- Email: [email protected]
Evaluation of lines of boron, phosphorus and sulfur by high-resolution continuum source flame atomic absorption spectrometry for plant analysis
Microchemical Journal 109 (2013) 134–138
Contents lists available at SciVerse ScienceDirect
Microchemical Journal journal homepage: www.elsevier.com/locate/microc
Evaluation of lines of boron, phosphorus and sulfur by high-resolution continuum source flame atomic absorption spectrometry for plant analysis Marcos André Bechlin a, José Anchieta Gomes Neto a,⁎, Joaquim Araújo Nóbrega b a b
UNESP, Univ. Estadual Paulista, Analytical Chemistry Department, PO Box 355, 14801-970, Araraquara SP, Brazil Federal University of São Carlos, Department of Chemistry, 13560-970, São Carlos SP, Brazil
a r t i c l e
i n f o
Article history: Received 25 November 2011 Received in revised form 6 March 2012 Accepted 19 March 2012 Available online 23 March 2012 Keywords: Boron Phosphorus Sulfur Medicinal plants HR-CS FAAS
a b s t r a c t The wavelength-integrated absorbance (WIA) and summation of absorbance (∑ lines) of different lines were evaluated to enhance sensitivity and determine B, P and S in medicinal plants by HR-CS FAAS. The lowest LOD for B (0.5 mg L− 1) and P (13.7 mg L− 1) was obtained by integration of lines 249.773 nm (3 pixels) and 247.620 nm (5 pixels), respectively. The ∑ lines for CS at 257.595 nm and 257.958 nm furnished LOD =30.5 mg L− 1, ca. 10% lower than the LOD obtained for the WIA using 257.595 nm and 5 pixels. Data showed the advantage of WIA over ∑ lines to improve sensitivity for all analytes. Under optimized conditions, calibration curves in the 1.0–100 mg L− 1 B and 50.0–2000 mg L− 1 P, S ranges were consistently obtained. Results obtained with the HR-CS FAAS method were in agreement at 98% and 95% confidence level with certified values for B and P, respectively. And results for S were in accordance to non-certified values. Concentrations of B, P, and S in 12 medicinal plants analyzed by the proposed method varied within the 19.4–34.5 mg kg− 1 B, 719–3910 mg kg− 1 P and 1469–7653 mg kg− 1 S ranges. © 2012 Elsevier B.V. All rights reserved.
1. Introduction Foliar diagnosis is an efficient tool in agriculture to manage the mineral nutrition of plants [1]. The monitoring of nutrients in crop leaves allows the identification of deficiency, sufficiency or excess of a given element, optimization of crop production and evaluation of fertilizer supplies [2]. Boron, phosphorus and sulfur are among the main nutrients usually required in foliar diagnosis. Spectrophotometry and inductively coupled plasma optical emission spectrometry are the customary analytical techniques employed in large-scale routine analysis of plants for B [3], P [4], and S [5]. Among the spectrometric techniques available for the determination of most macro- and micronutrients in plants, the line source flame atomic absorption spectrometry (LS-FAAS) is the oldest and most commonly used [6]. However, the difficulty of measuring absorbance of P within the 167.167 nm–178.765 nm, S within the 180.671 nm– 182.565 nm spectral ranges, and the poor sensitivity of non-resonance lines of P at 213.618 nm, 213.547 nm, and 214.914 nm, may be considered as the main drawbacks to the determination of phosphorus [7] and sulfur [8] by LS-FAAS. But these drawbacks can be circumvented measuring at molecular lines of PO and CS by high-resolution continuum source flame atomic absorption spectrometry (HR-CS FAAS) [9–12]. The low sensitivity of B determination by LS-FAAS (results of the small atomization degrees of B in flames) may be improved using the high
⁎ Corresponding author. Tel.: + 55 16 33019611; fax: + 55 16 33019692. E-mail address: [email protected] (J.A. Gomes Neto). 0026-265X/$ – see front matter © 2012 Elsevier B.V. All rights reserved. doi:10.1016/j.microc.2012.03.013
intensity xenon short-arc lamp of the HR-CS FAAS. The sequential multi-element analyses, the summation of absorbance signals [13] of main and secondary atomic lines resulting in a new calibration function and the integration of the absorbance signal over the center pixel by including part of the line wings [14] are facilities of the HR-CS FAAS to improve the sensitivity. For routine laboratories devoted to large-scale analyses, fast sequential multi-element determination is particularly helpful because time and analytical costs may be significantly reduced. Little attention has been given to the use of the summation of absorbance and integration of the absorbance to increase sensitivity in the sequential multi-element determination of B, P and S in plants by HR-CS FAAS. This paper reports an evaluation and application of the HR-CS FAAS technique in the determination of B, P and S for plant analysis. The influence of the nature of line (atomic or molecular), the summation of absorbance at these lines and the variation of number of pixels on sensitivity, accuracy and precision were studied. 2. Materials and methods 2.1. Instrumentation A contrAA 300 (Analytik Jena, Germany) high-resolution atomic absorption spectrometer equipped with a xenon short-arc lamp operating in the “hot-spot” mode as a continuum radiation source was used throughout the work. Air-acetylene flame was used for PO and CS production, and a fuel-rich nitrous oxide–acetylene flame was used for B atomization. High-purity acetylene (99.7% Air Liquid, SP,
M.A. Bechlin et al. / Microchemical Journal 109 (2013) 134–138
Brazil) and high purity nitrous oxide (99.5% Air Liquid) were used as fuel gas and oxidant gas, respectively. All measurements were carried out in seven replicates at aspiration rate fixed at 5.0 mL min − 1. The optimum nitrous oxide–acetylene flow rates were 462 and 245 L h − 1 (B). And the optimum air-acetylene flow rates were 594 and 100 L h − 1 (PO), 594 and 120 L h − 1 (CS). 2.2. Reagents and analytical solutions High purity de-ionized water obtained using a Millipore Rios 5® reverse osmosis and a Millipore Milli-Q™ Academic® deionizer system (resistivity 18.2 MΩ cm, Millipore, Bedford, MA, USA) was used throughout to prepare all solutions. A 5000 mg L − 1 P standard stock solution was prepared by dissolving 34.874 g Na2HPO3·5H2O (Riedel-De Häen, Seelze, Germany) in water and making the volume up to 1000 mL with water. A 5000 mg L − 1 S standard stock solution was prepared by dissolving 38.434 g MgSO4·7H2O (Spectrum, Gardena, CA, USA) in water and making the volume up to 1000 mL with water. This salt was standardized against gravimetry by making a precipitate of barium sulfate and weighing the isolated pure compound. A 1000 mg L− 1 B stock standard solution was prepared by dissolving 5.719 g H3BO3 (Spectrum, Gardena, CA, USA) in water, and making the volume up to 1000 mL with water. A 5000 mg L− 1 Ti stock standard solution was prepared by diluting the Titrisol® standard (Merck, Germany) in 200 mL water. Multi-element analytical solutions (1.0–100.0 mg L − 1 B, 50.0–2000 mg L− 1 P and 50.0–2000 mg L − 1 S) were daily prepared by proper dilution of single stock standard solutions. All solutions were stored in high-density polypropylene bottles (Nalgene®, Rochester, NY, USA). Plastic bottles and glassware materials were cleaned by soaking in 10% (v/v) HNO3 at least 24 h and rinsed abundantly in deionized water before use. 2.3. Sample preparation Medicinal plants of Peumus boldus, Matricaria chamomilla, Baccharis trimera, Echinodorus grandiflorus, Pimpinela anisun, Maytenus ilicifolia, Ginkgo biloba, Panax ginseng, Annona muricata, Mentha spp., Mentha pulegium and Cassia angustifolia were purchased at local market in Araraquara city, SP, Brazil. The certified standard reference materials Trace Elements in Spinach Leaves (1570a) and Apple Leaves (1515) (National Institute of Standards and Technology, Gaithersburg, MD) and samples were mineralized in triplicate using nitric acid and hydrogen peroxide. To 2000 g of dried powdered materials placed into 400-mL digestion tubes, 3.5 mL of 30% (v/v) H2O2 (Spectrum, Gardena, CA, USA) was added and let the mixture stand overnight. Thereafter, 10 mL HNO3 (Spectrum, Gardena, CA, USA) was added and the tubes were placed on a block digestor at 90 °C for 2 h. The temperature was increased to 150 °C and the solution was evaporated to near dryness. The residue was then taken up in water, filtered, added 5 mL of 5000 mg L − 1 Ti, and the volume was completed to 25 mL with water. 