JPC – Journal of Planar Chromatography – Modern TLC (2023) 36:335–349

https://doi.org/10.1007/s00764-023-00256-x

ORIGINAL RESEARCH PAPER

Gradient high‑performance thin‑layer chromatography

for characterizing complex hydrocarbon‑containing products
Carmen Jarne1,2 · Vicente L. Cebolla1 · Luis Membrado1 · José M. Escuín1 · Jesús Vela2

Received: 7 August 2023 / Accepted: 29 September 2023 / Published online: 10 November 2023
© The Author(s) 2023

Abstract
Classic saturates, aromatics, resins, and asphaltenes (SARA) analysis has long enabled the chromatographic separation of
complex hydrocarbon-containing products into these four families. This simplification often does not provide sufficient
or correct information on the relationship between the composition of the products and the process variables studied. The
obtention of profiles by gradient automated multiple development (AMD) is an interesting alternative to obtain extended
SARA separations with increasing level of complexity for heavy petroleum products that adequately represent the entirety
of the sample. An optimized 20-step gradient based on tetrahydrofuran (THF)‒dichloromethane‒n-heptane over a total
migration distance of 83 mm is proposed here to characterize all types of heavy oil products. The combination of ultraviolet
(UV)–densitometry (at 201, 228, and 273 nm), fluorescence detection by intensity changes using berberine cation, together
with on-plate recording of UV spectra of separated peaks and the comparison with those of standards, allow for a deep
characterization of chemical families. In the apolar zone of the chromatogram saturates, naphthenes, naphteno-aromatics,
alkyl-aromatics, and aromatics with different degrees of condensation are detected. The use of THF, both in the plate pre-
cleaning stage and in the first stages of the gradient, is responsible for the satisfactory resolution of the resins in the polar
zone. AMD chromatographic zones can be related to peaks from hydrocarbon groups in classical SARA. The application of
this gradient to different types of products allowed for the relation of the profiles of bitumen to their origin, differentiation
of the resin part trapped in asphaltenes, control performance of deasphalting in hydrorefining products, and the control of
the composition of base oils.

Keywords Automated multiple development · High-performance thin-layer chromatography (HPTLC) · Hydrocarbons ·

Petroleum products · SARA analysis

1 Introduction has more than 1080 isomers. With 15 hexagons of benzene,

74,107,910 condensed structures can be built [2].
Heavy petroleum fuels consist of complex mixtures of thou- The most powerful technique for petroleum characteriza-
sands of high-boiling nonpolar and polar compounds that tion is Fourier-transform ion cyclotron resonance mass spec-
include different molecule structures and sizes, polarities, trometry (FT–ICR MS). Recently, the development of tem-
chemical families, and also metals [1]. The degree of com- porally stable high-field (> 7 T) high-homogeneity magnets
plexity of these products increases with their boiling point. has led to the development of ultrahigh-resolution FT–ICR
It is hard to imagine the number of compounds present. For MS. With a high mass resolving power (m/Δm50%) > 300,000
example, a long, saturated hydrocarbon such as ­C167H336 and sub-parts-per-million mass accuracy, FT–ICR MS is
able to characterize thousands of species in petroleum prod-
ucts without any separation step [3]. Although this approach
* Vicente L. Cebolla
[email protected] also presents some problems and not all fuel components
have been identified, its inherent high resolving power and
1
Agencia Estatal Consejo Superior de Investigaciones high mass accuracy allow for baseline resolution of closely
Científicas (CSIC), Instituto de Carboquímica, spaced isobaric species, as well as molecular formula assign-
50018 Saragossa, Spain
ment through accurate mass determination.
2
Departamento de Química Analítica, Universidad de
Zaragoza, 50009 Saragossa, Spain

