|
V. GREGORIOU
Polaroïd Corporation
Photoacoustic Spectroscopy
The origins of Photoacoustic Spectroscopy (PAS) date
back to the discovery of the photoacoustic effect by Alexander Graham
Bell in 1880. Bell found that when light was focused onto thin diaphragms,
sound was emitted. In latter experiments, Bell studied the sounds produced
by the irradiation of various solid samples in a brass cavity sealed with
a glass window.
Practical use of the photoacoustic effect for condensed phase materials
had to wait for advances in instrumentation and theory. In 1973, PAS was
rediscovered simultaneously by A. Rosencwaig at Bell Laboratories and
by A. G. Parker at Johns Hopkins University. A general theory for the
photoacoustic effect was developed by Rosencwaig and Gersho and is commonly
referred to as the RG Model.
Photoacoustic spectroscopy is now commonly used in the analysis of a
variety of materials. It is a non-destructive technique that is applicable
to almost all types of samples. It offers minimal or no sample preparation,
the ability to look at opaque and scattering samples, and the capability
to perform depth profiling experiments. PAS can be used for both qualitative
and quantitative analysis. In particular, depth profiling experiments
are also useful for the characterization of surface-coated and laminar
materials and for studies of weathering, aging, curing, and the diffusion
of species into or out of a polymer matrix.
Introduction to Theory of PAS
In PAS the transformation of an optical event to an acoustic one takes
place. Initially, modulated light is absorbed by the sample located in
a sealed cell. The non-radiative decay of this absorbed light produces
a modulated transfer of heat to the surface of the sample. This modulated
thermal gradient produces pressure waves in the gas inside the cell that
can be detected by the attached microphone. This microphone signal, when
plotted as a function of wavelength, will give a spectrum proportional
to the absorption spectrum of the sample. One main advantage of PAS is
the ability to get information about the depth in the sample of the absorption.
The amount of the sample contributing to the PA signal is proportional
to the thermal diffusion depth. This thermal diffusion depth m, is inversely proportional to the modulation
frequency f. For example, for a typical organic polymer, a high
modulation frequency will probe a thin layer (f = 1 KHz; m
~ 4 mm,) while
at a low modulation frequency, a much thicker layer is probed (f =
100 Hz; m ~ 16 mm).
Figure 1 shows a depiction of the PA signal generation as it relates
to depth profiling. The model sample illustrated in this figure has a
thermally thin surface layer (thickness << m)
on a bulk substrate. After the light has been absorbed, the heat has to
diffuse from the point of absorption to the surface of the sample to be
detected. Since this thermal diffusion is a slow process relative to the
light absorption and non-radiative decay, an absorption in the bulk will
have a phase lag between the time of absorption and the thermal signal.
However, a surface absorption should not have a phase lag since the heat
doesn't have far to travel to generate the detected
pressure change in the transfer gas.
Figure 1. Photoacoustic Signal Generation
Experimental Considerations
For PAS, a FT-IR spectrometer capable of operating
in both continuous scan mode and in step-scan mode can be used. As will
be shown in the next section, the step-scan mode works very well for depth-profiling
experiments. A photoacoustic cell is generally available as an accessory
from most major suppliers of FT-IR instruments. Helium gas is used as
the transfer medium in the cell and to purge water vapor and carbon dioxide.
A 60% carbon black-filled polymer is often used as a reference sample.
Continuous Scan Mode
Most commercially available FT-IR spectrometers utilize the continuous-scan
mode of operation, where the moving mirror is scanning at constant velocity.
This type of scanning works very well for routine measurements. In the
continuous-scan mode of interferometry the laser fringe counter is used
to sense the accuracy of the scanning velocity. If a deviation is sensed,
correction signals are generated that assure the proper operation (constant
velocity). The consequence of this mode of operation is that each infrared
wavelength (l), is modulated
at its own particular Fourier frequency, given by the following Equation
1:
where v is the mirror velocity. Continuous-scan FT-IR
is the technique of choice when static spectral properties are determined.
Co-addition of successive scans increases the signal-to-noise ratio (S/N)
by a factor proportional to Ö t, where t is the
time that the signal is averaged at each collection point.
Step-Scan Mode
In step-scan FT-IR data are collected while the retardation is held constant
or is oscillated about a fixed value. Two types of experiments are possible
with step-scan interferometry. One type is the time-domain or time
resolved experiments where data are collected as a function of time at
each mirror position. The other type of experiment is the so-called frequency-domain
or synchronous modulation experiments. In these experiments, either amplitude
modulation (AM) or phase modulation (PM) can be used to modulate the intensity
of the infrared radiation in order to generate step-scan interferograms.
