Professional Documents
Culture Documents
a r t i c l e i n f o a b s t r a c t
Article history: This paper describes a dynamic heterogeneous one-dimensional model of trickle-bed reactor used for
Received 24 August 2011 catalytic hydrotreating of oil fractions. The model takes into consideration the main reactions occurring in
Received in revised form 3 February 2012 the hydrotreating process: hydrodesulfurization, hydrodenitrogenation, hydrodearomatization (mono-,
Accepted 10 February 2012
di-, and polyaroamatics), olens hydrogenation, and mild hydrocracking (gas oil, naphtha, and gases).
Available online 10 March 2012
Kinetic parameters were determined from experimental data obtained in an isothermal bench-scale reac-
tor during hydrotreating of atmospheric gas oil coming from a heavy crude oil over a commercial CoMo
Keywords:
catalyst. The developed model was used to predict the dynamic behavior of an industrial hydrotreat-
Dynamic modeling
Catalytic hydrotreating
ing reactor within a wide range of reaction conditions. Changes in concentration, partial pressure, and
Trickle-bed reactor temperature proles are simulated and discussed as a function of reactor axial position and time. The sim-
ulation results obtained with the proposed dynamic model showed good agreement with experimental
data.
2012 Elsevier B.V. All rights reserved.
0926-860X/$ see front matter 2012 Elsevier B.V. All rights reserved.
doi:10.1016/j.apcata.2012.02.034
14 F.S. Mederos et al. / Applied Catalysis A: General 425426 (2012) 1327
Nomenclature
Sp total geometric external surface area of catalyst par-
aL gasliquid interfacial area per unit reactor volume, ticle, cm2S
cm2I /cm3r t time, s
aS liquidsolid interfacial area per unit reactor volume, Tf temperature of f phase, K
cm2S /cm3r uf supercial velocity of f phase, cm3f /(cm2r s)
A ,b empirical constants for Bondis correlation, dimen- Vp total geometric volume of catalyst particle, cm3S
sionless wi weight fraction of compound i in the liquid phase,
Cpf specic heat capacity of f phase, J/(gf K) gi /gL
CiL molar concentration of component i in the bulk liq- Yi weight fraction for mild HCR reactions, gi /gTotal
uid phase, moli /cm3L z axial reactor coordinate, cmr
S
CSLi molar concentration of compound i at surface of
solid covered by liquid phase, moli /cm3L Greek symbols
dpe equivalent particle diameter, cmS Hads,H2 S adsorption enthalpy of H2 S, J/molH2 S
L
Dei effective ckian diffusivity in liquid phase of com- L
HR,j heat of reaction j in the liquid phase, J/moli
pound i inside porous solid, cm3L /(cmS s) B catalyst bed void fraction, cm3(G+L) /cm3r
L
DMi molecular diffusion coefcient of compound i in the
S catalyst particle porosity, cm3(G+L) /cm3S
liquid phase, cm3 L /(cmS s)
f external holdup of f phase, cm3f /cm3r
Ea,j activation energy for j reaction, J/moli
j L catalyst effectiveness factor of reaction j in the liquid
fw catalyst wetting efciency, cm2S,wet /cm2S
phase, dimensionless
F objective function to be optimized, moli /cm3L or
B catalyst bed density, gS /cm3cat
gi /gTotal
f density of f phase at process conditions, gf /cm3f
GmL supercial mass ow velocity of liquid phase,
gL /(cm2r s) i,j L stoichiometric coefcient of component i in reaction
hLS heat transfer coefcient for liquid lm surrounding j in the liquid phase, dimensionless
the catalyst particle, J/(s cm2S K) catalyst bed dilution factor, cm3cat /(cm3cat + cm3inert )
Hi Henrys law coefcient of component i, S shape factor (=surface area of a sphere of equal vol-
MPa cm3L /moli ume/solid surface area), dimensionless
k1,2,3 intrinsic reaction rate constants for mild HCR reac- Lj Thiele modulus of reaction j in the liquid phase,
1n dimensionless
tions, (moli /cm3L ) (1/s)
kapp,j apparent reaction rate constant for heterogeneous tortuosity factor for catalyst particle, cmL /cmS
m+n
reaction j, (cm3L ) /[moli (m+n1)
gS s] Subscripts
kin,j intrinsic reaction rate constant for heterogeneous app apparent
m+n
reaction j, (cm3L ) /[moli (m+n1) gS s] calc calculated
1n CH4 methane
k0,j frequency factor for reaction j, (moli /cm3L ) (1/s)
Di diaromatics
k0,j frequency factor for heterogeneous reaction j,
m+n exp experimental
(cm3L ) /[moli (m+n1) gS s] f phase (gas, liquid or solid); nal or outlet condition
kiS liquidsolid mass transfer coefcient of compound G gas phase
i, cm3L /(cm2S s) GO gas oil
KLi overall gasliquid mass transfer coefcient of com- HC hydrocarbon
pound i in the liquid phase, cm3L /(cm2I s) HCR hydrocracking reaction
Kj equilibrium constant for j (=HDAPoly/Di/Mono ) reac- HDA hydrodearomatization reaction
tion, dimensionless HDN hydrodenitrogenation reaction
KH2 S adsorption-equilibrium constant of H2 S on catalyst HDS hydrodesulfurization reaction
active sites, cm3L /molH2 S HGO olens hydrogenation reaction
K0,j pre-exponential factor for equilibrium constant of H2 molecular hydrogen
reaction j, dimensionless H2 S hydrogen sulde
K0,H2 S pre-exponential factor for adsorption-equilibrium i component index
of H2 S, cm3L /molH2 S in intrinsic
LB length of catalyst bed, cmr I gas-liquid interface
m,n reaction order, dimensionless j reaction index
MWL molecular weight of liquid phase, gL /molL k experiment index
NRL number of reactions in the liquid phase, dimension- L liquid phase
less Mono monoaromatics
pG
i
partial pressure of component i in the bulk gas N nitrogen compound
phase, MPa Naph naphthenes
rjL rate of reaction j per unit of volume in the liquid NH3 ammonia
phase, moli /(cm3L s) O olens
Poly polyaromatics
r Lj rate of reaction j per unit of catalyst mass in the
S sulfur compound; solid phase; condition at external
liquid phase, moli /(gS s)
surface of catalyst particle
R gas law constant, J/(moli K)
0 initial or inlet condition; reference condition
F.S. Mederos et al. / Applied Catalysis A: General 425426 (2012) 1327 15
The atmospheric gas oil feed and liquid product samples were
2. Experimental characterized using the following techniques.
The sulfur content was measured by an X-ray uorescence
2.1. Materials and experimental setup spectrometer following the ASTM D-4294 method. Before sulfur
analyzes the hydrotreated samples were treated (purged) with
The feed used for all experiments was a straight-run gas oil nitrogen gas in order to strip dissolved H2 S and NH3 . The total nitro-
obtained from a heavy crude oil, whose properties are presented in gen content was determined using a chemiluminescence method as
Table 1. The commercial catalyst used for experiments was a pre- per ASTM D-4629, while the basic nitrogen content was determined
sulded cobalt molybdenum supported on -alumina. The reactor by the color indicator titration method (UOP-313).
