PACKAGING TECHNOLOGY AND SCIENCE

Packag. Technol. Sci. 12, 19±28 (1999)

Component Migration From Adhesives Used

In Paper And Paperboard Packaging For
Foodstuffs
By Anita Gruner* and Otto Piringer

This paper presents the ®rst results of research on the migration of adhesive
components in the paper and paperboard packaging sector. It examines
components that can migrate from a variety of commonly used adhesives in the
framework of a `worst case' scenario in respect of the contents. The report also
explains the methodology used to analyse extraction and migration and the model
used for calculating the transfers. Copyright O 1999 John Wiley & Sons, Ltd.
Received 23 March 1998; Accepted 23 September 1998
KEY WORDS: migration; adhesive; food packaging; evaluation; paperboard

INTRODUCTION in very small quantities or because they do not

come into direct contact with the contents. One
Foodstuff packaging can consist of many indivi- example is adhesives used in paper and board
dual components. A single packaging material packaging. Adhesives, in turn, consist of a number
(paper, monoplastic, composite plastic, alumi- of individual components whose migration poten-
nium, tin, etc.) or composite packaging may be tial is examined in this paper.
covered with print, varnish, laminate, coatings and The transfer of substances from plastic packa-
bonded areas. These, in turn, can consist of a ging into foodstuffs is governed by Germany's
mixture of substances whose individual compo- LMBG of 10 April 1992, which enacts EU Directive
nents may contribute to the migration potential of 90/128 (including several amendments). As a
the packaging. result of these regulations many packaging mate-
Packaging must comply with paragraph 31 of rials have already been tested for their migration
the LMBG (Lebensmittel- und Bedarfsgegenstaen- potential.
degesetz), i.e., it must not affect the smell or The German Industrieverband Klebstoffe e.V.
¯avour of the contents. There are numerous has collaborated to select relevant types of adhe-
guidelines, recommendations and rules on the sives to produce typical packaging samples. The
use of individual substances in food packaging. following adhesives were used:
For many substances there are speci®ed areas of
use and limit values, which establish permissible • EVA hotmelt
limits for transfer into the contents of packages. • Dextrin adhesive
There are, however, other substances whose • Starch adhesive
migration potential has not yet been assessed. This • PVAc homopolymer dispersion adhesive
is mainly because, in packaging, they are used only • VAE copolymer dispersion adhesive

* Correspondence to: A. Gruner, Fraunhofer-Institut fuÈr Verfahrenstechnik und Verpackung, Giggenhauser Str. 35, D 85354
Freising, Germany
Contract/Grant Sponsor: Industrievereinigung fuer Lebensmitteltechnologie and Verpackung e.V. (IVLV).

Copyright  1999 John Wiley & Sons, Ltd. CCC 0894±3214/99/010019±10 $17.50
 A. GRUNER AND O. PIRINGER

