Maybe, your sample was polluted, such as have adsorbed some CO2 before the test. Suggest you re-do the test.

In my case I am also facing same thing. I would not say It is good or bad. Still I am unable to say reason,

Not normal, but not rare. Either your sample was not degassed enough before starting the test, or you have a small leak. Or both ! Look at your desorption branch: it goes through a maximum, which is physically meaningless. Therefore, you have an experimental problem. A leak makes the pressure increase linearly with time in the system. A degassing in a closed system makes the pressure increase slower and slower until saturation. You may measure the changes of pressure with time and deduce what is your problem.

Good luck

Alain

 Alain might be right,

but this could also be a case of diffusion controlled hysteresis (DCH). That would produce exactly an isotherm with a max in desorption like yours (and that max would then also have a physical meaning). This happens when diffusion is too slow for (part of) the absorptive to reach the sites where it would sorb. Upon desorption, part of the adsorbent isn't saturated at the respective relative pressure, so it sorbs more instead of desorbing, which results in the observed maximum. The reasons for DCH are usually a complex micropore (or lower meso) network (or particles too large) or some specific interaction of the adsorptive with the adsorbent is slowing down diffusion (I often see that for CO2 on clays).

Diffusion controlled hysteresis is explained very well in this paper:

https://www.researchgate.net/publication/226385955_Diffusion-Controlled_Hysteresis

The test is to redo the measurement with substantially increased equilibration time. In case of DCH, the hysteresis loop will become narrower or close.

I'm not that sure about the effect of a leak, maybe Alain can comment on this? In case of a leak, the pressure in the cell would be higher than expected, which for desorption would mean that the apparent desorbed amount is larger, i.e. that the amount of adsorbate still in/on the sample is underestimated, so the desorption branch would be lower than expected instead of higher as in this case?

Pieter

Article Diffusion-Controlled Hysteresis

Thank you all for your help. I decided to check the degas and analysis ports, and I noticed the filters and O-rings needed to be changed and I did. The degas ports filters had a lot of residue which I think contaminated my sample. Pieter, thanks for the recommended article.

I agree with Pieter-common in zeolitestudies

In my book: R. Roque-Malherbe, Adsorption and Diffusion of Gases in Nanoporous Materials, CRC Press-Taylor & Francis, Boca Raton, FL, USA, 2007. Capillary condensation is described as follows:

In practice, hysteresis is observed in materials consisting of slitlike pores, cylindrical-like pores, and spherical pores (i.e., ink-bottle pores). The adsorption mechanism in mesopores, at low relative pressures, is similar to the adsorption process on planar surfaces; however, after the completion of the monolayer formation, multilayer adsorption commences, capillary condensation occurs, fundamentally in the core of the pore. During adsorption when the pore is completely filled with liquid; the pore liquid is separated from the bulk gas phase by a hemispherical meniscus, hence pore evaporation takes place by a thinning meniscus at a pressure which is less than the pore condensation pressure; as described by the Kelvin’s equation; that is, a fluid confined in a pore condenses at a pressure lower than the saturation pressure at a given temperature. However, during desorption the pore liquid is separated from the bulk gas phase by a cylindrical meniscus, and pore evaporation takes place by a thinning meniscus at a pressure which is less than the pore condensation pressure; as described by the Kelvin-Cohan’s equation, but different in comparison to the adsorption branch.

The Kelvin-Cohan or modified Kelvin equation is derived from the Kelvin equation, which is based on the effect of the curvature of the surface on vapor pressure. This effect of curvature is best understood in terms of the pressure drop, across an interface which is described by the well-known Young-Laplace equation. The Kelvin-Cohan equation gives a correlation between the pore diameter and the pore condensation pressure. It predicts that pore condensation shifts to a higher relative pressure with increasing pore diameter and temperature.

Are you suggesting that Saki's case here can be explained by capillary condensation, Rolando?

It's not explicitly specified, but this seems to be a CO2 isotherm at 273K, which would mean that the max. p/p0 is about 0.04. Isn't that considerably below the onset of capillary condensation?

Does your book also describe the applicability of capillary condensation models as function of relative pressure for different sorptives at different temperatures? I've been looking for good literature on that. What I read is mostly that the entire concept is not applicable below about 4 nm. For N2 that is clear, from the forced closure, the tensile strength effect (which according to Thommes is not tensile strength related). But for other adsorptives like CO2 at 273K or water, I didn't find a lot.

regards,

Pieter

This happens if you have a free volume error and if your instrument sensitivity is very low (the pressure sensor). Also check the mode of the instrument, if you are using RTD and if your instrument is in other mode, then you will see such weird desorption curves.  If you are using a RTD probe, you should use right RTD probe, water for CO2 and the other one for N2. Do not use N2 probe for CO2 as both of these fluids have different density. For CO2 at 273 K and 298 K, you should use 'water RTD (float)'.

Try to clarify all the above points, probably you will get decent results. 

As similar for water vapour sorption process, a hypermolecular character of pore matter is important, i.e.  the amount of pores, their size and shape distribution but predominantly adhesion tension at pore-wall interfaces and  its distribution as well as. If absorption - desorption process is the condensation - evaporation process then the all process is controlled by Kelvin´s equation (see R. Roque-Malherbe´s answer). A simplest model as to explain sorption hysteresis is then introduced by conical or stepwise shape the cylindrical pore.

I would support what Pieter says: "The test is to redo the measurement with substantially increased equilibration time."

I think that the system not reaching equilibrium is the most likely explanation. If the instrument is moving on too early at each isotherm step, then the sample will still be adsorbing CO2 when the uptake is calculated for a particular "equilibrium" pressure. This will cause problems and could lead to apparent hysteresis in the data.

Also, as Pieter implied later, the pore size is important so it does depend on the material type; but capillary condensation as a cause seems highly unlikely. Other reasons for observing hysteresis in adsorption/desorption isotherms, apart from some of the experimental problems identified above, include sample swelling and/or structural phase transitions, which can occur in some metal-organic frameworks (MOFs). But all this, again, depends on the material type.

Regards,

Darren

May I give my suggestion? Would it be possible to make a repeated sorption-desorption cycle on the same sample or the sorption-desorption-sorption cycle? this could test whether the maximum is not an artifact of the frst measurement and possibly also to test the extent of attainment of equilibrium. I might miss but it si not clear what is the material and whether it was properly dried. Trace water may involve a complex reponse to CO2 sorption,,

 For the chars prepared from bamboo, this phenomenon is frequently observed by us. However, after the chars were activated by steam to  develop pores. This phenomenon disappeared. I think that Pieter's explanation by diffusion controlled hysteresis maybe be reasonable. So I want to read the original paper about DCH.