Preparation of the cadmium calibration solutions

Monoelemental calibration solutions are typically prepared under full gravimetric control by dissolving a high-purity metallic portion of the element in acid, and diluting the digest to a predefined elemental mass fraction ranging from 0.1 to 10 g kg−1. A small excess of nitric acid is added to enhance the stability of the CRMs. Each NMI prepared a cadmium solution with a nominal mass fraction of 1 g kg−1 dissolving pre-weighted cadmium metal in concentrated nitric acid and diluting the digest to the desired concentration with ultrapure water (resistivity > 18 MΩ cm). The nitric acid was purified in-house at each NMI by double sub-boiling distillation starting with acids of the highest commercial purity available (Suprapur®, Merck). INM(CO) used a molded PFA purification system (DST-1000-230, Savillex, USA) while TÜBİTAK-UME used quartz distillation units (duoPUR, Milestone, Italy). The acid was added in excess to a final mass fraction of approximately 2%. The solutions were thoroughly homogenized before packaging.

TÜBİTAK-UME prepared the sample UME-CRM-2211 using granulated, assayed, high-purity cadmium metal (Alfa Aesar, Puratronic, 1–3 mm shot, lot X15E013) and conducted substitution weighing for all steps. The UME-CRM-2211 was aliquoted into 125-mL high-density polyethylene (HDPE) bottles with approximately 100 g solution in each bottle. Conversely, INM(CO) prepared the sample INM-014-1 using high-purity cadmium metal foil (Sigma-Aldrich) pre-cleaned by etching in hydrochloric acid, water, and methanol and dried under argon flow. The INM-014-1 was aliquoted into sealed glass ampoules with a net volume of ca. 10 mL. Both NMIs provided each other with three units of their calibration solutions. The final presentation of the solutions is shown in the pictures at the bottom of Fig. 1.

Analytical methodology at TÜBİTAK-UME

TÜBİTAK-UME determined the purity of a granulated cadmium metal and certified it as a primary cadmium standard using the PDM described in the “Impurity assessment of high-purity cadmium” section. The high-purity cadmium metal was stored in an argon-filled glove box with controlled humidity and oxygen levels to prevent oxidation. This certified cadmium standard served both as the starting material for the gravimetric preparation of its cadmium monoelemental calibration solution (“Gravimetric preparation” section) and as a traceable calibrant for mass fraction measurements using the HP-ICP-OES method (“HP-ICP-OES measurements for the determination of cadmium mass fraction” section). The uncertainties associated with all measurements were estimated according to the Guide to the Expression of Uncertainty in Measurement (GUM) [11] using the software GUM Workbench [12].

The assigned mass fraction of cadmium in the UME-CRM-2211 calibration solution and its uncertainty were determined by combining gravimetric preparation and HP-ICP-OES measurement results using the Type B Model of Bias (BOB) procedure, which incorporates the bias in the average of the results of the methods as a source of uncertainty [13]. Combining gravimetric values with HP-ICP-OES measurements is a recognized method for assigning values to monoelemental calibration solutions.The HP-ICP-OES method was also used to measure the cadmium mass fraction in solution INM-014-1. The analytical routes chosen by TÜBİTAK-UME are presented on the left side of Fig. 1.

Impurity assessment of high-purity cadmium

TÜBİTAK UME followed case 3 of the “Roadmap for the purity determination of pure metallic elements” published by CCQM IAWG, which outlines an impurity assessment approach with expanded measurement uncertainties equal to or less than 0.01% [7, 14]. The impurity assessment approach primarily involves determining and subtracting all potential impurities from 100% [15]. TÜBİTAK-UME developed and validated high-resolution inductively coupled plasma mass spectrometry (HR-ICP-MS), inductively coupled plasma optical emission spectrometry (ICP-OES), and carrier gas hot extraction (CGHE) methods for quantitatively determining elemental impurities in high-purity cadmium metal. The validated methods enabled the quantification of 73 elements in the periodic table, excluding radioactive elements, halogens, and noble gases from the impurity assessment. Impurities present at levels below the corresponding limits of detection (LOD) were considered to have a mass fraction value equal to half the LOD, with expanded relative uncertainties set at 100%.

