Location via proxy:   [ UP ]  
[Report a bug]   [Manage cookies]                

Evaluation and Optimization of Seal Behavior Throu

Download as pdf or txt
Download as pdf or txt
You are on page 1of 10

Received: 7 December 2018 Revised: 18 March 2019 Accepted: 18 March 2019

DOI: 10.1002/pts.2442

PAPER SUBMITTED TO IAPRI PEER REVIEW STREAM

Evaluation and optimization of seal behaviour through solid


contamination of heat‐sealed films

Bram Bamps1 | Karlien D'huys2 | Ina Schreib3 | Benjamin Stephan3 |

Bart De Ketelaere2 | Roos Peeters1

1
Hasselt university, IMO‐IMOMEC, Packaging
Technology Center, Wetenschapspark 27, A method is presented to apply solid powder/granulate contamination (ground coffee
3590 Diepenbeek, Belgium
and blood powder) in between the heat conductive seals of flexible packaging mate-
2
Department of Biosystems, MeBioS, KU
Leuven, Kasteelpark Arenberg 30, B‐3001 rials. A response surface method is tested and validated to optimize seal strength of
Heverlee, Belgium heat conductive sealing with and without solid contamination. In this study, a maxi-
3
Division Processing Technology, Fraunhofer
mal seal strength is defined as optimal. Using these methods, three typical packaging
Institute for Process Engineering and
Packaging, Heidelberger Str. 20, 01189 films with varying seal layer composition (metallocene linear low‐density polyethyl-
Dresden, Germany
ene (LLDPE), plastomer, and sodium ionomer) are maximized towards contaminated
Correspondence seal strength. Contamination caused a decrease in seal strength and narrowed down
Bram Bamps, Hasselt university, IMO‐
the process window (seal temperature and time combinations) in which at least 90%
IMOMEC, Packaging Technology Center,
Hasselt 3500, Belgium. of the maximal strength is obtained. The influence of seal layer composition on the
Email: bram.bamps@uhasselt.be
clean and solid (ground coffee and blood powder) contaminated seal performance
Funding information (seal strength, process window, and leak tightness) was evaluated. The film with the
German Federal Ministry for Economic Affairs
plastomer‐based seal layer outperformed the other films with respect to the width
and Energy, Grant/Award Number: 172 EBR;
Agentschap Innoveren & Ondernemen, Grant/ of the process window. It also reached a higher seal strength and a higher amount
Award Number: 150817
of leak tight seals (evaluated with the dye penetration test) after optimization. The
hot tack test was evaluated as predictive test for the contaminated seal strength.
The results of this study do not support an indicative relationship.

K E Y W OR D S

hot tack, polyethylene, process window, response surface model, seal through contamination

1 | I N T RO D U CT I O N for heat seals. There are several technologies to heat a seal, such as
heat conductivity, inductivity, and ultrasonic vibrations. Conductive
The closure of flexible packages is mainly established by applying heating is the most popular technology for flexible packaging.1 Indus-
heat sealable materials at the inner side of the packaging material, trial research in the past showed that one‐third of sealed packages
which creates a seal after bringing the inner sides together, heating, are of an insufficient quality.2,3 Therefore, the tightness of the pack-
and compressing them. The materials melt, and polymer chains at age, which is in general essential to prevent microbial and biochemi-
the interface entangle, and after cooling, a strong seal can be cal degradation of food, cannot be guaranteed. For 65% of the
obtained. Besides chain entanglement, intermolecular and mechanical packages that show seal defects, contamination in between the seal
bonds are possible, but these adhesion mechanisms are less common layers is the major cause.3 Only 16% of the packers perform an

--------------------------------------------------------------------------------------------------------------------------------
This is an open access article under the terms of the Creative Commons Attribution‐NonCommercial‐NoDerivs License, which permits use and distribution in any
medium, provided the original work is properly cited, the use is non‐commercial and no modifications or adaptations are made.
© 2019 The Authors. Packaging Technology and Science Published by John Wiley & Sons Ltd

