The Journal of Supercritical Fluids 140 (2018) 101–115

Contents lists available at ScienceDirect

The Journal of Supercritical Fluids

journal homepage: www.elsevier.com/locate/supflu

Process simulation of organic liquid products fractionation in T

countercurrent multistage columns using CO2 as solvent with Aspen-Hysys
E.C. Costaa,b, C.C. Ferreiraa,b, A.L.B. dos Santosa,c, H. da Silva Vargensa,c, E.G.O. Menezesa,c,
⁎
V.M.B. Cunhaa, M.P. da Silvaa, A.A. Mâncioa,b, N.T. Machadoa,b, , M.E. Araújoa,c
a
Laboratory of Separation Processes and Applied Thermodynamic (TERM@), Faculty of Chemical Engineering-UFPA, Brazil
b
Graduate Program of Natural Resource Engineering-UFPA, Rua Augusto Corrêia Nº. 1, CEP: 66075-900, CP. 8619, Belém, Pará, Brazil
c
Graduate Program of Chemical Engineering-UFPA, Rua Augusto Corrêia Nº. 1, CEP: 66075-900, CP. 8619, Belém, Pará, Brazil

G R A P H I C A L A B S T R A C T

A R T I C LE I N FO A B S T R A C T

Keywords: Fractionation of liquid phase products (OLP), obtained by thermal catalytic cracking of palm oil, simulated in
Organic liquid products multistage countercurrent absorber/stripping columns using SC-CO2 as solvent with Aspen-Hysys, at 333 K, 140
Fractionation and 180 bar, and (S/F) = 12, 15, 17, 25, 30, 38. RK-Aspen EOS applied to correlate equilibrium data of binary
Carbon dioxide systems (OLP)-i/CO2. Composition of OLP used to predict thermophysical properties and acentric factor of OLP
EOS
compounds by group-contribution methods. The effect of pressure and (S/F) on process performance evaluated
Process simulation
by computing the yield and recoveries of hydrocarbons, olefins, oxygenates, carboxylic acids. The regressions
Aspen-Hysys
show that RK-Aspen EOS describe well phase equilibrium data, showing AAD between 0.8% and 1.25% and
0.01%– 0.66% for the liquid and gaseous phases. The best deacidifying conditions obtained at 333 K, 140 bar,
and (S/F) = 17. Simulations show that SC-CO2 was able to deacidify OLP and to obtain fractions with lower
olefins content.

⁎
Corresponding author at: Graduate Program of Natural Resource Engineering-UFPA, Rua Augusto Corrêia Nº. 1, CEP: 66075-900, CP. 8619, Belém, Pará, Brazil.
E-mail address: machado@ufpa.br (N.T. Machado).

https://doi.org/10.1016/j.supflu.2018.06.004
Received 16 December 2017; Received in revised form 5 June 2018; Accepted 6 June 2018
Available online 15 June 2018
0896-8446/ © 2018 Elsevier B.V. All rights reserved.
E.C. Costa et al.
1. Introduction
Pyrolysis and/or catalytic cracking is one of the most promising
process to convert triglycerides, the major compounds of vegetable oils
and animal fats [1,2], into liquid biofuels [3], as reported in the lit-
erature [3–25]. Both have the objective to obtain hydrocarbons for use
as fuels [3,4,6–24]. The reaction products obtained by pyrolysis and/or
catalytic cracking of oils, fats, grease, and fatty acids mixtures include
gaseous and liquid fuels, water and coke [6–8,14,15,17,21,7–24]. The
physicochemical properties and chemical composition of OLP depends
on the selectivity of the catalyst used [6,7,10,14–17,20,4–25]. The or-
ganic liquid products (OLP) consist basically of alkanes, alkenes, ring-
containing alkanes, ring-containing alkenes, cycle-alkanes, cycle-al-
kenes, and aromatics [8,11,12,16,17,21,7–24], corresponding to the
boiling point range of gasoline, kerosene and diesel fossil fuels, as well
as and carboxylic acids, aldehydes, ketones, fatty alcohols, and esters
[6–8,11,12,15,15,16,17,21,4–25]. The OLP can be refined and/or up-
graded by applying physical (filtration, decantation, and centrifuga-
tion) and thermal separation processes (distillation, liquid-liquid ex-
traction, and adsorption) to produce high quality green fuel-like
fractions with potential to substitute partially fossil fuels
[6,11,16,21,7–24,26–29]. The disadvantages and/or drawbacks of OLP
obtained by pyrolysis and/or catalytic cracking of oils, fats, greases, and
fatty acid mixtures remains on the high acid values [8,11,14,19,22],
and high concentrations of olefins, making it a corrosive and unstable
fuel [9,21].
In the last years, processes have been proposed to remove and/or
recover oxygenate compounds from OLP, particularly fractionation by
using single stage and multistage distillation to obtain hydrocarbons-
like fuels in the temperature boiling point range of gasoline, kerosene,
and diesel-like fractions [6,7,11,14,15,21–24,26–29]. In addition, bio-
mass-derived bio-oils have been upgraded to remove oxygenate by
applying several separation/purification processes including molecular
distillation to separate water and carboxylic acids from pyrolysis bio-
oils [30–32], fractional distillation to isolate/enrich chemicals and
improve the quality of bio-oil [33–38], liquid-liquid extraction using
organic solvents and water to recover oxygenate compounds of OLP and
bio-oils [26,39], and until non- bio-oils have been also applied [39].
However, until the moment no study has been reported in the literature
concerning the simulation of supercritical CO2 fractionation of organic
liquid products (OLP) using the commercial process simulator Aspen-
Hysys.
Multistage gas extraction in countercurrent columns, a process
identical in concept, to absorption [40] and/or liquid-liquid extraction
[41–45], is an alternative separation technique for the extraction and
fractionation of liquid mixtures [41–45]. In recent years, emphasis has
been focused on the separation of real systems such as tocopherols from
soya oil [46] and palm oil deodorizer distillates [41], among other
applications [42–45].
Despite the wide range of application of multistage gas extraction in
countercurrent packed columns, particularly in the food industry
[42,43], only a few number of works are devoted to simulation of
multistage gas extraction using self-made computer codes or commer-
cial chemical process simulators such as Aspen-Hysys, including the
simulation of the separation of mono-glycerides from a mixture of oleic
acid + glycerides as well as fatty acids from palm oil in countercurrent
columns [47], simulation of the separation of tocopherol from a syn-
thetic mixture of tocopherol + squalene + carboxylic acids using su-
percritical carbon dioxide as solvent within a flowsheet consisting of an
one stage extractor and a flash separator and the PR-EOS with the
LCVM mixing rule using the commercial simulator ASPEN+ [48], si-
mulation of deterpenation of lemon essential oil using supercritical
carbon dioxide as solvent by considering a one stage extractor using a
self-made academic computer code and the commercial simulator
ASPEN+ to simulate the fractionation in multiple stages represented by
a flowsheet consisting of a series of flash separators using the PR-EOS
102
The Journal of Supercritical Fluids 140 (2018) 101–115
with quadratic mixing rules [49], simulation process to recover provi-
tamin A (β-carotene) from palm oil (esterified) using carbon dioxide/
ethanol as the supercritical mixed solvent with the commercial simu-
lator HYSYS™ [50], fractionation of esterified olive oil deodorizer dis-
tillates, enriching of squalene from trans-esterified olive oil deodorized
distillates, separation of esters and sterols of soya oil deodorizer dis-
tillates esterified enzymatically and until the de-acidification of crude
olive oil using supercritical carbon dioxide as solvent in countercurrent
columns using an academic self-made computer code written in FOR-
TRAN [51–53] with the CG-EOS [54]. In addition, the literature reports
a dynamic model to simulate a supercritical fluid extraction plant
coupled to a countercurrent fractionation column, whose flowsheet was
modularly organized into a set of sub-models, including the extraction
packed column, the supercritical solvent heat exchanger, the product
recovery and solvent regeneration column, and the solvent make-up,
being the model validated against experimental data for the fractio-
nation of edible oil mixtures using supercritical carbon dioxide as the
solvent [55]. Recently, the modeling and simulation of the fractionation
of liquid streams from natural products, particularly vegetable oils, in
countercurrent columns absorber/stripping columns using SC-CO2 as
solvent with Aspen-Hysys, has been systematically investigated [40].
The objective of this work was to apply the commercial simulator
Aspen-Hysys 8.6 (Aspen One, 2015) to simulate the supercritical CO2
fractionation of organic liquid product (OLP), obtained by catalytic
cracking of palm oil at 450 °C, 1.0 atm, using 10% (wt.) Na2CO3 as
catalyst. The simulations were carried out by selecting the multistage
countercurrent absorber/stripping unit operation column. The Redlich-
Kwong Aspen (RK-Aspen) equation of state (EOS) was used as the
thermodynamic fluid package, applied to correlate the experimental
phase equilibrium data of binary organic liquid products (OLP) com-
pounds-i/CO2 systems available in the literature. The RK-Aspen tem-
perature-independent binary interaction parameters were obtained
using the Aspen Properties computational package from Aspen Plus 8.6
(Aspen One, 2015). The effect of pressure and the solvent to feed ratio
(S/F) on process performance were evaluated by analyzing the yield,
recovery of hydrocarbons, olefins, oxygenates, carboxylic acids in the
top and bottom streams, as well as the feasibility to obtain deacidify
OLP fractions with lower content of olefins.
2. Modeling and simulation methodology
2.1. Thermodynamic modeling
2.1.1. Prediction of thermo-physical properties of organic liquid products
compounds
Predictive methods were selected considering their applicability to
describe the chemical structure of molecules and ease of use, including
effects of the molecular weight of hydrocarbons. To perform a com-
parison and evaluation of all the methods applied to predict the thermo-
physical properties of organic liquid products compounds described in
Table 1, experimental data of normal boiling point (Tb) and critical
properties from Yaws [56], as well as vapor pressure data from Boublik
[57], were used.
Based on the composition of OLP described in Table 1, experimental
data of critical properties available in the literature were selected to the
following hydrocarbon functions: normal alkanes from C2 to C30, cyclic
from C3 to C17, alkenes with only one unsaturation C4 to C20 and aro-
matics from C6 to C15 [58]. The predictive methods by Joback and Reid
[59], Constantinou and Gani [60], Marrero-Marejón and Pardillo-
Fontdevila [61] and Marrero and Gani [62] were applied to estimate
the normal boiling temperature and critical properties of all the com-
pounds present in the OLP. In addition, the methods by Costantinou
et al. [63], Han and Peng [64], were applied to predict/estimate the
acentric factor directly, as well as an indirect method in which the
acentric factor is calculated using its definition from vapor pressure
data, based on Araújo and Meireles [65]. In this case, the correlation by
E.C. Costa et al. The Journal of Supercritical Fluids 140 (2018) 101–115

Table 1
Experimental composition of organic liquid products (OLP), expressed in mass fraction, used to predict the physical (ρ) and thermo-physical properties (Tb, Tc, Pc, Vc),
and the acentric factor ω of all the compounds present in OLP, obtained by thermal catalytic cracking of palm oil at 450 °C, 1.0 atm, with 10% (wt.) Na2CO3 as
catalyst [58].
Cas number ωi MW Tb ρ Tc Pc Vc ω [−]
[−] [°C] [kg/m3] [°C] [bar] [m3/kmol]

