Fuel 108 (2013) 647–655

Contents lists available at SciVerse ScienceDirect

Fuel
journal homepage: www.elsevier.com/locate/fuel

Monitoring the liquid/liquid extraction of naphthenic acids in brazilian crude

oil using electrospray ionization FT-ICR mass spectrometry (ESI FT-ICR MS)
Keroly A.P. Colati a, Guilherme P. Dalmaschio a, Eustáquio V.R. de Castro a, Alexandre O. Gomes b,
Boniek G. Vaz c,⇑, Wanderson Romão d,⇑
a
Department of Chemistry, Federal University of Espírito Santo, 29075-910 Vitória, ES, Brazil
b
Petróleo Brasileiro S/A – PETROBRAS, CENPES, 21941-915 Rio de Janeiro, RJ, Brazil
c
Chemistry Institute, Federal University of Goiás, 74001-970 Goiânia, GO, Brazil
d
Federal Institute of Education, Science and Technology of Espírito Santo, 29106-010 Vila Velha, ES, Brazil

h i g h l i g h t s

" Non-continuum liquid–liquid extraction of naphthenic acids at different pHs from two crude oil samples.
" The residue and the acid fraction were characterized using ESI() FT-ICR MS, FTIR and sulfur analysis.
" We observe a decreasing TAN in the original oil with increasing pH (7 ? 14), reaching a value of almost zero at pH 14.
" The most abundant naphthenic acids extracted are those with short alkyl chain lengths (<C44) and DBE = 3–4.

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

Article history: Although the term ‘‘naphthenic acids’’ was originally used to describe acids that contain naphthenic rings,
Received 6 November 2012 today this term is used in a more general sense and refers to all cyclic, acyclic, and aromatic acids in crude
Received in revised form 3 February 2013 oil. In crude oil, naphthenic acids exist as a complex mixture of compounds with broad polydispersity
Accepted 4 February 2013
with respect to both molecular weight and structure. Recently, there has been increasing interest in
Available online 26 February 2013
acidic fractions in crude oil because of the corrosion problems that these compounds cause during oil
refinery. This corrosion is associated with the total acid number (TAN). However, it has been argued that
Keywords:
there is no clear correlation between the TAN and the level of corrosion. Herein, naphthenic acids were
Petroleomics
Naphthenic acid
extracted from two crude oil samples (TAN = 4.95 and 3.19 mg KOH g1) using liquid/liquid extraction
Extraction process with alkaline solutions at three different pHs (pH 7, 10 and 14), thus evaluating the efficiency and selec-
Mass spectrometry tivity among different acidic extraction methods to prevent future corrosion processes in petroleum
FT-ICR MS industry such as regions of the refineries working. The original oil samples and their naphthenic acidic
fractions and residues (washed oil from the acidic extraction) were analyzed using negative-ion electro-
spray ionization (ESI) Fourier transform ion cyclotron mass spectrometry (FT-ICR MS) and Fourier trans-
form infrared spectroscopy (FTIR). The TAN and sulfur content were also determined. We observe a
decreasing TAN in the original oil with increasing pH (7 ? 14), reaching a value of almost zero at pH
14. Consequently, an intense band at approximately 1700 cm1 was observed for the naphthenic acidic
fraction. Fractions produced ESI()-FT-ICR MS spectra with average molecular weight distribution, Mw,
and m/z values ranging from m/z 200–350 and Mw = 270 (for pH 7) to m/z 200–650 and Mw = 390 (for
pH 14). Therefore, the acidic extraction method at pH 14 was more efficient and the ESI FT-ICR MS anal-
ysis reveals that the most abundant naphthenic acids are those with short alkyl chain lengths (<C44) and
DBE = 3–4. We associate these compounds as responsible for the TAN observed for the original oil.
Ó 2013 Elsevier Ltd. All rights reserved.

1. Introduction matic hydrocarbons (90%). The remainder (10%) consists of po-

lar compounds containing N, O, and S heteroatoms and metal
The chemical composition of crude oil consists predominantly atoms (only vanadium and nickel exist at concentrations >1 ppm)
of hydrocarbon compounds such as naphthenes, paraffins, and aro- [1]. Despite the small percentage of polar compounds, approxi-
mately 20,000 polar organic compounds with different elemental
compositions (CcHhNnOoSs) have been found in crude oil [2]. These
⇑ Corresponding authors. Tel.: +55 27 3335 2955.
polar compounds sometimes cause problems during the produc-
E-mail addresses: [email protected] (B.G. Vaz), wandersonromao@
gmail.com (W. Romão). tion, refining and storage of petroleum. These problems include

