Diamond is the hardest of all single-phase materials with hardness values around 100 GPa, depending on the crystalline direction. The hardness values for polycrystalline diamond (PCD) films range from 80 to 100 GPa, i.e. rather close to the single crystalline values.
The same is true even for ultrananocrystalline diamond (UNCD) and nanocrystalline (NCD) films, for which the reported hardness values were 95 GPa. Although there is a significant spread of hardness values for this type of films, systematic studies on the impact of important deposition parameters as well as the morphology and structure are missing.
While the UNCD films are a candidate for applications biomedicine and tribology, it is essential to know how the hardness as well as other tribological properties such as friction and adhesion depend on the deposition process and other film properties.
This article reveals the impact of the nucleation density on the morphology of the films, as well as the tribological and mechanical properties of the films. The nucleation density varied between 1 x 108 and 1 x 1010 cm-2 due to variation of the quantity of ultradisperse diamond powder (3-5 nm) added to the nanocrystalline diamond powder (250 nm) used for the ultrasonic pre-treatment of the silicon substrates. Scanning electron microscopy and atomic force microscopy were used to investigate structure and morphology of these films, while Nano Scratch and Nanoindentation tests were used for tribological and mechanical properties.
Deposition Set-Up and Conditions
Ultrananocrystalline diamond/amorphous carbon (UNCD/a-C) composite films were made through microwave plasma chemical vapor deposition (MWCVD) from 17% CH4 /N2 mixtures in a deposition set-up. The experiments were carried out at a substrate temperature of 600 °C, a MW plasma input power of 800 W, a working pressure of 22 mbar and the duration of the deposition was 390 min. The films were made from monocrystalline (100) silicon wafers, etched in NH4 F/HF and then pre-treated ultrasonically in a diamond powder suspension in n-pentane in order to enhance the nucleation density. The pre-treatment suspension contained 50 mg of NCD powder with a mean grain size of 250 nm, to which variable amounts (up to 80 mg) of ultradisperse diamond (UDD) powder with a mean grain size ranging from 3 to 5 nm were added.
Figure 1. Deposition principle
|CH4/N2 flow rate ratio:
|Input MW power:
Morphology and Topography
The UNCD/a-C films, which were deposited under the conditions explained above, have been comprehensively characterized with respect to their bonding structure, crystallinity and composition. Regardless of their different morphology (individual nodules or closed uniform films) as examined below, the UNCD/a-C films are made up of diamond nanocrystallites with sizes ranging from 3 to 5 nm as determined by XRD, which are embedded in an amorphous carbon matrix.
The ratio of the volume fraction of the two phases is very close to unity. The films, when studied with Raman spectroscopy, AES and XPS, demonstrated the presence of sp2-bonded carbon atoms (up to 15 at. %). H2 was not added in the precursor gas mixture, however, as revealed by nuclear reaction analysis, the UNCD/a-C films contain about 8 to 9 at.% H in depth, originating from the CH4 molecules and mainly bonded in the form of sp3 - CHx groups.
In the current experiment, the morphology of the UNCD/a-C films was investigated by using SEM. The films deposited on substrates that are pre-treated only with NCD powder (without UDD powder) are made up of individual nodules (Figure 2 (a)) with a diameter ranging from 700 to 800 nm, from which a nucleation density was determined on the order of 1–3×108 cm-2. The addition of 25 mg UDD powder to the pre-treatment suspension causes an increase in the nucleation density by over one order of magnitude (ca. 4.5×109 cm-2). As a result, the diameter of the nodule decreases to 130-170 nm. The film contains a column-like structure with some voids present at the interface with the Si substrate (Figure 2 (b)). For a mixture with 50 mg UDD powder, the nucleation density increases to approximately 6.5×109 cm-2. In this case, the growth has begun from individual nucleation sites until the growing nodules bonded together to form a closed film leaving some voids at the interface (Figure 2 (c)). Finally, the pre- treatment with 80 mg UDD powder produces a nucleation density on the order of 1×1010 cm-2, which reduces the required time to obtain a continuous film (Figure 2 (d)). AFM studies revealed that when the films are closed, the topography and the rms roughness on the order of 9-13 nm are independent on the nucleation density. The surface comprises of rounded features that are agglomerates of smaller substructures in almost all cases.
Morphology: Scanning Electron Microscope (SEM)
Figure 2. Cross-section SEM images of UNCD/a-C films on Si substrates pre-treated without (a) or with the addition of (b) 25 mg UDD powder, (c) 50 mg UDD powder and (d) 80 mg UDD powder.
- Nucleation density increases with the amount of UDD powder
- Development from individual nodules to closed films
- The higher the nucleation density is achieved, the sooner the films are closed
- Voids present at interface in all cases
Topography: Atomic force microscopy (AFM)
Nucleation density 3 x 108 cm-2
Nucleation density 1 x 1010 cm-2
Nucleation density extremely high (c-BN substrate)
- Morphology not influenced by nucleation density
- Surface roughness independent of the nucleation density
- Feature size independent of nucleation density
Anton Paar Nano Platform
All mechanical measurements were carried out on an Anton Paar Open platform with Nano Scratch head, Atomic Force Microscope, High Quality Optical Microscope and Nanoindentation module (Figure 3).
Several testing and imaging modules are set up together on the same platform. All measurement and imaging modules are “Positionally Synchronized” to each other and the optical microscope is included as a standard module on the platform.
