International Journal of Adhesion & Adhesives 40 (2013) 5663

Contents lists available at SciVerse ScienceDirect

International Journal of Adhesion & Adhesives

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

Determination of partition and diffusion coefcients of components

of two rubber adhesives in different multilayer materials
Cristina Nern a,n, Julio Gaspar a, Paula Vera a, Elena Canellas a, Margarita Aznar a, Peter Mercea b
a
b

Analytical Chemistry Department, GUIA Group, I3A, CPS, University of Zaragoza, Maria de Luna 3, 50018 Zaragoza, Spain
FABES GmbH, Schragenhofstr. 35, 80992 Munich, Germany

a r t i c l e i n f o

a b s t r a c t

Article history:
Accepted 5 July 2012
Available online 21 July 2012

Rubber adhesives both natural and synthetic are widely used in the manufacture of multilayer
laminates, commonly used as food packaging. For this reason, it is very important to identify the
compounds that could migrate from the laminate into the food. Twenty nine compounds were
identied in two different rubber adhesives, some of them with high toxicity levels according to the
theoretical model of Cramer such as benzene 4-cyanocyclohexene and benzene isothiocyanate.
Partition of these compounds between the adhesive and different substrates and diffusion in both
media were experimentally determined obtaining great variability of values depending on the
properties of the compounds and the substrates used. Finally, only three compounds were found to
migrate: benzene 4-cyanocyclohexene, butylated hydroxytoluene and 2-cyclopentyl-1.3.5-trimethylbenzene, with values of migration of 1.79, 22.9 mg/dm2 and 4.57 mg/dm2 respectively. These values
were below the specic migration limits (SML) enforced by the legislation of the European Union (EU)
and the values recommended by Cramer.
& 2012 Elsevier Ltd. All rights reserved.

Keywords:
Adhesives
Natural and synthetic rubber
Diffusion and partition coefcients
Migration

1. Introduction
Natural and synthetic rubber adhesives are elastomeric resins.
Usually, both have excellent peel strength but low shear strength
properties [1].They are generally supplied as solvent or water
solutions and as pressure sensitive tapes.
Natural rubber has been used in adhesives for more than a
hundred years and is highly suitable for the purpose because of its
excellent exibility and good tack retention properties. These
properties make it ideal for pressure sensitive adhesive applications. Natural rubber solutions are usually employed for bonding
paper, felt and textiles in the manufacture of stationery, carpets
and for packaging to attach metal foil to paper or wood. In
general, chemical catalysts or accelerators may be used at
ambient temperature, or heat curing can be used to vulcanize
the natural rubber to improve the strength and temperature
resistance. Natural rubber has good resistance to water, but poor
resistance to oils, organic solvents and chemical oxidizing agents.
Natural rubber adhesives set usually by solvent or water evaporation [1]. One natural rubber adhesive has been studied in
this work.

Corresponding author. Tel.: 349 761 873; fax: 349 762 388.
E-mail addresses: cnerin@unizar.es (C. Nern), pvera@unizar.es (P. Vera),
ecanellas@unizar.es (E. Canellas), marga@unizar.es (M. Aznar),
petermercea@gmail.com (P. Mercea).
0143-7496/$ - see front matter & 2012 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.ijadhadh.2012.07.003

Rubber adhesives can also be synthetic. Synthetic rubber

adhesives can be based on different substances such as butyl,
polyisobutylene, nitrile or styrene butadiene. In this work a
styrene butadiene rubber (SBR) was studied. SBR copolymers
are often used as the primary components in elastomeric adhesives, and they are also used as a major additive to adjust
properties of other resin-based adhesives. They are available in
a variety of formulations. Water dispersions are very common as
solvent solutions or as hot melt formulations. Often tackiers,
plasticizers, antioxidants and other additives are added to SBR
adhesives formulations. SBR adhesives are used as pressure
sensitive adhesives and laminating adhesives for paper, textile,
leather, plastic lms and metal foils [1].
As can be seen a wide variety of substances are added to these
adhesives to improve their properties. When these adhesives are
used for the manufacture of food packaging, these substances can
diffuse trough the packaging and migrate to food what represents
a potential risk for the health of the consumers. For this reason it
is very important to determine the potential of migration from
the adhesives. As it was described in the previous works [2,3], two
main mechanisms affect the migration processes from packaging
to food, partition and diffusion. In the case of multilayers
manufactured with an adhesive bonding two substrates it is
necessary to know the diffusion coefcients in each layer of the
laminate as well as the partition coefcient at each interface of
the laminate-food system. Among these coefcients, the partition
between the adhesive and its substrate KAS, plays an important

C. Nern et al. / International Journal of Adhesion & Adhesives 40 (2013) 5663

role in determining the level of migration from the laminate into

the food. Unfortunately, little is known yet about the KAS coefcients for rubber adhesives. Because of that one of the main aims
of this work was to determine them for natural and synthetic
rubbers and different substrate materials. These diffusion and
partition coefcients will be later used to implement the mathematical models for predicting migration values.
In practice, laminates with adhesives are manufactured not
only using plastic lms as substrates but also often adhesives are
used to stick plastic/paper or plastic/cardboard as well as to stick
paper/cardboard to paper/cardboard. Little is known about the
diffusion of substances which is contained in the rubber adhesives in cardboard or paper. Therefore in this work some results
are reported on the diffusion of several compounds coming from
rubber adhesives through different substrates.

