Validation of High-Performance Liquid Chromatography Method for Determination of Vitamin Bi in Powder Milk

Received:8th December 2019 Revised: 6th April 2020 Accepted: 20th April 2020 Online: 31st May 2020 Vitamin Bi plays an important role in the co-enzymatic reactions for energy-rich compounds called ATP (Adenosine Tri Phosphate). Therefore, it should be added to various food products, for example, milk powder. One method that can be used to determine vitamin Bi is SNI number 3751: 2009, but the method is intended for wheat flour. If the method is to be used for the analysis from other samples, such as milk powder , optimization, and validation, are needed. This experiment was carried out using HPLC, C18 column, and UV detector with a wavelength of 254 nm. The mobile phase used is methanol: acetic acid: bi-distilled water = 32:1:67 (v/v/v), flow rate = 1 mL/minute, isocratic, and reverse phased technique. Method validation parameters include tests of system suitability, linearity, the limit of detection , the limit of quantitation, precision (repeatability), and accuracy. The results showed that the system suitability test wasobtained relative standard deviations (% RSD) for retention time and peak area, tailing factor, resolution, separation factor was 0.297%, 1.476%, 1.113, 6.693, and 4.406 respectively. The validation test gets a correlation coefficient (R) of 0.9996, the limit of detection and limit of quantitation were 0.0122 mg/100 mL and 0.0244 mg/100 mL, respectively. The precision test obtained Horwitz 's ratio of 0.27%. Accuracy test using CRM obtained % recovery of 93.79-97.77%. All these results meet the requirements of method validation, so it can be concluded that the method of SNI number 3751: 2009 is valid for the determination of vitamin Bi in milk powder and can be used for routine analysis procedure.


Introduction
Milk powder is one product that is produced from processed fresh milk. This milk is in the form of powder with minimal moisture content. According to the BSN ( 2015 ) [1], powder milk is a dairy product obtained by reducing a large portion of water through the drying process of fresh milk with or without the addition of vitamins, minerals, and other food additives.
Vitamins are one of the micronutrients, which are usually added to milk powder. Vitamins are complex organic compounds that are needed in large quantities for the metabolic functions of the human body [2]. Vitamins can be grouped into two main groups, i.e., fat-soluble vitamins which include vitamins A, D, E, and K and water-because wheat flour and milk powder have the same characteristics. Both of them are powder and food products. According to Eurachem [8], a method should be validated when non-standard methods, laboratorydesigned / developed methods, standard methods used outside their intended scope, and amplifications and modifications of standard methods. SNI method number 3751 : 2009 is for wheat flour, so it is needed to be validated when we want to apply to milk powder as wheat flour and milk powder has different matrices.
The purpose of this experiment is to validate the testing method of vitamin Bi content milk powder and ensure that the analytical method is accurate, specific, and reliable so that this method can be used as a routine method for analyzing vitamin Bi in milk powder samples.

Material and Methods
This experiment consists of three stages, namely the preparation, testing, and processing of data. The preparation stage includes the preparation of reagents, mobile phase solutions, standard series, sample preparation, and conditioning of high-performance liquid chromatography (HPLC). The testing stages include a test of system suitability, linearity test , the limit of detection and limit of quantitation (LOD & LOQ), precision (repeatability) , and accuracy [ 7 ]. The data processing stage is done by calculating the results of the validation data and estimating the uncertainty of the test results and then comparing the results of the validation method with terms of acceptance. The method is valid and can be used as a routine analysis method if all method validation parameters meet the acceptability limit [8,9 ].

Preparation of a standard solution of Vitamin Bi
The standard vitamin Bi was weighed 0.0100 g into a 50 mL beaker glass, then added 50.00 mL of 0.1 N hydrochloric acids quantitatively. This standard solution is stirred with a magnetic stirrer , then heated in a water bath with a temperature of 95 -ioo°C for 30 minutes. This standard solution is cooled at room temperature ( ± 25°C ) then the pH of the solution is adjusted to pH 4.5 with sodium acetate 2M solution. This standard solution is put into a 100 mL volumetric flask and diluted with double distilled water to the mark, then homogenized and filtered with a filter membrane of 0.45 pm. A standard solution of 10 mg / 100 mL of vitamin Bi was taken as much as 5.00 mL and then put into a 50 mL volumetric flask to make a standard solution of 1 mg / 100 mL. The standard solution of 1 mg / 100 mL of vitamin Bi was taken ( 0.5; 1; 2; 6; 12) mL then put into 50 mL volumetric flask, respectively to make concentration (0.01; 0.02; 0.04 ; 0.12; 0.24 ) mg / 100 mL. This standard solution is diluted with double distilled water and then homogenized.

. Preparation of milk powder samples
Milk powder weighed 5.5000 g into a 50 mL beaker glass and added 0.1 N hydrochloric acid as much as 50.00 mL quantitatively. The sample solution is stirred with a magnetic stirrer then heated with a water bath at a temperature ( 95 -ioo)°C for 30 minutes. The sample solution is cooled at room temperature ( ± 25°C ) then the pH of the solution is adjusted to reach pH 4.5 with a 2 M sodium acetate solution. The sample solution is put into a 100 mL volumetric flask and diluted with double distilled water to the mark and then homogenized. The sample solution is filtered with Whatman filter number 42 and then filtered again with a filter membrane 0.45 pm.

