1 Introduction

Clear-cutting and ditch network maintenance (ditch cleaning) in drained forested peatlands markedly increase export of nutrients and suspended solids to watercourses (Joensuu et al. 2002; Nieminen 2004; Kaila et al. 2015; Nieminen et al. 2017a). Recent studies also indicate that old drained peatland forests may cause a longer term and much higher nutrient load than previously understood (Nieminen et al. 2017b). It is very likely that climate change will further escalate the problem by enhancing peat decomposition and nutrient release (Nieminen et al. 2017b). While current water protection methods such as sedimentation pits, sedimentation ponds, and peatland buffer areas can reduce the export of suspended solids, these methods are rather ineffective in reducing dissolved nutrient load (Joensuu et al. 2002; Liljaniemi et al. 2003; Nieminen et al. 2005). This calls for new methods in water protection. One solution could be adsorption-based active water purification using biochar-filled reactors in ditch drains.

Adsorption process, with activated carbon as adsorbent, is widely used in water purification in chemical engineering. The characteristics of activated carbon are well studied; the adsorption efficiency is based on the exceptionally high specific surface area and porosity. Similarly to activated carbon, biochar is a side product of pyrolysis i.e. heating organic material such as wood, straw, manure, or industrial waste (paper sludge and biosolids) under low oxygen supply (Lehmann and Joseph 2015). Pyrolysis produces energy in the form of biogas and bio-oil, and the remaining carbon-rich solid material is called biochar. Characteristics of biochar resemble those of activated carbon; it has porous structure, large specific surface area and high cation and anion exchange capacity, and it is thus capable of adsorbing nutrients and other dissolved compounds from water (Mohan et al. 2014; Tan et al. 2015; Yin et al. 2017). In some studies, the adsorption capacity of biochar has been even better than in activated carbon (Berger 2012; Inyang and Dickenson 2015; Dalahmeh 2016). Biochar has so far mainly been used as a soil amendment, but as a good adsorbent, it may provide a lower cost alternative to activated carbon in wastewater treatment (Foereid 2015; Gwenzi et al. 2016). Recently, adsorption of biochar has been utilized for inorganic nitrogen (N) and phosphorus (P) removal from dairy effluent, piggery slurry, wastewater, and urban runoff water (e.g. Zhang et al. 2012; Hollister et al. 2013; Sarkhot et al. 2013; Zhang et al. 2014; Kizito et al. 2015; Wang et al. 2015; Takaya et al. 2016). However, only very few attempts have been made to use biochar for nutrient reclamation from forest runoff water (e.g. Lafdani et al. 2020) so far.

To design a biochar reactor for water purification, it is necessary t know the adsorption rate and adsorption capacity of the biochar for the prevailing substances in the water. Previous laboratory experiments indicate that biochar can adsorb compounds from water; however, these experiments have typically been done for single compounds, whereas simultaneous adsorption of several compounds from natural waters has not been studied before. As multiple compounds compete simultaneously for the adsorption places in the adsorption process (Palviainen et al. 2018), all dissolved compounds in the water can affect the performance of biochar in nutrient reclamation. The chemical composition and nutrient concentrations of peatland forest runoff water differ significantly from agricultural or urban runoff waters (Oliveira et al. 2018). Therefore, the studies reporting biochar utilization in contaminant removal from agricultural and urban waters cannot be directly applied for forested peatlands.

This study was part of a research project aimed at developing biochar‐based water protection tools for forestry. Different scales, such as a small-scale laboratory experiment, meso-scale laboratory experiment, and a full-field scale study (biochar reactors installed into ditch drains), were applied during the project. The aim of this study was to determine the adsorption rate and adsorption capacity for two wood biochars (Norway spruce and silver birch) representing the most common raw biochar materials in Finland. We carried out a 10-day laboratory experiment where biochar was stirred with runoff water from a clear-cut peatland area at constant temperature. Changes in water pH, total nitrogen (TN), nitrate nitrogen (NO3−-N), ammonium nitrogen (NH4+-N), phosphorus (P) and total organic carbon (TOC) concentrations were measured throughout the experiment. Adsorption was quantified based on the concentration changes, and adsorption model containing the adsorption rate and adsorption capacity was fitted to the data. Implications of the results on the reactor design for forested peatland water protection purposes were discussed.

