105 - How much smoke is made in fires and how we measure that? with David Purser

With David , we dig quite a deep jump into toxicity of fires and why fires are toxic , and I've promised you there's more coming . So here we go . Here is more . This time I've asked some questions about how do we know what's produced in fires , how do we measure that and what actually is driving this production .

I've asked about the material properties and the ventilation , the burning conditions in which the burning occurs , and we ended up with quite a deep and interesting conversation that you're just about to hear , as in the previous episodes with David , by the way , i would highly recommend listening to those first to give you a basic idea about toxicology in fires .

But if you've missed those episodes , you're very welcome to listen to this one And then then move back if you enjoy this . As with all the episodes on this particular , very difficult subject I understand this are not not easy . There's a lot of chemistry . Not many people love chemistry , we tend to remove it from fire safety engineering as often as we can .

I guess there's not that many chemistry enthusiasts around in here , but , yeah , this is fundamental to our understanding of why fires are dangerous . When they are dangerous , how do we act on that and what's generally happening in in compartment fires . So , i believe , very important lessons in here , lessons that are difficult to find elsewhere .

Not that many great resources on this subject that you can easily learn from , so I hope you will still enjoy it . I've enjoyed it a lot . I don't know what that means about me , but I've certainly loved talking to David . Anyway , let's not prolong this anymore . Let's spin the intro and jump into the episode . Welcome to the fireside show .

My name is Vojci Wigciński and I will be your host , as usual . I would like to say thanks to the sponsor of this podcast , orfar Consultants . So this podcast is brought to you in collaboration with Orfar Consultants , a multi-award winning independent consultancy dedicated to addressing fire safety challenges . Orfar is the UK's leading fire risk consultancy .

Its globally established team has developed a reputation for preeminent fire engineering expertise , with colleagues working across the world to help protect people , property and planet .

In the UK that includes the redevelopment of the Printworks building in Canada Water , one of the tallest residential buildings in Birmingham , as well as historical structures like the National Gallery , national History Museum and National Portrait Gallery in London .

Internationally , the work ranges from Antarctic to the Atacama Desert in Chile , to a number of projects in Africa 2023, . Orfar is growing its team and it's keen to hear from industry professionals who want to collaborate on the fire safety futures this year . Get in touch at OrfarConsultantscom . Hello everybody , welcome to Fireside Show .

I'm today again with Professor David Purser . Hello David , hello , great to have you back on the show very quickly . I appreciate that a lot , but you've left me hanging . You left me on a cliffhanger the last interview , so we really have to finish this Today .

I would love to understand how smoke and toxic products in smoke are produced in fire And just let me give you a brief introduction of how I model that in my CFD And that's going to take me to sentences . I take heat release rate that my fire emits . I know the mass , the heat of combustion of that , so I know the mass loss rate of my fuel .

And then I apply the soot yield or a CO yield to that And I know how much I've emitted environment . So how good approximation that can be , given the complexity of fire physics .

The CFD combustion model does give you some information in the sense that it tells you what the major species are . So it tells you how much unburned fuel there is in that cell , how much carbon dioxide there is and how much oxygen you consumed on the temperature . So you've got a sort of very basic combustion model in each cell .

But I think , beyond that , when it comes to what CFD people call minor species , like which is all the ones I'm interested in , like carbon monoxide and polyametic high carbons and all these things , toxic species the model's still got some way to go And I know they are still working on this .

So I sort of approach that I would suggest is that you then , having got the basic conditions in the combustion conditions with the CFD analysis and the mass loss fuel concentration or effluent concentration in the analysis , you can then go and look up the empirical yields that people have measured in various kinds of test apparatus .

For if you know what the fuel is , you have to know what the fuel is . The fuel mix is that you're burning . Of course , what's been measured for that fuel under defined conditions and I'm going to come back to the defined conditions in a minute .

I think that I'm not jumping ahead a bit really but what we've tried to achieve in the world , we were doing because we realized that the yields of all these toxic products vary so much was different materials under different combustion conditions and this sort of thing When we wanted to find a test method where we could decompose off combustor material under defined

conditions and with fixing three or four major parameters that determine the products that will form , essentially the temperature , whether or not the material is flaming with a flame or not , or smoldering , and particularly the fuel-air-equivalence ratio , and then we could plot the yields of all the toxic products against those parameters defined . With those parameters .

If you then took those parameters and looked for them in full scale fire test or in other types of apparatus that were trying to achieve the same thing , did you get the same answer . In other words , when we measured something in a small scale test under these defined conditions ?

was it predictive of what you get in all other kinds of apparatus or burning things under those same conditions ? And ultimately I think and there's a lot more work needs being done on this , of course , but ultimately I think we're getting somewhere . You know , i think we sort of more or less have achieved , So I think that's a promising approach to take .

So at the moment , i know there are complex combustion models of some people use in conjunction with CFT , but I think at the moment a good way is to use the sort of expressions that we've been developing . Well , i've been fitting viable curves , you know , to my yield data which are defined in terms of equivalence ratio and things like that .

So you could take those expressions , algebraic expressions , apply them to the computed conditions in your CFT cells .

So so my colleagues Jeff Cox , stuart Miles and Siris Kumal did a CFD analysis to predict the conditions in the Mont Blanc tunnel during the actual incident And I was trying to do the toxicity analysis , the FED analysis , for people who might be moving in the tunnel , and one way we did this was to use the CFD .

So at that time what we could do was really quite limited in terms of predicting minor species . So what we decided was looking at the scenarios they were getting .

With the CFD analysis we felt that the conditions downstream of the main of the first burning HGV , of the big vehicle that was burning near the ceiling there , would have been initially not fully well ventilated because it's a confined dish sort of space in a very large fuel load , but it would be somewhat vitiated , somewhat under ventilated .

And we looked at the composition of the burning vehicle and its load , which was margarine in fact , and based on the sort of data that we had from the chip furnace and other sources on ULZ , under certain equivalence ratio type conditions , we came up with an estimate of the sort of CO , cyanide , etc . In the last few years .

We expected to get under those conditions and the CFD to work out the dilution factors and things , and so came up with time concentration curves in the tunnel , along the tunnel that we felt that the occupants would be lied to be exposed , and then from that I was able to do some estimates of how far people would be able to walk through the tunnel before they

would collapse , based on the CFD analysis and FED analysis . Now , in this particular incident we had another source of information , which was that our French colleagues had worked out . They looked at the time each vehicle that got stuck in the tunnel entered the tunnel and how far they travelled .

So they were able to calculate the time at which those people stopped in the tunnel And we know how far they walked in the tunnel before they collapsed . So we roughly knew at what time they collapsed And we could compare that with the predicted times based upon the CFD FED analysis , if you're with me .

So you've got two ways of coming at the same thing And we found that there was a reasonably good agreement between the two , in other words , the data from the time of the measured , if you like , time of arrival and collapse of the people in the tunnel agreed with what we would predict if the CFD modelling was correct , and so this is a sort of approach I'm

using quite a lot now to say that if we holistically take all these sources of information for an incident , can we use one to help validate another and arrive at a consensus understanding what's going on , and so that's one example where we have tried to apply yield data to us as CFD analysis . But it's that aspect of it was really quite simple .

I think we could do a lot better , but I know nobody's really run with this and tried to do it . It would be the greatest somebody would .

You understood that the smoke layer is under ventilated , so you have applied more complex models to understand the yields in that under ventilated layer to and , based on your empirical research in comparable conditions , you got the yields of minor toxic pollutants like HCN and other more dangerous things .

And then , knowing distance , the travel distance , you've calculated the doses that people took .

That's how we could do engineering , actually , because we do have some sort of a vacation models , though we don't really simulate behavior that much , and I guess that's going to be a major limiting factor in estimating dose because you just assume person walks away , not necessarily recalls the fire with their phone while inhaling smoke . Our models don't do that yet .

But yeah , i see that being a path for the future And I really like how you could , based on limited data from imperfect CFD like major toxic yields , figure out reliable numbers for the minor toxic pollutants . That's very interesting How much that is related to

material . We've touched a little bit that in the previous episode but maybe we can talk .