Packag Technol Sci. 2019;0–11. wileyonlinelibrary.com/journal/pts 1


2 BAMPS ET AL.

inspection of all produced goods. The majority just inspects samples 3 | MATERIALS AND METHODS
with an interval of 30 minutes.3
One way to prevent seal defects as a result of contamination is 3.1 | Materials
to work with materials that are able to seal through contamination
at particular seal settings to decrease the amount of defective pack- 3.1.1 | Commercial multilayer packaging films for
ages. Several polyethylene (PE)‐based packaging materials have been flowpack applications
developed with a good seal performance through contamination in
the last decades. Examples of these materials are metallocene‐ Table 1 shows the multilayer structure of three flowpack films, evalu-
catalysed linear low‐density PE (LLDPE), polyolefin plastomer, and ated and optimized in this study. Each film has a 12‐μm thick PE
ionomer.4 terephtalate (PET) outer layer. A 3‐layer blown film line with three
In scientific and technical papers, several tests are performed on nozzles is used for the production of the seal layers. The upper
packaging films with contaminated seals such as seal strength, leak 35 μm has the same composition for the three films, containing a
rate, and degree of particle encapsulation (caulkability).5-9 This study blend of low‐density PE (LDPE) and metallocene LLDPE (mLLDPE).
is focussed on seal strength. Hot tack tests are performed to The films differ mainly in the 15‐μm lower seal layer. Film 1 has a
evaluate the resistance of packaging films against spring back blend of LDPE and mLLDPE while film 2 has a blend of LDPE and
6,7,9,10
forces. The relation of seal through contamination and hot tack polyolefin plastomer (mLLDPE with a high amount of comonomer) in
performance is part of this study. In the last decades, the influence of that area. Both films have 2% processing aid in the lower seal layer.
seal material composition on the contaminated seal performance was Film 3 has a 5‐μm layer of acid copolymer resin between the 35‐μm
the topic of a limited amount of studies, and these studies did not PE and the 10‐μm ionomer layer to ensure the bonding of both layers.
include a well‐described application method for solid contamina-
tion.5,6,8,9 Moreover, there are no methods described in these studies 3.1.2 | Contamination types
to obtain the optimal or maximal seal performance through
contamination. Two types of solid contamination are used in this study: (a) sieved
ground coffee (Delhaize, Belgium; sieved to obtain a particle size
between 500 and 630 μm using a Fritsch analysette 3 sieve shaker
system) and (b) freeze‐dried pork blood powder (Solina, Germany; par-
2 | OBJECTIVES ticles with an average size of 100 μm).

The main objective of this study is to present a method to optimize


3.2 | Methods
the granular‐contaminated seal strength of packaging films. Firstly, a
protocol to apply solid contamination in a standardized way is 3.2.1 | Sample preparation contaminated seals
described, this protocol was missing in previous studies with powder
or granulate contamination. Secondly, an optimization method is pre- Films, sealed samples, and solid contamination are stored at a temper-
sented that is based on a previous study on ultrasonic sealing.11 In this ature of 23°C and a relative humidity of 50%. The precision balance
study, a similar methodology is used in order to receive information on OHAUS Explorer (Mettler‐Toledo International Inc, USA) with read-
the influence of all relevant seal parameters on the heat conductive ability of 0.0001 g is used for all weighings. Figure 1 shows an illustra-
sealing process based on a limited number of carefully selected tion of the sample preparation. The sample is cut in machine direction
experiments. (MD) with a width of 50 mm and an appropriate length to do a seal
A second objective is to evaluate the influence of variation in seal strength test (in this study the length exceeded 100 mm) (I). An area
layer composition (metallocene PE, plastomer, and ionomer) on the of 20 by 40 mm is then marked on the film sample. It is important that
seal through contamination performance (seal strength, width process the chosen length has sufficient extra margin compared with the seal
window, and leak tightness) by using the application and optimization length to ensure that the contamination is distributed over the full
methods of this study. Hot tack tests are evaluated as predictive tests length of the seal. In this case, 20 mm is chosen because the used
for contaminated seal performance. sealer produces 10 mm length seals. The required amount of solid

TABLE 1 Multilayer structure of the three packaging films

Film 1 Film 2 Film 3

PET 12 μm PET 12 μm PET 12 μm


80% LDPE‐20% mLLDPE 15 μm 80% LDPE‐20% mLLDPE 15 μm 80% LDPE‐20% mLLDPE 35 μm
80% LDPE‐20% mLLDPE 20 μm 80% LDPE‐20% mLLDPE 20 μm Acid copolymer resin 5 μm
68% LDPE‐30% mLLDPE 15 μm 68% LDPE‐30% plastomer 15 μm Sodium ionomer 10 μm

Abbreviations: LDPE, low‐density polyethyelene; mLLDPE, metallocene linear low‐density polyethylene; PET, polyethylene terephtalate.
BAMPS ET AL. 3

