5.1.1. Comparison of dissolution methods for sulphur and heavy metals analysis in plant materials (I)

Biological samples consist of a complex mixture of carbohydrates, proteins and lipids. It is therefore necessary to decompose the organic matter and release the metals from the sample matrix. The majority of the digestion procedures used to date involve the initial use of strong oxidising agents, such as nitric acid, to decompose the organic matrix of the sample. Many elements are then liberated as soluble nitrate salts. Other acids can be employed to break down the sample matrix further, depending on the elements to be determined and the analysis technique chosen. The use of hydrofluoric acid is always necessary for the determination of a number of elements that are associated with siliceous minerals (88). Losses of trace elements during dissolution will affect the accuracy of the final results. There are number of possible loss mechanisms during sample decomposition, including gaseous evolution, absorption or adsorption onto surfaces, precipitation and the persistence of undissolved material (87).

According to the results obtained in I, Tables III and IV, the widely used acids or acid mixtures such as HNO₃+H₂O₂, HNO₃+HClO₄ or HNO₃ alone, or the dry ashing digestion procedure with HF used for the destruction of organic material, gave widely varying results for trace elements in plant materials. Thus, a careful choice of suitable digestion procedures for plant material is of great importance in order to ensure that correct results are obtained.

According to Table 2, which shows the results of sulphur analyses from I, Tables III and IV, the acid procedure with HNO₃ gave lower results for sulphur than the HNO₃+H₂O₂ procedure. Thus, our results correspond with those of another study which reported that a combination of an effective oxidizing agent, hydrogen peroxide (H₂O₂), and HNO₃ gives more complete decomposition of organic components than nitric acid alone (170).

According to our results, the digestion procedure with a mixture of HNO₃+HClO₄ slightly underestimated the S concentrations of plant materials. The low recoveries for sulphur using the HNO₃+HClO₄ procedure can, according to the studies of Randal et al. (171) on plant vegetative material (i.e., pasture samples, rape, spinach and clover), be partly due to incomplete oxidation of sulphur.

In this context it is worth noting that gaseous losses of S can occur during the digestion of plant materials using HNO₃+HClO₄ in open vessels (171). According to the studies of Bethge (172) on wood and pulp samples, sulphur is lost in the form of sulphur dioxide (SO₂) and carbonyl sulphide (CS₂) during wet digestion with perchloric acid, but hydrogen sulphide (H₂S) was not detected.

According to Table 2, the losses of sulphur were the highest in the dry ashing digestion procedure (abbr. HF(DAC)) in the case of both reference samples BCR CRM 100, BCR CRM 101 and the “Kemi” and “Tornio” pine needle samples. This is obviously due to gaseous losses of sulphur at the high temperatures employed during ashing, and this phenomenon has also been reported by Huang et al. (170) and Randal et al. (171).

The results presented in Table 2 show that the Leco combustion technique with infrared (IR) detection gave good precision and accuracy for sulphur (S).The precision for sulphur in reference samples BCR CRM 100 and BCR CRM 101 and also in “Kemi” and “Tornio” pine needle samples were all within 1–3 %. In addition, Leco combustion gave excellent results for sulphur compared to the certified values in references materials BCR CRM 100, BCR CRM 101 and HUMH2.

Our results for Cr in reference material HUMH2 are listed in Table 3, and were originally presented in I, see Table IV. According to Table 3, both the HNO₃ and HNO₃+H₂O₂ procedures were especially effective for determining Cr in reference sample HUMH2 because the results for Cr agreed well with the certified value. However, the HNO₃+HClO₄ procedure gave lower result.

Cary et al. (174) and Greenberg et al. (87) also reported a high deficit of Cr in plant materials with the HNO₃+HClO₄ procedure that they used. They attributed this to the formation and volatilization of chromyl chloride (CrO₂Cl₂) at a temperature of 116 °C during the acid digestion step. According to our results, the HNO₃+HClO₄ digestion procedure also gave a low result for Cr in the “Tornio” pine needles: values for Cr about 72 % and 66 % lower than the values with HNO₃ or HNO₃+H₂O₂, respectively.

However, in our study, the HF and HF(DAC) acid procedures gave 1.6 and 1.4 times greater Cr values for reference sample HUMH2 and the uncertified “Tornio” pine needle sample, respectively. The high Cr results for the HF and HF(DAC) procedures in the “Tornio” pine needle sample are probably due to the fact that the dust emitted from the ferrochrome and steel works of the AvestaPolarit Stainless Oy at Tornio contains FeO-Cr₂O₃, which is difficult to dissolve with other acid procedures (175). According to the results for Cr in reference material HUMH2, the values obtained with the HF and HF(DAC) procedures were also higher than those for the other procedures used. In this context it is worth noting that validation of the chemical analyses was performed using blanks, standard samples, and control moss samples, and it is therefore not likely that contamination by metals occurred during the analyses. However, if we compare the HF and HF(DAC) procedures with each other, the HF(DAC) procedure gave slightly lower values for Cr than the HF procedure, probably due to the volatilization of Cr at high temperatures during ashing (176–178).

