Chemical Engineering Science 00 (2015) 000–000

Tulis judul di sini. Artikel ditulis dalam Bahasa indonesia

Author 1, Author 2, Author 3*
Program Studi Teknik Kimia, Universitas Islam Indonesia, Yogyakarta 55584, Indonesia

Corresponding author: email

Abstract

This study reviewed the value of the parameters of the kinetic models which were proposed by Akande and used the models to
evaluate the effect of mass transfer resistances. The values of parameter of kinetic models obtained in this study were nearly same
as those obtained by Akande. There was a typo at the value of activation energy on the original journal on the equation of kinetic
model based on mechanism. The equation should been written as –rA = (2.08 × 103e–4428/RTFA)/([1 + 3.83 × 107 FA]2. This typo
caused the model becomes invalid (Average Absolut Deviation (AAD) becomes 385.52%). During this study, it was found one
literature which refer to this journal wrote down this wrong equation. A questionable values was found on AAD for power law
model (–rA = 3.123 × 10–2 e–4400/RTFA0.43). Akande et al. (2006) reported the power law model had an AAD of 4.5%. This study
obtained the value of AAD for power law model of 8.98%. Mears’ criterion has been found to be a value of 5.59×10–3 thus the
external mass transfer resistance can be ignored. The internal diffusion has negligible effect to overall reaction rate since the Thiele
Modulus was found to be 0.57. The concentration difference between catalyst surface and catalyst core only ranged about 0.022%
– 0.032%.

Keywords: Journal review; Ni catalyst; Catalytic reforming; Mass transfer resistance; Internal diffusion

1. Introduction

Experimental study to obtain the model to describe the kinetic of hydrogen production using catalytic reforming of
crude ethanol over a 15%-Ni/Al2O3 catalyst has been performed in a packed bed reactor with inside diameter of 8 mm
and bed height of 32 mm, at temperature of 593, 693 and 793 K, at atmospheric pressure and ratios of catalyst weight
to mass flowrate of crude ethanol (W / FA0) of 779, 952, 1071, 1382, and 2143 kg cat. s / kg crude ethanol. The
composition of crude ethanol is shown in Table 1 (Akande et al., 2006)

Table 1 Composition of crude ethanol

Components Vol. percent Mole percent
Ethanol 12.0% 4.10%
Maltose 0.001% Trace
Glycerol 1.0% Trace
Lactic acid 1.0% Trace
Water 86.0% 95.90%
 Nomenclature

Ci0 Initial concentration of species i NSc Schmidt number

Cib Bulk concentration of species i (kmol m–3) -r'i Rate of formation of species i (kmol kg–1 s–1 )
Cis Conc. of species i (kmol m–3) at cat. surface Rc Catalyst radius (m)
Deff Effective diffusivity, (m2 s–1) v Superficial velocity of fluid (m s–1)
Fi Molar flowrate of species i (kmol s–1) ρb Bulk density of catalyst (kg m–3)
FT Total molar flowrate Φ2 Thiele Modulus
kc Mass transfer coefficient (m s–1) ρg Gas density (kg m–3)
n Order of reaction µ Dynamic viscosity (kg m–1s–1)
Ni Flux of species i (kmol m–1 s–1) ϵ Porosity
NRe Reynold number

The reaction of crude ethanol reforming can be represented by the following general equation

C2H6O + 3H2O  2CO2 + 6H2 (1)

Let C2H6O = A; H2O = B; CO2 = C; H2 = D.

Akande et al. (2006) proposed four models of the reaction mechanism which were developed with Eley Rideal
assumptions and one model based on power law. The 15 sets of kinetic data obtained from the experiments were used
to estimate the value of the parameters of the models based on mechanism and power law using a nonliniear regression
procedure with modified Levenberg-Marquardt algorithm.

Table 2 Kinetic models

Model No. Rate determining step Models AAD

1 Reaction mechanism 2080 e−4430/RT FA 6.0%
-rA =
[1+3.83 × 107 FA ]2

2 Power law rA =3.123 ×10-2 e-4410/RT FA0.43 4.5%

This study aims to reproduce the value of parameters of the best two models which were proposed by Akande et
al. (2006) as presented in Table 2. Average Absolute Deviation (AAD) method is used to describe the validity of the
models. The effect of mass transfer resistances also covered in this study. The value of parameters used to evaluate
mass transfer resistance obtained from previous work by Aboudheir, et al. (2005) before Akande et al. (2006) worked
on the kinetics model. The value of mass transfer resistance parameters is summarized in Table 3.

