Journal of Chromatographic Science, Vol.

37, June 1999

Quantitative Applications of Fluorescence and

Ultraviolet Scanning Densitometry for Compositional
Analysis of Petroleum Products in Thin-Layer
Chromatography

Downloaded from https://academic.oup.com/chromsci/article-abstract/37/6/219/315785 by guest on 22 March 2019

‡
Vicente L. Cebolla*, Luis Membrado, María P. Domingo, Patrick Henrion†, Rosa Garriga , and Paula González
Departamento de Procesos Químicos, Institute de Carboquímica, CSIC, María de Luna, 12. 50015 Zaragoza, Spain

Fernando P. Cossío* and Ana Arrieta

Kimika Fakultatea, Euskal Herriko Unibertsitatea, P.K. 1072, 20080 San Sebastián-Donostia, Spain

Jesus Vela
Grupo de Espectroscopía Analítica y Sensores, Departamento de Química Analítica, Facultad de Veterinaria, Universidad de Zaragoza,
Avenida Miguel Servet 177, 50013 Zaragoza, Spain

Abstract ing berberine and an alkane is applied on the thin layer, an

enhancement of the fluorescence signal in the visible region
Direct application of the fluorescence response enhancement occurs when the system is irradiated with monochromatic
(derived from the interaction of the berberine cation with ultraviolet (UV) light (2,3). This phenomenon was reported in
saturated hydrocarbons) to petrochemical analysis is achieved. a qualitative manner (visual inspection) in 1947 and 1980,
Hydrocarbon types are successfully determined in middle (gasoil) when the modern technology of densitometers was not avail­
and heavy (lubricants, vis-breaking fuel, heavy oil) petroleum able. In a previous work (4), the present authors proposed a
distillates with adequate precision and sensitivity. The novelty of theoretical explanation for it, laying the groundwork for
this method resides in the fact that a single fluorescence scanning
designing new interaction systems of analytical interest. A
is only needed for the determination of saturates and aromatics in
direct application can be the analysis of compounds in the
the same chromatogram. In spite of their chemical inertness,
saturated hydrocarbons can be quantitatively determined using
field of petroleum derivatives.
thin-layer chromatographic densitometry. The fluorescent response Analysis of compound classes, also called hydrocarbon
depends on the alkane structure. A reliable analysis depends on group-type analysis, is important in petrochemistry and car-
the adequate choice of variables that influence fluorescence bochemistry for the choice of process conditions, process mon­
response, precision, and sensitivity: sample load, application itoring, compliance with environmental regulations, and
volume, beam size, and impregnation conditions (berberine evaluation of fuel quality, catalyst performance, and feed
concentration and impregnation time). Sensitivity of the analysis can processability (5). This analysis involves the separation and
be tailored to a certain extent through control of the last parameter.
quantitation of the corresponding hydrocarbon types in the
The quantitative results agree with those provided using other
sample; therefore, chromatographic techniques are frequently
well-established techniques in the petrochemical industry.
used. The boiling point range and polarity of the sample to be
analyzed (6) are used as selection criteria for a particular chro­
matographic technique and method.
Introduction Among the available chromatographic techniques, TLC (or
planar chromatography) has the advantages of being compar­
atively simple, rapid, robust, and inexpensive. Nevertheless, it
Fluorescence enhancement of a broad variety of solutes has
has not been widely used for petroleum analysis, mainly
been used for a long time in thin-layer chromatography (TLC) (1).
because it is not considered quantitative or sensitive enough,
If a TLC silicagel plate is impregnated with a solution contain-
and the detection of saturated hydrocarbons (classically inert
*Authors to whom correspondence should be addressed: e-mail [email protected],
molecules that do not exhibit UV or fluorescence spectra in the
[email protected]. usual wavelength ranges) is not immediate using scanning
†
On leave from Laboratoire de Thermodynamique et d'Analyse Chimique, Faculé des Sciences,
densitometry.
Université de Metz, France.
‡
On leave from Departamento de Química Orgánica y Química Física, Facultad de Ciencias,
However, TLC presents several advantages: high sample
Universidad de Zaragoza, Spain. throughput, simultaneous processing of standards and sam-