2.4. Analytical procedure Measurements were carried out at the main and secondary atomic lines for B (249.773 nm and 249.677 nm), at PO molecular lines for P (246.400; 247.620 and 247.780 nm) and CS molecular lines for S (257.595; 257.958 nm). In HR-CS AAS it is possible to apply the wavelength-integrated absorbance (WIA) over the line core including part of the line wings to enhance sensitivity. The influence of WIA on the linear dynamic range (LDR), sensitivity, linear correlation coefficient (R), characteristic concentration (Co), limit of detection (LOD), limit of quantification (LOQ) and precision was evaluated by varying different number of
135
pixels used for detection: 1 pixel (central pixel, CP), 3 pixels (CP ± 1), 5 pixels (CP ± 2), and 7 pixels (CP ± 3). The integrated area has little differences in width due to the resolution of the equipment; each wavelength has a bandwidth per pixel. Table 1 shows spectrometric data to all elements determined. The WIA was applied to individual and summed lines of all analytes. For 5 mL min − 1 sampling rate, analytical curves in the 1.0–100 mg L − 1 B, 50.0–2000 mg L − 1 P and 50.0–2000 mg L − 1 S concentration ranges were consistently obtained using the lines 249.773 nm, 247.620 nm, and 257.595 nm, and integration equivalent to 3, 5 and 5 pixels, respectively. The limit of detection (LOD) and limit of quantification (LOQ) for all analytes were calculated according to the IUPAC recommendation [15]. 3. Results and discussion The high signal-to-noise ratio of the continuum radiation source, the accessibility to alternate lines for PO and CS molecules, and the possibility to enhance sensitivity with the WIA [14] and summation of absorbance of different lines [13] are favorable attributes to investigate the feasibility of fast sequential multi-element determination of B, P, and S in a single aliquot of plant digests by HR-CS FAAS. The instrumental operating conditions were then maximized employing the most sensitive lines for B (doublet at 249.773 and 249.677 nm), PO (246.400, 247.620 and 247.780 nm) and CS (257.595 and 257.958 nm) [7–9]. The sample preparation procedures were adapted to higher masses of herbal plants taking into consideration the low contents of B, P and S usually found in these plant tissues. The main figures of merit (slope and linear correlation coefficient of the calibration curve, characteristic concentration, limit of detection and relative standard deviation) related to individual lines and summation of lines (∑ lines) at different number of pixels for detection (CP, CP ± 1, CP ± 2, CP ± 3) are described in Tables 2, 3 and 4 for B, P and S, respectively. Analysis of Table 2 reveals that analytical curves built up in the 1.00–100 mg L − 1 B ranges for 249.773 nm, and 2.00–100 mg L − 1 to 249.677 nm and ∑ lines presented linear correlation coefficients better than 0.998. When the number of pixels was ranged from 1 to 7, slopes of analytical curves for 249.773 nm, 249.677 nm and ∑ lines increased from 1.872·10 − 4 to 8.693·10 − 4, 9.661·10 − 5 to 4.387·10 − 4, and 2.839·10 − 4 to 1.310·10 − 3, respectively. And the characteristic concentrations (Co) reduced from 23.5 to 5.1 mg L − 1 (249.773 nm), 45.5 to 10.0 mg L − 1 (249.677 nm), and 15.5 to 3.3 mg L − 1 (∑ lines). For any number of pixels used for detection, the lowest LOD was obtained for the main line at 249.773 nm, followed by the ∑ lines and 249.677 nm. The summation of lines did not contribute significantly to improve the LOD due to the different sensitivity of B atomic lines. In this situation, noise was the major contribution of the secondary line to the main line at 249,773 nm. However the main use of lines with different sensitivities is the expansion of the linear range [16]. The main line at 249.773 nm and WIA equivalent to 3 pixels (CP ± 1) was selected for B determination in further works taking into Table 1 Spectrometric data for the determined elements. Element
Wavelength (nm)
Bandwidth per pixel/pm
B
249.773 249.677 246.400 247.620 247.780 257.595 257.958
1.45 1.45 1.29 1.40 1.40 1.42 1.42
P (PO)
S (CS)
136
M.A. Bechlin et al. / Microchemical Journal 109 (2013) 134–138
Table 2 Linear range (mg L− 1), linear correlation coefficient, characteristic concentration (C0, mg L− 1), limit of detection (LOD, mg L− 1), and relative standard deviation (RSD, %) for B at main and secondary lines and different number of pixels. ∑ lines mean summation of absorbance signals of atomic lines. Wavelength (nm) 1 pixel (CP) 249.773 249.677 ∑ lines 3 pixels (CP ± 1) 249.773 249.677 ∑ lines 5 pixels (CP ± 2) 249.773 249.677 ∑ lines 7 pixels (CP ± 3) 249.773 249.677 ∑ lines
Linear range
Slope/R
C0
LOD
RSD
1.0–100.0 2.0–100.0 2.0–100.0
1.9·10− 4/0.9989 9.7·10− 5/0.9990 2.8·10− 4/0.9989
23.5 45.5 15.5
0.5 3.4 1.2
4.6 5.4 6.2
1.0–100.0 2.0–100.0 2.0–100.0
5.2·10− 4/0.9989 2.7·10− 4/0.9990 7.8·10− 4/0.9989
8.5 16.5 5.6
0.5 2.9 1.0
4.4 5.3 6.9
1.0–100.0 2.0–100.0 2.0–100.0
7.5·10− 4/0.9989 3.8·10− 5/0.9992 1.1·10− 3/0.9990
5.9 11.5 3.9
0.6 3.0 1.1
4.3 5.2 6.4
1.0–100.0 2.0–100.0 2.0–100.0
8.7·10− 4/0.9988 4.4·10− 4/0.9993 1.3·10− 3/0.9990
5.1 10.0 3.3
1.0 3.3 1.3
4.1 5.3 5.7
consideration the lowest LOD (0.5 mg L − 1) and RSD (4.4%). The lowest LOD associated to ∑ lines was 1.0 mg L − 1, almost twofold the LOD observed for the main line at 3 pixels. This emphasizes that the WIA-based strategy is more efficient than the summation of lines to improve sensitivity for B. It is interesting to emphasize that the standard deviation (∑SD blank) used in the equation 3xSD blank/slope to calculate the LOD associated to the ∑ lines was: ∑SD blank = (SD 2 λ1 + SD 2 λ2) 1/2, were λ1 and λ2 were the doublet of B (249.773 nm, 249.677 nm). The proposed sample preparation procedure combined with WIA detected B at a minimum concentration at 5.8 mg kg − 1, almost 7–12 folds lower than the optimum B contents of the leaves for most crops [17]. It should be mentioned that the calculated characteristic concentration (Co) obtained with 3 pixels at the main line (8.5 mg L − 1) is close to that furnished by the manufacturer (6.6 mg L − 1). For comparison purposes, this characteristic concentration is almost 1.5-fold lower than that obtained by LS FAAS (13 mg L − 1) at the same wavelength. It should be mentioned that despite summing atomic lines did not bring advantage to improve limit of detection for B, the summation of several molecular lines may be feasible due to their similarity in
