Phytooprensning af metaller

Summary and conclusions

As a part of the development and testing of technologies in the "Programme for Development of Technologies", initiated by the Danish EPA (Miljøstyrelsen), techniques for remediation of metal contaminated soils are in focus. Under Danish conditions, methods for remediation of moderate concentrations of several metals in mixture and lead contaminated soils are most needed. The methods must also be feasible for soils with a high content of clay.

In 1997, a project, reviewing the possible technologies in relation to Danish conditions with respect to handling and treatment of metal contaminated soil pointed to phytoextraction as one of the feasible technologies with a promising potential.

On this basis, the Danish EPA (Miljøstyrelsen) has given preliminary acceptance of testing of the phytoextraction technology on 4 metal contaminated sites located in the counties of Bornholm, Funen and Ringkøbing and in the municipality of Copenhagen. Thus, a growth chamber project involving soil samples from all four sites was initiated. The aim of the project was to establish an improved basis, partly regarding phytoextraction in practice and selection of plants that are feasible under Danish conditions, partly regarding the importance of soil conditions and the characteristics of the contaminants for the efficiency of the technology. Considering the possible subsequent testing of the technique in the field, the project was undertaken as a fast short-term project. This report presents the results of the growth chamber project carried out by VKI in collaboration with the Danish Institute of Agricultural Sciences.

In the introductory parts of the report, the potential for use and improvement of the method under Danish conditions is discussed in addition to the principles for remediation of metal contaminated soil by plants. This part of the report is based partly on the previously undertaken review of technologies for remediation of metal contaminated soils in Denmark but updated with recent and in some cases not published knowledge.

Based on the experiences from literature and through contacts, primarily international, but also Danish research groups and plant breeders, 9 plants were collected. Four of these plants and 2 soil samples from each of the 4 sites were used for growth experiments. The 4 plants were Thlaspi caerulescens, Brassica juncea, Salix burjatica "Germany" and Amaranthus retroflexus (two different ecotypes). The 4 selected plants differ with respect to appearance, growth, growth period and characteristics in relation to subsequent harvesting and treatment of the biomass.

After sampling, the soil samples were characterised for general soil parameters such as pH, texture, plant nutrients and contents of the trace elements arsenic, lead, cadmium, copper, chromium, nickel and zinc. This characterisation showed that the soil samples were representative for a range of different soil types with respect to texture, pH, content of calcium carbonate, organic carbon and plant nutrients. Regarding content of metals, the samples are typical for many of the contaminated sites found in Denmark, both in terms of level of concentration and composition. Nickel was the only metal not found in concentrations exceeding background level. Concentrations of each of the metals varied over the sites within the ranges: Arsenic, 4–1800 mg As/kg; lead, 83-1100 mg Pb/kg; cadmium, 0,2–4,7 mg Cd/kg; copper 7-11000 mg Cu/kg; chromium, 6-1000 mg Cr/kg; nickel, 3-41 mg Ni/kg; zinc, 29-2700 mg Zn/kg.

The experiment was carried out in growth chambers over a period of 2-3 months. The plants were sown or planted in containers with a volume of approximately 30 litre (depth 30 cm and surface area 0.1 m²). The growth chambers were run according to a fixed 24-hour schedule throughout the growth period. The applied day and night temperatures corresponded approximately to Danish July mean-maximum and mean-minimum temperatures.

Generally, the plants grew satisfactorily, except for the plants grown in soil samples from Funen County that were highly polluted with arsenic and copper. Here the 4 selected plants either died or showed poor growth, so, additional experiments with Agrostis capillaris were carried out. Agrostis capillaris is found to be tolerant to arsenic. This plant grew satisfactorily in the samples highly polluted with arsenic and copper.

After harvest, the plant material was weighed. Subsequently the root stock of the plant, in some occasions also the root material, were analysed for the content of the 7 trace elements. The yields of Salix burjatica and Brassica juncea were generally of the same order of magnitude whereas the yields of the other plant species were a factor of 4-5 lower. The largest fraction of root material compared to top material was obtained from Brassica juncea. For Brassica juncea the fraction was 16-24% (weight) and for Salix burjatica less than 10%. Regarding Thlaspi and Amaranthus, the root yields were small but made up more than 10% of the top material. Agrostis capillararis had a very high fraction of roots because it has a fibrous root system.