was loaded with 99.43 g (100 ml) of powdered catalyst previously The FIA (ASTM D-1319) method was used to separate feed and
crushed and sieved, which properties are given in Table 1. liquid product samples into saturate and aromatic fractions. The
The bench-scale reactor is operated in downow and isothermal aromatic part in liquid samples was analyzed by supercritical uid
mode provided with independent temperature control of a three- chromatography according to ASTM D-5186 method in order to
zone electric furnace. The internal diameter of the reactor is 2.54 cm obtain the content of aromatic structures (poly-, di-, and monoaro-
and at the center a thermowell of external diameter of 0.635 cm was matics). The saturate fraction of the feed was used to measure the
placed. Catalytic length was of 25.2 cm. initial naphthenes content by eld-desorption mass spectrometry
according to ASTM D-2425 method.
2.2. Experimental tests The extent of hydrocracking (HCR) of gas oil to lighter products
was determined through material balance and gas chromato-
Prior to experimental runs, the catalyst was activated by the graphic analysis of gas and liquid product samples. The bromine
procedure reported elsewhere [30]. After catalyst sulding, the number was determined by ASTM D-1159 and the olen content
experiments were carried out at constant pressure of 5.3 MPa was estimated from the bromine number.
and H2 /oil ratio of 356 std m3 /m3 (2000 scf/bbl), varying the reac- Other physical properties of the feed and products were deter-
tor temperature from 340 to 380 C and liquid hourly space mined by standard methods: liquid density by U-tube density
cell according to ASTM D-4052 method, viscosity by a Stabinger
viscosity cell based on ASTM D-7042 method, and boiling point
Table 1
Properties of the gas oil feedstock and catalyst.
distribution by ASTM D-86 method. The gas product was analyzed
by on-line gas chromatograph. Physical properties of the catalyst
Gas Oil Value such as bulk density of aggregates and total pore volume using
Density the BrunauerEmmettTeller (BET) technique were determined by
at 15.6 C and 586 mmHg (g/cm3 ) 0.873 ASTM C-29 and ASTM-4222 methods, respectively.
at 20 C and 586 mmHg (g/cm3 ) 0.869
Viscosity
at 20 C and 586 mmHg (cP) 7.2554 3. Formulation of the reactor model
at 40 C and 586 mmHg (cP) 4.0161
at 60 C and 586 mmHg (cP) 2.5455
The model developed to simulate a TBR for HDT of oil fractions
Total sulfur (wt%) 2.19
Total nitrogen (wppm) 330 at bench and commercial scales is dynamic one-dimensional het-
Basic nitrogen (wppm) 136 erogeneous, which is based on the three-phase steady-state reactor
Nonbasic nitrogen (wppm) 194 model reported in the literature [11,18].
Olens (wt%) 4.83
Total aromatics (wt%) 34.9
Polyaromatics (wt%) 6.6 3.1. Assumptions
Diaromatics (wt%) 7.8
Monoaromatics (wt%) 20.5 In the present work the following assumptions were made:
Total saturates (wt%) 65.1
Naphthenes (wt%) 46.17
Parafns (wt%) 18.93 1. The reactors (bench and commercial) operate in dynamic
Atmospheric distillation ( C) regime.
IBP 284.8
2. Both gas and liquid phases behave as plug ow.
5 vol% 291.4
10 vol% 292.8 3. The bench-scale reactor is isothermally and isobarically oper-
90 vol% 318.8 ated.
95 vol% 323.7 4. The commercial adiabatic reactor is isothermal only in the
FBP 328.7 radial direction.
5. No liquid evaporation or vapor condensation along the catalytic
Catalysta
bed.
Co (wt%) 3.47 6. In the bench-scale reactor the gas-side of the gasliquid mass
Mo (wt%) 14.70
transfer resistance is assumed to be negligible.
-Al2 O3 (wt%) 81.83
Particle shape Crushed Trilobe 7. Liquidsolid lm mass transfer is the rate limiting resistance
extrudate in bench-scale reactor, while intraparticle mass transfer is the
Equivalent particle diameter (cm) 0.0230 0.2540 rate limiting resistance in commercial reactor.
Bulk density (g/cm3 ) 0.9943 0.9200
8. Effectiveness factors are used to estimate the intraparticle mass
a
Sulded catalyst; crushed particles sieved with 60/70 mesh size. transfer resistances in commercial reactor.
16 F.S. Mederos et al. / Applied Catalysis A: General 425426 (2012) 1327
9. Chemical reactions occur only in the liquid phase when con- where i = H2 , H2 S, NH3 , and CH4 . sign is for cocurrent operation
tacting with the solid phase surface. and the + sign is for countercurrent operation.
10. The liquid volume (or liquid holdup) in the bench-scale and
commercial reactors remains constant. 3.2.2. Liquid phase
11. The catalyst particles are completely wetted in the commercial The unsteady state mass-balance equation in the catalyst bed
reactor. for the gaseous compounds in the liquid phase is
12. Catalyst activity does not change with time because of the use
of fresh catalyst and the relatively short time required during CiL CiL pG
i
L = uL + KLi aL CiL kiS aS (CiL CSLi
S
) (2)
experiments. t z Hi
13. There are not intraparticle temperature differences (isothermal
where i = H2 , H2 S, NH3 , and CH4 .
solid phase).
The model assumes that the organosulfur, organonitrogen, aro-
14. The energy balance for gas phase is not taken into consider-
matic, and olen compounds, as well as the liquid hydrocarbons,
ation because the heat capacity of the gas phase is much lower
are nonvolatile; therefore, the unsteady state mass-balance for the
compared with that of the liquid phase.
liquid compounds is
Regarding assumption 5, it should be mentioned that modeling CiL CiL
the feed vaporization in TBRs is in fact a difcult task. Up until L = uL kiS aS (CiL CSLi
S
) (3)
t z
now, only one paper dealing with this issue has used this con-
where i = S, N, Poly, Di, Mono, Naph, O, GO, and Naphtha.
