Short-Term Nitrogen Uptake of Barley from Differently Processed Biogas Digestate in Pot Experiments

: The use of biogas digestate as fertilizer is limited by the farm nutrient balance. Mechanical separation and drying of digestate increases its transport worthiness as well as the economic feasibility of nutrient export. This study compares the fertilizer effect of four treatments of digestate originating from two biogas plants: untreated digestate, liquid and solid fraction of separated digestate and dried solid fraction of separated digestate. Pot experiments with barley were performed with two fertilization levels for different digestate variants. Above-ground biomass yield, nitrogen (N) and phosphorus (P) content in biomass and plant uptake efﬁciency were highlighted. The results showed that all variants have higher above-ground biomass yield than the control. Due to the reduced amount of easily available N, short-term N uptake of barley from solid fractions of digestate was low. The treatments with the dried solid fraction at low fertilization level showed up to 59% lower N removal from soil and, at high fertilization level, up to 83% lower N removal compared to the respective fresh solid fraction (100%). Depending on the feedstock of biogas plants and processing of digestate, N availability varied and inﬂuenced the short-term N uptake. It is recommended that digestate processing should be combined with ammonia recovery to prevent N losses to the environment.


Introduction
Biogas plants are typically operated to treat organic wastes from agriculture, households or industry and to provide renewable energy at the same time. Feedstock can be readily biodegradable fractions of municipal and industrial waste, manure from livestock farming or energy crops such as maize, specifically grown for the purpose of energy production. The digestate, i.e., the residue of the anaerobic fermentation process, contains numerous nutrients and can therefore be used as a substitute for mineral fertilizers in agriculture or urban greening. The share of ammonium in the digestate from agricultural biogas plants is relatively higher compared to conventional organic agricultural fertilizers such as slurry and manure resulting in a positive effect for plant growth [1,2]. Regarding nitrogen (N), the application as fertilizer is limited at European level with the European Nitrate Directive to prevent nutrient accumulation in soil and groundwater as well as to avoid environmental pollution [3]. Furthermore, the application of digestate and other fertilizers is restricted per hectare and season of digestate and mineral fertilizer. Another study with pot experiments focused on the recycling of P from semi-liquid manure and digestate as an alternative fertilizer for ornamentals. The substrates investigated also included air-dried digestate. Compared to the other investigated substrates (P-salts, steam-dried digestate), air-dried digestate was not suitable to meet the plants' needs. However, when air-dried digestate and P-salt were combined, there was a synergetic effect on the biomass yield of the plants investigated [40].
To the best of the authors' knowledge, there is no comparative investigation where conventional digestate, its conversional fractions and its dried solid fraction are directly compared to each other. The aim of our study was (i) to investigate the short-term N and P uptake of barley from processed digestate compared to the untreated digestate originating from the same fermentation plant, and (ii) to determine the impact of different digestate treatments regarding quantitative parameters like biomass yield of the plants. The focus of this study was especially on the dried solid fraction, because additional energy is required for this treatment and the knowledge of how to use this fraction as a short-term N source is lacking.

