ORIGINAL PAPER/PEER-REVIEWED

Influence of Fiber Type On

Fiberweb Properties in
High-Speed Carding
By N. B. Doguc, Abdelfattah M. Seyam, William Oxenham, Nonwovens Cooperative Research Center,
North Carolina State University, Raleigh, NC

Abstract changing the production rates and products. Although

The performance of four different fiber types processed at regarded as a conventional fiberweb forming system, the
two different carding speeds of 85 and 120 m/min, under carding process has undergone significant evolution to meet
fixed operating conditions, was assessed. The totally novel today’s high production needs for nonwoven manufacturing
approach of fixing the fiber diameter was used in the experi- lines. Meltblowing and spunbonding can, however, achieve
ments. To serve this purpose, we designed four different production rates up to 600 m/min [1,4] and therefore, in order
fibers. In addition, we fixed the number of fibers per unit area to remain competitive, the productivity of nonwoven cards
of the carded web so as to create constant carding conditions needs to be increased. Recently machine manufacturers were
for all experimental fibers. The fixed number of fibers per unit successful in producing high-speed carding machines
area was achieved by carding the same number of fibers per through the use of better machine design (materials and con-
unit time for a given speed. We took samples from different figuration), feed uniformity control systems, and novel doff-
parts of the card and analyzed them in order to enumerate the ing techniques. The fact that some fibers cannot be processed
effect of fiber type on cardability. Fiberweb uniformity was at high speeds as compared to others in carding represents a
regarded as a key parameter for assessing the cardability. serious challenge for carding machine producers, fiber finish
Several other web and fiber parameters were measured in manufacturers, and nonwoven fabric producers.
order to augment the understanding of the role of fiber type In relation to the above mentioned concern, discussions
in high-speed carding. In general, data revealed that fiberweb were held with industry members of Nonwovens
uniformity did not significantly differ among fiber types at Cooperative Research Center (NCRC) at NC State University
the two carding speeds 85 m/min and 120 m/min. and it was indicated that the general experience was that
polypropylene (PP) fibers could be carded at much higher
Keywords: High-speed Carding, Fiberweb Uniformity, speed than polyester (PET) fibers. Others, however, indicated
Cardability, Bicomponent Core/Sheath Fiber that their experience revealed that PET fibers could be carded
faster than PP fibers. This mixed experiences could be due to
Introduction the fact that the fiber type was never been studied in a sys-
Among all fiberweb forming systems, carding provides tematic way. The cardability of a fiber depends on numerous
high flexibility by potentially using a wide array of materials interdependent parameters and a change in any one of these
to manufacture products with different features. Examples of parameters could cause a significant difference in cardability.
these are the use of recycled and natural fibers in numerous Carding parameters include fiber length, fiber crimp, fiber
textile products such as car interiors, wipes, and medical non- fineness, feed matt uniformity and openness, card wires types
woven products. Other benefits of the carding process are and density, machine configuration, number of fibers
high web uniformity, excellent fiber mixing and blending, processed/unit time, relative speeds of card rollers, fiber fin-
control of MD/CD strength ratio, and high flexibility in ish and finish level, etc. Previous researchers conducted card-

