Effects of tree harvesting time and tannin cold/hot-water extraction procedures on the performance of spruce tannin biocoagulant for water treatment

This study investigated the effect of seasonal variation (i.e. winter and summer) on tannin extracts from spruce bark as biocoagulants for water treatment applications. Tannins were extracted through three different water extraction procedures: cold-water extraction (21 ◦ C), cold-water plus hot-water extraction (85 ◦ C) and direct hot-water extraction (85 ◦ C). 1 H nuclear magnetic resonance spectra showed that the cold-water plus hot-water extractions possessed the highest proportion of phenols. Electrospray ionization-mass spectrometry indicated that a preliminary cold-water extraction sufficiently reduced the non-phenolic constituents (monosaccharides and sugars) from the winter bark in the cold-water plus hot-water extraction but not from the summer bark. The synthesis of tannin-based coagulants was performed through the Mannich reaction using formaldehyde and ethanolamine. Coagulants with the highest charge densities (3.379 ± 0.012 meq/g) were produced with tannin extracts obtained from the winter bark cold-water plus hot-water extractions. Water treatment experiments with the coagulants from the winter and summer bark cold-water plus hot-water extractions demonstrated that they were effective for particle settling. The study proved that the tree harvesting season and extraction procedure play a critical role in obtaining high-quality spruce tannin for the synthesization of tannin-based biocoagulants that provide better coagulative performance.


Introduction
One of the most underutilized side streams of the forest industry is tree bark.About seven million m 3 of tree bark produced as side streams from various Finnish forest industries was used for energy production in 2020 [1].Unfortunately, bark has a relatively low heating value and contains valuable chemicals such as tannin and other phenolic components [2].Tannin can be extracted from bark through a water extraction process and is continuously being investigated for different industrial and commercial applications worldwide [3].More importantly, tannin can be used for the synthesis of tannin-based biocoagulants for water treatment applications.Tannin can be easily modified through aminomethylation also known as the Mannich reaction.During the Mannich reaction, an iminium ion is produced through the condensation of an aldehyde with an amino compound, which then substitutes hydrogen in the polyphenolic matrix of tannin [4].
Several studies have alluded to the coagulative efficacy of tanninbased coagulants in effluents with diverse characteristics [5].Some notable practical applications of tannin biocoagulants in water treatment include the sequestration of heavy metals from surface water [6], decolouration of dye effluents [7], turbidity reduction of industrial process waters [8] and removal of humic substances from peat extraction runoff waters [9].On the other hand, a significant difference in their coagulative performance due to variation in the chemical composition of tannins from various sources has also been widely reported [10][11][12].Arismendi et al. [12] aminomethylated Acacia mearnsii, Schinopsis balansae and Castanea sativa extracts and observed that the carboxyl or ester groups in Castanea sativa extracts tend to prohibit or slow down the reaction between the iminium ions and tannin phenols by deactivating the electrophilic characteristics of the aromatic ring in the tannin matrix.Furthermore, the presence of hydrolysable tannin constituents in Castanea sativa extract could limit active hydrogen sites for iminium ion replacement, leading to poor coagulative performance by the Mannich modified coagulant.Hydrolysable tannin has been reported to occupy ionizable sites in the complex polyphenolic tannin matrix [5].
Spruce is a common tree species in the Nordic countries and accounts for over 30 percent of forest resources in Finland [13].Several studies have confirmed that hot-water extraction is a viable method for the extraction of spruce tannin [10,[14][15][16], and tannin extracted with hot water from spruce bark has been proven suitable for synthesizing biocoagulants [8,10].However, the main limitation in the application of spruce bark as a tannin source for tannin-based coagulants is the low tannin yield and the presence of a significant amount of co-extracted impurities such as sugars, minerals and phenolic non-tannin compounds [10,17].The presence of impurities is not desirable during aminomethylation and could facilitate the production of competitive byproducts during the Mannich reaction [12].Several studies have established that mono-and disaccharide sugars are the dominant nonphenolic compounds in spruce bark water extracts [18][19][20].The percentages of these sugars in extracts are at their peak in October and are at a minimum in May or June [20,21].
In recent times, some research studies have targeted improving the quality and yield of tannin extract from spruce bark by optimizing the hot-water extraction procedure [17,20,22].Kemppainen [20] claimed an increase in tannin yield and a relative reduction in the amount of carbohydrates in hot-water extracts by using sodium bisulphite and sodium carbonate as extraction chemicals.Nonetheless, the additional extraction chemicals reduce the purity of the tannin and the sustainability of the extraction process.Ding et al. [17] and Bianchi et al. [22] recorded an increase in the purity of tannin from spruce bark by employing a sequential cold-and hot-water extraction method.Extracts with higher tannin purity were obtained during the subsequent hotwater extraction, as free mono-and oligosaccharides were effectively removed in the preliminary cold-water extraction (30 • C) [17].The current study also uses the cold-and hot-water extraction approach and examines how this extraction procedure affects the properties and the performance of biocoagulants.Although some studies have been published on optimizing the tannin extraction process [17,19,20,22], no study has been conducted to explore the correlation between tannin extraction optimization and the coagulative performance of tanninbased coagulants.In this study, tannins were extracted from the spruce bark of winter-and summer-harvested trees using the cold-water (CW), sequential cold-and hot-water (CHW) and hot-water (HW) extraction methods.The composition of the extracts was investigated by electrospray ionization-mass spectrometry (ESI-MS), proton nuclear magnetic resonance ( 1 H NMR) spectroscopy and X-ray photoelectron spectroscopy (XPS).The extracts were pulverized by freeze drying before being cationized through the Mannich reaction.The influence of seasonal variations and extraction methods on the properties and performance of the synthesized coagulant was studied through charge density measurements and water purification experiments.

