«Dissertation zur Erlangung des Grades “Doktor der Naturwissenschaften” Im Promotionsfach Geowissenschaften Am Fachbereich Chemie, Pharmazie und ...»
0.2 0.1 Arsenic absorbance in HG-AAS (absorbances obtained in average of 3 sub-samples) against Arsenic concentration in samples Fig. (17) Graph plot shows actual arsenic sample concentration measured by the method of standard addition with addition of 1% cysteine in the sample
The plausible mechanism of this observed interference could be due to the interferent is precipitated as the metal (e.g. copper) and the mechanism probably involves capture and decomposition of the gaseous hydride at the freshly precipitated metal (Welz et al., 1984). The interference could be caused by a chemical reaction rather than physical adsorption and then followed by chemical reaction. All of the above matrix effect observed in the HG-AAS measurement in all steps and all samples could be minimized successfully with addition of 1% cysteine as masking agent in all steps all sub-samples (Appendix B, 9.2.2-9.2.3) to complex the trace amounts of metal. The above figure (17) demonstrated that such effect was reduced and the recorded arsenic absorbances were found to be ascending in right proportion compared with the standard calibration line.
(5.2) The sequential extraction of arsenic The following table (17) and figure (18) show the amount of arsenic, copper and manganese being partitioned in the five different sequential extraction steps. Further details of each subsample content and percentage distribution of As, Cu and Mn can be referred to Appendix (A, 9.1.1-9.1.5).
Fig. (18) Figure shows distribution of As, Fe and Mn (% range) in the sequential extraction scheme The observed As content under sequential extraction and XRF were found to be highly consistent (table 17). The global average content of arsenic being found in limestone is around
2.6 mg/kg according to Baur and Onishi (1969) and the As present in pellet limestone for soil pH amendment were found to be between 2-4 mg/kg (McBride and Spiers, 2001; Price and Pichler, 2006). It was also noticed that As in MAP/DAP phosphate fertilizers were in a typical range of 10-16 mg/kg (Charter et al., 1995; Raven and Loeppert, 1997). All of the measured sub-samples demonstrated much higher As content than the above mentioned usual limit.
Results and discussions
The measured Fe and Mn concentration in FAAS were 25% and 20% lower compared with direct XRF measurement of all lime samples included the reference material (table 17). The reason could be due to Fe/Mn bearing minerals (e.g. silicates) are not soluble in HCl (Step 2 and
5) and such missing Fe and Mn could still be left in the residue.
The geochemical similarity between P and As has leaded to an assumption that both P and As should be associated with similar constituents in the soils. The theories underlying this phosphate extraction step is due to similar electronic configuration and ability to form triprotic acids with similar dissociation constants followed by competitive exchange between phosphate and arsenate in soils. The nature of PO43- contains smaller size and higher charge density than arsenate would cause arsenate to be preferably desorbed in competition with phosphate (Wenzel et al., 2001; Keon et al., 2001) and these were revealed in the result’s figure (18) that a significant 17 ± 4% of As was mobilized by the first phosphate step.
Both of the mobilized Fe and Mn (both 4%) were minimal and could be neglected in the first extraction step. It was also noticed that only less than 3% of arsenic was mobilized in second step even though 25% and 42% total Fe and Mn were dissolved in this HCl-soluble step.
The valences of these elements (e.g. Fe2+ or Fe3+, and Mn2+ or Mn4+) were not known and it was therefore not possible to clearly determine the extraction yield was from a carbonate or as an amorphous oxide species.
The third oxalate moderately acid-reducing extraction extracted 18%, 33% and 14% of Fe, Mn and As respectively. The fourth strongly-reducing Ti step dissolved 46% (4.0 ± 2.2 g/kg) of the Fe content and 64% of the total As were also be dissolved, with a molar ratio As/Fe =
1.1·10-2. With the observed results, it demonstrated that the Fe fraction is dominated with crystalline oxides like goethite. On the contrary, only less than 1g/kg of Mn was mobilized in step 4.
