Abstract
A widely used industrial felted foam was compared to a volumetrically heated and compressed foam in this study in terms of cushion performance. The off-the-shelf melamine open cell foam and polyurethane open cell foam were triaxially heated and compressed via an in-house-made device, and energy dissipation was then compared to the felted off-the-shelf open cell melamine and the polyurethane foam. The hysteresis cycle compression test and impact test, in conjunction with a high-speed video camera, were conducted to measure the energy dissipation, G-value, and the Poisson’s ratio. Scanning electron microscopy was used to investigate the cellular morphology before and after the felting and volumetric heat and compression. The study showed that both felted foam and triaxial compressed foam demonstrated reasonably well the cushion performance due to the presence of auxetic and microcell structures.
Introduction
Polymer foams are cellular materials. They are commonly produced through an extrusion foam process where a foaming agent is introduced into a base material creating void cell structures in the foam. The polymer foam is classified into an open cell structure or closed cell structure. The open cell structure is produced by destroying the liquid films of gas bubbles during the foaming process, and air fills all of the spaces within the material, making the foam soft or weak. In a closed cell structure, the air or gas is trapped in the cell, giving strength to the foam.
Felted foam is the foam that is compressed from lower density conventional foam through a commercial process. In order to create felted foam, the foam is placed between a pair of parallel platens. The platens are heated to a temperature that enables them to generate a skin layer when compressing the foam. Thin aluminum sheets are placed on the opposite sides of the polymer foam sample that are normal to the direction of the compression from the two platens. A stop with adjustable dimension is fixed on one of the platens so that the movement of platens is within limits. After compression, the foam is removed and cooled in an ambient environment. The felted foam is widely used in acoustic and thermal insulation, especially for polyurethane (PU) foam. In recent years, doctors have applied felted foams in plantar pressure releasing for diabetic foot ulcers. 1
In 1987, Lakes 2 converted conventional foam with positive Poisson’s ratio to an auxetic, i.e. negative, Poisson’s ratio foam through a volumetrically heated and compression process in the laboratory. Poisson's ratio ν is defined as the negative transverse strain of a stretched or compressed object divided by its axial strain. All commonly known cellular materials (natural as well as synthetic) have a convex cell shape and exhibit a positive Poisson’s ratio, typically 0.2–0.4. The re-entrant polymer foams developed by Lakes have a Poisson’s ratio of –0.7, and the cell shape is concave. Based on Lakes’ method, Chan and Evans 3 developed a multi-stage fabrication method allowing a relatively large sample such as 20 × 20 × 5 cm to be produced. The fabrication methods involved squeezing a large foam sample into a smaller sized mold by hand. The mold was then heated above the softening point of the foam and cooled to room temperature. The process was repeated until the foam reached a final size. This kind of process can only be carried out manually in the laboratory. Brandel and Lakes 4 transformed successfully the large cell sized low-density polyethylene (LDPE) foam into re-entrant materials exhibiting a negative Poisson’s ratio via a combination of air pressure and vacuum treatments.
Several two-dimensional models have been proposed to explain the auxetic behavior of the foams. Masters and Evans 5 suggested that elastic hinges at the cell joints could enable the cell to deform and restore when the load is applied and removed. Grima et al. 6 proposed a theory that the auxetic cellular structure behaves like “rigid units,” and the major deformations occur along the length of the more flexible ribs that form “kinks” at their centers as a result of the extensive buckling.
Lakes and Elms 7 calculated energy absorption for the dynamic impact based on the strength and stiffness derived from the holographic indentation test. They concluded that the cushion performance is considerably higher for the re-entrant method than conventional foam. Scarpa et al. 8 studied the dynamic properties and concluded that the auxetic foam has a lower Young’s modulus than conventional foam. Bezazi and Scarpal 9 conducted compressive cycle loading for auxetic foam and conventional foam and found that the energy dissipated per unit volume for auxetic foam is 16 times higher than the conventional foam. The investigation conducted by Smith et al.10,11,12 showed that re-entrant foams had higher yield strength and less stiffness than conventional foams with the same original relative density. It also has been further proven that re-entrant foams indeed densify under indentation due to increase in shear stiffness. It is not clear in the reviewed literatures that how the density plays a role in cushion performance. The auxetic foam is obtained from lower density foam; its density is higher than that of conventional foam. As long as the volume compression ratio is within the limits of the range, the auxetic foam will exhibit better cushion performance than the conventional foam due to increased density. Because the densities of both foams are different, they cannot be compared equally. The auxetic foam should be compared to conventional foam with the same density.
