TGA (Thermal Gravimetric Analysis) generally refers to a combination of gravimetric and mass-spectrometry analysis. As such, it can be used to detect the presence of chemisorbed gases. This is typically negligible on most carbon adsorbents due to van der Waals repulsion, but it can occur in unique geometries like single-walled carbon nanotubes. TGA can be used effectively in tandem with isotope substitution to characterize and validate hydrogen spillover. An example for a spillover sample dosed first with H2 followed by D2, is shown in Figure 1. This has been interpreted as successive H2 and D2 dissociation on the spillover sites, recombination at the interface between the two isotope layers, and desorption at the same spillover sites.

Figure 1: TGA desorption trace for a spillover sample that was dosed first with H2 followed by D2.

The decomposition of the off-board rechargeable chemical hydride polyamino borane (Second dehydrogenation step of Ammonia borane) is given by: 

Equation 1: (NH2BH2)n → (NHBH)n + H2

Typical TGA-curves for this decomposition at different heating rates (1, 5 and 10 °C min−1) are shown in Figure 2a). A single mass loss step was observed at all heating rates. It is clearly evident that a mass loss is detectable in the same temperature range in which hydrogen release and heat evolution were determined. The final value of the mass loss depends significantly on the heating rate used. With rising heating rate the final mass loss increases from 7.1 wt.% at β = 0.1 °C/min to 20.3 wt.% at β = 10 °C/min. The formation of boron nitride at a temperature of 250°C is not probable, as follows from the volumetrically detected release of only 1.1 mol hydrogen per mol H₂BNH₂.

In Figure 2b) results of thermogravimetric and volumetric investigations are compared. Volumetric results were converted into mass loss data. The release of 1.1 mol H2, which was detected by volumetric investigations, corresponds to a mass loss of only 7.6 wt.%. The results for the mass loss from thermogravimetric and volumetric investigations are nearly in agreement at a heating rate of 1 °C/min. The amount of gaseous products evolved in addition to hydrogen should be very small at 1 °C/min. The thermal decomposition of polymeric aminoborane (H₂BNH₂)x is accompanied by the evolution of different gaseous products as follows by volumetric and thermogravimetric investigations.

Figure 2: a) Mass loss on the thermal decomposition of polymeric aminoborane (H2BNH2)X versus temperature b) Experimental data for the mass loss (TG, line) in comparison with the mass loss data calculated from the released amount of hydrogen (volumetric measurements, points), (heating rates 1 and 10 °C/min).

Gravimetric measurements are often used for measuring capacity for many hydrogen storage materials. However, in ammonia borane it is very difficult to obtain accurate information about the capacity using simple TGA measurements. The key issues that could skew measurements are:

• Foaming: Ammonia borane foams extensively in the range of 90-110 °C during hydrogen release. This often creates mechanical errors in the instrument balance thus reporting an erroneous weight loss.

• Sublimation: Sublimation of the storage material is more common on samples which are heated under TG conditions. Undetected loss of some of the sample via sublimation may result in a significant error in estimating hydrogen capacity.

• Non-hydrogen volatiles: Other volatile intermediates or byproducts such as ammonia and borazine also contribute to an overestimation of hydrogen capacity if not properly accounted for.

These errors though significant under certain conditions can be minimized by use of smaller sample sizes in larger sample cups equipped with a lid (with a small orifice) that can minimize sublimation and foaming issues. Volatile byproducts or intermediates can be quantified by other techniques mentioned in section dealing with gas composition analysis and corrected for those measurements.

Figure 3: Comparison between mass loss data detected thermogravimetrically (line) and calculated from volumetric results (points) (heating rate 1 K/min (a), 5 K/min (b))

TGA is less accurate than either the volumetric and gravimetric methods and is not typically used for measuring capacity.