Materials and Design 137 (2018) 298–304

Contents lists available at ScienceDirect

Materials and Design

journal homepage: www.elsevier.com/locate/matdes

Evaluation of novel temperature-stable viscoelastic polyurea foams as

helmet liner materials
B.J. Ramirez ⁎, V. Gupta ⁎
Department of Mechanical and Aerospace Engineering, University of California, Los Angeles, CA 90095, USA

H I G H L I G H T S G R A P H I C A L A B S T R A C T

• Viscoelastic polyurea foams were

manufactured that exhibit excellent
temperature stability from -15°C to
50°C.
• Polyurea foams were tested using the
NOCSAE and FMVSS 218 standards for
the football and motorcycle helmets,
respectively.
• Polyurea foam resulted in a reduction of
22% in peak g’s and 25% in the Severity
Index and Head Injury Criteria.
• Polyurea foam reduced the headform’s
peak g’s by 26% at -15°C and 20% at
50°C.

a r t i c l e i n f o a b s t r a c t

Article history: Viscoelastic polyurea foams with densities of 98, 170, and 230 kg/m3 were manufactured and integrated into hel-
Received 26 June 2017 met shells and tested using the NOCSAE and FMVSS 218 standards for football and motorcycle helmets, respec-
Received in revised form 26 September 2017 tively. For football helmet testing, a Riddell Revolution helmet shell was used. The helmets with a foam liner
Accepted 13 October 2017
thickness of 28 mm were dropped using a NOCSAE medium headform (4.9 kg) from heights ranging from
Available online 14 October 2017
0.305 to 1.524 m. All impact tests were done on the crown of the helmet and dropped on a 12.7 mm modular elas-
Keywords:
tomer programmer pad. Use of PU98 foam resulted in a reduction of 22% in peak g's and 25% in Severity Index and
Viscoelastic foams Head Injury Criteria values when compared to helmets with the original VN600 foams under ambient conditions.
Polyurea These tests were repeated under varying temperatures along with a reduced liner thickness of 22.2 mm. In these
Football helmets tests, PU98 foam reduced the peak g's by 18% at 23 °C, 26% at −15 °C, and 20% at 50 °C when compared to VN600
Impact mitigation helmets. The FMVSS motorcycle helmet tests demonstrated the success of using an additional layer of polyurea
Strain rate sensitivity foam liner on top of the existing expanded polystyrene liner in reducing the peak g's by 17%.
© 2017 Elsevier Ltd. All rights reserved.

1. Introduction event whereas motorcycle and bicycle helmets use plastic crushable
foams to protect the wearer from occasional impacts and requires re-
Use of elastomeric and plastic foams is ubiquitous in mechanical, placement of helmets after an accident [5–12]. These foams dissipate
civil, and aerospace structures and all protective head and body gears impact energy through local viscoelastic or plastic crushing processes
[1–4]. Most sports applications use viscoelastic elastometric foams be- that involve bending, twisting, and buckling of the struts that form the
cause of the need to take multiple impacts within the same sporting skeleton of their structure [13–18]. The compressive stress-strain be-
havior of a typical viscoelastic or a plastic foam exhibits three distinct re-
⁎ Corresponding authors. gions: (i) a linear elastic region associated with cell wall bending, (ii) a
E-mail addresses: [email protected] (B.J. Ramirez), [email protected] (V. Gupta). stress plateau region that results from continuous collapse of the foam's

https://doi.org/10.1016/j.matdes.2017.10.037
0264-1275/© 2017 Elsevier Ltd. All rights reserved.
 B.J. Ramirez, V. Gupta / Materials and Design 137 (2018) 298–304 299