Reproduction (photocopying) of editorial content of this journal is prohibited without publisher's permission. 219
 Journal of Chromatographic Science, Vol. 37, June 1999

ples, analysis of the whole sample without previous fractiona­ ferent proportions were analyzed. Saturates and aromatic frac­
tion or deasphalting, and the possibility of physically storing tions were obtained from CEPSA from the raw lubricant by the
the developed plates. strict application of a liquid chromatography-based ASTM
Conventional TLC silicagel and alumina plates (with devel­ D2007 standard method (21).
opments using solvents of increasing polarity) have been In the case of the vis-breaking fuel, compound class com­
mostly used for either qualitative and semiquantitative hydro­ position was also provided by CEPSA using TLC-flame ioniza­
carbon-type analysis or the preparative isolation of hydro­ tion detection (FID). Detailed descriptions of the technology,
carbon families in heavy distillates in order to carry out either method, and quantitative conditions can be found elsewhere
further analysis or the gravimetric determination of collected (19).
fractions (7-12). In general, nonquantitative analyses have
involved the visual detection of colorless substances using
Analytical TLC separation
inspection under UV illumination or the use of chromogenic
Silicagel plates (aluminium sheets, 20 × 20 cm, 5- to 25-μm
or fluorogenic spray reagents (13). At present, only few works

Downloaded from https://academic.oup.com/chromsci/article-abstract/37/6/219/315785 by guest on 22 March 2019

particle size, 60-Åpore size, 0.2-mm layer; Panreac, Barcelona,
deal with quantitative hydrocarbon-type analysis of fossil
Spain) and silicagel-impregnated plates with berberine sulfate
fuels using UV and fluorescence scanners (14,15).
(Across Chimica, Geel, Belgium) were used. A solution of
In our opinion, TLC has been underfrautilized in the case
berberine sulphate in methanol (6 mg/100 mL) was used for
of petroleum. The recent progress (16-18) in layer manu­
impregnation for 20 s. Plates were subsequently dried hori-
facturing (including small-sized high-performance TLC
zontally overnight at 40°C.
plates), sample application devices, development instru­
Samples were spotted at least in duplicate onto TLC plates
mentation, and automated UV and fluorescence scanning
using an SES 3202/IS-02 automatic spotter (Germany). Applied
densitometers should allow an improvement to be achieved
volumes of 0.4 μL were typical. Sample loads usually ranged
with regard to the separation and quantitation of petro­
from 1 to 20 μg.
leum-derived samples.
After application, samples were eluted using a sequence of
In 1991, Marsh and Hiekane (14) carried out an analysis of
eluants. The normal elution scheme for the lubricant and
bitumen using fluorescence (for saturates) and UV densitom­
gasoil was n-hexane (9 min) and dichloromethane (DCM) (4.5
etry (for the other peaks). However, neither the nature nor
min). A normal elution scheme for the vis-breaking fuel and
the influence of the different analytical parameters were
the heavy oil was n-hexane (30 min), toluene (8 min), and
studied, and the possible application to other kinds of samples
DCM (3 min). These eluants were obtained from Scharlau
was not considered.
(Barcelona, Spain). In the cases where short development
A more complete study of the application of TLC-UV-
times were employed, both sides of the plates were often used.
fluorescence scanning densitometry to the compositional
Plates were dried at room temperature after hexane elution.
analysis of petroleum products, including middle distillates as
Plates were always dried at room temperature when toluene
gasoil, is presented in this work. Quantitative, sensitive, and
was not used as eluent. When using toluene, plates were dried
precise methods based exclusively on fluorescence enhance­
at l00°C for 3 min.
ment for the compositional analysis of saturates and aro-
matics are presented, along with alternative methods for
aromatics using UV detection. The variables that influence Fluorescence and UV scanning densitometry
sample response, precision, and sensitivity of analysis have After development, a CS9301 TLC scanner (Shimadzu,
been studied. Japan) was used in either the fluorescence or UV mode to pro­
vide detection of the peaks.
In the case of berberine, fluorescence scanning densitometry
was used, usually under the following conditions: λ 365 nm;
exo,

Experimental beam size, 1 × 1 mm; filter 3 (visible zone); linear scanning.

