Recovering rhamnogalacturonan-I pectin from sugar beet pulp using a sequential ultrasound and microwave-assisted extraction: Study on extraction optimization and membrane purification

Bioresource Technology Recovering rhamnogalacturonan-I pectin from sugar beet pulp using a sequential ultrasound and microwave-assisted extraction:study on extraction optimization and membrane purification --Manuscript Draft-- Manuscript Number: BITE-D-23-06332R3 Article Type: Original research paper Section/Category: New & emerging biomass-based processes Keywords: Pectooligosaccharides; Box-Behnken design; Green extraction; ultrafiltration; diafiltration Corresponding Author: Susana Lucas, Ph. D. University of Valladolid Valladolid, Valladolid SPAIN First Author: Esther del Amo-Mateos Order of Authors: Esther del Amo-Mateos Berta Cáceres Mónica Coca M. Teresa García-Cubero Susana Lucas, Ph. D. Abstract: This research focuses on the integrated recovery of rhamnogalacturonan-I (RG-I) pectin from sugar beet pulp (SBP). First, the extraction of RG-I pectin through sequential ultrasound-assisted extraction (UAE) and microwave-assisted extraction (MAE) was assessed. Optimization using a response surface methodology identified the optimal conditions as initial pH 4, 10 minutes of UAE, and 157 °C for MAE, achieving a 66.0% recovery of pectooligosaccharides (POS). Additionally, purification through continuous diafiltration and concentration via ultrafiltration of the POS using membranes with different molecular weight cut-offs (MWCO) was explored. In contrast to previous research using discontinuous diafiltration, the use of continuous diafiltration allowed a decrease in the extract viscosity and obtained higher yields using a higher MWCO membrane. The refined RG-I pectin solids exhibited a high global yield (39 – 40 g pectin/100 g SBP), and high-methoxyl characteristics, as well as purity levels (70 – 80%) similar to commercial prebiotics. Powered by Editorial Manager® and ProduXion Manager® from Aries Systems Corporation Sugar beet pulp Impurities Extraction process Sequential UAE –MAE Box-Behnken Design: Initial pH, UAE-time, MAE-temperature Purification process Continuous Diafiltration + Ultrafiltration Membranes: 3and 5 kDa Water Characterization 100 g 0 20 40 60 0.1 10 1000 MV Mw (kDa) Freeze-drying RG-I rich pectin 39.7 g POS Others POS* yield: 66.0 % Ultrasound-assisted extraction (UAE) Microwave-assisted extraction (MAE) POS recovery: 90.0 % DF UF POS: pectooligosaccharides Graphical Abstract (for review) Efficient RG-I pectin recovery with sequential UAE/MAE and membrane purification. Use of short-time UAE decreased MAE time and temperature. Optimal recovery of 66.0 % of pectooligosaccharides from sugar beet pulp. The 90 % of POS were recovered using continuous DF and UF with 3 kDa MWCO membrane. High-purity RG-I pectin with high-methoxyl attributes similar to prebiotics. Highlights (for review) 1 Recovering rhamnogalacturonan-I pectin from sugar beet pulp using a sequential 1 ultrasound and microwave-assisted extraction: study on extraction optimization and 2 membrane purification 3 Esther del Amo-Mateosa,b, Berta Cáceresb, Mónica Cocaa,b, M. Teresa García-Cuberoa,b, 4 Susana Lucasa,b,* 5 a Institute of Sustainable Processes. University of Valladolid, Spain 6 b Department of Chemical Engineering and Environmental Technology, School of Industrial 7 Engineering, University of Valladolid, Dr. Mergelina, s/n, Valladolid, Spain 8 *Corresponding author. E-mail address: susana.lucas.y[email protected] (S. Lucas) 9 10 E-mail addresses: [email protected] (E. Del Amo-Mateos), 11 [email protected]a.es (B. Cáceres), [email protected] (M. Coca), 12 [email protected] (M. T. García-Cubero), [email protected] (S. Lucas) 13 Abstract 14 This research focuses on the integrated recovery of rhamnogalacturonan-I (RG-I) pectin 15 from sugar beet pulp (SBP). First, the extraction of RG-I pectin through sequential 16 ultrasound-assisted extraction (UAE) and microwave-assisted extraction (MAE) was 17 assessed. Optimization using a response surface methodology identified the optimal 18 conditions as initial pH 4, 10 minutes of UAE, and 157 °C for MAE, achieving a 66.0 % 19 recovery of pectooligosaccharides (POS). Additionally, purification through continuous 20 diafiltration and concentration via ultrafiltration of the POS using membranes with different 21 molecular weight cut-offs (MWCO) was explored. In contrast to previous research using 22 discontinuous diafiltration, the use of continuous diafiltration allowed a decrease in the 23 extract viscosity and obtained higher yields using a higher MWCO membrane. The refined 24 RG-I pectin solids exhibited a high global yield (39 – 40 g pectin/100 g SBP), and high25 methoxyl characteristics, as well as purity levels (70 – 80 %) similar to commercial 26 prebiotics. 