ESTIMATION OF MEASUREMENT UNCERTAINTY : A CASE STUDY FOR DETERMINATION OF CALCIUM ( CA ) CONTENT IN POWDERED TONIC FOOD DRINK USING ION

In participation of the laboratory on Proficiency Testing Scheme: Analysis of Proximates (Moisture, Fat, Protein, and Ash) and Minerals (Iron, Calcium, and Sodium) in Powdered Tonic Food Drink held by Food Nutrition Research Institute, Department of Science and Technology, Philippines (FNRI-DOST), Research Centre for Chemistry-LIPI conducted the estimation of uncertainty measurement in the determination of calcium (Ca) content in powdered tonic food drink using ion chromatography. The estimation of uncertainty measurement of laboratory participant of proficiency is needed for comparison with the standard deviation for proficiency assesment (σP


INTRODUCTION
The nutrition scenario in the South-East Asian region has changed dramatically in the last two decades.Especially among the urban segments of the community, diet related chronic diseases have been on the rise.The main cause of the deaths in most of these developing countries in the region is coronary heart disease.With the increase in these diseases in the region, there has been greater focus on the role of nutrition in the disorders.Consumers are paying greater attention to the nutritional value of their diets.The food industries are also increasing concentration of some nutrients while decreasing some of the others.Government agencies have intensified efforts to promote healthy dietary practices.
Nutrition labeling of foods is one of the strategies adopted to assist consumers in adopting healthy dietary practices.The primary objective of nutrition labeling is to describe the nutritional qualities of food product factually and informatively.It is aimed at providing a means for conveying information of the nutrient content on the label, thereby assisting consumers in making better food choices when planning daily meals.Nutrition labeling is equally important to the food industry as labeling provide a means for food manufacturers and retailers to become more aware of the nutritional properties of their products, and be encouraged to emphasize these properties to the consumers.Food manufacturers have a social responsibility to contribute positively to the healthy lifestyle programs of the health authorities.There is increasing interest in developing nutrition labeling around the world, even for developing countries, and a wide range of types and approaches are now being practiced.
The Food and Agriculture Organization of the United Nation (FAO) and the World Health Organization (WHO) has published the Codex Alimentarius standards and guidelines on food labeling that consists of collection of internationally adopted food standards, guidelines, codes of practice and other recommendation, with the purposes of protecting the health of consumers and ensuring fair practices in the food trade.In Indonesia, specific requirements for nutrition labeling in foods were provided in government regulation number 69:1999 pertaining to food labeling and advertisement.Nutrition labeling is mandatory for certain types of foods namely baby foods, dietary foods, milk and milk products and other foods as specified by the Director-General.The regulation also apply to foods making claims that they contain specific nutrients, including energy, protein, fat, and carbohydrate content, as well as levels of vitamins and minerals.Nutrition labeling is also mandatory for foods that required to be fortified or enriched with specific nutrients as required by the national legislations.The regulations are also applicable to the voluntary labeling of all other types of foods.
In order to support the implementation of the requirements of food labeling, the nutrients content in food product should be accurately determined using the certain analytical method by the food testing laboratories not only in industry but also in governmental agency that responsibility to monitor food products selling in public so that the valid information about the nutritional of food products can be provided to consumer factually.Therefore, the proficiency testing (PT) held by Food Nutrition Research Institute, Department of Science and Technology, Philippines (FNRI-DOST) covers seven nutrition labeling-related food components, namely moisture, fat, protein, ash, iron, calcium, and sodium.The objective of this PT is to evaluate the performance of participant laboratories in the analysis of moisture, fat protein, ash, iron, calcium, and sodium in powdered tonic food drink.This evaluation enables individual laboratories to compare their analytical result with those of other laboratories and demonstrate the validity and reliability of tests that they undertake.In addition, the regular participation of testing laboratories in wellorganized PTs is one of the quality assurances procedures that are included in the requirements for sustained accreditation to the ISO/IEC 17025 standard.
Participants in a PT scheme are required to report quantity values consistent with the definition of the measurand, which means that they need to apply a correction to the quantity value obtained directly from measurement to compensate for any systematic effect associated with differences between the actual measurement conditions and those specified for the measurand.Thus, the uncertainty of measurement indicating the analytical variability of a result has to be evaluated.It demonstrates how well the result represents the value of the quantity being measured in the test portion and also allows an assessment of the reliability of the result.The uncertainty provides an interval within which the value of measurand is believed to lie with a higher level confidence.Therefore, PT can no longer be limited to ensuring that participants's quantity values have satisfactory accuracy; it must include an evaluation of the reliability of the reported uncertainty associated with each measurement.
This paper describes the evaluation of uncertainty of measurement in determination of calcium (Ca) content in powdered tonic food drink using ion chromatography.The specification of measurand, source of uncertainty, standard uncertainty, combined uncertainty and expanded uncertainty from this measurement were evaluated and accounted.The purpose of the evaluation of uncertainty in this measurement is to provide the bias taken place which depended on the various components or measurands that effect to the measurement.This is required for the laboratory participating in the Proficiency Testing Scheme: Analysis of Proximates (Moisture, Fat, Protein, and Ash) and Minerals (Iron, Calcium, and Sodium) in Powdered Tonic Food Drink held by Food Nutrition Research Institute, Department of Science and Technology, Philippines (FNRI-DOST).
The estimation of uncertainty measurement of each laboratory is needed for comparison with the standard deviation for proficiency assesment (σ P ) that represents fitness-for-purposes over a whole application sector in proficiency testing scheme.

