Model-free kinetics applied to regeneration of coked alumina

Article abstract of DOI:10.1023/B:JTAN.0000027163.44593.4a

Thermogravimetry is proposed to study the alumina catalyst regeneration deactivated by coke, after being used in the transformation of styrene in a fixed bed continuous flow reactor. The model-free kinetic approach has been applied to data for the thermal oxidation of carbonaceous deposits on the catalyst. The activation energy (E) was calculated as a function of α (conversion) and T (temperature), by using Vyazovkin model-free kinetic method, allowing to estimate time required to remove coke at a given temperature.

Full text of DOI:10.1023/B:JTAN.0000027163.44593.4a

Journal of Thermal Analysis and Calorimetry, Vol. 75 (2004) 687–692
MODEL-FREE KINETICS APPLIED TO
REGENERATION OF COKED ALUMINA
1
G. J. T. Fernandes , A. S. Araújo , V. J. Fernandes Jr. and Cs. Novák
1
*1
2
1
Universidade Federal do Rio Grande do Norte, Departamento de Química, C.P. 1662,
5
9072-970 Natal, RN, Brazil
2
Hungarian Academy of Sciences, Budapest University of Technology and Economics, Research
Group of Technical Analytical Chemistry, H-1111 Budapest, Szt. Gellért tér 4, Hungary
Abstract
Thermogravimetry is proposed to study the alumina catalyst regeneration deactivated by coke, after
being used in the transformation of styrene in a fixed bed continuous flow reactor. The model-free
kinetic approach has been applied to data for the thermal oxidation of carbonaceous deposits on the
catalyst. The activation energy (E) was calculated as a function of α (conversion) and T (tempera-
ture), by using Vyazovkin model-free kinetic method, allowing to estimate time required to remove
coke at a given temperature.
Keywords: alumina deactivation, catalyst regeneration, model-free kinetics
Introduction
One of the major problems related to the operation of heterogeneous catalysis is the cat-
alyst loss of activity with time-on-stream, denominated deactivation. Mechanisms of
carbon deposition and coke formation on catalysts have been detailed in several re-
views [1–4]. One of the main causes of catalysts deactivation is the formation and de-
position of coke [5], which results from the nucleation of non-reactive polymeric spe-
cies on the catalyst surface, blocking the channels and cages [6]. Catalyst regeneration
is generally achieved by gradual heating under oxidizing atmosphere [7]. The applica-
tion of thermal analysis for the investigation of catalytic systems and catalysts has been
reported in several papers [8, 9]. In this work, thermogravimetry is used for studying
the kinetic parameters of alumina regeneration employing integral TG curves and a
model-free kinetic method to calculate the activation energy, the conversion rates and,
in particular, the coke removal time as a function of temperature.
*
Author for correspondence: E-mail: vjfj@superig.com.br
1
388–6150/2004/ $ 20.00
Akadémiai Kiadó, Budapest
©
2004 Akadémiai Kiadó, Budapest
Kluwer Academic Publishers, Dordrecht

