Polyphenol extraction from industrial water by-products: a case study of the ULTIMATE project in the fruit processing industry

Water is an essential resource, crucial for sustaining life, supporting ecosystems, and enabling various industrial processes. Its scarcity and contamination pose significant threats to global health, food security, and economic stability, necessitating efficient management and conservation strategies.

Through this perspective, water stemming from industrial use should not be viewed as waste but as a valuable by-product that can be treated and recycled by implementing advanced treatment technologies and closed-loop systems, enhancing water sustainability and promoting industrial efficiency. A smart approach in water management and use would ideally involve a circular water system that minimizes water losses, captures and exploits the value in water, and fosters water security, sustainability, and resilience, that way meeting current ecological, social, and economic needs (Water Europe 2023; Ramin et al. 2024).

The EU-funded project ULTIMATE [https://ultimatewater.eu/the-project/] focuses on the above-mentioned approach enhancing sustainability by converting wastewater into a valuable resource through Water Smart Industrial Symbiosis. It aims to optimize the recycling and reuse of water, energy, and materials within industrial sectors, thereby creating economic value and reducing environmental impact. The project is structured in nine large-scale case studies in nine countries.

Food industry water streams vary widely depending on the type of production process involved (Franzen Ramos et al. 2023; He et al. 2023). In the juice production industry, water streams typically arise washing, processing, equipment cleaning, and by-product handling. Each type of these streams has a different composition and may be rich in soil, organic matter, pesticides as well as fibers, sugars, organic acids, nutrients, polyphenols, and other value-added compounds. In particular, citrus fruit processing leads to water streams that have a high content of sugars, organic acids, proteins and amino acids, essential oils, and phenolic compounds such as flavonoids and anthocyanins (Viuda-Martos et al. 2010; Satari & Karimi 2018; Suri et al. 2021; Franzen Ramos et al. 2023).

Polyphenols, due to their high and diverse biological activity, are utilized across a wide range of scientific and industrial fields, from healthcare and food production to environmental management and material science (Georgé et al. 2005; Viuda-Martos et al. 2010; Ajila et al. 2011; Abad-García et al. 2012; Satari & Karimi 2018; Franzen Ramos et al. 2023). To name a few of these uses, emphasis can be placed on the food industry, where phenolic compounds are extensively used due to their health benefits (functional foods) and ability to improve food quality (preservatives, flavor enhancers). Moreover, the antioxidant and anti-inflammatory properties of polyphenols make them valuable in cosmetic formulations.

The extraction of polyphenols from water streams is therefore of major environmental and scientific importance (Satari & Karimi 2018; Suri et al. 2021). Large-scale extraction involves a combination of physical, chemical, and biological processes, like industrial-scale solid phase extraction (SPE) units (using polymeric resins or activated carbon) or countercurrent liquid–liquid extraction (LLE) where water and organic solvent flow in opposite directions to enhance the extraction efficiency. The recovery of polyphenols from the extract is a determinative step of the procedure yielding final products of various concentrations and purity (Ajila et al. 2011; Franzen Ramos et al. 2023).

Other advanced and quite attractive extraction techniques involve the use of subcritical water extraction (SWE) and supercritical fluid extraction (SFE). These methods offer several advantages, including high efficiency, selectivity, and environmental friendliness (Espinosa-Pardo et al. 2017). SWE, also known as hot water extraction, involves using water at temperatures between 100 and 374 °C and pressures sufficient to maintain it in the liquid state, in which conditions, water exhibits unique properties that enhance the extraction of polyphenols. SFE employs fluids in their supercritical state. Supercritical carbon dioxide (sCO₂) is most commonly used, due to its properties and environmental safety, since the process can be carried out at relatively low temperatures, preserving the integrity of heat-sensitive polyphenols.

Polyphenol extracts have been analyzed and constituents have been identified using LC-HRMS and LC-MS/MS (Ajila et al. 2011; Franzen Ramos et al. 2023; Wang et al. 2024), including flavonols/flavanones (Avula et al. 2016; Gómez-Mejía et al. 2019; Svilar et al. 2019; Luo et al. 2023; Wang et al. 2024), terpenes (Avula et al. 2016; Wang et al. 2024), coumarins (Abad-García et al. 2012; He et al. 2021; Luo et al. 2023), and anthocyanins (He et al. 2021; Rapisarda et al. 2022; Custodio-Mendoza et al. 2024). Liquid chromatography – high resolution mass spectrometry (LC-HRMS) is the technique of choice, since it provides adequate separation of analytes, followed by their unequivocal mass-based detection, incorporating extremely high efficiency of mass separation, thus increasing the certainty of identification (Schymanski et al. 2014).