Table 3 Linear range (mg L− 1), linear correlation coefficient, characteristic concentration (C0, mg L− 1), limit of detection (LOD, mg L− 1), and relative standard deviation (RSD, %) for P at different PO lines and different number of pixels. ∑ lines mean summation of absorbance signals of molecular lines. Wavelength (nm) 1 pixel (CP) 246.400 247.620 247.780 ∑ lines 3 pixels (CP ± 1) 246.400 247.620 247.780 ∑ lines 5 pixels (CP ± 2) 246.400 247.620 247.780 ∑ lines 7 pixels (CP ± 3) 246.400 247.620 247.780 ∑ lines
Linear range
Slope/R
C0
LOD
RSD
50.0–2000.0 50.0–2000.0 50.0–2000.0 50.0–2000.0
4.6·10− 6/0.9985 3.0·10− 6/0.9968 3.8·10− 6/0.9988 1.1·10− 5/0.9984
964.1 1171.8 1455.0 388.0
32.9 32.0 39.7 18.5
4.8 2.6 4.8 3.5
50.0–2000.0 50.0–2000.0 50.0–2000.0 50.0–2000.0
1.3·10− 5/0.9986 8.4·10− 6/0.9974 1.1·10− 5/0.9989 3.3·10− 5/0.9985
332.3 394.0 523.9 134.8
27.2 16.4 39.3 15.6
4.7 2.0 4.1 3.1
50.0–2000.0 50.0–2000.0 50.0–2000.0 50.0–2000.0
2.0·10− 5/0.9984 1.2·10− 5/0.9974 1.7·10− 5/0.9989 5.0·10− 5/0.9985
211.3 251.1 358.0 87.8
28.1 13.7 48.8 17.4
4.5 2.2 4.8 3.3
50.0–2000.0 50.0–2000.0 50.0–2000.0 50.0–2000.0
2.6·10− 5/0.9981 1.5·10− 5/0.9978 2.3·10− 5/0.9989 6.3·10− 5/0.9985
172.1 191.3 301.6 69.7
30.5 15.6 76.1 22.3
5.1 3.1 5.7 4.1
Table 4 Linear range (mg L− 1), linear correlation coefficient, characteristic concentration (C0, mg L− 1), limit of detection (LOD, mg L− 1), and relative standard deviation (RSD, %) for S at different CS lines and different number of pixels. ∑ lines mean summation of absorbance signals of molecular lines. Wavelength (nm) 1 pixel (CP) 257.595 257.958 ∑ lines 3 pixels (CP ± 1) 257.595 257.958 ∑ lines 5 pixels (CP ± 2) 257.595 257.958 ∑ lines 7 pixels (CP ± 3) 257.595 257.958 ∑ lines
Linear range
Slope/R
C0
LOD
RSD
50.0–2000.0 50.0–2000.0 50.0–2000.0
2.2·10− 6/0.9997 1.7·10− 6/0.9983 3.9·10− 6/0.9997
1955.5 2614.4 1119.0
40.0 53.5 30.5
5.9 14.6 10.4
50.0–2000.0 50.0–2000.0 50.0–2000.0
6.2·10− 6/0.9996 4.4·10− 6/0.9980 1.0·10− 5/0.9994
636.3 1006.2 416.3
38.7 54.9 31.2
4.4 10.5 6.7
50.0–2000.0 50.0–2000.0 50.0–2000.0
8.6·10− 6/0.9996 5.3·10− 6/0.9969 1.4·10− 5/0.9994
509.1 832.9 315.9
34.7 73.8 39.3
4.6 10.7 6.2
50.0–2000.0 50.0–2000.0 50.0–2000.0
9.6·10− 6/0.9997 4.1·10− 6/0.9963 1.4·10− 5/0.9993
470.1 1069.0 320.0
49.8 138.5 54.5
5.4 22.8 9.0
terms of sensitivity [18,19]. Taking this into consideration, only the most sensitive molecular lines were employed for evaluation. The main figures of merit for P determination are described in Table 3. Analytical curves built up in the 50.0–2000 mg L − 1 P ranges for 246.400 nm, 247.780 nm, 247.620 nm, and ∑ lines presented linear correlation coefficients better than 0.997. Slopes of analytical curves increased ca. 5.6-folds (246.400 nm), 4.8-folds (247.780 nm) and 6.1-folds (247.620 nm) when the number of pixels was ranged from 1 to 7. Sensitivities for ∑ lines also increased accordingly: as expected, the highest slope was obtained for analytical plot built up by summation of 246.400 nm, 247.780 nm and 247.620 nm lines. When the summation of absorbance is not considered, the lower LOD values were obtained the sequence: 247.620 nm b 246.400 nm b 247.780 nm, independently of the number of pixels. For the most sensitive line 247.620 nm, the LOD varied from 13.7 to 15.6 mg L − 1 for WIA in the 3–7 pixels range. It is interesting to note that close LODs were only attained if the strategy of absorbance summation is adopted: for ∑ lines 247.620 nm, 246.400 nm and 247.780 nm, and 3 or 5 pixels, the LOD = 15.6 mg L − 1 or 17.4 mg L − 1, respectively. The line at 247.620 nm and WIA equivalent to 5 pixels (CP ± 2) was selected for P determination in further works taking into consideration the lowest LOD (13.7 mg L− 1) and low RSD (ca. 2%). The lowest LOD associated to ∑ lines was 15.6 mg L− 1, ca. 14% higher the LOD observed for the most sensitive line for PO at 5 pixels. These findings also emphasize the beneficial aspects of WIA over the summation of lines to improve sensitivity for P. The line at 247.620 nm and WIA equivalent to 5 pixels allowed the determination of P in plant tissues containing at least 172 mg kg− 1 (LOD in dry basis), which corresponds to around 5-fold lower the usual content of P in plant leaves for most crops [17]. With regard to sulfur, Table 4 comprises the main figures of merit for CS lines. Analytical curves within the 50.0–2000 mg L− 1 S interval for 257.595 nm, 257.958 nm, and ∑ lines presented linear correlation coefficients better than 0.996. Slopes of analytical curves increased ca. 4.3-folds (257.595 nm), 2.4-folds (257.958 nm) and 3.5-folds (∑ lines) when the number of pixels ranged from 1 to 7. And the characteristic concentrations (Co) reduced from 1955.5 to 470.1 mg L− 1 (257.595 nm), 2614.4 to 832.9 mg L− 1 (257.958 nm), and 1119.0 to 315.9 mg L− 1 (∑ lines). For individual lines, the lower LOD values were obtained for the line 257.595 nm, independently of the number of pixels. For this line, the better LOD was around 35 mg L − 1 for WIA equivalent to 5 pixels range. On the other hand, the summation of absorbance (∑ lines at 257.595 nm and 257.958 nm) using 1 or 3 pixels gave a LOD
M.A. Bechlin et al. / Microchemical Journal 109 (2013) 134–138
close to 31 mg L − 1. It should be emphasize that this LOD was ca. only 10% lower than the LOD obtained for the line 257.595 nm and WIA equivalent to 5 pixels. Additionally, precision associated to the single line using WIA is better than those RSD obtained for ∑ lines. Taking into consideration the practical aspects in using only the single line at 257.595 nm and WIA equivalent to 5 pixels, these conditions were adopted for S determination in further works. Again, these findings reinforce the advantage of WIA over summation of lines to improve sensitivity for S determination. The optimized method allowed the determination of S in plant tissues containing at least 440 mg kg − 1 (LOD in dry basis) which corresponds to around sixfold lower the usual content of S in plant leaves for most crops [17]. The summation of molecular lines to improve LODs for both PO and CS was observed for 1 and 3 pixels in a factor of square root of n, where n are the number of lines summed [14]. Nevertheless, higher integrated areas deteriorated the limit of detection for summed lines. This sensitivity reduction occurred because some lines (PO: 247.780 nm, 246.400 nm; CS: 257.958 nm) are narrower than others.
137
On that situation, the use of more pixels increases noise and makes the strategy unfavorable in comparison to WIA. The summation of great number of molecular lines is required to obtain a significant improvement in the LOD [18,19]. Moreover, the spent time in reading different absorbance and treating data makes the summation unfavorable to improve sensitivity and was not adopted in this work. Shown in Fig. 1 are spectra of studied molecular lines. Analysis of figure reveals that lines 247.620 nm (PO) and 257.595 nm (CS) present broaden peak widths and allowed the use of higher integration area to improve LOD. It should be stressed that the mode to set the baseline is also an important parameter to be considered when monitoring molecular lines. In the present work all measurements employed the same baseline adjustment based on a reference spectrum from a blank solution. The full-width at half maximum (FWHM) may help to understand this subject. The FWHM defines the dispersion of the absorption profile over pixels of an analytical line and the magnitude of the absorbance values at peripheral pixels. For narrow peaks, the majority of analytical signal is within small
Fig. 1. Spectra detail of PO (a, b, c) and CS (d, e) in air/acetylene flame around 246.400 nm (a), 247.620 nm (b), 247.780 nm (c), 257.595 nm (d) and 257.958 nm (e).