2. Materials and methods

2.1. Reagents and solutions
The standards 4-vynil-1-cyclohexane, 1,2,4-trimethylcyclohexane, ethylbenzene, 4-tert-butyltoluene, phenylcyclohexane,
4-cyanocyclohexene, 1-phenyl-1-cyclohexene, benzene isothiocyanate, styrene, 1,3-diethylbenzene, 1-methylethylbenzene,
butylated
hydroxytoluene,
1-methyl-4-(1-methylpropyl)benzene, toluene, 4-tert-butyltoluene, alkanes standard solution
C8C20 and p-tertbutylphenol were purchased from SigmaAldrich Qumica S.A. (Madrid, Spain). All of them were of
analytical quality. Dichloromethane, methanol and hexane were
purchased from Scharlau Chemie S.A. (Sentmenat, Spain). All of
them were HPLC grade. Puried water was obtained from a MilliQ 185 Plus system from Millipore (Bedford, MA, USA).
Solutions of p-tertbutylphenol of 190 mg/g in methanol
(solution A), 190 mg/g in hexane (solution B), 10 mg/g in methanol
(solution C) and 3 mg/g in hexane (solution D) were used as
internal standard solutions.

57

Laminates with the structure [substrateadhesivesubstrate]

were manufactured in the laboratory. The laminates structures
were as follows:

 Laminate 1: [paperSR (9.671.1 g/m2)paper]

 Laminate 2: [PVCNR (22.8 70.0 g/m2)PVC]
In laminate 1, paper used as substrate was a glossy paper, treated
with a white kaolin for having a smooth surface that makes the
paper more suitable for the printing inks, with a grammage of 66.7 g/
m2, a thickness of 52 mm and a density of 1.284 g/cm3. The adhesive
layer had a thickness of 10 mm and a density of 0.932 g/cm3.
In laminate 2, PVC used as substrate was a PVC-H lm with a
grammage of 44 g/m2, a thickness of 30 mm and a density of
1.458 g/cm3. The adhesive layer had a thickness of 20 mm and a
density of 0.95 g/cm3.
For the application of the adhesive on the substrate at
laboratory scale different Meyer rods with different notchs were
used. After having prepared the laminates, they were passed
through a laminator at 90 1C. Finally, the laminates were cut into
10 cm by 10 cm cut-outs and wrapped in aluminum foil, leaving
them for 1 week to reach the equilibrium.
The grammage of adhesive expressed as g/m2 was obtained by
weight over the dried sample.
2.3. Gas chromatographymass spectrometry
The equipment used was a CTC Analytics CombiPal autosampler
coupled to a HP 6890 N gas chromatography with a mass spectrometer MS 5975B detector, all of them from Agilent Technologies
(Madrid, Spain). The capillary column used was a HP-5 MS (30 m 
0.25 mm  250 mm) from Agilent Technologies (Madrid, Spain). The
oven program was as follows: 40 1C for 2 min, with rate of 10 1C/min
up to 250 1C, maintained for 3 min. The mass detector was set at
SCAN mode (in the range m/z 45350) for the identication of the
compounds and in SIM mode for quantication purposes. Quantication ions are shown in Table 1.

2.2. Samples and laminates

2.4. Optimization of HS-SPME conditions

Two rubber adhesives with very different composition were

studied, a water- based synthetic rubber adhesive (SR) and a
solvent-based natural rubber adhesive (NR). Samples were provided by two adhesive companies. SR was an aqueous dispersion
of carboxylated styrene butadiene copolymer (SBR) with a 53%
solid content and with an aqueous dispersion of acrylate and
butadiene copolymer as a thickening agent. NR was a solution of
natural rubber in an organic solvent. No more precise information
about formulation can be supplied due to condentiality reasons.
All the adhesives were studied in cured form.

The rst step was to select the most appropriate solid phase
microextraction (SPME) ber for the HS-SPME-GCMS analysis.
The ber selection was based on obtaining the maximum number
of peaks and a high signal intensity for the detected peaks.
Five different bers were tested 100 mm polydimethylsiloxane
(PDMS), 65 mm polydimethylsiloxane/divinylbenzene (PDMS/
DVB), 50/30 mm divinylbenzene/carboxen/polydimethylsiloxane
(DVB/Carboxen/PDMS), 100 mm polyacrylate and 70 mm carbowax/
divinylbenzene (CW/DVB). All of them were from Sigma-Aldrich
Qumica S.A. (Madrid, Spain).

Table 1
Identied compounds in the synthetic rubber adhesive, toxicity according to Cramer (Tox), quantication ions (QI), quantication standards, analytical parameters of the
quantication method and initial concentration prole (CP0).
Compounds

Tox

QI

Quantication standard

Calibration range (mng/g)

Linearity (R2)

LOD (npg/g)

RSD (%)

4-Vinylcyclohexene
Ethylbenzene
Styrene
1-Methylethylbenzene
4-Cyanocyclohexene
4-Phenylcyclohexene
Cyclohexane,1-phenyl-3,4-divinyl_1
Cyclohexane,1-phenyl-3,4-divinyl_2
Cyclohexane,1-phenyl-3,4-divinyl_3
Cyclohexane,1-phenyl-3,4-divinyl_4
Cyclohexane,1-phenyl-3,4-divinyl_5

I
I
I
I
III
II
II
II
II
II
II

79.1
91.0
104.1
105.1
79.1
104.1
104.0
104.0
104.0
104.0
104.0

4-Vinylcyclohexene
Ethylbenzene
Styrene
1-Methylethylbenzene
4-Cyanocyclohexene
1-Phenyl-1-cyclohexene
Phenylcyclohexane
Phenylcyclohexane
Phenylcyclohexane
Phenylcyclohexane
Phenylcyclohexane

0.69750.0
0.61350.0
0.27650.0
0.08150.0
62.30250
0.01249.8
0.00750.4
0.00750.4
0.00750.4
0.00750.4
0.00750.4