cept in simulations of hydrotreating reactor [31]. In that study
the authors have concluded that only slight differences are found
in temperature and hydrodearomatization proles by considering 3.2.3. Solid phase
the vaporization of hydrocarbon (vapor liquid equilibrium, VLE). The components transported between liquid and solid phases
Also although they have found that hydrodesulfurization within are consumed or produced by the chemical reaction at the wet
the reactor can be signicant, the effect of vaporization becomes catalyst surface, according to the following rst-order ordinary dif-
smaller for high levels of conversion. Higher HDS is predicted by ferential equations (ODEs):
using VLE calculations, which provokes an increase in temper- S
CSLi
NRL
ature with the consequent enhanced vaporization. Hence these L L L
S (1 B ) = kiS aS (CiL CSLi
S
) + fw B i,j S
j r j (CSLi , TSS )
phenomena are strongly linked, and ignoring them may cancel its t
j=1
effect. This is in agreement with theoretical calculations of Akger- (4)
man et al. [32] who pointed out that at high conversions, the
difference between the model that takes into account the volatiles where i = H2 , H2 S, NH3 , CH4 , S, N, Poly, Di, Mono, Naph, O, GO, and
and that that does not diminishes due to depletion of the limit- Naphtha; j = HDS, HDN, HDAPoly , HDADi , HDAMono , HDANaph , HGO,
ing reactant. In our study the level of HDS was greater than 95%, and HCR.
thus according to these previous reports, the effect of vaporiza-
tion is expected to be minimal. In addition, the feed used in our 3.3. Unsteady state heat balances
experiments is heavy oil-derived gas oil, which apart from exhibit-
ing high contents of sulfur and aromatics, such compounds are of Since the HDT reactions are exothermic in nature, the energy-
high-molecular weight as compared with those present in typical balance equations for commercial HDT reactor modeling operating
gas oil. This heaviness and complexity of these impurities make under adiabatic no isothermal conditions are given by the following
our feedstock of less volatility. In summary, although VLE calcula- terms [7]:
tions are welcomed the assumption that liquid and gas phase are
in equilibrium is not well established [7]. (i) Liquid phase
TL TL
3.2. Unsteady state mass balances L L CpL = uL L CpL hLS aS (TL TSS ) (5)
t z
The mass transfer of the compounds in the TBR is described with (ii) Solid phase
the following set of partial differential equations (PDEs), taking into
account the previous assumptions. TSS
NRL
Lj r j (CSLi
L
S S CpS = hLS aS (TL TSS ) + fw B S
, TSS )(HR,j
L
)
t
j=1
3.2.1. Gas phase
(6)
The unsteady state mass-balance equation in the catalyst bed
for the compounds present in the gas phase is the following:
G G
3.4. Boundary conditions
G pi uG pi pG
i
= KLi aL CiL (1)
RTL t RTL z Hi
Since the mathematical model is a system of PDEs and ODEs with
time and spatial coordinate as independent variables, it is necessary
to dene the following initial and boundary conditions for the gas,
liquid, and solid phases for cocurrent operation mode:
For t = 0, at z = 0, pG
i
= (pG ) , i = H2 , H2 S, NH3 , and CH4
i 0
L L
Ci = (Ci )0 and
(7)
S = 0,
CSLi i = H2 , H2 S, NH3 , CH4 , S, N, Poly, Di, Mono, Naph, O, GO, and Naphtha
TL = TSS = T0
F.S. Mederos et al. / Applied Catalysis A: General 425426 (2012) 1327 17
at 0 < z < LB , pG
i
= 0, i = H2 , H2 S, NH3 , and CH4
CiL = 0 and
S = 0,
(8)
CSLi i= H2 , H2 S, NH3 , CH4 , S, N, Poly, Di, Mono, Naph, O, GO, and Naphtha
TL = TSS = T0
at z = LB , pG
i
= 0, i = H2 , H2 S, NH3 , and CH4
CiL = 0 and
(9)
S = 0,
CSLi i = H2 , H2 S, NH3 , CH4 , S, N, Poly, Di, Mono, Naph, O, GO, and Naphtha
TL = TSS = T0
(ii) Boundary conditions:
For t > 0 at z = 0, pG
i
= (pG ) ,
i 0
i = H2 , H2 S, NH3 , and CH4
S = 0,
CiL = (CiL )0 and CsLi
(10)
i = H2 , H2 S, NH3 , CH4 , S, N, Poly, Di, Mono, Naph, O, GO, and Naphtha
TL = (TL )0 ; TSS = (TSS )0
When a high-purity hydrogen stream without gas recycle is 3.6.2. Hydrodenitrogenation reaction
used, such as in the case of laboratory reactor, or when the recy- As in the case of HDS reaction, all the organic nitrogen com-
cle gas has been subject to purication process in commercial HDT pounds in the feedstock are grouped in one lump. The following
units, the values of partial pressures (pG i
) and liquid molar concen- generalized stoichiometric equation was used to represent all
trations (CiL ) of H2 S, NH3 , and CH4 at the entrance of the catalytic hydrodenitrogenation (HDN) reactions:
bed (z = 0) are equal or very close to zero. k
HDN
Ar N + 3H2 Ar H + NH3 (14)
3.5. Integration method
where Ar N is the nitrogen-bearing compound, Ar H is the aro-
matic compound free of nitrogen, and NH3 is ammonia. The
The reactor model was solved numerically by applying the
following pseudo-rst-order kinetics with respect to the concen-
method of lines [33,34]. The set of PDEs describing the mass and
tration of total nitrogen was assumed for the HDN reaction [14]:
heat transfer in the reactor were transformed into a set of rst-
order ODEs by discretizing the spatial partial derivatives in the L B r LHDN S
rHDN = = kin,HDN (CSLN ) (15)
axial direction using the backward nite difference expressions [ S (1 B )]
and leaving the independent variable time (time partial deriva-
Although nitrogen compounds can inuence HDS activity due
tives) without discretize. The nal system of ODEs obtained was
to competitive adsorption, that effect was not considered in this
then solved with respect to time using the fourth-order Runge-
study by the following reasons:
Kutta method. The software developed in this work to calculate the
chemicalphysical properties, transport properties, hydrodynamic
parameters and simultaneously to solve the system of ODEs was The heavy oil-derived gas oil used in our experiments possesses
a stand-alone program coded in Visual Fortran 90 programming much lower nitrogen content in comparison with sulfur and aro-
language that can be run on a personal computer. matics.
The CoMo supported catalyst used for the test is more selective
3.6. Reaction kinetic models for sulfur removal.
The effect of nitrogen compounds on HDS and other reactions is
3.6.1. Hydrodesulfurization reaction more notorious when dealing with studies for producing ultra-
To model hydrodesulfurization (HDS) reaction the following low sulfur diesel, which is not our case. Only in such cases the
generalized representation was assumed: competitive adsorption of nitrogen on HDS is important [37].