So what inhalation toxicologists I was an inhalation toxicologist at the time what they like to do or when they're and he's talking really about chemistry but also animal experiments here is that particularly for animal experiments , what you try to do is generate an atmosphere of some material .

We would do this for all sorts of things , not just for files , mainly for dust and gases and novel chemicals , industrial chemicals and things like this .

What you try to do as well as you can is to generate an atmosphere in a chamber of some sort that will remain of constant composition for a defined time And typically combustion toxicity tests would be 30 minutes . Standard acute toxicity tests would tend to standard international standards for a four hour period .

If somebody sends you a newly invented chemical that might be a dust that might be generated in the atmosphere , you would normally take groups of rats and expose them for four hours to different concentrations of that in the atmosphere . But you always try and make sure your atmosphere was steady state , called it steady state atmosphere .

So you put in the thing you're testing , you put it into the chamber at a constant rate and you dilute it with a constant stream of diluent air to get a defined concentration for the whole period of time .

So when people started to first work on this with combustion products , i sort of get the impression that somebody came to them and said look , we've got to do combustion , how are we going to do it ?

So I said , well , we need to find some way of generating this smoke atmosphere And I'll go and have a look in the cupboard of the old lab at the end of the building and see what we've got .

And they came up with all sorts of different bits of heating equipment and say , well , yeah , well , we can use this , we can stick our material here and they'll get hot enough . Or put it in the oven . It'll start to decompose and we can see what it does . Or chemically , with GCMS , we can see what chemicals are given off .

And so it was really done without too much thought about what actual conditions are we generating ? So when I first started working on this for the fire search station , i was asked to look at three materials wood , so a pine of Sylvesterus , just ordinary pine wood .

Polypropylene Polypropylene is a good place to start with our ideas because it's a polypropylene polyethylene oil . It's about as simple as you can get in a solid material . It's just carbon and hydrogen in a long chain . And the third material was one flexible polyurethane foam , because that was a big issue at the time .

It still is , and I was asked to test these under pyrolytic conditions , under nitrogen . Well , now I have to talk about what we mean by pyrolysis right Now . To me , pyrolysis is when you heat your material under a stream of nitrogen that you don't give it any oxygen .

There's another term called non-flaming oxidative decomposition , or some people call it oxidative pyrolysis , which is where you heat up a specimen , a material , in air , but it's not burning with a flame , so it's non-flaming , but you heat it up to a temperature where it starts to pyrolyze and then oxidize .

What do we mean by smoldering , which we could also talk about , although I'm worried about over the years What simple oxidative decomposition ? there'd be no glow , you'd just be heating it up . You got to a temperature where you've got pyrolysis , and pyrolysis just means the molecular structure of the material is breaking down due to heat .

That's where things start off . On the CFD people are very keen on pyrolysis rates And the origin . The first step in pyrolysis is that heat is applied and it starts to break down the material . But once it starts to break down , if there's any air around , you can get some degree of oxidation of those products .

No flames yet , but you're starting to get oxidation And these distinctions are really quite important . So the time I started doing my work there were really two or three major sorts of apparatus around the people were using for this work .

So the National Bureau of Standards had the NBS Cup furnace method And basically what they had was a 200-litre sealed box Perspex box with a one-end underneath it . They had a crucible , you know where , the volume of about 200-300 ml , and they actually did start to think a little bit about the combustion conditions .

So the tests that they did they put the samples of material into this , dropped them into this crucible at a temperature I'll forget the exact amount , but it was a few 10 or 20 degrees or something below the water-ognition temperature . So they're heating their material in air , but at a temperature where it will pyrolyze , may oxidize , but should not ignite .

And then they would expose rats to that for half an hour And in the standard test you then keep the rats for 14 days afterwards the ones that survive the immediate exposure to see if any further toxic effects occur . So that's the standard toxicity approach .

And then they would repeat the test at I'll forget the figure now , but it would say 20 degrees , 10 or 20 degrees above the auto-ognition temperature . So in this case the specimen would ignite and burn with a flame and you would expose the animals to that . So you're already starting to get some kind of distinction between these two states of non-flaming and flaming

. Another apparatus that was used quite widely and got some very interesting data from was the University of Pittsburgh . It's mostly in the states this is all going on . They used a sort of much larger sort of oven type situation where they took a specimen of a material and they heated it with a ramp temperature profile .

