OBJ Datastream
OBJ Datastream
OBJ Datastream
and Micro-motor Activated High Pressure Optical Cell for Bio-Physical Studies
A Thesis
of
Drexel University
by
of
Doctor of Philosophy
Philadelphia, PA
October 2007
i
Drexel University
Office of Graduate Studies
Thesis Approval Form
(For Masters and Doctoral Students)
Hagerty Library will bind a copy of this form with each copy of your Thesis/dissertation.
This thesis, entitled __Design, Simulation and Experimental Study of Shape Memory Alloy
_____and Micro-motor Activated High Pressure Optical Cell for Bio-Physical Studies___
______________________________________________________________________________
Signatures:
________________________________ ________________________________
Committee Members:
________________________________ ________________________________
________________________________ ________________________________
________________________________ ________________________________
________________________________ ________________________________
Dedications
ii
Acknowledgements
I would like to express my sincere gratitude to my advisors, Professors Jack G.
Zhou from Drexel University and Parkson Chong from Temple University, for their
support and advice throughout the course of this research. Without their encouragement
and guidance, not only in my research and study but also in my new life in United States,
this doctoral research and my other researches would not have been accomplished.
I would like to thank Mark and all other Machine Shop members for fabricating
my experimental tools and modifying them numerous times. Also I would like to thank
glass shop artists, Tom from Drexel University, David from Temple University and
Karen from University of Pennsylvania for making various cuvettes for me.
In addition, I would like to thank my colleagues, Lin Lu, Yi Ma, John Lin, Tie Hu,
Chunlong Hu, for their caring and discussion. I would like to say that I enjoy sharing my
feeling with you. Your warm caring and encouragement, every word during my sad and
happy time, had given me more confidence.
I would like to express my sincerely thanks to my mom for her endless love,
support and encouragement throughout my PhD study and my life, to my brothers and
sisters, to my family members for their love, caring and support. Also I would like to
thank my friends I met in this country, Robert Harcum, Curtis Holloman, Reggie
Garrison, Bryan Skyes, Jingfu Tan, Jay Li and many more who cheered me up when I am
down and shared my joy of my happy time.
Above all, no single word can express my sincerely thanks deeply in my heart. I
would like to say again that this research work cannot be accomplished without your
support, help and caring.
iii
Table of Content
ABSTRACT..................................................................................................................... xiv
1 Introduction................................................................................................................. 1
1.4.1 Design of the New SMA Activated Dual Chamber HPOC ...................... 11
1.4.2 Design of Shape Memory Alloy Activated Stopper and Mixing Systems 11
iv
2.1.5 Thermoelastic Martensite Transformation................................................ 20
2.1.9 The Limitations of Shape Memory Effect and Superelastic Behavior ..... 30
2.2 Challenges and Technical Approaches on SMA Based Actuator Design ........ 32
Actuator 62
v
4.3 HPOC Design: .................................................................................................. 99
4.6 Our New HPOC Fabrication collaborated with ISS ....................................... 111
8 Appendix................................................................................................................. 144
vi
8.1.1 Commercial SMA spring test.................................................................. 144
8.2.3 SMA spring and micro-motor mixing results (1 atm) raw results .......... 153
8.2.4 SMA spring and micro-motor mixing result under 300 bars (normalized)
155
vii
List of Figures
Figure 2-6 One-Way SMA actuator, dual SMA springs design ....................................... 38
Figure 2-10 Constant force loading of a shape memory compression spring (Warram
1993) ................................................................................................................................. 41
Figure 2-13 Typical stress-strain behavior of Shape Memory Alloy material ................. 43
Figure 2-15 Force of SMA at two phases (N/mm) D=0.135” d=0.019” .......................... 45
Figure 2-16 Force(N) vs Extension at high and low temperature (D=0.156”; d=0.018”) 46
Figure 2-17 Force(N) vs Extension at high and low temperature (D=0.156”; d=0.017”) 46
viii
Figure 2-20 Heating and cooling time histories of actuators subjected to step voltage
changes of 4 volts with actuator wire diameter of: (a) 0.75mm (b) 0.875mm (Baz, Iman et
Figure 2-21 Effect of cooling strategy on heating and cooling time histories of 0.4mm
Figure 2-22 Heating and cooling time histories of 0.4mm actuators subjected to
sinusoidal voltage excitations with air cooling velocity =7.2 m/s: (a) at 1/36Hz (b) at
Figure 2-24 (a) Circuit board and relay (b) Two SMA spring testing in water ................ 54
Figure 2-26 Dual SMA springs actuator with tensile and compressive SMA springs ..... 57
Figure 2-28 Surface response for Af Temperature and ultimate tensile stress ................. 59
Figure 2-29 Femlab heat transfer FEM Analysis for two SMA springs........................... 60
Figure 2-31 SMA actuator with tensile SMA and biasing compressive spring................ 63
ix
Figure 2-38 0.25 second, 5A heating and cooling process ............................................... 70
Figure 2-39 one coil cross-section of SMA wire close up heating condition................... 68
Figure 2-40 Temperature line plot of crossing SMA center 5A, 5seconds heating.......... 69
Figure 2-41 0.25 second heating, 1Hz and 10 cycles condition ....................................... 71
Figure 2-43 line plot at 5A 0.25 second heating 0.5Hz condition .................................... 73
Figure 2-44 Heat transfer analysis result for two coils of SMA wire close up................. 73
Figure 2-45 Line plot between 2 coils of 5A, 0.25 seconds heating, 0.5 Hz .................... 74
Figure 2-46 Temperature point plot at one wire diameter away from SMA coil (5A, 0.25
Figure 2-47 Temperature point plot at 10 times of wire diameter away from SMA coil
Figure 2-48 Temperature point plot at the center of coil (5A, 0.25s heating, 0.5 Hz) ..... 76
Figure 2-49 Temperature point plot (5A 0.33s heating 0.5Hz center).............................. 76
Figure 2-54 Mixing result of SMA spring & micro-motor mixer .................................... 81
Figure 3-2 A cuvette of 8cm height and 2cm diameter with 3 quarter necking ............... 84
x
Figure 3-4 Necking examination ...................................................................................... 86
Figure 3-6 Point plots along the center line from necking to bottom of the cuvette ........ 87
Figure 3-7 Location of stirrer with motor within apparatus: 1) motor winding 2) rotor 3)
Figure 3-9 Drop stirrer: 1) coil 2) head 3) core 4) shaft 5) body of apparatus ................. 92
Figure 3-10 Stirrer with reversing rotation:1) core 2) axle 3) finger 4) tube.................... 92
Figure 4-4 Exploded assembly view and assembly electrical lead................................. 102
Figure 4-5 Kalashnikov electrical lead: 1) body 2) holes 3) conductors 4) thread......... 103
Figure 4-6 Exploded view and assembly a window plug ............................................... 105
Figure 4-7 Stress analysis for cell assembly by Ansys ................................................... 108
Figure 4-8 First principle stress of cell assembly by Femlab ......................................... 109
xi
Figure 5-3 ISS K2 Fluorescence Measurement .............................................................. 118
Figure 5-7 Mixing with SMA and biasing spring actuator agitation .............................. 122
Figure 5-9 Mixing result under 1atm first measurement ................................................ 124
Figure 5-11 Mixing result under 1atm third measurement ............................................. 125
Figure 5-14 Mixing result under 300bars, third measurement ....................................... 127
Figure 6-2 Photo Bleaching vs Enzyme effect @ 12k light intensity............................. 132
Figure 6-3 Photo Bleaching vs Enzyme effect @ 100k light intensity........................... 133
Figure 8-1 Low temperature tensile testing force vs time (1-4) ..................................... 144
Figure 8-2 Low temperature tensile testing force vs time (5-8) ..................................... 144
Figure 8-3 Low temperature tensile testing force vs time (9) unplugging force test (10)
......................................................................................................................................... 145
Figure 8-5 High Temperature Af force vs step extension loading (1-4) ......................... 146
Figure 8-6 High Temperature Af force vs step extension loading (5-7) ......................... 146
Figure 8-7 High Temperature Af force vs step extension loading .................................. 147
xii
Figure 8-8 High Temperature Af force vs step extension loading (1-3) failed............... 147
Figure 8-9 Current response (3) (by increasing current, the peak of force increases).... 148
xiii
ABSTRACT
Design, Simulation and Experimental Study of Shape Memory Alloy
and Micro-motor Activated High Pressure Optical Cell for Bio-Physical Studies
Oliver Hongchun Xie
Prof. Jack G. Zhou
Prof. Parkson Chong
During the past two decades, bio-physicists have had an increasing interest in
finding out what happens when two bio-material solutions are mixed under high pressure.
pressure produces only volume changes under isothermal conditions, and pressure results
(HPOC) technology.
Window-type HPOC such as the one designed by Paladini and Weber have
relationships of biological molecules. However, the conventional HPOC is only good for
single solution testing and does not allow for quick mixing and stirring of additional
components while the specimen is under pressure. To mix two solutions under pressure,
Zhou and Chong developed a laser activated dual chamber HPOC. However, the
expensive laser device and its unavailability in most laboratories make the application
difficult. Zhou and Chong also introduced shape memory alloy (SMA) as an actuator to
unplug a urethane stopper with a biasing spring for agitation. The drawback of the
biasing spring is that the spring blocks the observing light beam and creates the creation
xiv
of unwanted reflections, and the SMA spring actuator method has never been put to
practice.
actuators were designed, simulated and tested for unplugging and mixing purposes. To
conduct this research, SMA helical springs were fabricated in house according to the
compressive spring and biasing spring, significant ranges of vibration were developed.
unplug the stopper and micro-motor as a stir device to agitate the solutions was
developed. Rapid mixing of 95% of total solution in 10 seconds was achieved under 300
bars. A new HPOC was designed according to the new cuvette with its new unplug and
mixing mechanism. Our industrial partner, ISS, further modified our design for easy
manufacturing reason and fabricated the HPOC which made SMA actuator mixing test
under pressure possible. Since the inside chamber height was increased to accommodate
the new cuvette and electrical leads were introduced through the cell body to achieve
remote control outside of the HPOC, FEA Method was used to assure the HPOC’s safety
under high pressure. Since the device is designed for bio-physical study, bio-
compatibility of Nitinol SMA with protein and phospholipids was studied. As SMA
actuator generates heat to function, heat transfer simulation was conducted to understand
the heat influence. The enzyme behavior under of applying different current to SMA
spring actuator was tested. A complete testing of the new HPOC system to observe bio-
reagent mixing and reaction under high pressure was conducted and the results were
satisfactory.
xv
1 Introduction
studies and fluorescence measurement is widely used bio-analytical tool with highest
sensitivity and versatility. The current high pressure optical cell (HPOC) doesn’t allow
two or multiple solutions in the same high pressure chamber and mix multi-solution
consists of SMA activator and micro-motor mixer and to build a SMA-activated dual
chamber HPOC, which can sustain a pressure up to ~3 kbar. The specific aims are (1) to
shorten the response time of the shape change of the SMA-plug mechanism so that the
initial rates of most enzymatic reactions can be detected under pressure, (2) to minimize
or mitigate the local heat effect caused by the electrically induced activation of SMA so
that the system being studied is subject to negligible perturbation on thermal equilibrium,
(3) to construct a convenient and reliable SMA plug for the use in the final dual chamber
HPOC, (4) if the proposed SMA activator cannot fulfill the mixing requirements, an
of the SMA-plug mechanism and validity of the design for fluorescence measurements
under pressure and to demonstrate the applicability of the final cell in the pressure studies
of sterol superlattice in membranes. The Ni-Ti Alloy composition in the SMA plug will
be systematically varied. Various shapes of the SMA plug will be constructed and tested.
The interacting effects of pressure and cholesterol on the activity of snake venom
model.
mutagenesis, is a unique and powerful tool for studying molecular recognition because
pressure produces a volume change in the interfacial region where molecular recognition
occurs. (Robinson & Sligar, 1995). Molecular recognition plays a vital role in cellular
flawed control of gene regulation and cellular profileration (resulting in cancer), and
between molecular recognition and human disease, pressure studies should have a
fundamental and broad impact on our understanding of human health related problems.
The new pressure cell can also make important contributions in understanding
organism living under hyperbaric conditions. Since the late 1980s, several deep-sea
2
hydrothermal vents were discovered where surprisingly large amount of organism are
thriving. Using the type of HPOC proposed here we may eventually gain a fundamental
surface dwelling organism cannot. There are also microorganisms living under varying
depths from soil surface. Clearly, pressure studies have a wide biological relevance.
underutilized. This situation arises mainly because the current pressure cells are
cumbersome, expensive and having limited versatility (e.g. unable to introduce additional
components into the sample chamber under pressure and unable to provide adequate
mixing and stirring). High-pressure stopped-flow devices (Goldsack et. Al., 1969;
Heremans et al., 1980; Hui Bon Hoa et al., 1982; Balny et al., 1984; Hara & Morishma,
1988; Bugnon et al., 1996) allow for the determination of the initial rate of the
biochemical reactions under pressure. However they are cumbersome to use because they
contain several components, all of which need to be carefully assembled each time before
flow device is expensive (e.g. ~$70,000) and it not readily compatible with most of the
fluorometers commonly used in research laboratories. Due to the high cost and structure
flow cells.
The SMA-activated HPOC offers a simple and economic way to solve these
$10,000. The new cell will allow the addition of one or more bio-chemicals into a sample
while all the components are under the same pressure and temperature. This design also
3
allows for a thorough mixing of those components within a short period of time. In doing
so, the effects of cofactors and regulators on a given biological molecule at high
or light scattering. Since we use the properties of smart materials, rather than complex
mechanisms, our design makes the mixing and stirring process simple, convenient,
reliable and easy to operate. This design will greatly increase the scope and versatility of
the pressure cell originally designed by Paladini and Weber. Since the Paladini/Weber
cell has already generated so much useful information, as mentioned earlier, the new
pressure cell should become an even more powerful tool in biological research.