6
88
FERNANDES et al.: ALUMINA REGENERATION
Experimental
The alumina catalyst was synthesized by the impregnation method with aluminum ni-
trate solution, followed by calcination at 823 K under nitrogen flow. The resulting cata-
lyst, which is a white powdery solid, was used in the transformation of styrene in a
fixed bed continuous flow reactor at 450°C, with MHSV (mass hourly space velocity)
–1
of 50 h , for 4 h. Under these conditions, the catalysts became deactivated by coking.
To study the regeneration of the coked catalyst, a Mettler TG/SDTA 851
thermobalance was calibrated over all heating rates, using a gas purge, in the same
conditions as those of the analysis.
Samples of the deactivated catalysts, containing 17.0% (w/w) of organics, were
submitted to a pre-treatment under a dry air atmosphere at 30°C. They were then heated
–1
in the temperature range 30–900°C using heating rates of 5.0, 10 and 20°C min , with
–1
3
a controlled dry air flow of 120 cm min . 0.01 g of coked catalyst was used for each
experiment.
Results and discussion
The reaction rate of a chemical reaction depends on conversion (α), temperature (T)
and time (t). The reaction rate represented as a function of conversion f(α) is different
for each process and must be determined experimentally. For simple reactions, the
th
evaluation of f(α) with the n order is possible. For complex reactions the function of
th
(
α) is complicated and generally unknown; in this cases the n order algorithm causes
unreasonable kinetic data. With the model-free kinetics more accurate evaluations of
complex reactions can be performed, as a trustworthy way of obtaining reliable and
consistent kinetic information about the overall process [10].
Vyazovkin [11, 12] developed an integral kinetic method where no model has to
be selected (model free-kinetics) which allows to evaluate both simple and complex
reactions, using multiple heating rates. This method has been used for thermal stabil-
ity evaluation of different systems [13]. The theory is based on that f(α) and that the
activation energy E(α) is constant for a certain conversion (isoconversional method).
A chemical reaction is measured at least at three different heating rates (β) and
the respective conversion curves are calculated out of the TG measured curves.
2
For each conversion α, ln(β/T ) is plotted vs. 1/T
, giving a straight line with the
α
slope –E /R, therefore the activation energy is obtained as a function of conversion.
α
α
Taking the reaction rate equation, presented as f(α), and dividing by the heating
rate β=dT/dt:
∂
∂
α
t
∂α
∂T
= k f (α)
=
kf (α)→
(1)
β
–
1
–1
where ∂α/∂t – reaction rate (s ), k – rate constant (s ), α – conversion, β – heating
–
1
rate (K s ).
Substituting k by the Arrhenius expression k=Ae
–
E/RT
and rearranging gives:
J. Therm. Anal. Cal., 75, 2004

FERNANDES et al.: ALUMINA REGENERATION
689
1
k0 –E/ RT
∂
α=
e
∂T
(2)
f (α)
β
Integrating up to conversion, α (at the temperature T) gives:
α
T
1
k0
–E/ RT
∂
α=g(α) =
e
∂T
(3)
(4)
(5)
∫
∫
f (α)
β
0
T0
Since E/(RT)>>>1, the temperature integral can be approximated by:
T
–
E/ RT
R
2 –E/ RT
e
∂T ≈ T e
∫
E
T0
Substituting the temperature integral, rearranging and logarithming, gives:
β
 Rk0
=ln

Eα
1
ln
 –

2

Tα
E g(α)
R Tα

α

This is defined as dynamic equation, which is used for the determination of the
activation energy for all conversion values (α).
The first step in the TG curve, from 30 to 450°C, is due to hydrocarbons ad-
sorbed during the cracking reaction (Fig. 1).
–
1
3
Fig. 1 TG and DTG curves of coked alumina. Heating rate 10 K min ; dry air flow 120 cm min
–1
In order to determine accurately the temperature range of coke thermal oxida-
tion, TG and DTG curves were analysed for each heating rate. For example, using a β
–
1
value of 10°C min , the thermal oxidation of coke occurs from 450 to 750°C (Fig. 1).
It can be noted that the thermal region of coked removal is a direct function of
the temperature in the samples and directly proportional to the heating rate employed
(Fig. 2), which must be considered for pilot plant estimations. Conversion curves as a
function of temperature are also shown in Fig. 3.
J. Therm. Anal. Cal., 75, 2004

