Overall migration tests

In the first stage of the research, overall migration tests were performed at Ghent University. All films used in the storage tests described in Chapters 3 and 4, except for the Paper/AlOx/PLA and the skalax film (not enough sample anymore), and some of the rigid materials described in Chapter 6 were tested (commercially available materials). In the second stage of the research,

OM 8 Food simulant E for 2 hours at _{175 °C and food simulant D2} for 2 hours at 100 °C

High temperature applications only

OM 9 Food simulant E for 2 hours at _{175 °C and food simulant D2} for 10 days at 40 °C

High temperature applications including long term storage at room temperature

159

overall migration tests on the heat resistant films described in Chapter 5 as well as on PLA + Joncryl™ (Chapter 6) were performed at the Belgian Packaging Insitute (BVI). This latter is an accredited laboratory regarding overall migration tests, but they were only involved as an active research partner during the second stage of the research (in the framework of a second project).

2.2.2.1 Tests performed at Ghent University

Overall migration tests were performed in simulant A (10% ethanol) and simulant D2 (vegetable oil or alternative) for 10 days at 40°C or for 2 hours at 70°C. These simulants were selected, since according to EC 10/2011, compliance with the overall migration limit for all types of food (except acidic food products), can be demonstrated by testing in these two simulants. As described in Chapter 3 and 4, no acidic food products were tested in this research. The first test condition (10 days at 40°C) was selected, since according to EC 10/2011, this condition (OM 2) is valid for any long term storage at room temperature or below, including a heating step of 70°C for up to 2 hours or a heating step of 100°C for up to 15 minutes. As described in chapters 3 and 4, the tested food products fall within this description. The second test condition (2 hours at 70°C, OM 3) was selected to see the influence of a higher temperature on the migration behavior of the materials. The overall migration was tested for the Natureflex™N913 and N931 films, the Natureflex™/PLA film, the Cellophane™/M/PLA film, the Natureflex™NK film, the printed and non-printed multilayer PLA film, a PLA tray, PP (beaker), PLA (beaker), sc-PLA (beaker), PLA + fiber (extrusion plate) and PLA/PHB blend (extrusion plate).

Overall migration tests in vegetable oil (olive oil) were performed by making use of stainless steel migration cells. Films (1.4 dm2) were conditioned at 50% RH (178 ml H2SO4/l water) in a dessicator and weighted every 24 hours. If the difference in mass between two time intervals was less than 5 mg, the starting mass of the sample could be determined. The sample was then placed in a stainless steel migration cell, filled with olive oil (Bertolli dal 1865 classico, Bertolli, Tavarnella val di pesa, Italy) and stored during 10 days at 40°C. Two blank samples (films not in contact with olive oil) were also tested. After 10 days, the olive oil was removed and the samples conditioned, first at 80% RH (102 ml H2SO4 /l water) for 24 hours and then at 50% RH and weighted every 24 hours. If the difference in mass between two time intervals was less than 5 mg, the end mass of the sample could be determined. The absorbed oil was extracted from the film with pentane. The foil was cut (± 30 mm²) and brought into an extraction tube (extraction thimbles, 33 × 118 mm, Whatman international, Maidstone, UK) which was closed

160

by cottonwool. The tube was placed in an soxhlet tube and was placed onto a flat bottom flask (250 ml) which was ¾ filled with pentane and 10 ml internal standard (20,35 mg triheptadecanoin in 10 ml heptane). Extraction took place during 7 or 8 hours, after which the flat bottom flasks were removed and replaced by new flat bottom flasks for a second extraction. The pentane was then evaporated and the oil esterified by refluxing for 10 min after addition of 10 ml n-heptane and 10 ml KOH solution (11g/ l MeOH) and refluxing for 2 min after addition of 5 ml boron-trifluoride-methanol complex (Merck). After addition of saturated NaCl-solution (40 g NaCl in 100 ml distilled water), a clear upper phase was obtained. This upper phase was pipetted in a test tube and subsequently in a GC-vial. Chromatographic analysis was performed in a Varian CP-3380 (Varian Analytical Instruments, Mitchel Drive, USA). The sample (1 µl) was introduced into the injector operating at 320°C and the separation was carried out in an Varian CP-Sil 8 CB (30m, 0.32mm, 0.25µm) capillary column. The oven temperature was programmed from 100°C (held for 2 min) to 290°C at a rate of 15°C/min (10 min). Triheptadecanoin (20.35mg/10 ml heptane) was used as an internal standard and a calibration curve was made. From the chromatograms, the peak areas of C16:0, C18:0, C18:1, and C18:2 olive oil methyl ester peaks were determined as well as the area of the C17:0 internal standard methyl ester peak. The test with coconut oil instead of olive oil was performed in a similar way (with the peak areas of C12:0 and C14:0 methyl esters and C17:0 (internal standard) methyl ester). Tests with coconut oil were performed because interference with the blank sample in olive oil was encountered for some samples. Tests were performed in quadruplicate, then averaged and a standard deviation was determined using Microsoft Excel.