Results and discussions
The last 5th step extracted further 9% of the Fe content and negligible fractions of As and Mn (2%). None of the As had been remained after 5 steps of extraction and however there was still 25% of both Fe and Mn being contained in the non-dissolved residual portion compared between the AAS and XRF analyses.
The above description stated how all the three elements distributed across five different steps. The major discovery in above was a comparatively strong binding between As and crystalline Fe phases. It was worth noticing that a significant one third of As managed to be mobilized under exchangeable to mildly acid-reducing conditions (sum of steps 1-3: 35 ± 6%, or
24.0 ± 6.1 mg/kg). This occurrence was because of readsorption of released metals is a common process in sequential extraction with potential sorbents are not exhausted and pH conditions are above the adsorption edge of the respective metal (Wenzel et al., 2001; Van Herreweghe et al., 2003). It was undesired for arsenate oxoanions to be adsorbed even under moderately acidic conditions and these observed results present uncertainty about whether the As load on the residual Fe oxide is the original or there is a considerable repartitioning of As between dissolved Mn and residual Fe oxides in the first three extractions steps and hence the predominance of As binding to crystalline Fe oxides over Mn oxides could not be assured in above circumstances.
Therefore, direct X-ray spectroscopic analysis was considered to be essential for further analysis.
(5.3) The sequential extraction of copper The following table (18) and figure (19) show the amount of copper, iron and manganese being partitioned in the four different sequential extraction steps. Further details of each subsample content and percentage distribution of Cu, Fe and Mn can be referred to Appendix A (9.1.6-9.1.10).
Fig. (19) Bar chart shows average percentage recoveries for the six lime sub-subsamples under the sequential extraction scheme of Cu and the matrix elements of Fe, Mn and Ca.
Total acid (HCl) digestion of the six different lime charges yielded bulk Fe, Mn and Cu concentrations of 8.5 ± 2.2 g/kg, 5.9 ± 1.8 g /kg, and 98.7 ± 69.8 mg/ kg, respectively (average of six subsamples). Each of the six individual sub-sample indicated Cu concentrations of 12, 41, 69, 137, 140 and 174 mg/kg. The vast standard deviation of Cu content is due to inhomogeneous lime composition but not analytical variance. All of the above sub-samples Cu concentrations showed higher copper concentrations than Cu content of 2-8 mg/kg limestone commonly used for soil pH amendment (McBride and Spiers, 2001). The copper concentrations of three subsamples demonstrated Cu contents exceeding the recommended limit of 100 mg/kg soil health tolerance level (Korthals et al., 1996). XRF analysis showed all the lime sub-samples contained 25% higher total Fe concentrations (11.3% ± 2.6%) and Mn concentrations (7.4 ± 2.2%), the extra Fe and Mn could be contained in HCl-insoluble Fe and Mn-bearing silicate minerals although no significant amount of Cu being presented in these residual phases according to XRF results (95 ± 64 mg/kg).
The average percentage recoveries for the six lime sub-subsamples under the sequential extraction scheme of Cu and the matrix elements of Fe, Mn and Ca were shown in Figure (19) and each of six sub-samples recoveries could be referred to Appendix A (126.96.36.199) (p.212-214).
The total amount of Cu extracted by the sequential procedure (Σsteps 1-4) was in good accordance with both the results of a separate single microwave-assisted aqua regia digestion and the direct XRF analyses. Copper was partitioned almost evenly among four steps. Around one third of Cu was mobilized by the first acetic acid step which is generally considered to target at exchangeable and carbonate-bound metal. The Cu associated with carbonate-bound metal was confirmed by the approximate 90% recovery for the carbonate matrix element of Ca (Figure 19).
The result showed carbonate minerals could be important for Cu fixation. An association of contaminant Cu with dolomite fragments in soil in form of malachite precipitates has already been reported (e.g. by McBride and Bouldin 1984). The above figure (19) showed that over 30% of the total Fe and 60% of Mn were dissolved along with the first step. The dissolved Fe and Mn contents could be due to Fe/Mn-bearing carbonates and also from poor stability of nanocrystalline oxide under the acidic condition of 1.0 M acetic acid in first step (pH ~ 2.5;
Whalley and Grant, 1994).