2 Materials and methods

2.1 Study site

Study site was located in Heinävesi, Eastern Finland (62° 29′ 56.0″ N, 28° 35′ 56.0″ E, Fig. 1). The mean annual temperature in the area is 3.6 °C and the mean annual precipitation is 638 mm (Pirinen et al. 2012). The site is a drained fertile spruce fen (Päivänen and Hånell 2012) that was clear-cut in August 2018. In the clear-cut, a total of 319 m3 (124 m3 ha−1) Norway spruce (Picea abies Karst.) and 239 m3 (93 m3 ha−1) silver birch (Betula pendula Roth.) were harvested. Total catchment area was 31.72 ha, and the size of the clear-cut area was 2.57 ha. The depth of the peat layer was 0.5–0.7 m and it was underlain by clay.

Fig. 1
figure 1

Study site in Heinävesi, Eastern Finland (PaITuli—spatial data for research and teaching; available at https://avaa.tdata.fi/en_US/web/paituli/latauspalvelu)

Full size image

Water for the adsorption experiment was collected from a ditch that drains the clear-cut area in two sampling occasions dated in October and November 2018. The initial nutrient concentrations of water were higher in November (1.4 ± 0.05 mg l−1 TN, 0.6 ± 0.09 mg l−1 NO3− -N, 0.1 ± 0.02 mg l−1 NH4+-N and 28 ± 0.01 mg l−1 TOC) than in October (0.9 ± 0.03 mg l−1 TN, 0.2 ± 0.01 mg l−1 NO3− -N, 0.05 ± 0.03 mg l−1 NH4+-N and 25 ± 0.2 mg l−1 TOC). The runoff from the clear-cut site flows into an oligo-mesotrophic, clear-water Lake Kermajärvi in the Vuoksi main catchment area. Nutrient exports following clear-cut from fertile peatland sites may jeopardize the good ecological status of oligo-mesotrophic clear-water lakes. Active water purification methods can be particularly suitable for sites where balancing between high economic income and high environmental risk is required.

2.2 Experimental design

We used two different commercial biochar types (Norway spruce and silver birch, Table 1) produced by slow pyrolysis at 600 °C (Carbofex Ltd., Tampere, Finland). In recent studies, high biomass pyrolysis temperature (600 °C) has been shown to increase the adsorption capacity (e.g. Yao et al. 2012). To determine the effect of biochar particle size and dose on the adsorption rate and capacity, we selected two different particle sizes (fine < 4 mm and coarse 4–6 mm) and two different doses (3 g and 12 g). Largitte and Pasquier (2016) identified three steps in the adsorption process: (1) the external mass transfer of the adsorbate from the bulk solution to the external surface of the adsorbent, (2) the internal diffusion of the adsorbate to the sorption sites and (3) the actual sorption. Particle size affects especially the first two steps. All treatments were conducted in four replicates. The runoff water collected in October was used for replicates 1 and 2, and runoff water collected in November was used for replicates 3 and 4. Before the experiment, water temperature was allowed to stabilize to room temperature (+ 21 °C) which fit within the temperature range in field conditions (from − 25 °C to + 25 °C). Temperature was standardized among the replicates and kept constant throughout the experiment, because temperature is known to affect the adsorption process (Mizuta et al. 2004). We added biochar into 3000 ml glass jars and four glass jars remained blank controls without biochar. Thereafter, 2500 ml of water from the field site was added to the jars, and 60 ml of water was taken to measure the initial element concentration in each jar. Jars were covered with aluminum foil and placed on a platform shaker (New Brunswick™ Innova® 2300, Eppendorf Nordic A/S, Denmark). Glass jars were shaken at 105 rpm for 10 days, and 60 ml water sample was taken from each jar at following points of time: 1, 2.33, 5.5, 25, 28, 46, 49, 70, 145, 169, 196, and 215 h from the beginning of the experiment. The samples were filtered through Filtration Assembly with Whatman GF/F Glass Microfiber Filters (pore size 0.7 µm, GE Healthcare Bio-Sciences, Marlborough, MA, USA). The samples were stored at + 4 °C until further analysis.

