More information from scratching around during Covid lockdown…
One of the underlying concerns with heat pasteurising is “bottle bombs”. There seems to be little data about how much pressure “standard” beer bottles can handle… lots of opinions but not much published information. However, I have come across some factual stuff that I thought worth sharing. As always, make up your own mind about how to use this information.
Basically, the situation is,
The fine print…
How much pressure does pasteurising generate
As outlined in some of my earlier posts, Andrew Lea’s website has a downloadable carbonation spreadsheet that calculates the pressure generated by different levels of carbonation at different temperatures (http://www.cider.org.uk/carbonation_table.xls). With a typical heat pasteurising process of 3 volumes of CO2, at 70C (158F), bottle pressure would be around 150 psi. With only 2.5 volumes of CO2, the pressure would be even less at 125 psi.
What research has been done on bottle strength
The short answer seems to be “not much!”
The only definitive stuff that I could find was a research study carried out at Ohio State University back in 1953* which examined surface stress of glass bottles under internal pressure. (* “An investigation of stresses in glass bottles under internal hydrostatic pressure”. PhD dissertation by Jo Morgan Teague Jr, B.S., M.S. Ohio State University. Google Ohiolink 5902326.pdf)
Although it looked at issues such as wall thickness, design and stress points, essentially the study looked at how much pressure bottles could take before they failed. There may be other studies, but this was the only one that I could find that specifically looked at the bursting pressure of bottles.
The study found that common 12 oz beer bottles had an average failure pressure of 396 psi for “Export Shape” bottles manufactured by the flow process, and 266 psi for “Select Shape” manufactured by the suction process. There isn’t much information about the bottle manufacturing processes except that they sound similar to the blow process used in the present day.
The bottles in the study were about 20% heavier than bottles in use today (8.5 oz bottles in 1953 vs 7.0 oz bottles in 2020). The Export shape was a bit like like “long necks” and the Select shape was a bit like “stubbies” but apart from the production process there didn’t seem to be much difference between them.
Comparing 1953 bottles with today’s bottles is difficult because both glass and manufacturing technologies may well have improved, with stronger glass formulations leading to getting the same results with less glass. Nevertheless, a conservative approach to what happens with today’s bottles might be to downgrade the 1953 results by 20% since today’s bottles are 20% lighter than the study bottles.
Of course, it isn’t the average bottle that might cause a bottle bomb, it is the weakest bottle which may have flaws or be underweight. The study for the “longneck” bottles had a failure spread of 200 psi to 600 psi. The “stubby” bottles had a failure spread of 300 psi to 700 psi. So, in the worst case, discounting these results by 20%, suggests that even the weakest bottle should stand 160 psi.
One approach to this issue might simply be to only use bottles weighing near to or above the 200g (7oz) “standard weight” of current 333ml (12 fl oz) bottles when heat pasteurising. i.e discard underweight bottles and use them for bottling cider that isn’t going to be heat pasteurised and subjected to high pressure.
Assuming that bursting pressure is related to the amount of glass in the bottle, then with this approach the bursting pressure of 333 ml (12 fl oz) bottles weighing 200g (7 oz) and above should be over 200 psi.
What actually happens
I ran a pressure test trial on some bottles of commercial beer that are normally carbonated to 2.5 volumes of CO2. These bottles were heated to a bit above 90C (195F) at which point the calculated bottle pressure was 180 psi. There were no failures. The water was just about boiling at our altitude (1000 metres) so I couldn't get them any hotter.
A further pressure test on salvaged James Squire, Peroni and Heinekin bottles with 3.0 volumes of CO2 resulted in a pressure of 230 psi with the only failure being a cap blowing off a Peroni Bottle. This seems to be consistent with the 1953 study results.
Interestingly a quick check of my assorted collection of salvaged bottles showed reasonable consistency in weight (i.e. the amount of glass that was in a bottle). The unbranded bottles were generally around 200 grams (7 oz), Premium James Squire around 195 grams (6.9 oz), Heinekin 205 grams (7.2 oz), and Peroni 185 grams (6.5 oz). There were however some bottles scattered among the others that were as light as 185 grams, so not all bottles are “created equal”.
Some other stuff
A post on HBT (by Beaudoin, 9 April 2013) reported that Grolsch claim that their bottles (which weigh 235g or 8.3 oz, which is about the same as the bottles in the 1953 OSU study) are rated at 290 psi but that the seals will “blow off” or leak at 70-80 psi.
It is interesting to note that Claude Jolicoeur refers to this at the end of his chapter on Ice Cider “I like to use beer bottles with a ceramic closure (e.g. Grolsch bottles), as these will let excess pressure exit and thus avoid bursting”.
ShangHai Misa Glass Co Ltd refers to China National Standard GB4544 “Beer Bottles” which indicates that new bottles should withstand the following pressures. The current standard for China is 1.2 MPa (174 psi), for Japan 1.8 MPa (261 psi) and for all other developed countries 1.6 MPa (232 psi). There is also an inference that recycled bottles will only achieve 75% of these figures.
(Beer glass bottle quality standards - MISA)
The U.S. Department of Commerce, National Institute of Standards and Technology. has published a Voluntary Product Standard for “Glass bottles for Carbonated Soft Drinks”. The minimum internal pressure that bottles should withstand is 1380 kPa (200 psi) for non-refillable bottles and 1550 kPa (225 psi) for refillable bottles.
(Full text of "Glass bottles for carbonated soft drinks")
One of the underlying concerns with heat pasteurising is “bottle bombs”. There seems to be little data about how much pressure “standard” beer bottles can handle… lots of opinions but not much published information. However, I have come across some factual stuff that I thought worth sharing. As always, make up your own mind about how to use this information.
