Acta Materialia 50 (2002) 5115–5126 www.actamat-journals.com
Shape memory polymer nanocomposites Ken Gall a,∗, Martin L. Dunn a, Yiping Liu a, Dudley Finch a, Mark Lake b, Naseem A. Munshi b a
Department of Mechanical Engineering, University of Colorado, Boulder, CO 80309, USA b Composite Technology Development, Inc., Lafayette, CO 80026, USA Received 23 April 2002; accepted 22 August 2002
Abstract The paper describes the fabrication and characterization of composites with a shape memory polymer matrix and SiC nanoparticulate reinforcements. Composites based on a SMP matrix are active materials capable of recovering relatively large mechanical strains due to the application of heat. The composites were synthesized from a commercial shape memory polymer resin system and particulate SiC with an average diameter of 300 nm. Composites with weight fractions of 10%, 20%, 30%, and 40% nanoparticulate SiC were fabricated by casting samples with sizes ranging from a few hundred microns to several millimeters. The former size scale is consistent with a microcasting process for manufacturing microelectomechanical systems. The micro-hardness and elastic modulus of the nanocomposites increased by approximately a factor of 3 with the addition of 40 wt% SiC into the base resin. Unconstrained strain recoverability of the nanocomposites was found to depend on the fraction of SiC. For 180° bend tests, the recoverability of the nanocomposites was perfect for SiC weight fractions below 40%. For 40 wt% SiC, permanent bend strains were discovered. Constrained bending recovery force in the nanocomposites was shown to increase by 50% with the addition of 20 wt% SiC. 2002 Acta Materialia Inc. Published by Elsevier Science Ltd. All rights reserved. Keywords: Shape memory polymer; Active material; Nano-particulate SiC; Composites
1. Introduction Shape memory polymer (SMP) materials have the capacity to recover large strains imposed by mechanical loading. The unconstrained recoverable strain limits in SMP materials are on the order of 100%, in contrast to shape memory metals or
Corresponding author. Tel.: +1-303-735-2711; fax: +1303-492-3498. E-mail address:
[email protected] (K. Gall). ∗
ceramics, which can recover about 10% and 1% strain, respectively. Traditionally, a shape recovery thermo-mechanical cycle for SMP materials consists of the following steps as illustrated in Fig. 1 [1–4]: 1. Deform the polymer at a temperature above the glass transition temperature, Tg. 2. Fix the deformed polymer shape and cool below Tg. 3. Upon the completion of cooling, remove the constraint from the polymer.
1359-6454/02/$22.00 2002 Acta Materialia Inc. Published by Elsevier Science Ltd. All rights reserved. PII: S 1 3 5 9 - 6 4 5 4 ( 0 2 ) 0 0 3 6 8 - 3
5116
Fig. 1.
K. Gall et al. / Acta Materialia 50 (2002) 5115–5126
Schematic of the idealized thermo-mechanical cycle leading to unconstrained strain recovery for a shape memory polymer.