Sampling and Analysis of Digestate Variants
The digestate investigated came from two wet fermentation biogas plants in Germany (biogas plant 1, further called BP1 and biogas plant 2, further called BP2) both operated by private owners. BP1 was fed with liquid manure (75%) and a mixture of energy crops (about 13% maize silage, 8% crops and 4% grass silage) whereas BP2 was fed exclusively with energy crops (about 80% maize silage, 20% grass silage, less than 1% other crops). In BP2, the liquid fraction from separated digestates was recirculated and mixed with the silage of maize and grass in a ratio 50:50 to ensure sufficient moisture for the fermentation process and a good mixing behavior in the digester. In both biogas plants, the digestate was separated into a solid fraction and a liquid fraction by using a screw extruder. The liquid fraction was collected in storage tanks, and the solid fraction was stored on a paved platform.
Five samples were collected in 2-L jars each from untreated digestate (UD), liquid digestate (LD) and solid digestate (SD) at both biogas plants. The jars were completely filled without leaving a headspace, sealed tightly and stored at 4 • C. One sample each of the solid digestate from each biogas plant was dried in a laboratory dryer (dryer "HD 75", Robert Hildebrand Maschinenbau GmbH, Oberboihingen, Germany) for 24 h at a temperature of 60 • C to produce the solid dried digestate variants (SDD), Table 1. Prior to chemical analysis, the individual samples per variant were combined and mixed. The UD-and LD-variants were homogenized using a macerator (Büchi Mixer B-400, BÜCHI Labortechnik GmbH, Essen, Germany) to ensure that the solids were homogenously distributed within the sample. The SDD-variants were ground in a rotary mill (Brabender®Rotary Mill, Brabender ® GmbH & Co. KG, Duisburg, Germany) to a particle size of 1.5 mm.

Chemical Composition of Digestate Variants
Dry matter content (DM) was determined by the standard oven method at 105 • C for 24 h [41]. The pH value of the liquid variants was measured in the original sample and in the solid variants in a CaCl 2 -suspension [42]. Organic carbon (C org ) was analyzed by the Dumas method using an elemental analyzer (varioELcube, Elementar Analysensysteme GmbH, Langenselbold, Germany). Total nitrogen (N t ) and ammonium nitrogen (NH 4 -N) were determined according to the standard methods [43,44]. Phosphorus (P), potassium (K), calcium (Ca), magnesium (Mg), sulfur (S), copper (Cu) and zinc (Zn) were determined by inductively coupled plasma optical emission spectrometry (Vista, Varian, Australia). Analysis of elements was based on methods of the Association of German Agricultural Analytic and Research Institutes (VDLUFA) [45]. The chemical composition of the different digestate variants is listed in Table 2.

Plant Growth Experiments
The plant growth experiments were carried out in pots (100 mm diameter and 220 mm height) according to standard methods as described in [46,47]. Subsoil (C-horizon) of a loess derived luvisol free of humus was taken as soil substrate for the experiments. Each pot was filled with 1 kg of soil. To provide a minimum of plant nutrients in the soil, a basic fertilizer with N, P, K and Mg was added to each pot before planting, as presented in Table 3. Digestate variants have been added in certain quantities to the soil substrate to achieve two levels of N fertilization, namely 300 and 500 mg N kg −1 soil, which are called a low (L) and high (H). The N equivalent of the digestate variants was calculated according to [18,46,48], assuming 100% availability for NH 4 -N and 30% availability for other N components (N t minus NH 4 -N), Table 4. Due to the different mass of added digestate per variant, each pot was conditioned with Styrofoam and filled with silica sand to ensure the same filling level and weight, respectively. Pots without additional N-fertilization except the basic fertilization served as the control.
The plant growth experiments were performed with spring barley (Hordeum vulgare Var. Belana) with eleven plants per pot. Treatments and control were set up on two tables in a greenhouse and tested in four replicates in a completely randomized block design. The average air temperature during the experiment was 27.5 • C during the day and 14.2 • C at night. Four weeks after sowing, yellow panels were fixed to the pots to control Sciara hemerobioides. In the last third of the trial, a fungicide treatment was applied due to mildew infestation. After a growing phase of 54 days, the barley plants were harvested and the above-ground biomass dry matter yield (Y DM ) per pot was measured after oven drying at 60 • C. Determination of the total C and N of plant biomass was done by Dumas combustion using an elemental analyzer (Vario MAX CN, Elementar Analysensysteme GmbH, Langenselbold, Germany). Determination of P content in the samples was done by a spectrophotometer (Spectrophotometer U-3300, Hitachi Europe GmbH, Düsseldorf, Germany), preparing the samples according to [49]. The nitrogen and P content of the above-ground biomass %N and %P were calculated in % of DM. Nitrogen removal N re from soil substrate in mg per pot was calculated as:  UD1  90  407  254  249  228  114  155  684  420  321  248  124  LD1  129  375  268  199  238  103  223  629  444  235  266  104  SD1  94  562  189  440  322  239  155  950  307  893  411  340  SDD1  66  804  87  946  847  859  113  1358  131  1407  1307  1400  UD2  81  512  210  198  227  106  140  866  344  233  247  110  LD2  91  460  230  204  229  106  173  780  381  243  250  111  SD2  100  599  171  376  302  133  158  1018  278  541  377  157  SDD2  57  909  43  666  1050  373  98  1544  55  1037  1661  570 Plant uptake efficiency of N (NUE) was calculated as described in [23,50,51]: with N re , control as N removal of control and N f as available N applied with digestate variant.

