Powell presents research at UBICOMP 2007 Workshop in Innsbruck, Austria

Mikael Powell an American university professor presented research at the Transitive Materials workshop of the 9th International Conference on Ubiquitous Computing (UbiComp 2007), held in Innsbruck, Austria, in September 2007. Mr. Powell co-authored "SMA Variables: Directing Kinesis", which is documentation of the research performed with Dido Tsigaridi at the Harvard Graduate School of Design in fulfillment of their graduate degree requirements. The research examines the variables inherent in Nickel/Titanium shape-changing phenomenon and suggests a new method of experimenting with shape memory alloys (SMA) to direct kinesis. This method widens the range of potential interactive applications beyond the traditional “on-off” state. The resulting flexibility permits personalization throughout the course of the transformation, enhancing the use of the material as infrastructure for designed spaces or for interface for pervasive computer systems or mechanical devices.

The workshop, Transitive Materials: Towards an
Integrated Approach to Material Technology, is founded on the thesis that “The domains of architecture, product design, fashion and
ubiquitous computing are rapidly converging. Shape-changing
polymers, parametric design, e-textiles, sensor networks, and
intelligent interfaces are now positioned to provide the
underpinnings of truly ubiquitous interactivity. Seamless and
effective integration will determine our ability to create more
meaningful environments that respond to our personal activities
and social needs”. Indeed, the Ubicmp 2007 conference is an international forum in which to present research in all areas relating to pervasive computing technologies from a wide variety of disciplines and geographic areas.
Mr. Powell, a registered architect and registered interior designer, joined the faculty in Fall 2007 after graduating with distinction from the Harvard Graduate School of Design and a career as a design professional for over 25 years. The following is an excerpt from the paper:

Directing Kinesis
Received: June 15, 2007 / Accepted: July 3, 2007
Abstract This paper examines the variables inherent in the
shape-changing phenomenon and suggests a new method of
experimenting with shape memory alloys (SMA) to direct
kinesis. This method widens the range of potential interactive
applications beyond the traditional “on-off” state. The resulting
flexibility permits personalization throughout the course of the
transformation, enhancing the use of the material as
infrastructure or for interface for pervasive computer systems
or mechanical devices.
Keywords Shape Memory Alloy (SMA) · directed kinesis ·
localized heating · transitive material · ubiquitous interaction ·
personalization · transformation
1 Introduction
Since the early 1930’s researchers have observed an unusual
temperature activated shape-changing phenomenon in various
alloys. While the effect seemed hopeful to provide physical
movement for machinery, the metals themselves proved
problematic for mass application. Early alloys were either too
expensive or too toxic. Modern researchers continue to test
alloys in regards to their efficacy and unique transformation
characteristics. Our particular interest is not in the metallurgy,
but rather, the variables inherent in the shape-changing
phenomenon that allow for variations in the transformation
trajectory of the material and enhance the designer’s palette.
Indeed, the ubiquitous nature of electronic computing suggests
an interconnectivity that is both flexible and personal, having
evolved from a central mainframe where users had to go to it,
through the onset of personal desk computers, to pervasive
technologies. Building upon the metaphor of ubiquitous
interactivity, as expressed through materiality, we regard a
shape memory alloy (SMA) as a transitive material, an
adequate tool to reflect current technologies. We will therefore
review the material used in our research, test the variables,
suggest new exploration which enhances personalization and
integration, and offer concluding remarks.
2 Related Work
Many artists and scientists have experimented with shape
memory alloys and claimed fascination by their potential to
merge the sensual, architectural and corporal worlds.
Etienne Krähenbühl's ‘Onibaba’, a field of waving reeds
fabricated of shape memory alloys, and the ‘Christmas dream’,
which is arms powered by a shape memory mechanism that
periodically deploy and stretch a ribbon heavenward,
exemplify the use of SMA’s as actuators and levers [1]. As
part of a component, the shape memory wire is stressed to
simulate muscular movement. Comparable to these are the use
of SMAs in contemporary projects, such as the nickel-titanium
alloy (NiTinol) spring skeleton that heats the liquid crystalline
elastomer (LCE) in Simon Biggs’s prototypes [2].
On the other hand, Jean-Marc Philippe’s “Hermaphrodite”,
“The Totem of the Future” and “L’Arbre de la Nouvelle
Alliance” seem to be the first art concepts that pay attention to
the full transformational process of the alloys installing SMA
as the feature, as opposed to the actuator of the piece [3]. That
is also evident in the function of the kinetic flower on the
“Kukkia” garment or the kinetic hemline in the “Vilkas” dress
developed by the Extra-Soft lab [4]. Nevertheless, they still
depend on uniform heating to achieve the desired
transformation.