4. Heat the polymer above Tg to recover original shape. The polymer must first be processed (molded and cured) to the desired shape. In Step 1 (Fig. 1), the as-processed material is deformed at a temperature above Tg. In Step 2 (Fig. 1), the polymer is held in its final deformed shape and cooled below Tg, typically to a temperature near room temperature. Removal of the constraint required to bend the polymer at the higher temperature is performed in Step 3 (Fig. 1). Generally, the polymer can hold the deformed shape in Step 3 indefinitely, until recovery is necessary [5]. In Step 4 (Fig. 1), the polymer is heated above Tg to recover the original undeformed as-processed shape. If partial constraint remains on the SMP in Step 3, then full strain recovery will not occur, but rather the material will generate a counteracting force. It should be noted that recent work [6], and the results presented here, demonstrate that some SMP materials can be deformed well below Tg and still recover strains when heated towards Tg. Such a thermo-mechanical cycle results in behavior analogous to that of shape memory alloys. Several investigators [7–11] have exploited the large strain recovery behavior of SMP materials to create novel medical devices or Biological MicroElectroMechanical Systems (Bio-MEMS). The SMP-based bio-MEMS typically perform invivo functions such as gripping [9] or releasing [10] of therapeutic medical devices within blood vessels. Bio-MEMS [7–11] have marked a significant increase in minimally invasive surgery technology due to their rapid, hence accurate, blind release and placement capacity. Commensurate with these maturing applications, recent work has examined the bio-compatibility of SMP materials
[12]. Aside from medical applications, SMP-based micro-grippers have potential use in other industrial applications where objects must be manipulated in an inaccessible location such as complex machinery or microsystem assembly [9]. Although SMP materials have found a niche application as an actuation material in MEMS, they have not fully reached their technological potential. A significant drawback of unreinforced SMP materials is their low stiffness compared to metals and ceramics. The low stiffness of SMP resins results in a relatively small recovery force under constraint (actuation force) compared to alternative active actuation materials or schemes. On the other hand, adding reinforcements to the SMP matrix allows tailoring of the material stiffness and attainable clamping force for MEMS applications. A few researchers have studied macroscale composite materials based on a shape memory polymer matrix [5,13,14]. Preliminary work in the late 90s [13] demonstrated that fiberglass and Kevlar reinforcements increased the stiffness of the SMP resins and reduced recoverable strain levels. Moreover, discontinuous fiber reinforced composites showed shape recovery in all directions, while continuous fiber reinforced composites only showed recoverability under transverse tension or bending [13]. Subsequent work [14] has focused on studying the relationship between fiber volume fraction and recoverability for discontinuous fiber reinforced SMP composites fabricated by injection molding. The composite stiffness and recoverable strain levels were found to depend strongly on the volume fraction of the discontinuous reinforcement. The addition of a 50% fraction of chopped glass increased stiffness by a factor of 4.0 and decreased recoverable strains by a factor of about 2.5. Recov-
K. Gall et al. / Acta Materialia 50 (2002) 5115–5126
erable strains in the SMP composites were also discovered to slightly degrade as a function of the number of thermo-mechanical cycles for both pure resin and reinforced materials [14]. Recent work by the present authors [5] on continuous carbon fiber reinforced shape memory polymer composites has focused on materials for the deployable space structure industry. All previous work on SMP based composites has considered macro-scale composite materials. In the present paper we discuss SMP-based composite materials fabricated at µm to mm size scales with nanoparticulate reinforcements. These materials are particularly attractive for MEMS applications where structure size scales are on the order of microns. In order to tailor material properties for micron-sized structures, the heterogeneity of the reinforcement structure must be on the order of nanometers. Polymers have seen considerably less use in MEMS compared to traditional integrated circuit materials due to the transfer of batch processing technologies from the integrated circuit industry. However, as an example, recent work [15] has concluded that certain MEMS sensor applications will benefit greatly from the use of polymers. Thus, motivated by emerging applications, several research groups have made progress towards developing techniques for micro-fabricating polymer MEMS [15–23]. The techniques for the microfabrication of polymer structures include etching [16], ion-beam lithography [17], soft-lithography [18], stereo-lithography [19], hot-embossing [20], and micro-molding [22]. One study [21] considered the effect of nanoparticulate oxide reinforcements on the resulting properties of solgel ceramic materials. However, the bulk of previous work on polymers used in MEMS has focused on unreinforced polymer materials. Moreover, previous work has not considered an active shape memory polymer matrix with reinforcements. In the present paper we describe the fabrication and characterization of composites based on a shape memory polymer matrix and nanoparticulate SiC reinforcements. We begin by discussing the experimental tools used to fabricate and analyze SMP composite materials with sizes ranging from µms to mms. We then present experimental results
5117
on the property characterization (Tg, E, hardness, and stress/strain recoverability) of the base resin material and the nanocomposites as a function of SiC volume fraction. The paper ends with a discussion of the results and succinct conclusions.
