Yam, Kit L. (ed.). The Wiley encyclopedia of packaging technology

Подождите немного. Документ загружается.

Table 1. Typical Properties of 3-mm PETG Co

polyester Sheet

Units

Typical Value

Property

Conditions

ASTM Method

U.S. Customary

General

Density

231C (731F)

D1505

kg/m

g/cm

1270

1.27

Water absorption

231C (73

1F), 24-h

immersion

D570

0.2

Mechanical

Tensile stress at yield

50 mm/min (2 in./min)

D638

MPa

psi

7700

Tensile stress at break

50 mm/min (2 in./min)

D638

MPa

psi

3800

Elongation at yield

50 mm/min (2 in./min)

D638

4.8

Elongation at break

50 mm/min (2 in./min)

D638

Tensile modulus

5.0 mm/min (0.2 in./min)

D638

MPa

psi

2200

3.2

Flexural modulus

1.27 mm/min (0.05 in./min)

D790

MPa

psi

2100

3.1

Flexural strength

1.27 mm/min (0.05 in./min)

D790

MPa

psi

11,200

Rockwell hardness

—

D785

R scale

115

Izod impact strength, notched

231C (73

1F)

D256

J/m

ft . lbf/in.

1.7

01C (32

1F)

D256

J/m

ft . lbf/in.

1.2

301C(

221F)

D256

J/m

ft . lbf/in.

0.7

Impact strength, unnotched

231

C (731F)

D4812

J/m

ft . lbf/in.

01C (32

1F)

D4812

J/m

ft . lbf/in.

301C(

221F)

D4812

J/m

ft . lbf/in.

Impact resistance—puncture,

energy at maximum load

231C (73

1F)

D3763

ft . lbf

01C (32

1F)

D3763

atJ

ft . lbf

101C (14

1F)

D3763

ft . lbf

201C(

41F)

D3763

ft . lbf

30

1C(221F)

D3763

ft . lbf

Thermal

Deﬂection temperature

0.455 MPa (66 psi)

D648

1F7

4 1

1.82 MPa (264 psi)

D648

1F7

0 1

Vicat softening temperature

1 kg

D1525

1F8

3 1

Underwriters Laboratory (UL)

ﬂammability classiﬁcation

—

UL94

—

Flammability (France)

—

NFP 92501

—

Flammability (Germany)

—

DIN4102, Part 1

—

Flammability (Great Britian)

—

BS476, Part 7

—

Oxygen index

—

D2863

Coefﬁcient of linear thermal

expansion

—D

696

5

/1F

6.8

3.8

1098

Optical

5

/ 1C

Haze

—

D1003

Light transmission

Specular

D1003

Diffuse

D1003

Gloss

601 angle

D523

Units

159

Color

CIELAB, Illuminant D6500,

101 observer

E308

Units

Yellowness index

CIELAB, Illuminant D6500,

101 observer

D1925

Units

o1.5

Units

Refractive index,

—

D542

—

1.57

Electrical

Dielectric constant

1 kHz

D150

—

2.6

1 MHz

D150

—

2.4

Dissipation factor

1 kHz

D150

—

0.005

1 MHz

D150

—

0.02

Arc resistance

—

D495

158

Volume resistivity

—

D257

ohm . cm

Surface resistivity

—

D257

ohms/square ohms/square

Dielectric strength short time

500 V/s rate of rise

D149

kV/mm

V/mil

16.1

410

Properties reported here are typical of average

lots. Eastman makes no representation that the

material in any particular shipment will conform

exactly to the values given.

Unless noted otherwise, all tests are run at

231C (731F) and 50% rh, using specimens machined

from extruded sheeting with a thickness as indicated.

Nonbreak as deﬁned in ASTM D4812 using

specimens with thickness as indicated.

Not tested by UL but would expect to be the

same as Spectar copolyester 14471 [94V-2 for thicknesses

of 3.13–12.7 mm (0.123–0.5 in.) and 94HB for thickne

sses of

o3.13 mm (0.123 in.)].

Tests performed on 5-mm sheeting gave a rating

of M2.

Dripping and warpage of samples during testing

can cause erratic test results.

1099

metal or paint. Letters, designs, and trademarks are

examples of hot-stamped processes. PETG sheet can also

be printed on conventional processes such as letterpress,

offset lithography, rotogravure, silk screen, and stenciling.

Application of a protective lacquer is often applied over

the hot-stamped and printed areas (see the Decorating

articles).

PETG copolyester can be extruded in a variety of

thicknesses, coextruded with or without tie layers, and

formed or fabricated, to meet the needs of ever-increasing

consumer and industrial markets. The price, performance,

and availability of PETG copolyester make it a very

important material in the plastics market today.

BIBLIOGRAPHY

General References

References include MBS-80J, Eastar PETG Copolyester 6763.

DSS-429A, Physical Properties of Spectar Copolyester 14471

Plastic Sheet.

TRS-94F, Secondary Fabrication Techniques for Sheet Extruded

from Spectar Copolyester.

TRS-131, Extrusion of Plastic Sheet from Spectar Copolymer.