Basically, the situation is,
Common 333 ml or 12 fl oz bottles weighing more than 7 oz will handle 200 psi pressure
Pasteurising cider with 3 volumes of CO2 at 70C (158F) generates about 160 psi in the bottle
Pasteurising cider with 3 volumes of CO2 at 70C (158F) generates about 160 psi in the bottle
The fine print…
How much pressure does pasteurising generate
As outlined in some of my earlier posts, Andrew Lea’s website has a downloadable carbonation spreadsheet that calculates the pressure generated by different levels of carbonation at different temperatures (http://www.cider.org.uk/carbonation_table.xls). With a typical heat pasteurising process of 3 volumes of CO2, at 70C (158F), bottle pressure would be around 150 psi. With only 2.5 volumes of CO2, the pressure would be even less at 125 psi.
What research has been done on bottle strength
The short answer seems to be “not much!”
The only definitive stuff that I could find was a research study carried out at Ohio State University back in 1953* which examined surface stress of glass bottles under internal pressure. (* “An investigation of stresses in glass bottles under internal hydrostatic pressure”. PhD dissertation by Jo Morgan Teague Jr, B.S., M.S. Ohio State University. Google Ohiolink 5902326.pdf)
Although it looked at issues such as wall thickness, design and stress points, essentially the study looked at how much pressure bottles could take before they failed. There may be other studies, but this was the only one that I could find that specifically looked at the bursting pressure of bottles.
The study found that common 12 oz beer bottles had an average failure pressure of 396 psi for “Export Shape” bottles manufactured by the flow process, and 266 psi for “Select Shape” manufactured by the suction process. There isn’t much information about the bottle manufacturing processes except that they sound similar to the blow process used in the present day.
The bottles in the study were about 20% heavier than bottles in use today (8.5 oz bottles in 1953 vs 7.0 oz bottles in 2020). The Export shape was a bit like like “long necks” and the Select shape was a bit like “stubbies” but apart from the production process there didn’t seem to be much difference between them.
Comparing 1953 bottles with today’s bottles is difficult because both glass and manufacturing technologies may well have improved, with stronger glass formulations leading to getting the same results with less glass. Nevertheless, a conservative approach to what happens with today’s bottles might be to downgrade the 1953 results by 20% since today’s bottles are 20% lighter than the study bottles.
Of course, it isn’t the average bottle that might cause a bottle bomb, it is the weakest bottle which may have flaws or be underweight. The study for the “longneck” bottles had a failure spread of 200 psi to 600 psi. The “stubby” bottles had a failure spread of 300 psi to 700 psi. So, in the worst case, discounting these results by 20%, suggests that even the weakest bottle should stand 160 psi.
One approach to this issue might simply be to only use bottles weighing near to or above the 200g (7oz) “standard weight” of current 333ml (12 fl oz) bottles when heat pasteurising. i.e discard underweight bottles and use them for bottling cider that isn’t going to be heat pasteurised and subjected to high pressure.
Assuming that bursting pressure is related to the amount of glass in the bottle, then with this approach the bursting pressure of 333 ml (12 fl oz) bottles weighing 200g (7 oz) and above should be over 200 psi.
What actually happens
I ran a pressure test trial on some bottles of commercial beer that are normally carbonated to 2.5 volumes of CO2. These bottles were heated to a bit above 90C (195F) at which point the calculated bottle pressure was 180 psi. There were no failures. The water was just about boiling at our altitude (1000 metres) so I couldn't get them any hotter.
A further pressure test on salvaged James Squire, Peroni and Heinekin bottles with 3.0 volumes of CO2 resulted in a pressure of 230 psi with the only failure being a cap blowing off a Peroni Bottle. This seems to be consistent with the 1953 study results.
Interestingly a quick check of my assorted collection of salvaged bottles showed reasonable consistency in weight (i.e. the amount of glass that was in a bottle). The unbranded bottles were generally around 200 grams (7 oz), Premium James Squire around 195 grams (6.9 oz), Heinekin 205 grams (7.2 oz), and Peroni 185 grams (6.5 oz). There were however some bottles scattered among the others that were as light as 185 grams, so not all bottles are “created equal”.
Some other stuff
A post on HBT (by Beaudoin, 9 April 2013) reported that Grolsch claim that their bottles (which weigh 235g or 8.3 oz, which is about the same as the bottles in the 1953 OSU study) are rated at 290 psi but that the seals will “blow off” or leak at 70-80 psi.
It is interesting to note that Claude Jolicoeur refers to this at the end of his chapter on Ice Cider “I like to use beer bottles with a ceramic closure (e.g. Grolsch bottles), as these will let excess pressure exit and thus avoid bursting”.
ShangHai Misa Glass Co Ltd refers to China National Standard GB4544 “Beer Bottles” which indicates that new bottles should withstand the following pressures. The current standard for China is 1.2 MPa (174 psi), for Japan 1.8 MPa (261 psi) and for all other developed countries 1.6 MPa (232 psi). There is also an inference that recycled bottles will only achieve 75% of these figures.
(Beer glass bottle quality standards - MISA)
The U.S. Department of Commerce, National Institute of Standards and Technology. has published a Voluntary Product Standard for “Glass bottles for Carbonated Soft Drinks”. The minimum internal pressure that bottles should withstand is 1380 kPa (200 psi) for non-refillable bottles and 1550 kPa (225 psi) for refillable bottles.
(Full text of "Glass bottles for carbonated soft drinks")
Last edited: