|
|
|
|
wave generated in the sample to be transmitted to the
gas where the PAS signal is generated.
Samples
with very low absorption or with very high absorption require special
considerations. Low-absorption
conditions are common in near infrared analyses and in low-concentration
measurements. In both cases, it is
necessary to use low mirror velocities to increase the sampling depth and the
fraction of the IR beam energy that excites the PAS signal.
In the case of very highly absorbing sample conditions, such as analysis
of adsorbates on carbon black, very high signal-to-noise ratio spectra are
required, which are then normalized using the method of self-referencing
described in the discussion on normalization of spectra.
If the signal-to-noise ratio is not sufficient, the bands of interest
will be dominated by noise when the spectrum is expanded on the ordinate axis to
observe these weak features. In
some instances, absorbance bands of adsorbates on very strongly absorbing
materials appear reversed in spectra as negative-pointing bands.
The mechanism for this has not been fully explored, but it appears that a
PAS signal from the black substrate, which produces a transmission spectrum
after the IR beam passes through the adsorbate, is larger than the PAS signal
from the adsorbate itself.
Microsamples,
such as films that are thinner than 2pL
and free-standing (gas on both sides), display magnitude and phase
responses that
are very different from thick samples. Figure
18 shows the calculated magnitude and phase signals as a function of modulation
frequency for a range of film thicknesses with an absorption coefficient of 104cm-1,
a typical value for a strong absorbance peak.
The magnitude signals are plotted as a ratio relative to the signal of a
2-mm-thick sample to best illustrate the signal enhancement effect common
to
thin samples. The enhancement
is due to the multiple passes that the thermal waves make within the sample as a
result of the very high reflectivity of solid-to-gas interfaces.
The multiple passes lead to multiple thermal transfers into the gas and a
larger signal. This enhancement
effect is what makes microsample analyses possible in PAS in spite of the very
small area that single fiber and single particle samples present to the
IR beam. Single fibers, 10
m
The
phase signals shown in Figure 18 also show the effect of multiple thermal-wave
reflection within thin samples. In
the case of phase, the signal is delayed in its evolution at lower frequencies
because 2pL
increases as the frequency decreases. This
results in many passes of the thermal waves back and forth within the
sample at
low frequencies before decaying away and a corresponding delay in the signal
evolution. The phase signal is seen
to vary between 90°
and 135° as expected for a single homogeneous layer.
A
final comment on thin film samples is that PAS spectra of these materials are
free of the optical interference fringes, which are observed in transmission and
reflection spectra. The absence of
fringes in PAS spectra makes it much easier to observe weak features due to
additives in polymer film spectra.
Gradient
and layered samples are discussed in detail in the second section and, in the
case of layers, in the next section, and are, therefore, not covered here.