SHELF LIFE

KENNETH S. MARSH

Kenneth S. Marsh and

Associates, Ltd., Woodstock

Institute for Science in Service

to Humanity, Seneca, South

Carolina

DEFINITION OF SHELF LIFE

Shelf life is the time after production and packaging that a

product remains acceptable under deﬁned environmental

Table 2. Thermal Properties of 3-mm PETG Copolyester Sheet

Test Method

Property,

Units Typical Value ASTM ISO

Deﬂection temperature at 1.83-MPa (264-psi) ﬁber stress

1C 64

1F 147 D648 75

at 0.46-MPa (66-psi) ﬁber stress

1C 70

1F 158

Vicat softening point

1C 85 D1525 306

1F 185

Thermal conductivity

W/m K 0.19 C177 —

(Btu in.)/(h ft

1F) 1.3

Glass-transition temperature

1C 81

1F 178

Speciﬁc heat, kJ/(kg K) [Btu/ (b 1F)]

at 601C (1401F) 1.30 (0.31)

at 1001C (2121F) 1.76 (0.42)

at 1501C (3021F) 1.88 (0.45)

at 2001C (3921F) 1.97 (0.47)

at 2501C (4821F) 2.05 (0.49)

Coefﬁcient of linear thermal 5.1 10

D696 —

expansion, mm/(mm. 1C) [in./(in. 1 F)] (2.8 10

)

Flammability [3.2-mm (1/8-in.)-thick specimen] cm/min o2.5

in./min o1 D635 —

UL ﬂammability classiﬁcation

3.2 mm (0.125 in.) 94V-2

1.6 mm (0.0625 in.) 94HB UL94 —

Oxygen index % 24 D2863 —

Melt density

at 2001C, g/cm

1.19 — —

at 2501C, g/cm

0.98

Properties reported here are typical of average lots. Eastman makes no representation that the material in any particular shipment will conform exactly to

the values given.

Unless noted otherwise, all tests were run at 231C(731F) and 50% rh.

Glass-transition temperature and speciﬁc heat were determined by DSC.

UL ﬂammability classiﬁcations are formula-speciﬁc. Spectar copolyester 14471 has the ratings shown.

1100 SHELF LIFE

conditions. It is a function of the product, the package, and

the environment through which the product is trans-

ported, stored, and sold. Each of these inﬂuencing factors

will be discussed below.

The shelf life for any given product can vary greatly,

depending upon the formulation, the packaging system,

and the distribution environment. In general, use of

preservatives, increased barrier packaging, and colder

distribution environments will extend shelf lives. These

options typically also increase costs. The ideal shelf life

from the company perspective is one that matches the

distribution and use of the product to the company’s

inventory. The ideal shelf life from a consumer point of

view allows them to fully use the product. Speciﬁcation for

overly lengthy shelf lives usually increases cost in terms of

materials, whereas overly short shelf lives usually in-

creases cost in terms of wasted or discarded product or

in terms of increased liability. It is therefore prudent to

deﬁne a reasonable shelf life and to assure that the shelf

life speciﬁcation is met.

The principles of shelf-life testing hold for virtually any

product/package combination. Products that are packaged

in impermeable materials, such as glass or metal, degrade

primarily through mechanisms that are inherent in the

chemistry of the product. Since these mechanisms are

essentially product-dependent and vary with the product,

evaluations of the package would be unproductive. Nota-

ble exceptions to this are situations in which the package

allows or contributes to deterioration of the product. For

example, glass allows light transmission that can promote

oxidation reactions, but can be tinted to ﬁlter wavelengths

that promote these reactions. Metal cans can react with

products with either (a) the metal substrate itself (e.g.,

pinholes in tinplate) or (b) components of the can coating.

Improvements in can and coating technology are designed

to preclude these interactions. (For further discussion

refer to sections of this Encyclopedia on the appropriate

container.) Since the reactions of these impermeable con-

tainers are speciﬁc to the containers, further discussion of

packaging inﬂuences will be directed toward semiperme-

able and permeable materials.

FACTORS THAT INFLUENCE SHELF LIFE

Product

Products differ greatly in their susceptibility to degrada-

tion by various agents. Some products will become unac-

ceptable from a change in moisture content. For example,

a moisture gain in ready-to-eat breakfast cereals will

destroy the crisp texture; a moisture loss in an intrave-

nous product will increase the declared dosage per unit

volume. Snack items often are sensitive to oxidation of the

oils absorbed during frying and can become rancid. Snacks

also can be sensitive to moisture change. The mode of

failure of the product will have a direct inﬂuence on the

type of protection which is sought from the packaging

material. The match between product susceptibilities and

package protection will have an impact on the shelf life.

Once the mode of deterioration of the product is deter-

mined, acceptance criteria need to be deﬁned. Acceptance

criteria is the range of the critical component for which the

product is considered acceptable. This is often done by a

sensory panel. Ideally, sensory scores can be correlated to

analytical data such that result of that analysis can

provide the index for product quality.

The acceptability criteria for any product has a

direct inﬂuence on the measurement of shelf life. Some

products have a clear point of acceptability. Others have

complex deteriorations that make it difﬁcult to identify

a critical point. This can be further complicated by pro-

ducts that have multiple or interacting modes of failure.

In these cases, subjective decisions must be made to set

the acceptance criteria. It deserves mention that the

ultimate acceptability is that of the consumer. Concor-

dance between a standardized deﬁnition of acceptability

(as used within the company) and actual consumer accep-

tance is crucial to the commercial success of the product.

The determination of this concordance is more properly

within the domain of marketing research, but the degree

of correlation determines the validity of the research

effort.

The product that is tested for shelf life must be

speciﬁed. This may sound trivial, but reformulations and

ingredient cost reductions can alter product characteris-

tics without changing the product name. The result is that

assumptions may be made on the basis of past perfor-

mance which may not hold for the present product. There-

fore, the test label should include product name,

formulation (or reference to speciﬁc formulation in a lab

notebook), date of manufacture, date of test, conditions of

test, and principal researcher. The packaging system must

also be identiﬁed in terms of key factors as described

below.

