Validation and Estimation of Analytical Uncertainty of the 1

ASTM E1898-21 Method for the Determination of Silver by

ed
2

FAAS 3

iew
Rodrigo Martínez-Peñuñuri1,2*, Paulina Vázquez-Bustamante1, Marcela Montoya-Blumenkron1, Dayanna Itzajara 4
Estrada-González1, José Refugio Parga-Torres2, Jorge Chávez-Orduño3 and Alejandro Monserrat García-Alegría 3,* 5

1 Metallurgical Technological Laboratory. Hermosillo, Sonora, México. rmartinez@metalurgicos.com.mx;

pvazquez@metalurgicos.com.mx; mblumenkron@metalurgicos.com.mx;
analisisquimico@metalurgicos.com.mx
2 Deparment of Materials and Metallurgy. Instituto Tecnológico de Saltillo. Tecnológico Nacional de

v
México. rmartinez@metalurgicos.com.mx; drjrparga@hotmail.com
3 University of Sonora. Hermosillo, Sonora, México. jorge.chavez@unison.mx; monserrat.garcia@unison.mx
12

re
* Correspondence: rmartinez@metalurgicos.com.mx (RMP). Tel. +52 6622510503
monserrat.garcia@unison.mx (AMGA). Tel.. +52 6622592163

Abstract: Silver is one of the most important precious metals in mining and metallurgy. Thus, the
methods used to determine its concentration are critical. The objective of this research work was to
er
validate and estimate the expanded uncertainty of the ASTM-E1898-21 method for silver
determination by flame atomic absorption spectroscopy (FAAS). The methodology used for the
validation and estimation of the expanded uncertainty is recommended by international
regulations. The validation results showed precision values under repeatability conditions with
pe
variations in the range of 1.67%–2.04%, whereas the accuracy showed values of 0.0131–0.1445 for
the bias evaluated through Z-score and recovery percentages from 100.7 to 104.8. The validation of
the analytical method was confirmed using a certified reference material (CDN-GS-1Z), and
precision values under repeatability conditions of 1.677 for the percentage of variation and recovery
percentages of 100.65 were obtained. Further, the evaluation of the relative combined uncertainty
showed values from 0.0046 to 0.1166 for the different sources of uncertainties evaluated. The
ot

percentage of relative uncertainty was 2.62% and the expanded uncertainty obtained was 1.2 mg/L
of silver with a reliability factor of 95% (ĸ = 2).

Keywords: Validation; uncertainty; silver; aqua regia.

tn

30
1. Introduction
Silver is a precious metal that is of great interest in the mining and metallurgy industries
[1,2]. In nature, it is found in free form, although it is more commonly found in galena-
type minerals, such as argentite/acanthite (Ag2S), as well as proustite (Ag3AsS3),
rin

pyrargyrite (Ag3SbS3), and chlorargyrite (AgCl) [3–7]. Silver is obtained from copper,
gold, lead, zinc, and nickel, among others. The extraction and/or digestion methods of
silver, as well as methods to determine its concentration or grade, in different minerals
vary [10–13]. One of the conventional methods for mineral digestion is the aqua regia
method (hydro-chloric acid nitric acid mixture with a 3:1 ratio, vol/vol) [14–16] or its
ep

variants [17]. Associated with the aqua regia method, conventional techniques to
determine the concentration of silver involve flame atomic absorption spectroscopy
(FAAS) or using a graphite furnace. In addition, official methods such as ASTM E1898-21
have been established [18]. However, little work has been performed regarding the
validation and/or verification of the analytical method of measurement and the estimation
Pr

of analytical uncertainty. The objective of this work is the validation and estimation of the
analytical uncertainty of the silver measurement method using flame atomic absorption
spectroscopy. For assays laboratories, conducting these analytical procedures is essential
since it grants certainty and credibility to their results and allows them to receive

This preprint research paper has not been peer reviewed. Electronic copy available at: https://ssrn.com/abstract=4347436
 2 of 18

accreditation and certification established by national and international regulations (ISO

ed
17025) [19].

2. Materials and Methods

The silver determination method used in this work based on aqua regia is the method

iew
recommended by ASTM E1898-21.

2.1 Reagents, Standards, and Reference Materials

The CDN-GS-1Z reference material was used, which contained 89.5 ± 4.4 g/t of silver,
Resource Laboratories Ltd. (Langley, B.C., Canada), as well as a reference standard that
contained 1000 ± 3 mg/L of silver (High Purity Standard).

v
2.2 Laboratory Equipment
The measurements to determine the concentration of silver were conducted by FAAS

re
(G8432A, Agilent).

2.3 FAAS Equipment Calibration

The FAAS equipment was verified for calibration using the 1000 mg/dL silver standard,
which was analyzed 30 times using a 0.2 mg/L concentration. The expected response of
er
the equipment should be 0.0267 according to the sensitivity recommended by the sup-
plier (Agilent Technology, 2021) [20].

2.4. Preparation of Glassware

pe
The material was subjected to strict quality control to avoid contamination by trace
elements, following the recommended methodology [21]. The glass work materials (sero-
logical and volumetric pipettes and volumetric flasks, among others) were thoroughly
washed with a commercial liquid detergent and then subjected to a 20% HNO3 solution
(J.T. Baker) for three days. Subsequently, they were rinsed with deionized water. Once
dried, the material was stored in plastic containers with hermetic lids until use.
ot

2.5. Analytical Balance and Volumetric Material Calibration

To verify the calibration of the volumetric material, the gravimetric method was used
[22,23]. An Ohaus analytical balance (Adventurer Pro AV64C), which was previously ver-
tn

ified for calibration using a set of standard masses (B0035TEJAC American Weight Scales
Class 2, NewItem) was used, the mass used was 20 mg, and its mass was determined
through 30 readings. Subsequently, an Acura micropipette of variable volume (10 to 100
and 100 to 1000 µL, Scoorex model 825) was used to measure the weight of 10 µL and 20
µL volumes with 30 repetitions [24]. From the mass values of the analytical balance and
rin

the micropipette, the standard deviation (SD) was obtained, and this value was used to
estimate both uncertainties, as described in the corresponding equations.

2.6. Method Validation

The mathematical methods used for the validation process of the measurement method
ep

were recommended in the Eurachem (2014) international guides. The mathematical

models used for the validation process of the measurement method are the models
recommended by the international Eurachem guidelines (2014) [25].

Linearity Evaluation
Pr

The linearity evaluation was conducted by means of the elaboration of three standard
curves from five concentrations of silver: 0.5, 1.0, 3.0, 6.0, and 9.0 mg/L. With the absorb-
ance average and the concentration of each point, the linearity was graphed, and we ob-
tained the Pearson correlation coefficient (R), the linear regression coefficient (R2), and the
line equation, which is used to analyze residues [26]. These parameters were obtained by

This preprint research paper has not been peer reviewed. Electronic copy available at: https://ssrn.com/abstract=4347436
 3 of 18

plotting concentration data against linearity absorbance and plotting the silver concentra-

ed
tion against the residues obtained in Excel version 2010 [27].

Determination of LOD (Limit of Detection) and LOQ (Limit of Quantification)

The LOD and LOQ were evaluated using the blank of the digestion. To do this, the blank
was read 25 times (n) and from the data obtained, the sd and the corrected standard

iew
deviation (sd´) were calculated. The equations used to calculate the LOD and LOQ are the
following:
sd´= sd/√n (1)

LOD=3sd´ (2)

LOQ=10sd´ (3)

v
Precision evaluation under repeatability conditions
The repeatability (r) was determined by the percentage of the coefficient of variation

re
(%CV). The silver concentrations used were 1.0, 3.0, and 6.0 mg/L. These values represent
the intermediate point of the standard curve, and for this purpose, 25 repetitions were
conducted for each concentration. The percentage of the CV was determined using the
following equation:
𝑠𝑑
% 𝐶𝑉 = *100 (4) 118
er 𝑦

where sd is the standard deviation, and ȳ y is the average of the 25 replicants. The ac-
119

ceptance criteria for instrumental methods are %CV ≤2 [28].

pe
Veracity Evaluation by Bias and Recovery Rate
The veracity by bias was determined by equations 6 and 7, for which the silver con-
centration of 3.0 mg/dL was analyzed 25 times. From the data obtained, the average of the
concentrations (ӯ) was determined. The test statistic used was the Z distribution with a
confidence level of 95% (Z95%). The value of sr is the value of the typical SD. The sr value
is the SD, and n is the number of replicate measurements (n = 25).
ot

𝒁𝟗𝟓% . 𝒔𝒓
[𝒚 ‒ 𝝁] ≤ 𝒏
(5)
tn

𝒔𝒓 = 𝟏.𝟓 + (𝟎.𝟏 ∗ 𝒚) (6) 130

The acceptance criteria are as follows: if ӯ–µ ≤ Z95%.sr/√n, the bias is equal to zero, and
the method is true or exact [28].
The veracity from the recovery percentage (%Rec) was obtained from the silver
concentration of 100 mg/dL, which was analyzed 30 times. This parameter was evaluated
rin

from the following equation:

𝒚
%𝑹𝒆𝒄 = 𝝁
∗ 𝟏𝟎𝟎 (7)
The measured (true) value represents the average (ӯ) of the concentration of 30 replicates,
ep

while µ is the theoretical value of the reference standard concentration (100 mg/dL). The
acceptance criteria are 100% ± 15% (85%–115%) [28].