FIGURE 1 Contaminated seal preparation

contamination is weighed. In order to facilitate the sealing with con- divided by seal width (25 mm) to obtain the hot tack strength. Seal
tamination, a simple cardboard tool is cut out of a cardboard sheet time, seal pressure, and cooling time were kept constant at respective
with an inner hole with larger width and length than the seal jaw. After values of 1.0 second, 0.3 N mm−2, and 0.1 second while seal temper-
one film sample is fixed with plastic tape to this cardboard tool, the ature was varied. Samples are tested in threefold, and average values
contamination is applied with a small spoon in the designated area. and standard deviations are shown.
In this study, 0.020 g is applied in a 20 × 40‐mm2 region to achieve
a 25 g m−2 contamination density. When the contamination is applied
and evenly distributed by eye, the second film sample is used to cover 3.2.3 | Seal characterization
the contamination and fixed with plastic tape (II). The cardboard tool
with contaminated film samples is manually placed between the seal Seal strength is tested according to ASTM F88 on 15‐mm wide sam-
jaws, and the seal is formed. Seals are produced with the Labthink ples. These samples are tested unsupported. Clamp distance is
HST‐H3 heat seal tester (Labthink Instruments Co Ltd, People's 10 mm, and tensile speed is 300 mm min−1. The seal strength is
Republic of China). This sealer has two flat aluminium jaws covered obtained by dividing maximum force with seal width (15 mm).
with silicon tape to prevent the contamination from sticking to the The dye penetration test uses an aqueous solution with 0.05%
jaws. After sealing, the amount of contamination, which is not trapped indicator dye (toluidine blue) and 3% wetting agent (PE glycol
within the seal area, is carefully removed using a small brush (III). octylphenyl ether) to determine if there are leaks. It allows to detect
and locate channel leaks, which are equal to or greater than channels
caused by a wire with diameter of 50 μm. The qualitative information
3.2.2 | Film characterization
(leak or no leak) delivers complementary information to seal strength.
Differential scanning calorimetry It is performed according to ASTM‐F3039 on samples as shown in
To characterize the thermal behaviour of the packaging material and Figure 2. The edges of the (contaminated) seal samples need to be
to relate this to the heat conductive sealing performance, differential sealed to have a reservoir where 1‐mL dye can be poured in (I). After
scanning calorimetry (DSC) measurements were executed with the removing the sealed edges at one side of the seal, the seal can be
instrument 2920 MDSC V.2.6A (TA instruments, USA). The three film pressed to an absorbing white paper after adding the dye. The pack-
samples and the main components of the seal layer (granulate form of age is held in a vertical position for 1 minute so that the dye can pen-
LDPE, mLLDPE, plastomer, acid copolymer resin, and sodium ionomer) etrate through possible channel leaks by gravity (II). If a stain is visible
were tested in a sequence of two controlled heatings and one cooling on the paper, the seal is reported as leaker. This test can be done prior
down stage within the range of 10°C ➔ 200°C at a heating/cooling to performing the seal strength test if the seal strength of a sample is
speed of 10°C min−1. The heating does not exceed 200°C to prevent
the PET layer from melting as this study focusses on the components
of the seal layer and the seal layer as its whole. The first heating cycle
is performed to delete the thermal history. The second heating cycle is
used to obtain the melting peak temperature and the melting onset
temperature (intersection of the tangent of the peak and the extrapo-
lated baseline). Both of these temperatures are used to compare the
materials.

Hot tack
Twenty‐five‐millimetre‐wide samples were tested with a J&B Hot
Tack Tester model 5000 MB (Vived‐Management, Belgium) according
to ASTM F1921 at a tensile speed of 200 mm s−1. Maximum force is FIGURE 2 Dye penetration test
4 BAMPS ET AL.

not influenced by the penetrating dye. This was confirmed by prelim- criterion (BIC). For a more detailed description of the model selection,
inary tests (results not shown). the reader is referred to the previous study on ultrasonic seals.11
Samples that pass this dye penetration test are considered as leak In a fourth step, the response surface model was utilized for opti-
tight. mizing the input variables towards the response (seal strength). In this
study, maximum seal strength is defined as an optimal result and was
3.2.4 | Maximization of contaminated seal quality thus assigned a desirability = 1. In addition to determining one optimal
parameter setting, a process window can be generated, which for
In order to assess the effect of the sealing parameters (temperature, example excludes parameter combinations resulting in seal strength
time, and pressure) on the seal quality of both clean and contaminated below a certain threshold. In this study, process windows were gener-
(coffee and blood powder) seals, the approach presented in D'huys ated containing only those parameter combinations at which at least
et al was followed. This approach is based on the concepts of design 90% of the maximum seal strength is reached.
of experiments (DOE) and response surface modelling. The steps will In a fifth and last step, the optimum was experimentally validated
be briefly described below for the case considered in this study. by performing repeated measurements (n = 10) at the optimal settings
First, a design space was defined that includes the three most to check if the predicted optimum lies within the confidence interval
important seal parameters: jaw temperature (°C), seal time (s), and seal calculated from the measured values. The success of the confirmation
pressure (N mm−2). The effect of these parameters is studied within is assessed by following an approach based on a confidence interval
certain limits. There are no strict rules to set these limits. They can calculation of the confirmation runs (CICon approach) as suggested
be based on film specifications, on the limits of the sealing process, in previous research.12,13 For details, the reader is referred to the pre-
on the relevance for the application, and/or on results of preliminary vious study on ultrasonic seals.11
tests. In this study, the limits considered for jaw temperature, seal
time, and pressure are, respectively, 120°C to 180°C, 0.4 to 1.0 s,
and 1.0 to 4.0 N mm−2. 4 | RESULTS AND DISCUSSION
Secondly, an experimental design is set up. In this study, a 20‐point
I‐optimal design was selected, rather than the Box‐Behnken design of 4.1 | Film characterization
a previous study.11 This I‐optimal design allows to include corner
points of the design space, which represent extreme parameter combi- It is not always possible to compare DSC results of packaging films
nations. Moreover, it allows to include a third‐order effect of seal with blown extruded films with 100% pure material such as
pressure in the response surface model, which was shown to possibly plastomer because of low viscous behaviour of this substance.
be of interest based on preliminary experiments. Because of this, the results of the seal layers of the three packaging
The third step involves fitting a response surface model with three films are compared with the individual components in granulate
input variables (temperature, time, and pressure) to the seal strength form. This allows to identify similarities and to suggest explanations
values obtained at the 20 experimental runs. The following quadratic in differences and similarities of the three packaging films. The melt-
model with interactions was fitted: ing onset and peak temperatures of the films and granulates in this
study are shown in Table 2. The values of film 1 are in between
y ¼ β0 þ β1 x1 þ β2 x2 þ β3 x3 þ β12 x1 x2 þ β23 x2 x3 þ β13 x1 x3 the values of its main components LDPE and mLLDPE in granulate
þ β11 x21 þ β22 x22 þ β33 x23 þ β333 x33 þ ε; form. In a previous study on blended films of LDPE and mLLDPE,
it was found that the melting point of the blended monolayer was
where x1, x2, and x3 are the three input parameters, seal temperature, between the values observed for mLLDPE and LDPE films.14 The
time, and pressure, y is the seal strength and ε is the error term. Besides melting onset temperature of film 2 is decreased with 5° compared
the main effects, the interaction terms and quadratic terms were also with film 1. The presence of plastomer instead of mLLDPE in the
considered in the model. Moreover, for pressure, a third‐order effect lower 15 μm of the seal layer is suggested as explanation since
was also included in the model. Nonsignificant effects are removed plastomer has a lower melting point than mLLDPE. This is indicated
from the model by an all possible subsets procedure. The model with by melting peak temperatures of the components in granulate form.
the best fitting subset of effects is selected. The criteria of this selection The melting peak temperature of the seal layer of film 2 is close to
are R2, Akaike information criterion (AIC), and Bayesian information the value of film 1. Acid copolymer and sodium ionomer presence,