HDSk
Ar S + 2H2 Ar H + H2 S (11) 3.6.3. Hydrodearomatization reaction
where Ar S is the sulfur-containing compound, H2 is hydrogen, The HDA reactions network was modeled with the following
Ar H is the corresponding aromatic compound free of sulfur, and stoichiometric equations as reported by Chowdhury et al. [5]:
H2 S is hydrogen sulde. The kinetic model to describe the HDS kPoly
reaction in this work is based on the following previously reported Polyaromatics + H2 Diaromatics (16)
kPoly-
equation:
S ) m1 S m2 kDi
kin,HDS (CSLS (CSLH ) Diaromatics + 2H2 Monoaromatics (17)
r HDS =
L 2
2
(12) kDi-
S
(1 + KH2 S CSLH ) k
2S Mono
Monoaromatics + 3H2 Naphthenes (18)
kMono-
The exponent two in the denominator of Eq. (12) means the
It is reported in the literature that in multiring aromatic
number of active sites participating in adsorption for hydrogen
compounds, the ring with the lowest aromaticity is rst hydro-
sulde. The adsorption equilibrium constant of hydrogen sulde
genated. For instance the low aromaticity character of anthracene
(KH2 S ) is also a function of temperature and can be estimated using
middle ring causes this compound to be hydrogenated to pro-
the vant Hoff equation:
duce 9,10-dihydroanthracene, and phenantrene is hydrogenated
Hads,H2 S to 9,10-dihydrophenanthrene over NiMo/alumina catalyst [35,36].
KH2 S = K0,H2 S exp (13)
RTSS Based on these experimental evidences, Chowdhury et al. [5] for-
mulate Eq. (16), which was also used for our model.
18 F.S. Mederos et al. / Applied Catalysis A: General 425426 (2012) 1327
Fig. 1. Reaction scheme of the three-lump kinetic model for mild HCR reactions.
n
r HDAMono = kin,HDA
L
(pG 2 S
H ) (CSLDi ) + kin,HDA
S
(CSLMono )
Di 2 Di -
n3
Hydrogenation of olens was also assumed to be pseudo-rst-
+ kin,HDA (pG S S
H2 ) (CSLMono ) kin,HDAMono- (CSLNaph ) order with respect to the total concentration of olens, as given by
Mono
the following reaction rate equation [14]:
(21a)
r HGO = kin,HGO
L S
(CSLO ) (31)
n
r HDANaph = r HDAMono = kin,HDA
L L
(pG 3 S
H ) (CSLMono )
3.6.5. Mild hydrocracking reactions
Mono 2
The mild HCR reactions were simulated using the three-lump
S
+ kin,HDA (CSLNaph ) (22a) model shown in Fig. 1, which includes gas oil, naphtha, and light
Mono -
gases (hydrocarbon gases like C1 C4 ) [14]. The reactions were rep-
resented by pseudo-rst-order kinetic equations as follows:
As the solution of this set of equations is quite formidable even
for the rst-order case, the following simplied set of rst-order L fw B r LGO S S
rGO = = k1 (CSLGO ) + k2 (CSLGO ) (32)
reaction rate expressions were used to describe the HDA reactions ( S S )
[5,14]:
L
fw B r LNaphtha S S
n
r HDAPoly = kin,HDA
L
(pG 1 S S rNaphtha = = k1 (CSLGO ) + k3 (CSLNaphtha ) (33)
Poly H ) (CSLPoly ) kin,HDA
2 Poly -
(CSLDi ) (19b) ( S S )
r HDADi = kin,HDA
L
(pG 2 S n S fw B r LLG
H ) (CSLDi ) kin,HDA
Di 2 Di -
(CSLMono ) (20b) L
rLG = S
= k2 (CSLGO S
) k3 (CSLNaphtha ) (34)
( S S )
n
r HDAMono = kin,HDA
L
(pG 3 S
H ) (CSLMono ) kin,HDA
S
(CSLNaph ) 3.7. Estimation of parameters
Mono 2 Mono -
(21b)
3.7.1. Kinetic parameters
Once the reactor equations and assumptions have been estab-
n lished, and HDS, HDN, HDA, and HGO reaction rate expressions
r HDANaph = r HDAMono = kin,HDA
L L
(pG 3 S
H ) (CSLMono ) have been dened by differential analysis of the experimental data
Mono 2
Table 2
Kinetic and hydrodynamic parameters for gas oil HDT reactions.
L
Reaction j k0,j a Ea,j (kJ/moli ) K0,i/j b Hads,i (kJ/moli ) HR,j (kJ/mol H2 )
Temperature ( C) A b
The HDS reaction rate was taken according to the kinetic model product yields. The objective function to estimate these parameters
given by Eq. (12) with m1 = 1.8 and m2 = 0.96. was dened as follows:
For HDA reactions, in order to determine the backward reaction
rate constants, the equilibrium constants were dened as follows:
Nexp Nlump
The HCR reaction rate constants were estimated by minimizing 6(Vp /Sp )
dpe = (49)
the difference between the experimental and predicted values of s
20 F.S. Mederos et al. / Applied Catalysis A: General 425426 (2012) 1327
Sulfur, mol/cm 3
ues of for each reaction [10,43,47]: 4.0E-04
for Lj < 3
3.0E-04
tanh(Lj )
Lj = (50)
Lj 2.0E-04
for Lj 3
1.0E-04
1
Lj = (51)
Lj 0.0E+00
0 5 10 15 20 25 30
Reactor length z, cm
3.7.3. Hydrodynamic parameters
According to Sattereld [48] the catalyst wetting efciency is Fig. 2. Sulfur concentration along the catalyst bed length in the bench-scale trickle-
dened with the following ratio: bed reactor at different times. () 60 s, ( ) 400 s, () 900 s, () 1700 s, ( )
experimental value.
kapp,HDS
fw = (52)
kin,HDS
For the remaining organic compounds it was assumed that they
The catalyst wetting efciency in experimental reactors has have the same molecular weight as the whole sample, so that it is
been reported to be in the range of 0.120.6, while for commercial possible to use their weight fractions to calculate concentrations:
reactors in the range of 0.71.0 [14,48]. (L )0
In the present work, the following partial wetting model of (CiL )0 = (w ) (58)
MWL i 0
Bondi [49] modied by Sattereld [48], was used to correlate the
apparent rate constant, k app,HDS , with the supercial liquid mass where i = Poly, Di, Mono, Naph, GO, and Naphtha. The estimation of
velocity, GmL , and the intrinsic rate constant, k in,HDS : the nal weight fractions is done with:
1 105 MWL L
kapp,HDS = (53) (wi )f = (C ) (59)
b (L )f i f
(1/kin,HDS 105 ) + (A /(GmL ) )
The initial liquid concentration of the gaseous compounds is
The fw ReL curve given by this correlation must be obtained by estimated with the following expression:
varying the liquid mass velocity for only one reactor temperature.
The values of the empirical constants A and b for different tem- (pG )
i 0
(CiL )0 = (60)
peratures are given in Table 2. The intrinsic reaction rate constant Hi
(k in,HDS ) is a temperature function that may be further developed
into the Arrhenius equation: where i = H2 , H2 S, NH3 , and CH4 . Methane was taken as the repre-
sentative compound for light hydrocarbons (C1 C4 ) kinetics.
Ea,HDS
kin,HDS = k0,HDS exp (54)
RTSS 4.1. Dynamic simulation of an isothermal HDT bench-scale
reactor
The model makes use of correlations reported in the open liter-
ature in order to calculate transport and thermophysical properties Simulation results of the bench-scale reactor are based on the
[50]. experimental operating conditions and parameters given earlier.
