Food analysis Packaging Materials

Further Reading

AOAC (1990) AOAC Ofﬁcial Methods of Analysis, 15th

edn. Arlington, VA: AOAC.

Codex Alimentarius (1969) Codex Alimentarius Sampling

Plan for Prepackaged Foods. Arlington, Virginia CAC/
RM 42-1969.

Hulme AC (1970) The Biochemistry of Fruits and Their

Products. London: Academic Press.

Kirk RS and Sawyer R (1991) Pearson’s Composition and

Analysis of Foods, 9th edn. London: Longman Scientific
and Technical.

Kratochvil B and Taylor JK (1981) Sampling for chemical

analysis. Analytical Chemistry 53(8): 924A–938A.

Nagy S and Attaway JA (1980) Citrus Nutrition and

Quality. Washington DC: American Chemical Society.

Pomeranz Y and Meloan CE (1987) Food Analysis, Theory

and Practice, 2nd edn. New York: AVI.

Packaging Materials

A W Lord

, Pira International, Leatherhead, UK

This article is a revision of the previous-edition article by Philip
Tice, pp. 3698–3706,

& 1995, Elsevier Ltd.

Introduction

Analytical measurements on food packaging materi-
als are generally carried out for three main purposes:

To identify the components of the packaging.
To identify and measure substances present that

could migrate into the packaged foods and cause a
health hazard to the consumer of the food. This work
is often accompanied by measurements of the migra-
tion of particular substances into either the actual
packaged foods, or into food simulants.

To identify and measure substances present that

could migrate into the packaged foods and result in
adverse effects on the organoleptic properties, such
as taste or odor.

Food Packaging Materials

The main categories of basic materials used for food
packaging are:

plastics,

regenerated cellulose ﬁlms,

paper and board,

metal, and

glass.

Of these, plastics are the most widely used and with-
in this category there are also the largest numbers of
variants. Many packaging materials are, however,
multilayered with either different layers of plastics or
combinations of plastics with paper/board, metal, or

glass. The individual properties of the different ma-
terials are used to produce food packaging with the
required characteristics. For example, in a packaging
material with two layers of different plastics, one
layer might provide the basic strength whilst the
other layer enables the packaging to be easily heat-
sealed. Coatings are also often added to the basic
plastic packaging material to provide additional bar-
riers to the permeation of oxygen and water vapor.
These coatings can be polymeric or vacuum depos-
ited aluminum.

With some metal cans used for foods and

beverages there is an inner lacquer (plastics) coating
for the purpose of either preventing corrosion of
the metal by the food/beverage, or preventing con-
tamination of the food/beverage by the can metal.

A combination of a polymer layer with a board is

used to package liquids such as milk, where the
plastic layer provides the barrier to contain the milk
within the package and the board the basic strength.
Where it is necessary to store the beverage for long
periods, such as fruit juices, additional barrier prop-
erties are required to prevent permeation of oxygen
into the food product. To achieve this additional
protection, an aluminum layer is incorporated within
a plastic/board composite.

With many of the multilayer packaging materials

adhesives are used to bind the layers together. The
printing on the outside is a further important com-
ponent of food packaging.

Identifying the Components of
Packaging Materials

It is often necessary to identify or conﬁrm the basic
composition of the packaging materials. This applies
particularly with plastics due to the range of polymer
types that are used. Six major polymer types are used
for packaging and these are shown in Table 1, with

FOOD AND NUTRITIONAL ANALYSIS

/ Packaging Materials

341

typical uses. In addition to these basic polymer types,
various copolymers are also used. For example, eth-
ylene is copolymerized with vinyl acetate to produce
ethylene vinyl acetate, styrene is copolymerized with
both acrylonitrile and butadiene to produce the ter-
polymer ABS.

On simple, one- or two-layer structures, identiﬁ-

cation of the polymer type is conveniently achieved
using

Fourier

transform

infrared

spectroscopy

(FTIR). Each of the major polymer classes or copol-
ymers has unique infrared spectra and are easily
identiﬁed by comparison of the spectra to reference
spectra.

The infrared spectrum of a plastic packaging ma-

terial is most easily obtained when the sample is in
the form of a ﬁlm sufﬁciently thin to allow the in-
frared spectrum to be obtained through the ﬁlm.
However, with most plastic packaging, even those
that are used in the form of ﬁlms, the thickness is
usually too great to obtain a good transmission spec-
trum. With some ﬁlm materials stretching can ade-
quately reduce the thickness. Alternatively, it may be
possible to produce a solvent solution and cast a ﬁlm
of the required thickness. Thin ﬁlms may also be
pressed from the sample by careful melting on a hot-
plate. Caution should be exercised with melt pres-
sing, as apart from polymer degradation, there is the
risk of altering the structure, for example, by skewing
a polymer layer away from the region of analysis.

Infrared spectra may be obtained from surfaces

using a variety of techniques. These included atten-
uated total internal reﬂectance (ATR) and specular
and diffuse reﬂectance. These techniques involve the
infrared beam passing through only the outer few
micrometers of the sample. The most widely appli-
cable is ATR. A typical two-layer plastic material
used for lidding on plastic food trays consists of po-
lyethylene and poly(ethylene terephthalate) bound
together with an adhesive. ATR infrared spectra of
the two surfaces will easily identify one surface as

polyethylene and the other surface as poly(ethylene
terephthalate) by the very different spectra obtained.
With the standard KRS5 (thallium bromide/iodide)
ATR crystal the depth of penetration of the infrared
radiation is a few micrometers. This is less than the
individual polymer layers and consequently the ad-
hesive is not ‘shown’ in either of the ATR spectra.