High-Pressure Optical Cells has been a very import tool in biophysical studies for
more than twenty years ((Neuman, Kauzmann et al. 1973; Heremans 1982; Silva and
Weber 1986; Wong 1987; Lester 1989). Pressure studies provide a fundamental
isothermal conditions, pressure only produces volume changes while under isobaric
conditions, while temperature not only produces volume changes but also changes in
thermal energy. Pressure results are more straightforward and easier to understand. Under
isothermal conditions, the reaction rate of equilibrium of the reaction can be studied
according to volume changes which is yielded by pressure (Royer, Weber et al. 1986;
Royer, Chakerian et al. 1990). The studies of biological functions under different
pressure can demonstrate which functions are dependent on the volume change of the
system and which are independent of it. For example, membrane transport phenomena is
somewhat dependant on volume change (Chong, Cossins et al. 1983; Chong and Weber
1983), but ordinary chemical reactions in which covalent bond exchanges takes place
4
exclusively are relatively insensitive to pressure (or volume change) (Weber and
maybe revealed at high pressure. Wong and Mantsch discovered five gel phases of
pressures (Wong and Mantsch 1985). Without pressure, it would be impossible to reveal
cell is required for these measurements. Generally, there are two kinds of cells: the
diamond anvil cell is mainly used for high-pressure infrared and two-photon microscopy
absorption and scattering measurements in the cuvette. Since fluorescence is used for bio-
analysis due to its highest sensitivity and versatility, this research focuses on window-
The history of high pressure research started from the living organisms found at
600 bars when the discovery of living organism at 6,000 meters below the surface of the
sea in 1882 triggered the curiosity of P. Regnard and A. Certes from France (Paladini,
1980). Following this, Regnard and Certes studied the behavior of whole animals under
pressure, and the effects on some organic materials such as milk, eggs, etc. From 1900 to
1950 more biological materials such as protein, viruses, antigens and antibodies, tumors
5
and tissues were studied under high pressure (Paladini, 1980). The effects of the pressure
on the materials were evident. But temperature was rarely considered in analyzing results.
after the pressure was released and the effects were mostly irreversible. However the
actual physical changes in the system under pressure were ignored. The research in that
period only proved that pressure had an effect over the systems studied but didn’t use
systems subjected to it. Hydrostatic pressure is a scalar magnitude without direction when
it is applied to the media. Real time study when pressure applied to the system is
important to understand pressure influence on the system while is it applied. It has been
found that high pressure creates volume change. Several investigators studied application
dilute protein (Villas-Boas, Silva et al. 1993). These studies mainly relied on the
observation of changes of absorption or emission of visible and ultraviolet light while the
4.3 ml/mole (Heremans 1982). It also has been approved that the covalent links between
the partners play no significant role in the interactions that lead to formation of the
stacked complex. Covalent bond lengths and angles remain unchanged over a range of
pressure much larger than the one employed. This un-compressibility of the covalent
bonds lengths is of fundamental importance in any analysis of the effect of pressure upon
6
convalently bonded chain. Inclusion complexes of benzenic or fluorescence changes
motivated Weber to design and construct a high-pressure bomb to meet the experimental
requirements. The bomb was designed to meet 3 criteria: hold pressure up to 2 kbars;
with four optical ports and one hydrostatic port to permit its pressurization; easy loading
and unloading samples and the prevention from making contact with either the inner
metallic bomb walls or the pressure transmitting fluid. For the efficiency of the pressure-
delivery pump consideration, the smallest internal volume is desired. This high pressure
bomb is made of a special steel alloy called Vascomax 300, developed by Teledyne
Vasco Company, Chicago. The steel can be hardened after machining to achieve
Rockwell C58 to C62, thus allowing the pressure cell dimensions to be small in
comparison with the dimensions required for standard steel were to be used instead of
monitored is on the inside temperature of the bomb instead of sample temperature since
an extra plug just for temperature monitoring is costly and difficult. The windows were
made of fused quartz T19 Suprasil 1, birefringence. The seal between the window and the
plug port was made by mirror polishing the metal port surface upon which the window
will rest. The seal is achieved when pressure press both surfaces against each other. The
mirror polishing of the plug surface is critical since any gap between quartz and plug port
7
Paladini and Weber’s HPOC has been widely used in fluorescence studies of
proteins, which have led to the discovery of conformational drift of oligomeric proteins
(King and Weber 1986; King and Weber 1986; Weber 1986; Ruan and Weber 1988;
Ruan and Weber 1989; Ruan and Weber 1993). It has also been used in protein folding
(Foguel and Weber 1995; Eftink, Ramsay et al. 1996; J. and A. 1998), protein-DNA
interactions (Royer, Ropp et al. 1992; Villas-Boas, Silva et al. 1993), enzyme kinetics of
membrane proteins (Chong, Fortes et al. 1985) and subunit-subunit interaction of proteins
Despite its great contributions to the biophysical research, the type of pressure
cell made by Paladini and Weber has certain limitations. It does not allow quick mixing
and stirring of additional components under pressure. For example, when using the
fragments at high pressures (Chong, 1985), the enzymatic hydrolysis reaction was
initiated by mixing the substrates, cofactors and enzyme together at atmospheric pressure.
Then the mixture was loaded into the quartz cuvette and the polyethylene tubing (used as
a cap of the cuvette). The cuvette plus the polyethylene tubing formed a closed system,
which was then transferred into the pressure cell and immediately thereafter pressure was
raised to the desired level. The entire process of sample loading, cuvette assembling, and
pressure generation took about several minutes. Finally, the time dependence of the
optical density from the reaction products was recorded which was reflective of the
8
The above operation has the following problems: (1) unless the optical change
with time remains in a linear fashion for >10 min., the operation not permit the
observation of the initial rates of the enzyme reactions. This is a serious problem
particularly for the studies of membrane surface acting enzymes because the activities of
those enzymes depend upon the initial membrane lipid composition (Liu & Chong, 1999);
after the reaction proceeds for a prolonged time, the exact lipid composition in the
membrane is not known. (2) This operation does not start with a true thermal equilibrium
and the enzymatic reaction proceeds prior to change become questionable. (3) With this
operation, maintaining a constant temperature for the sample before and after loading it
into the pressure cell is problematic. (4) The Paladini/Weber cell does not allow the
addition of other components into the sample while the sample is under pressure, which
(Zhou, Ostrow et al. 1998). This device made it possible to mix two samples under high
pressure. However, this device requires the use of an UV laser which is costly and not
readily available in most laboratories. Another drawback is that this device does not
provide sample mixing. To further deal with above problems, Zhou et al proposed to use
shape memory alloy (SMA) as an actuator for both unplugging stopper and mixing
solutions, they came out two designs: a one-way SMA spring plus a bias extension spring;
and, a two-way SMA spring (Zhou, Chong et al. 2000). Both designs have drawbacks in
that the first design suffers from obstruction of optical observation, due to the use of the
9
bias spring blocked the observation window, and both designs were never materialized,
A new HPOC was designed based on Paladini’s high pressure cell (Figure 1-1).
The new cell accommodates a dual chamber cuvette which can separate two solutions
after the system is assembled and pressure is raised to desired level. A smart activation
system resides in the cuvette upper chamber. Electrical lead plug provides electrical
10
1.4.1 Design of the New SMA Activated Dual Chamber HPOC
The design motivation of the cell is the introduction of biological samples until
the experimenter is ready to mix them. Prior mixing, all the samples experience an
identical pressure and the same rate of pressure increase. A specially designed dual-
chamber quartz cuvette holds two samples and locates at the center of the pressure
chamber. Several cuvettes were designed to fit each of the actuator design. A final HPOC
was designed to fit the latest cuvette. To help us to fabricate the HPOC, ISS Modified our
design to fit with their existing cell body. ISS used their existing HPOC as the base,
modified the top inlet plug to accommodate both Electrical lead and inlet. This design
saved a lot of manufacture cost, and made our final testing under pressure possible.
1.4.2 Design of Shape Memory Alloy Activated Stopper and Mixing Systems
combination of SMA springs and micro-motor actuators. The main rule is that the design
should be as simple as possible: if less components are introduced to the cuvette, it will
be easier to operate in bio-physical studies. The first design involves the use of a two-way
SMA wire as an actuator. A two-way SMA remembers both a hot (austenite) and a cold
(martensite) shape. This property is used in order to eliminate the extension spring used
in Zhou’s dual spring design. With this design, it is understood that the agitation of the
two samples can be accomplished without the use of extension spring. This two-way
SMA is trained to a hot length and a cold length. The wire will first unplug the stopper
and then act as an agitator by moving up and down using its Two-Way Shape Memory
Effect (TWSME). TWSME of SMA wire can be trained to remember its high
temperature shape and also its low temperature shape (Johnson 1984). The main
11
drawbacks are that the vibration range is too small and TWSME is not reliable. The
second design goes back to SMA springs as actuator with commercially available
resources. This design uses two tensile SMA springs as unplugging and agitation. This
design can do unplugging and provide vibrations, but has its shortcoming of interference
in the lower observation path, since one SMA spring resides in the lower chamber of the
cuvette. This design is mainly to test SMA spring’s possibility as actuator. Both tensile
and compressive SMA springs were bought for testing and tensile SMA spring actuator
worked well both on unplugging and vibration. Unfortunately, the compressive SMA
spring wasn’t usable at all because it lost memory effect very easily. The third design
comes with tensile SMA to pull plug and compressive SMA spring as biasing spring to
create vibration. Since commercially available SMA springs are very limited and
compressive SMA spring is very unreliable, we manufactured our own SMA springs. By
alternatively applying current to either SMA tensile spring or SMA compressive spring,
their phase was switched from their martensite to their austenite phase. Therefore using
both SMA springs we are then able to maximize the vibration range by fully using their
spring actuator in fourth design which lead to the other cooling spring not being able to
cool down, the fourth design uses a SMA tensile spring with a compressive biasing
spring. As above SMA actuators didn’t achieve rapid and complete mixing, we went with
our alternatively design, the fifth design which introduced a micro-motor as a mixer. The
design combined SMA tensile works as unplugging the stopper and a micro-motor as the
stopper as well as a mixer. The design achieved unplugging and mixing in one. The
12
complete mixing only took 10 seconds. And the heat generation can be controlled to
After SMA spring and micro-motor activated dual chamber cuvette functioned, a
new HPOC should be designed and fabricated according the new cuvette design. Since
the dual chamber is used in cuvette design and SMA based actuator needs space to
activate, the new HPOC size is increased. In addition to the ports required for the optical
observation of the samples, two additional ports are required for the transmission of
electrical current into the HPOC in order to activate the SMA actuator. These ports are
called electrical leads which can be achieved using a heat shrinkable plastic tubing (Terry
& Ruoff, 1972). A FEM analysis of stress and distortion will be conducted according to
the changes to guarantee the new cell can still hold pressure up to 3000 kbars.
After the SMA based actuator can function as an unplugger and a stirrer, and dual
chamber cuvette as well as a new HPOC is fabricated, the system will be assembled and
tested under pressure with ISS K2 fluorescence measurement. The assembly of the
complete system involves the sealing of cuvette, the transfer of pressure into cuvette, and
function of SMA based actuator, etc. The final test is basically the whole system function
test, leakage of cuvette sealing under pressure, the unplugging under high pressure,
working of the stirrer on mixing under pressure, and optical observations of mixed bio-
solutions.
13
2 Design of SMA Actuator
This Chapter will explore SMA as an actuator. From two-way shape memory
actuator to SMA spring actuators, five kinds of SMA based actuators were developed in
improving rapid and complete mixing. Before getting into the design and fabrication of
SMA based actuators, some theory reviews of shape memory alloy are very necessary.
SMA is often referred as smart material. Smart material has a prominent role in
the future for self-repairing and intelligent material (Hodgson, 1975). Smart materials are
a class of materials and composite media having inherent intelligence together with self-
constitute a few subsets of the material family that “manifest their own functions
The first reported steps towards the discovery of the shape memory effect were
taken in 1930s. According to Otsuka and Wayman (Otsuka and Wayman 1998), A.
Olander discovered the pseudoelastic behavior of the Au-Cd alloy in 1932. Greninger &
decreasing and increasing the temperature of a Cu-Zn alloy. The basic Phenomenon of
the memory effect governed by the thermoelastic behavior of the martensite phase was
widely reported a decade later by Kurdiunov & Khandros. Workers at the U.S. naval
14
nickel and titanium, which can be considered an alloy named Nitinol (Nickel-Titanium
Naval Ordnance Laboratory). Many efforts had been made since then, but not until the
mid-1990s, that the first breakthrough in medicine utilizing SMA in commercial stent
Two main characteristics of Shape Memory Alloys are shape memory effect and
pseudoelasticity. SMA is a unique class of alloys which are able to remember their shape
and are able to return to that shape even after being deformed. The phenomenon is called
return to its original shape upon unloading after a substantial deformation. (Funankubo
1987)
amount of possible “elastic” strain, which is more than 20 times higher compared to
austenite in the initial condition and which transform to the martensite (this phase is well-
known from steel hardening) if external stresses affect the material. The phase
transformation is related to the appearance of a pseudo-elastic strain, giving the effect its
name superelasticity. After releasing the stress the material recovers the initial shape and
temperature changes are not necessary for the occurrence of the effect. (Suzuki 1976;
Owen and al. 1979; Delaey and Chandrasekaran 1982; Tamura 1987)
15
At present, most interesting applications of the superelastic effect can be found in
medical implants and instruments, where for example the availability of thin-walled
tubing in the market of stents has opened for the memory alloys. Due to high elasticity
the stents can be folded into catheter sheeth and pushed into the vessel, where they
expand and provide a force to keep the vessel open. (Muddle 1990; Liu, Kunsmann et al.
the driving force behind shape memory alloys. When nickel and titanium atoms are
present in the alloy, in a given proportion, the material forms a crystal structure, which is
capable of undergoing a change from one form of crystal structure to another. The
transformation from one form of crystalline structure to another creates the mechanism
by which the shape change occurs in SMAs. This change involves transition from a
monoclinic crystal form (martensite) to an ordered cubic crystal form (Austenite). The
Austenite phase is stable at high temperature, and the martensite is stable at lower
temperature. In the martensite phase, atoms orient themselves in rows that are tilted left
or right. This phenomenon is referred to as twining, as the atoms form mirror images of
themselves or twins. The martensite twins are able to flip their orientation, in a simple
shearing motion, to the opposite tilt, creating a cooperative movement of the individual
twins. The atoms reorient themselves to straight up the lattice when the deformed SMA is
heated. Ordinary metals can’t reverse the deformation once deformation occurs by
dislocation motion and atomic planes sliding over one another, taking on a new crystal
position and results in permanent damage to the crystalline order. Unlike regular metals,
SMAs deform by detwinning which changes the tilt orientation of twins and doesn’t
16
cause any dislocation. Detwinning allows the martensite phase to absorb dislocation to a
given extent. Shape recovery only occurs to detwinning, but not to a deformation process
involving slip which is not irreversible. (Chu and Tu 1994; Ren and Otsuka 1997)
SMA in its Austenite phase exhibits a high elastic behavior. This allows the
material to deform up to 7% and then fully recover the resulting strain by simply
it is called stress-induced martensite. Applications using the Shape Memory Effect results
composition of the alloy and can vary from 100°C down to cryogenic temperatures. (kato
and Mori 1976; Otsuka and Shimizu 1982; Pelton, Hodgson et al. 1995)
temperature (Ms), Martensite finish temperature (Mf), Austenite start temperature (As)
and Austenite finish temperature (Af). Martensite phase starts to form when temperature
drops below Ms and completes transformation when temperature drops below Mf.