27 Title Page with Author Information 1 Recovering rhamnogalacturonan-I pectin from sugar beet pulp using a sequential 1 ultrasound and microwave-assisted extraction: study on extraction optimization and 2 membrane purification 3 Abstract 4 This research focuses on the integrated recovery of rhamnogalacturonan-I (RG-I) pectin 5 from sugar beet pulp (SBP). First, the extraction of RG-I pectin through sequential 6 ultrasound-assisted extraction (UAE) and microwave-assisted extraction (MAE) was 7 assessed. Optimization using a response surface methodology identified the optimal 8 conditions as initial pH 4, 10 minutes of UAE, and 157 °C for MAE, achieving a 66.0 % 9 recovery of pectooligosaccharides (POS). Additionally, purification through continuous 10 diafiltration and concentration via ultrafiltration of the POS using membranes with different 11 molecular weight cut-offs (MWCO) was explored. In contrast to previous research using 12 discontinuous diafiltration, the use of continuous diafiltration allowed a decrease in the 13 extract viscosity and obtained higher yields using a higher MWCO membrane. The refined 14 RG-I pectin solids exhibited a high global yield (39 – 40 g pectin/100 g SBP), and high15 methoxyl characteristics, as well as purity levels (70 – 80 %) similar to commercial 16 prebiotics. 17 Keywords 18 Pectooligosaccharides, Box-Behnken design, green extraction, ultrafiltration, diafiltration 19 1 Introduction 20 Pectin is a valuable polysaccharide that can be extracted from various plant sources. It 21 has many applications in the food, pharmaceutical, and biomedical sectors due to its 22 gelling, thickening, stabilizing, and emulsifying properties, as well as its health benefits as a 23 dietary fiber and prebiotic (Chandel et al., 2022). However, pectin is not a homogeneous 24 substance, but rather a complex mixture of different structural elements, such as 25 Revised manuscript (clean version) Click here to view linked References 2 homogalacturonan (HG), rhamnogalacturonan-I (RG–I), and rhamnogalacturonan-II (RG–II) 26 (del Amo-Mateos et al., 2023). Of these, the RG–I region is a branched pectic domain that 27 consists of a backbone of alternating galacturonic acid (GalA) and rhamnose units, with 28 various side chains of arabinose and galactose (Y. Mao et al., 2019). RG–I is particularly 29 interesting for its potential applications in food and pharmaceutical products, as it can 30 modulate the rheological properties of pectin gels, enhance the stability of emulsions, and 31 exhibit biological activities such as anti-inflammatory, anti-cancer, and immunomodulatory 32 effects (G. Mao et al., 2019). Therefore, recovering pectin from plant sources and 33 optimizing the extraction methods to preserve the RG–I structure are important research 34 topics that can lead to the development of novel and functional pectin-based products such 35 as pectooligosaccharides (POS). POS are known for their potential prebiotic properties with 36 a superior ability to regulate the human intestine microbiota (Y. Mao et al., 2019). 37 Pectin is extracted mainly from citrus and apple fruits due to the composition (Chandel et 38 al., 2022). However, using waste or by-products can enhance the industry’s sustainability 39 and feasibility. Sugar beet pulp (SBP), a by-product generated after sugar extraction, has a 40 global production of approximately 120 million tons (90 % humidity) (Bonnin et al., 2009). 41 SBP is a fibrous material rich in cellulose, hemicellulose, and pectin, with a predominant 42 presence of RG–I pectin (Y. Mao et al., 2019). Due to its availability and price, it has been 43 used for low-value applications such as animal feed, while SBP stands as a valuable raw 44 material for RG–I pectin extraction. 45 Extracting RG–I pectin is challenging, as conventional extraction of HG pectin using 46 acidic hot solutions (pH 1.5 – 3, 60 – 100 °C) over several hours is ineffective (G. Mao et 47 al., 2019). Under these conditions, the side chains present in the hairy region are degraded. 48 Moreover, commercial extraction requires large amounts of solvent and high energy and 49 water consumption (Chandel et al., 2022). Emerging technologies such as ultrasound and 50 3 microwave-assisted extraction have received research attention to address the 51 conventional extraction limitations. 52 Ultrasound-assisted extraction (UAE), a non-thermal extraction method, utilizes high53 frequency sound waves to create cavitation bubbles that disrupt plant tissues. Thus, the 54 solvent can easily penetrate the cells and facilitate pectin release. Previous research 55 employed UAE for pectin recovery from waste grapefruit peel using HCl (pH 1.5) as solvent 56 obtaining a yield of 23.49 % with an RG-I content of 38.31 % (Wang et al., 2017). 57 Microwave-assisted extraction (MAE), a thermal extraction technique, employs 58 electromagnetic waves to heat the sample rapidly, promoting the breakdown of cell walls 59 and the release of pectin (Marić et al., 2018). Some advantages of MAE over conventional 60 heating are homogeneous energy dispersion and fast heating. MAE has been studied for 61 pectin extraction from SBP under alkaline conditions (pH 13) at 90 °C for 120 min achieving 62 a yield of 23.4 % rich in sugars (60.09 %) (Y. Mao et al., 2019). 