MCC and TGA data don't fit together

[Tristan Hehnen]

Hello everybody!

When plotting the reaction rates from TGA and MCC data of the same material, they don't fit together. For example, Ding et al. describe this here (https://www.sciencedirect.com/science/article/abs/pii/S0010218019300136?via%3Dihub). They dealt with this discrepancy, by shifting the MCC data by a couple of Kelvin such that the peaks would align. They also state that the magnitude of the shift depends on the material. They did run TGA and MCC experiments at the same heating rate of 10 K/min.

1. For all institutes that have used MCC data to determine their pyrolysis kinetics, please explicitly indicate in your submitted data sets:
    - if you have adjusted the MCC data
    - how you did adjust the data (procedure)
    - by which value you have shifted the data

2. It would be interesting to understand which procedure you are using to adjust the data. In the article mentioned above, they have the same heating rates for both methods (TGA and MCC), but in the MaCFP repo there are no heating rates that fit together.

3. It would be very helpful, if the institutes that have access to an MCC apparatus would provided experiment data for different heating rates. Explicitly the following heating rates are of interest: 2.5 K/min, 5 K/min, 10 K/min, 15 K/min, 20 K/min and 50 K/min.

4. It would be very helpful, if the institutes that have access to a TGA apparatus would provided experiment data for 60 K/min.


Best,
Tristan

[Hostikka Simo]

Hello Tristan,

Very important point. In our 2017 paper, we noticed that the MCC heating rate had not been what was specified, but something different. So we had to afterwards adjust the heating rate, not shift it.

http://dx.doi.org/10.1080/00102202.2017.1295959

 

Simo

--
You received this message because you are subscribed to the Google Groups "MaCFP Discussions" group.
To unsubscribe from this group and stop receiving emails from it, send an email to macfp-discussi...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/macfp-discussions/c1daf386-5f65-439b-a417-7e19e2d04002n%40googlegroups.com.

[Tristan Hehnen]

Hello Simo,

thank you very much for the link, I'll have a look at the article!

Best,

Tristan

[Morgan Bruns]

Hi, Tristan. Thank you for bringing up this important issue. It's not something we've talked about in any of the virtual meetings because we've been focused on gasification. But modelers will need heat of combustion values for simulating the SBI and parallel panel cases so we should make this information available.

Maybe an experimentalist could chime in here, but my understanding of the issue is that MCC uses relatively large samples at relatively large heating rates such that the sample is not isothermal and the reported sample temperature is not actually representative of the temperature of the material. My approach to handling this issue has depended on knowing the pyrolysis kinetics ahead of time (from TGA) and then doing a simple temperature shift in the MCC data to estimate heats of combustion. This seems to work pretty well.

In response to your numbered items above,

1. Determine pyrolysis kinetics from MCC data doesn't seem like a good idea to me since you don't really know the actual sample temperature from the reported data. The difference between sample temperature and reported temperature will depend on the material and so we can't just assume a constant delta T for the apparatus. Ideally, one could develop a detailed model of the MCC to calculate the true sample temperature, but this would require knowledge of the thermal properties of the material (k, rho, and c).

2-4. It seems like the 60 K/min heating rate for MCC is pretty standard, and I'm unsure of the reasons why. I wouldn't recommend doing TGA much faster than 10 K/min because then you'd be dealing with the same type of temperature shift that we're talking about with MCC---the average sample temperature will no longer correspond to the reported thermocouple temperature.

Morgan

[Dietenberger, Mark - FS, WI]

Hi All. Thank you for keeping me in the loop.

Another word of caution about the MCC is its reliance on oxygen consumption correlation in a stoichimetric  burn, to derive the net heat of combustion. The reality is that there are some fuel that when pyrolyzed will present volatiles that do deviate from the simple oxygen consumption correlation for the stoichiometric HOC and do burn in a flame with such production of soot and CO that reduces the flaming net heat of combustion quite a lot. This is especially true in the case of solid polystyrene fuel.