138
M.A. Bechlin et al. / Microchemical Journal 109 (2013) 134–138
Table 5 Results (in mg kg−1 ± SD) for B, P and S determination (n = 3) in samples and CRMs by the proposed method. Samples/CRMs
B
P
S
Peumus boldus Matricaria cham. Baccharis trimera Echinodorus gran. Pimpinela anisun Maytenus ilic. Ginkgo biloba Panax ginseng Annona muricata Mentha spp. Mentha pulegium Cassia angu. CRM 1a CRM 2b
b LOQ b LOQ b LOQ b LOQ 19.8 ± 1.9 20.5 ± 1.8 34.5 ± 3.0 b LOQ b LOQ 19.4 ± 2.5 b LOQ 30.1 ± 2.5 24.7 ± 0.6 45.0 ± 2.9
818 ± 103 3910 ± 130 b LOQ 2955 ± 25 2420 ± 56 754 ± 11 1346 ± 61 2032 ± 219 719 ± 165 1479 ± 223 2755 ± 143 b LOQ! 1490 ± 70 5700 ± 300
bLOQ 3143 ± 182 bLOQ bLOQ 2860 ± 35 1542 ± 36 1469 ± 80 2032 ± 40 1862 ± 166 7653 ± 133 2503 ± 83 1486 ± 51 2150 ± 140 5300 ± 400
a 1515 apple leaves. Certified values: 27.0 ± 2.0 mg kg− 1 B; 1590 ± 110 mg kg− 1 P. Non-certified value: 1800 mg kg− 1 S. b 1570a trace elements in spinach leaves. Certified values: 37.6 ± 1.0 mg kg− 1 B; 5180 ± 110 mg kg− 1 P. Non-certified value: 4600 mg kg− 1 S.
wavelength intervals. For higher integration area, the main contribution of peripheral pixels to analytical signal is noise, contributing then to increase the LOD. On the other hand, the dispersion of the absorption profile of broad peaks allows that pixels distant from the core line contribute considerably to gain in analytical signal for higher integration areas. In these situations, better LODs may be attained using more pixels. After selecting the optimum conditions, the method was applied to the determination of B, P, and S in 12 samples of medicinal plants. The found concentration varied from 19.4–34.5 mg kg − 1 B, 719–3910 mg kg − 1 P and 1469–7653 mg kg − 1 S (Table 5). These found levels of B, P, and S in this work were close to those obtained for medicinal plants reported in the literature [1,17,20]. In the determination of P in plants containing high levels of calcium, it is a good practice add Ti in sample solutions in order to avoid the formation of Ca3(PO4)2, that impairs the formation of PO molecules in the flame [21]. Accuracy was checked for B, P and S determination in two plant certified reference materials (Table 5). A paired t-test showed that results obtained with the HR-CS FAAS method were in agreement at 98% and 95% confidence level with certified values for B and P, respectively. And results for S were in accordance to noncertified values for S. The relative standard deviations (n = 12) for a sample (M. ilicifolia) containing 20.5 mg kg− 1 B, 754.9 mg kg− 1 P, and 1542.2 mg kg− 1 S were 8%, 2%, and 2%, respectively. Limits of detection (LOD), calculated according to IUPAC recommendation [15], were 0.5 mg L − 1 B, 13.7 mg L − 1 P, and 34.7 mg L − 1 S, which correspond to (LOD in dry basis) 5.8 mg kg− 1 B, 172 mg kg− 1 P, and 440 mg kg− 1 S, making possible the determination of these elements in low concentrations in medicinal plants. 4. Conclusions The WIA and summation of absorbance signals of different lines may increase sensitivity for B, P and S determination by HR-CS FAAS. The lowest LOD for B and P were obtained after integrating the absorbance at lines 249.773 nm and 247.620 nm using 3 or 5 pixels, respectively. The summation of absorbance for CS lines at 257.595 nm and 257.958 nm furnished a limit of detection only 10% lower than that obtained for the line 257.595 nm using 5 pixels. The use of only WIA was enough and feasible for the determination of B, P and S in single aliquots of plant tissues by HR-CS FAAS.
Acknowledgments The authors would like to thank the FAPESP for financially supporting this work. The authors are also grateful to CNPq for the fellowship to M.A.B., and research ships to J.A.G.N. and J.A.N.
References [1] M.E. Farago, A. Mehra, Plants and the Chemical Elements, Wiley-VCH, Weinheim, 1994. [2] A. Kabata-Pendias, H. Pendias, Trace Elements in Soil and Plants, CRC Press, Boca Raton, FL, 1985. [3] R.N. Sah, P.H. Brown, Boron determination—a review of analytical methods, Microchem. J. 56 (1997) 285–304. [4] Z. Marczenko, Separation and Spectrophotometric Determination of Elements, 2nd ed. Ellis Horwood, Chichester, United States, 1986. [5] I. Novozamsky, R. Vaneck, J. Vanderlee, V. Houba, E. Temminghoff, Determination of total sulfur and extractable sulfate in plant materials by inductively-coupled plasma atomic emission-spectrometry, Commun. Soil Sci. Plant Anal. 17 (1986) 147–1157. [6] M.S. Cresser, Flame Spectrometry in Environmental Chemical Analysis: A Practical Guide, The Royal Society Chemistry, Cambridge, 1994 108 p. [7] M.D. Huang, H. Becker-Ross, S. Florek, U. Heitmann, M. Okruss, Determination of phosphorus by molecular absorption of phosphorus monoxide using a highresolution continuum source absorption spectrometer and an air-acetylene flame, J. Anal. At. Spectrom. 21 (2006) 338–345. [8] A. Virgílio, J.L.R. Júnior, A.A. Cardoso, J.A. Nóbrega, J.A. Gomes Neto, Determination of total sulfur in agricultural samples by high-resolution continuum source flame molecular absorption spectrometry, J. Agric. Food Chem. 59 (2011) 2197–2201. [9] B. Welz, H. Becker-Ross, S. Florek, U. Heitmann, High-Resolution Continuum Source AAS: The Better Way to Do Atomic Absorption Spectrometry, Wiley-VCH, Weinheim, 2005. [10] M.D. Huang, H. Becker-Ross, S. Florek, U. Heitmann, M. Okruss, Determination of sulfur by molecular absorption of carbon monosulfide using a high-resolution continuum source absorption spectrometer and an air-acetylene flame, Spectrochim. Acta Part B 61 (2006) 181–188. [11] B. Welz, F.G. Lepri, R.G.O. Araujo, S.L.C. Ferreira, H. Becker-Ross, M.D. Huang, M. Okruss, Determination of phosphorus, sulfur and the halogens using high-temperature molecular absorption spectrometry in flames and furnaces—a review, Anal. Chim. Acta 647 (2009) 137–148. [12] U. Heitmann, H. Becker-Ross, S. Florek, M.D. Huang, M. Okruss, Determination of non-metals via molecular absorption using high-resolution continuum source absorption spectrometry and graphite furnace atomization, J. Anal. At. Spectrom. 21 (2006) 1314–1320. [13] V.R.A. Filho, V.P. Franzini, J.A. Gomes Neto, Exploring potentialities of the HR-CS FAAS technique in the determination of Ni and Pb in mineral waters, Ecletica Quim. 33 (2008) 49–56. [14] U. Heitmann, B. Welz, D.L.G. Borges, F.G. Lepri, Feasibility of peak volume, side pixel and multiple peak registration in high-resolution continuum source atomic absorption spectrometry, Spectrochim. Acta Part B 62 (2007) 1222–1230. [15] L.A. Currie, Nomenclature in evaluation of analytical methods including detection and quantification capabilities (IUPAC Recommendations 1995), Anal. Chim. Acta 391 (1999) 105–126. [16] M. Resano, L. Rello, M. Flórez, M.A. Belarra, On the possibilities of high-resolution continuum source graphite furnace atomic absorption spectrometry for the simultaneous or sequential monitoring of multiple atomic lines, Spectrochim. Acta Part B 66 (2011) 321–328. [17] E. Malavolta, Manual de nutrição mineral de plantas, Editora Agronômica Ceres, São Paulo, 2006 638 pp. [18] M. Resano, J. Briceño, M.A. Belarra, Direct determination of phosphorus in biological samples using a solid sampling-high resolution-continuum source electrothermal spectrometer: comparison of atomic and molecular absorption spectrometry, J. Anal. At. Spectrom. 24 (2009) 1343–1354. [19] M. Resano, M. Flórez, Direct determination of sulfur in solid samples by means of high-resolution continuum source graphite furnace molecular absorption spectrometry using palladium nanoparticles as chemical modifier, J. Anal. At. Spectrom. 27 (2012) 401–412. [20] R.A. Street, M.G. Kulkarni, W.A. Stirk, C. Southway, J. Van Staden, Variation in heavy metals and microelements in South African medicinal plants obtained from street markets, Food Addit. Contam. 25 (2008) 953–960. [21] M.D. Huang, H. Becker-Ross, S. Florek, U. Heitmann, M. Okruss, The influence of calcium and magnesium on the phosphorus monoxide molecular absorption signal in the determination of phosphorus using a continuum source absorption spectrometer and an air–acetylene flame, J. Anal. At. Spectrom. 21 (2006) 346–349.
