Without much detail about the materials being tested, it next to impossible to give a good scientific answer.

But with my little experience with polymer, a learned guess would be cross-linking due to thermal activation. Looking at the temp range over which the weight changes it looks more epoxy based materials where the crosslinking density has direct relationship with temperature.

Usually a mass increase of a sample is done on purpose. For example, performing a redox reaction, where instead of an inert atmosphere, you deliberately introduce specific reactant gases. Mass increase is due to adsorption from substance in the atmosphere. For example, if you have CO2 atmosphere, and your sample is some oxide material, it is likely that the mass increase is due to CO2 adsorption, turning the oxide into a carbonate, thus a mass increase.

So I would suggest making sure that it really is just N2 in the atmosphere. It is unlikely to be N2 that has adsorbed on the sample surface. Culprits could be carbon in the line.

The other thing to check the crucible you put your sample into, or the cleanliness of the apparatus. Again, it could be due to impurities in the system that has lead to this. You might want to check what the people before you have used in their experiments.

A mass increase usually is interesting, but since you say that it is in N2 atmosphere, you may want to double-check all this. Your sample may have transformed and changed density, but don't forget that it's a mass differential, not volume. A sample may change density, but that just means the volume has changed thus the mass should be more or less be the same.

It is hard to make good comments without information about your sample composition. In fact in pure N2 the behaviour of your sample seems a little strange. Are you sure it is pure N2 without any traces of oxygen? Do you have an open or closed cap?

I agree with Jerry. If there is not an experimental mistake, the mass increase phenomenon is intersting. But it has to be verified. Most of the case, an oxidation reaction occurs, in the TGA cell. Or there is a contamination in your crucible. Repeat the experiment with a new crucible and verify if you have already N2 withouth contamination (N2 bottle is empty!!)

I am also facing the same phenomenon of increase mass while performing the experiment of TGA for a thermostable protein. I didn't find any scientific reason, and I also repeat the experiment for several times and got the same result.

Hi Trinath,

Did you observe an increase in mass, or did the mass spring back to original values after boserving a mass decrease?

I observed the later in my TGA experiments with thermoset polymers in inert athmosphere. After observing a decrease in the mass once the degradation stage was reached, a sudden increase in the mass occurred. I noticed that the material inside the crucible overflowed and showed pyrolization. My hypothesis is that a certain amount of decomposition products fell from the crucible to the bottom of the weighing scale, thus showing this sudden increase of weight.

May this have happened to you?

I with Jerry.

Dear all,

I have carboxylate group in my coordination polymer. Is there any chance of adsorption of CO2 which already expelled from the sample during TGA?

My first thought is that some partial oxidation could be taking place.

I can think of 3 reasons:

Firstly, I would keep in mind that N2 is not an inert gas..

Second: baseline drift (buoyancy effect)...try running an identical empty pan and subtracting that baseline from your data

Third: Its quite possible that some decomposition products did not get carried away and got deposited back...try changing the gas flowrate in this case.

Nagaparadeep N.,

One thing you are also not clear on. Does the mass increase become larger than the initial mass of the polymer sample, or is it just an increase that doesn't surpass the initial mass?

If you are implying that the mass increase due to the carboxylate groups being decomposed to CO2 followed by re-adsorption by the coordination polymer, it is very unlikely for the following reasons.

I think now that you have mentioned that you are analyzing polymers, I assume other people in your group are possibly analyzing the same thing, which leads me to believe that you have very likely a contamination in the system, left behind by residual carbon from prior users/experiments which allows for the effect of added mass to your sample during the heating.

The contamination could possibly be from the crucible, the interior of the chamber, or the gas lines. Regardless, you should run a blank and analyse the gas in order to see if you indeed have carbon in the system.

Dear Nagapradeep N.,

From my experience on TGA, the small increment of mass of your sample that you are observing at these temperatures is the Archimedes effect. When you hang an object in a fluid, the fluid tends to force the object upwards. Before you start your TGA analysis, you put your sample in the analyzer and the fluid (N2) force the sample upwards. It is when you tare your analyzer. When the analysis starts, the density of the fuild is sligthly decreased (by increasing the temperature) and consequently the hanged sample goes down. Then you record these changes as mass increment but they are produced by a small change in the density of the sourrounding fluid. You should observe the same effect with an inert sample (a piece of glass?) and same temperature program. This effect is worse at higher heating rates.

