Measuring Film Properties
FTIR (Fourier transform infrared) spectroscopy is a powerful analytical tool for characterizing and identifying organic molecules. Chemical bonds and the molecular structure of polymers, and many other materials can be identified by their IR spectrum. When infrared light passes through a material, bonds will only absorb light with the frequency equal to its characteristic vibrational frequency. By scanning a range of IR wavelength, peaks in the absorption spectrum will appear only when bonds are present. The resulting absorption spectra are most useful in identifying the presence of molecules or the formation, destruction, or straining of bonds with processing or mechanical stressing. When polymer materials are heated, exposed to light, or mixed, changes in their molecular structure occur over time that affects the IR "signature" of the bonds. All these changes can be tracked with an FTIR.
FTIR is more than a research tool, it is one of the least expensive analytical tool that should be used in any manufacturing or R&D facilities handling polymers or other thin-film processing materials. FTIR not only provides a mean to check materials at incoming inspection but also a way to check the evolution of polymers through processing and curing stages. It gives a non-destructive way to acquire information on daily processing variations or on degradation, due to strained bonds, oxidation, contamination etc. FTIR helps elucidate processing issues, identify contaminants, byproduct materials evolving during curing of polymers or even identify a competitors' product.
Adhesion is typically a difficult physical property to quantify. No device reliability can be expected if adhesion of thin films is left to guess work. In spite of its obvious fundamental importance to device fabrication, quantitative measurements of adhesion are not often performed in manufacturing. Adhesion test can quantify force, energy or strength, but the results often require an educated interpretation, rendering their use impractical in a manufacturing environment. In other words, there does not seem to be a universal measurement adequate for all situations.
The common qualitative adhesion "Scotch" test, inherited from the Military Standard (MIL-STD-883C) specifications, is unreliable and, at best, provides only a low value adhesion threshold. The only information obtained is that, if the film peeled, the bonding was very weak, therefore completely unacceptable.
The 180º peel test (per ASTM D903-98) is another qualitative test, superior to the "Scotch tape" test because it produces a number. A few mm wide strip is defined on the surface of the film to be tested. The strip is then peeled by applying a force, perpendicular to the surface, in a slow and uniform motion. The test is quantitatively understood and reproducible and is widely used in the PCB industry, but practical problems arise when attempting to pull thin films.
Other adhesion test techniques most frequently encountered are the stud pulling test, the blister test, the interfacial indenture tests and the scratch tests.
The stud pull test consists in bonding a metal stud on the DUT (Device Under Test) with epoxy. The stud is then pulled at a slow programmable rate until failure occurs. The result is the value of the breaking force exerted to separate the film under test from its substrate. The broken surfaces must be examined to understand where the failure occurred and interpret the validity of the results. For instance, the delamination may have occurred at the film-substrate interface, the film may have experienced cohesive failure, the substrate may have fractured, or the failure may have occurred in the stud or the epoxy. In spite of all its pitfalls, this method yields a reasonably repeatable tensile bonding strength. Independent of the operator, it provides a numerical value of the force per unit surface area necessary to reach tensile failure.
The blister test is performed by etching a hole in the substrate supporting the film to be tested. The film is then left tented over the hole. Hydrostatic pressure is applied on the film from the substrate side until delamination or tearing of the film occurs. The energy necessary to separate the film from its substrate and the adhesion strength can be calculated from the hydrostatic force and the work applied on the back of the film. The technique is difficult to implement, and, in practice, is reserved to theoretical research studies.
The indentation test presses a sharp stylus on the film until it buckles under the pressure causing localized delamination. The stressed film' ridges and buckles are measured to calculate the forces used in de-bonding the film. Like the blister test, this test is totally impractical for routine measurements.
Scratch testing is performed by dragging a sharp diamond or metal tipped tool on the surface of the film. The pressure exerted on the film is increased until the coating fails. The cracking of the film under load is acoustically detected. By analyzing the signature of the failure parameters for known substrates and film combination, the adhesion, hardness and modulus of a specific film can be found. This is a tool suitable for development work but not practical for routine manufacturing process control.
A profilometer is a basic tool that allows to quickly measure the physical thickness of films, providing a sharp edge of the film is available.
A stylus profilometer uses a diamond tip dragged on a surface with very light pressure to get information about surface topography. The movement of the tip actuates a LVDT (Linear Variable Differential Transducer) that converts movement in electrical signal. A LVDT is a very sensitive transducer but the tip of the profilometer is conical and has a finite rounded shape, which interacts with the sample being scanned. The vertical sensitivity is in the nanometer range but steep edge profiles are distorted because of the shape of the tip as indicated in the illustration. Small radius tips are better but more likely to be damaged by mishandling.
Four-point probes are used to measure the sheet resistivity of metal films. Sheet resistivity is a function of material, thickness and morphology of a metal film. A four-point probe is a Kelvin probe with the external contacts passing a current and measuring the voltage in the center probes. A good theory of operation is available on the Berkeley Microfabrication Technology site. Because of the current spread, measurements near the edges must be corrected. The corrections are automatic in the computer controlled automatic instruments.
Dielectric materials useful to RDL fabrication are transparent in the visible and near infrared. Taking advantage of this fact, the thickness is measured by reflectometry. By varying the wavelength of the light illuminating the film surface, enough data can be acquired to derive the real and imaginary part of the index of refraction, and the thickness of the film. Unlike an ellipsometer, this type of instrument does not measure the rotation of polarized light in the dielectric but measures the power of the light reflected from a film at normal incidence. Newer instruments have a wider range spectrometer and more computing power that allows the measurement of thick polymer films in the low tens of microns.