Hydrocarbons 0.8924
Normal parafins 0.4193
n-undecane 1120-21-4 0.0641 156.31 195.93 796.14 365.85 19.50 0.6850 0.5303
n-dodecane 112-40-3 0.0315 170.34 216.32 806.72 384.85 18.20 0.7550 0.5764
n-tridecane 629-50-5 0.0526 184.37 235.47 815.94 401.85 16.80 0.8260 0.6174
n-tetradecane 629-59-4 0.0925 198.39 253.58 824.20 419.85 15.70 0.8970 0.6430
n-pentadecane 629-62-9 0.1259 212.42 270.69 831.72 434.85 14.80 0.9690 0.6863
n-hexadecane 544-76-3 0.0274 226.45 286.86 838.67 449.85 14.00 1.0400 0.7174
n-heptadecane 629-78-7 0.0132 240.47 302.15 845.13 462.85 13.40 1.1100 0.7697
n-octadecane 593-45-3 0.0121 254.50 316.71 851.24 473.85 12.70 1.1900 0.8114
Olefins 0.2534
1-decene 872-5-9 0.0469 140.27 170.60 781.54 343.45 22.23 0.5840 0.4805
1-dodecene 112-41-4 0.0985 168.32 213.00 805.05 383.95 18.88 0.7250 0.5620
1-tridecene 2437-56-1 0.0019 182.35 232.84 814.71 401.65 17.49 0.7950 0.6074
1-tetradecene 1120-36-1 0.0052 196.38 251.10 823.09 417.85 16.27 0.8650 0.6503
1-pentadecene 13360-61-7 0.0043 210.40 268.46 830.76 434.85 15.70 0.9360 0.6815
1-heptadecene 6765-39-5 0.0331 238.46 300.33 844.36 462.85 14.10 1.0800 0.7503
1-nonadecene 18435-45-5 0.0082 266.51 329.02 856.39 486.85 12.80 1.2260 0.8325
(Z)-2-Decene 20348-51-0 0.0047 140.27 173.85 783.52 347.85 21.80 0.5900 0.4752
(2E)-2-Dodecene 7206-13-5 0.0038 168.32 217.85 807.48 389.85 18.90 0.7260 0.5693
(E)-9-octadecene 7206-25-9 0.0061 252.48 319.51 852.41 484.85 11.18 1.1138 0.6525
1,4 - undecadiene 55976-13-1 0.0056 152.28 190.15 794.89 360.95 20.88 0.6125 0.5422
(E) -5-Eicosene 74685-30-6 0.0008 280.50 323.90 854.27 496.80 10.62 1.1440 0.8933
3-hexadeceno 34303-81-6 0.0234 224.45 276.03 834.03 447.50 13.72 0.9115 0.7120
6-dodecyne 6975-99-1 0.0109 166.31 212.75 804.93 390.85 19.51 0.6894 0.5258
Naphthenics 0.2197
Isobutyl cyclohexane 1678-98-4 0.0059 140.27 171.30 781.97 368.95 25.85 0.5190 0.3536
n-butyl cyclohexane 1678-93-9 0.0038 140.27 180.98 787.75 393.85 25.70 0.5340 0.2743
Cyclododecane 294-62-2 0.0048 168.32 243.99 819.88 481.85 28.79 0.6053 0.3512
n-decyl cyclopentane 1795-21-7 0.0594 210.40 279.35 835.46 450.65 15.20 0.8790 0.6540
n-nonyl cyclopentane 2882-98-6 0.0364 196.38 262.15 828.00 437.35 16.52 0.7774 0.6100
n-nonyl cyclohexane 2883-02-5 0.0517 210.40 281.19 836.24 460.85 16.26 0.8733 0.6138
1-Ethyl-2-heptyl cyclopropane 74663-86-8 0.0048 168.30 213.90 805.50 391.80 20.14 0.6719 0.5346
1-Pentyl-2-propyl cyclopropane 41977-33-7 0.0197 154.30 195.70 796.02 372.90 21.98 0.6156 0.5239
1-butyl-1-cyclohexene 3282-53-9 0.0072 138.25 180.58 787.52 376.85 27.20 0.5296 0.4151
1-hexyl-1-cyclopentene 4291-99-0 0.0057 152.28 204.87 800.89 389.85 23.86 0.6052 0.5262
1-Heptyl-1-cyclohexene 15232-86-7 0.0032 180.30 237.30 816.80 385.70 18.59 0.6515 0.9056
1-nonyl-1-cyclohexene 15232-88-9 0.0144 208.40 266.19 829.77 429.86 15.89 0.7955 0.5813
1-decylcy-1-clohexene 62338-41-4 0.0027 222.42 278.53 835.11 574.30 14.54 1.1010 0.7168
Oxygenated 0.1076
Carboxylic acids 0.0263
Palmitic Acid 57-10-3 0.0175 256.43 339.00 860.56 507.23 15.10 1.0150 1.1750
Oleic Acid 112-80-1 0.0088 282.50 353.66 866.69 579.25 13.88 1.0570 1.0787
Alcohols 0.0351
O-cresol 95-48-7 0.0014 108.14 191.00 793.44 424.40 50.10 0.2820 0.4339
(Z)-9-Hexadecen-1-ol 143-28-2 0.0337 268.48 349.64 865.01 517.85 13.15 1.1286 0.8107
Ketones 0.0462
2-pentadecanone 2345-28-0 0.0462 226.40 309.79 848.34 476.85 15.57 0.8815 0.8121

Vetere [66] was used, that makes it possible to estimate the vapor
pressure based on the molecular structure of a pure substance. The Table 2
same criteria used by Melo et al. [67] and Araújo and Meireles [65], Experimental gaseous-liquid equilibrium data for the binary systems organic
were applied to decide on the best methods to predict the thermo- liquid products compounds-i/CO2 used to compute the binary interaction
physical properties (Tb, Tc, Pc, Vc) and the acentric factor (ω). These parameters of RK-Aspen equation of state (EOS) [68–75].
criteria are based on statistical analysis (measurements of central ten- CO2 + Points Temperature Pressure Data
dency and dispersion). [K] [bar]

Decane 29 319.11–372.94 34.85–160.60 Jimenez-Gallegos

2.1.2. Correlation of phase equilibrium data for the binary system organic et al. [68]
liquid products compounds-i/CO2 Undecane 18 314.98–344.46 23.73–133.88 Camacho-Camacho
The thermodynamic modeling applied to describe the organic liquid et al. [69]
products (OLP) fractionation in a multistage countercurrent absorber/ Tetradecane 2 344.28 155.54–162.99 Gasem et al. [70]
Pentadecane 22 293.15–353.15 5.60–139.40 Secuianu et al. [71]
stripping column using carbon dioxide as solvent, was performed by the
Hexadecane 12 314.14–333.13 80.65–148.70 D’Souza et al. [72]
Redlich-Kwong Aspen (RK-Aspen) equation of state (EOS). The RK- Octadecane 12 534.86–605.36 10.16–61.90 Kim et al. [73]
Aspen equation of state was applied to correlate the binary systems Palmitic acid 10 423.20–473.20 10.10–50.70 Yau et al. [74]
organic liquid products compounds-i/CO2 available in the literature, as Oleic acid 16 313.15–353.15 101.70–300.20 Bharath et al. [75]
described in Table 2. The RK-Aspen EOS with the van der Waals mixing

103
E.C. Costa et al. The Journal of Supercritical Fluids 140 (2018) 101–115

rules and RK-Aspen combining rules for two temperature-independent of experimental phase equilibrium data described in Table 2. The pro-
binary interaction parameters are given by the following equations: gram provides several options for how regression is computed, in-
cluding several different types of objective functions. The default ob-
RT a (T )
P= − jective function is the Maximum likelihood objective function, given by
V −b V (V + b) (1)
Eq. (14). To obtain the binary interaction parameters in Aspen Prop-
0.42748(R2Tc2 ) erties®, the following procedure is used, regardless of the type of system
a= × α (mi , ηi , Tri ) and model to which the data will be correlated.
Pc (2)

0.08664(RTc ) 1 Choice of components;

b=
Pc (3) 2 Specification of the method (where the model applied for the re-
gression of the experimental data is chosen);
1
3 Introduction or choice of experimental data (T-xy, P-xy, TP-x, T-x,
α (mi, ηi , Tri) = ⎡1 + mi ⎛1−Tri2 ⎞−ηi (1−Tri )(0.7−Tri ]2
⎜ ⎟
⎢ ⎝ ⎠ (4) TP-xy, T-xx, P-xx, TP-xx, TP-xxy, etc.) depending on the type and
⎣
information of the system; at this stage it is possible to either search
mi = 0,48508 + 1.55171ωi−1.15613ωi2 (5) for the compounds from the Aspen Properties® data base or enter
experimental data manually;
a= ∑ ∑ xi xj aij (6) 4 Regression of data: In this step the type of parameter, the para-
meters (according to the coding of the program) to be adjusted, the
b= ∑ ∑ xi xj bij (7) initial estimate and the limits for the regression are chosen.
1
aij = (aii ajj ) 2 (1−kij ) (8) 2.2. Process simulation

(bii bjj ) The organic liquid products (OLP) fractionation in multistage

bij = (1−lij )
2 (9) countercurrent absorber/stripping columns, using carbon dioxide as
solvent, was simulated using Aspen Hysys 8.6 (Aspen One, 2015). The
a= ∑ ∑ xi xj ai aj (1−kaij )
simulations carried out by selecting the multistage countercurrent ab-
i j (10)
sorber/stripping unit operation considering the dissolution/solubiliza-
bii + bjj tion of gases (CO2) in liquids (OLP), and that a fraction of the coexisting
b= ∑ ∑ xi xj 2
(1−kbij ) liquid phase dissolves in the gas phase, like a gas (CO2)-liquid (OLP)
i j (11)
equilibrium behavior. In the multistage countercurrent absorber/strip-
0 1 T ping column, the fraction collected at the top stream is the extract (light
kaij = kaij + kaij
1000 (12) phase) and the fraction collected at the bottom stream is the raffinate
(heavy phase).
0 1 T The Aspen Hysys 8.6 (Aspen One, 2015) executes the simulations
kbij = kbij + kbij
1000 (13)
for countercurrent multistage countercurrent column unit operation by
0 0 applying the inside-out algorithm to solve simultaneously the materials
Where kaij = and kbij = are the binary interaction parameters,
kaij kbij
considered as temperature-independent. The RK-Aspen binary interac- and energies balances through the columns, coupled with a rigorous
tion parameters were obtained using the Aspen Properties computa- thermodynamic model (MESH equations), for the operating conditions
tional package from Aspen Plus. The program uses the Britt-Lueck al- described in Table 3. The RK-Aspen equation of state, with the binary
gorithm, with the Deming parameters initialization method, to perform interaction parameters adjusted, as described in Table 4, was selected
a maximum like-hood estimation of the following objective function: as fluid package to compute the mixture properties for the multi-
component mixture organic liquid products/CO2.
exp−T calc 2 exp−P calc 2 exp−x calc 2
OF = ∑ ⎛T ⎜
⎞ +
⎟ ∑ ⎛P ⎜
⎞ +
⎟ ∑ ⎛x⎜
⎞
⎟

i ⎝ σT ⎠ i ⎝ σP ⎠ i ⎝ σx ⎠ 2.2.1. Simulation methodology and procedures

2 In general, the process simulations using Aspen-Hysys 8.4 consist of
y exp −y calc
+ ∑ ⎜⎛ ⎞
⎟ the following steps/and or procedures: 1. Selection of component list
i ⎝ σy ⎠ (14)
type: 2. Formulation of the list of components/Registration of hy-
Where, xi, yi are the compositions of compound i in the coexisting liquid pothetical components: 3. Basis formulation/Choose of thermodynamic
and vapor phases, respectively, and σ the standard deviations, applied package/Input of binary interaction parameters: 4. Flowsheet assembly
to the state conditions (T, P) and compositions (xi, yi), observed ex- of the process. In order to perform the simulations, initially the hy-
perimentally. The average absolute deviation (AAD) was calculated to drocarbons compounds identified in OLP composition were represented
evaluate the agreement between the binary systems measured experi- in the Aspen-Hysys 8.6 database library as hypothetical compounds,
mental data and the calculated results for each phase.
Table 2 shows the experimental gaseous-liquid equilibrium data for Table 3
the binary systems organic liquid products compound-i/CO2 used to Simulation process conditions of organic liquid products fractionation in
compute the binary interaction parameters. The estimated/predicted multistage countercurrent absorber/stripping columns using CO2 as solvent.
values for all the physical (ρ), thermo-physical properties (Tb, Tc, Pc, Operating Conditions
Vc), and acentric factor ω of organic liquid products compounds de-
scribed in Table 1, were those predicted by the best group contribution Column T-100 T = 333 K
Feed (OLP) = 100 [kg/h] P = 140, 180 bar
methods based on the minimization of central tendency and dispersion. Numbers of stages = 10 (S/F) = 12, 15, 17, 25
For the binary pairs organic liquid products compounds-i/CO2 not
0 0 Column T-102 T = 333 K
available in the literature, kaij = kaij and kbij = kbij were set equal to zero
Numbers of stages = 10 P = 140 bar
inside the matrix of binary interaction parameters. Feed = Raffinate 1 (S/F) = 30, 38
The Aspen Properties® package program was used for the regression

104
E.C. Costa et al. The Journal of Supercritical Fluids 140 (2018) 101–115

Table 4 3. Results and discussions

Optimized binary parameters of RK-Aspen-EOS for the binary systems hydro-
carbons-i/CO2 and carboxylic acids-i/CO2. 3.1. Prediction of thermo-physical properties of organic liquid products
CO2 + T [K] Kaij0 Kbij0 AADx AADy compounds