Raw materials and chemicals
Two batches of winter and summer spruce bark were obtained directly after dry debarking at Stora Enso Oyj Veitsiluoto and Vuokila Wood Oy Haukipudas mills in northern Finland.The winter bark batches were sampled in February 2021 and 2022 from Stora Enso Oyj and Vuokila Wood Oy mills, respectively, while the summer batches were collected from both mills in September 2021.It is essential to mention that bark from the Vuokila Wood Oy mill was collected only to verify the consistency of condensed tannin in the extracts during the different harvesting seasons.Hence, most investigations were performed with the spruce bark from the Stora Enso mill.The dry matter content of the fresh spruce bark was approximately 50% and was determined by drying two samples of 10 g at 60 • C for 24 h.The average time between tree harvesting and debarking was 45-60 days and the bark samples used for this study were collected within a day of the debarking process.For the kaolin/river water mixture used in the coagulation experiment, kaolin was collected from Pihlajavaara in the Kainuu region of Finland, and the river water was taken from the River Oulu, also in Finland.The ethanolamine (ETH) used during the aminomethylation of tannin was supplied by Sigma Aldrich Chemicals, Germany.Formaldehyde (37% v/w) was purchased from VWR International, France, and HCl was manufactured by Merck KGaA, Germany.The NaOH used for pH adjustment was provided by VWR chemicals, while Milli-Q (Merck Millipore) was used throughout the studies except when indicated.The polydiallyldimethylammonium chloride (PolyDADMAC) and sodium polyethylene sulphonate (PesNa) used to measure the charge density of coagulants and surface charge of water sample was provided by BTG Instruments (Sweden).

Tannin extraction and pulverization
The bark samples were not subjected to further processing, and no manual separation was required as the bark from the mills was woodfree.Tannins were extracted within 48 h of delivery to avoid the negative effect of prolonged storage on the quality of the tannin extracts (immobilization and degradation).A flow diagram of the extraction procedures is shown in Fig. 1.CW extracts were obtained by mixing 1 kg of fresh spruce bark with 10 L of deionized water for 2 h at 21 • C to give a bark/water ratio of 1:10 (10% w/v).The bark/water ratio was selected based on earlier studies by Kemppainen [20], who had reported a 1:10 bark/water ratio as the optimum for the batch extraction of tannin from spruce bark.
After the 2 h period, the extract was separated from the bark by decanting through a 90 mm Buchner funnel.For CHW extracts, wet bark from the CW extraction was transferred into 10 L of deionized water.Next, the mixture was rapidly heated to 85 • C with a VWR magnetic hotplate stirrer equipped with a temperature sensor.The mixture was heated for 2 h under 100 rpm magnetic stirrer agitation.It is important to mention that 85 • C was used for the hot water extraction because it has been reported that a higher temperature leads to a minimal increase in tannin yield and a surge in the amount of bound sugars in the extract [16].After heating, the bark extracts were decanted through a funnel.A similar extraction method and the bark/deionized water ratio (1:10) used for CHW was used for the HW extraction, with the exception that fresh spruce bark was employed.
The concentration of condensed tannin in the extracts was analysed by acid-butanol assay according to Gessner and Steiner [23], who reported the concentration in quebracho tannin equivalent.The liquid extracts were pulverized by freeze drying with a Christ Alpha freeze dryer.Before freeze drying, water was removed from the extracts with the aid of a 10 l Buchi rotary evaporator to increase the tannin concentration and minimize the time required for freeze drying.For each vacuum evaporation, 4.75 l of water content was removed from 5 l of extract, translating to a water content removal of 95%.After vacuum evaporation, the concentrated tannins were centrifuged at 2500 rpm for 10 min to separate insoluble residues from the concentrated extracts.Next, the extracts were frozen in a Christ K40 freezing bath before being transferred to the freeze dryer; a completely freeze-dried tannin product was obtained after 12 h.Tannin yields were calculated by dividing the mass of pulverized products by the dry mass of the spruce bark used during the extraction process, and results were expressed in grams per kilogram of dry bark.It is important to highlight that the tannin extractions and pulverizations for the different seasonal batches were performed simultaneously, immediately after the spruce bark had been collected from the wood mills.The subsequent biocoagulant modifications were carried out after all the pulverized tannins for both seasons had been obtained.

