About Us | Help Videos | Contact Us | Subscriptions

Journal of Environmental Quality - Special Section: Environmental Benefits of Biochar

Simple Biotoxicity Tests for Evaluation of Carbonaceous Soil Additives: Establishment and Reproducibility of Four Test Procedures


This article in JEQ

  1. Vol. 41 No. 4, p. 1023-1032
    Received: Mar 31, 2011

    * Corresponding author(s): claudia.kammann@uni-giessen.de
Request Permissions

  1. Daniela Buschab,
  2. Claudia Kammann *ac,
  3. Ludger Grünhage and
  4. Christoph Müllerac
  1. a Dep. of Plant Ecology, Justus-Liebig Univ., Heinrich-Buff-Ring 26-32, D-35392,Giessen, Germany
    b Institute of Soil Biogeochemistry, Martin-Luther University, Halle-Wittenberg, Von-Seckendorff-Platz 3, 06120 Halle (Saale), Germany
    c School of Biology and Environmental Sciences, Univ. College Dublin, Dublin, Ireland. Assigned to Associate Editor James Ippolito


Biochar derived from pyrolysis has received much attention recently as a soil additive to sequester carbon and increase soil fertility. Hydrochar, a brown, coal-like substance produced via hydrothermal carbonization, has also been suggested as a beneficial soil additive. However, before soil application, both types of char need to be tested for potential toxic effects. The aim of this study was to develop simple, inexpensive, and easy-to-apply test procedures to identify negative effects of chars but not to provide false-negative results. The following tests, based partly on ISO norm biotoxicity test procedures, were chosen: (i) cress germination test for gaseous phytotoxic emissions; (ii) barley germination and growth test; (iii) salad germination test; and (iv) earthworm avoidance test for toxic substances. Test reproducibility was ensured by carrying out each test procedure three times with the same biochar. Several modifications were necessary to adapt the tests for biochars/hydrochars. The tested biochar did not induce negative effects in any of the tests. In contrast, the beet-root chip hydrochar showed negative effects in all tests. In an extension to the regular procedure, a regrowth of the harvested barley shoots without further nutrient additions yielded positive results for the hydrochar, which initially had negative effects. This implies that the harmful substance(s) must have been degraded or they were water soluble and leached. Tests with a biochar and hydrochar showed that the proposed modified quick-check test procedures provide a fast assessment of risks and effects of char application to soils within a short period of time (<2 wk).


    ANOVA, analysis of variance; BGG, barley germination and growth; EE0, unfertilized standard peat-based substrate; PAH, polyaromatic hydrocarbons; PCB, polychlorinated biphenyls; WHC, water holding capacity

Biochar has recently been proposed as a global change mitigation tool because of its recalcitrant nature and benefits when biochar is applied to low fertility soils (Gaunt and Lehmann, 2008; Hansen et al., 2008; Lal, 2009; Lehmann, 2007; Woolf et al., 2010). The idea originates in investigations of Amazonian dark earths or terra preta soils with high charcoal contents, which revealed that these soils are characterized by greater soil fertility compared with adjacent Ferralsols (Glaser, 2007; Glaser et al., 2001; Steiner et al., 2007). However, biochar applied to soils should be free of toxic substances (Amonette and Joseph, 2009; Bridgwater et al., 1999; Krull et al., 2009; Novak et al., 2009). For example, wood gasifier biochar is known to often carry larger amounts of polyaromatic hydrocarbons (PAHs) (Schimmelpfennig, 2010). Depending on feedstock composition and production conditions, biochar can contain polychlorinated biphenyls (PCB) or dioxins, and may acquire PAH by sorption processes (Smernik, 2009).

Moreover, for a recently emerging and new form of carbonized materials termed hydrochars (Libra et al., 2011) or hydrothermal carbonization biocoals, not much is known regarding their suitability as soil amendments and possible toxic contaminants. Hydrochar is produced by hydrothermal carbonization of dry or wet feedstock during several hours in liquid water under pressures and temperatures ∼20 bar and 200°C, respectively (Bergius 1913; Funke and Ziegler, 2010; Titirici et al., 2007); it was originally intended for energy use (as brown coal substitute).

Different feedstock, process conditions, and production techniques can result in large differences in biochar or hydrochar properties (Amonette and Joseph, 2009; Chan and Xu, 2009; Downie et al., 2009; Funke and Ziegler, 2010; Krull et al., 2009; Libra et al., 2011; Novak et al., 2009). To evaluate the chemical and physical properties of the chars and predict char effects on plant growth and soil function (Libra et al., 2011), standardized test protocols are required.

Biochars, and in particular hydrochars, can emit strong odors when freshly produced. It is not entirely known what gaseous species they emit (e.g., various volatile organic compounds) and whether these emissions have a toxic effect (Funke and Ziegler, 2010; Titirici et al., 2008). Thus, the cress germination test as a biological test procedure was selected. It investigates gaseous phytotoxic emissions (Kehres et al., 2006).