Package

The package offers protection for the product against an

agent that degrades the product. For moisture-sensitive

products, this protection is in terms of the water vapor

transmission rate (WVTR); for oxygen-sensitive products

it is in terms of the oxygen transmission rate (OTR); and

so on. The mode of protection can change for the same

product with different packaging. For example, a snack

product that is sensitive to moisture gain (loses crispness)

and oxygen (becomes rancid) could be called ‘‘moisture-

sensitive’’ if the texture degrades before the rancidity

becomes objectionable. This same product, if packaged in

a sufﬁcient moisture barrier, would be identiﬁed as ‘‘oxy-

gen-sensitive’’ in terms of package criteria.

Permeation rates for chosen packaging materials need

to be determined for the agent in question. WVTR and

OTR are measurable by a variety of standard procedures

(see Testing, permeation and leakage). Permeation rates

for ﬂavors and other vapors can be determined by organic

permeation instruments (see Aroma barrier testing) as

well as by gas chromatographic, mass spectral, infrared,

and

other techniques.

The

size of the package also inﬂuences the shelf life. As

the package size increases, the surface-to-volume ratio

SHELF LIFE 1101

decreases. The result is that the amount of permeant that

comes through the package increases as a square function,

but the volume of product that absorbs this permeant

increases as a cube function. In other words, the permeant

is diluted with more product in a larger package. The

barrier requirements, therefore, decrease with the pack-

age size, all other factors remaining equal.

In summary, therefore, the packaging parameters that

must be speciﬁed for shelf life testing are the material

designation and source, the appropriate transmission

rate, the surface area, and the net weight of the enclosed

product.

Environment

Product distribution causes stress on the product as a

function of product sensitivities and all conditions experi-

enced by the product as it is carried through various

distribution networks. Actual conditions vary with where

it is shipped, how it is shipped, the seasons, warehouse

conditions, and so on. It is impossible to account for this

variety of conditions by using a single storage condition.

Therefore, the manufacturer of sensitive products (food,

pharmaceuticals, cosmetics, tobacco, chemicals) must de-

ﬁne the criteria that will be used for the shelf-life deter-

minations. Pharmaceutical manufacturers must package

for the worst-case situation, so packaging will be designed

to survive under the most severe conditions in the dis-

tribution chain. Food products are often packaged for less-

than-worst-case scenarios. Products that are not rotated

on the supermarket shelf (such as if a product is replen-

ished in front of existing stock) may fail, but this loss is

considered economically preferrable to increasing packa-

ging costs for the entire line to accommodate this situa-

tion. Conditions that allow 80% of the product to survive

distribution is not uncommon. Improved distribution and

product rotation can reduce the quantity of outlying

product. Once the representative distribution mode is

chosen, useful statistical indices can be derived which

represent this mode.

The shelf life impacts greatly on the distribution sys-

tem and vice versa. Products with short shelf lives require

a more rapid distribution system. For example, Frito Lay

supplies the nation’s grocery stores with snack items via

over 200 major distribution sites that in turn are supplied

from 32 production facilities (1). This allows potato chips

and other snacks, which are susceptible to both moisture

and oxidative degradation, to reach the consumer in a

fresh and crisp form. In contrast, Proctor and Gamble

Company distributes Pringles from one production

facility (Jackson, TN). A casual examination of Pringles

versus Frito’s potato chip packaging will describe for the

reader the consequences of the product/distribution stra-

tegies. Pringles uses a multilayered composite can struc-

ture with foil barrier as compared to the simpler ﬂexible

pouch in the Frito system. But when potato chips were

brought to the top of Mt. Everest, only Pringles made it

intact (2).

Refrigerated or frozen distribution is costly, but pro-

vides additional protection for the products. Additional

processing/packaging such as hot ﬁll, aseptic, or retort

packaging extend the shelf life and eliminate the need for

refrigeration. In Europe and Asia, where typical home

refrigerator and freezer space is more limited than in

typical U.S. households, retort and aseptic packaging are

far more important. The Brik-Pack, for example, was

developed in Europe to ﬁll the gap that the United States

ﬁlls with refrigeration.

The distribution environment also impacts shelf

life from a physical perspective. Lack of quality roads,

suspension systems on trucks, sway potential on railcars,

and so on, can inﬂuence product distribution substan-

tiallty. Potholes, for example, can lead to crushing, bruis-

ing, and impact damage. Packaging choice can partially

compensate against this damage, typically at a greater

cost. But poor infrastructure will inﬂuence either packa-

ging choice, product quality, or both. Two examples:

Crates can limit weight experienced by a truckload of

fresh produce on bumpy roads, and they were found to

reduce crushing damage from over 20% to 5% in

Sri Lanka; steel two-piece cans are used for bevereages

in Africa in locales in which aluminum two-piece cans

had inadequate strength to survive transport. Another

factor for determining acceptability or unacceptability of

products, especially food products, is attack by insects,

birds, rodents, and other animals. Packaging, storage, and

distribution environments all play a role in limiting

this damage. Such ‘‘shelf-life’’ limitations from physical

forces are typically addressed in distribution discussions

and components such as cushioning, and biological

attacks are typically addressed by materials choice and

warehouse and distribution conditions. Many people

in the developed world underestimate these losses (3),

which become increasingly important in less developed

countries in which food insecurity is a major problem.

Packaging serves to protect products, but also allows

products to be more competitive (4). Most of the following

discussion concentrates on chemical and microbiological

inﬂuences.

The measurement of shelf life is becoming increasingly

more important. The competitive nature of contemporary

business and increased energy costs have resulted in

renewed interest in reducing packaging costs. An abun-

dance of new packaging materials facilitates the choices,

but also adds to the complexity of the choice. In addition to

the above, governmental requirements for open dating,

required for pharmaceuticals and some foods (5), necessi-

tate shelf-life testing. Open dating is the addition of a date

on the package which indicates by what date the product

should be consumed. This can take different forms as

shown below.

A brief synopsis of different types of open dating is in

order

. Dating of

products online has been the practice of

companies producing food, pharmaceutical, and other

sensitive products. Each company developed its own pro-

prietary coding system to enable identiﬁcation of produc-

tion dates and production facility when necessary. These

codes, which were designed to be unreadable to consumers

and competing producers, identify production dates when

necessary. Open dating changes this.

Various types of dating are possible. A direct conversion

of production codes into consumer readable form gives the

1102 SHELF LIFE

date of manufacture. A freshness date (‘‘use by’’) indicates

the date by which the product should be consumed to

assure maximum freshness. A pull date is the date after

which the product should no longer be sold. Milk and dairy

products use pull dates.

It must be noted that all of the above dates assume

storage under speciﬁc conditions. For example, dairy

products are assumed to be refrigerated. If conditions

are not held as anticipated, product quality will not be

reﬂected as implied in the dates.

Dating of products will beneﬁt the consumer, but also

has the potential of beneﬁting both manufacturer and

retailer. It would improve stock rotation (as mentioned

above), which could ultimately reduce the packaging

requirements necessitated by existence of old product.

TRADITIONAL APPROACHES TO SHELF-LIFE TESTING

Shipping Tests

One common procedure for testing shelf life is a shipping

test. Shipping tests are often used as an adjunct to other

shelf-life testing, but sometimes are considered a replace-

ment. This is a mistake, because shipping is only a

component of the testing environment. This type of testing

is characterized by uncontrolled and often unknowable

storage conditions, although recording sensors now exist

to measure conditions. A product is shipped to a ware-

house, stored for a speciﬁed time, and then returned to

headquarters (or R&D) and examined for acceptability.

Shipping tests are also used to check for product dete-

rioration from shock and vibration during shipment. At

least one company uses shipment through the Chicago

post ofﬁce as the ultimate test for product endurance.

Shock and vibration effects are important for assessing

product quality. They are abuse-dependent rather than

time-dependent, however, and are discussed separately

(see Testing, product fragility).

Shipping tests show product quality changes with time

under real-world situations. The speciﬁc experience of a

speciﬁc test may be representative of part of the distribu-

tion chain but rarely, if ever, represents all conditions that

the product will experience. The test carries the illusion of

being a representative real-world test. However, the chan-

ging conditions, as well as factors such as the difference in

product experience between shipment in carton versus

pallet quantities, will necessitate careful evaluation of the

results.

Storage Tests

Storage tests consist of (a) maintaining product in a static

facility and (b) evaluating product quality with time.

Storage conditions can be uncontrolled or controlled.

Warehouse storage is often characterized by uncontrolled

conditions. Storage conditions can be controlled through

the use of storage cabinets that maintain temperature

(and often humidity) at speciﬁed values. This results in a

test that can offer more reproducible data. Various condi-

tions can be tested, including accelerated conditions that

allow for more rapid results. A typical ambient condition

(for the United States as deﬁned by TAPPI) is 731F(231C)/

50% relative humidity (R.H.).

Accelerated conditions vary, but a typical condition

may be 391C (951F)/80% R.H. Accelerated storage speeds

product degradation (6). However, the factor that deﬁnes

this increase is related to the kinetics of the chemical

reactions which deﬁne the degradation. This varies with

product and with mode of degradation. Therefore, any

standard multiplication factor is meaningless.

Accelerated studies are useful if an initial evaluation is

performed at both ambient and accelerated conditions and

then compared. This comparison can be through actual

kinetic studies (such as Q

analysis described below) or

by shelf-life comparison. It is imperative to ensure that

discontinuities do not exist between ambient conditions

and accelerated conditions. If either the product or packa-

ging passes through a transition, the results of the

accelerated test will have little bearing on comparisons

with ambient. Transitions that affect packaging include

the glass-transition temperature (T

). For example, per-

meation values for polypropylene cannot be extrapolated

to frozen conditions due to a T

at 101C (141F). Those for

products include melting/freezing of any ingredient

(usually water or fats), recrystallizations, and so on. In

addition, activation energies must be considered, because

certain reactions will not occur below certain tempera-

tures. Extrapolation below these temperatures is there-

fore not recommended.

analysis involves testing products at varying tem-

peratures, as well as deﬁning the difference in reaction

rate for a 101C (181F) temperature increase. A Q

of 2

means that the reaction rate will double for each 101C

increase in temperature. This is based on the Arrhenius

relationship, which is discussed in any text on kinetics or

physical chemistry. The Arrhenius relationship is perhaps

an oversimpliﬁcation for the complex chemistries of most

products. A straight-line relationship with a minimum of

three points (over a limited temperature range) indicates

that Q

can provide a useful indicator for storage dura-

tion. Caution relating to transitions remains, and it also

deserves mention that the Q

itself is temperature-de-

pendent. However, products are usually studied under a

limited range of temperatures through which a constant

can be assumed.