Where the polymer material is a copolymer it is

often possible to obtain a measurement of the re-
lative amounts of the various monomer components
from an infrared spectrum. For example, with an
ethylene–vinyl acetate copolymer the relative heights
of absorption bands from both the ethylene and vinyl
acetate are measured and ratioed with the spectrum
recorded in the absorbance mode. The most conveni-
ent absorbance bands are: 720 cm

for polyethyl-

ene and 1235 or 1740 cm

for vinyl acetate.

Copolymers of known composition are required for
calibration. It is possible to obtain an assessment on
the butadiene and acrylonitrile contents in styrene/
butadiene/acrylonitrile

copolymers.

The

bands

usually used are: for styrene 1600 cm

, for acrylo-

nitrile 2240 cm

, and for butadiene 996 cm

Modern packaging materials are very often mul-

tiple-layered structures. If the packaging material is a
laminate or coextrusion each layer will produce an
infrared spectrum. The resulting composite spectrum
becomes difﬁcult to interpret. In most cases lami-
nates are manufactured using adhesive to bond the
layers together. It is sometimes possible to select a
solvent to dissolve the adhesive thereby enabling
the individual polymer layers to be separated. The
separated polymers can then be identiﬁed by their
infrared absorption spectra. Spectra from a polyeth-
ylene/poly(ethylene terephthalate) laminate and the
separated layers are shown in Figures 1A–1D.
Polyurethane-based adhesives are widely used to
bond poly(ethylene teraphthalate) to polyoleﬁns.
Hot benzylalcohol is a good solvent for a range of
polyurethanes. Other solvents include tetrahydro-
furan and chloroform for acrylate-based adhesives.
This approach also enables the adhesive to be
identiﬁed.

An approach that can be applied to laminates and

coextrusions is to selectively dissolve and remove
polymer layers by careful selection of solvents. Thus,
the nylon layer in a polyethylene/nylon/polyethylene
coextrusion can be isolated by boiling in xylene. Al-
ternatively, the nylon could be removed by boiling
in formic acid. Solvents for the selective removal of
polymers are listed in Table 2. Acids or alkalis should
be avoided on some polymers where there is a risk of
reaction with the polymer. An example would be the
use of concentrated sodium hydroxide solution on a
metallized ﬁlm comprising certain acrylic/ethylene

Table 1

Types of plastics used for food packaging and typical

uses

Plastic type

Typical uses

Polyethylene

Bags and bottles

Polypropylene

Wrapping ﬁlms and pots

Poly(vinyl chloride),

unplasticized

Trays, bottles, and containers

Poly(vinyl chloride), plasticized

Wrapping ﬁlm and cling ﬁlm

Polystyrene

Trays, pots, and containers

Poly(ethylene terephthalate)

Lidding ﬁlms and oven

containers

Polyamide (nylon)

Laminated with polyethylene,

‘boil-in-bag’ pouches

342

FOOD AND NUTRITIONAL ANALYSIS

/ Packaging Materials

ionomer types, where marked alterations to the
infrared spectra can result from such treatment.

It is sometimes difﬁcult to obtain thickness meas-

urements on layers due to swelling of layers with the
solvent or partial break up of thin layers. If the den-
sity of the polymer is known or measured the thick-
ness of a layer may be calculated from the weight of
the polymer.

Although infrared spectroscopy is a useful tech-

nique for identifying polymers in packaging materi-
als it is important to emphasize that as a result of the
development of packaging technology, it is now

rarely possible or wise to use the technique on its
own without recourse to other analytical techniques.
There has in recent years been a trend toward the use
of coextrusions, and away from adhesive bonded
laminates. Coextrusions of a wide range of polymers
can be produced using thin tie layers (a few microm-
eters) of polymer with compatibility for the different
polymer types. For example, polypropylene and eth-
ylene vinyl alcohol can be extruded into ﬁlms and
bottles. Layers within coextrusions cannot be readily
separated using solvents. In addition, the time and
cost constraints upon the analysis of coextrusions
mean that physical isolation of layers is often
impractical.

The most efﬁcient approach to establishing the

structure of packaging is a combination of optical
microscopy, differential scanning calorimetry (DSC),
and infrared spectroscopy. The ﬁrst stage in estab-
lishing the construction of an unknown packaging
material is to subject it to the following optical mi-
croscopy techniques. A section (typically 5–10

mm) is

cut from a 10

10 mm area using a microtome. It is

important that the sample is held rigid but strain free
and cut with a very sharp knife. The best knife for
packaging material is usually a freshly made glass

Figure 1

Infrared transmission spectra of: (A) polyethylene/poly(ethylene terephthalate) laminate; (B) separated poly(ethylene

terephthalate) layer; (C) separated polyethylene layer plus adhesive; and (D) separated polyethylene layer with adhesive removed.