Table 1 Characteristics of the spruce and birch biochars
Full size table

2.3 Measurements and analyses

Carbon (C) and nitrogen (N) concentration change was measured to determine the adsorption of compounds onto the biochar surfaces (Eq. 1).

$$Q_{t\_i} = \frac{{\left( {C_{{{\text{ini}}\_i}} V_{{{\text{ini}}}} - C_{t\_i} V_{t} } \right) - \mathop \sum \nolimits_{{k = {\text{ini}}}}^{t} \left( {C_{{{k\_i}}} V_{{{\text{sample}}}} } \right)}}{{m_{{{\text{biochar}}}} }},$$
(1)

where Qt_i is the cumulative adsorption of substance i (mg g−1 biochar), Cini_i is the initial concentration of substance i (mg l−1), Vini is the water initial volume (l), Ct_i is the concentration of substance i in time t (mg l−1), Vt is water volume at time t, Ck_i is the concentration of substance i in previous sampling occasion at time k, Vsample is the volume of water sample in each sampling occasion (60 ml), and mbiochar is the dry mass of biochar in the jar.

Electrical conductivity (EC) and pH of water and biochars (on a 1:2.5 v:v biochar/water solution) were measured from each sample using WTW pH/cond 340i and WTW pH 340 m (WTW GmbH, Weilheim, Germany), respectively. Total organic carbon (TOC) and total nitrogen (TN) concentrations were determined with Multi N/C® 2100 (Analytik Jena AG, Jena, Germany). The concentrations of nitrate (NO3−) + nitrite (NO2−) nitrogen (hereafter referred to as NO3−-N) and ammonium nitrogen (NH4+-N) were determined colorimetrically applying the methods by Miranda et al. (2001) and Fawcett and Scott (1960), respectively. The concentrations of phosphorus (P) were determined by ICP atomic emission spectrophotometer (ARL 3580 OES, Fison Instruments, Valencia, USA). The C and N concentrations of biochars were analyzed with an elemental analyzer (Vario Max CN elemental analyser, Elementar Analysensysteme GmbH, Germany).

2.4 Data analyses

Adsorption rate and capacity were determined by fitting commonly used integral form of pseudo-second-order adsorption model to the data (Largitte and Pasquier 2016, Eq. 2)

$$Q_{i} = \left( {\frac{{k_{{{\text{ad}}\_i}} 2_{{Q_{{{\max}\_i}} }} t}}{{1 + k_{{{\text{ad}}\_i}} Q_{{{\max}\_i}} t}}} \right) ,$$
(2)

where Qi is cumulative adsorption of substance i (TN, NH4+-N, NO3−-N and TOC, mg g−1 biochar) at time t (h), kad_i is adsorption rate (g mg−1 h−1), and Qmax_i (mg g−1) is adsorption capacity for substance i.

The effect of treatments (biochar type, particle size and dose) on kad and Qmax were analyzed using nonlinear mixed-effect models (Online Resource 1). At first, a basic model, where all treatments for compound i were included to the analysis was calculated, and the performance of the model expressed as information criteria (AIC, BIC, -log likelihood) were recorded. Then different treatments were addressed as fixed effects of kad and/or Qmax in the analysis and the model performance was evaluated against the previous step model. Replicates were considered as random effects in the model.

Our analysis revealed that kad and Qmax were different for experiments done in October and November, indicating that the initial concentration affects the adsorption. This allowed us to use the initial concentration as an extra treatment in the analysis. Because the solution of nonlinear mixed-effect model is iterative, the model does not converge if the variance structure greatly deviates between the treatments. Therefore, we did separate analyses for different biochar doses and initial concentrations.

3 Results

3.1 pH and EC

Biochar increased water pH throughout the experiment (Fig. S1). Increases in pH were higher among treatments with 12 g dose of biochar (increase 16% from the initial) compared to 3 g dose (increase 12%). Electrical conductivity (EC) increased in all biochar treatments by 20% (Fig. S2), and the changes were similar in both biochar doses. Water pH and EC did not differ between birch and spruce biochar treatments.