Once the Q

is determined (and shown valid through a

linear Arrhenius relationship), the value can be used to

predict the time necessary to test a product to be equiva-

lent to the shelf life at ambient temperature. For example,

a product with a Q

equal to 2.4 can be stored for 10

weeks at 331C (911F) to obtain results similar to 24 weeks

storage at 231C (731F). (One can view these studies as a

test in which the conditions extended to the package are

exaggerated in an attempt to simulate the accelerated

passage of time. The underlying assumption is not always

valid, and the technique is only as good as that underlying

assumption.)

An alternative type of ‘‘accelerated’’ test is a study

in which the permeation rate of the packaging material

is ‘‘accelerated.’’ A product that is being tested for sensi-

tivity to moisture is packaged in a low-moisture/high-

oxygen barrier material (such as polyacrylonitrile or

SHELF LIFE 1103

polyacrylonitrile copolymers). With a known permeability,

area, and net weight, the amount of moisture permeating

the package can be determined. Product can be sampled

for acceptability until failure is reached. This deﬁnes the

amount of moisture transfer that can be tolerated. Calcu-

lations with higher moisture barrier materials can provide

performance criteria for the additional materials. A simi-

lar study can be performed for oxygen-sensitive products

by using a low-oxygen/high-moisture barrier material

(such as polyoleﬁns). In essence, this test purposefully

introduces a known amount of the critical agent to accel-

erate the time to degradation without changing the con-

ditions experienced by the product. A cautionary note is

important. The accelerated test described above includes

the inherent assumption that the changes in the product

occur faster than the transmission across the package. If

an accelerated test allows sufﬁcient permeation such that

product changes are slower than the inﬂux of permeant

(i.e., is diffusion dependant), the results will indicate a

shelf life that is longer than the extrapolation to a reason-

able barrier material. This is due to a critical permeant

level being obtained faster than the evidence of product

degradation.

COMPUTER MODELS FOR SHELF LIFE

The shipping and storage tests require that the choice of

packaging be made before the study. Chosen materials are

then submitted for evaluation with time as the test

proceeds. Successful completion of the test veriﬁes that

the chosen package will last at least as long as the chosen

test duration under the conditions of the test. Failure to

survive the test conditions demonstrates insufﬁcient pro-

tection. The deﬁciencies of this procedure are that failure

is not known until demonstrated, and overprotection is

not shown at all because testing usually terminates after

desired time is reached.

The use of a computer to simulate a storage test

provides a means to pretest materials before actual sto-

rage testing. By performing these calculations rapidly, as

only a computer can do, many iterations are possible

within a reasonable time, providing a powerful support

tool for packaging development. The simulations are

accomplished by separating the product and packaging

characteristics. The product is analyzed for changes that

occur on exposure to the shelf-life-limiting parameter,

which is identiﬁed as the mode of failure for the product.

The packaging material is analyzed for its barrier proper-

ties against that parameter. The computer is then used to

combine the protective aspects of the package with the

sensitive properties of the product. Since packaging, pro-

duct, and storage parameters are entered independently,

it is possible to test the effect of different packages or

conditions simply by entering the new variables.

Computer simulation yields the most cost-effective

packaging design. Later real-time storage studies serve

the purposes of legal requirements and conﬁrmation of

predicted results. Differences between predicted and ac-

tual results can be used as indices to check initial assump-

tions. For example, poor seals will shorten actual shelf life

achieved with an otherwise-adequate barrier material.

In this example, the false assumption is that the material

barrier value is equivalent to the package permeation (see

Testing, permeation and leakage).

Simulated studies extend beyond the one-package/

one-condition study. Parameters can be changed to affect

another study. A survey of potential new packag-

ing materials can be accomplished without need to retest

the product. Only the changes in material parameters

need be entered to test performance. This is especially

important when limited amounts of product material

are available, as in the product-development stage.

Changing storage conditions follow a similar proce-

dure, but the caution for passing transitions remains.

Computer modeling has been used successfully for foods

(7–9), over-the-counter products (10), and pharmaceuti-

cals (11).

Product Evaluation

The mode of failure for the product must be known prior to

using any simulated approach. This could be moisture

gain or loss, oxidation, CO

loss, ﬂavor loss, chemical

degradation, and so on. Once the critical factor is known,

testing is conducted to show product changes with change

in the critical factor. For example, a moisture-sensitive

product is evaluated for change in moisture content as the

relative humidity in the storage environment is varied.

This test is the moisture isotherm. The isotherm can be

completed in usually one to four weeks, depending on the

rapidity at which the samples equilibrate. For oxygen-

sensitive products, the testing consists of evaluating pro-

duct changes with oxygen uptake and can be performed by

inverse gas chromatography, respirometry, Warburg, To-

tox, or a variety of tests for degree of oxidation. For

carbonated beverages packaged in permeable bottles, the

test would be to measure carbonation loss (soda), or

oxidation (beer) or both.

The product testing should be performed at various

temperatures if temperature effects are to be modeled.

The evaluation of differences in reaction rates versus

temperature supplies important information concer-

ning the kinetics of the reaction and can be used to

investigate performance for different distribution envir-

onments. A minimum of three test temperatures are

recommended. The caution concerning transitions re-

mains valid.

Package Evaluation

For computer modeling purposes, the chief packaging

characteristics are barrier properties. Materials that do

not change signiﬁcantly with machining and ﬂexing can

be tested in ﬂat sheet form for water vapor transmission

rate (WVTR), oxygen transmission rate (OTR), or any

other permeant that is critical for the product. For materi-

als that change with handling (waxed glassine, foil, etc.,)

full package testing is recommended. The computer model

measures the change of permeant level in the package as a

function of the transmission rate. If the ﬂat sheet value is

used in a system that is poorly sealed, or has developed

1104 SHELF LIFE

fractures in handling, the predicted shelf life will be

erroneously long.

