Application Note | Wafer Coatings

Mechanical Properties of Silicon Carbide Wafer Coatings

Microindentation, Microscratch and Micro-Wear of ~4.5 µm Metal Coatings on Silicon Carbide

Coated silicon carbide wafer

Research & Experimental Testing

Jocelyn Esparza

Visual Design & Editorial

Andrew Shore

Introduction

Single crystal silicon carbide (c-SiC) wafers are widely used in the production and manufacturing of microelectronic devices such as integrated circuits (IC) or photovoltaic (PV) cells. Most wafers are formed from highly pure single crystals and can range in size from 25.4 mm to 300 mm. The wafer acts as the substrate for the micro-devices that are fabricated in, and over, the wafer. Doping or ion implantation, photolithographic patterning, deposition of different materials, and etching are all examples of the many micro-fabrication process steps a wafer can go through.

Importance of Mechanical Testing in Determining Hardness, Adhesion/Scratch Resistance and Tribological Properties of Wafer Coatings

The fabrication process for microelectronic devices can have over 300 different processing steps and can take anywhere from six to eight weeks. During this process, the wafer substrate must be able to withstand the extreme conditions of manufacturing, since a failure at any step would result in the loss of time and money. The hardness, adhesion/scratch resistance and COF/wear rate of the wafer must meet certain requirements in order to survive the conditions imposed during the manufacturing and application process to ensure a failure will not occur.

Measurement Objective

In this application, the NANOVEA Mechanical Tester’s Micro module is used to measure the hardness (GPa)/Young’s Modulus (E), Adhesive/Cohesive Failures and COF/Wear Rate of coatings on silicon carbide wafers. The wafers have a ~4.5 µm coating on top of a silicon carbide substrate. By determining the hardness and Young’s Modulus of these coatings the role the substrate plays in these properties can be examined.

By determining the adhesive/cohesive resistance of the coatings, we will be able to see the force at which they begin to fail. Lastly, determining the COF and wear rate of the coatings will allow us to see the rate at which material is lost and their frictional properties.

NANOVEA PB1000 Large Platform

Mechanical Tester

Nanoindenter and scratch tester platform NANOVEA PB1000 with nano and micro modules

Microindentation

Test Conditions and Procedure

Test parameterValue
Maximum force25 mN
Loading rate50 mN/min
Unloading rate50 mN/min
Computation methodASTM E2546 and Oliver & Pharr
Indenter typeBerkovich

Table 1: Microindentation test parameters.

Results

Five indents were made per sample in order to calculate an average Vickers Hardness, Hardness (GPa) and Young’s Modulus. The load vs depth curves that each indent made were analysed in order to assure appropriate initial point of contact and unloading curve slope. The data below clearly shows a higher average hardness for the Control coating versus the Au Alloy Coating and a higher Young’s Modulus for the Au alloy coating than the Control coating.

TestHardness (Vickers)Hardness (GPa)Young’s modulus (GPa)Max depth (nm)
13663.87113.1571
23183.37103.4609
33253.44106.8602
43163.34103.0611
53213.40107.6604
Average3293.48106.8599
Standard deviation210.224.116

Table 2: Microindentation results for the Au alloy coating.

Load versus depth curve for the Au alloy coating
Fig. 1: Load vs depth for the Au alloy coating.
TestHardness (Vickers)Hardness (GPa)Young’s modulus (GPa)Max depth (nm)
18549.0487.8448
27988.4492.9450
38218.6991.4448
48779.2894.6437
58458.9490.7446
Average8398.8891.5446
Standard deviation300.322.55

Table 3: Microindentation results for the control coating.

Load versus depth curve for the control coating
Fig. 2: Load vs depth for the control coating.

The NANOVEA Mechanical Tester’s Micro module demonstrates reproducibility and precise indentation at low levels. The controlled and closely monitored environment allows the measurement of hardness to be used as a quantitative value for comparing a variety of samples. This test also shows that it can measure other characteristics, such as Young’s modulus and maximum depth. Here we have shown the ability to control the specific load being applied to test the thin coatings on top of silicon carbide wafers.

Microscratch

Test Conditions and Procedure

Test parameterValue
Load typeProgressive
Initial load0.1 N
Final load10 N
Loading rate20 N/min
Scratch length3 mm
Scratching speed6 mm/min
Indenter geometry90° cone
Indenter material (tip)Diamond
Indenter tip radius20 µm

Table 4: Microscratch test parameters.

Results

The cohesive failure (A) and adhesive failure (B) of the coatings are reported as the load at which these phenomena occur. The data shows that the Au alloy coating has a higher cohesive and adhesive strength than the control coating. Three scratches were performed for each sample and a graph containing the load, depth and frictional force are shown as well as a full-length image for one of the scratches to visually show the failures.

It is important to mention that the repeatability of the scratches can vary due to the uniformity of the coating and substrate.

