Development of catalyst libraries for total oxidation of methane: A case study for combined application of "holographic research strategy and artificial neural networks" in catalyst library design

A. Tompos, József L. Margitfalvi, E. Tfirst, Lajos Végvári, Mohyeddin A. Jaloull, Hamza A. Khalfalla, Mohammed M. Elgarni

Research output: Contribution to journalArticle

33 Citations (Scopus)

Abstract

Using tools of combinatorial catalysis and high-throughput experimentation techniques new multi-component catalysts have been designed and tested for total oxidation of methane. The compositions of catalysts were optimized by means of holographic research strategy (HRS). In the first catalyst generation, the best hit resulted in 44% conversion of methane. However, after designing and testing 167 compositions in five generations, the best catalysts resulted in practically complete conversion of methane at 350 °C. The supports of the best catalysts found by HRS consist of mostly Ce oxide and small amount of La. In the best catalyst, the concentration of Pt, Pd and Au is 2.3, 2.3 and 0.1%, respectively. In order to obtain the catalytic activity versus composition relationship, artificial neural networks (ANNs) have been trained using catalytic results of the HRS optimization. Upon combining HRS and ANNs, "virtual" catalytic experiments were performed in order to (i) find "virtual" optimum compositions and (ii) map the full experimental space in two dimensions. Results obtained in this study proved that HRS is a very powerful tool both in catalyst library design and visualization of the experimental space. The combination of HRS with ANNs appeared to be an excellent method for knowledge extraction. In this way, further new information can be obtained about the catalytic system investigated.

Original languageEnglish
Pages (from-to)65-78
Number of pages14
JournalApplied Catalysis A: General
Volume285
Issue number1-2
DOIs
Publication statusPublished - May 10 2005

Fingerprint

Methane
Neural networks
Oxidation
Catalysts
Chemical analysis
Catalyst supports
Oxides
Catalysis
Catalyst activity
Visualization
Throughput
Testing

Keywords

  • Artificial neural networks
  • Catalyst library design
  • Combinatorial catalysis
  • High-throughput experimentation
  • Information mining
  • Methane oxidation
  • Multi-component catalysts

ASJC Scopus subject areas

  • Catalysis
  • Process Chemistry and Technology

Cite this

Development of catalyst libraries for total oxidation of methane : A case study for combined application of "holographic research strategy and artificial neural networks" in catalyst library design. / Tompos, A.; Margitfalvi, József L.; Tfirst, E.; Végvári, Lajos; Jaloull, Mohyeddin A.; Khalfalla, Hamza A.; Elgarni, Mohammed M.

In: Applied Catalysis A: General, Vol. 285, No. 1-2, 10.05.2005, p. 65-78.

Research output: Contribution to journalArticle

@article{e1b62f1e97ba47c1901ed6eedf3e8572,
title = "Development of catalyst libraries for total oxidation of methane: A case study for combined application of {"}holographic research strategy and artificial neural networks{"} in catalyst library design",
abstract = "Using tools of combinatorial catalysis and high-throughput experimentation techniques new multi-component catalysts have been designed and tested for total oxidation of methane. The compositions of catalysts were optimized by means of holographic research strategy (HRS). In the first catalyst generation, the best hit resulted in 44{\%} conversion of methane. However, after designing and testing 167 compositions in five generations, the best catalysts resulted in practically complete conversion of methane at 350 °C. The supports of the best catalysts found by HRS consist of mostly Ce oxide and small amount of La. In the best catalyst, the concentration of Pt, Pd and Au is 2.3, 2.3 and 0.1{\%}, respectively. In order to obtain the catalytic activity versus composition relationship, artificial neural networks (ANNs) have been trained using catalytic results of the HRS optimization. Upon combining HRS and ANNs, {"}virtual{"} catalytic experiments were performed in order to (i) find {"}virtual{"} optimum compositions and (ii) map the full experimental space in two dimensions. Results obtained in this study proved that HRS is a very powerful tool both in catalyst library design and visualization of the experimental space. The combination of HRS with ANNs appeared to be an excellent method for knowledge extraction. In this way, further new information can be obtained about the catalytic system investigated.",
keywords = "Artificial neural networks, Catalyst library design, Combinatorial catalysis, High-throughput experimentation, Information mining, Methane oxidation, Multi-component catalysts",
author = "A. Tompos and Margitfalvi, {J{\'o}zsef L.} and E. Tfirst and Lajos V{\'e}gv{\'a}ri and Jaloull, {Mohyeddin A.} and Khalfalla, {Hamza A.} and Elgarni, {Mohammed M.}",
year = "2005",
month = "5",
day = "10",
doi = "10.1016/j.apcata.2005.02.019",
language = "English",
volume = "285",
pages = "65--78",
journal = "Applied Catalysis A: General",
issn = "0926-860X",
publisher = "Elsevier",
number = "1-2",

