Mechanical behavior of thru air bonded nonwonven structures

http trj sagepub com Textile Research Journal DOI 10 11770040517507081313 2007; 77; 417 Textile Research Journal Amit Rawal, Stepan Lomov, Thanh Ngo, Ignaas Verpoest and Jozef Vankerrebrouck Mechanical Behavior of Thru air Bonded Nonwoven Structures http trj sagepub comcgicontentabstract776417 The online version of this article can be found at Published by http www sagepublications com can be found at Textile Research Journal Additional services and information for http trj sagepu.

http://trj.sagepub.com Textile Research Journal

DOI: 10.1177/0040517507081313

2007; 77; 417

Textile Research Journal

Amit Rawal, Stepan Lomov, Thanh Ngo, Ignaas Verpoest and Jozef Vankerrebrouck

Mechanical Behavior of Thru-air Bonded Nonwoven Structures

http://trj.sagepub.com/cgi/content/abstract/77/6/417

The online version of this article can be found at:

Published by:

http://www.sagepublications.com

can be found at:

Textile Research Journal

Additional services and information for

http://trj.sagepub.com/cgi/alerts

Email Alerts:

http://trj.sagepub.com/subscriptions

Subscriptions:

http://www.sagepub.com/journalsReprints.nav

Reprints:

http://www.sagepub.com/journalsPermissions.nav

Permissions:

http://trj.sagepub.com/cgi/content/refs/77/6/417

SAGE Journals Online and HighWire Press platforms):

(this article cites 25 articles hosted on the

Citations

Mechanical Behavior of Thru-air Bonded Nonwoven Structures 1

Amit Rawal2, Stepan Lomov, Thanh Ngo and Ignaas Verpoest

Department of Metallurgy and Material Engineering, Katholieke Universiteit Leuven, Kasteelpark Arenberg,

44, B-3001 Leuven, Belgium Jozef Vankerrebrouck Libeltex bvba, Marialoopsteenweg 51, 8760 Meulebeke, Belgium

Nonwoven fabrics are defined as a sheet, web or batt of

directionally or randomly oriented fibers/filaments, bonded

either by friction, and/or cohesion and/or adhesion In

gen-eral, nonwoven fabric formation is a two-step process i.e

web formation (laying up the fibers with certain orientation

characteristics) and bonding the fibers by mechanical,

ther-mal or chemical means This two-step process forms the

basis of classification of nonwoven structures i.e carded, air

laid, spunbonded, meltblown, needlepunched,

hydroentan-gled, adhesive bonded, thermal bonded, stitch bonded, etc

Thermal bonded structures are amongst the most widely

used nonwovens, with applications ranging from baby

dia-pers to high performance geotextiles These structures

undergo various modes of deformation during their

end-use performance For example, geotextiles when placed

under the soil exhibit high levels of tensile and compressive modes of deformation.1 Characterization of tension, shear, compressional and bending resistance is needed for accu-rate prediction of fabric draping during the formation of a three-dimensional (3D) shaped composite part for many automotive applications Therefore, the mechanical behav-ior of these structures is a topic of fundamental importance

to achieve the desired level of performance

Abstract Thermally bonded structures are amongst

the most widely used nonwoven structures, with

applications ranging from baby diapers to high

per-formance geotextiles These structures undergo

various modes of deformation during their

per-formance In this study, the mechanical properties

namely, tension, shear, bending and compression

of thermal bonded polyester fabrics, have been

investigated, including directional dependence of

the tensile and bending tests, revealing anisotropic

characteristics In addition, the fiber orientation

distribution was determined using

two-dimen-sional (2D) image analysis of fabric cross-sections,

a technique that is generally used for measuring fiber

orientation in short-fiber reinforced composites

The ambiguities, errors and corrections employed in

the measurements of fiber orientation are

dis-cussed The initial tensile response of thermal

bonded nonwovens has been predicted and

com-pared with the experimental stress-strain curves

obtained in various test directions Bending and

compression measurements were done on KES-F

and shear measurements were carried out using

picture frame test up to a shear angle of 45°

Key words anisotropy, fiber orientation,

mecha-nical properties, modeling, nonwoven, thermal

bonded, thru-air bonded

1 This paper was presented at the Nonwovens Research Academy conference, Leeds, UK, 29–30 March 2007.

2 Corresponding author Current address: The University of Bolton, Deane Road, Bolton, BL3 5AB, UK Tel.: +44 1204 903 429; fax: +44 1204 399 074; e-mail: amitrawal77@hotmail.com