TEST DESCRIPTION Gas chromatography

Equipment: Hewlett Packard model 6890 with
Premise auto sampler.
Column: DB-1; length: 20 m; inner diameter: 0.18
In order to measure or calculate the migration of mm; ®lm thickness: 0.18 mm.
components from adhesives into contents, we Injection: 5 ml; split: 1:20.
must ®rst establish Ð by extraction analysis Ð Injector temperature: 300°C; detector tempera-
the composition of the adhesive, i.e., the type and ture: 325°C.
quantity of the migratable substances it contains. Temperature programme: 50°C (2 minutes)±
Then we must determine Ð by migration tests Ð 340°C/rate 5° C/minute.
the quantity of components actually migrating. We Carrier gas: hydrogen.
are then able to calculate migration into the
contents from these test results and from data on
the application quantities and adherend surfaces
used in the bonding process. Test results
1. EVA hotmelt
Implementation
The chromatograms of the iso-octane and ethanol
In order to determine the concentration of migra- extracts (Figure 1) show that the main components
table substances in the adhesives, adhesives were of the hotmelt are non-polar and are, therefore,
applied to aluminium foil as the carrier material extracted better by iso-octane. Oligomer hydro-
and extracted with the FhILV-fast extraction carbons and complex paraf®n waxes appear as
method using iso-octane and ethanol.1 The quan- undissolved peaks. The resulting pattern is similar
tities applied were higher than commonly used in for both mediums of extraction, although with iso-
commercial manufacture as the production of test octane the extraction quantities are higher. How-
samples on a laboratory scale made it impossible ever, only the iso-octane extract produces a typical
to achieve a thinner application. distribution pattern of aliphatic paraf®ns, which
For quanti®cation, the extract solutions were are not extracted by the polar ethanol. All relevant
spiked with an internal standard solution consist- substances appear within a retention period of 20
ing of BHA (3-tert butyl-4-hydroxy-anisole) and to 65 minutes which, under these test conditions,
Tinuvin 234, concentrated, and qualitatively and corresponds to a molecular weight range of 200 to
quantitatively analysed by gas chromatography. 700.
In order to determine the amount of substances, In the chromatogram of the tenax migration, the
actually migrating adhesives were applied to less volatile oligomer peaks and the paraf®n
paperboard and analysed with the FhILV tenax pattern are almost imperceptible. Here the high-
method.2,3 This method was selected as adhesives volatile peak groups (4 to 10 minutes and 20 to 28
do not usually come into direct contact with liquid minutes) are predominant. This corresponds to a
foodstuffs and solvent migration is therefore not a molecular weight range of up to 300. The peaks
consideration. To simulate migration from paper- marked `i' are the added internal standards.
board at room temperature, we chose 24 hours at For the tests a hotmelt ®lm with 470 g/m2 was
60°C as the optimum migration conditions. Pre- used. For a quantitative analysis of the tenax
paratory tests had shown that these test conditions migration the total retention period of 1 to 50
best simulated migration at room temperature minutes was integrated as this is the standard
over a longer period. For quanti®cation, the procedure for determining overall migration
migration solutions were spiked with an internal values in the tenax method. Because of the
standard solution consisting of BHA (3-tert.-butyl- complexity of the chromatogram, a division into
4-hydroxy-anisole) and Tinuvin 234, concentrated, individual peaks was not thought to be a mean-
and qualitatively and quantitatively analysed by ingful option. An overall migration value of
gas chromatography. 300 mg/m2 hotmelt ®lm was found.

Copyright O 1999 John Wiley & Sons, Ltd. 20 Packag. Technol. Sci. 12, 19±28 (1999)
COMPONENT MIGRATION FROM ADHESIVES

Figure 1. Gas chromatograms of hotmelt.

2. Starch and dextrin adhesive the borax which is also present in this dextrin
adhesive.
The chromatogram resulting from the tenax
As these two types of adhesive consist mainly of method corresponds to the chromatogram of the
starch and dextrin, respectively, which are not paperboard carrier. There is only one conspicuous,
registered by this method of analysis, the chroma- very-high unidenti®ed peak at a retention time of
tograms of the iso-octane and the ethanol extrac- 17 minutes, which appears only in the tenax
tion of both types of adhesive (Figures 2 and 3) chromatogram of the dextrin adhesive.
show few peaks. The peaks marked `i' are added The starch adhesive was applied to the alumi-
internal standards; the peak appearing at a nium foil in a quantity of 72.5 g/m2 for the ethanol
retention time of 36 minutes is made up of di-iso- extraction and of 82.4 g/m2 for the iso-octane
ethylhexyl-phthalate which, because of its rela- extraction. From this an original concentration of
tively small quantity, is likely to be the result of 1.8 g/m2 can be calculated for the glycerine peak in
solvent contaminations. The very high peak the ethanol extraction. For the tenax migration the
registering at a retention period of 4 to 14 minutes paperboard was coated with 60 g/m2. For the peak
in the ethanol extract of the starch adhesive is appearing at a retention time of 35.5 minutes the
glycerine, which makes up 5 per cent of this quantity of migrated substance was calculated as
sample. Although glycerine also makes up 5 per 2.7 mg/m2, and for various smaller peaks as
cent of the tested dextrin adhesive, the correspond- between 0.3 and 0.6 mg/m2.
ing peak in the ethanol extraction (retention period The dextrin adhesive was applied to the
4 to 8 minutes) is markedly smaller. It is likely that aluminium foil in a quantity of 86.4 g/m2 for the
here the glycerine has already largely reacted with ethanol extraction and of 112.4 g/m2 for the iso-