63 A growing research interest is integrating ultrasound and microwave-assisted extraction 64 (UMAE) to combine their advantages. This hybrid approach can accelerate extraction and 65 minimize energy consumption, resulting in higher pectin yields and enhanced quality 66 (Gharibzahedi et al., 2019). Based on their mechanism extraction, the sequential use of 67 UAE followed by MAE may reduce MAE operation times and temperatures, thereby 68 minimizing RG-I pectin degradation and energy consumption (Liew et al., 2016). 69 Consequently, the combination of these emerging technologies offers a promising path 70 forward for the efficient and sustainable production of RG-I pectin. Gharibzahedi et al. 71 (2019) compared the acidic hot water and sequential UMAE extraction method for pectin 72 recovery from fig skin increasing the yield from 6.05 % to 13.97 %. Liew et al. (2016) 73 compared MAE, UAE, sequential UMAE and microwave-ultrasound extraction of pectin 74 from pomelo peels achieving the highest yield using UMAE (36.33 %) and the lowest with 75 4 UAE (14.25 %). However, to the best of the author’s knowledge, RG-I region extraction has 76 not previously been reported using the combination of these emerging technologies. 77 After pectin extraction, a purification method is essential to remove impurities that may 78 be present in the extract. Diafiltration (DF) and ultrafiltration (UF) are promising options due 79 to their cost-effectiveness, scalability, and low energy and solvent requirements (Ramos80 Andrés et al., 2021). However, pectin extracts are complex and can lead to membrane 81 fouling, reducing the permeate flux (Jin et al., 2022). Some research studies have applied 82 membrane processes to purify pectin by discontinuous DF (Gómez et al., 2013; Ramos83 Andrés et al., 2021). Continuous DF can mitigate membrane fouling by reducing the feed 84 solution's viscosity, while subsequent UF processes can be employed to concentrate the 85 extract. 86 This study develops an integrated process for RG-I pectin extraction from SBP through 87 sequential UMAE and a purification process of the extract obtained by continuous DF and 88 UF. The extraction process of RG-I pectin employing sequential UAE an MAE was 89 optimized using a Box-Behnken design. The purification of the optimized extract based on a 90 continuous DF and UF process was assessed and, to that end, the comparison of the 91 performance of membranes with different MWCO, according to compound recovery and 92 membrane fouling, was carried out. The characterization of the RG-I pectin-rich solids 93 obtained was studied in terms of composition, molecular weight, degree of esterification 94 and the chemical structure of the surface. To the best of author’s knowledge, this is the first 95 work proposing an integrated process using sequential UMAE and subsequent purification 96 by a continuous DF and UF membrane process, with promising results for the recovery of 97 RG-I pectin, with potential prebiotic applications, from SBP. This research proposes a 98 significant advance in the development of efficient technologies for POS recovery from a 99 5 byproduct, which is in line with the sustainable development goal 12: Ensure sustainable 100 consumption and production patterns and goal 7: Affordable and clean energy. 101 2 Materials and methods 102 2.1 Raw material 103 AB Azucarera Iberia kindly supplied SBP, which was washed, dried at 60 °C and ground 104 (particle size < 1 mm) before use. 105 2.2 Sequential UMAE for RG–I pectin extraction 106 Pectin from SBP was extracted using a two-step approach involving UAE followed by 107 MAE. UAE was conducted in ultrasound equipment with direct sonification (20 kHz) 108 (Hielscher Ultrasound Technology UIP1000hd transducer, Hielscher Ultrasonics GmbH, 109 Germany). Based on previous studies, the amplitude was 90 %, and the operating 110 temperature was 70 °C (Fernández-Delgado et al., 2023). Once the UAE was completed, 111 the mixture was subjected to MAE, performed in a closed microwave-assisted reaction 112 system (Multiwave PRO SOLV reactor 50 Hz with a Rotor type 16HF100, Anton Paar 113 GmbH, Austria, Europe). The slurry from the UAE was placed into vessels made of PTFE114 TFM (volume capacity of 100 mL) provided with magnetic stirrers. The temperature and 115 pressure of each vessel were continuously recorded by an infrared sensor, and a 116 temperature/pressure sensor controlled the microwave power (del Amo-Mateos et al., 117 2022). The MAE operation time was 9.4 min, corresponding to the optimum time extraction 118 previously established for OGalA (del Amo-Mateos et al., 2022). 