The core team, a Jeep and an elephant...

Up to this moment, five compression tests have taken place, with altered variables the number of tested bottles (9/ 12), the bottles’ volume (0,5L/ 1,5L), the design specifications (For carbonated soft drinks CSD/ For non-CSD), the bottles’ condition (Used/ New), the cap sealing method (Teflon tape/ Epoxy adhesive) and the bottles’ distribution (Hexagonal/ Orthogonal array). The internal pressure of the bottles was kept constant for all samples, at a level of 6 atm, in order for the results to be comparable. The level of deformation of the samples and their mass loss were rated after the tests, and the factors leading to failure were analysed.

The new bottles for CSD that were sealed by applying epoxy adhesive around their caps’ threads, performed considerably better, regardless their volume. Specifically, during the fourth compression test, nine 0,5L PET bottles withstood a compression load of 7tn, and presented a total displacement of 140mm, until the bottles could no be seen anymore from the front Plexiglas opening of the safety container. At that point, the test was stopped for safety reasons (the Plexiglas sheet on the corner of the safety container started to bend outwards), but up to then, there was no bursting noted nor significant leaking that would lead to rigidity loss. After releasing the load, the bottles recovered their previous shape and remained rigid. All samples presented minor plastic deformation at the base (pleats on the petaloid’s surface), but no serious damage in any case.

Similarly, during the fifth compression test, nine new 1,5L PET bottles for CSD (adhesive sealing), withstood 8tn of loading and were displaced by 160mm. In this case, as well, the test was stopped for security reasons and no bursting or serious leaking was detected. The bottles, after unloading, returned to their previous shape but their base contained damages.

The results from the last two tests are very encouraging, yet more tests will be conducted the following months so that safe conclusions can be drawn.

The equivalent of the 4th compression test 

First compression test

A dozen of type “3” bottles were filled with 6 gr. of dry ice and sealed with Teflon tape around the threads and duct tape around the cap. Measurements of their mass and max. diameter where taken so leakage can be determined. Three days later the measurements were repeated and the bottles showed an average leakage of 0,2%. 

Placing the bottles in the safety container

Afterwards, the bottles were placed in the safety container, in a hexagonal pattern, and were tested under compression. The bottles behaved surprisingly well but the test was not completed because the screws holding the Plexiglas sheet on the safety container were starting to fail. After having been loaded with a force of 20KN (2000kg) and having as a system around 8cm of displacement, the bottles returned to their previous condition once the force was removed, and no bursting took place, nor were they plastically deformed. The bottles situated at the upper 2 rows, in fact, remained intact. Some of the bottles at the lower 2 rows presented minor scratches along the contact points with the adjacent bottles, due to friction, but not severe fissures in any case.

2000 kg on the bottles only caused minor scratches

 After the compression test was competed, the bottles’ mass was re-measured. The bottles at the top 2 rows had no leakage, yet, the ones at the bottom had an average leakage of 0.17%. The measurements on CO2 leakage indicated that the sealing method needs to be optimised in order for the bottles to remain long-term rigid.

Mass measurements to determine C02 leakage

Boys with toys + two conclusions

The remains of a PET bottle

Twenty minutes of absence from the lab was enough time for two of my male colleagues to start experimenting with the dry ice leftovers, or better said, to start exploding bottles! Two useful conclusions were drawn, however, by pushing these poor bottles to their extremes.

Samples 31 and 32 didn't make it

Bottle type "6" (no.31), for example, was an overdosed sample sealed normally (no Teflon tape around the threads). In antithesis with case 8 and 9, where the same type of bottle was sealed with Teflon tape and exploded due to cap failure, here the cap stayed in place and the base bursted. Without the tape, the sealing was firmer and so the collapse occurred at the base of the bottle, which proved to be weaker than the walls. The failure of the base was intensified by the large quantity of inserted dry ice that was in contact with this surface, causing it to frost (this sample was not placed in warm water to accelerate sublimation). The temperature of the solid CO2 (-79 C) exceeded the minimum service temperature of the PET bottle (-58 to -38 C) and thus the base became brittle and incapable to withstand the internal pressure.

This test showed therefor that type "6" should not be excluded from the samples list (the PET surface itself is strong), but rather a new way of sealing the cap should be introduced, to prevent leakage. Furthermore, it indicated that special attention should be given during the sublimation of the dry ice in the bottles, so that frost is avoided.

