Constant-Current Coulometry and Ion Chromatography Bromide Determination to Characterize the Purity of the Potassium Chloride

A coulometria é considerada o método de maior qualidade metrológica para medição de quantidade de substância de padrões primários. Entretanto, a contaminação de sais de elevada pureza está presente em níveis traços sendo necessária a sua identificação e quantificação. Este trabalho apresenta os estudos desenvolvidos pelo Inmetro (Instituto Nacional de Metrologia, Normalização e Qualidade Industrial, Brasil) para determinar a pureza do cloreto de potássio (KCl) por titulação coulométrica a corrente constante, além da determinação do brometo como seu principal contaminante por cromatografia iônica (IC). Esses estudos foram necessários para desenvolver material de referência certificado (MRC) de KCl para garantir a qualidade dos resultados das medições químicas principalmente as de titulometria. A caracterização do candidato a MRC de KCl apresentou pureza de 99,9662%, a qual foi obtida após a correção do brometo, em fração mássica, obtido por IC, de 182,47 ± 3,03 mg kg-1 (incerteza expandida, k = 2, para um nível de confiança de aproximadamente de 95%).


Introduction
The potassium chloride is an essential vegetable nutrient used as a fertilizer and also an important chemical reagent used in large scale in chemical and pharmaceutical industries for various purposes.The purity required for most of these applications is about 98%. 1 The salt is used in various types of industrial products such as inks, soaps, detergents, foodstuffs, pharmaceuticals, as well as raw material for the preparation of other chemical substances such as potassium hydroxide and potassium carbonate.In addition, it is research foundation for pH, electrolytic conductivity and mono-element standard solutions. 2etrology in chemistry has been currently acknowledged as a relevant area in researches of analytical and measuring chemistry.Its importance comes from the need to produce reliable and uniform measurements in science, technology and international trade. 3In the analytical chemistry area, a great number of methods uses titration as the main technique to determine the concentration of a substance in a specific analyte.In order to carry out this analysis with accuracy, the laboratories need to use primary standards. 4,5owever, the primary standards should have properties such as high purity and homogeneity, definitely stable, precise composition, ability to give a thorough, simple and stoichiometric reaction with a titrant, easy to dry and handle. 6,7ne of the main functions of a National Metrology Institute (NMI) is to certify primary standards, e.g. to determine the purity of primary reference materials in order to develop the certified reference materials (CRMs).Inmetro (National Institute of Metrology, Standardization and Industrial Quality), the NMI of Brazil, aims to ensure traceability and reliability to the purity of salts, promoting the improvement of the quality of products and thus, the competitiveness of Brazilian industries.
Coulometric methods are capable of high precision and accuracy and are totally described by Faraday's law. 8oulometry is considered to have the potential to become a primary method 9 by the Consultative Committee for Amount of Substance (CCQM), Metrology in Chemistry.Inmetro has established a primary system of coulometry 10 in its Electrochemistry Laboratory (Label) to characterize and certify high-purity reference materials such as primary standards in order to guarantee the reliability of the chemical analysis performed by Brazilian laboratories.These reference materials can be used as primary standards in acid/base, precipitation, redox and complexometric titration techniques [11][12][13] to determine the concentration of diverse chemical substances.
Judging by the analysis of the KCl, 14 the main substance considered by the authors as an impurity to this work was bromide (0.05%).During the coulometric titration of the KCl, bromide ions are titrated indifferently as chloride ions.Therefore, the final result (which is related only to the concentration of chloride ion) would be over evaluated if both the titrations of the bromide ions and the chloride ions in the salt are considered.Hence, due to the electrochemical activity of the bromide ions, they should be analyzed with accuracy by using a sensitive and selective technique in order to be considered as impurity in the characterization of the candidate CRM of KCl by coulometric titration.The selected technique was ion chromatography (IC) using ion-exchange separation mode coupled with a membrane suppressed conductivity detector without sample pretreatment.The final result will then be considered as 100% of the concentration value obtained by coulometry minus the value of bromide concentration obtained from IC.
6][17] However, the column is overloaded when the chloride concentration is very large (which is often present as a 100-fold excess in environmental samples) interfering in the elution for being very close to the bromide, 18 and thus making the bromide quantification impossible when implementing common elutes like carbonate or borate buffers. 19Bicarbonate eluents (also mixed carbonate-bicarbonate eluents) have been used as the mainstay of IC for many years.The problem with these eluents is that the carbonic acid formed after suppression is susceptible to partial dissociation, leading to a background conductance that is somewhat higher than desired. 20Sodium carbonate is considered as an inorganic additive and its presence in the mobile phase influences the analytes retention times 21 and causes unstable baselines in the chromatograms. 20he ideal eluent is hydroxide since it forms water after suppression which has virtually zero conductance and therefore provides the perfect conductivity baseline. 20his work aims at presenting the studies which have been developed using the primary system of coulometry from Inmetro to characterize primary standards.For determining the purity of a KCl sample, coulometric titration was used to determine all the chloride contained in the sample as well as bromide determination as the main impurity through ion chromatography.The final result was obtained by the difference between the total chlorides from coulometric titration and the total bromides determined by ion chromatography.Initially, the preliminary studies will be shown to determine the purity of potassium chloride (KCl) using a standard reference material (SRM) from the National Institute of Standard and Technology (NIST).Then, there will be a focus on the results of an international comparison organized by CCQM to test the abilities of the metrology institutes to measure the amount content of chloride in KCl, called CCQM-K48.Finally, the characterization of the purity of KCl from a candidate CRM will be shown.

