Editor’s note: The temperatures mentioned in this chapter are very approximate. The temperatures mentioned here are chosen to be a round number in the middle of the ideal temperature range. Conversions between Celsius and Fahrenheit are approximate rather than exact for this reason.
Nothing about nitinol causes more frustration and difficulty than the process of setting the memory shape. Researchers around the world spend tremendous amounts of time determining how to properly heat treat nitinol to get the properties that are needed for their particular project. However, there are at least a dozen independent variables that influence the properties of nitinol. Therefore, no mathematical model exists to predict the heat treatment of nitinol.
Now, there are two basic aspects of a heat treatment that are needed to properly set the memory shape into nitinol: the right temperature and the right length of time. Thin nitinol wires can have the memory shape set very, very quickly because they are so thin. Thicker samples of nitinol require longer duration at the appropriate temperature as it takes longer for the crystal structure to rearrange.
With thicker samples of nitinol, the longer the heat treatment, the stronger the shape memory effect is. This is because more and more of the nitinol is twinned and capable of accepting deformation. If a nitinol sample isn’t heated very long, the shape memory effect must spend the bulk of its energy fighting against the untreated nitinol, which wants to behave like a normal material. Therefore, it is important that nitinol is heated long enough to exhibit as strong of a shape memory effect as needed.
The second part of heat treating nitinol–the right temperature–is much easier to discern what temperature must be used to set the memory shape. While some researchers anneal nitinol samples at temperatures as high as 1575°F (850°C) for short periods of time, it is generally agreed upon that the ideal temperature for setting the memory shape is approximately 1000°F (525°C).
The ideal way to set the memory shape is to use a calibrated lab furnace with a digital interface that will read back the temperature accurate to 1°F. A used lab furnace can be picked up for around $500-1,000 from marketplaces online such as eBay. New, they cost upwards of $5,000. Another method that yields excellent results over and over again is a fluidized salt bath. However, these run upwards of $10,000. So, what cost effective ways are there to set the memory shape? That is the purpose of this chapter.
Ok, so we just have to hold the wire in the shape we want and heat the wire to 1000°F (525°C) and then quench in water. Sounds simple enough right? Yeah, right. The biggest problem is measuring the temperature of the metal, not the air surrounding it. Without expensive instrumentation, the only real way to determine the temperature of the wire is to look for a red glow. Now, nitinol just barely begins to glow red at about 1100°F (600°C), more than 100°F (50°C) hotter than the ideal temperature. Videos on YouTube make it look like it’s easy to set a memory shape using a propane torch. However, a propane torch burns at an incredible 3600°F (2000°C), far hotter than needed. It is important that you use a cooler flame to set the memory shape.
Candles made with paraffin wax are great for setting the memory shape because they burn at about 1100°F, just a little above the ideal shape setting temperature. Caution must still be taken to not overheat smaller wires, but for everything 0.020″ (0.5mm) and thicker, this is not a big problem.
Now, after this initial heat treatment, called annealing, the nitinol will have one of three different properties. It will be either stiff and springy, soft and leaden, or brittle. If the wire is brittle, this means that the wire has been overheated and is no longer any good. This brittleness is caused when the titanium atoms combine with carbon and nitrogen in the atmosphere to form titanium carbide and titanium nitride crystals within the crystal lattice. This prevents twinning from occurring and makes the lattice very stiff and brittle. No shape memory effect can be seen from this nitinol, ever.
If the nitinol is soft and leaden, this means that the memory shape has been set and the transition temperature is above room temperature and the nitinol is ready to use. If the nitinol is soft and springy, this means that the memory shape has been set but the transition temperature is near or below room temperature. If superelastic nitinol is desired, then the job of heat treating the nitinol is complete! However, if the shape memory effect is expected, then a second heat treatment is needed.
The second heat treatment, called aging, is when nitinol is heated to 700°F (375°C). This causes the transition temperature to rise, but does not re-set the memory shape, so the nitinol sample does not need to be constrained for this part of the process.
Now, it’s hard enough to set the memory shape using tools found at home, how can you expect to age nitinol at home? Well, it turns out that if you have an electric kitchen oven, you can set the thermostat to ‘broil’ and put the nitinol on a cookie sheet on the top rack of the oven. This will be within about 50°F of the ideal value and it doesn’t take any effort to age consistently for very long times (if needed). Aging can raise the transition temperature by 70°F (40°C).