13
Vol.:(0123456789)
336 JPC – Journal of Planar Chromatography – Modern TLC (2023) 36:335–349

However, this technique is not operational for quality con- limitations, quantification problems linked to the irrevers-
trol. A complete molecular identification of heavy petro- ible adsorption of a substantial part of Asph and Res into
leum products based using this technique is not feasible. the columns, a good part of which are not eluted. Moreover,
There is little equipment available for this technique and not detection of Sat using differential refractometry has a low
everyone can have the necessary equipment in their labora- sensitivity. The number of samples per day to be processed
tory. Likewise, the experiments and data analysis are very is necessarily low due to the inherent complexity associated
complex, expensive, and time-consuming. But, regardless with HPLC manipulation and calibration of these type of
of this, industry needs rapid and straightforward techniques samples.
for quality control. Although high-performance thin-layer chromatography
At present, fuels are characterized in terms of chemical (HPTLC) has been scarcely used for analyzing heavy petro-
family composition, using a simple classical hydrocarbon leum products, the inherent advantages of HPTLC, such as
group type analysis such as SARA [4, 5], the acronym of high sample throughput, minimal sample preparation, and
saturates (Sat), aromatics (Ar), resins (Res), and asphaltenes the presence of the whole sample on the layer, together with
(Asph). recent development of reliable instruments, make it a tech-
In petrochemistry, SARA has been useful for: obtain- nique with a future for the characterization of fuels, either
ing information on the conversion of raw material and the from petroleum or other renewable biosources (e.g., bio-
quality of the conversion product, evaluating the variables mass) [12, 13].
of conversion processes, elucidating reaction pathways and An important general problem to develop HPTLC-based
kinetics, linking the stability of the product to the composi- SARA methods was how to detect and quantify saturates
tion, diagnosing the source of plant upsets, and evaluating on the plate. Our group has been studying some phenom-
the effect of changes in composition on product performance ena related to fluorescence enhancements due to apolar
properties [4, 5]. molecules for some time [14, 15]. Specifically, we devel-
There are different techniques and methods developed for oped an HPTLC-based method for classic SARA of heavy
SARA, mostly based on chromatographic techniques. As petroleum products [16]. This method is based on a con-
the repartition of compounds in SARA groups is depending ventional sequential elution using three solvents: n-heptane
on the procedure, SARA must be performed by strictly fol- (C7), dichloromethane (DCM), and tetrahydrofuran (THF),
lowing the selected experimental conditions for a reliable in order of increasing polarity, to separate Sat, Ar, Res,
comparison of results. and Asph. In this way, Ar, Res, and Asph were detected by
However, a problem is that classical SARA has been UV‒densitometry. Saturates were detected and quantified
performed in the past by techniques that present important by induced fluorescence using berberine-impregnated plates.
practical limitations. The reference is ASTM D2007, an Long hydrocarbon chains of Saturates induce an increase
open-column chromatography method based on preferen- in berberine fluorescence through nonspecific noncovalent
tial adsorption of Asph and Res on clay, and Ar on silica gel ion–dipole interactions [14, 15]. In consequence, saturates
[6]. These fractions are recovered using appropriate solvents, are detected and quantified by fluorescence densitometry
and Sat are eluted using n-pentane. The separated fractions with regard to the fluorescent baseline provided by berberine
are determined gravimetrically after evaporation of solvents. fluorophore. This detection technique is suitable for non-
This method is laborious, time-consuming, not always pro- absorbing or poorly absorbing molecules, such as Sat.
viding pure fractions, and there is an incomplete recovery of Nevertheless, the main problem with the SARA analysis
them from the adsorbents. Likewise, it needs deasphalting is conceptual. The classic SARA scheme is severely flawed
of the starting product before the start. for its use as a predictive tool because it uses only four pseu-
Thin-layer chromatography with flame ionization detec- docomponents for a fuel. Therefore, it is not surprising that
tion (TLC‒FID, better known as Iatroscan) is carried out on on many occasions it does not provide sufficient or correct
10-cm silica gel glass sinterized chromarods, and is therefore information to correlate the chemical differences of the prod-
limited from the point of view of separation, sufficient for ucts with the variables studied. It would be necessary, and
a simple SARA analysis but insufficient to achieve a higher it is the objective of this paper, to develop detailed SARA
degree of resolution if necessary [7, 8]. Likewise, FID sup- methods to obtain separations with a higher degree of com-
posed quantitativity (constant response factor for the differ- plexity, with densitometric techniques that allow that the
ent separated groups) has been questioned [9, 10]. Moreover, product is detected in its entirety.
the sample is destroyed during FID detection. Another prob- Automated multiple development (AMD) is a fully auto-
lem is its limited availability in terms of instrumentation, mated technique for HPTLC development that combines
associated consumables, and repair. incremental multiple development solvent gradient in short
High-performance liquid chromatography (HPLC) steps (1–2 mm) and repetition of elemental steps [17]. This
was also used for SARA [11]. Its use has, among other provides band focusing and improves sample separation. The

13
JPC – Journal of Planar Chromatography – Modern TLC (2023) 36:335–349 337

use of AMD in combination with a robotic, spray-on, and (80%; [834884-59-2] CAS), violanthrone-79 ([85652-50-2]
fully automatic sample applicator contributes to improve the CAS), perylene-66 (40%; [MFCD07785507] MDL), violan-
repeatability and sensitivity. AMD is especially adequate for throne-78 ([82145-74-2] CAS) as condensed heterocyclic
samples that comprise a wide range of polarity and there- compounds. All standards were purchased from Sigma-
fore appears as an interesting alternative to obtain different Aldrich Inc. (St. Louis, MO, USA), except for perhydroflu-
separations with increasing level of complexity for heavy orene and androstane, which were purchased from Chiron
petroleum products with regard to those provided by simple AS (Trondheim, Norway).
SARA. In this paper, different possibilities of gradient sepa- Standards were dissolved in toluene.
rations using AMD were explored. Although the above listed standards do not necessarily
have to be in the studied products, we intend to cover the
presence of most types of compounds that may be present,
2 Experimental with regard to functional groups, aromaticity/hydroaromatic-
ity, heteroatomic content, and molecular weight. Combined
2.1 Materials data on their migration (m.d. or RF) and UV spectra on silica
gel can be found in the Supplementary Information.
Berberine hemisulfate (> 95%; CAS number: 633-66-9) was
from Sigma-Aldrich (Steinheim, Germany). HPLC-grade
n-heptane (C7, 99.1%), dichloromethane (DCM, 99.9%), 2.3 Chromatographic plates
methanol (MeOH, 99.9%), tetrahydrofuran (THF, 99.9%),
and toluene (99.7%) were from Panreac (Barcelona, Spain). HPTLC silica gel plates (10 cm × 20 cm, pore size 60 Å,
THF was used without stabilizer. It was passed through layer thickness 0.20 mm) were used from Merck (Darmstadt,
­ eSO4 indicator before
a column of activated carbon with a F Germany). They were precleaned by developing 90 mm with
doing the HPTLC analysis, to avoid the peroxide formation, THF in a vertical developing chamber.
and then was kept under nitrogen.