In the present study, a mixed water stream originating from the various processing stages of sweet orange (Citrus sinensis L.), during juice-making procedure, was passed through a packed bed of a sorptive material on which polyphenols are selectively adsorbed, followed by three different methods for the extraction of the compounds of concern; (a) conventional extraction technique using mixtures of water and organic solvents (b) SWE and (c) SFE. The main objectives of the study were to (i) investigate the parameters that affect the extraction process, (ii) assess the most effective and environmentally appropriate extraction method, (iii) identify extracted polyphenols, and finally (iv) showcase the financial potential of the extraction for the industry.

For the adsorption of polyphenols, Amberlite FPX66 resin (DuPont, USA) was used, which is a food-grade adsorptive material, effective in adsorbing the substances under study. For the polyphenol extractions, Milli-Q water was produced in the laboratory by a Milli-Q purifying system (Millipore, Billerica, MA, USA). Also ethanol (EtOH) (95%, Merck, Germany), methanol (MeOH) (99%, Merck, Germany), EtOH–water, and methanol–water mixtures were used as extraction solvents. Gallic acid (>97.5%), Folin–Ciocalteu phenol reagent (2N), and sodium carbonate (>99%) (Merck, Germany) were used for total polyphenol analysis. Finally, food-grade CO₂ (>99.9%) was purchased by Revival, Greece.

For the separation and identification of polyphenols using UHPLC-Q-Orbitrap-MS/MS, methanol (MeOH) of HPLC grade (99.99%) was obtained from Fischer Scientific (Leicestershire, UK), acetonitrile (ACN) of LC-MS grade was obtained from Merck (St. Louis, MO, USA). High-purity water (18.2 MOhm) Milli-Q water was produced in the laboraory by a Milli-Q purifying system (Millipore, Billerica, MA, USA). Formic acid (FA) (LC-MS grade) was purchased by Riedel-d Haen (Seelze, Germany). Quercetin (>95%) was purchased as reference materials from Merck (Germany).

The concentration of total polyphenols was determined using a UV-vis spectrophotometer Shimadzu UV-1800 (Duisburg, Germany). Two hundred microlitres of the liquid samples were mixed with 200 μL of Folin–Ciocalteu reagent (2N). Then, 3 mL of ultrapure water and 0.6 mL of sodium carbonate solution of 20% were added. The absorbance of the obtained mixture was measured at 760 nm after incubation for 15 min at 50 °C. Because gallic acid was used as the standard, the content of polyphenols was expressed as gallic acid equivalent (GAE). The following regression model expresses the relation between the concentrations of standard gallic acid solutions and the corresponding absorbances observed: ⁠, where A765 is the absorbance at 765 nm and CGA is the concentration of gallic acid (mg L⁻¹).

For the selective adsorption of value-added compounds, the orange juice by-product was filtered using a bag filter to remove remaining solids and pumped through vessel that was packed with 20 kg of resin. The flow rate is 7 L/min thus the contact time for this process is 27 min. Once the resin reaches saturation, the vessels are taken out from the experimental setup (VesperX) and transferred to our laboratory. Here, we proceed with the extraction of polyphenols from the resin.

Saturated resin was extracted with different solvents, including MeOH, EtOH and water using a magnetic stirrer. Five grams of resin was extracted with 25 mL of water, methanol or ethanol at 60 °C under ultrasound for 30 min. The mixtures of methanol–water and ethanol–water of various concentration (10, 20, 50, 100% of the organic solvent) were also tested in order to compare the results. The extracts were used to measure total phenolic content (TPC) as previously described. The extraction yield was calculated using the formula ⁠, taking into account the total volume of the extract and by-product and express it in mass of polyphenols after the treatment and before.