TABLE 2 Melting onset and melting peak temperatures of films and granulates, tested with DSC

Film Granulate of Film Component


1 2 3 LDPE mLLDPE Plastomer Acid Copolymer Sodium Ionomer

Tmelt onset (°C) 100 95 98 102 95 87 77 70


Tmelt peak (°C) 112 113 112 112 111 102 98 90

Abbreviations: DSC, differential scanning calorimetry; mLLD, metallocene linear low density; LDP, low‐density polyethyelene.
BAMPS ET AL. 5

which have lower melting points than LDPE, mLLDPE, and both films, the hot tack strength decreases strongly in a similar way at
plastomer, does not decrease the melting peak and onset tempera- elevated temperatures (greater than or equal to 150°C). Film 3, with
ture of film 3. These components are present in low amount in pro- the sodium ionomer seal layer, shows a larger standard deviation com-
portion to LDPE and mLLDPE, and their melting temperatures are pared with the other films. It has a low hot tack initiation temperature
probably too low to influence the tangent line, which is used to (90°C), then the hot tack strength slowly increases until a high peak
obtain the onset temperature. value (0.41 ± 0.07 N mm−1) is reached at 115°C. The hot tack window
The hot tack strength is relevant for sealing through solid particles is very broad (100°C‐greater than or equal to 180°C), indicating that
as these particle can push the seal layers away from each other this film keeps a large portion of its strength at seal temperatures
directly after opening the hot jaws when the seal is still hot.15 This greater than or equal to 150°C. The superior hot tack performance of
spring back effect is similar when wrinkles are present. The effect is ionomers (high hot tack strength at low seal times and under a wide
discussed in several papers.6,7,9,10 Figure 3 shows the hot tack results range of seal temperatures) was previously described in literature.16
of all films. The hot tack initiation temperature (temperature where a Both films 2 and 3 are evaluated as good hot tack performers because
low but measurable hot tack strength is obtained), peak value, and of a combination of hot tack properties (low initiation temperature, high
window (temperature range where a relatively high hot tack strength peak value, and wide window).
is obtained) are discussed to compare the three packaging films. As
there is currently no clear definition of the hot tack initiation temper-
ature, it is defined in this study as the minimum temperature (°C) 4.2 | Evaluation and optimization of contaminated
where a seal with low‐hot tack strength is produced, a threshold value seal strength
of 0.03 N mm−1 must be exceeded. The hot tack window was defined
as the temperature range (°C) from minimum to maximum tempera- The experimental design of the three films of this study is shown in
ture where seals with medium hot tack strength are produced, a Table 3. At each of the parameter combinations defined by the design,
threshold value of 0.1 N mm−1 must be exceeded. both clean and contaminated (coffee and blood powder) seals were
Film 1 has a relatively high hot tack initiation temperature (105°C) created, and their seal strength was measured. Other responses such
compared with the other films. This can be a result of the absence of as leak tightness, seal energy, optical aspects, and seal thickness are
low melting main components such as plastomer or sodium also possible in a single or multi‐response model but were not consid-
ionomer in the lower 15 μm of the seal layer. The peak value ered in this study. All clean and contaminated seal strengths of films 1,
(0.29 ± 0.01 N mm−1) is low compared with other films. The peak value 2, and 3, produced at the 20 parameter combinations (temperature,
is reached at 110°C, which is in accordance with the melting peak time and pressure) defined by the experimental design are also shown
temperature of film 1 and the individual granulates of the two main in the table. These seal strengths served as an input to build a model
components, LDPE, and mLLDPE. The hot tack window is narrow that predicts the clean and contaminated seal strength at all possible
(110°C‐140°C) compared with the other films. Film 2 has a relative parameter settings within the defined design space.
low hot tack initiation temperature (90°C) making it suitable for high In Tables 4, 5, and 6, the coefficients of the terms included in the
speed sealing applications. This can be a result of the presence of models for each film are summarized for clean seals, coffee contaminated,
plastomer6 in the lower 15 μm of the seal layer. Compared with film 1, and blood powder contaminated seals, respectively. Nonsignificant terms
the peak value (0.43 ± 0.03 N mm−1) is high, suggesting more and/or were not included in the model, and therefore no coefficient is shown in
9
deeper entanglement, this was previously described in literature. For the table. Based on these models, the settings of temperature, time, and
pressure resulting in maximum seal strength were determined for each
film‐contaminant combination as described in the Methods section.
As a validation of the optima of the three films, the predicted opti-
mal parameter settings, predicted maximum seal strengths, and limits
of the confidence intervals calculated based on the validation experi-
ments (CICon approach) are shown in Table 7.
The predicted values are a good indication of the measured values,
but the model tends to slightly overestimate the optimized seal
strength. A higher accuracy could be reached by adding repetitions
to the test or augmenting the experimental design with additional
experiments. This can be a subject for further research. Contamination
decreases the seal strength, even when optimized. The rate of
decrease is dependent on the used seal material (blend
LDPE/mLLDPE, blend LDPE/plastomer, and sodium ionomer) and
the applied contamination (ground coffee and blood powder). For films
FIGURE 3 Hot tack graph with variation in seal temperature, seal
parameters: 1.0 s seal time‐0.3 N mm−2 seal pressure‐0.1 s cool time 1, 2, and 3, the degrees of decrease, based on the measured average
(n = 3) for three packaging films values (not shown in table), are, respectively, 25%, 16%, and 63% for
6 BAMPS ET AL.

