Macro to nano control in plastics molding David Kazmer, PE, PhD Professor, University of Massachusetts Lowell October 31 st , 2008
Dec 20, 2015
Macro to nano controlin plastics molding
David Kazmer, PE, PhDProfessor, University of Massachusetts Lowell
October 31st, 2008
Great Things for Akron• Goodyear Headquarters to stay• Prof. Kennedy’s 100 patents• Dean Cheng to National Academy of
Engineering
• PolymerEngineeringis vital
Is U.S. Manufacturing in Decline?
1950 1960 1970 1980 1990 2000 20100
5
10
15
20
25
30
35
Year
Man
ufac
turin
g E
mpl
oym
ent
(% o
f U
S W
orkf
orce
)
Is U.S. Manufacturing in Decline?
1950 1960 1970 1980 1990 2000 20100
100
200
300
400
500
600
700
800
900
Year
Man
ufac
turin
g ou
tput
(%
of
Y19
50 O
utpu
t)
U.S. Manufacturing Productivity
1950 1960 1970 1980 1990 2000 2010
1
1.1
1.2
1.3
1.4
1.5
1.6
1.7
1.8
1.9
2
Year
Out
put
per
Uni
t of
Lab
or C
ost
(Y20
00=
100%
)
US Industry Historical Data
Historical 0.8% Productivity IncreaseRecent 1.5% Productivity Increase
Manufacturing Competitiveness • Manufacturers need 1.5% annual productivity
gains to remain competitive
Cost CategoryTypical
PlantOverseas
PlantAutomated
Plant
Direct materials (resin, sheet, fasteners, etc.) 0.50 0.48 0.50
Indirect material (supplies, lubricants, etc.) 0.03 0.03 0.03
Direct labor (operators, set-up, supervisors, etc.) 0.25 0.08 0.05
Indirect labor (maintenance, janitorial, etc.) 0.05 0.05 0.02
Fringe benefits (insurance, retirement, vacation, etc.) 0.07 0.03 0.03
Other manufacturing overhead (rent, utilities, machine depreciation, etc)
0.10 0.08 0.10
Shipping (sea, rail, truck, etc.) 0.00 0.05 0.00
“Landed” product cost 1.00 0.80 0.73
Manufacturing Competitiveness
10,000 m2
500 m2
DLH IndustriesCanton, OH
Fawer VisteonChangchun, China
ObsoleteCompetitive
Some Manufacturing Research• Macro control
– Real time polymer melt pressure control
• Nano control– Polymer self-assembly
with a functionalizedsubstrate
The Molding Process
Conventional Molding
BarrelHeaters
Reciprocating Screw
Check valveInjectionCylinder
ClampingCylinder
Operator Interface
Stationary PlatenMoving PlatenMold
Pellets
PolymerMelt
Process ControllerHydraulic
Power Supply
Clamping Unit Injection Unit
Tie Rods
• Limited control– Static mold geometry – Open loop process w.r.t.
polymer– So use simulation to
optimize design
Dynamic Feed• System to control
polymer melt in real time– Sensors to monitor
pressure– Movable valve to adjust
flow restriction– Servo control of valve
position from closed loop controller
Dynamic Feed
Dynamic Feed• Two primary issues
– Cost• Pressure transducers for feedback control• Hydraulic servovalves or large servomotors• Increased size of mold components
– Reliability• Pressure transducer longevity & drift• Hydraulic hoses & cylinders
– Too much control energy
Self-Regulating Valve Design
cylinder
valve
Intensification Ratio 100A
A
• Two significant forces:– Top: control force – Bottom: pressure force
• Forces must balance– Pin moves to equilibrium– Melt pressure is proportional to control force– Intensification factor related to valve design
– With high intensification ratio, able to:» Use low cost pneumatic or motors» Eliminate pressure transducers & controller
3D Flow Analysis
Pin Positioning
0
2
4
6
8
10
12
14
16
18
0 0.5 1 1.5 2 2.5 3
Pin Position (mm)
Pre
ssu
re d
rop
(M
Pa
)Q=1cc/secQ=5cc/secQ=25cc/sec
Scaling Laws
0
2
4
6
8
10
12
0 2 4 6 8 10 12
Valve Outer Diameter (mm)
Pre
ssu
re D
rop
(M
Pa
)
2.5 mm
5 mm 10 mm
5.4
690
P
Validation• All validation was performed with a
two cavity hot runner mold– Mold Masters Ltd (Georgetown, Ontario)
• Mold produced binder separators– 1.8 mm thick by 300 mm long– 10 g weight
• Three control schemes investigated– Convention molding– Open loop control– Closed loop control
with pressure feedback
Open Loop Pressure Control
0
5
10
15
20
25
30
35
40
0 2 4 6 8 10
Cylinder Air Pressure (V)
Mel
t P
ress
ure
(MP
a)Cavity 1, Hyd=400, Air=50
Cavity 1, Hyd=800, Air=50
Cavity 1, Hyd=400, Air=85
Cavity 1, Hyd=800, Air=85
Saturated melt pressure
Process Sensitivities
Conventional Weight
7.1
7.2
7.3
7.4
7.5
7.6
7.7
7.8
Conventional W
eig
ht
Conventional Molding
Open Loop Weight
7.7
7.8
7.9
8
8.1
8.2
8.3
8.4
Open L
oop W
eig
ht
Open Loop Melt Valve
• Use of valves reduced both machine sensitivity (main effects) and intra-run variation (whiskers)
Product Consistency