ScratchAdhesive coating failure (N)Cohesive substrate failure (N)
11.864.258
21.6664.399
32.2144.455
Average1.9134.371
Standard deviation0.2780.102

Table 5: Scratch results for the Au alloy coating.

Scratch graph and full-length scratch image of the Au alloy coating with failures A and B marked
Fig. 3: Scratch test on the Au alloy coating: load, friction and depth, and the full-length scratch with failures A and B.
ScratchAdhesive coating failure (N)Cohesive substrate failure (N)
11.151.860
21.1711.851
31.1541.793
Average1.1581.835
Standard deviation0.0110.036

Table 6: Scratch results for the control coating.

Scratch graph and full-length scratch image of the control coating with failures A and B marked
Fig. 4: Scratch test on the control coating: load, friction and depth, and the full-length scratch with failures A and B.

The NANOVEA Mechanical Tester’s ability to apply a wide range of loads allows the user to produce repeatable low to high load scratches on a material. This ability is valuable for the wafer industry because manufacturers of these wafers can see at which loads the coating will fail.

Micro-Wear

Test Conditions and Procedure

Test parameterValue
Load typeConstant
Load1 N
Scratch length1 mm
Scratching speed8 mm/min
Total distance traveled16 mm
Indenter geometry90° cone
Indenter material (tip)Diamond
Indenter tip radius20 µm

Table 7: Microwear test parameters.

Results

The following data sets represent the tribology study and tests performed on the Au and Control coating with the NANOVEA Mechanical Tester. Linear reciprocating scratch tests were performed on both coatings with the same testing parameters in order to compare the COF and wear rate of each. The wear rate for Au alloy coating was greater than the wear rate for the control coating. However, the average COF was the same for both coatings. The wear rate was calculated by scanning the post worn area with the NANOVEA 3D profilometer and calculating the volume loss.

SampleMax COFMin COFAverage COFWear rate x 10⁻⁹ (mm³/Nm)
Au alloy coating0.06590.03970.0470335.938
Control coating0.05780.02770.04220.411

Table 8: COF and wear rate of the two coatings.

Reciprocating wear graph of normal force, frictional force and depth for the Au alloy coating
Fig. 5: Reciprocating wear test on the Au alloy coating.
Volume-of-a-hole analysis of the Au alloy wear track: volume 5375 cubic micrometers, maximum depth 4.038 micrometers
Fig. 6: Wear track volume of the Au alloy coating.
Reciprocating wear graph of normal force, frictional force and depth for the control coating
Fig. 7: Reciprocating wear test on the control coating.
Volume-of-a-hole analysis of the control wear track: maximum depth 74.04 nanometers
Fig. 8: Wear track volume of the control coating.

Conclusion

The NANOVEA Mechanical Tester’s micro module is able to not only apply progressive loads but constant loads as well in order to perform tribological studies on thin wafer coatings. By using the same testing parameters, the user is able to compare various materials in order to see how they will perform. The sensitivity that the instrument has allows us to see even the smallest COF values and wear rates.

In conclusion, it was shown that the NANOVEA Mechanical Tester’s micro module with its unmatchable load range (0.003 to 20N and 0.03 to 200N) can fully test coatings using nanoindentation to micro scratch and wear levels. Many other tests could be performed such as controlled fracture toughness and yield strength measurements among many others all on a single system.

Frequently Asked Questions About Testing Wafer Coatings

Why test the mechanical properties of wafer coatings?

Device fabrication can involve over 300 processing steps over six to eight weeks, and a failure at any step costs time and money. Hardness, scratch adhesion and wear of the coatings must meet requirements to survive manufacturing and use.

Which coating was harder?

Indented to 25 mN with a Berkovich tip, the control coating averaged 8.88 ± 0.32 GPa against 3.48 ± 0.22 GPa for the Au alloy coating; the Au alloy had the higher Young’s modulus, 106.8 against 91.5 GPa.

Which coating adhered better?

In progressive scratches from 0.1 to 10 N, the Au alloy coating failed adhesively at 1.913 ± 0.278 N and the substrate cohesively at 4.371 ± 0.102 N, against 1.158 and 1.835 N for the control coating.

How was wear measured on such thin coatings?

With linear reciprocating scratches at a constant 1 N over 16 mm, followed by a 3D profilometer scan of the worn area to calculate the volume lost. The Au alloy wore faster, while the average COF was similar (0.0470 and 0.0422).

Does the substrate affect the results?

Determining hardness and modulus of a ~4.5 µm coating shows the role the substrate plays in these properties, and scratch repeatability can vary with the uniformity of the coating and substrate.

What load range does the Micro module cover?

The PDF gives 0.003 to 20 N and 0.03 to 200 N, from low-load indentation to micro scratch and wear on a single system.

Qualifying Coatings on Wafers or Thin Films?