Valorization of Pinus taeda bark: source of phenolic compounds, tannins and fuel

Abstract

Pine bark is a byproduct of wood processing which is usually burnt for energy. This article analyzes the liquid–solid extraction of Pinus taeda bark to obtain phenolic compounds by using response surface methodology to determine extraction conditions. The independent variables studied were temperature, ethanol concentration and solid–liquid ratio, and the variables to be optimized were total extractives yield, phenolic content, antioxidant capacity and condensed tannins yield. In addition, the extract was characterized by Fourier transform infrared spectroscopy and UV–Vis spectroscopy. Furthermore, extraction kinetics were modeled, and mass transfer mechanisms were studied. The extraction condition that maximizes extraction yields was defined at 50 °C, with a solid–liquid ratio of 1/10 and with an ethanol concentration of 50%. The condensed tannins yield was 4.01 g catechin equivalent (CE)/100 g pine bark dry base (d.b.), the total extractive yield was 9.83 g extract/100 g pine bark d.b. and the Stiasny number was 77. The extract showed a FRAP antioxidant concentration of 20.89 mmol ascorbic acid equivalent (AAE)/100 g pine bark d.b and 0.35 mmol trolox (TRE)/g pine bark d.b. for ABTS assay. The results showed that extended Fick’s law was adequate to describe the extraction kinetics. The extraction did not significantly affect the calorific value of bark (21 kJ/g d.b.). After extraction, the ashes were reduced by 13% and potassium (K) by 48%. The extraction of Pinus taeda to obtain phenolic content is technically feasible, and this paper provides the scientific ground for further scaling the process.

1 Introduction

Currently, waste management has become a great challenge for many countries because there is a pressing need to preserve the required ecological standards without affecting the quality of life, despite the fact of the growing world’s population (Hoornweg et al. 2013). The reinsertion of mass and energy, formerly considered waste, in the economy is a goal to achieve a sustainable production system. Therefore, a circular economy must be designed, capable of allocating wastes and subproducts to sustainable value chains for further valorization of materials.

The pulp and sawmill industries use pine wood as a raw material for their processes and bark is one of the byproducts obtained, usually burnt to produce energy (Aspé and Fernández 2011). Uruguay has 160,000 ha of pine plantations (Boragno and Boscana 2019), mainly Pinus taeda destined to manufacture construction and carpentry products, which, considering the local growth of the species and the forestry cycles involved, could supply 3 mill m³ of logs yearly (Dieste et al. 2019). Assuming that on average 5% of that volume is bark, and that pine bark basic density is 380 kg/m³, it can be estimated that Uruguay has a potential stock of 60,000 annual dry ton of pine bark.

This solid organic byproduct must then be managed to comply with environmental standards; thus landfilling is not an option. In Uruguay, the most common use of pine bark is fuel to generate electricity (Dieste et al. 2016). That solution is not entirely satisfactory for a number of reasons: (1) in recent years, in Uruguay, the cost of electricity production has been drastically reduced due to the expansion of wind mill capacity, rendering biomass electricity generation barely competitive (Dieste et al. 2016); (2) pine bark contains a relatively high content of inorganic compounds, which increases the system ion conductivity, making the corrosion process faster in boilers, and reducing the melting point of ashes (Melissari 2012); (3) Uruguayan pines are grown in sandy soils, and therefore the contamination of bark with sand is highly common, causing blockage of the air nozzles of the furnace bed due to the coalescence and agglomeration of partially melted ash with sand particles (Melissari 2012).Therefore, it is necessary to find another economically feasible solution to valorize this secondary current.

Pine bark main components are lignin, cellulose and hemicelluloses. Currently, studies have focused on the composition of extractives (aromatic and aliphatic chemical compounds such as resins, tannins, phenolic compounds, dyes, lignans, glucosides, terpenes, alcohols, fats and waxes) paying particular attention to phenolic compounds due to its antioxidant properties (Sousa et al. 2018; Masendra et al. 2019). The bark differs from wood due to its high content of extractives soluble in water and/or organic solvents, ashes and phenolic compounds, being tannins their main fraction (Fradinho et al. 2002). Lignocellulose feedstock can potentially produce an array of chemicals, fuels and bioproducts, being phenolic compounds one of them. Phenolic compounds constitute a very small fraction (1–5%) of the biomass although they might have high commercial and economic value due to their proved antioxidant and antimicrobial properties (Xavier et al. 2017a; Neiva et al. 2020; Chmelová et al. 2020). It should be noted that very high fractions of phenolic compounds have been reported in some species, for example, the bark of Eucalyptus sideroxylon (22%) (Miranda et al. 2016).

There are many health benefits that can be attributed to antioxidant consumption, such as heart disease and gastro-intestinal cancer prevention (Ness and Powles 1997; Johnson 2004; Willcox et al. 2008). Tannins, which contain polyphenolic molecular structures are also presented in the pine species, and have many applications in many industrial activities. Conifers are scattered all over the world, that is why this source of tannins is highly important (Fradinho et al. 2002). Tannins are classified into two groups: hydrolysable tannins and condensed tannins (Kiser et al. 2013; Chupin et al. 2015). They are widely used in wine industry as an additive to improve desirable properties (Cheynier et al. 2006). Nowadays, tannins are also used to produce plywood adhesives, in a way that can reduce urea–phenol–formaldehyde in the production of such wood products (Kim 2009; Solt et al. 2019). In this way, fewer toxic gases are emitted during that process, and after the product is finished, contributing to the goal of minimal formaldehyde emission. To evaluate the ability of the extracts to be used in the formulation of adhesives, a parameter used is the Stiasny number which measures the amount of material capable of copolymerizing with formaldehyde in an acid medium (Yazaki and Hillis 1980; Vázquez et al. 2009).

Phenolic compounds can be extracted by different solvents and different extraction methods, such as conventional extraction, Soxhlet extraction, microwave-assisted extraction, or supercritical fluid extraction methods, among others (Venkatesan et al. 2020). The most common method for extracting phenolic compounds out of natural products is conventional solid–liquid extraction with solvents by maceration in a stirred tank. Phenolic compounds are both polar and non-polar substances, therefore either total extractives yield or antioxidant activity is highly dependent on the chosen solvent. Scientific research shows that mixtures of alcohol and water are more efficient for the extraction of the compounds of interest than mono component solvent systems. Ethanol in particular is GRAS (generally recognized as safe) and less expensive compared to other alcohols (Dai and Mumper 2010; Venkatesan et al. 2020). Furthermore, ethanol is suggested as a non-toxic solvent by the United States Food and Drug Administration (US FDA), for the extraction of phenolic compounds from plant materials (Venkatesan et al. 2019).