119 The SBP and the diluted acid solution, prepared by adjusting the pH using H2SO4, were 120 mixed in a solid-to-liquid ratio of 10 % (w/V) (15 g dried SBP and 150 mL diluted acid 121 solution). After the UMAE, the slurry was vacuum filtrated. The solid fraction was dried at 122 60 °C and weighed for solid recovery determination (g spent solid/g SBP). The liquid was 123 12 acid/L, 1.8 g acetic acid/L and 0.7 g furfural/L). Moreover, as can be seen from Fig. 2a-e, 260 Run 7 also exhibited the highest monomer concentration accounting for 30.1 g/L (1.9 g 261 galacturonic acid/L; 0.9 g glucose/L; 4.9 g galactose/L; 1.8 g rhamnose/L and 20.6 g 262 arabinose/L). Based on this, the observed high concentration of degradation compounds 263 may be attributed to the degradation of sugars and acetyl groups linked to the 264 oligosaccharides due to the severity of the extraction conditions (pH 1, MAE temperature 265 160 °C). The same tendency was found in the study of oligosaccharides extraction from 266 Robinia pseudoacacia wood by MAE carried out by Pérez-Pérez et al. (2023), where a 267 higher concentration of degradation compounds was obtained under more severe 268 extraction conditions. Furthermore, concentrations exceeding 1 g/L were found for the 269 experimental runs performed at initial pH 1 or MAE temperature of 160 °C (Runs 3, 11, 13, 270 15), suggesting that the UAE time did not influence the concentration of these compounds. 271 POS composition in the extracts can be found in Fig. 2f. The average concentrations at 272 the central point were: 10.3 ± 0.1 g GalA/L, 1.9 ± 0.1 g galactose/L, 1.9 ± 0.1 g rhamnose/L, 273 12.2 ± 0.6 g arabinose/L, 1.0 ± 0.0 g glucose/L, and 25.3 ± 0.9 g POS/L. GalA is the main 274 component of pectin. The OGalA concentration ranged from 7.1 (Run 14) to 10.6 g/L (Run 275 10). On the other hand, the POS concentration ranged from 10.5 (Run 7) to 31.6 g/L (Run 276 3). OGalA and AraOS were the main components of the POS. Nevertheless, the 277 composition of the POS was indeed influenced by the initial pH. The degradation of sugars 278 to their monomeric form was observed in the runs conducted at pH 1 (Runs 2, 7, 10 and 279 13). Consequently, under acidic conditions, the POS composition mainly comprised OGalA. 280 The highest concentration of POS (31.6 g/L) was observed in Run 3, where AraOS 281 comprised 60.1 % of the composition, while OGalA accounted for 24.3 %. The observed 282 composition distribution aligns with previous research findings (del Amo-Mateos et al., 283 2023), where it was also pointed out that higher temperatures during the extraction process 284 13 result in more sugar extraction than GalA. The results suggest that the composition of the 285 POS is influenced by the pH and MAE temperature. However, no clear relationship was 286 found between the extraction of POS and the UAE time. 287 3.3 Optimization of POS extraction from SBP by UMAE 288 A second-order polynomial equation for the POS concentration was proposed to relate 289 the response with the independent variables (Eq. 10) 290 𝑃𝑂𝑆 (𝑔/𝐿) = −143.236 − 21.6033 · 𝑋1+ 0.425 · 𝑋2+ 2.423 · 𝑋3+ 0.105 · 𝑋1· 𝑋2+ 0.221 · 𝑋1 · 𝑋3+ 1.40 ·10−4𝑋2· 𝑋3− 1.610 · 𝑋1 2− 0.021 · 𝑋2 2− 0.010 · 𝑋3 2 (10) R2 = 0.944; R2adjusted = 0.844 291 where X1 is the initial pH, X2 is the UAE time (min) and X3 the MAE temperature (°C). 292 The quadratic model exhibited a significant fit to the POS concentration (p < 0.05), 293 indicating that the model significantly impacts the variability of POS extraction. Additionally, 294 the lack of fit test yielded non-significant results (p > 0.05), suggesting that the model 295 adequately captures the observed data. Among the parameters investigated, initial pH and 296 MAE temperature significantly influenced POS extraction at a confidence level of 95 %. 297 However, UAE time did not significantly affect POS extraction (p > 0.05). The response 298 surface graphs are shown in Fig. 3a and 3b. 299 The initial pH had a significant effect on the extraction of POS. Previous studies have 300 demonstrated that extracting pectin under acidic conditions raises HG pectin yields. For 301 instance, Liew et al., (2019) observed increased pectin yield from 3.68 to 36.33 g 302 pectin/100 g pomelo powder when the extraction was conducted at pH 1.8 using UMAE. 303 This increase is attributed to the breakdown of protopectin, an insoluble pectin precursor 304 found in plants, into soluble pectin molecules. Acidic solutions facilitate the hydrolysis of 305 protopectin, resulting in water-soluble pectin molecules. However, as mentioned in section 306 3.2, the use of acid solutions can degrade sugars and consequently reduce the yield of 307 14 POS extraction. Fig. 3a illustrates the relationship between pH, MAE temperature, and POS 308 concentration. Lower pH values enhance POS extraction at low temperatures. Conversely, 309 at higher temperatures, a pH of 1 resulted in the lowest POS concentration due to the 310 pronounced degradation of sugars caused by the severity of the extraction conditions. 311 The UAE time was assessed to determine its impact on the extraction process. The 312 statistical analysis concluded that the UAE time was insignificant for the model (p > 0.05) in 313 the range tested. Liew et al. (2016) conducted a study on the extraction of pectin using a 314 sequential UMAE approach. As in the current study, their findings indicated that the 315 duration of UAE (12 – 28 min) did not significantly impact the yield of pectin. Fig. 3b depicts 316 the combined effect of the MAE temperature and UAE time on the POS concentration. The 317 graph illustrates that the application of UAE for around 10 – 15 min results in a slight 318 increase in POS concentration when compared to 5 min of application. This suggests that 319 increasing the UAE time from 5 to 10 – 15 min has a modest positive impact on the 320 efficiency extraction of POS. The mechanism of UAE is based on high frequency sound 321 waves that disrupt plant materials through acoustic cavitation. This process generates 322 cavitation bubbles that implode, causing fragmentation, erosion, pore formation, shear 323 forces, and increased absorption. These mechanisms reduce particle size, boost surface 324 area, and enhance solubilization of bioactive compounds in the solvent. UAE also improves 325 water absorption, diffusivity, and swelling index in plant tissues. All these mechanisms 326 collectively led to increase the extraction yield (Kumar et al., 2021). 