0016-2361/$ - see front matter Ó 2013 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.fuel.2013.02.007
648 K.A.P. Colati et al. / Fuel 108 (2013) 647–655

corrosion, the formation of emulsions, the poisoning of catalysts, is associated with the total acid number (TAN), which is defined as
coke formation, the development of poisonous and carcinogenic the mass of potassium hydroxide (KOH) in milligrams required to
characteristics, and contamination. Oil consumption is continu- neutralize 1 g of crude oil. However, it has been argued that there
ously growing, creating demand to use the limited oil reservoirs is no clear correlation between the TAN and the level of corrosion
and heavier fractions more efficiently. Therefore, it is important [6].
to know the composition of crude oils from different origins to Crude oil contains both neutral and basic nitrogen compounds.
optimize the refining processes. Neutral nitrogen compounds include, for example, carbazoles, in-
Among the polar components of petroleum containing hetero- doles, and pyrroles and correspond to less than 30% of all organic
atoms, naphthenic acids and phenols are the two most common nitrogen compounds [7]. Basic nitrogen compounds include, for
oxygen-containing compound classes in crude oil. There are other example, pyridine and quinoline derivatives. Nitrogen compounds
minor acidic classes, such as aromatic, olefinic, hydroxyl, and diba- are harmful in oil refining because they decrease the efficiency of
sic acids [3]. Naphthenic acids are defined as carboxylic acids that catalytic processes and increase product instability during storage
include one or more saturated ring structures, with five- and six- [8].
membered rings being the most common. In addition to ring-con- Recently, there has been growing interest in the chemical char-
taining acids, linear carboxylic acids are often included in the acterization of naphthenic acids and the acidic fractions of crude
naphthenic acid class [4]. Naphthenic acids are known to be a sig- oils due to the problems that these components cause for the oil
nificant source of corrosion in oil-refining equipment [5]. Corrosion refinery business [9]. Naphthenic acids are predominantly found

Table 1
The TAN and sulfur content of the two crude oil samples and their respective residual oils.

Sample TAN (mg of KOH/g) Reduction in the TAN (%) Sulfur (%)
Crude oil I 4.95 0 0.97
Residual oil I – pH 7 3.79 23 0.93
Residual oil I – pH 10 3.08 38 0.93
Residual oil II – pH 14 0.40 92 0.71
Crude oil II 3.19 0 0.59
Residual oil II – pH 7 2.27 29 0.59
Residual oil II – pH 10 2.28 28 0.58
Residual oil II – pH 14 0.46 85 0.51

Fig. 1. Scheme of the liquid–liquid extraction of naphthenic acids. (FA – fraction acid, FAR – fraction acid residue).
 K.A.P. Colati et al. / Fuel 108 (2013) 647–655 649

in low mature heavy crude. Hence, they are assumed to be gener- extraction process in terms of TAN reduction, thus providing, a de-
ated from the in-reservoir biodegradation of petroleum hydrocar- tailed characterization of chemical composition of acidic polar spe-
bons. Naphthenic acids can be obtained from crude oils by cies in function of pHs of extraction (pH 7 ? 14).
extraction procedures such as liquid/liquid extraction with alkaline
solutions and solid/liquid extraction or by ion exchange chroma-
tography [9–11]. These isolated naphthenic acid fractions and 2. Experimental
other polar fractions have been examined by Fourier transform
infrared spectroscopy (FTIR), nuclear magnetic resonance, and 2.1. Reagents
mass spectrometry (MS) [9,12]. However, most methods (with
exception of MS) do not allow the determination of the elemental Ethanol, petroleum ether, diethyl ether, chloridric acid, HCl,
compositions of compounds. Fourier transform ion cyclotron reso- anhydrous propan-2-ol, toluene and potassium hydroxide, KOH,
nance mass spectrometry (FT-ICR MS) offers the highest available (analytical grades with purity higher than 99.5%) were used for
mass resolution, mass resolving power, and mass accuracy, en- acidic extraction method and TAN measurements, being supplied
abling the analysis of complex petroleum mixtures on a molecular by Vetec Química Fina Ltda., Brazil. Ammonium hydroxide, NH4OH,
level [13]. High-resolution MS data have shown that it is possible and sodium trifluoroacetate, NaTFA, were purchased from Sigma–
to discriminate different compounds [14] because of the different Aldrich Chemicals USA and used for ESI()-FT-ICR MS measure-
ionization efficiencies of crude oil constituents [15]. ments. All reagents were used as received.
The extraction and characterization of the acids compounds
from crude oil is necessary in order to have clear information on 2.2. Acidic extraction
the role of these acids. Herein, two acidic crude oil samples
(TAN = 4.95 and 3.19 mg KOH g1, respectively) were submitted Two samples of crude oil (termed crude oil I and crude oil II;
to a liquid/liquid extraction process modified [9] with alkaline TAN > 3 mg KOH g1, Table 1) were submitted to an acidic extrac-
solutions at three different pHs (pH 7, 10 and 14), yielding acidic tion procedure according to the scheme shown in Fig. 1. Briefly, (i)
fractions and residual oil samples. Both, acidic fractions and resid- 100 mL of crude oil was washed five times with 200 mL of a 70% (v/
ual oil samples were analyzed using negative-ion electrospray ion- v) ethanol aqueous solution at pH 7, 10 or 14, producing an aque-
ization FT-ICR MS and FTIR. The sulfur (S) content and TAN also ous phase and a residual phase (termed the residual oil – pH 7, 10
were determined. The main objective of this study is to evaluate and 14). The three aqueous phases from each wash were then ex-
the efficiency and selectivity of non-continuous liquid/liquid tracted with two times 100 mL of (ii) petroleum ether (30–60 °C).

