*3.1. Static Adsorption/Desorption*

The adsorption capacity and desorption ratio are typically regarded as the two main benchmarks for selecting resins. We first selected the best resin type and amount for adsorbing polyphenolic compounds from hazelnut skin extracts. The adsorption ratios and capacities of the resins according to their amounts are shown in Tables 2 and 3, respectively. The absorption ratio increased with increasing amounts of resin, and the maximum value was obtained using 5 g of resin. XAD 16 and XAD 4 showed better adsorption ratios and capacities compared to that of XAD 7 for all amounts of resin evaluated. Using 5 g of XAD 16, 65.06 ± 0.14% of polyphenols in the extract were adsorbed, and this resin showed the highest adsorption capacity (40.05 ± 0.55 mg GAE/g dry resin). The adsorption capabilities of resins are related to both the target compound and absorbent properties, such as polarity, particle size, surface area, pore diameter, and chemical structure. Particularly, polyphenol compounds can be absorbed by macroporous resins via physical mechanisms, such as van der Waals forces, hydrogen bonds, and π-π conjugation between phenolics and the benzene rings of resins [16]. Polyphenols contain hydrogen groups and benzene rings and, depending on their structure, exhibit different polarities. Although XAD 16 and XAD 4 have similar polarities, XAD 16 provides a higher surface area and pore volume size and absorbs more polyphenols.

**Table 2.** Adsorption ratios (A%; mean ± standard deviation) of polyphenols from an extract of roasted hazelnut skin by different resin types and amount and results of ANOVA with Duncan's test.


Means in each column with the same uppercase letter are not significantly different according to Duncan's test (*p* < 0.05); means in each row with the same lowercase letter are not significantly different according to Duncan's test (*p* < 0.05); \*\*\* *p* < 0.001.


**Table 3.** Adsorption capacity (qa; mean ± standard deviation) of polyphenols from an extract of roastedhazelnutskinbydifferentresintypesandamountandresultsofANOVAwithDuncan'stest.

Means in each column with the same uppercase letter are not significantly different according to Duncan's test (*p* < 0.05); means in each row with the same lowercase letter are not significantly different according to Duncan's test (*p* < 0.05); \*\*\* *p* < 0.001.

The desorption ratios of absorbed polyphenols from resin using different concentrations of ethanol (50%, 70%, and 99.9% *v/v*) are shown in Table 4. A significant two-way interaction was observed (*p* < 0.001), confirming that the change in the number of resins or concentration of solvent affected the amount of polyphenol desorption for each resin type. Because a significant difference was found between each of the two variables and the main effects were significant, one-way ANOVA was performed to compare the differences within each group. During the static adsorption stage, XAD 4 and XAD 16 exhibited the maximum desorption ratio when 70% *v/v* ethanol was used, whereas XAD 7 showed the highest desorption when 50% *v/v* of ethanol was used as the solvent. Non-polar resins showed better adsorption and desorption of polyphenols compared to the slightly polar resin. Using a 70% *v/v* ethanol solution, 76.64% and 81.17% of the polyphenols were desorbed that had been absorbed by 5 g of XAD 4 and XAD 16, respectively, whereas the lowest efficiency was observed when 99.9% ethanol was used to recover the polyphenols from 1 g of XAD 7. Similarly, Wang et al. [10] compared different concentrations of ethanol solution (10–100%) to recover polyphenols adsorbed by HPD-300 (non-polar) resin. They observed that the highest content of polyphenol was recovered using 60% aqueous ethanol, which was eight-fold that of the crude extract. In addition, approximately 95% of the polyphenol was present in the 60% and 80% ethanol fractions. As explained by Xi et al. [17], pure ethanol increases the desorption of some impurities, but polyphenols are not completely dissolved at lower ethanol concentrations. Leyton et al. [11] obtained similar results in a comparison of different Amberlite XAD resins for purification of phlorotannins from *Macrocystis pyrifera*, and XAD 16 N showed good results with a desorption ratio of 38.2%.

**Table 4.** Static desorption ratios (mean ± standard deviation) of XAD 16, XAD 4, and XAD 7 Amberlite resins using different ethanol solutions, and ANOVA results with Duncan's test.


Means with same uppercase letter are not significantly different between ethanol concentration for each resin type, according to the Duncan's test (*p* < 0.05); means in each row with the same lowercase letter are not significantly different according to Duncan's test (*p* < 0.05); \*\*\* *p* < 0.001.

The adsorption process consists of different stages and does not remain stable during the adsorption phase. Generally, target molecules are absorbed through their mass transfer from the boundary layer, diffusion into the pores of the adsorbent, and/or adsorption at the surface-active sites of the adsorbent [18]. The adsorption rate is directly correlated with the duration, solvent, and adsorbent material used.

The adsorption kinetics of polyphenols from the roasted hazelnut skin extracts obtained from the three Amberlite macroporous resins are shown in Figure 1. The adsorption quantity increased over time and adsorption was faster in the initial stages. The TPC value decreased by approximately 50% during the first 30 min by XAD 16 and XAD 4, and 1 h by XAD 7. After 1 h, the rate of adsorption gradually decreased; after 120 min of contact, only a minor change was observed because the surface binding sites of the macroporous resin were mostly saturated. The system may have reached equilibrium after 120 min. This trend is similar to that reported by Park and Lee [19]. Le et al. also observed that the adsorption equilibrium of polyphenols (sinapine) from rapeseed meal protein isolate by-products was obtained after 120 min on Amberlite XAD 16 resin [19]. Hou and Zhang reported that polyphenol adsorption equilibrium was reach after 4 h using a highly polar resin (NKA–II) [9]. The surface area of this resin is very close to that of XAD 16 and XAD 4; hence, use of nonpolar or slightly polar resin accelerates the adsorption process. As shown in Table 5, the adsorption kinetics are not best-described by a pseudo-first-order model because the absorbance capacity values were inconsistent with the values predicted using the first-order model [19].

**Figure 1.** Adsorption kinetic curve of polyphenols from extracts of roasted hazelnut skin with Amberlite microporous resins.

**Table 5.** Kinetic parameters of static adsorption phase evaluated using two model equations.


In contrast, there was good agreemen<sup>t</sup> between the experimental and calculated absorbance capacities predicted by the pseudo-second-order model, and the correlation coefficients were close to unity (R<sup>2</sup> > 0.99) for all types of resins. This suggests that the pseudo-second-order model can be applied to predict the kinetics of polyphenol adsorption from hazelnut skin extract using macroporous resins.

In addition, previous studies reported that the pseudo-second-order model is suitable for the adsorption of polyphenols from extracts [20–22]. Wang et al. reported that the adsorption patterns of polyphenols from *Eucommia ulmoides* Oliv. leaves on HPD-300, HPD-600, D-3250, X-5, D-140, NKA-9, D-101, and AB-8 resins fit well to pseudo-secondorder kinetics [10]. Soto et al. also demonstrated that the pseudo-second-order model fits better than the pseudo-first-order model with the adsorption process of phenols from wine vinasses by SP700 and XAD16 HP resins. They suggested that the process fits pseudosecond-order models better when the sorption system is controlled by chemisorption mechanisms [23].