TABLE 3 Experimental design with parameters (seal temperature [T], time [t], and pressure [p]) and responses (clean and contaminated seal
strength) of films 1, 2, and 3 (F1, F2, and F3)

Response: Seal Strength (N mm−1)

Ground Coffee Blood Powder


Clean Contamination (25 g m−2) Contamination (25 g m−2)

pseal
Run Tjaw (°C) F1 Tjaw (°C) F2 Tjaw (°C) F3 tseal (s) (N mm−2) F1 F2 F3 F1 F2 F3 F1 F2 F3

1 149.3 141.2 143.9 0.7 1.9 2.4 3.0 1.4 1.4 2.3 0.6 0.7 1.4 0.3
2 181.5 181.5 181.5 0.4 3.2 2.3 3.1 1.8 1.9 2.0 0.3 0.3 0.4 0.3
3 119.6 104.1 109.2 0.5 1.6 0.6 1.9 0.1 0.0 0.0 0.0 0.0 0.0 0.0
4 150.5 142.7 145.3 0.7 3.3 2.6 2.8 1.7 1.4 1.9 0.6 0.7 1.7 0.3
5 150.5 142.8 145.4 0.7 3.2 2.5 3.1 1.8 1.5 2.1 0.5 0.7 1.5 0.6
6 162.6 157.8 159.4 1.0 3.4 2.2 2.6 1.3 2.2 2.3 0.8 0.4 0.8 0.5
7 144.1 134.7 137.9 1.0 1.9 2.5 3.0 1.5 1.4 2.6 0.6 1.1 1.4 0.5
8 119.6 104.1 109.2 0.7 3.1 2.3 2.9 0.4 0.6 0.9 0.0 0.5 0.1 0.0
9 181.5 181.5 181.5 0.7 1.0 2.2 2.9 1.8 1.9 2.9 0.5 0.9 0.6 0.2
10 148.8 140.6 143.3 0.7 1.8 2.4 2.9 1.7 2.0 2.9 0.6 0.7 0.9 0.3
11 144.3 135.0 138.1 0.4 4.0 2.3 3.2 0.5 1.2 1.7 0.1 0.7 1.0 0.2
12 119.6 104.1 109.2 1.0 4.0 2.2 2.9 0.4 1.4 1.5 0.1 1.1 0.6 0.3
13 181.5 181.5 181.5 0.4 1.9 2.3 2.9 1.7 1.1 1.4 0.5 0.3 0.3 0.4
14 181.5 181.5 181.5 1.0 1.0 2.2 2.9 1.7 1.3 3.0 0.5 0.6 0.5 0.4
15 150.5 142.8 145.4 0.4 1.0 2.4 3.1 0.7 1.0 2.1 0.3 0.4 1.4 0.2
16 119.6 104.1 109.2 1.0 1.0 2.1 3.0 0.3 1.6 1.1 0.0 0.8 0.3 0.1
17 181.5 181.5 181.5 1.0 2.4 2.2 3.6 1.8 2.0 3.2 0.5 0.4 1.6 0.4
18 181.5 181.5 181.5 0.7 4.0 3.1 2.6 2.2 2.0 2.9 0.7 0.3 0.3 0.4
19 119.6 104.1 109.2 0.4 3.3 0.7 1.1 0.1 0.0 0.0 0.0 0.0 0.0 0.0
20 150.8 143.1 145.6 0.7 1.8 2.3 2.9 1.5 1.2 2.5 0.5 0.6 2.0 0.3

