Because the concentration of hydrogen in a sample is not measured directly, this important parameter necessarily relies on indirect determination from directly measured variables. Accurate measurement of direct variables is important for all hydrogen storage properties and perhaps more so for the determination of concentration than any other.

All storage properties are determined by the relationships between the variables: concentration, sample weight, temperature, pressure, cycle and time. The important distinction between direct (weight, temperature, pressure, time, cycle) and indirect (concentration) variables is made to differentiate between variables that can be measured directly from those that cannot. The last five (mass, temperature, pressure, cycle and time) are termed direct variables because they can be directly measured with traditional measuring devices such as balances, thermocouples and transducers. Concentration is considered an indirect variable because it cannot be directly measured and must be determined by correlating a direct (measurable) variable with concentration through the use of an empirical relation. Because of its dependence on measurable variables and its explicit ties to capacity, concentration will be addressed in detail in the Concentration and Capacity chapter.

The direct variables present problems in accurately collecting and interpreting data. Weight, temperature and pressure are difficult to accurately measure in all testing methods because errors can be introduced by testing equipment, sample holder design and secondary effects such as buoyancy. Errors associated with measuring the direct variables affect all property investigations and therefore the sources of these errors will be reviewed in the introductory chapter. In addition to this section on the direct variables, the effects of testing and variable collecting methods unique to a particular storage property (e.g. capacity, kinetics) will be addressed in the property chapters that follow.

• Measureable Direct Variables

Measureable or “direct” variables (weight, temperature, pressure, time, cycle) affect the intrinsic adsorption/absorption behavior of a material: 

Where Na is the number of moles adsorbed. In addition, the measurement of direct variables is needed to determine “indirect” variables such as capacity. The following are the direct variables that are typically measured to determine the hydrogen storage properties of a material. While the means and accuracies by which these variables are measured may vary (e.g., volumetric, gravimetric, electrochemical….) the influence on the intrinsic properties and performance of the materials should be the same.

•  Variable: Weight

The weight variable measured in the gravimetric method is often considered a direct measurement of hydrogen concentration; unfortunately, this is not the case. The balance in the gravimetric method (used to measure sample weight) actually measures the resultant force of a number of discrete forces: the weight of the sample, the buoyancy force caused by the displaced gas and the forces associated with mechanical disturbances and gas convection. The resultant force is often referred to as the apparent weight of the sample in order to differentiate between the measured weight and the actual weight of the sample. In order to determine the actual weight of the sample (from which concentration can be determined), it is necessary to account for and minimize the extraneous forces acting on the balance. The buoyancy force can be determined by an understanding of the skeletal density of the material while proper vibration damping can minimize mechanical forces. Forced and free convective forces are caused by pressure changes during charging/discharging and thermal gradients, respectively, and are transient in nature; patience is essential to minimize convective forces. Gravimetric measurements often use constant gas flow past the sample. In these cases, the flow drag forces must also be known and taken into account. Generally the flow drag force changes with the gas density, and therefore is a function of both temperature and pressure. Proper calibration of gas flow forces is necessary if the flow forces are significant in the particular apparatus or experimental arrangement.

In addition to the significant measurement considerations discussed above, the accuracy and precision of measuring weight depends on the type of balance used. A common weight-measuring system is an electronic strain gauge attached to a cantilever or balance. The accuracy and precision of the strain gauge and cantilever/balance system are based on the properties of the strain gauge and the cantilever/balance material. The deflection varies with the modulus of elasticity of the material and the weight of the sample; a low modulus of elasticity leads to greater deflection and a more sensitive instrument but the instrument is more susceptible to noise due to external forces. The modulus of elasticity will vary with temperature and this must be taken into account when making measurements at different temperatures or when ramping temperatures. The sensitivity of the balance must be considered along with other trade-offs when choosing gravimetric instruments. With respect to hydrogen storage materials, the hydrogen storage capacity is usually given on a weight basis. Therefore accurate measurement of a samples weight is critical. Both chemisorption and physisorption materials may be highly air and water reactive, therefore samples are usually prepared and weighed in an inert atmosphere such as an argon glove-box. In addition, both types of material may require significant out-gassing of residual solvents, or adsorb impurities such as water before hydrogen testing. Thus for correct hydrogen capacity measurements it is vital that the sample is weighed after degassing. This can be performed either before or after hydrogen testing and preferably in an inert gas atmosphere.

