A Kinetic Model for Catalytic N-Butane Oxidative Dehydrogenation under Oxygen-Free Reaction Conditions in a Fluidized CREC Riser Simulator

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

Comments for author File: Comments.pdf

Comments on the Quality of English Language

Author Response

Reply to Reviewer #1

In the paper, the authors develop a kinetic model for the oxidative dehydrogenation of n-butane on a V/MgO-Al2O3 catalyst. They conduct experiments in a fluidized CREC Riser Simulator at four different temperatures. In the kinetic model, they consider cracking, oxy-dehydrogenation, and complete oxidation reactions as part of the mechanism, and employ Langmuir-Hinshelwood-type kinetics for the reaction kinetics. Overall, the science in the paper is acceptable.

We appreciate the positive evaluation of our work by the Reviewer.

However, the paper is not well-written, with disorganized sections that make it difficult to follow.

Our native English-speaking editor checked the paper thoroughly before submission. However, in view of the comment by the Reviewer, the manuscript was checked again. We believe that with this additional revision, the updated manuscript complies with the high writing standards of the Catalysts journal.

Major Revisions:

1. The order of the section’s arrangement is totally a disaster. Please double check before the submission. Section 3 should be in front of section 2.

Regarding the issue raised by the Reviewer, we would like to state that it is advised by the Catalysts journal to organize manuscript chapters, as follows: a) Introduction, b) Results, c) Methods, d) Conclusions. This is exactly what we did in the first version of the manuscript. However, and to better comply with the Catalysts journal guidelines, we made some minor changes to the titles of various sections as follows:

• Introduction
• Results
• Experimental Methods
• Kinetic Modelling Methods
• Conclusions

2. There are two Figures 4.

It is true. There were two figures designated as Figure 4. The error is now rectified in the revised manuscript. Furthermore, all figures following Figure 4 were also updated.

3. How the adsorption energy parameters listed in table 3 are calculated are unclear. Are

they fitted from TPD experiments or obtained from some literature.

 The adsorption parameters were not taken from the literature. They were calculated, as reported on page 8 of the manuscript, using Eqs (4) and (5). This is also described further, in a recently published paper (Bin Sulayman et al (17)). However, to address the issue raised by the Reviewer, the text on page 8 was expanded, as follows: 

“To address this issue, studies were developed in the mini-fluidized CREC Riser Simulator where adsorption constants for the various chemical species on the g-alumina support, were determined. The amounts of adsorbed species including n-butane, butene, and carbon oxides were also established, at 475°C, 500°C, 525°C, and 550 °C, in this lab-scale unit. Furthermore, our research team calculated the adsorbed species amounts by solving numerically Eq(4), and (Eq 5), and by using a fourth order Runge-Kutta least square method, as reported in (17).”.

 

And later:

 

“Regarding T_m , the selected median temperature, it was 513 °C. The adsorption parameter regression fitting process was conducted by using a NonlinearModelFit built function in Wolfram Mathematica (17). Table 3 reports the adsorption constants and heats of adsorption, for the dominant species involved in the ODH butane conversion.

“

4. The fitting shown in figure 4 line 282 are too good. Usually when do the fitting, there are always experimental error or some missing kinetics by the model leading to some deviation between model and experiments. Especially for a model with only three reactions as mechanism.

The experimental data reported in Figure 4, are average values of 3-4 complete repeat runs. There were 6 injections of n-butane per run and 5 out of 6 consecutive injections (except the first one) were considered. The catalyst was regenerated after every complete run. Thus, every symbol in this figure represents between 15-20 data points.

Every figure symbol (average values of 15-20 data points) has an associated standard deviation, as reported in the caption of Figure 4.  Thus, the average experimental data reported falls within a probability band with a reported standard deviation (e.g +/-6-8%). To clarity this matter further, the caption of Figure 4 was modified as follows:

Figure 4: Comparison Between the Average Experimental Data and Kinetic Model Predictions, at 475° C, 500°C, 525°C, 550 °C. Typical standard deviation for 3 repeat runs: +/- 6.05-8.15 %”. Notes: a) The average experimental data points (  repeat runs with   being calculated for 4 out of 6 consecutive injections (excluding the first and second one), b) The ODH catalyst was regenerated after every repeat 6 injection run, b) the “i” subscript refers to the n-butane, butene and Cox species under consideration.