This work aims to contribute to valorize pine bark before using it as a fuel, and therefore proposes a system in which the quality of fuel will not be affected, and more value will be recovered from the raw material, creating a value chain. The objective of this work is to study the phenolic content obtained by the extraction process from Pinus taeda bark and to understand its extraction mechanism. This will find a base for future designing and scaling of the process. To do this, the characterization of pine bark was carried out. Then, the extraction of phenolic compounds in pine bark was optimized using the response surface methodology and its antioxidant capacity and the total extractives yield were evaluated. Three variables were studied: ethanol concentration, temperature and solid–liquid ratio. The kinetics of the extraction was studied by analyzing mathematical models considering different temperatures. The diffusion coefficients of phenolic compounds and extractives inside pine bark particles were estimated. The extract was chemically characterized for two applications: (1) adhesive (condensed tannins yield, Stiasny number); (2) antioxidant properties (ferric reducing antioxidant power (FRAP) method and 2,2′-azino-di(3-ethylbenzothiazoline-6-sulfonic acid (ABTS) assay). Additionally, the calorific value of the extracted bark was determined, since the extraction process valorizes bark, but the bulk of the material has still to be burned for two purposes: (1) energy; (2) volume reduction.

2 Material and methods

2.1 Materials

Pine bark samples were supplied by Lumin S.A. a plywood mill located in Tacuarembó, Uruguay. The samples were dried in an air flow dryer until a moisture of approximately 0.12 g H₂O/g dry basis (d.b.) was reached. The dried bark was then milled and particles with size between 0.3 and 1 mm were selected.

2.2 Reagents

All reagents were American Chemical Society (ACS) reagents quality or higher. Gallic acid, sulphuric acid, sodium carbonate, l-ascorbic acid, Folin–Ciocalteu reagent, iron(III) chloride hexahydrate, hydrochloric acid, acetic acid, sodium acetate 3-hydrate, potassium persulfate, acetic acid and acetonitrile were purchased from Merck (Darmstadt, Germany). ABTS, formaldehyde, methanol, ethanol and HPLC standards were acquired from Sigma (Steinheim, Germany). Trolox (TRE), 2,4,6-tri(2-pyridyl)-S-triazine (TPTZ), catechin hydrate were purchased from Fluka (Steinheim, Germany).

2.3 Extraction procedure

Milled dry bark was put in contact with ethanol–water mixture inside the incubator with proper solid–liquid ratio, temperature, and ethanol concentration according to the experimental design. The pine bark was placed in 250 mL Erlenmeyer flask together with the solvent inside the incubator. The extraction time and the stirring speed of the incubator were in all cases 90 min and 200 rpm, respectively. The resulting liquid extract was vacuum filtered using a Buchner funnel. The liquid extract and solid were refrigerated for subsequent analysis. Total extractives yield, phenolic content, antioxidant activity and condensed tannins yield were determined in the liquid extract. In addition, the liquid extract was characterized by Fourier transform infrared spectroscopy (FTIR) and UV–Vis spectroscopy. The total extractives yield was measured by drying a sample of the liquid extract up to constant weight. The results were expressed as g dry extract/100 g pine d.b. Higher heating value, ashes, and potassium content were measured in the extracted solid.

The value of the total amount of phenolic content of pine bark was obtained conducting a multi-stage extraction procedure. Firstly, solid phase was put in contact with solvent (ethanol–water mixture) at 80 °C, for an extraction time of 90 min and 200 rpm until equilibrium conditions were reached. Once in equilibrium state, a certain amount of liquid phase was taken away from the flask, leaving the solid phase and a known amount of liquid phase in the extraction flask. Then, new solvent was added to the extraction flask, and the process started again. At each stage of extraction, the phenolic content was determined in the liquid phase. The method was repeated until the detection limit of the analytical method was reached. Total amount of phenolic compounds in pine bark was determined by summing the quantity of phenolic content present in each stage, according to Eq. (1):

$$n_{m{\acute{a}}x} = \frac{{\mathop \sum \nolimits_{i = 1}^{n} V_{i} c_{i} }}{M}$$

(1)

where n_máx is the maximum extractable phenolic content per gram of dry pine bark (g GAE/g pine bark d.b.). V_i is the extraction volume expressed in liters (L) for stage i, c_i is the phenolic concentration in liquid phase in mg GAE/L for stage I, and M is the dry pine bark mass used for extraction in g d.b.

The samples were first extracted with hexane to evaluate the effect of bark resinous substances on the extraction. Successive extractions of pine bark were carried out first with hexane and then with 50% ethanol water. The extracted samples were dried and again extracted with water–ethanol at 50 °C with a solid–liquid ratio of 1/10 and 200 rpm. These conditions were selected for being one of the best found for this experimental design. Total extractives yield and phenolic content were determined in the liquid extract.

2.4 Analytical methods

2.4.1 Composition and higher heating value of native pine bark

The composition of pine bark (extractives, lignin, carbohydrates and ash) were determined following NREL protocols (Sluiter et al. 2008a, b, 2012). Firstly, the extractives were determined by water–ethanol extractions (Sluiter et al. 2008a). The structural components were quantified in the biomass free of extractives through an acid hydrolysis, from which a liquid and a solid fraction were obtained. Total lignin is the sum of soluble and insoluble lignin. Insoluble lignin corresponds to the solid fraction of hydrolysis. Soluble lignin was quantified in the liquid fraction with a Shimadzu UV spectrophotometer using a wavelength of 240 nm and an absorptive factor of 12 L/g cm. The extractives with hexane were determined in an accelerated extraction apparatus (SOX406 Fat analyzer, Hannon Instrument) for 3 h at 120 °C (Bikovens et al. 2013).

The sugars (glucose and the sum of xylose, galactose and mannose) were also determined in the liquid fraction, with a Shimadzu HPLC using Bio-Rad Aminex HPX-87H column and a refractive index detector (Sluiter et al. 2012).Total carbohydrates are the sum of the four sugars. All components were expressed in g/100 g of dry bark. The ashes were determined in a muffle at 575 °C before and after extraction (Sluiter et al. 2008b).

The basic density of bark was measured following the SCAN-CM 43:95 (1995) standard. Higher heating value of milled bark was measured in an adiabatic calorimeter by incineration in pure oxygen atmosphere according to UNE 164001:2005 EX standard method. The experiment was carried out with 1 g of milled bark before and after extraction.

For the determination of potassium, an atomic absorption spectroscopy by flame was used. The sample was calcined and the resulting ashes were treated with hydrochloric acid (SCAN-CM 63:05 2005).

2.4.2 Determination of phenolic content

Folin–Ciocalteu method was used to determine phenolic content in the extract as described by Xavier et al. (2017b). This method consists of a reaction between the extract and the Folin–Ciocalteu reagent and sodium carbonate. It is a non-specific test, which allows any type of phenol to be determined. It is based on the reducing power of phenolic hydroxyl groups (Sun et al. 1998). The results were measured by spectrophotometry at 760 nm and were expressed as g gallic acid equivalent (GAE)/100 g pine bark d.b or g GAE/100 g extract (Singleton and Rossi 1965).

2.4.3 Determination of condensed tannins yield

The vanillin method was used for the determination of condensed tannins (Sun et al. 1998). To 0.5 mL of extract was added 3.0 mL of vanillin solution in methanol (4% m/v) and 1.5 mL of concentrated hydrochloric acid (37%). Absorbance was measured at 500 nm in a spectrophotometer. The calibration curve was prepared with catechin (0–200 mg/L) and the results were expressed as g catechin equivalent (CE)/100 g pine bark d.b.

