Content uploaded by David Harrison
Author content
All content in this area was uploaded by David Harrison on Jul 02, 2014
Content may be subject to copyright.
INVESTIGATIONS
OF
GENERIC SELF DISASSEMBLY
USING SHAPE MEMORY ALLOYS
J.D.
Chiodo,
E.H. Billett,
D.J.
Harrison
and
P.
Harry
Cleaner
Electronics
Research
Brunel
University,
Runnymede Campus
Egham, Surrey,
Tw20
OJZ,
UK
Abstract
-
Industrial recycling is a practice of growing
importance while impending ‘Take Back’ European
legislation and economic pressures are increasing. Landfill
sites are becoming exhausted and the industry could benefit
from a novel approach to recycling pre and post consumer
waste. Cost constraints limit the number of different products
that can be recycled. Recyclers are worlung
on
broadening the
range of reusable components from this waste stream, but the
proposed approach would significantly increase the volume of
recyclable material used in manufacturing new products. This
alternative could potentially reduce recycling cost per product
in the event of mandatory recycling as a wide variety of
consumer electronics could be actively or self disassembled on
the same generic dismantling line. The use of Shape Memory
Alloy
(SMA)
actuators in a wide variety
of
consumer
electronic products
in
the same dismantling facility was tested.
The candidate products had undergone a multi-stage
hierarchical temperature regime on their macro and
subassembly disassemblies and results reported. Two forms
of SMA actuators were employed in the designs of actuators;
these were one-way Nickel-Titanium (NiTi) and two-way
Copper-Zinc-Aluminum (CuZnAl) actuators.
I.
INTRODUCTION
A novel form of disassembly was tested on more than one type
of product
in
the same dismantling facility. The internal
designs
of
products with active or self disassembly in mind are
altered to include Shape Memory Alloy (SMA) devices
incorporated into their assembly. At the end of their useful
consumer life upon collection, these products could enter the
dismantling line at the recycler’s facility, where the product’s
self-disassembly would be triggered by an appropriate variable
temperature regime. This would be possible using products
designed for ‘Active Disassembly using Smart Materials’
[
13,
the topic
of
earlier
work.
With SMA actuators inside the candidate products, the
products would enter a sequence of controlled temperature
changes in which the SMA devices would be activated at
appropriate stages. Active disassembly would then occur
allowing different components to be sorted after, through
conventional mechanical sorting technologies. This eliminates
the need for product specific robotic disassembly, a research
approach
so
far unable to provide an economic disassembly
model. Active disassembly however, could provide an
increase in the range
of
profitable recyclable products
[2].
With this disassembly approach demonstrated in the earlier
work, a larger variety of consumer products manufactured by
various consumer electronics manufacturers could be recycled
at the same facility.
Novel findings include multi stage generic disassembly
through a temperature hierarchy
(70,
85,
100 and 120 degrees
Celsius) on product macro and sub assemblies with a variety
of force provisions available from a number of NiTi and
CuZnAl SMA actuators.
II. BACKGROUND
Currently, robotic disassembly is cost prohibitive. Hand
disassembly is only economic for a small proportion
of
the
input material
[3].
In a global survey
of
manufacturers and
recyclers, the majority of manufacturers felt that neither
robotic nor hand disassembly was technically or economically
feasible at the present time
[4].
The active disassembly
approach to recyclability and reuse of constituent components
would widen the narrow band of economically feasible
recyclable products.
Future trends in product design engineering point towards
recycling as an integral part of the life cycle of electronic
consumer products. Automation of dismantling the post
consumer product is still seen as a product specific endeavor.
As
the amount and diversity of electronic products in our lives
increase dramatically, current models of production and
‘unproduction’ seem outdated. This system would also enable
manufacturers
to
separate toxic and dissimilar components and
allow their reuse or safer disposal. This work, while focussed
on electronic products, has the potential to lead to more
generic applications in a wide range of industries.
This paper observes the initial results in the application of
SMA devices in the active disassembly of assembled products.