Copyright O 1999 John Wiley & Sons, Ltd. 21 Packag. Technol. Sci. 12, 19±28 (1999)
 A. GRUNER AND O. PIRINGER

Figure 2. Gas chromatograms of starch adhesive.

octane extraction. From this an original concentra- extraction is mainly identical to that of the ethanol
tion of 57 mg/m2 can be calculated for the extraction.
glycerine peak in the ethanol extraction. For the The three peaks and the softening agent from the
tenax migration the paperboard was coated with extraction also appear in the tenax chromatogram.
100 g/m2. Here an overall migration value of 41 The homopolymer adhesive was applied to the
mg/m2 was found. aluminium foil in a quantity of 105 g/m2 for the
ethanol extraction. This indicates an original
3. Dispersion adhesives concentration of 15 g/m2 for the softening agent
and of 8 and 32.7 mg/m2, respectively, for the
3.1 Homopolymer dispersion adhesive. In the peaks at the corresponding retention times of 14
chromatogram for the ethanol extraction of the and 26.5 minutes.
softened homopolymer (Figure 4), the high peak of For the tenax migration the paperboard was
the softening agent at a retention time of 25 coated with 120 g/m2, which resulted in an overall
minutes stands out. In addition to the standard migration value of 0.57 g/m2.
peaks marked `i' there are three larger relevant
peaks that were not identi®ed. This method of 3.2 Copolymer dispersion adhesive. In the chro-
analysis did not detect a silicone-based defoaming matogram for the ethanol extraction of the un-
agent which, according to the formula, should softened copolymer adhesive (Figure 5) hardly any
have been present. The declared concentration of peaks were found, apart from the marked stan-
the preservative is so low that this method could dard peaks. There is a peak at a retention time of 36
not detect it. The chromatogram for the iso-octane minutes and an undissolved peak cluster can be

Copyright O 1999 John Wiley & Sons, Ltd. 22 Packag. Technol. Sci. 12, 19±28 (1999)
COMPONENT MIGRATION FROM ADHESIVES

Figure 3. Gas chromatograms of dextrin adhesive.

seen at a retention period of 24 to 40 minutes. The CALCULATION OF

peak cluster comprises hydrocarbon-containing
components. Here, too, the declared preservative
TRANSFER
concentration is so low that this method could not
detect it. The chromatogram of the iso-octane We selected ®ve packaging samples, bonded with
extraction is largely identical to that of the ethanol a variety of adhesives and including inner as well
extraction. as outer packaging.
The undissolved peak also appears in the tenax From the analytical test results we calculated the
migration. maximum potential transfer from the packaging to
The copolymer adhesive was applied to the the contents, taking into account the quantity of
aluminium foil at 41.6g/m2 for the ethanol extrac- adhesive applied, the adherend surface and the
tion. This indicates an original concentration of quantity of contents. Our evaluation of the
70mg/m2 for the peak at a retention period of 26 to calculation results also took account of the differ-
42 minutes and 2.6 or 1 mg/m2 for the peaks ence between inner and outer packaging and the
registering at retention times of 36 and 40.5 fact that bonded areas are usually covered.
minutes, respectively.
For the tenax migration the paperboard was Maximal migration quantities: premise
coated with 60 g/m2, which results in an overall
migration value of 33.3 mg/m2. The tenax migration results for the adhesives,

Copyright O 1999 John Wiley & Sons, Ltd. 23 Packag. Technol. Sci. 12, 19±28 (1999)
 A. GRUNER AND O. PIRINGER