METHODOLOGY 2.1 Proficiency Testing Scheme
All the participant laboratories were requested to analyze the FNRI-DOST powdered tonic food drink test samples, for any or all the following analytes: moisture, fat, protein, ash, iron, calcium, and sodium, using the method of analysis routinely used in their respective laboratories.The participant laboratories were requested to report values with their corresponding measurement uncertainty, for all analytes.In order to ensure that the results obtained from this activity are valid and useful for subsequent analysis, participants were requested to adhere the instructions which contained the information related to proficiency testing such as sample, storage and handling, analysis, test method, reporting, laboratory code number, and submission of documents.

Analytical method selection
The preparation sample method of tonic food drink sample chosen by Research Centre for Chemistry-LIPI laboratories was dry digestion according to AOAC Official Method 999.11.Subsequently, the calcium content in sample was determined by using ion chromatography based on the proposed method for analysis ion potassium, sodium, calcium and magnesium by Basta, N.T, and Tabatabai, M.A., (1985).These analytical methods are used by Research Centre for Chemistry-LIPI laboratories for the routine analysis of calcium content in water and food samples.

EXPERIMENTAL 3.1 Reagents
All chemicals were purchased from MERCK and for analysis grade.NANOpure deionized water (17.8MΩcm, Barnstead) was used for all solution preparation.
The dilute of HNO 3 solution was used in the preparation of powdered tonic food drink sample such as 0.5 M of solution.Standard of Ca of 1000 mg/L as stock solution was prepared by the dilution of 1.350 g CaCl 2 .2H 2 O.The test sample is powdered tonic food drink which prepared and distributed by FNRI-DOST.Solution of 1.8 mM methyl sulfonic acid was used as mobile phase for elution the component from the column of ion chromatograph.The flow rate of mobile phase was set to 1 mL/min as condition optimum of separation in ion chromatography.

Instrumental
Hotplate Cymarec 2 and muffle furnace Sibata were used in the preparation of powdered tonic food drink sample by means of dry digestion.Dionex Ion Chromatography equipped by cation exchange column of Dionex CS12 A (4x250 mm) and conductivity detector was utilized to determine the concentration of Ca contained in sample.

Sample Preparation
The powdered tonic food drink sample was prepared by dry digestion method according to AOAC Official Method 999.11.
Drying step.The 5 g of sample was weighed in the crucible and heated on a hot plate at temperature 100° C until the sample blacken was formed.
Ashing step.The crucible was put in the furnace at initial temperature not higher than 100° C and the temperature of furnace was slowly raised to 450° C at rate no more than 50° C/hour.The sample was ashed for at least 8 hour or overnight.Then, the crucible was taken out from the furnace and let it cool.The ash was wetted with 1-3 mL water and evaporated on hotplate.The crucible was put back in furnace with temperature set up the same as the first step of ashing above.The procedure was repeated until the sample was completely ashed, ash should be white/grey or slightly colored.
Dissolving step.The ash was dissolved in 5 mL of 0.5 M HNO 3 .Then, the residue was filtered.Finally, the solution was transferred into volumetry flask and diluted by the aqua demineralized until 50 mL.

