The Role of Hydrogen Bond Donor on the Extraction of Phenolic Compounds from Natural Matrices Using Deep Eutectic Systems

Recently, deep eutectic systems (DESs) as extraction techniques for bioactive compounds have surfaced as a greener alternative to common organic solvents. In order to study the effect of these systems on the extraction of phenolic compounds from different natural sources, a comprehensive review of the state of the art was carried out. In a first approach, the addition of water to these systems and its effect on DES physicochemical properties such as polarity, viscosity, and acidity was investigated. This review studied the effect of the hydrogen bond donor (HBD) on the nature of the extracted phenolics. The effects of the nature of the HBD, namely carbon chain length as well as the number of hydroxyl, methyl, and carbonyl groups, have shown to play a critical role in the extraction of different phenolic compounds. This review highlights the differences between DES systems and systematizes the results published in the literature, so that a more comprehensive evaluation of the systems can be carried out before any experimental trial.


Introduction
Most phenolic compounds are secondary metabolites [1], found in a wide number of plants, fruits, and vegetables, that are responsible for plant growth and development while also displaying sensorial, structural, and defensive properties [2,3]. From a chemical point of view, they are characterized by the presence of one or more hydroxyl groups bound to an aromatic ring [2,4,5]. Depending on the number of carbon atoms, they can be classified into several groups, including phenolic acids (hydroxybenzoic and hydroxycinnamic acids), flavonoids (flavones, flavanones), and condensed tannins (lignin) [2]. Phenolic compounds have strong antioxidant activities [6,7], granting them desirable biological and pharmacological properties that include but are not limited to anti-inflammatory, neuroprotective, antimutagenic, and anticarcinogenic properties [6][7][8][9]. For these reasons, phenolic compounds have found a wide range of applications in several areas of interest such as cosmetics, food additives, nutraceuticals, and pharmaceuticals [4,10,11]. Phenolic compounds can be obtained either via synthetic processes or from natural sources. The synthesis of phenolic compounds is done through organic synthesis, which usually requires the preparation of intermediary species and multiple reaction steps. These individual steps require long reaction times, high temperatures, the use of toxic and harmful reagents, and the need for further purification in order to remove unwanted by-products and waste material, resulting in low yields [12]. For natural sources such as plant and vegetable matrices, an extraction step is required to obtain the desired compounds. There are a wide variety of natural matrices, however; wastes from the agricultural and food industries [13] such as wine lees and olive pomace [14][15][16][17] can be used as source materials, thus reducing wastes from the agricultural and food industries [13] such as wine lees and olive pomace [14][15][16][17] can be used as source materials, thus reducing the environmental impact of the industries. The extraction step is crucial, since these compounds can be degraded and therefore have their activity compromised by the processing methodologies employed.
The extraction of phenolic compounds from solid plant matrices can be carried out using either conventional methods or alternative methods [6]. Conventional extraction methodologies include solid/liquid and Soxhlet extractions, maceration, and percolation, which are associated with long extraction periods and high temperatures that can lead to phenolic thermal degradation and low yields, as well as the use of extensive amounts of toxic and harmful organic solvents and the production of high quantities of waste products [4,7,18].
Alternative extraction methodologies like enzyme-assisted extraction, ultrasoundassisted extraction (UAE), microwave-assisted extraction (MAE), subcritical and supercritical extractions, and high-pressure-assisted extractions [19][20][21] can be used to overcome some of the problems associated with conventional extraction methods, since they offer lower extraction times, higher yields, and use of lower amount of solvents, whether they be green or organic solvents [6]. However, despite the advantages modern methods may have over conventional methods, both groups of methods have problems regarding solvent toxicity, thermal instability, polarity, solubility, and poor selectivity [4], as well as the use of specialized and high-cost equipment [6,22]. Furthermore, common organic solvents such as methanol, ethanol, and acetone [6,23] are highly volatile and flammable, making them health and environmental hazards [24,25]. Considering that the applications of most of the extracted phenolic compounds are in the food, nutraceutical, and pharmaceutical industries, where the use of organic solvents such as ethanol, methanol, and dichloromethane is heavily regulated [26], there is a need to find greener and safer options that are harmless for human consumption. Recently, deep eutectic solvents have been gaining a lot of attention as green extraction solvents. As it can be seen in Figure 1, research on DES extractions has grown exponentially since 2012, with places like China, Iran, and Turkey producing the most research on the topic.