The smart materials considered
in
this study are alloys of
Nickel-Titanium (NiTi) and Copper-Zinc-Aluminum
(CuZnAl). In active disassembly, two distinct approaches can
be taken when incorporated into existing product designs
(retrofitted) or incorporated in the product during the design
phase (design for active disassembly). In this initial
investigation, the feasibility of retrofitting actuators to
assembled products has been explored. The range of
permissible ambient temperatures and the actuator
transformation temperatures (Tg) will be considered. Future
research will include design for active disassembly.
The
design
of
the actuators employed in these product trials is not
0-7803-4295-X/98/$10.00
82
to be regarded as optimal,. The cost effectiveness of these
components, the range of permissible ambient temperatures,
and the actuator transformation temperatures
[5]
are
considered key parameters
to
be optimized in future research.
111.
METHOD
Prior to the generic disassembly experiments a further series
of experiments concerned the disassembly
of
product
housings, in which the metal assembly screws were removed
to permit the required disassembly forces to approximate
those, provided by the actuators. Details of snap fasteners
included in these successful experiments described the
parameters of force requirements and forces that needed
to
be
surpassed
[6].
All cases tested in these disassembly
experiments proved successful and useful for this regime of
generic disassembly tests.
Further preparations were made to accommodate the
generic active disassembly experiments. As the NiTi alloys
were superelastic in typical ambient temperatures
171
(or
whilst in their 100% Martensite state
[SI),
it was necessary to
anneal [9] the samples to rid them of any potential stresses
experienced since manufacturing. Annealing and training
[IO]
of NiTi SMA actuators are {described in Table
1.
The NiTi SMA actuators were one-way
[ll]
Shape
memory effect (SME) but can be mechanically deformed
[
121
(whilst in 100% Martensite or below martensite finishing
temperature
(MO
[13])
after their first and subsequent
actuations. The CuZnAl SMA actuators did not require
training as these were pre-trained off the shelf items. These
actuators were of a helical design exhibiting two-way effect
[
141. These devices did not require mechanical deformation
after first and subsequent actuations. Table
1
details the
annealing and training processes for the SMA actuators.
Three designs of actuators were used, Fig
1
and Fig
2.
Fig.
I.
SMA
NiTi
Rods
and
Disc
were
used
in
some
of
the
candidate products, refer to Table
4.
First, all NiTi I-way
actuators were trained and mechanically deformed, then
inserted into the products$w disassembly experiments.
TABLE 1
Annealing and Training of SMA Actuators
Deformation
I
I
ldependent
4
SME
(Tg:
As-f)
(alloy specific
lspecific
***
I
<OS
sec
SMA
=
shape
memory
alloy
SME
=
shape memory effect
Tg = Transformation temperature
that
is alloy specific
As-f
=
Ausenite starting to finishing temperature
*
=
Imposing
100%
Martensite structure (surpassing
Mf)
from
100%
Ausenite
structure.
**
=
Dependent
on
extent
of
required
form
change.
In
many
cases,
helical
samples
were
jig
formed
and
All the SMA actuators would be in the martensite state and
consequently a lower displacement shape while in a product in
use at a typical ambient temperature range of approximately
-50
-
+90
C.
As these SMA actuators are trained and
mechanically deformed, they are ready for repeated SMEs.
For this to happen, they must be heated to the Af temperatures,
Table
3.
Before incorporation into products, cycling through
multiple
SME
trials successfully tested all the actuators.
In
the
first series of active disassembly trials, the actuators were
incorporated into product housings and were heated to exceed
Fig.
2.
2-Way
SMA
CuZnAl
helical coil actuators were used
in some
of
the candidate products, Table
4.
All
CuZnAl
actuators did not require training or mechanical deformation
as they were pretrained.
83
their Af temperatures. Following actuation the actuators
returned in less than one second to their previously trained
shapes, hence ‘Shape Memory‘. The SME is time independent
but takes a noticeable time
in
the experiment, as heat must be
conducted through the entire actuator for it to undergo SME.