ICP-OES and HR-ICP-MS measurements for purity determination

Commercial multi-element standard solutions (HPS, solutions A, B, and C) were used as calibrants for the impurity measurements using ICP-OES (Arcos 2, Spectro, Germany) and HR-ICP-MS (Element 2, Thermo Finnigan, Germany). Solution A contains Al, As, Ba, Be, Bi, B, Cd, Ca, Ce, Cs, Cr, Co, Cu, Dy, Er, Eu, Gd, Ga, Ho, In, Fe, La, Pb, Li, Lu, Mg, Mn, Nd, Ni, P, K, Pr, Re, Rb, Sm, Sc, Se, Na, Sr, Tb, Tl, Th, Tm, U, V, Yb, Y, and Zn in 2% HNO3. Solution B contains Sb, Ge, Hf, Mo, Nb, Si, Ag, Ta, Te, Sb, Ti, W, and Zr in 2% HNO3 and trace amount of hydrofluoric acid (HF). Solution C contains Au, Ir, Os, Pd, Pt, Rh, and Ru in 15% hydrochloric acid (HCl). NIST Standard Reference Materials (SRM) 3133 and 3181 were used as calibrants for mercury (Hg) and sulfur (S), respectively. All samples and standards were prepared gravimetrically. Analysis of trace impurities in high-purity metals with ICP-MS is subject to matrix effects; hence, standard addition calibration with internal standards (tungsten for Solution A, yttrium for Solutions B and C) was performed to mitigate these effects. Given the rapid exothermic reactions of nitric acid with most metals, perfluoroalkoxy (PFA) tubes (Savillex) were chosen over polypropylene (PP) tubes (Brand) for dissolving cadmium metal due to their greater resistance to hot nitric acid and lower risk of metal leaching from the inner surface.

For HR-ICP-MS measurements, three sets of cadmium stock solutions were prepared for the analysis of elements in solutions A (including Hg and S), B, and C, respectively. Each set consisted of six replicate solutions prepared by dissolving 0.5 g cadmium metal with 2.5 mL HNO3 in 15-mL PFA tubes and diluting to a final mass of 10 g with ultrapure water. Before diluting to the final mass, 0.1 mL HF and 0.5 mL HCl were added to the solution sets for elements in solutions B and C, respectively. The standard addition calibration solutions and sample solutions were prepared by diluting the stock cadmium solutions after the addition of internal standard and different spiking levels of the multi-element standard solutions. All calibration standards and sample solutions contained 500 mg kg−1 of cadmium and 5 µg kg−1 internal standard. Multi-element standard additions were made at 1, 2, 5, 10, 20, and 50 µg kg−1. Due to the high-matrix content of the solutions, certain parts of the HR-ICP-MS instrument (sampler cone, skimmer cone, extraction lens, torch, injector, spray chamber, and peristaltic pump tubing) required cleaning after the study. Proper cleaning of the sampler cone, skimmer cone, and extraction lens is critical for HR-ICP-MS after measurements of high Cd-containing samples, as it is necessary to ensure accurate determination of low levels of Cd. The extraction lens in HR-ICP-MS consists of two parts made of graphite, which were soaked in 5% HNO3, rinsed with ultrapure water, then dried, and reused. HR-ICP-MS operating conditions for impurity measurements are given in Table S1 (see Electronic Supplementary Material Table S1).

Similarly, for ICP-OES measurements, another three sets of cadmium stock solutions were prepared for the analysis of elements in solutions A (including Hg and S), B, and C, respectively, with three replicates for each set. The solutions were prepared by dissolving 1 g cadmium metal with 5 mL HNO3 in 15-mL PFA tubes and diluting to a final mass of 10 g with ultrapure water. Before diluting to the final mass, 0.1 mL HF and 0.5 mL HCl were added to the solution sets for elements in solutions B and C, respectively. The standard addition calibration solutions and sample solutions were prepared by diluting the stock cadmium solutions after the addition of internal standard and different spiking levels of the multi-element standard solutions. All calibration standards and sample solutions contained 10,000 mg kg−1 of cadmium and 0.3 mg kg−1 internal standard. Multi-element standard additions were made at 10, 20, 50, 100, and 200 µg kg−1. ICP-OES measurements were performed in the axial view. In axial view analysis of high-matrix samples, metal ions can accumulate on the surface of the cone and torch. Failure to clean these components thoroughly can lead to arcing during plasma ignition, resulting in torch deformation. Therefore, the cone and torch were cleaned after each sequence. ICP-OES operating conditions for impurity measurements are shown in Table S2 (see Electronic Supplementary Material Table S2).

