RIKILT - Institute of Food Safety is part of the international knowledge organisation Wageningen UR (University & Research centre).

RIKILT conducts independent research into the safety and quality of food. The institute is specialised in detecting and identifying
substances in food and animal feed and determining the functionality and effect of those substances.

RIKILT advises national and international governments on establishing standards and methods of analysis. RIKILT is available
24 hours a day and seven days a week in cases of incidents and food crises.

The research institute in Wageningen is the National Reference Laboratory (NRL) for milk, genetically modified organisms, and
nearly all chemical substances, and is also the European Union Reference Laboratory (EU-RL) for substances with hormonal
Examination of packaging materials in
effects.

RIKILT is a member of various national and international expertise centres and networks. Most of our work is commissioned by the
bakery products
Dutch Ministry of Economic Affairs, Agriculture and Innovation and the new Dutch Food and Consumer Product Safety Authority.
Other parties commissioning our work include the European Union, the European Food Safety Authority (EFSA), foreign
governments, social organisations, and businesses. A validated method for detection and quantification

RIKILT Report 2012.007

More information: www.rikilt.wur.nl L.W.D. van Raamsdonk, V.G.Z. Pinckaers, J.J.M. Vliege and H.J. van Egmond
Examination of packaging materials in
bakery products
A validated method for detection and quantification

L.W.D. van Raamsdonk, V.G.Z. Pinckaers, J.J.M. Vliege and H.J. van Egmond

Report 2012.007 May 2012

Project number: 122.71.666.01

BAS-code: WOT-02-004-014
Project title: Development of methods of analysis for animal feed and feed
ingredients

Project leader: T. Zuidema

RIKILT – Institute of Food Safety

Wageningen UR (University & Research centre)
Akkermaalsbos 2, 6708 WB Wageningen, The Netherlands
P.O. Box 230, 6700 AE Wageningen, The Netherlands
Tel. +31 317 480 256
Internet: www.rikilt.wur.nl
Copyright 2012, RIKILT – Institute of Food Safety.
The client is allowed to publish or distribute the full report to third parties. Without prior written
permission from RIKILT – Institute of Food Safety it is not allowed to:
a) publish parts of this report;
b) use this report or title of this report in conducting legal procedures, for advertising, acquisition
or other commercial purposes;
c) use the name of RIKILT – Institute of Food Safety other than as author of this report.

This research was funded by the Dutch Ministry of Economic Affairs, Agriculture and Innovation
(WOT 02 Animal Feed).

Distribution list:
• Ministry of Economic Affairs, Agriculture and Innovation (EL&I; E.R. Deckers, G.J. Greutink)
• Netherlands Food and Consumer Product Safety Authority (NVWA; W. Ooms, R. Theelen,
R. Herbes, M. Pelk, H.A. van der Schee)
• International Association for Feeding stuff Analysis (IAG), Section Feeding stuff Microscopy
(I. Paradies-Severin, F. Wernitznig, G. Frick, J. Vancutsem)
• European Commision (J. Moynagh)

Cover: oil paint impression of remnants of packaging materials.

This report from RIKILT – Institute of Food Safety has been produced with the utmost care.
However, RIKILT does not accept liability for any claims based on the contents of this report.
Summary

Methods for the detection and quantification of packaging materials in feed materials are
necessary for the control of the prohibition of these materials according to Regulation (EC)
767/2009. A method has been developed and validated at RIKILT for bakery products, including
sweet bread and raisin bread. This choice is based on the situation that this category of former
food products has the highest volume for re-use as feed ingredient.

The method of analysis can briefly be summarised as: 1) visual selection of undesired ingredients
which can be identified as remnants of packaging materials, 2) weighing of the selected materials,
3) defatting, 4) dehydration, 5) final weighing, 6) reporting of weight and percentage. In all cases
the total amount of the sample material was investigated, which is usually 500 grams. This
procedure prevents inhomogeneity of the sample to be a problem.