Table 2

Solvents for plastics

Plastic type

Solvents

Polystyrene and copolymers

Chloroform, ethyl acetate,

ketones

Polyethylene

Decalin, hot toluene

Polypropylene

Decalin, hot xylene

Poly(vinyl chloride)

Tetrahydrofuran,

cyclohexanone

Poly(ethylene terephthalate)

o-Chlorophenol, trichloroacetic

acid

Polyamide

Formic acid, phenols

FOOD AND NUTRITIONAL ANALYSIS

/ Packaging Materials

343

knife. Thicker sections (

42 mm) require use of a

steel knife. It is sometimes beneﬁcial to cool samples
below their glass transition temperature (T

) in order

to provide a more rigid structure to section. It may
also be necessary to mount samples in potting resin
for support prior to sectioning through the sample
and resin.

The section is then examined under polarized light

(using cross-polars). A tint plate is useful to provide
color differentiation of layers. This enables the
number and thickness of each layer to be estab-
lished. Thickness measurements are made by calcu-
lation after measuring the total thickness of the
sample using a micrometer. The ratio of total thick-
ness to layer thickness is calculated in arbitrary units
using the scale graduations on the microscope. The
thickness of each layer is then calculated from the
ratios knowing the true total thickness. An example
of a typical cross-section of a coextrusion is shown
in Figure 2.

Information on the composition of individual lay-

ers in the structure can then be obtained by observat-
ions of the layer under the microscope. Different
types of polymers have a recognizable morphology.
Polypropylene has very large spherulites distinct
from many other polymers. Figure 3 shows a
photograph obtained from polypropylene cooled
slowly from a melt. Compounded poly(vinyl chlo-
ride) (PVC) can also be quite characteristic due to the
different phase regions arising from the presence of
impact modiﬁers as well as pigment specks from col-
or adjusters. Calcium carbonate ﬁller and talc anti-
blocking agent have recognizable morphologies. It is
also possible to determine whether a layer contains

recycled process scrap. It is quite common for off-
cuts to be recycled in an inner layer of a coextrusion.

Infrared spectra are then obtained from the sur-

faces of the packaging material after solvent removal
of any print or lacquer. Spectra are best obtained by
ATR. A further portion of the packaging material is
then subjected to DSC. This is a technique where a
few milligrams of the sample is subjected to a
programmed temperature ramp in a speciﬁed atmos-
phere inside a sample chamber. The heat ﬂow (pow-
er) to the sample is monitored against temperature as
the sample is subjected to the heating ramp. For the
purposes discussed here this provides a trace showing
the melting points of the polymers present. Typical
melting ranges for common packaging polymers are
tabulated in Table 3.

The technique cannot be used to obtain melting

points for amorphous polymers. The sample polymer
is heated and cooled and then reheated at a control-
led rate to record the meting points. This procedure
removes hysterisis effects that may be present in the
polymer as a result of the manufacturing process and
which may alter the perceived melting point. DSC is
capable of identifying polymers and polymer blends

Foil

Polyurethane
adhesive

PET

Polyurethane
adhesive

Nitrocellulose
lacquer/print

Surlyn

(LD-MD) PE

Ethylene-co-acrylic
acid (ionomer)

Figure 2

Cross-section of a coextrusion viewed through a tint

plate on an optical microscope. PET

¼ Poly(ethylene terepha-

late), (LD-MD) PE

¼ ? (Reprinted with permission from Mr R

Musgrove, Pira International.)

Figure 3

Polypropylene spherulites viewed through a tint

plate. (Reprinted with permission from Mr R Musgrove, Pira
International.)

Table 3

Typical melting ranges for common polymers

Polymer

Melting range (

1C)

Linear low-density polyethylene

115–130

Low-density polyethylene

100–115

Ethylene vinyl acetate

100–110

Polyamide

210–260

Poly(ethylene terephthalate)

240–260

Poly(vinylidene chloride)

220

Polypropylene

160–170

Ethylene propylene random co polymer

149

344

FOOD AND NUTRITIONAL ANALYSIS

/ Packaging Materials

not readily identiﬁable in packaging materials using
infrared spectroscopy. Examples of these include,
low-density polyethylene, linear low-density poly-
ethylene, high-density polyethylene, and blends of
these polymers. The blend ratios of these polyoleﬁns
can be estimated after calibration using the pure
polymers.

The construction of the packaging material is then

determined by comparing the data obtained from all
the analytical techniques. Any layers that are difﬁcult
to identify are then identiﬁed by either detailed ana-
lysis on the isolated layer using FTIR and other
analytical techniques, or by applying additional
analytical techniques to the whole structure. For
example, pyrolysis gas chromatography–mass spec-
trometry (GC–MS) can conﬁrm the presence of a
styrene/acrylate copolymer adhesive or vinylidene
chloride/acrylate copolymer coating. The pyrolysis
causes depolymerization, often to the starting mon-
omers, which are then identiﬁed from their mass
spectra. The technique can be applied to the whole
construction. This is useful when FTIR analysis is not
conclusive, or where the layer cannot be isolated for
FTIR analysis. In the absence of a pyrolyser instru-
ment, it is possible to perform the technique by
briefly heating the sample in an inert atmosphere in a
sealed headspace vial over a gentle Bunsen burner
ﬂame. Static headspace GC–MS analysis of the
pyroloysate is then carried out. Hot stage micros-
copy is a particular useful technique for identifying
layers. A key advantage of the technique is the ability
to identify layers that cannot be isolated for analysis.
In this technique, the packaging material cross-
section is heated at a controlled rate under the mi-
croscope. The melting ranges of the individual layers
can be observed and compared with the melting
ranges observed by DSC. It should be noted that
there is a difference of a few degrees centigrade in
the melting ranges observed by DSC and hot stage
microscopy:

The strategy for identiﬁcation of a packaging

material construction is summarize:

1. Examine a cross-section by optical microscopy;

determine the number of layers and their thick-
ness, tentatively identify the polymers in the lay-
ers.