Coulometric titration procedure for KCl assay
The primary system of coulometry from Inmetro consists of equipment of higher metrological quality (higher stability, sensibility and precision).All the equipment that constitutes the coulometric system was published elsewhere. 10or determining the purity of KCl, the constant-current coulometric titration was used.In high-precision constantcurrent coulometry, charge is added in three stages. 10,13,22In the initial titration (pretitration), a small amount of analyte (spike) is added, and small charge increments at low current are used to locate the initial end point.The electrolyte is then prepared (without impurities) for the second titration called main titration, where the sample is added and 99.95% of the electrochemistry reaction is carried out by using higher current than the initial titration.In the final titration, the process is ended where the end point for the titration of the sample is detected, using small charge increments as in the initial titration.By Faraday's law, the charge consumed between the two end points is equivalent to the amount of the sample.The measurement equation is shown below as equation 1: The end-point detection was carried out by using the potentiometric technique with glass and silver-silver chloride electrodes.The end points were measured by non-linear regression analysis of the titration curves.The sample of KCl was dried in a muffle furnace at 500 °C for 6 h without crushing or grinding.After, it was placed in desiccator with silica gel to be cooled.The coulometric glass cell had a vertical design with two separate compartments: a cathodic and an anodic one.The cathodic compartment was inside an intermediate compartment which was separated from the anodic one through a silicate plug.The supporting electrolyte (250 mL) was a solution of 3.5 mol L -1 of perchloric acid (Merck, 70-72%).In perchloric acid, the current efficiency is close to 100% for current densities up to 100 mA cm -2 . 23Whenever not mentioned, a high purity silver rod (99.999%) was the anode with an initial area of ca. 8 cm 2 .Pt electrode (1 cm 2 area) was the cathode.The necessary current for the coulometric titration of KCl sample (weighing on the range of 300 mg by using a microbalance, Mettler Toledo, UMX 5, resolution of 0.1 mg) passed through 3 different stages: in the initial titration, it was passed a current of 10 mA; in the main titration, a current of 200 mA and for the final titration, 10 mA.All the weighing was corrected for buoyancy and the coulometric cell was maintained at 23 °C during the experiment.Firstly, argon gas of high purity (99.999%) was passed through the anodic compartment for deaeration.After, it was maintained above the anolyte during the titrations.All the titrations were carried out without the presence of light (with the coulometric cell shielded) to avoid dissolving the formed precipitate as well as photochemical decomposition of silver chloride which can cause errors in the titration. 13,18,24During the whole process, the anolyte solution was magnetically stirred.Figure 1 shows the schematic diagram of the coulometric cell used to determine the KCl purity linked to the primary system of coulometry.

