'Barrier Polymers'. In: Encyclopedia of Polymer Science ...nguyen.hong.hai.free.fr/EBOOKS/SCIENCE...

BARRIER POLYMERS

Introduction

Barrier polymers can be broadly defined as macromolecules having the ability tosignificantly restrict the passage of gases, vapors, and liquids. Since all polymersrestrict the transport of penetrants to some degree and the barrier performanceof polymers to different penetrants depends on a variety of factors, it is difficult toprovide a concise, objective definition. In a practical sense, however, the definitionof a barrier polymer depends upon the end-use requirements, and a material thatprovides sufficient barrier for a particular application can be considered to be abarrier polymer for that purpose. In the present discussion, polymers that haveresistance to transport of gases, vapors, and liquids as one of their key attributeswill be considered to be barrier polymers.

Polymers have found wide acceptance as alternatives to traditional materialssuch as glass, paper, and metals, in food, beverage, and other packaging indus-tries. A key characteristic of glass and metals as packaging materials is their totalbarrier to transport of gases and vapors. While polymers can provide an attrac-tive balance of properties such as flexibility, toughness, light weight, formability,and printability, they do allow the transport of gases and vapors to some extent.Unfortunately, an inexpensive, recyclable polymeric material possessing high bar-rier properties to every gas or vapor in addition to good mechanical, thermal, andoptical properties is not available. For this reason, the selection of a barrier poly-mer for a particular application typically involves tradeoffs between permeation,mechanical, and aesthetic properties as well as economic and recycling consider-ations. Additionally, there is an ongoing interest in optimizing property sets ofbarrier polymers to provide an efficient and economical method for packaging andfor extending the shelf life of packaged foods and beverages.

198Encyclopedia of Polymer Science and Technology. Copyright John Wiley & Sons, Inc. All rights reserved.

Vol. 5 BARRIER POLYMERS 199

This article discusses various types of barrier polymers and structures, theirpermeability characteristics and the underlying phenomena involved, measure-ment techniques, ways to predict and improve barrier properties, and current aswell as potential future applications for barrier polymers.

Fundamentals of Permeation in Polymers

The permeability coefficient, or simply permeability, characterizes the steady-state rate of mass transport of penetrant molecules through polymers. In a densepolymer film, the permeability P is defined as the molar flux of penetrant throughthe polymer relative to a fixed coordinate system NA normalized by the film thick-ness L and the difference between the upstream (p2) and downstream (p1) partialpressures (1):

P = NAL(p2 − p1)

(1)

Accordingly, permeability has dimensions of quantity of penetrant (eithermass or moles) times thickness divided by area, time, and pressure. Several unitshave been used to report permeability of gases and water vapor in the literature.In the United States, a commonly used unit for permeability of gases in barrierpolymers is (cm3(STP)·mil)/(100 in.2·day·atm). Table 1 provides conversion factorsfor several permeability units, including the SI unit (mol·m)/(m2·s·Pa), which iscommonly preferred in technical encyclopedias.

The steady-state transport properties of water vapor in barrier polymers arecharacterized by water vapor transmission rate (WVTR). The dimensions of WVTRare quantity of water transmitted through a film times thickness divided by areaand time, and a common unit for WVTR is (g·mil)/(100 in.2·day). Table 2 providesconversion factors for some WVTR units, including the SI unit (mol·m)/(m2·s).WVTR can be converted to water vapor permeability by dividing by the waterpartial pressure difference (which can be calculated from the specified relativehumidity and temperature).

Penetrant transport through polymers is described by the so-called solution-diffusion model (1). According to this model, permeation through a flat sheet orfilm occurs in three steps: penetrant dissolves into the upstream (ie, the highpartial pressure or high thermodynamic activity) side of the film, diffuses throughthe film, and desorbs from the downstream (ie, the low partial pressure or lowthermodynamic activity) side of the film. The rate-limiting step in this processis diffusion through the film. In one dimension, penetrant diffusion through apolymer typically follows Fick’s law:

NA = −DdCdx

(2)

where D is the effective diffusion coefficient for the penetrant in the polymer anddC/dx is the local concentration gradient of the penetrant.

When the downstream side penetrant partial pressure and concentrationare negligible relative to those on the upstream face of the film, using Fick’s law of

Table 1. Table of Common Gas Permeability Units with Conversion Factorsa

To obtain

Given Barrer cc·cmcm2·s·cm Hg

cc·cmcm2·s·atm

cc·cmcm2·s·Pa

mol·cmcm2·s·cm Hg

mol·mm2·s·Pa

cc·mil100 in.2·day·atm

cc·20µmm2·day·atm

Barrer 1 1.00 × 10− 10 7.60 × 10− 9 7.501 × 10− 14 4.461 × 10− 15 3.346 × 10− 16 1.668 × 102 3.283 × 103

cc·cmcm2·s·cm Hg

1.00 × 1010 1 76 7.501 × 10− 4 4.461 × 10− 5 3.346 × 10− 6 1.668 × 1012 3.283 × 1013

cc·cmcm2·s·atm

1.316 × 108 1.316 × 10− 2 1 9.869 × 10− 6 5.87 × 10− 7 4.403 × 10− 8 2.195 × 1010 4.32 × 1011

cc·cmcm2·s·Pa

1.333 × 1013 1.333 × 103 1.013 × 105 1 5.948 × 10− 2 4.461 × 10− 3 2.224 × 1015 4.377 × 1016

mol·cmcm2·s·cm Hg

2.241 × 1014 2.241 × 104 1.703 × 106 16.81 1 7.501 × 10− 2 3.738 × 1016 7.359 × 1017

mol·mm2·s·Pa

2.988 × 1015 2.988 × 105 2.271 × 107 2.241 × 102 13.33 1 4.984 × 1017 9.81 × 1018

cc·mil100 in.2·day·atm

5.996 × 10− 3 5.996 × 10− 13 4.557 × 10− 11 4.497 × 10− 16 2.675 × 10− 17 2.007 × 10− 18 1 19.68cc·20µm

m2·day·atm3.046 × 10− 4 3.046 × 10− 14 2.315 × 10− 12 2.285 × 10− 17 1.359 × 10− 18 1.019 × 10− 19 5.08 × 10− 2 1

aGiven permeability in the units shown in one element of the first column, convert it to the units shown in one element of the first row by multiplying theoriginal permeability by the factor at the intersection of the row and column of interest. For example, a value of 2 (cc·20 µm)/(m2·day·atm) is equal to (2 ×0.0508) or 0.1 (cc·mil)/(100 in.2·day·atm). In this table and throughout the article cc (or cm3) has been used to denote cubic centimeters of gas as measuredat standard temperature and pressure (STP) conditions, which are 0◦C and 1 atm.

200


Table 2. Table of Common WVTR Units with Conversion Factorsa

To obtain

Given mol·mm2·s

g·mil100 in.2·day

g·cmm2·day

mol·mm2·s 1 3.95 × 109 1.55 × 108

g·mil100 in.2·day

2.53 × 10− 10 1 3.94 × 10− 2

g·cmm2·day

6.45 × 10− 9 25.4 1

aGiven a WVTR value in units shown in one element of the first column,convert it to the units shown in one element of the first row by multiplyingthe original WVTR value by the factor at the intersection of the row andcolumn of interest. For example, a value of 2 (g·cm)/(m2·day) is equal to(2 × 25.4) or 50.8 (g·mil)/(100 in.2·day).

diffusion (eq. 2), permeability can be expressed as product of the effective diffusioncoefficient D and the solubility coefficient S, which is the ratio of the equilibriumpenetrant concentration in the polymer at the upstream side of the film dividedby the penetrant partial pressure or activity in the contiguous phase (1):

P = D× S (3)

According to equation 2, the diffusion coefficient is a kinetic term characterizingthe mass flux of penetrant through a polymer film in response to a concentrationgradient (1). Diffusion coefficients have units of (length)2/time, and are often ex-pressed in cm2/s. The solubility or partition coefficient is a thermodynamic factorthat links the equilibrium penetrant concentration in the polymer, C, with thepenetrant partial pressure contiguous to the polymer surface, p (1):

C = S× p (4)

When the penetrants of interest are vapors, liquids, or solids, the partial pressureis often replaced by penetrant activity. For an ideal gas, penetrant activity is equalto the ratio of penetrant partial pressure to its saturation vapor pressure (1). Fornonideal systems thermodynamic models must be used to estimate penetrantactivity (2).

The diffusion process of penetrants in polymers can be broadly classified intotwo categories: Fickian (which obeys the Fick’s law of diffusion) and non-Fickian.Penetrants in rubbery polymers and at low activities in glassy polymers typicallyexhibit Fickian behavior (3). The signature of Fickian diffusion in a thin polymerfilm contacted on both faces with a constant partial pressure (or activity) of pen-etrant is a weight increase due to penetrant absorption that is initially a linearfunction of the square root of the contact time and then asymptotically approachesa fixed equilibrium value (3). For Fickian diffusion-controlled kinetics of penetranttransport in a plane film whose thickness (L) is much smaller than its width orlength, the time required to reach steady state in a permeation experiment, tSS,is given by (3):

tss = L2

D(5)

202 BARRIER POLYMERS Vol. 5

The diffusion coefficients of large penetrants (eg, flavor and aroma compounds) inbarrier polymers, which can be of the order of 10− 14 cm2/s or less, coupled with atypical film thickness of 10 µm result in extremely large time scales (over 3 years)to reach steady-state transport. Hence, over the shelf life of packaged products(typically of the order of several months), flavor and aroma compounds may neverachieve equilibrium or steady-state conditions. In such cases, the steady-statepermeability does not provide sufficient information to predict package shelf life.As discussed later, more detailed knowledge of penetrant solubility and diffusivityis required to accurately predict migration of such compounds.

In glassy polymers, deviations from Fickian behavior can occur. These devia-tions are generally believed to arise as a consequence of the finite rate of polymerstructure reorganization in response to penetrant-induced swelling during thesorption–diffusion process (4). An example of so-called non-Fickian diffusion be-havior is the penetrant sorbing into the polymer in two stages, an initial Fickian-like stage followed by a protracted, slow drift toward the final equilibrium sorptionvalue (5). In such cases, the time required to achieve steady-state transport maybe much higher than that predicted based on equation 5. This type of diffusionbehavior is often observed when organic vapors at relatively high activity sorbinto amorphous glassy polymers (6). For example, toluene vapors exhibit Fickiandiffusion behavior in amorphous poly(vinyl chloride) (PVC) films at activities upto 0.4 and show increasingly non-Fickian behavior at higher activities (7).

Mechanism of Penetrant Transport in Dense Polymers

The rate-limiting step for penetrant diffusion is the creation of transient “gaps” inthe polymer matrix via local scale polymer segmental dynamics involving severalpolymer chains (8). Penetrant molecules vibrate inside local cavities in the polymermatrix at frequencies much higher than the frequency of polymer chain motionrequired to open a gap of sufficient size to accommodate the penetrant. These stepsare shown schematically in Figure 1. In Figure 1a, a penetrant molecule is showndissolved in a polymer matrix. The penetrant vibrates inside a gap or molecularscale cavity in the polymer matrix at very high frequency (ca 1012 vibrations/s or1 vibration/ps) (8). The polymer molecules do not occupy the entire volume of thepolymer sample. Because of packing inefficiencies and polymer chain molecularmotion, some of the volume in the polymer matrix is empty or “free” and thisso-called free volume is redistributed continuously as a result of the random,thermally stimulated molecular motion of the polymer segments (1).

In Figure 1b, local polymer segmental motion has opened a connecting chan-nel between two free-volume elements in the polymer matrix and the penetrantmolecule can, as a result of its own Brownian motion, explore the entire corridorbetween the initial free-volume element which it occupied and the second free-volume element which is connected to it via the opening of a transient gap inthe polymer matrix. Eventually, local segmental motion of the polymer segmentscloses the connection between the two free-volume elements and if the penetranthappens to be away from its original position, as shown in Figure 1c, when thegap in the polymer matrix is closed, the penetrant will be trapped in another free-volume element in the polymer matrix and will have executed a diffusion step. The


Local, penetrant-sized gapin polymer due to chainmotion

Penetrant dissolved inpolymer matrix

(a)

Opening of transient,penetrant-sized gap inpolymer matrix due tocooperative segmentalmotion

(b)

Gap closes behind penetrant,which has executed adiffusion step to a new regionof polymer matrix

(c)

Fig. 1. Schematic depicting mechanism of penetrant transport in polymers (8).

process shown in Figure 1 has been called the “Red Sea” mechanism of penetranttransport in polymers (8).

Figure 2 shows a schematic of two polymer chains undergoing coordinatedlocal segmental motion as a result of random, thermally stimulated movementsof the polymer chains to open a gap between the polymer chains of sufficientsize to permit passage of a penetrant molecule from one free-volume element toanother (9). This schematic emphasizes that the polymer segmental motion is therate-controlling step in penetrant diffusion.

Figure 3 provides a very simplistic schematic of the molecular processes in-volved in the local segmental motions of polymers that contribute to the formationof transient gaps in the polymer matrix important for penetrant diffusion. Thisfigure shows what is believed to be a typical example of intramolecular coopera-tive local segmental motion of the polymer backbone of a polyethylene chain. This


Prior to jump Penetrant

Polymer chain

Activated state

After jump completedld

ld

Fig. 2. Schematic depicting polymer chain position before, during, and after a diffusionstep by a penetrant molecule (9).

so-called crankshaft motion requires the cooperation of several adjacent ethyleneunits and can act to create gaps in the polymer matrix of sufficient size to accom-modate small penetrant molecules. It should be emphasized that the detailed un-derstanding of the molecular level motions in polymers that contribute to diffusionis evolving rapidly as a result of more detailed atomistic simulation of penetranttransport in polymers. As a result, more realistic descriptions of the importantmolecular processes for penetrant diffusion in polymers should be expected in thenear future.

The motion of polymer segments to produce a gap of sufficient size to accom-modate a penetrant molecule is much slower than the vibration of the penetrantin a gap in the polymer matrix. For example, computer simulations of oxygen dif-fusion in a polyimide reveal that the oxygen molecules execute a diffusion stepapproximately once every nanosecond (1 ns = 1000 ps) (10). The rate of produc-tion of gaps of sufficient size to accommodate penetrant molecules decreases withincreasing size of the penetrant. That is, there are fewer gaps produced per unittime in a polymer matrix of sufficient size to accommodate methane (kinetic di-ameter = 0.38 nm), for example, than there are for hydrogen (kinetic diameter =0.289 nm).

Computer simulation results of diffusion in poly(2,6-dimethyl-1,4-phenyleneoxide) (PPO) have been reported (11). Figures 4 and 5 present the results forthe displacement of an oxygen molecule and a nitrogen molecule, respectively, as a


1 2

3

4

56

7

12

3

4

5

Fig. 3. Crankshaft motion—An example of local segmental dynamics in polyolefins. Eachball represents, for example, a methylene (ie, CH2) unit in PE and the solid lines representthe covalent chemical bonds between neighboring CH2 groups. Such molecular motions arebelieved to be important in providing pathways for diffusion of small molecules in polymers.The crankshaft motion requires the simultaneous rotation of several contiguous methyleneunits about bonds 1 and 7 or bonds 1 and 5 (9).

0 100 200 300 400 500 600 700 800 900 1000

O2Diffusion jumps

0

0.5

1.0

1.5

2.0

2.5

3.0

Dis

plac

emen

t, nm

Time, ps

Fig. 4. Computer simulation of the displacement of an oxygen molecule in poly(2,6-dimethyl-1,4-phenylene oxide) as a function of time (11).


0

0.5

1.0

1.5

2.0

2.5

3.0

0 100 200 300 400 500 600 700 800 900 1000

N2

Diffusionjump

Time, ps

Dis

plac

emen

t, nm

Fig. 5. Computer simulation of the displacement of a nitrogen molecule in poly(2,6-dimethyl-1,4-phenylene oxide) as a function of time (11).

function of time. There is about one diffusion jump for oxygen (kinetic diameter =0.346 nm) every 300–350 ps but only one diffusion step for nitrogen (kinetic di-ameter = 0.364 nm) over the entire 1000 ps duration of the computer simulation(11). The oxygen molecule spends most of its time rattling within a small cage (orfree-volume element) with average displacements of the order of 0.2 nm or so. Thediffusion jumps occurring approximately every 300–350 ps involve displacementsof the oxygen atom of the order of 0.4–0.5 nm. As shown in Figure 5, the diffusionjump length for nitrogen is longer (approximately 1 nm) and the jumps occur lessfrequently. However, these results are obtained for very short periods of time ow-ing to computational limitations, and very long simulations would be required togenerate precise estimates of jump lengths and jump frequencies. Because of theextremely demanding computational resources required for such molecular-levelsimulations, they are only now becoming possible for small molecule migration inrelatively permeable polymers. As yet, computers are not fast enough to providerealistic simulations of phenomena such as migration of large flavor molecules inhigh barrier polymers.

Factors Affecting Permeability, Diffusivity, and Solubility

Free Volume. The dependence of penetrant transport properties on chainpacking in polymers is often described using correlations involving the fractionalfree volume (FFV) of polymers. FFV is the fraction of volume in a polymer that isavailable to assist in penetrant transport, and does not include volume occupied bypolymer molecules and volume in the polymer matrix that is otherwise unavailable


for penetrant transport. It is often estimated using group contribution methods.One popular method for estimating FFV is based on the following expression (1):

FFV = V − V0

V(6)

where V is the polymer specific molar volume and V0 is the so-called occupiedvolume that is not available to assist in penetrant transport. The occupied volumeis usually estimated by Bondi’s method as follows (12):

V0 = 1.3Vw (7)

where Vw is van der Waals volume of the molecule. A good estimate of Vw can beobtained from bond radii, van der Waals radii of constituent atoms, and geometricfactors. Bond radii are nearly constant from one molecule to another since thesame chemical bond will always have the same radius. The most complete list ofVw values is available in the compilation given in Reference 13. The dependenceof diffusion coefficients on FFV can be expressed as (14):

D= A exp( −B

FFV

)(8)

where A and B are empirical constants. The higher the FFV, the larger the dif-fusion coefficient. Figure 6 shows the effect of free volume on gas diffusion coeffi-cients in a series of substituted polysulfones (15).

