Craft & Design Technology
How-To: Work with Shape-Memory Alloy

We’re very pleased to have Jie Qi on MAKE for our Advanced Materials month (which has been extended until Feb 6). Her bio is so impressive, I thought I’d post the entire thing:

I was born in China and moved to the US when I was 6… Fast forward to college at Columbia University. I started out a pre-medical/biomedical engineering major and spent a semester and summer doing tissue engineering research in the MBL group. That same summer, I got an internship in Brooklyn building sculptures out of bottles for Aurora Robson, through which I fell in love with art again. So I changed my major to mechanical engineering (because they get access to the machine shop!) and discovered new media art. At this point, I got started on the littleBits project with Ayah Bdeir and learned to design and build electronics. She introduced me to the Media Lab, and I immediately fell in love with the place. So I got a summer internship in Leah Buechley’s group. That summer, I discovered the Arduino and the joys of paper + electronics. After I went back to Columbia for my senior year, I continued working on littleBits and spent a semester attempting to make microscale artwork using photolithography and microfluidics in the LMTP research group. Turns out it’s really hard to make art when you can’t see it! After graduating (with a BS in mechanical engineering), I came straight to the Media Lab and started my masters in the High-Low Tech group. I’m still here, now in my second year, and learning tons every day!

Jie put together this wonderful introduction to shape-memory alloy (SMA) for us. Thanks, Jie! Great to have you aboard. -Gareth

You’ve likely heard about shape-memory alloys (SMAs), metals that change shape when heated to an activation temperature. When cool, they are malleable and can be shaped like a typical metal. However, when heated to activation, they return to their preset shape. At the atomic level, the crystalline structure of an SMA changes with heat from one regular structure to another. However, while all metals will change shape with heat (i.e. melt), SMAs change shape all in solid phase and this change is reversible. For example:

The most commonly used SMA is nitinol (nickel titanium). Commercially it can come in unset form, meaning it has no “memory” yet, as well as pre-trained shapes like muscle wire which contracts when heated (hence the name).

Because SMAs are silent, lighter than traditional actuators, and can be set to create many kinds of motion, there are a ton of cool things you can do with them that are hard to achieve with other actuators. Marcelo Coelho has experimented with combining custom-set nitinol with textiles to create actuated fashions as well as responsive and soft interfaces like the Shutters Project:

Hylozoic Soil, a sculptural installation by Philip Beesley, uses muscle wire to actuate portions of this “living” architecture.

Yet another magical sculpture is the Robotany project by Jill Coffin, John Taylor, and Daniel Bauen in which the branches of a living tree are made to sway when someone walks by. Since the SMA wire is completely silent and hidden, the tree appears to be moving to a virtual breeze.

In my projects, I’ve found that SMA wires are perfect for actuating papers, which are rigid enough to hold interesting structures but soft and light enough to be moved by tiny SMA wires. Some examples are:

Venus fly trap pop-up page:

Animated vines:

I/O self-folding paper:

SMA Primer
If you’re using SMAs for the first time, I recommend starting with Flexinol muscle wire (which is available on Robotshop, or directly from Dynalloy). It is reliable and simple enough to hook up to a circuit. When you run current through, the resistance of the wire causes it to heat up and the wire will contract about 10% of its original length. Since there is a layer of oxidation around the wire, you cannot solder directly to the wire itself. To attach the flexinol wire to a circuit, simple fold the end of the wire into a “u” and use a jewelry crimp bead around the ends. The crimp bead holds onto the wire mechanically and also makes it possible to solder the ends for robust electronic connections.

The hardest part about using muscle wire is controlling the amount of current running through the wire. You want to give it enough for a dramatic effect, but not so much current that the wire burns out (and stops contracting). Flexinol wire has a consistent resistance per length and an optimum current as specified in the flexinol technical data.

One simple technique is to look at the target current from the data sheet and then use Ohm’s law (voltage = current x resistance) to calculate the length of wire that is needed to maintain this amount of current based on the power supply you have. Since these wires generally require hundreds of milliamps, I recommend getting a strong lithium-ion battery or use a wall power supply. For example, if I were using the 0.006″ diameter wire, which needs 0.400 Amps, and I have a 5V power supply, I would need a total resistance of 5/0.4 = 12.5 ohms. Since the resistance of this particular wire is 1.3 ohms/in, I would need 12.5/1.3 = 9.5 inches.

In general, always test your mechanism using low power and turn on the wire in short intervals. If you see a jerking motion, chances are the wire has gotten too much power and might in fact have burned out. If you have an Arduino, you can hook up the wire using a transistor and PWM the power. Start with a low duty cycle and work your way up until you get a strong enough movement. Note that once the metal is getting enough power to change shape, adding more power won’t make the movement more dramatic, it will only make the shape change more quickly.

Now that you have the basic electronics, here are some simple mechanisms you can use to get various motions out of your wire. One of the most dramatic mechanisms is the curling mechanism where you simply sew nitinol to the paper. In this case, when the wire contracts, the paper must curl around it to make up for the shorter length of wire.

To make a self-folding flap, anchor the ends of wire on the main side of the crease and attach the center of the wire very close to the crease on the flap. This mechanism uses the wire as a muscle to pull the flap.

