#12 from R&D Innovator Volume 1, Number 4          November 1992 

Getting Warmer: The Nitinol Engine
by Ridgway Banks 

Mr. Banks, an independent machinist and inventor in Point Richmond, California, develops novel heat engines using a peculiar alloy called nitinol.  He has taught junior high school music and science, published chamber music, and worked as a technical associate for Lawrence Berkeley Laboratory. 

For the past 20 years, I have spent most of my free time working on an engine that uses low-grade heat and strange mechanical linkages to create linear or rotary motion.  Although I have devoted more time to this project than to paying activities, the excitement of reaching a goal has provided strong incentive to overcome obstacles.  Much of this excitement comes from exploring the peculiar thermal-mechanical properties of the alloy I use as the power element in my engines.   

The present invitation to contribute a discovery story is disconcertingly timely.  While it's more comfortable to write such an account after the significance of a discovery is substantiated, I suspect that the very discomforts that actually spur research are the ones that are most easily forgotten after success is achieved.  Even a painstaking effort at truthful reconstruction is subject to some unconscious editing when one writes from the vantage point of certainty.   

The unwitting ancestor of the nitinol engine was one of a series of "nonsense machines" I've made over the years as presents for children -- or for no particular reason.  This toy was an attempt to harness the back-and-forth action of solenoids into a continuous rotary motion. 

I built the toy from coat hangers and household junk, and no doubt it would be long forgotten except for a series of events at the Lawrence Berkeley Laboratory, where I was working as a technician.  In 1971, after Harry Heckman completed the pioneering run at the Bevalac particle accelerator, I wanted to give him a personal memento, and the solenoid motor seemed suitably "high tech." 

This toy amused the physicists about as much as it annoyed the electronics maintenance folks who were forced to repair it.  I guess its appeal lay in its total lack of mechanical finesse--it rattled and swayed, sparks flew and solenoids groaned while it ran.  Impressive?  Perhaps--you couldn't help but be impressed that the contraption worked at all! 

Experience with that toy combined with another of my interests, small steam engines, to produce my concept for the nitinol engine.  I thought about integrating the steam engine into an auxiliary, solar-powered domestic generating system, but a glance at the economics of solar collectors convinced me to turn my back on this idea.  It's simply too expensive to produce steam using the diffuse energy of sunlight.   

As I searched for a lower-temperature substitute for steam, I realized that compounds like paraffin have outstanding coefficients of thermal expansion but relatively poor heat-transfer characteristics.  Perhaps a paraffin bellows could replace the pistons and cylinders.  At any rate, during my search for a metallic envelope for paraffin, I began considering bimetallic components--springs made of two metals with different coefficients of thermal expansion--as possible power elements for my engine.  (If this meandering preamble seems self-indulgent, that's because I want to describe the process, not just the result.  A lifetime's familiarity with creative work convinces me of the absolute propriety of nonlinear thinking under certain circumstances.  To put it another way, a premature rush to "get to the point" can guarantee missing the point altogether.) 

Nitinol Has a "Kick"

I played with a bimetallic coil from a kitchen thermometer, then sketched an engine— based on my solenoid toy—that used 200 bimetallic springs to link a vertical rotating wheel to a stationary crankshaft.  Several of my co-workers also were designing bimetallic systems to recover mechanical work from hot water or sunlight.   

The lab had a sample of nitinol, a homogeneous nickel-titanium alloy that changes shape due to a change in crystal structure in an effect unrelated to thermal expansion and contraction.  In its natural state, a wire of this alloy is about as drab as a metal can be--it has a dull gray surface and bends with an unenthusiastic sort of elasticity.  But I was intrigued by what I'd read about the material's unique "shape-memory."   Nitinol "remembers" the shape it had when it was formed at a high temperature.  When cooled, it becomes more malleable (unlike most metals, which harden).  But when the wire is heated, it immediately springs back to its original form.  

According to the instructions, shape memory must be imprinted on the metal by heating it to cherry red.  As someone had borrowed my only torch, I had no convenient way to heat it and just left the wire on my desk.  After staring at the wire for a few days, I wondered if it might already have enough shape memory, from heating during production, to do something interesting.  I made a U-shaped bend and dipped it into my coffeepot. 

After almost two decades of working with nitinol, I still cannot adequately describe my reaction at feeling an inanimate piece of metal spring to life in my hand.  Although the force and speed of the response can be measured, they must be felt to be believed.  I haven't recovered from that experience, nor from the recognition that nitinol has an infinitely higher specific work potential than bimetallic materials. 

Twenty Million Revolutions and Still Going Strong

I made a prototype engine in 1973 from loops of nitinol wire.  The engine had two pans, one containing cold water, the other hot, and it worked from the start.  It's now made over 20 million revolutions and still turns as briskly as ever, just as long as we maintain the temperature differential.  This led to my first patent on a nitinol engine.

As the temperature differential between the heat source and the heat sink need only be 20°C, I can use geothermal hot springs, solar-heated water, or even Arctic Ocean water combined with ambient air as sources of energy.  I was certain I could build practical, useful nitinol engines, particularly since the country was seriously thinking about energy and environment during the mid-1970s.  My engine, after all, produced usable work from diffuse, low-grade heat—the most abundant source of energy on earth.  It had few mechanical parts:  The wire was boiler, expander and condenser all in one.  It had neither valves nor seals, and the materials were cheap.  Best of all, the engine was nonpolluting.  I confidently predicted I'd have a commercial prototype within a year or two. 

Not Yet Ready to Commercialize--But Soon

Too bad seeing a model work is not the same as developing an engine with real-world applications.  Now, two decades later, I've almost reached my goal.  Why the delay?  Partly, it is due to excessive adherence to conventional wisdoms of machine design.  After my first demonstration appeared in the press, everyone—large corporations, the government and others—joined in the act.  We all formed the wires into a coil or loop that would straighten after being heated.  The thermodynamic conversion efficiencies were poor, but efficiency is not an overriding factor when the energy source is practically infinite and free.  We made many embarrassing machines in that period, some of which provided valuable insights.  

After most of my competitors stopped working with nitinol engines, I continued because I felt the potential had not been fully explored.  Over the years, I've tested many designs, continually learning and redesigning.  My latest model has benefited from abandoning some conceptual blinders that I brought from experience with other engines.   

The insight that permitted my new generation of nitinol engines occurred unexpectedly, while I was working alone.  I liken the process to staring at a black box long enough to allow a hunch to transform itself into a rational course of action.  I now use air as the cooling medium (heat sink), deforming nitinol wires in pure linear tension (rather than the coil or loop configuration) and letting them shrink back to their original length in solar-heated water.  This uses energy much more efficiently, even if it does not make a machine that runs "like clockwork." 

I have finally reached the stage where I "know" my efforts will be recognized by commercial success.  Early applications are likely to be in the area of on-site power production for such purposes as agricultural water pumping, but it is possible to envision (in detail) engines producing many hundreds of horsepower for applications ranging from refrigeration to primary power production. 

While 20 years ago I was confident that nitinol would make a commercially valuable engine, this was a subjective conviction.  It has been a great satisfaction to see this conviction substantiated by reality.