Light Ends Popcorn
Popcorn polymer is a notorious troublemaker in the ethylene industry, and recent developments indicate the problem is far more widespread than previously understood. [1, 2] This post is the second in a series of articles about popcorn polymer. For more background information, see the first post in this series of articles about popcorn polymer.
The overhead condenser from a DeC4 unit fouled with popcorn polymer. Note that it is not white and fluffy like BD popcorn.
Areas of Concern
A butadiene unit is not the only place where popcorn polymer can grow. Around 2014, Nalco conducted a survey of ethylene plants around the world, and received data from 40 of them. They found that contrary to popular belief at the time, popcorn polymer can also grow in de-ethanizers (DeC2), depropanizers (DeC3) and debutanizers (DeC4). [1] Bulk BD concentrations in light end units are generally below 70 wt%, yet dozens of crackers around the world reported finding deposits of popcorn in one or more of these types of towers.
Graphs from a 2014 EPC paper by Nalco scientists [1]. Popcorn deposits were found in many towers around the world with < 70% bulk BD, contrary to historical beliefs, primarily located in the vapor space areas such as reboilers and the undersides of distillation trays
The maximum bulk BD concentrations listed in the survey responses are different for each type of tower:
DeC2 - BD concentrations <= 30%
DeC3 - BD concentrations <= 50%
DeC4 - BD concentrations <= 70%
These recent observations contradict the historical criteria defining where to implement popcorn mitigation programs. Before the survey results were published in 2014, most ethylene producers believed popcorn polymer could not be initiated when butadiene concentrations were below 70%. Lab seed initiation experiments did support this belief: lab experiments only observed seed formation down to 70-80 wt% BD. [2] How can popcorn be forming in light end units with BD well below 70%, but not in the lab?
Almost none of the popcorn observations in light ends were in areas where BD levels were above the historical lower concentration limit [1]
There is a long list of differences between conditions in any lab experiment versus a 60 meter (~200 ft) distillation tower, so it’s tempting to just wave our hands, chock it up to the chemist’s bad experimental design, and leave things be. Of course, that’s not what happened. After EEPC members discussed the survey findings during a popcorn prevention workshop in 2014, they decided to develop an updated set of guidelines for the global olefins community, ultimately issued in 2019. With those guidelines came a plausible hypothesis that can explain the apparent discrepancy.
Localized BD concentration effect
In light end units such as the depropanizer (DeC3), stagnant vapor space is more vulnerable to fouling than areas where bulk liquid is normally present. This is consistent with the fact that piping to safety relief valves commonly plugged with popcorn based on the survey findings. Piping to safety relief valves are considered stagnant zones/dead legs because only during an emergency does that line see any flow.
For a multicomponent system, the vapor composition is generally different than the liquid composition because different molecules have different boiling points. Butadiene has a lower boiling point than all butene isomers as well as butanes. That means that a liquid rich in C4s, such as depropanizer bottoms, will have a higher concentration of BD in the vapor.
In stagnant lines, particularly when insulation is poor, the walls of the pipe tend to cool further away from the hot bulk material. The lighter vapor starts to condense and flows down along the wall as a liquid while heating up the pipe itself. The rising and condensing vapor wets the inside of the pipe, and the condensed liquid trickling down the wall carries away some of the heavier components. The vapor in the line forms a new equilibrium, further concentrating the butadiene. This cycle repeats itself over time, forming localized areas of high BD concentrations. [2]
Left: An infrared image of a reboiler and pipe connections illustrates the temperature gradient along a line going to a safety valve that contains stagnant vapor. Right: Calculated equilibrium vapor compositions vs temperature for three different liquids at 4 bar illustrates how BD can become more concentrated in a stagnant vapor line with a temperature gradient. [2]
This staged mechanism explains the discrepancy with lab experiments, as well as historical experience based on BD units. While certain aspects about the popcorn formation mechanism are unclear, lab experiments did find specific conditions that promote rapid seed initiation. Contaminants such as molecular oxygen from air ingress, rust inside equipment and, curiously, water are known to be involved in peroxide formation and/or catalyze peroxide decomposition. Peroxides decompose when the O-O bond breaks, forming free radicals able to initiate polymerization chain reactions that could create new popcorn seeds. Seed initiation is a topic planned for the next article in this series.