Contents lists available at SciVerse ScienceDirect
Microchemical Journal journal homepage: www.elsevier.com/locate/microc
Evaluation of lines of boron, phosphorus and sulfur by high-resolution continuum source flame atomic absorption spectrometry for plant analysis Marcos André Bechlin a, José Anchieta Gomes Neto a,⁎, Joaquim Araújo Nóbrega b a b
UNESP, Univ. Estadual Paulista, Analytical Chemistry Department, PO Box 355, 14801-970, Araraquara SP, Brazil Federal University of São Carlos, Department of Chemistry, 13560-970, São Carlos SP, Brazil
a r t i c l e
i n f o
Article history: Received 25 November 2011 Received in revised form 6 March 2012 Accepted 19 March 2012 Available online 23 March 2012 Keywords: Boron Phosphorus Sulfur Medicinal plants HR-CS FAAS
a b s t r a c t The wavelength-integrated absorbance (WIA) and summation of absorbance (∑ lines) of different lines were evaluated to enhance sensitivity and determine B, P and S in medicinal plants by HR-CS FAAS. The lowest LOD for B (0.5 mg L− 1) and P (13.7 mg L− 1) was obtained by integration of lines 249.773 nm (3 pixels) and 247.620 nm (5 pixels), respectively. The ∑ lines for CS at 257.595 nm and 257.958 nm furnished LOD =30.5 mg L− 1, ca. 10% lower than the LOD obtained for the WIA using 257.595 nm and 5 pixels. Data showed the advantage of WIA over ∑ lines to improve sensitivity for all analytes. Under optimized conditions, calibration curves in the 1.0–100 mg L− 1 B and 50.0–2000 mg L− 1 P, S ranges were consistently obtained. Results obtained with the HR-CS FAAS method were in agreement at 98% and 95% confidence level with certified values for B and P, respectively. And results for S were in accordance to non-certified values. Concentrations of B, P, and S in 12 medicinal plants analyzed by the proposed method varied within the 19.4–34.5 mg kg− 1 B, 719–3910 mg kg− 1 P and 1469–7653 mg kg− 1 S ranges. © 2012 Elsevier B.V. All rights reserved.
1. Introduction Foliar diagnosis is an efficient tool in agriculture to manage the mineral nutrition of plants [1]. The monitoring of nutrients in crop leaves allows the identification of deficiency, sufficiency or excess of a given element, optimization of crop production and evaluation of fertilizer supplies [2]. Boron, phosphorus and sulfur are among the main nutrients usually required in foliar diagnosis. Spectrophotometry and inductively coupled plasma optical emission spectrometry are the customary analytical techniques employed in large-scale routine analysis of plants for B [3], P [4], and S [5]. Among the spectrometric techniques available for the determination of most macro- and micronutrients in plants, the line source flame atomic absorption spectrometry (LS-FAAS) is the oldest and most commonly used [6]. However, the difficulty of measuring absorbance of P within the 167.167 nm–178.765 nm, S within the 180.671 nm– 182.565 nm spectral ranges, and the poor sensitivity of non-resonance lines of P at 213.618 nm, 213.547 nm, and 214.914 nm, may be considered as the main drawbacks to the determination of phosphorus [7] and sulfur [8] by LS-FAAS. But these drawbacks can be circumvented measuring at molecular lines of PO and CS by high-resolution continuum source flame atomic absorption spectrometry (HR-CS FAAS) [9–12]. The low sensitivity of B determination by LS-FAAS (results of the small atomization degrees of B in flames) may be improved using the high
⁎ Corresponding author. Tel.: + 55 16 33019611; fax: + 55 16 33019692. E-mail address: [email protected] (J.A. Gomes Neto). 0026-265X/$ – see front matter © 2012 Elsevier B.V. All rights reserved. doi:10.1016/j.microc.2012.03.013
intensity xenon short-arc lamp of the HR-CS FAAS. The sequential multi-element analyses, the summation of absorbance signals [13] of main and secondary atomic lines resulting in a new calibration function and the integration of the absorbance signal over the center pixel by including part of the line wings [14] are facilities of the HR-CS FAAS to improve the sensitivity. For routine laboratories devoted to large-scale analyses, fast sequential multi-element determination is particularly helpful because time and analytical costs may be significantly reduced. Little attention has been given to the use of the summation of absorbance and integration of the absorbance to increase sensitivity in the sequential multi-element determination of B, P and S in plants by HR-CS FAAS. This paper reports an evaluation and application of the HR-CS FAAS technique in the determination of B, P and S for plant analysis. The influence of the nature of line (atomic or molecular), the summation of absorbance at these lines and the variation of number of pixels on sensitivity, accuracy and precision were studied. 2. Materials and methods 2.1. Instrumentation A contrAA 300 (Analytik Jena, Germany) high-resolution atomic absorption spectrometer equipped with a xenon short-arc lamp operating in the “hot-spot” mode as a continuum radiation source was used throughout the work. Air-acetylene flame was used for PO and CS production, and a fuel-rich nitrous oxide–acetylene flame was used for B atomization. High-purity acetylene (99.7% Air Liquid, SP,
M.A. Bechlin et al. / Microchemical Journal 109 (2013) 134–138
Brazil) and high purity nitrous oxide (99.5% Air Liquid) were used as fuel gas and oxidant gas, respectively. All measurements were carried out in seven replicates at aspiration rate fixed at 5.0 mL min − 1. The optimum nitrous oxide–acetylene flow rates were 462 and 245 L h − 1 (B). And the optimum air-acetylene flow rates were 594 and 100 L h − 1 (PO), 594 and 120 L h − 1 (CS). 2.2. Reagents and analytical solutions High purity de-ionized water obtained using a Millipore Rios 5® reverse osmosis and a Millipore Milli-Q™ Academic® deionizer system (resistivity 18.2 MΩ cm, Millipore, Bedford, MA, USA) was used throughout to prepare all solutions. A 5000 mg L − 1 P standard stock solution was prepared by dissolving 34.874 g Na2HPO3·5H2O (Riedel-De Häen, Seelze, Germany) in water and making the volume up to 1000 mL with water. A 5000 mg L − 1 S standard stock solution was prepared by dissolving 38.434 g MgSO4·7H2O (Spectrum, Gardena, CA, USA) in water and making the volume up to 1000 mL with water. This salt was standardized against gravimetry by making a precipitate of barium sulfate and weighing the isolated pure compound. A 1000 mg L− 1 B stock standard solution was prepared by dissolving 5.719 g H3BO3 (Spectrum, Gardena, CA, USA) in water, and making the volume up to 1000 mL with water. A 5000 mg L− 1 Ti stock standard solution was prepared by diluting the Titrisol® standard (Merck, Germany) in 200 mL water. Multi-element analytical solutions (1.0–100.0 mg L − 1 B, 50.0–2000 mg L− 1 P and 50.0–2000 mg L − 1 S) were daily prepared by proper dilution of single stock standard solutions. All solutions were stored in high-density polypropylene bottles (Nalgene®, Rochester, NY, USA). Plastic bottles and glassware materials were cleaned by soaking in 10% (v/v) HNO3 at least 24 h and rinsed abundantly in deionized water before use. 2.3. Sample preparation Medicinal plants of Peumus boldus, Matricaria chamomilla, Baccharis trimera, Echinodorus grandiflorus, Pimpinela anisun, Maytenus ilicifolia, Ginkgo biloba, Panax ginseng, Annona muricata, Mentha spp., Mentha pulegium and Cassia angustifolia were purchased at local market in Araraquara city, SP, Brazil. The certified standard reference materials Trace Elements in Spinach Leaves (1570a) and Apple Leaves (1515) (National Institute of Standards and Technology, Gaithersburg, MD) and samples were mineralized in triplicate using nitric acid and hydrogen peroxide. To 2000 g of dried powdered materials placed into 400-mL digestion tubes, 3.5 mL of 30% (v/v) H2O2 (Spectrum, Gardena, CA, USA) was added and let the mixture stand overnight. Thereafter, 10 mL HNO3 (Spectrum, Gardena, CA, USA) was added and the tubes were placed on a block digestor at 90 °C for 2 h. The temperature was increased to 150 °C and the solution was evaporated to near dryness. The residue was then taken up in water, filtered, added 5 mL of 5000 mg L − 1 Ti, and the volume was completed to 25 mL with water. 