|NDT.net March 2004 Vol. 9 No.03|
| 2nd MENDT Proceedings
Thin Transparent Films as Laser-Based Ultrasound DetectorsAbdulaziz M Aljalal
King Fahd University of Petroleum & Minerals
Dhahran 31261, Saudi Arabia
Corresponding Author Contact:
SynopsisIn recent years, a new generation of ultrasound detectors is being developed to be capable of nondestructively examining near-surface blood vessels and other tissue structures in human beings. This kind of contact sensors is based on a very thin film used as a Fabry-Perot interferometer addressed remotely by a laser beam. It can have a very broadband response in the range of tens of MHz and it can achieve phenomenal spatial resolution in the range of tens of micrometers. The physical principles behind this type of detectors will be presented along with some practical suggestions to improve its sensitivity. This type of sensors has a great potential in the field of nondestructive test of materials as well. An experimental demonstration of using a thin film sensor on a metallic surface will be given and its performance will be compared with that of a conventional PZT resonance transducer.
IntroductionUltrasonic pulses can be generated inside human bodies by very short light pulses from a low energy focused near infrared laser. Due to relative transparency of human tissues to near infrared wavelengths, a laser pulse penetrates several centimeters deep inside the body. When the laser pulse is absorbed by some tissues, for example blood, these tissues are heated up for a very short period of time and, as a result, they expand and contract over a very short period of time. Consequently, a wideband ultrasonic pulse is generated . A new attractive type of ultrasound sensors based on transparent thin polymer films is being developed to detect these weak ultrasound pulses generated inside the body .
Transparent film ultrasound sensors are attractive because they can be chosen to be also transparent to the exciting pulsed laser and hence exciting and detection can be made on the same spot and form the same side. Moreover, they have very broadband responses in the range of tens of MHz, and they can achieve very high spatial resolutions in the range of tens of microns .
This paper discusses the physical principles of these sensors and explores the feasibility of using them in another ultrasound sensing applications, namely, nondestructive testing of materials.
Principle of operationA typical thin-film ultrasonic detector consists of a transparent thin film, a laser beam and a fast light detector as shown in Fig. 1. The intensity of the light reflected towards the light detector depends on the optical phase difference between the beam reflected from one surface of the film and the other beam reflected from the other surface of the film. The optical phase difference is proportional to the thickness and refractive index of the film. Since an ultrasonic wave passing through the film modulates the film thickness and its refractive index, the ultrasonic pulse also modulates the intensity of the reflected light. Thus, the presence of the ultrasonic wave in the film can be monitored simply by monitoring the intensity of the reflected light.
The overall sensitivity of this detector can be expressed as a product of two sensitivities. The first sensitivity measures the ability of the sensor to convert the ultrasonic energy into change in optical phase and it will be called the acoustic-phase sensitivity As. The second sensitivity reflects the ability of the sensor to convert the optical phase shift into change in light intensity and it will be called the phase-intensity sensitivity Is. The acoustic-phase sensitivity depends on the elastic and photo-elastic properties of the film as well as on the properties of its surroundings. Polymers such as parylene and polyethylene Terephthalate (PET) have relatively large acoustic-phase sensitivity compared, for example, with glass or other dielectric films. The phase-intensity sensitivity depends on the power of the laser beam, reflectance of the light from each of the film surfaces and on the optical quality of the film surfaces.  Thus the overall sensitivity S is
In the following, two simple models will be presented to explore different factors affecting the performance of this detector. The first model focuses on the main factors affecting the acoustic-phase sensitivity while the other model focuses on the main factors affecting the phase-intensity sensitivity.
Acoustic-phase modelThe acoustic-phase model consists of a sinusoidal ultrasonic pressure wave incident on a thin film of thickness L which has a very large surface area (See Fig. 2). Let the incident pressure wave travel from the negative x direction with an amplitude of P0 and a wavelength of la. The side from which the ultrasonic wave coming will be called the substrate side and the other side will be called the backing-material side. The ultrasonic wave affects both the thickness of the film L and the refractive index of the film n. However, the effect on the refractive index is much smaller than that on the thickness for polymer films , and hence, only the effect on the thickness will be discussed.
It can be shown that the amplitude of the change in the thickness |dL| this sinusoidal ultrasonic pressure wave is given by 
Here ka=2p/la is the acoustic wave number, E is the Young's modulus of the film. T0, RL, and R0 are the pressure amplitude transmission and reflection coefficients at x = L and x = 0. They are related to the acoustic impedance of the substrate Zs, the acoustic impedance of the film Zf, and the acoustic impedance of backing material Zb by
Fig. 3 shows the change in the thickness as a function of frequency due to a sinusoidal wave with an amplitude of 1 kPa for three different films of thicknesses 25 µm, 50 µm, and 100 µm. These calculations are performed for a PET film facing water and backed by glass.
