The first publications on the evaluation of thin-layer chromatograms by image pro- cessing appeared as early as the mid-1980s [118]. These methods, also known as “digi- tal evaluation”, were presented to the scientific world in the lectures of Proˇsek, who developed his own software for the purpose in 1991 [119], although only the enormous developments in the hardware (video cameras, computers, printers etc.) have made video evaluation of thin-layer chromatograms economically justifiable. Proˇsek pre- sented his paper on the validation of quantitative TLC by video camera in 1997 [120].
What is “digital photography”?
Digital images consist of information about the intensity and color of light at thou- sands of separate points on the image. These data consist of numbers, encoded as a se- ries of zeros and ones, that are transferred in the form of electrical impulses to or from the microprocessor of a computer and are then translated back into an image on the computer screen.
No digital image is as good as the human eye’s view of an actual object. The smaller the number of stored points of the image (known in computer terminology as pixels) the poorer is its quality.
The second criterion of quality is the depth of color of a digital photograph. This de- pends on the number of colors encoded: for 256 colors, eight digital characters (bits) per pixel are necessary. A color intensity of 16 bits enables 65 536 colors to be repre- sented, which is known as “high color”. An almost photographically realistic represen- tation (“true color”) is only possible with 16.7 million colors, and this requires 24 bits per pixel [121].
Modular systems of the latest type for the production of images and documentation of thin-layer chromatograms consist of four components.
Illumination systemwith the following kinds of light:
– 254-nm short-wave UV incident light – 366-nm long-wave UV incident light
– 400-nm to 750-nm visible incident light and/or transmitted light
The tubular lamps are driven by 25-kHz to 30-kHz high-frequency voltage. This not only enables the light efficiency to be optimized, but also eliminates synchronization problems with the CCD camera. Furthermore, a light-proof hood and camera bellows can be used to exclude all extraneous light from the system, enabling photographs to be taken even in a room that is not completely darkened.
174 7 Evaluation After Derivatization
Camera system,consisting of:
– 1-chip or, better, 3-chip video camera1
1 Digital cameras do not contain a film. Instead, a so-called CCD (charge-coupled device) chip takes the picture. This chip consists of separate sensors with red, green and blue filters so that they are sensitive to various colors. The image is not stored by the chip but in a separate mem- ory. The image is then immediately available to be printed out or displayed on a computer screen.
with zoom objective
– macro-objective with converter to give a full-format image of a 5 × 5 cm TLC plate
– supplementary lenses (various focal lengths) – UV filter combination
– adjustable camera stand (e.g. tripod)
Data processing equipment.In order to run the software, the minimum operating system is a PC with a Pentium II processor and a 64-MB memory and Windows 95/
98/NTTM.
Theimage output systemshould consist of:
– video printer with 256 shades of gray, or
– color printer, i.e. Windows-driven laser or ink-jet printer. The images which this produces on special paper (720 dpi) have a more limited gray range compared with video prints, but do not require heat-sensitive paper and can be stored al- most indefinitely
The items listed under the above four headings can currently be purchased individu- ally without difficulty from specialist suppliers, but it is becoming somewhat more dif- ficult to obtain functional software. We would therefore recommend the purchase of a complete package from a company such as CAMAG (VideoStore) or DESAGA (Densitometer) in which the software has been developed in accordance with GMP guidelines and excludes the possibility of tampering with images. Professional image processing systems automatically “flag up” any images that have been altered. The processed images are referred to as “derivatives” (the same word as that used to de- scribe the results of chemical reactions) [123].
7.3.1 Qualitative Video Evaluation
Many of the illustrations in this book were produced by a video system. Not only can qualitative results be obtained using such systems, but semiquantitative evaluations are also possible from a good video shot of the chromatogram (see Fig. 72 in Section 5.1.2). Practical tips for the preparation of video shots are closely linked to the docu- mentation and given described in detail in Section 8.4.
175 7.3 Evaluation Using a Video System
7.3.2 Quantitative Video Evaluation
Both of the manufacturers of video systems mentioned in Section 7.3 are putting a great deal of effort into the development of their own software for the quantitative evaluation of thin-layer chromatograms. The first examples of this type of product were presented at conferences in early 1997. In a paper by Maria Müller, the individ- ual zones of the chromatogram in Fig. 72 were quantified using the DESAGA system.