The physical and mechanical properties of thermal

bonded structures are highly dependent upon fiber

proper-ties, type of web structure, binder fiber characteristics, the

bonding (temperature and velocity of air) and cooling

con-ditions [1–3] Various studies have been reported

formulat-ing the relationship between fiber and fabric properties

Dharmadhikary et al [4] have shown that thermal bonded

structures produced from high modulus fibers result in

lower tensile strength compared to fibers having low

mod-ulus A recent review of thermal bonded nonwovens

indi-cated that the maximum level of bonding is achieved in a

random structure, as the maximum number of

fiber-to-fiber contacts is realized [5] The fiber-to-fiber orientation dictates

the anisotropy of various in-plane mechanical properties,

such as tensile modulus, maximum stress in tension,

elon-gation at maximum stress, shear modulus and bending

rigidity [3, 6]

Backer and Petterson [7] carried out an extensive study

to determine the anisotropy of the tensile moduli and

strength in a thermal bonded nonwoven structure based

upon fiber network theory Subsequently, the theory was

extended by incorporating the effect of fiber curl [8]

Win-chester and Whitwell [9] measured the thermal bonded

web responses i.e elastic, handle and rupture properties to

variations in base fiber, binder, web formation and bonding

conditions and developed empirical models for these

responses van Wyk [10] in his pioneering work calculated

the number of fiber-to-fiber contacts, which is

fundamen-tally important in determining the mechanical properties

of fibrous assemblies Subsequently, the theory was

expanded by predicting the number of fiber contacts based

upon any fiber orientation distribution, length and

arbi-trary shape of the fiber cross-section [11, 12] The initial

tensile response, compression and shear deformations of

“general” fibrous assemblies were predicted based upon

these theories [13–17] Recently advanced variants of fiber

network theory based on direct Monte-Carlo modeling of

the fiber placement were also proposed [18] However,

these theories lack extensive validation, which is partly

caused by scarceness of the experimental data on

anisot-ropy of tensile properties in relation with the fiber

orienta-tion distribuorienta-tion

The present paper aimed at supplementing these studies

with an investigation of mechanical properties namely,

ten-sion, bending, shear and compression of thermal bonded

nonwoven structures, in addition to the measurement of

fiber orientation distribution in the fabric (the thermal

bonded structure here refers to the thru-air bonded

struc-ture that has excluded the effect of calendaring) We also

propose a simple micromechanical model for predicting the

initial tensile response under uniaxial loading in various test

directions, based upon Pan et al [16] analysis, by calculating

the number of bond points in a unit volume of fibrous

assemblies

Production of Thru-air Bonded Nonwoven Structures

The thru-air bonded nonwoven structures were produced

by initially opening and blending the homofil (melting perature: 250°C) and bicomponent (sheath melting tem-perature: 110°C) polyester fibers Subsequently, the fibers were orientated in the machine direction by carding How-ever, the mechanism of randomizing fiber arrangement in

a thermal bonded nonwoven structure was similar to the process shown by Lin et al [19] In carding, two sets of ran-dom rollers (R1 and R2) were placed between cylinder (C) and doffer (D), as shown in Figure 1a The objective of these random rollers was to minimize the anisotropy of thermal bonded nonwoven structures In addition, the maximum level of bonding was achieved in a random struc-ture, as the maximum number of fiber-to-fiber contacts was realized [5] The ratio of speeds between random roll-ers (R1 and R2) and cylinder (C) ranged between 0.5 and 0.7 This resulted in overfeeding of the fibers between the cylinder (C) and rollers (R1 and R2) and the preferential orientation of the fibers in the machine direction was mini-mized In other words, the lower the speed ratio between the rollers (R1 and R2) and cylinder (C), the higher the disturbance or randomization caused between the fibers Finally, the web was passed through a hot air oven in order

to melt the sheath of bicomponent fibers for the formation

of required bonds in the nonwoven structure Figure 1b shows the flow chart for the production of thru-air bonded nonwoven structures

Experimental

The reported work is based upon two thermal bonded non-wovens, produced by Libeltex and labeled here as TB1 and TB2 (Figure 2) These nonwoven structures were produced using homofil (melting temperature: 250°C) and bicompo-nent (sheath melting temperature: 110°C) polyester fibers

in equal proportions by weight The linear densities of pol-yester fibers used in the production of TB1 were 2.2 dtex (bicomponent) and 3.3 dtex (homofil), whereas in TB2 were 4.4 dtex (bicomponent) and 12 dtex (homofil) The stress-strain curves of these fibers were determined using micro-tensile tester at a crosshead speed and gauge length

of 5 mm/min and 10 mm, respectively (see Figure 3) The mass per unit area and thickness of the fabrics TB1 and TB2 are shown in Table 1