Some chars, in particular those produced at higher temperatures or from manures (Chan and Xu, 2009; Novak et al., 2009), can contain high ash contents, which may either supply beneficial micronutrients or macronutrients, or have a negative effect on seedling germination due to salt stress (Chan et al., 2007; Chan et al., 2008; Rondon et al., 2007) or presence of toxic heavy metals. A toxicity test will identify if dangerous toxic components are present. To test for toxic substances and salt stress, two different germination tests were selected: (i) barley germination and growth (BGG) test, where char is mixed into an organic medium (standardized, unfertilized peat substrate) and the plant is not sensitive to salt stress; and (ii) salad germination test (based on ISO–17126), where the char is mixed with an “inert” fine-sand medium, but where the germination of a salt-sensitive species (Lactuca sativa L.) is evaluated. Thus, negative effects at high application levels on salad germination, but not barley germination and growth, may indicate high ash contents but no harmful substances, whereas effects on barley germination would indicate potential toxic effects.

As a representative for testing effects of char addition on the soil fauna, the earthworm avoidance test (ISO–17512) was chosen because (i) it is simple and quick to conduct; (ii) the test species Eisenia foetida is very sensitive to a broad spectrum of different pollutants; (iii) this behavioral test method was as sensitive as a reproductive test, which takes longer (see ISO–17512 background; Hund-Rinke et al., 2005; Schaefer, 2003); (iv) it has been used for biochar testing (e.g., Chan et al., 2008; Van Zwieten et al., 2010); and (v) an increase in the earthworm activity in agricultural soils due to biochar presence may provide additive beneficial effects (e.g., Noguera et al., 2010).

All four test procedures were selected from a broad range of tests designed for compost testing or assessment of contaminated soils, and then adapted for the assessment of chars. We developed the tests with a biochar that was known to be nontoxic: (i) to identify effects arising from the test procedures that can potentially generate a false-negative result, thus providing an indication how to optimize/modify the test procedure, and (ii) to investigate if a positive result (i.e., char does not impart toxic effects) would be reproducible. The requirements for the test procedures are that they are general enough to be suitable for different char types, that they can be performed in a short period of time, and are easy to use. Thus, the primary goal of this study was to develop inexpensive, reproducible, and easy-to-adopt test procedures so that tests can be performed in many regions of the world, including developing countries with limited funds for research. Biochar technology is particularly suitable for small subsistence farmers.

Materials and Methods

Soils, Substrates, and Chars

For the salad germination test, fine-washed silica sand and coarse sand were used as growth/mixing medium and cover sand, respectively (see below). In the earthworm avoidance test, LUFA standard soil 2.2 was used (loamy sand, according to DIN and USDA classifications). In the cress and barley test, an unfertilized standard peat-based substrate (EE0) was used as control substrate.

For standardization, a toxic-free biochar (Eprida, Athens, GA, USA) was used that had been produced from peanut hull residues at 498°C maximum process temperature, 550°C steam activation, resulting in C and N contents of 71.6 and 1.84%, respectively, and an alkaline pH of 7.6 and 8.1 (1:4 in CaCl2 and H2O, respectively) (further details can be found in Kammann et al., 2011). Its concentrations of PAHs, PCBs, and dioxins were less than critical levels and mostly less than the detection limit. The hydrochar used in this study was produced from wet beet-root chip feedstock (residue following sugar beet processing) with a carbonization time of 2 h at 1.6 MPa and 203°C in water (Hydrocarb GmbH, Ohmes, Germany, now Revatec GmbH, Geeste, Germany). The hydrochar had C and N contents of 49.2 and 2.12%, respectively, a typical pH of 4.4 (1:4 in either CaCl2 or H2O; Libra et al., 2011), and a conductivity of 2.56 dS m−1 (1:20 in H2O). The hydrochar (water content ∼5.5–6 g H2O g−1 char) was used production fresh in the cress germination test but was dried at 105°C to a constant weight and then ground to <2 mm in all other tests (the fresh hydrochar started to develop mold quickly).

In the following, the original test procedures are described below with a focus on modifications that had to be made to enable char (instead of compost or soil) testing. Chars were ground to a particle size 1.5 to 2 mm by milling (1.5 mm) or passed through a 2-mm sieve with a pestle.

Cress Germination Test for Gaseous Phytotoxic Substances

The cress (Lepidium sativum L.) germination test has been designed, e.g., for compost testing (Kehres et al., 2006). In the original method, 0.50-g cress seeds are dispersed onto a wetted cotton pad and hung over a 250-mL container filled with the test material within a glass vessel with a lid. The EE0 substrate serves as control substrate. The lid was kept open (1-mm slit) with a tiny wire. The seeds were allowed to germinate for 7 d in the glass house at standardized conditions (18–20°C; 3.000 Lux for 12 h d−1). Based on the original guidelines, a cress fresh weight of >80% of the control indicates that the test was passed, i.e., no critical level of toxic gaseous components has been emitted (original: n = 2 replicates).

