LED based infrared scanning white light interferometry (IR-SWLI) permits non-destructive imaging of embedded
MEMS structures. We built an IR-SWLI instrument featuring a custom-built IR-range LED-based light source, capable
of stroboscopic use. The source combines multiple separately controllable LEDs with different wavelengths into a
collimated homogenous beam offering an adjustable spectrum. We employ software-based image stitching to form
millimeter-size 3D images from multiple high magnification scans. These images delineate three layers in a MEMS
cavity covered by silicon and reveal a micron-size inlet inside the channel.
Scanning White Light Interferometry (SWLI) provides high vertical precision for measuring step-like structures in
microelectromechanical systems (MEMS). The SWLI performance depends on its light source. A rapidly modulated
light source with a broad bandwidth inside the infrared (IR) region is necessary to measure layered MEMS that move.
Typical SWLI light sources - light emitting diodes (LEDs) and Halogen (HG) bulbs - fulfill only one of these
requirements.
To overcome this shortcoming we equipped our SWLI setup with a supercontinuum (SC) light source produced by
Fiberware Gmbh (Ilum 100 USB II). We tested our setup by measuring in plane and out of plane oscillating thermal
bridges with visible light, as well as top and bottom surfaces of silicon structures using IR light. The wide SC spectrum
creates localized interferograms. This allowed us to measure top and bottom surfaces of a thin (4 μm) bridge. The
stroboscopically measured profiles of oscillating thermal bridges were comparable to those measured using a white LED.
The results of static measurements were similar to those achieved with an HG lamp.
Solid state light sources are replacing a tungsten filament based bulbs in Scanning White Light Interferometers. White
LEDs generate little heat, feature short switching times, and have long lifetimes. Phosphor-based white LEDs produce a
wide spectrum but have two separate peaks which cause interferogram ringing. This makes measuring multi layered
structures difficult and may degrade measurement precision even when measuring a single reflecting surface. Most non
phosphor white LEDs exhibit a non Gaussian spectrum, but multi-LED based white LEDs can achieve switching times
and stability similar to those of single color LEDs. By combining several LEDs and by controlling their input current
independently it is possible to create almost an arbitrary spectrum.
We designed a new light source by combining a non phosphor white LED (American Opto Plus LED, L-513NPWC-
15D) and single color LEDs. This allowed us to fill the spectral gap between the blue and yellow peaks of the non
phosphor white LED. By controlling the input current of the LEDs individually a nearly Gaussian shaped spectrum was
achieved. This wide continuous spectrum creates short interferograms (FWHM ~1.4 μm) without side peaks. To
demonstrate the properties of this source we measured through a 5 μm thick polymer film. The well localized
interference created by the source allows measuring both surfaces of thin films simultaneously. We were able to pulse
the source at 5.4 MHz.
We apply a hybrid light source with adjustable spectrum to Scanning White Light Interferometric MEMS device
characterization. The source combines light from a blue laser (409 nm), a fluorescent material (emission peak 521 nm),
amber LED (597 nm) and cyan LED (505 nm) to cover the visible wavelengths. The Gaussian spectrum of the light
source reduces interference ringing and improves surface localization, which is important when imaging diffuse surfaces
or layered structures. The new light source allows both stroboscopic illumination and spectrum shaping during a
measurement. Changing the illumination spectrum allows one to maximize the reflection from the measured surface -
compared to reflections from other surfaces - as a mean to improve signal-to-noise-ratio.
To validate the source we measured static MEMS samples featuring known step heights using the light source at three
different mean wavelengths (508 nm, 524 nm and 579 nm). The measured step heights (7.029 ± 0.045 μm,
7.002 ± 0.041 μm and 7.005 ± 0.056 μm) were close to those measured using a halogen lamp (7.025 ± 0.020 μm).
Interferograms without the side lobes typical for white LEDs were achieved. The FWHM of the interferogram of the
combined light source was (1.859 ± 0.008 μm).
Scanning White Light Interferometry (SWLI) allows surface characterization of MEMS components. With transparent
samples SWLI can image multiple stacked layers. However, since silicon is opaque to visible wavelengths, only the top
layer can be measured using visible light. We combined multiple infrared light emitting diodes (IR-LEDs) to achieve
adjustable IR illumination. This allows simultaneous measurement of top and bottom surface topographies of silicon
samples - such as MEMS membranes- using a SWLI equipped with an IR camera. This advances the state of the art of
the field of MEMS characterization by allowing looking under membranes of these devices during operation.
The chronological order of creation of crossing lines scratched into a copper surface was determined using 3D profiles
measured with SWLI and CM. As the methods used are based only on the deformations of the surface and since the
imaging techniques can be used for different materials, the proposed methods are potentially effective also on other
materials.
Determining the chronological order of orthogonally crossing lines is studied in forensic science. The order of creation of
such lines allows in some cases determination of the history of an object without comparing it to other objects.. Methods
based on two dimensional (2D) imaging have been used for this task, but such methods are ineffective if the lines are
made with a similar tool. We apply Scanning White Light Interferometry (SWLI) and Confocal Microscopy (CM) to
study crossing lines on a copper surface scratched with a scratching device. Both SWLI and CM quantitatively measure
the 3D surface profiles with sufficient accuracy for forensic applications. 3D image processing allows removing
unimportant features, such as surface form and roughness, as well as measurement noise from the measured profiles.