According to the results in Paper IV (Table III) for the determination of Zn in reference material BCR CRM 101, both the HF and HF(DAC) digestion procedures gave values within the certified value. In contrast, the other procedures for Zn in BCR CRM 101 gave results lower than the certified values. According to the results in Paper IV (Table IV), the determination of Ni seemed to be less critical since a wide range of reagent combinations gave good results.

The low recovery for K in reference materials BCR CRM 100 and BCR CRM 101 with the HNO₃+HClO₄ procedure was also significant (see Paper I, Table III). A similar phenomenon has been reported in other studies, and is probably due to the formation of potassium perchlorate (KClO₄), which has a low solubility (173, 179). According to the results in Paper I (Table III), when the HF(DAC) digestion procedure was used for reference material BCR CRM 100, some K was evidently lost; this is probably due to volatilization at the high temperatures (99, 180).

5.1.2. Sulphur concentrations in pine needles (II)

Sulphur accumulation in pine needles around the pulp and paper mills was clearly higher than that at other points in the Kemi area. The highest sulphur concentrations occurred in the northern part of Kemi in the vicinity Oy Metsä-Botnia Ab Kemi Mills, which is the main area affected by sulphur deposition derived from these mills (Fig. 2).

In 1999, the sulphur concentrations of pine needles varied between 699–1090 mg/kg in (C) needles and between 605–1100 mg/kg in (C+1) needles. The highest individual sulphur concentration (1090 mg/kg) in (C) needles occurred at sampling site 12 (Vilmilä), and the highest individual sulphur concentration (1100 mg/kg) in (C+1) needles at sampling site 10 (Sotisaari). These values were 46 % and 45 % higher than those in the corresponding background samples (C-needles: 746 mg/kg, C+1-needles: 759 mg/kg) collected in Kuivaniemi at a distance about 25 km from Kemi.

The average sulphur concentration calculated from pine needles collected at sampling sites 6 (Mäntylä), 9 (Vähäkuivanuorontie), 10 (Sotisaari), 12 (Vilmilä), 14 (Elijärventie) was 954 mg/kg for (C) needles and 953 mg/kg for (C+1) needles. These values are 28 % higher for (C) needles and about 26 % higher for (C+1) needles than those in the corresponding background samples collected in Kuivaniemi. Oy Metsä-Botnia Ab Kemi Mills and sampling points 6 (Mäntylä), 9 (Vähäkuivanuorontie), 10 (Sotisaari) and 12 (Vilmilä) are located inside the area where the sulphur concentration was between 1000–1100 mg/kg (see also Fig 2 in II). Correspondingly, the average sulphur concentration of needles at sampling sites 6, 9, 10, 12 and 14, compared to the average sulphur concentration for needles at all sampling sites, was 11 % higher than the average sulphur concentration in (C) needles and 18 % higher than in (C+1) needles.

In the southern part of Kemi the impact of pollution from StroraEnso Oyj Veitsiluoto Mill was most clearly evident at sampling site 21 (Järppi) and at sampling site 22 (Hepola). However, the district heating plant in the vicinity of sampling site 22 (Hepola) probably also has an influence on the sulphur concentrations of needles at this sampling site. In conclusion, sulphur deposition and the accumulation of sulphur in pine needles around the pulp and paper mills were higher than at other points in the Kemi area. Thus our results are in good agreement with the studies of Määttänen (22) on the environmental effects of the pulp mills of Enocell Oy in North Karelia. Similar results have also been reported in many other studies using pine needles as a bioindicator for sulphur deposition in the areas surrounding point sources, such as Kekäläinen et al. (181) and Pesonen et al. (182).

In addition, the results of the needle sulphur survey carried out in 1999 also correspond well with the latest computer simulation study made in 2001 by the Finnish Meteorological Institute on the aerial distribution pattern of SO₂ and TRS in the area around the pulp and paper mills of Oy Metsä-Botnia Ab Kemi Mills (17). This model is based on the use of the UDM-FIM (Urban Dispersion Modelling System-Finnish Meteorological Institute) computer simulation program, which uses a Gaussian model to predict the dispersion of a plume in the vertical and horizontal directions (183). According to this simulation program, the modelled concentrations of SO₂ and TRS were the highest in the immediate vicinity of the mills, i.e. within a radius of 1–3 km around the mills of Oy Metsä-Botnia Ab Kemi Mills.