2. Obtain a DSC of the whole material. This will

identify all the crystalline polymers present in the
structure.

3. Obtain infrared spectra from the surfaces. This

will conﬁrm the composition of the outer layers.
Isolate fractions of the construction and obtain
their infrared spectra to conﬁrm the identiﬁcations
made by DSC and optical microscopy.

Optical microscopy techniques can also be applied

to packaging failure problems. Sections can be taken
through a heat seal region to establish the integrity of
the seal. Molecular orientation, melt ﬂow, blend
homogeneity, and crystallinity can be observed that
can reveal the cause of stress cracking and other
types of packaging failures.

Analysis of Substances Related to
Food Safety

It is sometimes necessary for technological reasons to
use chemicals that have toxic properties in the man-
ufacture of a food packaging materials. Also, there is
the possibility that some of the chemicals and com-
ponents used for food packaging materials can con-
tain trace levels of toxic contaminants. Where toxic
substances are unavoidably present in a food packa-
ging material for any of these reasons, it is necessary
to ensure that levels of these substances are restricted
so that any transfer to packaged food does not
exceed safety limits. The national regulations of
individual countries control the safety of food
packaging with respect to the substances with known
toxic properties. In some countries the specific re-
strictions are contained in ofﬁcial recommendations
or codes of practice. The primary restrictions are on
the levels that migrate or transfer to the packaged
food and are designated ‘specific migration limits’.
However, in some cases the restriction is a permitted
level in the packaging material, while for others the
restriction is a limit on the quantity, which can be
extracted. As might be expected where the safety of
food is concerned, the set limits are often low re-
quiring sensitive analytical methods.

Food Contact Plastics

Vinyl chloride monomer used for the manufacture
of PVC plastics intended for contact with foods
provides an example where there is a low ‘specific
migration limit’, plus a low limit on the level allowed
to be present in the packaging material. These limits
are contained in an EC Directive 78/142/EEC on
PVC plastics and are: 0.01 milligrams per kilogram
of food (10 ppb), and 1 milligram per kilogram of
polyvinyl chloride.

Vinyl chloride is a gas at ambient temperature and

the ofﬁcial EC analytical methods for both determi-
nations use headspace GC with a ﬂame ionization
detector (FID). Where a determination exceeds the
legislation limit, conﬁrmation is required with head-
space GC using either a different chromatography
column, or a different detector, or with the gas chro-
matograph coupled to a mass spectrometer.

FOOD AND NUTRITIONAL ANALYSIS

/ Packaging Materials

345

For the determination of free vinyl chloride mon-

omer in plastics, the test sample is dissolved or dis-
persed in N,N-dimethylacetamide in a sealed vial and
then equilibrated at 60

1C before sampling the head-

space. When determining vinyl chloride monomer in
foods or food simulants, N,N-dimethylacetamide is
again used with the sample in a sealed vial, with
equilibration at 60

1C.

Other volatile plastics monomers with similar

migration limits, such as acrylonitrile and butadiene,
are also determined by the headspace GC technique.
For the measurement of ‘nonvolatile’ monomers in
plastic food packaging and in foods or food simul-
ants due to migration, high-performance liquid chro-
matography (HPLC) and ion-exchange chroma-
tography techniques can be employed. Food and
food simulants may give rise to interference problems
with the analysis. Sample clean-up procedures are
widely used such as solid-phase extraction of inter-
ference from extracts or size exclusion chromato-
graphy to remove fats and oils. Selective detectors
such as mass spectrometers are now widely used for
both liquid and gas chromatography.

Figure 4 is an ion-exchange chromatogram of

phthalic acids. Terephthalic acid and isophthalic acid
are monomers of polyester plastics. Orthophthalic
acid is the internal standard. There are migration
limits for both terephthalic acid and isophthalic acid
in EC regulations of 7.5 and 5.0 mg per kg of food.
The other main classes of substances in the ‘safety’
category, which can be present in food packaging
plastics and for which analytical measurements are
required, are the plastics additives. These substances
perform the functions of plasticizers, antioxidants,
antistatics, slip agents, and stabilizers. Of these add-
itives, the plasticizers have been most extensively
studied and analytical methods developed for their
determination in the plastics and in packaged foods.
The techniques are usually based on GC either with
an FID or with a mass spectrometer as the detector.
Key advantages offered by a mass spectrometer are
selectivity of response and the ability to add a de-
uterated internal standard to the sample to compen-
sate for the incomplete and variable recovery of the
analyte in the analysis.

Migration testing of plastics packaging prior to use

for compliance with any legislation limits is usually
carried out with food simulants rather than actual
foods. First, the analytical task is more often simple
and, second, testing with a food simulant or simul-
ants for a class of foods covers use with all foods in
that class. The food simulants are simple liquids that
represent different classes of foods. For foods where
the liquid phase is largely water, distilled water is
used as the simulant. For acidic foods (typically pH

4.5 or less) the simulant is an acetic acid aqueous
solution, and for alcoholic beverages and other foods
containing alcohol, the simulant is an ethanol aque-
ous solution with strengths more or less equal to the
alcoholic concentration in the beverage or food.

