|NDT.net June 2004 Vol. 9 No.06|
Photoacoustic spectroscopy as an NDT tool for thin filmsS. Mahalakshmi, P. Palanichamy1 and K. Ramachandran
School of Physics, Madurai Kamaraj University, Madurai - 625 021.
1 DPEND, Indira Gandhi center for Atomic Research, Kalpakkam - 603 102.
Corresponding Author Contact:
AbstractThe technique of Photo acoustic spectroscopy is applied here for non-destructive testing (NDT) of thin films. The Photo acoustic signal shows a change in magnitude and phase wherever the crack or void or the defect is present in thin films.
Keywords: NDT, Photo acoustic spectroscopy (PAS), Thermal diffusivity, Photo acoustic absorption spectrum, NDE.
Introduction:The photo acoustics (PA) technique has been effectively used to characterize materials, because of its great versatility as a non-destructive and non-invasive method for evaluation of materials. Here Photo acoustic spectroscopy (PAS) is used in the form of Photoacoustic microscopy (PAM), which offers minimal or no sample preparation, the ability to look at opaque and scattering samples and the capability, to detect the defects on the surface of thin film .There is a limited information on the application of photoacoustic spectroscopy as a non-destructive method, eventhough it provides optical properties as well as the defect states associated with semiconductors effectively.
Cadmium sulphide, a semiconductor, has been an important subject of research, as it possesses many industrial applications. Nano CdS and thin film of CdS play vital roles in the fabrication of new Opto electronic devices. Before the fabrication of any device, it is necessary to characterize the material, particularly for surface defects. Several types of NDT methods, such as thermography and Ultrasonics, are well known to study the defects apart from the usual SEM and AFM. On the other hand, photothermally-based NDT methods, such as photoacoustic and photothermal imaging, are some of the candidates of new NDT methods of this decade [3,4].
In the present paper, the Photoacoustics technique is applied to view the surface defects on Cadmium sulphide thin film deposited on a glass substrate and to study quantitatively some thermal properties of this film for NDT.
In PA microscopy, the depth profiling analysis is carried out by measuring the PA signal at different modulation frequencies at different locations of the sample. Subsurface locations are then differentiated by the change of penetration depth of the heat wave (i.e., the thermal diffusion length ) with the modulation frequency.
The PAM enables to receive the images of subsurface structure of objects, which cannot be received by the traditional non-destructive evaluation methods. The PAM is used for detection and analysis of subsurface heterogeneities, such as defects, cracks, inclusion, delamination etc. Now PAM technique is also used to detect the non-uniformity in thin films. Most of the time the thickness of the films will be in the order of microns and so the defects, if exist, will be mostly from surface and since PA is very effective and useful for surface studies, any defects or non-uniformity in the deposition on the substrate can be quantitatively analyzed with PA spectroscopy. Since the scanning is a non-destructive testing, the thin film will not be affected in any other way. Recently Dillenz  have used ultra sound lock-in thermography to detect cracks in several materials. Here we have employed PAS to detect defects and a quantitative study on Cadmium sulphide thin film.
Conventional thermal wave methods are sensitive to all thermal boundaries and surfaces. Eventhough such studies on thick composite materials (Dilenz etal) are reported by Photoacoustics, works on thin films by these methods are rarely available in the literature and here we present our photoacoustics investigation on thin films.
Photo acoustics:In this technique, a sample is placed inside a specially designed cell called photoacoustic cell, containing a suitable gas and a sensitive microphone. The sample is then illuminated with chopped radiation as shown in Fig. 1. Light absorbed by the sample is converted in part into heat by non-radiative deexcitation processes within the sample. The resulting periodic heat flow from the sample to the surrounding gas creates pressure fluctuations in the cell, which is detected by the microphone as an acoustic signal which is phase coherent at the chopping frequency. The resulting photo acoustic signal is directly related to the amount of light absorbed by the sample .
Photo acoustic spectrometer:The present PA spectrometer consists of a 400W Xe-lamp (Jobin Yvon) as the light source. The sample is placed in the PA cell and the mike is placed very close to the sample. To get the modulated light, a mechanical chopper (Model PAR 650) is used with the source. The PA signal from the sample is fed to a lock in amplifier (Model Perkin Elmer 7225 DSP). The light is allowed to fall on the sample through a monochromator (Model Triax 180, Jobin Yvon) as shown in Fig. 1, when wavelength dependence is intended.
Photoacoustic Microscopy (PAM):This technique uses the fact that the thermal properties (specific heat capacity and thermal conductivity) of a non-uniform surface or a void buried inside a sample are different from the bulk of the sample. Since the PA signal from a sample is a strong function of the thermal properties of the irradiated region of the sample, the differences in the uniformity get reflected in the PA signal. In PAM, one scans the entire surface of the sample with an intensity-modulated beam of light using X-Y scanner and records the PA amplitude and phase as a function of position. Differences in thermal properties within the thermal diffusion length get reflected in the measured PA amplitude and phase. The usefulness of these methods results from sensitivity to thermal properties, such as diffusivity and boundary conditions, near the point of the surface being probed. Through this sensitivity to localized thermal properties, scanning PAM can examine beneath the surface of an opaque material for voids, closed or open cracks, and inclusions. The maximum probe depth is approximately 2mm, and the technique is particularly sensitive to flaws at depth from about 1 micron to 300 microns .
