Food Chemistry 132 (2012) 549–553

Contents lists available at SciVerse ScienceDirect

Food Chemistry
journal homepage: www.elsevier.com/locate/foodchem

Analytical Methods

Investigation of antiradical activity of plant material by thin-layer

chromatography with image processing
Marta Olech a, Łukasz Komsta b, Renata Nowak a,⇑, Łukasz Cieśla c,d, Monika Waksmundzka-Hajnos c
a
Department of Pharmaceutical Botany, Medical University of Lublin, 1 Chodźki Street, 20-093 Lublin, Poland
b
Department of Medicinal Chemistry, Medical University of Lublin, 4 Jaczewskiego Street, 20-090 Lublin, Poland
c
Department of Inorganic Chemistry, Medical University of Lublin, 6 Staszica Street, 20-081 Lublin, Poland
d
Department of Biochemistry, Institute of Soil Science and Plant Cultivation – State Research Institute, 8 Czartoryskich Street, 24-100 Puławy, Poland

a r t i c l e i n f o a b s t r a c t

Article history: A novel, easy, and cheap technique for preliminary quantitative evaluation of antiradical activity, based
Received 17 December 2010 on HPTLC, has been proposed. This method combines chromatographic separation of polar compounds,
Received in revised form 16 May 2011 present in plant extracts, with data analysis by means of image processing software. Bleaching of the pur-
Accepted 22 October 2011
ple DPPH colour, caused by substances with antiradical activity, was observed and recorded using a
Available online 9 November 2011
photo camera. ImageJ, a free and open source image processing program was used for quantitative mea-
surements. For evaluation of assay efficiency, the antiradical activity of rose flower extracts (from Rosa
Keywords:
rugosa Thunb.) was expressed as Standard Activity Coefficients (SACs), which are relative measures of
RP-HPTLC
DPPH
the activity to the four well known antioxidants; i.e., quercetin, gallic acid, protocatechuic acid, and Trol-
ImageJ ox. The method uses small amounts of free radical and is easily applicable – only a digital camera with
Radical scavenging activity freely available open source software is required.
Rosa rugosa Thunb. Ó 2011 Elsevier Ltd. All rights reserved.

1. Introduction
A variety of methods are used to determine the radical scaveng-
ing capacity of pure individual compounds as well as crude plant
extracts. In vitro tests comprise a large group of antiradical capacity
determinations and they still arouse scientists’ interest, as they are
very useful even for preliminary analysis (Karadag, Ozcelik, & Sa-
ner, 2009; Paixão, Perestrelo, Marques, & Câmara, 2007). Generally,
the most popular are spectrophotometric methods, based on the
colour reaction between an antioxidant and a relatively stable rad-
ical, for example ABTS+ (2,20 -azinobis-(3-ethylbenzothiazoline-6-
sulphonic radical cation) or DPPH (2,2-diphenyl-1-picrylhydrazyl
radical) (Brand-Williams, Cuvelier, & Berset, 1995; Re et al.,
1999). Their history dates back to the 1950s when the DPPH rad-
ical was described for the first time (Blois, 1958). Since then, many
variants of antiradical measurements, based on bleaching of the
purple colour of DPPH, have been suggested. However, determina-
tion of the antioxidant potential and radical scavenging activity for
each individual compound contained in a complex mixture is
hardly a possible task. Therefore, the search for new, easy, rapid,
and cheap methods for screening a free-radical scavenging poten-
tial of individual compounds contained within complex mixtures
continues.
⇑ Corresponding author. Tel.: +48 81 742 37 03; fax: +48 81 742 38 05.
E-mail address: renata.nowak@umlub.pl (R. Nowak).
TLC is a method of choice when screening plant extracts for the
presence of biologically active compounds (Waksmundzka-Hajnos,
Sherma, & Kowalska, 2008). The role of bioautographic assays
using TLC plates is still growing due to several advantages; these
include speed of method development, high sample throughput
and flexibility. TLC also can provide quick access to information
concerning both the activity and the localisation of the activity in
complex plant matrices (Simões-Pires, Hmicha, Marston, & Host-
ettmann, 2009). TLC combined with DPPH application for the
detection of free-radical scavengers was first introduced by Glav-
ind and Holmer (1967). However, quantification of the free-radical
scavenging activity in situ, after dipping TLC plates in a DPPH solu-
tion, is somewhat difficult. Conventional slit-scanning densitome-
ters operate serially; thus, they cannot be used for the quantitative
determination of radical-scavenging activity directly from the
plate as concluded by Yrjönen, Peiwu, Summanen, Hopia, and Vuo-
rela (2003). The reaction with DPPH is very time-instable and den-
sitometric processing of such plates is almost unavailable, because
it brings unpredictable results. The authors of the aforementioned
paper have proposed a video-documentation system based on a
CCD camera for detection and quantification of the radical scav-
enging capacity. However, due to the lack of detailed descriptions
related to quantitative processing techniques of such images, new
solutions need to be sought.
The videoscanning concept of TLC is a well-established topic in
the context of quantitative determination (Vovk, Prosek, & Kaiser,
2001). Although densitometry generally has a better performance,