0.999
0.999
0.997
0.999
0.997
0.996
0.999
0.999
0.999
0.999
0.999

209
18.4
82.73
24.2
18700
3.81,8
2.01
2.01
2.01
2.01
2.01

6.7
4.7
5.4
7.0
10.1
3.4
3.8
3.8
3.8
3.8
3.8

CP0 (mg/g)
3.177 0.31
5.337 0.40
5.487 0.43
3.617 0.42
27.607 1.50
9.407 0.32
0.04 7 0.00
0.697 0.04
0.937 0.05
0.257 0.02
o LOQ

C. Nern et al. / International Journal of Adhesion & Adhesives 40 (2013) 5663

58

The second step was the optimization of the SPME extraction

conditions. Three different parameters were optimized: extraction temperature (Text) from 40 to 80 1C, extraction time (text)
from 10 to 30 min, and desorption time (Tdes) from 1 to 5 min. The
optimization procedure was carried out using an experimental
design by means of Modde software. For this purpose, a response
surface design method (RSM) was used. Final HS-SPME extraction
conditions were: Text 80 1C, text 30 min and Tdes 1 min. The ber
selected for both adhesives was the PDMS ber.

joined, spiked with 20 mL of solution B and concentrated using a

N2 current up to 1 g solution. The PVC substrates were also
extracted following the same procedure as with the laminate.
The calibration curve was prepared in hexane at seven concentration levels in the calibration range shown in Table 2 for
each compound. Three replicates of each point were prepared and
analyzed.

2.5. Identication of the compounds present in the rubber adhesives

and toxicity classication

Partition and diffusion experiments were carried out in laminates 1 and 2.

The partition coefcient of a compound C between the adhesive and the substrate KA,S, can be calculated according to Eq. 1.

For the identication of the compounds present in SR and NR

adhesives, 1 g of pure adhesive was placed in a 20 ml vial and
analyzed by HS-SPME-GCMS.
The identication of the compounds was based on the comparison of the retention time and the mass spectrum with the
pure standards. The identied compounds were then classied
according to the toxicity Cramer rules with the software Toxtree
v1.51 (Ideaconsult Ltd.).
2.6. Determination of the initial concentration prole (CP0)
of rubber adhesives
For the determination of the CP0 in the SR adhesive the
samples were water diluted (1/200 w/w) prior to the analysis.
Three aliquots of 5 g of the water diluted sample were spiked
with 10 mL of solution A and analyzed by HS-SPME-GCMS.
Due to the difculty of nding all the necessary standards,
some of the identied compounds were quantied with standards
that had a similar chemical structure.
The calibration curve was prepared by diluting the standards
in water at different concentration levels. Three replicates of each
point of the calibration curve were prepared and analyzed.
The NR adhesive was not available as pure adhesive without
substrate. For this reason, it was necessary to work with the
laminates for the CP0 determination of this adhesive. The laminate contained the cured adhesive. A liquid extraction of laminate
2 was carried out. For this purpose laminate 2 was cut into small
cut-outs of approximately 0.5 cm  0.2 cm. The rst step was the
selection of the most appropriate extraction solvent. Three
different polarity solvents were used: hexane, methanol and
dichloromethane. For each solvent, 1 g of laminate 2 was
extracted with 6 g of solvent during 24 h at 40 1C. Three consecutive extractions were made on each sample and afterwards
the three extracts were joined in a vial and concentrated using a
N2 current up to 1 g [3]. Final extraction protocol was as follows:
1 g of laminate 2 cut-outs was three consecutive times extracted
with 6 g of hexane during 24 h at 40 1C, the three extracts were

2.7. Determination of the partition and diffusion coefcients

K A,S

C eq A
C eq S

where [Ceq]A and [Ceq] S are the concentrations of compound C

units mg/cm3 in the adhesive and in the substrate at equilibrium respectively.
As it was impossible to separate the substrates from the adhesive
at equilibrium since they were glued, the methodology used by
Canellas et al. [2] and Vera et al. [3] for calculating KA,S was used. In
this methodology the laminate was sandwiched between two substrates, identical to those used to manufacture the laminate:

 Paper[paperSR paper]paper
 PVC[PVCNRPVC]PVC
These sandwiches were placed in a migration cell that consist
of two aluminum plates of 1 dm by 1 dm of surface which can be
tightened together with a controlled torque of 0.8 Nm. The cells
were kept closed in a constant temperature oven at 40 1C for
30 days in order to assure that the compounds reached the
equilibrium concentration in each layer of these sandwiches. All
the experiments were carried out in duplicate. After this period,
cut-outs from the central part of the two sandwiched substrates
(5 cm by 5 cm) were spiked with 10 mL of solution C (for paper
samples) or solution D (for PVC samples), left equilibrate for 24 h
at room temperature and analyzed by HS-SPME-GCMS. Afterwards, the concentration of the compound at equilibrium in the
sandwiched substrate was determined [Ceq]S.
The [Ceq]A was calculated with a simple balance equation
proposed by Canellas et al. [2] and straightforward the KA,S
coefcients were determinate with Eq. (1).
To determine diffusion coefcients of the compounds in the
adhesives and substrates, the laminates 1 and 2 were sandwiched
between 10 sheets of virgin substrates placed at each side of the
laminate, and these sets were kept in the migration cells used
previously for the partition experiments at 40 and 60 1C during

Table 2
Compounds identied in the natural rubber adhesive, toxicity according to Cramer (Tox), quantication ions (QI), quantication standards, analytical parameters of the
quantication method and initial prole concentration prole (CP0).
Compounds

Tox

QI

Quantication standard

Calibration range
(nmg/g)

Linearity (R2)

