Anders F. Henriksen, Sc.D
President and CEO
Ceramco Inc.
Center Conway, N.H.
Back in the late 1980s, I had my company make a couple hundred ceramic hex-headed nuts and bolts Sizes 2-56 and 1/4-20 out of alumina to use as promotional giveaways at the National Design Show in Chicago. Despite a nice booth showcasing the company's ability to form complex ceramic parts with tight tolerances, many of the engineers seemed more interested in the ceramic nuts and bolts. They wanted to know their properties, what other designers were using them for, and what was their availability.
Since then, that trade-show gimmick has grown into a three-catalog line of alumina and zirconia fasteners ranging from 1 /4-20 to 0-80, and up to 3-in. long. We also manufacture metric fasteners for European markets (2 to 6 mm and up to 75-mm long). So it seems there is a definite market for re-liable fasteners that are electrically nonconductive, nonmagnetic, and capable of withstanding temperatures to 1,600°C (2,900°F).
CERAMICS GO AGAINST THE GRAIN
Many engineers are inclined to reject the idea of ceramic fasteners on the grounds that they must be impossible to torque and, therefore, impossible to install. Ceramics do have little resistance to torque, and they do not degrade or break down gracefully. Overtorque them by just a hair, drop them, or strike an installed nut with a hammer, and they shatter. And if even a small fracture starts, you can be sure the fastener will soon be in pieces. But this just makes alumina and zirconia fasteners tricky to install, not impossible.
Alumina and zirconia also have lower coefficients of thermal expansion than do metals. And with engineers mixing and matching, using ceramic nuts on metal bolts and vice versa, they must take into account operating temperatures and thermal expansion. So if you plan to use an alumina nut on a threaded, metal rod, be aware it could split if the metal expands more than the alumina at higher temperatures.
What overcomes engineers' reluctance to use such brittle materials for nuts and screws is usually a challenging project. They have to attach a critical component to a device or another component, but the fastener can't conduct electricity and provide a path for a short circuit. Or the nuts and bolts must ignore high magnetic fields in an MRI machine, for example, where magnetic fields could push and pull metal fasteners.
Ceramics are also inert and practically immune to corrosion. The only acid that affects them, for example, is hydrofluoric acid. This lets them be used in plating operations. Ceramics also don't outgas and are impervious even in helium atmospheres, so ceramic fasteners can be used in vacuum environments. And although fasteners made of nylon, PTFE, Delrin, or PEEK might do the job, they will fail if the device is expected to operate at relatively modestly elevated temperatures, (above 480°F for PEEK, above 180°F for Delrin, and above 150°F for nylon, for example).
WORKING WITH ALUMINA AND ZIRCONIA
In general, ceramics encompass all materials that are neither organic or metals. The four basic classes include oxides (MgO, Al2O3, and mullite), nonoxides (carbides, nitrides, and several combinations of C, Si, and N), glasses (SiO2 and B2O3), and various salts (chlorides, sulfates, and nitrates). For now, however, commercial fasteners are only made out of alumina (Al2O3, which is also known as corundum or polycrystalline sapphire) and zirconia (ZrO2).
We manufacture ceramic fasteners using a proprietary low-pressure (100-psi or less) injection-molding technique. Other manufacturers use high-pressure molding to make ceramic fasteners. They typically mix ceramic powder with a polymer binder, which makes a feedstock flexible enough to be injected in molds at pressures up to 5,000 psi. They use a minimum amount of binder because of the difficulty removing it after the part has been formed in the mold. As a result, the feed-stock is stiff, and it takes high pressures to push it into the mold's nooks and crannies.
The downside of this approach is that high-pressure equipment is more expensive and the feed-stock of powder and binder is abrasive enough to erode expensive carbide tooling, wearing them out quickly. The high pressures also create density variations in the part, with lower densities in sections farthest from the injectionmolding gates. This, in turn, creates parts that do not shrink uniformly when fired as denser sections shrink less. Therefore, molds must be adjusted to compensate for the shrinkage variations, making tooling even more expensive.