2.2 Samples and standards 2.4 Sample application

Samples include ten bitumen from different sources and, Samples were dissolved in THF and applied using the Auto-
within each source, obtained at different distillation tem- matic TLC Sampler 4 (CAMAG, Muttenz, Switzerland) as
peratures, and two Asphaltenes, three hydrorefined vacuum 4-mm bands. A total of 25 bands of samples were applied
gas oils, one heavy crude and its hydrorefined derivative, on the same plate with a distance of 2.9 mm between tracks.
and one base oil. They all were provided by TOTAL RM Two additional tracks were always kept free of application
(France). as blank runs. The first application position was 10 mm
The following model compounds were used as migra- (x coordinate), and the distance from lower edge of plate
tion distance (m.d.) qualitative standards: n-tetracosane was 10 mm (y coordinate). Typical application volumes for
­C24 (≥ 99%; [646-31-1] CAS), n-octacosane ­C28 (≥ 98%; petroleum samples were between 0.2 and 0.8 μL. Typical
[630-02-4] CAS), and n-dotriacontane ­C32 (≥ 96%; [544- effective masses applied were between 0.5 and 3 μg for UV,
85-4] CAS) as paraffins; 1-octadecene (≥ 99.5%; [112-88- and between 3 and 10 μg for fluorescence evaluation.
9] CAS) and cis-9-tricosene (97%; [27519-02-4] CAS) as
olefins; ethyl-benzene (99.8%; [100-41-4] CAS), pentyl-
benzene (99%; [538-68-1] CAS), and 2-ethylnaphthalene 2.5 Chromatography
(≥ 99%; [939-27-5] CAS) as alkyl-aromatics; cis-decahy-
dronaphthalene (99%; [439-01-6] CAS), 1,2,3,4-tetrahy- Chromatographic development was performed with an
dronaphthalene (99%; [119-64-2] CAS), 9,10- dihydro- AMD2 system (CAMAG) using different conditions. Gra-
phenanthrene (≥ 94%; [776-35-2] CAS), perhydrofluorene dient composition in each step, number of steps, distance
([5744-03-6] CAS), and 5-α (H)-androstane ([438-22-2] per step (mm s­ tep‒1), and total developing distance (mm)
CAS) as naphthenes; naphthalene (≥ 99.7%; [91-20-3] are detailed in Tables 1, 2, 3, and 4: a three solvent (C7‒
CAS), phenanthrene (≥ 98%; [85-01-8] CAS), anthracene DCM‒THF) classic SARA sequence (Table 1), a 23-step,
(≥ 99%; [120-12-7] CAS), pyrene (≥ 99%; [129-00-0] CAS), THF‒DCM‒C7 gradient (Table 2), a 20-step, THF‒DCM‒
benzo[k]fluoranthene (98%; [207-08-9] CAS), benzo[e] C7 gradient (Table 3), and a 20-step DCM‒C7 gradient
pyrene (99%; [192-97-2] CAS), benzo[ghi]perylene (98%; (Table 4). Migration of standards under different conditions
[191-24-2] CAS), and coronene (99%; [191-07-1] CAS) and their recorded on-silica gel UV spectra can be found in
as condensed hydro- and polyaromatics; and quinolin-65 the Supplementary Information.

13
338 JPC – Journal of Planar Chromatography – Modern TLC (2023) 36:335–349

Table 1  AMD conditions for classic SARA (sequential) Table 3  AMD conditions for 20-step THF‒DCM‒C7 gradient
THF (vol%) DCM (vol%) C7 (vol%) Migra- THF (vol%) DCM THF C7 (vol%) Migration
tion (mm)
(mm)
Step 1 100 0 0 14
Step 1 0 0 100 60 Step 2 100 0 0 18
Step 2 0 100 0 40 Step 3 0 100 0 22
Step 3 100 0 0 20 Step 4 0 100 0 26
Step 5 0 80 20 30
Step 6 0 80 20 34
Step 7 0 60 40 38
Table 2  AMD conditions for 23-step THF‒DCM‒C7 gradient
Step 8 0 60 40 42
THF (vol%) DCM (vol%) C7 (vol%) Migration (mm) Step 9 0 40 60 46
Step 1 100 0 0 14 Step 10 0 40 60 50
Step 2 100 0 0 18 Step 11 0 20 80 54
Step 3 0 100 0 22 Step 12 0 20 80 58
Step 4 0 100 0 26 Step 13 0 15 85 62
Step 5 0 80 20 30 Step 14 0 0 100 65
Step 6 0 80 20 34 Step 15 0 0 100 68
Step 7 0 60 40 38 Step 16 0 0 100 71
Step 8 0 60 40 42 Step 17 0 0 100 74
Step 9 0 40 60 46 Step 18 0 0 100 77
Step 10 0 40 60 50 Step 19 0 0 100 80
Step 11 0 20 80 54 Step 20 0 0 100 83
Step 12 0 20 80 58
Step 13 0 15 85 62
Step 14 0 10 90 64
Step 15 0 8 92 66
Step 16 0 6 94 68
Step 17 0 4 96 70
Step 18 0 2 98 72 Table 4  AMD conditions for 20-step DCM‒C7 gradient
Step 19 0 1 99 74 DCM (vol%) C7 (vol%) Migration (mm)
Step 20 0 0 100 76
Step 21 0 0 100 78 Step 1 100 0 14
Step 22 0 0 100 80 Step 2 100 0 18
Step 23 0 0 100 82 Step 3 100 0 22
Step 4 100 0 26
Step 5 80 20 30
Step 6 80 20 34
2.6 UV and fluorescence densitometry Step 7 60 40 38
Step 8 60 40 42
A TLC Scanner 3 (CAMAG) was used in UV and fluores- Step 9 40 60 46
cence modes. First, UV‒densitometry at different wave- Step 10 40 60 50
lengths (201, 228, and 273 nm) was recorded, as well as Step 11 20 80 54
UV spectra of separated peaks. Subsequently, HPTLC Step 12 20 80 58
silica gel plates were postimpregnated with a 60 mg ­L−1 Step 13 15 85 62
solution of berberine hemi-sulphate in MeOH, using a Step 14 0 100 65
CAMAG Chromatogram Immersion Device 3. Fluores- Step 15 0 100 68
cence mode was then used for peak detection by excitation Step 16 0 100 71
at 365 nm. Emission was collected at longer wavelengths Step 17 0 100 74
than 400 nm. This detection was referred to as fluores- Step 18 0 100 77
cence detection by intensity changes (FDIC). Step 19 0 100 80
Step 20 0 100 83

