A PEAT REFERENCE MATERIAL FOR TRACE ELEMENT ANALYSES

Carlo Barbante* (Department of Environmental Sciences, University of  Venice, Italy), William Shotyk (Geological Institute, University of Berne, Switzerland),  Harald Biester (Institute of Environmental Geochemistry, University  of Heidelberg, Germany), Andriy Cheburkin (EMMA Analytical, Elmvale, Ontario, Canada), Hendrik Emons (Environmental Specimen Bank, JRC, Juelich, Germany), John Farmer (Department of Chemistry, University of Edinburgh, Scotland), Eric Hoffman (ACTLABS, Ancaster, Ontario, Canada), Antonio  Martinez Cortizas (Department of Soil Sciences, University of  Santiago de Compostela, Spain), Jörg Matschullat (Interdisciplinary for Environmental Research Centre, Technical University of Freiberg, Germany), Stephen Norton  (Department of Geology, University of Maine, USA), Fiona Roos (Geological Institute, University of Berne, Switzerland),  James Schweyer (Ontario Geological Survey, Sudbury, Ontario, Canada), Eiliv Steinnes (Department of Chemistry,  University of Trondheim, Norway)

 

ABSTRACT

During the evaluation of analytical results, strict control of the entire analytical procedure and the determination of accuracy are essential parts of good laboratory practice. Given the growing interest in using of peat bogs as environmental archives of atmospheric trace elements, a collaborative inter‑laboratory exercise was organized to improve the quality of laboratory analyses of trace elements in peat, and to contribute to a general improvement in our understanding of peat chemistry. Twelve institutions in seven countries participated in the project and measured, using a wide range of analytical techniques, the concentrations of many trace elements of environmental interest. This new candidate reference material consists of a Carex (sedge) fen peat. The results, treated under strict statistical protocol allow us to compare the different analytical methods used and discuss the possible sources of error linked to a specific laboratory or to a specific method.

 

INTRODUCTION

There is growing interest in the use of peat bogs as archives of atmospheric metal pollution (Shotyk et al., 1996). We now know that Pb is effectively immobile in peat cores from ombrogenic bogs (Shotyk et al., 1998), and there is increasing interest that the same is true of Hg (Cortizas et al., 1999) and Cu. Because the silicate minerals derived from soil dust do not weather appreciably, lithogenic elements such as Sc, Ti, Y, Zr, Hf  and REE can be used to quantify atmospheric fluxes of soil dust, identify long-term weathering trends, and reconstruction Holocene climate change. High quality analyses of major and trace elements are needed both to understand the trophic status of peat profiles, and for the applications described above. However, thus far there has been no peat reference material available to evaluate independently the accuracy of chemical analyses of peats. While analytical precision can be determined from replicate measurements of individual samples, accuracy can only be determined through blind analyses of a comparable material whose elemental concentrations are known (Barbante et al. 2000).

            Plant Certified Reference Materials (CRMs) are often used to determine accuracy, but Sphagnum peat from ombrogenic bogs contains not only nutrient elements taken up by plants from rainwater, but also soil-derived dust. Depending on the degree of humification of peat, the ash content (residual material upon combustion at 550 °C) can be considerably higher in peat compared to plants. Given that much of this ash may consist of silicate minerals, a complete acid dissolution of peat requires efficient destruction of both the organic and inorganic phases (Weiss et al., 1999). It has been shown that HF is needed to completely dissolve all of the silicates which are present in peat, but this reagent is generally not needed to digest plant samples. Thus, analytical methods optimized for measuring trace elements in plant materials may not necessarily provide good recoveries for peat samples. Most analytical methods for measuring trace elements in peats such as Flame and Graphite Furnace AAS, ICP-OES and ICP-MS, DPASV, and ID-TIMS require that the solid sample be brought into a liquid phase; to ensure total metal concentrations are obtained, the digestion must be complete.

Coal CRMs have both a biological component (mainly plant-derived) and an inorganic fraction. However, the biological part may have been subjected to the elevated temperatures and pressures of diagenesis and metamorphism, giving rise to highly condensed, aromatic, polymeric building-blocks which are difficult to decompose. Also, the inorganic fraction of coals is dominated by metal sulphides, especially those of Fe, but these are largely absent from ombrogenic peat.

            Here we described the results of an interlaboratory comparison of a peat reference material, OGS 1878P, which was produced by the Ontario Geological Survey in 1982.

 

METHODS

            The material consists of 500 kg of a Carex (sedge) fen peat from the Holland Marsh, Ontario, Canada. Because it is a fen peat, the concentration of mineral material was relatively high (ca. 20 % by weight ash). Considering that the Holland Marsh is a vegetable-producing agricultural area, Cu concentration is also fairly high because copper sulphate has long been used as a fertilizer supplement. Finally, one of Canada`s major highways intersects the Holland Marsh near the point of sample collection, so Pb too is also rather abundant in this material.

