Continue to Site

Eng-Tips is the largest engineering community on the Internet

Intelligent Work Forums for Engineering Professionals

  • Congratulations KootK on being selected by the Eng-Tips community for having the most helpful posts in the forums last week. Way to Go!

Secondary Decomposition Rxns. 1

Status
Not open for further replies.

JDI_eng

Chemical
Dec 13, 2023
3
Hello,

I am investigating decomposition relief scenarios and need to learn more about low-phi factor testing (i.e. VSP2) and high-phi factor testing (i.e. ARC). I understand that ARC tests can establish kinetics and develop kinetic models for decomposition relief. However, I recently learned that low-phi tests can induce different, secondary decomposition reactions compared to high-phi tests. Conceptually, this is sensible due to the difference in heat transfer conditions. In my case, the chemicals of interest do not have intuitive decomposition pathways (i.e. not like H2O2 decomp.)

Can anyone provide additional information/resources/personal experiences on this topic? Has anyone encountered situations where their ARC data drastically differs from their VSP2 data - perhaps showing different decompositions? Are these secondary decompositions unique to complex decompositions? If VSP2 is more representative of scale-up conditions due to its adiabatic nature, why use ARC testing in the first place? Perhaps ARC is better suited for less complex decompositions.

In the context of emergency relief systems (ERS) sizing, it is crucial to understand the kinetics of these scenarios and properly size the ERS. Moreover, from a legal perspective, it is required by OSHA PSM (see 1910.119(d)(1)(iv)) that "Process Safety Information (PSI) shall consist of reactivity data." Recent CSB findings for the Optima Belle decomposition catastrophe and reactive relief experts (i.e. ioMosaic) further reinforce this requirement. Therefore, I feel obligated to ensure that the right test is chosen to accurately depict these scenarios.

Thank you very much,
JDI
 
Replies continue below

Recommended for you

What chemicals, reaction press / temp region and what is the primary decomposition reaction ?
 
Reaction Pressure - 800-900 psig
Reaction Temps - 300F; decomp onset temps range from 350-425F
Chemicals - Amines and Nitriles

Various amines and nitriles exist in the portfolio, so there are various decompositions that have been theorized by previous ARC studies and gas-capture/gas chromatography analysis. For one example, a nitrile was theorized to decompose into hydrogen cyanide (HCN), hydrogen, oligomers of acrylonitrile (ACN). The previous kinetic Modelling Engineer proposed two pathways: 1st being a single reaction to oligomers and HCN ... and the 2nd being a trio of decompositions to ACN oligomers and hydrogen.
 
I dont know if I can help you here. I've read a bit about this from articles on the net. Decomposition of acetonitrile for example, seems to be initiated at much higher temperatures than 300-400degF you have here, its more like 900degC. So do you have these reactions at these low temperatures?

 
Run your VSP2 testing. I ran a VSP2 on a molecule similar to acetonitrile for an inadvertent mixing scenario. There appeared to be no reaction initially, but when the sample was heated slightly, still well under 200 C, a decomposition reaction appeared to occur. There was no initial reaction to boost temperature up into normal organic molecule decomp temp ranges of 300-800C.

As you can see from the values below, these decomp reactions essentially constitute an explosion. Do not play around. If you think there is a possibility of the scenario occurring - TEST!

VSP2_Excerpt_sykyp1.png


Fortunately for us, there is no heat source other than pool fire for heating of the mixed material like this, so the scenario of inadvertant mixing plus heating isn't valid. Still, this does mean we need to look further into the mechanics of this reaction because decomps don't normally occur at so low of a temperature.
 
JDI_eng said:
Can anyone provide additional personal experiences on this topic?
I had been working on a facility with reactors of the simple raction acrylonitrile (mentioned above) to acrylamide at atm pressure and 100±10°C. 2m3 SS continiously operating stirred ajitators: 2 in series, 2 in parralel, 4 in total. Nothing special.

We have faced with multiple explosion cases caused by a runaway reaction. When catalyst conditions slightly change heat removal turned out insufficient and reactors became overheated over seconds. Liquid swelled, rupture disc DN100 was inadequate to relief foam, ~50 studs M60 between shell and head stretched, jet beat out windows all around, studs ruptured one by one, vessel tore in half, head+gear+agitator+motor rocketed up, pierced reinforced concrete ceiling, and lied on the next floor. And such events were not sporadic.

Inspite of the design described has been provided quite long time ago by people I am not familiar with believe me - some R&D did was provided and designers involved did considered the relief measures are adequate. One involved in all stages of a process life (not in concept stage only) is able to find out that in reality lab conditions differ from industry one. This is what I have had to learn from my personal experience. Oppau 1921 case is one of the best expamples of a such "lab vs industry scale" difference.

Any reaction and its kinetic are some sort of oversimplifying intended to explain/report a measurement/study conducted. Even H2O2 decomposition at industry scale is as multiple factors are involved simultaneously. Similar situation as with the sound pressure - decibel vs octave pressure. Having been over both sides of this wall (design and operation) I became sceptic over such cases as the reaction modelling. Fortunatelly most scientists prefer to omit that a lab measurement is only one simple and cheap stage of the proper modelling of an industry process. One is able to find in academic literature a lot of versatile discussions how predictions/measurements/modelling should be provided and its pros and cons and nothing about how results are consisten with the real faility.

For info
Some details can be found in para. 3.4 HSG 136
HSG 143
 
'HCN or not HCN'
Note that HCN is included in the list of designated chemical warfare agents (instead of most amines and nitriles).
Is such info (data related to HCN risk) critical for process development and safety? I in person believe that substances having extremely low acute toxicity threshold are worth a comprehensive study. Seveso 1976 is an example of the attitude to process design when only primary reaction is taken into account and side reactions are ignored.
 
Status
Not open for further replies.

Part and Inventory Search

Sponsor