Gas storage

Gas storage applications exploit the enhancement in the density of an adsorbate in the pores of a material due to overlap of the adsorption potentials. For any given pressure, using nanoporous materials should, in principle, allow the storage of a gas at a higher density as compared to a compressed gas store or at a more practical temperature compared to a liquid. Two of the most prominent applications are the storage of supercritical hydrogen and methane for use as alternative energy carriers or fuels for transportation. In this case, the challenge is to find a material that can adsorb significant amounts of either species at ambient temperature and reasonable pressures (e.g., <10 MPa). This tends to rely on adsorption in micropores. Low temperatures are generally required for hydrogen adsorption, although other properties can be exploited to potentially raise the operational temperature, including the enhanced interaction of gases with specific surface sites and structural flexibility, which offers the possibility of trapping the adsorbate molecules within the pore structure.^{Reference Zhao, Xiao, Fletcher, Thomas, Bradshaw and Rosseinsky51}

Many nanoporous materials have been investigated for hydrogen storage.^{Reference Suh, Park, Prasad and Lim25,Reference Thomas45,Reference Broom46} An increasing number are MOFs, but zeolites, silicas, carbons, and MOPs have all been the subject of recent studies.^{Reference Thomas45,Reference Tedds, Walton, Broom and Book52} Investigations of methane storage are increasing, and this trend is likely to continue due to the renewed interest in natural gas as an energy source.^{Reference Voser53} Previous studies of methane storage tended to focus on carbon^{Reference Lozano-Castelló, Alcañiz-Monge, de la Casa-Lillo, Cazorla-Amorós and Linares-Solano54} and, to a lesser extent, zeolites and silicas.^{Reference Menon and Komarneni55} However, the emergence of MOFs as a potential storage material has inevitably led to an increased interest in their use for methane storage.^{Reference Konstas, Osl, Yang, Batten, Burke, Hill and Hill47,Reference Makal, Li, Lu and Zhou48,Reference Düren, Sarkisov, Yaghi and Snurr56–Reference He, Zhou, Krishna and Chen58} The materials exhibiting the highest hydrogen and methane storage capacities reported to date are listed in Table I. In practical applications, the compressed gas stored between the particles in the bed and in larger macropores, in which significant adsorption does not occur, also needs to be considered, together with the weight and volume of the storage system itself.

Table I.

Absolute hydrogen and methane storage capacities of the best performing nanoporous materials, with some activated carbon and zeolite capacities included for comparison.

AC, ACF, and Maxsorb-A are porous carbons; Na-X and Ca-X are zeolites; PPN-4 is a porous organic polymer; the others are types of metal-organic frameworks. For comparison, the 2017 US Department of Energy (DOE) storage system targets for hydrogen include a gravimetric capacity of 0.055 kg H₂ kg^–1 (55 mg g^–1 or 5.5 wt%) and a volumetric capacity of 0.04 kg H₂ L^–1 (40 g L^–1)^{Reference Suh, Park, Prasad and Lim25} in the (delivery) temperature range of 233 to 358 K. The US DOE methane target, set in 2000, is 180 cm^{Reference Talu3} (STP) cm^–3 at 298 K and 3.5 MPa (where cm³ (STP) cm^–3 is the standard temperature and pressure equivalent volume of methane per unit volume of the adsorbent).^{Reference Düren, Sarkisov, Yaghi and Snurr56,Reference Zhou57} Note: AC, activated carbon; ACF, activated carbon fiber; MOF, metal-organic framework; NOTT, Nottingham; NU, Northwestern University; PCN, porous coordination network; PPN, porous polymer network; UTSA, University of Texas at San Antonio.