2. Materials and methods Fabrication of the SMP nanocomposites was accomplished by casting in molds of various sizes. Composite Technology Development produced the shape memory polymer resin (designated CTDDP-7). The CTD-DP7 shape memory polymer is a thermo set epoxy system, in contrast to many existing SMP systems, which are thermo plastic resins. The 300 nm SiC nanopowder was mixed by hand at 50 °C into the viscous resin mixture prior to curing. Four different percentages of SiC by weight (10%, 20%, 30%, and 40%) were added to the resin to create different particulate composite materials. Prior to curing, the mixture was kept at approx. 50 °C and 25 in.Hg in a vacuum chamber for 1 h to degas. The materials were cured by using a cycle of 4 h at 80 °C, followed by 8 h at 120 °C, and a final post-cure of 4 h at 150 °C. The mold for casting micron sized beams with an approximate cross section of 300 µm by 200 µm was fabricated from a thin copper sheet. The resin mixture was poured over the copper sheet mold and then NaCl plates were pressed onto the top and bottom of the thin copper plate to force complete mold filling. After curing, the top and bottom NaCl mold plates were dissolved with water and the small beams were punched out of the copper sheet. Scanning Electron Microscope (SEM) images of several SMP beams taken in a low vacuum chamber without metal coating are shown in Fig. 2. Two backscatter SEM images of the freeze fracture surface of a sample containing 40 wt% SiC are shown in Fig. 3. Micro-indentation tests were conducted on the micro-beams using a Fischerscope-H100 and a diamond Vickers indentor with a face angle of 136°. The load measuring resolution of the machine is ±0.02 mN and the depth resolution is ±2 nm. All indentation tests were conducted at 25 °C under quasi-static loading conditions. The micro-indentor was also used to
5118
K. Gall et al. / Acta Materialia 50 (2002) 5115–5126
Fig. 2. Scanning Electron Microscope (SEM) images of several nanocomposite shape memory polymer mini beams.
perform micro cantilever and three-point beam bending (Fig. 4) tests at 25 °C to determine the elastic moduli of the different composite materials. Eqs. (1)–(3) show the formula used to calculate the elastic moduli from the bending force (F) versus deflection (v) data. (1) Three ⫺ Point Bend: E ⫽ kL3 / (48Izz) Cantilever bend: E ⫽ kL3 / (3Izz)
(2)
Stiffness: k ⫽ dF / dv
(3)
Moment of Inertia: Izz ⫽ wt3 / 12
(4)
Eqs. (1)– (4) assume beam theory is valid, which is reasonable based on the typical dimensions of the micro-beams. Base resin cylindrical specimens with lengths of 44.8 mm and diameters of 24.9 mm, were cast in a steel mold for compression testing. The purpose of macroscale testing was to compare measurements on samples with traditional size to measurements on micron sized samples. A small tension sample with dimensions of 44.8×19.5×5 mm was cut from the compression plug for tensile testing, and to provide a flat surface for strain measurement on the compression sample. Several small bending
Fig. 3. Backscatter SEM images a freeze fracture surface of a composite with 40 wt% SiC showing the distribution of the SiC in the matrix.
samples were also extracted from the larger compression plug to measure the spatial variation of the modulus. Compression and tensile tests were conducted on a servo-hydraulic mechanical testing load frame at ambient temperature. Small strains in the elastic regime were measured using Digital Image Correlation (DIC), while larger strains were measured via crosshead displacement. A flat surface on the specimen was painted with white and black spray paint to obtain a randomly distributed,
K. Gall et al. / Acta Materialia 50 (2002) 5115–5126
Fig. 4. Fixture used to measure the elastic modulus via (a) three-point bending and (b) a cantilever beam. The force and displacement was measured using an instrumented microindentor.
