Assessing the safety of recycled food contact materials 941 Source control

Source control is one of the most important steps for closed loop recycling of packaging plastics. As mentioned before, especially for the high diffusive polymers, the source control plays the major role because only non-contaminated packages previously used for food packaging should be introduced into the recycling process. In general, feedstock materials for recycling processes can be divided into the four quality classes:

Class 1: materials remaining from production by the manufacturing or converting industry where their history is well known. These materials typically are always under the control of the processor. Provided that good manufacturing practice is followed and contamination can be excluded, this material is as suitable for direct contact with foodstuffs as new material. Class 1 material can be defined as 'post-industrial recycled polymers' and corresponds to US FDA's primary recycling (pre-consumer scrap). Class 2: PCR material which had been used for food packaging for well-known applications and re-collected pure-grade by the recycler, for instance, via a deposit system. This material typically contains only post-consumer food packaging materials. Due to its post-consumer character, the recycler usually does not have complete control of the plastics material over the time period from its first use up to its return.

Class 3: impurified PCR material and possibly mixed plastics which have been used for certain applications outside the food packaging area that enters the recycling feedstream via mixed plastics collection. This material could include packaging materials from non-food packaging applications. Class 4: any class 1 to 3 material which had been chemically reprocessed by depolymerisation into monomers or oligomers from which, after purification, a new polymer has been regenerated.

Classes 2 and 3 correspond to US FDA's category 'Physical reprocessing: Secondary Recycling'. Class 4 corresponds to US FDA's category 'Chemical Reprocessing: Tertiary Recycling'.

Closed loop recycling materials from class 1 and class 4 can be considered to be safe and in compliance with the legal requirements due to the absence of contamination or the high purification effect of the de- and re-polymerisation steps. The materials from class 2 and 3 should have been sorted to a polymer purity of about 99%, which means also that materials from non-food packaging should also be separated. However, for the special case of PET, the FDA also allows non-food PET as an input material as long as the polymer is in compliance with 21 CFR 177.1630.

9.4.2 Challenge test

The cleaning efficiency of a recycling process should be tested with a so-

called 'challenge test'. Deliberately contaminated material with several surrogates is introduced in the recycling process. After the recycling process the residual concentrations of the surrogates are determined. The difference between the input and the output concentration represents the cleaning efficiency of the investigated recycling process in relation to the individual surrogate. Typically, a set of several surrogates is applied which represents the general four categories of compounds: volatile and non-polar, volatile and polar, non-volatile and non-polar and non-volatile and polar. In addition, several chemical functional groups are introduced, i.e., alcohols, esters, aromatic rings or chloro-organic compounds. Table 9.1 summarises some surrogates used for challenge tests.

A general recommendation for which chemical compounds should be applied in the challenge tests is difficult and depends in the end on the polymer type and on the recycling process being investigated. However, the surrogates should be stable during the recycling process and it should be possible to analyse the output material for the residual concentrations of the surrogates. For example, the metal organic compound copper-II ethylhexanoate was used in some challenge tests described in the literature. The cleaning efficiency of the investigated recycling processes was reduced due to the fact that the metal organic compound reacts with copper oxide (CuO), which could not be removed any more. In this case the cleaning efficiency was determined to be lower than in reality. On the other hand dimethylsulfoxide (DMSO) was discussed as a water soluble surrogate. During the high temperature typically used for recycling, DMSO is oxidised to dimethylsulfate. Therefore the determined cleaning efficiency for DMSO is higher than in reality with serious consequences regarding food law compliance evaluation.

The applied surrogates should also be non-toxic compounds. During recycling, high temperatures are applied which result in a volatilisation of the surrogates and this represents a hazardous risk to the recycling plant staff. In addition, if the challenge test is performed on a production line of the investigated recycling process, the use of hazardous compounds as surrogates leads to a contamination of the production line. Chemicals like lindane should therefore be substituted with other, non-hazardous compounds. In general, practical instructions and recommendations for challenge tests are given in refs 5 and 10.