 

As to the heating rates in TGA, we have found 3.5, 10, and 60 K/min for thin vegetative foliage material like leaves (and especially if including the fuel as wet and dry, and in inert gases and air, that add variations to kinetics analysis), to be suitable for kinetics analysis with an analytical solution method for implementation in Excel spreadsheet.  We just need to keep the samples thin enough and yet heavy enough for the TGA resolution. For a vegetative material that chars, I consider the MCC data that uses inert gases, as useless, because of the oxidative pyrolysis that occur in the real world. I would vouch for testing the fuel as a thin material in a specialized test holder with T/Cs in a cone calorimeter test in air to simulate the real world as much as possible (I have analyzed these data, just need to get it into publications). Further, there is value in obtaining also the proximate and ultimate analysis, to prove the oxygen consumption correlation, like I was able to do for live organic materials in this paper,

Summative and ultimate analysis of live leaves from southern U.S. forest plants for use in fire modeling | US Forest Service Research and Development (usda.gov)

 

Mark Dietenberger

From: macfp-di...@googlegroups.com <macfp-di...@googlegroups.com> On Behalf Of Morgan Bruns
Sent: Monday, June 26, 2023 9:44 AM
To: MaCFP Discussions <macfp-di...@googlegroups.com>

Subject: [External Email]Re: MCC and TGA data don't fit together

 

[External Email]
If this message comes from an unexpected sender or references a vague/unexpected topic;
Use caution before clicking links or opening attachments.
Please send any concerns or suspicious messages to: Spam....@usda.gov

To view this discussion on the web visit https://groups.google.com/d/msgid/macfp-discussions/da0b42b8-1270-4885-99a0-ba2050cea074n%40googlegroups.com.

This electronic message contains information generated by the USDA solely for the intended recipients. Any unauthorized interception of this message or the use or disclosure of the information it contains may violate the law and subject the violator to civil or criminal penalties. If you believe you have received this message in error, please notify the sender and delete the email immediately.

[Isaac Leventon]

Thank you, Mark, you highlighted an important distinction of the measured MCC heat of combustion, which more closely represents the heat of complete combustion of the gaseous volatiles produced by these fuels. Values of HOC calculated from other data (e.g., the cone) may include the impact of partially incomplete combustion; whether this slightly reduced value (from the cone) is the best option for larger scale flame spread modeling is not necessarily clear to me (and may also vary depending on the combustion model of choice).

To Tristan's original question regarding the MCC data available on the repo - it is nontrivial to conduct replicate experiments at six unique heating rates so I do not know that data can be easily provided. I would agree with Morgan's suggestion that TGA data is likely more suitable for use in the derivation/calibration of reaction mechanisms and associated kinetics (more direct measurement of your output quantity, Mass / Mass loss rate versus HRR) and I would also hesitate to use 60k/min data from the TGA for this purpose (as we would then be stressing the thermally thin approximation used there).

To Simo's note on heating rates - I'd agree this is important to confirm for all these mg-scale data. Actual versus nominal heating rate data is available in the preliminary summary of these experimental results [https://github.com/MaCFP/matl-db/releases/tag/v1.0.0] and by running the scripts in this repo.

The heating rate chosen for the MCC, 60K/min, allows for the calculation of other properties (e.g., fire growth capacity or heat of combustion, which is an integral quantity) that do not necessarily require the same assumptions / sample behavior during model calibration as kinetic property determination; also, at 60K/min you get a 12 minute test + similar cool down time, which is much more practical (for many applications with less stringent requirements) than a 1+ hour long test]. At progressively higher heating rates, yes you may see a shift in MLR/HRR towards higher temperatures; as we've also discussed before, the design of the MCC (with a hot furnace directly above / downstream of the pyrolyzer region) may also allow for heating of the sample and thus an apparent shift in HRR vs. the measured sample temperature, but in the *opposite* direction as typically seen just when going to higher heating rates (noted in the previously suggested references). It appears there is still some confusion on this effect, maybe I did not do a good job describing this last time we caught up and I'm worried a long text-only description will be insufficient. I will reach out to a colleague involved in the original design and production of this instrument and who has an "all things MCC YouTube channel" to see if it makes sense to prepare a more thorough/video response.

-Isaac

To view this discussion on the web visit https://groups.google.com/d/msgid/macfp-discussions/PH0PR09MB844162BE1173C02E29E06191CF26A%40PH0PR09MB8441.namprd09.prod.outlook.com.

