Hi Srujan,

Could you please give more info about your sample? (ie. micro-meso-macroporous materials), and if you could upload your N2 isotherm to here, I can help you.

Regards,

Hakan

These problems have been already addressed in detail in previous threads, you should have a look there. You may have a problem of insufficient degassing, or a leak in your apparatus. For carbon materials, you can heat up to 200°C, and even up to 250°C most of times, during at least 12H in secondary vacuum. For organic materials, you should limit the temperature to 80°C or even to 60°C, depending on the cases, but in this case the degassing should be done in excellent vacuum for at least one full day.

If the problem continues, it should be a leak in the device. In the case of leak, the pressure continuously increases, almost linearly. In the case of a degassing from the sample, the increase of pressure continuously slows down until a plateau is reached. So you can separate these 2 phenomena.

Finally, a negative C constant may also come from a badly chosen pressure range for doing the calculation, even in the absence of degassing or leak. I agree that, whatever you do, your apparatus will always provide a value of BET surface area, so bettter being careful.

Alain

Hi again,

Most of the mesoporous materials (i.e. MCM41, MCM48, SBA15, etc.) for degassing; I have been using 200°C during 12hours under vacuum and this method pretty works.

Like Alain says, there should be a leak in the equipment, check the vacuum pressure.

Regards.

My sample is carbonaceous material derived by carbonization of resorcinol formaldehyde polymer. Is working above 200 °C a safe option? Also should I consider going for a 10 point or 20 point BET rather than 40 point BET to achieve a better result in my case?

I have attached a picture of the curve. (Does this curve indicate any leakage in the device or I just need to use a suitable degassing temperature? )

I am agree with the answer of Prof. Dr. Alain.

Carbon from RF resin can be heated up to 250°C in vacuum without problem. We do this on a daily basis. Give it a try, and if this happens again, then for sure there is a leak.

Alain

carbonaceous material generally give negative value of 'c'. when you use BET equation to estimate the specific surface areas. Normally, this is due to not considering the free space. Therefore, you should (1) use a filler rod during N2 sorption measurement; (2) estimate the surface area by Langmuir adsorption isotherm instead of BET.

To establish the absence of instrument artifacts (leaks, etc.) you may consider running an experiment using an empty sample cell (and filler rod). Without sample you should get negligible gas uptakes in both adsorption and desorption isotherm branches. You should also consider running a standard reference material to confirm that the instrument is performing as specified. If those tests are successful and re-running your carbonaceous material gives repeatable results, consider extending the equilibration time for each point, as noted above. Your raw data suggests that, barring sample swelling, you may be dealing with small nanopores (or nanopore openings) that require long equilibration times to be filled and emptied when using nitrogen at 77 K. If you still do not achieve isotherm closure after extended equilibration times at 77 K, you may also consider using argon at liquid argon temperatures (87 K) or CO2 at 273 K for your experiments. 

Does your device has a procedure to perform a leak test? This is usually the case of commercial devices. This would confirm the problem of leak if there is one.

You can also check manually by connecting an empty cell on your machine and you put it under vacuum. When the ultimate vacuum of your pump is reached, you isolate the measuring cell by closing the valve just past it and you control the pressure measured in the cell: it should remain roughly stable or rising very slowly.

To ensure proper degassing, you can heat your samples up to 250 °C (as indicated by Professor Celzard) under secondary vacuum until the ultimate vacuum of your pump is reached. This ensures that the degassing is completed. This can take 4 hours or more, depending on your samples.

Make sure that the equilibrium time is sufficient for measurements. This is even more true for microporous samples whose first points may each need equilibrium times of several hours. Experiment with different equilibrium times: if isotherms are then superimposed, the equilibrium time is sufficient.

I'm not familiar with your samples or your device, but I think it might be necessary to begin the measures at a lower value of p/p0, eg 0.0001 (or lower if your machine allows).

If your sample is microporous, most of the information will be located before p/p0 = 0.05.