TABLE 4 Significant coefficients, terms, and regression significance of the seal strength model for clean seals of films 1, 2, and 3

Film 1 Film 2 Film 3


Coefficient Term Value P value Value P value Value P value

β0 Intercept −0.2435 .6765 1.5979 .0314 −2.0659 .0001


β1 T 0.0126 .0012 0.0071 .0523 0.0219 <.0001
β2 t 0.7350 .0295 0.6962 .1144 0.5145 .0364
β3 p 0.0981 .1577 −0.0479 .6028 0.0552 .2711
β12 T*t −0.0446 .0029 −0.0328 .0326 −0.0130 .1112
β23 t*p — — −0.1641 .6741 — —
β13 T*p — — −0.0012 .7122 — —
β11 T 2
−0.0005 .0089 −0.0003 .0857 −0.0001 .1203
β22 t2 −2.7627 .1164 — — −4.1615 .0046
β33 p 2
0.1299 .1344 — — −0.1272 .0535
β333 p 3
— — — — — —

ground coffee and 71%, 45%, and 79% for blood powder contamina- particles when a same mass of contamination is applied. Figure 4
tion compared with the clean seal strength. The samples contaminated shows that there are more clean areas with contaminated seal samples
with coffee reach a higher optimized seal strength than the samples with coffee particles compared with those with blood powder.
with blood powder. This can be a result of more binding spots Seal pressure has a slight or no influence on the seal strength as
between the seal layers because of the lower amount of coffee shown in Tables 4, 5, and 6. There is no significant effect of pressure
BAMPS ET AL. 7

TABLE 5 Significant coefficients, terms, and regression significance of the seal strength model for coffee contaminated seals of films 1, 2, and 3

Film 1 Film 2 Film 3

Coefficient Term Value P value Value P value Value P value

β0 Intercept −1.0186 .0085 −1.8826 .0177 −0.7098 .0111


β1 T 0.0106 <.0001 0.0250 <.0001 0.0086 <.0001
β2 t 1.2563 <.0001 1.0453 .0104 0.0202 .8879
β3 p 0.0893 .0347 −0.0446 .8187 −0.0254 .4254
β12 T*t −0.0459 <.0001 −0.0207 .0900 −0.0091 .0902
β23 t*p −0.3708 .0362 — — — —
β13 T*p — — — — −0.0029 .0232
β11 T2 −0.0006 <.0001 −0.0006 .0004 −0.0001 .1426
β22 t2 — — — — — —
β33 p 2
— — −0.0785 .4184 −0.0502 .2118
β333 p 3
— — 0.1347 .2000 — —

TABLE 6 Significant coefficients, terms, and regression significance of the seal strength model for blood powder contaminated seals of films 1, 2,
and 3

Film 1 Film 2 Film 3


Coefficient Term Value P value Value P value Value P value

β0 Intercept 0.4262 .1593 0.5214 .3284 −0.5112 .0049


β1 T −0.0016 .3138 0.0030 .3214 0.0037 .0004
β2 t 0.6953 .0006 0.5969 .1189 0.3370 .0025
β3 p −0.0194 .5617 — — 0.0404 .0582
β12 T*t −0.0276 .0010 — — −0.0046 .1595
β23 t*p −0.2909 .0587 — — — —
β13 T*p −0.0047 .0059 — — — —
β11 T2 −0.0003 .0013 −0.0007 <.0001 −0.0001 .0042
β22 t2 — — — — 0.4579 .3619
β33 p2 0.0684 .1176 — −0.0220 .3829
β333 p 3
— — — — — —

TABLE 7 Validation of statistical optimum (n = 10) + optimal settings

Seal Strength (N mm−1)

Clean Ground Coffee Contamination (25 g m−2) Blood Powder Contamination (25 g m−2)

Predicted CI measured Predicted CI measured Predicted CI measured

Film 1 3.0 2.2‐2.7 2.1 1.6‐2.0 1.1 0.6‐0.9


Film 2 3.3 3.0‐3.2 3.1 2.3‐2.8 1.6 1.4‐1.9
Film 3 2.2 1.7‐2.0 0.9 0.6‐0.8 0.5 0.3‐0.5
Optimal settings
Film 1 165°C_0.7 s_4.0 N mm−2 151°C_1.0 s_1.0 N mm−2 150°C_1.0 s_1.0 N mm−2
Film 2 144°C_1.0 s_1.0 N mm−2 161°C_1.0 s_4.0 N mm−2 147°C_1.0 s_2.0 N mm−2
Film 3 182°C_0.7 s_2.7 N mm−2 182°C_0.4 s_1.2 N mm−2 157°C_1.0 s_3.4 N mm−2

Abbreviation: CI, confidence interval.