Measurement of Fiber Orientation Angle

The fiber orientation angle was determined by a method that is generally used for measuring the fiber orientation

angle in short-fiber reinforced composites [20–22] The

fabric was impregnated in an epoxy resin (Araldite 564,

hardener HY 2944) and the whole assembly consisting of

matrix and fabric was cured for two hours in an oven at

120°C The specimens were then sectioned in two planes

(XZ and YZ, X being the machine direction of the fabric,

corresponding to the 0° direction), as shown in Figure 4

Subsequently, the grinding and polishing were used for

smoothening the surface of the specimens Typically, 700

fibers were sectioned for obtaining the fiber orientation

distribution The images of the specimens were captured

using an optical microscope and analyzed using LEICA Q

Win software In the 2D image analysis, a cross-section of a

fiber was assumed to be circular Sectioned by a plane, it

was registered as an ellipse with major and minor axes a

and b, respectively The image analysis software provided

the values of minor and major axes along with an in-plane

fiber orientation angle ϕ (here prime refers to the local

co-ordinate system associated with the sectioning plane)

The out of plane fiber orientation angle (θ ) was calculated

as shown below

θ = arccos(b/a) (1) Figure 5 shows the orientation of an elliptical cross-section

of the fiber expressed in polar coordinates by the two angles (θ , ϕ ) or in Cartesian coordinates (X Y Z ) by the

components of a fiber direction unit vector p (p x , p y , p z) The measurement of fiber orientation angles depended upon the direction of the sectioning plane The relation-ship between θ , ϕ and the global fiber orientation angles

θ, ϕ is shown in Table 2 The different fibers in the blend were segregated during the data processing by their

diam-eters i.e values of b.

However, there are certain errors and ambiguities involved in measuring the fiber orientation using 2D image analysis method [20–22] Firstly, there are two possible val-ues for fiber in the in-plane orientation angle (ϕ ) as the

Figure 1 Production of thermal bonded nonwovens (a) mechanism of random fiber arrangement: cylinder (C), doffer (D) and random rollers (R1 and R2); (b) flow chart for production of thru-air bonded nonwoven structures

′

′

′

′ ′ ′ ′ ′

′ ′ ′

′ ′

′

fibers with orientations ϕ and ϕ + 180 have identical

cross-sections Secondly, the probability of finding a fiber

with defined orientation (θ , ϕ ) such that it has an

ellipti-cal cross-section on the sectioning plane needs to be

deter-mined Thirdly, small errors in the measurement of elliptical

axes for the fibers oriented nearly perpendicular (θ 0)

to the sectioning plane can yield large errors in the

meas-ured value of θ Hence, the proportional number of fibers oriented perpendicular to the sectioning plane measured

by 2D image analysis is expected to be minimal In addi-tion, the effect of fiber crimp has been neglected during the measurement of fiber orientation angles

The first ambiguity of the fiber symmetry was overcome by

assuming the orientation distribution to be symmetrical,

Figure 2 Fabrics: left, TB1; right, TB2 Arrows show tensile test directions 0° corresponds to the machine direction Insets: SEM images of the bonds in the fabrics

Figure 3 Fiber stress-strain curves Error bars indicate standard errors

in ten tests

′ ′

′ ′

′ ≈

′

and the out of plane orientation angle was randomized by

using the following equation:

θ1 = (sgn (RAND (0,1) – 0.5)⋅ θ +180) (2) where θ1 is the corrected out of plane fiber orientation angle, RAND is a uniform random number generator in the interval (0,1) and sgn function is defined as follows:

Furthermore, the “probability of intersection” between

the fiber and the sectioning plane was determined by divid-ing the fiber orientation distribution with cosine of the

sec-tioning plane angle, in order to overcome the second ambiguity

[21, 22] for θ 90° For example, be the fiber

orien-tation frequency value for a bin (i, j) of the fiber

orienta-tion histogram, obtained from the image analysis while

Table 1 Physical properties of thru-air bonded nonwoven fabrics

Fabric sample IDs Mass per unit area (g/m 2 ) Nominal mass per unit area (g/m 2 ) Thickness (mm)*

TB1

TB2

31.09 ± 0.082 28.46 ± 0.031

30 30

0.44 ± 0.0012 0.43 ± 0.0038

* Thickness was measured at 0.1 PSI or 7.03 g/cm 2 (ASTM D 5729-1995).

