Nitinol is the trade name attributed to the nearly equiatomic binary mixture of nickel and titanium.  In plain language, this means that nitinol is binary (having only two elements: nickel and titanium) with approximately the same number of atoms of each element.  The name comes from the elemental symbols for nickel and titanium (NiTi) as well as the name of the laboratory at which it was discovered, the Naval Ordinance Laboratory (later renamed the Naval Surface Weapons Center). Nitinol became famous when it was discovered to exhibit the shape memory effect.  Nitinol was not the first material discovered which exhibits the shape memory effect, but it is the most dramatic and powerful.  The shape memory effect in nitinol arises from the existence of two temperature dependent crystal structures with dramatically different properties: martensite and austenite.Nitinol is not the first material discovered to exhibit martensite and austenite—these structures had been observed in steel 125 years prior to the discovery of nitinol.  However, martensite and austenite in steel are not as dramatically different from each other as they are in nitinol.  This largely arises from the very high degree of twinning found in martensitic nitinol.  Twinning is where the atomic bonds are rotated by ‘a partial atomic spacing’.  The technical term is that the bonds are floppy.  So, when the nitinol is martensitic, these bonds are free to move around without breaking, giving the perception that the sample is being deformed. Then, when the sample transforms to austenite, since none of the bonds have broken, they all re-form into the rigid B2 cubic structure, yielding a very stiff and springy sample.The term nitinol has come to be used to cover a whole field of ternary and quaternary (three and four elements, respectively) mixtures.

The term ‘transition temperature’ is a bit of a misnomer as there are several critical temperatures in the shape memory cycle.  The transformation and reverse transformation generally require a few °C from start to finish—giving us the austenite start and finish (A_s and A_f) temperatures as well as the martensite start and finish (M_s and M_f) temperatures.  Also of interest are the peak temperatures (M_p and A_p) when the transformation is occurring the fastest.  If you are dealing with low hysteresis SMA, then the equilibrium temperature (M₀) comes into play.  This is the temperature that must be passed in order for a transformation to occur—even if the start temperature has already been passed.  Lastly, when dealing with superelastic nitinol, the martensite difficult temperature (M_d) may play a role.  At temperatures above M_d, the stress induced martensitic transformation generally does not occur, causing nitinol to behave more similarly to most engineering materials. In standard, binary nitinol, the critical temperatures can be listed, from lowest to highest, as: M_f, M_p, M_s, M₀, A_s, A_p, A_f, M_d.  However, in low hysteresis nitinol, the ordering from lowest to highest can be: M_f, A_s, A_p, M₀, M_p, M_s, A_f, M_d.  Additionally, in wide hysteresis nitinol, the ordering from lowest to highest can be: M_f, M_p, M_s, M₀, M_d, A_s, A_p, A_f. A relatively simple at-home method of measuring transition temperature begins with placing the nitinol in the freezer to ensure that it has been chilled below M_f.  Carefully bend or deform the wire and then place it in cold water.  Gradually warm the water , using a digital thermometer to keep track of the temperature.  When the wire begins to move, you have determined A_s.  When the movement stops, you have determined A_f.  This is a modified form of ASTM-F2082, Standard Test Method for Determination of Transformation Temperature of Nickel-Titanium Shape Memory Alloys by Bend and Free Recovery.

Hysteresis is the gap between the transformation temperature and the reverse transformation temperature.  Different organizations use different methods to measure this, depending on what they are trying to accomplish.  The most liberal measurements are calculated by ε = As – Ms.  A more commonly accepted calculation is ε = Af – Mf. 

‘Trained’ is basically vernacular for ‘it’s already been given a heat treatment so that the wire will work  directly out of the package.  We specify that our wire is trained ‘generally’ straight.  It will have some curvature to it due to the fact that we don’t have a reel to reel furnace for annealing entire spools of wire.

Heat treatment establishes the shape memory effect.  It does this by altering the crystal structure of the nitinol.  Varying the heat treatment profile can have dramatic effects on the mechanical properties of the wire.