cell structure on itself through non-linear fully recoverable (viscoelastic became increasingly more efficient in dissipating the impact energy as
foams) or irrecoverable (plastic foams) buckling of its struts, and (iii) a the loading rate was increased. Tests done using the Split-Hopkinson
densification region where the already-collapsed cells begin to com- Bar setup showed that they dissipated more impact energy than the
press against each other to near full density, resulting in a steeply rising plastic expanded polystyrene (EPS) foam at strain rates N 2000 s− 1
stress. The amount of energy absorbed or dissipated by the foam is while maintaining the peak stress below 1.5 MPa. Because of the very
roughly equal to the area under its stress-strain curve [13–18]. low glass transition temperature of polyurea (Tg = − 50 °C) they
For a given foam thickness, which is typically constrained by the spe- were able to maintain their ambient impact properties at low and
cific application, the design process reduces to determining the lowest high temperatures [20–22]. Furthermore, even with their uniform mi-
density of the foam per unit volume that maximizes impact energy ab- crostructure, they performed similar to foams with modulated stiffness
sorption while keeping the peak load below the injury threshold level. and density. Finally, being viscoelastic, they recovered fully after each
For head health applications, the peak transmitted stress must be kept impact and were able to reproduce their energy absorbing characteris-
below 1.5 MPa [1,2]. The manufacturing technology has advanced sig- tics impact after impact that were spaced 30s apart. This unique behav-
nificantly over the past two decades that any polymer can be foamed ior was tied to their microstructure that comprised of large polyhedral
into the desired density by controlling the process variables and by cells (300 μm–500 μm) covered with perforated membranes with
use of various surfactants, foaming and gelling agents. The industrial small apertures (20 μm–70 μm) as shown in Fig. 1. This makes the
processes are also well developed to produce foams with uniform micro- polyurea (PU) foam strain rate sensitive as the rate at which the air es-
structures in large quantities at fairly low cost. Today foams are used in capes the cells through perforations depends upon the loading rate. At
practically every industry. Despite their ubiquity and maturity, there are lower strain rate, typically at the start of the loading event, big cells sim-
still areas where improvements in the foam technology can be made, es- ply collapse with air escaping freely through tiny perforations. This
pecially those that pertain to impact applications. These are discussed limits excess buildup of stress like any other viscoelastic foam. As the
next. loading proceeds, both strain rate and level of material stress increases.
Almost every current application utilizes foams with a uniform mi- During this phase, the rate at which the air escapes the cells cannot
crostructure. Even with the choice of the lowest density, these foams catch up with the rate of loading. Consequently, the air that remains in-
are not truly optimized. This is because the initial linear response and side each cell acts to stiffen the cell while the air that escapes adds to vis-
the plateau stress which determine the impact attenuation and momen- coelastic damping. Finally, when the impactor velocity has substantially
tum trapping properties of the foam depend strongly on the loading rate reduced and the strain rate has dropped, the remaining undeformed
[1–2,15–18]. The loading rate, however, changes within the same im- cells collapse slowly, giving plenty of time for the air to escape freely
pact event. The material strain rate is lower at the start of the loading through tiny perforations. Each cell thus acts like a time-dependent vis-
as the foam cells start to collapse, and then builds up quickly during coelastic damper on the microstructural scale. Therefore, the PU foams
the densification process, and finally, reduces significantly when the ve- are able to manage the varying material strain rate that occurs within
locity of the impactor has slowed down considerably. One way to further the same loading event without the need to modulate the material den-
optimize the foam's performance is to create a section with modulated sity or stiffness. In addition to the microstructure, energy is also dissi-
stiffness or density to account for this varying material strain rate within pated by dynamic bending, twisting and rotation of the cell walls
the same loading event. Such a foam structure cannot be manufactured during impact like any other foam. In PU foams, cell walls are made
in a single low-cost manufacturing step. It would require lamination using the polyurea elastomer which itself is capable of absorbing the im-
where foam layers of different density, stiffness, and thickness are pact energy through sliding and stretching of its flexible molecular
stacked up in a specific order. In addition to being expensive, the lami- chains. The micro-phase segregation of hard and soft domains along
nation process is not easily scalable as required for production of high with extensive hydrogen bonding present within the polyurea structure
volume commodity products. also grants it the ability to chemically tailor its mechanical toughness,
Another area of opportunity is improving the performance of foams resiliency, and hardness [23–28]. The polyurea chemistry chosen for
at low and high temperatures. Present viscoelastic foam pad systems the foams presented in this paper was already optimized for maximum
(e.g., vinyl nitrile, VN and thermoplastic polyurethane, TPU) become energy dissipation, as coating liners [29,30] and adhesives [31–33],
hard at cold temperatures and soft at high temperatures [19]. In both while exhibiting exceptional strain rate sensitivity [34–36] and strength
situations, the peak transmitted stress greatly exceeds the injury [37,38].
threshold limit. Ideal foam should have a Tg well below 0 °C so that it In this paper, the combined benefits of temperature stability and rate
can maintain its viscoelastic properties over a wide range of sensitivity of low density (≤200 kg/m3) viscoelastic PU foams are dem-
temperatures. onstrated by integrating them as liners into football and motorcycle hel-
Motivated by these goals, in a previous communication [3], we pre- mets, and evaluating their performance using the NOCSAE (National
sented polyurea-based strain rate sensitive viscoelastic foams that Operating Committee for Standards in Athletic Equipment) and

Fig. 1. Scanning electron micrographs of the microstructure of the PU Foams. The red arrow shows the cell diameter (“cell size”) while the yellow arrow shows the cell aperture (“cell
windows”).
300 B.J. Ramirez, V. Gupta / Materials and Design 137 (2018) 298–304

FMVSS 218 (Federal Motor Vehicle Safety Standards) standards, respec- Commercially available D3O material strips were used as a reference
tively [39,40]. The football helmet is taken as an example of protective for evaluating the performance of PU foam strips when used as an addi-
equipment in sports where the multiple hit capability is required tional liner layer on top of the EPS pads for motorcycle helmet
whereas the motorcycle helmet application serves as an example of application.
how to effectively improve the performance of existing EPS-based hel-
mets by adding a thin layer of the PU foam liner on top of the EPS pad. 3. Experimental procedures