high contrast, fine black-white dot pattern. During loading, digital images were captured with a Kodak 1.6 mega-pixel CCD camera and a Nikon micro lens with a 60 mm focal length. An aperture of 1/11 was used with an additional white light source to increase the focus depth and give a better image contrast. A typical image is made up of 1000×1500 pixels. In-house and commercial software packages (Yale Digital Image Strain Mapper) were both used to correlate digital images and provide full-field 2-D strain information. Numerous rectangular specimens with dimensions of approximately 22.0×4.0×1.2 mm were cut from a larger casting for Dynamic Mechanical Analysis (DMA). The glass transition temperature, Tg, was measured using the DMA on specimens with the above-mentioned geometries. Dual cantil-
5119
ever and three-point bending fixtures were both used. Based on the relatively low modulus of the resin, a small 10 mm span was required for threepoint bend to obtain reasonable modulus–temperature curves. The parameters for DMA testing of the materials were as follows: Three-point bend— static force=50 mN, dynamic force=40 mN, heating rate=5 °C/min, frequency=1 Hz, Dual Cantilever—oscillatory amplitude of 2.0 µm, heating rate=5 °C/min, frequency=1 Hz. For consistent comparison between tests, the value of Tg was taken to be the peak of the tan–delta curve, although other measures can be used. Recoverable strain tests were accomplished on the micro-beams by bending the beams around a mandrel with a consistent radius at approx. 120 °C and then cooling the beam while under constraint. The constraint was removed at 25 °C and the beams were slowly heated to achieve strain recovery. Recoverable force tests under constraint were conducted in the DMA apparatus using a mm sized specimen that was pre-deformed at 120 °C and cooled to room temperature. The pre-deformed “frozen” specimens were placed in the three-point bend fixture and the probe tip was brought just into contact with the inner surface of the specimen (Fig. 5). While the probe tip was kept fixed by displacement control, the temperature was increased, and recoverable force was measured. Base-resin specimens made from two separate resin batches cured at different times and locations were compared under DMA and compression testing. The results of these tests were identical within expected experimental error. Hence, it was concluded that the resin processing and cure procedures were repeatable and resulted in consistent material behavior.
Fig. 5. Setup used to test bending recovery forces under displacement constraint in the DMA.
5120
K. Gall et al. / Acta Materialia 50 (2002) 5115–5126
3. Results Selected moduli versus temperature curves from the DMA testing are shown in Fig. 6 for (a) the base SMP and (b) a SMP with 20 wt% SiC reinforcement. For both materials, the storage modulus drops continually starting from 25 °C, while the loss modulus begins to decrease around 60°C. The rate of decrease of the storage modulus markedly increases as the temperature approaches 60 °C. The onset of the tan delta curve for both materials appears around 40 °C. The peak of either modulus curve or the peak of the tan delta curve are often employed to define the glass transition temperature, Tg. Here we define Tg as the peak of the tan delta curve. It should be noted that the transition process does not occur at a discrete tempera-
Fig. 6. Modulus as a function of temperature from DMA testing for (a) the base resin and (b) a 20% SiC reinforced resin.
ture so caution should be exercised when using Tg values to interpret results. Based on an average of three tests from two material batches and two different test fixtures, Tg for the unreinforced SMP material was found to be 79.3 °C with a standard deviation of 2.7 °C. Although the onset temperatures are similar in the reinforced and unreinforced SMP materials, the transition occurs faster in the reinforced material. Consequently, the tan delta curve is narrower in the reinforced material and the measured Tg based on the tan delta peak is lower (Fig. 6). Representative displacement plots from DIC measurements on a compression sample are presented in Fig. 7. The purpose of the DIC measurements was to obtain the elastic modulus
Fig. 7. Data from the DIC measurements showing (a) displacement versus the y direction along the sample and (b) stress, axial, and transverse strain in the elastic region.
K. Gall et al. / Acta Materialia 50 (2002) 5115–5126
Table 1 Elastic constants of the base resin system as measured by mechanical loading and Digital Image Correlation (DIC) Stress State
Compression (DIC measurement) Tension (DIC measurement) Compression (Crosshead displacement)
Elastic Modulus, Poisson E (GPa) Ratio, n 0.67
0.50
0.69 0.69
0.46 n/a
and Poissons ratio on a macroscale base resin sample. Fig. 7(a) shows the displacement in the ydirection of the sample (inset), as a function of the y-direction. Full field measurements were taken on the sample surface, but only measurements from the extreme edges are shown in Fig. 7(a) in the interest of space. The displacement distribution was linear and consistent at all x positions for all load levels within the elastic regime. Fig. 7(b) shows incremental measurements in stress–strain space that were used to calculate the elastic modulus and Poissons ratio of the base resin. Table 1 summarizes the elastic constants under compression and tension using both DIC and crosshead position (Fig. 8). The modulus values are identical within experimental error. The strong agreement between the crosshead and DIC measurements are
Fig. 8. Room temperature stress–strain and recovery response of the SMP resin.