2.4.4 Antioxidant activity: FRAP method and ABTS assay

The antioxidant activity was determined by two methods: (1) FRAP method and (2) ABTS assay. These methods were selected since they are simple, precise, inexpensive and reproducible. The FRAP method measures the ability of the antioxidant to reduce the Fe³⁺–TPTZ complex to Fe²⁺–TPTZ at low pH. The obtained complex is colored and can be measured at 593 nm. The reducing power seems to be related to the degree of hydroxylation and the extension of the conjugation in the polyphenols. The ABTS method is based on the inhibition of the absorbance of the ABTS^·+ cation by antioxidants that show a main absorption peak at 415 nm and others with secondary absorption at 660, 734, and 820 nm (Singh and Singh 2008).

FRAP method was used to determine antioxidant power in extract as described by Xavier et al. (2017b). During the analytical procedure, FRAP reagent was prepared, by mixing a FeCl₃·6H₂O solution with water diluted acetate and TPTZ dissolved in HCl. FRAP reagent was then mixed with pine bark extract, producing a colored solution for further spectrometric analysis at 593 nm (Szegediensis et al. 2002). The results were expressed as mmol ascorbic acid equivalent (AAE)/100 g pine bark d.b.

ABTS aqueous solution (7 mM) was mixed with K₂S₂O₈ aqueous solution (2.45 mM). Such mixture induces the formation of radical ABTS^·+, which was then kept in the dark for 40 h at room temperature. After 40 h, the obtained solution containing radical ABTS was diluted with water until an absorbance of 0.7 was reached at 734 nm (Re et al. 1999). Then, 25 µL of sample was mixed with 2.5 mL of ABTS radical. Once the reagents were perfectly mixed, they were kept in total darkness for 6 min at room temperature. A blank was prepared with 25 µL of distilled water and 2.5 mL of ABTS reagent. Finally, absorbance was measured at 734 nm. The calibration curve was made using TRE (0.1–2 mM) as a standard. The extract concentration that inhibits 50% of the radical ABTS (EC₅₀) was determined as described by Piwowarska and González-Alvarez (2012). The results were expressed as mmol TRE/g pine bark d.b. and EC₅₀ as mg extract/mL.

2.4.5 UV–Vis spectroscopy analysis

UV–Vis spectrum of the extracts was carried out with a Shimadzu UVmini-1240 spectrophotometer. Extracts were re-dissolved in distilled water. Such water was also used as blank.

2.4.6 Infrared spectroscopy

FTIR spectrum of the extract was obtained with a Shimadzu IRAffinity-1S (Japan) spectrometer with 4 cm⁻¹ resolution. 32 scans were accumulated prior to making the Fourier transform. Three measurements were averaged to produce the obtained spectrum.

2.4.7 Adhesive extracts properties: Stiasny number determination

A quick way to assess the possibility of extracts to be used in the formulation of wood adhesives is using the Stiasny method. It is based on the reaction of flavonoid structures present in the tannins condensed with formaldehyde (Yazaki and Hillis 1980; Vázquez et al. 2009).

25 mL of pine bark extract were put to react with 5 mL of formaldehyde 38% and 2.5 mL of HCl 37% in a round bottom flask. This mixture was heated under reflux for 30 min (Yazaki and Hillis 1980). Then, the obtained solid was filtered with filter paper in a vacuum pump system. During this filtering, the solid was washed with distilled water until removing all formaldehyde. The washed solid was then dried at 105 °C until constant weight and the Stiasny number was determined (Eq. 2).

$$Stiasny \ Number \left( \% \right) = \frac{Dry \ mass \ of \ precipitate}{{Dry \ mass \ of \ extract}}100$$

(2)

The analytical results were reported as mean ± standard deviation (SD) of triplicate determinations.

2.5 Experimental design

An experimental design was implemented to evaluate how some of the selected independent variables (temperature, ethanol concentration and solid–liquid ratio) affect total extractives yield, phenolic compounds, condensed tannins yield and antioxidant capacity.

The experimental design was based on a three-factor and three level Box–Behnken design for fitting response surfaces. Extraction experiments were planned according to an incomplete factorial design, 3³ (Montgomery and Runger 2011). As shown in Table 1, the independent variables used were temperature, x₁, with levels 20, 40 and 60 °C; ethanol concentration, x₂, with levels 20%, 50% and 80% and solid–liquid ratio, x₃, with levels 1/5, 1/10 and 1/15. The dependent variables studied (Y_j) were total extractives yield, phenolic content, condensed tannins yield, and antioxidant activity measured in different ways: FRAP, ABTS and ABTS EC₅₀. This design is very efficient in terms of the number of runs required. The design consists of 12 experiments and three replicas at the central point to determine the experimental error (see Table 2) (Montgomery and Runger 2011).

Table 1 Experimental design in the original and coded form of the independent variables (x₁, temperature; x₂, ethanol concentration; x₃, solid–liquid ratio)

Table 2 Experimental design in the original form of the independent variables (x₁, temperature; x₂, ethanol concentration; x₃, solid–liquid ratio) and experimental results of total extractives yield, phenolic content, condensed tannins yield, FRAP, ABTS and ABTS EC₅₀

The experimental results were analyzed by polynomial regression applying the backward elimination method (Montgomery and Runger 2011) using Infostat software (version 2015, Grupo InfoStat, FCA, Córdoba, Argentina) and fitting data to the following equation:

$${\varvec{Y}}_{{\varvec{j}}} = {\varvec{a}}_{{0{\varvec{j}}}} + {\varvec{a}}_{{1{\varvec{j}}}} {\varvec{x}}_{1}^{\user2{*}} + {\varvec{a}}_{{2{\varvec{j}}}} {\varvec{x}}_{2}^{\user2{*}} + {\varvec{a}}_{{3{\varvec{j}}}} {\varvec{x}}_{3}^{\user2{*}} + {\varvec{a}}_{{12{\varvec{j}}}} {\varvec{x}}_{1}^{\user2{*}} {\varvec{x}}_{2}^{\user2{*}} + {\varvec{a}}_{{13{\varvec{j}}}} {\varvec{x}}_{1}^{\user2{*}} {\varvec{x}}_{3}^{\user2{*}} + {\varvec{a}}_{{23{\varvec{j}}}} {\varvec{x}}_{2}^{\user2{*}} {\varvec{x}}_{3}^{\user2{*}} + {\varvec{a}}_{{11{\varvec{j}}}} {\varvec{x}}_{1}^{{\user2{*}2}} + {\varvec{a}}_{{22{\varvec{j}}}} {\varvec{x}}_{2}^{{\user2{*}2}} + {\varvec{a}}_{{33{\varvec{j}}}} {\varvec{x}}_{3}^{{\user2{*}2}}$$

(3)

where $x_{1}^{*} ,x_{2}^{*} ,x_{3}^{*}$ are the independent variables coded in Table 1. Coding is done using the following equation:

$$x_{i}^{*} = \frac{{x_{i} - x_{M} }}{{\Delta x_{i} }}$$

(4)

being x_i the real value of the independent variable, x_M refers to the independent variable at the central point and Δx_i refers to the variation interval of the independent variable. The regression coefficients include the constant term $a_{0j}$, the linear terms $a_{1j}$, $a_{2j}$, $a_{3j}$, the interactions terms $a_{12j}$, $a_{13j}$, $a_{23j}$ and the quadratics terms $a_{11j}$, $a_{22j}$, $a_{33j}$.

Statistical study was carried out by making an ANOVA variance analysis, setting a confidence level of 95%. Furthermore, experimental values of each dependent variable Y_j were compared with the values obtained using the selected model. Standard deviation (σ) and the corresponding multiple determination coefficient (R²) were determined.

2.6 Extraction kinetic

To determine the extraction kinetics of phenolic compounds and extractable compounds, the experiments were carried out as detailed in Sect. 2.3 at different contact times (1, 2, 3, 4, 5, 10, 15, 20, 35, 60, 90 and 120 min). An ethanol concentration of 50% (w/w) with a solid–liquid ratio of 1/10, 200 rpm and a temperature of 50 °C was set as extraction conditions for being the best ones found in reported experimental design. To determine the dependence of extraction kinetics on temperature, a set of experiments at 20, 50 and 80 °C was carried out. Total extractives yield and phenolic content were determined in the liquid extract.

2.7 Kinetic models

To model the extraction process, the external mass transfer resistance was assumed to be negligible. Therefore, solid diffusion is the mechanism that determines the extraction kinetics, thus Fick’s second law was applied to model the process (Bird et al. 2006; Chan et al. 2014), assuming diffusion in spherical geometry with one-directional mass flow in radial direction and constant effective diffusion coefficient ($D_{eff}$):

$$\frac{\partial C}{{\partial t}} = D_{eff} \frac{{\partial^{2} C}}{{\partial r^{2} }}$$

(5)

where C is the extractable solute concentration in solid phase at time t and r is the radius of the spherical particle. The analytical solution of the differential equation considering spherical particles of radius R, solutes initially uniformly distributed in the solid and diffusion of the solutes with no interaction between them, obtained by Crank (1975) is:

$$\frac{{C - C_{e} }}{{C_{o} - C_{e} }} = \frac{{c_{\infty } - c}}{{c_{\infty } }} = \frac{6}{{\pi^{2} }}\mathop \sum \limits_{n = 1}^{\infty } \frac{1}{{n^{2} }}exp\left\{ { - \frac{{D_{eff} n^{2} \pi^{2} t}}{{R^{2} }}} \right\}$$

(6)

where C_e is the equilibrium concentration of extractable solutes in solid, C_o is the initial concentration of extractable solutes in solid, c is the extractable solute concentration in solvent at time t and c_∞ is the extractable solute concentration in equilibrium conditions (Pinelo et al. 2006).

A simplification of this solution (Eq. 6) can be applied to obtain a simple model considering only the first term of the series, named in this work as “Simplified Fick’s law model”.

$$\frac{{c_{\infty } - c}}{{c_{\infty } }} = \frac{6}{{\pi^{2} }}exp\left\{ { - \frac{{D_{eff} \pi^{2} t}}{{R^{2} }}} \right\}$$

(7)

In this case, the equation can be easily expressed as a linear correlation to determine the effective diffusivity or the relation $\left( {\frac{{D_{eff} }}{{R^{2} }}} \right)$ as a model parameter. The following equation was obtained when Eq. (7) was linearized (Pinelo et al. 2006):

$$ln\left( {\frac{{c_{\infty } }}{{c_{\infty } - c}}} \right) = \ln \left( {\frac{{\pi^{2} }}{6}} \right) + \frac{{D_{eff} \pi^{2} t}}{{R^{2} }}$$

(8)

Another model was evaluated based on the Osburn and Katz (1944) theory that considers the presence of two parallel diffusion processes in the porous particle and proposed Eq. (9) to describe the extraction process. It is known as Fick’s law extended analytical solution and in this work it is called “Fick’s law extended model”. The existence of two stages in the extraction process is proposed: washing stage and diffusion stage (Tao et al. 2014; Wright 2015). At the beginning of the extraction, high rates of extraction are observed, as the concentration gradient is maximum. Then, the extraction gets slower during the diffusion stage until equilibrium condition is reached.

$$\frac{{c_{\infty } - c}}{{c_{\infty } }} = \frac{6}{{\pi^{2} }}\left[ {f_{1} exp\left\{ {\frac{{ - \pi^{2} D_{{eff_{1} }} t}}{{R^{2} }}} \right\} + f_{2} exp\left\{ {\frac{{ - \pi^{2} D_{{eff_{2} }} t}}{{R^{2} }}} \right\}} \right]$$

(9)

where $D_{{eff_{1} }}$ and $D_{{eff_{2} }}$ are the effective diffusion coefficients during the washing stage and the diffusion stage, respectively. Besides, f₁ and f₂ are the solute fractions that are extracted in the correspondent stages.

In order to determine the kinetic model parameters, f₁, f₂, $D_{{eff_{1} }}$ and $D_{{eff_{2} }}$, it is assumed that the first term of the expression is negligible after long periods of extraction time. In this way, it is possible to linearize the function, so $D_{{eff_{2} }}$ and f₂ values can be determined (Hojnik et al. 2008; Chan et al. 2014; Wright 2015). By imposing an additional condition, knowing that at the beginning of the extraction process solute concentration in the liquid phase is null, f₁ is obtained (Wright 2015).

$$f_{1} + f_{2} = \frac{{\pi^{2} }}{6}$$

(10)

Once f₁, f₂ and $D_{{eff_{2} }}$ are determined, and considering no significant variation in the second term of Eq. (9), it is possible to determine $D_{{eff_{1} }}$ value, with the slope of Eq. (11):

$$\ln \left( {\frac{{c_{\infty } - c}}{{c_{\infty } }}} \right)\frac{{\pi^{2} }}{6} - f_{2} = \ln \left( {f_{1} } \right) - \pi^{2} \frac{{D_{{eff_{1} }} }}{{R^{2} }}t$$

(11)

Diffusion was assumed in spherical geometry, however, if the non-uniformity of particles geometry is considered, the model parameters determined will be f₁, f₂, $\frac{{D_{{eff_{1} }} }}{{R^{2} }}$ and $\frac{{{\text{D}}_{{{\text{eff}}_{2} }} }}{{{\text{R}}^{2} { }}}$.

To compare the different models and to evaluate the best fit to the experimental data, the standard deviation (σ) and the absolute average relative deviation percentage (AARD%) were determined.

$$\sigma = \sqrt {\frac{{\mathop \sum \nolimits_{i = 1}^{n} \left( {c_{exp} - c_{calc} } \right)^{2} }}{n}}$$

(12)

$$AARD\% = \frac{100}{n}*\mathop \sum \limits_{1}^{n} \left( {\frac{{c_{calc} - c_{exp} }}{{c_{exp} }}} \right)$$

(13)

where c_exp and c_calc are the experimental concentration data and the calculated concentration value by each model, respectively, and n is the total experimental points.