• Variable: Pressure

The pressure variable is intimately associated with hydrogen storage research and influences several important properties including concentration correlations, capacity and the kinetics of charging and discharging. Pressure is a commonly used measurable variable in concentration correlations and is therefore of practical interest to the measurement of all hydrogen storage properties. For example, the volumetric method measures changes in equilibrium pressure, along with volume and temperature information, to determine concentration. Changes in pressure are often used to control sorption/desorption reactions during both testing and practical operation. Isothermal measurements such as PCTs, one of the most common representations of capacity, use pressure variation to drive charging and discharging during storage property characterization. Pressure variation can also be used to control sorption and desorption in functioning storage systems, depending on the application. In fuel cell (FC) applications, hydrogen supplied at initially isothermal conditions (ambient) and elevated pressure is used to charge the storage system at the fueling station. The elevated pressure causes the reversible storage material to charge with hydrogen for future consumption by the fuel cell.

In addition to its obvious effect on reaction kinetics, pressure profoundly affects capacity measurements and practical capacities. Among the many capacity definitions are two based on thermodynamic considerations, reversible and usable capacity. Reversible capacity is of concern at the materials development level and is the measure of capacity available under feasible conditions. Reversible capacity is of particular importance to chemisorbing materials. Usable capacity is the capacity defined by the thermodynamic restrictions placed on the storage system by the environment and the end application. Once again, fuel cells present an excellent example of the importance of the thermodynamic variables on practical storage. The range of temperatures and pressures accessible to charge and discharge a storage system are restricted by availability; in classic PEMFC applications, this translates to operating temperatures from ambient to (~353K) and roughly ambient to a few bar pressures (assuming no external pressurizing or regulating equipment is used). It is highly probable that maximum capacity of a given storage system does not occur within these limited ranges; therefore it is important to quantify usable capacity.

Electronic pressure transducers are the most common means of measuring pressure in hydrogen storage testing equipment. The accuracy limitations of transducers are generally described by two different types of error bands, those based on a percent of the actual reading (Capacitance Manometers) and those based on a percent of the full scale (Strain Gauge Transducers) (see Figure 1). For this reason, percent reading error bands are more accurate at low pressures while percent full-scale error bands may be more accurate at high pressures depending on the total error rating of the transducer. For hydrogen storage testing, particularly volumetric methods, the low-pressure range (0-15 bar) is often the most critical range for investigating storage properties. This makes the percent reading error bands the preferred error band for hydrogen storage.

Figure 1: Depiction of the two types of error bands in pressure transducers

A simple calculation demonstrates the percent reading advantage at low pressures. Take two pressure transducers rated to 300 bar, one has a 1 % reading error band while the other has a 0.1 % full-scale error band; the ‘break even’ pressure for these two transducers is 30 bar. The percent reading transducer is more accurate for pressures below 30 bar, the range of pressures most often encountered in hydrogen storage experiments. For optimum accuracy, several percent reading transducers rated to different pressures should be used.

The sensitivity of pressure transducers requires that the pressure signal change from all other sources, ΔP0, is minimized. It is conceivable to potentially account for these other pressure signal changes and correct the data for their influence, but in practice it is more fruitful to minimize erroneous ΔP0 signals. Background pressure signal change can be caused by a number of factors including transducer sensitivity, zero drift, miss-calibration and hysteresis effects.

With a near instantaneous change in pressure, there may be some error introduced into the immediate response of the pressure transducer, especially with respect to pressure measurements made by measuring the strain or deflection of a gauge’s diaphragm. There are a couple of ‘tricks’ that can help determine the true measurement signal from a pressure transducer. After the transient from gas introduction has decayed, the measurement signal becomes pseudo-differential in the sense that the current signal can be compared to the measurements that came before and will come after. The signal curve is expected to show a smooth behavior and therefore noise superimposed on the curve can be “reduced” through standard procedures such as smoothing. If the sample shows marginal sorption/desorption during the transient, the pressure vs. time curve contains both the calibration data and the sorption/desorption data. In other words, since very little absorption occurred during the transient, the transient can be ignored and the pressure reading at t = 0 is the pressure with “no absorption”. Thus the ideal sample, ideal in the sense that it is easy to measure, would have no sorption/desorption signal until after the transient but would come to equilibrium before other longer-term error signals such as temperature fluctuations have a chance to impact the data.

