摘要:

ANALYST, SEPTEMBER 1993, VOL. 118 1175 Decomposition of Biological Samples for Inductively Coupled Plasma Atomic Emission Spectrometry Using an Open Focused Microwave Digestion System Antoaneta Krushevska, Ramon M. Barnes and Chitra Amarasiriwaradena University of Massachusetts, Department of Chemistry, Lederle Graduate Research Center, Amherst, MA 0 7 003-0035, USA A focused microwave digestion system operated at atmospheric pressure was applied to the preparation of milk, total parenteral nutrition, tissues (mussel, kidney, oyster and bovine liver) and urine. Reagent combinations (HN03, H2S04 and H202) and power-time programmes were examined with respect to the residual carbon content (RCC) and element recovery. Inductively coupled plasma atomic emission spectrometry was used to determine the residual carbon and analytes (As, Ba, Ca, Cd, Cu, Fe, K, Mg, Mn, Na, P, Pb, Sr and Zn).Different reagents (HN03, HCI and ethylenediaminetetraacetic acid) were investigated for the final digestion step in order to improve the accuracy in the determination of elements forming low-solubility sulfates. The RCC obtained with an open-focused microwave system was similar to that obtained with high-pressure digestion, but lower than with a closed, medium-pressure microwave system. Keywords: Trace analysis; inductively coupled plasma atomic emission spectrometry; open-focused microwave digestion; biological sample; residual carbon Sample digestion is an important step in elemental analysis, because of the preparation time and possibilities for system- atic errors it contributes to the analysis.Common digestion methods include conductive heating at atmospheric pressure (i.e., hot-plate) with mineral acids (e.g., HN03, HC104 and H2SO4) or oxidants (H202) or at elevated pressure and temperature with pressure containers or a commercial high- pressure digestion apparatus (HPA; Anton Paar, Graz, Austria)' with HN03. Microwave heating for sample preparation has advantages compared with hot-plate digestion: the digestion time is shorter, reagent consumption and blank level are lower, and accuracy and reproducibility are better.2 Two types of microwave system are used: closed vessels at elevated pressure and open vessels at atmospheric pressure. In the closed-vessel systems the microwave energy is dispersed throughout the cavity, whereas in the commercial open-vessel system the energy is focused on the part of the digestion vessel containing the sample.Versions of the open-focused micro- wave system can digest large sample masses (e.g., up to 5 g in the Microdigest and 10 g in Maxidigest).3 Many applications for biological materials in closed systems under pressure4-6 and in open systems at atmospheric pres- sure3,7,8 have been described. The most common reagents for this type of digestion are HN03, H2SO4 and H202.8 The effectiveness of decomposition of samples containing organic matter can be evaluated not only from the inorganic analyte recoveries, but also from the completeness of the destruction of the organic matter. Although the residual carbon content (RCC) is not crucial for some techniques such as inductively coupled plasma atomic emission spectrometry (ICP-AES) or flame atomic absorption spectrometry, RCC can introduce errors with other measuring techniques such as voltammetry or in preconcentration methods when extraction and hydride generation are applied.A comparison of different sample decomposition procedures for milk, based on the recoveries of RCC and Zn, was reported recently.9 Some systematic errors result from chemical reactions of the analytes during digestion. For example, elements forming sparingly soluble sulfates ( i e . , Ba, Pb and Sr) can precipitate whenever H2SO4 is used in the digestion. The loss of Pb when using HN03, HClO4 and H2SO4 in the decomposition of plant materials by conductive heating is due to the presence of Ba and the precipitation of sulfates.10 The problem of determina- tion of Pb in poly(viny1 chloride) (PVC) with a high alkaline-earth content has been resolved through the addition of ethylenediaminetetraacetic acid, sodium salt (NaEDTA)11J2 or HN0312 in the final stage.The influence of HN03 on the solubility of PbS04 was described in an enrichment technique for Pb.13 Hence, systematic errors in the determination of alkaline-earth elements resulting from the dissolution of biological materials with H2SO4 might be minimized with similar techniques. An open-focused microwave system was investigated for the determination of As, Ba, Ca, Cd, Cu, Fe, K, Mg, Mn, Na, P, Pb, Sr and Zn in biological materials. The procedure was compared with other sample-digestion techniques.Experimental Apparatus The apparatus used for sample digestion include an open- focused microwave system (Microdigest M 300, Prolabo, Paris, France), with 100% power at 300 W, and 50 ml glass vessels; a closed-vessel microwave system (RMS 150, Floyd, Lake Wylie, SC, USA) with double-wall, medium-pressure Teflon-PFA (perfluoroalkoxy) vessels in which pressure-con- trol mode is used and 100% power corresponds to 600 W; and the HPA with 70 ml quartz vessels. The operating conditions for each of these systems are listed in Table 1. The ICP-AES determinations of the analytes and carbon were performed with an automated, dual monochromator (Plasma 11, Perkin-Elmer, Norwalk, CT, USA). The working conditions are presented in Table 2. Myers-Tracy signal compensation14 was used to measure the background and the reference Sc I1 424.683 nm line from 50 pg ml-1 of Sc simultaneously.Four replicate readings were usually taken. An element analyser (Perkin-Elmer 240) was used to determine total carbon in the original samples. A vacuum freeze-dryer (Virtis, Model 10-800, Gardiner, NY, USA) was used to remove water from the samples. Reagents The sample weighing and reagent additions were performed in a Class 100 clean room. Vessels were cleaned by heating at 100 "C for 24 h in HN03 or for the carbon determination with a1176 ANALYST, SEPTEMBER 1993, VOL. 118 Table 1 Operating conditions for digestion systems Reagentdm1 System Weight/g HN03 H2S04 Prolabo M 300 l(10ml) 3.0 1.0 or or 3.75 0.75 Floyd RMS 150 0.3 (10 ml) 5.0 HPA 0.2 (5 ml) 3.0 H202 HCl EDTA Programme 8.0 10% power/4 min 40% power/4 min or until charring occurs 40% power/l6 min 10% power/4 min 3.0 10% power/4 min 3.0 10% power/4 min 20 psi/;! min 40 psiR min 60 psi/:! min 80 psi/2 min 100 psi/2 min 120 psU2 min 140 psi/2 min 160 psi/l5 rnin 40 "C/20 mid110 "C 110 "C/30 min/l 10 "C 110 "C/90 min/280 "C 280 "C/90 mid280 "C Table 2 Instrumental parameters for ICP-AES (Plasma 11) R.f ./MHz Incident powerikW Argon gas flow rated min-l Outer Intermediate Nebulizer Observation height/mm Read delayis Sampling time/ms Analytical wavelengthshm Reference wavelengthhm 27.12 1.0 15 1.0 l.OforC, KandNa 0.7 for others 9 for C 15 for others 60 for C 20 for others 100 As 1193.759 Ba I1 455.403 C I 193.091 Ca I1 393.366 Cd 11 214.438 Cu 1324.754 Fe I1 238.204 Mg I1 280.270 Mn I1 257.610 Pb I1 220.353 Sr I1 421.552 Zn 1213.856 Sc I1 424.683 mixture of concentrated HN03 and HZSO4 (1 + 1).They were rinsed with distilled and de-ionized water (18 MS2 cm-2 resistivity, Barnstead NANOpure, SybrodBarnstead, Bos- ton, MA, USA). Reagents were of high purity (H2SO4 and H202) or pre-purified by sub-boiling (HNO,) , isothermal distillation (aqueous ammonia) or recrystallization in 2% HN03 (EDTA). For the carbon determination, analytical reagent- grade chemicals were used.15 A 10% NaEDTA solution was prepared by dissolving 10 g of the reagent in 100 ml of de-ionized water. A 10% NbEDTA solution was prepared by dissolving 10 g of EDTA in a small volume of aqueous ammonia and diluting to 100 ml with de-ionized water.A 1000 pg ml-1 Sc stock solution was prepared by dissolving 0.7889 g of Sc2O3 (99.98%, Johnson Matthey, Ward Hill, MA, USA) in 10 ml of HN03 (1 + 1) and diluting, with de-ionized water, to 500 ml in a calibrated flask. Standard solutions for ICP-AES measurements were prepared in 10% v/v HN03. The following test standard solutions (v/v) were prepared to contain 50 pg ml-1 Sc: 5% HN03, 15% HN03, 12% HC1-3% HN03, 15% HCl and 1.2% m/v NH4EDTA-4% H2SO4 neutralized with aqueous ammonia. Carbon standards were prepared from mannitol or urea.15 A 1000 pg ml-1 As stock standard solution was prepared by dissolving 0.1428 g of C2H6AsNa02.3H20 in 50 ml of de-ionized water. An artificial urine sample was prepared by dissolving 2.083 g of urea, 0.125 g of creatinine, 1.41 g of NaCl, 0.28 g of KCl, 0.0794 g of CaC12.2H20, 0.088 g of MgS04-7H20 and 0.19 ml of aqueous ammonia in de-ionized water in a 100 ml calibrated flask.Samples The reference materials (RMs) investigated were National Institute of Standards and Technology (NIST) Standard Reference Material (SRH) 1577b Bovine Liver, NIST SRM 1566 Oyster Tissue, International Atomic Energy Agency (IAEA) H8 Horse Kidney, National Institute of Environmen- tal Studies (Japan) (NIES) No. 6 Mussel, NIST SRM 1549 Non-Fat Milk Powder and NIST SRM 2670 Toxic Metals in Freeze-Dried Urine (elevated level). Samples included pack- aged infant formula milk, total parenteral nutrition (TPN), a 24 h urine sample and artificial urine. Liquid samples (milk and TPN) were freeze-dried and used for the determination of carbon in the original sample as well as for sample-digestion studies.Aliquots of the urine sample were either evaporated to dryness in the digestion vessel used (HPA, Floyd) before the dissolution, or during the procedure (Prolabo) . Sample Preparation Open-focused microwave digestion (Prolabo M 300) A 1 g (dry) or 10 g (liquid) sample was weighed into a 50 ml Prolabo flat-bottomed glass vessel, then 5 ml of HN03, 1 ml of H2S04 and 5 ml of 250 mg 1-1 Sc solution were added. A Vigreux reflux column was attached. Power-time pro- grammes (Table 1) were applied until all of the HN03 was completely removed and the boiling-point of H2SO4 was reached. This corresponds usually to carbon black colour formation (i.e., charring).Thereafter, the heating was stopped after 1 min, approximately 0.5 ml of 30% H202 was added dropwise, and the solution was again heated (at 40% power) for 1 min. This procedure was repeated for 16 rnin (i.e., 8 ml of H202). The vessel was washed through the condenser with de-ionized water, and different reagents were added: 3.75 ml of HN03 or 3 ml of HCI plus 0.75 ml of HN03,ANALYST, SEPTEMBER 1993, VOL. 118 1177 or aqueous ammonia (until precipitation occurred at pH 9) and then 3 ml of 10% NH&DTA. The solution was diluted with de-ionized water to approximately 15 ml and again heated (at 10% power) for 4 min. After cooling, the solution was transferred to a Nalgene tube and diluted to 25 ml with de-ionized water. When the EDTA solution was added, the pH was checked again and adjusted if necessary with drops of aqueous ammonia.Medium-pressure microwave dissolution (Floyd RMS 150) A dry sample (0.3 g) or liquid urine (10 g) was weighed into an 85 ml PFA double-walled vessel (the latter sample was dried in the vessel), 5 ml of HN03 was added, and the vessel was closed. The sample was heated in the pressure-control mode (Table 1). After cooling, the vessel was opened, and 5 ml of 250 pg ml-1 Sc were added. The solution was transferred to a Nalgene tube and diluted to 25 ml with de-ionized water. High-pressure as hing A dry sample (0.2 g) or liquid urine (5 g) was weighed into a 70 ml quartz HPA vessel (the latter sample was dried in the vessel), and 3 ml of HN03 was added. The vessel was covered, and the sample was heated according to the programme presented in Table 1.After cooling, 2 ml of 250 pg ml-l Sc was added. The sample was transferred to a Nalgene tube and diluted to 10 ml with de-ionized water. Procedures ICP-A ES determination Arsenic, Ba, Ca, Cd, Cu, Fe, K, Mg, Mn, Na, P, Pb, Sr and Zn were determined by ICP-AES in the solutions obtained after sample decomposition against standard solutions in 10% HN03. Aliquots were taken for carbon determination and treated according to the method described previously.15 Sample temperature The importance of temperature control and internal pressure has been demonstrated for closed-vessel microwave systems for various samples and reagents.2 In the focused microwave system the microwave energy is directed only at the part of the digestion vessel containing the sample.Therefore, tempera- ture correlations found for the closed-vessel systems are not expected to be valid for the open-vessel microwave system. In an attempt to characterize the heating mechanism in the open-focused microwave system, solution temperatures were measured for different volumes of de-ionized water and by applying 15% power for 5 min and 30% power for 2 min. 3.5 cm I ( 1 3 I \ 13.5 cm -1 3.5 cm 3.5 cm - , 3.5 cm , Fig. 1 Digestion containers used for temperature measurements. Prolabo vessel with round bottom; glass cylinder; Prolabo vessel with flat bottom Three types of Prolabo glass vessel were compared (Fig. 1): 50 ml volume with a flat bottom, 50 ml with a round shape, and a cylindrical vessel with the same height (approximately 175 mm) and bottom diameter as the flat one.The Vigreux condenser was replaced by a Teflon cap to allow the rapid immersion of a thermometer. A 0-100 "C thermometer with 0.1 "C increments was immersed in the solution immediately after the microwave heating was stopped. The temperature (7'2) was measured after mixing the water well. The AT = T2 - TI, where T1 is the temperature of the water before heating, was calculated. Insoluble sulfate formation To characterize the formation of insoluble sulfates, test solutions were studied. Five ml of 250 pg ml-1 Sc, various amounts of Ba (2.5-250 pg), and 'spikes' of 25 or 125 pg of Pb, 10 or 25 yg of Sr and 25 or 2000 pg of Ca were placed in 50 ml glass beakers or Prolabo vessels. Approximately 5 ml of HN03 and 1 ml of H2S04 were added, and the solutions were evaporated on a hot-plate or in the open-focused microwave system until SO3 fumes evolved.After cooling, the solutions were diluted to 25 ml with: (a) de-ionized water; ( 6 ) 5, 10 or 15% v/v HN03; or (c) 11.25% HC1-3.75% v/v HN03, 11.6% HC1-2.8% HN03, 12% HC1-3% HN03 or 15% HCl, and heated. When EDTA salts were used, each sample (after cooling) was diluted with a small volume of de-ionized water, and aqueous ammonia was added until a precipitate appeared (pH 9), followed by 3 ml of 10% NaEDTA or 10% NH4EDTA. The solution was heated then diluted with de-ionized water to 25 ml. The pH was checked, and, if necessary, drops of aqueous ammonia were added to bring the pH to 9. Experiments with NaEDTA at pM 2 were also performed (at this pH, NH4EDTA precipitates). The concen- trations of Ba, Ca, Pb and Sr were measured by ICP-AES. Sample analysis Biological samples alone or spiked with Ba, Pb or Sr were analysed.The RCC and the calculated residual carbon (RC), which represent the effectiveness of the sample decomposi- tion,9 and element recovery were measured for the three digestion techniques. Results and Discussion ICP-AES Determination With the Myers-Tracy signal compensation technique for the ICP-AES determination,l4 10% v/v HN03 standard solutions can be used for all the procedures with test solutions containing 5% v/v HN03, 15% HN03, 12% HC1-3% HN03, or 1.2% m/v NH4EDTA4% v/v H2SO4 neutralized by aqueous ammonia. Sample Temperature Temperature changes for various water volumes heated in three microwave vessels are compared in Fig.2. The results are mean values from six replicates taken on different days. The distribution of the results was relatively low, especially for volumes above 2 ml (i.e., the relative standard deviation was 10% for 2 ml and 2% for > 2 ml). With a small volume of liquid (under 8 ml), the round-bottomed vessel provided the highest temperature rise. In round-bottomed vessels the same amount of liquid occupied a height in the focused microwave cavity greater than in the two other vessels. As a result, the sample probably absorbed more energy. When the height in the vessel, instead of a fixed volume occupied, was examined (Fig. 3), the difference between the vessels could be discoun- ted, especially for heights under 1.5 cm.For greater heights the vessel shape also influenced the temperature rise. TheseANALYST, SEPTEMBER 1993, VOL. 118 1178 80 60 u P a a 20 0 10 20 Volu me/m I 30 Fig. 2 Temperature change when heating different volumes of water in three type of vessels. A, Prolabo vessel with flat bottom; B, Prolabo vessel with round bottom; and C, glass cylinder 80 60 0 Q 40 20 0 A 1 2 3 4 5 6 7 Heig ht/cm Fig. 3 Temperature change as a function of liquid height in the vessel. A, Prolabo vessel with flat bottom; B, Prolabo vessel with round bottom; and C, glass cylinder 100 80 20 1 I ,A 0 1 2 3 4 5 Concentration of Ba/mg 1-1 Fig. 4 Recovery of A, Ba, B, Pb, and C, Sr as a function of Ba concentration in the presence of 4% H2SO4 results clarified our observation that, for round-bottomed vessels filled above a minimum volume, 30% power has to be used in the second heating stage instead of 40% power to avoid spilling from overheating.Also, at the same microwave power, usually a shorter time was necessary to evaporate the same amount of HN03 from the round-bottomed vessel. 100 80 I s $ 60 - > s 40 20 0 1 2 3 4 5 Concentration of Ba/mg I-’ Fig. 5 Recovery of Ba, Pb, and Sr as a function of Ba concentration in the presence of 4% H2SO4 + 5% HN03: A, Ba; B, Pb; and C, Sr; and 4% H2SO4 + 15% HN03: D, Ba; E, Pb; and F, Sr 100 80 s I $ 6o > 8 2 40 20 B \ F 0 1 2 3 4 5 Concentration of Ba/mg 1-1 Fig. 6 Recovery of Ba, Pb, and Sr as a function of Ba concentration in the presence of 4% H2SO4 + 12% HCl + 3% HN03: A, Ba; B, Pb, and C, Sr: and 4% H2S04 + 15% HCI: D, Ba; E, Pb; and F, Sr 100 80 &? 1 $ 60 s a 40 20 0 1 2 3 4 5 Concentration of Ba/mg 1-1 Fig. 7 Recovery of Ba, Pb, and Sr as a function of Ba concentration in the presence of 4% NH4S04 + 1.2% NaEDTA or NH4EDTA salts pH = 9): I, Ba; 0, Pb; and e, Sr; and 4% H2S04 + 1.2% NaEDTA pH = 2): 0, Ba; A , Pb; and A, Sr These results can be used as guidelines for specific analytical problems.In the subsequent investigations the flat-bottomed vessels were employed, because they are somewhat easier to handle and are commonly used.ANALYST, SEPTEMBER 1993, VOL. 118 1179 Table 3 Recovery of Ba, Pb and Sr from model solutions containing 10 pg ml-1 Ba and different reagents (n = 8) Recovery (%) Reagent 4% H2S04 4% H2S04-10% HN03 4% H2S04-11.25% HC1-3.75 HN03 4% H2S04-ll.6% HC1-2.8% HN03 4% H2S04-12% HC1-3% HN03 4% NH4S04-1 .2% NaEDTA 4% H2SO4-5% HN03 4% H2S04-15% HN03 4% H2S04-15% HCI 4% NH4S04-1.2% NH4EDTA Ba 0.3 f 0.1 0.7 k 0.1 2.0 f 0.8 2.4 f 1.2 0.8 f 0.4 0.8 k 0.4 0.8 f 0.4 0.5 rt 0.3 100 k 2.8 100 f 2.5 Pb 12.2 f 1.1 17.7 f 1.0 30.0 f 3.3 52.5 f 4.0 83.2 k 8.0 85.5 f 5.4 85.8 f 4.2 84.8 k 3.7 98.5 f 1.5 98.7 k 1.6 Sr 72.8 f 8.5 75.0 k 5.2 83.0 f 4.4 86.8 f 4.5 81.6 f 6.2 78.5 rt 4.0 79.5 f 5.3 76.7 f 5.3 100 k 3.0 100 f 2.0 Table 4 Analysis of standard reference materials using open-focused microwave system and ICP-AES.Values (n = 6) are given in pg g-1 except where percentages are shown. Non-certified values are given in parentheses Element Ca Cd c u Fe Mg Mn P (Yo) K(%) Na (%) Sr Zn NIST SRM 1577b Bovine Liver IAEA H8 Horse Kidney Element Found Reference Found Reference Ca 122 f 6 116 f 4 944 f 56 924 f 77 Cd 0.51 f 0.04 0.50 k 0.03 208 rt 17 189 f 4.5 c u 162 f 4 1 6 0 f 8 32.8 f 1.3 31.3 f 1.75 265 k 15 Fe 188 f 8 184 f 15 Mg 610 k 22 601 k 28 764 k 54 818 rt 75 Mn 10.8 f 0.8 10.5 f 1.7 5.99 f 0.14 5.73 f 0.275 P (%) 1.07 f 0.05 1.10 f 0.03 1.06 f 0.05 1.12 f 0.06 K (%) 1.04 f 0.008 0.994 f 0.002 1.22 f 0.02 1.17 f 0.075 Na (%) 0.218 f 0.015 0.242 f 0.006 0.91 f 0.06 0.96 f 0.03 Zn 124 f 7 127 rt 16 192 f 7 260 f 10 Sr 0.139 f 0.003 0.136 f 0.001 1.1 f 0.1 (1.1) 193 k 6 NIES No.6 Mussel NIST SRM 1566 Oyster Tissue NIST SRM 1549 Non-Fat Milk Powder Found Reference Found Reference Found Reference 0.135 rt 0.005% 0.13 k 0.01% 0.148 5~ 0.008% 0.15 rt 0.02% 1.30 rt 0.06% 1.30 k 0.05% 4 .0 0.82 f 0.03 3.7 f 0.5 3.5 f 0.4 4 . 0 0.0005 f 0.0002 5.1 f 0.6 4.9 f 0.3 63.7 f 2.1 63.0 f 3.5 0.7 f 0.2 0.7 f 0.1 0.198 f 0.006% 0.21 f 0.01% 0.126 f 0.002% 0.128 k 0.009% 0.117 f 0.03% 0.120 f 0.03% 15.6 f 0.9 16.3 f 1.2 17.5 f 0.5 17.5f 1.2 0.30 f 0.07 0.26 f 0.06 0.74 f 0.03 (0.77) 0.72 f 0.02 (0.81) 1.00 f 0.03 (1.05) 0.57 f 0.01 0.54 f 0.02 0.967 f 0.01 0.969 f 0.005 1.80 f 0.1 1.69 f 0.03 0.94 f 0.04 1.00 f 0.03 0.52 k 0.04 0.51 f 0.03 0.494 f 0.015 0.497 f 0.010 10.20 f 0.43 10.36 k 0.56 2.6 k 0.2 Not specified 846 f 31 852 f 14 46.4 f 0.8 46.1 f 2.2 152 rt 5 158 rt 8 194 f 6 195 k 34 2.7 rt 0.6 (2.1) 17 f 0.5 (17) 106 f 10 106 k 6 Table 5 Analysis of NIST SRM 2670 Toxic Metals in Freeze-Dried Urine (elevated level) with use of open-focused microwave digestion (n = 2) Element Cdmg ml-1 Cdpg ml-1 Cdpg ml-1 Wmg ml-1 Mg/mg ml-1 Mdpg ml- Ndmg ml- Found 0.109 f 0.01 0.075 k 0.006 0.35 4 0.02 1.4 f 0.1 0.060 f 0.002 0.320 k 0.01 2.50 k 0.20 Reference 0.105 f 0.005 0.088 k 0.003 0.37 rt 0.03 0.063 k 0.003 (0.33) 2.62 f 0.14 (1.5) Elements Forming Insoluble Sulfates Sulfuric acid is effective in decomposing organic matter, and the resulting solutions contain little residual carbon.9 Low residual carbon is important for some techniques.However, H2SO4 can restrict the determination of elements forming low-solubility sulfates. When about 3 pmol 1-1 of Ba was present in the digest (approximately 0.44% v/v H2S04), coprecipitation of Pb occurred.10 The ionic strength and formation of soluble complexes can strongly influence the Occurrence of a precipitate and coprecipitate. In biological samples, Pb as well as Ca and Sr have to be determined. A comparison of the ionic radii and the solubility product constants shows that coprecipitation with BaS04 can be expected for Pb and Sr.16 In the determination of Ba in waters containing sulfates, the addition of NaEDTA provided accurate results; NaEDTA was also added in the determina- tion of Pb in PVC11J2 by atomic absorption spectrometry, and 4% v/v HN03 was added to the final H2S04 digest for square-wave voltammetry.12 The increase of the lead chloride solubility in an excess of HCl and the use of an HCl and HN03 mixture for the analysis of a pure lead sample was reported.17 The 11.6% HC1-2.8% HN03 mixture was found to be the most appropriate for element compatibility and long-term stability.'* These investigations have been performed with conductive heating.Microwave energy interacts on a molecular level, and some differences in precipitation and occlusion processes can be expected. Therefore, test solutions were investigated with use of both types of heating. The results obtained were statistically indistinguishable, which indicates that ionic strength and complex formation are the major influencing factors. In all experiments with test solutions, the recovery of Ca was quantitative (95-102%). Some of the mean value results (n = 8) for both types of heating for the other elements are presented in Figs. 4-7. With only 4% H2S04 in the final solutions (Fig.4), the loss of Pb starts with the precipitation of1180 ANALYST, SEPTEMBER 1993, VOL. 118 Table 6 Analysis of formula milk and TPN (n = 6) with use of three digestion procedures Formula milk TPN HPA 3200 f 100 7.4 f 0.4 77.9 5 0.6 0.12 f 0.2 280 k 10 ~ 0 . 5 0 0.54 f 0.07 0.29 f 0.01 46.3 f 2.2 Floyd 3100 f 100 7.0 f 0.3 78.2 f 0.7 0.13 f 0.2 280 f 5 0.58 f 0.05 0.28 f 0.02 41.9 f 2.2 <0.80 Prolabo 3200 f 100 6.9 f 0.3 76.1 f 0.6 0.13 f 0.2 300f 10 0.72 f 0.02 0.54 f 0.06 0.29 0.01 41.3 f 2.0 HPA 67.5 k 3 <0.50 <0.50 0.040 k 0.003 32.7 f 1.8 c0.50 0.065 f 0.05 0.023 k 0.003 1.19 k 0.1 Floyd 68.4 f 4.5 <0.80 <0.80 0.045 f 0.004 32.6 f 2.0 (0.80 0.061 f 0.04 0.025 f 0.004 1.20 k 0.07 Prolabo <0.25 67.4 f 4 <0.25 0.043 k 0.004 32.3 f 2.0 0.060 f 0.05 0.024 f 0.003 1.17 k 0.05 ~ 0 .2 5 Table 7 Recovery (%) of As, Ba, Pb and Sr from samples, spiked with 125 pg each of As, Ba and Pb (5 pg 1-I), for three digestion proce- dures with open-focused microwave digestion (n = 3) Recovery (YO) Sample Bovine Liver Mussel Oyster Tissue Horse Kidney Non-Fat Formula Milk Urine Sample TPN Milk Powder (24 h) 15% HN03 As Ba Pb 98.3 5~ 1.5 10.4 f 2.5 62.2 f 2.5 96.2 5 1.2 13.0 f 3.1 55.7 f 1.8 100 f 3.0 12.0 f 2.8 67.5 f 2.3 100 f 2.5 14.0 k 3.2 67.2 f 1.8 100 & 1.7 23.0 f 4.0 68.5 f 2.8 99.3 5 1.8 9.4 f 1.3 55.7 f 2.3 98.8 f 2.3 9.4 f 1.8 55.8 k 1.6 99.4 k 2.8 19.4 f 2.8 65.0 f 2.4 Sr As 90.2 f 1.8 98.8 f 1.5 82.4 f 1.6 99.5 f 2.1 89.3 f 1.5 100 f 2.4 86.5 f 1.0 94.7 f 0.8 83.4 f 1.8 96.8 f 1.3 88.8 f 2.0 100 f 1.5 84.2 f 0.9 100 f 2.0 97.2 k 1.6 97.8 f 1.3 12% HC1-3% HN03 Ba Pb Sr As 2.4f0.5 87.5k3.0 84.5f3.3 99.22 1.0 4.0f 1.5 81.6f2.5 78.6k 1.9 100f2.3 3.0f0.6 88.2f 1.0 80.7f2.2 99.3 f 2 . 1 6.8f2.5 82.0f2.6 91.0f 1.6 99.5 f 1.8 8.4f1.1 92.1f2.3 81.6f2.1 98.8f1.4 4.4f1.0 93.2f1.9 92.3f2.0 100f1.6 2.8f0.3 80.8k1.5 93.2f1.6 100f2.3 14.4f3.0 90.551.8 89.0f2.0 100f1.8 1.2% NHdEDTA Ba Pb Sr 98.8 f 0.8 100 f 1.0 99.5 f 0.9 100 f 1.3 96.7 f 1.3 100 f 0.9 100 f 2.3 99.5 f 2.1 99.0 f 2.0 100 k 1.8 101 f 2.6 97.8 f 1.8 100 k 1.2 96.5 k 1.2 99.2 f 1.9 100 k 1.6 100 f 1.9 97.5 f 1.1 96.6 f 0.8 96.8 f 0.8 99.5 f 1.7 98.4 k 1.1 97.8 k 1.1 100 f 2.0 Table 8 Carbon content of original samples (c,,) Carbon (%) Sample Dry.material NIST SRM 1566 Oyster Tissue 46.64 NIES No.6 Mussel 44.24 IAEA H8 Horse Kidney 46.64 NIST SRM 1577b Bovine Liver 48.14 NIST SRM 1549 Non-Fat Milk Powder 41.12 Formula milk (Similac) 50.88 TPN 38.81 Artificial urine - NIST SRM 2670 Toxic Metals in Freeze Dried urine - Urine sample - Liquid material - - 4.02 4.90 0.46 0.56 0.49 Ba above 0.1 pg ml-1 Ba. Strontium is less prone to coprecipitate above 1 pg ml-1 Ba. The presence of 15% HN03 (Fig. 5) reduced the coprecipitation, especially for Sr, but loss of Pb occurred at a Ba concentration above 1 pg ml-1. The presence of HCl improved the recovery of Pb (Fig. 6), but recovery of Sr decreased slightly at high Ba concentrations. Addition of EDTA solutions at pH 9 afforded complete quantitative recovery of all the elements including Ba and Sr (Fig.7). At lower pH, only the Pb complexes were stable. The NbEDTA salt was preferred to the Na salt, because the alkaline-earth elements could be determined therewith and the salt was easily purified. The recoveries of Ba, Pb and Sr, with 10 pg ml-1 Ba present in the final solution, are reported in Table 3. Increasing the acid concentrations improves the recovery; HN03 favours more Ba and Sr, and HCl aids Pb. The EDTA dissolves all the precipitates and permits quantitative determination of all the elements investigated. Sample Analysis Based on investigations with these test solutions, RMs and samples were analysed to verify the efficacy of the open- focused microwave procedure. Less than 0.1 pg ml-1 Ba in the final solution was found in all investigated samples. For the open-focused microwave digestion, results obtained with 15% HN03, 12% HCI-3% HN03 and 1.2% NH4EDTA were statistically indistinguishable, and means are presented for the RMs in Tables 4 and 5.The values found are in good agreement with the certified data. For the analysis of TPN and formula milk samples, the results for nine elements are compared in Table 6 for digestions with the HPA and the two microwave systems. The results agree closely. The sample preparation and accuracy of determination were assessed for authentic samples, for Pb, Sr and Ca, which form low-solubility sulfates, and for As, which is prone to volatilization. Because of the low concentrations of Pb and As in the SRMs and the authentic samples and the poor ICP-AES sensitivity, samples were spiked with 125 pg of Pb and As, the latter element added as the Na salt of cacodilic acid.Spikes of 125 pg of Ba were also added. Element recoveries are summarized in Table 7. The recovery of As is quantitative for all samples. Similar results are obtained for the recovery of Ba, Pb and Sr from these samples and from test solutions with different reagents. These data confirm that, in the presence of a high concentration of Ba, the negative systematic error can be overcome by adding NH4EDTA in the final preparation step. Comparison of Sample-digestion Techniques The RCC in the final solutions was measured for each digestion approach, by use of the operating conditions and reagents cited in Table 1, but with EDTA salts being used in the Prolabo system. To calculate the effectiveness of the procedure, the original carbon content (tor) was measured by elemental analysis of the SRMs and freeze-dried TPN and formula milk.The original carbon content in SRM 2670 and a urine sample after 10-fold dilution was measured by ICP- AES.15 The procedure was verified by comparing results for artificially prepared urine. The original sample carbon con- tents (Table 8) were used to calculate the total carbon left undigested (RC%) = RCC X lOO/cOr.9 The results obtained are presented in Table 9. With both the open digestionANALYST, SEPTEMBER 1993, VOL. 118 1181 Table 9 Residual carbon (RC; for definition see text) in sample digested with use of three systems ( n = 2 to 6) RC (%) Sample NIST SRM 1577b Bovine Liver IAEA H8 Horse Kidney NIES No.6 Mussel NIST SRM 1566 Oyster Tissue NIST SRM 1549 Non-Fat Milk Powder Formula milk TPN Artificial urine with urea Artificial urine with urea and NIST SRM 2670 Toxic Metals in Urine sample creatinine Freeze-Dried Urine Prolabo 0.58 t- 0.4 0.36 k 0.2 0.54 k 0.2 0.38 & 0.1 0.15 k 0.1 0.10 k 0.05 0.15 rt 0.05 <0.2 <0.2 2.70 f 0.3 0.73 k 0.2 HPA 1.0 & 0.2 1.34 ?c 0.5 0.70 k 0.3 0.64 k 0.1 0.51 f 0.1 0.51 -e 0.1 0.36 k 0.1 <0.2 <0.2 2.25 rt 0.8 0.78 & 0.3 Floyd 16.4 zk 4.0 14.2 & 2.0 16.1 t 3.0 14.6 k 2.2 6.8 ?c 1.5 11.1 ?c 1.0 2.1 t 0.4 4 . 2 0.53 rt 0.2 9.7 & 3.5 7.4 f 2.0 microwave system and the HPA, the effectiveness of organic matter decomposition is similar and superior to that of the closed microwave digestion procedure.Unexpectedly high RC results were obtained for urine. Because of the low organic carbon content in the original samples (Table 8), complete decomposition of carbon was expected. This was observed for an artificial urine sample to which only urea was added (Table 9). For some organic compounds found in urine (Le., urea, uric acid, creatinine and benzoic acid),19 a three times higher RC was observed for creatinine, after digestion of each pure compound (0.1-0.5 g) in the open-focused system, than for the other compounds. With the addition of creatinine to the artificial urine at a concentration based on literature values,19 a detectable difference above the detection limit for carbon15 was observed only for the closed microwave system (Table 9). Furthermore, incomplete decomposition of carbon was observed for actual urine samples.The SRM 2670, based on the North American diet, yielded RC values higher than those for the local 24 h urine sample. Perhaps this disparity could be explained by the presence of very stable organometallic compounds, the concentrations of which can vary with the sample.20 However, additional studies are required. Conclusion The investigation with the open-focused microwave system has produced an optimized method for the dissolution of biological samples. A minimal volume of H2SO4 (i.e., 1 ml) was used to avoid nebulizer transport effects in ICP-AES, but it was sufficient to decompose the organic matter. With addition of 15% HN03 or 12% HC1-3% HN03, the accuracy in the determination of Pb and Sr can be improved.Addition of 1.2% NH4EDTA in the final digestion step improves the recovery of Ba. In authentic biological samples, the spiked Ba recovery was quantitative. The RC obtained with the open- focused microwave system was lower than that obtained with the closed medium-pressure microwave system, and was similar to that with the HPA. The established method can also be used with automated models of the open-focused microwave system, as well as for measuring techniques other than ICP-AES where complete decomposition of carbon is necessary. The results obtained are similar to those obtained recently for milk samples.9 We thank Prolabo, Floyd and Questron for their support and assistance in using their equipment. Research was sponsored by the ICP Information Newsletter. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 References Schramel, P., Hasse, S . , and Knapp, G., Fresenius 2. Anal. Chem., 1987,326, 142. Kingston, H. M., and Jassie, L. B. (eds.), Introduction to Microwave Sample Preparation, American Chemical Society, Washington, DC, 1988, pp. 25, 26, 76, 165. Grillo, A. C., Spectroscopy, 1989, 4, 16. Kingston, H. M., and Jassie, L. B., Anal. Chem., 1986, 58, 2534. Salt, R. N., and Miller, R. O., Anal. Chem., 1992, 64, 230. Stripp, R. A., and Bogen, D. C., J. Anal. Toxicol., 1989,13,57. Hocquellet, P., and Candillier, M.-P., Analyst, 1991, 116, 505. Feinberg, M. H., Analusis. 1901, 19, 47. Krushevska, A., Barnes, R. M., Amarasiriwaradena, C. J., Foner, H., and Martines, L., J . Anal. At. Spectrom., 1992, 7, 851. Temminghoff, E. J. M., and Novozamsky, I., Analyst, 1992, 117, 23. Belarra, M. A., Gallarta, F., Anzano, J. M., and Castillo, J. R.. J. Anal. At. Spectrom., 1986, 1 , 141. Peddy, R. V. C., Kalpana, G., and Koshy, V. J . , Analyst, 1992, 117, 27. Karabash, A. G., Bondarenko, L. S., Morozova, G. G., and Peizulaev, Sh. I., Zh. Anal. Khim., 1960, 15, 623. Meyers, S. A., and Tracy, D. H., Spectrochim. Ada, Part B, 1983, 38, 1227. Krushevska, A., Barnes, R. M., Amarasiriwaradena, C. J., Foner, H., and Martines, L., J . Anal. At. Spectrom., 1992, 7, 845. Weast, R. C. Astie, M. J., and Beyer, W. H. (eds.), Handbook of Chemistry and Physics, CRC Press, Boca Raton, FL, 65th edn., 1984, F-165 and B-222. Degtyareva, 0. F., and Ostrovskaya, M. F., Zh. Anal. Khirn., 1960, 20, 814. Que Hee, S. S . , Macdonald, T. J., and Boyle, J . R., Anal. Chem., 1985, 57, 1242. Lehninger, A. L., in Principles of Biochemistry, ed. Anderson, S., and Fox, J., Wort, New York, NY, 1982, p. 703. Hanna, C. P., Tyson, J . F., and McIntosh, S., Clin. Chem. (Winston-Salem, N . C.), in the press. Paper 21066675 Received December 16, 1992 Accepted March 4, 1993