Indeed, while much scientific research has been conducted on
the overall effects of generally heated and fully actuated
SMA’s, instances of experimentation on locally heated
applications for the enhancement of the trajectory geometry
has been sparse. R. S. Dennis developed a laser to locally heat

a NiTi SMA wire to compositionally change areas of the wire
to vary transition temperature along the length, allowing for
regions of SMA effects or no SMA effects [5]. However, his
work focuses on the elastic qualities of the phenomenon as
does much of the locally heated experimentation. Also,
“localized heating and cooling of a Shape Memory material
can provide a very effective means of damping vibrational
energy” as found by Kloucek, Reynolds and Seidman in their
research on NiTi Shape Memory wires [6]. There remains a
dearth of selectively heated trajectory exploration.
The method of experimentation that we employ adds
specificity to the heat application. In doing so, each shape of
SMA is activated sequentially resulting in a speed and
trajectory that can be ‘coded’, as you will, to predetermined
personalized patterns. This evolution of the phenomenon is
analogous to the departure from analog computing to
integrated variations of data. The following experimentation
considers SMA’s as transitive materials, capable of reflecting
the multifarious character of ubiquitous computing.
3 Materials
Shape memory alloys are materials whose microstructure
changes with an input of thermal energy. Although the
phenomenological result may suggest that shape memory
materials are energy-exchanging, they are actually phasechanging
[7]. Heat enables an alteration of the material’s
microstructure through a crystalline phase change.
The material we used for our experimentation is NiTi (Nickel
and Titanium alloy). NiTi can be temperature-activated by
direct heat or through resistance to electrical current [8]. For
our experimentation, two different shape memory alloy forms
were used. They are bar shape and NiTinol wire of circular
cross section. Their specifications are as follows:
a. Bar Shape #SMA-5A, 0.023” x 0.074” x 5”, by Johnson
Mathey Company.
b. HS-6 Nitinol Memory Wire, 0.0297” diameter, by
Educational Innovations Inc.
Both are Nickel-Titanium Alloys with a transition temperature
between 30°C and 50°C.
3.1 Variables
In our bibliographical research and experimentation with the
shape memory alloy, we identified different variables that
influence the phenomenological effects of ‘memory’ and
‘movement’. In our analysis, we define memory as the ability
to return to the austenitic phase. Movement is expressed by the
path of trajectory.
The following functions summarize the interconnectedness of
the variables, as learned through bibliographical research and
initial experimentation (see Table 1).
More specifically:
a. Temperature enables the shape transformation.
b. Temperature influences the velocity of the transformation (it
is a non- linear function).
c. The wire diameter is inversely proportional to the minimum
bend radius and the force released [9].
d. The duration of heat application affects reaction speed and
fatigue.
e. Re-set ability is dependant on an external or counterbalanced
force.
We selected to focus on speed and trajectory for our
examination. In doing so, we introduced selective zone heating
as a method of analysis and experimentation.
Table 1 Primary functions
4 Selective zone heating
Many manufacturers of shape memory wires propose
activating the material with general heating, either by
submersion in hot liquid or buffeting with hot air. They direct
that the wire should be uniformly heated for maximum effect.
The experiments presented in this paper inform us that by
heating particular zones of the wire, different effects are
exhibited. These allow further design flexibility.
4.1 Observation
Velocity is highly influenced by the point of heat application.
We noticed that when a looped NiTi wire was heated close to
the deformation, it immediately returned to the straight
position. When heat was applied farther from the bend, it took
much longer to return to the austenite state (see Fig. 1).
[Fig. 1 Velocity as a function of local resistive heat application.
We surmise that the lengthened period is related to the time
needed for the “critical curve” to fully reach the temperature
Shape = f (T)
Velocity = f (T) and Velocity = f (Hd)
Bend Radius = f (Wd)
Fatigue = f (Hd)
Force = f (Wd)
T: temperature, Hd: Heat application duration, Wd: Wire diameter]

required for activation. NiTi is a relatively poor conductive
material. In regards to resistive heating, the current (which
changes with the length and cross-section of the Nitinol wire)
controls the temperature required to fully attain the Af state.
Rather than consider this a material deficiency, we chose to
highlight its design potential. This was the catalyst for defining
the “critical curves” of shape memory wires and proceeding to
experimentation to explore the trajectory of a wire as it
progresses from shape ‘A’ to shape ‘B’ while selectively
heated.