Similar were the results from the rest damaged bottles that were collected from the lab's floor...

First Experiments: observations and conclusions

After testing the bottles, the following observations were made for each sample: 

Sample 1

Sample 2

Sample 3

Samples 4-5

Samples 6-8

Samples 9-11

Sample 12

Samples 13-15

Samples 16-18

Samples 19-21

Samples 22-26

In the chart below can be found the results of the above tests and the 
first measurements for the C02 leakage test:

First experiments: 14/10/11

Evaluating the results of the first experiments the following conclusions were drawn: 

·      Types “3”, “4”, “14” showed the best performance

·      The weakest areas of the bottles are located around the neck, at the base, seams caused by its production method

·      Bottles were surprisingly rigid even with a low content of CO₂. The optimal relation between rigidity and deformation needs to be defined, as well as the stability of the samples under compression

·      Screw caps for soft drinks, because they are designed to be pressure resistant, can withstand large values of pressure, whereas 1-start thread caps fail

First experiments: the process

During the first day of experiments (14^th October) the following procedures took place:

·         different types of bottles were tested for burst strength

·         bottles of the same type were filled with solid C0₂ ranging from 2 to 34,5 gr.

The stable samples will be used for the compression test scheduled on the 17^th of this month

·         leakage through the cap was evaluated and different sealing methods were attempted

·         first set of measurements was taken to determine CO₂ leakage.

      Next measurements will take place on the 17^th of this month

·         12 bottles of the same type were filled with 6gr of solid C0₂ and will be tested together under compression (lateral)

·         The elastic deformation of the different samples was observed

Types of bottles

List of measurements on the samples

 

Based on the outcome of the experiments:

·         A comparison between PET bottles regarding burst strength was formulated

·         The areas of the PET bottles most prone to cracking were located 

·         A first perception of the relationship between CO₂ content and the resulting rigidity was obtained

·         The performance of screw caps for carbonated and non-carbonated products

was compared

In short, the steps of the experiment were the following:

Insert dry ice in the bottles

Examine deformation and cracking

Repeat process varying the dosage, the bottle type 

and the sealing method

From the bin to the floating platform

Four bins were placed by coffee corners and restaurants of the faculties CiTG and BK, TU Delft, to gather the first set of PET bottles (0,5L) that will be used for the experiment. 

Bins places in CiTG and BK to gather PET bottles

Ideally all the bottles needed for the floating platform (that can reach the number of 18.165!) can be collected from the TU Delft campus and the student houses in Delft. 

I never saw, by the way, in TU Delft’s dining areas and coffee corners recycling bins for plastic. In that sense, planning this campaign to gather material for the ReVolt House is not only stressing the fact that the building industry can absorb this waste and make constructive use of it, but also raises the issue of recycling in the university’s premises.  

First set of gathered bottles

Dry ice: how much is too much?

Dry ice is solid carbon dioxide. At normal atmospheric pressure and temperature it sublimates (changes directly from a solid to a gas), resulting in a tremendous increase in volume.

To calculate the amount of pressure that occurs in a 0,5L PET bottle when a specific amount of dry ice is inserted and then heated up, we can apply the  Van der Waals equation, which is a revised version of the ideal gas equation, used to more accurately describe the behavior of real gasses.

Therefore, if, for example, we have a PET bottle of V=0.5L and room temperature 

T= 20° C +273= 293K, then 3gr of dry ice are going to produce 3.23atm pressure.

According to design specifications, a PET bottle for carbonated products can hold up to 13,6 atmospheres of pressure before it bursts. This level of pressure can be reached if 13,3 grams of powdered dry ice are inserted in the bottle.

The maximum level of pressure that a bottle can hold depends of course on the thickness of the PET vessel (ex. 0.5 liter Sprite PET bottle, 0,45 mm thick has burst pressure =13.6atm) and the configuration of the bottle (ex. a five-footed bottle base gets less distorted than a four-footed base). Nonetheless, 13gr of dry ice can be the starting value for the experiments.

Performance of 20g and 22gr 0,5L bottles in 4- and 5-foot base variations

Source: Raising the bar in PET bottle lightweighting - WRAP