3.       Introduction 

High-Performance Liquid Chromatography (HPLC) is a technique in analytical chemistry [1], which is used to identify, and quantify each component in a mixture by separating them. HPLC working principle means the high pressure pumping (about at 200 bar) of a liquid solvent containing the sample mixture through a column filled with a solid adsorbent material. Each component in the sample interacts slightly differently with the adsorbent material, causing different flow rates for the different components and leading to the separation of the components as they flow out the column. An example of instrument is shown on Figure 1.

HPLC has multiple uses in manufacturing and research fields. Among them, some examples are production process control of pharmaceutical products, analysis of food samples, or the analysis of the composition of a new artificial resin under research. For the particular tests, the HPLC instrument (chromatograph) works as a complex hardware-software system, as shown further.

A chromatograph consists of the following main modules (Figure 2) a pump [2] delivers the mobile phase (so-called eluent) through a high-pressure tube system to the further modules. The next module is the injection port, which is used for the injection of the sample solution into the eluent stream by a syringe and solenoid valve. Usually, syringe and injection port is integrated in a unit called autosampler, which can automatically inject samples sequentially as given in the test program. Then follows the chromatographic column, the module responsible of the separation. On this column the components of elution are separated by adsorption to the column. After the column follows the detector, which measures the current presence of a component in the trespassing eluent.

The signal of detector is recorded by a computer unit (shown as Display on Figure 2) as absorbance-time points. The time points are saved into a raw format specific to the instrument manufacturer’ software. Opening the raw format by the software plots the so-called chromatogram. The x axis of the chromatogram shows the run time and the y axis the detector signal. The recorded data appear as a continuous plot containing gaussian peaks. The time of the peaks (retention time) identifies the given component, and the peak area gives quantitative information about the component. Prior identification and quantitation, a chromatogram with standard material(s) is recorded for qualitative and quantitative calibration of the instrument. At the quantitation there are some peak integration parameters (for standard and sample too), which have to be varied casually for accurate quantitative results. This is for example the starting point and end point of the integration.

The aim of this paper is the elimination of the need of manufacturer specific software for the measurement data processing. The raw files, in which the chromatograms are stored, can be opened and processed by the instrument manufacturer’ costly software with its own algorithm. Thus, the access to the computer in the given laboratory is needed. However, in research, there may be a need to evaluate the results away of the laboratory. Having the guidelines of an accurate data processing [4,5], chromatogram raw files could be “taken away” and processed with own algorithm. There is an open-source application [6] to open and process chromatogram data files (so-called raw files) of the chromatogram obtained, however with strict limitations, e.g. processing Diode Array Detector, DAD is in progress yet. Thus, printing the given chromatograms and reprocessing them from the scanned “hard copy” is a useful way for out of the laboratory chromatographic data processing.

4.       Experiment 

4.1.  Result Set to Reprocess 

A common method for HPLC determination of vitamins contains a sample preparation by saponification [7], also described by ISO standard [8]. Saponification was optimized with a 3k Statistical Experimental Design (SED), and it was found that the optimal temperature and time of saponification are 82^°C and 18 minutes respectively [9]. At these conditions, five replicate tests were performed. Diode Array Detector (DAD) was used at 325 nm. Retention time of Vitamin A was of 4 minutes.

Five replicate test runs at the optimal saponification temperature and time settings were selected (further original result set), and their chromatogram was scanned as reprocessed result set. The criteria of successful reprocessing were the following: 

·         Original result set and reprocessed result set are the same, as compared by homoscedastic t test,

·         Homoscedasticity was determined by F test of the original result set and reprocessed result set. 

The same calibration was used for the reprocessing. 

4.2.  Scanning of Chromatograms and Image Processing

Chromatogram papers were scanned by a HP 3-in-1 printer and scanner at 300 DPI. Images were saved into JPEG format with intermediate compression and quality. The images were open with Get data Graph Digitizer [10]. Scale was set as follows: first the zero point was assigned on both axis, then one point was set on x and y axis at 14 minutes and 80 mA respectively. Then graph data points were acquired with d=0.01 and these data points were transferred to Excel.

In the Excel Table, peak area was calculated by the trapeze integration method.

4.3.  Integration Method

Trapeze integration method was used. Formula is shown by the following formula:

5.                   Results and Discussion 

5.1.  Data Points and Re-Plotted Chromatograms

Chromatograms replotted and reprocessed by [8] are visualized on Figure 2.

Some data points are shown by Table 1. 

Peak on Figure 3 was integrated by trapeze integration method, and vitamin A concentration was calculated by the existing calibration data. Their comparison is shown in Table 3.

Match between result set A and B was determined by t test (Table 3). Homoscedasticity of the result sets was identified by F test. Thus a two-sample, homoscedastic t-test was run on the two data series. If it is superior than t_crit=0.05, then there is match between result set A and B. As t is found much higher than t_crit, it can be concluded that the reintegration method gives the same test results as the original instrument integration method. 

6.       Conclusion 

Out-of-laboratory integration of chromatograms may be necessary for HPLC test result reprocessing in the research field. For this, an alternative process method of chromatogram printouts was developed. It was found that the scanned and reprocessed chromatogram results are in-line with the original test results. 

7.       Acknowledgement 

Test results were provided by Bálint Analitika Ltd. (Budapest, Hungary).

Figure 1: High pressure liquid chromatograph [2].

Figure 2: Schema of a high pressure liquid chromatograph [3].

Figure 2: Example of chromatogram printout.

Figure 3: Getting chromatogram data points.

Figure 4: Reprocessed chromatogram peak.

Table 1: Example of data points.

Table 2: Results of vitamin A content of margarine samples calculated by A) instrument software method and by B) integration of scanned chromatogram.

Table 3: Results of t-test and f-test between A and B.

1.       Katz E, Eksteen R, Schoenmakers P, Miller N (1998) Handbook of HPLC. Pg No: 1-3.

2.       https://gmi-inc.com/agilent-1100-hplc-system-vwd-with-quaternary-pump.html.

3.       High Performance Liquid Chromatography: Module 4-Types of Stationary Phases, lab-training.com [http://lab-training.com/landing/free-hplc-training-programme-6/].

4.       https://www.slideshare.net/deepakmishra56679015/integration-of-chromatographic-peaks.

5.       http://www.chromatographyonline.com/questions-quality-where-can-i-draw-line.

6.       https://www.openchrom.net/.

7.       Nollet LML (1996) Handbook of Food Analysis. 1: 611-613.

8.       MSZ-EN 12823-1:2014 (2014) Foodstuffs - Determination of vitamin A by high performance liquid chromatography - Part 1: Measurement of all-trans-retinol and 13-cis-retinol.

9.       Garai L (2017) Improving HPLC Analysis of Vitamin A and E: Use of Statistical Experimental Design, Procedia Computer Science. Pg No: 1500-1511.