So it started at room temperature , then they increased the temperature with a ramp linear ramp and it got hotter and hotter and hotter . At some point it would start to pyrolyze and then at a bit higher temperature it might ignite . So they didn't really report or detail what conditions they were getting .

They would just use this as a protocol And then they would expose mice to that . And you remember last time we were speaking about the mouse RD-50 test the purpose of their test was particularly to measure the irritancy of the smoke using mice , but they also looked at lethal effects of carbon oxide and stuff like that .

And then the other main apparatus we could now come over from the United States to Europe was the DN-53436 apparatus And the din . These are all bench scale tests . The din test used a typical I think it's called a Leigh big furnace . It's a tube furnace .

So they had a long tube of about 600 millimeters I can't remember the exact length though , but quite long tube , about a meter long And about 25 millimeters diameter , just a simple furnace tube . But what they did , which was quite useful , was they had an annular furnace outside the tube so they could heat up the tube .

But they had the annular furnace to move on , a movement device . So they tested their material as a long strip And they'd start the furnace at one end of the furnace , would move along the strip , and now we're starting to get a more constant decomposition rate .

You see , because they're putting air through the tube at a constant rate and they're decomposing the furor at a constant rate as the heated furnace moves along the tube , doing successive plots of the specimen And what comes out at the end . Then it depends on the ratio of the mass loss of the strip of material and the air supply .

So it's the fuel-air ratio and it remains fairly constant to the duration of the test . Once you've got things up and running , it's fairly constant . Because the flow is steady and the fuel is constant And the exposure rate of the fuel is constant , so the fuel ratio is constant . And so when I started my work I started to use a variant of that approach .

But because I didn't have that particular apparatus , i used a simpler version where I got a standard tube furnace about half a meter long and a furnace tube to put in it silica furnace tube in it , and then I put the specimen in a little surrounding boat over the long strip and I pushed it in at a constant rate . So I'm achieving the same thing as din .

So I have a constant flow of air in so many liters or grams per minute And I put the specimen in a standard rate , 70 grams per minute . So I've got a defined fuel-air ratio .

And when I first started doing this work , we were looking at pyrolysis under nitrogen And that was quite easy to do And we did it at three temperatures 300 degrees , 600 degrees and 900 degrees centigrade in the furnace , and it was fairly straightforward to do because there's no flames involved . So it's all fairly predictable .

But I hope I'm getting towards the question whilst , because what we found with this and I'm exposing animals to this , i'm sure , as well as trying to measure the composition in simple terms CO concentrations , cyanide concentrations , carbon dioxide and oxygen are the main parameters , but also smoke , smoke-particular concentration and optical density .

So what we found when I decomposed polypropylene in a stream of nitrogen , what we got was a sort of hydrocarbon mist of liquid droplet , very fine particles , sort of something you get with a theatrical smoke generator in fact . And when I exposed animals to this mist it was pretty well had no effect on them at all .

There was no oxygen , so we couldn't make any carbon monoxide . There's no nitrogen , so you've got no cyanide , so there's no nasty friction gases . And because there's no oxygen to oxidize the hydrocarbon coming off this , there's no irritants being formed And so it's effectively a relatively harmless atmosphere .

Now , at the same time as I was doing this , professor Woolley and his colleagues Peter Fardell were doing GC mass spec analysis of these atmospheres , and what they found , with polypropylene and nearly everything else as well in fact , was a long I think we talked about this a bit last time a long series of chemicals .

So if you think of the hydrocarbons , the fourth one you get is methane , ch4 . And then you'll get CH2 , ethane , ethylene or propane , hexane . You've got the series right until you get to octane and things like that . And they found all those . Unfortunately they didn't get the methane on ethane because the nature of the apparatus they couldn't capture it .

What they found was they got this long series of hydrocarbon fragments which are basically fairly harmless . But the other thing they did find and you get this more when there's option around as well is that you get cyclic products produced .