100 100

95 95
Transmitance (%)

Transmitance (%)

90 90

85 85

80 80
Crude Oil
Crude Oil Residual Oil - pH 7
75 Acid Fraction - pH 7 75
Residual Oil - pH 10
Acid Fraction - pH 10
Acid Fraction - pH 14
(a) Residual Oil - pH 14 (b)
70 70
4000 3500 3000 2500 2000 1500 1000 4000 3500 3000 2500 2000 1500 1000

Wavenumber (cm-1) Wavenumber (cm-1)

100 100

95
Transmitance (%)

Transmitance (%)

90 90

85

80 80
Crude Oil
Crude Oil Residual Oil - pH 7
75 Acid Fraction - pH 7
Residual Oil - pH 10
Acid Fraction - pH 10
Acid Fraction - pH 14 (c) Residual Oil - pH 14 (d)
70 70
4000 3500 3000 2500 2000 1500 1000 4000 3500 3000 2500 2000 1500 1000
Wavenumber (cm-1) Wavenumber (cm ) -1

Fig. 2. FT-IR spectra of crude oil I and crude oil II and their respective acidic fractions (a and c) and residues (b and d).
650 K.A.P. Colati et al. / Fuel 108 (2013) 647–655

The aqueous extracts were combined, and the ethanol was evapo- potassium hydroxide per gram of sample required to titrate a sam-
rated to prevent the formation of esters upon acidifying the solu- ple in a specified solvent to a specified end point (see Table 1) [16].
tion in the next step. The pH was then adjusted to pH 1 (using a
concentrated solution of HCl) before extraction with four times
100 mL of diethyl ether. The diethyl ether was then evaporated 2.4. Sulfur content
to obtain the polar acidic fraction (three acidic fractions termed
acidic fraction – pH 7, acidic fraction – pH 10 and acidic fraction The sulfur content was determined according to the standard
– pH 14). The acidic extraction method developed in this work ASTM method [17] (ASTM D4294) using an SLFA-2800 energy dis-
was based on that of Hemmingsen et al. [9] The initial crude oils, persive X-ray fluorescence spectrometer from Horiba Scientific
residual oils and acidic fractions were subsequently analyzed by (New Jersey, USA). First, calibration curves were constructed, and
FT-IR and ESI()-FT-ICR MS. The TAN and S content were deter- parameters such as the linearity and limits of detection were ob-
mined for the crude oils and residual oil samples. tained. The measurement time was 100 s. The sulfur contents of
the crude oil and residue samples were determined in triplicate.

2.3. Characterization 2.4.1. FTIR

For the FTIR studies, an ABB BOMEN IR (FTLA2000-102 model)
2.3.1. TAN coupled with a MIRacle attenuated total reflectance (ATR) acces-
TAN measurements were performed according to the standard sory was used. Crude oil samples and their respective fractions
ASTM method (ASTM D664-09) using a potentiometer (Metrohm and residues were placed under a single-reflection crystal plate,
Analytical Instruments and Accessories, USA) with an accuracy of and a total of 16 scans were taken. The spectra were recorded from
pH ± 0.003. In 250 mL beakers, the crude oil samples (5.0 ± 0.5 g) 4000 to 400 cm1 in transmission mode with a resolution of
and their respective residues (20 ± 2 g) were weighed, dissolved 4 cm1. The background was performed in air, and background
in 125 mL of water/anhydrous propan-2-ol/toluene (0.5/49.5/50% scans were acquired before each sample was analyzed. The ATR-
in volume), and then titrated, in duplicate, with a 0.1 mol L1 alco- FTIR spectra were acquired using GRAMS/AI software (Thermo
holic KOH solution. The results are expressed as milligrams of Galactic).

[C21H38O2 - H] [C21H38O2 - H]
O O
DBE 3 [C24H 21N - H] DBE 3 [C24H 21N - H]
R R 321.28 R
321.28 N DBE 15 R N
OH DBE 15
OH
H 322.16 H 322.16

322.28

322.16 321.28

321.28 322.16

321.28 321.28
322.16

322.16
322.28

321.28 322.16
321.28

322.16 322.28

321.0 321.5 322.0 322.5 321.0 321.4 321.8 322.2

(a) Crude Oil I (a) Crude Oil I

(b) Acidic fraction (e) Residual Oil

pH 7 pH 7

(c) Acidic faction (f) Residual Oil

pH 10 pH 10

(d) Acidic fraction (g) Residual Oil

pH 14 pH 14

200 300 400 500 600 700 800 900 200 300 400 500 600 700 800 900
m/z m/z

Fig. 3. ESI()-FT-ICR MS spectra ofcrude oil I (a) and its acidic fractions and residual oils when extracted at pH 7 (b and e), 10 (c and f) or 14 (d and g). The Mw distribution
broadened and shifted to higher m/z values with increasing of. The insert shows the naphthenic acid and carbazole structures with their respective m/z and DBE values. The
intensity of the[C21H38O2AH]ion at m/z 321.28 increased for the acidic fractions and decreased for the residual oils as a function of pH (7 ? 14).
 K.A.P. Colati et al. / Fuel 108 (2013) 647–655 651