The kinetic parameters estimated above by minimizing the differ-
4. Results and discussion ences between model predictions and experimental results, were
used to simulate the behavior of the bench-scale TBR.
The three-phase isothermal reactor model developed in this The variation of the sulfur concentrations in the liquid phase
work was applied to analyze and simulate the performance of along the length of the catalyst bed is presented in Fig. 2. The con-
a bench-scale reactor. The model solution for the experimental centration proles at different times were generated at a reactor
reactor is an initial-value problem as the concentrations of reac- temperature of 340 C, pressure of 5.3 MPa, LHSV of 2.5 h1 , and
tants and products are known at the reactor inlet. The model was H2 /oil ratio of 356 std m3 /m3 (2000 scf/bbl). The experimental sul-
solved with the kinetic parameters estimated from experiments as fur content at the exit of the reactor is well predicted. At short time
reported previously. (i.e., 60 s) the feed only reaches 20% of the total reactor length, that
The calculation of the initial molar concentrations for sulfur-, is why the sulfur content in the remaining 80% of the reactor length
nitrogen-containing compounds and olens can be made by the is zero, which means that this part of the reactor is empty. As the
following expressions: time passes the reactor is completely lled with the reacting mix-
(L )0 ture, reaching the steady-state condition at 1700 s. The variation of
(CSL )0 = (wS )0 (55) the hydrogen partial pressure in the gaseous phase and the concen-
32.066
tration in the liquid phase along the catalyst bed length are shown
(L )0
(CNL )0 = (wN )0 (56) in Fig. 3. It has been established by steady-state simulations that
14.00674 the curve shape of hydrogen concentration in the liquid phase is
(L )0 dictated by a balance between mass transfer and reaction rate. This
(COL )0 = (wO )0 (57)
247.06 is much clearly observed from dynamic simulation results.
F.S. Mederos et al. / Applied Catalysis A: General 425426 (2012) 1327 21
6.0 0.2
5.5
5.0
0.1
4.5
4.0
0 5 10 15 20 25 30 0
Reactor length z, cm 0 5 10 15 20 25 30
Reactor length z, cm
3.0E-04
7.0E-05
6.0E-05
2.0E-04 5.0E-05
4.0E-05
3.0E-05
1.0E-04
2.0E-05
1.0E-05
0.0E+00
0 5 10 15 20 25 30 0.0E+00
Reactor length z, cm 0 5 10 15 20 25 30
Reactor length z, cm
Fig. 3. H2 partial pressure and liquid molar concentration along the catalyst bed
length in the bench-scale trickle-bed reactor at different times. () 60 s, ( ) 400 s, Fig. 4. H2 S partial pressure and liquid molar concentration along the catalyst bed
() 900 s, () 1700 s. length in the bench-scale trickle-bed reactor at different times. () 60 s, ( ) 400 s,
() 900 s, () 1700 s.
The variation of H2 S partial pressure in the gaseous phase and its Fig. 6 shows the effect of reactor temperature on the hydro-
concentration in the liquid phase along the catalyst bed length are genation of aromatics and compares the simulation results with
shown in Fig. 4. Similarly to H2 partial pressure and concentration experimental measurements. The model output was found to agree
in the liquid phase, H2 S proles follow the well-known tendencies. reasonably well with the experimental data. The conversion of pol-
Since H2 S is not present at the inlet of the reactor as the time passes, yaromatics reached a maximum at a reactor temperature of 360 C.
its partial pressure and concentration are increased. At short times, At high temperature the differences between experimental and
H2 S is produced until certain length of the reactor because sulfur simulated results for poly- and diaromatics are more notorious,
is present and reacts with H2 . After that point there is no more
reacting mixture and the reaction stops.
1.8E-04
The simulation results at steady-state (1700 s) were compared
with experimental measurements for HDS. The developed model 1.6E-04
LHSV = 1/h
was found to simulate the performance of the bench-scale TBR 1.4E-04 LHSV = 1.5/h
with high accuracy, obtaining errors in sulfur conversion prediction LHSV = 2/h
Sulfur, mol/cm3
5.0E-05 500
480
4.0E-05
460
Liquid temperature, C
440
Sulfur, mol/cm3
3.0E-05
420
400
2.0E-05
380
1.0E-05 360
340
0.0E+00 320
0 500 1000 1500 2000 2500 3000 3500 4000 4500
300
Time, s
0 100 200 300 400 500 600 700 800 900
600
Reactor length z, cm
Liquid temperature, C
550 Fig. 9. Simulated temperature proles of the liquid phase down through the catalyst
bed of commercial reactor with time. () 60 s, ( ) 400 s, () 900 s, () 1700 s.
500
because the difference of temperature between the liquid and cat-
450 alyst was of only 0.03 C. This result validates the simulation of
the heat balance with only one equation for a pseudo-phase, as it
400 was reported elsewhere [28]. This gure also demonstrates that the
increase in reaction temperature is higher in the initial part of the
350
reactor, due to the greater removal of impurities that occurs in this
zone, which has been also reported by others [18].
The transient behavior of the sulfur concentration in the com-
300
0 500 1000 1500 2000 2500 3000 3500 4000 4500
mercial HDT reactor is illustrated in Fig. 10. It can be observed that
there is a distinctive maximum in transient values, which is high-
Time, s
est at the top of the reactor and progressively decreases down the
Fig. 7. Proles of () sulfur concentration and temperature on commercial reactor. Those transient sulfur concentration proles together with
reactor, ( ) experimental sulfur concentration at the outlet of catalyst bed, the concentration proles of the other compounds shown in Fig. 11,
( ) dynamic sulfur concentration in experimental reactor, (- - -) experimen- are important for on-line tuning of the controllers settings in the
tal isothermal temperature (Commercial reactor, 340 C; 5.3 MPa; z = 853.44 cm;
control system of hydrotreating processes [23,25].
uL = 0.72 cm/s; uG = 8.93 cm/s. Bench-scale reactor, 340 C; 5.3 MPa; z = 25.2 cm;
uL = 1.81 102 cm/s; uG = 0.24 cm/s).