So they got cyclohexane and styrene , styrene monomer and things like that , and this is a process called ring-cytylization which is very important in fires . So basically , the way I explain to my students and in simple terms that I understand it , is that you think of a long aliphatic hydrocarbon chain . The chain bites its tail .

I think the guy who discovered benzene dreamt about the snakes biting their tails . So you've got a ring formed And to start with the ring is called is cyclohexane , so it's six carbon atoms with a hydrogen but there are no double bonds in it . And then you get sort of progressive condensation reactions between adjacent molecules of benzene .

So you think of two foreign molecules of benzene in there coagulating together , and so these ringlets get bigger and bigger and bigger And so you get polyanomatic hydrocarbons . And then those get even bigger and they form smoke particles , and that's what smoke particles essentially are .

The plates are graphite , or most of them , but they start off as these condensing ring structures . So what you've got then is this long chemical series , from stuff like methane up to various types of aromatic hydrocarbons , finishing up with sort of actual smoke particles which are more or less pure carbon . They're just carbon .

When you've got rid of all the hydrogen , you just end up with carbon rings essentially . So that's what happens Now . Then later on , they start to depress and air down the furnace tube . Now our temperature is below the ignition point , so up to about 400 degrees C , something like that . So that's hot enough to get pyrolysis .

So you start to break down the materials in the way I've described .

But now you're able to get some oxidation reactions going on , and so you get a whole new set of products which are partially oxidized hydrocarbons and aromatic materials or aliphatic and aromatic hydrocarbons , and so you start with things like form aldehyde , a chrolyn , which is an unsaturated aldehyde , croton aldehyde , which is the next series .

These materials are carcinogens and they're also extremely irritant . You also get things like phenol and acetaldehyde and all these sorts of things , and Professor Woolley and his colleagues were measuring these in their GMC mass spectra .

They would come up with these two sets of big ranges of hydrocarbon-type organic products , a sort of aliphatic set , non-oxidized set and then an oxidized set And the oxidized sets getting worrying , quite toxic .

I spent many weeks going through the thing called the R-TEX database , the registry of toxic effects of substances , trying to find out the toxic effects of some of these weird , wonderful organic products that were appearing in these spectra And the basic , the bottom line , is that they're one of the obvious ones that we know about that are important acutely as

irritants or on damaging to the lungs And others that we know about with unknown carcinogens and things like that . Well , a lot we don't know much about at all , and in later experiments I exposed mice to some of these atmospheres . I'm looking at the irritancy now And I tried to get some GCMS work done at the same time as I'm exposing the mice .

So can I correlate the known atmosphere composition in terms of known irritant chemicals to the objective reduction in breathing in the mice ? And the answer was yes and no . Yeah , so if you produce an atmosphere like chloralized polypropylene which you've got no irritants in , then the mice aren't affected .

But as soon as you put oxygen in there and you get these oxidized compounds , they react very strongly . They find that atmosphere extremely irritant . But the potency of the irritancy , in terms of the mass you need to produce it , is minute . In other words , it's much , much .

The amount you need to decompose it to get the concentration mass loss concentration that caused a 50% decrease in breathing in the mice was tiny compared to the amount of the known products are in there . In other words , most of the observed irritant effects could not be accounted for by what we could measure in the atmosphere . And that's still the case today .

We still don't really know exactly what's going on there , so it's quite a complex situation .

And we're still talking about the same plastic material being in the tube , that is , polypropylene . Did you go all the way up to exposing mice to products of combustion of polypropylene , like fully burning it ?

And then of course that's much more complicated because polyurethane , when you heat it , breaks down into its two main components , which are a long chain polyole alcohol long chain alcohol which stays in the furnaces of sort of brown liquid And the eye toluene diisocyanate , the isocyanate component , which comes up off as a vapor and then reacts with itself to form a

kind of particulate smoke It's called yellow smoke which is highly irritant . So we were getting all these kinds of things going on . So we're getting a feel for what's going on .

But of course the elephant in the room , as it were , here is that nearly all fires that kill people In fact I would say all fires that kill , really that kill people are flaming farmers . Now it's not to say that people aren't exposed to smoldering and non-flaming situations , which can be hazardous .