4.2 Definition
We define a “critical curve” as the area of the material where
more stress is exerted. Usually this occurs at the deformed
curve in a straight shape memory wire, or the straight part of a
wire that is “set” to be bent in the austenite state. We define a
“critical point” as the peak of a “critical curve”.
4.3 Postulations
We assumed that:
a. In a NiTi wire of consistent density and cross-section,
directly applied heat at the axis of the critical point causes the
greatest transformational velocity. In a similar SMA wire, with
resistive heating, the proximity of the contacts to the axis of
the critical point is relational to the velocity of the
transformation.
b. Local heat application will not enable the shape memory
wire to fully reach the austenite state, unless it consists of a
single “critical curve”.
c. The sequence of local heat application for the full recovery
of a shape memory wire with multiple “critical curves” defines
the trajectory of its transition from the martensite to the
austenite state.
d. The variance from the temperature needed to achieve Af
affects the duration of the total transformation.
Experiments were conducted to support these claims. In order
to reliably chart the extent of transformation in the wire, we
added a fixed point as a new variable to isolate the resultant
movement.
4.4 Experimentation
Analysis
The Nickel-Titanium alloy bar stock was pre-set in the
austenite state to be linear. A 7 ½” length was deformed in the
martensitic state to have two opposing radius bends. Direct
heat application was used.
We constructed a neutral background and different colored
acrylic paints were applied to the SMA wire at each endpoint
and to the centerline of each critical curve (see Fig. 2). Heat
was directly applied systematically to each zone by a heat gun.
We charted the path of the points along the shape to the
austenite conclusion.
[Fig. 2 First Experiment. Shape memory in martensite state and
the two trajectories charted to return to the austenite state.]
The experiment was repeated twice for two different sequences
of heating the “critical curves”. For each sequence, the
experiment was conducted once with the end-point of the wire
fixed to the background and once with the center point fixed.
We noted that selective heating changes the trajectory of points
along the shape as well as the culminating shape. The order of
heat application determined the distinct route of each point.
Also, as a geometrical subsequence, it was noted that heat
application near the fixed end of the wire substantially effected
the movement at the distant colored unfixed endpoint. The
length of path was greater the farther the colored point was
from the fixed end. (see Fig. 3)
The zone heating method allows the wire to maintain both
martensitic and austenitic states contiguously.
[Fig. 3 a. Center fixed. b. End fixed. Each actuated in the same
sequence of local heating, applied to the axis of each “critical
curve”. The different trajectories are illustrated fully charted.]
Prototypes
We structured our apparatus to investigate the variable of
direct heat to various zones of the wire, to experience
phenomenon in three dimensions. Our prototypes are not
objects, rather, they are armatures to support different
applications and their distinct properties.
After reviewing our previous experimentation and research, we
compiled a chart to illustrate the variables explored with the
prototypes. The following table shows the interplay between
fixed parameters and both independent and dependant
variables in the new equations that we established for our
prototypes (see Table 2).
For our experiments, the temperature transferred to the wires is
fixed and produced by a 12 Volt current. The diameter of the
nickel-titanium alloy (Nitinol) wires used was chosen to be

0.0297” so as to be thin enough for tight bends. The heat
application time is an independent variable which fluctuates
upon the user’s discretion. The reset ability is established
either by an external force or by a self-deformation generated
according to the heat application time (a dependant variable).
Finally, selective zone heating is meant to be another
independent variable. The zones of heating were pre-selected,
but the way of their alteration for the production of the effect
was manipulated by the user.
[Table 2 The new set of variables, as explored by the
prototypes.]
In our second level of experimentation, we moved from the
single shape memory wire, and the 2D plane of its movement,
to a wire gridwork and a three dimensional plane of action.
The first prototype considers all the new variables, but is still
in a 2D format, while the following two are in 3D. A key
difference between our 2D and 3D experimentation is the fact
that the 3D tests do not have any fixed points. The wires are
free to move in a floating membrane and restraining forces are
only employed to counteract the friction or the weight of the
system. In contrast, in the 2D experiment each wire of the
parallel system of wires has one end fixed.
First Prototype – The soft sculpture
The Nitinol wires were pre-set in the austenite state to be
linear. Six 12” long wires were deformed in the martensitic
state to have five alternating 1” d. radius bends each. A wood
frame was constructed to support the base of the SMA wires.