}

TY - JOUR

T1 - Development of catalyst libraries for total oxidation of methane

T2 - A case study for combined application of "holographic research strategy and artificial neural networks" in catalyst library design

AU - Tompos, A.

AU - Margitfalvi, József L.

AU - Tfirst, E.

AU - Végvári, Lajos

AU - Jaloull, Mohyeddin A.

AU - Khalfalla, Hamza A.

AU - Elgarni, Mohammed M.

PY - 2005/5/10

Y1 - 2005/5/10

N2 - Using tools of combinatorial catalysis and high-throughput experimentation techniques new multi-component catalysts have been designed and tested for total oxidation of methane. The compositions of catalysts were optimized by means of holographic research strategy (HRS). In the first catalyst generation, the best hit resulted in 44% conversion of methane. However, after designing and testing 167 compositions in five generations, the best catalysts resulted in practically complete conversion of methane at 350 °C. The supports of the best catalysts found by HRS consist of mostly Ce oxide and small amount of La. In the best catalyst, the concentration of Pt, Pd and Au is 2.3, 2.3 and 0.1%, respectively. In order to obtain the catalytic activity versus composition relationship, artificial neural networks (ANNs) have been trained using catalytic results of the HRS optimization. Upon combining HRS and ANNs, "virtual" catalytic experiments were performed in order to (i) find "virtual" optimum compositions and (ii) map the full experimental space in two dimensions. Results obtained in this study proved that HRS is a very powerful tool both in catalyst library design and visualization of the experimental space. The combination of HRS with ANNs appeared to be an excellent method for knowledge extraction. In this way, further new information can be obtained about the catalytic system investigated.

AB - Using tools of combinatorial catalysis and high-throughput experimentation techniques new multi-component catalysts have been designed and tested for total oxidation of methane. The compositions of catalysts were optimized by means of holographic research strategy (HRS). In the first catalyst generation, the best hit resulted in 44% conversion of methane. However, after designing and testing 167 compositions in five generations, the best catalysts resulted in practically complete conversion of methane at 350 °C. The supports of the best catalysts found by HRS consist of mostly Ce oxide and small amount of La. In the best catalyst, the concentration of Pt, Pd and Au is 2.3, 2.3 and 0.1%, respectively. In order to obtain the catalytic activity versus composition relationship, artificial neural networks (ANNs) have been trained using catalytic results of the HRS optimization. Upon combining HRS and ANNs, "virtual" catalytic experiments were performed in order to (i) find "virtual" optimum compositions and (ii) map the full experimental space in two dimensions. Results obtained in this study proved that HRS is a very powerful tool both in catalyst library design and visualization of the experimental space. The combination of HRS with ANNs appeared to be an excellent method for knowledge extraction. In this way, further new information can be obtained about the catalytic system investigated.

KW - Artificial neural networks

KW - Catalyst library design

KW - Combinatorial catalysis

KW - High-throughput experimentation

KW - Information mining

KW - Methane oxidation

KW - Multi-component catalysts

UR - http://www.scopus.com/inward/record.url?scp=17344364219&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=17344364219&partnerID=8YFLogxK

U2 - 10.1016/j.apcata.2005.02.019

DO - 10.1016/j.apcata.2005.02.019

M3 - Article

AN - SCOPUS:17344364219

VL - 285

SP - 65

EP - 78

JO - Applied Catalysis A: General

JF - Applied Catalysis A: General

SN - 0926-860X

IS - 1-2

ER -