The method for establishing the level of contamination with remnants of packaging materials in
bakery products intended for animal feeding has been validated successfully at RIKILT with a
quantification limit of 0.01% w/w and an average recovery of 95.5% at a contamination level of
0.15% w/w. The standard deviation of the intra-laboratory reproducibility was S W, 0.15 = 0.012%
w/w. The repeatability of the method could not be established because of the semi-destructive
nature of the method. The method is accredited in 2010 by the Dutch Board for Accreditation
(RvA).

RIKILT Report 2012.007 3

4 RIKILT Report 2012.007
Contents

Summary ...................................................................................................................... 3

1 Introduction ............................................................................................................ 7

2 Method development ............................................................................................... 8

2.1 Background ........................................................................................................ 8
2.2 Method development ........................................................................................... 8
2.3 Method description ............................................................................................ 11

3 Results .................................................................................................................. 12
3.1 Method validation .............................................................................................. 12
3.1.1 Description of the experiments .................................................................. 12
3.1.2 A-priori set limits for performance parameters............................................. 13
3.2 Results of the validation experiments ................................................................... 13

4 Discussion en conclusions ...................................................................................... 15

5 Acknowledgement ................................................................................................. 16

References .................................................................................................................. 17

Annex I Raw data validation experiments .............................................................. 18

RIKILT Report 2012.007 5

6 RIKILT Report 2012.007
1 Introduction

In the current practice of food production, proper packaging of materials is provided for assuring
quality maintenance during transport and storage. Article 6 of Regulation (EC) 767/2009 of the
European Parliament and of the Council of 13 July 2009 , states that: ”Feed shall not contain or
consist of materials whose placing on the market or use for animal nutritional purposes is
restricted or prohibited”. In Annex III of this Regulation, packaging from the use of products from
the agri-food industry, and parts thereof, are mentioned as prohibited for placing on the market or
use for animal nutritional purposes.

The main or at least one of the main categories of former food products intended for feeding
purposes (FFPs) is bakery products. This category includes dried and ground meal from bread and
biscuit products. Biscuit meal comprises sweet bread, raisin bread, biscuits, treacle waffles,
chocolate bread (not confectionary), gingerbread, breakfast cereals, crisps, nuts, a.o. The
estimated volume of recycling in the Netherlands is approx. 150,000 metric ton (MT).
Approximately an identical volume is imported from surrounding countries, especially Germany,
which means that an estimated amount of 300,000 MT of bakery products are processed in the
Netherlands to yield animal feed. The processed bread is predominantly packaged, while the
majority of the biscuit products is processed unpacked (ca. 80%) (van Raamsdonk et al., 2011).
The importance of the share of bakery products is also indicated in previous studies (Kamphues,
2005).

From 2005 on, the Netherlands Food and Consumer Product Safety Authority (NVWA) has
conducted the official control of remains of packaging materials in FFPs that are intended to be
used as feed material. RIKILT Institute of Food Safety has carried out the analyses in the
framework of this monitoring program. The majority of these samples (160 out of 243 as analysed
between 2005 and 2010) consisted of dried and ground bakery products (bread and biscuit meal).
Some other categories are sweets (in the form of syrups), chocolate products, and dairy products
(predominantly milk and whey powders). A remainder category consists of a diversity of products,
ranging from vegetable products, potato products, dough for baking, to starch products (van
Raamsdonk et al., 2011).

In the framework of the Dutch monitoring program a method for detection and quantification of
packaging materials in bakery products intended as feed material has been developed and
validated. The scope of the method excludes the detection of remnants of packaging materials in
processed (e.g. pelleted) matrices. This report presents the method and the parameters of its
performance.

RIKILT Report 2012.007 7

2 Method development

2.1 Background
The type of packaging materials is not easy to establish. For food products relevant for reclaiming
and re-use in feed production, mainly plastics, and paper and board packaging materials are used.
Principally, the following types of materials can be present:

- polyolefins (PO, including polypropylene (PP) and polyethylene (PE); [1]),

- polyethylene terephthalate and its copolymers (PET; [2]),
- polystyrene (PS; [3]),
- rigid polyvinylchloride (PVC; [4]),
- regenerated cellulose (RC; [5]),
- paper and board [6],
- aluminium foil [7].

