This essay has been submitted by a student. This is not an example of the work written by our professional essay writers.
Although portable X-Ray Fluorescence (XRF) technology is widely accepted for environmental use in field screening test regarding the analytical approach, it needs to be evaluated with sufficient data and meet its performance characteristics to be employable for decision making purposes. Usually, for an XRF sample, the most interesting query is "How accurate is the XRF technique in detecting different targeted metals in soil?" This paper presents pairwise comparisons between the XRF and Inductively Coupled Plasma Atomic Emission Spectrometer (ICP-AES) results for individual elements of Ni, Cu, Zn, Pb, Cd, Cr, Hg, and As. The portable XRF analyzer was used to estimate the concentration levels of eight heavy metal elements, and then pairwise comparisons were made between the XRF and ICP-AES results. Results presented in this paper suggest that XRF testing can be used as a screening technique with high confidence on samples where the first group of metal element (Pb, Zn, Ni, and Cu) concentrations is well in excess of the threshold limits. The order of reliability of the XRF measurements is Pb > Zn > Ni > Cu and their relative proximity ranges from 85%~35%. On the other hand, another group of metal elements which includes Hg, Cd, Cr, and As shows poor correlation. Their relative proximity ranges from 25%~2.3%.
Keywords: XRF; heavy metal; ICP-AES; soil pollution; relative proximity.
During the year of 1970, industry was booming and great quantities of industrial wastes were dumped along the Erren River in Taiwan. Hazardous contaminants have been found along the riverbanks, and, through time, polluted contaminants have become embedded in the subsurface soil, therefore, necessitating a comprehensive survey within the river basin. Since the conventional laboratory experiments are time consuming and costly, the alternative X-Ray Fluorescence (XRF) technique provides a rapid, cost-effective solution for conducting field sampling. However, for the XRF operator, the most interesting query is "How accurate is the XRF technique in detecting different targeted metals in soil?"
Several laboratory experiments can detect heavy metal in soil. Two major methods used are the inductively coupled plasma atomic emission spectrometer (ICP-AES) and the atomic absorption spectrometer (AAS). Both of these methods require soil samples to be imported into the instrument as a solution in order to perform sample digestion or extraction (Radu and Diamond, 2009; Shefsky, 1997).
Especially important is rapid pollution monitoring in instances of pollution incidents since timely on-site analysis and fast decision making are highly important in protecting the health of local communities (Radu and Diamond, 2009). Field methods, sufficiently accurate and well-documented, can offer significant advantages over laboratory methods to support field decision-making. Field analysis requires less involvement of sample handling, transportation, and chain-of-custody documentation compared to that of laboratory analysis, and, therefore, is often less expensive per sample. Additionally, speedy analytical turn-around of the field method can be advantageous for instantaneous field decision-making and significantly reduce project cost. The lower cost-per-sample allows for denser, more complete sampling (Shefsky, 1997; Radu and Diamond, 2009; Shrivastava et al., 2005).
The field portable XRF has been extensively used in research and has several remarkable characteristics. Providing simultaneous analysis of up to 25 elements (Radu and Diamond, 2009), the portable device, an exemplary field method, offers extremely rapid, cost-effective screening of heavy metals in soil by in-situ measurements (Shefsky, 1997). In addition, portable XRF analyzers have been successfully utilized for lead-based pollutant screening (Clark et al., 1999; Markey et al., 2008; Morley et al., 1999; Drake et al., 2003) and can quickly and reliably provide lead concentration information for safety purposes (Drake et al., 2003). Further, this technique significantly cuts the time required for sample characterization (Radu and Diamond, 2009; Shrivastava et al., 2005; Markey et al., 2008; Morley et al., 1999; Bernick and Campagna, 1995; Song et al., 2001). Since XRF is completely non-destructive, any sample collected and measured in the field can be retained for verification in a laboratory (Radu and Diamond, 2009; Shefsky, 1997; Shrivastava et al., 2005; Kalnicky and Singhvi, 2001). Ideally, this portable instrument has the capability to perform direct, in-situ analysis of solid soil samples without the need for digestion, which would be a major step forward in contaminant analysis (Radu and Diamond, 2009).