• Variable: Temperature

Temperature is the second thermodynamic variable with important implications to hydrogen storage properties. It has many of the same hydrogen storage effects as pressure; it can be used to determine the binding energy of hydrogen in a sample (as in Differential Scanning Calorimetry), drive the sorption/desorption reactions in several testing methods and applications, and temperature influences a number of different definitions of capacity. To continue with the fuel cell example, temperature variation is the primary mechanism used to release hydrogen after initial charging because of the relatively limited pressures available due to FC structural considerations. Temperature affects practical capacity in much the same way as pressure. The two capacity definitions derived from thermodynamic considerations are based on temperature considerations as well as pressure considerations. Although, in hydrogen storage research, pressure and temperature are intimately linked, their measurements have their own, unique considerations.

Temperature-related measurement error is one of the most common errors associated with hydrogen storage measurements, particularly kinetics measurements. The accuracy of temperature measurements is typically limited to the accuracy of the thermocouples used and the heat transfer characteristics of the sample material and sample cell. Like pressure transducers, the accuracy of a thermocouple measurement depends on what type of thermocouple is used and the temperature regime of the measurement. Some thermocouple types have a wide temperature range but are less accurate compared to those with limited temperature ranges. Because most hydrogen storage experiments are conducted under isothermal conditions, the authors advise the use of limited-range durable thermocouples that offer greater accuracy in measurement.

An important consideration is that thermocouples of all types generally present nearly the same voltage at room temperature. It is only at elevated temperatures that the deviation in the type (J, K, and N for example) will become apparent. As most modern temperature measuring devices offer many thermocouple options, it is important not only to validate the devices thermocouple type settings but also to validate by using separate secondary temperature measuring devices and to run calibrations on measurement devices on a regular basis.

Note that most thermocouples (if not all) are at the limit of their operating ranges under cryogenic temperatures and can have large errors when within a few degrees of 77K. The temperature is preferably measured with Pt resistance or silicon diode at around or below 77K.

Heat transfer between the sample material, sample cell and thermocouple is one of the primary sources of error in hydrogen storage measurements. Sorption and desorption reactions can be highly exothermic and endothermic and the energy loads must be transported efficiently to maintain isothermal conditions during testing. Insufficient heat transfer can lead to pockets of sample at higher temperature than the temperature read by the thermocouple, effectively invalidating the isothermal assumption and conclusions based on that assumption. Thermocouples that are not in intimate contact with the sample, or perhaps not even in contact with the sample holder, can produce very large in-accuracies with respect to the actual sample temperature. On a system level where kilograms of storage material may be used, excellent heat transfer characteristics are required to supply and dissipate the significant amounts of energy necessary to charge and discharge a sample.

• Variable: Cycle

The desire for reusable hydrogen storage systems necessitates the ability to charge and discharge (cycle) repeatedly without loss of performance. Nearly all storage properties, particularly capacity and kinetics, vary with cycling and the variation can have a profound impact on system efficacy. Cycling phenomena include activation effects and poisoning (capacity) and retardation (kinetics) due to the gettering of impurities during cycling. In general a sample should be cycled until the capacity is stable; usually 10 cycles are sufficient for that purpose.

Some important considerations for the cycle variable are primarily aimed at minimizing activation and gas stream impurity effects on measurements and are not testing method-specific. For activation effects for both capacity and kinetics, the authors advise cycling the sample at least ten times in order to measure the intrinsic properties of the material. It is also critical to evaluate the effects of poisoning and retardation on performance, especially in metal hydrides; for practical application, storage systems will commonly be charged with hydrogen gas that contains impurities like CO2, H2O and NH3. The impurities adsorb to the material, occluding catalytic sites and diffusion pathways, and can be difficult to desorb because of their high thermodynamic affinity. As cycling increases with impure hydrogen gas, the impurities build to levels that poison capacity and retard kinetics.