Table 3 Parameters for evaluating internal diffusion (see Notation section for symbols)

Parameter Value Units

Deff 3.2 × 10–7 m s–1
ρb 375 kg m–3
Rc 3.0 × 10–4 m
ρg 0.31. kg m–3

2. Methodology

The review on the parameter value were conducted by fitting the kinetics experimental data reported by Akande,
et al. (2006) to the proposed models by using NonlinearModelFit provided by mathematical software
Mathematica©. The mass transfer resistance including film resistance and internal diffusion were earlier evaluated
using Mears’ criterion and Thiele Modulus respectively. This study focused on internal diffusion. The best model
obtained from parameter review was used to evaluate the internal diffusion with reaction. NDSolve by Mathematica©
was used to solve the partial differential equation describing diffusion with reaction inside the catalyst.
3. Result and discussion

3.1. Review on the proposed model

The values of parameter for kinetic models obtained in this study were nearly same with the values obtained by
Akande et al. (2006). Table 4 shows the result obtained. This small difference might be caused by the different
algorithm. For the reaction Model #1, there was a typo written at the value of activation energy in the original journal.
The activation energy should had minus sign, otherwise the model becomes totally inaccurate (AAD up to 385.52%).
Unfortunately, this typo led the same mistake to another literature which refer to this journal, such as Guczi, L., and
Erdôhelyi A (2012).

Table 4 Value of parameters obtained by two approaches

Parameters Akande, et al. (2006) Mathematica ©
Model #1
k0 2080 2079.69
E 4430 4428.19
KA 3.83E–7 3.83E–7
AAD 6.05% 6.04%
Model #2
k0 3.123E–2 3.123E–2
E 4410 4405
Α 0.43 0.43
AAD 8.98%*) 9.06%
*)Akande, et. al. (2006) reported the value of AAD is 4.5%

The questionable result was found on the Average Absolute Deviation (AAD) for Model #2. Akande et al. (2006)
reported the power law model is the best model fit the kinetics data with AAD of 4.5%. In this study AAD of Model
#2 has been found to be 8.98%. Therefore, based on this study the best model is Model #1 which is based on
mechanism.

3.2. External mass resistance

The external mass transfer coefficient kc has been calculated using Eq. (2) based on the data on Table 3. The
Reynold Number and Schimdt Number were calculated using Eq. (3) and (4) respectively (Geankoplis, 2003). The
porosity of bed was assumed to be 0.4 (Fogler, 2006) and the gas viscosity was assumed to be same as steam viscosity
(1.2 × 10–5 kg m-1 s-1 ) since the gas contains only 4.12%-mole of ethanol. The left hand side of Eq. (5) (Mears’
criterion) which is often considered to determine mass transport limitation in the film (Fogler, 2006) has been found
to be a value of 5.59 × 10–3 (less than 0.15). Hence the external mass transfer limitations can be ignored safely then
this study discussed internal diffusion in detail.
0.4548 N-0.4069
Re v
kc = (2)
ϵ N2/3
Sc

𝜌 vd
p
NRe = μ(1𝑔 − ϵ) (3)
μ
NSc = (4)
𝜌𝑔 Deff

rA ρb R n
<15 (5)
kc CA
3.3. Internal diffusion with chemical reaction

The value of Thiele Modulus (Eq. 6) which is usually used to consider the effect of resistance due to internal
diffusion (Fogler, 2006) has been found to be a value of 0.57. Hence, the internal diffusion does not take any part on
the overall rate of reaction.

rA R2
Φ2 = (6)
Deff CAs

The evaluation on internal diffusion and reaction takes place in the porous were assumed at steady state for
simplicity. The variation in temperature in the spherical geometric pellet is negligible since the size of the catalyst is
very small (3 × 10–4 m). The mole balance over the shell thickness Δr described as following

FA |r -FA |r+∆r +r'A ρb A Δr = 0 (7)

NA |r A-NA |r+∆r A+r’A ρb A Δr = 0 (8)

d[NA 4πr2 ]
− r’A ρb 4πr2 = 0 (9)
dr

The reaction was carried out at constant temperature and pressure, and involving diluted crude ethanol as reactant,
therefore we still can write the equation describing the flux as the following

dCA
NA =-Deff (10)
dr

Substituting Eq. (10) into Eq. (9),

d dCA
[−Deff r2 ] − r’A ρb r2 = 0 (11)
dr dr

Applying derivation on Eq. (11) and dividing through –r2Deff

2 dCA d2 CA r’A ρb
+ 2 − =0 (12)
r dr dr Deff

In this case, the rate of reaction was measured in form of molar flowrate (FA). Eq. (13) provides correlation between
molar flowrate and concentration (Fogler, 2006).

Fj P T
Cj =CT0 ( ) ( ) ( 0 ) (13)
FT P0 T

The compressibility factor was assumed to be negligible since the experiments were carried out isothermally at high
temperature and isobaric at low pressure (atmospheric pressure). Therefore the value of Z0 on Eq. (15) is equal to one,
and Eq. (13) becomes Eq. (15).

