3 Micro and Mesopore Size Determination by Gas Sorption First: Quantitative estimation of micropore volume and area…T-plot and DR methods.
4 Multilayer adsorption Type II, IVLow slope region in middle of isotherm indicates first few multilayers, on external surface including meso and macropores… before the onset of capillary condensationVolume adsorbedAfter the knee, micropores cease to contribute to the adsorption process.Relative Pressure (P/Po)
5 Estimation of Micropores... the t-plot method This method uses a mathematical representation of multi-layer adsorption. The thickness, t, of an adsorbate layer increases with increasing pressure. The t-curve so produced is very similar in appearance to a type II isotherm. For every value of P/Po, the volume adsorbed is plotted against the corresponding value of “t”.If the model describes the experimental data a straight line is produced on the t-plot...
6 The t-plot Resembles a type II Statistical thickness A statistical multilayerA statistical monolayerRelative Pressure (P/Po)
8 t-plot Method showing a “knee” CBSlope A - slope B = area contribution by micropores size C
9 ?QuizWhat is an s plot?s (for Ken Sing) is a comparison plot like the t-plot but its slope does not give area directly.A
10 Estimation of Micropores Dubinin-Radushkevich (DR) Theory W = volume of the liquid adsorbateW0 = total volume of the microporesB = adsorbent constant = adsorbate constantA linear relationship should be found between log(W) and log2(Po/P)...
12 Pore Size Determination Requires a recognition and understanding of different basic isotherm types.
13 t-plot Method (in the presence of micropores) Intercept = micropore volume
14 Types of Isotherms Type II Type I Volume adsorbed Type V Type IV Type IIIType IVRelative Pressure (P/Po)Volume adsorbedType V
15 Types of Isotherms Volume adsorbed Type I or pseudo-“Langmuir” Relative Pressure (P/Po)Volume adsorbedSteep initial region due to very strong adsorption, for example in micropores.Limiting value (plateau) due to filled pores and essentially zero external area.
16 Quiz Why pseudo Langmuir? A Langmuir applies to monolayer limit, not volume filling limit.A
17 Types of Isotherms Type II Volume adsorbed Relative Pressure (P/Po) Rounded knee indicates approximate location of monolayer formation.Absence of hysteresis indicates adsorption on and desorption from a non-porous surface..Low slope region in middle of isotherm indicates first few multilayers
18 Types of Isotherms Type III Volume adsorbed Relative Pressure (P/Po) Lack of knee represents extremely weak adsorbate-adsorbent interactionBET is not applicableExample: krypton on polymethylmethacrylate
19 Types of Isotherms Type IV Volume adsorbed Relative Pressure (P/Po) Rounded knee indicates approximate location of monolayer formation.Low slope region in middle of isotherm indicates first few multilayersHysteresis indicates capillary condensation in meso and macropores.Closure at P/Po~0.4 indicates presence of small mesopores (hysteresis would stay open longer but for the tensile-strength-failure of the nitrogen meniscus.
20 Types of Isotherms Type V Volume adsorbed Relative Pressure (P/Po) Example: water on carbon blackType VVolume adsorbedLack of knee represents extremely weak adsorbate-adsorbent interactionBET is not applicableRelative Pressure (P/Po)
21 Types of Hysteresis Large pores/voids Gel Volume adsorbed Mesopores MCMRelative Pressure (P/Po)
23 Analyzer measures volume of pores: Yes or No? QuizAnalyzer measures volume of pores: Yes or No?NO! It measures what leaves supernatent gas phaseA
24 Pore Size Distribution Hysteresis is indicative of the presence of mesopores and the pore size distribution can be calculated from the sorption isotherm.Whilst it is possible to do so from the adsorption branch, it is more normal to do so from the desorption branch...Mesopore (Greek meso = middle): 2nm - 50 nm diameterMacropore (Greek macro = large): >50 nm diameterMicropore (Greek micro = small): 0 nm - 2 nm diameter
25 Adsorption / Desorption multilayer formationDesorption =meniscus development
26 Kelvin* Equation* Lord Kelvin a.k.a. W.T. Thomson
27 Pore Size rp = actual radius of the pore rk = Kelvin radius of the poret = thickness of the adsorbed film
28 Statistical Thickness, t Halsey equationGeneralized HalseydeBoer equationCarbon Black STSA
32 Pore Size DataVolume and size of pores can be expressed from either adsorption and/or desorption data.The total pore volume, V, is taken from the maximum amount of gas adsorbed at the “top” of the isotherm and conversion of gas volume into liquid volume.The mean pore diameter is calculated from simple cylindrical geometry:where A is the BET surface area.