Selecting a simulant for foods containing fats and

oils has not been easy. In the USA, n-Heptane is
speciﬁed as the fatty food simulant in the Food and
Drug Administration (FDA) regulations, although it
is now recognized to give migration levels of specific
substances well in excess of those that occur with the
foods even after applying a suitable reduction factor.
In Europe, olive oil has been selected as the fatty

Figure 4

Separation of terephthalic acid (TPA), isophthalic acid

(IPA), and orthophthalic acid (OPA) by ion-exchange chro-
matography. (From Ashby et al., 1992.)

346

FOOD AND NUTRITIONAL ANALYSIS

/ Packaging Materials

food simulant with alternatives of sunﬂower oil or a
synthetic triglyceride, known as HB307, developed
specifically for migration testing. Alternatives to
olive oil were considered necessary as it was known
that various analytical problems arise with olive oil.
The food simulants required for regulatory migration
testing in the EC Member States on plastics packa-
ging intended for use with foods, are listed in EC
Directives 97/48/EEC and 85/572/EEC. These are
shown in Table 4 together with the classes of foods
and beverages that they represent. The 85/572/EEC
Directive also contains a table that speciﬁes the
simulant or simulants to be used for individual
categories foods and beverages, and the 97/48/EEC
Directive gives the test conditions of time and tem-
perature corresponding to the intended conditions of
use. Fatty foods contain various amounts of oils and
fats. For those that have high oil or fat content the
extent of migration of substances from the plastics is
often higher than for those with low oil or fat con-
tents. Consequently, reduction factors are applied to
migration values obtained with olive oil to allow for
the various levels of oil and fat in particular foods.

It has been generally accepted that migration from

packaging materials into dry foods will be low com-
pared to moist, liquid, or fatty foods. Accordingly,
more work has been completed in developing simul-
ants for these foods than dry foods. At the present
time the only regulatory or generally recognized si-
mulant for dry foods is Tenax (poly(2,6-diphenyl-p-
phenylene oxide)). Tenax has been adopted as a dry
food simulant because it is a dry porous polymer
with a large surface area that exhibits high-adsorp-
tion characteristics for a wide range of volatile
organic compounds. Tenax is therefore considered to
be a worst-case dry food simulant for a wide range of
dry foods. Other food simulants have been investi-
gated for use with paperboard. A semisolid food
simulant consisting of a mixture of diatomaceous
earth, water, and olive oil has been used, as well as
ﬁlter paper impregnated with olive oil.

In addition to the requirement to measure the

levels of individual specific substances that have
migrated from plastic food packaging into foods or
food simulants and have toxic properties, there is
sometimes also the necessity to measure the total
quantity of substances that are likely to transfer from
the plastics to the food. This is called overall migra-
tion or global migration. The tests are carried out
with food simulants as they are impracticable with
foods. No attempt is made to identify the nature of
the substances that have migrated from the plastic
material. In the current EC ‘food contact’ plastics
regulations there is an overall migration limit of
60 mg of substances from the plastics material per
kilogram of food simulant (60 mg kg

), or ex-

pressed alternatively as 10 mg of plastic substances
from 1 dm

surface area of plastics test specimen

(10 mg dm

). With the three aqueous based food

simulants – distilled water, acetic acid, and ethanol
solutions – the overall migration is measured by de-
termining gravimetrically the nonvolatile residue in
the simulant following exposure to the plastic test
specimen. The values normally obtained with these
aqueous-based food simulants are usually well below
the regulation limit and in the region of 6–
18 mg kg

(1–3 mg dm

). As the tests are most

often carried out on test specimens with a surface
area of 1–3 mg dm

the total quantities of migra-

ting substances are typically a few milligrams. Con-
sequently, care has to be taken with the gravimetric
measurement to ensure reliable results are obtained.
After evaporating the simulant to dryness the
nonvolatile residue is dried in an oven at 110

1C un-

til constant weight is obtained. It has been found,
however, that particular care has to be taken when
using glass evaporating dishes to ensure that there is
adequate time allowed for both the heating period in
the oven and the cooling period in the desiccator for
the mass to stabilize before each weighing. With
metal evaporating dishes the mass of the dish and
residue stabilize more quickly, but with the acetic
acid simulant it is necessary to use dishes made of
platinum, or a metal with similar chemical resist-
ance, to prevent additional errors from corrosion
products.

To measure overall migration with olive oil or alter-

native simulants, the method used with the aqueous-
based food simulants is obviously not applicable.
With oil-type simulants the test is carried out by
measuring the loss in mass of the test specimen after
exposure to the food simulant. However, most plas-
tics absorb some of the oil that then has to be ex-
tracted and quantitatively measured before the true
loss in mass can be calculated. The extraction solvent
that has been most commonly used in the past is

Table 4

EC food simulants for migration tests and the corre-

sponding classes of foods

Food simulant

Class of food or beverage

Distilled water

Aqueous foods and beverages

3% (w/v) aqueous solution of

acetic acid

Foods and beverages with

pH 4.5 or less

10% (v/v) aqueous solution of

ethanol

Foods and beverages with

15% or less alcohol

Olive oil or sunﬂower oil or

HB307

Foods containing fats and oils

For a food or beverage with an alcohol content greater than 10%

(v/v), a simulant with a similar ethanol concentration is used.