2.7. Uncertainty Estimation

Pr

The mathematical models used to estimate the analytical uncertainty of the

measurement method were recommended in the Eurachem guidelines (2012, 2014)
[29,30].

Establishment of the Sources of Uncertainty Using Ishikawa Diagram

This preprint research paper has not been peer reviewed. Electronic copy available at: https://ssrn.com/abstract=4347436
 4 of 18

The sources of uncertainty were established by the Ishikawa diagram (Figure 1), which

ed
included the concentration of silver, volumetric material, dilution factors, analytical
balance, repeatability, and reference material [31,32].

v iew
re
er
Figure 1.- Ishikawa diagram with the sources of uncertainty
considered for the estimation of analytical uncertainty.
pe
Standard Uncertainty of Calibration Curves
The standard uncertainty of the aluminum mass concentration through the instrument
response was calculated using the following equation:
𝑺 𝟏 𝟏 (𝜸(𝒙)𝒊 ‒ ȳ𝑴𝑹)𝟐
𝒖𝜸𝒙 = 𝒃𝟏 𝝆 + 𝒏
+ 𝑺𝒙𝒙
(8) 159
ot

Where:
𝒖𝜸𝒙 = uncertainty of the measurand (silver) obtained by the calibrated instrument.
𝐬 = residual standard deviation from the linear regression calculation.
tn

𝐛𝟏 = calculated slope.
𝛒 = number of replicates of the sample under study.
𝐧 = number of solutionns (i) used in the calibration curve multiplied by the number of
replicates (j) of each solution (total data) (i*j).
𝛄𝐱 = silver concentration under study (mg/dL).
rin

ȳ𝐌𝐑 = average of the mass concentrations of the solutions used in the calibration curves.
𝐒𝐱𝐱 = Sum of squares of the residuals of the concentrations obtained.

On the other hand, the relative standard uncertainty for the silver concentration obtained
from the calibration curve was determined by the following equation:
ep

𝒖𝜸𝒙 𝟐
𝒖𝒓𝜸𝒙 = ( ) 𝜸𝒙
(9)

174
Volumetric Material Uncertainty
Pr

Regarding the estimation of the uncertainty of the measurement of the volume of the
aliquot (V1) and capacity (V2) in any volumetric material, there are three main sources of
uncertainty [33]:
1.-) Volumetric material tolerance:

This preprint research paper has not been peer reviewed. Electronic copy available at: https://ssrn.com/abstract=4347436
 5 of 18

𝒕𝒐𝒍
(10)

ed
𝒖𝒕𝒐𝒍 = 𝟑
2.-) The effect of variability or repeatability of the volume measured by the analyst:
𝒔𝑽
𝒖𝑹 = 𝒏 (11)
3.-) The effect of temperature variation in relation to the calibration temperature:
[(𝑻 ‒ 𝑻𝟐𝟎) ∙ 𝜶 ∙ 𝑽]

iew
𝒖∆𝑻 = 𝟑
(12)
Where:
T = Water temperature °C at the time of measurement
T20 = Volumetric material calibration temperature: 20°C
α = Coefficient of expansion of the volume of water °C-1
V = Volume of the pipette or flask

v
Combined Uncertainty of Volumetric Material

re
The three uncertainties are combined, and the value of the combined standard uncertainty
for each volumetric material is calculated using the following equation:

𝒖𝒄𝒐𝒎𝒃 𝒎𝒂𝒕 𝒗𝒐𝒍 = (𝒕𝒐𝒍)𝟐 + (𝑹𝒆𝒑)𝟐 + (𝜟𝑻)𝟐 (13)

er
The relative combined uncertainty for each volumetric material is determined by the
following equation:
𝒖𝒄𝒐𝒎𝒃 𝒎𝒂𝒕 𝒗𝒐𝒍 𝟐
pe
𝒖𝒓𝒄𝒐𝒎𝒃 𝒎𝒂𝒕 𝒗𝒐𝒍 = ( 𝒗𝒐𝒍 ) (14)

The combined standard uncertainty for all volumetric materials used is determined by the
following equation:

(15)
ot

𝒖𝒄𝒐𝒎𝒃𝒗𝒐𝒍𝒎𝒂𝒕 = (𝒎𝒂𝒕𝒗𝒐𝒍𝟏)𝟐 + (𝒎𝒂𝒕𝒗𝒐𝒍𝟐)𝟐 + (𝒎𝒂𝒕𝒗𝒐𝒍𝟑)𝟐 + (𝒎𝒂𝒕𝒗𝒐𝒍𝟒)𝟐 + (𝒎𝒂𝒕𝒗𝒐𝒍𝟓)𝟐 + (𝒎𝒂𝒕𝒗𝒐𝒍𝟔)𝟐 206

207
The combined relative standard uncertainty for all bulk materials used is determined by
the following equation:
tn

𝒎𝒂𝒕𝒗𝒐𝒍𝟏 𝟐 𝒎𝒂𝒕𝒗𝒐𝒍𝟐 𝟐 𝒎𝒂𝒕𝒗𝒐𝒍𝟑 𝟐 𝒎𝒂𝒕𝒗𝒐𝒍𝟒 𝟐 𝒎𝒂𝒕𝒗𝒐𝒍𝟓 𝟐 𝒎𝒂𝒕𝒗𝒐𝒍𝟔 𝟐

𝒖𝒄𝒐𝒎𝒃𝒗𝒐𝒍𝒎𝒂𝒕 = ( 𝒗𝒐𝒍𝟏 ) ( + 𝒗𝒐𝒍𝟐 ) ( + 𝒗𝒐𝒍𝟑 ) + (𝒗𝒐𝒍𝟒
+ ) ( 𝒗𝒐𝒍𝟓 ) + (𝒗𝒐𝒍𝟔 ) (16)
212

Dilution Factor Uncertainty

rin

Calculation of the standard uncertainty due to the dilution factors.

𝒖𝒇𝒅𝒏 = (𝒖𝑽𝟏)𝟐 + (𝒖𝑽𝟐)𝟐 (17)

Calculation of the relative standard uncertainty due to dilution factors

ep

𝒖𝒇𝒅𝒏 𝒖𝑽𝟏 𝟐 𝒖𝑽𝟐 𝟐

𝒇𝒅𝒏
= ( ) +( )
𝑽𝟏 𝑽𝟐
(18)

The dilution factors are given by:

Pr

𝑽𝟐
𝒇𝒅𝒏 = 𝑽
𝟏
(19)
Where:
𝐟𝐝𝐧 = Dilución factor n
𝐕𝟐 = Capacity volume

This preprint research paper has not been peer reviewed. Electronic copy available at: https://ssrn.com/abstract=4347436
 6 of 18

𝐕𝟏 = Aliquot volume

ed
The uncertainty of the dilution factor is calculated from the following equation:

𝟐
𝒖𝒇𝒅𝒏 = (𝒖𝒇𝒅𝟏)𝟐 + (𝒖𝒇𝒅𝟐)𝟐 + (𝒖𝒇𝒅𝟑) + (𝒖𝒇𝒅𝟒)𝟐 + (𝒖𝒇𝒅𝟓)𝟐 (20)

iew
The relative uncertainty of the dilution factor is determined using the following equation:

𝒖𝒇𝒅𝒏 𝒖𝒇𝒅𝟏 𝟐 𝒖𝒇𝒅𝟐 𝟐 𝒖𝒇𝒅𝟑 𝟐 𝒖𝒇𝒅𝟒 𝟐 𝒖𝒇𝒅𝟓 𝟐

𝒗 = ( ) ( ) ( ) ( )
𝒗𝟏
+ 𝒗𝟐
+ 𝒗𝟑
+ 𝒗𝟒
+ ( )
𝒗𝟓
(21)