33 Pore size analysis of MCM 41 (Templated silica) by N2 sorption at 77 K
34 Pore size analysis of MCM 41: Calculations compared
38 HK & SF Horvath-Kawazoe & Saito-Foley Direct mathematical relationship between relative pressure (P/Po) and pore size. Relationship calculated from modified Young-Laplace equation, and takes into account parameters such as magnetic susceptibility. Based on slit-shape pore geometry (e.g. activated carbons). Calculation restricted to micropore region ( 2nm width).SFSimilar mathematics to HK method, but based on cylindrical pore geometry (e.g. zeolites). Calculation restricted to micropore region ( 2 nm diameter).
39 DA & DR Dubinin-Astakov and Dubinin-Radushkevic Closely related to DR calculation based on pore filling mechanism. Equation fits calculated data to experimental isotherm by varying two parameters, E and n. E is average adsorption energy that is directly related to average pore diameter, and n is an exponent that controls the width of the resulting pore size distribution. The calculated pore size distribution always has a skewed, monomodal appearance (Weibull distribution).DRSimple log(V) vs log2(Po/P) relationship which linearizes the isotherm based on micropore filling principles. “Best fit” is extrapolated to log2(Po/P) (i.e. where P/Po = 1) to find micropore volume.
40 BETThe most famous gas sorption model. Extends Langmuir model of gas sorption to multi-layer. BET equation linearizes that part of the isotherm that contains the “knee” , i.e. that which brackets the monolayer value. Normally solved by graphical means, by plotting 1/(V[(Po/P)]-1) versus P/Po. Monolayer volume (Vm) is equal to 1/(s+i) where s is the slope and i is the y-intercept. Usually BET theory is also applied to obtain the specific surface area of microporous materials, although from a scientific point of view the assumptions made in the BET theory do not take into account micropore filling. Please note, that for such samples the linear “BET” range is found usually at relative pressures< 0.1, in contrast to the classical BET range, which extends over relative pressures between 0.05 – 0.3.
41 LangmuirAdsorption model limited to the formation of a monolayer that does not describe most real cases. Sometimes can be successfully applied to type I isotherms (pure micropore material) but the reason for limiting value (plateau) is not monolayer limit, but due to micropore filling. Therefore type I physisorption isotherm would be better called “pseudo-Langmuir” isotherm.
42 t-plot Statistical Thickness Multi-layer formation is modeled mathematically to calculate a layer “thickness, t” as a function of increasing relative pressure (P/Po). The resulting t-curve is compared with the experimental isotherm in the form of a t-plot. That is, experimental volume adsorbed is plotted versus statistical thickness for each experimental P/Po value. The linear range lies between monolayer and capillary condensation. The slope of the t-plot (V/t) is equal to the “external area”, i.e. the area of those pores which are NOT micropores. Mesopores, macropores and the outside surface is able to form a multiplayer, whereas micropores which have already been filled cannot contribute further to the adsorption process.It is recommended to initially select P/Po range 0.2 – 0.5, and subsequently adjust it to find the best linear plot.
43 BJH & DH Barrett, Joyner, Halenda and Dollimore-Heal Modified Kelvin equation. Kelvin equation predicts pressure at which adsorptive will spontaneously condense (and evaporate) in a cylindrical pore of a given size. Condensation occurs in pores that already have some multilayers on the walls. Therefore, the pore size is calculated from the Kelvin equation and the selected statistical thickness (t-curve) equation.DHExtremely similar calculation to BJH, which gives very similar results. Essentially differs only in minor mathematical details.
44 Other Methods FRACTAL DIMENSION The geometric topography of the surface structure of many solids can be characterized by the fractal dimension D, which is a kind of roughness exponent. A “flat” surface is considered D is 2, however for an irregular (real) surface D may vary between 2 and 3 and expresses so the degree of roughness of the surface and/or porous structure. The determination of the surface roughness can be investigated by means of the modified Frenkel-Halsey Hill method, which is applied in the range of multilayer adsorption.