0308-8146/$ - see front matter Ó 2011 Elsevier Ltd. All rights reserved.
doi:10.1016/j.foodchem.2011.10.067
550 M. Olech et al. / Food Chemistry 132 (2012) 549–553
videoscanning is preferred in many fields, especially when the sep-
aration results on the TLC plate change in time after the develop-
ment. Because an illumination chamber and dedicated camera
are quite expensive, scientists search for cheaper alternatives pro-
viding similar performance. When the separated spots on TLC
plates can be detected under visible light, the plate can be scanned
by an office flatbed scanner or photographed with a digital camera.
A successful approach with a flatbed scanner was first presented by
Soponar, Mot, and Sarbu (2008), who used the equipment for
quantitative determination of some food dyes. Unfortunately, the
scanner cannot be used for creating digital images of wet plates,
just from the DPPH solution, as it is not a waterproof (nor chemi-
cal-proof) device. An interesting alternative could be processing of
pictures made with the use of a digital photo camera. Such a image
can be processed by various commercial and free software pack-
ages such as MATLAB (with Image Processing Toolbox), GNU R or
Scilab. Because the packages require programming (and chemo-
metric) knowledge by the operator, an interesting alternative pro-
posed by us is the application of ImageJ software. It is an open
source and free program written in Java (and therefore able to be
used almost with any operating system) developed at the National
Institutes of Health (NUH) in the United States Department of
Health and Human Services, especially for handling biological
images. This software is very easy to operate. In our study, it was
employed for handling pictures taken after separation and immer-
sion of TLC plates in a 2,2-diphenyl-1-picrylhydrazyl radical meth-
anolic solution. Such an approach for quantitative analysis of the
free-radical scavenging activity of crude plant extracts has not
been described earlier. ImageJ software contains many built-in
algorithms, which are – according to our investigation – sufficient
to process the images of TLC plates correctly.
In this report, we describe a useful method, with a new way of
data processing, to examine quantitatively the free radical scav-
enging activity of plant extracts. For this purpose, two types of Rosa
rugosa Thunb. (rugosa rose, Japanese rose) flower extracts were
analysed. R. rugosa Thunb. was chosen because a variety of antirad-
ical compounds have been reported in this species (Hashidoko,
1996; Nowak & Gawlik-Dziki, 2006). Furthermore, it is a cultivated
plant, commonly used for food, medicinal and cosmetic purpose.
Our previous study findings have shown that rugosa rose could
be of further interest (Nowak & Gawlik-Dziki, 2006).
2. Experimental section
2.1. Apparatus and reagents
Standards of gallic (GA), gentisic (GTA), protocatechuic (PCA),
and ferulic acids (FA), as well as quercetin (Q), Trolox (T) and
2,2-diphenyl-1-picrylhydrazyl radical (DPPH) were purchased
from Sigma–Aldrich Chemical Co. (St. Louis, MO, USA). Isoquercetin
(IzQ) was purchased from LGC standards (Dziekanów Leśny,
Poland). Ethanol, methanol, and hydrochloric acid were obtained
from Polish Reagents – POCH (Gliwice, Poland). All chemicals were
of analytical purity grade. TLC was performed on the 10 cm 
10 cm glass-backed RP-18W HPTLC plates (Merck, Darmstadt,
Germany).
Solutions of all test substances and plant crude extracts were
applied to the chromatographic plates band-wise by means of a
Desaga AS-30 automatic TLC sampler (Heidelberg, Germany) and
developed in the horizontal DS chambers (Chromdes, Lublin, Po-
land). Chromatograms were documented using a Canon EOS
350D digital camera with standard 18–35 mm lens (Tokyo, Japan).
The pictures so obtained were processed using the ImageJ 1.43u
image processing program (Wayne Rasband, National Institutes
of Health, USA; http://rsbweb.nih.gov/ij/).
2.2. Standard substance solutions and crude plant extracts
All reference substances were dissolved in methanol to prepare
the stock solutions (1.0 mg ml1). From the stock solutions, various
standard mixtures containing from 0.8 to 200 lg ml1 of known
compounds were prepared by appropriate dilution with methanol.
In addition, the methanolic solution (1 mg ml1) of Trolox (i.e., a
