- Email: [email protected]
Optimisation of digestion method for determination of arsenic in shrimp paste sample using atomic absorption spectrometry
Food Chemistry 134 (2012) 2406–2410
Contents lists available at SciVerse ScienceDirect
Food Chemistry journal homepage: www.elsevier.com/locate/foodchem
Analytical Methods
Optimisation of digestion method for determination of arsenic in shrimp paste sample using atomic absorption spectrometry C.W. Zanariah C.W. Ngah ⇑, Mohd Adib Yahya 1 Faculty of Science and Technology, Universiti Sains Islam Malaysia, Bandar Baru Nilai, 71800 Negeri Sembilan, Malaysia
a r t i c l e
i n f o
Article history: Received 13 January 2011 Received in revised form 17 November 2011 Accepted 7 April 2012 Available online 17 April 2012 Keywords: Microwave system Optimisation Half factorial design As Atomic absorption spectrometry Shrimp paste Design ExpertÒ 7.0
a b s t r a c t The microwave digestion method was developed and verified for the determination of arsenic in shrimp paste samples. Experimental design for five factors (HNO3 and H2O2 volumes, sample weight, microwave power and digestion time) were used for the optimisation of sample digestion. For this purpose, two level half factorial design, which involves 16 experiments, was adopted. The concentration of arsenic was analysed by graphite furnace atomic absorption spectrometry. Design ExpertÒ 7.0 software was used to interpret all data obtained. The combination of 2 mL HNO3 and 1 mL H2O2 volumes, 0.1 g sample weight, 1400 W power and 5 min digestion time was found to be the optimum parameters required to digest the shrimp paste samples. Tests with spiked samples presented good recoveries with relative standard deviations between 0.32% and 5.35%. Ó 2012 Elsevier Ltd. All rights reserved.
1. Introduction Arsenic (As) speciation in marine ecosystem has been the subject of much attention over the past 20 years. Seafood was also identified as a source of major exposure to As through human consumption (Leufroy, Noël, Dufailly, Beauchemin, & Gueˇrin, 2011). As is known to cause negative effects to human body even at a low intake level (Tuzen, Citak, Mendil, & Soylak, 2009). It may also be found as a result of industrial applications, including leather, wood treatment and pesticides. This element can be readily transformed by microbes, changes in geochemical conditions, and other environmental processes. Anthropogenic As contamination results from activities such as metals and alloys manufacturing, refining petroleum, and burning fossil fuels and wastes. These activities have created a strong legacy of As pollution throughout the world. As also bio-accumulates in the marine food chain resulting in higher levels of this element in bivalves and fish than in most other foods (Melamed, 2005). Sample preparation is one the most time consuming and critical step of an analytical procedure. Despite all recent advances it still requires further development to reach the same high standards of all instrumental techniques required for accurate determination of ⇑ Corresponding author. Tel.: +60 19476342; fax: +60 67986465. E-mail addresses: [email protected] (C.W.Z.C.W. Ngah), amirulsyazwan_ [email protected] (M.A. Yahya). 1 Tel.: +60 27647340; fax: +60 67986465. 0308-8146/$ - see front matter Ó 2012 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.foodchem.2012.04.032
analytes. Microwave digestion system has improved the digestion process in sample preparations (Costa, Santos, Hatje, Nobrega, & Korn, 2009; Soylak, Narin, Bezerra, & Ferreira, 2005). Apart from shorter time and better recovery including volatile elements, microwave technology also provide researchers with minimal contamination, volume of reagent consumption, good working environment (Agazzi & Pirola, 2000), and minimal residue or waste (Costa, Ferreiraa, Nogueirab, & Braz, 2005). Factorial design experiment and microwave-assisted digestion can be applied to accelerate the pretreatment steps and improve the results accuracy. Parameters like sample weight, volumes of acid, power applied, and time of digestion that can influence the process of digestion can be optimised by allowing the consideration of overall number of experiments and possible interactions effects between the involved parameters (Bezerra, Santelli, Oliveira, Villar, & Escaleira, 2008; Costa et al., 2005; Jalbani, Kazi, Arain, Jamali, Afridi, & Sarfraz, 2006). Shrimp paste or locally known as belacan is made from fermented small shrimp (Acetes and Mysad varieties) with salt, and then pulverised into a viscous paste. It is widely used as condiment or main ingredient in most Malaysian cooking (Sharif, Ghazali, Rajab, Haron, & Osman, 2008). The same products have also been produced in other countries in Asia with different names like trassi in Indonesia, bagoong in the Phillippines, padoc in Laos and prahoc in the Republic of Khmer (Lee, Steinkraus, & Reilly, 1993). The aim of this study was to optimise the digestion procedures for the determination of As in shrimp paste samples. Optimisation of the factors affecting the digestion process (HNO3 and H2O2
C.W.Z.C.W. Ngah et al, / Food Chemistry 134 (2012) 2406–2410
volumes, sample weight, microwave power and digestion time) using a complete factorial design requires many experiments. The two-level half factorial design allows the influences of each preparation to be observed at a variety of other variables levels, as well as interactions among the variables without a substantial loss of information. Moreover, this design is an adequate tool to optimise variables affecting sample digestion. Also, it can make the analyst’s task easier and quicker and provide a more reliable answer to the analytical question posed especially when a large number of variables need to be evaluated (Momen, Zachariadis, Anthemidis, & Straits, 2007). 