2. Materials 3.1. Material characterization

Open cell polyurea foams were fabricated by mixing modified Meth- Low strain rate (≤10−1 s−1) compression tests for material charac-
ylene Diphenyl Diisocyanate (MDI) compound with oligomeric diamine terization were conducted using an Instron Micro-Tester (Model
with different mix ratios. Polyureas of this stoichiometry are known to 5942) equipped with a 2 kN load frame. For these tests, disc samples
have behaviors within the viscoelastic regime [20–22]. It consisted of of 28 mm nominal diameter and 10 mm thickness were cut from a
preparing the B-side of the mix by taking the oligomeric diamine polyol large molded block. Samples were then compressed under stroke-
and mixing it with appropriate amounts of deionized water, surfactants, control with cross-head velocity of 24 mm/min in order to examine
cross-linkers, and catalysts to control the gelling and foaming reactions the material response over a range of quasi-static strain rates. These
so as to obtain a free-rise foam of predetermined density. The A-side of cross-head speed yielded a strain rate of 0.04 s−1. Force-displacement
the mix was the as-received modified Methylene Diphenyl Diisocyanate data obtained from the machine was converted into the stress-strain
(MDI). The A and B sides were mixed in 1:4 weight ratio and poured data using the sample dimensions.
into a 90 mm × 90 mm × 90 mm mold. The rise ratio of the foam was Intermediate strain rate (101–102 s−1) compression tests for materi-
between 7 and 8. Because of the high reaction exotherm, no curing al characterization were carried out using an Instron DynaTup (Model
was needed after the completion of the foaming and gelling reactions. 8250) drop-weight impact tester. Specimens with aforesaid dimensions
The foam was completely tack free after 3 min. The entire polyurea were tested by placing them on a 45 mm-diameter flat force plate and
foam block was removed from the mold, with sections cut from the cen- impacted using a 5.5 kg balanced mass attached to a 75 mm-diameter
ter of the molded block to prepare specimens for mechanical testing, flat stainless steel indenter head. The transmitted force was measured
characterization of physical and microstructural properties, and inser- using a force transducer (Kistler Instruments, Model 9041A) with a
tion into helmet shells. full scale output of 90 kN and sensitivity of 4.3 pC/N. Impact energy
Foam blocks with densities of 98, 170, and 230 kg/m3 were prepared was varied by changing the drop height of the indenter head. The dis-
using the same polyurea mix with a density of 1070 kg/m3. Hereinafter placement of the top sample surface during impact was also measured
these samples are referred to as PU98, PU170, and PU230, respectively. using high speed photography. The recorded force-displacement data
These densities were chosen because they fall within the range used in was converted into engineering stress-strain characteristics using the
protective helmets [5–11]. Fig. 1 shows the scanning electron micro- _ was calculated using:
initial sample dimensions. The strain rate ðεÞ
graphs of these foams. The PU foams consist of open spheroid cells    
. d l−lo 1 dl v
with wall apertures (perforations). The nominal cell sizes are 1000
ε_ ¼ dε ¼ ¼ ¼ ð1Þ
μm, 600 μm, and 400 μm for PU98, PU170, and PU230, respectively, dt dt lo lo dt lo
with corresponding nominal cell apertures of 300 μm, 200 μm, and
105 μm. where v is the impact velocity and lo is the specimen thickness. It should
Closed-cell vinyl nitrile (VN) foams were chosen as a reference ma- be noted that a constant strain rate is not attained in this experiment as
terial to compare the impact attenuation of PU foams in football hel- the velocity of the falling mass is reduced during foam compression.
mets. VN600 foam samples of 111 kg/m3 density were obtained from However, the strain rate computed using the initial impact velocity
a commercially available Riddell Revolution varsity helmet. The SEM and that calculated by differentiating the strain profile yielded only a
micrograph of VN600 in Fig. 2 shows that they have uniformly distribut- 10% difference below 60% strain. Therefore, the stress-strain response
ed irregular polyhedral closed cells with a nominal cell size of 150 μm. up to 60% strain can be considered constant for all practical purposes.
Data presented in this paper was obtained at a strain rate of 200 s−1.

3.2. Helmet testing as per NOCSAE standard test

Two helmet specimens were prepared and tested using the NOCSAE
Football Helmet Standard test [39,40] at the Dynamic Research, Inc. lo-
cated in Torrance, California. Polyurea foam material was placed inside
a Riddell Revolution (large/varsity) helmet shell from which the original
Riddell foams (vinyl nitrile-VN600 with density of 111 kg/m3) were re-
moved. The polyurea foams were of the same thickness (28 mm) and
geometry so they fit perfectly inside the shell (Fig. 3). A brand new
Riddell Revolution helmet containing VN600 foam with identical shell
material and geometry was used as a control. The two helmets, placed
on a NOCSAE medium headform (4.9 kg), were dropped on their
crown on a 12.7 mm modular elastomer programmer (MEP) pad, as
shown in Fig. 4, from heights ranging from 0.305 to 1.524 m even
though the standard requires dropping the helmet from 1.524 m only.
Lower drop height tests were motivated by the fact that there are over
200 hits during a regular NFL season that correspond to lower energies
[12]. Therefore, it is important that the helmet performs adequately at
lower energies as well. The challenge in foam design is that the foam
density is chosen to optimize the performance at the highest NOCSAE
Fig. 2. Scanning electron micrographs of the microstructure of vinyl nitrile foam VN600. drop height (1.524 m). They are not optimized for response at lower
 B.J. Ramirez, V. Gupta / Materials and Design 137 (2018) 298–304 301