2.4. Analytical procedure Measurements were carried out at the main and secondary atomic lines for B (249.773 nm and 249.677 nm), at PO molecular lines for P (246.400; 247.620 and 247.780 nm) and CS molecular lines for S (257.595; 257.958 nm). In HR-CS AAS it is possible to apply the wavelength-integrated absorbance (WIA) over the line core including part of the line wings to enhance sensitivity. The influence of WIA on the linear dynamic range (LDR), sensitivity, linear correlation coefficient (R), characteristic concentration (Co), limit of detection (LOD), limit of quantification (LOQ) and precision was evaluated by varying different number of
135
pixels used for detection: 1 pixel (central pixel, CP), 3 pixels (CP ± 1), 5 pixels (CP ± 2), and 7 pixels (CP ± 3). The integrated area has little differences in width due to the resolution of the equipment; each wavelength has a bandwidth per pixel. Table 1 shows spectrometric data to all elements determined. The WIA was applied to individual and summed lines of all analytes. For 5 mL min − 1 sampling rate, analytical curves in the 1.0–100 mg L − 1 B, 50.0–2000 mg L − 1 P and 50.0–2000 mg L − 1 S concentration ranges were consistently obtained using the lines 249.773 nm, 247.620 nm, and 257.595 nm, and integration equivalent to 3, 5 and 5 pixels, respectively. The limit of detection (LOD) and limit of quantification (LOQ) for all analytes were calculated according to the IUPAC recommendation [15]. 3. Results and discussion The high signal-to-noise ratio of the continuum radiation source, the accessibility to alternate lines for PO and CS molecules, and the possibility to enhance sensitivity with the WIA [14] and summation of absorbance of different lines [13] are favorable attributes to investigate the feasibility of fast sequential multi-element determination of B, P, and S in a single aliquot of plant digests by HR-CS FAAS. The instrumental operating conditions were then maximized employing the most sensitive lines for B (doublet at 249.773 and 249.677 nm), PO (246.400, 247.620 and 247.780 nm) and CS (257.595 and 257.958 nm) [7–9]. The sample preparation procedures were adapted to higher masses of herbal plants taking into consideration the low contents of B, P and S usually found in these plant tissues. The main figures of merit (slope and linear correlation coefficient of the calibration curve, characteristic concentration, limit of detection and relative standard deviation) related to individual lines and summation of lines (∑ lines) at different number of pixels for detection (CP, CP ± 1, CP ± 2, CP ± 3) are described in Tables 2, 3 and 4 for B, P and S, respectively. Analysis of Table 2 reveals that analytical curves built up in the 1.00–100 mg L − 1 B ranges for 249.773 nm, and 2.00–100 mg L − 1 to 249.677 nm and ∑ lines presented linear correlation coefficients better than 0.998. When the number of pixels was ranged from 1 to 7, slopes of analytical curves for 249.773 nm, 249.677 nm and ∑ lines increased from 1.872·10 − 4 to 8.693·10 − 4, 9.661·10 − 5 to 4.387·10 − 4, and 2.839·10 − 4 to 1.310·10 − 3, respectively. And the characteristic concentrations (Co) reduced from 23.5 to 5.1 mg L − 1 (249.773 nm), 45.5 to 10.0 mg L − 1 (249.677 nm), and 15.5 to 3.3 mg L − 1 (∑ lines). For any number of pixels used for detection, the lowest LOD was obtained for the main line at 249.773 nm, followed by the ∑ lines and 249.677 nm. The summation of lines did not contribute significantly to improve the LOD due to the different sensitivity of B atomic lines. In this situation, noise was the major contribution of the secondary line to the main line at 249,773 nm. However the main use of lines with different sensitivities is the expansion of the linear range [16]. The main line at 249.773 nm and WIA equivalent to 3 pixels (CP ± 1) was selected for B determination in further works taking into Table 1 Spectrometric data for the determined elements. Element
Wavelength (nm)
Bandwidth per pixel/pm
B
249.773 249.677 246.400 247.620 247.780 257.595 257.958
1.45 1.45 1.29 1.40 1.40 1.42 1.42
P (PO)
S (CS)
136
M.A. Bechlin et al. / Microchemical Journal 109 (2013) 134–138
Table 2 Linear range (mg L− 1), linear correlation coefficient, characteristic concentration (C0, mg L− 1), limit of detection (LOD, mg L− 1), and relative standard deviation (RSD, %) for B at main and secondary lines and different number of pixels. ∑ lines mean summation of absorbance signals of atomic lines. Wavelength (nm) 1 pixel (CP) 249.773 249.677 ∑ lines 3 pixels (CP ± 1) 249.773 249.677 ∑ lines 5 pixels (CP ± 2) 249.773 249.677 ∑ lines 7 pixels (CP ± 3) 249.773 249.677 ∑ lines
Linear range
Slope/R
C0
LOD
RSD
1.0–100.0 2.0–100.0 2.0–100.0
1.9·10− 4/0.9989 9.7·10− 5/0.9990 2.8·10− 4/0.9989
23.5 45.5 15.5
0.5 3.4 1.2
4.6 5.4 6.2
1.0–100.0 2.0–100.0 2.0–100.0
5.2·10− 4/0.9989 2.7·10− 4/0.9990 7.8·10− 4/0.9989
8.5 16.5 5.6
0.5 2.9 1.0
4.4 5.3 6.9
1.0–100.0 2.0–100.0 2.0–100.0
7.5·10− 4/0.9989 3.8·10− 5/0.9992 1.1·10− 3/0.9990
5.9 11.5 3.9
0.6 3.0 1.1
4.3 5.2 6.4
1.0–100.0 2.0–100.0 2.0–100.0
8.7·10− 4/0.9988 4.4·10− 4/0.9993 1.3·10− 3/0.9990
5.1 10.0 3.3
1.0 3.3 1.3
4.1 5.3 5.7
consideration the lowest LOD (0.5 mg L − 1) and RSD (4.4%). The lowest LOD associated to ∑ lines was 1.0 mg L − 1, almost twofold the LOD observed for the main line at 3 pixels. This emphasizes that the WIA-based strategy is more efficient than the summation of lines to improve sensitivity for B. It is interesting to emphasize that the standard deviation (∑SD blank) used in the equation 3xSD blank/slope to calculate the LOD associated to the ∑ lines was: ∑SD blank = (SD 2 λ1 + SD 2 λ2) 1/2, were λ1 and λ2 were the doublet of B (249.773 nm, 249.677 nm). The proposed sample preparation procedure combined with WIA detected B at a minimum concentration at 5.8 mg kg − 1, almost 7–12 folds lower than the optimum B contents of the leaves for most crops [17]. It should be mentioned that the calculated characteristic concentration (Co) obtained with 3 pixels at the main line (8.5 mg L − 1) is close to that furnished by the manufacturer (6.6 mg L − 1). For comparison purposes, this characteristic concentration is almost 1.5-fold lower than that obtained by LS FAAS (13 mg L − 1) at the same wavelength. It should be mentioned that despite summing atomic lines did not bring advantage to improve limit of detection for B, the summation of several molecular lines may be feasible due to their similarity in
Table 3 Linear range (mg L− 1), linear correlation coefficient, characteristic concentration (C0, mg L− 1), limit of detection (LOD, mg L− 1), and relative standard deviation (RSD, %) for P at different PO lines and different number of pixels. ∑ lines mean summation of absorbance signals of molecular lines. Wavelength (nm) 1 pixel (CP) 246.400 247.620 247.780 ∑ lines 3 pixels (CP ± 1) 246.400 247.620 247.780 ∑ lines 5 pixels (CP ± 2) 246.400 247.620 247.780 ∑ lines 7 pixels (CP ± 3) 246.400 247.620 247.780 ∑ lines