Due to the lack of reference materials for elemental impurities in high-purity cadmium metal on the market, accuracy testing was performed by adding a known concentration of the standards to dissolved cadmium solutions. For this purpose, 5 µg kg−1 and 50 µg kg−1 multi-element spiked cadmium solutions were prepared for HR-ICP-MS and ICP-OES, respectively.

CGHE measurements for purity determination

Measurements of carbon (C), oxygen (O), nitrogen (N), and hydrogen (H) impurities in the high-purity cadmium material were performed by CGHE, also called inert gas fusion. This technique involves melting the sample at high temperatures and measuring the analytes in the gas phase. In the CGHE instrument (Inductar EL Cube, Elementar, Germany), nitrogen is determined with a thermal conductivity detector (TCD), while an electronic hydrogen sensor (EHS) is utilized for hydrogen determination. Carbon and oxygen are first converted to carbon dioxide (CO2) and then determined with an infrared detector (IR). The instrument offers two measurement modes: ONH and C modes. The ONH mode allows for simultaneous measurement of O, N, and H, while the C mode is specifically for C measurements. Although switching between modes is not time-consuming, it is discouraged to do it more than once per day to ensure system stabilization, as the instrument requires purging. The CGHE method is commonly employed in the metal and ceramic industries, with limited scientific publications available on C, O, N, and H measurements in high-purity metals using this method [16,17,18]. Calibration and quality control for O, N, H, and C measurements were performed using the reference materials detailed in Table S3 and Table S4, respectively (see Electronic Supplementary Material Table S3 and Table S4). Since there is no reference material in the market with known amounts of C, O, N, and H in high-purity cadmium metal, the accuracy test was performed using steel reference materials.

Gravimetric preparation

TÜBİTAK-UME used metrological weighing to ensure high accuracy of cadmium mass fraction in the gravimetric preparation of UME-CRM-2211. This method involves placing reference standards (R) and test weights (T) on the weighing pan in a cycle R/T/T/R, determining the weighing difference between the test and reference as the average of multiple cycles. 60 g of Cd metal was carefully weighed into a pre-cleaned and pre-weighted 1-L PFA bottle. The weighing differences for the empty PFA bottle and the bottle with cadmium metal were measured by a balance with a mass resolution of 0.1 mg (MSA524S-100-DA, Sartorius AG, Germany). Sub-boiled concentrated nitric acid was then added to the bottle in a sufficient amount to achieve a mass fraction of approximately 2% in the final solution. Upon complete dissolution of the Cd metal, the solution was diluted to 1 L with ultrapure water, ensuring thorough homogenization. Subsequently, the solution was transferred to a pre-cleaned and pre-weighted 60-L HDPE bottle (Kautex, Germany). After rinsing the PFA bottle with ultrapure water several times, the rinse water was added to the HDPE bottle to ensure no residual material was left behind. The HDPE bottle was then filled with ultrapure water to a final mass of 60 kg. To achieve temperature equilibrium, the solution was homogenized in the same laboratory before the final weighing was performed. The weighing differences for the solution and the HDPE bottle were measured using an electromagnetically compensated balance with a mass resolution of 0.1 g (ID5, Mettler Toledo), which was integrated into the automatic weighing system. To ensure accurate measurements, the system was installed on an isolated concrete block to eliminate the transmission of vertical vibrations from the building to the system. Additionally, the system was housed in a plexiglass cabin to reduce airflow and minimize external influences on the measurements. In each measurement, the reference standards and the test weight were entirely removed from the weighing pan, and the balance indicator was zeroed using the zeroing function. For all weighings, OIML E2 class weight standards were utilized as references [19].

Cadmium mass fraction in the CRM solution was calculated using Eqs. 1–3. The symbols used in the equations are given in Table 1.

Table 1 The symbols and the definitions used in Eqs. 1 to 3

$$_^= \frac_}_} \times _$$

(1)

$$_= \frac_^ \times _^ - _ \times _+ _^}_}$$

(2)

$$_= \frac_^ \times _^ - _ \times _-_ \times _ + _^}_}$$

(3)