Results and discussions
An average of 18 ± 12% of Cu was mobilized by the second moderately reducing HAHC step. This step targets both Mn oxyhydroxides and amorphous Fe oxyhydroxides in which Fe was released more than Mn. However this second step in comparison only dissolved a minor proportion (12%) of total Fe but a significant proportion (34%) of total Mn. The calculated Cu/Mn ratios are the same for both of the first and second extraction step but not the case for the Cu/Fe ratios. The third step dissolved further 21 ± 10% on average of Cu in which associated with non-silicate, strongly reducible Fe oxides, with the Mn proportion contributed no significant role. The last step dissolved an additional 27 ± 8% on average of Cu, with one third of Fe and nearly nothing of Mn being left in this residual portion. This residue consists of silicates and other resistant minerals which usually play no significant role in environmental assessments. All of the copper was recovered as the result of total Cu in ∑1-4 steps matched with the direct XRF analysis.
The sequential extraction results suggest the copper can be mobilized under both acidic and reducing conditions. The results however do not reveal the exact phase Cu being associated with as the first acetic acid step can dissolve both Cu in the carbonate matrix and also the poorly crystalline oxides. Cu can possibly be hosted by the carbonate matrix or the oxide dendrite, or by the both phases. In addition it is important to distinguish Cu loaded between Mn and Fe hydroxide dendrites phases as Cu associated with the Mn hydroxide dendrites can be mobilized under reducing (e.g., submerged) soil conditions but not the Cu in the Fe hydroxide dendrites.
Direct X-ray spectroscopic speciation analysis should be introduced to resolve these uncertainties.
An attempted sequential extraction with modified 1st step with 1M sodium acetate and acetic acid buffer was applied (Appendix A, 188.8.131.52) (p.215) and successfully recovered all of the Cu in the six lime sub-samples (Appendix A, 184.108.40.206) (p.215) however it was unfortunate that the first step in which targeting exchangeable and carbonate-bound Cu metal did not reveal high percentage recovery of carbonate matrix element of Ca. The low percentage recovery of carbonate matrix element of Ca indicated the recovered Cu metal in step 1 did not successfully partition all the Cu metal which is associated with exchangeable and carbonate-bound Cu metal phase.
Wind direction Distribution March(%) Fig. (20) Figure shows wind and weather statistic of Münster/Osnabrück. Statistics based on observations taken between 7/2001 - 12/2010 daily from 7am to 7pm local time. (Section 8.3 Ref. 53)
The above figure (20) demonstrated the last 10 years statistical observation for the wind direction from 2001 to 2010 at Münster/Osnabrück region. The finding suggested the dominant wind direction throughout the years were mainly towards northeast. The sampling location 1 and 2 are positioned at the northeast of cement factory in Lengerich. All of the sampling points from location 1 and 2 were located less than 3km away from the centre of the radius of the cement factory.
Table (19): Table shows soil thallium (Tl) concentrations extracted by ammounium nitrate and aqua regia total acid digestion at location 1
The above table (19) showed the exchangeable thallium (extracted by ammonium nitrate) and total thallium concentrations in different sampling points and depths. The exchangeable thallium concentration of top and sub-soils could be compared.
A simple paired T-test of exchangeable and total thallium concentrations between top and sub-soils were carried out separately. The assumption for the observed data were from same subject (Location 1 in this case) and drawn from a population with a normal distribution. The test statistic was t with n-1 degrees of freedom. If the p-value associated with t was low ( 0.05), there was evident to reject the null hypothesis. As a result there was an evident that there was a difference in means across the paired observations.
The exchangeable thallium t-test value was 3.4998 with degree of freedom equal to 61 and p = 0.0009. The p-value associated with t was low ( 0.05). The small p-value indicated rejection of the null hypothesis. The paired exchangeable thallium concentration between top and sub-soils was considered to be statistically significant difference. The t-test was also applied for total thallium extracted by aqua regia acid digestion. T-test value was 3.5134 with degree of freedom equal to 59 and p = 0.0009. The paired total thallium concentration between top and sub-soils was also considered to be statistically significant difference.
Results and discussions