Acceleration of End-of-Life Plastic Chemical Recycling by Catalysis by Pierre Moreau

so thank you very much uh so I'm T I'm research engineer for plastic energy um so I'm going to talk first of plastic energy Who We Are The Catalyst solution we develop how did we improve the performance of The Catalyst or did we model the reactor at full scale and we are going to propose a reaction mechanism for the first time ever and then conclude and then talk about our future work so first plastic energy Who We Are so plastic energy is a Global Recycling Technology Company headquarter in London so not too far from sample we have two commercial Comm recycling plant in Spain so one in Amia one in SEI that have been operating for the past seven years so team of technologies have around 16e combine uh combine experience in developing a unique patented T process so that process convert plastic that would otherwise go into land Fielding it into the ocean would be incinerated into tackle which is a fit stock that can be used to actually create virgin great plastic and we are the only global company that has successfully and consistently produced enough fit stock or Tac oil uh to enable the conversion back into plastic so I think everybody is aware of the plastic crisis but just to give you a some figures so there's 1.8 billion ton of greenhouse gas that are produced every year due to the plastic life cycle which is 3.4% of the total emission every year there is 3 3190 million ton of plastic that produced globally of these 390 million 22 million link into the environment or into the ocean every year and overall so there's only 9% so less than 10% of the plastic waste that is recycled 50% is landfill 20 22% is meat managed and 90% inated so however there's some up so there's an increased momentum to tackle the plastic waste there's environmental and societal pressure there's some trade restriction of plastic was and there's policy and Vol voluntary commitment from different major companies so why chemical recycling so first we can increase the recycling rate so diver what is going into Landfield or leaking into the ocean we can increase the Virgin quality of what is recycled so we're able to produce food we recycle content uh we have got also super sustainability goal so reduce the dependency on fossil fuel and help the Bland uh to achieve their um recycled Target so what does plastic energy bring to the market so until now you are basically s that one yes you had the plastic production plant and then either it was non recyclable so going to incineration or leaking into the environment or you got a mechanical recycling Loop so you go to Mechanical recycling plant and then go back to the plastic plant use byash again then you create a loot a loop and so on and so on however not all plastic can be mechanically recycling so a fair part of the plasting It Go bring then to the T plant where it's converted into an oil or t oil that then can be used to produce more oine that can be used to produce more poly and then you create what we call a chemical recycling Loop and that can go on like that so plastic energy who we are so a global leader in chemical recycling uh we got a patent technology we've been developing that process over the past decade we got but very strong operational exp experience with two operational plants that have been operating for the last seven years and we demonstrated we're able to make plastic to plastic process we are product commercialized on the European market and got Global partnership with major uh players worldwide so what is our mission so it's basically first reduce pollution so improve Waste Management by diverting plastic from lville or incineration creating a circular economy and creating a plastic Loop increase the recycling content of what has been produced including food grade plastic so higher quality reduce oil dependency of the country we work with so ruce the fossil oil consumption and boost local economy and creating job so what is our fit stock so our fit stock you can see here it is basically a real underf plastic it is dirty thing is not virgin plastic um and Y okay and it is basically mainly ldp hdp PP PS all under of Life what we can process still but in lower quantity is p PVC EPS category 7 what are we producing so producing an oil you can see here that we call TAC oil and that tackle then can be taken by petrochemical company fed into their steam cracker to produce some o oline that then are used to produce some more polymer some circular polymer so here is a very simplified schematic of the plant so we have basically the plasting coming then it is melted then it is going to reactor at high temperature here then the plastic uh quck and basically form a vapor that then go through a contactor that is a bit like a distillation column then through condensation where the gas and the liquid are separated so the liquid go to storage and floating and the gas is going to basically eat up the Rector so what is the Catalyst solution we developed so first just a reminder so when we use a catalyst we decrease the activation energy open another reaction pathway if the past during my career JY speaking when you don't use a catalyst you don't see any product so I've been working for a very long time fish tropes and that was the case here in the slide I'm going to show you if you don't use a catalyst you still have a conversion you still do the fal treatment so I just wanted to point that out first before going to the next slide then what type of catalyst would been using so we've been using Zite we've been especially using zm5 Zite