Orhan Talu

Behavior and performance of adsorptive natural gas storage cylinders during discharge

Abstract Adsorbed natural gas (ANG) has the potential to replace compressed natural gas in mobile storage applications, such as in vehicles. Although a substantial research effort has been devoted to ANG, very few studies evaluate the impact of heat of adsorption on system performance. This paper concentrates on the impact of heat of adsorption on ANG performance during discharge, while the gas outflow rate is dictated by the energy demand of the application. The temperature drop and performance loss was measured with commercially available ANG cylinders under realistic conditions. Data show as high as a 37°C temperature drop at high discharge rate, with a performance loss approaching 25% of isothermal capacity. The performance loss is expected to be 15–20% at moderate discharge rates. Analysis of data and predictions of a simple model indicate that the ANG system is neither adiabatic nor isothermal during discharge; the thermal capacity of the vessel wall and external heat transfer conditions have a significant effect on system behavior. The poor thermal conductivity of packed adsorbent is a major obstacle for the utilization of these energy sources. Changing the flow direction during discharge from axial to radial by a perforated tube inserted at the center of the cylinder significantly reduces the performance loss by increasing the heat transfer from the wall to the central region. At intermediate discharge rates, where the inserted tube has the largest impact, the performance loss is reduced to 12% from 22% without the tube under identical conditions.

Net Adsorption: A Thermodynamic Framework for Supercritical Gas Adsorption and Storage in Porous Solids

The thermodynamic treatment of adsorption phenomena is based on the Gibbs dividing surface, which is conceptually clear for a flat surface. On a flat surface, the primary extensive property is the area of the solid. As applications became more significant, necessitating microporous solids, early researchers such as McBain and Coolidge implemented the Gibbs definition by invoking a reference state for microporous solids. The mass of solid is used as a primary extensive property because surface area loses its physical meaning for microporous solids. A reference state is used to fix the hypothetical hyperdividing surface typically using helium as a probe molecule, resulting in the commonly used excess adsorption; experimentalists measure this reference state for each new sample. Molecular simulations, however, provide absolute adsorption. Theoreticians perform helium simulations to convert absolute to excess adsorption, mimicking experiments for comparison. This current structure of adsorption thermodynamics is rigorous (if the conditions for reference state helium measurements are completely disclosed) but laborious. In addition, many studies show that helium, or any other probe molecule for that matter, does adsorb, albeit to a small extent. We propose a novel thermodynamic framework, net adsorption, which completely circumvents the use of probe molecules to fix the reference state for each microporous sample. Using net adsorption, experimentalists calibrate their apparatus only once without any sample in the system. Theoreticians can directly calculate net adsorption; no additional simulations with a probe gas are necessary. Net adsorption also provides a direct indication of the density enhancement achieved (by using an adsorbent) over simple compression for gas (e.g., hydrogen) storage applications.

Reference potentials for adsorption of helium, argon, methane, and krypton in high-silica zeolites

Quantitative agreement with experimental data for adsorption of argon, krypton, and methane in high-silica zeolites is achieved with a molecular model, which (1) assumes pairwise additivity of intermolecular forces; (2) ignores partially concealed silicon atoms and lumps their dispersion energy with the oxygen atoms; (3) neglects energy induced by the weak electric field inside silicalite; (4) uses the Lennard–Jones 12-6 potential for gas–gas and gas–solid interactions. The parameters of the Ar–O gas–solid potential are e/k=93.0 K and σ=3.335 A. The gas–solid potential parameters for krypton and methane were calculated from the Ar–O reference potential using Lorentz–Berthelot mixing rules and the principle of corresponding states. The pore volume of silicalite determined from the He–O potential is 0.175 cm3 g−1. The local density of gases adsorbed in the pores varies with the strength of the gas–solid potential; at 306 K and 10 bar, the maximum absolute density of methane is about one half the density of liquid methane at its normal boiling point.

Gibbs Dividing Surface and Helium Adsorption

All adsorption data is based on the definition of Gibbs dividing surface, which is a purely mathematical transformation. Adsorption measurements in microporous solids necessitate experimental determination of the dividing surface. An international protocol does not exist on how to perform this important measurement. Commonly, helium is assumed not to adsorb and used as a probe molecule for this measurement. Each experimentalist chooses an arbitrary set of conditions, often without even disclosing them, which adds to the confusion in adsorption literature. Here, a self-consistent method for the analysis of helium data is proposed which does not assume non-adsorbing helium. The method is compared to others using the extensive set of helium/silicalite data. The Gibbs dividing surface and hence the helium isotherms at all temperatures are determined.