By adding simple folds, you can change a mechanism completely. For example, going from curling to flapping only takes a couple of curved folds to each side of the curling mechanism.

You can add extra flaps to the folding mechanism to make a parallelogram, which makes surfaces rise off the page.

For these and other mechanisms, check out these projects from a recent paper electronics workshop

If you’re looking for a simple beginner project, try this flapping origami crane tutorial.

Happy making!

41 thoughts on “How-To: Work with Shape-Memory Alloy

  1. What are the differences between different types of nitinol? I see some nitinol for sale on ebay (see here: that is significantly cheaper than 1 meter of flexinol. It would be great to be able to get it for such a cheap price but I assume the cheaper version is inferior to the flexinol in some way when used as an actuator.

    1. The main difference is that the nitinol is untrained, which means it has no set memory. So when you heat it up, it wont actually contract or have any particular movement (other than perhaps straighten, which could still be fun…).

      1. Great tutorial. I am just wondering if flexinol (or musle wire) the only nitinol that is trained to contract? Are there other nitinol suppliers that carry similar products? Thanks.

        1. Good question! “Muscle wire” would be the generic term for contracting wire. I looked around a bit and Flexinol seems to be the predominant form. Apparently in addition to robotshop, you can also get it from Jameco (here). However, for a range of interesting shapes (springs, for example), Imagesco has some good options (here).

          If you do find other sources/brands of muscle wire, please do let us all know!

          1. I have extensive experience using Flexinol wire. Have take in several projects from a prototype into production. Anyone looking to work with this great technology feel free to contact me at


  2. Beginner question, sorry. I bought Nitinol, shaped it to a certain pattern, heated it up till it’s glowing red. I then straightened it back. When reheated, it shaped back to the pattern but not straightened back when cooled. How do you make the wire go back and forth repeatedly from one to another? Thanks, Jie Qi.

    1. First off, kudos for setting your own nitinol! Sounds like it worked pretty well :). Regarding your question, the nitinol gets malleable when it cools back down, but doesn’t actually have a restoring force when it’s cool. In my examples, I rely on gravity or the springiness of the paper to provide the restoring force. For example, in the curling papers the wire contracts when curls the paper, but when cool the paper pulls the wire so it expands, allowing the paper to flatten.

      So in your setup, you will need an external force (perhaps a spring or a rubber band?) to make the wire straighten back. You can also try using another nitinol wire that creates the opposing motion when heated, kind of like a coupled pair of muscles. Hope this helps!

  3. hi! love this stuff!!! i would like to create something like the crane, but multiple ones. would that be possible with one power source? sorry. i don’t know much about electronics…i would really appreciate your help!!!

      1. Yep! It doesn’t burn out as easily and moves pretty quickly– the wires contract when current runs through but they return to their natural shape through ambient cooling (which takes longer for thicker wires). And you could definitely power multiple pieces at the same time using one power source, but it does require a beefier source. In this case a 5v wall power supply would work better than a battery. However, if you’re just starting out I recommend making a single piece first and seeing how that goes.

  4. hi jie! you’re so awesome! thanks for the advice :) i’m definitely going to test it out as soon as I can get some flexinol that is. i’m currently in south korea and I can’t seem to find any here. a shame… maybe i’m not looking hard enough..

      1. I’ve already already talked to them but they don’t ship USPS which is the only way for me to receive shipments without paying a huge shipping fee. well, I’l just keep looking! Thanks again~ I might contact you again for more advice!!!

  5. Wow, how cool is that? This is awesome. This reminds me of a Jimmy Neutron episode where he was able to make his pants self fold. It would be awesome if we could do something like this. How much flexinol do you think it would need to accomplish something like folding a shirt? Oh, but now I realize that the current might burn the tissue. Well, nevermind.. Still, great work on the “living papers”, the videos are great!

  6. Arduino board able to control more flexinol wire?
    1 SMA is using around 200mA, if i use 22 SMA in arduino board, can arduino able to control it?

  7. Hi Jie Qi,
    Thanks for sharing, I’ve been purchased some flexinol wire 0,15″ LT. Just curious, can flexinol to be train like nitinol?


    1. Yep! You just have to make sure you have a jig to hold the nitinol in place during the high temperatures. For example, I’ve even seen people wrap the wire around a nail and run high current through to heat it up. For more controlled setting, you’ll need a temperature-controlled oven. Marcelo’s thesis goes into wonderful detail about this process:

  8. I realise this was posted quite a while ago, I can only hope that someone reads this. I am looking for a wire that will extend when a current is applied and ideally remain that way. it needs to be able to work in extreme temperatures. Im working on a low power, low complexity deployable sun shade for a satellite telescope. Any suggestions?

  9. I have read many articles about shape memory alloy and its applications and what not, but this is the first article I found on how to work with this alloy that too in interesting way. Videos are simply amazing.

Comments are closed.


I’m a tinkerer and life-long learner currently pursuing my PhD at the MIT Media Lab, in the High Low Tech and Responsive Environments groups. My research is all about combining electronics and programming with arts and crafts to creating expressive, imaginative and personally-meaningful technology.

my website

View more articles by Jie Qi