acidic catalyst further accelerates the hydrolysis process of cellulose, reduces its crystallinity and enhances its accessibility to reactive sites [59]. After the hydrolysis of cellulose, the presence of acidic conditions favors the dehydration of glucose into 5-HMF. In this step, the H⁺ ions facilitate the removal of water molecules from glucose through sequential dehydration reactions, leading to 5-HMF as the final product [21].

Attempts to fit the experimentally determined 5-HMF yield into a quadratic model (Equation (4)) yielded satisfactory results. The predicted values aligned well with the experimental data, falling within the acceptable range of experimental errors. The following quadratic model was calculated using multilinear least square regression based on the data obtained from this experimental series:

$${\mathrm{Y}}_{5-\mathrm{H}\mathrm{M}\mathrm{F}}=138.1-1.9{\mathrm{x}}_1-8.6{\mathrm{x}}_2-30.9{\mathrm{x}}_1^2-10.9{\mathrm{x}}_2^2-14.3{\mathrm{x}}_1{\mathrm{x}}_2$$

(4)

In the experimental series focused on optimizing 5-HMF production, the standard error for the coefficients was as follows: for b₀, b₁, and b₂, it was 0.38; for b₁₁ and b₂₂, it was 0.6; and for b₁₂, it was 0.76. The two-dimensional (a) and three-dimensional (b) representations of the 5-HMF yield (Y_5-HMF) in the WS fractions, as a function of the reaction temperature and holding time, were calculated using these models and are shown in Figure 3. The contour plots in Figure 3a represent constant value curves for Y5-HMF, as predicted by the quadratic model.

The optimal conditions for the hydrothermal production of 5-HMF from PSR were achieved in a T range of 198.7–202.8 °C and a t_h range of 0.5–4.5 min (t = 1.2–5.0 min). The highest estimated 5-HMF yield was 139.9 g kg⁻¹ (±0.4), obtained at 200.7 °C and a t_h of 2.5 min (t = 3.3 min). These predicted values were in good agreement with the experimental results, showing a high correlation coefficient (R² = 0.955).

This demonstrates the robustness of the quadratic model in predicting the optimal conditions for 5-HMF production from PSR, confirming that the temperature and holding time significantly influenced the yield, with well-defined optimal ranges.

3.4.1. Isolation of 5-HMF from Hydrothermal Conversion of Aqueous Product Fractions

5-HMF was removed and recovered from the WS fractions obtained under the optimal conditions of the experimental design series (series 2). The liquid–liquid extraction method, as described by Liu et al. [39], was employed for isolating 5-HMF. After extraction using an n-butanol–THF mixture, the 5-HMF content in the WS fraction significantly decreased, reducing to 49.0 g·kg⁻¹ (35.7% of the initial value). The subsequent evaporation of solvents from the organic phase allowed for the complete recovery of 5-HMF, achieving 100% efficiency in the process. This efficient recovery process underlines the effectiveness of the optimized extraction method, ensuring minimal loss of the valuable 5-HMF compound during the separation from the reaction mixture.

3.4.2. Composition of WS and WI Fractions—Products of Hydrothermal Treatment of PSR

The composition of the main selected components contained in the WS fractions resulting from the hydrothermal treatment of PSR in experimental series 2 is detailed in Figure 4.

It includes the yields of significant products such as monosaccharides and by-products like 5-HMF, which highlight the treatment’s effectiveness under varying conditions.

The yields of WI fractions after the hydrothermal treatment of PSR in experimental series 2 are summarized in

Table 4. The yield of WI fractions obtained in experimental series 2 and the content of unreacted hemicellulose, cellulose and sulfuric acid insoluble lignin in solid post-reaction residues as a function of the severity index and corrected severity factor.

Experiment	Experimental Series 2
	Log Ro	CSF	YWI (1) (g·kg−1)	Content in the WI Fractions (g·kg −1)
				H (2)	C (3)	L (4)
1	3.65	2.39	476.6	0.0	8.1	892.2
2	3.51	2.28	570.1	0.0	122.3	670.0
3	4.18	2.89	469.8	0.0	7.6	930.7
4	3.12	1.89	547.1	0.0	74.8	657.2
5	3.54	2.31	503.8	0.0	12.6	745.8
6	3.88	2.64	525.6	0.0	14.7	853.3
7 5	3.83	2.61	507.5	0.0	5.2	859.9

⁽¹⁾ yield of the WI fractions. ⁽²⁾ hemicellulose. ⁽³⁾ cellulose. ⁽⁴⁾ sulfuric acid insoluble lignin. ⁽⁵⁾ mean values for three repeated experiments.

