Magnesium and Doped Magnesium Nanostructured Materials for Hydrogen Storage - Low-cost Nanomaterials: Toward Greener and More Efficient Energy Applications - page 301

The method of determining PCIs, also known as Sievert's method, requires a
chamber of known volume [ 69 ]. When a sample is exposed to hydrogen, the
change in pressure can be accounted for absorption of hydrogen. Alternatively,
when a hydrogenated sample is exposed to vacuum, or a pressure lower than the
desorption equilibrium pressure, the pressure change can be attributed to desorp-
tion of hydrogen [ 78 ]. The recorded change in pressure, P, with the known vol-
ume, V, and temperature, T, of the chamber, can be applied to the ideal gas law,
PV = nRT, to determine the molar amounts, n, of absorption or desorption of
hydrogen (R is the gas constant). From this calculation, a percent hydrogenation
can be determined:

mass H 2
mass Mg þ mass H 2

Percent Hydrogenation ¼ ð 3 Þ

which can then be used to calculate a, where

H 2 weight%
Maximum

a ¼ ð 4 Þ

H 2 weight%

which can then be used in kinetic fittings.
Using a similar apparatus to the one used in Sievert's method, with the addition
of a microbalance, a GCI method measures the change in weight of the material
that is a result of the absorption and desorption of hydrogen [ 20 , 79 ]. Another main
difference in the method lies in the fact that because the change of mass is mea-
sured and not the change of pressure, the pressure of the chamber can be held
constant. This proves advantageous when a kinetic constant is calculated and
applied, for reasons explained later. Despite this mentioned advantage, corrections
must be made to the recorded mass to account for buoyancy effects from the
differences in air density within the chamber and the surroundings [ 79 ].
As mentioned earlier, for PCI and GCI data, the activation energy can be found
by plotting ln k versus 1/T. This plot is based on the Arrhenius equation:

ln ðÞ ¼ E a
R

1
T þ ln ð A Þ

ð 5 Þ

Before this can be done, however, the kinetic rate constant has to be deduced for
various temperatures.
The kinetic rate constant is found by fitting the experimental data with different
model equations, f(a) (Table 2 )[ 80 ]. Each function has a different rate-deter-
mining step, which is mentioned in the description in Table 2 . The best function
will produce a plot that is linear, with the slope that is equal to the rate constant,
k (Fig. 10 )[ 79 , 80 ]. However, because of the high surface area-to-volume ratio of
nanoscaled materials, fitting kinetic data to kinetic models based on bulk materials
proves arduous, if not impossible [ 81 , 82 ]. The two most common models used are
the Johnson-Mehl-Avrami (JMA) model of Nucleation and Growth (NG) [ 83 - 86 ]