2. Materials and methods 2.1. Instruments A Perkin-Elmer Analyst 800 atomic absorption spectrometer (AAS) equipped with graphite furnace (GFAAS) and AS 800 autosampler was used in this study. The operating conditions of the spectrometer for the determination of arsenic were: Lamp EDL; Current 380 mA; Wavelength 193.7 nm; Slit 0.7 nm; Mode AABG; Processing peak area; Read time 5.0 s; Replicates 3. Graphite furnace atomisation was conducted using 20 lL sample volume and 5 lL matrix modifier volume. The temperature program was: drying 30 s at 110 °C + 30 s at 130 °C; pyrolysis 20 s at 1200 °C (all with gas flow 250 mL/min); atomisation 5 s at 2000 °C with zero gas flow; cleaning 3 s at 2450 °C with 250 mL/min gas flow. A Perkin Elmer microwave reaction system model Anton Paar Multiwave 3000, programmable for time and power between 600 and 1400 W, equipped with eight high-pressure quartz vessels was used to digest the samples. 2.2. Chemicals and sample preparation 65% HNO3 and 30% H2O2 (Fisher Scientific) were used. Standard solutions of As were prepared by diluting a stock solution of 1000 mg L1 supplied by Perkin Elmer. Ultrapure water Mili-QÒ System Milipore (PURELAB classic, ELGA) was used throughout the experiment (Manjusha, Dash, & Karunasagar, 2007). Quartz vessels were cleaned by soaking overnight in 10% (v/v) HNO3 followed by rinsing with ultrapure water and dried before used. Shrimp paste (belacan) samples were purchased from local shops at Negeri Sembilan, Malaysia. The samples were homogenised using Panasonic blender (MX-335) and then dried overnight at temperatures 115–125 °F using an ExcaliburÒ Food Dehydrator Parallexx (USA). 2.3. Microwave digestion procedure Considering the microwave-assisted digestion, a two level half factorial design for five factors 25–1 (16 runs) was developed in order to determine the influence of the factors and their interactions to the concentration of As analysis. Design ExpertÒ 7.0 software was used to analyse all the data obtained. Regression model was developed to calculate the predicted values of the As concentration (Montgomery, 2001). Low () and high levels (+) of each factor were chosen according to literatures reviewed (Araujo, Gonzalez, Ferreira, Nogueira, & Nobrega, 2002; Costa et al., 2005; Gonzalez, ˇ aga Souza, Oliveira, Forato, Nobrega, & Nogueira, 2009; Sola-Larran & Navarro-Blasco, 2009). The low and high level values for all parameters or factors involved in digestion procedures were as follows: factor A (volume of HNO3) were 2 and 4 mL; factor B (volume of H2O2) were 1 and 2 mL; factor C (sample weight) were 0.1 and 0.5 g; factor D (power) were 600 and 1400 W and factor E (time) were 5 and 10 min, respectively.
2407
2.4. Arsenic measurement by GFAAS Palladium and magnesium (nitrate salts) in concentrations of 5 and 3 lg L1 were used as chemical modifiers (Bohrer, Becker, do Nascimento, Dessuy, & de Carvalho, 2006) for the determination of As in the shrimp paste after digestion. The steps involved in the optimisation procedure were done and the most effective procedure was selected to optimise the digestion process. For the recovery study, samples spiked with 2000 lg L1 of As standard solution were analysed by GFAAS following the conditions described in Section 2.1. The recovery study was done to verify the method to ensure that every measurement in the analysis is close enough to the unknown true value of the content of the analyte. Classical approaches to validation only check performance against reference values (de Souza, Pinto, & Junqueira, 2007; Gonzalez & Herrador, 2007; Polati et al., 2005), which do not reflect the nature of certain authentic samples such as used in this study. Therefore, recovery studies were conducted to verify the accuracy of the proposed method. 3. Results and discussion 3.1. Optimisation of the digestion factors To identify which main factors and interactions between 2-factors had the largest effects on the concentration of As in the samples, a total of 15 estimate effects were calculated. The estimate effects that were negligible were assumed to be normally distributed, with mean zero and variance r2 would tend to fall along a straight line on the plot shown in Fig. 1, whereas significant effects would have nonzero means and would not lie along the straight line (Montgomery, 2001). Only four estimate effects (C, D, CE, and DE) were found to be significant, and others were assumed to be insignificant. Therefore, more attention was taken for these four estimate effects during sample preparation steps to provide maximum analyte recovery. Five model terms; A, D, E, AD, and CE showed positive signs suggesting 33.3% of the total effects were found to have better performances at high levels. While the other ten, which were B, C, AB, AC, AE, BC, BD, BE, CD, and DE showed negative signs suggesting 66.7% of the total effects were found to have better performances at low levels. Percentages contribution values indicates that main factor C (sample weight at low level) dominated the process of digestion. The significant effects were analysed by ANOVA and the main factors C (sample weight), D (power) and the interactions of CE and DE were found to be significant with pvalues less than 0.005. The significant values suggest that factor C and D at low and high levels, respectively, contribute significantly to the digestion process of shrimp paste for As analysis. Table 1 summarises the experimental matrix, the observed, predicted and residual values of As concentrations. The observed values were the actual As concentration analysed in this experiment, and the predicted values were obtained from the regression model in Eq. (1). The residual values were the difference between the observed and predicted values. Highest predicted values (3.96 mg L1) were observed in experiments 5 and 6 that correlate well with the observed values of 3.88 and 4.03 mg L1, respectively. The average for predicted values was 2.66 mg L1, similar as the average of the observed values, indicating that the regression model used to obtain the predicted values was accurate.
^As ¼ 2:66 ð0:99Þx3 þ ð0:12Þx4 þ ð0:11Þx3 x5 ð0:079Þx4 x5 y
ð1Þ
where 2.66 is the average from the total 16 observations and the x3 represents Ceffect, x4 reperesents Deffect, x3x5 represents CEinteraction, ˆAs represents the predicted As and x4x5 represents DEinteraction. y concentration.