The information from this technique is limited to thermal diffusion length of the sample. However, by restricting the light modulation frequencies to low values, one can extend the thermal diffusion length to several hundred micrometers in samples with high thermal conductivity. Even though the technique is limited to physically thin samples, it finds special applications in non-destructive testing of thin bulk samples, thick films etc.
Calibration:To calibrate the set up, PAM is done on known material like glass. We have taken both a perfect sample and the other one with a fabricated defect. We found that at the places of defect the PA amplitude abruptly increases and in all other places the PA amplitude decreases as frequency increases. This is shown in Fig: 2a for glass. This unusual behavior namely the rapid increase in PA amplitude shows that there is a defect in the sample. It is known that whenever a defect is present in a perfect lattice , there will be an increase in absorption and consequently an increase in PA amplitude. This is repeated by allowing the chopped beam to fall on different positions of the sample. These measurements are shown in Fig: 2b. Here again the unusual rise in PA amplitude is seen at position C marked in the Fig: 2b. This proves the effectiveness of the technique.
Thin film:CdS thin film is a technologically useful material, as many devices based on CdS, including sensors have come up in the recent years. One of the most promising techniques for producing large areas of inexpensive CdS film for terrestrial photovoltaic application is spray pyrolysis and here we followed this method to synthesize the CdS film . The spray pyrolysis technique is an attractive, low- cost technique for uniform, thin film deposition over a large area with excellent mixing of the constituent elements. Eventhough this is a widely used technique before studying the thermal properties of thin films of CdS, grown by spray pyrolysis the uniformity and the defects, if any, should be first investigated.
Since CdS is highly absorbing, the photoacoustics measurements will yield better information than the other spectroscopic technique and the thermal diffusivity of the sample depends on the uniform deposition of the sample over the glass substrate. When the coating is non-uniform, the thermal diffusion decreases and subsequently the amplitude and phase of the photoacoustic signal will drastically change as of the beam determining depth and size of defect which determines the PA signal. The area of distortion corresponds to interval of depths, in which photoacoustic microscope will be sensitive to defect.
To start with, CdS thin film with uniform thickness is taken and the PA experiment is carried out. Here, as frequency increases the PA amplitude gradually decreases, even though it is not 1/f variation. This is shown in Fig: 2c. Then another CdS thin film sample is taken in which non-uniformity is present. We found that the PA amplitude increases when there is a defect on the surface of the film. This is shown in Fig: 2d for a specific case.
Results and discussions:Modification in Fig: 1 is done with X-Y mirror inserted between the chopper and the PA cell for the present measurements. It is manually operated in such a way that the beam falls on the sample at different points. With the Xenon lamp alone (no monochromator first) the photoacoustic signal is observed for different chopping frequencies, when a thin film is (with the glass as the substrate) placed inside the PA cell. Thermal diffusivity is then calculated from the characteristic frequency of the sample. (The characteristic frequency is defined as the frequency at which the sample goes from thermally thick region to thermally thin region). This is found out from the region where there is an 1/f dependence on the PA signal. If fc is the characteristic frequency of the sample of thickness l then the thermal diffusivity can be calculated from,
This is one of the important thermal property that will be affected by the presence of defects in thin films. Those results are given in Table: 1, for samples with and without defects. The comparison with other experiments for thermal diffusivity and thermal conductivity for CdS thin film is quite good for perfect samples, whereas for non-uniform thin film, the values are quite different.
Thus the non-uniformity in the thickness of the films grown and the presence of pinholes throughout the samples can be detected by this PA spectroscopy with good accuracy.
For example, if we compare the glass samples 1 and 2 in the Table: 1a, we find that there is a sharp increase in PA amplitude, the characteristic frequency and the thermal diffusivity where the modulated beam come across a defect in glass (this is done by scratching the glass with a sharp edge). A good glass sample 1 has a value of thermal diffusivity of 5.93x10-7m2/sec, which compares well with the literature value. But the glass with a fabricated defect on the surface (sample no.2) shows a value of 17.1x10-7m2/sec for the same diffusivity. This is clearly seen in the graph also (as in Fig: 2a). Similarly a glass with multiple defects is scanned at different locations. Whenever a defect is encountered, the characteristic frequency and the diffusivity increase rapidly as shown in Table: 1b. Except for sample 1, all the measurements shown in Table: 1b, correspond to a scratch (or defect) i.e., only when the beam encounters a defect there is an increase in the characteristic frequency. These measurements clearly show the effect of defects on the PA signal and hence the thermal diffusivity. Now, from Table: 1c, where the PA measurements are given for CdS thin film, an increase in the characteristic frequency from 40 Hz to 65 Hz is observed. This increase predicts the presence of the defect and so the increase in diffusivity.
All these CdS thin film samples are prepared by spray pyrolysis on glass substrate and for the present investigations, samples prepared at identical conditions are taken. That is, all the films are prepared at spraying time of 3hrs and of two layers thickness. The weights of these films were measured using a chemical balance and the accuracy is about 5%. Assuming no other defects, fine scratches were made on these films so that defects are artificially made. So the present PA measurements are specifically for the defects present, which are seen at the PA amplitude and the thermal properties of the film.