Modification of tannin samples
The aminomethylation of tannins was performed through the Mannich reaction, as described in the literature [7].The Mannich modification was initiated by gradually dissolving 2.5 g of tannin samples into 10 ml of ultrapure MQ water under ambient temperature.The temperature of the tannin solutions was brought up to 70 • C before adding 4.9 ml of ethanolamine.Then the pH of the solution was adjusted to 6.5 from an initial pH of ~ 11 with concentrated HCl (37%) under continuous agitation by a magnetic stirrer.The temperature of the solutions was again raised to 80 • C before a gradual addition of 1.38 ml of formaldehyde with the aid of a peristaltic pump over a 90-minute period.After that, the solutions were mixed and reacted at 85 • C for 180 min under continuous mechanical agitation.The reaction was then deactivated by adding 5 ml of MQ water and adjusting the pH of the products to 1.6 with HCl (37%).Next, the synthesized coagulants were standardized by transferring them into a 50 ml volumetric flask and filling to the mark with MQ water, which increased the pH to ~ 2. The solid content in the coagulant before and after standardization was ~ 14% (w/v) and ~ 5% (w/v), respectively.The dosage of coagulants was determined based on the active coagulant content of the standardized product.

X-ray spectroscopy
The XPS characterization of the spray-dried tannin samples was carried out using a Thermo Fischer ESCALAB 250xi with an X-ray source from a rotating monochromatic Al anode, generating an X-ray beam at 1486 eV.The emanating spectra were recorded perpendicularly to the surface and the energy was measured within a mean radius of 650 μm.
Wide-scan spectra were recorded in steps of 1 eV and a pass energy of 191 eV, whereas the high-resolution spectra were obtained in steps of 0.1 eV and a pass energy of 20 eV.The analysis of XPS data was performed with Avantage software and the binding energy for adventitious carbon was charge shifted to 284.8 eV.All spectra were fitted using a Shirley-type background.

Nuclear magnetic resonance analysis
The 1 H nuclear magnetic resonance (NMR) spectra of summer and winter extracts were measured by using a 600 MHz Bruker NMR spectrometer, equipped with a cryoprobe (Bruker Prodigy TCI 600 S3 H&F-C/N-D-05 Z) and an automatic Sample Jet sample changer.Prior to the NMR measurements, 400 µl of liquid sample was transferred to a 5 mm NMR tube followed by the addition of D 2 O (125 µl) containing 3-(trimethylsilyl)-propionic-d 4 acid (1.5 mM) as an internal standard of known concentration.Compounds were identified by using separately measured reference compounds. 1H NMR spectra were collected using an automation program with the following parameters: 90 • pulse angle, total relaxation delay of 13 s and 32 scans at 300 K. Detailed descriptions of the analytical procedure can be found in the supplementary document.

Electrospray ionization mass spectrometry
Tannin extracts were sonicated for 30 min and diluted 10-fold with 50% acetonitrile/50% 0.1% formic acid in water, centrifuged for 30 min at 14,000 g and filtered through a 0.45 syringe filter.The filtrates were analysed by MS on a Q-Exactive plus mass spectrometer (Thermo Scientific) equipped with a heated ESI source, by direct infusion at 20 µl/ min.The mass spectrometer was operated in positive polarity with a resolution of 70,000 and a m/z range of 70-2000.

Charge density measurements of coagulants
The charge density of the modified coagulants was measured with a Mütek PCD-05 particle charge detector (Hershing, Germany).Prior to measurement, 0.1 ml of the modified tannin coagulants was diluted in 250 ml of ultrapure MQ water (resulting pH was 4.6-4.8).The suspension was stirred briefly with a magnetic stirrer to obtain a completely homogeneous solution.Then 10 ml of the suspension was titrated with 0.001 N PesNa to the endpoint where the cationic solution was neutralized.The charge density of each sample was measured in triplicate, and the mean value was taken.The charge densities of the products were estimated as the charge per gram of solid tannin in the standardized volume of coagulants.