13
JPC – Journal of Planar Chromatography – Modern TLC (2023) 36:335–349 339

3 Results and discussion first separated (with C7 at 60 mm m.d.), then aromatics

(Ar, with DCM at 40 mm m.d.), resins (Res, with THF at
3.1 Classic SARA and SARA expanded by gradient 20 mm), and asphaltenes (Asph, non-developed peak at the
application point). Sat were detected after plate impreg-
Figure 1A and B show a classic SARA chromatogram of nation by berberine-induced fluorescence (λexc = 365 nm;
a Canadian bitumen, using a sequence of several elution λem > 400 nm) (Fig. 1B). Ar, Res, and Asph were detected
solvents in order of increasing polarity. In this case, the by UV at 273 nm, the wavelength that provided maximum
AMD instrument was used as a conventional develop- absorption for each of these groups (Fig. 1A).
ing chamber. As detailed in Table 1, saturates (Sat) were When a 23-step AMD gradient based on the same eluents
(THF‒DCM‒C7, see conditions in Table 2) was applied

Fig. 1  Classic SARA chromatogram of a Canadian bitumen: A UV 273 nm and B berberine-induced fluorescence 365 nm. A 23-step AMD gra-
dient of a Canadian bitumen: C UV 273 nm and D FDIC 365 nm

13
340 JPC – Journal of Planar Chromatography – Modern TLC (2023) 36:335–349

Fig. 2  A Chromatogram of a vacuum gas oil at different wavelengths: under the AMD conditions (Table 3, 20 steps); C zoom of apolar
UV 273 nm (–), UV 228 nm ( --- ), UV 201 nm ( --- ), and FDIC zone (55‒85 mm)
365 nm ( --- ) under the AMD conditions (Table 2, 23 steps), and B

on the same sample, the resulting chromatogram is shown combination with increasing developing distance. The compo-
in Fig. 1C and D. As we can observe, the simplicity of the sition of the mobile phase in each step does not have to be the
classical SARA chromatogram (Fig. 1A) is deceptive. The same and usually is of lower polarity that than of the previous
23-stage SARA (Fig. 1C) gives us a clearer idea of the com- one, adjusted by mixing of different proportions of the three
plexity of the sample. selected solvents. The stepwise gradient formed allows narrow
AMD operates in successive elemental elution steps peaks and more efficient separations to be obtained in a highly
(of several millimeters each), which are performed with reproducible way (± 0.45 mm). Repetitions of elemental steps
decreasing solvent strength, in our case THF > DCM > C7, in (twice) were performed to provide band refocusing.