Undecane 314.98 0.116458 −0.008014 0.0029 0.0003 1

The methods selected to estimate/predict the physical (ρ), thermos-
344.46 0.103282 −0.029465 0.0030 0.0036
physical properties (Tb, Tc, Pc, Vc), and the acentric factor ω of all the
Tetradecane 344.28 0.099874 −0.000546 0.0003 0.0011
Pentadecane 313.15 0.093344 0.026454 0.0125 0.0020 compounds present in organic liquid products, described in Table 1.
333.15 0.101805 0.014904 0.0067 0.0030
Hexadecane 314.14 0.083111 −0.075317 0.0117 0.0090 3.2. Thermodynamic modeling of phase equilibrium data for the binary
333.13 0.082146 −0.081056 0.0013 0.0034
system organic liquid products compounds-i/CO2
Octadecane 534.86 0.246616 0.073306 0.0010 0.0024
605.36 0.107125 0.015525 0.0002 0.0064
Palmitic acid 423.20 −0.179556 −0.042625 0.0037 8,93E-05 Table 4 presents the RK-Aspen EOS temperature-independent binary
473.20 −0.059218 −0.013329 0.0008 0.0002 interaction parameters adjusted with experimental phase equilibrium
Oleic acid 313.15 0.110902 0.132527 0.0039 0.0031 data for the binary systems hydrocarbons-i/CO2 and carboxylic acids-i/
333.15 0.116604 0.054485 0.0039 0.0035
CO2, as well as the absolute mean deviation (AAD) between experi-
353.15 0.117892 0.049413 0.0049 0.0046
mental and predicted compositions for both coexisting liquid and gas-
eous phases. The regressions show that RK-Aspen EOS was able to de-
scribe the high pressure gaseous-liquid experimental phase equilibrium
data for all the systems under investigation with AAD between 0.8%
using the physical and thermo-physical properties estimated/predicted and 1.25% to liquid phase and 89.10−5% to 0.9% to gaseous phase.
in Table 1. The selected thermo-physical properties of palmitic and
oleic acids were the same presented by Yau et al. [74] and Bharath et al. 3.3. Process simulations of OLP fractionation in multistage countercurrent
[75]. Then, the thermodynamic model chosen, and the fractionation of absorber/stripping columns using CO2 as solvent with Aspen-Hysys
organic liquid products (OLP) in multistage countercurrent absorber/
stripping columns using SC-CO2 as solvent, simulated. Afterwards, the The fractionation of organic liquid products (OLP), obtained by
column top and bottom streams were depressurized by using expansion thermal catalytic cracking of palm oil at 450 °C, 1.0 atm, with 10% (wt.)
valves and heat exchanges to separate the OLP fractions from carbon NaOH as catalyst, was simulated in countercurrent multistage ab-
dioxide within flash drums. sorber/stripping columns with 10 theoretical stages, numbered from
top (1) to bottom (10), using CO2 as solvent with Aspen-Hysys chemical
process simulator, for all operating conditions described in Table 3. The
2.2.2. Process flowsheet strategy RK-Aspen EOS with temperature-independent binary interaction para-
Based on preliminary simulations, two (02) process flowsheets were meters presented in Table 4 was used as fluid package, and the 02 (two)
proposed and/or analyzed. Flowsheet I consists of a classical flowchart different process flowsheets (I, II) presented in Figs. 1 and 2.
and includes one (01) multistage countercurrent absorber/stripping
column, two (02) flash drums and expansions valves, and two (02) heat 3.3.1. Simulations of process flowsheet I
exchanges. The proposed flowsheet I (column T-100) shows that, the Tables 5–8 illustrate the material and energy balances of organic
top stream 1, containing a fraction of initial OLP feed, was deacidified liquid products fractionation in multistage countercurrent absorber/
at 140 bar. In this context, to achieve a complete deacidification of OLP, stripping columns using CO2 as solvent for all the process streams of
the bottom stream 2 was used as feed in a second multistage counter- Flowsheet I, expressed in CO2 basis, at 333 K, 140 and 180 bar, (S/
current absorber/stripping column using SC-CO2 as solvent by applying L) = 12, 15, 17, and 25. Table 9 presents the simulations results of
the following strategy. Simulations were carried out within a second organic liquid product fractionation process for flowsheet I, as de-
multistage countercurrent absorber/stripping column (flowsheet II), scribed in Fig. 1. The top and bottom streams mass flows rates of
maintaining the same state conditions of column T-100 (P, T), by using column 1, expressed in solvent free basis, obtained by depressurization
the bottom stream 2 from column T-100, named here Raffinate 1, di- using flash drums as follows: the top stream 15 (V-103) and the bottom
rectly as feed, in top stream (Raffinate 1+ CO2) and CO2 make up as the stream 9 (V-101).
feed, in bottom stream of column T-102. Raffinate 1, assumes the same Figs. 3 and 4 illustrate the hydrocarbons and oxygenates con-
mass flow rate and composition of flash drum (V-101) stream 9 of centrations, in solvent free basis, computed at the top stream of column
flowsheet I. T-100, for all the simulations of process flowsheet I, at 333 K, 140 and
The effect of solvent-to-feed (S/F) ratios on the processes perfor- 180 bar, and (S/F) = 12, 15, 17, 25. It can be seen that the best dea-
mance were evaluated based on yields of hydrocarbons and oxygenates, cidifying operating condition, that is, the highest hydrocarbons and
as well as the recovery of hydrocarbons, olefins, oxygenates and car- lowest oxygenates composition at top stream 1 (Column T-100) of
boxylic acids in OLP fractions process streams. In addition, the com- flowsheet I, in CO2 free basis, of organic liquid products fractionation in
position profile of all the hydrocarbons and oxygenates chemical multistage countercurrent absorber/stripping columns using CO2 as
functions along the columns stages were analyzed. The yield and re- solvent was observed at 333 K, 140 bar, and (S/F) = 17. The higher the
covery of process stream i are computed as follows: pressure, maintaining the temperature constant, the higher the
m˙ i
Yieldi [%] =
m˙ OLP (15) 1
For n-alkanes, the method of Marrero-Gani selected for the prediction of Tc, Pc and Vc,
and that of Han-Peng for ω. For alkenes, the method of Marrero-Gani selected for the
prediction of Tb and Tc, Marrero-Pardillo for Pc and Vc, and Han-Peng for ω. For un-
ωj *˙mi substituted cyclic hydrocarbons the method of Constantinou-Gani selected for the pre-
Recoveryj [%] =
˙ *mOLP
ωj, OLP (16) diction of Tb, Marrero-Gani for Tc, Joback for Pc and Vc, and the undirected method of
Vetere for ω. For substituted cyclic hydrocarbons the method of Constantinou-Gani se-
lected for the prediction of Tb and Pc, Marrero-Gani for Tc and Vc, and the undirected
Where ṁi and ṁOLP are the mass flow rates of process stream i and OLP method of Vetere for ω. For aromatic hydrocarbon the method of Joback selected for the
feed stream, and ωj and ωj, OLP the mass fractions of hydrocarbons or prediction of Tb, Constantinou-Gani for Tc and Vc, Marrero-Gani for Pc, and the un-
oxygenates chemical functions of process stream i and OLP feed stream. directed method of Vetere for ω.

105
E.C. Costa et al. The Journal of Supercritical Fluids 140 (2018) 101–115

Fig. 1. Flowsheet I [absorber/stripping column 1 (T-100); flash drums (V-100, V-101, V-102, V-103); valves (VLV-100, VLV-101, VLV-102, VLV-103); heat exchanges
((E-100, E-101)] of organic liquid products fractionation in multistage countercurrent absorber/stripping columns using CO2 as solvent.

Fig. 2. Flowsheet II [absorber/stripping columns 1 and 2 (T-100, T-102); flash drums (V-100, V-101, V-102, V-103, V-100-2, V-101-2); valves (VLV100, VLV101,
VLV102, VLV103, VLV-100-2, VLV101-2); heat exchanges (E-100, E-101, E-102, E-100-2)] of organic liquid products fractionation in multistage countercurrent
absorber/stripping columns using CO2 as solvent (Raffinate 1 + CO2 as feed top stream and CO2 make up as feed bottom stream).

dissolution and/or the solubility of OLP in supercritical CO2. The higher
the dissolution and/or the solubility of OLP in supercritical CO2, the
lower the selectivity, being the results according to experimental data
of multistage gas extraction in countercurrent columns reported in the
literature [41,44,45]. In addition, higher operating pressures and
higher solvent-to-feed ratios are associated to higher energy
requirements not only to compression/depressurization of CO2 (CO2
cycle), but also to higher thermal energy necessary to heat the solvent
after depressurization, before entering the compressor. In this context,
all the simulations results for the column T-100 described in this work
are related to the operating conditions at 333 K, 140 bar, and (S/
F) = 17, by considering it was the best deacidifying operating

106
E.C. Costa et al. The Journal of Supercritical Fluids 140 (2018) 101–115

Table 5
Material and energy balances of organic liquid products fractionation in multistage countercurrent absorber/stripping columns using CO2 as solvent (Flowsheet I),
expressed in CO2 basis, at 333 K, 140 bar, (S/L) = 17.
Stream Nº ωi,Gas T P Molar Flow Mass Flow Liquid Vol. Flow Heat Flow ωCO2 Q-100 Q-101 Q-102
[ºC] [bar] [kmol/h] [kg/h] [m³/h] [kcal/h] [−] [kcal] [kcal] [kcal]

PLO 0 60 140 0.5196 100 0.1214 −43,708.4 5832 18957 590.8

CO2 1 60 140 38.4232 1691 2.0578 −3,672,925
1 1 147.94 140 20.0574 910.0584 1.1081 −1,872,018 0.95972
2 0.92522 77.78 140 18.8854 880.9416 1.0712 −1,794,359 0.92808
3 1 20.26 50 16.4788 725.2752 0.8826 −1,561,944
4 0 20.26 50 2.4066 155.6663 0.1885 −232,415
5 0.87256 20.26 50 18.8854 880.9416 1.0712 −1,794,359
6 0.69074 78.40 1.5 2.4066 155.6663 0.1885 −232,415
7 0.86819 25 1.5 2.4066 155.6663 0.1885 −226,583
8 1 25 1.5 2.0894 91.9792 0.1119 −196,537
9 0 25 1.5 0.3172 63.6871 7.66E-02 −30,045.5
10 0.98826 107.40 45 20.0574 910.058 1.1081 −1,872,018
11 1 38 45 19.5051 858.818 1.0451 −1,841,585
12 0 38 45 0.5523 51.2402 6.30E-02 −49,390.1
13 0.60206 −5.60 1.5 0.5523 51.2402 6.30E-02 −49,390.1
14 1 25 1.5 0.3379 14.8920 1,81E-02 −31,785.6
15 0 25 1.5 0.2143 36.3482 4.48E-02 −17,013.7
17 0.97246 38 45 20.0574 910.0584 1.1081 −1,890,975
18 0.61191 25 1.5 0.5523 51.2402 6.30E-02 −48,799.3

Table 6
Material and energy balances of organic liquid products fractionation in multistage countercurrent absorber/stripping columns using CO2 as solvent (Flowsheet I),
expressed in CO2 basis, at 333 K, 140 bar, (S/L) = 25.
Stream Nº ωi,Gas T P Molar Flow Mass Flow Liquid Vol. Flow Heat Flow ωCO2 Q-100 Q-101 Q-102
[ºC] [bar] [kmol/h] [kg/h] [m³/h] [kcal/h] [−] [kcal] [kcal] [kcal]