water-soluble vitamin E analogue) was used as an external stan-
dard antioxidant.
Flowers of R. rugosa Thunb. were collected in Elizówka (near
Lublin, Poland), in June 2009. The botanical material was authenti-
cated by Prof. Tadeusz Krzaczek and a voucher specimen was
deposited in the Department of Pharmaceutical Botany, Medical
University of Lublin, Poland. The plant material was dried at ambi-
ent temperature and powdered according to the 6th European
Pharmacopoeia (2007).
Twenty-five grams of dried and pulverised rose petals were ex-
tracted four times (for 24 h each), at ambient temperature with
150 ml of 80% (v/v) aqueous methanol. During the last extraction,
a sonication step at 50 °C (30 min) was applied in order to achieve
exhaustive extraction. Extracts were combined, filtered and evapo-
rated to dryness under vacuum. The residue was weighted and
redissolved into 80% (v/v) methanol to obtain stock methanolic
solutions (M) at a concentration of 100 mg ml1.
A part of the dry methanolic extract – 0.52 g (which corre-
sponds to 1 g of dried plant material) was hydrolysed in hot water
bath with 20 ml of 1.2 M HCl in 50% (v/v) methanol for 1 h (Nowak
& Tuzimski, 2005). Four millilitres of the hydrolysed sample were
diluted with 6 ml of water and passed through an octadecyl SPE
column (500 mg, JT. Baker Inc., Philipsburg, USA). First, the SPE col-
umns were conditioned with methanol, water, and 0.1 M HCl in
20% methanol and the diluted sample (10 ml) was transferred to
the column, washed with water to remove sugars liberated during
hydrolysis and then eluted with 5 ml of methanol. The collected
eluates were evaporated to dryness and redissolved in 80% (v/v)
methanol to obtain the hydrolysed extract H, at concentration of
100 mg of plant material ml1. Both extracts M and H were pre-
pared in duplicate.
2.3. Chromatography and DPPH radical derivatisation
All experiments were carried out at a stable temperature of
24 °C ± 1 °C. The plates were developed with methanol–water-o-
phosphoric acid (45:54:1 v/v) in a saturated twin trough chamber
(10  20 cm, DS, Poland) to the distance of 80 mm (developing
time 55 min), dried at ambient temperature for 15 min and im-
mersed for 3 s in freshly prepared 0.1% (m/v) DPPH radical meth-
anolic solution. After immersion and removal of DPPH excess, the
plates were immediately photographed daylight. The photos were
stored as JPG files for further image processing with method, as de-
scribed below.
3. Results and discussion
3.1. Optimisation of chromatographic conditions
The prepared extracts from petals of R. rugosa Thunb. contain
substances of high polarity, mainly polyphenols. Resolution of this
group of natural compounds raises several difficulties. Some of the
compounds present in crude extracts remain unresolved when sil-
ica gel layers were applied for their chromatography in preliminary
experiments. Highly polar compounds strongly adsorbed to the
surface of silica gel and did not move from the point of application.
Some substances also tailed on the TLC plates, despite the use of
acidic mobile phases. Thus, it was decided to separate the
 M. Olech et al. / Food Chemistry 132 (2012) 549–553
compounds on the surface of octadecylsilica gel (RP18, reversed-
phase system). First, the chromatographic conditions were opti-
mised for a prepared mixture of standards composed of Q, IzQ,
FA, GTA, PCA, GA in concentrations 0.2, 0.2, 0.2, 0.0008, 0.004,
and 0.1 mg ml1, respectively. Different concentrations of stan-
dards in the mixture were related to different radical scavenging
capacity presented by these compounds. The concentrations were
proportional to the recorded signals in a separated zone on the
TLC plate. Five microlitres of this standards’ mixture were applied
to RP-18W HPTLC plate as a 8-mm band. The best resolution was
obtained with the application of the following mobile phase: meth-
anol–water-o-phosphoric acid (45:54:1 v/v). The inclusion of a
strong mineral acid was essential to suppress the ionisation of
acidic compounds being chromatographed. The incorporation of
formic or acetic acid was insufficient to suppress the ionisation
of the analysed compounds in the aqueous mobile phase. The opti-
mised chromatographic conditions were then used to analyse both
standards and active constituents of R. rugosa Thunb. flower ex-
tracts. A good separation of active plant constituents was achieved