To apply the BET law for the determination of the specific area, be sure to observe the following criteria:

- The value of C must always be > 0 ,

- A linear fit to the BET transformed data should be obtained (R > 0,995),

- The value of  nads(1−p/po) should be increasing with p/po for the data selected to calculate the BET parameters,

- The monolayer capacity nm should be within the limits of data that were used to fit the BET parameters,

- The value of 1/(√C+1) ≈ p/po at the monolayer capacity.

If one of these criteria is not met, start again by changing the selection of points used to calculate the surface area.

Dear Mr. Srujan Singh,

Estimation of surface area accessible for adsorption and chemisorption (in the case of a chemical reaction) of some chemical individuals is much more complicated than it could be seen at first glance. A few comments and maybe recommendations (very similar to those written by other contributors):

1. The BET analysis is a slightly misleading term in the general sense. The BET isotherm was derived for precisely defined conditions (multilayer adsorption, no changes in the adsorption capability during and after covering neighbourhood sites with adsorbate, the adsorption heat for the first layer differs from the second and the next ones, etc. – see: Brunauer, S., Emmett, P. H., & Teller, E. (1938). Adsorption of gases in multimolecular layers. Journal of the American Chemical Society, 60(2), 309-319). It is better to use “Analysis of textural properties”.

2. If adsorption obeys “the BET rules” C constant cannot be negative. The negative value usually indicates presence of microspores, which are filled by capillary condensation and afterwards the BET multilayer adsorption starts. In this case you need to use a combination of approaches, see for example: (a) Schneider, P. (1995). Adsorption isotherms of microporous-mesoporous solids revisited. Applied Catalysis A: General, 129, 157–165. DOI: 10.1016/0926-860x(95)00110-7. (b) Horniakova, J., Králik, M., Kaszonyi, A., & Mravec, D. (2001). A practical approach to the treatment of adsorption–desorption isotherms, acidity and catalytic behaviour of zeolite catalysts. Microporous and mesoporous materials, 46(2), 287-298, or (c) Králik, M. (2014). Adsorption, chemisorption, and catalysis. Chemical Papers, 68(12), 1625-1638.

3. For estimation of isotherm parameters, a non-linear treatment of data is much better than utilization of lineralized forms (see above cited papers).

4. It is necessary to consider for what purpose you plan to use the investigated material (4 possibilities): (a) adsorption, or (b) catalysis, the (ac, bc) gas phase, or (ad, bd) liquid phase. For catalytic purposes a monolayer adsorption is important and reactants have to access catalytic sites in (micro-) pores. For more information about textural properties of catalytic materials I advice the paper: Lecloux, A. J. (1981). Texture of catalysts. In J. R. Anderson, & M. Boudart (Eds.), Catalysis: Science and technology (Vol. 2, pp. 171–230). Berlin, Germany: Springer.

5. Outgassing has to clean the surface and not to damage the structure of material, which can appear with increase in temperature and decrease in pressure. For example common organic (e.g. resin based) materials do not allow more than than 150 oC, inorganic materials, e.g. thermally stable zeolites can be treated even at temperatures higher than 500oC. To avoid structure ruptures due to decrease in pressure at increased temperature an inert gas, e.g. helium is often applied.

6. Take into account differences between adsorption in a gas phase environment (e.g. nitrogen, Kr as adsorptives, condensation in micropores, multilayer adsorption, ...) and liquid environment (monolayer adsorption, ion exchange, etc. ..). In the liquid phase adsorption a possible changes of material by interaction with solvent (in extreme case swelling) can play important role.

7. If you have a special new material, it is necessary to check the rate of increasing temperature and decreasing the pressure, the rate and time for outgassing. For a proper outgassed material adsorption- desorption isotherms cannot differ when time for outgassing is prolonged.

8. If your desorption isotherm does not meet the adsorption one, either the material was not properly outgassed, or material changed during measurements. Pay also attention to time for equilibrating of amount of adsorptive at a given pressure.

I wish you many successes in your research,

Yours sincerely,

Milan Kralik.

Milan, I cannot agree with point 8. This phenomenon is usual for mesoporous materials.

