ArticlePDF Available

Measurement of water vapour ingress in PET bottles and correlation with oxygen and carbon dioxide permeation

Authors:

Abstract

The beer and beverage industry is using ever more barrier enhanced plastic bottles for the filling of its products. The quality of the products can be considerably affected by the permeation of oxygen into the bottle and carbon dioxide out of it. The quality control of the bottles with particular emphasis on the gas barrier is thus of great importance. However, the conventional gas permeation measuring method needs too much time. In order to respond effectively and quickly to barrier defects, bottle production or incoming goods inspection measuring time must be shortened, for example by 2 hours. A physical problem of a quick measurement of oxygen is the comparably long unsteady state of permeation due to desorption of oxygen into the bottle after filling. In order to overbear this difficulty methods are tested which use other gases or as in this instance water vapour. Instead of a complete permeation only the migration of water from PET into the bottle inner is measured. The ruggedness of the method meets the requirements of the practical measurement conditions. The correlation of the water vapour migration rate with the permeation of carbon dioxide and oxygen measured with a real-time method is linear. Active barriers employing scavenger material can not be detected by the water vapour ingress measurement.
105 May / June 2008
Authors: Prof. Dr.-Ing. Jan Schneider, University of Applied Sciences
Ostwestfalen-Lippe, Department of Beverage Technology, Liebigstr. 87,
D-32657 Lemgo, Germany;
Dipl.-Ing. Ingrid Weber, Dipl.-Ing. Roland Pahl, Research Institute
for Engineering and Packaging in the Beverage Industry, VLB Berlin,
Seestrasse 13, D-13353 Berlin, Germany
Tables and fi gures see Appendix
J. Schneider, I. Weber and R. Pahl
Measurement of Water Vapour Ingress in PET
Bottles and Correlation with Oxygen and Carbon
Dioxide Permeation
The beer and beverage industry is using ever more barrier enhanced plastic bottles for the fi lling of its products. The
quality of the products can be considerably affected by the permeation of oxygen into the bottle and carbon dioxide out of
it. The quality control of the bottles with particular emphasis on the gas barrier is thus of great importance. However, the
conventional gas permeation measuring method needs too much time. In order to respond effectively and quickly to bar-
rier defects, bottle production or incoming goods inspection measuring time must be shortened, for example by 2 hours. A
physical problem of a quick measurement of oxygen is the comparably long unsteady state of permeation due to desorption
of oxygen into the bottle after fi lling. In order to overbear this diffi culty methods are tested which use other gases or as in
this instance water vapour. Instead of a complete permeation only the migration of water from PET into the bottle inner is
measured. The ruggedness of the method meets the requirements of the practical measurement conditions. The correlation
of the water vapour migration rate with the permeation of carbon dioxide and oxygen measured with a real-time method is
linear. Active barriers employing scavenger material can not be detected by the water vapour ingress measurement.
Descriptors: multilayer, oxygen, PET bottle, permeation, quick test, water vapour
1 Introduction
This article is the second part of a study of which the fi rst part
has been published before [15]. The role of plastic bottles and
the necessity of the quality control were discussed in the fi rst
paper already: an objective comparison of the barrier properties
of various bottles and closure types is an important prerequisite
in being able to forecast a product’s shelf life, and hence in selec-
ting the most appropriate container-cap combination, as several
authors dealing with this topic agree [7, 14, 18]. Since in the food
packaging industry plastic packaging materials have been used
for a long time a carrier gas method for fl exible food packaging
has become a matter of a German standard (DIN 53380-3) [3,
19]. A modifi cation of this method permits the measurement of
packaging systems such as bottles [14]. Beside other diffi culties
as the creep effect (expanding of carbonised bottles) the problem
is the minimum test time of this method. It is physically limited to
about 2 days. Other so-called short-time tests are the manometric
methods (absolute pressure method) which are not specifi c to a
specifi c kind of gas, and a method that employs gas chromatography
after taking samples [4, 5, 9, 10, 11, 12, 13, 16, 19].
In order to investigate PET bottles with a high degree of accuracy,
a so-called real time method can be used today [14, 19]. Real time
means that the test time corresponds to the real time of the product
shelf-life. In the case of PET bottle production or purchase thereof
no useful tool is existent. Even a test time of 2 days as is possible
with the DIN method is not quick enough. The aim is to provide
a method that is fast, rugged and accurate enough for decisions in
either bottle production control or incoming goods control.