The analysis of plant material showed large variations between the concentrations of the metals in both the plant species and the soil samples. This verified that certain plants may take up metals found in elevated concentrations in the soil. The uptake of metals found in elevated concentrations in the soil compared to the insignificant uptake of metals from the soil with low concentrations indicates that the experiments, the treatment and the analysis of the soil and plant samples have worked as intended.

The results document a certain uptake of arsenic in Agrostis and Brassica grown on the soils highly contaminated with arsenic. In some of the other samples, detectable concentrations of arsenic were found in the plant material, however only in concentrations less than 1 mg As/kg and thus insignificant in relation to the much higher concentrations found in the soil.

The concentrations of Cd in many of the plant samples were interestingly high. Salix and in some cases Amaranth contained high concentrations of cadmium, typically higher (up to 10 mg Cd/kg plant material) than the concentrations found in the soil. In plant material of Thlaspi from two of the sites, more than 100 mg Cd/kg were determined. These concentrations are equivalent to a accumulation of approximately a factor of 100 in relation to the concentrations in the soil. The highest concentrations of Cd were found in plant material coming from the soil samples containing the highest concentrations of Cd. It is known that the distribution of cadmium between soil and soil solution and thus the plant uptake is very dependent of pH. This is probably the explanation to the relatively high concentrations found in Thlaspi (16 and 36 mg Cd/kg) grown on the soil samples from Ringkøbing County. These samples contain cadmium at background level but have very low pH-values at approximately 4.5.

Copper and partly chromium are found in measurably elevated concentration in some of the plant samples, but the concentrations are much lower than the concentrations in the soil.

Since the concentrations of nickel are relatively low in the soil samples, the potential for testing the remediation of nickel is low. As expected, it is Thlaspi caerulescens, known to have the ability to hyperaccumulate nickel, that takes up the highest concentrations of nickel from the soils containing the highest concentrations of nickel.

In Denmark lead is the most abundant metal at polluted sites. It is known that lead has a very strong binding and low availability in the soil. In the experiment the highest concentrations of 158 and 177 mg Pb/kg plant material were found in the Amaranth grown on soil from Ringkøbing County. The concentrations were approximately 4 times higher than the soil quality criteria for lead (40 mg/kg) and the experiments demonstrated uptake from a plant, which is relatively unknown in this context. It is also worth noting that Salix took up lead in concentrations up to 40 mg Pb/kg plant from the soil samples from Ringkøbing County and gave a significantly higher yield than Amaranth. As previously mentioned, all soil samples contained elevated contents of lead but the highest uptakes by plants were not directly correlated to the concentrations in the soil. Again, the low pH-values in the soil samples from Ringkøbing County are presumably the explanation for the higher availability of the metals from these samples.

Zinc was the metal taken up in the highest concentrations, which was to be expected, since zinc is an essential element for the plants and abundantly available in the soils. Thlaspi caerulescens, that can hyperaccumulate zinc, contained concentrations of 1800 to 8200 mg Zn/kg plant material. High concentrations were also found in Salix and Amaranth, up to 1400 and 1100 mg Zn/kg respectively. Despite the significantly lower concentrations of zinc in the soil from Ringkøbing County, the corresponding concentrations in the plants were not low compared to other samples. Again, the reason, as discussed for cadmium and lead, must be related to the low pH-values in the soil that increased the plant availability of the metals.

The concentrations of metals in the selected root samples are mostly interesting with respect to lead because the highest concentrations (500-1000 mg Pb/kg) are approximately a factor of 5 higher than the highest concentrations of lead in the shoot material. The concentrations of arsenic, chromium, copper and nickel were moderately higher in root samples than in shoot samples. However, the concentrations are not sufficiently high to signify any contribution to the total uptake of metals from the soil. Generally, cadmium and zinc were found in slightly lower concentrations in the root material than in the shoot material. Presumably this is related to an effective translocation of these metals from the roots to the green parts of the plant.