Deep Eutectic Systems
Deep eutectic systems are the result of the complexation of a hydrogen bond acceptor (HBA) and a hydrogen bond donor (HBD), which are usually naturally occurring, and biodegradable compounds. For this reason, it is expected that DES will be biodegradable and present a lower toxicity when compared to other solvents. This novel class of solvents has a set of advantageous properties, such as low melting points, low volatility, nonflammability, low vapor pressure, polarity, chemical and thermal stability, miscibility,

Deep Eutectic Systems
Deep eutectic systems are the result of the complexation of a hydrogen bond acceptor (HBA) and a hydrogen bond donor (HBD), which are usually naturally occurring, and biodegradable compounds. For this reason, it is expected that DES will be biodegradable and present a lower toxicity when compared to other solvents. This novel class of solvents has a set of advantageous properties, such as low melting points, low volatility, nonflammability, low vapor pressure, polarity, chemical and thermal stability, miscibility, and high solubility [4,10,18,27,28], properties that make DESs a safer alternative to conventional ex-Molecules 2021, 26, 2336 3 of 32 traction solvents. Additionally, they can be produced at low costs and high yields without the need for further purification steps [18,24,25,27]. Since DESs are formed through intermolecular interactions such as hydrogen bonding, there is no chemical reaction involved, hence there is no production of secondary compounds. As such, every atom in the system is used, meaning that DES production should have 100% atom economy. Similarly, since DESs result from the molecular interactions of two or more compounds, no waste is usually produced, meaning that DESs have a very low E-Factor. The intermolecular interactions, mainly hydrogen bond interactions, between the HBA and the HBD are responsible for the physicochemical characteristics of the DES [10,29], which means that by changing either the HBA/HBD ratio or one of the components, the DES can be specifically tailored for different applications. This is a major advantage in terms of achieving desirable properties and improving the extraction efficiency [25,27]. The preparation of a DES can be done either via the stirring and heating method [22,[29][30][31], where the HBA and HBD are mixed in the desired ratio at high temperatures until a clear liquid forms, or by the freeze-drying method [22,25,29], where the HBA/HBD mixture is dissolved in water, brought to −80 • C, and lyophilized until a constant weight is reached. With the exponential growth in this field of research and the versatility of DESs, a question worth answering is whether the influence of the HBD plays a role in the extraction selectivity of phenolic compounds.
DESs are able to dissolve a wide range of compounds, both polar and nonpolar, making them a desirable extraction medium for phenolic compounds [16]. Furthermore, they are also able to form hydrogen bonds with phenolic compounds, improving their dissolution and extraction ability [32,33]. Extractions of phenolic compounds using DESs are often solid/liquid extractions in which a liquid solvent, the DES, is used to extract compounds of interest from a solid plant matrix. There are several extraction techniques [34], but the most used are heat stirring (H/S), ultrasound-assisted extraction (UAE), and microwave-assisted extraction (MAE). While the use of MAE is a more efficient method, producing higher yields in relatively shorter times, the uses of UAE and H/S are more prevalent, since the equipment required to perform these extractions, i.e., a stirring plate or an ultrasound bath, is easily accessible and less expensive than microwave equipment and offers better potential for large-scale processes [6,35].