In general, in processed FFP fibres (representing paper and board), plastic (including PP, PE, PET,
PS, PVC, RC), aluminium (aluminium foil as well aluminium coated paper that is usually used as
wrapping of sweets a.o.), and metal (clips) can be recognised. RC should not be classified as
plastic, but distinction is not feasible.

2.2 Method development

The original material is usually fractionated by the factory in a mill with 2 to 4 mm mesh size.
Therefore, the method applies to ground and dried meal from bread and biscuit products. Biscuit
meal comprises sweet bread, raisin bread, biscuits, treacle waffles, chocolate bread (not
confectionary), gingerbread, breakfast cereals, crisps, nuts, a.o. No further processing in the
laboratory, such as grinding, is applied to the samples before analysis.

The basic principle is to select and separate every particle that is not native to the matrix by bare
eye examination. It is necessary for a proper selection procedure to have particles of
approximately the same size. Furthermore, the difficulty of recognising and picking up of particles
depends on their size. Particles smaller than 1 mm are very difficult to handle. The effect of
sieving was tested by examination of eight samples from practice.

Table 1. Description of samples used for the method development.

RIKILT nr matrix Type packaging material

RIK208913 Bread meal Fibres, plastic, aluminium coated paper
RIK210280 Bread meal Fibres, plastic
RIK215359 Bread meal Fibres, plastic
RIK210282 Bread meal Fibres, plastic
RIK211732 Bread meal Fibres, plastic
RIK211735 Biscuit Fibres, plastic, aluminium coated paper
RIK211733 Biscuit Fibres, plastic, aluminium coated paper
RIK214297 Biscuit Fibres, aluminium coated paper

8 RIKILT Report 2012.007

The samples listed in Table 1 were sieved with sieves of mesh sizes of 4.75 mm, 2.0 mm, 1.0 mm
and 0.5 mm, resulting in five fractions. Of every fraction, the non-native particles were selected
and weighted separately. The resulting amounts of the fractions and of the selected portions were
presented in Figures 1 and 2.

800
RIK208913
700
RIK210280

600 RIK215359

500 RIK210282

400 RIK211732

RIK211735
300
RIK211733
200
RIK214297
100

0
> 4,75 > 2,0 > 1,0 > 0,5 rest < 0,5

Figure 1. Share of material of five size fractions of bakery products. X-axis: size limits of the particles of
each fraction. Y-axis: cumulative amount (g) of the fractions.

1.0
RIK208913
0.9
RIK210280
0.8

0.7 RIK215359

0.6 RIK210282

0.5
RIK211732
0.4
RIK211735
0.3
RIK211733
0.2

0.1 RIK214297

0.0
> 4,75 > 2,0 > 1,0 > 0,5 rest < 0,5

Figure 2. Share of amount of remnants of packaging material in five size fractions of bakery products. X-
axis: size limits of the particles of each fraction. Y-axis: cumulative amount (g) of the remnants.

The contribution of the fraction with particles larger than 4.75 mm to the total amount of material
in the samples was negligible (Figure 1). The share of other fractions differed between the
samples, but in all cases their contribution was significant. The major share of fragments of
packaging material was recovered from the three fractions with the larger particles (> 1 mm). The
other two fractions did not contribute to the final amount of packaging materials (Figure 2). It was
concluded that it is not necessary to separate the fraction with the largest particles (> 4.75 mm)

RIKILT Report 2012.007 9

and the fraction with the smallest particles (< 0.5 mm). Only sieves of mesh sizes of 2.0 mm and
1.0 mm need to be applied, resulting in three fractions.

The remaining fragments of packaging materials can absorb a certain amount of fat or grease,
and of water, depending on the type of matrix. In order to establish this effect and to assure that
the selected portion of fragments is a net amount, defatting and dehydrating experiments were
carried out. Four samples of 400 grams were contaminated artificially with 0.6 grams of fragments
of packaging materials (0.15% w/w) and placed at room temperature for a week. The four
samples were sieved, all fragments were selected from the several fractions and weighted (gross
weight). The fragments were defatted in tetrachloroethylene (TCE) for four hours, the remaining
TCE was decanted and the remaining portion was dried at 60 oC for at least four hours. The
remaining material was weighted again (net weight). The fragments were put back in the matrix,
thoroughly shuffled, and kept at room temperature for seven days. After that a second procedure
was started by sieving the sample, selecting the fragments, weighting, defatting, drying and
weighting again. The results are presented in Table 2.