We replaced cotton pads with filter paper (5.5-cm diam.) to prevent the impact of impurities of the cotton pads. Instead of the poorly defined wetness of the test substrate as “hand wet,” the moisture was adjusted to 30% of the maximum water holding capacity (WHC) in control and char substrates. At a WHC of 30%, problems were mostly avoided, e.g., hygroscopic behavior of chars that were too dry or mold development when chars were too wet. The lid was closed over the entire incubation to prevent the escape of phytotoxic gases.

A metal soil corer collar (8-cm diam., 250 mL) was closed at the bottom, filled with 250 mL of char or control substrate EE0, both wetted to 30% of their maximum WHC, and enclosed in 1.1-L Weck jars (Fig. 1). A filter paper on a tripod table was connected to a water reservoir at the bottom of the vessel below the 250-mL collar; filter paper strips were attached to the filter paper. Fifty cress seeds were placed on the filter paper and moistened with distilled water. All jars were closed and put in a glass house for 7 d with constant light regimes and temperatures as described in the initial test guideline. After 7 d, germinated cress seedlings were counted, collected from the filter paper with their roots, dried from adhering droplets, and weighed. The lengths of 10 randomly collected hypocotyls were determined. All tests were performed using four replicates.

Fig. 1.
Fig. 1.

Sketch of the setup for the cress germination test (see text for further explanations).


Water contents of fresh hydrochars can range from ∼100% to several hundred percent (Libra et al., 2011). To compare biochar and hydrochar in this study, the wet production-fresh hydrochar was dried, its WHC was determined, and dried char was rewetted to 30% of its maximum WHC. In contrast to biochar from dry pyrolysis, water content of fresh hydrochar can vary greatly, depending on the initial feedstock (e.g., wood chips versus swine manure), production pathway, and post-treatment. However, gaseous phytotoxic substances may not be found if the hydrochar is dried and rewetted. Therefore, two test procedures for hydrochar are proposed: (i) production fresh but without process water; and (ii) dried and rewetted with deionized water to 30% of its WHC (i.e., WHC of a dry, rewetted hydrochar, which is different from a production-fresh hydrochar) (Libra et al., 2011). Combining both procedures should provide evidence if harmful substances impede seed germination initially and how quickly this effect disappears.

Barley Germination and Growth Test in Standardized Peat Substrate

The germination and growth test with barley is designed for the evaluation of fresh compost substrates (Kehres et al., 2006). It is based on mixing volume ratios of fresh EE0 and the respective compost in ratios of zero (pure EE0, control), 10, 25, and 50% by volume (based on the fresh substrates, i.e., with varying water contents). Three 500-mL pots are filled per mixture and 50 summer barley seeds per pot are sown and covered with a handful of the mixture. The mixtures are adjusted to a moisture described as “hand wet” and fertilized only once with 100 mL of a 7.33 g l−1 Florey–3 full-compound fertilizer solution (Florey–3: 15% N, 10% P2O5, 15% K2O, 2% MgO) per pot at the start of the incubation. Frequent watering is required during the test run but without any detailed instructions for target water content. The barley seedlings are cut at the soil surface and the number of shoots and fresh weight are determined after 10 to 12 d of germination grown under greenhouse conditions of 18 to 20°C and at a light regime of 12 h d−1 with 3.000 Lux.

However, after the very first test run (not shown), it became evident that the procedure had to be adjusted for testing chars. To avoid reduced barley growth due to variations in the water supply, a constant moisture supply was used. The following modifications were introduced: (i) lower percentage of char was used to be mixed with EE0 (0, 5, 10, and 25% char, instead of 0, 10, 25, and 50% compost); the 50% ratio was omitted and a 5% (by volume) mixing ratio was included to achieve a gravimetric mixing ratio that would be equal to a realistic field application level of 12.3 t ha−1 (e.g., for a char with a density of 0.4 g mL−1, e.g., of woody origin), related to the pot surface; (ii) mixing ratios were based on the dry weight ratios to ensure repeatable mixing conditions for a broad variety of chars; hence, the gravimetric water content of EE0, as well as the density of dried EE0 substrate and of chars, was determined beforehand; subsequently, the fresh weights (e.g., EE0) necessary to achieve mixings of 0, 5, 10, and 25 dry volume (%) were calculated and weighed into mixing vessels and mixed homogeneously by hand; (iii) to ensure consistent soil moisture throughout incubation, water content was adjusted to 60% of the maximum WHC (WHCmax) in each mixture and was maintained by daily watering; the largest error is associated with the biomass at the end of the incubation, which may account for 9 to 10 g of the total pot weight (or ∼ 2.3–2.6% of the average pot weight); (iv) we used 400 mL instead of 500 mL test substrate mixture per pot because EE0 reduces its volume when wetted and mixed, but most chars did not; rather, they sometimes swelled; and (v) five replicates instead of three were used to increase statistical power.

At the final harvest after 10 d, biomass fresh weight and dry weight (65°C to consistent weight) were determined to evaluate potential effects of chars on the plant water content. In one regrowth test with both chars, the harvested barley seedlings were allowed to regrow without further addition of nutrients but frequent watering. After 4 wk, the relative chlorophyll content (as a measure for N content) of the regrown shoots was measured with a SPAD 502 device (Minolta, Osaka, Japan) on eight leaves per pot and then shoots were harvested, weighed, dried, and reweighed.