It can be noticed form Fig. 3 that the thicker the film the more responsive the film is to low ultrasonic frequencies. This should be expected since in the limit of very low ultrasonic frequencies, the ultrasonic wavelengths are very long such that the film can be considered under constant pressure gradient, and for any elastic material under constant pressure gradient, its elongation is proportional to its original length. Also, it can be noticed that each curve in Fig. 3 has local maxima and minima. The position of these maxima and minima are inversely proportional to the thickness of the film. The film of the least thickness has the largest bandwidth response. The positions of the local maxima occur when the thickness of film is equal to an odd number of quarter wavelengths of the ultrasonic wave while, the positions of the local minima occur when the thickness of the film is equal to half wavelengths of the ultrasonic wave. Since the acoustic impedance of water is much smaller than that of the film and the acoustic impedance of the film is much smaller than that of the film, the acoustic resonances of the film is similar to that of an air column with one end closed and the other end open . The responses that correspond to an odd number of half wavelengths of the ultrasonic wave, such as 11 MHz for 100-µm film, is not close to zero because the mismatch in impedance between the film and its backing material is not very large and hence the incident wave on the interface is only partially reflected and the reflected wave does not completely cancel the incident wave.
Eq. (2) indicates that if the ratio of change of the thickness of the film to its original thickness |DL|/L is plotted versus kaL, a universal curve is obtained which does not depend on the thickness of the film but depends only on the film and its surrounding materials.
Fig. 4 shows the effect of using different backing materials, namely, glass, air, and water. Here, the substrate of the PET film is water and the film has a thickness of 100 µm. For the case of air, RL ~ -1, and the incident wave on this interface is almost completely reflected and it acquires a phase shift of phi upon reflection. In the limit of low frequencies, the pressure gradient across the film is very small since the effect of the incident wave is almost completely cancelled by the reflected wave.
Since the acoustic impedances of water and air are much smaller than that of PET, the acoustic resonances behavior is similar to that of air column with both ends open . The positions of the local maxima occur when the thickness of film is equal to an odd number of half wavelengths of the ultrasonic wave.
The phase-intensity model is the second model that will be used to elucidate major factors affecting the phase-intensity sensitivity. It consists of a plane optical wave incident at an angle on a film with prefect plane parallel surfaces coated with thin dielectric layers to control the optical reflectivity on both film surfaces, as shown in Fig. 5. At each interface some light is reflected and some light is transmitted. The intensity of the reflected light resulting form multi-reflections inside the film is given by
where, Ii is the intensity of incident optical wave, F is the coefficient of finesse of the cavity formed by the two surfaces of the film and it is related to the reflection coefficient, r, by
is the phase difference between any two successive reflected beams. L is the thickness of the film, qt is angle of refraction, nf is the refractive index of the film and l is wavelength of the light. 
Fig. 6 shows the reflectance from the film (I/Ii) for three different surface reflectivity r2 as a function of the phase shift d. It is clear for the figure that the higher the surface reflectivity, the steeper the slope of the reflectance and hence the more pronounced the change in intensity due to a small change in the phase shift which is caused by either variations in film thickness or light wavelength. Thus, one way to increase the sensitivity of thin-film sensor is to use films with very high reflective surfaces. Table 1 shows the gain in sensitivity compared to an uncoated film as a function of surface reflectivity. The uncoated film surface reflectivity is assumed to be 0.04. However, there are important practical limitations which limit the values in table 1 quoted for the idealized case. These limitations include the quality of reflecting surface, parallelism of the two surfaces and diffraction of light beams of finite beam diameters. In one hand, the demand on the quality of the surface and parallelism of the two surfaces becomes more sever for laser beams with large diameters. On the other hand, the diffraction of the beam might be the limiting factor for laser beams with very small diameters. The angle of diffraction is inversely proportional to the light beam diameter. Even for a high quality film, one starts to observe deviation from the idealized case for surface reflectivity values as low as 0.95. Fig. 7 shows the reflectance of a 100-µm glass film which has surface flattens and parallelism better than l/100 over 10 mm diameter ( l=633 nm) for laser beam diameter of about 60 µm. It can be seen that the reflectance does not go to zero, in contradiction to what is predicted by the idealized model. It can be shown for this beam diameter that the main cause of this degradation can be attributed to the imperfect parallel surfaces.
Another way to improve the sensitivity of this detector is to increase the power of the probe laser. As can be seen from Eq. 4, the intensity of the reflected beam, I, is proportional of the intensity of the incident beam Ii. Consequently, the slope of I versus d curve is proportional to Ii. Since the noise associated with a light beam is proportional to the square root of its intensity , the overall increase the signal to noise ratio is proportional to the square root of the laser beam intensity
Thin film on aluminum plateIt would be very interesting for nondestructive testing of materials to demonstrate the ability of the thin-film ultrasonic sensor to detect ultrasonic pulses originating from a piece of metal, and to compare its performance with conventional PZT transducers. Fig. 8 shows our experimental step. A laser beam from a tunable external cavity diode laser, Environmental Optical Sensors (EOSI) model LCU-2001-A, is guided using two mirrors through a coupler to a single mode optical fiber which is used to clean up the spatial distribution of the laser beam to produce a TEM00 Gaussian beam. The laser can be tuned from 770 nm to 790 nm with a speed of 1nm/7sec. In addition to controlling the electrical current of the laser, the power of the laser beam emerging form the fiber can be controlled by density filters (DF) and slightly misaligning the laser beam incident on to the coupler. The beam coming out from the fiber is collimated and sent into a beam splitter (BS), which splits the beam into two parts, one moving straight into the thin film and the other is reflected into a reference power meter (the reference detector).