In the cases of both nonimpreganted silicagel and coated
Samples (impregnated) plates, UV scanning densitometry was used. The
Pure n-alkanes (Fluka, Basel, Switzerland; Ultrascientific, wavelength was 250 nm. Linear or zigzag scanning and different
North Kingstown, RI) were used. The gasoil, synthetic lubri­ beam sizes were used in order to obtain adequate sensitivity.
cant, and vis-breaking (a thermal treatment for fuels that Peak positions and distances of spotting between the dif­
decreases viscosity without additives) fuel were provided by ferent lanes were determined using either UV bidimensional
CEPSA (Madrid, Spain). The heavy oil was a deasphalted oil scanning (CS9301 software) or conventional scanning (per­
from a 450°+ vacuum distillate Brent residue (19). pendicular to the direction of development, at the y-coordinate
In the case of gasoil, the content in aromatics was provided corresponding to undeveloped peaks).
by CEPSA using a high-performance liquid chromatography- Peak area data were taken from the scanner, and all data
refractive index (HPLC-RI)-based IP 391 standard (20). were collected, displayed, and stored using Shimadzu CS9310
The synthetic lubricant was prepared from the main frac­ PC software. The Kubelka-Munk equation (22) was applied
tions (saturated and aromatic) of the original lubricant by through the software to each data point in order to linearize
mixing these fractions in known proportions in order to vali­ the area-mass response. An SX = 3 factor, recommended by the
date the results. Different lubricants obtained by mixing dif­ manufacturer for 0.2-mm-thick silicagel plates, was used.

220
Journal of Chromatographic Science, Vol. 37, June 1999

Preparative isolation of calibrating fractions and Results and Discussion

calibration procedure
Quantitative determination was carried out in each case by Quantitative determination of saturates using
external calibration using fractions from the corresponding TLC-fluorescence scanning densitometry
fuel itself. Berberine sulfate is a fluorescent, aromatic, heterocyclic
Preparative TLC for the isolation of saturate and aromatic alkaloid. If a TLC silicagel plate is impregnated with a solution
fractions from the studied products (except the lubricant) containing the berberine cation and an alkane is applied to the
was carried out using glass silicagel TLC plates (20 × 20 cm, thin layer, an enhancement of the fluorescence signal in the
2-mm layer, 5- to 25-μm particle size; 60-Á pore size). After visible region occurs when the system is irradiated with mono­
sample application, n-hexane was allowed to develop to the chromatic UV light (e.g., λ = 365 nm) (2,3,14). Figure 1 shows
end of the plate. The separation edge between saturates and chromatograms corresponding to the saturates and aromatics
aromatics in a preparative separation was monitored by determined in different petroleum products. Saturates are
either fluorescence scanning using a berberine-impregnated detected using fluorescence after elution with n-hexane. Aro­

Downloaded from https://academic.oup.com/chromsci/article-abstract/37/6/219/315785 by guest on 22 March 2019

plate or UV scanning. After choosing the cutting point, the matics are detected by UV after elution with toluene.
preparative isolation was carried out in a nonimpregnated It was demonstrated in a previous work using computa­
silicagel plate. After development, the zones corresponding tional calculations that an ion-molecule interaction between
to saturates and aromatics were scratched and subsequently alkanes and berberine cation is responsible for this phenom­
extracted from the silicagel in a Soxtec apparatus (Tecator, enon (4). The role of the alkane is to provide an apolar environ­
Sweden) under mild conditions (dichloromethane solvent) ment for the excited berberine cation that hinders alternative
and dried under N until constant weight. Purity of the
2 relaxation mechanisms and favors fluorescence emission. The
fractions was monitored by fluorescence and UV scanning phenomenon is also observed for compounds that have a par­
densitometry (vide infra). tially saturated structure, as long as they adopt extended con­
The saturate and aromatic fractions from the lubricant formations along the main berberine axis. Therefore, alkyl-
were provided by CEPSA. They were obtained by the applica­ substituted aromatics should also be detectable, provided the
tion of the ASTM D2007 standard method. length of the alkylic chain is long enough. This suggests the