. These yields represent the unreacted and insoluble products, primarily hemicellulose, cellulose, and lignin, as a function of the severity index and combined severity factor. The resulting WI fractions ranged from 469.8 g·kg⁻¹ to 570.1 g·kg⁻¹, showing a decrease due to the solubilization and hydrolysis of residual hemicellulose and cellulose, as well as complex transformations of lignin. Notably, hemicellulose was entirely absent in the WI fractions, and only trace amounts of cellulose were detected, except for the WI fraction obtained at the lowest temperature and shortest holding time (Exp. 2). The most effective cellulose solubilization and depolymerization occurred at log Ro values between 3.54 and 4.18 and CSF values between 2.31 and 2.64.

In the WI fractions, the content of unreacted and remaining insoluble sulfuric acid components was still relatively high, ranging from 657.2 g·kg⁻¹ to 930.7 g·kg⁻¹. These non-dissolved organic materials obtained post-hydrolysis included reaction products such as glucose char, cellulose char [60], and other undesirable products from the thermal decomposition of components in the reaction mixture.

3.5. Overall Mass Balance

The material balances for the hydrothermal treatment of 1.0 kg of WLS and 1.0 kg of PSR, as outlined in Figure 5, reveal key insights into the effects of various conditions on product yields and by-products.

In experimental series 1, milder hydrolysis conditions for WLS (Exps. 2 and 5) did not significantly promote the solubilization of the raw material. However, these conditions were optimal for achieving the highest recovery of monosaccharides, with the minimal formation of secondary products, losses, and gas fractions. This shows that hydrolysis was more selective for monosaccharide production under mild conditions.

Conversely, when higher reaction temperatures (Exps. 1 and 3) or longer holding times (Exp. 6) were applied, the process led to an increased yield of secondary hydrothermal treatment products, higher losses, and larger gas fractions. These more intense conditions reduced the yield of WI fractions and primary reaction products, as the harsher treatment likely accelerated degradation and side reactions, leading to the formation of by-products rather than desired monosaccharides.

For the second experimental series, which focused on PSR, the results showed that higher temperatures and longer holding times (Exps. 1 and 3) facilitated solubilization but did not yield satisfactory amounts of 5-HMF in the WS fractions. In contrast, the experiments that yielded the highest recovery of 5-HMF (Exps. 4, 5, and 7) resulted in the formation of the most solid reaction residue, indicating that optimal 5-HMF production conditions also increased the solid fraction remaining after the reaction. This suggests that achieving high 5-HMF yields requires a balance between the temperature and reaction time to minimize the breakdown of cellulose into undesired by-products while maximizing dehydration reactions that produce 5-HMF.

During the first step of the process, the hydrothermal treatment of WLS in the presence of H₂SO₄ ensured the efficient recovery of hemicellulose through the auto-hydrolysis of the biomass. These findings align with previous studies, as extensively reported in works such as [14,26,30,32,33,35,54]. The deconstruction process occurred under typical conditions for this type of reaction, within a temperature range of 160–260 °C, pressures of 0.7–4.8 MPa, and times shorter than 15 min. The process offers several advantages, including moderate energy requirements, feasibility for industrial-scale applications, a low production of degraded by-products, reduced cellulose crystallinity, and high sugar recovery [26,30,33].

At the same time, even though the use of dilute H₂SO₄ has several advantages, such as a low cost and appropriate conversion yields, this causes equipment corrosion. Consequently, the use of acid requires suitable reactors to prevent the corrosion of the equipment [14].

In the second experimental series, under acidic conditions, the cellulose in the PRS was readily hydrolyzed and dehydrated into 5-HMF, which was obtained with moderate yield. However, the commercial production of 5-HMF from glucose appears to be a more viable option [21].

4. Conclusions

White lupin straw, an agricultural by-product that is generally underutilized, holds significant potential for producing valuable compounds such as monosaccharides and 5-HMF. The composition of WLS makes it a promising raw material for chemical processing, aligning with the principles of biorefinery, circular bioeconomy, and biodiversity.

This study investigated the hydrothermal treatment of WLS under subcritical water conditions, specifically at 200 °C with trace amounts of H₂SO₄. The severity factor and the combined severity factor were crucial in influencing the course of WLS hydrothermolysis. Optimal monosaccharide and 5-HMF yields were achieved within the mid-range of severity values, balancing solubilization efficiency and product selectivity.

The highest monosaccharide yield, approximately 163 g·kg⁻¹, was achieved in the T range of 174–181 °C, with a t_h between 7 and 14 min. The highest yield of 5-HMF, about 139.9 g·kg⁻¹, was obtained at slightly higher temperatures, ranging from 199 to 203 °C, with a t_h between 0.5 and 4.5 min.

The monosaccharides can be further processed into bioethanol, as the levels of fermentation inhibitors are low and within acceptable limits. Meanwhile, 5-HMF is a versatile compound with numerous applications, including the production of chemicals, fuels, and energy.

Scaling up the biorefinery process for WLS can offer significant economic and environmental benefits. WLS can contribute to more sustainable agricultural practices. However, careful consideration of techno-economic feasibility, regulatory frameworks, and economic freedom will be necessary for successful commercialization and large-scale adoption. Balancing these factors will be critical to the long-term success and sustainability of WLS biorefinery operations.