The wires were fixed to the base in a linear series. A sheer,
synthetic yarn material was employed. The fabric’s corners
were sewn together to form a shroud and the mass was
mounded atop the wire assembly. A 25w., 120v., 60Hz device
was used to manually apply direct heat alternately to critical
curves in the wires to create a stimulating dichroic effect (see
Fig. 4).
As a result of this procedure, we noted that selective heating in
random sequences deformed the shape of the fabric in various
ways, in different speeds and through a variety of angles. The
smooth, visual three-dimensional effect was produced only by
transformations of the wires in six XY planes acting as
structure to the fabric surface. (see Fig. 5).
We chose a polyester/nylon blend iridescent organza for it’s
ability to accentuate the varying planes of geometry through
the color change effect and because slight changes in the
structure affected the stability of the sheer fabric causing
slipping billows of various and often accelerating speeds. We
mounted the fabric onto the SMA structure by gravity to avoid
bonding stress to the wires that inhibit transformation behavior
and to prevent damage to the fabric in the wire heating process
through proximity to the direct heat or wire convection.
We selected direct heat application to enhance the fast/ slow
transformations by positioning the heat source at varying
distances from the critical axis.
There were some limitations to the procedures used. It did not
allow for remote heating – an access area in the back was
required to actuate the wires, and the heating device was not
variable.
[Fig. 4 The soft sculpture.]
This did prove, however, to be a good experiment in
scalability. Large distinctive visual effects were accomplished
by small shifts in the structure. Indeed, the largest trajectory
transformation was produced by actuating zone 2 {Z2} on the
SMA wires (see Fig. 5) in the same effort as applying heat to
zone 6 {Z6} (the smallest trajectory).
Future experimentation should focus on ways to set the wires
to manipulate dichroic membranes to affect colors based on
preset patterns, sequences and speeds. It might be orchestrated
to present the array to specific areas in the field, derived from
the incidence of light on the membrane plane positioned by the
SMA support. Further research should explore ways to actuate
a membrane to direct light and control reflection and
diffraction.
[Fig. 5 Analysis of trajectories followed if heated in this
sequence. The endpoint is charted.]
Second Prototype – The uniform transit
Eight 8” long wires were configured in a grid pattern and
adhered to a gauze sheet 9” wide by 9” long. The Nitinol
wires were again pre-set in the austenite state to be linear. The
gauze was able to move freely in the table and the only force
applied to it was from its own weight.
[Fig. 6 Symmetrical deformation]
Specifically, the shape memory grid (XY), which was attached
to the gauze, was bent in the Z axis to create an uneven
surface. A hair dryer was then used to bring the system back to
its first flat condition. The experiment was repeated ten times
with the heating source in different angles. The last three times
the sheet was heated in the center. The process was videorecorded
as it deformed in a specific pattern (symmetrically) to
attest our hypothesis that the way of its restoration can be
predicted if the first shape of deformation and the heating
process are controlled (see Fig. 6).
[Fig. 7 Directed Kinesis]
Final Prototype – Directed Kinesis
For this prototype, Nitinol wires were again pre-set in the
austenite state to be linear. Eight 1-foot long wires were
configured in a grid pattern and adhered to a dichroic acetate
membrane 18” wide by 18” long. The corners of the
membrane were attached to a wooden frame with elastic bands.
Electrical wiring was installed to provide 16 circuits to the
shape memory wire grid (see Fig. 7).
Specifically, the assembly was deformed to provide radiuses in
the Z-axis direction. The electric circuit was powered and
attached to 16 specific bends of the surface. Switches enabled
the control of each point separately. The affected bends were
expected to go back to their straight shape, influencing the
vicinage, as it had happened in the previous experimental
prototype.
In the cases where two or more points were activated at the
same time, the voltage was divided between them reducing the
power exerted to each one (they were connected in parallel). A
series of technical problems was specified as responsible for
the partial success of this prototype. Specifically, the
membrane was stiff and loosely connected to the wires. Thus,
it soon became unable to follow the movement of the wires.
The circuit wires, on the other hand, were very heavy and
disabled the movement of the shape memory grid to which
they were attached. Finally, the shape memory wires
themselves should probably have been stronger (have a larger
diameter) in order to support the motion of such a wide and
heavy area, or be fragmented in multiple short pieces with less
“critical curves”. Proper adhesion of the wire to the membrane
must be explored to provide enough foundation for the
membrane without incurring bonding and debonding stresses
as sometimes occurs in NiTi wires embedded in membranes
and resistively heated [10] .