48 INJ Summer 2004

ability investigations utilizing commercial fibers and attempt- Table 1
ed to explain their results in terms of differences in fiber diam- VARIABLE AND CONSTANT PARAMETERS
eters, fiber crimp, fiber finish, number of fibers per unit time
processed through the card, etc.
The main goal of the research reported in this paper is to Parameters Constant Variable
Card Setting X
determine the effect of fiber type and process speed on fiber
Production Speed -85 m/min
cardability. To achieve the goal, we followed new approach to
-120 m/min
assess the cardability of different fiber types which have the Relative Speeds X
same diameter, fiber length, fiber crimp, and fiber finish. Number of Fibers X
Additionally, process these fibers at constant rate through the Fiber Types -PET
card for a given speed. This method could potentially reveal -PP
the reasons behind the dissimilar behavior of different fibers -Bicomponent1 (PET/PP)
in carding. -Bicomponent2 (PP/PET)
Fiber Diameter X
Experimental Approach Fiber Finish X
Experiments were designed in a way to allow the elimina- Fiber Crimp X
Fiber Length X
tion of many parameters, which affect the processability of
Card Clothing X
fibers. Table 1 shows the variables and constant parameters
used in the experiments. By this design, the effect of fiber type
could be assessed without any doubts since all other parame- only fixes the interrelation between fiber physics and wire
ters (except carding speed) that affect fiber processability were clothing geometry, but could also provide the ability of card-
kept unchanged. Thus at a given speed the impact of fiber ing a constant number of fibers per unit area of the fiberweb
type on the performance of fibers in carding could be clearly for each fiber type. In order to specify the above mentioned
understood. The card used in our experiment had a maximum fibers, PET fiber with linear density of 3 denier was chosen as
speed (doffer speed) of 120 m/min. While the design shown the base fiber and the diameter of this fiber was calculated as
in Table 1 does not allow the determination of the maximum the reference and the linear densities of the other three fibers
processing speed of each fiber, comparison of fiberweb uni- were subsequently calculated to match the diameter of the
formity at the two selected speeds may be used as a parame- PET fiber.
ter to indicate which fiber could handle faster carding speed. Fiber diameter df (cm) can be calculated in terms of fiber
The design of the experimental fibers was very crucial to denier N (g/9 km), and fiber density ρf (g/cm3). The following
this research. The main aspiration when designing the fibers equation can be derived by assuming a fiber with uniform cir-
was to create a carding condition which would leave the effect cular cross-section:
of fiber type isolated when carding under fixed operating con-
ditions. The specifications of the experimental fibers and fiber-
web basis weight, which corresponds to a fixed number of
fibers processed per unit time for a given speed, are illustrat- (1)
ed in Table 2.
Another essential aspect of the research was
the use of bicomponent fibers with a thick core
and sheath of minimum thickness. The dimen- Table 2
sions of the core and sheath of the bicomponent SPECIFICATIONS OF EXPERIMENTAL FIBERS
fibers are identical. This kind of structure
allows the observation of how a fiber’s behav-
ior in carding is affected by its total physical
structure, i.e. surface properties or as a whole.
100% PET fiber has the same surface character-
istics of bicomponent fiber 2, which is mostly
PP fiber (72% by volume or 63% by weight).
Additionally, 100% PP fiber has same surface Fiber Type PET PP Bicomponent 1 Bicomponent 2
characteristics as bicomponent fiber 1. Thus, Fiber Denier 3.00 1.96 2.70 2.25
the fiber specifications allow comparison of Outer Diameter, µ 17.54 17.54 17.54 17.54
100% PET fiber with 100% PP fiber and PP fiber Inner Diameter, µ - - 14.9 14.9
with the same surface characteristics as PET. % Weight, PET/PP 100/0 0/100 80/20 37/63
% Volume, PET/PP 100/0 0/100 72/28 28/72
Fiber Linear Density Calculation Web basis weight g/m2 30.0 19.6 27.0 22.5
It was determined that the fiber diameter Fiber Length: 63.5 mm (2.5 inch)
should be fixed for all fiber types, which not

49 INJ Summer 2004

 Substituting in Equation (1) for ρf = 1.38 and N = 3 for poly- be found from Equation 5. Substituting for the value of L
ester fiber, we obtain the value of polyester fiber diameter: (which is constant for all fiber types) and fiber linear density,
w for each fiber type can be calculated. The reason for calcu-
lating w for each fiber type is to adjust the carding machine
parameters to obtain the target basis weight for each fiber and
The polypropylene fiber denier can be calculated from hence ensure that all fiberwebs were obtained by processing
Equation (1), by substituting for df = 0.001754 (cm), and ρf = 0.9 same number of fibers per unit time through the card. The
(g/cm3). The result is shown in Table 2. result of calculating fiberweb basis weight for each fiber type
The bicomponent fiber denier can be calculated from the is shown in Table 2. Tables 3 and 4 show the target linear densi-
following two equations: ty and fiberweb basis weight versus the measured values for
each fiber type. The results of Tables 3 and 4 indicate that the
target fiber linear density and fiberweb basis weight were
(2) closely achieved.