Both felted and auxetic foam are made through a process that buckles the cell ribs inward, creating re-entrant geometry. The difference between the two is that the felted foam is fabricated through a single axial heat and compression commercial process, and the auxetic foam is derived through triaxial volumetric compression, and samples are produced in the laboratory. The felted foam is widely used in commercial insulation. The auxetic foam is only at the stage of laboratory-based research. In comparison to auxetic foam, very little literature can be found to study the dynamic performance of the felted foam, probably due to its wide industrial application.
The aim of this study is to make auxetic foams and compare their cushion performance to the felted off-the-shelf foams with the same density, specifically the G-level, during an impact test and compare the energy dissipated by the foams. This study is based on the hypothesis that the felted off-the-shelf foam will more likely conserve the similar cell geometry as the auxetic foam. This paper considers cellular morphology and Poisson’s ratio of the foams to be two important areas to interpret the cushion properties.
Materials and method
Materials
Two kinds of the original foam materials were examined, a gray melamine open cell foam with a density of 9.6 kg/m3 and a yellow PU open cell foam with a density of 44.9 kg/m3. These two materials are the base materials for felting as well as triaxial compression. The felted melamine open cell foam was manufactured by BASF Chemicals with the trade name Basotec®. It is felted foam with a measured density of 27.2 kg/m3. The felted PU open cell foam was supplied by Pinta Foamtec, with the density of 112.1 kg/m3. The trade name is Pinta 9700. Samples of felted foams were cut from bulk foam using a hot wire saw.
The laboratory triaxial densification was designed to permanently compress fully cured polymer foam in three directions. In order to fabricate sample size foam, an apparatus was designed to allow the fabrication process to be similar to a commercially produced process. The compression apparatus has a base to support the conventional foam, two side compression plates, and one top-to-bottom plate. Densification consisted of four stages. Firstly, the apparatus was preheated to a specific temperature in the oven. Secondly, the apparatus was taken out of the oven and the fully cured polymer foam was placed in between the compression plates of the device. Thirdly, the foam was manually compressed by one side plate pushing horizontally inward on the foam followed by the adjacent side plate repeating the action. While remaining in the mold, the compressed foam was then heated and cured in the oven for a set amount of time. Finally, the mold was removed from the oven and was opened immediately, exposing the foam to ambient temperature. Rapid cooling of the compressed foam was possible because the low density of the original polymer cell structure was maintained in the microcell structure, allowing hot air to be exchanged easily.
The initial sample size of the triaxial compressed foam was determined based upon the desired final sample size and compression ratio used in commercial felted foam in order to make a comparison. For example, a targeted 5.1 × 5.1 × 2.5 cm compressed foam with a density of 51.23 kg/m3 had an initial density of 8.8 kg/m3. Densification began with a sample of fully cured polymer foam that had the initial dimensions of 9.1 × 9.1 × 4.6 cm in the apparatus. The first side compression changed the dimensions to 5.1 × 9.1 × 4.6 cm followed by the adjacent side compressing the foam to 5.1 × 5.1 × 4.6 cm. Finally, the top-to-bottom compression finished foaming the foam into a 5.1 × 5.1 × 2.5 cm sample. Then the foam was secured to the base and placed in the oven to cure. The heating temperature and duration were dependent on the softening temperature of the foam type.
Densities and sizes of the felted off-the-shelf foam and triaxial compressed foam.
Method
Morphological analysis
In order to understand the differences in cell structure between the original foam, felted foam, and the triaxial compressed foam, a thin sliver of the foam was cut from the samples with a razor blade along the thickness direction. These samples were then examined by scanning electron microscopy (SEM).
Hysteresis cycle testing
Two tests were conducted to evaluate the cushion performance of the polymer, the hysteresis cycle compression test and the impact test. The room temperature at the measurement ranged from 20℃ to 21℃, and the humidity ranged from 30% RH to 60% RH. The equipment used was an INSTRON 5567 with a thermal chamber. The collected data are the averaged results of the five samples per material per ASTM D3574-05. The hysteresis cycle compression testing was performed according to ASTM D3574-05, represented by 5-cycle compression stress–strain loop curve. The area in the center of the hysteresis loop is the energy dissipated by the materials due to their plasticity. The stress–strain curve consists of three general regions. At region 1, when the foam is compressed less than 10%, the foam acts like a linear spring cushion. At region 2, when the foam is compressed between 10% and 70%, the foam cushion has a plastic deformation. Most cushion performances take place in this region. At region 3, when the foam cell is destroyed, the foam has no cushion effect but acts as a solid pad before rebound.
Prior to starting the hysteresis cycle compression testing, a sample of each material was compressed to 90% or until the foam could be compressed no further. The purpose of the testing was to create a “hit bottom” scenario to identify the three regions of the stress–strain curves for all materials.
Impact testing
Impact testing was conducted on a cushion tester. The purpose of the impact test was to measure peak acceleration and impact duration on a platen that free falls to the polymer foam. Five samples of each material with the same 5.1 × 5.1 × 2.5 cm dimensions were placed between the platen and the solid support platform. The weight of the upper plate was 4.5 kg with a drop height of 3.7 cm. The drop height 3.7 cm was chosen from trial drop heights to avoid hitting the bottom of the foam by the platen, which would result in a high G-level.
(a) The apparatus for fabricating the triaxial compressed foam and (b) The felted and triaxial compressed foam.
Two sets of data were collected from the impact testing. One set of data included the G-value on the platen, the impact duration, and the impact velocity when the platen impacts on the foam. The data were captured by an accelerometer installed on the platen that channeled the impact data into the Lansmont TestPartner™ software for analysis. TestPartner is a Windows-based software system that includes shock response spectrum (SRS) analysis with shock response animation in both 2D and 3D modes. Shock response analysis presents the peak acceleration G and the shock duration as well as velocity change during impact. The second set of data provided images of the foam’s deformation during impact via a high-speed video camera. From the deformation images during impact, a Poisson’s ratio, i.e. positive or negative value, can be determined through the observation of the deform direction of the foam. When a non-auxetic foam material is subjected to an impact, the cell material tends to move away from the impact area that leads to an expansion of the side panel of the foam. When an object hits an auxetic foam material, the cell materials compress into the place of the impact, and the side panel of the foam indents toward the area of the impact. The transverse strain can be obtained from dividing the foam indentation or expansion by the original width of the foam, and the axial strain can be calculated when the maximum deformation is divided by the original foam thickness.
Experimental set up for impact testing.
Results and discussion
Cellular morphology
Several phenomena in the SEM are noted. Firstly, the cell wall of the melamine foam is thinner than the PU foam, and the averaged cell size is smaller than the PU foam. The cells of both uncompressed foams, Figure 3(a) and Figure 4(a), are mixed with large and small cells. After the felting and volumetric heat and compression, the cells of both foams deform irregularly (Figure 3(b) to (f), Figure 4 (b) to (f)). The location of the deformation corresponds more to the spaced areas and large-size cell. Most ribs of the large-size cell of the tested foams appear to have broken extensively, whereas the medium-size cells buckle and twist along the ribs. Although the deformation of these medium-size cells failed to produce an ideal auxetic structure described in the literature,2,6 they present an inward cell structure that has essential features to exhibit an auxetic behavior when the foam is loaded. Interestingly, the smallest cells appear to have not been affected by the compression and curing and retained their shapes. These small cells, along with the wrinkled medium-size cells, produced compact microstructures that absorb more impact energy than the uncompressed foams. The locations of the small cells are mainly in the cell walls of the large cell. The compression forces destroyed the joints between the large cells and left the small cells untouched. Secondly, there is a significant presence of “Y-joints” or rib breakage in the cured melamine foam illustrated in Figure 4 (e) to (f). These “Y-shaped joints” behave like “rigid triangles,” as described by Grima et al.,
6
that rotate relative to each other. This finding echoed with Grima’s theory that the foam cell with “Y-shaped joints” exhibits auxetic behavior as a result of re-rotation of the “triangular joints” and unfolding of the “kinks” when the foam is uniaxially loaded. Although “Y-shaped joints” exist also in the cured PU foam illustrated in Figure 3 (e) to (f), the number of “Y-shaped joints” is less than the melamine foam. Lastly, a clearly aligned cell band was observed near the skin segment of the felted melamine foam as well as in the center part. The stretch direction is normal to the felting direction. The unique reoriented and hinged joints exist in both felted and triaxial compressed melamine foam.