LOD
(mg/g)

RSD (%)

CP0 (mg/g)

Toluene
1, 1, 2-Trimethylcyclohexane
2-Tert-butyltoluene
1, 3-Diethy-5-methyllbenzene
1-Methyl-4-(1-methylpropyl)-benzene
Benzene isothiocyanate
Butylated hydroxytoluene
2-Cyclopentyl-1, 3, 5-trimethylbenzene

I
I
I
I
I
III
II
II

91.0
69.0
133.0
133.0
119.0
135.0
205.0
159.0

Toluene
1, 2, 4-Trimethylcyclohexane
4-Tercbutyltoluene
1, 3-Diethyl-5-methylbenzene
1-Methyl-4-(1-methylpropyl)-benzene
Benzene isothiocyanate
Butylated hidroxytoluene
Phenylcyclohexane

0.084710.0
0.132910.8
0.06110.3
0.06410.2
0.05210.0
0.076710.3
0.054810.3
0.02310.2

0.995
0.998
0.999
0.999
0.999
0.995
0.999
0.994

0.01
0.04
0.02
0.02
0.01
0.02
0.01
0.01

10.3
9.3
7.8
7.3
8.2
2.2
7.0
1.7

154.6 7 60.3
10.37 2.4
13.6 7 2.0
7.407 1.2
18.5 7 3.1
79.2 7 11.3
165.5 7 24.7
41.7 7 6.4

C. Nern et al. / International Journal of Adhesion & Adhesives 40 (2013) 5663

24 h. After this period, cut-outs of 5 cm by 5 cm from the central

parts of the added 2  10 substrates were produced and they were
analyzed by HS-SPME-GCMS. The results obtained were mean
concentrations of migrants in each of the 2  10 added substrates
at time t, [Ct]S. The same procedure was followed with other
migration cells after 48 h respectively.
Calibration curves for quantifying paper and PVC samples were
built by spiking 5 cm by 5 cm cut-outs of paper and PVC virgin
samples with 10 mL of the compounds, in methanol and hexane
respectively at different concentration levels. Afterwards, the
samples were spiked with 10 mL of solution C (for paper samples)
or solution D (for PVC samples), left equilibrated for 24 h at room
temperature and analyzed by HS-SPME-GCMS.
Knowing for a sandwich made of the laminate and virgin
substrates the CP0 and the KA/S calculated previously, the concentration prole [Ct]S of a compound diffusing from the adhesive
into the stack of virgin substrates can be calculated with an
appropriate numerical solution of the diffusion equation [36].
The tting parameters in these calculations are the diffusion
coefcients DA and DS of the compound in the adhesive and the
substrate materials respectively. The iteration method used to
derive DA and DS from the experimental concentration proles
[Ct]S was presented previously [36]
2.8. Determination of migration results
Migration tests from these two laminates were carried out
using Tenax as food simulant due to these kinds of laminates are
commonly used for dry food packaging and Tenax is recommended for the migration test.
For migration tests the methodology proposed by Canellas and
Vera et al.[2,3] was used. Cut-outs of each laminate, size 1 cm by
8.5 cm, were covered with 0.34 g of Tenax forming an uniform
layer (UNE-EN 14338)[4]. This system was placed inside a Petri
dish and kept in the oven at 40 1C for 10 days. Then Tenax was
extracted two consecutive times with 3.4 mL each time of
acetone, solutions were put together, added with 10 mL of internal
standard solution, concentrated under a stream of N2 to 200 mL
and nally analyzed by GCMS.
To quantify the migration data different solutions of the
compounds at different concentration levels were prepared in
acetone and analyzed by GCMS. Three replicates of each concentration level were analyzed.
The migration was calculated by mg of compound that
migrated to Tenax per dm2 of laminate in contact with it.

3. Results and discussion

SR adhesive is commonly used for manufacturing metallic
complexes on porous media such as the adhesion of foil or
metallized plastic lm and paper (cardboard boxes or trays, food
exible packaging, candies wrapper).
The main function of NR adhesive was the sealing of cardboard
boxes through its application in the form of tape under pressure. It
was used primarily in the secondary packaging, although in many
cases the cardboard boxes are the primary packaging (candy, sweets,
etc.). The materials that NR usually joins are PVC and paper.
The rst step was to identify the substances present in the
adhesives and likely acting as migrants when in contact with food.
3.1. Identication of the compounds present in the rubber adhesives
and toxicity classication
Five different SPME bers with different polarities were used
during the optimization of the rubber adhesive analysis. For both

59

adhesives the best results were obtained with the PDMS ber,
which provided the best sensitivity. The compounds detected
were identied and classied according to Cramer toxicity rules.
These rules are not based on toxicity studies but on the capacity
of a compound for being metabolized by the organism depending
on its chemical structure. Three structural classes have been
established and a maximum daily intake (mg/person/day) has
been proposed as follows:

 Class I (low toxicity): 1.8 mg/person/day.