Figure 4. Gas chromatograms of homopolymer dispersion adhesive.

which provided the maximum potential migration quantity from the copolymer dispersion adhesive
quantities for a given application quantity, were is relatively small. This is, in any case, the outer
used as the basis for calculating the transfer into packaging and the contents are additionally
foodstuffs. Each of the tenax chromatograms had protected by blister packaging, so adhesive com-
to be reduced by the reference sample value of the ponents have virtually no access to the contents.
untreated paperboard. Table 1 shows the corre-
sponding values. Sample 2: outer packaging for chocolates. A see-
through panel is inserted into the packaging with a
Secondary/outer packaging bead-line of copolymer dispersion adhesive at 60
g/m2 and the packaging ¯aps are glued down
Sample 1: outer packaging for pharmaceutical with spots of hotmelt.
products. The packaging is bonded with a bead- (a) With a bead-line surface of 0.47 dm2, the
line application of copolymer dispersion adhesive maximum potential migration from the copolymer
at 50g/m2. adhesive is 157.4 mg of substance per pack.
With a bead-line surface of 5 cm2, the maximum For a 150g package, the maximum potential
potential migration is 14 mg of substance per transfer is 1 050 mg per 1 kg foodstuff.
package. (b) With a spot bonding surface of 2 cm2 the
For an 84 g pack containing 90 capsules, the maximum potential migration from the hotmelt is
maximum potential transfer is 167 mg of substance 60 mg of substance per pack.
per 1 kg contents or 0.16 mg of substance per For a 150g package, the maximum potential
capsule. transfer is 400 mg per 1 kg foodstuff.
At less than 1 mg/kg, the potential migration Again, this is outer packaging, as the chocolates

Copyright O 1999 John Wiley & Sons, Ltd. 24 Packag. Technol. Sci. 12, 19±28 (1999)
COMPONENT MIGRATION FROM ADHESIVES

Figure 5. Gas chromatograms of copolymer dispersion adhesive.

are in a covered tray, and transfer is possible only maximum potential migration from the hotmelt is
via the vapour area. A potential migration quantity 70 mg of substance per bag.
of 1 050 mg/kg from the copolymer dispersion For a 220 g chocolate content, the maximum
adhesive is not alarming, provided there are no potential transfer is 318 mg of substance per 1 kg of
toxicological restrictions, e.g., speci®c migration chocolates.
limits, for the individual substances. However, (b) With a side seam adhesive bead surface of
such an identi®cation and assessment would be 0.22 dm2, a base adherend surface of 0.56 dm2 and
made only if the adhesive was to be added to a a sight panel adhesive surface of 0.85 dm2, the
`positive' list. Transfers from the hotmelt are maximum potential migration from the copolymer
negligible, as the migration quantity of 400 mg/ adhesive is 454.4 mg of substance per bag.
kg is below the upper limit for paraf®ns currently For a 220 g chocolate content, the total maximum
under consideration by the EU. potential transfer is 2 065 mg of substance per 1 kg
of chocolates.
Sample 3: chocolate bag with see-through/sight The observations made in connection with
panel. The side seam of the bag is bonded with Sample 2, above, apply here too.
bead-lines of hotmelt and copolymer dispersion
adhesives, applied at 50 g/m2 each. The entire Primary/inner packaging
bottom surface is bonded with copolymer disper-
sion adhesive at 50 g/m2 and a sight panel is Sample 4: bakery bag. The bag is bonded with an
inserted into the bag with copolymer dispersion adhesive bead-line at 50 g/m2. Separate calcula-
adhesive at 50 g/m2 tions were made for homopolymer, copolymer and
(a) With a bonded seam area of 0.22 dm2, the dextrin adhesives.