Advanced Mechanical Testing
Previous mechanical studies with UNCD/a-C films with a nodule structure deposited at a higher substrate temperature of 770 °C revealed that in most of the cases, the interaction of the indenter with the film resulted in removal of individual nodules because of the discontinuous morphology. On the other hand, when the indenter is directed to the top of the nodules, the determined mechanical properties are very similar to those of the closed films. In the current study, only the closed UNCD (25), UNCD (50) and mechanical measurements of UNCD (80) films were performed with respect to their adhesion and nanoindentation.
As shown in Figure 4, all nanoindentation measurements produce comparable and reproducible results with typical load/displacement curves.
Figure 4. Typical load/displacement curves of UNCD/a-C films on Si substrates pre-treated with different amounts of UDD powder.
The penetration depth was reduced to 4 mN (approximately 90 nm of penetration) in all measurements in order to minimize the impact of the substrate. The Elastic modulus EIT and the Instrumented Hardness HIT were determined using the Oliver & Pharr method; and for all UNCD/a-C films under study, they were in the ranges of 262–271 GPa and 25-28 GPa, respectively, as shown in Figure 5. The elastic recovery was evaluated from the maximum, while the residual penetration depths were approximately 62–65%. On a first view, the values of EIT and HIT for UNCD/a-C films appear to be rather low in comparison with diamond, UNCD, NCD and PCD films.
Figure 5. Hardness and elastic modulus of UNCD/a-C films on Si substrates pre-treated with different amounts of UDD powder.
These results can be explained only with the presence of the amorphous carbon matrix, which forms approximately one half of the material. While on the other hand, the toughness, not the hardness, is decisive for many tribological or mechanical applications. It has been revealed that nanocomposites comprising of hard or superhard nanocrystallites embedded in an amorphous carbon matrix can add considerable toughness without losing most of the hard character. The values of EIT and HIT for UNCD/a-C films under investigation deposited at 600 °C are slightly lower than the films prepared at 770 °C.
This may be due to the slightly increased fraction of the matrix and decreased density caused by the lower deposition temperature. However, the reduced film thickness (1 µm rather than 4 µm) may also have caused this decrease of EIT and HIT.
In this context, it must be noted that for a 1 µm thick UNCD/a-C film deposited on a polycrystalline diamond film at 600 °C, higher values of EIT=362±49 GPa and HIT=33.7±4.1 GPa have been observed.
The Nanoindentation Tester was verified in accordance with the ISO 14577 standard by measurements on fused silica; and the elastic modulus calculated was71.3 ± 1.7 GPa, which is very close to the theoretical value of 72 ± 2 GPa.
Figure 6. AFM image of Nanoindentation on UNCD 25 sample.
A sphero-conical scratch indenter with an angle of 90° and a radius of 5 µm was used for the Nano Scratch Tests (NST, Anton Paar). Three scratch tests were carried out on each sample, from 0.05 mN to 300 mN with a 6 mm scratch length, the loading rate was 600 mN/min at a scanning speed of 6 mm/min.
As shown in Figures 7 to 9, the critical loads for full delamination were determined from the recorded normal force, frictional force and penetration depth curves along the scratch; the respective images have also been taken.
Figure 7. Nano Scratch datas for UNCD 25 sample.
Figure 8. Critical forces for delamination of UNCD/a-C films on Si substrates pre-treated with different amounts of UDD powder.
Figure 9. Comparison of the optical critical load.
Figure 10 shows a typical 3D image of the resulting scratch. Four different parts can be determined along with the scratch: (i) no penetration of the indenter in the UNCD/a-C film, (ii) increase of the penetration depth and starting of the delamination, (iii) complete delamination of the films, and (iv) damage of the silicon substrate.
Figure 10. 3D image of a 1 mm long scratch on UNCD/a-C film created with a nanoscratch tester. The profiles correspond to the four cases described in the text.
Once the delamination occurs, it is not limited to the width of the scratch, but happened in larger areas with a semicircular shape with a diameter of more than 20 µm. As shown in Figure 8, the normal load at which the full delamination occurs, i.e. the critical force, was determined for all samples. As it can be seen from the figure, no distinct impact of the nucleation density on the adhesion of the closed UNCD/a-C films on Si could be noticed.
However, in all those areas where complete delamination had happened, the underlaying silicon substrate is damaged severely and it proved the protecting nature of the UNCD/a-C coatings.
Ultrananocrystalline diamond / amorphous carbon composite films were prepared by MWCVD from CH4 /N2 mixtures on silicon substrates following different pre-treatments. The addition of ultradisperse diamond powder to the pre-treatment suspension can increase nucleation density by two orders of magnitude. Consequently, the morphology of the coatings change from individual nodules to closed and uniform films.
However, the nucleation density has no distinct impact on the mechanical properties, like elastic modulus, adhesion and microhardness, of the closed films. Although the presence of amorphous matrix reduces the hardness of composite films, improved toughness and the ability to prevent fatal brittle failure are the advantages. The nano scratch test results reveal the protecting nature of the coatings on the underlaying substrate.
The Authors wish to express their sincere gratitude to:
Dr. W. Kulisch (European Commission Joint Research Centre, Institute for Health and Consumer Protection, Ispra, Italy)
Prof. J.P. Reithmaier (Institute of Nanostructure Technologies and Analytics, University of Kassel, Germany)
This information has been sourced, reviewed and adapted from materials provided by Anton Paar TriTec SA.
For more information on this source, please visit Anton Paar TriTec SA.