ISSN:0003-2654    

DOI:10.1039/AN9931801175

出版商:RSC    

年代:1993

数据来源: RSC

摘要:

ANALYST, SEPTEMBER 1993, VOL. 118 1183 Determination of Trace Amounts of Thallium and Tellurium in Nickel-base Alloys by Electrothermal Atomic Absorption Spectrometry Suh-Jen Jane Tsai and Ching-Ching Jan Department of Applied Chemistry, Providence University, Taichung Hsien, Taiwan A method is described for determining trace amounts of TI in nickel-base alloys using pre-treatment with ammonia solution and electrothermal atomic absorption spectrometry. Thallium was coprecipitated when the sample solution of a nickel-base alloy was treated with ammonia solution. Background absorption was effectively eliminated in the new matrix. Nickel worked effectively as a chemical modifier for Te, and raised the charring temperature from 500 t o 1000 "C. Therefore, it was possible t o determine sub-nanogram levels of Te without any complicated pre-treatment. The accuracy and precision of the proposed method were elucidated through the analysis of two nickel-base alloys: spectroscopic standard certified reference material 346A IN100 Alloy and Standard Reference Material 899 Tracealloy C.There was good agreement between the expected values and the results obtained. For TI, the results found for IN100 and Tracealloy C were 1.94 k 0.09 pg 9-1 (s, = 5%) and 0.251 k 0.005 pg 9-1 (s, = 2%), respectively, the reference value for IN100 being 2 and certified value for Tracealloy C being 0.252 k 0.003 pg 9-1. The recoveries for these alloys were 100 k 4 and 102 k 3%, respectively. The detection limit was 15 pg g-1. For Te, the certified values for IN100 and Tracealloy C were 9 k 1 and 5.9 -+ 0.6 pg g-1, respectively.The Te contents determined by the proposed procedure were 9.03 k 0.09 pg 9-1 (s, = 1%) and 5.93 k 0.26 pg g-1 (s, = 4%), with recoveries of 98 k 4 and 97 k 4%, respectively. The detection limit was 35 pg g-1. Although the addition of Pd modifier gave a better detection limit (18 pg g-I), it led t o poorer results in terms of accuracy, precision and recovery. Keywords: Thallium and tellurium determination; nickel-base alloy; ammonia pre-treatment; nickel-base and palladium modifiers; electrothermal atomic absorption spectrometry The quality of nickel-base alloys is highly dependent on the amounts of trace elements present because the existence of trace elements has serious effects on the mechanical and physical properties of these high-temperature alloys.1-5. The determination of trace elements in nickel-base alloys had been a challenge owing to the complexity of the alloy matrices and therefore only a few studies on this subject have been reported .(+s As TI is a toxic element, the determination of trace concentrations of this element in environmental and biological samples is important. Electrothermal atomic absorption spectrometry (ETAAS) and inverse voltammetry are appro- priate methods for the determination of trace amounts of T1 with detection limits at the ng ml-1 level, as reported in a review published in 1988.9 Anodic stripping voltammetry had also been employed in the determination of trace amounts of TI. 1 0 The determination of TI at the pg ml-1 level or higher by flame AAS was usually accomplished by extracting the T1 complexes with an organic solvent such as diisopropyl ether.11 Both complexation and extraction processes had to proceed to completion in order to obtain reliable results.Trace amounts of were determined together with Gal1' and In"' by successive titrations and the detection limit thus obtained was 511 ppm.12 Trace amounts of TI were measured using radiochemical neutron activation analysis13 and isotope dilu- tion mass spectrometry14 in addition to ETAAS.15 Welcher et al.6 showed the potential application of non- flame AAS methods in determining ppm concentrations of TI in high-temperature alloys. There was good agreement between the results obtained using graphite furnace atomiza- tion and by emission spectrometry. However, the precision (s, = 11% for 3.4 ppm) could be improved.The matrix material of the nickel alloy caused about a 20% decrease in sensitivity. The very large background signals obtained need to be reduced before a satisfactory detection limit can be attained. A spiking technique was recommended by Dulski and Bixler8 in order to compensate for the serious matrix effects of nickel-base alloys. Difficulties were encountered in the determination of trace amounts of TI by ETAAS. Among them, serious signal depression caused by either HCF or HC1O41h both of which received much attention. The matrix interferences caused by HCl were eliminated by H2S0417 and those caused by HC104 could be reduced by using pyrolytic graphite coated graphite tubes.18 Palladium has also been confirmed as an effective chemical modifier for T1.19,20 Slavin and Manning21 confirmed that less signal depression was observed when the atomization was performed on a L'vov platform.Coprecipitation has sometimes been employed for the separation of the analyte from interfering ions and also for preconcentration. Hafnium hydroxide is an effective collector for trace amounts of Be.22 Niskavaara et al.23 proposed a method for the separation of trace amounts of Ag, Au, Pd, Pt, Rh, Se and Te from relatively complex geological samples; reductive coprecipitation was achieved with tin( 11) chloride as the reductant and mercury as the collector. The preconcentra- tion of 16 elements, including TI, from deep ocean sea-water and coastal water has also been achieved by tetrahydroborate reductive precipitation.24 Welcher et aZ.6 also determined ppm concentrations of Te in high-temperature alloys.As certified reference materials (CRMs) of nickel-base alloys were not available at that time, the accuracy of the ETAAS procedure was evaluated based on the addition of 5 ppm of Te to various nickel-base alloys. Headridge and Nicholson3 tried to analyse a nickel-base alloy via the analysis of a solid sample by ETAAS. The AAS signals of Te were monitored at 214.3 nm and the charring and the atomization temperatures were 1000 and 2900 "C, respec- tively. The detection limit thus obtained was 0.003 pg g-1. Tellurium was reduced to the elemental state with hypophos- phorous acid and collected by As before determination by ETAAS.25 A prior extraction step with isobutyl methyl ketone (IBMK) was employed to remove trace amounts of Te from the complex matrix when geochemical samples were analysed by ETAAS .26,27 Ion-exchange processes were required to remove interfering elements from the analyte in the determi- nation of Te in atmospheric aerosol samples by ETAAS.28,29 An increase in the Te signal by some mineral acids, including HCI, HN03 and H2S04, was reported by Kunselman and Huff,30 who consequently used an analyte addition technique.1184 ANALYST, SEPTEMBER 1993, VOL.118 Various metals, such as Ni, Pd, Mg and Cu31932 and Ir,24 have often been employed as effective chemical modifiers. Success- ful measurements of trace amounts of Te in geological samples or natural water were made using a Pd-coated graphite tube.33 Tellurium in copper-base alloys (30.0002%) was deter- mined by flame AAS with an air-acetylene flame following an extractive pre-treatment with trioctylphosphine oxide- IBMK.34 In Hubert and Chao’s work,35 IBMK was used instead. An extractive pre-treatment was required in deter- mining trace amounts of Te by flame AAS in order to obtain better sensitivity.A hydride generation system coupled with flame AAS was employed to determine Te in a lead alloy. However, a tedious pre-treatment was required before the metal reduction could be effected.25 The connection of a hydride generation system with ETAAS led to the satisfactory determination of Te in a sample of silicate ore? As serious interferences from nickel were reported when trace amounts of Te were determined by hydride generation AAS,37 an effective pre-treatment would be required if complex nickel- base alloys are to be analysed by hydride generation ETAAS.Other instrumental methods such as X-ray fluorescence spectrometry38 and differential-pulse polarography39.40 have also been employed in the determination of trace amounts of Te . The objective of this work was to study the determination of T1 and Te in CRMs of nickel-base alloys, namely Spectro- scopic Standard (SS) CRM 346A IN 100 Alloy and National Institute of Standards and Technology (NIST) Standard Reference Material (SRM) 899 Tracealloy C, by ETAAS. For TI determination, the effects of chemical modifiers including H2S04, Pd and coprecipitation in ammonia solution on the reduction of background signals were evaluated. Pre-treat- ment with ammonia solution was the most effective method.This work also demonstrates the merits of a stabilized temperature platform furnace in the determination of trace amounts of Te. In spite of the complexity of alloy matrices, it was found that trace amounts of Te in nickel-base alloys can be successfully determined by measurement against nickel-con- taining aqueous standard solutions. Experimental Apparatus A Perkin-Elmer Model 1100B atomic absorption spec- trometer fitted with an HGA-700 furnace and AS-70 auto- sampler was used. Integrated absorbance (peak area) values were used in measurements. Experimental parameters and peak profiles were recorded with an Epson EX-800 printer.An electrodeless discharge lamp (EDL) equipped with a Perkin-Elmer EDL power supply was used for Tl determina- tions. The EDL was also used for Te determinations. New pyrolytic graphite coated graphite tubes (Part No. B010-9322) and pyrolytic graphite platforms (Part No. B010-9324) were used. Maximum power heating was used for the atomization step. A Barnstead Nanopure I1 system was employed for water purification. Reagents and Standards Standard solutions of TI (TIN03 in 10% HN03, TI = 1001 pg ml-l), Te (TeCI4 in 30% HC1, Te = 1002 pg rnl-l), Pb [Pb(N03)2 in 2% HN03, Pb = 1001 pg ml-11, Se (H2Se03 in 2% HN03, Se = 998 pg ml-l), Ga [Ga(N03)3 in 2% HN03, Ga = 1000 pg ml-I] and In [In(N03)3 in 2% HN03, In = 1000 pg ml-11 were products of Inorganic Ventures.Hafnium (atomic absorption standard solution, in 10% HCI, Hf = 985 pg ml-I), Zr (atomic absorption standard solution, in 5% HCI, Zr = 1030 pg ml-1) and atomic absorption standard solutions of V, Mo, Nb and Ta were products of Aldrich. Standard solutions of Ni [Ni(N03)3-6H20 in HN03, 0.5 mol 1-11, Co [CO(NO~)~-GH~O in HN03, 0.5 mol 1-11, Ti (TiC14 in HCI, 0.5 mol I-l), Fe [Fe(N03)3.9H20 in HN03, 0.5 mol I-I], Cu [Cu(N03)2-3H20 in HN03, 0.5 mol 1-11, Ti (TiCI4 in HCI, Pd [Pd(N03)2 in 15% HN03, Pd = 10.0 k 0.2 g 1-I), K2Cr207, CrC13.6H20, aluminium foil and concen- trated ammonia solution were of analytical-reagent grade from Merck. Concentrated HCI, HF and HN03 were singly distilled acids from Eastar Chemical. Working standard solutions were prepared from 1000 mg 1 - 1 stock standard solutions by serial dilution.Distilled, de-ionized water (DDW) prepared with a Barnstead Nanopure I1 system was used for all of the determinations. The CRMs of nickel-base alloys, SSCRM 346A IN100 and SRM 899 Tracealloy C were obtained from the Bureau of Analysed Samples and NIST, respectively. Purified grade argon (99.99%) was used as the purge gas during sample analysis. Sample Decomposition Precise amounts of nickel-base alloys (0.04-0.12 g) were digested with 1.5 ml of HCI-HN03 (4 + 1 v/v). After digestion, the sample solutions were heated nearly to dryness and the residue was dissolved in a few millilitres of DDW. The heating and dissolution procedures were repeated in order to remove the excess amount of acids. For TI determinations, the resulting solutions were treated with ammonia solution.For Te determinations, the alloy solutions were diluted to 10 ml with 0.2% HN03. Thallium Determination The above alloy solutions were further treated with ammonia solution as follows before being subjected to ETAAS. Each sample solution prepared above was treated with 5 ml of concentrated ammonia solution and the solution together with the precipitate were transferred into a 10 ml test-tube, which was then filled with concentrated ammonia solution. After the solution had been centrifuged, the precipitate was decom- posed completely with 3-5 ml of HN03 (1 + 1 v/v). The volume of the sample solution was reduced to about 1 ml by heating gently. The above solution was diluted to 5.0 or 10.0 ml precisely with 0.2% HN03.These solutions were analysed by injecting 5-50 pl portions onto the platform located in a pyrolytic graphite coated graphite tube. The corresponding blank solutions were also analysed. In the measurements with a chemical modifier, 1% H2S04 or Pd (1000 pg ml-1 Pd), sample solution was followed by 5 pl of modifier solution. Aliquots of TI solution (0.02 pg ml-I), e.g., 3,5,10,15,20 and 25 pl, were injected sequentially into the graphite tube with the AS-70 autosampler to obtain the calibration graphs. Dilute HN03 (0.2%) was used for preparing the working solutions. Tellurium Determination The sample solutions prepared as described under Sample Decomposition were analysed directly by ETAAS. In the measurement with a Pd modifier, a volume of 5 pl of Pd modifier solution (1000 pg ml-1 Pd) was injected into the graphite tube into which sample solution had been injected. Thus, sample solutions were analysed together with the Pd modifier.The calibration graphs were determined by injecting several aliquots of Te solution (0.05 pg ml-1). Each of these aliquots contained 0.005 g of Ni, corresponding to the amount present in the nickel-base alloys. Dilute HN03 (0.2%) was used for preparing the working solutions. Interferences from Foreign Ions For interference studies, 10 pl of analyte solution (0.02 pg ml-1 TI or 0.05 pg ml-1 Te) and 10 pl of the foreign ionANALYST, SEPTEMBER 1993, VOL. 118 1185 were added separately to a graphite tube. Although there was little chance of direct overlap of atom lines, the AAS signals of either TI or Te was detected with only foreign ion in the graphite tube in order to check whether there were any spectral interferences.The optimum 'dry,' 'char' and 'atom- ize' HGA-700 programme developed in this laboratory was followed and the integrated areas of the absorbance peaks were recorded. The recommended analytical conditions and the temperature programmes for TI and Te are summarized in Table 1 Optimum analytical conditions for TI determination Supply power 6.5 W Background corrector Deuterium lamp Lamp current 5 mA Wavelength 276.8 nm Slit 0.7 nm Integration time 5.0 s Characteristic mass (pg/0.0044 A s) 7.0 Furnace Furnace step No. temperature/'C 1 110 2 600,* 700,t 1000* 3 1350,* 1500,t 1550t 4 2650 5 20 * TI standard solution. t IN100.4. Tracealloy C. Time/s Gas flow/ Ramp Hold mlmin-1 Readon/s 10 15 300 10 30 300 0 5 0 0.0 1 5 300 5 5 300 Table 2 Optimum analytical condition for Te determination Light source Te EDL Supply power 8.5 W Background corrector Deuterium lamp Technique Atomic absorbance - background Lamp current 5 mA Wavelength 214.3 nm Slit 0.2 nm Signal processing Integrated absorbance Internal gas Ar, high-purity (99.99%) Tube/site Pyrol ytic graphite/platform Characteristic mass (pg/0.0044 A s) 25.1 Integration time 5.0 s Time/s Furnace Furnace Gas flow/ step No. temperaturePC Ramp Hold ml min-1 Read on/s 1 110 10 15 300 2 500,* 1000,t 1250t 10 30 300 3 1700,t 1800*,t 0 5 0 0.0 4 2650 1 5 300 5 20 5 5 300 * Te standard solution. t IN100 and Tracealloy C. 1 INlOO + Pd and Tracealloy C + Pd.Tables 1 and 2, respectively. Blanks were run regularly and their values were subtracted from the gross values to obtain the net values reported. Results and Discussion Ammonia Pre-treatment, Chemical Modifier and Temperature Programme The sample solutions of the CRMs of nickel-base alloys, IN100 and Tracealloy C, were first analysed directly by ETAAS. As demonstrated in Table 3, the background signals were much higher than the AAS signals for TI. Without the addition of any chemical modifier, the AAS signals for T1 in INlOO and Tracealloy C were -0.023 and -0.063, whereas the corresponding background absorbances were 3.019 and 9.622, respectively. Apparently, the operation of the automatic deuterium background corrector did not correct the light- scattering signals or molecular absorption of the co-volatilized salts completely.The very high background absorbances prevented direct instrumental measurement of nickel-base alloys by ETAAS. Nickel metal boils at 2840 "C, NiCI2 sublimes at 973 "C8 and TI compounds are volatile, e . g . , TIC1 and TlN03 boil at 720 and 430 "C, respectively.9 It was not possible to remove the interfering molecules completely without loss of the analyte. Chemical modification techniques have been widely applied in trace analysis by ETAAS. The addition of chemical modifiers results in an enhancement of the volatility of unwanted elements and in stabilization of the analyte during the charring stage until most of the matrix components have been vaporized. Sulfuric acid and Pd have often been used as effective chemical modifiers for the determination of TI in environmen- tal samples, as mentioned earlier.The addition of H2SO4 allows TI to form the more stable oxide instead of the volatile chloride41. The effects of these chemical modifiers on the determination of trace amounts of TI in nickel-base alloys were, therefore, studied. The background signals were still much higher than the T1 absorption in the presence of either of the chemical modifiers, as shown in Table 3. Although background correction, chemical modification and thermal pre-treatment can often eliminate interferences in ETAAS, none of them reduced the background signals to an acceptable level when analysing nickel-base alloys. The CRM INlOO contains over 50% Ni. The composition of INlOO as follows (all values in mg g-1): C, 1.5; Cr, 100; Mo, 30; Al, 55; Co, 150; Ti, SO; V, 10; Pb, 0.022; and trace amounts of Bi, Ag, Se, Te, TI, Sb, As, Cd, Ga, Sn, Zn, Mg, Ca and In.The reference value for TI is 2 pg g-1. The composition of Tracealloy C is as follows (all values in mg g-1): C, 1.2; Cr, 120; Al, 20; Co, 85; Ti, 20; B, 0.1; W, 17.5; Nb, 9; Zr, 1; Ta, 17.5; and Hf, 12. These values are not certified but are provided for information only. The certified value for TI is 0.252 k 0.003 pg g-1. When ammonia solution was added to a solution of nickel-base alloys, most of the Nil[ would be precipitated as Ni(OH)2, which would dissolve and form [Ni(NH3)4]2+. The addition of ammonia solution also brought about the forma- ~~~~ ~ Table 3 Effects of ammonia solution pre-treatment and chemical modifiers on the AAS signals for TI and the corresponding background signals for nickel-base alloys Chemical modifer NH3, NH.3, IN 100 AA - BG -0.023 -0.016 -0.007 0.023 0.084 BG 3.019 0.189 1.512 0.004 0.004 Tracealloy C AA - BG -0.063 -0.013 0.020 0.012 0.052 BG 9.622 3.188 3.013 0.758 0.002 Alloy Method' None 1% HzS04 Pd HCI-HN03 HN03 * AA = atomic absorbance; BG = background.The integrated absorbance was recorded.1186 ANALYST, SEPTEMBER 1993, VOL. 118 tion of [ C O ( N H ~ ) ~ ] ~ + under the same conditions. Thallium was coprecipitated and separated from most of the interfering ions. The resulting mixture was centrifuged in order to avoid the adsorption of any precipitate onto the filter-paper. A volume of the decantate was injected onto the platform and analysed by ETAAS.No TI signal was detected, as expected. When the precipitates were dissolved in a mixture of HCI and HN03, the background signal for Tracealloy C was still much higher than the AAS signal or of TI. As shown in Table 3, the integrated absorbance of TI and the background signals were 0.012 and 0.758, respectively. This indicated the interference of chloride, as reported earlier.41-43 When the precipitates were dissolved in HN03 (1 + 1), a great improvement in the signal for TI in nickel-base alloys was found. In addition, the background signals were eliminated for both TNlOO and Tracealloy C, as shown in Table 3. Pre-treatment with ammonia solution provided an effective solution to the serious background problem in the determina- tion of trace amounts of TI in nickel-base alloys by ETAAS.Figs. 1(A) and 2(A) show the influence of charring and atomization temperatures on the integrated absorbances of TI in a standard solution. For the TI standard, the optimum temperatures for charring and atomization were 600 and 13.50 "C, respectively. Figs. 1(B) and (C) and 2(B) and (C) show the effects of charring and atomization temperatures on the integrated absorbances of TI in nickel-base alloys. The absorbances of IN100 and Tracealloy C were monitored at an atomization temperature of 1.500 "C. The appropriate maxi- mum charring temperatures for these two samples were 700 and 1000 "C, respectively. The signal intensity of the TI 0.1 v) . 2 0.08 5 0.06 m 2 m -u 2 0.04 h f 0.02 DJ 4- A - C 0 ' I I I I r 200 400 600 800 1000 1200 Charring temperature/"(= Fig.1 Absorbance of T1 as a function of charring temperature at an atomization temperature of 1500 "C. A, 0.40 ngT1 standard; B, 0.2088 mg IN100; and C, 1.6476 mg Tracealloy C v) 0.08 -. 0 c m I) & 0.06 I] m 0.02 standard reached a plateau when the temperature increased to 13.50 "C, whereas the signal for TI in Tracealloy C decreased rapidly above 1500 "C. The optimum atomization temperature of TI in IN100 was 15.50 "C. As nickel has been used to stabilize volatile elements in trace analyses by ETAAS,31?44 it was worthwhile differen- tiating the temperature dependence of the AAS signals for Te in a standard solution from that in nickel-base alloys. Hence, the optimum temperature for both charring and atomization steps could be found.Figs. 3 and 4 show the influence of charring and atomization temperatures on the integrated absorbances of Te in various solutions of standard Te, IN100 and Tracelloy C. Tellurium atoms in the standard solution were stable only up to SO0 "C; the dramatic decrease in the AAS signals at higher temperatures was due to the expulsion of Te from the ends of the graphite tube or the injection hole. In contrast, Te atoms in the alloys were stable up to 1000 "C. This indicated that the maximum permissible charring temper- ature had been increased from 500 to 1000 "C as a result of thermal stabilization of Te, which formed intermctallic compounds on the platform during the charring step. In addition, the nickel modifier brought about a 100 "C decrease in the atomization temperature.Palladium is also an effective chemical modifier in the determination of Te by ETAAS.33,44 With the addition of Pd, Te could be converted into a more thermally stable species. The mechanisms involved in thermal stabilization by chemical modifiers can be divided into two different concepts. One is the formation of solid solutions between the analyte and the modifier and/or the atoms (ions) of the analyte replace atoms (ions) in the crystal lattice of the modifier by isomorphous substitution. The other is the less complicated formation of chemical compounds that have defined properties and struc- ture between the analyte and the modifier.45 Different 0.1 I 1 $ 0.08 * 0.04 f 0.02 I 1 I I I 1 200 400 600 800 1000 1200 1400 Charring temperaturePC Fig.3 Absorbance of Te as a function of charring tcmpcrature at an atomization tempcrature of 1650 "C. A, 1 .0 ng Tc standard; B, 36.0 pg IN100; and C, 45.7 pg Tracealloy C 0.12 I I 0 1000 1200 1400 1600 1800 2000 Atomization temperature/"C Fig. 2 Absorbance of T1 as a function of atomization temperature at diffcrcnt charring temperatures. A, 0.4 ng T1 standard at 600 "C; B, 0.2088 mg IN100 at 800 "C; and C, 1.6476 mg Tracealloy C at 1100 "C 0 1400 1600 1800 2000 2200 Atomization temperaturePC Fig. 4 Absorbance of Te as a function of atomization tcmperaturc at different charring ternperaturcs. A, 1 .0 ng Tc standard at 500 "C; B, 36.0 pg IN100 at 750 "C; and C, 45.7 pg Tracealloy C at 750 "CANALYST, SEPTEMBER 1993, VOL.118 1187 chemical modifiers lead to different results. An example is the work reported by Volynsky et aZ.44 The reduction temperature of Ga203 and PbO with graphite was decreased by PdC12 whereas only the reduction temperature of Ga203 was decreased by NiC12. Although a nickel modifier had been shown to be effective in raising the charring temperature of Te, it was of interest to establish whether there would be further a improvement with the Pd modifier. Consequently, the charring temperature was increased to 1250 "C as shown in Fig. 5. However, there was a slight increase in the peak area, by 1.7 and 3.3% for IN100 and Tracealloy C, respectively, when the atomization temperature was increased from 1700 to 1800 "C. The atomization plots are shown in Fig.6. Fig. 7 shows the effect of Pd on the AAS signals. The AAS signals for Te in IN100 reached a maximum when 2 pg of Pd were added. Although only 0.2 pg of Pd was required for Tracealloy C to obtain the maximum absorbance, there was a slight fluctuation in the AAS signals. In order to ensure that sufficient Pd modifier was present, an amount of 5 pg of Pd was used for each measurement. Interference Studies In order to understand the effects of HN03, HCI, HF, ammonia solution and various metal ions on the determina- tion of TI and Te, a number of standard solutions spiked with different foreign ions were measured by ETAAS. The relative absorbance, defined as Relative absorbance = was monitored. For Te, an amount of 500 pg was determined instead of 200 pg. Fig.8 shows the variation in the relative absorbances of T1 as a function of the concentration of the acids and ammonia absorbance of 200 pg TI + foreign ions absorbance of 200 pg TI w !! 0.04 - a c w - 0.03 I I 300 500 700 900 1100 1300 1500 Charring temperature/"(= Fig. 5 Charring profiles of Te in nickel-base alloys with Pd modifier at an atomization temperature of 1700 "C. A, 36.0 pg IN100 + 5.0 pg Pd; and B, 43.2 pg Tracealloy C + 5.0 pg Pd 0.08 v) . $ 0.06 C m + :: 0.04 -0 w F 0.02 w C - 0 1500 1700 1900 21 00 Atomization temperaturePC Fig. 6 Atomization profiles of Te in nickel-base alloys with Pd modifier at a charring temperature of 1250 "C. A, 36.0 pg IN100 + 5.0 pg Pd; and B, 43.2 pg Tracealloy C + 5.0 pg Pd solution (0.020-4.60 mol 1 - 1 ) . As TlN03, which was decom- posed at 450 "C, could be converted into the more stable oxide,41 no losses were observed with the addition of HN03.This was consistent with the observation reported by Fuller.17 The very large depression of the TI signals in the presence of HF and HCI was due to the formation of relatively volatile thallium halides.9 Ammonia solution caused less than a 10% variation in the relative absorbances even though the concen- tration was increased from 0.026 to 2.60 mol 1-1. Potential interfering metal ions, which included Ni, V, Mo, Nb, Ta, Fe, Cu, Co, Al, Cr, Ti, Ga, In, Zr, Hf and Pb, were studied. The relative absorbances were monitored with various concentrations of foreign ions up to 800 pprn (Table 4). Nickel and Ga caused an enhancement of the AAS signals.The relative absorbance increased from 1.20 to 2.10 as the concentration of Ni increased from 25 to 800 ppm whereas Fe gave about 20% enhancement of the T1 signal. Gallium enhanced the TI signal to a smaller extent; the relative absorbance increased only from 1.51 to 1.72 in the same range of Ga concentration. Cobalt, Al, Zr and Hf caused severe inhibition of the TI signals whereas Ti had no detectable effect. However, HCl in the metal ion solutions also caused some inhibition on the T1 signals. The interference thus observed could represent contributions from both acids and metal ions. By comparing the interference thus observed in Table 4 with the corresponding net effect of acids in Fig. 8, it could be concluded that Hf had a negligible effect on the TI signals whereas Zr and Al caused serious depression of the T1 signal.The relative absorbance increased linearly with the concen- tration of In at first. It then flattened to a plateau when the concentration exceeded 300 ppm. Chromium gave about a 30% enhancement of the relative absorbance in the concentra- tion range 25-800 ppm whereas Pb caused about a 30% decrease in the same range. Tantalum caused a negligible effect up to 100 ppm. However, the TI signal decreased rapidly at higher concentrations of Ta. Vanadium, Nb and Cu also inhibited the TI signal to different extents. Molybdenum had a negligible effect on the signal for TI. s n m 0.06 0.05 E 0.04 -u w a - 1 1 I I I 0 2 4 6 8 10 Amount of Pd/Fg Affect of Pd on the AAS signal for Te. A, 54.4 pg IN100; and Fig.7 B, 80.8 pg 1.1 8 0.9 C m + 2 0.7 m a, .- A C : / 0.5 U.J 0 1.0 2.0 3.0 4.0 5.0 Concentration/mol I 1 Effect of ammonia solution and various acids on the signal for Fig. 8 TI (n r= 5). A, HNO,; B, HCI; C, HF; and D, NH,1188 ANALYST, SEPTEMBER 1993, VOL. 118 Table 4 Effects of various metal ions on the relative absorbances of T1 Compound Concentration Relative Ion added Ni c o Al Cr Ti TiC14 V Mo Nb Ta Fe c u Ga (ppm) absorbance* 25 50 100 300 500 800 25 50 100 300 500 800 25 50 100 300 500 600 800 25 50 100 300 400 500 600 700 800 25 50 100 300 500 800 25 50 100 300 500 800 25 50 100 300 500 800 25 50 100 300 500 800 25 50 100 300 500 800 25 50 100 300 500 800 25 50 100 300 500 800 25 50 100 1.20 1.19 1.28 1.39 2.10 2.10 1.02 0.94 0.66 0.32 0.19 0.08 1.14 1.17 1.17 1.09 0.99 0.84 0.59 1.20 1.27 1.27 1.38 1.40 1.36 1.38 1.41 1.33 0.96 0.97 1.02 1 .oo 1.02 1 .00 0.27 0.26 0.28 0.27 0.25 0.25 1.03 0.92 1.05 1.00 1.06 0.97 0.73 0.70 0.36 0.22 0.22 0.23 1.10 1.11 0.99 0.75 0.65 0.42 1.18 1.17 1.18 1.18 1.19 1.16 0.99 0.79 0.51 0.38 0.31 0.24 1.51 1.54 1.79 Integrated absorbance? 0.004 0.005 0.004 0.010 0.035 0.012 0.003 0.003 0.002 0.005 0.002 continued- Table "continued Ion In Zr Hf Pb Compound added Zr in 5% m/m HCI Hf in 10% m/m HCI Concentration Relative Integrated (ppm) absorbance* absorbance+ 300 1.79 500 1.71 600 1.71 700 1.78 800 1.72 0.002 25 1.02 50 1.23 100 1.32 300 1.43 500 1.43 800 1.43 -0.004 25 1.06 50 1.03 100 0.95 300 0.78 500 0.64 800 0.53 0.017 25 1.09 50 1.02 100 1.05 300 0.86 500 0.88 800 0.71 0.006 25 0.69 50 0.67 100 0.63 300 0.65 500 0.62 800 0.64 0.002 * The relative absorbances are averages of five determinations.absorbance of 200 pg TI + foreign ions absorbance of 200 pg TI Relative absorbance = t Integrated absorbances were determined with only foreign ion in the graphite tube (duplicate determinations). 1.1 a, C m I) $ 0.9 I) m a, 'E 0.7 a, [r - 0.5 ' 1 1 1 0 1.0 2.0 3.0 4.0 5.0 Concentration/mol 1-1 Fig. 9 Effects of ammonia solution and various acids on the absorbance signal for Te (n = 5 ) . A, HNO,; B, HCl; C, HF; and D, NH3 Among these foreign ions, Cr, Ti, V and Zr showed some degree of overlapping of the absorbance spectrum with the emission spectrum of T1. The peak areas measured with only foreign ions in the graphite tube are given in Table 4.The results for Te are given in Fig. 9 and Table 5. Both HN03 and HC1 had negligible effects on the relative absorbances (less than 10% deviation), even though the concentration was increased from 0.024 to 4.50 mol 1-1. Hydrofluoric acid caused about a 30% depression of the Te signals in the concentration range 0.046-6.90 mol I-'. Ammonia solution gave a major depression of the Te signals; the relative absorbance was only 0.51 in the presence of 2.60 mol 1-1 ammonia solution. The effects of foreign ions on the Te signal were also investigated. The relative absorbances were monitored withANALYST, SEPTEMBER 1993, VOL. 118 1189 Table 5 Effects of various metal ions on the relative absorbances of Te Table S-continued Compound Concentration Relative Integrated Compound Concentration Relative Integrated Ion added (ppm) absorbance* absorbance' Ion added (ppm) absorbance* absorbancet Ni c o Al Cr 100 300 500 800 Ti TiCL V Mo Nb Ta 100 300 500 800 25 50 100 300 500 800 K2Cr207 25 50 100 300 500 600 700 800 25 50 100 300 500 800 V in HN03 25 50 100 300 500 800 Mo in H20 25 50 100 300 500 800 Nb in H20 25 50 100 300 500 800 Ta in trace Hf 25 50 100 300 500 800 25 50 100 300 500 800 25 50 100 300 Al in 1% v/v HCl Zr 1.05 1.19 1.12 1.79 2.16 4.70 1.17 1.16 1.03 1.05 1.08 1.09 1.22 1.30 1.26 3.58 3.55 5.76 1.17 1.15 1.18 1.18 1.16 1.20 1.24 1.23 1.09 1.09 0.87 0.94 0.87 0.90 0.53 0.55 0.54 0.54 0.55 0.57 0.77 0.74 0.81 0.90 0.83 0.86 1.04 1.03 0.90 0.89 0.87 0.89 0.71 0.69 0.69 0.70 0.73 0.73 0.70 0.71 0.75 0.97 1.12 1.29 0.55 0.57 0.55 0.56 0.015 0.002 0.256 0.001 0.022 0.086 0.012 0.270 0.179 0.137 600 1.21 700 1.26 800 1.22 0.168 Hf Hf in 10% m/m 25 1.21 HCI 50 1.25 100 1.30 300 1.40 500 1.40 800 1.41 0.025 Pb Pb(N03)2 25 0.80 50 0.80 100 0.81 300 0.73 500 0.76 600 0.71 700 0.71 800 0.75 0.015 Se H2Se03 in 2% 25 0.93 HN03 50 0.96 100 0.98 300 1.02 500 0.99 800 1.02 0.007 * The relative absorbances are averages of five determinations.absorbance of 500 pg Te + foreign ions absorbance of 500 pg Te Relative absorbance = t Integrated absorbances were determined with only foreign ion in the graphite tube (duplicate determinations). various concentrations of foreign ions up to 800 ppm. Table 5 summarizes the effects of metal concentrations on the relative absorbances. Nickel and A1 enhanced the absorbance signals of Te.The relative absorbance increased from 1.05 to 2.16 as the concentration of Ni increased from 25 to 500 ppm. Further addition of Ni would increase both the Te signals and the background signals. With 800 pprn of Ni, the relative absorbance was 4.70 whereas the corresponding background was 1.30. The relative absorbance increased from 1.22 to 5.76 in the same range of Al concentrations. However, an extremely high background absorbance with a value of 1.49 appeared when 800 pprn of A1 were added. The relative absorbance increased gradually from 1.21 to 1.41 on addition of 25-800 ppm of Hf. Chromium gave about a 20% enhance- ment of the relative absorbance in the concentration range of 25-800 pprn and Zr gave a similar level of enhancement in the same concentration range.The effect of Se was negligible. Although there was a small degree of fluctuation in the relative absorbance in the presence of Co and Cr, these metal ions caused only about a 10% deviation in the relative absorbance. The effect of Pb was a 20-30% suppression of the Te signal in the concentration range 25-800 ppm. The addition of V, Nb, Ta, Fe, Cu and Mo also resulted in decreases in the Te signal. Peak areas (summarized in Table 5) measured with only foreign ions in the graphite tube showed that the Te emission spectrum overlapped with most of the absorbance spectra of foreign ions except those for Co, Cr and Se. Quantitative Analysis 500 0.57 The CRMs of nickel-base alloys were analysed to validate the 800 0.56 0.208 accuracy of the proposed procedure.The solid sample of IN100 or Tracealloy C was decomposed with a mixture of Zr in 5% m/m 25 1.02 1.05 HN03 and HCI. A precipitate containing trace amounts of T1 100 1.05 mn I n7 was formed in ammonia solution. Thallium was then deter- HCl 50 J"" I.", mined by ETAAS with the stabilized temperature platform furnace. Tellurium was determined directly with ETAAS. 500 1.19 continued-1190 ANALYST, SEPTEMBER 1993, VOL. 118 Table 6 Analytical parameters and results of TI determinations Calibration graph- TI (ng) = K x absorbance + B Linear rangdng K B R2 0.0604.500 2.3853 0.0027 0.9995 Anulytical results* - IN100 Tracealloy C Certified value/pg g- I 2; 0.252 k 0.003 Determined valuc/pg 6.1 1.94 k 0.09 0.251 * 0.005 Recovery (YO)) 99.8 -t 4.0 101.8 f 2.9 * Results of five determinations.+ This value was not certified but given for information only. + The amount of standard TI added ranged from 0.06 to 0.60 ng. Detection limit/pg 15 15 0.180 I i 0.144 m a, (u . 0.108 fl % a 0.072 a, i- F [5) a, + 0.036 0 0 0.2 0.4 0.6 0.8 1.0 Amount of Telng Fig. 10 Calibration graph for Te determinations. A, Established with standard Te solutions; and B, established with standard Te solutions + 0.005 mg Ni For T1 determinations, a calibration graph with a linear range from 0.060 to 0.50 ng was established [Tl (ng) = K x absorbance + B; K = 2.3853, B = 0.0027, correlation coefficient = 0.99951. There was good agreement between the certified values and the results obtained, as indicated in Table 6. The results found for IN100 and Tracealloy C were 1.94 k 0.09 pg g-1 (s, = 5%) and 0.251 k 0.005 pg g-1 (s, = 2%), which represent 3 and 0.4% differences from the reference value of 2 pg g-I and the certified value of 0.252 IL 0.003 pg g-l, respectively.By adding various amounts of standard TI to sample solutions (0.06-0.60 ng), the results gave recoveries of 100 31 4 and 102 k 3% for IN100 and Tracealloy C, respectively. The detection limit, defined as 3s (s = standard deviation of 12 consecutive measurements), was 15 The certified values for Te in IN100 and Tracealloy C are 9 k 1 and 5.9 k 0.6 ppm, respectively. The need to use Ni to stabilize the volatile Te is again verified in Fig. 10, which shows the calibration graphs for standard Te solutions obtained with and without Ni modifier.The linear range is up to 4 ng of Te. The calibration graph curves towards the concentration axis at higher concentrations. A compilation of the results is given in Table 7. The determined values obtained from the calibration graph, with a linear range from 0.100 to 1.00 ng [Te (ng) = K x absorbance + B; K = 5.7251, B = 0.0004, correlation coefficient = 0.99931, for IN100 and Tracealloy C were 9.03 2 Pg g-l. Table 7 Analytical parameters and results of Te determinations Calibration graph'- Te (ng) = K X absorbance + B Linear rangdng K B R2 5.725 1 0.0004 0.9993 0.100-1 .OO Analytical results- IN 100 Tracealloy C Certified value/pg g-' 9 f l (1) with Ni/yg g-1 ( n = 5 ) f s, (Yo) 1 Detection limit/pg g-1 35 Recovery (Yo) 1 Detect ion limi t/pg 18 Recovery (% ) 9.03 t.0.09 98 f 4 9.04 +- 0.31 92 f 5 12.4 t- I .9 (2) with Ni and Pd/pg g-1 ( n = 5 ) (3) With the standard additions method/pg g-1 ( n = 3) 5.9 k 0.6 5.93 * 0.26 4 35 97 -t 4 5.78 k 0.20 18 91 * 3 11.9+ 1.6 Calibration graph established with standard Te solutions + 0,005 mg of Ni. 1 II = Total number of determinations. The amount of standard Te added ranged from 0.10 to 0.60 ng. 0.09 and 5.93 IL 0.26 pg g-1, respectively. The results obtained in this work were in close agreement with the certified values. By adding various amounts of standard Te to sample solutions (0.10-0.60 ng), the results gave recoveries of 98 & 4 and 97 & 4% for IN100 and Tracealloy C, respectively. The detection limit was 35 pg g-1. Although the addition of Pd modifier resulted in a higher charring temperature, the quantitative results did not improve.As shown in Table 7, an absolute error of -0.12 pg g-' for Tracealloy C was obtained. The recoveries for the nickel-base alloys were 92 IL 5 and 91 k 3%, which were worse than those obtained without the addition of Pd modifier. Although the addition of Pd modifier gave a better detection limit (18 pg g-*), it led to worse results in terms of accuracy, precision and recovery. The Te contents determined by the standard additions method were 12.4 & 1.9 and 11.9 31 1.6 pg g-1 for IN100 and Tracealloy C, respectively. Apparently, the standard additions gave poor precision and accuracy. Conclusions Pre-treatment with ammonia solution effectively eliminated background signals in the determination of trace amounts of TI in nickel-base alloys by ETAAS.For Te determination, Ni worked effectively as a chemical modifier. After the solid sample of IN100 or Tracealloy C had been decomposed with a mixture of HN03 and HCl, the T1 present in the nickel-base alloys was coprecipitated with ammonia solution. The result- ing precipitates were dissolved and analysed by ETAAS. Trace amounts of Te were determined directly without ammonia pre-treatment. The calibration graphs for the determination of TI and Te were obtained using aliquots of HN0,-diluted metal stock solutions; however, Ni was added to the Te standard solution. This indicated that the matrix matching of samples to calibration standards was not neces- sary with the ammonia solution pre-treatment.The proposed method provided precise and accurate determinations of trace amounts of TI and Te in a nickel-base alloys with relatively low detection limits of 15 and 35 pg g-l, respectively. Although the addition of Pd modifier gave a better detection limit (18 pg g-1) for Te, it resulted in poor accuracy, precision and recovery. The financial support of this work by a grant from the National Science Council of the Republic of China is gratefully acknowledged.ANALYST, SEPTEMBER 1993, VOL. 118 1191 I 2 3 4 5 6 7 8 9 I 0 11 12 13 14 15 16 17 18 19 20 21 22 23 References Andrcws, D. G., and Hcadridgc, J . B., Anulyst, 1977,102,436. Backman, S . , and Karlsson, R. W., Analyst, 1979, 104, 1017. Hcadridgc, J . B., and Nicholson, R. A., Analyst, 1982, 107, 1200.Ford, D. A . , Met. Technol., 1984. 11, 438. Lowe, D. S . . Analyst, 1985, 110, 583. Wclchcr, G. G., Kricge, 0. H., and Marks, J. Y., And. Chem., 1974, 46. 1227. Forrcstcr, J . E., Lehccka, V., Johnston, J . R., and Ott, W. L., At. Ahsorpt. Newsl., 1979, 18, 73. Dulski, T. R., and Bixlcr, R. R., Anal. Chim. Acta, 1977, 91, 199. Gricpink, B., Sager, M., and Tolg, G., Pure Appl. Chem., 1988, 60, 1425. Labuda, J., and Vanickova, M., And. Chirn. Acta, 1988, 208, 219. Murti, S. S., Sambasiva Rao, I. V., and Rajan, S. C. S., Tulanta, 1989, 36, 601. Hafcz, M. A. E., Abdallah. A. M. A . , and Wahdan,T. M. A. E . , Anulysf, 1991, 116, 663. Henke, G., Fresenius’ J. Anal. Chem., 1991, 339, 245. Waidmann, E., Stocppler, M., and Hciningcr, P., Analyst, 1992, 117, 295. Flanjak, J., and Hodda, A. E., Anal. Chim. Acta, 1988, 207, 283. Koirytohann, S. R., Glass, E. D., and Lichtc, F. E., Appl. Fuller, C. W., Anul. Chim. Acta, 1976, 81, 199. Slavin, W., Carnrick, G. R., and Manning, D. C., And. Chim. Acta, 1982, 138, 103. Shan, X., Ni, Z., and Zhang, L., Tulanta, 1984, 31, 150. Manning, D. C., and Slavin, W., Spectrochirn. Acta, Part B , 1988, 43, 1157. Slavin, W., and Manning, D. C., Spectrochim. Actu, Part B, 1980. 35, 701. Ucda, .I., and Kitadani, T., Analyst, 1988. 113, 581. Niskavaara, H., and Kontas, E., Anul. Chim. Acta, 1990, 231, 273. Sl,ectro.sc. , 198 1 , 35, 22. 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 Nakashima, S., Sturgeon, R. E., Willie, S. N . , and Berman, S. S., Anul. Chim. Acta, 1988, 207, 291. Fox, G. J . , At. Spectrosc., 1990, 11, 13. Beaty, R. D., At. Ahsorpt. Newsl., 1974, 13, 38. Sighinolfi, G. P., Santos, A. M., and Martinelli, G., Tuluntu, 1979, 26, 143. Chiou, K. Y., and Manuel, 0. K., Anal. Chem., 1984,56,2721. Muangnoicharocn, S . , Chiou, K.-Y., and Manuel, 0. K., Tulanta, 1988, 35, 679. Kunsclman, G. C., and Huff, E. A., At. Ahsorpt. Newsl., 1976, 15, 29. Edigcr, R. D., At. Ahsorpt. Newsl., 1975, 14, 127. Donaldson, E. M., and Lcavcr, M. E., Talanta, 1990, 37, 173. Zhang, L., Ni, Z.-M., and Shan, X.-Q., Spectrochim. Actu, Purr B, 1989, 44, 7.5 1. Bedrossian, M., Anal. Chem., 1984, 56, 31 1. Hubert, A. E., and Chao, T. T., Taluntu, 1985, 32, 568. Greenland, L. P., and Campbell, E. Y., Anal. Chim. Aclu, 1976, 87, 323. Wickstrom, T., and Lund, W., and Bye, R., Anal. Chim. Actu, 1988, 208, 347. Corbett, J . A., and Godbcer, W. C., Anal. Chim. Actu, 1977, 91, 211. Fcrri, T., Morabito, R., Pctronio, B. M., and Pitti, E., Talantu, 1989, 36, 1259. Trivcdi. B. V., and Thakkar, N. V., Talunta, 1989. 36, 786. Leloux, M. S., Lich, N. P., and Claude, J . R., At. Spectrosc., 1987, 8, 71. Web, B . , Schlcmmer, G . , and Mudakavi. J . R., Anal. Chem., 1988, 60, 2567. Hamid, H. A., Al Joboury, M. I., and Mohammed, A. K., Anal. Chim. Actu, 1991, 243, 239. Volynsky, A., Tikhomirov, S . , and Elagin, A., Analyst, 1991, 116, 145. Mandjukov, P. B., Vassileva, E. T., and Simeonov, V. D., Anal. Chem., 1992, 64, 2596. Paper 31007831 Received February 9, 1993 Accepted April 19, 1993