Figure 1. TLC chromatograms corresponding to saturate fractions determined using fluorescence scanning densitometry on berberine-coated silicagel plates
(1) and aromatic fractions determined using UV scanning densitometry on the same plates (2): heavy oil (A), vis-breaking fuel (B), lubricant (C), gasoil (D).

221
 Journal of Chromatographic Science, Vol. 37, June 1999

possibility of performing a complete analysis of petrochemicals The addition of each new - C H - group results in an increase
2

using TLC-fluorescence scanning densitometry, as will be in the fluorescent response. The curves of response corre­
explained later. sponding to the fractions of the saturates from the analyzed
Variables that influence fluorescence, chromatographic petrochemicals show coherence with regard to the results
response, precision, and sensitivity of determinations are obtained from the n-alkanes (Figure 2). The lubricant-satu­
sample load, length of paraffinic chain, application volume, rate fraction that consists, on average, of long alkanes (> C ) 2 4

development length, beam size, and impregnation condi­ provided higher response than C . In turn, the response of the
2 4

tions. gasoil-saturate fraction is near the C response. Therefore,

18

For a given wavelength and concentration of berberine in the the map of the fluorescence responses of n-alkanes may be
plate, the fluorescent response is proportional to the mass of used to obtain some clues about their nature, an interesting
the alkane applied and also depends on the alkane structure. possibility for unknown environmental samples in order to
identify types of pollutants.
The scanning beam size and the impreg­

Downloaded from https://academic.oup.com/chromsci/article-abstract/37/6/219/315785 by guest on 22 March 2019

nation conditions have a clear influence
on the sensitivity of the fluorescent detec­
tion and also affect precision. The larger
the beam size, the lower the sensitivity.
For the volumes (0.4 μL), sample loads,
and chromatographic conditions typically
used, 1 × 1 mm was the minimum beam
size that allowed the totality of the peak to
be scanned. As expected, it provided the
highest sensitivity, as seen in Figure 3.
Smaller beam sizes did not cover the peaks
and therefore were not included in the
study. Development length also has a clear
influence on peak width. Therefore, elu-
tion and detection conditions must be
chosen with some care in order to achieve
a good compromise between peak separa­
Figure 2. Fluorescent chromatographic responses with sample load for pure n-alkanes (1, n-C ; 2, n-
12 tion and sensitivity for a given sample.
C ; 3, n-C ; 5, n-C ; 7, n-C ) and saturates from the lubricant (6) and the gasoil (4). Conditions:
16 18 24 30
Because the fluorescence response of
development, 9 min with n-hexane; impregnation, 2 mg berberine/100 mL methanol; volume of appli­
berberine was used to detect saturated
cation, 0.4 μL; beam size, 1 × 1 mm.
hydrocarbons, the berberine concentration
on silicagel resulting from the impregna­
tion process will directly influence the
response. The sensitivity of the analysis will
be determined by berberine concentration
in the impregnating solution and impreg­
nation time. Three concentrations of ber­
berine (2, 4, and 6 mg/100 mL methanol)
and two impregnating times (20 and 50 s)
were studied in duplicate for a gasoil satu­
rate fraction (Figure 4). Plates impregnated
for 50 s gave worse repeatabilities and base­
lines t h a n those obtained using 20-s
impregnated plates, probably because of a
deterioration of the silicagel-aluminium
sheets. In the case of the 20-s impregna­
tion time, the saturation of fluorescence
signal occurred at sample loads higher than
100 μg for concentrations of 2 and 4 mg
berberine/100 mL methanol and at 40 μg
for 6 mg berberine/100 mL methanol.
Figure 3. Influence of scanning beam size on the fluorescent response for the gasoil saturate fraction:
The simultaneous selection of impreg­
1, 1 × 1 mm; 2, 1 × 2 mm; 3, 0.4 × 5 mm; 4,1 × 5 mm; 5,0.4 × 10 mm; 6,1 × 10 mm. Conditions:
nating conditions, application volume (in
development, 9 min with n-hexane; impregnation, 2 mg berberine/100 ml methanol; volume of appli­
this case, 0.4 μL), development time (e.g., 9
cation, 0.4 μL; beam size, 1 × 1 mm.
min), and beam size (allowing the totality of