Statistical Analysis
Statistical analysis was conducted by SAS program (SAS Institute Inc., Cary, USA). The data set was processed by multivariate analysis of variance in which the origin of the digestate (BP1, BP2), digestate variant, and N-fertilization level were considered as factors. Furthermore, data was tested for normality and homogeneity. In case of not fulfilling these prerequisite, logarithmic transformations of original values were carried out. The Tukey Test was used to test the differences among means (p = 0.05).

Visual Observation of Plant Development
Germination of the plants occurred 3-5 days after sowing, except for treatments LD1L (liquid digestate, biogas plant 1, low fertilization level) and LD1H (liquid digestate, biogas plant 1, high fertilization level), where germination was delayed by 2-4 days. Twenty-eight days after sowing, yellow leaf tips were observed and stalks started to show violet discolorations in treatments SDD2L (solid digestate dried, biogas plant 2, low fertilization level), SDD2H (solid digestate dried, biogas plant 2, high fertilization level) and the control (see Figure 1), which might be an indication of chlorosis caused by N deficiency due to the relatively low NH 4 -N content in the pot. After 35 days, yellow leaf tips and violet stalks also occurred in treatments SDD1L (solid digestate dried, biogas plant 1, low fertilization level), SDD1H (solid digestate dried, biogas plant 1, high fertilization level) and SD2L (solid digestate, biogas plant 2, low fertilization level). At harvest, 54 days after sowing (booting stage), yellow leaf tips were visible in almost all treatments and the stalks of nearly all of the barley plants had turned violet, which indicates N and/or P deficiency [46,52].  [46,52].

Above-Ground Biomass Yield
Above-ground biomass yield Y DM for the different treatments is shown in Figure 2.