Table 2. Results of two experiments using the same set of four samples for establishing the effect of
defatting and of drying on the remnants of packaging materials in feed.

EXPERIMENT 1 EXPERIMENT 2
Sample Gross weight Net weight of Percentage Gross weight Net weight of Percentage
of fragments fragments recovered of fragments fragments recovered
before after before after
defatting and defatting and defatting and defatting and
drying drying drying drying
Bread 1.22 g 0.7250 g 120.8% 0.82 g 0.7210 g 120.2%
meal 1
Bread 3.22 g 0.6107 g 101.7% 2.79 g 0.5750 g 95.8%
meal 2
Biscuit 1.66 g 0.7107 g 118.5% 1.14 g 0.6490 g 108.2%
meal 1
Biscuit 1.22 g 0.6343 g 105.7% 1.08 g 0.6654 g 110.9%
meal 2

The difference in gross weight between the fragments in the first experiment and the second
experiment is obvious. Apparently the fragments could not absorb the same amount of fat and/or
water. Although the net weights of both experiments show a good reproducibility, the recovered
amounts were between 95% and 121% of the original amount of contaminating material.

The method was improved by decanting over a sieve with a mesh size identical to the fraction
(either 1 mm or 2 mm). The result was that small grains of the matrix, originally adhering to the
fragments of packaging material but released during the defatting process, could be removed
together with the TCE. The clean fragments of packaging material were then collected on top of
the sieve and dried. Several periods of time for the defatting and drying steps were tested,
resulting in an optimization of the method.

10 RIKILT Report 2012.007

2.3 Method description
The description of the method of analysis is:

1. Principally approx. 500 grams of material (former food product) is taken for investigation,
without further pre-treatment. The entire sample is weighted. The material is sieved at mesh
sizes of 2 mm and 1 mm. The fraction with particles larger than 2 mm, the fraction with
particles between 1 and 2 mm and the fraction with particles less than 1 mm are all
weighted individually. The two largest fractions (> 2 mm and 1-2 mm) are investigated
separately for remnants of presumed packaging materials in a large plate with upright
borders. All particles that are not native to the matrix are selected and kept apart in two
separate portions. The selected particles are picked up by a pair of tweezers. If necessary, a
large magnifying glass can be used. These two portions of remnants of presumed packaging
materials are kept separately for the entire process.
2. The two portions of selected material are weighted (gross weight).
3. Defatting: the two portions each are placed in a beaker together with 50 ml of
tetrachloroethylene (TCE) for 10 min. The TCE is decanted over a sieve of the appropriate
mesh size (2 mm or 1 mm, respectively). The remnants are each placed in a sieve tray for
drying overnight.
4. Dehydration: the two portions in the sieve trays are placed in an oven at 60 degrees Celsius
for 4 hours.
5. The two portions are weighted for establishing the final amount (net weight). The amounts,
together with the results of step 1), are used to calculate the percentage (w/w) of the
amount of remnants of presumed packaging materials per fraction and for the entire sample.
6. If necessary and depending on the nature of the sample, selected portions of 1 gram can be
taken from the third fraction with particles smaller than 1 mm for a further investigation
under the microscope for the presence of remnants of presumed packaging materials.
Quantification of these small particles is at least time consuming to get a rough estimate at
the best. During the validation study it was concluded that if any small particles are present
in this third fraction, the share in the total weight is insignificant. If desired the nature of the
remnants found can be established (e.g. paper, board, plastic aluminium, foil etc.).
7. Reporting of net weight and percentage (w/w).

RIKILT Report 2012.007 11

3 Results

3.1 Method validation

Validation of a method of analysis at RIKILT is implemented according to the Dutch guidelines
NEN 7777 and NEN 7779. For a newly developed method it is requested to establish the values for
the parameters reproducibility, accuracy, limit of detection, selectivity and robustness. The
repeatability was not examined due to the situation that the full samples are investigated and that
the method as described is semi-destructive. Duplicate samples and, hence, duplicate results were
therefore not available.