Austenite phase does the opposite and start to form when the temperature is raised up to
As and finish transformation when temperature is above Af. There is a gap between Ms
and Af, it is called hysteresis. The temperature range for the martensite-to-austenite
transformation, i.e. soft-to hard transformation that takes place upon heating is somewhat
higher than that for the reverse transformation upon cooling. The difference between the
generally defined as the difference between the temperature at which the material is 50%
transformed to austenite upon heating and 50% transformed to martensite upon cooling.
17
The difference of heating and cooling temperature can be up to 20-30°C, which means
(Af < 37°C ) would require cooling to about as low as 7°C to fully retransform into
The only two alloys to have achieved level of commercial exploitation are NiTi
Alloys and the copper-base alloys. Properties of the two alloys are quite different. The
Niti alloys have greater shape memory strain (up to 8% versus 4 to 5% for the copper
base alloys), tend to be much more thermally stable, have excellent corrosion resistance
stress-corrosion cracking, and have much higher ductility. On the other hand, the copper-
base alloys are much less expensive, can be melted and extruded in air with ease, and
have a wider range of potential transformation temperatures. The two alloys both have
has moderate solubility range for excess nickel or titanium, as well as most other metallic
elements, and it also exhibits ductility comparable to most ordinary alloys. The solubility
allows alloying with many of the elements to modify both the mechanical properties and
transformation properties. Heat treating to impart the desired memory shape is often done
at 500 to 800°C (950 to 1450°F), but it can be done at 300 to 350°C (600 to 650°F) if
18
sufficient time is allowed. The SMA component may need to be restrained in the desired
memory shape during the heat treatment; otherwise it may not remain there.
such as boron, cerium, cobalt, iron, titanium, vanadium and zirconium are also added for
grain refinement. The thermal stability of copper-base alloys is ultimately limited by the
decomposition kinetics. For this reason, prolonged exposure of CuZnAl and CuAlNi
alloys at temperatures above 150°C and 200°C respectively, should be avoided. Aging at
lower temperature may also shift the transformation temperatures. In the case of aging the
beta phase, the results from the change is in long-range order (Schofield and Miodownik
1980). When aged in the martensitic state, the alloys exhibit an aging induced martensite
What make SMAs unlike other metallic materials which induce large amount of
deformation without permanent deformation? How can they sustain large deformations
and recover their strain upon temperature change or loading? The unique characteristic of
SMAs is due to their phase transformation mechanism. Some essential metallurgy can
explain these characteristics. When the atomic percentage of Nickel in Nitinol is close to
50%, the structure is called near-equiatomic. The equilibrium Ni-Ti alloys make the
shape memory effect and transformation pseudoelasticity occurs. The atomic structure
structure to the phase with a monoclinic B19’ structure, or to more often the two-step
transformation from the β to a trigonal phase (so called R-phase) and then to the B19’
19
phase. Ni content, aging thermo-mechanical treatment and addition of alloying elements
Equilibrium Ni-Ti phase diagram in the vicinity of NiTi shows the B2 region is
very narrow at the temperature below 923K (650°C). B2 region below 923K is generally
accepted only between 50.00 and 50.5 at% Ni. When nickel contents exceeds 50.5at%,
temperature below 973K (700°C) after quenching from a high temperature, the final
product of the composition is a mixture of TiNi3 and TiNi. In the early stage of aging at
low temperatures, a metastable phase Ti3Ni4 forms as fine platelets with coherency to the
matrix and affects the properties of the TiNi alloys. Ti3Ni4 has a rhombohedral structure,
the space group R3. The unit cell of this structure contains 18 titanium atoms and 24
nickel atoms. The precipitate particles of Ti3Ni4 are lenticular shape and coherent to the
matrix in early stage and produce strain fields around them. These strain fields produce
the all-round shape memory effect. Precipitation of Ti3Ni4 strengthens the matrix B2
Fully annealed near-equiatomic Ti-Ni alloys transform from the B2 parent phase
Ni alloys transform in two steps, from the B2 parent to the R-phase and then to the B19’
phase. B19’ martensite has a monoclinic structure and R-phase has a trigonal structure or
20
martensitic transformation. The thermoelastic martensitic transformation causing the
shape memory recovery is a result of the need of the crystal lattice structure to
accommodate to the minimum energy state for a given temperature. In NiTi, the relative
symmetries between the two phases lead to a highly ordered transformation, where the
shape change on a macroscopic scale. The crystal structure of martensite is relatively less
symmetric compared to that of the parent phase. If a single crystal of parent phase is
equivalent habit planes are generally created. There is however, only one possible parent
phase (austenite) orientation, and all martensitic configurations revert to that single
defined structure and shape upon heating above Af. The mechanism by which single
symmetry displacement of atoms across a particular atom plane, the twinning plane. Not
like most metals deform by slip or dislocation, NiTi responds to stress by simply
changing the orientation of its crystal structure through the movement of twin boundaries.
A NiTi specimen will deform until it consists only of the correspondence variant which
produces maximum strain. However, deformation beyond this will result in classical
plastic deformation by slip, which is irrecoverable and therefore has no “memory effect”.
If the deformation is halted in the midway, the specimen will contain a parent phase with
an orientation identical to that existing prior to the lattice correspondences between the
original parent phase and each variant martensite has a more complex rhombic structure.
This phenomenon causes the specimen to revert completely to the shape it had before the
deformation.
21
Shape memory alloys (frequently referred to as smart materials), such as nickel-
heated from the weaker, low temperature form (martensite) to the stronger, high
temperature form (austenite). Nitinol with 55% Ni and 45% Ti has mechanical properties
as shown in Table 1. It can be seen from Table 1 that nitinol is much easier to deform at
low temperature than that at high temperature due to its low shear modulus and Young’s
modulus. Figure 2-1 shows phase transformation versus temperature curve. Hysteresis of
austenite starts to form when temperature reach As (austenite start temperature) and
finish temperature).
22
Table 1 Nitinol mechanical properties
Figure 2-2 shows stress and strain curve of transformation. In heating process, as
temperature increases, strain recovers slightly before As and recovers significantly after
Af .
The shape memory transformation has been observed for many years. (Buehler;
Bobrowsky 1968; Hodgson, Wu et al. 1989) This inherent phase transformation is the
basis for the unique properties of these alloys, in particular, shapes memory and super-
elasticity. Generally, these materials can be plastically deformed at some relatively low
temperature, and upon exposure to some higher temperature will return to their original
shape prior to the deformation. The martensite is easily deformed to several percent of its
23
original length and shape at quite a low stress, whereas the austenite possessing a higher
shape memory alloy Figure 2-3, the austenite phase A and two martensitic twin variants
uniaxial loading forms a layered structure (Chang and Read 1951), each layer being in
one of the three above phases. Motivated by this observation, the SMA shape
transformation process can be described as follows Figure 2-4. The far left is in
martensite phase at low temperature with M+ and M-twined next to each layer. When a
uni-axial force is applied, the twined structure will change to a single direction, either M+
or M- Figure 2-3. By heating up, the phase transformation will be achieved following
Figure 2-1 and Figure 2-2. In the heating process, Nitinol material can generate high
force and strain recovery. In the cooling process, A is simply turned to M+ and M- and
24
Figure 2-4 The shape transformation process (Ryhänen)
In most cases, the memory effect is one-way. That is, upon cooling, a shape
memory alloy does not undergo any shape change, even though the structure changes to
martensite. When the martensite is strained up to several percent, however, that strain is
retained until the material is heated, at which time shape recovery occurs. Upon cooling,
the material does not spontaneously change shape, but must be deliberately strained if
shape recovery occurs. Upon recooling, the material does not spontaneously change
shape, but must be deliberately strained if shape recovery is again desired (Foos, Frantz et
memory ability in which the shape change occurs upon both heating and cooling. The
25
amount of this shape change is always significantly less than the change obtained with
one-way shape memory, and very little stress can exerted by the alloy as it tries to assume
its low-temperature shape. The heating shape change can still exert very high forces, as
to create the two-way shape memory effect (Johnson 1972; Wilson 1972). All these
methods rely on the introduction of microstructural stress concentrations which cause the
martensite plates to initiate particular directions when they form upon cooling, resulting
Two-way shape memory effect is that the shape memory element will return to a
However during cooling with the two-way effect, the material simply recovers its low
temperature shape but can not exert a force to the external mechanical components
(Warram 1993; Takagi, Khovalio et al. 2003). Without the external force, it is difficult to
achieve the two-way shape memory effect. In this case, two-way shape memory is
unreliable.
A two-way shape memory alloy can be trained in following process: the alloy is heated
under a first training stress to a temperature which is above the upper operating
temperature of the alloy, then cooled to a cold temperature below the zero-force transition
temperature of the alloy, then deformed while applying a second training stress which is
greater in magnitude than the stress at which the alloy is to be operated, then heated back
to the hot temperature, changing from the second training stress back to the first training
26
stress. Huang and Toh also showed the feasibility of using reheat treatment to obtain two-
way shape memory effect (Huang and Toh 2000). By conducting reheat treatment on
shape memory alloy, two-way shape memory was introduced into initially one-way shape
memory alloy as indicated by the change of the diameter cycling. By embedding SMA
wire with two-way shape memory effect in silicone elastomer cylindrical rods, the shape
Thin-film NiTi SMA membrane exhibits the two-way effect (Ma and Wang 2000;
Gill and Ho 2002). Thin-film NiTi SMA wasis sputtered-deposited onto a silicon wafer in
Nitinol and other shape memory alloy mill products - bar, wire, ribbon and sheet
are normally finished by cold working to achieve dimensional control and enhance
surface quality. Cold working suppresses the shape memory response of these alloys. It
also raises the strength and decreases the ductility. However, cold work does not raise
the stiffness of the material. Heat treating after cold working diminishes the effects of
cold working and restores the shape memory response of these alloys. Therefore, in
order to optimize the physical and mechanical properties of a Nitinol product and achieve
shape memory and /or superelasticity, the material is cold worked and heat treated.
The mill product supplier normally provides the material in the cold worked
condition. The maximum practical level of cold work will be limited by the alloy and by
the product section size. Binary superelastic NiTi alloy fine wires with As in the range of
27
–25 to +95°C are typically supplied with cold reduction after the final anneal in the range
of 30 to 50%. Higher reductions are sometimes used for very fine wires. These same
alloys will be limited to about 30% maximum cold reduction in larger diameter bar sizes.
Binary NiTi alloys with very low As in the range of –50 to –60°C will not sustain the
Both superelastic and shape memory properties are optimized by cold work and
heat treatment. This thermo-mechanical process is applied to all Nitinol alloys although
different amounts of cold work and different heat treatments may be used for different
alloys and property requirements. Shape setting is accomplished by deforming the Nitinol
to the shape of a desired component, constraining the Nitinol by clamping and then heat
treating. This is normally done with material in the cold worked condition, for example
cold drawn wire. However, annealed wire may be shape set with a subsequent lower
In shape setting cold worked material, care must be taken to limit deformation
strain to prevent cracking of the material. Another approach is to partially anneal the
wire prior to shape setting. Yet another option is to shape set in incremental steps. Smith
et al. reviewed the types of furnaces and fixture hardware or mandrels that have been
used in heat treatment. Many types of furnaces have been used including box furnaces,
continuous belt hearth furnaces, tube furnaces, heated platen presses, vacuum furnaces,
induction heaters, salt baths and fluidized bed furnaces. The electrical resistance of
Nitinol makes it a good candidate for self heating by electric current. Nitinol will be
oxidized when heat treated in air. Therefore, surface requirements and atmosphere
28
Shape setting can be done over a wide temperature range from 300°C to 900°C.
However, heat treating temperatures for binary NiTi alloys are usually chosen in the
Consideration must be given to the mass of the heat treating fixture as well as the mass of
the product. Sufficient time must be allowed in the furnace to get the entire mass to the
desired temperature.
The shape setting heat treatment changes the physical and mechanical properties
of Nitinol. Morgan and Broadly mapped the effect of temperature and time at
temperature on shape set wire properties. (Morgan and Broadley 2003) Their response
curves illustrate that physical and mechanical properties do not always change in the
same direction. Also, some properties are not monotonic functions of time at temperature.
For example, upper plateau stress goes through a minimum as a function of time at
temperature when heat treating a superelastic alloy in the range of 450°C to 550°C. This
can be understood in terms of the complex precipitation response of the nickel rich Ni –
Ti alloys.
Precipitation processes in nickel rich NiTi were studied in detail by Nishida et al.
Their TTT diagram shows that in a Ti – 52 atomic % Ni alloy heat treated below 820°C,
precipitation starts as fine Ti11Ni14 transitions over time to Ti2Ni3 and terminates after a
long time as TiNi3 in equilibrium with the NiTi matrix. All the while, the NiTi ratio in
the matrix is being shifted towards higher Ti content and higher transformation
isothermal heat treatment on the physical and mechanical properties of Nitinol wire. This
29
work suggests that for a 50.8 atomic % Ni alloy Ti11Ni14 dissolves at about 500°C and
properties as NiTi ratio in the matrix goes through a peak. Furthermore, this analysis
suggests that the temperature for transition from Ti11Ni14 precipitation to Ti2Ni3
SMAs can be attained. Human body temperature (~37°C) SMA is widely used for
martensite and austenite states. Body temperature SMA works best from room
(37°C).
About 8% strain can be recovered by unloading and heating. Strain above the
limiting value will remain as a permanent plastic deformation. The operating temperature
for shape memory devices must not move significantly away from the transformation
range, or else the shape memory characteristics maybe altered. A shape memory NiTi
30
deformation limit determined by distinctive implant design (sharp angles, etc) and the
range of motion, often for many cycles. Such an application is the circuit board edge
connector made by Beta Phase Inc. (Krumme 1987). In this electrical connector system,
the SMA component is used to force open a spring when the connector is heated. This
allows force –free insertion or withdrawal of a circuit board in the connector. Upon
cooling, the NiTi actuator becomes weaker and the spring easily deforms the actuator
while it closes tightly on the circuit board and forms the connections. Based on the same
principle, CuZnAl shape memory alloys have found several applications in this area.
