|NDT.net March 2004 Vol. 9 No.03|
| 2nd MENDT Proceedings
PERFORMANCE OF A LASER-BASED ULTRASONIC DETECTOR USING CON FOCAL FABRY-PEROT INTERFEROMETERAbdulaziz M Aljalal, Ph.D.
King Fahd University of Petroleum & Minerals
Dhahran 31261, Saudi Arabia
Mohammed Aslam Khan, Ph.D.
Senior Research Scientist/Professor
Laser Research Section, Center for Applied Physical Sciences
The Research Institute,
King Fahd University of Petroleum & Minerals
Dhahran 31261, Saudi Arabia
SynopsisA detector based on a confocal Fabry-Pérot interferometer (CFPI) is being developed at KFUPM. This uses a He-Ne laser in conjunction with a scanning CFPI. The He-Ne laser and the CFPI have to be stabilized in unison to counter any drifts occuring due to thermal effects. Additional experimental aspects include synchronization with the ultrasonic pulse initially launched into the specimen, retrieval of the signal, and post-acquisition processing. The working principles of the detector are presentd along with some prelimiary results on laser mode structure and comparison with other non-contact detectors.
IntroductionLaser-based ultrasonic techniques are finding increasing use in non-contact inspection of material defects and risk assessments . These techniques involve launching of ultrasonic pulses into the specimen as well as detection of transmitted and/or reflected pulses. The generation of ultrasonic pulses at any surface is a simple and straightforward process. It involves rapid absorption of some laser energy in an area of sub-millimeter dimensions of the material where a local hot spot is formed while the remaining surface of the specimen remains essentially unaffected . The resulting thermal gradient is responsible for instantaneous creation of all sorts of bulk and surface ultrasonic pulses. Enhancements in the intensity of these pulses can be achieved by several techniques such as adding a surface layer of highly absorbing material or by using higher laser energies which evaporate a minute amount of surface material that is accompanied by oppositely directed much stronger ultrasonic pulse [i, ii]. The non-contact detection of ultrasound with laser, on the other hand, is a far more complex proposition. This is primarily due to very small amplitude of the emerging ultrasonic pulse and the prevailing noisy vibration-loaded ambience in the real-life environment .
Laser-based detection of ultrasound can be accomplished using any one of several different effects. For instance, the displacement of surface under ultrasonic stimulation can be monitored by a simple photo-detector positioned to monitor the intensity of the secularly reflected beam of a laser. Another way is to use a Michelson-type interferometer for monitoring the displacement of the surface.[1,2] However, for both these detectors, the environmental noise from sources such as machines in the vicinity or air-conditioning can seriously degrade the signal-to-noise ratio. Furthermore, the surface under examination should be fairly polished and hence reflecting. A third possibility that is not sensitive to the two factors noted above uses a CFPI. This monitors the velocity of the vibrating surface (rather than its displacement) through Doppler shift in the frequency of light scattered by the surface [2,5].
A project is in progress at KFUPM to develop a fully non-contact and remote laser-based ultrasonic system for material inspection. We have previously reported the use of laser generated ultrasound for inspecting partially submerged objects . In this paper, a detector based on a CFPI being developed in our laboratory is discussed. This uses a He-Ne laser in conjunction with a scanning CFPI. The working principles of the detector are presentd along with some prelimiary results on laser mode structure, and comparison with other non-contact detectors.
Working Principles of Confocal Fabry Perot InterferometerA CFPI consists of two identical spherical mirrors facing each other and separated by a distance equal to their common radius of curvature [7,8]. When illuminated from outside by a laser beam close to the axis, the light transmitted by the first mirror is reflected by the second mirror. At the same time, a part of the incident light is also transmitted by the second mirror. The reflected light goes back to the first mirror and undergoes a further reflection as well as partial transmission. The process of reflection back and forth between the two mirrors continues many times. Simultaneously, there are an equivalent number of transmitted rays both in the forward and back ward direction. Due to the confocal geometry of the two mirrors, a ray takes four reflections to come back to its starting point (Fig.1) Accordingly, the optical phase difference between any two consecutive transmitted rays is given by
where l, f and c are the wavelength, frequency, and speed of light, respectively; and d is the separation between the two mirrors. The transmitted intensity is 
Here Ii is the intensity of the incident laser beam and R is the mirror reflectivity.