8 BAMPS ET AL.

FIGURE 4 A, Raw images of coffee (left) and blood powder (right) contaminated samples of film 2, sealed at 181.5°C, 0.7 s, and 1.0 N mm−2 with
a high resolution digital imaging set up with LED backlight illumination for high contrast. These raw images are converted to B, binary images
where clean areas are black

on seal strength with clean seals. Previous research on clean seals of solid contamination regarding seal strength. These results are
described the very limited influence of seal pressure on seal inconsistent with the good hot tack performance (low seal initiation,
strength.17,18 With ground coffee contamination, a slight effect of high peak value, and wide window) of this film. One possible explana-
pressure is seen as the first order and t*p term of film 1 and T*p term tion would be that for this film, only the lower part of the seal layer
of film 3 have significant effects. With blood powder contamination, participates in the encapsulation of the particles. The lower thickness
there is only a significant effect of the T*p term on seal strength for (5‐μm acid copolymer +10‐μm ionomer) of the effective seal layer
film 1. Higher significance is observed for temperature and time and compared with the particle size could decrease the seal through con-
the combination of both parameters on clean and contaminated seal tamination performance. Another possible explanation can lie within
strength. This result is in line with previous studies that state that the flow behaviour of the seal material. In previous research,9 flow
temperature and time are the most important factors influencing seal ability was related to the encapsulation of milk powder particles.
strength.17,18 These parameters were used in process windows These particles can be isolated if the seal material can flow around
within which at least 90% of the maximum seal strength is obtained. them. Both topics can be interesting for further research to gain bet-
The process windows for the three films are shown in Figure 5. Films ter insight into the clean and contaminated seal performance of pack-
1 and 2 have the widest process window when seals are clean, and aging films.
process windows become narrow when solid contamination is pres- All samples that were optimized in seal strength are tested for
ent. Process windows for clean and coffee contaminated seals are their leak tightness by a dye penetration test prior to the seal
wider for film 2 than for the other films. Even at low seal times of strength test. Samples are tested in tenfold. All clean optimized sam-
0.5 to 0.6 s, it is possible to produce strong seals if the temperature ples were leak tight. In the case of coffee contamination, films 1 and
is set at 170°C. In an industrial context, this is an advantageous film 3 have, respectively, 8 and 7 out of 10 leak tight samples. All sam-
property with respect to production speed. Blood powder contami- ples of film 2 were leak tight at optimal settings. In the case of blood
nated seals need higher seal time to produce strong seals. Taken into powder contamination, film 1 has 9 out of 10 leak tight seals, and
account the optimal values of Table 7 and the process windows of films 2 and 3 are leak tight at the optimal settings. Comparing the
Figure 4, film 1 (seal layer blend of LDPE and mLLDPE) is less toler- three seal layers, the plastomer‐based seal layer in film 2 has the
ant for solid contamination than film 2 (seal layer blend of LDPE and best seal through contamination performance regarding leak tight-
plastomer) regarding seal strength. The results of this comparison are ness at optimal settings. A previous study15 suggests that viscous,
in line with the comparison of hot tack performance between both hot tack, and mechanical properties of seal materials are related with
films (lower initiation, wider window, and higher peak value for film the encapsulation of solid contamination. Low zero shear viscosity, a
2). For film 3, all process windows are narrow compared with films high hot tack strength window (=area under hot tack curve between
1 and 2. There is almost no overlap between the process windows. hot tack initiation temperature and actual seal jaw temperature), and
Taken into account the seal strengths of Table 7 and the process high resistance to elongation under stress were beneficial for
windows of Figure 5, this film has the worst tolerance for these types preventing leaks.15
BAMPS ET AL. 9

FIGURE 5 Temperature vs time process windows for clean, coffee, and blood powder contaminated samples. The process windows indicate
those combinations of temperature and time that result in a seal strength of at least 90% of the optimum value. Pressure was kept constant at
2.5 N mm−2

5 | C O N CL U S I O N S Innoveren & Ondernemen (VLAIO‐TETRA nr. 150817)) and German


government (German Federal Ministry for Economic Affairs and
In this study, a method to optimize the granular contaminated seal Energy (BMWi, IGF project no. 172 EBR)).
strength of packaging films is presented. The optimal values predicted
by the response surface method are experimentally validated. Pre- ORCID
dicted values are a good indication of clean and contaminated seal Bram Bamps https://orcid.org/0000-0002-8992-5098
strength, although there is a tendency of overestimation by the model.
Augmenting the initial experimental design or including repetitions in RE FE RE NC ES
the design could improve this. Besides giving optimal values at one
1. Hishinuma K. Heat Sealing Technology And Engineering For Packaging:
specific parameter setting, process windows for clean and contami- Principles And Applications. Lancaster: DEStech Publications inc;
nated seals can be obtained by doing a limited amount of tests. These 2009:30‐42 ISBN: 978–1–932078‐85‐5.
process windows are highly relevant for practical use in industry. Solid 2. Tauschitz B, Washüttl M, Wepner B, Tacker M. MAP‐Verpackungen:
contamination causes a decrease in the maximal seal strength and nar- ein Drittel nicht optimal. PACKaktuell, DE. 2003;04:6‐8.