Therefore, it is important to develop materials that can withstand the effects of gas stream impurities in order to avoid performance deterioration.

Several other phenomena occur when cycling a hydrogen storage material. In intermetallic compounds, decrepitation, self-pulverization due to stresses caused by lattice expansion upon hydriding, and disproportionation, dissociation of a compound into its fundamental components during repeated cycling (e.g. LaNi5+La2Mg17→3La+5Mg2Ni+7Mg), cause variation in storage properties and system performance. Decrepitation and disproportionation are generally considered activation phenomena and initial cycling mitigates their effects.

Note: UHP Ultra-High purity (99.999% purity) hydrogen gas is recommended. It is also highly recommended that the source hydrogen gas is always tested for purity, especially for oxygen and water contaminants. A good solution is to have a hydrogen purifier system just before the gas enters the testing system.

• Variable: Time

The effect of time on hydrogen storage measurements is manifested by the rate at which measurements are taken. However, some confusion surrounds how measurement rates affect data representation. Measurement data is collected in two distinct steps: the data acquisition hardware converts the continuous analog output from the measuring device (e.g. thermocouple, pressure transducer) to a digital value at a certain rate R1 and the computer software samples the digital value from the data acquisition hardware at another rate R2.

Figure 2: Schematic of a typical hardware software set-up and the flow of information in a hydrogen storage testing system

As Figure 2 illustrates, information is exchanged between two interfaces but not necessarily at the same rate. The sampling rate of the data acquisition hardware can be faster or slower than the sampling rate of the computer software. The difference between the rates may lead to the collection of multiple data points at the same value of the measured parameter and/or a step-wise rather than continuous change in the measured parameter. Both of these effects are most pronounced at small time steps that approach the limits of the sampling rates and are an artifact of data collection. The average results of the measured parameters are still representative of the homogeneous change in the properties being measured.

Figure 3 illustrates a number of data collection artifacts. At the beginning of the experiment, the R2 sampling rate of the computer software is faster than the analog to digital conversion rate R1 of the data acquisition hardware. Therefore, the computer software samples the data acquisition hardware several times before the hardware updates. This leads to several consecutive data points collected at the same pressure, an artifact that disappears as the R2 sampling rate decreases (by design) below the data conversion rate R1 as the experiment progress. The step-wise rather than continuous change in the data at the beginning of the experiment is caused by the difference between the sampling rates and the initial gradient in the measured parameter. Figure 3 shows the measured pressure directly after the sample is dosed with hydrogen; the pressure gradient is initially large but slowly levels off as the sample approaches equilibrium. This is reflected in the data as the pressure step size becomes less and less and the data eventually appears continuous. These kinds of data collection artifacts demonstrate the resolution limits of the instrumentation but generally do not have a significant impact on the accuracy of the measurement and the conclusions that can be drawn. However, without understanding what causes them, they may be misinterpreted as a problem with data acquisition or instrument operation when first observed.

Figure 3: Example of constant and step data collection artifacts during desorption. The ordinate is measured pressure

Reference List:

• Gross, K.J. “Characterization of Advanced Hydrogen Storage Materials”, MRS Spring Hydrogen Tutorial, (2006).

• Dantzer, P., “Metal-Hydride Technology: A Critical Review”, Topics in Applied Physics - Hydrogen in Metals III, 73 (1997) p. 279-340.

• Sandrock, G.D., Murray, J.J., Post, M.L., and Taylor, J.B., “State-of-the-art Review of Hydrogen Storage in Reversible Metal Hydrides for Military Fuel Cell Applications”, Final Report for ONR Contract N00014-97-M-0001, NTIS Order No. AD-A328 073/2INZ. (1997).

• Gross, K.J., “Intermetallic Materials for Hydrogen Storage”, PhD thesis. Institut de Physique, Université de Fribourg, (1998).

• Gross, K.J., Measurement performed by author.