Fig. 5. Stress-strain behavior of PU and VN foams under quasi-static compression.

demonstrated later, the PU foams overcome this challenge through

Fig. 3. Polyurea foams inside the Riddell Revolution helmet. their strain-rate dependent behavior.
In addition to ambient (25 °C) testing, tests corresponding to drop
heights between 0.305 and 1.219 m were conducted at cold (−15 °C)
impact heights and velocities. Consequently, the chosen foam density is and hot (50 °C) temperatures. These low and high temperature tests
higher than the optimal density and therefore higher forces are trans- were limited to the best performing PU98 foam sample under ambient
mitted through the helmet at lower drop heights and velocities. As conditions. The performance was compare against the VN600 foam
samples that were obtained from a brand new commercially available
Riddell Revolution helmet. In these tests, the foam thickness was cut
down to 22.2 mm (7/8 in.), which is often the thinnest foam liner
used in such helmets [5]. This also allowed us to further study the effect
of foam thickness on the impact performance.
The helmet performance was evaluated using several injury criteria,
including the peak linear deceleration at the center of gravity of the
headform [41], the Severity Index (SI) [42,43] and the Head Injury
Criteria (HIC) [43,44]. The latter two criteria integrate the effects of
both the impulse duration and magnitude of head g's as:

ZT
SI ¼ aðt Þ2:5 dt ð2Þ
0

8 2 32:5 9
>
< Zt 2 >
=
6 aðt Þ 7
HIC ¼ ðt 2 −t 1 Þ4 dt 5 ð3Þ
>
: ðt 2 −t 1 Þ >
;
t1
max

where a(t) is the linear deceleration at the center of gravity of the

headform in g's, T is the impact duration in seconds, and t1 and t2 are

Fig. 4. NOCSAE standard test step-up for crown testing football helmets. Fig. 6. Stress-strain behavior of PU and VN foams under dynamic compression.
302 B.J. Ramirez, V. Gupta / Materials and Design 137 (2018) 298–304

Table 1
Peak g's, SI, and HIC for different types of PU foams in a Riddell Revolution helmet shell.

Foam material 0.914 m drop height 1.524 m drop height

(impact velocity 3.7 m/s) (impact velocity 4.8 m/s)

Peak g SI HIC Peak g SI HIC

PU98 52.9 151 136 83.3 398 352

PU170 64.5 202 180 105.2 565 492
PU230 70.4 234 202 115.2 640 553

the limits of integration in seconds spanning the impact duration and

further defined in FMVSS 218. The SI should not exceed 300 for impact
velocities of 3.4 m/s and never exceed 1200 for any impact while the HIC
should be limited to 600 for moderate injury and 1000 for severe injury
[39–44].
Fig. 7. Head deceleration of NOCSAE tests using polyurea foam (PU98) compared to
VN600 foam from a drop of 1.524 m.
3.3. Federal motor vehicle safety standard (FMVSS) liner tests