that is basically an Amic Zite belonging to the pantas family of Zite it is characterized by the acid side silic ratio there are some internal Poe and external surface there is a porosity with micropore and mesopore and different acid number different also by strs it is largely used industrially um for different reaction like carbon carbon bone cracking isomerization delation it is interesting for us it produce at tone scale every year and so this material consequently a lot of potential uh for catalyzing under life plastic paralysis so did we start we started simply with five commercial Z that I'm sure a lot of the lab in the UK around which are basically the five zm5 you find on the website from zist International they differ by mainly by verment ratio from from 23 to 280 so we tested that a very small scale first on our microt reactor which is basically equip with a 300 M reactor the sm5 were cign first in air at 500° C see between 125 and 250 Micron and there was basically mix with plastic polymer mix at 0.5% weight and the polymer mix is basically a mix of ldp hdp PP PS PVC and this represent what we get in our fit stock the reactor was obviously Purge overnight with nitrogen to remove any presence of oxygen and then we do the test for two hours at high temperature and what do we observe so to our great surprise we saw that the most acidic zoites so the sm5 23 and 30 are actually uh the lowest activity even if we compare to the thermal doesn't show more activity so it's basically showing no Catalyst activity on the contrary the Z5 50 and 80 show the best activity almost fully converted the batch then the Z 5280 which is the less acidic one is basically umow showing lower activity so to our great surprise a silic ter ratio is 50 for activity and selectivity to T so at the time but I will come back to that later in the presentation uh we had the hypothesis a low secure Terio we got fast cooking that completely deactivate the Zite and high silica Termino we got less acid sight which basically um reduce the activity of the catalyst so I always come back to that slide to try to understand what is happening so for circle of catalis first you synthetize the material then you characterize it then you test it on your microp pilot then you model it on the you model it uh doing some kinetic modeling and then at the center you get what is called knowledge so did we start our journey to understanding the results so first I use a software that is called GMP software that is a statistical software that is very good that when you got different results trying to relate the results to the physical chemical property of the Zite so I've been putting the soft to a lot of different parameter acidity silic ratio anything I could find and but basically the two parameters that work the best were actually mesopor volume and activity and external surface era and that correlate quite well with activity and what does it tell us that basically the Zite with highest mesopor volume shows the highest activity this suggests consequently the peris reaction Happ by po move catalysis so it give us directly a pathway on how to improve the performance of the Zite by basically increasing the external surface area or the mesop volume by doing some desilication so how did we get on to that so I've been simply using um recipe from the literature doing 3 silication a different constraction of cium hydroxide 80° C 500 RPM for 30 minutes then I did free exchange with amonium nitrate one M solution calcination 550 for hour the5 is poti C and between 125 to 50 Micron and what You observe so for all treatment basically the mesopor um volume is significantly higher so between 1.7 to 2.1 at the same time the micro po room is reducing so this is expecting so you can have blocking by silicate then we've been trying to understand what was the impact on the reaction rate of the reaction so for this we did 3 TJ measurement at 5 10 and 20° C trying to determinate the free parameter of the kinetic equation that are the P exponential Factor activation energy order of reaction we use first we use superal method so first we use systematically the kinger method to determine the activ energy and the active exponential factor C meod to determine the order fraction it was always the first order when the Catalyst was used and then we did the final optimization to minimize the error using the Excel server and what do we observe when we compare the pound Zite uh with a three dislocated so as I said before it's all first order which is expected and then activation energy didn't change too much to be honest but the pron Factor change a lot so we got even um a ratio of um of almost 100 between deicated one Zite and the m550 so this show that we haven't changed fundamentally the mechanism of reaction but what we change is really the way the how that the reactant are actually diffusing much quicker to the active side of the Zite then we try to understand um to compare those results with the thermal case which is basically no Catalyst then we've been calculating the rate of reaction every time that is the initial RTI of fraction for concentration of ldp of one and then we've been normalizing the the results first to the thermal case which is basically no catalyst and as you can see whatever Zite use even the pal Zite is significantly more active than the than the thermal case then in the third column we normalize compared to the parite here we see also there's an improvement and what is interesting is that the Improvement is bigger as the intensity um as the intensity of the treatment is increasing so Al all of that is leading us to say that basically the reaction happen mainly by per Mo and kilo catalis so this was introduced mainly at the beginning of the year 2000 by Professor Yan mer with alization of anen and he been showing that basically the molecule mainly um um reacting at the entrance of the PO or on the external surface and then we didn't stop there so we try to model the process at full scale so at the scale of the planted Spain so did we get on that so we use another software that called Aspen plus diamic so we are to basically tweak the software so it was not especially designed for plants so there's been a lot of addon onto that on the normal software we have input different data so the kinetic equation enp of fraction polymer property polymer M properties product distribution Rector geometry three dimension and then you can get a lot of output data like the batch time the Rector polymer field LEL at Mass uh the flate of Rector Vapor actual Rector temperature and pressure and the [Music] conversion so we've been trying to compare three case so a case the thermal case with no sm5 so we've been normalizing all the results to that case the 1% weight sm5 loading and the 10% weight that loading so what to observe so as the quantity of Zite tested increase we have been we see that basically The Bash time is reducing which is very good for us and then we can increase the foodp of the plant at the same time as expected the maximum field level of polymer me in the reactor reduce th divided by two for the 10% case obviously the freight of Rector Vapor increase and this is um actually all right right see it's still within the maximum allowable FL rate through the contactor so we didn't stop there and we try to see okay what happen if we double the flow rate of hand of Life plastic coming into the reactor what happen in that case that's interesting so for normal feed rate we get an improvement of 2.4 which is already very big and we are very very pleased with but then if we double the field rate it's increased to 3.3 which is even more interesting so this significant they show for the pl that with Thal par the reaction rate is liit is really the limiting factor well with the catalytic paralysis then the feeding rate becomes the limiting factor so from now on uh we try in parallel also um in parallel or all the scale of work etc try to understand what is the reaction mechanism what is happening exactly so I have to say that the first time I show this results okay so first fraction first first slide on that reaction startup there's been perhaps a lot of discussion on how the reaction is starting up if it's starting Cally or if it's starting in a formal way uh you will see we have an opinion on that um so here I show just an example I was able to show so basically we did two TJ so one virgin ldp 20 mgram another one with Virgin ldp and 10% 10% um 10% the M5 so 20 milligram ldp 2 millgram m550 as you can see so here we show the percentage m versus the temperature the um basically the the reaction initiating a much lower temperature for the cataly case compared to the thermal case I have to say this is true at any scale we've tested so far so for example on our micropilot we follow the gas and a tackle production and I can tell you that there is roughly 50 degrees C difference between the thermal case and the catalytic case so it is really substantial and for me Beyond do um also another important parameter is that that whatever also at any scale we do the reaction gas is always produced first you will see that has an importance for the rest and that gas so we did some online GC analysis is very oine W especially a rich in ethylene and propylene which is obviously interesting for us but those molecule are highly reactive even just by them cell and even more reactive if you add a Zite so actually so what is happening so we've been um we've been with our partner doing some experiment where we fed actually just aine so no P but just aine to the same five the same five from the restal uh so we use 300 milligram every time see to between 200 and 400 Micron we use a fixed bed Rector 7° C at room pressure and we purge re of with nitrogen and then and start theine nitrogen e to ratio volume to volume 20 so this is called by the way the alpha test um and what do we see so the most acidic Zite which is in R here is actually dropping very the activity is dropping very fast so there is a very quick deactivation then the second least um acidic is actually dropping fast little bit slower than the 521 then the at21 show lower initial activity than the activation and then the least acidic is show almost no activity all those Zite should C deposit except the 28021 and once again the sm550 seems to be the best compromise between activity initial activity and deactivation so this lead us to things that the Moline basically react initially when we do the reaction with polymer and that the basically condition the Zite for the rest of its life then we've been looking at what is happening post reaction so we've been doing a lot of characterization post reaction so we've been doing the reaction for 16 hour for the same F the same five right so did 16 to be sure it was fully converted we were able to unload the Catalyst been using two instruments a microtic reflex instrument to measure poity and then a chns Elemental Anis from element so what do we see so whatever the use that is always POS loss so that is in Gray here so percentage loss versus a differen light in yellow is the microporosity so you see the micros loss is basically increasing as you increase the acidity of the Zite to the point that the most acidic has completely lost this micros on the contrary the biggest MOS loss is actually the 280 to18 and also so it's not displayer but we did the percentage carbon at the end measurement and basically the percentage carbon is increasing as the zeolite is becoming more and more acidic which lead us to say that the code deposit is the main reason for Activation so yeah I'm not showing the re basically it's underly plastic but it's still true with underly plastic and this is despite there is a lot of poison within that within that end of life plastic so what was also very interesting we did that same experiment for only two hours since we wanted to compare primarily the different activity of the Zite but actually two of the Zite has fully converted the batch in two hours so they were very active that wasm 550 and 80 so we've been doing the same measurement and what we observe is that for the sm50 the POS is almost completely lost the m micros is completely lost mesopor is almost fully lost but that Zite is still active we're still able to reuse it several time um but then we obser we compare the result with the 16 hours what we observe that a large part of the mesoporosity and the microporosity is recovered which shows that the Zite is still active and can actually when there's no more ldp can actually crack the the coke then sm550 sorry 80 is not too much affected by power blockage and that doesn't change too much after 16 hours so also an important parameter so we did with our partner from the Catalyst up some cared spectroscopy so we've been collected different and Spectra at different temperature uh with a mixture of ldp Zite one to three and then there was um a m spectrometer on the outlet and uh we are able to do Mass spectrometry uh so basically to determine the content of the product so what we saw is that basically we have a build up of carbon double bond carbon carbon containing H carbon the light surface as the plastic start to crack so we see that basically with the blue line the ldp vibration weakens and then we got a build up of aromatic monocyclic aromatics on the surface so so we can see actually those aromatic being produced with the mass spectrometry that important show us that basically we are getting the same product that we get in our micropilot and pilot so mainly laen and alen a alen and aromatic products and finally we've been looking so what is the composition of the coke post reaction when we recover the C so for this uh we've been using a method developed by my PhD supervisor menu at P University where we dissolve the coolite in a concentrated fuic acid then we do an extraction with deor methan and then we inject that for the soluble part onto the gcms and what we finding so different molecule of naline substituted monod D3 naline fluorine fenine anine deyl Pine and you can see you can grow the cycle of aromatic you can go to very very large molecule here the only limit is how much space there is around the molecule and uh basically the supply of oine um of ethine and propylene finally looking at aromatic so you're still on our micropilot at 400 been doing reusing the same Zite over and over and over again and something that came quite striking was that the yield of aromatic ver is run number so as you can see it's produced quite high level initially but then it is decreasing to the point that after six reuse we don't see any more aromatics and also we were managed to fit an equation to with a very good correlation factor of 99.52% so that shows that the aromatics are produced certainly initially within the poe we know the poe are getting blow microporosity is getting blocked up to the point that no aromatic are produced so let's get on with the reaction mechanism so I divided that into different stage um and here we've got a framework so the sm5 framework that is Virgin so is nothing within initially and there is an ldp molecule here I just remind you that ldp is not forc linear so it can be Branch so what happen initially so from our different experimental evidence we think that the ldp is reacting on the external surface initially or in the PO move as was shown here by y merant and start to form some gas that gas is very rich in ethylene and propylene so this is the step one startup then we got into step two that's called conditioning where basically ldp is still there so you still have some ldp in the reactor but you have formed some gas you start to form some tackle after around 20 minutes and that gas very rich in ethine and propylene now that ethine and propylene can contrary to the ldp can diffuse within the power of the Zite and that ethine and propine is going to do some reaction within the power of the Zite and what is going to do dimerization alization Etc polymerization Etc so you start to form some gas some Tac oil including some aromatic and there a last reaction you form is basically you're producing Coke so you're producing some polyaromatic molecule that start to form either very close to the external surface or really deep within the the same side then we go to state three which I call steady state so stage the coke is growing the C is always growing so long f f stock and space to grow we still have got some ldp that basically um is converted on the external surface now it cannot go inside see basically the micros is fully blocked by the Coke molecule so this space here is becoming completely unavailable to the reaction and then it does react and then it does still form some gas and some tacle and some o that fraction here is not form anymore so we don't