WO2008017843A1

WO2008017843A1 - Recycling process for polyethylene terephthalate (pet)

Info

Publication number: WO2008017843A1
Application number: PCT/GB2007/003022
Authority: WO
Inventors: Edward Ireneusz Kosior
Original assignee: The Waste And Resources Action Programme
Priority date: 2006-08-09
Filing date: 2007-08-09
Publication date: 2008-02-14
Also published as: GB0615765D0; EP2054207A1

Abstract

The present application relates to a process for recycling PET from waste material comprising PET. The process comprises steps of: a) optionally washing said waste material; b) sorting at least some non-PET material from said waste material, where present; c) grinding the waste material to form flakes; d) optionally washing said flakes; e) sorting and separating said flakes on the basis of their surface area; f) heating the flakes in order to decontaminate the flakes; and g) melting the decontaminated flakes and extruding the melt. The process may be used to prepare food grade PET from waste material comprising PET.

Description

Recycling Process for Polyethylene Terephthalate (PET)

The present invention relates to an improved process for the recycling of PET. Many recycling processes for PET exist. These range from chemical processes such as methanolysis whereby the PET is depolymerised to provide dimethyl terephthalate (DMT) and ethylene glycol to physical processes which involve combinations of standard mechanical recycling processes with non- mechanical procedures such as high temperature washing.

A typical physical PET recycling process includes washing PET bottles and containers, sorting the bottles and containers and removing any metal therefrom. The bottles and containers are then ground and washed before being dried and extruded and then pelletised.

As the customer becomes more environmentally aware, the demand for recycled materials increases. The estimated demand for recycled PET in Western Europe is 666 kilo tonnes per annum, of which 50% is estimated to be in food contact applications.

In addition, the demand for recycled PET in food contact applications is being driven by a number of other factors including:

• the imminent requirements of European Directives; and • the potential for reducing the cost of bottle grade PET resin.

In order for recycled PET to be suitable for food contact applications, it must meet criteria set out by various administrative authorities and/or legislation. As the definition of food grade PET varies throughout the world so too do the aforementioned criteria. That said, the testing for establishing these criteria is largely the same and is based upon migration of contaminants from the recycled PET to the food contained in the recycled PET package.

To comply with US Food & Drug Administration (USFDA) requirements, the migrating substances must either only be present in dietary concentration at very low levels, i.e. less than 0.5 ppb, or the level is less than 1% of the acceptable daily intake (ADI) for those substances.

European legislation defines the overall migration limit as "Plastic Materials and articles shall not transfer their constituents to foodstuffs in quantities exceeding lOmg dm^"2 of surface area of material or articles or όOmgkg^"1 of food for containers with a capacity of 0.5 to 10 litres" (Council Directive 86/572/EEC, Art. 2).

Another consideration of a recycling process is the effective removal of surrogate contaminants, that is contaminants to which the PET container is exposed to as a result of consumer misuse, for example, the storage of petrol or oil in a bottle which originally stored beverages, say. These surrogate chemicals are selected to represent the wide range of volatile and non- volatile, polar and non-polar materials that can be encountered in the consumer environment. Clearly, an effective recycling process must remove such surrogate contaminants if the recycled material is intended for food contact.

The food packaging industry represents a significant market for recycled PET. Therefore, in order for PET recycling to be commercially viable it is imperative that the recycled PET meets the criteria laid down by the various administrative authorities and/or legislation. Moreover, some large corporations demand recycled PET to be even less contaminated than the legislative parameters which are in place. WO 01/21373 discloses a process for preparing food contact grade PET. The process involves sorting at least some of the non-PET materials from waste material, followed by dividing the PET material into flakes. The flakes are then washed in a hot aqueous medium containing alkaline materials and surfactants. Following washing, the flakes are dried such that absorbed contaminants are removed by heating and stirring the flakes under vacuum. The flakes are then melted, extruded and pelletised. This process achieves decontamination after extrusion which precludes the opportunity to use a flake product which would be cheaper in the next application. There is therefore a need for a process which achieves decontamination to very high levels at the flake stage allowing the flake to be directly used if required. Further decontamination and building of intrinsic viscosity (IV) can be achieved during the extrusion step.