Salad Germination Test in Fine Sand

This test is based on the ISO–17126 guideline to evaluate toxic substances and their effects in soils. The test substance (soil) was mixed with fine sand (0.4–0.8 mm) in (quadratically) increasing mixing ratios (e.g., 2, 4, 8, 16, and 32 g) in 100-g mixture (n = 3 per mixture treatment). Forty salad seeds were dispersed on 100 g of the fine sand (mixture) in large Petri dishes (15-cm diam., dispersal with 1-cm space to lip), wetted to 85% of its maximum WHC, and covered with 80 g of coarse (0.7–1.2 mm), dry sand. The dishes were placed in a greenhouse or growth chamber, and remain enclosed in 3-L plastic zipper bags to prevent evaporation. They were covered with dark, opaque plastic films during the first 48 h to induce germination. Subsequently, the sheets were removed and seedlings were allowed to grow in the light (4.300 lux, 12/12 h day/night cycle). After 7 d, the seedlings that emerged through the coarse sand were counted. Before the start and at the end of the test, pH and conductivity were measured on 50 g of the lowest (zero) and highest mixture level.

For salad seedlings on char-sand mixtures, an upper level of char addition had to be defined because chars with a low density cannot be accommodated in a Petri dish. The upper mixing level was therefore set to a maximum of 1.6:1 (vol vol−1) char:sand ratio. According to ISO–17126, WHC of the single components were determined and resulting water application amounts were calculated. The fresh weight of the aboveground seedling biomass in the hydrochar test was harvested with a sharp scalpel to determine if the aboveground fresh weight of the seedlings would provide a more sensitive measure than the number counts.

Earthworm Avoidance Test

The ISO–17512 guideline of the earthworm avoidance test aims to determine the quality of soils and effects of chemicals on the avoidance behavior of Eisenia foetida and/or E. andrei in a two-compartment vessel (ca. 20-cm diam and 5-cm height). Fivefold replicated vessels were used and the soil adjusted to 60% of its maximum WHC. Control and effect (char) side must be marked on each vessel. At the start, 10 mature clitellate animals were set on the line separating the halves, the lid was closed, and vessels were incubated at 18 to 20°C at a day–night cycle of 8 to 16 h in climate chambers.

After 2 d, a separator was inserted again along the middle line and vessels were turned upside down to count the number of animals in both sides. The percentage of avoidance was calculated according to ISO–17512:where Xaviod. is the avoidance in percent, nc is the number of worms in the control soil (either per vessel or in all five replicates), nt is the number of worms in the test soil (as above), and N is the total number of worms.

As vessels, metal boxes (18.2-cm diam. and 7-cm height; GeckoPac, Michelstadt, Germany) were used. A quadratic window of 10 cm by 10 cm was cut into each lid and closed with a fine mesh to allow aeration and light but prevented animals from escaping. Standard soil (soil 2.2, LUFA Speyer, Germany; recommended by ISO–17512) was used. The soil was a loamy sand (note: soil with a high clay and low soil organic matter content can lead to large worm mortalities; Natal-da-Luz et al., 2008), which was characterized by a pH of 5.5, organic carbon and nitrogen contents of 1.93% and 0.17%, respectively, and bulk density of 1.247 g cm−3. The worms were always collected shortly (1–2 d) before the test was performed on suitable sites, where high population densities of Eisenia existed (e.g., horse manure compost piles) and kept in the laboratory at room temperature in their original substrate.

Two application levels were used with a depth of 7 cm, 10 and 30 t ha−1. These amounts corresponded to 1.4 and 4.3%, respectively, on a weight basis, when added to the selected soil (LUFA standard soil 2.2, bulk density 1.247 g l−1). The chosen char application rates represent upper limits of reasonable field application rates. The maximum WHC was determined for all char–soil mixtures and for pure soil.

Reproducibility was tested with four different treatments: (i) true control where the same unamended soil was in both vessel halves; (ii) positive control where 50 mg of boric acid powder was applied to the effect side (Stegger et al., 2011); (iii) 10 t ha−1 amendment with the peanut hull biochar used for the reproducibility test; and (iv) 30 t ha−1 amendment with the peanut hull biochar used for the reproducibility test. Treatments 1 and 2 were included to verify that the test procedure was successful. Required amounts to fill a vessel half full were calculated from the standard soil’s bulk density and weighed into a 2-L aluminum pan. For the effect side, soil was mixed with char (12.7 and 38.2 g for 10 and 30 t ha−1, respectively). Then, the pure soil and char-amended soil were poured into the two halves of a vessel and divided by a separator made from polyethylene (7 cm by 18 cm by 0.25 mm). Subsequently, on both sides, the necessary amount of water was slowly applied across the entire surface, the separator was removed, and 10 mature worms were placed along the middle line. The animals were sprayed once with a water atomizer to enable them to disappear quickly into the soil.