Figure 4 Sectioning of fabric samples

′

x

( ) sgn

1 x, >0

0 x, = 0

1 x, <1 –











=

Figure 5 Fiber orientation measurements (a) 2D image analysis software; (b) scheme for the calculation of fiber orienta-tion angles

′ ≈ νij

XZ

sectioning the sample in the XZ plane Thus, the corrected

distribution was computed as follows:

(3)

where are local out of plane and in-plane fiber

ori-entation angles, respectively

The corrected global fiber orientation angle

distribu-tion, (θi, ϕi), could be calculated as shown below

(4)

where N XZ and N YZ are the total number of fibers sec-tioned in XZ and YZ planes, respectively

Finally, the error of finding minimum number of fibers

elucidated by fitting a normal or Gaussian distribution to

the measured values of in plane fiber orientation angles The fitted histograms of normal distribution consisted of maximum proportional of fibers oriented at 0° (machine direction), as expected from the process (also described previously) Figure 6 shows the histograms and normal fit-ted curves of relative frequency of fibers that have been used in the production of thermal bonded nonwovens TB1 and TB2

Table 2 Calculation of out of plane, θ, and in-plane, ϕ, fiber orientation angles

Section

plane Local coordinates on the section plane Vector components Fiber orientation angles

y’ = y z’ = z

p x= sin θ’cosθ’

p y= sin θ’sinφ’

p z= cos θ’

φ = φ’

θ = θ’

y’ = –z z’ = y

p x = p x‘= sin θ’cosφ’

p y = p z‘= cos θ’

p z = –p y‘= –sin θ’sinφ’

φ = arctan = arctan ’ ’ ’

θ = arccosp z= arccos(sin θ’ sinφ’)

y’ = z

z’ = x

p x = p z= cos θ’

p y = p x‘= sin θ’cosφ’

p z = p y‘= sin θ’sinφ’

φ = arctan = arctan ’ ’ ’

θ = arccosp z= arccos(sin θ’ sinφ’)

p y

p x

- cos θ

θ cos θ sin

-p y

p x

- sin θ cos θ

θ cos

-νij corrected XZ νXZ

θi′ ϕ, j′ ( )

θi′ cos

-=

θi′ ϕ, j′

νij global

νij

global

θi,ϕj

XZ

N XZ

N YZ

× +

N XZ+N YZ

-=

′ ≈

Tensile Characteristics of Thermal Bonded

Nonwoven Structures

The fabric strips of 20 × 15 cm were tested on an Instron

tensile testing machine under uniaxial loading at a strain

rate of 10 mm/min in a “grab” test with the width and gauge

length of 50 and 150 mm, respectively These fabrics were

tested in various test directions, 0, 22.5, 45, 67.5 and 90°,

with respect to the machine direction Tests were repeated

for ten specimens of nonwoven fabric in each test direction

Poisson’s ratio was also determined by measuring the

con-traction in the center of the specimen relative to the strain

in the test direction

The stress-strain curves of thermal bonded structures

TB1 and TB2 are shown in Figure 7 The slope of the

curves decreased with the angle of test, showing high

degree of anisotropy of the tensile resistance that further

reflected the anisotropy of the fiber orientation, also

evi-dent by the visual observation and confirmed by the

meas-urements of the orientation distribution, as shown in

Figures 2 and 6, respectively The ratio of tensile strength

in the machine direction to the cross-machine direction for

thermal bonded structures TB1 and TB2 was found to be

3.3 and 3.5, respectively A similar order of magnitude of

the tensile anisotropy has been observed by Lin et al [19]

The difference in the moduli was much more pronounced,

reaching the ratio of 20–50 between the secant moduli in 0°

and 90° directions, as shown in Figure 8

It should be noted that the thermal bonded structure TB1 had higher tensile strength in comparison to TB2 in all the test directions, although the fibers (4.4 and 12 dtex) used in the production of TB2 had high tensile strength in comparison to the fibers (2.2 and 3.3 dtex) used in TB1, as shown in Figure 3 In addition, both fabrics had the same nominal area density (Table 1) It is well-known in thermal bonded structures that fibers with higher initial modulus and lower elongation can produce fabric of similar or even lower strength as that of fibers with lower modulus and

higher elongation due to better “load sharing” by high

elon-gation fibers [1, 4] Moreover, tensile deformation along the preferential i.e machine direction requires significant fiber extension [3] This can also be seen by comparing the secant modulus of two thermal bonded structures obtained

at 3.5 % strain, as shown in Figure 8 Here, the slope of the curves steeply decreased between 0 and 22.5°, showing that high proportions of fibers were oriented in the machine direction (0° indicates machine direction)

Shear Properties of Thermal Bonded Nonwoven Structures

Shear measurements were carried out using picture frame that was mounted on Instron tensile testing machine with a