3 Results and discussion

3.1 Characterization of native pine bark

The chemical composition of the studied raw material is shown in Table 3. Bark’s components are similar to other wood types. However, chemical structure and composition are quite different from wood. Bark generally contains more extractives, ash and lignin. Many of the mentioned extractives are phenolic compounds, which can present antioxidant properties among others (Frankó et al. 2017). If we compare the values for the pine bark reported by Pan et al. (2013) with those obtained in this work (Table 3), it was observed that the contents of glucose (23.1%) and ash (1.1%) were similar, while the lignin content was higher (43.5%). The ashes represent 1.5% of the biomass, and are associated with soluble salts of the bark (Bianchi et al. 2015). The potassium content was 1774 ± 145 ppm. The results are consistent with values reported by others (Pan et al. 2013; Liu et al. 2018). The pine bark has total extractives of 14.5%, the majority of which are polar and can be removed with water and ethanol (12.7%). The lipophilic compounds removed with hexane represent only 1.8%. Miranda et al. (2017) report that 91% of pine bark extractives are extracted with polar solvents. It has been shown that the lipophilic fraction content of bark ranges between 0.7% and 4%, independently of the pine species (Nunes et al. 1999; Hafizoǧlu et al. 2002; Miranda et al. 2017; Masendra et al. 2018; Sousa et al. 2018). Comparing the values of extractives content removed with water and ethanol with other pine species particularly in the bark, it is observed that they are lower than the values found in the literature for Pinus pinea (18.8%) (Miranda et al. 2017) and Picea abies (20.3%) (Neiva et al. 2018), and higher than those found for Pinus pinaster (9.1%) (Nunes et al. 1996). Similar values for total extractives were also found by Pan et al. (2013), who obtained values of 13.2% for Pinus taeda using Soxhlet extraction with methanol–water as extraction solvent.

Table 3 Composition of native pine bark

Comparing the lignin content with different conifers, we see that the values are higher: for example, Picea abies (27%), (Neiva et al. 2018); Pinus pinaster (33.2%) (Fradinho et al. 2002) and similar to those of Pinus pinea (42.2%) (Miranda et al. 2017). The hemicelluloses that were detected in the chemical characterization of the bark were mannose, galactose and xylose, while the content of arabinose was negligible. When the sum of these three sugars is compared with the reported hemicellulose content for pine bark by Pan et al. (2013) (14.1%), it is observed to be significantly less. As can be seen in Table 3, the total carbohydrate content is 27.6%. This value is lower than that of other pine species: for example, Pinus sylvestris (37.6%) (Miranda et al. 2012); Pinus pinaster (48.4%) (Fradinho et al. 2002); Pinus pinea (37 and 36.8%) (Nunes et al. 1999; Miranda et al. 2017). This could be favorable for obtaining products with higher purity, due to the lower presence of sugars in the extract (Pan et al. 2013).

Furthermore, this study included the determination of the higher heating value of pine bark, which was 21.49 ± 0.32 kJ/g d.b. The calorific value of lignocellulosic materials is proportional to the content of lignin and extractives (polar and nonpolar). In particular, it is considered that the calorific value of softwood has a higher calorific value than hardwoods because it contains resin and higher levels of extractives (White 1987). As already mentioned, the bark has a high content of lignin, which contributes to the higher heating value. The calorific values of pine and oak bark were reported at 18.3 and 19 kJ/g d.b, respectively (Ingram et al. 2008).

3.2 Effects of extraction conditions: response surfaces

The effects of extraction conditions (temperature, ethanol concentration and solid–liquid ratio) on total extractives yield, phenolic content, condensed tannins yield and antioxidant activity were analyzed. The experimental conditions and results of the experimental designs are shown in Table 2.

These results were adjusted to a quadratic correlation (Eq. 3). Table 4 shows the regression coefficients of each dependent variable obtained based on the independent codified variables $x_{1}^{*}$ (temperature), $x_{2}^{*}$ (ethanol concentration), $x_{3}^{*}$ (solid–liquid ratio). The coefficients of multiple determination (R²) for the regression models for total extractives yield, phenolic compounds, condensed tannins yield, FRAP, ABTS and ABTS EC₅₀ which were 0.93, 0.81, 0.75, 0.93, 0.98, 0.93, respectively; representing a generally acceptable fitting of the model (see Table 4). A value of R² > 0.75 indicates acceptance of the model (Le Man et al. 2010).

Table 4 Regression coefficients of polynomial models (Eq. 3)

The mathematical models that describe the dependent variables (total extractives yield, phenolic content, condensed tannins yield, FRAP, ABTS and ABTS EC₅₀) as a function of temperature ($x_{1}^{*}$), ethanol concentration ($x_{2}^{*}$) and solid–liquid ratio ($x_{3}^{*}$) were calculated only considering the significant correlation coefficients.

$$Total\,extractives\, yield = 9.91 + 1.21x_{1}^{*} + 1.01x_{2}^{*} - 2.13x_{2}^{*2}$$

(14)

$$Phenolic\, compounds = 6.02 + 1.1x_{2}^{*} - 1.8x_{2}^{*2}$$

(15)

$$Condensed\, tannins\, yield = 4.68 - 0.40x_{1}^{*2} - 0.86x_{2}^{*2} - 0.46x_{3}^{*2}$$

(16)

$$FRAP = 20.21 + 4.19x_{1}^{*} + 2.01x_{3}^{*} + 2.99x_{1}^{*} x_{3}^{*} - 2.37x_{1}^{*2} - 2.92x_{2}^{*2} - 2.45x_{3}^{*2}$$

(17)

$$ABTS = 0.53 + 0.03x_{2}^{*} - 0.08x_{3}^{*} - 0.04x_{2}^{*} x_{3}^{*} - 0.19x_{1}^{*2} - 0.23x_{2}^{*2} - 0.10x_{3}^{*2}$$

(18)

$$ABTS EC_{50} = 0.28 + 0.26x_{3}^{*} + 0.45x_{1}^{*2} + 0.36x_{2}^{*2}$$

(19)

To study the effects that extraction conditions have on total extractives yield, phenolic content, condensed tannins yield and antioxidant activity, surface responses were determined as shown in Figs. 1, 2, 3, 4, 5 and 6.

Fig. 1

Response surfaces for total extractives yield (g extract/100 g pine bark d.b.) as a function of ethanol concentration and temperature

Fig. 2

Response surface for phenolic content (g GAE/100 g pine bark d.b.) as a function of ethanol concentration

Fig. 3

Response surfaces for condensed tannins yield (g CE/100 g pine bark d.b.) as a function of ethanol concentration and solid–liquid ratio for an extraction temperature of a 20 °C, b 40 °C, c 60 °C

Fig. 4

Response surfaces for antioxidant activity FRAP (mmol AAE/100 g pine bark d.b.) as a function of ethanol concentration and solid–liquid ratio for an extraction temperature of a 20 °C, b 40 °C, c 60 °C

Fig. 5

Response surfaces for antioxidant activity ABTS (mmol TRE/g pine bark d.b.) as a function of ethanol concentration and solid–liquid ratio for an extraction temperature of a 20 °C, b 40 °C, c 60 °C

Fig. 6

Response surface for antioxidant activity ABTS EC₅₀ (mg extract/mL) as a function of ethanol concentration and solid–liquid ratio for an extraction temperature of a 20 °C y 60 °C, b 40 °C