Micrographs comparing representative popcorn polymer samples recovered from (A) a butadiene purification tower and (B) a debutanizer reboiler. [2]
Steam crackers have been operating around the world now for the better part of a century, including light end units and BD purification towers. Why did it take so many years to acknowledge popcorn polymer was more widespread than a few very high BD areas? One reason could be that popcorn samples from different areas look different: BD polymer is white and fluffy before it starts turning yellow as it oxidizes in air, but other areas can form sandy brown deposits of popcorn, or a mixture of brown popcorn and regular glassy polymer in the same chunk. The phrase “pseudo-popcorn” has been used by some in the past to describe light ends popcorn findings because it looks different but still bends metal when it grows, albeit much more slowly.
Popcorn seed growth rate increases nonlinearly with butadiene concentration, with a sharp increase above about 70-80% BD. [3]
Even under the microscope they can appear different at first glance, but popcorn from light end units do show signs of florets under the microscope. All macroscopic particles of popcorn show a unique “cauliflower” morphology on a microscopic level. That cauliflower appearance is a consequence of the “inside out” growth mechanism causing mechanical strain until it “pops,” breaking bonds to create new radicals along the fault line. Fresh monomer diffusing through the particle reacts with these new radicals to create new growth sites. Each half of the fault grows outward and imparts force against its sibling, creating new florets.
Dow developed what they call the Styrene Popcorn Polymer Test to positively identify whether or not a deposit sample is popcorn polymer and how active it is compared to a synthetic reference seed. [4] Light end deposits do in fact seed growth in uninhibited styrene, but samples that have no popcorn characteristics such as quench system foulant do not grow. Since the same type of mechanism is involved for BD popcorn growing in a depropanizer
A 10 mL Schlenk tube partially filled with polymer after a ~ 1 cm x 1 cm deposit sample was added to de-inhibited styrene under nitrogen and allowed to grow undisturbed at room temperature for a few days. Note the height of the solids grew past the original liquid level.
Until 2019, there had never been a loss of containment event in any light ends unit in the world. Unfortunately, that streak came to an end on July 31 when an explosion and fire destroyed a depropanizer after popcorn growth ruptured a 20” pipe. That event demonstrated beyond a shadow of a doubt the very real dangers posed by popcorn in light end units.
Depropanizer tower burning at ExxonMobil’s Baytown Olefins Plant July 31, 2019 after popcorn polymer growing in a stagnant zone caused loss of containment, vapor cloud detonation and a fire that lasted hours. Thankfully nobody was fatally injured.
Don’t Panic
Popcorn polymer isn’t going away any time soon, but there are several steps that facility operators can take to manage the risks to acceptable levels. Butadiene popcorn mitigation efforts have generally been effective at mitigating fouling, so manufacturers with BD units were already familiar with popcorn management programs.
Future blog posts in this series will discuss what we currently understand about seed initiation mechanisms, and mitigation options. If you would like to receive emails when new blog entries are published, there is a signup form near the bottom of the page. Feel free to email me for questions or comments.
References
D. Rossana, J.M. Hancock, “Popcorn polymer formation in light ends units,” AIChE Ethylene Producers Conference Proceedings, April 2014.
EEPC Popcorn Issue Group, "Recommendations for Preventing Popcorn in Steam Crackers and Butadiene Plants," European Ethylene Producer's Conference Proceedings, March 2019.
International Institute of Synthetic Rubber Producers, Inc., “Butadiene Popcorn Polymer Resource Book,” 2006.
S.J. Korf, S.S. Seifert, “The identification of butadiene popcorn polymer,” AIChE Ethylene Producers Conference Proceedings, April 2013.