2 Equipment and methods
2.1 Measurement of water vapour migration into PET
bottles
The measuring apparatus was developed as a quick tester by the
company SIG Corpoplast (Hamburg, Germany). A humidity sensor
inside of a bottle measures the humidity that migrates from the
bottle wall through a perhaps existing inner coating layer into the
bottle inner (Fig. 1). With the help of the tester the barrier layer is
supposed to be evaluated within a few minutes. The increase of
humidity is measured. Water is comparably well soluble in PET
and through a defi ned exposition of the bottle in a climatic chamber
the moisture content can be increased and adjusted with a high
precision. Afterwards the bottle is fl ushed with compressed air
wile one minute in order to obtain a dry atmosphere in the inner
of the bottle. At this point a dew point humidity sensor (Fa 300-2,
CS-Messtechnik) is introduced according to Figure 1. The wor-
May / June 2008 106
king range is from –80 °C to 20 °C. Simultaneously the inner of
the bottle is tightened to the environment. The measuring time
including the fl ushing is 6 minutes. Every 20 seconds a measuring
of the humidity takes place.
2.2 Real-time method
As a reference method the real-time measurement oxygen ingress
and carbon dioxide loss is used. This method described separately
[19, 15] has gained acceptance in the brewing industry because of
its high reliability and the accordance to the product shelf-life.
2.3 Tested PET bottles
A number of exemplary monolayer and barrier enhanced PET
bottles were used in order to challenge the test method with a wide
spread of measuring values. The characterisation of the bottles is
described in Table 1. Multilayer bottles consist of more than one
layer, typically 3 or 5 layers. The inner and outer layers provide
the mechanical stability and the enclosed layer or layers represent
the gas barrier. Active barriers (scavengers) are also implemented
additionally in the passive barrier. Scavengers are able to chemically
bind passing oxygen. However an active carbon dioxide barrier
is not available. Bottles subsequently called “coating” consist of
a monolayer PET bottle with the addition of an inorganic (used
here) or organic coating such as a gas barrier.
3 Results and discussion
3.1 Development of a measurement category
In Figure 2 two series of measurements demonstrate exemplary
that the sensor can detect the changes in humidity properly. In
the initial phase a steep raise of humidity is displayed through the
measuring device. This is due to the technique of the measuring
sensor and could not be avoided. The glass bottle was therefore
exposed to a temperature of 23 °C and 50 % relative humidity
for 1 h (repeated determination). The experiment with the glass
bottles show in difference to plastic bottles (Fig. 3) a steady drop
of humidity within the measuring time of 320 s. These measure-
ments were used for the setting and the calibration of the drying
of the bottle material.
During the exposition of the bottle in defi ned climatic conditions
the PET takes up water according to the solubility. The curve
progressing in Figure 3 represents the water that migrates out of
the PET and that actually permeates through the internal coating
layer. Apparently a time of 220 s is necessary before the linear
increase of the humidity can be measured. It seems to make
sense to use a fi xed period between t1 = 220 s and t2 = 320 s as
measuring time. The change of humidity in this period was set
as the fi nal measuring size “H” (water vapour migration rate) for
the following experiments.
3.2 Repeatability (r)
Selected PET bottle (Coating 1 to 3) were measured daily and
each measurement were repeated 25 times. The data shown in
Figure 4 for the three selected bottles show a good repeatability
between rCoating 1-3= 0,069 and rCoating 1-3= 0,095.
3.3 Impact of the conditioning parameters
The impact of the conditioning parameters (temperature, relative
humidity, conditioning time) on the water vapour migration rate
was investigated. These measurements were carried out with the
coating bottle 2. As shown in Figure 5 there is no signifi cant dif-
ference in the measuring result by varying the relative humidity
between 50 % and 70 % (moisture of wet air 8 g/kg and 27 %
higher). By varying the temperature between 23 °C and 40 °C
(8 g/kg and 24 g/kg) the difference is as well not signifi cant (Fig.
6) corresponding to a signifi cance level of 95 % probability.
As soon as both conditioning parameters, temperature and relative
humidity are varied a signifi cant difference appears. As Figure 7
indicates that the incubation with a lower temperature and relative
humidity results in a signifi cantly smaller humidity migration
compared with the conditioning on a higher level in temperature
and relative humidity.
The conditioning time plays an important role as shown in
Figure 8: whilst there is no signifi cant difference in the cases of
more than 30 min conditioning time the trial with 15 min incubation
results in a signifi cantly higher amount of migrated vapour.