Fig. 2 shows the intensity of transmitted light as a function of the phase difference for a mirror reflectivity of R=0.95. The intensity distribution consists of sharp maxima separated by regions of near-zero intensity. The sharpness of the peaks depends on the reflectivity of the mirrors; higher reflectivities, yielding sharper peaks.
Noting that the phase difference is a function of the separation d between the mirrors as well as the frequency f of the incident light, we can change the intensity of transmitted light by changing the separation between the mirrors or the frequency of incident light. It is important to remark that the maximum slope of the intensity profile occurs at the phase difference corresponding to half-intensity points of the transmission peaks. Accordingly, to obtain maximum change in intensity for a small change in frequency, we have to fix the distance d between the mirrors to correspond to the half intensity point.
A sensitive light detector can be used to monitor any variation in intensity of the transmitted light and this could form the basis for using CPFI as a detector of ultrasound (see Fig. 3).
Detecting ultrasonic pulses with CPFIBasically a laser beam is incident on the surface of a specimen to be probed and the light scattered by the specimen is fed into a CFPI. The CFPI is operated at the half-maximum intensity point. When the specimen is under the stimulus of an ultrasonic wave, the frequency of the scattered laser beam will be Doppler-shifted and this will result in intensity fluctuation of the transmitted light imaging the ultrasonic stimulus.
It is known that the size of the frequency shift Df is proportional to the velocity v of the specimen surface and it is related to the wavelength of the laser light l and the incident angle q by [i]
For a sinusoidal ultrasonic wave, the displacement of the surface d(t) is given by
Here, U is the amplitude of the displacement ultrasonic wave, fu is its frequency and Y is some phase constant. The accompanying shift in the frequency is
The phase shift df due to the frequency shift Df can be found from Eq. (1). Thus,
Usually this phase shift is much smaller than the FWHM of the transmission peak. Typical change of the intensity of transmitted beam of a CFPI operating at half-maximum intensity point in response to an ultrasonic signal of displacement amplitude of 1 nm is shown in Fig. 4 as a function of the frequency of the ultrasonic wave.
Special FeaturesThe main features of the CFPI detector of ultrasound waves are non-contact scanning of real life specimen, high spatial resolution and broadband response. The system can be used for inspecting specimens located at a distance of up to several meters. This should be compared with other known non-contact transducers such as electromagnetic acoustic transducer (EMAT) that can only work for some special materials e.g. metals and requires good surface finish and have to be very close to the specimen within few millimeters. [3,9] Furthermore, the CFPI detector can achieve high resolution limited by the spot size of the laser which can be focused down to a few micrometers. In addition, it provides a large bandwidth to enable detection of frequencies up to hundreds of MHz with a single detector. Broadband detection also facilitates discrimination between closely spaced pulses in time.
One major limitation is low sensitivity of CFPI compared with conventional resonance contact-type transducers even though a reasonable signal to noise ratio has bean achieved by CFPI detectors using higher power lasers and bigger scattered-light collecting optics. One further point to mention is the higher level of skills needed to operate CFPI detectors.
Figure 5 shows a schematic layout of the experimental setup. Basically it consists of a He-Ne laser, a CFPI with one mirror mounted on a PZT-driven actuator, polarizing optics and fast photodiodes. Our He-Ne laser operates in three longitudinal modes separated in frequency by about 750 MHz. The gain on these modes is continuously changing due to change in the length of the laser tube resulting from thermal effects. Luckily the third mode is usually much smaller in intensity compared to the other two. If the temperature of the laser tube is stabilized, this drift can be controlled to a great extent. Typically for frequency drift of 1 MHz, the change in the temperature of the tube should remain less than 0.7x10-3 ºC corresponding to a change in length of the tube of the order of 2 nm. Accordingly, very high precision temperature stabilization scheme is needed. At the same time, the CFPI has to be synchronized with the stabilized laser.