62 Pore filling pressures for nitrogen in cylindrical pores at 77 K, (Gubbins et al. 1997)
63 Pore filling pressures for nitrogen in cylindrical silica pores at 77 K (Neimark et al., 1998)
64 Pore size analysis of MCM 41 by silica by N2 sorption at 77 K
65 Gas- and liquid density profiles in a slit pore by GCMC (Walton and Quirke,1989)
66 NLDFT / GCMC (Monte Carlo) Kernel File Applicable Pore Diameter Range ExamplesNLDFT– N2 - carbon kernel at 77 K based on a slit-pore model0.35nm-30 nmCarbons with slit-like pores, such as activated carbons and others.NLDFT– N2 – silica equilibrium transition kernel at 77 K, based on a cylindrical pore model0.35nm- 100nmSiliceous materials such as some silica gels, porous glasses, MCM-41, SBA-15, MCM-48 and other adsorbents which show type H1 sorption hysteresis.NLDFT– N2 - silica adsorption branch kernel at 77 K, based on a cylindrical pore model0.35nm-100nmSiliceous materials such as some controlled pore glasses, MCM-41, SBA-15, MCM-48, and others. Allows to obtain an accurate pore size distribution even in case of type H2 sorption hysteresisNLDFT– Ar zeolite/silica equilibrium transition kernel at 87 K based on a cylindrical pore model0.35nm -100nmZeolites with cylindrical pore channels such as ZSM5, Mordenite, and mesoporous siliceous materials (e.g., MCM-41, SBA-15, MCM-48, some porous glasses and silica gels which show type H1 sorption hysteresis).
67 NLDFT / GCMC (Monte Carlo) Kernel File Applicable Pore Diameter Range ExamplesNLDFT – Ar-zeolite/silica adsorption branch kernel at 87 K based on a cylindrical pore model0.35nm-100nmZeolites with cylindrical pore channels such as ZSM5, Mordenite etc., and mesoporous siliceous materials such as MCM-41, SBA-15, MCM-48, porous glasses some silica gels etc). Allows to obtain an accurate pore size distribution even in case of H2 sorption hysteresis.NLDFT – Ar-zeolite / silica equilibrium transition kernel based on a spherical pore model (pore diameter < 2 nm) and cylindrical pore model (pore diameter > 2 nm)Zeolites with cage-like structures such as Faujasite, 13X etc. , and mesoporous silica materials (e.g., MCM-41, SBA-15, porous glasses, some silica gels which show H1 sorption hysteresis).NLDFT – Ar-zeolite / silica adsorption branch kernel at 87 K based on a spherical pore model (pore diameter < 2 nm) and cylindrical pore model (pore diameter > 2 nm)Zeolites with cage-like structures such as Faujasite, 13X, and mesoporous silica materials (e.g., MCM-41, SBA-15, controlled-pore glasses and others). Allows to obtain an accurate pore size distribution even in case of H2 sorption hysteresis.
68 NLDFT / GCMC (Monte Carlo) Kernel File Applicable Pore Diameter Range ExamplesNLDFT – Ar - carbon kernel at 77 K based on a slit-pore model0.35 nm - 7 nmCarbons with slit-like pores, such as activated carbons etc.NLDFT - CO2 - carbon kernel at 273 K based on a slit-pore model0.35nm-1.5 nmGCMC – CO2 - carbon kernel at 273 K based on a slit-pore model
69 RECENT ADVANCES IN THE PORE SIZE ANALYSIS OF MICRO- AND MESOPOROUS MOLECULAR SIEVES BY ARGON GAS ADSORPTION
70 Micropore Size Characterization Physical adsorption in micropores, e.g. zeolites occurs at relative pressures substantially lower than in case of adsorption in mesopores.Adsorption measurements using nitrogen at 77.4 K is difficult, because the filling of nm pores occurs at P/Po of 10-7 to 10-5, where the rate of diffusion and adsorption equilibration is very slow.
71 Advantages of Using Argon Advantage to analyze such narrow micropores by using argon at liquid argon temperature (87.3 K).Argon fills these micropores (0.5 – 1nm) at much higher relative pressures (i.e., at relative pressures 10-5 to 10-3) compared to nitrogen.
72 Advantages of Higher Temperature & Pressure Accelerated diffusion.Accelerated equilibration processes.Reduction in analysis time.
73 Argon Adsorption at 87.3 K versus Nitrogen Adsorption at 77.4 K The different pore filling ranges for argon adsorption at 87.3K and nitrogen adsorption at 77.4K in faujasite-type zeolite are illustrated above.