3.4 Ruggedness of the method
The impact of these parameters was verifi ed in a ruggedness test.
A rugged method must provide reliable results even in the case
of expected deviation in the infl uencing parameters. In Table 2
the factors are listed with their adjusted variation. The effects
documented too. The effects indicate a ranking of the factors
concerning the extent of their infl uence.
The result discovers a greater infl uence of the conditioning time
and the quantity of measuring points compared with the relative
humidity and the temperature. The raise of the conditioning tem-
perature from 95 min up to 105 min leads to mean increase of
the measuring result by the factor 3 according to a signifi cance
level of 95 %. The confi dence intervals of the effects range in the
case of all factors in a span covering zero indicating that there
is no signifi cant impact on the measuring result and hence the
method is rugged.
3.5 Measurement of selected bottles and correlation with
the oxygen and carbon dioxide permeation measure-
ments
In Figure 9 a comparison of the water vapour migration and the
real-time-permeation measurements (oxygen and carbon dioxide)
is shown. There is a clear tendency of qualitative correspondence
of the compared methods. A quantitative correlation between water
vapour and oxygen or carbon dioxide ingress can not derived from
these results. Regarding the coated bottles (coating 1, coating 2
and coating 3) a signifi cant difference between the bottles is re-
vealed with the reference method but has not been detected with
the proposed method. Multilayer bottle 1 indicates a high water
107 May / June 2008
vapour ingress in difference to the permeation of oxygen a carbon
dioxide. In case of the oxygen the explanation can be found in
the active barrier which infl uences the oxygen permeation but not
the water vapour. The correlation analysis by Bravais-Pearson
discovers a linear dependency between both methods. The corre-
lation coeffi cient is rxy= 0,67 for the comparison of water vapour
migration and oxygen permeation. If the results for multilayer
bottle 1 (active barrier) is taken out of account the correlation
coeffi cient is rxy= 0,94 for the comparison with the oxygen and
rxy= 0,96 with the carbon dioxide permeation.
4 Conclusion
A new quick method for the determination of the gas barrier of
plastic bottles measures the water vapour migration out of the
plastic material through the inner surface into the inner of a bottle.
The bottles must be incubated under defi ned climatic conditions
in order to obtain constant moisture in the plastic material. The
ruggedness and repeatability allow the use of the new method
considering the deviation of the infl uencing parameters under
practical conditions. The comparison with an established real-time-
permeation measurement (oxygen and carbon dioxide) reveals a
linear correlation. The water vapour migration test can physically
not be sensitive to active barriers. The test can be alternatively be
employed in case of incoming good test or in the process control
of bottle coating machines.
5 References
1 DIN 53380 Teil 3 07.98: Prüfung von Kunststoffen– Bestimmung der
Gasdurchlässigkeit– Teil 3: Sauerstoffspezifi sches Trägergasverfahren
zur Messung an Kunststofffolien und Kunststoffformteilen.
2 Draaijer, A., König, J. W., de Gans, O., Jetten, J., Douma, A.C.: A
novel optical method to determine Oxygen in Beer Bottles. EBC
Congress 1999, Cannes, France
3 Göbel, S.: Precise evaluation of barrier., PETplanet insider, Vol. 5,
(2004), no. 7/8, pp. 38-39/26-30.
4 Hartung, J., Elpelt, B., Klösner, K.-H.: Statistik. Lehr- und Handbuch
der angewandten Statistik. Wien: Oldenbourg, 2002.
5 Hertlein, J., Bornarova, K., Weisser, H.: Eignung von Kunststoff-
aschen für die Bierabfüllung. Brauwelt, 137 (1997), no. 21/22, pp.
860-866.
6 Jetten, J.: Evaluation of beer packaging. E.B.C. Symposium (Mono-
graph 30) im November 2000 in Oslo, Norwegen. Fachverlag Hans
Carl, Nürnberg, 2002.
7 Lundquist, L., Pelletier, C., Wyser, Y.: Oxygen permeability – Oxygen
transmission rate measurement using oxygen sensitive fl uorescent
tracers. Verpackungs-Rdsch. 11 (2004), pp. 69-72.