But the bottom line is that nearly all the situations where you get smoldering or non-flaming . It's a very , very slow process And hopefully the effluents that come off during smoldering will be picked up by some detector or somebody will smell it or become aware And there's plenty of time to escape .

So I don't think many people become seriously injured or killed in non-flaming fires , smoldering fires Most people certainly die in flaming fires because they grow so quickly And the mass loss rate is exponential And you can very quickly get to very bad conditions in a flaming fire .

So what we really need to look at if we want to understand hazards to people is flaming conditions . So we started to move on to flaming conditions And what we found there was as soon as I burnt things with a flame .

Of course all these pyrolysis products are combusted in the flame And if you've got efficient combustion they're all burnt up And then the atmosphere becomes non-irritant by quite a large fact of 10 or even 100 .

So if it's got no halogens or anything in it like that , then the question comes in what does flaming do to the yields that you're going to get in these atmospheres ?

Well , moving on , moving on as we did , got more into this we started to realize it's fairly obvious , especially now that the fuel-air ratio of combustion , combustion conditions vary the products that you get .

So if you've got , we identified something we called well-ventilated flaming , where you get mostly carbon dioxide , water and heat from a simple polymer or wood type material And you don't really get a little bit of smoke particulates , which are fairly benign . You know the sort of smoke you get off from well-ventilated wood .

If you go and look at the deposits after the event , you find this a very dry , dusty particulate , whereas when you burn things under ventilated conditions and you look at the smoke this deposit from that , it's sticky , oily , nasty stuff .

So anyway , we've got well-ventilated flaming which is very efficient , can be very efficient and relatively low toxicity in terms of atmosphere

. And we're talking about one thing that we used to define things in terms of was the CO2-CO volume ratio , or some engineers like to use it the other way . So in a well-ventilated form , co2 concentration divided by CO concentration , mole fraction , is going to be something like 100 . You get hardly any CO and lots of carbon dioxide .

It's usually about , i don't know , 304 to 1 , co2 divided by CO And you're getting carbon monoxide yields of 0.2 gram per gram . Mass loss , fuel mass loss , right , that's kind of slow .

So we started to think about these under-ventilated fliers and measure the composition of these kind of atmospheres And we ended up putting this into ISO quite a lot with saying well , if you want to look at toxicity of your atmospheres , you need to consider 304 basic fire types .

So one type would be non-flaming oxidative decomposition , so when something is being decomposed but it hasn't ignited . Another scenario is where you've got a relatively small but well-ventilated fire , so a pre-flash-over well-ventilated fire kind of condition .

And the third main sort of condition would be something approaching flash-over , where you've got maybe fairly high compartment temperatures , 800 , 900 centigrade in the upper layer , something like that .

So these fires in compartment fires in real buildings tend to be under-ventilated , because when something goes to flash-over in a room with limited ventilation you're involving a huge fuel mass , so you tend to get a very high fuel content in your atmosphere .

And so we thought of post-flash-over fires as being typically under-ventilated fires , but that's something that's been some arguments about . If you like , can you have such a thing as a well-ventilated post-flash-over fire ? And we thought about you could get various degrees of under-ventilation in a post-flash-over fire And you could also get small under-ventilated fires .

So with a small-scale fire where you don't have high upper-layer temperatures in particularly , you can have one small item of burning like an armchair in a room and you can get quite under-ventilated conditions there . So there's a whole range of different conditions you can get And we really need to get a handle on what these conditions do to yield .

And it's round about that point that I was introduced to the concept of equivalence ratio . If something maps onto a mixture fraction to some extent , there's a way of handling it in CFD , is that right ?

So I think with Victor Barbauskas , who first introduced me to the idea of equivalence ratio , and I started looking at the papers by PITs and people like that were on global equivalence ratio and clean equivalence ratio .

And of course I then started to look at Architurosums wonderful body of work , and he was using a small scale or varying scale apparatus , the ASTM apparatus .

And there are plenty of different conditions in which the materials can burn And we often define them by the ratio of fuel to air , which is the equivalence ratio . Did I get it correct ?