2408
C.W.Z.C.W. Ngah et al, / Food Chemistry 134 (2012) 2406–2410
Fig. 1. Normal% probability plot of effect estimate for arsenic analysis.
Table 1 Experimental matrix, observed, predicted and residual values for As concentration. Experimental runs
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16
Experimental matrix A
B
C
D
E
+ + + + + + + +
+ + + + + + + +
+ + + + + + + +
+ + + + + + + +
+ + + + + + + +
Observed value, yarsenic
ˆarsenic Predicted value, y
ˆarsenic Residual, e = yarsenic–y
3.68 1.36 3.49 3.56 3.88 4.03 3.55 1.70 1.71 1.69 1.85 3.44 1.81 1.90 3.57 1.33
3.58 1.36 3.50 3.56 3.96 3.96 3.58 1.74 1.76 1.82 1.74 3.50 1.76 1.82 3.56 1.36
0.10 0.00 0.01 0.00 0.08 0.07 0.03 0.04 0.05 0.13 0.11 0.06 0.05 0.08 0.01 0.03
The results can be graphically analysed by constructing the response surface and contour plots. These plots would display the numbers of predicted values which were found from the regression model. Fig. 2a shows the response surface of CE interaction, horizontal axes represent factors C and E while vertical axis represent the predicted As concentration. It indicates that the lowest concentration of As would be obtained when both factors were at low levels. Whereas in contour plot of CE interaction (Fig. 2b), horizontal axes represent both factors C and As concentration, while vertical axis represent the factor E. The displayed values represent the predicted As concentrations. Focusing on factor C axis, the predicted values decreases with the increase of sample weight which mean that the recovery of As was higher when small sample weight was digested. Fig. 3a shows the response surface of DE interaction, horizontal axes represents factors D and E whereas vertical axis represents the predicted As concentrations. The predicted As concentration was achieved at the area in which factors D was at high level and factor E at low level, respectively. This agrees with the re-
sponse surface of DE which shows that factor E is better operated at low level. In contour plot of DE interaction (Fig. 3b), horizontal axes represents the factor D and As concentration, and vertical axis represents the factor E. The contour curves were facing to the high level of factor D with the increasing values of As concentrations. This indicates that factor D (at low and high levels) had significant effects to the digestion process as confirmed by the ANOVA. Evidently, high level of factor D contributed to the effectiveness of the digestions process of shrimp paste samples which contributed to higher concentration of As analysed in the samples. Since factors A and B had no significant effect on the digestion process, therefore it is best to use low levels of both factors as the best option to reduce the chemical consumption. The low and high level values of both factors had no significant difference indicating that the As concentration could be the same for both levels. Consequently, 2 mL HNO3 (A: low level), 1 mL of H2O2 (B: low level), 0.1 g of sample weight (C: low level), 1400 W of power (D: high level) and duration of 5 min (E: low level) were suggested as the optimum parameters for use in the digestion procedure for
2409
C.W.Z.C.W. Ngah et al, / Food Chemistry 134 (2012) 2406–2410
Fig. 2. (a) Response surface for CE interactions (b) contour plot of CE interaction.
determination of As in shrimp paste. Since shrimp paste matrices comprises of organic and inorganic compositions (Sharif et al., 2008), and previous research have shown the need of employing different sample digestion methods according to the sample matrix (Bizzi, Barin, Müller, Schmidt, Nobrega, & Flores, 2010; Gonzalez, Souza, Oliveira, Forato, Nobrega, & Nogueira, 2009) therefore the digestion parameter obtained here could be applied to analyse As in different samples that have similar kind of matrices.
Fig. 3. (a) Response surface for DE interactions (b) contour plot of DE interaction.
Table 2 Results of As recovery studies.
3.2. Recovery studies b
The precision of the digestion method was determined by replicate determination of seven shrimp paste samples (purchased from different shops) spiked with 2000 lg L1 standard As. Spiking was done before the digestion process; each spiked sample was digested according to the parameter optimised in Section 3.1. The calculated R.S.D. for all samples was found in the range between 0.32% and 5.35% (Table 2). The results indicate that the developed digestion method offers good precision for arsenic recoveries in shrimp paste samples. 4. Conclusion The microwave digestion method which was developed and verified for determination of As in shrimp paste samples proved to be simple and effective (2 mL HNO3, 1 mL H2O2, 0.1 g of sample weight, 1400 W power, and 5 min digestion time). This method requires minimal amount of chemicals and the combination of small amounts of sample (0.1 g) with higher microwave power (1400 W), rapid (total digestion time <30 min), and good recoveries (R.S.D.
Samples
Recovery 1 (%)
Recovery 2 (%)
Mean (%)
Standard deviation
RSD (%)b
1 2 3 4 5 6 7
111.00 123.72 128.87 92.06 106.84 111.41 110.78
110.50 118.38 127.20 97.83 99.05 116.02 112.68
110.75 121.05 128.04 94.95 102.95 113.72 111.73
0.35 3.78 1.18 4.08 5.51 3.26 1.34
0.32 3.12 0.92 4.30 5.35 2.87 1.20
Relative standard deviation.
<5.5%) for determination of As in shrimp paste samples. Also, it can be considered as a useful technique for routine analysis of As in fermented seafood and dried seafood samples possibly with similar matrices. The application of experimental design may contribute to a good relation of cost and benefit and also permit the analyst to estimate the optimum conditions to reassure quality including accuracy and precision. The two level half factorial design was proven to be a powerful tool to optimise the variables affecting samples digestion. This design allows the influences of each preparation to be observed at a variety of other variables levels, as well as interactions among the variables without a substantial loss of information. Acknowledgement We would like to thank the Universiti Sains Islam Malaysia (USIM) for the Research fund (Grant PPPP (V) 2006).