PLO 0 60 140 0.5196 100 0.12147 −43,708.4

CO2 1 60 140 56.8055 2500 3.0423 −5,430,108
1 1 74.93 140 53.6042 2403.322 2.9252 −5,094,310 0.97568
2 0.80782 105,66 140 3.7209 196.678 0.2385 −350,509 0.78865
3 1 64.93 50 3.3325 146.7319 0.1785 −313,686
4 0 64.93 50 0.3884 49.9460 6.00E-02 −36,822.1
5 0.89560 64.93 50 3.7209 196.678 0.2385 −350,509
6 0.48339 46.36 1.5 0.3884 49.9460 6.00E-02 −36,822.1
7 0.47873 25 1.5 0.3884 49.9460 6.00E-02 −37,277.1
8 1 25 1.5 0.1859 8.1859 9.96E-03 −17,492.6
9 0 25 1.5 0.2024 41.7601 5.00E-02 −19,784.5
10 0,86679 16.79 50 53.6042 2403.322 2.9252 −5,094,310
11 1 16.79 50 46.4636 2045.047 2.4887 −4,407,324
12 0 16.79 50 7.1405 358.2754 0.4365 −686,986
13 0.55926 −81.51 1.5 7.1405 358.2754 0.4365 −686,986
14 1 25 1.5 6.8118 300.0567 0.3651 −640,657
15 0 25 1.5 0.3287 58.2186 7.14E-02 −272,46.9
17 0.95396 25 1.5 7.1405 358.2754 0.4365 −667,903

condition. solvent-to-feed ratios are below the approximate theoretical relation

Based on the composition of OLP described in Table 1, consisting
basically of alkanes, alkenes, and naphthenic compounds, as well as the
choice of CO2 as solvent, the selected state (333 K, 140 bar) and oper-
ating conditions (S/F = 17) are according to high pressure phase
equilibrium data available in the literature for the systems alkanes/
CO2, including n-undecane/CO2 [69], tetradecane/CO2 [70], pentade-
cane/CO2 [71,76], hexadecane/CO2 [72,76], and decane/CO2 [77],
whose phase envelop at 313 K close between 160 and 200 bars, thus
meaning the organic liquid products fractionation in multistage coun-
tercurrent absorber/stripping columns using CO2 as solvent operates
within the two phase region for the selected state conditions (333 K,
140 bar) [78], even though the phase envelop for the systems naph-
thenics/CO2, cyclopentane/CO2 [79,80], cyclohexane/CO2 [81], me-
thylcyclohexane/CO2 [82], n-propylcyclohexane/CO2 [73], are below
100 bars for T = 333 K, as described by Vitu et al. [83]. In addition, by
analyzing the phase equilibrium data for system n-undecane/CO2 [69],
a key compound present in OLP, at T = 344 K, a temperature close to
333 K, one may observe in the gaseous phase a yn-undecane ≈ 0.97,
giving a theoretical solvent-to-feed ratio (S/F) ≈ 32, so that selected
given by phase equilibrium data.
The simulation carried out for (S/F) = 17 presented a top stream
yield of 36.65% with 96.95% (wt.) hydrocarbons and 3.05% (wt.)
oxygenates, in solvent free basis, showing a decrease on the oxygenates
content in OLP feed from 10% (wt.), with 2.63% (wt.) carboxylic acids,
to 3.05% (wt.) oxygenates, with 0.52% (wt.) carboxylic acids, as well as
an enrich of hydrocarbons in OLP in top stream of 8.64% related to the
initial concentration of OLP in feed. The bottom stream with a yield of
63.35%, presents a OLP fraction with 84.78% (wt.) hydrocarbons and
15.22% (wt.) oxygenates, containing 3.85% (wt.) carboxylic acids and
the lowest content of olefins, equal to 18.95% (wt.).
Table 9 and Figs. 5 and 6 illustrate the simulation results for the
recovery of OLP fractions (Flowsheet I, column T-100, Top Stream 1,
Bottom Stream 1), in CO2 free basis, of organic liquid products frac-
tionation in multistage countercurrent absorber/stripping columns
using CO2 as solvent, at 333 K, 140, and (S/F) = 17. The highest hy-
drocarbons recovery, equal to 60.19%, for (S/F) = 17, was obtained in
the bottom stream of column T-100, showing an olefin recovery of
47.37%, an oxygenates recovery of 89.60%, and carboxylic acids

107
E.C. Costa et al. The Journal of Supercritical Fluids 140 (2018) 101–115

Table 7
Material and energy balances of organic liquid products fractionation in multistage countercurrent absorber/stripping columns using CO2 as solvent (Flowsheet I),
expressed in CO2 basis, at 333 K, 180 bar, (S/L) = 12.
Stream Number ωi,Gas T P Molar Flow Mass Flow Liquid Vol. Flow Heat Flow ωCO2 Q-100 Q-101 Q-102
[ºC] [bar] [kmol/h] [kg/h] [m³/h] [kcal/h] [−] [kcal] [kcal] [kcal]

PLO 0 60 180 0.5196 100 0.1214 −43,584.4 12,878.8 1135.3

CO2 1 60 180 27.2666 1200 1.4603 −2,614,316
1 6.43E-03 88.09 180 2.3436 127.0283 0.1547 −220,949 0.75363
2 1 81.53 180 25.4426 1172.972 1.4271 −2,422,068 0.94142
3 1 25.59 60 20.0853 884.162 1.0759 −1,906,755
4 0 25.59 60 5.35733 288.8096 0.3511 −515,313
5 0.78943 25.59 60 25.4426 1172.972 1.4271 −2,422,068
6 0.64675 −82.07 1.5 5.3573 288.8096 0.3511 −515,313
7 0.93294 25 1.5 5.3573 288.8096 0.3511 −502,434
8 1 25 1.5 4.9980 220.0829 0.2678 −470,104
9 0 25 1.5 0.3592 68.7267 8.33E-02 −32,330.5
10 0.70218 36.01 55 2.3436 127.0283 0.1547 −220,949
11 1 36.01 55 1.6456 72.4554 8.82E-02 −155682
12 0 36.01 55 0.6979 54.5729 6.65E-02 −65,267.4
13 0.72509 −36.54 1.5 0.6979 54.5729 6.65E-02 −65,267.4
14 1 25 1.5 0.5244 23.1003 2.81E-02 −49,325.6
15 0 25 1.5 0.1735 31.4726 3.84E-02 −14,806.4
17 0.751376 25 1.5 0.6979 54.5729 6.65E-02 −64,132

Table 8
Material and energy balances of organic liquid products fractionation in multistage countercurrent absorber/stripping columns using CO2 as solvent (Flowsheet I),
expressed in CO2 basis, at 333 K, 180 bar, (S/L) = 15.
Stream Nº ωi,Gas T P Molar Flow Mass Flow Liquid Vol. Flow Heat Flow ωCO2 Q-100 Q-101 Q-102
[ºC] [bar] [kmol/h] [kg/h] [m³/h] [kcal/h] [−] [kcal] [kcal] [kcal]

PLO 0 60 180 0.5196 100 0.1214 −43,584.4 32454 1389.7

CO2 1 60 180 34.0833 1500 1.8254 −3,267,895
1 1.81E-03 86.29 180 2.7233 148.5321 0.1808 −256,980 0.74779
2 1 68.66 180 31.8796 1451.468 1.7660 −3,047,260 0.95691
3 1 23.61 60 18.8016 827.5973 1.0071 −1,786,054
4 0 23.61 60 13.0780 623.8706 0.7589 −1,261,206
5 0.58976 23.61 60 31.8796 1451.468 1.7660 −3,047,260
6 0.60407 −81.56 1.5 13.0780 623.8706 0.7589 −1,261,206
7 0.97506 25 1.5 13.0780 623.8706 0.7589 −1,228,751
8 1 25 1,5 12.7519 561.5169 0.6833 −1,199,412
9 0 25 1.5 0.3260 62.3537 7.56E-02 −29,338.9
10 0.69247 34.92 55 2.7233 148.5321 0.1808 −25,6980
11 1 34.92 55 1.8858 83.0253 0.1010 −178,435
12 0 34.92 55 0.8374 65.5067 7.98E-02 −78,544.9
13 0.72794 −38.02 1.5 0.8374 65.5067 7.98E-02 −78,544.9
14 1 25 1.5 0.6319 27.8350 3.39E-02 −59,440.1
15 0 25 1.5 0.2055 37.6717 4.59E-02 −17,715
16 0.75461 25 1.5 0.8374 65.5067 7.98E-02 −77,155.1

Table 9 recovery of 92.70%, being necessary, because of the low yield of OLP in
Simulation of organic liquid products fractionation in multistage counter- top stream of column T-100, to insert a second multistage counter-
current absorber/stripping columns using CO2 as solvent (Flowsheet I, Column current absorber/stripping column to recover the OLP fraction in
T-100), expressed in mass fraction (CO2 free basis), at 333 K and 140 bar, (S/ bottom stream of column T-100. In addition, the simulations results for
F) = 17, from flash V-103 stream 15 (top) and flash V-101 stream 9 (bottom). OLP fraction in top stream of column T-100, at 333 K and 140 bar,
OLP Top Bottom presented chemical composition similar to those of kerosene-like frac-
tion obtained by fractional distillation of OLP, produced in pilot scale as
Mass flow rate [kg/h] 100.00 36.65 63.35
described in the literature [28]. The kerosene-like fraction presents
Yield [%] 36.65 63.35
ωI [−] (CO2 free basis)
93.63% (wt.) hydrocarbons (24.89% paraffins, 42.62% olefins, and
Hydrocarbons 0.8924 0.9695 0.8478 26.12% naphthenics) and 6.37% (wt.) oxygenates. The simulations for
Olefins 0.2534 0.3639 0.1895 (S/F) = 17 ratio, presenting a top stream yield of 36.65%, showed OLP
Oxygenates 0.1076 0.0305 0.1522 fraction with 96.95% (wt.) hydrocarbons (39.14% paraffins, 36.39%
Carboxylic acids 0.0263 0.0052 0.0385
olefins, and 21.42% naphthenics) and 3.05% (wt.) oxygenates, with
Alcohols 0.0351 0.0086 0.0505
Ketones 0.0462 0.0167 0.0632 0.52% (wt.) of carboxylic acids.
Recovery [%]
Hydrocarbons – 39.81 60.19
Olefins – 52.63 47.37
3.3.1.1. Composition profiles in multistage countercurrent absorber/
Oxygenates – 10.40 89.60 stripping column T-100. Figs. 10 and 11 present the results for the
Carboxylic acids – 7.30 92.70 simulated OLP composition profiles, expressed in weight fraction (CO2
Alcohols – 8.94 91.06 basis), in the coexisting liquid and gaseous phases, of organic liquid
Ketones – 13.28 86.72
products fractionation along with the multistage countercurrent
absorber/stripping column using CO2 as solvent, at 333 K, 140, and

108
E.C. Costa et al. The Journal of Supercritical Fluids 140 (2018) 101–115

Fig. 5. Recovery of OLP fractions (%) (Flowsheet I, Column T-100, Top Stream
Fig. 3. Hydrocarbons concentration (wt.%) (Flowsheet I, Column T-100, Top
1), in CO2 free basis, of organic liquid products fractionation in multistage
Stream 1), in CO2 free basis, of organic liquid products fractionation in multi-
countercurrent absorber/stripping columns using CO2 as solvent, at 333 K,
stage countercurrent absorber/stripping columns using CO2 as solvent, at
140 bar, and (S/F) = 17.
333 K, 140 and 180 bar, and (S/F) = 12, 15, 17, 25.

