Coke deposition mechanisms of propane dehydrogenation on different sites of Al2O3 supported PtSn catalysts

Jianhao Jiao, Ye Yang, Maojie Yuan, Xuqi Tang, Mengfan Shi, Kai He, Haijuan Zhang, Yanfeng Bi, Yucai Qin, Lijuan SongChemical 2025 · DOI

TL;DR

Researchers have mapped out the "coke" formation pathways on PtSn catalysts used for propane dehydrogenation, revealing that not all catalyst sites are created equal. By using in situ infrared spectroscopy, the team discovered that while large platinum clusters are quickly "choked" by carbon deposits, ultra-small, low-coordination platinum sites remain highly resistant to deactivation, especially when supported by hydrogen "spillover" from the metal to the support.

What Problem Does This Paper Address?

Propane dehydrogenation (PDH) is a critical industrial process for producing propylene—the building block for everything from plastics to car parts. Platinum-tin (PtSn) catalysts supported on alumina ($Al_2O_3$) are the industry standard, but they suffer from a major Achilles' heel: coking.

During the reaction, side reactions produce heavy carbon deposits (coke) that physically block the active sites, killing the catalyst's productivity. While we know coke forms, we haven't fully understood the "molecular choreography" of how it starts on different parts of the catalyst (the metal, the promoter, or the support). This paper aims to pinpoint exactly where and how coke initiates to help engineers design more "coke-proof" catalysts.

Methods

The authors employed a suite of advanced characterization tools, with the star of the show being in situ Fourier Transform Infrared Spectroscopy (FTIR).

  • Pulse Experiments: They "fed" the catalyst pulses of propane at 600 °C and watched the chemical bonds evolve in real-time.
  • Model Catalysts: To isolate the effects of different components, they compared four systems: pure $Al_2O_3$, $Sn/Al_2O_3$, $Pt/Al_2O_3$, and the full $Pt/Sn-Al_2O_3$ catalyst.
  • Complementary Tech: They used STEM (Scanning Transmission Electron Microscopy) to see the size of metal clusters, XPS to check oxidation states, and Raman spectroscopy to analyze the "graphitic" nature of the resulting coke.

Datasets

As an experimental chemistry paper, this study does not use large-scale digital datasets in the machine-learning sense. Instead, it generates and analyzes:

  • Vibrational Spectra: A library of FTIR peaks corresponding to propyl, propylidene, and various aromatic coke precursors.
  • Thermogravimetric (TG) Data: Measuring the exact weight of coke formed on different catalyst configurations.
  • Catalytic Performance Curves: Tracking propane conversion and product selectivity over time (up to 100 hours for stability tests).

Key Insights and Ideas

  • The Propyl Starting Point: Regardless of the site, the "original sin" of coking is the excessive stripping of hydrogen from propane to form propyl species. These species then act as the seeds for larger carbon structures.
  • Site-Specific "Coke Recipes":
    • On Alumina ($Al_2O_3$): Propyl species cyclize into complex monocyclic and bicyclic aromatics.
  • On Tin ($SnO_x$): The process is milder, mostly forming monocyclic aromatics.
  • The "Good" vs. "Bad" Platinum: This is the most significant finding. Platinum sites with "saturated coordination" (larger clusters) are coke magnets. However, ultra-small, low-coordination Pt sites are remarkably resistant to coking. They catalyze the "clean" PDH reaction without letting the carbon stick.
  • Hydrogen as a Cleaning Agent: Platinum doesn't just make propylene; it generates "active hydrogen" that spills over onto the alumina support. This migrated hydrogen acts like a chemical broom, regulating and inhibiting the formation of heavy coke precursors on the support surface.

Key Results

  • Stability: The $Pt/Sn-Al_2O_3$ catalyst showed significantly better stability than $Pt/Al_2O_3$. While the latter saw a 25.3% drop in activity within just one hour, the tin-promoted version was much more resilient.
  • Coke Quantity: TG analysis showed that pure $Al_2O_3$ and $Sn/Al_2O_3$ actually accumulate very little coke (1.4–1.9 wt%) because they lack the strong dehydrogenation power to start the process. $Pt/Al_2O_3$ accumulated a massive 18.5 wt% coke, whereas adding tin reduced this to 5.0 wt%.
  • Hydrogen Co-feed: Adding $H_2$ to the propane feed inhibited coke formation and resulted in a higher Hydrogen-to-Carbon (H/C) ratio (3.96) in the precursors, meaning the coke that did form was less "deadly" and easier to manage.

Limitations and Future Work

The authors acknowledge that while they've mapped the initiation mechanisms, industrial conditions involve complex feedstocks and long-term thermal stresses not fully captured in short-term pulse experiments.

Future work could explore:

  1. Single-Atom Catalysis: Given that ultra-small Pt sites are the most resistant to coke, the next step is perfecting the synthesis of truly single-atom Pt catalysts that are thermally stable.
  2. Regeneration Cycles: Investigating how these specific coke species (monocyclic vs. bicyclic) respond to oxidative regeneration (burning off the coke) to optimize catalyst lifespan.
  3. Alternative Promoters: Applying this in situ FTIR methodology to other promoters beyond Tin to see if similar "hydrogen spillover" cleaning effects can be amplified.

Summary updated

Abstract

Propane dehydrogenation (PDH) Pt-based catalysts are facing the serious challenge of coke deactivation. The locations would greatly influence the coke formation, while the detailed mechanism is not fully explored. Herein, the coke mechanisms on different locations including Al2O3, Sn, Pt, and Pt-Sn sites were deeply investigated via in situ Fourier transform infrared spectroscopy (FTIR) technology, and the key factors triggering catalyst coke deactivation were proposed. Excessive dehydrogenation of propyl species is a crucial initial step in the formation of coke, whether at metal sites or supports. These propyl species on Al2O3 supports then cyclize to form monocyclic aromatic and bicyclic aromatic species, while those on SnOx sites cyclize to form monocyclic aromatic species. As for the Al2O3 supported PtSn catalysts, the strong dehydrogenation function and the interaction between Pt and Al2O3 supports trigger the complex coke formation mechanism. The surface Pt sites with saturated coordination are prone to coke deposition, leading to rapid deactivation at the initial stage of the reaction. However, the low-coordination Pt sites with ultra-small size are found to be highly resistant to coke formation in the PDH reaction, which selectively catalytic the PDH. Owing to the metal-support interaction, the extensive active hydrogen species generated from Pt can regulate the formation of coke precursors on the Al2O3 supports. Furthermore, the effect of hydrogen co-feed on coke deposition is also explored. The hydrogen co-feed inhibits the coke formation and results in a higher H/C ratio (3.96) for aromatic coke precursors. This study can enhance the understanding of the coke formation in PDH, which is important to designing efficient PDH Pt-based catalysts.
BibTeX
@inproceedings{jiao2025coke,
  author = {Jianhao Jiao and Ye Yang and Maojie Yuan and Xuqi Tang and Mengfan Shi and Kai He and Haijuan Zhang and Yanfeng Bi and Yucai Qin and Lijuan Song},
  title = {{Coke deposition mechanisms of propane dehydrogenation on different sites of Al<sub>2</sub>O<sub>3</sub> supported PtSn catalysts}},
  booktitle = {Chemical},
  year = {2025},
  doi = {10.20517/cs.2024.43},
}
Figure