3.2 The concentration changes and adsorption of TN

Biochar decreased TN concentrations in water (Fig. 2). The largest decrease in TN was observed during the first 70 h of the experiment. The adsorption of TN was detected in all biochar treatments (Fig. 3). The differences among the treatments were the most evident in Qmax (Table 2). Birch biochar has predominantly higher Qmax than spruce biochar (Fig. 3). The adsorption process was more consistent in higher biochar dose and higher concentration, which can be seen from Fig. 3 and residual variance in Table 2. Particle size did not have an effect on kad, while Qmax increased with decreasing particle size. Experiment revealed that adsorption increases with increasing initial concentration. Qmax was in most cases about two times higher for initial concentration 1.4 mg l−1 compared to 0.9 mg l−1.

Fig. 2
figure 2

The effect of biochar on TN concentrations in different treatments: a 12 g of birch biochar and b 12 g of spruce biochar (initial concentration 1.4 mg l−1), c 12 g of birch biochar and d 12 g of spruce biochar (initial concentration 0.9 mg l−1), e 3 g of birch biochar and f 3 g of spruce biochar (initial concentration 1.4 mg l−1), and g 3 g of birch biochar and h 3 g of spruce biochar (initial concentration 0.9 mg l−1). Values are mean ± SD. Blue dots represent coarse particle size (4–6 mm), red dots represent fine particle size (< 4 mm), and black dots are blank water samples without biochar (colour figure online)

Full size image
Fig. 3
figure 3

The adsorption of TN in different treatments: a 12 g of birch biochar and b 12 g of spruce biochar (initial concentration 1.4 mg l−1), c 12 g of birch biochar and d 12 g of spruce biochar (initial concentration 0.9 mg l−1), e 3 g of birch biochar and f 3 g of spruce biochar (initial concentration 1.4 mg l−1), and g 3 g of birch biochar and h 3 g of spruce biochar (initial concentration 0.9 mg l−1). The solid lines represent pseudo-second-order adsorption model (Eq. 1) fitted using non-linear mixed effect model. Parameter values for adsorption rate (kad) and adsorption capacity (Qmax) are shown in each figure. Blue dots and lines represent coarse particle size (4–6 mm) and red dots and lines represent fine particle size (< 4 mm) (colour figure online)

Full size image
Table 2 Nonlinear mixed-effect models for TN in different treatments
Full size table

3.3 The concentration changes and adsorption of inorganic N

The NO3−-N concentrations decreased in all biochar-treated samples throughout the experiment (Fig. 4). The steepest decline in NO3−-N was detected in the samples with 12 g of birch biochar (Fig. 4a, c). Furthermore, Qmax for NO3−-N was the highest in birch biochar treated samples and generally increased with increasing initial concentration (Fig. 5; Table 3). The biochar particle size did not have a clear effect on NO3−-N adsorption.

Fig. 4
figure 4

The effect of biochar on NO3−-N concentrations in different treatments: a 12 g of birch biochar and b 12 g of spruce biochar (initial concentration 0.6 mg l−1), c 12 g of birch biochar and d 12 g of spruce biochar (initial concentration 0.2 mg l−1), e 3 g of birch biochar and f 3 g of spruce biochar (initial concentration 0.6 mg l−1), and g 3 g of birch biochar and h 3 g of spruce biochar (initial concentration 0.2 mg l−1). Values are mean ± SD. Dashed lines represent the detection limit (0.02 mg l−1). Blue dots represent coarse particle size (4–6 mm), red dots represent fine particle size (< 4 mm), and black dots are blank water samples without biochar (colour figure online)

Full size image
Fig. 5
figure 5

The adsorption of NO3−-N in different treatments: a 12 g of birch biochar and b 12 g of spruce biochar (initial concentration 0.6 mg l−1), c 12 g of birch biochar and d 12 g of spruce biochar (initial concentration 0.2 mg l−1), e 3 g of birch biochar and f 3 g of spruce biochar (initial concentration 0.6 mg l−1), and g 3 g of birch biochar and h 3 g of spruce biochar (initial concentration 0.2 mg l−1). The solid lines represent pseudo-second-order adsorption model (Eq. 2) fitted using non-linear mixed effect model. Parameter values for adsorption rate (Kad) and adsorption capacity (Qmax) are shown in each figure. Blue dots and lines represent coarse particle size (4–6 mm) and red dots and lines represent fine particle size (< 4 mm) (colour figure online)