2410
C.W.Z.C.W. Ngah et al, / Food Chemistry 134 (2012) 2406–2410
References Agazzi, A., & Pirola, C. (2000). Fundamentals, methods and future trends of environmental microwave sample preparation. Microchemical Journal, 67, 337–341. Araujo, G. C. L., Gonzalez, M. N., Ferreira, A. G., Nogueira, A. R. A., & Nobrega, J. A. (2002). Effect of acid concentration on closed-vessel microwave-assisted digestion of plant materials. Spectrochimica Acta Part B, 57, 2121–2132. Bezerra, M. A., Santelli, R. E., Oliveira, E. P., Villar, L. S., & Escaleira, L. A. (2008). Response surface methodology (RSM) as a tool for optimization in analytical chemistry. Talanta, 76, 965–977. Bizzi, C. A., Barin, J. S., Mülle, R. E. I., Schmidt, L., Nobrega, J. A., & Flores, E. M. M. (2010). Available from: http://dx.doi.org/10.1016/j.talanta.2010.10.015. Bohrer, D., Becker, E., do Nascimento, P. C., Dessuy, M., & de Carvalho, L. M. (2006). Comparison of graphite furnace and hydride generation atomic absorption spectrometry for the determination of selenium status in chicken meat. Food Chemistry. http://dx.doi.org/10.1016/j.foodchem.2006.10.008. Costa, L. M., Santos, D. C. M. B., Hatje, V., Nobrega, J. A., & Korn, M. A. A. (2009). Focused-microwave-assisted acid digestion: Evaluation of losses of volatile elements in marine invertebrate samples. Journal of Food Composition and Analysis, 22, 238–241. Costa, L. M., Ferreiraa, S. L. C., Nogueirab, A. R. A., & Braz, J. A. N. J. (2005). Use of factorial design for optimization of microwave-assisted digestion of lubricating oil. Chemical Society, 16(6A), 1269–1274. de Souza, S. V. C., Pinto, C. T., & Junqueira, R. G. (2007). In-house method validation: Application in arsenic analysis. Journal of Food Composition and Analysis, 20, 241–247. ˘ . (2007). A practical guide to analytical method Gonza˘lez, A. G., & Herrador, M. A validation, including measurement uncertainty and accuracy profiles. Trends in Analytical Chemistry, 26(3), 227–238. Gonza˘lez, M. H., Souza, G. B., Oliveira, R. V., Forato, L. A., Nobrega, J. A., & Nogueira, A. R. A. (2009). Microwave-assisted digestion procedures for biological samples with diluted nitric acid: Identification of reaction products. Talanta, 79, 396–401. Jalbani, N., Kazi, T. G., Arain, B. M., Jamali, M. K., Afridi, H. I., & Sarfraz, R. A. (2006). Application of factorial design in optimization of ultrasonic-assisted extraction of aluminum in juices and soft drinks. Talanta, 70, 307–314.
Lee, C. H., Steinkraus, K. H. & Reilly, P. J. A. (1993). Fish fermentation technology. Available from: http://books.google.com.my/books?id=Gt_QnYUQbWcC& printsec=backcover&dq=shrimp+paste&lr [Accessed 31.12.07]. Leufroy, A., Noël, L., Dufailly, V., Beauchemin, D., & Gueˇrin, T. (2011). Determination of seven arsenic species in seafood by ion exchange chromatography coupled to inductive coupled plasma–mass spectrometry following microwave assisted extraction; method validation and occurrence data. Talanta, 83, 770–779. Manjusha, R., Dash, K., & Karunasagar, D. (2007). UV-photolysis assisted digestion of food samples for the determination of selenium by electrothermal atomic absorption spectrometry (ETAAS). Food Chemistry, 105, 260–265. Melamed, D. (2005). Monitoring arsenic in the environment: A review of science and technologies with the potential for field measurements. Analytica Chimica Acta, 532, 1–13. Momen, A. A., Zachariadis, G. A., Anthemidis, A. N., & Straits, J. A. (2007). Use of fractional factorial design for optimization of digestion procedures followed by multi-element determination of essential and non-essential elements in nuts using ICP–OES technique. Talanta, 71, 443–451. Montgomery, D. C. (2001). Design and analysis of experiments (5th ed.). USA: John Wiley and Sons. Polati, S., Roz, M., Angioi, S., Gianotti, V., Gosetti, F., Marengo, E., et al. (2005). Statistical evaluation of recovery of 3,4-dichloroanilline in soil as function of particle size and analyte concentration. Talanta, 68(1), 93–98. Sharif, R., Ghazali, A. R., Rajab, N. F., Haron, H., & Osman, F. (2008). Toxicological evaluation of some Malaysian locally processed raw food products. Food and Chemical Toxicology, 46, 368–374. Sola-Larranˇaga, C., & Navarro-Blasco, I. (2009). Optimizations of slurry dispersion method for minerals and trace elements analysis in infant formulae by ICP and FAAS. Food Chemistry, 115, 1048–1055. Soylak, M., Narin, I., Bezerra, M. D. A., & Ferreira, S. L. C. (2005). Factorial design in the optimization of preconcentration procedure for lead determination by FAAS. Talanta, 65, 895–899. Tuzen, M., Citak, D., Mendil, D., & Soylak, M. (2009). Arsenic speciation in natural water samples by co-precipitation-hydride generation atomic absorption spectrometry combination. Talanta, 78, 52–56.
