Multi-wavelength imaging of HPTLC plates using a constructed illumination chamber with a smartphone camera as the detector
Graphical abstract
Introduction
In recent years, thin-layer chromatography (TLC) has been considerably improved. It offers improved detection limits [1,2]. It can be combined with advanced instrumental techniques such as, e.g. matrix assisted laser desorption ionization time-of-flight mass spectrometry detection [3] or high-resolution mass spectrometry detection [4] to provide comprehensive chromatographic data, which eventually increases the pool of candidate analytes that can be screened during a single chromatographic run. Several samples can be analyzed in parallel in one chromatographic run. A plate can be developed more than once using different chromatographic conditions and/or in two directions [5,6] and/or use orthogonal systems to enhance separation. One can explore the bioactivity potential of the separated constituents of a given mixture, including, for instance, its antibacterial, antifungal, or antioxidant activity [[7], [8], [9], [10]]. All of these features make the TLC technique attractive from the analytical point of view. It can deliver reproducible chromatographic separations and, thus, reliable analytical data sufficient to answer the research questions that arise in different branches of science and technology [3,4,11].
Unfortunately, over time, instrumentation and maintenance have become expensive. Therefore, inexpensive instruments [12], new tools [13], and efficient data handling strategies [14] are always sought after to decrease the cost of a TLC analysis and to extend the areas in which it can be applied.
In TLC, the enhancement of detection is an essential area of further improvement. Among different detection methods, including visual inspection, densitometry plays a crucial role. It assumes scanning the surface of a chromatographic plate along with the developed track. At each measurement point, the intensity of the reflected monochromatic light is recorded. When the densitometric scan is completed, all of the measurements form the so-called densitograms. It describes changes in the optical density and intensity of the signal proportional to the concentration of a mixture component. Routine densitometric detection involves a single scan of the TLC plate. However, for mixtures with differently absorbing compounds, multiple scans (i.e., using different detection wavelengths), are beneficial because they can reveal their presence. A significant advantage of multi-wavelength detection is the possibility to assess peak purity and to collect comprehensive information about the absorbance of the individual chemical components.
On the other hand, multiple densitograms form comprehensive analytical data that are multivariate and complex. Therefore, the relevant information is often hidden in the data and not readily accessible. Its extraction requires chemometric modeling. Despite the advantages of multidimensional measurements, recording the multi-wavelength densitograms requires relatively expensive instrumentation, and each track has to be scanned several times. Many measurements are carried out at each location in order to obtain densitograms that have a high signal-to-noise ratio. Moreover, to describe the mixture comprehensively, as many detection wavelengths as possible should be considered, and the spatial resolution (sampling) must be high. These detection assumptions make the scanning procedure for a single sample time-consuming.
Chromatographic plates are also documented using specially designed digital cameras and flatbed scanners. They enable the rapid scanning of a surface, with the possibility to illuminate plates with different light sources. Nonetheless, their cost is relatively high. Regardless of the available at hand instrument, preprocessing of obtained instrumental signals (densitograms and images), including improvement of signal-to-noise ratio (baseline removal and noise suppression), peak alignment, and normalization, is a crucial step, mainly when many samples are compared. The quality of the instrumental signal depends on the detection settings and the inherent properties of a mixture. Therefore, improving the signal's quality is not always possible by simple modifications of measurement settings, or one can find these efforts demanding, bearing in mind the expected outcome.
In this article, a multi-source illumination chamber equipped with a smartphone camera as the detector is introduced as an inexpensive alternative to TLC documentation systems and densitometers. It combines the imaging potential and allows for performing multi-wavelength scans. Moreover, a data handling strategy that is suitable for processing and analyzing the recorded images of TLC plates (including reducing instrumental signals and their further chemometric modeling) is proposed. The originality of this work primarily stems from the concept of a comprehensive analysis of multiple images expressing separation onto TLC plates, followed by further chemometric modeling of the relevant data. They are collected using UV-A, UV-C and LED illumination sources, and the resulting images are analyzed using three different detection channels of a smartphone camera sensor (three color components: red (R), green (G) and blue (B)). A significant improvement is that a few illumination sources replace the detection wavelengths and the surface image is described by the RGB channels, which, in contrast to the multi-wavelength scan mode, does not provide any direct physical interpretation. Because of the differences in the spectroscopic/absorbance properties (e.g. fluorescence), the RGB triplet of images is captured in a specific UV range and can reveal hidden information not detected using another light source (for instance, in a visible spectral range). A possible synergy effect is achieved when the diverse information is contained in a set of measurements, and its extraction is followed by the appropriate data processing. This facilitates solving the analytical classification/regression challenges that were impossible to tackle using a simple RGB image data. The potential of the new device is demonstrated for exploring the differences among various brands of water-based inks (a forensic application). The results were compared with the results that were obtained using a commercially available densitometer enabling multi-wavelength detection.