5121
due to the extreme stiffness of the large load frame compared to the compliant polymer specimen. Larger strain studies required the used of crosshead position to measure strain since full field DIC measurements were difficult to obtain at larger strains due to the chosen speckle pattern [Fig. 7(a)]. Spaced dot patterns have been successfully used to measure large strains in SMP materials with image correlation [6], however, these results are not presented here. Fig. 8 shows the overall compressive stress–strain response of the base resin at 25 °C. The sample was deformed to approx. 10% strain, unloaded, and heated to recover the imposed strain. The same sample was then subsequently loaded to approx. 22% strain, unloaded, and heated to recover the imposed strain. The loading stress–strain response shows an elastic region followed by some type of inelastic deformation. During unloading, the curve is initially linear followed by a nonlinear instantaneous recovery at lower stress values. Upon heating, the imposed 10% and 22% strains were fully recovered. It should be noted that the test temperature of 25 °C is below the Tg of the material as measured by any standard (Fig. 6). Fig. 9 presents micro-indentation load versus depth curves for SMP micro-beams with various weight fractions of SiC. The left portions of the curves are created during loading, while the right
Fig. 9. Load–depth curves from the instrumented micro-hardness of the composite materials with different weight fraction of SiC.
5122
K. Gall et al. / Acta Materialia 50 (2002) 5115–5126
portions are created during unloading as indicated in Fig. 9. Hardness and elastic modulus can be inferred from the micro-indentation load versus depth curves. Hardness is proportional to the indentor load over the indentation depth squared, so smaller indentation depths at a given load reveal higher hardness values. The elastic modulus of a material dictates the initial unloading response during indentation. Consequently, the initial unloading slope is a measure of the elastic modulus of a given material—stiffer materials show steeper unloading slopes. Based on the load–depth curves, the addition of SiC increases both the material hardness and the elastic modulus of the base SMP resin. More quantitative elastic modulus measurements on the micro-beams were conducted using the micro-indentor and the flexure configurations in Fig. 4. The elastic modulus was calculated using indentor tip force, F, and deflection measurements, v, along with the specimen geometry and Eq. (1)– (4). Fig. 10(a) shows typical indentor force– deflection curves for the cantilever and three-point micro-bending tests. The linear portion of the curves was used to calculate the elastic moduli of various micro-beams. Fig. 10(b) shows the elastic moduli as a function of the wt% of SiC for cantilever and three-point bending tests. The elastic modulus of the unreinforced micro beam material was measured to be approx. 1 GPa. Increasing the fraction of SiC to 40 wt% increases the modulus to approx. 3.0 GPa. The error bars in Fig. 10(b) represent 5% beam size measurement uncertainty. The data reduction for the cantilever bending assumes a rigid fixture to the substrate. The similarity in the three point bending and the cantilever results assure that the compliance in the end fixture does not strongly influence the results. The most critical properties of the active SMP nanocomposites are their strain and stress recovery capacities. Fig. 11 shows the unconstrained strain recovery of SMP micro-beams that were bent around a small mandrel of a consistent radius. The first row shows the beams as processed in the micro-mold. The second row shows the microbeams after being bent at 120 °C, constrained, and cooled to 25°C. The recovery progression of the different beams as a function of time is shown in rows 3 through 6. The beams with a higher fraction
Fig. 10. (a) Typical force–displacement curve for micro cantilever and three-point bend tests. (b) Elastic modulus as a function of the weight fraction of SiC as determined by micro bend tests.
of SiC generally recover faster, but to a lesser extent. Full shape recovery is seen for the beams with 0 wt% to 30 wt% SiC, while the beam with 40 wt% SiC shows permanent strains. Fig. 12 shows the fully constrained recovery force as a function of temperature for the base SMP resin and a composite with 20 wt% SiC (DMA specimen sizes). The data in Fig. 12 was generated by heating pre-deformed bent beams under a fixed displacement as shown in Fig. 5. As seen in the DMA data (Fig. 6) and the free recovery tests (Fig. 11), the composite SMP material reacts faster compared
K. Gall et al. / Acta Materialia 50 (2002) 5115–5126
Fig. 11.