O R O
[C24H42O2 - H]
DBE = 4 N R
361.31
OH [C24H42O2 - H] OH
H DBE = 4 N
R R
361.31 H
[C27H26N - H]
DBE= 16 [C27H26N - H]
362.19 DBE= 16
362.31 362.19
362.16
361.31 361.31

362.31 362.19
362.19

361.31 361.31

362.19
362.31

361.31 362.19
361.31

362.31

361.2 361.4 361.6 361.8 362.0 362.2 362.4 361.2 361.4 361.6 361.8 362.0 362.2

(a) Crude Oil II (a) Crude Oil II

(b) Acidic fraction (e) Residual Oil

pH 7 pH 7

(c) Acidic faction (f) Residual Oil

pH 10 pH 10

(d) Acidic fraction (g) Residual Oil

pH 14 pH 14

200 300 400 500 600 700 800 900 1000 200 300 400 500 600 700 800 900 1000
m/z m/z

Fig. 4. ESI()-FT-ICR MS spectra of crude oil II (a) and its respective acid fractions and residual oils when extracted at pH 7 (b and e), 10 (c and f) or 14 (d and g). The Mw
distribution broadened and shifted to higher m/z values with increasing pH. The insert shows the naphthenic acid and carbazole structures with their respective m/z and DBE
values. The intensity of the [C24H42O2AH]ion at m/z 361.31 increased for the acidic fractions and decreased for the residual oils as a function of pH (7 ? 14).

2.4.2. ESI()-FT-ICR MS (Bruker Daltonics, Bremen, Germany). Using the Composer soft-
Petroleum samples were analyzed by ESI()-FT-ICR MS. Briefly, ware (Sierra Analytics, Pasadena, CA, USA), the MS data were han-
each sample was diluted to 1.2 mg mL1 in 50:50 (v/v) toluene/ dled, and the elemental compositions of the compounds were
methanol (containing 0.1% m/v of NH4OH). The resulting solution determined by measuring the m/z values [18]. To help summarize,
was directly infused at a flow rate of 5 ll min1 into the ESI source. visualize, and interpret the MS data, three classical plots were con-
The mass spectrometer (model 9.4 T Solarix, Bruker Daltonics, Bre- structed: (i) the distribution of heteroatom-containing com-
men, Germany) was set to operate in negative ion mode (ESI()) pounds; (ii) a plot of the carbon number vs the double bond
over a mass range of m/z 200–2000. The ESI source conditions were equivalents (DBE); (iii) and the DBE distributions for N1 and O2
as follows: nebulizer gas pressure of 0.5–1.0 bar, capillary voltage species, where DBE is defined as the number of rings plus the num-
of 3–3.5 kV, and transfer capillary temperature of 250 °C. The ion ber of double bonds in a molecular structure. The aromaticity of a
accumulation time in the hexapole was 0.01 s and was followed petroleum component can be deduced directly from its DBE value
by transport through a hexapole ion guide to the ICR cell. Each according to the following equation:
spectrum was acquired by accumulating 200 scans of time-domain
transient signals in four mega-point time-domain data sets. The DBE ¼ c  h=2 þ n=2 þ 1 ð1Þ
front and back trapping voltages in the ICR cell were 0.60 V and
where c, h, and n are the numbers of carbon, hydrogen, and nitrogen
0.65 V, respectively. All mass spectra were externally calibrated
atoms, respectively, in the molecular formula.
using a NaTFA (m/z from 200 to 1200) after they were internally
recalibrated using a set of the most abundant homologous alkyl-
ated compounds for each sample. A resolving power (m/ 3. Results and discussion
Dm50%  530,000, in which Dm50% is the full peak width at half-
maximum peak height) of m/z 400 and a mass accuracy of Table 1 shows the TANs for the two crude oil samples (I and II)
<1 ppm provided unambiguous molecular formula assignments and their respective residual oils (obtained from the original oil
for singly charged molecular ions. Mass spectra were acquired washed at different pHs: pH 7, 10 and 14, Fig. 1). First, the TANs
and processed using the software package Compass Data Analysis for crude oil I and crude oil II were 4.95 and 3.19 mg KOH g1,
652 K.A.P. Colati et al. / Fuel 108 (2013) 647–655