I hope this explanation could help you.

This effect is called buoyancy effect.

To check, if it is related to this, you need to record a TGA profile of an empty container.

If the TGA curve of your empty container shows an increase, you got it !!

Then you will subtract the TGA curve of the empty container from the TGA curve(s) of your sample.

You might need to redo this each time the analysis conditions are changed (temp rate, type of container, gas flow...).

One has to record the TG curve of thermally inert material under identical conditons (mass and  heating rate) used that for sample. When TG curve of sample is substracted from theTG curve of  thermally inert material, then TG curve of actual mass change due the sample is obtained. Thus, TG curve of thermally inert material nullifies the buoyancy effect, convection current etc..

Other then the buoyancy effect, who effects on the measurements, one thing that we forget is that N2 is not a inert gas. It is also be verified if there is a reaction between your sample(and container) and N2 during heating...    

I agree if it is a large change than it is the N2 gas reacting with the sample (very unusual but I have seen it once) or you have contaminated gas.  However, small changes will be most likely due to buoyancy.  One way to resolve the issue is to purge with He or argon.  Please make sure you calibrate when switching gasses as thermal conductivity of the gasses are very different.

Taking into consideration the temperature range, the oxygen impurity of N2 gas might be the reason for this behaviour. Did you purify the N2 gas before using or used directly from the container ? 

Hi Dear Nagapradeep N.,

A slight weight increased in the early stage of degradation could be explained due to buoyancy effect of TGA equipment.There are few fundamental thermal analysis text book and papers explained about the buoyancy effect in TGA equipment:

 Gabbott, P.. Principles and applications of thermal analysis. (Ed.), (2008), UK, John Wiley & Sons.

 Menczel, J. D., & Prime, R. B. Thermal analysis of polymers: fundamentals and applications. (Eds.). (2014), UK,  John Wiley & Sons.

 Vaimakis, Tiverios C. "Thermogravimetry (TG) or Thermogravimetric Analysis (TGA)."

 In TGA analysis buoyancy is the upward force on the sample produced by the surrounding atmosphere, which will affect the apparent mass during TGA experiment. The buoyancy effect occurs as the density of the atmosphere in the balance decreases with increasing temperature, resulting in apparent mass gain. Consequently, it depends on the volume of the sample and the density of the atmosphere.

The root cause are as follows:

1.)During an experiment, the gas in the system is heated, primarily by conduction at low temperatures and by radiation at higher temperatures. In the initial stage the gas temperature is in advance of the sample/crucible temperature. The buoyancy of the gas itself decreases and therefore the sample/crucible appears to gain weight. This is, in part, due to the differences in thermal conductivity, density, and heat capacity for the gas and the sample/crucible. This effect becomes less at higher temperatures as radiation becomes the primary route for heat transfer. Thus the largest effect is seen at lower temperatures, usually at the start of the experiments, with only smaller increases being seen at higher temperatures.

 2.)On initial heating in a furnace, the wall of the furnace is hotter than the sample area. The gas at the furnace wall heats up first and starts to rise. Cold gas is forced down the centre of the furnace and this impinges on the crucible/sample giving an increase in weight. This convection effect is minimized in modern systems by careful design and by having a small volume within the furnace.

Hi Dear Nagapradeep N.,

A weight gained at 70 Celsius and 140 Celsius could also be due to oxidation of the tested materials. Please find discussion regarding this issue in one of my paper as follows:

Pickering, K. L., Khimi, S. R., & Ilanko, S. (2015). The effect of silane coupling agent on iron sand for use in magnetorheological elastomers Part 1: Surface chemical modification and characterization. Composites Part A: Applied Science and Manufacturing, 68, 377-386.

Apparent mass changes in TG measurements are well understood on the basis of bouyancy and 'Thermomolecular flow effect.The latter is however not applicable in the present case as the data are obtained at the normal laboratory .pressure. Decreasing density of the gas around the sample region during heating results in the sinking of the pan containing the sample. Moreover , the viscous drag experienced by the hangdown wire and the sample pan during decent of the cold gas by convection could also contributeThe to the apparent mass gain. The mass error resulting from these effects can however be corrected by recording a blank run under identical experimental condition and appropriately correcting the curve recorded with the sample.The present day commercially available TG instruments incorporate software for such corrections.

Usually, this happens due to reactions such as oxidation