Regarding the contamination conditions and the initial concentrations of the surrogates, there are some different procedures recommended. The US FDA10 recommends a soaking procedure using contact conditions of 14 days at 40 °C. The surrogates should be dissolved in the solvents heptane and iso-propanol. The material, which should be contaminated, is totally immersed with this solution. A 100% feedstock of contaminated material is then recycled with the investigated recycling process. The initial concentrations are in the range of 49 ppm for, e.g., benzophenone, to 1100 ppm for trichloroanisol and up to 4860 ppm for the solvents chloroform and diethyl ketone.7 In Europe

Table 9.1 Examples of chemicals to be used as surrogates in a challenge test surrogate


Functional group



Toluene Chlorobenzene


1,1,1 -Trichloroethane

Methyl salicylate

Tetracosane Phenyl cyclohexane Benzophenone Methyl stearate Lindane


CH3CCl3 (133.4)


C24H50 (338.7) C12H16 (160.3) C13H10O (182.2) C!9H38O2 (298.5) C6H6Cl6 (290.8)

Aromatic hydrocarbon Halogenated aromatic hydrocarbon

Halogenated hydrocarbon Halogenated hydrocarbon


Hydrocarbon Aromatic hydrocarbon Aromatic ketone Aliphatic ester Halogenated hydrocarbon

Volatile, non-polar, liquid Volatile, medium-polar, liquid, aggressive to PET Volatile, medium-polar, liquid, aggressive to PET Volatile, medium-polar, liquid, aggressive to PET

Non-volatile, polar, liquid

Non-volatile, Non-volatile, non-polar, solid non-polar, liquid

Non-volatile, polar, solid Non-volatile, polar, solid Non-volatile, medium-polar, solid

Metal organic compound Non-volatile, solid

Environmentally hazardous compound, restricted use in Europe, should be substituted surrogate for limonene

Hazardous compound, should be substituted Not stable during recycling

nt o fo

a solventless contamination procedure is recommended using contact conditions of seven days at 50 °C;5 100% of contaminated feedstock in form of flakes (or bottles) should be used. The initial concentrations of feedstock material range from 350 ppm to 500 ppm (which are understood to be the concentrations of contaminants after a conventional recycling treatment). When comparing these surrogate concentration ranges with contamination levels of postconsumer contaminants found in PET11 safety factors 18 to 25 (e.g. for flavour compounds such as limonene) or 120 to 170 for unknown compounds, can be discussed.5 It should be noted that these safety factors do not include the effect of the super-clean process which reduces the contamination concentrations to non-detectable levels.

In conclusion, science and practice have demonstrated that both the US FDA soaking procedure (14 days at 40 °C) and the solventless contamination procedure (seven days at 50 °C) are suitable to evaluate decontamination technologies with respect to their potential for producing regulatory compliant food grade recycled PET. The preferred procedure may be selected case by case according to the particular requirements of the technological process and the end user (customer).

9.4.3 Migration estimation

For several of the super-clean recycling processes cleaning efficiency data or residual surrogate concentrations can be found in the literature. However, the recycling company normally do not know how much recyclate will be applied in the packaging material. In addition the packaging dimensions (volume, surface area) are not known. Therefore a general estimation of the migration on the basis only of the cleaning efficiency is not possible, which means that every application should be experimentally tested. In the pre-market or pilot plant phases such a procedure is not feasible.

Migration models can be used for estimating the maximum migration of organic compounds out of the packaging materials into contact media. A generally recognised migration model based on diffusion coefficient estimation or organic chemical substances in polymers has been validated within a European project.12 For migration estimation of surrogates from PET, this migration model is applicable and provides estimated migration values on a 95% probability level. For PET such a migration calculation is illustrated by Fig. 9.1, which models the migration from PET after ten days at 40 °C depending on the molecular weight of a migrating PET constituent or contaminant and its residual concentration of the compound in the PET bottle wall (CP,0). The migration model applied for this calculation is described in ref. 12. Figure 9.1 shows that the higher the molecular weight of a compound the lower is the migration into a contact media. According to Fig. 9.1 and assuming that CP,0 = 10 ppm for any potential migrant of a PET constituent, toluene (molecular weight MW = 92) as a surrogate would give a migration value of approximately 7 ppb whereas methyl stearate (MW = 298) migrates

Molecular weight of migrant

Fig. 9.1 Molecular weight dependent relationship between residual content CP0 of an organic chemical compound in PET and its migration after ten days at 40 °C into a food simulant or food with high solubility for the substance. For a substance with MW = 200, for instance, a CP0 of 10 ppm corresponds to a migration of 4 ppb.