(see Fig. 1). The active antiradical constituents appeared as yellow-
ish, white spots, produced by bleaching the purple colour of the
DPPH reagent. A stronger radical scavenging capacity resulted in
more intensive area around the bleached spot on the TLC plate.
3.2. The image processing methodology
As already mentioned, the results can be documented using a
compact digital camera. However, the analyst must adhere to some
requirements in order to retain the usability of the data:
1. The chromatograms visualised on different photos cannot be
used for comparative analysis because they do not ‘‘match’’ in
the chemometric sense due to spatial differences (e.g., a dis-
tance between the camera and plate) (Russ, 2006). To make
them match, one would have to put the camera on the tripod,
placing the plates precisely in the same place and distance.
The photos would them have to be taken in the same exposition
(i.e., no automatic exposure time or aperture) under exactly the
Fig. 1. Radical-scavenging constituents of Rosa rugosa Thunb. flower extracts.
Chromatogram after treatment with DPPH solution, Track assignment: 1 – extract
M; 2 – mixture of standards; 3 – extract H. Mobile phase: methanol–water-o-
phosphoric acid (45:54:1 v/v) (RP18 HPTLC).
551
same light conditions. In practice, it means that one would have
to construct the illumination chamber analogous to commercial
videoscanning equipment. In the proposed approach such a
step is not needed, because all the chromatograms (i.e.,
obtained those for calibration purposes as well as for samples
and references) are documented on the same photo. Several
plates can be photographed on the same picture, but they must
be developed with DPPH in the same time as the colour reac-
tion changes quickly. A photo cannot be overexposed.
2. The plates must be photographed parallel to the photo borders.
A photo cannot be skewed. The focal length should be sufficient
to avoid a barrel distortion. The light should be enough to avoid
a noise or movement, and must be as ambient as possible. Inho-
mogeneous illumination will cause a significant baseline drift
and result in quantitative problems (Vovk & Prosek, 1997).
The drifting baseline can lead to problems in manual integrat-
ing of the areas under the common peaks.
3. The photo resolution, in megapixels, plays a secondary role
here. However, the photos should be saved in the best quality
possible (smallest compression) to minimise JPG compression
artifacts.
4. The chromatograms must be applied as bands, not spots,
because in the latter case the width of the track would play a
dominant role in evaluation of peak areas.
After dipping the plates in the DPPH solution, the chromato-
gram should be immediately documented by means of a camera
Fig. 2. An example of a raw univariate densitogram extracted from image before
denoising (a), after denoising (b) and after denoising and baseline drift removal (c).
552 M. Olech et al. / Food Chemistry 132 (2012) 549–553
and for data handling, the ImageJ software is applied. The built-in
gel option should be used, which was originally designed to ana-
lyse one-dimensional electrophoretic gels.
Two main problems can deal with a well done photo and in-
clude the noise removal and baseline drift correction. A CCD noise
(present both in videoscans and photographs) is almost white and
easy to cope with (Komsta, 2009a). The extensive comparative
study of many popular algorithms recommends the use of a med-
ian filter as the best (and very simple) approach (Komsta, 2009b).
The ImageJ software has the median filter (Process/filters/median)
implemented; therefore, applying such a filter with a 5–10 pixels
width should be the first step in image processing (Fig. 2a and b).
The next preprocessing step is to remove the baseline drift
caused by inhomogeneous illumination. Many solutions have been
reported in the literature and the most promising technique is
based on multidimensional tensors of splines (Kaczmarek, Walc-
zak, de Jong, & Vandeginste, 2005), but these complicated algo-
rithms are necessary only in the case of 2D electrophoresis. For
peak-based quantification, a rollerball algorithm or high-pass 2D
filter is fast and sufficient. Yet, the rollerball algorithm did not
work well in this study. In fact, it requires only black and white