--

--------------------------------------

Isaac Leventon, PhD
National Institute of Standards and Technology

Fire Research Division

Building 224 - Room A265

Gaithersburg, MD 20899

(301) 975-6887

[Stanislav I. Stoliarov]

The MCC pyrolyzer was not designed to operate below 5 K/min. The heating rate control of the original version of the instrument was geared toward high heating rates (60 K/min or above). The instrument is supposed to maintain the prescribed heating rate of the crucible well after a brief overshoot. However, the commercial versions of the instrument do not always do that due to flaws associated with their own implementations of the instrument and controls. Therefore, some commercial instruments are unable to deliver prescribed heating rates. Even in a well designed version of MCC, the temperature for the sample itself is not as well controlled as in well-calibrated and correctly setup TGA. Therefore, for accurate kinetics TGA is prefered.

Stas   

--------------------------------------------------------
Dr. Stanislav I. Stoliarov, Professor
University of Maryland
Department of Fire Protection Engineering
3104C J.M. Patterson Bldg.
4356 Stadium Dr.
College Park, MD 20742
Phone: 301.405.0928
Fax: 301.405.9383
Email: sto...@umd.edu

--

[Tristan Hehnen]

Hello everybody,

thank you all very much for your time and feedback!

I'm still confused about the shifting of the data.

In the article mentioned by Simo [1], they found that the true heating rate of the furnace was off, compared to the set value. For set values of 20 K/min and 60 K/min, the actual heating rates were 33 K/min and 85 K/min, respectively.
@Simo, I've got a few questions, could you please elaborate:
- What is meant with the "furnace temperature", is this the supposed sample temperature (measured below the crucible) or the temperature of the heater?
- Did you also see heat release before the mass release in the rate-vs-temp-plot, as reported by Ding et al. [2]?

Ding et al. state that the heat release and mass loss rate profiles were shifted [2], see example below. Specifically, the HRR profile was determined to be shifted towards lower temperatures. This was deduced from observable heat release in the absence of mass loss. Thus, they adjusted the HRR profile, assuming the MLR profile was "correct". This shift is reported to be between 4-20 K, depending on the material.

The difference in heating rate reported in [1] seems large. Thus, I was wondering what this translates to, in terms of distances of the reaction peaks in the rate-vs-temp-plot and did set up a small test. Using our MaCFP-3 parameter set, I did recreate the described setup of [1] in FDS. Note: [1] used wood and I used PMMA and assumed that the "furnace temperature" in [1] refers to the temperature recorded next to the sample.
This translates to a shift of 7-10K between peaks in the rate-vs-temp-plot, see example below. This difference is well in line with the differences reported in [2].

Looking at the MCC data in the MaCFP repo (NIST, 60 K/min). When determining the heating rate from the reported temperature, it seems feasible to get reasonably close to the desired heating rate.
Thus, even if the combustion chamber introduces extra heat into the sample this seems to not necessarily show itself in the measured temperature below the sample. I would also find it a bit surprising if the combustion chamber would change the heating rate so much, to be honest (even though I'm not an experimentalist and I've not operated such apparatuses myself). For the heating rates reported in [1], this would mean that the extra energy increases the heating rate measurably by 15% to 65%.
It is interesting that the difference in the rate-vs-temp-plot in my assessment for [1] and from [2] are notably close and the heating rate seems to be recreated in NIST MCC data quite well.

The question arises, why we should expect that the MLR and HRR would overlap in the first place, if recorded under the same conditions. Wouldn't this only be true, if the heat of combustion (HOC), as a constant factor, would be fully accounting for the difference between both curves?
I would argue that the effective heat of combustion of the released gas mixture is not constant over the course of the experiment. Simple examples would be materials like wood that contains some moisture, or technical materials designed to be fire retarded by releasing carbon dioxide early on.
For our recent article about estimating PMMA material parameter sets (presently under review), I was not able to match 10 K/min MLR with 60 K/min HRR in the simulation. There, I did assume that the HOC is constant throughout the experiment. Near matching data series could only be arrived at, by using heating rates that are closer together, i.e. 50 K/min MLR with 60 K/min HRR. 