(≤ 10− 1 s− 1) but significantly affects the behavior at higher rates

Foam strips of 6 mm thickness were cut from the molded block and
(101–102 s−1) where the trapped air provides stiffness and the escaping
bonded on the inside surface of 10 motorcycle helmets on top of their
layer adds to viscoelastic dissipation.
existing EPS foam liners. For these tests only PU98 foams were used as
The stress-strain behavior of VN600 is similar to that of PU98 up to
they performed the best compared to other PU foams in the football hel-
30% strain under quasi-static deformation rate. Beyond 30% strain, the
met testing. Complete motorcycle helmets with an expanded polysty-
air pressure within the closed cells of the VN foam begins to contribute
rene (EPS) liner of density 52 kg/m3 were obtained from the O'Neal
to the plateau strength. Unlike the open cell structure of the PU foams
Motocross Company. The helmets were placed on the 5 kg DOT size-C
which allow the air to move freely within the foam, the compression
dummy headform and dropped from a height of 1.93 m (6 m/s) onto
of the gas inside the closed cell requires larger stresses to continue to
a flat steel anvil. As per the FMVSS 218 test, helmets were impacted
collapse the foam. Thus, the VN600 foam has higher plateau stress
with two successive identical impacts on the same spot, spaced 90 s
than PU98 foam after 30% strain.
apart. The peak vertical deceleration of the headform was measured
after each impact for the 10 helmets with PU strips and 10 control hel-
4.2. NOCSAE helmet test results
mets with EPS only. All these tests were conducted at the ACT Labs in
Los Angeles, CA. To compare the performance of PU foams with other
Table 1 provides values of peak g's, SI, and HIC for different types of
commercially available impact resistant materials, 10 helmets made
PU foams in a Riddell Revolution helmet shell for drop heights of 0.914
using strips of D3O polymer (density of 1100 kg/m3) of the same thick-
and 1.524 m. Despite having the lowest density, the PU98 foam displays
ness (6 mm) were also tested. A brand new O'Neil helmet with the exact
the lowest peak g's, SI, and HIC values. Thus, the density of PU98 is op-
same model number and shell as used for making helmets with PU and
timized for football helmets where the design constrains the foam
D3O foam strips was used as a control. The average headform decelera-
liner to have a maximum thickness of 28 mm to limit the overall helmet
tion was measured and recorded for 10 motorcycle helmets with PU
size. In the 0.914 m drop tests, the SI and HIC values are below 300 and
foam strips, 10 helmets with D3O foam strips, and 10 control helmets
250, respectively, for all PU foams. Thus, based on these metrics PU
with only the EPS liner.
foams can possibly provide adequate protection against mTBI injuries.
The data also shows that for the expected highest impact energies in
4. Results and discussion the football game which corresponds to a maximum drop height of
1.524 m in the NOCSAE test, the peak g's are below the 125 g threshold
4.1. Compression test results for TBI injury. Since PU98 performed the best among all PU foams, it was
chosen for further testing against VN600 foam.
Figs. 5 and 6 show the stress-strain behaviors of the PU (PU98, Table 2 shows the values for peak g's, SI, and HIC parameters for
PU170, and PU230) and VN600 foams at quasi-static ( ε_ ¼ 0:04s−1 ) PU98 and VN600 foams when placed inside the Riddell Revolution hel-
and dynamic (ε_ ¼ 200s−1 ) strain rates, respectively. For PU foams, an met shell and dropped from NOCSAE heights ranging from 0.305 to
increase in the Young's modulus and plateau stress, and a decrease in 1.524 m. On an average, the PU98 foam liner represents a 22% reduction
the densification strain can be observed with increasing density and de- in peak g's, 26% reduction in SI, and a 25% reduction in HIC values across
formation rates. This behavior is due to the very high strain rate sensi- the board. Fig. 7 shows a typical deceleration vs. time curve comparing
tivity of the solid polyurea that forms their cell structure but more PU98 and VN 600 at 1.524 m drop height. The reduction in peak g's and
importantly to the size of the cell wall apertures. As discussed above, SI values, can lead to a 15–25% reduction in concussion risk based on the
the aperture size has negligible effect at low rates of deformation Pellman curve [9–11]. This is a significant improvement over current VN

Table 2
Peak g's, SI, and HIC parameters for PU98 and VN600 foams when placed inside the Riddell Revolution helmet shell and dropped from NOCSAE heights ranging from 0.305 to 1.524 m.

Drop height Impact velocity Peak g SI HIC

ft. (m) ft/s (m/s)
VN600 PU98 % reduction VN600 PU98 % reduction VN600 PU98 % reduction

1 (0.305) 6.8 (2.1) 27.6 25.0 9.4 33 25 24.2 30 23 23.3

2 (0.610) 9.8 (3.0) 50.8 39.3 22.6 107 81 24.3 94 74 21.3
3 (0.914) 12.0 (3.7) 70.6 52.9 25.1 217 151 30.4 190 136 28.4
4 (1.22) 14.0 (4.3) 84.6 65.5 22.6 344 249 27.6 304 222 27.0
5 (1.52) 15.7 (4.8) 103.0 83.3 19.1 517 398 23.0 459 352 23.3
 B.J. Ramirez, V. Gupta / Materials and Design 137 (2018) 298–304 303

Table 3
Temperature effect on the peak g's experienced by the headform's center of gravity at drop heights of 0.305, 0.610, and 0.914 m.

Drop height ft. (m) Impact velocity ft/s (m/s) Peak g – ambient (23 °C) Peak g – cold (−15 °C) Peak g – hot (50 °C)

VN600 PU98 % reduction VN600 PU98 % reduction VN600 PU98 % reduction

1 (0.305) 6.8 (2.1) 61.0 50.0 18.0 97 78 19.6 55 54 1.8

2 (0.610) 9.8 (3.0) 79.0 75.0 5.1 153 113 26.1 77 74 3.9
4 (0.914) 14.0 (4.3) 132 130 1.5 229 156 31.9 178 143 19.7