Solutions and reagents
Ultra pure water was obtained by an Elix 10 reverse osmosis system (Milli-Q gradient A10 system (Millipore, Bradford, USA)) having a minimum resistivity of 18.2 MΩ cm.This water was used to prepare all the solutions.
The calibration solutions of bromide were prepared through gravimetric dissolution of appropriate masses of the standard reference material (SRM) 3184 from NIST 25 in order to have the six calibration solutions with concentration ranging from 0.05 to 1.50 mg kg -1 .The masses of bromide solutions were weighed in 50 mL capacity polypropylene (PP) flasks and completed to approximately 50 g in the same flasks.These solutions were prepared immediately before analysis.

Equipment
The used chromatographic system was a Dionex Model ICS 2000 equipped with an Ion Pac AS 18 analytical column and an Ion Pac AG 18 precolumn with internal diameter of 5 µm and dimensions of 4 mm × 250 mm and 4 mm × 50 mm, respectively.These columns were coupled with an ASRS ultra 4 mm membrane conductivity suppressor using a 99 mA current.A gradient pump device GS 50 was used.Sample injection volumes of 25 µL were used in a system coupled with an auto sample AS40.The separation was based on ion exchange chromatography using an isocratic eluent generator of sodium hydroxide.The eluent generator of KOH coupled with a continuously regenerated trap column was used to remove trace level anionic contaminants from the carrier deionized water, as well as to protect the ion exchange columns and system components.Conductivity signals were measured by a stabilized conductivity cell DS 6, which was controlled by an electrochemical detection module ED 50.
The chromatograms were obtained with the aid of a 6.5v Chromeleon software.Equipment and software were supplied by Dionex (Dionex Corporation, Sunnyvale, USA).
The analytical balance used to prepare all solutions was a Satorius, model ME235S, of 0.01 mg resolution.

Sample preparation
The sample from CCQM-K48 (called in this work as unknown sample) and the quality control (QC) sample (SRM 999b) were dried at 500 °C for 6 h without crushing or grinding.After being dried, they were placed in a desiccator with silica gel and cooled to room temperature before weighing it.The same procedure was used to sample preparation candidate CRM of KCl.
Aliquots of the samples were taken from the bottle to the weighing flask using polytetrafluorethylene (PTFE, Teflon ® , DuPont) coated spatula.Two aliquots of approximately 0.5 g of each sample (unknown, candidate CRM and QC) were taken to 50 mL capacity of the polypropylene flasks for the determination of bromide.The samples were diluted gravimetrically with ultra pure water to a final mass of about 50 g in the same flask.The salts were weighed and then homogenized for about 1 h.The samples were filtered using filter caps on the sample vials before introducing the chromatographic system by auto-sampler.Sixteen replicates of each aliquot were analyzed.The analyses were performed immediately after dissolution.

Assay of KCl purity from NIST SRM
In order to evaluate the accuracy of the results of the measurement using the primary system of coulometry from Inmetro, a constant-current coulometry technique was applied to determine the purity of potassium chloride using the SRM 999b from NIST.In Table 1, these preliminary studies related to the measurement results of the purity of the KCl by coulometric titrations are shown.
In Table 1, the results show good agreement with the certified value.However, considering the certified value of SRM 999b (99.977%), which is in the NIST certificate, 26 the measurement presented good precision (RSD 0.003%), but it was 0.025% biased.Therefore, this result suggests that the measurement needs to be improved in order to increase its sensitivity as well as its accuracy.One possibility is to use a silver electrode with high purity, since in the prior analysis these assays were carried out with a silver electrode with only 99.99% purity.This may have affected the current efficiency of the coulometric process due to the presence of less noble metals which can interfere in the generation of Ag + ion. 27Other possibility may be the non consideration of the bromide ion as the main impurity in the KCl. 26