Full size image
Table 3 Nonlinear mixed-effect models for NO3−-N in different treatments
Full size table

NH4+-N concentrations were low and decreased rapidly in all biochar treatments below the detection limit (0.02 mg l−1; Fig. 6). The decrease in NH4+-N was more evident in the samples with 12 g of biochar (Fig. 6a–d). In the situation where the adsorbing compound is depleted from the water, fitting the adsorption model is not meaningful, because then Qmax would not represent the real adsorption capacity of biochar.

Fig. 6
figure 6

The effect of biochar on NH4+-N concentrations in different treatments: a 12 g of birch biochar and b 12 g of spruce biochar (initial concentration 0.1 mg l−1), c 12 g of birch biochar and d 12 g of spruce biochar (initial concentration 0.05 mg l−1), e 3 g of birch biochar and f 3 g of spruce biochar (initial concentration 0.1 mg l−1), and g 3 g of birch biochar and h 3 g of spruce biochar (initial concentration 0.05 mg l−1). Values are mean ± SD. Dashed lines represent the detection limit (0.02 mg l−1). Blue dots represent coarse particle size (4–6 mm), red dots represent fine particle size (< 4 mm), and black dots are blank water samples without biochar (colour figure online)

Full size image

3.4 The concentration changes of TOC and P

The concentrations of TOC decreased in samples with high initial concentrations (28 mg l−1) during the first 2 days of the experiment (Fig. S3). After 145 h, TOC concentrations did not differ substantially between biochar and control treatments, and there were increases in TOC among all treatments.

The concentrations of P were below the detection limit (0.025 mg l−1) throughout the experiment, and, therefore, it was not possible to determine the adsorption parameters for P.

4 Discussion

Our study indicated that spruce and birch biochar can adsorb both organic and inorganic N. The review study by Yin et al. (2017) about biochar utilization for inorganic N removal from water showed that biochar can adsorb between 0.7 and 140 mg inorganic N g−1 biochar. Our results were below this range, being maximum at 0.14 mg g−1 biochar for TN, although wood biochars have generally higher NO3−-N adsorption capacity than non-wood biochars (Kameyama et al. 2016). The lower adsorption capacity in our study was probably due to the low initial N concentrations in the water. Our results support the findings of Ahmadvand et al. (2018), showing that with an increase in the initial N concentration of solution, the adsorption capacity of biochar tends to increase.

The low initial concentrations of NH4+-N (< 0.1 mg l−1) and P (< 0.025 mg l−1) were also the reason why we could not determine the adsorption parameters for these compounds, but based on the literature, it is known that biochar is an effective adsorbent also for NH4+-N and P (Yin et al. 2017). However, NH4+-N was adsorbed initially until the concentrations decreased below the detection limit (0.02 mg l−1; Fig. 6). Small fraction of the NH4+-N may have been lost through volatilization (Sha et al. 2019), but the significance of this process is probably negligible because water NH4+-N concentrations were low. Even though the surfaces of carbon-based materials are often thought to be negatively charged (Eberhardt et al. 2006; Krishnan and Haridas 2008; Yao et al. 2011), the adsorption of both NO3−-N and NH4+-N indicates that biochar is capable to adsorb both anions and cations.

Adsorption rates were similar between different biochar treatments (Figs. 3, 5). The largest differences among the treatments can be seen in the adsorption capacity. Qmax decreased with the increasing biochar particle size, which could be due to that the surface area where adsorption can take place is smaller when the particle size increases (Eberhardt and Min 2008). However, despite higher specific surface area of spruce biochar (320 m2 g−1) compared to birch biochar (260 m2 g−1), Qmax was higher with birch biochar. While the increase in the biochar surface area may improve N adsorption (Zhang et al. 2012), the adsorption capacity is not exclusively dependent on the biochar surface area (Zhang et al. 2014; Takaya et al. 2016).