RESULTS AND DISCUSSION 4.1 Specification of Measurand
The specification of measurand, first step of estimation uncertainty processes, is carried out to describe the measurement procedure.This specification was conducted by a comprehensive description of the several stages of the analytical method and by providing the equation of measurand.The several stages to determine of Ca content in powdered tonic food drink using ion chromatography was shown in the following flowchart Figure 1.According to the manufacturer, the volumetric flask has been calibrated at temperature of 20 °C, whereas the laboratory temperature varies between the limits of ±3°C when dissolving process.The uncertainty due to this temperature effect can be determined from the estimation of temperature difference and the coeficient of the volume expansion.The volume expansion of the liquid (2.1×10 -4 °C-1 at 20°C, for water) is considerably greater than that flask (10×10 -6 °C-1 for borosilicate glass flask), so only the former needs to be considered.The standard uncertainty due to the temperature effect was calculated using the assumption of rectangular distribution for the temperature variation, in equation ( 5).(5) The two contributions of uncertainty, calibration and effect temperature, were combined to give the standard uncertainty µ(V) of the volume (V), in equation ( 6).

Dilution Factor
Due to high the concentration of sample, the sample solution was diluted 50 times with aqua demineralized in 50 mL volumetric flask to obtain the peak area that can be compared against the standard.Contribution of uncertainty due to the variation within specification limits and temperature effect were determined and combined for each type of glassware (1 mL of pippete and 50 mL of volumetry flask) Table 1. is a summary of the calculation of uncertainties arising from the variation within specification limits and temperature effect.There was an uncertainty associated with the initial and final volumes taken, so the dilution factor was associated with them.The uncertainty from the dilution factors, µ(F d ), were calculated as:

Concentration of Ca standard
Preparation of standard Ca of 1000 mg/L as stock solution involved several stages of procedure such as weighing and dissolving of standard CaCl 2 .2H 2 O, as described in Figure 1.Subsequently, the concentration of standard Ca was calculated by the following formula.
In addition, the dilution of standard Ca of 1000 mg/L solution was conducted for injection purposes in ion chromatograph.Hence, the uncertainty due to concentration of Ca standard depended on the estimation of standard uncertainty of weighing, dissolving, molecular weight, and purity of standard CaCl 2 .2H 2 The same method in previous section was used for calculation of standard uncertainty of weighing, dissolving and dilution of standard Ca.Furthermore, the uncertainty of purity of CaCl O as well as the dilution of standard Ca solution.
O quoted in the bottle is 99.5± 0.5%.There is no further information from the supplier related the uncertainty of purity in the bottle.Consequently, this uncertainty, 0.005, was taken as having a rectangular distribution and calculated as:  2.  From the Table 3, the standard uncertainty of molecular weight, µ(MW), was calculated by the square root of the sum of the squares of the uncertainty contribution from each element, as follows: So the uncertainty due to concentration of Ca standard was obtained by combining the standard uncertainty of weighing, dissolving, molecular weight, and purity of standard CaCl 2 .2H 2 O as well as the dilution of standard Ca solution, as summarized in the Table 4.The standard uncertainty of peak area of sample and standard were simply determined by dividing their standard deviation of peak area with the square root of the amount of each replication, as calculated in Table 6.

Linearity
Since the quantification of concentration Ca in sample using one point calibration, the linearity of the ion chromatograph response in certain range concentration between standard and sample was determined.The non linearity response would contribute to the accuracy of result.A series of standard of Ca was prepared and analyzed by ion chromatograph to obtain the peak response from several concentration of standard, as summarized in Table 7.The highest difference between theoretical and observed value of Ca standards represent the deviation caused by the linearity of measurement.Table 7 shows that the maximum deviation is 0.258.The linearity contribution is assumed to show a rectangular distribution and is converted to a standard uncertainty by calculation in equation 11.

Repeatability
In order to investigate the standard uncertainty arising from random effects, the repeatability experiments was conducted and the standard deviation of the measured value was quantified.This standard deviation represents the uncertainty due to the variability observed of measurement result within a laboratory, over a short time, using a single operator, item of equipment etc.The result showed a repeatability of the determination of Ca by ion chromatography of 3.662% (as % rsd), as shown in Table 8.