If we compare the results of the bioindicator study made in 1999 to those carried out earlier, i.e. in 1979 by Huttunen et al. (184) and in 1989 by Vanhatalo (185), there is a clear decreasing trend in the size of the sulphur dispersion area (km²) during 1979–1999 (see Table 4). In 1979 and 1989 sulphur was determined by XRF, and in 1999 by ICP-AES.

It is reasonable to suppose that this decreasing trend in the size of the sulphur dispersion area is due to decreased total sulphur emissions in the Kemi area during 1980–1998. The decreasing trends for the maximum and mean sulphur concentrations in needles also seem to be reasonable in the light of the decreased sulphur emissions (see Paper II, Fig. 1, Table III, Table IV). Thus, our results are similar to those reported for Oulu in the areas around the pulp and paper mills works (186). In conclusion, although the total sulphur emissions in Kemi have decreased tremendously during the past two decades from a value of 4500 t (S) in 1980 to a value of 990 t (S) in 1998, pine needles still appear to be useful bioindicators for assessing the distribution patterns of aerial sulphur emissions derivated from the pulp and paper industry.

5.1.3. Concentrations of Fe, Zn, Ca, V and Pb in pine needles (II)

The highest needle iron (Fe) concentrations occurred relatively close to the pulp and paper mills of Oy Metsä-Botnia Ab at sampling sites 6 (Mäntylä) and 12 (Vilmilä), and close to the mill of StoraEnso Oyj at sampling sites 51 (Rivinokka) and 52 (Haukkari). However, high values also occurred close to roads. It is thus highly likely that the high Fe concentrations are derived from dust from the surrounding land rather than from the mills. The fact that the highest results for zinc (Zn) occurred along roadsides and near roads also indicated that it is derived from the same sources as Fe.

High concentrations of calcium (Ca) occurred especially in needles collected in the northern part of Kemi in the vicinity of the Oy Metsä-Botnia Ab Kemi Mills. Thus it is likely that airborne pollutants from the Oy Metsä-Botnia Ab Kemi Mills have had a strong influence on needle Ca concentrations especially at sampling site 9 (Vähäkuivanuorontie), where the Ca concentration was 4960 mg/kg in (C+1) needles, as well as at sampling sites 10 (Sotisaari), 12 (Vilmilä), 15 (Ristikangas), 16 (Junko) and 32 (Lautiosaari). The Ca concentration in (C+1) needles at sampling site 9 (Vähäkuivanuoro) was 48 % higher than the average Ca concentration calculated from all (C+1) needles. Thus it is likely that Ca emissions, which are typical for pulp and paper mills, reached this sampling point located close to Oy Metsä-Botnia Kemi Mills, and that part of the Ca in the needles is derived from the mills.

According to Table 5, there is poor correlation between the concentrations of individual elements in (C+1) needles. The non-significant correlations are probably partly due to the extremely low emissions of heavy metals from the pulp and paper mills, and also due to the physiological properties of needles in accumulating metals (187).

However, although pine needles do not appear to be as appropriate a method for monitoring the deposition of heavy metals as mosses and lichens, Laaksovirta et al. (187) used the chemical analysis of pine needle as a method for monitoring the accumulation of airborne pollutants emitted from pulp and paper mills at Valkeakoski, and Määttänen (22) in the areas around the pulp and paper mills of Enocell Oy in North Karelia (47).

In our study, the vanadium concentrations, which are a good indicator of fuel oil burning, were < 1 mg/kg at all the sampling sites, and the corresponding lead concentrations were < 5 mg/kg. Thus, needles did not readily accumulated these elements; this phenomenon has also been reported in another study (187). Therefore the low Pb concentrations in the needles is also partly due the fact that Pb emissions from traffic (0 t/1999) is no longer a problem, because nowadays the petrol used in cars in Finland is unleaded (188). However, even in the mid 1970s and mid 1980s, the average Pb concentrations in the Helsinki metropolis area were 22.8 mg/kg and 8.5 mg/kg in (C) and (C+1) needles, respectively (189).

5.1.4. Accumulation of Cr, Ni and Zn in mosses (III)

The regional distribution patterns of Cr and Ni in mosses in the Kemi-Tornio area in 2000 showed clearly that the most polluted area (Cr > 200 g/g and Ni > 20 g/g) appears to lies within a few kilometres of the ferrochrome and stainless steel works of AvestaPolarit Stainless Steel Oy (see Paper III, Fig. 1). Within this area, the Cr concentrations in mosses were 4–13 times higher than those outside the urban area of Tornio. In 2000, the highest individual Cr concentration (2700 g/g) occurred at a distance of 1.9 km from to the southeast from the works. Slightly polluted areas (Cr < 50 g/g and Ni < 9.9 g/g) were located farther away at a distance of about 12–14 km from the works.