. So , simply speaking , equivalence ratio is the actual fuel concentration divided by the actual air ratio in your test , as a function of the stoichiometric fuel air ratio , which is the ratio of fuel , that was just exactly the right amount of air to completely combust the fuel .

So put it simply if your equivalence ratio is one that means that in your particular test you have stoichiometric conditions , you have exactly the right amount of air to burn the fuel that you've got in your atmosphere .

If you have an equivalence ratio of 0.1 , which is what happens in early stages of fire sometimes then you've got a real big excess of air , So you've got 10 times as much oxygen as you need to burn your fuel . And if you've got an equivalent ratio greater than one , then you haven't got enough air to burn the fuel .

And those are the underventilated conditions that we get in many certain stages of many component fires , but we're still within the flammability limits , so it still can combust .

But as far as I can see he's getting non-flaming conditions then . So he still thinks that the equivalent ratio has some relevance while he's under conditions there where the flame will have extinguished So it's not actually burning with the flame . And that limit may vary a bit depending on your apparatus .

But essentially in large-scale component fires tests that I've done we're usually in the well below three . If you're getting up into that sort of area then your fires like to self-extinguish most cases . So the interesting , I'll say the interesting zone is from point one up to one and then up to about two and a half to three .

Equivalent ratio is the interesting band , if you like , in real fire conditions . And I do commend you to look at Arty's work . So in Arty's work he used a difference of apparatus to me .

He used a flat plate specimen placed inside a vertical cylindrical tube glass tube And then he radiated from outside with a halogen heat source And his specimen is on the load cell . So Arty controls the air going up the tube . So when he collects it's about the effluent up to top . So it's quite a nice piece of apparatus .

I'll forget the figure 2805 or something 2805 . Astm .

And he's got some lovely plots and data in his chapter . What he found was that when you decompose things and I've found exactly the same thing in my apparatus when you decompose things under well-ventilated conditions of five less than one , then you get very low yields of carbon monoxide , and these products of inefficient combustion is what we're talking about .

You mostly get carbon , but you get very high yields , stoichiometric yields in terms of carbon dioxide , and you get subwater and heat . And then as you increase the equivalence ratio in separate runs I do mine in set In my apparatus each equivalence ratio is maintained for one furnace test run .

So as you increase the equivalence ratio , nothing much changes until you get start to approach an equivalent ratio of one When . So , as you get to near stoichiometric conditions , then you start to see some signs of inefficient combustion . So you get more smoke particles , you get more carbon monoxide .

And one of the things I can do with my apparatus is do a complete , if you like , recovery and accounting of the fate of the fuel , because we know what the fuel composition is And so what we can calculate is the total on-bird hydrocarbon content .

We don't know what form it's in , but we know how much carbon and hydrogen is in or in an organic form in the atmosphere we're producing .

So all of these things start to increase slightly as you approach an equivalence ratio of one stoichiometric conditions And then once you get to one and above , that curve gets very , very steep and you get a very big increase in the yields of these products of poor ventilation . So you get big increases in smoke yields , smoke particulate yields .

You get big increases in carbon monoxide If you've got nitrogen in your fuel , you get big increases in cyanide produced , but what decreases is the products of well-ventilated combustion , which is carbon dioxide . So one thing we look at is the total fate of all the carbon in the fuel .

Where does the carbon go , and we can do an exact accounting of where all the carbon's gone .

So when you get to an equivalence ratio of two , which you've got half enough oxygen to burn your fuel , then you find , not surprisingly , that the carbon dioxide yield is more or less halved , but the carbon monoxide yield goes up to something like towards 0.2 ram per gram , bearing a bit depending on the material . So the carbon monoxide is going up .

I mean , and it's quite interesting to compare what goes on with organic nitrogen and what goes on with organic carbon . And what I found in my experiments with various types of materials was you can express the yields of these gases as actual yields . So how many grams of carbon monoxide do we get per gram of fuel burned ?

Think of something simple like polypropylene . So how many grams of carbon monoxide released per gram of polypropylene ? go on ? Or we can think of it in a normalized sense , saying of the carbon content of the starting fuel , polypropylene , how much is released in the form of carbon monoxide , how much in the form of carbon dioxide ? So that's the normalized yield .