see anymore the aromatic um and that goes on and until you run out of ldp then when you R of ldp so what is happening so call that cooking so you you just have your zeolite and some coke molecule the zeolite is still active if you put more ldp that will convert but given that there's no ldp and there just Coke molecule it start to convert the coke molecule into gas so we don't see any T it's only gas at that stage and that gas is still very rich in ethylene and propylene then what you can do you stop your reactor you can actually do an oxidative regeneration at high temperature of your Zite so what we observe when we do otive regeneration so of course before regeneration the light is black after it looks of course very white it's not as white as the fragilite I would say it's more cream color and when we do chns and you see a bit of carbon so we think that carbon is really deep inside the molecule since when we do the reaction again honestly from a reaction R point of view we don't see the difference at all then if you put more ldp but you do the same you react on the external surface or in the porou then you form some more gas some more tackle some more oline that can diffuse inside you form some more coke that can produce some more gas some more D some more Ara and then you go through the cycle and you can do several cycle of reaction regeneration reaction regeneration up to the point that Z Light is basically not act not active enough to be interesting so to conclude on that slide you have a minority of C that is refractory to Regeneration but it is not a problem for us because we get the same activities ini then I try to think but so is if there's in the literature um any other example of such reaction with medor poite which zsm5 is with hydrocarbon so there's two example um but there might be more so there is your stud Station witha 1 Zite where it's been showing that basically there was um a mechanism of fraction by um absor from the external surface of the crystalite and then go the reaction is the PO move on the external surface and then the second example is basically my PhD that I did between 2002 and 2005 with ipn so I've been studing aine mization the idea was to form xylene to have form talic acid to form P PT bottle uh so they wanted to maximize the xylene from the CTIC cut in the refineries and we found um a Zite that was called eu1 Zite that has basically um some interesting properties so what we're observing it was deactivating very very quickly initially to show up yeah so the open open symbol of the initial quick deactivation and then stabilization so initially you've got both inside the micropor on the external surface um and you got cooking to the point you fully block the microporosity and then you got only reaction on the external surface so you form more you increase your selectivity in isomerization which is actually the product you want so that was a very interesting project so that is certainly not limited to that uh but that was to main example I could think of the conclusion the future work so so man to show that the 25210 is the most interesting one this silication for us uh is an efficient and econic method to improve significantly the um the rate of reaction the food put for the plant um we think that the reaction happen mainly by power mov and kilo catalyses the Rector model has confirmed the results we're getting from the kinetic rate so we can improve the food pool by 2.4 for f rate 3.3 for double field weate and this is a very good news that mean we can process more and of Life plastic potentially show that the rtion limited by reaction rate for thermal pares and feeding rate for catalytic paralyses the final point so yeah there's there's a lot of optimization of the plant that will needs to be done to accommodate the Catalyst obviously and I have to say also so there is there is of course what was the purpose of the project which was to increase the food putut of the plant I have to say there's a lot of detail advantage of using C compared to doing formal power I'm not going to get into the details it's not only the increasing the frood putut for future work so the zm5 formulation has been actually scale up at our third party this year to 20 kilogram that was successful characterization testing of the micropilot came back positive um on we have a slot that will be tested on our 30 L pilot plant Reactor with and of Life plastic um and then if it's positive we manufacture that to turnone size of our third party and then do a demonstration our plant in Spain uh for the theic 58 so all the people who have been helping me to put together the presentation and generate the result so thank you Emil uh Daria ABI George Alex Andrew David Anton Loro and London P Unity Isabel Ander for catalist characterization and C analysis loveb University Sam sndy and Simon for material synthesis and several advice catalis edma and and Bill for Catal synthesis modification C cation well thank you and I'm open to any question thank you very much indeed here questions yes yeah thanks s so much and great to see the progress as I heard you speak earlier this year um many of your tests were done on a mixture of three oants so ethylin propin spiring what is your commercial Target and how sensitive is the is the process to um different amounts of those so I mean there's been a lot of test done I mean a lot yeah I mean I say a decade it's really a decade I mean they started we started part the test on the pilot plant in Ireland in 2008 so show you the scale of the work that has been done um so basically that's what you found in your