A parameter that affects the extraction of phenolic compounds is solvent polarity. This is due to the existence of a great variety of phenolic compounds with different polarities. Therefore, by varying the solvent, the extracted phenolic compounds and its antioxidant capacity also change (Thoo et al. 2010). By analyzing the regression coefficients (Table 4), it was found that the quadratic effect of ethanol concentration ($x_{2}^{*2}$) is significant in all the evaluated responses (p < 0.05). Results also show that this term is negative for all responses except in ABTS EC₅₀. This means that dependent variables first increase and then decrease as ethanol concentration increases. On the other hand, ABTS EC₅₀ has a minimum at an ethanol concentration of 50%. ABTS EC₅₀ is the extract concentration needed to reduce the free radicals present in the system, meaning that low values of such variable is related to high antioxidant activity (Radojković et al. 2012). The linear effect $\left( {x_{2}^{*} } \right)$ has an increased response with ethanol concentration for total extractives yield, phenolic content and ABTS antioxidant activity. Interaction of ethanol concentration with temperature ( $x_{1}^{*} x_{2}^{*}$) is not significant (p > 0.05) in none of the dependent variables.

Figures 1, 2, 3, 4, 5 and 6 show that maxima obtained for total extractives yield, phenolic content, tannins yield, FRAP and ABTS antioxidant capacity were found at ethanol concentrations of 58, 59, 50, 50, and 53%, respectively. The optimums of the different dependent variables were at ethanol concentrations between 50 and 59%. These results are similar to the ones found in previous researches (Thoo et al. 2010; Derkyi et al. 2011; Kaiser et al. 2013).

Regarding the temperature effect on the response variables, we see in Table 4 that the quadratic effect of temperature ($x_{1}^{*2}$) is rather significant for condensed tannins yield and in the different antioxidant activities: FRAP, ABTS and ABTS EC_50. The linear effect ($x_{1}^{*}$) is also significant when related to total extractives yield and FRAP. This performance can be seen in the similar responses that were obtained for condensed tannins yield (Fig. 3), ABTS (Fig. 5) and ABTS EC₅₀ (Fig. 6). As temperature increases, responses also increase until reaching an optimum value, after which responses start to decrease. Figures 3 and 5 show that the maximum condensed tannins yield and antioxidant activity ABTS are reached at 40 °C. In the case of ABTS EC₅₀ (Fig. 6), the minimum is the one obtained at 40 °C. Moreover, maximum total extractives yield (Fig. 1) and FRAP (Fig. 4) antioxidant capacities are obtained at 60 °C. By increasing the temperature, total extractives yield and FRAP antioxidant capacity increase by $x_{1}^{*}$. The increased yield of these variables with temperature can be explained due to the fact that the solvents viscosity decreases with temperature, making the solid–liquid interaction easier and increasing the solubility of the retained components. Compounds solubility also increases (Derkyi et al. 2011; Tham and Liew 2012; Wright 2015). However, not all phenolic compounds are stable at high temperatures as they can be decomposed. On the other hand, temperature was found not to be significant (p > 0.05) in the phenolic content (Fig. 2). With a confidence level of 95% the temperature effect is not observable, this may be because in the range of the selected independent variables, the most significant impact is observed with the ethanol concentration.

Solid–liquid ratio presents a negative quadratic effect ($x_{3}^{*2}$) in condensed tannins yield, FRAP and ABTS (Table 4). As solid–liquid ratio gets higher values, extraction increases due to higher concentration gradients, meaning mass transfer potential also increases, making extraction yields higher. On the other hand, the solid–liquid relationship was found not to be significant for the total extractives yield and phenolic content. Interaction of solid–liquid and temperature, ($x_{1}^{*} x_{3}^{*}$) has positive effects on the antioxidant activity FRAP. On the other hand, solid–liquid ratio and solvent concentration interaction $(x_{2}^{*} x_{3}^{*}$) have negative effects on the antioxidant activity ABTS. As solid–liquid ratio increases, ABTS EC₅₀ also does, as the extract is more diluted in this case. This fact is reflected by a positive linear effect ($x_{3}^{*}$) in the model.

3.3 Extraction conditions selection and verification of the model

The maximum values for each response variable were determined from response surface figures. Figure 1 shows that the maximum total extractives yield (11.2 g extract/100 g pine bark d.b.) was obtained at the highest tested temperature (60 °C) at an ethanol concentration of 57.5%.Venkatesan et al. (2020) reported yields of 10.99 and 16.86 g extract/100 g biomass d.b. in Pinus densiflora needle and bark, using ethanol (40%) as solvent. Shakya et al. (2016) reported Soxhlet total extractives yield of Acacia nilotica of 8.16 g extract/100 g biomass d.b. using ethanol and 8.05 g extract/100 g biomass d.b using water as a solvent. Extraction yields found for Pinus densiflora bark are similar to those obtained for Pinus taeda. By using water and ethanol as solvents, yields were reported in the range of 8.34–11.76 g extract/100 g biomass d.b. (depending on the ethanol concentration). Nowadays, food supplements and other commercial products made up of Pinus densiflora extract are available in the market (Venkatesan et al. 2019). Luís et al. (2012) reported extraction yields of Acacia dealbata bark, with values in the range of 3.26–9.51 g extract/100 g biomass d.b (depending on the solvent used). These values are similar to the ones reported in the present article. Miranda et al. (2016) reported extraction yields significantly higher (50 g extract/100 g biomass d.b.) for Eucalyptus sideroxylon, also using water–ethanol as extraction solvent.

Figure 2 shows that the maximum of phenolic content (6.02 g GAE/100 g bark d.b) predicted by the model was reached at an ethanol concentration of 59%. Rosales-Castro et al. (2015) reported a result of 8.2 g GAE/100 g biomass d.b. using ethanol as solvent and 5.1 g GAE/100 g biomass d.b. with hot water out of pink cedar bark. These results agree with values reported by González et al. (2017), who obtained 4.8 g GAE/100 g biomass d.b. with methanol 50% out of Eucalyptus nitens.

In terms of maximum condensed tannins yield, Fig. 3 shows that the maximum value (4.7 g CE/100 g pine bark d.b.) was obtained at 40 °C, with a solid–liquid ratio of 1/10 and an ethanol concentration of 50%. Bocalandro et al. (2012) reported a lower tannins yield than those reported in this work (1.6 CE/100 g Pinus radiata bark d.b. in ethanol solutions at 120 °C for 120 min at a solid–liquid ratio of 1/20). Condensed tannins can be found by its UV/Vis spectra. Condensed tannins have only one absorption maximum between 260 and 270 nm. Pine bark extracts presented only one absorption maximum at 270 nm, which is characteristic of proantocianidins (Vázquez et al. 2008).

The maximum FRAP antioxidant activity (24 mmol AAE/100 g pine bark d.b.) was obtained at 60 °C with a solid–liquid ratio of 1/15 and an ethanol concentration of 50% (Fig. 4). Such antioxidant activity was higher than the ones reported by Piwowarska and González-Alvarez (2012), who obtained values of 12.3 mmol AAE/100 g d.b. at 75 °C from forestry biomass, using methanol 50% as extraction solvent and solid–liquid ratio of 1/10.