HP-ICP-OES measurements for the determination of cadmium mass fraction

ICP-OES stands out as one of the most common elemental analysis techniques renowned for its high precision and accuracy. In the determination of cadmium mass fraction in the solutions, the high-performance ICP-OES method (HP-ICP-OES), developed by NIST, was utilized [20,21,22]. This method incorporates internal standardization to minimize short-term instrumental noise, a drift-correction procedure to minimize instrumental drift, gravimetric sample preparation, and a robust experimental design [23]. Furthermore, to enhance accuracy and reduce measurement uncertainty, Winchester et al. employed an exact matching procedure [24, 25]. This involves meticulously aligning analyte mass fractions and internal standard mass fractions among the calibration standards and sample solutions, ensuring close matching of analyte/IS mass fractions, in addition to matrix compositions. All measurements were conducted using the Spectro Arcos 2 ICP-OES. A crossflow nebulizer with a Scott-type double-pass spray chamber was preferred over a concentric nebulizer with a cyclonic spray chamber to achieve more stable signals. Additionally, a small internal diameter (0.5 mm ID) autosampler probe was chosen to reduce the take-up time and facilitate efficient sample flow. The concentrations of cadmium and the IS (scandium) in the solutions, as well as the measurement wavelength values, were determined based on Salit et al. [20,21,22]. Signal intensities for 10 mg kg−1 Cd and 0.5 mg kg−1 Sc were compared at the selected wavelengths to ensure compatibility. The operating conditions for HP-ICP-OES measurements are provided in Table S5 (see Electronic Supplementary Material Table S5).

The instrument underwent a preconditioning period of at least 45 min before the measurement to ensure stability. However, it was observed that the instrumental drift in the initial measurements was larger compared to subsequent ones, independent of the preconditioning time. To mitigate this issue and avoid unstable measurements at the beginning of the sequence, the sequence was halted after the first five measurements and then restarted from the beginning to obtain more stable signals. Autosampler probe rinsing between samples in the sequence was omitted for two reasons: firstly, to reduce the total analysis time, and secondly, to eliminate potential carry-over between the rinse solution and the sample solutions. Given that all solutions in the sequence had similar elemental compositions and mass fractions, the risk of carry-over between solutions was minimized by employing long take-up times. Five independent calibration solutions were gravimetrically prepared from primary Cd standard by dissolving 0.1 g Cd metal with sufficient HNO3 to achieve a final mass fraction of approximately 2% HNO3 in 125-mL HDPE bottles and diluting to a final volume of 100 mL with ultrapure water. Two 50 g calibration solutions were then prepared from each primary Cd standard by aliquoting 0.5 g into 50-mL PP centrifuge tubes (Brand), spiked with 0.5 g IS stock solution (50 mg kg−1 Sc), and diluted with ultrapure water. Similarly, three sub-samples from each of the 3 CRMs and 3 verification solutions were prepared in the same manner as the calibration solutions. To mitigate evaporation effects during sample preparation, the caps of the tubes were kept closed. Buoyancy correction was not applied in the weighings. The prepared calibration, sample, and verification solutions (28 solutions in total) were analyzed using a “randomized complete block sequence” as described by Salit and Turk [22]. This sequence design, crucial for the drift correction procedure—one of the four pillars of the HP-ICP-OES method—ensures each sample is measured once in a random order, repeated up to the desired number of measurements. In this study, each sample was measured six times. Approximately 32 mL of solution was used for six replicates under the given conditions. Since the ICP-OES software (Smart Analyzer Vision) does not allow multiple measurements from a single sample position in the same measurement sequence, the X, Y, and Z coordinates of the autosampler probe had to be redesigned. The analyte-to-internal standard signal ratios produced by ICP-OES were re-evaluated using the drift correction procedure. Cadmium mass fractions in the solutions were calculated using Eqs. 4–8, with the symbols defined in Table 2.

Table 2 The symbols and the definitions used in Eqs. 4 to 8

$$_= \frac_^_} \times _$$

(5)

$$_= \frac_ \times _^}_^}$$

(6)

$$_= \frac_^_} \times _$$

(7)

$$_= \frac_ \times _^}_}$$

(8)

The accuracy of the measurement results was verified using NIST SRM 3108 Cadmium Standard Reference Material. Three independent solutions were prepared by gravimetrically diluting the NIST SRM 3108 at a cadmium mass fraction of 1 g kg−1 and measured using the HP-ICP-OES method on three different days. Three sub-samples from each of the three NIST SRM 3108 solutions were prepared each day for analysis.