13
JPC – Journal of Planar Chromatography – Modern TLC (2023) 36:335–349 341

3.2 Gradient optimization: from 23‑step SARA conditions. UV spectra of several standards are shown in the
to 20‑step SARA​ Supplementary Information.
The most polar and heaviest compounds (Asph) remain
Figure 2 shows the 23-step chromatogram (Table 2) of a at the application point (m.d. = 10 mm) and are directly
hydrorefined vacuum gas oil at different UV wavelengths detected on the plate by UV‒densitometry. Resins zone
and FDIC. AMD chromatographic zones can be related to (18‒58 mm m.d. zone, 10 steps, from 100 vol% DCM to
peaks from hydrocarbon groups in classical SARA. The 20:80, V/V, DCM‒C7) is particularly well resolved using
peak in the gradient chromatogram appearing in the 100 the described gradient conditions. As we will show below,
vol% THF development zone (up to 14 mm m.d.; one step) the role of THF is crucial to achieve an adequate resolution
can be considered as asphaltenes (non-developed by THF). of this zone.
The peaks eluted in the 18‒58 mm m.d. zone (10 steps, from Peaks eluted in the 58‒83 mm zone, seven steps at 100
100 vol% DCM to 20:80, V/V, DCM‒heptane), correspond vol% C7, are Aromatics and Saturates. This zone is espe-
to resins. This zone is particularly well resolved using the cially interesting and rich in aromatic and saturated com-
described gradient conditions and may provide more use- pounds in hydro-refining products such as this one, as can
ful information than that obtained using classical methods. be seen by comparing the profile of this product with that of
The peaks eluted in the 58‒83 mm zone, from 15:85, V/V, a bitumen in Fig. 1B.
DCM‒heptane to 100 vol% n-heptane, are usually consid- Figure 2C shows a detailed view of different chemical
ered aromatics and saturates. families in this zone of the chromatogram where several
An inefficient zone with regard to separation (57–67 mm, peaks are coeluted. Identification of component groups has
m.d.) was identified in this chromatogram. As in general been possible by using selective detection, direct recording
AMD practice, an unresolved chromatogram can be super- of UV spectrum of each separated peak, and comparing UV
imposed with the corresponding diagram of the gradient spectra of peaks with those of standards that were eluted on
employed, changes in solvent composition can be made, and the plate under the same gradient conditions.
a number of steps can be done to improve gradient separa- The peak eluted at 83.3 mm with the front of solvent
tion. In this way, the gradient conditions were modified to is due to saturated hydrocarbons. This peak has no UV
20 steps (see Fig. 2B and conditions in Table 3) to optimize response and is detected by FDIC‒berberine [5]. An ion-
separation zones on the plate. Thus, separation can be fine- induced dipole interaction between n-alkanes and berberine
tuned and amplified for each segment of the chromatogram. cation is responsible for an increase in berberine emission
In our case, the separation in the apolar zone was improved in the presence of alkanes [5]. The shoulder at 81 mm in the
without losing resolution in the asphaltenes‒resins zone and FDIC‒berberine chromatogram, which coelutes with satu-
by removing the inefficient plate zone. To do this, steps 12 rated hydrocarbons and has a response in the presence of
to 17, which included DCM‒C7 gradient, were removed, berberine, may be attributed to naphthenes. This shoulder is
and seven final steps of C7 (100%) were added. In total, the encompassed in the peak eluted at 80.3 mm, which has been
sequence went from having 23 to 20 steps. Details of both selectively detected by UV at 201 nm. UV spectrum sug-
gradients are described in Tables 2 and 3. gests that this peak consists of naphthenic and naphthene‒
The saving of three steps in the chromatographic develop- aromatic structures.
ment involved going from 5 h and 20 min to 5 h of analysis The inspection of the chromatogram obtained by UV
time. Regardless of the final number of steps and despite the detection at 228 nm, and the UV spectrum of the peak eluted
duration of these analyses, the number of substances pro- at 78.2 mm, are coherent with the presence of alkyl-aromatic
cessed per plate compensates the time invested with regard structures in this peak. Likewise, the chromatogram obtained
to other chromatographic techniques. The AMD instrument by UV at 273 nm together with UV spectra of peaks eluted
can be left programmed to perform the analysis overnight at 73.2, 68.5, and 64.3 mm suggest the presence of aromatic
in automatic mode. structures, probably with different degrees of condensation.
Migration distances of standards under these conditions and
3.3 Heavy product characterization using the corresponding UV spectra are given in Supplementary
optimized 20‑step THF‒DCM‒C7 SARA Information.
gradient Therefore, this optimized 20-step THF‒DCM‒C7 gradi-
ent can be used for profiling a variety of samples that covers
Figure 2B shows the chromatogram of the mentioned the whole range of petroleum heavy products.
hydrorefined vacuum gas oil under these AMD conditions
at different UV wavelengths and FDIC. On-silica UV spec-
tra of standards are similar to those obtained in dissolution
and provide useful information for selecting densitometric

13
342 JPC – Journal of Planar Chromatography – Modern TLC (2023) 36:335–349

Fig. 3  A Bitumen Norooz +420 °C, B Bitumen Norooz +450 °C, +440 °C, H Lokele +470 °C, and I Lokele +490 °C. 20-Step AMD
C Bitumen Norooz +490 °C, D Oguendjo/Olende +495 °C, E gradient, UV 273 nm
Oguendjo/Olende +520 °C, F Oguendjo/Olende +560 °C, G Lokele

3.4 Applications of expanded SARA gradient for evaluating the effect of process variables on products.
We show here some specific applications to different heavy
The application of multi-step gradient sequence to different products.
heavy products provides qualitatively different chromato-
grams among products. Although this separation does not 3.4.1 Bitumen
provide a complete molecular resolution of heavy petroleum
products, the resulting profiles adequately represent the com- Bitumen is a black, viscous mixture of hydrocarbons
plexity of the analyte, and can provide a useful information obtained naturally or as a residue from petroleum distillation

13
JPC – Journal of Planar Chromatography – Modern TLC (2023) 36:335–349 343

[18]. It is is too thick to be pumped from the ground or application (in Iatroscan and in HPTLC) [13], or with the
through pipelines. part that remains irreversibly retained in the stationary
Different qualitative profiles were obtained by application phase of chromatographic columns (in the case of HPLC
of 20-step expanded SARA gradient (THF‒DCM‒C7) in and open column chromatography, as in the D2007 stand-
the function of their origin, regardless of their distillation cut ard [6]).
temperature (DCT; in degrees celsius, °C). Figure 3 shows When Asph are gravimetrically isolated they tend
different profiles for bitumen from Iran (Norooz, DCT at to “catch” some Res, structures that resemble those of
+420, +450, and +490 °C), Gabon (Oguendjo/Olende, DCT Asph with a lower molecular weight and that are strongly
at +495, +520, and +560 °C), and Cameroon (Lokele, DCT attracted by them. Figure 4 shows that the 23-step gradi-
at +440, +470, and +490 °C). ent reveals different “resinic part” in two different types
It seems possible to obtain fingerprints for bitumen of Asph. This gradient is also useful to detect trapped Sat
characterization by HPTLC in a similar way to those in Asph structures (not shown).
obtained in the case of characterization of plants, therefore
relating bitumen composition and properties. 3.4.3 Hydrorefining products