The temperature range required for a complete SME is termed
as ‘Austenite start to finish‘ (As-f). This temperature range is
different than the temperature range required to induce a stable
low temperature state or 100% Martensite state. The range at
which
this
takes effect is known as Martensite start to finish
(Ms-f). The difference between these two ranges is known as
Hysteresis
[
151.
These actuators returned close to their trained
shapes as dimensional changes were about 2/3rd to 415th of
their original shapes after SME for the NiTi actuators. This
characteristic and force return in SME are subject
to
the their
trained shapes, cross section, force applied in mechanical
deformation, composition
of
alloy and Tg. The actuators can
be subjected to repeated deformations and SME cycles. NiTi
actuators were trained, mechanically deformed and tested
prior to insertion into candidate products, Fig.
3
and Fig.
4.
Fig.
3.
SMA NiTi helical coils were used in some
of
the
candidate products, Table
4.
Displacement is shown prior to
SME tests.
Fig.
4.
Displacement after SME tests.
The generic disassembly experiments were conducted on a
variety of products ranging from cell phones to
a
small
electronic game. There were three different design styles in
the fifty-five actuators employed in total. Before the
experiments, all the products were disassembled manually
without destruction to snap fasteners or product casings. The
SMA actuators were then inserted into the macro and
subassemblies of the test products. A small number of trials
were performed previous to the experiments for appropriate
placement, combinations of actuators in place and hierarchy of
anticipated subassembly outcomes. This consisted of placing
lower actuation temperature SMA actuators in the macro
assemblies and higher temperature actuators in the
subassemblies of the candidate products.
Upon insertions of actuators
in
the candidate products, the
entire sample base was placed in a wooden and glass chamber.
The chamber was then heated with air heating apparatus from
room temperature quickly raising the average temperature
from
20
to
75
degrees Celsius
(C)
in the first stage of the
disassembly experiment with results reported.
The hierarchy in temperature regime continued from 70 C
to
115
C, see Table
2.
These temperatures were chosen to
surpass the Af of actuators by at least
5
C.
After 75 degrees
C, vector heating was employed as difficulty was experienced
with higher average temperatures in the chamber available.
General actuator placements for the corresponding
hierarchical disassembly experiments are characterized for the
candidate products, Table
4.
Also
noted are the Af
temperatures of the actuators, their design types and particular
product assembly separation required pertinent
to
the
disassembly. Table
4,
“Candidate Products” may have more
than one of the same product and
in
some cases, the same type
of product but by different manufacturers; this
is
for
comparison reasons. These comparisons will be used in the
investigation of design guidelines of generic disassembly
using smart materials such as SMA employed in these
experiments in future work. Disassembly hierarchies range
from a one product one time self-disassembly with one
actuator inside to a 3-time self-disassembly within
one
produce
housing four actuators. Table
3
is an actuator description of
those employed
in
the experiments.
TABLE
2
Hierarchical Temperature Regime
of
SMA
Actuator Employment by Stage
84
IV. DESCRIPTIONS
TABLE
3
Actuator Descriptions
-AfTemp. Pre Post
T"'
degrees(C)
SME
SME
Height Height
2.3 6
2.3 6
70
3
1.2
-
n.a.
=
not applicable Dimensi
(N)
Deflext'n Temp.
(gr)
SME
>
2200
1.8
70
0.96 13
63.7
0.828 NM 60
0.7
1
2.2
63.7
0.828 NM 60
0.57
2.2
V.
RESULTS
TABLE
4
Hierarchical Actuator Placement by Product and
Results of Generic Disassembly Experiments
Hierarchical Actuator Placement by Product:
Results:
Placement Af(C)at Actuator Hierarchy Disassembly
stages
FuUy
i
t-
Disassembly Type
Within
Technique
of
Successful
I
n.a. Ifailed n.a n.a.
I
n.a.
I
n.a. lfaill
n.a.
I
n.a. Ifailed
I
n.a.
I
n.a.
I
I
I
no
ed n.a. n.a.