The dependence of solubility on FFV is usually weaker than that of diffusiv-ity, especially in amorphous polymers (16). Therefore, permeability often followsa similar dependence on free volume as penetrant diffusivity. Attempts have beenmade to correlate FFV of polymers with gas permeability (17,18). As shown inFigure 7 (17), a nearly linear correlation was found to exist between the logarithmof oxygen permeability coefficients and the inverse of FFV in several families ofamorphous, glassy polymers and high barrier liquid crystalline polymers. Manybarrier polymers are glassy materials, since their use temperature is below theirglass-transition temperature. In glassy polymers, which are nonequilibrium ma-terials, free volume can be altered to some extent by the processing history ofthe sample (19). For instance, higher rates of cooling create higher free volumein the glassy state, and vice versa. A more effective way to alter free volume is tovary the chemical structure by, for example, adding or removing pendant groupson the polymer backbone (19). The presence of polar groups with low specific vol-umes can reduce the free volume (and hence, penetrant diffusion coefficients) byfacilitating more efficient packing of the polymer chains due to stronger interac-tions between them (19). For example, barrier polymers such as polyacrylonitrile(PAN) and poly(vinyl alcohol) (PVOH) have polar pendant groups, which lead tostrong energetic interactions between the polymer chains, efficient polymer chainpacking, low free volume, and, in turn, low permeability coefficients.

Free volume in polymers can be measured using probes such as elec-trochromic, photochromic, and fluoroscent probes, as well as xenon nuclear mag-netic resonance, small-angle x-ray scattering, density measurements, and positron


10−10

10−9

10−8

10−7

6 87 9 10

Diff

usio

n co

effic

ient

s, c

m2 /

s

1/FFV

CO2

CH4

O2

N2

Fig. 6. Correlation of gas diffusion coefficients with inverse of polymer fractional freevolume in a series of substituted polysulfones (15).

annihilation lifetime spectroscopy (pals) (20). Each method has its strengths andlimitations, and a simple, direct measure of FFV is not available. The pals tech-nique has, however, emerged in the past several years as a valuable nondestructiveprobe of free volume in polymers (20,21). PALS uses orthoPositronium (oPs) as aprobe of free volume in the polymer matrix. oPs resides in regions of reduced elec-tron density, such as free-volume elements, that typically range in radius from0.2 to 0.4 nm. This range of cavity radii compares well with nonbonded inter-atomic distances in polymers and molecular radii of diffusing penetrants (20).PALS permits an estimate of both the size and concentration of free-volume el-ements in the polymer matrix. Transport properties of barrier polymers, copoly-mers, and polymer blends have been well correlated with FFV as measured by pals(20–26).

Temperature. The temperature dependence of permeability and diffusiv-ity are usually modeled using Arrhenius equations of the following forms (1):

P = P0 exp( −Ep

RT

)(9)

D= D0 exp( −ED

RT

)(10)


10−2

10−1

100

101

102

103

104

1/FFV

5 6 9 107 8

Oxy

gen

perm

eabi

lity,

(cm

3�

mil)

/(10

0 in

.2�

day

� at

m)

Fig. 7. Correlation of oxygen permeability with inverse of polymer fractional free volumefor several families of amorphous, glassy, and liquid crystalline polymers (17). Polystyrene(35◦C), � polycarbonates (35◦C), polyesters (30◦C), � polyamides (25◦C), ♦ liquid crys-talline polymers (35◦C).

where Ep and ED are activation energies for permeation and diffusion, and P0 andD0 are preexponential factors. The effect of temperature on solubility is usuallyexpressed by a van’t Hoff relationship (1):

S= S0 exp( −�HS

RT

)(11)

where S0 is a preexponential factor and �Hs is the heat of sorption of penetrant inthe polymer. Since steady-state permeability is the product of diffusivity and solu-bility, the activation energy of permeation can be defined as the sum of activationenergy of diffusion and the heat of sorption (1):

Ep = Ed + �HS (12)

Ed is always positive; �Hs is often positive for light gases (such as H2, O2, and N2),but can be negative for larger, more soluble penetrants (such as C3H8 and C4H10).For polymers such as low density polyethylene (LDPE) and poly(vinyl chloride)(PVC), Ep is always positive (1). Therefore, permeability increases with increasingtemperature. To illustrate this behavior, Figure 8 shows the effect of temperatureon oxygen permeability of four widely used barrier polymers (27).


10−3

10−2

10−1

100

101

102

Temperature, °C

PET

AN

PVDC

EVOH 27

0 10 20 30 40 50

Oxy

gen

perm

eabi

lity,

(cm

3�

mil)

/(10

0 in

.2�

day

� at

m)

Fig. 8. Effect of temperature on oxygen permeability at 75% RH. PET is poly(ethyleneterephthalate), AN is an acrylonitrile–styrene copolymer, PVDC is vinylidene chloride–vinyl chloride copolymer (coextrusion resin grade), and EVOH 27 is ethylene vinyl alcoholcopolymer containing 27 mol% ethylene (27).

At temperatures far from the transition temperatures (eg, glass-transitiontemperature and melting point) the Arrhenius relationship (eq. 10) is obeyed,and with a known activation energy of diffusion, ED, the diffusion coefficient of apenetrant in a polymer can be estimated at any temperature. In cases where EDis not reported, it can be estimated using a known correlation (28,29) between D0and ED:

lnD0 = aED

RT− b (13)

where a and b are independent of penetrant type. The parameter a is indepen-dent of polymer type and has a universal value of 0.64 (30). b has a value of9.2 [–ln(10− 4 cm2/s)] for rubbery polymers (ie, polymers above their Tg) and11.5 [–ln(10− 5 cm2/s)] for glassy polymers (ie, polymers below their Tg) (13).Equation 13 is often referred to as a “linear free-energy” relation. Similar rela-tions between D0 and ED are observed for viscosity of organic liquids, molten salts,and metals (31) and for first-order chemical reaction kinetics (32), which are alsoactivated processes described by the Arrhenius equation. Additionally, commonbarrier polymers such as poly(ethylene terephthalate) (PET) and polycarbonate


10−2

10−1

100

101

ED/RT

D0,

cm

2 /s

5 10 15 20 25

Fig. 9. Correlation of D0 with ED/RT for various penetrants in glassy polymers (33).◦ Poly(ethylene terephthalate), bisphenol-A polycarbonate, � tetramethyl bisphenol-APC.

(PC) are known to follow this relation both above and below the glass-transitiontemperature, as shown in Figure 9 (33). Combining equations 10 and 13 gives

D= exp[

− b− (1 − a

) ED

RT

](14)

On the basis of a single value of the diffusion coefficient at one temperature, EDcan be estimated from equation 14. This equation can then be used to estimate thediffusion coefficient at other temperatures, provided that the two temperaturesdo not traverse thermal transitions (such as glass transition) and the polymermorphology is otherwise unchanged.

Chemical Structure. The presence of polar groups on or in polymer chainsoften increases chain rigidity, which can increase glass-transition temperatureand improve mechanical properties, and increases packing density (1). Conven-tional barrier polymers, such as PAN, have very low gas permeability as a result,in part, of restricted chain mobility due to the presence of polar groups. Polymerchain interactions can be quantified in terms of cohesive energy density (CED),and CED has a strong influence on penetrant diffusion. CED of a polymer is thesquare of its solubility parameter and characterizes the strength of attraction (orinteractions) between the polymer chains. It can be estimated using group contri-bution techniques (13). In a simple model of penetrant diffusion in polymers, theactivation energy for diffusion is directly proportional to the CED of a polymer


10−3

10−2

10−1

100

10+1

10+2

10+3

0.2 0.4 0.6 0.8 1.0

CED, kJ/cm3

PE PS

PVA

PVC

PAN

PVOH

Oxy

gen

perm

eabi

lity,

(cm

3�

mil)

/(10

0 in

.2� da

y� at

m)

Fig. 10. Correlation of oxygen permeability (23◦C, 0% RH) with CED for six barrier poly-mers: PE, PS, PVA, PVC, PAN, PVOH (36,37).

(34). On the basis of this model and the linear free-energy relation (35), the loga-rithm of penetrant diffusion coefficients should decrease linearly with increasingCED. Solubility of relatively nonpolar penetrants usually has a weaker depen-dence on CED than diffusivity, and hence, the logarithm of permeability shoulddecrease linearly with increasing CED. As shown in Figure 10, a nearly linearcorrelation was found between oxygen permeability and CED of barrier polymerswith permeability values ranging over 5 orders of magnitude (36,37).

Penetrant diffusivity, and hence permeability, can also be decreased byadding substituents to the polymer chain that reduce chain flexibility. Bulky sidegroups or rigid linkages such as aromatic groups decrease chain flexibility andhence reduce penetrant diffusion coefficients. Flexible linkages, such as ether ormethylene groups, produce the opposite effect (1). Several methods are used tocharacterize chain flexibility (1). The glass-transition temperature is a measureof long range or bulk molecular motion, and high Tg materials are usually rigidand inflexible. Sub-Tg relaxations are another indication of molecular motion, andcorrelations of O2 and CO2 permeability with sub-Tg relaxations have been ob-served within a family of amorphous polyesters and copolyesters (38). However,the exact nature of molecular motions which control penetrant diffusion are com-plex and unclear, and hence Tg and sub-Tg relaxations do not, in general, providepredictive correlations with penetrant diffusivity (1).

Changes in molecular structure of the polymer often affect more than onefactor influencing permeability and the net effect can be difficult to anticipate.


For example, addition of bulky side groups can stiffen the polymer chains, whichis expected to reduce the diffusion coefficients. However, the same modificationcould also decrease chain packing in the amorphous phase and reduce the levelof crystallinity in the polymer, which are expected to increase the diffusion coef-ficients (39). The net result of these competing effects can be difficult to predict apriori.

Crystallinity. Increasing crystallinity in a polymer generally decreases gaspermeability (16) (see SEMICRYSTALLINE POLYMERS). Crystallinity influences bothsolubility and diffusion coefficients. For most polymers and penetrants of interest,crystalline regions, which are much more dense and well ordered than amorphousregions, preclude penetrant sorption, thereby reducing penetrant solubility. Ad-ditionally, the presence of impermeable crystallites in a polymer matrix acts asbarriers to diffusion, increasing the path length for diffusion and, in some cases,increasing chain rigidity, which also reduces diffusion coefficients (16). Althoughcrystallite size, shape, and orientation do not usually influence solubility in poly-mers significantly, these factors can be important in penetrant diffusion. The effectof crystallinity on penetrant diffusion can be expressed using the following model(39):

D= Da

τβ(15)

where Da is the penetrant diffusion coefficient in the amorphous polymer, τ is ageometric impedance (ie, tortuosity) factor, and β is a chain immobilization factor.Impermeable crystalline regions force penetrants to follow a tortuous pathwaythrough the permeable amorphous regions. This effect is captured by the factor τ ,which is the ratio of the average distance traveled by a penetrant molecule to thethickness of the sample (39). τ can be a complex function of crystalline contentas well as crystallite size, shape, and orientation (16,40). Crystallites can alsorestrict segmental mobility by acting as physical cross-links. This effect is takeninto account by the factor β and is generally more pronounced in flexible rubberypolymers such as polyethylene (PE). In glassy polymers such as PET, the inherentrigidity of the chain backbone imposes more impedance to chain mobility than thecrystallites and hence, β is one (16) (see AMORPHOUS POLYMERS).

A two-phase model is often used to describe penetrant solubility in asemicrystalline polymer (41):

S= Saφa (16)

where Sa is solubility coefficient in the amorphous regions of the polymer and φa isamorphous phase volume fraction. This model assumes that the solubility of thecrystalline regions is zero, and that the presence of crystallites does not changethe amorphous phase solubility coefficient. For polymers used in barrier packagingapplications, the assumption of zero solubility in the crystalline regions is gener-ally accurate (42). The second assumption that the amorphous phase solubility isindependent of crystalline content is not necessarily obeyed, particularly in glassypolymers, whose state of amorphous phase structural organization may be influ-enced significantly by common processing protocols (eg, orientation, stretching,


100

101

102

103

104

0 0.1 0.2 0.3 0.4 0.5

Amorphous phase volume fraction

CO2

O2

N2

Polyethylene, T = 25°C

Gas

per

mea

bilit

y, (

cm3

� mil)

/(10

0 in

.2� d

ay� at

m)

Fig. 11. Effect of crystalline content on gas permeability in PE at 25◦C and 0% RH (47).

annealing, and contact with crystallization-inducing agents) (43–46). Neverthe-less, the simplest and most commonly used model for the effect of crystallinity onsteady-state permeability is based on these assumptions and is expressed by thefollowing relationship:

P = (Saφa)(

Da

τβ

)(17)

Figure 11 shows the effect of crystallinity on gas permeability in PE at 25◦C. Per-meability decreases with increasing crystallinity primarily because of decreasingdiffusion coefficients. The effect of crystallinity is more pronounced on tortuosityfactor τ than the chain immobilization factor β (47). In glassy polymers, the mostwidely used approximations for τ and β are τ = φa

− 1 and β = 1. Introducing thesevalues into the above equation yields P = Paφa

2, where Pa is the amorphous phasepermeability.

In certain polymers, the simple assumptions of the two-phase model do nothold. For example, poly(4-methyl-1-pentene) (PMP), which is a highly permeablepolymer, has a very low density crystal structure, and hence, penetrant moleculescan sorb into its crystalline phase (42). In PET, it has been observed that thepresence of crystalline regions increased the concentration of sorption sites in the


amorphous regions remaining in the polymer after crystallization (41). Othershave recently examined this phenomenon in PET in more detail (48,49). It hasbeen reported (43) that exposing amorphous PET to a strongly sorbing penetrantat high activity results in penetrant-induced crystallization. This process affectsthe amorphous phase penetrant solubility of PET, resulting in a marked increasein overall solubility with increasing crystallinity. For example, acetaldehyde sol-ubility in PET increased by more than 300% as a result of penetrant-inducedcrystallinity of about 36 wt% (43). This effect was attributed to the creation ofmicrovoids in the polymer as a by-product of penetrant-induced crystallization(43). Evidence for microvoid formation in PET due to exposure to strongly sorbingpenetrants has been presented in studies investigating solvent treatments to im-prove the dyeability of PET yarns (44), and in studies investigating the effect ofcrystallizing liquids on the morphology of PET (45,46). For example, it has beenreported that dye uptake in PET fibers exposed to dimethylformamide (DMF) wasfour to five times higher than in unexposed fibers (44). The dye diffusion coefficientwas also more than 2 orders of magnitude higher in the DMF-exposed samplesthan in the unexposed samples (50).

Chain Orientation. Stretching or drawing of polymer films can improvemechanical properties, and under certain conditions, barrier properties (see FILMS,ORIENTATION). The degree of chain orientation achieved is dependent on the drawratio and other process conditions (16). Orientation is usually characterized bybirefringence and quantified by the Herman’s orientation function f (51):

f = 12

(3 cos2θ − 1) (18)

where θ is the average angle between the polymer chain axis and the drawdirection.

Depending on the mode of deformation and the physical processes that oc-cur during orientation, permeability may either increase or decrease with in-creasing orientation (16). Impermeable polymer crystallites may become orientedinto plate-like structures during deformation, and this process generally de-creases penetrant diffusivity by increasing tortuosity (16). In addition, drawing ofsemicrystalline polymers can improve barrier properties through stress-inducedcrystallization and orientation of the remaining amorphous phase. Hence, the re-duction in permeability caused by orientation of crystallizable polymers can begreater than that in noncrystallizable polymers (16). The dramatic effect of orien-tation is supported by oxygen permeability data for PET in the literature. A 4 ×biaxial orientation (ie, draw ratio = 4 in each axis) decreased the permeabilityof oxygen in PET by a factor of about 2 (52). For other systems, however, in-creases in permeability upon biaxial orientation have also been reported. A vinyli-dene chloride (VDC)–vinyl chloride copolymer, for example, showed an increasein oxygen permeability from 0.2 (cm3·mil)/(100 in.2·day·atm) to 0.3 (cm3·mil)/(100 in.2·day·atm) upon 2.5 × biaxial orientation (53). The permeability increasewas attributed to microvoid development during orientation of the polymer chainsafter crystallinity was fully developed. Table 3 shows the effect of orientation onoxygen permeability of semicrystalline and amorphous barrier polymers (53). Insemicrystalline VDC copolymer and nylon MXD-6 polymers, under the conditionsstudied, the orientation process results in a slight increase in oxygen permeability


Table 3. Effect of Orientation on Oxygen Permeability Characteristics ofSemicrystalline and Amorphous Barrier Resinsa

Barrier polymer resins Oxygen permeability,b (cm3·mil)/(100 in.2·day·atm)

Semicrystalline resinsVDC copolymerc

Compression molded film 0.20 ± 0.02Extrusion cast film 0.20 ± 0.01Biaxially oriented – 2.5 × 0.30 ± 0.01Aromatic Nylon MXD-6d

Extrusion cast 0.37 ± 0.09Biaxially oriented – 2 × 0.39Amorphous resinsAmorphous Nylon Selar PA 3426e

Extrusion cast film 1.40 ± 0.31Uniaxially oriented – 2.5 × 1.14 ± 0.07Biaxially oriented – 2.5 × 1.01 ± 0.01Polyacrylic-imide XHTA-50A f

Extrusion cast 3.12 ± 0.17Uniaxially oriented – 2 × 2.95 ± 0.04Uniaxially oriented – 2.5 × 2.84Biaxially oriented – 2 × 2.76 ± 0.03aRef. 53.bAt 23.5◦C, 65% RH.cDow Chemical Company’s vinylidene chloride–vinyl chloride copolymer (experimental grade XU32009.02).dTrademark of Mitsubishi Gas Chemical Co., Japan.eTrademark of E. I. du Pont de Nemours & Co., Wilmington, Del.f Trademark of Rohm & Haas Co., Philadelphia, Pa.

with chain orientation, whereas the reverse is true for amorphous Selar 3426 andpolyacrylic-imide barrier polymers.