Water analyses and coagulation experimental procedure
A kaolin/river water mixture was used as the water sample to determine the coagulative effectiveness of the tannin coagulants; the preparation of this mixture has been reported in detail in an earlier publication [8].The turbidity and conductivity of the water samples were measured with a Hach 2100Q turbidity meter and a Mettler Toledo conductivity meter.The total surface charge (TSC) of the water samples was determined with a Mütek particle charge detector (PCD-05) by titrating 10 ml of the water samples with PesNa or PolyDADMAC titrant (0.001 N).The pH of the solutions was measured with a Metrohm 744 pH meter.Dissolved organic carbon (DOC) measurement was performed with a Sievers 900 portable total organic carbon (TOC) analyser, while the ultraviolent absorbance (UV) of water samples was measured at a wavelength of 254 nm with a Shimadzu ultraviolet spectrometer.Before the DOC and UV 254 measurements, the water samples were filtered through a 0.2 µm VWR polyethersulphone filter membrane.The specific ultraviolet absorbance (SUVA) value was determined according to equation (1), which is in line with guidelines from the United States Environmental Protection Agency [24]: Coagulation experiments were conducted with a Kemira flocculator 2000 jar test apparatus equipped with 1 L glass beakers.The beakers were filled to the 800 ml mark with the kaolin/river water mixture and dosed with different concentrations of the tannin coagulants.The suspensions were stirred rapidly at 150 rpm for 1 min, followed by slow mixing at 40 rpm for 20 min and then sedimentation for 30 min.After sedimentation, 250 ml of the supernatant was extracted from a point 5 cm below the surface of the test water sample for further analysis.

Extraction yield for condensed tannin
Condensed tannin is suitable for synthesizing cationic coagulants due to its nucleophilic characteristics [10,25].As shown in Fig. 2, extracts from the bark of the winter-harvested spruce from the various extraction procedures contained a higher amount of condensed tannin.The tannin content in the CHW and HW extracts from the winter-sampled spruce bark was ~ 135% higher than that collected during the summer.Furthermore, a higher amount of tannin was recorded in cold-water extracts from winter bark when compared to the summer counterpart.The low tannin content in the cold-water extracts indicates that only a negligible fraction of tannin is extracted with cold water, and a more aggressive extraction method is required to extract a significant amount of tannin from spruce bark.This is in agreement with previous studies by Ding et al. [22] and Bianchi [26], who reported a low concentration of tannin when cold water was used as the solvent for spruce bark.In their studies, Ding et al. [22] carried out extraction with water with a temperature of 10 • C while Bianchi [26] used 30 • C water.
The lower condensed tannin yields of the summer batches can be attributed to the rapid degradation of hydrophilic constituents in the bark of summer-harvested spruce.Halmemies et al. [27] have reported that hydrophilic extracts such as tannin and other phenolic compounds decreased by 43% 12 weeks after tree felling, while the hydrophilic extractives in the bark of winter-harvested trees remained unaffected during the same time frame.The underlying reason for the cause of rapid degradation of hydrophilic extracts from bark sourced during the summer season was ascribed to higher UV radiation, and high microbial and insect activity on harvested trees.However, factors such as tree age, soil quality, pollution and plant parts have also been reported to influence the level of tannin extraction [20,28].Due to the limitation of the scope of this study, consideration was only given to the effects of seasonal variation and extraction procedures.
The yield of pulverized tannin for extracts from the Stora Enso mill spruce bark is presented in Table 1.The mass of pulverized tannin obtained from the extracts ranged from 13.6 to 41.6 g/kg of dry bark.Similar yield ranges have been reported for successive cold-and hotwater extractions for spruce bark [17,22].A detailed analysis of the results showed a lower yield range from summer extraction (13.6 -37.9 g/kg of dry bark) compared to its winter counterpart, which produced a yield range of between 26.1 and 41.6 g/kg of dry bark.It is worth mentioning that visual observation of the pulverized tannin showed that the CHW and HW tannins were composed of homogeneous and flaky particles.In contrast, the CW tannin was composed of heterogeneous powdered particles with twig-like structures.Comparing the low tannin content in the CW extracts (Fig. 2) to their relatively higher mass of pulverized tannin given in Table 1, it is logical to conclude that the CW extracts contain a larger amount of other hydrophilic extractives (impurities) aside from tannin.Bianchi [26] reported a greater amount of non-phenolic constituents in a 30 • C cold-water tannin extraction from spruce bark than in extractions obtained with 90 • C hot water.These impurities have been reported to reduce the quality of tannin [16].Therefore, the CW tannins cannot be considered suitable for the synthesis of coagulants for water treatment applications.Nevertheless, CW tannin was investigated further in this study for comparative purposes.

XPS characterization of tannin extracts
XPS characterization was performed on the tannin samples to investigate the difference in the chemical compositions and elemental states of the extracts.The XPS wide survey showed that all tannin samples were composed mainly of carbon (C1s ~ 285 eV) and oxygen (O1s ~ 533 eV) peaks, which was in agreement with previous studies [8,10].Small amounts of inorganics (K, P, Cl, Si, N and Ca) were also detected in most of the tannin samples, and especially the summer batches showed a higher amount of potassium.Earlier studies on the inorganic constituents of spruce bark have reported that it is mainly composed of potassium, nitrogen and calcium [18,29].
The C1s peak was analysed in detail to elucidate the functional groups, and the overlaid spectra for the summer and winter extractions are presented in Fig. 3a and b, respectively.For the SCW, summer CHW  (SCHW) and summer HW (SHW) tannins in Fig. 3a, the C1s peaks can be fitted into three components centred at 284.8 eV, 286.5 eV and 288.1 eV assigned to C-C/C=C/C-H (C1), C-OH/C-O (C2) and C=O (C3) functional groups, respectively (Fig. S2) [30].Furthermore, the overlaid C1s spectra for the summer samples displayed visible doublet peaks at 292.8 eV and 295.6 eV and were ascribed to K2p 3/2 and K2p 1/2 of potassium [31,32].The peak at 292.8 eV also contains an overlapping π-π* shakeup satellite attributed to sp 2 aromatic carbon [33].The C1 and C2 components in the C1s spectrum for the winter tannin samples were located in sequential order at 284.8 eV and 286.5 eV.For the C3 component, the fitted peaks for the winter CW (WCW) and winter HW (WHW) samples were located at 288.1 eV, while that of the winter CHW (WCHW) sample was centred at 287.8 eV.This implies that the --C=O attributed to the presence of non-stilbene glycosidic and polymeric sugars in spruce tannin extracts [10,16] differs by 0.3 eV in the WCHW.The position of this peak at 287.8 eV indicates the uniqueness of the C3 component in the WCHW tannin compared to other samples.By analysing the proportion of the C3 component in the C1s peak, it was observed that its proportion was low in the WCHW extract (1%) compared to the WCW and WHW extracts, which possessed 7% and 4%, respectively.The percentual proportion of the C3 component in the SCW, SCHW and SHW samples was 15%, 5% and 7% in sequential order.This low proportion demonstrates that the WCHW extract contains the lowest amount of organic impurities in the form of glycosides and polymeric sugars.As shown in Fig. 3b, the characteristic doublet peak ascribed to potassium in the summer tannins was not evident in the winter samples; potassium was probably present but could not be distinctly observed in the survey.Hence, it was suggested that the  excitation observed around the region could be a combination of a π-π* shake-up satellite and K2p.

Nuclear magnetic resonance of tannin extracts
Fig. 4 is a collection of the measured 1 H NMR spectra for summer and winter bark extracts compared to TSP (trimethylsilyl propanoic acid) reference signals.Table 2 lists the raw integrals obtained from the NMR spectrum of each sample.The main differences between the summer and winter bark extracts are the aromatic compounds/phenols, carbohydrates/alcohols and hydrocarbon molecules.Moreover, ethanol and methanol signals at 2 ppm and 3.5 ppm are present only in the summer samples, while acetyl group (MeCO-) signals at ~ 2.3 ppm only exist in the winter samples.For example, there is an expansion of the anomeric sugar area (5.5-5.0 ppm) in the WHW sample (see Fig. S2), showing different amounts and types of carbohydrates.
As can be seen in Table 2, the amount of compounds in the coldwater extracts is the lowest while that of hot-water is the highest.An estimation of the percentage of aromatic compounds/phenolic content in individual summer extracts showed that there was 21%, 24% and 20% in SCW, SCHW and SHW, respectively.For the winter extracts, WCW, WCHW and WHW were recorded as containing 17%, 27% and 14% aromatic compounds/phenolic constituents in sequential order.This result indicates that the SCHW and WCHW extracts possessed the highest percentage of aromatic compounds/phenolic content.High phenolic content in tannin extracts has been associated with high tannin quality [16].Furthermore, the 1 H NMR analysis indicates that the coldwater extracts (SCW and WCW) contained about 30% carbohydrates/ alcohol constituents.For the cold-plus hot-water extractions, the carbohydrates/alcohol constituent in the extracts was evaluated at 59% and 53% for the SCHW and WCHW extracts, respectively.The carbohydrates/alcohol content in the hot-water extraction was approximately 53% and 63% in sequential order for the SHW and WHW extracts.These results suggest that the WHW extracts contain a higher concentration of impurities in the form of carbohydrates or alcohol.Some of these carbohydrates could be free mono and dimeric sugars such as glucose, while the alcohols could be aliphatic alcohols [34].The carbonyl content in the summer extracts was considerably higher than in the winter sample extracted with the same procedure.With regard to spruce extracts, carbonyls are associated with the presence of glycosides and sugars [18,20].Further purification of extracts could be achieved through membrane-based separation techniques [22,35]; however, it is essential to note that such purification techniques will increase the cost of extraction.
Table 3 shows the integrals of some individual compounds that can be assigned from the spectra.These compounds are ethanol (EtOH), methanol (MeOH), acetic acid (AcOH), 2,3-butanediol (Bu-diol) and propionic acid (EtCO 2 H).There are also two compounds with characteristic chemical shifts: a para-substituted phenol (Ar-OH, e.g. a compound like salidroside) and a CH 3 COR compound, possibly acetone.These integrals are taken from baseline-corrected spectra in which the other overlapping signals have been removed.The values of exact concentrations in millimol/l are given in Table S2, since the number of protons in each compound is known.According to the results, the summer samples contained the highest number of assignable compounds.Also, it could be deduced that SCW and SHW extracts possessed a higher volume of combined assignable compounds, although no study was found that reported the effect of these compounds on tannin quality.However, these compounds could form competitive reactions during the aminomethylation process of tannin coagulant.For example, an alcohol functional group can interact with protonated amine under acidic conditions to produce corresponding ethers [36].