Through the experience gained by this prototype, the
knowledge for a successful future model lies in specific
modifications. These modifications include: a lighter and
flexible membrane properly attached to the shape memory
grid, a lighter and shorter length circuit system, higher voltage
(or a non-parallel circuit), slightly thicker SMA wire for
greater electrical force. In regards to the heating system, we
surmise that replacing the circuit with conductive thread or
with custom flexible heaters would exhibit better motion and
higher flexibility. More specifically, the heaters, which use
silicone or polyester with multiple circuits that proportion
wattage to distribute varying heat levels, would be the
following step for a fully directed kinesis.
5 Conclusions and Extensions
Being aware of the numerous interconnected variables inherent
in shape memory materials, we understand that our research
provides only an introduction to the possibilities of these
alloys. Foremost is the knowledge that changing even a small
variable can significantly influence the whole behavior of the
material. A small alteration can produce a great effect. This
can lead to a variety of design applications.
Evaluating this project requires open-minded attention to both
process and outcome. We never saw the prototypes as end
products. They served only as fields of experimentation to
prove our claims with regards to local heating application for
the manipulation of kinesis.
We explored a way of controlling shape memory wires and
SMA meshes by locally heating them, both directly and
remotely (using resistive heating).The potential to control the
transformation trajectory and direct kinesis is promising.
Establishing the full criteria for such a direction will enable the
following advancement - having a desired trajectory (or
transformation) as a given, correctly heating a material to
achieve it, resetting back to the austenite state and then
reheating the same material following a different sequence to
achieve another predetermined path. This would enable, for
instance, a device, made “universal” by the sequential codes of
transformations. Further exploration to direct kinesis might
systematically record and analyze the shape memory material
transformation in space (in lieu of the 2D plane). Also, more
attention should be paid to the concept of critical point.
Research to explore heat application offset from the point will
add the 4th dimension of Time to the application. Thus, each

sequenced trajectory could be manipulated in terms of
transformation period – fast/slow reactions.
Finally, it is crucial to take into account, when viewing this
material as a reflection of contemporary ubiquitous nature, that
each material has unique properties for exploitation and its
optimal use is not independent from scale. The high cost, low
strength, relatively low thermal conductivity or slow reaction
SMA wire may have worked for a long time as impediments in
the use of the item in the architectural environment. However,
we believe it is up to the designer to study their properties
more precisely and use them wisely – by accepting the
differences they imply, taking advantage of them and
respecting the smaller scale they introduce to the design by
their dimensions. These materials should not be used as
substitutes to replace conventional materials, but as something
novel with great potential to influence an integrated
architecture.
Acknowledgements We thank Professor Michelle Addington
for her support and insightful remarks. Our work was produced
within the framework of her course, “Smart Materials: Design
Issues and Applications”, at the Harvard Graduate School of
Design.
References
1. http://sb2.epfl.ch/instituts/Gotthardt/Art.html
2. Blackwell, A., Biggs, S., Making Material Culture,
Leonardo, v.39, pp. 471-473 (2006).
3. Philippe, J-M., Art and Shape-Memory Alloy, Leonardo,
v.22, pp.117-120 (1989).
4. Berzowska, J. Coelho, M. , Kukkia and Vilkas: kinetic
electronic garment, Ninth IEEE International Symposium
on Wearable Computers, 2005. Proceedings, pp. 82-85
(2005).
5. Dennis, Ryan Shaun. Development of a Localized Heat
Treatment Tool for Shape Memory Alloy Wires using an
Ytterbium Fiber Laser. Under the direction of Dr. Juei-
Feng Tu. (2005).
6. Kloucek, P., Reynolds, D. R., Seidman, T. I. “Thermal
stabilization of shape memory alloy wires” Smart
Structures and Materials 2003: Modeling, Signal
Processing, and Control. Edited by Smith, Ralph C.
Proceedings of the SPIE, Volume 5049, pp. 24-34 (2003).
7. Addington, M., Schodek, D., Smart Materials and
Technologies, Architectural Press, Oxford, pp.16-17
(2005).
8. Otsuka, K., Wayman, C.M., Shape Memory Materials,
Cambridge University Press, Cambridge (1998).
9. Gilbertson, R.G., Muscle Wires: Project Book, Mondo-
Tronics, San Anselmo (1994).
10. K.D. Jonnalagadda, N.R.. Sottos, M.A. Qidwai and D.C.
Lagoudas “Transformation of Embedded Shape Memory
Alloy Ribbons” Journal of Intelligent Material Systems
And Structures, Vol. 9 (1998).
(COPYRIGHT © 2007 MIKAEL POWELL. All Rights Reserved)