WO 2005/037514 discloses a further process for recycling used PET bottles. The process involves washing and grinding the bottles into flakes, said flakes are then separated based upon their mass/density/wall thickness. Interestingly, WO 2005/037514 teaches that contamination is related to the thickness of the flakes, i.e. thicker flakes are more deeply contaminated. Thus, flakes are separated on the basis of their mass/density/wall thickness.

Despite there being up to 60 recycling processes having USFDA approval for use with food contact applications, very few of these are in use due to the economics being less favourable compared to virgin resin.

Therefore, it is an object of the present invention to provide an improved process for the efficient and cost effective recycling of PET suitable for food contact applications. Thus, according to a first aspect of the present invention there is provided a process for recycling PET from waste material comprising PET, said process comprising the following steps of: a) optionally washing said waste material; b) sorting at least some non-PET material from said waste material, where present; c) grinding the waste material to form flakes; d) optionally washing said flakes; e) sorting and separating said flakes on the basis of their surface area; f) heating the flakes in order to decontaminate the flakes; and g) melting the decontaminated flakes and extruding the melt.

The process according to this aspect of the present invention may be used to prepare food grade PET from waste material comprising PET.

Advantageously, the process of the present invention gives rise to a more efficient and cost effective process since only the flakes that are rapidly decontaminated need be used in the latter stages of the process.

A further advantage of the process of the present invention is that solid stating is not required. By eliminating a separate solid stating step from the recycling significant capital costs savings are made and energy costs for operating the process are reduced.

Solid stating is typically used as a way of advancing the intrinsic viscosity of PET resin. This is usually done with virgin resin as a normal way of converting fibre grade resin into bottle grade resin, and involves heating PET granules at elevated temperatures i.e. 200 °C in an anhydrous atmosphere, which may be air or an inert atmosphere. This is typically achieved in a separate step either in a rotating dryer or in a vertical column with control on the residence time which is typically related to the increment in IV that is being achieved by the resin.

Where necessary, washing of the waste material prior to sorting may be performed using any suitable solvent. This initial washing step is often desirable in order to reduce contaminant levels. Preferred solvents include any of the following either alone or in combination: water, ionic or non-ionic detergents up to 0.2% by wt, dilute caustic soda up to 1.0 % by wt.

Where present, non-PET material separated out from the remaining waste material typically includes labels and metal objects, such as caps, etc. It will be appreciated that the present invention may be carried out using pre-sorted waste from which at least some non-PET material has been removed (and which may optionally have been washed).

The non-PET material is separated out by means well known to those skilled in the art. For example, sieving, air elutriation, sink-float separation, colour separation.

Once some non-PET material is separated out from the waste material, the waste material is ground to form flakes. Typically, the flakes are ground such that they have an average size in the range of from 2mm to 20mm. Preferably, the flakes have an average size in the range of from 4mm to

12mm, more preferably from 5mm to 10mm. The 'size' of a flake is intended to mean the size as measured across the widest part of the flake.

The size distribution of the flakes may vary. Ideally, at least 80% (± 10%) of the flakes fall within the above mentioned size ranges. It has been realised that upon grinding two types of flakes are produced. These flake types differ in respect of their surface area per unit mass.

The thinner sections of the waste PET, for example, walls of bottles, form flakes having a larger surface area than the thicker sections of the waste PET, for example neck and base regions of the bottles.

As referred to herein, larger surface area means flakes having a surface area to mass ratio of about 15mm²/g or greater. Therefore, flakes having a smaller surface area are those having a surface area to mass ratio less than about 15mm²/g.

As the majority of the surface area of a bottle comprises thinner wall sections, the thinner PET waste is generally more contaminated than the less abundant thicker neck and base sections.

Therefore, separation of the two areas of a bottle leads to a more efficient recycling process as the less contaminated material can be used for food contact applications, whilst the more contaminated material can be recycled and used for non food contact applications.