327 Finally, the MAE temperature plays a crucial role on POS extraction, with higher 328 temperatures resulting in enhanced extraction of POS as can be observed in Fig. 3. The 329 MAE temperature is evolved in two of MAE mechanisms, the penetration of solvent into the 330 plant matrix and the elution and dissolution of the bioactive compounds. Higher extraction 331 temperature leads to better solvent penetration and diffusivity of the pectin during elution 332 15 and dissolution into the solvent (Chan et al., 2017). Moreover, particle size is also related to 333 the two mechanisms mentioned above (Chan et al., 2017). The reducing of particle size 334 during UAE due to cavitation, increased the surface area which could enhance the pectin 335 extraction. However, it is important to note that excessively high temperatures can lead to 336 the degradation of pectin and oligosaccharides into smaller molecules, resulting in a loss of 337 their structural integrity and functional properties. Additionally, using a short MAE time (9.4 338 min) could mitigate pectin degradation by reducing exposure to high temperatures, as was 339 found in the study of Liew et al. (2016). 340 Based on the model proposed, the optimal conditions to maximize POS extraction were 341 initial pH 4, 10 min and 157 °C. Three confirmatory runs were carried out under these 342 conditions to validate the model. The composition of the extract obtained under optimal 343 extraction conditions is summarized in Table 2. The t-test concluded that there were no 344 significant differences between the experimental result for POS concentration (32.5 ± 0.5 345 g/L) and the predicted value of 31.6 g/L (p > 0.05). Thus, a good agreement was 346 established between the model and the experimental results. Under the given optimal 347 conditions, there was a solid recovery of 59.7 ± 2.5 % and the extract pH was 3.9 ± 0.1. 348 The recovery of POS in the extract from SBP was 66.0 ± 1.0 %. This yield exceeded the 349 results of previous studies conducted with either MAE or UAE. The POS extraction yield 350 from SBP obtained by hydrothermal MAE (165 °C, 12 min) reached 59.7 % (del Amo351 Mateos et al., 2022). Additionally, the extract pH was similar in both studies (extract pH ≈ 352 4), suggesting that the higher yield achieved in this study was attributed to the use of 353 sequential UMAE. Furthermore, the results of Fernández-Delgado et al. (2023) focused on 354 GalA extraction from SBP by UAE showed a recovery of around 20 % of total GalA (pH 4, 355 UAE amplitude 90 % and 90 min) compared to the 46.9 ± 2.3 % of OGalA recovered in the 356 current research. The results obtained reveal that the combination of UMAE may be a 357 16 suitable technology to enhance POS recovery from SBP. Moreover, energy consumption is 358 a key factor in the development of new technologies. It seems that the use of UAE for a 359 short duration before MAE can effectively disrupt the cell walls, leading to higher yields at 360 lower temperatures and shorter MAE times. Consequently, this approach may have the 361 potential to decrease energy consumption during the extraction process. 362 3.4 POS purification and concentration 363 During the extraction of RG-I pectin from SBP, other non-targeted compounds, such as 364 monomers, organic acids, or furfural and HMF from pentose and hexose dehydration, can 365 be found in the extracts (Moure et al., 2006). Thus, these small molecules should be 366 removed to increase the purity of the extract. Among the techniques available for pectin 367 purification and concentration, the use of membranes (DF and UF) avoids the use of 368 harmful organic substances, requires little space and is energy efficient (Gómez et al., 369 2013). However, the MWCO is a key parameter to consider when selecting a membrane. 370 The process yield and filtration flux are affected by the MWCO. Higher MWCO leads to 371 higher filtration flux, which means shorter operation times, but a lower process yield, since 372 more targeted molecules can pass through the membrane. 373 This study compared the recovery yield of POS and the filtration flux using two 374 membranes with different MWCO (3 and 5 kDa). Table 2 summarizes the composition of 375 the extract obtained under UMAE optimal conditions, the composition of the retentates, and 376 the yield of the membrane processes calculated using Eq. (1). 