The formula used in ISO–17512 (Eq. [1]) was modified slightly by multiplying the result by −1, which would provide a negative value that is more intuitive for avoidance behavior.


A t-test was used to test for significant effects in the cress germination test between EE0 control and char within the same test run. In the barley germination and salad germination growth tests, one-way analysis of variance (ANOVA) and the (posthoc) Student−Newman−Keuls test were used to evaluate if the applications of increasing amounts of char affected germination or growth, and to identify char addition levels that were different from each other. For the earthworm avoidance test, the Fischer’s exact test was used to investigate if the behavior of the animals was affected by char amended to one side of the vessel but not to the other (ISO–17512). A result was considered significant at p < 0.05 and is reported as a tendency with p < 0.1. All statistical tests were performed with SigmaPlot (Systat Inc., Chicago, IL, USA).


Cress Germination Test for Gaseous Phytotoxic Emissions

Based on the initial test procedure, the test would be considered to be passed when the cress fresh weight is >80% of the control (Kehres et al., 2006). Every water regime but the first (no water application, original procedure) would have resulted in a cress fresh weight >80% of the control (Fig. 2), i.e., the test would have been passed. In all tests, the germination rate was >95%.

Fig. 2.
Fig. 2.

Cress fresh weight (A) and hypocotyl length (B), means + standard deviation (n = 4); asterisks indicate significant difference between control and char within a test run (= tr). The numbers in the gray bars in A give the char mean as percentage of the control. tr 1: without extra water supply; tr 2: water drips added through a septum every 2 d; tr 3: water drops added through a septum daily but not during the weekend; tr 4 and 5: permanent watering (see Fig. 1). BC = biochar (peanut hull); HC = hydrochar. HC was tested at three different ages: production-fresh (7-d old) wet, dried and rewetted, and dried hydrochar 2 wk older and rewetted.


The fresh hydrochar completely prevented germination of the cress seeds (arrows in Fig. 2). In addition, fungal growth was observed on the char surface at the end of the incubation, which also occurred when the hydrochar was dried first and then rewetted. With the dried and rewetted hydrochar, 36 (out of 200) seeds in two out of four jars germinated; the seedlings were too small for hypocotyl measurement (Fig. 2, right). However, when the dried hydrochar was stored for additional 2 wk at 22 ± 2°C in closed plastic vessels, the cress fresh weight was not significantly impacted, but the hypocotyls’ length still was significantly reduced (Fig. 2, right).

Barley Germination and Growth Test

In all three repeated test runs, peanut hull biochar did not significantly impact barley germination or growth (fresh or dry weight), and none of the biochar additions resulted in a significant reduction or increase (Fig. 3). When the data of the three test runs were pooled and analyzed with a two-way ANOVA, the factor “test run” was significant (p < 0.001); for “fresh weight” as second factor, biochar addition did not have a significant effect (p = 0.123); but for “mean seedling weight” as second factor, biochar had a significant positive effect on seedling fresh weight (p = 0.001; no significant interaction with “test run” for either combination). However, the absolute increases in seedling weight were small (+7, +13, and +7% in the 5, 10, and 25% biochar treatments, respectively), i.e., the largest positive effect was observed in the +10% volume biochar addition.

Fig. 3.
Fig. 3.

Barley germination and growth test results: germination (A), fresh weight (B), and dry weight (C) of three repeated test runs (tr) and one run with a beet-root chip hydrochar (together with tr 2); means plus one standard deviation are given. Different letters indicate significant differences between char addition rates within the same test run.


In contrast, hydrochar reduced germination even in the lowest application rate. At the 5% level, the seedling weight was not significantly lower than the control seedling weight (not shown), whereas the fresh and dry weights were both significantly reduced (Fig. 3B, C). With larger hydrochar applications, the reduced fresh and dry weights were a result of both reduced germination (Fig. 3A) and reduced seedling weight. No significant effect on shoot water content was observed.

After test run 2, all control, biochar, and hydrochar pots were left to regrow and were harvested again (Fig. 4). Due to a die off of some seedlings, the variability became larger than in the first growth. At the second barley harvest, the fresh weight tended to be greater with biochar (144–176% of the control), although the leaf-N concentration was significantly less, in particular, with 10% volume addition. For the second harvest, the hydrochar significantly promoted barley growth (Fig. 4A), although none of the nongerminated seeds germinated at this stage. Instead, the remaining live plants grew stronger (e.g., 5% hydrochar led to 261% of the control dry weight with fewer plants, compare Fig. 3A) and had larger leaf-N concentrations (Fig. 4B).

Fig. 4.
Fig. 4.

Proposed extension of the barley test: regrowth without further fertilization; fresh weight (A) and relative chlorophyll content (B) of regrown barley seedlings 3 wk after tr 2 with either 0, 5, 10, or 25 vol. % biochar (BC, dark gray), or hydrochar (HC, light gray) mixed into the unfertilized standard peat-based substrate (EE0). The dashed lines mark the respective control values. Different letters indicate significant differences (one-way analysis of variance [ANOVA], least significant difference [LSD] test) between the control (ctrl) and biochar, or hydrochar mixtures, respectively.