74 Micropore Size Calculation Difficulties are associated with regard to the analysis of micropore adsorption data.Classical, macroscopic, theories  like DR and semiempirical treatments such those of HK and SF do not give a realistic description of micropore fillingThis leads to an underestimation of pore sizes for micropores and even smaller mesopores . F. Rouquerol, J. Rouquerol & K. Sing, Adsorption by Powders & Porous Solids, Academic Press, 1999[ 2 ] P. I Ravikovitch, G.L. Haller, A.V. Neimark, Advcances in Colloid and Interface Science , 203 (1998)
75 New CalculationTo overcome the above mentioned problems we introduce a new method for micropore analysis based on a Non-local Density Functional Theory (NLDFT) model by Neimark and Co-workers [3-5].The new DFT-method is designed for micro-mesopore size characterization of zeolitic materials ranging in size from 0.44 to 20 nm using high-resolution low-pressure argon adsorption isotherms at 87.3 K. P.I. Ravikovitch, G.L. Haller, A.V. Neimark, Advances in Colloid and Interface Science, 76 – 77 (1998), A.V. Neimark, P.I Ravikovitch, M. Gruen, F. Schueth, and K.K. Unger, J. Coll. Interface Sci., 207, (1998) 159 A.V. Neimark, P.I. Ravikovitch, Microporous and Mesoporous Materials (2001) 44-45, 697
76 Systematic, Experimental Study To evaluate the application of argon sorption for micro- and mesopore size analysis of zeolites and mesoporous silica materials including novel mesoporous molecular sieves of type MCM-41 and MCM-48.The sorption isotherms were determined using a static volumetric techniqueSamples were outgassed for 12 h under vacuum (turbomolecular pump) at elevated temperatures (573 K for the zeolites and 393 K for MCM-41/MCM-48).
77 ResultsArgon adsorption isotherms at 87 K on MCM-41, ZSM-5 and their mixture.
82 Pore Size Distribution 0.010.020.030.040.050.060.070.080.090.11101001000D, [Å]dV/dD [cm3/g]0.20.30.40.50.6Vcum [cm3/g]histogramintegral
83 DiscussionArgon sorption at 77 K is limited to pore diameters smaller than 12 nm.i.e. no pore filling/pore condensation can be observed at this temperature for silica materials containing larger pores.This lack of argon condensation for pores larger than ca. 12 nm is associated with the fact, that 77 K is ca. 6.8 K below the bulk triple point [4,5] . M. Thommes, R. Koehn and M. Froeba, J. Phys. Chem. B (2000), 104, 7932 M. Thommes, R. Koehn and M. Froeba, Stud. Surf. Sci. Catal., (2001),
84 DiscussionThese limitation do not exist for argon sorption at its’ boiling temperature, i.e. ca. 87 K.Pore filling and pore condensation can be observed over the complete micro- and mesopore size range .
85 DiscussionResults of classical, and semi-empirical methods (e.g., BJH, SF etc) indicate that these methods underestimate the pore size considerably.Deviations from the DFT-results are often in a range of ca. 20 % for pore diameters < 10 nm.
86 SummaryOur results indicate that argon sorption data at 87 K combined with the new NLDFT-methods provides a convenient way to achieve an accurate and comprehensive pore size analysis over the complete micro-and mesopore size range for zeolites, catalysts, and mesoporous silica materials.
87 AcknowledgementsSpecial thanks go to Alex Neimark and Peter Ravikovitch at TRI Princeton, New Jersey, USA.
88 References to research work of nitrogen, argon and krypton in MCM-48/MCM-41 materials (1) M. Thommes, R. Koehn and M. Froeba, “ Systematic Sorption studies on surface and pore size characteristics of different MCM-48 silica materials”, Studies in Surface Science and Catalysis 128, 259 (2000)(2) M. Thommes, R. Koehn and M. Froeba, “Sorption and pore condensation behavior of nitrogen, argon and krypton in mesoporous MCM-48 silica materials” J. Phys. Chem. B 104, 7932 (2000)(3)M. Thommes, R. Koehn and M. Froeba, “Sorption and pore condensation behavior of pure fluids in mesoporous MCM‑48 silica, MCM-41 silica and controlled pore glass, Studies in Surface Science and Catalysis, 135, 17 (2001)(4)M. Thommes, R. Koehn and M. Froeba, “Characterization of porous solids: Sorption and pore condensation behavior of nitrogen, argon and krypton in ordered and disordered mesoporous silica materials (MCM-41, MCM-48, SBA-15, controlled pore glass, silica gel) at temperatures above and below the bulk triple point”, Proceedings of the first topical conference on nanometer scale science and engineering” (G.U. Lee, Ed) AIChE Annual Meeting, Reno, Nevada, November 4-9, 2001(5)M. Thommes, R. Koehn and M. Froeba, “Sorption and pore condensation behavior of pure fluids in mesoporous MCM-48 silica, MCM-41 silica and controlled pore glass at temperatures above and below the bulk triple point”, submitted to Applied Surface Science, (2001)
89 Rapid Micropore Size Analysis by CO2 Adsorption
91 RAPID MICROPORE ANALYSIS The advantages of micropore analysis with Quantachrome’s Density Functional Theory (DFT) and CO2 include:Speed of analysis; with the higher diffusion rate at K, analysis times are reduced as much as 90%.Carbon dioxide at K permits probing pores from about 2 angstroms (0.2 nm).