Several official methods such as the Environmental Protection Agency (EPA) Method 6200 (Environmental Protection Agency, 2007) and the National Institute for Occupational Safety and Health (NIOSH): Method 7702 (Drake et al., 2003; NIOSH, 1998) now involve the use of the portable XRF technology. In addition, it is being increasingly highlighted by numerous researchers for the determination of metals in soil (Radu and Diamond, 2009; Shefsky, 1997; Clark et al., 1999; Markey et al., 2008; Bernick et al., 1995; Carr et al., 2008; Makinen et al., 2005). Since XRF instruments have been extensively used for site measurements, the operator may be interested in its analysis results for various metals in comparison with the laboratory analysis (Markey et al., 2008; Drake et al., 2003; Song et al., 2001; Kalnicky and Singhvi, 2001; Carr et al., 2008).
This paper presents how reliable and accurate the XRF tests are by comparing 60 in-situ samples with eight heavy metal elements of the XRF analysis results with those of the ICP-AES experiments. The field XRF tests were conducted in the area of the Erren River Basin of Taiwan, and sixty samples were carried out by ICP-AES experiments simultaneously.
Materials and methods
The portable field X-ray fluorescence (XRF) model used in this work is the NITON XL-722, which is equipped with a Cd-109 radioisotope source and Am-241 radioisotope source. The Cd-109 source with a measurement time of 800 s is used for detecting elements of chromium (Cr), manganese (Mn), iron (Fe), cobalt (Co), nickel (Ni), copper (Cu), zinc (Zn), arsenic (As), selenium (Se), strontium (Sr), zirconium (Zr), molybdenum (Mo), mercury (Hg), lead (Pb), and rubidium (Rb). The Am-241 source with a measurement time of 200 s is used for elements of cadmium (Cd), silver (Ag), tin (Sn), antimony (Sb), and barium (Ba). With the confirmatory analysis using an inductively coupled plasma atomic emission spectrometer (ICP-AES), the laboratory-based aqua regia acid digestion (Radu and Diamond, 2009) was performed on the soil samples collected in the study area. The soil samples were digested using the method of standard ISO11466.2 (International Organization for Standardization, 1995) (known in Taiwan by EPA as NIEA S321.63B).
In order to comprehensively evaluate the raw data of XRF vs. ICP-AES for various elements, Relative Proximity (RP) is used. The definition of Relative Proximity takes into consideration only the samples with values over the controlled threshold limit (i.e., the permissible exposure level for pollutants). Therefore, the number of detected field samples of the ICP-AES results over the threshold limit divided by the number of detected XRF results over the threshold limit determines the Relative Proximity; that is,
The Sampling Site
With approximately a 339-square kilometer drainage area, the Erren River is about 62.5 kilometers in length. It flows through Tainan County and Kaohsiung County, past Tainan City, and runs into the Taiwan Strait. Five tributaries branching off of the Erren River from upstream to downstream are Ngau-Liao Creek, Ngau-Chou-Po Creek, Song-Zi-Jiao Creek, Shen-Keng-Zi Creek, Kang-Wei-Kuo Creek, and San-Yeh Creek (Fig.1).
An effective framework of site selection for identifying and locating potential pollution sites are within the Erren River basin (Wu et al., 2010; Lee et al., 2011). Collected soil and water samples were tested for possible contamination. Soil samples were analyzed for possible metal ion concentrations or major pollutants based on the industrial activity in the surrounding area.
Field screening tests were carried out on forty-one selected sites in the area of the Erren River Basin of Taiwan (Wu et al., 2010; Lee et al., 2011). One-hundred seventeen soil samples were collected using the XRF in the field screening tests. From the collected, on site environmental samples, sixty top ranking samples within forty-one sites were delivered to the laboratory for further ICP-AES testing. The collected soil samples were found to be denser with contaminants between Kang-Wei-Kuo Creek and the estuary. Sampling locations are mapped in Fig. 2. All XRF data were collected with a NITON XL-722 equipped with a Cd-109 radioisotope source and Am-241 radioisotope source.
Discrete sampling, where physical removal of a sample from soil, was carried out for analyzing the soil samples by the XRF technique, limiting the number of measurements normally performed in a site activity. The benefit is that analytical accuracy and precision are generally improved for prepared samples compared to in situ measurements (Kalnicky and Singhvi, 2001). Soil samples undergoing the XRF technique were analyzed through plastic bags. The measurements of the soil samples in freezer bags were measured according to an empty freezer bag analyzed as a blank sample and all sample measurements were blank-corrected (Radu and Diamond, 2009). Prior to the sample measurement, an internal instrument calibration was performed (Radu and Diamond, 2009).