 Class II (moderate toxicity): 0.54 mg/person/day.
 Class III (high toxicity): 0.09 mg/person/day.
This classication was selected in order to compare the
theoretical toxicity between the compounds identied.
SR adhesives are obtained from processing hydrocarbons. One
of the most commonly used SR is the adhesive used in this work, a
copolymer of styrene and 1,3-butadiene (SBR). Fig. 1 shows the
chromatogram obtained in the HS-SPME-GCMS analysis of the
SR adhesive with a PDMS ber. Fourteen different compounds and
the internal standard (p-tertbutylphenol) were detected, 11 were
identied and three were unknown compounds. Most of the
identied compounds were aromatic compounds. Sulfenamide
accelerators and phenylendiamine antidegradants were not present in the cured adhesive. Table 1 shows the toxicity of the
compounds according to theoretical model of Cramer. One compound was classied as class III toxicity, 4-cyanocyclohexene, a
reagent used as hardening in adhesives and coatings. Six
compounds with class II toxicity were obtained such as 4-phenylcyclohexene and the ve isomers of cyclohexane,1-phenyl-3,
4-divinyl and the rest of the compounds were classied as class I
toxicity. Among the class I compounds, a monomer used in the
formation of SBR polymer, styrene, and some solvents such as
ethyl benzene and 1-methylethylbenzene were detected.
NR is solid polyisoprene which is extracted from a certain type
of plants after it is tapped. Latex comes out of this tree after
tapping and then solid rubber is separated from it. Fig. 2 shows
the chromatogram obtained in the HS-SPME-GCMS analysis of
the NR adhesive with a PDMS ber. A total of 13 compounds were
detected and eight of them were identied. A class III toxicity
compound (Table 2), benzene isothiocyanate and two class II
toxicity compounds were found, a solvent (toluene) and a common
antioxidant (BHT).
A high partition coefcient (KAS) together with low diffusion
coefcients (DA and DS) of the compounds may reduce the level of
migration below the limits of detection. For this reason, it is very
important to evaluate the physicochemical properties of these
compounds and especially to determine these coefcients in the
different materials used in the laminates. Once KAS, DA and DS are
known for a given laminate one can apply mathematical modeling to predict migration into a food or a food simulant.
3.2. Determination of the initial concentration prole (CP0)
of rubber adhesives
Once the compounds were identied, a quantication method
was developed in order to nd out the amount of the compounds
in the cured adhesives.
Table 1 shows the analytical parameters and the concentration of
the compounds found in the SR adhesive. Samples were analyzed
by HS-SPME and analytical parameters were satisfactory in terms
of linearity, limits of detection and reproducibility. For some
compounds like phenylcyclohexane the limits of detection found
were very low 2.01 picogram per gram of hexane (pg/g). Moreover,
concentration of the compounds ranged from 0.04 to 27.60 mg/g.
The highest concentration was found for 4-cyanocyclohexene

60

C. Nern et al. / International Journal of Adhesion & Adhesives 40 (2013) 5663

Fig. 1. Chromatogram of the synthetic rubber adhesive sample analyzed by HS-SPME-GCMS using a PDMS ber, compounds identication and index toxicity ( I: low
toxicity, II: medium toxicity, and III: high toxicity).

Fig. 2. Chromatogram of the natural rubber adhesive sample analyzed by HS-SPME-GCMS using a PDMS ber, compounds identication and toxicity index ( I: low
toxicity, II: medium toxicity, and III: high toxicity).

(27.6 mg/g). This compound was classied in the Cramer class III, the
most toxic according to this classication, so the study of the
behavior of this compound in the laminate is very relevant. Nevertheless, the lowest concentrations were found for the isomers of

cyclohexane, 1-phenyl-3,4-divinyl classied as medium toxicity

according to Cramer rules.
Table 2 shows the analytical parameters and concentration of
the compounds found in the NR adhesive. Analytical parameters

C. Nern et al. / International Journal of Adhesion & Adhesives 40 (2013) 5663

61

Tables 5 and 6 show the partition coefcients of some of the

compounds identied in the adhesives under study and some of
their properties, like their octanolwater partitioning coefcient
(log POW), molecular weight (Mw), and number of hydrogen
acceptors and donors.
As it can be seen, only one compound, cyclohexane,1-phenyl3,4-divinyl, 1R,3trans,4trans)_1 had a higher tendency to remain
in the paper rather than in the adhesive as the Kadhesive/paper was
below 1. Moreover, 4-cyanocyclohexene showed the higher tendency to remain in the adhesive with a Kadhesive/paper 200.7. This
compound had the lowest log POW and therefore it is the most
polar of the substances studied. According to the composition
given in Section 2.2, SR has a polar character so the polar
compounds such as 4-cyanocyclohexene had a preference to stay
in it. This is a relevant data since this substance was classied in
the class III, the most toxic class according the Cramer classication. As a result 4-cyanocyclohexene will have a low tendency to
migrate to the food in contact with the laminate while cyclohexane,1-phenyl-3,4-divinyl, 1R,3trans,4trans)_1 will have a higher
tendency to reach the paper in contact with food. Moreover, it can
be seen that not only polarity but also molecular structure
inuenced the partition coefcients as this value was different
in all the different cyclohexane, 1-phenyl-3,4-divinyl isomers
studied.
Table 6 shows the partition coefcients of some of the
compounds identied in the natural rubber adhesive between
the adhesive and PVC. The adhesive studied is commonly used to
bond PVC in laminates for food contact so it was the reason to
study the partition coefcients Kadhesive/PVC. Table 6 shows that
one of the compounds, butylated hydroxytoluene had a high

were satisfactory in terms of linearity, limits of detection and

reproducibility. For most of the compounds the limits of detection
found were below 0.02 mg/g. Moreover, the concentration of the
compounds ranged from 7.40 to 165 mg/g. Then, in general the
concentration of compounds in NR was higher than those analyzed in SR. The highest concentration was found for BHT
(165.5 mg/g). This compound was classied in the medium toxicity class by Cramer so the study of the behavior of this
compound in the laminate was very important.
3.3. Determination of the partition and diffusion coefcients
Once the concentration of the compounds in both SR and NR
was obtained it was necessary to analyze their amount in the
substrates (paper and PVC) at equilibrium in order to calculate
partition and diffusion coefcients. As a result, an analytical
method using SPME was designed. Table 3 shows the analytical
parameters for this method applied to the analysis of the
compounds in paper and Table 4 shows the analytical parameters
for this method applied to the analysis of the compounds in PVC.
As can be seen, very good limits of detection were obtained for
both. In the case of paper analysis very low limits of detection
(LOD) were reached since phenylcyclohexane had only 0.07 ng/g.
In PVC analysis even lower LOD were reached, as two compounds
(toluene and 4-tert-butyltoluene) had a LOD of 0.01 ng/g.
3.3.1. Determination of the partition coefcients
Partition coefcients between two polymers are mainly inuenced by different factors like polarity, hydrogen bonds formation,
temperature, chemical structure and size of the compound [5].
Table 3
Analytic parameters for the HS-SPME-GCMS method applied to the analysis of paper.
Compound

