#168 from R&D Innovator Volume 4, Number 7          July 1995

Running an Idea Backwards
by Martin V. Sussman, Ph.D.

Dr. Sussman, a professor of chemical engineering at Tufts University, has written two textbooks on thermodynamics and has twenty patents.

Recently, I invented a process to that improves the physical properties and manufacturing productivity of textile and industrial fibers.  That a specialist in thermodynamics came up with this process shows how seemingly unrelated experiences can serve as the germ for innovation.

The story starts over 20 years ago, when I had heard that an Israeli biophysicist, Aharon Katchalsky, was interested in non-equilibrium thermodynamics.  Classical thermodynamics, my specialty, concerns systems that go from an initial equilibrium or rest state to another such equilibrium state, while non-equilibrium thermodynamics has to do with continually changing systems or systems that are experiencing dynamic inflows and outflows, even when at rest. 

Non-equilibrium thermodynamics is applicable to such biological systems as muscle movement.  I contacted Dr. Katchalsky and arranged to spend my sabbatical in his lab to better understand his attempts to apply thermodynamics to biology and to what are sometimes called “steady state” systems.

The Israeli lab workers were fascinated by the properties of collagen fiber.  Collagen is a basic building material of tendons.  Under the name “catgut,” it serves as an absorbable suture in surgery.  Collagen fiber contracts forcefully to half its length when placed in certain salt solutions.  In fresh water, it expands again, and the expansion-contraction process can be cycled over and over again.

Running the Engine

Here was a material that changed its linear dimension upon the "chemical command" of a salt solution.  If you can get something to change dimension, it doesn't take too much ingenuity to use it to drive an engine.  In fact, that's how humans have always built engines, starting with the steam engines of Mr. Watt and his predecessors.  They didn't use a chemical command, but instead used heat to change the volume of a gas (steam), and the volume change of steam moved a piston.  The use of heat to produce motion is effective, but inherently inefficient.  The direct conversion of chemical energy to mechanical energy would be more efficient and is routine in muscle, but man-made machines are generally not driven this way.

Scientists in Katchalsky's lab assembled a “chemical” engine in which a loop of collagen fibers ran around four pulleys.  Two pulleys were immersed in fresh water and two in salt solution.  By simple gearing, the pulleys were interconnected and constrained to move at rates matching the extent of contraction and expansion of the collagen.  As the fiber contracted in the salt water and expanded in fresh water, it drove the pulleys slowly. 

I said, "That's clever, but you guys are so absorbed in non-equilibrium thermodynamics that you've ignored classical thermodynamics.  An efficient engine always moves against a resisting force only infinitesimally smaller than the force it generates."  (For example, to raise a weight with a rope and pulley, an efficient process would use the minimal force that would move the weight.)  And they said, "Oh?  Well, how do we do this with a fiber?"  My suggestion was that they replace the pulleys with two cones with collagen fiber wrapped in a helix about the cones.  The cones, unlike the pulleys, allowed the collagen to contract gradually in many small decrements; and thus release its contractile energy very efficiently as the fiber wrapped around the pulleys moved from the large to the narrow ends.  We built such a device.  It was a novel mechanochemical turbine that rotated as smoothly and quietly as an electric clock!  The output power was taken from a pulley on the cone's shaft.

Our little engine simulated, efficiently but crudely, the chemical roots of animal muscle action.  All man-made engines use heat to expand something.  Instead, our "mechanochemical turbine" derived its energy directly from a chemical reaction.  We described the work in Science (167:  45, 1970) and our article attracted wide attention from the popular press.

Unhappily, our engine wasn’t practical.  The amount of work it could perform was very low.  You'd need a truckful of salt water to power a bicycle.  Also the collagen lost its strength after two days of engine operation.  It didn't seem worth patenting.  We hoped, however, that our success would stimulate ideas for other types of mechanochemical engines.  Perhaps, the thermodynamics of our engine would provide insights into biological systems like muscle—or more promising—perhaps someone would synthesize a better contractile polymer fiber.

Running the Engine Backwards

One day, while watching the collagen engine run, I asked myself, "What if we were to run it backwards?  What use could that have?  Could we do something by stretching the fibers in this manner?  In the engine, the contracting fiber rotated the cones.  So, how about rotating the cones specifically to elongate a fiber?"

These thoughts brought back memories of fiber production at du Pont, where I used to work.  When nylon is made, it is stretched on rotating cylinders to align the molecules and strengthen the fiber.  Now that I was working with rotating cones (rotating due to collagen shrinking), it seemed that cones may have a distinct advantage over cylinders to produce fibers with more finely controlled molecular alignment and structure.

I returned to Tufts and started working on the idea, which I call the "Incremental Draw Process."  Remember, the mechanochemical turbine worked because small incremental contractions forced the collagen fiber to rotate the metal cone.  Working backwards, I now incrementally stretch the fiber as it moves in an expanding helix from the smaller to the larger circumference of a pair of rotating cones. 

It was also possible to impose a temperature gradient on the cones, allowing the polymer molecules to remain mobile as their orientation along the fiber's length increased.  This allows increased stretching and gives the fiber improved properties, including greater tensile strength.  We now could vary the deformation rate, temperature, and temperature gradient to control molecular structure far more precisely than is possible in conventional fiber processing, making the incremental draw process applicable to many kinds of industrial and textile fibers.  I published this work in Polymer News (18:  70, 1993), and elsewhere.  This attracted quite a bit of attention from the fiber industry.

I worked with a number of fiber companies to test the concept.  It was difficult to obtain the commitment needed for our realization of a commercial process.  However, I persisted, as did Allied Signal and finally a viable manufacturing process was achieved.  As a consequence, other companies are now showing interest.

In my lab, I'm collecting a "library" of fibers made by changing parameters during the incremental draw process.  We analyze this library using x-ray diffraction, and other tools, to try to quantify and prepare mathematical models of structure formation in crystalizable polymers as a function of processing history. 

Developing both the mechanochemical turbine and the incremental draw process were rewarding experiences.  When I think about a new concept or device, I go out of my way to consider what would happen if I ran it backwards.  It's a powerful way to trigger new ideas.