Over the last few years, several review articles have been published relating to deep eutectic systems, as well as their use in the extraction of polyphenolics. While the review of Clarke et al. [36] is geared towards green and sustainable solvents in general, such as ionic liquids, DES, liquid polymers, and supercritical CO 2 , important considerations regarding toxicity are still brought up, since even if the components of a DES have low toxicities or are nontoxic by themselves, the same is not always true when they are mixed together [37,38]. In more specific cases regarding eutectic systems, the work by Liu et al. [29] is a comprehensive review related to the fundamentals of natural deep eutectic systems (NaDESs), mainly regarding how the hydrogen bonding interactions between the HBD and the HBA, studied by NMR and FTIR, relate to the system's physicochemical properties, such as conductivity, viscosity, polarity, solubility, stability, and biocompatibility, while also discussing potential applications of NaDESs as extraction media, chromatographic media, and biomedical carriers. Works from Bubalo et al. [39], Cunha et al. [34], and Zainal-Abidin et al. [4] discuss the fundamentals of DESs while exploring the application of DESs in the extraction of phenolic compounds from natural matrices. The work of Bubalo et al. [39] is focused on the comparison of different extraction green solvents such supercritical CO 2 , subcritical water, and NaDESs, while Cunha et al. [34] deal with different extraction methods, such as ultrasound-assisted extraction or microwave-assisted extractions. Zainal-Abidin et al. [4] reviewed and, more recently, Fernández et al. [40] focused on the extraction of different plant metabolites and how their physicochemical properties may influence the extraction efficiencies of DESs. All of the previously mentioned articles show the versatility of DESs in the extraction of different bioactive compounds from natural matrices.
While building on the ideas in some of the reviews discussed, the intent of this work was to study the extraction selectivity of different phenolic acids and flavonoids using DESs, by trying to find a correlation between the functional group of the HBD and the chemical structure of the extracted compounds. To do so, a review of the state of the art was conducted. Since 2015, around 155 articles on the extraction of phenolic compounds using DESs have been published ( Figure 2). All of the previously mentioned articles show the versatility of DESs in the extraction of different bioactive compounds from natural matrices. While building on the ideas in some of the reviews discussed, the intent of this work was to study the extraction selectivity of different phenolic acids and flavonoids using DESs, by trying to find a correlation between the functional group of the HBD and the chemical structure of the extracted compounds. To do so, a review of the state of the art was conducted. Since 2015, around 155 articles on the extraction of phenolic compounds using DESs have been published ( Figure 2). In a first approach to the available data, articles were selected if they presented colorimetric and/or HPLC data regarding the extraction of different phenolic acids or flavonoids using a DES and a reference organic solvent such as water, ethanol, and methanol. Due to the focus on phenolic acids and flavonoids, articles pertaining to the extraction of other compounds such as terpenes, lignocellulosic compounds, or metals from plant and vegetable matrices were not considered. Using the described criteria, around 28 articles were found for study, which comprised 18% of articles published in the last six years. The articles reviewed showed that at least 86 individual chemical compounds were identified ( Figure 3) and extracted from 32 different plant matrices.  In a first approach to the available data, articles were selected if they presented colorimetric and/or HPLC data regarding the extraction of different phenolic acids or flavonoids using a DES and a reference organic solvent such as water, ethanol, and methanol. Due to the focus on phenolic acids and flavonoids, articles pertaining to the extraction of other compounds such as terpenes, lignocellulosic compounds, or metals from plant and vegetable matrices were not considered. Using the described criteria, around 28 articles were found for study, which comprised 18% of articles published in the last six years. The articles reviewed showed that at least 86 individual chemical compounds were identified ( Figure 3) and extracted from 32 different plant matrices.
All of the previously mentioned articles show the versatility of DESs in the extraction of different bioactive compounds from natural matrices.