FOOD AND NUTRITIONAL ANALYSIS

/ Packaging Materials

347

1,1,2-trichlorotriﬂuoroethane, as it is a good solvent
for olive oil but does not dissolve most plastics. As
this solvent is a chloroﬂuorocarbon its main use and
supply is being phased out and pentane and diethyl
ether are now used instead for the extractions.

Once the extraction has been completed the olive

oil is quantitatively analyzed by hydrolyzing to the
fatty acids, methylating to form the methyl esters,
and measurement by GC. The value obtained is sub-
tracted from the mass of the test specimen after ex-
posure to the olive oil, to allow the overall migration
value to be calculated. The test method has a rep-
utation of poor precision and reliability not only due
to the complex procedure, described above, but also
due various other factors that are known to inﬂuence
results, such as moisture absorption by the plastic
test specimens, incomplete extraction of the olive oil,
and change in composition of the extracted olive oil.
The overall and specific migration analytical test
methods have been established as Standards by the
European Committee for Standardisation (CEN).
Reference plastics are also available with certiﬁed
migration values in olive oil (Institute for Reference
Materials and Measurements BCR, Geel, Belgium,
sales@irmm.jrc.be).

Paper and Board

Most paper and board food packaging materials are
not used in direct contact with liquid foods and con-
sequently migration tests with liquid food simulants
are not considered appropriate. Paper and board
packaging does, however, come into direct contact
with various moist and fatty foods where migration
of substances into the food can sometimes occur. No
simulants have yet been selected to specifically rep-
resent these classes of foods in migration tests on
paper and board. To ensure that the paper and board
material is suitable and safe to package foods, ex-
traction tests are often carried out. The extraction
test is performed with cold or hot water, or some-
times with dilute acids and solvents and is considered
to be a more severe test than a migration test. An-
alytical measurements are then carried out for
specific substances on the extraction liquid. For ex-
ample, tests are carried out for free formaldehyde,
which can arise from wet-strength additives of the
melamine formaldehyde or urea formaldehyde types.
A typical analytical method for the determination of
free formaldehyde in the extracts is based on colori-
metric procedures using chromotropic acid or ace-
tylacetone (pentane-2,4-dione). With paper and
board products there is concern that there could be
toxic contaminants present that in turn could trans-
fer to the food when used as packaging. These

possible contaminants include: the toxic heavy ele-
ments arsenic, mercury, lead, cadmium, and chro-
mium, plus chlorophenols and polychlorinated
biphenyls (PCBs). With extraction liquids such as
water or dilute aqueous acid solutions, the toxic
heavy elements can be analytically determined using
atomic absorption spectroscopy. Arsenic can be
measured with the hydride generation technique,
mercury with the cold vapor technique, and the other
metals by the standard ﬂame technique. Inductively
coupled plasma atomic emission spectroscopy is now
also widely used. Pentachlorophenol and other
chlorophenols can be determined by either GC or
HPLC. When using GC the chlorophenols are best
derivatized to form the methyl or acetyl derivatives
in order to improve the chromatographic perform-
ance and the analytical precision. These analytical
techniques have also been used in the detection and
analysis of chlorophenols suspected of being respon-
sible for odors and food tainting, as described later.
The PCBs are determined in extracts from paper and
board materials by GC with an electron capture de-
tector or mass spectrometer.

Two compounds are currently of particular inter-

est in paper and board. Diisopropyl naphthalene
(DIPN) is a mixture of isomers that until recently
were widely used in carbonless copy papers as ink
solvents. Although it is currently being replaced it
occurs as a persistent contaminant in recycled paper
and board. Various studies have shown that it is able
to migrate from paperboard into food. There is a
draft CEN analytical method available. This method
involves acetone extraction and quantiﬁcation by
GC–MS using diethyl naphthalene as an internal
standard. There is currently no limit for DIPN but
levels are being monitored to reduce concentrations
in recycled paperboard.

Two related compounds are 3-monochloropro-

pane-1,2-diol (3-MCPD)

and

dichloropropanol.

These arise in paper board due to the hydrolysis of
epichlorohydrin-based wet strength agents. 3-MCPD
can occur in food from hydrolyzed vegetable protein.
The limit in food is 120 ppb. In packaging the spe-
cific migration limit is 12 ppb in the food. The di-
chloropropanol does not at present have a limit.
However, the German BGVV recommendations
(widely accepted as useful guidelines) list a limit of
2 ppb in a hot water extract. A convenient method of
analysis is to extract the two compounds with water.
The water extract is then totally absorbed onto a
diatomaceous earth cartridge. The cartridge is then
washed with a large volume of diethyl ether. The
water is retained on the cartridge and the 3-MCPD
and dichloropropanol extracted and eluted by
the diethyl ether. The ether is then concentrated by

348

FOOD AND NUTRITIONAL ANALYSIS

/ Packaging Materials

evaporation and the two compounds derivatized and
injected for analysis by GC–MS.

Metal Packaging

Cans are widely used to pack food. In some cases
tinplated steel cans are used, for example, for
packing fruit. Predominantly, the cans are internally
coated with a polymeric coating to prevent corrosion
or food spoilage. A considerable amount of work has
been done in recent years investigating the extent to
which compounds present in the lacquers migrate
into food.

Attention has focused on the chemical compounds

bisphenol A, BADGE, and BFDGE. Bisphenol A is
manufactured from the reaction of phenol with
acetone. The bisphenol A is further reacted with epic-
hlorohydrin to produce bisphenol A diglycidyl ether
(BADGE). BADGE is then polymerized and cross-
linked in a stoving process to produce an epoxy phe-
nolic coating that has high chemical and mechanical
resistance. These coating are called bisphenol
A-epoxies.

Alternatively, phenol may be reacted with formal-

dehyde to generate bisphenol F. Unlike bisphenol A,
bisphenol F is a mixture of isomers rather than a
discreet compound. Bisphenol F can be subjected to a
condensation process in which a polymeric resin
called Novalac is produced. The Novolac can be re-
acted with epichlorohydrin to produce a polyglcidyl-
ether and these are called novolac glycidyl ethers
(NOGE). NOGE is not used to produce epoxyphe-
nolic coatings, as is the case with BADGE.

Organosol coatings are dispersions of PVC in sof-

tener, solvent, and other resins. Solid contents are
typically 40–80%. The coating is stoved to evaporate
off the solvents and cure the resin. BADGE is often
added as an additive to scavenge for hydrochloric
acid generated from the PVC during curing. Alter-
natively, NOGE is used as an additive instead of
BADGE.

The most widely used lacquer types for food cans,

where the food is retorted in the can to ensure pre-
servation are:

epoxyphenolic and

organosol

Epoxyphenolic lacquers are universally used for both
can bodies and ends for two- and three-piece con-
structions, although more usually for shallow draw
cans. Beverage can bodies are commonly epoxyami-
no coated, and the ‘easy-open’ end and deeper draw
two-piece cans are organosol coated. Coatings may
contain residual BFDGE and bisphenol F arising
from the NOGE in organosol coatings and bisphenol

A as well as BADGE arising from the use of BADGE
in organosol and epoxy phenolic coatings. Possible
residues remaining in the coatings are listed below,
all of which have been found to contaminate the
food:

BFDGE

BADGE

Bisphenol A

Bisphenol F

BADGE and BFDGE undergo hydrolysis and addi-
tion of hydrogen chloride released from the PVC
organosol in aqueous foods and a series of reaction
products result. Concern has been raised over these
reaction products. These are listed below:

BADGE

HCl

BADGE

2HCl

BADGE

HCl

BADGE

These decomposition products result from the ring
opening on the epoxy group of which there are two.
The legislation (Directive 2002/16/EC, February 20,
2002, on epoxy food contact materials) speciﬁes a
migration limit of 1 ppm in the food. This limit is the
total of all the reaction products and BADGE added
to the BFDGE and its reaction products. In addition,
there is a requirement of no detectable migration of
NOGE at a detection limit of 0.2 mg kg

in the

food or 0.2 mg/6 dm

in the can. The decomposition

product BADGE

O in food is ignored as this is

not of toxicological signiﬁcance. However, it must be
included if the migration test is done on food simul-
ants as there is the risk of forcing decomposition
through to the BADGE

O and underestimating

the other compounds. The legislation is due for
review in 2004 as the toxicity of the chlorohydrins is
not at present established.

Analysis of Substances Causing Taint

Taint from food packaging is very rare when one
considers the tonnage of packaged food consumed
each year. Tainting chemical compounds present or
derived from the food packaging are often volatile
compounds. With plastics these odorous volatiles can
be: monomer residues, reaction by-products from
the polymerization process, breakdown products of
certain additives and contaminants. For example,
with polystyrene plastics high levels of styrene mono-
mer produce a very characteristic odor and a number
of incidents of tainting from styrene monomer
have been reported. With polyethylene terephthalate
(PET) plastics, acetaldehyde can be formed during

FOOD AND NUTRITIONAL ANALYSIS

/ Packaging Materials

349

the polymerization process and when PET is used for
beverage bottles the acetaldehyde can cause tainting
of the beverage. With paper and board materials the
volatiles arise mainly from natural lipids and resins
originating from the wood raw material, but some
can come from synthetic resins used in surface coa-
tings that are applied for improved printability and
appearance. The predominant volatiles originating
from the wood lipids and resins are usually alde-
hydes, carboxylic acids, and alcohols. The odors of
some of these aldehydes are not unpleasant being
described as ‘grassy’, but others have ‘rancid’ odors.
Many of the carboxylic acids and alcohols have
strong sharp odors. The synthetic resin binder used in
the surface coatings is typically a styrene/butadiene
copolymer that can contain odorous reaction by-
products such as 4-phenyl cyclohexene. If solvent-
based adhesives are used in sealing the packaging or
to bind layers together, and if the ﬁnished packaging
is printed with solvent-based inks, solvent residues
can add to the list of volatile substances. If the more
odorous of these print solvents and volatile sub-
stances are present in sufﬁcient quantities they can
cause the packaging to be odorous and in turn result
in tainted foods.