(Waldbusser 1987) Among the SMA force actuator, SMA spring is mostly used. (Baz,
mechanism by using only a selected portion of the recovery because the transformation
occurs over a range of temperature rather than at a single temperature. A device has been
developed by Beta Phase Inc. (Hodgson 1988) in which a valve controls the rate of fluid
flow by carefully heating a shape-memory alloy component just enough to close the
A number of products have been brought to market that use the psudoelastic
absorb large deformations without damaging the frames are now marketed, and guide
wires developed for steering catheters into vessel in the body using NiTi wire, which
31
resists permanent deformation if bent severely. Arch wires for orthodontic correction
using NiTi have been used for many years to give greater rapid movement of teeth.
temperature of the Ni-Ti Alloy (composition: 51% Ni and 49% Ti) used in Zhou’s SMA
proposal was 80°C, which was too high for normal biological studies. We have to find
body temperature SMA commercially available with an economic price. The temperature
transformation hysteresis, T1, for most Nitinol alloy is 20°C, which make the control of
temperature before and after activated extremely tight. As body temperature SMA is
selected and the activation temperature is around 37.5°C, the temperature needs to be
Fortunately, body temperature SMA wires were found, and more exciting things is that
with some additives, the temperature transformation hysteresis is reduced to 15°C from
some companies. The 5°C makes a lot of difference since 23.5°C is room temperature.
lower transformation temperature of the alloy would also allow for a faster cycling
between heating and cooling, hence facilitating a thorough mixing of the samples.
Specially, in the future experiments on biological systems, we will use Ni-Ti alloys with
transformation rate at which heat can be delivered to the structure. Since we are using
electrical lead to generate and conduct the heat, it can be very quick. Shape recovers less
32
than one millisecond has been observed (Hodgson, 1988). If in the future study we can
develop a thick SMA wire which has a response/recovery time of 10 miliseconds, then a
thorough mixing and agitation, for example total of 10 agitation cycles can be finished in
100 miliseconds. Since the SMA and the surrounding biological molecules will be
should have a negligible effect on the thermal equilibrium of the biological samples
examined. Notice that there will be a circulating bath controlling the temperature of the
pressure cell, the local heating problem can be eliminated or greatly reduced by selecting
an appropriate Ni-Ti alloy. The heat transfer analysis will be conducted to find out if the
heating is local and also the heating time and cooling time
The third challenge is two-way SMA. Although the two-way SMA has been
discussed for a long time and it has many advantages and potential applications, the two-
way SMA is still not commercially available. The main reasons are that the material
property and behavior of two-way SMA are still not clear and not easy to control, and
there is not a standard way to fabricate two-way SMA. In this study we will conduct
extensive materials study to find the composition relations, mechanics properties, and
structure behaviors of the two-way SMA. To create two-way SMA in Ni-Ti alloys
involves a somewhat complex training process and can use the following methods
(Duerig, Metton, 1989): overdeformation method, shape memory cycling method, and
The fouth challenge is pulling-up force. The pulling-up force produced by the
SMA wire can be calculated. The calculation helps to select a right SMA wire size and to
program the SMA change that is required to overcome the weight of liquid on the plug
33
and the frictional force between the urethane plug and the cuvette. In this calculation we
need to consider the sealing pressure and the friction coefficient between the cuvette neck
and the urethane plug. We also need to carefully design a SMA wire in terms of its length,
wire diameter, and material strength, so it will be able to pull off the plug and generate
agitation. Since in the cuvette, the top and bottom of the plug are under equal pressure,
The design of SMA Actuator starts at the Two-Way Shape Memory Effect
(TWSME). The SMA spring comes into the design with different combinations.
Commercial available SMA springs were purchased from Robot Store to test the concept.
Due to its size and available compressive spring wasn’t usable for use, Nitinol wire was
The first design (Figure 2-5) is using two-way shape memory effect (TWSME) to
design a simple structure with one SMA string to pull the stopper and then using the
TWSME effect to make the stopper vibrate up and down by supplying square wave
current to the SMA string. While the string is heating up, it will shrink and pull the
stopper up. While the current is cut off and the temperature will drop, the SMA wire will
return to its Martensite phase alternate shape which is longer and drop the stopper
downward.
34
Figure 2-5 Two-Way SMA actuator design
The ability of shape memory alloys to recover a preset shape upon heating above
its transformation temperatures and return to an alternate shape upon cooling is known as
two-way memory. Two-way memory is unique in that the material “remember” different
high temperature and low temperature shapes. Much has been written about this unique
phenomenon in the literature and it has triggered many ideas in the minds of product
z Constrained aging for long periods of time (to yield the all-round shape
35
effect which is slightly different than two-way memory).
z The memory can be erased with very slight overheating (as low as 250˚C).
z The long-term fatigue and stability characteristics are not well known.
Wherever possible, it is better to modify the device design to make use of one
way memory with biasing force acting against the shape memory element to return it
upon cooling. All of above limitations are addressed by this approach except for the
inherent temperature hysteresis, which remains. Two-way actuators using one-way shape
memory elements acting against bias forces have demonstrated large strains, high forces
in both heating and cooling directions, and excellent long-term stability equal to millions
of cycles.
Two-way shape memory effect is that the shape memory element will return to a
However, during cooling with the two-way effect, the material simply recovers its low
temperature shape but can not exert a force to the external mechanical components.
Without the external force, it is difficult to achieve the two-way shape memory effect. In
36
this case, two-way shape memory is unreliable for this design. Another limitation is the
deformation of 2% which is way too less for our application, which requires more than
15% at the least. This brings us back to the SMA springs which provide larger
deformation difference.
This design is not practical main for two reasons: 2% shape change will lead to a
very long cuvette to unplug a 5 mm stopper (at least 250 mm long for the upper chamber);
the shape memory effect is not stable on cooling process and its cooling shape
transformation doesn’t produce any force, which is a big problem for accelerate the
motion of the stopper and the agitation cannot be realized. The failure of this design leads
product from the robotics store. This design (Figure 2-6) was used only to test the
concept using two tensile SMA springs as an actuator. By apply current alternatively to
either spring, one is heating up while the other one is cooling down, the maximum
vibration range is achieved. The motivation to use two SMA springs instead of one SMA
spring and one biasing spring is to utilize SMA spring’s difference between high Young’s
modulus on high temperature Austenite phase and the low Young’s Modulus on low
temperature Marstensite phase. The difference of Young’s Modulus between two phases
is eight times.
37
Figure 2-6 Dual SMA spring actuator design
The use of a shape memory alloy in the form of a helical spring appears to have
the most promise in actuator applications. A shape memory spring has reasonable
force/motion characteristics, a compact size, and simplicity of design. The force that a
spring of any material produces at a given deflection depends linearly on the shear
modulus (rigidity) of the material. Shape memory alloys exhibit a large temperature
dependence on the material shear modulus, which increases from low to high temperature.
Therefore, as the temperature is increased the force exerted by a shape memory spring
dimensional plot of stress, strain, and temperature, as shown in Figure 2-7. In fact,
simultaneous variation of stress, strain, and temperature results in two surfaces, one for
heating and the other for cooling; although only one surface is shown in Figure 2-7 for
38
clarity. The two surfaces enclose a three dimensional envelope.
When the envelope is cut by a plane perpendicular to the strain axis, the points of
intersection form the stress-temperature hysteresis curve at the particular value of the
strain where the plane lies. In a shape memory actuator, this hysteresis would correspond
mechanical properties). When the envelope is cut by a plane perpendicular to the stress
axis, the points of intersection form the strain-temperature hysteresis curve for the stress
39
Figure 2-8 Constant deflection configuration (Warram 1993)
hysteresis curve, under a constant force loading (refer to Figure 2-10, Figure 2-11 and
Figure 2-12).
40
Figure 2-10 Constant force loading of a shape memory compression spring
(Warram 1993)
41
Figure 2-12 Superposition of linear Force-Deflection biasing spring (Warram, 1993)
of stress, strain, and temperature occurs, as in the case of a shape memory spring working
against a biasing force with a corresponding variation of force, motion, and temperature
Figure 2-12. For engineering purposes, it is convenient to plot a number of shear stress -
shear strain curves measured at various temperatures, rather than to deal with a three
dimensional plot. These stress – strain curves are referred to as isotherms, since they are
42
Figure 2-13 Typical stress-strain behavior of Shape Memory Alloy material
(Warram, 1993)
behavior of a test spring, at several temperatures, and by then converting the force-
dδ (2-1)
Shear _ strain = γ =
πD 2 n
2 D ⎛ ∆P ⎞
shear _ stress = T = 3 ⎜
δ + 3P ⎟
πd ⎝ ∆δ ⎠ (2-2)
d = wire diameter
P = force
43
δ = deflection
∆P
∆δ = slope of force-deflection curve
Figure 2-14 shows how the data are measured for one point on a force-deflection
curve. Higher values of strain in Figure 2-13 and Figure 2-14, the stress-strain curves are
nonl-linear, thus facilitating the use of the slope parameter to account for the curvature.
The design of shape memory actuators which utilize the non-linear portion of the stress-
transformation in their crystal structure when it is heated from the weaker, low
temperature martensite form to the stronger, high temperature austenite form. Nitinol
with 55% Ni and 45% Ti has mechanical properties as shown in table 1. It can be seen
from table 1 that nitinol is much easier to deform at low temperature than at high
44
temperature due to its low shear modulus and Young’s modulus. Different diameter
spring forces are simulated using Matlab Figure 2-15, Figure 2-16 and Figure 2-17. The
minimum force needed to unplug the stopper is 3N. By using safety factor of 1.5, we
need 4.5N minimum to unplug. The bigger the wire diameter, the more heat generated.
Minimum wire diameter is more desirable. From Figure 2-15, to generate 4.5 N of force,
the extension of the spring needed is less than 10 mm for 10 coil spring with diameter of
0.135” of wire diameter of 0.019”. Increasing spring diameter and decreasing wire
diameter can both decrease force generation ability (refer to Figure 2-16). Figure 2-17
45
Figure 2-16 Force vs Extension at high and low temperature (D=0.156”; d=0.018”)
Figure 2-17 Force vs Extension at high and low temperature (D=0.156”; d=0.017”)
46
2.3.2.3 SMA Spring Actuator Dynamics
The equation differs from regular spring dynamic in that its spring constant can change
⎛ (K − K ) (K ) ⎞ + K
⎜ h l
f ( vt ) + h
⎟ ( x − 10 )
l
⎝ 2 ⎠ 2
(2-3)
2
dx d x
− K biase ( x − 30 ) + λ = m
dt dt 2
Where, Kh denotes spring constant of SMA spring at its high temperature phase,
Kl denotes spring constant of SMA spring at its low temperature phase, and Kbiase denotes
Gh d 4
Kh = = 0.538 N / mm
8nD 3
Gl d 4
Kl = = 0.0710 N / mm
3nD 3
K biase = 0.28 N / mm
K=Gd/8C3N
E
G – shear Modulus µ=
2(1 + ν )
47
D=0.135 d=0.019
D= 0.156; d = 0.018;
D = 0.156 d = 0.017;
Equation below is a simplified version and uses sin(vt) equation to simulate the
force change.
dx d 2x
( K sma sin(vt ) + K )( x − 10) − K biase ( x − 30) + λ =m 2 (2-4)
dt dt
This system can be illustrated as Figure 2-18. By solving the above equation in
Matlab, vibration can be simulated as Figure 2-19, in which damping ratio of 0.1 was
selected. The simulation gives very encouraging result with 250Hz frequency and 25mm
vibration range. If we can apply input current on its natural frequency, resonance can be
48
Figure 2-19 Spring dynamics Matlab simulation plot
Due to the limitation of current slow relay response and extra electrical lead, the
testing result was with 3mm vibration range and 2Hz frequency. A quick response relay
and elimination of electrical lead will improve the vibration. The following figures and
information show a part of the investigations concerning SMA springs. Different SMA
springs have been produced and tested for several companies. These two images indicate
the load-deformation-behavior of two different SMA springs with different diameters and
transition temperatures. At high temperatures (here 100°C) the spring acts like an
ordinary steel spring, but at low temperatures (here 23°C) the spring yields at a constant
load.
49
2.3.2.3.2 Spring Responses to Electrical Current in the Air
The most common way to heat up SMA wire is applying electrical current. To
understand response of SMA wire to electrical current helps to design SMA actuators.
The extension responses to current before the temperature reach austenite states. Figure
2-20 shows two SMA wires response to a 4V step voltage, and it can be seen that the
Figure 2-20 Heating and cooling time histories of actuators subjected to step voltage
changes of 4 volts (a) 0.75mm (b) 0.875mm (Baz, Iman et al. 1990)
50
Cooling is a very important issue with SMA actuator design. If the application is
in the air, circulation helps cooling efficient. Figure 2-21 shows that by increasing the air
velocity from 0 to 7.2, the cooling time reduces from 58 seconds to less than 10 seconds.
On a repeating heating and cooling cycle, cooling condition is usually set for application
Figure 2-21 Effect of cooling strategy on heating and cooling time histories of 0.4mm
51
Figure 2-21 shows 0.4mm actuators subjected to sinusoidal voltage excitations
with air cooling velocity = 7.2 m/s, as the frequency increases, the wire can’t fully cool
to its low temperature and its low temperature phase can’t fully recovered. By increasing
any amount frequency over 1/36Hz, it is sacrificing its full working displacement range.
But since when cooling approach its low temperature it slows down, the selection of
Figure 2-22 Heating and cooling time histories of 0.4mm actuators subjected to
sinusoidal voltage excitations with air cooling velocity =7.2 m/s: (a) at 1/36Hz (b) at
52
From the above illustration, it seems that cooling is big issue on SMA response
frequency. Also the above result is based on air cooling, our media is water based protein
and a cycling cooling bath. Our cooling is much faster, a heat transfer analysis or
temperature monitoring experiment should be conducted to find out our cooling time.
To test two commercial spring actuator concept and to provide power and
power supply (Tenna with max 30volt or 30ampere), Relay(Omron 12VDC, 10Ampere),
circuit board, etc. To test the response of SMA spring, a circuit is built as in Figure 2-23.
The frequency can be adjusted by selecting R1, R2 and C values. R stands for resistor and
C stands for capacitor. The ratio of R1 over R2 will control the percentage of peak value.
TDS2024), Amplifier Circuit Board (Figure 2-23), Thermo couple with 0.005 mm in
53
diameter on the tip, DAQ card data collection with Labview analysis. To measure the
temperature, two methods will be used: Oscilloscope monitoring and DAQ card
collecting data with Labview. Both methods will use the following amplifier to convert
measured signal to a measurable value and cold junction compensation built in the circuit.