The confirmatory method should be an element method in order to be compatible with the XRF method, which is also an element method. The laboratory confirmatory method should as nearly as possible match the field method. The soil samples, collected in the area, went through the aqua regia acid digestion in the laboratory and then were analyzed by an inductively coupled plasma atomic emission spectrometer (ICP-AES). The ICP-AES experiments require the soil sample to be imported as a solution into the instrument, but, beforehand, the lab must perform a sample extraction or digestion (Shefsky, 1997). A statistical analysis of data was performed using the Matlab statistical toolbox. Linear regression was used to correlate the XRF and ICP-AES data, and each data set was checked for potential outliers (Radu and Diamond, 2009; Kalnicky and Singhvi, 2001).
Results and discussion
Individual elements interpretation
Pairwise comparisons between the XRF and ICP-AES results were made for elements Ni, Cu, Zn, Pb, Cd, Cr, Hg, and As; these eight heavy metal elements are presented in Fig. 3 to Fig. 18. In Fig. 3, Fig. 5, Fig. 7, Fig. 9, Fig. 11, Fig. 13, Fig. 15, and Fig. 17, the green dash-dotted line represents the threshold limit of the concentrations higher than the pollution threshold limit in Table 1. On the other hand, Fig. 4, Fig. 6, Fig. 8, Fig. 10, Fig. 12, Fig. 14, Fig. 16, and Fig. 18 represent the XRF value vs. ICP-AES value with their regression lines for these eight elements. The regression lines (black solid line) describe the minimized distance from the line to the data points of the individual methods. The blue dash line is the 1:1 line. The purpose for presenting the1:1 line is to show how well the slope of regression line compares to the 1:1 line. Ideally, the scatter points lying on the 1:1 line means that the XRF measurements and ICP-AES experiments are exactly the same, and the regression line should fall on the 1:1 line. Pairwise comparisons between the XRF and ICP-AES results for each element are analyzed as follows.
(1) Element nickel (Ni)
In Fig. 3, the overall trends measured by the XRF technique are higher than the ICP-AES results for the element Ni. The values higher than the threshold limit (the green dash-dotted line in Fig. 3 and the value in Table 1) were detected by the XRF method; however, the lower values of XRF below the threshold limit show a downward parallel gap with ICP-AES. The soil samples have a linear regression slope of 1.144 and the correlation coefficient (R2 value) of 0.7281 as shown in Fig. 4. As can be seen, although the R2 value of Ni is higher than all other elements, the regression line has shifted away from the 1:1 line.
(2) Element copper (Cu)
Similar to Ni, the overall trends of Cu measured through the XRF technique are higher than (except one instance shows equal to) the ICP-AES results (Fig. 5). The values higher than the pollution threshold limit were detected by the XRF method, but the lower values regarding the ICP-AES results differ from the XRF results. The soil samples have a linear regression slope of 1.184 and R2 of 0.4095 as shown in Fig. 6 and this line veers away from the 1:1 line.
(3) Element zinc (Zn)
The result of the element Zn shows that sensitivity and accuracy of the XRF measurements compared with those of the ICP-AES are acceptable and that any interference by other factors is unlikely (Fig. 7). The regression line (Fig. 8) and the 1:1 line at the small values almost merge together but diverge as the value increases. The soil samples have a linear regression slope of 0.9468, and the R2 value is 0.6551. If controlling only the values over the pollution threshold limit, 80.0% of the data exist in close proximity of each other (Table 2). The screening test of high values can be used as a reference for selecting sampling points.
(4) Element lead (Pb)
The overall trend of the XRF measurements agrees with the ICP-AES experiments for the element Pb (Fig. 9). The regression line with a slope of 0.9781 (Fig. 10) matches fairly well with the 1:1 line, and the R2 value is 0.6689. If controlling only the values over the pollution threshold limit, 85.17% of the data are positioned in close proximity of each other (Table 2). This result is superior to all the other seven elements. Excellent sensitivity and accuracy of Pb have a very small potential for interference by other factors, unless a very unique situation would occur. High screen measurement values of the above information indicate a considerable degree of reliability.
(5) Element cadmium (Cd)
Cd shows poor accuracy. There is no significant relationship between the XRF measurements and the ICP-AES experiments (Fig. 11). As can be seen in Fig. 11, the XRF measurements are all higher than the ICP-AES experiments. The relative proximity is only 5.77% (Table 2). The regression line with a slope of 0.9106 and the R2 value of 0.07823 diverges greatly from the 1:1 line that the 1:1 line cannot be shown on the figure (Fig. 12).