Mass analysis of the mass fraction of tannins
The mass spectra of the tannin extracts are shown in Fig. 5.They show that the selected extraction procedures and seasonal variation influence the class of extractable compounds in the extracts.For the SCW and WCW spectra shown in Fig. 5a and b, it can be seen that the summer and winter tannin extracts are characterized by compounds of distinctive mass fractions.The SCW fractions are predominately within a very narrow mass range of 180-184 m/z ascribed to free monomers of sugars [37].On the other hand, the mass spectrum for the WCW extract presented in Fig. 5b, depicted an extract with more diverse compounds compared to SCW.Mass fractions of ions in the spectrum range from 168 m/z to 574 m/z, and the most abundant fraction in the spectrum gives an ion at 364 m/z, which was suspected to be an abduct of sucrose [M + Na] - [38].Despite having very low intensities in the SCW spectrum, the mass spectra for the CW extracted samples showed that both extracts possessed mass fractions in ion regions corresponding to procyanidin dimers (494-577 m/z) [10,39].A prominent mass fraction in the spectrum of both samples is located at 532 m/z [M− H] -, assigned to trihydroxyflavan dimers [39].The detection of these compounds indicates the presence of condensed tannin in the cold water extracts.
Fig. 5c and 5d present the mass fractions of compounds present in the CHW extraction from summer and winter spruce bark.In contrast to the pattern observed in the CW spectra, the CHW extract of the summer batch tends to contain compounds of more diverse mass fractions than the winter extract (Fig. 5c).A noticeable difference in their spectra is the almost complete lack of visibility of mass fractions <200 m/z in the spectrum for the WCHW extract (Fig. 5d).The compounds in the lower mass fractions have been ascribed to moieties of monomeric sugars in the extract [37].In relevance to this study, sugars in the extract are regarded as impurities, and they have been implicated in reducing the quality of tannin from spruce bark in previous studies [20].The phenomena behind the selective elimination of monomeric sugars and simple carbohydrates from the WCHW extract is not fully understood.However, one conclusion from the comparison between the summer and winter spectrum of the CHW extracts was that the cold water extraction acted as a pre-treatment for removing free monosaccharides and sugars from the spruce bark of the winter batch.Furthermore, some notable mass units in the CHW samples associated with condensed tannin were detected at 290 m/z and 578 m/z, ascribed to monomers and dimers of procyanidin units, respectively [40,41].It is important to note the imposing mass fraction located at 304 m/z in the SCHW, identified as taxifolin [42].Taxifolin is a bioactive compound produced in significant amounts by spruce as a defence response to pathogen infections and herbivory attacks [43], which are more prevalent during the summer season.This supports the claim by Halmemies and coworkers [27], who attributed the difference in the extractives from winter and summer spruce bark to the variation in seasonal bioactivity.
The spectra for the HW extractions from the summer and winter batches are displayed in Fig. 5e and 5f.The HW spectra show that direct hot-water extractions from the seasonal batches of spruce bark produced ions with extremely distinctive mass numbers and intensities but located within a similar mass region (135-750 m/z).The SHW spectrum contains more mass fraction units, indicating an extract with more diverse compounds.Due to the limitation in references for mass spectrometrybased metabolomics, identifying flavonoids through mass fractions is challenging [44].In general, the spectra substantiate the findings from the NMR analysis that the spruce tannin extracts from the two harvesting seasons are characterized by a unique set of individual compounds.