While building on the ideas in some of the reviews discussed, the intent of this work was to study the extraction selectivity of different phenolic acids and flavonoids using DESs, by trying to find a correlation between the functional group of the HBD and the chemical structure of the extracted compounds. To do so, a review of the state of the art was conducted. Since 2015, around 155 articles on the extraction of phenolic compounds using DESs have been published ( Figure 2). In a first approach to the available data, articles were selected if they presented colorimetric and/or HPLC data regarding the extraction of different phenolic acids or flavonoids using a DES and a reference organic solvent such as water, ethanol, and methanol. Due to the focus on phenolic acids and flavonoids, articles pertaining to the extraction of other compounds such as terpenes, lignocellulosic compounds, or metals from plant and vegetable matrices were not considered. Using the described criteria, around 28 articles were found for study, which comprised 18% of articles published in the last six years. The articles reviewed showed that at least 86 individual chemical compounds were identified ( Figure 3) and extracted from 32 different plant matrices.  Regarding the solvents used, 173 DESs with different HBA/HBD ratios were identified, with 93 of these being unique combinations of an HBA/HBD. From the 173 combinations used, 140 used an ammonium salt as the HBA, such as choline chloride (105 systems) or betaine (35 systems), which accounted for 80% of the analysed DESs. Despite its wide use, the European Commission for Cosmetic Ingredients [26] has banned choline chloride use due to its irritant properties. As a possible alternative, betaine has been proposed [33,41]. Aroso et al. [41] showed that using the same HBD, DESs based on betaine proved more difficult to prepare, often requiring the use of water as a ternary component. The structural differences between choline chloride and betaine will lead to different interactions with the bioactive compounds extracted, either hindering or promoting the extraction yield. Fanali et al. [42] tested similar DESs based on choline chloride and betaine in the extraction of spent coffee grounds. They found that in this case, betaine-based systems had higher extraction yields than their choline-chloride-based counterparts. While this shows that the HBA has an influence in extraction yields, for the purpose of this work the effect of the HBA was not considered, with greater importance given to the effect of the HBD.
The HBDs of the reported systems can be divided according to their functional groups ( Figure 4), with systems based on alcohols (79 systems), organic acids (46 systems), and sugars (30 systems) being primarily used.  Regarding the solvents used, 173 DESs with different HBA/HBD ratios were identified, with 93 of these being unique combinations of an HBA/HBD. From the 173 combinations used, 140 used an ammonium salt as the HBA, such as choline chloride (105 systems) or betaine (35 systems), which accounted for 80% of the analysed DESs. Despite its wide use, the European Commission for Cosmetic Ingredients [26] has banned choline chloride use due to its irritant properties. As a possible alternative, betaine has been proposed [33,41]. Aroso et al. [41] showed that using the same HBD, DESs based on betaine proved more difficult to prepare, often requiring the use of water as a ternary component. The structural differences between choline chloride and betaine will lead to different interactions with the bioactive compounds extracted, either hindering or promoting the extraction yield. Fanali et al. [42] tested similar DESs based on choline chloride and betaine in the extraction of spent coffee grounds. They found that in this case, betaine-based systems had higher extraction yields than their choline-chloride-based counterparts. While this shows that the HBA has an influence in extraction yields, for the purpose of this work the effect of the HBA was not considered, with greater importance given to the effect of the HBD.
The HBDs of the reported systems can be divided according to their functional groups ( Figure 4), with systems based on alcohols (79 systems), organic acids (46 systems), and sugars (30 systems) being primarily used. Thirty-nine individual compounds were used as HBDs to form DESs, with compounds such as glucose (13 systems), lactic acid (14 systems), citric acid (6 systems), sucrose (8 systems), and glycerol (8 systems) being some of the most used. Choline chloride was one of the first HBAs reported by Abbot et al. [31] in 2004 and systems using it, alongside some of the HBDs mentioned, have been studied and characterized [27,28,30,31,34,43]. Since the systems are described in the literature, their use in new applications such as extractions can be considered as a starting point or a comparison term before testing new systems.
While it is important to study the extraction conditions used, such as extraction time and temperature, these are often set between 45 to 60 min and 40 °C to 65 °C and will not be discussed here.  Thirty-nine individual compounds were used as HBDs to form DESs, with compounds such as glucose (13 systems), lactic acid (14 systems), citric acid (6 systems), sucrose (8 systems), and glycerol (8 systems) being some of the most used. Choline chloride was one of the first HBAs reported by Abbot et al. [31] in 2004 and systems using it, alongside some of the HBDs mentioned, have been studied and characterized [27,28,30,31,34,43]. Since the systems are described in the literature, their use in new applications such as extractions can be considered as a starting point or a comparison term before testing new systems.
While it is important to study the extraction conditions used, such as extraction time and temperature, these are often set between 45 to 60 min and 40 • C to 65 • C and will not be discussed here.