Fig. 12 shows the predicted concentration proles at steady-
state of the main compounds in liquid phase along the commercial
reactor: sulfur, nitrogen, poly-, di-, monoaromatics, naphthenes,
increasing catalytic bed temperature observed in the liquid phase and olens. From this gure it is possible to observe that the feed
of the adiabatic reactor. containing several types of mono-, di-, and condensed polyaro-
Fig. 9 reports the predicted evolution of temperature of the liq- matic structures, exhibits hydrogenation reactivities considerably
uid phase along the industrial reactor at different times. In this different. Since more moles of H2 are involved in the nal ring
gure the catalyst (or solid phase) temperature prole is not shown, hydrogenation (3 mol for monoaromatics compared with 1 or
6.0E-04 1.2E-04
1.1E-04
5.0E-04 1.0E-04
9.0E-05
Sulfur, mol/cm3
Sulfur, mol/cm3
4.0E-04 8.0E-05
7.0E-05
3.0E-04 6.0E-05
5.0E-05
2.0E-04 4.0E-05
3.0E-05
1.0E-04 2.0E-05
1.0E-05
0.0E+00 0.0E+00
0 100 200 300 400 500 600 700 800 900 0 200 400 600 800 1000 1200 1400 1600 1800
Reactor length z, cm Time, s
Fig. 8. Concentration proles of sulfur down through the catalyst bed of the com- Fig. 10. Dynamic sulfur concentration proles in the liquid phase of commercial
mercial reactor with time. ( ) Experimental bench-scale value, (lines) predicted reactor. ( ) z = 213.4 cm, ( ) z = 284.5 cm, ( ) z = 426.7 cm, (- - -)
values for commercial reactor. () 60 s, ( ) 400 s, () 900 s, and () 1700 s. z = 640.1 cm, () z = 853.4 cm, and ( ) experimental bench-scale value.
24 F.S. Mederos et al. / Applied Catalysis A: General 425426 (2012) 1327
1.4E-05 1.6E-04
Diaromatics, mol/cm3
Polyaromatics, mol/cm3
1.3E-05 1.8E-04
1.4E-04
1.2E-05 1.6E-04
1.1E-05 1.2E-04 1.4E-04
1.0E-05
9.0E-06 1.0E-04 1.2E-04
8.0E-06 1.0E-04
7.0E-06 8.0E-05
6.0E-06 8.0E-05
5.0E-06 6.0E-05
4.0E-06 6.0E-05
4.0E-05
3.0E-06 4.0E-05
2.0E-06 2.0E-05 2.0E-05
1.0E-06
0.0E+00 0.0E+00 0.0E+00
0 200 400 600 800 1000 1200 1400 1600 1800 0 200 400 600 800 1000 1200 1400 1600 1800 0 200 400 600 800 1000 1200 1400 1600 1800
Time, s Time, s Time, s
8.0E-04 1.7E-03 1.5E-04
1.6E-03 1.4E-04
7.0E-04 1.5E-03
Monoaromatics, mol/cm3
1.3E-04
1.4E-03 1.2E-04
6.0E-04 Naphthenes, mol/cm3 1.3E-03
1.1E-04
Olefins, mol/cm3
1.2E-03
1.1E-03 1.0E-04
5.0E-04 9.0E-05
1.0E-03
9.0E-04 8.0E-05
4.0E-04
8.0E-04 7.0E-05
7.0E-04 6.0E-05
3.0E-04 6.0E-04 5.0E-05
5.0E-04 4.0E-05
2.0E-04 4.0E-04
3.0E-04 3.0E-05
1.0E-04 2.0E-04 2.0E-05
1.0E-04 1.0E-05
0.0E+00 0.0E+00 0.0E+00
0 200 400 600 800 1000 1200 1400 1600 1800 0 200 400 600 800 1000 1200 1400 1600 1800 0 200 400 600 800 1000 1200 1400 1600 1800
Time, s Time, s Time, s
Fig. 11. Dynamic nitrogen, polyaromatics, diaromatics, monoaromatics, naphthenes, and olens concentration proles in the liquid phase of commercial reactor. ( ) z =
0 cm, ( ) z = 213.4 cm, ( ) z = 426.7 cm, (- - -) z = 640.1 cm, () z = 853.4 cm, and ( ) experimental bench-scale value.
2 mol for hydrogenation of poly- or diaromatics, respectively), The variations in gas oil and naphtha dynamic concentration
hydrogenation of the rst ring in polyaromatics is usually less within the commercial reactor are shown in Fig. 13. The concentra-
thermodynamically favored (with lower equilibrium constant val- tion proles of liquid compounds involved in mild HCR reactions
ues) than hydrogenation of the nal ring of the monoaromatics in the catalyst bed at steady-state are also given in Fig. 13.
at typical HDT conditions. On the other hand, it has also been Due to the lower acidity of the carrier (-Al2 O3 ), the commer-
observed the rst-ring hydrogenation of polyaromatics to be most cial CoMo catalyst used in our experiments produces a low yield
favored kinetically. The hydrogenation rates of subsequent rings of naphtha by HCR. From Table 2, it is observed, according to pre-
tend to become lower (see Table 2), and hydrogenation of the last exponential factor values, that HCR of gas oil to naphtha is not as
ring (corresponding to monoaromatics) proceeds with consider- easy as HCR of gas oil to light gases. This is because breaking side
able difculty compared with the initial hydrogenation steps [35]. chains of one to four carbons to produce light gases is easier than
For this study, monoaromatics hydrogenation was limited by ther- breaking long chains with more carbons into naphtha.
modynamic equilibrium, which favors the reverse reaction, as the The transient variation of hydrogen, hydrogen sulde, and
monoaromatics concentration increased along the catalytic bed ammonia partial pressures in the gaseous phase at different bed
instead of decreasing. It means that the diaromatics to monoaro- locations are shown in Fig. 14. In general, hydrogen partial pres-
matics forward reaction was much faster than the monoaromatics sure decreases along the reactor as a result of H2 consumption and
to naphthenes forward reaction, and the reduction of total aromat- increase in solubility, while hydrogen sulde and ammonia partial
ics is therefore small. pressures have the opposite behavior due to sulfur and nitrogen
removals, respectively. Methane partial pressure proles where
omitted because its yield by mild HCR reactions was negligible.
1.6E-03 Partial pressure proles of the main gaseous compounds in the
1.5E-03 simulated commercial HDT reactor at steady-state are presented
Liquid concentration, mol/cm3
1.4E-03
1.3E-03
Sulfur in Fig. 15. This gure illustrates the predicted steady-state partial
Nitrogen
1.2E-03 Poly-aro
pressures of H2 , H2 S, NH3 , and CH4 in the gas phase along the axial
1.1E-03 Di-aro position of the reactor. It can be seen that the H2 partial pressure
1.0E-03 Mono-aro in the gas phase decreases gradually, while the partial pressure of
9.0E-04 Naphthenes
H2 S in the gas phase increases along the reactor as HDS is occurring.
8.0E-04 Olefins
7.0E-04 According to Melis [17], the practically constant H2 partial pressure
6.0E-04 along the reactor means that HDT reaction was conducted under
5.0E-04 H2 excess. The partial pressure of CH4 presents a limited increase
4.0E-04 along the catalytic bed due to the beginning of the mild HCR reac-
3.0E-04
2.0E-04
tions (mainly formed by cracking of the alkyl groups), as reactor
1.0E-04 temperature also increased along the reactor. The production of
0.0E+00 NH3 by HDN reaction is negligible compared with H2 S formed from
0 100 200 300 400 500 600 700 800 900 HDS reaction, which could be explained by the low conversion of
Reactor length z, cm nitrogen compounds observed in Fig. 12.