Based on the knowledge of the produced plant biomass and the concentrations of metal in the plant material and the soil, the relative removal of each of the metals was estimated. The removal is given in percent in relation to the fraction of the content in soil that exceeds the soil quality criteria. Generally, the estimate only includes the metal found in the above ground plant material, but where possible the content in root material was included.

Relative to the concentrations in the soil, cadmium is the metal taken up (accumulated) in the largest quantities by the plants. The removal of cadmium was for Thlaspi caerulescens typically 3 to 4%, and up to more than 9% were removed by this single harvest carried out after one growth season. Salix is the other plant which is able to take up cadmium in quantities exceeding 0.1% of the content in the soil. However, the removal by Salix did not exceed 1% of the content in the soil in any of the experiments. Thus, it seems feasible by use of Thlaspi, possibly in combination with Salix, to carry out a remediation under field conditions. Besides cadmium, zinc is the metal with the best possibility or successful remediation by phytoextraction. The plants are capable of taking up high concentrations of zinc, but zinc is also found in high concentrations in the soil. This means that the relative removal is low. Only Thlaspi and Salix removed zinc in quantities equivalent to 0.1-0.6% of the soil contents. All other plants revealed relative removals of 0.1% or less. Since zinc is toxic for humans and plants only at very high concentrations, this element will presumably only very rarely determine the need for remediation of a contaminated site. Under all circumstances, it is concluded that the experienced relative removals of zinc are too low to be directly useful for practical remediation purposes.

The results of the experiment carried out over a period of 2 to 3 months can not directly be applied to field conditions. However, it is likely that the harvested biomass per area within a factor of 2 will correspond to the yield expected from one growing season in the field. The experiments provide an estimate of the expected removal of metal from the soil within the first growth season. Estimates of the removal over several years involve extrapolations of the results. In the literature linear calculations (same annual removal per year) have been used to estimate the necessary remediation period for phytoextraction. This assumption is probably fair for the first years of remediation, but hereafter the availability of the metals is likely to decrease as the most available metal is removed.

With respect to cadmium, it is likely that effective remediation by phytoextraction may be carried out over a period of maximum 10-20 years. For all other metals investigated, the perspectives in terms of necessary remediation periods are significantly longer.

Since the direct uptake of other metals was limited, it is obvious to look into the possibilities for further development of the technology, which has the advantage of being a low-tech, "green" and relatively low-cost technology.

In the long term, plants with higher metal concentrations and larger biomass yields will be developed, which will lead to considerable higher rates of removal than the results of these investigations have shown. Optimising the uptake in plants by manipulation of the soil and soil solution is another possibility, which may be further developed. Based on the experiments carried out, it is shown that metals such as cadmium and zinc, probably controlled by sorption processes in the soil, at low pH-values may be removed from the soil down to very low concentrations. For these metals it can be concluded that adjustment of pH is one possibility for optimising the uptake of metal in the plants. However, it should be taken into account that the establishment of plants themselves may again change the pH-conditions in the soil.

Addition of ligands or chelators that can form complexes with the metals in the soil solution and thereby increase the plant uptake is another possibility for optimising phytoextraction. This possibility may however increase the risk of metal leaching. The additions must therefore be controlled very precisely in relation to the variation in infiltration and the retention and degradability of the ligands and chelators.

Phytoextraction as a remediation technique is known to influence primarily the most available forms of metal in the soil and thus also the most undesired effects associated with the metals. This may be utilised in case of lead, which is strongly bound in the soil. This approach will demand more knowledge of the time dependent changes in removal and the influence of plant growth on the availability of the metals.

In conclusion the experiments document a direct potential for using phytoextraction with the tested plants under Danish conditions for remediation of cadmium polluted soil. Among the tested plants Thlaspi caerulescens, possibly in combination with Salix, was most effective. A more widespread use of phytoextraction as a remediation technique demands further development of the technique.