Linear range (lng/g)

Linearity (R2)

LOD (ng/g)

LOQ (ng/g)

RSD (%)

4-Vinylcyclohexene
Ethylbenzene
Styrene
1-Methylethylbenzene
4-Cyanocyclohexene
1-Phenyl-1-cyclohexene
Phenylcyclohexane

22.40.0231.42
11.80.0112.360
0.002.202.410
0.0043.802.360
0.8722.385
0.004.52.350
0.000.072.360

0.996
0.998
0.992
0.990
0.990
0.999
0.999

6.71
3.53
0.60
1.2
2621.6
1.30
0.02

22.4
11.8
2.20
3.8
8721.7
4.50
0.07

26.7
23.5
18.1
19.7
8.3
6.7
4.4

Table 4
Analytic parameters of the HS-SPME-GCMS method applied to the analysis of PVC.
Compound

Linear range (lg/g)

Linearity (R2)

LOD (lng/g)

LOQ (lng/g)

RSD (%)

Toluene
1, 2, 4-Trimethylcyclohexane
4-Tert-butyltoluene
1,3-Diethylbenzene
1-Methyl-4-(1-methylpropyl)-benzene
Benzene isothiocyanate
Phenylcyclohexane
Butylated hydroxytoluene

0.0226.10
0.8114.37
0.0426.09
0.3006.18
0.2526.26
0.57036.23
0.1856.26
0.2816.20

0.982
0.992
0.993
0.994
0.991
0.998
0.999
0.996

0.01
0.24
0.01
0.09
0.07
0.17
0.06
0.08

0.02
0.81
0.04
0.30
0.25
0.57
0.18
0.28

31.1
34.1
28.9
25.9
21.3
7.0
8.7
8.5

Table 5
Molecular weight (Mw), LogPOW, partition coefcient adhesive/paper (KAS) and diffusion coefcients in paper (DS) and in adhesive (DA) at 2 different temperatures (40 and
60 1C) of the identied compounds in the synthetic rubber (SR) adhesive.
Compound
4-Cyanocyclohexene
4-Phenylcyclohexene
Cyclohexane,1-phenyl-3,4-divinyl,
Cyclohexane,1-phenyl-3,4-divinyl,
Cyclohexane,1-phenyl-3,4-divinyl,
Cyclohexane,1-phenyl-3,4-divinyl,

Mw (g/mol) Log POW

107,1
158,2
1R,3trans,4trans)_1 212,2
1R,3trans,4trans)_2 212,2
1R,3trans,4trans)_3 212,2
(1R,3cis,4cis)_1
212,2

KAS40 1C

DA 40 1C (cm2/s) DA 60 1C (cm2/s) DS 40 1C (cm2/s) DS 60 1C (cm2/s)

1,70 70,27 200,7 7 4,1 1,65E 07

4,04 70,41 27,0 7 0,0 1,69E 07
5,15 70,47
0,17 0,0 6,74E 08
5,15 70,47
4,1 7 0,0 8,67E 08
5,15 70,47
8,9 7 0,0 5,33E 08
5,15 70,47
2,8 7 0,0 8,54E 08

1,13E 06
6,04E 07
5,13E 07
4,66E 07
6,46E 07
5,39E 07

3,18E 08
1,32E 08
1,12E 08
1,16E 08
4,90E 09
7,33E 09

1,87E  07
5,25E  08
3,90E  08
2,67E  08
2,36E  08
3,13E  08

62

C. Nern et al. / International Journal of Adhesion & Adhesives 40 (2013) 5663

Table 6
Molecular weight, LogPOW, number of hydrogen acceptors and donors, partition coefcients adhesive/PVC (KAS) and diffusion coefcients in PVC (DS) and in the adhesive
(DA) at 2 different temperatures of the identied compounds in the natural rubber (NR) adhesive.
Compound

Mw
(g/mol)

Log POW

N1 hydrogen
acceptors

N1 hydrogen
donors

DA (cm2/s,
40 1C)

DA (cm2/s,
60 1C)

KAS40 1C

DS (cm2/s,
401C)

DS (cm2/s,
601 C)

1-Methyl-4-(1-methylpropyl)-benzene
Benzene isothiocyanate
2-Cyclopentyl-1.3.5-trimethylbenzene
Butylated hydroxytoluene

148,2
135,2
184,1
220,3

4,28 70,39
3,18 70,25
5,06 70,6
5,06 70,37

0
1
0
1

0
0
0
1

ND
ND
ND
2,675E  08

ND
ND
ND
1,945E  07

44.9 7 8.0
2.1 7 0.2
43.2 7 7.2
779.9 7 24.6

ND
ND
ND
5.53E  09

ND
ND
ND
4.73E  09

ND: Non-detected compound in diffusion experiments.

tendency to remain in the adhesive as its partition coefcient

Kadhesive/PVC was 779.9. This compound is the only one that had a
hydrogen donor and therefore, it could form hydrogen bonds with
some compounds that have hydrogen donors in the structure of
this adhesive such as butylated hydroxytoluene