Contents lists available at SciVerse ScienceDirect
Food Chemistry journal homepage: www.elsevier.com/locate/foodchem
Analytical Methods
Optimisation of digestion method for determination of arsenic in shrimp paste sample using atomic absorption spectrometry C.W. Zanariah C.W. Ngah ⇑, Mohd Adib Yahya 1 Faculty of Science and Technology, Universiti Sains Islam Malaysia, Bandar Baru Nilai, 71800 Negeri Sembilan, Malaysia
a r t i c l e
i n f o
Article history: Received 13 January 2011 Received in revised form 17 November 2011 Accepted 7 April 2012 Available online 17 April 2012 Keywords: Microwave system Optimisation Half factorial design As Atomic absorption spectrometry Shrimp paste Design ExpertÒ 7.0
a b s t r a c t The microwave digestion method was developed and verified for the determination of arsenic in shrimp paste samples. Experimental design for five factors (HNO3 and H2O2 volumes, sample weight, microwave power and digestion time) were used for the optimisation of sample digestion. For this purpose, two level half factorial design, which involves 16 experiments, was adopted. The concentration of arsenic was analysed by graphite furnace atomic absorption spectrometry. Design ExpertÒ 7.0 software was used to interpret all data obtained. The combination of 2 mL HNO3 and 1 mL H2O2 volumes, 0.1 g sample weight, 1400 W power and 5 min digestion time was found to be the optimum parameters required to digest the shrimp paste samples. Tests with spiked samples presented good recoveries with relative standard deviations between 0.32% and 5.35%. Ó 2012 Elsevier Ltd. All rights reserved.
1. Introduction Arsenic (As) speciation in marine ecosystem has been the subject of much attention over the past 20 years. Seafood was also identified as a source of major exposure to As through human consumption (Leufroy, Noël, Dufailly, Beauchemin, & Gueˇrin, 2011). As is known to cause negative effects to human body even at a low intake level (Tuzen, Citak, Mendil, & Soylak, 2009). It may also be found as a result of industrial applications, including leather, wood treatment and pesticides. This element can be readily transformed by microbes, changes in geochemical conditions, and other environmental processes. Anthropogenic As contamination results from activities such as metals and alloys manufacturing, refining petroleum, and burning fossil fuels and wastes. These activities have created a strong legacy of As pollution throughout the world. As also bio-accumulates in the marine food chain resulting in higher levels of this element in bivalves and fish than in most other foods (Melamed, 2005). Sample preparation is one the most time consuming and critical step of an analytical procedure. Despite all recent advances it still requires further development to reach the same high standards of all instrumental techniques required for accurate determination of ⇑ Corresponding author. Tel.: +60 19476342; fax: +60 67986465. E-mail addresses: [email protected] (C.W.Z.C.W. Ngah), amirulsyazwan_ [email protected] (M.A. Yahya). 1 Tel.: +60 27647340; fax: +60 67986465. 0308-8146/$ - see front matter Ó 2012 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.foodchem.2012.04.032
analytes. Microwave digestion system has improved the digestion process in sample preparations (Costa, Santos, Hatje, Nobrega, & Korn, 2009; Soylak, Narin, Bezerra, & Ferreira, 2005). Apart from shorter time and better recovery including volatile elements, microwave technology also provide researchers with minimal contamination, volume of reagent consumption, good working environment (Agazzi & Pirola, 2000), and minimal residue or waste (Costa, Ferreiraa, Nogueirab, & Braz, 2005). Factorial design experiment and microwave-assisted digestion can be applied to accelerate the pretreatment steps and improve the results accuracy. Parameters like sample weight, volumes of acid, power applied, and time of digestion that can influence the process of digestion can be optimised by allowing the consideration of overall number of experiments and possible interactions effects between the involved parameters (Bezerra, Santelli, Oliveira, Villar, & Escaleira, 2008; Costa et al., 2005; Jalbani, Kazi, Arain, Jamali, Afridi, & Sarfraz, 2006). Shrimp paste or locally known as belacan is made from fermented small shrimp (Acetes and Mysad varieties) with salt, and then pulverised into a viscous paste. It is widely used as condiment or main ingredient in most Malaysian cooking (Sharif, Ghazali, Rajab, Haron, & Osman, 2008). The same products have also been produced in other countries in Asia with different names like trassi in Indonesia, bagoong in the Phillippines, padoc in Laos and prahoc in the Republic of Khmer (Lee, Steinkraus, & Reilly, 1993). The aim of this study was to optimise the digestion procedures for the determination of As in shrimp paste samples. Optimisation of the factors affecting the digestion process (HNO3 and H2O2
C.W.Z.C.W. Ngah et al, / Food Chemistry 134 (2012) 2406–2410
volumes, sample weight, microwave power and digestion time) using a complete factorial design requires many experiments. The two-level half factorial design allows the influences of each preparation to be observed at a variety of other variables levels, as well as interactions among the variables without a substantial loss of information. Moreover, this design is an adequate tool to optimise variables affecting sample digestion. Also, it can make the analyst’s task easier and quicker and provide a more reliable answer to the analytical question posed especially when a large number of variables need to be evaluated (Momen, Zachariadis, Anthemidis, & Straits, 2007). 