beIN basically uh as I say we try can I yeah can go back actually so so first ldp is by far the main polymer that is produced worldwide and then you find PP and PS so we've we I can't disclose too too much about that because there has been a lot of Ip link to that so if a customer come to us we've got a guideline of like the the fix stock needs to have no more than this polymer no less than that basically and it's based on test that we did doing different mix and seeing what happened can tell you there's been lot of RoR blockage if you use certain polymer there been a lot of learning made and that's what what was interesting to have also this plant in Spain is to be able to test different configuration so but basically our as I say can go back quickly so basically ldp P ldp hdp PP PS is what you find mainly so the one we don't really convert is p the one who is on this one is special so it has oxygen Etc if you do the Thal case so this one we can process it but you shouldn't access there's a maximum level we can tolerate in our plant than interesting um and it's really impressive that you get a 3.3 times increasing throughput that's Catalyst for you um the other thing the catalysts do is is they can um change the selectivity do you see changes in so yeah yeah there's there's um so there's a lot lot of talk about the literature and what is happening when you use Catalyst versus so I would say the selectivity selectivity depend on a lot of parameter and so it will depend which Catalyst you use first so the good thing is that actually the deselected catalyst is more selective to liquid um then that depend of the operating condition be careful for the thermal Catal the optimum temperature is like 400 currently the Catal so active so that in order not to be at 100% with micropal now are 340 so that show you 60° just huge uh but to come back to your question be careful the gas is a product the gas is actually very interesting see it's rich in ethine propine com composition doesn't change too much when you do the reaction on the reaction started once you got that initial Flash in in gas it's very consistent what you produce and it is very rich in ethine propine just what you want to make more more poly so so there will be two type of plants so there will be the type of plant like s just by themselves so those one try to maximize the liquid because the gas the gas well you need to compress it after if you want to to send it somewhere else liquid is much easier to you just put into a tun that's fine but we got a plant like got so we got a plant in Netherland in Gin that will be directly linked to a steam cracker and we are going to produce gas so I mean the obviously thing to do is to send that to the steam cracker directly so it can be I mean the gas can be actually interesting if you're close to a petrochemical site since then basically what what we get as gas is very close to what they get from what is the output of the steam cracking and be careful don't remember you need to fill the old chain so there's a steam cracking but then after there you don't outside the steam cracking you don't get a single flow of ethylene or propylene it's all so then we obliged to do a CR cogenic distillation to separate the different um oin so that is actually very interesting so that is also a possibility I saw the question around when we use the right if we could maximize the gas which is actually possible but that's that's two way two different way to drive the process so I guess your process economics you will be required or you are required to regenerate zit so therefore I guess the optimum the optimum operation may not be to run the Zid all the way through to loss of activity You may wish to regenerate it earlier than that I guess you have thought about that and I wonder if you thought about trying to recover any energy from the coke that you're forming um good question uh so let's start sorry let's start so yeah you you you you you think you pay for the Zite so I'm always in comparison with the thermal case so we're doing technoeconomic to show that actually is the problem you pay for the Zer light here you're going to increase your Opex you need to show that is a significant increase compared to the Thal case so that it is um so that it is Prof so you still make a profit while paying the supplier the supplier needs to make a profit also so the old chain needs to make a profit uh what we observe is that so we managed to make the Zer very active I have to say so now we're faking gating very low in um in the we won't go to 10% I can tell you there been so much progress these results are even a bit old so we made some further progress since so you're not oblig to use that much to the point that would be the devil advocate you could say you can get away without any however however a company we say all about circularity Etc so we'll do regeneration we to use a third party I'm in contact with third party to the generation so do we use up to the point so the point we won't go to the point where the Z is completely inactive so you don't want to get to the point to the theral case if the Zite is completely inactive that is a thermal case so that is not interesting anymore so you better S out for Generation surveys actually the well we started last year the economic model and it is very positive I can't get into that because it's very secretive but it is it is very positive it is yeah I mean it's we get the price get a price one from the supplier at the Autumn and we're in a very good position okay I think we've run out of time thank you very much indeed here thank you