Copyright O 1999 John Wiley & Sons, Ltd. 25 Packag. Technol. Sci. 12, 19±28 (1999)
 A. GRUNER AND O. PIRINGER

Table 1.
Adhesive quantity Applied Overall migrate Speci®c migration values

EVA-hotmelt 50 g/m2 32 mg/m2 Not determined

Starch adhesive 50 g/m2 3.3 mg/m2 2.2 mg/m2 (main substance)
0.7 mg/m2 (unidenti®ed peak)
0.4 mg/m2 (unidenti®ed peak)
Dextrin adhesive 50 g/m2 23 mg/m2 20 mg/m2 (main substance)
0.4 mg/m2 (unidenti®ed peak)
0.1 mg/m2 (unidenti®ed peak)
PVAc-homopolymer 50 g/m2 239.5 mg/m2 237.5 mg/m2 (softener)
dispersion adhesive 0.7 mg/m2 (unidenti®ed peak)
0.7 mg/m2 (unidenti®ed peak)
VAE-copolymer 50 g/m2 27.9 mg/m2 27.3 mg/m2 (main substance)
dispersion adhesive 0.3 mg/m2 (unidenti®ed peak)
0.2 mg/m2 (unidenti®ed peak)
0.5 mg/m2 (other peaks)

(a) With a copolymer adhesive, the maximum application of 50 g/m2 copolymer dispersion
potential migration from a bead-line surface of 0.5 adhesive and the bottom surface is bonded ¯at
dm2 is 140 mg of substance per bag. with 50 g/m2 of hotmelt.
For a 1 kg loaf, the maximum potential transfer (a) With a bead-line surface of 0.2 dm2 and an
is 140 mg of substance per 1 kg of bread. inner bag surface application of 1.2 dm2, the
(b) With a homopolymer adhesive, the max- maximum potential migration from the copolymer
imum potential migration from a bead-line surface adhesive is 391 mg of substance per bag.
of 0.5 dm2 is 1 198 mg substance per bag, or per 1 kg Based on a 225 g rusk content, the maximum
of bread. potential transfer is 1 738 mg of substance per 1 kg
(c) With a dextrin adhesive, the maximum of rusk.
potential migration from a bead-line surface of (b) With a bead-line surface of 0.1 dm2 and a
0.5 dm2 is 115 mg of substance per bag, or per 1 kg base adherend surface of 0.56 dm2, the maximum
of bread. potential migration from the hotmelt is 212 mg of
Calculations were made for the use of three substance per bag.
different types of adhesive. The bakery bag is Based on a 225 g rusk content, the maximum
primary packaging, but it is used only for brief potential transfer is 938 mg of substance per 1 kg of
storage or transport periods. A transfer of adhesive rusk.
components cannot, however, be ruled out. This sample is primary packaging. Here again
The potential transfers for dextrin and copoly- the observations made in Sample 2, above, about a
mer dispersion adhesives are very small. When a potential transfer apply. Again, the migration
homopolymer dispersion adhesive is used, the quantity from the hotmelt remains below the
softening agents may transfer at a rate of up to 1 prescribed upper limit. Because of the low volati-
188 mg/kg. Again, in the current example, using a lity of the paraf®ns and their use at room
phthalate softening agent, the transfer quantity temperature, any actual transfer should be limited
remains below the speci®ed migration limit for to a fraction of the potential transfer quantity.
phthalates of 3 mg/kg.
Use of adhesives in the catering sector
Sample 5: double-layered rusk bag. The bag is
bonded with bead-lines of hotmelt and copolymer In addition to their typical use in packaging,
dispersion adhesive, applied at 50 g/m2 each. The adhesives are also used as laminating agents
inner bag is ®xed in place with an overall surface between the two layers of a composite, for