Section snippets
Interpolation of the baseline of a signal
Estimating the baseline of a signal is a challenging preprocessing step. The most intuitive approach fits a polynomial of a given degree to manually selected signal points from the baseline. Then, the estimated baseline is subtracted from the instrumental signal. Although the procedure is straightforward, it is time-consuming and subjective.
There are many algorithms for baseline estimation [15,16], which are focused on minimizing a non-quadratic criterion. The one that was used in this study
Chemicals, reagents and samples
Mixtures of different commercially available inks were chromatographed using HPTLC silica gel 60 plates (20 × 20 cm) as the stationary phase aluminum (Merck KgaA, Darmstadt, Germany). The chemical reagents that were used during the experiments: acetone (≥99.8%), n-butanol (≥99.5%) and ethanol (≥99.9%) were also supplied by Merck, KgaA, Darmstadt, Germany. A total of five black and five blue inks were purchased from five different manufacturers. The different inks were coded as follows: 1) Hero
Acquisition of the images
The most important feature of imaging is the possibility to visually inspect the data, which enables the straightforward detection of any spatial artifacts. The measurement procedures that are described in detail in the section ‘Measurement principles and settings’ ensure the correspondence of the pixels in the nine-channel images of a given plate. However, minor translations may arise in the vertical or horizontal directions because of the positioning of the successive plates in the chamber.
Conclusions
Based on the experimental data, it was proved that a low-cost illumination chamber provided measurements with a similar quality compared to the ones that were obtained from a commercially available densitometer. This was possible after the proper processing of the collected data and the elimination of the baseline. For an effective baseline estimation in multi-wavelength signals, symmetric or asymmetric variants of the Huber cost function were used. Their minimization enabled an automatic
Credit author statement
Lukasz Pieszczek, Conceptualization, Methodology, Software, Formal analysis, Investigation, Data curation, Writing - original draft, Writing - review & editing, Visualization, Project administration, Funding acquisition. Michal Daszykowski, Conceptualization, Methodology, Formal analysis, Investigation, Writing - review & editing, Supervision
Declaration of competing interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
Acknowledgments
This research was supported by the National Science Centre, Poland (research grant no. UMO-2018/29/N/ST4/01547).
References (30)
- et al.
Stationary phase thickness determines the quality of thin-layer chromatography/matrix-assisted laser desorption and ionization mass spectra of lipids
Anal. Biochem.
(2014) - et al.
Bioautography detection in thin-layer chromatography
J. Chromatogr., A
(2011) - et al.
Characterization of drug authenticity using thin-layer chromatography imaging with a mobile phone
J. Pharmaceut. Biomed. Anal.
(2016) - et al.
The start-to-end chemometric image processing of 2D thin-layer videoscans
J. Chromatogr., A
(2011) - et al.
Asymmetric least squares for multiple spectra baseline correction
Anal. Chim. Acta
(2010) - et al.
Background removal from spectra by designing and minimising a non-quadratic cost function
Chemometr. Intell. Lab. Syst.
(2005) - et al.
Target selection for alignment of chromatographic signals obtained using monochannel detectors
J. Chromatogr., A
(2007) - et al.
A comparison of three algorithms for chromatograms alignment
J. Chromatogr., A
(2006) - et al.
Automated alignment of one-dimensional chromatographic fingerprints
J. Chromatogr., A
(2010) - et al.
Aligning of single and multiple wavelength chromatographic profiles for chemometric data analysis using correlation optimised warping
J. Chromatogr., A
(1998)