2. Materials and methods 2.1. Instruments A Perkin-Elmer Analyst 800 atomic absorption spectrometer (AAS) equipped with graphite furnace (GFAAS) and AS 800 autosampler was used in this study. The operating conditions of the spectrometer for the determination of arsenic were: Lamp EDL; Current 380 mA; Wavelength 193.7 nm; Slit 0.7 nm; Mode AABG; Processing peak area; Read time 5.0 s; Replicates 3. Graphite furnace atomisation was conducted using 20 lL sample volume and 5 lL matrix modifier volume. The temperature program was: drying 30 s at 110 °C + 30 s at 130 °C; pyrolysis 20 s at 1200 °C (all with gas flow 250 mL/min); atomisation 5 s at 2000 °C with zero gas flow; cleaning 3 s at 2450 °C with 250 mL/min gas flow. A Perkin Elmer microwave reaction system model Anton Paar Multiwave 3000, programmable for time and power between 600 and 1400 W, equipped with eight high-pressure quartz vessels was used to digest the samples. 2.2. Chemicals and sample preparation 65% HNO3 and 30% H2O2 (Fisher Scientific) were used. Standard solutions of As were prepared by diluting a stock solution of 1000 mg L1 supplied by Perkin Elmer. Ultrapure water Mili-QÒ System Milipore (PURELAB classic, ELGA) was used throughout the experiment (Manjusha, Dash, & Karunasagar, 2007). Quartz vessels were cleaned by soaking overnight in 10% (v/v) HNO3 followed by rinsing with ultrapure water and dried before used. Shrimp paste (belacan) samples were purchased from local shops at Negeri Sembilan, Malaysia. The samples were homogenised using Panasonic blender (MX-335) and then dried overnight at temperatures 115–125 °F using an ExcaliburÒ Food Dehydrator Parallexx (USA). 2.3. Microwave digestion procedure Considering the microwave-assisted digestion, a two level half factorial design for five factors 25–1 (16 runs) was developed in order to determine the influence of the factors and their interactions to the concentration of As analysis. Design ExpertÒ 7.0 software was used to analyse all the data obtained. Regression model was developed to calculate the predicted values of the As concentration (Montgomery, 2001). Low () and high levels (+) of each factor were chosen according to literatures reviewed (Araujo, Gonzalez, Ferreira, Nogueira, & Nobrega, 2002; Costa et al., 2005; Gonzalez, ˇ aga Souza, Oliveira, Forato, Nobrega, & Nogueira, 2009; Sola-Larran & Navarro-Blasco, 2009). The low and high level values for all parameters or factors involved in digestion procedures were as follows: factor A (volume of HNO3) were 2 and 4 mL; factor B (volume of H2O2) were 1 and 2 mL; factor C (sample weight) were 0.1 and 0.5 g; factor D (power) were 600 and 1400 W and factor E (time) were 5 and 10 min, respectively.
2407
2.4. Arsenic measurement by GFAAS Palladium and magnesium (nitrate salts) in concentrations of 5 and 3 lg L1 were used as chemical modifiers (Bohrer, Becker, do Nascimento, Dessuy, & de Carvalho, 2006) for the determination of As in the shrimp paste after digestion. The steps involved in the optimisation procedure were done and the most effective procedure was selected to optimise the digestion process. For the recovery study, samples spiked with 2000 lg L1 of As standard solution were analysed by GFAAS following the conditions described in Section 2.1. The recovery study was done to verify the method to ensure that every measurement in the analysis is close enough to the unknown true value of the content of the analyte. Classical approaches to validation only check performance against reference values (de Souza, Pinto, & Junqueira, 2007; Gonzalez & Herrador, 2007; Polati et al., 2005), which do not reflect the nature of certain authentic samples such as used in this study. Therefore, recovery studies were conducted to verify the accuracy of the proposed method. 3. Results and discussion 3.1. Optimisation of the digestion factors To identify which main factors and interactions between 2-factors had the largest effects on the concentration of As in the samples, a total of 15 estimate effects were calculated. The estimate effects that were negligible were assumed to be normally distributed, with mean zero and variance r2 would tend to fall along a straight line on the plot shown in Fig. 1, whereas significant effects would have nonzero means and would not lie along the straight line (Montgomery, 2001). Only four estimate effects (C, D, CE, and DE) were found to be significant, and others were assumed to be insignificant. Therefore, more attention was taken for these four estimate effects during sample preparation steps to provide maximum analyte recovery. Five model terms; A, D, E, AD, and CE showed positive signs suggesting 33.3% of the total effects were found to have better performances at high levels. While the other ten, which were B, C, AB, AC, AE, BC, BD, BE, CD, and DE showed negative signs suggesting 66.7% of the total effects were found to have better performances at low levels. Percentages contribution values indicates that main factor C (sample weight at low level) dominated the process of digestion. The significant effects were analysed by ANOVA and the main factors C (sample weight), D (power) and the interactions of CE and DE were found to be significant with pvalues less than 0.005. The significant values suggest that factor C and D at low and high levels, respectively, contribute significantly to the digestion process of shrimp paste for As analysis. Table 1 summarises the experimental matrix, the observed, predicted and residual values of As concentrations. The observed values were the actual As concentration analysed in this experiment, and the predicted values were obtained from the regression model in Eq. (1). The residual values were the difference between the observed and predicted values. Highest predicted values (3.96 mg L1) were observed in experiments 5 and 6 that correlate well with the observed values of 3.88 and 4.03 mg L1, respectively. The average for predicted values was 2.66 mg L1, similar as the average of the observed values, indicating that the regression model used to obtain the predicted values was accurate.
^As ¼ 2:66 ð0:99Þx3 þ ð0:12Þx4 þ ð0:11Þx3 x5 ð0:079Þx4 x5 y
ð1Þ
where 2.66 is the average from the total 16 observations and the x3 represents Ceffect, x4 reperesents Deffect, x3x5 represents CEinteraction, ˆAs represents the predicted As and x4x5 represents DEinteraction. y concentration.