8 Mang, K.-P.: Sauerstoffmessung-Prinzip, Anforderungen an die
Einrichtung, neue Entwicklungen von Meßsystemen. Brauindustrie,
3 (2003), pp. 26-28
9 Manger, H.-J.: Die Sauerstoffmessung in der Brauerei und Geträn-
keindustrie. Brauerei Forum, 4 (2004), pp. 126-128
10 Manger, H.-J.: Kohlendioxid als Messobjekt in der Getränkeindust-
rie. Brauerei Forum, 12 (2003), pp. 337-339; 1 (2004), pp. 10-12; 2
(2004), pp. 38-40; 3 (2004), pp. 70–71
11 Müller, K.: O2-Durchlässigkeit von Kunststofffl aschen und Verschlüs-
sen- Messung und Modellierung der Stofftransportvorgänge (Diss.).
TU München, Lehrstuhl für Brauereianlagen und Lebensmittel-Ver-
packungstechnik, 2003.
12 Murer, G., Gautsch, J.: Selective CO2 measurement for beverages with
the new mutliple volume expansion method. Brauwelt International,
22 (2004), no. 3, pp. 176-178
13 Orzinski, M., Weber, I., Schneider, J.: Hohe Genauigkeit und gute
Reproduzierbarkeit. Permeationsmessung an der VLB Berlin. Brau-
industrie 11 (2004), pp. 66-71
14 Orzinski, M., Weber, I., Schneider, J.: New Requirements of the
Measurements of Permeation through Plastic Bottles and Closures.
MBAA TQ, vol. 42 (2005), no. 4, pp. 346-351
15 Orzinski, M.; Hunger, W.; Schneider, J.: Determination of Oxygen
Ingress and Carbon Dioxide Loss through Plastic Bottles using Per-
meation Simulation with Hydrogen. Brewing Sciences no. 2, 2007,
pp. 55-59
16 Orzinski, M.: Untersuchung der Permeation von anorganischen Gasen
und organischen Verbindungen durch barriereverbesserte Kunststoff-
aschen und ihrer messtechnischen Erfassung. pH. D. Thesis, TU
Berlin, 2007
Received 9 April, 2008, accepted 3 June, 2008
May / June 2008 108
Appendix
Table 1 PET bottles used as measuring objects
bottle description1 body height weight base wall barrier
identifi cation diameter [mm] [g] thickness thickness material2
[mm] [mm] [mm]
Multilayer 1 PET/PA+Sc/PET 65.5 238.3 27.9 0.24 0.32 10 %
Coating 1 PET + internal coating 66.5 239.4 27.8 0.24 0.38 0.14 µm
Coating 2 PET + internal coating 63.4 238.5 27.6 0.25 0.34 0.15 µm
Coating 3 PET + internal coating 65.4 238.1 27.8 0.28 0.32 0.15 µm
Coating 4 PET + internal coating 65.0 238.6 27.8 0.27 0.35 0.15 µm
Monolayer 1 PET 63.4 247.0 28.0 0.28 0.29
Monolayer 2 PET 63.4 245.0 28.0 0.29 0.28
1 PET (poly ethylene therephthalat), PA (poly amide), Sc (Scavenger)
2 Barrier material: mass percentage or thickness of coating layers
Table 2 Ruggedness verifi cation on the measurement category FH2; signifi cance level 95 %
variation of the infl uencing ruggedness (effects)
parameters
Factor lower limit upper limit unit effect lower upper
(infl uencing [ppm/min] confi dence confi dence
parameter) – + limit limit
[ppm/min] [ppm/min]
relative humidity 48 52 % 0,9 39,7 37,9
temperature 21 25 °C 1,1 39,9 37,7
time 95 105 min 2,9 35,9 41,7
quantity of quantity 36,2 41,4
measuring points 6 8 2,6
bottle
humidity sensor
Sealing
dew point measurment device
bottle
humidity sensor
Sealing
dew point measurment device
0%4
#OATING
3ENSOR
0%4
#OATING
3ENSOR
Fig. 1 Principle of the water vapour
migration (left) and detec-
tion with a humidity sensor
(right) [16]
109 May / June 2008
0
200
400
600
800
1000
1200
1400
1600
1800
2000
0 50 100 150 200 250 300
time [s]
Humidity [mg/m³]
measurment 1
measurment 2
n=25; alpha=0,05
0
200
400
600
800
1000
1200
1400
1600
0 50 100 150 200 250 300 350
time [s]
Humidity [mg/m³]
time t1
time t2
Fig. 2 Humidity measurements in the inner of 0.5 liter glass bottle (NRW) after a treatment in climatic chamber
(23 °C, 50 % relative humidity)
Fig. 3 humidity measured inside of an exemplary PET bottle with a internal coating barrier after a treatment in