IPC
Mouse jcent(:r
rear
I
60
lhelical
coil NiTi
I
2
stage Isnap fit expansion
Isharp
Calculator ltop end
I
60
lhelical coil NiTi
I
2
stage ISnap fit expansion
I
middle
85, 85
helical coil
CuZnAI
2
YeS
70
disc
NiTi No Snap fit expansion
1
no
68,70
helical coil
CuznAl
2
stage
2
no
85
VI. CONCLUSIONS
Most
of
the candidate products proved successful
in
the
temperature/hierarchical
generic disassembly experiments,
From the beginning, only the A4 (17”) CRT monitor, one PC
keyboard and two
of
the four cell phones were not fit for
disassembly as actuator placement proved difficult in the time
allotted. Of the fifty-five actuators used, all provided
SME
successfully with some of the 2-way CuZnAl actuators over
stressed and thus, not able to provide
SME
to the designed
specifications repeatedly, fig.
8.
Of the twenty-one products
chosen, four were unable to be tested. Of the remaining
seventeen, twelve products successfully dismantled with
nineteen
SME
disassembly occurrences since some of the
products were
of
a multi-stage nature. Of the five products
unsuccessful, six failures occurred in total. Seven of the
products were multi-stage and one, the Kodak S.U.Camera-2,
was a 3-stage all within the temperature regime “Stage of
Disassembly 2”. The SMA devices in this application
successfully dismantled the camera at
70,
72 and 73 degrees
C.
This camera’s result exhibits some accuracy potential
within an active or self-disassembly system as a generic
process. Table
5
describes the nature of experiments
throughout the four stages of the hierarchical
disassembly/temperature regimes.
The experiments made it clear that exposure to ambient
temperature was insufficient. This was especially the case in
the unsuccessful experiments, Fig
6.
As SME is temperature
dependent, it is crucial that allowances be made in the product
sub and macro assemblies for ambient temperatures to affect
the SMA devices before destroying the product, Fig
5.
Fig.
5.
CuZnAl
actuators were placed inside the camera.
The disassembly was successful (Stage
2)
however; the
product was not removed until after the
full
batch
of
products
were exposed past the Stage
4
temperature regime.
Fig.
6.
NiTi actuators were placed inside the battery charger
for disassembly (Stages
2
&
4),
Table
4.
The disassembly was
unsuccessful, but destroyed the casing. The product still
requiredforceful prying to dismantle.
Product design must include some changes to the housing of
the intended products if active or self-disassembly were to
take place otherwise, significant damage can result, Fig.
7.
This adjustable keyboard was significantly destroyed as the
NiTi actuator was unable to receive a significant increase in
ambient temperature to react before the polymer casing began
to give way, not allowing a clean break apart at the snap
fastening.
Fig.
7.
A
NiTi actuator was placed inside the adjustable
keyboard for disassembly at Stage
1,
a low temperature
actuation, Table
4.
The disassembly was unsuccessful. Whilst
the polymer housing was broken, the product still required
forceful prying to dismantle.
86
Fig.
8.
periods
of
time
no
longer fit for repeated
SME.
CuZnAl actuators exposed past their
Af
for extended
Fig.
9.
SMA
NiTi and
CuZnAl
actuators were used in the
2
Stage disassembly
of
this successfully disassembled PC
mouse. The cord was easily removed and all internal
components remain intact.
Earlier work has found similar requirements in design changes
necessary for self-disassembling products
[
161.
Other
observations suggest SMA devices should not be exposed to
higher temperatures than their Af for an extended period of
time as this affects cyclical values, Fig.
8.
Future work will
address these and other issues. Most of the products
disassembled in this study proved successful, Fig.
9.
The overall research
is
part of on going step change studies
[
171 attempting to tackle environmental impact reduction in
consumer electronics in part through diversified needs
scenarios
[18,
191.
Preliminary findings were based on the
study at Brunel University 1201.
ACKNOiWLEDGMENT
Paul Harry thanks for help in data collection at Brunel
University
UK.
Teresa O’Toole thanks for the use of the lab,
also for data collection at Durham College, Oshawa ON
Canada. MA” Recycling (Eric Tickner) thanks for the
facility observations and cell phones. Paul Simpson thanks for
the use of testing and observation equipment at Brunel.