Separating inherent features in the crossing area, from other surface characteristics allows one to determine the sequence
of creation of the lines even on a rough and wavy surface.
Scanning white light interferometry (SWLI) allows dynamic full-field 3D profiling of MEMS devices. With stroboscopic
illumination periodic out-of-plane oscillation can be characterized, but in-plane movement is unresolved. We combine
stroboscopic SWLI with image processing to concurrently characterize periodic out-of-plane and in-plane displacement.
A difference in frequency is induced between the sample excitation and stroboscopic illumination signals. The difference
frequency is chosen to allow recording the surface movement at video rate. The stroboscopic image is thus no longer
frozen in time, but moves at frequency equal to the difference in stroboscopic frequencies. This motion is captured with a
CCD camera. The surface velocity is extracted from the apparent motion using optical flow algorithms. For concept
validation we characterize the in-plane and out-of-plane movement of thermal microbridges fabricated on silicon-oninsulator
by deep reactive ion etching. The microbridge geometry was designed for in-plane movement with minor outof
plane deflection.
Stroboscopic scanning white light interferometry is a method for dynamic nanometer range profilometry that is widely
applied for quality control in the MEMS industry. Monochromatic and phosphor coated (PC) white LEDs produce short
light pulses for stroboscopy. The time resolution of a stroboscopic setup depends on its capability to produce short light
pulses with duty cycles less than 5%. The peak wavelength and the spectral shape of PC white light diodes change with
duty cycle. The spectrum of a PC white light LED was measured using Czerny-Turner-type monochromator (Jobin Yvon
H 25) with an optical power meter (Ando AQ-1125). A custom made pulse amplifier drove the LED with a square wave
voltage at 120 Hz. The blue peak wavelength of the white diode was blue-shifted by 7 nm when the duty cycle was
reduced from 10% to 0.5%. The impact of the spectral change on the vertical resolution of the stroboscopic measurement
was characterized through simulating the change in measurement uncertainty. The results were applied to characterize
out-of-plane vibration of thermal MEMS bridges manufactured from SOI wafers. The simulated increase in
measurement uncertainty was 1 nm, when the spectrum shifted 10 nm towards blue. Noise from background vibration
obscured the effect of spectral shift. Although literature says that temperature increase shifts the spectrum of LED, and
although our simulations indicate the existence of such a shift, our experimental results indicate that the deletory effect is
negligible (it does not introduce bias or uncertainty to profiling measurement).
Surface deformation inflicted on two different kinds of thin layered polymer films was investigated under static indentation and dynamic loading (plowing) at room temperature. Affecting the surface were polished spherical steel tips of 0.5-1.6 mm radii moving at 0.1-16.0 mm/s along the surface. A load of 2-5 N was applied on the tip normal to the surface. The surface response was measured with a scanning white light interferometer. The relation between groove
parameters (width and depth) and the deformation tool velocity as well as the tip diameter and the applied load were obtained from white light interferometer scans. In order to cover a large groove area, a series of 3D groove profiles were stitched together with a piece of software. The profile of the entire groove was compared to friction data recorded by the scratching device. The dependence of groove parameters on the input parameters (tip radius, load force and velocity) indicates that white light interferometry can be used to determine mechanical properties such as 'scratchability'
(abrasing resistance) of polymer surfaces without sample contact.
We have built a customized vertical scanning white light interferometer (SWLI) for characterization of freely suspended (no substrate) transparent 5 µm thick monolayer polymer films. These films that are used by the electronics industry are found in e.g. film capacitors. Quality control (QC) of such films is important with regards to capacitance accuracy and process yield.
Devices currently used to characterize polymer films scan the surface point by point whereas we measure, using wide-field SWLI, the total focal volume (0.9 mm by 0.7 mm by 0.01 mm, with 1.5 μm lateral resolution using 6x magnification.) in one rapid (1 s) scan. The thickness profile of the sample is obtained as the difference between the first and second surface profiles.
Inclusions, e.g. defects or intentionally introduced entities also carry interest with regards to the QC of polymer (capacitor) films. We obtained the inclusion distribution within the film volume (with a depth resolution of 20 nm and a lateral resolution comparable to that on the surface) from the changes in the refractive index observed within the film.
The measurement generates data presented in multiple formats: 3-D picture of the scanned volume (revealing inclusions), surface topographic plots, cross-section of the surface, and depth profiles.
We report on using a Scanning White Light Interferometer (SWLI) for quality control of aluminum lead single-point Tape Automated Bonding (spTAB). A spTAB process was used to connect 14 μm thick, 42 μm wide aluminum leads on a 12 μm thick polyimide layer to a micro chip. Three different bonding process parameters were varied in order to maximize the pull force: bond force, ultrasonic power, and ultrasonic time.
A custom built SWLI was used to measure the topography of the bonds in order to find features that correlate with the tensile bond force. This force was obtained in a destructive way by a pull test.
By keeping the bond height within 3±1.5 μm, bonds with acceptable tensile forces in excess of 54 mN were obtained. This was verified by a separate validation measurement where the pull force of bonds complying with the height requirement was recorded.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
INSTITUTIONAL Select your institution to access the SPIE Digital Library.
PERSONAL Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.