Carding
The experimental fibers were conditioned for at least 24
hours prior to the trial. Because of the limited quantity of
(3) experimental fibers, a “cleaning fiber” was run between each
trial. This “cleaning fiber” was a 2 denier PP fiber (supplied by
Fiber Vision) which had the same finish as the experimental
fiber. Before each trial run, a bale of cleaning fibers was run
Where, wc /ws is the weight ratio of core fiber to sheath fiber, through the card. After the card was ready for the trial, exper-
dc is fiber core diameter, ρc density of core polymer, and ρs imental fibers were fed to the card through the fiber opening
density of sheath polymer. system and chute feed.
The bicomponent fiber 1 denier was calculated in two steps. Experiments were conducted at NSC-USA’s (Schlumberger-
The first step is to determine dc from Equation 2 by substitut- Thibeau Cards) Nonwoven Systems Showroom located at Fort
ing for wc /ws = 80/20, df = 0.001754 (cm), ρc = 1.38, and ρs = Mill, SC. The card used in the experiments was a Thibeau CA-
0.9. The second step is to calculate N from Equation 3 since all
the parameters on the right hand side of the equation are
known. Table 3
The denier of bicomponent fiber 2 can be calculated from COMPARISON BETWEEN TARGET AND
Equation 3. Then the % polymer component by weight can be MEASURED FIBERWEB BASIS WEIGHT
determined from Equation 2. The results are shown in Table 2. Fiber Process Target Measured
Parameters of Table 2 were used as processing variables to PET/PP Speed, m/min Weight, g/m2 Weight, g/m2
manufacture the fibers. The fibers were produced by Fiber 100/0 85 30.0 29.4
Innovation Technologies using a finish supplied by Goulston 100/0 120 30.0 29.5
Technologies. 0/100 85 19.6 20.9
0/100 120 19.6 20.8
Calculation of Web Weights 80/20 85 27.0 27.6
Carded web basis weight w (g/m2) for PET can be calculat- 80/20 120 27.0 28.3
ed from the empirical equation: 37/63 85 22.5 21.1
37/63 120 22.5 21.3
(4)

Where, k = 17 for double doffer cards [2].

Since it was intended that the number of fibers per unit area Table 4
of fiberweb would be constant for every fiber type, the total COMPARISON BETWEEN TARGET AND
length of fibers per unit area of fiberweb must be constant for MEASURED FIBER LINEAR DENSITY
all fiber types.
Fiber Target Measured
The basis weight of the fiberweb can be expressed in terms
PET/PP dpf dpf
of fiber denier N and the total length of fibers in one square
100/0 3.00 2.99
meter of fiberweb L as:
0/100 1.96 1.95
80/20 2.70 2.70
(5) 37/63 2.25 2.40

Now w can be calculated for PET from Equation 4 and L can

50 INJ Summer 2004

10 (2255PP) dynamic roller-top card. This had a dou- Figure 1
ble doffer configuration and was equipped with lat- FIBER STRENGTH AT DIFFERENT SPEEDS AND
est technology systems such as Servo-X input auto- PROCESS STAGES
leveler, LDS (Linear Doffing System) and WID (Web
Introduction Device). The maximum speed of this
card was 120 m/min and the working width was 2.5
meters.

Design of Experiment
To study the influence of the fiber type and card-
ing speed the four fibers and two speeds shown in
Tables 1 and 2 were used. As can be seen from the
tables, the design requires a total of eight runs (4
fibers X 2 speeds). For each experimental run, a fiber
was processed through the card until the card
reached a steady state (10-15 minutes) before taking
any samples. After that, the card was suddenly
stopped using the “emergency brakes” to eliminate
any progressive damage or changes to the fiberweb
that may have been caused by the slowing of the Figure 2
card. Afterwards, samples from the feed matt and FIBER MODULES AT DIFFERENT SPEEDS AND
the fiberweb were collected in order to carry out PROCESS STAGES
planned tests. The samples were collected from dif-
ferent locations across the card.