The temperature proles in the commercial reactor for various
Fig. 12. Simulated concentration proles of liquid compounds along the commercial
reactor length at steady-state.
positions in the axial direction are given in Fig. 16. These proles
F.S. Mederos et al. / Applied Catalysis A: General 425426 (2012) 1327 25
3.5E-03 5
3.0E-03 4.5
4
2.5E-03
Gas oil, mol/cm3
3.5
Hydrogen, MPa
2.0E-03
3
1.5E-03 2.5
1.0E-03 2
1.5
5.0E-04
1
0.0E+00
0 200 400 600 800 1000 1200 1400 1600 1800 0.5
Time, s
0
9.0E-06 0 200 400 600 800 1000 1200 1400 1600 1800
8.0E-06 Time, s
7.0E-06 0.4
Naphtha, mol/cm3
6.0E-06 0.35
5.0E-06
3.0E-06 0.25
2.0E-06 0.2
1.0E-06
0.15
0.0E+00
0 200 400 600 800 1000 1200 1400 1600 1800 0.1
Time, s
3.0E-03 0.05
7.0E-06
0
2.9E-03 6.0E-06 0 200 400 600 800 1000 1200 1400 1600 1800
Gas oil, mol/cm3
Time, s
5.0E-06
Naphtha, mol/cm3
2.8E-03 0.008
4.0E-06
0.007
2.7E-03 3.0E-06
0.006
Ammonia, MPa
2.0E-06
2.6E-03 Gas Oil 0.005
Naphtha 1.0E-06
0.004
2.5E-03 0.0E+00
0 100 200 300 400 500 600 700 800 900
0.003
Reactor length z, cm
0.002
Fig. 13. Dynamic gas oil and naphtha concentration proles in the liquid phase of
commercial reactor and at steady-state along commercial catalyst bed. ( ) z = 0 cm,
( ) z = 213.4 cm, ( ) z = 426.7 cm, (- - -) z = 640.1 cm, () z = 853.4 cm,
0.001
and ( ) experimental bench-scale value.
0
0 200 400 600 800 1000 1200 1400 1600 1800
show the simulated change in liquid phase temperature with time, Time, s
starting from an initial temperature in the catalyst bed of 340 C,
over a time period of about 1700 s. In the axial direction, there are Fig. 14. Dynamic H2 , H2 S, and NH3 partial pressure proles in the commer-
strong temperature gradients. However, the temperature proles cial reactor. ( ) z = 0 cm, ( ) z = 213.4 cm, ( ) z = 426.7 cm, (- - -)
show a fairly smooth distribution in z-direction. z = 640.1 cm, () z = 853.4 cm, and ( ) experimental bench-scale value.
5.0
4.0
2.0
1.0
0.0
0 100 200 300 400 500 600 700 800 900
Reactor length z, cm
Fig. 15. Simulated partial pressure proles of gaseous compounds along the commercial reactor at steady-state.
500 1.1
450 z = 640.1 cm
z = 853.4 cm 0.7
0.6
400 0.5
z = 0 cm
0.4 z = 213.4 cm
z = 426.7 cm
0.3
z = 640.1 cm
0.2 z = 853.4 cm
350
0.1
0
0 200 400 600 800 1000 1200 1400 1600 1800
300 Time, s
0 200 400 600 800 1000 1200 1400 1600 1800
Time, s Fig. 17. Transient effectiveness factor for HDS reaction in the commercial reactor
at different bed positions.
Fig. 16. Transient temperature proles in the commercial reactor at different bed
positions.
When the concentration of limiting reactant within pores
becomes less by chemical reaction along the catalyst bed, the rate of
is mainly due to the consideration of rst order of reaction, and diffusion becomes higher than the reaction rate and then the effec-
thus independency of Thiele modulus on concentration of species tiveness factor value increases along the catalyst bed. Fig. 17 shows
[54]. Therefore, the prevailing effect is the rate constant, provoking that steady-state inside the catalyst pores was achieved faster than
decreased effectiveness factor for those positions far from reactor in the liquid phase, which justies to approximate the catalyst par-
entrance. ticle mass balance with an effectiveness factor [55].
Fig. 17 shows the transient behavior of the effectiveness fac-
tor for HDS reaction, assuming a typical reaction order of 1.7, at 5. Conclusions
different positions in the catalyst bed.
From the simulations with the developed dynamic HDT reactor
model, the following conclusions can be pointed out:
Table 3
Values of effectiveness factors for the various reactions at steady-state.a It is possible to simulate the most important HDT reactions simul-
Reaction Axial position (cm) taneously, provided that adequate kinetic model parameters are
determined from experimental data.
0 213.4 426.7 640.1 853.4
No wrong-way behavior is found when the reactant proles along
HDS 0.64 0.70 0.84 0.90 0.93 the catalyst bed in the rst section of the reactor have a pro-
HDN 0.88 0.81 0.71 0.62 0.58
nounced reduction or sharply slope.
HDAPoly 0.90 0.81 0.75 0.70 0.67
HDADi 0.88 0.82 0.77 0.73 0.70
It is not necessary to take into account the gas phase in the heat
HDAMono 0.98 0.98 0.98 0.98 0.98 balance, since only the solid and liquid phase heat balance equa-
HGO 0.88 0.92 0.93 0.93 0.93 tions or even with a single pseudohomogeneous heat balance
HCR 1.0 1.0 1.0 1.0 1.0
equation is enough to predict the increase of reactor temperature
a
T0 = 340 C, P = 5.3 MPa, LHSV = 2.5 h1 , gas/oil ratio = 356 std m3 /m3 . along the reactor length.
F.S. Mederos et al. / Applied Catalysis A: General 425426 (2012) 1327 27
The proposed dynamic heterogeneous model can be used as a [28] F.S. Mederos, M.A. Rodrguez, J. Ancheyta, E. Arce, Energy Fuels 20 (3) (2006)
tool to estimate kinetic parameters for further simulation of HDT 936945.
[29] F.S. Mederos, J. Ancheyta, Appl. Catal. A General 332 (1) (2007) 821.
reactors. [30] M.S. Rana, J. Ancheyta, P. Rayo, S.K. Maity, Catal Today 98 (2004) 151160.
[31] J. Chen, V. Mulgundmath, N. Wang, Ind. Eng. Chem. Res. 50 (2011) 15711579.
[32] A. Akgerman, G.M. Collins, B.D. Hook, Ind. Eng. Chem. Fundam. 24 (1985)
Acknowledgments 396401.
[33] S.M. Walas, Modeling with Differential Equations in Chemical Engineering,
The authors thank Instituto Mexicano del Petrleo for its nan- Butterworth-Heinemann, Boston, MA, USA, 1991, 450 pp.
[34] W.E. Schiesser, The Numerical Method of Lines Integration of Partial Differ-
cial support. F.S.M. and I.E. also thank the nancial support provided
ential Equations, Academic Press, Inc., San Diego, CA, USA, 1991, 326 pp.
by CONACyT. [35] A. Stanislaus, B.H. Cooper, Catal. Rev. Sci. Eng. 36 (1) (1994) 75123.