6
90
FERNANDES et al.: ALUMINA REGENERATION
The activation energy (E) was calculated as a function of α (conversion) and T
temperature), by using Vyazovkin model-free kinetic method, varying from 150
to 250 kJ mol [α=0.1 to 0.9] (Fig. 4a).
(
–
1
Conversion of coke deposited on alumina as a function of time, for different
temperatures, was calculated, noting that the conversion time decreases considerably
as a function of temperature (Fig. 4b). It was also possible to estimate temperature for
removing coke as a function of time, allowing to estimate time required to remove
coke for a given temperature (Table 1). It was observed that, to remove 90% of coke
from the alumina in a period of 10 h, it would be necessary to carry out thermal oxida-
tion at 416°C, with the same purge flow.
Table 1 Temperature (°C) for removing coke as a function of time, for different conversions
Conversion/%
Time/min
1
0
25
50
75
90
95
–
6
0
235.30
290.72
275.01
266.22
260.15
255.54
251.83
248.73
246.07
243.76
241.70
239.85
238.18
236.65
235.24
233.94
232.72
231.59
230.52
229.52
228.57
227.67
226.82
226.00
–
374.76
360.10
351.83
346.09
341.72
338.19
335.23
332.69
330.47
328.50
326.73
325.12
323.65
322.29
321.03
319.86
318.77
317.74
316.76
315.85
314.97
314.15
313.36
432.51
416.12
406.88
400.47
395.58
391.64
388.34
385.52
383.04
380.84
378.87
377.07
375.43
373.92
372.52
371.21
369.99
368.84
367.76
366.74
365.77
364.84
363.97
–
–
1
1
2
3
3
4
4
5
6
6
7
7
8
9
9
20
80
40
00
60
20
80
40
00
60
20
80
40
00
60
222.34
–
–
215.07
–
–
210.03
–
–
206.20
445.62
441.55
438.15
435.22
432.67
430.39
428.35
426.49
424.80
423.23
421.78
420.43
419.16
417.97
416.85
415.79
414.79
413.83
412.92
–
–
203.11
200.53
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
1
1
1
1
1
1
1
1
020
080
140
200
260
320
380
440
J. Therm. Anal. Cal., 75, 2004

FERNANDES et al.: ALUMINA REGENERATION
691
Fig. 2 Multiple heating rate thermogravimetric integral curves for different coke re-
moval rates in % (w/w). Heating rates: a – 5, b – 10 and c – 20 K min
–1
Fig. 3 a – Multiple heating rate TG curves; b – conversion vs. temperature for coke
thermal oxidation in alumina
Fig. 4 a – Activation energy curve and b – conversion of coke (in %) deposited on alu-
mina vs. time for different temperatures
J. Therm. Anal. Cal., 75, 2004

6
92
FERNANDES et al.: ALUMINA REGENERATION
This work was funded by the CNPq (National Council of Scientific and Technological Development)
whose aid is gratefully acknowledged. The authors are very grateful to Petroflex SA and Micronal.
References
1
2
3
4
5
6
7
8
9
J. Haber, J. H. Block and B. Delmon, Pure Appl. Chem., 67 (1995) 1257.
C. N. Satterfield, in: Heterogeneous Catalysis in Industrial Practice, McGraw-Hill, New York.
C. H. Bartholomew, Chem. Eng., 12 (1984) 97.
E. B. Maxted, Adv. Catal., 3 (1951) 129.
R. Hughes, Deactivation of Catalysts, Academic Press, London 1984, p. 3.
P. Magnoux, M. Guisnet, S. Mignard and P. Cartraud, J. Catal., 117 (1989) 495.
P. G. Menon, J. Mol. Catal., 59 (1990) 207.
S. Vyazovkin and C. A. Wight, Thermochim. Acta, 340–341 (1999) 53.
Z. Sarbak and K. Surma, J. Therm. Anal. Cal., 72 (2003) 159.
1
1
1
1
0 K. Wieczorek-Ciurowa, K. Gamrat and Ju. G. Shirokov, J. Therm. Anal. Cal., 72 (2003) 323.
1 S. Vyazovkin and V. Goryachko, Thermochim. Acta, 194 (1992) 221.
2 S. Vyazovkin and A. I. Lesnikovich, Russ. J. Phys. Chem., 62 (1988) 2949.
3 V. J. Fernandes Jr., A. S. Araújo and G. J. T. Fernandes, J. Therm. Anal. Cal., 64 (2001) 807.
J. Therm. Anal. Cal., 75, 2004