Hydrorefining is a process for improving the quality of a

3.4.2 Asphaltenes petroleum product by hydrogenating in the presence of a
catalyst at a temperature below that at which decomposi-
The structure of asphaltenes is still a challenge for ana- tion takes place [20]. This improves the H/C ratio of the
lytical chemistry. From the point of view of solubility, product. Thus, a consequence of hydrorefining, deasphalt-
asphaltenes are insoluble in C7 and soluble in toluene. ing is observed (a decrease in the Asph peak), and also an
This is the way to obtain them gravimetrically. Recent increase in the Saturates fraction. This can be observed in
research suggests that they are “hand” shaped entities, Fig. 5 where the 23-step gradient chromatograms of a heavy
with a central core consisting of a hydro/polyaromatic sys- Canadian oil and its derived hydrorefined product are shown.
tem (for example, around seven cores with a few N and S
heteroatoms) with peripheral alkanes [19]. The molecular 3.4.4 Base oils
masses of the asphaltenes are between 300 and 1400 Da,
with the majority of species between 400 and 800 Da. Base oils, produced from refining crude oil, are used in
Chromatographically, asphaltenes are associated the formulation of lubricant products. They are complex
with the undeveloped peak that remains at the point of mixtures containing hydrocarbon families of very different

Fig. 4  A, B Asphaltenes from different origins. 23-Step AMD gradient, UV 273 nm

13
344 JPC – Journal of Planar Chromatography – Modern TLC (2023) 36:335–349

Fig. 5  The 23-step AMD gradient of heavy Canadian oil: A UV 273 nm, B FDIC 365 nm. Derived hydrorefined of heavy Canadian product: C
UV 273 nm, D FDIC 365 nm

polarity, such as paraffins, isoparaffins, naphthenes (cyclo- detail. This zone corresponds to the last seven stages of the
paraffins), hydroaromatics, aromatics with different conden- gradient, with a composition 100% C7, according to Table 3.
sation degrees, and a variety of S-, O-, and N-heterocyclic The families in base oil D were detected by UV detection
structures, with molecular weight distributions ranging from at 201, 228, and 273 nm and FDIC, and UV spectra of the
300 to 600 Da, although in general, they are asphaltene-free separate peaks were consistent with those of the model com-
and have a lower polarity than other heavy products [21]. pounds studied (Supplementary Information).
Base oils only have compounds in the apolar zone of the Although aromatic compounds present some coelution
20-step chromatogram (Fig. 6). in this zone of chromatogram, a selective UV detection
Under the 20-step AMD conditions, it is possible to dis- at different wavelengths and the UV spectra of obtained
tinguish the families of naphthenes, naphthenoaromatics, peaks allowed us to identify several types of compounds
and alkylaromatics between 75 and 83 mm, with a degree of in base oil D: saturates (m.d. = 84.3 mm), naphthenes

13
JPC – Journal of Planar Chromatography – Modern TLC (2023) 36:335–349 345

system, i.e., using DCM‒C7 gradient, and DCM was the

washing solvent, resolved peaks disappeared, and resins
compounds were eluted to the apolar part of the chromato-
gram (Fig. 7A). Using this same gradient, resolved peaks in
the resins zone are present when THF has been used in the
preconditioning, washing step (Fig. 7B).
Figure 7C shows that a low-resolution chromatogram,
similar to that of Fig. 7A, is obtained when THF is used
in the washing step but is removed from the plate heating
(heating for 30 min at 105 °C) before sample application.
Likewise, Fig. 7D gives an idea of the importance of the
precleaning operation. When no washing step is included,
an intense peak of impurities is eluted by the gradient at a
migration distance of 72 mm.
Results obtained from the 20-step THF‒DCM‒C7
gradient from the same vacuum oil sample are given in
Fig. 8A–D. In all cases, resolved chromatographic profiles
were obtained, with an adequate degree of resolution for the
compounds eluted in the resins zone when THF was present
Fig. 6  The 20-Step AMD gradient of base oil: (–): UV 273 nm, ( --- ): either in the gradient mobile phase or as a washing solvent.
UV 228 nm, ( --- ): UV 201 nm ( --- ): FDIC 365 nm The effect of washing with DCM (Fig. 8A) or THF
(Fig. 8B) gives a different distribution proportion of devel-
(m.d. = 82.5 mm, a shoulder in saturates peak), heavy naph- oped compounds. Likewise, the removal of THF by heating
thene/naphthene-aromatics (m.d. = 80.4 mm), and alkyl-aro- (Fig. 8C) shows a chromatogram similar to that of Fig. 8A
matics (m.d. = 78.3 mm). In the case of the peak at 80.4 mm, (with DCM as the precleaning solvent). Moreover, Fig. 8D
there would be a greater contribution from heavy naphthenes shows that using the plate with previous cleaning provided
than from aromatics, since little absorbance is observed at different repartition and proportion of resins onto the silica
273 nm. gel plate.
It is well known that THF adsorbs on silica gel surfaces,
3.5 The role of THF in the separation and that distinct mechanisms of adsorption (nonspecific,
specific, molecular, and chemical adsorption) were observed
We have found that THF plays a crucial role in the degree [22]. Likewise, it has been found that adsorption isotherms
of resolution of extended separations of heavy petroleum of polar solvents on silica gel does fit abi-layer adsorption
products, and particularly in the good separation obtained equation rather than the mono-layer one [23]. According
in the resins zone. to Scott [23], such layers can impede the interaction of the
When THF is used either as a plate precleaning solvent solute directly with the stationary phase as in sorption.
in preconditioning step, or as a mobile phase in the gradi- A combined and interactive mechanism of sorption and
ent, good resolutions are obtained in the resulting chroma- displacement was proposed by Scott when the silica surface
tograms. On the contrary, when THF is neither used as a was covered with a layer of molecules from one or more
washing solvent nor as a mobile phase, profiles completely different solvents (Fig. 9).
lose their resolution. In our case, THF is the solvent that interacts directly with
For evaluating the effect of THF as a mobile phase in the the silica surface in the precleaning step, and consequently
gradient, a 20-step DCM‒C7 gradient was used (Table 4) to constitutes the first layer. Thus, elution began using pure
compare with the gradient containing THF (Table 3). There- THF, and the next layer was a second layer of THF. Thus,
fore, the first two steps eluted with THF were changed to a monolayer or bilayer of THF may be formed on the silica
two steps eluted with DCM. Moreover, the effect of DCM surface with which the solute interacts by sorption (mecha-
as plate preconditioning solvent was compared with that of nisms a and b). In general, the stronger the polarity of the
THF. The sample studied was a vacuum gas oil. solute, the more likely it is to interact with the surface by
Figure 7 summarizes the results obtained for DCM‒C7 displacement (mechanism c), even to the extent of displac-
gradient. When THF was not present in the chromatographic ing both layers of THF (mechanism d). This influences the
separation of the resins in the petroleum products studied.