climatic chamber (23 °C, 50 % relative humidity); confi dence intervals (n = 25, α = 5 %)
May / June 2008 110
0
0,2
0,4
0,6
0,8
1
1,2
1,4
1,6
0 5 10 15 20 25
numbers of measurment
vapour migration rate H [mg/(s m³)]
Monolayer 2
Coating 1
Coatin
g
2
n=15; alpha=0,05
-
200
400
600
800
1.000
1.200
1.400
1.600
200 220 240 260 280 300 320 340 360
time [s]
Humidity [mg/m³]
T = 23°C / rel.H.= 50 % / t = 60 min
T = 23°C / rel.H.= 60 % / t = 60 min
T = 23°C / rel.H.= 70 % / t = 60 min
Fig. 4 Repeatability (r) measurements of the water vapour migration rate H
Fig. 5 Impact of conditioning with different relative humidity on the migrated water vapour; conditioning time: 1 h,
temperature: 23 °C
111 May / June 2008
n=15; alpha=0,05
-
200
400
600
800
1.000
1.200
1.400
1.600
200 220 240 260 280 300 320 340 360
time [s]
Humidity [mg/m³]
T = 23 °C/ rel.H.= 50 %/ t = 60 min
T = 30 °C/ rel.H.= 50 %/ t = 60 min
T = 40 °C/ rel.H.= 50 %/ t = 60 min
n=15; alpha=0,05
-
200
400
600
800
1.000
1.200
1.400
1.600
200 220 240 260 280 300 320 340 360
time [s]
Humidity [mg/m³]
T=23°C/ rel.H.= 50 %/ t = 60 min
T=40°C/ rel.H.= 70 %/ t = 60 min
Fig. 6 Impact of the conditioning temperature on the migrated water vapour; conditioning time: 1 h, relative humidity
50 %
Fig. 7 Impact of the combined variation of the conditioning parameter temperature and relative humidity on the
migrated water vapour, conditioning time 1 h
May / June 2008 112
n=15; alpha=0,05
-
200
400
600
800
1.000
1.200
1.400
1.600
1.800
200 220 240 260 280 300 320 340 360
time [s]
Humidity [mg/m³]
T= 23°C/ rel.H.= 50 %/ t = 15 min
T= 23°C/ rel.H.= 50 %/ t = 30 min
T= 23°C/ rel.H.= 50 %/ t = 60 min
T= 23°C/ rel.H.= 50 %/ t = 180 min
0
0,5
1
1,5
2
2,5
Monolayer 1 Multilayer 1 Coating 1 Coating 2 Monolayer 2 Coating 3 Coating 4
vapour migration rate H [mg/(s m³)]
-1,0
0,0
1,0
2,0
3,0
4,0
5,0
6,0
7,0
8,0
O
2
ingress [mg/l]
CO
2
loss [g/l]
water vapour (320s-220s)
O2 (VLB)
CO2 (VLB)
Fig. 8 Impact of the conditioning time on the migrated water vapour; conditioning temperature: 23 °C, relative
humidity 50 %
Fig. 9 Results of the water vapour ingress measurements as beside the oxygen and carbon dioxide permeation measu-
rements by VLB method at selected bottles
ResearchGate has not been able to resolve any citations for this publication.
Article
Full-text available
The beer and beverage industry is using ever more barrier enhanced plastic bottles for the filling of its products. The quality of the products can be considerably affected by the permeation of oxygen into the bottle and carbon dioxide out of it. The quality control of the bottles with particular emphasis on the gas barrier is thus of great importance. However, the conventional gas permeation measuring method needs too much time. In order to respond effectively and quickly to barrier defects, bottle production or incoming goods inspection measuring time must be shortened, for example by 2 hours. A physical problem of a quick measurement of oxygen is the comparably long unsteady state of permeation due to desorption of oxygen into the bottle after filling. Hydrogen as a test gas can overbear this difficulty because of its high molecular mobility and low viscosity so a method for the use of hydrogen was established. The ruggedness of the method meets the requirements of the practical measurement conditions. The correlation of the hydrogen ingress rate with the permeation coefficient of carbon dioxide and oxygen measured with a real-time method is not linear but can be used to differentiate between bottles with good or poor passive barriers. Active barriers employing scavenger material can not be detected by the hydrogen ingress measurement.