Analytical methodology at INM(CO)

INM(CO) performed complexometric gravimetric titrations using an EDTA salt, with the endpoint indicated potentiometrically via ion-selective electrodes. The methodology consists of two steps: first, characterizing the EDTA salt in terms of the amount of complexing agent concentration, and second, determining the cadmium mass fraction in the sample solutions using EDTA solutions of known concentration. This procedure is similar to the “classical metrological approach” reported by Felber et al. for the complexometric characterization of copper calibration solutions [26]. However, the titration reported by Felber et al. is weight-volumetric, where most of the titrant is added gravimetrically at the beginning and then volumetrically near the endpoint [26]. In contrast, the approach presented here employs gravimetric burettes modified to deliver small portions of titrant solution, allowing for the entire titration to be performed gravimetrically.

The characterization of the EDTA salt utilizes the lead nitrate NIST SRM 928 as the standard for the amount of substance [27]. In summary, standard lead solutions were prepared to a known lead concentration of approximately 4.8 mmol/kg in 2% HNO3 and EDTA sample solutions were prepared gravimetrically to an estimated concentration of 10 mmol/kg in ultrapure water. For the EDTA solutions preparation, the disodium dihydrate form of EDTA was dried at 80 °C until constant weight was achieved (Halogen Moisture Analyzer HC103, Mettler Toledo, scale division 0.1 mg), and complete dissolution of the salt was ensured by placing the solutions in an ultrasonic bath for at least 30 min. The gravimetric preparation data relates the EDTA concentration in the titrated samples to the purity of the starting reagent. Air buoyancy correction factors were applied for the gravimetric preparation of solutions. Then, in a single titration, a weighted aliquot of the lead standard solution (approximately 5 g) was placed in a PP titration vessels, and sodium tartrate (240 µmol) and ammonium hydroxide (10 mmol) were added. Ammonium hydroxide creates an alkaline medium, while the sodium tartrate prevents the precipitation of lead hydroxide. The gravimetric burettes were PP syringes with nominal capacity of 5 mL (Precision Dispenser Tips, Brand), adapted with PP capillary tips. The burette was filled with EDTA solution, and its weight was set as the initial tare weight. The mass of added EDTA solution was measured by the difference between the current mass and the initial tare weight. The electric potential difference at a lead selective electrode was continuously monitored under magnetic stirring. Approximately 99.5% of the equivalent EDTA solution mass was added initially, with very small increments made near the endpoint. The endpoint was determined by the steepest change in the potentiometric titration curve. The result is calculated as shown in Eq. 9. The symbols used in the equation are provided in Table 3. The amount of complexing agent concentration is reported as the mass fraction of the dihydrate disodium EDTA salt in the solid reagent.

Table 3 The symbols and the definitions used in Eq. 9

$$_= \frac_\times _\times _}_^\times _\times _}\times \frac}_}}_-}_}\times \frac}_\times }_\times _}}_\times }_}\times }_$$

(9)

A potentially relevant source of uncertainty in the EDTA salt characterization using a lead nitrate SRM arises from the atomic weight of lead, which is used in the measurement model shown in Eq. 9. The reported standard atomic weight of lead spans a broad range due to the natural variability of its isotopic composition [28]. To minimize the impact of this uncertainty on the uncertainty budget, the lead atomic weight applicable to NIST SRM 928 was estimated by measuring the corresponding isotopic ratios using a quadrupole ICP-MS instrument (NexION 300D, Perkin-Elmer, PA, USA). Instrumental isotope fractionation effects were corrected through sample-standard bracketing (SSB) with the lead isotopic standard NIST SRM 981 [27].

After establishing the purity of the EDTA salt in terms of the amount of complexing agent concentration, it can be utilized as a transfer standard for determining the amount of metallic cations in solution. This second titration step follows a similar procedure to the EDTA characterization described earlier, using a cadmium-selective combination electrode instead of the lead-selective electrode. The cadmium mass fraction in the solution is calculated as shown in [4]. The symbols used in Eq. 10 are provided in Table 4. All uncertainty calculations were performed according to the GUM using the R package propagate [29].

Table 4 The symbols and the definitions used in Eq. 10

$$_=\frac_\times _ \times _}_\times _\times _}\times \frac_-_}_}\times _^\times _\times ^ +_+_$$

(10)

Accuracy of the cadmium mass fraction titration results was verified using the NIST SRM 3108 cadmium standard solution. Standard solutions were prepared by gravimetrically diluting NIST SRM 3108 to a cadmium mass fraction close to 1 g kg−1 in 2% HNO3 and titrated alongside the samples. Between two and four replicate titrations were performed for the solutions on four different days.