Figure 2-24 (a) Circuit board and relay (b) Two SMA spring testing in water
It is very hard to measure the temperature of SMA wire, since the ambient cooling
effect is very fast. This experiment tried to get some idea of temperature response to
electrical current. Figure 2-25 shows the temperature in the adjoining water of SMA
54
spring, the temperature response to electrical current is not the temperature of SMA wire,
but the ambient. The result shows the temperature change, but it is far from 15 degree
different. It is not possible to inquire temperature data, so we need to use FEA method to
find out the temperature response to electrical current. FEA simulation will be shown in
following sections. It took carefully designing of FEM model and boundary condition to
(Table 1). When the period is decreased to 110 ms, the vibration range can’t be increased
calculated as follow:
55
Where R=0.8ohm; K unknown, but small, ∆L=5mm
Ignoring heat lose in the air. I, applied current; R, SMA spring resistor; K, SMA
spring coefficient; ∆L, spring travel length; h, convective heat transfer coefficient; ∆T,
2.3.2.5 Conclusion
The SMA spring can unplug a stopper; the actuator can generate 10 mm vibration
range in open water container; the reaction/response frequency can achieve to 1 Hz; the
ambient water is heated up after 15-20 cycle to 5°C. Measuring temperature of SMA
spring is difficult since water cools the surface of SMA wire much quicker than air does.
To overcome the defect of second design, we put both springs on the same side to
leave the bottom chamber of cuvette clear for observation. The new design includes two
SMA springs, one tensile and one compressive. This will solve the problem of the
previous design in which the bottom tensile SMA spring blocks observation light and
interfere with the readout. We ordered some tensile and compressive SMA springs from
the Robotics Store, the tensile spring which had been utilized on previous design
wouldn’t fit into our small cuvette. The commercial compressive spring didn’t perform
very good SME, and lose its SME easily by a little stretch. Also, its diameter is way too
big for our utilization. After consulting with a few companies, which charge $2000 per
customized SMA springs, we decided to make our own springs according to our surveyed
research paper and patent. Both tensile and compressive springs were made by ourselves
56
and functioned well separately. When they work together, the heat transfer analysis
shows that it was very critical to set both springs position to avoid either of them work
improperly.
In this design, we have two self made SMA springs, one is tensile and the other is
compressive. The diagram of this design is show as Figure 2-26. With both spring on one
side, the bottom portion of the cuvette is cleared for observation light to pass through.
Figure 2-26 Dual SMA springs actuator with tensile and compressive SMA springs
To make SMA springs, a fixture will be used to keep the spring shape and heat
treatment will be conducted. Fixture materials must withstand the temperatures needed
for shape setting to provide the strength needed to maintain required geometries without
losing strength and hardness. At the same time, corrosion loss and oxidation must be
57
minimized. Figure 2-27 a) shows how to make a SMA spring: a SMA wire is wound to a
mandrel fixture according to desired pitch and both ends must be secured to fix the shape
during heat treatment (Smith and Hodgson 2003). After heat treatment, the wire will be
Heat treatment will influence the Austenite final temperature (Af) and oxidation
(Morgan and Broadley 2003). The longer the heat treatment, the lower the Af and the
more oxidation. Figure 2-28 shows the relationship of Af with both heat treatment
58
temperature and time. The higher the heat treatment temperature, the lower the Af, The
composition of nickel in nitinol determines mostly the Af, but heat treatment temperature
can do small adjustment. Heat treatment time can also slightly adjust the Af in a very
limited range but will create more corrosion and oxidation. Heat treatment temperature
and time can also modify ultimate tensile stress Figure 2-28 (Morgan and Broadley 2003).
Figure 2-28 Surface response for Af Temperature and ultimate tensile stress
2.3.3.3 FEM Heat Transfer Analysis for SMA Tensile and Compressive Spring
The heat transfer process is simulated using FEMLab. A 2-D for 3-D problem
simplified structure is used for the model: two cylinder shape cross-sections represent
two Nitinol springs (the heat generator), quartz cuvette (close to insulator), metal cap,
59
Figure 2-29 Femlab heat transfer FEM Analysis for two SMA springs
The left is axial symmetric insulation boundary, the right and the top are air
convection cooling and the bottom is cooled by a cooling water pump. By applying
alternative heat flux to either of the springs, each spring will go through its heating and
cooling process. The distance of the tensile spring to the center and the distance of the
compressive spring to the quartz cuvette need to be adjusted carefully to achieve this
periodic high and low temperature pattern. Temperature changes are most likely to occur
60
on both springs. Fluid and cell body do not have significant temperature rise. From the
simulation, the spring temperature in some cycles can reach to 370K which is undesirable
and much higher than 313K. The heat source needs to be optimized for each cycle and
variable amplitude of electrical current will be used. The close sine curve is the
temperature change of point on the SMA spring according to current change, the near
point still carry the sine curve with much smaller amplitude, and the far point is almost
not influenced buy cooling down with the cool bath. According to the simulation, the heat
can be dissipated to the environment and the testing solution temperature is kept (Figure
2-30).
61
2.3.3.4 Conclusion
In this stage, SMA springs were fabricated by ourselves and their shape memory
response were very stable and repeatable both on tensile and compressive SMA springs.
We also got wide range of body temperature SMA wire from 0.012” to 0.18” and we can
make springs according to our needs and select the best fit diameter. With numerous
testing, we found out 0.014” in diameter wire performs the best in our design. It can
provide enough force to unplug and cooling faster due to its smaller surface area to
volume ratio. The actuator can produce 10mm range vibration with 1 Hz frequency in
open water. Comparing to commercial SMA spring, the spring length is much shorter, 10
rounds compare to 20 rounds, the vibration range was doubled and also the heat
generation is much smaller due to its smaller diameter wire. Assembly test wasn’t very
successful. When the actuator was assembled in cuvette, the heat generation was way
faster, after 5 to 8 cycle, the vibration stopped and this agrees with heat transfer analysis
result which shows that tensile and compressive springs were ever sensitive to their
Actuator
Due to SMA compressive spring heat interfere with SMA extensive spring from
cooling down and reheat up, SMA compressive spring was replaced with a biasing
62
Figure 2-31 SMA actuator with tensile SMA and biasing compressive spring
To investigate the heat transfer effect of SMA to the system, a more detailed FEM
model is build with Femlab Figure 2-32. The heat source is from electrical current going
through the SMA spring wire. Water based protein solution conduct convective and
conductive heat transfer process, also ethanol between cuvette and HPOC. HPOC is
cooled by a water bath in the bottom, with air circulation on the top and side. Since this is
63
Figure 2-32 Heat Transfer analysis result (whole layout)
The objective of this simulation is to study the heat transfer process and to find an
optimum heat source input to achieve the best performance. Overheating will destroy
protein, and under heating won’t make the spring function. First different currents from
1A to 5A is conducted to find out how fast the temperature of SMA wire can reach from
room temperature(22°C) to body temperature (37.5°C). 1A and 2 A was used for SMA
wire in the air usually (Figure 2-33).shows that using 1A to reach 40°C will take far more
than 5 seconds in water solution. Figure 2-34 shows that using 2A to reach 37.5°C takes a
64
Figure 2-33 1A current, 5 seconds heating response
65
3A electrical current is the turning point, Figure 2-38 shows that the heating time
is shorten to 2 second. This is more useful, but not good enough for our application in
that the heating response is not fast enough. 4A electrical current heats up the SMA wire
much faster, refer to Figure 2-36, it takes 0.5 second to get to 37.5°C. 4A electrical
current will be useful to keep the cycle under 1second. 5A electrical current response is
even more promising, refer to Figure 2-37, it takes around 0.25 second to reach body
temperature. We can keep these two current selections in mind for later application.
66
Figure 2-36 4A current, 5 seconds heating response
67
We feel confident that 5A electrical current for heating can produce promising
result. A closer look at the model, Figure 2-38 shows the result after 5 seconds by heating
up with 5A current. Figure 2-39 shows the temperature line plot, which the cross line
resides on the red line as show on Figure 2-38. After 10 seconds heating with 5A
electrical current, the center of the wire reaches 80°C, but its surrounding area
temperature drops dramatically as the distance is a few times of wire diameter. The
heating effect is considerably local. On the other hand, 5 seconds heating up constantly
with 5A current is undesirable, since the center reaches over 40°C which is unfavorable
to protein solution.
Figure 2-38 one coil cross-section of SMA wire close up heating condition
68
Figure 2-39 Temperature line plot of crossing SMA center 5A, 5seconds heating
Now, let’s examine the cooling process. The SMA wire is heated up with 5A
electrical current for 0.25 second, and let the wire to cool down. Figure 2-40 shows that
the temperature of SMA wire reaches to 39°C in 0.25 second, the temperature drops
down to 25°C in 0.75 second. It looks that a whole cycle of heating and cooling can be in
exactly 1 second, which means that 1Hz of vibration frequency can be achieved.
69
Figure 2-40 0.25 second, 5A heating and cooling process
Now, let’s examine 5A heat transfer analysis result. Another heat transfer model
is to simulate a periodic supply of 5A electrical current for 0.25 second each cycle for 10
second running time. The structure was kept as the previous one, so does boundary
conditions. The change is only the input. Figure 2-41 shows that cooling from 40°C to
25°C needs more than 0.75 seconds, while the heating up needs 0.25 seconds. The total
time comes to 1 second. The cooling time is the limitation to increase agitation frequency.
Referring (Figure 2-1) and (Figure 2-1), the direct current heating and cooling behavior
shows that there is a delay period which SMA absorbs heat to prepare phase
transformation. But the period is not either defined or provided with experimental data
for liquid cooling system. We give 15seconds more for both heating and cooling period.
70
Figure 2-41 shows 1Hz heat transfer plot of the wire center temperature. From the second
cycle, the low temperature is higher than 25°C (Martensite Temperature of our Nitinol
Body Temperature wire), which means, that it would not fully return its martensite phase,
In Figure 2-42, frequency is reduced to 0.5Hz, the wire is fully cooled down
under Martensite transformation temperature 25°C in the first cycle. After 5 cycles, the
peak temperature is only a little higher than 45°C, comparing 1Hz with reaching almost
50°C instead. The ideal frequency is 0.5Hz instead of 1Hz due to the heat effect.
71
Figure 2-42 5A 0.25 second heating at 0.5Hz condition
Figure 2-43 shows the temperature line plot of temperature change of 5 cycle of
5A/0.5Hz/0.25s heating condition. The line plot crosses the cross-section 1 of SMA
spring wire from the center to the ambient as show in Figure 2-44.The ambient
temperature in solution is less influenced comparing to that in Figure 2-39. Figure 2-45
SMA wires parallel to the previous line as show in Figure 2-44. The center of the cuvette
is less influenced, and the high of temperature only reaches 22.7°C, less than 1°C rise.
72
Figure 2-43 line plot at 5A 0.25 second heating 0.5Hz condition
Figure 2-44 Heat transfer analysis result for two coils of SMA wire close up
73
Figure 2-45 Line plot between 2 coils of 5A, 0.25 seconds heating, 0.5 Hz
Figure 2-46 Temperature point plot at one wire diameter away from SMA coil (5A,
74
Figure 2-46 shows close area temperature change. The point selected in within
distance of the diameter of the wire. The change of temperature follows the wire center,
but the amplitude is much smaller. Figure 2-47 show the temperature change out of 2
wire diameter distance, the heating process has almost no influence at all. Figure 2-48
shows the center line temperature ramp. The center is heated up by 5°C after 5 cycles
heating.
Figure 2-47 Temperature point plot at 10 times of wire diameter away from SMA
There is some fine tuning of peak portion to keep the heating to minimum. By
reducing the peak time to 1/3 second, FEMLab won’t cooperate well, it refuse to produce
cycles after 3 rounds Figure 2-49. And the peak temperature doesn’t reduce much. Finer
mesh and smaller time interval was tried, but didn’t produce better result. So we are
75
Figure 2-48 Temperature point plot at the center of coil (5A, 0.25s heating, 0.5 Hz)
Figure 2-49 Temperature point plot (5A 0.33s heating 0.5Hz center)
76
2.3.4.3 Mixing Test using Fluorescence Measurement
The heat transfer analysis shows that the best strategy is to use 5A current with
0.25 second heating and 0.5 Hz to create the biggest vibration. The design assembly is
ready for mixing test. Since the HPOC will be fabricate after the finalization of the
cuvette, the mixing test is conducted without HPOC. The set up can be referred in Figure
2-50. Carboxy-fluoresceine (CF) of 300 nM was filled in lower chamber of the cuvette
and the upper chamber was filled with pure water. The exciting light wavelength is 495
nm and the emission wavelength is 520 nm. The excitation light is hit in the middle of the
lower chamber, so is the observation lens is aligned in the middle of the lower chamber.
77
The mixing test result can be referred as Figure 2-51. Due to the low frequency of
the vibration created by SMA tensile spring and compressive biasing spring, the mixing
result wasn’t very satisfactory. According to the volume ratio of lower chamber and
upper chamber of the cuvette, the drop of the intensity after mixing didn’t reach target
point.
To further confirm our prediction, the solution was left over night and continued
to test after 10 hours. Even after 10 hours, the intensity level didn’t drop much. After
78
manual mixing, the intensity of observation dropped to predict level (Figure 2-52). This
2.3.4.4 Conclusion
The design can provide unplugging and stable vibration range. FEM heat transfer
analysis shows that heat generation is local and controllable. Mixing test was not
satisfactory due to the slow vibration frequency. The cooling time is the bottle neck in
increasing frequency.
The final design introduced a micro-motor for better mixing result. Since the
electrical lead plug is very costly, the motor shares electrical input with SMA actuator
79
Figure 2-53. A resistor is added to control the distribution of current between SMA
specifications. The biggest concern of micro-motor is if it can work under high pressure.
Different shape and size micro motor was selected and constructed and a vibration motor
commonly used for pager and cell phone was selected. This micro-motor is compact with
4 mm in diameter and 10 mm in length. Since it is widely used in cell phone industry, the
price is incredibly affordable since mass production made the cost to the minimum. A
The mixing test used the same set up as previous design. With ISS Fluorescence
80
selected according to CF property. The mixing result of this setup can be referred in
Figure 2-54 and Figure 2-55. Figure 2-54 shows the stabilization before and after the
mixing start point and Figure 2-55 shows the close-up view around the mixing starting
point. The result was very promising in that the intensity drop is close to the volume ratio
81
Figure 2-55 Close-up view at mixing start point
2.3.5.3 Conclusion
This design combined tensile SMA spring and micromotor and made both
unplugging and mixing possible. Since SMA spring is only heated for 0.25 second, the
heat generation was controlled to the minimum. Micromotor functions both as a mixer
and stopper, saves space, and keeps design simple. The mixing result was very promising.
Intensity drop shows that the two solutions can be completely mixed in 10 seconds from
totally separation.
82
3 Cuvette Design and Mixing Process
The Cuvette is designed for the mixing of two solutions, with its necking to fit a
plug that separatse the two solutions, the cuvette is illustrated as Figure 3-1. The design
concept is to minimize the size but to meet the measurement requirement. The biggest
size of a fluorescence spot is about 3mm in diameter and varies by different experimental
setup. Since some of the assays are very expensive, the minimum size is highly favored.