 

5. In line 238. It is unclear how the parameter estimation is built. Are they using all experimental data points to formulate the objective function in equation 6, then what do they mean “every experimental condition being used repeatedly 3 times.”

 

The meaning of this statement is now clarified as follows:

As reported in Table 4, a high DOF (degrees of freedom) was used in this analysis, to adequately calculate the six model parameters. The degree-of freedom (DOF) was 570 and accounted for 576 experimental data points minus the six kinetic constants numerically regressed. This was the case given that every experimental condition considered included the third to the sixth butane consecutive injections (excluding the first and second ones), with runs being repeated 3 times. The catalyst was regenerated after every complete run.

 

6. In line 277, the author claim “reveals that the estimated parameters are weakly correlated. This further validates the reliability of our proposed kinetic model.”. It is unclear to me why weakly correlated parameters lead to better model.

Regarding the point raised by the Reviewer, we would like to state that in a multiparameter estimation problem, there is more than one possible parameter set that minimize theoretical prediction and experimental values. Let us consider the adsorption parameters, in Table 3.  One must minimize their connections, in order to reduce the determination parameter uncertainty.

One possible approach, as described by Equation 5, is the use of the Arrhenius equation using the  centering temperature (Tm). One can reduce in this way, parameter interaction, and calculate parameters with narrow 95% confidence spans (or little uncertainty), restricting  as a result their possible variability to narrow ranges. See Table 3 below.

Parameter (atm-1)	Estimated value with 95% confidence spans	Parameter (KJmol-1)	Estimated value with 95% confidence spans
	0.83 ± 0.03		31.2 ± 0.72
	0.41 ± 0.012		60.8 ± 1.20
	0.50 ± 0.018		54.0 ± 1.05

Table 3. Adsorption Parameters of Different Species.

Thus, in order to clarify this, the following text was included, on page 8 (line 229):

“The narrow 95% confidence spans show the limited parameter variation and as a result, their potential phenomenological value.”.

Furthermore, in the case of the kinetic model for ODH, there are 6 parameters (3 energies of activation and 3 pre-exponential factors) that need to be adjusted numerically. In this situation, there are two requirements to calculate the kinetic constants satisfactorily: a) Narrow confidence intervals for all parameters, b) Cross-correlation coefficients smaller than 1, in the cross-correlation matrix. Both of these conditions are met with the data obtained and shown in Table 4. To further explain this matter, the clarification of “coefficients being smaller than one”, is now included in the revised manuscript on line 259, as follows:

“Table 4 shows that the estimated parameters were defined with reduced and desirable 95% confidence intervals. The other important information that can be seen in Table 4 is the small degree of cross-correlation between parameters (smaller than 1), with all reported cross-correlation coefficients being below 0.89.”.  

7. In figure 3, it is missing leading that the summing of the selectivity of O4-olefin and COx

is not 100% at different temperatures.

There is nothing missing in Figure 3, as suggested by the Reviewer. The addition of C4H8 and COx selectivities does not add up 100%. The addition of these two selectivities while close to 100% is not 100%. This is the case given that CH4, C2, C3 and coke species selectivities (refer to Table 2) are small and cannot be reported  very effectively, in Figure 3 given the difference in magnitude. All these details are provided, in Table 3.

Minor:

1. There are some mathematic notion errors in the documents, such as in “z 5 wt%

V/MgO−γAl2O3 Catalyst” in line 233 (what is z?). and the name of the catalyst is

changing from between “5 wt% V/MgO−γAl2O3” and “5 V/MgO−γAl2O3”, etc. Please

proofread before submission

Thank you for the observations: a) There was an error. “5wt%V/MgO−γAl2O3” is correct and has been rectified, b) The 5wt%V/MgO−γAl2O3catalyst designation is now employed throughout the article.