2408
C.W.Z.C.W. Ngah et al, / Food Chemistry 134 (2012) 2406–2410
Fig. 1. Normal% probability plot of effect estimate for arsenic analysis.
Table 1 Experimental matrix, observed, predicted and residual values for As concentration. Experimental runs
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16
Experimental matrix A
B
C
D
E
+ + + + + + + +
+ + + + + + + +
+ + + + + + + +
+ + + + + + + +
+ + + + + + + +
Observed value, yarsenic
ˆarsenic Predicted value, y
ˆarsenic Residual, e = yarsenic–y
3.68 1.36 3.49 3.56 3.88 4.03 3.55 1.70 1.71 1.69 1.85 3.44 1.81 1.90 3.57 1.33
3.58 1.36 3.50 3.56 3.96 3.96 3.58 1.74 1.76 1.82 1.74 3.50 1.76 1.82 3.56 1.36
0.10 0.00 0.01 0.00 0.08 0.07 0.03 0.04 0.05 0.13 0.11 0.06 0.05 0.08 0.01 0.03
The results can be graphically analysed by constructing the response surface and contour plots. These plots would display the numbers of predicted values which were found from the regression model. Fig. 2a shows the response surface of CE interaction, horizontal axes represent factors C and E while vertical axis represent the predicted As concentration. It indicates that the lowest concentration of As would be obtained when both factors were at low levels. Whereas in contour plot of CE interaction (Fig. 2b), horizontal axes represent both factors C and As concentration, while vertical axis represent the factor E. The displayed values represent the predicted As concentrations. Focusing on factor C axis, the predicted values decreases with the increase of sample weight which mean that the recovery of As was higher when small sample weight was digested. Fig. 3a shows the response surface of DE interaction, horizontal axes represents factors D and E whereas vertical axis represents the predicted As concentrations. The predicted As concentration was achieved at the area in which factors D was at high level and factor E at low level, respectively. This agrees with the re-
sponse surface of DE which shows that factor E is better operated at low level. In contour plot of DE interaction (Fig. 3b), horizontal axes represents the factor D and As concentration, and vertical axis represents the factor E. The contour curves were facing to the high level of factor D with the increasing values of As concentrations. This indicates that factor D (at low and high levels) had significant effects to the digestion process as confirmed by the ANOVA. Evidently, high level of factor D contributed to the effectiveness of the digestions process of shrimp paste samples which contributed to higher concentration of As analysed in the samples. Since factors A and B had no significant effect on the digestion process, therefore it is best to use low levels of both factors as the best option to reduce the chemical consumption. The low and high level values of both factors had no significant difference indicating that the As concentration could be the same for both levels. Consequently, 2 mL HNO3 (A: low level), 1 mL of H2O2 (B: low level), 0.1 g of sample weight (C: low level), 1400 W of power (D: high level) and duration of 5 min (E: low level) were suggested as the optimum parameters for use in the digestion procedure for
2409
C.W.Z.C.W. Ngah et al, / Food Chemistry 134 (2012) 2406–2410
Fig. 2. (a) Response surface for CE interactions (b) contour plot of CE interaction.
determination of As in shrimp paste. Since shrimp paste matrices comprises of organic and inorganic compositions (Sharif et al., 2008), and previous research have shown the need of employing different sample digestion methods according to the sample matrix (Bizzi, Barin, Müller, Schmidt, Nobrega, & Flores, 2010; Gonzalez, Souza, Oliveira, Forato, Nobrega, & Nogueira, 2009) therefore the digestion parameter obtained here could be applied to analyse As in different samples that have similar kind of matrices.
Fig. 3. (a) Response surface for DE interactions (b) contour plot of DE interaction.
Table 2 Results of As recovery studies.
3.2. Recovery studies b
The precision of the digestion method was determined by replicate determination of seven shrimp paste samples (purchased from different shops) spiked with 2000 lg L1 standard As. Spiking was done before the digestion process; each spiked sample was digested according to the parameter optimised in Section 3.1. The calculated R.S.D. for all samples was found in the range between 0.32% and 5.35% (Table 2). The results indicate that the developed digestion method offers good precision for arsenic recoveries in shrimp paste samples. 4. Conclusion The microwave digestion method which was developed and verified for determination of As in shrimp paste samples proved to be simple and effective (2 mL HNO3, 1 mL H2O2, 0.1 g of sample weight, 1400 W power, and 5 min digestion time). This method requires minimal amount of chemicals and the combination of small amounts of sample (0.1 g) with higher microwave power (1400 W), rapid (total digestion time <30 min), and good recoveries (R.S.D.
Samples
Recovery 1 (%)
Recovery 2 (%)
Mean (%)
Standard deviation
RSD (%)b
1 2 3 4 5 6 7
111.00 123.72 128.87 92.06 106.84 111.41 110.78
110.50 118.38 127.20 97.83 99.05 116.02 112.68
110.75 121.05 128.04 94.95 102.95 113.72 111.73
0.35 3.78 1.18 4.08 5.51 3.26 1.34
0.32 3.12 0.92 4.30 5.35 2.87 1.20
Relative standard deviation.
<5.5%) for determination of As in shrimp paste samples. Also, it can be considered as a useful technique for routine analysis of As in fermented seafood and dried seafood samples possibly with similar matrices. The application of experimental design may contribute to a good relation of cost and benefit and also permit the analyst to estimate the optimum conditions to reassure quality including accuracy and precision. The two level half factorial design was proven to be a powerful tool to optimise the variables affecting samples digestion. This design allows the influences of each preparation to be observed at a variety of other variables levels, as well as interactions among the variables without a substantial loss of information. Acknowledgement We would like to thank the Universiti Sains Islam Malaysia (USIM) for the Research fund (Grant PPPP (V) 2006).
