Increasing adhesion of current-conducting coatings on alumina substrates
ower electronic modules, produced by domestic industry, contain elements of metal-ceramic compounds in the design. These compounds must meet the requirements for adhesion strength and resistance to temperature changes. At the moment, they are obtained using DBC (Direct Bonded Copper) technology. The transition to thin-film technology to create the basis of soldered elements of power modules will reduce the percentage of rejects and improve the quality of connections. The appearance of a typical switching card and a power module is shown in Fig.1.
In the process of operation, the power modules are subjected to shock, thermal and vibration loads. The chemical and thermomechanical incompatibility of the substrate and coating materials leads to the detachment of the conductive paths, the destruction of the coating, the overheating of the electronic components of the module and, in some cases, can cause them to burn out. Various types of ceramics are used as substrate materials for the fabrication of a switching card: aluminum oxide, beryllium oxide, aluminum nitride. The choice of material is determined by the required thermal conductivity, which should be sufficient to divert the heat generated by the elements of the module during operation. Also an important parameter is the coefficient of thermal expansion of the substrate. If it is very different from the coefficient of thermal expansion of the metallization material, during the heating process the difference in the size of the substrate and the film will lead to an additional load on the joint of the film-ceramic connection.
To study the technology of formation of conductive coatings on ceramic substrates, the laboratory machine for thin-film deposition was modernized. The modernization consisted in the installation of a gas and magnetron sputtering systems, which provided the technological base for the development of thin-film deposition processes.
As a result, the machine received two independent sources of deposited materials: a thermal evaporator and a magnetron sputtering system. The substrate holder is designed in such a way that the test samples can be introduced into the zone of each of the sources. This makes it possible to form two-layer coatings in a single vacuum cycle.
A feature of this magnetron system is the possibility of sputtering with a hot target. The magnetron is constantly supplied with chilled water in a closed loop, which prevents the device from overheating and the failure of its magnetic system. However, only cooling of magnets is provided, and the target is heated up to a temperature of about 900 °C during the operation under the influence of ion bombardment.
Fig.2 shows the scheme of the upgraded system.
After the completion of the commissioning, a series of experiments was conducted to determine the range of modes in which it is possible to form films by means of a thermal evaporator and a magnetron sputtering system.
In the first series of experiments, samples of copper films on alumina ceramics were obtained. Coatings were formed by means of a thermal evaporator at a chamber pressure of 1.3 ∙ 10–3 mbar for 5 minutes. The obtained samples were tested for the adhesive strength by the detachment method. A metal "fungus" of 1 mm2 area was soldered to the film, and then the tear force of the "fungus" from the substrate was measured with the use of a tearing machine and the character of the destruction was investigated. All samples of a single-layer copper coating showed an adhesion strength value of about 5 MPa, which is insufficient for the use of such coatings as conductive in switching cards. In all cases, the film was peeled off from the ceramic substrate.
In the second series of experiments, the coatings obtained using a magnetron sputtering system were examined. Titanium was used as a target material, and alumina ceramic was used as the substrate.
Before the start of the deposition, the target was trained for 5 minutes. Sign of the end of training was a sharp increase in current with its subsequent stabilization. Simultaneously with the cleaning, the target was heated to a temperature of about 800 °C. After the completion of the cathode training, the samples were introduced into the process zone of the magnetron sputtering system and hot cathode deposition was performed. The dependence of the cathode current of the magnetron sputtering system during deposition is shown in Fig.3.
In the course of preliminary studies, the operating pressure in the chamber and the power at which the coatings have no visual defects were determined. For titanium, the pressure in the chamber was from 4.0 ∙ 10–2 to 4.5 ∙ 10–2 mbar, the power of the magnetron sputtering system was from 350 to 420 W. Then, in these modes, a series of experiments was carried out to apply a titanium film to an alumina substrate.
The obtained samples were examined for adhesion by the method of net incisions (Fig.4) and for peeling. In the first case the samples were scratched in two directions, and in the second case a glued "fungus" was tearing away from the film measuring the tear force and investigating the nature of the destruction of the film.
According to the method of net incisions, the maximum value was 5 V, and in the peel test, the force was 65 ± 15 MPa, with the destruction occurring along the adhesive layer with which the "fungus" was attached to the coating, hence the actual values of the adhesive strength are better than the values obtained.
The results of the second experiment allow us to talk about the possibility of using the selected modes for applying titanium as an adhesive layer for the materials used in soldering the components (copper, tin-gold, etc.).
The third series of experiments was aimed at studying the formation of conductive copper coatings and testing the possibility of using titanium layers as an adhesive sublayer under a copper film.
The copper thin film was deposited on the substrate by thermal evaporation, and the adhesive layer was obtained using the magnetron sputtering method. The substrate surface of Al2O3 was previously cleaned in an ultrasonic bath for 10 minutes at a temperature of 25 °C in isopropyl alcohol. To remove moisture before coating, the samples were heated to a temperature of 200 °C.
The adhesive strength of the obtained samples was measured by the detachment method. First, a metal "fungus" measuring 1 mm2 was soldered in the center of the surface of the samples, and then, using a tearing machine, it was torn off perpendicular to the plane of the surface and the value of the tearing force was recorded. The measured value of adhesive strength was 45 ± 5 MPa, which is 10 times greater than that of a single-layer copper coating in the first series of experiments (Fig.5).
The results obtained make it possible to talk about the possibility of using a titanium sublayer as an adhesive layer for materials that are used for soldering components (copper, tin-gold, etc.). At present, investigations of coatings obtained under other application modes are continuing. ■