images. The high-pass 2D filters worked well and are implemented
in ImageJ without any plugins (Process/FFT/Bandpass Filter, the
second parameter should be set to 0); they can be applied after
denoising by the median filter. In our case, the cutoff value of
‘‘large structures’’ was chosen in the range of 50–100 (Fig. 2c).
After denoising and removing the baseline drift, the rectangular
selection tool was used to outline the desired track on the plate’s
picture. The width should be as large as possible, while preserving
the homogenous thicknesses of the bands (i.e., not marking round
borders of the spots). The ImageJ software can compensate auto-
matically for differences in the track width. The option ‘Select First
Lane’ was therefore applied. Furthermore, the ‘Plot Lanes’ option
was used to generate the line profile plots, followed by the
‘Straight Line Selection Tool’ to draw baselines so that each peak
Fig. 3. An example of quantitative integration of peaks in the case of complex and
not fully-separated plant extracts.
Table 1
Statistical evaluation of the linearity response of standard solutions and plant extracts.
Sample Linear range [lg] Slope
M 15.0–62.5 297.1
H 2.50–15 1169
GA 0.25–3.0 4956
Q 0.25–3.0 4937
T 0.50–2.5 2399
PCA 0.01–0.2 11274
of interest defines a closed area. The area under the individual
peaks was measured by clicking inside the peak with the ‘Wand
Tool’.
After extracting the univariate videodensitogram, the baseline
should be marked by a line tool as one line from the start to the
end (without connecting the valleys between peaks, as shown in
Fig. 3).
3.3. Evaluation of radical scavenging capacity
Image processing of the chromatograms denotes radical scav-
enging constituents as negative peaks. The proposed quantitative
methodology is based on measurements of TEA (Total Extract
Area); i.e., the total area under a videoscan of an extract (approxi-
mation of the sum of all the peaks) which can be considered as the
sum of antiradical activities of all extract ingredients. The TEA va-
lue is compared to areas obtained for standards, and then calcu-
lated using the experimentally determined formulae. Only one
requirement is working inside the linear range of the peak area re-
sponse, both for the standard and extract measured.
To determine the linear range, different amounts per band of
four standards (GA, PCA, Q, T) and rose extracts (M and H) were
chromatographed to examine the peak area response against the
applied amount of antioxidant. All aliquots were applied in tripli-
cate. The reference standards (GA, PCA, Q, T) were selected because
of their well-known radical scavenging capacity and frequent use
as reference substances during the estimation of antioxidant activ-
ity. The experimental linear ranges for the calibration curves and
the statistical evaluation are given in Table 1, which also contains
the limits of detection (LOD) for the reference compounds, evalu-
ated as the smallest amounts causing the visibility of the spot.
Having the linear calibration curve for a reference compound and
working in the linear range of an extract, one can compute the equiv-
alent amount of the standard, giving the same antiradical activity
(TEA) as the extract. This can be called the standard activity coeffi-
cient (SAC). For example if the SAC equals 2, this mean that the dry
extract is twice as active as the pure reference compound.
mS
SAC ¼
mE
where mS and mE are the quantities of standard and extract, respec-
tively in lg/spot; mS, mE are in the linear range, and both have the
same antiradical activity TEA.
The standards and crude extracts cannot be applied at levels
resulting in the same TEA, because of different linearity ranges.
Additionally, at lower concentrations, some antioxidant com-
pounds can be applied below the detection limit. Furthermore,
both samples are spotted in the middle of their linearity range
and the SAC is calculated from the formula containing the TEA of
an extract and calibration curve coefficients of a standard:
TEA  I
SAC ¼
S  mE
where TEA – total extract area; I – intercept from the standard’s
curve equation; S-slope from the standard’s curve equation;
Intercept p (Mandel’s test) LOD [lg] r
585.8 0.09 – 0.9527
302.4 0.43 – 0.9979
11418 0.39 0.25 0.9683
3026 0.25 0.25 0.9920
543.1 0.87 0.5 0.9897
4054 0.21 0.01 0.9822
 M. Olech et al. / Food Chemistry 132 (2012) 549–553 553