Dear Srujan:

Please check the paper shown below to see if there is anything helpful?

Langmuir 29(42): 12982-12989.

Carlos gives excellent answer to solve the negative c problem.

As to the degassing temperature, I suggest you run a TG first, and then choose the highest temperature that DO NOT decompose your sample as the degassing temperature.

Ideally for new sample take one TGA. Temperature of plateau region of TGA curve can be taken as Degassing temperature. regarding time, longer the better but it should not be too long. it is trade off between the accuracy you get and the time you spend on analysis. four to Six hours is more than sufficient. 

I agree with XIN Zhou 100%. The most important thing you should not destroy the sample structure. As far as you know there is moisture water and structure water. Regarding the BET we are interested to get rid of the moisture content only. Therefore, you have to know the material very well before BET test. I have to if the material has organic content or gypsum or clay or silicate. Do not hesitate to ask if you have a sample and you do not know at which Temp. you should treat it.

Do not worry about the negative C value, it is common to microporous material due to the nature of the BET equation is not appropriate.

The problem is clearly experimental: You can see that desorption branch amount is increasing when pressure decreases, this can only happen due to leaks and is extremely unlikely to happen due to narrow pores (desorption is orders of magnitude faster than adsorption).

I recommend:

• Fix the equipment and after test a reference sample (SiO2, alumina or similar; it should be included with your supplies from equipment manufacturer)
• Are you using enough sample? I can see some 'steps' in the isotherm which may occur to very low sample amount (small adsorbed amount divided in small sample amount gives steps in cc/g)
• Start the measurement at lower relative pressures, there is no data between 0-0.05 P/P0. Then you can check that the desorption branch effectively goes to zero.

Dear all,

First of all thanks for all your responses. I have uploaded a pic which shows 4 different isotherms of my samples numbered 1,2,3 and 4. In all the 4 samples my matrix(sample) is the same but only the loading of a certain element let's say X is changed. The loading being the least in 1 and the maximum being in 4.

2,3 and 4 show similar trend in isotherm, then why 1 is so different? Well this is a previous run of the samples(before I uploaded my question), based on many of your recommendation I have used a higher degassing temperature of 2500C (previously I was using 1700C) for 12 hours(increased the degassing time too). I have obtained better results i.e the two branches are meeting now (forming a closed loop) but the c still comes negative.

Basically I am more interested in finding out the pore size distribution for my sample rather than the surface area. So I wanted to know how much of error can I expect in my analysis if c is negative? Also all of you are suggesting my sample to be microporous but aren't the isotherms belonging to Type 4 ( I mean why the step in the desorption isotherm) suggesting mesoporous nature?

Plot 1 is completely wrong due to a huge leak.

Other plots typical of micro-meso-macroporous  material with some leak still present.

Some test with Hg porosimeter recommended.

Dear Srujan,

   what is the expected surface area of your materia? if it lies below 300 sq.mper gram, i recommend to use more amount of sample so that you adsorption and desorption curve will meet. the reason is while measuring the desorption curve there may not be enough N2 that is being desorbed to be detected by the pressure transducer. and this may also inprove the c-value.

Dear Talla

I expect the surface area to be more than 300 m2/g as it is porous carbonaceous material but till now none of my BET runs have given me area above this value, so I wont know for sure because the isotherms have been erratic so far. 

If one more try with modifications in degassing temp. and time and time given for equilibration should would give some consistent value, I can be sure. But anyways how much amount do you think is appropriate for a proper run, 50 mg should do or more?

And also is it the weight of the sample or the volume (filling up a decent volume of the bulb in case of low density samples) that matters?

Could you give us more information about the device used for the measures?

The amount of sample needed depends on its specific surface and sensitivity of the pressure sensors of the apparatus used. For example, if the device is capable of measuring a minimum total area of 10 m² (data provided by the manufacturer), the minimum amount necessary for an estimated sample of 300 m²/g will be 33 mg (300 x 0.033 = 10m²).