222
Journal of Chromatographic Science, Vol. 37, June 1999

sample to be scanned) provides adequate area-mass linear cent relative standard deviation (for 5 measurements) was
regressions and precisions. (Tables I and II). lower than 5.
The direct determination of saturates is possible with this Under the impregnation conditions employed, fluorescence
method on other samples as well. Table I shows results con­ responses were linear with adequate precision at least in the
cerning the quantitative determination of saturates for a lubri­ range of 1-20 μg for saturates and 1-13 μg for aromatics.
cant, a heavy oil, and a vis-breaking fuel. In each case, the Table III also shows that the saturate and aromatic contents
application of different sample loads of the corresponding from the fluorescence chromatograms agree quite well with
preparative fraction of saturates was used to
perform the calibration curve.
Quantitative results agree quite well with
those using TLC-FID for the studied prod­
ucts. Under the impregnation conditions
used (6 mg/100 mL methanol), sensitivity is

Downloaded from https://academic.oup.com/chromsci/article-abstract/37/6/219/315785 by guest on 22 March 2019

adequate, and the linear mass range for
these products is between 3 and 16 μg.

Quantitative compound class analysis of

middle and heavy petroleum distillates
using TLC-fluorescence scanning
densitometry
Typical analyses of petrochemicals include
the determination of saturates, aromatics,
and polar materials. Figure 1 shows the
chromatograms of several analyzed prod­
ucts. In this case, samples were scanned
twice: saturates were obtained using fluo­
rescence scanning under the conditions
already described, and the other peaks were Figure 4. Influence of impregnating conditions on the fluorescent response for the gasoil saturate frac­
obtained using UV scanning at 250 nm. In tion: 1, 2 mg berberine/100 ml methanol (time, 20 s); 2, 2 mg berberine/100 mL methanol (time, 50
essence, this procedure is a direct applica­ s); 3, 4 mg berberine/100 mL methanol (time, 20 s); 4, 6 mg berberine/100 mL methanol (time, 20 s);
5, 4 mg berberine/100 mL methanol (time, 50 s).
tion of the previous work of Marsh and
Hiekane (14) to different samples.
Nevertheless, TLC-fluorescence scan­ Table I. Percentage of Saturates Fraction (Weight %) of Different Petroleum
ning densitometry on berberine-impreg- Products Determined Using TLC-Scanning Fluorescence Densitometry with
nated plates can also be used to perform a Berberine-Coated Silicagel Plates* and Other Techniques
quantitative determination of both satu­
rates and aromatics in these samples, Sample TLC-fluorescence Other techniques
because the aromatic fraction usually
includes alkyl aromatic compounds, re­ Heavy oil 35.2 36.1 (TLC-FID)
moving the need to scan the plates twice. Vis-breaking fuel 12.3 12.1 (TLC-FID)
Lubricant (spiked sampled ) †
56.2 56.8 (Spiked)
The particular characteristics of the alkylic
chains in the alkyl aromatics will greatly * Impregnation, 6 mg/100 mL methanol.
influence the detection limit for these com­ Samples spiked from their previously isolated saturate and aromatic fraction according to ASTM D2007.
pounds. In this case, polars are determined
by difference. The aromatic fraction could
also show an intrinsic fluorescence re­ Table II. Precision and Linear Regression Data for TLC-Fluorescence Analysis
sponse, especially in the case of samples of Different Petroleum Products Using Berberine-Coated Silicagel Plates* and
that include condensed aromatics. These Other Techniques
samples can be analyzed as well, although Equation †