For the overall migration in 10% and 95% ethanol the method of total immersion (PLA/PHB and PLA + fiber) or filling (beakers of PP, PLA and sc-PLA) was used. For the first method, the multilayer bioplastics were cut (1 dm²), after which they were immersed in 200ml of 10% or 95% ethanol (10 ml or 95 ml ethanol in respectively 90 ml or 5 ml distilled water) and stored for 10 days at 40°C or 2 hours at 70°C in glass jars with a lid. For the second method, beakers were filled with 100ml of simulant and covered with parafilm to avoid evaporation during the incubation period. The contact surface (CS) could be determined by the diameter of the beaker (db) and the used volume of simulant (Vs).

ℎ =

(𝜋𝑟2₎ (equation 7.1, with r = db/2 and h = hight of simulant in beaker)

161

Each time, two blank samples (food simulant not in contact with the tested material) were also tested. After 10 days or 2 hours, the films were removed (method of total immersion) or the simulant was transferred to an erlenmeyer (method of filling) and the ethanol was evaporated by the following method: little aluminum or stone jars were left in an oven (105 °C) overnight, then put in a dessicator for 15 minutes and weighted, the solutions were then evaporated in these jars on a hot plate until a small amount remained, which was put in the oven till dryness, cooled down in a dessicator for 15 minutes and weighted. This last procedure (oven + dessicator) was repeated until a stable weight was obtained. The overall migration in 10% and 95% ethanol was determined in triplicate, then averaged and a standard deviation was determined using Microsoft Excel.

It was decided to test a third alternative fatty food simulant, iso-octane as well, since already 2 alternative fatty food simulants were used, namely coconut oil and 95% ethanol. 95% ethanol and iso-octane are the two alternative fatty food simulants described in Directive 97/48/EC (European Commission, 1997). The migration test in iso-octane was performed by the method of total immersion as well (as described in previous paragraph) but during 2 days at 20°C instead of during 10 days at 40°C (according to table 4 in directive 97/48/EC).

2.2.2.2 Tests performed at BVI

Since no specific food groups were targeted for the materials in the second part of the research (focusing on heat resistance), migration tests were performed in all 5 food simulants (3% acetic acid, 10, 20, 50 and 95% ethanol/olive oil). A first test condition (10 days at 40°C) was selected, since this condition covers a lot of applications were heat resistant packaging is required (e.g. in package pasteurization) The second test condition (2 hours at 100°C, OM 5) was selected to see the influence of a higher temperature on the migration behavior of the materials and to see whether the packaging materials are suited for sterilization applications (up to 121°C). The worst case scenario of 2 hours at 175°C (OM 7) could not be tested, because a preliminary test showed decomposition of the film at this condition. The overall migration was tested for the Natureflex™NVS, NVR, NK and NE and CelloTherm™T films.

Tests in olive oil were performed making use of stainless steel migration cells, similar as previously described (2.2.2.1). Tests in 3% acetic acid, 10, 20, 50 and 95% ethanol were performed by the method of total immersion, similar as previously described (2.2.2.1).

162

3 Results and Discussion

3.1 Printability tests

The printing of the Natureflex™N913 film in different colors gave no problems and the printed film is now commercially used as rice and pasta packaging (figure 7.1).

Figure 7.1: Lima’s cellulose-based rice packaging

Printing of the multilayer PLA film also gave no problems and the scratch test, tape tests and tensile test showed no adhesion problems. This means a perfect adhesion from the printed inks onto the film. When a higher volume of ink (more cm³ ink per m² of film) was applied on the surface, a better adhesion of the ink was obtained. In general these tests showed that the quality of the print was good and that the print will not come off during storage.

The residual solvent content analysis (Annex B) showed that the total concentration of residual solvents in several samples was above 20 mg/m2_{, indicating that the PLA film is very receptive} to solvents. This could lead to odor and flavor defects in food products, if these solvents can migrate into the packed food product. The largest share in the residual solvent analysis were retarders (1-etoxy-2-propanol and 1-propoxy-2-propanol), which are present in the inks. These retarders are added to let the ink dry more slowly, which is only necessary for detailed printing (e.g. fine dots). Therefore, it could be agreed with the supplier to leave them out when no printing in detail is needed. If these retarders are left out, the total concentration of residual solvents will be much lower. Furthermore, the determination of the dry weight showed that the

163

total residual amount of both single components (cyan or magenta) exceeded the maximum of 0.5% (EN 13432 Packaging). The overlap sample (cyan and magenta) was in accordance with the standard (5%). It is noticed that the limit is much more stringent for single components than for overlap samples (respectively 0.5% and 5%). These tests showed that the choice of ink (e.g. with or without retarders) and the print design are very important when printing PLA films. The amount of ink remaining on the surface and the amount of residual solvents should not be too high in order to avoid negative effects on the packed food product (by migration).

It can be concluded that (the coatings of) the cellulose-based and PLA-based surfaces can be easily printed and that this print is of good quality. However for PLA surfaces, the choice of ink and print design is important, since these can have an effect on the migration towards the food product (effect on flavor or odor). This migration will not be seen in the in the overall migration tests, since these only take the migration of non-volatile compounds into account.

3.2 Migration tests performed at Ghent University