Penetrant Concentration (or Partial Pressure). The influence of pen-etrant concentration on solubility, diffusivity, and, in turn, permeability varies,depending on the penetrant–polymer system. Rubbery and glassy polymers typ-ically show little or no concentration dependence for solubility, diffusivity, andpermeability of light gases such as H2, N2, and O2. Consistent with this notion,Figure 12a shows essentially no influence of pressure on H2 permeability in PE(19). Gases such as CO2, which are more soluble than light gases, typically havea permeability–presssure response in glassy polymers similar to that shown inFigure 12b. Permeability decreases monotonically with increasing pressure aspredicted by the dual-mode sorption model (54). The magnitude of the permeabil-ity decrease depends upon the amount of so-called nonequilibrium excess volumein the polymer, which can increase with increasing Tg, the affinity of the pene-trant for the nonequilibrium excess volume, and the mobility of the penetrant inthe nonequilibrium excess volume relative to its mobility in the equilibrium freevolume (1). The permeability of a rubbery polymer to an organic vapor often ex-hibits the behavior shown in Figure 12c. The monotonic increase in permeabilityis often due to increases in penetrant solubility with increasing pressure coupledwith increases in diffusivity with increasing pressure (1). The response shown in


(b)CO2/PC

800

1200

1600

0 100 200 300

Per

mea

bilit

y

Pressure, psia

(d )Acetone/EC

0 4 8 12

Per

mea

bilit

y

Pressure, psia

3�105

4�105

5�105

6�105

1500

1600

1700

1800

0 100 200 300

Per

mea

bilit

y

Pressure, psia

(a)H2/PE

(c)C3H8/PE

4000

8000

12000

0 40 80 120

Per

mea

bilit

y

Pressure, psia

Fig. 12. Typical permeability–pressure dependence in rubbery and glassy polymers. (a)Hydrogen in PE at 30◦C, (b) carbon dioxide in PC at 35◦C, (c) propane in PE at 20◦C,(d) acetone in ethyl cellulose (EC) at 40◦C (19). The permeability values have units of(cm3·mil)/(100 in.2day·atm).

Figure 12d is typical for strongly interacting penetrants (eg, organic vapors) inglassy polymers at sufficiently high penetrant partial pressures. It can be viewedas a superposition of the behaviors in Figures 12b and 12c (1). The sharp increasein permeability begins as the penetrant plasticizes the polymer. Plasticizationoccurs when penetrant molecules dissolve in the polymer matrix at sufficient con-centration to force polymer chain segment separation, thereby increasing the freevolume, and in turn, facilitating polymer segmental motion. This increase in seg-mental mobility, which may be observed by the depression in Tg, results in anincrease in penetrant diffusion coefficients and, in turn, permeability (1).

Humidity. The absorption of water can increase, decrease, or have no effecton gas permeability of barrier polymers (27). Increasing the relative humidity(RH) from 0 to 50% increases the oxygen permeability of cellophane (regeneratedcellulose) by an order of magnitude, and exposure to 90% RH removes it fromthe class of high barriers by further increasing the permeability by more than an


10−3

10−2

10−1

100

101

102

0 20 40 60 80 100

Relative humidity, %

AmNY

BON

MXD-6

EVOH 44

EVOH 32

Oxy

gen

perm

eabi

lity,

(cm

3�

mil)

/(10

0 in

.2�

day

� at

m)

T = 20°C

Fig. 13. Effect of relative humidity on oxygen permeability of hydrophilic barrier poly-mers. AmNY is amorphous nylon (Selar), BON is biaxially oriented nylon-6, MXD-6 isoriented poly(metaxylylenediamine–adipic acid), EVOH 44 and 32 are ethylene vinyl alco-hol copolymers containing 44 and 32 mol% ethylene (27).

order of magnitude (27). For packaging of foods that require protection againstoxygen ingress, cellophane is coated or laminated with water barriers such aspolyolefins (27). Other hydrophilic barrier polymers, with the exception of certainamorphous polyamides, also lose their barrier properties with increasing RH, asshown in Figure 13. This is because water acts as a plasticizer and increases thefree volume of the polymer (55). However, at low to moderate RH, amorphouspolyamides and PET show slightly improved barrier properties with increasingRH (27). This behavior has been explained as the water molecules not swelling thepolymer, but occupying some of the polymer free-volume sites instead, resultingin reduction in permeability of other gases (56). VDC copolymers, acrylonitrilecopolymers, and polyolefins show essentially no effect of RH on gas permeability(57).

Techniques for Measuring Transport Properties

The determination of permeability, solubility, and diffusivity requires direct orindirect measurement of mass transfer under controlled conditions. The perme-ability of barrier polymers can be determined directly by measuring the pressurechange or other physical evidence of transfer or indirectly by using an indicatorof permeation, such as chemical reaction of the transferring gas with another


substance. The preferred methods of measurement differ for different classesof penetrants: light gases, water vapor, condensable vapors, and food flavor andaroma compounds.

There are two basic methods for measuring permeability: isostatic and quasi-isostatic (55). Isostatic methods employ a continuous flow on both sides of the poly-mer film to provide constant penetrant concentrations. Quasi-isostatic methodsuse a continuous flow to maintain constant penetrant concentration only on theupstream side and allow penetrant accumulation on the downstream side of thefilm. However, this accumulation is limited to a very low concentration, and hencethe penetrant partial pressure difference can be approximated as a constant (55).Figure 14a typically shows the course of an isostatic permeability experiment fora barrier polymer film of uniform thickness exposed to constant penetrant par-tial pressure p on the upstream side and constant removal of penetrant that haspermeated through the film to the downstream side. Using a specified initial con-dition (concentration in the film uniformly equal to zero) and boundary conditions(constant penetrant concentration C at the upstream side and zero penetrant con-centration at the downstream side), this situation can be described by the followingmathematical expression (58):

q = DCL

(t − L2

6D

)− 2LC

π2

∞∑n= 1

( − 1)n

n2exp

( − Dn2π2tL2

)(19)

where q is the total mass of penetrant permeating per unit film area in time t, Dis the diffusion coefficient, L is film thickness, and C is penetrant concentration atthe upstream side in equilibrium with the upstream penetrant partial pressurep. When steady state is reached, t becomes large enough to make the exponentialterm negligibly small, and the above equation reduces to

q = DCL

(t − L2

6D

)(20)

A plot of q vs t yields a straight line whose slope is the steady-state penetrant flux(NA = DC/L) and whose x-axis intercept is called the time lag (tL).

tL = L2

6D(21)

The time lag can be related to the time required to achieve steady state (tSS ∼2.7tL) (27). The diffusion coefficient can be calculated by rearranging the aboveequation (58):

D= L2

6tL(22)

Permeability can be calculated from equation 1 and the steady-state flux value(P = NA L/p = DC/p). It should be noted that for concentration-dependentdiffusion coefficients, tL will vary with the pressure difference across thepolymer film, and as a result, this simple time lag analysis may yield significant


tL

Steady state

t

0

q

t1/2 tSS

1.0

0.5

0

(a)

(b)

NA/(

NA) S

S

Fig. 14. (a) Mass of permeating penetrant per unit film area (q) as a function of time(providing a measure of time lag tL). (b) Normalized penetrant flux (NA) as a function oftime (providing a measure of halftime t1/2). tSS is time required to achieve steady state.

errors in the diffusion coefficients estimated using equation 22 (58). More generalexpressions for the time lag have been developed by including the concentrationdependence of D (59). Alternatively, a ‘concentration-averaged’ diffusion coeffi-cient can be obtained by plotting normalized penetrant flux (ie, flux at any timet divided by the steady-state flux) as a function of time (Fig. 14b). The diffusioncoefficient can be estimated using the following relationship (60):

D= L2

7.2t1/2(23)


Moist N2

Moist O2

Inside chamber Outside chamber

Barrier film

RH probe

RH probe

To sensor

Fig. 15. A schematic of a permeation cell in the Ox-Tran (MOCON, Inc.) oxygen trans-mission rate measurement system (62).

where t1/2 is the halftime (ie, time required for the penetrant flux to reach half ofits steady-state value). Thus, permeability, diffusivity, and hence solubility (S =P/D) can be determined from a single experiment. Given the ready availability ofcomputing power, it is now possible to use any of a variety of numerical techniquesto fit the entire response and extract the desired parameters and, when applicable,their concentration dependence.

Oxygen and Carbon Dioxide Permeation. The most widely used com-mercial instrument for measuring oxygen transmission rates of flat films andpackages is the Ox-Tran (Modern Controls Inc., Minneapolis, Minn.), and mea-surements are made in accordance with ASTM method D3985 (61). In this isostaticcoulometric method, flat film samples are clamped into a diffusion cell, which isthen purged of residual oxygen using an oxygen-free carrier gas such as N2. Thecarrier gas is routed to the instrument sensor until a stable zero has been estab-lished. Pure oxygen is then introduced into the outside chamber of the diffusioncell (see Fig. 15) (62). Oxygen molecules diffusing through the film to the insidechamber are conveyed to the sensor by the carrier gas. The Ox-Tran system usesa patented coulometric sensor (Coulox) to detect oxygen transmission throughboth flat films and packages. This sensor provides parts-per-billion sensitivity tooxygen even in the presence of water vapor. Digital pressure and flow controlsallow for RH control. Alternative instruments for measuring oxygen transmis-sion rates include Oxygen Permeation Analyzers from Illinois Instruments Inc.(Ingleside, Ill.).

Modern Controls, Inc. (MOCON) also makes instruments for measuring car-bon dioxide permeation. Their Permatran-C line of instruments uses an infrareddetector to detect carbon dioxide that permeates through the test film.


Water Vapor Permeation. WVTRs can either be measured by the tra-ditional gravimetric “cup” method (63) or by newer electronic instruments. Thenewer method (eg, ASTM method F1249 (64)) uses infrared detection to mea-sure water vapor transmission through barrier films. One of the most widely usedcommercial WVTR systems is Permatran-W (Modern Controls Inc., Minneapolis,Minn.). The newest model of this system (Permatran-W 3/31) uses a patentedmodulated infrared sensor to detect water vapor transmission through flat filmsand packages. It provides a sensitivity in the range of parts per million. Variousmodels are available with different temperature and RH capabilities.

Lyssy AG (Zollikon, Switzerland) also manufactures automatic water vaporpermeability testers designated as the L80 line of instruments. The L80-5000 isthe newest member and the fifth generation in this series.

In the traditional method (ASTM method E96) (27), a sample cell containingeither a desiccant or distilled water is covered with the sample film and placed ina controlled atmosphere. Typical conditions for the desiccant method are 100◦F(37.8◦C) and an external RH of 90%, although the standard also allows for temper-atures between 70 and 90◦F (21 and 32◦C) at (50 ± 2)% RH. The cell assembly isweighed periodically until steady state is reached. WVTR can be calculated fromthe steady-state rate of change in the weight of the cell.

Flavor and Aroma Compounds. Measurement of transport rates of fla-vor and aroma vapors in plastics is more complicated than that of either watervapor or light gases. Elaborate equipment and sensitive analytical devices are re-quired to obtain reliable results. Since the transport behavior of these compoundsis often strongly concentration dependent, measurements must be made in theactivity range in which the compounds are present in practice. Thus, some of themajor complexities include providing precisely mixed quantities of the condens-able vapor in an inert carrier such as nitrogen or argon at very low concentrations,typically a few parts per million, and assuring that the concentration is main-tained. Temperature must also be carefully controlled to prevent condensation onequipment surfaces (27).

As such, no single instrument has gained the widespread acceptance notedabove for instruments for O2, CO2, and water vapor, though several methods havebeen used to measure transport properties of flavor and aroma compounds inbarrier polymers. These include isostatic permeation techniques and gravimet-ric techniques (65). The permeation techniques directly yield permeability anddiffusivity of flavors in barrier polymers. Solubility can then be calculated indi-rectly using the relation P = D × S. The MAS 2000 Organic Vapor PermeationTest System (MAS Technologies Inc., Zumbrota, Minn.) is an example of a sys-tem that was commercially available for flavor permeation measurements. Massspectrometry and flame ionization detection have also been successfully used asvapor concentration detectors (65). In contrast, gravimetric techniques permit di-rect and independent measurements of both solubility and diffusivity (66). It ispossible to measure sorption and desorption of organic flavors in barrier polymersusing sensitive gravimetric sorption instruments such as the McBain spring bal-ance and the Rubotherm magnetic suspension balance (67–70). A schematic ofa McBain spring balance assembly is shown in Figure 16. The polymer sampleis suspended from a sensitive helical quartz spring inside the sorption chamber.After introducing the penetrant, the spring position relative to a fixed reference


Tovacuum

Vent

Liquid nitrogencold trap

Penetrantgas/vapor

inlet

To pressure read-outand power supply

Transducer(0−1000 torr)

12-Lgas/vaporreservoir

CCD camera

Water in

Water out

Quartz spring

Referencerod

Polymersample

Water-jacketedsorption cell

Computer

Fig. 16. A schematic of a McBain spring balance apparatus for measuring sorption anddesorption of organic vapors in barrier polymers (66).

rod hanging inside the chamber is recorded using a CCD camera. From the kineticuptake data, solubility and diffusivity values are estimated (71). Acceptable lev-els of agreement have been reported for solubility coefficients of ethyl acetate inLDPE, linear LDPE (LLDPE), and ionomer films obtained from gravimetric andisostatic permeation techniques (65).

Techniques for Predicting Transport Properties

Modeling Transport Properties of Gases and Condensable Vaporsin Polymers. The permeation of low molecular weight gases such as O2, N2,CO2 as well as large flavor and aroma compounds is an essential considerationin the selection and schematic of food packages and containers. Predictive modelsfor permeation would minimize the number of experiments required in packagematerial selection and development. Perhaps, more importantly, they also providean insight into the underlying factors controlling permeation in barrier polymers.

The permachor method has been used to predict permeabilities of low molec-ular weight penetrants in barrier polymers (72). Although originally developed


for O2, N2, and CO2, this method can be extended to other gases and vapors, pro-vided there is no specific interaction between the penetrant and the polymer. Ithas been successfully used for over 60 different polymers (72). In this method,numerical values (ie, group contributions) are assigned to polymer segments. Anaverage numerical value can then be obtained for the polymer, which is referredto as the permachor value of the polymer. A simple equation is used to relate gaspermeability P to polymer permachor value π :

P = A e− sπ (24)

where A and s are temperature-dependent constants. This method also takes intoaccount the reduction in permeability caused by the orientation of crystallinepolymers using the following expressions:

P =(

Aτ0

)e − sπ (25)

τ0≈1.13√φa

(26)

where τ 0 is tortuosity related to crystallite orientation and φa is the amorphousphase volume fraction. A good agreement has been reported between experimen-tal and model predictions of O2, N2, and CO2 permeability values in a varietyof polymers (72). This method has also been modified for predicting liquid per-meation through polymers (63). The permachor method works well for polymersand copolymers, but is not applicable to polymer blends (27). As with any groupcontribution method, caution should be exercised when attempting to performpredictions which are outside the data set used to generate the correlation.

Other methods for correlating gas permeability in barrier polymers withpolymer molecular structure have been developed using free-volume theory(17,18,73). Group contribution techniques can be used to estimate polymer free vol-ume from densities and intrinsic volumes of various polymer components. In themethod proposed in Reference 18, polymer-specific free volume was used, whichwas defined as (V−V0)/M, where V is specific volume, V0 is specific occupied vol-ume, and M is polymer molecular weight. V0 can be calculated according to Bondi’smethod from van der Waals volumes of the various groups in the polymer struc-ture (18). In this model, free volume was defined on a unit weight basis so thatvarious molecular structures could be compared on the same weight basis. Themodel predicts a linear relationship between logarithm of gas permeability andthe reciprocal of polymer-specific free volume. Other improvements to this modelhave been suggested (17,73). Gas permeability has often been correlated with FFV(as defined in eq. 6) using the following relation (73):

P = A exp( −B

FFV

)(27)

where A and B are constants for a particular gas. It has been observed that whenthis model is limited to a specific family of polymers, eg, polyesters and polyamides


(17), a reasonably good correlation can be obtained. However, when the correla-tion is broadened to include a wider range of polymer types, there is consider-able scatter in the data, particularly at low values of gas permeabilities. Eventhough these free-volume-based models have some fundamental basis for corre-lating transport properties, they have the following limitations (18,73): (1) theassumption of solubility being independent of free volume and polymer structureis clearly an approximation, (2) the concept of free volume cannot capture allthe factors affecting gas permeability (such as chain flexibility and CED), and (3)there may be errors in values of van der Waals volumes available in the literature.Attempts have been made to refine these models by introducing more empiricalparameters and making them more predictive, and these efforts have resulted insignificant improvements in the accuracy of the correlations (73).

Larger and more condensable penetrants, eg, flavor and aroma compounds,can have extremely low diffusion coefficients in common barrier polymers, result-ing in extremely large time scales to achieve steady state. For example, d-limonenehas diffusion coefficients of the order of 10− 14 cm2/s in PET (74). When coupledwith a typical film thickness of 10 µm, this leads to time scale to reach steady stateof more than 3 years. Hence, as noted before, over the shelf life of the packagedproduct (typically of the order of several months), flavor and aroma compoundsmay never reach steady-state transport. Therefore, independent predictions oftheir diffusion and solubility coefficients become necessary. Several methods havebeen proposed for predicting solubility coefficients. A widely used method is basedon a thermodynamic approach that relates penetrant sorption to solubility param-eters of the penetrant and the polymer (75). This dependence can be expressed asfollows:

S= S0 exp[(�Hvap − �Hmix)/RT ]

�Hmix = ν1φ2 (δ1 − δ2)2 (28)

where S0 is a constant for a particular polymer, ν1 is partial molar volume of thepenetrant, φ2 is volume fraction of polymer in the mixture, δ1 and δ2 are solubilityparameters of penetrant and polymer respectively, and �Hmix is enthalpy changeon mixing of penetrant molecules with polymer segments. The values of ν1, δ1, andδ2 can be obtained from the literature (36). The enthalpy change on vaporizationof the penetrant (�Hvap) can be calculated from the penetrant boiling point byusing available correlations (2). Reasonably good agreements have been reportedbetween model-predicted and experimentally observed solubility coefficients ofseveral penetrants in vinylidene chloride–vinyl chloride copolymers and LDPE at85 and 30◦C respectively (76).