Table 2
SAC values of the extracts.

Sample M H
Standard 20 lg 25 lg 40 lg Mean RSD% 10 lg 12 lg 14 lg Mean RSD%
GA 0.35 0.41 0.38 0.38 8.45 0.74 0.80 0.73 0.76 5.12
Q 0.43 0.48 0.43 0.45 6.14 0.92 0.94 0.85 0.91 5.47
T 0.94 1.03 0.91 0.96 6.14 2.01 2.03 1.82 1.95 5.83
PCA 0.18 0.21 0.19 0.19 6.22 0.39 0.41 0.37 0.39 5.34

mE-applied microgram of extract. Due to introduction of parameters References

from the standard’s curve equation, the SAC becomes independent
of the applied amount of extract.
For further analysis, only data obtained from linearity was
processed. To obtain the SAC value of extracts, the aliquots of
nonhydrolysed (M, 25 lg/spot) and hydrolysed (H, 12 lg/spot)
extracts were applied as 8-mm bands on an HPTLC plate. The
SAC values for four reference compounds were obtained (Table 2).
Table 2 also contains the computed values from other amounts of
spiked extracts as a means to evaluate the robustness of the
method. The robustness is expressed as the relative standard devi-
ation (RSD) between three SACs obtained with different spiked
amounts of an extract.
The method for the estimation of the radical scavenging capac-
ity of plant extracts is easy and convenient in practice. The chro-
matographic response is a direct function of quenched DPPH
(and an observed negative peak is indeed a peak of DPPH, but its
area depends on the antiradical capacity of the compound). Thus,
all peaks can be directly compared against their antiradical
activity; the total antiradical activity can be linearly added and ex-
pressed as some equivalent of the reference compounds. Moreover,
there is no need to work in the UV range, and the plate can be pho-
tographed in visible light.
4. Conclusions
The method is simple, fast and efficient for free-radical scaveng-
ing activity analysis of phytochemicals and crude plant extracts. It
could be an easy way to assess antiradical potential of herbal ex-
tracts and any other plant food products, enabling separation and
detection of their active constituents simultaneously. This method
is easy, rapid, cheap and could be available for almost every analyt-
ical laboratory. Further advantages of the procedure are its relatively
high reproducibility as well as short analysis time. Furthermore, in
HPTLC DPPH decreased amounts of dangerous radical can be used
during analysis compared to the classical UV-based method.
One of the most important drawbacks is the necessity to pre-
pare fresh DPPH solution during each analysis, which, however
is typical of all experimental work related to radicals.
Acknowledgement
Blois, M. S. (1958). Antioxidant determinations by the use of a stable free radical.
Nature, 4617, 1199–1200.
Brand-Williams, W., Cuvelier, M. E., & Berset, C. (1995). Use of a free radical method
to evaluate antioxidant activity. Food Science and Technology, 28, 25–30.
European Pharmacopoeia (2007). 6th ed., Strasbourg: Council of Europe, cop.
Glavind, J., & Holmer, G. J. (1967). Thin-layer chromatographic determination of