The reference detector is used to monitor the power of the laser. The beam incident on the film is focused by a lens of 6 cm focal length, and a portion of the beam reflected back from the film is directed with the help of the beam splitter into a 35-MHz bandwidth photodiode detector (the signal detector). Signals from the reference and signal detectors are collected by a digital oscilloscope. A CDD camera is used to study the spatial distribution of the laser beam. The digital oscilloscope and the camera are interfaced to a laptop. The transparent film, a PET film of 120 µm thickness, is mounted on one side of an aluminum block using a coupling material (more bout this later). On the other side of the aluminum block, a PZT ultrasonic transducer with 1" diameter and 3.5 MHz resonance frequency is mounted. Johnson and Johnson K-Y lubricating Jelly is used as the coupling material between the aluminum block and the transducer. The transducer is powered by Panametrics pulser-receiver model 500PR. The film, the aluminum block, and the transducer are all mounted on a rotational-translational stage to control the position and direction of the film relative to the incident laser beam.
Fig, 9 shows four curves of the powers of different laser beams as a function of the wavelength of the laser. Curves A, B, and C are for a film hanged in air, which is not in contact with any coupling materials. Curve D is for a film in contact with a coupling material from one side, and air form the other side. Curve A shows the power detected by the signal detector while curve B shows the power detected by the reference detector. The modulations in the power are due to some mechanisms by which the laser is tuned in wavelengths. To minimize these modulations in the laser power and get a smooth curve, the signal form the signal detector is divided by that from the reference detector as shown by curve C.
Curve D in Fig. 9 shows the degradation of the slope when the film is in contact with the Johnson and Johnson K-Y lubricating Jelly coupling material. This worsens the sensitivity of the thin-film sensor. The reason for this degradation is that the refractive index of the coupling material is quite close to that of the film. This lowers the reflectivity from the surface since the optical reflectivity is proportional to the difference between the refractive indices of the materials on both sides of the surface. The coupling material is essential to couple the ultrasound pulses from the aluminum block to the film and a solution is needed to increase the sensitivity of the sensor in the presence of the coupling material. One solution is to coat the film with reflecting coatings such as very thin layer of silver or dielectric layers. This might be an expensive solution. Another solution is to use colored coupling materials such as colored paints in hope that some of the light will be reflected due to the color of the paint. Since we are working with wavelengths around 780 nm, which is near infrared, one would guess that red paint might provide enough reflectivity in the near infrared to get reasonable slope. We used acrylic cadmium red from Liquitex. The maximum slope increases by a factor of 13 compared with the case when the jelly coupling material is used.
An example of an ultrasonic signal is shown in Fig. 10. This signal is obtained by tuning the laser to a point of maximum slope on the reflectance-wavelength curve, and an ultrasonic pulse from the PZT transducer. The signal corresponds to a change in the thickness of the 120-µm film by 0.3 nm and is obtained by averaging over 128 traces. The signal-to-noise ratio is about 20 and it should be compared with the signal-to-noise ratio of 30 obtained by our resonance PZT transducer from one trace. Thus the sensitivity of our PZT transducer is about 17 times larger than that of the uncoated thin film. It should be noticed that a performance comparable with the PZT transducer can be obtained achieved even with out coating the film with reflecting materials. For example, for ultrasonic pulses of 3.5 MHz frequency, an increase in the signal-to-noise by a factor of seven can be accomplished by using a 320-µm thick film instead of the 120-µm thick film. Moreover, an increase by another factor of three can be obtained by increasing the power of the incident laser on the film form 1 mW to 10 mW.
ConclusionThe physical principles behind a laser-based transparent thin-film ultrasonic sensor are highlighted. Different factors affecting the performance of this sensor are illustrated with the help of two simple models. An experiment on detecting ultrasonic pluses produced by 3.5 MHz PZT ultrasonic transducer by a 120-µm uncoated PET film is presented and compared with the performance of our PZT ultrasonic transducer.
AcknowledgementsThe author would like to thank Dr. Paul Beard, University College London, for the many discussions about thin-film sensors and for allowing him to use his lab during the summer of 2003 to do the experiments related to this work. Also, he would like to thank King Fahd University of Petroleum and Minerals, and the Postdoctoral Summer Research Programme 2003 (funded by BAe Systems / Al-Yamamah Programme and administrated by the British Council) for the financial supports to visit Dr. Beard.