the origin of the fluorescence signal would

Sample % RSD Linear interval m b r
be different.
Figure 5 shows t h e complete c h r o ­
Heavy oil 6.85 3-16 567.5 4574 0.999
matograms corresponding to the separation Vis-breaking fuel 1.85 4-20 979.7 4876 0.996
of the lubricant and gasoil studied with this Lubricant (spiked sample ) ‡
2.21 4-16 306.8 2464 0.993
method. The linearity intervals of the fluo­
rescence response for the aromatic fraction
*
Impregnation, 6 mg/100 mL methanol.
†
Area = m mass + b. Regression coefficient = r.
of these samples were similar to those found ‡
Sample spiked from their previously isolated saturate and aromatic fraction according to ASTM D2007.
for saturates (Table III). In all the cases, per­

223
 Journal of Chromatographic Science, Vol. 37, June 1999

those found using other reference techniques. In the particular Alternative determination of aromatics in middle and heavy
case of gasoil, reference values are obtained using HPLC-RI, distillates using TLC-UV scanning densitometry
according to the IP391 standard (20). This standard allows the The complete analysis of the sample using only fluorescence
calculation of the saturate content by difference, whereas the scanning has clear, practical advantages, but it obviously
present method, using detection based on the fluorescence depends on the nature and abundance of the aromatic fraction.
response, provides a direct determination of this value. It It cannot be assumed that the aromatic fraction will always be
should be noted that quantitative analysis of compound classes detectable, although impregnation conditions allow the mag­
of gasoils using TLC have not previously been reported in the nitude of response to be tailored to a certain degree, and the
literature, as far as the authors of this work know. aromatic fraction might also have an intrinsic fluorescence
According to the previously proposed explanation, the ob­ response. In any case, the alternative determination of aro­
served higher response of aromatics from the lubricant, when matics using UV detection remains available because it is
compared with those from gasoil, will show the contribution clearly suitable for this kind of detection, and there is no need
of the presence of longer alkylic chains. The particular char­ to perform a second analysis using a nonimpregnated plate

Downloaded from https://academic.oup.com/chromsci/article-abstract/37/6/219/315785 by guest on 22 March 2019

acteristics of the alkylic chain (in the case of alkyl aromatics) because the aromatic fraction can also be quantitated in the
will also greatly influence the detection limit for these com­ berberine-impregnated plates. Figure 6 and Table IV illustrate
pounds. this possibility.

Figure 5. TLC chromatograms corresponding to saturates and aromatics, both determined using fluorescence scanning densitometry on berberine-coated silicagel
plates: lubricant (A) and gasoil (B).

Table III. Quantitative (Weight %) Determination of Compound Classes for the Studied Lubricant and Gasoil
Using TLC-Scanning Fluorescence Densitometry on Berberine-Coated Silicagel Plates*
†
Equation
Sample Reference values TLC-flourescence Linear interval
compound classes (weight %) (weight %) (μg) m r

Lubricant
saturates 56.8 56.2 4-16 306.8 2464 0.993
aromatics 43.2 43.8 3-16 432.5 4754 0.992

Gasoil
‡
saturates 70.8 69.8 5-20 357.6 4818 0.990
‡
aromatics 29.2 30.2 1-13 242.7 212 0.98

* Conditions: development, 9 min with n-hexane and 4.5 min with dichloromethane-methanol (95:5, v/v); impregnation, 6 mg berberine/100 mL methanol; volume of
application, 0.4μL;beam size, 1 × 1 mm.
†
Area = m mass + b. Regression coefficient = r.
‡
Determined using HPLC-RI according to standard IP391; aromatics are directly determined by this technique, but saturates are obtained by difference.