The current cuvette used by Chong for his ISS HPOC measurement is about 1ml in
volume and the size of the spot light is well centered, so we are going to keep the bottom
part the same volume as the current design. The top part of the cuvette will be decided by
.60
1.00
Heating Cooling
Heating
83
3.1.1.1 FEM Simulation of Free Diffusion vs. Neck Opening
How long it will take the two solutions to diffuse without any agitation? A
Femlab simulation predicts that it takes almost 8000 seconds or more than 2 hours for an
8cm in height and 2cm in diameter cuvette with a three fourth reduction in necking
(Figure 3-2 and Figure 3-3). The diffusion model used water’s diffusion coefficient since
the water is the common base for many solution mix. Water’s diffusion coefficient is
around 1E-8. Material’s diffusion coefficient is proportional to its molecule size. The
bigger the diameter, the smaller the diffusion coefficient and the onger it takes to diffuse.
According to Philip Robert, it takes 30days for salt to be fully mixed in the bottom of
Figure 3-2 A cuvette of 8cm height and 2cm diameter with 3 quarter necking
84
Figure 3-3 Diffusion time of liquid
The necking design is based in diffusion analysis. Let’s examine three condition
of necking from one eighth of reduction in diameter to half reduction in diameter (Figure
3-2). Figure 3-5 shows that the different necking will influence the diffusion time. The
point plots are set in the middle of the lower chamber of the cuvette, the same spot where
the observation light will go through. It takes about a hour for half opening and three
quarter opening with different diffusion speed, but it takes almost 2 hour for a quarter
opening. The necking influence is one of the main considerations on cuvette design.
85
Figure 3-4 Necking examination
0.8
Concentration
0.7
0.6
0.5
0.4
0 1000 2000 3000 4000 5000 6000 7000 8000 9000 10000
Second
86
To further examine the diffusion along the center line, we can find out that the
close necking area can reach the equilibrium fast and moving along the center line, the
diffusion time getting closer and closer (Figure 3-6). The diffusion curve slows down as
it moves to the bottom of the cuvette, but reaches equilibrium about the same time. The
plot is based on cuvette 8cm in height and 2cm in diameter with 3 quarter necking or one
quarter opening.
0.95
0.9
0.85
0.8
Concentration
0.75
0.7
0.65
0.6
0.55
0.5
0 1000 2000 3000 4000 5000 6000 7000 8000 9000 10000
Second
Figure 3-6 Point plots along the center line from necking to bottom of the cuvette
From the simulation above, the bigger the neck opening, the faster the mixing
process takes. Many variations of necking were designed, fabricated and tested. The final
design eliminates necking with “step down” necking design. This necking design will
87
maximize the diffusion process. The lower chamber of the cuvette is kept for 1 ml for
observation requirement while the upper chamber is designed according to pulling micro-
To understand the mixing process, we need to start from diffusion. When you add
cream or sugar to your coffee, you need a stirrer, a plastic or a wooden stick from
Starbucks or a tiny golden or silver spoon from your granny’s. Without a stick or a spoon,
it takes forever to taste the sweetness or to get rid of the bitterness in coffee. A simple
stir will accelerate the diffusion process dramatically. But how does this happen? What
88
make the forever waiting to a few seconds? It seems so simple that nobody will question
If you work in a chemical or bio lab, many different kinds of stirrer might be
familiar. But to mix some solution under high pressure might only be seen in a science
fiction movie or James Bonds’ adventure. I remember that a big piece of metal block was
enclosed in the glass bell. It is rotating with a suspense background soundtrack. It will
explode at some point and a hero is going to save the scene, but you never know how.
The difficulty of mixing under pressure is space limitation. High pressure will
also increase the viscosity leading to an increased in back force. The choice of proper
methodology is the key component with respect to different application. The main mixing
Mixing methods under pressure are mechanical agitators, rotating and rocking
One of the oldest and least-preferred methods of mixing is to agitate with motors located
outside of apparatus. By introducing a stirrer into high pressure apparatus with a shaft
passing through a gland, the motor drives the stirrer. Although improvements have been
made in the design of glands and friction methods, they now can be reduced without the
breaking the hermetic seal, its use allows friction to be reduced only at low surface
rubbing speeds. (Flom and Porill 1955) Another mechanical agitator is stirrers with
motors located inside apparatus. When the internal dimensions of the apparatus is large
enough, the stirrer and motor can be placed inside the apparatus, introducing the motor
power current through electrical leads. (Kiebler 1945) This requires that the motor
winding can withstand the action of the medium inside the apparatus. A deficiency of this
89
design is the possibility of contamination of the medium being mixed by the lubricant of
the motor, as well as the necessity of cooling the motor, which can be easily overheat
(Figure 3-8).
Figure 3-8 Location of stirrer with motor within apparatus: 1) motor winding 2)
Rotating and rocking apparatus is suitable for apparatus not too cumbersome and
if the general location of the apparatus does not present a problem, mixing maybe
Electromagnetic Stirrers are widely used for agitation of the contents of vessels
operating at pressure of a few thousand atmospheres. This design has generally a few
stirrer power supply of powerful solenoid has interrupter, magnetic starter, transformer,
autotransformer and solenoid (Figure 3-9). A drop stirrer is attached to the head 2 of the
90
apparatus, made of nonmagnetic steel (Figure 3-10). When the current is passed through
the coil, the core returns symmetrically relative to the coil, raising the stirrer. When the
current is turned off, the stirrer drops by its own weight to the bottom of the apparatus.
Thus, agitation is achieved by turning the current on and off. This type of agitation
requires a considerable amount of time, but gives good results, especially if the amplitude
of the stirrer oscillations is rather great. The intensity of the mixing action depends also
on the frequency with which the stirrer is dropped and the form of its blades. A rotating
electromagnetic stirrer is another option of this kind of stirrer. The reciprocating motion
of a stirrer can be combined with rotation. As show in Figure 3-11, core 1 is joined to a
stirrer shaft 2 with finger 3 which moves in a spiral cut in tube 4, fastened firmly into the
apparatus. As the core is draw upward, the motion of the axle will include rotation, due to
finger 3.
91
Figure 3-10 Drop stirrer: 1) coil 2) head 3) core 4) shaft 5) body of apparatus
Figure 3-11 Stirrer with reversing rotation:1) core 2) axle 3) finger 4) tube
92
In the above mixing processes, there are two motions involved: up and down, and
rotation. When drinking a cup of tea made from a tea bag, one will pull the string of the
tea bag up and down and the tea can be dissolved immediately. For coffee with sugar and
mathematical elegance, it is an idealized subject and often not very useful in handling
phenomna. Since fluid elements carrying different physical quantities move around
randomly under turbulence, we expect turbulence to enhance the transport processes. The
theory of turbulent diffusion was first formulated by Taylor (1921). Even if you have
never heard of Taylor you make use of the motion of turbulent diffusion when you stir
coffee to mix sugar in it. If one puts sugar in the coffee and does not stir it, then the sugar
mixes with the coffee only by molecular diffusion, which takes a very long time. On the
other hand, if the coffee is stirred, then generates turbulence inside it, turbulent diffusion
mixes up the sugar much more efficiently than molecular diffusion. (Graham).
Turbulence is the main factor to accelerate the mixing process. Despite many
years of intensive research into turbulent diffusion, however, our ability to predict means
contaminant distributions is often quite crude and our attempt to predict statistical
efficient in rapidly decreasing the concentrations of contaminants that are released into
93
diffusion, particularly engineering models, predict only time-averaged concentrations.
kinetic energy. Since the nature of turbulence is unpredictable, statistical measures is the
best method to describe the motion. The velocity is described as a time-averaged value
(3-1)
Where T is a time much longer than the longest turbulent fluctuations in the flow.
A time record such as shown in Figure 3-12 is called statistically stationary if the mean
94
3.4 Mechanism of mixing in turbulent flows
Turbulent flows contain irregular motions over a wide range of length and time
scales. How do these motions contribute to mixing? Mass transport due to molecular
diffusion can be described by Fick’s Law. In a flowing fluid, the flow itself contributes to
the transport mechanism. The mixing process is therefore due to distortion, stretching,
and convolution of the original patch whereby the original volume is distributed
irregularly over a larger volume, so that the concentration, averaged over some finite
volume, decreases. In the absence of molecular diffusion, however, such a process would
not reduce actual peak concentrations at a point; the reduction of these peaks is therefore
very dependent on molecular diffusion. G. I. Taylor (Taylor 1921) published one of the
most important results in turbulent diffusion theory, which provides a link between the
eddy diffusion coefficients and turbulent flow properties, in 1921. Since turbulent
velocity fluctuations are irregular, the results of each trial differ. On average, however,
the particles wander apart from each other, and the rate at which they wander apart can be
particles, released from the coordinate origin at different times. Many scientists and
engineers tried to understand how turbulent flow accelerates mixing processes (McComb
The expressions for the molecular transport coefficient from kinetic theory have
different values. For example, compare expressions for thermal conductivity and
viscosity. The coefficients for different turbulent transport processes, however, are
believed to be the same. The cause for this contrast between molecular and turbulent
transport is that molecular motions are related to the quantities they transport. Molecules
95
carrying more thermal energy move faster. This is not true for a turbulent fluid. The
temperature of a blob of fluid does not have a direct functional relation with the turbulent
velocity of the blob. Hence all the diffusion coefficients for the passive transport of any
quantity in a turbulent fluid have the approximate value as following equation, the
1
DT = v 2τ cot (3-2)
3
This equation best explains micromotor mixes much better than SMA spring
vibrator, since micromotor can produce a velocity change from 0 rpm to 2500 rpm in
seconds but the SMA spring vibrator can only produce 0.5 Hz to 1 Hz due to cooling
limitation. The equation also shows that velocity fluctuation contributes to accelerate
mixing process, once the flow is fully developed, the velocity won’t help anymore.
96
4 High Pressure Optical Cell Design
requirements. The bomb can withstand pressures up to 2 kbars, it has four optical ports
and one hydrostatic port to permit its pressurization (Paladini 1980). The cell allow for
easy loading and unloading samples and prevent them from making contact with either
the inner metallic bomb walls or the pressure transmitting fluid. The inner chamber was
kept as small as possible to increase the efficiency of the pressure-delivery pump which
97
4.2 ISS’s current HPOC
ISS continued Weber and Paladini’s design concept and improved finishes. The
ISS HP Cell is designed for high pressure absorption and fluorescence experiments, as
well as for Raman spectroscopy studies. The cell can be pressurized up to 3,000 bar
(43,500 psi) when using quartz windows and up to 40,000 bar (58,000 psi) when using
sapphire windows. Three 10mm diameter windows are placed in the optical plane; the
10mm clear aperture allows spectroscopy studies using either a laser or a lamp as light
excitation sources.
The cell is made of a stainless steel alloy, which has excellent thermal
temperature of the cell can be monitored by a thermocouple inserted in the side of the cell.
Ethanol is the medium to pass hydraulic pressure from the pump to the cell. The ethanol
98
is required with low fluorescence background. A hand pump generates hydrostatic
pressure. The finish of ISS HP Cell is more polished than the Paladini and Weber’s
prototype. The cell shape is rather cubic than cylinder shape, this makes alignment of
Our design is based on ISS HPOC. Since the SMA spring is heated up by
applying electrical current, an electrical plug assembly is intruded to the current HPOC.
The two electrical plugs increase the height of the HPOC and require a new pressure
analysis. The high pressure optical cell is made from the cell body, observation windows,
electrical lead plug and pressure inlet plug. The cuvette assembly containing the SMA-
based actuator will be placed inside the inner channel of the cell body. The cell body is
made of Vascomax C-300 which has favorable strength and toughness characteristics and
is commonly accepted as a high pressure cell material. Although the ideal shape of a
pressure cell is sphere (Bobrowsky 1968), a cylindrical shape provides the best
Stresses are present in a material under load, causing deformation. The design of
the apparatus operating at low pressures entails primarily determining the wall thickness
sufficient for the apparatus to withstand the load to be applied without permanent
utilized in construction. When the stresses in the metal exceed the yield strength, the
99
metal goes over into the plastic state. The internal layer of the wall of the cylinder, where
the pressure exceeds the yield strength, will be in the plastic state, and the equivalent
stresses in it will have a certain value, uniform throughout the entire plastic layer. With a
further increase in pressure, the stress in this layer will be not increased, but the thickness
of the plastic layer will increase. This agrees with the concepts of the nature of plastic
deformation presented above. The high-pressure cylinder is thus, as it were, divided into
two concentric layers; the internal (plastic) layer, and the external (elastic) layer which
protects the plastic layer increases, and the thickness of the elastic layer consequently
decreases until rupture finally occurs. Experiments have shown that rupture of a high-
To select the right electrical leads is important to the HPOC design. First, the
electrical lead plug must hold 3 kbar of pressure; second, the cost is another issue to
1000atm using various insulating materials. Sealing wax or other cement, melted onto the
rod to be insulated can be used for this purpose. Good results can be obtained using fibra
The insulated lead is compressed by the sealing material, and friction prevents the lead
from being expelled. If the conductor’s diameter is greater than 0.5mm (such as when
100
high current must be passed) and the pressure in the apparatus is over 1000atm, this
method is unsuitable.
A more reliable seal for an electrical lead is the inverted cone. A metal cone with
body. The pressure acts on the large base of the cone and forces it into the seat. The
insulator used may be any one of numerous materials such as ivory, plastic, pyrophyllite,
mica, etc. the selection of the material to be used depends on the pressure and
the cone is a simple and reliable method allowing the use of an electrical lead at
temperatures up to 800-900˚C and has been successfully used for many years at pressures
of 18,000-20,000atm.
An Electrical Plug Assembly (Figure 4-3) is used to provide the SMA wire with
electricity power under high pressure. A heat shrinkable plastic tubing (Terry 1972) has
been proved to be effective in both gas and liquid systems, and it is capable of
withstanding pressure up to 10kbar (or 1000Mpa). A Kynar tubing is shrunk over a 16°
cone using a heat gun. This assembly is inserted in a cone drilled in the pressure cell,
with the large side of the cone facing the pressure side. This electrical lead design will be
101
Figure 4-3 Exploded assembly view and assembly electrical lead
suggested by Ya. A. Kalashnikov (Kalashnikov and Zh. 1954). Body, made of 10Kh steel
contains two holes whose length depends on the diameter of the wire to be brought out of
the apparatus and pressure to be contained. In particular, for a pressure of 5000atm and a
conductor diameter of 0.5mm, the channel length should be at least 18mm (Figure 4-4).