Interestingly, it has been found that although the flakes having a larger surface area per unit weight are more contaminated, the contamination is predominantly surface contamination and so is easier to remove than the contaminants of the flakes having a smaller surface area per unit weight. The waste PET material is ground such as to form discrete pieces or flakes.

The grinding process produces two distinct types of flakes which differ in their surface area.

The grinding step may be carried out by any of the techniques known in the art. For instance, grinding is generally performed by cutting the bottles against a fixed size aperture screen, for example 4 mm to 15 mm, with rotary blades. As the bottles are cut, the fragments pass through the screen. The thicker sections of the bottle such as the neck and base end up with generally these dimensions of screen size and thickness. Grinding in this manner ensures that flakes formed from the thinner wall sections have a larger surface area than those formed from the thicker base and neck sections. This is the key to the separation of the flakes which is discussed below.

The flakes are optionally washed after grinding. This may be desirable to reduce surface contamination, such as inorganic deposits. The flakes are typically washed at temperatures above 80°C with 85-90°C being the preferred temperature. The flakes may be washed using any suitable solvent, such as water, an aqueous solution of caustic soda at 1 to 1.5% by wt, or an aqueous solution of an ionic or non ionic detergent at 0.1 to 0.5 % by wt.

Following washing, the flakes are preferably dried. The flakes may be dried by any conventional drying technique, for example using a fluidised bed drier, recirculating air drier or and desiccant drier. Suitable conditions used for the drying step are well known to those skilled in the art.

A key step in the process of the present invention is the separation of the more contaminated flakes from the less contaminated flakes. The flakes having a larger surface area to mass ratio, which are more contaminated, are separated from the less contaminated flakes which have a smaller surface area to mass ratio.

Separation may be achieved by any suitable means, such as by the use of destoners or air classifiers. A destoner effectively separates heavier-than-product debris, such as glass, stones and metal from a large amount of lighter product. The principle is that an inclined plate is vibrated and under set pressure the lighter flakes move down the plate. In the case of the present invention the lighter flakes are the flakes having the larger surface area to mass ratio.

Preferably, separation is achieved by air classifiers.

Air classifiers work by using the principle of terminal velocity for a specific product to classify and separate particles. The flakes having a larger surface area have a different terminal velocity from the flakes having a smaller surface area. By using multiple aspirators, and variable air flow rates appropriate separation can be achieved.

The separation technique used in the present invention may include one or more passes. For example, each pass lifts approximately 75% of the flakes having a larger surface area. This leaves a balance of 25% of the flakes having a large surface area in the through stock. On a unit having two passes a second pass would also remove approximately 75% of the particles having a larger surface area (i.e. 75% of the remaining 25%), leaving 6.25% of the flakes having a larger surface area in the through stock. Therefore, a dual pass system has 93.75% efficiency.

Therefore, the efficiency of the separation process can be maximised by tailoring the separation conditions.

The invention permits the selection of only the less contaminated PET flakes, i.e. those having a smaller surface area to mass ratio, for decontamination in the next step of the process. Heat treatment to decontaminate the flakes may be carried out at any suitable temperature. By way of example, the heat treatment may be carried out at temperatures of 140⁰C or above. Higher temperatures are generally preferred, as the heating time is reduced as the temperature increases, which is economically beneficial. More favourable results are achieved at temperatures of 16O⁰C or above, preferably 17O⁰C or above, more preferably 180⁰C or above, and still more preferably 190⁰C or above. A temperature of about 200⁰C (±5°C, more preferably ±2°C) has been shown to be particularly suitable.

The heating time varies depending upon the temperature of the heat treatment. At around 200⁰C, by way of example a heating time of about 4 hours has been found to give good results. As the temperature is reduced, the heating time should be increased accordingly. By way of example, at around 190⁰C the flakes may be heated for about 8 hours, at around 180⁰C the flakes may be heated for about 16 hours, at around 170⁰C the flakes may be heated for about 32 hours, at around 160⁰C the flakes may be heated for about 32 hours, etc. The heating conditions may be routinely varied by the person skilled in the art.