Related applications include the sequestration of carbon dioxide and the exploitation of shale gas and coal bed methane. In the former, CO₂ is removed from the atmosphere by injecting it into locations such as shale formations and non-economic coal seams.^{Reference White, Smith, Jones, Goodman, Jikich, LaCount, DuBose, Ozdemir, Morsi and Schroeder32,Reference Sakurovs, Day, Weir and Duffy66–Reference Sakurovs, Day and Weir68} This is related to the use of coal bed methane^{Reference Busch and Gensterblum34} and shale gas.^{69,Reference Kargbo, Wilhelm and Campbell70} The latter is classified as an unconventional gas resource. Recent advances in hydraulic fracturing and horizontal drilling^{Reference Kargbo, Wilhelm and Campbell70} have made exploitation economically viable.⁶⁹ This has been successfully demonstrated in the United States and has had a dramatic effect on the natural gas market.^{Reference Kuhn and Umbach71} The small total pore volume of shale causes a problem for gas adsorption studies because of the low uptakes, which are difficult to measure. However, in practice, the low sorption capacities of shale are counterbalanced by the prevalence of shale resources.⁶⁹

For geological applications, it is important to assess the capacity of different formations, which can contain both compressed gas in macropores and an adsorbed phase in nanopores. Different materials have different storage capacities, and accurate assessment of this is an important challenge in adsorption research. However, for the exploitation of coal bed methane and shale gas, it is also important to understand the depletion of the resource, which is dependent on both the thermodynamics and kinetics of adsorption.^{Reference Busch and Gensterblum34}

Separation and purification

Adsorption by nanoporous materials is used extensively for gas separation^{Reference Ruthven1–Reference Sircar6} and is often the technology of choice for the production and purification of industrial gases.^{Reference Häring72} Examples include H₂ purification, N₂ and O₂ production from air, and He production from natural gas. The most common mechanism used is the difference in the equilibrium uptakes of different gas or vapor species by the same adsorbent.^{Reference Sircar, Myers, Auerbach, Carrado and Dutta5,Reference Ruthven44} In fixed bed adsorption, a gas mixture is passed through an adsorbent column, in which one component is more strongly adsorbed. The bed is then regenerated by reducing the pressure or by heating. The former is known as pressure swing adsorption (PSA) and is the most commonly used method;^{Reference Sircar, Myers, Auerbach, Carrado and Dutta5,Reference Sircar6} the latter is known as temperature swing adsorption.

A key factor for any separation process is the selectivity, the definition of which depends on the mechanism. If separation is based on differences in equilibrium uptake, a simple selectivity can be calculated by comparing the isotherm uptakes at specific pressures. In the case of kinetic separations,^{Reference Ruthven1,Reference Yang2,Reference Sircar, Myers, Auerbach, Carrado and Dutta5,Reference Ruthven44} the ratio of the linear driving force rate constants can be compared. For air separation by PSA using carbon molecular sieves, for example, the equilibrium amounts of adsorbed O₂ and N₂ are very similar, whereas the linear driving force rate constant for O₂ is at least 20 times that of N₂, which results in highly selective adsorption of O₂.^{Reference Chagger, Ndaji, Sykes and Thomas20,Reference O’koye, Benham and Thomas73–Reference Reid and Thomas75} However, for both types of separation, the working capacity of the adsorbent is also important. This may be defined as the difference between the uptake at the adsorption and desorption or purge pressures. Other important considerations include the ease of regeneration, the kinetics of adsorption/desorption, and the thermal stability of the material. The ease of regeneration requires a relatively low enthalpy of adsorption, to ensure the adsorbate is not too strongly bound, and the kinetics must be rapid enough to not inhibit the adsorption and regeneration processes.

CO₂ is widely used in a range of industries and can be produced from many different sources, the most common being synthesis gas, or syngas, a term that covers a variety of gas mixtures consisting mainly of H₂, CO, and N₂, with lower concentrations of hydrocarbons, CO₂, and Ar.^{Reference Häring72} Syngas itself can be produced in a number of ways, including the gasification of fossil fuels and biomass and steam reforming of natural gas.^{Reference Häring72} However, an emerging application is carbon capture and storage (CCS) for the reduction of greenhouse gas emissions from power plants.^{Reference Figueroa, Fout, Plasynski, McIlvried and Srivastava76–Reference Espinal and Morreale79} This involves CO₂ removal from both pre- and post-combustion gas streams, and adsorption is one of the proposed methods for its capture. It has potential advantages over absorption by liquids, such as amine solutions, and a variety of porous adsorbents are therefore being considered.^{Reference Choi, Drese and Jones80–Reference Liu, Thallapally, McGrail, Brown and Liu84}