ISSN:0003-2654    

DOI:10.1039/AN9931801183

出版商:RSC    

年代:1993

数据来源: RSC

摘要:

ANALYST, SEPTEMBER 1993, VOL. 118 1193 Determination of Citric Acid and Oxalacetic Acid in Foods by Enzymic Flow Injection Milagros Plants, Fernando Lazaro, Rosa Puchades and Angel Maquieira" Department of Chemistry, Polytechnic University of Valencia, 46071 Valencia, Spain A reversed-flow injection method for the determination of citric acid in foods is proposed. Two enzymes, citrate lyase and malate dehydrogenase (MDH), are used, the latter being immobilized. This method, based on the decrease in the absorbance of reduced nicotinamide adenine dinucleotide, has a linear range between 1 and 20 mg dm-3, with a relative standard deviation (RSD) of 2.2%, an average recovery of 100.9% and a sampling frequency of 20 h-1. Additionally, a flow injection method for the determination of oxalacetic acid in fruit juices using immobilized MDH is also proposed. The features of the method are as follows: linear range of the calibration graph 1-40 mg dm-3, RSD 1.4%, average recovery 96.8% and sampling frequency 40 h-1.Keywords: Citric acid; oxalacetic acid; enzymic flow injection; food Citric and oxalacetic acids are essential components in all living beings. They directly take part in the production of energy through the tricarboxylic acid cycle, as well as in several processes associated with the metabolism of fatty acids, carbohydrates and certain amino acids.1 Citric acid is widespread within the plant kingdom and is often predominant in the total carboxylic acid content. Therefore, it is a commonly used parameter in the analysis of plants and their derivatives.Furthermore, owing to its innocuity it is often used as an antibacterial substance and as an additive for the control of pH.2 Several methods have been proposed for the determination of citric acid, such as that based on oxidation with KMn04,3 conductimetric methods4 or spectrophotometric methods based on general reactions with carboxylic acids.5.6 Two of these methods have involved flow injection (FI) techniques.4.6 However, when a high specificity is required, these methods cannot be applied, as citric acid often occurs together with many organic acids. Therefore, chromatographic7 or enzymic8 methods are used in these instances, but these methods are time consuming . On the other hand, oxalacetic acid is not commonly determined as it occurs at low concentrations.In the litera- ture, only generic methods for a-ketoacids,5 chromatographic methods7 and bioluminescence FI methodsg710 have been found. The aim of this work was to establish methods for the determination of both acids in foods by using immobilized enzymes with three purposes: (u) to provide this enzymic determination with the advantages of FI;I1 ( b ) to minimize enzyme consumption during the analysis by means of its immobilization; and (c) to simplify the proposed method,s reducing the number of enzymes used. The classical enzymic methods for the determination of citrate makes use of citrate lyase (CL), lactate dehydrogenase (LDH) and malate dehydrogenase (MDH), the consumed reduced nicotinamide adenine dinucleotide (NADH) being monitored spectrophotometrically at 340 nm.It is based on the following reactions: CL Citrate oxalacetate + acetate (1) Oxalacetate + pyruvate + COz (2) OD LDH Pyruvate + NADH + H+ + L-lactate + NAD+ (3) * To whom correspondence should be addressed. It admits the existence of oxalacetate decarboxylase (OD) together with the enzyme CL. In order to ensure that all oxalacetate reacts (despite the low amount of OD present), MDH is added: MDH Oxalacetate + NADH + H+ + L-malate + NAD+ (4) The method consists of two steps: (i) A mixture of LDH, MDH, NADH and buffer is added to the sample to remove all oxalacetate and pyruvate present. Reactions (3) and (4) take place. It lasts 5 min. (ii) Citrate lyase, with a certain amount of OD, is added, so that the citrate formed reacts according to reactions (1)-(4), giving rise to a decrease in the absorbance of NADH (analytical signal). Afterwards, a blank assay is carried out to measure the possible decrease in absorbance occurring due to the degrada- tion of the NADH during the long time of analysis (10-15 min) .The proposed method for citrate determination is based on reactions (1) and (4) and avoids the use of OD and LDH for two reasons: ( a ) according to the manufacturers, the commer- cial product of CL contains some specified impurities, but not OD; (b) in the FI technique, the reactions involved are not usually required to proceed to completion.ll As long as sufficient reaction occurs to achieve the desired sensitivity, any additional reactions resulting from impure reagents, e.g., OD in the CL, will be constant and accounted for by the standard calibration.The proposed method for oxalacetate determination is exclusively based on reaction (4), the decrease in the absorbance of NADH being controlled at 340 nm. This reaction catalysed by MDH has already been studied, though in the opposite direction (malate -+ oxalacetate) to that for the determination of malic acid;12 hence, a more exhaustive study has been carried out. Experimental Apparatus A four-channel Gilson Minipuls-3 peristaltic pump (Gilson, Villiers, France), a variable-volume Rheodyne 5041 injection valve (Rheodyne, Cotati, CA, USA), a Rheodyne 5302 diversion valve, a laboratory-made poly(methy1 methacrylate) mixing point, a Hellma 178.712QS flow cell (Hellma, Jamaica, NY, USA), with an inner volume of 18 mm3 and a Philips PU-8625 spectrophotometer (Philips Analytical, Cambridge, UK), connected to a Perkin-Elmer 56 recorder (Perkin- Elmer, Analytical Instruments, Norwalk, CT, USA), were1194 ANALYST, SEPTEMBER 1993, VOL.118 used in the determinations. Poly(viny1 chloride) pump tubing of different diameters, suited to the required flow rate, and poly(tetrafluoroethy1ene) coils of 0.5 mm i.d., were also used. Reagents Malate dehydrogenase (E.C.1.1.1.37), approximately 1200 U mg-1 protein (1 U = 16.67 nkat), CL (E.C.4.1.3.6), approximately 8 U mg-1 protein, and NADH were supplied by Boehringer Mannheim, Mannheim, Germany. Controlled-pore glass (CPG 0240,80/120 mesh, 240 A mean pore diameter) (Sigma, St. Louis, MO, USA), 3-(amino- propy1)triethoxysilane (Sigma), protected from moisture and stored at 4 "C, and 25% glutaraldehyde (Sigma) were used to immobilize MDH and CL.Standard solutions of citric and oxalacetic acids were prepared from citric acid monohydrate, analytical-reagent grade (Merck, Darmstadt, Germany), and oxalacetic acid, approximately 98% (Sigma), respectively. The oxalacetic acid solution was prepared weekly, owing to its instability. For the determination of oxalacetic acid, a stock buffer solution of 0.5 mol dm3 KH2P04 (Merck), adjusted to pH 7.8 with KOH (Merck), was prepared, and further diluted 1 + 5 to obtain the buffer solution B (Fig. 1). The same concentration of buffer must be included in the reagent solution, R (260 mg dm-3 NADH) and the sample solution, S, in order to avoid problems in the analytical signal resulting from differ- ences in the refractive index.For the determination of citric acid, a stock buffer solution of 0.5 mol dm-3 glycylglycine (Boehringer Mannheim) was prepared, the pH being adjusted to 7.8 with KOH and further diluted 1 + 5 to obtain the buffer solution, B' (Fig. 2). The sample solution, S', and the reagent solution, R' [0.67 U cm-3 CL-0.5 mmol dm-3 ZnC12 (Merck)4.51 g dm-3 NADH] must have the same concentration of buffer as in buffer solution, B'. Configurations and Procedures The determination of oxalacetic acid was performed with the manifold shown in Fig. 1. It is based on the insertion of a sample volume, V , into a carrier stream, B, which then merges with the reagent solution, R. As the sample plug passes through the immobilized MDH reactor the enzyme catalyses the above described reaction (4), giving rise to an FI peak on passage of the reacting plug through the detector.This peak is negative, as it corresponds to a decrease in the absorbance of the baseline. In food analysis it is necessary to carry out a blank assay for each sample, by changing the reagent solution R (which U Pump t I& w Fig. 1 Configuration for oxalacetic acid determination. B, Buffer; R, reagent; S, sample, V, injection valve; and W, waste. The values for the flow rates ( q , , q2) and the size of the MDH reactor are shown in Table 1 Fig. 2 Configuration for citric acid determination. B', Buffer; R', reagent; S', sample; V, injection valve; DV; deviation valve; and W, waste.The values for the flow rate ( q l ) and the sizes of the tubing reactor (r,) and the MDH reactor are shown in Table 1 contains NADH) for the buffer solution B, in order to eliminate the matrix influence on the analytical signal. For the determination of citric acid, the manifold depicted in Fig. 2 was used. The reagent solution, R', which contains CL, can be injected into a stream of sample solution, S', or into a buffer solution, B' (which is used as the carrier stream), by using the diversion valve (DV, reverse FI configuration). When the carrier is the buffer solution B', the signal obtained corresponds to the absorbance (Ao) of NADH injected with the reagent solution plug, which has not reacted. On the other hand, when the carrier is the sample solution, S', the signal obtained corresponds to the absorbance (Al) of unconsumed NADH after the reactions (1) and (4) have been accomplished (in the rl and MDH reactor, respectively).The analytical signal (proportional to the concentration of citric acid) will be derived from the decrease in the absorbance (Ao - A * ) . In food analysis it is not necessary to carry out a blank assay because the baseline coincides with the absorbance of the matrix. Sample Pre-treatment Liquid samples only require filtration and suitable dilution, but solid samples require a prior extraction step with distilled water. Immobilization of Enzymes The glass, after cleaning with 30% m/m HN03, was alkylami- nated with 3-(aminopropyl) triethoxysilane and the cross-link- ing agent (glutaraldehyde) coupled as described by Masoom and Townshend.13 Immobilization was performed by the following procedure: MDH, 80 mg (5600 U), or CL, 150 mg (60 U), was dissolved in 2.0 cm3 of cold (4 "C) deoxygenated phosphate buffer (0.1 mol dm-3, pH 6.0) and added to 300 mg of activated glass. This solution was kept at 4 "C overnight in a nitrogen atmosphere, then the glass was packed in a labora- tory-made glass reactor [1.5 mm i.d. and the required length (200 cm)] and washed with cold phosphate buffer solution to remove unlinked enzyme. Results and Discussion Immobilization of Enzymes The concentration of immobilized protein was above 99% for MDH, whereas with CL only 60% was achieved. The reactor of immobilized MDH could be used daily for at least 3 months without a noticeable loss of activity, if stored at 4 "C and in phosphate buffer solution after use. The CL reactor was found to be very unstable and, according to Fig.3, its activity decreased exponentially with time in such a way that the initial value was reduced to 50% after 60 min and to 70% after 180 min. Hence, the CL reactor was not useful from the analytical point of view, and it was necessary to use the CL enzyme in solution. 0.08 0.06 0 co 0 a 0.04 0.02 0 100 t h i n 200 Fig. 3 Loss of activity of immobilized CL reactor versus timcANALYST, SEPTEMBER 1993, VOL. 118 0.08 0.07 2 0.06 0.05 a, m 2 2 0.04 0.03 0.02 Determination of Oxalacetic Acid The optimization of the chemical and FI variables affecting the system was performed by a univariate method; the values taken as optimum are listed in Table 1.A study of the pH of the phosphate buffer showed a constant analytical signal in the range 6.4-8.1, above which it significantly decreased (Fig. 4). As regards the buffer concentration, it did not exert any influence on the analytical signal in the range studied (0.005-0.45 mol dm-3). However, the concentration of NADH was very important, as the absorbance increased up to a value of 260 mg dm-3, after which it remained constant. Flow rate ( q l = q2) had no influence on the analytical signal over the range 0.25-1.00 cm3 min-1. Lower values gave rise to higher analytical signals, at the cost of an excessive time of analysis. The analytical signal increased with increasing injected volume, although this meant a higher sample consumption and reaction time.A volume of 130 mm3 was chosen as a compromise. The analytical signal increased with MDH reactor lengths up to 6.0 cm, after which the signal remained constant. I - - - - - d ! I Table 1 Optimum values of the variables Variable Oxalacetic acid* [NADH]/mg dm-3 [KH2P04]/mol dm-3 [Glycylglycine]/mol dm-3 [CL]/U cm-3 [ZnC12]/mmol dm-3 Buffer pH q,/crn' min-1 q2/cm3 min-1 r,/cm MDH reactor/cm Sample volume/mm3 * Sec Fig. 1. 1 See Fig. 2. * 1.5 mm i.d. 260 0.1 - - - 7.8 0.50 0.50 6+ - 130 Citric acid1 510 - 0.1 0.67 0.5 7.8 0.28 - 200 70 6+ 0.26 0.24 a f m 9 0.22 s a 0.20 I] n.18 1 I I I _. . _ 6 7 8 9 1 0 1 1 PH Fig. 4 of oxalacetic acid Influence of pH on the analytical signal in the determination 0.6 0.5 8 0.4 9 0.3 m 2 2 0.2 0.1 0 10 20 30 40 50 Oxalacetic acid concentration/mg dm-3 Fig.5 Calibration graph for determination of oxalacetic acid 1195 Under these working conditions, the calibration graph (Fig. 5 ) between 1.0 and 40.0 mg dm-3 of oxalacetic acid was a straight line given by absorbance decrease = 0.0043 + 0.01419 [oxalacetic acid]; correlation coefficient (r2) = 0.9986. The relative standard deviation (RSD) was 1.4% for 12 determina- tions of 15.5 mg dm-3 oxalacetic acid. The sampling frequency was 40 h-1. This method was applied to the determination of oxalacetic acid in several types of fruit juice. The results obtained are presented in Table 2. The recovery was studied by adding two different amounts (10.0 and 20.0 mg dm-3) of oxalacetic acid to the same pre-treated fruit juices to obtain a mean recovery of 96.8% (range 91.8-103.9%).Determination of Citric Acid The initial purpose was to use immobilized CL and MDH for the determination of citric acid with these aims: ( a ) to save time and effort in the preparation of the reagents; ( b ) to re-use the same enzyme many times, thereby decreasing the cost per analysis; (c) to achieve a higher reproducibility in the results by always using the same concentration of enzyme; and (d) to simplify the FI configuration with the subsequent enhance- ment of sensitivity (lower dilution of the sample plug).l4 However, owing to its low stability (Fig. 3), the CL reactor was not useful and it was necessary to use the CL in solution. Although in FI systems it is common to inject the sample into a reagent stream, in this instance a reverse-FI configura- tion15 has been used, as CL is rather expensive.In this configuration, the reagent (enzyme) is injected into the sample stream (Fig. 2). The variables studied and the optimum values selected are summarized in Table 1. The influence of pH is small, a pH of 7.8 yielding the highest signal (Fig. 6). All buffers of this pH are not suitable, as Zn2+ existing in the reagent solution could precipitate with the P043- that occurs in many natural samples. Glycylglycine is used in order to overcome this problem. Table 2 Study of recovery of added amounts of oxalacetic acid in spiked samples Content of oxalacetic acid/ Sample mg dm-3 Apple 1 50.2 Applc 2 2.3 Pineapple 53.3 Orange 100.2 Amount added/ mg dm-3 10.00 20.00 10.00 20.00 10.00 20.00 10.00 20.00 Amount recovered/ mg dm-3 9.85 18.54 10.35 20.78 9.29 18.36 9.56 19.06 Recovery 98.5 92.7 103.5 103.9 92.9 91.8 95.6 95.3 ("/I1196 ANALYST, SEPTEMBER 1993, VOL.118 0.3 g 0.2 2 % 2 0.1 m I I I I 0 10 20 30 Citric acid concentration/mg dm-3 Fig. 7 Calibration graph for determination of citric acid Table 3 Study of recovery of added amounts of citric acid in spiked samples Content of Sample citric acid Chewing gum 1.23% m/m Caramel 1.77% m/m Asparagus* 0.82 g dm-3 Artichoke* 6.79 g dm--3 Orange (juice) 10.96 g dm-3 Lemont 8.50 g dm-3 Sweet lime? 3.20 g dm--3 * Canned. t Soft drink. Amount added/ mg dm-3 5.00 10.00 5.00 10.00 5.00 10.00 5.00 10.00 5.00 10.00 5.00 10.00 5.00 10.00 Amount recoveredl mg dm-3 5.13 9.75 5.12 10.08 5.05 9.63 4.96 10.37 5.33 10.24 5.10 10.50 4.88 9.60 Recovery 102.6 97.5 102.4 100.8 101.0 96.3 99.2 103.7 106.6 102.4 102.0 105.0 97.6 96.0 (%) The concentration of NADH does not exert any influence on the analytical signal over the range 0.2-0.7 g dm-3; higher values saturate the detector. On the other hand, the analytical signal increases with the concentration of CL; the maximum signal is reached with 1.4 U cm-3, as the amount of NADH considered as optimum has already been consumed.As for the presence of ZnClz (very important to avoid the loss of activity of CLZ), the analytical signal decreases when its concentration increases. This effect can be caused by the complex formed between Zn2+ and oxalacetate, as Zn2+ hinders the reaction with MDH,2 but it is not significant in concentrations up to 0.6 mmol dm-3.As regards FI variables, the analytical signal dramatically increases with decreasing flow rates. A value of 0.28 cm3 min-1 was chosen to avoid high residence times. On the other hand, the signal increases linearly with the injected volume until the injection of the blank produces saturation of the detector. The maximum analytical signal is obtained with a 300 cm reactor (rl). This reflects the fact that the reaction catalysed by CL needs a certain time to proceed (10 min are required to attain the maximum signal). However, a value of 200 cm was chosen because it provides a good analytical response and a significant reduction in the analysis time. The length of the MDH reactor exerts a similar influence to that observed in the method for the determination of oxalacetic acid.Under these working conditions, the calibration graph (Fig. 7) was linear over the range 1.0-20.0 mg dm-3 of citrate (absorbance = 0.0190 + 0.0111 [citric acid]; r2 = 0.9973). The RSD was 2.2% (for 11 determinations of 10.0 mg dm-3), and a sampling frequency of about 20 h-1 was achieved. Table 4 Comparative study of the use of conventional and FI methods for the determination of citric acid Parameter RSD (%) Linear range ratio* Detection limit CL consumption per MDH consumption per Mean recovery (%) Sampling frequency/h- 1 Blank? Number of enzymes used (as A = O.Ol)/mg dm-3 anal ysis/U anal ysis/U Conventional Proposed method FI method 1.5-2.0 20 4.8 0.8 12 100.7 -4 Yes 3 2.2 20 1 .o 0.4 <o. 1 100.9 =20 No 2 * Ratio of the upper to the lower limit of the calibration graph. t Additional measurement in order to subtract the absorbance of the matrix at the monitoring wavelength.Table 3 shows the results obtained using the procedure previously described for the determination of citric acid in different food samples, where the citric acid is either present naturally or added. The recoveries achieved by adding 5.0 and 10.0 mg dm-3 citric acid to the assayed samples was in the range 96.0- 106.5%; the mean recovery was 100.9%. Conclusions A fast and simple method for the determination of oxalacetic acid in fruit juices has been established; it avoids the need for using sophisticated instrumentation and a greater number of enzymes, as well as the complexity of bioluminescence methods.10 It also combines the specificity of the enzymic methods with the rapidity of the FI technique, whereas the existing chromatographic methods are much slower, with more expensive instrumentation, a longer stabilization time for the system, a more complicated sample pre-treatment and poor peak resolution.In all the methods considered, the detection limits are similar. Finally, a comparative study between conventional2 and proposed FI enzymic methods for citric acid determination (Table 4) leads to the following conclusions. The detection limit, defined as the concentration corresponding to a variation of the absorbance of 0.01, is lower in the FI method. The proposed method is valid for most samples as their content of oxalacetic acid is generally negligible when compared with that of citric acid; the conventional method is applicable to all types of samples.The number of enzymes involved has been reduced, as well as their consumption. Sampling frequency is five times higher in the FI method. In this method, the carrier is the sample and, therefore, no blank assay is required as the baseline coincides with the absorbance of the matrix. The chromatographic methods proposed for the determina- tion of citric acid show the same drawbacks as those mentioned above, together with a poorer detection limit. The authors gratefully acknowledge financial support from the Comision Interministerial de Ciencia y Tecnologia, Project ALI 90-0633. References 1 Lehninger, A. L., Curso Breve de Bioquimica, Ediciones Omega, Barcelona, 1979, pp. 214-223.ANALYST, SEPTEMBER 1993, VOL. 118 1197 2 3 4 5 6 7 8 9 Moellering, H., in Methods of Enzymatic Analysis, ed. Bergmeyer, H. V., Verlag Chemie, Weinheim, 1983, pp. 2-12, Ribereau-Gayon, J., Peynaud, E., Sudraud, P., and RibCreau- Gayon, P., Trait6 d’Enologie. Sciences et Techniques du Vin, Dunod, Paris, 1976, vol. I, pp. 182-284. Matsumoto, K., Ishida, K., Nomura, T., and Osajima, Y., Agric. Biol. Chem., 1984, 48, 2211. Pesez, M., and Bartos, J., Colorimetric and Fluorimetric Analysis of Organic Compounds and Drugs, Marcel Dekker, New York, 1974, pp. 62,295,298. Dunemann, L., Anal. Chim. Acta, 1989, 221, 19. Saag, K., in HPLCin Food Analysis, ed. Macrae, R., Academic Press, London, 1982, p. 230. Moellering, H., and Gruber, W., Anal. Biochem., 1966, 17, 369. Kurkijarvi, K., Heinonen, T., Lovgren, T., Lavi, J., and Rannio, R., Anal. Appl. Biolumin. Chemilumin. (Proc. Int. Symp.) 3rd 1984, p. 125. 10 11 Kurkijarvi, K., Vierijoki, T., and Korpela, T., Ann. N. Y. Acad. Sci., 1990, 585, 394. Valcarcel, M., and Luque de Castro, M. D., Flow Injection Analysis. Principles and Applications, Ellis Horwood, Chichester, 1987, p. 39. 12 Puchades, R., Herrero, M. A., Maquieira, A., and Atienza, J., Food Chem., 1991,42, 167. 13 Masoom, M., and Townshend, A., Anal. Chim. Acta, 1984, 166, 111. 14 Lazaro, F., Luque de Castro, M. D., and Valcarcel, M., Anal. Chim. Acta, 1988,214, 217. 15 Lazaro, F., Luque de Castro, M. D., and Valcarcel, M., Fresenius’ 2. Anal. Chem., 1985, 321, 467. Paper 2105890A Received November 4, 1992 Accepted March 1, 1993