P0
CT0 = (14)
Z0 RT0

FT
FA =CA ( ) (15)
CT0
Substituting equation of rate of reaction Model #1 into Eq. (12) yields,

2 dCA d2 CA 2080 e−4430/RT FA ρb

+ − =0 (16)
r dr dr2 [1+3.83 x 107 FA ]2 Deff

By substituting Eq. (15) into Eq. (16) we get

𝐹𝑇
d2 CA 2080 e−4430/RT 𝐶𝐴
2 dCA 𝐶𝑇0 ρb
+ − 𝐹 (17)
r dr dr2 [1+3.83 x 107 𝐶𝐴 𝑇 ]2 Deff
𝐶𝑇0

The boundary conditions are:

dCA
= 0, r=0 (18)
dr
CA = CAb, r = R (19)

Concentration at surface of the catalyst was assumed to be same with concentration at bulk because the effect of
external mass transfer can be neglected. The higher temperature leads higher consumption of crude ethanol inside the
catalyst due to increasing of the rate of reaction. The concentration different between the surface and core of the
catalyst is inversely proportional with the molar flowrate of reactant (concentration of ethanol), as presented in Table
5. This caused by the driving force of mass transfer. The higher concentration at surface the higher driving force for
diffusion. The internal diffusion resistance has very small effect toward concentration variation inside the catalyst.
This study reported the concentration difference between surface and core only ranged about 0.022% – 0.032%. Fig.
1 shows the decrease of concentration difference inside the catalyst is linearly proportional with the increase of
concentration of crude ethanol in bulk (CAb). The concentration profile of ethanol inside the catalyst with bulk
concentration CAb of 5.99 × 10-4; 6.92 × 10-4; and 8.37 × 10-4 kmol s-1 are presented in Fig. 2.

Table 5 Concentration difference of crude ethanol between surface and core

CA 593 K 693 K 793 K

5.9980E-04 0.025% 0.029% 0.032%
6.9258E-04 0.024% 0.027% 0.030%
8.3777E-04 0.022% 0.025% 0.028%

0.035%
0.030%
0.025%
0.020%
ΔCA

0.015%
0.010% 793 K
0.005% 693 K
593 K
0.000%
5.5E-04 6.5E-04 7.5E-04 8.5E-04
CA
Fig. 1. The decrease of concentration difference (ΔCA) vs bulk concentration of A (CAb)
 0.00059985

0.00059980

CA 0.00059975 593 K 693 K 793 K

0.00059970

0.00059965

0.00000 0.00005 0.00010 0.00015 0.00020 0.00025 0.00030

r
(a)
0.00069255

0.00069250 593 K 693 K 793 K

CA
0.00069245

0.00069240

0.00000 0.00005 0.00010 0.00015 0.00020 0.00025 0.00030

r
(b)
0.00083780

0.00083775

0.00083770 593 K 693 K 793 K

CA
0.00083765

0.00083760

0.00083755

0.00000 0.00005 0.00010 0.00015 0.00020 0.00025 0.00030

r
(c)
Fig. 2. Concentration profile inside the catalyst at CAb of (a) 5.99 × 10-4; (b) 6.92 × 10-4; and (c) 8.37 × 10-4 kmol s-1
4. Conclusion

The value of parameters obtained by Akande et al. (2006) has been confirmed by using NonlinearModelFit
provided by mathematical software Mathematica©. There was a typo in original journal on the kinetic model based
on the mechanism. The value of activation energy should has minus sign (–rA = (2.08 × 103e–4428/RT NA)/([1 + 3.83 ×
107 NA]2). This mistake caused the model becomes invalid (AAD becomes 385.52%). During this study, there was one
literature which refer to this journal write the equation with the mistake. A questionable values is found on AAD for
power law model of -ra = 3.123 × 10–2 e–4400/RTNA0.43. Akande et. al. (2006) reported the power law model had an
AAD of 4.5%. This study obtained the value of AAD for power law model is 8.98%. Mears’ criterion has been found
to be a value of 5.59 × 10–3 thus the resistance of external mass transfer can be ignored. The profile of crude ethanol
concentration inside the catalyst was predicted based on the assumption of steady state and no temperature variation
in spherical geometric. The internal diffusion resistance has no effect toward overall reaction rate. The value of Thiele
Modulus has been found to be 0.57. The concentration difference between catalyst surface and catalyst core only
ranged about 0.022% – 0.032%.

References

1. A. Akande, A. Aboudheir, R. Idem, A. Dalai, International Journal of Hydrogen Energy. 31 (2006) 1707 – 1715
2. A. Aboudheir, A. Akande, R. Idem and A. Dalai, International Journal of Hydrogen Energy. 31 (2006) 752 – 761
3. Geankoplis, C.H., 2003. Transport Process and Separation Process Principles 4 th ed., Wiley, New York
4. H. S. Fogler, 2006. Elements of Chemical Reaction Engineering 4thed, Pearson Education, New Jersey
5. L. Guczi, and A. Erdôhelyi (eds.), 2012. Catalysis for Alternative Energy Generation, Springer, New York