Charge densities of coagulants
Many studies on the application of tannin-based coagulants in water treatment applications have established that charge neutralization plays a critical role in the destabilization of colloids and humic substances [8,12,45].Thus, coagulants with higher charge densities are generally preferred during tannin modification.According to the charge density results presented in Fig. 6, the coagulants modified with the winter extracts were observed to possess higher charge densities, except for the HW extracts, which showed similar values.Regarding the cold-water extractions, the charge densities of the coagulants synthesized with the WCW spruce extract were approx.40% higher than SCW.Comparing the charge densities of coagulants modified with CHW extraction, that of the WCHW was 38% (3.38 ± 0.01 meq/g) higher than that of the SCHW (2.45 ± 0.04 meq/g).The results also showed that tannin from the WCHW extract produced coagulants with the highest charge density among all the extracts.
For the summer bark, the results shown in Fig. 6 indicate that SCW produced the coagulant with the lowest charge density (1.58 ± 0.01 meq/g) among the summer extracts.In contrast to the winter extracts, there was not much difference between the charge densities of coagulants modified from SCHW (2.45 ± 0.04 meq/g) and SHW (2.48 ± 0.01 meq/g) extracts.Furthermore, the variation in their charge densities was confirmed to be statistically insignificant (t-test, p < 0.05, two-tailed).With the exclusion of winter hot-water extraction (WHW), the pattern of charge densities in most coagulant modifications is similar to the condensed tannin content measured in the extract for both seasonal batches.The charge density displayed for the WHW coagulant showed a divergence from the conventional positive correlation between the condensed tannin in extracts and charge densities of modified coagulants.Aside from the initial cold-water extraction before hot-water extraction for the CHW extracts, CHW and HW extracts were subjected to the same experimental conditions.It is reasonable to suggest that the cold-water extraction was beneficial to the WCHW coagulants by eliminating impurities (free sugars) that could have been detrimental to the aminomethylation process.Bianchi et al. [17] and Ekam and Ebong [46] have reported that glycosides and carbohydrates are effectively removed by cold-water extraction.More importantly, Bianchi et al. [17] demonstrated that a preliminary cold-water extraction before hot-water extraction could substantially extract free mono-and oligosaccharides (mainly glucose, fructose and sucrose) from spruce bark, therefore increasing the purity of extracts from the bark in subsequent extraction procedures.Interestingly, a prior cold-water extraction was not beneficial in the case of the summer batch.This was due to the faster degradation of hydrophilic compounds (tannin, stilbenes and carbohydrates) in the bark during the warm summer season [27].The rapid loss of the hydrophilic constituents in the summer batch eroded the advantage of a cold-water treatment for the SCHW coagulant.
Before using a Mannich-modified tannin coagulant in a real water treatment application, an essential factor to consider is the concentration of unreacted formaldehyde in the product.Formaldehyde is vital for the cross-linking between amine and tannin constituents during the cationization process.Unfortunately, there is sufficient evidence in the literature on the carcinogenicity of formaldehyde [47] to require minimal residual formaldehyde for these coagulants to be viable.The weight percentage of residual formaldehyde in the modified coagulants is shown in Table 4.By comparison, the formaldehyde percentage recorded in the modified coagulants was even lower than the percentage values reported earlier for tannin-based coagulants [8].More interestingly, these percentage values were found to be relatively lower than the background level of formaldehyde detected in some food commodities such as dried fish (0.02 %) and mushrooms (0.04 %) [48].In order to support the viability of the modified coagulants in real applications, a scenario was considered where the dosage demand of a treated water sample was 100 mg/l for the coagulant with the highest wt% of residual formaldehyde (WCW), assuming all residual formaldehyde stays in the treated sample.This extreme scenario estimated that a mere 0.1 mg/l (0.00001 wt%) of residual formaldehyde would stay in the treated water sample.

