| NDT.net - April 2002, Vol. 7 No.04 |
Thermo-mechanically induced deformations and stresses, especially shear stress, are one of
the key issues concerning fatigue and wear of microsystems. In this paper we report on the
development of a combined set-up for measuring temperature by a thermographic camera as
well as deformation gradients by 3D-shearography. The 3D-shearographic camera based on a
multi-wavelength set-up is adapted to the microscope optics in order to view the same area of
interest and the same state of operation of the sample under test as the thermographic camera.
The measurement system works for quasi-stationary states.
The system will be used for characterisation of microsystems under real operating conditions,
without being restricted to destructive sample preparation. We will present preliminary results
of measurements and compare temperature distribution with thermo-mechanical deformation
gradients.
Systems and products are more and more miniaturised and integrated to microsystems. Current developments in microelectronics include an increasing number of joints per surface area. Higher semiconductor integration in combination with larger dice is rapidly increasing the power dissipation per volume unit. Packaging related reliability becomes increasingly important. Thermo-mechanically induced deformations and stresses, especially shear stress, are one of the key issues concerning fatigue and wear.
We expect a combination of optical methods for deformation analysis [1] and thermography to be best suited for thermo-mechanical investigations on these issues. Infrared detectors can quantify the transmission of heat between e.g. the substrate and the IC through the bonding and assess the temperature distribution on the sample. This heat conduction or temperature field would be correlated to the bonding quality (a lack of adhesion should decrease drastically the heat transfer for that bonding). As heat diffusion to the surface is dominant for IR-detectors, ESPI or shearography is thought to be sensitive to the same effect, but detecting thermally induced deformation rather than temperature distributions. A key issue, to our opinion, is the measurement of deformations and temperature under real operating conditions, without being restricted to destructive sample preparation.
In this paper we describe a combined set-up for measuring temperature by a thermographic camera as well as deformation gradients by 3D-shearography [2]. Imaging optics for both systems are truly collinear, but with separate beam paths and not interfering.
The 3D-shearographic camera was developed in the Brite EuRAM project Multi-Wavelengths Shearography (MuWaS) [3]. To acquire all information necessary for strain analysis three IR lasers with different wavelengths are used to illuminate the object simultaneously from different directions. The sensor head houses three synchronised B/W-CCD-cameras, each of which grabs the speckle fields of a single laser only. A standard temporal phase stepping procedure is applied. This allows for simultaneous measurement of three phase maps corresponding to three components of deformation gradients. Diode lasers from Spectra Diode Labs (SDL) emitting at wavelengths of 810, 830 nm and 850 nm are used. They operate in single longitudinal mode with an output power between 90 and 200 mW. Object illumination is performed without using any collimating optics.
![]() Fig 1: Shearographic 3D-camera with beam splitter BS, bandpasss filters Fi and quarter waveplate l/4. Using a liquid crystal device (LCD) and a polarising beam splitter PBS two shear directions (via mirrors M2 and M3) can be selected. Details are given in the text. |
The camera set-up is given schematically in Fig.1. A f/1.9 CCD lens with a focal length of 35
mm (Schneider/Germany) serves as entrance lens of the sensor head. A Michelson
interferometer serves as the basic shearing element. In order to implement the phase stepping
procedure, the reference beam mirror M1of the Michelson interferometer is mounted on a
piezo actuator.
For switching of the shear directions a method developed by ILM was used, where a liquid
crystal device (LCD), a polarising beam splitter and an additional mirror M3 were added to
the Michelson interferometer [4]. The mirrors M2 and M3 induce a shear along the x- and y-direction,
respectively. Hence, the two polarisation directions separated by the PBS are
sheared in different directions. The LCD acts as a switchable l/2 wave plate. In its neutral
state the LCD does not affect the polarisation. When switched on, the LCD rotates the
polarisation by 90°.
Another f = 35 mm transfers the image of the entrance lens at unit conjugate ratio through the
colour separation optics to the CCD cameras. The horizontally polarised light passes the
polarising beam splitter cube (PBS), and the band pass filter F2 centred at 850 nm transmits
this light onto CCD 2. The vertically polarised light is directed towards CCD 1. A band pass
filter F1 with centre wavelength of 810 nm transmits the light of the corresponding diode and
reflects all other wavelengths towards CCD 3. This combination yields high speckle
intensities on all three CCD cameras. For partially depolarised light, the filters ensure that
each camera records a speckle image corresponding to one laser source only, preventing ghost
images of the other laser wavelengths. The diode lasers were temperature tuned to the centre
wavelengths of the respective band pass filters, which have a FWHM of Dl = 10 ± 2 nm and a
high transmission of approximately 80-90% at their centre wavelengths.