Figure 5 shows that the maximum value found for ABTS is 0.55 mmol TRE/g pine bark d.b. with a solid–liquid ratio of 1/8, 40 °C and a solvent concentration of 53%, while Fig. 6 shows that for ABTS EC₅₀, the minimum (0.02 mg/mL) is obtained at a solid–liquid ratio of 1/5, 40 °C and a solvent concentration of 50%.

The ABTS antioxidant activity of Pinus densiflora bark extract, using ethanol (40%) as extraction medium, was 0.27 mmol TRE/g biomass (d.b.), which is about half of the results obtained in this work (Venkatesan et al. 2020).

As expected, the extraction conditions did not maximize the different variables responses concurrently, and therefore the optimum conditions were selected within the ranges obtained for all the independent variables. The response surfaces showed that the extraction conditions that maximize extraction yields as well as antioxidant activity were at 50 °C, solid–liquid ratio of 1/10 and an ethanol concentration of 50%. Table 5 shows experimental and model values at the selected extraction conditions and the corresponding standard deviation. This model shows to be acceptable for predicting the extraction variables responses of pine bark.

Table 5 Experimental and predicted values by the model at T: 50 °C; S/L: 1/10; ethanol concentration: 50%

At the selected extraction conditions, the total phenols expressed as g GAE/100 g extract reached the value 50.7. This value is similar to those reported for Maritime pine bark (57.2 g GAE/100 g extract). These extracts are commercialized for pharmaceutical formulations with the name Pycnogenol^® (Ustun et al. 2012). Pinus densiflora bark extracts (which are also available in the market) reported values in the range of 14.6 and 25.7 g GAE/100 g extract, presenting different values depending on the extraction solvent used (Venkatesan et al. 2019). Miranda et al. (2016) reported lower extraction values per gram of extract (44.1 g GAE/100 g extract, but higher values per gram of dry pine bark, due to its high extraction yields (22.1 g GAE/100 g biomass d.b.).

The removal of the resinous matter with hexane prior to extraction with water and ethanol was carried out with the objective of evaluating the impact on total extractives yield and phenolic content. After successive extractions, the total extractives yield was 6.52 g extract/100 g pine bark d.b and the phenolic content was 5.01 g GAE/100 g bark d.b. If we compare these results with the values obtained without the prior extraction with hexane (Table 5), the extraction of pine bark with non-polar solvents prior to the extraction with water–ethanol decreased the yield of the total extractives and the extraction of phenolic compounds remained unchanged. Therefore, no further improvement in the phenolic compounds yield was found by running a previous extraction with hexane. Furthermore, the extraction performance decreases, which could be attributed to the low resin content of pine bark (1.8%) in comparison with other pine species, such as Pinus nigra (3.3%) or Pinus sylvestris (4.2%) (Hafizoǧlu et al. 2002; Miranda et al. 2012).

3.4 FTIR spectra

Figure 7 shows the FTIR spectra of the extract, extracted 90 min at an ethanol concentration of 50% (w/w) with a solid–liquid ratio of 1/10, 200 rpm and a temperature of 50 °C. There are some characteristic bands which can be associated with some specific organic functional groups. Firstly, the extended band in the range of 3500–3100 cm⁻¹ indicates the presence of OH phenolic groups. Between 2950 and 2880 cm⁻¹, some low intensity peaks appear, indicating C–H methoxil group. Between 1600 and 1500 cm⁻¹, characteristic vibrations related to aromatic structures appear. The spectrum shows C–C bonds related to phenolic compounds, which appear at 1500–1400 cm⁻¹. Around 1200 cm⁻¹, another middle intensity peak appears, which is a strong clue of polyphenol C–O group in its structure (Santos et al. 2018). The spectrum shows a high intensity narrow peak with a maximum of absorbance around wavelength 1100 cm⁻¹, indicating the presence of an aromatic-CH bending. Lastly, around wavelength 750 cm⁻¹, a slightly extended peak appears, being a clue for the presence of C=C in aromatic structures. This spectrum would allow a primary identification of the phenolic compounds extracted, comparing the spectrum obtained with the literature (Vázquez et al. 2008; Chupin et al. 2013).

Fig. 7

FTIR spectra of pine bark extract

3.5 Pine bark energetic characterization after extraction

In optimum conditions, higher heating value was measured before and after extraction. Higher heating value before extraction reached a value of 21.49 ± 0.32 kJ/g d.b, while after extraction, the measured value was 21.30 ± 0.12 kJ/g d.b. The difference was found to be negligible. On the other hand, there is approximately 10% mass loss because of the extraction and therefore, the potential energy output of the extracted bark as fuel is reduced. This aspect should be taken into consideration for the scale up of the operation. Little variation in the calorific value before and after extraction would imply a greater dependence on lignin content than on the concentration of extractives, supported by the higher content of lignin in the bark of Pinus taeda. Therefore, as expected, extraction does not affect the calorific value of bark (White 1987). After extraction, the ashes from the extracted bark were reduced by 13%. Potassium was reduced by 48%. This reduction in potassium is highly beneficial, because it generates corrosion problems (Krogell et al. 2012) and reduces the melting point of ashes. Therefore, the extraction would contribute to improve the use of this bark as fuel.

3.6 Extraction kinetics and model: total extractives yield and phenolic content

The evolution of the kinetic curves of total extractives yield and phenolic content over time at different temperatures is shown in Figs. 8 and 9, respectively. The kinetics of both total extractives yield and phenolic content behave similarly. At first, the extraction rate increases significantly during the first 4–5 min of extraction, decreasing its extraction rate from minute 5–60, until equilibrium concentration is reached. This behavior has already been observed for different lignocellulosic materials (Wright 2015; González et al. 2017).

Fig. 8

Extraction curves of extractives at different temperatures: 20 °C (filled circle), 50 °C (filled square) and 80 °C (filled triangle). Symbols represent experimental data and solid lines the Fick’s law extended model

Fig. 9

Extraction curves of phenolic compounds at different temperatures: 20 °C (filled circle), 50 °C (filled square) and 80 °C (filled triangle). Symbols represent experimental data and solid lines the Fick’s law extended model

Concentrations of extractives increase as temperatures get higher, reaching maxima of 5.9 g extract/100 g pine bark d.b. at 20 °C, 7.7 g extract/100 g pine bark d.b. at 50 °C and 8.4 g extract/100 g pine bark d.b. at 80 °C in equilibrium conditions. As seen in extractives extraction kinetic study, concentration of phenolic compounds in the liquid phase is higher as extraction temperature increases, obtaining a maximum of 3.55 g GAE/100 g pine bark d.b. at 20 °C, 4.43 g GAE/100 g pine bark d.b. at 50 °C and 5.99 g GAE/100 g pine bark d.b. at 80 °C reached in equilibrium conditions. As discussed in Sect. 3.2, these results are comparable to others reported in the literature. The experimental design included extraction experiences in the range of 20–60 °C. In this case, the temperature's effect is not observable in phenolic content. Extraction kinetics’ temperatures varied in a wider range (20–80 °C) at a constant ethanol concentration. In this case, the results showed an increase in phenolic content when reaching equilibrium conditions.