Thermal stabilization scheme:It should be noted that a strong ultrasonic pulse launched into the specimen by a pulsed laser will cause a Doppler shift of the order of 100 kHz . However, because of the selected position of the operation of the CFPI at half-maximum intensity point, a measurable change in intensity can be recorded for our laser stabilized with a fraction of the width of the transmission peak of the CFPI (about 6 MHz). Therefore a control of thermal drift to within 1 `MHz should be adequate for our system.
The circuit for thermal stabilization for the laser is based on the following physical principle. When there is a small change in the length of the tube, the intensity of two consecutive modes will change in different directions, one increasing the other decreasing. If we compare the difference between the intensity of these two modes and change the length of the tube to keep their difference fixed, then we can stabilize the laser (See Fig. 6) .
Since the modes of the laser are spatially overlapping, it is not possible to directly measure their intensity behavior. Luckily, the polarizations of any two consecutive modes are perpendicular to each other. This provides an easier method to separate these two modes. One way to achieve this is to use a polarizing beam splitter in conjunction with a half-wave plate (Fig 5). The intensity difference between the two emerging beams is fed, after amplification, to a heater wrapped around the laser tube. The current through the heater is increased when the length decreases and vice versa.
The original beam from the He-Ne laser is split into two beams one is sent to the specimen to detect the ultrasonic signal (signal beam) and the other is sent directly into the CFPI to tune it (reference beam). The intensity of the reference beam is measured by a photodiode and the maximum intensity of the reference beam is found by scanning the CFPI using its PZT drive. The output of the photodiode is compared to an external voltage signal which is adjusted at the half-maximum of the photodiode signal. The difference between these two signals is fed through a differential amplifier to the PZT drive. Initially, the difference between these two signals will be different from zero and accordingly the CFPI will be tune by the output of the amplifier until the difference between these two signals becomes zero. This is the automatic locking position for the half-maximum intensity point of CFPI .
The signal beam Doppler-shifted in frequency by the ultrasonic signal is fed to the locked CFPI and seen by another photodiode. Because of the difference in frequency between the signals in the absence and presence of the ultrasonic pulse, the intensity seen by the photodiode will be modulated when the ultrasonic pulse is present.
Typical resultsModes of our He-Ne laser
Figure 7 shows the modes of our He-Ne laser measured by a spectrum analyzer (Spectra Physics 470-03 free spectrum range = 2 GHz). The three different modes are clearly vsible separated by 750 MHz. It is noted that the drift rate which is reposible for changes in the intensity of each individual mode varies with time. This is primarily because initally the whole laser tube is at room temperature and it is heated very slowly by the elelctical discharge in the He-Ne gas mixture. Slowly the heat spreads through the entire tube and the temperature begins to stabilize typically after about 90 minutes. In particular, when the laser is first turned on, the thermal drift is rather huge, typically in the range of a few GHz/min. However, after 30 minutes, this drift reduces to about 50 MHz/min while after 90 minutes, this becomes about 5 MHz/min. By adding the thermal satiblization scheme, the time to reach a stable operating point was reduced to about 10 minutes. Eventually the drift was restricted to within 1 MHz.
Comparison of CFPI with other non-contact detectors of ultrasoundA set of typical traces showing the arrival of ultrasonic Rayleigh pulses using five different detectors including the CFPI, optical beam deflection (OBD), EMAT (in-plane and out-of-plane configurations) and Photo-refractive crystal is given Fig. 8. These are from an aluimum sample. The ultrasonic pulses were generated by a Q-switched Nd:YAQ laser delivering 45 mJ energy in 12 nsec pulses. While the EMATs appear to give a considereably better signal-to-noise ratio, it should be kept in view that the EMATs are very close to the surface and have much lower bandwidth. On the other hand, the three other detectors offer nearly the same level of signal-to-noise ratio. Futhermore, the CFPI is located much farther thatn any other detector.
ConclusionThe operating principles of a detector of ultrasound based on a confocal Fabry-Perot interferometer in conjunction with a He-Ne laser are discussed. The role of thermal stabilization in controling the mode drift is highlighted. A comparison of performance with some other non-contact detectors is also presented from previously published work.
AcknowledgementThe authors would like thank the Research Institute at King Fahd University of Petroleum Minerals for supporting this work through an Applied Research Incomes Grant API-018.