foam technology. Most importantly, PU98 foam outperforms the VN600 headform decelerations with the D3O material strip (density of
foam at all drop heights. Since large numbers of helmet impacts are ac- 1100 kg/m3) of the same thickness as the PU98 foam were 162 g ±
tually at lower heights and velocities, PU foams can perform better 2.6 g and 193.5 g ± 2.1 g for the first and second hits, respectively.
under repetitive sub-concussive and concussive hits than VN foams. Based on these tests one can conclude that polyurea-based foams signif-
The dissipation of the impact energy in the PU foams start at the mo- icantly improves the impact performance of the reference helmet by in-
lecular level through reversible viscoelastic shear deformations be- troduction of a thin layer of viscoelastic foam and furthermore they
tween the hard and the soft phases of the polyurea structure when outperform liners made from the D3O material that is presently consid-
the micrometer-sized struts, beams, and thin plates of the foam cellular ered the state of the art.
structure buckle, bend, and twist to generate very large local shear
strains during foam compression. This intrinsic molecular-level shear
dissipation process in polyurea foams is synergistically coupled to an 5. Conclusions
additional energy and momentum trapping mechanism that includes
controlled collapse of a novel cell microstructure that comprises of The potential of previously developed PU foams was demonstrated
very large cells (300 to 500 μm), each covered with a perforated mem- by integrating them into football and motorcycle helmet shells and test-
brane with apertures 20 to 70 μm in diameter (Fig. 1). At very low strain ing them using the NOCSAE and FMVSS 218 DOT helmet standards, re-
rates, air escapes easily from large cells through perforations, allowing spectively. In the football helmet tests, the pre-exiting foam liners from
easy collapse of the microstructure to limit the peak force below the in- a Riddell Revolution helmet were replaced by the PU foams of the same
jury threshold levels. At higher velocity impacts or higher strain rates, geometry, whereas for motorcycle helmet testing, 6 mm thick strips of
air starts to provide viscoelastic damping and substantial strengthening PU98 foams were placed as additional liner on top of the existing EPS
to the cell walls as its rate of escape through tiny perforations cannot liner. The performance of PU foams was compared with VN600 foams
catch up with the higher rate of loading. Consequently, the same low that are presently utilized in the Riddell Revolution football helmets.
density material can now support a much higher dynamic load without For motorcycle helmets, strips of D3O material of density 1100 kg/m3
the need to increase the density of the material. Every other commer- were used as the reference. On an average, use of PU98 foam resulted
cially available foam today requires a substantial increase in the density in a reduction of 22% in peak g's, 26% in SI, and 25% in HIC value when
(or reduction in cell size with higher thickness of cell walls) to support compared to helmets with VN600 foams. The thickness of the foams
the enhanced dynamic loads that arise at higher strain rates. Thus, this in these tests was 28 mm and all drops were conducted under ambient
microstructure makes the polyurea foams “strain rate dependent,” conditions from heights ranging from 0.305 to 1.524 m. These tests
with the performance becoming better at increasing strain rates and were repeated with reduced liner thickness of 22.2 mm to study the ef-
thus, it is able to outperform VN foams at every impact velocity. fect of thickness as well as the temperature on impact performance. In
Table 3 shows the temperature effect on the peak g's experienced by these tests, PU98 foam reduced the peak g's by as much as 18% under
the headform's center of gravity at drop heights of 0.305, 0.610, and ambient temperature (23 °C), 25.8% at − 15 °C, and 19.7% at 50 °C
0.914 m. For these temperature tests, the foam thickness was reduced when compared to VN600 helmets at drop heights of 0.305 to
to 22.2 mm (7/8 in.) to examine the effect of foam thickness as well. 0.914 m. These results are noteworthy as the PU98 foam is also lighter
At ambient temperatures (23 °C) the PU98 foam reduced the peak g's by about 12% compared with VN foams.
by as much as 18% when compared to VN600. At cold temperatures The motorcycle helmet tests demonstrated the success of using an
(−15 °C) PU98 foam reduced the peak g's by an average of 26% when additional layer of PU foam liner on top of the existing EPS liner in re-
compared to VN600 across all drop heights. Furthermore, at hot tem- ducing the peak g's by 17% in FMVSS 218 DOT tests. These tests also
peratures (50 °C), PU98 reduced the peak g's by as much as 19.7% demonstrated that the PU foams perform better than the state of the
when compared to VN600. This is a significant improvement in the im- art D3O material in such a bilayer design.
pact attenuation of VN foams at low and hot temperatures. Viscoelastic
foam systems are susceptible to temperature variations, becoming
harder at cold temperatures and softer at high temperatures. This affects Acknowledgements
their energy dissipation performance. Since the glass transition temper-
ature of polyurea is −50 °C it is able to retain its superior impact prop- This work was supported by the Office of Naval Research under grant
erties over a range of temperatures (−15 °C to 50 °C). As expected, the number N0008 for which we are grateful to Dr. Roshdy Barsoum of the
reduction in padding thickness (from 28 mm to 22.2 mm) increased the agency.
peak g's for both VN and PU foams at ambient condition. This can be
seen by comparing the values of peak g's in Tables 2 and 3. References
[1] L.J. Gibson, M.F. Ashby, Cellular Solids: Structure and Properties, Cambridge Univer-
4.3. FMVSS test results sity Press, 1999.
[2] N. Mills, Polymer Foams Handbook: Engineering and Biomechanics Applications
and Design Guide, Butterworth-Heinemann, 2007.
The use of polyurea foam strips resulted in headform deceleration of
[3] B.J. Ramirez, O.T. Kingstedt, R. Crum, C. Gamez, V. Gupta, Tailoring the rate-
146.7 g ± 0.6 g when compared with 176.0 g ± 1.0 g recorded for the sensitivity of low density polyurea foams through cell wall aperture size, J. Appl.
reference helmet. When the same helmet was impacted for a second Phys. 121 (2017) 225107.
time on the same spot, headform decelerations of 177.3 g ± 1.5 g [4] V. Gupta, G. Youssef, Orientation-dependent impact behavior of polymer/EVA bilay-
er specimens at long wavelengths, Exp. Mech. 54 (2014) 1133–1137.
were recorded. These are well below 198 g ± 3.6 g that were recorded [5] A. Bartsch, E. Benzel, V. Miele, V. Prakash, Impact test comparisons of 20th and 21st
with the reference helmets under identical conditions. The measured century American football helmets, J. Neurosurg. 116 (2012) 222–233.
304 B.J. Ramirez, V. Gupta / Materials and Design 137 (2018) 298–304

[6] J.T. Gwin, J.J. Chu, S.G. Diamond, P.D. Halstead, J.J. Crisco, R.M. Greenwald, An inves- [27] K. Holzworth, Z. Jia, A.V. Amirkhizi, J. Qiao, S. Nemat-Nasser, Effect of isocyanate con-
tigation of the NOCSAE linear impactor test method based on in vivo measures of tent on thermal and mechanical properties of polyurea, Polymer 54 (2013)
head impact acceleration in American football, J. Biomech. Eng. 132 (2010) 011006. 3079–3085.
[7] F. Hernandez, P.B. Shull, D.B. Camarillo, Evaluation of a laboratory model of human [28] Z. Jia, A.V. Amirkhizi, K. Holzworth, S. Nemat-Nasser, The effect of stoichiometric
head impact biomechanics, J. Biomech. 48 (2015) 3469–3477. ratio on viscoelastic properties of polyurea, in: B. Antoun (Ed.), Challenges in Me-
[8] W.C. Moss, M.J. King, Impact Response of US Army and National Football League Hel- chanics of Time-Dependent Materials and Processes in Conventional and Multifunc-
met Pad Systems, No. LLNL-SR-464951, Lawrence Livermore National Lab. CA, 2011. tional Materials, Conference Proceedings of the Society for Experimental Mechanics
[9] E.J. Pellman, D.C. Viano, A.M. Tucker, I.R. Casson, A.B. Valadka, J.C. Maroon, J.E. Bailes, SeriesVol. 2, Springer, Cham 2014, pp. 17–20.
D.F. Kelly, M.L. Levy, Concussion in professional football: location and direction of [29] M.R. Amini, J.B. Isaacs, S. Nemat-Nasser, Experimental investigation of response of
helmet impacts - part 2, Neurosurgery 53 (2003) 1328–1341. monolithic and bilayer plates to impulsive loads, Int. J. Impact Eng. 37 (2010) 82–89.
[10] E.J. Pellman, D.C. Viano, C. Withnall, N. Shewchenko, C.A. Bir, P.D. Halstead, Concus- [30] M.R. Amini, J.B. Isaacs, S. Nemat-Nasser, Investigation of effect of polyurea on re-
sion in professional football: helmet testing to assess impact performance - part 11, sponse of steel plates to impulsive loads in direct pressure-pulse experiments,
Neurosurgery 58 (2006) 78–95. Mech. Mater. 42 (2010) 628–639.
[11] E.J. Pellman, D.C. Viano, A.M. Tucker, I.R. Casson, J.F. Waeckerle, J.C. Maroon, M.R. [31] A. Jain, V. Gupta, Construction and characterization of stainless steel/polyurea/E-
Lovell, M.W. Collins, D.F. Kelly, A.B. Valadka, R.C. Cantu, J.E. Bailes, M.L. Levy, Concus- glass composite joints, Mech. Mater. 46 (2012) 16–22.
sion in professional football: reconstruction of game impacts and injuries, Neurosur- [32] A. Jain, G. Youssef, V. Gupta, Dynamic tensile strength of polyurea-bonded steel/E-
gery 53 (2003) 799–814. glass composite joints, J. Adhes. Sci. Technol. 27 (2013) 403–412.
[12] S. Rowson, S.M. Duma, Development of the STAR evaluation system for football hel- [33] H. Kim, J. Citron, G. Youssef, A. Navarro, V. Gupta, Dynamic fracture energy of
mets: integrating player head impact exposure and risk of concussion, Ann. Biomed. polyurea-bonded steel/E-glass composite joints, Mech. Mater. 45 (2012) 10–19.
Eng. 39 (2011) 2130–2140. [34] C.M. Roland, J.N. Twigg, Y. Vu, P.H. Mott, High strain rate mechanical behavior of
[13] K.C. Rusch, Energy-absorbing characteristics of foamed polymers, J. Appl. Polym. Sci. polyurea, Polymer 48 (2007) 574–578.
14 (1970) 1433–1447. [35] J. Shim, D. Mohr, Using split Hopkinson pressure bars to perform large strain com-
[14] K.C. Rusch, Load-compression behavior of flexible foams, J. Appl. Polym. Sci. 13 pression tests on polyurea at low, intermediate and high strain rates, Int. J. Impact
(1969) 2297–2311. Eng. 36 (2009) 1116–1127.
[15] A. Nagy, W.L. Ko, U.S. Lindholm, Mechanical behavior of foamed materials under dy- [36] T. Jiao, R.J. Clifton, S.E. Grunschel, Pressure-sensitivity and constitutive modeling of
namic compression, J. Cell. Plast. 10 (1974) 127–134. an elastomer at high strain rates, in: M. Elert, M.D. Furnish, W.W. Anderson, W.G.
[16] K.Y. Jeong, S.S. Cheon, M.B. Munshi, A constitutive model for polyurethane foam Proud, W.T. Butler (Eds.), Shock Compression of Condensed Matter 2009: Proceed-
with strain rate sensitivity, J. Mech. Sci. Technol. 26 (2012) 2033–2038. ings of the American Physical Society Topical Group on Shock Compression of Con-
[17] S. Ouellet, D. Cronin, M. Worswick, Compressive response of polymeric foams under densed Matter, AIP Publishing 1195 2009, pp. 1229–1232.
quasi-static, medium and high strain rate conditions, Polym. Test. 25 (2006) [37] T. Jiao, R.J. Clifton, S.E. Grunschel, Pressure-sensitivity and tensile strength of an elas-
731–743. tomer at high strain rates, in: M. Elert, M.D. Furnish, R. Chau, N. Holmes, J. Nguyen
[18] J. Zhang, Z. Lin, Z. Wong, N. Kikuchi, V.C. Li, A.F. Yee, G.S. Nusholtz, Constitutive (Eds.),Shock Compression of Condensed Matter 2007, AIP Conference Proceedings,
modeling and material characterization of polymeric foams, J. Eng. Mater. Technol. 955, 2007, pp. 707–710.
119 (1997) 284–291. [38] G. Youssef, V. Gupta, Dynamic tensile strength of polyurea, J. Mater. Res. 27 (2012)
[19] M.R. Shariatmadari, R. English, G. Rothwell, Effects of temperature on the material 494–499.
characteristics of midsole and insole footwear foams subject to quasi-static com- [39] National Operating Committee on Standards for Athletic Equipment, Standard Per-
pressive and shear force loading, Mater. Des. 37 (2012) 543–559. formance Specification for Newly Manufactured Football Helmets: NOCSAE Stan-
[20] J. Yi, M.C. Boyce, G.F. Lee, E. Balizer, Large deformation rate-dependent stress-strain dard 002-98m03, 1998.
behavior of polyurea and polyurethanes, Polymer 47 (2006) 319–329. [40] National Operating Committee on Standards for Athletic Equipment, Standard Drop
[21] A.V. Amirkhizi, J. Isaacs, J. McGee, S. Nemat-Nasser, An experimentally-based visco- Test Method and Equipment Used in Evaluating the Performance Characteristics of
elastic constitutive model for polyurea, including pressure and temperature effects, Protective Headgear: NOCSAE Standard 001-04m05, 2004.
Philos. Mag. 86 (2006) 5847–5866. [41] H.R. Lissner, M. Lebow, F.G. Evans, Experimental studies on the relation between ac-
[22] S.S. Sarva, S. Deschanel, M.C. Boyce, W. Chen, Stress-strain behavior of a polyurea celeration and intracranial pressure changes in man, Surg. Gynecol. Obstet. 111
and a polyurethane from low to high strain rates, Polymer 48 (2007) 2208–2213. (1960) 329–338.
[23] Y. He, D. Xie, X. Zhang, The structure, microphase-separated morphology, and prop- [42] C.W. Gadd, Use of a weighted-impulse criterion for estimated injury hazard, SAE
erty of polyurethanes and polyureas, J. Mater. Sci. 49 (2014) 7339–7352. Technical Paper, 660793, 1966, pp. 164–174.
[24] M. Grujicic, B. Pandurangan, T. He, B.A. Cheeseman, C.F. Yen, C.L. Randow, Computa- [43] J. Versace, A review of the severity index, SAE Technical Paper, 710881, 1971,
tional investigation of impact energy absorption capability of polyurea coatings via pp. 771–796.
deformation-induced glass transition, Mater. Sci. Eng. A 527 (2010) 7741–7751. [44] W.N. Hardy, T.B. Khalil, A.I. King, Literature review of head injury biomechanics, Int.
[25] D. Fragiadakis, R. Gamache, R.B. Bogoslovov, C.M. Roland, Segmental dynamics of J. Impact Eng. 15 (1994) 561–586.
polyurea: effect of stoichiometry, Polymer 51 (2010) 178–184.
[26] T. Choi, D. Fragiadakis, C.M. Roland, J. Runt, Microstructure and segmental dynamics
of polyurea under uniaxial deformation, Macromolecules 45 (2012) 3581–3589.