The effect of biochar on water quality was visible in elevated pH (Fig. S1). Water pH is an important factor regulating ion exchange capacity and, therefore, N adsorption on biochar (Yin et al. 2017). The adsorption capacity of biochar has been shown to increase as the pH value of the solution increases from 4 to 8 (Yin et al. 2017). The initial pH in our water samples was moderately acidic (approx. 6), which is typical for waters draining from boreal peatlands (Åström et al. 2001). At low pH values, H+ and NH4+ can compete for the active sites on the biochar surfaces (Novak et al. 2010), which slows down N adsorption. On the other hand, NO3−-N is poorly adsorbed under basic conditions due to the competition between hydroxide ions (OH−) and NO3−-N ions for the adsorption sites on the biochar surfaces (Chintala et al. 2013; Iida et al. 2013).

Our results did not indicate desorption of N from biochar during the experiment (215 h). Again, this was probably due to limited availability of N in the water. Chintala et al. (2013) reported that increases in the initial NO3−-N concentrations lead to lower bond energy and more exchangeable ions, which increases desorption of NO3−-N from biochar to the water.

So far, studies of the removal of organic C from water using biochar are few. Although it has been found that biochar can remove dissolved organic carbon (DOC) from river water (6.4 mg DOC g−1 C; Lee et al. 2018), our results imply that TOC concentrations were increased with biochar addition. This can be due to biochar itself releasing organic C into water. Different types of biochar contain a range of organic C and nutrient forms with different release rates (Mukherjee and Zimmerman 2013), which should be considered in further experiments.

Current water protection methods used in forested peatlands, such as sedimentation ponds, sedimentation pits, overland flow fields and peak-flow control methods, retain suspended solids, but are not capable to effectively remove NO3−-N from runoff water (Joensuu et al. 2002; Liljaniemi et al. 2003; Nieminen et al. 2005). Furthermore, riparian buffer zones and overland flow fields require large areas, and nutrient retention through plant uptake is restricted to the growing season (Liljaniemi et al. 2003; Nieminen et al. 2005). Available physico-chemical and biological technologies developed for NO3−-N removal in turn tend to be expensive and generate additional by-products and NO3−-N concentrated waste streams (Bhatnagar and Sillanpää 2011). Our results demonstrating adsorption-based removal of NO3−-N thus provide the first evidence on potential solutions to reduce NO3−-N leaching from fertile peatlands, where N exports may significantly increase after the clear-cutting (between 0.05 and 0.22 mg l−1 increase in NO3−-N; Nieminen 2004). The detected positive relationship between the initial concentration and adsorption capacity also indicates that higher nutrient exports would lead to even more efficient adsorption of nutrients, which is ideal for the operability of the biochar reactor in the field conditions. Since the concentration decrease was the most evident during the first 2 days of the experiment, also relatively short water residence time in the biochar reactor can be sufficient for the adsorption of nutrients onto biochar surfaces in the field conditions. On the other hand, the use of biochar reactor might not be beneficial in the field sites with very low inorganic N concentrations, because the runoff water would probably run out of NO3−-N and NH4+-N during the first week after installation.

As a regenerable adsorbent, biochar could be used as a soil amendment following the adsorption-based water purification. Biochar addition has been found to improve soil properties and the growth of crops and trees (Biederman and Harpole 2013; Lehmann and Joseph 2015; Palviainen et al. 2018), and nutrients adsorbed onto biochar are easily available for plants when placed in the soil (Taghizadeh-Toosi et al. 2012). Therefore, the utilization of biochar as an adsorbent can further improve its potential as a soil amendment.

5 Conclusions

Our results indicated efficient adsorption of N onto biochar surfaces. The highest Qmax was achieved with maximum TN levels, using fine birch biochar. Increases in the export of terrestrial dissolved organic N (DON) to boreal aquatic ecosystems further emphasize the need for effective water protection to which adsorption-based N removal with biochar seems to respond well. Unlike the traditional water protection methods currently used in forestry, biochar is also capable of adsorbing NO3−-N from forest runoff water, which makes biochar a promising tool in development of new water protection methods.