377 In the continuous DF process, the objective was to remove small molecules and refine 378 the extract; while the UF process aimed to clean the small molecules remaining after DF 379 and to concentrate the extract. As shown in Table 2, after the membrane process, the 380 concentration of POS increased in both retentates (RUF – 3: 61.2 g/L and RUF – 5: 53.5 g/L), 381 indicating a successful concentration. Simultaneously, the monomers and degradation 382 17 compounds passed through the membranes and were removed from the retentate. The 383 POS recovery yield was significantly higher in the RUF – 3 retentate than in the RUF – 5. This 384 difference may be attributed to the presence of some small OGalA and AraOS molecules, 385 the major components of POS, which could pass through the membrane of 5 kDa. In both 386 cases, monomers and degradation compounds were effectively removed, as their highest 387 concentration was only 0.1 g/L from the 2.0 g monomers/L and 1.4 g/L of degradation 388 compounds present in the extract. 389 The recovery yields (Table 2) obtained with the membrane process were higher when 390 compared to previous studies. Gómez et al. (2013) refined a pectin extract from lemon peel 391 using a cellulose membrane with an MWCO of 1 kDa through discontinuous DF and UF, 392 reporting recovery yields of 96.1, 59.3 and 79.8 % for OGalA, AraOS and GalOS, 393 respectively; whereas, up to the 26.1 % of the monomers remained in the retentate after the 394 membrane process. In contrast, the recovery yields of the oligomers obtained in the current 395 study and the removal of monomers, using an MWCO membrane of 3 kDa, were higher 396 (Gómez et al., 2013). This suggests that using a higher MWCO membrane, such as the 3 397 kDa membrane used in this study, does not result in a loss of targeted compounds and can 398 increase the elimination of small molecules when employing a continuous DF process. The 399 reason for the improved results in the current study could be the viscosity of the feed. In the 400 case of the discontinuous DF process, the viscosity increased due to concentration. A 401 higher viscosity solution could reduce the efficiency of the process (Field and Wu, 2022). 402 Furthermore, the highest recovery yield for OGalA reported by Jin et al. (2022) was 403 approximately 75 % during UF using a PES membrane with an MWCO of 3 kDa. This yield 404 is significantly lower than the value obtained in this research using the same MWCO 405 membrane (3 kDa). The inclusion of the previous DF step in the current study may have 406 contributed to the higher recovery yield of OGalA. 407 18 Additionally, the change in the permeate flux over time is depicted in Fig. 4a. The trend 408 observed for both membranes is similar, with the permeate flux reaching a steady state 409 after approximately 1.5 hours of operation. However, there was a notable difference in the 410 permeate flux. The permeate flux using the MWCO membrane of 5 kDa was 6.7·10-6 411 m3/(m²·s); whereas, for the 3 kDa MWCO membrane, it only reached 1.7·10-6 m3/(m²·s). 412 This discrepancy is an important factor to consider, since the operation time required to 413 refine pectin using the smaller MWCO membrane would be much longer. Similar results 414 were reported in the study of Jin et al. (2022). 415 Fouling is another important parameter that must be considered. The fitting models for 416 membrane fouling have been used (see supplementary material). An R2 higher than 0.91 417 was found for the three models for both membranes, confirming the suitability of the fouling 418 models proposed and suggesting that the membrane, pore blocking, and cake formation 419 resistances were limited in the first 1.5 h of operation. During the initial 1.5 hours of UF (Fig. 420 4b), there is a significant contribution of total resistance. The total resistance increased to 421 60.6 % during the UF process using the 3 kDa MWCO membrane, but only 23.7 % using 422 the 5 kDa MWCO membrane. Thus, the fouling resistance was much more pronounced in 423 the case of the 3 kDa MWCO membrane. This trend is consistent with the decline in the 424 permeate flux (Fig. 4a), which was significantly higher, as mentioned above. Based on the 425 calculated constants of the filtration resistance models (equations 6 – 8) (see 426 supplementary material), there is a statistical difference between the constant parameters 427 for the two membranes used. Nevertheless, the difference is notably higher in the case of 428 KC followed by Km. Thus, the much lower permeate flow observed during operation with the 429 3 kDa MWCO membrane could be due to the resistance of the membrane and the cake 430 formation, and somewhat less to the pore blocking. 431 19 The findings suggest that the membrane process employed in this study yields better 432 results than previous research, particularly in terms of the recovery yields of the targeted 433 compounds. The inclusion of a continuous DF step in the process likely contributed to 434 achieving higher recovery yields, while efficiently removing the undesired compounds. 435 Additionally, the selection of the appropriate MWCO membrane is crucial, as recovery 436 yields may be compromised due to the lower permeate flux when using a smaller MWCO 437 membrane. 