5123
Image demonstrating the unconstrained bend recoverability as a function of SiC fraction and time at temperature.
Fig. 12. Constrained bend recovery force as a function of temperature for reinforced and unreinforced SMP materials.
to the base SMP resin. In addition, the base SMP resin can only generate 300 mN of force, while the composite with 20 wt% SiC generates 450 mN of force for identical constraint conditions and specimen geometry.
4. Discussion Previous studies, [24,25] among others, have examined nanoparticulate reinforcements in polymer matrices. However, the experimental results presented here are the first to investigate nanocomposites with an active shape memory polymer
matrix. When fabricating polymer-based nanocomposites, preceding work [24] has emphasized the importance of particle dispersion and wetting. These interrelated factors are influenced by processing conditions and base material chemistry, and can strongly affect the mechanical properties of the nanocomposites. We have discovered that for the present shape memory polymer resin and nanoparticulate SiC, dispersion and wetting are not major factors for the mixing techniques used at weight fractions up to 30% (Fig. 3). Some particle agglomeration was discovered in the 40 wt% SiC composites. Techniques to circumvent particle agglomeration at higher volume fractions are currently under investigation. We also note that future work will include different matrix and reinforcement materials. The dispersion and wetting characteristics of these new material systems will invariably be different and will require further investigation. Analogous to the Differential Scanning Calorimeter for shape memory alloys, the DMA provides baseline state transition information for shape memory polymer materials. The present results have shown that the glass transition temperature, as measured by the tan delta peak, is sensitive to the presence of nanoparticulate reinforcements. In general, Tg is lower in the reinforced material compared to the unreinforced material (Fig. 6). One possible explanation for this difference is that the kinetics of the glass transition are altered by the
5124
K. Gall et al. / Acta Materialia 50 (2002) 5115–5126
presence of the particles. The particle–matrix interface may help to increase the velocity or number of transition growth fronts on the nanoscale, resulting in a more rapid overall transition. The narrower width of the tan delta peak but the similar tan delta onset points for the reinforced materials (Fig. 11) indicate that a kinetic argument may be a plausible explanation for a shift in Tg. The addition of nanoparticuate SiC reinforcements increases both the hardness (Fig. 9) and elastic modulus [Fig. 10(b)] of the base resin material. The increase in both of these material properties is a direct consequence of the relatively high hardness/modulus of the SiC particles relative to the polymer matrix. The hardness and modulus increases are directly proportional to the weight fraction of SiC. The hardness and modulus of the composite material can be tailored for a given application by altering the weight fraction of the SiC, or using alternative reinforcement materials/architectures. Work is currently underway in this area and results will be presented in due course. An apparent anomaly was discovered when comparing the modulus of the minibeams [Fig. 10(b)] to the modulus of the macroscopic compression sample (Table 1) for the base resin material. For numerous duplicate tests, the elastic modulus of the macroscopic sample (under tension and compression) was consistently lower than determined from the bending of minibeams, 0.7 GPa versus 1.0 GPa, respectively. The difference in modulus for the samples of various size scales was investigated by extracting numerous small mini beams at the top, center, and bottom of the large compression sample (Fig. 7). These three locations can experience different degrees of cure since they are located in regions with diverse heat transfer characteristics. The bottom of the sample is in contact with a metal mold wall, the top is in contact with air, and the middle is contained within a polymer matrix. The elastic moduli, as determined by the three-point bending (Fig. 4) of mini beams, was determined to be 0.52 GPa, 0.66 GPa, and 1.15 GPa for the middle, top, and bottom of the compression sample, respectively. Clearly these values span the average compression modulus in Table 1 and also the modulus of the unreinforced mini beams in Fig. 10(b).