antioxidants by the stable free radical alpha, alpha-diphenyl-betapicrylhydrazyl.
Journal of American Oil Chemists Society, 44, 539–542.
Hashidoko, Y. (1996). The Phytochemistry of Rosa rugosa. Phytochemistry, 43,
535–549.
Kaczmarek, K., Walczak, B., de Jong, S., & Vandeginste, B. G. M. (2005). Baseline
reduction in two dimensional gel electrophoresis images. Acta
Chromatographica, 15, 82–96.
Karadag, A., Ozcelik, B., & Saner, S. (2009). Review of methods to determine
antioxidant capacities. Food Analytical Methods, 2, 41–60.
Komsta, Ł. (2009a). Suppressing the charged coupled device noise in univariate thin
layer videoscans: A comparison of several algorithms. Journal of
Chromatography A, 1216, 2548–2553.
Komsta, Ł. (2009b). Dealing with charged-coupled device noise in thin-layer
videodensitometry. Optimization of several image-denoising techniques. Acta
Chromatographica, 21, 355–367.
Nowak, R., & Gawlik-Dziki, U. (2006). Polyphenols of Rosa L. leaves extracts and
their radical scavenging activity. Zeitschrift für Naturforschung C – Journal of
Bioscience, 62, 32–38.
Nowak, R., & Tuzimski, T. (2005). A solid-phase extraction-thin-layer
chromatographic-fiber optical scanning densitometric method for
determination of flavonol aglycones in extracts of rose leaves. Journal of
Planar Chromatography, 18, 437–442.
Paixão, N., Perestrelo, R., Marques, J. C., & Câmara, J. S. (2007). Relationship between
antioxidant capacity and total phenolic content of red, rose and white wines.
Food Chemistry, 105, 204–214.
Re, R., Pellegrini, N., Proteggente, A., Pannala, A., Yang, M., & Rice-Evans, C. (1999).
Antioxidant activity applying an improved ABTS radical cation decolorization
assay. Free Radical Biology and Medicine, 26, 1231–1237.
Russ, J. C. (2006). The Image Processing Handbook (6th ed.). Boca Raton: CRC Press.
Simões-Pires, C. A., Hmicha, B., Marston, A., & Hostettmann, K. (2009). A TLC
bioautographic method for the detection of a- and b-glucosidase inhibitors in
plant extracts. Phytochemical Analysis, 20, 511–515.
Soponar, F., Mot, A. C., & Sarbu, C. (2008). Quantitative determination of some food
dyes using digital processing of images obtained by thin-layer chromatography.
Journal of Chromatography A, 1188, 295–300.
Vovk, I., & Prosek, M. (1997). Reproducibility of densitometric and image analysing
quantitative evaluation of thin-layer chromatograms. Journal of Chromatography
A, 768, 329–333.
Vovk, I., Prosek, M., & Kaiser, R. E. (2001). A retrospective view for the third
millennium. In Sz. Nyiredy (Ed.), Planar Chromatography (pp. 464–488).
Budapest: Springer Verlag.
Waksmundzka-Hajnos, M., Sherma, J., & Kowalska, T. (2008). Thin Layer
Chromatography in Phytochemistry. Boca Raton: CRC Press, Francis and Taylor.
Yrjönen, T., Peiwu, L., Summanen, J., Hopia, A., & Vuorela, H. (2003). Free radical-
scavenging activity of phenolics by reversed-phase TLC. Journal of the American
Oil Chemists’ Society, 80, 9–14.

This work was financially supported by the Polish Ministry of

Science and Higher Education (Grant No. N N405 617938).