Besides the experimental conditions used during extraction, there is also a need to take into consideration the physicochemical properties of the DES. Viscosity and polarity are properties that surely play a role in the type of bioactive compound extracted; however, this is information that is not always readily available. While scarce, authors such as Dai et al. [32] have published experimental data regarding some physicochemical proper-ties and, recently, Haghbakhsh et al. [44,45], Bakhtyari et al. [46], and Taherzadeh et al. [47] developed mathematical models to describe these properties. The problem with this information relates to the fact that it is obtained for pure DESs and it is common practice to add water to the systems (Figure 4), which will modulate the properties in each individual case. As such, it is not possible to accurately explore the effects of viscosity and polarity on the types of bioactive compounds extracted.
High viscosities, which are a characteristic of most DESs, are a drawback of these systems, since they act as a hindrance in mass transfer phenomena and thus lower the extraction efficiencies of the DES [48]. Viscosity is a physicochemical property highly dependent on temperature, meaning that upon increasing the temperature by even a few degrees, the viscosity of the system is reduced drastically and hence, mass transfer phenomena is improved. Nonetheless, as some phenolic compounds are thermolabile, an increase in temperature may not always be the solution. This drawback can also be overcome by the addition of water to the systems [22,25,49]. Figure 5 shows the amount of water which has been added to the DESs used to perform extractions. The water content of the system varies between 10% and 80%, with the most commonly used being 20% to 30% w/w.
Besides the experimental conditions used during extraction, there is also a need to take into consideration the physicochemical properties of the DES. Viscosity and polarity are properties that surely play a role in the type of bioactive compound extracted; however, this is information that is not always readily available. While scarce, authors such as Dai et al. [32] have published experimental data regarding some physicochemical properties and, recently, Haghbakhsh et al. [44,45], Bakhtyari et al. [46], and Taherzadeh et al. [47] developed mathematical models to describe these properties. The problem with this information relates to the fact that it is obtained for pure DESs and it is common practice to add water to the systems (Figure 4), which will modulate the properties in each individual case. As such, it is not possible to accurately explore the effects of viscosity and polarity on the types of bioactive compounds extracted.
High viscosities, which are a characteristic of most DESs, are a drawback of these systems, since they act as a hindrance in mass transfer phenomena and thus lower the extraction efficiencies of the DES [48]. Viscosity is a physicochemical property highly dependent on temperature, meaning that upon increasing the temperature by even a few degrees, the viscosity of the system is reduced drastically and hence, mass transfer phenomena is improved. Nonetheless, as some phenolic compounds are thermolabile, an increase in temperature may not always be the solution. This drawback can also be overcome by the addition of water to the systems [22,25,49]. Figure 5 shows the amount of water which has been added to the DESs used to perform extractions. The water content of the system varies between 10% and 80%, with the most commonly used being 20% to 30% w/w. Figure 5. Amount of water used to aid extraction of phenolic compounds from different natural matrices, with n.a. Meaning information not available or that water was not added.
The amount of water in a DES has been a topic of discussion due to the possibility of water disrupting the intermolecular interactions between the HBA and the HBD. Recent studies [50][51][52] have shown that water is able to form hydrogen bonds with the HBA and the HBD, and thus help to modulate DES properties. However, as the amount of water in the system increases, the intermolecular interaction between the HBA and the HBD weakens until the mixture becomes a solution, with Zhekenov et al. [51] suggesting that this limit is at 50% mol. fraction of water.
The addition of water plays a very important role in DES properties such as viscosity, polarity, and acidity and, consequently, on the extraction yield of the DES.
In one of previous study, Mansur et al. [53] found that the addition of water to a DES significantly improved the extraction yields of flavonoids when compared with Figure 5. Amount of water used to aid extraction of phenolic compounds from different natural matrices, with n.a. Meaning information not available or that water was not added.
The amount of water in a DES has been a topic of discussion due to the possibility of water disrupting the intermolecular interactions between the HBA and the HBD. Recent studies [50][51][52] have shown that water is able to form hydrogen bonds with the HBA and the HBD, and thus help to modulate DES properties. However, as the amount of water in the system increases, the intermolecular interaction between the HBA and the HBD weakens until the mixture becomes a solution, with Zhekenov et al. [51] suggesting that this limit is at 50% mol. fraction of water.