13
346 JPC – Journal of Planar Chromatography – Modern TLC (2023) 36:335–349

Fig. 7  Vacuum gas oil, 20-step DCM‒C7 gradient: A DCM as washing solvent, B THF as washing solvent, C THF as washing solvent removed
by plate heating for 30 min at 105 °C, D without washing step

4 Conclusion migration distance of 83 mm. Each step is repeated twice

for refocusing the HPTLC bands. UV detection at several
Extended SARA profiles (20 and 23 AMD steps) of any selected wavelengths and berberine-induced fluorescence,
type of heavy petroleum product were obtained under combined with on-silica recording of UV spectra of sepa-
fixed conditions by AMD gradient on a silica gel HPTLC rated peaks and standards, allowed obtaining qualitative
plate. These gradients are based on 20 (optimal) or 23 steps information of interest on the composition of these heavy
with different proportions of C7‒DCM‒THF over a total petroleum products (bitumen, base oils, hydrorefining

13
JPC – Journal of Planar Chromatography – Modern TLC (2023) 36:335–349 347

Fig. 8  Vacuum gas oil/hydrorefined product, 20-step THF‒DCM‒n-heptane gradient: A DCM as washing solvent, B THF as washing solvent,
C THF as washing solvent (plate heating for 30 min at 105 °C), D without washing step

products, asphaltenes, etc.) and the variables involved in In addition to the inherent advantages of HPTLC (e.g.,
the corresponding processes. complete detection of the sample avoiding column irre-
Expansion of separation of selected zones of the chro- versible adsorptions), a high number of samples on the
matograms can be achieved by simple modification of same plate were processed, under strictly identical condi-
AMD development conditions. The use of THF, coming tions, with minimal solvent requirements, reduced labor
from the plate precleaning step or from the AMD gradient times, and important cost savings.
is crucial to improve separation in the resins zone.

13
348 JPC – Journal of Planar Chromatography – Modern TLC (2023) 36:335–349

Asphaltenes, heavy oils, and petroleomics. Springer, New York.

