If the measurements are correct, I can think of three hypothesis:

1. The regeneration was incomplete and small amounts of impurities continue desorbing throughout the adsorption/desorption cycle.

2. The material is degrading. Are you determining O2 isotherms? Maybe some CO2 is being produced.

3. The sample is being sucked during evacuation. Make sure the initial and final sample masses are the same.

However, the hypothesis raised by Ka Lok Chiu is the more probable one. Make sure your machine is working properly by testing a well characterised material.

Theoretically,  desorption curve is higher than adsorption curve in low pressure area if the material is mesoporous. If the material is microporous, both curves should be matched together. In your case, the problem may come from your machine and you may conduct a trial test with a standard.

If the measurements are correct, I can think of three hypothesis:

1. The regeneration was incomplete and small amounts of impurities continue desorbing throughout the adsorption/desorption cycle.

2. The material is degrading. Are you determining O2 isotherms? Maybe some CO2 is being produced.

3. The sample is being sucked during evacuation. Make sure the initial and final sample masses are the same.

However, the hypothesis raised by Ka Lok Chiu is the more probable one. Make sure your machine is working properly by testing a well characterised material.

This is an impossible situation in a well measured curve, unless some of the artifacts  are occurring. I think the ones that Rui Alfonso mentioned, would change the adsorbed amount before beginning the experiment, but would have no effect on putting desorption over adsorption.

There are almost no details in your question to try to suggest a better answer: Adsorption of what on what? In which conditions? In which kinda equipment? How much sample? Which approximate surface area? What is the particle size? etc...

Hi Fernando,

Thanks for the remark on my comment. I think that, at the end of your 1st paragraph, you meant "below", not "over". Anyway, I do not agree with your point.

By definition, no. 2 does not occur prior to the beginning of the experiment.

No. 1 can occur if the impurity desorption is occurring slowly enough as to not be detected, neither before the experiment nor during adsorption or desorption curve determination, but quickly enough to alter the results. Naturally, this error will only go undetected if the operator is not careful in assessing the 100 % uptake value. However, I would not be surprised if an unexperienced person was unable to detect such a situation.

No. 3 also depends on how careful the operator has been during the regeneration stage when compared to the desorption stage. Desorption involves evacuation, so, in my mind, the risk of sample drag is always present.

I hope to have clarified my previous points.

Rui Afonso (WITH NO "L")

When you say the lowest you mean the presence of a hysteresis loop? If so the cause could be the presence of micropores or bottle-shaped pores. However it is not common in carbonaceous samples. Perhaps the sample is not well degassed for analysis. I recommend making a little bit longer or at a higher degassing temperature and repeat the analysis to verify that there is indeed a difference in the curves.

greetings

Fernando

The most obvious answer is usually the correct one. The most obvious here, as stated above, is that the sample has been drawn out of the burette during the outgassing procedure. Usually the problem can be the oposite when the equipment performs a dead space calculation prior to the experiment with helium which then helps to clean out the pores better than the original out gassing.

Best regards,

Malcolm

maybe, the N2 liquid (or any cooling fluid you use) itś evaporating before finishing your analysis that could increase the temperature and in consequence rise the rate of desorption. (I Know that sounds a little unlikely but you never know).

It is necessary to check the reproducibility of the desorption. Does the ratio of adsorption and desorption peaks changed from one experiment to other. The absent of  reproducibility means that part of gase didn't desorb from microporous. It is realy possible in the case of coal sorbents

Correction of your machine you can control by testing adsorbents without microporous (like silica) 

If we consider that there were only physical adsorption, one would expect identical curves for adsorption and desorption. However, if a small fraction of the particles (or gases) suffer chemical adsorption or imprisonment on infinite "cages" common in carbon-based materials with multi-channel systems it is possible that the desorption curve is smaller than that of adsorption

In case of including pores of various sizes of porous materials, adsorption and desorption isotherms do not follow the same path. Because, adsorption and desorption starts from narrow pores and large pores, respectively.

If we consider that there were only physical adsorption, one would expect identical curves for adsorption and desorption. However, if a small fraction of the particles (or gases) suffer chemical adsorption or imprisonment on infinite "cages" common in carbon-based materials with multi-channel systems it is possible that the desorption curve is smaller than that of adsorption

Dr. R. Júnior,

Interesting perspective. I suppose it is indeed possible that an irreversible reaction could be taking place. I hadn't thought of that. But this would mean that the isotherm is vertical at P=0, and horizontal for P>0. I'm not sure if this can be considered adsorption. Regardless, the isotherm shape should allow an instant assessment of whether there is an irreversible chemical reaction taking place.