438 3.5 Refined POS characterization 439 After the membrane process, the refined extracts (RUF – 3 and RUF – 5) were freeze440 dried. Although alcoholic precipitation is a widely used method for HG pectin recovery from 441 extracts, it is not efficient in recovering RG-I pectin, particularly arabinose (del Amo-Mateos 442 et al., 2022). In contrast, freeze-drying is a commercial technology used for the dehydration 443 of food-grade products (García-Velásquez and van der Meer, 2023) and allows the 444 recovery of all the components present in the extract. 445 The global yield (g pectin/100 g SBP), composition and degree of esterification can be 446 found in Table 3. The global yields were 39.7 and 39.0 % for Pectin – 3 and Pectin – 5, 447 respectively, indicating that the MWCO of the membrane used did not influence the global 448 yield. Both yields were higher than those reported for conventional pectin extraction and/or 449 ethanol precipitation. For instance, Adiletta et al. (2020) reported a pectin yield from SBP of 450 25 % using hot acidic extraction (pH 1.5, 90 °C, 4 h, solid-to-liquid ratio 1:30 g/mL) followed 451 by ethanol precipitation and the yield obtained by Y. Mao et al. (2019) after MAE (90 °C, 452 120 min) and alcohol precipitation was 23.4 %. 453 The highest performance in the membrane process was achieved with the 3 kDa 454 MWCO membrane, resulting in a 90 % recovery of POS. This has resulted in a slightly 455 higher purity of the final pectin solid from RUF – 3 (80.2 % POS) as compared to 72.9 % 456 20 POS in Pectin–5. It is worth noting that the purity of both pectins were in the range of 457 commercial prebiotics. The POS content in pectins was much higher than those reported in 458 previous research (60.9 %), where the extract was not purified before being subjected to 459 freeze-drying (del Amo-Mateos et al., 2023). This result allowed to conclude the suitability 460 of a membrane purification process to increase the content of POS. As in the extract, 461 AraOS was the major component in both pectins, while OGalA accounted for the 22.4 and 462 17.0 % in Pectin – 3 and Pectin – 5, respectively. In both pectins, a small amount of protein 463 was detected, comprising 3.7 % (Pectin – 3) and 4.7 % (Pectin – 5) of the composition. 464 Based on the degree of esterification, pectin can be classified as high methoxyl (HM) and 465 low methoxyl (LM). The degree of esterification was similar in both pectins and can be 466 considered as high-methoxyl pectin (degree of esterification > 50 %), indicating that pectins 467 obtained in this study can form gels at low pH and in the presence of sugars. High-methoxyl 468 pectins are applied as stabilizers, as a rheology modifier, and in sugary products (Abboud 469 et al., 2020). 470 The molecular weight distribution curves of the two pectins (see supplementary material) 471 showed a similar pattern, indicating a comparable distribution. Both pectins exhibited a 472 major peak at a molecular weight of 377.4 kDa. Additionally, a secondary peak was 473 observed, with a relatively lower intensity, at a molecular weight of 3.0 kDa in Pectin – 3 474 and at 4.3 kDa in Pectin – 5, suggesting the presence of smaller molecules in the pectins. 475 The surface structure of pectins was analyzed by FTIR for their functional groups in the 476 range of 4000 – 400 cm-1 (see supplementary material). The patterns obtained showed the 477 typical pectin chemical composition (Concha Olmos and Zúñiga Hansen, 2012). The peaks 478 found at 1740 and 1650 – 1680 cm-1 were attributed to the C=O stretching vibration of 479 esterified and ionic carboxyl groups, respectively (Jiang et al., 2012). The higher 480 21 absorbance intensity of the ester carboxyl group corroborated that both pectins belong to 481 the high-methoxyl category. 482 3.6 Future perspectives of the integrated process for RG-I pectin recovery: technical and 483 economic considerations 484 A preliminary study of the energy consumption of the sequential UMAE was carried out. 485 The ultrasound and microwave equipment continuously monitored the power consumption. 486 Thus, the total amount of energy required for each extraction step was calculated according 487 to equation 11: 488 𝑄 = ∫𝑃(𝑡) · 𝑑𝑡 (11) where Q is the energy required (kW·h), P is the power dissipated (kW) and t is the 489 extraction time (h). 490 The total energy consumption to produce 1 g of pectin was 0.02 kW·h for both pectin – 3 491 and pectin – 5 (see supplementary material). This value was significantly lower compared 492 to the one obtained by Liew et al. (2019) for conventional extraction (1.05 kW·h). This lower 493 value may be explained by the much shorter operation time required during sequential 494 UMAE (10 min UAE and 9.4 min MAE) compared to the one used for conventional 495 extraction (141.4 min). 