For penetrants that interact with the polymer matrix primarily via disper-sion (ie, van der Waals) forces, penetrant solubility scales with measures of pene-trant condensability such as penetrant boiling point, critical temperature, or theforce constant in the Lennard–Jones potential model (43). The following relationbetween penetrant critical temperature and penetrant solubility has been derivedusing a classical thermodynamics model (13,77):

lnSa = N+ MTc (29)


0 100 200 300 400 500 600 70010−5

10−4

10−3

10−2

10−1

100

101

102

Critical temperature Tc, K

MIPKMEK

CH2 Cl2

(CH3)2 CO

CH3 OH C6H6 H2O

CH3 COOC2H5

C2H4O

n-C5H12

n-C4H10

i-C4H10

i-C5H12

CO2

CH4

O2

N2

He

Ar

C6 H5 CH3

Infin

ite d

ilutio

n pe

netr

ant s

olub

ility

, cm

3 (S

TP

)/(c

m3

� cm

Hg)

Fig. 17. Correlation of infinite dilution penetrant solubility with penetrant critical tem-perature in PET ((71) and unpublished data). Nonpolar penetrants, ◦ polar and quadrupo-lar penetrants, ♦ aromatic penetrants. MEK is methyl ethyl ketone and MIPK is methylisopropyl ketone. The slope M = 0.019 ± 0.001 K− 1 and intercept N = −9.6 ± 0.4.

In this expression, N is a parameter that depends primarily on polymer–penetrantinteractions and polymer free volume. Tc is the penetrant critical temperature,which is widely tabulated for many penetrants of interest (2). M is constant andhas a value of approximately 0.016 K− 1 for gas dissolution in liquids and in rub-bery and glassy polymers (13). While N varies from polymer to polymer, averagevalues of −9.7 and −8.7 for rubbery and glassy polymers at 35◦C, respectively,when solubility is expressed in cm3(STP)/(cm3·cm Hg) have been recommended(13). Penetrants with strong dipole or quadrupole moments may be more solublein a polar polymer matrix, such as PET, than predicted based on equation 29 (71).Although equation 29 is strictly valid for penetrant sorption in equilibrium ma-trixes, such as liquids or rubbery polymers, it also provides an excellent descriptionof equilibrium solubility in glassy polymers (71). Figure 17 shows the correlationof penetrant solubility in PET with penetrant critical temperature (71). Nonpolarpenetrants show excellent agreement with the model presented in equation 29 andpolar or quadrupolar penetrants exhibit significant scatter around the correlationline.

Over wider ranges of critical temperature, it has been suggested that pene-trant solubility coefficients may be better correlated with the square of reciprocalreduced temperature (Tc/T)2 (78):


ln Sa = n+ m(

Tc

T

)2

(30)

where T is the temperature of the experiment, and m and n are the slope and in-tercept of the correlation line, respectively. This equation may also be derived fromfundamental thermodynamic considerations (43). Other semiempirical methods,eg, UNIFAC group contribution model, have also been proposed for predictingpenetrant solubility in polymers (79,80).

Several predictive and correlative methods have been developed for diffusioncoefficients of penetrants in polymers. An empirical relationship has been devel-oped for correlating diffusion coefficients with penetrant critical volume (81):

Da = τ

Vηc

(31)

where Da is amorphous phase diffusion coefficient, Vc is penetrant critical volume,and τ , η are adjustable constants. This equation has been proposed based onanalogy with correlations of diffusion coefficients with critical volume of smallmolecules in liquids. For larger penetrants (eg, long-chain hydrocarbons), diffusionsteps may occur via motion of only part of the molecule, and critical volume isnot expected to capture the effective size of a penetrant unit participating in adiffusion step (43). In such cases, diffusion coefficients would be less sensitive topenetrant size than indicated in the above equation. Also, critical volume failsto capture the effect of penetrant shape on diffusion coefficients (71). Figure 18shows a plot of D vs. Vc for PET at 25◦C. Diffusion coefficients of penetrants (up tomolecular weights of 100 Da) in PVC, PS (polystyrene), and PMMA [poly(methylmethacrylate)] have been correlated empirically with other measures of penetrantsize, such as molecular diameter (82).

A theoretical model based on polymer free volume, temperature, and pene-trant size and shape has been developed (83). According to this model, diffusioncoefficients of large penetrants in amorphous rubbery polymers are given by

D= V2f

6

(eRTM

)1/2(

1

1Ac21

+ 1

2Ac22

+ 1

3Ac23

)exp

( −ED

RT

)(32)

where V f is average free volume per polymer chain segment, e is the base ofnatural logarithm, M is penetrant molecular weight, i is length of penetrantmolecule along a given direction i (taken as the principal axis of inertia), Aci iseffective penetrant molecular cross-sectional area perpendicular to the direction i,and ED is the activation energy of diffusion. V f can be calculated from the followingequation (83):

Vf = V [0.025 + αf(T − Tg2 + kw1)] (33)

where V is the total volume per mole of the polymer repeat unit, αf is tempera-ture coefficient of free-volume expansion, Tg2 is glass-transition temperature ofthe polymer, k is plasticizing efficiency of the penetrant for the polymer, and w1 is


10−16

10−14

10−12

10−10

10−8

10−6

0 100 200 300 400 500

Critical volume Vc, cm3/mol

He

O2

N2

CO2

H2OAr

CH4

C3H8

n-C4H10

n-C5H12

C6H5CH3

CH2Cl2

CH2Cl2

C6H6MIPK

i-C5H12

(CH3)2 CO MEK

Infin

ite d

ilutio

n pe

netr

ant d

iffus

ivity

, cm

2/s

Fig. 18. Effect of penetrant size on infinite dilution, amorphous phase diffusion coeffi-cients in PET at 25◦C ((71) and unpublished data). � Branched pentrants. MEK is methylethyl ketone and MIPK is methyl isopropyl ketone. The best-fit parameters of equation 31are τ = (5.7 ± 1.2) × 108 (cm2/s) (cm3/mol)9.1 and η = 9.1 ± 0.9.

weight fraction of the penetrant in the polymer. Equation 32 is strictly valid forlarge penetrants satisfying the size criterion Vs � V f, where Vs is effective stericvolume of the penetrant (83). The penetrant molecular shape dependence is rep-resented by the three-termed expression within the parentheses of equation 32.The equation can be applied to penetrants of a wide variety of molecular shapes(83). This model has been tested using data for diffusion of plasticizers in PVC(83).

“Migration Modeling” of Polymer Additives into Packaged Foodsand Beverages. One of the key applications of barrier polymers is food andbeverage packaging. Several low molecular weight components, eg, monomersand oligomers, as well as additives such as lubricants, stabilizers, and plasticizers,which are necessary for processing and stability, can be present in polymers usedfor packaging. Hence, there exists a potential for permeation (or migration) ofthese additives into the food or beverage with subsequent contamination (84–86). To ensure the safety of packaged food components, the U.S. Food and DrugAdministration (FDA) established a ‘Threshold of Regulation’ approach whichsets upper limits on the additive concentrations in the food (84). Since traditionalmigration testing methods are time consuming, expensive, and the analysis can


be difficult (especially at low penetrant concentrations), the FDA has developedmodels for predicting the additive concentration in the food simulant and rate oftransport of additives (84):

q = 2C0ρ

√Dptπ

(34)

where q is total mass of permeating species per unit surface area, C0 is initialadditive concentration in the polymer, ρ is polymer density, Dp is additive diffu-sion coefficient, and t is the package shelf life. This equation assumes that (84) (1)permeation is diffusion-controlled and follows Fick’s law, (2) no solubility-limitedpartitioning occurs between the polymer and the food, and (3) other external phasemass transfer resistances (eg, mixing, reaction with food) are negligible. The fol-lowing empirical equation has been developed for predicting diffusion coefficients(84):

Dp = 104 × exp[

Ap − a× MW − b(

1T

)](35)

where Dp is additive diffusion coefficient (cm2/s), Ap is constant for a particularpolymer, MW is additive molecular weight (g/mol), and a, b are correlation con-stants with values of 0.01 mol/g and 10,450 K respectively. The Ap values are 9for LDPE, −3 for PET, and 5 for HDPE and PP (84). A semiempirical model forpredicting diffusion coefficients has also been developed (87):

lnDP = lnA+ α(MW

)1/2 − K(MW

)1/3

T(36)

where A, α, and K are constants determined from experimental data. The diffusioncoefficients follow Arrhenius-type behavior and are taken to be independent ofpenetrant concentration (84). These models do not explicitly account for the effectof polymer crystallinity or orientation on additive diffusion coefficients.

The approach of the European Commission (EC), on the other hand, has beento assign ‘specific migration limits’ to different substances with adverse toxicologi-cal properties (86). A Fickian diffusion-based model, called the Piringer migrationmodel, which uses the ‘Migratest Lite’ program, has been used (85,86), and itsmathematical form is given below:

Mt = C0ρL(

α

1 + α

)[1 −

∞∑n= 1

2α(1 + α

) (1 + α +α2q2

n

)exp

(− Dptq2

n

L2

)],

α = 1K

VF

Vp, tan qn = − αqn

(37)

where L is film thickness, K is partition coefficient, and VF and Vp are the volumesof food and polymer, respectively. The diffusion coefficient can be estimated fromequation 35.

The sorption and transport of flavor and aroma compounds from the foodsimulant into the packaging walls can affect the migration of additives from the


walls into the food simulant. Neither approach (FDA or EC) takes this effectinto account. In general, these migration models provide conservative estimates(from a safety viewpoint) of additive concentrations in the food simulant and theirdiffusion coefficients. However, there are certain cases when they can fail (84) (eg,in cases where the flavor and aroma compounds from the food simulant plasticizethe polymer, or if the additive reacts either with the polymer or with the foodsimulant to produce a species that is not detectable), and attempts are beingmade to improve them.

Chemical Structures and Properties of Barrier Polymers

Barrier polymers can be broadly classified as high barrier and moderate to lowbarrier polymers, depending on the degree to which they restrict the passage ofgases such as O2 or CO2 and water vapor. The boundaries between these classifica-tions, while somewhat arbitrary, are based on the effect of the barrier properties ofthe polymer on the shelf life of the packaged products. This section discusses var-ious properties of different classes of barrier polymers. The selection of a barrierpolymer for a particular packaging application depends not only on its barrierproperties but also on other physical properties, and a comparison of physical,mechanical, and optical properties of some commonly used barrier polymers is,therefore, presented in Table 4 (88). Permeabilities of light gases (O2 and CO2)and water vapor are presented in Tables 5 and 6 respectively (27), and Table 7presents permeability, diffusivity, and solubility of flavor and aroma compounds invarious high and moderate barrier polymers (63). Permeability data of light gasesand water vapor in barrier polymers (93–97) are more widely available than thoseof flavor and aroma compounds (98–100). Proper caution should be exercised whenusing or comparing data from different sources since, as discussed in the previoussections, polymer permeability values depend on a wide variety of factors.

High Barrier Polymers. High barrier polymers are generally understoodto be those polymers which offer a high resistance to gas transmission. Thereare no specific limits for the gas transmission rates, but this category comprisespolymers with gas permeabilities low enough to significantly prolong the shelf lifeof packaged products.

Ethylene–Vinyl Alcohol Copolymers. The general structure of ethylene–vinyl alcohol (EVOH) resins is as follows (101):

EVOH resins are random copolymers of ethylene and vinyl alcohol madeby the hydrolysis of ethylene vinyl acetate copolymers (101). The leading manu-facturers are Kurray (EVALCA) and Nippon Gohsei (57). In commercial gradesof EVOH used in packaging, ethylene concentration ranges from 29 to 44 mol%(EVAL Co., U.S.). No additives are required in their processing as the presenceof ethylene units renders the otherwise intractable vinyl alcohol melt process-able in conventional molding and extrusion equipments. A high concentration of

Table 4. Physical, Mechanical, Optical, and Chemical Properties of Some Commonly Used Barrier Polymersa

Property HDPE LDPE PP PET PVC PS PVDC Nylon-6 EVOH

Density, g/cm3 0.945–0.967 0.915–0.925 0.90 1.4 1.22–1.36 1.05 1.6–1.7 1.14 1.14–1.19Glass-transition −55 −25 −20 80 ∼80 100 −17 50

temperature, ◦CYield, Pa− 1, 1 mil 4.1 4.2 4.4 2.9 2.8 3.8 2.4 3.5 3.3–3.5Tensile strength, MPa 17.2–41.4 10.3–34.5 137.8–206.7 172.2–227.4 27.5–55.1 55.1–82.7 55.1–110.2 172.2–255 8.2–11.7Tensile modulus, 0.9 0.1–0.3 2.4 4.8 2.4–4.1 2.7–3.3 0.3–1 1.7–2 2.0–2.6

1% secant, GPaElongation at break, % 200–600 200–600 50–275 70–130 100–400 2–30 50–100 70–120 120–280Tear strength, kN/mb — 17.5–87.5 175–263 175–350 17.5–52.5 52.5–175 0.35 87.5–140 —Chemical resistance Inert Inert Inert Inert Inert Inert (except Inert, Inert

oils, greases) sorbs waterHaze, % 3 5–10 3 2 1–2 1 1–5 1.5 1–2Light transmission, % — 65 80 88 90 92 90 88 90Heat-seal temperature 135–155 120–175 90–150 135–175 135–170 120–175 120–150 120–175 175–200

range, ◦CService temperature −40 to 120 −55 to 80 5 to 120 −70 to 150 −30 to 65 −60 to 80 −15 to 135 −70 to 200 −15 to 150

range, ◦FaRef. 88.bTo convert kN/m to lb/in., divide by 0.175.

231


Table 5. Oxygen and Carbon Dioxide Permeabilities of Various High and ModerateBarrier Polymersa

Gas permeability,b (cm3·mil)/(100 in.2·day·atm)

Barrier polymer Oxygen Carbon dioxide

EVOH, 27 mol% ethylene 0.03 (20◦C) 0.04 (20◦C, 65% RH)P(VDC–AN) barrier coating 0.04 0.1Liquid crystalline polymerc 0.06 (100% RH)EVOH, 44 mol% ethylene 0.07 (20◦C) 0.2 (20◦C, 65% RH)PVDC coextrusion resin 0.10 0.25Nylon MXD-6 oriented film 0.17BLOX 4000 series PHAE resind 0.2 (80% RH)Nitrile resin 0.65 1.6Amorphous nylon 1.2 4.0 (30◦C, 80% RH)PEN polyestere 1.2Nylon-6, biaxially oriented 2.6 5.8 (dry)PET, 25% crystalline (bottle wall) 4.8 24PVC, rigid 5.0 20Nylon-6, unoriented 6.6 10.2Aclar 33C PCTFE film f 7.0 (0% RH) 16 (0% RH)PP, biaxially oriented 150 548HDPE, molded 185 580MDPE, molded 250 1000PS film, oriented 365 900LDPE, molded 498 2500aRef. 27.bAt 75% RH and 23–25◦C except as noted.cVectra A950 LCP film (Celanese AG) (89).dBLOX is a trademark of The Dow Chemical Co. (Midland, Mich.) (90).eRef. 57.f Aclar is a trademark of Honeywell (Morristown, N.J.) (91).

ethylene is also recommended for thermoforming (102). At low to moderate RH,EVOH copolymers provide an excellent barrier to gases. The hydroxyl units (or OHgroups) contribute strongly to increasing chain cohesive energy density and im-proving barrier properties, and so the greater the fraction of OH groups, the lowerthe permeability (103). Figure 19 shows the effect of ethylene concentration onoxygen and water vapor permeability of EVOH (27). This figure also presents per-meation properties of poly(vinyl alcohol) (PVOH) [ie, fully hydrolyzed poly(vinylacetate)] and HDPE (which has no OH groups) for comparison with the EVOHseries (see VINYL ALCOHOL POLYMERS). As RH increases, the barrier properties ofEVOH copolymers decrease (cf Fig. 13). The OH groups are also responsible forthe hydrophilic nature of the polymer. Hence, the greater the percentage of vinylalcohol units, the greater the influence of humidity on gas barrier properties ofEVOH copolymers (27). Conversely, high proportions of ethylene units improveresistance to moisture (as shown in Fig. 13) and decrease the water vapor trans-mission rate, as shown in Figure 19. Moisture sensitivity can also be decreasedsomewhat by biaxial orientation (57). EVOH also offers very high barrier to fla-vor and aroma compounds and these barrier properties are not as sensitive to


Table 6. Water Vapor Transmission Rates of Various High and Moderate BarrierPolymersa

Barrier polymer Water vapor transmission rate,b (g·mil)/(100 in2·day)

P(VDC–AN) barrier coating 0.02Aclar 33C PCTFE filmc 0.025PVDC coextrusion resin 0.09Liquid crystalline polymerd 0.15 (100% RH)PP, biaxially oriented 0.25HDPE, molded 0.3MDPE, molded 0.7PVC, rigid 0.9LDPE, molded 1.0Nylon MXD-6 oriented film 1.2EVOH, 44 mol% ethylene 1.4PET, 25% crystalline (bottle wall) 1.8Nitrile resin 4.0EVOH, 27 mol% ethylene 5.7PS film, oriented 7.1Nylon-6, biaxially oriented 10Amorphous nylon 10Nylon-6, unoriented 15aRef. 27.bAt 90% RH, 37.8◦C.cAclar is a trademark of Honeywell (Morristown, N.J.) (91).dVectra A950 LCP film (Celenese AG) (89).

moisture as its oxygen barrier (57). In the majority of commercial applications,EVOH is used in a multilayer structure with moisture barrier and/or structurallayers on each side, a typical example being multilayer bottles with polypropylene(PP) for ketchup (57). EVOH has also been used as a flavor barrier on the inside ofPE-coated paperboard containers, a typical application being packaging of orangejuice, where it minimizes the loss of limonene from the juice into the PE layer(57). EVOH is used as a barrier layer in many other rigid and flexible packagingapplications.