 NiTi is your standard binary nitinol that everyone talks about, together with all of its benefits and drawbacks.  NiTi has a moderate hysteresis (30-40°C) but has the greatest recovery (10%).  Unfortunately, fatigue sets in quickly in NiTi (50% of original recovery strength is lost after 10% of cycle life) and total fatigue life is relatively limited (10,000 cycles at 5% recovery)  NiTiCu is similar to NiTi but a portion of the nickel has been replaced by copper.  The primary benefit that people like to take advantage of is the lowered hysteresis (10-20°C).  This permits higher cycle rates, lower transition temperatures, and greater efficiency.  The fatigue life is excellent (>10 million cycles at 2.5% strain) and fatigue sets in slowly (~5% loss at 50% of fatigue life).  The primary drawback is that NiTiCu shouldn’t be stressed much beyond 3%  NiTiFe is similar to NiTi but a portion of the nickel has been replaced by iron.  This results in a very high ultimate tensile strength (UTS)—exceeding most titanium alloys for strength.  This, coupled with the superelastic effect yields a very durable device that will outlast almost any other material.  Most applications of NiTiFe use the superelastic effect rather than shape memory.           

Nitinol changes phases at a specific series of temperatures.  So, anything you do to change the temperature will actuate nitinol.  The most obvious of these is to apply direct heat.  This may be from a flame, warm water, sunlight, or the daily change in air temperature—whatever is convenient for you.  Another option is to use electricity to heat the wire to its transition temperature to initiate the transformation.

There is a basic physical principle, known as Joule heating that allows us to very accurately calculate how much current is needed to activate a nitinol device in the amount of time we want it to. NOTE: This does not take into account radiant, conductive, or convective cooling so it only works for short cycle times (2 seconds or less). The Joule heating equation is P = I²R where P is power, I is current, and R is resistance.  Resistance can be calculated using resistivity: R = ρL/A, where ρ is resistivity, L is the length of the mechanism, and A is the cross-sectional area.  For nitinol, ρ = 7.6 x 10^-5 ohm/cm.  To determine the amount of power that needs to be delivered to the nitinol mechanism, P = E * t, where E is energy and t is time in seconds.  Nitinol has a latent heat of transformation of approximately 20J/g and a specific heat of 0.01 J/g*C. Example:  we have a nitinol mechanism, weighing 13g with a cross-sectional area of 0.75cm^2 and a length of 2.7cm.  The energy required to heat the sample from ambient (20°C) to the transition temperature (45°C) is 3.25J + 260J of latent heat for a total energy of 263J.  In order to cycle this in one second, we need to apply 263W.  The resistance is calculated to be R = 2.7 x 10^-4 Ohms.  Now, we only have one unknown variable, I.  I² = 263/2.7 x 10^-4 = 974000.  Taking the square root of both sides, we find I = 987A.

Unfortunately, we can’t tell you because the mechanical properties of nitinol can be varied so broadly through heat treatments that it’s not really possible to predict in advance.  However, a good starting point is to assume that nitinol generates 25,000 PSI in pure tension.  From there, we can adjust the heat treatment to make the nitinol behave exactly as needed.

While determining the time required to heat your nitinol wire is a relatively simple engineering problem (discussed above), heating the wire thermally requires everyone’s favorite heat transfer partial differential equations (and textbooks).  However, if a rule of thumb is good enough, then the following  values are close enough for back of the envelope calculations: 0.25mm wire: 40°C per second 0.5mm wire: 10°C per second 1mm wire: 3°C per second 2mm wire: 1°C per second 3mm wire: 3 seconds per °C

Every time you heat treat a nitinol wire, you change the transition temperature.  If the transition temperature is low enough, it won’t overcome the thermal hysteresis before reaching ambient temperature.  A good measure is to use a freezer to chill to 0°F/-10°C and then gradually warm the wire until you see it start to move.  This is the austenite start temperature.  Keep heating the wire until it completely restores to its original shape.  This is the austenite finish temperature—which is what is generally referred to as the transition temperature.

You want to heat your wire at 200-400°C to gradually raise the transition temperature.  If you’re not fortunate enough to be equipped with a precisely controlled laboratory furnace, a toaster oven does a fine job as well.  Set the thermostat to HIGH and begin cooking at five or ten minute intervals.  This will allow you to tune the transition temperature to exactly what you want it to be.