In order to meet the exacting standards required by large corporations, standard high temperature decontamination was found to be insufficient.

Therefore, the process of the present invention preferably employs high temperature decontamination, which is preferably conducted in an inert atmosphere. The inert atmosphere protects the polymer, minimising degradation and discoloration of the polymer at high temperatures. It is especially desirable to employ an inert atmosphere at heating temperatures of 17O⁰C and or where there is prolonged heating. The inert atmosphere may be, for example, nitrogen or argon, and is preferably nitrogen. The inert atmosphere may take the form of a blanket covering the flakes, thereby preventing oxidation.

Advantageously, decontamination conducted at about 200⁰C in an inert atmosphere reduces the contaminants in the flakes by approximately 99.8 to 99.9 %. Of course, these percentages differ depending upon the contaminant, but nevertheless demonstrate the efficiency of the process of the present invention.

Following decontamination, the flakes are melted and extruded into strands which are typically pelletised. A twin screw extruder may be used to extrude the flakes. Extrusion may take place at any suitable temperature, for example at a temperature within the range of from 280⁰C to 290 C. The extruder may be twin or triple vented. During extrusion, the melt may be filtered to remove residual particles having a diameter above 75 microns.

Advantageously, the extrusion step avoids solid stating and it increases the intrinsic viscosity of the PET thereby eliminating the need to use rotary vacuum dryers which in turn reduces the capital cost of the process of the present invention. In fact, the material obtained from the process of the present invention, i.e. the output material, has a higher intrinsic viscosity that the material used in the process, i.e. the input material. The more contaminated flakes having a larger surface area can still be processed as food grade, but may not meet the exacting standards required by large corporations.

According to a second aspect of the present invention there is provided recycled PET produced by a process as described herein. According to a third aspect of the present invention there is provided the use of recycled PET produced by a process as described herein for the preparation of PET containers.

The containers may be for food contact applications. In particular, the containers may be bottles for beverages.

The present invention will now be described by way of example only and with reference to the following results.

One of the assumptions underlying the process of the present invention was that two flakes types having different degrees of contamination can be obtained from PET containers, etc. The two flake types are those having a large surface area to mass ratio, i.e. more contaminated, light flakes (LF), and those having a small surface area to mass ratio, i.e. less contaminated, heavy flakes (HF). It was also assumed that separation of these flake types will improve the decontamination process. These assumptions were tested by analysing the two flake types from the 'challenge test'; after they had been washed and then before and after a drying step using an infra red rotary (IRD) dryer. The samples were analysed without separation of the flake types up to the wash stage and with separation of the flake types thereafter. The results can be seen in Table 1 below.

Table 1

The results show that initially the 'light' flakes possess higher levels of surrogate chemicals (1.6 to 2.5 times) than the 'heavy' flakes and after one step of drying for 10 minutes in the IRD dryer the 'light' flakes had lower levels in general (0.6 to 0.9 with the exception of benzophenone) than the 'heavy' flakes.

This initial result is very significant since it indicates that the 'light' flakes in the Challenge test are more highly contaminated. The only way that this could happen would be if the contaminants were concentrated at the surface rather than being distributed uniformly through the particles.

The surface area of the two types of flakes is significantly different, with a flat particle having at least 4 times the surface area of a spherical particle of equal weight.

(The theoretical ratio of the surface area is defined by the ratio D/3T, where D is the diameter of a sphere (i.e. a heavy flake) and T is the thickness of a flake (i.e. a light flake) of equal weight). This calculation is based on considering 'heavy' particles with a diameter of 4mm and 'light' flakes with a wall thickness of 0.3mm as a typical example of the sizes that exist in the flake samples. Given that the 'heavy' flakes are not spherical but a range of irregular shapes with higher surface areas, the surface area ratio will be less that the theoretical value for the idealised geometry. This feature (surface area differences) means that the Challenge test procedure used by the

Fraunhofer IVV results in unavoidable differentiation between the contamination levels imposed on the two different types of particles.