Ion chromatography of the bromide impurity in KCl
In this study an analytical method for the determination of bromide impurity in KCl salt is provided.The best method optimization was obtained by using an eluent flow rate of 1.2 mL min −1 and eluent composition of 30 mmol L −1 KOH.In these conditions, the background conductivity obtained in the detector was around 0.55 mS cm −1 and 15.7 × 10 3 kPa (2243 psi) system pressure.A typical chromatogram of KCl sample using the described optimized conditions can be seen in Figure 2. The samples unknown (CCQM-K48) and candidate CRM were diluted 100 and 400 times, respectively, prior to analysis since the concentration of chloride was extremely high (about 100,000 times bigger than the bromide).Neither the resolution nor the efficiency was affected.No sample pretreatment was performed.
Analytical curves were obtained in a range between 0.05 and 1.5 mg kg -1 by using six calibration solutions.The correlation coefficient was better than 0.9996 and the sensitivity obtained was 0.43 mg kg -1 .
The limit of detection (LOD) 0.1 mg kg -1 and limit of quantification (LOQ) 0.3 mg kg -1 of the method were obtained from the standard deviation of the blank curve.Eight replicates were measured and the LOD was considered to be 3 times the standard deviation (SD) of these replicates.In an analogous approach, the LOQ was considered to be 10 times the SD. 28The method accuracy was evaluated by the relative deviation of the measurement from a control solution.The average of bromide mass fraction determined in the sample was 31.32 (mg kg -1 ) with expanded uncertainty of 0.41 mg kg -1 (k = 2) for a confidence level of approximately 95%.The uncertainty budget was estimated based on the guides EURACHEM/CITAC 29 and ISO GUM, 30 and the sources of uncertainty considered were: the dilution of calibration standards, the repeatability and the calibration curve, as can be seen in Table 2.The uncertainty associated to the repeatability of measurements was obtained from the standard deviation of the peak areas used to determine the bromide concentration in the samples.
One of the main uncertainty sources was the uncertainty associated to the calibration curve, representing the major contribution.This uncertainty is within the expected limits and represents about 0.5% of the final result.The authors understand that this is almost the possible limit with the approach used.

International comparison CCQM-K48
With the improvements in the coulometric titration process, and also considering the bromide ion as the main impurity in the KCl sample, Inmetro has participated in an international key-comparison named as CCQM-K48 for the determination of the purity of KCl.Table 3 shows the results of coulometric titration for KCl samples.The uncertainty budget of the coulometric titration result 2 for this comparison is shown in Table 4.According to the data, the main source of uncertainty came from the current.
The results for bromide with the used measurement methods are shown in Table 5. 2 The authors believe that even though the results were considered satisfactory by the organizers, they might be a little biased.The main source of bias is believed to be the presence of different chemical species of bromide in the salt.A chemical treatment of the sample may be necessary to determine the total bromine through ion chromatography.One possible change in sample preparation should be the use of more time for it to be equilibrated in solution before analysis.Other approaches should be studied in further works.The uncertainty expanded of 0.41 mg kg -1 is consistent with the repeatability and other uncertainty sources considered in the proposed matrix.
The used reference value in this international comparison, CCQM-K48, was the median of all results submitted by the participating laboratories.The results are shown with bromide correction, since it is the major contaminant in the KCl.The bromide assay was 0.00039 mol kg -1 ± 1.3% with coverage factor (k = 2), for a confidence level of 95%. 2 Then, the KCl purity after bromide correction (determined by Inmetro in this exercise) was 99.9787%, which is equivalent to an amount content of 13.41077 ± 0.0029 mol kg -1 , (k = 2) for a confidence level of 95%.
There was comparability among the participant metrology institutes, though the result from Inmetro, as  shown in Figure 3, is somehow lower than the median.These results showed that the measurements carried out by Inmetro needed some improvements.Assays should be carried out with increased sample weighs to select the best weigh which can reach better current efficiency and, consequently, better accuracy.Another problem could be the co-precipitation/adsorption of the chloride ion on AgCl precipitated. 2In order to avoid this problem, the dissolved KCl in water should be added gradually into the cell, instead of adding solid KCl sample at once into the coulometric cell.