ISSN:0003-2654    

DOI:10.1039/AN9931801193

出版商:RSC    

年代:1993

数据来源: RSC

摘要:

ANALYST, SEPTEMBER 1993. VOL. 118 1199 Flow Injection Spectrophotometric Method for the Speciation of Aluminium in River and Tap Waters M a Jose Quintela, Mercedes Gallego and Miguel Valcarcel" Department of Analytical Chemistry, Faculty of Sciences, University of Cordoba, E- 74004 Cordoba, Spain An automatic flow injection method for aluminium speciation in waters at the nanogram per millilitre level is reported, based on a Pyrocatechol Violet chelation-ion-exchange method. Data are presented on total reactive aluminium, total monomeric aluminium and non-labile monomeric aiuminium. The proposed method is sensitive (detection limit, 5 pg I-'), rapid and reproducible. The results obtained in the analysis of six river and tap water samples were consistent with those provided by the manual counterpart of the method.Keywords: Aluminium; speciation; flow injection; water; spectrophotometry Aluminium is the third most abundant element in the Earth's crust. Concentration of dissolved aluminium in most neutral waters are low because of the relatively low solubility of natural aluminium minerals (e.g., feldspars in metamorphic and igneous rocks). Several authorsl.2 have reported average aluminium concentrations of 10 pg 1-1 for the hydrosphere and 240 pg 1-1 for freshwaters. There is growing concern over the environmental consequences of the geochemical mobility of aluminium in soils and aquatic systems, where its presence has increased considerably as a result of atmospheric precipitation in the form of acid rain.3 Aluminium chemistry in acid waters is receiving increasing attention owing to the toxic effects of this element on aquatic and terrestrial organisms.Recent studies suggest a possible correlation between high concentra- tions of aluminium, higher than 50 pg l-1,4 in drinking waters with the incidence of senile dementia.5 The toxicity of aluminium is dependent on its speciation chemistry in water. Thus, Driscoll and co-workers6.7 have shown positively charged aluminium hydroxy species to be much more toxic to fish than organic aluminium complexes. This has fostered the development of aluminium speciation methods.8 Bertsch and Anderson9 classify aluminium speciation methods into three groups: (u) computation that uses thermodynamic-based geochemical models; ( b ) analytical separation of aluminium species by differential reaction kinetics with chelating agents and/or the physicochemical separation of aluminium species by size or charge; and (c) combination of one or more analytical separation techniques and a geochemical speciation model.Driscoll7 distinguishes three forms of aluminium based on an 8-hydroxyquinoline colorimetric method10 combined with ion-exchange separation and liquid-liquid extraction. This allows the determination of acid-soluble aluminium, non-labile monomeric aluminium and labile monomeric alu- minium. More recently, complete aluminium speciation in lake water with high suspended solids content (>20 mg 1-1) from Llyn Brianne reservoir (14 km north of Llandovery, Dyfed, UK), by using a colorimetric procedure based on Pyrocatechol Violet (PCV) was accomplishedll including a study of the influence of sorption processes. Another alumi- nium speciation method12 is based on a parabolic model that assumes the formation of three complexes between the metal and Morin.The speciation of aluminium in natural waters9 and forest soil13 using ion chromatography has been addressed by using post-column derivatization with Tiron9 and 8-hydroxyquinoline-5-sulfonic acid13 for identification of flu- oro-, oxalato-, and citrato-aluminium complexes, and by employing a highly sensitive and selective short column for the direct determination of aluminium species of the general formula AI(OH),(~-X)+, as well as fluoro species.14 * To whom correspondence should be addressed. Two automatic methods for fractionation and determina- tion of aluminium species in natural waters use Technicon AutoAnalyzers.One method15 is based on the ion-exchange procedure developed by Driscoll et aZ.6 and identifies total monomeric aluminium and non-labile monomeric aluminium. The other method16 provides fractionated inorganic monomeric and organic monomeric aluminium only. Neither method was investigated for the effect of chemical and flow variables. Finally, several flow-injection (FI) methods for the determination of total aluminium in waters,17 dialysis concen- trates18.19 and soil extracts,20 or monomeric aluminium species in waters21 have been developed using different detectors. In this work we automated a PCV method, an adaptation of the original Dougan and Wilson22 procedure.The proposed method allows three aluminium fractions to be determined using an FI system, which includes an Amberlite IR-120 ion-exchange mini-column. The combined use of FI and ul traviole t-visible (UV-VIS) spectropho tome try for the speciation of aluminium in river and tap waters yields reproducible results with lower risks of contamination or analyte losses, and in a shorter analysis time, relative to the manual method. Experimental Instruments and Apparatus A GBC UV-VIS Model 911 spectrophotometer connected to a Knauer recorder and equipped with a Hellma flow-through cell (10 mm light path, 1 mm i.d., 18 pl) was used. A Perkin-Elmer Model 1100-B atomic absorption spectrometer furnished with an HGA-700 graphite furnace and an AS-70 autosampler, and equipped with an aluminium hollow cathode lamp for the determination of total aluminium was also used.The wavelength was set to 309.3 nm and the spectral slit-width to 0.7 nm. Pyrolytic graphite coated graphite tubes were employed. The graphite furnace temperature pro- gramme is shown in Table 1. A Gilson Minipuls-2 peristaltic pump fitted with poly(viny1 chloride) (PVC) and Solvaflex (for buffer solutions) pump tubing was utilized. The injector consisted of a Rheodyne 5041 four-way valve to which a loop of the required volume was fitted; poly(tetrafluoroethy1ene) tubing (0.5 mm i.d.) was also used. A laboratory-constructed PVC column (10 cm x 2.5 mm i.d.) was employed for ion-exchange separations. A FIAtron 721 flow-cell accommo- dating a glass-calomel microelectrode connected to a Rad- iometer Model 62 pH meter was used for pH measurement.Reagents A 5 x 10-3 moll-' stock PCV solution was prepared in water. The iron-masking reagent was prepared from 2.55 X 10-21200 ANALYST, SEPTEMBER 1993, VOL. 118 AlJpH 3.5) Table 1 Graphite furnace temperature programme used for the determination of aluminium in some river and tap waters Time/s Argon flow rate/ Step TemperaturePC Ramp Hold ml min-1 1 100 2 800 3 1250 4 2500 5 2650 lnstrurnental conditions: Light source Lamp current Wavelength Spectral bandwidth Signal processing paramcter Integrating time Injected volume 25 20 300 15 15 300 15 15 300 0 4 0 (read) 1 1 300 Hollow cathode lamp 25 mA 309.3 nm 0.7 nm Peak area mode 5 s 20 1.11 sample rnol 1-1 (0.6% m/v) l,l0-phenanthrolinium chloride (mono- hydrate) and 4.3 rnol 1-1 (30% m/v) hydroxylammonium chloride in water.Two 1.5 mol 1-1 (20% m/v) hexamethyl- enetetramine buffers in water (pH 8.2) or ammonia solution (pH 8.8) were used. All reagents were purchased from Merck and remained stable for at least two weeks at room tempera- ture. A 1000 mg 1-1 aluminium solution was prepared by dissolving 1.000 g of aluminium wire in 20 ml of concentrated H2S04 and 50 ml of concentrated HN03, and subsequently diluting to 1 I with water. High-purity (ultrapure) water (Milli-Q) was used in all instances. For continuous introduc- tion of the PCV and iron-masking reagents into the FI system, both the stock solutions were diluted 10-fold with water. All other reagents used were of analytical-reagent grade.The Amberlite IR-120 Plus (wet mesh 16-50) strong cation- exchange resin (Sigma) was washed twice with water and then with 1 X 10-3 rnol dm-3 sodium chloride until the supernatant was clear. Materials and Cleaning All sample collections, storage and preparation of sample and reagent solutions were carried out in polyethylene containers (Aldrich), which are widely used to store environmental and trace component samples. All vessels were decontaminated by soaking in 10% HN03, for 48 h followed by rinsing five times with ultrapure water and filling with ultrapure water until used .23 Sample Collection and Preparation The river and tap water samples were collected into poly- ethylene containers and stored at room temperature in the dark. The pH of the river and tap water at the time of collection was 6.7 and 6.5, respectively. The total aluminium content in the untreated river water (sample A) and the tap water (sample D) was about 20 and 40 pg 1-1, respectively, as determined by electrothermal atomization atomic absorption spectrometry (ETAAS) using the method proposed by Sanz-Medel et al.24 The aluminium content in these natural waters was increased in order to make it adequate for quantitative speciation. Thus, we used six samples: unspiked river water (sample A), tap water (sample D), river water sample spiked with aluminium (as nitrate salt) to 70 (sample B) and 210 pg 1-1 (sample C) and tap water sample spiked with aluminium to 220 (sample E) and 600 pg 1-1 (sample F), respectively.The six water samples were analysed by ETAAS and their total aluminium concentration was found to remain virtually constant for up to 30 d, after which it decreased to a different extent depending on the type of water concerned.mI 7 min-1 i R 0.4 PCV 1.0 B 0.8 U Fig. 1 Schematic representation of the FI system used for the speciation of aluminium in waters. R, iron-masking reagent (2.55 mmol 1-1 1,lO-phenanthroline and 0.43 rnol 1-1 hydroxylammonium chloride); PCV, 5 x 10-4 mol 1-1 Pyrocatechol Violet; B, 1.5 moll-' hexamethylenetetramine buffer (pH 8.2 or 8.8); Al, total monomeric aluminium; AI, total reactive aluminium; Al,, non-labile monomeric aluminium; resin, Amberlite IR-120 P; B and S, blank and sample signals, respectively Conditioning of the Ion-exchange Column The Amberlite IR-220 P cation-exchange resin was used to determine the concentration of non-labile monomeric forms of aluminium in the water samples.The resin, in its sodium form, was packed in a 10 cm X 2.5 mm i.d. PVC column, washed with 3 ml of 1 moll-' HCl, then with 1 rnol 1-1 NaCl until the pH of effluent was 5.5, and finally conditioned by passing through a NaCl solution with the same conductivity and pH as the sample to be treated until the pH of the effluent was equal (k0.2) to that of the sample. The conductivity (pS cm-I), pH of the sample and the concentration of the NaCl ( ~ 1 0 - 3 rnol 1-1) used to achieve the same in the effluent, were: sample A , 203, 7.5, 1.55; sample B, 207, 7.9, 1.55; sample C, 239, 6.7, 1.80; sample D, 430,7.2, 3.40; sample E, 430, 7.8, 3.40; and sample F, 469, 7.1, 3.7, respectively.The column was reconditioned each time it was used. Procedures Manual Driscoll Method Total reactive aluminium (Air). A volume of 3.5 ml of water sample was acidifed to pH 1.0 by adding 50 p1 of concentrated HCl. After 1 h, the sample was analysed by the PCV method (see below). Total monomeric aluminium ( A f J . No sample pre-treat- ment before PCV detection of aluminium was required. However, 50 pl of concentrated HCI must be added simul- taneously with the hexamethylenetetramine buffer solution prior to analysis by the PCV method (see below). Non-labile monomeric aluminium (Al,). The sample was aspirated through the conditioned column at a flow rate of 4.0 ml min-1, the first 10 ml of sample was discarded, after which, the following 3.5 ml was collected and treated as for total monomeric aluminium.PCV method To a 3.5 ml volume of sample, whether standard or blank [acidified to pH 1 for 1 h (for Al, only)] 0.1 ml of the iron- masking reagent (0.1% 1,10 phenanthroline and 10% hydroxylammonium chloride) and 0.2 ml of a 0.0375% m/v PCV solution were added and mixed, then 1.0 ml of buffer solution (30% hexamethylenetetramine and 1.65% ammonia solution) and 50 pl of concentrated HCl were added (for the Al, and Al, fractions). After 10 min, the absorbance was measured at 580 nm. The linear determination range of the method was found to be between 5 and 400 pg I-'.ANALYST, SEPTEMBER 1993, VOL. 118 1201 Table 2 Fractionation of aqueous aluminium Total reactive aluminium (Al,) J \ \ Total monomeric aluminium (Al,) cation- exchange treated monomeric Acid soluble aluminium (Ale). aluminium (Alr0).aluminium (Ali). (monomeric alumino- (free aluminium, (colloidal polymeric organic complexes) monomeric aluminium; strong aluminium sulfate; alumino-organic fluoride, and complexes) hydroxide complexes) Automatic Driscoll Method Total reactive aluminium (Air). A 10 ml volume of water sample was acidified to pH 1.0 by adding 150 yl of concen- trated HCI. After 1 h, the sample was injected into the system shown in Fig. 1. Total monomeric aluminium ( A l J . The water sample was rapidly injected into the FI system (Fig. 1) after adjusting the pH to 3.5. Non-labile monomeric aluminium (Al,). The column was positioned vertically and the water sample was aspirated at a flow rate of 4.0 ml min-1 through it.The first 10 ml of sample was discarded; and the following 10 ml was collected and injected into the system (Fig. 1) after adjusting the pH to 3.5. Flo w-injection Method The proposed FI configuration for aluminium speciation is shown in Fig. 1. A 280 p1 volume of sample containing between 10 and 1000 yg 1-1 of aluminium was injected into a carrier solution including the iron-masking reagent (R) after merging with the PCV reagent, and then with the hexamethy- lenetetramine buffer solution of pH 8.2 (for the Al, and Al,, fractions) or pH 8.8 (for Air). Complex formation took place in the 280 cm coil (at pH 6.1-6.2). The blue colour was measured at 580 nm against an ultrapure water blank or a 0.1 mol 1-1 HCI solution (for Al,.).Results and Discussion As stated above, in this work we automated the method originally developed by Driscoll7 for the speciation of alu- minium in waters, but used PCV instead of 8-hydroxyquinol- ine as a chelating reagent.22 A schematic representation in the form of a flow chart is given in Table 2. The aluminium fractions determined in water samples were defined as fo!lows: total reactive aluminium (Al,.), i.e., the aluminium that was detected colorimetrically after a sample was acidified with 0.1 rnol 1-1 HCl for 1 h. Total monomeric aluminium (Al,), i. e., the aluminium detected colorimetrically in unacidi- fied effluent from the ion-exchange column. Exchangeable or labile monomeric aluminium (Al',), was quantified as the difference between Al, and Al,.Also, the difference between total reactive aluminium and total monomeric aluminium was called acid-soluble aluminium (Ali) and is an estimate of the aluminium that requires acid dissolution (pH 1.0) for its determination; this fraction would include colloidal alumi- nium and very strongly bound organic aluminium forms. If polymeric aluminium were not adsorbed onto the cation- exchange resin, then polymers might interfere with the reaction by preventing acidification of the effluent from the 100 300 500 100 300 500 r' Injected volume/$ Coil length/crn m 2 (TJ a 0 0.2 0.4 0.6 0.8 1 .o 1.2 PCV flow rate/rnl rnin-1 Fig. 2 Influence of the (u) injected volume, ( b ) coil length and (c) flow rate on the peak height. Sample: 0.5 pg ml-1 AP+; 1 cm = absorbance of 0.045 ion-exchange column to pH 1.Also, polymers are positively charged so they can adsorb, just as inorganic monomeric aluminium does. Consequently, the non-labile monomeric Al fraction (Al,) is an estimate of monomeric aluminium that is organically complexed, while the labile monomeric aluminium fraction (Al',) includes free monomeric aluminium and aluminium sulfate, fluoride and hydroxide complexes accord- ing to Driscoll.7 Optimization of the Flow-injection Manifold First, the FI manifold for the determination of aluminium by the PCV method was optimized and then was fitted to a cation-exchange column for speciation. Thus, a standard sample containing 0.5 pg ml-1 of aluminium was injected into the system at a pH of 2.0-6.0 (the optimal pH was 2.5-4.1).A sample pH of 3.5 was chosen for further experiments. The blank signal remained constant over the pH range 2.6-6.0; therefore, ultrapure water could be used as the blank. The concentration of PCV was varied between 1.0 X 10-4 and 14.5 X moll-'. The sensitivity to aluminium increased with an increase in the concentration of PCV up to 5 x 10-4 moll-1, beyond which the signal remained more or less constant. Obviously, a high molar excess of PCV over aluminium was desirable in order to minimize interferences from anions which form complexes with aluminium. However, the blank signal and background absorption increased in parallel with the PCV concentration, so 5 x rnol 1-1 (0.01875% m/v) was chosen to minimize changes in the background signal. Hexamethylenetetramine is the most commonly used buffer for the determination of aluminium by this method.The absorbance difference was constant over the range 1-2.5 mol 1-1 hexamethylenetetramine (pH 8.2), so a 1.5 mol 1-1 concentration was chosen. The flow variables studied were the sample volume, coil lengths and flow rates. The effect of the injected volume on the peak height and shape at a constant flow rate was studied over the range 100-500 yl. As can be seen in Fig. 2(a), above 280 p1, the absorbance difference remained virtually constant up to 400 pl, beyond which the signal decreased owing to the formation of three zones in the inserted sample plug: two at the ends, which merge with the reactants, and one in the middle, where no chemical reaction takes place. This was confirmed by the appearance of two FI peaks per injected sample volume exceeding 400 yl.A sample volume of 280 pl was chosen in order to ensure that blanks yielded low signals1202 ANALYST, SEPTEMBER 1993, VOL. 118 and hence increased reproducibility. Peak heights were scarcely affected by the length of the mixing coil used for mixing the sample and PCV, over the range 0-200 cm, because no reaction (only mixing) took place in it as the pH was still unsuitable (pH <4.0); therefore, no coil was used. As can be seen in Fig. 2(b), the coil length for the sample-PCV mixture and the hexamethylenetetramine buffer mixing was critical because it determined the optimal pH (6.1-6.2) required for complex formation; the length should not exceed 330 cm in order to avoid decreased signals.The influence of the flow rates of the PCV and hexamethylenetetramine solutions was investigated. Fig. 2(c) shows the dependence of the absor- bance on the flow rate of the PCV ( 5 X 10-4 moll-]) solution. The signal difference between sample and blank initially remained virtually constant and then increased slightly, after which it remained constant again up to 1.0 ml min-1. A PCV flow rate of 1.0 ml min-1 was chosen as optimua. The development of the aluminium-PCV complex depended on the flow rate of the hexamethylenetetramine buffer because the optimal pH in this coil must be maintained between 6.1 and 6.2. The optimal flow rate was between 0.4 and 0.8 ml min-1. At higher flow rates, the difference between the sample and blank signal was smaller owing to increased dispersion of the complex.Speciation of Aluminium In the Driscoll method,7 total reactive aluminium is deter- mined after the water sample has been acidified with 0.1 mol 1-1 HCI for 1 h. As shown above, the optimal pH for the sample ranged between 2.5 and 4.1 when the FI manifold was used, so the pH of the buffer solution must be higher in order to allow the most acidic sample (pH 1) to be buffered. Several hexamethylenetetramine buffers of different pH values were assayed for this purpose and a flow pH meter was included in front of the flow cell in the final flow system. Also, a standard sample containing 1.0 pg ml-1 of aluminium at pH 1.0 was injected into the system because the signal obtained with 0.5 pg ml-1 of aluminium was too low owing to dispersion of the complex in the mixing chamber of the pH meter.Finally, various 1.5 mol 1-1 hexamethylenetetramine buffers contain- ing different aliquots of ammonia solution (0-170 PI) were assayed. Complex formation was favoured at a pH of the sample-PCV mixture and buffer between 6.1 and 6.2 (ie., the pH obtained when standards of pH 2.5-4.1 and a hexamethy- lenetetramine buffer of pH 8.2 were used). Therefore, for samples of pH 1.0, the pH of the buffer solution must be increased to 8.8 in order to ensure a pH of 6.1-6.2 in the reaction coil. In summary, for the determination of Al,, a 1.5 moll-1 hexamethylenetetramine buffer of pH 8.8 in ammonia solution must be used; Al, was determined by directly injecting the water sample after adjusting the pH to 3.5 with 1 mol 1-1 HCI, into the FI system; Al, was determined in the unacidified effluent from the ion-exchange column.For this purpose, after an initial sample volume (10 ml, to displace the eluent) was discarded, 10 ml was collected in a vial and adjusted to pH 3.5 prior to analysis by the proposed FI method. The cation-exchange resin had a strong affinity for aluminium. As solutions passed through the cation-exchange column, there was competition between aqueous ligands and the strong cation-exchange resin for aluminium. The affinity of the resin for aluminium was studied by using a 0.5 pg ml-1 of aluminium solution (from aluminium wire dissolved in an acid), and another solution containing 0.5 pg ml-1 of aluminium plus a strong organic ligand (100 pg ml-l sodium citrate), both of which were passed through the cation- exchange column (at pH 6, similar to that of the water samples).In this experiment, detection of aluminium in the presence of citrate was complicated by the citrate competing with the PCV for aluminium and thus interfering with the determination of the metal. Citrate interferes with the proposed FI method at concentrations above 10 pg ml-I; however, because a concentration of 100 pg ml-1 resulted in half of the original sensitivity (slope of the calibration graph for aluminium) aluminium can still be determined in spite of such a decrease. Aluminium was completely removed from the solution of inorganic aluminium (the concentration of aluminium found in the effluent was lower than the detection limit, 5 pg 1 - 1 ) and transferred through the cation-exchange resin in the presence of citrate.Therefore, the resin effectively retained the inorganic aluminium and eluted the organic aluminium. Determination of Aluminium Two calibration graphs were obtained under the optimum conditions given in Fig. 1. One graph was linear over 10-80 pg 1 - 1 of aluminium and the second between 80 and 1000 pg 1 - 1 of aluminium. The corresponding equations are absorbance = 0.003 + 9.2 X x (x: 10-80 pg 1 - l of aluminium) absorbance = 0.054 + 2.3 X 10-4 x (x: 80-1000 pg 1-1 of aluminium) The slope of the calibration graph for 10-80 pg 1-1 of aluminium was higher than it was for 80-1000 pg 1-1 of aluminium owing to the presence of excess PCV (5 X 10-4 mol 1-1) relative to aluminium.The detection limit ( 5 pg 1-1 aluminium) was calculated as three times the absorbance for 30 injections of the ultrapure water blank by using the calibration graphs for 10-80 pg 1-1 aluminium. The precision of the automatic method obtained for 11 samples containing 50 pg 1-1 or 500 pg 1-1 (intermediate concentrations of the two linear ranges) expressed as relative standard deviation, was 1.8 and 0.65%, respectively. The sample throughput, exclud- ing the sample pre-treatment (i.e., acidification, ion exchange), was 40-50 h-1. The potential interference of iron was studied by injecting samples containing 0.5 pg ml-1 of aluminium and different concentrations of iron(I1i). The metal ion was found to interfere at the same concentration level as aluminium(in) when the iron-masking reagent carrier solution (Fig. 1) was replaced with a water carrier.Iron-masking reagent solutions containing 0.03, 0.06 or 0.1% 1,lO-phenanthroline plus 3% hydroxylammonium chloride suppressed the interference from iron up to 1.5, 5.5 and 9 pg ml-1, respectively. An iron-masking reagent solution containing 0.06% 1,10-phenan- throline (2.55 mmol l-1) and 3% hydroxylammonium chloride (0.43 moll-1) was chosen in order to avoid the interference of up to about 5.5 pg ml-1 of iron, which was sufficient for most waters. This iron-masking reagent also prevented interfer- ences from Cu2+, Ni2+, Zn2+, Pb2+ and Co*+ at concentra- tions of up to at least 15 pg ml-1. Analysis of Water Samples The proposed method was applied to the speciation of aluminium in freshwaters (samples A and D) as received, and those spiked with aluminium (samples B, C, E and F).The results obtained with the proposed FI method and the manual Driscoll method7 are given in Table 3. The total aluminium measured by ETAAS was higher than the total reactive aluminium provided by the PCV method as the determination of total aluminium by the PCV method would require pre-digestion of samples with peroxodisulfate, but pre-diges- tion is impossible for aluminium speciation. Conclusions The manual method is roughly three-times more sensitive than the proposed automatic FI method because of the increased sample dispersion in the flow system; however, the FI methodANALYST, SEPTEMBER 1993, VOL. 118 1203 Table 3 Concentration of aluminium (pg 1-1)* in freshwater samples by the proposed automatic (x) FI and manual (y) Driscoll methods Al'" St Ali X Y X Y X Y X Y A 20 19.6(5.3) 15.9(8.5) 16.5(5.2) 13.5(1.1) 5.4(4.8) 4.1 (1.1) 11.1 (0.4) 9.4 (0.0) B 70 65.2 (1.0) 59.7 (6.3) 45.6 (3.7) 42.3 (2.1) 28.3 (1.9) 14.1 (1.6) 17.3 (1.8) 28.2 (0.5) C 210 208.7(3.3) 139.0(1.3) 39.9(6.3) 33.6(7.9) 13.9(5.5) lO.O(l.1) 26.0(0.8) 23.6(6.9) D 40 35.2 (5.1) 29.7 (6.0) 25.6 (4.4) 22.3 (3.8) 18.3 (2.9) 14.1 (1 .l) 7.3 (1.5) 8.2 (2.7) E 220 204.3 (0.8) 196.0 (5.9) 127.1 (1.0) 99.1 (4.7) 49.9 (0.2) 45.4 (0.8) 77.2(0.8) 53.7 (3.9) F 600 597.7 (1.1) 507.7 (3.3) 507.3 (5.0) 467.3 (5.6) 78.5 (4.2) 45.5 (4.3) 428.8(0.8) 421.8 (1.3) * Relative standard deviation for five measurements of each water sample are given in brackets (%).t A , B and C are river waters; D, E and F are tap waters; samples B, C, E and F spiked with aluminium.-I Total aluminium measured by ETAAS. is more rapid and reproducible. As can be seen in Table 3, the manual method provides slightly lower concentrations because the reactants are placed directly in the sampling tubes, so losses by adsorption or precipitation on the sample tube walls are to be expected. Because of the similarities between the proposed method and that of Henshaw et al. ,*I only the most salient differences between the two are commented on. Thus, Henshaw et al. only determine the fraction of monomeric aluminium because they only analyse the percolate from the Amberlite IR-120 cation-exchange column in an FI system similar to that used in this work.Acidification of the sample to pH 1.0 or 3.5 prior to reaction with PCV allows the determination of total reactive aluminium (Al,) and total monomeric aluminium (Ala). In addition, conditioning of the ion-exchange resin permits the determination of non-labile monomeric aluminium (Al,) . Also, the total aluminium content is obtained by ETAAS. In summary, as can be seen from Table 2, the proposed method, like that of Driscoll, allows up to five aluminium fractions to be determined. The results obtained in the speciation of aluminium in waters by using the manual and automatic method are compared. Financial support from the BCR [BCR Contract No. 531 351/90/02-BCR-E( lo)] is gratefully acknowledged. References 1 Stumm, W., and Morgan, J . J . , Aquatic Chemistry, Wiley York, 1970.11/91 New 2 Bowen, H. J. M., Trace Elements in Biochemistry, Academic Press, New York, 1966. 3 Campbell, P. G. C., Bisson, M., Boagie, R., Tessier, A., Villeneuve, J. P., Anal. Chem., 1983, 55, 2246. 4 Standard Methods for the Examination of Water and Wastewater, American Public Health Association, American Water Works Association, Water Pollution Control Federation, Washington, DC, 17th Edition, 1989. 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 Martyn, C. N., Osmond, C., Edwardson, J . A., Barker, D. J. P., Harris, E. C., and Lacey, R. F., Lancet, 1989, January 14, p. 59. Driscoll, C. T., Baker, J. P. , Bisogni, J . J . , and Schofield, C. L., Nature (London), 1980, 284, 161. Driscoll, C. T., Int. J. Environ. Anal. Chem., 1984, 16, 267. Salbu, B., Environ. Geochem. Health, 1990, 12, 3. Bertsch, P. M., and Anderson, M. A., Anal. Chem., 1989,61, 535. Barnes, R. B., Chem. Geol., 1976, 15, 177. Goenaga, X., Bryant, R., Willians, D. J. A., Anal. Chem., 1987, 59, 2673. Browne, B. A., McColl, J . G., and Driscoll, C. T., J. Environ. Anal., 1990, 19, 65. Michalas, F., Glavac, V., and Parlar, H., Fresenius' J . Anal. Chem., 1992,343,308. Jones, P., Int. J. Environ. Anal. Chem., 1991, 44, 1. Rogeberg, E. J., and Henriksen, A., Vatten, 1985,41, 48. Lazerte, B. D., Chun, C., and Evans, D., Environ. Sci. Technol., 1988, 22, 1106. Royset, O., Anal. Chim. Acta, 1986, 185, 75. Pereiro Garcia, M. R., L6pez Garcia, A., Diaz Garcia, M. E., and Sanz-Medel, A., J. Anal. At. Spectrom., 1990, 5, 15. Fernandez, P., Perez Conde, C., Gutierrez, A., and Cgmara, C., Talanta, 1991, 38, 1387. Downard, A. J., Kipton, H., Powell, J., and Xu, S . , Anal. Chim. Acta., 1992, 256, 117. Henshaw, J. M., Lewis, T. E., and Heithmar, E. M., Int. J. Environ. Anal. Chem., 1988, 34, 119. Dougan, W. K., and Wilson, A. L., Analyst, 1974,99,413. Wilhelm, M., and Ohnesorge, F. K., J . Anal. Toxicol., 1990,14, 206. Sanz-Medel, A., Rodriguez Roza, R. M., Gonzalez Alonso, R., Noval, A., and Cannata, J. B., J . Anal. At. Spectrom., 1987, 2, 177. Paper 21064661 Received December 4, 1992 Accepted May 4, 1993

ISSN:0003-2654    

DOI:10.1039/AN9931801199

出版商:RSC    

年代:1993

数据来源: RSC

摘要:

ANALYST, SEPTEMBER 1993, VOL. 118 1205 Selective and Sensitive Spectrophotometric Determination of Tetrafluoroborate in Waste Water After lon-pair Extraction Using Bis[2-(5-chloro-2-pyridylazo)-5-diethylaminophenolato]cobalt(111) as a Counter Ion lssei Kasahara, Satoshi Hosokawa, Noriko Hata, Shigeru Taguchi and Katsumi Goto Department of Chemistry, Faculty of Science, To yama University, Go fuku 3190, To yama 930, Japan A simple, selective and sensitive method has been proposed for the ion-pair extraction and spectrophotometric determination of tetrafluoroborate in waste water. Tetrafluoroborate is extracted as an ion pair with bis[2-(5-chloro-2-pyridylazo)-5-diethylaminophenolato~cobalt(iii) into chlorobenzene. The absorbance of the organic phase is measured at 567 nm against a reagent blank.Beer's law is obeyed overthe concentration range 0.051 .O mg 1-1 of tetrafluoroborate and the detection limit (three times the standard deviation of the blank) was 0.009 mg 1-1. Relative standard deviations were 0.6-2.0%. Nitrate seriously interferes with the determination but can be decomposed by shaking with zinc powder after adjusting the acidity with 0.05 mol I-' sulfuric acid. The proposed method can be applied satisfactorily to the determination of tetrafluoroborate in industrial and laboratory waste waters containing high concentrations of fluoride, chloride, nitrate, sulfate, borate, silicate and many other metal ions. Keywords: Tetrafluoroborate in waste water; ion-pair extraction; spectrophotometry; bis[2- (5-ch yrid ylazo)-5-dieth ylamino p hen ola tolco ba Itliii) Fluorine compounds, both organic and inorganic, have their specific use in modern science and technology.Water that contains fluorine is discharged from metal plating, soldering, semiconductor and fine chemicals industries and laboratories. Fluoride ion is relatively easily removed from waste water by precipitation as calcium fluoride by adding calcium salts. If, however, fluorine is present as the tetrafluoroborate ion (BF4-), fluorine removal is more difficult. This is mainly owing to the inertness of the tetrafluoroborate ion. In the process of water quality control or fluorine removal treatment of such kinds of waste water, it is, therefore, necessary to determine fluoride and tetrafluoroborate ions separately. Fluoride and tetrafluoroborate ions can be determined by ion-chromatography (IC) , but difficulties arise in its applica- tion to the analysis of waste water, which contains large amounts of foreign ions and suspended solids, because of overlapping peaks and deterioration of the column.Methods of ion-pair extraction with spectrophotometric detection seem promising for the determination of tetrafluo- roborate in waste water. Several cationic dyes1-4 have already been proposed as counter ions for the determination of boron after converting the boron into the tetrafluoroborate ion. However, they are not suitable for waste water analysis because ion-pair extraction with these dyes is successful only within a narrow pH range and the reagent blank is usually high. During the course of investigation of ion-pair extraction systems,S it was found that complexes of cobalt(I1) with pyridylazoaminophenols are rapidly oxidized to their corre- sponding CO'~' complexes, some of which are inert, and some of which are negatively or positively charged and have excellent properties for use as counter ions in ion-pair extraction with spectrophotometric determination of ionic species in water. They are intensely coloured, some have molar absorption coefficients (E) exceeding 104 m2 mol-1.They have no acidic or basic groups, so that their ability for use as extraction reagents is independent of pH over a wide range. Exploiting these advantages these complex ions were applied to the ion-pair extraction and spectrophotometric determina- tion of anionic6 and cationic7 surfactants in water.In this paper, the bis[2-(5-chloro-2-pyridylazo)-5-diethyl- aminophenolato]cobalt(~rr) ion, (Co-5-Cl-PADAP)+ , is pro- posed for use as a counter ion for the ion-pair extraction and spectrophotometric determination of tetrafluoroborate. The proposed method is sensitive and reproducible and relatively free from interferences. Experimental Apparatus A Hitachi U-3200 double-beam automatic recording spectro- photometer with 10 mm quartz cells was used for measure- ment of absorbance. A Taiyo-kagaku Type SR-I1 reciprocal shaker was used for equilibrating the aqueous and organic phases. A Hitachi centrifuge Model 05P-21 was used for phase separation. Measurements of pH and fluoride concentration were carried out with a Horiba Model N-SF ion meter.A Hitachi Model 306 Super Scan ICP emission analysis system was used for the , determination of metals by inductively coupled plasma atomic emission spectrometry (ICP-AES). The operating parameters of the ICP spectrometer were as follows: forward power, 1.0 kW; plasma gas flow rate 16 1 min-1; auxiliary gas flow rate, 0.5 I min-1; nebulizer gas flow rate, 0.5 1 min-1; and nebulizer, concentric flow type. A Dionex Model 2000i ion chromatograph was used for the determination of anions. Reagents Bis [2-(5-chloro-2-pyridylazo)-5-diethylaminophenolato] cobalt(m) chloride (Co-5-Cl-PADAP)+Cl- solution. The (Co-5-CI-PADAP)+ C1- was purchased from Dojindo Labor- atory. This reagent can also be obtained from Fluka. A 0.05% m/v aqueous solution was prepared and the solution was washed three times with chlorobenzene to remove impurities.All other reagents and solvents were of analytical-reagent grade and were used as received. Tetrafluoroborate ( BF4-) standard stock solution (1000 mg 1-1). Dissolve 0.1265 g of sodium tetrafluoroborate in 100 ml of 0.05 mol 1-1 sulfuric acid containing 0.42 g of sodium fluoride. Recommended Procedure Take 20 ml of sample solution containing less than 20 pg of tetrafluoroborate into a poly(propy1ene) centrifuge tube with1206 ANALYST, SEPTEMBER 1993, VOL. 118 a screw cap. Add 2 ml of 0.5 moll-' sulfuric acid and 0.1 g of zinc powder. Shake the mixture for 5 min and centrifuge for 3 min (2500 rev min-1; 11OOg). Transfer a 10 ml portion of the supernatant into another centrifuge tube.Add 1 ml each of sodium sulfate solution and (Co-5-CI-PADAP)t- solution, and 10 ml of chlorobenzene. Shake for 3 min and centrifuge for 3 min (2500 rev min-1; 11OOg). Transfer the chlorobenzene layer into a cell and measure the absorbance at 567 nm against a reagent blank. Calculate the concentration from the calibration graph. Results and Discussion Structure and Stability of the (Co-5-Cl-PADAP)+ Fig. 1 shows the structure of (Co-5-CI-PADAP)+. Cobalt(I1) forms a 1 : 2 (metal : ligand) complex with 5-Cl-PADAP, and is then oxidized to CO(III) in the complex by dissolved oxygen. The CO(III) complex is an inert, singly charged cation, stable over a wide pH range (1-13); the absorption spectrum remains the same over this pH range. Its aqueous solution gives a high absorbance at 584 nm (E = 9.0 x 103 m2 mol-1).Extraction and Spectral Characteristics Various water-immiscible solvents were examined. Benzene, toluene and xylene did not extract the ion pair of (co-5-cI- PADAP)+ and tetrafluoroborate. o-Dichlorobenzene, chlo- roform and benzyl chloride extracted the complex cation even in the absence of tetrafluoroborate. Chlorobenzene was chosen as the solvent because the absorbance of the ion pair was highest and that of the reagent blank was lowest in this solvent. In addition, the solvent extracted the ion pair quantitatively. The absorption spectrum of (co-5-cI- PADAP)+ extracted into chlorobenzene as an ion pair with t Fig. 1 Structure of (Co-5-Cl-PADAP)+ A 1 I I - 450 500 550 600 650 Wave le ngt h/n m Fig. 2 Absorption spectra of the ion pairs extracted with (Co-5-Cl- PADAP)+ as counter ion: A, 0.5 mg 1-1 of BF4-; B, reagent blank.Conditions as for the proposed method with chlorobenzene as reference tetrafluoroborate is shown in Fig. 2 together with that of a reagent blank. The absorption spectrum has a peak at 567 nm and this wavelength was selected for the determination. Preservation of Stock Standard Solution The rate of hydrolysis of tetrafluoroborate in aqueous solution is very low.8>9 If, however, the standard solution of tetrafluo- roborate is prepared with distilled water only, the tetrafluoro- borate gradually hydrolyses and hence decreases in concentra- tion, as shown in Fig. 3. To prevent the hydrolysis of tetrafluoroborate and to maintain a constant concentration, sulfuric acid and sodium fluoride were added at concentra- tions of 0.05 and 0.1 mol 1-1, respectively.Comparison of the Results Obtained by the Proposed Method and by Ion Chromatography A solution of sodium tetrafluoroborate in distilled water was prepared and the changes in concentration of tetrafluoro- borate measured by the proposed method as well as by IC. Fig. 3 shows the relation between the results obtained by the proposed method and by IC. It can be seen in this figure that r L 0.6 I . 7 r" 0.5 F E 2 6 0.3 II 0 0.4 C 0 .- b 0.2 2 0.1 C 0 .- 4- C 0 C 0 " 0 0.1 0.2 0.3 0.4 0.5 0.6 Concentration by proposed method/mg I-' Fig. 3 Comparison of the results obtained by the proposed method and ion chromatographic method. 1, Just after preparation of standard solution; and after setting aside for: 2, 3 d; 3, 20 d; and 4, 30 d. A 1000 mg 1-l tetrafluoroborate solution was prepared by dissolving sodium tetrafluoroborate in distilled water.The solution was stored at room temperature and diluted 2000-fold with water for the analysis. Slope of line = 1.0 Table 1 Example of analyses of waste water samples (concentration given in mg 1-1) Determinand Sample A Sample B Sample C Sample D Element*- Aluminium Boron Barium Calcium Iron Magnesium Silicon Strontium Zinc 400 540 140 41 6.5 4.0 4.5 5 800 72 110 80 27 1.9 1.1 1.6 1400 130 350 2.3 1.6 1.5 0.8 4 500 190 - 4.5 90 - 7.7 0.4 5.1 24 - 6.0 Anion?- Fluoride 27000 20000 54000 160 Chloride 400 150 1 000 17 Nitrate 6 900 4 000 19 000 10 14 Sulfate * Measured by ICP-AES after proper dilution.t Measured by IC after proper dilution. - - -ANALYST, SEPTEMBER 1993, VOL. 118 1207 Table 2 Effect of diverse substances on the determination of BF4- (0.5 mg 1-I) Substance None - F- NaF c1- NaCl s o p Na2S04 NO3 - KN03 c103 - NaCIOJ DS-7 Na salt Amount added1 Added as mg 1-1 1000 1000 1000 1 100 0.1 0.1 - Na+ K+ Ca2+ Mn2+ Fez+ Fe3+ Ni2+ cu2+ AF+ Mg2+ * Without reduction procedure. i DS- = dodecyl sulfate. Set aside for 1 h after addition of aluminium. 650 63 1000 100 100 100 100 100 100 100 100 Absorbance at 567 nm 0.384 0.388 0.385 0.384 0.492 0.389 0.390 0.385 0.385 0.389 0.384 0.381 0.382 0.383 0.381 0.384 0.386 0.375 0.308 Error (Yo 1 - +1.0 +0.3 +0.0 +28* +1.3 +1.6 +0.3 +0.3 +1.3 +0.0 -0.8 -0.5 -0.3 -0.8 +o.o -0.3 -2.3 - 20.0j: 0.8 '.Oi 0.4 -&b ---& _________________ -&_ 0.2 0 r' 10 20 30 Time for reductiodmin Fig.4 Effect of time of reduction with different concentrations of nitrate. Conditions: 0.5 mg 1-1 of BF4-. The broken line represents the absorbance by the proposed method in the absence of nitrate. Nitrate concentration: A, 10; B, 100; and C, 500 mg I-' correlation between the two methods is excellent (correlation coefficient, 0.999), suggesting that only tetrafluoroborate is determined by the proposed method. Example of Determination of Fluoride in Waste Waters The results of analyses of different kinds of waste waters that contained fluoride and borate are shown in Table 1. Samples A and B were from a university laboratory for electronic materials. Samples C and D were waste waters discharged from metal plating and soldering factories, respectively.The table shows diversity in quality of fluoride containing waste waters. It is to be noted that in addition to fluoride, boron and silicon are present at relatively high concentrations. Table 3 Reproducibility at different BF4- levels. Number of measure- ments = 5 , sample volume = 20 ml and reference = reagent blank Concentration of BF4-/mg 1-1 0.05 0.10 0.30 0.50 0.70 1 .oo Absorbance at 567 nm sr* (Yo) 0.040 2.0 0.077 1 .o 0.245 1.2 0.388 1.0 0.548 0.6 0.783 0.7 * Relative standard deviation. Table 4 Analysis of several waste water samples and results of recovery test. Each sample was analysed after dilution of the original waste water sample shown in Table 1. A' = a 100-fold diluted solution of sample A.C' = a 1000-fold diluted solution of sample C. D' = a 200-fold diluted solution of sample D Proposed method Recovery of added BF4- Sample BF4- added/ BF4- found/ name mgl-l mg 1-1 mgl-* YO A' None 0.100 0.300 0.500 C' None 0.100 0.300 0.500 - - 0.387 0.481 0.094 94.0 0.695 0.308 103 0.884 0.497 99.4 0.259 0.357 0.098 98.0 0.572 0.313 1 04 0.755 0.496 99.2 - - - - D' None 0.450 0.100 0.554 0.104 104 0.300 0.747 0.297 99.0 0.500 0.953 0.503 101 Effect of Diverse Ions Table 2 shows the effects of diverse ions on the determination of tetrafluoroborate by the proposed method. Aluminium combines with the fluoride ion of the tetrafluoroborate at pH values of below 3; this is illustrated when aluminium salts are used to decompose tetrafluoroborate.Nearly all other cations have little effect. Fluoride, chloride and sulfate at high concentrations do not affect the determination. Nitrate ion has nearly the same extractability as the tetrafluoroborate ion, causing positive errors, but it can be decomposed by reduction with zinc powder; as will be described in the next section. The proposed method is very sensitive, so that original samples can be diluted for analysis by more than 100-fold in order to reduce interferences. Tetrafluoroborate seems to hydrolyse slowly to trifluorohydroxyborate,"." which, however, does not interfere with the determination of tetrafluoroborate. Removal of Nitrate Interference As shown in Table 2, nitrate seriously interferes with the determination of tetrafluoroborate by being similarly extrac-1208 ANALYST, SEPTEMBER 1993, VOL.118 Table 5 Application of proposed method to the analysis of waste water samples (concentrations given in mg 1-l) Calculated total fluoride Measured Sample* BF,- found (A) Fin BF4-(B) F-(C) (D) = (B) + (C) total fluoridet 1 33.1 29.0 92.4 121.4 122 2 30.8 27.0 1.8 28.8 28.2 3 31.7 27.8 8.7 36.5 39.0 4 11.3 9.9 4.9 14.8 17.3 5 5.0 4.4 7.7 12.1 13.4 * Sample size = 20 ml of a 100-fold diluted sample. t Total fluoride concentration was measured spectrophotometrically using lanthanum Alizarin Complexone after separation of fluoride by distillation. ted. As hydrofluoric acid is used as a mixture with nitric acid in etching and other operations, nitrate interference causes a serious problem in the analysis of waste waters.In order to attempt to decompose nitrate by reduction to other species some common reductants, such as copper powder, Devalda’s alloy, hydroxylammonium chloride, hydrazinium sulfate and zinc powder, were examined. It was found that zinc powder was the most satisfactory because of low reagent blanks obtained and its ease of handling. An IC study showed that none of the nitrate remains in the samples treated with zinc powder. Fig. 4 shows the effect of time of reduction on the determination of tetrafluoroborate at differ- ent concentrations of nitrate. It can be clearly seen that up to 100 mg 1-1 of nitrate does not interfere with the determina- tion. Nearly identical calibration graphs were obtained both with and without the reduction process. Precision and Detection Limit The precision of determination is shown in Table 3 and the detection limit (defined as the concentration corresponding to a three times the standard deviation of the reagent blank) was found to be 0.009 mg 1-1.The absorbance of the reagent blank for this method is about 0.02. Application to Waste Water Samples and Recovery Test The proposed method was applied to the analysis of various waste water samples. Table 4 shows the results of the analyses of the original samples and samples to which known amounts of tetrafluoroborate had been added. The recovery of the spiked tetrafluoroborate was almost quantitative. Tetrafluoroborate in waste water samples treated by different methods of fluoride removal was also determined by the proposed method.Determination of total fluoride and free fluoride ion was also carried out. Total fluoride was determined spectrophotometrically using lanthanum Alizarin Complexone after separation of the fluoride by distillation. Free fluoride was determined potentiometrically by using an ion-selective electrode. The results are summarized in Table 5. Sample 1 is the original waste water sample and the others are treated waters. The sum of the concentrations of free fluoride and that in the form of tetrafluoroborate agrees very well with the total fluoride concentration determined after separation of total fluoride by distillation. This result indicates that most of the fluorine in these samples is present as free fluoride and tetrafluoroborate. Other forms of fluoroborates (i. e., hydroxyfluoroborates), if any, must be hydrolysed rapidly.819 Conclusion The proposed method is applicable for the determination of tetrafluoroborate in waste water with good precision. References Ducret, L., Anal. Chim. Acta, 1957, 17, 213. Pasztor, L., Bode, J . D., and Fernando, Q., Anal. Chem., 1960, 32,277. Cizek, Z . , and Studlarova, V., Talanta, 1984, 31, 547. Mikasa, H., Shirouzu, M., and Hori, Y . , Bunseki Kagaku, 1991, 40, 749. Kasahara, I., Ohgaki, Y., Matsui, K., Kano, K., Taguchi, S . , and Goto, K., Nippon Kagaku Kaishi, 1986, 894. Taguchi, S . , Kasahara, I . , Fukushima, Y., and Goto, K., Talanta, 1981, 28, 616. Kasahara, I . , Kanai, M., Taniguchi, M., Kakeba, A., Hata, N., Taguchi, S . , and Goto, K., Anal. Chim. Acta, 1989, 219, 239. Wamser, C. A., J. Am. Chem. SOC., 1951, 73,409. Mesmer, R. E., Palen, K. M., and Baes, C. F., Znorg. Chem., 1973, 12, 89. Paper 3102488A Received April 30, 1993 Accepted May 20, 1993

ISSN:0003-2654    

DOI:10.1039/AN9931801205

出版商:RSC    

年代:1993

数据来源: RSC

摘要:

ANALYST, SEPTEMBER 1993, VOL. 118 1209 Spectrophotometric Determination of Lipohydroperoxides and Organic Hydroperoxides by Use of the Triiodide-Hexadecylpyridinium Chloride Micellar System M a Loreto Lunar, Soledad Rubio and Dolores Perez-Bendito Department of Analytical Chemistry, Faculty of Sciences, University of Cordoba, 14004 Cordoba, Spain The association complex formed between triiodide ion and hexadecylpyridinium chloride (cetylpyridinium chloride; CPC) was used to develop a spectrophotometric method for the determination of hydroperoxides, based on the ability of hydroperoxides to oxidize iodide ion to iodine in an acetic acid medium. The triiodide ion thereby produced associates with CPC cationic micelles, which results in maximum absorption at 500 nm, in addition to substantially increased absorptivity and stability constant for the triiodide complex.The micellar medium allows the determination of various hydroperoxides (hydrogen peroxide, cumene hydroperoxide and tert-butyl hydroperoxide) at concentrations between 5 x 10-7 and 2.5 X 10-6 moll-1, with a molar absorptivity for triiodide ion of (6.51 f 0.08) x 103 m2 mol-1 (Le., about three times higher than those typical of methods implemented in aqueous media). The proposed method was successfully applied to the determination of lipohydroperoxides in five commercially available oil samples (olive, sunflower seed, corn, cod liver and linseed) and of various organic hydroperoxides in commercial samples (the recovery of hydroperoxides from heptane ranged between 98 and 106%). The results obtained in the determination of lipohydroperoxides are consistent with those provided by iodimetric titration.Keywords: Spectrophotometry; micellar media; organic h ydroperoxide; lipoh ydroperoxide; oil Methods for the identification and determination of peroxides are used in a number of major industrial, environmental and clinical applications, yet few of them have proved suitable or reliable for dealing with peroxides at concentrations as low as 1 pmol1-1.1*2 The most sensitive methods for this purpose are based on fluorimetric detection. Hence, organic hydroperox- ides and lipohydroperoxides at concentrations as low as 8 nmol 1-1 have been determined by fluorimetric oxidation of dichlorofluorescein in the presence of haematin. The assay requires the dichlorofluorescein-hydroperoxide mixture to be incubated in an argon atmosphere for 50 min.3 Oxidation of iodide is the most widely used of all reactions in peroxide determinations.Various procedures involving iodide have been designed in order to overcome problems posed by earlier alternatives.2 Methods based on this reac- tion involve oxidation of potassium iodide by hydrogen peroxide or organic peroxides, in aqueous or organic medium, in the presence or absence of ammonium molyb- date as catalyst. Released iodine is determined by spectro- photometry as triiodide at 352 nm (molar absorptivity 2.3 X 103 m2 mol-1) or by titration with thiosulfate ion (this procedure is only sensitive to concentrations of approxi- mately 5 x 10-4 mol 1-1, as lower concentrations give rise to errors of up to 50%4).The primary objective of this work was to develop a quantitative spectrophotometric method for the determination of hydroperoxides at concentrations below 1 pmol 1 - 1 by using the oxidation of iodide in a micellar medium. It is well known that micellar systems can enhance existing analytical methods by considerably improving their sensitivity and selectivity.5 As regards the spectrophotometric determin- ation of triiodide in micellar media, a direct method for the determination of iodine in the range 1-10 p moll-1, based on the increase in the maximum absorbance observed at 360 nm on addition of cetyltrimethylammonium bromide to solutions containing iodine and potassium iodide, was developed.6 The local concentration of both reagents in the micellar pseudo- phase also increases the apparent equilibrium formation constant of triiodide ion by a factor of about 50 relative to the aqueous medium. Similar effects have been observed in the presence of other alkylammonium cationic surfactants such as dodecyltrimethylammonium chloride.7 In this context, it is worth noting the analytical use of the interaction between triiodide ion and hexadecylpyridinium chloride (cetylpyridi- nium chloride; CPC) micelles.8 Addition of CPC to aqueous triiodide solutions has been shown to result in a bathochromic shift from 350 nm (the absorption wavelength of the triiodide complex) to 500 nm (the maximum absorption wavelength of the 1,--CPC association complex) , in addition to a substantial increase in the absorptivity and stability constant of the triiodide complex.These effects can allow one to overcome completely or at least minimize the selectivity and sensitivity problems confronting many of the original spectrophoto- metric procedures that involve aqueous media for monitoring iodine. In this work, such effects were used for the determina- tion of hydrogen peroxide and organic hydroperoxides. Under the experimental conditions used, the molar absorptivity of the triiodide ion was found to be (6.51 & 0.08) x 103 m2 mol-1 (i.e., approximately 2.8 times that in water). The proposed method is considerably more sensitive than the recommended enzymic and non-enzymic spectrophotometric methods for the determination of hydroperoxides in the range 1-10 pmol 1-1.1 To the authors’ knowledge, only the spectrophotometric methods using the leuco base of Methylene Blue (E = 7720 k 390 m2 mol-1) or its N-benzoyl derivative (E = 15500 m2 mol-1) are more sensitive.9710 These methods, however, suffer from serious drawbacks, so many analytical chemists in need of reliable trace-level analyses for peroxides have turned away from them.Hence, preparing pure enough leuco Methylene Blue reagent is rather difficult; also, oxygen and steam cause the dilute reagent solution to decay during storage and must, therefore, be scrupulously removed. On the other hand, both methods are highly sensitive to light, which must be excluded during the determination; in addition, the reaction with peroxides is relatively slow2 (benzoyl peroxide takes about 30 h to react fully in the method based on the N-benzoyl derivative).The proposed method was succesfully applied to the determination of lipohydroperoxides in oil samples and of organic hydroperoxides in hydrocarbon samples, with a detection limit of about 2 x 10-7 mol 1-1 hydroperoxide.1210 ANALYST, SEPTEMBER 1993, VOL. 118 Experimental Apparatus Kinetic measurements were performed on a Hitachi (Tokyo, Japan) U-2000 spectrophotometer fitted with 1 cm pathlength cells. The spectrophotometer cell compartment was thermo- statically controlled by circulating water from a Neslab (Newington, NH, USA) Model RTE bath with a temperature stability of kO.1 "C throughout. Reagents All the chemicals used were of analytical-reagent grade and were utilized as purchased, without further purification.Distilled water was used throughout. Solutions (1.25 x 10-2 rnol 1-1) of hydrogen peroxide and tert-butyl hydroperoxide were prepared in water. A 1.25 X 10-2 rnol 1-1 solution of cumene hydroperoxide was prepared in ethanol. More dilute solutions (5 X 10-4 mol I-*) were prepared from these stock solutions before each set of experiments by appropriate dilution with doubly distilled water. A 4.1 rnol 1-1 potassium iodide solution was prepared and stored in a dark glass bottle. A 1.4 x 10-3 rnol 1-1 solution of CPC was prepared by dissolving the required amount of the surfactant in distilled water. Analytical-reagent grade glacial acetic acid and chloro- form were also used. General Procedure for Calibration With Hydrogen Peroxide A calibration graph was obtained as follows: to a 10 ml calibrated flask were added, in sequence, between 0.2 and 1 ml of 5 X 10-4 moll-' hydrogen peroxide, 1.5 ml of glacial acetic acid, 0.1 ml of 4.1 moll-1 potassium iodide and distilled water up to a final volume of 2.6 ml.The mixture was swirled gently for 1 min, allowed to stand in the dark for 5 rnin and diluted to the mark with distilled water. A 0.5 ml aliquot of this solution was placed in a I 0 ml calibrated flask. The triiodide content of each aliquot was determined as follows: 0.25 ml of 1.4 x 10-3 rnol 1-1 CPC was added, the stop-clock was started and the mixture was diluted to the mark with water. A portion of the reaction mixture was then transferred into a cell kept at 20 k 0.1 "C, and the absorbance of the solution at 500 nm was measured as a function of time until the maximum absorbance signal was attained.Measurements were started exactly 1 min after the CPC was added and the maximum absorbance signal was generally obtained within 4 min. A blank solution containing no hydrogen peroxide was prepared and analysed similarly for each series of samples, and the resulting absorbance was subtracted from that yielded by the samples. Procedure for the Determination of Lipohydroperoxides in Oil Samples An accurately weighed amount of the oil (0.02-0.5 g) was placed in a 25 ml beaker and to this sample were added, in sequence, 1 ml of chloroform, 1.5 ml of glacial acetic acid and 0.1 ml of 4.1 rnol 1-1 potassium iodide. The contents of the beaker were swirled gently for 1 rnin and allowed to stand in the dark for 5 min, after which, 8.4 ml of distilled water were added.Immediately afterwards between 0.2 and 1 ml of the aqueous phase was pipetted into a 10 ml calibrated flask, and the triiodide content of this solution was determined as described above for the determination of hydrogen peroxide. A similarly prepared blank solution containing no oil was also analysed and the resulting absorbance was subtracted from that yielded by the samples. The calibration graph established for hydrogen peroxide determination was used to determine lipohydroperoxides, which were expressed as 'active peroxide concentrations' (mequiv kg-1 of peroxide in oil). (The conversion factor of mmol to mequiv is 2.) Procedure for the Determination of Hydroperoxides in Hydrocarbon Samples A 10 ml hydrocarbon sample was extracted with two 2 ml portions of 1% sodium hydroxide in ethanol.If the hydrocar- bon and ethanolic phases proved to be miscible, the required phase was reclaimed by extraction with 1% sodium hydroxide in ethanol-water (1 + 2 or 1 + 3). The extracts were collected together in a 10 ml calibrated flask and diluted to the mark with distilled water. A solution volume of <1 ml was pipetted into a 10 ml calibrated flask and to this were added, in sequence, 1.5 ml of glacial acetic acid, 0.1 ml of 4.1 rnol 1 - 1 potassium iodide and distilled water up to a final volume of 2.6 ml. The contents of the flask were swirled gently for 1 min, allowed to stand in the dark for 15 rnin and diluted to the mark with distilled water.The triiodide content of a 0.2-1 ml aliquot of this solution was determined as described above for hydrogen peroxide. A blank solution containing the required concentration of ethanolic sodium hydroxide was prepared and analysed similarly and the resulting absorbance was subtracted from that yielded by the samples. The calibration graph established for hydrogen peroxide determination was used to determine hydroperoxides, which were expressed as 'active peroxide concentrations' (mmol 1- 1 of peroxide in the sample). Results and Discussion Study of the Hydroperoxide-Iodide System in the CPC Micellar Medium In order to establish the optimal experimental conditions for determining hydroperoxides by reaction with iodide in the presence of CPC micelles, hydrogen peroxide, cumene hydroperoxide and tert-butyl hydroperoxide were used as test materials.Acetic acid was chosen as the reaction medium as it is in this solvent that iodide exhibits its reducing power in full .2 Iodimetric procedures exhibit an essentially constant molar response for all reactive peroxide types, thereby allowing a general calibration graph to be obtained for any conveniently available known peroxide. Inasmuch as the reactivity of hydroperoxides is a function of their structure, quantitative reduction is crucial to obtaining a constant absolute response. Under the experimental conditions used to apply the proposed procedures, the reaction between hydrogen peroxide and iodide completed quantitatively in 5 rnin at room temperature, while cumene and tert-butyl hydroperoxide took 15 min.The surfactant (CPC) had no appreciably accelerating effect on these reactions, so it was only added to the reaction medium when triiodide ion was yielded quantitatively. This allowed the hydroperoxide reduction reactions to occur under condi- tions of higher acidity (a previous dilution was made before measurements because pH values below 2 resulted in rapid decomposition of the triiodide-CPC association complex) and hence more rapidly. Two steps were considered in optimizing the reaction conditions for the determination of hydroperoxides in the presence of CPC: production of triiodide and formation of the triiodide-CPC association complex. Fig. 1 shows the variation of the absorbance at 500 nm as a function of time when CPC was added to a solution containing triiodide ion.The time at which maximum absorbance was reached was between 2 and 4 rnin , depending on the hydroperoxide concentration. Production of triiodide ion in the reaction between hydroperoxides and iodide was essentially affected by the iodide and acetic acid concentrations and the temperature.' Fig. 2 shows the final absorbance obtained for a 2 x 10-6 rnol 1-1 hydrogen peroxide concentration as a function of the iodide (u) and acetic acid ( 6 ) concentration after a reaction time of 5 min. Quantitative reaction was only obtained at iodide and acetic acid concentrations above approximatelyANALYST, SEPTEMBER 1993, VOL. 118 I I 1 I _ 1211 0.14 and 9 rnol 1-1, respectively. Iodide and acetic acid concentrations of 0.16 and 10 moll-1 also assured quantitative reaction of cumene and tert-butyl hydroperoxide within 15 min.Higher concentrations of these reagents resulted in high blank signals and no further effect on the rate of formation of triiodide ion. The selected concentration of iodide ion (0.16 rnol 1-1) was high enough relative to that of iodine to avoid errors arising from volatilization of iodine and from addition of iodine to unsaturated materials.11 Increased temperatures raised the rate of both the hydroperoxide and blank reactions. No net gain in absorbance was achieved, though, as the reaction temperature was increased. Room temperature was, therefore, believed to be the most suitable. The experimental conditions under which the triiodide- CPC association complex shows maximum absorbance and stability are reported elsewhere.8 The iodide concentration must be at least three times higher than the iodine concentra- tion to cause a bathocromic shift in the triiodide spectrum.On the other hand, iodide ion forms an insoluble salt with cetylpyridinium with a solubility product of approximately 5.2 x 10-7 mo12 1-2. Under the experimental conditions described above for the determination of hydroperoxides, the iodide concentration in the solution prepared for measuring the absorbance of triiodide was 2 X 10-3 rnol 1-1. This concentration was high enough for the triiodide-CPC associa- tion complex to be formed and for precipitation of the iodide-cetylpyridinium salt to be avoided. The maximum absorbance signal at 500 nm for the triiodide-CPC association complex was observed at pH values between 2 and 10.Although triiodide was produced below pH 2 under the proposed experimental conditions, the solutions o-2 I I 1 3 5 Time/mi n Fig. 1 Absorbance versus time plot for the triiodide-CPC associa- tion. [H202] = 1 x 10-6 rnol I-'. (For experimental conditions, see text .) 0.2 la' 0.1 1 L I used to form the triiodide-CPC association complex had pH values of approximately 3 after pertinent dilution. Increased temperatures had an adverse effect on the absorbance at 500 nm of the triiodide-CPC association complex, which remained virtually constant between 10 and 20 "C. Fig. 3 shows the absorbance maximum obtained as a function of the CPC concentration. A break.point appeared at about 1.8 X 10-5 mol 1-1 CPC, above which the absorbance remained constant throughout the range examined.A 3.5 X 10-5 rnol 1-1 CPC concentration was chosen as optimal. Analytical Figures of Merit Absorbance versus peroxide concentration calibration graphs were established for hydrogen peroxide, cumene hydroperox- ide and tert-butyl hydroperoxide (Fig. 4). The experimental conditions used to determine the organic hydroperoxides were identical with those used for the determination of hydrogen peroxide, except that a time of 15 min was needed for the respective redox reactions to go to completion. The absor- bances for the different hydroperoxides fell roughly on the same line, so the same calibration plot was used for all the samples. Alternatively, a regression equation could be used.As hydrogen peroxide was the most reactive species, it was used as a calibration standard. The determination of these hydroperoxides was feasible over the linear range (5-25) X 10-7 rnol 1-1. The standard error of the estimate was 7 x 10-3 absorbance units and the correlation coefficient was 0.992 (n = 7). The molar absorptivity was (6.51 2 0.08) x 103 m2 mol-1. The detection limit (30) was 2 x 10-7 rnol 1-1 hydroperoxide. Finally, the precision of the proposed method, expressed as the relative standard deviation, was 2.3% ( n = 11) for a 1 X 10-6 rnol 1-1 hydroperoxide concentration. Determination of Lipohydroperoxides in Oil Samples The iodide method most widely used for the determination of peroxides in fats and oils involves use of potassium iodide in 0 2 4 6 8 [CPC]/10-5 mol 1-1 Fig.3 Influence of the concentration of CPC on the absorbance of the triiodide ion produced from the iodide-hydrogen peroxide system at a fixed time of 5 min. (For experimental conditions, see text.) 5 15 25 [Hydroperoxidel/lO-7 rnol 1-1 Fig. 4 Calibration gra hs for (0) hydrogen peroxide; (A) cumene hydroperoxide; and (Op tert-butyl hydroperoxide. Times of triiodide production: (0) 5 min; and ( A , 0) 15 min. (For experimental conditions, see text.)1212 ANALYST, SEPTEMBER 1993, VOL. 118 acetic acid-chloroform (3 + 2 or 2 + 1) at room temperature. The spectrophotometric determination of the triiodide pro- duced has a serious drawback: the high absorption of fat and oil matrices at 360 nm, where the triiodide species absorbs maximally.For this reason, titration of triiodide with thiosul- fate is the most commonly used procedure to determine the peroxide content in these types of sample.12 The batho- chromic shift that CPC causes in the maximum absorption of triiodide ion (A,,, = 500 nm) makes the iodimetric spectro- photometric method selective enough to permit reliable determinations of the peroxide content of fats and oils. In addition, the molar absorptivity for the absorption at 500 nm of the triiodide-CPC association complex (approximately 6510 m2 mol-1) makes it highly suitable for this application. The reactivity of the lipohydroperoxides towards iodide ion was found to be similar to that of hydrogen peroxide. Accordingly, the calibration graph for hydrogen peroxide determination was also used for the lipohydroperoxides, values being expressed as 'active peroxide concentrations' (mequiv kg-1 of peroxide in oil). Acetic acid was used instead of acetic acid-chloroform (3 + 2) to obtain triiodide from lipohydroperoxides in oils in preparing the calibration stan- dards both media provided equivalent results in the determi- nation of hydrogen peroxide.The use of a relatively large excess of iodide (which assured iodine was completely converted into triiodide) and a volume of aqueous phase ten times higher than that of the organic phase ensured virtually quantitative extraction of the iodine, produced in the chloro- form, into the aqueous phase. As the chloroform phase was slightly coloured after a few minutes of equilibration between the aqueous and organic phases, the aqueous aliquot used for determination of the triiodide ion was immediately separated from this extraction system, once the aqueous phase (8.4 ml) was added, in order to obtain maximum accuracy. Table 1 lists the results obtained in the determination of lipohydroperoxides in five commercially available oil samples by the triiodide-CPC method and shows the excellent agreement with those provided by classical iodimetric titra- tion.12 Determination of Hydroperoxides in Hydrocarbon Samples The presence of hydroperoxides in many hydrocarbon distil- lates has been associated with sediment formation, odour and colour, and studies of hydrocarbon oxidation are in need of sensitive measurements of the hydroperoxide concentration.Hydroperoxides can be quantitatively extracted from alkyl and aromatic hydrocarbon solutions by using ethanolic sodium hydroxide. Dialkyl and diaroyl peroxides are less acidic than hydroperoxides and do not respond t o this treatment. We used this extraction procedure with slight modifications (viz., the volume of hydrocarbon sample and extracts was reduced ten times, i.e. , 10 ml instead of 100 ml); ethanol was used instead of methanol, and two extractions were required to complete the recovery of hydrogen peroxide, Table 1 Determination of the lipohydroperoxide content of oil samples Lipohydroperoxide content/mequiv kg- 1 Triiodide-CPC Oil method* lodimetric titration Olive 9.99 9.72 Corn 4.38 4.44 Sunflower seed 8.29 8.2" Cod liver 8.73 8.73 Linseed 5.9" 7.65 * Oil samples of less than 0.2 g were used in every instance.cumene hydroperoxide and tert-butyl hydroperoxide (four extractions are needed in the original extraction procedure). In order to assess the applicability of the triiodide-CPC method to the determination of hydroperoxides in hydrscar- bon samples, commercially available analytical reagent-grade hydrocarbon samples were analysed. Hydroperoxides were undetectable in all instances. The analytical recoveries achieved at three different concentrations (7 X l O - 5 , 1 x 10-4 and 2 x 10-4 mol 1-1) of hydrogen peroxide, cumene hydroperoxide and tert-butyl hydroperoxide added individu- ally to heptane (Merck, Darmstadt, Germany) ranged between 98 and 106% (average of three determinations for each concentration of hydroperoxide). The results show the applicability of the triiodide-CPC method to the determina- tion of hydroperoxides in hydrocarbon samples.Conclusions The use of CPC micelles improves on the sensitivity and selectivity of the most commonly used method for the determination of hydroperoxides, which is based on the oxidation of iodide to iodine by these compounds. The proposed analytical system allows hydroperoxide concentra- tions below 1 pmol I-* to be determined with a high precision owing to the high molar absorptivity of the triiodide-CPC association complex, which surpasses that of most spectropho- tometric methods available for the determination of hydroper- oxides. The bathochromic shift undergone by triiodide ion in the micellar medium overcomes one of the most serious problems confronting the spectrophotometric determination of this ion in aqueous media, namely, the strong absorption of many samples (e.g., oils) at 350 nm. Although starch can also be used as a spectral shift reagent, it has so far been limited in its spectrophotometric applications because its molar absorp- tion coefficient depends on the chain length of the starch used.The proposed method is a valid alternative to the iodimetric titration method recommended by the American Oil Che- mists' Society for the determination of lipohydroperoxides in oils.12 Hence, based on the same principle, the triiodide-CPC method is more sensitive, rapid and economic (consumption of reagents and sample is considerably reduced) and less cumbersome than the titrimetric method. The authors gratefully acknowledge financial support from the CICyT (Project No. PB91-0840). 1 2 3 4 5 6 7 8 9 10 11 12 13 References Frew, J. E., Jones, P., and Scholes, G., Anal. Chim. Acta, 1983, 155, 139. Mair, R. D., and Hall, R. T., in Treatise on Analytical Chemistry, eds. Kolthoff, I. M . , and Elving, P. J., Wiley- Interscience, New York, 1971, vol. 14, Part 11, p. 295. Cathcart, R., Schwiers, E., and Ames, B. N., Anal. Biochem., 1983, 134, 111. Heaton, F. W.. and Uri, N., J . Food Sci. Agric., 1958, 9, 781. Hinze, W. L., in Solution Chemistry of Surfactants, ed. Mittal, K. L., Plenum, New York, 1979, vol. 1, p. 79. Cuccovia, M., and Chaimovich, H., Anal. Chem., 1982,54,789. Hayakawa, K., Kanda, M., and Satake, I., Bull. Chem. SOC. Jpn., 1979, 52, 3171. Lunar, M. L., Rubio, S., and PCrez-Bendito, D., Anal. Chim. Acta, 1992, 268, 145. Sorge, G., and Ueberreiter, K., Angew. Chem., 1956, 68,486. Eiss, M. I., and Giesecke, P., Anal. Chem., 1959,31, 1558. Wagner, C. D., Smith, R. H., and Peters, E. D., Anal. Chem., 1947, 19,976. Official Methods of Analysis of the Association of Official Analytical Chemists, ed. Horwitz, W., AOAC, Washington. DC, 1975, 12th edn., p. 489, 28.022-28.023. Pobiner, H., Anal. Chem., 1961,33, 1423. Paper 3100996C Received February 19, 1993 Accepted April 6, 1993