Consequently, the degree of cure, and subsequent mechanical properties, is quite sensitive to the local heat transfer characteristics. These results have ramifications on the curing of shape memory polymer materials for macro and micro applications. For example, a MEMS device fabricated from a shape memory polymer may have unanticipated mechanical properties due to size scale effects influenced by both processing conditions and deformation mechanisms. The recoverability of the composites depends on the presence and volume fraction of the SiC reinforcement. Experimental results demonstrate that the addition of SiC lowers the unconstrained recoverable strain (displacement) limit and increases the attainable constrained recovery stress (force). The decrease in the recoverable strain limit is caused by the inability of the finite fraction of SiC particles to exhibit shape memory characteristics. Instead, the particles are forced to store internal elastic strain energy during loading and freezing. The same mechanism that restricts the generation of large recoverable strain, imparts higher recoverable force levels. The stored elastic strain energy in the particles can be released during the deployment cycle. The release of the stored elastic strain energy in the particles results in a more rapid and ultimately more powerful deployment of the polymer (Fig. 12). The tradeoff between recoverable displacement and force is an important attribute of the shape memory polymer nanocomposites. The recovery parameters can be tailored to optimize recovery force or displacement for specific applications. In closing, we will briefly discuss the appearance of the shape memory effect at a temperature well below Tg, as observed in Fig. 8. The sample in Fig. 8 was compressed at a temperature of 25 °C, well below the Tg for this resin. Strain recovery was accomplished by heating the sample towards Tg. Most previous work on shape memory polymers has not demonstrated inelasticity and recovery so far below Tg. In the present resin system, the unique recovery response far below Tg was only observed in compression, but not tension (or bending) where brittle fracture precedes large strain inelasticity. Under tension and bending, large strain recoverability was only realized by the
K. Gall et al. / Acta Materialia 50 (2002) 5115–5126
traditional thermo-mechanical cycle described in the introduction and used for the bending of mini beams. However, there are resins under development that show recoverability under tension even when initially deformed well below Tg [6].
[3]
[4]
5. Conclusions
[5]
1. Shape memory polymer composites with nanoparticulate SiC (weight fractions of 10%, 20%, 30%, and 40%) were successfully fabricated by micro-casting. Electron microscopy showed that the nanoparticulate reinforcements were well dispersed throughout the shape memory polymer matrix. 2. The micro-hardness and elastic modulus of the nanocomposites increased by approximately a factor of 3 with the addition of 40 wt% SiC into the base resin. 3. Unconstrained strain recoverability of the nanocomposites was found to depend on the fraction of SiC. For 180° bend tests, the recoverability of the nanocomposites was perfect for SiC weight fractions below 40%. For 40 wt% SiC, permanent bend strains were discovered. 4. Constrained bending recovery force in the nanocomposites was shown to increase by 50% with the addition of 20 wt% SiC.
[13]
Acknowledgements
[14]
The work is funded by the National Science Foundation, DMII, Nanomanufacturing, grant number DMI-0200495. The authors gratefully thank Roy Geiss of National Institute of Standards and Technology (NIST) in Boulder, CO for his help with SEM imaging of composite samples.