Pre-combustion capture refers to the separation of CO₂ from syngas that has undergone the water-gas shift reaction, in which H₂O reacts with CO, to convert the CO into CO₂ and produce further H₂. The mixture therefore consists primarily of H₂ and CO₂, so the capture process mainly involves CO₂/H₂ separation. Depending on the gasification or reforming process, the pressure is typically a few MPa.^{Reference Plasynski, Litynski, McIlvried and Srivastava77,Reference Kanniche, Gros-Bonnivard, Jaud, Valle-Marcos, Amann and Bouallou85} Post-combustion capture predominantly involves the removal of CO₂ from N₂-rich flue gases at ambient pressure.^{Reference Figueroa, Fout, Plasynski, McIlvried and Srivastava76,Reference Drage, Snape, Stevens, Wood, Wang, Cooper, Dawson, Guo, Satterley and Irons86} It therefore requires CO₂/N₂ separation, normally in the presence of water vapor, and so hydrothermal stability is crucial.^{Reference Choi, Drese and Jones80–Reference D’Alessandro, Smit and Long82,Reference Liu, Thallapally, McGrail, Brown and Liu84,Reference Drage, Snape, Stevens, Wood, Wang, Cooper, Dawson, Guo, Satterley and Irons86,Reference Kizzie, Wong-Foy and Matzger87} Both hydrothermal and thermal stability are issues for MOFs and MOPs, because they are typically less stable than traditional adsorbents. However, some types, such as zeolitic imidazolate frameworks (e.g., ZIF-8), are reported to be relatively stable.^{Reference Li, Sculley and Zhou49,Reference Liu, Thallapally, McGrail, Brown and Liu84} Materials must also be resistant to other contaminants in flue gas.^{Reference Keskin, van Heest and Sholl81,Reference Liu, Thallapally, McGrail, Brown and Liu84,Reference Drage, Snape, Stevens, Wood, Wang, Cooper, Dawson, Guo, Satterley and Irons86,Reference Granite and Pennline88}

New materials for CCS should therefore satisfy the stringent performance and long-term stability criteria for large-scale industrial use, involving real gas mixtures under cyclic conditions, as well as exhibiting high selectivities and working capacities, as determined under laboratory conditions. It is important to note that, due to its large quadrupole moment in comparison with other permanent gases, such as H₂, CH₄, and N₂,^{Reference Reid and Thomas75} CO₂ will generally adsorb more strongly, so most nanoporous materials could, in principle, be used for its separation.^{Reference Keskin, van Heest and Sholl81} Demonstrable CO₂ selectivity alone is therefore not enough to qualify a material as a strong candidate for CCS applications. Economic and production scale-up considerations are also critical, which is particularly relevant, for example, when expensive MOFs are considered.^{Reference Liu, Thallapally, McGrail, Brown and Liu84}

Adsorption measurement

Accurate determination of the adsorption behavior of a material is clearly important, regardless of the application. Data are required by chemical engineers for process design,^{Reference Sircar7} but such characterization is also critical for new materials discovery and development. Understanding the most important physical or chemical properties of a material and their effect on the thermodynamics and kinetics of adsorption will clearly help guide the materials development process.

Single component adsorption measurement techniques

Single component measurements most commonly involve the determination of adsorption isotherms, which are plots of the uptake of a particular gas versus pressure at constant temperature. In addition to providing the capacity of a material, isotherms measured at a range of temperatures also allow the determination of the enthalpy of adsorption at zero surface coverage,^{Reference Chen, Zhao, Putkham, Hong, Lobkovsky, Hurtado, Fletcher and Thomas89} which provides a fundamental measure of the adsorbate-adsorbent interaction. At higher loadings, adsorbate-adsorbate interactions can modify the enthalpy of adsorption.