The package material characteristics of sealability,

printability, compatability, cost, mechanical strength, etc

remain the province of the packaging engineer and are not

usually incorporated into computer models. Nonbarrier

parameters are not included in this discussion of shelf life.

Those parameters that do affect shelf life, but are not in

the model (e.g., light transmission for oxygen-sensitive

products), must be considered in evaluating any computer

printouts.

Model Fundamentals

The internal environment of a package changes as per-

meants enter or leave the package. In terms of the

mathematics, components which enter the package from

the outside environment, and those which permeate from

the inside to the outside, are identical. Transmission will

always occur from the high concentration to the lower

concentration side of the material. Therefore, any discus-

sion or examples of permeation rates are valid for either

circumstance.

The transmission rate is standardized for package

area, ﬁlm thickness, time of measurement, and difference

of partial pressure of permeant across the package. In

general, the terms are amount of permeant multiplied by

ﬁlm thickness per package area per time per unit of

partial pressure differential. The measurement is also

made at a temperature that is implied but not usually

stated in the units. In use, the standardized transmission

rate is combined with area, thickness, and time in the

model to deﬁne how much permeant can enter or leave the

package. The effect on the product is a function of the net

weight of the product upon which the permeant acts. The

partial pressure differential across the package becomes

difﬁcult to measure, and it changes as the internal envir-

onment of the package changes.

The partial-pressure-differential change with per-

meation is the major reason for the use of computers in

shelf-life modeling. If this change did not occur, one

could simply multiply transmission rate by time and

area and divide by the thickness to obtain amount of

permeation for any time. This assumed straight-line

relationship overestimates transmission because the per-

meation process itself reduces the differential of partial

pressure across the ﬁlm and therefore reduces the trans-

mission. In addition to the driving-force change due to the

transmission of permeant, some of the permeant is ab-

sorbed by the product and further complicates the system.

In engineering terms, as the permeation process proceeds,

the driving force of this process (the partial pressure

differential) decreases. This change can be described by

a differential equation. The computer models incorporate

these effects and yield a representative manifestation of

the process.

Computer models are usually based on a relatively

simple mass transfer equation. Many models have

been developed that have not been reported in a

usable form in the literature. These include models on

oxidation-sensitive products, beverages, and products in

which product characteristics result in product-dependent

models.

The general models are based on a standard

differential equation that describes mass transfer across

a permeable barrier. This is shown in the following

equation (12):

 AðP

out

 P

where

t = time

W = change in weight of the critical component

k = permeability coefﬁcient for the packaging material

l = thickness of the packaging material

A= area of the package

out

= partial pressure of permeant outside the package

= partial pressure of permeant inside the package

This equation can be converted into a usable form by

substituting measurable values for the W and P

terms.

The P

out

term is readily obtainable: For oxygen, it is 0.21

atm (21.3 kPa); For moisture, it is the saturated vapor

pressure (13) times the RH of storage; for ﬂavors and

volatile components, it is essentially zero.

The P

term requires knowledge of the reaction of the

permeant with the product. In the case of moisture

change, this is expressed as the moisture isotherm that

describes change in product moisture versus storage

relative humidity. An expression that describes the water

activity over the product for any moisture value can be

used to deﬁne the P

term. (Water activity, Aw, is the

partial pressure of water above the product divided by the

saturated vapor pressure of water at the same tempera-

ture.) First, the isotherm is determined. Second, an ex-

pression is derived which describes Aw as a function of

moisture content. Finally, P

at any moisture content can

be described as the expression for Aw times the saturated

vapor pressure. (Note: Water activity is traditionally used.

However, this is only valid at the temperature used in the

isotherm study. The preferred procedure would study

moisture content as a function of vapor pressure directly,

instead of as a function of Aw. The model converts Aw into

pressure terms.) Models for permeants other than moist-

ure require analogous expressions to relate internal pres-

sure to product changes.

With a substituted expression for the P

term in

concentration terms rather than weight terms, the W

term needs to be expressed in concentration terms. An

expression is therefore needed which describes the change

in concentration as a function of weight change.

The above expressions are then substituted into the

equation, they are rearranged to combine variables onto

one side of the equation, and the new equation is inte-

grated to form the model. The limits of integration are (a)

permeant concentration from initial to critical values and

(b) time from initial to shelf life (time to reach critical

amount of permeant).

SHELF LIFE 1105

The various models described in the literature use

approaches related to the above procedure. Simplifying

assumptions (such as linearity in the isotherm) are used to

develop simpler models. These models can be used to

advantage as long as the assumptions are kept in mind.

The fewer the restrictions used in deriving the expres-

sions, the more universal the model. However, the mathe-

matics become increasingly complex.

The models do not replace the need for actual storage

studies. Storage is still necessary to verify product

changes, ﬁnd unusual effects, and, in some cases, meet

compliance criteria with government regulations. The use

of the models serves to (a) reduce the time and effort

necessary to identify optimal packaging protection and (b)

eliminate storage studies on materials with insufﬁcient

barrier properties to be reasonable candidates for the

product.