[6]
[7]
[8]
[9]
[10]
[11]
[12]
[15]
[16]
[17]
References [1] Tobushi H, Hayashi S, Kojima S. Mechanical-properties of shape memory polymer of polyurethane series—(basic characteristics of stress–strain–temperature relationship). JSME J. Ser. I-Sol. Mech. Str. Matl 1992;35:296–302. [2] Tobushi H, Hara H, Yamada E, Hayashi S. (1996) Thermomechanical properties in a thin film of shape memory
[18] [19]
[20]
5125
polymer of polyurethane series. Smrt Matl Struc 1996;5:483–91. Tobushi H, Hashimoto T, Ito N, Hayashi S, Yamada E. Shape fixity and shape recovery in a film of shape memory polymer of polyurethane series. J. Int. Mat. Sys. Struc. 1998;9:127–36. Tobushi H, Ito N, Takata K, Hayashi S. Thermomechanical constitutive modeling of polyurethane-series shape memory polymer. Shape Mem. Matl. 2000;327(3):343–6. Gall K, Mikulas M, Munshi NA, Tupper M. Carbon fiber reinforced shape memory polymer composites. J. Int. Mat. Sys. Struc. 2000;11:877–86. Abrahamson E, Lake M, Munshi N, Gall K. Shape memory polymers for elastic memory composites. 43rd AIAA/ASME/ASCE/AHS/ASC Structures, Structural Dynamics, and Materials Conference, Denver, CO, 2002. Benett WJ, Krulevitch PA, Lee AP, Northrup MA, Folta JA. Miniature plastic gripper and fabrication method. United States Patent 5,609,608, 1997. Lee AP, Northrup MA, Ahre PE, Dupuy PC. Polymer micromold and fabrication process. United States Patent 5,658,515, 1997. Lee AP, Fitch JP. Micro devices using shape memory polymer patches for mated connections. United States Patent 6,086,599, 2000. Lee AP, Northrup MA, Ciarlo DR, Krulevitch PA, Benett WJ. Release mechanism utilizing shape memory polymer material. United States Patent 6,102,933, 2000. Ferrera DA. Shape memory polymer intravascular delivery system with heat transfer medium. United States Patent 6,224,610, 2001. Langer R. Shape memory biocompatible plastics. Adv. Eng. Mater. 2001;3:441–. Liang C, Rogers CA, Malafeew E. Investigation of shape memory polymers and their hybrid composites. J. Int. Mat. Sys. Struc. 1997;8:380–6. Ni QQ, Ohsako N, Ohki T, Sakaguchi M, Wang WX, Iwamoto M. Development of smart composites based on shape memory polymer. International Symposium on Smart Structures and Microsystems, Hong Kong, 2000. Pedersen M, Olthuis W, Bergveld P. Development and fabrication of capacitive sensors in polyimide. Sens. Matl. 1998;10:1–20. Hirano T, Fan LS, Gao JQ, Lee WY. MEMS milliactuator for hard-disk-drive tracking servo. J. MEMS 1998;7:149–55. Sanchez JL, van Kan JA, Osipowicz T, Springham SV, Watt FA. High resolution beam scanning system for deep ion beam lithography. Nuc. Inst. Meth. Phys. B 1998;138:385–9. Xia YN, Whitesides GM. Soft lithography. Ann Rev. Mat. Sci. 1998;28:153–84. Zhang X, Jiang XN, Sun C. Micro-stereolithography of polymeric and ceramic microstructures. Sens. Actu. A 1999;77:149–56. Becker H, Heim U. Hot embossing as a method for the
5126
K. Gall et al. / Acta Materialia 50 (2002) 5115–5126
fabrication of polymer high aspect ratio structures. Sens. Actu. A 2000;83:130–5. [21] Chan CM, Cao GZ, Stoebe TG. Net shape ceramic microcomponents by modified sol-gel casting. Microsystem Tech. 2000;6:200–4. [22] Liew LA, Zhang W, An L, Shah S, Luo R, Liu Y, Cross T, Dunn ML, Bright VM, Daily JW, Raj R, Anseth K. Ceramic MEMS—new materials, innovative processing and futuristic applications. American Ceramic Society Bulletin 2001;80(5):25–30.
[23] Yang H, Deschatelets P, Brittain ST, Whitesides GM. Fabrication of high performance ceramic microstructures from a polymeric precursor using soft lithography. Advanced Engineering Materials 2001;13(1):54–8. [24] Reid CG, Greenberg AR. Influence of silica reinforcement upon the glass transition behavior of acrylic polymers. J. App. Poly. Sci. 1990;39:995–1014. [25] Bhattacharya SK, Tummala RR. Epoxy nanocomposite capacitors for application as MCM-L compatible integral passives. J. Elec. Pack. 2002;124:1.