The addition of water plays a very important role in DES properties such as viscosity, polarity, and acidity and, consequently, on the extraction yield of the DES.
In one of previous study, Mansur et al. [53] found that the addition of water to a DES significantly improved the extraction yields of flavonoids when compared with extractions where no water was added. They also studied the effect of different amounts of water content in the DES, from 20% to 80%, and concluded that despite the addition of water improving the extraction of flavonoids when compared with a DES with no water, 20% water provided the best extraction results. Besides affecting the viscosity of the systems, the addition of water will also change other physicochemical properties, such as polarity and acidity of the different solvents, which in turn also play an important role in the results obtained. Regarding polarity, Xu et al. [21] showed that fine-tuning the system polarity can be used to improve the extraction yield of different flavonoids. In their work, they performed multiple extractions of citrus flavonoids using choline-chloride-based systems with different HBDs. By keeping the HBA constant, the polarity of the systems was made dependent on the different HBDs used and polarity was measured using the n-octanol/water partition coefficient of the HBD. The authors found that systems with lower polarities extracted higher amounts of citrus flavonoids such as hesperidin, which have low polarities. The higher extraction efficiencies can be explained by the similar polarities between the DES and phenolic. Besides playing an important role in the viscosity and polarity of DESs, the addition of water can also be used to tune DES acidity. DESs are a mixture of compounds complexed by intermolecular interactions, therefore, there are no free protons in the liquid media; thus, measuring the system's acidity using a conventional pH electrode does not render a value with physical meaning. Abbott et al. [54] published work showing a spectrophotometric method that could be used to measure DES acidity; however, the technique described is not straightforward. Nonetheless, upon adding water to a DES, which will be free and not part of the DES structure, there will be protons in solution and therefore the pH can be measured using a conventional electrode.
Postprocessing and isolation of the active compounds after extraction is not an easy task. It is difficult to separate the DES from the extracted phenolics due its low vapor pressure and usually high viscosity. As such, most purification methods involve the use of membrane and antisolvent processes which can be expensive [16,33]. An alternative that has recently been gaining attention [22] is the use of tailored-made DESs that can both extract phenolic compounds from plant matrices and act as stabilizing agents, being incorporated into the final product, namely for further applications in cosmetic and pharmaceutical formulations. In this sense, the idea is to formulate a bioactive composition that can be incorporated into a final product. Regarding the stability of bioactives, Dai et al. [55] showed that flavonoids such as quercetin are able to form hydrogen bonds with a DES made of choline chloride and sucrose, which increased their stability in the DES. Other researchers [55,56] have also proven, following the degradation kinetics, that a DES can be used to stabilize phenolic compounds.

Discussion
As previously discussed, many different DESs have been used for extraction; however, a clear relationship between the DES and the nature of extracted phenolic compounds is still difficult. This work sought to shed light on the issue by studying the type of HBD used and their effects on different phenolic compounds by conducting a review of the state of the art. Due to the large amount of information available, several selection criteria were defined to refine the information collected. The selection criteria were defined as follows: • Systems were selected as data points if the DES was a binary mixture; • Ternary eutectic mixtures were only considered if the third component was water; • If a DES was shown to be able to extract more than one compound from the same plant matrix, then a data point was considered for each compound extracted. For example, Viera et al. [57] reported, among others, a choline chloride:citric acid system (2:1) that was able to extract three types of flavonoids, querticin-3-O-glucoside, querticin-Opentoside, and 3-O-caffeyolquinic acid, from walnuts and therefore three data points were considered; • Data points were only selected if there data were reported on the yield of extraction using the DES and one of three conventional extraction solvents: water, ethanol, or methanol, which were used for comparison purposes.