Nitrile Polymers. The general structure of nitrile copolymers is shown below(101):

CH2 C

H

CN

CH2 C

H

Rx y

where R = or COOCH3

Application of PAN in packaging started in the 1960s. Despite processingdifficulties, PAN was used because of its excellent barrier properties. A groupof copolymers were developed in the 1970s. Comonomers are methyl acrylateand styrene at concentration up to 20 mol% (103). They are efficient oxygenbarriers and have high grease and oil resistance, strength, and stiffness. Lopac(Monsanto Co.), Barex (Sohio), and Cycopac (Borg-Warner Chemicals) are three


Table 7. Permeability, Diffusivity, and Solubility Coefficients of Flavor and AromaCompounds in Various High and Moderate Barrier Polymers (63) at 25◦C and 0% RH

Permeability, Diffusivity, Solubility,Flavor/aroma compound 10− 22 (kg·cm)/(cm2·s·Pa)a cm2/s kg/(cm3·Pa)

Ethylene vinyl alcohol copolymerb

Ethyl hexanoate 0.41 3.2 × 10− 14 1.3 × 10− 9

Ethyl 2-methylbutyrate 0.3 6.7 × 10− 14 4.7 × 10− 10

Hexanol 1.2 2.6 × 10− 13 4.6 × 10− 10

trans-2-Hexenal 110 6.4 × 10− 13 1.8 × 10− 8

d-Limonene 0.5 1.1 × 10− 13 4.5 × 10− 10

3-Octanone 0.2 1.0 × 10− 14 2.0 × 10− 9

Propyl butyrate 1.2 2.7 × 10− 13 4.5 × 10− 10

Vinylidene chloride copolymerb

Ethyl hexanoate 570 8.0 × 10− 14 7.1 × 10− 7

Ethyl 2-methylbutyrate 3.2 1.9 × 10− 13 1.7 × 10− 9

Hexanol 40 5.2 × 10− 13 7.7 × 10− 9

trans-2-Hexenal 240 1.8 × 10− 13 1.4 × 10− 7

d-Limonene 32 3.3 × 10− 13 9.7 × 10− 9

3-Octanone 52 1.3 × 10− 14 4.0 × 10− 7

Propyl butyrate 42 4.4 × 10− 14 9.4 × 10− 8

Dipropyl disulfide 270 2.6 × 10− 14 1.0 × 10− 6

Low density polyethyleneEthyl hexanoate 4.1 × 106 5.2 × 10− 9 7.8 × 10− 8

Ethyl 2-methylbutyrate 4.9 × 105 2.4 × 10− 9 2.3 × 10− 8

Hexanol 9.7 × 105 4.6 × 10− 9 2.3 × 10− 8

trans-2-hexenal 8.1 × 105

d-Limonene 4.3 × 106

3-Octanone 6.8 × 106 5.6 × 10− 9 1.2 × 10− 7

Propyl butyrate 1.5 × 106 5.0 × 10− 9 3.0 × 10− 8

Dipropyl disulfide 6.8 × 106 7.3 × 10− 10 9.3 × 10− 7

High density polyethylened-Limonene 3.5 × 106 1.7 × 10− 9 2.5 × 10− 7

Menthone 5.2 × 106 9.1 × 10− 9 4.7 × 10− 7

Methyl salicylate 1.1 × 107 8.7 × 10− 10 1.6 × 10− 6

Polypropylene2-Butanone 8.5 × 103 2.1 × 10− 11 4.0 × 10− 8

Ethyl butyrate 9.5 × 103 1.8 × 10− 11 5.3 × 10− 8

Ethyl hexanoate 8.7 × 104 3.1 × 10− 11 2.8 × 10− 7

d-Limonene 1.6 × 104 7.4 × 10− 12 2.1 × 10− 7

Poly(ethylene terephthalate)c

d-limonene 1.5 6.0 × 10− 13

aTo convert the permeability values to (cm3·mil)/(100 in.2·day·atm), multiply the values provided inthe table by (4.98 × 1022)/MW, where MW is penetrant molecular weight (g/mol).bValues have been extrapolated from higher temperatures.cAt 25◦C; values are not expected to show any significant variations with RH (92).

commercial nitrile copolymers that have been used in packaging applications.Their compositions are shown in Table 8. Barex and Cycopac are rubber-modifiedfor improved mechanical properties, and the barrier properties of these copolymersare relatively insensitive to moisture. Concerns about the possible migration of


10−1

100

101

10−3

10−2

10−1

100

101

102

103

0 20 40 60 80 100

mol% EthylenePVOH HDPE

EVOH

WV

TR

, (g

� m

il)/(

100

in.2

� da

y)

Oxygen perm

eability, (cm3

� mil)/(100 in. 2

� day� atm

)

Fig. 19. Effect of ethylene content on oxygen permeability (20–23◦C, 65% RH) and WVTR(40◦C) of EVOH. The figure also shows data for PVOH and HDPE for comparison (27).

acrylonitrile monomer, a toxic compound (57), have limited the use of nitrilepolymers in food contact applications (see ACRYLONITRILE AND ACRYLONITRILE

POLYMERS).Vinylidene Chloride Copolymers. Copolymers of VDC with vinyl chloride

and acrylonitrile were among the first high barrier polymers to be widely used.During their commercial appearance in the late 1930s, they had the lowest perme-abilities among plastics to gases and water vapor (103). Poly(vinylidene chloride)(PVDC) homopolymer is soluble only in hot dichlorobenzene (among common sol-vents) and has a melting point only a few degrees below its decomposition temper-ature (103). These characteristics make it difficult to fabricate by melt-processing

Table 8. Compositions of Commercial, High Barrier Nitrile Copolymersa

Polymer Manufacturerb Chemical compositionc

Lopac Monsanto Co. 70% acrylonitrile + 30% styreneBarex Sohio 74% acrylonitrile + 26% methyl methacrylate

+ 10% butadiene graft rubberCycopac Borg-Warner 74% acrylonitrile + 26% styrene

Chemicals + 10% butadiene graft rubberaRef. 63.bManufacture are listed from Ref. 63. However, these copolymers are now manufactured by BPAmoco Chemicals.cData from FDA regulations for corresponding materials.


techniques. Copolymers were synthesized to overcome these drawbacks. Acrylateswere found to be among the most useful comonomers, along with vinyl chlorideand acrylonitrile (103). By adding comonomers, the melting point can be decreasedto a range of 140–175◦C (as compared to 198–205◦C for PVDC), thus making meltprocessing feasible (104). These copolymers are semicrystalline and soluble inonly a limited range of solvents. The most notable attributes of VDC copolymersare their chemical resistance and extremely low permeabilities to gases and wa-ter vapor (cf Tables 5 and 6). The structure of the most widely used vinylidenechloride–vinyl chloride copolymer is shown below:

where x is 85–90 mol% (103). VDC copolymers are commercially available undera variety of trade names such as Saran (The Dow Chemical Co.), Daran (W. R.Grace), Amsco Res (Union Oil), and Serfene (Rohm and Haas) within the UnitedStates; Haloflex (Neoresins), Diofan (Solvin), Ixan (Solvin), and Polyidene (Scott-Bader) in Europe (see VINYLIDENE CHLORIDE POLYMERS). Copolymers are availablein the following forms (103): (1) lattices of approx. 1000–1500 A which can beapplied as coatings to paper and plastic films to improve their barrier proper-ties, (2) soluble resins for coating plastic films (especially cellophane) to improvetheir barrier properties, (3) melt-processable resins for extrusion, coextrusion, andmolding, and (4) clear, transparent films for commercial packaging applications.Small amounts of processing aids and heat stabilizers are added to extrusion andmolding resins.

Polyamides. The standard semicrystalline polyamides (PA) (nylon-6,nylon-6,6, etc) used in packaging have medium gas barrier properties, and theirbarrier properties are affected by humidity (101). However, specialty grades ofpolyamides with higher gas barrier properties are available. Commercial gradesof Selar amorphous polyamides (AmPA) (E. I. du Pont de Nemours & Co., Wilm-ington, Del.), for example, exhibit good O2 barrier and reduced dependence of gasbarrier properties on RH. In fact, their gas barrier properties improve with in-creasing RH (cf Fig. 13). At an RH of 80% or more, their O2 barrier is similar tothat of PAN (102). Moreover, at 95–100% RH, the O2 barrier is equivalent to thatof EVOH at similar conditions, and substantially better than that of nylon-6. Theamorphous nature of Selar results in a much broader range of processing con-ditions than those of semicrystalline nylon-6 (105). The mechanical and barrierproperties of AmPA can be improved by orientation (see POLYAMIDES, PLASTICS;POLYAMIDES, AROMATIC).

Another high barrier polyamide is MXD-6 resin (Mitsubishi Gas and Chem-ical Co., and Toyobo, Japan), which was developed in the 1970s. It is made fromthe reaction of m-xylylenediamine and adipic acid (57):


It provides improved clarity, mechanical, thermal, and barrier properties rel-ative to standard nylons. It has better gas barrier properties than nylon-6 and PETat all humidities, and is better than EVOH at 100% RH. The barrier properties ofMXD-6 are relatively unaffected by moisture up to an RH of 70% (57). Because ofits cost and the lack of a domestic source, MXD-6 has found limited applications inthe United States. It has a much wider market in Japan, appearing in commercialapplications such as nonpasteurized plastic beer bottles and carbonated soft drinkbottles (57).

Polyesters. The most widely used member of the polyester family for foodand beverage packaging applications is PET (see POLYESTERS, THERMOPLASTIC).However, PET offers a moderate barrier to gases and water vapor. Poly(ethylenenaphthalate) (PEN) offers a much higher barrier to gases and water vapor thanPET and can be classified as a high barrier polyester.

PEN is a homopolymer of dimethyl-2,6-naphthalene dicarboxylate (NDC)and ethylene glycol (106):

The rigid double-ring structure in the polymer backbone results in increasedmechanical strength, heat stability, and barrier properties as compared to PET.As with PET, orientation produces a substantial reduction in gas permeability,and the oxygen permeability of oriented PEN is a factor of 5 lower than that oforiented PET (57). Teijin (Japan) and ICI (U.K.) have been the leaders in man-ufacturing PEN films. Teijin manufactures both resin and films, the latter un-der the trade name Teonex. Since 1990, DuPont and Teijin have had a world-wide joint venture in polyester films, including PEN. DuPont has also acquiredICI’s Melinex PET film and Kaladex PEN film operations. The main disadvan-tages of PEN are its high cost, and unsettled sources of monomer technologyand supply (57). One of the leading manufacturers of NDC, BP Amoco Chemi-cals (Decatur, Ala.) has developed a technology that uses o-xylene rather thannaphthalene as the feedstock, which could reduce the manufacturing cost of PENresin (57). Other newer NDC technologies from Kobe Steel (Japan) and MobilChemicals could also lead to reduced cost for PEN. Another way of addressingthe price issue is by using blends or copolymers of PEN with PET. Gas and va-por permeability has been found to decrease continually as PEN is added to PET(107,108).

Liquid Crystalline Polymers. Liquid crystalline polymers, main-chain (qv)(LCPs) offer excellent thermal and chemical resistance, and exhibit very highbarrier properties that are almost unmatched by existing barrier polymers(16,109,110). They also offer adequate mechanical properties for certain packag-ing applications (111). LCPs are very efficiently packed, highly oriented, and oftensemicrystalline materials. Commercially available LCPs such as Vectra (CelaneseAG), Zenite (E. I. du Pont de Nemours & Co.), and Xydar (Solvay Advanced


Polymers, Alpharetta, Ga.) are aromatic copolyesters that have a significantlyhigher degree of chain orientation than typical polyesters such as PET (111).

LCPs of this kind were first introduced to the market in the 1980s.Table 9 shows a comparison of properties of biaxially oriented LCP films withthose of PET films (112). Figure 20 compares barrier properties of LCPs withother barrier polymers. LCPs offer the best combination of water vapor and O2barrier properties among all known classes of polymers. They also offer excellentbarrier to CO2, N2, and other gases and vapors (111). However, commercial ap-plications of LCPs have been limited primarily because of their high cost, lackof transparency, and processing characteristics (111). One key to unlocking thepotential of LCPs in barrier packaging is to be able to process thin uniform layersin coextruded multilayer structures. Recently, three-layered structures (PET–tielayer–LCP) have been coextruded (Superex Polymer, Inc., Waltham, Mass.) witha total thickness of 25–50 µm and 10–30% LCP layer thickness. This multilayerfilm is claimed to offer a high performance–cost ratio (111).

Poly(hydroxy amino ethers). Poly(hydroxy amino ethers) (PHAE) are anew family of high barrier epoxy-based thermoplastics. The general chemi-cal structure of amide-containing PHAE, which are formed by the reaction of

Table 9. Comparison of Properties of Biaxially Oriented LCP and PET Filmsa

Biaxially oriented Biaxially orientedProperty LCP filmb PET filmc

Tensile strength, MPad 240 170Tensile modulus, GPae 12.4 3.5Oxygen permeability f , 0.05 4.8

(cm3·mil)/(100·in.2·day·atm)Water vapor permeability,g 0.02 1.7

(g·mil)/(100·in.2·day·atm)Density, g/cm3 1.4 1.4Upper use temperature, ◦C Over 250 120Tear resistance, kN/mh

Initiation 595 35Propagation 175–525 9–53aRef. 112.bVectra (Celanese AG) isotropic LCP film (orientation angle = 45◦).cMylar (DuPont) isotropic PET film (orientation angle = 45◦).dTo convert MPa to psi, multiply by 145.eTo convert GPa to psi, multiply by 145,000.f Permeability value at 25◦C (and unspecified RH).gPermeability value at 25◦C and 90% RH.hTo convert kN/m to lb/in., divide by 0.175.


10−2

10−1

100

101

102

ThermotropicLCP

EVOH

Nylon-6

MXD-6 PAN

PET

PEN

PVDC PCTFE

Biax PP

HDPE

10210−3 10−2 10−1 100 101 103

Wat

er v

apor

per

mea

bilit

y, (

g �

mil)

/(10

0 in

.2�

day

� at

m)

Oxygen permeability, (cm3 � mil)/(100 in.2 � day � atm)

Fig. 20. A comparison of oxygen and water vapor barrier properties of various high andmoderate barrier polymers at 23◦C (111).

amide-containing bisphenols with aromatic diglycidyl ethers (113), is shownbelow:

where R can be (CH2)n (n = 1, 2, 3. . .) or an aromatic group, and Ar designatesan aromatic moiety.

These polymers are amorphous, with Tg values ranging from 90 to 133◦C.They incorporate both amide and hydroxyl moieties on the chain backbone. Inorder to prevent cross-linking, the hydroxyl groups are generated during the poly-merization of preferred amide-containing monomers (113). The presence of aro-matic groups between the amide groups in the polymer backbone can result inhigh Tg and good barrier properties. It has also been observed that the presenceof m-phenylene units instead of p-phenylene units can reduce O2 permeabilities(as much as 30–40%) by increasing the chain packing efficiency of the polymer(113). This phenomenon is, in fact, rather general among aromatic polymers. Of-ten, meta-linked aromatic rings in polymer backbones have lower permeabilitycoefficients than their para-linked analogues (114,115). Lower O2 permeabili-ties can also be obtained when hydrogen bonding interactions in the polymerbackbone are increased either by reducing the number of nonpolar methyleneunits or by increasing the population density of polar amide groups (113). Un-like other hydroxyl-containing polymers such as EVOH, the barrier properties


of amide-containing PHAE improve with increasing RH (19,113). For example,a decrease in O2 permeability from 1.4 (cm3·mil)/(100 in.2·day·atm) at 5% RH to0.8 (cm3·mil)/(100 in.2·day·atm) at 75–80% RH has been reported. On the basisof preliminary work involving density and pals studies, it has been postulatedthat water molecules not only occupy free-volume elements in these materials butalso enhance interchain cohesion, thus inhibiting the transport of other nonpolargases (116).

A series of polymers from the PHAE family has recently been commercial-ized by Dow Chemical Co. (Midland, Mich.) under the trade name BLOX Adhesiveand Barrier Resins (90,117). They offer high gas barrier properties, excellent ad-hesion to a variety of substrates, high optical clarity, and good mechanical proper-ties. For example, BLOX 4000 series resins exhibit an oxygen transmission rateof 0.1 (cm3·mil)/(100 in.2·day·atm) at 23◦C and 60% RH (90). The BLOX resinshave found some commercial applications in barrier packaging, starch-based foampackaging, and powder coatings (117).

Polychlorotrifluoroethylene. Polychlorotrifluoroethylene (PCTFE) is a flex-ible thermoplastic made from fluorinated–chlorinated resins. It was first discov-ered in the 1950s and has been commercially produced since the 1960s (118). Itschemical structure is as shown below:

The key characteristics of this polymer are its high optical clarity and ex-cellent moisture barrier properties (118). In commercial Aclar resins (Honeywell,Morristown, N.J.), the polymer is generally modified by copolymerization, result-ing in a semicrystalline material with Tg of about 45◦C and melting point of about190◦C (104). It shows excellent thermal and chemical stability, and high watervapor and O2 barrier properties (120) (cf Tables 5 and 6). For example, Aclar filmstypically have water vapor transmission rates less than 0.04 (g·mil)/(100 in2·day)(23◦C) and oxygen transmission rates less than 14 (cm3·mil)/(100 in2·day) (25◦C)(119). PCTFE is most widely used for producing blister packs in pharmaceuticalapplications (118). The moisture barrier properties can be tailored by varying thePCTFE layer thickness in coextruded multilayer structures, thereby adjustingthe performance-to-cost ratio. An example of a commercial coextruded structureis Aclar NT AE-1 (Honeywell, Morristown, N.J.), which is 33 µm thick, and con-tains about 8.4 µm of PCTFE (118). PCTFE can also be laminated with PP, PAN,PET, HDPE, LDPE, and PVC and then formed into blister packs. Other appli-cations of the same technology include electronic component packaging wherelower moisture barrier may be acceptable, resulting in more favorable economics(118).

Moderate Barrier Polymers. Polymers in this category includepolyesters, polyolefins, poly(vinyl chloride), polystyrene, and certain semicrys-talline polyamides. The polymers included in this section are typically more widelyavailable, have more manufacturers, and are less expensive than the specialtybarrier resins described in the previous section.

Polyesters. Polyesters represent a class of versatile barrier plastics. PETis by far the most important member in this family from a commercial


viewpoint. It is extensively used in the food and beverage packaging industriesand is especially known for its widespread use in bottles for carbonated beverages(120) (see POLYESTERS, THERMOPLASTIC). The other members in this category arepoly(trimethylene terephthalate) (qv) (PTT) and Polylactide (PLA).