3.3.2. Determination of diffusion coefcients

The diffusion coefcients for the SR adhesive and the paper
substrate used to manufacture the laminate 1 are shown in
Table 5. It can be observed that the magnitude of these coefcients depends mainly on the molecular weight of the migrants,
the temperature and the matrix where the diffusion occurs
(adhesive or paper)
The literature has shown that the diffusion of a molecule in a
polymer is related to some characteristics of the polymer (matrix)
such as its molecular weight, degree of crystallinity, glass transition temperature, the temperature of the environment where the
diffusion occurs as well as the size, the shape, the chemical nature
and the polarity of the diffusing molecule[68]. In general the
diffusion coefcients are higher for small molecules and for
polymers with a low crystallinity degree [6,9].
Table 5 shows that, as expected, diffusion coefcients values in
adhesive SR at 40 1C for the compounds with the lowest molecular weight such as 4-cyanocyclohexene and 4-phenylcyclohexene were higher than those of the highest molecules. The similar
trend was found for the diffusion coefcients of the compounds in
the paper substrate.
As shown in Table 5, the temperature is an important factor
affecting to the diffusion in the adhesive and the substrate. The
diffusion in rubber polymers is a thermally active process which
depends on the temperature according to the Arrhenius equation
[8], where the diffusion of the migrants at high temperatures is
facilitated by the expansion related with the increase of free
volume in the polymer produced at high temperatures [10,11]. In
this case, it was shown that diffusion coefcients were signicantly higher at 60 1C than at 40 1C (p o0.01).
Only few publications dealing with the diffusion processes in
paper were found. The diffusion in this medium depends on
properties such as its porosity, tortuosity and permeability
[1214]. Thus, in fact, the DS data listed in Table 5 for paper
should be regarded as macroscopic apparent diffusion coefcients. The diffusion of all compounds in paper (DS) was lower
than in the adhesive (DA) at both temperatures.
Table 6 shows the diffusion of butylated hydroxytoluene in
natural rubber and PVC substrate used to make the laminate 2.
For the rest of compounds, the diffusion in both media could not
be calculated as the compounds did not diffuse under the test
conditions listed above.
Butylated hydroxytoluene was also found in a laminate manufactured with a hot melt adhesive and cardboard as substrate
studied by Vera et al. [3]. Its diffusion in this adhesive was lower

than in the NR adhesive at two temperatures. In contrast, its

diffusion in the cardboard substrate was higher than that in the
PVC substrate also at both temperatures.
Pure PVC is a brittle crystalline polymer with a glass transition
temperature (Tg) of about 70 1C. Diffusion coefcients in rigid PVC
are very low [15]. When PVC resin is mixed with a plasticizer the
polymer becomes more exible, reduces its Tg and at the same
time diffusion coefcients may increase several orders of magnitude [16]. From 20 to 60 1C the PVC substrates used in this work
were highly exible. This indicates that they were plasticized
samples (about 3545% w/w) and implies that their glass transition temperature Tg layed obviously below ambient temperature.
Such PVC samples usually exhibit diffusion coefcients in the
same range low density polyethylene (LDPE). In fact the DS
coefcients reported in Table 6 for butylated hydroxytoluene
are in the same range as its diffusion coefcient in LDPE [17].

3.4. Determination of migration results

Table 7 shows the migrant compounds found for the laminates
1 and 2 using Tenax as food simulant. Migration results were
expressed as micrograms of migrant compound per dm2 of
laminate in contact with the simulant. The analytical parameters
of GCMS for standard solution in acetone used to quantify the
different migrant compounds were also shown.
Only three compounds migrated as 4-cyanocyclohexene, butylated hydroxytoluene and 2-cyclopentyl-1.3.5-trimethylbenzene
with values of migration corresponded to 1.79, 22.9 mg/dm2 and
4.57 mg/dm2 respectively.
In order to study the possible risks of these migrant compounds, migration values were compared with the specic migration limits (SML) stipulated by the EU legislation. Only butylated
hydroxytoluene compound has a SML of 3 mg/Kg [18]. For NR the
migration result was 22.9 mg/dm2, corresponding to 0.14 mg/Kg
of food (assuming a cube with a surface area of 6 dm2 in contact
with 1 kg of food [19]), and therefore this value was below of
its SML.
For the rest of the compounds the estimated daily intake (EDI)
was calculated according to the U.S. Food & Drug Administration
(FDA) equation




mg
mg
EDI
migration value
person  day
kg
 3 kg total food intake per person per day  CF
where CF is the consumption factor, what means the fraction of
the daily diet expected to be in contact with a specic packaging
material (for adhesives CF is 0.14).
As the Table 7 shows none of the migration values of EDI
exceeded the recommended Cramer exposure values for compounds of classes I, II and III.