SWE was performed in a 1-L Supercritical CO₂/Subcritical Water Extractor (ExtrateX, France). Six hundred seventy grams saturated with orange juice by-products resin was placed in the extraction vessel. Water or water–20% methanol mixture was pumped by means of an HPLC pump set at 10 mL min⁻¹. The extractor was set at 120 °C and 12 bar pressure to keep the water in liquid state. Extraction was carried out for 1 h for water (sample subW) and one more hour for water methanol mixture (sample subWM) collecting 120 mL of extract in total.

Supercritical CO₂ extraction was performed in the same 1 L Supercritical CO₂/Subcritical Water Extractor. The 670 g of resin previously subjected to SWE were then placed again into the extraction vessel. The flow of CO₂ was set 50 g min⁻¹, with CO₂ recirculation, while the pressure and temperature were set at 150 bar and 70 °C, respectively. The extraction was performed for 24 h (sample super24) collecting two fractions (55 mL of SuperA from separator A and 47 mL of Super B from separator B), the two separators operate with ΔP = 100 bar. Samples were also obtained 2 h (sample Super2) within the process. Extraction under higher temperature or higher-pressure conditions did not perform, as high temperature destroys polyphenols, and high pressures damage the resin, making it impossible to regenerate and reuse.

The workflow for compound identification included the following steps: (1) A number of polyphenols were specifically selected and targeted, due to their abundance and frequent occurrence in citrus by-product water according to literature, (2) their molecular formula, exact theoretical m/z identification ions, exact experimental detected m/z identification ions, Δppm (mass accuracy) and MS/MS fragment m/z were calculated (Table 1) in order to be separated chromatographically, and extract the respective ion chromatograms using Full MS Spectrum. (3) For the semiquantification of targeted compounds, two compounds were spiked to the initial samples at a predefined concentration. Quercetin was used for compounds identified with positive MS and gallic acid for compounds identified using negative ESI.

Semiquantification was carried out using the following approach (Malm et al. 2021; Gutiérrez-Martín et al. 2023):

The main assumption in both cases was that targeted compounds have either similar chemical structure or are closely eluted to the spiked compound, thus the response of the structurally similar compound will be similar to that of the targeted, and therefore, the RF of the similar standard (e.g. quercetin) can be used to estimate the concentration of the unknown.

A Dionex 3000 UHPLC system (Thermo Scientific, Bremen, Germany) was used for the chromatographic separation. The system consisted of a vacuum degasser, a high-pressure binary pump, an autosampler with a temperature-controlled sample tray set at 7 °C and a column oven set at 30 °C. Chromatographic separation was performed at 30 °C using a Zorbax Eclipse Plus C18 column (100 × 2.1 mm i.d., 1.8 μm particle size; Agilent Technologies). The mobile phase consisted of 5 mM ammonium formate in 0.02% FA (solvent A) and a mixture of acetonitrile:water (90:10 v/v) containing 5 mM ammonium formate and 0.02% aqueous FA (solvent B). A gradient elution program was employed at a constant flow rate of 0.2 mL/min with solvent B starting at 5% for 3 min, initially increasing to 30% in 4 min, then increasing to 90% in 11 min and finally, set back to 5% in 11.5 min. Post-run equilibrium time was 3.5 min. The injection volume was 20 μL. Diverter valve was programed to send LC eluent in waste for the first 4 min.

A Q Exactive benchtop Orbitrap-based mass spectrometer (Thermo Scientific) operated in the positive polarity mode equipped with a heated ESI (HESI) source. Source parameters were: sheath gas (nitrogen) flow rate, auxiliary gas (nitrogen) flow rate and sweep gas flow rate: 40, 10 and 1 AU, respectively, capillary temperature: 250 °C, ESI heater temperature: 20 °C, spray voltage: +4.0 kV (positive polarity). The instrument operated in full scan mode from m/z 100–1,000 at 17,500 resolving power and injection time of 100 ms and in MS/MS mode from m/z 100–1,000 at 17,500 resolving power and injection time of 62 ms (product ion mode). The automatic gain control (AGC) was set at 10E⁶ ions. Mass calibration of the Orbitrap instrument was evaluated in both positive and negative modes weekly and external calibration was performed prior to use following the manufacturer's calibration protocol.

Compound identification was achieved by: (a) t_R of compounds, (b) Δppm between theoretical and experimental accurate mass of precursor m/z < 10 ppm and (c) the presence of two characteristic precursor/product ion transitions from bibliographic references.