rows down the parameter region of time and temperature in the pro- 3. Dudbridge M, Turner R. Seal integrity and the impact on food waste.
cess window where 90% of that maximal strength is obtained http://www.wrap.org.uk/sites/files/wrap/Household_food_and_drink_
waste_in_the_UK_‐_report.pdf, date of access:22/11/2018. WRAP
compared with clean seals. When an equal mass of coffee and blood
2009. ISBN: 1–84405–430‐6.
powder is applied, blood powder has a more negative impact on the
4. Morris B. The Science And Technology Of Flexible Packaging: Multilayer
maximum strength than coffee powder. Films From Resin And Process To End Use. Amsterdam: Elsevier inc
The influence of variation in seal layer composition on the seal 2016; 2017:216‐218 249–250. ISBN: 978–0–323‐24273‐8.
through contamination performance of packaging films is evaluated. 5. Simpson MF, Presa JL. (1996): A comparison of AFFINITY plastomers
The film with the plastomer‐based seal layer outperformed the other produced via INSITE Technology with ethylene/acrylic acid copoly-
films with a higher seal strength, wider process windows, and a higher mers, ionomers, and ULDPE. In: Future‐Pak 1996, Chicago, USA.

degree of leak tightness (evaluated with the dye penetration test). This 6. Mesnil P, Rohse N, Arnauts J. Halle, RW. (2000): Seal trough contami-
film also has a better clean seal performance than the other ones. nation performance of metallocene plastomers. In: TAPPI polymers,
laminations, & coatings conference 2000, Chicago, USA.
Regarding the use of hot tack as a predictive test for the contaminated
7. Morris B. Sure ways to reduce package sealing failure. http://www2.
seal strength, it can be concluded that there are similarities in the com-
dupont.com/Packaging_Resins/en_US/assets/downloads/white_
parison of films with metallocene and plastomer‐based seal layer, such papers/Reduce_Package_Sealing_Failure_Barry_Morris_November_
as the hot tack initiation temperature, window, and peak value. The 2010.pdf, date of access: 22/11/2018. Dupont 2010.
hot tack results of the film with sodium ionomer, however, were not 8. Bach S, Thürling K, Majschak J‐P. Ultrasonic sealing of flexible packag-
predictive for the contaminated seal strength. ing films—principle and characteristics of an alternative sealing method.
The influence of seal technology, jaw geometry, and effective Packag Technol Sci. 2011;25(4):233‐248. https://doi.org/10.1002/
pts.972
thickness and flow behaviour of the seal layer on contaminated seal
performance can be subjects of further research. 9. Sadegi F, Ajji A. Application of single site catalyst metallocene polyeth-
ylenes in extruded films: effect of molecular structure on sealability,
flexural cracking and mechanical properties. Can J Chem Eng 2014.
2014;92(7):1181‐1188. https://doi.org/10.1002/cjce.21964
ACKNOWLEDGEMEN TS
10. Bilgen M, Van Dun J. Hermetic sealing of flexible packages. In: Interna-
This research was performed within the CORNET project tional polyolefins conference 2014, Houston, USA.
“EVOCOSEAL: Evaluation and Optimization of Contaminated Seal 11. D'huys, K, Bamps B., Peeters R, De Ketelaere B. Multi‐criteria evalua-
Performance for Food Packaging,” funded by the Flemish (Agentschap tion and optimization of the ultrasonic sealing performance based on
10 BAMPS ET AL.

design of experiments and response surface methodology. Accepted 17. Theller HW. Heatsealability of flexible web materials in hot‐bar sealing
for publication in packaging technology and science. applications. J Plast Film Sheeting. 1989;5(1):66‐93. https://doi.org/
12. Antony J. Design of Experiments for Engineers and Scientists. Amster- 10.1177/875608798900500107
dam: Elsevier; 2003 ISBN: 9780750647090. 18. Stehling FC, Meka P. Heat sealing of semicrystalline polymer films. I.
13. Jensen WA. Confirmation runs in design of experiments. J Qual Technol. Calculation and measurement of interfacial temperatures: effect of
2016;48(2):162‐177. https://doi.org/10.1080/00224065.2016.11918157 process variables on seal properties. J Appl Polym Sci.
1994;51(1):89‐103. https://doi.org/10.1002/app.1994.070510111
14. Sierra JD, Noriega MD, Nicolais V. Effect of metallocene polyethylene
on heat sealing properties of low density polyethylene blends. J Plast
Film Sheeting. 2000;16(1):33‐42. https://doi.org/10.1106/YYFG‐
9KH1‐R7QU‐VK9H How to cite this article: Bamps B, D'huys K, Schreib I,
15. Ward M, Li M. Seal‐through‐contamination and caulkability—an evalu- Stephan B, De Ketelaere B, Peeters R. Evaluation and optimi-
ation of sealants' ability to encapsulate contaminants in the seal area.
zation of seal behaviour through solid contamination of heat‐
In: Tappi place conference 2016, Fort Worth, USA.
sealed films. Packag Technol Sci. 2019;1–10. https://doi.org/
16. Suenaga S, Hirasawa E. Connection between hot tack force and form‐
fill‐seal performance of packaging films. JSTD. 2004;16(7):450‐458. 10.1002/pts.2442
https://doi.org/10.4325/seikeikakou.16.450

You might also like