Uncertainty of Analytical Balance Calibration

v
𝒔𝒃𝒂𝒍𝒂𝒏𝒛𝒂
𝒖𝒃𝒂𝒍𝒂𝒏𝒛𝒂 = (22)

re
𝟔
The relative uncertainty of the analytical balance is determined by the following equation:

𝒖𝒃𝒂𝒍𝒂𝒏𝒛𝒂
𝒖𝒓𝒃𝒂𝒍𝒂𝒏𝒛𝒂 = 𝒎 (23)
m is the standard mass used to verify the calibration of the analytical balance.
er
The combined standard uncertainty for the six masses tested is determined by the
following equation:
pe
𝒖𝒄𝒐𝒎𝒃 𝒃𝒂𝒍𝒂𝒏𝒛𝒂 = (𝒖𝒎𝟏)𝟐 + (𝒖𝒎𝟐)𝟐 + (𝒖𝒎𝟑)𝟐 + (𝒖𝒎𝟒)𝟐 + (𝒖𝒎𝟓)𝟐 + (𝒖𝒎𝟔)𝟐 (24) 237

238
The combined relative uncertainty for the six masses used is determined by the following
equation:
ot

𝒖𝒎𝟏 𝟐 𝒖𝒎𝟐 𝟐 𝒖𝒎𝟑 𝟐 𝒖𝒎𝟒 𝟐 𝒖𝒎𝟓 𝟐 𝒖𝒎𝟔 𝟐

𝒖𝒓𝒄𝒐𝒎𝒃 𝒃𝒂𝒍𝒂𝒏𝒛𝒂 = ( ) ( ) ( ) ( ) ( ) ( )
𝒎𝟏
+ 𝒎𝟐
+ 𝒎𝟑
+ 𝒎𝟒
+ 𝒎𝟓
+ 𝒎𝟔 (25)
tn

Uncertainty of Precision Under Conditions of Repeatability and Intermediate Precision

According to the uncertainty guidelines [27,34], the uncertainty related to the repeatability
of the measurements must be included, and it is calculated according to the following
equation:
𝑺𝜸𝒓𝒆𝒑 𝟐
( )
rin

𝒖𝜸𝒓𝒆𝒑 = 𝒏
(26)

Importantly, the precision under repeatability and intermediate precision conditions were
determined for silver concentrations of 1.0, 3.0, and 6.0 mg/L (rep1, rep2, and rep3), S
represents the deviation standard, and n is 25. On the other hand, the relative standard
ep

uncertainty for repeatability is calculated using the following equation:

𝒖𝜸𝒓𝒆𝒑𝟏 𝟐
𝒖𝒓𝜸𝒓𝒆𝒑= ( )
𝜸𝒓𝒆𝒑𝟏
(27)
Pr

The combined standard uncertainty for the three concentrations with which the
repeatability was evaluated is determined as follows:

𝒖𝜸𝒄𝒐𝒎𝒃 𝒓𝒆𝒑 = (𝒖𝒓𝒆𝒑𝟏)𝟐 + (𝒖𝒓𝒆𝒑𝟐)𝟐 + (𝒖𝒓𝒆𝒑𝟑)𝟐 (28)

This preprint research paper has not been peer reviewed. Electronic copy available at: https://ssrn.com/abstract=4347436
 7 of 18

The combined relative standard uncertainty for the three concentrations with which the

ed
repeatability was evaluated is calculated by the following equation:

𝒖𝒓𝒆𝒑𝟏 𝟐 𝒖𝒓𝒆𝒑𝟐 𝟐 𝒖𝒓𝒆𝒑𝟑 𝟐

𝒖𝒓𝜸𝒄𝒐𝒎𝒃 𝒓𝒆𝒑 = ( ) +( ) +( )
𝟏.𝟎 𝟑.𝟎 𝟔.𝟎
(29) 258

iew
259
Uncertainty of the Certified Reference Material

The reference material insert indicates that the silver concentration is 1000 ± 3 µg/mL. The
value of ±3 represents the expanded uncertainty, which is why the expanded uncertainty
(U) must be divided by the coverage factor (k = 2) to obtain the standard uncertainty, as

v
described in the following equation:

re
𝒖𝑨𝒈 = 𝑼𝒆𝒙𝒑/𝒌 (30)

Once the standard uncertainty was obtained, it was divided by √ 3 since a rectangular
distribution was assumed to obtain the estimate of the real standard uncertainty of the
reference material [35], as indicated in the following equation:
er 𝒖𝜸𝑨𝒈 =
𝑼𝒆𝒙𝒑/𝒌
3
(31)

whereas the relative standard uncertainty for the reference material was obtained using
pe
the following equation:

𝒖𝜸𝑨𝒈 𝟐
𝒖𝒓𝜸𝑨𝒈= ( )
𝒚𝑨𝒈
(32)

Combined Standard Uncertainty (CSU)

The uncertainty sources considered in this study were combined in a quadratic form using
ot

the following equation:

𝒖𝒄𝜸𝑨𝒈 = (𝒖𝜸𝑪𝑪)𝟐 + (𝒖𝒗𝒐𝒍𝒎𝒂𝒕)𝟐 + (𝒖𝒇𝒅𝒏)𝟐 + (𝒖𝒃𝒂𝒍𝒂𝒏𝒄𝒆)𝟐 + (𝒖𝜸𝒓𝒆𝒑𝒆𝒂)𝟐 + (𝒖𝜸𝑨𝒈)𝟐 (33)

tn

Once the CSU was obtained, it was divided by the silver concentration to obtain the
combined relative uncertainty (CRU) using the following equation:
𝒖𝒓𝒄𝜸𝑨𝒈 =
𝒖𝜸𝒙 𝟐 𝒖𝒄𝒐𝒎𝒃 𝒎𝒂𝒕 𝒗𝒐𝒍 𝟐 𝒖𝒇𝒅 𝟐 𝒖𝒎 𝟐 𝒖𝜸𝒓𝒆𝒑 𝟐 𝒖𝜸𝑨𝒈 𝟐
( ) ( + ) +( ) +( ) +( ) +( ) (34)
rin

𝜸𝒙 𝒗𝒐𝒍 𝒇𝒅 𝒎 𝜸𝒓𝒆𝒑 𝒚𝑨𝒈

On the other hand, the percentage of the combined relative uncertainty was determined
by the following equation:
ep

% 𝒖𝒓𝒄𝜸𝑨𝒈 = 𝒖𝒓𝒄𝜸𝑨𝒈x 100 (35)

Expanded Uncertainty (EU)
The combined standard uncertainty obtained was multiplied by the coverage factor (κ =
2) to obtain the expanded uncertainty, whose value was within a 95% confidence interval
[36], as follows:
Pr

𝑼𝒆𝒙𝒑𝑨𝒈 = 𝒖𝒄𝜸𝑨𝒈* 2 (36)

3. Results and Discussion

3.1 Method Validation

This preprint research paper has not been peer reviewed. Electronic copy available at: https://ssrn.com/abstract=4347436
 8 of 18

ed
Linearity Evaluation
A visual inspection to Figure 2 indicates that the calibration curve showed a linear
relationship between concentration and absorbance readings. That is, it is observed that
when the concentration increases, the absorbance increased proportionally [26]. Thus, the

iew
results obtained for the Pearson correlation coefficient (R) and the regression coefficient
(R2) are acceptable since these values must be ≥ 0.9950 to evaluate linearity [26]. Mortada
et al., (2013) [36] obtained Pearson correlation coefficient values of 0.9990, which are similar
to those obtained by our group. On the other hand, recent investigations obtained slightly
lower values (R2 = 0.9900 to 0.9980) than those obtained by our group, but they are within
the acceptable range [12,37,38]. The analysis of the residuals (Figure 3) showed that these
values had a random distribution and did not show a particular trend; thus, they are

v
acceptable for the evaluation of linearity [26,37].

re
306
Absorbance (328.1 nm)

1
0.9
0.8
er 0.7
0.6
0.5
pe
0.4
0.3
0.2
0.1
0
0 1 2 3 4 5 6 7 8 9 10
ot

Concentration of silver (mg/L)

Figure 2.- Calibration curve with the average of three absorbance

tn

readings for each concentration.