The measuring range chosen is 0.02 – 0.15% w/w. The 0.15% contamination level was based on
unofficial proposals for a reasonable tolerance limit (Kamphues 2005; van Raamsdonk et al.,
2011). The 0.02% level was the assumed limit of detection. The procedure chosen was an intra-
laboratory validation, which means that only the intra-laboratory reproducibility was established.

3.1.1 Description of the experiments

The mentioned guidelines request a series of experiments to be carried out:

- Experiment A: Eight samples from practice (four bread meals, two biscuit meals, two other
matrices within the scope, all contaminated with an unknown amount of packaging
materials) are measured twice for establishing the intra-laboratory reproducibility. The
selected fragments of packaging material were replaced in the sample after the first
examination. The waiting time between the two examinations was one week in order to
allow the fragments of packaging materials to re-assimilate fat and moist from the sample.
- Experiment B: Eight laboratory samples (four bread meals, four biscuit meals) were
artificially contaminated with 0.15% of fragments of relevant packaging materials. The
results were used to establish the accuracy.
- Experiment C: Eight laboratory samples (four bread meals, four biscuit meals) were
artificially contaminated with 0.02% of fragments of relevant packaging materials for
determining the limit of detection.
- Experiment D: Two samples of bread meal were contaminated with 0.15% of fragments of
packaging materials together with 0.15% of chocolate sprinkles, and two biscuit meals with
0.15% of fragments of packaging materials together with 0.15% of raisin fragments. These
samples were used to examine the selectivity.
- Experiments E1 and E2: Two samples of bread meal contaminated with 0.15% of fragments
of packaging materials were defatted in TCE at a shorter time (5 minutes instead of
10 minutes), and two samples of bread meal contaminated with 0.15% of fragments of
packaging materials were dehydrated during a shorter period (3 hours instead of 4 hours).
These experiments were used to establish the robustness.

Experiments B, C, D and E are based on artificially made matrices without remnants of packaging
materials, produced by RIKILT. Adulteration was carried out according to the descriptions of the
experiments. All experiments were carried out during the months November and December 2009.

12 RIKILT Report 2012.007

3.1.2 A-priori set limits for performance parameters

The criteria used for approving the performance of each of the parameters are a-priori set based
on the guidelines or on internal RIKILT standards. The guidelines and standards are originally
developed for chemical method of analysis. Modification is applied when necessary since the
current method is based on visual inspection with a semi-destructive nature. The latter implied
that the parameter repeatability could not be established.

The Dutch guidelines NEN 7777 and NEN 7779 provide equations for calculating whether
performance parameters are within limits, but does not provide the actual values for these limits.
EU legislation provides guidelines (EC, 2002) for setting limits for performance parameters.
A working group of the European Commission developed guidelines for pesticide analysis,
including criteria for performance parameters (EC, 2009b). In addition, International Standard
ISO 5725 (ISO, 1994) has been used to approve further the strategy chosen in this study for
fixing the parameters limits. The result of the justification of the chosen a-priori parameter limits
based on these literature sources is presented in Table 3.

The minimal limits which can be chosen for accuracy or recovery differ for different circumstances.
The range is wider for screening methods (60 - 140%) than for confirmation methods
(70 - 120%; EC, 2009b). In addition, for lower levels of contamination modified ranges might be
used (EC, 2002). RIKILT, based on EC (2002), applies as common procedure to apply a range of
80 - 110%. This range was set as a-priori criterion for the accuracy.

The intralaboratory reproducibility can be defined in relation to the level of contamination (EC,
2002). The given examples range from 1 µg/kg to 1000 µg/kg, which is a factor of 1000 below the
levels in the current study. The reproducibility is defined in EU (2009b) as precision (RSD wR ) and
should be lower than or equal to 20%. This statement, which is applicable for chemical analysis,
cannot clearly be transformed to semi-destructive visual methods of analysis. In the current study
the second analysis of a sample was applied on a version of the test sample that was restored as
good as possible (re-integration of the remnants of the packaging materials in the total sample,
storage for one week to allow assimilation of fat and moist). Under these circumstances a limit of
S w = 0.02 has been used.