Above-Ground Biomass Yield
Above-ground biomass yield YDM for the different treatments is shown in Figure 2.  Table 1; n = 4, whiskers show standard deviation, different letters indicate significant differences, p = 0.05).
All treatments except SDD2H showed a significantly higher biomass yield than the control. Treatment SDD2H, with 1.2 g per pot, was not significantly different from the control, with 0.9 g per pot, and the increased fertilization did not result in a higher biomass yield compared to treatment SDD2L with 1.8 g per pot. Gunnarsson et al. [23] stated a higher biomass yield for ryegrass with untreated digestate at two fertilization levels (75 mg N and 150 mg N per dm² pot surface area) compared to the control similar to our study. The above-ground biomass yield was higher with increased fertilization. This is in contrast to the present study with a lower above-ground biomass yield when comparing untreated digestate from biogas plant 1 (UD1) in the lower (L) and higher (H) fertilization level (UD1L, UD1H), respectively. The findings regarding the biomass yield of the liquid and solid fractions compared with the control are in line with the findings of Grigatti et al. [22], who performed pot experiments with Italian ryegrass, except for LD1H, which had a very low biomass yield. The maximum above-ground biomass yield in our study was achieved in the treatment SD1H with 9.1 g per pot. As expected, the higher N fertilization level with 500 mg N per kg soil resulted in a higher biomass yield, with the exception of the treatments LD1H, UD1H and SDD2H, which showed a very low dry mass yield. This might be attributed to a delayed growth in the treatment LD1H due to the high moisture content in the soil substrate as a result of the low dry matter content of this digestate variant. This means that the low biomass yield of treatment LD1H  Table 1; n = 4, whiskers show standard deviation, different letters indicate significant differences, p = 0.05).
All treatments except SDD2H showed a significantly higher biomass yield than the control. Treatment SDD2H, with 1.2 g per pot, was not significantly different from the control, with 0.9 g per pot, and the increased fertilization did not result in a higher biomass yield compared to treatment SDD2L with 1.8 g per pot. Gunnarsson et al. [23] stated a higher biomass yield for ryegrass with untreated digestate at two fertilization levels (75 mg N and 150 mg N per dm 2 pot surface area) compared to the control similar to our study. The above-ground biomass yield was higher with increased fertilization. This is in contrast to the present study with a lower above-ground biomass yield when comparing untreated digestate from biogas plant 1 (UD1) in the lower (L) and higher (H) fertilization level (UD1L, UD1H), respectively. The findings regarding the biomass yield of the liquid and solid fractions compared with the control are in line with the findings of Grigatti et al. [22], who performed pot experiments with Italian ryegrass, except for LD1H, which had a very low biomass yield. The maximum above-ground biomass yield in our study was achieved in the treatment SD1H with 9.1 g per pot. As expected, the higher N fertilization level with 500 mg N per kg soil resulted in a higher biomass yield, with the exception of the treatments LD1H, UD1H and SDD2H, which showed a very low dry mass yield. This might be attributed to a delayed growth in the treatment LD1H due to the high moisture content in the soil substrate as a result of the low dry matter content of this digestate variant. This means that the low biomass yield of treatment LD1H was not caused by fertilization effects, but by an insufficient aeration of the soil. Comparing the dried variants SDD1 and SDD2 with the corresponding fresh variants SD1 and SD2, the results show a significantly lower above-ground biomass yield. Therefore, the results also reveal a negative impact of drying the digestate. Figure 3 shows the %N content of the above-ground biomass of the different treatments and the control, which was in a range of between 1.3% and 5.2%. was not caused by fertilization effects, but by an insufficient aeration of the soil. Comparing the dried variants SDD1 and SDD2 with the corresponding fresh variants SD1 and SD2, the results show a significantly lower above-ground biomass yield. Therefore, the results also reveal a negative impact of drying the digestate. Figure 3 shows the %N content of the above-ground biomass of the different treatments and the control, which was in a range of between 1.3% and 5.2%.  Table 1; n = 4, whiskers show standard deviation, different letters indicate significant differences, p = 0.05).