0.2
rin
Residues

0.1

0
0 1 2 3 4 5 6 7 8 9 10
-0.1
ep

-0.2

-0.3
Pr

Concentration of silver (mg/L)

Figure 3.- Residue analysis for the three concentrations

experimentally obtained from the three calibration curves.

This preprint research paper has not been peer reviewed. Electronic copy available at: https://ssrn.com/abstract=4347436
 9 of 18

ed
LOD (Limit of Detection) and LOQ (Limit of Quantification)

𝒅𝒔 𝟎.𝟎𝟑𝟎𝟒
𝒅𝒔´ = 𝒏
= 𝟐𝟓 = 0.00608

iew
LOD = 3ds´ = 3*0.00608 = 0.0182 mg/L

LOQ = 10ds´ = 10*0.00608 = 0.0608 mg/L

321
The LOD and LOQ values for silver are exceptionally low and acceptable, considering that
they were determined by FAAS. Díaz et al. (2017) [12] obtained the LOD and LOQ for

v
silver, between 1.1 and 3.4 and between 3.6 and 11.1 µg/g, respectively, using laser-
induced breakdown spectroscopy (LIBS) and atomic emission spectroscopy. Saha et al.

re
(2015) [8] obtained an LOQ of 5.0 µg/kg, determined by GF-AAS (Graphite Furnace-
Atomic Absorption Spectroscopy).

Accuracy evaluation under repeatability conditions

The evaluation of the precision under repeatability conditions was conducted with the
er
three intermediate concentrations with which the calibration curve was built, 1.0, 3.0, and
6.0 mg/L of silver, and % CV was calculated for the three concentrations (% CV1, % CV2
and % CV3).
334
pe
𝑑𝑠 0.0175
% 𝐶𝑉1 = 𝑦
*100 = 1.0131
∗ 100 = 1.73
336

𝑑𝑠 0.0641
% 𝐶𝑉2 = 𝑦
*100 = 3.1445
∗ 100 = 2.04 337

𝑑𝑠 0.1010
ot

% 𝐶𝑉3 = 𝑦
*100 = 6.0440
∗ 100 = 1.67 338

The results obtained are acceptable since the acceptance criterion for precision under
repeatability conditions is %CV ≤2% for instrumental methods [28]. Mortada et al. (2013)
tn

[36] obtained a %CV of 3.16% using the aqua regia method, determined by Perkin Elmer
GF-AAS. Resano et al. (2006) [39] obtained precision values from 6.1 to 13.5% RSD using
different reference materials when determining silver concentration by AAS-GF. These
values are well above the values obtained by our group.
rin

Veracity Evaluation by bias and percentage recovery

The evaluation of the veracity by calculating the bias and the recovery percentage was
performed using the three intermediate concentrations by which the calibration curve was
ep

built, 1.0, 3.0, and 6.0 mg/L of silver. The bias (𝑠1, 𝑠2, 𝑠3) and the recovery percentage (%
Rec1, % Rec2, and % Rec3) were calculated for the three concentrations.
Bias (s)
𝒔𝒓 = 𝟏.𝟓 + (𝟎.𝟏 ∗ 𝟏.𝟎𝟏𝟑𝟏) = 1.6013
Pr

𝟏.𝟗𝟔 . 𝟏.𝟔𝟎𝟏𝟑
𝒔𝟏 [𝟏.𝟎𝟏𝟑𝟏 ‒ 𝟏.𝟎𝟎𝟎𝟎] ≤ 𝟐𝟓 = 0.0131 ≤ 0.6277

𝒔𝒓 = 𝟏.𝟓 + (𝟎.𝟏 ∗ 𝟑.𝟏𝟒𝟒𝟓) = 1.8144

This preprint research paper has not been peer reviewed. Electronic copy available at: https://ssrn.com/abstract=4347436
 10 of 18

𝟏.𝟗𝟔 . 𝟏.𝟖𝟏𝟒𝟒
= 0.1445 ≤ 0.7112

ed
𝒔𝟐 [𝟑.𝟏𝟒𝟒𝟓 ‒ 𝟑.𝟎𝟎𝟎𝟎] ≤ 𝟐𝟓

𝒔𝒓 = 𝟏.𝟓 + (𝟎.𝟏 ∗ 𝟔.𝟎𝟒𝟒𝟎) = 2.1044 356

𝟏.𝟗𝟔 . 𝟐.𝟏𝟎𝟒𝟒
𝒔𝟐 [𝟔.𝟎𝟒𝟒𝟎 ‒ 𝟔.𝟎𝟎𝟎𝟎] ≤ 𝟐𝟓 = 0.0440 ≤ 0.8249

iew
The bias value was found to be less than the value of the Z distribution test statistic (Z-
score) for the three concentrations evaluated; therefore, there is no bias, indicating that the
method is true [28].
Recovery percentage (%Rec) 361

𝟏.𝟎𝟏𝟑𝟏
%𝑹𝒆𝒄𝟏 = 𝟏.𝟎𝟎𝟎𝟎 * 100 = 101.3 362

v
𝟑.𝟏𝟒𝟒𝟓
%𝑹𝒆𝒄𝟐 = * 100 = 104.8 363

re
𝟑.𝟎𝟎𝟎𝟎

𝟔.𝟎𝟒𝟒𝟎
%𝑹𝒆𝒄𝟑 = 𝟔.𝟎𝟎𝟎𝟎 * 100 = 100.7 364

The obtained recovery percentage results (100% ± 5%) are acceptable, indicating that there
is truth in the evaluation method [28]. De la Calle et al. (2011) [12] obtained silver recovery
percentages between 81% and 107% when they used acid digestion, while the recovery
er
percentages improved from 97% to 106% when they used thiourea as the extracting agent
using SRM 2710 and 2782. Mortada et al. (2013) [36] obtained silver recovery percentages
ranging from 98.4% to 104.5% for different types of biological samples and/or wastewater,
evaluated with GF-AAS, whereas Naeemullah et al. (2014) [40] found recovery percentage
pe
values on the order of 98%, obtained from different water samples.

Precision and veracity evaluation from CDN-GS-1Z certified reference material

To evaluate and compare the results obtained in the measurement method validation
process, a certified reference material (CDN-GS-1Z) was used. The certified reference
ot

material had a silver concentration of 89.5 ± 4.4 g/T. With it, the precision was evaluated
under repeatability conditions (%CV) and the veracity was found from the recovery
percentage (%Rec) using 20 silver determinations. The values obtained experimentally
were substituted in equations 4 and 7, respectively.
tn

1.677
% 𝐶𝑉= 90.088
∗ 100 = 1.677

90.08
% Rec = 89.50 * 100 = 100.65
rin

The results obtained are acceptable since they are within the acceptance percentages. The
acceptance values are %CV ≤2% and %Rec = 100% ± 5% [28]. With these results, the
validation process of the ASTM E1898-21 measurement method was concluded.
ep

3.2 Uncertainty estimation

Concentration uncertainty from calibration curves
By solving for values in equation 8, the standard uncertainty was obtained for the
concentration obtained from the calibration curve as a response of the measurement
Pr

instrument.

0.00998308 1 1 (3.9011655 ‒ 3.90)2

𝑢𝑦𝑥 = * + 15 + = 0.04506 mg/L
0.1144 5 51.2
393

10

This preprint research paper has not been peer reviewed. Electronic copy available at: https://ssrn.com/abstract=4347436
 11 of 18

Meanwhile, the relative standard uncertainty was obtained using equation 9.

ed
2
𝑢𝑟𝑦𝑥 = (0.04506
3.90116)
= 0.0115

Volumetric material uncertainty

iew
Clearing values in equation 10:
1.-) Volumetric material tolerance
𝟎.𝟎𝟎𝟓
𝒖𝒕𝒐𝒍 = 𝟑
= 0.0028 for micropipette of 1 mL
𝟎.𝟎𝟐
𝒖𝒕𝒐𝒍 = 𝟑
= 0.0115 for micropipette of 2 mL
𝟎.𝟎𝟐
= 0.0115 for micropipette of 5 mL

v
𝒖𝒕𝒐𝒍 = 𝟑
𝟎.𝟎𝟑
𝒖𝒕𝒐𝒍 = = 0.0173 for pipette of 10 mL

re
𝟑
𝟎.𝟎𝟒
𝒖𝒕𝒐𝒍 = 𝟑
= 0.0231 for pipette of 20 mL
𝟎.𝟏
𝒖𝒕𝒐𝒍 = 𝟑
= 0.0577 for volumetric flask of 100 mL

2.-) Effect of variability or repeatability of the volume measured by the analyst.

er
Clearing values in equation 11:
𝟎.𝟎𝟎𝟓𝟓
𝒖𝑹 = 𝟑𝟎
= 0.0010 for micropipette of 1 mL
𝟎.𝟎𝟏𝟐𝟏
𝒖𝑹 = = 0.0022 for micropipette of 2 mL
pe
𝟑𝟎
𝟎.𝟎𝟏𝟔𝟔
𝒖𝑹 = 𝟑𝟎
= 0.0031 for micropipette of 5 mL
𝟎.𝟎𝟏𝟖𝟓
𝒖𝑹 = 𝟑𝟎
= 0.0034 for pipette of 10 mL 412
𝟎.𝟎𝟐𝟏𝟕
𝒖𝑹 = 𝟑𝟎
= 0.0039 for pipette of 20 mL 413
𝟎.𝟏𝟑𝟏𝟏
= 0.0024 for volumetric flask of 100 mL
ot

𝒖𝑹 = 𝟑𝟎

3.-) Effect of temperature variation in relation to the calibration temperature.