Molecular weight of migrant

Fig. 9.1 Molecular weight dependent relationship between residual content CP0 of an organic chemical compound in PET and its migration after ten days at 40 °C into a food simulant or food with high solubility for the substance. For a substance with MW = 200, for instance, a CP0 of 10 ppm corresponds to a migration of 4 ppb.

at approximately 2.4 ppb only. Or, when defining the maximum initial concentration (MIC in ppm) of a surrogate in PET which would correspond to a migration value of 10 ppb in food or food simulant then the ratios of MIC values given in Table 9.2 can be derived for contact conditions of ten days at 40 °C. It should be noted that these MIC values are still conservative due to the applied migration model.

9.4.4 Quality assurance and compliance testing

Analytical determination of contaminants

Suitable analytical monitoring programmes are recommended to ensure continued product quality as will have been demonstrated by the challenge test. Useful, and in practice feasible, approaches have been developed and published in the literature. Possible methods and techniques include sniffing devices for returned used bottles as well as instrumental analysis techniques such as headspace or thermodesorption gas chromatography coupled to flame ionisation (FID) or mass spectrometry (MS) detectors.7'9'11'1317 Other suitable methods may also be established. These analytical methods can be comfortably implemented into the production process for checking either the input quality to allow early sorting out of any inconvenient post-consumer qualities from conventional recycling as well as for super-clean product control.

Table 9.2 Surrogate-dependent MIC values (in ppm) corresponding to a migration value of 10 ppb

Surrogate (mol. weight)

MIC (ppm)

Toluene (92)


Chlorobenzene (113)


Phenyl cyclohexane (160)


Benzophenone (182)


Methyl stearate (298)


fictive substance (400)


fictive substance (500)


fictive substance (750)


Migration aetermination

To enable migration testing as the most direct evaluation step, a model food contact article should be manufactured from the particular challenge test product. The model article should be manufactured as close as possible to the real industry scale conditions. However, technical difficulties may occur due to relatively high contamination levels and also mechanical adverse effects, or optical impairments may be observed on the final model articles for the same reasons. Nevertheless, these articles can be used for migration testing since they rather generate a worse case concerning the diffusion of surrogates. Similarly, when different types and geometries of food contact articles are likely, it is recommended that the type which is expected to have the highest diffusion rate should be manufactured. For instance, amorphous sheets have higher diffusion rates than bottles.

Migration testing is generally recommended but not always necessary. Instead of verification of the assessment criteria by migration testing, this requirement can be checked via determination of residual surrogate content in the recycling product (recycled PET pellets or bottles and other articles) or in the surrogate article, in connection with a scientifically recognised method for migration estimation. If the concentrations of the surrogates in the output material (e.g. pellets) are such that under the assumption that 100% migration of the whole surrogate amount will not lead to concentrations above 10 ppb in the foodstuff, no migration testing is necessary. The foodstuff/ PET relation and the amount of recyclates in the bottle wall (e.g. 25% recycled material and 75% virgin PET or other ratios) should be taken into account.

In any case where the challenge test product fails the above-mentioned criteria, or in cases of doubt, migration testing is obligatory and needs to be carried out according to the provisions laid down in EU Directives 97/48/EC and 85/572/EEC and their amendments. The conditions of foreseeable use of the PCR PET containing article do influence the extent of possible migration into food. The migration rate determining parameters are contact time and temperature as well as the nature of the real filled product respectively, and the corresponding test conditions according to the above mentioned EU Directives. With regard to the conditions of use it also has to be considered whether the recycled PET is in direct contact with the foodstuff or separated by a functional barrier. In cases of doubt, it must be guaranteed that migration testing is carried out under worst-case conditions.

The assessment criterion to decide whether the challenge test has passed the crucial requirement of efficient removal of potential contaminants is defined by a maximum migration rate leading to a concentration of 10 ppb (mg l-1) in the food simulant. It must be noted that initial surrogate concentrations introduced by a challenge test into a super-clean process range several orders of magnitude higher compared to what can be found in reality. Therefore, reduction of these initially high concentrations to such low levels in the challenge test product, or in the model food contact article, which correspond with or lead to migration values smaller than or equal to 10 ppb, demonstrates the deep-cleansing efficiency of the technology and is not connected to any consumer exposure considerations.

Sensorial evaluation

To comply with the general requirements of Article 3 of the EU Regulation 1935/200418 sufficient sensory inertness of the PCR PET products as food contact articles needs to be assured. Therefore appropriate sensory testing of food contact articles made from super-clean products is recommended. As worse case test conditions for this purpose, storage of the article in direct contact with water for ten days at 40 °C have been generally accepted. However, depending on the particular application, modified tests may be more suitable.

0 0

Post a comment