Testing and Evaluation

Several tests were executed to measure various
characteristics of the collected samples. Fibers, feed
matt and web parameters were measured and com-
pared in order to reveal potential effects of the fiber
type and carding speed on carding performance and
quality. The fibers were evaluated for their linear
density, crimp stability, strength, and modulus before
and after carding. The carded webs from different
fibers and speeds were evaluated for their uniformi-
ty, thickness and fiber orientation. The dominant
angle was used to represent fiber orientation.
Numerous testing techniques and equipment were used for all experiments), respectively.
employed to complete the testing phase of the research. State Statistical Analysis System (SAS) was used to reveal any
of the art fiber measurement devices, latest image analysis statistically significant similarities or differences in the mea-
techniques and a variety of analytical testing instruments sured parameters among the fiber types and between the two
were utilized and details of these are reported elsewhere [3]. production speeds.
Since most of the tests carried out did not have established
standards, the number of samples was determined by statisti- RESULTS AND DISCUSSION
cal means. Before each test, a preliminary test was carried out Fiber Properties
and the number of samples needed was calculated using the One of the parameters to assess cardability was the change
following formula: in fiber strength and modulus after carding as compared to
the raw materials. Figure 1 shows the results of the fiber
strength at different speeds and processing stages. The results
(6)
indicate that in general the change in fiber strength is insignif-
icant. Further, statistical analysis using General Linear Model
(GLM) procedure revealed that the effects of process speed
and process speed/fiber type interaction are not significant.
Where n, za/2 , σe and E are sample size, constant (1.96 for These findings were supported by paired comparison using F-
95% CI), estimated standard deviation and error (10% was test of fiber strength for each fiber at the two carding speeds.

51 INJ Summer 2004

 Figure 2 illustrates the fiber modulus results Figure 3
measured at different processing stages and FIBER CRIMP STABILITY AT DIFFERENT SPEEDS AND
speeds. The GLM procedure and paired compari- PROCESS STAGES
son F-test showed that fiber modulus was not
influenced significantly by process speed and
there was no effect of first order interaction of
speed/fiber type on fiber modulus.
The fiber crimp stability, level, and geometry
are important parameters for converting fibers
into webs. The carding performance is affected by
fiber crimp characteristics, especially in high-
speed carding. Crimp allows fiber bundles to be
more easily held in structure of the fiberweb and
fiber separation into individual fibers. This fiber-
web cohesion is achieved by the hooks of the
crimp. The fiber cohesion has positive contribu-
tion to fiberweb uniformity and carding efficiency
since incoherent fiberweb breaks and causes fre-
quent stops. Figure 4
Figure 3 depicts the results of crimp stability FIBER LINEAR DENSITY (DENIER/FILAMENT OR
(the ability of fiber to maintain its crimp level after
DPF) AT DIFFERENT SPEEDS AND PROCESS STAGES
it undergoes loading cycles) at different carding
speed and process stages. The only fiber with sig-
nificantly lower crimp stability as compared to the
rest of its values at different stages is
37%PET/63%PP fiber processed at carding speed
of 120 m/min. This fiber lost 15.8% of its crimp as
compared to the raw material stage.
All fibers were evaluated for their linear densi-
ties at different process stages to investigate
whether change in linear density may take place
as a result of processing the fibers since the fibers
undergo stretch and heat during processing.
Figure 4 shows the results of the fiber linear den-
sity. The results of Figure 4 supported by statisti-
cal analysis prove that there is no significant
change in fiber linear density as a result of pro-
cessing.
Figure 5
Fiberweb Properties
EFFECT OF FIBER TYPE AND CARDING SPEED ON
Carded web thickness, fiber orientation, and uni- FIBERWEB THICKNESS
formity are important structural parameters that
influence the final product’s properties. Figures 5-7
show the results of these properties.
Figure 5 depicts the thickness data of eight fiber-
webs (4 fibers X 2 carding speeds) processed. The
behavior of fiberweb thickness can be explained by
the crimp stability behavior shown in Figure 3. As
it can be seen from Figure 3, the PP rich fibers show
lower crimp stability than the PET rich fibers. The
PP fibers lost significant part of their crimp due to
processing while the PET rich fibers maintain their
crimp. As a result of crimp loss and lower crimp
stability, fiberwebs from PP rich fibers possessed
lower thickness than those of fiberwebs from PET
rich fibers. The reason behind such behavior is that