The multi-stage extraction was carried out in order to determine the total amount of phenolic compounds contained in pine bark, which resulted in 9.5 ± 0.4 g GAE/100 g pine bark d.b.. This value was obtained after n extraction stages. Phenolic content obtained in equilibrium condition at 80 °C represents 63% of the total amount of phenolic compounds extracted. Therefore, it can be seen that the highest percentage of extraction is obtained in the first stage.

Two models were used: “Simplified Fick’s law model” and “Fick’s law extended model” to model the extraction kinetics. Table 6 shows the fitting parameters of both models with their respective σ and AARD%. The Fick’s law extended model fitted better than the Simplified Fick’s law model. For each model σ and AARD% values were lower for the Fick’s law extended model than for the Simplified Fick’s law model, meaning that a better fitting was reached for such case. The same model was suitable for the extraction of extractives from Pinus contorta bark (Wright 2015). The Fick’s law extended model and the experimental data of total extractives yield and phenolic content are plotted in Figs. 8 and 9, respectively.

Table 6 Models parameters, AARD% and σ for the extractives and phenolic compounds at different temperatures

3.6.1 Model parameters analysis: extraction of extractives

As shown in Table 6, the parameters, $\frac{{D_{eff} }}{{R^{2} }},$ obtained for the diffusion of extractives according to the Simplified Fick’s law model at different temperatures were in the order of 10E−5 (s⁻¹). These parameter values were similar to those obtained from Pinus contorta bark (3.77E−5 s⁻¹ for 60 °C, 3.48E−5 s⁻¹ for 80 °C and 13.16E−5 s⁻¹ for 100 °C) (Wright 2015).

In the Fick’s law extended model, $\frac{{D_{{eff_{1} }} }}{{R^{2} }}$ and $\frac{{D_{{eff_{2} }} }}{{R^{2} }}$ are also similar to those obtained from Pinus contorta bark ($\frac{{D_{{eff_{1} }} }}{{R^{2} }}$: 15.11E−4 s⁻¹ for 60 °C, 8.88E−4 s⁻¹ for 80 °C and 5.70E−4 s⁻¹ for 100 °C and $\frac{{D_{{eff_{2} }} }}{{R^{2} }}$: 2.88–5 s⁻¹ for 60 °C, 2.47E−5 s⁻¹ for 80 °C and 5.70E−5 s⁻¹ for 100 °C) (Wright 2015). The reaction rate is higher at the beginning of the extraction for all temperatures, represented by a steep slope at the start of the extraction (Fig. 8). Furthermore, the parameter values in the washing stage, $\frac{{D_{{eff_{1} }} }}{{R^{2} }}$ for all temperatures were greater than in the diffusive stage, $\frac{{D_{{eff_{2} }} }}{{R^{2} }}$ (Hojnik et al. 2008). Regarding the parameter values, no matter the experiment, nor the extraction stage, no trends with temperature were observed. This lack of tendency was also reported by Čretnik et al. (2005), Hojnik et al. (2008) and Wright (2015).Thus, no way to predict the diffusivity variation with temperature was tested.

3.6.2 Model parameters analysis: extraction of phenolic compounds

Table 6 shows that the parameters, $\frac{{D_{eff} }}{{R^{2} }},$ determined for the extraction of phenolic compounds according to the Simplified Fick’s law model at different temperatures were also in the order of 10-5.

The obtained parameter values of Simplified Fick’s law model for the extraction of phenolic compounds and its tendency were in accordance with those obtained by González et al. (2017), who worked with Eucalyptus nitens samples using methanol 50% (w/w) as extraction solvent, with a solid–liquid ratio of 1/60 (w/v) (8.82E−5 s⁻¹ for $\frac{{D_{eff} }}{{R^{2} }},$ at 53 °C). This model did not have a satisfactory fitting at the beginning of the extraction time.

The parameters for the Fick’s law extended model determined were similar to those obtained by González et al. (2017) with Eucalyptus globulus (2.07E−4 s⁻¹ for ${ }\frac{{D_{{eff_{1} }} }}{{R^{2} }}{ }$ and 7.63E−5 s⁻¹ for $\frac{{D_{{eff_{2} }} }}{{R^{2} }}$ at 46 °C). Again, the model parameters in the washing stage,$\frac{{D_{{eff_{1} }} }}{{R^{2} }}$, were higher than in the diffusive stage, $\frac{{D_{{eff_{2} }} }}{{R^{2} }}$. The first stage of extraction is assumed to be faster than the second stage because the surface layers of the crust may have suffered mechanical damage when ground. Thus, phenolic compounds are easily washed from the surface layer. The following solutes have to diffuse through the intact tissue within the particles. Therefore, slow diffusion begins to control the rate of extraction (Hojnik et al. 2008).

From the model’s parameters determined and taking into account the range of particle size (297–1000 µm), the diffusion coefficients can be calculated. The diffusion coefficients of extractives and phenolic compounds were between 10E−13 and 10E−11 m² s⁻¹, a similar range of values was obtained in a similar process by other researchers. Furthermore, these values are in the range of magnitude order of diffusion in solids (Gironi and Piemonte 2011; Tsibranska et al. 2011; Chan et al. 2014).

3.7 Adhesive extracts properties: Stiasny number

The potential adhesive property of the extract was evaluated from the Stiasny number. According to Tahir et al. (2002), a Stiasny number of 65 is the minimum to produce high-quality adhesives. The Stiasny number for the pine bark was 36.3, 77.6 and 82.0 at extraction temperature of 20, 50 and 80 °C, respectively. As observed, at 50 and 80 °C, the Stiasny number results were higher than the minimum required to produce high-quality adhesives. These results are comparable with those obtained by Vázquez et al. (2009) for chestnut shell.

The Stiasny number increase from 20 to 50 °C was higher than its variation from 50 to 80 °C. One possible explanation is that there is a higher fraction of condensed tannins within the phenolic compounds due to an increase in solubility between 20 and 50 °C.

4 Conclusion

The extraction of phenolic compounds from pine bark was encouraging, since acceptable extraction yields were observed. Satisfactory regression models for total extractives yield, phenolic content, tannins yield and antioxidant activity were found in the range of the independent variables assayed. An optimal work point was determined by evaluating the antioxidant capacity and the amount of compounds obtained. Considering the application of the extract as antioxidant, the results gave evidence that it could be used for that application.

The kinetic model parameters for extraction of extractives and phenolic compounds from pine bark were represented in a theoretical extraction model that allows the prediction of the extraction yield under different conditions. Fick’s law extended model showed to be the best one to fit experimental data. This is the first need to scale up the process in an industrial operation.

In addition, the extract could be used for adhesive formulations since the results showed high Stiasny number and acceptable extraction yield value of phenolic compounds. Future research lines such as the study of formaldehyde-free adhesive formulation can be carried out. Finally, the higher heating value obtained after the extraction of phenolic compounds supports the idea of valorizing the bark in two stages: (1) extraction of extractives and (2) using the extracted bark as fuel. The decrease in inorganic content and particularly the potassium will contribute to limiting the concentration of metals that causes the melting point reduction of ash, and also to limit corrosion problems. The economic evaluation of the entire process scaled at different yields will be the subject of further research.