KLA InstrumentsTMMini Lesson
Regularly share various technical information, application notes, and usage guides of KLA Instruments' products.
Our products include probe/optical profilometers, nanoindentation testers, thin film thickness gauges, square resistance measuring instruments, and other measuring instruments.
This course
Using KLA's nanoindentation technique to evaluate the brittleness of substrates and film layers
(Keywords: nanoindentation; brittleness; elastic modulus; hardness; fracture toughness; scratch testing)
Many materials used in semiconductor manufacturing may crack during processing and application, exhibiting brittle failure. Although the concept of "brittleness" is easy to understand, it is not easy to give a "brittleness" value for a certain material or material system because "brittleness" is not a material property [1]. Here, we propose a concise combination of nanoindentation experiments aimed at evaluating the brittleness of substrates and epitaxial layers, and providing feedback to semiconductor manufacturers to reduce potential defects that may occur and propagate during the manufacturing process.
To evaluate brittleness, we propose the "brittleness triangle" shown in the following figure. By measuring (1) the ratio of hardness to elastic modulus, (2) fracture toughness, and (3) critical scratch load, we will be able to understand how a material system will behave under the external forces shown in the figure.
Brittleness is not a physical quantity that can be calculated through the equation of these three measurement results. For complex material systems, the brittleness is most suitable for application when the three parameters in the triangle are "matched". Brittleness is a combination of three measurement parameters, just like a beautiful face is made up of its various elements.
In this article, we will mainly discuss two measurements based on nanoindentation testing: (1) elastic modulus and hardness, and (2) fracture toughness.
The first measure of brittleness is the ratio of hardness to elastic modulus, and we need to measure H (hardness) and E (elastic modulus) to calculate their ratio. In fact, H and E are the most common and possibly the most important mechanical properties in nanoindentation testing. Usually, these two parameters are obtained simultaneously in the same test. Control the pressure head to press into the material and continuously record information such as load, depth, and contact stiffness. Using the measured data and the classic Oliver Pharr model, H and E can be calculated.
However, the presence of substrate effects makes testing difficult when measuring thin films. This refers to the fact that the measured value is easily influenced by the underlying substrate, in other words, the measured value is a mixed result of the film and its substrate. The greater the depth of compression, the greater the influence of the substrate. Thin films are ubiquitous in semiconductor devices. Due to the substrate effect, it is difficult to evaluate the impact of changes in process or material composition on the mechanical properties of thin films.
Our solution is AccuFilm Ultra. Jennifer Hay and Crawford Bryan from the KLA nanoindentation team proposed a clever method to subtract the influence of the substrate from the measured values, thereby obtaining the modulus of the film itself. The model and equation are as follows.
Using this equation, we can calculate the target parameter - film modulus - based on the measured values (apparent modulus). The other input parameters of this equation are usually known values, including film thickness, substrate modulus, contact depth, etc. In our software, all calculations are automatically completed by the AccuFilm Ultra method.
To evaluate brittleness, we propose the "brittleness triangle" shown in the following figure. By measuring (1) the ratio of hardness to elastic modulus, (2) fracture toughness, and (3) critical scratch load, we will be able to understand how a material system will behave under the external forces shown in the figure.
So, what are the advantages of AccuFilm Ultra technology? Let's understand this through examples.
The above image shows the test results of a low-k film with a thickness of about 500nm on a silicon wafer. The orange curve represents the modulus results influenced by the substrate. Its value increases with depth because the influence of the substrate gradually increases. In the curve, usually people can only follow the guidance of industry standards and take the value of the plateau at about 10% of the film thickness as the film performance, which is about 7.5GPa.
The flat colored curve in the figure represents the film modulus obtained using the AccuFilm method. When the pressing depth reaches 40% of the film thickness, the test results can still avoid the influence of the substrate. This allows us to obtain modulus values at greater depths (such as 25% of the film thickness), which is approximately 6GPa.
It is worth noting that taking values at deeper locations can provide practical assistance. Measurement of indentation at very small depths (such as 10 nanometers) is more susceptible to factors such as sample surface roughness, environmental noise, and needle tip defects. As the measurement depth increases, the standard deviation decreases from 0.36GPa to zero point one7GPa, which means that the measurement accuracy has improved by more than 50%.