respectively. Decreases in the TANs of both crude oils were ob- showed that 90% of the acidic compounds consisted of carboxylic
served with increasing pH, reaching values of almost zero at pH acids with molecular weights in the range 300–800 Da and a
14. Additionally, the TANs observed for the oils at washed pH 7 DBE = 3 being the most abundant. Therefore, their results agree
and 10 were closer. This behavior can be explained by a possible with the results obtained in this work. The main difference is re-
variation in the pH due to the hydrolysis processes that may occur lated to the MS spectra for the residual oil samples, where the spe-
during the extraction process, primarily at pH 10, because that no cies distributed in the range of m/z 600–900 corresponded to an
buffer was used during the first step of this process (Table 1). How- asphaltene fraction in the oil washed at pH 14. This difference is
ever, the acidic extraction process at pH 14 was more efficient for due to the continuous acidic extraction process, in which the same
both crude oils, with reductions in the TAN of 92% and 85%, residual oil was washed consecutively with aqueous ethanol solu-
respectively. tions at pH 7, 10 and 14.
For the sulfur content, the values remained constant from the Similar ESI()-FT-ICR MS results were also obtained for another
crude oil to the residual oil at pH 7 and 10 (S = 0.97 ? 0.91% for crude oil sample (crude oil II, Fig. 4a–g), thus confirming the con-
crude oil I and S = 0.59 ? 0.58% for crude oil II; Table 1). A reduc- tinuous extraction of the same compound types at each pH evalu-
tion of 20–30% was observed only for the residual oil at pH 14. ated. Note that the Figs. 3a and 4a were duplicated in order to
Given that sulfur species, such as thiols, have pKa values of approx- compare changes such as Mw and polar species between crude oils
imately 10.5, this reduction is expected when performing the and their acidic fraction and residual oils.
extraction procedure at pH > 10. In 2012, Freitas et al. [18] studied the correlation between the
Fig. 2a–d shows the FT-IR spectra for the samples of crude oil I composition of heavy distillation cuts and the corrosion process.
and crude oil II and their fractions (2a and c) and residual oils (2b The cuts presented boiling temperatures lower than 315 °C and
and d). For the FTIR spectra of the crude oils, three main regions TAN values changes from 0.47 to 1.83 mg g1. According to the re-
were identified: (i) 2923 cm1 and 2853 cm1: asymmetric and sults of ESI()-FT-ICR MS, naphthenic acid species were only de-
symmetric CAH stretching, respectively, of CH3 and CH2 groups; tected in the seven cuts studied, with m/z < 300 and DBE values
(ii) 1452 cm1 and 1376 cm1: CH2 and CH3 deformations (bend- ranging from 2 to 4. The relative abundance of heavy naphthenic
ing); and (iii) 900–700 cm1: aromatic CAH deformation. All these acids and the TAN increase with distillation cut temperature. This
vibration frequencies are typical of hydrocarbons [19,20]. In con- information correlated with the corrosion mechanism for AISI 1020
trast, in the FTIR spectra corresponding to the acidic fractions steel in the heavy oil and distillation cuts, where the corrosion is
(Fig. 2a and c), an intense band in the regions of 1700 cm1 and evident in oil cuts being characterized as alveolar and pitting. In
3380 cm1 were observed due to the carbonyl (C@O), hydroxyl 2007, Silva and Senna [22] have also demonstrated that the reduc-
(OH) and amine groups (NAH) present in the heteroatom-contain- tion of the TAN from light petroleum cuts affected directly the
ing compounds, which were primarily naphthenic acids, carbazoles
and pyridines analogues [19–21]. However, with increasing pH
(7 ? 14), a distinct behavior was observed for both bands: increas-
ing and decreasing intensities were observed for the C@O and NAH
groups, respectively (see the enlargement shown in Fig. 2a and c).
This behavior is indicative of the increasing concentration of naph-
thenic acids in the acidic extracts, primarily at pH 10 and 14. For
the residual oils, an increase in the concentration of nitrogen spe-
cies was observed due to the removal of acidic groups from the ori-
ginal crude oil (see the enlargement of the region around
3380 cm1, Fig. 2b and d).
Fig. 3a–g shows the ESI()-FT-ICR MS spectra for crude oil I (3a)
and its acidic fractions (3b–d) and residues (3e–g). Heteroatom-
containing species were detected as deprotonated molecules, that
are [MH] ions. For crude oil I (3a), an ESI()-FT-ICR MS spec-
trum with an m/z range from 200 to 900 and a maximum centered
at m/z 460 was observed, corresponding primarily to naphthenic
acids and carbazole analog species. For the acid fraction (3b–d), a
lower average molecular weight distribution, Mw, was observed
for each sample (m/z range: 200–700). Note that with increasing
pH, the Mw distribution broadened and shifted to higher m/z val-
ues: pH 7: m/z 200–350 and Mw = 270 Da; pH 10: m/z 200–450
and Mw = 300 Da; and pH 14: m/z 200–650 and Mw = 390 Da.
Therefore, low alkyl chain lengths acid species, m/z 200–650, are
responsible by most TAN observed (from 85% to 92%).
As consequence of the removal of naphthenic acid species from
crude oil, a decreasing in the Mw was observed in the mass spectra
for the residual oils (Fig. 3e–g). The enlarged area around m/z 321
(Fig. 3a–g) highlights the selectivity of the extraction of acid spe-
cies. The relative intensity of the [C21H38O2AH] ion, which has
an m/z of 321.28, increased for the acidic fractions and decreased
for the residual oils as a function of pH. Similar to the TAN and FTIR
results, the ESI()-FT-ICR MS spectrum also demonstrated that the
extraction process at pH 14 was more efficient and selective.
In 2006, Hemmingsen et al. [9] studied the molecular composi-
tion of different acidic fractions produced from a North Sea acidic Fig. 5. Heteroatom-containing compound class distribution for crude oil I (5a) and
crude oil (NAT = 2.9 mg of KOH g1). Their FT-ICR MS results II (5b) and their acid and residue fractions at pH 7, 10 and 14.
 K.A.P. Colati et al. / Fuel 108 (2013) 647–655 653