Linear range
Slope/R
C0
LOD
RSD
50.0–2000.0 50.0–2000.0 50.0–2000.0 50.0–2000.0
4.6·10− 6/0.9985 3.0·10− 6/0.9968 3.8·10− 6/0.9988 1.1·10− 5/0.9984
964.1 1171.8 1455.0 388.0
32.9 32.0 39.7 18.5
4.8 2.6 4.8 3.5
50.0–2000.0 50.0–2000.0 50.0–2000.0 50.0–2000.0
1.3·10− 5/0.9986 8.4·10− 6/0.9974 1.1·10− 5/0.9989 3.3·10− 5/0.9985
332.3 394.0 523.9 134.8
27.2 16.4 39.3 15.6
4.7 2.0 4.1 3.1
50.0–2000.0 50.0–2000.0 50.0–2000.0 50.0–2000.0
2.0·10− 5/0.9984 1.2·10− 5/0.9974 1.7·10− 5/0.9989 5.0·10− 5/0.9985
211.3 251.1 358.0 87.8
28.1 13.7 48.8 17.4
4.5 2.2 4.8 3.3
50.0–2000.0 50.0–2000.0 50.0–2000.0 50.0–2000.0
2.6·10− 5/0.9981 1.5·10− 5/0.9978 2.3·10− 5/0.9989 6.3·10− 5/0.9985
172.1 191.3 301.6 69.7
30.5 15.6 76.1 22.3
5.1 3.1 5.7 4.1
Table 4 Linear range (mg L− 1), linear correlation coefficient, characteristic concentration (C0, mg L− 1), limit of detection (LOD, mg L− 1), and relative standard deviation (RSD, %) for S at different CS lines and different number of pixels. ∑ lines mean summation of absorbance signals of molecular lines. Wavelength (nm) 1 pixel (CP) 257.595 257.958 ∑ lines 3 pixels (CP ± 1) 257.595 257.958 ∑ lines 5 pixels (CP ± 2) 257.595 257.958 ∑ lines 7 pixels (CP ± 3) 257.595 257.958 ∑ lines
Linear range
Slope/R
C0
LOD
RSD
50.0–2000.0 50.0–2000.0 50.0–2000.0
2.2·10− 6/0.9997 1.7·10− 6/0.9983 3.9·10− 6/0.9997
1955.5 2614.4 1119.0
40.0 53.5 30.5
5.9 14.6 10.4
50.0–2000.0 50.0–2000.0 50.0–2000.0
6.2·10− 6/0.9996 4.4·10− 6/0.9980 1.0·10− 5/0.9994
636.3 1006.2 416.3
38.7 54.9 31.2
4.4 10.5 6.7
50.0–2000.0 50.0–2000.0 50.0–2000.0
8.6·10− 6/0.9996 5.3·10− 6/0.9969 1.4·10− 5/0.9994
509.1 832.9 315.9
34.7 73.8 39.3
4.6 10.7 6.2
50.0–2000.0 50.0–2000.0 50.0–2000.0
9.6·10− 6/0.9997 4.1·10− 6/0.9963 1.4·10− 5/0.9993
470.1 1069.0 320.0
49.8 138.5 54.5
5.4 22.8 9.0
terms of sensitivity [18,19]. Taking this into consideration, only the most sensitive molecular lines were employed for evaluation. The main figures of merit for P determination are described in Table 3. Analytical curves built up in the 50.0–2000 mg L − 1 P ranges for 246.400 nm, 247.780 nm, 247.620 nm, and ∑ lines presented linear correlation coefficients better than 0.997. Slopes of analytical curves increased ca. 5.6-folds (246.400 nm), 4.8-folds (247.780 nm) and 6.1-folds (247.620 nm) when the number of pixels was ranged from 1 to 7. Sensitivities for ∑ lines also increased accordingly: as expected, the highest slope was obtained for analytical plot built up by summation of 246.400 nm, 247.780 nm and 247.620 nm lines. When the summation of absorbance is not considered, the lower LOD values were obtained the sequence: 247.620 nm b 246.400 nm b 247.780 nm, independently of the number of pixels. For the most sensitive line 247.620 nm, the LOD varied from 13.7 to 15.6 mg L − 1 for WIA in the 3–7 pixels range. It is interesting to note that close LODs were only attained if the strategy of absorbance summation is adopted: for ∑ lines 247.620 nm, 246.400 nm and 247.780 nm, and 3 or 5 pixels, the LOD = 15.6 mg L − 1 or 17.4 mg L − 1, respectively. The line at 247.620 nm and WIA equivalent to 5 pixels (CP ± 2) was selected for P determination in further works taking into consideration the lowest LOD (13.7 mg L− 1) and low RSD (ca. 2%). The lowest LOD associated to ∑ lines was 15.6 mg L− 1, ca. 14% higher the LOD observed for the most sensitive line for PO at 5 pixels. These findings also emphasize the beneficial aspects of WIA over the summation of lines to improve sensitivity for P. The line at 247.620 nm and WIA equivalent to 5 pixels allowed the determination of P in plant tissues containing at least 172 mg kg− 1 (LOD in dry basis), which corresponds to around 5-fold lower the usual content of P in plant leaves for most crops [17]. With regard to sulfur, Table 4 comprises the main figures of merit for CS lines. Analytical curves within the 50.0–2000 mg L− 1 S interval for 257.595 nm, 257.958 nm, and ∑ lines presented linear correlation coefficients better than 0.996. Slopes of analytical curves increased ca. 4.3-folds (257.595 nm), 2.4-folds (257.958 nm) and 3.5-folds (∑ lines) when the number of pixels ranged from 1 to 7. And the characteristic concentrations (Co) reduced from 1955.5 to 470.1 mg L− 1 (257.595 nm), 2614.4 to 832.9 mg L− 1 (257.958 nm), and 1119.0 to 315.9 mg L− 1 (∑ lines). For individual lines, the lower LOD values were obtained for the line 257.595 nm, independently of the number of pixels. For this line, the better LOD was around 35 mg L − 1 for WIA equivalent to 5 pixels range. On the other hand, the summation of absorbance (∑ lines at 257.595 nm and 257.958 nm) using 1 or 3 pixels gave a LOD
M.A. Bechlin et al. / Microchemical Journal 109 (2013) 134–138
close to 31 mg L − 1. It should be emphasize that this LOD was ca. only 10% lower than the LOD obtained for the line 257.595 nm and WIA equivalent to 5 pixels. Additionally, precision associated to the single line using WIA is better than those RSD obtained for ∑ lines. Taking into consideration the practical aspects in using only the single line at 257.595 nm and WIA equivalent to 5 pixels, these conditions were adopted for S determination in further works. Again, these findings reinforce the advantage of WIA over summation of lines to improve sensitivity for S determination. The optimized method allowed the determination of S in plant tissues containing at least 440 mg kg − 1 (LOD in dry basis) which corresponds to around sixfold lower the usual content of S in plant leaves for most crops [17]. The summation of molecular lines to improve LODs for both PO and CS was observed for 1 and 3 pixels in a factor of square root of n, where n are the number of lines summed [14]. Nevertheless, higher integrated areas deteriorated the limit of detection for summed lines. This sensitivity reduction occurred because some lines (PO: 247.780 nm, 246.400 nm; CS: 257.958 nm) are narrower than others.
137
On that situation, the use of more pixels increases noise and makes the strategy unfavorable in comparison to WIA. The summation of great number of molecular lines is required to obtain a significant improvement in the LOD [18,19]. Moreover, the spent time in reading different absorbance and treating data makes the summation unfavorable to improve sensitivity and was not adopted in this work. Shown in Fig. 1 are spectra of studied molecular lines. Analysis of figure reveals that lines 247.620 nm (PO) and 257.595 nm (CS) present broaden peak widths and allowed the use of higher integration area to improve LOD. It should be stressed that the mode to set the baseline is also an important parameter to be considered when monitoring molecular lines. In the present work all measurements employed the same baseline adjustment based on a reference spectrum from a blank solution. The full-width at half maximum (FWHM) may help to understand this subject. The FWHM defines the dispersion of the absorption profile over pixels of an analytical line and the magnitude of the absorbance values at peripheral pixels. For narrow peaks, the majority of analytical signal is within small
Fig. 1. Spectra detail of PO (a, b, c) and CS (d, e) in air/acetylene flame around 246.400 nm (a), 247.620 nm (b), 247.780 nm (c), 257.595 nm (d) and 257.958 nm (e).