1 kN load cell Three shear cycles were performed for each sample consisting of five specimens The picture frame is Figure 6 In-plane fiber orientation distribution (a) 2.2 dtex, TB1; (b) 3.3 dtex, TB1; (c) 4.4 dtex, TB2; (d) 12 dtex, TB2

generally employed for measuring the shear resistance at

higher shear angles [23] Figure 9a shows the first cycle of

shear that was significantly different from second and third

cycles A similar effect has been observed for the first cycle

of shear by Lomov et al [23] for heavy woven fabrics due

to the initial misalignments and pretensions caused by mounting the sample onto the frame Hence, the first shear cycle described the sample installation on the frame rather than the actual material behavior The subsequent cycles showed similar behavior as that of woven fabrics displaying the consistent behavior Initial high shear stiffness (up to shear angles of 3°) represented bond resistance that was sufficient to prevent rotation of the fibers in the thermal bonded nonwoven structure Later, bond resistance was overcome and the shear resistance was primarily domi-nated by the rotation of structural elements of the fabric Scanning electron microscopy (SEM) images, as shown in Figure 2, revealed spaces between the fibers to accommo-date the rotation without significant restriction from the neighboring fibers This produced the region of low shear stiffness (up to ~12°) that led to an increase in the local densities of fibrous assemblies An increase in the density

of the fibrous assembly resulted in an increase in the shear stiffness

Figure 9b shows a comparison between shear force per unit width and shear angle of fabrics TB1 and TB2 for third cycle of loading TB1 required more shear force than non-woven TB2 at higher shear angles, although both fabrics had similar area density and thickness This may be attrib-uted to the fact that finer fibers (2.2 and 3.3 dtex) were used

in the production of TB1, whereas coarser fibers (4.4 and 12 dtex) were used in TB2 The proportion of finer fibers was higher in TB1, which led to a higher number of thermal bonds in the structure and lower pore sizes, that resulted in jamming of the fibers at higher shear angles The more

“cloudy” structure of TB1 can be clearly seen in the images

of the fabrics (Figure 2) The (un)evenness of the shear strain field was also studied using optical full field strain measurement technique (ARAMIS® system) [23] Figure 9b shows an example of the registered shear strain field The scatter of the local shear angle of nonwoven TB1 was high i.e 16° to 32° at a frame shear angle of 30°

Compressional Properties of Thermal Bonded Nonwoven Structures

The compressional properties were determined using

KES-F B3 (Compression Tester) During the compression meas-urements, the pressure was increased gradually and reached

up to 50 g/cm2 Three cycles were repeated for each sample

at three distinct positions Lee and Lee [24] have reported that compressional forces are generally transmitted through the contact points These contact points can be classified in two categories, “slipping” or “nonslipping” contacts [13], and the latter is reported to be stronger [25] Therefore, in

a thermal bonded structure the “thermal bonds” are the nonslipping contacts that can transmit the compressional load In addition, bending, torsional, tensile and frictional characteristics of the constituent fibers are also key factors

Figure 7 Stress-strain curves of thermal bonded

nonwo-vens (a) TB1; (b) TB2 Error bars indicate the standard

deviation in ten tests

Figure 8 Secant moduli of thermal bonded structures

TB1 and TB2 at 3.5 % strain in various test directions

influencing the compressional properties of the fibrous

assemblies Figure 10 shows the relationship between the

pressure and thickness of fabrics TB1 and TB2 for the first

cycle of loading

It was found that the three cycles produced similar results in all the three positions At lower pressure (up to

5 g/cm2), both thermal bonded structures followed the same curve However at higher pressure, the thermal bonded structure TB2 was slightly more compressible than TB1 This may be attributed to the fact that larger and more compressible bonds were formed in TB2 as coarser fibers were used in comparison to TB1

Bending Properties of Thermal Bonded Nonwoven Structures

The bending properties were determined using KES-F B2 (Bending Tester) The hysteresis was measured by bending the fabrics between the curvature of –2.5 and 2.5 cm–1 The three cycles were repeated for each sample in the machine (0°) and cross-machine (90°) directions

Figure 11 and Table 3 show the comparison between the bending diagrams and parameters of the bending resistance of fabrics TB1 and TB2 in the machine and cross-machine directions Several observations can be made:

Figure 9 Shear tests (a) typical

dia-grams for the three shear cycles

(TB1); (b) comparison of the

aver-aged shear diagrams (third cycle)

for TB1 and TB2 Insets: nonwoven

fabric TB1 on the picture frame; von

Mises strain field at a frame shear

angle of 30° showing minimum

and maximum shear angles

Figure 10 Relationship between the pressure and

thick-ness of thermal bonded nonwovens TB1 and TB2 for first

cycle of loading