Salad Germination Test

The repeatedly tested peanut hull biochar did not reduce or increase the germination of Lactuca sativa in either of the three repeated test runs, even at a rate where one-third (32 g) or 1.6:1 (vol vol−1) of the mixture was char:sand (Fig. 5). The hydrochar, in contrast, reduced germination significantly, even at the lowest application rate and was reduced to zero at the highest hydrochar application rate (Fig. 5). Determining the fresh weight of the salad seedlings in the hydrochar test gave a more sensitive indication than germination of the emerging negative effects with increasing application rates. While the germination was reduced to 82% in the lowest (2 g) addition, the fresh weight was reduced to almost zero because the seedlings hardly penetrated the surface and were too small for harvesting; the same was true in all higher hydrochar additions.

Fig. 5.
Fig. 5.

Salad germination test: germination rate of three repeated test runs (tr) and one run with a beet-root chip hydrochar with additions of 0, 2, 4, 8, 16, and 32 g to 100 g of fine sand; means plus one standard deviation. Different letters indicate significant differences (one-way analysis of variance [ANOVA]) among application levels within the same run. HC beet = beet-root chip hydrochar.


Earthworm Avoidance Test

In all four test runs, the earthworm behavior was completely repeatable in the “true” control (i.e., soil 2.2 on both vessel sides, without further amendment) and there was no significant effect of the “vessel side” (p > 0.05). The positive control with boric acid on one side was repeatable and always showed a significant avoidance (Fig. 6). In none of the four test runs did earthworms show avoidance toward the peanut hull biochar-amended side. However, in three out of four test runs, the earthworms significantly preferred the side with the peanut hull biochar over the control side (Fig. 6). In all tests with the peanut hull biochar, the addition of 10 and 30 t ha−1 (1.4 and 4.3% wt/wt) showed a similar preference (Fig. 6, Fig. 7). However, the earthworms significantly avoided the hydrochar side at both application levels (Fig. 7).

Fig. 6.
Fig. 6.

Earthworm avoidance test performed on four subsequent dates, using earthworms (Eisenia foetida) freshly extracted before the test from horse-manure composting sites. Mean counts per vessel half (n = 5) on the effect side (hatched) or control side (plain), plus standard deviation. Boric acid was used as a positive control and biochar applications of 10 and 30 t ha−1 (7-cm depth) were used. Symbols: results of the Fisher’s least significance test (ns, not significant; * p ≤ 0.05; ** p ≤ 0.01; *** p ≤ 0.001).

Fig. 7.
Fig. 7.

Proposed visualization example for char testing with the earthworm avoidance test: including “preference” in addition to “avoidance” (results of control, boric acid and BC taken from Fig. 7;HC = hydrochar). Symbols: results of the Fisher’s least significance test (ns, not significant; * p ≤ 0.05; ** p ≤ 0.01; *** p ≤ 0.001).



The reproducibility of all established test procedures obtained with a nontoxic biochar was satisfactory. The finding “no negative effect” was always observed in the repeated test runs. Several modifications were made to adapt the original procedures to the testing of biochars and hydrochars. Care was taken to adjust the water supply protocol in the cress and barley tests, and ensure that the testing protocols would fit both biochar and hydrochar. Reproducibility could not be investigated with hydrochar since it showed ageing effects (cress test); thus, results changed over time, as discussed below.

Cress Germination Test

The cress germination test (Kehres et al., 2006) was modified to avoid producing a false (“toxic gaseous emissions”) test result. The biochar was produced 1 yr before use in test run 1 (Fig. 2). Hence, volatile substances may have disappeared. Furthermore, in another study where dry biochar was rewetted, this only generated brief emission peaks in the first few hours after application for CO2, CH4, and N2O (Kammann et al., 2012), indicating that volatile organic carbons, for which CH4 may be an indicator, were likely not continuously emitted. Moreover, separate analyses showed that peanut hull biochar was practically free of toxic compounds, such as PAH, PCB, or dioxins. Hence, other causes than toxic substances had to be responsible for the observed negative effect of the biochar. We observed a gradual alleviation of the negative effect with a higher continuous water supply. This all points to a moisture effect, rather than a toxic effect.

Biochars can be hydrophobic when they are fresh; with age, they develop more functional groups and hence lose the initial hydrophobicity (Cheng et al., 2006; Lehmann et al., 2009). However, several different biochars and, in particular, hydrochars changed from hydrophobic to hydrophilic behavior when they were wetted (unpublished observations). The latter likely occurred with the biochar used here. The water that was provided to the biochar (30% of the maximum WHC) may have not been sufficient to fill the microspores and mesopores (<50 nm; Downie et al., 2009), and so the char absorbed water vapor from the jar atmosphere. This is in line with the observed filter paper drying on which the seeds germinated and in which we tried to compensate with the watering regimes. A solution would have been to set moisture to similar water potentials, but this would have required information on the biochar moisture characteristic (i.e., matric potential versus volumetric water content). Another solution would have been to increase the water supply, but this might have intensified the mold problem we encountered with the hydrochar. The reason filter papers did not become as dry with the hydrochar was likely to be its much lower porosity and lower surface area (Libra et al., 2011).