Copyright O 1999 John Wiley & Sons, Ltd. 26 Packag. Technol. Sci. 12, 19±28 (1999)
COMPONENT MIGRATION FROM ADHESIVES

example, for disposable tableware where several preparation used. If an unsoftened adhesive is
layers of paper, paper board or cellulose are used, or a softening agent for which there is no
laminated together. This inevitably involves a speci®c migration limit value, the maximum
considerably larger bonded surface than common permitted transfer is controlled by the overall
for primary or secondary packaging, so a higher migration limit. It is the packer or laminator's
transfer of components cannot be ruled out. responsibility to select a suitable adhesive.
A laminated paper plate, as commonly used at
fast food stands, was chosen for a sample calcula-
tion.
Cellulose paper and paper board are bonded CONCLUSIONS
over the entire surface with 8±11 g/m2 of homo-
polymer PVAc adhesive. The new LMBG prescribes a limit of 60 mg per 1 kg
For a plate surface of 2.8 dm2, the maximum on transfers of substances contained in packaging
potential migration is 1 463 mg of substance per materials to foodstuffs. Our analysis shows that,
plate. even in a `worst case' scenario, components from
On the standard convention that 1 kg of food- the adhesives tested are well below this limit
stuff is packed in 6 dm2 of packaging, the maxi- value.
mum potential transfer is 3 135 mg of substance per In addition to the overall migration, speci®c
1 kg of foodstuff (=3.1 mg/kg). migration limits apply to a number of substances.
A migration analysis of the paper plate found 6 In our tests of such registered substances, the
mg of substance/dm2. For the model calculation, measured quantities lay below the speci®c limit
following the convention, this equates to a transfer values.
of 36 mg substance per 1 kg of foodstuff. It can be assumed that the migration values in
The discrepancy between the measured and the commercial packaging would be markedly lower,
estimated transfer value could be due to the fact as adhesives are generally applied to covered
that the application quantity used as the basis for seams or surfaces and direct contact with food-
the model was 10 times higher than is common in stuffs is extremely rare.
practice. In the tenax determination this may have
led to a distribution balance between the adhesive
ACKNOWLEDGEMENTS
and the tenax which could have distorted the basis
for the model calculation. We wish to express our thanks to the project's ®nancial
Despite this dif®culty, this transfer is not critical sponsors, the Industrievereinigung fuer Lebensmittel-
from the point of view of the overall migration technologie und Verpackung e.V. (IVLV) and the
value which permits a transfer of 60 mg/kg, Forschungsgemeinschaft fuer Lebensmitteltechnologie
und Verpackung (FoGe).
provided the migrating substances are not covered
We are also grateful to the Industrieverband Kleb-
by the prescribed limits for toxicologically relevant
stoffe e.V. for supplying sample materials, additional
substances. ®nancing for the project, and a stimulating discussion of
The substance tested in this case was a phthalate the topic.
softening agent for which there is a migration limit The article was published in The Original German
of 3 mg/kg. This limit is just reached in the model version in the October 1998 edition of Lebensmittel
calculation and far exceeded in the migration Verfahrens-und Verpackungstechnik.
analysis. But it is important to recognize that the
test conditions for the migration analysis did not
correspond to conditions in commercial practice,
where a fast-food plate is generally not totally
covered with food and the contact period is very REFERENCES
brief. Tests under more realistic conditions are
likely to result in lower values.
1. Berghammer, A., Buecherl, T. and Malter, C.
The excessive transfer was caused in this case `Schnellextraktionsverfahren zur Bestimmung der
not by the laminating adhesive, nor by the potentiell migrierfaehigen Substanzen aus ¯exiblen
lamination itself, but by the speci®c adhesive Verpackungen und beschichteten und/oder lack-

Copyright O 1999 John Wiley & Sons, Ltd. 27 Packag. Technol. Sci. 12, 19±28 (1999)
 A. GRUNER AND O. PIRINGER

ierten Metallen', Verpackungundschau, 45 (7), 41±46 3. CEN-Entwurf ‘Test methods for the determination
(1994). of overall migration at high temperatures (up to 175
2. Piringer, O., Wolff, E. and Pfaff, K. ‘Use of high Celsius) using modified polyphenylene oxide as an
temperature-resistant sorbents as simulants for adsorbant,’ prepared by Fraunhofer Institute, 30
testing’, Food Additives and Contaminants, 10 (6), November 1994.
621–629 (1993).

Copyright O 1999 John Wiley & Sons, Ltd. 28 Packag. Technol. Sci. 12, 19±28 (1999)