Poly(ethylene terephthalate). PET is a linear thermoplastic made from ethy-lene glycol and terephthalic acid, or ethylene glycol and dimethyl terephthalate(101). The structure of PET is shown below:

PET is used in many rigid food and beverage containers because of a goodbalance of physical and mechanical properties, barrier properties, processibilityand formability, ecological and toxicological characteristics, and economics (121).As a result, PET bottles have virtually replaced glass packages for carbonatedsoft drinks in the United States. In the glassy state, it is strong, stiff, ductile,and tough. It can be oriented by stretching during molding and extrusion, whichfurther increases its strength, stiffness, and barrier properties. One of the earlydrawbacks of PET was its low heat distortion temperature of 60◦C (140◦F), whichprevented it from being used in applications requiring filling at elevated tem-peratures (121). However, careful heat-treatment increases the heat distortiontemperature of crystalline PET containers and recently commercialized processesclaim resistance to temperatures of up to 90–92◦C (194–196◦F) (121). It has mod-erate barrier properties for light gases, but is a good barrier for flavors and aromacompounds. Its barrier properties can be improved by increasing crystallinity andorientation (121). A 4× biaxial orientation of amorphous PET at temperaturesnear 100◦C produces significant strain-induced crystallization and decreases thepermeability of O2 and CO2 by a factor of about 2 (71).

Oriented and heat-set PET films have also found use in a broad range offlexible packaging applications because of their high strength, good barrier, highclarity, heat resistance, and good metalizability. There are more than 50 specificapplication areas for PET films (122). Manufacturers have met the product re-quirements of each of the diverse end-use markets by tailoring formulations andprocess conditions. For example, in the food and beverage packaging industries,many types of PET films have been developed, including metallized PET filmsfor packaging of coffee, wine, and meats, poly(vinylidene chloride)-coated PETfilms for meat and cheese packaging, and coextruded multilayer PET films forheat-sealable packaging (122). Commercial manufacturers of PET films includeDuPont Teijin Films (Mylar, Melinex, Tetoron) and Mitsubishi Polyester Film,LLC. (Diafoil, Hostaphan).

Poly(trimethylene terephthalate). PTT is made from the polycondensationreaction of trimethylene glycol (also called 3G) with either terephthalic acid ordimethyl terephthalate (57). Although this polymer was first synthesized in 1941,it was never commercialized because of lack of an economical source of 3G. In the


early 1990s, Shell Chemical Co. announced a catalyst breakthrough to make 3Geconomically by hydroformylation of ethylene oxide (57).

PTT is an engineering resin and has been targeted mainly toward injection-molding applications. However, there have been claims about its barrier properties(57), and it is therefore included in this article for completeness.

Because of the presence of an odd number of methylene units in the chainbackbone, PTT has physical properties different from PET (57). Its O2 permeabilityis about 6 (cm3·mil)/(100 in.2·day·atm) at 0% RH, and moisture absorption rateis typically 0.03% after 24 h and 0.15% after 14 days (57). Typical properties ofPTT are as follows: it has a melting point of 228◦C, a Tg of 45–65◦C, and can beproduced with crystallinity values up to 45 wt%.

Polylactide. Polylactide (PLA) is a semicrystalline, linear thermoplasticmade from lactic acid, especially as derived from corn (maize). Polylactide (qv)is new to the commercial market (123). Its applications are still being explored.However, its barrier is adequate for some food packaging uses, especially for man-agement of flavor and aroma profiles. It can be heat sealed and thermoformed.It has a glass-transition temperature of 55–65◦C. The principal manufacturer ofPLA is Cargil Dow, LLC (NatureWorks).

Polyamides. This category of moderate barrier polymers includes nylon-6and nylon-6,6. Nylon-6 is made by the polymerization of caprolactam, and nylon-6,6 is made by the reaction of hexamethylenediamine and adipic acid (101). Theirchemical structures are shown below:

In general, nylons have good gas and aroma barrier properties, but poor mois-ture barrier properties (57) (cf Tables 5 and 6). Absorbed water has a plasticizingeffect that leads to a reduction in tensile strength and increase in impact strength.Uniaxial and biaxial orientations can improve their flex-crack resistance, mechan-ical, and barrier properties (101). Biaxially oriented nylons offer better gas barrierproperties, softness, and puncture resistance, compared to oriented PET, whichoffers better rigidity and moisture barrier properties (101). Nylons are less widelyused in the packaging industry than polyolefins or PET, with the majority of theapplications being blow-molded bottles (coextrusion and blending of nylon-6 withPE), for industrial and household chemical markets (57).

Polyolefins. PE and PP are two of the most widely used polymers in the foodand beverage packaging industry. These polymers find use as films, moldings,coatings, adhesives, and closures (124). They are available in a wide variety of


Density HighLow

Property

Low

High

StiffnessTensile strengthChemical resistance

Gas and vapor transmissionElongation at breakLTIESCR

Fig. 21. Typical effect of density on various properties of PE. LTI is low temperatureimpact strength and ESCR is environmental stress-crack resistance (124).

types and grades. Although they have much higher permeabilities to gases thanmany other barrier polymers, they are very good moisture barriers (cf Tables 5and 6).

PE, whose structure is shown below, was one of the first olefinic polymers tobe used commercially in the packaging industry:

It is classified on the basis of density (see ETHYLENE POLYMERS, HDPE;ETHYLENE POLYMERS, LDPE; ETHYLENE POLYMERS, LLDPE; ETHYLENE POLYMERS,VLDPE). Figure 21 shows the effect of density on various properties of PE (124).A branched structure for low density PE (LDPE) results from exceptionally hightemperature and pressure during its manufacture. It is tough, flexible, can beeasily melt processed, and has good moisture barrier properties. It is a semicrys-talline polymer with crystallinity typically in the range of 40%. Medium densityPE (MDPE) is stronger, stiffer, and has better barrier properties than LDPE. Highdensity PE (HDPE) is essentially unbranched and is the strongest and most rigidpolymer in this family. It offers barrier properties to moisture and gases that aresuperior to those of LDPE and MDPE.

If unsaturated comonomers such as butene, hexene, or octene are added tothe HDPE polymerization process in the presence of a stereospecific catalyst, it re-sults in the formation of a linear polymer with short branch-like pendant groups(104). Its density is in the same range as LDPE, but the degree of branchingis greatly reduced. This polymer is called linear LDPE (LLDPE) and its den-sity depends on the amount of comonomer added. The larger the amount of thecomonomer, the lower the density of the copolymer (104). LLDPE combines the


clarity and excellent heat-seal properties of LDPE with the strength and tough-ness of HDPE. It is often blended with LDPE in order to optimize the benefitobtained from both materials (104).

PP, whose chemical structure is shown below, can be made by the catalyticpolymerization of propylene at high temperature and pressure (124):

Isotactic PP (with all methyl groups on same side of the polymerchain) is the commercially desired form for packaging applications (124) (seePROPYLENE POLYMERS (PP)). The concentration of atactic PP (with irregular ar-rangement of methyl groups) is kept low by suitable catalysts and polymerizationconditions (124). PP offers high resistance to water vapor permeation and is widelyused in rigid as well as flexible food packaging applications. PP films can be ori-ented, which improves their barrier properties, mechanical strength, and opticalproperties. For example, oriented PP has about three times higher resistance towater vapor transmission than unoriented PP (27). These properties can be variedover a wide range by the choice of the manufacturing process.

Poly(vinyl chloride). Poly(vinyl chloride), also known as vinyl or PVC, ismade by low pressure free-radical polymerization of vinyl chloride at temperaturesin the 38–71◦C range (101). Its chemical structure is shown below:

PVC is a versatile polymer that can be formulated to meet the requirementsof many applications in packaging and other industries (88) (see VINYL CHLORIDE

POLYMERS). Its properties can be tuned over a very wide range by incorporatingcomonomers, plasticizers, and other additives. When used as a rigid sheet or bot-tle, little modification is required, but PVC requires the addition of plasticizersto make it useful as a barrier film for flexible packaging (88). The plasticizersincrease chain flexibility and reduce the processing temperature of PVC. For ex-ample, addition of 40 vol% dioctyl phthalate plasticizer reduces the Tg of PVCfrom 100◦C to about 5◦C (125). The increase in chain flexibility of plasticized PVCalso results in a reduction in its gas barrier properties primarily as a result ofhigher diffusion coefficients in plasticized films. Figure 22 presents a comparisonof diffusion coefficients of several penetrants in plasticized and unplasticized PVCfilms (126) (see PLASTICIZERS).

PVC has some drawbacks as a food-packaging material (88): Vinyl chloridemonomer is an animal carcinogen and causes liver cancer in humans. Thus theamount of monomer in the finished polymer should be brought down to 1 ppm orless. There is concern about the toxicity of plasticizers and other additives usedin PVC. Moreover, there have also been concerns that incineration of chlorine-containing plastics can possibly lead to the formation of dioxin, a chlorinatedtoxic molecule. Finally, the release of plasticizers over long periods of time canlead to gradual embrittlement of PVC films. However, none of these drawbackssignificantly affect the functionality of PVC and it is used to make flexible films,


10−18

10−16

10−14

10−12

10−10

10−8

10−6

1.21.00.2 0.4 0.6 0.8

Penetrant size, nm

Diff

usio

n co

effic

ient

s, c

m2 /

s

Unplasticized

Plasticized

Fig. 22. Diffusion coefficients in unplasticized and plasticized PVC at 30◦C as a functionof penetrant molecular diameter (126).

rigid sheets, and bottles for a variety of food (eg, fresh fruits, vegetables, andpoultry packaging in the United States) and nonfood packaging applications (104).

Polystyrene. Polystyrene (PS) is made by the peroxide-catalyzed bulk orsuspension polymerization of styrene (101) (see STYRENE POLYMERS). The poly-merization reaction takes place at low pressure and temperature in the range of250–400◦F, and the polymer chemical structure is given by

Polymer molecular weight, which affects the processing characteristics ofthe resin, is often in the range of 40,000–220,000; variations in molecular weightcan be obtained by changing the catalysts and polymerization conditions. PS isan amorphous, clear, hard, brittle, low-strength material with poor impact re-sistance (101). It has low to moderate moisture and gas barrier properties (88).Copolymerization with butadiene or other rubbers increases its impact strengthand decreases stiffness (88). This copolymer is commonly referred to as high im-pact polystyrene (HIPS). Comonomers such as α-methyl styrene can improve theheat resistance of PS by increasing the heat distortion temperature to 100◦C or


higher (88). PS can be foamed by adding foaming agents such as pentane to thereaction mixture during the suspension polymerization. This so-called expandedpolystyrene (EPS) is a very low density, yet highly rigid material that is used formaking egg cartons, and trays for meat, poultry, and other products (88). However,EPS has poor gas barrier properties.

Improving Barrier Properties of Polymers

Barrier Structures. As mentioned in the previous sections, the principalapplication of barrier polymers is in the food and beverage packaging industries.Combining two or more polymers, or other materials, can achieve performanceadvantages not available in any of the materials taken alone. In many cases, toachieve better barrier performance, it is more efficient and economical to use athin layer of an expensive high barrier polymer (eg, EVOH) sandwiched betweenlayers of less expensive, moderate barrier, structural polymers (eg, PP, PET) thanto increase the monolayer thickness of the moderate barrier polymer. Multilayerstructures can be obtained by coextrusion, lamination, and coating. Barrier poly-mers may also be combined to form miscible and immiscible blends.

Multilayer Structures. Steady-state barrier properties of multilayer filmscan be described by their permeability, which can be calculated by treating theindividual layers as resistances in series (63):

1P′ = Lt

Pt= L1

P1+ L2

P2+ · · · + Ln

Pn(38)

where Pt, Lt, and P′ are permeability, thickness, and permeance of the compos-ite structure, and P1, P2, . . . , Pn and L1, L2, . . . , Ln are the permeabilities andthicknesses of the individual layers. The permeability Pt (or permeance P′) canbe used to evaluate the performance of the composite structure for packagingapplications. Figure 23 illustrates the effect of barrier layer thickness on per-meability of a hypothetical two-layer sheet, the permeability of the barrier layerbeing 0.1 (cm3·mil)/(100 in.2·day·atm) and that of the nonbarrier layer being 100(cm3·mil)/(100 in.2·day·atm). For moisture-sensitive polymers (eg, EVOH or ny-lon), it is important to use the permeability that corresponds to the effective RHthat the layer will experience in the composite structure.

Coextrusion (qv) is one of the most cost-effective techniques for producingmultilayer barrier polymer films (127). Although coextruded films typically have3–7 layers, as many as 11 layers can be extruded simultaneously (128). Thistechnique allows for the thinnest possible layers of the individual polymer resinswithin the structure. Resins that do not bond well can be bonded together withan adhesive ‘tie’ layer (128). The materials must be compatible in terms of theirmelt temperatures and viscosities in order to undergo simultaneous coextrusion.Figure 24 presents a comparison of processing temperatures of various commonlyused barrier polymers (129). Figure 25 shows a schematic of a typical 9-layer co-extruded structure. A coextrusion line will have several different extruders, eachextruder responsible for supplying individual polymer resins (127).


10−1

100

101

102

0 0.2 0.4 0.6 0.8 1

Barrier layer thickness fraction

Per

mea

bilit

y, (

cm3

� m

il)/(

100

in.2

� da

y �

atm

)

Fig. 23. Effect of thickness of barrier layer on permeability of a hypothetical two-layerbarrier composite structure.

100

150

200

250

300

350

MXD-6Nylon-6

Nylon-6,6 PET PC

PE

PPEVOH

Pro

cess

ing

tem

pera

ture

, °C

Fig. 24. A comparison of processing temperatures of various commercially used barrierpolymers (129).

Coating and lamination are two additional processes for producing multi-layer structures and are especially useful in applications where a nonpolymericmaterial is part of the structure. Coatings can be melt extrusions of a polymer ontoa base film or can be made by applying solutions or dispersions of polymers to the


Skin layer (15−30%)

Adhesive layers (2−3% ea.)

Barrier layers(5−15%)

Adhesive layers

Recycle layer (30−50%)

Skin layer (15−30%)

Fig. 25. An example of a typical symmetrical nine-layer coextruded barrier compositestructure (128).

base film (128). Barrier polymers that resist water vapor and provide gas barrierare often laminated to paper and paperboard. Cellophane, a flexible transparentgas barrier polymer, can also be made ‘moisture-proof ’ by coating or laminatingwith other polymer films (27). A commercial application of solvent coating withbarrier polymers is VDC resin dissolved in a polar solvent and coated onto cel-lophane or PET (63). Water-based emulsion coatings of VDC are used for plasticPET beer bottles, primarily in the United Kingdom. They can typically lower O2permeability of PET bottles by nearly 60% (57). Several companies produce PVDCemulsions for coating films such as nylon, PE, PP, and PET to improve their O2and moisture barrier properties (57). In the mid-1990s, PPG Industries, Inc. in-troduced Bairocade external epoxy-amine organic coatings for improving barrierproperties of PET bottles (130,131). It is claimed that these coatings can reduceO2 permeability of PET bottles by a factor of 6 or more with the level of barrierimprovement depending on the thickness and formulation used, and that coatedbottles extend the shelf life of carbonated beverages by a factor of 3, and beer bya factor of 20 relative to uncoated bottles (131).

Another way of improving barrier properties of polymers is by coating themwith thin inorganic layers (57,132,133). This can typically improve barrier prop-erties by a factor of 100, whereas the thickness of the barrier coating applied isless than 0.5% of the base film (57). Laminations and coatings of aluminum foil oraluminum oxide on barrier polymers can provide significant improvements in gasbarrier properties (57). These metallized films contain an extremely thin layerof aluminum, which not only enhances the barrier properties of the base film,but also provides a shiny metallic appearance. Because the layer is very thin, itdoes not appreciably affect the strength and flexibility of the base film (134). Thebarrier properties of metallized films can approach those of pure aluminum foil.However, unlike aluminum foil, metallized films are not subject to flex-crackingand hence, are better at maintaining their barrier properties (134). Barrier poly-mers that have been successfully metallized include PP, nylons, PET, and unplas-ticized PVC. As an example, 99% decrease in oxygen permeability and 98.5% de-crease in water vapor permeability have been reported for metallized PET films ascompared to unmetallized PET films (135). Metallized films are widely used in the


flexible packaging industry, an example being potato chip packages consisting ofmultilayer structures of oriented PP, PE, and metallized PP (136).

However, one of the main disadvantages of metallized structures is that theymay not be transparent—a desirable feature for many packaging applications. Inpart to address this shortcoming, several alternative high barrier coating tech-nologies, typically producing amorphous carbon or glass-like layers, are beinginvestigated. These newer coatings have achieved limited market penetration todate, with most of the commercial activity in Europe and Japan. It is still un-clear what their eventual importance will be, but these technologies are beingactively pursued. Among the various glass and ceramic materials used as barriercoatings, silicon dioxide (silica) has been the most widely used (57,133). Thesebarrier composite structures are clear, microwaveable, and recyclable. Electron-beam treatment and plasma-enhanced chemical vapor deposition (PECVD) aretwo of the most widely used techniques for depositing thin silica layers on barrierpolymers (57,137). In the electron-beam treatment, a high energy electron-beamsource is used to vaporize silica, which then precipitates onto the polymer film,forming a continuous coating as the film passes through a vacuum chamber. Theresulting coatings are uniform and can be as thin as 0.04 µm. PECVD has theadvantage of being able to control coating density and thickness, by changingthe process variables (137). Figure 26 presents the effect of coating thickness onoxygen permeability of a PECVD silica-coated PET film (137). Barrier propertiesof silica-coated polymers can be superior to those of high barrier polymers such

10−7

10−6

10−5

10−4

10−3

10−2

10 100

Coating thickness, nm

Oxy

gen

tran

smis

sion

rat

e, c

m3 /

(100

in.2 �

day

� at

m)

Fig. 26. Effect of plasma-deposited silica coating thickness on steady-state oxygen trans-mission rates (33◦C, 0% RH) of PET films that have an uncoated thickness of 13 µm (137).


10−3

10−2

10−1

100

101

102

103

102

1.3 1.0

<0.003

OPP

SiOx ORMOCER

OPPOPPSiOx

OPP

ORMOCER

Oxy

gen

tran

smis

sion

rat

e, c

m3 /

(100

in.2 �

day

� at

m)

Fig. 27. Influence of ORMOCER coatings on oxygen transmission rates (OTRs) of SiOx-coated oriented polypropylene (OPP) (OTR measured at 23◦C and 70% RH) (138).

as PVDC and EVOH. Moreover, they are not influenced by moisture and temper-ature (57). However, because of poor adhesion and mechanical properties, for allpractical applications silica-coated films have to be laminated.

