John Fernie joined TWI in January 1990 as a Senior Research Engineer in the Ceramics and Precision Processes Department. Before this he obtained a degree in metallurgy and microstructural engineering at Sheffield City Polytechnic, going on to research oxynitride-glass ceramics at the Centre for Advanced Materials Technology at Warwick University.
His TWI activities include investigating techniques to bond ceramics to metals, ceramics and composites. He is also involved in projects on ceramic coatings, the application of lasers to advanced ceramics, and superconducting ceramics.
The advantageous engineering properties of ceramics are in growing demand, increasing the need to find satisfactory means of joining them to themselves and to metals. John Fernie takes up the story.
Ceramics have been used as structural and basic engineering materials for hundreds of years. Early ceramics were usually naturally-occurring minerals, fired after a crude shaping operation - indeed, the word ceramic is derived from the Greek Keramos meaning potter's clay. Today, there is still a need for such traditional materials in the refractory and whiteware industries. Alumina-based ceramics are used as a cheap source of thermal and electrical insulation or as inexpensive building materials where the application calls for hardness, wear resistance and chemical inertness.
Applications
Modern ceramics are widely used in the electronics and engineering industries; applications range from cutting tools, to wear-resistant parts, to heat-engine components. Compared with metals, ceramics generally exhibit higher strengths particularly at high temperatures, greater hardness and dimensional stability, and better resistance to corrosion and wear. Ceramics also have lower densities, so their strength-weight ratio is high.
Engineering ceramics are generally based on simple combinations of oxygen, nitrogen and carbon with aluminium, silicon and zirconium. These elements are abundant and are relatively easy to extract from the earth's atmosphere or crust; moreover, they are nonstrategic, unlike some of the exotic elements found in high-performance metallic alloys.
The automotive, aerospace and electronics industries are particularly involved in ceramics research. Engineering ceramics are used in gas turbines and reciprocating engines where their properties are very attractive; they can be used as turbine blades, shroud rings and valve components, and in turbocharger motors. Ceramic coatings are becoming increasingly important in high-temperature engines, where they provide both thermal protection and wear resistance. Other uses of ceramics include cutting tools, die parts, nozzles and ceramic armour. The electronics industry - the biggest user of advanced ceramics - uses them as substrate material. In recent years ceramics have been used as safety devices in the nuclear industry and since 1986 much research has centred on superconducting ceramics.
Joining methods
In use, most ceramic components need to be joined to a metal or another ceramic part, and there are many possible methods of producing metal-ceramic joints. These may be categorised as mechanical or chemical; there are advantages and disadvantages to both, but the main problem encountered in all ceramic-metal bonding is difference in thermal-expansion coefficients. In general, metals expand much more than ceramics when heated, and contract more when cooled. This causes an increasing build up of strain at the metal-ceramic interface as the temperature changes, and can lead to failure.
Mechanical joints
A range of mechanical techniques can be used to incorporate ceramic components into metallic structures. Technically, mechanical joints are simple in design, using bolts or screw threads. However, even simple design requires complex diamond machining, which makes such joints expensive; moreover, they will not be gas tight.
Shrink and taper fitting exploit the difference in thermal expansion between the metal and ceramic components. The joint is formed at high temperatures; the metal surrounds the ceramic component and contracts on cooling. As the temperature falls the metal clamps the ceramic, so forming a strong joint. Obviously, this kind of bond can be used only at low temperatures and can be subject to failure at local stress concentrations.
A metal flow process can produce a mechanical bond around a ceramic component. The metal, under applied stress and pressure, deforms plastically and flows into recesses in the ceramic surface. Such mechanical keying can form a reasonably strong gas-tight bond.
Friction welding
In this process, a ceramic component is held stationary in contact with a rapidly spinning metallic component; this generates a great deal of heat and the metal flows plastically and can create considerable flash (Figure 1). The process is completed within seconds and the bond can be considered an intermediate between mechanical and chemical bonding.
Chemical bonding
A chemical bond is best achieved by producing intimate atomic contact at the interface. Chemical bonding can be divided into solid- and liquid-phase bonding.
Solid-phase bonding is achieved by atomic diffusion (diffusion bonding). In the absence of a liquid phase, the joint has to be made at sufficiently high temperature and pressure to produce intimate contact and to provide thermal energy to cause diffusion and chemical reaction. Surface flatness and cleanliness are therefore critical. Direct bonding should be used only when the metal and ceramic have similar thermal-expansion coefficients. When materials of dissimilar coefficients are to be joined using solid-phase methods, a system of graded interlayers is used (Figure 2).