Preparation of mobile phase solution
The mobile phase solution consists of methanol: acetic acid: water ( 32: 1: 67 , v / v / v) prepared as much as 500 mL. 5 mL of acetic acid and 335 mL of double-distilled water (1: 67 ) were put into a 500 mL beaker glass and then added 0.4718 g hexane-isulfonic acid sodium salt. The solution is stirred until dissolved then filtered with filter membrane 0.45 pm. The acetic acid solution and the water are treated for 20 minutes then put into the mobile phase container. Methanol is prepared and put into other mobile phase containers. The mobile phase solution is mixed in the HPLC instrument with a ratio of methanol and a mixture of acetic acidwater = 32:68 (v / v)

Validation parameters
Validation of the analytical method deals to ensure that the developed method is acceptable and reliable for its deliberate purpose. As per ICH guidelines, the developed HPLC method was validated for different validation characteristics such as system suitability, selectivity, linearity, precision, accuracy, LOD, and LOQ.

1. System suitability
The system suitability ensures the validity and specificity of the developed method. System suitability test was performed by measuring chromatographic parameters such as peak area, retention time, theoretical plates as well as tailing factor and by determining the standard deviation ( RSD) for each parameter. The percent RSD of peak area ( < 2), tailing factor (< 2) , and theoretical plates ( N > 2000) were within the acceptable limits [  The specificity of a method is the ability of a method to measure certain substances, carefully and thoroughly, in the presence of other components that might be in the sample matrix. For the selectivity test, the substance to be tested must be determined first. Then made a standard solution , test solution and the blank solution is then tested by the method , whether the method can separate the analyte from the sample matrix

Linearity
Linearity is the ability of the analytically-developed procedure to check the result of the sample analytes, whether it is linear to its concentration. The linearity of the method was tested by analyzing different concentrations range of the analytes. The regression analysis data for the calibration curve shows a linear relation over this concentration range for analytes. The correlation coefficient value was found to be R > 0.995 [ 9 , 12], which suggested excellent correlation and good linearity for the optimized method.

Precision (Repeatability)
The precision was carried out by analyzing samples of 7 replications using HPLC. From the seven replications, the value of % RSD was calculated and compared with % RSD (Horwitz ), so the Horwitz ratio was obtained. Horwitz ratio can be accepted if the value < 2.

Accuracy (%Recovery)
The accuracy test is done by analyzing the Certified Reference Material

. System suitability test
Before conducting the analysis, we need to do a system suitability test to ensure that the HPLC is functioning properly. The standard solutions of vitamin Bi were measured to know the retention time of the thiamin. After we know the retention time of the thiamin, the sample was injected eight times to determine the repeatability of the area and retention time. From the chromatogram in Figure 1, we can get the system suitability data as follows: The curve shows the relationship between concentration and peak area. The line connecting the points on the curve above is called the regression line or correlation line. The parameter that states the relationship between two variables (on the x-axis and yaxis) is called the correlation coefficient. The magnitude of the correlation coefficient (r) on the calibration curve above is 0.9996 . This value meets the requirements of r > 0.9950 [ 9 , 12]. The slope value on the calibration curve can be used to see sensitivity. The slope value on the curve above is 398871.78 indicating high sensitivity. Intercept is the intersection point at y when x is 0. Ideally, the intercept is 0, but the intercept on the curve above is negative. When y is 0, and the intercept is interpolated to the regression equation , an x value of 0.002878 mg / 100 mL is obtained.

. Precision (repeatability)
Precision is useful to see the repeatability of the analytical method carried out under the same conditions of the tool, place, and analyst, in the next test period.
Measurements were made with seven replications. The results of the vitamin Bi precision test can be seen in

. Specificity
Based on the chromatogram in Figure 1, it can be concluded that a retention time of 8.624 is specific for thiamin. It means that this method has good selectivity because the retention time of analyte is far from the other peak ( R = 6.693 ), and the retention time of the analyte is not disturbed by the sample matrix.

. Linearity
Linearity tests were carried out by measuring the peak area of a standard series of vitamin Bi at concentrations (0.0100; 0.0200; 0.0400 ; 0.1200; 0.2400 ) mg / 100 mL. The standard series concentrations are made in the range ( 50 -150 )% of the estimated or expected concentration of the analyte in the test sample. Standard series solutions were prepared at five different concentrations with solution preparation three times repetition. The chromatogram and calibration curve for standard vitamin Bi can be seen in Figure 2.

Content of Vitamin Bi
(mg / 100 g)

Weight of sample (g)
Requirement [11] No. Area Based on precision results, the Horwitz ratio of the sample is 0.27 , it is still less than 2, so the precision is still within the acceptable limit

Accuracy
Accuracy testing is performed to determine the closeness between the measured value and the accepted real value, which is expressed as percent recovery.
Accuracy testing is carried out using Certified Reference Material, whose purity is known. CRM is treated according to the test sample preparation method. The accuracy testing results can be seen in Table 4 .

. Estimation of uncertainty [ 14 , 15 ]
The first step in determining the uncertainty is the identification and analysis of sources of uncertainty through a Fishbone diagram. The Fishbone diagram of the estimation of the uncertainty of determining vitamin Bi content in powder milk can be seen in Figure 3 .

. Limit of detection (LOD) and limit of quantitation (LOQ)
The theoretical LOD and LOQ are obtained from statistical calculations from the calibration curve data. From the theoretical LOD and LOQ, the next step was confirmed by making and analyzing standards with a concentration of theoretical LOD and LOQ. The theoretical LOD and LOQcalculations can be seen in Table 5.   Table 5 , theoretical LOD and theoretical LOQ were obtained for 0.0116 mg / 100 mL and 0.0385 mg / 100 mL, respectively. From this theoretical LOD, it is necessary to confirm these values by making and analyzing standard solutions at these concentrations, but these concentrations can be raised or lowered until a peak that is well-read (positive response), while the LOQvalue must meet the requirements of precision and accuracy.