Nitrogen Content of Above-Ground Biomass
These values are within the range reported by Finck [47]. Neubert et al. [53] defined an N content between 2.9% and 5% as sufficient for barley plants at an early growing stage. Only treatments UD1H and LD1H showed a higher N content than 5%, with values of 5.1% and 5.2% respectively. Generally, the control and low-fertilization treatments showed values below the critical value of 2.9%, except LD1L with 3.2%. The high-fertilization treatments show an N concentration in the liquid treatments UD1H, LD1H, UD2H and LD2H above the critical value of 2.9% N, whereas treatments based on solid digestate (SD1H and SD2H), as well as the corresponding dried variants (SDD1H and SDD2H), were below the critical value. This indicates that the plants were not sufficiently supplied with N. For the variants with the dried solid digestate, this might be explained by ammonia volatilization during processing and the resulting loss of readily plant available NH4-N [11] so that the readily available N is not present in sufficient quantity for the plants and it needs time to mineralize N for plant uptake into an inorganic form. The results are supported by the findings of Möller et al. [18] and show that the liquid fraction is a suitable N fertilizer and the solid fraction should be used as a P fertilizer (see also Figure 3). Additionally, Grigatti et al. [22] describe  Table 1; n = 4, whiskers show standard deviation, different letters indicate significant differences, p = 0.05).
These values are within the range reported by Finck [47]. Neubert et al. [53] defined an N content between 2.9% and 5% as sufficient for barley plants at an early growing stage. Only treatments UD1H and LD1H showed a higher N content than 5%, with values of 5.1% and 5.2% respectively. Generally, the control and low-fertilization treatments showed values below the critical value of 2.9%, except LD1L with 3.2%. The high-fertilization treatments show an N concentration in the liquid treatments UD1H, LD1H, UD2H and LD2H above the critical value of 2.9% N, whereas treatments based on solid digestate (SD1H and SD2H), as well as the corresponding dried variants (SDD1H and SDD2H), were below the critical value. This indicates that the plants were not sufficiently supplied with N. For the variants with the dried solid digestate, this might be explained by ammonia volatilization during processing and the resulting loss of readily plant available NH 4 -N [11] so that the readily available N is not present in sufficient quantity for the plants and it needs time to mineralize N for plant uptake into an inorganic form. The results are supported by the findings of Möller et al. [18] and show that the liquid fraction is a suitable N fertilizer and the solid fraction should be used as a P fertilizer (see also Figure 3). Additionally, Grigatti et al. [22] describe similar results regarding the solid and the liquid Energies 2019, 12, 696 9 of 14 fraction and confirm the results that the N content in plants is higher when fertilized with the liquid treatments than with the solid ones. Figure 4 shows the %P content of the above-ground biomass of the different treatments and control, which was in a range between 0.4% and 0.9%.

Phosphorus Content of Above-Ground Biomass
Energies 2019, 12, x FOR PEER REVIEW 9 of 14 similar results regarding the solid and the liquid fraction and confirm the results that the N content in plants is higher when fertilized with the liquid treatments than with the solid ones. Figure 4 shows the %P content of the above-ground biomass of the different treatments and control, which was in a range between 0.4% and 0.9%.  Table 1; n = 4, whiskers show standard deviation, different letters indicate significant differences, p = 0.05).

Phosphorus Content of Above-Ground Biomass
Neubert et al. [53] defined %P content from 0.29% to 0.6% as sufficient for barley plants in an early growth stage. All treatments showed values above the critical value of 0.29%. That means that the P availability was not a limiting factor in the experiment. However, it is remarkable that the control showed the highest %P content. This might be explained by the low biomass development, where P took a larger relative share in the dry matter. In addition, the %P content in most treatments based on the solid fraction of separated digestate is higher than that of treatments based on liquid fractions, except SDDL1. Again, this might be caused by the lower biomass development, but also by the higher P content of those digestate variants. The results support the findings from other studies which show that mechanical separation of digestate into a liquid and solid fraction influences the subsequent use as fertilizer [12].

Nitrogen Removal from Soil Substrate and Plant Uptake Efficiency of Nitrogen
The nitrogen removal Nre reflects the combination of biomass yield and its %N content. All treatments showed a significantly higher N removal than the control, Figure 5. Treatments based on the solid fraction of separated digestate showed a lower N removal than the treatments with a liquid fraction (UD and LD) in most cases, with the exception of SD1H and the corresponding LD1H. The  Table 1; n = 4, whiskers show standard deviation, different letters indicate significant differences, p = 0.05).
Neubert et al. [53] defined %P content from 0.29% to 0.6% as sufficient for barley plants in an early growth stage. All treatments showed values above the critical value of 0.29%. That means that the P availability was not a limiting factor in the experiment. However, it is remarkable that the control showed the highest %P content. This might be explained by the low biomass development, where P took a larger relative share in the dry matter. In addition, the %P content in most treatments based on the solid fraction of separated digestate is higher than that of treatments based on liquid fractions, except SDDL1. Again, this might be caused by the lower biomass development, but also by the higher P content of those digestate variants. The results support the findings from other studies which show that mechanical separation of digestate into a liquid and solid fraction influences the subsequent use as fertilizer [12].