However, if the composition of the released gas mixture is allowed to change during the experiment, this will lead to a variable effective heat of combustion. Then it is possible to reproduce a wide set of experiment data well, with a single parameter set. This is provided in the discussion of said article as a possible solution and is encapsulated in the parameter set we submitted to MaCFP-3, where we actually tried it out.

A couple of days ago, I found this article [3]. Here, the author explicitly compares TGA mass loss rates with MCC heat release rates to determine pyrolysis reaction kinetics. Figure 3.a shows how I naively would expect MLR and HRR to be related, if the HOC is not constant and the heating rate could be recorded accurately. They show the variability of the effective HOC in figure 3.b, comparing it also to the constant effective HOC. They are also able to reproduce more complex decomposition behaviour (PVC) with a single reaction, by having A and E depended on some global conversion for the different materials. Figure 3.a shows that the HRR starts at higher temperatures than MLR and they also do not mention to have observed a shift in the experiment data between both methods. It's an interesting article and I highly recommend reading it.

Something, somewhere seems to be wrong here.

With respect to the heating rates: 

It makes sense to me, that for higher heating rates the peaks move to higher rates and higher temperatures in the rate-vs-temp-plot. 

I can also follow the argument, that if the sample receives more energy from the combustion chamber, the actual heating rate of the sample would be larger than the recorded one, due to some transmission delay. Thus, it would look like the decomposition takes place at lower temperatures.
As outlined above, the extra energy seems to translate into quite a large difference in heating rate, thus:
- How do we know that this effect (energy introduced from combustion chamber) exists and is not user error in operating the device?
- If we know that this effect exists, why is it not accounted for?


In articles [1,2], the shift is mentioned more in passing than being object of the investigation itself. The article which explicitly assesses the combination of TGA and MCC [3] does not report a shift and shows reasonably matching data series. Does someone of you know of literature that explicitly deals with the reported shift and would be so kind to link it here?

[1] https://www.tandfonline.com/doi/abs/10.1080/00102202.2017.1295959?journalCode=gcst20
[2] https://www.sciencedirect.com/science/article/abs/pii/S0141391016301446?via%3Dihub
[3] https://www.sciencedirect.com/science/article/abs/pii/S0040603114003311?via%3Dihub

Best,

Tristan

[Stanislav I. Stoliarov]

Tristan,

I think it is important to understand that the shift in temperature in the MCC data reported in Ref. [2] (the work of my group) is not a consequence of the difference between the set and overall actual heating rate of the sample. This difference is due to an existence of a spatial variation in temperature in the sample/sample holder system of the MCC. The sample (especially its top surface) is systematically hotter than the thermocouple located below the crucible. After a short induction period at the start of the heating ramp, they both attain the same heating rate, but because the top of the sample and crucible are exposed to additional radiation from the combustor, the difference in temperature between the sample and the thermocouple persists throughout the heating process. You can qualitatively simulate this effect, but to do it, you need to include transient heat transfer through the sample and crucible and all contact resistance effects into your simulations.

Stas     

--------------------------------------------------------
Dr. Stanislav I. Stoliarov, Professor
University of Maryland
Department of Fire Protection Engineering
3104C J.M. Patterson Bldg.
4356 Stadium Dr.
College Park, MD 20742
Phone: 301.405.0928

Email: sto...@umd.edu

To view this discussion on the web visit https://groups.google.com/d/msgid/macfp-discussions/163ec282-bc55-4b9f-8583-08118f935912n%40googlegroups.com.

[Hostikka Simo]

Tristan,

• I think we meant the sample (crucible) TC temperature.
• I would not pay too much attention to the shifts in this case. Too many uncertainties. But systematic rate error was important.

 

Thanks Stas for the good explanation of what is going on inside MCC

 

Simo

 

 

From: macfp-di...@googlegroups.com <macfp-di...@googlegroups.com> On Behalf Of Tristan Hehnen

Sent: Sunday, 6 August 2023 14:27
To: MaCFP Discussions <macfp-di...@googlegroups.com>

Subject: Re: MCC and TGA data don't fit together

 

Hello everybody,

thank you all very much for your time and feedback!