2410
C.W.Z.C.W. Ngah et al, / Food Chemistry 134 (2012) 2406–2410
References Agazzi, A., & Pirola, C. (2000). Fundamentals, methods and future trends of environmental microwave sample preparation. Microchemical Journal, 67, 337–341. Araujo, G. C. L., Gonzalez, M. N., Ferreira, A. G., Nogueira, A. R. A., & Nobrega, J. A. (2002). Effect of acid concentration on closed-vessel microwave-assisted digestion of plant materials. Spectrochimica Acta Part B, 57, 2121–2132. Bezerra, M. A., Santelli, R. E., Oliveira, E. P., Villar, L. S., & Escaleira, L. A. (2008). Response surface methodology (RSM) as a tool for optimization in analytical chemistry. Talanta, 76, 965–977. Bizzi, C. A., Barin, J. S., Mülle, R. E. I., Schmidt, L., Nobrega, J. A., & Flores, E. M. M. (2010). Available from: http://dx.doi.org/10.1016/j.talanta.2010.10.015. Bohrer, D., Becker, E., do Nascimento, P. C., Dessuy, M., & de Carvalho, L. M. (2006). Comparison of graphite furnace and hydride generation atomic absorption spectrometry for the determination of selenium status in chicken meat. Food Chemistry. http://dx.doi.org/10.1016/j.foodchem.2006.10.008. Costa, L. M., Santos, D. C. M. B., Hatje, V., Nobrega, J. A., & Korn, M. A. A. (2009). Focused-microwave-assisted acid digestion: Evaluation of losses of volatile elements in marine invertebrate samples. Journal of Food Composition and Analysis, 22, 238–241. Costa, L. M., Ferreiraa, S. L. C., Nogueirab, A. R. A., & Braz, J. A. N. J. (2005). Use of factorial design for optimization of microwave-assisted digestion of lubricating oil. Chemical Society, 16(6A), 1269–1274. de Souza, S. V. C., Pinto, C. T., & Junqueira, R. G. (2007). In-house method validation: Application in arsenic analysis. Journal of Food Composition and Analysis, 20, 241–247. ˘ . (2007). A practical guide to analytical method Gonza˘lez, A. G., & Herrador, M. A validation, including measurement uncertainty and accuracy profiles. Trends in Analytical Chemistry, 26(3), 227–238. Gonza˘lez, M. H., Souza, G. B., Oliveira, R. V., Forato, L. A., Nobrega, J. A., & Nogueira, A. R. A. (2009). Microwave-assisted digestion procedures for biological samples with diluted nitric acid: Identification of reaction products. Talanta, 79, 396–401. Jalbani, N., Kazi, T. G., Arain, B. M., Jamali, M. K., Afridi, H. I., & Sarfraz, R. A. (2006). Application of factorial design in optimization of ultrasonic-assisted extraction of aluminum in juices and soft drinks. Talanta, 70, 307–314.
Lee, C. H., Steinkraus, K. H. & Reilly, P. J. A. (1993). Fish fermentation technology. Available from: http://books.google.com.my/books?id=Gt_QnYUQbWcC& printsec=backcover&dq=shrimp+paste&lr [Accessed 31.12.07]. Leufroy, A., Noël, L., Dufailly, V., Beauchemin, D., & Gueˇrin, T. (2011). Determination of seven arsenic species in seafood by ion exchange chromatography coupled to inductive coupled plasma–mass spectrometry following microwave assisted extraction; method validation and occurrence data. Talanta, 83, 770–779. Manjusha, R., Dash, K., & Karunasagar, D. (2007). UV-photolysis assisted digestion of food samples for the determination of selenium by electrothermal atomic absorption spectrometry (ETAAS). Food Chemistry, 105, 260–265. Melamed, D. (2005). Monitoring arsenic in the environment: A review of science and technologies with the potential for field measurements. Analytica Chimica Acta, 532, 1–13. Momen, A. A., Zachariadis, G. A., Anthemidis, A. N., & Straits, J. A. (2007). Use of fractional factorial design for optimization of digestion procedures followed by multi-element determination of essential and non-essential elements in nuts using ICP–OES technique. Talanta, 71, 443–451. Montgomery, D. C. (2001). Design and analysis of experiments (5th ed.). USA: John Wiley and Sons. Polati, S., Roz, M., Angioi, S., Gianotti, V., Gosetti, F., Marengo, E., et al. (2005). Statistical evaluation of recovery of 3,4-dichloroanilline in soil as function of particle size and analyte concentration. Talanta, 68(1), 93–98. Sharif, R., Ghazali, A. R., Rajab, N. F., Haron, H., & Osman, F. (2008). Toxicological evaluation of some Malaysian locally processed raw food products. Food and Chemical Toxicology, 46, 368–374. Sola-Larranˇaga, C., & Navarro-Blasco, I. (2009). Optimizations of slurry dispersion method for minerals and trace elements analysis in infant formulae by ICP and FAAS. Food Chemistry, 115, 1048–1055. Soylak, M., Narin, I., Bezerra, M. D. A., & Ferreira, S. L. C. (2005). Factorial design in the optimization of preconcentration procedure for lead determination by FAAS. Talanta, 65, 895–899. Tuzen, M., Citak, D., Mendil, D., & Soylak, M. (2009). Arsenic speciation in natural water samples by co-precipitation-hydride generation atomic absorption spectrometry combination. Talanta, 78, 52–56.