Measurement and analysis of oxygen/nitrogen/ 5A-zeolite adsorption equilibria for air separation

Multicomponent adsorption equilibrium data are essential for the reliable design of processes and equipment for gas separation by adsorption. We discuss techniques for the measurement and analysis of multicomponent adsorption equilibrium data, and present a comprehensive set of equilibrium data for the adsorption of oxygen and nitrogen on 5A-zeolite.

Activity coefficients of adsorbed mixtures

Experimental and simulated data for adsorption of gas mixtures on energetically heterogeneous surfaces like activated carbon and zeolites exhibit negative deviations from ideality. The deviations are large in some cases, with activity coefficients at infinite dilution equal to 0.1 or less. Similar molecules form ideal mixtures, but molecules of different size or polarity are nonideal. Equations for bulk liquid mixtures (Wilson, Margules, etc.) do not apply to isobars for adsorbed mixtures. A two-constant equation for activity coefficients as a function of composition and spreading pressure is in good agreement with theory, simulation, and experiment.

An Overview of Adsorptive Storage of Natural Gas

Abstract The Adsorbed Natural Gas (ANG) storage technology is critically examined. The present status, theoretical limits and operational problems are discussed. Research areas which may make a significant impact are also indentified.

Long range corrections for computer simulations of adsorption

Long range corrections are routinely applied to simulations of bulk fluids by assuming that the radial distribution function is unity beyond a certain cutoff radius for pairwise interactions. Similar long range corrections for gas-solid interactions in adsorption frequently are ignored because of the anisotropic structure of the solid. However, the error associated with assuming isotropy beyond the cutoff radius is small compared with the magnitude of the long range correction. The long range correction to the Henry constant for a cutoff radius of 13 Å is 14% for CH4 and 70% for SF6 for adsorption in silicalite at 298 K. The large errors incurred by neglecting long range corrections can be concealed by increasing the well depth of the gas-solid interaction, but this approximation reduces the accuracy and portability of the potential parameters. Consistency in the cutoff radius is more important than the inclusion or neglect of long range corrections to the energy.

Structural effect on molecular simulations of tight-pore systems

Adsorption of benzene and p-xylene on silicalite at infinite dilution at 20 °C has been investigated via direct integration and Monte Carlo techniques. A small uncertainty in crystal structure causes a very large difference in the values of the Henry constants, while the internal energy of adsorption is less sensitive to this uncertainty. The sensitivity of the Henry constant to structural detail is magnified in these systems due to the tight fit of these molecules in the silicalite pore structure. This high sensitivity casts a shadow of doubt on the applicability of the common practice of determining potential parameters from data at infinite dilution while assuming that the structure is rigid and completely defined. The calculations reveal that the preferential adsorption sites for both benzene and p-xylene are the channel intersections. Benzene is somewhat mobile between straight channels and intersections while p-xylene is almost completely localized at the intersections at 20 °C. In tight-pore systems the most preferred adsorption site is not necessarily the location with the highest adsorption potential, contrary to adsorption on flat surfaces and in large pores.

Electrodeposition of Nickel Nanowires and Nanotubes Using Various Templates

Nickel nanotubes and nanowires are grown by galvanostatic electrodeposition in the pores of 1000, 100, and 15 nm polycarbonate as well as in anodised alumina membranes at a current density of 10 mA cm−2. The effects of pore size, porosity, electrodeposition time, effective current density, and pore aspect ratio are investigated. Nickel nanotube structures are obtained with 1000 nm pore size polycarbonate membrane without any prior treatment method. At the early stages of electrodeposition hollow nickel nanotubes are produced and nanotubes turn into nanowires at longer depositon times. As effective current density accounting for the membrane porosity decreases, the axial growth direction is favoured yielding nanowires rather than nanotubes. However, for smaller pore size polycarbonate membranes, nanowires are obtained even though effective current densities were higher. We believe that when the pore diameter is below a critical size, nanowires grow regardless of current density since narrow pores promote layer by layer growth of nanorods due to smaller surface area of the pore bottom compared to pore walls. Pore size has a dominant effect over effective current density in determining the structure of the fibres produced for small pores. Nickel nanowires are also obtained in the small pores of anodised alumina, which has higher aspect ratios. High aspect ratio membranes favour the fabrication of nanowires regardless of current density.