Fig. 6. Recovery of OLP fractions (%) (Flowsheet I, Column T-100, Bottom

Stream 1), in CO2 free basis, of organic liquid products fractionation in multi-
Fig. 4. Oxygenates concentration (wt.%) (Flowsheet I, Column T-100, Top stage countercurrent absorber/stripping columns using CO2 as solvent, at
Stream 1), in CO2 free basis, of organic liquid products fractionation in multi- 333 K, 140, and (S/F) = 17.
stage countercurrent absorber/stripping columns using CO2 as solvent, at
333 K, 140 and 180 bar, and (S/F) = 12, 15, 17, 25.
the top of column T-100, being the results according to experimental
data of multistage gas extraction in countercurrent columns reported in
(S/F) = 17. For the liquid phase, the composition profile of
the literature [41,44,45]. In fact, the more the light compounds are
hydrocarbons along with the column T-100 increases from 0.06 at the
dissolved in the gaseous phase (SC-CO2), the most the liquid OLP phase
bottom to 0.64 at the top, while that of olefins from 0.014 at the bottom
is enriched with heavier compounds, being the dissolution/
to 0.174 at the top, showing that pure CO2 entering at the bottom of
solubilization of the remaining heavier compounds lower in SC-CO2,
column T-100 is continuously dissolved within the OLP liquid phase
so that at every new stage from bottom to the top of column T-100, the
along with the column stages, being the CO2 dissolution in the OLP
OLP fraction in the gaseous phase decreases. The weight fraction (CO2
phase maximum at stage 1, that is, at the top of column T-100. The
basis) of OLP along with the column stages explains the lower OLP yield
weight fraction (CO2 basis) of OLP along with the column stages
fraction in the top stream of column T-100. In addition, one may
explains the higher OLP fraction yield in the bottom stream of
observe that the composition profiles of oxygenates decreases from
column T-100. In addition, one may observes that the composition
0.00586 to 0.00123, while that of carboxylic acids decreases from
profiles of carboxylic acids varies from 0.002 to 0.02, showing a smooth
0.00143 to 2.11088E-4, showing a smooth variation between stage 4
variation between stage 2 and 10 from 0.014 to 0.02, while that of
and 9, and the composition of olefins increases from 0.0086 to 0.01465.
oxygenates increases from 0.011 to 0.085 between stages 1 and 10. For
the gaseous phase, the composition profile of OLP fraction along with
the column T-100 decreases from 0.04483 at the bottom to 0.04027 at 3.3.2. Simulations of process flowsheet II
the top, showing that pure CO2 entering at the bottom of column T-100 The main idea behind the proposal by the simulation of process
dissolves the OLP along with the column stages, being the OLP flowsheet II is basically to use CO2 as a co-solvent by introducing
dissolution in the gaseous phase (CO2) minimum at stage 1, that is at (Raffinate 1 + CO2) from flash V-101 stream 9 (bottom) as feed top
stream in column T-102. Tables 10 and 11 illustrate the material and

109
E.C. Costa et al. The Journal of Supercritical Fluids 140 (2018) 101–115

Table 10
Material and energy balances of organic liquid products fractionation in multistage countercurrent absorber/stripping columns using CO2 as solvent (Flowsheet II, for
S/F = 17 in Column T-100), expressed in CO2 basis, at 333 K, 140 bar, (S/L) = 30.
Stream Nº ωi,Gas T P Molar Flow Mass Flow Liquid Vol. Flow Heat Flow ωCO2 Q-100 Q-101 Q-102 Q-103
[ºC] [bar] [kmol/h] [kg/h] [m³/h] [kcal/h] [−] [kcal] [kcal] [kcal] [kcal]

PLO 0 60 140 0.5196 100 0.1214 −437,08.4 2237.2 19,905.9 763.3 154.1
CO2 1 60 140 38.4232 1691 2.0578 −3672,925
1 1 147.94 140 20.0574 910.0584 1.1081 −1,872,018 0.95972
2 0.92522 77.78 140 18.8854 880.9416 1.0712 −1,794,359 0.92808
5 1 87.91 140 36.6942 1639.885 1.9955 −3,474,999 0.98028
6 0.92295 97.30 140 6.6619 318.0111 0.3862 −630,030 0.90246
CO2#2 1 77.78 140 24.4707 1076.954 1.3106 −2,324,576
3 1 13.83 40 35.7958 1575.429 1.9172 −3,389,196
4 0 13.83 40 0.8984 64.4557 7.84E-02 −85,802.8
5-2 0.97551 13.83 40 36.6942 1639.885 1.9955 −3,474,999
6-2 0.67472 −72.55 1.5 0.8984 64.4557 7.84E-02 −85,802.8
7 0.80897 25 1.5 0.8984 64.4557 7.84E-02 −83,565.5
8 1 25 1.5 0.7268 32.0003 3.89E-02 −68,364
9 0 25 1.5 0.1716 32.4554 3.94E-02 −15,201.6
10 0.98826 107.40 45 20.0574 910.0584 1.1081 −1,872,018
11 1 35 45 19.4630 856.909 1.0428 −1,838,423
12 0 35 45 0.5943 53.1494 6.53E-02 −53,500.7
13 0.62478 −14.02 1.5 0.5943 53.1494 6.53E-02 −53,500.7
14 1 25 1.5 0.3794 16.7190 2.03E-02 −35,684.9
15 0 25 1.5 0.2149 36.4304 4.49E-02 −17,052.4
17 0.97036 35 45 20.0574 910.0584 1.1081 −1,891,924
18 0.63841 25 1.5 0.5943 53.1494 6.53E-02 −52,737.3
3-2 1 45.22 50 6.3116 277.8009 0.3380 −595,764
4-2 0 45.22 50 0.3503 40.2102 4.82E-02 −34,266.3
5-3 0.94740 45.22 50 6.6619 318.0111 0.3862 −630,030
6-3 0.58290 15.35 1.5 0.3503 40.2102 4.82E-02 −34,266.3
7-2 0.58526 25 1.5 0.3503 40.2102 4.82E-02 −34,112.2
8-2 1 25 1.5 0.2050 9.0252 1.10E-02 −19,289.2
9-2 0 25 1.5 0.1453 31.1849 3.72E-02 −14,822.9

Table 11
Material and energy balances of organic liquid products fractionation in multistage countercurrent absorber/stripping columns using CO2 as solvent (Flowsheet II, for
S/F = 17 in Column T-100), expressed in CO2 basis, at 333 K, 140 bar, (S/L) = 38.
Stream Nº ωi,Gas T P Molar Flow Mass Flow Liquid Vol. Flow Heat Flow ωCO2 Q-100 Q-101 Q-102 Q-103
[ºC] [bar] [kmol/h] [kg/h] [m³/h] [kcal/h] [−] [kcal] [kcal] [kcal] [kcal]

PLO 0 60 140 0.5196 100 0.1214 −43,708.4 19.21 19905.9 763.34 612.44
CO2 1 60 140 38.4232 1691 2.0578 −3,672,925
1 1 147.94 140 20.0574 910.0584 1.1081 −1,872,018 0.95972
2 0.92522 77.78 140 18.8854 880.9416 1.0712 −1,794,359 0.92808
5 1 109.28 140 46.0238 2056.118 2.5020 −4,336,693 0.98077
6 0.95661 89 140 8.9835 414.5419 0.5037 −851,561 0.94256
CO2#2 1 77.78 140 36.1219 1589.718 1.9346 −3,431,363
3 1 46.47 40 45.6288 2008.652 2.4444 −4,300,445
4 0 46.47 40 0.3949 47.4658 5.76E-02 −36,248.2
5-2 0.99141 46.47 40 46.0238 2056.118 2.502041 −4,336,693
6-2 0.48241 24.02 1.5 0.3949 47.4658 5.76E-02 −36,248.2
7 0.48266 25 1.5 0.3949 47.4658 5.76E-02 −36,229
8 1 25 1.5 0.1906 8.3927 1.02E-02 −17,931.2
9 0 25 1.5 0.2043 39.0730 4.74E-02 −18,297.8
10 0.98826 107.40 1.5 20.0574 910.0584 1.1081 −1,872,018
11 1 35 45 19.4630 856.909 1.0428 −1,838,423
12 0 35 45 0.5943 53.1494 6.53E-02 −53,500.7
13 0.62478 −14.02 1.5 0.5943 53.1494 6.53E-02 −53,500.7
14 1 25 1.5 0.3794 16.7190 2.03E-02 −35,684.9
15 0 25 1.5 0.2149 36.4304 4.49E-02 −17,052.4
17 0.97036 35 45 20.0574 910.0584 1.1081 −1,891,924
18 0.63841 25 1.5 0.5943 53.14948 6.53E-02 −52,737.3
3-2 1 30.24 50 8.6057 378.7497 0.4609 −814,200
4-2 0 30.24 50 0.3778 35.7921 4.29E-02 −37,361
5-3 0.95794 30.24 50 8.9835 414.5419 0.503779 −851,561
6-3 0.69577 −23.01 1.5 0.3778 35.7921 4.29E-02 −37,361
7-2 0.71213 25 1.5 0.3778 35.7921 4.29E-02 −36,748.6
8-2 1 25 1.5 0.2690 11.8415 1.44E-02 −25,309.2
9-2 0 25 1.5 0.1087 23.9506 2.84E-02 −11,439.4

110
E.C. Costa et al. The Journal of Supercritical Fluids 140 (2018) 101–115

Table 12 (wt.%) in the top stream of column T-102 (Flowsheet II, Column T-102,
Simulation of organic liquid products fractionation in multistage counter- Top Stream 1, for S/F = 17 in Column T-100), in CO2 basis, of organic
current absorber/stripping columns using CO2 as solvent (Flowsheet II, liquid products fractionation in multistage countercurrent absorber/
Columns T-100 and T-102), (Raffinate 1 + CO2) as feed top stream and CO2 stripping columns using CO2 as solvent, (Raffinate 1 + CO2) as feed top
make up as feed bottom stream, expressed in mass fraction (solvent free basis),
stream and CO2 make up as feed bottom stream, at 333 K, 140 bar, and
at 333 K and 140 bar.
(S/F) = 30, 38. It can be seen that the best deacidifying operating
Column (T-100) Column (T-102) condition, regarding the OLP yield, hydrocarbons and oxygenates
composition, as well as hydrocarbons, oxygenates, and carboxylic acids
(S/F) 17 30 38
recoveries was observed at 333 K, 140 bar, and (S/F) = 38. The higher
OLP Bottom Top Bottom Top Bottom the solvent-to-feed ratio, maintaining the pressure and temperature
(Raffinate1) constants, the higher the amount of solute (OLP) carried by the gaseous
phase (SC-CO2), producing higher yields in the top stream of column T-
Mass Flow [kg/h] 100 63.35 32.33 31.02 40.15 23.20
102, being the results according to experimental data of multistage gas
ωI [−] (CO2 free basis)
Yield [%] 51.04 48.96 63.39 36.61 extraction in countercurrent columns reported in the literature
Hydrocarbons 0.8924 0.8478 0.9281 0.7641 0.9271 0.7106 [41,44,45]. The drawbacks of using higher solvent-to-feed ratios re-
Normal paraffins 0.4193 0.4354 0.4622 0.4075 0.4662 0.3823 mains on the higher energy requirements not only to compression/de-
Olefins 0.2534 0.1895 0.2318 0.1454 0.2187 0.1389
pressurization of CO2 (CO2 cycle), but also to higher thermal energy
Naphthenics 0.2197 0.2229 0.2341 0.2112 0.2422 0.1894
Oxygenates 0.1076 0.1522 0.0719 0.2359 0.0729 0.2894
necessary to heat the solvent after depressurization, before entering the
Carboxylic acids 0.0263 0.0385 0.0147 0.0633 0.0149 0.0794 compressor. On the other hand, using (Raffinate 1 + CO2) as feed top
Alcohols 0.0351 0.0505 0.0181 0.0842 0.0178 0.1071 stream in column T-102 makes it possible to operate the multistage gas
Ketones 0.0462 0.0632 0.0391 0.0884 0.0403 0.1030 extraction in countercurrent absorber/stripping columns with higher
Recovery [%]
solvent-to-feed ratios, as the density of liquid phase (Raffinate 1 + CO2)
Hydrocarbons – – 55.87 44.13 69.31 30.69
Olefins – – 62.44 37.56 73.17 26.83 is lower compared to the case the OLP fraction would be introduced as
Oxygenates – – 24.11 75.89 30.36 69.64 feed top stream in column T-102, so thus, even though phase equili-
Carboxylic acids – – 19.52 80.48 24.49 75.51 brium data for system n-undecane/CO2 [69], a key compound present
Alcohols – – 18.27 81.73 22.30 77.70 in OLP, at T = 344 K, a temperature close to 333 K, shows yn-undecane ≈
Ketones – – 31.56 68.44 40.36 59.64
0.97 in the gaseous phase, giving a theoretical solvent-to-feed ratio (S/
F) ≈ 32, the column may still operates in countercurrent mode without
flooding.
The simulation carried out for (S/F) = 38 presented a top stream
yield of 63.39% with 92.71% (wt.) hydrocarbons and 7.29% (wt.)
oxygenates, showing a decrease on the oxygenates content in OLP
fraction (Raffinate 1) feed from 15.22% (wt.), with 3.85% (wt.) car-
boxylic acids, to 7.29% (wt.) oxygenates, with 1.49% (wt.) carboxylic
acids, that is, a decrease of 52.10% and 61.30% on the oxygenates and
carboxylic acids contents, respectively. The bottom stream with a yield
of 36.61%, presents a OLP fraction with 71.06% (wt.) hydrocarbons
and 28.94% (wt.) oxygenates, showing an increase on the oxygenates
content in OLP fraction (Raffinate 1) feed from 15.22% (wt.), with
3.85% (wt.) carboxylic acids, to 28.94% (wt.) oxygenates, with 7.94%
(wt.) carboxylic acids, that is, an increase of 90.14% and 106.23% on
the oxygenates and carboxylic acids contents, respectively. In fact, by

Fig. 7. Top stream concentrations (wt.%) in column T-102 (Flowsheet II,

Column T-102, Top Stream 1, for S/F = 17 in Column T-100), in CO2 free basis,
of organic liquid products fractionation in multistage countercurrent absorber/
stripping columns using CO2 as solvent, (Raffinate 1 + CO2) as feed top stream
and CO2 make up as feed bottom stream, at 333 K, 140 bar, and (S/F) = 30, 38.

energy balances of organic liquid products fractionation in multistage

countercurrent absorber/stripping columns using CO2 as solvent for all
the process streams of Flowsheet II, expressed in CO2 basis, at 333 K,
140 bar, (S/L) = 30 and 38.
Table 12 presents the simulations results of organic liquid product
fractionation process for flowsheet II, as described in Fig. 2. The top
stream mass flows rate of column T-102, expressed in solvent free basis,
obtained by depressurization using flash drum V-101, split in top
stream 8 and bottom stream 9, while the bottom stream mass flows rate
of column T-102, expressed in solvent free basis, obtained by de- Fig. 8. Recovery of OLP fractions (%) in column T-102 (Flowsheet II, Column T-
pressurization using flash drum V-101-2, split in top stream 8-2 and 102, Top Stream 1, for S/F = 17 in Column T-100), in CO2 free basis, of organic
bottom stream 9-2. liquid products fractionation in multistage countercurrent absorber/stripping
columns using CO2 as solvent, (Raffinate 1 + CO2) as feed top stream and CO2
Fig. 7 illustrates the hydrocarbons and oxygenates concentrations
make up as feed bottom stream, at 333 K, 140 bar, and (S/F) = 30, 38.