[36] S.B. Jaffe, Ind. Eng. Chem. Proc. Des. Dev. 13 (1) (1974) 3439.
[37] T.C. Ho, D. Nguyen, Chem. Eng. Comm. 193 (2006) 460477.
References
[38] O. Levenspiel, Chemical Reaction Engineering, third ed., John Wiley and Sons,
Inc., New York, USA, 1999, 668 pp.
[1] European standard for gasoline EN 228:1999 and for diesel EN 590:2004. [39] Software Polymath, Version 5.1, Numerical Solutions for Engineering and
[2] R.L. Irvine, D.M. Varraveto, PTQ Summer (1999) 3744. Science Problems, Cache Corp., Internet direction: http://www.polymath-
[3] A.P. Lamourelle, D.E. Nelson, J. McKnight, PTQ Summer (2001) 5156. software.com (accessed January 15, 2010).
[4] R.E. Palmer, S.P. Torrisi, PTQ Revamps Operations (February) (2004) 1518. [40] B. Carnahan, H.A. Luther, J.O. Wilkes, Applied Numerical Methods, John Wiley
[5] R. Chowdhury, E. Pedernera, R. Reimert, AIChE J. 48 (1) (2002) 126135. and Sons, Inc., New York, USA, 1969, 604 pp.
[6] M. Bhaskar, G. Valavarasu, A. Meenakshisundaram, K.S. Balaraman, Pet. Sci. [41] B.H. Cooper, B.B.L. Donnis, App. Catal. A General. 137 (2) (1996) 203223.
Technol. 20 (34) (2002) 251268. [42] Software Matlab, Release 2008a, The Language of Technical Computing, The
[7] G.F. Froment, G.A. Depauw, V. Vanrysselberghe, Kinetic Modeling, Ind. Eng. MathWorks Inc., Internet direction: http://www.mathworks.com (accessed
Chem. Res. 33 (12) (1994) 29752988. January 15, 2010).
[8] D.E. Mears, Chem. Eng. Sci. 26 (9) (1971) 13611366. [43] G.F. Froment, K.B. Bischoff, J. De Wilde, Chemical Reactor Analysis and Design,
[9] H.C. Henry, J.B. Gilbert, Ind. Eng. Chem. Proc. Des. Dev. 12 (3) (1973) 328334. 3rd ed., Wiley, New York, USA, 2010, 860 pp.
[10] M.P. Dudukovic, AIChE J. 23 (6) (1977) 940944. [44] B.H. Cooper, B.B.L. Donnis, B. Moyse, Hydroprocessing Conditions Affect Cata-
[11] H. Korsten, U. Hoffmann, AIChE J. 42 (5) (1996) 13501360. lyst Shape Selection, OGJ 84 (December (49)) (1986) 3944.
[12] D.G. Avraam, I.A. Vasalos, Catal. Today 7980 (14) (2003) 275283. [45] R.B. Bird, W.E. Stewart, E.N. Lightfoot, Transport Phenomena, John Wiley & Sons,
[13] E. Pedernera, R. Reimert, N.L. Nguyen, V. van Buren, Catal. Today 7980 (14) New York, USA, 2002, 895 pp.
(2003) 371381. [46] C.N. Sattereld, Mass Transfer in Heterogeneous Catalysis, MIT Press, Cam-
[14] M. Bhaskar, G. Valavarasu, B. Sairam, K.S. Balaraman, K. Balu, Ind. Eng. Chem. bridge, MA, 1970, 267 pp.
Res. 43 (21) (2004) 66546669. [47] M.O. Tarhan, Catalytic Reactor Design, McGraw-Hill, New York, 1983, 372 pp.
[15] Z.-M. Cheng, X.-C. Fang, R.-H. Zeng, B.-P. Han, L. Huang, W.-K. Yuan, Chem. Eng. [48] C.N. Sattereld, AIChE J. 21 (2) (1975) 209228.
Sci. 59 (2223) (2004) 54655472. [49] A. Bondi, Chem. Tech. 1 (1971) 185188.
[16] M.J. Macas, J. Ancheyta, Catal. Today 98 (12) (2004) 243252. [50] F. Mederos, I. Elizalde, J. Ancheyta, Catal. Rev. Sci. Eng. 51 (2009) 485607.
[17] S. Melis, L. Erby, L. Sassu, R. Baratti, Chem. Eng. Sci. 59 (2223) (2004) [51] J. Carberry, A. Varma, Chemical Reaction and Reactor Engineering. Chemical
56715677. Industries/26, Marcel Dekker, New York, USA, 1987, 1069 pp.
[18] M.A. Rodrguez, J. Ancheyta, Energy Fuels 18 (3) (2004) 789794. [52] F. Jimnez, K. Ojeda, E. Snchez, V. Karafov, R. Maciel Filho, in: V. Plesu,
[19] H. Yamada, S. Goto, Korean J. Chem. Eng. 21 (4) (2004) 773776. P.S. Agachi (Eds.), 17th European Symposium on Computer Aided Process
[20] K. Sertic-Bionda, Z. Gomzi, T. Saric, Chem. Eng. J. 106 (2) (2005) 105110. EngineeringESCAPE17, Elsevier B.V., The Netherlands, 2007, pp. 16.
[21] G.D. Stefanidis, G.D. Bellos, N.G. Papayannakos, Fuel Process. Technol. 86 (16) [53] Z. Liu, Y. Zheng, W. Wang, Q. Zhang, L. Jia, Appl. Catal. A General 339 (2) (2008)
(2005) 17611775. 209220.
[22] M. Oh, E.J. Jang, J. Korean Inst. Chem. Eng. (in Korean) 35 (5) (1997) 791798. [54] J.G. Schwartz, G.W. Roberts, Ind. Eng. Chem. Proc. Des. Dev. 12 (3) (1973)
[23] J. Chen, Z. Ring, T. Dabros, Ind. Eng. Chem. Res. 40 (15) (2001) 32943300. 262271.
[24] Y.-C. Chao, J.-S. Chang, Chem. Eng. Comm. 56 (16) (1987) 285309. [55] T. Salmi, J. Wrn, S. Toppinen, M. Rnnholm, J.P. Mikkola, Braz. J. Chem. Eng.
[25] M.A. Hastaoglu, B.E. Jibril, Chem. Eng. Comm. 190 (2) (2003) 151170. 17 (47) (2000) 10231035.
[26] I. Iliuta, Z. Ring, F. Larachi, Chem. Eng. Sci. 61 (4) (2006) 13211333. [56] S. Ueda, S. Yokohama, T. Ishil, K. Makino, G. Takeya, Ind. Eng. Chem. Proc. Des.
[27] P.R. Gunjal, V.V. Ranade, Chem. Eng. Sci. 62 (1820) (2007) 55125526. Dev. 14 (4) (1975) 493499.