DISTRIBUTION MODELS

The above models are based on storage conditions with a

single temperature and relative humidity. Computers can

be used to model changing conditions, either predicted or

measured, to more accurately reﬂect real-world conditions

(14). One such model applied average environmental

conditions for key cities in the United States and abroad

to predict warehoused product performance in these dif-

ferent locations. Not surprisingly, products distributed in

colder climates exhibited longer shelf lives than similar

products warehoused in warmer climates. The study also

showed that products manufactured in cooler seasons

exhibited longer shelf life than did the same products

produced during the summer months. It was found that

higher temperatures experienced earlier in the distribu-

tion cycle had a more pronounced impact than did the

same temperatures experienced later. Such information

could be used to either modify the packaging or the

distribution environment (such as storage time or ware-

house conditions) or even the shelf-life requirement

(shorter for warmer climates, longer for cooler ones) for

economic advantage.

The development of sensors and RFID tags offers

additional capabilities for shelf-life modeling. Sensors

currently exist to measure real-world conditions and

transmitthesedatatoawarehousecomputer.Computers

could be programmed to (a) recalculate remaining shelf

lif e for stored products base d upon thei r actual distribu-

tion and storage experience and (b) allow product pulls

based upon remaining shelf life rather than actual

storagetime(FirstIn,FirstOut).Inthismanner,pro-

ducts experience temperature abuse could be moved

out more rapidly (or pulled if unacceptable and

thereby improving quality and reducing consumer com-

plaints). At the time of this writing, RFID technology and

computer programs exist to initiate this proposed pro-

gram (15), but the sesnsors are currently considered too

costly for general product distribution. This scenario

could easily change by the time this encyclopedia is

published.

OTHER MODELS

Modeling of permeation through packaging ﬁlms can be

extended to evaluate more complicated scenarios than

storage of a product in an ambient atmosphere. As long

as permeation of gases and vapors remains the key

shelf-life determining parameter, modeling of mass trans-

fer is applicable. Modiﬁed atmosphere packaging (MAP)

(see Modiﬁed atmosphere packaging) utilizes a gas mix-

ture to extend the life of the product. The gas mixtures

are composed of gases with different permeation rates

and different concentrations in and out of the pacakge.

Computer models can predict how the concentrations

will change with time. Shelf life can be directly determined

if the shelf life relates directly to a gas concentration.

For example, if an oxygen-sensitive product would oxi-

dize rapidly above a critical oxygen concentration,

the time required for the oxygen concentration within

the package to attain that level would determine the

shelf life.

Another computer model that relates to shelf life

involves controlled atmosphere packaging (CAP). CAP

implies that the atmosphere is controlled during the life

of the product and is typically employed in warehouse

situations, particularly with fresh produce. CAP can also

be used for a consumer package in which a combination of

permeation characteristics of a ﬁlm and respiration rates

of a fresh product combine to create the desirable packa-

ging environment. As the produce respires (takes in

oxygen and expels carbon dioxide), the concentration of

oxygen within the package will be depleted, and that of

carbon dioxide will be enhanced. An impermeable

package would quickly reach conditions of low oxygen

and high carbon dioxide that would asphyxiate most

produce. A correctly chosen permeable ﬁlm, however,

can result in a packaging atmosphere that slows

senescence without killing the product, thereby enhancing

shelf life. Computer models in this case are used to

calculate the gas concentrations that result from the

dual respiration and permeation effects. The concentra-

tions are then compared with ideal CAP conditions for the

particular product and thereby used to extend the shelf

life of the product.

Various models have been developed to predict shelf life

of products that are limited by microbial growth. Exam-

ples of some of these models are referenced in the biblio-

graphy (16–23).

BIBLIOGRAPHY

1. ‘‘21 Years of Manufacturing Excellence,’’ Food Engineering,

September 10, 2003.

2. J. Fassl, ‘‘Editor’s Note: Keeping It Real, Keeping It Simple in

Food Packaging,’’ Food Engineering, April 15, 2003.

3. K. S. Marsh, ‘‘Acceptance of Unacceptable Losses,’’ Food

Technology, ‘‘Back Page’’ 57(3), 96. 2003.

4. K. S. Marsh, ‘‘Value Added Packaging Can Make You More

Competitive: Packaging & Economic Development,’’ Written

for and published by the United Nations, International Trade

Center, Geneva, Switzerland, 2005.

1106 SHELF LIFE

5. USDA Food Labeling Fact Sheet website, updated February 8,

2007, http://www.fsis.usda.gov/Fact_Sheets/Food_Product_

Dating/index.asp updated.

6. D. Olds, ‘‘Shelf-Life Predictions Using Temperature as an

Acceleration Factor,’’ Food Quality, 28, 000–000 (1996).

7. S. Gyeszly, ‘‘Shelf-Life Simulation Predicts Package’s ‘Total

Performance’,’’ Package Engineering 00, 70–73 (1980).

8. K. S. Marsh, ‘‘Computer Simulation Speeds Shelf Life Assess-

ment,’’ Package Engineering 00, 72 (1983).

9. K. S. Marsh, ‘‘Prediction of Packaged Product Shelf Life: Case

Histories of Shortened Time to Market,’’ in Proceedings from

the 1994 IoPP Technology of Packaging Conference, Institute

of Packaging Professionals, Herndon, VA, 1995.

10. K. S. Marsh, T. J. Ambrosio and D. K. Morton, ‘‘The Determi-

nation of Moisture Stability of a Dynamic System Under

Different Environmental Conditions,’’ in D. Henyon, ed.,

Food Packaging Technology, ASTM STP 1113, American

Society for Testing and Materials, Philadelphia, PA, 1991, p.

13.