Nitrogen Removal from Soil Substrate and Plant Uptake Efficiency of Nitrogen
The nitrogen removal N re reflects the combination of biomass yield and its %N content. All treatments showed a significantly higher N removal than the control, Figure 5. Treatments based on the solid fraction of separated digestate showed a lower N removal than the treatments with a liquid fraction (UD and LD) in most cases, with the exception of SD1H and the corresponding LD1H. The reason might be the lower content of NH 4 -N as already explained above. The treatment SD1H with a high N removal of 232.9 mg per pot constitutes an exception, which is mainly governed by the high biomass yield of this variant. Another exception is the treatment LD1H, where the low N removal is caused by the low biomass yield as explained above. The results concur with the results of Grigatti et al. [22], that solid variants show lower N removal. In the experiments of the aforementioned study, a nutrient immobilization occurs with the solid treatments.
Energies 2019, 12, x FOR PEER REVIEW 10 of 14 reason might be the lower content of NH4-N as already explained above. The treatment SD1H with a high N removal of 232.9 mg per pot constitutes an exception, which is mainly governed by the high biomass yield of this variant. Another exception is the treatment LD1H, where the low N removal is caused by the low biomass yield as explained above. The results concur with the results of Grigatti et al. [22], that solid variants show lower N removal. In the experiments of the aforementioned study, a nutrient immobilization occurs with the solid treatments.  Table 1; n = 4, whiskers show standard deviation, different letters indicate significant differences, p = 0.05).
The nitrogen uptake efficiency (NUE) of the different digestate variants is shown in Table 5. Generally, the digestate variants of BP1 showed a higher NUE than the corresponding variants of BP2, with the exception of LD1H and LD2H, where LD1H has a very low above-ground biomass yield as a result of a delayed growth, as explained before, so that the NUE is, accordingly, 13%. Liquid digestate variants (UD, LD) showed a higher NUE than solid variants. These results correspond to the amount of NH4-N added to the soil substrate as shown in Table 4. For the liquid variants, the resulting NH4-N added to the soil via digestate was higher than the solid ones. The nitrogen uptake efficiency (NUE) of the different digestate variants is shown in Table 5. Generally, the digestate variants of BP1 showed a higher NUE than the corresponding variants of BP2, with the exception of LD1H and LD2H, where LD1H has a very low above-ground biomass yield as a result of a delayed growth, as explained before, so that the NUE is, accordingly, 13%. Liquid digestate variants (UD, LD) showed a higher NUE than solid variants. These results correspond to the amount of NH 4 -N added to the soil substrate as shown in Table 4. For the liquid variants, the resulting NH 4 -N added to the soil via digestate was higher than the solid ones.

Conclusions
The objective of this research was to investigate the N availability of dried solid digestate compared to other mechanically separated fractions of digestate and the untreated variants. The results show that an untreated biogas digestate is a suitable source of plant nutrients. Mechanical separation and drying of digestate decreases the NH 4 -N content in the solid fractions considerably. Due to the reduced amount of easily available N, the short-term N uptake of barley from solid fractions of digestate was low. The results show that the %N content in above-ground biomass was higher in the liquid and untreated variants than the solid and the dried solid variants. The results of %P show an inverse behavior. Furthermore, the mechanical separation of digestate effected the partitioning of plant nutrients according to their solubility in water, i.e., their prevalence in the liquid or solid fraction of the processed material.
The short-term N availability of the dried solid digestate, expressed as NUE, is lower than that of the other variants from the same biogas plant. However, this is not so when comparing the digestate obtained from the two biogas plants. Depending on the feedstock of the biogas plants and the processing method of the digestate, the N availability varies and influences the short-term N uptake. In summary, mechanical separation and drying of digestate is an essential step in digestate processing to increase the plant nutrient concentration and decrease transport costs for the solid fractions. In order to preserve the N in the digestate during drying, digestate processing should be combined with ammonia recovery to prevent N losses to the environment.