The traditional methods are the volumetric and gravimetric techniques.^{Reference McBain and Bakr90–Reference Kiefer and Robens94} The most common form of the volumetric technique is the manometric method, in which the sorbed quantity is determined by measuring the change in pressure in a system of fixed volume. This is an indirect measurement in which gas is dosed from a calibrated volume into the sample cell. The pressure drop in the absence of adsorption can be calculated from the ratio of the dosing volume to the total dead volume of the system. Any further reduction in pressure is attributed to adsorption, the magnitude of which is calculated from a molar balance expression. This approach is commonly applied to the measurement of adsorption at sub-ambient pressure by commercial surface area and pore size distribution analyzers^{Reference Lowell, Shields, Thomas and Thommes95} used as standard characterization tools for new nanoporous materials. The gravimetric method, on the other hand, is a direct measurement that uses a microbalance. A measurement begins at vacuum, and any weight gain at higher pressure is then attributed to adsorption.

In both techniques, isotherms are determined by sequentially dosing gas at increasing pressure. The calculated adsorption data must allow for the presence of the sample in the measurement chamber. The gravimetric method therefore requires corrections for the buoyancy effect, whereas the manometric method requires so-called dead volume or dead space corrections to account for the difference between the volume of the empty and loaded sample cell.

In principle, both techniques also allow for the determination of adsorption kinetics,^{Reference Kärger, Ruthven and Theodorou96} but the gravimetric approach has the significant advantage that kinetic measurements can be performed under isobaric (constant chemical potential) conditions. The data are then amenable to the application of kinetic models, such as the linear driving force,^{Reference Chagger, Ndaji, Sykes and Thomas20,Reference Reid and Thomas97} stretched exponential^{Reference Klafter and Shlesinger98–Reference Fletcher, Uygur and Thomas100} and double exponential models,^{Reference Fletcher, Cussen, Bradshaw, Rosseinsky and Thomas101} and solutions to the diffusion equation.^{Reference Chagger, Ndaji, Sykes and Thomas20,Reference Kärger, Ruthven and Theodorou96,Reference Crank102}

Other methods have also been reported for the measurement of single component gas adsorption, although these are not as widely applied because they are either less convenient to implement or have lower accuracy. For example, there are a number of variations on the volumetric technique, which include differential volumetric measurement^{Reference Zielinski, Coe, Nickel, Romeo, Cooper and Pez103,Reference Qajar, Peer, Rajagopalan and Foley104} and flowing methods.^{Reference Rouquerol, Rouquerol and Sing92,Reference Nelsen and Eggertsen105} Frequency response techniques have also been used to study kinetics.^{Reference Sircar7,Reference Kärger, Ruthven and Theodorou96}

Multicomponent (competitive) adsorption measurement techniques

Multicomponent adsorption can be determined from variations of the volumetric and gravimetric techniques, but both require an additional method of determining the amount of each species adsorbed. Flowing gravimetric systems can be used, with the gas phase composition beyond the sample determined using a mass spectrometer.^{Reference Fletcher, Benham and Thomas106} Different volumetric approaches are also available, including both open and closed systems.^{Reference Talu3,Reference Talu107} Closed system methods typically involve the circulation of a mixture through an adsorbent bed and the analysis of the gas phase composition after equilibration. Open system approaches involve equilibrating an adsorbent bed with a mixture and either using a molar balance expression to calculate the amount of each component adsorbed, or desorbing the entire contents of the bed and subsequently analyzing its composition.^{Reference Sircar7}

The volumetric and gravimetric techniques can also be combined such that the total adsorption is determined both volumetrically and gravimetrically. Providing that the adsorbates under investigation differ sufficiently in molecular weight, the amount of each component adsorbed can be calculated. The isotope exchange technique, which is a closed loop cycle approach involving isotopic tracers, can also be used.^{Reference Sircar7,Reference Rynders, Rao and Sircar108}