https://​doi.​org/​10.​1007/0-​387-​68903-6_3
4. Barman BN, Cebolla VL, Membrado L (2000) Chromatographic
techniques for petroleum and related products. Crit Rev Anal
Chem 30:75–120. https://​doi.​org/​10.​1080/​10408​34009​11641​99
5. Hammami A, Ratulowski J (2007) Precipitation and deposition
of Asphaltenes in production systems: a flow assurance over-
view. In: Mullins OC, Sheu EY, Hammami A, Marshall AG (eds)
Asphaltenes, heavy oils, and petroleomics. Springer, New York.
https://​doi.​org/​10.​1007/0-​387-​68903-6_​23
6. ASTM D2007–19 (2020) Standard test method for characteristic
groups in rubber extender and processing oils and other petro-
leum-derived oils by the clay-gel absorption chromatographic
Fig. 9  Mechanism of sorption and displacement of THF on silica gel method. American Society for Testing and Materials, West
surface ( O: Solute and O: THF): a solute interacting with first layer Conshohocken
7. Ranny M (1987) Thin-layer chromatography with flame ionization
of THF by sorption, b solute interacting with second layer of THF by detection. D. Reidel Publishing Company, Dordrecht
sorption, c solute interacting with first THF layer by displacing THF 8. Barman BN (1996) Hydrocarbon-type analysis of base oils and
from second layer, d solute interacting with silica surface by displac- other heavy distillates by thin-layer chromatography with flame-
ing THF from second layer lonization detection and by the clay-gel method. J Chromatogr Sci
34:219–225. https://​doi.​org/​10.​1093/​chrom​sci/​34.5.​219
9. Vela J, Membrado L, Cebolla VL, Ferrando AC (1998) Suitabil-
Supplementary Information The online version contains supplemen-
ity of thin-layer chromatography-flame ionization detection with
tary material available at https://d​ oi.o​ rg/1​ 0.1​ 007/s​ 00764-0​ 23-0​ 0256-x.
regard to quantitative characterization of different fossil fuel prod-
ucts. II. Calibration methods concerning quantitative group-type
Acknowledgements Authors thank TOTAL RM (France) for financial
analysis. J Chromatogr Sci 36:487–493. https://​doi.​org/​10.​1093/​
support for this project.
chrom​sci/​36.​10.​487
10. Jiang CQ, Larter SR, Noke KJ, Snowdon LR (2008) TLC–FID
Funding Open Access funding provided thanks to the CRUE-CSIC
(Iatroscan) analysis of heavy oil and tar sand simples. Org Geo-
agreement with Springer Nature.
chem 39:1210–1214. https://​doi.​org/​10.​1016/j.​orgge​ochem.​2008.​
01.​013
Declarations 11. Pei P, Britton J Jr, Hsu S (1983) Hydrocarbon type separation of
lubricating base oil in multigram quantity by preparative HPLC.
Conflict of interest The authors declare no conflict of interest. The sec- J Liq Chromatogr 6:627–645. https://d​ oi.o​ rg/1​ 0.1​ 080/0​ 14839​ 1830​
ond author, V.L.C., is a member of the editorial board of the journal. 80760​73
Therefore, the submission was handled by a different member of the 12. Cebolla VL, Membrado L, Matt M, Gálvez EM, Domingo MP
editorial board, and V.L.C. did not take part in the review process in (2002) Thin-layer chromatography for hydrocarbon charac-
any capacity. terization in petroleum middle distillates. In: Hsu CS (ed) Ana-
lytical advances for hydrocarbon research. Kluwer Academic/
Open Access This article is licensed under a Creative Commons Attri- Plenum Press, New York, pp 95–112. https://​doi.​org/​10.​1007/​
bution 4.0 International License, which permits use, sharing, adapta- 978-1-​4419-​9212-3_5
tion, distribution and reproduction in any medium or format, as long 13. Matt M, Gálvez EM, Cebolla VL, Membrado L, Bacaud R, Pessa-
as you give appropriate credit to the original author(s) and the source, yre S (2003) Improved separation and quantitative determination
provide a link to the Creative Commons licence, and indicate if changes of hydrocarbon types in gas oil by normal phase high-performance
were made. The images or other third party material in this article are TLC with UV and fluorescence scanning densitometry. J Sep Sci
included in the article’s Creative Commons licence, unless indicated 26:1665–1674. https://​doi.​org/​10.​1002/​jssc.​20030​1465
otherwise in a credit line to the material. If material is not included in 14. Cossío FP, Arrieta A, Cebolla VL, Membrado L, Domingo
the article’s Creative Commons licence and your intended use is not MP, Henrion P, Vela J (2000) Enhancement of fluorescence in
permitted by statutory regulation or exceeds the permitted use, you will thin-layer chromatography induced by the interaction between
need to obtain permission directly from the copyright holder. To view a n-alkanes and an organic cation. Anal Chem 72:1759–1766.
copy of this licence, visit http://creativecommons.org/licenses/by/4.0/. https://​doi.​org/​10.​1021/​ac991​302q
15. Cebolla VL, Mateos E, Garriga R, Membrado L, Cossío FP,
Gálvez EM, Matt M, Delgado-Camón A (2012) Changes in fluo-
rescent emission due to non-covalent interactions as a general
References detection procedure for thin-layer chromatography. ChemPhy-
sChem 13:291–299. https://​doi.​org/​10.​1002/​cphc.​20110​0590
1. Altgelt KH, Boduszinski M (2014) Composition and analysis of 16. CAMAG Application Note A-101.1. Determination of SARA
heavy petroleum fractions. Taylor & Francis, Boca Raton (Saturates, Aromatics, Resins and Aaphaltenes) in bitumen by
2. Elguero J (2021) Algunas reflexiones sobre el futuro de la quimica HPTLC. Muttenz
computacional. An Quím 117(3):203–208 17. Membrado L, Cebolla VL, Jarne C, Garriga R, Bernard-Savary P,
3. Rodgers RP, Marshall AG (2007) Petroleomics: advanced char- Vela J (2023) Automated multiple development. In: Poole CF (ed)
acterization of petroleum-derived materials by Fourier Trans- Instrumental thin-layer chromatography. Handbook of separation
form Ion Cyclotron Resonance Mass Spectrometry (FT-ICR science, 2nd edn. Elsevier, Amsterdam, pp 81–110. https://​doi.​
MS). In: Mullins OC, Sheu EY, Hammami A, Marshall AG (eds) org/​10.​1016/​C2021-0-​01004-4

13
JPC – Journal of Planar Chromatography – Modern TLC (2023) 36:335–349 349

18. Holýa M, Remisováa E (2019) Analysis of influence of bitumen 21. Abdul Jameel AG, Sarathy SM (2018) Lube products: molecular
composition on the properties represented by empirical and vis- characterization of base oils. Petroleum and liquid fossil fuels
cosity test. Transp Res Procedia 40:34–41. https://​doi.​org/​10.​ analysis. Encyclopedia of analytical chemistry. Wiley, Hoboken
1016/j.​trpro.​2019.​07.​007 22. Merkku P, Yliruusi J, Vuorela H, Hiltunen R (1994) Determi-
19. Sharma A, Mullins OC (2007) Insights into molecular and aggre- nation of specific heats and adsorption energies of ternary TLC
gate structures of Aaphaltenes using HRTEM. In: Mullins OC, eluents and silica gel. J Planar Chromatogr 7:305–308
Sheu EY, Hammami A, Marshall AG (eds) Aaphaltenes, heavy 23. Scott RPW (1980) The silica gel surface and its interactions with
oils, and petroleomics. Springer, New York. https://​doi.​org/​10.​ solvent and solute in liquid chromatography. J Chromatogr Sci
1007/0-​387-​68903-6_8 18:297–306. https://​doi.​org/​10.​1093/​chrom​sci/​18.7.​297
20. Fahim MA, Alsahhaf TA, Elkilani A (2010) Hydroconversion.
In: Fahim MA, Alsahhaf TA, Elkilani A (eds) Fundamentals of
petroleum refining. Elsevier, Amsterdam, pp 153–198

13