The speckle patterns are recorded with 3 synchronised b/w OEM board CCD cameras
connected to the RGB input channels of a colour frame grabber. For image acquisition we
used the IC-PCI from Imaging Technology. The code for the image acquisition and
processing board was included as a dynamic link library in a LabVIEW environment
(National Instruments), which was used for controlling the shearographic system and the
thermography system.
The thermography camera is part of a LN2-cooled infrared scanning microscope (Computherm III, EDO Barnes) operating in the wavelength range of 3–5 mm, and is used in reliability assessment of microelectronic components down to a size of 2x2 mm2 . It features scan sizes of up to 128x128 points. Calibration of emissivity is done in each pixel using a heater stage (Peltier element). The measurement system works for quasi-stationary states only, since the scan time for one shot is about 2.5s – notably more than for the acquisition time of the full 3D-shearographic information.
The shearographic camera is attached to the viewer optics of the thermoemission microscope in order to observe the same area of interest and the same state of operation of the sample under test as the thermographic camera. Fig. 2 shows the schematic arrangement of the two systems, while Fig.3 gives an overall and a detailed view of the set-up.
![]() Fig 2: Experimental set-up, schematic. The IR ring optics transfers radiation towards the IR-camera; the laser light is imaged through the axial aperture and mirror onto the shearographic sensor. Two of the three illuminating diode lasers are shown. Details are given in the text. | ![]() Fig. 3: Experimental set-up, overall view (top). In the detailed view, the three illuminating diode lasers and the object in the centre are shown (bottom). Lasers are arranged at the corners of a rectangle with 170 x 230 mm2. |
The test device is loaded with voltage/current. At each loading step the thermal equilibrium is waited for, after which time the phase fringe patterns of the three lasers are measured. Transformation to the in- and out-of-plane phase maps is done for each successive pair of images. The following figures show preliminary results of measurements for a heater foil. Fig. 4 shows the temperature distribution on the resistor band (10V, 0.2A).
Fig. 4: Temperature distribution at equilibrium for
a resistive heat foil. Area scanned in
128x128 points is 16x16 mm 2 . Temperature
span is approximately 10°C. |
Fig. 5–7 show the corresponding phase gradient maps. The sensitivity is calculated by simulation software, taking into account the laser wavelengths and geometry of the set-up. This results in a sensitivity of 0.58, 0.81, and 0.28 mm/fringe for the u, v, and w-component of the deformation, respectively. The shear distance is approximately 15% of the image width, corresponding to about 3mm (both directions). Therefore, the gradients in Figs. 5–7 are scaled to 0.19, 0.27, and 0.09 parts per thousand per fringe.
![]() Fig 5: Deformation component u, sheared in x- (left) and y-direction (right). | ![]() Fig. 6: Deformation component v, sheared in x- (left) and y-direction (right). | ![]() Fig 7: Deformation component w (out-of-plane component), sheared in x- (left) and y-direction (right). |
First results with the combined shearography and thermography measurement system are very promising. Although it is in the present stage not suitable for real micro-measurements on system diameters below 1 mm, it will yet have a strong application in thermo-mechanical and reliability analyses of microelectronic components and packages. Future improvements will on one hand concentrate on dynamic measurements in order to assess heat dissipation effects and load cycling. On the other hand, we think of replacing shearography by ESPI in order to have a better localisation of the deformation with respect to the object geometry.
The development of the shearographic sensor was funded within the Brite/EuRAM Project MuWaS by the European Commission and the Ministry of Education and Science (BBW) of Switzerland.
This Paper was presented at Fringe 2001 "The 4th International Workshop on Automatic Processing of Fringe Patterns" held in Bremen, Germany, 17-19 September 2001. Proceedings edited by Wolfgang Osten, BIAS, Germany. Please contact Wolfgang Osten for full set of proceedings at wolfgang@uni-bremen.de.
| © NDT.net - info@ndt.net | |Top| |