Interlayers are usually thin wafers of metal whose thermal-expansion coefficients lie between those of the bulk metal and the ceramic. Appropriate stacking of different interlayers reduces the strain gradient in the joint. Interlayers formed from glasses can be used in the same way; glasses have the advantage that their compositions can be tailored to produce a desired thermal-expansion coefficient, allowing greater flexibility in design. Glasses are also used as interlayers in bonding ceramics to ceramics.
Metals can be bonded directly to glasses by electrostatic (or field-assisted) bonding. Similar to diffusion bonding, electrostatic bonding requires intimate contact between the two surfaces to be joined, but in addition requires an applied voltage, which induces ion migration in the glass and so produces electrostatic attraction between the glass and the metal. For this technique to work the non-metallic component must exhibit ion migration, and good contact across the two surfaces is vital.
Direct bonding of ceramics to metals using a eutectic liquid is a commercial practice in the electronics industry. In bonding copper to alumina, the materials are placed together and heated in an oxidising atmosphere to 1060-1080°C. A eutectic liquid is formed at the copper surfaces, and reacts with the alumina to form a chemical bond.
Liquid-phase bonding or brazing is an alternative to solid-state diffusion techniques for metallic-ceramic bonding. At the joint interface, a braze must form a liquid which must wet both surfaces.
One problem often encountered is that ceramics are relatively inert and are generally not wetted by metals. Two methods are used to overcome this. The first is to metallise the ceramic component prior to conventional brazing. Metallisation treatments such as molybdenum-manganese or titanium are painted or sputtered on to the surface of the ceramic; this allows the braze to react and wet the ceramic.
The second method uses 'active metal brazes' which contain deliberate additions of wetting agents. A good example is a silver-copper braze which contains titanium as the active metal. The Ag-Cu braze metal acts as a carrier for the titanium which diffuses into, and reacts with, the surface of the ceramic, so enhancing wetting.
Brazing is carried out in a vacuum to avoid formation of metallic oxides. Indium can be added to reduce the eutectic temperature, if required. If the thermal mismatch between the metal and ceramic is too great, interlayers may be used in the same way as for solid-state bonding.
The Ag-Cu bronzes are not suitable for use at temperatures above 500°C; for high temperature applications brazes based on Ni-Cr and Ti are available.
Adhesive bonding
There is interest in using adhesives for joining ceramics to themselves or metals, as some adhesives are particularly ductile and can therefore accommodate large mismatches between materials to be joined. Generally, adhesive bonds exhibit only moderate to low strength and retain this strength only at relatively low temperatures (approximately 200°C). The bond strength of the adhesive can be increased by premetallising the ceramic components.
Another family of adhesives, the so-called ceramic adhesives, offers high-temperature reliability. These inorganic adhesives are easy to apply and the curing cycle is simple. As little information is available about the properties of ceramic adhesives, their use is mainly in non-critical areas where they hold components in place and do not have to withstand severe thermal or mechanical shock.
Summary
Engineering ceramics, eg silicon nitride, silicon carbide, alumina and zirconia, have the following advantages over metals:
- high strength-weight ratio;
- retention of strength (and other mechanical properties) at high temperatures (1000°C);
- low coefficient of thermal expansion, ie high thermal stability;
- high temperature corrosion resistance;
- high electrical resistivity;
- high thermal resistivity;
- low coefficient of friction.
In joining metals and ceramics, the most important factor is joint design. There is usually considerable mismatch in thermal expansion between the metal and ceramic which can lead to considerable strain at the interface. It is possible to overcome this problem with careful design of the joint.
Current interest has directed attention to joining ceramics both to themselves and to metals and techniques for this have been discussed above. The most important currently are diffusion bonding and brazing, although friction welding seems to have a promising future in joining ceramics to metals.
One method that has not been touched upon is microwave processing of ceramics. The reason is that, so far, the technique is in its infancy in the research laboratory. But it is easy to envisage a massive industrial interest when suitable hardware is available.
Each method of bonding has pros and cons, depending on the final service requirements of the component. Many materials scientists regard ceramics as the 'material of the future' and the various advantages they have will drive the development of their bonding techniques. In the years to come ceramics will play an increasingly important role in engineering.