111
E.C. Costa et al. The Journal of Supercritical Fluids 140 (2018) 101–115

1
Hydrocarbons
2 Olefins
Oxigenates
3
Carboxylic acids
4

Stage Number [-]

5

10
0.00 0.01 0.02 0.03 0.04 0.05 0.06
Gaseous Phase Composition [-]
Fig. 9. Recovery of OLP fractions (%) in column T-102 (Flowsheet II, Column T-
102, Bottom Stream 1, for S/F = 17 in Column T-100), in CO2 free basis, of Fig. 11. Simulated OLP composition profile, expressed in weight fraction, in
organic liquid products fractionation in multistage countercurrent absorber/ gaseous phase (Flowsheet I, Column T-100), in CO2 basis, of organic liquid
stripping columns using CO2 as solvent, (Raffinate 1 + CO2) as feed top stream products fractionation in multistage countercurrent absorber/stripping columns
and CO2 make up as feed bottom stream, at 333 K, 140 bar, and (S/F) = 30, 38. using CO2 as solvent, at 333 K, 140, and (S/F) = 17.

necessary to insert a third multistage countercurrent absorber/stripping

column to recover the OLP fraction in bottom stream 9-2 of column T-
102. The highest oxygenates recovery, equal to 75.89%, for (S/F) = 30,
was obtained in the bottom stream of column T-102, showing hydro-
carbons, olefins, oxygenates, and carboxylic acids recoveries of 44.13%,
37.56%, 75.89%, and 80.48%, respectively. In addition, the simulations
results for OLP fraction in top stream of column T-102, at 333 K and
140 bar, presented chemical composition similar to those of light diesel-
like fraction obtained by fractional distillation of OLP, produced in pilot
scale as described in the literature [28]. The light diesel-like fraction
presents 92.71% (wt.) hydrocarbons (46.62% paraffins, 21.87% olefins,
and 24.22% naphthenics) and 7.29% (wt.) oxygenates with a low
content of carboxylic acids, equal to 1.49% (wt.). In fact, the insertion
of a third multistage countercurrent absorber/stripping column to re-
cover the OLP fraction in bottom stream 9-2 of column T-102, will
probably produce a heavy diesel-like fraction, as the heavy hydro-
carbons compounds are dissolved in the gaseous phase (SC-CO2) at last.

Fig. 10. Simulated OLP composition profile, expressed in weight fraction, in

liquid phase (Flowsheet I, Column T-100), in CO2 basis, of organic liquid pro-
ducts fractionation in multistage countercurrent absorber/stripping columns
using CO2 as solvent, at 333 K, 140, and (S/F) = 17.

comparing the results of simulations performed at 333 K, 140 bar, and

(S/F) = 30 and 38 described in Table 7 and Fig. 7, regarding to the
compositions of OLP fractions, in solvent free basis, of bottom streams 9
and 9-2, one can observe that the key factor to decide the best deaci-
difying operating condition was the OLP fraction yield.
Table 12 and Figs. 8 and 9 illustrate the simulation results for the
recovery of OLP fractions (%) in column T-102 (Flowsheet II, Column T-
102, Top Stream 1, for S/F = 17 in Column T-100), in CO2 free basis, of
organic liquid products fractionation in multistage countercurrent ab-
sorber/stripping columns using CO2 as solvent, (Raffinate 1 + CO2) as
feed top stream and CO2 make up as feed bottom stream, at 333 K,
140 bar, and (S/F) = 30, 38.
The highest hydrocarbons recovery, equal to 69.31%, for (S/
F) = 38, was obtained in the top stream of column T-102, showing an Fig. 12. Simulated OLP composition profile, expressed in weight fraction, in
olefin recovery of 73.17%, an oxygenates recovery of 30.36%, and liquid phase (Flowsheet II, Column T-102, for S/F = 17 in Column T-100), in
carboxylic acids recovery of 24.49%, while the hydrocarbons, oxyge- CO2 basis, of organic liquid products fractionation in multistage countercurrent
nates, and carboxylic acids recoveries in the bottom stream of column absorber/stripping columns using CO2 as solvent, (Raffinate 1 + CO2) as feed
T-102 were 30.69%, 69.64%, and 75.51%, respectively, being top stream and CO2 make up as feed bottom stream, at 333 K, 140 bar, and (S/
F) = 30.

