Pi. Ravikovitch et al., Unified approach to pore size characterization of microporous carbonaceousmaterials from N-2, Ar, and CO2 adsorption isotherms, LANGMUIR, 16(5), 2000, pp. 2311-2320
We present a unified approach to pore size characterization of microporous
carbonaceous materials such as activated carbon and carbon fibers by nitrog
en, argon, and carbon dioxide adsorption at standard temperatures, 77 K for
N-2 and Ar and 273 K for CO2. Reference isotherms of N-2, Ar, and CO2 in a
series of model slit-shaped carbon pores in the range from 0.3 to 36 nm ha
ve been calculated from the nonlocal density functional theory (NLDFT) usin
g validated parameters of intermolecular interactions. Carbon dioxide isoth
erms have also been generated by the grand canonical Monte Carlo (GCMC) met
hod based on the 3-center model of Harris and Yung. The validation of model
parameters includes three steps: (1) prediction of vapor-liquid equilibriu
m data in the bulk system, (2) prediction of adsorption isotherm on graphit
e surface, (3) comparison of the NLDFT adsorption isotherms in pores to tho
se of GCMC simulations, performed with the parameters of fluid-fluid intera
ctions, which accurately reproduce vapor-liquid equilibrium data of the bul
k fluid. Pore size distributions are calculated by an adaptable procedure o
f deconvolution of the integral adsorption equation using regularization me
thods. The deconvolution procedure implies the same grid of pore sizes and
relative pressures for all adsorbates and the intelligent choice of regular
ization parameters. We demonstrate the consistency of our approach on examp
les of pore structure characterization of activated carbons from adsorption
isotherms of different gases and from different models (NLDFT and GCMC). S
ince the CO2 isotherms measured up to 1 atm are not sensitive to pores wide
r then 1 nm, the NLDFT method for CO2 has been extended to high-pressure CO
2 adsorption up to 34 atm. The methods developed are suggested as a practic
al alternative to traditional phenomenological approaches such as DR, HK, a
nd BJH methods.