REFERENCES
[I]
E.H. Billett, Active Disassembly using SMART Materials
(ADSM), unconventional ideas EPSRC research grant,
Brunel University, Runnymede, Surrey, UK, 28 April 1996
to 28 April 1997.
[2] C.B. Boks, Delft University of Technology, from
conversation, The Netherlands, 12 February 1997.
[3] MANN Recyclers:
E.
Tickner (Special Projects Manager), A.
Thomas (Senior Production Engineer). From conversation,
Ross on Wye, UK, 16 April 1997.
[4] C.B. Boks, E. Tempelman, “Future disassembly and
recycling technology for the electronics and automotive
industry,” Delphi Study Report
-
K370, Delft University of
Technology, The Netherlands, April 1997, pp.
IO.
[5]
R.G.
Gilbertson,
Muscle Wires Project
Book.
Mondo-tronics,
Inc., San Anselmo, CA., 1994, pp. 2-7.
[6]
J.D. Chiodo, J.D. Chiodo,
T,
Anson, E.H. Billett, D.J.
Harrison and M. Perkins, “Eco-design for active disassembly
using smart materials,” Shape Memory and Superelastic
Technologies (SMST) conference, Asilomar Conference
Center, Pacific Grove, CA., USA.,
2-6
March 1997.
[7]
R.G.
Gilbertson,
Muscle Wires Project
Book.
Mondo-tronics,
Inc., San Anselmo, CA., 1994, pp. 2-7.
[8]
R.A. Higgins,
Engineering Metallurgy.
Edward Arnold,
London, 1983, pp. 157.
[9] R.G. Gilbertson,
Muscle Wires Project
Book.
Mondo-tronics,
Inc., San Anselmo, CA., 1994, pp. 2-4.
[
lo] D.E. Hodgson, Ming H. Wu and Robert J. Biermann. (1997).
Shape Memory Alloys, SMA Inc. http://www.sma-
inc.com/SMAPaper.html
[
1
I]
P. Tautzenberger,
Engineering Aspects
of
Shape Memory
Alloys.
(P. Tautzenberger, D. Stockel and C.M.Wayman,
ed.), Butterworth-Heinemann, New York, 1990, pp. 208.
[
121 RA. Higgins,
Engineering Metallurgy.
Edward Arnold,
London, 1983, pp. 90.
[
131 R.G. Gilbertson,
Muscle Wires Project
Book.
Mondo-tronics,
Inc., San Anselmo, CA., 1994, pp. 2-7.
[
141 R.F. Gordon,
Engineering Aspects
of
Shape Memory Alloys.
(T.W Duerig, K.N. Melton, D. Stockel and C.M. Wayman,
ed.), Butterworth-Heinemann, New York, 1990, pp. 245.
[
151 R.G. Gilbertson,
Muscle Wires Project
Book.
Mondo-tronics,
Inc., San Anselmo, CA., 1994, pp. 2-6.
1161 J.D. Chiodo and E.H. Billett, “Preliminary outline guidelines
for active disassembly,” Product Design Education 1997
PDE97, Conference, July 1997. Brunel University,
Runnymede, Surrey, UK.
[17] J.D. Chiodo, P.J. Simpson and
E.H.
Billett, “Socially
responsible design eco-needs scenario comparison,” Product
Design Education 1997 PDE97, Conference, July 1997.
Brunel University, Runnymede, Surrey, UK.
[
181 ibid.
[19] J.D. Chiodo, B.J. Ramsey and P.J. Simpson, “The
development
of
a step change design approach to reduce
environmental impact through provision of altemative
processes and scenarios for industrial designers,” ICSIDP7
The Humane Village Congress, 1997, Conference, August
1997. The Design Exchange, Toronto, ON, Canada. In press
[20] J.D. Chiodo and E.H. Billett, “Active disassembly-
investigation,” Journal of Sustainable Design, Refereed
Journal, Surrey, UK,
1988.
In press.
87