112
E.C. Costa et al.
Fig. 13. Simulated OLP composition profile, expressed in weight fraction, in
gaseous phase (Flowsheet II, Column T-102, for S/F = 17 in Column T-100), in
CO2 basis, of organic liquid products fractionation in multistage countercurrent
absorber/stripping columns using CO2 as solvent, (Raffinate 1 + CO2) as feed
top stream and CO2 make up as feed bottom stream, at 333 K, 140 bar, and (S/
F) = 30.
Fig. 14. Simulated OLP composition profile, expressed in weight fraction, in
liquid phase (Flowsheet II, Column T-102, for S/F = 17 in Column T-100), in
CO2 basis, of organic liquid products fractionation in multistage countercurrent
absorber/stripping columns using CO2 as solvent, (Raffinate 1 + CO2) as feed
top stream and CO2 make up as feed bottom stream, at 333 K, 140 bar, and (S/
F) = 38.
3.3.2.1. Composition profiles in multistage countercurrent absorber/
stripping column T-102. Figs. 12–15 present the results for the
simulated OLP composition profiles, expressed in weight fraction, in
liquid phase (Flowsheet II, Column T-102, for S/F = 17 in Column T-
100), in CO2 basis, of organic liquid products fractionation in
multistage countercurrent absorber/stripping columns using CO2 as
solvent, (Raffinate 1 + CO2) as feed top stream and CO2 make up as
feed bottom stream, at 333 K, 140 bar, and (S/F) = 30 and 38.
For the liquid phase with (S/F) = 30, the composition profile of
hydrocarbons along with the column T-102 increase from 0.074 at the
bottom to 0.41 at the top, while that of olefins increases from 0.0142 at
the bottom to 0.0883 at the top, showing that CO2 make up entering at
the bottom of column T-102 is instantly dissolved within the OLP liquid
phase along with the column stages from the stage 9, maintaining an
almost constant value between stages 8 and 2. Even though, the weight
113
The Journal of Supercritical Fluids 140 (2018) 101–115
Fig. 15. Simulated OLP composition profile, expressed in weight fraction, in
gaseous phase (Flowsheet II, Column T-102, for S/F = 17 in Column T-100), in
CO2 basis, of organic liquid products fractionation in multistage countercurrent
absorber/stripping columns using CO2 as solvent, (Raffinate 1 + CO2) as feed
top stream and CO2 make up as feed bottom stream, at 333 K, 140 bar, and (S/
F) = 38.
fraction (CO2 basis) of OLP along with the column stages is lower
compared to that in Flowsheet I, when increasing the solvent-to-feed
from 17 to 30 in column T-102 of Flowsheet II, explains the higher OLP
fraction yield in the top stream of column T-102, as greater amounts of
OLP fractions will be dissolved in the gaseous phase (SC- CO2). In ad-
dition, one may observes that the composition profiles of oxygenates
increases from 0.023 to 0.0884, showing a smooth variation between
stage 7 and 2 from 0.117 to 0.112, while that of carboxylic acids in-
creases from 0.0062 to 0.02297 between stages 1 and 10, showing also
a smooth variation between stage 8 and 2. For the gaseous phase, the
composition profile of hydrocarbons, olefins, oxygenates, and carbolic
acids decreases along with the column T-102 from 0.04023 at the
bottom to 0.0183 at the top, from 0.0072 at the bottom to 0.0046 at the
top, from 0.0101 at the bottom to 0.0014 at the top, and from 0.0026 at
the bottom to 0.0003 at the top, respectively. This explains the drastic
decrease of oxygenates and carbolic acids contents in the top stream of
column T-102, being the results according to experimental data of
multistage gas extraction in countercurrent columns reported in the
literature [41,44,45].
For the liquid phase with (S/F) = 38, the composition profile of
hydrocarbons, olefins, oxygenates, and carbolic acids increases along
with the column T-102 from 0.1323 at the bottom to 0.4816 at the top,
from 0.0243 at the bottom to 0.1050 at the top, from 0.0390 at the
bottom to 0.0832 at the top, and from 0.0103 at the bottom to 0.0220 at
the top, respectively, causing an enriching of oxygenates and carbolic
acids in the bottom stream of column T-102, as well as an impover-
ishment of hydrocarbons and olefins. In addition, the drastic change of
hydrocarbons and oxygenates composition between stage 10 and 9
shows that CO2 make up entering at the bottom of column T-102 is
instantly dissolved within the OLP liquid phase. Again, even though,
the weight fraction (CO2 basis) of OLP along with the column stages is
lower compared to that in Flowsheet I, when increasing the solvent-to-
feed from 17 to 38 in column T-102 of Flowsheet II, explains the higher
OLP fraction yield in the top stream of column T-102, as greater
amounts of OLP fractions will be dissolved in the gaseous phase (SC-
CO2).
For the gaseous phase, the composition profile of hydrocarbons and
olefins increase along with the column T-102 from 0.1477 at the bottom
to 0.2037 at the top, and from 0.0295 at the bottom to 0.0419 at the
top, respectively, while those of oxygenates and carboxylic acids
E.C. Costa et al.
decrease along with the column T-102 from 0.0704 at the bottom to
0.0355 at the top, and from 0.0186 at the bottom to 0.0085 at the top,
respectively. This explains the drastic decrease of oxygenates and car-
bolic acids contents in the top stream of column T-102, as well as the
enriching of hydrocarbons in the top stream of column T-102, being the
results according to experimental data of multistage gas extraction in
countercurrent columns reported in the literature [41,44,45].
4. Conclusions
The methods selected to estimate and/or predict the physical (ρ),
thermos-physical properties (Tb, Tc, Pc, Vc), and the acentric factor ω of
all the compounds present in organic liquid products, were chosen
based on statistical analysis that minimizes the measurements of central
tendency and dispersion. The regressions show that RK-Aspen model
was able to describe the high pressures experimental phase equilibrium
data for all the binary systems under investigation with AAD between
0.8% and 1.25% to liquid phase and 89.10-5% to 0.9% to gaseous
phase.
For all the simulations of process flowsheet I, at 333 K, 140 and
180 bar, and (S/F) = 12, 15, 17, 25, the best deacidifying operating
condition was observed at 333 K, 140 bar, and (S/F) = 17. Based on the
composition of OLP, the choice of CO2 as solvent, and the selected state
(333 K, 140 bar) and operating conditions (S/F = 17), the organic li-
quid products fractionation in multistage countercurrent absorber/
stripping columns operates within the liquid-gas phase region. The si-
mulation carried out for (S/F) = 17 presented a top stream yield of
36.65% with 96.95% (wt.) hydrocarbons and 3.05% (wt.) oxygenates,
in solvent free basis, showing a decrease on the oxygenates content in
OLP feed of 10% (wt.), with 2.63% (wt.) carboxylic acids, to 3.05%
(wt.) oxygenates, with 0.52% (wt.) carboxylic acids, as well as an en-
rich of hydrocarbons in OLP in top stream of 8.64% related to the initial
concentration of OLP in feed. Phase equilibrium data for system n-un-
decane/CO2 [69], a key compound present in OLP, at T = 344 K, a
temperature close to 333 K, shows gaseous phase yn-undecane ≈ 0.97,
giving a theoretical solvent-to-feed ratio (S/F) ≈ 32, above the selected
solvent-to-feed ratio of 17. In addition, the simulations results for OLP
fraction in top stream of column T-100, at 333 K and 140 bar, presented
chemical composition similar to those of kerosene-like fraction ob-
tained by fractional distillation of OLP, produced in pilot scale as de-
scribed in the literature [28].
For the simulations of OLP fractionation of Flowsheet II (Column T-
102, for S/F = 17 in Column T-100), the best deacidifying operating
condition was observed at 333 K, 140 bar, and (S/F) = 38, presenting a
top stream yield of 63.39% with 92.71% (wt.) hydrocarbons and 7.29%
(wt.) oxygenates, as well as a decrease on the oxygenates content in
OLP (Raffinate 1) feed of 15.22% (wt.), with 3.85% (wt.) carboxylic
acids, to 7.29% (wt.) oxygenates, with 1.49% (wt.) carboxylic acids.
The simulation results of OLP fractionation in multistage counter-
current absorber/stripping columns using SC-CO2 as solvent with the
chemical process simulator Aspen-Hysis, show that SC-CO2 was able not
only to deacidify OLP, but also to obtain OLP fractions with lower
olefins content.
Acknowledgements
E.C. Costa wish to acknowledge CAPES (Coordenação de
Aperfeiçoamento de Pessoal de Nivel Superior) for the scholarship. The
authors would like to thank ELETROBRAS S/A (Centrais Elétricas
Brasileiras S.A.-142796/2009-3) and CNPq (Conselho Nacional de
Desenvolvimento Científico e Tecnológico-486780/2012-0) for the fi-
nancial support.
References
[1] M.J. Haas, Animal fats, 6th edition, in: F. Shahidi (Ed.), Bailey’s Industrial Oil and
114
The Journal of Supercritical Fluids 140 (2018) 101–115
Fat Products, vol. 5, John Wiley & Sons, Inc., Hoboken, New Jersey, 2005, pp.
267–284.
[2] F.D. Gunstone, The Chemistry of Oils and Fats: Sources, Composition, Properties
and Uses, Blackwell Publishing-CRC Press, UK, 2004.
[3] K.D. Maher, D.C. Bressler, Pyrolysis of triglyceride materials for the production of
renewable fuels and chemicals, Bioresour. Technol. 98 (2007) 2351–2368.
[4] E. Vonghia, D.G.B. Boocock, S.K. Konar, A. Leung, Pathways for the deoxygenation
of triglycerides to aliphatic hydrocarbons over activated alumina, Energy Fuels 9
(6) (1995) 1090–1096.
[5] R.O. Idem, S.P.R. Katikaneni, N.N. Bakhshi, Thermal cracking of canola oil: reaction
products in the presence and absence of steam, Energy Fuels 10 (6) (1996)
1150–1162.
[6] L. Dandik, H.A. Aksoy, A. Erdem-Senatalar, Catalytic conversion of used oil to
hydrocarbon fuels in a fractionating pyrolysis reactor, Energy Fuels 12 (6) (1998)
1148–1152.
[7] L. Dandik, H.A. Aksoy, Pyrolysis of used sunflower oil in the presence of sodium
carbonate by using fractionating pyrolysis reactor, Fuel Process. Technol. 57 (1998)
81–92.
[8] D.G. Lima, V.C.D. Soares, E.B. Ribeiro, D.A. Carvalho, E.C.V. Cardoso, F.C. Rassi,
Diesel-like fuel obtained by pyrolysis of vegetable oils, J. Anal. Appl. Pyrolysis 71
(2004) 987–996.
[9] T. Hua, L. Chunyi, Y. Chaohe, S. Honghong, Alternative processing technology for
converting vegetable oil and animal fats to clean fuels and light olefins, Chin. J.
Chem. Eng. 16 (3) (2008) 394–400.
[10] C.M.R. Prado, N.R. Antoniosi Filho, Production and characterization of the biofuels
obtained by thermal cracking and thermal catalytic cracking of vegetable oils, J.
Anal. Appl. Pyrolysis 86 (2009) 338–347.
[11] V.R. Wiggers, H.F. Meier, A. Wisniewski Jr., A.A. Chivanga Barros, M.R. Wolf
Maciel, Biofuels from continuous fast pyrolysis of soybean oil: a pilot plant study,
Bioresour. Technol. 100 (2009) 6570–6577.
[12] X. Junming, J. Jianchun, S. Yunjuan, C. Jie, Production of hydrocarbon fuels from
pyrolysis of soybean oils using a basic catalyst, Bioresour. Technol. 101 (2010)
9803–9806.
[13] N. Taufiqurrahmi, S. Bhatia, Catalytic cracking of edible and non-edible oils for the
production of biofuels, Energy Environ. Sci. 4 (2011) 1087–1112.
[14] E. Buzetzki, K. Sidorová, Z. Cvengrošová, J. Cvengroš, Effects of oil type on products
obtained by cracking of oils and fats, Fuel Process Technol. 92 (2011) 2041–2047.
[15] E. Buzetzki, K. Sidorová, Z. Cvengrošová, A. Kaszonyi, J. Cvengroš, The influence of
zeolite catalysts on the products of rapeseed oil cracking, Fuel Process. Technol. 92
(2011) 1623–1631.
[16] S. Yan, C. Dimaggio, H. Wang, S. Mohan, M. Kim, L. Yang, S.O. Salley, K.Y. Simon
Ng, Catalytic conversion of triglycerides to liquid biofuels through transesterifica-
tion, cracki ng, and hydrotreatment processes, Curr. Catal. 1 (2012) 41–51.
[17] F. Yu, L. Gao, W. Wang, G. Zhang, J. Ji, Bio-fuel production from the catalytic
pyrolysis of soybean oil overMe-Al-MCM-41 (Me = La, Ni or Fe) mesoporous ma-
terials, J. Anal. Appl. Pyrolysis 104 (2013) 325–329.
[18] V.P. Doronin, O.V. Potapenko, P.V. Lipin, T.P. Sorokina, Catalytic cracking of ve-
getable oils and vacuum gas oil, Fuel 106 (2013) 757–765.
[19] V.R. Wiggers, G.R. Zonta, A.P. França, D.R. Scharf, E.L. Simionatto, L. Ender,
H.F. Meier, Challenges associated with choosing operational conditions for trigly-
ceride thermal cracking aiming to improve biofuel quality, Fuel 107 (2013)
601–608.
[20] L. Li, K. Quan, J. Xu, F. Liu, S. Liu, S. Yu, C. Xie, B. Zhang, X. Ge, Liquid hydro-
carbon fuels from catalytic cracking of waste cooking oils using basic mesoporous
molecular sieves K2O/Ba-MCM-41 as catalysts, ACS Sustain. Chem. Eng. 1 (11)
(2013) 1412–1416.
[21] S.A.P. Mota, A.A. Mancio, D.E.L. Lhamas, D.H. de Abreu, M.S. da Silva, W.G. dos
Santos, D.A.R. de Castro, R.M. de Oliveira, M.E. Araújo, L.E.P. Borges,
N.T. Machado, Production of green diesel by thermal catalytic cracking of crude
palm oil (Elaeis guineensis Jacq) in a pilot plant, J. Anal. Appl. Pyrolysis 110 (2014)
1–11.
[22] H. da Silva Almeida, O.A. Correa, J.G. Eid, H.J. Ribeiro, D.A.R. de Castro,
M.S. Pereira, L.M. Pereira, A. de Andrade Mancio, M.C. Santos, J.A. da Silva Souza,
Luiz E.P. Borges, N.M. Mendonca, N.T. Machado, Production of biofuels by thermal