The comparison of different eluent compositions for the extraction of polyphenol from the resin reveals notable differences in efficiency (Figure 2). One hundred percent water exhibited the lowest extraction efficiency, 5.2%. At 100% concentration, ethanol (EtOH) did not perform significantly better, with 8% extraction yield, while methanol (MeOH) was more efficient with 26% yield. At concentrations of 50% aqueous solutions with MeOH or EtOH resulted in the highest extraction yield. The use of EtOH exhibited significant improvement, achieving 27% extraction efficiency, whereas MeOH exhibited the highest efficiency observed at 70%. In all other compositions the efficiency ranged from 5 to 20%. At 5% concentration, EtOH achieved a low extraction percentage of 7%, while MeOH performed slightly better at 18%. At 10% EtOH exhibited similar efficiency to 5% EtOH, at 7%, while 10% MeOH achieved 8% extraction. Increasing the concentration to 20%, EtOH exhibited some improvement, reaching 8%, while MeOH remained at 9%. In general, MeOH outperformed EtOH in extracting polyphenols across all concentrations. The most efficient eluent composition was 50% MeOH, achieving an extraction percentage close to 70%. EtOH showed moderate efficiency at 50% concentration but did not perform as well as methanol at any concentration. 100% water was the least effective eluent for polyphenol extraction.

SWE enhanced the yield fivefold achieving 25% recovery of polyphenols while the use of solvent, water in this case, was five time less than in conventional extraction with pure water. Addition of 20% of MeOH as a co-solvent increased the yield to 33%. Moreover, the addition of MeOH enabled the extraction of polyphenols that were not extracted with water, such as 3-desmethyloxytangeretin, enhancing at the same time the extraction of other, less polar, polyphenols like hesperetine (Supplementary material).

Supercritical CO₂ extraction doubled the extraction yield, reaching 75%, without using any additional co-solvent. After the extraction, when pressure is reduced, CO₂ escapes as gas resulting in a solvent free extract.

From the wide range of targeted compounds the most abundant signals and the widest range of compounds were obtained using SFE, with abundances increasing given longer extraction times.

The extraction technique and solvent clearly influence the abundance and profile of compounds extracted from the sample.

In the case of flavanols, the most abundant compounds identified in the sample SuperB, were rutin, kaempferol and iso-kaempferide, while SubW mainly showed kaempferol and SubWM kaempferol and diosmin (Supplementary material).

The main coumarins and phenolic acids obtained using SFE were p-coumaric acid, sinapinic acid, umbelliferon, 5,7dihydroxycoumarin, 5,6,7trihydroxycoumarin and 4-hydroxycoumarin. On the other hand, far less abundant were the compounds obtained using SubW, providing only p-coumaric acid, sinapinic acid, umbelliferon and 5,7dihydroxycoumarin. SubWM included mainly p-coumaric acid, sinapinic acid, umbelliferon and 5,7dihydroxycoumarin at higher abundances than SubW (Supplementary material).

In the case of terpenes, isolimonic acid, nomilinic acid and limonin were mainly identified in Super2 sample, while SubW included the same compounds at far less abundance, while SubWM also included isolimonic acid, nomilinic acid and limonin, as well as resveratrol which was identified only in this sample.

Anthocyanins peonidin, malvidin and petunidin were the main identified members of this class in Super2, while SubW and SubWM contained mainly peonidin, at far less abundance, as shown in Supplementary material.

The use of SFE provided extracts (SuperA) with significantly higher concentration of the targeted compounds, as shown by previous studies (Brijesh kumar et al. 2013). SubW produced extracts with fewer compounds, mostly water soluble. The SubW extract did not include hesperidin, narirutin, naringin, diosmin, trimethocyxoumarin, nomilin, obacunone or myricetin, while the overall quantity of the extracted compounds accounted for <10% as compared to SuperA. The addition of MeOH in the eluent (SubWM), resulted in the presence of semi-water soluble compounds, and higher compound abundancies (Figure 5(a)).

For example, the contribution of SubW in the extraction of p-coumaric acid, sinapinic acid and ferulic acid is quite significant, while tangeretin, nobiletin, diosmetin, hexamethylquercetangetin are mainly extracted through SuperA and SuperB (Figure 5(b))

The results of semiquantification are presented in Supplementary material, Table S1.