Since the physicochemical properties such as density, viscosity, polarity, pH, etc. of DESs are a result of the molecular interactions [58] between the HBA and the HBD, and a wide range of HBDs have been used for extraction, the functional group of the HBD was used as a reference to interpret the results reported in the literature, as it ultimately influences the final physicochemical properties of the systems. In a first approach, the DES were grouped by the HBD functional group (alcohol, amide, amino acid, organic acid, and sugar), and then each group was divided according to the nature of the extracted compounds (flavonoid, phenolic acids, total flavonoids, and total phenolics; Figure 6). methanol, which were used for comparison purposes. Since the physicochemical properties such as density, viscosity, polarity, pH, etc. of DESs are a result of the molecular interactions [58] between the HBA and the HBD, and a wide range of HBDs have been used for extraction, the functional group of the HBD was used as a reference to interpret the results reported in the literature, as it ultimately influences the final physicochemical properties of the systems. In a first approach, the DES were grouped by the HBD functional group (alcohol, amide, amino acid, organic acid, and sugar), and then each group was divided according to the nature of the extracted compounds (flavonoid, phenolic acids, total flavonoids, and total phenolics; Figure 6). Considering the functional group of the HBD and the type of phenolic extracted, we observed that flavonoids have mostly been extracted using HBDs based on organic acids, alcohols, and sugars, while phenolic acids have been mostly extracted using alcohol-and organic-acid-based HBDs. To better understand the role of HBDs, the extraction yields of the DES were compared with the extraction yields of common organic solvents to determine the extraction efficiency (EE) of the DESs. Extraction efficiencies were defined as the ratio between the extraction yield of the DES and the extraction yield of the organic solvent used as control Equation (1), when data for both solvents were reported by the authors. The data were selected for the optimized parameters when calculating the EE.

=
(1) It is important to note that since the EE is the ratio between DES yields and organic solvent yields, in studies with an EE lower than 1, the DES had a lower extraction yield than the organic solvent. The following tables present the EE values of different target compounds extracted using DESs, using HBAs based on ammonium salts such as choline chloride grouped by the functional group of the HBDs, extraction technique, and natural matrices. Due to the high number of data points, the information was condensed by the phenolic extracted. Table 1 shows the effect of different HBDs on the EE of flavonoids from several matrices using water as the comparison organic solvent, while Table 2 and Table 3 present similar information using ethanol and methanol respectively as the extraction solvent for comparison. Unless otherwise mentioned, the DESs presented in the Considering the functional group of the HBD and the type of phenolic extracted, we observed that flavonoids have mostly been extracted using HBDs based on organic acids, alcohols, and sugars, while phenolic acids have been mostly extracted using alcohol-and organic-acid-based HBDs. To better understand the role of HBDs, the extraction yields of the DES were compared with the extraction yields of common organic solvents to determine the extraction efficiency (EE) of the DESs. Extraction efficiencies were defined as the ratio between the extraction yield of the DES and the extraction yield of the organic solvent used as control Equation (1), when data for both solvents were reported by the authors. The data were selected for the optimized parameters when calculating the EE.
It is important to note that since the EE is the ratio between DES yields and organic solvent yields, in studies with an EE lower than 1, the DES had a lower extraction yield than the organic solvent. The following tables present the EE values of different target compounds extracted using DESs, using HBAs based on ammonium salts such as choline chloride grouped by the functional group of the HBDs, extraction technique, and natural matrices. Due to the high number of data points, the information was condensed by the phenolic extracted. Table 1 shows the effect of different HBDs on the EE of flavonoids from several matrices using water as the comparison organic solvent, while Tables 2 and 3 present similar information using ethanol and methanol respectively as the extraction solvent for comparison. Unless otherwise mentioned, the DESs presented in the tables are based on choline chloride. Natural matrices were found to be in the form of a freeze-dried powder before use.        