The level of quantification is defined (EC, 2009b) as the level where quantified results can be
achieved complying to the criteria for accuracy (70 - 120%) and precision (RSD wR <= 20%). It
was chosen in the current validation study to use a contamination level of 0.02% as presumed
level of detection. The experimental results will show if this limit can be validated as level of
quantification.

For robustness two aspects of the method have been altered. The experimental results are
expected to maintain within the limits for accuracy (80 - 110%).

3.2 Results of the validation experiments

The full data of the five experiments are presented in Annex I. The summary of the results of the
validation experiments with the performance parameters of the method is shown in Table 3.

The standard deviation of the intra-laboratory reproducibility was S W,0.15 = 0.012% w/w. The
relative deviation between the duplicate analyses of each sample was R REL = 32.75%. These

RIKILT Report 2012.007 13

results comply with the a-priori set criterion, and are acceptable considering the qualification of
the method as semi-destructive. The relative deviation of the accuracy for the eight samples at
the contamination level of 0.15% w/w was d T,rel = -4.48, indicating that more than 95% w/w of
the contaminated material was recovered. The recovery in seven out of eight samples was
between 92.7% and 102.1% w/w. The eighth sample showed a lower recovery (86.2% w/w).
Notwithstanding this result, the recovery is well within the set limits. Also at the level of
contamination of 0.02% w/w the relative deviation and the precision are within the set limits,
although the deviation was slightly larger than at the level of 0.15%.

Table 3. Performance parameters of the method for quantification of remnants of packaging materials in
bakery products.

Parameter Result Criterion

Intra-lab reproducibility R rel = 32.75%, S W,0.15 = 0.012% w/w S W = 0.02% w/w
Accuracy 0.15% J = 95.5% (86.2% - 102.1%), d T,rel = -4.48 80 – 110% w/w
Accuracy 0.02% J = 94.3% (85.0% - 105.5%), d T,rel = -5.73 80 – 110% w/w
Quantification limit S AG = 0.001246, AG W = 0.004% w/w AG W = 0.02% w/w
Selectivity (chocolate, raisins) RA = -2.99%, RA = -0.57% 5% *
Robustness (defatting, drying) RA = -0.552%, RA = 0.734% 5% *

*: based on: 86.2% (lowest recovery percentage) - 5% > 80%;

102.1% (highest recovery percentage) + 5% < 110%.

The calculation of the limit of detection was based on the analysis of eight samples conta-
minated at a level of 0.02% w/w. Because of the semi-destructive nature of the method it was
impossible to comply to the guideline by analysing the same sample at a series of subsequent
days. The standard deviation of the eight results was S AG = 0.001246, resulting in a detection
limit of AG W = 3 * S AG = 0.004% w/w (calculation according to NEN 7777). As shown, the
accuracy at the 0.02% level is within the set limits, which allows the conclusion that a reliable
quantification is sufficiently reached at this level. Since the method as presented excludes the
quantification of particles in the smallest fraction (particles < 1 mm), it is reasonable to adjust
the limit of quantification to AG W = 0.01% w/w. RIKILT applies the rule for the limit of
quantification as BG W = 6 * S AG . In the current practice of a visual method, however, it can be
argued that every fragment of the contaminant that is detected will add to the final amount. In
any case, the limit of quantification is well below the criterion of 0.02% w/w (see Table 3).

The relative deviations after adding confusing materials (chocolate sprinkles and ground raisins)
for determining selectivity and after changing some method parameters (shorter drying time and
shorter defatting time) for determining robustness are low in all cases (table 3). Because of the
exclusive nature of the presented method parameters, setting criteria is difficult. A deviation of
5% w/w is set as criterion, since this value limits the results including the deviations between the
range for recovery of 80 - 110% w/w. The results for selectivity and robustness are well within
this criterion.