catalytic cracking of scum from grease traps in pilot scale, J. Anal. Appl. Pyrolysis
118 (2016) 20–33.
[23] H. da Silva Almeida, O.A. Corrêa, J.G. Eid, H.J. Ribeiro, D.A.R. de Castro,
M.S. Pereira, L.M. Pereira, A. de Andrade Aâncio, M.C. Santos, S.A.P. da Mota,
J.A. da Silva Souza, Luiz E.P. Borges, N.M. Mendonça, N.T. Machado, Performance
of thermochemical conversion of fat, oils, and grease into kerosene-like hydro-
carbons in different production scales, J. Anal. Appl. Pyrolysis 120 (2016) 126–143.
[24] H. da Silva Almeida, O.A. Corrêa, C.C. Ferreira, H.J. Ribeiro, D.A.R. de Castro,
M.S. Pereira, A. de Andrade Aâncio, M.C. Santos, S.A.P. da Mota, J.A. da Silva
Souza, Luiz E.P. Borges, N.M. Mendonça, N.T. Machado, Diesel-like hydrocarbon
fuels by catalytic cracking of fats, oils, and grease (FOG) from grease traps, J.
Energy Inst. 90 (2017) 337–354.
[25] J.A. Melero, J. Iglesias, A. Garcia, Biomass as renewable feedstock in standard re-
finery units. Feasibility, opportunities and challenges, Energy Environ. Sci. 5 (2012)
7393–7420.
[26] Andréia de Andrade Mâncio, Production, Fractionation and De-Acidification of
Biofuels Obtained by Thermal Catalytic Cracking of Vegetable Oils. PhD Thesis,
April, UFPA, CDD 22, Ed. 660.2995 Graduate Program of Natural Resources
Engineering, 2015, http://proderna.propesp.ufpa.br/ARQUIVOS/teses/Andreia.
pdf.
[27] A.A. Mâncio, K.M.B. da Costa, C.C. Ferreira, M.C. Santos, D.E.L. Lhamas, S.A.P. da
Mota, R.A.C. Leão, R.O.M.A. de Souza, M.E. Araújo, L.E.P. Borges, N.T. Machado,
E.C. Costa et al.
Process analysis of physicochemical properties and chemical composition of organic
liquid products obtained by thermochemical conversion of palm oil, J. Anal. Appl.
Pyrolysis 123 (2017) 284–295.
[28] C.C. Ferreira, E.C. Costa, D.A.R. de Castro, M.S. Pereira, A.A. Mânio, M.C. Santos,
D.E.L. Lhamas, S.A.P. Mota, A.C. Leão, S.J. Duoisin, M.E. Araújo, L.E.P. Borges,
N.T. Mahado, Deacidification of organic liquid products by fractional distillation in
laboratory and pilot scales, J. Anal. Appl. Pyrolysis 127 (2017) 468–489.
[29] M.C. Santos, R.M. Lourenço, D.H. de Abreu, A.M. Pereira, D.A.R. de Castro,
M.S. Pereira, H.S. Almeida, A.A. Mãncio, D.E.L. Lhamas, S.A.P. da Mota, J.A. da
Silva Souza, S.D. Junior, M.E. Araújo, L.E.P. Borges, N.T. Machado, Gasoline-like
hydrocarbons by catalytic cracking of soap phase residue of neutralization process
of palm oil (Elaeis guineensis Jacq), J. Taiwan Inst. Chem. Eng. 71 (2016) 106–119.
[30] Zuo-gang Guo, Shu-rong Wang, Ying-ying Zhu, Zhong-yang Luo, Ke-fa Cen,
Separation of acid compounds for refining biomass pyrolysis oil, J. Fuel Chem.
Technol. 37 (2009) 49–52.
[31] X. Guo, S. Wang, Z. Guo, Q. Liu, Z. Luo, K. Cen, Pyrolysis characteristics of bio-oil
fractions separated by molecular distillation, Appl. Energy 87 (2010) 2892–2898.
[32] Z. Guo, S. Wang, Y. Gu, G. Xu, X. Li, Z. Luo, Separation characteristics of biomass
pyrolysis oil in molecular distillation, Sep. Purif. Technol. 76 (2010) 52–57.
[33] Earl D. Christensen, Gina M. Chupka, Jon Luecke, Tricia Smurthwaite, Teresa
L. Alleman, Kristiina Iisa, James A. Franz, Douglas C. Elliott, Robert L. McCormick,
Analysis of oxygenated compounds in hydrotreated biomass fast pyrolysis oil dis-
tillate fractions, Energy Fuels 25 (11) (2011) 5462–5471.
[34] S.-Q. Li, Q. Yao, Y. Chi, J.-H. Yan, K.-F. Cen, Pilot-scale pyrolysis of scrap tires in a
continuous rotary kiln reactor, Ind. Eng. Chem. Res. 43 (17) (2004) 5133–5145.
[35] Ji-Lu Zheng, Qin Wie, Improving the quality of fast pyrolysis bio-oil by reduced
pressure distillation, Biomass Bioenergy 35 (2011) 1804–1810.
[36] Xue-Song Zhang, Guang-Xi Yang, Hong Jiang, Wu-Jun Liu, Hong-Sheng Ding, Mass
production of chemicals from biomass-derived oil by directly atmospheric distilla-
tion coupled with co-pyrolysis, Sci. Rep. 3 (2013) 1120, http://dx.doi.org/10.1038/
srep01120.
[37] Anil Kumar Sarma, D. Konwer, Feasibility studies for conventional refinery dis-
tillation with a (1:1) w/w of a biocrude blend with petroleum crude oil, Energy
Fuels 19 (4) (2005) 1755–1758.
[38] Yaseen Elkasabi, Charles A. Mullen, AkwasiA. Boateng, Distillation and isolation of
commodity chemicals from bio-oil made by tail-gas reactive pyrolysis, Sustain.
Chem. Eng. 2 (2014) 2042–2052.
[39] T.P. Vispute, G.W. Huber, Production of hydrogen, alkanes and polyols by aqueous
phase processing of wood-derived pyrolysis oils, Green Chem. 11 (2009)
1433–1445.
[40] H.V. da Silva, Modelagem e Simulação do Fracionamento de Correntes Liquidas de
Produtos Naturais em Colunas em Contracorrente Usando Dióxido de carbono
Supercrítico. Master Thesis, Faculty of Chemical Engineering, UFPA, Belém-Pará-
Brazil, 2011.
[41] N.T. Machado, Fractionation of PFAD-Compounds in Countercurrent Columns
Using Supercritical Carbon Dioxide as Solvent. Doctoral Thesis, TU-Hamburg-
Harburg, 1998.
[42] G. Brunner, Industrial process development countercurrent multistage gas extrac-
tion (SFE) processes, J. Supercrit. Fluids 13 (1998) 283–301.
[43] G. Brunner, Counter-current separations, J. Supercrit. Fluids 47 (2009) 574–582.
[44] N.T. Machado, G. Brunner, Separation of saturated and unsaturated fatty acids from
Palm fatty acids distillates in continuous multistage countercurrent colums with
supercritical carbon dioxide as solvent: a process design methodology, Ciênc.
Tecnol. Aliment. 17 (1997) 361–370.
[45] G. Brunner, N.T. Machado, Process design methodology for fractionation of fatty
acids from palm fatty acid distillates in countercurrent packed columns with su-
percritical CO2, J. Supercrit. Fluids 66 (2012) 96–110.
[46] C. Säure, Untersuchungen zur Anreicherung von Squalen und Tocopherolen mittels
Gegenstromextraktion mit überkritischem Kohlendioxid. Doctoral Thesis, TU-
Hamburg-Harburg, Germany, 1996.
[47] M. Moricet, Simulierung von Gasextraktion in Bodenkolonnen am Beispiel der
Abtrennung von Monoglycerid aus einen Ölsäureglyceridgemisch sowie der Freien
Fettsäure aus Palmöl. Doctoral Thesis, Universität Erlagen-Nürnberg, 1982.
[48] M.F. Mendes, A.M.C. Uller, F.L.P. Pessoa, Simulation and thermodynamic modeling
of the extraction of tocopherol from a synthetic mixture of tocopherol, squalene and
CO2, Braz. J. Chem. Eng. 17 (2000) 761–770.
[49] F. Benvenuti, F. Gironi, L. Lamberti, Supercritical deterpenation of Lemon essential
oil, experimental data and simulation of semicontinuous extraction process, J.
Supercrit. Fluids 20 (2001) 29–44.
[50] EleniseBannwart de Moraes, Mario Eusebio Torres Alvarez, Maria Regina
Wolf Maciel, Rubens Maciel Filho, Simulation and optimization of a supercritical
extraction process for recovering provitamin A, Appl. Biochem. Biotechnol. 13
(2006) 1041–1050 1–3.
[51] L. Vásquez, C.F. Torres, T. Fornari, F.J. Senôrans, G. Reglero, Recovery of squalene
from vegetable oil sources using countercurrent supercritical carbon dioxide ex-
traction, J. Supercrit. Fluids 40 (2007) 59–66.
[52] T. Fornari, L. Vazquez, C.F. Torres, E. Ibanez, F.J. Senôrans, G. Reglero,
Countercurrent supercritical fluid extraction of different lipid-type materials: ex-
perimental and thermodynamic modeling, J. Supercrit. Fluids 45 (2008) 206–212.
[53] L. Vásquez, A.M. Benavides-Hurtado, G. Reglero, T. Fornari, E. Ibánez,
F.J. Senorans, Deacidification of olive oil by countercurrent supercritical carbon
dioxide extraction: experimental and thermodynamic modeling, J. Food Eng. 90
115
The Journal of Supercritical Fluids 140 (2018) 101–115
(2009) 463–470.
[54] S. Skjold-Jorgensen, Gas solubility calculations. II. Application of a new group
contribution equation of state, Fluid Phase Equilib. 16 (1984) 317–351.
[55] J.B. Fernandes, P.F. Lisboa, J.P.B. Mota, P.C. Simões, Modelling and simulation of a
complete supercritical fluid extraction plant with countercurrent fractionation
column, Sep. Sci. Technol. 46 (2011) 2088–2098.
[56] C.L. Yaws, Thermophysical Properties of Chemicals and Hydrocarbons, 2nd edition,
Willian Andrew Inc., New York, 2014 809 pp..
[57] T. Boublik, V. Fried, E. Hala, The Vapor Pressure of Pure Substances, 2nd edition,
Elsevier, New York, 1984 980 pp..
[58] A.A. Mancio, K.M.B. da Costa, C.C. Ferreira, M.C. Santos, D.E.L. Lhamas, S.A.P. da
Mota, R.A.C. Leão, R.O.M.A. de Souza, M.E. Araújo, L.E.P. Borges, N.T. Machado,
Thermal catalytic cracking of crude palm oil at pilot scale: effect of the percentage
of Na2CO3 on the quality of biofuels, Ind Crops Prod. 91 (2016) 32–43.
[59] K.G. Joback, R.C. Reid, Estimation of pure-component properties from group-con-
tributions, Chem. Eng. Commun. 57 (1987) 233–243.
[60] L. Constantinou, R. Gani, New group contribution method for estimating properties
of pure compounds, AIChE J. 40 (10) (1994) 1697–1710.
[61] J. Marrero-Morejon, E. Pardillo-Fontdevila, Estimation of pure compound proper-
ties using group-interaction contributions, AIChE J. 45 (3) (1999) 615–621.
[62] J. Marrero, R. Gani, Group-contribution based estimation of pure component
properties, Fluid Phase Equilib. 183–184 (2001) 183–208.
[63] L. Constantinou, R. Gani, J.P. O’Connell, Estimation of the acentric factor and the
liquid molar volume at 298 K using a new group contribution method, Fluid Phase
Equilib. 103 (1995) 11–22.
[64] B. Han, D.-Y. Peng, A group-contribution correlation for predicting the acentric
factors of organic compounds, Can. J. Chem. Eng. 71 (1993) 332–334.
[65] M.E. Araújo, M.A.A. Meireles, Improving phase equilibrium calculation with the
Penge Robinson EOS for fats and oils related compounds/supercritical CO2 systems,
Fluid Phase Equilib. 169 (2000) 49–64.
[66] A. Vetere, Predicting the vapor pressures of pure compounds by using the Wagner
equation, Fluid Phase Equilib. 62 (1991) 1–10.
[67] S.A.B.V. Melo, M.F. Mendes, F.L.P. Pessoa, An oriented way to estimate Tc and Pc
using the most accurate methods available, Braz. J. Chem. Eng. 13 (1996) 192–199.
[68] R. Jimenez-Gallegos, L.A. Galicia-Luna, O. Elizalde-Solis, Experimental vapor-liquid
equilibria for the carbon dioxide + octane and carbon dioxide + decane systems, J.
Chem. Eng. Data 51 (2006) 1624–1628.
[69] L.E. Camacho-Camacho, L.A. Galicia-Luna, O. Elizalde-Solis, Z. Martinez-Ramirez,
New isothermal vapor-liquid equilibria for the CO2 + n-nonane, and CO2 + n-
undecane systems, Fluid Phase Equilib. 259 (2007) 45–50.
[70] K.A.M. Gasem, K.B. Dickson, P.B. Dulcamara, N. Nagarajan, R.L.J. Robinson,
Equilibrium phase compositions, phase densities, and interfacial tensions for CO2
+ hydrocarbon systems. 5. CO2 + n-Tetradecane, J. Chem. Eng. Data (1989)
191–195.
[71] C. Secuianu, V. Feroiu, D. Geanã, Phase behavior for the carbon dioxide + N-
pentadecane binary system, J. Chem. Eng. Data 55 (2010) 4255–4259.
[72] R. D’Souza, J.R. Patrick, A.S. Teja, High-pressure phase equilibria in the carbon
dioxide-n-hexadecane and carbon dioxide-water systems, Can. J. Chem. Eng. 66
(1988) 319.
[73] H. Kim, H.M. Lin, K.-C. Chao, Vapor-liquid equilibrium in binary mixtures of carbon
dioxide + n-propylcyclohexane and carbon dioxide + n-octadecane, AIChE Symp.
Ser. 81 (244) (1985) 96–101.
[74] J.S. Yau, Y.Y. Chiang, D.S. Shy, F.N. Tsai, Solubilities of carbonn dioxide in car-
boxylic acids under high pressures, J. Chem. Eng. Jpn. 25 (5) (1992) 544–548.
[75] R. Bharath, H. Inomata, T. Adschiri, K. Arai, Phase equilibrium study for the se-
paration and fractionation of fatty oil components using supercritical carbon di-
oxide, Fluid Phase Equilib. 81 (1992) 307–320.
[76] H. Tanaka, Y. Yamaki, M. Kato, Solubility of carbon-dioxide in pentadecane, hex-
adecane, and pentadecane plus hexadecane, J. Chem. Eng. Data 38 (1993)
386–388.
[77] G.F. Chou, R.R. Forbert, J.M. Prausnitz, High-pressure vapor-liquid-equilibria for
CO2/n-decane, CO2/tetralin, and CO2/n-decane/tetralin at 71.1ºC and 104.4 ºC, J.
Chem. Eng. Data 35 (1990) 26–29.
[78] Mariana B. Oliveira, AntónioJosé Queimada, Georgios M. Kontogeorgis, A.P. João,
Coutinho, Evaluation of the CO2 behavior in binary mixtures with alkanes, alco-
hols, acids and esters using the Cubic-Plus-Association Equation of State, J.
Supercrit. Fluids 55 (2011) 876–892.
[79] C.J. Eckert, S.I. Sandler, Vapor–liquid equilibria for the carbon dioxide–cyclo-
pentane system at 37.7, 45.0, and 60.0 ºC, J. Chem. Eng. Data 31 (1986) 26–28.
[80] N.N. Shah, J.A. Zollweg, W.B. Streett, Vapor–liquid equilibrium in the system
carbon dioxide + cyclopentane from 275 to 493 K at pressures to 12.2 MPa, J.
Chem. Eng. Data 36 (2) (1991) 188–192.
[81] N. Nagarajan, R.L. Robinson, Equilibrium phase compositions, phase densities, and
interfacial tensions for carbon dioxide + hydrocarbon systems. CO2 + cyclo-
hexane. CO2 + benzene, J. Chem. Eng. Data 32 (3) (1987) 369–371.
[82] H.J. Ng, D.B. Robinson, The equilibrium phase properties of selected naphthenic
binary systems: carbon dioxide–methylcyclohexane, hydrogen sulfide–methylcy-
clohexane, Fluid Phase Equilib. 2 (1979) 283–292.
[83] Stéphane Vitu, Romain Privat, Jean-Noël Jaubert, Fabrice Mutelet, Predicting the
phase equilibria of CO2 + hydrocarbon systems with the PPR78 model (PR EOS
and kij calculated through a group contribution method), J. Supercrit. Fluids 45
(2008) 1–26.