The area most polluted by the opencast chromium mining complex of AvestaPolarit Chrome Oy Kemi Mine (Cr > 200 g/g and Ni < 20 g/g) appeared to be in the immediate vicinity of the complex. The slightly polluted area (5.0 g/g < Ni < 9.9 g/g) occurred within a radius of 2–6 km around the mining complex. However, there was a small anomaly to the south of the mining complex at Järppi, probably due to Ni emissions from local oil-burning sources.

According to Figure 3, the deposition and accumulation of Cr and Ni in mosses clearly decreased with increasing distance from the works. However, the Zn concentrations in mosses did not decrease with increasing distance from the pollution source (works). Thus, our results for Cr and Ni correspond to those obtained in another study carried out in the same area by Kansanen and Venetvaara (190) at the end of the 1980s.

The Zn concentrations were similar throughout the study area and there is therefore no distribution pattern at all for Zn. According to Rühling and Tyler, this phenomenon is probably due to the fact that the retention of Cr and Ni from atmospheric deposition by mosses is almost total, while that of Zn is relatively low; mosses have distinct accumulation rates for individual trace metals as a result of their physiological properties (34,191). Similar results have also been reported in other studies in which mosses were used as a bioindicator for heavy metal deposition (36, 38, 165, 192). Another possible explanation could be that metallic Zn has a relatively low boiling point (907 °C, as opposed to 2672 °C for Cr and 2730 °C for Ni), and a high proportion of the Zn emitted from the ferrochrome and stainless steel works of AvestaPolarit Stainless Oy is in the form of metal vapour. As a result, Zn emissions are likely to be spread over a great distance (27). Differences in the translocation of heavy metals in the atmosphere have been reported in (2, 193).

Pilegaard (41) used transplanted lichens (Hypogymnia Physodes) for the biological monitoring of air pollution emitted from a Danish steelwork in Frederiksvaerk, and he reported low deposition of Zn on lichens; the lichens initially accumulated Zn, but they were not able to retain large amounts of Zn when deposition decreased. In addition, the “wash out” of heavy metals in mosses, lichens and needles as well as also sulphur in needles by water (i.e. rain or washing) has been reported in (28, 194–197).

The 3-D scatterplot results in Fig. 4 for Cr, Ni and Zn also partly support the finding that the highest Cr and Ni concentrations in mosses occurred close to the pollution sources (0.6–1.9 km and 0.6–3.5 km, respectively), but not for Zn, because the highest Zn concentration (78 g/g) occurred at a distance of 12.3 km from the nearest pollution source (AvestaPolarit Chrome Oy Kemi Mine).

The 3-D scatterplots also illustrate that, although the heavy metal deposition emitted from both the ferrochrome and stainless steel works of AvestaPolarit Stainless Oy and the opencast chromium mining complex of AvestaPolarit Chrome Oy Kemi Mine overlap in many parts of the study area, the main deposition areas are close to the works and the mine.

The Kendall correlations in Table 6 show that Cr and Ni were highly correlated in mosses, thus indicating that they are originated from the same emission sources. Although the correlation between Zn and Cr and Ni was not very high, it is clear that the Zn deposition in the mosses was primarily derived from the ferrochrome and stainless steel works of AvestaPolarit Stainless Oy and the opencast chrome mine of AvestaPolarit Chrome Oy Kemi Mine, because the nearest point sources that emit significant amounts of these metals are located more than 150 km away.

According to the descriptive statistical results for the moss survey in 1990, 1995 and 2000 (Paper III, Table I and II), it can be concluded that the average Ni concentration (mean) in mosses has increased since 1990, which reflects the increased Ni emissions from the ferrochrome and stainless steel works during the same period. The Zn emissions from works were at approximately the same level in 1990 and in 1999 but, according to our results, the average Zn concentration (mean) in mosses has decreased. However, the average Cr concentration (mean) for all the moss samples in 2000 was 2.2 times higher than the corresponding value in 1990 and 1995 even though the Cr emissions have decreased since 1990 from a value of 20 t to 15.2 t in 1999.

There are a large number of uncontrollable factors, e.g. weather parameters, such as wind direction and speed, precipitation, humidity, the uncontrolled spread of emissions, the composition and variations in the amount of emissions, and the efficiency of the dust removal processes, that cannot be “standardized” before sampling. All these factors may have an influence on the amounts of heavy metals emitted from point sources, and how they are spread and deposited after release from the point sources. Thus, these factors could partly explain why the average Cr concentrations (mean) in mosses have increased since 1990 even though the Cr emissions have decreased.