14 RIKILT Report 2012.007

4 Discussion en conclusions

The proposed method for detection and quantification of remnants of packaging materials in
bakery products intended as feed ingredient is a non-chemical and semi-destructive method.
Guidelines or standard procedures for validation of these methods are currently not available. It is
also important to stretch that European guidelines for intralaboratory validation are not available
as well. As a next best approach the Dutch guidelines NEN 7777 and NEN 7779 are applied and
modified when necessary. In addition, guidelines for setting a-priori limits or performance
parameters have been consulted.

The method is developed in line with other visual methods (see IAG section Microscopy for
examples [8]). Nevertheless, modifications appeared to be necessary with respect e.g. to the
fraction of the matrix with particles smaller than 1 mm and to the cleaning of the particles of the
contaminant. Fibres of paper and board, fragments of plastic, aluminium foil and metal, chips of
plastic clips and metal wires are the types of materials that can be recovered from the samples.
These types are indicated along the way of appearance. The indication "plastic" refer to several
types of packaging materials with each a different chemical background and application (van
Raamsdonk et al., 2011).

The method for establishing the level of contamination with remnants of packaging materials in
bakery products intended for animal feeding has been validated successfully at RIKILT with a
quantification limit of 0.01% w/w and an average recovery of 95.5% at a level of 0.15% w/w.
The standard deviation of the intra-laboratory reproducibility was S W,0.15 = 0.012% w/w. The
repeatability of the method was not established because of the semi-destructive nature of the
method. The method is accredited in 2010 by the Dutch Board for Accreditation (RvA).

RIKILT Report 2012.007 15

5 Acknowledgement

This research was funded by the Dutch Ministry of Economic Affairs, Agriculture and Innovation.
The authors thank the RIKILT colleagues Th. Polman, J. de Jong, T. Zuidema and L. van Ginkel for
their support.

16 RIKILT Report 2012.007

References

European Commission, 2002. Commission Decision 2002/657/EC of 14 August 2002,

implementing Council Directive 96/25/EC concerning the performance of analytical methods
and the interpretation of results. Official Journal L 221, 17.8.2002, p. 8-36.
European Commission, 2009a. Regulation (EC) No 767/2009 of the European Parliament and of
the Council of 13 July 2009 on the placing on the market and use of feed, amending
European Parliament and Council Regulation (EC) No 1831/2003 and repealing Council
Directive 79/373/EEC, Commission Directive 80/511/EEC, Council Directives 82/471/EEC,
83/228/EEC, 93/74/EEC, 93/113/EC and 96/25/EC and Commission Decision 2004/217/EC.
Official Journal L 299, 14.11.2009, p. 18–54.
European Commission, 2009b. Method validation and quality control procedures for pesticide
residues analysis in food and feed. Document SANCO/10684/2009.
ISO, 1994. ISO 5725: Accuracy (trueness and precision) of measurement methods and results,
part 1-6. International Organization for Standardization, Geneva, www.iso.org.
Kamphues, J., 2005. Expert report on Foreign Materials in Feedstuffs. Institut für Tierernährung,
Hannover, Germany, 26 pp.
NEN 7777, 2003. Milieu - Prestatiekenmerken van meetmethoden. Environment - Performance
characteristics of measurement methods. The Netherlands Standardization Institute, Delft,
www.nen.nl.
NEN 7779, 2008. Milieu - Meetonzekerheid. Environment - Measurement uncertainty. The
Netherlands Standardization Institute, Delft, www.nen.nl.
Raamsdonk, L.W.D. van, R. Rijk, G.P.J. Schouten, W. Mennes, G.A.L. Meijer, A.F.B. van der Poel,
and J. de Jong, 2011. A risk evaluation of traces of packaging materials in former food
products intended as feed materials. RIKILT Report 2011.002 (69 pages).