I'm still confused about the shifting of the data.

In the article mentioned by Simo [1], they found that the true heating rate of the furnace was off, compared to the set value. For set values of 20 K/min and 60 K/min, the actual heating rates were 33 K/min and 85 K/min, respectively.
@Simo, I've got a few questions, could you please elaborate:
- What is meant with the "furnace temperature", is this the supposed sample temperature (measured below the crucible) or the temperature of the heater?
- Did you also see heat release before the mass release in the rate-vs-temp-plot, as reported by Ding et al. [2]?

Ding et al. state that the heat release and mass loss rate profiles were shifted [2], see example below. Specifically, the HRR profile was determined to be shifted towards lower temperatures. This was deduced from observable heat release in the absence of mass loss. Thus, they adjusted the HRR profile, assuming the MLR profile was "correct". This shift is reported to be between 4-20 K, depending on the material.

 

 

 

The difference in heating rate reported in [1] seems large. Thus, I was wondering what this translates to, in terms of distances of the reaction peaks in the rate-vs-temp-plot and did set up a small test. Using our MaCFP-3 parameter set, I did recreate the described setup of [1] in FDS. Note: [1] used wood and I used PMMA and assumed that the "furnace temperature" in [1] refers to the temperature recorded next to the sample.
This translates to a shift of 7-10K between peaks in the rate-vs-temp-plot, see example below. This difference is well in line with the differences reported in [2].

 

 

 

Looking at the MCC data in the MaCFP repo (NIST, 60 K/min). When determining the heating rate from the reported temperature, it seems feasible to get reasonably close to the desired heating rate.
Thus, even if the combustion chamber introduces extra heat into the sample this seems to not necessarily show itself in the measured temperature below the sample. I would also find it a bit surprising if the combustion chamber would change the heating rate so much, to be honest (even though I'm not an experimentalist and I've not operated such apparatuses myself). For the heating rates reported in [1], this would mean that the extra energy increases the heating rate measurably by 15% to 65%.
It is interesting that the difference in the rate-vs-temp-plot in my assessment for [1] and from [2] are notably close and the heating rate seems to be recreated in NIST MCC data quite well.

 

 

 

The question arises, why we should expect that the MLR and HRR would overlap in the first place, if recorded under the same conditions. Wouldn't this only be true, if the heat of combustion (HOC), as a constant factor, would be fully accounting for the difference between both curves?
I would argue that the effective heat of combustion of the released gas mixture is not constant over the course of the experiment. Simple examples would be materials like wood that contains some moisture, or technical materials designed to be fire retarded by releasing carbon dioxide early on.
For our recent article about estimating PMMA material parameter sets (presently under review), I was not able to match 10 K/min MLR with 60 K/min HRR in the simulation. There, I did assume that the HOC is constant throughout the experiment. Near matching data series could only be arrived at, by using heating rates that are closer together, i.e. 50 K/min MLR with 60 K/min HRR. 

 

 

 

 

However, if the composition of the released gas mixture is allowed to change during the experiment, this will lead to a variable effective heat of combustion. Then it is possible to reproduce a wide set of experiment data well, with a single parameter set. This is provided in the discussion of said article as a possible solution and is encapsulated in the parameter set we submitted to MaCFP-3, where we actually tried it out.

 

 

 

 

A couple of days ago, I found this article [3]. Here, the author explicitly compares TGA mass loss rates with MCC heat release rates to determine pyrolysis reaction kinetics. Figure 3.a shows how I naively would expect MLR and HRR to be related, if the HOC is not constant and the heating rate could be recorded accurately. They show the variability of the effective HOC in figure 3.b, comparing it also to the constant effective HOC. They are also able to reproduce more complex decomposition behaviour (PVC) with a single reaction, by having A and E depended on some global conversion for the different materials. Figure 3.a shows that the HRR starts at higher temperatures than MLR and they also do not mention to have observed a shift in the experiment data between both methods. It's an interesting article and I highly recommend reading it.

 

To view this discussion on the web visit https://groups.google.com/d/msgid/macfp-discussions/163ec282-bc55-4b9f-8583-08118f935912n%40googlegroups.com.