496 Previous research has shown the potential prebiotic properties of POS obtained from 497 SBP (Prandi et al., 2018). Traditional pectin extraction methods typically target GalA 498 recovery from the HG region, which has market prices around $21/kg (Moslemi, 2021). In 499 contrast, the market price range for prebiotics has been reported to be significantly higher, 500 ranging from 120 to 850 €/kg (www.consumerlab.com). Furthermore, the market of 501 prebiotics and pectin were estimated at $6.0 billion and $1.5 billion, respectively with an 502 annual growth rate for 2030 of 11 % for prebiotics and 5 % for pectin 503 (www.rearchandmarkets.com). Although a deep economic analysis should be done to 504 28 Figure captions 651 Figure 1. Scheme of the sequential UMAE extraction and membrane purification 652 Figure 2. RSM-BBD results: composition of the extracts and POS. POS: 653 pectooligosaccharides; OGalA: Oligogalacturonoides; GalOS: galactooligosaccharides; 654 RhaOS: rhamnooligosaccharides; AraOS: arabinooligosaccharides 655 Figure 3. Response surface of RSM-BBD: effect of the independent variables on the 656 concentration of POS in the extracts 657 Figure 4. Time courses of permeate flux (J) (a) and total resistance (RT) (b) during the 658 ultrafiltration process 659 29 Tables 660 Table 1. RSM-BBD: experimental conditions, pH of extracts and solid recovery 661 Run Independent variables Extract pH Solid recovery (%) x1 X1 (Initial pH) x2 X2 (UAE-t 1, min) x3 X3 (MAE-T 2, °C) 1 0 2.5 1 25 -1 120 4.0 ± 0.1 74.0 ± 1.1 2 -1 1 0 15 -1 120 1.5 ± 0.3 55.4 ± 0.9 3 1 4 0 15 1 160 3.8 ± 0.1 46.5 ± 1.5 4 0 2.5 0 15 0 140 3.8 ± 0.2 64.9 ± 1.7 5 1 4 1 25 0 140 4.0 ± 0.2 58.7 ± 1.2 6 0 2.5 0 15 0 140 3.8 ± 0.3 62.2 ± 0.9 7 -1 1 0 15 1 160 1.6 ± 0.2 42.1 ± 0.8 8 0 2.5 0 15 0 140 3.9 ± 0.1 66.8 ± 0.9 9 1 4 0 15 -1 120 4.2 ± 0.2 73.4 ± 1.1 10 -1 1 -1 5 0 140 1.5 ± 0.2 42.3 ± 1.4 11 0 2.5 1 25 1 160 3.8 ± 0.3 44.9 ± 0.8 12 1 4 -1 5 0 140 4.0 ± 0.1 62.0 ± 1.2 13 -1 1 1 25 0 140 1.5 ± 0.2 61.7 ± 1.3 14 0 2.5 -1 5 -1 120 4.0 ± 0.3 74.1 ± 0.9 15 0 2.5 -1 5 1 160 4.0 ± 0.2 53.1 ± 1.5 1UAE-t: Ultrasound-assisted extraction time; 2MAE-T: Microwave-assisted extraction 662 temperature 663 30 Table 2. Composition of the extract obtained under optimal UAE and MAE conditions and 664 the composition of the final retentates obtained after the purification membrane processes 665 (RUF – 3, RUF – 5). Recovery yields after the DF/UF membrane processes 666 Concentration (g/L) Recovery yield (%) Extract RUF – 3 RUF – 5 RUF – 3 RUF – 5 Monomers 2.0 ± 0.0 0.0 ± 0.0 0.1 ± 0.0 0.9 2.8 Degradation compounds1 1.4 ± 0.1 0.1 ± 0.0 0.1 ± 0.0 2.4 3.7 OGalA2 8.8 ± 0.4 18.0 ± 0.7 12.4 ± 0.1 97.5 64.0 GalOS3 3.9 ± 0.1 7.9 ± 0.4 8.7 ± 0.1 95.7 ∼100 RhaOS4 1.6 ± 0.1 2.9 ± 0.1 3.9 ± 0.0 82.3 ∼100 AraOS5 18.3 ± 0.3 32.5 ± 1.5 28.6 ± 1.4 84.9 71.2 GlcOS6 0.9 ± 0.0 0.0 ± 0.0 1.4 ± 0.0 0.0 84.4 POS7 32.5 ± 0.4 61.2 ± 2.7 53.5 ± 1.6 90.0 74.6 1Degradation compounds: formic acid, acetic acid, HMF and furfural. 2OGalA: 667 Oligogalacturonoides; 3GalOS: galactooligosaccharides; 4RhaOS: rhamnooligosaccharides; 668 5AraOS: arabinooligosaccharides; 6GlcOS: glucooligosaccharides; 7POS: 669 pectooligosaccharides 670 31 Table 3. Pectin yield, composition, and structural characteristics of final RG-I pectin-rich 671 solids: Pectin – 3 and Pectin – 5 672 Pectin – 3 Pectin – 5 Yield (%, g pectin/g SBP) 39.7 ± 0.5 39.0 ± 0.3 Degree of esterification (%) 73.4 ± 1.0 69.6 ± 0.9 Composition (%) OGalA1 22.4 ± 0.1 17.0 ± 0.3 GalOS2 11.1 ± 0.0 12.0 ± 0.5 RhaOS3 3.7 ± 0.1 4.4 ± 0.2 AraOS4 43.1 ± 0.0 39.6 ± 1.3 GlcOS5 0.1 ± 0.0 1.2 ± 0.0 POS6 80.2 ± 0.0 72.9 ± 1.9 Protein 3.7 ± 0.2 4.7 ± 0.2 1OGalA: Oligogalacturonoides; 2GalOS: galactooligosaccharides; 3RhaOS: 673 rhamnooligosaccharides; 4AraOS: arabinooligosaccharides; 5GlcOS: 674 glucooligosaccharides; 6POS: pectooligosaccharides 675 32 Figures 676 677 33 678 0 2 4 6 8 10 12 14 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 Galacturonic acid (g/L) Run Monomers Oligomers (a) 0 1 2 3 4 5 6 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 Galactose (g/L) Run Monomers Oligomers 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 12345678910 11 12 13 14 15 Rhamnose (g/L) Run Monomers Oligomers (c) 0 5 10 15 20 25 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 Arabinose (g/L) Run Monomers Oligomers 0.0 0.5 1.0 1.5 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 Glucose (g/L) Run Monomers Oligomers (e) 0 5 10 15 20 25 30 35 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 POS (g/L) Run OGalA GalOS RhaOS AraOS (b) (d) (f) 34 679 680 681 (a) (b) 35 682 1E-06 2E-06 3E-06 4E-06 5E-06 6E-06 7E-06 8E-06 9E-06 1E-05 03000 6000 9000 12000 15000 18000 J (m3/(m2·s)) Time (s) 3 kDa 5 KDa 0E+00 1E+13 2E+13 3E+13 4E+13 5E+13 6E+13 7E+13 8E+13 9E+13 03000 6000 9000 12000 15000 18000 RT (m-1) Time (s) 3 kDa 5 kDa (a) (b) Credit Author Statements Esther del Amo Mateos  Investigation, methodology, supervision, writing-original draft Berta Cáceres  Investigation, methodology, visualization Mónica Coca  Conceptualization, formal analysis, supervision María Teresa García-Cubero  Conceptualization, formal analysis, supervision Susana Lucas  Conceptualization, writing -review & editing, project administration Credit Author Statement Electronic Annex Click here to access/download Electronic Annex Supplemetary_material_R3.docx