92 DFT ADVANTAGEDFT has recently been applied to describe the behavior of fluids that are confined in small pores. The current popular gas sorption models, e.g. BJH, HK, SF, DA, etc., assume that the density of the adsorbed phase remains constant, regardless of the size of the pores that are being filled. Packing considerations suggest that these models are less than satisfactory for analyses of pores less than 2 nm.
93 DFT “Fitting”For a given adsorbate-adsorbent system, DFT calculates the most likely summation of "ideal isotherms“ calculated from "ideal pores" of fixed sizes needed to match the experimental results.
94 CO2 for Speed!Typically, micropore analyses with nitrogen as adsorbate will require 24 hours or more to run.Using carbon dioxide as adsorbate provides several advantages.Carbon dioxide molecules are slightly thinner than nitrogen molecules (2.8 angstroms radius vs angstroms) and will fill smaller pores than nitrogen.The use of carbon dioxide allows the measurements to be made at K, typically with an ice/water bath.There is no longer any need to provide and maintain or replenish a level of liquid nitrogen during the analysis.
95 CO2 BenefitsAt this temperature, the diffusion rate of molecules moving through small and tortuous micropores is much higher than at 77.35K. This so-called "activated adsorption" effect led to the popularization of the use of carbon dioxide to characterize carbonaceous material since the early 1960s.
96 CO2 BenefitsThis higher diffusion rate is responsible for reducing the analysis time to a few hours for a complete adsorption experiment. The faster rate also provides for the possibility of using larger samples than with nitrogen adsorption, thus reducing sample weighing errors.Pore size distributions thus obtained are comparable to those from a 24-hour nitrogen/77.35K analysis.
99 CO2 Adsorption at 0oCDensity Functional Theory Micropore Distribution
100 CO2 Adsorption at 0oCMonte Carlo Simulation Micropore Distribution
101 How to do it? Hardware requirements for this new method are minimal: a wide- mouth dewar anda water-level sensor.The proprietary Quantachrome Autosorb® software provides the DFT data reduction capabilities to do the rest. Pore size distributions from about 2 angstroms can be determined from the data taken at K.Currently, calculation parameters are optimized for studies on carbon surfaces.
102 BIBLIOGRAPHY for Rapid Micropore Size Analysis by CO2 Adsorption 1. J. Garrido, A. Linares-Solano, J.M. Martin-Martinez, M. Molina-Sabio, F. Rodriguez-Reinoso, R. Torregosa Langmuir, 3, 76, (1987)2. F. Carrasco-Martin, M.V. López-Ramón, C. Moreno-Castilla. Langmuir, 9, 2758 (1993)3. P. Tarazona. Phys.Rev.A 31, 2672 (1985)4. N.A. Seaton, J.P.R.B. Walton, N. Quirke. Carbon, 27, 853 (1989)5. C. Lastoskie, K.E. Gubbins, N. Quirke. J.Phys.Chem., 97, 4786 (1993)6. J.J. Olivier. Porous Materials 2, 9 (1995)7. P.I. Ravikovitch, S.C. Ó Domhnaill, A.V. Neimark, F. Schüth, K.K. Unger. Langmuir, 11, 4765 (1995)8. A.V. Neimark, P.I. Ravikovitch, M. Grün, F. Schüth, K.K. Unger. COPS-IV, 1997 (in press)9. P.I. Ravikovitch P.I., D. Wei, W.T. Chuen, G.L. Haller,A.V. Neimark. J.Phys.Chem., May 199710. E.J. Bottani, V. Bakaev, W.A. Steele. Chem.Eng.Sci. 49, 293 (1994)11. M.M. Dubinin. Carbon 27, 457 (1989)