We use a specially formulated wax as a binder with alumina powders that have a median grain size of 0.8 m. The mixture, which is 85% solids by weight, gets heated to 200°F, turning it into a homogenous paste with a taffy-like consistency. A vacuum pulls out most air bubbles, leaving just the feedstock, which is placed in our proprietary molding machines. There, low-pressure air forces it into an aluminum (6016) mold without compacting the mixture. These molds are a magnitude less expensive than carbide molds. And low pressures do not compact the feedstock appreciably, so there are no density variations in "green" parts. This means fired parts shrink uniformly when fired and there is no need to adjust molds.
A dewaxing process — sorry, trade secret — removes most of the wax and opens the material's porosity. We burn the rest of the wax out, then fire the part to 1,650°C, sintering it. The uniform density leads to uniform shrinkage in all dimensions.
Between removing the binder and sintering, alumina parts shrink 16% (linear) in size; while zirconia parts shrink 23% (linear). With proper controls and planning, we mold fasteners with dimensional tolerances of ±1%. For example, to make sure our ceramic bolts work with any Class 2 thread, we make the pitch diameter of the threads 0.001 to 0.0015 below the Class 2's minimum-allowable diameter. This way, even though the pitch diameter on the threads are designed slightly smaller than on a normal Class 2 bolt, they still work with any Class 2 thread. Similarly, we ensure the pitch diameter on our ceramic nuts are a little larger than normal so they work with regular Class 2 bolts.
The limitation of this approach is tolerance stack up. To avoid this problem, we recommend engineers never have designs that have the bolt engage more than five or six threads at once.
Injection molding makes it easy for us to put different heads on the fasteners by just changing molds. Machining a hex socket or Torx head into a finished bolt would be totally impractical. It is also impossible to machine internal threads on inserts or onto parts. (Machining alumina requires diamond tooling, and it is actually a grinding process. And there is no such thing as a diamond tap for internal threading.) But with molding, we just make a mandrel of the thread type with the proper shrink factor and insert it in the mold to create internal threads.
The bolts don't need any secondary machining after sintering. And a 1/4-20 bolt will have 8,000 psi of tensile strength, making them stronger than plastics, but not as strong as metals. They can withstand 14 lb-in. of torque, are reusable, and will survive temperatures to 1,650°C. The fasteners are nonconductive, non-magnetic, and relatively lightweight, having a density of about 4 gm/cm 3 , about half that of steel.
In the future, it's possible we will be adding carbon nanotubes into the ceramic powders and binders before forming ceramic parts, including fasteners. This could strengthen the fasteners, make them more resistant to impacts, and capable of handling more torque.
CERAMIC FASTENERS | |||
Size | Destructive torque (lb-in.) | Tensile strength (psi) | Maximum temperature (°C) |
4-40 | 2.4 | 4,400 | 1,600 |
6-32 | 3.3 | 5,000 | 1,600 |
8-32 | 7.4 | 7,000 | 1,600 |
10-32 | 11.7 | 8,000 | 1,600 |
1/4-20 | 14.6 | 8,000 | 1,600 |
PROPERTIES OF ALUMINA AND ZIRCONIA | ||
Property | Alumina (A998) | Zirconia (PSZ) |
Composition by weight | 99.8% Al2O3 | 5.4% Y2O3 |
Density gm/cm | 3.89 | 6.0 |
Porosity (volume %) | Impervious | Impervious |
Color | Ivory | Off-white |
Hardness (GPa Knoop) | 11.1 | 11.7 |
Four-point bend strength (MOR) (MPa) | 330 | 760 |
(psi) | 48,000 | 110,000 |
Maximum temperature (°C) | 1,650 | 2,200 |
Thermal expansion coefficient (from 25 to 700°C) (1/ °C) | 7.5 X 10-6 | 11.2 X 10-6 |
Thermal conductivity (@20°C) (W/m°K) | 35 | 2 |
((Btu in.)/(ft2 hr °F)) | 242 | 14 |
Dielectric strength (V/mil) | 200 | 51 |
Dielectric constant (@1 MHz and 20°C) | 10.0 | 28 |
Volume resistivity: -cm @ 25°C | Greater than 10-14 | Greater than10-13 |
-cm @ 300°C | 2.5 X 10-11 | N/A |
-cm @ 500°C | 5.5 X 10-8 | 4.8 X 10-3 |