ISSN:0003-2654    

DOI:10.1039/AN9931801209

出版商:RSC    

年代:1993

数据来源: RSC

摘要:

ANALYST, SEPTEMBER 1993, VOL. 118 1213 Determination of Free State Manganese(i1) in a Decoction of a Traditional Chinese Medicine Based on a Kinetic Spectrophotometric Method Liu Jianli," Wang Budong and Zhang Fenyan De pa rtm en t o f Ch em ica I Engineering, North western U n ive rsit y, Xia n, S haa nxi, 7 I 0069, Peo p I e 's Rep u b lic of China The chemical state of trace elements in decoctions of traditional Chinese medicines is a new research subject in the chemical studies of traditional Chinese medicine. A simple and sensitive kinetic spectrophotometric method is proposed for the determination of free manganese(i1). The method is based on the catalytic effect of manganese(i1) on the oxidation of Malachite Green by potassium periodate. The reaction was followed spectrophotometrically by measuring the rate of change in the absorbance of the coloured Malachite Green at 615 nm.Trace amounts of manganese(i1) (0.2-2 ng) can be determined with good accuracy and reproducibility; the detection limit is about 0.1 ng. Only iron(li1) and aluminium(lll) interfere in the determination but these can be masked by fluoride. This method has been used t o analyse decoctions of traditional Chinese medicines for trace amounts of free manganese(l1) without separation. Keywords: Free manganese(i1) determination; kinetic spectrophotometry; traditional Chinese medicine The therapeutic benefit of traditional Chinese medicine (TCM), for the treatment ofdiseases, is based on the chemical constituents of the individual TCMs. There are thousands of organic constituents in the TCMs, some of which have biological activity, but none act independently and cannot replace the functions of TCM as a whole.Analysis of TCMs revealed that they were rich in many trace elements. It was suggested that this was the major factor in the functions of the TCMs. Trace elements were then determined in different kinds of TCM and compared with each other, in order to find the relationship between the content of trace elements and the functions of the individual TCM. However, the situation is complex, because the functions of the trace elements are critically dependent on their chemical states and environment within the TCM. Different states have different functions, toxicity and percentage absorption by the body. Therefore, research into the functions of trace elements in TCMs must consider their chemical states.' The chemical states in which trace elements are found are free, complexed and organically bound.The major com- ponents of TCM are organic compounds, many of which are complex agents. Trace elements co-exist with these organic compounds in a decoction of a TCM, most will be bound to organic compounds; therefore, the concentration of the free trace elements can be very low. In order to find evidence to support the prediction that most trace elements in a TCM exist in the bound state, the free trace elements should be determined initially. There are many sensitive methods for determining the total concentration of the trace elements present, but these do not differentiate between free and bound states.It was suggested2 that an ion-selective electrode should be used to determine the free trace elements; however, this was not found to be sensitive enough. We have been unable to find a suitable method in the literature for determining these very low concentrations of trace elements in a decoction of a TCM. This paper presents a kinetic spectrophotometric method to determine the free state of magnanese(r1) in decoctions of TCMs, based on its catalytic effect on the oxidation of Malachite Green by potassium periodate.3.4 Only iron(irr) and aluminium(ii1) influence the determination and these can be masked by fluoride. Various experimental conditions are * Present address: Department of Chemistry, University of Manchester, Oxford Road, Manchester, UK M13 9PL.discussed in detail. The method has been successfully applied to the determination of free manganese(n) in decoctions of TCMs. Experimental Reagents and Apparatus All chemicals were of analytical-reagent grade. De-ionized water was used throughout. Manganese(i1) stock solution, 1000 pg ml-1. Prepared by dissolving the appropriate amount of manganese sulfate (MnSO4.H20) in hydrochloric acid. The final concentration was determined by titration with ethylenediaminetetraacetic acid. Working manganese(i1) standard solutions, 10 and 0.1 pg ml-1. Prepared by diluting the stock solution. Malachite Green solution. Prepared by dissolving 0.1572 g of Malachite Green in de-ionized water (1 I). Potassium periodate. Prepared by dissolving 4.00 g of potassium periodate in de-ionized water (1 1).Sodium fluoride solution. Prepared by dissolving 0.913 g of sodium fluoride in de-ionized water (1 1). Buffer solution. Prepared by mixing 10 moll- 1 ethanoic acid with 2.6 moll-' ammonium acetate (1.6 + 1); the pH was 3.8. A spectrophotometer (Model 721; Shanghai Third Instru- ment Co.) was used for the absorbance measurements and an electronic thermostated water-bath was used to control the temperature of the reagent solution. Preparing the Sample Solution Place 300 g of the TCM in a beaker, add 2000 ml of de-ionized water, decoct until reduced to 1200 ml then filter through four layers of absorbent gauze to remove the solid. Add 2000 ml of de-ionized water to the solid, decoct again until reduced to 1200 ml, filter through four layers of absorbent gauze then combine the filtrates and allow t o stand overnight.After- wards, filter the decoction with a microfilter to remove the granules. The filtrate was used as the sample solution. Calibration Procedure Transfer a series of manganese(i1) standard solutions into 50 ml calibrated test-tubes fitted with ground-glass stoppers. Add1214 ANALYST, SEPTEMBER 1993, VOL. 118 2.0 ml of buffer solution, 0.8 ml of potassium periodate, 1.0 ml of sodium fluoride, 0.8 ml of Malachite Green solution and dilute to 50 ml with de-ionized water. Place the test-tube into the thermostated water-bath at 40 "C for 30 min. Then, measure the absorbance at 615 nm and prepare a calibration graph of absorbance versus concentration of manganese(I1). Procedure for the Determination of Free Manganese(I1) By using 2.0 ml of the sample solution instead of the standard solution the above procedure was repeated and the absor- bance measured ( A ) .Take 2.0 ml of the sample solution, omitting the potassium periodate, and measure the absor- bance ( A ' ) . The change of absorbance (A' - A ) is due to catalysis by free manganese(I1). Calculate the amount of free manganese(i1) from the calibration graph prepared pre- viously. Procedure for the Determination of Total Manganese(I1) Weigh 2.0 ml of the sample solution into a 100 ml beaker, add 10 ml of concentrated nitric acid and heat [Caution: Take care to avoid a violent reaction] until a tranquil solution is produced. Cool, then slowly add 5 ml of 60-70% perchloric acid, and heat until brown fumes of nitrogen dioxide appear. If the solution has a dark colour, add 2 ml of concentrated nitric acid and continue heating until white fumes of perchloric acid appear.Continue heating until evaporated to near dryness. Then, dissolve the residue in dilute hydrochloric acid (0.1 moll-'). Use this sample instead of the standard solution in the above procedure, and measure the absorbance. Calculate the total amount of manganese(i1) from the calibra- tion graph. Results and Discussion Absorption Spectrum An absorption spectrum was obtained by plotting absorbance against wavelength. The spectrum exhibited an absorption maximuin at 615 nm. 0.4 I 3 3.5 4 4.5 5 PH Fig. 1 ng ml-1 Mn; and C, 0.8 ng ml-1 Mn Relation between pH and absorbance. A, Blank; B, 0.4 0.5 I I 20 30 40 50 Te m pe rat u re/"C Relation between temperature and absorbance.A, Blank; B, Fig. 2 0.4 ng ml-1 Mn; and C, 0.8 ng ml-l Mn Effect of pH The influence of pH on the determination was studied by altering the pH while keeping the other conditions constant. The relation between the pH value and the rate of the catalytic reaction is shown in Fig. 1. The optimum pH range was 3.54.5. Outside this range, the rate of the catalytic reaction is slow and the difference between the different concentrations of manganese(i1) becomes less. Effect of Temperature The effect of temperature was studied between 20 and 60 "C while keeping the other conditions constant. The relation between temperature and the rate of the catalytic reaction is shown in Fig.2. The optimum temperature was 35-45 "C. If the temperature is below 35 "C, the rate of the catalytic reaction is slow and there is only a small absorbance difference between the different concentrations of manganese(i1). If the temperature is higher than 45 "C, the rate of the uncatalysed reaction increases and the absorbance becomes too small to determine accurately. Effect of Reaction Time The influence of reaction time was studied by using a 2 ng manganese(i1) standard solution and measuring the absor- bance, as above, at different times. The relation between reaction time and the rate of the catalytic reaction is shown in Fig. 3. The optimum reaction time is about 30 min. If the reaction time is too short, there is insufficient change of absorbance to determine accurately.If the reaction time is too long, the absorbance is too low to determine accurately. Effect of Reagent Volume The effects of the volumes of reagents were studied by altering each one in turn while keeping the others constant. The volumes chosen were those which yielded the maximum slope for the calibration. The optimum conditions obtained were as follows: potassium periodate, 0.8 ml; Malachite Green, 0.8 ml; NaF, 1.0 ml; and buffer solution, 2.0 ml. Interference Studies The interference effects of many cations and anions on the kinetic determination of manganese(I1) were examined: 10000-fold excesses of Na+, C1-, Pb2+, NO3-, K+, Br-, S042-, Cu2+, Hg2+ , Mo"', 2000-fold excesses of Co2+ , Ca2+ , Zn2+, Mg2+, Cr3+, I- and 1000-fold excesses of Ag+, V", Ni2+, did not interfere with the determination of man- ganese(i1).The only interference observed was by Fellr and Al"' and these can be masked by fluoride. There are many kinds of organic compounds in a decoction of a TCM. If these compounds react with potassium periodate or Malachite Green and form new compounds that absorb at 615 nm they will interfere. In the procedure for the determina- 0.8 I 1 Ti rn e/rn i n Relation between the reaction time and absorbance Fig. 3ANALYST, SEPTEMBER 1993, VOL. 118 1215 tion of free manganese(ii), A' - A subtracts the interference from the reaction between the compounds in the decoction and the Malachite Green. As there was no difference in absorbance between omitting the Malachite Green and omitting both the Malachite Green and the potassium periodate in the above procedures, this confirmed that there was no reaction, between compounds in the decoction and the potassium periodate, that produced an absorbance at 615 nm.Therefore, there was no interference in the determination of free manganese( 11) by using this procedure. Applications of the Method A calibration graph was prepared under the optimum condi- tions described above. The range of linearity was 0.2-2 ng ml-1, the detection limit was 0.1 ng ml-1, the relative standard deviation was 3.6% for nine determinations of 1.0 ng ml-1 of manganese(i1). Recovery experiments were carried out by adding 2.0 pg of manganese(r1) to decoctions of the TCMs. The total amount of Table 1 Recovery of manganese from TCMs Mn Mn Mn present/ added/ found/ Recovery Sample State pg ml-1 pg ml-1 pg ml-1 (%) cassia Presl Total 2.45 2.00 4.21 90 Cinnamoum Free 0.039 Free(%) 1.6 Zingier officinafe Free 0.002 Rosc. Total 5.1 2.00 6.83 95 Free (YO) 0.04 Pueonia lactiflora Free 0.004 Pall Total 2.32 2.00 4.37 102 Free(%) 0.17 Glycyri-&a Free 0.015 urafensisFisch Total 1.90 2.00 3.84 97 Free (YO) 0.78 manganese(i1) was determined and the recovery calculated; this is the total recovery.The recovery of the separated free and bound states could not be obtained, because of the complex equilibrium reaction between the free and com- plexed states of the manganese(i1). When free manganese(r1) was added to the decoctions, some free manganese(i1) might change to the bound state. The results are shown in Table 1. The proposed method was used to determine the free and total amount of manganese(i1) in decoctions of the TCMs. The results in Table 1 show that most of the manganese(r1) in the decoctions exists in the bound state, only very low concentra- tions are in the free state. Therefore, a study of the functions of trace elements in a TCM must consider the chemical states. The bound state might play a more important role than the free state. Conclusion A spectrophotometric method was used to determine quanti- tatively the free manganese(1i) in decoctions of TCMs without separation. The method has a low detection limit and does not suffer interference from inorganic ions and organic com- pounds. I t might be suitable for the determination of free manganese(i1) in other complex systems. This project was supported by the National Natural Science Foundation of China. The authors thank Janet Large for the revision of English in this manuscript. References 1 2 3 4 Liu, J . , Weifiang Yuansu, 1991, Suppl. 38. Zhu, T., Zhong Chao Yao, 1990, 21(10), 37. Liu, Z . , Huaxue Shijie, 1983, 45. Zheng, Z . , Wang, Y . , and Han, L., Fenxi Huaxue, 1989, 17, 160. Paper 2104940F Received September 1.5, 1992 Accepted April 30, 1993

ISSN:0003-2654    

DOI:10.1039/AN9931801213

出版商:RSC    

年代:1993

数据来源: RSC

摘要:

ANALYST, SEPTEMBER 1993, VOL. 118 1217 Nitrogen Factors for Beef: A Reassessment Analytical Methods Committee* Royal Society of Chemistry, Burlington House, Piccadilly, London, UK W l V OBN Nitrogen factors (the percentage of nitrogen on a fat-free basis) for beef were estimated using a total of 43 clean beef carcases and 30 cull cow carcases from six commercial abattoirs. The sample included carcases representative of the European Community classification ranges for fatness and conformation. The nitrogen factors for the lean meat with intermuscular fat for the whole sides of clean beef and cull cow beef were 3.65 and 3.70, respectively. For beef generally, a nitrogen factor of 3.65 is recommended for use in the analysis of beef and beef products. Nitrogen factors for different commercial joints differed significantly and the appropriate joint value should be used where the nature of the joint is known.Those for cull cow beef were generally higher than those for clean beef. All factors estimated were higher than 3.55, the value recommended for beef by the Analytical Methods Committee in 1963. Fat, moisture, ash, hydroxyproline and nitrogen contents in each joint and also in the side are reported in this paper. Keywords: Beef; nitrogen factor; fat; moisture; h ydrox yproline The Analytical Methods Committee has received and has approved for publication the following report from its Meat Factors Sub-committee. Report The constitution of the Sub-committee responsible for the preparation of this report was Professor R. A . Lawrie (Chairman), Miss 1.Agater (from March 1991), Mr. R. A. Evans, Mr. D. J . Favell, Dr. G. Finney (until December 1991), Mr. M. W. Fogden, Professor R. S. Hannan (until July 1989), Mr. A. J. Harrison, Mr. N. Harrison, Dr. G. Hodson [Ministry of Agriculture, Fisheries and Food (MAFF) Project Officer, until April 19891, Dr. R. B. Hughes, Dr. A. J. Kempster, Mr. R. S. Kirk, Miss D . B. Lowe, Dr. R. L. S. Patterson, Mrs. K. Swan, Dr. R. Wood (MAFF Project Officer) and Dr. M. L. Woolfe (MAFF Project Officer, from April 1989), with Mr. J. J. Wilson as Secretary. Introduction The nitrogen factor (the percentage of nitrogen on a fat-free basis) currently used in estimating the meat content of beef and beef products was established by the Analytical Methods Committee of the Society for Analytical Chemistry in 1963.1 The recommendation was that an average nitrogen factor of 3.55 was the best compromise for general use.This factor was based on the fat-free nitrogen of the raw comminuted lean meat with intermuscular fat, in the whole side. As there have been major changes in beef production over the past 20 years, with greater use of continental beef breeds and more Canadian Holstein blood in dairy-bred calves, this study was carried out to provide an up-to-date factor and to examine the chemical composition of beef joints in more detail than before. It was part of a larger study carried out in order to determine the over-all chemical composition of clean beef (cattle not used for breeding) and cull cow beef (beef from cows that are no longer economically productive), which included lipid content and fatty acid profiles.These aspects will be the subject of a further paper by the Meat and Livestock Commission (MLC). * Correspondence should be addressed to the Secretary, Analytical Methods Committee, Analytical Division, Royal Society of Chemistry, Burlington House, Piccadilly, London, UK W1V OBN. Experimental The Meat Factors Sub-committee worked closely with the MLC, the Lancashire County Council Laboratory (LCC) and five specified independent laboratories on the design and implementation of the study in all respects as it related to nitrogen factors for beef. Particular attention was paid to full compliance with written detailed protocols (approved by the Meat Factors Sub-committee) for carcase selection, dissec- tion, sample preparation, packaging and labelling, analytical test methods and quality control.Forty-three clean beef carcases and 30 cull cow carcases were included in the trial. Carcases were selected at the abattoir to fall into specific conformation and fat class cells. Conformation is a subjective assessment of the surface shape of a carcase, made on the slaughterline; a carcase of conformation ‘U’ has more convex surfaces and thicker musculature than a carcase of conformation ‘P’. An MLC 15-point numerical scale is used in the selection of carcases with an acceptability range for each European Community (EC) class. Fat classes describe the amount of fat cover on the surface of the carcase. Approximate percentage subcutaneous fat cover (SFe) allows acceptability ranges for each fat class.Details of the E C classifications are given in ref. 2. For statistical analysis the fat and conformation classes of each carcase were converted to 15-point scales (see ref. 3 and Tables 1 and 2). Clean beef carcases of all three sexual types, viz., steers, heifers and young bulls, were selected from six commercial abattoirs representing a geographical spread. The carcases covered the main EC conformation and fatness classes (Table Cull cow carcases were sampled from three commercial abattoirs. The abattoirs supplied carcases from 5-10 year old 3). Table 1 Relationship between the fat classes used in commercial classification and a visual assessment of carcase subcutaneous fat content to the nearest percentage unit (SFe) (from ref.3) SFe EC fat class Range Mean <4.5 3.00 1 2 3 4.5-7.4 6.00 4L 7.5-8.9 8.25 4H 9.0-1 0.4 9.75 5L 10.5-13.4 12.00 5H >13.5 15.001218 ANALYST, SEPTEMBER 1993, VOL. 118 Table 2 Relationship between the conformation classes used in commercial classification and conformation on a 15-point scale (C15) (from ref. 3) C15 EC confirmation - class P 0- O+ R U- U+ E Range 1 2-3 4-6 7-9 10-12 13-14 >15 Mean 1 .o 2.5 5.0 8.0 11.0 13.5 15.0 Table 3 Sampling design of the clean beef trial. AC (acceptability criteria) for fat class as estimated percentage of subcutaneous fat in the carcase (SFe), and conformation on standard MLC 15-point scale. Total sample: 11 young bulls, 21 steers and 11 heifers Fat class Conformation class u+/u- (AC: 10-13) R/O+ (AC: 4-9) 0-/P (AC: 1-3) 2 (AC: 3-5) 2 young bulls 1 steer 2 young bulls 4 steers 2 young bulls 3 steers 4L 5L (AC: 8-9) AC: 11-13) 2 young bulls 2 steers 2 steers 2 heifers 2 heifers 2 young bulls 2 steers 4 steers 4 heifers 1 heifer 1 young bull 3 steers 2 heifers Table 4 Sampling design of the cull cow trial.Each cell consists of three carcases of each of two age groups (4-7 and 8-10 years). AC (acceptability criteria) for fat class as estimated percentage of subcutaneous fat in the carcase (SFe), and conformation on standard MLC 15-point scale Fat class Conformation 112 4L 5L/5H class (AC: 1-5) (AC: 8-9) (AC: 11-15) P+/P- 0-/P+ - (AC: 1-2) 6 6 6 (AC: 1-3) - - O+/O- (AC : 2-6) - (AC: 4-9) - - 6 R/O+ 6 - cows, covering the complete range of EC fat classes and the poorer conformation classes, viz., R, 0 and P (Table 4).Throughout this report the term ‘side weight’ refers to half of the mass of the dressed cold carcase. The fatness and conformation scales are referred to as SFe and (215, respect- ively. Dissection and Chemical Analysis The left side of each carcase was divided into 15 joints: brisket, jacobs ladder, fore rib, chuck, thin flank, topside, shin, leg, clod, sticking, loin, rump, thick flank, silverside and the fillet, as described by Kempster et aL4 and shown in Fig. 1. Each joint was separated into its component tissues, viz., lean, intermuscular fat, subcutaneous fat and bone, and the tissues, excluding bone , were macerated. The macerated tissues of shin and leg, and of clod and sticking, were then recombined proportionally by weight on an individual tissue Topside (media I ) and silverside (lateral) Jacobs ladder Sticking Clod Fig.1 Standardized joints used in MLC beef dissection technique Table 5 Grouping of joints into pistola and residual fore quarter Pistola Residual fore quarter Chuck Leg Topside Sticking Thick flank Clod Silverside Shin Rump Jacobs ladder Fillet Brisket Loin Thin flank Fore rib 3.75 / Young bull 3.6 I I I I I I S Fe t 1 I 1 1 I I Fat class Fig. 2 Fat-free nitrogen in clean beef sides by sexual type in relation to SFe (fat class). Solid lines = lean tissue; broken lines = lean with fatty tissues 0 2 4 6 8 1 0 1 2 1 4 112 3 4L 4H 5LANALYST, SEPTEMBER 1993, VOL. 118 1219 basis on all carcases. The loin, rump and fillet, and the thick flank and silverside from cull cow carcases only were recombined in the same way.The samples, three separate tissues of each joint, were homogenized, stored at - 18 "C, and then, in the context of the results reported in this paper, were analysed for lipid, moisture, ash, hydroxyproline and nitrogen contents. The LCC and MLC were the principal laboratories involved in this study. The estimated chemical compositions of the lean with intermuscular fat, and the lean with intermuscular and subcutaneous fats, were constructed mathematically from the components of the individual tissues. The tabulated results therefore show five tissue types. In addition, the estimated chemical compositions of the combined loin, rump and fillet, and the combined thick flank and silverside, were calculated for the clean beef carcases to allow comparison with those of the cull cows.Industrial practice often separates carcases into pistola and Table 6 Lean tissue: chemical composition and content in each clean beef joint as a percentage of total lean (standard errors in parentheses) Chemical composition (% of tissue) Brisket Jacobs ladder Fore rib Chuck Thin flank Shin and leg Clod and sticking Topside Loin, rump and fillet Thick flank and silverside Side Pistola Fore quarter % of total 8.58 (0.114) 8.66 (0.129) 3.78 (0.055) 15.25 (0.145) 6.90 6.03 (0.061) 10.12 (0.151) 9.14 (0.065) 15.00 (0.109) 16.54 100.00 (0.100) (0.112) 48.32 (0.177) 51.67 (0.177) Lipid 8.9 (0.33) 5.3 (0.19) 7.4 (0.37) 5.9 6.9 (0.22) (0.31) 2.5 (0.15) 3.8 (0.16) 2.8 (0.14) 5.8 3.4 (0.22) (0.15) 5.1 (0.17) 4.3 (0.16) 5.9 (0.19) Moisture 69.9 (0.28) 72.5 (0.15) 70.6 (0.29) 72.6 70.9 (0.26) 74.2 (0.15) 73.8 (0.16) 73.8 (0.14) 71.5 (0.19) 73.7 (0.14) 72.5 (0.15) 72.9 (0.17) 72.2 (0.17) (0.21) Ash 1.01 1.06 (0.007) 1.03 (0.007) 1.01 (0.009) 1.04 (0.008) 1.04 (0.007) 1.06 (0.007) 1.11 (0.008) 1.06 1.08 (0.005) 1.05 (0.004) 1.07 (0.005) 1.03 (0.005) (0.008) (0.008) Hydroxy- proline 0.27 0.27 0.24 (0.006) 0.29 0.25 (0.009) 0.44 (0.0 lo) (0.01 0) (0.01 1) (0.013) 0.35 (0.009) 0.18 (0.006) 0.23 (0.009) 0.28 (0.026) 0.28 (0.006) 0.26 (0.007) 0.30 (0.0 10) Nitrogen 3.25 (0.0 19) 3.45 (0.016) 3.39 3.36 3.43 (0.024) 3.61 3.45 3.60 (0.0 17) 3.45 (0.0 17) 3.52 3.45 (0.015) 3.51 (0.01 6) 3.39 (0.015) (0.020) (0.02 1 ) (0.018) (0.021) (0.02 1 ) Fat-free nitrogen 3.57 3.64 3.66 3.57 3.68 3.71 (0.0 18) (0.015) (0.01 8) (0.022) (0.022) (0.020) 3.59 (0.022) (0.020) 3.71 3.66 (0.014) 3.64 3.64 3.67 3.61 (0.01 5 ) (0.02 1 ) (0.0 15) (0.016) Table 7 Intermuscular fat: chemical composition and content in each clean beef joint as a percentage of total intermuscular fat (standard errors in parentheses) Chemical composition (% of tissue) Brisket Jacobs ladder Fore rib Chuck Thin flank Shin and leg Clod and sticking Topside Loin, rump and fillet Thick tlank and silverside Side Pistola Fore quarter of total 19.88 (0.28 1) 12.12 (0.228) 5.38 (0.106) 11.14 (0.206) 10.56 (0.282) 3.05 (0.129) 11.62 (0.219) 3.19 (0.090) 12.48 (0.257) 10.58 (0.176) 100.00 33.64 (0.376) 66.35 (0.376) Lipid 74.7 (0.77) 73.4 (0.78) 76.1 (0.83) 71.2 (0.72) 75.1 (1.06) 50.1 (1.27) 69.9 (0.67) 64.9 (0.97) 76.8 (0.54) 67.6 (0.81) 72.2 (0.65) 71.2 (0.63) 72.7 (0.67) Moisture 18.9 (0.57) 20.4 (0.59) 17.5 (0.59) 21.8 (0.52) 18.0 (0.66) 36.6 (0.86) 23.0 26.2 (0.70) 16.9 (0.39) 25.2 (0.62) 20.9 (0.47) 21.5 (0.46) 20.5 (0.49) (0.50) Ash 0.25 (0.009) 0.26 0.31 (0.026) 0.31 0.23 (0.008) 0.42 0.27 (0.009) 0.34 (0.013) 0.27 0.30 (0.009) 0.27 (0.007) 0.30 (0.008) 0.26 (0.007) (0.01 0) (0.010) (0.011) (0.0 10) Hydroxy- proline 0.46 (0.033) 0.44 (0.027) 0.41 (0.025) 0.53 (0.039) 0.047 (0.039) 1.13 (0.063) 0.51 (0.030) 0.56 (0.030) 0.39 (0.0 17) 0.56 (0.025) 0.49 0.40 0.48 (0.023) (0.02 1) (0.018) Nitrogen 1.01 (0.039) 0.97 (0 .O40) 1 .oo (0 .O41) 1.08 (0.036) 1.10 (0.064) 2.07 (0 .081 ) 1.13 1.35 (0.049) 0.98 1.12 1.08 1.12 (0.028) 1.07 (0.037) (0.026) (0.009) (0.030) (0.034) Fat-free nitrogen 4.01 3.66 (0.130) 4.17 (0.093) 3.71 (0.087) 4.41 (0.136) 4.14 (0.098) 3.73 (0.084) 3.84 (0.093) 4.24 3.45 (0.056) 3.89 (0.050) 3.90 3.89 (0.1 12) (0.052) (0 "3) (0.066)1220 ANALYST, SEPTEMBER 1993, VOL.118 residual fore quarter parts. The joints included in these are laboratories (see Appendix). The methods used in the shown in Table 5. The chemical composition of the pistola, the interlaboratory calibration were those of British Standard residual fore quarter and the side was additionally calculated. 44015 or approved variants that had been shown to give The analyses carried out by the LCC and MLC laboratories equivalent results. Additionally, the MLC used CEM6 were separately calibrated against the five independent methods for fat and moisture determination and these results laboratories appointed by the Meat Factors Sub-committee to were reconciled with those obtained from the British Standard be representative of enforcement, industrial and research methods.5 Table 8 Subcutaneous fat tissue: chemical composition and content in each clean beef joint as a percentage of the total subcutaneous fat (standard errors in parentheses) Chemical composition (Yo of tissue) Brisket Jacobs ladder Fore rib Chuck Thin flank Shin and leg Clod and sticking Topside Loin, rump and fillet Thick flank and silverside Side Pistola Fore quarter % of total 14.53 (0.349) 6.60 3.97 (0.141) 7.13 (0.192) 14.41 (0.41 0) 6.73 (0.418) 7.96 (0.285) 7.27 (0.329) 19.85 (0.405) 11.55 (0.396) 100.00 (0.359) 47.58 (0.606) 52.42 (0.606) Lipid 71.7 (0.97) 77.8 (1.28) 85.1 (1.37) 77.3 (1.18) 74.7 (1.16) 49.3 (1.73) 68.5 (1.39) 79.6 (0.77) 82.8 (0.73) 68.4 (1.26) 74.3 (0.95) 75.7 (0.99) 72.9 (0.98) Moisture 21.7 15.8 (0.82) 10.5 (0.93) 15.7 (0.87) 19.1 (0.79) 33.7 (0.70) (1.08) 22.0 (0.84) 14.2 (0.45) 12.0 (0.42) 21.3 (0.84) 18.3 (0.60) 16.7 (0.61) 19.8 (0.65) Ash 0.22 0.18 (0.019) 0.13 (0.01 9) 0.24 0.19 0.39 (0.013) 0.26 (0.017) 0.16 (0.087) 0.12 0.23 (0.01 3) 0.20 0.18 (0.0 16) 0.22 (0.012) (0.021) (0.02 1) (0.01 0) (0.012) (0 * 0 12) Hydroxy- proline 0.67 (0.033) 0.60 (0.044) 0.39 (0.033) 0.65 (0.044) 0.58 (0.033) 1.68 (0.105) 0.86 (0.05 1) 0.64 (0.039) 0.53 (0.032) 0.97 (0.052) 0.72 (0.036) 0.75 (0.060) 0.69 (0.034) Nitrogen 1.04 1.02 (0.077) 0.66 (0.082) 1.04 (0.052) 0.95 (0.064) 2.76 (0.124) 1.44 1 .OO 0.81 1.58 (0.080) 1.16 (0 .062) 1.20 (0.068) 1.12 (0.060) (0.059) (0.099) (0.054) (0.05 9) Fat-free nitrogen 3.61 4.50 (0.103) 4.50 (0.2 13) 4.55 (0.138) 3.70 5.39 (0.125) 4.47 4.82 (0.092) 4.63 (0.140) 4.92 (0.119) 4.41 (0.085) 4.85 4.07 (0.083) (0.120) (0.120) (0.122) (0.101) Table 9 Lean and intermuscular fat: chemical composition and content in each clean beef joint as a percentage of the total lean and inter- muscular fat (standard errors in parentheses) Chemical composition (YO of tissue) Brisket Jacobs ladder Fore rib Chuck Thin flank Shin and leg Clod and sticking Topside Loin, rump and fillet Thick flank and silverside Side Pistola Fore quarter YO of total 10.23 (0.147) 9.17 (0.130) 4.03 (0.057) 14.65 (0.137) 7.45 (0.108) 5.58 (0.067) 10.34 (0.148) 8.27 (0.064) 14.63 (0.113) 15.65 (0.106) 100.00 46.15 (0.190) 53.84 (0.190) Lipid 27.6 (0.69) 18.4 (0.57) 20.9 (0.64) 13.1 (0.40) 21.1 (0.67) 6.2 (0.27) 14.7 (0.48) 6.3 (0.25) 14.8 (0.45) 9.6 (0.34) 14.9 (0.44) 11.4 (0.36) 17.9 (0.51) Moisture 55.4 (0.53) 62.5 (0.43) 60.2 67.0 (0.34) 59.9 (0.51) 71.4 (0.49) (0.22) 65.5 (0.39 71.2 64.7 (0.36) 69.1 (0.27) 65 .0 67.4 (0.29) 62.9 (0.40) (0.20) (0.34) Ash 0.80 0.91 (0.008) 0.89 (0.009) 0.93 (0.006) 0.87 1 .OO (0.008) 0.93 (0.008) 1.07 (0.007) 0.96 (0.008) 1 .oo (0.006) 0.94 (0.005) 0.99 (0.0O6) 0.90 (0.006) (0.010) (0.0 10) Hydroxy- proline 0.32 0.30 0.27 (0.006) 0.31 (0.013) 0.29 0.49 (0.014) 0.37 (0.007) 0.20 (0.006) 0.25 (0.007) 0.31 (0.008) 0.31 0.28 (0 .009) 0.33 (0.007) (0.012) (0.009) (0.01 0) (0.005) Nitrogen 2.61 (0.028) 2.97 (0.025) 2.92 (0.025) 3.10 2.94 (0.032) 3.49 3.07 3.48 (0.0 18) 3.14 3.29 (0.023) 3.10 3.26 2.97 (0.023) (0.022) (0.020) (0.026) (0.020) (0.02 1 ) (0.01 9) Fat-free nitrogen 3.61 (0.018) 3.64 (0.01 7) 3.69 (0.019) 3.57 (0.02 1 ) 3.73 (0.023) 3.72 3.59 (0 .023) 3.71 (0.01 7) 3.68 (0.014) 3.64 3.65 (0.015) 3.68 (0.0 15) 3.62 (0.016) (0.020) (0.02 1 )ANALYST, SEPTEMBER 1993, VOL.118 1221 Results Clean Beef An initial data analysis showed that the response of the chemical determinants to changes in the fat index (SFe) depended on the sexual type of the carcase, hence the model for data analysis (including effect of sexual type, regression on side weight and (21.5) had separate regressions on SFe for each type- Initially, predictions were made for representative carcases of each sexual type.These were specified as follows: Steer: side weight 158 kg, conformation R, fat class 4L Heifer: side weight 130 kg, conformation R/O+ fat class 4L Young bull: side weight 142 kg, conformation R, fat class 3 The national slaughter population during the sampling period consisted of approximately 57% steers, 26% heifers Table 10 Lean and fatty tissues: chemical composition and content in each clean beef joint as a percentage of total lean and fatty tissues (standard errors in parentheses) Chemical composition (YO of tissue) Brisket Jacobs ladder Fore rib Chuck Thin flank Shin and leg Clod and sticking Topside Loin, rump and fillet Thick flank and silvcrside Side Pistola Fore quarter % of total 10.57 8.96 (0.135) 4.03 (0.058) 14.05 (0.130) 8.04 (0.119) 5.63 (0.067) 10.15 (0.138) 8.17 (0.075) 15. 12 (0.116) 15.28 100.00 46.26 (0.175) 53.74 (0.175) (0.152) (0.110) Lipid 32.4 (0.75) 22.1 (0.62) 25.9 (0.74) 15.8 (0.47) 28.8 (0.85) 9.9 (0.43) 18.2 11.6 (0.47) 22.2 (0.63) 13.2 (0.44) 19.7 (0.56) 16.8 22.3 (0.61) (0.59) (0.50) Moisture 51.7 (0.57) 59.6 (0.46) 56.3 (0.57) 64.8 (0.39) 54.1 (0.65) 68.1 (0.31) 62.7 (0.46) 67.0 (0.35) 59.0 (0.49) 66.2 (0.34) 61.2 (0.43) 63.2 (0.39) 59.5 (0.47) Ash 0.73 0.86 (0.008) 0.83 0.91 (0.007) 0.77 0.95 (0.008) 0.89 (0.009) 1 .oo (0.008) 0.87 0.95 (0.007) 0.88 (0.007) 0.92 (0.007) 0.85 (0.007) (0.0 10) (0.01 0) (0.011) (0.0 10) Hydroxy- proline 0.36 0.31 (0.008) 0.28 0.33 0.33 (0.009) 0.59 (0.0 17) 0.40 (0 .008) 0.23 (0.007) 0.27 (0.007) 0.34 (0.009) 0.34 (0.005) 0.31 (0.009) 0.35 (0.007) (0.01 2) (0.005) (0.012) Nitrogen 2.43 (0.030) 2.84 (0.026) 2.74 3.02 2.65 (0.037) 3.43 (0.026) 2.96 (0.030) 3.29 (0.023) 2.88 (0.027) 3.18 (0.025) 2.94 (0.024) 3.08 2.82 (0.026) (0.029) (0.022) (0,022) Fat-free nitrogen 3.60 (0.019) 3.65 (0.017) 3.70 3.58 3.73 (0.023) 3.80 3.61 (0.023) 3.73 (0.017) 3.70 (0.014) 3.66 3.66 3.70 3.63 (0.0 16) (0.020) (0.02 1) (0.021) (0.02 1) (0.015) (0.015) Table 11 Regression data for fat-free nitrogen (Yo of tissue): clean beef sides by sexual type Tissue Lean lntermuscular fat Subcutaneous fat Lean with intermuscular fat Lean with fatty tissues Sexual type Steer Heifer Young bull Steer Heifer Young bull Steer Heifer Young bull Steer Heifer Young bull Steer Heifer Young bull Constant 3.654 3.678 3.SO6 4.330 4.288 4.145 6.111 5.916 6.116 3.688 3.723 3.552 3.743 3.779 3.610 Side weight/kg X1 0.0001 0.0001 0.0001 - 0.0014 -0.0014 - 0.0014 - 0.0047 - 0.0047 - 0.0047 - 0.0001 - 0.0001 -0.0001 -0.0001 -0.0001 -0.0001 C15 x2 0.001 0.001 0.001 -0.006 -0.006 -0.006 0.005 0.005 0.005 0.000 0.000 0.000 0.001 0.001 0.001 SFe (YO) x3 -0.003 -0.008 0.014 -0.021 -0.027 0.019 -0.137 -0.122 -0.073 -0.002 -0.009 0.014 -0.004 -0.011 0.014 Fat free nitrogen is given by the following expression: constant + ( X , x side weight) + ( X , x Cl5) + (X3 x SFe) Example: The fat-free nitrogen content in the lean with intermuscular fat of a young bull with side weight = 150 kg, conformation R* (equivalent to C15 = 8) and fat class 2* (equivalent to SFe = 3) is given by 3.552 + (-0.0001 X 150) + (0.000 X 8) + (0.014 X 3) = 3.579 * See Tables 1 and 2 for conversion from EC grades to 15-point scales.1222 ANALYST, SEPTEMBER 1993, VOL. 118 Table 12 Fat-free nitrogen content for the tissues of each clean beef joint and the side (% of tissue) Brisket Jacobs ladder Fore rib Chuck Thin flank Shin and leg Clod and sticking Topside Loin, rump and fillet Thick flank and silverside Least significnat difference* Side Pistola Forc quarter Lean 3.57 3.64 3.66 3.57 3.68 3.71 3.59 3.71 3.66 3.64 0.07 3.64 3.67 3.61 Intermuscular fat 4.01 3.66 4.17 3.71 4.41 4.14 3.73 3.84 4.24 3.45 0.34 3.89 3.90 3.89 Subcutaneous fat 3.61 4.50 4.50 4.55 3.70 5.39 4.47 4.82 4.63 4.92 0.47 4.41 4.85 4.07 * The smallest difference between any two joints that is statistically significant. - 3'75 I - I I \ \ \Lean tissue \ \ '\ 3.65 Lean and fatty tissues 3 .6 l I I I ' I 0 2 4 6 8 I 0 1 2 1 4 S Fe L I I I I I I I t 2 3 4L 4H 5L Fat class Fig. 3 Fat-free nitrogen in cull cow sides in relation to SFe (fat class) and 17% young bulls.7 The means of the three sexual types, weighted in this way, formed the basis for the consolidated clean beef results reported in this paper. Side weight generally explained little of the statistical variation in the fitted models, but C15 and interactions of SFe with sexual type both contributed significantly to the models. The lipid in the fatty tissues increased at a much greater rate with SFe for the steers and heifers than for the young bulls.The fat-free nitrogen in the lean, and in the lean with fatty tissues, of the side decreased with increasing SFe for steers and heifers but increased for the young bulls, although a typical carcase of each sexual type differed little in its fat-free nitrogen content (Fig. 2). With the exception of the chuck and fore rib, each individual joint showed similar responses, i.e., the fat-free nitrogen increasing with SFe for young bulls while decreasing with SFe for the other two sexual types. Tables 6-10 show the chemical composition of the five tissue types with the contribution made by each joint to that tissue type in the side. The regressions of fat-free nitrogen in the side on weight, conformation and fatness are shown in Table 11.Table 12 gives the fat-free nitrogen for each joint and tissue with approximate least significant differences. The joints differed significantly in their fat-free nitrogen content, in particular the brisket, chuck, clod and sticking had consis- tently lower levels in the lean and in the lean with fatty tissues Lean and intermuscular fat 3.61 3.64 3.69 3.57 3.73 3.72 3.59 3.7 1 3.68 3.64 0.07 3.65 3.68 3.62 Lean, inter- muscular and subcutaneous fat 3.60 3.65 3.70 3.58 3.73 3.80 3.61 3.73 3.70 3.66 0.07 3.66 3.70 3.63 than the other joints. All determined nitrogen factors were greater than the currently recommended factor of 3.55. Cull cows Initial data analysis showed a curved response to differences in SFe for some chemical determinants, especially for lipid.Models with regression on side weight, conformation and both linear and quadratic terms for SFe were therefore fitted. Predictions were made, and are presented, for a typical cull cow carcase of conformation 0-, fat class 4L and 135 kg side weight. The response of fat-free nitrogen in lean and in lean with fatty tissues, with fatness level is shown in Fig. 3. Tables 13-17 show the chemical composition of the five tissue types, again with the contribution made by each joint to the total of each tissue type in the side. The relationships between fat-free nitrogen in the side and the independent variables are shown in Table 18. These regressions show how the fat-free nitrogen factors can be calculated from the important variables. Table 19 gives the fat-free nitrogen content for each joint and each tissue with approximate least significant differences.Again the joints differed in their fat-free nitrogen contents, the pattern being similar to that for clean beef with the exception of topside, which showed a low value compared with the other joints. In general the nitrogen factors for cull cow beef were greater than those for clean beef, which probably reflects the greater age of the former animals. As with the clean beef data all exceeded the currently rec- ommended figure of 3.55. This figure was markedly higher than the value, namely 3.4, recommended by the Analytical Methods Committee in 19528 and it is evident that the nitrogen factor for beef found in the present work of between 3.65 and 3.70, represents an increment of a similar order over the Analytical Methods Committee recommendation made in 1963.Currently, the estimated national kill of beef animals consists of 20% cull cows and 80% clean beef.9 Although such a proportion may be rather less appropriate for beef products (in which relatively more cull cow beef may be used), it is suggested that this proportion should be the present basis for establishing the general factor for beef and the factor thus derived from a mathematical recombination of the analytical results obtained is 3.65 (rounded as usual to the nearest 0.05). Beef for processing is often traded on the basis of a prescribed lean-to-fat ratio, e.g., 90% visual lean. A comparison of the fat-free nitrogen figures for lean and lean with fatty tissues shows little differences (Tables 12 and 19).It can therefore beANALYST, SEPTEMBER 1993, VOL. 118 1223 Table 13 Lcan tissue: chemical composition and content in each cull cow joint as a percentage of total lean (standard errors in parentheses) Chemical composition (YO of tissue) Brisket Jacobs ladder Fore rib Chuck Thin flank Shin and leg Clod and sticking Topside Loin, rump and fillct Thick flank and silverside Side Pistola Fore quarter Yo of total 8.85 (0.187) 8.96 (0.170) 3.57 (0.084) 15.42 7.37 (0.143) 5.76 9.91 (0.244) 8.90 (0.106) 15.53 (0.174) 15.73 (0.186) 100.00 (0.21 2) (0.101) 47.42 (0.330) 52.58 (0.330) Lipid 9.0 5.5 (0.23) 8.5 (0.68) 6.2 (0.39) 3.8 2.6 4.0 2.1 (0.24) 5.1 (0.46) 2.7 5.0 (0.24) 3.8 (0.29) 6.1 (0.27) (0.51) (0.50) (0.52) (0.25) (0.37) Moisture 69.9 (0.38) 72.4 (0.24) 69.7 (0.58) 72.4 (0.39) 71.3 (0.39) 74.7 (0.39) 73.9 73.4 71.3 (0.34) 73.6 (0.24) 72.4 72.6 72.2 (0.29) (0.35) (0.21) (0.21) (0.20) Ash 1 .00 (0.024) 1.07 (0.0 19) 1.01 0.96 (0.045) 0.98 (0.045) 1.03 (0.031) 1 .05 (0.016) 1.12 1.04 (0.0 14) 1.02 (0 .040) 1.03 (0.0 17) 1.04 (0.0 19) 1.01 (0.024) (0.067) (0.020) Hydroxy- prolinc 0.26 0.26 0.22 (0.014) 0.27 (0.008) 0.26 0.39 (0.01 6) 0.30 (0.009) 0.19 (0.009) 0.22 0.23 0.25 0.23 (0.007) 0.27 (0.006) (0.009) (0.01 0 ) (0.012) (0.009) (0.01 0) (0.005) Nitrogen 3.38 (0.038) 3.56 3.45 (0.029) 3.42 (0.026) 3.48 (0.034) 3.71 3.54 (0.03 1) 3.58 3.47 3.52 (0.030) 3.50 3.53 3.47 (0 I 025) (0.033) (0.033) (0.028) (0.022) (0.025) (0.022) Fat-free nitrogen 3.71 3.76 (0.027) 3.77 (0.030) 3.64 (0.030) 3.73 (0.032) 3.81 (0.034) 3.68 (0.036) 3.66 (0.030) 3.66 (0 424) 3.62 3.68 3.66 (0.024) 3.70 (0.024) (0.034) (0.034) (0.022) Table 14 Intermuscular fat: chemical composition and content in each cull cow joint as a percentage of total intermuscular fat (standard errors in parentheses) Chemical composition (YO of tissuc) Brisket Jacobs ladder Fore rib Chuck Thin flank Shin and leg Clod and sticking Topside Loin, rump and fillet Thick flank and silverside Side Pistola Fore quarter % of total 18.82 (0.625) 12.14 (0.379) 5.67 (0.295) 11.13 (0.442) 10.89 (0.5 18) 2.74 (0.232) 11.58 3.13 13.64 (0.3 10) 10.26 (0.4 19) 100.00 (0.522) (0.102) 34.60 (0.673) 65.40 (0.673) Lipid 73.5 (1.43) 73.2 (1.98) 77.1 ( 1 3 3 ) 70.5 (1.35) 70.4 (1.94) 53.0 (1.99) 71.2 67.0 (1.37) 75.7 68.4 71.8 (1.28) 71.7 (1.14) 71.9 (1.45) (1.79) (1.00) (1.57) Moisture 18.6 19.8 (1.68) 15.3 (3.31) 21.8 (1.14) 19.7 (1.40) 33.8 (1.12) (1.52) 21.2 (1.42) 24.3 (1.05) 17.2 (0.78) 23.9 20.2 20.4 (0.97) 20.1 (1.35) (1 .0S) (1.21) Ash 0.29 (0.026) 0.32 0.32 0.37 (0.031) 0.31 (0.024) 0.43 (0.027) 0.32 0.3.5 (0.0 19) 0.29 (0.0 16) 0.28 0.31 (0.01 7) 0.30 (0.015) 0.26 (0.036) (0.021) (0.025) (0.025) (0.020) Hydroxy- proline 0.47 (0.032) 0.48 (0.033) 0.42 0.61 (0.060) 0.62 (0 .OS9) 1.09 (0.075) 0.60 0.52 (0.053) 0.44 (0.033) 0.58 (0 .O36) 0.54 (0.024) 0.52 (0.026) 0.63 (0.078) (0 429) (0.051) Nitrogen 1.06 1.04 (0.059) 0.95 (0.060) 1.13 (0.050) 1.29 1.91 (0.132) 1.14 (0.070) 1.26 (0.069) 0.97 (0.040) 1.08 (0.052) 1.11 (0.039) 1.07 (0.033) 1.13 (0.050) (0.061 ) (0.1 20) Fat-free nitrogen 4.01 (0.170) 3.99 (0.183) 4.13 (0.157) 3.86 4.27 4.04 4.02 (0.137) 3.82 (0.106) 4.04 3.48 (0.185) 3.97 3.84 4.04 (0.117) (0.118) (0.22 1) (0.180) (0.112) (0.100) (0.101)1224 ANALYST, SEPTEMBER 1993, VOL.118 ~ ~ ~~ ~~~ ~ Table 15 Subcutaneous fat: chemical composition and content in each cull cow joint as a pcrcentage of total subcutaneous fat (standard errors in parentheses) Chemical composition (Yo of tissue) Brisket Jacobs ladder Fore rib Chuck Thin flank Shin and leg Clod and sticking Topside Loin, rump and fillet Thick flank and silverside Side Pistola Fore quarter % of total 13.83 (1.055) 6.74 (0.430) 3.61 (0.408) 6.43 (0.496) 13.66 (0.845) 5.80 (0.631) 7.42 (0.652) 6.47 (0.815) 24.71 (1 -361) 11.33 (0.743) 100.00 50.54 (1.743) 49.46 (1.743) Lipid 72.2 (2.57) 80.2 87.0 (3.31) 79.7 (2.74) 73.9 (2.47) 51.5 (2.09) 72.6 (2.54) 78.9 (2.28) 83.8 (3.42) 70.9 (2.48) 76.3 (2.44) 78.1 (2.49) 74.3 (2.53) (3.55) Moisture 20.0 (1.99) 12.9 (2.09) 7.9 (1.91) 13.2 (1.48) 19.0 (1.87) 31.1 (1.85) 18.9 14.0 (1.53) 10.0 (1.87) 19.3 (1.53) 16.2 (1.67) 14.5 (1.50) 18.0 (1.83) (2.00) Ash 0.23 (0.024) 0.26 (0.041) 0.18 (0.060) 0.27 (0.162) 0.25 (0.032) 0.42 (0.032) 0.30 (0.032) 0.19 (0.026) 0.17 (0.034) 0.28 (0.039) 0.23 (0.027) 0.21 (0.027) 0.26 (0.035) Hydroxy- proline 0.63 (0.085) 0.49 (0.147) 0.29 (0.106) 0.56 (0.136) 0.65 (0.089) 1.51 (0.144) 0.72 (0.078) 0.66 (0.068) 0.47 (0.133) 0.92 (0.113) 0.65 0.66 (0.097) 0.63 (0.078) (0.081) Nitrogen 1.18 (0.099) 1 .00 (0.236) 0.65 (0.248) 1.11 1.11 (0.104) 2.64 1.22 1.29 (0.141) 0.96 (0.274) 1.72 1.21 (0.137) 1.26 (0.183) 1.16 (0.222) (0.211) (0.1 12) (0.200) (0.116) Fat-free nitrogen 4.2 1 (0.171) 4.92 (0.224) 4.83 (0.242) 5.31 (0.182) 4.22 (0.204) 5.43 (0.285) 4.56 (0.204) 6.19 (0.272) 5.86 (0.228) 5.91 (0.355) 5.06 (0.137) 5.71 (0.169) 4.50 (0.145) Table 16 Lean and intermuscular fat: chemical composition and content in each cull cow joint as a percentage of total lean and intermuscular fat (standard errors in parentheses) Chemical composition (YO of tissue) Brisket Jacobs ladder Fore rib Chuck Thin flank Shin and leg Clod and sticking Topside Loin, rump and fillet Thick flank and silverside Side Pistola Fore quarter Yo of total 10.29 (0.224) 9.43 (0.167) 3.88 (0.093) 14.80 (0.218) 7.88 (0.164) 5.32 10.14 (0.261) 8.08 (0.095) 15.26 (0.165) 14.92 (0.177) 100.00 (0.102) 45.56 (0.279) 54.44 (0.279) Lipid 26.1 (1.19) 18.2 (0.87) 23.0 (1.18) 13.2 (0.63) 19.5 6.2 (1.11) (0.71) 15.1 (0.83) 5.7 (0.36) 14.2 (0.68) 9.1 (0.40) 14.6 (0.58) 11.2 (0.44) 17.5 (0.77) Moisture 56.3 (0.94) 62.5 (0.78) 58.2 (0.96) 66.9 (0.55) 60.9 (0.88) 71.8 (0.51) 65.2 (0.75) 70.6 (0.29) 64.3 (0.52) 68.7 (0.29) 64.8 (0.48) 66.9 63.1 (0.66) (0.33) Ash 0.81 (0.023) 0.93 0.87 (0.05 1) 0.90 (0.040) 0.85 (0.036) 0.98 (0.028) 0.93 (0.0 16) 1.07 0.95 (0.0 15) 0.94 (0.036) 0.92 (0.01 6) 0.96 (0.017) 0.89 (0.018) (0.21) (0.21) Hydroxy- proline 0.31 0.30 (0.009) 0.26 (0.01 5 ) 0.30 (0.009) 0.33 (0.013) 0.44 (0.017) 0.35 0.21 0.25 0.26 0.29 (0.005) 0.26 (0.008) 0.32 (0.006) (0.0 10) (0.010) (0.0 10) (0.01 0) (0.01 1) Nitrogen 2.76 (0.049) 3.09 (0.033) 2.92 (0.042) 3.17 (0.030) 3.04 (0.045) 3.59 (0.038) 3.14 (0.034) 3.45 (0.033) 3.15 (0.026) 3.28 (0.027) 3.16 (0.023) 3.26 3.07 (0.029) (0.02 1 ) Fat-free nitrogen 3.74 (0.038) 3.78 (0.032) 3.80 (0.03 1) 3.65 (0.030) 3.77 (0.036) 3.82 (0.033) 3.70 (0.038) 3.66 (0.029) 3.67 (0.025) 3.61 (0.035) 3.70 (0.023) 3.67 (0.025) 3.72 (0.026)ANALYST, SEPTEMBER 1993, VOL.118 1225 Table 17 Lean and fatty tissues: chemical composition and content in each cull cow joint as a percentage of total lean and fatty tissues (standard errors in parentheses) Chemical composition (YO of tissue) Brisket Jacobs ladder Fore rib Chuck Thin flank Shin and leg Clod and sticking Topside Loin, rump and fillet Thick flank and silverside Side Pistola Fore quarter YO of total 10.60 (0.259) 9.21 (0.167) 3.85 (0.099) 14.08 (0.227) 8.39 (0.206) 5.34 (0.109) 9.90 (0.294) 7.90 (0.137) 16.15 (0.254) 14.58 (0.196) 100.00 46.00 (0.31 3) 54.00 (0.3 13) Lipid 31.2 (1.35) 22.3 (1.16) 28.4 (1.19) 15.9 (0.86) 27.4 (1.41) 10.4 (0.97) 18.8 (1.06) 10.5 23.9 13.2 (0.71) 20.1 (0.92) 17.6 (0.80) (1.33) (0.85) 22.2 (1.06) Moisture 52.3 59.3 (0.99) 54.0 (0.96) 64.7 (0.72) 54.9 68.1 (0.67) 62.3 (0.91) 66.9 (0.58) 56.9 (0.98) 65.4 (0.47) 60.6 (0.71) 61.9 (0.61) 59.4 (0.85) (1.02) (1.12) Ash 0.75 0.89 (0.01 9) 0.81 (0.046) 0.87 (0.038) 0.77 (0.03 1) 0.93 (0.025) 0.89 (0.0 17) 1.01 0.84 0.90 (0.034) 0.86 (0.016) 0.89 0.84 (0.01 9) (0.022) (0.021) (0.021) (0.0 18) Hydroxy- proline 0.34 0.30 0.26 0.31 (0 .008) 0.37 (0.0 17) 0.54 (0.023) 0.37 0.23 0.27 0.30 0.32 (0.006) 0.29 (0.009) 0.34 (0.006) (0.012) (0.008) (0.0 18) (0.010) (0.0 10) (0.01 0) (0.0 12) Nitrogen 2.58 (0.056) 2.94 (0.043) 2.72 (0.043) 3.08 (0.034) 2.76 (0.058) 3.49 (0.056) 3.02 (0.036) 3.31 (0.044) 2.83 (0.046) 3.17 (0.036) 2.97 (0.035) 3.06 (0.033) 2.91 (0.040) Fat-free nitrogen 3.76 (0.042) 3.80 (0.033) 3.80 (0.030) 3.66 (0.030) 3.80 (0.041) 3.90 (0.034) 3.72 (0.039) 3.70 (0.03 1) 3.73 (0.028) 3.66 (0.035) 3.73 (0.024) 3.72 (0.025) 3.74 (0.028) Table 18 Regression data for fat-free nitrogen in cull cow sides (YO of tissue) SFe* (%) Side weightlkg C15 Linear Quadratic Tissue Const ant x1 x2 x3 x4 Lean 3.583 0.0008 0.005 -0.001 -0.000 Intermuscular fat 3.910 -0.0022 -0.031 0.104 -0.006 Subcutaneous fat 6.543 - 0.0057 -0.023 -0.077 -0.000 -0.000 Lean with intermuscular fat 3.596 0.0007 0.003 0.004 Lean with fatty tissues 3.634 0.0006 0.002 0.005 -0.001 Fat-free nitrogen is given by the following expression: constant + (X, x side weight) + (X2 X C15) + (X3 x SFe) + (X, X SFe X SFe) Example: The fat-free nitrogen content of the lean with fatty tissues of a cull cow side of weight = 140 kg, conformation P (equivalent to C15 = 1) and fat class 4L (equivalent to SFe = 8.25) is given by: * SFe has both a linear and a quadratic term.3.634 + (0.0006 x 140) + (0.002 X 1.0) + (0.005 x 8.25) + (-0.001 x 8.25 x 8.25) = 3.693 Table 19 Fat-free nitrogen content for the tissues of each cull cow joint and the side (YO of tissue) Brisket Jacobs ladder Fore rib Chuck Thin flank Shin and Icg Clod and sticking Topside Loin, rump and fillet Thick flank and silverside Least significant difference* Side Pistola Fore quarter Lean 3.71 3.76 3.77 3.64 3.73 3.81 3.68 3.66 3.66 3.62 0.11 3.68 3.66 3.70 Intermuscular fat 4.01 3.99 4.13 3.86 4.27 4.04 4.02 3.82 4.04 3.48 0.57 3.97 3.84 4.04 Subcutaneous fat 4.21 4.92 4.83 5.31 4.22 5.43 4.56 6.19 5.86 5.91 0.86 5.06 5.71 4.50 Lean and intermuscular fat 3.74 3.78 3.80 3.65 3.77 3.82 3.70 3.66 3.67 3.61 0.12 3.70 3.67 3.72 Lean, inter- muscular and subcutaneous fat 3.76 3.80 3.80 3.66 3.80 3.90 3.72 3.70 3.73 3.66 0.12 3.73 3.72 3.74 * The smallest difference between any two joints that is statistically significant.1226 ANALYST, SEPTEMBER 1993, VOL.118 assumed that a single nitrogen factor is suitable for use with different lean-to-fat ratios.Recommendations On the basis of these results the Meat Factors Sub-committee makes the following recommendations, to be applied as appropriate. A nitrogen factor of 3.65, on a fat-free basis for the lean meat with intermuscular fat, is appropriate when applied to beef generally and should be used in the analysis of beef products. Nitrogen factors of 3.65 and 3.70, on a fat-free basis, should be used when applied to clean beef and cull cow beef, respectively. The factors shown in Tables 12 and 19 are applicable to specific joints and to sides of clean beef and cull cows, respectively, when the source of the beef is thus known. The data given in Tables 11 and 18 allow the calculation of the nitrogen factors for sides of clean beef and cull cows, respectively, when carcase weight and EC fatness and conformation classes are known.Changes to individual joints cannot be calculated from these data. Further advice can be obtained from the MLC. 1. 2. 3. 4. results obtained from the LCC and MLC laboratories. The participating laboratories were: ( i ) Unilever Research, Sharn- brook; (ii) Laboratory of the Government Chemist, Tedding- ton; (iii) Institute of Food Research, Bristol Laboratory; (iv) Avon County Council Scientific Services Laboratory, Bristol; and ( v ) Tayside Regional Council Laboratory, Dundee. Tissues from six cull cow carcases and eight clean beef carcases representing the three sexual types and a range of conforma- tion and fatness levels were each analysed by either LCC or MLC and by one external laboratory for lipid, moisture, nitrogen, ash and hydroxyproline contents. The difference was used to adjust all LCC and MLC analytical results so that they were centred around the mean of the interlaboratory calibration laboratories for all determinations. All adjust- ments were small. The Analytical Methods Committee gratefully acknowledges the financial support given by the Ministry of Agriculture, Fisheries and Food, and by the Meat and Livestock Commis- sion, to the work of this Sub-committee. 6 Appendix Statistical Analysis of the Difference Between the LCC and MLC Laboratories and the Data From the Five External Laboratories Five external laboratories, appointed by the Meat Factors Sub-committee and taken as a representative sample of all laboratories, were involved in the process of calibrating the 7 8 9 References Analytical Methods Committee. Analyst, 1963, 88, 422. Meat and Livestock Commission, Reef Yearbook, MLC, Milton Keynes, Buckinghamshire, 1991, p. 91. Kempster, A. J , , Cook, G. L., and Grantley-Smith, M., Meat Sci., 1986, 17, 107. Kempster, A. J . , Cook, G. L., and Smith, R. J., J. Agric. Sci., 1980, 95,431. British Standards Institution, BS4401: Analytical Methods for Meat and Meat Products, Part I : 1980, Determination of Ash; Part 2: 1980, Determination of Nitrogen, Part 3: 1980, Determi- nation of Moisture; Part 4: 1970, Determination of Total Fat; and Part 11: 1979, Determination of L-( -)-Hydroxyproline, BSI, Milton Keynes, Buckinghamshire. CEM AVC-80, Moisture and Fat Analyser, CEM Corp., Matthews, NC, USA. Meat and Livestock Commission, Beef Yearbook, MLC, Milton Keynes, Buckinghamshire, 1990, p. 113. Analytical Methods Committec, Analyst, 1952, 77, 543. Meat and Livestock Commission, Beef Yearbook, MLC, Milton Keynes, Buckinghamshire, 1991, p. 9. Paper 3102842 K Received May 18, 1993

ISSN:0003-2654    

DOI:10.1039/AN9931801217

出版商:RSC    

年代:1993

数据来源: RSC

摘要:

ANALYST, SEPTEMBER 1993, VOL. 118 COMMU NlCATlON 1227 Material for publication as a Communication must be on an urgent matter and be of obvious scientific importance. Rapidity of publication is enhanced if diagrams are omitted, but tables and formulae can be included. Communications receive priority and are usually published within 5-8 weeks of receipt. They are intended for brief descriptions of work that has progressed to a stage at which it is likely to be valuable to workers faced with similar problems. A fuller paper may be offered subsequently, if justified by later work. Manuscripts are usually examined by one referee and inclusion of a Communication is at the Editor's disc re tio n . Permeation Tubes for Calibration in Flow Injection Analysis Stuart J. Chalk and Julian F.Tyson" Department of Chemistry, University of Massachusetts, Amherst, MA, 01003, USA Don C. Olson FIA Solutions, P.O. Box 670786, Houston, TX, 77267, USA The use of a permeation tube for the production of liquid stream calibration standards in the flow injection determination of ammonia was investigated. By varying the flow rate from 0.5 to 4.0 ml min-1, calibration standards over the range 1.5-0.18 ppm could be produced. The relationship between concentration of the resulting solution and the reciprocal of the flow rate was shown to be linear. The inherent high temperature dependence of the release rate necessitated the tubes being used under thermostated conditions. The release rate was found to vary between 136 and 158 ng min-1 cm-1 over a 7 d period.Taking all sources of uncertainty into account, the 95% confidence interval for the release rate varied between 12,O and 16.2 ng min-1 cm-1 over the same 7 d period. The device was used in a manifold for the determination of ammonia in pond water in which the method of standard additions was employed. Keywords: Permeation tube; flow injection; calibration; ammonia determination Flow injection (FI) techniques have proved to be useful for the adaptation of a variety of determinations based on aqueous solution reaction chemistry to an automated format. There is considerable interest in the use of FI procedures as the basis of process analysers for on-line monitoring of liquid process streams and waste effluents. A feature of such analysers is the requirement for the production of suitable calibration stan- dards over the relatively long periods of unattended opera- tion.A number of innovative FI calibration procedures have been described in the past few years.'-10 Most of these procedures are based on the use of a single concentrated standard from which a set of calibration standards can be produced by a number of dilution procedures that are an inherent feature of the FI dispersion process. Various gradient concentration profiles have been exploitedl-8 for continuous dilution as well as, more recently, the use of a recirculating loop to perform controlled step-wise dilution.9.10 However, these techniques still require the production of a stock solution on a routine basis and their adaptation to the standard additions format (if needed) is not necessarily straightfor- ward.In this paper the use of an ammonia permeation tube for calibration of a solution spectrophotometric procedure has been investigated. For the calibration of procedures in which an analyte in a gaseous sample is determined, a range of gas permeation tubes are commercially available (Dynacal per- meation tubes, VICI Metronics, Santa Clara, CA, USA). The tubes contain a concentrated solution of the volatile analyte * To whom correspondence should be addressed. which diffuses, at a known rate, across the tube walls into an acceptor stream containing a diluent gas. The partial pressure of the analyte is inversely proportional to the acceptor stream flow rate. Thus changing the flow rate of the acceptor stream produces different concentrations of analyte.There are several reports of the determination of ammonia by FI procedures. Spectrophotometric,11-'3 diffusion14.15 and various other methodsl6-1s have been reported. In this paper a method based on the use of Nessler's reagent was employed.11 An important parameter for determining whether these devices are useful for on-line calibration is the day-to-day reproducibility of the release rate at a certain temperature. Ideally the device would be calibrated only once, at the beginning of its life. For release into a gaseous stream, VICT Metronics quotes +1% for the first 3-4 months. After that the stability is dependent on the dryness of the gas used and can vary by up to +5%. It is likely that over the lifetime of the tube (estimated as 7 months) stability will not be as good because of the release into aqueous solutions.Day-to-day reproducibility of the release rate was investigated along with the variation of release rate with temperature. Experimental Apparatus The manifold (see Fig. 1) was constructed from 0.8 mm i.d. Teflon tubing (Omnifit, New York, NY, USA), polyethylene ethyl ketone (PEEK) nuts and ferrules (UpChurch Scientific, Oak Harbor, WA, USA), and Teflon unions (Omnifit). A six-port rotary injection valve (PhaseSep, Norwalk, CT, USA) with 0.8 mm i.d. Teflon tubing connections was used for1228 ANALYST, SEPTEMBER 1993, VOL. 118 Reagents and Samples Nessler's reagent was made up and stored according to the procedure previously described .20 The potassium iodide, sodium hydroxide and mercury(i1) chloride were all of ACS analytical-reagent grade (Fisher Scientific) .Carrier stream and permeation tube stream solutions of 0.01 mol 1-1 nitric acid were made by diluting 0.629 ml of concentrated nitric acid (ACS analytical-reagent grade) to 1 1. Sodium hydroxide carrier stream (3 mol I-*) was made by dissolving 120 g of solid in 500 ml of E-pure water, transferred to a 1 1 flask and made up to volume. A 100 ppm stock solution of ammonia (as ammonium ion) was made by dissolving 0.3163 g of ammonium chloride (ACS analytical-reagent grade) in 250 ml of 0.01 moll-' nitric acid. Standards were made by appropriate dilution of the stock solution. De-ionized 18 MQ E-pure water was used throughout. A sample of water was taken from the University of Massa- chusetts campus pond and was acidified (pH 2) and filtered (0.45 pm) before use.P2 I 3 C P1 1 I ' I / R 410 nm Fig. 1 Schematic diagram of the manifold used for the determination of ammonia with Nessler's reagent. Components are as follows: variable speed peristaltic pumps, P1 and P2 (numbers are approxi- mate flow rates in ml min-1); injection port, I; 25 cm mixing coil, MC; detector, D; waste, W; carrier stream, C; Nessler's reagent, R; tube stream solution, TS; permeation tube, PT; 1.5 m heat exchanger coil, H; thermometer, T; and water-bath, WB. Pulse dampers used in each flow line are not shown sample injection. Reagent, carrier and permeation tube solutions were pumped using two Ismatec MS Reglo variable speed pumps (Cole-Parmer, Chicago, IL, USA) which were fitted with Tygon pump tubing (Cole-Parmer).The manifold was interfaced to the spectrometer using a 1 cm (18 PI) pathlength flow cell (Pye Unicam, Cambridge, UK). A Perkin-Elmer chemifold block (Perkin-Elmer, Norwalk, CT, USA) was used to merge the reagent and carrier stream. Pulse dampers were made for each line from a T-piece (Omnifit), a piece of Tygon pump tubing and a pump tube connector (Omnifit) which crimped the Tygon tube at one end. An ammonia, controlled release, Dynacal permeation tube with an active length of 5 cm (6 mm 0.d.) (VICI Metronics) was inserted into a 10 cm, 8 mm i.d. glass column (Omnifit) which was fitted with U4.28 end-fittings. The permeation device was allowed to rest against the wall of the tube on the rubber seals at each end.The nominal release rate into a gaseous acceptor stream at atmospheric pressure, according to the manufacturer's specifications, was 295 ng min-1 cm-1 (+lo%) at 30 "C. For temperature control a water-bath (Fisher Scientific, Pittsburgh, PA, USA) was used in conjunction with a 22-28 "C thermometer (Emil Greiner, New York, NY, USA) which was accurate to kO.01 "C. Real-time data acquisition of the reaction product peaks was obtained using a Lambda 6 UVNIS spectrophotometer (Perkin-Elmer). The instrument was interfaced to an IBM PS/2 (IBM, Armonk, NY, USA) running Perkin-Elmer Computerized Spectroscopy Software (PECSS) Version 3.26. A Macintosh SE computer (Apple Computer Inc. , Cupertino, CA, USA) was used for data evaluation. Lambda 6 .SP files were converted to .DX files in the PECSS software and then transferred onto Mac using Apple file exchange (Apple).They were then processed using Excel (Microsoft Corp., Redmond, WA, USA) (removal of header and footer) and imported as text into PeaksDemo (Analog Digital Instruments, Milford, MA, USA) for height/area evaluation. To aid peak identification a 29-point (optimized for this work) median filter recently described19 was implemented using Excel before files were imported into PeaksDemo. Each calibration and sample solution was injected five times. Solutions were de-gassed using nitrogen just prior to use as was the solvent used to make up the off-line standards. Any accumulated precipitate was flushed from the manifold by the passage of 2 mol 1-1 hydrochloric acid followed by E-pure water.Flow rates were calculated by timing the collection of the streams in 10 ml calibrated flasks. Method Development Reaction Chemistry Initial experiments were conducted with an acid-base indica- tor (Bromothymol Blue) reaction and without temperature control. The majority of the experiments were carried out using Nessler's reagent. Most of the experimental work was conducted with an acid acceptor stream flowing around the permeation tube. The effects of the variations of flow rate of the acceptor stream and of its temperature were investigated. Baseline Noise Reduction A number of methods for the reduction of baseline noise were investigated, including, the addition of pulse dampers, varia- tion of the mixing device (coil, packed-bed reactor, single bead string reactor, alternating helical reactor and stirred mixing chamber), use of a median filter for post data- acquisition smoothing, and the use of an alkaline acceptor stream.Calibration for Ammonia A series of calibration standards (0.2, 0.5, 1.0 and 2.0 ppm) were prepared off-line and five replicate injections of each solution were made. The solutions were prepared by dilution of the 100 ppm standard made up to volume with 0.01 mol I-' nitric acid and were prepared for each calibration of the tube. The slope of the calibration was calculated by an un- weighted least squares procedure €or data which did not include a blank value. Variation of Concentration With Flow Rate For four values of the flow rate of the acceptor stream, over the range from approximately 0.5 to 4.0 ml min-1, five replicate injections of the resulting ammonia solutions were made.The slope of the plot of concentration (obtained from the absorbance versus concentration of the standards pre- pared off-line) versus the reciprocal of the flow rate was calculated by an unweighted least squares procedure for data which did not include blank values. Calculation of the Release Rate The release rate of the permeation tube may be calculated from the slope of the concentration versus reciprocal flow rate plot, the length of the permeation tube and a conversion factorANALYST, SEPTEMBER 1993, VOL. 118 1229 for the mass [eqn. (l)]. This gives a result with units that can be compared directly to the manufacturer's nominal value of the release rate into a gas stream.Release rate (ng min-1 cm-1) = (1) calibration slope (pg min-1) x 1000 (ng pg-1) 5 (cm) Variation of the Release Rate With Temperature The change in release rate as a function of temperature may be calculated from eqn. (2) log (P2) = log (Pl) + 4 T 2 - T1) (2) where a is 0.034 "C-1, and P1 and P2 are the respective release rates at TI and T2. From the data supplied by the manufac- turer, it may be calculated that, if the release rate into a liquid acceptor stream is the same as that into a gaseous stream, the release rate would be 157 ng min-1 cm-1 at 22 "C. It can also be seen from eqn. (2) that temperature has a significant effect on the release rate. The release rate changes by approximately 11% for every 1 "C increase. Thus a water-bath was used to keep both the permeation tube and the acceptor solution at a steady known temperature for each experiment (see Fig.1). A temperature variation study was performed between 22 and 28 "C. Day-to-day Release Rate Variation An initial study over a period of 1 week showed possible problems. Five experiments performed on different days produced values varying from 140 to 240 ng min-1 cm-1. Modifications to the manifold were made by: ( a ) de-gassing the carriedtube solutions and the water used for the calibra- tion standards; (b) shortening, as much as possible, the tubing out of the water-bath to and from the pump; and ( c ) adding a heat exchanger coil (1.5 m) after the pump, immersed in the water-bath. This minimized any variation in the dissolved gas concentration in the carriedtube streams and eliminated the possible heating effect of the pump on the tube solution.Analysis of Pond Water Samples To test that the system worked with real samples, 500 ml of campus pond water were taken for analysis. Determination of the ammonia concentration was performed using both normal calibration and standard additions (acceptor flow rates of 0.5, 1.0,2.0 and 4.0 ml min-1 in both cases). Standard additions is readily performed using the permeation tube as the different additions can be generated by simply passing the sample over the tube at the different flow rates indicated. An unweighted least squares procedure was again used in the calibrations. Results and Discussion Reaction Chemistry Initial experiments, performed with nitric acid and Bromo- thymol Blue, gave peak heights which varied with flow rate as expected.However , significant non-zero intercepts were obtained. This and the non-selectivity of the reaction promp- ted a change to a quantitative chemical system with selectivity for ammonia. Nessler's reaction was chosen because it has a lower detection limit and faster reaction rate in comparison with the indophenol method." Baseline Noise Reduction Throughout the experimentation significant baseline noise was observed due to the viscosity difference between the Nessler's reagent and the acid carrier stream (which causes difficulties in obtaining rapid radial mixing downstream of the Table 1 Peak height versus concentration for ammonia solutions Concentration Peak height ( P P d (absorbance) 0.20 0.0235 0.50 0.0576 1 .oo 0.1170 2.00 0.2500 Table 2 Peak height versus reciprocal flow rate for an ammonia permeation tube ( M o w rate)/ Peak height min ml-1 (absorbance) 1.770 0.2700 0.898 0.1307 0.460 0.0710 0.238 0.0380 confluence point), and irregular generation and movement of bubbles through the flow cell. Indeed this is likely to be the reason for the large standard deviation (SD) of the individual results seen in Table 4 as the repeatability of injection of sample and standard solutions was typically poor (5-10%).An investigation of different mixing devices showed that removal of some of the noise could be achieved. However, this was directly related to an increase in the mixing device volume and thus also resulted in a significant decrease in the sensitivity of the manifold.In addition, the devices did not improve the repeatability of injections and the bubble problem was not removed. The 50 cm coil used in the experiment provided the best manifold stability albeit with a large amount of baseline noise (standard deviation approximately 0.01 absorbance). Replacing the carrier and acceptor solutions with 3 moll-' sodium hydroxide improved the repeatability of injection (3-5%). Also the sensitivity was increased by a factor of 2, due to the elimination of a decrease in the hydroxide concentration at the confluence point. No bubbles were seen which suggests that the bubbles seen in the acid manifold were due to de-gassing of the samples as they were basified at the confluence point (the pH of the resulting solution is over 13).Calibration for Ammonia A typical example calibration for ammonia (Table 1) had a slope of 0.1255 ppm-1 and an intercept of --0.005 with a correlation coefficient of 0.9993. No blank was included in the regression analysis. Variation of Concentration With Flow Rate Plotting peak height versus the reciprocal of the flow rate of the acceptor stream gave good linearity in all experiments (typical r2 = 0.999). Table 2 shows a typical calibration with a slope of 0.1513 ml min-1 and an intercept of 7 X 10-5. Using the peak heights from this calibration, concentrations of the solutions could be calculated from the off-line calibra- tion. From a plot of concentration versus the reciprocal of the flow rate (Fig. 2), it can be seen that the permeation tube shows the same behaviour with a liquid acceptor stream as it does with a gaseous one.Calculation of the Release Rate Taking the slope from the concentration versus reciprocal flow rate plot (1.20 pg min-1) the release rate of the permeation tube (and the associated error) can be calculated according to eqn. (1). In this example the release rate was found to be 240 k 23 ng min-1 cm-1.1230 ANALYST, SEPTEMBER 1993, VOL. 118 Variation of Release Rate With Temperature Table 3 shows the results of the release rate variation with temperature. As can be seen the logarithm of the release rate increases linearly with temperature change (compared with the 22 "C value) and the slope is close (0.051 "C-1, r2 = 0.990) to that obtained when using the tube with a gas stream (0.034 "C-1).Day-to-day Release Rate Variation Table 4 shows the results with this modified manifold configuration (Fig. 1) which was the final one used (carrier and acceptor solutions are 0.01 mol 1-1 nitric acid). These results are much more reproducible and the average release 0 1 (1Mlow rate)/min mi-' 2 Fig. 2 Concentration versus reciprocal flow rate for ammonia permeation tube, y = 0.0396 + 1.2043~; r2 = 0.998 rate (145 k 15 ng min-1 cm-1 at 95% confidence) is now closer to the value calculated from eqn. (2). Table 5 shows the results of the same manifold using 3 moll-' sodium hydroxide instead of 0.01 mol 1-1 nitric acid for the carrier and acceptor solutions. Surprisingly, the release rate (153 k 22 ng min-1 cm-1 at 95% confidence) is equally as stable with a basic acceptor stream as it is with an acidic one.Indeed, as mentioned before, the repeatability of injection was improved, resulting in the lower standard deviations of the calculated release rates. Analysis of Pond Water Samples Pond water ammonia concentrations using normal calibration were found to be 0.82, 0.69 and 0.71 ppm. Using standard additions the same sub-samples were found to have 0.57,0.55 and 0.57 ppm ammonia, respectively. This clearly shows the possibility of using these tubes for on-line standard additions calibrations. Conclusion The use of permeation devices for FI calibration has been shown to be a viable proposition provided certain precautions are taken. The most important of these are the stability and control of the temperature of the tube, and the amount of dissolved gases present in the solutions passing across the tube.However, other important considerations should be taken into account. The stability of the pump used has a direct impact on the concentration of the solutions produced. In these experiments, an eight-roller peristaltic pump was used and a pulse damper had to be installed to minimize pulsations caused by the rollers. Transfer line tubing from the end of the permeation tube column to the injection valve was made of Teflon, the same material as the permeation tube. To reduce any possible losses in this part of the manifold it would be advisable to use a material less permeable to gases, e.g., Table 3 Variation of release rate with temperature Release rate/ Calibration Tube solutions ng min-1 cm-1 Log (release TemperaturePC Slope SD Slope SD Value SD A T/"C rate) 22.25 0.1137 0.0034 0.0803 0.0043 141.2 8.7 0 2.150 24.13 0.1137 0.0034 0.1072 0.0069 188.6 13.4 1.88 2.276 26.01 0.1104 0.0052 0.1251 0.0049 226.6 13.9 3.76 2.355 27.97 0.1104 0.0052 0.1555 0.0050 281.7 16.0 5.72 2.450 Table 4 Variation of release rate over a period of days for 0.01 mol 1-1 nitric acid carrier and acceptor solutions (using the manifold in Fig.1) Release rate/ Calibration Tube solutions ng min-1 cm-1 Day TemperaturePC Slope SD Slope SD Value SD 1 22.08 0.1070 0.0059 0.0789 0.0047 147.5 12.0 2 22.10 0.1082 0.0079 0.0733 0.0044 135.5 12.8 3 22.01 0.1170 0.0068 0.0922 0.0055 157.6 13.1 4 22.01 0.1105 0.0095 0.0773 0.0060 139.9 16.2 Table 5 Variation of release rate over a period of days for 3 mol 1-1 sodium hydroxide carrier and acceptor solutions (using the manifold in Fig.1) Release rate/ Calibration Tube solutions ng min-1 cm-1 SD 1 22.06 0.2160 0.0051 0.1684 0.0103 156.0 10.2 2 22.08 0.2016 0.0065 0.1726 0.0061 171.3 8.2 3 22.01 0.2046 0.0056 0.1447 0.0046 141.4 5.9 142.4 12.1 4 22.01 0.1988 Value Day Temperature/"C Slope SD Slope SD 0.0070 0.1415 0.0110ANALYST, SEPTEMBER 1993, VOL. 118 1231 PEEK tubing. The device may also be useful for standard additions calibrations, and it has a wide tolerance to the pH (2-14) of the acceptor stream. Clearly this approach is only applicable to those analyte species for which a suitable permeation tube is available. The manufacturer’s current list of approximately 110 species includes a number that are of relevance to industrial processes such as S2-, SO2, HCN methanol and ethanol.As a variety of tube lengths are available, the release rate may be selected to allow calibration over a wide range of concentrations. Financial support by the Shell Development Company is gratefully acknowledged and Dr. Stephen Bysouth is thanked for helpful discussions. References Muller, H., and Kramer, J., Fresenius’ 2. Anal. Chem., 1989, 335, 205. Muller, H., and Kramer, J., Fresenius’ 2. Anal. Chem., 1989, 335,210. Yang, J., Ma., C . , Zhang, S., and Shen, Z., Anal. Chirn. Acta, 1990, 235, 323. Fan, S. H., and Fang, Z . L., Anal. Chim. Acta, 1990, 241, 15. Sperling, M., Fang, Z., and Welz, B., Anal. Chem., 1991, 63, 151. Baron, A., Guzman, M., RGiiCka, J., and Christian, G. D., Analyst, 1992, 117, 1839. MacLaurin, P., and Worsfold, P. J., Microchem J., 1992, 45, 178. Starn, T. K., and Hieftje, G. M., J. Anal. At. Spectrom., 1992, 7, 335. 9 10 11 12 13 14 15 16 17 18 19 20 Agudo, M., Rios, A., and Valcarcel, M., Anal. Chirn. Acta, 1992, 264,265. Tyson, J. F., Bysouth, S. R., Grzeszczyk, E. A., and Debrah, E., Anal. Chirn. Acta, 1992, 261,75. Stewart, J. W. B., RGiiEka, J., Bergamin Filho, H., and Zagatto, E. A. G., Anal. Chirn. Acta, 1976, 81, 371. Krug, F. J., RBiiEka, J., and Hansen, E. H., Analyst, 1979,104, 47. Bergamin Filho, H., Reis, B. F., Jacintho, A. O., and Zagatto, E. A. G., Anal. Chirn. Acta, 1980, 117, 81. Van Son, M., Schothorst, R. C., and Den Boef, G., Anal. Chirn. Acta, 1983, 153, 271. Nakata, R., Kawamura, T., Sakashita, H., and Nitta, A., Anal. Chirn. Acta, 1988, 208, 81. Mikasa, H., Motomizu, S., and Toei, K., Bunseki Kagaku, 1985, 34, 518. Ishibashi, N., and Imato, T., Fresenius’ 2. Anal. Chern., 1986, 323, 245. Hara, H., Motoike, A., and Okazaki, S., Analyst, 1988, 113, 113. Moore, A. W., Jr., and Jorgenson, J. W., Anal. Chern., 1993, 65, 188. Bassett, J., Denney, R. C., Jeffrey, G. H., and Mendham, J., in Vogel’s Textbook of Quantitative Inorganic Analysis, Longman, London, 4th edn., 1983, p. 731. Paper 31038001 Received July 1, 1993 Accepted July 27, 1993

ISSN:0003-2654    

DOI:10.1039/AN9931801227

出版商:RSC    

年代:1993

数据来源: RSC

摘要:

ANALYST. SEPTEMBER 1993. VOL. 118 1233 CUMULATIVE AUTHOR INDEX JANUARY-SEPTEMBER 1993 Aarkrog, Asker, 1101 Aboal-Somoza, Manuel, 665 AbramoviC, Biljana F., 899 Adams, Michael J., 229 Akasaka, Kazuaki, 765 Alder, J. F., 395 Allag, Houssein, 401 Al-Masri, M. S . , 873 Alwarthan, Abdulrahman A., Amarasiriwaradena, Chitra, Analytical Methods Committee, Anderson, David R., 449 Andrew, B. E., 153 Andrews, William J., 425 Aoki, Nobumi, 909 Armstrong, Fraser A., 973 Arrigan, Damien W. M., 355 Ashok Kumar, T., 293 Avidad, Ramiro, 303 Bae, Yea-Ling, 297, 301 Baj, Stefan, 1081 Banerjee, Arun B., 937 Bangar Raju, G., 101 Bannon, Thomas, 361 Barclay, David A., 245 Barjat, Herv6, 73 Barker, Philip G., 347 Barnard Howie, Judith A., 35 Barnes, Ramon M., 1175 Bartle, Keith D., 737 Bartlett, Philip N., 371 Banvick, Ian M., 489 Baxter, Douglas C., 495, 1007 Bayo, Javier, 171 Beauchemin, Diane, 815 Bell, Jimmy D., 241 Belton, Peter S ., 73 Benmakroha, Farida, 401 Beone, T., 979 Bermejo Martin-Lazaro, A., 917 Bermejo-Barrera, Pilar, 665 Bernal Suarez, M. M., 917 Bhaskar, Nilam, 1 Biondi, Cinzia, 183 Birmingham, John. J., 1 Blackburn, R., 873 Blaih, Salah M., 577 Blair, Neil, 371 Bond, Alan M.. 973 Boriitsky, Juri A., 859 Bos, Albert, 323 Bos, Martinus, 323 Boudjerda, Tarik, 401 Boufenar, Rabah, 401 Bovara, Roberto, 849 Bradbury, Michael W. B., 533 Brand, Tom L., 1101 Breen, William, 415 Brereton, Richard G., 779 Brienza, Sandra Maria Boscolo, Brinkman, Udo A. Th., 11 Brossier, Pierre, 1021 Brough, Paul A., 753 Brown, Marc B., 407 Bruns, Roy E., 213 Biichi, Felix N., 973 Budong, Wang, 1213 Bui, Liin N., 463 Cai, Xiaohua, 53 Calokerinos, Antony C., 627, Campanella, L., 979 Campiglio, Antonio, 545 Canela, Ramon, 171 Capitan-Vallvey, Luis Fermin, Carlson, Robert G., 257 Carrea, Giacomo, 849 Casey, Vincent, 389 639 1175 1089, 1217 719 633 303 Cassidy, John F., 415 Catterick, Timothy, 791 Celeste, M., 895 Cepas, Juana, 923 Cerda, V., 895 Cermhk, Josef, 79 Chadima, Radko, 79 Chaisuksant, Rasamee, 179 Chalk, Stuart J., 1227 Chan, Wing-Hong, 863, 869 Chang, Qing, 839 Chartier, A., 157 Chatergoon, Lutchminarine, 947 Chen, Liang, 277 Chen, Qiang, 1131 Chen, Rong-mei, 1055 Chen, Tianyu, 541 Chokshi, Hitesh P., 257 Chowdhury, Bhabadeb, 937 Chun-Xiang, He, 1077 Cladera, A., 895 Clark, Alastair, 601 Clarke, Colin G., 229 Clifford, Anthony A., 737 Clinton, Cathriona, 415 Comber, Sean, 505 Cooke, Michael, 449 Cooper, Andrew I., 1111 Corbini, Gianfranco, 183 Cordero, Bernard0 Moreno, 209 Corti, Piero, 183 Costa Garcia, Agustin, 649 Cotsaris, Evangelo, 265 Crane, Michael, 617 Crean, G.M., 429 Crooks, Steven R. H., 447 Crosby, Neil T., 489 Crouch, Stanley R., 695 Crubellati, Ricardo O., 529 Cruz Ortiz, M., 801 Csizkr, Eva, 609 Cummins, Diane, 1 Cummins, Phillip G., 1 Cunningham, K., 341 Daenens, P., 137 Dams, Richard, 1015 Danielsson, Bengt, 845 Das, Sulobh K., 1153 Davis, Willard E., 249 Day, J. Philip, 1101 Dazhong, Shen, 1143 de Andrade, Jog0 Carlos, 213 de la Guardia, Miguel, 23, 1043, de la Rosa, Francisco F., 643 de Paula Eiras, Sebastigo, 213 Dean, John R., 747 Deasy, Brian, 355 Debrabandere, Lode, 137 Deftereos, Nikolaos T., 627 Delves, H.Trevor, 533 Dempsey, Eithne, 411 Dhaneshwar, Ramesh G., 1153 Diamond, Dermot, 347,1127, Diaz, Susana, 171 Diewald, Wolfgang, 53 Djerboua, Ferhat, 401 Domanskf, Karel, 335 Dominguez, Lucas, 171 Dominguez-Gonzalez, Raquel , dos Reis, Boaventura Freire, 719 Dowle, Chris J., 17 Dreassi, Elena, 183 Duran-Meras, Isabel, 807 Dvinin, Alexei, 859 Economou, Anastasios, 47 Edmonds, Tony E., 407,443 Efstathiou, Constantinos E., 627 Egan, Denise A., 201, 411 Eggins, Brian R., 439 Elliott. ChristoDher T.. 447 1167 1131 665 El-Yazbi, Fawzy A., 577 Emteborg, HBkan, 1007 Escobar, Rosario, 643 Espinosa-Mansilla, Anunciacion, 89, 807 Estela, J. M., 895 Fabre, H., 1061 Faure, Uta, 475, 481 Fearn, Tom, 235 Fenyan, Zhang, 1213 Fernandez Laespada, M".Esther, 209 Ferri, Elida, 849 Feygin, Ilya, 281 Fieldcn, Peter R., 47 Finglas, Paul M., 475, 481 Fitzgerald, Catherine, 361 Flaherty, T., 429 Fogg, Arnold G., 1157, 1163 Foster, Robert, 415 Fox, C. G., 157 Fraidias Becerra, Antonio J., Frech, Wolfgang, 495, 1007 Friel, Sharon, 371 Frutos, G., 59 Fu, Chengguang, 269 Gaal, Ferenc F., 899 Gaind, Virindar S . , 149 Gallagher, Timothy, 753 Gallego, Mercedes, 1199 Gallegos, Robyn Dahl, 1137 Gallignani, Maximo, 1043, 1167 Gangadharan, S . , 1085 Gangemi, Gianni, 849 Garcia Gomez de Barreda, Garcia-Lopez, Trinidad, 303 Garcia-Mesa, J. A., 891 Gardner, Julian W., 371 Garrigues, Salvador, 1043, 1167 Georges, J . . 157 Gcrakis, Athanasios M., 1001 Ghijsen, Rudy T., 1 I Ghini, Severino, 849 Gibney, Patrick M., 425 Gilner, Danuta, 1081 Giosuk, Maria Antonietta, 849 Girotti, Stcfano, 849 Givens, Richard S ., 257 Glennon, Jeremy D., 355 Godovski , D. Yu., 997 Gomez-Hens, Agustina, 707 Goodfellow, Brian J., 73 Gorog, Sandor, 609 Goto, Katsumi, 1205 Grau, Harald, 689 Greenfield, Stanley, 443 Greenway, Gillian, 17 Gregory, Donald P., 1 Grob, Robert, 1 1 Gu, Xiao-Hong, 863 Gu, Zhi-cheng, 105, 1055 Guiraum, Alfonso, 643 Haegel, Franz-Hubert , 703 Halbig, Peter, 689 Halls, David J., 821 Halvatzis, Stergios A . , 633 Hamano, Takashi, 909 Hamnett, Andrew, 973 Han, Chuan-qiang, 1055 Hanai, Toshihiko, 769, 773 Hara, Hirokazu, 549 Harris, S. J., 341 Harris, Stephen J., 1127 Hartnett, Margaret, 347 Haswell, Stephen J., 245 Hata, Noriko, 1205 Hauser, Peter C., 991 Hawkesworth, K., 395 Hawkins, Peter, 35 Hembree, Jr., Doyle M., 249 Hernandez Garcia..I.. 917 175 Daniel, 175 Hidalgo de Cisneros, Jos6 L. Hill, H. Allen O., 973 Hiraidc, Masataka, 537 Hirose, Tsuyoshi, 517 Hodgkinson, Mark, 1049 Hokari, Norihisa, 219 Hollman, Pctcr C. H., 475, 481 Hornig, Jamcs F., 033 Horvith, KornClia S . , 899 Hosokawa, Satoshi, 1205 Howard, Vyvyan C., 1 Howdle, Steven M., 11 11 Huang, Ka-Lin. 205 Hughes, Catherine, 11 1 1 Hunt, Tcrcncc P., 17 Idriss, Kamal A., 223 lizuka. Ryuji, 165 Imai, Kamhiro, 759 Inui, Syn-ya, 1031 Ishibashi, Mumio, 759 Ishida, Junichi, 165 Ishida, Ryoei, 1071 Ito, Yoshio, 909 Ivaska, Ari, 885 Iwachido, Tadashi, 273 Iwata, Tetsuharu, 517 lyer, R. M., 929 Izquierdo, Pilar, 707 Itumi, Sigcru, 553 Jan, Ching-Ching, 1183 Janata, Jiii, 335 Jedrzejczak, Kazik, 149 Jefferics, Terry M., 753 Jianli, Liu, 1213 Jobling, Margaret, 11 1 I Johnston, Brian, 355 Johnston, Kcith P., 1 1 11 Jones, Carol L., 1 Josowicz, Mira, 335 Ju, Dowcon, 253 Kalcher, Kurt, 53 Kallury, Krishna M.R., 309 Kalman, Pctcr G., 463 Kanc, Jean S . , 953 Karayannis, Miltiades I., 7 1 1, Kasahai-a, Issei, 1205 Kasumimoto, Hanae, 131 Katz, Stanley E., 281 Kawaguchi, Hiroshi. 537 Kazarian, Sergci G., 11 11 Kcsslcr, Margalith, 235 Keycs, Emmetine T., 385 Khayyami, Masoud, 845 King, Bernard, 587 Kinoshita, Toshio, 161, 769, 773 Kiranas, Efstratios R., 727 Kiss, Attila, 661 Kobayashi, Atsushi, 273 Kobayashi, Shouichi, 131 Koqak, Ali, 657 Koh, Tomozo, 669 Kojima, Nobuaki, 909 Kokot, Serge, 1049 Koltypin, E. A., 997 Konidari, Constantina N., 71 1 Konig, Monika, 703 Koshino, Yukihiro, 827.1027 Kostov. Yordan, 987 Kotrly, Stanislav, 70 Koupparis, Michael A., 100 1 Kovanic, Psvel, 145 Krishan Puri, Ral, 85 Krushevska, Antoaneta, 1175 Kubal. Gina, 241 Kulicki, Zdzislaw, 1081 Kulkarni, Achyut V., 1 153 Kumar, Manjeet, 103 Kumar, Sanjiv, 1085 Kundu, Dipali, 905 Kvalheim, Olav M.. 779 Lan, Chi-Rcn, 189 Hidalgo, 175 723, 7271234 ANALYST, SEPTEMBER 1993, VOL. 118 Lang, Mark J., 425 Lannon, A. Martin, 973 Larsson, Per-Olof, 845 Lauko, Anna, 609 Lazaro, Fernando, 1193 Lcdesma, Aricl G.. 529 Ledingham, Kenneth W. D., 601 Lee, Albert Wai-Ming, 869 lcGras, Christopher A. A., 1035 Lettington, Olwen C., 973 Lev, Ovadia, 557 Li, Ronghua, 563 Li, Xiang-Ming, 289 Liang, Wei-An, 97 Liang, Yi-Zeng, 779 Lihua, Nie, 1143 Lin, Qingxiong, 643 Lin.Yuche, 277 Littlejohn, David, 541,821, 1065 Lopez Palacios, Jesus, 801 Lopez Ruiz, B., 59 Lowdon, John, 747 Lowe, Roger D., 613 Lu, Bing-liang, 1055 Lu, Jianmin, 1131 Lunar, Ma Loreto, 1209 Lunar, Maria Loreto, 715 Lunte, Susan M., 257 Luque dc Castro, Maria Dolores, 593, 891 Lyons, Cormac H., 361 Lyons, Michael E. G., 361 Mc Monagle, James B., 389 McArdle, Fiona A., 419 McCallum, John J., 401 McCarrick, Mary, 1127 McCaughey, William J., 447 McClean, Stephen, 511 MacCraith, Brian D., 385 McDonagh, Colette M., 385 MacDonald, Robert C., 913 McEvoy. John D. G., 447 McGilp, John F., 385 MacKay, Graham A., 741 McKeown, Neil B., 463 McKervey, M. A., 341 MacLaurin, Paul, 617 McLeod, Cameron W., 449 Magee, Robert J., 53 Maguire, Michael, 1107 Maillols, H., 1061 Malone, Michael A., 649 Mandrou, B., 1061 Maquieira, Angel, 855, 1193 Marshall, Archibald, 601 Martelli, Patricia Benedini, 719 Martin, J.P., 59 Martinez-Lozano, C., 567 Martinez-Vado, Annabelle, Masotti, Piero, 849 Mathieu, Jacques, 11 Matsubara, Chiyo, 553 Meguro, Hiroshi, 765 Mellidis, Antonios S., 179 Mertens, Bart, 235 Midgley, Derek, 41 Miller, James N., 407, 455 Miller, Richard M., 1 Mills, Andrew, 839 Mitsuhashi, Yukimasa, 909 Miura, Yasuyuki, 669 Mizuno, Takayuki, 1031 Mohan, Hari, 929 Monks, Cheryl D., 623 Moreno, Miguel A., 171 Mori, Yuichi, 553 Moriyama, Youichi, 29 Moskvina, M. A., 997 Moss, Martin C., 1 Mottola, Horacio A., 675 Moulder, Robert, 737 Munakata, Toshihide, 1071 Muiioz de la Pefia, Arsenio, 807 Muiioz Leyva, Juan A., 175 Munro, C.H., 731 1043 Nabekura, Tomiko, 273 Nacapricha, Duangjai, 623 Nagahiro, Tohru, 85 Nakagawa, Genkichi, 219 Nakai, Chic, 769, 773 Nakamura, Kayoko, 29 Nakamura, Masaru, 517 Nan, Zhou, 1077 Nanos, Christos G., 711 Narayanaswamy, Ramaier, 317 Narukawa, Akira, 827, 1027 N a v a h , Alberto, 303 Neuhold, Christian, 53 Ni, Yongnian, 1049 Niazi, Shahida B., 821 Nicholson, Brenton C., 265 Nickel, Ulrich, 689 Nimura, Noriyuki, 161, 769,773 Norman, Philip, 617 Nukatsuka, Isoshi, 1071 Nwosu, Titus, 845 O’Beirn, Brendan, 389 O’Donoghue, Eilish, 415 Ohkubo, Hiromi, 549 Ohrui, Hiroshi, 765 Ohta, Kiyohisa, 1031 Ohzeki, Kunio, 1071 Oji, Yoshikiyo, 909 Okabe, Katsuaki, 669 Okada, Tetsuo, 959 O’Kane, Edward, 511 O’Keeffe, Gerard, 385 O’Kelly, Brendan, 385 O’Kennedy, Richard, 201,411 Olson, Don C., 1227 O’Neill, Robert D., 433 O’Sullivan, Ciara, 411 Oughton, Deborah H., 1101 Palaniappan, R., 293 Pandey, Amita, 941 Papageorgiou, Vassilios P., 179 Pasha, Akmal, 777 Paukert, TomaS, 145 Paynter, J., 379 Pearce, Timothy C., 371 Perez Pavon, Jose Luis, 209 Perez-Bendito, Dolores, 707, 715, 923, 1209 Perez-Ruiz, T., 567 Peris Cardells, Empar, 23 Persaud, Krishna C., 419 Petelenz, Danuta, 335 Petrukhin, Oleg M., 859 Pinatel, Henri, 831 Pitre, Krishna S ., 65 Planta, Milagros, 1193 Poliakoff, Martyn, 1111 Pramauro, Edmondo, 23 Preston, Gaynor, 245 Prevot, Alessandra Bianco, 23 Prieta, Javier, 171 Proietti, Daniela, 183 Puchades. Rosa, 855, 1193 Pyo, Dongjin, 253 Qi, Kang, 1143 Quencer, Brett M., 695 Quintela, Ma Jos6, 1199 Radulovic, Stojan, 241 RadunoviC, Aleksandar, 533 Rahmani, Ali, 779 Ramachandran, Venkataraman Ramesh, A., 945 Rauch, Pavel, 849 Raurich, Josep Garcia, 197 Reckhow, David A., 71 Reed, Peter I ., 877 Reid, Helen J., 443 Remy, Isabelle, 1021 Repasi, Janos, 661 Reviejo, A. Julio, 1149 Riley, David P., 407 Roda, Aldo, 849 Roe, Merrion P., 425 Romaschin, Alex D., 463 Roy, S. K., 905 N., 511 Ruan, Fu-Chang, 289 RubeSka, Ivan, 145 Rubio Leal, Amparo, 89 Rubio, Soledad, 715, 1209 RGiiCka, Jaromir, 885 Sabot, Jean-Frangois, 831 Sadler, Peter J., 241 Saez, Jose A., 801 Salbu, B., 1101 Saleh, Magda M. S., 223 Salinas, Francisco, 89, 807 Salvatore, Michael J., 281 Sammartino, M. P., 979 Sanchis, Vicente, 171 Sander, Joseph, 601 Santana Rodriguez, J. J., 917 Sanyal, Asis K., 937 Sanz, A., 567 Sanz Pedrero, P., 59 Satake, Masatada, 85 Savarino, Piero, 23 Sawai, Kaori, 549 Schneider, Jeffery A,, 933 Schneider, Siegfried, 689 Schwuger, Milan Johann, 703 Scott, Steven, 1117 Seare, Nichola J., 407 Selby, Mark, 1049 Shallow, A., 429 Sharma, Devender K., 941 Shepherd, Lindsey A., 1111 Sheppard, Robert C., 1 Shibata, Masaru, 909 Shimoishi, Yasuaki, 273 Shiu, Kwok-Keung, 863, 869 Shortt, Desmond H., 447 Shouzhuo, Yao, 1143 Silva, Manuel, 681, 923 Simpson, Michael, 449 Singhal, Ravi P., 601 Singleton, Scott, 1 Slangen, Jean H., 475, 481 Slater, Jonathan M., 379 Smith, Clayton, 947 Smith, Roger M., 741 Smith, W.E., 731 Smyrl, Norman R., 249 Smyth, Malcolm R., 411, 649 Smyth, W. Franklin, 511 Snook, Richard D., 613 Somer, Guler, 657 Song, Lin, 1143 Soto-Ferreiro, Rosa M., 665 $outhgate, David A.T., 475,481 Sramkova, Jitka, 79 Srivastava, P. K., 193 Stalikas, Constantine D., 723 Stewart, Jr., Charles W., 1123 Su, Hongbo, 309 Suarez, Guillermo, 171 Subbarao, Nanda K., 913 Suliman, Fakhr Eldin O., 573 Sultan, Salah M., 573 Sun, S. W., 1061 Svehla, Gyula, 341, 355 Svendsen, C. N., 123 Swadesh, Joel K., 1123 Taguchi, Shigeru, 1205 Takamura, Kiyoko, 553 Tan, Susie S. S., 991 Tang, Gui-Na, 205 Taniguchi, Hirokazu, 29 Tanweer, Ahmad, 835 Taylor, Colin G., 623 Teasdale, P. R. , 329 Terao, Tadao, 759 Thakur, Hari K., 941 Thompson, Michael, 309, 463 Thompson, Michael (Birkbeck), 235, 1107 Timotheou-Potamia, Meropi M., 633 Tomas, C., 895 Tomas, V., 567 Tomassetti, M., 979 Torrades, Francesc, 197 Torro, Luis, 855 Toyo’oka, Toshimasa, 257, 759 Tsai, Suh-Jen Jane, 297,301, Tsionsky, Michael, 557 Tsuzuki, Wakako, 131 Tucker, Alan, 241 Tuiion Blanco, Paulino, 649 Tyson, Julian F., 1227 Tzonkov, Stoyan, 987 Tzouwara-Karayanni, Stella M., Uchida, Takaaki, 537 Uden, Peter C., 1123 Urusov, Yuri I ., 859 ValcArcel, Miguel, 593,891,1199 Van Allemeersch, Franqoise, Van Boven, M., 137 van der Linden, Willem E., 323 Vanhoe, Hans, 101.5 Veiro, Jeffrey A., 1 Verma, Balbir C., 941 Verma, Neerja, 65 Verma, Rakesh, 1085 Versieck, Jacques, 1015 Vijaya Raju, K., 101 Vijayashankar, Yadathora N., Vilchez, JosC Luis, 303 Viscardi, Guido, 23 Volkov, A. V., 997 Vos, Johannes G., 385 Voulgaropoulos, Anastasios, Wada, Hiroko, 219 Wagstaffe, Peter J., 475, 481 Wallace, G. G., 329 Walton, Philip W., 425 Wang, Bao-Ning, 205 Wang, Joseph, 277,411, 1131, Warwick, Peter, 489 Watt, E. J., 379 White, Peter C., 731, 791 Whiting, Robin, 947 Williams, David M., 249 Williams, Kathleen E., 245 Winefordner, James D., 1031 Wong, Kwok-Yin, 289 Wong, Wai-Cheong, 869 Worsfold, Paul J., 617 Worswick, Richard, 583 Wuchner, Klaus, 11 Xie, Bin, 845 Xie, Yuefeng, 71 Xu, Guoping, 877 Xu, Hongda, 269 Yamaguchi, Masatoshi, 165,517 Yamanaka, Nobuhiro, 1031 Yamauchi, Shuji, 161, 769, 773 Yan, Hsiao-Tzu, 521 Yang, Mengsu, 309 Yoshida, Tomohiko, 29 Yoshioka, Hiroshi, 553 Yotova, Ljubov, 987 Yuchi, Akio, 219 Zagatto, Elias Ayres Guidetti, Zanoni, M. Valnice B., 1157, Zawadiak, Jan, 1081 Zenki, Michio, 273 Zhang, D., 429 Zhang, Peixun, 1065 Zheng, Minghui, 269 Zhou, Jie, 97 Zhou, Zhauro, 563 Zhu, Zhong-Iiang, 105, 1055 Zoski, Cynthia G., 973 Zotou, Anastasia, 753 Zou, Shi-Fu, 97 521, 1183 723, 727 1015 777 179 1149 719 1163

ISSN:0003-2654    

DOI:10.1039/AN9931801233

出版商:RSC    

年代:1993

数据来源: RSC