Clearing values in equation 11:
tn

[(𝟐𝟖 ‒ 𝟐𝟎) ∙ 𝟎.𝟎𝟎𝟎𝟐𝟏 ∙ 𝟏]

𝒖∆𝑻 = 𝟑
= 0.00009 for micropipette of 1 mL
[(𝟐𝟖 ‒ 𝟐𝟎) ∙ 𝟎.𝟎𝟎𝟎𝟐𝟏 ∙ 𝟐]
𝒖∆𝑻 = 𝟑
= 0.00194 for micropipette of 2 mL
[(𝟐𝟖 ‒ 𝟐𝟎) ∙ 𝟎.𝟎𝟎𝟎𝟐𝟏 ∙ 𝟓]
𝒖∆𝑻 = 𝟑
= 0.00489 for micropipette of 5 mL
rin

[(𝟐𝟖 ‒ 𝟐𝟎) ∙ 𝟎.𝟎𝟎𝟎𝟏 ∙ 𝟏𝟎]

𝒖∆𝑻 = 𝟑
= 0.00969 for pipette of 10 mL 421
[(𝟐𝟖 ‒ 𝟐𝟎) ∙ 𝟎.𝟎𝟎𝟎𝟏 ∙ 𝟐𝟎]
𝒖∆𝑻 = 𝟑
= 0.01939 for pipette of 20 mL 422
[(𝟐𝟖 ‒ 𝟐𝟎) ∙ 𝟎.𝟎𝟎𝟎𝟏 ∙ 𝟏𝟎𝟎]
𝒖∆𝑻 = 𝟑
= 0.09699 for volumetric flask of 100 mL
ep

Combined standard uncertainty of the volumetric material

The three uncertainties are combined and their value is calculated using equation 13:
𝒖𝒄𝒐𝒎𝒃 𝒎𝒂𝒕 𝒗𝒐𝒍𝟏 = (𝟎.𝟎𝟎𝟐𝟖)𝟐 + (𝟎.𝟎𝟎𝟏𝟎)𝟐 + (𝟎.𝟎𝟎𝟎𝟎𝟗)𝟐 = 0.0029 for micropipette of 1 mL 427
𝒖𝒄𝒐𝒎𝒃 𝒎𝒂𝒕 𝒗𝒐𝒍𝟐 = (𝟎.𝟎𝟏𝟏𝟓)𝟐 + (𝟎.𝟎𝟎𝟐𝟐)𝟐 + (𝟎.𝟎𝟎𝟏𝟗𝟒)𝟐 = 0.0118 for micropipette of 2 mL 428
Pr

𝒖𝒄𝒐𝒎𝒃 𝒎𝒂𝒕 𝒗𝒐𝒍𝟑 = (𝟎.𝟎𝟏𝟏𝟓)𝟐 + (𝟎.𝟎𝟎𝟑𝟏)𝟐 + (𝟎.𝟎𝟎𝟒𝟖𝟗)𝟐 = 0.0128 for micropipette of 5 mL 429
𝒖𝒄𝒐𝒎𝒃 𝒎𝒂𝒕 𝒗𝒐𝒍𝟒 = (𝟎.𝟎𝟏𝟕𝟑)𝟐 + (𝟎.𝟎𝟎𝟑𝟒)𝟐 + (𝟎.𝟎𝟎𝟗𝟔𝟗)𝟐 = 0.0220 for pipette of 10 mL 430
𝒖𝒄𝒐𝒎𝒃 𝒎𝒂𝒕 𝒗𝒐𝒍𝟓 = (𝟎.𝟎𝟐𝟑𝟏)𝟐 + (𝟎.𝟎𝟎𝟑𝟗)𝟐 + (𝟎.𝟎𝟏𝟗𝟑𝟗)𝟐 = 0.0304 for pipette of 20 mL 431
𝒖𝒄𝒐𝒎𝒃 𝒎𝒂𝒕 𝒗𝒐𝒍𝟔 = (𝟎.𝟎𝟓𝟕𝟕)𝟐 + (𝟎.𝟎𝟎𝟐𝟒)𝟐 + (𝟎.𝟎𝟗𝟔𝟗𝟗)𝟐 = 0.1128 for volumetric flask of 100 mL 432
433

11

This preprint research paper has not been peer reviewed. Electronic copy available at: https://ssrn.com/abstract=4347436
 12 of 18

To obtain the estimate of the combined standard uncertainty of all the volumetric materials

ed
used, equation 15 was used:

𝒖𝒄𝒐𝒎𝒃 𝒗𝒐𝒍 𝒎𝒂𝒕 = (𝟎.𝟎𝟎𝟐𝟗)𝟐 + (𝟎.𝟎𝟏𝟏𝟖)𝟐 + (𝟎.𝟎𝟏𝟐𝟖)𝟐 + (𝟎.𝟎𝟐𝟐𝟎)𝟐 + (𝟎.𝟎𝟑𝟎𝟒)𝟐 + (𝟎.𝟏𝟏𝟐𝟖)𝟐 =

0.1201

iew
To obtain the combined relative standard uncertainty of all volumetric materials used,
equation 16 was used:

𝟎.𝟎𝟎𝟐𝟗 𝟐 𝟎.𝟎𝟏𝟏𝟖 𝟐 𝟎.𝟎𝟏𝟐𝟖 𝟐 𝟎.𝟎𝟐𝟐𝟎 𝟐 𝟎.𝟎𝟑𝟎𝟒 𝟐 𝟎.𝟏𝟏𝟐𝟖 𝟐

𝒖𝒓𝒄𝒐𝒎𝒃 𝒗𝒐𝒍 𝒎𝒂𝒕 = ( 𝟏 ) ( + 𝟐 ) (+ 𝟓 ) + ( 𝟏𝟎 ) ( + 𝟐𝟎 ) + ( 𝟏𝟎𝟎) = 0.0075
444

v
Dilution factor uncertainty
To estimate the standard uncertainty due to dilution factors, it was solved in equation 17:

re
𝒖𝒇𝒅𝟏 = (𝟎.𝟎𝟎𝟐𝟗)𝟐 + (𝟎.𝟏𝟏𝟐𝟖)𝟐 = 0.1128
𝒖𝒇𝒅𝟐 = (𝟎.𝟎𝟏𝟏𝟖)𝟐 + (𝟎.𝟏𝟏𝟐𝟖)𝟐 = 0.1134
𝒖𝒇𝒅𝟑 = (𝟎.𝟎𝟏𝟐𝟖)𝟐 + (𝟎.𝟏𝟏𝟐𝟖)𝟐 = 0.1135
𝒖𝒇𝒅𝟒 = (𝟎.𝟎𝟐𝟐𝟎)𝟐 + (𝟎.𝟏𝟏𝟐𝟖)𝟐 = 0.1149
𝒖𝒇𝒅𝟓 = (𝟎.𝟎𝟑𝟎𝟒)𝟐 + (𝟎.𝟏𝟏𝟐𝟖)𝟐 = 0.1168
er
To estimate the combined standard uncertainty of the dilution factors, Equation 20 was
used:

𝟐
𝒖𝒇𝒅𝒏 = (𝟎.𝟏𝟏𝟐𝟖)𝟐 + (𝟎.𝟏𝟏𝟑𝟒)𝟐 + (𝟎.𝟏𝟏𝟑𝟓) + (𝟎.𝟏𝟏𝟒𝟗)𝟐 + (𝟎.𝟏𝟏𝟔𝟖)𝟐 = 0.1653
pe
456
To estimate the combined relative uncertainty of the dilution factors, values were
calculated using equation 21:
𝒖𝒇𝒅𝒏 𝟐 𝟐 𝟐 𝟐 𝟐
𝒗 = (𝟎.𝟏𝟏𝟐𝟖
𝟏 )
+ ( 𝟐 ) + ( 𝟓 ) + ( 𝟏𝟎 )
𝟎.𝟏𝟏𝟑𝟒 𝟎.𝟏𝟏𝟑𝟓 𝟎.𝟏𝟏𝟒𝟗
+ (𝟎.𝟏𝟏𝟔𝟖
𝟐𝟎 )
= 0.1166
ot

Uncertainty of analytical balance calibration

To obtain the standard uncertainty of the analytical balance, values of each mass used were
solved using equation 22:
tn

𝟎.𝟎𝟏𝟏𝟑
𝒖𝒃𝒂𝒍𝒂𝒏𝒄𝒆 = 𝟔 = 0.0046 mg for mass of 1 g

𝟎.𝟎𝟎𝟗𝟔
𝒖𝒃𝒂𝒍𝒂𝒏𝒄𝒆 = 𝟔 = 0.0039 mg for mass of 2 g
rin

𝟎.𝟎𝟎𝟗𝟎
𝒖𝒃𝒂𝒍𝒂𝒏𝒄𝒆 = 𝟔 = 0.0036 mg for mass of 5 g

𝟎.𝟎𝟎𝟕𝟗
𝒖𝒃𝒂𝒍𝒂𝒏𝒄𝒆 = 𝟔 = 0.0032 mg for mass of 10 g
ep

𝟎.𝟎𝟑𝟓𝟗
𝒖𝒃𝒂𝒍𝒂𝒏𝒄𝒆 = 𝟔 = 0.0146 mg for mass of 20 g

𝟎.𝟎𝟓𝟐𝟒
𝒖𝒃𝒂𝒍𝒂𝒏𝒄𝒆 = 𝟔 = 0.0213 mg for mass of 100
Pr

The combined uncertainty for the calibration of the analytical balance with standard
masses was found using equation 24:

𝒖𝒄𝒐𝒎𝒃 𝒃𝒂𝒍𝒂𝒏𝒛𝒂 = (𝟎.𝟎𝟎𝟒𝟔)𝟐 + (𝟎.𝟎𝟎𝟑𝟗)𝟐 + (𝟎.𝟎𝟎𝟑𝟔)𝟐 + (𝟎.𝟎𝟎𝟑𝟐)𝟐 + (𝟎.𝟎𝟏𝟒𝟔)𝟐 + (𝟎.𝟎𝟐𝟏𝟑)𝟐 = 0.0052

12

This preprint research paper has not been peer reviewed. Electronic copy available at: https://ssrn.com/abstract=4347436
 13 of 18

For the combined relative standard uncertainty of the analytical balance, values in

ed
Equation 25 were calculated:

𝟎.𝟎𝟎𝟒𝟔 𝟐 𝟎.𝟎𝟎𝟑𝟗 𝟐 𝟎.𝟎𝟎𝟑𝟔 𝟐 𝟎.𝟎𝟎𝟑𝟐 𝟐 𝟎.𝟎𝟏𝟒𝟔 𝟐 𝟎.𝟎𝟐𝟏𝟑 𝟐

𝒖𝒓𝒄𝒐𝒎𝒃 𝒃𝒂𝒍𝒂𝒏𝒛𝒂 = ( 𝟏 ) ( + 𝟐
+) ( 𝟓 ) + ( 𝟏𝟎 ) (
+ 𝟐𝟎 ) + ( 𝟏𝟎𝟎) = 0.0046

iew
Uncertainty of precision under repeatability conditions
Precision under repeatability and intermediate precision conditions were determined for
silver concentrations of 1.0, 3.0 and 6.0 mg/dL (rep1, rep2, and rep3), s represents the
deviation standard, and n is 25. Therefore, values are cleared in equation 26:

𝟎.𝟏𝟔𝟎𝟗𝟓𝟖𝟑𝟑𝟕 𝟐
( ) = 0.0322

v
𝒖𝜸𝒓𝒆𝒑𝟏 = 𝟐𝟓

re
𝟎.𝟐𝟎𝟐𝟒𝟒𝟎𝟐𝟖𝟓 𝟐
𝒖𝜸𝒓𝒆𝒑𝟐 = ( 𝟐𝟓 ) = 0.0405

𝟎.𝟏𝟑𝟔𝟗𝟏𝟓𝟓𝟐𝟗 𝟐
er 𝒖𝜸𝒓𝒆𝒑𝟑 = ( 𝟐𝟓 ) = 0.0274

The combined standard uncertainties for the three concentrations by which the
repeatability was evaluated were determined by solving for values in equation 28:
pe
𝒖𝜸𝒄𝒐𝒎𝒃 𝒓𝒆𝒑 = (𝟎.𝟎𝟑𝟐𝟐)𝟐 + (𝟎.𝟎𝟒𝟎𝟓)𝟐 + (𝟎.𝟎𝟐𝟕𝟒)𝟐 = 0.0346

The relative standard uncertainties for the three concentrations were evaluated by solving
for values in equation 29:

𝟎.𝟎𝟑𝟐𝟐 𝟐 𝟎.𝟎𝟒𝟎𝟓 𝟐 𝟎.𝟎𝟐𝟕𝟒 𝟐

𝒖𝒓𝜸𝒄𝒐𝒎𝒃 𝒓𝒆𝒑 = ( ) (
+ + ) ( ) = 0.0324
ot

𝟏.𝟎 𝟑.𝟎 𝟔.𝟎

Uncertainty of the certified reference material

The silver concentration of the reference material is 1000 ± 3 µg/mL. To obtain the standard
tn

uncertainty, values were solved in equation 30:

𝒖𝑨𝒈 = 𝟑/𝟐 = 1.5

Once the standard uncertainty was obtained, it was divided by √3 since a rectangular
distribution was assumed. To obtain the estimate of the real standard uncertainty of the
rin

certified reference material [20], values are cleared in equation 31:

𝟏.𝟓
𝒖𝜸𝑨𝒈 = = 0.8660
3
Therefore, the value for the relative standard uncertainty of the certified reference material
ep

was obtained by isolating values in equation 32:

𝟎.𝟖𝟔𝟔𝟎 𝟐
𝒖𝒓𝜸𝑨𝒈= ( 𝟏𝟎𝟎𝟎.𝟎 ) = 0.00086
Pr

Combined Standard Uncertainty (CSU)

The sources of uncertainty considered for this study were combined in quadratic form,
clearing values in equation 33:

𝒖𝒄𝜸𝑨𝒈 = (𝟎.𝟎𝟒𝟓𝟏)𝟐 + (𝟎.𝟏𝟐𝟎𝟏)𝟐 + (𝟎.𝟏𝟔𝟓𝟑)𝟐 + (𝟎.𝟎𝟎𝟒𝟔)𝟐 + (𝟎.𝟎𝟑𝟒𝟔)𝟐 + (𝟎.𝟖𝟔𝟔𝟎)𝟐 = 0.8380

13

This preprint research paper has not been peer reviewed. Electronic copy available at: https://ssrn.com/abstract=4347436
 14 of 18

Once the CSU was obtained, it was divided by the silver concentration to obtain the

ed
combined relative uncertainty (CRU), considering the working concentration for silver, for
which values were cleared in equation 34:

𝒖𝒓𝒄𝜸𝑨𝒈 = (𝟎.𝟎𝟏𝟏𝟓)𝟐 + (𝟎.𝟎𝟎𝟕𝟓)𝟐 + (𝟎.𝟏𝟏𝟔𝟔)𝟐 + (𝟎.𝟎𝟎𝟒𝟔)𝟐 + (𝟎.𝟎𝟑𝟐𝟒)𝟐 + (𝟎.𝟎𝟎𝟎𝟖𝟔)𝟐 = 0.0262

iew
On the other hand, the percentage of the combined relative uncertainty was determined
by solving for values in equation 35:

% 𝒖𝒓𝒄𝜸𝑨𝒈 = 0.0262 x 100 = 2.62%

Expanded uncertainty (EU)
The combined standard uncertainty obtained was multiplied by the coverage factor κ = 2

v
to obtain the expanded uncertainty, whose value was within a 95% confidence interval
[35]. For this, values were cleared in equation 36:

re
𝑼𝒆𝒙𝒑𝑨𝒈 = 0.8380 * 2 = 1.176 ≈ 1.2
Fraser (2012) [41] obtained a standard uncertainty value of 0.089 g/T for gold, obtained
from a certified reference material (AMIS0013) digested with aqua regia and determined
by atomic absorption spectroscopy. This value is higher than the one obtained by our
group for silver since the expanded uncertainty value we found was 1.2, considering six
er
sources of uncertainty, with a significance level of 5% and a confidence interval of 95% (k
= 2). Leśniewska et al. (2016) [42] obtained combined standard uncertainty values ranging
from 3.22 to 7.94 and expanded uncertainties in the range of 7.9–11.7 mg/kg for zinc,
pe
determined by FAAS from three fractions of soil extracts. Singh et al. (2014) [43] obtained
an expanded uncertainty of 0.085 using a new gravimetric fire test method for the
determination of silver. When comparing their results against the method recommended
by ASTM, they obtained an expanded uncertainty of 0.2 with a confidence level of 95%
and a coverage factor equal to 2 (k = 2). Korshed et al. (2012) [44] obtained expanded
uncertainty values of 3.1 and 5.4% for chromium and cobalt, respectively, while they
obtained values from 8.6 to 38.6 for other metals. The estimation of uncertainty allows to
ot

objectively determine the reliability of the results obtained based on the measurement
methods used. In addition, it offers a magnificent tool for developing projects and/or
continuous improvement plans, impacting the internal certification and/or accreditation
tn

processes of testing or service laboratories [45].

Contribution of sources of uncertainty

Figure 4 shows that the dilution factor contributes with greater relative uncertainty, in
relation to the other sources of uncertainties evaluated. Our team obtained similar results
rin

when evaluating the methods for determining the concentrations of other chemical
elements using FAAS or GF-AAS [46,47].
ep
Pr

14

This preprint research paper has not been peer reviewed. Electronic copy available at: https://ssrn.com/abstract=4347436
 15 of 18

ed
Relative uncertainty
0.15

iew
0.1

0.05

0
Uncertainty Uncertainty Uncertainty Uncertainty Uncertainty Uncertainty
of CC of Vol Mat of DF of Balance of Repea of CRM

v
re
Sources of uncertainty

Figure 4.- Contribution with relative uncertainty for each of the sources of
uncertainty. er
4. Conclusions
It was possible to validate and estimate the analytical uncertainty of the ASTM-
pe
E1898-21 method for the determination of silver using FAAS by adhering to the
recommendations established by international regulations. The relevant results of the
validation process showed precision values ranging from 1.67 to 2.04 (%CV) under
repeatability conditions, while the veracity showed values from 0.0131 to 0.1445 for the
bias evaluated through Z-score and the percentages of accuracy from 100.7 to 104.8. The
validation of the analytical method was corroborated with a certified reference material
ot

(CDN-GS-1Z) with which precision values were obtained under repeatability conditions
of 1.677 (%CV) and recovery percentages of 100.65 (% rec). On the other hand, the
evaluation of the relative combined uncertainty showed values from 0.0046 to 0.1166 for
the different sources of uncertainty evaluated, whereas the percentage of relative
tn

uncertainty was 2.62% and expanded uncertainty obtained was 1.2 mg/L for silver with a
reliability factor of 95% (ĸ = 2).

Author Contributions: Conceptualization, resources and project administration, Martínez-

Peñuñuri, R.; methodology, supervision, Vázquez-Bustamante, P.; validation, and formal analysis,
rin

Montoya-Blumenkron, M., Estrada-González, D.I., Chávez-Orduño J., and Parga-Torres J.R.;

investigation, data curation, writing—original draft preparation, writing—review and editing,
García-Alegría, A.M. All authors have read and agreed to the published version of the manuscript.”
Please turn to the CRediT taxonomy for the term explanation. Authorship must be limited to those
who have contributed substantially to the work reported.
ep

Funding: This research received no external funding.

Data Availability Statement: Not applicable.
Acknowledgments: The authors appreciate the support granted by the Metallurgical Technological
Laboratory, S.A de C.V. Hermosillo, Sonora, Mexico, to carry out this research work.
Pr

Conflicts of Interest: The authors declare no conflict of interest.

15

This preprint research paper has not been peer reviewed. Electronic copy available at: https://ssrn.com/abstract=4347436
 16 of 18

References 582

ed
1. Liu, Y.H.,;Wan, B.; Xue, D.S. Sample Digestion and Combined Preconcentration Methods for the Determination of Ultra- 58
Low Gold Levels in Rocks. Molecules. 2019. 8;24(9):1778.
2. Palyanova, G.A. Gold and Silver Minerals in Sulfide Ore. Geol. Ore Deposits 62, 383–406 (2020). 58
https://doi.org/10.1134/S1075701520050050
3. Huang, D., Ding, X., Wu, C. and Chang, C. (1991), Mineralogy and Mode of Occurrence of Gold, Silver and Bismuth of the 58

iew
Caijiaying Lead-Zine-Silver Deposit, Hebei Province. Acta Geologica Sinica - English Edition, 4: 371-385.
https://doi.org/10.1111/j.1755-6724.1991.mp4004003.x
4. Calderon-Rodarte, J., López-Valdivieso, A., Aragón-Piña, A., Reyes-Bahena, J. L., Gallegos-García, M. I. L., Zapata- 59
Velázquez, A., Robledo-Cabrera, A. (2017). Mineralogy and silver distribution in argentiferous manganese ores from La
Encantada mines in Mexico. Physicochemical Problems of Mineral Processing, 53(1), 591-600.
https://doi.org/10.5277/ppmp170146
5. Birch, W. D. (2017). Mineralogy of the Silver King deposit, Omeo, Victoria. Proceedings of the Royal Society of Victoria, 59

v
129(1), 41-52.
6. Drif B, Taha Y, Hakkou R, Benzaazoua M. Recovery of Residual Silver-Bearing Minerals from Low-Grade Tailings by Froth 59
Flotation: The Case of Zgounder Mine, Morocco. Minerals. 2018; 8(7):273. https://doi.org/10.3390/min8070273

re
7. Swinkels, L. J., Burisch, M., Rossberg, C. M., Oelze, M., Gutzmer, J., & Frenzel, M. (2021). Gold and silver deportment in 59
sulfide ores–A case study of the Freiberg epithermal Ag-Pb-Zn district, Germany. Minerals Engineering, 174, 107235.
8. Saha, M. C., Baskey, R., & Lahiri, S. (2015). Determination of Ag and Cd in Soil and Sediment Samples by Graphite Furnace 60
Atomic Absorption Spectrometry (GFAAS). Atomic Spectroscopy, 36(4), 177.
9. Khalid, M. (2017). Gold cyanidation: gold associated with silver minerals embedded within base-metal sulphide matrices. 60
Thesis. Université Laval. Quebec, Canada.
10.
er
De La Calle, I., Cabaleiro, N., Costas, M., Pena, F., Gil, S., Lavilla, I., & Bendicho, C. (2011). Ultrasound-assisted extraction
of gold and silver from environmental samples using different extractants followed by electrothermal-atomic absorption
60

spectrometry. Microchemical Journal, 97(2), 93-100.