SEM micrographs of the felted and triaxial compressed melamine foams. (a) Uncompressed melamine foam (90 × magnification). (b) Triaxial compressed melamine foam (90 × magnification). (c) Skin segment of the felted (90 × magnification). (d) Inner segment of the felted off-the-shelf melamine foam off-the-shelf melamine foam (90 × magnification). (e) The triaxial compressed melamine foam (250 × magnification) and (f) Skin segment of the felted off-the-shelf melamine foam (250 × magnification). SEM micrographs of the felted and triaxial compressed PU foams. (a) Uncompressed foam PU foam (80 × magnification). (b) The triaxial compressed melamine foam (80 × magnification). (c) Skin segment of the felted off-the-shelf PU foam (80 × magnification). (d) Inner segment of the felted off-the-shelf PU foam (80 × magnification). (e) The triaxial compressed PU (foam 250 × magnification) and (f) Skin segment of the felted off-the-shelf PU foam (250 × magnification).
Hysteresis cycle testing
Comparison of felted and triaxial compressed foam.
Impact testing
Summary of the impact G-values of different foams.
Measured strain and Poisson’s ratio ν.
The negative Poisson’s ratio of the volumetrically heated and compressed foam in Table 4 appears to be smaller than the reported values of –0.6 to –0.7 in previous studies.2,3,12 This is due to the differences in measuring methods of Poisson’s ratio and the fabricating method of the auxetic foam. In this study, Poisson’s ratio was obtained from the maximum transverse and axial deformations during impact. This Poisson’s ratio reflects the actual dynamic cushion performance in the field. Considering sample size and impact speed, it is perhaps not practical to achieve a compression Poisson’s ratio of –0.6 to –0.7. The sequentially triaxial compression differentiates itself also from the manually squeezing, multi-stage method reported in the literature. 3 Overall, the felting and triaxial compressing processes enable us to bring the Poisson’s ratio of original foams down to near zero and negative values despite the difference of the materials and densities. The Poisson’s ratio for felted and volumetrically compressed foams is negative and close. The SEM for melamine foams showed that the felting process changed the morphology not only in skin layers but also created similar morphology in its inner layer that is close to the structure of volumetrically compressed foams. This might be the reason why both Poisson’s ratios are close.
Conclusion
A laboratory-based manufacturing method is introduced in this study to produce foam with a negative Poisson ratio via a triaxial compressing method. Compared to manually squeezing the foam into different-sized molds, this mechanical device leads to possible automation by connecting a plunge to each of the three platens driven by a hydraulic mechanism.
The image captured by the high-speed video camera during the impact test of both felted and triaxial compressed melamine foam demonstrated characteristics of negative Poisson’s ratio. Although the calculated negative value of the felted foam is smaller than the triaxial compressed foam, it proved the hypothesis that the felting process produces similar auxetic structure for certain materials. The SEM micrographs of both PU and melamine foam also showed a similar cell geometry between the skin segment of the felted foam and triaxial compressed foam.
From the SEM micrographs, the densification through triaxial compression results in a more uniformly distributed cell size. The study concluded that the deformable cells contributed largely to the auxetic behavior due to their inward structure generated by re-entrant method, namely the felting and volumetric compression process. In addition, the “Y-shaped joint” cells in both foams present also auxetic behavior. A large number of “Y-shaped joints” and the microstructure in melamine foam are two major factors that led to a negative Poisson’s ratio. More investigation is needed to better understand the quantitative relationship between cell size and auxetic foam behavior.
From the hysteresis cycle compression testing and impact testing, both felted foam and triaxial compressed foam exhibited more energy dissipation compared to the conventional-based foam. The better cushion performance of both compressed foams is not contributed to by the re-entrant structure alone but by a combined contribution from three factors: density, materials, and the re-entrant structure. For example, the G-value difference between the felted and triaxial compressed foam is 4.2% for the PU foam and 6% for the melamine foam on the first drop, and 7% for PU foam and 21% for melamine foam on the fifth drop.
The simple felted foam manufacturing process has a great advantage over the triaxial compression when it comes to commercialization. A further study is recommended to focus on developing a foam felting process that leads to creation of a more auxetic foam structure and microcell structure.
The images of maximum deformations of the foams during the impact. (a) The triaxial compressed melamine foam. (b) The felted melamine foam. (c) The triaxial compressed PU foam and (d) The felted PU foam.
Footnotes
Acknowledgements
This work was conducted in part under a NASA research project, “Protective and flexible packaging for critical space hardware” in 2009. NASA’s grant and technical assistance is gratefully acknowledged.
Funding
This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.