Fig. 6. Plot of the DBE vs the carbon number for the O2 and N classes of crude oil II (a and h) and their acid extracts (b–d and i–k) and residues (e–g and l–n) at pHs 7, 10 and
14. The maximum carbon number decreased from C32 (crude oil, 6a) to C25 (acid fraction obtained at pH 14, 6 g) for the O2 class. Generally, carbazole species (N class) were
negligible in the acid extracts, being present in high abundance only in the original oil and residues, in which the carbon number ranged from C16AC64(6 h) to C18AC52 (6 l, m,
n).

reduction in corrosion yields. Therefore, the liquid/liquid extrac- It should be noted that the relative abundance of O2 compounds
tion method can be applied to prevent corrosion process in regions increased after the acid extraction process (pH 7 ? 14) from 63%
of the refineries working at temperatures above 100.0 °C, since that to 81% for crude I (Fig. 5a) and from  64 to 99% for crude oil II
the most acid species are obtained from acidic fractions (Fig. 5b) (compare the histograms for the original oil and the acidic
(Mw = 270–390 and m/z 200–650), thus covering, all acid species extracts). Consequently, a decrease in the relative abundance of the
detected from heavy distillation cuts as reported by Freitas et al. O2 class was observed for the residual samples, with decreases
[18]. In general, it is always important to be careful in the correla- from 63% to 36% for crude oil I (5a) and from 64% to 49%
tion between the TAN and the corrosion process. The corrosion for crude oil II (Fig. 5b). This result clearly indicates the greater effi-
mechanism of naphthenic acids has not been fully elucidated, ciency of the extraction process in alkaline media with increasing
being known that the corrosive nature of oil is not directly related pH.
to the TAN but to the presence of a specific group of acids and sul- Plots of the DBE vs the carbon numbers have proven to be useful
fur compounds present in oil [18,23]. tools for the differentiation of complex organic mixtures based on
Generally, petroleum samples have chemical compositions that chemical composition. It has been shown that these plots allow the
differ significantly from one another [24]. One way to display the presentation of all signals from a specific class in a simple and fea-
similarities or differences between the signal patterns of crude sible way [25,26]. For this reason, to qualitatively compare the
oil samples is to construct certain types of plots, such as plots of crude oils and their acid fractions, DBE vs carbon number plots
the relative abundances of different classes of compounds and of were generated, showing the distribution of all components from
the DBE vs the carbon number. a specific class. Each line in the diagram represents a homologous
In the present study, for the class profile diagrams, the relative series with a specific DBE value; each line differs by one CH2 unit.
abundance was calculated by summing the peak magnitudes for all To simplify the discussion and allow the better understanding of
ions of each class and dividing by the summed peak magnitudes for the changes in the profiles of different compound classes, we will
ions of all assigned masses. The relative intensity of N, NO2 and O2 focus this discussion on the data obtained for crude oil II and its
compounds present in the crude oils and their fractions are pre- respective acidic and residue factions (Fig. 6).
sented in Fig. 5a and b. For both oils, the O2 compounds class Fig. 6a–n show plots of the DBE vs the carbon number for the O2
was the most abundant class. The N class was the second most and N classes of crude oil II (a and h) and their acidic (b–d and i–k)
abundant, followed by the NO2 class. and residue fractions (e–g and l–n). For the O2 class, there were
654 K.A.P. Colati et al. / Fuel 108 (2013) 647–655

Fig. 7. Relative abundances of the O2 and N compound classes in the ESI()-FT-ICR MS spectra of crude oil II and its (a) acid and (b) residue fractions produced at pH 7, 10 and
14. Naphthenic acids with DBEs of 2, 3 4, 7 and 8 were the main compounds removed from the original oil (7a).

