Membrane Gas Separation

4.1.2
Details of the Photochromic Probe Method 
As anticipated above, photochromism of molecular probes is a phenomenon which is 
sensitive to the distribution of local free volume in polymer glasses [9] . Photochromic 
molecular probes in glassy polymers were fi rst used as far back as 1968 [10] . Since then 
they have been successfully used in evaluating the FV of various polymer matrices 
[11 – 14] . The principle of the method is based on the space required for photoisomeriza-
tion of probe molecules. For this purpose the photochromic or photo 
- 
isomerizable 
molecules, often stilbenes and azobenzenes, are dispersed homogeneously into the 
polymer matrix. Upon irradiation with UV or visible light with the proper wavelength 
and/or upon heating, the photochromic probes may undergo trans – cis isomerization and 
back. The crucial hypothesis at the basis of this technique is that photoisomerization in 
the glassy state requires a minimum critical size of local free volume in the vicinity 
of the chromophore. This isomerization requires a certain amount of extra volume during 
the isomerization of the molecule from the trans to the cis confi guration [9] not only 
because the molecular volume of the cis and the trans isomers may be different, but also 
because the structural rearrangement requires some freedom of motion. This is schemati-
cally displayed in Figure 4.3 . Only if this volume is available, i.e. if the photochromic 
λ
λ
’
, Δ
t
Figure 4.2 Photo - induced trans – cis isomerization of 4 - (4 - nitrophenylazo)aniline 
(4 - amino - 4 ’ - nitro - azobenzene, Disperse orange 3). The back reaction to the more 
stable trans isomer may be photo - induced or thermally induced
Figure 4.3 Illustration of the total volume required for the isomerization of the 
photochromic probe 4 - (4 - nitro - phenylazo)aniline (4 - amino - 4 ’ - nitro - azobenzene, Disperse 
orange 3)

---

Amorphous Glassy Perfl uoropolymer Membranes of Hyfl on AD®
63
molecule is located in a suffi ciently large free volume element, then photoisomerization 
will occur.
An important and generally accepted assumption is that the dye molecule does not 
infl uence its surroundings, i.e. it does not interact with the polymer matrix and it does 
not change the free volume itself. The polymer chain dynamics must be suffi ciently slow 
that there is no signifi cant relaxation of the free volume elements on the time scale 
required for the trans – cis transition. The latter makes this method only suitable for glassy 
polymers at temperatures suffi ciently far below the glass transition. 
In practice, the amount of probe photoisomerization in the polymer is measured relative 
to that in dilute solution in a non - viscous model solvent, where free volume is not a 
constraint to isomerization. The ratio between the two is an indication for the available 
local free volume of a given size and it is usually plotted as a function of the volume 
required for the photoisomerization of different probes. The cis fraction in solution rep-
resents the maximum possible degree of isomerization under the given conditions and the 
value in the polymer phase can range from 0% to 100% of the value in solution. For the 
same probe compound irradiated under the same conditions, a lower degree of isomeriza-
tion in the polymer phase compared to the solution indicates that a fraction of the mol-
ecules is trapped in too small free volume elements to enable successful isomerization. 
This procedure gives a fi rst qualitative impression of the distribution of local free volume 
if a series of photochromic molecules is used which require each a different volume for 
the isomerization. Under the proper conditions a quantitative free volume distribution can 
be obtained. If the concentration or number of probes ( n
probe
) is much lower than the 
number of FVEs ( n
hole
) in the polymer matrix, and if we further assume that the probes 
are homogeneously distributed in the polymer matrix, that is, without aggregation, then 
the relative degree of isomerization in fi lm compared to solution is equal to the fraction 
of probe molecules located in a FVE larger than or equal to the minimum isomerization 
volume of the probe molecule. Since n
hole
>> n
probe
, we can invert this relation and conclude 
that the relative degree of isomerization is also equal to the fraction of FVEs larger than 
the minimum volume required for the isomerization reaction. This represents a point in 
the cumulative free volume distribution curve. Using a series of different probe molecules 
with a suffi ciently wide range of isomerization volumes we will thus obtain several points, 
which together make up the entire cumulative distribution curve. 
This method has been successfully applied to various glassy polymers. Torkelson and 
co - workers studied the local free volume distribution in polystyrene (PS) [9] , poly(methyl 
methacrylate) (PMMA) [15] as well as bisphenol - A polycarbonate (PC) and poly(vinyl 
acetate) (PVAc) [16] . An interesting feature was that they were able to follow the process 
of physical ageing. The overall distribution is larger in PMMA than in PS and in both 
polymers physical ageing leads to a larger decrease in the fraction of large free volume 
elements than in the fraction of small free volume elements. In PS, on the other hand, a 
rearrangement of the entire FVD takes place and both the small and the large FVEs are 
affected. Results of PVAc demonstrated that the method cannot be successfully applied 
near the glass transition, where the typical chain relaxation times are of the same order 
of magnitude as the time required to reach the photostationary state of the probe molecules 
upon irradiation [16] . In this case, changes of the FVEs during the photoisomerization 
reaction compromise the interpretation of the data. In general the method can be applied 
to a wide range of systems. The matrix may be a thermoplastic polymer [11] , as in the 

---

64
Membrane Gas Separation
studies above, or a gel [13,14] and the probes may not only be dispersed in the polymer 
matrix but also chemically bound [12,14,17] . 
A fundamental requirement for the system to be suitable for the application of the 
photochromic probe method is that the polymer is suffi ciently transparent to UV or visible 
light in the spectral range of the photochromic probe molecules, fi rst of all to induce 
the photo - isomerization reaction upon irradiation of the sample, and secondly to allow 
the determination of the absorption spectrum of the isomers before and after the irradia-
tion, and thus to calculate the degree of isomerization. In this respect, Hyfl on AD is an 
excellent polymer because it is highly transparent in the visible spectrum and in most of 
the UV spectrum, the reason for its original application in optical devices.