Rui Afonso

I have kind of similar pattern of isotherm linear plot (desorption gets lower of adsorption curve, it crosses over at the P/P0 0.84). I repeated the measurement using N2 adsorption technique. The material is clay. I have several heat treated materials and the extreme one gets this curve. Although I did twice to make sure operational error is not occurring, I would do again. But, I guess something like Renato W. Junior said. Chemisorption might be the case. 

OK thanks Dr Trunfio. I use Gemini V (Micromeritics) using N2 and He gas at 77K. So far the instrument is cable to measure >0.1 m2 surface area. the strange desorption one got very low BET surface area (2.89 m2/g) but the sample amount I took satisfies it to be 0.54 m2, which is >0.1 m2 limit. The degassing was performed at 70 deg C overnight (~12 h). I am also thinking to take higher amount of sample for another repetition. Please also consider that, the structure of this material has been collapsed after extreme heating (this is my treatment), which is ~60% collapse in SSA. Thanks.

Your outgassing procedure may be the problem or just the limitations of your instrument. Usually, the desorption curve can be higher than the adsorption one due to activated entry or very slow adsorption at very low relative pressures if the sample is microporous. However, with less than 3 m2/g we can rule out microporosity. Thus, insufficient outgassing leads to a slightly dirty sample that with the Helium dead volume calibration can get cleaner, but this would only lead to you underestimating the area and would not lead to the desorption branch crossing below the adsorption one. I think that the best solution is to use much more sample as suggested by Giuseppe Trunfio.

On the point of the importance of outgassing, I once had a colleague who was working with a series of non-microporous aluminas. On outgassing at 140°C overnight the areas were about 200m2/g but did not follow any logical order. Since the samples had been heat treated at 500°C I decided to remeasure the surface areas after outgassing at 300°C and achieved consistent results. His problem was the very strong adsorption of moisture from the air which outgassing at 140°C was not capable of removing even though the samples were not microporous.

In any vacuum-volumetric (manometric) sorption analyzer that gives desorption data below adsorption we usually start by considering instrument-related problems (leaks). Now in this case, the adsorbed amount is precious little (I'm assuming that the isotherm should be type II - no pores), so the desorption data appearing below the adsorption data might simply be due to experimental uncertainties and nothing to worry about.

Nevertheless, if one is concerned that the error is greater than anticipated, it is easy to run an empty cell (use 1g as the apparent sample mass... the effective adsorbed/desorbed amounts then represent absolute volumes of gas). This will evaluate instrument performance in the absence of sample... and can be expected to be present in sample runs too (even when not directly observed because it has been overwhelmed by sample uptake)!

Coolant level control is an important factor in these instruments (to maintain a constant cold free space/void volume) and if the particular method that is being used by the instrument is not functional, then desorption data will tend to cross under the adsorption data, but with decreasing difference as the P/Po values approach zero; this type of free space/void volume error (time dependent) is a function of pressure in the cell... least error at low P/Po, greatest at high P/Po. However, it is extremely unusual in the Gemini instrument to see this because of its "balance tube" approach: a constant coolant level is not required, since the evaporation takes place around both tubes (sample cell and empty balance tube) and the adsorbed amount is effectively the difference between dosed amounts into both tubes. However this is only true if the tube stems are identical... same i.d. and both using filler rod, or both not using filler rod. Filler rods are always recommended during analysis unless they exacerbate elutriation of fine powders (and when no other precaution against that is viable because of some peculiar cell geometry).

As Fernando pointed out it is uncommon in carbons, but it could be a hysteresis loop.

See e.g. http://www.microtrac-bel.com/en/tech/bel/seminar16.html

am recommending rapid heating method, for example, after removing the glass with liquid nitrogen, the sample is dipped into the water 70-80gr C. The desorption peak is a more accurate and close to the adsorption

Hi! I also have a follow up question about this. . what if the desorption curve is higher than the adsorption curve? does this mean that the gas was not able to escpae from the material along the desorption process? Thank you!