The measurement of breakthrough curves for packed beds, which show the time required for the components of the challenge (inlet) mixture to reach a preset fraction of the input concentration (C ₀) at the outlet (see Figure 2), can also be used to study multicomponent adsorption.^{Reference Keskin, van Heest and Sholl81} Breakthrough curves can provide both a qualitative and quantitative comparison between materials, but the data can also be fitted to expressions, such as the Wheeler–Jonas^{Reference Wood109,Reference Lodewyckx, Wood and Ryu110} and constant pattern linear driving force^{Reference LeVan, Rodrigues, LeVan and Tondeur111} models, and others based on assumptions applicable to the adsorbate-adsorbent system under investigation. These expressions allow the quantification of kinetic parameters, as well as the breakthrough time (see Figure 2). The assumptions can include models for the variation in flow velocity across the reactor, isothermal or non-isothermal conditions, and linear or nonlinear isotherm behavior.^{Reference Ruthven1,Reference Yang2} Model breakthrough curves for single component and binary mixtures are shown in Figure 2.

Figure 2.

Model breakthrough curves for (a) a single component in a carrier gas and (b) a two component mixture in a carrier gas.^{Reference Wood109}C/C ₀ is the breakthrough fraction, or the relative concentration of the adsorbate in the effluent gas stream, where C is the concentration at any given time, and C ₀ is the inlet (challenge) concentration. A curve can be characterized by the breakthrough time, t _B, which is the time taken for C/C ₀ to reach a particular threshold (A), and the stoichiometric time, t _s, which is the time taken to reach C/C ₀ = 0.5 (B). The gradient of the red line is dependent on interparticle diffusion and intraparticle sorption kinetics, axial flow dispersion, and heat transfer effects.^{Reference Ruthven1} For the binary mixture, C _m(1) is the maximum concentration in the bed reached by component 1, which is greater than its concentration at the inlet, C ₀₍₁₎. C ₀₍₂₎ is the inlet concentration of component 2, for which the adsorbent has a higher adsorption capacity.

Experimental challenges

Despite its conceptual simplicity, adsorption is surprisingly difficult to measure from a practical standpoint. Although low pressure single component gas adsorption measurements are relatively straightforward, care is still required to avoid common pitfalls and hence perform accurate measurements. However, some of the potential error sources increase significantly at higher pressures, which are often required in order to analyze the performance of materials for practical applications. The measurement of multicomponent adsorption, on the other hand, is challenging regardless of pressure.

Challenges in single component adsorption measurement

Single component gas adsorption measurements are common, partly due to the popularity of BET (Brunauer–Emmett–Teller) analysis of N₂ adsorption data at 77 K to provide a comparative measure of the surface area of powders and porous materials, in particular, in catalysis, separation, and purification applications.^{Reference Rouquerol, Rouquerol and Sing92,Reference Lowell, Shields, Thomas and Thommes95} This approach, and the use of N₂, Ar, and CO₂ adsorption for the determination of pore volumes and pore size distributions, typically involves the measurement of adsorption at sub-ambient pressure using a manometric apparatus.^{Reference Lowell, Shields, Thomas and Thommes95} These measurements can still be subject to significant experimental error,^{Reference Sing, Everett, Haul, Moscou, Pierotti, Rouquérol and Siemieniewska19,Reference Badalyan, Pendleton and Wu112,Reference Pendleton and Badalyan113} but at elevated pressures, the errors increase notably due to a number of factors, including those related to the calculation of the adsorbed quantity, such as error propagation, and experimental problems, such as leaks in the apparatus and the presence of impurities in the gas phase at higher partial pressures.

Apart from the practical difficulties of working at higher pressures, there are two main issues: the definition or determination of the volume of the solid and the description of the behavior of real gases. Both issues affect the dead volume and buoyancy effect corrections, which are equivalent.^{Reference Fuller, Poulis, Czanderna and Robens114} The magnitudes of both corrections depend on the density of the displaced gas and therefore increase with pressure. They also depend on the sample density and, for a given sample mass, increase with decreasing material skeletal density.

The measured quantity in an adsorption experiment is the excess adsorption.^{Reference Sircar115,Reference Sircar116} In the volumetric case, this is defined as the difference between the amount of gas present in the dead (unoccupied) volume of the apparatus in the event of adsorption and the amount of gas that would otherwise be present in its absence. The latter is typically determined using helium, under the assumptions that helium adsorption is negligible and the volume accessible to the helium is equal to the volume accessible to the adsorbate. Neither assumption is likely to be completely satisfied, and any associated error will contribute to the error in the calculated excess adsorption. Application of the buoyancy effect correction in the gravimetric case, without accounting for the adsorbed phase, also yields the excess adsorption.