It is important to note that the flakes have been separated on the differences in surface area per unit weight (i.e. specific surface area) in the aerodynamic separation step performed in the air classifiers. This factor will also be expected to play a significant role in the decontamination process since contaminants will initially be removed from the surface by diffusion into the air stream.

The results of the short decontamination step used above show that the flakes with the higher specific surface area decontaminate the surrogate chemicals to a greater extent, i.e. 71 to 89% compared with 39 to 67% for flakes with a lower specific surface area. The results show that the higher specific surface area flakes

(light flakes) decontaminate at 1.3 to 1.8 times that rate of the lower specific surface area flakes (heavy flakes) as shown in Table 2 below.

Table 2

These results suggest that both the process of Challenge test contamination and the process of decontamination are controlled by specific surface area of the flakes.

Therefore, these findings validate the key principle of using an aerodynamic separation process that separates the flakes into two fractions prior to processing to a decontamination step. Table 3 shows that by varying the volumetric flow rate of the air, it is possible to achieve a wide range of separation percentages. The results shown in Table 3 were generated using a multi aspirator air classifier to separate five different samples of 500g.

Table 3

Table 3 shows that by find adjustment of the air setting, the desired percentage of liftings can be achieved accurately and consistently.

Table 4 shows the efficiency of decontamination of PET bottles having had various surrogate chemicals stored therein prior to recycling. The data presented is for recycled PET at various stages in the process of the present invention, i.e. initial contaminated flake, flake after washing step, flake after high temperature decontamination, pellet after vacuum extrusion and the test bottles made from 100% recycled PET according to the present invention. The flakes used to generate the data of Table 4 are light flakes (i.e. high surface area to mass ratio flakes).

Table 4 Table 5 shows the decontamination efficiencies as percentages.

Table 5

The most significant decontamination step is the heat treatment step at about

200⁰C for 4 hours which shows a dramatic improvement in the removal of all components to very low levels in the region of lppm.

The data for the surrogate chemicals in the flake is worthy of noting. The levels are very low and would potentially indicate that this process may well meet the requirements of beverage companies based on the levels of volatile chemical being well under 3ppm and the non- volatile chemicals being well under lOppm.

The results for the test bottles show that the levels of surrogate contaminants are present at very low levels and based on USFDA submissions it is expected that these bottles will meet the requirements of the migration tests. The data for the pellet contamination levels shows slightly higher levels of contaminants compared to the flake. It is thought that this variation is due to the arrangement used during testing where the flake was sampled from the top of the bed of flake being dried indicating that the distribution of gas flowing over the flakes may not have been ideal during the trial. This situation would not be expected under normal production conditions where each flake would have the same residence time and consequently the same residual surrogate contaminants. This anomaly means that the level of contaminants in the bottles would be even lower than recorded in the current tests and shown in Table 3.

The key issue is the migration of any residual surrogate chemicals into the 95% ethanol solution, which is an aggressive extractant for PET used as a food stimulant.

The Fraunhofer Institute have calculated the maximum residual concentrations in ppm that would correspond to a migration limit equal to or smaller than lOppb based on conservative diffusion modelling data in olive oil at 4O⁰C for 10 days. These levels are shown in Table 6.

Table 6

These results show that a migration test which could have been done on the bottles produced from the surrogate containing pellets obtained from the challenge test (Table 3) could not lead to migration values in food stimulants exceeding 10 ppb.

In addition, application of migration modelling to the data shows that the results are applicable to hot fill conditions of 2 hours at 7O⁰C or 30 minutes at 100⁰C followed by regular cold fill storage conditions of 10 days at 4O⁰C.

Therefore, the process of the present invention produces food grade recycled PET that meets the ILSI and EU conditions and that the resin can be used at filling conditions up to 100⁰C. Table 7 shows melt flow and intrinsic viscosity (IV) of the recycled PET.

Table 7

These results show that the flakes having a large surface area are substantially increased in IV from an initial value of typically 0.75 to 0.76 to a final value of 0.81 to 0.86.

This means that an IV increase of 0.6 to 1.09 may be achieved by the drying process. This would imply substantial elimination of water molecules and quite likely any other molecules such as surrogate contaminant molecules.