Processing RelativeVariable Variance Valve Gates Open Loop Closed Loop Valve Gates Open Loop Closed LoopMelt Temp 0.0025 0.1479 0.0240 0.0487 5.47E-05 1.44E-06 5.92E-06Mold Temp 0.0025 0.0812 -0.0082 0.0319 1.65E-05 1.66E-07 2.54E-06Inj Pres 0.0025 0.0308 0.0065 0.0109 2.37E-06 1.06E-07 2.99E-07Inj Velocity 0.0025 0.0000 -0.0211 -0.0818 1.66E-12 1.11E-06 1.67E-05Pack Pres 0.0025 0.2667 0.0158 0.0176 1.78E-04 6.21E-07 7.72E-07Pack Time 0.0025 0.1348 0.0826 0.0589 4.55E-05 1.71E-05 8.67E-06
Estimated long run standard deviations (g) 0.0172 0.0045 0.0059
Estimated short term standard deviations (g) 0.0096 0.0039 0.0078
Estimated total standard deviations (g) 0.0197 0.0060 0.0098
Relative process capability, Cp 1.000 3.806 2.915
VariancesSensitivities
2
2
1
m
jj j
dyxy dx
• Significant increase in process capability index
6P
USL LSLC
Flexibility Example• Switch mold inserts to make different cavities
– Varying sizes & thicknesses• Use pressure
valve to controlweights & size
Set max pressure and times for packing stage
Design mold withmultiple valves
For each zoneSet valve to
fully open, closeother valves.
Determine bestmachine settings
for one zone
Add all flow ratesand shot sizes for
filling stage
Mold with optimalsettings for
all zones
Optimalmoldings?
Adjust individual zones
7.7
7.8
7.9
8
8.1
8.2
8.3
8.4
0 5 10 15 20 25 30
Time (min)
Big
Pa
rt W
eig
ht (
g)
0
0.5
1
1.5
2
2.5
Pro
cess
Ca
pab
ility
Ind
ex,
Cp
k.
Large Cavity Control• Adjustments 2, 5, & 6 made for large cavity
– More melt flow and cavity pressure
Small Cavity Control• Adjustments 1, 3, 4, & 6 made for small cavity
– High melt flow rate but lower maximum pressure
6.08
6.09
6.1
6.11
6.12
6.13
6.14
6.15
0 5 10 15 20 25 30
Time (min)
Litt
le P
art
We
igh
t (g
)
0
0.5
1
1.5
2
2.5
Pro
cess
Ca
pa
bili
ty In
de
x, C
pk.
Pre
ssur
e (M
Pa)
100
80
60
40
20
0
0 5 10 15 20 25 Time (s)
Pressure Profile Phasing• The filling of each cavity may be offset in time• By offsetting pressures, the moment of
maximum clamp force is offset• Slight extensions in cycle time can yield drastic
reductions in clamp tonnage
Pre
ssur
e (M
Pa)
100
80
60
40
20
0
0 5 10 15 20 25Time (s)
Pre
ssur
e (M
Pa)
100
80
60
40
20
0
0 5 10 15 20 25Time (s)
Machine Optimization• Machine requirements can be greatly reduced
by optimizing and decoupling each zone
20
25
30
35
40
45
50
24.5 25 25.5 26 26.5 27 27.5
Cycle Time (sec)
Ton
nage
Pre
ssur
e (M
Pa)
100
80
60
40
20
0
0 5 10 15 20 25Time (s)
Pre
ssur
e (M
Pa)
100
80
60
40
20
0
0 5 10 15 20 25Time (s)
Pre
ssur
e (M
Pa)
100
80
60
40
20
0
0 5 10 15 20 25Time (s)
Pre
ssur
e (M
Pa)
100
80
60
40
20
0
0 5 10 15 20 25Time (s)
Summary• The concept of adding degrees of freedom to
polymer processing is very powerful– Real-time melt control is one example– Many other examples exist
Some Manufacturing Research• Macro control
– Real time polymer melt pressure control
• Nano control– Polymer self-assembly
with a functionalizedsubstrate
Flory-Huggins Free Energy• The bulk free energy
i: lattice volume fraction of component i
– ij : interaction parameter of i and j
– mi : degree of polymerization of i
– R : gas constant– T : absolute temperature
Phase diagram of ternary blends
Unguided
Template directed assembly
Highly ordered structures
Polymer A Polymer B
Template Guided Polymer Assembly
Fundamentals
• The total free energy of the ternary system (Cahn-Hilliard equation),
– F : total free energy– f : local free energy
– Ci : the composition of component i
– i: the composition gradient energy coefficient
• The mass flux, Ji is:
– Ci: Composition of component i
– Mi: is the mobility of component i
– i: is the chemical potential of component i
• This leads to a system of 4th order PDEs:
Mass FluxFundamentals
Numerical Method
• Discrete cosine transform method for PDEs
– and are the DCT of and – is the transformed discrete Laplacian,
Simulation Parameters
Validation Experiments• Chemically heterogeneous substrate on Au surface
– Ebeam lithography followed by self-assembly of alkanethiol monolayer– Hydrophylic strips covered by 11-Amino-1-undecanthiol (NH2)– Hydrophobic strips covered by 1-octadecanethiol (ODT)
• Ternary system of polymers used– Polyacrylic acid (PAA): Negative static electrical force– Polystyrene (PS): Hydrophobic– Dimethylformamide (DMF): Solvent, on the order of 98% volume