Therefore, by using the AccuFilm algorithm, the sensitivity has been significantly improved. This algorithm is very useful when there are slight changes in the film formula that result in differences in the mechanical properties that need to be measured.
When the film thickness is further reduced to ultra-thin, such as 100 nanometer gold film, our thin film method can still provide reliable test results.
The above figure (left) shows the untreated apparent modulus results. The measured values continue to rise, and there is no stable platform for reading the performance of the film. But after AccuFilm analysis, it is easy to identify a platform at a depth of about 10% of the film thickness. The value of this platform is very consistent with the nominal modulus of gold, and accurate results can be obtained even at depths as small as 8nm.
The above is about the ratio of hardness to modulus, which is the first indicator for measuring brittleness. Next is the second indicator for measuring brittleness - fracture toughness.
During the use of devices and materials, cracks may form and 'grow'. Fracture toughness is a key parameter related to crack propagation, which is related to the energy required for crack propagation after formation. In short, materials with high fracture toughness often have smaller cracks.
As shown in the above figure, the definition of fracture toughness is very intuitive. Usually, a cube corner indenter or a Bose indenter is used to press into the material to generate cracks. The geometric shape of the indenter determines the constant in the formula, which is 0.036 for a cube corner indenter. Then, measure the crack length c generated under the load P. In addition, it is necessary to use a Brinell indenter to measure the elastic modulus E and hardness H of the sample. Through these four values, the fracture toughness can be calculated.
Fracture toughness is an inherent property of materials, therefore many known materials already have relevant data. However, the fracture toughness values in literature are usually based on bulk and polycrystalline materials. For single crystals or epitaxial layers, traditional material toughening mechanisms are no longer applicable and often make these types of materials more brittle, requiring additional testing.
Next, let's take a look at some examples.
TEOS can be used for wafer to wafer bonding in chip manufacturing and advanced packaging processes. It serves as a protective layer and buffer layer. The sedimentation methods of TEOS are diverse. The next generation technology will adopt layered deposition, with each layer having a thickness of about 5 microns and stacked up to 40 microns. Each layer can alternately apply compressive or tensile stress to make the final coating stress approach zero. Subsequently, the coating was subjected to thinning treatment through chemical mechanical polishing (CMP). The main problem during CMP process is the generation and propagation of cracks inside TEOS and between TEOS and the underlying chip.
In addition, TEOS may also be used to deposit into the grooves between chips. The sedimentation characteristics of the sidewalls and corners are different from those of the plane, so the resulting material properties may also be different. Testing is crucial for understanding the deformation characteristics and fracture toughness of materials.
Sample | Fracture Toughness （MPa m1/2） |
1 | zero point nine one±zero point one |
2 | zero point seven one±0.1 |
3 | zero point nine nine±0.1 |
4 | one point three±0.1 |
5 | one point four±zero point two |
The above figure shows the test results of TEOS thin films prepared using different processes. Under the same compressive load, the crack length of sample 5 is significantly smaller than that of sample 2, indicating that it has better fracture toughness.
Not only for thin film samples, but also when the composition or crystal structure of bulk materials is different from known materials, measuring fracture toughness is crucial. For example, in the thermal mechanical processing of SiC wafers, accurate determination of their fracture toughness and elastic modulus is crucial, as elastic modulus is used to evaluate the stress state, while fracture toughness is used to determine the stress limit under a given defect distribution state. The values in the literature are not consistent and therefore cannot be accepted, so it is necessary to measure the characteristics of specific wafers. Nanoindentation technology can provide accurate and reproducible measurement results of fracture toughness, Young's modulus, etc., for use in simulation software.
The above introduces some key technologies and methods for measuring brittleness. For more research on material brittleness in semiconductor manufacturing and information on related nanoindentation products, please contact us.
References:
[1] B. R. Lawn, and D. B. Marshall, “Hardness, toughness, and brittleness: an indentation analysis,” J. Am. Ceram. Soc., vol. 62, pp. 347-350, July 1979.
WeChat Official
Support
