184
P
oitevin
: J
ournalof
AOAC I
nternational
V
ol
. 95, N
o
. 1, 2012
(e.g., for calcium, magnesium, phosphorus, potassium, and
sodium) commercial stock standard solutions. However, it is
also acceptable to use commercially prepared custom-blended
stock standard mixtures containing all of the nine elements at
appropriateconcentrations.Anumberof companiesprovide this
stock standard service.
(
c
)
Intermediate stock solution
.—(Suggested composition
of the intermediate stock standard solution, in mg/kg: Ca =
1500; Cu=10; Fe=50;K=2000;Mg=500;Mn=0.25;Na=
1000; P= 1000; Zn= 20).Add into a 500mLvolumetric flask,
75mLcalcium10 000mg/kg, 5mLcopper 1000mg/kg, 25mL
iron 1000 mg/kg, 100 mL potassium 10 000 mg/kg, 25 mL
magnesium 10 000 mg/kg, 0.125 mLmanganese 1000 mg/kg,
50mLsodium10 000mg/kg, 50mLphosphorus10 000mg/kg,
and 10 mL zinc 1000 mg/kg. Add 10 mLHNO
3
and dilute to
volumewithH
2
O.
(
d
)
Working standard solutions
.—Standards prepared from
intermediate stock standard solution are designed to have the
same acid concentration as digested test solutions (i.e., 10%,
v/v, HNO
3
) for MDC or 15% (v/v) for MDO using combined
acids (HNO
3
, H
2
O
2
, andHCl).
(1) Std6
.—Pipet 15.0 mL intermediate stock standard
solution into a 100 mL acid-washed volumetric flask. Add
10 mL HNO
3
(MDC) or 15 mL combined acids (MDO),
dilute to volume with H
2
O, mix, and transfer to acid-washed
polyethylene bottle.
(2) Std5
.—Pipet 10mL intermediate stock standard solution
into a 100mLacid-washedvolumetricflask.Add10mLHNO
3
(MDC) or 15mLcombinedacids (MDO), dilute tovolumewith
H
2
O,mix, and transfer to acid-washed polyethylene bottle.
(3) Std4
.—Pipet 5.0mL intermediate stock standard solution
into a 100mLacid-washedvolumetricflask.Add10mLHNO
3
(MDC) or 15mLcombinedacids (MDO), dilute tovolumewith
H
2
O,mix, and transfer to acid-washed polyethylene bottle.
(4) Std3
.—Pipet 2.0mL intermediate stock standard solution
into a 100mLacid-washedvolumetricflask.Add10mLHNO
3
(MDC) or 15mLcombinedacids (MDO), dilute tovolumewith
H
2
O,mix, and transfer to acid-washed polyethylene bottle.
(5) Std2
.—Pipet 1.0mL intermediate stock standard solution
into a 100mLacid-washedvolumetricflask.Add10mLHNO
3
(MDC) or 15mLcombinedacids (MDO), dilute tovolumewith
H
2
O,mix, and transfer to acid-washed polyethylene bottle
(6) Std1
.—Pipet 0.5mL intermediate stock standard solution
into a 100mLacid-washedvolumetricflask.Add10mLHNO
3
(MDC) or 15mLcombinedacids (MDO), dilute tovolumewith
H
2
O,mix, and transfer to acid-washed polyethylene bottle.
(7) Blank.—
Add 10mLHNO
3
(MDC) or 15mL combined
acids (MDO) into a 100 mL acid-washed volumetric flask,
dilute to volume with H
2
O, mix, and transfer to acid-washed
polyethylene bottle. All calibration solutions are stable for
1week in glass volumetricflasks.
(
e
)
Sampler wash solution, 10% HNO
3
(v/v)
.—Dilute
100mL tracemetal gradeHNO
3
to 1000mLwithH
2
O.
H. Determination
Make a calibration curve using either weighted linear or
quadratic regression with correlation coefficients of at least
0.9999 from seven standards prepared from intermediate
standard solution, including a blank and six suggested
concentrations of the standard solution (Std1–Std6) displayed
inTable
2011.14J
, and expressed inmg/kg.
Analyze test solutions using an ICP-OES instrument
calibratedwith theworking standard solutions. Insert aworking
standard or other suitable quality control solution every 10
test portions to monitor for instrument drift. The inclusion of
a digestion blank, a sample duplicate, and known reference
materials is highly encouraged.
I. Calculations
The
concentration (C) of each element, in mg/kg, is
calculated as follows:
m
F x V x a C
=
where C = concentration in the test portion sample (mg/kg), a
= concentration (mg/L) of the element in the digest solution as
obtained from instrument,V=volume (mL) of the test solution
after being made up (i.e., 50 mL for MDC and 100 mL for
MDO), F = dilution factor of the test solution, andm =weight
of the test portion (g).
Reference:
J. AOAC Int.
95
, 177 (2012)
Results andDiscussion
All data that are displayed in the
J. AOAC Int.
paper (1) and
used for selectivity, accuracy, and precision performance tests
were treated using robust statistics based in the concept of
Rousseew andCroux (2). The presence of some suspect values
(outliers) can strongly distort classical estimations; however,
resultsmust not be eliminatedwithout a valid justification. For
that reason, robust statistics, that providegood estimations even
without the eliminationof suspect values, havebeenused in the
SLV and ring trial. These robust estimations are insensitive to
extreme values and depend only slightly on data distribution.
It is then neither necessary to test for outliers nor to exclude
suspect values.Themedianhasbeenusedas a robust estimation
of the central value. All data displayed in Tables
2011.14A
–
I
and treated using classical statistics are not significantly
different from those displayed and treatedwith robust statistics
in the
J. AOAC Int.
paper (1).
SLVRobust Statistics
Validation of this method involved an SLV, including a
ruggedness study in which the method was applied in parallel
by at least twodifferent operators in three different laboratories
after open- and closed-vessel digestions on different ICP-
OES equipment with axial, radial, and dual view grating
configurations using Cs 0.1% (w/v) as minimal ionization
buffer concentration.
Linearity
The calibration curves constructed by plotting element
concentration versus peak ratio response (element/IS)
showed
good linearity either in linear or in weighted nonlinear
regression. Weighted nonlinear regression used during SLV
for all elements gave the best regression coefficients with R
2
MTE-01
FORWORKINGGROUP/ERPUSEONLY
DONOTDISTRIBUTE
