3. AOACRIFeedsFertilzerMethods-2018Awards

33

and when an interferent is present in a particular line, the result for that line is omitted from the mean value reported. ( 2 ) Some ICP software has the capability of mathematically modeling potential interferents and deconvoluting the instrument response into an analytical element portion and an interferent portion. ( 3 ) Interelement correction is an alternative mathematical technique to use with instruments for which mathematic modeling is not available, or where direct spectral overlap negates use of the deconvolution technique. The following lines, if used, must utilize one of the correction techniques; corrections for other wavelengths may be applied as needed and appropriate: ( a ) As 188.980: Correct for Cr interference, or verify that Cr is not present at an interfering level in the test portion analyzed. ( b ) As 193.696: Fe affects the arsenic peak. Remove with an Fe model, or verify that Fe is not present at an interfering level in the test portion analyzed. ( c ) Cd 214.439 and 226.502: Fe, present in many fertilizers, interferes with both suggested Cd wavelengths. Mathematically correct instrument Cd response for the interference, or verify analytically that Fe is not present at an interfering level in the test portion analyzed. ( d ) Pb 220.353: Mathematically correct instrument Pb response for Fe interference, or verify that Fe is not present at an interfering level in the test portion analyzed. ( e ) Se 196.026: Mathematically correct instrument Se response for Fe interference, or verify that Fe is not present at an interfering level in the test portion analyzed. ( d )  ICP instrument calibration .—Prepare Group A working standard solutions from 1000 mg/L commercial stock standards. Custom blended multielement stock standard in an acid ratio (9% HNO 3 :3% HCl) is acceptable. Working standards should be prepared at concentrations listed in Table  2017.02B , if they fit the sensitivity of the available instrumentation. Calibration concentrations should be adjusted to match the sensitivity of an instrument. However, linear curves should have correlation coefficients of at least 0.999 and a standard error of no more than 10%. Quadratic calibrations should have a correlation coefficient of at least 0.999, a standard error of less than 10%, a curvature of no more than 25%, and an upward curvature of no more than 400%. G. Reagents (Option 2): Dual Acid Digestion ( a )  Water .—18 Megaohm water. ( b )  HNO 3 .—Trace metal grade HNO 3 (nitric acid – HNO 3 , 67 – 70%, OmniTrace grade; EMD Chemicals). ( c )  HCl .—Trace metal grade HCl (hydrochloric acid – HCl, 35 – 38%, trace metal grade; Cat. No. A508-500, Fisher Scientific, Pittsburgh, PA, USA). ( d )  Triton X-100 solution .—Triton X-100 .— Octyl phenol ethoxylate (J.T. Baker Chemicals, Center Valley, PA, USA). ( e )  0.5% Triton X-100 solution .—Dilute 0.5 mL Triton X-100, G ( d ), to 100 mL with H 2 O, G ( a ). ( f )  Cesium chloride .—Formula weight 168.36, trace metal basis, purity >99.999%, Cat. No. 203025-50G (Sigma-Aldrich, St. Louis, MO, USA). ( g )  1000 mg/L Sc standard .—In 4%HNO 3 , Product No. 100048-1 (High Purity Standards). ( i )  Ionization buffer/internal standard solution .—Weigh 8.0 g CsCl, G ( f ), into a 1000 mL acid-washed volumetric flask. Add 3 mL each of ICP grade scandium, G ( g ), and beryllium, G ( h ), 1000 mg/L stock solution, as internal standards. Also add 1 mL of 0.5% Triton X-100, G ( e ), dilute to volume, and mix. Store in a polypropylene bottle. ( Note : Reagent concentrations assume the use (High Purity Standards, Charleston, SC, USA). ( h )  1000 mg/L Be standard .—In 4% HNO 3 , Product No. 1005-1

Table 2017.02A. Recommended inductively coupled plasma- optical emission spectrometry wavelengths for Group A and B

metals Element

Wavelength(s)

188.980 a , 193.696 a

As Ca Cd Co Cu Fe Mg Mn Mo Cr

183.944, 318.127, 430.253 214.439 a , 226.502 a , 228.802 228.615, 230.786, 258.033 205.560, 267.716, 276.653

217.895, 222.778, 324.754, 327.395 234.350, 238.204, 240.489, 259.837 277.983, 278.297, 285.213, 383.829 260.568, 261.815, 263.817, 293.931

202.032, 204.598

Ni

216.555, 222.486, 231.604

Pb Se Zn

220.353 a 196.026 a

206.200, 213.857, 334.502, 472.215

a  Wavelengths with potential spectral interference.

buffer/internal standard solution of 8000 mg/kg cesium chloride and 3 mg/kg scandium and/or beryllium internal standard(s); and pump tubes of white/white (1.02 mm id) sample and orange/white (0.64 mm id) internal standard, the white/white contributing about 72%, and the orange/white contributing about 28%, to the final nebulized solution.] All analytical wavelengths should be corrected using an internal standard wavelength. However, best practice is to utilize similar transitions between analyte and internal standard. For example, the 188.980 wavelength is from arsenic in the atomic state, so the internal standard wavelength used for correction should also be from the atomic state, such as Sc 361.383. Conversely, match ionic sample lines with ionic internal standard lines. ( Note : Do not use yttrium as an internal standard, since it is found native at low levels in some phosphate ore sources.) ( b )  ICP wavelengths .—A number of wavelengths may be used for analysis of the elements of interest, depending on the capability of the analytical instrument used. At a minimum, select at least two wavelengths for each element of interest, and report the averaged value of closely agreeing results, with the exception of lead and selenium, for which there is only one reliable wavelength available. Table 2017.02A provides a list of suggested wavelengths, not in any order of preference, that have been found acceptable formost fertilizer materials. Other lines of appropriate sensitivity free of interferences or corrected for interferences may be just as acceptable. However, it is imperative that instrument response (both the wavelength peak scan and the calculated concentration) be reviewed for each test solution and element. Fertilizer materials are extremely variable in composition, and a wide concentration range of potential interfering elements is expected, so no single wavelength will work in every instance. Occasionally, data with an interference will inevitably be found and must be eliminated from inclusion in the mean calculation result for that particular element and sample. ( c )  Wavelength interference treatment.— Interelement interference can cause substantial error in analytical result. Error can be minimized by several techniques: ( 1 ) Three or more analytical lines may be used for a given element,

© 2017 AOAC INTERNATIONAL

03/10/2019