Table 8 shows melt flow and IV data for pellets made by the process of the present invention.

Table 8

The data shows that the final IV of pellets made by the process of extrusion is in the range 0.78 to 0.82 with an average of 0.80. This result is consistent with solid stating of the flake to a level of 0.81 to 0.86 and having a reduction during melt processing of 0.04.

The final IV is in the range of 0.80±0.02 is a very satisfactory result since the IV of virgin resin is typically also 0.80. It is of course to be understood that the invention is not intended to be restricted to the details of the above embodiments which are described by way of example only.

Claims

1. A process for recycling PET from waste material comprising PET, said process comprising the following steps of: a) optionally washing said waste material; b) sorting at least some non-PET material from said waste material, where present; c) grinding the waste material to form flakes; d) optionally washing said flakes; e) sorting and separating said flakes on the basis of their surface area; f) heating the flakes in order to decontaminate the flakes; and g) melting the decontaminated flakes and extruding the melt.

2. A process according to claim 1 for preparing food grade PET from said waste material comprising PET.

3. A process according to claim 1 or claim 2, wherein steps (f) and (g) are carried out only on a selected population of flakes sorted and separated on the basis of their surface area.

4. A process according to claim 3, wherein the flakes in said selected population of ftakes have a surface area to mass ratio of less than about 15mm²/g.

5. A process according to any preceding claim, wherein the heating step (f) is carried out at a temperature of at least 14O⁰C to 190⁰C.

6. A process according to claim 5, wherein the heating step (f) is carried out at a temperature of about 200⁰C.

7. A process according any preceding claim, wherein the heating step (f) is carried out in an inert atmosphere.

8. A process according to claim 7, wherein the inert atmosphere comprises nitrogen or argon.

9. A process according to claim 8, wherein the inert atmosphere comprises nitrogen.

10. A process according to any preceding claim, wherein the flakes formed in step (c) have an average size in the range of from 2mm to 20mm.

11. A process according to claim 10, wherein the flakes formed in step (c) have an average size in the range of from 4mm to 12mm.

12. A process according to claim 11, wherein the flakes formed in step (c) have an average size in the range of from 5mm to 10mm.

13. A process according to any one of claims 10 to 12, wherein at least 80% (± 10%) of the flakes fall within the average size range.

14. A process according to any preceding claim, wherein the flakes are washed in step (d).

15. A process according to claim 14, wherein the flakes are washed at a temperature of 80⁰C or above.

16. A process according to claim 15, wherein the flakes are washed at a temperature in the range of from 85°C to 90⁰C.

17. A process according to any one of claims 14 to 16, wherein the flakes are washed in a solvent selected from one or more of: water, caustic soda, and ionic or non-ionic detergents.

18. A process according to any one of claims 14 to 17, wherein the flakes are dried after washing step (d).

19. A process according to any preceding claim, wherein the separation of the flakes in step (e) includes two or more passes.

20. A process according to any preceding claim, wherein separation step (e) is carried out in an air classifier.

21. A process according to any one of claims 1 to 19, wherein separation step (e) is carried out in a destoner.

22. A process according to any preceding claim, wherein the extrusion of step (g) is carried out at a temperature within the range of from 280⁰C to 290⁰C.

23. A process according to any preceding claim, wherein the extrusion of step (g) includes filtering of the melt to remove residual particles having a diameter above 75 microns.

24. A process according to any preceding claim, wherein the flakes are washed in step (a).

25. A process according to claim 24, wherein the flakes are washed in a solvent selected from one or more of: water, ionic or non-ionic detergents, and caustic soda.

26. A process according to any preceding claim, wherein the waste material comprising PET includes bottles.

27. Recycled PET produced by a process according to any preceding claim.

28. The use of recycled PET produced by the process of any one of claims 1 to 26 for the preparation of a PET container for a food contact application.

29. The use according to claim 28, wherein the container comprises a bottle.

30. The use according to claim 29, wherein the bottle is for containing a beverage.