To maximize the sensitivity of the device, it is also necessary to use glass inserts placed inside the sample tubes to reduce the free volume.

Dear Srujan,

if the surface area is below 300 then i recomend to take an sample of weight not less thatn 100 mg. also an degassing time of atleast 10 h will be good

Hi all,

is there any publication that study the effect of degassing temperature on the isotherm curve and BET data?

Thank you

What would be the best degassing temperature for BET analysis?

There are (at least) three more threads about degassing temperature for adsorption:

https://www.researchgate.net/post/what_is_the_influence_of_degassing_process_on_BET_Surface_area_measurement

https://www.researchgate.net/post/Can_anyone_explain_how_to_determine_degassing_temperature_for_the_preparation_of_a_material_for_BET_analysis

https://www.researchgate.net/post/How_can_I_decide_the_degassing_temperature_and_the_duration_for_degassing_process_for_doing_the_BET_analysis_of_a_sample

seems to me that your dewar was empty. Check the p° values for the points where it started going down.

could be a leak or something like that. Always check reproducibility before wasting time on interpreting an unusual measurement. If it isn't reproducible, measure a standard, that usually shows what went wrong.

The negative value of C-constant of BET-isotherm is not results of not-proper degasig. If you use the standard range of BET-linearization (P/P0=0.05-0.30, or even 0.05-0.20), the negative value of C-constat is indication of the presence of micropores, in your case zeolite structure. For ANY microporous solid you obtain by this standard BET-method negative C-constatt, and at the same time this fact indicates that the value of S(BET) has no physical meaning. In such case for the evaluation of texture you must treat the adsorption data by t-plot (or alpha-plot) method. I attach my paper in Stud.Surf.Sci.Catal. 142, 2002, 1587-1594 dealing with this problem.

The answer of Pavel is quite to the point, you have to adjust the rang of pressures to which you apply the BET eq.For Microporous surfaces the classical range (0.05 to 0.3) is not adequate but should be reduced at the high end.

Note that a negative C is physically irrealistic, so your linearized BET equation should for microporous surface intersect the Y-axis very close to zero (at the positive side). Skip the high pressure values that cause your linearized BET line to negative intercepts.

Luuk, I would like to comment your recommendation to decrease the higher-p/p0 points of adsorption isotherm to receive possitive C-constant of BET-isotherm. I know that in many papers such solution is used, but in my opinion, that can not result in proper specific surface area. BET-isotherm was developed under several conditions that are not fulfilled for microporous materials, mainly multilayer adsorption, that can not proceed im micropores. So we can to modify the range for BET-linearization, but it is still BET-isotherm that is not valid for microporous materials, and the resulted value od S(BET) has no physical meaning in any case. The most used method for zeolites, active carbon and other microporous solids is t-plot method, that distingiush between external surface area together with area of mesopores (if present) - that is real surface area in m2/g, and volume of micropores in cm3/g, both values very near to reality.

I'm sorry Pavol, but I think your comment and paper are outdated.

Nobody these days would use a 'standard' range for the BET fit of a microporous material, I think, since Rouquerol plots have been implemented in most of the common softwares. The paper explaining this and also the physical meaning of the BET area dates from 2007: Article Is the BET Equation Applicable to Microporous Adsorbents?

Also, there are better methods than t-plot nowadays to determine the area of a zeolite.

Pieter, I know very well the todays methods of BET-modifying for reaching the possitive value of C-constant, in false supposal that in such case the result - specific surface area obtained by modified range of BET is meaningful. My opinion is that for microporous materials is BET isotherm not suitable in any modified form! Such value has not any physical meaning, micropores have not the surface that could be evaluated by monolayer. Micropores can be evaluated only by t-plot or alpha-plot methods (from adsorption data obtained from standard equipments), or by DFT using new equipments with lower transducer measuring the micropores and new software.

So my answer to Srujan Singh was to explain, that negative value of C-constant in BET for microporous materials is nothing exceptional, that is only indication of micropore presence, and results of BET-isotherm could be omitted, and to evaluate textute by other suitable method.