138
M.A. Bechlin et al. / Microchemical Journal 109 (2013) 134–138
Table 5 Results (in mg kg−1 ± SD) for B, P and S determination (n = 3) in samples and CRMs by the proposed method. Samples/CRMs
B
P
S
Peumus boldus Matricaria cham. Baccharis trimera Echinodorus gran. Pimpinela anisun Maytenus ilic. Ginkgo biloba Panax ginseng Annona muricata Mentha spp. Mentha pulegium Cassia angu. CRM 1a CRM 2b
b LOQ b LOQ b LOQ b LOQ 19.8 ± 1.9 20.5 ± 1.8 34.5 ± 3.0 b LOQ b LOQ 19.4 ± 2.5 b LOQ 30.1 ± 2.5 24.7 ± 0.6 45.0 ± 2.9
818 ± 103 3910 ± 130 b LOQ 2955 ± 25 2420 ± 56 754 ± 11 1346 ± 61 2032 ± 219 719 ± 165 1479 ± 223 2755 ± 143 b LOQ! 1490 ± 70 5700 ± 300
bLOQ 3143 ± 182 bLOQ bLOQ 2860 ± 35 1542 ± 36 1469 ± 80 2032 ± 40 1862 ± 166 7653 ± 133 2503 ± 83 1486 ± 51 2150 ± 140 5300 ± 400
a 1515 apple leaves. Certified values: 27.0 ± 2.0 mg kg− 1 B; 1590 ± 110 mg kg− 1 P. Non-certified value: 1800 mg kg− 1 S. b 1570a trace elements in spinach leaves. Certified values: 37.6 ± 1.0 mg kg− 1 B; 5180 ± 110 mg kg− 1 P. Non-certified value: 4600 mg kg− 1 S.
wavelength intervals. For higher integration area, the main contribution of peripheral pixels to analytical signal is noise, contributing then to increase the LOD. On the other hand, the dispersion of the absorption profile of broad peaks allows that pixels distant from the core line contribute considerably to gain in analytical signal for higher integration areas. In these situations, better LODs may be attained using more pixels. After selecting the optimum conditions, the method was applied to the determination of B, P, and S in 12 samples of medicinal plants. The found concentration varied from 19.4–34.5 mg kg − 1 B, 719–3910 mg kg − 1 P and 1469–7653 mg kg − 1 S (Table 5). These found levels of B, P, and S in this work were close to those obtained for medicinal plants reported in the literature [1,17,20]. In the determination of P in plants containing high levels of calcium, it is a good practice add Ti in sample solutions in order to avoid the formation of Ca3(PO4)2, that impairs the formation of PO molecules in the flame [21]. Accuracy was checked for B, P and S determination in two plant certified reference materials (Table 5). A paired t-test showed that results obtained with the HR-CS FAAS method were in agreement at 98% and 95% confidence level with certified values for B and P, respectively. And results for S were in accordance to noncertified values for S. The relative standard deviations (n = 12) for a sample (M. ilicifolia) containing 20.5 mg kg− 1 B, 754.9 mg kg− 1 P, and 1542.2 mg kg− 1 S were 8%, 2%, and 2%, respectively. Limits of detection (LOD), calculated according to IUPAC recommendation [15], were 0.5 mg L − 1 B, 13.7 mg L − 1 P, and 34.7 mg L − 1 S, which correspond to (LOD in dry basis) 5.8 mg kg− 1 B, 172 mg kg− 1 P, and 440 mg kg− 1 S, making possible the determination of these elements in low concentrations in medicinal plants. 4. Conclusions The WIA and summation of absorbance signals of different lines may increase sensitivity for B, P and S determination by HR-CS FAAS. The lowest LOD for B and P were obtained after integrating the absorbance at lines 249.773 nm and 247.620 nm using 3 or 5 pixels, respectively. The summation of absorbance for CS lines at 257.595 nm and 257.958 nm furnished a limit of detection only 10% lower than that obtained for the line 257.595 nm using 5 pixels. The use of only WIA was enough and feasible for the determination of B, P and S in single aliquots of plant tissues by HR-CS FAAS.
Acknowledgments The authors would like to thank the FAPESP for financially supporting this work. The authors are also grateful to CNPq for the fellowship to M.A.B., and research ships to J.A.G.N. and J.A.N.
References [1] M.E. Farago, A. Mehra, Plants and the Chemical Elements, Wiley-VCH, Weinheim, 1994. [2] A. Kabata-Pendias, H. Pendias, Trace Elements in Soil and Plants, CRC Press, Boca Raton, FL, 1985. [3] R.N. Sah, P.H. Brown, Boron determination—a review of analytical methods, Microchem. J. 56 (1997) 285–304. [4] Z. Marczenko, Separation and Spectrophotometric Determination of Elements, 2nd ed. Ellis Horwood, Chichester, United States, 1986. [5] I. Novozamsky, R. Vaneck, J. Vanderlee, V. Houba, E. Temminghoff, Determination of total sulfur and extractable sulfate in plant materials by inductively-coupled plasma atomic emission-spectrometry, Commun. Soil Sci. Plant Anal. 17 (1986) 147–1157. [6] M.S. Cresser, Flame Spectrometry in Environmental Chemical Analysis: A Practical Guide, The Royal Society Chemistry, Cambridge, 1994 108 p. [7] M.D. Huang, H. Becker-Ross, S. Florek, U. Heitmann, M. Okruss, Determination of phosphorus by molecular absorption of phosphorus monoxide using a highresolution continuum source absorption spectrometer and an air-acetylene flame, J. Anal. At. Spectrom. 21 (2006) 338–345. [8] A. Virgílio, J.L.R. Júnior, A.A. Cardoso, J.A. Nóbrega, J.A. Gomes Neto, Determination of total sulfur in agricultural samples by high-resolution continuum source flame molecular absorption spectrometry, J. Agric. Food Chem. 59 (2011) 2197–2201. [9] B. Welz, H. Becker-Ross, S. Florek, U. Heitmann, High-Resolution Continuum Source AAS: The Better Way to Do Atomic Absorption Spectrometry, Wiley-VCH, Weinheim, 2005. [10] M.D. Huang, H. Becker-Ross, S. Florek, U. Heitmann, M. Okruss, Determination of sulfur by molecular absorption of carbon monosulfide using a high-resolution continuum source absorption spectrometer and an air-acetylene flame, Spectrochim. Acta Part B 61 (2006) 181–188. [11] B. Welz, F.G. Lepri, R.G.O. Araujo, S.L.C. Ferreira, H. Becker-Ross, M.D. Huang, M. Okruss, Determination of phosphorus, sulfur and the halogens using high-temperature molecular absorption spectrometry in flames and furnaces—a review, Anal. Chim. Acta 647 (2009) 137–148. [12] U. Heitmann, H. Becker-Ross, S. Florek, M.D. Huang, M. Okruss, Determination of non-metals via molecular absorption using high-resolution continuum source absorption spectrometry and graphite furnace atomization, J. Anal. At. Spectrom. 21 (2006) 1314–1320. [13] V.R.A. Filho, V.P. Franzini, J.A. Gomes Neto, Exploring potentialities of the HR-CS FAAS technique in the determination of Ni and Pb in mineral waters, Ecletica Quim. 33 (2008) 49–56. [14] U. Heitmann, B. Welz, D.L.G. Borges, F.G. Lepri, Feasibility of peak volume, side pixel and multiple peak registration in high-resolution continuum source atomic absorption spectrometry, Spectrochim. Acta Part B 62 (2007) 1222–1230. [15] L.A. Currie, Nomenclature in evaluation of analytical methods including detection and quantification capabilities (IUPAC Recommendations 1995), Anal. Chim. Acta 391 (1999) 105–126. [16] M. Resano, L. Rello, M. Flórez, M.A. Belarra, On the possibilities of high-resolution continuum source graphite furnace atomic absorption spectrometry for the simultaneous or sequential monitoring of multiple atomic lines, Spectrochim. Acta Part B 66 (2011) 321–328. [17] E. Malavolta, Manual de nutrição mineral de plantas, Editora Agronômica Ceres, São Paulo, 2006 638 pp. [18] M. Resano, J. Briceño, M.A. Belarra, Direct determination of phosphorus in biological samples using a solid sampling-high resolution-continuum source electrothermal spectrometer: comparison of atomic and molecular absorption spectrometry, J. Anal. At. Spectrom. 24 (2009) 1343–1354. [19] M. Resano, M. Flórez, Direct determination of sulfur in solid samples by means of high-resolution continuum source graphite furnace molecular absorption spectrometry using palladium nanoparticles as chemical modifier, J. Anal. At. Spectrom. 27 (2012) 401–412. [20] R.A. Street, M.G. Kulkarni, W.A. Stirk, C. Southway, J. Van Staden, Variation in heavy metals and microelements in South African medicinal plants obtained from street markets, Food Addit. Contam. 25 (2008) 953–960. [21] M.D. Huang, H. Becker-Ross, S. Florek, U. Heitmann, M. Okruss, The influence of calcium and magnesium on the phosphorus monoxide molecular absorption signal in the determination of phosphorus using a continuum source absorption spectrometer and an air–acetylene flame, J. Anal. At. Spectrom. 21 (2006) 346–349.