(6) Element chromium (Cr)
Cr presents poor accuracy. There is a parallel gap of the XRF measurements below the threshold limit compared with the ICP-AES experiment results (Fig. 13). Most of the XRF measurements are all higher than the results of the ICP-AES experiments. The relative proximity is only 16.67% (Table 2). The regression line with a slope of 0.5778 and the R2 value of 0.1504 diverges from the 1:1 line extensively that the 1:1 line cannot be shown on the figure (Fig. 14).
(7) Element mercury (Hg)
Similar to Cr, Hg presents poor accuracy. The XRF measurements parallely shift away from the ICP-AES experiment results (Fig. 15). As can be seen in Fig. 15, the XRF measurements are all higher than the results of the ICP-AES experiments. The relative proximity is only 2.30% (Table 2). The regression line with a slope of 0.1076 and the R2 value of 0.01143 totally diverges from the 1:1 line and cannot be shown on the figure (Fig. 16). This sample of the Hg screen test data shows uncertainty regarding other factors which may cause serious interference. In the case where the Hg screening test shows abnormality without any special performance of the other elements, this information can be useful as a basis for choosing sampling sites.
(8) Element arsenic (As)
Similar to Cr and Hg, the trend of element As parallely descends for both the XRF measurements and the ICP-AES experiment results. As can be seen in Fig. 17, the XRF measurements are all higher than the results of the ICP-AES experiments. The sensitivity and accuracy of As compared with the existing laws and regulations are not satisfied (Fig. 17), and 25% of the data are in close proximity of each other (Table 2). The regression line, with a slope of 0.4208 and the R2 value of 0.3449, turns greatly away from the 1:1 line and, thus, cannot be shown on the figure (Fig. 18).
Comprehensive interpretation was carried out by overall evaluations of both the raw data of XRF vs. ICP-AES and their regression lines. Table 2 presents the summary of eight elements measured by the XRF method and the results obtained form the ICP-AES experiments for 60 samples from the Erren River watershed.
Based on the interpretation of the individual elements and the summary table (Table 2), the eight heavy metal elements can be approximately divided into two groups. The first group includes Pb, Zn, Ni, and Cu, which the XRF measurements show to have better agreement with the ICP-AES experiment results. The order of reliability by the XRF measurements is Pb > Zn > Ni > Cu. The second group includes As, Cr, Cd, and Hg. The order of reliability by the XRF measurements is As > Cr > Cd > Hg. Table 2 shows that the first group Cu, Ni, Zn, and Pb measured values have good correlation (relative proximity ranges from 35%~85%). The second group, which includes Hg, Cd, Cr, and As, shows poor correlation (relative proximity ranges from 2.3%~25%). On the other hand, the R2 values of the regression line range from 0.4095~0.7281 for the first group and from 0.01143~0.3349 for the second group.
It is also observed that the XRF results are able to sensitively and accurately detect the first group of elements (Pb, Zn, Ni, and Cu), especially, for those values above the threshold limit. However, in some cases, the values which are below the threshold limit do not always mean that the ICP-AES results will be below the threshold limit also. For instance, in Fig. 9 and Fig. 13 (brown-dotted-circle), the values obtained by the ICP-AES experiments are above the threshold limit, while the values obtained by the XRF measurements are below the threshold limit.
å-å…¸ - æª¢è¦-è©³ç´°å-å…¸
The XRF instrument is a powerful tool that can be very effective in the validation of both the absence and presence of certain metal elements. In principle, this instrument could be employed to provide rapid in-situ detection of the presence of toxic metals such as Pb, Ni, As, Cr, Cd, Cu, Zn, and Hg in soil samples (Radu and Diamond, 2009). Data presented in this paper suggest that XRF measurements can be used as a screening technique with high confidence in samples where the first group of element (Pb, Zn, Ni, and Cu) concentrations is well in excess of the threshold limits (Shrivastava et al., 2005). The order of reliability by the XRF measurement for metals is Pb > Zn > Ni > Cu. However, the second group of elements that include Hg, Cd, Cr, and As show poor correlation between the XRF measurements and the ICP-AES experiments. Due to the limitations of the XRF and ICP-AES analytical test methods, the most effective method needs to be verified and calibrated via the supplier's specifications (Shrivastava et al., 2005).