When using water as a control (Table 1), an overall improvement in EE can be seen, while the improvement is not as noticeable when using ethanol and methanol. This effect may be related to solute-solvent interactions, particularly the effect of solvent polarity in extraction. While this is an important factor, there are other factors to take into consideration that are also observable when using ethanol and methanol as comparison solvents (Tables 2 and 3). In the case of alcohol-based HBDs, there are two main factors that influence EE: carbon chain length and the number of hydroxyl groups. Carbon chain length appears to have an inverse relationship with EE, since a decrease of the number of carbons in the chain (1,4-butanediol < 1,2-propanediol < ethylene glycol) increased the EE of compounds such as apigenin, hesperidin, and luteolin. On the other hand, looking at the effect of sugar alcohols in the same compounds, an increase of EE can be seen with an increase of the number of hydroxyl groups present, as in glycerol < xylitol < sorbitol. When working with amide-based systems, a clear relationship is harder to see, since EE values are similar; however, a higher degree of methylation appears to improve EE, especially in the case of hesperidin. The major effect of organic-acid-based systems in the extraction of flavonoids seems to be tied to the degree of carboxylation of the HBD. Looking at the extraction of quercetin glycosides, Vieira et al. [57] showed that the EE decreases with increasing degrees of carboxylation. This trend can also be observed in the extraction of hesperidin from Nobis tangerine; however, in the case of luteolin and apigenin, di and tricarboxylic acids produced an increase of the EE. The main difference could be attributed to the fact that hesperidin, quercetin-3-O-glucoside, and quercetin-O-pentoside are glycosides, while luteolin and apigenin are aglycones. As such, it can be inferred that when targeting aglycones, a higher degree of carboxylation may be desired, while the reverse is favored when trying to extract glycosides. Using sugar based HBD's, the complexity of the sugar behaves like the carboxylation degree in organic acids. In this case simpler sugars such as fructose and glucose show that the use of DES increases the EE of different flavonoids.
A similar approach was taken to look for correlations between the functional group of the HBD and the extraction of phenolic acids. These correlations can be found in Tables 4-6, with water, ethanol, and methanol as the respective comparison solvents.         Looking at the effect of the HBDs in the extraction of phenolic acids, similar trends regarding alcohol chain length, hydroxylation, amide methylation, organic acid carboxylation, and sugar complexity can be observed. Despite the existence of outliers, alcohol-based systems have increased extraction efficiencies. This increase can be related to the systems' acidity. When the HBA remains constant, the acidity of the DES is determined by the HBD. Since alcohols generally have higher pKa [78] values than organic acids, systems formed with these compounds will be more basic in nature. As such, it is possible that the phenolic acids have acid-base interactions with the alcohol-based systems, promoting their extraction.
Another interesting research avenue is the application of the Hildebrand Hansen solubility parameters, which measures the solute-solvent interactions between the DES and the bioactive compounds. While data are still limited, Salehi et al. [79] used molecular dynamics to shown that DESs with choline chloride and urea, ethylene glycol, glycerol, malonic acid, and oxalic acid have high solubility parameters, indicating that these can be considered polar solvents.

Conclusions
In recent years, research into DESs and their application for the extraction of phenolic compounds has grown exponentially. This growth can be attributed to their ease of preparation and use, as well as the green and safer properties inherent to DESs. When used as extraction media, three main extraction methodologies have been used, namely UAE, U/S, and MAE, with UAE and H/S being the most used. The amount of water in the system has been shown to play a critical role in the tuning of different physicochemical properties, namely viscosity, polarity, and acidity. Using the HBD functional group as a starting point for comparison purposes, this review evaluated the extraction efficiencies reported in the literature as a comparison tool to understand the effects of the HBD on the types of phenolic compounds extracted and the yield of extraction. The extracted compounds fell into under two major phenolic families, phenolic acids and flavonoids, mainly anthocyanins, flavones, and flavonols. Taking into consideration the nature of the HBDs and the type of phenolic extracted, some correlations can be drawn regarding the effectiveness of the HBDs. When using alcohol-based systems, small carbon chains and a high degree of hydroxylation are desirable, while a high degree of methylation is preferable in amide-based systems. In organic-acid-based systems, there are two factors which play a major role in the extraction yield: the degree of carboxylation and whether the extracted compound is an aglycone or not. If the extracted compound is an aglycone, a higher degree of carboxylation is more desirable, while the reverse is true for glycosides. If the desired compound is a phenolic acid, alcohol-based systems, which are more basic in nature, are desired, with the same rules previously discussed still applying. In the future, the influence of critical factors such as amount of water, pH, and polarities, which have not yet been reported in the literature, could also help develop our understanding of the effect of HBDs in the extraction efficiency and selectivity of these solvents towards particular classes of phenolic compounds.