Two classes of highly odorous substances that have

been known to contaminate food packaging and
result in food tainting are the halophenols and
the related haloanisoles (methylated chlorophenols
and bromophenols). In the past contamination was
invariably with the chlorophenols and corresponding
anisoles from wooden pallets and surfaces treated
with wood preservers or phenol-based disinfectants.
The ansioles are generated by microorganisms such
as molds from the phenols. Recently, there has been a
noticeable trend toward increased contamination
with the bromophenol and corresponding anisoles.
This reﬂects the substitution of bromophenols
for chlorophenols in wood treatments. The odor
threshold for tribromoanisole in water is 8 pg l

g l

or 8 parts per trillion, ppt). Low-

detection limits are therefore required for such taint
investigations. Concentrations above 1 ppb in the
packaging are often sufﬁcient for tainting to occur.
Polyethylene is the most widely used polymer in
contact with food, usually in the form of a thin inner
layer of a food pack. It therefore only requires con-
tamination of a few sacks of polyethylene granules
with the ansiole to result in tainting of a large
amount of food.

Analytical measurements and investigations are

therefore carried out to detect and measure volatile
odorous substances in food packaging either for
quality control purposes or when odor and taint
problems arise. The technique of choice is GC–MS.

For odor investigation the chromatograph is ﬁtted
with an odor port so that the ﬂow from the analytical
capillary column is split via a T piece to an odor port.
By this means it is possible to smell compounds elu-
ting from the capillary column simultaneously with
their detection by the mass spectrometer. This ena-
bles an odorous compound to be identiﬁed and
quantiﬁed in food and packaging.

Isolation and concentration of tainting compounds

from the rejected food and packaging prior to ana-
lysis is usually the most challenging step in the
investigation. Dynamic headspace sampling is widely
used. In this technique, a sample of the packaging
material is placed in a vessel that is closed, heated to
a temperature of

B701C, and then purged with an

inert gas such as nitrogen or helium. Volatiles re-
leased from the packaging are removed by the purge
gas, trapped, and concentrated on a porous polymer
such as Tenax. Transfer of the volatiles from the po-
rous polymer to the gas chromatograph is performed
by thermal desorption or by solvent elution and in-
jection as a solvent solution. The chromatogram in
Figure 5 shows volatile substances that have been
collected by the dynamic headspace technique from a
printed carton-board that had caused tainting in a
packaged cake. The tainting was attributed to the
benzophenone that appears as the large peak at just
below 16 min. Benzophenone is used as an initiator
in ultraviolet radiation cured printing inks. The peak
at 3.1 min is the aldehyde, hexanal, which originated
from the pulp used to make the board. The cluster of
peaks from

B5.5 to B7 min is volatiles from the

synthetic resin binder in the board coating. None of
these substances produced detectable odors.

The Likens–Nickerson extraction technique can

also be used as a concentration technique, particu-
larly for those volatile substances that are steam
volatile such as the chlorophenols and chloroani-
soles, and also when carrying out an analysis for the
packaging volatiles in foods. The sample is boiled in
a ﬂask with water. Consideration must be given to
the pH of the sample in the water. Basic compounds
will be present as water-soluble involatile salts in
boiling water at low pH, and acids as the corre-
sponding salts in boiling water at high pH. The pro-
cedure is therefore best carried out under basic
conditions and then repeated after acidiﬁcation with
a few drops of nitric acid. The steam is condensed
and continuously extracted with a suitable nonwater-
miscible solvent, any solvent-soluble volatile sub-
stances being transferred to the solvent. After con-
centration of the solvent solution by evaporation of
the solvent with a Kuderna–Danish apparatus, the
analysis is again performed using a gas chro-
matograph coupled to a mass spectrometer.

350

FOOD AND NUTRITIONAL ANALYSIS

/ Packaging Materials

Solid-phase micro extraction is a useful technique

in which volatiles are partitioned from the sample
onto ﬁbers coated with polar or nonpolar bonded
phases. The ﬁber is then placed directly into the
heated injection port of the GC where compounds
are volatilized and carried onto the capillary column.
The technique is less sensitive than the techniques
described above as it is an equilibrium process.
However, modern ion trap mass spectrometers have
increased in sensitivity and the technique is becoming
widely used.

Residual Solvents

Solvent residues from printing and adhesives have
the potential to cause food tainting. Typical solvents
used in printing with characteristic, easily detectable
odors are aliphatic esters, such as ethyl acetate, iso-
propyl acetate, and n-propyl acetate, the alcohols
isopropyl alcohol and n-propyl alcohol and hydro-
carbon mixtures, particularly aromatics. Regular
tests are carried out by printers of food packaging
to ensure that the concentrations of solvent residues
are maintained below the odor and tainting thresh-
old levels. The most widely used analytical technique
to measure the levels of solvent residues is GC head-
space analysis. Portions of the packaging are placed
in sealed vials or other suitable containers and heated

at a set temperature for a prescribed period of time.
The headspace is then sampled by means of a gas
syringe or automatic sampling unit and injected into
a suitable gas chromatograph with an FID. As the
measurements are often carried out for quality con-
trol purposes, short heating times are sometimes used
with external calibration and the measurements do
not always accurately determine the solvent residues
in the packaging, but do give reproducible results.
This is the case with UK, BSI Standard BS6455 –
Monitoring the levels of residual solvents in ﬂexible
packaging materials – and also the corresponding
American ASTM Standard F 151-86. There are two
draft EN standard methods in existence, prEN
13628-1 (absolute method) and prEN 13628-2 (in-
dustrial quality control monitoring method).

See also: Adhesives and Sealants. Food and Nutri-
tional Analysis: Oils and Fats. Infrared Spectroscopy:
Overview; Sample Presentation; Industrial Applications.
Liquid Chromatography: Food Applications. Plastics.
Sensory Evaluation.