• Experimental procedure– Polymer solution placed on substrate by pipette – 6 minutes quiescence at room temperature and low humidity– Polymer solution spin coated at varying RPM for in 30 seconds
Validation Experiments• Investigated factors
– Spin coating RPM: 100, 3000, and 7000 RPM– Pattern substrate width: 100 to 1000 nm– PS/PAA ratio: 30/70, 50/50, 70/30– PAA molecular weight: 2k, 50k, 450 k
• Image acquisition– Field emission scanning electron microscopy (JEOL 7401)– Atomic force microscopy (non-contact mode, Veeco NanoScopella)– Fourier transform analysis (PSIA, v. 1.5)
• Model parameters then tuned by inspection of experimental and simulation results
Evolution of Domain Size, R
– The domain size, R(t), is proportional to t1/3
Phase Separation with Solvent Evaporation
Lz=L0-L·exp(-a*t), where t is the time, a is a constant, and Lz is the thicknessof the film at time t, and L0 is the thickness at t=0
Polymer 1 Polymer 2 Solvent
Time
Determination of M and
After comparison of the simulation and the experimental results
M=3.63·10-21 m5/(J*s)=1.82·10-7J/m
Experimental condition:• Spin coating speed: 3000 rpm• Time: 30 seconds• Solvent w%: 99%• PS/PAA (weight) : 7:3
Characteristic length, R=0.829m
Experiment
Experiment
Different Pattern Strip Widths
The simulation results generally matches the experimental behavior The pattern size has to match the intrinsic
domain size
Different PS:PAA Weight Ratios
The volume ratio of PS/PAA has to match the functionalized pattern area ratio
Effects of PAA Molecular Weight
The molecular weight of PAA will affect the shape of the Flory-Huggins local free energy Smaller molecular weight results in a more compatible pattern
Summary
3D simulation for ternary polymer system is established The evolution mechanism is investigated, with the
evolution of the domain proportional to t1/3 The condensed system has a faster agglomeration pace.
Simulation is validated by the experimental results Parameters are estimated, such as the mobility and gradient
energy coefficient. Effects of experimental factors are investigated.
The numerical results matches the experimental results in general, and the model can be used to assist the experiment and theoretical work. Incorporation with high rate plastics manuacturing is the
next focus.
Conclusions
The United States is no longer the R&D super power US R&D was 30% of global R&D in 1970 US R&D is now only 10% of global R&D These facts do not indicate that the US in in decline, but
rather that the rest of the world has made progress Manufacturing will remain a vital source of wealth creation
Competitive advantages are still evolving Natural and human resources
Logistical access to end-users US manufacturers must continue focused R&D
New product innovation Process productivity improvements Employee recruitment, growth, & retention
Acknowledgements
• Melt Control Research• Dynisco, Synventive, Mold-Masters• National Science Foundation (grant #NSF-0245309)
• Simulation of Polymer Self Assembly • Centre of High rate Nano-manufacture at UMass Lowell• National Science Foundation (grant #NSF-0425826)
• Prof. Isayev and the University of Akron
The Effects of the Rotation Speed
The faster rotation speed results in a smaller R value, due to the effects of the faster solvent evaporation
Validation with the Experiments-- with the Patterned Substrate
Measure of the compatibility parameter, Cs
Experiment: SEM images are compared with the template patterns
Simulation: Comparison of result pattern and substrate template are compared element by element
s1(k) - the parameter in the surface energy expression for polymer oneSk - the quantitative representation of the substrate attraction.
, and the greater the better
Determination of Controlling Factors• Huggins Interaction parameter,
– 12,C
: critical interaction parameter. 12,C
for spinodal
decomposition to occur.
– Determines the miscibility of the polymer pair
– Bigger D. P., easier phase separation
–
–
– i: solubility parameter of component I
– The difficulties to obtain accurate solubility parameters.
Determination of Controlling Factors
• Gradient energy coefficient,
– a : Monomer size, the affecting radius of van de
Waals force– Determines the domain size and interface thickness– – D: Diffusivity– Determines the kinetics of the phase transaction.
The values of k and D are estimated by benchmarking with the experimental results, as shown later.