Combined Standard Uncertainty
The value of parameters for calculation of Ca concentration in sample, equation (1), their standard uncertainties and their relative standard uncertainties were summarized in Table 9.  13), the combined standard uncertainties were used as follows  The combined standard uncertainty given in equation 14 is not including the contribution of linearity into account.Hence, the following calculation was carried out to consider the uncertainty due to the linearity.Figure 4 shows the evaluation of results of proficiency testing by involving the uncertainty of measurement from each laboratory.Research Centre for Chemistry LIPI, coded as laboratory 38, demonstrated that the result of measurement was within the range of consensus value ± its expanded uncertainty (X±2µ X : 526.1±28.6)which is 497.5-554.7 mg/100g.Therefore, the result of laboratory 38 was categorized as "satisfactory" or in-lier result in this proficiency testing.
Moreover, the expanded uncertainty of laboratory 38 was large and overlapped within the area of X±2µ X .It indicated that overall of the sources of uncertainty especially the major contribution of uncertainty were taken into account to the measurement by the laboratory.Meanwhile, some of laboratories reported the results with the small uncertainty which no overlap with X±2µ X such as laboratory 14 and 24, as shown in Figure 4.It could be happened when the laboratories was unable to identify all of the sources of uncertainty especially the uncertainty that majorly contribute to the measurement.

Figure 1
Figure 1 Determination of Ca Content in Powdered Tonic Food Drink Using Ion Chromatography m of Uncertainty All of the possible uncertainty sources were identified an analyzed in determination of Ca content in powdered tonic food drink using ion chromatography.In order to describe the influence of uncertainty sources from each parameter to the value of the measurand, the cause-effect diagram or Ishikawa fishbone diagram was created, as shown in Figure2.The uncertainty sources of determination Ca by ion chromatography comprises the uncertainties of peak area of Ca in sample and standard obtained from ion chromatograph, the uncertainties of volume of sample, the uncertainties of mass of sample, the uncertainties of dilution, the uncertainties of concentration of Ca standard, the uncertainties of linearity, and the uncertainties of repeatability.

Figure 2
Figure2The Sources of Uncertainty in The Determination of Fe Content in Powdered Tonic Food Drink Using Graphite Furnace Atomic Absorption Spectrometry by combining the uncertainty in the atomic weights of its constituent elements.The information of atomic weights including uncertainty estimates was obtained from the latest International Union of Pure and Applied Chemistry (IUPAC) Table.The standard uncertainty for each elements are generated by treating the IUPAC quoted uncertainty as rectangular distribution.Thus, the corresponding uncertainty is obtained by dividing those values by .The atomic weights and listed uncertainties for the constituent elements of CaCl 2 .2H 2 O are shown in Table In chromatography, the quantification of concentration sample was carried out by comparison between peak area of sample (A ) X ) and standard (A ST obtained by multiplying the combined standard uncertainty by a coverage factor 2 (at confidence level 95%), in equation (16).
of different parameters were shown in Figure 3.The largest contribution of uncertainty arises from the repeatability of method measurement of Ca by ion chromatography.

Figure 3
Figure 3 Uncertainty Contributions in The Determination of Ca Content in Powdered Tonic Food Drink Sample 4.12 Proficiency testing resultFigure4shows the evaluation of results of proficiency testing by involving the uncertainty of measurement from each laboratory.Research Centre for Chemistry LIPI, coded as laboratory 38, demonstrated that the result of measurement was within the range of consensus value ± its expanded uncertainty (X±2µ X : 526.1±28.6)which is 497.5-554.7 mg/100g.Therefore, the result of laboratory 38 was categorized as "satisfactory" or in-lier result in this proficiency testing.Moreover, the expanded uncertainty of laboratory 38 was large and overlapped within

Table 1
Uncertainties Due To The Manufacture's Spesification and Effect Temperature

Table 2
The Atomic Weights and Listed Uncertainties for The Constituent Elements of Cacl 2 .2H 2 Table 2 by the number of atoms.The table 3 listed the separate element contribution to the molecular weight, together with the uncertainty contribution for each.

Table 3
The Separate Element Contribution to The Molecular Weight and Its Uncertainty Contribution

Table 4
Uncertainties Due To The Concentration of Ca Standard

Table 5
Peak Area of Sample and Standard Obtained from Chromatogram Ion Chromatography Instrument).The peak area of sample and Ca standard obtained from chromatogram of the ion chromatography instrument was shown in the Table5.

Table 6
Standard Uncertainty of Peak Area From Sample and Standard

Table 7
Linearity Verification of The Ion Chromatograph Using Standard of Ca *maximum deviation (μ)

Table 8
Repeatability of Ca Measurement in Sample By Ion Chromatography

Table 9
Uncertainties in The Determination of Ca in Powdered Tonic Food Drink by Ion Chromatography