Internet links
[1] http://en.wikipedia.org/wiki/Polyolefins (26 Jan 2011)

[2] http://en.wikipedia.org/wiki/Polyethylene_terephthalate (26 Jan 2011)

[3] http://en.wikipedia.org/wiki/Polystyrene (26 Jan 2011)

[4] http://en.wikipedia.org/wiki/PVC (26 Jan 2011)

[5] http://en.wikipedia.org/wiki/Cellophane (26 Jan 2011)

[6] http://en.wikipedia.org/wiki/Paper_bag (26 Jan 2011)

[7] http://en.wikipedia.org/wiki/Aluminium_foil (26 Jan 2011)

[8] http://www.iag-micro.org/ (7 Dec 2011)

RIKILT Report 2012.007 17

Annex I
Raw data validation experiments

Experiment A: Intralaboratory reproducibility

sample type Sample nr. day 1 day 2 day 3 day 4 day 5 day 8 day 9 day 10 day 11 day 12 day 15 day 16 day 17 day 18
bread meal 1-217774 0.055% 0.057%
bread meal 2-230626 0.123% 0.111%
bread meal 3-229078 0.022% 0.025%
bread meal 4-232700 0.053% 0.052%
biscuit meal 5-233622 0.042% 0.051%
biscuit meal 6-227212 0.027% 0.037%
other 7-227211 0.084% 0.067%
other 8-217778 0.640% 0.598%

Experiment B: Accuracy

sample type Sample nr. day 1 day 2 day 3 day 4 day 5 day 8 day 9 day 10 day 11 day 12 day15 day 16 day 17 day 18
bread meal B1 0.143%
bread meal B2 0.144%
bread meal B3 0.141%
bread meal B4 0.129%
biscuit meal K1 0.147%
biscuit meal K2 0.153%
biscuit meal K3 0.148%
biscuit meal K4 0.139%

18 RIKILT Report 2012.007

Experiment C: Quantification limit

sample type Sample nr. day 1 day 2 day 3 day 4 day 5 day 8 day 9 day 10 day 11 day 12 day15 day 16 day 17 day 18
bread meal B5 0.018%
bread meal B6 0.019%
bread meal B7 0.019%
bread meal B8 0.019%
biscuit meal K5 0.018%
biscuit meal K6 0.021%
biscuit meal K7 0.017%
biscuit meal K8 0.020%

Experiment D: Selectivity

sample type Sample nr. day 1 day 2 day 3 day 4 day 5 day 8 day 9 day 10 day 11 day 12 day15 day 16 day 17 day 18
bread meal* B1 0.143%
bread meal* B3 0.141%
bread meal B9 0.143%
bread meal B10 0.143%
biscuit meal* K1 0.147%
biscuit meal* K4 0.139%
biscuit meal K9 0.147%
biscuit meal K10 0.151%

* values taken from Experiment B on the appropriate days.

RIKILT Report 2012.007 19

Experiments E1 and E2: Robustness

sample type Sample nr. day 1 day 2 day 3 day 4 day 5 day 8 day 9 day 10 day 11 day 12 day 15 day 16 day 17 day 18
(5'TCE) (3 hrs (5'TCE) (3 hrs
drying) drying)
bread meal B11 0.144%
bread meal B12 0.144%
bread meal B13 0.153%
bread meal B14 0.151%

20 RIKILT Report 2012.007

RIKILT - Institute of Food Safety is part of the international knowledge organisation Wageningen UR (University & Research centre).
RIKILT conducts independent research into the safety and quality of food. The institute is specialised in detecting and identifying
substances in food and animal feed and determining the functionality and effect of those substances.

RIKILT advises national and international governments on establishing standards and methods of analysis. RIKILT is available
24 hours a day and seven days a week in cases of incidents and food crises.

The research institute in Wageningen is the National Reference Laboratory (NRL) for milk, genetically modified organisms, and
nearly all chemical substances, and is also the European Union Reference Laboratory (EU-RL) for substances with hormonal
Examination of packaging materials in
effects.

RIKILT is a member of various national and international expertise centres and networks. Most of our work is commissioned by the
bakery products
Dutch Ministry of Economic Affairs, Agriculture and Innovation and the new Dutch Food and Consumer Product Safety Authority.
Other parties commissioning our work include the European Union, the European Food Safety Authority (EFSA), foreign
governments, social organisations, and businesses. A validated method for detection and quantification

RIKILT Report 2012.007

More information: www.rikilt.wur.nl L.W.D. van Raamsdonk, V.G.Z. Pinckaers, J.J.M. Vliege and H.J. van Egmond