Thus, the negative effect observed with fresh hydrochar was most likely associated with the emission of phytotoxic volatile substances, such as formic or acetic acid (Titirici et al., 2008; E. Röcker, personal communication), which prevented germination. However, the inhibitory effect quickly vanished with age and when the hydrochar was dried (Fig. 2). This is in accordance with pure hydrochar, which contained lower volatile substances as compared with older hydrochar (Kammann et al., 2012). In summary, we conclude that the cress germination test is suitable for detecting harmful phytotoxic emissions but also their decline in freshly produced chars when a possible hygroscopic behavior of a (bio)char is accounted for. However, more work is needed to solve the watering versus mold problem to obtain standardized protocols, and to identify the phytotoxic gaseous emission from the hydrochar.

Barley Germination and Growth Test

In the barley germination and growth test, problems similar to that observed with the cress test (a physiological drier substrate) were overcome by adjusting WHC of the substrate–biochar mixtures. Subsequently, the three test runs with the same biochar did provide statistically reproducible results for all measured parameters.

Even when the peanut hull biochar contributed 25% of the volume of the growth mixture, no negative effect was observed (Fig. 3). The same biochar did significantly increase WHC in a poor sandy soil and stimulated the yield of crop quinoa (Chenopodium quinoa Willd.) significantly at “sufficient water supply” and “moderate water stress” treatments, respectively (Kammann et al., 2011). However, since: (i) BGG test was much shorter (10 d), (ii) WHC of the peat substrate exceeded that of the biochar, and (iii) a full-fertilizer nutrient solution was applied at the start, no significant increase in the barley biomass was observed with increasing biochar application (Fig. 3).

In contrast, the hydrochar caused a significant reduction in the germination and growth responses with increasing concentrations. Rillig et al. (2010) also observed significantly reduced growth of Taraxacum sp. with stepwise-increasing hydrochar rates from 10 to 80% (vol vol−1). The reduction in shoot weight observed by Rillig et al. (2010) was between 20 and 30% hydrochar additions, similar to the reduction observed here (25% vol vol−1 dry). However, the authors also did not wash the hydrochar to remove potentially harmful substances.

In the BGG test, the fresh/dry weight was the most sensitive parameter to indicate negative hydrochar effects (comparing statistical results in Fig. 3), which is in line with observations by Bagur-González (2010) that root elongation (usually well correlated to root biomass) was more sensitive to low-degree contaminations than germination.

There was a tendency that regrowth was stimulated by biochar and significantly stimulated by hydrochar. In the biochar treatments, this was possibly associated with the nutrient-retention abilities (Atkinson et al., 2010; Ding et al., 2010; Liang et al., 2006; Singh et al., 2010; Steiner et al., 2010; Steiner et al., 2008). However, such stimulation was unexpected for the hydrochar, which has shown prior signs of a strong toxic effect at levels of 10 and 25% hydrochar mixing ratios. Furthermore, the regrowth of barley plants on hydrochar mixtures had higher relative chlorophyll contents, providing evidence that nutrients associated with the hydrochar must have been responsible. Due to the lower carbonization temperature and watery medium, hydrochars will usually retain more nutrients in a plant-available form than the same feedstock pyrolyzed to biochar (Libra et al., 2011). However, it is unclear if “washing” or aging had removed toxic substance(s), enabling the beneficial effects of the hydrochar to emerge. This observation prompts further detailed investigations.

Salad Germination Test

The results of the salad germination test were also fully reproducible (Fig. 5). The peanut hull biochar did not reduce the germination of the salad seedlings, even at application rates that would never be applied in the field (i.e., ratios of biochar:sand at 1.6:1). In contrast, hydrochar inhibited the germination of Lactuca sativa seeds, even at the lowest application rate of 2%. Thus, this test was more sensitive to the negative, unknown substance(s) present in the hydrochar than the BGG test. Although salt stress may have been involved (to which BGG is likely to be less sensitive), this cannot be the only cause; the conductivity of the hydrochar was not particularly high. However, this hydrochar contained five to six times its own weight in the form of process water, which dried on the char. Therefore, it is possible that the hydrochar contained osmotically active substances other than salts, such as dissolved organic carbon, which can be elevated in the process water (Libra et al., 2011).

Furthermore, with hydrochar, the seedling fresh weight was the more sensitive parameter (see Barley Germination and Growth Test; Bagur-González et al., 2010). Hence, although this is not part of the initial ISO–17126 guideline, seedling fresh weight (total or aboveground clipped) may serve as a sensitive parameter for the detection of biotoxic substances in char testing.