Comparision of coagulation performance
Previous studies have indicated that charge density is an important indicator for evaluating the coagulative performance of tannin-based coagulants in effluents with a positive charge demand [10,49].Hence, based on their charge densities SCHW and WCHW were chosen for the investigation of the coagulative performance of the modified tannin coagulants.The biocoagulants were tested in a kaolin/river water mixture.Table 5 presents the characteristics of the river water and the kaolin/river water mixture.
The dosage curves for the residual turbidity, total surface charge (TSC), UV 254 absorbance and SUVA value of the treated water sample are depicted in Fig. 7.As shown in Fig. 7a, both coagulants were able to reduce the turbidity of the kaolin/river water mixture significantly.The WCHW coagulant recorded a turbidity reduction of ~ 90% with an optimal coagulant dosage of 15 mg/l, whereas the SCHW coagulant achieved a minimum turbidity of > 90% reduction with a 20 mg/l dosage.The lower dosage required by the WCHW coagulant was attributed to its higher charge density.It is important to note that these optimal dosages for turbidity reduction were lower than the 50 mg/l reported in our earlier studies for a spruce tannin coagulant modified with the same amine and using the same modification conditions [10].The impressive lower coagulant dosages recorded in this study were attributed to the optimized extraction process and a faster freeze drying and pulverization procedure, producing spruce tannin coagulants with superior charge densities (>57%) compared to those obtained in our previous study.Furthermore, as can be seen in Fig. 7a, the turbidity of the water sample treated with WCHW coagulant increased after the coagulant dosage exceeded 100 mg/l.This increase in turbidity at excess coagulant concentration was attributed to charge reversal, a phenomenon peculiar to the overdosing of coagulants with high cationic properties [50].On the other hand, the SCHW coagulant maintained a relatively stable residual turbidity even at extreme coagulant dosages.
The TSC of the water samples after the coagulation experiment with the SCHW and WCHW coagulants is shown in Fig. 7b.The results reveal that the TSC of the treated water samples decreased (TSC becomes less negative) as the coagulant dosage increased.Both treated water samples exhibited an identical TSC reduction pattern, but the TSC reduction was more pronounced in the WCHW treated water sample.A complete charge neutralization in the treated water sample was achieved by the WCHW and SCHW coagulants with a dosage of 40 mg/l and 80 mg/l, respectively.These dosages were higher than those required for optimal turbidity reductions.Previous studies have reported that excess tannin   coagulants stay dissolved in treated effluents [8].Hence, it is rational to expect a higher surface charge in supernatants treated with the WCHW coagulant due to its superior cationic charge density when compared with SCHW (Table 4).The plot for UV 254 absorbance as a function of the dosage of the modified coagulants is displayed in Fig. 7c.UV 254 absorbance is widely used in water treatment as an indicator to characterize the organic constituents in a water sample.A detailed analysis of the plot illustrates a 49% and 24% decrease in UV 254 absorbance by the WCHW and SCHW coagulants with 25 mg/l and 35 mg/l dosage concentrations, respectively.This trend was in agreement with earlier studies with the same water (kaolin/river water mixture), which showed that tannin-based coagulants with higher charge densities were more effective in reducing UV 254 absorbance [8,10].Also, Chang et al. [51] and Bolto et al. [52] have observed superior performance in UV 254 absorbance reductions by polymers with higher cationic charges in municipal wastewater and natural waters, respectively.
The SUVA value is known to provide an overview of natural organic matter (NOM) in water samples.NOM exists in water as either hydrophobic or hydrophilic components, but most water samples are characterized by both components [53].The dosage curve of the SUVA values obtained from the treated water samples at different coagulant dosages is illustrated in Fig. 7d.The initial SUVA of the water was above 3.5 l/mg-m, which indicates that the kaolin/river water mixture is characterized by aquatic humics and a mixture of NOM [14,15].The SUVA value obtained from the treated water sample at optimal dosages was<2 l/mg-m for both tannin coagulants (Fig. 7d).A SUVA value < 3 l/ mg-m implies that the treated kaolin/river water mixture sample was now largely composed of non-humics and high in hydrophilic material [50,53].It is imperative to mention that the dissolved organic matter (DOC) of the treated samples increased slightly at optimal dosages (Fig. S3), and the pattern of dosage curves indicates that residual coagulants also affected the other measured water quality indicators.Nevertheless, the increase in SUVA value at higher coagulant concentrations was a factor of both an increase in UV 254 absorbance and a decrease in DOC values.

Conclusion
The study showed that extraction procedures and seasonal variation influenced the composition of the extracts and this in turn affected the quality of the tannin coagulant.Winter-harvested spruce contained a higher amount of condensed tannin in various extraction procedures, and with regard to the synthesized tannin-based coagulants, WCHW tannin was considered the most viable.This was attributed to a slower degradation of phenolic content during the winter season and an advantageous cold-water pre-extraction procedure.The cold-water preextraction procedure significantly reduced monomeric sugars and simple carbohydrate compounds in the WCHW spectrum.Coagulation studies were performed with coagulants obtained from the SCHW and WCHW extracts and the jar test experiment showed that the WCHW coagulant slightly outperformed the SCHW coagulant in the studied kaolin/river water mixture due to its higher charge density.The study demonstrated that a cold-and hot-water extraction procedure increased the quality of tannin extract from winter-harvested spruce bark, but no similar benefit was observed with summer-harvested spruce.Other

Fig. 1 .
Fig. 1.Flow diagram of bark processing for different tannin extraction methods.

Fig. 2 .
Fig. 2. Tannin content in extracts from different water extraction processes for summer and winter spruce bark from industrial processes.Error bars represent the deviation in tannin concentration of extracts from the Stora Enso and Vuokila Wood Oy mills.

A
.Bello et al.

Fig. 6 .
Fig. 6.Charge densities of coagulants obtained from cold-water (CW), coldand hot-water (CHW) and hot-water (HW) extractions from the bark of summer-and winter-harvested spruce (error bars represent the deviation in charge density measurements of two repeated modifications).

Fig. 7 .
Fig. 7. Effect of coagulant dosage on a) turbidity b) total surface charge c) UV 254 absorbance and d) SUVA of the kaolin/river water mixture using SCHW and WCHW coagulants.

Table 1
Pulverized tannin yield of extracts from summer and winter spruce bark from Stora Enso mill.
BatchExtraction method Yield (g/kg of dry bark)

Table 2
Raw integrals from each 1 H NMR spectrum.

Table 3
The integrals of assignable individual compounds in the 1 H NMR spectra.

Table 4
Residual formaldehyde and charge densities of coagulants after two months of storage.

Table 5
Characteristics of river water and kaolin/river water mixture.