224
journal of Chromatographic Science, Vol. 37, June 1999

orescence method does allow the determi­

nation of saturates to be directly performed
without the technical problems associated
with the use of an RI detector.
A l t h o u g h t h e f l u o r e s c e n t response
depends on the chain length and amount of
the corresponding hydrocarbon, the mag­
nitude of the response can be modulated, if
necessary, using an adequate scanning
beam size or through the modification of
the conditions of berberine impregnation.
As explained previously, in the case of
aromatics determination, at least a part of

Downloaded from https://academic.oup.com/chromsci/article-abstract/37/6/219/315785 by guest on 22 March 2019

the fluorescent response comes from the
alkylic chain. Even in the case in which
fluorescence sensitivity is not adequate for
determining short-chain or low-abundance
alkyl-aromatics in some products, an alter­
native UV scanning reading allows aroma­
Figure 6. Responses (area counts) versus mass of aromatics for the lubricant and gasoil using fluores­ tics to be detected on the same berberine-
cence and UV scanning densitometry (λ = 250 nm) on silicagel (2 and 3) and berberine-coated silicagel coated silicagel plate.
(1,4, and 5) plates: 1, lubricant, fluorescence; 2, gasoil, UV; 3 and 4, lubricant, UV; 5, gasoil, fluo­
rescence. Impregnation conditions: 6 mg berberine/100 mL methanol.

Acknowledgments
Table IV. Comparison Between Aromatics Determination
(Weight %) Using Either Fluorescence or UV Scanning
Densitometry This work was supported by the Plan Nacional de I+D
(Spanish CICYT, project QUI98-0852 and the Gobierno
Sample TLC-fluorescence* TLC-UV †
Vasco/Eusko Jaurlaritza (project GV 170.215-EX97/11).
P. Henrion gratefully acknowledges a grant from the Agence
Lubricant 43.8 42.1 Française pour la Maitrise de l'Energie (ADEME, France) and
Gasoil 30.2 27.0 the Rotary Club (Hagondange, France). Dr. J. Lázaro, Dr. J.
Mazón, and CEPSA are acknowledged for the gift of petro­
* On berberine plates (6 mg/100 mL methanol).
†
On berberine plates (6 mg/100 mL methanol); UV zigzag scanning (beam size, chemical samples and the saturate and aromatic fraction from
0.4×0.4 mm)
the lubricant. V.L. Cebolla would like to thank Dr. B.N. Barman
for the fruitful discussions on petrochemical analysis.
Conclusion

A weak interaction between the berberine cation and satu­ References

rated hydrocarbons allows them to be detected and quanti-
tated using TLC-fluorescence scanning densitometry, as was 1. W . R . G . Baeyens. Fluorescence and phosphorescence of phar­
demonstrated in a previous work. The proposed theoretical maceuticals, in Molecular Luminiscence Spectroscopy. Methods
explanation for this phenomenon predicts fluorescent response and Applications: Part 1, S.G. Schulman, Ed. (Vol. 77 in Chemical
for other analytes (e.g., alkyl-aromatics) with partially saturated Analysis, P.J. Elving and J . D . Winefordner, Eds.). Wiley Inter-
science, New York, NY, 1985, Chapter 2.2.
structure, adopting extended conformations along the main
2. L. Mamlok. Technical note: berberine hydrochloride for detection
axis of the berberine molecule. Therefore, direct application of (as a detector) in thin-layer chromatography. J. Chromatogr. Sci.
this phenomenon to petrochemical analysis has been achieved 19: 53 (1981).
in this work. Hydrocarbon types have been successfully deter­ 3. A.Y. Huc and J . G . Roucaché. Quantitative thin-layer chromatog­
mined in middle (gasoil) and heavy (lubricants, vis-breaking raphy of sedimentary organic matter. Anal. Chem. 53: 914-16
(1981).
fuel, heavy oil) petroleum distillates with adequate precision
4. F.P. Cossío, A. Arrieta, V.L. Cebolla, L. Membrado, M.P. Domingo,
and sensitivity. The novelty of this method resides on the fact P. Henrion, and J . Vela. Enhancement of fluorescence in thin-layer
that only a single fluorescence scanning is needed for the chromatography induced by the interaction between n-alkanes
determination of saturates and aromatics in the same chro- and an organic cation. Submitted for publication in Anal. Chem.
matogram. 5. Chromatography in the Petroleum Industry, E.R. Adlard, Ed. Else­
vier, Amsterdam, The Netherlands, 1995.
Quantitative results agree with those provided using other
6. V.L. Cebolla, L. Membrado, and J . Vela. Applications of liquid
well-established techniques: TLC-FID in the case of heavy dis­ chromatography to petroleum, in Encyclopedia of Separation
tillates and HPLC-RI, according to IP391 standard, in the case Science; I.D. Wilson, T.R. Adlard, C.F. Poole, and M . Cooke,
of gasoil. In this last case, it should be highlighted that our flu- Eds. Academic Press, London, England, in press.