Pavol and Pieter, one should realize that the N2-BET surface area is a standard method that gives a relative value to the specific surface area (SSA) as it is based on a model and assumptions regarding the N2 way of adsorption. For microporous surfaces a high C value applies, so the linearized BET eq. can be drawn through the origin and followed up to deviations from linearity to obtain the N2-BET-SSA of the sample. An applicant should know that the method gives a relative area, but it is still a good way to compare the SSA of materials of the same chemistry. In the case of microporous surfaces it is just a number that is related to the pore volume and not to the pore area but as long as it is compared between microporous surfaces of the same nature it still is usful information.

Of course it is better to use for microporous surfaces a more dedicated model, but such models are only available for specific materials with specific pore geometries.

Dear Luuk and Pieter,

I started this discussion because I use nitrogen adsorption for evaluation of texture properties of different micro- and meso-porous materials and their mixtures for about 35 years. And thanks to ASAP 2400 equipment (Micromeritics) that is capable to measure 6 samples at the same time (it means about 30 samples weekly) I was able to measure planty of samples - mixtures of pure mesoporous material (alumina) and pure zeolite (NaY, Na-mordenite, Na-ZSM-5) from 0 to 100 wt.% with the aim to study the sensitivity of micropore volume on the value of C-constant in BET-isotherm. As you can see from Fig.3 in attached conference paper (Stud.Surf.Sci.Catal,174, 2008,981-984) the value of C-constant increases at first to very high positive values, and at the micropore volume about 0.017-0.020 cm3/g immediately changes to very high negative values, goint to value negative about C=-41 for pure zeolite. All measurings and all evaluation were realised at the same conditions (13 adsorption points in standard BET-range 0.05-0.30, and linearization always in the same range). In such case, the value of C-constant in standard evaluated BET-isotherm is for me also certain quantitative indication of volume of micropores, mainly in the range up to about 20-30% of zeolite in mixture with mesoporous matrix.

In my papers, for micropore-containing materials I always give also the specific surface area by standard BET-isotherm also with C-constant value (even if it has no physical meaning, but for comparison with other measurements), and at the same as main texture characterization by t-plot.

Thank you for your participation in this discussion, I hope that it will be helpful mainly to the initiator of this discussion - Srujan Singh

Dear Pavol and Luuk,

an interesting discussion indeed.

There are a couple of things I'd like to add, and hear your opinion on:

• First of all, the original question by Srujan Singh is 4 years old, and wasn't about zeolites as far as I understood. He mentioned low pressure hysteresis, which was an equilibration issue. The zeolite isotherms were posted by Nor Naimah Rosyadah Ahmad who hijacked the thread a couple of days ago, with something completely different, probably experimental problems, not related to negative C-constants. But that shouldn't keep us from having an interesting discussion here I think.
• Concerning the physical meaning of the BET area. I'm also not fond at all of modified BET models, Pavol. But, in the paper I linked above, Rouquerol and co show, by comparison to microcalorimetry, that when the (original!) BET regression is applied from the lowest p/p° to the top of the n_a (p° - p) vs. p/p° plot, then 'the apparent BET monolayer content exactly corresponds to the most strongly adsorbed portion of the adsorbate'. One of the samples used is a zeolite. From this they go on to conclude (I'm just copying): "Everybody agrees on the fact that the sole concept of “BET monolayer content” is inadequate in the presence of micropores where a monolayer has no clear physical or theoretical meaning. Now, if we do not want to lose, in the case of microporous materials, the benefits of the BET equation, we need to introduce another concept. This is fortunately offered by the calorimetric experiments which suggest that the “BET monolayer content” physically corresponds to an energetically strong retention. This quantity, provided by the BET equation, could therefore be called safely the “BET strong retention capacity”. This quantity includes two parts, which are the “micropore capacity” and the “monolayer content”on the non-microporous portions of the surface. The latter, which provides the “external (i.e. non-microporous) surface area” is easily assessed by means of the αS or t methods, without even requesting the very low part of the adsorption isotherm.
• So from Rouquerol's work, I understood that the BET method is not merely an industry standard, giving a relative area, but that it does have a physical meaning, i.e. it corresponds to the most strongly retained fraction of the sorbate. When the regression is done in the suggested way, this property is accurate and reproducible for any material, which means it can be used for comparing different materials. Assigning a surface area to that BET strong retention capacity might be somewhat arbitrary, but that arbitrariness is in every method, and it doesn't affect the use as a 'relative' property. I guess that's basically what Luuk meant?
• Personally I'd use DFT or one of the related methods for zeolites. I'm in business for not even 20 years, but I don't consider them new anymore, those methods. They are implemented in all popular softwares and easy to obtain/use (so easy in fact that they are very often used without much consideration). The kernels are available for zeolites, because they have regular and know geometries. Also, I don't often get the chance to use DFT, since I mostly work on disordered micro/meso/macro porous materials.
• I'm not sure anything measured with N2 at 77K on a microporous zeolite is actually very physically relevant. Some years ago Dr D. Klank showed me the attached figure. What they did there is charge a zeolite (4Å) with N2 for 24h at 760 Torr at different temperatures. Between 100K and 150K, the adsorbed amount rises from 0.6 to 110 cc/g. This is of course an extreme example, but I do think that kinetic effects are seriously under-appreciated by the community. There are a lot of papers out comparing N2 at 77K with Ar at 87K by now. Personally I'd also measure zeolites with Ar, or CO2 if they aren't mesoporous.