 

Author Response File: Author Response.pdf

Reviewer 2 Report

Comments and Suggestions for Authors

Review of the article «A Kinetic Model for Catalytic N-Butane Oxidative Dehydrogenation under Oxygen-Free Reaction Conditions in a Fluidized CREC Riser Simulator», A. B. Sulayman, H. de Lasa

The production of low molecular weight olefins such as butenes (1-butene, cis- and trans-butene-2, butadiene) by dehydrogenation of n-butane is of great industrial interest. However, the direct dehydrogenation reaction of lower alkanes is thermodynamically limited and highly endothermic. The work is devoted to the promising process of oxidative dehydrogenation of n-butane. A lot of interesting data was obtained, but there are a few comments:

1.     Under these reaction conditions, the isomerization reaction of n-butane is thermodynamically favorable, but there is no data on the formation of isobutane in the work. Tell me, was the formation of isobutane observed?

2.     In what state is vanadium on the surface of the support? Which vanadium oxide particles are the active centers in the process of oxidative dehydrogenation of n-butane?

3.     How stable are these catalytic systems at one temperature?

Author Response

Reply to Reviewer 2

The production of low molecular weight olefins such as butenes (1-butene, cis- and trans-butene-2, butadiene) by dehydrogenation of n-butane is of great industrial interest. However, the direct dehydrogenation reaction of lower alkanes is thermodynamically limited and highly endothermic. The work is devoted to the promising process of oxidative dehydrogenation of n-butane. A lot of interesting data was obtained, but there are a few comments:

We would like to thank this Reviewer for the positive assessment of our work.

1. Under these reaction conditions, the isomerization reaction of n-butane is thermodynamically favorable, but there is no data on the formation of isobutane in the work. Tell me, was the formation of isobutane observed?

The n-butane gas sample used in the experiments contained 99% butane and 1% iso-butane. We did not observe additional isobutane formation with the catalyst of the present research. This is consistent with the findings of others, who used fluidizable catalysts of the same family (20). The type of ODH catalyst used in the present study appears to operate with very restricted isomerization activity.

2. In what state is vanadium on the surface of the support? Which vanadium oxide particles are the active centers in the process of oxidative dehydrogenation of n-butane?.

This issue was addressed in a recent paper of our team (17), co-authored by Bin Sulayman et al. In that manuscript, the 5wt%V/MgO−γAl2O3 ODH catalyst was fully characterized using a number of techniques: BET, TPD/TPR, MH3-TPD, XRD, FTIR, LRS and XPS. In particular, XPS analysis showed that dominant catalyst oxidation states, of the V₂O₅ species on the MgO-γAl₂O₃ catalyst surface, during ODH, are likely a blend of V⁺⁴ and V⁺⁵ species.

In order to clarify this matter, the following statement, was introduced, on page 2 of the revised manuscript.

“Vanadium oxide-based catalysts are among the metal/support ODH catalysts that have been examined in previous research [14–16]. Furthermore, a series of magnesium oxide-modified vanadium oxide-based catalysts for alkane ODH, have also been recently considered by our research group. These catalysts were characterized using BET, TPD/TPR, MH3-TPD, XRD, FTIR, LRS and XPS. In particular, XPS analysis showed that the dominant catalyst oxidation species on the MgO-γAl₂O₃ catalyst, are likely a blend of V⁺⁴ and V⁺⁵ species[17].”.

3. How stable are these catalytic systems at one temperature?

The catalyst developed is very stable, as shown in Table 2. In order to clarify this matter further, the following statement was introduced in the revised manuscript:

“Table 2 reports the various chemical species selectivities and butane conversions obtained for one complete run, at 10s reaction time, and 500°, 525°C, 550°C. Every run was developed using six consecutive injections,with catalyst regeneration being conducted, after every run. All runs as the ones reported in Table 2, were repeated three times. One can observe that for injections 3 to 6, reproducible butane conversions and C2, C3, C4 olefin selectivities were obtained. All this points to the adequate and stable performance of the 5wt%V/MgO-γAl₂O₃ catalyst. It has to be mentioned that only injection 1 and 2 yielded some extra butane conversion, likely due to the butane reaction with the catalyst labile oxygen, and as result were not used in the kinetic model analysis”.

 

 

 

Author Response File: Author Response.pdf

Round 2

Reviewer 2 Report

Comments and Suggestions for Authors

All comments have been eliminated, the publication can be accepted in its present form.

Author Response

Thanks for the time to review our manuscript. I have replied to all your comments.