11. Das, A., & Saha, M. C. (2013). Rapid procedure for the quantitative determination of gold and base metals by graphite 60
pe
furnace and flame atomic absorption spectrometry. Atomic Spectroscopy, 34(2), 60-65.
12. Díaz, D., Hahn, D. W., & Molina, A. (2017). Quantification of gold and silver in minerals by laser-induced breakdown 60
spectroscopy. Spectrochimica Acta Part B: Atomic Spectroscopy, 136, 106-115.
13. Drif, B., Taha, Y., Hakkou, R., & Benzaazoua, M. (2018). Recovery of residual silver-bearing minerals from low-grade tailings 61
by froth flotation: The case of zgounder mine, morocco. Minerals, 8(7), 273.
14. Petrović, N., Buđelan, D., Cokić, S., & Nešić, B. (2001). The determination of the content of gold and silver in geological 61
samples. Journal of the Serbian Chemical Society, 66(1), 45-52.
ot

15. Wang, Y., Baker, L. A., & Brindle, I. D. (2016). Determination of gold and silver in geological samples by focused infrared 61
digestion: A re-investigation of aqua regia digestion. Talanta, 148, 419-426.
16. Shen, R. (2021). Determination of Silver in Copper Concentrate by Atomic Absorption Spectrometry. Advances in Material 61
Science, 5(1).
tn

17. Helmeczi, W., Helmeczi, E., Baker, L. A., Wang, Y., & Brindle, I. D. (2018). Development of a general acid method for the 61
digestion of gold ore samples together with a comparison of extraction solvents for gold and determination by microwave-
induced plasma-atomic emission spectrometry (MIP-AES). Journal of Analytical Atomic Spectrometry, 33(8), 1336-1344.
18. ASTM E1898-21. Standard Test Method for Determination of Silver in Copper Concentrates by Flame Atomic Absorption 62
Spectrometry. (2021). ASTM International.
19. ISO/IEC 17025:2017. General requirements for the competence of testing and calibration laboratories. (2017). 62
rin

20. User manual. Agilent TecnologyFlame Atomic Absorption Spectroscopy. Method Development ePrimer. (2021). Santa 62
Clara, CA. United State.
21. Bohrer D., do Nascimento P.C., Binotto R. y Becker E.J. (2003). Influence of the glass packing on the contamination of 62
pharmaceutical products by aluminum. Part II1: Interaction container-chemicals during the heating for sterilization. J. Trace
Elem. Med. Biol. 17(2), 107-115. DOI: 10.1016/S0946-672X(03)80006-8
ep

22. Kadis, R. Evaluation of measurement uncertainty in volumetric operations: the tolerance-based approach and the actual 63
performance-based approach. Talanta, 64 (2004) 167–173.
23. Swati, T.S.S., Saxena, R.K., and Gupta, P.K. Development and validation of method with evaluation of measurement 63
uncertainty for the speciation analysis of chromium by ion chromatography, MAPAN-J. Metrol. Soc India, 30(2) (2015) 131–
137.
24. Batista, E., Pinto, L., Filipe, E., and van der Veen, A.M.H. Calibration of micropipettes: test methods and uncertainty analysis. 63
Pr

Measurement, 40 (2007) 338–342.

25. Magnusson, B. and Ornemark, U. Eds, Eurachem guide: the fitness for purpose of analytical methods, a laboratory guide to 63
method validation and related topics, United Kingdom, 2nd edition, (2014) 45. http://www.eurachem.org,
26. Barwick, V. Preparation of calibration curves. A guide to best practice, Milestone Reference: KT2/1.3 LGC/VAM/2003/032 63
(2003) 1–27.

16

This preprint research paper has not been peer reviewed. Electronic copy available at: https://ssrn.com/abstract=4347436
 17 of 18

27. Nielsen,J.J. Microsoft Excel, Microsoft Official Academic Course, Wiley, (2016). 64

ed
28. Thompson, M., Ellison, S.L.R., and Wood, R. International union of pure and applied chemistry technical report. 64
Harmonized Guidelines for Single-Laboratory Validation of Methods of Analysis (IUPAC Technical Report). Pure and
Applied Chemistry, 74(5) (2002) 835–855.
29. Ellison, S.R.L. Implementing measurement uncertainty for analytical chemistry: the Eurachem guide for measurement 64

uncertainty. Metrologia, 51 (2014) 199–205.

iew
30. Ellison S.L.R., and Williams, A. EURACHEM/CITAC guide CG 4, quantifying uncertainty in analytical measurement, 64

United Kingdom, 3rd edition, (2012) 48.

31. Park, J., Nam, G., and Choi, J.O. Parameters in cause and effect diagram for uncertainty evaluation. Accreditation and 64
Quality Assurance, 16 (2011) 325–326.
32. Basak, S., and Kundu, D. Evaluation of measurement uncertainty in determination of lead in glass materials by a standard 65
complexometric method. MAPAN-J. Metrol. Soc India, 27 (2012) 175–182.

v
33. Solaguren-Beascoa, M., Ortega, V., and Serrano, R. On the uncertainty evaluation for repeated measurements. J. Metrol. 65
Soc. India, 29 (2014) 19–28.
34. Gates, K., Chang, N., Dilek, I., Jian, H., Pogue, S., and Sreenivasan, U. The Uncertainty of reference standards, a guide to 65

re
understanding factors impacting uncertainty, uncertainty calculations, and vendor certifications. J. Anal. Toxicol., 33 (2009)
532–539.
35. González, A.G., and Herrador, M.A. A practical guide to analytical method validation, including measurement uncertainty 65
and accuracy profiles. Trends Anal. Chem., 26 (2007) 227–238.
36. Mortada, W. I.; Ali, Z.A. A.; & Hassanien, M. M. (2013). Cloud point extraction of Pd (II), Au (III), and Ag (I) prior to their
er 66
determination by graphite furnace atomic absorption spectrometry. Canadian Journal of Chemistry, 91(12), 1219-1224.
37. De La Calle, I., Cabaleiro, N., Costas, M., Pena, F., Gil, S., Lavilla, I., & Bendicho, C. (2011). Ultrasound-assisted extraction 66
of gold and silver from environmental samples using different extractants followed by electrothermal-atomic absorption
spectrometry. Microchemical Journal, 97(2), 93-100.
38. Abolhasani, J., Amjadi, M., & Ghorbani, K. E. (2013). Ultra-trace determination of copper and silver environmental samples 66
pe
by using ionic liquid-based single drop microextraction-electrothermal atomic absorption spectrometry. Journal of
Chemical Health Risks 3(3): 29-42, 2013.
39. Resano, M., Garcia-Ruiz, E., Aramendía, M., & Belarra, M. A. (2005). Solid sampling-graphite furnace atomic absorption 66
spectrometry for Hg monitoring in soils. Performance as a quantitative and as a screening method. Journal of Analytical
Atomic Spectrometry, 20(12), 1374-1380.
40. Naeemullah, K. T., Afridi, H. I., Shah, F., Arain, S. S., Brahman, K. D., Ali, J., & Arain, M. S. (2016). Simultaneous 67
determination of silver and other heavy metals in aquatic environment receiving wastewater from industrial area, applying
ot

an enrichment method. Arab J Chem 9 (1): 105–113.

41. Fraser, A. W. (2015). Minimising Uncertainty in Measurement and Improving Limit of Detection in Gold Bearing Materials 67
from Concentrations Predicted by Linear Regression in Atomic Absorption Spectrometry (Doctoral dissertation, University
of Johannesburg (South Africa)).
tn

42. Leśniewska, B., Kisielewska, K., Wiater, J. et al. Fast and simple procedure for fractionation of zinc in soil using an 67
ultrasound probe and FAAS detection. Validation of the analytical method and evaluation of the uncertainty budget.
Environ Monit Assess 188, 29 (2016).
43. Singh N. A rugged, precise and accurate new gravimetry method for the determination of gold: an alternative to fire assay 68
method. Springerplus. 2012 Aug 16;1:14.
rin

44. Khorshed, M. A. (2012). Method validation of trace elements in water by atomic absorption spectrometer determination. 68
Journal of Plant Protection and Pathology, 3(3), 239-251.
45. Damasceno, J. C. , Couto, P. R. . Methods for Evaluation of Measurement Uncertainty. In: Anil, A. , editor. Metrology 68
[Internet]. London: IntechOpen; 2018 [cited 2022 Oct 12]. Available from: https://www.intechopen.com/chapters/59818 doi:
10.5772/intechopen.74873
46. Garcia-Alegria, A. M., Canez-Carrasco, M. G., Serna-Felix, M., Encinas Soto, K. K., & Gomez-Alvarez, A. (2018). Estimation 68
ep

of uncertainty in the determination of serum electrolytes (Na, K, Ca, Mg) by flame atomic absorption spectroscopy.
MAPAN, 33(2), 99-112.
47. García-Alegría, A. M., Gómez-Álvarez, A., Anduro-Corona, I., Burgos-Hernández, A., Ruiz-Bustos, E., Canett-Romero, R., 69
& Astiazarán-García, H. F. (2017). Estimation of the expanded uncertainty of an analytical method to quantify aluminum in
tissue of Sprague Dawley rats by FAAS and ETAAS. MAPAN, 32(2), 131-141.
Pr

693
694
695
696
697
698

17

This preprint research paper has not been peer reviewed. Electronic copy available at: https://ssrn.com/abstract=4347436
 18 of 18

699

ed
v iew
re
er
pe
ot
tn
rin
ep
Pr

18

This preprint research paper has not been peer reviewed. Electronic copy available at: https://ssrn.com/abstract=4347436