52 INJ Summer 2004

with high fiber crimp the air trapped inside the struc-
ture is high a matter that leads to bulkier (or thicker)
Figure 6
fiberwebs than those processed using low fiber crimp. EFFECT OF FIBER TYPE AND CARDING SPEED ON
Figure 6 shows the effect of fiber type and carding FIBER ORIENTATION
speed on the web fiber orientation represented by the
dominant angle. Here the angle 90O represents the
machine direction (MD). The results of the eight fiber-
webs indicate that: (1) Most of the fibers are oriented
approximately in MD, (2) The process speed did not
cause significant change in fiber orientation except in
the case of PP which showed the highest difference
between the two speeds (5.6O). While the 5.6O fiber ori-
entation difference is statistically significant per
paired F-test, its practical implication for most non-
wovens applications is not significant.
Figure 7 depicts the effect of fiber type and carding
speed on fiberweb uniformity index. In general, as the
carding speed increases the fiberweb uniformity gets
lower. The paired comparison using F-test showed
that there is no significant effect of carding speed on Figure 7
fiberweb uniformity except in the case of EFFECT OF FIBER TYPE AND CARDING SPEED ON
80%PET/20%PP fiber. This is unexpected since the
FIBERWEB UNIFORMITY
results of Figures 1-4 prove that 80%PET/20%PP fiber
did not undergo significant changes in properties due
to carding.

Conclusions
We have employed a novel approach to investigate
the impact of carding speed and fiber type on fiberweb
properties. Four fibers were designed and extruded
with same fiber diameter, length, crimp, and finish.
The fibers were processed using state-of-the-art high-
speed card. The processing parameters were kept con-
stant including the number of fibers per tooth of a card
element. With this setting, the effects of all the para-
meters (except carding speed and fiber type) were nul-
lified. The fiber properties (strength, modulus, crimp
stability, and linear density) were measured and
reported to investigate whether fiber properties would Acknowledgement
change as a result of processing. For all four fibers, fiber The current research was funded by the Nonwovens
strength, modulus, and linear density did not significantly Cooperative Research Center at NC State University and their
change by carding as compared to the raw fiber. Additionally, support and the assistance of industry members is gratefully
PET rich fibers did not significantly change their crimp and acknowledged.
crimp stability by processing while PP rich fibers showed sig-
nificant loss in crimp and lower crimp stability due to card- References
ing. Despite the crimp loss and lower crimp stability of PP 1. Chan, K., ITMA 99 Survey 21: Nonwovens Equipment,
rich fibers as compared to PET rich fibers, the fiberweb uni- Textile Asia, 31-34, November, 1999.
formity and fiber orientation were not affected by fiber type 2. Cozon, J. N., Schlumberger Inc., Personal
and carding speed. The crimp loss and the lower crimp sta- Communications.
bility of PP rich fibers caused reduction in fiberweb thickness 3. Doguc, N.B., “Influence of Fiber Types on Fiberweb
as compared to those fiberwebs made from PET rich fibers. Properties in High-Speed Carding,” M.S. Thesis, North
Our investigation showed that PP and PET fibers cardabili- Carolina State University, 2002.
ty are almost identical within the experimental range studied. 4. Fuchs, H., “Nonwovens Production Strategic Objectives
Following ITMA,” AVR- Allgemeiner Vliessstoff-Report 1/2000,
38-41. — INJ

53 INJ Summer 2004