Recently, inorganic–organic hybrid polymers have been developed, which canbe used as laminating agents and in conjunction with silica to enhance barrierproperties of polymers (138,139). An example is ORMOCER or organically modi-fied ceramic (Fraunhofer Gesellschaft, Germany). Such materials can be used ascoatings as well as high barrier laminating agents in multilayer structures. Theuse of ORMOCER as a top layer on silica-coated PP can significantly improve itsoxygen barrier properties, as shown in Figure 27. The composite structure alsooffers good water vapor barrier properties. Similar improvements in barrier prop-erties of PET have been reported (138). Figure 28 shows a comparison of O2 andwater vapor barrier properties of different barrier polymer composite structures(140).

Another potential candidate as an effective barrier coating is diamond-likecarbon (DLC) (132). DLC refers to a group of amorphous, hard, and chemicallyinert materials consisting of carbon bonded partially as diamond (sp3) and par-tially as graphite (sp2) and containing 0–40% hydrogen atoms. These coatings aretransparent, flexible, extremely impermeable, biocompatible, and adhere well toa wide range of polymers.

Among the most recent commercial applications of barrier polymeric con-tainers coated with thin inorganic layers is plastic beer packaging. Glaskin (Tetra


10−2

10−1

100

101

102

10−3

BOPP (met)/PE

EVOH composites

PA/PE composites

PET/Met./PEPET (met)/PE

Multilayer bottles

PET–PEN (met)/PE

ORMOCERswith inorganic layersAl composites

10−2 10−1 100 10110−310−4

Oxy

gen

tran

smis

sion

rat

e, c

m3 /

(100

in.2 �

day

� at

m)

Water vapor transmission rate, g/(100 in.2 � day)

Fig. 28. A comparison of oxygen and water vapor transmission rates of various bar-rier polymer composite structures (140). BOPP (met) is biaxially oriented, metallizedpolypropylene; PET (met.) is metallized poly(ethylene terephthalate); PEN (met) is metal-lized poly(ethylene naphthalate); Al composites are aluminum composites.

Pak, Geneva, Switzerland) is a proprietary technology that utilizes a vacuum de-position process for coating clear, extremely thin layers of silicon oxide on theinside of blown plastic bottles (141). It offers excellent barrier properties for O2,CO2, and flavor compounds, and it is completely recyclable. For beer packaging, ashelf life of 6 months has been claimed using this technology. BESTPET (BarrierEnhanced Silica Coated PET) is another silica-based coating technology developedby the Coca Cola Co. (Atlanta, Ga.) in conjunction with Applied Films (Longmont,Colo.) and Krones (Nevtraubling, Germany). The exterior surface of PET bottles iscoated with a thin silica layer and improvements in barrier properties of PET by afactor of at least 2 have been claimed, which results in a shelf life of over 6 monthsfor packaged beer (142). A carbon-based high-barrier bottle coating technologyis ACTIS (Amorphous Carbon Treatment on Internal Surface technology) (Sidel,France) (143). In this process, the internal surface of plastic bottles is coated witha 0.15 µm thick layer of highly hydrogenated amorphous carbon obtained from afood-safe gas (eg, acetylene) in its plasma state. This technology has been claimedto improve O2 barrier properties of PET bottles by a factor of 30, and CO2 barrierproperties by a factor of 7. Another carbon-based DLC coating technology (KirinBrewery, Japan) claims to improve the barrier properties of PET for O2, CO2, andH2O by factors of 20, 7, and 8 respectively (142).

Miscible and Immiscible Blends. Polymer blending offers an alternative,simple, unique, and economical approach for improving barrier properties forseveral applications (see POLYMER BLENDS). In general, the goal is to add small


amounts of a high barrier polymer (generally more expensive) to a selected ma-trix polymer (generally low cost) (40). High cost polymers such as PEN and LCPscan be blended with lower cost polymers such as PET to achieve a balance ofbarrier properties and cost. A common objective of blending is to attenuate the de-ficiencies while maintaining as much as possible the desirable properties of eachcomponent. Occasionally, synergistic effects result in blend properties better thanthose of the individual components (40). Reactive blending is also possible, forexample, with the mixing of different polyester polymers (transesterification), orpolyester and nylon (polyesteramide formation) (40).

In general, polymer blends can be broadly classified as homogenous or misci-ble blends and multiphase or immiscible blends. The permeability coefficient P ofmiscible blends as well as copolymers often follows an empirical semilogarithmicadditivity rule (40):

lnP = φ1 lnP1 + φ2 lnP2 (39)

where φi is volume fraction of the ith component and Pi is the component’s per-meability coefficient. This simple additivity rule is generally obeyed only if thereare no interactions between the components. Deviations from this rule can ei-ther be positive or negative depending on the nature and magnitude of inter-actions (40). For example, a miscible blend of styrene–acrylonitrile copolymers(SAN) containing 9% acrylonitrile (AN) and tetramethyl bisphenol A polycar-bonate (TMPC) shows negative deviations from the linear additivity rule, indi-cating strong polymer–polymer interactions. On the other hand, blends of SANcontaining 13.5 and 28% AN, and PMMA show positive deviation from the linearadditivity rule (40).

When immiscible polymers are blended, or when inorganic filler is added toa polymer matrix, it results in the formation of a dispersion of one component ina continuous matrix of the other. Figure 29 shows several schematic examples ofsuch systems (144). Immiscible polymer blends are far more common than miscibleblends (40). Barrier properties of an immiscible blend depend on the permeabilitiesof the individual components, their volume fractions, phase continuity, and theaspect ratio of the dispersed (or discontinuous) phase (40). The aspect ratio L/Wrefers to the shape of the particles in the dispersed phase. Spheres and cubes havean aspect ratio of 1, whereas platelets and rods have higher aspect ratios.

The presence of an impermeable dispersed phase lowers the permeabilityby increasing tortuosity. The Maxwell model can be used to calculate the perme-ability of a polymer blend, P, with impermeable spherical particles dispersed in acontinuous phase (40):

P = Pm(1 − φd

)1 + φd/2

(40)

where Pm is permeability of the continuous polymer phase and φd is volumefraction of the dispersed phase. Different models for permeation in a heteroge-neous medium wherein the dispersed phase is impermeable and represented by


(a) (b)

(c) (d)

Fig. 29. Examples of heterogeneous, immiscible blends. (a) Random spheres in a dis-persed phase, (b) aggregated spheres in a dispersed phase, (c) oriented platelets in a dis-persed phase, and (d) oriented rods in a dispersed phase (144).

different geometrical shapes have been reviewed (144). Several modifications tothe Maxwell model have been made to describe permeation behavior in plateletsand ellipsoid-shaped particles dispersed in a more permeable continuous phase(40). For example, oxygen barrier properties of blends of oriented PET and EVOHhave been successfully modeled with one such modification (Fricke model) (145).Robeson extended the Maxwell model by applying it to blends in which both thepolymers contribute to the continuous phase (40). This model considers the prac-tical implications of attempts to increase the barrier performance of a moderatebarrier polymer by adding small amounts of a high barrier polymer to it. Accord-ing to this model, a multilayer structure provides the highest barrier followed bya blend in which the high barrier polymer is the continuous phase (40).

Figure 30 shows the improvement in O2 barrier properties of PE–EVOHblends as a function of EVOH volume fraction. Blends of PET with MXD-6 nylon


10−2

10−1

100

101

102

103

0 0.1 0.2 0.3 0.4 0.5

EVOH volume fraction

a

b

Oxy

gen

perm

eabi

lity,

(cm

3�

mil)

/(10

0 in

.2�

day

� at

m)

Fig. 30. Effect of EVOH content on oxygen permeability of PE–EVOH blends that exhibit(a) discontinuous morphology and (b) co-continuous lamellar morphology (40).

(Mitsubishi Gas Chemical Co.) have shown significant reduction in O2 and CO2permeability coefficients relative to those of PET and are being investigated for usein plastic containers (40). PET–EVOH blends have been promoted (Kuraray Co.)for potential applications in the beverage industry (57). The transport propertiesof PET–LCP blends have also been studied, and Table 10 shows the reductionsin permeabilities of O2, N2, and CO2 that have been reported for blown filmscontaining 2, 10, and 30 wt% LCP (146).

Polymer Nanocomposites. Polymer nanocomposites are immiscible blendsmade by adding nanometer-size particles to barrier polymers. Since nanometer-size grains, fibers, and plates have dramatically increased surface area compared

Table 10. Gas Permeabilitiesa of LCP–PET blendsb

%LCPc in PET CO2 permeability O2 permeability N2 permeability

0 114.5 21.3 5.12 78.1 14.9 2.610 65.8 12.0 3.130 34.5 5.5 2.0

a(cm3·mil)/(100 in.2·day·atm).bRef. 40.cRodrun LC3000 LCP.


to conventional-size materials, the chemistry of nanosized materials is differentfrom other conventional materials. Polymers filled with nanometer-size particleshave significantly different properties than those filled with conventional inor-ganic materials (147). Properties of nanocomposites such as high tensile strengthcan be achieved by using higher conventional filler loading, but other proper-ties such as improved clarity cannot be duplicated by filled resins at any loading(148).

Polymer nanocomposites were developed in the late 1980s, and were firstcommercialized by Toyota, which used nanocomposite parts in one of its car mod-els for several years (57). Initial developments focussed on the use of nylon resinsand very fine smectite clay particles, with surface area of about 750 m2/g, asfillers (57). Ube Industries developed its first nylon nanocomposite in 1989 for anautomotive timing belt cover. They have also developed other nylon nanocompos-ites called nylon clay hybrids (57). More recently, novel nanocomposite nylon-6resins developed by Honeywell Engineered Applications and Solutions (Morris-town, N.J.) have been claimed to improve O2 and CO2 barrier properties by a factorof 3–4 (149). It has also been claimed that these nanocomposites double the heatresistance of nylon-6 and improve other mechanical properties by 30–50% (149).Others such as Nanocor, Inc. (150) are also actively developing nanocomposites forenhanced barrier performance. Commercial products are available for PP and flu-oropolymers. Some of the other potential candidate polymers for nanocompositesinclude polyesters, PS, and ethylene vinyl acetate copolymers (57). Several poly-mer nanocomposites, including amorphous nylon and EVOH matrixes, intendedfor high barrier applications are in various stages of development.

Oxygen-Scavenging Systems. Another strategy for improving the bar-rier properties of polymers is the introduction of reactive groups in the polymer.These groups can reduce the transmission of penetrants such as oxygen and wa-ter vapor and the term “active barrier” is often used to describe this approachto distinguish it from “passive barrier” packages that rely on reduced permeabil-ity to decrease gas transmission. By using an oxygen scavenger, which absorbsthe residual oxygen after packaging, quality changes of oxygen-sensitive foodscan be minimized (151–153). Although oxygen-scavenging technology is rapidlyevolving and encompasses a wide variety of chemistries, the majority of currentcommercial oxygen-scavenger packages probably still employ sachets that removeoxygen from the headspace via iron oxidation. Ageless (Mitsubishi Gas Chemi-cal Co., Japan) oxygen absorbers are the most commonly used sachets that areplaced inside the food package (57). If the initial oxygen concentration and oxy-gen permeability of the packaging polymer is known, then an oxygen scavengercan be chosen with a higher capacity than the theoretically needed capacity, andnear total absence of oxygen can be maintained during the expected shelf life ofthe product (151). Other iron-based oxygen-scavenger sachets are ATCO (StandaInd., France), Freshilizer (Toppan Printing Co., Japan), Vitalon (Toagosei Chem.Ind., Japan), Freshpax (Multisorb Technologies Inc., U.S.), and Sanso-cut (FinetecCo., Japan) (151).

An alternative to sachets is to incorporate the oxygen scavenger into thebarrier polymer structure itself (151). An example of this strategy is Oxbar(Crown Cork and Seal, Philadelphia, Pa.), which involves cobalt-catalyzedoxidation of nylon MXD-6 polymer used in multilayer PET bottles for packaging of


0

0.05

0.1

0.15

0.2

0 100 200 300

Time, days

No scavenger

(a) 50 ppm Co

(b) 200 ppm Co

Bot

tle w

all o

xyge

n pe

rmea

nce,

cm

3 /(1

00 in

.2 � da

y �

atm

)

Fig. 31. Reduction in oxygen transmission rates due to oxygen scavenging in blends ofPET and nylon MXD-6 (157). The measurements were made at 23◦C and 50% RH. (a) 4 wt%MXD-6 and 50 ppm cobalt (as metal); (b) 4 wt% MXD-6 and 200 ppm cobalt (as metal). Forcomparison, the dotted line indicates oxygen permeance in the absence of cobalt catalyst(no scavenging).

beverages. Another example of a polymer-based absorber is Amosorb (BP AmocoChemicals), which can reportedly be incorporated into various rigid and flexi-ble packaging structures (151). This is a rapidly evolving field and several othercompanies have also introduced or announced oxygen-scavenging resins. Amongthose active in this area are EVAL Co., U.S., Kuraray Co., Japan and DarexCo., U.S., with DarEval (154); Owens-Illinois/Continental PET Technologies withCPTX-312; Honeywell International, Inc. with scavenging nylon resin (155); Cry-ovac with OS 1000; and Chevron Phillips Chemical Co. with their OSP scavengers(156). While the speed and capacity of oxygen-scavenging films are considerablylower than iron-based oxygen-scavenger sachets (151), oxygen ingress into thepackage can be reduced to very low levels with appropriately designed scavengerstructures. Figure 31 presents an example of the reduction in oxygen transmis-sion rate due to oxygen scavenging in blends of PET and nylon MXD-6 (157).Scavengers can also be incorporated into the liner of bottle closures where theycan significantly reduce the ingress of oxygen into the package through the closureliner (158).


A recent study (159) has shown that incorporation of inorganic fillers as wellas reactive groups in barrier polymer films can significantly improve the shelf lifeof packages. According to that theory, the presence of immobile reactive groupsdramatically increases the time lag (ie, the time required to achieve steady-statepermeability), but do not affect steady-state transport of penetrants across barrierpolymer films. For example, in LDPE and PVDC films containing 10% flakes ofmica or clay and linolenic acid as the oxygen-scavenging species, the time lag forpermeation of O2 dramatically increased by about 3 orders of magnitude to 40 hand 3 years, respectively (159).

Conclusions

Barrier polymers are widely used in food, beverage, and other packaging indus-tries. Some of the their advantages over traditional packaging materials such asglass, paper, and metals are flexibility, light weight, toughness, versatility, andprintability. The selection of a polymer for a particular packaging application de-pends on its barrier as well as other physical properties. A comparison of theseproperties and current as well as potential future applications for different typesof barrier polymers has been presented. Additionally, important factors govern-ing permeability (eg, penetrant size, polymer chemical structure, temperature,humidity) as well as ways to measure and predict permeability have also beendiscussed.

Unlike glass and metals, no polymer offers an infinite gas barrier. Despitethis limitation, monolayer polymer structures in many instances satisfy the bar-rier requirements for a package. In other situations, combinations of different poly-mers, or polymers with inorganic materials, in the form of multilayer structuresor blends, can provide cost-effective barrier for the intended shelf life of packagedproducts. As a result, plastic packaging is ubiquitous. Market pressures, however,drive the need for continual improvements in packaging materials. Hence, there isan ongoing interest in improving the barrier properties of polymers used in pack-aging, and the search for improved barrier polymers and structures is ongoing.Inorganic materials such as silicon and aluminum oxides and clays can be usedto significantly enhance gas barrier and other mechanical properties of polymers.One area of recent activity is polymer nanocomposites, which involves disper-sion of nanoscale barrier particles in a polymer matrix. A considerable amount ofresearch is also being focussed on techniques for developing thin layers of inor-ganic coatings on barrier polymer films and containers. In addition, work in thefield of oxygen-scavenging technologies (so-called active packaging systems) hasresulted in several developments. While advances in these newer barrier technolo-gies have recently opened up several commercial applications, with plastic beerpackaging being perhaps the most publicized, it is much too early to accuratelypredict their eventual success in the market place. Constructions with established“passive” barrier polymers currently dominate the barrier packaging market, andthese materials are expected to be commercially important for many years tocome.


BIBLIOGRAPHY

“Barrier Polymers” in EPST 1st ed., Vol. 1, pp. 65–95, by S. P. Nemphos, M. Salame, andS. Steingiser, Monsanto Co.; “Barrier Polymers” in EPSE 2nd ed., Vol. 4 pp. 176–192, byW. A. Combellick, Goodyear Tire & Rubber Co.

1. K. Ghosal and B. D. Freeman, Polym. Adv. Technol. 5, 673 (1994).2. R. C. Reid, J. M. Prausnitz, and B. E. Poling, The Properties of Gases and Liquids,

McGraw Hill, New York, 1988, p. 230.3. J. Crank, The Mathematics of Diffusion, 2nd ed., Clarendon Press, Oxford, 1975,

p. 414.4. C. E. Rogers, in J. Comyn, ed., Polymer Permeability, Elsevier Applied Science,

London, 1988, p. 11.5. C. C. McDowell, B. D. Freeman, and G. W. McNeely, Polymer 40, 3487 (1999).6. A. H. Windle, in Ref. 4, p. 75.7. A. R. Berens, J. Appl. Polym. Sci. 37, 901 (1989).8. F. Muller-Plathe, Acta Polym. 45, 259 (1994).9. W. J. Koros and M. W. Hellums, in J. I. Kroschwitz, ed., Encyclopedia of Polymer Sci-

ence and Engineering, John Wiley & Sons, Inc., New York, 1990, Supplement Volume,p. 724.

10. D. Hofman and co-workers, Polymer 37, 4773 (1996).11. J. R. Fried, J. Membr. Sci. 149, 115 (1998).12. A. Bondi, J. Phys. Chem. 68, 441 (1964).13. D. W. van Krevelen, Properties of Polymers: Their Correlation with Chemical Structure;

Their Numerical Estimation and Prediction from Additive Group Contributions, 3rded., Elsevier, Amsterdam, 1990, p. 875.

14. M. H. Cohen and D. Turnbull, J. Chem. Phys. 31, 1164 (1959).15. K. Ghosal, R. T. Chern, and B. D. Freeman, J. Polym. Sci., Polym. Phys. Ed. 33, 657

(1995).16. D. H. Weinkauf and D. R. Paul, in W. J. Koros, ed., Barrier Polymers and Structures

(ACS Symposium Series, Vol. 423), American Chemical Society, Washington, D.C.,1990, p. 60.