Characterization of a candidate CRM
Table 6 shows the results of the candidate CRM that was assayed by coulometric titration, considering the bromide ion as the main impurity determined by ion chromatography.The mass fraction of the bromide ion in the KCl sample was 182.47 mg kg -1 with expanded   uncertainty of 3.03 mg kg -1 (k = 2; 95%).The contribution of the bromide ion in the characterized value by using constant-current coulometric titration was 0.017%.It corresponds to minus 0.0028 mol kg -1 on the purity value of the candidate CRM.Therefore, the final value for the characterization of the candidate CRM of KCl with its expanded uncertainty was 13.4090 ± 0.0032 mol kg -1 (k = 2; 95%), which corresponds to 99.9662% in purity.

Conclusions
The ion chromatographic method was successful for determining bromide as the main impurity in high purity KCl and all the chromatographic parameters were satisfactory for the analysis on this level.
The characterization results of KCl primary reference material using coulometric titration with ion chromatography will guarantee Inmetro the competence of measuring amount content with accuracy and reliability which is needed to develop CRM for acid, bases and salts.
After the preliminary studies carried out to characterize the purity of KCl and also to participate in an international comparison, the value was obtained as amount content (with its expanded uncertainty) of 13.4090 ± 0.0032 mol kg -1 (k = 2, considering 95% of confidence level) which corresponds to 99.9662% in purity for the characterization of the candidate CRM developed by Inmetro.This candidate CRM, after its certification, can be used to provide quality measurements results with reliability and traceability to the international system of units (SI) when carried out by chemical laboratories in Brazil, mainly those which use titrimetric analysis techniques.

( 1 )
where, n = amount content, I 1 = current during the initial titration, I 2 = current during the main titration, I 3 = current during the final titration, t 1 = total time of initial titration from the initial end-point, t 2 = time of the main titration, t 3 = total time of final titration up to the final end-point, z = charge number of the analyte, F = Faraday constant (96485.3383C mol -1 ), m = sample weight, cor = sample weight to air buoyancy correction and cor 1 = correction of the difference between the end point and time of filling the intermediate compartment of the cell.

Figure 1 .
Figure 1.Schematic diagram of the coulometric cell with the primary system of coulometry from Inmetro.

Figure 2 .
Figure 2. Typical chromatogram obtained from the separation of chloride and bromide in KCl sample.Mobile phase of 30 mmol L −1 KOH and flow rate of 1.2 mL min −1 .

Figure 3 .
Figure 3.Comparison results of the purity of KCl for CCQM-K48 with bromide correction. 2 The error bar means 2×u c (u c is the combined uncertainty).

Table 1 .
Purity of the KCl from SRM 999b determined by coulometric titration

Table 2 .
Uncertainties considered in the bromide measurements by IC for the CCQM-K48

Table 3 .
2esults of coulometric titration for purity of KCl in CCQM-K482 a n, measurement; b RSD, relative standard deviation.

Table 4 .
2ncertainty budget for purity of KCl by coulometry for CCQM-K482 a U, expanded uncertainty (coverage factor k = 2 for a confidence level of 95%).

Table 5 .
2esults of bromide measurements obtained from participants for CCQM-K482

Table 6 .
Coulometric titration results for a candidate CRM of KCl a n, number of measurements; b SD, standard deviation.