Kaempferol and diosmetin were the main compounds found in the total extract, exceeding 20 and 7 mg in the resin, as shown in Figure 6(b). In the case of coumarins, ferulic acid is the most abundant representative of the class (>100 mg in the resin), with p-coumaric acid and the hydroxy-derivatives of coumarin at higher abundance (Figure 6(c)).

Nomillinic acid, limonin and isolimonic acid were the dominant compounds in the terpene group (Andrade et al. 2023), exceeding 10 mg in the processed resin (Figure 6(d)).

Anthocyanins were detected at significantly lower concentrations, in agreement with Brito et al. (2014), as expected by previous studies, with peonidin found at the highest concentration (Figure 6(e)).

In total, the most abundant classes of polyphenols that were extracted from the stream, are flavanones/flavones and coumarins (Figure 6(f)). Based on the semiquantitative technique the concentration of the selected targeted polyphenols extracted exceeds 1.3 g in the resin, with the calculated total polyphenol content was calculated >20 g.

The exploitation of polyphenols presents a great business opportunity as a result of the underlying huge business potential. This potential arises, as previously discussed, from the fact that polyphenols exhibit very high biological activity, that is the capacity of a specific molecular entity to achieve a defined biological effect (Jackson et al. 2007). They have been traditionally used for a variety of uses ranging from the non-specific, e.g. as natural colorants or preservative in the food and beverage industry to the specific such as cancer (Cháirez-Ramírez et al. 2021) or COVID-19 treatment (Paraiso et al. 2020).

Readily available in fruit and vegetables, they have been extensively used as extracts for non-specific (F&B, food supplements, cosmetics, etc.) applications due to the facts that (i) they need to be isolated from complex matrices, and (ii) their complex structure makes it hard to be produced synthetically, which has kept their price extremely high (as shown in Table 2) for specific uses.

Within this study, we have achieved a relatively low cost, especially viewed through the perspective of the market price of pure polyphenols, removal of the complex matrices of polyphenols sources which have practically zero cost of acquisition. We have also shown that chemical isolation and purification can be performed chromatographically, hence achieving high isolated compounds of purity. We expect that this will make polyphenols more available at this purity level, at a lower cost than today. We strongly believe that this will be a disrupting factor for the polyphenol market both in range of use and market size not only leading the wider exploitation of their properties, but also creating new markets altogether.

The current study deals with sustainable water resource management in the fruit processing industry, focusing on a case study located in a juice factory in Nafplio, Greece. Results of the study reveal significant implications for resource recovery and environmental sustainability. By evaluating the recovery of polyphenols from water by-product using a mobile wastewater treatment unit, the research highlights the potential for converting waste into valuable resources. Among the extraction methods tested – conventional solvent extraction, SWE, and SFE – SFE emerged as the most efficient, yielding the highest polyphenol recovery at 75%.

The diverse polyphenol profile obtained through SFE, including flavonoids, terpenes, coumarins, and anthocyanins, underscores the economic viability of this approach for applications in the food, cosmetics, and pharmaceutical industries. The semiquantification of the polyphenols recovered indicates that a significant amount of valuable compounds, >20 g of total polyphenols and >1.3 g of targeted identified polyphenols, can be extracted from the water by-product. This study demonstrates that advanced extraction technologies can significantly enhance resource recovery, promoting circular economy practices and reducing environmental impact in the fruit processing sector.

In conclusion, this research provides a convincing case for adopting sustainable water management practices in the industry. The successful recovery of high-value polyphenols from wastewater aligns with environmental sustainability goals and offers substantial economic benefits. Future research should focus on scaling up these extraction processes to industrial levels, integrating them with existing wastewater treatment systems, and assessing long-term environmental and economic impacts. Additionally, optimizing extraction conditions and exploring other valuable compounds in by-product streams will be essential for maximizing resource recovery and sustainability in the industry.

The results were obtained and evaluated in the frame of the project ‘Ultimate’ funded by the European Union's Horizon 2020 research and innovation program under grant agreement number 869318. The authors would like to thank Dr Dimitris Kletsas for the hospitality at the research infrastructure of the Institute of Biosciences and Applications of the National Center for Scientific Research ‘Demokritos’.

All relevant data are included in the paper or its Supplementary Information.

The authors declare there is no conflict.