This was performed in two series of measurements, as the individual zones showed concentration differences that were too great. The results obtained from this quantita- tive video evaluation are given in Table 21, enabling a visual evaluation to be per- formed. It can be seen that the visual evaluation of zone 1 (known impurity) is in good agreement with the result of the video evaluation. This should correspond with the cal- ibration of this substance on the plate (lanes 1–3). The three unknown substances (Z 2, 3 and 5) were calculated with the aid of the measured values of the limit value con- centrations on lanes 4 and 5. The discrepancies between the visual and the video evalu- ations are sometimes appreciable here.
Table 21: Comparison of semiquantitative visual and quantitative video evaluation of Fig. 72 No. of the zone in Visual evaluation Video evaluation
Fig. 72 Lane 6 Lane 7 Lane 6 Lane 7
declaration in % of 400 g
Z 1 = known related compound lane 7 0,5 0,38 0,44
Z 2 = unknown compound lane 7 0,5 0,12 0,09
Z 3 = unknown compound trace 0,5 – 0,20
Z 4 Zone of the active ingredient
Z 5 = unknown compound 0,5 0,5 0,63a) 0,23
Secondary spots total 1,25 1,75 1,13 0,96
Unknown compounds total 0,75 1,25 0,75 0,52
a)Limit test acc. to DAB exceeded
Visual evaluation is described in Section 5.1.2 “Detection with 254 nm UV Light”.
Comments on video evaluation can be found in Section 7.3.2.
The book accompanying InCom 1996 includes an example of the “Video- Densitometric Evaluation of Thin-Layer Chromatograms” [124]. Thomas Mall con- cluded that, compared with the usual TLC evaluation methods, the video-densi- tometric method was
1. quicker,
2. linear over a wider range, and
3. represented an ideal combination of evaluation with documentation.
The evaluation software used was the BASys 1D-PAGE program available from BIOTEC-FISCHER, Reiskirchen (Germany). This program, which was developed for gel electrophoresis, was adapted to the requirements of TLC. The software gives the possibility of performing a so-called “background correction” by subtracting the curve 176 7 Evaluation After Derivatization
that represents the background, thus compensating for optical effects such as nonuni- form illumination of the TLC plate. Figure 97a gives an example of the calibration curve of the substance BM 96.0102 over the range 0.1–1.0 gwithoutbackground cor- rection, and the same TLC platewithbackground correction is shown for comparison in Fig. 97b. These documents were made available by Thomas Mall.
7.3.3 Comparison of the TLC Scanner With Video Evaluation
A comparison between quantitative evaluation by TLC scanner and by a video system was made by the author. For this, the foil carrying the coffee chromatograms (Fig. 12) was scanned with the CD60, and a video shot, obtained with the VD 40, was evaluated with the ProResult software (all from DESAGA). The results show good agreement between similar samples, but there were differences when the matrices were different.
This environmentally friendly TLC was developed especially for teaching in schools (sorbent, TLC aluminum foil RP-18 F254s, solvent system: methanol + water, 60 + 40 v/
v). The results are given in Table 22, and a detailed description of the experiments can be found in [124a].
Table 22: Assay of caffeine [%] in various samples of coffee and tea
Video system TLC scanner
Tchibo “best beans” 1,18 1,08
Jacobs “light” 0,68 0,71
Jacobs “free” 0,06 0,03
Nestl ´e “our best” 2,02 1,50
Darjeeling tea 1,68 1,60
Comparison of determinations using a video system and a TLC scanner
Comparative measurements have also been made on a phytopharmaceutical prod- uct. In the investigation of ginseng roots and of the medicaments prepared from this, the ginsenoside Rg1was determined by both 1-dimensional and 2-dimensional devel- opment. The results were presented in a poster in Vienna in September 1998 [124b].
In its house journal (CBS), CAMAG has published many comparisons between densitometric and video evaluation [124c], and has given a clear account of the current scope and limitations of the two systems:
Classical densitometry utilizes the whole spectral range between 190 nm and 800 nm with high spectral selectivity. Absorption spectra can be produced covering the whole range.
The video technology in general use today operates exclusively in the visible range.
Therefore, substances can only be detected if they are either colored or can be con- verted into colored derivatives or if they fluoresce in the visible range on excitation by long-wave UV light (365 nm). Indirect detection is also possible on plates with 177 7.3 Evaluation Using a Video System
fluorescence indicators by excitation by UV light (254 nm or 365 nm). It is not possi- ble to produce spectra using video technology.
The comparisons of the various methods of evaluation made in this Chapter are valid up to mid-1999, but the very rapid developments in the multimedia market will lead to further improvements in the future. A summarized comparison of quantitative evalua- tion methods for TLC up to the present time can be found in [124d].