compounds with DBEs ranging from 2 to 15 and carbon numbers hindrance [11,12a]. This conclusion indicates that aromatic naph-
ranging from C16 to C68, with the most abundant compounds being thenic acid species are mainly detected when higher pH values
C32 with a DBE of 4. The most abundant O2 compounds were are used during extraction.
mainly composed of three naphthenic rings, in agreement with Fig. 6e–g shows the distribution of N compounds for crude oil II.
the results reported by Schaub et al. [27]. A notable difference Generally, the range of DBEs was notably higher and wider for the
among the acidic fractions was observed when the O2 class was residual fractions than the acid fractions, with DBEs ranging from 9
analyzed. For this class, the DBE and carbon distributions increases to 24 and carbon numbers ranging from C16 to C64. These results
as a function of pH (pH 7 ? 14, Fig. 6a–n). Initially, acid fraction – are in good agreement with the acidity of N-containing compounds
pH 7 contained a series of compounds with DBEs ranging from 1 to such as benzocarbazoles (DBE = 12) and dibenzocarbazoles
10 Fig. 6b, including compounds with a narrow carbon distribution (DBE = 15), which have high pKa values and fully aromatic struc-
(from C12 to C38). However, for acid fraction – pH 14, Fig. 6d, a wide tures [28]. The results obtained for crude oil I are presented in
DBE distribution is observed, containing compounds with carbon the Supplementary material (Fig. S1).
distribution from C12 to C48. It is noteworthy that, at high pH, the Fig. 7 shows the relative abundance distributions for the O2 and
extraction process remove majority a large amount of O2 com- N compound classes in crude oil II and its acidic (7a) and residue
pounds due to the deprotonating mechanism that occurs within (7b) fractions prepared at pH 7, 10 and 14. Naphthenic acids with
this pH range. DBEs of 2, 3, 4, 7 and 8 were primary compounds removed from the
It is well-known that carboxylic acids exhibit strong acidities original oil (Fig. 7a). Consequently, an increase in the relative abun-
because of their abilities to form resonance hybrids. The equal dis- dance distribution of the N class was observed for the residual
tribution of the negative charge between the two oxygen atoms samples (mainly with DBE = 12–15, Fig. 7b). The results for crude
significantly reduces their attraction to a charged proton. Due to oil I are presented in Fig. S2 (Supplementary material).
their complex chemical composition, naphthenic acids exhibit rel-
atively considerable variation in their acidity. The calculated pKa
values for most naphthenic acids range from 3.0 to 20.0. This value 4. Conclusion
is influenced by the hydrocarbon structure: cyclic vs acyclic and
saturated vs aromatic rings. For aromatic structures, the pKa sub- A method for the liquid–liquid extraction of naphthenic acids in
stantially increases up to 1.0 unit. For saturated cyclic and acyclic alkaline solutions has been adapted and was applied to two acidic
structures, the magnitude of the change in acidity is reduced on crude oil samples (TAN = 4.95 and 3.19 mg KOH g1). We observe a
the order of 0.2–0.4 units, primarily due to the effects of steric decreasing TAN in the original oil with increasing pH (7 ? 14),
 K.A.P. Colati et al. / Fuel 108 (2013) 647–655 655