C. Nern et al. / International Journal of Adhesion & Adhesives 40 (2013) 5663

63

Table 7
Migrant compounds in the synthetic and natural rubber, quantication ions (QI), quantications standards, analytical parameters of the quantications method, values of
migration (mg/dm2), the estimated daily intake (EDI) expressed as mg/(person x day) and the recommended values Cramer exposure values.
Compound

4-Cyanocyclohexene
4-Phenylcyclohexene
Cyclohexane,1-phenyl-3,4divinyl, 1R,3trans,4trans)_1
Cyclohexane,1-phenyl-3,4divinyl, 1R,3trans,4trans)_2
Cyclohexane,1-phenyl-3,4divinyl, 1R,3trans,4trans)_3
Cyclohexane,1-phenyl-3,4divinyl, (1R,3cis,4cis)_1
1-Methyl-4-(1-methylpropyl)benzene
Benzene isothiocyanate
Butylated hydroxytoluene
2-Cyclopentyl-1.3.5trimethylbenzene

QI

Migration EDI (mg/

(lg/dm2) (person 
day))

Cramer (mg/
(person  day))

1.34
0.89
0.77

1.79
o LOD
o LOD

0.0045

0.09

3.74

0.77

o LOD

0.999

3.74

0.77

o LOD

2.5610.3

0.999

3.74

0.77

o LOD

0.1311.5

0.997

2.41

0.04

o LOD

0.1712.3
0.27104

0.996
1

6.81
7.24

0.05
0.08

o LOD
22.9

0.06

0.54

2.5610.3

0.999

3.74

0.12

4.57

0.11

0.54

Quantication
Standard

Calibration
range (lg/g)

79.1
104.1
104.0

4-Cyanocyclohexene
1-Phenyl-1-cyclohexene
Phenylcyclohexane

4.469.56
2.9660.4
2.5610.3

0.996
0.998
0.999

6.73
4.25
3.74

104.0

Phenylcyclohexane

2.5610.3

0.999

104.0

Phenylcyclohexane

2.5610.3

104.0

Phenylcyclohexane

119.0

1-Methyl-4-(1methylpropyl)-benzene
Benzene isothiocyanate
Butylated
hydroxytoluene
Phenylcyclohexane

135.0
205.0
159.0

4. Conclusion
Two rubber adhesives (synthetic and natural) commonly used
in food packaging materials have been studied. Twenty seven
compounds were identied, most of them had a similar chemical
structure. Only two of the compounds determined showed a high
toxicity according to the theoretical model of Cramer, benzene
isothiocyanate and 4-cyanocyclohexene nevertheless only
4-cyanocyclohexene migrated, with a migration value below the
recommended Cramer exposure values as well as the rest of
migrant compounds. Partition coefcients between adhesive and
substrate used to manufacture of the laminate showed a wide
variation of values mainly depending on the nature of the
migrating compound and the substrates used in the laminates
were found.
References
[1] Petrie EM. Handbook of adhesive and sealants. 2nd ed.. New York: McGraw
Hill handbooks; 2000.
[2] Canellas E, Aznar M, Nerin C, Mercea P. Partition and diffusion of volatile
compounds from acrylic adhesives used for food packaging multilayers
manufacturing. J Mat Chem 2010;20:51009.
[3] Vera P, Aznar M, Mercea P, Nerin C. Study of hotmelt adhesives used in food
packaging multilayer laminates. Evaluation of the main factors affecting
migration to food. J Mater Chem 2011;21:42031.
[4] Paper and board intended to come into contact with foodstuffsconditions
for determination of migration from paper and board using modied
polyphenylene oxide (MPPO) as a simulant, U. E. 14338:2004, 2004.
[5] Tehrany EA, Desobry S. Partition coefcients in food/packaging systems:
a review. Food Addit Contam 2004;21:1186202.

Linearity
(R2)

RSD (%)

LOD
(lg/dm2)

[6] Limam M, Tighzert L, Fricoteaux F, Bureau G. Sorption of organic solvents by

packaging materials: polyethylene terephthalate and TOPASs. Polym Test
2005;24:395402.
[7] Dole P, Feigenbaum AE, De la Cruz C, Pastorelli S, Paseiro P, Hankemeier T,
et al. Typical diffusion behaviour in packaging polymersapplication to
functional barriers. Food Addit Contam 2006;23:20211.
[8] Crank J, Park G, editors. Diffusion in polymers. A. Press: Academic Press;
1968.
[9] Escobal A, Iriondo C, Katime I. Organic solvents adsorbed in polymeric lms
used in food packaging: determination by head-space gas chromatography.
Polym Test 1999;18:24955.
[10] Gnanasekharan V, Floros JD. Migration and sorption phenomena in packaged
foods. Crit Rev Food Sci Nutr 1997;37:51959.
[11] Klopffer MH, Flaconneche B. Transport properties of gases in polymers:
bibliographic review. Oil & Gas Science and Technology-Revue De L Institut
Francais Du Petrole 2001;56:22344.
[12] Dury-Brun C, Jury V, Guillard V, Desobry S, Voilley A, Chalier P. Water barrier
properties of treated-papers and application to sponge cake storage. Food Res
Int 2006;39:100211.
[13] Kacem M, Salvador S, Quintard M. Textural characterization of media
composed of compacted pieces of cardboard and polyethylene using a gas
tracer method. Waste Manag 2009;29:6607.
[14] Triantafyllou VI, Akrida-Demertzi K, Demertzis PG. Determination of partition behavior of organic surrogates between paperboard packaging materials
and air. J Chromatogr A 2005;1077:749.
[15] Berens AR. the diffusion of gases and vapors in rigid PVC. J Vinyl Tech
1979;1:8.
[16] Shailaja D, Ahmed SM, Yaseen M. Comparative study of release kinetics of
pheromone from polymer dispensers. J Appl Polym Sci 1997;64:137380.
[17] Mercea P, Piringer O, Barner AL. Plastic Packaging. Wiley-VCH; 2008 525
Appendix I.
[18] Gilbert SG, Miltz J, Giacin JR. Transport considerations of potential migrants
from food-packaging materials. J Food Process Pres 1980;4:2749.
[19] Giacin JR. Factors affecting permeation, sorption, and migration processes in
package-product systems. Food and packaging materials-chemical interac
tions. Ackermann Paul, Jagerstad
Margaretha, Ohlsson Thomas, editors.
Cambrigde royal society of chemistry; 1995. p. 1222.