At higher pressures, it is necessary to account for real gas behavior.^{Reference Zhou, Wu, Hartman and Yildirim117} The molar balance expressions used to calculate adsorption in the manometric method include the compressibility factor, Z, of the gas. Similarly, the gas density is also required for the buoyancy effect corrections. The determination of Z requires the use of an accurate equation of state, but also high accuracy measurement of both pressure, P, and temperature, T. The effect of errors in P, T, and Z depends on the gas species because of differences in their behavior. The problems are particularly acute for CO₂^{Reference Busch and Gensterblum34,Reference van Hemert, Bruining, Rudolph, Wolf and Maas118} because of the strong dependence of Z on temperature and pressure near its critical point (7.38 MPa, 304 K), which is a region of practical interest. Any error in Z will again contribute to the error in the calculated excess adsorption.

A further consideration is the conversion of the excess adsorption, n _ex, to the absolute adsorption, n _abs, which is defined as the total amount of adsorbate in the adsorbed phase. This can be calculated using the following equation,

(1)$$n_{abs} {\rm{ = }}\,n_{{\rm{ex}}} \,{\rm{ + }}\,V_{\rm{a}} \rho _{\rm{g}} {\rm{,}}$$

where V _a is the volume of the adsorbed phase, and ρ_g is the bulk gas phase density. At low pressures, V _aρ_g will be small and can be insignificant when there is a significant amount of adsorption. However, this is not the case at elevated pressures and supercritical temperatures, so careful attention must be paid to the conversion under these conditions. The difference between n _ex and n _abs is illustrated schematically in Figure 3. It can be seen that n _ex reaches a peak at elevated pressures and then decreases as the V _aρ_g term becomes more significant.^{Reference Sakurovs, Day, Weir and Duffy66–Reference Sakurovs, Day and Weir68,Reference Sircar115,Reference Gensterblum, van Hemert, Billemont, Busch, Charriére, Li, Krooss, de Weireld, Prinz and Wolf119} The difficulty is that V _a is unknown.^{Reference Sircar115,Reference Sircar116}

Figure 3.

A schematic representation of the difference between the excess and absolute adsorption as a function of pressure at constant temperature. The pores on the left-hand side are shown filling up as a function of the total pressure. The blue and orange regions represent the magnitude of the excess adsorption and the V _aρ_g term, respectively. At low total pressure (A), the majority of the absolute adsorption is due to the excess, but at high pressures (D), the V _aρ_g term contributes significantly to the absolute value. The excess begins to decrease once the V _aρ_g term reaches a sufficient magnitude. The pressure at which this occurs is usually above ambient, but its actual value is dependent on the measurement temperature and the adsorbate-adsorbent system. Note: V _a, volume of the adsorbed gas; ρ_g, bulk gas phase density.

The crystallographic structures of zeolites and MOFs can be determined using x-ray or neutron diffraction. Both techniques give a good indication of the pore volume, V _pore, which consists mainly of micropores. In this case, it seems reasonable to assume that V _a = V _pore, and this approach is used in the recent literature on hydrogen storage by MOFs.^{Reference Furukawa, Miller and Yaghi120,Reference Murray, Dincă and Long121} However, this is only an assumption, and it is possible that this approach may overestimate the absolute adsorption at lower pressures, in particular, and for larger pore materials, in general.