Earthworm Avoidance Test

The earthworm avoidance test has been most frequently used to assess the ecotoxicological effects of biochar (e.g., Chan et al., 2008; Van Zwieten et al., 2010). The earthworm avoidance test has several advantages over reproductive tests with soil animals. It is often more sensitive to various contaminants, not too sensitive to soil type, and is easy to use within a short time period (Hund-Rinke and Wiechering, 2001; Hund-Rinke et al., 2005; Hund-Rinke et al., 2003; Loureiro et al., 2005; Natal-da-Luz et al., 2008; Schaefer, 2003). Our results demonstrated that the test was completely reproducible in four different test runs with the same biochar and animals taken from different field sites. The biochar results were reproducible with regard to: (i) true control, where no effect occurred as required by theory, (ii) toxic effect of boric acid that always induced a significant avoidance reaction (demonstrating that the test/animals functioned as they should; see also Stegger et al., 2011), and (iii) confirmation that the biochar used here never produced an avoidance reaction in the earthworms.

However, it was surprising that in three out of four test runs, Eisenia significantly preferred the biochar-amended side over the unamended control side. This is in line with Chan et al. (2008) who observed a preference of earthworms for a nonactivated poultry litter biochar but not for an activated biochar that had been produced at a higher temperature. Also, Van Zwieten et al. (2010) reported that earthworms significantly preferred a papermill waste biochar applied to acidic Ferrosol with an associated pH increase of 1.2 to 1.7 units. However, for the same biochar, no preference was observed when applied to a slightly alkaline Calcarosol. In contrast to the Ferrosol, the Calcarosol did not change its pH after biochar addition. This suggests that a pH effect may have caused the earthworm reaction to biochar in soils, provided that no harmful substances were present (Hund-Rinke et al., 2005; Loureiro et al., 2005; Saterbak et al., 1999). However, the loamy sand we used was less acidic (pHCaCl2 5.5) and the biochar less alkaline than the substrates and biochars used by Chan et al. (2008) or Van Zwieten et al. (2010). Thus, it is unclear if the preference behavior is simply a reaction to a pH change of alkalinizing biochars.

Furthermore, the earthworm reaction may be related to their physiological state. While Eisenia showed only a weak preference in the first test run in spring when animals had been collected after a long-lasting winter (frost) period, a much more pronounced preference behavior was observed when earthworms were collected in summer when they were (likely) more active. Thus, more work is required to understand if Eisenia will show annual seasonality in its preference for biochar, which might be related to their physiology and activity state. It is important to evaluate how reliable and comparable “biochar preference” results are throughout the season. However, this did not impact the original aim of the test to identify if a toxic substance was present.

As with the other tests, hydrochar caused significant avoidance in the earthworms in both the 10 and 30 t ha−1 treatments, with avoidance increasing with hydrochar amount. The lower pH of hydrochar (4.4) than biochar (7.6) would fit into the pH theory outlined above. However, even if the pH had a negative effect, it cannot be excluded that toxic substances were responsible. Physiological drought (desiccation) caused by the char, as observed by Li et al. (2011), cannot be the reason for the avoidance reaction because we used the same WHC, determined beforehand for both the control and char-amended soil (as demanded by the ISO guideline). Thus, neither the demonstrated earthworm preference nor avoidance for the biochar- or hydrochar-amended soil should have been caused by hydrological alterations due to char addition.


Before biochar or hydrochars are applied to soils, it is necessary to test these soil additives for unwanted or toxic ingredients. We have developed a range of quick tests that would help in the evaluation, based on standard (ISO) test procedures. With certain necessary modifications, all methods were suitable for char testing and reproducible in a nontoxic biochar. The use of a known nontoxic char ensured that there were no “pseudo-toxicity” effects caused by the experimental settings, such as the water supply. Particular care must be taken with the cress test when using quickly molding hydrochars to ensure that the equipment is not contaminated. Disposable sterile equipment may be useful. All four tests agreed in their assessment of the hydrochar having negative effects. For hydrochar, the BGG and salad tests showed that seedling fresh weight was the more sensitive parameter than germination. The BGG test, however, revealed that the interaction with soil organic matter would not alleviate the hydrochar-associated problem. However, despite the negative test results obtained for hydrochar, it displayed a potentially positive effect on soil fertility when the BGG test plants were allowed to regrow. Hence, allowing regrowth with excess watering and zero fertilization may become a valuable test extension to identify the potential of the respective char as a soil amendment.

Testing with a large range of chars is required to verify the methodology or assess a range of potential toxic effects. Physicochemical properties should also be determined to deduce from combined test results which kinds of toxic substances may be present. Finally, a closer inspection of dose-response relationships, in particular with harmful chars, should be used to assist in defining char categories with regard to their safe use in soils.


The authors are grateful for the help of Angelika Bölke, Gerhard Mayer, and Jürgen Franz in the laboratory and greenhouse, and for the preliminary procedure testing and method-discussion efforts. Thanks to Diedrich Steffens for accommodating our earthworm tests in his climate chambers whenever needed. The authors gratefully acknowledge the financial support of the Hessian Agency for the Environment and Geology.




Be the first to comment.

Please log in to post a comment.
*Society members, certified professionals, and authors are permitted to comment.