17. D. H. Weinkauf and D. R. Paul, J. Polym. Sci., Polym. Phys. Ed. 30, 837(1992).

18. W. M. Lee, Polym. Eng. Sci. 20, 65 (1980).19. A. Singh and W. J. Koros, in Permeation Processes in Barriers and Membranes: Differ-

ences and Similarities, Polymers, Laminations & Coatings Conference, San Francisco,Calif., 1998, p. 361.

20. A. J. Hill and co-workers, Eur. Polym. J. 32, 843 (1996).21. B. D. Freeman and A. J. Hill, in M. R. Tant and A. J. Hill, eds., Structure and Properties

of Glassy Polymers, Vol. 710, American Chemical Society, Washington, D.C., 1997,p. 306.

22. E. A. McGonigle and co-workers, Polymer 42, 2413 (2001).23. V. P. Shantarovich and co-workers, Macromolecules 33, 7453 (2000).24. J. E. Robertson, T. C. Ward, and A. J. Hill, Polymer 41, 6251 (2000).25. G. Dlubek and co-workers, J. Phys.: Condens. Matter 10, 10443 (1998).26. H. Yang and Y. C. Jean, Mater. Sci. Forum 255–257, 40 (1997).27. W. E. Brown, in W. E. Brown, ed., Plastics in Food Packaging, Marcel Dekker, Inc.,

New York, 1992, p. 292.28. R. M. Barrer, Trans. Faraday Soc. 38, 322 (1942).29. G. J. van Amerongen, J. Appl. Phys. 17, 972 (1946).30. R. M. Barrer and G. Skirrow, J. Polym. Sci. 3, 549 (1948).31. R. M. Barrer, Trans. Faraday Soc. 39, 48 (1943).


32. J. E. Leffler, J. Org. Chem. 20, 1202 (1955).33. W. J. Koros, in Ref. 16, p. 1.34. P. Meares, J. Amer. Chem. Soc. 76, 3415 (1954).35. T. K. Kwei and W. Arnheim, J. Chem. Phys. 37, 1900 (1962).36. J. Brandrup, E. H. Immergut, and E. A. Grulke, Polymer Handbook, 4th ed., Wiley-

Interscience, New York, 1999, p. VI–198.37. M. Salame, in K. M. Finlayson, ed., Plastic Film Technology, Vol. 1, Technomic Pub-

lishing Co., Inc., Lancaster, Pa., 1989, p. 132.38. R. R. Light and R. W. Seymour, Polym. Eng. Sci. 22, 857 (1982).39. A. S. Michaels, W. R. Vieth, and J. A. Barrie, J. Appl. Phys. 34(1), 13 (1963).40. P. M. Subramanian and I. G. Plotzker, in D. R. Paul and C. B. Bucknall, eds., Polymer

Blends, Vol. 2, John Wiley & Sons, Inc., New York, 2000, p. 359.41. A. S. Michaels, W. R. Vieth, and J. A. Barrie, J. Appl. Phys. 34(1), 1 (1963).42. A. C. Puleo, D. R. Paul, and P. K. Wong, Polymer 30, 1357 (1989).43. G. E. Serad and co-workers, Polymer 42, 6929 (2001).44. H. D. Weigmann and co-workers, Text. Res. J. 46, 574 (1976).45. H. Jameel, J. Waldman, and L. Rebenfeld, J. Appl. Polym. Sci. 26, 1795 (1981).46. H. Jameel, H. D. Noether, and L. Rebenfeld, J. Appl. Polym. Sci. 27, 773

(1982).47. A. S. Michaels and R. B. J. Parker, J. Polym. Sci. 61, 53 (1959).48. N. Qureshi and co-workers, J. Polym. Sci., Polym. Phys. Ed. 38, 1679 (2000).49. D. J. Sekelik and co-workers, J. Polym. Sci., Polym. Phys. Ed. 37, 847 (1999).50. H. D. Weigmann and co-workers, Text. Res. J. 47, 745 (1977).51. P. H. Hermans and P. Platzek, Kolloid Z. 88, 68 (1939).52. M. R. Tant and co-workers, Polym. Mater. Sci. Eng. 81, 374 (1999).53. R. Shastri, H. C. Roehrs, C. N. Brown, and S. E. Dollinger, in Ref. 16,

p. 239.54. D. R. Paul, Ber. Bunsenges. Phys. Chem. 83, 294 (1979).55. R. J. Hernandez, S. E. M. Selke, and J. D. Culter, Plastics Packaging: Properties, Pro-

cessing, Applications, and Regulations, Hanser Gardner Publications, Inc., Cincinnati,Ohio, 2000, p. 313.

56. R. J. Hernandez, J. Food Eng. 22, 495 (1994).57. G. Strupinsky and A. L. Brody, in A Twenty-Year Retrospective on Plastics: Oxygen

Barrier Packaging Materials, Polymers, Laminations & Coatings Conference, SanFrancisco, Calif., 1998, p. 119.

58. W. R. Vieth, Diffusion in and through Polymers: Principles and Applications, OxfordUniversity Press, New York, 1991, p. 20.

59. H. L. Frisch, J. Phys. Chem. 62, 401 (1958).60. K. D. Ziegel, H. K. Frensdorff, and D. E. Blair, J. Polym. Sci. 7, 809 (1969).61. Annual Book of ASTM Standards, Standard D3985, Vol. 15.09, American Society for

Testing and Materials, West Conshocken, Pa., 1994, p. 542.62. Ox-Tran® Oxygen Transmission Rate Measurement System, MOCON, Inc., Minneapo-

lis, Minn., 1998.63. P. DeLassus, in J. I. Kroschwitz, ed., Encyclopedia of Polymer Science and Technology,

Vol. 3, John Wiley & Sons, Inc., New York, 1992, p. 931.64. Annual Book of ASTM Standards, Standard F1249, Vol. 15.09, American Society for

Testing and Materials, West Conshocken, Pa., 1994, p. 1051.65. C. D. Barr, J. R. Giacin, and R. J. Hernandez, Packag. Technol. Sci. 13, 157 (2000).66. C. C. McDowell, D. T. Coker, and B. D. Freeman, Rev. Sci. Instr. 69, 2510

(1998).67. P. Hernandez-Munoz, R. Gavara, and R. J. Hernandez, J. Membr. Sci. 154, 195

(1999).


68. M. Moaddeb and W. J. Koros, J. Appl. Polym. Sci. 57, 687 (1995).69. B. I. Chaudhary and A. I. Johns, J. Cell. Plast. 34, 312 (1998).70. O. Pfannschmidt and W. Michaeli, in Determination of the Solubility and Diffusivity

of Gases in Polymers by using a High Pressure Magnet-Suspension-Balance, An-nual Technical Conference of the Society of Plastics (ANTEC), Atlanta, Ga., 1998,p. 1918.

71. S. N. Dhoot and co-workers, J. Polym. Sci., Polym. Phys. Ed. 39, 1160 (2001).72. M. Salame, Polym. Eng. Sci. 26, 1543 (1986).73. J. Y. Park and D. R. Paul, J. Membr. Sci. 125, 23 (1997).74. R. Franz, Packag. Technol. Sci. 6, 91 (1993).75. S. C. Fayoux, A. Seuvre, and A. J. Voilley, Packag. Technol. Sci. 10, 145

(1997).76. G. Strandburg, P. T. DeLassus, and B. A. Howell, in S. J. Risch and J. H. Hotchkiss,

eds., Food and Packaging Interactions II, Vol. 473, American Chemical Society,Washington, D.C., 1991, p. 133.

77. R. M. Barrer and G. J. Skirrow, J. Appl. Polym. Sci. 3, 564 (1948).78. S. A. Stern, J. T. Mullhaupt, and P. J. Gareis, AIChE J. 15, 64 (1969).79. S. Li and J. S. Paik, Trans. ASAE 39, 1013 (1996).80. J. S. Paik and M. A. Tigani, J. Agric. Food Chem. 41, 806 (1993).81. B. D. Freeman and I. Pinnau, in B. D. Freeman and I. Pinnau, eds., Polymeric Materials

for Gas Separations: Chemistry and Materials Science, American Chemical Society,Washington, D.C., 1999, p. 1.

82. R. C. Mason and co-workers, Tappi J. 75(6), 163 (1992).83. C. S. Coughlin, K. A. Mauritz, and R. F. Storey, Macromolecules 24, 1526

(1991).84. T. H. Begley, Food Addit. Contam. 14, 545 (1997).85. O. Piringer and co-workers, J. Agric. Food Chem. 46, 1532 (1998).86. A. O’Brien, A. Goodson, and I. Cooper, Food Addit. Contam. 16, 367 (1999).87. W. Limm and H. C. Hollifield, Food Addit. Contam. 13, 949 (1996).88. W. A. Jenkins and J. P. Harrington, in W. A. Jenkins and J. P. Harrington, eds.,

Packaging Foods with Plastics, Technomic Publishing Co., Lancaster, Pa., 1991,p. 46.

89. Permeability and Other Film Properties of Plastics and Elastomers, Plastics DesignLibrary, Norwich, N.Y., 1995, p. 143.

90. BLOX TM Adhesion and Barrier Resins: Product Information, The Dow Chemical Com-pany, May 2001.

91. A. B. Robertson and K. R. Habermann, in M. Bakker, ed., The Wiley Ency-clopedia of Packaging Technology, John Wiley & Sons, Inc., New York, 1986,p. 311.

92. J. Quezada-Gallo, F. Debeaufort, and A. Voilley, in S. J. Risch, ed., Food Packag-ing: Testing Methods and Applications (ACS Symposium Series, Vol. 753), AmericanChemical Society, Washington, D.C., 1999, p. 130.

93. Ref. 89, p. 706.94. V. T. Stannett, Polym. Eng. Sci. 18, 1129 (1978).95. R. J. Ashley, in J. Comyn, ed., Polymer Permeability, Elsevier Applied Science Pub-

lishers Ltd., Essex, England, 1985, Chapt. “7”.96. M. Salame, in Ref. 91, p. 48.97. S. Pauly, in J. Brandrup and E. H. Immergut, eds., Polymer Handbook, 3rd ed., Wiley-

Interscience, New York, 1989, p. IV–435.98. P. T. DeLassus, in J. I. Gray, B. R. Harte, and J. Miltz, eds., Food

Product-Package Compatibility, Technomic Publishing Co., Lancaster, Pa., 1987,p. 229.


99. J. Landois-Garza and J. H. Hotchkiss, in J. H. Hotchkiss, ed., Food and Packag-ing Interactions (ACS Symposium Series, Vol. 365), American Chemical Society,Washington, D.C., 1988, p. 42.

100. M. G. R. Zobel, in Aroma Barrier Properties of Coextruded Films, Proceedings of 4thAnnual International Conference on Coextrusion Markets and Technology, SchotlandBusiness Research, Princeton, N.J., 1984, p. 75.

101. G. L. Robertson, in H. A. Hughes, ed., Food Packaging: Principles and Practice, Vol.6, Marcel Dekker, Inc., New York, 1993, p. 9.

102. R. Catala and R. Gavara, Food Sci. Tech. Int. 2, 281 (1996).103. Ref. 27, p. 132.104. Ref. 55, p. 89.105. Selar PA Processing Guide, E. I. du Pont de Nemours & Co., Inc., Wilmington, Del.,

2001.106. P. M. Morse, Chem. Eng. News 8 (Nov. 10, 1997).107. T. M. McGee and A. S. Jones, in The Effect of Processing Parameters on Physical

Properties of PET/PEN Blends for Bottle Applications, 11th Annual High PerformanceBlow Molding Conference: Technical Innovations in Blow Molding, Cleveland, Ohio,1995, p. 89.

108. C. C. McDowell and co-workers, J. Polym. Sci., Polym. Phys. Ed. 36, 2981(1998).

109. G. R. Cantrell and co-workers, in C. Carfagna, ed., Liquid Crystalline Polymers, Perg-amon Press, Oxford, 1994, p. 233.

110. N. R. Miranda and co-workers, J. Membr. Sci. 94, 67 (1994).111. R. Lusignea, Tappi J. 80, 205 (1997).112. R. Lusignea, in Flexible Multilayer Packaging with Oriented LCP Barrier Layer,

Polymers, Laminations & Coatings Conference, San Francisco, Calif., 1998,p. 889.

113. D. J. Brennan and co-workers, Macromolecules 29, 3707 (1996).114. K. Ghosal and co-workers, Macromolecules 29, 4360 (1996).115. M. R. Pixton and D. R. Paul, in D. R. Paul and Y. P. Yampolskii, eds., Polymeric Gas

Separation Membranes, CRC Press: Boca Raton, Fla., 1994, p. 83.116. C. H. Silvis, Trends Polym. Sci. 5(3), 75 (1997).117. T. Glass, H. Pham, and M. Winkler, in New Thermoplastic Adhesive and Barrier

Resins, ANTEC 2000: Plastics, The Magical Solution, Orlando, Fla., 2000.118. W. Erickson, in New Coextruded PCTFE-Based Moisture Barrier Films for High Per-

formance Packaging, Polymers, Laminations & Coatings Conference, San Francisco,Calif., 1998, p. 355.

119. Aclar Fluoropolymer Films: Product Information, Honeywell International,2001.

120. P. T. DeLassus, D. L. Clarke, and T. Cosse, Mod. Plast. 86 (Jan. 1983).121. Ref. 27, p. 125.122. E. Werner, S. Janocha, M. Hopper, and K. Mackenzie, in J. I. Kroschwitz, ed., Ency-

clopedia of Polymer Science and Technology, Vol. 12, Wiley-Interscience, New York,1992, p. 193.

123. N. Whiteman, P. T. DeLassus, and J. Gunderson, in TAPPI Proceedings of PLC Con-ference, 2001.

124. Ref. 27, p. 105.125. J. L. Throne, in Ref. 91, p. 529.126. A. R. Berens, Makromol. Chem., Macromol. Symp. 29, 95 (1989).127. J. H. Levy, in Nylon 6 Barrier Coextrusions: A Cost-effective Packaging Route,

Polymers, Laminations & Coatings Conference, San Francisco, Calif., 1998,p. 163.


128. Ref. 88, p. 11.129. MXD-6 Nylon Polyamide Property Brochure, Mitsubishi Gas Chemical Co., Inc.,

2001.130. D. A. Abramowicz and P. J. Heyes, in Comparing the Performance of Barrier-Coated

versus Multilayer PET Containers for Oxygen Sensitive Products, Nova-Pack Amer-icas 2000, Orlando, Fla., 2000.

131. T. Hart, in New Effective Barrier Coating for Plastic Containers, Bev-Pak Americas’97, Directions 21, Inc., 1997.

132. D. S. Finch and co-workers, Packag. Technol. Sci. 9, 73 (1996).133. N. Inagaki, in S. J. Clarson, J. J. Fitzgerald, M. J. Owen, and S. D. Smith, eds., Silicones

and Silicone-Modified Materials, ACS Symposium Series, Vol. 729, Oxford UniversityPress, Washington, D.C., 2000, p. 544.

134. Ref. 55, p. 217.135. Ref. 101, p. 111.136. Ref. 88, p. 195.137. A. S. Da Silva Sobrinho, G. Czeremuszkin, M. Latreche, and M. R. Wertheimer, in W.

W. Lee, R. d’Agostino, and M. R. Wertheimer, eds., Plasma Deposition and Treatmentof Polymers, Materials Research Society Symposium Proceedings, Vol. 544, MaterialsResearch Society, Boston, 1998, p. 245.

138. K. H. Haas and co-workers, Surf. Coat. Tech. 111, 72 (1999).139. M. Hoffmann and co-workers, J. Sol Gel Sci. Tech. 1/2, 141 (1998).140. H. Langowski, Kunstoffe 90(6), 88 (2000).141. Packaging World 2 (May 2000).142. I. Bucklow and P. Butler, Materials World 14 (Aug. 2000).143. Packaging World 2 (June 1999).144. R. M. Barrer, in J. Crank and G. S. Park, eds., Diffusion in Polymers, Academic Press,

New York, 1968, p. 165.145. K. M. Kit, J. M. Schultz, and R. Gohil, Polym. Eng. Sci. 35, 680 (1995).146. O. Motta and co-workers, Polymer 37, 2373 (1996).147. E. Petrovicova and co-workers, J. Appl. Polym. Sci. 78, 2272 (2000).148. Polymer Nanocomposites, Business Communications Co., Inc., Apr. 2000.149. Packaging World (Nov. 6, 2000).150. K. Kamena, in Emerging Nanocomposite Technologies for Barrier and Thermal

Improvements in PET Containers, Nova-Pack Americas ’99, Orlando, Fla., 1999,p. 1.

151. L. Vermeiren and co-workers, Trends Food Sci. Tech. 10, 77 (1999).152. W. Scholl, in Active Barrier Packaging: A Technology for the 21st Cen-

tury, Polymers, Laminations & Coatings Conference, Atlanta, Ga., 1999,p. 671.

153. M. L. Rooney, in M. L. Rooney, ed., Active Food Packaging, 1st ed., Blackie Academic& Professional, Glasgow, U.K., 1995, p. 74.

154. S. Lambert, in New EVOH-Based Oxygen Scavenging Resins for Use inCo-Injected PET Bottles, Nova-Pack Europe 2000, Neuss, Germany, 2000,p. 163.

155. E. P. Socci, M. K. Akkapeddi, and D. C. Worley, in New High Bar-rier, Oxygen Scavenging Polyamides for Packaging Applications, ANTEC,2001.

156. B. D. Rodgers, in New High Capacity Oxygen Scavenging Polymer ForUse in Flexible Packaging, Flex-Pak Europe 2000, Amsterdam, Nov. 2000,p. 61.

157. U. S. Pat. 5,639,815 (1997), M. A. Cochran and co-workers.

Vol. 5 BIODEGRADABLE POLYMERS, MEDICAL APPLICATIONS 263

158. F. N. Teumac, B. A. Ross, and M. R. Rassouli, MBAA Tech. Q. 27, 122(1990).

159. C. Yang, E. Nuxoll, and E. Cussler, AIChE J. 47, 295 (2001).

SUSHIL N. DHOOT

University of Texas at AustinBENNY D. FREEMAN

University of Texas at AustinMARK E. STEWART

Eastman Chemical Company

Date post:	25-Mar-2018
Category:	Documents
Upload:	trinhkhue
View:	232 times
Download:	7 times

'Barrier Polymers'. In: Encyclopedia of Polymer Science ...nguyen.hong.hai.free.fr/EBOOKS/SCIENCE...

Documents