102
Figure 4-4 Kalashnikov electrical lead: 1) body 2) holes 3) conductors 4) thread
After passing wires 3 through the holes, body 1 is filled with a styrene monomer,
then it is poly-merized in a special bomb for two days at 1500atm and 80°C. As a result
of this, the bond between the leads, polystyrene, and body wall becomes so strong that
the pressure cannot force the conductor out of the plastic or the plastic out of the body.
This force out is hindered also by threaded diameter 4 in which the styrene is
polymerized. The excess polymerized styrene is cut out on a lathe. After this, the
electrical lead is ready to use. This method’s advantage is multi-electrical lead at the same
of a somewhat different design, also made in the form of an inverted cone (Vereshchagin,
Kuznetsov et al. 1956). The Inverted cone is the most convenient and popular used
electrical lead method. In our design, we used the inverted cone method.
103
4.3.3 Observation Window Design
The lower window plug assemblies are used for collecting the spectroscopy data.
The port assembly is designed to hold a fused silica or quartz window through which the
surface. A threaded brass cap that provides only light mechanical support for the window
is installed at the end of the assembly. The cap has a 6.4 mm diameter hole to allow the
transmitting fluid to enter and press upon the exposed window surface. The window itself
is made of Corning 7940 fused silica that can withstand the required pressures as well as
pass the UV-visible wavelengths. The window arrangement used in this assembly is
support is based on the Bridgeman (Bridgeman 1931) unsupported area principle. In brief,
sealing is accomplished because the viewing aperture creates a surface area imbalance
between the fluid side and the support side of the window, resulting in a pressure
differential, which ensures a strong seal. This sealing method, however, is heavily
dependent upon the accurate machining of an optically flat surface on both the window
and window support and is not reliable for very small apertures. The methods for
checking the surface quality of windows and supports have been reported by Sherman
104
Figure 4-5 Exploded view and assembly a window plug
Windows for visual observation at high pressures are made of glass, transparent
quartz, sapphire, diamond, and even cooking salt. Glass windows are extremely sensitive
to the action of water. It is known that water is dissolved into the surface layer of glass,
increased and water penetrates even more deeply into the glass. The simplest viewing
window design consists of a cone made of glass or transparent quartz, lapped to the metal
seal. Sometimes, packing is placed between the glass and metal. The large end of the
cone is place toward the inside of the apparatus, so that the pressure forces the cone into
the seat. The reliability of this type of construction depends on the care with which the
lapping is done and the quality of the glass or quartz cone. Even very well-made viewing
windows of the type are burst due to chipping of the edge of the cone. These windows are
with optically parallel ends is pressed against the polished surface of the nipple. The cap
will retain the glasses. The nipple is screwed into the body of the apparatus and packed
using packing. The pressure created in the apparatus acts on the area. The central hole is
105
drilled into the stock for the nipple and is enlarged with a reamer. The part is place
between the centers of a lathe and turned, after which, while mounted in the chuck, the
thread and final turning are performed in one pass. The surfaces of the parts which are to
be polished are hardened. The end of the nipple is heated until dark red, then tempered
for 2-3 minutes in warm (30-35˚C) water and cooled in oil. After polishing, the viewing
window with its optically flat surface held against the dry surface of metal should show 2
to 3 Newton rings.
The inlet plug seal is achieved by using a combination of three sealing rings. One
lead ring, which is sandwiched between two brass rings, deforms to fill any voids created
by wear or poor surface quality. The type of sealing method has been used successfully in
other hydrostatic pressure cells. The pressure inlet plug uses a similar external
configuration and three-ring sealing method as used on the four window assemblies in the
lower chamber.
high-pressure apparatus.
106
4.3.6 Machine Finishes
finishing of areas directly in contact with the medium creating the pressure, as well as
finishing of threads, packing surfaces, etc., must be especially accurate. Burrs and
grooves cannot be tolerated. Sharp angles should be smoothed. Special requirements are
placed on threads. Threads in high-pressure apparatus are usually made short, i.e. with
smoothes sharp angles, with prevents “seizing”. Special attention must be turned to the
centering of threads in holes that are sealed by rods, pins, or shafts. Methods of
FEM software ANSYS is used for cell body assembly stress and deformation
analysis. Since the cell body assembly is symmetric we only take half for analysis to
save computer time. The pressure is 3000 bar in the chamber and 1 bar outside. As
showed in Figure 4-6, the maximum stress is 8893psi, which converted to metric is
613bars around the window assembly. The yield strength of the Vascomax C-300 is 1915
MPa, which is three times larger than that of the cell’s maximum stress.
107
Figure 4-6 Stress analysis for cell assembly by Ansys
To verify ANSYS simulation results, FEM software FEMLAB is used for stress and
deformation analysis of cell body assembly. Since the cell body assembly is axial
symmetric we only take half for analysis to save computer time. Pressure is 3000 bars in
the chamber and 1 bar outside. As showed in Figure 4-7, the maximum stress (824.5 MPa)
and deformation (6.14E-5m) are in the radial direction around the window assembly. The
yield strength of the Vascomax C-300 is 1915 MPa and the safety factor is around 2.35.
108
Figure 4-7 First principle stress of cell assembly by Femlab
Difference FEA method showed different results, this bring up the uncertainty of
FEA methods. A thoroughly FEA study can be performed using TRUEGrid and ANSYS
together. Both the safety factors weren’t very high, around 2 to 3. But the highest stress is
below its yield strength which means no plastic deformation will occur. Refer to previous
discussion of high pressure cell design, for very high pressure condition, plastic
deformation will occur in inner layer, the stress in the layer won’t increase according to the
pressure increase, which will lead to plastic deformation layer increase. Our design is under
that stage, a plastic deformation won’t occur for pressure up to 3500bars. The cell thickness
is determined from numerous previous designs. The height increase didn’t introduce higher
stress concentration.
109
4.5 The Final HPOC Design
The final HPOC design involves height change to existing design, and also
introduces two new openings to accommodate Electrical Lead Plug design (Figure 4-8).
The height increase is a potential risk of increasing hoop deformation. And new plugs
110
4.6 Our New HPOC Fabrication collaborated with ISS
manufacture commercial HPOC based on Weber and Paladini’s Original High Pressure
Bomb design. We would like to collaborate with them to have technology transfer our
new designed HPOC to become a product. After receiving our new HPOC design, ISS
proposed some modifications to reduce cost and fit to existing set-up. Basically, the
modified design kept all features of their current HPOC body, but modified the inlet plug
by introducing electrical leads into the chamber. The design eliminated two electrical
lead plugs which are the most costly. Besides the inductions of electrical lead, the top
plug still serves as an inlet plug. With the electrical leads taking the top side of the plug,
and the opening for pressure inlet is set on the side (Figure 4-9 c). To accommodate the
increase of the height of new cuvette, the plug was hollowed in the bottom side. Two
pieces of 3/16 in diameter rod was cut to serve as electrical leads. Heat shrink tubing are
approximately the same length as the brass electrical leads. Using a standard heat gun, the
tubes are shrunk onto the Brass Electrical leads. Then the leads are inserted into the
threaded electrical lead holders as shown in Figure 4-9 a) and the heat shrunk tubes are
trimmed to expose brass electrical lead rods. Using the supplied wrench, we inserted the
leads and holders into the Top Plug as shown, finger tight.
111
Figure 4-9 Electrical lead and inlet plug combo
The new ISS design also modified the fixture of the cuvette. The our design and
ISS current HPOC, a fixture was used to hold cuvette in position at the bottom of the
HPOC inner chamber. The fixture positions the cuvette and aligns the observation path
through the same spot of the cuvette on each measurement. Since the electrical lead is on
the top, which will turn as the plug is screwed to tighten up to seal the HPOC, electrical
lead will be tangled for 9 turns if the cuvette stays in current fixture. ISS designed a
fixture which connected to the top plug (Figure 4-10 (a), (b)). The fixture is also served
as cuvette cap (Figure 4-10 (c)). There are two plugs that serve to seal the cuvette: red
silicone plug seals the cuvette fixture cap and pink vinyl plug seals the bottom cuvette.
112
Figure 4-10 Cuvette and its cap
The ISS design has many advantages: budget saving, no tangle of wires, compact,
etc. But it also has some shortcomings compare to our own design. After a few
experiments, the O-rings used to guide both the cuvette and cuvette cap become frictional
which makes mounting and dismounting cuvette to cuvette cap and cuvette cap to inlet
and electrical lead plug combo very difficult. The friction increases because the O-rings
are soaked in ethanol which is the media of transferring the pressure into the cell. The
mounting of cuvette from the top causes problem on the positioning of the cuvette to the
light path. The tilt of the cuvette causes the initial measurement of value to be different
from each measurement. This can be improved upon by some modification on the Cell.
113
114
5 Complete System Assembly and Final Testing
The final complete HPOC assembly is shown as Figure 5-1. The new design
HPOC has 3 observation windows, the third window didn’t show in the figure is
perpendicular to the other two windows and at the same level. The excitation light will
pass through the third window and hit on the center of the lower part of the cuvette
chamber. As the assay is excited by certain wavelength light, the emission light is
collected from the right side window and the left side window is used for comparison
purpose mostly. The top plug has two functions: the yellow lines indicate electrical leads
which pass the current to both SMA spring and micro-motor; an opening on the side of
plug passes hydraulic pressure into the chamber of HPOC. Inside of HPOC chamber, a
dual chamber cuvette resides in the cuvette holder connected it to the top plug. Through
the cuvette holder, electrical lead is connected to SMA spring and micro-motor via a
socket. The micro-motor has a Teflon O-ring which functions as a stopper to separate two
different solutions before the pressure reaches desired level. A flexible polyethylene tube
made fan is attached to the shaft of the micro-motor which starts turning once the plug is
opened by SMA spring. A red soft plug seals the cuvette cap from the top to pass
pressure from inner HPOC chamber into the cuvette, so does the pink soft plug in the
115
Figure 5-1 Final Complete HPOC Assembly
Figure 5-2). Also from the figure shows the hydraulic pressure system. High grade
ethanol is required for eliminating any unwanted fluorescence interference from the
media. Before adding pressure to the HPOC, the manual pressure pump needs to fill with
ethanol. By cutting off the valve to the HPOC, opening the incoming valve and turning
back the screw of manual pump, the whole pressure control system is filled with ethanol.
After closing the incoming valve and opening the valve to the HPOC, turn the manual
screw forward and let ethanol start to flow out from the tip of the tube and this shows the
pressure system is filled with ethanol. Then connect the tip to top plug of the HPOC, the
pumping process can be started. While turning the wheel of manual pressure pump, the
pressure can be monitored from pressure gauge. During the pressurization process,
leakage from all valves and connections need to be checked since the leakage can
116
produce pressure drop which can be monitored from pressure gauge. Once the pressure is
Figure 5-3. ISS K2 is widely used for steady-state fluorescence measurement. Its optical
design and automatic instrument control are “state-of the-art” and time-resolved
package make the instrument easy to use. The measurement has a light source, light path
through which desired wavelength selection and modulation are made, enclosed HPOC
117
Figure 5-3 ISS K2 Fluorescence Measurement
Figure 5-4 shows its optical diagram. Fluorescent light is generated from a Lamp
CF will emit 520nm wavelength of fluorescent light. The receptor will measure the
intensity of the light, around 520nm wavelength. The wavelength of emission is always
bigger than that of excitation. Without any CF in the cuvette, the intensity is around 1000,
but when 300nM CF solution is measured, the intensity is around 300,000. The
measurement number is relative, it can be adjusted. The upper limit is above 1,000,000.
To receive the most sensitivity of comparison, the higher the intensity, the more desired.
But high intensity can cause photo bleaching which cause the intensity to drop gradually.
Our measurement is set around 300,000 which will be a good resolution and minimum
photo bleaching.
118
Figure 5-4 Optical diagram of ISS K2
The plug design has two considerations: to seal the two solutions well before
activation; and keep the size compact. Since the diffusion process takes a long time for
free diffusion, once the two solutions are separated, the leakage won’t influence much.
To keep the plug compact reduce the travel distance of the plug which will reduce the
upper part of the cuvette as well. The leakage testing is set up with ISS fluorescence
and filled at the bottom part of the cuvette and water is filled in the top part. The
measurement result is show as Figure 5-5. Raw data is normalized and fitted with
polynomial method. The measurement put water solution in the lower cuvette chamber
119
and CF solution in the upper chamber. There is more fluctuation on the measurement
since the value is very small, around 1000 intensity. But if leakage happens, the CF will
keep the measurement running for a few hours, no leakage found either.
Leakage Testing
0.6
0.4
0.2
0
0 200 400 600 800 1000 1200 1400 1600 1800
Time (s)
The diffusion test was set up with water in the bottom of the cuvette and 300nM
CF on the top. A plug is sealed in the middle. After stabilize, the plug is pushed down,
the diffusion process starts. Figure 5-6 shows the diffusion results. The results show that
it takes more than 3 hours to reach the equilibrium point. Since CF molecule is much
larger than water molecule, its diffusion coefficient is much smaller since diffusion
120
coefficient is reverse to its size. The simulation of water diffusion shows that it takes
around 2 hours to reach equilibrium. The CF diffusion takes much more time than water.
In the experiment, the push down of the plug gave a turbulence push and the process is
chapter.
5-1
VO2 denotes molecule volume of O2. From the equation, we can see that diffusion
Diffusion Testing
800
700
600
500
Intensity
200
100
0
0 2000 4000 6000 8000 10000 12000 14000
Time(s)
121
.
5
x 10 Mixing with SMA and biasing spring actuator agitation
5
data 1
4.5 6th degree
3.5
3
Intensity
2.5
1.5
0.5
0
0 1000 2000 3000 4000 5000 6000 7000
time(s)
Figure 5-7 Mixing with SMA and biasing spring actuator agitation
experiment designed to start the measurement from high intensity to low intensity. By
putting 300 nM CF in lower chamber of the cuvette, the starting intensity is set around
300,000. With known volume ration of CF solution in lower chamber to water in higher
chamber, the final intensity of well mixed solution can be predicted. To test the dilution
and intensity relation, we did some dilution test. By doubling the volume with water each
time, the intensity from 300 nM and stopped at 37.5 nM is measurement. The result
almost follows the linear trend (Figure 5-8). Using this result, with known volume ratio,
122
Figure 5-8 Intensity vs nM
First, the mixing fluorescent measurement under 1 atm pressure is recorded and
the result is shown in Figure 5-9, Figure 5-10, and Figure 5-11. The target was set to 10
second with 90% drop in intensity. All 3 cases reach the target. The mixing ratios
123
Figure 5-9 Mixing result under 1atm first measurement
124
Figure 5-11 Mixing result under 1atm third measurement
When the pressure is raised to 300bars, about 300 times atmosphere, the mixing
result is as Figure 5-12, Figure 5-13 and Figure 5-14. The mixing trends varied a little
from each other, 90% drop all achieved in three cases, but all reached the equilibrium
feature. The diffusion coefficient is related to the flow velocity, but only can be described
unpredictable. This experiment proved that SMA spring with micro-motor actuator can
perform unplugging and mixing work under pressure. More experiments are undergoing
for higher pressure such as 1000bar or 3000bar. Higher pressure experiments encounter
some unstable problem, the cause and solution are under investigation.