225
 Journal of Chromatographic Science, Vol. 37, June 1998

7. A. Herold. Planar chromatography of coal products and oils. chromatography with U V scanning detection for quantitative
J. Planar Chromatogr. 7: 180-96 (1994). analysis of coal-derived products. J. Planar Chromatogr. Modern
8. T. Kalmal and A. Matsunaga. Determination of sulfur compounds TLC 11: 277-82 (1988).
in high-boiling petroleum distillates by ligand-exchange thin- 16. C.F. Poole and S.K. Poole. Multidimensionality in planar chro­
layer chromatography. Anal. Chem. 50: 268-70 (1978). matography. J. Chromatogr. A 703: 573-612 (1995).
9. T.G. Harvey, T.W. Matheson, and K.C. Pratt. Chemical class sep­ 17. S. Ebel. Quantitative analysis in TLC and HPTLC. J. Planar Chro­
aration of organics in shale oil by thin-layer chromatography. matogr. Modern TLC 9: 4-15 (1996).
Anal. Chem. 56: 1277-81 (1984). 18. Handbook of Thin Layer Chromatography, J . Sherma and B. Fried,
10. Y.Y. Wang and T.F. Yen. J. Planar Chromatogr. 3: 376-80 (1990). Eds. Marcel Dekker, New York, NY, 1991.
11. A.A. Herod and R. Kandiyoti. Fractionation by planar chro­ 19. J . Vela, V.L. Cebolla, L. Membrado, and J.M. Andrés. Quantitative
matography of a coal-tar pitch for characterization by size exclu­ hydrocarbon group type analysis of petroleum hydroconversion
sion chromatography, UV, fluorescence and direct-probe mass products using an improved TLC-FID system. J. Chromatogr. Sci.
spectrometry. J. Chromatogr. A 708: 143-60 (1995). 33: 417-25(1995).
12. V.R. Dhole and G.K. Ghosal. Adulteration of petrol with kerosene: 20. Standard Methods for Analysis and Testing of Petroleum and
detection and semiquantitative determination by TLC: part II. Related Products 1992, Vol. 2. Institute of Petroleum, London,

Downloaded from https://academic.oup.com/chromsci/article-abstract/37/6/219/315785 by guest on 22 March 2019

Planar Chromatogr. Modern TLC 8: 80-81 (1995). England, 1992.
13. B. Fried and J . Sherma. Thin-layer chromatography—techniques 21. ASTM D 2007, in Manual on Hydrocarbon Analysis, 4th ed.,
and applications. In Chromatographic Science, Vol. 35. Marcell A . W . Drews, Ed. American Society for Testing of Materials,
Dekker, New York, NY, 1986. Philadelphia, PA, 1989, ρ 353.
14. C.M. Marsh and C.J. Hiekane. Quantitative analysis of bitumen by 22. P. Kubelka and F. Munk. Ein Beitrag zur Optik Farbanstriche.
conventional high performance thin-layer chromatography. Z.Tech. Phys. 12: 593 (1931).
J. Planar Chromatogr. Modern TLC 4: 293-98 (1991).
15. J . Vela, V.L. Cebolla, L. Membrado, and A.C. Ferrando. Thin-layer Manuscript accepted April 30, 1999.

226