For PAS, phase modulation is used. Phase modulation is accomplished by
oscillating one of the interferometer mirrors about each step position
at a specified amplitude and frequency. To detect the phase modulated
signal, either a lock-in amplifier or a digital signal processor (DSP)
is required. The advantages of step-scan operation include the ability
to apply virtually any modulation frequency to the infrared radiation
and to carry out multiple modulation experiments. Since the frequency
of modulation is not a function of any retardation velocity (e.g., mirror
scan speed), there is no dependence on radiation wavelength. An advantage
of lock-in amplifier or DSP detection is the easy retrieval of the signal
phase. This is possible due to the fact that the beamsplitter (instrumental)
phase is identical for the in-phase and quadrature (90° out of phase)
components of the signal. These components are easily obtained as the
in-phase (I) and quadrature (Q) outputs of a two-phase lock-in amplifier.
As a result, not only the magnitude M, but also the phase F
can be easily obtained by following Equations 2 to 5:
(2)
F =
arctan(Q/I) (3)
I = M cosf (4)
Q = M sinf (5)
The advantages of Step-scan vs. continuous Scan
for PAS
In step-scan FT-IR spectroscopy, the same modulation frequency can be
applied to the entire spectral range. This is particularly useful for
photoacoustic (PA) detection since, in the absence of saturation, the
PA response at all wavelengths in step-scan FT-IR PA spectroscopy will
correspond to the same depth in the sample. In a typical continuous-scan
FT-IR PA spectrum each wavelength corresponds to a different sampling
depth. This results in inefficient probing at shorter wavelengths.
A step-scan FT-IR PA spectral depth profile can be obtained by changing
the modulation frequency for different scans. Alternatively, depth information
can be obtained from the phase of the PA signal. The step-scan mode of
data collection permits easier access to the PA phase than does the continuous-scan
mode. The combined use of modulation-frequency variation and analysis
of the phase allows a wider range of depths to be probed. This type of
spectrally resolved depth information is particularly useful when applied
to polymeric materials with special surface properties.
Depth Profiling
One method for depth profiling uses the phase of the PA signal to distinguish
between depths of different absorptions. Theoretically, this can be done
two different ways, by using the phase spectrum or by looking at different
components of the signal by detecting at different phases with respect
to the infrared light modulation. It has been shown
theoretically that the signal from a weakly absorbing thermally thin layer
should be separated by 45° from the signal of the weakly absorbing
substrate.
In order to eliminate instrumental phase contributions, a reference phase
must be established. This is done by placing a complete absorber into
the PA cell and rotating the phase of the lock-in until all of the signal
is in one channel of the lock-in, (e.g. in the quadrature channel, Q).
By replacing the reference with the real sample and collecting the in-phase
(I) and quadrature (Q) spectra, the bulk signal (B/Ö 2) should be in one channel (I), while information proportional
to the surface signal (S+B/Ö 2) should be in the other channel (Q). As a consequence, the surface
signal can be found by rotating the original interferograms by 45o and
obtaining S/Ö 2. Figure 2 illustrates the relationship between the various components
of the step-scan photoacoustic signal.

Figure 2. Photoacoustic phase relationships for a
weakly absorbing thermally thin layer.
Examples of Depth Profiling
The sub-micron resolution capability of the technique has been demonstrated
for a coated system having a sub-micron layer as the top layer. The system
under study was a micron thick multi-layered structure of a mixture of
acrylic polymers coated on a poly (ethylene terephthalate) (PET) base.
The base is several tenths of microns thick, substantially thicker than
that of the coated layer.
Figure 3 shows the evolution of the photoacoustic signal as a function
of the degree of rotation for the experiment performed with 750 Hz phase
modulation. The evolution of the overtone of the carbonyl band at 3428
cm-1 that belongs to the PET substrate can be clearly followed as we move
to phase delays that correspond to deeper depths. In addition, it is also
easy to see the appearance of the C-H aromatic stretching modes above
3000 cm-1, as the phase rotation angle increases. These bands were due
to the PET base and, as expected, their intensity increased as the overall
signal was dominated by the contribution of the polyester substrate. These
results demonstrate the ability of technique to completely separate the
IR signatures of the top layer from the bulk material.
Figure 3. Evolution of the photoacoustic signal as
a function of degrees of phase rotation, showing spectra originating
from different depths: 750Hz phase-modulation frequency; 0.5 lHeNe
phase-modulation amplitude.
|