125
Figure 5-12 Mixing result under 300bars, first measurement
126
Figure 5-14 Mixing result under 300bars, third measurement
5.4 Conclusion
The complete system was assembled, tested and ready for bio-physical study. The
new designed HPOC can hold up to 3 kbar with easy assembling, loading and unloading.
providing electrical supply. A dual chamber cuvette can separate two solutions before the
pressure was added up to desired level. SMA spring and micromotor achieved
127
6 Bio-Safety test of Heat Generation of SMA
The compatibility of the Ti-Ni shape memory alloy (SMA) with biological
molecules such as phospholipids and proteins have been tested. After the SMA Flexinol
the concentration of phospholipids remained in the bulk solution was determined using
the Bartlett method [15]. To our surprise, there was a small increase in phospholipids
concentration with increasing the surface area of Flexinol wires (Table 2). At present, the
origin of the increase in [POPC] is unclear. It is unlikely due to wire contamination. The
wires were thoroughly rinsed with light detergent and deionized water prior to use, and
that we can make corrections for this unexpected increase in phospholipid concentration
Table 2 The change of phospholipid concentration with SMA wire surface area
There are no significant changes in the concentration of the protein samples after
the addition of the shape memory alloy Flexinol wires. Three proteins, namely, human
128
albumin, bovine serum albumin, and bovine gamma-globulin, were tested. The protein
concentration was determined using the Lowry (Lowry, Rosbrough et al. 1951)method
generates heat. The heat generation can damage protein that’s around when temperature
reach too high. The experiment of heat study is conducted with interplay between
enzymes. The effect of the membrane cholesterol mole fraction on the initial rate of
seconds. 2 seconds is for SMA spring to unplug the stopper. And 5A is the effective
reduced due to heat transfer is very efficient in a bathed controlled cell. From Table 3, it
shows that the 5A applied for 2 seconds reduce the dropping time factor by 19% and 5A
applied for 5 seconds reduce the dropping time factor by 25%. From the testing result,
heating. To control the temperature minimum less than 40°C is very critical to many
membrane proteins. Referring to Figure 2-37, heat can be controlled to a minimum, less
than 2 seconds. By applying current quarter second periodically and giving the cooling
time for a second, it can minimize the heating effect and provides desired force. For 5A
129
10 seconds testing, the result wasn’t consistent at 2 seconds and 5 seconds condition, the
influence was less. But when the solution was examined, residue was found in enzyme
the light emission intensity of most protein, lipid, etc was low, photo bleaching misleads
130
the result of measurement. When bio-safety experiment was conducted, photo bleaching
effect was encountered. The Experiment Design was using 1.8ml Buffer (Water); 10µl
DHE; 10µl Enzyme (Cox). The ISS setup: Excitation Wavelength: 325nm; Right
stabilization.
The first comparison is for different light intensity. From 350K to 12k, photo
bleaching reduces while light intensity drops. Figure 6-1 shows the red line with intensity
of 350k at its starting point; blue with 100k and pink with 12k.
131
The second test was comparing photo bleaching and enzyme effect. Fig. 2 and Fig.
3 are 2 runs with different start point. Fig. 2 starts from 12k for the photo bleaching test;
just after 15min, 10µl Enzyme is added in. From Figure 6-2, the drop of photo bleaching
Figure 6-3 is another photo Bleaching and enzyme effect comparison at higher
light intensity 100k. It produces similar results as Figure 6-2 result, but larger scale.
132
Figure 6-3 Photo Bleaching vs Enzyme effect @ 100k light intensity
We are going to compare enzyme treated under SMA influence with control
enzyme. It seems to me that the control enzyme doesn’t produce a good steep slope and
the photo bleaching effect is more significant. We ran different ratio between DHE and
enzyme to increase the slope drop with ratio of DHE to enzyme from 100:10; 25:10 and
The photo bleaching effect was minimized by reducing the light passing amount.
Photo bleaching always exist in fluorescent measurement, and can’t be eliminated. But it
can be minimized to the level where it can be neglected within the measurement frame.
133
7 Conclusion and future work
In this project, SMA as an actuator was studied and designed. SMA spring was
made and tested. Self made SMA springs both tensile and compressive performs
according to body temperature SMA properties. Tensile SMA can produce reliable force
for unplugging. SMA spring actuator can produce 5mm to 10mm range vibration with
1Hz to 0.5Hz under water based media cooling. Mixing process was studied both
Turbulent Diffusion process were studied to help with the understanding of the mixing
Fluorescence Measurement provided a very good understanding how well the mixing
process were done. 95% mixing were achieved both in 1atm and 300bars. The mixing
time to 90% took less than 10 seconds for both conditions. 25 seconds is the time for
reaching equilibrium. Compare to free diffusion and unreliable perturbation, which can
take a few hours or days, it is a very reliable mixing under pressure. It is a huge
achievement.
A new HPOC with electrical lead was designed and manufactured. A complete
HPOC with smart activation mechanism were tested with fluorescence measurement. The
whole system is tested under pressure. It proved SMA spring with micro-motor stirrer
work for high pressure mixing. The mixing of 95% was achieved under 300 bar.
There are still some problems existing in the current design. The mixing of two
solution wasn’t 100%. The performance of SMA spring under pressure wasn’t reliable.
The SMA spring doesn’t pull the micro-motor which function as both a stopper and a
stirrer to its highest position. These problems are due to our manufacture facility
134
limitations. Professional manufacture process will improve the performance. Quartz
cuvette fabrication has the similar problem, individual glass shop cannot control cuvette
shape as our requirement and a professional glass factory can improve the dimension and
shape control and produce unified products. Optimum design is needed to compromise
Other solutions such as ultrasonic, piezoelectric motor and stir bar are very
promising on mixing process. From our investigation, the problem of ultrasonic is that
commercially available is hard to fit in HPOC due to its size. Piezoelectric motor was
contamination. Stir bar is a widely used in bio-chemical labs, the difficulty of using stir
bar is how to construct magnetic field around or inside HPOC. Heat generation is a
common problem for all above applications and the minimization of heat generation with
135
Reference
Baz, A., K. Iman, et al. (1990). "The Dyanmics of Helical Shape Memory Actuators."
Chong, P. L.-G., A. R. Cossins, et al. (1983). "A differential polarized phase fluoremetric
study of the effects of high hydrostatic pressure uppon the fluidity of cellular
Chu, Y. and H. Tu (1994). Proc. Int. Symp on Shape Memory Materials Beijing, China.
136
Delaey, L. and S. Chandrasekaran (1982). J. de Phys. Proc. ICOMAT-82.
Foos, M., C. Frantz, et al. (1975). Shape Memory Effects Alloys Penum Press.
Funankubo, H. (1987). Shape Memory Alloys. NY, Gordon Breach Sci. Pub.
George, E. P., R. Gotthardt, et al. (1997). Materials for Smart Systems II. SRS Symp.
Proc.
77.
http://astro.berkeley.edu/~jrg/ay202/node168.html.
Press.
137
Hodgson, D. (1988). Proc. of Eng. Aspects of SMA, Lansing, MI.
Hodgson, D. E., M. H. Wu, et al. (1989). American Society for Metals Metals Park, OH.
Huang, W. and W. Toh (2000). "Training two-way shape memory alloy by reheat
Johnson, A. D. (1972).
Kesten, H. and G. C. Papanicolaou (1979). "A limit Theorem for Tubulent Diffusion."
Lee, H. J. and J. J. Lee (2000). "Evaluation of the characteristics of a shape memory alloy
138
Liu, C. T., H. Kunsmann, et al. (1992). Shape Memory Materials- Fundamental Aspect
Lowry, O. H., N. J. Rosbrough, et al. (1951). "Protein measurement with the Folin phenol
Ma, C. C. and R. Wang (2000). "Frequency Response of TiNi Shape Memory Alloy Thin
370-374.
Morgan, N. B. and M. Broadley (2003). Taking the art out of smart! - forming process
and durability issues for the application of NiTi memory alloys in medical devices.
Proceedings from the materials & processes for medical devices conference.
Otsuka, K. and K. Shimizu (1982). Proc. Int. Conf. on Solid-Solid Phase Transformations,
AIME, NY.
University Press.
Payne, I. W., C. W. Sreed, et al. (1958). Ind. Eng. Chem. 50(1): 47.
139
Pelton, A. R., D. Hodgson, et al. (1995). Proc. 1st Int. Conf. on Shape Memory and
Royer, C. A., T. Ropp, et al. (1992). "Solution studies of the interactions between the
histone core proteins and DNA using fluorescence spectroscopy." Biophys Chem
43(2): 197-211
Royer, C. A., G. Weber, et al. (1986). "Dissociation of the lactose repressor protein
2153.
6295-6301.
Ryhänen, J.
140
Sherman, W. F. and A. A. Stadtmuller (1987). Experimental Techniques in High-
Smith, S. A. and D. E. Hodgson (2003). Shape setting nitinol. Proceedings from the
Sendai, Japan.
Takagi, T., V. Khovalio, et al. (2003). "Two-way shape memory effect and Mechanical
Terry, R. E. (1972). "Improved electrical leads for high pressure system." Rev. of Instrum
43: 1379-1380.
75(1): 471-476.
141
Villas-Boas, M. S., J. L. Silva, et al. (1993). Pressure studies on protein-DNA interactions.
Warram, T. (1993). Actuator Deisgn Using Shape Memory Alloys. Hamilton, Ontario,
Canada.
Weber, G. and H. G. Drickamer (1983). "The effect of high pressure upon proteins and
Wilson, J. R. (1972).
Zhou, J. G., P. L.-G. Chong, et al. (2000). "Shapme Memory Alloy activated high-
pressure opitical Cell for biophysical studies." Re. of Sci. Instrum. 71: 4249-4256.
Zhou, J. G., A. Ostrow, et al. (1998). "New instrument design of high pressure optical
cell for biophysical and chemical research." Re. of Sci. Instrum. 69: 3958-3965.
142
143
8 Appendix
Specimen 1 to 4
3.0
2.0
Load (N)
Specimen #
1.0 1
2
3
4
0.0
-1.0
0 10 20 30 40 50 60 70 80 90 100 110
Time (sec)
Specimen 5 to 8
3.0
2.0
Load (N)
Specimen #
1.0 5
6
7
8
0.0
-1.0
0 10 20 30 40 50 60 70 80 90 100 110
Time (sec)
144
Specimen 9 to 11
3
Load (N)
2 Specimen #
9
1 10
11
-1
0 10 20 30 40 50 60 70 80 90 100
Time (sec)
Figure 8-3 Low temperature tensile testing force vs time (9) unplugging force test
(10)
Specimen 1 to 4
10
6
Load (N)
Specimen #
4 1
2
3
2 4
-2
-1 0 1 2 3 4 5 6 7 8 9 10 11 12
Extension (mm)
145
8.1.2 In house made SMA spring testing
Specimen 1 to 4
8
7
6
5
Load (N)
Specimen #
4
1
3 2
3
2 4
1
0
-1
0 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 160
Time (sec)
Specimen 5 to 7
6
Load (N)
5 Specimen #
5
4 6
7
3
1
0 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150
Time (sec)
146
Specimen 5 to 8
4
Load (N)
Specimen #
3 5
6
7
2 8
0
0 10 20 30 40 50 60 70 80 90 100 110 120 130
Time (sec)
Specimen 1 to 4
4
Load (N)
Specimen #
3
1
2
2
3
4
1
-1
0 10 20 30 40 50 60 70 80 90 100 110 120 130
Time (sec)
Figure 8-8 High Temperature Af force vs step extension loading (1-3) failed
147
Specimen 1 to 3
12
10
8
Load (N)
Specimen #
6 1
2
4 3
0
0 20 40 60 80 100 120 140 160 180 200 220
Time (sec)
Figure 8-9 Current response (3) (by increasing current, the peak of force increases)
Specimen 1 to 1
3
Load (N)
2
Specimen #
1 1
-1
0 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 160
Time (sec)
148
Specimen 1 to 1
1.6
1.4
1.2
1.0
Load (N)
0.8
Specimen #
0.6 1
0.4
0.2
0.0
-0.2
0 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 160
Time (sec)
149
8.2 ISS measurement test results
1200
1000
800
Intensity
600
400
200
0
0 500 1000 1500 2000
Time
150
8.2.2 Free diffusion
600000
500000
400000
Intensity
300000 Series2
200000
100000
0
0.01
314
673
1274
1587
1897
2211
2525
2835
3149
3463
3773
4087
4400
4714
5246
time(s)
600000
500000
400000
Intensity
300000
200000
100000
0
311.38 909.84 1246.87 1584.06 1920.96 2254.06 2590.9 3093.39 3426.85
time(s)
151
diffusion tube feeding 1st time
600000
500000
400000
Intensity
300000
200000
100000
0
139.69
306.63
476.32
654.95
821.73
988.45
1155.2
1321.67
1488.75
1655.32
1818.29
1984.95
2151.64
2318.75
2485.55
2652.31
2819.1
2991.99
time(s)
800000
700000
600000
500000
Intensity
400000 Series1
300000
200000
100000
0
0 200 400 600 800
Time
152
8.2.3 SMA spring and micro-motor mixing results (1 atm) raw results
153
Figure 8-19 3rd result
154
8.2.4 SMA spring and micro-motor mixing result under 300 bars
(normalized)
155
Figure 8-21 2nd Result
156
Figure 8-22 2nd result
157
Figure 8-23 3rd result
158
Figure 8-24 4th result
159
8.2.5 Photo Bleaching
160
Figure 8-26 Photo Bleaching at different intensity
161
Figure 8-28 Comparison of Photo Bleaching rates
162
Figure 8-29 Intensity vs Concentration (nM)
163
Vita
Mr. Oliver Hongchun Xie was born on February 9, 1972 in Qixingpao, Baoqing
University of Science and Technology in 1994. After college graduation, his first job was
China which brought the booming industry of VCD market and established Chinese new
consumer electronics era. He served as a process engineer for Compact Disc Mechanism
Oliver Hongchun Xie came to the United States for his M.S study at Villanova
University in 2001. In July 2003, he finished his M.S. and started his Ph.D. study in the
Process I & II. His research includes squeeze film bearing force analysis; FIB method in
making single digit nano scale hole for bio-applications; Shape Memory Alloy actuator
design, etc.
164