11. K. S. Marsh, J. Bitner, P. David and A. Rao, ‘‘Shelf Life

Prediction Software ﬁnds Application with Ethical Drugs,’’

Packaging Technology & Science 12, 173 (1999).

12. K. S. Marsh, ‘‘Computer-Aided Shelf Life Prediction,’’ 1984

Polymers, Laminations and Coatings Conference—Book 1,

TAPPI Press, Atlanta, GA, 1984.

13. Handbook of Chemistry and Physics, CRC Press, Inc., West

Palm Beach, FL (any edition will have saturated vapor

pressure charts).

14. K. S. Marsh, ‘‘Computer Model Looks at the Environment to

Predict Shelf Life,’’ Food Engineering 00, 58 (1985).

15. K. S. Marsh, ‘‘Shelf Life Modeling in Real Time and Real

Conditions,’’ in SLIM 2006 Shelf Life International Meeting,

special issue of the Italian Journal of Food Technology,

Chiriotti Editori, Italy, 2007.

16. K. R. Anantha-Narayanan; A. Kumar, and G. R. Patil,

‘‘Kinetics of Various Deteriorative Changes During Storage

of UHT—Soy Beverage and Development of a Shelf-Life

Prediction Model,’’ Lebensmittel-Wissenschaft-und-Technolo-

gie 26(3), 191–197 (1993).

17. M. Ferrer-Gimenez, ‘‘[Determination of the Commercial

[Shelf]-Life of a Dehydrated Food.]’’, Alimentaria 169, 43–48

(1986).

18. S. S. Kurade and J. D. Baranowski, ‘‘Prediction of Shelf-Life of

Frozen Minced Fish in Terms of Oxidative Rancidity as

Measured by TBARS Number,’’ Journal of Food Science

52(2), 300–302, 311 (1987).

19. H. Einarsson, ‘‘Predicting the Shelf Life of Cod (Gadus

morhua) Fillets Stored in Air and Modiﬁed Atmosphere at

Temperatures Between 4 Degree and +16 Degree,’’ in

Quality Assurance in the Fish Industry, 1992, pp. 479–488.

20. B. R. Jorgensen, D. M. Gibson, and H. H. Huss, ‘‘Microbiolo-

gical Quality and Shelf Life Prediction of Chilled Fish,’’

International Journal of Food Microbiology 6(4), 295–307

(1988).

21. G. C. Mead, ‘‘Predictive Aspects of Spoilage in Fresh Poultry

Meat Including the Use of Biochemical Parameters and the

Application of Mathematical Models,’’ in Seventh European

Symposium on Poultry Meat Quality, 1985, pp. 171–186.

22. G. J. Newell, ‘‘A New Shelf-Life Failure Model,’’ Journal of

Food Protection 44(8),

580, 582 (1984).

23. M. G

. Geoule, Jr., and J. J. Kubala, ‘‘Statistical Models for

Shelf Life Failures,’’ Journal of Food Science 40(2), 404–409

(1975).

SHOCK

ROBERT M. FIEDLER , CPP

Robert Fiedler & Associates,

Minneapolis, Minnesota

The term shock describes a rapid change in something

over a very short period of time. In the packaging ﬁeld,

this could relate to mechanical movements, sudden stops,

or even rapid changes in environmental conditions such as

thermal shock, when temperatures change rapidly with

movement of products from inside to outside. We focus our

discussions here only on the mechanical shocks related to

the physical movement of products and packages.

The most common shocks experienced by packages and

products result from handling drops. It is not the fall that

hurts, but the sudden stop on impact with the solid ground.

The suddenness of the shock is described as the duration of

the impact. It is normally expressed in terms of millise-

conds (ms) or

1000

th of seconds. Typical packaged product

impact durations range from 2 to 50 ms from drops onto

hard ﬂoors. The duration of the shock is dependent pri-

marily on the amount of cushioning provided by the

packaging materials used to protect the product and to

some extent on the rigidity of the surface impacted.

The term cushioning is used to describe the effect of

slowing the rate of change of the shock by increasing the

time duration of the shock event. This is accomplished by

allowing the product to stop over a longer period of time by

beginning to stop it at a higher level above the point of

ﬁnal rest. Cushioning requires distance to provide the

added time. Other articles in this Encyclopedia address

the need of how to best design and shape the cushion to

effectively use the available stopping distance provided by

the cushion thickness to obtain maximum cushioning.

The rate of change of the speed of an item when it

impacts is called the acceleration rate (or deceleration

rate) of the impact. The greater the time duration of the

impact, the lower the acceleration rate of the shock for

impacts of the same speed. The impact acceleration rate is

expressed in terms of G values, which are the dimension-

less ratio between the acceleration in length per time

squared and the acceleration of gravity in the same

units—in effect, multiples of the acceleration rate of

gravity (1g) on earth. The gravitational attraction of the

earth is deﬁned as 386 in./(s s), 32 f

/(s s), or 9.8 m/(s s).

Measured impact acceleration levels as discussed de-

pend on the duration of the impact shock resulting from

the cushion’s effect. With no or very limited cushioning,

acceleration levels can easily reach 500–1000G. With

effective cushioning, acceleration levels can be reduced

to 15–100G. How much cushioning is provided depends on

the products’ need for protection and the expected height

from which the product will be dropped.

In simple terms, if an item is dropped from 30 in., it will

fall at a rate of 1G, due to the effect of gravity. To stop the

item at a rate of 1G would require an additional stopping

distance of 30 in., assuming that you could bring it to a

stop at a constant rate of 1G throughout the stopping

distance. Fortunately, most products can be stopped at a

SHOCK 1107