regards,

Pieter

Dear Pieter and Pavol,

The Rouquerol work has indeed shaped my opinion and is still the best way to understand ot the benefits and drawbacks of the BET SSA.

With respect to the use of the standard isotherm in the t or alfa-s method for the discrimination between micro- and meso-pore contribution we should also realize that the standard isotherm used affects the results and that the standard is based on the idea that the multilayer formation of N2 is purely a nitrogen property independent of the underlying surface. Therefore, also the obtained distinction in micro and meso pores is 'relative' and not necessarily true.

DFT is better, but only available for a given surface chemistry and geometry, so we probably will in many cases need the less accurate (relative) methods to obtain information.

Pavol, your discussion of the C constant is mainly useful when one studies specfic materials with small changes in structure. The general implications are very limited. Negative C values should therefore be excluded when a BET SSA is calculated. When we use the standard pressure range to obtain the BET SSA and obtain a negative C this SSA value should be rejected. Instead only rel. pressure points should be used that provide a positive C. The linearized BET should run through the origin and the linearized BET model is only valid to obtain the monolayer capacity (and 'relative' "SSA") for the pressure range that the points don't deviate from the straight line.

I m getting surface area in -ve value.

please give me answer why it is so.

There can be many reasons for that. It is hard to say something useful without seeing your data. So please share your measurement (isotherm plot).

That is definitely not a measurement on which you should attempt a BET fit.

I see 3 potential causes, in order of likelihood:

• Leaking sample cell: remeasure, maybe replace the o-ring
• Sample not properly outgassed: outgas longer and at higher T if possible and remeasure
• Something else wrong with the instrument: measure a standard or blank, contact Micromeritics service

Also the mass of sample you use is low. That would be ok if you expect a very large sorption capacity, high surface area. If not, increase sample mass

regards,

Pieter

I agree with Pieter, there is something wrong with the instrument. The adsorbed volume can not to decrease with increasing relative pressure. Also saturation pressure about 820 torr seems to me unreally high.

After seeing your file, i think that the analysis is done by an operator, who is well equipped with the instrument, so the chance of leaks is vey less.

is it that only your sample is showing such behaviour? or all submitted samples are showing negative values?

if all are showing -Ve then definetly it is the problem with instrument. if not

can you provide the following details

what is the material type?

what is the expected surface area?

did you run a TGA of the sample?, if yes at what temperature you have observed any weight loss?

or is it a calcined sample? if so what is the calcination temperature

the first two question can give me an idea of why you are getting the negative value and the 3rd and 4th questions will help in deciding the outgassing temperature and type