            The bulk material was milled, homogenized, split and bottled in 1982 by the Ontario Geological Survey (sample code OGS 1878P). It has been made available by the Geoscience Laboratories, Sudbury, Ontario. Aliquots of approximately 1 kg each were delivered to all the laboratories participating to the interlaboratory study, where further treatments were carried out following strict procedures.

            Each participant was asked to homogenize the material by shaking the sample container jar for one hour, then dry the material for 2 hours at 105 °C until constant weight. Five aliquots of sample (from 50 to 250 mg each) were weighed and digested following the methods in use in each laboratory participating in the exercise. The method we recommended includes HF and provides a complete digestion of both the organic and inorganic fractions of peat (Weiss et al., 1999). In laboratories using direct analytical methods such as XRF, INAA, pyrolysis, solid- or slurry-sampling GFAAS, no chemical or microwave assisted digestion of the samples was necessary.

            Twelve laboratories, using ten different analytical techniques (ICP-OES, ICP-MS, ICP-SFMS, DPASV, ETAAS, HGAAS, INAA,  XRF, ID-TIMS and AAS) analysed the aliquots of the sample independently and without knowing in advance the concentration of the elements in the material. All the laboratories participating in the exercise were requested to analyse all the elements which they usually measure following their own usual procedures. They were selected as expert laboratories in the field, having already in place the necessary quality assurance and quality control systems before participating in the interlaboratory exercise. In other words the methods used already had been routinely validated for peat and compared with analyses of certified reference materials, such as: NIST Apple Leaves - NBS 1515; NIST Pine Needles - NBS 1575; BCR 281 Rye-Grass; IAEA 336 Trace Elements in Lichen.

            Finally a strict protocol for the analysis was followed in order to have a homogeneous set of data. Data were then collected and evaluated using a rigorous statistical approach.

 

RESULTS AND DISCUSSION

            Several elements have been analysed thanks to the analytical techniques used in the different laboratories, but only for few of them the values were obtained by combining data from two or more independent analytical methodologies. Recommended values obtained for some trace elements are reported in Table 1. The relevant uncertainties derive from the spread of the mean values for a given element after the appropriate statistical treatment. The recommended value for a given element is the mean of means for accepted laboratories.

 

Table 1. Details of OGS 1878P peat bog Candidate Reference Material for trace metals;

concentration in mg g-1.

 

 

Element                Certified                                           Element            Certified                                  

 

      As                  8.9 ± 0.75a                                       Mn                   206 ± 30         

      Cr                  8.0 ± 1.0                                          Ni                    6.1 ± 0.7                     

      Cu                 195 ± 16                                          Sb                    0.34 ± 0.05                 

      Co                 1.8 ± 0.2                                          Se                    0.73 ± 0.08                 

      Cd                 0.53 ± 0.03                                      Sc                    1.04b   

      Fe                  8700 ± 600                                      Hg                    0.089b             

      Pb                  78.8 ± 2.9                                        V                     9.7 ± 1.2         

      Zn                  43    ± 3                                          

 

a  95% confidence interval. b Information value only.

 

 

An example of the bar graphs used to illustrate the results obtained by the participants is shown in figure 1. In this case, after a statistical test it was decided to not consider the values submitted by laboratories No 10 and 11, because their values were considered as outlayers. The low value obtained by XRF (10) can be explained by possible inhomogeneity of the sample or inefficiency of the technique due to the high content of mineral material. In the case of the measurements carried out by AAS (11) the laboratory involved used 8 M nitric acid and not HF in the digestion procedure, having then a not complete digestion of the sample (Weiss et al, 1999). A similar approach, using bar graph and statistical evaluation, was applyed to all the elements considered.

 


 

 


Figure 1.           Concentration values for Pb in 1878P peat bog candidate reference material submitted by different laboratories (1- 11). Values of laboratories 10 and 11 were not accepted for certification. Error bars such as standard deviation are also reported. The techniques used were: ICP-AES (2, 3), DPASV (4), ID-TIMS (5), ETAAS (6, 7), XRF (10), ICP-QMS (1), ICP-SFMS (8, 9) and AAS (11). Certified value (————) and 95% confidence interval ( —    —    — ) are also shown.

 

REFERENCES

 

Barbante C, Cozzi G, Capodaglio G, Cescon P (2000), J.Anal. At. Spectrom.. 15:377-382.

Cortizas AM, Pontevedra Pomba X, Garcia-Rodeja E, Novoa Munoz JC and W Shotyk (1999), Science. 284:939-942.

Shotyk W (1996), Environmental Reviews. 4(2):149-183.

Shotyk W, Norton SA Farmer JG (1997), Water Air and Soil Pollution. 100 (3/4): 213-219.

Shotyk W, Weiss D, Appleby PG, Cheburkin AK, Frei R, Gloor M, Kramers JD, Reese S, van der Knaap WO, (1998), Science. 281:1635-1640.

Weiss D, Shotyk W, Schäfer H, Loyall U, Grollimund E, and Gloor M (1999), Fresenius Journal of Analytical Chemistry. 363:300-305.