On the other hand, the structure of amorphous materials, such as activated carbons, polymers of intrinsic microporosity, and hyper-cross-linked polymers, as well as natural adsorbents, including coals and shales, cannot be determined crystallographically. These materials also possess a distribution of pore sizes, which cannot be determined precisely. The only methods currently available for pore volume and pore size distribution determination are based on gas adsorption, and these have their own limitations.^{Reference White, Smith, Jones, Goodman, Jikich, LaCount, DuBose, Ozdemir, Morsi and Schroeder32,Reference Hall, Brown, Fernandez and Calo122–Reference Lozano-Castelló, Cazorla-Amorós and Linares-Solano124} Therefore, the use of the assumption V _a = V _pore is more problematic for these materials, and the use of other approaches^{Reference Murata, El-Merraoui and Kaneko125} may be more appropriate. The main alternative is to assume that the adsorbed phase has a constant density, but its exact nature, particularly at supercritical temperatures, is unknown.^{Reference Sakurovs, Day, Weir and Duffy66–Reference Sakurovs, Day and Weir68,Reference Sircar115,Reference Sircar116}

This issue is likely to remain an active area of research, particularly as the absolute adsorption isotherm is also required for the comparison of experimental data with simulation results,^{Reference Talu and Myers126} which is an area of significant current interest.^{Reference Düren, Bae and Snurr127–Reference Xiang, Cao, Lan, Wang and Broom129} It is also crucial for the exploitation of shale gas and coal bed methane because, at higher pressures, although n _abs may decrease as the resource is depleted, n _ex may increase. Understanding the relationship between these two quantities is therefore important.

Challenges in multicomponent (competitive) adsorption measurement

The difficulty with multicomponent adsorption measurement is the determination of the concentration of each species in the adsorbed phase. Total uptakes from gas mixtures may be measured gravimetrically, but the composition of the adsorbed phase can only be inferred indirectly by assessing the amount of each species removed from the gas phase. This approach is limited by the sample size and the amount of adsorption exhibited by the material. Desorption of the adsorbed phase is then required to establish the amounts of each component present.^{Reference Fletcher, Benham and Thomas106} Similarly, volumetric techniques rely on the determination of the composition of a gas mixture in equilibrium with the solid, rather than the direct determination of the amount of each adsorbed component.

At elevated pressures, both volumetric and gravimetric measurements of multicomponent adsorption are affected by the same issues as single component adsorption, namely the definition and determination of the volume of the solid and the description of real gas behavior. For process design, the excess quantities may be used,^{Reference Talu3,Reference Sircar6,Reference Sircar7,Reference Sircar115,Reference Sircar116} so for some applications the excess to absolute conversion is not necessarily required. However, multicomponent volumetric measurements have additional issues. According to Talu,^{Reference Talu3,Reference Talu107} closed systems have significant advantages in terms of accuracy. The open system methods require accurate measurement of both the flow and gas composition at the outlet, which is difficult to achieve. Closed system measurements provide higher accuracy because they rely on the measurement of equilibrium pressures rather than flow, but it is important to ensure that the gas composition throughout the system is under equilibrium. Accuracy considerations require the use of relatively large sample sizes, which are not always available for newly synthesized materials, and there is also no way of controlling the equilibrium pressure of each component.^{Reference Sircar7,Reference Rynders, Rao and Sircar108}

In addition, multicomponent measurements are laborious. The time required can be appreciated by considering that one measurement yields a single data point, in terms of pressure, temperature, and composition (of either the bulk gas or adsorbed phases), yet equilibrium may take several hours or days, depending on the species, the adsorbent and, to a certain extent, the measurement method, because closed systems equilibrate more slowly than open systems.^{Reference Talu3}

The above points constitute serious practical challenges to the efficient and accurate measurement of multicomponent gas adsorption and have undoubtedly contributed to the shortage of comprehensive multicomponent adsorption data in the literature.^{Reference Sircar6} The number of materials being considered for adsorption applications and the laborious nature of multicomponent measurements means that the amount of multicomponent data available is likely to remain small compared with the number of possible adsorbate-adsorbent combinations. Models for the prediction of the adsorption of binary or higher mixtures from single component data, such as the ideal adsorbed solution theory,^{Reference Ruthven1,Reference Yang2} offer an alternative for the determination of multicomponent adsorption equilibria, but these models still require experimental datasets for their validation. The measurement of breakthrough data offers an attractive alternative, as this technique is more rapid than the equilibrium methods. However, more work is required in order to correlate multicomponent breakthrough data with detailed single component equilibrium and kinetic measurements.^{Reference Bell, Angus, Todd and Thomas130}