So the normalized yield varies depending on the carbon content of the starting material . So the normalized yield for wood , the actual yield of carbon oxide for wood , might be 0.1 . The normalized yield might be 0.2 because there's only 50% carbon in wood .

So your lung is going to get a lot more carbon monoxide than your hydrogen cyanide because you're starting with much more carbon than your nitrogen

. When you normalize the yields , the interesting thing you find is the normalized future plot of the normalized yield of carbon in the form of carbon monoxide as a function of equivalence ratio . well , against the conversion of fuel nitrogen to hydrogen cyanide , they pretty well exactly match on a one-to-one basis .

So if you convert 1% of the carbon to carbon monoxide , you will also have converted 1% of the nitrogen to hydrogen cyanide . And that's very nice because hydrogen cyanide is quite difficult and challenging to measure . Most people have got an apparatus to measure carbon monoxide in their tests , but not many people can cope with things like cyanide .

So I'm giving the engineer now a way of getting a pretty good computation of if you've measured the CO and you know the elemental composition of your starting material and you know the nitrogen carbon content of your starting material and your fuel , you can pretty well calculate how much cyanide you might be getting as a function of equivalence ratio , right ?

So I think we start to come together .

And as I talk with you for hours now , i understand more why It's really a complicated world in which we'll not have an exact solution for everything , even if we get the yields and stuff exactly . It's doubtful we will get perfect data on LC50 , on primates , for any toxicant species there is , because it's just unavailable at this point .

But what you propose in here is to use science and research and data to have the second next thing , the second next approximation . So if you cannot measure HCN , well , you figured out how to do that based on the material composition , the elemental components of the material , and it must be like that because that's the elemental physics and chemistry .

I just wanted to say I absolutely love this way of thinking .

Although there's complexity there , and I think most of these sort of topic areas , when we first start to look at them , we think they're fairly simple , like how much carbon monoxide you produce in a fire .

You might think that's a fairly simple thing to do And then the more we go into it , as we've been describing the complexity starts to emerge , and then my mission I'm hoping to get across to you is that we can sort of come out the other side a bit and start to find simplicity re-emerging , or that we can measure and control certain parameters .

Then we can get a pretty good handle on what's going to be produced in our fire and how we can bring it into our CFD analysis .

So I'm trying to persuade you in a way that we can deal with this or produce tools that NGNIC can use that are good enough to give us a good understanding of what's happening in our fire instance , which is essentially what we really want to do And , ladies and gentlemen , i am really sorry but that's it for today .

I think it's a whole another hour of interesting deep dive into toxicity in fires with David . So you feel very invited to the next week's episode . I hope this one already brought you a lot of value for me In a way , fascinating and perhaps disappointing , devastating .

I mean we focus so much on simple material properties like suit heels and CO2 heels And yeah , sure , they describe the properties of our materials and to what extent they can produce some products in very specific combustion conditions , but do we even have these conditions on our buildings ? really , i feel like we opened a can of worms in here .

I mean I understood that having different equivalence ratio will lead to different species productions , but it's really such a vast difference if you had under or over ventilated fire and just putting a single value on the products It just simply doesn't make sense . I mean , i got to rethink this .

I need to understand how I can incorporate this in my engineering , in my studies . I need to understand if I need to do that And if I do , if I find it out , i'll share it with you . I mean David , in the beginning of the episode talking about mole blank fire , has given quite a good idea how this approach can be incorporated .

Knowing only major species , if you know relevant minor species productions in certain equivalence ratios , We can figure them out . And because in CFD we conserve the yields and species transport and mass , you can figure out , based on ratios , what's the local concentration of any pollutant if you know the ratios . So there are ways not easy , perhaps worth it .

That's really high level engineering and I am super thankful to David for coming into Fire Science Show and sharing this with me and you . So thank you for today . Next Wednesday , we know what you can expect More chemistry just before summer . I guess that's exactly what you need . I feel this .

Interviews with David are perhaps a little harder in terms of the amount of high level knowledge found in here . In my professional career I never learned that much about toxicity as I did in these four hours talking to David , so I hope for some of you . At least it's the same experience . Thank you for being here . See you here next week . Bye .