reaching values of almost zero at pH 14. Additionally, the TANs ob- cyclotron resonance mass spectrometry: the effect of suppression. Energy
Fuels 2007;21:266–73.
served for the oils washed at pH 7 and 10 were closer due to the
[8] Wu Z, Rodgers RP, Marshall AG. Comparative compositional analysis of
variation in the pH during the extraction process, especially at untreated and hydrotreated oil by electrospray ionization fourier transform
pH 10. Therefore, the extraction process in alkaline solution at ion cyclotron resonance mass spectrometry. Energy Fuels 2005;19:1072–7.
pH 14 was most efficient for both crude oils, which exhibited [9] Hemmingsen PV, Kim S, Pettersen HE, Rodgers RP, Sjoblom J, Marshall AG.
Structural characterization and interfacial behavior of acidic compounds
reductions in the TAN of 92% and 85%, respectively. In the FTIR extracted from a North Sea Oil. Energy Fuels 2006;20:1980–7.
spectra, an intense band in the region of 1700 cm1 was observed [10] Seifert WK, Howells WG. Interfacially active acids in a california crude oil.
for the naphthenic acidic fractions, indicating that there was an Isolation of carboxylic acids and phenols. Anal Chem 1969;41:554–62.
[11] Saab J, Mokbel I, Razzouk AC, Ainous N, Zydowickz N, Jose J. Quantitative
increasing concentration of acid species, primarily at pHs 10 and extraction procedure of naphthenic acids contained in crude oils.
14. For the residual oils, an increase in the concentration of nitro- Characterization with different spectroscopic methods. Energy Fuels
gen species (3380 cm1) was observed due to the removal of acidic 2005;19:525–31.
[12] (a) Tomczyk NA, Winans RE, Shinn JH, Robinson RC. On the nature and origin of
groups from the original crude oil. The primary acid species acidic species in petroleum 1. Detailed acid type distribution in a california
responsible for the variation in the TAN were identified based on crude oil. Energy Fuels 2001;15:1498–504;
the ESI()-FT-ICR MS spectra. These acid species exhibited average (b) Islas-Flores CA, Buenrostro-Gonzalez E, Lira-Galeana C. Comparisons
between open column chromatography and HPLC SARA fractionations in
molecular weight distribution, Mw, and m/z values from m/z 200– petroleum. Energy Fuels 2005;19:2080–8.
350 and Mw = 270 Da (at pH 7) to m/z 200–650 and Mw = 390 Da [13] Hsu CS, Hendrickson CL, Rodgers RP, McKenna AM, Marshall AG. Petroleomics:
(at pH 14), with a maximum alkyl chain length of C44 and a high advanced molecular probe for petroleum heavy ends. J Mass Spectrom
2011;46:337–43.
abundance of DBE = 4. This result suggested a structure that in-
[14] Gaspar A, Schrader W. Expanding the data depth for the analysis of complex
cludes one carboxylic group and three naphthenic rings. In addi- crude oil samples by Fourier transform ion cyclotron resonance mass
tion to the high abundance of acidic polar species (O2 class), spectrometry using the spectral stitching method. Rapid Commun Mass
other species were also found in lower abundances, including Spectrom 2012;26:1047–52.
[15] Klein GC, Angstrom A, Rodgers RP, Marshall AG. Use of Saturates/Aromatics/
N-containing compounds and NO2 compounds. Resins/Asphaultenes (SARA) fractionation to determine matrix effects in crude
oil analysis by electrospray ionization fourier transform ion cyclotron
Acknowledgments resonance mass spectrometry. Energy Fuels 2006;20:668–72.
[16] ASTM D664-09. Standard test method for acid number of petroleum products
by potentiometric titration; 2006.
This research was generously funded by PETROBRAS/CENPES, [17] ASTM D 4294. Standard test method for sulfur in petroleum and petroleum
CNPq, CAPES and FINEP. products by energy dispersive X-ray fluorescence spectrometry; 2008.
[18] Freitas S, Malacarne MM, Romão W, Dalmaschio GP, Castro EVR, Celante VG,
Freitas MBJG. Analysis of the heavy oil distillation cuts corrosion by
Appendix A. Supplementary material electrospray ionization FT-ICR mass spectrometry, electrochemical
impedance spectroscopy, and scanning electron microscopy. Fuel
2013;104:656–63.
Supplementary data associated with this article can be found, in [19] Castro LV, Vazquez F. Fractionation and characterization of mexican crude oils.
the online version, at http://dx.doi.org/10.1016/j.fuel.2013.02.007. Energy Fuels 2009;23:1603–9.
[20] Watson JS, Jones DM, Swannell RPJ, Van Duina ACT. Formation of carboxylic
acids during aerobic biodegradation of crude oil and evidence of microbial
References oxidation of hopanes. Org Geochem 2002;33:1153–69.
[21] Silverstein RM, Webster FX, Kiemle DJ. Identificação espectrométrica de
[1] (a) Hughey CA, Rodgers RP, Marshall AG, Walters CC, Qian K, Mankiewicz P. compostos orgânicos. 7th ed. Rio de Janeiro: Editora LTC; 2006.
Acidic and neutral polar NSO compounds in smackover oils of different [22] Silva JP, Senna LF, Do Lagoa DCB, Junior PFS, Dias EG, de Figueiredo MAG,
thermal maturity revealed by electrospray high field Fourier transform ion Chiaro SSX. Characterization of commercial ceramic adsorbents and its
cyclotron resonance mass spectrometry. Org Geochem 2004;35:863–80; application on naphthenic acids removal of petroleum distillates. Mater Res
(b) Hughey CA, Rodgers RP, Marshall AG, Qian K, Robbins WK. Identification of 2007;10:219–25.
acidic NSO compounds in crude oils of different geochemical origins by [23] Laredo GC, Lópes CR, Álvarez RE, Cano JL. Naphthenic acids, total number and
negative ion electrospray Fourier transform ion cyclotron resonance mass sulfur content profile characterization in Isthmus and Maya crude oils. Fuel
spectrometry. Org Geochem 2002;33:743–59. 2004;83:1689–95.
[2] Marshall AG, Rodgers RP. The next grand challenge for chemical analysis. Acc [24] Mapolelo MM, Rodgers RP, Blakney GT, Yenc AT, Asomaning S, Marshall AG.
Chem Res 2004;37:53–9. Characterization of naphthenic acids in crude oils and naphthenates by
[3] Brient JA. Commercial utility of naphthenic acids recovered from petroleum electrospray ionization FT-ICR mass spectrometry. Int J Mass Spectrom
distillates. In: Symposium on Acidity in Crude Oil, Presented Before the 2011;300:149–57.
Division of Petroleum Chemistry, Inc., 215th National Meeting. Dallas (TX): [25] Hughey CA, Hendrickson CL, Rodgers RP, Marshall AG. Kendrick mass defect
American Chemical Society; 1998. spectroscopy: a compact visual analysis for ultrahigh-resolution broadband
[4] Barrow MP, Headley JV, Peru KM, Derrick PJ. Data visualization for the mass spectra. Anal Chem 2001;73:4676–81.
characterization of naphthenic acids within petroleum samples. Energy Fuels [26] Van Krevelen DW. Graphical-statistical method for the study of structure and
2009;23:2592–9. reaction processes of coal. Fuel 1950;29:269–84.
[5] Wu Z, Rodgers RP, Marshall AG. Compositional determination of acidic species [27] Schaub TM, Jennings DW, Kim S, Rodgers RP, Marshall AG. Heat-exchanger
in Illinois #6 coal extracts by electrospray ionization Fourier transform ion deposits in an inverted steam-assisted gravity drainage operation. Part 2.
cyclotron resonance mass spectrometry. Energy Fuels 2004;18:1424–8. Organic acid analysis by electrospray ionization fourier transform ion
[6] Barrow MP, McDonnel LA, Feng X, Walker J, Derrick PJ. The continued battle cyclotron resonance mass spectrometry. Energy Fuels 2007;21:185–94.
against corrosion. Anal Chem 2003;75:860–6. [28] Balon M, Carmona MC, Muñoz MA, Hidalgo J. The acid–base properties of
[7] Teräväinem MJ, Pakarinem JMH, Wickström K, Vainiotalo P. Comparison of the pyrrole and its benzologs indole and carbazole. A reexamination from the
composition of Russian and north sea crude oils and their eight distillation excess acidity method. Tetrahedron 1989;45:7500–4.
fractions studied by negative-ion electrospray ionization fourier transform ion